Fundamentals PDF

Fundamentals of
Vacuum Technology
FUNDAMENTALS
dV
S =
dt
D00 E 19.06.2001 21:35 Uhr Seite 1
General
Fundamentals of Vacuum Technology
D00.02 LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002
Fundamentals of
Vacuum Technology
revised and compiled by
Dr. Walter Umrath
with contributions from
Dr. Hermann Adam , Alfred Bolz, Hermann Boy, Heinz Dohmen,
Karl Gogol, Dr. Wolfgang Jorisch, Walter Mnning,
Dr. Hans-Jrgen Mundinger, Hans-Dieter Otten, Willi Scheer,
Helmut Seiger, Dr. Wolfgang Schwarz, Klaus Stepputat, Dieter Urban,
Heinz-Josef Wirtzfeld, Heinz-Joachim Zenker
Preface
A great deal has transpired since the final reprint of the previous edition of Fundamen-
tals of Vacuum Technology appeared in 1987. LEYBOLD has in the meantime introduced
a number of new developments in the field. These include the dry-running ALLex
chemicals pump, the COOLVAC-FIRST cryopump systems with quick regeneration fea-
ture, turbomolecular pumps with magnetic bearings, the A-Series vacuum gauges, the
TRANSPECTOR and XPR mass spectrometer transmitters, leak detectors in the UL
series, and the ECOTEC 500 leak detector for refrigerants and many other gases. Moreo-
ver, the present edition of the Fundamentals goes into much greater detail on some
topics. Among these are residual gas analyses at low pressures, measurement of low
pressures, pressure monitoring, open- and closed-loop pressure control, and leaks and
their detection. Included for the first time are the sections covering the devices used to
measure and control the application of coatings and uses for vacuum technology in the
coating process. Naturally LEYBOLDs Vacuum Technology Training Center at Cologne
was dependent on the invaluable support of numerous associates in collating the litera-
ture on hand and preparing new sections; I would like to expressly thank all those indi-
viduals at this juncture. A special word of appreciation is due the Communications
Department for its professional contribution in preparing this document for publishing.
Regrettably Dr. Hermann Adam, who at one time compiled the very first version of the
Fundamentals, did not live to see the publication of this edition. Though he had been
in retirement for many years, he nonetheless labored over the corrections and editing of
this current edition until shortly before his death.
I hope that this volume will enjoy a response as favorable as the previous version.
Dr. Walter Umrath
Cologne, August, 1998
D00 E 19.06.2001 21:35 Uhr Seite 2
Contents
D00
1. Vacuum physics Quantities,
their symbols, units of
measure and definitions . . .7
1.1 Basic terms and concepts
in vacuum technology . . . . . .7
1.2 Atmospheric air . . . . . . . . .10
1.3 Gas laws and models . . . . .11
1.3.1 Continuum theory . . . . . . . .11
1.3.2 Kinetic gas theory . . . . . . . .11
1.4 The pressure ranges in
vacuum technology and
their characterization . . . . . 12
1.5 Types of flow and
conductance . . . . . . . . . . . .12
1.5.1 Types of flow . . . . . . . . . . .12
1.5.2 Calculating conductance
values . . . . . . . . . . . . . . . . .13
1.5.3 Conductance for piping
and openings . . . . . . . . . . .14
1.5.4 Conductance values for
other elements . . . . . . . . . .15
2. Vacuum Generation . . . . .16
2.1 Vacuum pumps: A survey .16
2.1.1 Oscillation displacement
vacuum pumps . . . . . . . . . .17
2.1.1.1 Diaphragm pumps . . . . . . .17
2.1.2 Liquid sealed rotary
displacement pumps . . . . . .17
2.1.2.1 Liquid ring pumps . . . . . . . .17
2.1.2.2 Oil sealed rotary
2.1.2.2.1 Rotary vane pumps
(TRIVAC A, TRIVAC B,
TRIVAC E, SOGEVAC) . . . . .18
2.1.2.2.2 Rotary plunger pumps
(E-Pumps) . . . . . . . . . . . . .20
2.1.2.2.3 Trochoid pumps . . . . . . . . .21
2.1.2.2.4 The gas ballast . . . . . . . . . .21
2.1.3 Dry compressing rotary
2.1.3.1 Roots pumps . . . . . . . . . . .24
2.1.3.2 Claw pumps . . . . . . . . . . . .27
2.1.3.2.1 Claw pumps with internal
compression for the
semiconductor industry
(DRYVAC Series) . . . . . . .28
2.1.3.2.2 Claw pump without internal
compression for chemistry
applications (ALLex) . . . .31
2.1.4 Accessories for oil-sealed
rotary displacement
pumps . . . . . . . . . . . . . . . .33
2.1.5 Condensers . . . . . . . . . . . .33
2.1.6 Fluid-entrainment pumps . .35
2.1.6.1 (Oil) Diffusion pumps . . . . .36
2.1.6.2 Oil vapor ejector pumps . . .38
2.1.6.3 Pump fluids . . . . . . . . . . . .39
2.1.6.4 Pump fluid backstreaming
and its suppression
(Vapor barriers, baffles) . . .39
2.1.6.5 Water jet pumps and
steam ejectors . . . . . . . . . .40
2.1.7 Turbomolecular pumps . . . .41
2.1.8 Sorption pumps . . . . . . . . .45
2.1.8.1 Adsorption pumps . . . . . . .45
2.1.8.2 Sublimation pumps . . . . . . .46
2.1.8.3 Sputter-ion pumps . . . . . . .46
2.1.8.4 Non evaporable getter
pumps (NEG pumps) . . . . .48
2.1.9 Cryopumps . . . . . . . . . . . . .49
2.1.9.1 Types of cryopump . . . . . . .49
2.1.9.2 The cold head and its
operating principle . . . . . . .50
2.1.9.3 The refrigerator cryopump .51
2.1.9.4 Bonding of gases to cold
surfaces . . . . . . . . . . . . . . .51
2.1.9.5 Pumping speed and position
of the cryopanels . . . . . . . .52
2.1.9.6 Characteristic quantities of a
cryopump . . . . . . . . . . . . . .53
2.2 Choice of pumping
process . . . . . . . . . . . . . . .56
2.2.1 Survey of the most usual
pumping processes . . . . . . .56
2.2.2 Pumping of gases
(dry processes) . . . . . . . . .57
2.2.3 Pumping of gases and
vapors (wet processes) . . . .58
2.2.4 Drying processes . . . . . . . .60
2.2.5 Production of an oil-free
(hydrocarbon-free)
vacuum . . . . . . . . . . . . . . .60
2.2.6 Ultrahigh vacuum working
Techniques . . . . . . . . . . . . .61
2.3 Evacuation of a vacuum
chamber and determination
of pump sizes . . . . . . . . . . .62
2.3.1 Evacuation of a vacuum
chamber (without additional
sources of gas or vapor) . . .62
2.3.1.1 Evacuation of a chamber
in the rough vacuum
region . . . . . . . . . . . . . . . . .62
2.3.1.2 Evacuation of a chamber
in the high vacuum
region . . . . . . . . . . . . . . . . .63
2.3.1.3 Evacuation of a chamber in
the medium vacuum
region . . . . . . . . . . . . . . . . .63
2.3.2 Determination of a suitable
backing pump . . . . . . . . . . .64
2.3.3 Determination of
pump-down time from
nomograms . . . . . . . . . . . .65
2.3.4 Evacuation of a chamber
where gases and vapors
are evolved . . . . . . . . . . . . .66
2.3.5 Selection of pumps for
drying processes . . . . . . . .66
2.3.6 Flanges and their seals . . . .67
2.3.7 Choice of suitable valves . . .68
2.3.8 Gas locks and seal-off
fittings . . . . . . . . . . . . . . . .69
3. Vacuum measurement,
monitoring, control and
regulation . . . . . . . . . . . .70
3.1 Fundamentals of low-
pressure measurement . . . .70
3.2 Vacuum gauges with
pressure reading that is
independent of the type
of gas . . . . . . . . . . . . . . . . .72
3.2.1 Bourdon vacuum gauges . . .72
3.2.2 Diaphragm vacuum
gauges . . . . . . . . . . . . . . . .72
3.2.2.1 Capsule vacuum gauges . . .72
3.2.2.2 DIAVAC diaphragm vacuum
gauge . . . . . . . . . . . . . . . . .72
3.2.2.3 Precision diaphragm
vacuum gauges . . . . . . . . . .72
3.2.2.4 Capacitance diaphragm
gauges . . . . . . . . . . . . . . . .73
3.2.3 Liquid-filled (mercury)
vacuum gauges . . . . . . . . . .74
3.2.3.1 U-tube vacuum gauges . . . .74
3.2.3.2 Compression vacuum
gauges (according to
McLeod) . . . . . . . . . . . . . . .74
3.3 Vacuum gauges with
gas-dependent pressure
reading . . . . . . . . . . . . . . . .75
3.3.1 Spinning rotor gauge
(SRG) (VISCOVAC) . . . . . . .75
3.3.2 Thermal conductivity
vacuum gauges . . . . . . . . . .76
3.3.3 Ionization vacuum gauges . .77
3.3.3.1 Cold-cathode ionization
vacuum gauges (Penning
vacuum gauges) . . . . . . . . .77
3.3.3.2 Hot-cathode ionization
vacuum gauges . . . . . . . . . .78
3.4 Adjustment and calibration;
DKD, PTB national
standards . . . . . . . . . . . . . .81
3.4.1 Examples of fundamental
pressure measurement
methods (as standard
methods for calibrating
vacuum gauges . . . . . . . . . .81
3.5 Pressure monitoring, control
and regulation in vacuum
systems . . . . . . . . . . . . . . .83
D00 E 19.06.2001 21:35 Uhr Seite 3
Contents
3.5.1 Fundamentals of pressure
monitoring and control . . . .83
3.5.2 Automatic protection,
monitoring and control
of vacuum systems . . . . . . .83
3.5.3 Pressure regulation and
control in rough and
medium vacuum
systems . . . . . . . . . . . . . . .84
3.5.4 Pressure regulation in
high and ultrahigh
vacuum systems . . . . . . . . .87
3.5.5 Examples of applications
with diaphragm
controllers . . . . . . . . . . . . .88
4. Analysis of gas at low
pressures using mass
spectrometry . . . . . . . . . .89
4.1 General . . . . . . . . . . . . . . . .89
4.2 A historical review . . . . . . . .89
4.3 The quadrupole mass
spectrometer
(TRANSPECTOR) . . . . . . . .89
4.3.1 Design of the sensor . . . . . .90
4.3.1.1 The normal (open)
ion source . . . . . . . . . . . . .90
4.3.1.2 The quadrupole
separation system . . . . . . . .91
4.3.1.3 The measurement
system (detector) . . . . . . . .92
4.4 Gas admission and
pressure adaptation . . . . . .93
4.4.1 Metering valve . . . . . . . . . .93
4.4.2 Pressure converter . . . . . . .93
4.4.3 Closed ion source (CIS) . . .93
4.4.4 Aggressive gas monitor
(AGM) . . . . . . . . . . . . . . . .94
4.5 Descriptive values in
mass spectrometry
(specifications) . . . . . . . . . .94
4.5.1 Line width (resolution) . . . .94
4.5.2 Mass range . . . . . . . . . . . . .95
4.5.3 Sensitivity . . . . . . . . . . . . . .95
4.5.4 Smallest detectable
partial pressure . . . . . . . . . .95
partial pressure ratio
(concentration) . . . . . . . . . .95
4.5.6 Linearity range . . . . . . . . . .95
4.5.7 Information on surfaces
and amenability to
bake-out . . . . . . . . . . . . . . .95
4.6 Evaluating spectra . . . . . . . .96
4.6.1 Ionization and fundamental
problems in gas analysis . . .96
4.6.2 Partial pressure
measurement . . . . . . . . . .100
4.6.3 Qualitative gas analysis . . .100
4.6.4 Quantitative gas analysis . .101
4.7 Software . . . . . . . . . . . . . .102
4.7.1 Standard SQX software
(DOS) for stand-alone
operation(1 MS plus,
1 PC, RS 232) . . . . . . . . . .102
4.7.2 Multiplex/DOS software
MQX (1 to 8 MS plus
1 PC, RS 485) . . . . . . . . . .102
4.7.3 Process-oriented software
Transpector-Ware for
Windows . . . . . . . . . . . . .102
4.7.4 Development software
TranspectorView . . . . . . . .102
4.8 Partial pressure
regulation . . . . . . . . . . . . .102
4.9 Maintenance . . . . . . . . . . .103
5. Leaks and their
detection . . . . . . . . . . .104
5.1 Types of leaks . . . . . . . . . .104
5.2 Leak rate, leak size,
mass flow . . . . . . . . . . . . .104
5.2.1 The standard helium
leak rate . . . . . . . . . . . . . .106
5.2.2 Conversion equations . . . .106
5.3 Terms and definitions . . . .106
5.4 Leak detection methods
without a leak detector
unit . . . . . . . . . . . . . . . . . .107
5.4.1 Pressure rise test . . . . . . .107
5.4.2 Pressure drop test . . . . . .108
5.4.3 Leak test using vacuum
gauges which are sensitive
to the type of gas . . . . . . .108
5.4.4 Bubble immersion test . . .109
5.4.5 Foam-spray test . . . . . . . .109
5.4.6 Vacuum box check bubble .109
5.4.7 Krypton 85 test . . . . . . . . .109
5.4.8 High-frequency vacuum
test . . . . . . . . . . . . . . . . . .109
5.4.9 Testing with chemical
reactions and dye
penetration . . . . . . . . . . . .110
5.5 Leak detectors and how
they work . . . . . . . . . . . . .110
5.5.1 Halogen leak detectors
(HLD 4000, D-Tek) . . . . . .110
5.5.2 Leak detectors with
mass spectrometers (MS) .110
5.5.2.1 The operating principle
for a MSLD . . . . . . . . . . . .111
5.5.2.2 Detection limit, background,
gas storage in oil (gas ballast),
floating zero-point
suppression . . . . . . . . . . .111
5.5.2.3 Calibrating leak detectors;
test leaks . . . . . . . . . . . . .112
5.5.2.4 Leak detectors with
quadrupole mass
spectrometer
(ECOTEC II) . . . . . . . . . . .113
5.5.2.5 Helium leak detectors
with 180 sector mass
spectrometer
(UL 200, UL 500) . . . . . . .114
5.5.2.6 Direct-flow and
counter-flow leak
detectors . . . . . . . . . . . . .115
5.5.2.7 Partial flow operation . . . .115
5.5.2.8 Connection to vacuum
systems . . . . . . . . . . . . . .116
5.5.2.9 Time constants . . . . . . . .116
5.6 Limit values / Specifications
for the leak detector . . . . .117
5.7 Leak detection techniques
using helium leak
detectors . . . . . . . . . . . . .117
5.7.1 Spray technique
(local leak test) . . . . . . . . .117
5.7.2 Sniffer technology
(local leak testing using
the positive pressure
method) . . . . . . . . . . . . . .118
5.7.3 Vacuum envelope test
(integral leak test) . . . . . . .118
5.7.3.1 Envelope test test
specimen pressurized
with helium . . . . . . . . . . . .118
a) Envelope test with
concentration measur
ement and subsequent
leak rate calculation . . .118
b) Direct measurement of
the leak rate with the
leak detector
(rigid envelope) . . . . . .118
5.7.3.2 Envelope test with test
specimen evacuated . . . . .118
a) Envelope =
plastic tent . . . . . . . .118
b) Rigid envelope . . . . . . .119
5.7.4 Bombing test,
Storage under pressure .119
5.8 Industrial leak testing . . . .119
6. Thin film controllers and
control units with quartz
oscillators . . . . . . . . . . .120
6.1 Introduction . . . . . . . . . . .120
6.2 Basic principles of coating
thickness measurement
with quartz oscillators . . . .120
6.3 The shape of quartz
oscillator crystals . . . . . . .121
6.4 Period measurement . . . . .122
6.5 The Z match technique . . .122
D00 E 19.06.2001 21:35 Uhr Seite 4
Contents
6.6 The active oscillator . . . . .122
6.7 The mode-lock oscillator . .123
6.8 Auto Z match technique . .124
6.9 Coating thickness
regulation . . . . . . . . . . . . .125
6.10 INFICON instrument
variants . . . . . . . . . . . . . .127
7. Application of vacuum
technology for coating
techniques . . . . . . . . . .128
7.1 Vacuum coating
technique . . . . . . . . . . . . .128
7.2 Coating sources . . . . . . . .128
7.2.1 Thermal evaporators
(boats, wires etc.) . . . . . . .128
7.2.2 Electron beam
evaporators
(electron guns) . . . . . . . . .129
7.2.3 Cathode sputtering . . . . . .129
7.2.4 Chemical vapor
deposition . . . . . . . . . . . . .129
7.3 Vacuum coating
technology/coating
systems . . . . . . . . . . . . . .130
7.3.1 Coating of parts . . . . . . . .130
7.3.2 Web coating . . . . . . . . . . .130
7.3.3 Optical coatings . . . . . . . .131
7.3.4 Glass coating . . . . . . . . . .132
7.3.5 Systems for producing
data storage disks . . . . . . .132
8. Instructions for vacuum
equipment operation . . . .134
8.1 Causes of faults where the
desired ultimate pressure
is not achieved or is
achieved too slowly . . . . .134
8.2 Contamination of vacuum
vessels and eliminating
contamination . . . . . . . . . .134
8.3 General operating infor-
mation for vacuum
pumps . . . . . . . . . . . . . . .134
8.3.1 Oil-sealed rotary vacuum
pumps (Rotary vane
pumps and rotary piston
pumps) . . . . . . . . . . . . . . .135
8.3.1.1 Oil consumption, oil
contamination, oil
change . . . . . . . . . . . . . . .135
8.3.1.2 Selection of the pump oil
when handling
aggressive vapors . . . . . . .135
8.3.1.3 Measures when pumping
various chemical
substances . . . . . . . . . . . .136
8.3.1.4 Operating defects while
pumping with gas ballast
potential sources of error
where the required ultimate
pressure is not achieved . .137
8.3.2 Roots pumps . . . . . . . . . .137
8.3.2.1 General operating
instructions, installation
and commissioning . . . . . .137
8.3.2.2 Oil change, maintenance
work . . . . . . . . . . . . . . . . .137
8.3.2.3 Actions in case of
operational disturbances . .138
8.3.3 Turbomolecular pumps . . .138
8.3.3.1 General operating
instructions . . . . . . . . . . . .138
8.3.3.2 Maintenance . . . . . . . . . . .138
8.3.4 Diffusion and vapor-jet
vacuum pumps . . . . . . . . .139
8.3.4.1 Changing the pump fluid
and cleaning the pump . . .139
8.3.4.2 Operating errors with
diffusion and vapor-jet
pumps . . . . . . . . . . . . . . .139
8.3.5 Adsorption pumps . . . . . .139
8.3.5.1 Reduction of adsorption
capacity . . . . . . . . . . . . . .139
8.3.5.2 Changing the molecular
sieve . . . . . . . . . . . . . . . . .139
8.3.6 Titanium sublimation
pumps . . . . . . . . . . . . . . .140
8.3.7 Sputter-ion pumps . . . . . .140
8.4 Information on working
with vacuum gauges . . . . .140
8.4.1 Information on installing
vacuum sensors . . . . . . . .140
8.4.2 Contamination at the
measurement system and
its removal . . . . . . . . . . . .141
8.4.3 The influence of magnetic
and electrical fields . . . . . .141
8.4.4 Connectors, power pack,
measurement systems . . .141
9. Tables, formulas,
nomograms, diagrams
and symbols . . . . . . . . .142
Tab I Permissible pressure units
including the torr and its
conversion . . . . . . . . . . . .142
Tab II Conversion of pressure
units . . . . . . . . . . . . . . . . .142
Tab III Mean free path . . . . . . . . .142
Tab IV Compilation of important
formulas pertaining to the
kinetic theory of gases . . .143
Tab V Important values . . . . . . . .143
Tab VI Conversion of pumping
speed (volume flow rate)
units . . . . . . . . . . . . . . . . .144
Tab VII Conversion of throughput
(a,b) Q
pV
units; leak rate
units . . . . . . . . . . . . . . . . .144
Tab VIII Composition of
atmospheric air . . . . . . . . .145
Tab IX Pressure ranges used in
vacuum technology and their
characteristics . . . . . . . . . .145
Tab X Outgassing rate of
materials . . . . . . . . . . . . . .145
Tab XI Nominal internal diameters
(DN) and internal diameters
of tubes, pipes and apertures
with circular cross-section
(according to PNEUROP). .146
Tab XII Important data for
common solvents . . . . . . .146
Tab XIII Saturation pressure and
density of water . . . . . . . .147
Tab XIV Hazard classificationof
fluids . . . . . . . . . . . . . . . .148
Tab XV Chemical resistance of
commonly used elastomer
gaskets and sealing
materials . . . . . . . . . . . . .149
Tab XVI Symbols used invacuum
technology . . . . . . . . . . . .152
Tab XVII Temperature comparison
and conversion table . . . . .154
Fig. 9.1 Variation of mean free path
l (cm) with pressure for
various gases . . . . . . . . . .154
Fig. 9.2 Diagram of kinetics of
gases for air at 20 C . . . .154
Fig. 9.3 Decrease in air pressure
and change in
temperature as a
function of altitude . . . . . .155
Fig. 9.4 Change in gas composition
of the atmosphere as a
function of altitude . . . . . .155
Fig. 9.5 Conductance values for
piping of commonly used
nominal internal diameters
for molecular flow . . . . . . .155
Fig. 9.6 Conductance values for
piping of commonly used
nominal internal diameters
for molecular flow . . . . . . .155
Fig. 9.7 Nomogram for
determination of
pump-down time tp of a
vessel in the rough
vacuum pressure range . . .156
D00
D00 E 19.06.2001 21:35 Uhr Seite 5
Contents
determination of the
conductance of tubes with
a circular cross-section
for air at 20 C in the
region of molecular flow . .157
determination of
conductance of tubes in
the entire pressure range .158
Fig. 9.10 Determination of
pump-down time in the
medium vacuum range
taking into account the
evolution of gas from
the walls . . . . . . . . . . . . . .159
Fig.9.11 Saturation vapor
pressure of various
substances . . . . . . . . . . . .160
Fig. 9.12 Saturation vapor pressure
of pump fluids for oil and
mercury fluid entrainment
pumps . . . . . . . . . . . . . . .160
of major metals used in
vacuum technology . . . . . .160
Fig. 9.14 Vapor pressure of
nonmetallic sealing
materials (the vapor
pressure curve for fluoro
rubber lies between the
curves for silicone
rubber and Teflon). . . . . . .161
ps of various substances
relevant for cryogenic
technology in a
temperaturerange of
T = 2 80 K. . . . . . . . . . . .161
Fig. 9.16 Common working ranges
of vacuum pumps . . . . . . .161
Fig. 9.16a Measurement ranges of
common vacuum
gauges . . . . . . . . . . . . . . .162
Fig. 9.17 Specific volume of
saturated water vapor . . . .163
Fig. 9.18 Breakdown voltage
between electrodes for air
(Paschen curve) . . . . . . . .163
Fig 9.19 Phase diagram of water . . .164
10. The statutory units used in
vacuum technology . . . . .165
10.1 Introduction . . . . . . . . . . .165
10.2 Alphabetical list of variables,
symbols and units frequently
used in vacuum technology
and its applications . . . . . .165
10.3 Remarks on alphabetical
list in Section 10.2 . . . . . .168
10.4 Tables . . . . . . . . . . . . . . . .170
10.4.1 Basic SI units . . . . . . . . . .170
10.4.2 Derived coherent SI units
with special names
and symbols . . . . . . . . . . .170
10.4.3 Atomic units . . . . . . . . . .170
10.4.4 Derived noncoherent SI
units with special names
and symbols . . . . . . . . . . .170
11. National and international
standards and
recommendations
particularly relevant to
vacuum technology . . . . .171
11.1 National and international
standards and
recommendations of
special relevance to
vacuum technology . . . . . .171
12. References . . . . . . . . . .174
13. Index . . . . . . . . . . . . . .184
D00 E 19.06.2001 21:35 Uhr Seite 6
Vacuum Physics
1. Quantities,
their symbols,
units of
measure and
definitions
(cf. DIN 28 400, Part 1, 1990,
DIN 1314 and DIN 28 402)
1.1 Basic terms and
concepts in
vacuum technology
Pressure p (mbar)
of fluids (gases and liquids). (Quantity:
pressure; symbol: p; unit of measure:
millibar; abbreviation: mbar.) Pressure is
defined in DIN Standard 1314 as the
quotient of standardized force applied to a
surface and the extent of this surface
(force referenced to the surface area).
Even though the Torr is no longer used as
a unit for measuring pressure (see Section
10), it is nonetheless useful in the interest
of transparency to mention this pressure
unit: 1 Torr is that gas pressure which is
able to raise a column of mercury by 1 mm
at 0 C. (Standard atmospheric pressure is
760 Torr or 760 mm Hg.) Pressure p can
be more closely defined by way of sub-
scripts:
Absolute pressure pabs
Absolute pressure is always specified in
vacuum technology so that the abs index
can normally be omitted.
Total pressure p
t
The total pressure in a vessel is the sum of
the partial pressures for all the gases and
vapors within the vessel.
Partial pressure p
i
The partial pressure of a certain gas or
vapor is the pressure which that gas or
vapor would exert if it alone were present
in the vessel.
Important note: Particularly in rough
vacuum technology, partial pressure in a
mix of gas and vapor is often understood
to be the sum of the partial pressures for
all the non-condensable components
present in the mix in case of the parti-
al ultimate pressure at a rotary vane
pump, for example.
Saturation vapor pressure p
s
The pressure of the saturated vapor is
referred to as saturation vapor pressure
p
s
. p
s
will be a function of temperature for
any given substance.
Vapor pressure p
d
Partial pressure of those vapors which can
be liquefied at the temperature of liquid
nitrogen (LN
2
).
Standard pressure p
n
Standard pressure pn is defined in DIN
1343 as a pressure of p
n
= 1013.25 mbar.
Ultimate pressure p
end
The lowest pressure which can be achie-
ved in a vacuum vessel. The socalled ulti-
mate pressure p
end
depends not only on
the pumps suction speed but also upon
the vapor pressure p
d
for the lubricants,
sealants and propellants used in the pump.
If a container is evacuated simply with an
oil-sealed rotary (positive displacement)
vacuum pump, then the ultimate pressure
which can be attained will be determined
primarily by the vapor pressure of the
pump oil being used and, depending on
the cleanliness of the vessel, also on the
vapors released from the vessel walls and,
of course, on the leak tightness of the
vacuum vessel itself.
Ambient pressure pamb
or (absolute) atmospheric pressure
Overpressure p
e
or gauge pressure
(Index symbol from excess)
p
e
= p
abs
p
amb
Here positive values for p
e
will indicate
overpressure or gauge pressure; negative
values will characterize a vacuum.
Working pressure p
W
During evacuation the gases and/or vapors
are removed from a vessel. Gases are
understood to be matter in a gaseous state
which will not, however, condense at
working or operating temperature. Vapor
is also matter in a gaseous state but it may
be liquefied at prevailing temperatures by
increasing pressure. Finally, saturated
vapor is matter which at the prevailing
temperature is gas in equilibrium with the
liquid phase of the same substance. A
strict differentiation between gases and
vapors will be made in the comments
which follow only where necessary for
complete understanding.
Particle number density n (cm
-3
)
According to the kinetic gas theory the
number n of the gas molecules, referenced
to the volume, is dependent on pressure p
and thermodynamic temperature T as
expressed in the following:
p = n k T (1.1)
n = particle number density
k = Boltzmanns constant
At a certain temperature, therefore, the
pressure exerted by a gas depends only on
the particle number density and not on the
nature of the gas. The nature of a gaseous
particle is characterized, among other
factors, by its mass m
T
.
Gas density (kg m
-3
, g cm
-3
)
The product of the particle number densi-
ty n and the particle mass m
T
is the gas
density :
= n m
T
(1.2)
The ideal gas law
The relationship between the mass m
T
of a
gas molecule and the molar mass M of this
gas is as follows:
M = N
A
m
T
(1.3)
Avogadros number (or constant) N
A
indicates how many gas particles will be
contained in a mole of gas. In addition to
this, it is the proportionality factor between
the gas constant R and Boltzmanns
constant k:
R = N
A
k (1.4)
Derivable directly from the above
equations (1.1) to (1.4) is the correlation
between the pressure and the gas
density of an ideal gas.
R T
p = (1.5)
M
In practice we will often consider a certain
enclosed volume V in which the gas is
present at a certain pressure p. If m is the
mass of the gas present within that
volume, then
D00
D00 E 19.06.2001 21:35 Uhr Seite 7
ture this is always equivalent to reducing
the gas pressure p. Explicit attention must
at this point be drawn to the fact that a
reduction in pressure (maintaining the
volume) can be achieved not only by
reducing the particle number density n but
also (in accordance with Equation 1.5) by
reducing temperature T at constant gas
density. This important phenomenon will
always have to be taken into account
where the temperature is not uniform
throughout volume V.
The following important terms and
concepts are often used in vacuum
technology:
Volume V (l, m
3
, cm
3
)
The term volume is used to designate
a) the purely geometric, usually predeter-
mined, volumetric content of a vacuum
chamber or a complete vacuum system
including all the piping and connecting
spaces (this volume can be calculated);
b) the pressure-dependent volume of a
gas or vapor which, for example, is
moved by a pump or absorbed by an
adsorption agent.
Volumetric flow (flow volume) q
v
(l/s, m
3
/h, cm
3
/s)
The term flow volume designates the
volume of the gas which flows through a
piping element within a unit of time, at the
pressure and temperature prevailing at the
particular moment. Here one must realize
that, although volumetric flow may be
identical, the number of molecules moved
may differ, depending on the pressure and
temperature.
Pumping speed S (l/s, m
3
/h, cm
3
/s)
The pumping speed is the volumetric flow
through the pumps intake port.
dV
S = (1.8a)
dt
If S remains constant during the pumping
process, then one can use the difference
quotient instead of the differential quoti-
ent:
V
S = (1.8b)
t
(A conversion table for the various units of
measure used in conjunction with
pumping speed is provided in Section 9,
Table VI).
m
= (1.6)
V
The ideal gas law then follows directly
from equation (1.5):
m
p V = R T = R T (1.7)
M
Here the quotient m / M is the number of
moles u present in volume V.
The simpler form applies for m / M = 1, i.e.
for 1 mole:
p V = R T (1.7a)
The following numerical example is
intended to illustrate the correlation
between the mass of the gas and pressure
for gases with differing molar masses,
drawing here on the numerical values in
Table IV (Chapter 9). Contained in a
10-liter volume, at 20 C, will be
a) 1g of helium
b) 1g of nitrogen
When using the equation (1.7) there
results then at V = 10 l , m = 1 g,
R = 83.14 mbarlmol
-1
K
-1
,
T = 293 K (20 C)
In case a) where M = 4 g mole
-1
(monatomic gas):
In case b), with M = 28 g mole
-1
(diato-
mic gas):
The result, though appearing to be
paradoxical, is that a certain mass of a
light gas exerts a greater pressure than the
same mass of a heavier gas. If one takes
into account, however, that at the same
gas density (see Equation 1.2) more
particles of a lighter gas (large n, small m)
will be present than for the heavier gas
(small n, large m), the results become
more understandable since only the
particle number density n is determinant
for the pressure level, assuming equal
temperature (see Equation 1.1).
The main task of vacuum technology is to
reduce the particle number density n insi-
de a given volume V. At constant tempera-
Vacuum Physics
D00.08
Quantity of gas (pV value), (mbar l)
The quantity of a gas can be indicated by
way of its mass or its weight in the units of
measure normally used for mass or
weight. In practice, however, the product
of p V is often more interesting in vacu-
um technology than the mass or weight of
a quantity of gas. The value embraces an
energy dimension and is specified in
millibarliters (mbarl) (Equation 1.7).
Where the nature of the gas and its
temperature are known, it is possible to
use Equation 1.7b to calculate the mass m
for the quantity of gas on the basis of the
product of p V:
(1.7)
(1.7b)
Although it is not absolutely correct,
reference is often made in practice to the
quantity of gas p V for a certain gas.
This specification is incomplete; the
temperature of the gas T, usually room
temperature (293 K), is normally implicitly
assumed to be known.
Example:
The mass of 100 mbar l of nitrogen (N
2
)
at room temperature (approx. 300 K) is:
Analogous to this, at T = 300 K:
1 mbar l O
2
= 1.28 10
-3
g O
2
70 mbar l Ar = 1.31 10
-1
g Ar
The quantity of gas flowing through a
piping element during a unit of time in
accordance with the two concepts for gas
quantity described above can be indica-
ted in either of two ways, these being:
Mass flow q
m
(kg/h, g/s),
this is the quantity of a gas which flows
through a piping element, referenced to
time
m
q
m
= or as
t
m
mbar g mol
mbar mol K K
=

100 28
83 300
1
1 1
g g =
=
=
2800
300 83
0113 .
m
p V M
R T
=

p V
m
M
R T =
mbar
609
p
g mbar mol K K
K g mol
1 8314 293
10 4
1 1
1
.
mbar
87
p
g mbar mol K K
K g mol
1 8314 293
10 28
1 1
1
.
LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002

D00 E 19.06.2001 21:35 Uhr Seite 8
Vacuum Physics
pV flow q
pV
(mbarls
-1
).
pV flow is the product of the pressure and
volume of a quantity of gas flowing
through a piping element, divided by time,
i.e.:
p V d (p V)
q
pV
= =
t dt
pV flow is a measure of the mass flow of
the gas; the temperature to be indicated
here.
Pump throughput q
pV
The pumping capacity (throughput) for a
pump is equal either to the mass flow
through the pump intake port:
m
q
m
= (1.9)
t
or to the pV flow through the pumps inta-
ke port:
p V
q
pV
=
t
It is normally specified in mbarls
-1
. Here
p is the pressure on the intake side of the
pump. If p and V are constant at the intake
side of the pump, the throughput of this
pump can be expressed with the simple
equation
q
pV
= p S (1.10a)
where S is the pumping speed of the pump
at intake pressure of p.
(The throughput of a pump is often indica-
ted with Q, as well.)
The concept of pump throughput is of
major significance in practice and should
not be confused with the pumping speed!
The pump throughput is the quantity of
gas moved by the pump over a unit of
time, expressed in mbar l/s; the pum-
ping speed is the transportation capacity
which the pump makes available within a
specific unit of time, measured in m
3
/h
or l/s.
The throughput value is important in deter-
mining the size of the backing pump in
relationship to the size of a high vacuum
pump with which it is connected in series
in order to ensure that the backing pump
will be able to take off the gas moved by
the high vacuum pump (see Section 2.32).
Conductance C (l s
-1
)
The pV flow through any desired piping
element, i.e. pipe or hose, valves, nozzles,
openings in a wall between two vessels,
etc., is indicated with
q
pV
= C(p1 p2) = p C (1.11)
Here p = (p1 p2) is the differential bet-
ween the pressures at the inlet and outlet
ends of the piping element. The proportio-
nality factor C is designated as the con-
ductance value or simply conductance. It
is affected by the geometry of the piping
element and can even be calculated for
some simpler configurations (see Section
1.5).
In the high and ultrahigh vacuum ranges,
C is a constant which is independent of
pressure; in the rough and medium-high
regimes it is, by contrast, dependent on
pressure. As a consequence, the calculati-
on of C for the piping elements must be
carried out separately for the individual
pressure ranges (see Section 1.5 for more
detailed information).
From the definition of the volumetric flow
it is also possible to state that: The
conductance value C is the flow volume
through a piping element. The equation
(1.11) could be thought of as Ohms law
for vacuum technology, in which q
pV
corresponds to current, p the voltage and
C the electrical conductance value. Analo-
gous to Ohms law in the science of elec-
tricity, the resistance to flow
1
R =
C
has been introduced as the reciprocal
value to the conductance value. The
equation (1.11) can then be re-written as:
1
q
pV
= p (1.12)
R
The following applies directly for connec-
tion in series:
R
= R1 + R2 + R3 . . . (1.13)
When connected in parallel, the following
applies:
1 1 1 1
= + + + (1.13a)
R
R1 R2 R3
Leak rate qL (mbarls
-1
)
According to the definition formulated
above it is easy to understand that the size
of a gas leak, i.e. movement through unde-
sired passages or pipe elements, will
also be given in mbarls
-1
. A leak rate is
often measured or indicated with atmos-
pheric pressure prevailing on the one side
of the barrier and a vacuum at the other
side (p < 1 mbar). If helium (which may be
used as a tracer gas, for example) is pas-
sed through the leak under exactly these
conditions, then one refers to standard
helium conditions.
Outgassing (mbarl)
The term outgassing refers to the liberati-
on of gases and vapors from the walls of a
vacuum chamber or other components on
the inside of a vacuum system. This quan-
tity of gas is also characterized by the pro-
duct of p V, where V is the volume of the
vessel into which the gases are liberated,
and by p, or better p, the increase in
pressure resulting from the introduction of
gases into this volume.
Outgassing rate (mbarls
-1
)
This is the outgassing through a period of
time, expressed in mbarls
-1
.
Outgassing rate (mbarls
-1
cm
-2
)
(referenced to surface area)
In order to estimate the amount of gas
which will have to be extracted, knowledge
of the size of the interior surface area, its
material and the surface characteristics,
their outgassing rate referenced to the sur-
face area and their progress through time
are important.
Mean free path of the molecules (cm)
and collision rate z (s
-1
)
The concept that a gas comprises a large
number of distinct particles between which
aside from the collisions there are no
effective forces, has led to a number of
theoretical considerations which we sum-
marize today under the designation kine-
tic theory of gases.
One of the first and at the same time most
beneficial results of this theory was the
calculation of gas pressure p as a function
of gas density and the mean square of
velocity c
2
for the individual gas molecules
in the mass of molecules m
T
:
D00
D00 E 19.06.2001 21:35 Uhr Seite 9
Vacuum Physics
1 --- 1 ---
p = c
2
= n m
T
c
2
(1.14)
3 3
where
k T
c
2
= 3 (1.15)
m
T
The gas molecules fly about and among
each other, at every possible velocity, and
bombard both the vessel walls and collide
(elastically) with each other. This motion of
the gas molecules is described numerical-
ly with the assistance of the kinetic theory
of gases. A molecules average number of
collisions over a given period of time, the
so-called collision index z, and the mean
path distance which each gas molecule
covers between two collisions with other
molecules, the so-called mean free path
length , are described as shown below as
a function of the mean molecule velocity c
-
the molecule diameter 2r and the particle
number density molecules n as a very
good approximation:
C
Z = (1.16)
where
(1.17)
and
(1.18)
Thus the mean free path length for the
particle number density n is, in accordan-
ce with equation (1.1), inversely proportio-
nal to pressure p. Thus the following rela-
tionship holds, at constant temperature T,
for every gas
p = const (1.19)
Used to calculate the mean free path
length for any arbitrary pressures and
various gases are Table III and Fig. 9.1 in
Chapter 9. The equations in gas kinetics
which are most important for vacuum
technology are also summarized (Table IV)
in chapter 9.
( )
=

1
2 2
2
n r
c
k T
m
R T
M
T
=

8 8

Impingement rate z
A
(cm
-2
s
-1
) and
monolayer formation time (s)
A technique frequently used to characteri-
ze the pressure state in the high vacuum
regime is the calculation of the time requi-
red to form a monomolecular or monoato-
mic layer on a gas-free surface, on the
assumption that every molecule will stick
to the surface. This monolayer formation
time is closely related with the so-called
impingement rate z
A
. With a gas at rest the
impingement rate will indicate the number
of molecules which collide with the surfa-
ce inside the vacuum vessel per unit of
time and surface area:
(1.20)
If a is the number of spaces, per unit of
surface area, which can accept a specific
gas, then the monolayer formation time is
(1.21)
Collision frequency z
v
(cm
-3
s
-1
)
This is the product of the collision rate z
and the half of the particle number density
n, since the collision of two molecules is to
be counted as only one collision:
(1.21a)
z
n
z
V
=
2
= =

a
z
a
n c
A
4
z
n c
A
=

4
1.2 Atmospheric air
Prior to evacuation, every vacuum system
on earth contains air and it will always be
surrounded by air during operation. This
makes it necessary to be familiar with the
physical and chemical properties of
atmospheric air.
The atmosphere is made up of a number of
gases and, near the earths surface, water
vapor as well. The pressure exerted by
atmospheric air is referenced to sea level.
Average atmospheric pressure is
1013 mbar (equivalent to the atmos-
phere, a unit of measure used earlier).
Table VIII in Chapter 9 shows the compo-
sition of the standard atmosphere at relati-
ve humidity of 50 % and temperature of
20 C. In terms of vacuum technology the
following points should be noted in regard
to the composition of the air:
a) The water vapor contained in the air,
varying according to the humidity level,
plays an important part when evacua-
ting a vacuum plant (see Section 2.2.3).
b) The considerable amount of the inert
gas argon should be taken into account
in evacuation procedures using sorpti-
on pumps (see Section 2.1.8).
c) In spite of the very low content of heli-
um in the atmosphere, only about 5
ppm (parts per million), this inert gas
makes itself particularly obvious in
ultrahigh vacuum systems which are
sealed with Viton or which incorporate
glass or quartz components. Helium is
able to permeate these substances to a
measurable extent.
The pressure of atmospheric air falls with
rising altitude above the earths surface
(see Fig. 9.3 in Chapter 9). High vacuum
prevails at an altitude of about 100 km and
ultrahigh vacuum above 400 km. The com-
position of the air also changes with the
distance to the surface of the earth (see
Fig. 9.4 in Chapter 9).
D00 E 19.06.2001 21:35 Uhr Seite 10
Vacuum Physics
1.3 Gas laws and
models
1.3.1 Continuum theory
Model concept: Gas is pourable (fluid)
and flows in a way similar to a liquid. The
continuum theory and the summarization
of the gas laws which follows are based on
experience and can explain all the proces-
ses in gases near atmospheric pressure.
Only after it became possible using ever
better vacuum pumps to dilute the air to
the extent that the mean free path rose far
beyond the dimensions of the vessel were
more far-reaching assumptions necessary;
these culminated in the kinetic gas theory.
The kinetic gas theory applies throughout
the entire pressure range; the continuum
theory represents the (historically older)
special case in the gas laws where atmos-
pheric conditions prevail.
Summary of the most important gas laws
(continuum theory)
Boyle-Mariotte Law
p V = const.
for T = constant (isotherm)
Gay-Lussacs Law (Charles Law)
for p = constant (isobar)
Amontons Law
for V = constant (isochor)
Daltons Law
Poissons Law
p V
k
= const
(adiabatic)
Avogadros Law
m
V
m
V
M M
1
1
2
2
1 2
: : =
p p
total i
i
=
p p t = +
0
1 ( )
V V t = +
0
1 ( )
Ideal gas Law
Also: Equation of state for ideal gases
(from the continuum theory)
van der Waals Equation
a, b = constants
(internal pressure, covolumes)
Vm = Molar volume
Also: Equation of state for real gases
Clausius-Clapeyron Equation
L = Enthalpy of evaporation,
T = Evaporation temperature,
V
m,v,
V
m,l
= Molar volumes of vapor
or liquid
1.3.2 Kinetic gas theory
With the acceptance of the atomic view of
the world accompanied by the necessity
to explain reactions in extremely dilute
gases (where the continuum theory fails)
the kinetic gas theory was developed.
Using this it is possible not only to derive
the ideal gas law in another manner but
also to calculate many other quantities
involved with the kinetics of gases such
as collision rates, mean free path lengths,
monolayer formation time, diffusion con-
stants and many other quantities.
Model concepts and basic assumptions:
1. Atoms/molecules are points.
2. Forces are transmitted from one to
another only by collision.
3. The collisions are elastic.
4. Molecular disorder (randomness)
prevails.
A very much simplified model was develo-
ped by Krnig. Located in a cube are N
particles, one-sixth of which are moving
toward any given surface of the cube. If
the edge of the cube is 1 cm long, then it
will contain n particles (particle number
density); within a unit of time n c t/6
molecules will reach each wall where the
change of pulse per molecule, due to the
L T
dp
dT
V V
m v m l
= ( )
, ,
( ) ( ) p
a
V
V b R T
m
m
+ =
2
p V
m
M
R T R T = =
change of direction through 180 , will be
equal to 2 m
T
c. The sum of the pulse
changes for all the molecules impinging on
the wall will result in a force effective on
this wall or the pressure acting on the wall,
per unit of surface area.
where
Derived from this is
Ideal gas law (derived from the kinetic
gas theory)
If one replaces c
2
with c
2
then a compari-
son of these two general gas equations
will show:
or
The expression in brackets on the left-
hand side is the Boltzmann constant k; that
on the right-hand side a measure of the
molecules mean kinetic energy:
Boltzmann constant
Mean kinetic energy of the molecules
thus
In this form the gas equation provides a
gas-kinetic indication of the temperature!
The mass of the molecules is
m
M
N
Mass mol
Molecules mol
T
A
= =
/
/
p V N k T N E
kin
= =
2
3
E
kin
m
T
c
=

2
2
k
m
T
R
M
J
K
=

=

138 10
23
.
p V N
m
T
R
M
T N
m
T
c
=

=

( ) ( )
2
3
2
2
p V
m
M
R T N m
T
c = =
1
3
2
p V N m
T
c =
1
3
2
n
N
V
=
n
c m
T
c n c m
T
p
6
2
1
3
2
= =
D00
D00 E 19.06.2001 21:35 Uhr Seite 11
Vacuum Physics
where N
A
is Avogadros number
(previously: Loschmidt number).
Avogadro constant
N
A
= 6.022 1023 mol
1
For 1 mole, and
V = Vm = 22.414 l (molar volume);
Thus from the ideal gas law at standard
conditions
(T
n
= 273.15 K and p
n
= 1013.25 mbar):
For the general gas constant:
1.4 The pressure
ranges in vacuum
technology and their
characterization
(See also Table IX in Chapter 9.) It is com-
mon in vacuum technology to subdivide its
wide overall pressure range which spans
more than 16 powers of ten into smaller
individual regimes. These are generally
defined as follows:
Rough vacuum (RV) 1000 1 mbar
Medium vacuum (MV) 1 10
-3
.mbar
High vacuum (HV) 10
-3
10
-7
mbar
Ultrahigh vacuum (UHV)
10
-7
(10
-14
) mbar
This division is, naturally, somewhat arbit-
rary. Chemists in particular may refer to
the spectrum of greatest interest to them,
lying between 100 and 1 mbar, as inter-
mediate vacuum. Some engineers may
not refer to vacuum at all but instead speak
of low pressure or even negative pres-
sure. The pressure regimes listed above
can, however, be delineated quite satisfac-
torily from an observation of the gas-kine-
tic situation and the nature of gas flow. The
operating technologies in the various ran-
ges will differ, as well.
R
mbar mol
K
mbar
mol K
=

=
=
1013 25 22 4
27315
8314
1
. .
.
.
p V
m
M
R T =
z
n
V
=
2
1.5 Types of flow and
conductance
Three types of flow are mainly encounte-
red in vacuum technology: viscous or con-
tinuous flow, molecular flow and at the
transition between these two the so-cal-
led Knudsen flow.
1.5.1 Types of flow
Viscous or continuum flow
This will be found almost exclusively in the
rough vacuum range. The character of this
type of flow is determined by the interac-
tion of the molecules. Consequently inter-
nal friction, the viscosity of the flowing
substance, is a major factor. If vortex moti-
on appears in the streaming process, one
speaks of turbulent flow. If various layers
of the flowing medium slide one over the
other, then the term laminar flow or layer
flux may be applied.
Laminar flow in circular tubes with para-
bolic velocity distribution is known as Poi-
seuille flow. This special case is found
frequently in vacuum technology. Viscous
flow will generally be found where the
molecules mean free path is considerab-
ly shorter than the diameter of the pipe:
d.
A characteristic quantity describing the
viscous flow state is the dimensionless
Reynolds number Re.
Re is the product of the pipe diameter, flow
velocity, density and reciprocal value of the
viscosity (internal friction) of the gas
which is flowing. Flow is turbulent where
Re > 2200, laminar where Re < 2200.
The phenomenon of choked flow may also
be observed in the viscous flow situation.
It plays a part when venting and evacua-
ting a vacuum vessel and where there are
leaks.
Gas will always flow where there is a diffe-
rence in pressure p = (p1 p2) > 0. The
intensity of the gas flow, i.e. the quantity of
gas flowing over a period of time, rises
with the pressure differential. In the case
of viscous flow, however, this will be the
case only until the flow velocity, which also
rises, reaches the speed of sound. This is
always the case at a certain pressure diffe-
rential and this value may be characterized
as critical:
(1.22)
A further rise in p > p
crit
would not
result in any further rise in gas flow; any
increase is inhibited. For air at 20,C the
gas dynamics theory reveals a critical
value of
(1.23)
The chart in Fig. 1.1 represents sche-
matically the venting (or airing) of an eva-
cuated container through an opening in the
envelope (venting valve), allowing ambient
air at p = 1000 mbar to enter. In accordan-
ce with the information given above, the
resultant critical pressure is
p
crit
= 1000 (10.528) mbar 470 mbar;
i.e. where p > 470 mbar the flow rate will
be choked; where p < 470 mbar the gas
flow will decline.
Molecular flow
Molecular flow prevails in the high and
ultrahigh vacuum ranges. In these regimes
the molecules can move freely, without
any mutual interference. Molecular flow is
present where the mean free path length
for a particle is very much larger than the
diameter of the pipe: >> d.
p
p
crit
2
1
0528
_
,
= .
p p
p
p
crit
crit
=
_
,
1
]
1
1
2
1
1
p
mbar
q
m
%
1
2
1000
q
m
p
470
0
25
50
75
100
s
venting time (t)
1 Gas flow rate qm choked = constant
(maximum value)
2 Gas flow not impeded, qm drops to p = 0
Fig. 1.1 Schematic representation of venting an
evacuated vessel
D00 E 19.06.2001 21:35 Uhr Seite 12
Vacuum Physics
Knudsen flow
The transitional range between viscous
flow and molecular flow is known as Knud-
sen flow. It is prevalent in the medium
vacuum range: d.
The product of pressure p and pipe diame-
ter d for a particular gas at a certain tem-
perature can serve as a characterizing
quantity for the various types of flow.
Using the numerical values provided in
Table III, Chapter 9, the following equiva-
lent relationships exist for air at 20 C:
Rough vacuum Viscous flow
p d > 6.0 10
-1
mbar cm
Medium vacuum Knudsen flow
6 10
-1
> p d > 1.3 10
-2
mbar cm
High and ultrahigh vacuum Molecular
flow
p d < 1.3 10
-2
mbar cm
In the viscous flow range the preferred
speed direction for all the gas molecules
will be identical to the macroscopic direc-
tion of flow for the gas. This alignment is
compelled by the fact that the gas particles
are densely packed and will collide with
one another far more often than with the
boundary walls of the apparatus. The
macroscopic speed of the gas is a group
velocity and is not identical with the ther-
mal velocity of the gas molecules.
In the molecular flow range, on the other
hand, impact of the particles with the walls
predominates. As a result of reflection (but
also of desorption following a certain resi-
dence period on the container walls) a gas
particle can move in any arbitrary direction
in a high vacuum; it is no longer possible
to speak of flow in the macroscopic
sense.
It would make little sense to attempt to
determine the vacuum pressure ranges as
a function of the geometric operating
situation in each case. The limits for the
individual pressure regimes (see Table IX
in Chapter 9) were selected in such a way
that when working with normal-sized labo-
>
d
2
d d
100 2
< <
<
d
100
ratory equipment the collisions of the gas
particles among each other will predomi-
nate in the rough vacuum range whereas
in the high and ultrahigh vacuum ranges
impact of the gas particles on the contai-
ner walls will predominate.
In the high and ultrahigh vacuum ranges
the properties of the vacuum container
wall will be of decisive importance since
below 10
-3
mbar there will be more gas
molecules on the surfaces than in the
chamber itself. If one assumes a mono-
molecular adsorbed layer on the inside
wall of an evacuated sphere with 1 l volu-
me, then the ratio of the number of
adsorbed particles to the number of free
molecules in the space will be as follows:
at 1 mbar 10
-2
at 10
-6
mbar 10
+4
at 10
-11
mbar 10
+9
For this reason the monolayer formation
time t (see Section 1.1) is used to charac-
terize ultrahigh vacuum and to distinguish
this regime from the high vacuum range.
The monolayer formation time is only a
fraction of a second in the high vacuum
range while in the ultrahigh vacuum range
it extends over a period of minutes or
hours. Surfaces free of gases can therefo-
re be achieved (and maintained over lon-
ger periods of time) only under ultrahigh
vacuum conditions.
Further physical properties change as
pressure changes. For example, the ther-
mal conductivity and the internal friction of
gases in the medium vacuum range are
highly sensitive to pressure. In the rough
and high vacuum regimes, in contrast,
these two properties are virtually indepen-
dent of pressure.
Thus, not only will the pumps needed to
achieve these pressures in the various
vacuum ranges differ, but also different
vacuum gauges will be required. A clear
arrangement of pumps and measurement
instruments for the individual pressure
ranges is shown in Figures 9.16 and 9.16a
in Chapter 9.
1.5.2 Calculating
conductance values
The effective pumping speed required to
evacuate a vessel or to carry out a process
inside a vacuum system will correspond to
the inlet speed of a particular pump (or the
pump system) only if the pump is joined
directly to the vessel or system. Practically
speaking, this is possible only in rare situa-
tions. It is almost always necessary to
include an intermediate piping system com-
prising valves, separators, cold traps and
the like. All this represents an resistance to
flow, the consequence of which is that the
effective pumping speed S
eff
is always less
than the pumping speed S of the pump or
the pumping system alone. Thus to ensure
a certain effective pumping speed at the
vacuum vessel it is necessary to select a
pump with greater pumping speed. The
correlation between S and S
eff
is indicated
by the following basic equation:
(1.24)
Here C is the total conductance value for
the pipe system, made up of the individual
values for the various components which
are connected in series (valves, baffles,
separators, etc.):
(1.25)
Equation (1.24) tells us that only in the
situation where C = (meaning that the
flow resistance is equal to 0) will S = S
eff
.
A number of helpful equations is available
to the vacuum technologist for calculating
the conductance value C for piping sec-
tions. The conductance values for valves,
cold traps, separators and vapor barriers
will, as a rule, have to be determined empi-
rically.
It should be noted that in general that the
conductance in a vacuum component is
not a constant value which is independent
of prevailing vacuum levels, but rather
depends strongly on the nature of the flow
(continuum or molecular flow; see below)
and thus on pressure. When using con-
ductance indices in vacuum technology
calculations, therefore, it is always neces-
sary to pay attention to the fact that only
the conductance values applicable to a cer-
tain pressure regime may be applied in
that regime.
1 1 1 1 1
1 2 3
C C C C C
n
= + + + . . .
1 1 1
S S C
eff
= +
D00
D00 E 19.06.2001 21:35 Uhr Seite 13
Vacuum Physics
1.5.3 Conductance for piping
and orifices
Conductance values will depend not only
on the pressure and the nature of the gas
which is flowing, but also on the sectional
shape of the conducting element (e.g. cir-
cular or elliptical cross section). Other fac-
tors are the length and whether the ele-
ment is straight or curved. The result is
that various equations are required to take
into account practical situations. Each of
these equations is valid only for a particu-
lar pressure range. This is always to be
considered in calculations.
a) Conductance for a straight pipe, which
is not too short, of length l, with a cir-
cular cross section of diameter d for the
laminar, Knudsen and molecular flow
ranges, valid for air at 20 C (Knudsen
equation):
(1.26)
where
d = Pipe inside diameter in cm
l = Pipe length in cm (l 10 d)
p
1
= Pressure at start of pipe
(along the direction of flow) in mbar
p
2
= Pressure at end of pipe
(along the direction of flow) in mbar
If one rewrites the second term in (1.26) in
the following form
(1.26a)
with
(1.27)
it is possible to derive the two important
limits from the course of the function f
(d p
):
Limit for laminar flow
(d p
> 6 10
-1
mbar cm):
(1.28a)
Limit for molecular flow
(d p
< 6 10
-2
mbar cm):
C
d
l
p s = 135
4
/
( ) f d p
d p d p
d p
=
+ +
+
1 203 2 10
1 237
3 2 2
.78
( ) C
d
l
f d p = 12 1
3
.
p
p p
=
+
1 2
2
C
d
l
p = + 135
4
d
l
d p
d p
s
+
+
12 1
1 192
1 237
3
. /
(1.28b)
In the molecular flow region the conduc-
tance value is independent of pressure!
The complete Knudsen equation (1.26) will
have to be used in the transitional area
10
2
< d p
< 6 10
-1
mbar cm. Conduc-
tance values for straight pipes of standard
nominal diameters are shown in Figure 9.5
(laminar flow) and Figure 9.6 (molecular
flow) in Chapter 9. Additional nomograms
for conductance determination will also be
found in Chapter 9 (Figures 9.8 and 9.9).
b) Conductance value C for an orifice A
(A in cm
2
): For continuum flow (viscous
flow) the following equations (after
Prandtl) apply to air at 20 C where
p
2
/p
1
= :
for 0.528 (1.29)
for 0.528 (1.29a)
and for 0,03 (1.29b)
= 0.528 is the critical pressure
situation for air
Flow is choked at < 0.528; gas flow is
thus constant. In the case of molecular
flow (high vacuum) the following will apply
for air:
C
mol
= 11,6 A l s
-1
(A in cm
2
) (1.30)
Given in addition in Figure 1.3 are the
p
p
crit
2
1
_
,
C A
s
visc
= 20

C
A
s
visc
=
20
1
C
A
s
visc
=
76 6 1
1
0.712 0.288
.
C
d
l
s = 12 1
3
. /
pumping speeds S*
visc
and S*
mol
refer-
enced to the area A of the opening and as
a function of = p
2
/p
1
. The equations
given apply to air at 20 C. The molar mas-
ses for the flowing gas are taken into con-
sideration in the general equations, not
shown here.
When working with other gases it will be
necessary to multiply the conductance
values specified for air by the factors
shown in Table 1.1.
Fig. 1.2 Flow of a gas through an opening (A) at high
pressures (viscous flow)
l s
1
cm
2
Fig. 1.3 Conductance values relative to the area,
C*
visc
, C*
mol
, and pumping speed S*
visc
and S*
mol
for an orifice A, depending on the
pressure relationship p
2
/p
1
for air at 20 C.
Gas (20 C) Molecular flow Laminar flow
Air 1.00 1.00
Oxygen 0.947 0.91
Neon 1.013 1.05
Helium 2.64 0.92
Hydrogen 3.77 2.07
Carbon dioxide 0.808 1.26
Water vapor 1.263 1.7
Table 1.1 Conversion factors (see text)
D00 E 19.06.2001 21:35 Uhr Seite 14
Vacuum Physics
Nomographic determination of conduc-
tance values
The conductance values for piping and
openings through which air and other
gases pass can be determined with nomo-
graphic methods. It is possible not only to
determine the conductance value for
piping at specified values for diameter,
length and pressure, but also the size of
the pipe diameter required when a pum-
ping set is to achieve a certain effective
pumping speed at a given pressure and
given length of the line. It is also possible
to establish the maximum permissible pipe
length where the other parameters are
known. The values obtained naturally do
not apply to turbulent flows. In doubtful
situations, the Reynolds number Re (see
Section 1.5.) should be estimated using
the relationship which is approximated
below
(1.31)
Here q
pV
= S p is the flow output in mbar
l/s, d the diameter of the pipe in cm.
A compilation of nomograms which have
proved to be useful in practice will be
found in Chapter 9.
1.5.4 Conductance values for
other elements
Where the line contains elbows or other
curves (such as in right-angle valves),
these can be taken into account by assu-
ming a greater effective length l
eff
of the
line. This can be estimated as follows:
(1.32)
Where
l
axial
: axial length of the line (in cm)
l
eff
: Effective length of the line (in cm)
d : Inside diameter of the line (in cm)
: Angle of the elbow
(degrees of angle)
l l d
eff axial
= +
133
180
.

Re = 15
q
d
pV
The technical data in the Leybold catalog
states the conductance values for vapor
barriers, cold traps, adsorption traps and
valves for the molecular flow range. At hig-
her pressures, e.g. in the Knudsen and
laminar flow ranges, valves will have about
the same conductance values as pipes of
corresponding nominal diameters and
axial lengths. In regard to right-angle val-
ves the conductance calculation for an
elbow must be applied.
In the case of dust filters which are used to
protect gas ballast pumps and roots
pumps, the percentage restriction value
for the various pressure levels are listed in
the catalog. Other components, namely the
condensate separators and condensers,
are designed so that they will not reduce
pumping speed to any appreciable extent.
The following may be used as a rule of
thumb for dimensioning vacuum lines:
The lines should be as short and as wide
as possible. They must exhibit at least the
same cross-section as the intake port at
the pump. If particular circumstances pre-
vent shortening the suction line, then it is
advisable, whenever this is justifiable from
the engineering and economic points of
view, to include a roots pump in the suc-
tion line. This then acts as a gas entrain-
ment pump which reduces line impedance.
D00
Axial length
D00 E 19.06.2001 21:35 Uhr Seite 15
Vacuum Generation
2. Vacuum
generation
2.1. Vacuum pumps: A
survey
Vacuum pumps are used to reduce the gas
pressure in a certain volume and thus the
gas density (see equation 1.5). Conse-
quently consider the gas particles need to
be removed from the volume. Basically
differentiation is made between two
classes of vacuum pumps:
a) Vacuum pumps where via one or
several compression stages the gas
particles are removed from the volume
which is to be pumped and ejected into
the atmosphere (compression pumps).
The gas particles are pumped by means
of displacement or pulse transfer.
b) Vacuum pumps where the gas particles
which are to be removed condense on
or are bonded by other means (e.g.
chemically) to a solid surface, which
often is part of the boundary forming
volume itself.
A classification which is more in line with
the state-of-the-art and practical applica-
tions makes a difference between the
following types of pumps, of which the
first three classes belong to the compres-
sion pumps and where the two remaining
classes belong to the condensation and
getter pumps:
1. Pumps which operate with periodically
increasing and decreasing pump cham-
ber volumes (rotary vane and rotary
plunger pumps; also trochoid pumps)
2. Pumps which transport quantities of
gas from the low pressure side to the
high pressure side without changing
the volume of the pumping chamber
(Roots pumps, turbomolecular pumps)
3. Pumps where the pumping effect is
based mainly on the diffusion of gases
into a gas-free high speed vapor jet
(vapor pumps)
4. Pumps which pump vapors by means of
condensation (condensers) and pumps
which pump permanent gases by way
of condensation at very low tempera-
tures (cryopumps)
5. Pumps which bond or incorporate ga-
ses by adsorption or absorption to
surfaces which are substantially free of
gases (sorption pumps).
A survey on these classes of vacuum
pumps is given in the diagram of Table 2.1.
Adsorption
pump
Fluid entrainment
vacuum pump
Ejector
vacuum pump
Liquid jet
vacuum pump
Gas jet
vacuum pump
Vapor jet
vacuum pump
Diffusion pump
Self-purifying
diffusion pump
Fractionating
diffusion pump
Diffusion ejector
pump
Drag
vacuum pump
Gaseous ring
vacuum pump
Turbine
vacuum pump
Axial flow
vacuum pump
Radial flow
vacuum pump
Molecular drag
vacuum pump
Turbomolecular pump
Rotary
vacuum pump
Liquid sealed
vacuum pump
Liquid ring
vacuum pump
Rotary vane
vacuum pump
Multiple vane
vacuum pump
Rotary piston
vacuum pump
Rotary plunger
vacuum pump
Dry compressing
vacuum pump
Roots
vacuum pump
Claw
vacuum pump
Scroll pump
Reciprocating
positive displacement
vacuum pump
Diaphragm
vacuum pump
Piston
vacuum pump
Ion transfer
vacuum pump
Getter pump
Bulk getter pump
Sublimation
pump
Getter ion pump
Evaporation ion pump
Sputter-ion pump
Cryopump
Condenser
Vacuum pump
(Operating principle)
Kinetic
vacuum pump
Gas transfer
vacuum pump
Positive displacement
vacuum pump
Entrapment
vacuum pump
Table 2.1 Classification of vacuum pumps
D00 E 19.06.2001 21:35 Uhr Seite 16
Vacuum Generation
2.1.1 Oscillation displacement
vacuum pumps
2.1.1.1 Diaphragm pumps
Recently, diaphragm pumps have
becoming ever more important, mainly for
environmental reasons. They are alterna-
tives to water jet vacuum pumps, since
diaphragm pumps do not produce any
waste water. Overall, a diaphragm vacuum
pump can save up to 90 % of the operating
costs compared to a water jet pump.
Compared to rotary vane pumps, the
pumping chamber of diaphragm pumps
are entirely free of oil. By design, no oil
immersed shaft seals are required.
Diaphragm vacuum pumps are single or
multi-stage dry compressing vacuum
pumps (diaphragm pumps having up to
four stages are being manufactured). Here
the circumference of a diaphragm is
tensioned between a pump head and the
casing wall (Fig. 2.1). It is moved in an
oscillating way by means of a connecting
rod and an eccentric. The pumping or
compression chamber, the volume of
which increases and decreases periodi-
cally, effects the pumping action. The
valves are arranged in such a way that
during the phase where the volume of the
pumping chamber increases it is open to
the intake line. During compression, the
pumping chamber is linked to the exhaust
line. The diaphragm provides a hermetic
seal between the gear chamber and the
pumping chamber so that it remains free
of oil and lubricants (dry compressing
vacuum pump). Diaphragm and valves are
the only components in contact with the
medium which is to be pumped. When
coating the diaphragm with PTFE (Teflon)
and when manufacturing the inlet and
exhaust valves of a highly fluorinated
elastomer as in the case of the DIVAC from
LEYBOLD, it is then possible to pump
aggressive vapors and gases. It is thus
well suited for vacuum applications in the
chemistry lab.
Due to the limited elastic deformability of
the diaphragm only a comparatively low
pumping speed is obtained. In the case of
this pumping principle a volume remains
at the upper dead center the so called
dead space from where the gases can
not be moved to the exhaust line. The
quantity of gas which remains at the ex-
haust pressure expands into the expanding
pumping chamber during the subsequent
suction stroke thereby filling it, so that as
the intake pressure reduces the quantity of
inflowing new gas reduces more and
more. Thus volumetric efficiency worsens
continuously for this reason. Diaphragm
vacuum pumps are not capable of
attaining a higher compression ratio than
the ratio between dead space and
maximum volume of the pumping cham-
ber. In the case of single-stage diaphragm
vacuum pumps the attainable ultimate
pressure amounts to approximately 80
mbar. Two-stage pumps such as the
DIVAC from LEYBOLD can attain about
10 mbar (see Fig. 2.2), three-stage pumps
can attain about 2 mbar and four-stage
diaphragm pumps can reach about
510
-1
mbar.
Diaphragm pumps offering such a low
ultimate pressure are suited as backing
pumps for turbomolecular pumps with fully
integrated Scroll stages (compound or wide
range turbomolecular pumps, such as the
TURBOVAC 55 from LEYBOLD). In this way
a pump system is obtained which is ab-
solutely free of oil, this being of great im-
portance to measurement arrangements
involving mass spectrometer systems and
leak detectors. In contrast to rotary vane
pumps this combination of pumps for leak
detectors offers the advantage that
naturally no helium is dissolved in the
diaphragm pump thereby entirely avoiding
a possible build up of a helium background.
2.1.2 Liquid sealed rotary
displacement pumps
2.1.2.1 Liquid ring pumps
Due to the pumping principle and the
simple design, liquid ring vacuum pumps
are particularly suited to pumping gases
and vapors which may also contain small
amounts of liquid. Air, saturated with water
vapors or other gases containing conden-
D00
(1) Casing lid
(2) Valves
(3) Lid
(4) Diaphragm disk
(5) Diaphragm
(6) Diaphragm support disk
(7) Connecting rod
(8) Eccentric disk
Fig. 2.1 Schematic on the design of a diaphragm pump stage (Vacuubrand) Fig. 2.2 Principle of operation for a two-stage diaphragm pump (Vacuubrand)
Opening and closing of the valves, path and pumping mechanism
during four subsequent phases of a turn of the connecting rod (a-d)
a)
b)
c)
d)
1st stage 2nd stage
IN EX
D00 E 19.06.2001 21:35 Uhr Seite 17
Vacuum Generation
sable constituents, may be pumped
without problems. By design, liquid ring
pumps are insensitive to any contamina-
tion which may be present in the gas flow.
The attainable intake pressures are in the
region between atmospheric pressure and
the vapor pressure of the operating liquid
used. For water at 15 C it is possible to
attain an operating pressure of 33 mbar. A
typical application of water ring vacuum
pumps is venting of steam turbines in
power plants. Liquid ring vacuum pumps
(Fig. 2.3) are rotary displacement pumps
which require an operating liquid which
rotates during operation to pump the gas.
The blade wheel is arranged eccentrically
in a cylindrical casing. When not in opera-
tion, approximately half of the pump is
filled with the operating fluid. In the axial
direction the cells formed by the blade
wheel are limited and sealed off by
control discs. These control discs are
equipped with suction and ejection slots
which lead to the corresponding ports of
the pump. After having switched on such a
pump the blade wheel runs eccentrically
within the casing; thus a concentrically
rotating liquid ring is created which at the
narrowest point fully fills the space
between the casing and the blade wheel
and which retracts from the chambers as
the rotation continues. The gas is sucked
in as the chambers empty and compres-
sion is obtained by subsequent filling. The
limits for the intake or discharge process
are set by the geometry of the openings in
the control discs.
In addition to the task of compression, the
operating fluid fulfills three further
important tasks:
1. Removal of the heat produced by the
compression process.
2. Uptake of liquids and vapors
(condensate).
3. Providing the seal between the blade
wheel and the casing.
2.1.2.2 Oil sealed rotary displacement
pumps
A displacement vacuum pump is generally
a vacuum pump in which the gas which is
to be pumped is sucked in with the aid of
pistons, rotors, vanes and valves or
similar, possibly compressed and then
discharged. The pumping process is
effected by the rotary motion of the piston
inside the pump. Differentiation should be
made between oiled and dry compressing
displacement pumps. By the use of sealing
oil it is possible to attain in a single-stage
high compression ratios of up to about
10
5
. Without oil, inner leakiness is
considerably greater and the attainable
compression ratio is correspondingly less,
about 10.
As shown in the classification Table 2.1,
the oil sealed displacement pumps include
rotary vane and rotary plunger pumps of
single and two-stage design as well as
single-stage trochoid pumps which today
are only of historic interest. Such pumps
are all equipped with a gas ballast facility
which was described in detail (for details
see 2.1.2.2.4) for the first time by Gaede
(1935). Within specified engineering
limits, the gas ballast facility permits
pumping of vapors (water vapor in
particular) without condensation of the
vapors in the pump.
2.1.2.2.1 Rotary vane pumps
(TRIVAC A, TRIVAC B,
TRIVAC E, SOGEVAC)
Rotary vane pumps (see also Figs. 2.5 and
2.6) consist of a cylindrical housing
(pump-ing ring) (1) in which an eccentri-
cally suspended and slotted rotor (2) turns
in the direction of the arrow. The rotor has
vanes (16) which are forced outwards
usually by centrifugal force but also by
springs so that the vanes slide inside the
housing. Gas entering through the intake
(4) is pushed along by the vanes and is
finally ejected from the pump by the oil
sealed exhaust valve (12).
The older range of TRIVAC A pumps (Fig.
2.5) from LEYBOLD has three radial vanes
offset by 120. The TRIVAC B range (Fig.
2.6) has only two vanes offset by 180. In
both cases the vanes are forced outwards
by the centrifugal forces without the use of
springs. At low ambient temperatures this
possibly requires the use of a thinner oil.
The A-Series is lubricated through the
arising pressure difference whereas the B-
Series pumps have a geared oil pump for
pressure lubrication. The TRIVAC B-Series
is equipped with a particularly reliable anti-
suckback valve; a horizontal or vertical
Fig. 2.3 Liquid ring vacuum pump, schematic
(Siemens)
1 Rotor
2 Rotor shaft
3 Casing
4 Intake channel
5 Liquid ring
6 Flex. discharge channel
Fig. 2.4 Arrangement of the sealing passage in rotary
vane pumps
also known as duo seal
Constant, minimum clearance a for the entire
sealing passage b
Fig. 2.5 Cross section of a single-stage rotary vane
pump (TRIVAC A)
1 Pump housing
2 Rotor
3 Oil-level sight glass
4 Suction duct
5 Anti-suckback valve
6 Dirt trap
7 Intake port
8 Lid of gas ballast
valve
9 Exhaust port
10 Air inlet silencer
11 Oil filter
12 Exhaust valve
13 Exhaust duct
14 Gas ballast duct
15 Oil injection
16 Vane
D00 E 19.06.2001 21:35 Uhr Seite 18
Vacuum Generation
arrangement for the intake and exhaust
ports. The oil level sight glass and the gas
ballast actuator are all on the same side of
the oil box (user friendly design). In
combination with the TRIVAC BCS system
it may be equipped with a very com-
prehensive range of accessories, designed
chiefly for semiconductor applications.
The oil reservoir of the rotary vane pump
and also that of the other oil sealed
displacement pumps serves the purpose
of lubrication and sealing, and also to fill
dead spaces and slots. It removes the heat
of gas compression, i.e. for cooling
purposes. The oil provides a seal between
rotor and pump ring. These parts are
almost in contact along a straight line
(cylinder jacket line). In order to increase
the oil sealed surface area a so-called
sealing passage is integrated into the
pumping ring (see Fig. 2.4). This provides
a better seal and allows a higher com-
pression ratio or a lower ultimate pres-
sure. LEYBOLD manufactures three
different ranges of rotary vane pumps
which are specially adapted to different
applications such as high intake pressure,
low ultimate pressure or applications in
the semiconductor industry. A summary of
the more important characteristics of
these ranges is given in Table 2.2. The
TRIVAC rotary vane pumps are produced
as single-stage (TRIVAC S) and two-stage
(TRIVAC D) pumps (see Fig. 2.7). With the
two-stage oil sealed pumps it is possible
to attain lower operating and ultimate
pressures ompared to the corresponding
single-stage pumps. The reason for this is
that in the case of single-stage pumps, oil
is unavoidably in contact with the
atmosphere outside, from where gas is
taken up which partially escapes to the
vacuum side thereby restricting the
attainable ultimate pressure. In the oil
sealed two-stage displacement pumps
manufactured by LEYBOLD, oil which has
already been degassed is supplied to the
stage on the side of the vacuum (stage 1 in
Fig. 2.7): the ultimate pressure lies almost
D00
TRIVAC A TRIVAC B TRIVAC BCS TRIVAC E SOGEVAC
Vanes per stage 3 2 2 2 3 (tangential)
1 1.5 1.6 16 25
Pumping speed 2 4 4 8 16 25 2.5 40 100
[m
3
/h] 8 16 16 25 40 65 180 280
30 60 40 65 585 1200
Sealing passage yes yes yes yes no
Ultimate pressure, < 2 10
2
< 2 10
2
< 2 10
2
< 5 10
1
single-stage [mbar]
Ultimate pressure < 2.5 10
4
< 1 10
4
< 1 10
4
< 1 10
4
two-stage [mbar]
Oil supply Pressure difference Gear pump Gear pump Eccentric pump Pressure difference
Slots Comparable for all types: about 0.01 to 0.05 mm
Bearing/lubrication Axial face / oil Axial face / oil Axial face / oil Ball / grease Ball / oil
Special Hydropneumatic Coated parts in Many Cost-effective
characteristics anti-suckback valve contact with medium accessories
Media No ammonia Clean to Aggressive and Clean to Clean
light particles corrosive light particles
Main areas of Multi- Multi- Semiconductor Multi- Packaging
application purpose purpose industry purpose industry
Fig. 2.6 Cross section of a single-stage rotary vane
pump (TRIVAC B)
1 Intake port
2 Dirt trap
3 Anti-suckback
valve
4 Intake duct
5 Vane
6 Pumping cham-
ber
7 Rotor
8 Orifice, connec-
tion for inert gas
ballast
9 Exhaust duct
10 Exhaust valve
11 Demister
12 Spring
13 Orifice; connec-
tion for oil filter
Leaf
spring
of the
valve
I High vacuum stage
II Second forevacuum stage
Fig. 2.7 Cross section of a two-stage rotary vane
pump, schematic
Valve stop
Table 2.2 Rotary vacuum pump ranges
D00 E 19.06.2001 21:35 Uhr Seite 19
Vacuum Generation
in the high vacuum range, the lowest
operating pressures lie in the range
between medium vacuum / high vacuum.
Note: operating the so called high vacuum
stage (stage 1) with only very little oil or
no oil at all will in spite of the very low
ultimate pressure in practice lead to
considerable difficulties and will signifi-
cantly impair operation of the pump.
2.1.2.2.2 Rotary plunger pumps
(E-Pumps)
Shown in Fig. 2.9 is a sectional view of a
rotary plunger pump of the single block
type. Here a piston (2) which is moved
along by an eccentric (3) turning in the
direction of the arrow moves along the
chamber wall. The gas which is to be
pumped flows into the pump through the
intake port (11), passes through the intake
channel of the slide valve (12) into the
pumping chamber (14). The slide valve
forms a unit with the piston and slides to
and fro between the rotatable valve guide
in the casing (hinge bar 13). The gas
drawn into the pump finally enters the
compression chamber (4). While turning,
the piston compresses this quantity of gas
until it is ejected through the oil sealed
valve (5). As in the case of rotary vane
pumps, the oil reservoir is used for
lubrication, sealing, filling of dead spaces
and cooling. Since the pumping chamber
is divided by the piston into two spaces,
each turn completes an operating cycle
(see Fig. 2.10). Rotary plunger pumps are
manufactured as single and two-stage
pumps. In many vacuum processes
combining a Roots pump with a single-
stage rotary plunger pump may offer more
advantages than a two-stage rotary
plunger pump alone. If such a combination
or a two-stage pump is inadequate, the
use of a Roots pump in connection with a
two-stage pump is recommended. This
does not apply to combinations involving
rotary vane pumps and Roots pumps.
Motor power
The motors supplied with the rotary vane
and rotary plunger pumps deliver enough
power at ambient temperatures of 12 C
and when using our special oils to cover
the maximum power requirement (at about
400 mbar). Within the actual operating
range of the pump, the drive system of the
warmed up pump needs to supply only
about one third of the installed motor
power (see Fig. 2.11).
Fig. 2.8a Cross section of a two-stage rotary vane
pump (TRIVAC E)
Fig. 2.8b SOGEVAC pump SV 300 with three tangen-
tial vanes
Fig. 2.9 Cross section of a single-stage rotary plun-
ger pump (monoblock design)
1 Casing
2 Cylindrical piston
3 Eccentric
4 Compression
chamber
5 Oil sealed pressure
valve
6 Oil-level sight glass
7 Gas ballast channel
8 Exhaust pot
9 Gas ballast valve
10 Dirt trap
11 Intake port
12 Slide valve
13 Hinge bar
14 Pumping chamber
(air is flowing in)
Fig. 2.10 Operating cycle of a rotary plunger pump (for positions 1 to 9 of the plunger)
1 Upper dead point
2 Slot in suction channel of
slide valve is freed begin-
ning of suction period
3 Lower dead point slot in
suction channel is quite free,
and pumped-in gas (arrow)
enters freely into the pum-
ping chamber (shown sha-
ded)
4 Slot in suction channel is clo-
sed again by swivelling hinge
bar end of suction period
5 Upper dead point maximum
space between rotating
piston and stator
6 Shortly before beginning of
compression period, the front
surface of the rotating plun-
ger frees gas ballast opening
commencement of gas bal-
last inlet
7 Gas ballast opening is quite
free
8 End of gas ballast inlet
9 End of pumping period.
D00 E 19.06.2001 21:35 Uhr Seite 20
Vacuum Generation
2.1.2.2.3 Trochoid pumps
Trochoid pumps belong to the class of so
called rotary piston pumps, which (see
overview of Table 2.1) in turn belong to the
group of rotary pumps. With rotary piston
pumps, the pistons center of gravity runs
along a circular path about the rotational
axis (hence rotary piston machines). A
rotary piston pump can in contrast to the
rotary plunger pump be completely
balanced dynamically. This offers the
advantage that larger pumps can operate
without vibration so that they can be
installed without needing foundations.
Moreover, such pumps may be operated at
higher speed, compared to rotary plunger
pumps (see below). The volume of the
pumping chamber with respect to the
volume of the entire pump the so called
specific volume is, in the case of
trochoid pumps, approximately twice of
that of rotary plunger pumps. Larger
rotary plunger pumps run at speeds of
500 rpm. The trochoid pump may run at
1000 rpm and this applies also to larger
designs. It is thus about four times smaller
compared to a rotary plunger pump having
the same pumping speed and runs without
producing any vibrations. Unfortunately
the advantages in the area of engineering
are combined with great disadvantages in
the area of manufacturing, so that today
LEYBOLD does not produce trochoid
pumps any more. Operation of such a
pump is shown in the sectional diagram of
Fig. 2.12.
2.1.2.2.4 The gas ballast
The gas ballast facility as used in the rotary
vane, rotary plunger and trochoid pumps
permits not only pumping of permanent
gases but also even larger quantities of
condensable gases.
The gas ballast facility (see Fig. 2.13)
prevents condensation of vapors in the
pump chamber of the pump. When
pumping vapors these may only be
compressed up to their saturation vapor
pressure at the temperature of the pump.
If pumping water vapor, for example, at a
pump temperature of 70 C, the vapor may
only be compressed to 312 mbar
(saturation vapor pressure of water at
70 C (see Table XIII in Section 9)). When
compressing further, the water vapor
condenses without increasing the
pressure. No overpressure is created in
the pump and the exhaust valve is not
opened. Instead the water vapor remains
as water in the pump and emulsifies with
the pumps oil. This very rapidly impairs
the lubricating properties of the oil and the
pump may even seize when it has taken up
too much water. The gas ballast facility
developed in 1935 by Wolfgang Gaede
inhibits the occurrence of condensation of
the vapor in the pump as follows. Before
the actual compression process begins
(see Fig. 2.13), a precisely defined quantity
of air (the gas ballast) is admitted into
the pumping chamber of the pump. The
quantity is such that the compression ratio
of the pump is reduced to 10:1 max. Now
vapors which have been taken in by the
pump may be compressed together with
the gas ballast, before reaching their
condensation point and ejected from the
pump. The partial pressure of the vapors
which are taken in may however not
exceed a certain value. It must be so low
that in the case of a compression by a
factor of 10, the vapors can not condense
at the operating temperature of the pump.
When pumping water vapor this critical
value is termed the water vapor
tolerance.
Shown schematically in Fig. 2.14 is the
pumping process with and without gas
ballast as it takes place in a rotary vane
pump when pumping condensable vapors.
D00
Fig. 2.11 Motor power of a rotary plunger pump (pum-
ping speed 60 m
3
/h) as a function of intake
pressure and operating temperature. The cur-
ves for gas ballast pumps of other sizes are
similar.
1 Operating temp.
curve 1 32C,
2 Operating temp.
curve 2 40C,
3 Operating temp.
curve 3 60C,
4 Operating temp.
curve 4 90C,
5 Theoretic curve for
adiabatic compres-
sion
6 Theoretic curve for
isotherm compres-
sion
Pressure [mbar]
Fig. 2.12 Cross section of a trochoid pump
1 Toothed wheel fixed to the driving shaft
2 Toothed wheel fixed to the piston
3 Elliptic piston
4 Inner surface of the pump
5 Driving shaft
6 Eccentric
P
o
w
e
r

o
f

t
h
e

d
r
i
v
e

m
o
t
o
r

[
W
a
t
t
]
Suction
C
o
m
p
r
e
s
s
i
o
n
G
a
s
b
a
lla
s
t in
le
t
Gas ballast
inlet
Position
of the
leading
vane
D
is
c
h
a
r
g
e
Fig. 2.13 Working process within a rotary vane pump
with gas ballast
12 Suction
25 Compression
34 Gas ballast inlet
56 Discharge
D00 E 19.06.2001 21:35 Uhr Seite 21
Vacuum Generation
Two requirements must be met when
pumping vapors:
1) the pump must be at operating
temperature.
2) the gas ballast valve must be open.
(With the gas ballast valve open the
temperature of the pump increases by
about 10 C. Before pumping vapors the
pump should be operated for half an hour
with the gas ballast valve open).
Simultaneous pumping of gases and
vapors
When simultaneously pumping permanent
gases and condensable vapors from a
vacuum system, the quantity of permanent
gas will often suffice to prevent any
condensation of the vapors inside the
pump. The quantity of vapor which may be
pumped without condensation in the
pump can be calculated as follows:
(2.1)
Where:
p
vapor
= is the partial pressure of
vapor at the intake of the
pump
p
perm
= is the total pressure of all
pumped permanent gases
at the intake of the pump
p
vapor, sat
= is the saturation pressure
of the pumped vapor,
depending on temperatu-
re (see Fig. 2.15)
p
sum
= p
exhaust
+ p
valve
+
p
exhaust filter
p
valve
= is the pressure difference
across the exhaust valve
which amounts depen-
ding on type of pump and
operating conditions to
0.2 ... 0.4 bar
p
p p
p
p
perm
vapour sat
sum
vapour
vapour
+
<
,
p
exhaust filter
= is the pressure difference
across the exhaust filter
amounting to 0 ... 0.5 bar
Example 1:
With a rotary vane pump with an external
oil mist filter in series, a mixture of water
vapor and air is being pumped. The follo-
wing values are used for applying eq. (2.1):
p
exhaust
= 1 bar
p
valve
+ p
exhaust filter
= 0.35 bar,
temperature of the pump 70 C
Hence:
p
sum
= 1.35 bar; p
vapor sat
(H
2
O)
= 312 mbar
(see Table XIII in chapter 9)
Using eq. (2.1) follows:
The pressure of the water vapor in the
air/water vapor mixture must not exceed
23 % of the total pressure of the mixture.
Example 2:
Ethanoic acid is to be pumped with a
rotary plunger pump.
p
exhaust
= 1.1 bar (taking into
consideration the flow
resistance of the pipes)
p
valve
= 0.25 bar
p
exhaust filter
= 0.15 bar
(pressure loss in the oil
mist trap)
Hence:
p
sum
= 1.5 bar.
By controlled cooling the pump and oil
temperature is set at 100 C. The saturati-
on pressure of the acid therefore is see
Fig. 2.15 p
vapor, sat
= 500 mbar.
From eq. (2.1) follows:
Returning to the question of pumping
water vapor in a mixture with air, the ratio
3 parts of permanent gases to 1 part of
water vapor, as indicated in example 1, can
be for guidance only. In actual practice it is
p
p p
vapour, acid
vapour, acid air
.
. +
= < =
0 5
15
1
3
p
p p
vapour , H
2
O
vapour , H
2
O air
.
312
1350
0 23
+
< =
Fig. 2.14 Diagram of pumping process in a rotary vane
pump without (left) and with (right) gas bal-
last device when pumping condensable sub-
stances
a) Without gas bal-
last
1) Pump is connected
to the vessel, which
is already almost
empty of air
(70 mbar) it must
thus transport
mostly vapor
particles
2) Pump chamber is
separated from the
vessel compres-
sion begins
3) Content of pump
chamber is already
so far compressed
that the vapor con-
denses to form
droplets over-
pressure is not yet
reached
4) Residual air only
now produces the
required overpres-
sure and opens the
discharge valve,
but the vapor has
already condensed
and the droplets
are precipitated in
the pump.
b) With gas ballast
1) Pump is connected
to the vessel, which
is already almost
empty of air
(70 mbar) it must
thus transport
mostly vapor
particles
2) Pump chamber is
separated from the
vessel now the
gas ballast valve,
through which the
pump chamber is
filled with additio-
nal air from outsi-
de, opens this
additional air is cal-
led gas ballast
3) Discharge valve is
pressed open, and
particles of vapor
and gas are pushed
out the overpres-
sure required for
this to occur is rea-
ched very early
because of the sup-
plementary gas
ballast air, as at the
beginning the enti-
re pumping pro-
cess condensation
cannot occur
4) The pump dischar-
ges further air and
vapor
Fig. 2.15 Saturation vapor pressures
Temperature [ C]
S
a
t
u
r
a
t
i
o
n

v
a
p
o
r

p
r
e
s
s
u
r
e

[
T
o
r
r
]
S
a
t
u
r
a
t
i
o
n

v
a
p
o
r

p
r
e
s
s
u
r
e

[
m
b
a
r
]
D00 E 19.06.2001 21:35 Uhr Seite 22
Vacuum Generation
recommended to run up a rotary pump of
the types described hitherto always with
the gas ballast valve open, because it takes
some time until the pump has reached its
final working temperature.
From eq. (2.1) follows for the permissible
partial pressure p
vapor
of the pumped
vapor the relation
(2.2)
This relation shows that with p
perm
= 0 no
vapors can be pumped without condensa-
tion in the pump, unless the gas-ballast
concept is applied. The corresponding for-
mula is:
(2.3)
Where:
B = is the volume of air at
1013 mbar which is admitted
to the pump chamber per unit
time, called in brief the gas
ballast
S = is the nominal speed of the
pump (volume flow rate)
p
sum
= is the pressure at the
discharge outlet of the pump,
assumed to be a maximum at
1330 mbar
p
vapor, sat
= Saturation vapor pressure of
the vapor at the pumps
exhaust port
p
vapor, g.b.
= is the partial pressure of any
vapor that might be present
in the gas used as gas ballast
(e.g. water vapor contained in
the atmospheric air when
used as gas ballast)
p
perm
= is the total pressure of all per-
manent gases at the inlet port
of the pump
Eq. (2.3) shows that when using gas
ballast (B 0) vapors can also be pumped
without condensation if no gas is present
at the intake of the pump. The gas ballast
may also be a mixture of non-condensable
gas and condensable vapor as long as the
partial pressure of this vapor (p
vapor, g.b.
) is
less than the saturation pressure p
vapor, sat
of the pumped vapor at the temperature of
the pump.
p p
B
S
(p
p
p p
p
p p
p
vapour sum
Va pour, sat vapour , g.b.
sum vapour , sat
vapour sat
sum vapour sat
perm

+
+
)
p
p
p p
p
vapour
vapour, sat
sum vapour, sat
perm
Water vapor tolerance

An important special case in the general
considerations made above relating to the
topic of vapor tolerance is that of pumping
water vapor. According to PNEUROP water
vapor tolerance is defined as follows:
Water vapor tolerance is the highest
pressure at which a vacuum pump, under
normal ambient temperatures and
pressure conditions (20 C, 1013 mbar),
can continuously take in and transport
pure water vapor. It is quoted in mbar. It
is designated as p
W,O
.
Applying eq. (2.3) to this special case
means:
p
perm
= 0 and p
vapor, sat
= p
s
(H
2
O), thus:
(2.4)
If for the gas ballast gas atmospheric air of
50 % humidity is used, then
p
vapor, g.b.
= 13 mbar; with B/S = 0.10
a usual figure in practice and p
sum
(total exhaust pressure) = 1330 mbar, the
water vapor tolerance p
W,0
as function of
the temperature of the pump is represen-
ted by the lowest curve in diagram Fig.
2.16. The other curves correspond to the
pumping of water vapor-air mixtures,
hence p
perm
= p
air
O), indicated by the
symbol p
L
in millibar. In these cases a hig-
her amount of water vapor partial pressu-
re p
w
can be pumped as shown in the dia-
gram. The figures for p
W,0
given in the
catalogue therefore refer to the lower limit
and are on the safe side.
According to equation 2.4 an increase in
p p
B
S
p H O p
p p H O
W sum O
s vapour, g.b.
sum s
,
( )
( )
=

2
2
the gas ballast B would result in an increa-
sed water vapor tolerance p
W,0
. In prac-
tice, an increase in B, especially in the case
of single-stage gas ballast pumps is
restricted by the fact that the attainable
ultimate vacuum for a gas ballast pump
operated with the gas ballast valve open
becomes worse as the gas ballast B
increases. Similar considerations also
apply to the general equation 2.3 for the
vapor tolerance p
vapor
.
At the beginning of a pump down process,
the gas ballast pump should always be
operated with the gas ballast valve open.
In almost all cases a thin layer of water will
be present on the wall of a vessel, which
only evaporates gradually. In order to
attain low ultimate pressures the gas
ballast valve should only be closed after
the vapor has been pumped out. LEYBOLD
pumps generally offer a water vapor tole-
rance of between 33 and 66 mbar. Two-
stage pumps may offer other levels of
water vapor tolerance corresponding to
the compression ratio between their
stages provided they have pumping
chamber of different sizes.
Other gases as ballast
Generally atmospheric air is used as the
gas ballast medium. In special cases,
when pumping explosive or toxic gases,
for example, other permanent gases like
noble gases or nitrogen, may be used (see
Section 8.3.1.3).
D00
Fig. 2.16 Partial pressure p
W
of water vapor that can
be pumped with the gas ballast valve open
without condensation in the pump, as a fun-
ction of the pump temperature for various
partial pressures p
L
of air. The lowest curve
corresponds to the water vapor tolerance
p
W, O
of the pump
Temperature of the pump
p
W
[
m
b
a
r
]
p
L
[mbar]
D00 E 19.06.2001 21:35 Uhr Seite 23
Vacuum Generation
2.1.3 Dry compressing rotary
displacement pumps
2.1.3.1 Roots pumps
The design principle of the Roots pumps
was already invented in 1848 by Isaiah
Davies, but it was 20 years later before it
was implemented in practice by the
Americans Francis and Philander Roots.
Initially such pumps were used as blowers
for combustion motors. Later, by inverting
the drive arrangement, the principle was
employed in gas meters. Only since 1954
has this principle been employed in
vacuum engineering. Roots pumps are
used in pump combinations together with
backing pumps (rotary vane- and rotary
plunger pumps) and extend their operating
range well into the medium vacuum range.
With two stage Roots pumps this extends
into the high vacuum range. The operating
principle of Roots pumps permits the
assembly of units having very high
pumping speeds (over 100,000 m
3
/h)
which often are more economical to
operate than steam ejector pumps running
in the same operating range.
A Roots vacuum pump (see Fig. 2.17) is a
rotary positive-displacement type of pump
where two symmetrically-shaped impel-
lers rotate inside the pump casing past
each other in close proximity. The two
rotors have a cross section resembling
approximately the shape of a figure 8 and
are synchronized by a toothed gear. The
clearance between the rotors and the
casing wall as well as between the rotors
themselves amounts only to a few tenths
of a millimeter. For this reason Roots
pumps may be operated at high speeds
without mechanical wear. In contrast to
rotary vane and rotary plunger pumps,
Roots pumps are not oil sealed, so that the
internal leakage of dry compressing
pumps by design results in the fact that
compression ratios only in the range 10
100 can be attained. The internal leakage
of Roots pumps, and also other dry
compressing pumps for that matter, is
mainly based on the fact that owing to the
operating principle certain surface areas of
the pump chamber are assigned to the
intake side and the compression side of
the pump in alternating fashion. During the
compression phase these surface areas
(rotors and casing) are loaded with gas
(boundary layer); during the suction phase
this gas is released. The thickness of the
traveling gas layer depends on the
clearance between the two rotors and
between the rotors and the casing wall.
Due to the relatively complex thermal
conditions within the Roots pump it is not
possible to base ones consideration on
the cold state. The smallest clearances and
thus the lowest back flows are attained at
operating pressures in the region of
1 mbar. Subsequently it is possible to
attain in this region the highest com-
pression ratios, but this pressure range is
also most critical in view of contacts
between the rotors and the casing.
Characteristic quantities of roots pumps
The quantity of gas Q
eff
effectively pumped
by a Roots pump is calculated from the
theoretically pumped quantity of gas Qth
and the internal leakage QiR (as the
quantity of gas which is lost) as:
Q
eff
= Q
th
Q
iR
(2.5)
The following applies to the theoretically
pumped quantity of gas:
Q
th
= p
a
S
th
(2.6)
where pa is the intake pressure and Sth is
the theoretical pumping speed. This in turn
is the product of the pumping volume V
S
and the speed n:
S
th
= n V
S
(2.7)
Similarly the internal leakage Q
iR
is
calculated as:
Q
iR
= n V
iR
(2.8)
where p
V
is the forevacuum pressure
(pressure on the forevacuum side) and S
iR
is a (notional) reflow pumping speed
with
S
iR
= n V
iR
(2.9)
i.e. the product of speed n and internal lea-
kage volume V
iR
.
Volumetric efficiency of a Roots pumps is
given by
(2.10)
By using equations 2.5, 2.6, 2.7 and 2.8
one obtains
(2.11)
When designating the compression p
v
/p
a
as k one obtains
(2.11a)
Maximum compression is attained at zero
throughput (see PNEUROP and
DIN 28 426, Part 2). It is designated as k
0
:
(2.12)
k
0
is a characteristic quantity for the Roots
pump which usually is stated as a function
of the forevacuum pressure p
V
(see Fig.
2.18). k
0
also depends (slightly) on the
type of gas.
For the efficiency of the Roots pump, the
generally valid equation applies:
(2.13)
Normally a Roots pump will be operated in
connection with a downstream rough
vacuum pump having a nominal pumping
speed S
V
. The continuity equation gives:
S
V
p
V
= S
eff
p
a
= S
th
p
a
(2.14)
= 1
k
k
o
k
S
S
th
iR
0
0
= = ( )
= 1 k
S
S
iR
th
= 1
p
p
S
S
V
a
iR
th
=
Q
Q
eff
th
Fig. 2.17 Schematic cross section of a Roots pump
1
2 2
3
5
4
1 Intake flange
2 Rotors
3 Chamber
4 Exhaust flange
5 Casing
Fig. 2.18 Maximum compression k
0
of the Roots pump
RUVAC WA 2001 as a function of fore vacu-
um pressure p
V
D00 E 19.06.2001 21:36 Uhr Seite 24
Vacuum Generation
From this
(2.15)
The ratio S
th
/S
V
(theoretical pumping
speed of the Roots pump / pumping speed
of the backing pump) is termed the
gradation k
th
. From (2.15) one obtains
k = k
th
(2.16)
Equation (2.16) implies that the compres-
sion k attainable with a Roots pump must
always be less than the grading k
th
between Roots pump and backing pump
since volumetric efficiency is always < 1.
When combining equations (2.13) and
(2.16) one obtains for the efficiency the
well known expression
(2.17)
The characteristic quantities to be found in
equation 2.17 are only for the combination
of the Roots pump and the backing pump,
namely maximum compression k
0
of the
Roots pump and gradation k
th
between
Roots pump and backing pump.
With the aid of the above equations the
pumping speed curve of a given combi-
nation of Roots pump and backing pump
may be calculated. For this the following
must be known:
a) the theoretical pumping speed of the
Roots pump: S
th
=
+
k
k k
o
th
0
k
p
p
S
S
V
a
th
V
= =
b) the max. compression as a function of
the fore vacuum pressure: k
0
(p
V
)
c) the pumping speed characteristic of the
backing pump S
V
(p
V
)
The way in which the calculation is carried
out can be seen in Table 2.3 giving the data
for the combination of a Roots pump
RUVAC WA 2001 / E 250 (single-stage ro-
tary plunger pump, operated without gas
ballast). In this the following is taken for
S
th
:
S
th
= 2050 2.5 % = 2000 m
3
/h
The method outlined above may also be
applied to arrangements which consist of a
rotary pump as the backing pump and
several Roots pumps connected in series,
for example. Initially one determines in
line with an iteration method the
pumping characteristic of the backing
pump plus the first Roots pump and then
considers this combination as the backing
pump for the second Roots pump and so
on. Of course it is required that the
theoretical pumping speed of all pumps of
the arrangement be known and that the
compression at zero throughput k
0
as a
function of the backing pressure is also
known. As already stated, it depends on
the vacuum process which grading will be
most suitable. It may be an advantage
when backing pump and Roots pump both
have the same pumping speed in the
rough vacuum range.
Power requirement of a roots pump
Compression in a Roots pump is perfor-
med by way of external compression and
is termed as isochoric compression.
Experience shows that the following equa-
tion holds approximately:
N
compression
= S
th
(p
v
p
a
) (2.18)
In order to determine the total power (so-
called shaft output) of the pump, mecha-
nical power losses N
V
(for example in the
bearing seals) must be considered:
N
tot
= N
compression
+ N
V
(2.19)
The power losses summarized in N
V
are
as shown by experience approximately
proportional to S
th
, i.e.:
N
v
= const S
th
(2.20)
Depending on the type of pump and its
design the value of the constant ranges
between 0.5 and 2 Wh / m
3
.
The total power is thus:
N
tot
= S
th
(p
v
p
a
+ const.)
The corresponding numerical value equa-
tion which is useful for calculations is:
N
tot
= S
th
(p
v
p
a
+ const.) 3 10
-2
Watt
(2.21)
with p
v
, p
a
in mbar, S
th
in m
3
/ h and
the constant const. being between
18 and 72 mbar.
D00
Forevacuum Pumping speed k
th
= S
th
/ S
v
k
0
(p
v
) of = k
0
/ k
0
+k
th
Intake pressure
pressure S
v
of the = 2001/S
v
the RUVAC (Volumetric. S
eff
= S
th
p
a
= p
v
S
v
/ S
eff
P
v
E 250 WA 2001 efficiency) (equation 2.14)
100 250 8.0 12.5 0.61 1.220 21
40 250 8.0 18 0.69 1.380 7.2
10 250 8.0 33 0.8 1.600 1.6
5 250 8.0 42 0.84 1.680 0.75
1 250 8.0 41 0.84 1.680 0.15
5 10
1
220 9.1 35 0.79 1.580 7 10
2
1 10
1
120 16.6 23 0.6 1.200 1 10
2
4 10
2
30 67 18 0.21 420 3 10
3

Pumping speed characteristic
for the combination
WA 2001 / E250
Table 2.3
The values taken from the two right-hand columns give point by point the pumping speed curve for
the combination WA 2001/E250 (see Fig. 2.19, topmost curve)
D00 E 19.06.2001 21:36 Uhr Seite 25
Vacuum Generation
Load rating of a roots pump
The amount of power drawn by the pump
determines its temperature. If the tempe-
rature increases over a certain level,
determined by the maximum permissible
pressure difference p
V
p
a
, the danger
exists that the rotors may seize in the
casing due to their thermal expansion. The
maximum permissible pressure difference
p
max
is influenced by the following fac-
tors: forevacuum or compression pressure
p
V
, pumping speed of the backing pump
S
V
, speed of the Roots pump n, gradation
k
th
and the adiabatic exponent of the
pumped gas. p
max
increases when p
V
and S
V
increase and decreases when n and
k
th
increase. Thus the maximum difference
between forevacuum pressure and intake
pressure, p
V
p
a
must during conti-
nuous operation not exceed a certain
value depending on the type of pump.
Such values are in the range between 130
and 50 mbar. However, the maximum
permissible pressure difference for conti-
nuous operation may be exceed for brief
periods. In the case of special designs,
which use gas cooling, for example, high
pressure differences are also permissible
during continuous operation.
Types of motors used with roots pumps
Standard flange-mounted motors are used
as the drive. The shaft feedthroughs are
sealed by two oil sealed radial shaft seals
running on a wear resistant bushing in
order to protect the drive shaft. Flange
motors of any protection class, voltage or
frequency may be used.
Integral leak tightness of this version is
< 10
-4
mbarls
-1
.
In the case of better leak tightness
requirements of < 10
-5
mbarls
-1
the
Roots pump is equipped with a canned
motor. The rotor is seated in the vacuum
on the drive shaft of the pump and is
separated from the stator by a vacuum-
tight non-magnetic tube. The stator coils
are cooled by a fan having its own drive
motor. Thus shaft seals which might be
subject to wear are no longer required. The
use of Roots pumps equipped with canned
motors is especially recommended when
pumping high purity-, toxic- or radioactive
gases and vapors.
Maintaining the allowed pressure
difference
In the case of standard Roots pumps,
measures must be introduced to ensure
that the maximum permissible pressure
difference between intake and exhaust port
due to design constraints is not exceeded.
This is done either by a pressure switch,
which cuts the Roots pump in and out
depending on the intake pressure, or by
using a pressure difference or overflow
valve in the bypass of the Roots pumps
(Fig. 2.20 and 2.21). The use of an
overflow valve in the bypass of the Roots
pump is the better and more reliable
solution. The weight and spring loaded
valve is set to the maximum permissible
Fig. 2.19 Pumping speed curves for different pump combinations with the corresponding backing pumps
Intake pressure p
a

P
u
m
p
i
n
g

s
p
e
e
d

S
WAU 2001
SOGEVAC SV 1200
Fig. 2.20 Cross section of a Roots pumps with
bypass line
Fig. 2.21 Vacuum diagram Roots pump with integra-
ted bypass line and backing pump
D00 E 19.06.2001 21:36 Uhr Seite 26
Vacuum Generation
pressure difference of the particular pump.
This ensures that the Roots pump is not
overloaded and that it may be operated in
any pressure range. In practice this means
that the Roots pump can be switched on,
together with the backing pump, at
atmospheric pressure. In the process any
pressure increases will not adversely affect
combined operation, i.e. the Roots pump
is not switched off in such circumstances.
Pre-admission cooling (Fig. 2.22)
In the case of Roots pumps with pre-
admission cooling, the compression pro-
cess basically is the same as that of a nor-
mal Roots pump. Since greater pressure
differences are allowed more installed
power is needed, which at the given speed
and the pressure difference between inlet
and discharge port is directly proportional
and is composed of the theoretical work
done on compression and various power
losses. The compression process ends
normally after opening of the pumping
chamber in the direction of the discharge
port. At this moment warmed gas at higher
pressure flows into the pumping chamber
and compresses the transported volume of
gas. This compression process is perfor-
med in advance in the case of pre-admis-
sion cooling. Before the rotor opens the
pumping chamber in the direction of the
discharge port, compressed and cooled
gas flows into the pumping chamber via
the pre-admission channel. Finally the
rotors eject the pumped medium via the
discharge port. The cooled gas, which in
the case of single-stage compression is
taken from the atmosphere and admitted
from the pre-admission cooler, and which
in the case of multi-stage pump systems is
taken from downstream gas coolers,
performs a pre-compression and removes
by inner cooling the heat of compression
at the point of time it occurs.
2.1.3.2 Claw pumps
Like Roots pumps, claw pumps belong to
the group of dry compressing rotary
piston vacuum pumps (or rotary vacuum
pumps). These pumps may have several
stages; their rotors have the shape of
claws.
The design principle of a claw pump is
explained by first using an example of a
four-stage design. The cross section inside
the pumps casing has the shape of two
partly overlapping cylinders (Fig. 2.23).
Within these cylinders there are two freely
rotating rotors in each pump stage: (1)
with their claws and the matching recesses
rotating in opposing directions about their
vertical axes. The rotors are synchronized
by a gear just like a Roots pump. In order
to attain an optimum seal, the clearance
between the rotor at the center of the
casing and the amount of clearance with
respect to the inside casing wall is very
small; both are in the order of magnitude
of a few 0.01 mm. The rotors periodically
open and close the intake and discharge
slots (5) and (4). At the beginning of the
work cycle in position a, the right rotor just
opens the intake slot (5). Gas now flows
into the continually increasing intake space
(3) in position b until the right rotor seals
off the intake slot (5) in position c. After
both claws have passed through the center
position, the gas which has entered is then
compressed in the compression chamber
(2) (position a) so long until the left rotor
releases the discharge slot (4) (position b)
thereby discharging the gas. Immediately
after the compression process has started
(position a) the intake slot (5) is opened
simultaneously and gas again flows into
the forming intake space (3) (position b).
Influx and discharge of the gas is
performed during two half periods. Each
rotor turns twice during a full work cycle.
Located between the pumping stages are
intermediate discs with flow channels
which run from the discharge side of the
upper stage to the intake side of the next
stage, so that all inlet or exhaust sides are
arranged vertically above each other (Fig.
2.24). Whereas in a Roots pump the in-
coming gas is pumped through the pump
at a constant volume and compression is
only performed in the forevacuum line
(see Section 2.1.3.1), the claw pump
compresses the gas already within the
pumping chamber until the rotor releases
the discharge slot. Shown in Fig. 2.25 are
the average pressure conditions in the
individual pumping stages of a DRYVAC at
an intake pressure of 1 mbar. In order to
meet widely differing requirements
LEYBOLD manufactures two different
D00
Fig. 2.22 Diagram of a Roots pump with pre-admis-
sion cooling
1 Intake port
2 Discharge port
3 Gas cooler
4 Flow of cold gas
1
2
3
4
4
Fig. 2.23 Principle of operation
a
b
c
1 Rotors
2 Compression
chamber
3 Intake space
4 Exhaust slot
5 Intake slot
6 Intermediate stage purge
gas
P
Z
P
Z
P
Z
P
Fig. 2.24 Arrangement of the pumps and guiding of
the gas flow. P = Pump stage Z = Interme-
diate ring
D00 E 19.06.2001 21:36 Uhr Seite 27
Vacuum Generation
series of claw pumps, which chiefly differ
in the type of compression process used:
1) Pumps with internal compression,
multi-stage for the semiconductor
industry (DRYVAC Series) and
2) Pumps without internal compression,
two-stage for chemistry applications
(ALLex).
Figs. 2.26 and 2.27 demonstrate the
differences in design. Shown is the course
of the pressure as a function of the volume
of the pumping chamber by way of a pV
diagram.
Fig. 2.26 shows the (polytropic) course of
the compression for pumps with internal
compression. The pressure increases until
the discharge slot is opened. If at that
point the exhaust back pressure has not
been reached, the compression space is
suddenly vented with hot exhaust gas. As
the volume is reduced further, the gas at
exhaust pressure is ejected. The work
done on compression is represented by
the area under the p-v curve 1-2-3-4. It is
almost completely converted into heat. In
the case of dry compressing vacuum
pumps not much of this heat can be lost to
the cooled casing due to the low density of
the gas. This results in high gas tempe-
ratures within the pump. Experiences with
claw pumps show that the highest tem-
peratures occur at the rotors.
Shown in Fig. 2.27 is the principle of
isochoric compression in a p-v diagram.
Here the compression is not performed by
reducing the volume of the pumping
chamber, but by venting with cold gas
which is applied from the outside after
completion of the intake process. This is
similar to the admission of a gas ballast
when opening the gas ballast valve after
completion of the intake phase. From the
diagram it is apparent that in the case of
isochoric compression the work done on
compression must be increased, but cold
gas instead of hot exhaust gas is used for
venting. This method of direct gas cooling
results in considerably reduced rotor
temperatures. Pumps of this kind are dis-
cussed as ALLex in Section 2.1.3.2.2.
2.1.3.2.1 Claw pumps with internal
compression for the
semiconductor industry
(DRYVAC series)
Design of DRYVAC Pumps
Due to the work done on compression in
the individual pumping stages, multi-stage
claw pumps require water cooling for the
four stages to remove the compression
heat. Whereas the pumping chamber of
the pump is free of sealants and lubri-
cants, the gear and the lower pump shaft
are lubricated with perfluoropolyether
(PFPE). The gear box is virtually herme-
tically sealed from the pumping chamber
by piston rings and a radial shaft seal. The
bearings in the upper end disk are
lubricated with PFPE grease. In order to
protect the bearings and shaft seals
against aggressive substances, a barrier
gas facility is provided. A controlled water
cooling system allows the control of the
casing temperature over a wide range as
the pump is subjected to differing gas
loads coming from the process. The four
stage design is available in several
pumping speed and equipment grades of
25, 50 and 100 m
3
/h DRYVAC pumps:
a) as the basic version for non-aggressive
clean processes: DRYVAC 25 B, 50 B
and 100 B (Fig. 2.29a)
b) as a version for semiconductor
processes: DRYVAC 25 P, 50 P and
100 P (Fig. 2.29b)
c) as a system version with integrated self
monitoring: DRYVAC 50 S and 100 S
d) as a system version with integrated self
Fig. 2.25 Pressures in pump stages 1 to 4
Stage 1
Stage 2
Stage 3
Stage 4
IN
IN
IN
IN
OUT
OUT
OUT
OUT
1 mbar
3 mbar
150 mbar
15 mbar
3 mbar
15 mbar
150 mbar
1000 mbar
P
V
W
Compr.
Fig. 2.26 Compression curve for a claw pump with
internal compression
1
2
3
4
P
V
W
Compr.
Fig. 2.27 Compression curve for a claw pump without
internal compression
(isochoric compression)
1
2
3
Fig. 2.28 DRYVAC pump
1 Intake port
2 Front panel / electronics
3 Main switch
1
2
3
D00 E 19.06.2001 21:36 Uhr Seite 28
Vacuum Generation
monitoring offering an increased
pumping speed in the lower pressure
range: DRYVAC 251 S and 501 S (Fig.
2.29c)
The ultimate pressure attainable with the
DRYVAC 251 S or 501 S is compared to
the versions without integrated Roots
pump by approximately one order of
magnitude lower (from 210
-2
mbar to
310
-3
mbar) and the attainable throughput
is also considerably increased. It is of
course possible to directly flange mount
LEYBOLD RUVAC pumps on to the
DRYVAC models (in the case of semi-
conductor processes also mostly with a
PFPE oil filling for the bearing chambers).
The pumps of the DRYVAC family are the
classic dry compressing claw vacuum
pumps that are preferably used in the se-
miconductor industry, whereby the pumps
need to meet a variety of special require-
ments. In semiconductor processes, as in
many other vacuum applications, the
formation of particles and dusts during the
process and/or in the course of compres-
sing the pumped substances to atmos-
pheric pressure within the pump, is una-
voidable. In the case of vacuum pumps
operating on the claw principle it is
possible to convey particles through the
pump by means of so called pneumatic
conveying. This prevents the deposition
of particles and thus the formation of
layers within the pump and reduces the
risk that the claw rotor may seize. Care
must be taken to ensure that the velocity of
the gas flow within the individual pumping
stages is at all times greater than the
settling speed of the particles entrained in
the gas flow. As can be seen in Fig. 2.31,
the settling speed of the particles depends
strongly on their size. The mean velocity
(V
Gas
) of the flowing gas during the
compression phase is given by the
following equation:
(2.22)
q
pV
= gas throughput
p = pressure
A = surface area
One can see that with increasing pressure
the velocity of the pumped gas flow slows
down and attains the order of magnitude
of the settling speed of the particles in the
gas flow (Fig. 2.32). This means that the
risk of depositing particles in the operating
chamber of the pump and the resulting
v
Gas
q
p A
mbar s
mbar cm
q
p A
m
s
pV
pV

1
2
10
D00
Fig. 2.29a Vacuum diagram for the DRYVAC B
Cooling
water
Exhaust
line
Intake
line
Casing
suction
Fig. 2.29b Vacuum diagram for the DRYVAC P
Inert
gas
Cooling
water
Exhaust
line
Intake
line
Casing
suction
100 P only
Fig. 2.29c Vacuum diagram for the DRYVAC S
TSH
PSL
PSH
FSL
MPS
PT 100
CS
EPS
To be provided by the customer
Temperature switch
Pressure switch
Pressure switch
Flow switch
Motor protection switch
Temperature sensor
For DRYVAC with LIMS
Current sensor
Exhaust pressure sensor
Fig. 2.30 Key to Figures 2.29a 2.29c
Fig. 2.31 Settling speed as a function of pressure p.
Parameter: particle size
Pressure
Particle size
L
i
m
i
t

s
p
e
e
d
Fig. 2.32 Mean gas velocity v
g
during compression
without purge gas (left) and with purge gas
(right) in stages 2, 3 and 4
Throughput cross sec-
tion 6.5 cm
2
, constant
Throughput in
mbarl/s
Stage 2
2 500 mbarl/s
Stage 3
8 300 mbarl/s
Stage 4
20 000 mbarl/s
G
a
s

v
e
l
o
c
i
t
y
Pressure
D00 E 19.06.2001 21:36 Uhr Seite 29
Vacuum Generation
impairment increases with increasing
pressure. In parallel to this the potential
for the formation of particles from the
gaseous phase increases at increasing
compression levels. In order to keep the
size of the forming particles small and thus
their settling speed low and to maintain a
high velocity for the gas, an additional
quantity of gas is supplied into the pump
via the individual intermediate discs (purge
gas). For this, the inflowing quantity of
purge gas is matched to the pressure
conditions prevailing in the individual
pumping stages (see top right part of Fig.
2.32). This keeps the velocity of the gas
flow high enough within the entire pump
by so-called pneumatic pumping. Through
the way in which the gas is lead within the
pump, i.e. from the intake through the four
pumping stages with the related intermedi-
ate discs to the exhaust, it is possible to
reduce the influence of the purge gas on
the ultimate pressure to a minimum. Test
results (Fig. 2.33) indicate that the in-
fluence of purge gas in the fourth stage is
as to be expected of the lowest level
since there are located between this stage
and the intake side the three other pum-
ping stages. The admission of purge gas
via the second and third stages (Fig. 2.33)
has a comparatively small influence on the
ultimate pressure as can be seen from the
pumping speed curve in Fig. 2.34. Finally it
can be said that the formation of particles
is to be expected in most CVD processes.
When using dry compressing claws
vacuum pumps, the controlled admission
of purge gas via the individual intermediate
discs is the best approach to avoid the
formation of layers. When applying this
method several effects can be noted:
The admitted purge gas dilutes the
pumped mixture of substances, par-
ticle-forming reactions will not occur, or
are at least delayed.
The risk of an explosion through self-
igniting substances is significantly
reduced.
Particles which have formed are con-
veyed pneumatically through the pump
Losses in pumping speed and a reduc-
tion in ultimate pressure can be kept
very small due to the special way in
which the gas is made to pass through
the pump.
Fig. 2.33 Ultimate pressure of the DRYVAC 100S as a function of pure gas flow
in stages 2 4
mbar
U
l
t
i
m
a
t
e

p
r
e
s
s
u
r
e
Fig. 2.34 Pumping speed with and without purge gas
Stage 2
Stage 3
Stage 4
Pressure
P
u
m
p
i
n
g

s
p
e
e
d
2 500 mbar l/s
8 300 mbar l/s
20 000 mbar l/s
Fig. 2.35 Simple arrangement of the dry compression ALLex pump
Intake port
Claws
1st stage
2nd stage
Axial face seals
Motor
Coupling
Gear, complete
with shafts and
bearings
Purge gas flow
D00 E 19.06.2001 21:36 Uhr Seite 30
Vacuum Generation
2.1.3.2.2 Claw pumps without internal
compression for chemistry
applications (ALLex )
The chemical industry requires vacuum
pumps which are highly reliable and which
do not produce waste materials such as
contaminated waste oil or waste water. If
this can be done, the operating costs of
such a vacuum pump are low in view of
the measures otherwise required for pro-
tecting the environment (disposal of waste
oil and water, for example). For operation
of the simple and rugged ALLex pump
from LEYBOLD there are no restrictions as
to the vapor flow or the pressure range
during continuous operation. The ALLex
may be operated within the entire pressure
range from 5 to 1000 mbar without
restrictions.
Design of the ALLex pump
The design of the two-stage ALLex is
shown schematically in Fig. 2.35. The gas
flows from top to bottom through the
vertically arranged pumping stages in
order to facilitate the ejection of con-
densates and rinsing liquids which may
have formed. The casing of the pump is
water cooled and permits cooling of the
first stage. There is no sealed link between
gas chamber and cooling channel so that
the entry of cooling water into the
pumping chamber can be excluded. The
pressure-burst resistant design of the
entire unit underlines the safety concept in
view of protection against internal explo-
sions, something which was also taken
into account by direct cooling with cold
gas (see operating principle). A special
feature of the ALLex is that both shafts
have their bearings exclusively in the gear.
On the pumping side, the shafts are free
(cantilevered). This simple design allows
the user to quickly disassemble the pump
for cleaning and servicing without the
need for special tools.
In order to ensure a proper seal against
the process medium in the pumping
chamber the shaft seal is of the axial face
seal type a sealing concept well proven
in chemistry applications. This type of seal
is capable of sealing liquids against
liquids, so that the pump becomes
rinseable and insensitive to forming
condensate. Fig. 2.36 shows the compo-
nents supplied with the ALLex, together
with a gas cooler and a receiver.
Operating principle
Isochoric compression, which also serves
the purpose of limiting the temperature
ultimately attained during compression,
especially in the stage on the side of the
atmosphere, and which ensures protection
against internal explosions, is performed
by venting the pumping chamber with cold
gas from a closed refrigerating gas cycle
(Fig. 2.37). Fig. 2.38-1 indicates the start of
the intake process by opening the intake
slot through the control edge of the right
rotor. The process gas then flows into the
intake chamber which increases in size. The
intake process is caused by the pressure
gradient produced by increasing the
volume of the pumping chamber. The
maximum volume is attained after 3/4 of a
revolution of the rotors (Fig. 2.38-2). After
the end of the intake process, the control
edge of the left rotor opens the cold gas
inlet and at the same time the control edge
of the right rotor opens the intake slot (Fig.
2.38-3) once more. In Fig. 2.38-4 the
control edge of the left rotor terminates the
discharge of the gas which has been
compressed to 1000 mbar with the cold
gas; at the same time the control edge of
the right rotor completes an intake process
again.
The total emissions from the system are
D00
Fig. 2.37 Circulation of the cold gas in the ALLex
with cooler / condenser
Cooling gas
Exhaust gas
Fig. 2.36 ALLex pump
3
4
5
6
1
2
1 Motor
2 Pump
3 Intake port
4 Exhaust port
5 Exhaust cooler
6 Cooling water
connection
7 Cooling receiver
7
m h
3 -1
.
2
10
2
8 6 4
4
6
8
10 1
1
100
1000
Ansaugdruck
mbar
100 1000
S
a
u
g
v
e
r
m
g
e
n
Fig. 2.39 Pumping speed characteristic of an ALLex 250
Intake pressure
P
u
m
p
i
n
g

s
p
e
e
d
D00 E 19.06.2001 21:36 Uhr Seite 31
Vacuum Generation
not increased by the large quantities of cold
gas, since a closed refrigerating cycle is
maintained by way of an externally arran-
ged gas cooler and condenser (Fig. 2.37).
The hot exhaust gas is made to pass
through the cooler and is partly returned in
the form of cold gas for pre-admission
cooling into the pump. The pump takes in
the quantity of cold process gas needed for
venting the pumping chamber back into the
compression space on its own. This pro-
cess, however, has no influence on the
pumping speed of the ALLex because the
intake process has already ended when the
venting process starts. Designing the
cooler as a condenser allows for simple
solvent recovery. The method of direct gas
cooling, i.e. venting of the pumping cham-
ber with cold gas supplied from outside
(instead of hot exhaust gas) results in the
case of the ALLex in rotor temperatures
which are so low that mixtures of substa-
nces rated as ExT3 can be pumped reliably
under all operating conditions. The
ALLex thus fully meets the requirements
of the chemical industry concerning the
protection against internal explosions. A
certain degree of liquid compatibility makes
the ALLex rinseable, thus avoiding the
formation of layers in the pump, for examp-
le, or the capability of dissolving layers
which may already have formed respective-
ly. The rinsing liquids are usually applied to
the pump after completion of the connec-
ted process (batch operation) or while the
process is in progress during brief blocking
phases. Even while the ALLex is at
standstill and while the pumping chamber
is completely filled with liquid it is possible
to start this pump up. Shown in Fig. 2.39 is
the pumping speed characteristic of an
ALLex 250. This pump has a nominal
pumping speed of 250 m
3
/h and an
ultimate pressure of < 10 mbar. At
10 mbar it still has a pumping speed of
100 m
3
/h. The continuous operating
pressure of the pump may be as high as
1000 mbar; it consumes 13.5 kW of
electric power.
V
max
P
P
P
P
V
min
100 1000
(10) (100)
V
max
V
min
100 1000
(10) (100)
V
max
V
min
100 1000
(10) (100)
V
max
V
min
100 1000
(10) (100)
Cold gas inlet
Exhaust slot
Exhaust slot
Cold gas inlet
Cold gas inlet
Beginning of admitting cold gas
Intake
slot
Volume of the pump
chamber starts to increase
Suction
Volume of the pump
chamber at maximum
End of suction
Volume of the pump chamber
stars to decrease (without
compression). Pressure
increase to 1000 mbar
only by admitting cold gas.
Ejection of the mixture
composed of sucked
in gas and cold gas.
mbar
mbar
mbar
mbar
4
3
2
1
Fig. 2.38 Diagrams illustrating the pumping principle of the ALLex pump (claw pump without inner compression)
D00 E 19.06.2001 21:36 Uhr Seite 32
Vacuum Generation
2.1.4 Accessories for
oil-sealed rotary
displacement pumps
During a vacuum process, substances
harmful to rotary pumps can be present in
a vacuum chamber.
Elimination of water vapor
Water vapor arises in wet vacuum proces-
ses. This can cause water to be deposited
in the inlet line. If this condensate reaches
the inlet port of the pump, contamination
of the pump oil can result. The pumping
performance of oil-sealed pumps can be
significantly impaired in this way. More-
over, water vapor discharged through the
outlet valve of the pump can condense in
the discharge outlet line. The condensate
can, if the outlet line is not correctly ar-
ranged, run down and reach the interior of
the pump through the discharge outlet
valve. Therefore, in the presence of water
vapor and other vapors, the use of
condensate traps is strongly recommen-
ded. If no discharge outlet line is connec-
ted to the gas ballast pump (e.g., with
smaller rotary vane pumps), the use of
discharge filters is recommended. These
catch the oil mist discharged from the
pump.
Some pumps have easily exchangeable fil-
ter cartridges that not only hold back oil
mist, but clean the circulating pump oil.
Whenever the amount of water vapor pre-
sent is greater than the water vapor tole-
rance of the pump, a condenser should
always be installed between the vessel and
the pump. (For further details, see Section
2.1.5)
Elimination of dust
Solid impurities, such as dust and grit,
significantly increase the wear on the
pistons and the surfaces in the interior of
the pump housing. If there is a danger that
such impurities can enter the pump, a dust
separator or a dust filter should be in-
stalled in the inlet line of the pump. Today
not only conventional filters having fairly
large casings and matching filter inserts
are available, but also fine mesh filters
which are mounted in the centering ring of
the small flange. If required, it is recom-
mended to widen the cross section with KF
adaptors.
Elimination of oil vapor
The attainable ultimate pressure with oil-
sealed rotary pumps is strongly influenced
by water vapor and hydrocarbons from the
pump oil. Even with two-stage rotary vane
pumps, a small amount of back-streaming
of these molecules from the pump interior
into the vacuum chamber cannot be
avoided. For the production of hydrocar-
bon-free high and ultrahigh vacuum, for
example, with sputter-ion or turbomolecu-
lar pumps, a vacuum as free as possible of
oil is also necessary on the forevacuum
side of these pumps. To obtain this, me-
dium vacuum adsorption traps (see Fig.
2.40) filled with a suitable adsorption
material (e.g., LINDE molecular sieve 13X)
are installed in the inlet line of such oil-
sealed forepumps. The mode of action of a
sorption trap is similar to that of an ad-
sorption pump. For further details, see
Section 2.1.8. If foreline adsorption traps
are installed in the inlet line of oil-sealed
rotary vane pumps in continuous opera-
tion, two adsorption traps in parallel are
recommended, each separated by valves.
Experience shows that the zeolite used as
the adsorption material loses much of its
adsorption capacity after about 10 14
days of running time, after which the
other, now-regenerated, adsorption trap
can be utilized; hence the process can
continue uninterrupted. By heating the
adsorption trap, which is now not connec-
ted in the pumping line, the vapors
escaping from the surface of the zeolite
can be most conveniently pumped away
with an auxiliary pump. In operation, pum-
ping by the gas ballast pump generally
leads to a covering of the zeolite in the
other, unheated adsorption trap and thus
to a premature reduction of the adsorption
capacity of this trap.
Reduction of the effective pumping
speed
All filters, separators, condensers, and
valves in the inlet line reduce the effective
pumping speed of the pump. On the basis
of the values of the conductances or
resistances normally supplied by
manufacturers, the actual pumping speed
of the pump can be calculated. For further
details, see Section 1.5.2.
2.1.5 Condensers
For pumping larger quantities of water
vapor, the condenser is the most econo-
mical pump. As a rule, the condenser is
cooled with water of such temperature that
the condenser temperature lies sufficiently
below the dew point of the water vapor and
an economical condensation or pumping
action is guaranteed. For cooling, however,
media such as brine and refrigerants (NH
3
,
Freon) can also be used.
When pumping water vapor in a large
industrial plant, a certain quantity of air is
always involved, which is either contained
in the vapor or originates from leaks in the
plant (the following considerations for air
and water vapor obviously apply also in
general for vapors other than water vapor).
Therefore, the condenser must be backed
D00
Fig. 2.40 Cross section of a medium vacuum
adsorption trap
1 Housing
2 Basket holding the sieve
3 Molecular sieve (filling)
4 Sealing flanges
5 Intake port with small flange
6 Upper section
7 Vessel for the heater or refrigerant
8 Connection on the side of the pump with
small flange
D00 E 19.06.2001 21:36 Uhr Seite 33
Vacuum Generation
by a gas ballast pump (see Fig. 2.41) and
hence always works like the Roots pump
in a combination. The gas ballast pump
has the function of pumping the fraction of
air, which is often only a small part of the
water-vapor mixture concerned, without
simultaneously pumping much water
vapor. It is, therefore, understandable that,
within the combination of condenser and
gas ballast pump in the stationary con-
dition, the ratios of flow, which occur in the
region of rough vacuum, are not easily
assessed without further consideration.
The simple application of the continuity
equation is not adequate because one is no
longer concerned with a source or sink-free
field of flow (the condenser is, on the basis
of condensation processes, a sink). This is
emphasized especially at this point. In a
practical case of non-functioning of the
condenser gas ballast pump combination,
it might be unjustifiable to blame the
condenser for the failure.
In sizing the combination of condenser
and gas ballast pump, the following points
must be considered:
a) the fraction of permanent gases (air)
pumped simultaneously with the water
vapor should not be too great. At partial
pressures of air that are more than
about 5 % of the total pressure at the
exit of the condenser, a marked accu-
mulation of air is produced in front of
the condenser surfaces. The condenser
then cannot reach its full capacity (See
also the account in Section 2.2.3 on the
simultaneous pumping of gases and
vapors).
b) the water vapor pressure at the
condenser exit that is, at the inlet side
of the gas ballast pump should not
(when the quantity of permanent gas
described in more detail in Section
2.2.3 is not pumped simultaneously) be
greater than the water vapor tolerance
for the gas ballast pump involved. If
as cannot always be avoided in practice
a higher water vapor partial pressure
is to be expected at the condenser exit,
it is convenient to insert a throttle bet-
ween the condenser exit and the inlet
port of the gas ballast pump. The
conductance of this throttle should be
variable and regulated (see Section
1.5.2) so that, with full throttling, the
pressure at the inlet port of the gas
ballast pump cannot become higher
than the water vapor tolerance. Also,
the use of other refrigerants or a de-
crease of the cooling water temperature
may often permit the water vapor pres-
sure to fall below the required value.
For a mathematical evaluation of the
combination of condenser and gas ballast
pump, it can be assumed that no loss of
pressure occurs in the condenser, that the
total pressure at the condenser entrance
ptot 1, is equal to the total pressure at the
condenser exit, p
tot 2
:
p
tot 1
= p
tot 2
(2.23)
The total pressure consists of the sum of
the partial pressure portions of the air p
p
and the water vapor p
v
:
p
p1
+ p
v1
= p
p2
+ p
v2
(2.23a)
As a consequence of the action of the
condenser, the water vapor pressure p
D2
at
the exit of the condenser is always lower
than that at the entrance; for (2.23) to be
fulfilled, the partial pressure of air p
p2
at
the exit must be higher than at the
entrance p
p1
, (see Fig. 2.43), even when
no throttle is present.
The higher air partial pressure p
p2
at the
condenser exit is produced by an accumu-
lation of air, which, as long as it is present
at the exit, results in a stationary flow
equilibrium. From this accumulation of air,
the (eventually throttled) gas ballast pump
in equilibrium removes just so much as
streams from the entrance (1) through the
condenser.
All calculations are based on (2.23a) for
which, however, information on the
quantity of pumped vapors and permanent
gases, the composition, and the pressure
should be available. The size of the
condenser and gas ballast pump can be
calculated, where these two quantities are,
indeed, not mutually independent. Fig. 2.42
represents the result of such a calculation
as an example of a condenser having a
condensation surface of 1 m
2
, and at an
inlet pressure p
v1
, of 40 mbar, a
condensation capacity that amounts to
15 kg / h of pure water vapor if the fraction
of the permanent gases is very small. 1 m
3
of cooling water is used per hour, at a line
overpressure of 3 bar and a temperature of
12 C. The necessary pumping speed of the
gas ballast pump depends on the existing
operating conditions, particularly the size of
the condenser. Depending on the efficiency
Fig. 2.42 Condensation capacity of the condenser
(surface area available to condensation 1 m
2
)
as a function of intake pressure p
D1
of the
water vapor. Curve a: Cooling water tempera-
ture 12 C. Curve b: Temperature 25 C.
Consumption in both cases 1 m
3
/h at 3 bar
overpressure
Intake pressure p
D1
C
o
n
d
e
n
s
a
t
i
o
n

c
a
p
a
c
i
t
y

[
k
g

h
-
1
]
Fig. 2.43 Schematic representation of the pressure dis-
tribution in the condenser. The full lines corre-
spond to the conditions in a condenser in
which a small pressure drop takes place
(p
tot 2
< p
tot 1
). The dashed lines are those for
an ideal condenser (p
tot 2
p
tot 1
). p
D
: Partial
pressure of the water vapor, p
L
: Partial pressu-
re of the air
P
v1
P
v2
P
p1
P
v2
P
p2
P
p2
Fig. 2.41 Condenser (I) with downstream gas ballast
pump (II) for pumping of large quantities of
water vapor in the rough vacuum range (III)
adjustable throttle
1 Inlet of the condenser
2 Discharge of the condenser
3 See text
D00 E 19.06.2001 21:36 Uhr Seite 34
Vacuum Generation
of the condenser, the water vapor partial
pressure p
v2
lies more or less above the
saturation pressure p
S
which corresponds
to the temperature of the refrigerant. (By
cooling with water at 12 C, p
S
, would be
15 mbar (see Table XIII in Section 9)).
Correspondingly, the partial air pressure
p
p2
that prevails at the condenser exit also
varies. With a large condenser,
p
v2
p
S
, the air partial pressure p
p,2
is thus
large, and because p
p
V = const, the
volume of air involved is small. Therefore,
only a relatively small gas ballast pump is
necessary. However, if the condenser is
small, the opposite case arises:
p
v2
> p
S
p
p2
, is small. Here a relatively
large gas ballast pump is required. Since
the quantity of air involved during a
pumping process that uses condensers is
not necessarily constant but alternates
within more or less wide limits, the
considerations to be made are more
difficult. Therefore, it is necessary that the
pumping speed of the gas ballast pump
effective at the condenser can be regulated
within certain limits.
In practice, the following measures are
usual:
a) A throttle section is placed between the
gas ballast pump and the condenser,
which can be short-circuited during
rough pumping. The flow resistance of
the throttle section must be adjustable
so that the effective speed of the pump
can be reduced to the required value.
This value can be calculated using the
equations given in Section 2.2.3.
b) Next to the large pump for rough pum-
ping a holding pump with low speed is
installed, which is of a size corres-
ponding to the minimum prevailing gas
quantity. The objective of this holding
pump is merely to maintain optimum
operating pressure during the process.
c) The necessary quantity of air is ad-
mitted into the inlet line of the pump
through a variable-leak valve. This
additional quantity of air acts like an
enlarged gas ballast, increasing the
water vapor tolerance of the pump.
However, this measure usually results
in reduced condenser capacity. More-
over, the additional admitted quantity of
air means additional power consump-
tion and (see Section 8.3.1.1) increased
oil consumption. As the efficiency of the
condenser deteriorates with too great a
partial pressure of air in the condenser,
the admission of air should not be in
front, but generally only behind the
condenser.
If the starting time of a process is shorter
than the total running time, technically the
simplest method the roughing and the
holding pump is used. Processes with
strongly varying conditions require an
adjustable throttle section and, if needed,
an adjustable air admittance.
On the inlet side of the gas ballast pump a
water vapor partial pressure p
v2
is always
present, which is at least as large as the
saturation vapor pressure of water at the
coolant temperature. This ideal case is
realizable in practice only with a very large
condenser (see above).
With a view to practice and from the stated
fundamental rules, consider the two
following cases:
1. Pumping of permanent gases with
small amounts of water vapor. Here the
size of the condenser gas ballast pump
combination is decided on the basis of the
pumped-off permanent gas quantity. The
condenser function is merely to reduce the
water vapor pressure at the inlet port of
the gas ballast pump to a value below the
water vapor tolerance.
2. Pumping of water vapor with small
amounts of permanent gases. Here, to
make the condenser highly effective, as
small as possible a partial pressure of the
permanent gases in the condenser is
sought. Even if the water vapor partial
pressure in the condenser should be
greater than the water vapor tolerance of
the gas ballast pump, a relatively small gas
ballast pump is, in general, sufficient with
the then required throttling to pump away
the prevailing permanent gases.
Important note: During the process, if the
pressure in the condenser drops below the
saturation vapor pressure of the conden-
sate (dependent on the cooling water
temperature), the condenser must be
blocked out or at least the collected
condensate isolated. If this is not done, the
gas ballast pump again will pump out the
vapor previously condensed in the
condenser
2.1.6 Fluid-entrainment pumps
Basically, a distinction is made between
ejector pumps such as water jet pumps
(17 mbar < p < 1013 mbar), vapor ejector
vacuum pumps
(10
-3
mbar < p < 10
-1
mbar) and diffusion
pumps (p < 10
-3
mbar). Ejector vacuum
pumps are used mainly for the production
of medium vacuum. Diffusion pumps
produce high and ultrahigh vacuum. Both
types operate with a fast-moving stream of
pump fluid in vapor or liquid form (water
jet as well as water vapor, oil or mercury
vapor). The pumping mechanism of all
fluid-entrainment pumps is basically the
same. The pumped gas molecules are
removed from the vessel and enter into the
pump fluid stream which expands after
passing through a nozzle. The molecules
of the pump fluid stream transfer by way
of impact impulses to the gas molecules in
the direction of the flow. Thus the gas
which is to be pumped is moved to a space
having a higher pressure.
In fluid-entrainment pumps corresponding
vapor pressures arise during operation
depending on the type of pump fluid and
the temperature as well as the design of
the nozzle. In the case of oil diffusion
pumps this may amount to 1 mbar in the
boiling chamber. The backing pressure in
the pump must be low enough to allow the
vapor to flow out. To ensure this, such
pumps require corresponding backing
pumps, mostly of the mechanical type. The
vapor jet cannot enter the vessel since it
condenses at the cooled outer walls of the
pump after having been ejected through
the nozzle.
Wolfgang Gaede was the first to realize
that gases at comparatively low pressure
can be pumped with the aid of a pump
fluid stream of essentially higher pressure
and that, therefore, the gas molecules
from a region of low total pressure move
into a region of high total pressure. This
apparently paradoxical state of affairs
develops as the vapor stream is initially
entirely free of gas, so that the gases from
a region of higher partial gas pressure (the
vessel) can diffuse into a region of lower
partial gas pressure (the vapor stream).
This basic Gaede concept was used by
Langmuir (1915) in the construction of the
first modern diffusion pump. The first
diffusion pumps were mercury diffusion
pumps made of glass, later of metal. In the
Sixties, mercury as the medium was
D00
D00 E 19.06.2001 21:36 Uhr Seite 35
Vacuum Generation
almost completely replaced by oil. To
obtain as high a vapor stream velocity as
possible, he allowed the vapor stream to
emanate from a nozzle with supersonic
speed. The pump fluid vapor, which con-
stitutes the vapor jet, is condensed at the
cooled wall of the pump housing, whereas
the transported gas is further compressed,
usually in one or more succeeding stages,
before it is removed by the backing pump.
The compression ratios, which can be
obtained with fluid entrainment pumps,
are very high: if there is a pressure of 10
-9
mbar at the inlet port of the fluid-
entrainment pump and a backing pressure
of 10
-2
mbar, the pumped gas is com-
pressed by a factor of 10
7
!
Basically the ultimate pressure of fluid
entrainment pumps is restricted by the
value for the partial pressure of the fluid
used at the operating temperature of the
pump. In practice one tries to improve this
by introducing baffles or cold traps. These
are condensers between fluid entrain-
ment pump and vacuum chamber, so that
the ultimate pressure which can be
attained in the vacuum chamber is now
only limited by the partial pressure of the
fluid at the temperature of the baffle.
The various types of fluid entrainment
pumps are essentially distinguished by the
density of the pump fluid at the exit of the
top nozzle facing the high vacuum side of
the pump:
1. Low vapor density:
Diffusion pumps
Oil diffusion pumps
(Series: LEYBODIFF, DI and DIP)
Mercury diffusion pumps
2. High vapor density:
Vapor jet pumps
Water vapor pumps
Oil vapor jet pumps
Mercury vapor jet pumps
3. Combined
oil diffusion/ vapor jet pumps
4. Water jet pumps
Cooling of fluid entrainment pumps
The heater power that is continuously
supplied for vaporizing the pump fluid in
fluid-entrainment pumps must be dissipa-
ted by efficient cooling. The energy
required for pumping the gases and
vapors is minimal. The outside walls of the
casing of diffusion pumps are cooled,
generally with water. Smaller oil diffusion
pumps can, however, also be cooled with
an air stream because a low wall tempe-
rature is not so decisive to the efficiency as
for mercury diffusion pumps. Oil diffusion
pumps can operate well with wall tempera-
tures of 30 C, whereas the walls of mer-
cury diffusion pumps must be cooled to
15 C. To protect the pumps from the
danger of failure of the cooling water
insofar as the cooling-water coil is not
controlled by thermally operated protec-
tive switching a water circulation moni-
tor should be installed in the cooling water
circuit; hence, evaporation of the pump
fluid from the pump walls is avoided.
2.1.6.1 (Oil) Diffusion pumps
These pumps consist basically (see Fig.
2.44) of a pump body (3) with a cooled
wall (4) and a three- or four-stage nozzle
system (A D). The oil serving as pump
fluid is in the boiler (2) and is vaporized
from here by electrical heating (1). The
pump fluid vapor streams through the
riser tubes and emerges with supersonic
speed from the ring-shaped nozzles (A
D). Thereafter the jet so-formed widens
like an umbrella and reaches the wall
where condensation of the pump fluid
occurs. The liquid condensate flows
downward as a thin film along the wall and
finally returns into the boiler. Because of
this spreading of the jet, the vapor density
is relatively low. The diffusion of air or any
pumped gases (or vapors) into the jet is so
rapid that despite its high velocity the jet
Fig. 2.44 Mode of operation of a diffusion pump
1 Heater
2 Boiler
3 Pump body
4 Cooling coil
5 High vacuum flange
6 Gas molecules
7 Vapor jet
8 Backing vacuum connection
A
B
C Nozzles
D
D00 E 19.06.2001 21:36 Uhr Seite 36

Vacuum Generation
becomes virtually completely saturated
with the pumped medium. Therefore, over
a wide pressure range diffusion pumps
have a high pumping speed. This is
practically constant over the entire
working region of the diffusion pump
( 10
-3
mbar) because the air at these low
pressures cannot influence the jet, so its
course remains undisturbed. At higher
inlet pressures, the course of the jet is
altered. As a result, the pumping speed
decreases until, at about 10
-1
mbar, it
becomes immeasurably small.
The forevacuum pressure also influences
the vapor jet and becomes detrimental if
its value exceeds a certain critical limit.
This limit is called maximum backing
pressure or critical forepressure. The
capacity of the chosen backing pump must
be such (see 2.3.2) that the amount of gas
discharged from the diffusion pump is
pumped off without building up a backing
pressure that is near the maximum
backing pressure or even exceeding it.
The attainable ultimate pressure depends
on the construction of the pump, the vapor
pressure of the pump fluid used, the
maximum possible condensation of the
pump fluid, and the cleanliness of the
vessel. Moreover, backstreaming of the
pump fluid into the vessel should be
reduced as far as possible by suitable
baffles or cold traps (see Section 2.1.6.4).
Degassing of the pump oil
In oil diffusion pumps it is necessary for
the pump fluid to be degassed before it is
returned to the boiler. On heating of the
pump oil, decomposition products can
arise in the pump. Contamination from the
vessel can get into the pump or be
contained in the pump in the first place.
These constituents of the pump fluid can
significantly worsen the ultimate pressure
attainable by a diffusion pump, if they are
not kept away from the vessel. Therefore,
the pump fluid must be freed of these
impurities and from absorbed gases.
This is the function of the degassing
section, through which the circulating oil
passes shortly before re-entry into the
boiler. In the degassing section, the most
volatile impurities escape. Degassing is
obtained by the carefully controlled
temperature distribution in the pump. The
condensed pump fluid, which runs down
the cooled walls as a thin film, is raised to
a temperature of about 130 C below the
lowest diffusion stage, to allow the volatile
components to evaporate and be removed
by the backing pump. Therefore, the re-
evaporating pump fluid consists of only
the less volatile components of the pump
oil.
Pumping speed
The magnitude of the specific pumping
speed S of a diffusion pump that is, the
pumping speed per unit of area of the
actual inlet surface depends on several
parameters, including the position and
dimensions of the high vacuum stage, the
velocity of the pump fluid vapor, and the
mean molecular velocity c
-
of the gas being
pumped (see equation 1.17 in Section
1.1). With the aid of the kinetic theory of
gases, the maximum attainable specific
pumping speed at room temperature on
pumping air is calculated to
S
max
= 11.6 l s
-1
cm
-2
. This is the spe-
cific (molecular) flow conductance of the
intake area of the pump, resembling an
aperture of the same surface area (see
equation 1.30 in Section 1.5.3). Quite
generally, diffusion pumps have a higher
pumping speed for lighter gases compared
to heavier gases.
To characterize the effectiveness of a
diffusion pump, the so called HO factor is
defined. This is the ratio of the actually
obtained specific pumping speed to the
theoretical maximum possible specific
pumping speed. In the case of diffusion
pumps from LEYBOLD optimum values
are attained (of 0.3 for the smallest and up
to 0.55 for the larger pumps).
The various oil diffusion pumps manufac-
tured by LEYBOLD differ in the following
design features (see Fig. 2.45 a and b).
a) LEYBODIFF series
This series of pumps is equipped with a
fractionating device. The various consti-
tuents of the pump fluid are selected so
that the high vacuum nozzle is supplied
only by the fraction of the pump fluid that
has the lowest vapor pressure. This
assures a particularly low ultimate pres-
sure. Fractionating occurs because the
degassed oil first enters the outer part of
the boiler, which serves the nozzle on the
backing vacuum side. Here a part of the
more volatile constituents evaporates.
Hence the already purified pump fluid
reaches the intermediate part of the boiler,
which serves the intermediate nozzle. Here
D00
Fig. 2.45 Diagram showing the basic differences in LEYBOLD oil diffusion pumps
1
2
a) LEYBODIFF-Pump with fractionating
facility
1 Center
2 Middle section
3 Outer part of the boiler
(fractionation)
b) DI pump; side view on to the
internal heater
1 Thermostat sensor
2 Heating cartridge
D00 E 19.06.2001 21:37 Uhr Seite 37
Vacuum Generation
the lighter constituents are evaporated in
greater quantities than the heavier con-
stituents. When the oil enters the central
region of the boiler, which serves the high
vacuum nozzle, it has already been freed of
the light volatile constituents.
b) DI series
In these pumps an evaporation process for
the pump fluid which is essentially free of
bursts is attained by the exceptional heater
design resulting in a highly constant
pumping speed over time. The heater is of
the internal type and consists of heating
cartridges into which tubes with soldered
on thermal conductivity panels are
introduced. The tubes made of stainless
steel are welded horizontally into the
pumps body and are located above the oil
level. The thermal conductivity panels
made of copper are only in part immersed
in the pump fluid. Those parts of the
thermal conductivity panels are so rated
that the pump fluid can evaporate
intensively but without any retardation of
boiling. Those parts of the thermal
conductivity panels above the oil level
supply additional energy to the vapor.
Owing to the special design of the heating
system, the heater cartridges may be
exchanged also while the pump is still hot.
2.1.6.2 Oil vapor ejector pumps
The pumping action of a vapor ejector
stage is explained with the aid of Fig. 2.46.
The pump fluid enters under high pressure
p
1
the nozzle (1), constructed as a Laval
nozzle. There it is expanded to the inlet
pressure p
2
. On this expansion, the
sudden change of energy is accompanied
by an increase of the velocity. The
consequently accelerated pump fluid
vapor jet streams through the mixer region
(3), which is connected to the vessel (4)
being evacuated. Here the gas molecules
emerging from the vessel are dragged
along with the vapor jet. The mixture,
pump fluid vapor gas, now enters the
diffuser nozzle constructed as a Venturi
nozzle (2). Here the vapor gas mixture is
compressed to the backing pressure p
3
with simultaneous diminution of the
velocity. The pump fluid vapor is then con-
densed at the pump walls, whereas the
entrained gas is removed by the backing
pump. Oil vapor ejector pumps are ideally
suited for the pumping of larger quantities
of gas or vapor in the pressure region
between 1 and 10
-3
mbar. The higher
density of the vapor stream in the nozzles
ensures that the diffusion of the pumped
gas in the vapor stream takes place much
more slowly than in diffusion pumps, so
that only the outer layers of the vapor
stream are permeated by gas. Moreover,
the surface through which the diffusion
occurs is much smaller because of the
special construction of the nozzles. The
specific pumping speed of the vapor
ejector pumps is, therefore, smaller than
that of the diffusion pumps. As the pum-
ped gas in the neighborhood of the jet
under the essentially higher inlet pressure
decisively influences the course of the flow
lines, optimum conditions are obtained
only at certain inlet pressures. Therefore,
the pumping speed does not remain con-
stant toward low inlet pressures. As a
consequence of the high vapor stream
velocity and density, oil vapor ejector
pumps can transport gases against a rela-
tively high backing pressure. Their critical
backing pressure lies at a few millibars.
The oil vapor ejector pumps used in
present-day vacuum technology have, in
general, one or more diffusion stages and
several subsequent ejector stages. The
nozzle system of the booster is construc-
ted from two diffusion stages and two
ejector stages in cascade (see Fig. 2.47).
The diffusion stages provide the high
pumping speed between 10
-4
and 10
-3
mbar (see Fig. 2.48), the ejector stages,
the high gas throughput at high pressures
(see Fig. 2.49) and the high critical backing
pressure. Insensitivity to dust and vapors
dissolved in the pump fluid is obtained by
a spacious boiler and a large pump fluid
Fig. 2.46 Operation of a vapor jet pump
1 Nozzle (Laval)
2 Diffuser nozzle (Venturi)
3 Mixing chamber
4 Connection to the vacuum chamber
1 High vacuum port
2 Diffusion stages
3 Ejector stages
Fig. 2.47 Diagram of an oil jet (booster) pump
Fig. 2.48 Pumping speed of various vapor pumps as a
function of intake pressure related to a nomi-
nal pumping speed of 1000 l/s.
End of the working range of oil vapor ejector
pumps (A) and diffusion pumps (B)
Intake pressure p
a
P
u
m
p
i
n
g

s
p
e
e
d

[
l

1
]
Fig. 2.49 Pumping speed of various vapor pumps
(derived from Fig. 2.48)
Intake pressure p
a
P
u
m
p
i
n
g

s
p
e
e
d

[
T
o
r
r

1
]
P
u
m
p
i
n
g

s
p
e
e
d

[
m
b
a
r

1
]
D00 E 19.06.2001 21:37 Uhr Seite 38
Vacuum Generation
reservoir. Large quantities of impurities
can be contained in the boiler without de-
terioration of the pumping characteristics.
2.1.6.3 Pump fluids
a) Oils
The suitable pump fluids for oil diffusion
pumps are mineral oils, silicone oils, and
oils based on the polyphenyl ethers.
Severe demands are placed on such oils
which are met only by special fluids. The
properties of these, such as vapor pres-
sure, thermal and chemical resistance,
particularly against air, determine the
choice of oil to be used in a given type of
pump or to attain a given ultimate vacuum.
The vapor pressure of the oils used in
vapor pumps is lower than that of mer-
cury. Organic pump fluids are more sen-
sitive in operation than mercury, because
the oils can be decomposed by long-term
admission of air. Silicone oils, however,
withstand longer lasting frequent admis-
sions of air into the operational pump.
Typical mineral oils are DIFFELEN light,
normal and ultra. The different types of
DIFFELEN are close tolerance fractions of a
high quality base product (see our cata-
log).
Silicone oils (DC 704, DC 705, for examp-
le) are uniform chemical compounds
(organic polymers). They are highly
resistant to oxidation in the case of air
inrushes and offer special thermal stability
characteristics.
DC 705 has an extremely low vapor pres-
sure and is thus suited for use in diffusion
pumps which are used to attain extremely
low ultimate pressures of < 10
-10
mbar.
ULTRALEN is a polyphenylether. This fluid
is recommended in all those cases where a
particularly oxidation-resistant pump fluid
must be used and where silicone oils
would interfere with the process.
APIEZON AP 201 is an oil of exceptional
thermal and chemical resistance capable
of delivering the required high pumping
speed in connection with oil vapor ejector
pumps operating in the medium vacuum
range. The attainable ultimate total pres-
sure amounts to about 10
-4
mbar.
b) Mercury
Mercury is a very suitable pump fluid. It is
a chemical element that during vaporiza-
tion neither decomposes nor becomes
strongly oxidized when air is admitted.
However, at room temperature it has a
comparatively high vapor pressure of
10
-3
mbar. If lower ultimate total pressures
are to be reached, cold traps with liquid
nitrogen are needed. With their aid,
ultimate total pressures of 10
-10
mbar can
be obtained with mercury diffusion pumps.
Because mercury is toxic, as already
mentioned, and because it presents a
hazard to the environment, it is nowadays
hardly ever used as a pump fluid.
LEYBOLD supplies pumps with mercury as
the pump fluid only on request. The vapor
pressure curves of pump fluids are given in
Fig. 9.12, Section 9.
2.1.6.4 Pump fluid backstreaming and
its suppression (vapor
barriers, baffles)
In the vapor stream from the topmost
nozzle of a diffusion pump, pump fluid
molecules not only travel in the direction
of streaming to the cooled pump wall, but
also have backward components of
velocity because of intermolecular colli-
sions. They can thus stream in the direc-
tion of the vessel. In the case of
LEYBODIFF and DI pumps, the oil-back-
streaming amounts to a few micrograms
per minute for each square centimeter of
inlet cross-sectional area. To reduce this
backstreaming as much as possible,
various measures must be undertaken
simultaneously:
a) the high vacuum-side nozzle and the
shape of the part of the pump body
surrounding this nozzle must be
constructed so that as few as possible
vapor molecules emerge sideways in
the path of the vapor stream from the
nozzle exit to the cooled pump wall.
b) the method for cooling the pump wall
must allow as complete as possible
condensation of the pump fluid vapor
and, after condensation, the fluid must
be able to drain away readily.
c) one or more pump-fluid traps, baffles,
or cold traps must be inserted between
the pump and the vessel, depending on
the ultimate pressure that is required.
Two chief requirements must be met in the
construction of baffles or cold traps for oil
diffusion pumps. First, as far as possible,
all backstreaming pump-fluid vapor mole-
cules should remain attached to (conden-
sed at) the inner cooled surfaces of these
devices. Second, the condensation sur-
faces must be so constructed and geome-
trically arranged that the flow conductance
of the baffles or cold traps is as large as
possible for the pumped gas. These two
requirements are summarized by the term
optically opaque. This means that the
particles cannot enter the baffle without
hitting the wall, although the baffle has a
high conductance. The implementation of
this idea has resulted in a variety of
designs that take into account one or the
other requirement.
A cold cap baffle is constructed so that it
can be mounted immediately above the
high vacuum nozzle. The cold cap baffle is
made of metal of high thermal conductivity
in good thermal contact with the cooled
pump wall, so that in practice it is main-
tained at the cooling-water temperature or,
with air-cooled diffusion pumps, at ambi-
ent temperature. In larger types of pumps,
the cold cap baffle is water cooled and
permanently attached to the pump body.
The effective pumping speed of a diffusion
pump is reduced by about 10 % on
installation of the cold cap baffle, but the
oil backstreaming is reduced by about 90
to 95 %.
Shell baffles consist of concentrically
arranged shells and a central baffle plate.
With appropriate cooling by water or
refrigeration, almost entirely oil vapor-free
vacua can be produced by this means. The
effective pumping speed of the diffusion
pump remains at least at 50 %, although
shell baffles are optically opaque. This type
of baffle has been developed by LEYBOLD
in two different forms: with a stainless-
steel cooling coil or in the so-called
Astrotorus baffles with cooling inserts of
copper. The casing of the former type is
made entirely of stainless steel.
For the smaller air-cooled, oil diffusion
pumps, plate baffles are used. The air-
cooled arrangement consists of a copper
plate with copper webs to the housing
wall. The temperature of the plate baffle
remains nearly ambient during the
operation of the diffusion pump.
Hydrocarbon-free vacuum
If extreme demands are made on freedom
from oil vapor with vacuum produced by
oil diffusion pumps, cold traps should be
used that are cooled with liquid nitrogen
D00
D00 E 19.06.2001 21:37 Uhr Seite 39
Vacuum Generation
so that they are maintained at a tempera-
ture of -196 C.
Low-temperature baffles or cold traps
should always be used with a cold cap in
place. On this the greatest part of the
backstreaming oil is condensed, so that
the inevitable loss of pump fluid from the
condensation of the pump fluid on the low-
temperature surface is kept at a minimum.
With longer-term operation it is always ad-
visable to install, in place of the cold cap, a
water-cooled shell or chevron baffle
between the diffusion pump and the low-
temperature baffle or cold trap (see Fig.
2.50).
LEYBOLD manufactures cold traps made
of metal so called LN
2
cold traps. These
cold traps are to be used in those cases
where a cold trap is to be operated for
prolonged periods of time without re-
quiring a filling facility for liquid nitrogen.
The temperature increase at the vessel
containing the refrigerant is so slight over
the operating time that as the liquid level
drops no significant desorption of the
condensate takes place. Located on the
pumping side is an impact panel made of
copper. The low temperature of this panel
ensures that the greater part of the
condensed pump fluid remains in the
liquid state and may drip back into the
pump. Today the oils used to operate dif-
fusion pumps have a very low vapor pres-
sure at room temperature (for example
DIFFELEN light, 2 10
-8
mbar; DC 705,
4 10
-10
mbar). The specified provisions
with a liquid-nitrogen-cooled baffle or cold
trap would enable an absolutely oil-free
vacuum to be produced.
In practice, however, complete suppres-
sion of oil-backstreaming is never at-
tained. There are always a few pump-fluid
molecules that, as a result of collisions
with one another, reach the vessel without
having hit one of the cooled surfaces of the
baffle or the cold trap. Moreover, there are
always a few highly volatile components of
the pump fluid that do not remain attached
to the very low temperature surfaces. The
temperature and the vapor molecules
adsorbed at the surface of the vessel
determine exactly the pressure in the
vessel. If, the surfaces are not fully cove-
red with adsorbed pump-fluid molecules
after a bake-out process, their vapor
pressure contributes only insignificantly to
the pressure in the vessel.
After a certain time, the stay-down time,
a continuous layer of oil molecules builds
up, and the ultimate pressure is practically
determined by the vapor pressure of the
pump fluid at the temperature of the vessel
walls. This stay-down time can even
amount to several hours, indeed even to
days, with the use of low-temperature
baffles.
Oil can reach the vessel not only as vapor,
but also as a liquid film, because oil wets
readily and thus creeps up the wall.
By installation of an anticreep barrier (see
Fig. 2.50) made of Teflon polymer, a mate-
rial that is not wetted by oil and can stand
a bake-out temperature up to 200 C,
further creeping of the oil can be effecti-
vely prevented. It is most appropriate to
arrange the anticreep barrier above the
upper baffle (see Fig. 2.50).
Note:
It must be noted that data on back-
streaming as specified in catalogs apply
only to continuously-operated oil diffusion
pumps. Shortly after starting a pump the
uppermost nozzle will not eject a well
directed vapor jet. Instead oil vapor
spreads in all directions for several se-
conds and the backstreaming effect is
strong. When switching a diffusion pump
on and off frequently the degree of oil
brackstreaming will be greater.
2.1.6.5 Water jet pumps and steam
ejectors
Included in the class of fluid-entrainment
pumps are not only pumps that use a fast-
streaming vapor as the pump fluid, but
also liquid jet pumps. The simplest and
cheapest vacuum pumps are water jet
pumps. As in a vapor pump (see Fig. 2.46
or 2.51), the liquid stream is first released
from a nozzle and then, because of
turbulence, mixes with the pumped gas in
the mixing chamber. Finally, the movement
of the water gas mixture is slowed down
in a Venturi tube. The ultimate total
pressure in a container that is pumped by
a water jet pump is determined by the
vapor pressure of the water and, for
example, at a water temperature of 15 C
amounts to about 17 mbar.
Essentially higher pumping speeds and
lower ultimate pressures are produced by
steam ejector pumps. The section
through one stage is shown in Fig. 2.51.
The markings correspond with those
shown in Fig. 2.46. In practice, several
pumping stages are usually mounted in
cascade. For laboratory work, two-stage
pump combinations are suitable and
consist of a steam ejector stage and a
water jet (backing) stage, both made of
glass. The water jet backing stage enables
operation without other backing pumps.
With the help of a vapor stream at
overpressure, the vacuum chamber can be
evacuated to an ultimate pressure of about
3 mbar. The condensate from the steam is
led off through the drain attachment. The
water jet stage of this pump is cooled with
water to increase its efficiency. Steam
ejector pumps are especially suitable for
work in laboratories, particularly if very
aggressive vapors are to be pumped.
Steam ejector pumps, which will operate
at a pressure of a few millibars, are
Fig. 2.50 Schematic arrangement of baffle, anticreep
barrier and cold trap above a diffusion pump
1 Diffusion pump with cold
cap baffle (cooled by contact),
2 Shell or chevron baffle
3 Anticreep barrier
4 Sealing gasket
5 Bearing ring
6 LN
2
cold trap,
7 Vacuum chamber
D00 E 19.06.2001 21:37 Uhr Seite 40
2.1.7 Turbomolecular pumps
The principle of the molecular pump well
known since 1913 is that the gas par-
ticles to be pumped receive, through
impact with the rapidly moving surfaces of
a rotor, an impulse in a required flow direc-
tion. The surfaces of the rotor usually
disk-shaped form, with the stationary
surfaces of a stator, intervening spaces in
which the gas is transported to the backing
port. In the original Gaede molecular
pump and its modifications, the inter-
vening spaces (transport channels) were
very narrow, which led to constructional
difficulties and a high degree of sus-
ceptibility to mechanical contamination.
At the end of the Fifties, it became possible
through a turbine-like design and by
modification of the ideas of Gaede to
produce a technically viable pump the so-
called turbomolecular pump. The
spaces between the stator and the rotor
disks were made in the order of
millimeters, so that essentially larger
tolerances could be obtained. Thereby,
greater security in operation was achieved.
However, a pumping effect of any
significance is only attained when the
circumferential velocity (at the outside
rim) of the rotor blades reaches the order
of magnitude of the average thermal
velocity of the molecules which are to be
pumped. Kinetic gas theory supplies for c
-
of the equation 1.17:
in which the dependency on the type of gas
as a function of molar mass M is contained.
The calculation involving cgs-units
(where R = 83.14 10
6
mbar cm
3
/mol K)
results in the following Table:
Gas Molar Mean thermal
Mass M velocity (m/s)
H
2
2 1761
He 4 1245
H
2
O 18 587
Ne 20 557
CO 28 471
N
2
28 471
Air 28.96 463
O
2
32 440
Ar 40 394
CO
2
44 375
CC
13
F (F11) 134.78 68
Table 2.4 c
-
as a function of molar mass M
c
R T
M
=

Vacuum Generation
especially recommended for pumping
laboratory distillation apparatus and
similar plants when the pressure from a
simple water jet pump is insufficient. In
this instance, the use of rotary pumps
would not be economical.
Even in spite of their low investment costs
water jet pumps and steam ejectors are
being replaced in the laboratories more
and more by diaphragm pumps because of
the environmental problems of using water
as the pump fluid. Solvent entering the
water can only be removed again through
complex cleaning methods (distillation).
Whereas the dependence of the pumping
speed on the type of gas is fairly low
(S c
_
1 / E
__
M ), the dependence of the
compression k
0
at zero throughput and
thus also the compression k, because
of k
0
e
M
log k
0

__
M , is greater as
shown by the experimentally-determined
relationship in Fig. 2.55.
Example:
from theory it follows that
this with k
0
(He) = 3 10
3
from Fig. 2.55
results in:
log k
0
(N
2
) = 2.65 log (3 10
3
) = 9.21
or k
0
(N
2
) = 1.6 10
9
.
This agrees as expected well (order of
magnitude) with the experimentally
determined value for k
0
(N
2
) = 2.0 10
8
from Fig. 2.55. In view of the optimizations
for the individual rotor stages common
today, this consideration is no longer
correct for the entire pump. Shown in Fig.
2.56 are the values as measured for a
modern TURBOVAC 340 M.
In order to meet the condition, a circumfe-
rential velocity for the rotor of the same
order of magnitude as c
-
high rotor speeds
are required for turbomolecular pumps.
They range from about 36,000 rpm for
pumps having a large diameter rotor
(TURBOVAC 1000) to 72,000 rpm in the
case of smaller rotor diameters
(TURBOVAC 35 / 55). Such high speeds
naturally raise questions as to a reliable
bearing concept. LEYBOLD offers three
concepts, the advantages and disadvan-
tages of which are detailed in the follo-
wing:
Oil lubrication / steel ball bearings
+ Good compatibility with particles by
circulating oil lubricant
- Can only be installed vertically
+ Low maintenance
Grease lubrication / hybrid bearings
+ Installation in any orientation
+ Suited for mobile systems
Air cooling will do for many applica-
tions
+ Lubricated for life (of the bearings)
log ( )
log ( ) .
k He
k N
0
0 2
4
28
1
7
1
2 65

log ( ) . log ( ) k N k He
0 0
2 65
2

D00
Fig. 2.51 Schematic representation of the operation
of a steam ejector pump
1 Steam inlet
2 Jet nozzle
3 Diffuser
4 Mixing region
5 Connection to the vacuum chamber
D00 E 19.06.2001 21:37 Uhr Seite 41
Vacuum Generation
Free of lubricants / magnetic suspen-
sion
+ No wear
+ No maintenance
+ Absolutely free of hydrocarbons
+ Low noise and vibration levels
+ Installation in any orientation
Steel ball bearings / hybrid ball bearings
(ceramic ball bearings): Even a brief tear
in the thin lubricating film between the
balls and the races can if the same type
of material is used result in microwelding
at the points of contact. This severely re-
duces the service life of the bearings. By
using dissimilar materials in so called hy-
brid bearings (races: steel, balls: ceramics)
the effect of microwelding is avoided.
The most elegant bearing concept is that
of the magnetic suspension. As early as
1976 LEYBOLD delivered magnetically
suspended turbomolecular pumps the
legendary series 550M and 560M. At that
time a purely active magnetic suspension
(i.e. with electromagnets) was used. Ad-
vances in electronics and the use of per-
manent magnets (passive magnetic
suspension) based on the System KFA
Jlich permitted the magnetic suspen-
sion concept to spread widely. In this
system the rotor is maintained in a stable
position without contact during operation,
by magnetic forces. Absolutely no lubri-
cants are required. So-called touch down
bearings are integrated for shutdown.
Fig. 2.52 shows a sectional drawing of a
typical turbomolecular pump. The pump is
an axial flow compressor of vertical
design, the active or pumping part of
which consists of a rotor (6) and a stator
(2). Turbine blades are located around the
circumferences of the stator and the rotor.
Each rotor stator pair of circular blade
rows forms one stage, so that the
assembly is composed of a multitude of
stages mounted in series. The gas to be
pumped arrives directly through the
aperture of the inlet flange (1), that is,
without any loss of conductance, at the
active pumping area of the top blades of
the rotor stator assembly. This is equip-
ped with blades of especially large radial
span to allow a large annular inlet area.
The gas captured by these stages is
transferred to the lower compression
stages, whose blades have shorter radial
spans, where the gas is compressed to
backing pressure or rough vacuum pres-
sure. The turbine rotor (6) is mounted on
the drive shaft, which is supported by two
precision ball bearings (8 and 11), accom-
modated in the motor housing. The rotor
shaft is directly driven by a medium-fre-
quency motor housed in the forevacuum
space within the rotor, so that no rotary
shaft lead-through to the outside atmos-
phere is necessary. This motor is powered
and automatically controlled by an external
frequency converter, normally a solid-state
frequency converter that ensures a very
low noise level. For special applications,
for example, in areas exposed to radiation,
motor generator frequency converters are
used.
The vertical rotor stator configuration
provides optimum flow conditions of the
gas at the inlet.
To ensure vibration-free running at high
rotational speeds, the turbine is dyna-
mically balanced at two levels during its
assembly.
Fig. 2.52 Schematic diagram of a grease lubricated TURBOVAC 151
turbomolecular pump
7 Pump casing
8 Ball bearings
9 Cooling water connection
10 3-phase motor
11 Ball bearings
1 High vacuum inlet flange
2 Stator pack
3 Venting flange
4 Forevacuum flange
5 Splinter guard
6 Rotor
Fig. 2.52a Cross section of a HY.CONE turbomolecular pump
Turbomole-
cular
pump stage
Siegbahn
stage
1
2
3
4
5
6
8
7
5 Bearing
6 Motor
7 Fan
8 Bearing
1 Vacuum port
3 Rotor
4 Stator
D00 E 19.06.2001 21:37 Uhr Seite 42
Vacuum Generation
The pumping speed (volume flow rate)
characteristics of turbomolecular pumps
are shown in Fig. 2.53. The pumping
speed remains constant over the entire
working pressure range. It decreases at
intake pressures above 10
-3
mbar, as this
threshold value marks the transition from
the region of molecular flow to the region
of laminar viscous flow of gases. Fig. 2.54
shows also that the pumping speed
depends on the type of gas.
The compression ratio (often also simply
termed compression) of turbomolecular
pumps is the ratio between the partial
pressure of one gas component at the
forevacuum flange of the pump and that at
the high vacuum flange: maximum com-
pression k
0
is to be found at zero
throughput. For physical reasons, the
compression ratio of turbomolecular
pumps is very high for heavy molecules
but considerably lower for light molecules.
The relationship between compression
and molecular mass is shown in Fig. 2.55.
Shown in Fig. 2.56 are the compression
curves of a TURBOVAC 340 M for N
2
, He
and H
2
as a function of the backing pres-
sure. Because of the high compression
ratio for heavy hydrocarbon molecules,
turbomolecular pumps can be directly
connected to a vacuum chamber without
the aid of one or more cooled baffles or
traps and without the risk of a measurable
partial pressure for hydrocarbons in the
vacuum chamber (hydrocarbon-free
vacuum! see also Fig. 2.57: residual gas
spectrum above a TURBOVAC 361). As the
hydrogen partial pressure attained by the
rotary backing pump is very low, the
turbomolecular pump is capable of
attaining ultimate pressures in the 10
-11
mbar range in spite of its rather moderate
compression for H
2
. To produce such ex-
tremely low pressures, it will, of course, be
necessary to strictly observe the general
rules of UHV technology: the vacuum
chamber and the upper part of the turbo-
D00
10
6
2 4 6 8
10
5
10
4
p 10
3
10
2
10
1
10000
1000
500
200
100
10
S
mbar
340M
50/55
361
600
151
1000/1000 MC
Fig. 2.53 Pumping speed for air of different turbomolecular pumps
l s
1
Fig. 2.54 Pumping speed curves of a TURBOVAC 600 for H
2
, He, N
2
and Ar
l s
1
Fig. 2.55 TURBOVAC 450 Maximum compression
k
0
as a function of molar mass M
k
0
M
Fig. 2.56 Maximum compression k
0
of a turbomolecular
pump TURBOVAC 340 M for H
2
, He and N
2
as
a function of backing pressure
Fig. 2.57 Spectrum above a TURBOVAC 361
M = Mass number = Relative molar mass at
an ionization 1
I = Ion current
D00 E 19.06.2001 21:37 Uhr Seite 43
Vacuum Generation
molecular pump must be baked out, and
metal seals must be used. At very low
pressures the residual gas is composed
mainly of H
2
originating from the metal
walls of the chamber. The spectrum in Fig.
2.57 shows the residual gas composition
in front of the inlet of a turbomolecular
pump at an ultimate pressure of 7 10
-10
mbar nitrogen equivalent. It appears that
the portion of H
2
in the total quantity of
gas amounts to approximately 90 to 95 %.
The fraction of heavier molecules is con-
siderably reduced and masses greater than
44 were not detected. An important crite-
rion in the assessment of the quality of a
residual gas spectrum are the measurable
hydrocarbons from the lubricants used in
the vacuum pump system. Of course an
absolutely hydrocarbon-free vacuum
can only be produced with pump systems
which are free of lubricants, i.e. for exam-
ple with magnetically-suspended turbo-
molecular pumps and dry compressing
backing pumps. When operated correctly
(venting at any kind of standstill) no hy-
drocarbons are detectable also in the spec-
trum of normal turbomolecular pumps.
A further development of the turbomo-
lecular pump is the hybrid or compound
turbomolecular pump. This is actually two
pumps on a common shaft in a single
casing. The high vacuum stage for the mo-
lecular flow region is a classic turbo-
molecular pump, the second pump for the
viscous flow range is a molecular drag or
friction pump.
LEYBOLD manufactures pumps such as
the TURBOVAC 55 with an integrated
Holweck stage (screw-type compressor)
and, for example, the HY.CONE 60 or
HY.CONE 200 with an integrated Siegbahn
stage (spiral compressor). The required
backing pressure then amounts to a few
mbar so that the backing pump is only
required to compress from about 5 to 10
mbar to atmospheric pressure. A sectional
view of a HY.CONE is shown in Fig. 2.52a.
Information on the operation of turbomo-
lecular pumps
Starting
As a rule turbomolecular pumps should
generally be started together with the
backing pump in order to reduce any
backstreaming of oil from the backing
pump into the vacuum chamber. A delayed
start of the turbomolecular pump, makes
sense in the case of rather small backing
pump sets and large vacuum chambers. At
a known pumping speed for the backing
pump S
V
(m
3
/h) and a known volume for
the vacuum chamber (m
3
) it is possible to
estimate the cut-in pressure for the
turbomolecular pump:
Simultaneous start when
and delayed start when
at a cut-in pressure of:
(2.24)
When evacuating larger volumes the cut-in
pressure for turbomolecular pumps may
also be determined with the aid of the
diagram of Fig. 2.58.
Venting
After switching off or in the event of a
power failure, turbomolecular pumps
should always be vented in order to
prevent any backdiffusion of hydrocarbons
from the forevacuum side into the vacuum
chamber. After switching off the pump the
cooling water supply should also be
switched off to prevent the possible
condensation of water vapor. In order to
protect the rotor it is recommended to
comply with the (minimum) venting times
stated in the operating instructions. The
p e mbar
V Start
S
V
V
,

=
_
,
6
S
V
h
v
<

40
1
S
V
h
v
>

40
1
pump should be vented (except in the case
of operation with a barrier gas) via the
venting flange which already contains a
sintered metal throttle, so that venting may
be performed using a normal valve or a
power failure venting valve.
Barrier gas operation
In the case of pumps equipped with a
barrier gas facility, inert gas such as dry
nitrogen may be applied through a
special flange so as to protect the motor
space and the bearings against aggressive
media. A special barrier gas and venting
valve meters the necessary quantity of
barrier gas and may also serve as a ven-
ting valve.
Decoupling of vibrations
TURBOVAC pumps are precisely balanced
and may generally be connected directly to
the apparatus. Only in the case of highly
sensitive instruments, such as electron
microscopes, is it recommended to install
vibration absorbers which reduce the
present vibrations to a minimum. For mag-
netically suspended pumps a direct
connection to the vacuum apparatus will
usually do because of the extremely low
vibrations produced by such pumps.
For special applications such as opera-
tion in strong magnetic fields, radiation
hazard areas or in a tritium atmosphere,
please contact our Technical Sales Depart-
ment which has the necessary experience
and which is available to you at any time.
Fig. 2.58 Determination of the cut-in pressure for
turbomolecular pumps when evacuating
large vessels
D00 E 19.06.2001 21:37 Uhr Seite 44
Vacuum Generation
2.1.8 Sorption pumps
The term sorption pumps includes all
arrangements for the removal of gases and
vapors from a space by sorption means.
The pumped gas particles are thereby
bound at the surfaces or in the interior of
these agents, by either physical tempera-
ture-dependent adsorption forces (van der
Waals forces), chemisorption, absorption,
or by becoming embedded during the
course of the continuous formation of new
sorbing surfaces. By comparing their ope-
rating principles, we can distinguish
between adsorption pumps, in which the
sorption of gases takes place simply by
temperature-controlled adsorption pro-
cesses, and getter pumps, in which the
sorption and retention of gases are
essentially caused by the formation of
chemical compounds. Gettering is the
bonding of gases to pure, mostly metallic
surfaces, which are not covered by oxide
or carbide layers. Such surfaces always
form during manufacture, installation or
while venting the system. The mostly
metallic highest purity getter surfaces are
continuously generated either directly in
the vacuum by evaporation (evaporator
pumps) or by sputtering (sputter pumps)
or the passivating surface layer of the
getter (metal) is removed by degassing the
vacuum, so that the pure material is
exposed to the vacuum. This step is called
activation (NEG pumps NEG = Non Eva-
porable Getter).
2.1.8.1 Adsorption pumps
Adsorption pumps (see Fig. 2.59) work
according to the principle of the physical
adsorption of gases at the surface of
molecular sieves or other adsorption
materials (e.g. activated Al
2
O
3
). Zeolite
13X is frequently used as an adsorption
material. This alkali aluminosilicate
possesses for a mass of the material an
extraordinarily large surface area, about
1000 m
2
/g of solid substance. Corres-
pondingly, its ability to take up gas is con-
siderable.
The pore diameter of zeolite 13X is about
13 , which is within the order of size of
water vapor, oil vapor, and larger gas
molecules (about 10 ). Assuming that the
mean molecular diameter is half this value,
5 10
-8
cm, about 5 10
18
molecules are
adsorbed in a monolayer on a surface of
1 m
2
. For nitrogen molecules with a
relative molecular mass Mr = 28 that
corresponds to about 2 10
-4
g or 0.20
mbar l (see also Section 1.1). Therefore,
an adsorption surface of 1000 m
2
is
capable of adsorbing a monomolecular
layer in which more than 133 mbar l of
gas is bound.
Hydrogen and light noble gases, such as
helium and neon, have a relatively small
particle diameter compared with the pore
size of 13 for zeolite 13X. These gases
are, therefore, very poorly adsorbed.
The adsorption of gases at surfaces is
dependent not only on the temperature,
but more importantly on the pressure
above the adsorption surface. The
dependence is represented graphically for
a few gases by the adsorption isotherms
given in Fig. 2.60. In practice, adsorption
pumps are connected through a valve to
the vessel to be evacuated. It is on
immersing the body of the pump in liquid
nitrogen that the sorption effect is made
technically useful. Because of the different
adsorption properties, the pumping speed
and ultimate pressure of an adsorption
pump are different for the various gas
molecules: the best values are achieved for
nitrogen, carbon dioxide, water vapor, and
hydrocarbon vapors. Light noble gases are
hardly pumped at all because the diameter
of the particles is small compared to the
pores of the zeolite. As the sorption effect
decreases with increased coverage of the
zeolite surfaces, the pumping speed falls
off with an increasing number of the
particles already adsorbed. The pumping
speed of an adsorption pump is, therefore,
dependent on the quantity of gas already
pumped and so is not constant with time.
The ultimate pressure attainable with
adsorption pumps is determined in the
D00
Fig. 2.59 Cross section of an adsorption pump
5 Thermal conducting
vanes
6 Adsorption material
(e.g. Zeolith)
1 Inlet port
2 Degassing port
3 Support
4 Pump body
Fig. 2.60 Adsorption isotherms of zeolite 13X for nitrogen at -195 C and 20 C, as well as for helium and
neon at -195 C
Pressure [Torr]
A
d
s
o
r
b
e
d

q
u
a
n
t
i
t
y

o
f

g
a
s

p
e
r

q
u
a
n
t
i
t
y

o
f

a
d
s
o
r
b
e
n
t

[
m
b
a
r

1
]
Pressure [mbar]
D00 E 19.06.2001 21:37 Uhr Seite 45
Vacuum Generation
first instance by those gases that prevail in
the vessel at the beginning of the pumping
process and are poorly or not at all
adsorbed (e.g. neon or helium) at the
zeolite surface. In atmospheric air, a few
parts per million of these gases are
present. Therefore, pressures < 10
-2
mbar
can be obtained.
If pressures below 10
-3
mbar exclusively
are to be produced with adsorption pumps,
as far as possible no neon or helium
should be present in the gas mixture.
After a pumping process, the pump must
be warmed only to room temperature for
the adsorbed gases to be given off and the
zeolite is regenerated for reuse. If air (or
damp gas) containing a great deal of water
vapor has been pumped, it is recom-
mended to bake out the pump completely
dry for a few hours at 200 C or above.
To pump out larger vessels, several
adsorption pumps are used in parallel or in
series. First, the pressure is reduced from
atmospheric pressure to a few millibars by
the first stage in order to capture many
noble gas molecules of helium and neon.
After the pumps of this stage have been
saturated, the valves to these pumps are
closed and a previously closed valve to a
further adsorption pump still containing
clean adsorbent is opened so that this
pump may pump down the vacuum
chamber to the next lower pressure level.
This procedure can be continued until the
ultimate pressure cannot be further
improved by adding further clean adsorp-
tion pumps.
2.1.8.2 Sublimation pumps
Sublimation pumps are sorption pumps in
which a getter material is evaporated and
deposited on a cold inner wall as a getter
film. On the surface of such a getter film
the gas molecules form stable com-
pounds, which have an immeasurably low
vapor pressure. The active getter film is
renewed by subsequent evaporations.
Generally titanium is used in sublimation
pumps as the getter. The titanium is
evaporated from a wire made of a special
alloy of a high titanium content which is
heated by an electric current. Although the
optimum sorption capacity (about one
nitrogen atom for each evaporated tita-
nium atom) can scarcely be obtained in
practice, titanium sublimation pumps have
an extraordinarily high pumping speed for
active gases, which, particularly on star-
ting processes or on the sudden evolution
of greater quantities of gas, can be rapidly
pumped away. As sublimation pumps
function as auxiliary pumps (boosters) to
sputter-ion pumps and turbomolecular
pumps, their installation is often indi-
spensable (like the boosters in vapor
ejector pumps; see Section 2.1.6.2).
2.1.8.3 Sputter-ion pumps
The pumping action of sputter-ion pumps
is based on sorption processes that are
initiated by ionized gas particles in a Pen-
ning discharge (cold cathode discharge).
By means of paralleling many individual
Penning cells the sputter ion pump
attains a sufficiently high pumping speed
for the individual gases.
Operation of sputter-ion pumps
The ions impinge upon the cathode of the
cold cathode discharge electrode system
and sputter the cathode material (titanium).
The titanium deposited at other locations
acts as a getter film and adsorbs reactive
gas particles (e.g., nitrogen, oxygen, hydro-
gen). The energy of the ionized gas par-
ticles is not only high enough to sputter the
cathode material but also to let the
impinging ions penetrate deeply into the
cathode material (ion implantation). This
sorption process pumps ions of all types,
including ions of gases which do not
chemically react with the sputtered titanium
film, i.e. mainly noble gases.
The following arrangement is used to
produce the ions: stainless-steel, cylindri-
cal anodes are closely arranged between,
with their axes perpendicular to, two
parallel cathodes (see Fig. 2.61). The
cathodes are at negative potential (a few
kilovolts) against the anode. The entire
electrode system is maintained in a strong,
homogeneous magnetic field of a flux
density of B = 0.1 T, (T = Tesla = 10
4
Gauss) produced by a permanent magnet
attached to the outside of the pumps
casing. The gas discharge profduced by
the high tension contains electrons and
ions. Under the influence of the magnetic
field the electrons travel along long spiral
tracks (see Fig. 2.61) until they impinge on
the anode cylinder of the corresponding
cell. The long track increases ion yield,
which even at low gas densities (pres-
sures) is sufficient to maintain a self-
sustained gas discharge. A supply of elec-
trons from a hot cathode is not required.
Because of their great mass, the
movement of the ions is unaffected by the
magnetic field of the given order of mag-
nitude; they flow off along the shortest
path and bombard the cathode.
The discharge current i is proportional to
the number density of neutral particles n
0
,
the electron density n-, and the length l of
the total discharge path:
i = n
0
n
l (2.25)
The effective cross section s for ionizing
collisions depends on the type of gas.
According to (2.25), the discharge current
i is a function of the number particle
density n0, as in a Penning gauge, and it
can be used as a measure of the pressure
in the range from 10
-4
to 10
-8
mbar. At
lower pressures the measurements are not
reproducible due to interferences from
field emission effects.
In diode-type, sputter-ion pumps, with an
electrode system configuration as shown
in Fig. 2.62, the getter films are formed on
the anode surfaces and between the
sputtering regions of the opposite
cathode. The ions are buried in the
cathode surfaces. As cathode sputtering
proceeds, the buried gas particles are set
free again. Therefore, the pumping action
for noble gases that can be pumped only
by ion burial will vanish after some time
Fig. 2.61 Operating principle of a sputter-ion pump
PZ
Direction of motion of the ionized gas
molecules
Direction of motion of the sputtered
titanium
- - Spiral tracks of the electrons
PZ Penning cells
D00 E 19.06.2001 21:37 Uhr Seite 46
Vacuum Generation
and a memory effect will occur.
Unlike diode-type pumps, triode sputter-
ion pumps exhibit excellent stability in
their pumping speed for noble gases
because sputtering and film forming
surfaces are separated. Fig. 2.63 shows
the electrode configuration of triode
sputter-ion pumps. Their greater efficiency
for pumping noble gases is explained as
follows: the geometry of the system favors
grazing incidence of the ions on the
titanium bars of the cathode grid, whereby
the sputtering rate is considerably higher
than with perpendicular incidence. The
sputtered titanium moves in about the
same direction as the incident ions. The
getter films form preferentially on the third
electrode, the target plate, which is the
actual wall of the pump housing. There is
an increasing yield of ionized particles that
are grazingly incident on the cathode grid
where they are neutralized and reflected
and from which they travel to the target
plate at an energy still considerably higher
than the thermal energy 1/2 k T of the
gas particles. The energetic neutral
particles can penetrate into the target
surface layer, but their sputtering effect is
only negligible. These buried or implanted
particles are finally covered by fresh
titanium layers. As the target is at positive
potential, any positive ions arriving there
are repelled and cannot sputter the target
layers. Hence the buried noble gas atoms
are not set free again. The pumping speed
of triode sputter-ion pumps for noble
gases does not decrease during the
operation of the pump.
The pumping speed of sputter-ion pumps
depends on the pressure and the type of
gas. It is measured according to the
methods stated in DIN 28 429 and
PNEUROP 5615. The pumping speed
curve S(p) has a maximum. The nominal
pumping speed S
n
is given by the maxi-
mum of the pumping speed curve for air
whereby the corresponding pressure must
be stated.
For air, nitrogen, carbon dioxide and water
vapor, the pumping speed is practically the
same. Compared with the pumping speed
for air, the pumping speeds of sputter-ion
pumps for other gases amount to
approximately:
Hydrogen 150 to 200 %
Methane 100 %
Other light
hydrocarbons 80 to 120 %
Oxygen 80 %
Argon 30 %
Helium 28 %
Sputter-ion pumps of the triode type excel
in contrast to the diode-type pumps in
high-noble gas stability. Argon is pumped
stably even at an inlet pressure of
1 10
-5
mbar. The pumps can be started
without difficulties at pressures higher than
1 10
-2
mbar and can operate continuously
at an air inlet producing a constant air
pressure of 5 10
-5
mbar. A new kind of
design for the electrodes extends the
service life of the cathodes by 50 %.
Influence on processes in the vacuum
chamber by magnetic stray fields and
stray ions from the sputter-ion pump.
The high-magnetic-field strength required
for the pumping action leads inevitably to
stray magnetic fields in the neighborhood
of the magnets. As a result, processes in
the vacuum chamber can be disturbed in
some cases, so the sputter-ion pump
concerned should be provided with a
screening arrangement. The forms and
kinds of such a screening arrangement can
be regarded as at an optimum if the
processes taking place in the vacuum
chamber are disturbed by no more than
the earths magnetic field which is present
in any case.
Fig. 2.64 shows the magnetic stray field at
the plane of the intake flange of a sputter-
ion pump IZ 270 and also at a parallel
plane 150 mm above. If stray ions from the
discharge region are to be prevented from
reaching the vacuum chamber, a suitable
screen can be set up by a metal sieve at
opposite potential in the inlet opening of
the sputter-ion pump (ion barrier). This,
however, reduces the pumping speed of
the sputter-ion pump depending on the
mesh size of the selected metal sieve.
D00
Fig. 2.62 Electrode configuration in a diode sputter-ion pump
Titanium atoms
Gas molecules
Ions
Electrons
B Magnetic field
Fig. 2.63 Electrode configuration in a triode sputter-ion pump
Titanium atoms
Gas molecules
Ions
Electrons
A Anode cylinder
(same as in the diode
pump)
B Magnetic field
F Target plate
(pump housing)
as the third electrode
K Cathode grid
D00 E 19.06.2001 21:37 Uhr Seite 47
Vacuum Generation
2.1.8.4 Non evaporable getter pumps
(NEG Pumps)
The non evaporable getter pump operates
with a non evaporable, compact getter
material, the structure of which is porous
at the atomic level so that it can take up
large quantities of gas. The gas molecules
adsorbed on the surface of the getter
material diffuse rapidly inside the material
thereby making place for further gas
molecules impinging on the surface. The
non evaporable getter pump contains a
heating element which is used to heat the
getter material to an optimum temperature
depending on the type of gas which is
preferably to be pumped. At a higher
temperature the getter material which has
been saturated with the gas is regenerated
(activated). As the getter material, mostly
zirconium-aluminum alloys are used in the
form of strips. The special properties of
NEG pumps are:
constant pumping speed in the HV and
UHV ranges
no pressure restrictions up to about 12
mbar
particularly high pumping speed for
hydrogen and its isotopes
after activation the pump can often
operate at room temperature and will
then need no electrical energy
no interference by magnetic fields
hydrocarbon-free vacuum
free of vibrations
low weight
NEG pumps are mostly used in
combination with other UHV pumps (tur-
bomolecular and cryopumps). Such com-
binations are especially useful when
wanting to further reduce the ultimate
pressure of UHV systems, since hydrogen
contributes mostly to the ultimate pres-
sure in an UHV system, and for which NEG
pumps have a particularly high pumping
speed, whereas the pumping effect for H
2
of other pumps is low. Some typical
examples for applications in which NEG
pumps are used are particle accelerators
and similar research systems, surface
analysis instruments, SEM columns and
sputtering systems. NEG pumps are
manufactured offering pumping speeds of
several l/s to about 1000 l/s. Custom
pumps are capable of attaining a pumping
speed for hydrogen which is by several
orders of magnitude higher.
Fig. 2.64 Stray magnetic field of a sputter-ion pump
in two places parallel to the inlet flange
(inserts) curves show lines of constant
magnetic induction B in Gauss.
1 Gauss = 1 10
4
Tesla
D00 E 19.06.2001 21:37 Uhr Seite 48
Vacuum Generation
2.1.9 Cryopumps
As you may have observed water
condenses on cold water mains or
windows and ice forms on the evaporator
unit in your refrigerator. This effect of
condensation of gases and vapors on cold
surfaces, water vapor in particular, as it is
known in every day life, occurs not only at
atmospheric pressure but also in vacuum.
This effect has been utilized for a long time
in condensers (see 2.1.5) mainly in
connection with chemical processes;
previously the baffle on diffusion pumps
used to be cooled with refrigerating
machines. Also in a sealed space (vacuum
chamber) the formation of condensate on
a cold surface means that a large number
of gas molecules are removed from the
volume: they remain located on the cold
surface and do not take part any longer in
the hectic gas atmosphere within the
vacuum chamber. We then say that the
particles have been pumped and talk of
cryopumps when the pumping effect is
attained by means of cold surfaces.
Cryo engineering differs from refrigeration
engineering in that the temperatures
involved in cryo engineering are in the
range below 120 K (< -153 C). Here we
are dealing with two questions:
a) What cooling principle is used in cryo
engineering or in cryopumps and how
is the thermal load of the cold surface
lead away or reduced?
b) What are the operating principles of the
cryopumps?
2.1.9.1 Types of cryopump
Depending on the cooling principle a
difference is made between
Bath cryostats
Continuous flow cryopumps
Refrigerator cryopumps
In the case of bath cryostats in the most
simple case a cold trap filled with LN
2
(li-
quid nitrogen) the pumping surface is
cooled by direct contact with a liquefied
gas. On a surface cooled with LN
2
(T 77 K) H
2
O and CO
2
are able to
condense. On a surface cooled to 10 K
all gases except He and Ne may be
pumped by way of condensation. A
surface cooled with liquid helium
(T 4.2 K) is capable of condensing all
gases except helium.
In continuous flow cryopumps the cold
surface is designed to operate as a heat
exchanger. Liquid helium in sufficient
quantity is pumped by an auxiliary pump
from a reservoir into the evaporator in order
to attain a sufficiently low temperature at
the cold surface (cryopanel).
The liquid helium evaporates in the heat
exchanger and thus cools down the
cryopanel. The waste gas which is
generated (He) is used in a second heat
exchanger to cool the baffle of a thermal
radiation shield which protects the system
from thermal radiation coming from the
outside. The cold helium exhaust gas
ejected by the helium pump is supplied to
a helium recovery unit. The temperature at
the cryopanels can be controlled by
controlling the helium flow.
Today refrigerator cryopumps are being
used almost exclusively (cold upon
demand). These pumps operate basically
much in the same way as a common
household refrigerator, whereby the
following thermodynamic cycles using
helium as the refrigerant may be employed:
Gifford-McMahon process
Stirling process
Brayton process
Claude process
The Gifford-McMahon process is mostly
used today and this process is that which
has been developed furthest. It offers the
possibility of separating the locations for
the large compressor unit and the
expansion unit in which the refrigeration
process takes place. Thus a compact and
low vibration cold source can be designed.
The cryopumps series-manufactured by
LEYBOLD operate with two-stage cold
heads according to the Gifford-McMahon
process which is discussed in detail in the
following.
The entire scope of a refrigerator cryopump
is shown in Fig. 2.65 and consists of the
compressor unit (1) which is linked via
flexible pressure lines (2) and thus
vibration-free to the cryopump (3). The
cryopump itself consists of the pump
casing and the cold head within. Helium is
used as the refrigerant which circulates in a
closed cycle with the aid of the compressor.
D00
Fig. 2.65 All items of a refrigerator cryopump
1 Compressor unit
2 Flexible pressure
lines
3 Cold head (without
condensation
surfaces)
3
1
2
D00 E 19.06.2001 21:37 Uhr Seite 49
Vacuum Generation
2.1.9.2 The cold head and its opera-
ting principle (Fig. 2.66)
Within the cold head, a cylinder is divided
into two working spaces V
1
and V
2
by a
displacer. During operation the right space
V
1
is warm and the left space V
2
is cold. At
a displacer frequency f the refrigerating
power W of the refrigerator is:
W = (V
2,max
V
2,min
) (p
H
p
N
) f (2.26)
The displacer is moved to and fro pneuma-
tically so that the gas is forced through the
displacer and thus through the regenerator
located inside the displacer. The regenera-
tor is a heat accumulator having a large
heat exchanging surface and capacity,
which operates as a heat exchanger within
the cycle. Outlined in Fig. 2.66 are the four
phases of refrigeration in a single-stage
refrigerator cold head operating according
to the Gifford-McMahon principle.
The two-stage cold head
The series-manufactured refrigerator
cryopumps from LEYBOLD use a two-
stage cold head operating according to the
Gifford-McMahon principle (see Fig. 2.67).
In two series connected stages the
temperature of the helium is reduced to
about 30 K in the first stage and further to
about 10 K in the second stage. The
attainable low temperatures depend
among other things on the type of
regenerator. Commonly copperbronze is
used in the regenerator of the first stage
and lead in the second stage. Other
materials are available as regenerators for
special applications like cryostats for
extremely low temperatures (T < 10 K).
The design of a two-stage cold head is
shown schematically in Fig. 2.67. By
Fig. 2.66 Refrigerating phases using a single-stage cold head operating according to the Gifford-McMahon process
Phase 1:
The displacer is at the left dead center;
V
2
where the cold is produced has its
minimum size. Valve N remains closed,
H is opened. Gas at the pressure pH
flows through the regenerator into V
2
.
There the gas warms up by the pressure
increase in V
1
.
Phase 2:
Valve H remains open, valve N closed:
the displacer moves to the right and
ejects the gas from V
1
through the rege-
nerator to V
2
where it cools down at the
cold regenerator.; V
2
has its maximum
volume.
Phase 3:
Valve H is closed and the valve N to the
low pressure reservoir is opened. The
gas expands from p
H
to p
N
and thereby
cools down. This removes heat from the
vicinity and it is transported with the
expanding gas to the compressor.
Phase 4:
With valve N open the displacer moves
to the left; the gas from V
2,max
flows
through the regenerator, cooling it down
and then flows into the volume V
1
and
into the low pressure reservoir. This
completes the cycle.
V2 (cold) V1 (warm)
Regenerator Displacer
V2 (cold) V1 (warm)
V2 (cold) V1 (warm)
V2 (cold) V1
D00 E 19.06.2001 21:37 Uhr Seite 50
Vacuum Generation
means of a control mechanism with a
motor driven control valve (18) with
control disk (17) and control holes first the
pressure in the control volume (16) is
changed which causes the displacers (6)
of the first stage and the second stage (11)
to move; immediately thereafter the
pressure in the entire volume of the
cylinder is equalized by the control
mechanism. The cold head is linked via
flexible pressure lines to the compressor.
2.1.9.3 The refrigerator cryopump
Fig. 2.68 shows the design of a cryopump.
It is cooled by a two-stage cold head. The
thermal radiation shield (5) with the baffle
(6) is closely linked thermally to the first
stage (9) of the cold head. For pressures
below 10
-3
mbar the thermal load is
caused mostly by thermal radiation. For
this reason the second stage (7) with the
condensation and cryosorption panels (8)
is surrounded by the thermal radiation
shield (5) which is black on the inside and
polished as well as nickel plated on the
outside. Under no-load conditions the
baffle and the thermal radiation shield
(first stage) attain a temperature ranging
between 50 to 80 K at the cryopanels and
about 10 K at the second stage. The
surface temperatures of these cryopanels
are decisive to the actual pumping
process. These surface temperatures
depend on the refrigerating power
supplied by the cold head, and the thermal
conduction properties in the direction of
the pumps casing. During operation of the
cryopump, loading caused by the gas and
the heat of condensation results in further
warming of the cryopanels. The surface
temperature does not only depend on the
temperature of the cryopanel, but also on
the temperature of the gas which has
already been frozen on to the cryopanel.
The cryopanels (8) attached to the second
stage (7) of the cold head are coated with
activated charcoal on the inside in order to
be able to pump gases which do not easily
condense and which can only be pumped
by cryosorption (see 2.1.9.4).
2.1.9.4 Bonding of gases to cold
surfaces
The thermal conductivity of the condensed
(solid) gases depends very much on their
structure and thus on the way in which the
condensate is produced. Variations in
thermal conductivity over several orders of
magnitude are possible! As the condensa-
te increases in thickness, thermal resis-
tance and thus the surface temperature
increase subsequently reducing the pum-
ping speed. The maximum pumping speed
of a newly regenerated pump is stated as
its nominal pumping speed. The bonding
process for the various gases in the
cryopump is performed in three steps: first
D00
Fig. 2.67 Two-stage cold head
1 Electric connections and current
feedthrough for the motor in the
cold head
2 He high pressure connection
3 He low pressure connection
4 Cylinder, 1st stage
5 Displacer, 1st stage
6 Regenerator, 1st stage
7 Expansion volume, 1st stage
8 1st (cooling) stage (copper
flange)
9 Cylinder, 2nd stage
10 Displacer, 2nd stage
11 Regenerator, 2nd stage
12 Expansion volume, 2nd stage
13 2nd (cooling) stage
(copper flange)
14 Measurement chamber for the
vapor pressure
15 Control piston
16 Control volume
17 Control disk
18 Control valve
19 Gauge for the hydrogen vapor
pressure thermometer
20 Motor in the cold head
Fig. 2.68 Design of a refrigerator cryopump (schematic)
2 Pump casing
3 Forevacuum flange
4 Safety valve for gas discharge
5 Thermal radiation shield
6 Baffle
7 2nd stage of the cold head
(10 K);
8 Cryopanels
9 1st stage of the cold head
( 50 80 K)
10 Gauge for the hydrogen vapor
pressure thermometer
11 Helium gas connections
12 Motor of the cold head
with casing and electric
connections
D00 E 19.06.2001 21:37 Uhr Seite 51
Vacuum Generation
the mixture of different gases and vapors
meets the baffle which is at a temperature
of about 80 K. Here mostly H
2
O and CO
2
are condensed. The remaining gases
penetrate the baffle and impinge in the
outside of the cryopanel of the second
stage which is cooled to about 10 K. Here
gases like N
2
, O
2
or Ar will condense. Only
H
2
, He and Ne will remain. These gases
can not be pumped by the cryopanels and
these pass after several impacts with the
thermal radiation shield to the inside of
these panels which are coated with an
adsorbent (cryosorption panels) where
they are bonded by cryosorption. Thus for
the purpose of considering a cryopump
the gases are divided into three groups
depending at which temperatures within
the cryopump their partial pressure drops
below 10
-9
mbar:
1st group:
p
s
< 10
-9
mbar at T 77K (LN
2
): H
2
O, CO
2
2nd group:
p
s
< 10
-9
mbar at T 20K: N
2
, O
2
, Ar
3rd group:
p
s
< 10
-9
mbar at T < 4.2K: H
2
, He, Ne
A difference is made between the different
bonding mechanisms as follows:
Cryocondensation is the physical and re-
versible bonding of gas molecules through
Van der Waals forces on sufficiently cold
surfaces of the same material. The bond
energy is equal to the energy of vaporiza-
tion of the solid gas bonded to the surface
and thus decreases as the thickness of the
condensate increases as does the vapor
pressure. Cryosorption is the physical and
reversible bonding of gas molecules
through Van der Waals forces on suffici-
ently cold surfaces of other materials. The
bond energy is equal to the heat of
adsorption which is greater than the heat
of vaporization. As soon as a monolayer
has been formed, the following molecules
impinge on a surface of the same kind
(sorbent) and the process transforms into
cryocondensation. The higher bond energy
for cryocondensation prevents the further
growth of the condensate layer thereby
restricting the capacity for the adsorbed
gases. However, the adsorbents used, like
activated charcoal, silica gel, alumina gel
and molecular sieve, have a porous
structure with very large specific surface
areas of about 10
6
m
2
/kg. Cryotrapping is
understood as the inclusion of a low
boiling point gas which is difficult to pump
such as hydrogen, in the matrix of a gas
having a higher boiling point and which
can be pumped easily such as Ar, CH
4
or
CO
2
. At the same temperature the
condensate mixture has a saturation vapor
pressure which is by several orders of
magnitude lower than the pure condensate
of the gas with the lower boiling point.
2.1.9.5 Pumping speed and position of
the cryopanels
Considering the position of the cryopanels
in the cryopump, the conductance from
the vacuum flange to this surface and also
the subtractive pumping sequence (what
has already condensed at the baffle can
not arrive at the second stage and
consume capacity there), the situation
arises as shown in Fig. 2.69.
The gas molecules entering the pump pro-
duce the area related theoretical pumping
speed according the equation 2.29a with T
= 293 K. The different pumping speeds
have been combined for three representa-
tive gases H
2
, N
2
and H
2
0 taken from each
of the aforementioned groups. Since water
vapor is pumped on the entire entry area of
the cryopump, the pumping speed
measured for water vapor corresponds
almost exactly to the theoretical pumping
speed calculated for the intake flange of
the cryopump. N
2
on the other hand must
first overcome the baffle before it can be
bonded on to the cryocondensation panel.
Depending on the design of the baffle, 30
to 50 percent of all N
2
molecules are
reflected.
H
2
arrives at the cryosorption panels after
further collisions and thus cooling of the
gas. In the case of optimally designed
cryopanels and a good contact with the ac-
tive charcoal up to 50 percent of the H
2
which has overcome the baffle can be
bonded. Due to the restrictions regarding
access to the pumping surfaces and
cooling of the gas by collisions with the
walls inside the pump before the gas
reaches the pumping surface, the measu-
red pumping speed for these two gases
amounts only to a fraction of the theoreti-
cal pumping speed. The part which is not
pumped is reflected chiefly by the baffle.
Moreover, the adsorption probability for
H
2
differs between the various adsorbents
and is < 1, whereas the probabilities for the
condensation of water vapor and N
2
1.
Three differing capacities of a pump for the
gases which can be pumped result from
the size of the three surfaces (baffle,
condensation surface at the outside of the
second stage and sorption surface at the
inside of the second stage). In the design
of a cryopump, a mean gas composition
(air) is assumed which naturally does not
apply to all vacuum processes (sputtering
processes, for example. See 2.1.9.6
Partial Regeneration).
Fig. 2.69 Cryopanels temperature and position define the efficiency in the cryopump
Hydrogen Water vapor Nitrogen
Area related conductance of the intake flange in l / s cm
2
:
43.9 14.7 11.8
Area related pumping speed of the cryopump in l / s cm
2
:
13.2 14.6 7.1
Ratio between pumping speed and conductance:
30 % 99 % 60 %
D00 E 19.06.2001 21:37 Uhr Seite 52
Vacuum Generation
2.1.9.6 Characteristic quantities of a
cryopump
The characteristic quantities of a cryo-
pump are as follows (in no particular
order):
Cooldown time
Crossover value
Ultimate pressure
Capacity
Refrigerating power and net
refrigerating power
Regeneration time
Throughput and maximum pV flow
Pumping speed
Service life / duration of operation
Starting pressure
Cooldown time: The cooldown time of
cryopumps is the time span from start-up
until the pumping effect sets in. In the case
of refrigerator cryopumps the cooldown
time is stated as the time it takes for the
second stage of the cold head to cool
down from 293 K to 20 K.
Crossover value: The crossover value is a
characteristic quantity of an already cold
refrigerator cryopump. It is of significance
when the pump is connected to a vacuum
chamber via an HV / UHV valve. The
crossover value is that quantity of gas with
respect to T
n
= 293 K which the vacuum
chamber may maximally contain so that
the temperature of the cryopanels does
not increase above 20 K due to the gas
burst when opening the valve. The
crossover value is usually stated as a pV
value in in mbar l.
The crossover value and the chamber
volume V result in the crossover pressure
pc to which the vacuum chamber must be
evacuated first before opening the valve
leading to the cryopump. The following
may serve as a guide:
(2.27)
V = Volume of the vacuum chamber (l),
Q
.
2
(20K) = Net refrigerating capacity in
Watts, available at the second stage of the
cold head at 20 K.
p
c
V
Q K mbar

35
20
2
( )
Ultimate pressure p
end
: For the case of
cryocondensation (see Section 2.1.9.4)
the ultimate pressure can be calculated by:
(2.28)
p
S
is the saturation vapor pressure of the
gas or gases which are to be pumped at
the temperature T
K
of the cryopanel and T
G
is the gas temperature (wall temperature in
the vicinity of the cryopanel).
Example: With the aid of the vapor
pressure curves in Fig. 9.15 for H
2
and N
2
the ultimate pressures summarized in
Table 2.6 at T
G
= 300 K result.
The Table shows that for hydrogen at
temperatures T < 3 K at a gas temperature
of T
G
= 300 K (i.e. when the cryopanel is
exposed to the thermal radiation of the
wall) sufficiently low ultimate pressures
can be attained. Due to a number of inter-
fering factors like desorption from the wall
and leaks, the theoretical ultimate pres-
sures are not attained in practice.
Capacity C (mbar l): The capacity of a
cryopump for a certain gas is that quantity
of gas (p
V
value at T
n
= 293 K) which can
be bonded by the cryopanels before the
pumping speed for this type of gas G
drops to below 50 % of its initial value.
The capacity for gases which are pumped
by means of cryosorption depends on the
quantity and properties of the sorption
agent; it is pressure dependent and
generally by several orders of magnitude
lower compared to the pressure
independent capacity for gases which are
pumped by means of cryocondensation.
Refrigerating power Q
.
(W): The refrige-
rating power of a refrigeration source at a
temperature T gives the amount of heat
that can be extracted by the refrigerating
source whilst still maintaining this
temperature. In the case of refrigerators it
has been agreed to state for single-stage
cold heads the refrigerating power at 80 K
and for two-stage cold heads the
( ) p p T
T
T
end s K
G
K
=
refrigerating power for the first stage at 80
K and for the second stage at 20 K when
simultaneously loading both stages
thermally. During the measurement of
refrigerating power the thermal load is
generated by electric heaters. The refrige-
rating power is greatest at room tempe-
rature and is lowest (Zero) at ultimate tem-
perature.
Net refrigerating power Q
.
(W): In the
case of refrigerator cryopumps the net
refrigerating power available at the usual
operating temperatures
(T1 < 80 K, T2 < 20 K) substantially defines
the throughput and the crossover value.
The net. refrigerating power is depen-
ding on the configuration of the pump
much lower than the refrigerating power of
the cold head used without the pump.
pV flow see 1.1
Regeneration time: As a gas trapping
device, the cryopump must be regenerated
after a certain period of operation. Rege-
neration involves the removal of condens-
ed and adsorbed gases from the cryopa-
nels by heating. The regeneration can be
run fully or only partially and mainly differs
by the way in which the cryopanels are
heated.
In the case of total regeneration a diffe-
rence is made between:
1. Natural warming: after switching off the
compressor, the cryopanels at first
warm up only very slowly by thermal
conduction and then in addition
through the released gases.
2. Purge gas method: the cryopump is
warmed up by admitting warm purging
gas.
3. Electric heaters: the cryopanels of the
cryopump are warmed up by heaters at
the first and second stages. The
released gases are discharged either
through an overpressure valve (purge
gas method) or by mechanical backing
pumps. Depending on the size of the
pump one will have to expect a
regeneration time of several hours.
D00
T
K
(K) Ultimate pressure Ult. press. (mbar) Ult. press.(mbar)
(according to equ. 2.28) H
2
N
2
2.5 10.95 p
s
3.28 10
14
immeasurably low
4.2 8.66 p
s
4.33 10
9
immeasurably low
20 3.87 p
s
3.87 10
+3
3.87 10
11
Table 2.6 Ultimate temperatures at a wall temperature of 300 K
D00 E 19.06.2001 21:37 Uhr Seite 53
Vacuum Generation
Partial regeneration: Since the limitation
in the service life of a cryopump depends
in most applications on the capacity limit
for the gases nitrogen, argon and
hydrogen pumped by the second stage, it
will often be required to regenerate only
this stage. Water vapor is retained during
partial regeneration by the baffle. For this,
the temperature of the first stage must be
maintained below 140 K or otherwise the
partial pressure of the water vapor would
become so high that water molecules
would contaminate the adsorbent on the
second stage.
In 1992, LEYBOLD was the first manufac-
turer of cryopumps to develop a method
permitting such a partial regeneration.
This fast regeneration process is micro-
processor controlled and permits a partial
regeneration of the cryopump in about 40
minutes compared to 6 hours needed for a
total regeneration based on the purge gas
method. A comparison between the typical
cycles for total and partial regeneration is
shown in Fig. 2.70. The time saved by the
Fast Regeneration System is apparent. In a
production environment for typical sput-
tering processes one will have to expect
one total regeneration after 24 partial
regenerations.
Throughput and maximum pV flow:
(mbar l/s): the throughput of a cryopump
for a certain gas depends on the pV flow of
the gas G through the intake opening of
the pump:
Q
G
= q
pV,G;
the following equation
applies
Q
G
= p
G
S
G
with
p
G
= intake pressure,
S
G
= pumping capacity for the gas G
The maximum pV flow at which the
cryopanels are warmed up to T 20 K in
the case of continuous operation, de-
pends on the net refrigerating power of the
pump at this temperature and the type of
gas. For refrigerator cryopumps and
condensable gases the following may be
taken as a guide:
Q
max
= 2.3 Q
.
2 (20 K) mbar l/s
Q
.
2
(20 K) is the net refrigerating power in
Watts available at the second stage of the
cold heat at 20 K. In the case of inter-
mittent operation, a higher pV flow is per-
missible (see crossover value).
Pumping speed S
th
: The following applies
to the (theoretical) pumping speed of a
cryopump:
(2.29)
A
K
Size of the cryopanels
S
A
Surface area related pumping speed
(area related impact rate according to
equations 1.17 and 1.20, proportional
to the mean velocity of the gas
molecules in the direction of the
cryopanel)
a Probability of condensation
(pumping)
p
end
Ultimate pressure (see above)
p Pressure in the vacuum chamber
The equation (2.29) applies to a cryopanel
built into the vacuum chamber, the surface
area of which is small compared to the
surface of the vacuum chamber. At
sufficiently low temperatures = 1 for all
gases. The equation (2.29) shows that for
S A S
p
p
th K A
end
=
_
,
1
p >> pend the expression in brackets
approaches 1 so that in the oversaturated
case
p >> pend > Ps so that:
S
th
= A
K
S
A
(2.29a)
with
T
G
Gas temperature in K
M Molar mass
Given in Table 2.7 is the surface area-
related pumping speed S
A
in l s
-1
cm
-2
for some gases at two different gas
temperatures T
G
in K determined accor-
ding to equation 2.29a. The values stated
in the Table are limit values. In practice the
condition of an almost undisturbed
equilibrium (small cryopanels compared
to a large wall surface) is often not true,
because large cryopanels are required to
attain short pumpdown times and a good
end vacuum. Deviations also result when
the cryopanels are surrounded by a cooled
baffle at which the velocity of the
penetrating molecules is already reduced
by cooling.
Service life / duration of operation t
op
(s):
The duration of operation of the cryopump
for a particular gas depends on the
equation:
with
C
G
= Capacity of the cryopump for the
gas G
Q
G
(t) =Throughput of the cryopump for
the gas at the point of time t
If the constant mean over time for the
throughput Q
G
__
is known, the following
applies:
(2.30)
After the period of operation t
op,G
has
elapsed the cryopump must be rege-
nerated with respect to the type of gas G.
Starting pressure p
o
: Basically it is
possible to start a cryopump at atmosphe-
ric pressure. However, this is not desirable
for several reasons. As long as the mean
free path of the gas molecules is smaller
t
C
Q
C
p S
op G
G
G
G
G G
,
= =
C Q t dt
G G
t
op G
( )
,
0
S
R T
M
c
T
M
s cm
A
G G
=

= =
2 4
365
2
. /
Fig. 2.70 Comparison between total (1) and partial (2) regeneration
Time
(2)
(1)
about 6 hours 40 minutes
T
e
m
p
e
r
a
t
u
r
e

(
K
)
2 1
D00 E 19.06.2001 21:37 Uhr Seite 54
Vacuum Generation
than the dimensions of the vacuum cham-
ber (p > 10
-3
mbar), thermal conductivity
of the gas is so high that an unacceptably
large amount of heat is transferred to the
cryopanels. Further, a relatively thick layer
of condensate would form on the cryo-
panel during starting. This would markedly
reduce the capacity of the cryopump
available to the actual operating phase.
Gas (usually air) would be bonded to the
adsorbent, since the bonding energy for
this is lower than that for the condensation
surfaces. This would further reduce the
already limited capacity for hydrogen. It is
recommended that cryopumps in the high
vacuum or ultrahigh vacuum range are
started with the aid of a backing pump at
pressures of p < 5 10
-2
mbar. As soon as
the starting pressure has been attained the
backing pump may be switched off.
D00
M S
A
S
A
T
S
Triple point
Symbol Gas Molar at 293 K at 80 K Boiling point (= melting point)
mass gas temp. gas temp. 1013 mbar T
t
p
t
g/mol l/s cm
2
l/s cm
2
K K mbar
H
2
Hydrogen 2.016 43.88 22.93 20.27 13.80 70.4
H
e
Helium 4.003 31.14 16.27 4.222 2.173 50.52
CH
4
Methane 4.003 15.56 8.13 111.67 90.67 116.7
H
2
O Water 18.015 14.68 373.15 273.16 6.09
N
e
Neon 20.183 13.87 7.25 27.102 24.559 433.0
CO Carbon monoxide 28.000 11.77 6.15 81.67 68.09 153.7
N
2
Nitrogen 28.013 11.77 6.15 77.348 63.148 126.1
Air 28.96 11.58 6.05 80.5 58.5
O
2
Oxygen 31.999 11.01 5.76 90.188 54.361 1.52
Ar Argon 39.948 9.86 5.15 87.26 83.82 687.5
Kr Krypton 83.80 6.81 3.56 119.4 115.94 713.9
Xe Xenon 131.3 5.44 2.84 165.2 161.4
Table 2.7 Surface-related pumping speeds for some gases
D00 E 19.06.2001 21:37 Uhr Seite 55
Vacuum Generation
2.2 Choice of pumping
process
2.2.1 Survey of the most usual
pumping processes
Vacuum technology has undergone rapid
development since the Fifties. In research
and in most branches of industry today, it
is indispensable.
Corresponding to the many areas of ap-
plication, the number of technical proce-
dures in vacuum processes is extraordi-
narily large. These cannot be described
within the scope of this section, because
the basic calculations in this section cover
mainly the pumping process, not the
process taking place in the vessel. A
survey of the most important processes in
vacuum technology and the pressure
regions in which these processes are
chiefly carried out is given in the diagrams
(Figures 2.71 and 2.72).
Generally, the pumping operation for these
processes can be divided into two
categories dry- and wet vacuum proce-
dures, that is, into processes in which no
significant amounts of vapor have to be
pumped and those in which vapors
(mostly water or organic) arise.
Distinctions between the two categories
are described briefly:
Dry processes work primarily in a narrow
and limited pressure region.
The system is usually evacuated to a
suitable characteristic pressure before the
actual working process begins. This
happens, for example, in plants for evapo-
rative coating, electron-beam welding, and
crystal pulling; in particle accelerators,
mass spectrometers, electron micro-
scopes; and others.
Further, there are dry processes in which
degassing in vacuum is the actual
technical process. These include work in
induction- and arc furnaces, steel degas-
sing plants, and plants for the manufacture
of pure metals and electron tubes.
Wet processes are undertaken primarily
in a prescribed working operation that
covers a wider pressure region.
This is especially important in the drying of
solid materials. If, for instance, work is
undertaken prematurely at too low a
pressure, the outer surfaces dry out too
quickly. As a result, the thermal contact to
the moisture to be evaporated is impaired
and the drying time is considerably
increased. Predominantly processes that
are carried out in drying, impregnating,
and freeze-drying plants belong in this
category.
In the removal of water vapor from liquids
or in their distillation, particularly in
degassing columns, vacuum filling, and
resin-casting plants, as well as in
molecular distillation, the production of as
large a liquid surface as possible is
important. In all wet processes the provi-
sion of the necessary heat for evapo-
ration of the moisture is of great impor-
tance.
Basic pumping procedures are given in the
following paragraphs.
If you have specific questions, you should
get in touch with a specialist department in
LEYBOLD where experts are available to
you who can draw on many years of
experience.
Classifications of typical vacuum proces-
ses and plants according to the pressure
regions.
Rough vacuum 1013 mbar 1 mbar
Drying, distillation, and steel degassing.
Medium vacuum 1 -10
-3
mbar
Molecular distillation, freeze-drying,
impregnation, melting and casting
furnaces, and arc furnaces.
High vacuum 10
-3
10
-7
mbar
Evaporative coating, crystal pulling,
mass spectrometers, tube production,
electron microscopes, electron beam
plants, and particle accelerators.
Ultrahigh vacuum: < 10
-7
mbar
Nuclear fusion, storage rings for
accelerators, space research, and
surface physics.
Rough vacuum Medium vacuum High vacuum Ultrahigh vacuum
Mass spectrometers
Molecular beam apparatus
Ion sources
Particle accelerators
Electron microscopes
Electron diffraction apparatus
Vacuum spectographs
Low-temperature research
Production of thin films
Surface physics
Plasma research
Nuclear fusion apparatus
Space simulation
Material research
Preparations for
electron microscopy
10
13
10
10
10
7
10
3
10
0
10
3
Pressure [mbar]
Fig. 2.71 Pressure ranges (p < 1000 mbar) of physical and chemical analytical methods
D00 E 19.06.2001 21:37 Uhr Seite 56
Vacuum Generation
2.2.2 Pumping of gases
(dry processes)
For dry processes in which a non conden-
sable gas mixture (e.g., air) is to be
pumped, the pump to be used is clearly
characterized by the required working
pressure and the quantity of gas to be
pumped away. The choice of the required
working pressure is considered in this
section. The choice of the required pump
is dealt with in Section 2.3.
Each of the various pumps has a charac-
teristic working range in which it has a
particularly high efficiency. Therefore, the
most suitable pumps for use in the follo-
wing individual pressure regions are des-
cribed. For every dry-vacuum process, the
vessel must first be evacuated. It is quite
possible that the pumps used for this may
be different from those that are the
optimum choices for a process that is
undertaken at definite working pressures.
In every case the choice should be made
with particular consideration for the
pressure region in which the working
process predominantly occurs.
a) Rough vacuum (1013 1 mbar)
The usual working region of the rotary
pumps described in Section 2 lies below
80 mbar. At higher pressures these pumps
have a very high power consumption (see
Fig. 2.11) and a high oil consumption (see
Section 8.3.1.1). Therefore, if gases are to
be pumped above 80 mbar over long
periods, one should use, particularly on
economic grounds, jet pumps, water ring
pumps or dry running, multi-vane pumps.
Rotary vane and rotary piston pumps are
especially suitable for pumping down
vessels from atmospheric pressure to
pressures below 80 mbar, so that they can
work continuously at low pressures. If
large quantities of gas arise at inlet
pressures below 40 mbar, the connection
in series of a Roots pump is recommen-
ded. Then, for the backing pump speed
required for the process concerned, a
much smaller rotary vane or piston pump
can be used.
b) Medium vacuum (1 10
-3
mbar)
If a vacuum vessel is merely to be eva-
cuated to pressures in the medium vacu-
um region, perhaps to that of the required
backing pressure for diffusion or sputter-
ion pumps, single- and two-stage rotary
pumps are adequate for pressures down to
10
-1
and 10
-3
mbar, respectively. It is es-
sentially more difficult to select the
suitable type of pump if medium vacuum
processes are concerned in which gases
or vapors are evolved continuously and
must be pumped away. An important hint
may be given at this point. Close to the
attainable ultimate pressure, the pumping
speed of all rotary pumps falls off rapidly.
Therefore, the lowest limit for the normal
working pressure region of these pumps
should be that at which the pumping speed
still amounts to about 50 % of the nominal
pumping speed.
Between 1 and 10
-2
mbar at the onset of
large quantities of gas, Roots pumps with
rotary pumps as backing pumps have
optimum pumping properties (see Section
2.1.3.1). For this pressure range, a single-
stage rotary pump is sufficient, if the chief
working region lies above 10
-1
mbar. If it
lies between 10
-1
and 10
-2
mbar, a two-
stage backing pump is recommended.
Below 10
-2
mbar the pumping speed of
single-stage Roots pumps in combination
with two-stage rotary pumps as backing
pumps decreases. However, between 10
-2
and 10
-4
mbar, two-stage Roots pumps (or
two single-stage Roots pumps in series)
with two-stage rotary pumps as backing
pumps still have a very high pumping
speed. Conversely, this pressure region is
the usual working region for vapor ejector
pumps. For work in this pressure region,
they are the most economical pumps to
purchase. As backing pumps, single-stage
rotary positive displacement pumps are
suitable. If very little maintenance and val-
veless operation are convenient (i.e., small
vessels in short operation cycles are to be
pumped to about 10
-4
mbar or large
vessels are to be maintained at this pres-
sure unattended for weeks), the previously
mentioned two-stage Roots pumps with
two-stage rotary pumps as backing pumps
are the suitable combinations. Allthough,
such a combination does not work as
D00
10
10
10
7
10
3
10
0
10
3
Pressure [mbar]
Annealing of metals
Melting of metals
Degassing of molten metals
Steel degassing
Electron-beam melting
Electron-beam welding
Evaporation coating
Sputtering of metals
Zone melting and crystal growing in high-vacuum
Molecular distillation
Degassing of liquids
Sublimation
Casting of resins and lacquers
Drying of plastics
Drying of insulating papers
Freeze-drying of mass materials
Freeze-drying of pharmaceutical products
Production of incandescent lamps
Production of electron tubes
Production of gas-discharge tubes
Fig. 2.72 Pressure ranges of industrial vacuum processes
D00 E 19.06.2001 21:37 Uhr Seite 57
Vacuum Generation
economically as the corresponding vapor
ejector pump, it can operate for a much
longer time without maintenance.
c) High vacuum (10
-3
to 10
-7
mbar)
Diffusion, sputter-ion, and turbomolecular
pumps typically operate in the pressure
region below 10
-3
mbar. If the working
region varies during a process, different
pumping systems must be fitted to the
vessel. There are also special diffusion
pumps that combine the typical properties
of a diffusion pump (low ultimate pres-
sure, high pumping speed in the high
vacuum region) with the outstanding
properties of a vapor ejector pump (high
throughput in the medium vacuum region,
high critical backing pressure). If the
working region lies between 10
-2
and 10
-6
mbar, these diffusion pumps are, in
general, specially recommended.
d) Ultrahigh vacuum (< 10
-7
mbar)
For the production of pressures in the
ultrahigh vacuum region, sputter-ion, and
sublimation pumps, as well as turbomole-
cular pumps and cryopumps, are used in
combination with suitable forepumps. The
pump best suited to a particular UHV
process depends on various conditions
(for further details, see Section 2.5).
2.2.3 Pumping of gases and
vapors (wet processes)
When vapors must be pumped, in addition
to the factors working pressure and
pumping speed, a third determining factor
is added namely the vapor partial pressure
which may vary considerably in the
course of a process. This factor is decisive
in determining the pumping arrangement
to be installed. In this regard, the con-
densers described in Section 2.15 are very
important adjuncts to rotary displacement
pumps. They have a particularly high
pumping speed when pumping vapors.
The next section covers pumping of water
vapor (the most frequent case). The
considerations apply similarly to other
non-aggressive vapors.
Pumping of Water Vapor
Water vapor is frequently removed by
pumps that operate with water or steam as
a pump fluid, for example, water ring
pumps or steam ejector pumps. This de-
pends considerably on circumstances,
however, because the economy of steam
ejector pumps at low pressures is gene-
rally far inferior to that of rotary pumps.
For pumping a vapor gas mixture in
which the vapor portion is large but the air
portion is small, the vapor can be pumped
by condensers and the permanent gases,
by relatively small gas ballast pumps (see
Section 2.1.5).
Comparatively, then, a pump set con-
sisting of a Roots pump, condenser, and
backing pump, which can transport 100
kg/h of vapor and 18 kg/h of air at an inlet
pressure of 50 mbar, has a power
requirement of 4 10 kW (depending on
the quantity of air involved). A steam
ejector pump of the same performance
requires about 60 kW without altering the
quantity of air involved.
For the pumping of water vapor, gas ballast
pumps and combinations of gas ballast
pumps, Roots pumps, and condensers are
especially suitable.
Pumping of water vapor with gas ballast
pumps
The ratio of vapor partial pressure p
v
to air
partial pressure p
p
is decisive in the
evaluation of the correct arrangement of
gas ballast pumps, as shown previously by
equations 2.2 and 2.3. Therefore, if the
water vapor tolerance of the gas ballast
pump is known, graphs may be obtained
that clearly give the correct use of gas
ballast pumps for pumping water vapor
(see Fig. 2.73). Large single-stage rotary
plunger pumps have, in general, an
operating temperature of about 77 C and
hence a water vapor tolerance of about 60
mbar. This value is used to determine the
different operating regions in Fig. 2.73. In
addition it is assumed that the pressure at
the discharge outlet port of the gas ballast
pump can increase to a maximum of 1330
mbar until the discharge outlet valve
opens.
Region A: Single-stage, rotary plunger
pumps without gas ballast inlet.
At a saturation vapor pressure p
S
of 419
mbar at 77 C, according to equation 2.2,
the requirement is given that p
v
< 0.46 p
p
,
where
p
v
is the water vapor partial pressure
p
p
is the partial pressure of air
p
v
+ p
p
= p
tot
total pressure
This requirement is valid in the whole
working region of the single-stage rotary
plunger pump hence, at total pressures
between 10
-1
and 1013 mbar.
Region B: Single-stage rotary plunger
pumps with gas ballast and an inlet
condenser.
In this region the water vapor pressure
exceeds the admissible partial pressure at
the inlet. The gas ballast pump must, Fig. 2.73 Areas of application for gas ballast pumps and condensers pumping water vapor (o.G. = without gas ballast)
W
a
t
e
r

v
a
p
o
r

p
a
r
t
i
a
l

p
r
e
s
s
u
r
e

p
v
Air partial pressure p
p
p
v
/
p
p

=

0
.
4
6
D00 E 19.06.2001 21:37 Uhr Seite 58
Vacuum Generation
therefore, have a condenser inserted at the
inlet, which is so rated that the water vapor
partial pressure at the inlet port of the
rotary pump does not exceed the
admissible value. The correct dimensions
of the condenser are selected depending
on the quantity of water vapor involved.
For further details, see Section 2.1.5. At a
water vapor tolerance of 60 mbar, the
lower limit of this region is
p
v
> 6O + 0.46 p
p
mbar
Region C: Single-stage rotary plunger
pumps with a gas ballast.
The lower limit of region C is characterized
by the lower limit of the working region of
this pump. It lies, therefore, at about
p
tot
= 1 mbar. If large quantities of vapor
arise in this region, it is often more
economical to insert a condenser: 20 kg of
vapor at 28 mbar results in a volume of
about 1000 m
3
. It is not sensible to pump
this volume with a rotary pump. As a rule
of thumb:
A condenser should always be inserted at
the pumps inlet if saturated water vapor
arises for a considerable time.
As a precaution, therefore, a Roots pump
should always be inserted in front of the
condenser at low inlet pressures so that
the condensation capacity is essentially
enhanced. The condensation capacity does
not depend only on the vapor pressure,
but also on the refrigerant temperature. At
low vapor pressures, therefore, effective
condensation can be obtained only if the
refrigerant temperature is correspondingly
low. At vapor pressures below 6.5 mbar,
for example, the insertion of a condenser
is sensible only if the refrigerant tempera-
ture is less than 0 C. Often at low pres-
sures a gas vapor mixture with
unsaturated water vapor is pumped (for
further details, see Section 2.1.5). In
general, then, one can dispense with the
condenser.
Region D: Two-stage gas ballast pumps,
Roots pumps, and vapor ejector pumps,
always according to the total pressure
concerned in the process.
It must again be noted that the water vapor
tolerance of two-stage gas ballast pumps
is frequently lower than that of correspon-
ding single-stage pumps.
Pumping of water vapor with roots pumps
Normally, Roots pumps are not as
economical as gas ballast pumps for
continuous operation at pressures above
40 mbar. With very large pump sets, which
work with very specialized gear ratios and
are provided with bypass lines, however,
the specific energy consumption is indeed
more favorable. If Roots pumps are
installed to pump vapors, as in the case of
gas ballast pumps, a chart can be given
that includes all possible cases (see Fig.
2.74).
Region A: A Roots pump with a single-
stage rotary plunger pump without gas
ballast.
As there is merely a compression between
the Roots pump and the rotary plunger
pump, the following applies here too:
p
v
< 0.46 p
p
The requirement is valid over the entire
working region of the pump combination
and, therefore, for total pressures between
10
-2
and 40 mbar (or 1013 mbar for Roots
pumps with a bypass line).
Region B: A main condenser, a Roots
pump with a bypass line, an intermediate
condenser, and a gas ballast pump.
This combination is economical only if
large water vapor quantities are to be
pumped continuously at inlet pressures
above about 40 mbar. The size of the main
condenser depends on the quantity of
vapor involved. The intermediate con-
denser must decrease the vapor partial
pressure below 60 mbar. Hence, the gas
ballast pump should be large enough only
to prevent the air partial pressure behind
the intermediate condenser from ex-
ceeding a certain value; for example, if the
total pressure behind the Roots pump
(which is always equal to the total
pressure behind the intermediate conden-
ser) is 133 mbar, the gas ballast pump
must pump at least at a partial air pressure
of 73 mbar, the quantity of air transported
to it by the Roots pump. Otherwise, it must
take in more water vapor than it can
tolerate. This is a basic requirement: the
use of gas ballast pumps is wise only if
air is also pumped!
With an ideally leak-free vessel, the gas
ballast pump should be isolated after the
required operating pressure is reached and
pumping continued with the condenser
only. Section 2.1.5 explains the best
possible combination of pumps and
condensers.
Region C: A Roots pump, an intermediate
condenser, and a gas ballast pump.
The lower limit of the water vapor partial
pressure is determined through the
compression ratio of the Roots pump at
the backing pressure, which is determined
by the saturation vapor pressure of the
condensed water. Also, in this region the
intermediate condenser must be able to
reduce the vapor partial pressure to at
least 60 mbar. The stated arrangement is
D00
Fig. 2.74 Areas of application for Roots pumps and condensers pumping water vapor (o.G. = without gas ballast)
p
W
a
t
e
r

v
a
p
o
r

p
a
r
t
i
a
l

p
r
e
s
s
u
r
e

p
v
p
p v
/
p p
=

0
.
4
6
D00 E 19.06.2001 21:38 Uhr Seite 59
Vacuum Generation
suitable when cooling the condenser
with water at 15 C for water vapor
pressures between about 4 and 40 mbar.
Region D: A roots pump and a gas ballast
pump.
In this region D the limits also depend
essentially on the stages and ratios of
sizes of the pumps. In general, however,
this combination can always be used
between the previously discussed limits
therefore, between 10
-2
and 4 mbar.
2.2.4 Drying processes
Often a vacuum process covers several of
the regions quoted here. In batch drying
the process can, for example (see Fig.
2.74), begin in region A (evacuation of the
empty vessel) and then move through
regions B, C, and D in steps. Then the
course of the process would be as follows:
A. Evacuating the vessel by a gas ballast
pump and a Roots pump with a bypass
line.
B. Connecting the two condensers
because of the increasing vapor pressure
produced by heating the material.
The choice of the pumping System is
decided by the highest vapor partial
pressure occurring and the lowest air
partial pressure at the inlet.
C. Bypassing the main condenser
It will now not have an effect. Instead it
would only be pumped empty by the
pumping system with a further drop in
vapor pressure.
D. Bypassing the intermediate condenser
Roots pumps and gas ballast pumps alone
can now continue pumping. With short-
term drying, the separation of the
condenser filled with condensed water is
particularly important, because the gas
ballast pump would continue to pump
from the condenser the previously
condensed water vapor at the saturation
vapor pressure of water.
With longer-term drying processes, it
suffices to shut off the condensate
collector from the condenser. Then only
the remaining condensate film on the
cooling tubes can reevaporate. Depending
on the size of the gas ballast pump, this
reevaporation ensues in 30 60 min.
E. If the drying process should terminate
at still lower pressures
When a pressure below 10
-2
mbar is
reached a previously bypassed oil vapor
ejector pump should be switched on in
addition.
Drying of solid substances
As previously indicated, the drying of solid
substances brings about a series of further
problems. It no longer suffices that one
simply pumps out a vessel and then waits
until the water vapor diffuses from the
solid substance. This method is indeed
technically possible, but it would intorably
increase the drying time.
It is not a simple technical procedure to
keep the drying time as short as possible.
Both the water content and the layer thick-
ness of the drying substance are im-
portant. Only the principles can be stated
here. In case of special questions we
advise you to contact our experts in our
Cologne factory.
The moisture content E of a material to be
dried of which the diffusion coefficient
depends on the moisture content (e.g. with
plastics) as a function of the drying time t
is given in close approximation by the
following equation:
(2.31)
E
0
where E is the moisture content before
drying
q is the temperature-dependent coef-
ficient. Thus equation (2.31) serves
only for the temperature at which q
was determined
K is a factor that depends on the
temperature, the water vapor partial
pressure in the vicinity of the material,
the dimensions, and the properties of
the material.
With the aid of this approximate equation,
the drying characteristics of many
substances can be assessed. If K and q
have been determined for various
temperatures and water vapor partial
pressures, the values for other
temperatures are easily interpolated, so
that the course of the drying process can
be calculated under all operating
conditions. With the aid of a similarity
transformation, one can further compare
the course of drying process of a material
( )
E
E
K t
q
=
+
0
1
%
with known properties with that of a
material with different properties.
Fundamentally, in the drying of a
material, a few rules are noteworthy:
Experience has shown that shorter drying
times are obtained if the water vapor
partial pressure at the surface of the
material is relatively high, that is, if the
surface of the material to be dried is not
yet fully free of moisture. This is possible
because the heat conduction between the
source of heat and the material is greater
at higher pressures and the resistance to
diffusion in a moist surface layer is smaller
than in a dry one. To fulfill the conditions
of a moist surface, the pressure in the
drying chamber is controlled. If the neces-
sary relatively high water vapor partial
pressure cannot be maintained perma-
nently, the operation of the condenser is
temporarily discontinued. The pressure in
the chamber then increases and the
surface of the material becomes moist
again. To reduce the water vapor partial
pressure in the vessel in a controlled way,
it may be possible to regulate the re-
frigerant temperature in the condenser. In
this way, the condenser temperature
attains preset values, and the water vapor
partial pressure can be reduced in a con-
trolled manner.
2.2.5 Production of an oil-free
(hydrocarbon-free)
vacuum
Backstreaming vapor pump fluids, vapors
of oils, rotary pump lubricants, and their
cracking products can significantly disturb
various working processes in vacuum.
Therefore, it is recommended that certain
applications use pumps and devices that
reliably exclude the presence of hydro-
carbon vapors.
a) Rough vacuum region
(1013 to 1 mbar)
Instead of rotary pumps, large water jet,
steam ejector, or water ring pumps can be
used. For batch evacuation, and the pro-
duction of hydrocarbon-free fore vacuum
for sputter-ion pumps, adsorption pumps
(see Section 2.1.8.1) are suitable. If the
use of oil-sealed rotary vane pumps
cannot be avoided, basically two-stage
D00 E 19.06.2001 21:38 Uhr Seite 60
Vacuum Generation
rotary vane pumps should be used. The
small amount of oil vapor that back-
streams out of the inlet ports of these
pumps can be almost completely removed
by a sorption trap (see Section 2.1.4)
inserted in the pumping line.
b) Medium vacuum region
(1 to 10
-3
mbar)
For the pumping of large quantities of gas
in this pressure region, vapor ejector
pumps are by far the most suitable. With
mercury vapor ejector pumps, completely
oil-free vacua can be produced. As a pre-
caution, the insertion of a cold trap chilled
with liquid nitrogen is recommended so
that the harmful mercury vapor does not
enter the vessel. With the medium vacuum
sorption traps described under a), it is
possible with two-stage rotary vane
pumps to produce almost oil-free vacua
down to below 10
-4
mbar.
Absolutely oil-free vacua may be produced
in the medium vacuum region with
adsorption pumps. Since the pumping
action of these pumps for the light noble
gases is only small, vessels initially filled
with air can only be evacuated by them to
about 10
-2
mbar. Pressures of 10
-3
mbar
or lower can then be produced with
adsorption pumps only if neither neon nor
helium is present in the gas mixture to be
pumped. In such cases it can be useful to
expel the air in the vessel by first flooding
with nitrogen and then pumping it away.
c) High- and ultrahigh vacuum region
(< 10
-3
mbar)
When there is significant evolution of gas
in the pressure regions that must be pum-
ped, turbomolecular pumps, or cryo-
pumps should be used. A sputter-ion
pump is especially suitable for maintaining
the lowest possible pressure for long
periods in a sealed system where the
process does not release large quantities
of gas. Magnetically-suspended turbomo-
lecular pumps also guarantee hydrocar-
bon-free vacua. However, while these
pumps are switched off, oil vapors can
enter the vessel through the pump. By
suitable means (e.g., using an isolating
valve or venting the vessel with argon),
contamination of the vessel walls can be
impeded when the pump is stationary. If
the emphasis is on generating a hydroc-
arbon-free vacuum with turbomolecular
pumps, then hybrid turbomolecular
pumps with diaphragm pumps or classic
turbomolecular pumps combined with
scroll pumps should be used as oil-free
backing pumps.
2.2.6 Ultrahigh vacuum
working Techniques
The boundary between the high and ultra-
high vacuum region cannot be precisely
defined with regard to the working me-
thods. In practice, a border between the
two regions is brought about because
pressures in the high vacuum region may
be obtained by the usual pumps, valves,
seals, and other components, whereas for
pressures in the UHV region, another
technology and differently constructed
components are generally required. The
border lies at a few 10
-8
mbar. Therefore,
pressures below 10
-7
mbar should ge-
nerally be associated with the UHV region.
The gas density is very small in the UHV
region and is significantly influenced by
outgasing rate of the vessel walls and by
the tiniest leakages at joints. Moreover, in
connection with a series of important
technical applications to characterize the
UHV region, generally the monolayer time
(see also equation 1.21) has become
important. This is understood as the time t
that elapses before a monomolecular or
monatomic layer forms on an initially
ideally cleaned surface that is exposed to
the gas particles. Assuming that every gas
particle that arrives at the surface finds a
free place and remains there, a convenient
formula for is
(p in mbar)
Therefore, in UHV (p < 10
-7
mbar) the
monolayer formation time is of the order
of minutes to hours or longer and thus of
the same length of time as that needed for
experiments and processes in vacuum.
The practical requirements that arise have
become particularly significant in solid-
state physics, such as for the study of thin
films or electron tube technology. A UHV
system is different from the usual high
vacuum system for the following reasons:
a) the leak rate is extremely small (use of
metallic seals),
b) the gas evolution of the inner surfaces
=
3 2
10
6
.
p
s
of the vacuum vessel and of the
attached components (e.g., connecting
tubulation; valves, seals) can be made
extremely small,
c) suitable means (cold traps, baffles) are
provided to prevent gases or vapors or
their reaction products that have origi-
nated from the pumps used from
reaching the vacuum vessel (no
backstreaming).
To fulfill these conditions, the individual
components used in UHV apparatus must
be bakeable and extremely leaktight.
Stainless steel is the preferred material for
UHV components.
The construction, start-up, and operation
of an UHV system also demands special
care, cleanliness, and, above all, time. The
assembly must be appropriate; that is, the
individual components must not be in the
least damaged (i.e. by scratches on
precision-worked sealing surfaces). Fun-
damentally, every newly-assembled UHV
apparatus must be tested for leaks with a
helium leak detector before it is operated.
Especially important here is the testing of
demountable joints (flange connections),
glass seals, and welded or brazed joints.
After testing, the UHV apparatus must be
baked out. This is necessary for glass as
well as for metal apparatus. The bake-out
extends not only over the vacuum vessel,
but frequently also to the attached parts,
particularly the vacuum gauges. The
individual stages of the bake-out, which
can last many hours for a larger system,
and the bake-out temperature are arranged
according to the kind of plant and the
ultimate pressure required. If, after the
apparatus has been cooled and the other
necessary measures undertaken (e.g.,
cooling down cold traps or baffles), the
ultimate pressure is apparently not
obtained, a repeated leak test with a
helium leak detector is recommended.
Details on the components, sealing
methods and vacuum gauges are provided
in our catalog.
D00
D00 E 19.06.2001 21:38 Uhr Seite 61
Vacuum Generation
2.3. Evacuation of a
vacuum chamber
and determination
of pump sizes
Basically, two independent questions
arise concerning the size of a vacuum
system:
1. What effective pumping speed must the
pump arrangement maintain to reduce
the pressure in a given vessel over a
given time to a desired value?
pump arrangement reach during a
vacuum process so that gases and
vapors released into the vessel can be
quickly pumped away while a given
pressure (the operating pressure) in the
vessel, is maintained and not
exceeded?
During the pumping-out procedure of
certain processes (e.g., drying and hea-
ting), vapors are produced that were not
originally present in the vacuum chamber,
so that a third question arises:
pump arrangement reach so that the
process can be completed within a
certain time?
The effective pumping speed of a pump
arrangement is understood as the actual
pumping speed of the entire pump ar-
rangement that prevails at the vessel.
The nominal pumping speed of the pump
can then be determined from the effective
pumping speed if the flow resistance
(conductances) of the baffles, cold traps,
filters, valves, and tubulations installed
between the pump and the vessel are
known (see Sections 1.5.2 to 1.5.4). In the
determination of the required nominal
pumping speed it is further assumed that
the vacuum system is leaktight; therefore,
the leak rate must be so small that gases
flowing in from outside are immediately
removed by the connected pump arrange-
ment and the pressure in the vessel does
not alter (for further details, see Section
5). The questions listed above under 1., 2.
and 3. are characteristic for the three most
essential exercises of vacuum technology
1. Evacuation of the vessel to reach a
specified pressure.
2. Pumping of continuously evolving
quantities of gas and vapor at a certain
pressure.
3. Pumping of the gases and vapors
produced during a process by variation
of temperature and pressure.
Initial evacuation of a vacuum chamber is
influenced in the medium-, high-, and
ultrahigh vacuum regions by continually
evolving quantities of gas, because in
these regions the escape of gases and
vapors from the walls of the vessel is so
significant that they alone determine the
dimensions and layout of the vacuum
system.
2.3.1 Evacuation of a vacuum
chamber (without
additional sources of gas
or vapor)
Because of the factors described above, an
assessment of the pump-down time must
be basically different for the evacuation of
a container in the rough vacuum region
from evacuation in the medium- and high
vacuum regions.
2.3.1.1 Evacuation of a chamber in the
rough vacuum region
In this case the required effective pumping
speed S
eff
, of a vacuum pump assembly is
dependent only on the required pressure
p, the volume V of the container, and the
pump-down time t.
With constant pumping speed S
eff
and
assuming that the ultimate pressure p
end
attainable with the pump arrangement is
such that p
end
<< p, the decrease with time
of the pressure p(t) in a chamber is given
by the equation:
(2.32)
Beginning at 1013 mbar at time t = 0, the
effective pumping speed is calculated
depending on the pump-down time t from
equation (2.32) as follows:
(2.33a)
(2.33b)
n
p S
eff
V
t
1013
=
dp
p
p
S
eff
V
t
1013
=
=
dp
dt
S
eff
V
p
(2.34)
Introducing the dimensionless factor
(2.34a)
into equation (2.34), the relationship
between the effective pumping speed S
eff
,
and the pump-down time t is given by
(2.35)
The ratio V/S
eff
is generally designated as
a time constant . Thus the pump-down
time of a vacuum chamber from
atmospheric pressure to a pressure p is
given by:
t = s (2.36)
with
and
The dependence of the factor from the
desired pressure is shown in Fig. 2.75. It
should be noted that the pumping speed of
single-stage rotary vane and rotary piston
pumps decreases below 10 mbar with gas
ballast and below 1 mbar without gas
ballast. This fundamental behavior is
different for pumps of various sizes and
n
p
1013
=
V
S
eff
S
eff
V
t
=
= n
p p
1013
2 3
1013
. log
S
eff
V
t
n
p
V
t p
= =
1013
2 3
1013
. log
Fig. 2.75 Dependency of the dimensionless factor s for
calculation of pumpdown time t according to
equation 2.36. The broken line applies to
single-stage pumps where the pumping
speed decreases below 10 mbar
Dimensionless factor
P
r
e
s
s
u
r
e

D00 E 19.06.2001 21:38 Uhr Seite 62

Vacuum Generation
types but should not be ignored in the
determination of the dependence of the
pump-down time on pump size. It must be
pointed out that the equations (2.32 to
2.36) as well Fig. 2.75 only apply when the
ultimate pressure attained with the pump
used is by several orders of magnitude
lower than the desired pressure.
Example: A vacuum chamber having a
volume of 500 l shall be pumped down to
1 mbar within 10 minutes. What effective
pumping speed is required?
500 l = 0.5 m
3
; 10 min = 1/6 h
According to equation (2.34) it follows
that:
For the example given above one reads off
the value of 7 from the straight line in Fig.
2.75. However, from the broken line a
value of 8 is read off. According to
equation (2.35) the following is obtained:
under consideration of the fact that the
pumping speed reduces below 10 mbar.
The required effective pumping speed thus
amounts to about 24 m
3
/h.
high vacuum region
It is considerably more difficult to give
general formulas for use in the high vacu-
um region. Since the pumping time to
reach a given high vacuum pressure
depends essentially on the gas evolution
from the chambers inner surfaces, the
condition and pre-treatment of these
surfaces are of great significance in
vacuum technology. Under no circums-
tances should the material used exhibit
porous regions or particularly with
regard to bake-out contain cavities; the
inner surfaces must be as smooth as
possible (true surface = geometric sur-
face) and thoroughly cleaned (and
degreased). Gas evolution varies greatly
with the choice of material and the surface
S
eff
m h or = =
0 5
1
6
7 21
3
.
/
S
eff
m h = =
0 5
1
6
8 24
3
.
/
S
eff
m h
=
= =
0 5
2 3
1013
1
3 2 3 3 01 20 8
3
1 6
.
. log
. . . /
/
condition. Useful data are collected in
Table X (Section 9). The gas evolution can
be determined experimentally only from
case to case by the pressure-rise method:
the system is evacuated as thoroughly as
possible, and finally the pump and the
chamber are isolated by a valve. Now the
time is measured for the pressure within
the chamber (volume V) to rise by a
certain amount, for example, a power of
10. The gas quantity Q that arises per unit
time is calculated from:
(2.37)
(p = measured pressure rise )
The gas quantity Q consists of the sum of
all the gas evolution and all leaks possibly
present. Whether it is from gas evolution
or leakage may be determined by the
following method:
The gas quantity arising from gas evolu-
tion must become smaller with time, the
quantity of gas entering the system from
leakage remains constant with time.
Experimentally, this distinction is not
always easily made, since it often takes a
considerable length of time with pure
gas evolution before the measured pres-
sure-time curve approaches a constant (or
almost a constant) final value; thus the
beginning of this curve follows a straight
line for long times and so simulates
leakage (see Section 5, Leaks and Leak
Detection).
If the gas evolution Q and the required
pressure pend are known, it is easy to de-
termine the necessary effective pumping
speed:
(2.38)
Example: A vacuum chamber of 500 l may
have a total surface area (including all
systems) of about 5 m
2
. A steady gas
evolution of 2 10
-4
mbar l/s is assumed
per m
2
of surface area. This is a level
which is to be expected when valves or
rotary feedthroughs, for example are
connected to the vacuum chamber. In
order to maintain in the system a pressure
of 1 10
-5
mbar, the pump must have a
pumping speed of
S
eff
mbar s
mbar
s =

5 2 10
4
1 10
5
100
/
/
S
eff
Q
p
end
=
Q
p V
t
=

A pumping speed of 100 l/s alone is
required to continuously pump away the
quantity of gas flowing in through the
leaks or evolving from the chamber walls.
Here the evacuation process is similar to
the examples given in Sections 2.3.1.1.
However, in the case of a diffusion pump
the pumping process does not begin at
atmospheric pressure but at the forevacu-
um pressure pV instead. Then equation
(2.34) transforms into:
At a backing pressure of p
V
= 2 10
-3
mbar
compression K is in our example:
In order to attain an ultimate pressure of
1 10
-5
mbar within 5 minutes after
starting to pump with the diffusion pump
an effective pumping speed of
is required. This is much less compared to
the effective pumping speed needed to
maintain the ultimate pressure. Pump-
down time and ultimate vacuum in the
high vacuum and ultrahigh vacuum ranges
depends mostly on the gas evolution rate
and the leak rates. The underlying mathe-
matical rules can not be covered here. For
these please refer to books specializing on
that topic.
medium vacuum region
In the rough vacuum region, the volume of
the vessel is decisive for the time involved
in the pumping process. In the high and
ultrahigh vacuum regions, however, the
gas evolution from the walls plays a
significant role. In the medium vacuum
region, the pumping process is influenced
by both quantities. Moreover, in the
medium vacuum region, particularly with
rotary pumps, the ultimate pressure pend
attainable is no longer negligible. If the
quantity of gas entering the chamber is
known to be at a rate Q (in millibars liter
per second) from gas evolution from the
walls and leakage, the differential equation
(2.32) for the pumping process becomes
S
eff
s
=

500
5 60
2 3 200 9 . log

K =

=
2 10
3
1 10
5
200
S
eff
V
t p
V
t
K
p
V
= = n n
D00
D00 E 19.06.2001 21:38 Uhr Seite 63
Vacuum Generation
(2.39)
Integration of this equation leads to
(2.40)
where
p
0
is the pressure at the beginning of the
pumping process
p is the desired pressure
In contrast to equation 2.33b this equation
does not permit a definite solution for S
eff
,
therefore, the effective pumping speed for
a known gas evolution cannot be deter-
mined from the time pressure curve
without further information.
In practice, therefore, the following
method will determine a pump with
sufficiently high pumping speed:
a) The pumping speed is calculated from
equation 2.34 as a result of the volume
of the chamber without gas evolution
and the desired pump-down time.
b The quotient of the gas evolution rate
and this pumping speed is found. This
quotient must be smaller than the
required pressure; for safety, it must be
about ten times lower. If this condition
is not fulfilled, a pump with corres-
pondingly higher pumping speed must
be chosen.
2.3.2 Determination of a
suitable backing pump
The gas or vapor quantity transported
through a high vacuum pump must also be
handled by the backing pump. Moreover,
in the operation of the high vacuum pump
(diffusion pump, turbomolecular pump),
the maximum permissible backing
pressure must never, even for a short time,
be exceeded. If Q is the effective quantity
of gas or vapor, which is pumped by the
high vacuum pump with an effective
pumping speed S
eff
at an inlet pressure p
A
,
this gas quantity must certainly be trans-
ported by the backing pump at a pumping
speed of S
V
at the backing pressure p
V
. For
the effective throughput Q, the continuity
equation applies:
Q = p
A
S
eff
= p
v
S
V
(2.41)
t
V
S
eff
p
o
p
end
Q S
eff
p p
end
Q S
eff
=
_
,
_
,
n
/
/
dp
dt
S
eff
p p
end
Q
V
=

_
,
The required pumping speed of the
backing pump is calculated from:
(2.41a)
Example: In the case of a diffusion pump
having a pumping speed of 400 l/s the
effective pumping speed is 50 % of the
value stated in the catalog when using a
shell baffle. The max. permissible backing
pressure is 2 10
-1
mbar. The pumping
speed required as a minimum for the
backing pump depends on the intake
pressure p
A
according to equation 2.41a.
At an intake pressure of p
A
= 1 10
-2
mbar
the pumping speed for the high vacuum
pump as stated in the catalog is about 100
l/s, subsequently 50 % of this is 50 l/s.
Therefore the pumping speed of the
backing pump must amount to at least
At an intake pressure of p
A
= 1 10
-3
mbar
the pump has already reached its nominal
pumping speed of 400 l/s; the effective
pumping speed is now S
eff
= 200 l/s; thus
the required pumping speed for the
backing pump amounts to
If the high vacuum pump is to be used for
pumping of vapors between 10
-3
and
10
-2
mbar, then a backing pump offering a
nominal pumping speed of 12 m
3
/h must
be used, which in any case must have a
3
/h at a pressure of
2 10
-1
mbar. If no vapors are to be
pumped, a single-stage rotary vane pump
operated without gas ballast will do in
most cases. If (even slight) components of
vapor are also to be pumped, one should
in any case use a two-stage gas ballast
pump as the backing pump which offers
also with gas ballast the required
pumping speed at 2 10
-1
mbar.
If the high vacuum pump is only to be
used at intake pressures below 10
-3
mbar,
a smaller backing pump will do; in the case
of the example given this will be a pump
offering a pumping speed of 6 m
3
/h. If the
continuous intake pressures are even
lower, below 10
-4
mbar, for example, the
required pumping speed for the backing
pump can be calculated from equation
2.41a as:
S
V
s m h =

= =
1 10
3
2 10
1
200 1 3 6
3
/ . /
h S
V
s m =

=
1 10
2
2 10
1
50 2 5 9
3
. / /
S
V
p
A
p
V
S
eff
=
Theoretically in this case a smaller backing
pump having a pumping speed of about 1
m
3
/h could be used. But in practice a
larger backing pump should be installed
because, especially when starting up a
vacuum system, large amounts of gas may
occur for brief periods. Operation of the
high vacuum pump is endangered if the
quantities of gas can not be pumped away
immediately by the backing pump. If one
works permanently at very low inlet pres-
sures, the installation of a ballast volume
(backing-line vessel or surge vessel)
between the high vacuum pump and the
backing pump is recommended. The
backing pump then should be operated for
short times only. The maximum admis-
sible backing pressure, however, must
never be exceeded.
The size of the ballast volume depends on
the total quantity of gas to be pumped per
unit of time. If this rate is very low, the rule
of thumb indicates that 0.5 l of ballast
volume allows 1 min of pumping time with
the backing pump isolated.
For finding the most adequate size of
backing pump, a graphical method may be
used in many cases. In this case the
starting point is the pumping speed cha-
racteristic of the pumps according to
equation 2.41.
The pumping speed characteristic of a
pump is easily derived from the measured
pumping speed (volume flow rate)
characteristic of the pump as shown for a
6000 l/s diffusion pump (see curve S in
Fig. 2.76). To arrive at the throughput
characteristic (curve Q in Fig. 2.76), one
must multiply each ordinate value of S by
its corresponding pA value and plotted
against this value. If it is assumed that the
inlet pressure of the diffusion pump does
not exceed 10
-2
mbar, the maximum
throughput is 9.5 mbar l/s
Hence, the size of the backing pump must
be such that this throughput can be
handled by the pump at an intake pressure
(of the backing pump) that is equal to or
preferably lower than the maximum
permissible backing pressure of the dif-
fusion pump; that is, 4 10
-1
mbar for the
6000 l/s diffusion pump.
After accounting for the pumping speed
S
V
s m h =

= =
1 10
4
2 10
1
200 0 1 0 36
3
. / . /
D00 E 19.06.2001 21:38 Uhr Seite 64
Vacuum Generation
characteristics of commercially available
two-stage rotary plunger pumps, the
throughput characteristic for each pump is
calculated in a manner similar to that used
to find the Q curve for the diffusion pump
in Fig. 2.76 a. The result is the group of Q
curves numbered 1 4 in Fig. 2.76 b,
whereby four 2-stage rotary-plunger
pumps were considered, whose nominal
speeds were 200, 100, 50, and 25 m
3
/h,
respectively. The critical backing pressure
of the 6000 l/s diffusion pump is marked as
V.B. (p = 4 10
-1
mbar). Now the maximum
throughput Q = 9.5 mbar l/s is shown as
horizontal line a. This line intersects the
four throughput curves. Counting from
right to left, the first point of intersection
that corresponds to an intake pressure
below the critical backing pressure of
4 10
-1
mbar is made with throughput
characteristic 2. This corresponds to the
two-stage rotary plunger pump with a
3
/h.
Therefore, this pump is the correct backing
pump for the 6000 l/s diffusion pump
under the preceding assumption.
However, if the pumping process is such
that the maximum throughput of 9.5
mbar l/s is unlikely, a smaller backing
pump can, of course, be used. This is self-
explanatory, for example, from line b in
Fig. 2.76 b, which corresponds to a
maximum throughput of only 2 mbar l/s.
In this case a 25 m
3
/h two-stage rotary-
plunger pump would be sufficient.
2.3.3 Determination of pump-
down time from
nomograms
In practice, for instance, when estimating
the cost of a planned vacuum plant, cal-
culation of the pump-down time from the
effective pumping speed S
eff
, the required
pressure p, and the chamber volume V by
formulas presented would be too trouble-
some and time-consuming. Nomograms
are very helpful here. By using the nomo-
gram in Fig. 9.7 in Section 9, one can
quickly estimate the pump-down time for
vacuum plants evacuated with rotary
pumps, if the pumping speed of the pump
concerned is fairly constant through the
pressure region involved. By studying the
examples presented, one can easily under-
stand the application of the nomogram.
The pump-down times of rotary vane and
rotary piston pumps, insofar as the pum-
ping speed of the pump concerned is con-
stant down to the required pressure, can
be determined by reference to example 1.
In general, Roots pumps do not have
constant pumping speeds in the working
region involved. For the evaluation of the
pump-down time, it usually suffices to
assume the mean pumping speed.
Examples 2 and 3 of the nomogram show,
in this context, that for Roots pumps, the
compression ratio K refers not to the
atmospheric pressure (1013 mbar), but to
the pressure at which the Roots pump is
switched on.
In the medium vacuum region, the gas
evolution or the leak rate becomes
significantly evident. From the nomogram
9.10 in Section 9, the corresponding
calculations of the pump-down time in this
vacuum region can be approximated.
In many applications it is expedient to
relate the attainable pressures at any given
time to the pump-down time. This is easily
possible with reference to the nomogram
9.7 in Section 9.
As a first example the pump-down cha-
racteristic that is, the relationship pres-
sure p (denoted as desired pressure p
end
)
versus pumping time tp is derived from
the nomogram for evacuating a vessel of
5 m
3
volume by the single-stage rotary
plunger pump E 250 with an effective
pumping speed of S
eff
= 250 m
3
/h and an
ultimate pressure p
end,p
= 3 10
-1
mbar
when operated with a gas ballast and at
p
end,p
= 3 10
-2
mbar without a gas ballast.
The time constant = V / S
eff
(see equation
2.36) is the same in both cases and
amounts as per nomogram 9.7 to about
70 s (column 3). For any given value of
p
end
> p
end,p
the straight line connecting
the 70 s point on column 3 with the
(p
end
p
end,p
) value on the right-hand
scale of column 5 gives the corresponding
t
p
value. The results of this procedure are
shown as curves a and b in Fig. 2.77.
D00
Fig. 2.76 Diagram for graphically determining a suitable backing pump
Intake pressure p
a
[mbar]
T
h
r
o
u
g
h
p
u
t

Q

[
m
b
a
r
1
]
Q

[
m
b
a
r
1
]
P
u
m
p
i
n
g

s
p
e
e
d

S

[
l
1
]
a) Pumping speed characteristic of a 6000 l/s diffusion pump b) Series of throughput curves for two-stage rotary
plunger pumps (V.B. = critical forevacuum pressure)
D00 E 19.06.2001 21:38 Uhr Seite 65
Vacuum Generation
It is somewhat more tedious to determine
the (p
end,tp
) relationship for a combination
of pumps. The second example discussed
in the following deals with evacuating a
vessel of 5 m
3
volume by the pump
combination Roots pump WA 1001 and
the backing pump E 250 (as in the
preceding example). Pumping starts with
the E 250 pump operated without gas
ballast alone, until the Roots pump is
switched on at the pressure of 10 mbar. As
the pumping speed characteristic of the
combination WA 1001/ E 250 in contrast
to the characteristic of the E 250 is no
longer a horizontal straight line over the
best part of the pressure range (compare
this to the corresponding course of the
characteristic for the combination WA
2001 / E 250 in Fig. 2.19), one introduces,
as an approximation, average values of
S
eff
, related to defined pressure ranges. In
the case of the WA 1001/ E 250
combination the following average figures
apply:
S
eff
= 800 m
3
/h in the range 10 1 mbar,
S
eff
= 900 m
3
/h in the range 1 mbar
to 5 10
-2
mbar,
S
eff
= 500 m
3
/h in the range 5 10
-2
to
5 10
-3
mbar
The ultimate pressure of the combination
WA 1001 / E 250 is: P
end,p
= 3 10
-3
mbar.
From these figures the corresponding time
constants in the nomogram can be
determined; from there, the pump-down
time tp can be found by calculating the
pressure reduction R on the left side of
column 5. The result is curve c in Fig. 2.77.
Computer aided calculations at LEYBOLD
Of course calculations for our industrial
systems are performed by computer
programs. These require high per-
formance computers and are thus usually
not available for simple initial calculations.
2.3.4 Evacuation of a chamber
where gases and vapors
are evolved
The preceding observations about the
pump-down time are significantly altered if
vapors and gases arise during the
evacuation process. With bake-out pro-
cesses particularly, large quantities of
vapor can arise when the surfaces of the
chamber are cleared of contamination. The
resulting necessary pump-down time
depends on very different parameters.
Increased heating of the chamber walls is
accompanied by increased desorption of
gases and vapors from the walls. However,
because the higher temperatures result in
an accelerated escape of gases and vapors
from the walls, the rate at which they can
be removed from the chamber is also in-
creased.
The magnitude of the allowable tempera-
ture for the bake-out process in question
will, indeed, be determined essentially by
the material in the chamber. Precise pump-
down times can then be estimated by
calculation only if the quantity of the
evolving and pumped vapors is known.
However, since this is seldom the case
except with drying processes, a quantita-
tive consideration of this question is
abandoned within the scope of this
publication.
2.3.5 Selection of pumps for
drying processes
Fundamentally, we must distinguish bet-
ween short-term drying and drying pro-
cesses that can require several hours or
even days. Independently of the duration
of drying, all drying processes proceed
approximately as in Section 2.24
As an example of an application, the drying
of salt (short-term drying) is described,
this being an already well-proven drying
process.
Drying of salt
First, 400 kg of finely divided salt with a
water content of about 8 % by mass is to
be dried in the shortest possible time
(about 1 h) until the water content is less
than 1 % by mass. The expected water
evolution amounts to about 28 kg. The salt
in the chamber is continuously agitated
during the drying process and heated to
about 80 C. The vacuum system is
schematically drawn in Fig. 2.78.
During the first quarter of drying time far
more than half the quantity of water vapor
is evolved. Then the condenser is the
actual main pump. Because of the high
water vapor temperature and, at the
beginning of the drying, the very high
water vapor pressure, the condensation
efficiency of the condenser is significantly
increased. In Fig. 2.78 it is understood that
two parallel condensers each of 1 m
2
con-
densation surface can together condense
about 15 l of water at an inlet pressure of
100 mbar in 15 min. However, during this
initial process, it must be ensured that the
water vapor pressure at the inlet port of
the rotary piston pump does not exceed 60
mbar (see Section 2.15 for further details).
Since the backing pump has only to pump
away the small part of the noncondensable
gases at this stage, a single-stage rotary
piston pump TRIVAC S 65 B will suffice.
With increasing process time, the water
vapor evolution decreases, as does the
Fig. 2.77 Pumpdown time, t
p
, of a 5 m
3
vessel using a
rotary plunger pump E 250 having a nominal
3
/h with (a) and
without (b) gas ballast, as well as
Roots/rotary plunger pump combination WA
1001 / E250 for a cut-in pressure of 10 mbar
for the WA 1001 (e)
D00 E 19.06.2001 21:38 Uhr Seite 66
Vacuum Generation
water vapor pressure in the condenser.
After the water pressure in the chamber
falls below 27 mbar, the Roots pump (say,
a Roots pump RUVAC WA 501) is
switched in. Thereby the water vapor is
pumped more rapidly out of the chamber,
the pressure increases in the condensers,
and their condensation efficiency again
increases. The condensers are isolated by
a valve when their water vapor reaches its
saturation vapor pressure. At this point,
there is a water vapor pressure in the
chamber of only about 4 mbar, and
pumping is accomplished by the Roots
pump with a gas ballast backing pump
until the water vapor pressure reaches
about 0.65 mbar. From experience it can
be assumed that the salt has now reached
the desired degree of dryness.
Drying of paper
If the pumps are to be of the correct size
for a longer process run, it is expedient to
break down the process run into charac-
teristic sections. As an example, paper
drying is explained in the following where
the paper has an initial moisture content of
8 %, and the vessel has the volume V.
1. Evacuation
The backing pump must be suitably rated
with regard to the volume of the vessel and
the desired pump-down time. This pump-
down time is arranged according to the
desired process duration: if the process is
to be finished after 12 15 h, the pump-
down time should not last longer than 1 h.
The size of the backing pump may be
easily calculated according to Section
2.3.1.
2. Predrying
During predrying depending on the
pressure region in which the work is
carried out about 75 % of the moisture is
drawn off. This predrying should occupy
the first third of the drying time. The rate at
which predrying proceeds depends almost
exclusively on the sufficiency of the heat
supply. For predrying 1 ton of paper in 5 h,
60 kg of water must be evaporated; that is,
an energy expenditure of about 40 kWh is
needed to evaporate water. Since the paper
must be heated to a temperature of about
120 C at the same time, an average of
about 20 kW must be provided. The mean
vapor evolution per hour amounts to
12 kg. Therefore, a condenser with a capa-
city of 15 kg/h should be sufficient. If the
paper is sufficiently preheated (perhaps by
air-circulation drying) before evacuation,
in the first hour of drying, double vapor
evolution must be anticipated.
3. Main drying
If, in the second stage, the pressure in a
further 5 h is to fall from 20 to about
5.3 mbar and 75 % of the total moisture
(i.e., 19 % of the total moisture of 15 kg)
is to be drawn off, the pump must, accor-
ding to equations (2.37) and (2.38), have a
pumping speed of
According to equation 1.7, 15 kg of water
vapor corresponds at 15 C to a quantity of
water vapor of
Hence the Roots pump RUVAC WA 1001
would be the suitable pump. The
permissible remaining moisture in the
product determines the attainable ultimate
S m h
eff
=
=
20000
5 5 3
750
3
.
/
V p
m R T
M
mbar m subsequently
=

=

15 83 14 288
18
20000
3
.
S
eff
V p
t p
=

pressure. The relationship between the

ultimate pressure and the remaining
moisture is fixed for every product but dif-
ferent from product to product. LEYBOLD
has many years of experience to its record
regarding applications in this area.
Assume that a 0.1 % residual moisture
content is required, for which the neces-
sary ultimate pressure is 6 10
-2
mbar.
During the last 5 h the remaining 6 % of
the moisture content, or 5 kg of water, is
removed. At a mean pressure of about
0.65 mbar, 2000 m
3
/h of vapor is evolved.
Two possibilities are offered:
a) One continues working with the above-
mentioned Roots pump WA 1001. The
ultimate total pressure settles at a value
according to the water vapor quantity
evolving. One waits until a pressure of
about 6.5 10
-2
mbar is reached, which
naturally takes a longer time.
b From the beginning, a somewhat larger
Roots pump is chosen (e.g., the
RUVAC WA 2001 with a pumping speed
of 2000 m
3
/h is suitable). For larger
quantities of paper (5000 kg, for
example) such a pumping system will
be suitable which at a pumping speed
for water vapor of up to 20,000 m
3
/h
automatically lowers the pressure from
27 to 10
-2
mbar. The entire time need
for drying is significantly reduced when
using such pumps.
2.3.6 Flanges and their seals
In general, demountable joints in metallic
vacuum components, pumps, valves,
tubulations, and so on are provided with
flanges. Vacuum components for rough,
medium, and high vacuum from LEYBOLD
are equipped with the following stan-
dardized flange systems:
Small flanges (KF) (quick-action con-
nections to DIN 28 403) of nominal
widths 10, 16, 20, 25, 32, 40 and 50
mm. The values 10, 16, 25 and 40 are
preferred widths according to the
PNEUROP recommendations and the
ISO recommendations of the technical
committee ISO/TC 112 (see also Sec-
tion 11). For a complete connection of
two identical flanges one clamping ring
and one centering ring are required.
Clamp flanges (ISO-K) of nominal
widths 65, 100, 160, 250, 320, 400,
D00
Fig. 2.78 Vacuum diagram for drying of salt. Pump
combination consisting of Roots pump, con-
denser and rotary plunger pump for stepwise
switching of the pumping process (see text)
1 Vacuum chamber with salt filling
2 RUVAC WA 501
3 Condensers
4 Throttle valve
5 Rotary plunger pump
D00 E 19.06.2001 21:38 Uhr Seite 67
Vacuum Generation
500 and 630 mm. Also, these flanges
correspond to the nominal widths and
construction of the PNEUROP and
ISO/TC 112 recommendations. Clamp
flanges are joined together by clamps
or collar rings. Centering rings or gas-
kets are needed for sealing.
Bolted flanges (ISO-F) for the same
nominal widths as above (according to
PNEUROP and ISO/TC 112). In special
cases bolted flanges having a smaller
nominal width are used. Clamp flanges
and bolted flanges are in accordance
with DIN 28 404.
The nominal width is approximately equal
to the free inner diameter of the flange in
millimeters; greater deviations are excep-
tions, so the clamp flange DN 63 has an
inner diameter of 70 mm. See also Table XI
in Section 9).
High vacuum components are made of
aluminum or stainless steel. Stainless
steel is slightly more expensive but offers
a variety of advantages: lower degassing
rate, corrosion resistant, can be degassed
at temperatures up to 200 C, metal seals
are possible and stainless steel is much
more resistant to scratching compared to
aluminum.
Ultrahigh vacuum components are made
of stainless steel and have CF flanges
bakeable to a high temperature. These
components, including the flanges, are
manufactured in a series production,
starting with a nominal width of 16 up to
250 mm. CF flanges are available as fixed
flanges or also with rotatable collar
flanges. They may be linked with CONFLAT
flanges from almost all manufacturers.
Copper gaskets are used for sealing
purposes.
Basically, the flanges should not be
smaller than the connecting tubes and the
components that are joined to them. When
no aggressive gases and vapors are
pumped and the vacuum system is not
exposed to a temperature above 80 C,
sealing with NBR (Perbunan) or CR (Neo-
prene) flange O-rings is satisfactory for
work in the rough-, medium-, and high
vacuum regions. This is often the case
when testing the operation of vacuum
systems before they are finally assembled.
All stainless steel flanges may be degassed
at temperatures up to 200 C without
impairment. However, then Perbunan
sealing material is not suitable as a flange
sealant. Rather, VITILAN
(a special FPM)
sealing rings and also aluminum seals,
which allow heating processes up to
150 C and 200 C respectively, should be
used. After such degassing, pressures
down to 10
-8
mbar, i.e. down to the UHV
range, can be attained in vacuum systems.
Generating pressures below 10
-8
mbar
requires higher bake-out temperatures. As
explained above (see Section 2.2.6) work
in the UHV range requires a basically
different approach and the use of CF
flanges fitted with metallic sealing rings.
2.3.7 Choice of suitable valves
Vacuum technology puts great demands
on the functioning and reliability of the
valves, which are often needed in large
numbers in a plant. The demands are
fulfilled only if correct shut-off devices are
installed for each application, depending
on the method of construction, method of
operation, and size. Moreover, in the
construction and operation of vacuum
plants, factors such as the flow
conductance and leak-tightness of valves
are of great importance.
Valves are constructed so that they will not
throttle pumping speed. Hence, when
opened fully, their conductance in the
rough and medium vacuum regions equals
that of corresponding tube components.
For example, the conductance of a right-
angle valve will equal the conductance of a
bent tube of the same nominal bore and
angle. Similarly, the conductance of the
valve for molecular flow (i. e., in the high
and ultrahigh vacuum regions), is so high
that no significant throttling occurs. Actual
values for the conductance of various
components are given in the catalog.
To meet stringent leak-tightness demands,
high-quality vacuum valves are designed
so that gas molecules adhering to the
surface of the valve shaft are not
transferred from the outer atmosphere into
the vacuum during operation. Such valves
are therefore equipped with metal bellows
for isolating the valve shaft from the
atmosphere, or alternatively, they are fully
encapsulated, that is, only static seals exist
between atmosphere and vacuum. This
group is comprised of all medium and high
vacuum valves from LEYBOLD that are
operated either manually or electro-
pneumatically (Fig. 2.80) and (Fig. 2.79).
The leak rate of these valves is less than
10
-9
mbar l/s.
Valves sealed with oil or grease can be
Fig. 2.79 Right angle vacuum valve with
solenoid actuator
1 Casing
2 Valve disk
Fig. 2.80 Right angle vacuum valve with
electropneumatic actuator
4 Compressed air supply
5 Piston
3 Compression spring
4 Solenoid coil
1 Casing
2 Valve disk
3 Bellows
D00 E 19.06.2001 21:38 Uhr Seite 68
Vacuum Generation
used for highly stringent demands. Their
leakage rate is also about 10
-9
mbar l/s.
However, a special case is the pendulum-
type gate valve. Despite its grease-covered
seal, the leak rate between vacuum and
external atmosphere is virtually the same
as for bellows-sealed valves because when
the valve is in operation the shaft carries
out only a rotary motion so that no gas
molecules are transferred into the vacuum.
Pendulum-type gate valves are not manu-
factured by LEYBOLD.
For working pressures down to 10
-7
mbar,
valves of standard design suffice because
their seals and the housing materials are
such that permeation and outgassing are
insignificant to the actual process. If
pressures down to 10
-9
mbar are required,
baking up to 200 C is usually necessary,
which requires heat resistant sealing
materials (e.g., VITILAN
) and materials
of high mechanical strength, with prepared
(inner) surfaces and a low outgassing rate.
Such valves are usually made of stainless
steel. Flange connections are sealed with
aluminum gaskets, so permeation pro-
blems of elastomer seals are avoided. In
the UHV range these issues are of special
significance so that mainly metallic seals
must be used. The gas molecules bonded
to the surface of the materials have, at
pressures below 10
-9
mbar, a very great
influence. They can only be pumped away
within a reasonable period of time by
simultaneous degassing. Degassing
temperatures up to 500 C required in UHV
systems, pose special requirements on the
sealing materials and the entire sealing
geometry. Gaskets made of gold or copper
must be used.
The various applications require valves
with different drives, that is, valves that are
manually operated, electropneumatically-
or magnetically-operated, and motor
driven, such as variable-leak valves. The
variety is even more enhanced by the
various housing designs. In addition to the
various materials used, right-angle and
straight-through valves are required.
Depending on their nominal width and
intended application, flanges fitted to
valves may be small (KF), clamp (ISO-K),
bolted (ISO-F), or UHV (CF).
In addition to the vacuum valves, which
perform solely an isolation function (fully
open fully closed position), special
valves are needed for special functions.
Typical are variable leak valves, which
cover the leakage range from 10
-10
cm
3
/s
(NTP) up to 1.6 10
3
cm
3
/s (NTP). These
valves are usually motor driven and
suitable for remote control and when they
are connected to a pressure gauge, the
process pressures can be set and
maintained. Other special valves fulfill
safety functions, such as rapid, automatic
cut-off of diffusion pumps or vacuum
systems in the event of a power failure. For
example, SECUVAC valves belong to this
group. In the event of a power failure, they
cut off the vacuum system from the
pumping system and vent the forevacuum
system. The vacuum system is enabled
only after a certain minimum pressure
(about 200 mbar) has been attained once
the power has been restored.
When aggressive gases or vapors have to
be pumped, valves made of stainless steel
and sealed with VITILAN
sealant are
usually used. For nuclear technology,
valves have been developed that are sealed
with special elastomer or metal gaskets.
We will be pleased to provide further
design information for your area of
application upon request.
2.3.8 Gas locks and seal-off
fittings
In many cases it is desirable not only to be
able to seal off gas-filled or evacuated ves-
sels, but also to be in a position to check
the pressure or the vacuum in these
vessels at some later time and to post-eva-
cuate or supplement or exchange the gas
filling.
This can be done quite easily with a seal-off
fitting from LEYBOLD which is actuated via
a corresponding gas lock. The small flange
connection of the evacuated or gas-filled
vessel is hermetically sealed off within the
tube by a small closure piece which forms
the actual valve. The gas lock required for
actuation is removed after evacuation or
filling with gas. Thus one gas lock will do to
actuate any number of seal-off fittings.
Shown in Fig. 2.81 is a sectional view of
such an arrangement. Gas locks and seal-
off fittings are manufactured by LEYBOLD
having a nominal width of DN 16 KF, DN 25
KF and DN 40 KF. They are made of stain-
less steel. The leak rate of the seal-off fit-
tings is less than 1 10
-9
mbar l/s. They
can sustain overpressures up to 2.5 bar,
are temperature resistant up to 150 C and
may be protected against dirt by a standard
blank flange.
Typical application examples are double-
walled vessels with an insulating vacuum,
like Dewar vessels, liquid gas vessels
(tanks) or long distance energy pipelines
and many more. They are also used for
evacuation or post-evacuation of reference
and support vacua in scientific instruments
seal-off fittings with gas locks are often
used. Previously it was necessary to have a
pump permanently connected in order to
post-evacuate as required. Through the
use of gas locks with seal-off fittings a
vacuum-tight seal is provided for the
vessel and the pump is only required from
time to time for checking or post-
evacuation.
D00
DN
a
h
a
h
1
h
2
d
Fig. 2.81 Gas lock with centering ring and seal-off
fitting, sectional view
D00 E 19.06.2001 21:38 Uhr Seite 69
Vacuum Measurement
3. Vacuum
measurement,
monitoring,
control and
regulation
The pressures measured in vacuum tech-
nology today cover a range from 1013
mbar to 10
-12
mbar, i.e. over 15 orders of
magnitude. The enormous dynamics invol-
ved here can be shown through an analo-
gy analysis of vacuum pressure measure-
ment and length measurement, as depic-
ted in Table 3.1.
Measuring instruments designated as
vacuum gauges are used for measurement
in this broad pressure range. Since it is
impossible for physical reasons to build a
vacuum gauge which can carry out quanti-
tative measurements in the entire vacuum
range, a series of vacuum gauges is avai-
lable, each of which has a characteristic
measuring range that usually extends over
several orders of magnitude (see Fig.
9.16a). In order to be able to allocate the
largest possible measuring ranges to the
individual types of vacuum gauges, one
accepts the fact that the measurement
uncertainty rises very rapidly, by up to
100 % in some cases, at the upper and
lower range limits. This interrelationship is
shown in Fig. 3.1 using the example of the
VISCOVAC. Therefore, a distinction must
be made between the measuring range as
stated in the catalogue and the measuring
range for precise measurement. The
measuring ranges of the individual vacu-
um gauges are limited in the upper and
lower range by physical effects.
3.1 Fundamentals of
low-pressure
measurement
Vacuum gauges are devices for measuring
gas pressures below atmospheric pressu-
re (DIN 28 400, Part 3, 1992 issue). In
many cases the pressure indication
depends on the nature of the gas. With
compression vacuum gauges it should be
noted that if vapors are present, condensa-
tion may occur due to the compression, as
a result of which the pressure indication is
falsified. Compression vacuum gauges
measure the sum of the partial pressures
of all gas components that do not conden-
se during the measurement procedure. In
the case of mechanically compressing
pumps, the final partial pressure can be
measured in this way (see 1.1). Another
way of measuring this pressure, is to free-
ze out the condensable components in an
LN
2
cold trap. Exact measurement of par-
tial pressures of certain gases or vapors is
carried out with the aid of partial pressure
measuring instruments which operate on
the mass spectrometer principle (see sec-
tion 4).
Dependence of the pressure indication
on the type of gas
A distinction must be made between the
following vacuum gauges:
1. Instruments that by definition measure
the pressure as the force which acts on
an area, the so-called direct or absolu-
te vacuum gauges. According to the
kinetic theory of gases, this force,
which the particles exert through their
impact on the wall, depends only on the
number of gas molecules per unit volu-
me (number density of molecules n)
and their temperature, but not on their
molar mass. The reading of the measu-
ring instrument is independent of the
type of gas. Such units include liquid-
filled vacuum gauges and mechanical
vacuum gauges.
2. Instruments with indirect pressure
measurement. In this case, the pressu-
re is determined as a function of a pres-
sure-dependent (or more accurately,
density-dependent) property (thermal
conductivity, ionization probability,
electrical conductivity) of the gas.
These properties are dependent on the
molar mass as well as on the pressure.
Analogy analysis
Determination by Absolute
means of pressure Length
empirical world
of human beings 1 bar 1 m
simple measuring
methods > 1 mbar > 1 mm
mechanical
measuring
methods > 10
3
mbar > 1 mm
indirect
methods 10
9
mbar 1/100
atom
extreme indirect 0.18
methods 10
12
mbar electron
Table 3.1
favorable measuring range
(relative measurement uncertainty < 5%)
20
15
10
5
1
10
6
10
5
10
4
10
3
10
2
10
1
1
R
e
l
a
t
i
v
e

m
e
a
s
u
r
e
m
e
n
t

u
n
c
e
r
t
a
i
n
t
y

(
%
)
Pressure (mbar)
Fig. 3.1 Measurement uncertainty distribution over the measuring range: VISCOVAC
D00 E 19.06.2001 21:38 Uhr Seite 70
Vacuum Measurement
The pressure reading of the measuring
instrument depends on the type of gas.
The scales of these pressure measuring
instruments are always based on air or
nitrogen as the test gas. For other gases or
vapors correction factors, usually based
on air or nitrogen, must be given (see
Table 3.2). For precise pressure measure-
ment with indirectly measuring vacuum
gauges that determine the number density
through the application of electrical energy
(indirect pressure measurement), it is
important to know the gas composition. In
practice, the gas composition is known
only as a rough approximation. In many
cases, however, it is sufficient to know
whether light or heavy molecules predomi-
nate in the gas mixture whose pressure is
to be measured (e.g. hydrogen or pump
fluid vapor molecules).
Example: If the pressure of a gas essenti-
ally consisting of pump fluid molecules is
measured with an ionization vacuum
gauge, then the pressure reading (applying
to air or N
2
), as shown in Table 3.2, is too
high by a factor of about 10.
Measurement of pressures in the rough
vacuum range can be carried out relatively
precisely by means of vacuum gauges with
direct pressure measurement. Measure-
ment of lower pressures, on the other
hand, is almost always subject to a num-
ber of fundamental errors that limit the
measuring accuracy right from the start so
that it is not comparable at all to the
degree of accuracy usually achieved with
measuring instruments. In order to mea-
sure pressure in the medium and high
vacuum ranges with a measurement
uncertainty of less than 50 %, the person
conducting the experiment must proceed
with extreme care. Pressure measure-
ments that need to be accurate to a few
percent require great effort and, in general,
the deployment of special measuring
instruments. This applies particularly to all
pressure measurements in the ultrahigh
vacuum range (p < 10
-7
mbar).
To be able to make a meaningful statement
about a pressure indicated by a vacuum
gauge, one first has to take into account at
what location and in what way the measu-
ring system is connected. In all pressure
areas where laminar flows prevail
(1013 > p > 10
-1
mbar), note must be
taken of pressure gradients caused by
pumping. Immediately in front of the
pump (as seen from the vessel), a lower
pressure is created than in the vessel. Even
components having a high conductance
may create such a pressure gradient.
Finally, the conductance of the connecting
line between the vacuum system and the
measuring system must not be too small
because the line will otherwise be evacua-
ted too slowly in the pressure region of
laminar flow so that the indicated pressure
is too high.
The situation is more complicated in the
case of high and ultrahigh vacuum.
According to the specific installation fea-
tures, an excessively high pressure or, in
the case of well-degassed measuring
tubes, an excessively low pressure may be
recorded due to outgassing of the walls of
the vacuum gauge or inadequate degas-
sing of the measuring system. In high and
ultrahigh vacuum, pressure equalization
between the vacuum system and the mea-
suring tubes may take a long time. If pos-
sible, so-called nude gauges are used. The
latter are inserted directly in the vacuum
system, flange-mounted, without a
connecting line or an envelope. Special
consideration must always be given to the
influence of the measuring process itself
on the pressure measurement. For exam-
ple, in ionization vacuum gauges that work
with a hot cathode, gas particles, especial-
ly those of the higher hydrocarbons, are
thermally broken down. This alters the gas
composition. Such effects play a role in
connection with pressure measurement in
the ultrahigh vacuum range. The same
applies to gas clean-up in ionization vacu-
um gauges, in particular Penning gauges
(of the order of 10
-2
to 10
-1
l/s). Contami-
nation of the measuring system, interfe-
ring electrical and magnetic fields, insula-
tion errors and inadmissibly high ambient
temperatures falsify pressure measure-
ment. The consequences of these avoida-
ble errors and the necessary remedies are
indicated in the discussion of the individu-
al measuring systems and in summary
form in section 8.4.
Selection of vacuum gauges
The desired pressure range is not the only
factor considered when selecting a suita-
ble measuring instrument. The operating
conditions under which the gauge works
also play an important role. If measure-
ments are to be carried out under difficult
operating conditions, i.e. if there is a high
risk of contamination, vibrations in the
tubes cannot be ruled out, air bursts can
be expected, etc., then the measuring
instrument must be robust. In industrial
facilities, Bourdon gauges, diaphragm
vacuum gauges, thermal conductivity
vacuum gauges, hot cathode ionization
vacuum gauges and Penning vacuum gau-
ges are used. Some of these measuring
instruments are sensitive to adverse ope-
rating conditions. They should and can
only be used successfully if the above
mentioned sources of errors are excluded
as far as possible and the operating
instructions are followed.
D00
Given the presence Correction factor
of predominantly based on N
2
(type of gas) (nitrogen = 1)
He 6.9
Ne 4.35
Ar 0.83
Kr 0.59
Xe 0.33
Hg 0.303
H
2
2.4
CO 0.92
CO
2
0.69
CH
4
0.8
higher
hydrocarbons 0.1 0.4
Table 3.2 Correction factors
D00 E 19.06.2001 21:38 Uhr Seite 71
Vacuum Measurement
3.2 Vacuum gauges
with pressure
reading that is
independent of the
type of gas
Mechanical vacuum gauges measure the
pressure directly by recording the force
which the particles (molecules and atoms)
in a gas-filled space exert on a surface by
virtue of their thermal velocity.
3.2.1 Bourdon vacuum gauges
The interior of a tube bent into a circular
arc (so-called Bourdon tube) (3) is
connected to the vessel to be evacuated
(Fig. 3.2). Through the effect of the exter-
nal air pressure the end of the tube is
deflected to a greater or lesser extent
during evacuation and the attached pointer
mechanism (4) and (2) is actuated. Since
the pressure reading depends on the exter-
nal atmospheric pressure, it is accurate
only to approximately 10 mbar, provided
that the change in the ambient atmosphe-
ric pressure is not corrected.
3.2.2 Diaphragm vacuum
gauges
3.2.2.1 Capsule vacuum gauges
The best-known design of a diaphragm
vacuum gauge is a barometer with an
aneroid capsule as the measuring system.
It contains a hermetically sealed, evacu-
ated, thin-walled diaphragm capsule made
of a copper-beryllium alloy. As the pressu-
re drops, the capsule diaphragm expands.
This movement is transmitted to a point by
a lever system. The capsule vacuum
gauge, designed according to this princi-
ple, indicates the pressure on a linear
scale, independent of the external atmos-
pheric pressure.
3.2.2.2 DIAVAC diaphragm vacuum
gauge
The most accurate pressure reading possi-
ble is frequently required for levels below
50 mbar. In this case, a different diaphragm
vacuum gauge is more suitable, i.e. the
DIAVAC, whose pressure scale is consider-
ably extended between 1 and 100 mbar.
The section of the interior in which the
lever system (2) of the gauge head is loca-
ted (see Fig. 3.3) is evacuated to a referen-
ce pressure pref of less than 10
-3
mbar.
The closure to the vessel is in the form of a
corrugated diaphragm (4) of special steel.
As long as the vessel is not evacuated, this
diaphragm is pressed firmly against the
wall (1). As evacuation increases, the diffe-
rence between the pressure to be measu-
red px and the reference pressure decrea-
ses. The diaphragm bends only slightly at
first, but then below 100 mbar to a greater
degree. With the DIAVAC the diaphragm
deflection is again transmitted to a pointer
(9). In particular the measuring range bet-
ween 1 and 20 mbar is considerably exten-
ded so that the pressure can be read quite
accurately (to about 0.3 mbar). The sensi-
tivity to vibration of this instrument is
somewhat higher than for the capsule
vacuum gauge.
3.2.2.3 Precision diaphragm vacuum
gauges
A significantly higher measuring accuracy
than that of the capsule vacuum gauge and
the DIAVAC is achieved by the precision
diaphragm vacuum gauge. The design of
these vacuum gauges resembles that of
capsule vacuum gauges. The scale is line-
ar. The obtainable degree of precision is
the maximum possible with present-day
state-of-the-art equipment. These instru-
ments permit measurement of 10
-1
mbar
with a full-scale deflection of 20 mbar. The
greater degree of precision also means a
higher sensitivity to vibration.
Capsule vacuum gauges measure pressu-
re accurately to 10 mbar (due to the linear
scale, they are least accurate at the low
pressure end of the scale). If only pressu-
res below 30 mbar are to be measured, the
DIAVAC is recommended because its rea-
ding (see above) is considerably more
accurate. For extremely precise measuring
accuracy requirements precision dia-
phragm vacuum gauges should be used. If
low pressures have to be measured accu-
rately and for this reason a measuring
range of, for example, up to 20 mbar is
selected, higher pressures can no longer
be measured since these gauges have a
linear scale. All mechanical vacuum gau-
ges are sensitive to vibration to some
extent. Small vibrations, such as those that
arise in the case of direct connection to a
backing pump, are generally not detrimen-
tal.
Fig. 3.2 Cross-section of a Bourdon gauge
1 Connecting tube to
connection flange
2 Pointer
3 Bourdon tube
4 Lever system
Fig. 3.3 Cross-section of DIAVAC DV 1000
diaphragm vacuum gauge
1 Base plate
2 Lever system
3 Connecting flange
4 Diaphragm
5 Reference
pressure pref,
6 Pinch-off end
7 Mirror sheet
8 Plexiglass sheet
9 Pointer
10 Glass bett
11 Mounting plate
12 Housing
D00 E 19.06.2001 21:38 Uhr Seite 72
Vacuum Measurement
3.2.2.4 Capacitance diaphragm
gauges
Deflection of a diaphragm can also be elec-
trically measured as strain or as a chan-
ge in capacitance. In the past, four strain
gauges, which change their resistance
when the diaphragm is deflected, i.e.
under tensile load, were mounted on a
metallic diaphragm in a bridge circuit. At
LEYBOLD such instruments have been
given a special designation, i.e.
MEMBRANOVAC. Later, silicon dia-
phragms that contained four such strain
resistances directly on their surface were
used. The electrical arrangement again
consisted of a bridge circuit, and a con-
stant current was fed in at two opposite
corner points while a linear voltage signal
proportional to the pressure was picked up
at the two other corner points. Fig. 3.4 illu-
strates the principle of this arrangement.
Such instruments were designated as
PIEZOVAC units and are still in use in
many cases. Today the deflection of the
diaphragm is measured as the change in
capacitance of a plate capacitor: one elec-
trode is fixed, the other is formed by the
diaphragm. When the diaphragm is deflec-
ted, the distance between the electrodes
and thus capacitance of the capacitor is
altered. Fig. 3.5 illustrates the principle of
this arrangement. A distinction is made
between sensors with metallic and those
with ceramic diaphragms. The structure of
the two types is similar and is shown on
the basis of two examples in Fig. 3.6.
Capacitance diaphragm gauges are used
from atmospheric pressure to 110
-3
mbar
(from 10
-4
mbar the measurement uncer-
tainty rises rapidly). To ensure sufficient
deflection of the diaphragms at such low
pressures, diaphragms of varying thickn-
esses are used for the various pressure
levels. In each case, the pressure can be
measured with the sensors to an accuracy
of 3 powers of ten:
1013 to 1 mbar
100 to 10
1
mbar
10 to 10
2
mbar und
1 to 10
3
mbar.
If the pressures to be measured exceed
these range limits, it is recommended that
a multichannel unit with two or three sen-
sors be used, possibly with automatic
channel changeover. The capacitance dia-
phragm gauge thus represents, for all
practical purposes, the only absolute pres-
sure measuring instrument that is inde-
pendent of the type of gas and designed
for pressures under 1 mbar. Today two
types of capacitive sensors are available:
1) Sensors DI 200 and DI 2000 with
aluminum oxide diaphragms, which are
particularly overload-free, with the
MEMBRANOVAC DM 11 and DM 12
units.
2) Sensors with Inconel diaphragms CM 1,
DM 10, CM 100, CM 1000 with extre-
mely high resolution, with the DM 21
and DM 22 units.
D00
Fig. 3.4 Piezoelectric sensor (basic diagram) Fig. 3.5 Capacitive sensor (basic diagram)
1 2
p1 p2
C1 C2
Fig. 3.6 Capacitive sensors (basic diagram)
Diaphragm
(Inconel)
Reference chamber closure
C ~ A/d
C = capacitance
A = area
d = distance C ~ A/d
C = capacitance
A = area
d = distance
Diaphragm
(ceramic)
Grid + reference chamber
closure
Capacitor plates
Signal converter
+ amplifier
24 V DC, 4 20 mA
Sensor body
(ceramic)
Electrode
Amplifier
+ 15 V DC
15 V DC
0 10 V
left: CAPACITRON (Inconel diaphragm) right: MEMBRANOVAC (aluminum oxide diaphragm)
D00 E 19.06.2001 21:38 Uhr Seite 73
Vacuum Measurement
3.2.3 Liquid-filled (mercury)
vacuum gauges
3.2.3.1 U-tube vacuum gauges
U-tube vacuum gauges filled with mercury
are the simplest and most exact instru-
ments for measuring pressure in the rough
vacuum range (1013 to a few mbar).
Unfortunately their use in technical plants
is limited because of their size and pro-
neness to breakage (see 3.4.1a).
In the evacuated limb of the U-tube vacu-
um gauge a constant pressure is maintai-
ned equal to the vapor pressure of mercu-
ry at room temperature (about 10
-3
mbar).
The other limb is connected to the volume
in which the pressure is to be measured.
From the difference in the levels of the two
columns, the pressure to be measured can
be determined on the mbar scale provided.
The reading is independent of the atmos-
pheric pressure.
3.2.3.2 Compression vacuum gauges
(according to McLeod)
The compression vacuum gauge develo-
ped by McLeod in 1874 is a very rarely
used type of vacuum gauge today. In its
refined form the instrument can be used
for absolute pressure measurement in the
high vacuum range down to 10
-5
mbar. In
the past it was frequently used as a refe-
rence instrument for the calibration of
medium and sometimes also of high vacu-
um gauges. For such measurements,
however, numerous precautionary rules
had to be taken into account before it was
possible to assess the measuring accu-
racy. The pressure is measured by com-
pressing a quantity of gas that initially
occupies a large volume into a smaller
volume by raising a mercury level. The
increased pressure obtained in this man-
ner can be measured in the same way as in
a U-tube manometer and from it the origi-
nal pressure is calculated (see equations
below).
According to the type of scale division, a
distinction is made between two forms of
compression vacuum gauges: those with a
linear scale (see Fig. 3.7) and those with a
square-law scale (see Fig. 3.8). In the case
of the compression vacuum gauges of the
McLeod linear-scale type, the ratio of the
enclosed residual volume Vc to the total
volume V must be known for each height
of the mercury level in the measurement
capillary; this ratio is shown on the scale
provided with the instrument. In the case
of compression vacuum gauges with a
square-law scale, the total volume and the
capillary diameter d must be known.
Nowadays a shortened McLeod type
compression vacuum gauge according to
Kammerer is used to measure the partial
final pressure of mechanically com-
pressing pumps. Through the high degree
of compression the condensable gas com-
ponents (vapors) are discharged as liquid
(the volume of the same mass is then
smaller by a factor of around 10
5
and can
be neglected in the measurement) so that
only the pressure of the permanently
gaseous components is measured (this is
where the expression permanent gases
comes from).
Principle of measurement with compres-
sion vacuum gauges
If h is the difference in the mercury level
between the measurement capillary and
the reference capillary (measured in mm),
then it follows from the Boyle-Mariotte
law:
p V = (p + h) Vc (3.1)
(3.1a)
p measured in mm of mercury (= torr).
If V
c
<< V, then:
(3.1b)
V
c
and V must be known, h is read off
(linear scale).
These relationships remain unchanged if
the difference in level is read off a scale
with mbar division. The pressure is then
obtained in mbar:
h in mm (3.1c)
If during measurement the mercury level
in the measurement capillary is always set
so that the mercury level in the reference
capillary corresponds to the upper end of
the measurement capillary (see Figs. 3.7
and 3.8), the volume Vc is always given by:
(3.1d)
h ....difference in level, see Fig. 3.5
d ....inside diameter of measurement
capillary
V
c
h d =
4
2
p h
V
c
V
=
4
3
p h
V
c
V
=
p h
V
c
V V
c
=
Fig. 3.7 McLeod compression vacuum gauge with linear scale (equation 3.1b)
Volume V [cm
3
]
torr
P
r
e
s
s
u
r
e

p
m
e
a
s
u
r
i
n
g
r
a
n
g
e
Lower limit for:
Vc
min.
= 5 10
3
cm
3
h
min.
= 1 mm
Upper limit for:
Vc
max.
= 0.1 cm
3
h
max.
= 100 mm
Upper limit for:
Vc
max.
= 1 cm
3
h
max.
= 100 mm
m
e
a
s
u
r
i
n
g
r
a
n
g
e
Fig. 3.8 McLeod compression vacuum gauge with square-law scale (equation 3.1f)
Volume V [cm
3
]
torr
Lower limit for:
d = 2.5 mm
Lower limit for:
d = 1 mm
Upper limit for:
d = 2.5 mm
Upper limit for:
d = 1 mm
m
e
a
s
u
r
i
n
g
r
a
n
g
e
m
e
a
s
u
r
i
n
g
r
a
n
g
e
D00 E 19.06.2001 21:38 Uhr Seite 74
Vacuum Measurement
If this term is substituted for V
c
in equati-
on (3.1b), the result is:
(3.1e)
that is, a square-law scale in mm (torr) if d
and V are measured in mm or mm
3
. If the
scale is to be divided into mbar, then the
equation is:
(3.1f)
where h in mm
d in mm
and V in mm
3
Compression vacuum gauges ensure a
reading of the sum of all partial pressures
of the permanent gases, provided that no
vapors are present that condense during
the compression procedure.
The measuring range between the top and
bottom ends is limited by the maximum
and minimum ratios of the capillary volu-
me to the total volume (see Figs. 3.7 and
3.8). The accuracy of the pressure mea-
surement depends to a great extent on the
reading accuracy. By using a vernier and
mirror, pressure measurements with an
accuracy of 2 % can be achieved. In the
low pressure range, where h is very small,
this accuracy is no longer attainable, chief-
ly because small geometric deviations
have a very noticeable effect at the closed
end of the capillary (systematic error).
The presence of vapors that may conden-
se during compression influences the
measurement, often in an indefinite man-
ner. One can easily determine whether
vapors having a pressure that is not negli-
gible are present. This can be done by set-
ting different heights h in the measure-
ment capillary under constant pressure
while using the linear scale and then cal-
culating p according to equation 3.1b. If no
vapors are present, or only those whose
pressure is negligible at room temperature
(such as mercury), then the same value of
p must result for each h.
The scale of compression vacuum gauges
can be calculated from the geometric
dimensions. This is why they were used in
the past by official calibration stations as
normal pressure (see equation 3.4.1a).
p h
d
V
=
2
3
2
p h
d
V
=
2
4
2
3.3 Vacuum gauges

with
gas-dependent
pressure reading
This type of vacuum gauge does not mea-
sure the pressure directly as an area-rela-
ted force, but indirectly by means of other
physical variables that are proportional to
the number density of particles and thus to
the pressure. The vacuum gauges with
gas-dependent pressure reading include:
the decrement gauge (3.3.1), the thermal
conductivity vacuum gauge (3.3.2) and the
ionization vacuum gauge having different
designs (3.3.3).
The instruments consist of the actual sen-
sor (gauge head, sensor) and the control
unit required to operate it. The pressure
scales or digital displays are usually based
on nitrogen pressures; if the true pressure
p
T
of a gas (or vapor) has to be determi-
ned, the indicated pressure p
I
must be
multiplied by a factor that is characteristic
for this gas. These factors differ, depen-
ding on the type of instrument, and are eit-
her given in tabular form as factors inde-
pendent of pressure (see Table 3.2) or, if
they depend on the pressure, must be
determined on the basis of a diagram (see
Fig. 3.11).
In general, the following applies:
True pressure
p
T
= indicated pressure p
I
correction factor
If the pressure is read off a nitrogen
scale but not corrected, one refers to
nitrogen equivalent values.
In all electrical vacuum gauges (they inclu-
de vacuum gauges that are dependent on
the type of gas) the increasing use of com-
puters has led to the wish to display the
pressure directly on the screen, e.g. to ins-
ert it at the appropriate place in a process
flow diagram. To be able to use the most
standardized computer interfaces possi-
ble, so-called transmitters (signal conver-
ters with standardized current outputs) are
built instead of a sensor and display unit
(e.g. THERMOVAC transmitter, Penning
transmitter, IONIVAC transmitter). Trans-
mitters require a supply voltage (e.g. +24
volts) and deliver a pressure-dependent
current signal that is linear over the entire
measuring range from 4 to 20 mA or
0 10 V. The pressure reading is not pro-
vided until after supply of this signal to the
computer and processing by the appro-
priate software and is then displayed direc-
tly on the screen.
3.3.1 Spinning rotor gauge
(SRG) (VISCOVAC)
Pressure-dependent gas friction at low gas
pressures can be utilized to measure pres-
sures in the medium and high vacuum
range. In technical instruments of this kind
a steel ball that has a diameter of several
millimeters and is suspended without
contact in a magnetic field (see Fig. 3.9) is
used as the measuring element. The ball is
set into rotation through an electro-
magnetic rotating field: after reaching a
starting speed (around 425 Hz), the ball is
left to itself. The speed then declines at a
rate that depends on the prevailing pressu-
re under the influence of the pressure-
dependent gas friction. The gas pressure
is derived from the relative decline of the
speed f (slowing down) using the follo-
wing equation:
(3.2)
p = gas pressure
r = radius of the ball
= density of the ball material
c
-
= mean speed of the gas particles,
dependent on type of gas
= coefficient of friction of the ball, inde-
pendent of the type of gas, nearly 1.
=

f
df
dt
p
c r
10
D00
Fig. 3.9 Cross-section of the gauge head of a
VISCOVAC VM 212 spinning rotor gauge
(SRG)
1 Ball
2 Measuring tube,
closed at one end,
welded into
connection flange 7
3 Permanent magnets
4 Stabilization coils
5 4 drive coils
6 Bubble level
7 Connection flange
D00 E 19.06.2001 21:38 Uhr Seite 75
Vacuum Measurement
As long as a measurement uncertainty of
3 % is sufficient, which is usually the case,
one can apply = 1 so that the sensitivity
of the spinning rotor gauge (SRG) with
rotating steel ball is given by the calculable
physical size of the ball, i.e. the product
radius x density r (see equation 3.2).
Once a ball has been calibrated, it is sui-
table for use as a transfer standard, i.e.
as a reference device for calibrating ano-
ther vacuum gauge through comparison,
and is characterized by high long-term sta-
bility. Measurements with the VISCOVAC
are not limited to measurement of the
pressure, however. Other variables invol-
ved in the kinetic theory of gases, such as
mean free path, monolayer time, particle
number density and impingement rate, can
also be measured. The instrument permits
storage of 10 programs and easy change-
over between these programs. The measu-
ring time per slowing-down operation is
between 5 seconds for high pressures and
about 40 seconds for lower pressures. The
measurement sequence of the instrument
is controlled fully automatically by a
microprocessor so that a new value is dis-
played after every measurement (slowing-
down procedure). The programs additio-
nally enable calculation of a number of sta-
tistical variables (arithmetic mean, stan-
dard deviation) after a previously specified
number of measurements.
While in the case of the kinetic theory of
gases with VISCOVAC the counting of par-
ticles directly represents the measuring
principle (transferring the particle pulses
to the rotating ball, which is thus slowed
down). With other electrical measuring
methods that are dependent on the type of
gas, the particle number density is measu-
red indirectly by means of the amount of
heat lost through the particles (thermal
conductivity vacuum gauge) or by means
of the number of ions formed (ionization
vacuum gauge).
3.3.2 Thermal conductivity
vacuum gauges
Classical physics teaches and provides
experimental confirmation that the thermal
conductivity of a static gas is independent
of the pressure at higher pressures (par-
ticle number density), p > 1 mbar. At lower
pressures, p < 1 mbar, however, the ther-
mal conductivity is pressure-dependent
(approximately proportional 1 /
M). It de-
creases in the medium vacuum range star-
ting from approx. 1 mbar proportionally to
the pressure and reaches a value of zero in
the high vacuum range. This pressure
dependence is utilized in the thermal con-
ductivity vacuum gauge and enables preci-
se measurement (dependent on the type of
gas) of pressures in the medium vacuum
range.
The most widespread measuring instru-
ment of this kind is the Pirani vacuum
gauge. A current-carrying filament with a
radius of r1 heated up to around 100 to
150 C (Fig. 3.10) gives off the heat gene-
rated in it to the gas surrounding it
through radiation and thermal conduction
(as well as, of course, to the supports at
the filament ends). In the rough vacuum
range the thermal conduction through gas
convection is virtually independent of
pressure (see Fig. 3.10). If, however, at a
few mbar, the mean free path of the gas is
of the same order of magnitude as the fila-
ment diameter, this type of heat transfer
declines more and more, becoming depen-
dent on the density and thus on the pres-
sure. Below 10
-3
mbar the mean free path
of a gas roughly corresponds to the size of
radius r
2
of the measuring tubes. The sen-
sing filament in the gauge head forms a
branch of a Wheatstone bridge. In the
THERMOTRON thermal conductivity gau-
ges with variable resistance which were
commonly used in the past, the sensing
filament was heated with a constant cur-
rent. As gas pressure increases, the tem-
perature of the filament decreases because
of the greater thermal conduction through
the gas so that the bridge becomes out of
balance. The bridge current serves as a
measure for the gas pressure, which is
indicated on a measuring scale. In the
THERMOVAC thermal conductivity gau-
ges with constant resistance which are
almost exclusively built today, the sensing
filament is also a branch of a Wheatstone
bridge. The heating voltage applied to this
bridge is regulated so that the resistance
and therefore the temperature of the fila-
ment remain constant, regardless of the
heat loss. This means that the bridge is
always balanced. This mode of regulation
involves a time constant of a few millise-
conds so that such instruments, in con-
trast to those with variable resistance, res-
pond very quickly to pressure changes.
The voltage applied to the bridge is a mea-
sure of the pressure. The measuring volta-
ge is corrected electronically such that an
approximately logarithmic scale is obtai-
ned over the entire measuring range. Ther-
l r
1
-
l r
2
-
l r
1
-
r
1
r
2
l r r
2 1
-
I II III
10
5
10
4
10
3
10
2
10
1
Druck [mbar]
A
b
g
e
f
h
r
t
e
r

W
r
m
e
f
l
u
100 10 1
Fig. 3.10 Dependence of the amount heat dissipated by a heated filament (radius r
1
) in a tube (radius r
2
) at a constant
temperature difference on the gas pressure (schematic diagram)
Pressure [mbar]
H
e
a
t

l
o
s
s
I Thermal dissipation due to radiation and conduction in the metallic ends
II Thermal dissipation due to the gas, pressure-dependent
III Thermal dissipation due to radiation and convection
D00 E 19.06.2001 21:38 Uhr Seite 76
Vacuum Measurement
mal conductivity vacuum gauges with con-
stant resistance have a measuring range
from 10
-4
to 1013 mbar. Due to the very
short response time, they are particularly
suitable for controlling and pressure moni-
toring applications (see section 3.5). The
measurement uncertainty varies in the dif-
ferent pressure ranges. The maximum
error at full-scale deflection is about 1 to
2 %. In the most sensitive range, i.e. bet-
ween 10
-3
and 1 mbar, this corresponds to
around 10 % of the pressure reading. The
measurement uncertainty is significantly
greater outside this range.
As in all vacuum gauges dependent on the
type of gas, the scales of the indicating
instruments and digital displays in the
case of thermal conductivity vacuum gau-
ges also apply to nitrogen and air. Within
the limits of error, the pressure of gases
with similar molecular masses, i.e. O
2
, CO
and others, can be read off directly. Cali-
bration curves for a series of gases are
shown in Fig. 3.11.
An extreme example of the discrepancy
between true pressure p
T
and indicated
pressure p
I
in pressure measurement is
the admission of air to a vacuum system
with argon from a pressure cylinder to
avoid moisture (pumping time). According
to Fig. 3.11, one would obtain a p
I
reading
of only 40 mbar on reaching an Ar atmos-
pheric pressure p
T
with a THERMOVAC as
a pressure measuring instrument. Argon
might escape from the vessel (cover
opens, bell jar rises). For such and similar
applications, pressure switches or vacuum
gauges that are independent of the type of
gas must be used (see section 3.2).
3.3.3 Ionization vacuum
gauges
Ionization vacuum gauges are the most
important instruments for measuring gas
pressures in the high and ultrahigh vacu-
um ranges. They measure the pressure in
terms of the number density of particles
proportional to the pressure. The gas
whose pressure is to be measured enters
the gauge heads of the instruments and is
partially ionized with the help of an electric
field. Ionization takes place when electrons
are accelerated in the electric field and
attain sufficient energy to form positive
ions on impact with gas molecules. These
ions transmit their charge to a measuring
electrode (ion collector) in the system. The
ion current, generated in this manner (or,
more precisely, the electron current in the
feed line of the measuring electrode that is
required to neutralize these ions) is a mea-
sure of the pressure because the ion yield
is proportional to the particle number den-
sity and thus to the pressure.
The formation of ions is a consequence of
either a discharge at a high electric field
strength (so-called cold-cathode or Pen-
ning discharge, see 3.3.3.1) or the impact
of electrons that are emitted from a hot
cathode (see 3.3.3.2).
Under otherwise constant conditions, the
ion yield and thus the ion current depend
on the type of gas since some gases are
easier to ionize than others. As all vacuum
gauges with a pressure reading that is
dependent on the type of gas, ionization
vacuum gauges are calibrated with nitro-
gen as the reference gas (nitrogen equiva-
lent pressure, see 3.3). To obtain the true
pressure for gases other than nitrogen, the
read-off pressure must be multiplied by
the correction factor given in Table 3.2 for
the gas concerned. The factors stated in
Table 3.2 are assumed to be independent
of the pressure, though they depend
somewhat on the geometry of the electro-
de system. Therefore, they are to be regar-
ded as average values for various types of
ionization vacuum gauges (see Fig. 3.16).
3.3.3.1 Cold-cathode ionization
vacuum gauges (Penning
vacuum gauges)
Ionization vacuum gauges which operate
with cold discharge are called cold-catho-
de- or Penning vacuum gauges. The
discharge process in a measuring tube is,
in principle, the same as in the electrode
system of a sputter ion pump (see section
2.1.8.3). A common feature of all types of
cold-cathode ionization vacuum gauges is
that they contain just two unheated
electrodes, a cathode and an anode, bet-
ween which a so-called cold discharge is
initiated and maintained by means of a d.c.
voltage (of around 2 kV) so that the
discharge continues at very low pressures.
This is achieved by using a magnetic field
to make the paths of the electrons long
enough so that the rate of their collision
with gas molecules is sufficiently large to
form the number of charge carriers requi-
red to maintain the discharge. The magne-
tic field (see Fig. 3.12) is arranged such
that the magnetic field lines of force cross
the electric field lines. In this way the elec-
trons are confined to a spiral path. The
positive and negative charge carriers pro-
D00
Fig. 3.11 Calibration curves of THERMOVAC gauges for various gases, based on nitrogen equivalent reading
Indicated pressure [mbar] p
I
T
r
u
e

p
r
e
s
s
u
r
e

[
m
b
a
r
]

p
W
p
I
> p
T
p
I
< p
T
N
2
, O
2
air p
I
= p
T
D00 E 19.06.2001 21:38 Uhr Seite 77
Vacuum Measurement
duced by collision move to the correspon-
ding electrodes and form the pressure-
dependent discharge current, which is
indicated on the meter. The reading in
mbar depends on the type of gas. The
upper limit of the measuring range is given
by the fact that above a level of several
10
-2
mbar the Penning discharge changes
to a glow discharge with intense light out-
put in which the current (at constant volta-
ge) depends only to a small extent on the
pressure and is therefore not suitable for
measurement purposes. In all Penning
gauges there is considerably higher gas
sorption than in ionization vacuum gauges
that operate with a hot cathode. A Penning
measuring tube pumps gases similarly to a
sputter ion pump (S 10
-2
l/s). Here again
the ions produced in the discharge are
accelerated towards the cathode where
they are partly retained and partly cause
sputtering of the cathode material. The
sputtered cathode material forms a gette-
ring surface film on the walls of the gauge
tube. In spite of these disadvantages,
which result in a relatively high degree of
inaccuracy in the pressure reading (up to
around 50 %), the cold-cathode ionization
gauge has three very outstanding advanta-
ges. First, it is the least expensive of all
high vacuum measuring instruments.
Second, the measuring system is insensi-
tive to the sudden admission of air and to
vibrations; and third, the instrument is
easy to operate.
3.3.3.2 Hot-cathode ionization vacuum
gauges
Generally speaking, such gauges refer to
measuring systems consisting of three
electrodes (cathode, anode and ion collec-
tor) where the cathode is a hot cathode.
Cathodes used to be made of tungsten but
are now usually made of oxide-coated iri-
dium (Th
2
O
3
, Y
2
O
3
) to reduce the electron
output work and make them more resi-
stant to oxygen. Ionization vacuum gauges
of this type work with low voltages and
without an external magnetic field. The hot
cathode is a very high-yield source of elec-
trons. The electrons are accelerated in the
electric field (see Fig. 3.13) and receive
sufficient energy from the field to ionize
the gas in which the electrode system is
located. The positive gas ions formed are
transported to the ion collector, which is
negative with respect to the cathode, and
give up their charge there. The ion current
thereby generated is a measure of the gas
density and thus of the gas pressure. If i
-
is the electron current emitted by the hot
cathode, the pressure-proportional current
i
+
produced in the measuring system is
defined by:
i
+
= C i p und (3.3)
(3.3a)
The variable C is the vacuum gauge con-
stant of the measuring system. For nitro-
gen this variable is generally around
10 mbar
-1
. With a constant electron cur-
rent the sensitivity S of a gauge head is
defined as the quotient of the ion current
and the pressure. For an electron current
of 1 mA and C = 10 mbar
-1
, therefore, the
sensitivity S of the gauge head is:
S = i
+
/ p = C i
-
= 10 mbar
-1
1 mA
= 10 mbar
-1
10
-3
A
= 1 10
-2
A/mbar.
Hot-cathode ionization vacuum gauges
also exhibit gas sorption (pumping
action), which, however, is considerably
smaller than with Penning systems, i.e.
approx. 10
-3
l/s. Essentially this gas sorp-
tion takes place on the glass wall of the
gauge head and, to a lesser extent, at the
ion collector. Here use is made of nude
gauges that are easy to operate because an
external magnet is not needed. The upper
limit of the measuring range of the hot-
cathode ionization gauge is around
10
-2
mbar (with the exception of special
designs). It is basically defined by the
scatter processes of ions at gas molecules
due to the shorter free path at higher pres-
sures (the ions no longer reach the ion
collector = lower ion yield). Moreover,
uncontrollable glow or arc discharges may
form at higher pressures and electrostatic
discharges can occur in glass tubes. In
these cases the indicated pressure p
I
may
deviate substantially from the true pressu-
re p
T
.
At low pressures the measuring range is
limited by two effects: by the X-ray effect
and by the ion desorption effect. These
effects results in loss of the strict propor-
tionality between the pressure and the ion
current and produce a low pressure thres-
hold that apparently cannot be crossed
(see Fig. 3.14).
The X-ray effect (see Fig. 3.15)
The electrons emitted from the cathode
impinge on the anode, releasing photons
(soft X-rays). These photons, in turn,
trigger photoelectrons from surfaces they
p
i
i C
=
Fig. 3.12 Cross-section of PENNINGVAC PR 35 gauge

1 Small flange
DN 25 KF;
DN 40 KF
2 Housing
3 Ring anode with
ignition pin
4 Ceramic washer
5 Current leadthrough
6 Connecting bush
7 Anode pin
8 Cathode plate
Fig. 3.13 Schematic diagram and potential curve in a
hot-cathode ionization vacuum gauge
i+: ion current
i-: electron current
Anode
U
C
(+ 50V)
U
A
(+ 200V)
U
A
U
C
Cathode Ion collector
D00 E 19.06.2001 21:38 Uhr Seite 78
Vacuum Measurement
strike. The photoelectrons released from
the ion collector flow to the anode, i.e. the
ion collector emits an electron current,
which is indicated in the same manner as
a positive ion current flowing to the ion
collector. This photocurrent simulates a
pressure. This effect is called the positive
X-ray effect, and it depends on the anode
voltage as well as on the size of the surfa-
ce of the ion collector.
Under certain circumstances, however,
there is also a negative X-ray effect. Pho-
tons which impinge on the wall surroun-
ding the gauge head release photoelec-
trons there, which again flow towards the
anode, and since the anode is a grid struc-
ture, they also flow into the space within
the anode. If the surrounding wall has the
same potential as the ion collector, e.g.
ground potential, a portion of the electrons
released at the wall can reach the ion
collector. This results in the flow of an
electron current to the ion collector, i.e. a
negative current flows which can compen-
sate the positive ion current. This negative
X-ray effect depends on the potential of the
outer wall of the gauge head.
The ion desorption effect
Adsorbed gases can be desorbed from a
surface by electron impact. For an ionizati-
on gauge this means that, if there is a layer
of adsorbed gas on the anode, these gases
are partly desorbed as ions by the impin-
ging electrons. The ions reach the ion
collector and lead to a pressure indication
that is initially independent of the pressure
but rises as the electron current increases.
If such a small electron current is used so
that the number of electrons incident at the
surface is small compared to the number of
adsorbed gas particles, every electron will
be able to desorb positive ions. If the elec-
tron current is then increased, desorption
will initially increase because more elec-
trons impinge on the surface. This finally
leads to a reduction in adsorbed gas par-
ticles at the surface. The reading falls again
and generally reaches values that may be
considerably lower than the pressure rea-
ding observed with a small electron cur-
rent. As a consequence of this effect in
practice, one must ascertain whether the
pressure reading has been influenced by a
desorption current. This can be done most
simply by temporarily altering the electron
current by a factor of 10 or 100. The rea-
ding for the larger electron current is the
more precise pressure value.
In addition to the conventional ionization
gauge, whose electrode structure resemb-
les that of a common triode, there are
various ionization vacuum gauge systems
(Bayard-Alpert system, Bayard-Alpert
system with modulator, extractor system)
which more or less suppress the two
effects, depending on the design, and are
therefore used for measurement in the
high and ultrahigh vacuum range. Today
the Bayard-Alpert system is usually the
standard system.
a) The conventional ionization vacuum
gauge
A triode of conventional design (see Fig.
3.16 a) is used as the gauge head, but it is
slightly modified so that the outer electro-
de serves as the ion collector and the grid
within it as the anode. With this arrange-
ment the electrons are forced to take very
long paths (oscillating around the grid
wires of the anode) so that the probability
of ionizing collisions and thus the sensiti-
vity of the gauge are relatively high. Becau-
se the triode system can generally only be
used in high vacuum on account of its
strong X-ray effect, the gas sorption (pum-
ping) effect and the gas content of the
electrode system have only a slight effect
on the pressure measurement.
D00
Fig. 3.14 Apparent low pressure limit due to X-ray
effect in a normal ionization vacuum gauge
Actual pressure mbar
I
n
d
i
c
a
t
e
d

p
r
e
s
s
u
r
e

m
b
a
r
Fig. 3.15 Explanation of the X-ray effect in a conventio-
nal ionization gauge. The electrons e
-
emitted
by the cathode C collide with anode A and
trigger a soft X-ray radiation (photons) there.
This radiation strikes, in part, the ion collec-
tor and generates photoelectrons e
s
there
C Cathode
A Anode
I Ion collector
I Pressure reading without X-ray effect
II Apparent low pressure limit due to X-ray effect
III Sum of I and II
Fig. 3.16 Schematic drawing of the electrode
arrangement of various ionization vacuum
gauge measuring systems
I ion collector
Sc screen
M modulator
A anode
C cathode
R reflector
a)
b)
c)
d)
e)
Conventional
ionization
vacuum gauge
system
ionization vacu-
um gauge
system for higher
pressures (up to
1 mbar)
Bayard-Alpert
ionization
vacuum gauge
system
Bayard-Alpert
ionization
vacuum gauge
system with
modulator
extractor
ionization
vacuum gauge
system
D00 E 19.06.2001 21:38 Uhr Seite 79
Vacuum Measurement
b) The high-pressure ionization vacuum
gauge (up to 1 mbar)
A triode is again used as the electrode
system (see Fig. 3.16 b), but this time with
an unmodified conventional design. Since
the gauge is designed to allow pressure
measurements up to 1 mbar, the cathode
must be resistant to relatively high oxygen
pressure. Therefore, it is designed as a so-
called non-burnout cathode, consisting of
an yttria-coated iridium ribbon. To obtain a
rectilinear characteristic (ion current as a
linear function of the pressure) up to a
pressure of 1 mbar, a high-ohmic resistor
is installed in the anode circuit.
c) Bayard-Alpert ionization vacuum
gauge (the standard measuring
system used today)
To ensure linearity between the gas pres-
sure and the ion current over as large a
pressure range as possible, the X-ray
effect must be suppressed as far as possi-
ble. In the electrode arrangement develo-
ped by Bayard and Alpert, this is achieved
by virtue of the fact that the hot cathode is
located outside the anode and the ion
collector is a thin wire forming the axis of
the electrode system (see Fig. 3.16 c). The
X-ray effect is reduced by two to three
orders of magnitude due to the great
reduction in the surface area of the ion
collector. When pressures in the ultrahigh
vacuum range are measured, the inner
surfaces of the gauge head and the
connections to the vessel affect the pres-
sure reading. The various effects of ad-
sorption, desorption, dissociation and flow
phenomena cannot be dealt with in this
context. By using Bayard-Alpert systems
as nude gauge systems that are placed
directly in the vessel, errors in measure-
ment can be extensively avoided because
of the above mentioned effects.
d) Bayard-Alpert ionization vacuum gauge
with modulator
The Bayard-Alpert system with modulator
(see Fig. 3.16 d), introduced by Redhead,
offers pressure measurement in which
errors due to X-ray and ion desorption
effects can be quantitatively taken into
account. In this arrangement there is a
second thin wire, the modulator, near the
anode in addition to the ion collector insi-
de the anode. If this modulator is set at the
anode potential, it does not influence the
measurement. If, on the other hand, the
same potential is applied to the modulator
as that on the ion collector, part of the ion
current formed flows to the modulator and
the current that flows to the ion collector
becomes smaller. The indicated pressure
p
I
of the ionization gauge with modulator
set to the anode potential consists of the
portion due to the gas pressure p
g
and that
due to the X-ray effect p
:
p
A
= p
g
+ p
(3.4)
After switching the modulator from the
anode potential over to the ion collector
potential, the modulated pressure reading
p
M
is lower than the p
I
reading because a
portion of the ions now reaches the modu-
lator. Hence:
p
M
= p
g
+ p
(3.5)
with < 1.
The p
g
share of the X-ray effect is the same
in both cases. After determining the diffe-
rence between (3.4) and (3.5), we obtain
the equation for the gas pressure p
g
:
(3.6)
can immediately be determined by expe-
riment at a higher pressure (around
10
-6
mbar) at which the X-ray effect and
thus p
are negligible. The pressure corre-

sponding to the two modulator potentials
are read off and their ratio is formed. This
modulation method has the additional
advantage that the ion desorption effect is
determined in this way. It permits pressu-
re measurements up to the 10
-11
mbar
range with relatively little effort.
e) Extractor ionization vacuum gauge
Disruptive effects that influence pressure
measurement can also be extensively eli-
minated by means of an ion-optical system
first suggested by Redhead. With this
extractor system (see Fig. 3.16 e) the ions
from the anode cylinder are focused on a
very thin and short ion collector. The ion
collector is set up in a space, the rear wall
of which is formed by a cup-shaped elec-
trode that is maintained at the anode
potential so that it cannot be reached by
ions emanating from the gas space. Due to
the geometry of the system as well as the
potential of the of individual electrodes,
the disruptive influences through X-ray
effects and ion desorption are almost com-
pletely excluded without the need of a
modulator. The extractor system measures
pressures between 10
-4
and 10
-12
mbar.
p
g
p
A
p
M
=

1
Another advantage is that the measuring
system is designed as a nude gauge with a
diameter of only 35 mm so that it can be
installed in small apparatus.
D00 E 19.06.2001 21:38 Uhr Seite 80
Vacuum Measurement
3.4 Adjustment and
calibration; DKD,
PTB national
standards
Definition of terms: Since these terms are
often confused in daily usage, a clear defi-
nition of them will first be provided:
Adjustment or tuning refers to the correct
setting of an instrument. For example, set-
ting 0 and 100 % in THERMOVACs or set-
ting the mass spectrometer to mass 4 in
the helium leak detector.
Calibration inspection refers to com-
parison with a standard in accordance with
certain statutory regulations by specially
authorized personnel (Bureau of Stand-
ards). If the outcome of this regular ins-
pection is positive, an operating permit for
the next operation period (e.g. three years)
is made visible for outsiders by means of a
sticker or lead seal. If the outcome is nega-
tive, the instrument is withdrawn from
operation.
Calibration refers to comparison with a
standard in accordance with certain
statutory regulations by specially authori-
zed personnel (calibration facility). The
result of this procedure is a calibration cer-
tificate which contains the deviations of
the readings of the instrument being cali-
brated from the standard.
Calibration facilities carry out this calibra-
tion work. One problem that arises is the
question of how good the standards are
and where they are calibrated. Such stan-
dards are calibrated in calibration facilities
of the German Calibration Service (DKD).
The German Calibration Service is mana-
ged by the Federal Physical-Technical
Institute (PTB). Its function is to ensure
that measuring and testing equipment
used for industrial measurement purposes
is subjected to official standards. Calibrati-
on of vacuum gauges and test leaks within
the framework of the DKD has been assig-
ned to LEYBOLD, as well as other compa-
nies, by the PTB. The required calibration
pump bench was set up in accordance
with DIN 28 418 (see Table 11.1) and then
inspected and accepted by the PTB. The
standards of the DKD facilities, so-called
transfer standards (reference vacuum
gauges), are calibrated directly by the PTB
at regular intervals. Vacuum gauges of all
makes are calibrated on an impartial basis
by LEYBOLD in Cologne according to
customer order. A DKD calibration certifi-
cate is issued with all characteristic data
on the calibration. The standards of the
Federal Physical-Technical Institute are the
so-called national standards. To be able to
guarantee adequate measuring accuracy
or as little measurement uncertainty as
possible in its calibrations, the PTB largely
carries out its measurements through the
application of fundamental methods. This
means, for example, that one attempts to
describe the calibration pressures through
the measurement of force and area or by
thinning the gases in strict accordance
with physical laws. The chain of the recali-
bration of standard instruments carried
out once a year at the next higher qualified
calibration facility up to the PTB is called
resetting to national standards. In other
countries as well, similar methods are car-
ried out by the national standards institu-
tes as those applied by the Federal Physi-
cal-Technical Institute (PTB) in Germany.
Fig. 3.17 shows the pressure scale of the
PTB. Calibration guidelines are specified in
DIN standards (DIN 28 416) and ISO pro-
posals.
3.4.1 Examples of fundamental
pressure measurement
methods (as standard
methods for calibrating
vacuum gauges)
a) Measuring pressure with a reference
gauge
An example of such an instrument is the
U-tube vacuum gauge, with which the
measurement of the pressure in the mea-
surement capillary is based on a measure-
ment of the weight over the length of the
mercury column.
In the past the McLeod vacuum gauge was
also used for calibration purposes. With a
precision-made McLeod and carefully exe-
cuted measurements, taking into account
all possible sources of error, pressures
down to 10
-4
mbar can be measured with
considerable accuracy by means of such
an instrument.
Another reference gauge is the VISCOVAC
decrement gauge with rotating ball (see
3.3.1) as well as the capacitance dia-
phragm gauge (see 3.2.2.4).
b) Generation of a known pressure; static
expansion method
On the basis of a certain quantity of gas
whose parameters p, V and T are known
exactly p lies within the measuring range
of a reference gauge such as a U-tube or
D00
10
12
10
9
10
6
10
3
10
0
10
3
0.1
0.3
1
3
10
30
R
e
l
a
t
i
v
e

u
n
c
e
r
t
a
i
n
l
y

o
f

t
h
e

p
r
e
s
s
u
r
e

d
e
t
e
r
m
i
n
a
t
i
o
n

[
%
]
Pressure [mbar]
Dynamic
expansion
Molecular
beam
Static expansion
U-Tube
Fig. 3.17 Pressure scale of Federal Physical-Technical Institute (PTB), Berlin, (status as at August 1984) for inert gases,
nitrogen and methane
D00 E 19.06.2001 21:38 Uhr Seite 81
Vacuum Measurement
McLeod vacuum gauge a lower pressure
within the working range of ionization gau-
ges is reached via expansion in several
stages.
If the gas having volume V1 is expanded to
a volume (V
1
+ V
2
), and from V
2
to
(V
2
+ V
3
), etc., one obtains, after n stages
of expansion:
(3.7)
p
1
= initial pressure measured directly in
mbar
p
n
= calibration pressure
The volumes here must be known as pre-
cisely as possible (see Fig. 3.18) and the
temperature has to remain constant. This
method requires that the apparatus used
be kept very clean and reaches its limit at
pressures where the gas quantity can be
altered by desorption or adsorption effects
beyond the permissible limits of error.
According to experience, this lower limit is
around 5 10
-7
mbar. This method is cal-
led the static expansion method because
the pressure and volume of the gas at rest
are the decisive variables.
p
n
p
V
V V
V
V V
V
n
V
n
V
n
=
+

+

+
1
1
1 2
2
2 3
1
1
c) Dynamic expansion method
(see Fig. 3.19)
According to this method, the calibration
pressure p is produced by admitting gas at
a constant throughput rate Q into a va-
cuum chamber while gas is simultaneous-
ly pumped out of the chamber by a pump
unit with a constant pumping speed S. At
equilibrium the following applies accor-
ding to equation 1.10 a:
p = Q/S
Q is obtained either from the quantity of
gas that flows into the calibration chamber
from a supply vessel in which constant
pressure prevails or from the quantity of
gas flowing into the calibration chamber at
a measured pressure through a known
conductance. The pressure in front of the
inlet valve must be high enough so that it
can be measured with a reference gauge.
The inlet apertures of the valve (small
capillaries, sintered bodies) must be so
small that the condition d << is met, i.e.
a molecular flow and hence a constant
conductance of the inlet valve are obtained
(see Section 1.5). The quantity of gas is
then defined by p
1
L
1
, where
p
1
= pressure in front of the inlet valve and
L
1
= conductance of the valve. The pum-
ping system consists of a precisely mea-
sured aperture with the conductance L
2
in
a wall that is as thin as possible (screen
conductance) and a pump with a pumping
speed of PS
p
:
(3.8)
and thus
(3.9)
This method has the advantage that, after
reaching a state of equilibrium, sorption
effects can be ignored and this procedure
can therefore be used for calibrating gau-
ges at very low pressures.
p
2
p p ) (1 +
L
S
L L
L S
= =
1 1
1 1 2
2
p
S
L S
p
L S
p
L
L
S
p
=

+
=
+
2
2
2
1
2
+
+
+ + +
+
+
IM IM
13000 cm
3
V =
4
p
4
p
3
p
2
p
1
V = 25 cm
3
3
V = 25 cm
1
3
V = 1000 cm
2
3
Fig. 3.18 Generation of low pressures through static expansion Fig. 3.19 Apparatus for calibration according to the dynamic expansion method
p2 = p1 (Sp >> L2)
L
1
S
1 Volume 1
2 Volume 2
3 Inlet valve
(conductance L1)
4 Aperture with
conductance L2
5 Valve
6 to pump system
7 Valve
8 to gas reservoir
9 Valve
10 LN
2
cold trap
11 to pump system
12 U-tube vacuum
gauge
13 McLeod vacuum
gauge
14 Valve
15 Calibrated
ionization gauge
tube
16 to pump
(pumping speed
PSp)
17 Gas inlet
18 Mass spectrometer
19,20 Gauges to be
calibrated
21 Nude gauge to be
calibrated
22 Bake-out furnace
D00 E 19.06.2001 21:38 Uhr Seite 82
Vacuum Measurement
3.5 Pressure
monitoring, control
and regulation in
vacuum systems
3.5.1 Fundamentals of pressure
monitoring and control
In all vacuum processes the pressure in
the system must be constantly checked
and, if necessary, regulated. Modern plant
control additionally requires that all mea-
sured values which are important for
monitoring a plant are transmitted to cen-
tral stations, monitoring and control cen-
ters and compiled in a clear manner. Pres-
sure changes are frequently recorded over
time by recording equipment. This means
that additional demands are placed on
vacuum gauges:
a) continuous indication of measured
values, analog and digital as far as pos-
sible
b) clear and convenient reading of the
measured values
c) recorder output to connect a recording
instrument or control or regulation
equipment
d) built-in computer interface (e.g. RS 232)
e) facility for triggering switching operati-
ons through built-in trigger points
These demands are generally met by all
vacuum gauges that have an electric mea-
sured value display, with the exception of
Bourdon, precision diaphragm and liquid-
filled vacuum gauges. The respective con-
trol units are equipped with recorder out-
puts that supply continuous voltages bet-
ween 0 and 10 V, depending on the pres-
sure reading on the meter scale, so that
the pressure values can be recorded over
time by means of a recording instrument.
If a pressure switching unit is connected to
the recorder output of the gauge, swit-
ching operations can be triggered when
the values go over or below specified set-
points. The setpoints or switch threshold
values for triggering switching operations
directly in the gauges are called trigger
values. Apart from vacuum gauges, there
are diaphragm pressure switches that trig-
ger a switching operation (without display
of a measured value) via a contact ampli-
fier when a certain pressure is reached.
Valves, for example, can also be controlled
through such switching operations.
3.5.2 Automatic protection,
monitoring and control of
vacuum systems
Protection of a vacuum system against
malfunctions is extremely important. In
the event of failure, very high material
values may be at risk, whether through
loss of the entire system or major compo-
nents of it, due to loss of the batch of
material to be processed or due to further
production down time. Adequate operatio-
nal control and protection should therefo-
re be provided for, particularly in the case
of large production plants. The individual
factors to be taken into account in this
connection are best illustrated on the basis
of an example: Fig. 3.20 shows the sche-
matic diagram of a high vacuum pump
system. The vessel (11) can be evacuated
by means of a Roots pump (14) or a diffu-
sion pump (15), both of which operate in
conjunction with the backing pump (1).
The Roots pump is used in the medium
vacuum range and the diffusion pump in
the high vacuum range. The valves (3), (8)
and (16) are operated electropneuma-
tically. The individual components are
actuated from a control panel with push-
buttons. The pump system is to be protec-
ted against the following malfunctions:
a) power failure
b) drop in pressure in the compressed air
network
c) failure of cooling water to the diffusion
pump
d) fault in diffusion pump heating system
e) failure of backing pump
f) pressure rise in the vessel above a
maximum permissible value
g) pressure rise above a maximum
backing pressure (critical forepressure
of the diffusion pump)
The measures to be taken in order to fore-
stall such malfunctions will be discussed
in the same order:
a) Measures in the event of power failure:
All valves are closed so as to prevent
admission of air to the vacuum vessel
and protect the diffusion pump against
damage.
b) Protection in the event of a drop in
pressure in the compressed air net-
work: The compressed air is monitored
by a pressure monitoring device (5). If
D00
Fig. 3.20 Schematic diagram of a high vacuum pump system with optional operation of a Roots pump or
a diffusion pump
1 Backing pump
2 Backing pressure
monitoring device
3 Electropneumatic valve
4 Compressed air connection
5 Pressure monitoring device
6 Temperature monitoring device
7 Cooling water monitoring device
9 Recorder
10 High-vacuum monitoring device
11 Vessel
12 High-vacuum gauge
13 Limit switches
14 Roots pump
15 Diffusion pump
17 Venting valve
D00 E 19.06.2001 21:39 Uhr Seite 83
Vacuum Measurement
the pressure falls under a specified
value, a signal can initially be emitted or
the valves can be automatically closed.
In this case, a sufficient reserve supply
of compressed air is necessary (not
shown in Fig. 3.20), which allows all
valves to be actuated at least once.
c) Measures in the event of failure of coo-
ling water to the diffusion pump: The
cooling water is monitored by a flow or
temperature monitoring device (6) and
(7). If the flow of cooling water is ina-
dequate, the heater of the diffusion
pump is switched off and a signal is
given; the valve (8) closes.
d) Protection against failure of the diffusi-
on pump heater: Interruption of the dif-
fusion pump heating system can be
monitored by a relay. If the temperature
rises above a maximum permissible
value, a temperature monitoring device
(6) responds. In both cases the valve
(8) closes and a signal is given.
e) Protection in the event of backing pump
failure: Belt-driven backing pumps
must have a centrifugal switch which
shuts down the entire system in the
event of belt breakage or another
malfunction. Monoblock pumps for
which the drive is mounted directly on
the shaft can be monitored by current
relays and the like.
f) Protection against a pressure rise in the
vessel above a certain limit value: The
high vacuum monitoring device (10)
emits a signal when a specified pressu-
re is exceeded.
g) Ensuring the critical forepressure of the
diffusion pump: When a certain backing
pressure is exceeded, all valves are clo-
sed by the backing pressure monitoring
device (2), the pumps are switched off
and again a signal is given. The position
of the valves (3), (8) and (16) is indica-
ted on the control panel by means of
limit switches (13). The pressure in the
vessel is measured with a high vacuum
gauge (12) and recorded with a recor-
der (9). Protection against operating
errors can be provided by interlocking
the individual switches so that they can
only be actuated in a predetermined
sequence. The diffusion pump, for
example, may not be switched on when
the backing pump is not running or the
required backing pressure is not main-
tained or the cooling water circulation is
not functioning.
In principle, it is not a big step from a
system protected against all malfunctions
to a fully automatic, program-controlled
plant, though the complexity of the electri-
cal circuits, of course, increases signifi-
cantly.
3.5.3 Pressure regulation and
control in rough and
medium vacuum systems
Control and regulation have the function of
giving a physical variable in this case the
pressure in the vacuum system a certain
value. The common feature is the actuator
which changes the energy supply to the
physical variable and thus the variable its-
elf. Control refers to influencing a system
or unit through commands. In this case
the actuator and hence the actual value of
the physical variable is changed directly
with a manipulated variable. Example:
Actuation of a valve by means of a pressu-
re-dependent switch. The actual value may
change in an undesirable way due to addi-
tional external influences. The controlled
unit cannot react to the control unit. For
this reason control systems are said to
have an open operating sequence. In the
case of regulation, the actual value of the
physical variable is constantly compared
to the specified setpoint and regulated if
there is any deviation so that it completely
approximates the setpoint as far as possi-
ble. For all practical purposes regulation
always requires control. The main differen-
ce is the controller in which the setpoint
and the actual value are compared. The
totality of all elements involved in the con-
trol process forms the control circuit. The
terms and characteristic variables for
describing control processes are stipula-
ted in DIN 19226.
Generally a distinction is made between
discontinuous control (e.g. two-step or
three-step control) with specification of a
pressure window, within which the pressu-
re may vary, and continuous control (e.g.
PID control) with a specified pressure set-
point, which should be maintained as pre-
cisely as possible. We have two possible
ways of adjusting the pressure in a vacu-
um system: first, by changing the pumping
speed (altering the speed of the pump or
throttling by closing a valve); second,
through admission of gas (opening a
valve). This results in a total of 4 procedu-
res.
Discontinuous pressure regulation
Although continuous regulation undoub-
tedly represents the more elegant proce-
dure, in many cases two-step or three-step
regulation is fully adequate in all vacuum
ranges. To specify the pressure window,
two or three variable, pressure-dependent
switch contacts are necessary. It does not
matter here whether the switch contacts
are installed in a gauge with display or in a
downstream unit or whether it is a pressu-
re switch without display. Fig. 3.21 illust-
rates the difference between two-step
regulation through pumping speed thrott-
ling, two-point regulation through gas
Time Time Time
Pressure Pressure Pressure
p
Atm
p
Atm
p
Atm
p
max
p
max
p
min
p
min
p
mitte
p
min
pumping speed throttling
and gas admission
pumping speed throttling
Two-step regulation through Three-step regulation through
gas admission
Fig. 3.21 Schematic diagram of two-step and three-step regulation
D00 E 19.06.2001 21:39 Uhr Seite 84
Vacuum Measurement
admission and three-point regulation
through a combination of pumping speed
throttling and gas admission. Figures 3.22
and 3.23 show the circuit and structure of
the two two-step regulation systems. In
the case of two-step regulation through
pumping speed throttling (Fig. 3.22), vol-
tage is supplied to pump valve 4, i.e. it is
open when the relay contacts are in the
release condition. At a level below the
upper switching point the valve remains
open because of the self-holding function
of the auxiliary relay. Only at a level below
the lower switching point is the relay lat-
ching released. If the pressure subse-
quently rises, the valve is opened again at
the upper switching point.
In the case of two-step regulation through
gas admission, the inlet valve is initially
closed. If the upper pressure switching
point is not reached, nothing changes;
only when the pressure falls below the
lower switching point, do the make
contacts open the gas inlet valve and
actuate the auxiliary relay with self-holding
function simultaneously. Return to the idle
state with closing of the gas inlet valve is
not effected until after the upper switching
point is exceeded due to the release of the
relay self-holding function.
Fig. 3.24 shows the corresponding three-
step regulation system which was created
with the two components just described.
As the name indicates, two switching
points, the lower switching point of the
regulation system through pumping speed
throttling and the upper switching point of
the gas inlet regulation system, were com-
bined.
To avoid the complicated installation with
auxiliary relays, many units offer a facility
for changing the type of function of the
built-in trigger values via software. Initially
one can choose between individual swit-
ching points (so-called level triggers)
and interlinked switching points (interval
triggers). These functions are explained
in Fig. 3.25. With interval triggers one can
also select the size of the hysteresis and
D00
Fig. 3.22 Two-step regulation through pumping speed throttling
' Gauge with two
switching points
^ Throttle valve
Vacuum pump
Pump valve
I Vacuum vessel
Fu Fuse
R, Mp Mains connection 220 V/50 Hz
S
max
Switching point for maximum value
S
min
Switching point for minimum value
PV Pump valve
R1 Auxiliary relay for pump valve
K1 Relay contact of R1
M Measuring and switching device
Fig. 3.23 Two-step regulation through gas admission
' Gauge with two
switching points
^ Variable-leak valve
Inlet valve
Gas supply
I Throttle valve
I Vacuum pump
I Vacuum vessel
Fu Fuse
S
max
S
min
EV Inlet valve
R2 Auxiliary relay for inlet valve
Fig. 3.24 Three-step regulation system
' Gauge with three
switching points
^ Variable-leak valve
Variable-leak valve
Inlet valve
I Gas supply
I Throttle valve
I Vacuum pump
I Pump valve
^ Vacuum vessel
Fu Fuse
S
max
S
mitte
Switching point for mean value
S
min
T TORROSTAT S 020
PV Pump valve
EV Inlet valve
R1 Auxiliary relay for pump interval
R2 Auxiliary relay for inlet interval
Fig. 3.25 Diagram of level triggers and interval triggers
D00 E 19.06.2001 21:39 Uhr Seite 85
Vacuum Measurement
the type of setpoint specification, i.e. either
fixed setting in the unit or specification
through an external voltage, e.g. from 0
10 volts. A three-step regulation system
(without auxiliary relay), for example, can
be set up with the LEYBOLD MEMBRANOVAC
of the A series. Fig. 3.26 shows different
units of the new LEYBOLD A series, which,
although they function according to diffe-
rent measuring methods, all display a uni-
form appearance.
Continuous pressure regulation
We have to make a distinction here bet-
ween electric controllers (e.g. PID con-
trollers) with a proportional valve as actua-
tor and mechanical diaphragm controllers.
In a regulation system with electric con-
trollers the coordination between control-
ler and actuator (piezoelectric gas inlet
valve, inlet valve with motor drive, butterf-
ly control valve, throttle valve) is difficult
because of the very different boundary
conditions (volume of the vessel, effective
pumping speed at the vessel, pressure
control range). Such control circuits tend
to vibrate easily when process mal-
functions occur. It is virtually impossible to
specify generally valid standard values.
Many control problems can be better sol-
ved with a diaphragm controller. The func-
tion of the diaphragm controller (see Fig.
3.27) can be easily derived from that of a
diaphragm vacuum gauge: the blunt end of
a tube or pipe is either closed off by means
of an elastic rubber diaphragm (for refe-
rence pressure > process pressure) or
released (for reference pressure < process
pressure) so that in the latter case, a
connection is established between the pro-
cess side and the vacuum pump. This ele-
gant and more or less automatic regula-
tion system has excellent control characte-
ristics (see Fig. 3.28).
To achieve higher flow rates, several dia-
phragm controllers can be connected in
parallel. This means that the process
chambers and the reference chambers are
also connected in parallel. Fig. 3.29 shows
such a connection of 3 MR 50 diaphragm
controllers.
To control a vacuum process, it is fre-
quently necessary to modify the pressure
in individual process steps. With a dia-
phragm controller this can be done either
manually or via electric control of the refe-
rence pressure.
Electric control of the reference pressure
of a diaphragm controller is relatively easy
because of the small reference volume that
always remains constant. Fig. 3.31 shows
such an arrangement on the left as a pic-
Fig. 3.26 LEYBOLD-A series, equipment with level and interval triggers Fig. 3.27 Principle of a diaphragm controller
reference pressure
adjustment valve
pump connection
control chamber
controller seat
reference chamber
diaphragm
measuring
connection
for reference
chamber
measuring
connection
for process pressure
process chamber
connection
Fig. 3.28 Control characteristics of a diaphragm controller
P
1
= process pressure, P
2
= pressure in pump, P
ref
= reference pressure
D00 E 19.06.2001 21:39 Uhr Seite 86
Vacuum Measurement
ture and on the right schematically, see
3.5.5 for application examples with
diaphragm controllers.
To be able to change the reference pressu-
re and thus the process pressure towards
higher pressures, a gas inlet valve must
additionally be installed at the process
chamber. This valve is opened by means of
a differential pressure switch (not shown
in Fig. 3.31) when the desired higher pro-
cess pressure exceeds the current process
pressure by more than the pressure diffe-
rence set on the differential pressure
switch.
3.5.4 Pressure regulation in
high and ultrahigh vacu-
um systems
If the pressure is to be kept constant wit-
hin certain limits, an equilibrium must be
established between the gas admitted to
the vacuum vessel and the gas simulta-
neously removed by the pump with the aid
of valves or throttling devices. This is not
very difficult in rough and medium vacu-
um systems because desorption of adsor-
bed gases from the walls is generally neg-
ligible in comparison to the quantity of gas
flowing through the system. Pressure
regulation can be carried out through gas
inlet or pumping speed regulation. Howe-
ver, the use of diaphragm controllers is
only possible between atmospheric pres-
sure and about 10 mbar.
In the high and ultrahigh vacuum range,
on the other hand, the gas evolution from
the vessel walls has a decisive influence on
the pressure. Setting of specific pressure
values in the high and ultrahigh vacuum
range, therefore, is only possible if the gas
evolution from the walls is negligible in
relation to the controlled admission of gas
by means of the pressure-regulating unit.
For this reason, pressure regulation in this
range is usually effected as gas admission
regulation with an electric PID controller.
Piezoelectric or servomotor-controlled
variable-leak valves are used as actuators.
Only bakeable all-metal gas inlet valves
should be used for pressure regulation
below 10
-6
mbar.
D00
Fig. 3.29 Triple connection of diaphragm controllers
Fig. 3.30 Control of vacuum drying processes by regu-
lation of the intake pressure of the vacuum
pump according to the water vapor tolerance
DC Diaphragm controller
P Vacuum pump
PS Pressure sensor
V1 Pump valve
V2 Gas inlet valve
TH Throttle
RC Reference chamber
PC Process chamber
CV Internal reference pressure control valve
Fig. 3.31 Diaphragm controller with external automatic reference pressure regulation
DC Diaphragm controller
PS Process pressure sensor
RS Reference pressure sensor
V1 Gas inlet valve
V2 Pump valve
V3 Gas inlet variable-leak valve
TH Throttle
PP Process pump
RC Reference chamber
PC Process chamber
AP Auxiliary pump
CV Internal reference pressure control valve
D00 E 19.06.2001 21:39 Uhr Seite 87
Vacuum Measurement
3.5.5 Examples of applications
with diaphragm
controllers
1) Regulation of a drying/distillation pro-
cess, taking into account the maximum
water vapor tolerance of a vane type
rotary pump
In a drying process it is frequently desira-
ble to carry out drying solely by means of
vacuum pumps without inserting conden-
sers. In view of the limited water vapor
tolerance of vacuum pumps approx.
30 mbar as a rule this would result in
condensation of the vapors produced wit-
hin the vacuum pump, given non-throttled
or non-regulated pumping speed. One can
avoid this through process-dependent
remote control of a diaphragm controller
with auxiliary control valves and a measu-
ring and switching device with a pressure
sensor at the inlet connection of the vacu-
um pump if the intake pressure is adapted
to the pumps water vapor tolerance
through automatic monitoring of the inta-
ke pressure of the vacuum pump and by
throttling the pumping speed. Fig. 3.30
shows the principle of this arrangement.
Mode of operation: Starting from atmos-
pheric pressure with the process heating
switched off, valve V1 is initially open
(maximum switching point exceeded) so
that atmospheric pressure also prevails in
the reference chamber.
The diaphragm controller is therefore clo-
sed. When the system is started up, the
connecting line between the vacuum pump
and pump valve V2 is first evacuated. As
soon as the pressure drops below the
maximum switching point, valve V1 clo-
ses. When the pressure falls below the
minimum switching point, valve V2 opens.
In this manner the pressure in the referen-
ce chamber is slowly lowered, the thrott-
ling of the diaphragm controller is reduced
accordingly and thus the process pressure
is lowered until the quantity of process gas
is greater than the quantity conveyed by
the pump so that the minimum switching
point is again exceeded. Valve V2 closes
again. This interaction repeats itself until
the pressure in the process chamber has
dropped below the minimum switching
point. After that, valve V2 remains open so
that the process can be brought down to
the required final pressure with a comple-
tely open diaphragm controller.
The material to be dried is usually heated
to intensify and speed up the drying pro-
cess. If a certain amount of water vapor is
produced, the intake pressure rises above
the two switching points. As a result, valve
V2 first closes and V1 opens. Through
incoming air or protective gas the pressu-
re in the reference chamber is raised and
the throughput at the diaphragm controller
thus throttled until the intake pressure of
the vacuum pump has dropped below the
set maximum switching point again. Then
valve V1 closes.
Depending on the quantity of vapor that
accumulates, the throughput of the
diaphragm controller is set by increasing
or decreasing the reference pressure in
each case so that the maximum permissi-
ble partial water vapor pressure at the inlet
connection of the vacuum pump is never
exceeded.
As soon as the pressure in the process
chamber drops below the set minimum
switching point towards the end of the dry-
ing process, valve V2 opens and remains
open. In this way the unthrottled cross-
section of the diaphragm controller is avai-
lable again for rapid final drying. At the
same time the final drying procedure can
be monitored by means of the pressure
sensor PS.
2) Pressure regulation by means of dia-
phragm controller with external
automatic reference pressure adjust-
ment (see Fig. 3.31)
For automatic vacuum processes with
regulated process pressure, presetting of
the desired set pressure must often func-
tion and be monitored automatically. If a
diaphragm controller is used, this can be
done by equipping the reference chamber
with a measuring and switching device and
a control valve block at the reference
chamber. The principle of this arrange-
ment is shown in Fig. 3.31.
Mode of operation: Starting with atmos-
pheric pressure, gas inlet valve V1 is clo-
sed at the beginning of the process. Pump
valve V2 opens. The process chamber is
now evacuated until the set pressure,
which is preset at the measuring and swit-
ching device, is reached in the process
chamber and in the reference chamber.
When the pressure falls below the set swit-
ching threshold, pump valve V2 closes. As
a result, the pressure value attained is
caught as the reference pressure in the
reference chamber (RC) of the diaphragm
controller (DC). Now the process pressure
is automatically maintained at a constant
level according to the set reference pres-
sure by means of the diaphragm controller
(DC). If the reference pressure should rise
in the course of the process due to a leak,
this is automatically detected by the mea-
suring and switching device and corrected
by briefly opening pump valve V2. This
additional control function enhances the
operational reliability and extends the
range of application. Correcting the increa-
sed reference pressure to the originally set
value is of special interest for regulated
helium circuits because the pressure rise
in the reference chamber (RC) of the dia-
phragm controller can be compensated for
through this arrangement as a consequen-
ce of the unavoidable helium permeability
of the controller diaphragm of FPM.
To be able to change the reference pressu-
re and thus increase the process pressure
to higher pressures, a gas inlet valve must
be additionally installed at the process
chamber. This valve is opened by means of
a differential pressure switch (not shown
in Fig. 3.31) when the desired higher pro-
cess pressure exceeds the current process
pressure by more than the pressure diffe-
rential set at the differential pressure
switch.
D00 E 19.06.2001 21:39 Uhr Seite 88
D00
Mass Spectrometry
4. Analysis of gas
at low
pressures
using mass
spectrometry
4.1 General
Analyses of gases at low pressures are
useful not only when analyzing the residu-
al gases from a vacuum pump, leak testing
at a flange connection or for supply lines
(compressed air, water) in a vacuum. They
are also essential in the broader fields of
vacuum technology applications and pro-
cesses. For example in the analysis of pro-
cess gases used in applying thin layers of
coatings to substrates. The equipment
used for qualitative and/or quantitative
analyses of gases includes specially deve-
loped mass spectrometers with extremely
small dimensions which, like any other
vacuum gauge, can be connected directly
to the vacuum system. Their size distin-
guishes these measurement instruments
from other mass spectrometers such as
those used for the chemical analyses of
gases. The latter devices are poorly suited,
for example, for use as partial pressure
measurement units since they are too
large, require a long connector line to the
vacuum chamber and cannot be baked out
with the vacuum chamber itself. The in-
vestment for an analytical mass spectro-
meter would be unjustifiably great since,
for example, the requirements as to reso-
lution are far less stringent for partial pres-
sure measurements. Partial pressure is
understood to be that pressure exerted by
a certain type of gas within a mix of gases.
The total of the partial pressures for all the
types of gas gives the total pressure. The
distinction among the various types of
gases is essentially on the basis of their
molar masses. The primary purpose of
analysis is therefore to register qualita-
tively the proportions of gas within a
system as regards the molar masses and
determine quantitatively the amount of the
individual types of gases associated with
the various atomic numbers.
Partial pressure measurement devices
which are in common use comprise the
measurement system proper (the sensor)
and the control device required for its ope-
ration. The sensor contains the ion source,
the separation system and the ion trap.
The separation of ions differing in masses
and charges is often effected by utilizing
phenomena which cause the ions to reso-
nate in electrical and magnetic fields.
Initially, the control units were quite clum-
sy and offered uncountable manipulation
options. It was often the case that only
physicists were able to handle and use
them. With the introduction of PCs the
requirements in regard to the control units
became ever greater. At first, they were fit-
ted with interfaces for linkage to the com-
puter. Attempts were made later to equip a
PC with an additional measurement circuit
board for sensor operation. Todays sen-
sors are in fact transmitters equipped with
an electrical power supply unit attached
direct at the atmosphere side; communica-
tion with a PC from that point is via the
standard computer ports (RS 232, RS
485). Operating convenience is achieved
by the software which runs on the PC.
4.2 A historical review
Following Thomsons first attempt in 1897
to determine the ratio of charge to mass
e/m for the electron, it was quite some
time (into the 1950s) before a large num-
ber and variety of analysis systems came
into use in vacuum technology. These
included the Omegatron, the Topatron and
ultimately the quadrupole mass spectro-
meter proposed by Paul and Steinwedel in
1958, available from INFICON in its stan-
dard version as the TRANSPECTOR (see
Fig. 4.1). The first uses of mass spectro-
metry in vacuum-assisted process techno-
logy applications presumably date back to
Backus work in the years 1943 / 44. He
carried out studies at the Radiographic
Laboratories at the University of California.
Seeking to separate uranium isotopes, he
used a 180 sector field spectrometer after
Dempster (1918), which he referred to as
a vacuum analyzer. Even today a similar
term, namely the residual gas analyzer
(RGA), is frequently used in the U.S.A. and
the U.K. instead of mass spectrometer.
Todays applications in process monitoring
are found above all in the production of
semiconductor components.
4.3 The quadrupole
mass spectrometer
(TRANSPECTOR)
The ion beam extracted from the electron
impact ion source is diverted into a qua-
drupole separation system containing four
rod-shaped electrodes. The cross sections
of the four rods form the circle of curvatu-
re for a hyperbola so that the surrounding
electrical field is nearly hyperbolic. Each of
the two opposing rods exhibits equal
potential, this being a DC voltage and a
superimposed high-frequency AC voltage
a: High-performance sensor with Channeltron
b: Compact sensor with Micro-Channelplate
c: High-performance sensor with Faraday cup
a
b
c
Fig. 4.1a TRANSPECTOR sensors
Fig. 4.1b TRANSPECTOR XPR sensor
D00 E 19.06.2001 21:39 Uhr Seite 89
Mass Spectrometry
(Fig. 4.2). The voltages applied induce
transverse oscillations in the ions traver-
sing the center, between the rods. The
amplitudes of almost all oscillations
escalate so that ultimately the ions will
make contact with the rods; only in the
case of ions with a certain ratio of mass to
charge m/e is the resonance condition
which allows passage through the system
satisfied. Once they have escaped from the
separation system the ions move to the ion
trap (detector, a Faraday cup) which may
also take the form of a secondary electron
multiplier pick-up (SEMP).
The length of the sensor and the sepa-
ration system is about 15 cm. To ensure
that the ions can travel unhindered from
the ion source to the ion trap, the mean
free path length inside the sensor must be
considerably greater than 15 cm. For air
and nitrogen, the value is about
p = 6 10
3
mbar cm. At p = 1 10
-4
bar this corresponds to a mean free path
length of = 60 cm. This pressure is gene-
rally taken to be the minimum vacuum for
mass spectrometers. The emergency shut-
down feature for the cathode (responding
to excessive pressure) is almost always
set for about 5 10
-4
mbar. The desire to
be able to use quadrupole spectrometers
at higher pressures too, without special
pressure convertors, led to the develop-
ment of the XPR sensor at Inficon (XPR
standing for extended pressure range). To
enable direct measurement in the range of
about 2 10
-2
mbar, so important for sput-
ter processes, the rod system was reduced
from 12 cm to a length of 2 cm. To ensure
that the ions can execute the number of
transverse oscillations required for sharp
mass separation, this number being about
100, the frequency of the current in the
XPR sensor had to be raised from about 2
MHz to approximately 6 times that value,
namely to 13 MHz. In spite of the reduction
in the length of the rod system, ion yield is
still reduced due to dispersion processes
at such high pressures. Additional electro-
nic correction is required to achieve per-
fect depiction of the spectrum. The dimen-
sions of the XPR sensor are so small that
it can hide entirely inside the tubulation
of the connection flange (DN 40, CF) and
thus occupies no space in the vacuum
chamber proper. Fig. 4.1a shows the size
comparison for the normal high-perfor-
mance sensors with and without the Chan-
neltron SEMP, the normal sensor with
channel-plate SEMP. Fig. 4.1b shows the
XPR sensor. The high vacuum required for
the sensor is often generated with a
TURBOVAC 50 turbomolecular pump and
a D 1.6 B rotary vane pump. With its great
compression capacity, a further advantage
of the turbomolecular pump when
handling high molar mass gases is that the
sensor and its cathode are ideally pro-
tected from contamination from the
direction of the forepump.
4.3.1 Design of the sensor
One can think of the sensor as having been
derived from an extractor measurement
system (see Fig. 4.3), whereby the
separation system was inserted between
the ion source and the ion trap.
4.3.1.1 The normal (open) ion source
The ion source comprises an arrangement
of the cathode, anode and several baffles.
The electron emission, kept constant, cau-
ses partial ionization of the residual gas,
into which the ion source is immersed as
completely as possible. The vacuum in the
vicinity of the sensor will naturally be influ-
enced by baking the walls or the cathode.
The ions will be extracted through the baf-
fles along the direction of the separation
system. One of the baffles is connected to
a separate amplifier and entirely inde-
pendent of ion separation provides con-
tinuous total pressure measurement (see
Fig. 4.4). The cathodes are made of iridium
wire and have a thorium oxide coating to
reduce the work associated with electron
discharge. (For some time now the thori-
um oxide has gradually been replaced by
yttrium oxide.) These coatings reduce the
electron discharge work function so that
the desired emission flow will be achieved
even at lower cathode temperatures. Avai-
lable for special applications are tungsten
cathodes (insensitive to hydrocarbons but
sensitive to oxygen) or rhenium cathodes
(insensitive to oxygen and hydrocarbons
but evaporate slowly during operation due
to the high vapor pressure).
Shielding
Cathode
Anode
Focussing plate
(extractor diaphragm) Ion source exit
diaphragm
(total pressure measurement)
Quadrupole exit
diaphragm
Ion detector Quadrupole separation system Ion source
Fig. 4.2 Schematic for quadrupole mass spectrometer Fig. 4.3 Quadrupole mass spectrometer Extractor ionization vacuum gauge
Extractor measurement system
Transpector measurement head
Anode
Cathode
Ref l ec-
tor
Ion trap
Fig. 4.4 Open ion source
Cathode
Shielding
Extractor
diaphragm
Total pressure
diaphragm
Anode
D00 E 19.06.2001 21:39 Uhr Seite 90
Mass Spectrometry
4.3.1.2 The quadrupole separation
system
Here the ions are separated on the basis of
their mass-to-charge ratio. We know from
physics that the deflection of electrically
charged particles (ions) from their trajec-
tory is possible only in accordance with
their ratio of mass to charge, since the
attraction of the particles is proportional to
the charge while the inertia (which resists
change) is proportional to its mass. The
separation system comprises four cylin-
drical metal rods, set up in parallel and iso-
lated one from the other; the two opposing
rods are charged with identical potential.
Fig. 4.2 shows schematically the arrange-
ment of the rods and their power supply.
The electrical field inside the separation
system is generated by superimposing a
DC voltage and a high-frequency AC volta-
ge:
= (U + V cos t) (x
2
y
2
) / r
0
2
r
0
= radius of the cylinder which can be
inscribed inside the system of rods
Exerting an effect on a single charged ion
moving near and parallel to the center line
inside the separation system and perpen-
dicular to its movement are the forces
The mathematical treatment of these equa-
tions of motion uses Mathieus differential
equations. It is demonstrated that there
are stable and unstable ion paths. With the
stable paths, the distance of the ions from
the separation system center line always
remains less than r
o
(passage condition).
With unstable paths, the distance from the
axis will grow until the ion ultimately colli-
des with a rod surface. The ion will be
discharged (neutralized), thus becoming
unavailable to the detector (blocking con-
dition).
Even without solving the differential equa-
tion, it is possible to arrive at a purely phe-
nomenological explanation which leads to
an understanding of the most important
characteristics of the quadrupole sepa-
ration system.
If we imagine that we cut open the se-
F
e
r
x t
x
=
2
0
2
cos ( )
F
e
r
y t
y
=
2
0
2
cos ( )
F
z
= 0
paration system and observe the deflec-
tion of a singly ionized, positive ion with
atomic number M, moving in two planes,
which are perpendicular one to the other
and each passing through the centers of
two opposing rods. We proceed step-by-
step and first observe the xz plane (Fig.
4.5, left) and then the yz plane (Fig.4.5,
right):
1. Only DC potential U at the rods:
xz plane (left): Positive potential of +U
at the rod, with a repellant effect on the
ion, keeping it centered; it reaches the
collector (passage).
yz plane (right): Negative potential on
the rod -U, meaning that at even the
tiniest deviations from the center axis
the ion will be drawn toward the nearest
rod and neutralized there; it does not
reach the collector (blocking).
2. Superimposition of high-frequency
voltage V cos t:
xz plane (left):
Rod potential +U + V cos t. With
rising AC voltage amplitude V the ion
will be excited to execute transverse
oscillations with ever greater amplitu-
des until it makes contact with a rod
and is neutralized. The separation
system remains blocked for very large
values of V.
yz plane (right):
Rod potential -U -V cos t. Here again
superimposition induces an additional
force so that as of a certain value for V
the amplitude of the transverse oscilla-
tions will be smaller than the clearance
between the rods and the ion can pass
to the collector at very large V.
3. Ion emission i
+
= i
+
(V) for a fixed mass
of M:
xz plane (left): For voltages of V < V
1
the
deflection which leads to an escalation
of the oscillations is smaller than V
1
, i.e.
still in the pass range. Where V > V
1
the deflection will be sufficient to
induce escalation and thus blockage.
yz plane (right): For voltages of V < V
1
the deflection which leads to the dam-
ping of the oscillations is smaller than
V
1
, i.e. still in the block range. Where
V > V
1
the damping will be sufficient to
settle oscillations, allowing passage.
4. Ion flow i
+
= i
+
(M) for a fixed ratio of
U / V:
Here the relationships are exactly oppo-
site to those for i
+
= i
+
(V) since the
influence of V on light masses is great-
er than on heavy masses.
xz plane: For masses of M < M
1
the
deflection which results in escalation of
the oscillations is greater than at M
1
,
which means that the ions will be
blocked. At M > M
1
the deflection is no
longer sufficient for escalation, so that
the ion can pass.
yz plane: For masses of M < M
1
the
deflection which results in damping of
the oscillations is greater than at M
1
,
which means that the ion will pass. At
M > M
1
the damping is not sufficient to
calm the system and so the ion is
blocked.
5. Combination of the xz and yz planes. In
the superimposition of the ion currents
i
+
= i
+
(M) for both pairs of rods (U / V
being fixed) there are three important
ranges:
Range I : No passage for M due to the
blocking behavior of the xz pair of rods.
Range II : The pass factor of the rod
systems for mass M is determined by
the U/V ratio (other ions will not pass).
We see that great permeability (corre-
sponding to high sensitivity) is bought
at the price of low selectivity (= resolu-
tion, see Section 4.5). Ideal adjustment
of the separation system thus requires
a compromise between these two prop-
D00
xz plane yz plane
+
+
+
+
+
+
-
-
+
-
-
+
Superimposition of the xy and yz planes
Rod:
+U
Transmission:
full
Rod:
+U+V, cos
Transmission:
low-pass
i
+
i
+
i
+
V V1 V1 V
M
i
+
M M1 M1
i
+
M
I III
II
yz xz
Rod:
U
Transmission:
none
Rod:
UV cos
Transmission:
high-pass
( )
U
V
fixed
U
V
.. Selectivity (resolution) Sensitivity
1
2
3
4
5
Fig. 4.5 Phenomenological explanation of the
separation system
D00 E 19.06.2001 21:39 Uhr Seite 91
Mass Spectrometry
erties. To achieve constant resolution,
the U/V ratio will remain constant over
the entire measurement range. The
atomic number M (see 4.6.1) of the
ions which can pass through the sepa-
ration system must satisfy this condi-
tion:
V = High-frequency amplitude,
r
O
= Quadrupole inscribed radius
f = High-frequency
As a result of this linear dependency
there results a mass spectrum with li-
near mass scale due to simultaneous,
proportional modification of U and V.
Range III : M cannot pass, due to the
blocking characteristics of the yz pair of
rods.
4.3.1.3 The measurement system
(detector)
Once they have left the separation system
the ions will meet the ion trap or detector
which, in the simplest instance, will be in
the form of a Faraday cage (Faraday cup).
In any case the ions which impinge on the
detector will be neutralized by electrons
from the ion trap. Shown, after electrical
amplification, as the measurement signal
itself is the corresponding ion emission
stream. To achieve greater sensitivity, a
secondary electron multiplier pickup
(SEMP) can be employed in place of the
Faraday cup.
m
e
M
V
f r
o

14.438
2 2
Channeltrons or Channelplates can be
used as SEMPs. SEMPs are virtually iner-
tia-free amplifiers with gain of about 10
+6
at the outset; this will indeed drop off
during the initial use phase but will then
become virtually constant over a long peri-
od of time. Fig. 4.6 shows at the left the
basic configuration of a Faraday ion trap
and, on the right, a section through a
Channeltron. When recording spectra the
scanning period per mass line t
0
and the
time constants of the amplifier t should
satisfy the condition that t
0
= 10 . In
modern devices such as the
TRANSPECTOR the otherwise unlimited
selection of the scanning period and the
amplifier time constants will be restricted
by microprocessor control to logical pairs
of values.
Amplifier
Amplifier
Connection
to front end
of the inside
surface
Collector
Positive ion
Negative high voltage
Resistance 10
8
Resistance of the inner surface

R 4 10
6
Separation system output

Electron suppressor
Faraday cup
Fig. 4.6 Left: principle of the Faraday cup; right: configuration of the Channeltron
S
S compensates for L No segregation
eff 2
Mass
spectrometer
L molecular d
2 L

2
S L S ~
eff 1 eff

L
2
~
1
M
L
1
~
1
1
M
M
Non-segregating gas inlet system
Stage B Stage A
p=1... 10mbar p=10... 1000mbar p 10 mbar
4
Capillary
No
segregation
Laminar flow Molecular flow
L
2
L
3
L
1
S
eff Q
Pumping
Q Q
HV Pumping
Q
HV
(Transition laminar/molecular)
Fig. 4.7 Principle of the pressure converter (stage B only in the single-stage version and stages A and B in
two-stage units)
D00 E 19.06.2001 21:39 Uhr Seite 92
Mass Spectrometry
4.4 Gas admission and
pressure
adaptation
4.4.1 Metering valve
The simplest way to adapt a classical mass
spectrometer to pressures exceeding
1 10
-4
mbar is by way of a metering
valve. The inherent disadvantage is, howe-
ver, that since the flow properties are not
unequivocally defined, a deviation from the
original gas composition might result.
4.4.2 Pressure converter
In order to examine a gas mix at total pres-
sure exceeding 1 10
-4
mbar it is neces-
sary to use pressure converters which will
not segregate the gases. Figure 4.7 is used
to help explain how such a pressure con-
verter works:
a. Process pressure < 1 mbar: Single-
stage pressure converter.
Gas is allowed to pass out of the vacuum
vessel in molecular flow, through a dia-
phragm with conductance value L
2
and
into the sensor chamber (with its own
high vacuum system). Molecular flow cau-
ses segregation but this will be indepen-
dent of the pressure level (see Section
1.5). A second diaphragm with molecular
flow, located between the sensor chamber
and the turbomolecular pump, will com-
pensate for the segregation occurring at
L
2
.
b. Process pressure > 1 mbar: Two-stage
pressure converter. Using a small (rotary
vane) pump a laminar stream of gas is
diverted from the rough vacuum area
through a capillary or diaphragm (conduc-
tance value L
3
). Prior to entry into the
pump, at a pressure of about 1 mbar, a
small part of this flow is again allowed to
enter the sensor chamber through the dia-
phragm with conductance value L
2
, again
as molecular flow.
A falsification of the gas composition
resulting from adsorption and condensati-
on can be avoided by heating the pressure
converter and the capillary.
To evaluate the influence on the gas com-
position by the measurement unit itself,
information on the heating temperature,
the materials and surface areas for the
metallic, glass and ceramic components
will be needed along with specifications on
the material and dimensions of the catho-
de (and ultimately regarding the electron
impact energy for the ion source as well).
4.4.3 Closed ion source (CIS)
In order to curb or avoid entirely influ-
ences which could stem from the sensor
chamber or the cathode (e.g. disturbance
of the CO-CO
2
equilibrium by heating the
cathode) a closed ion source (CIS) will be
used in many cases (see Fig 4.8).
D00
Process: 10
-2
mbar Process: 10
-2
mbar
Impact chamber
Cathode chamber
Exit diaphragm
10
-5
10
-3
10
-5
10
-5
Exit diaphragm
Pump Pump
10
-5
10
-5
10
-5
10
-5
(Elastomer) (Metal) Valve
Inlet diaphragm
Example of the sputter
process
To be detected is 1 ppm N as
contamination in argon,
the working gas
2
1 ppm N
2
at the inlet:
10
-6
10
-5
mbar = 10
-11
mbar
Background:
Residual gas (valve closed) 10
-6
mbar total
total, including 1% by mass 28 : 10
-8
mbar
Background noise
Signal at 1 of background
cannot be detected
1 ppm N
2
at the inlet:
10
-6
10
-3
mbar = 10
-9
mbar
Background:
Residual gas (valve closed) 10
-7
mbar total
total, including 1% by mass 28 : 10
-9
mbar
1 ppm
Background noise
Signal twice the background noise
amplitude; can just be clearly detected
Fig. 4.8 Open ion source (left) and closed ion source (right)
D00 E 19.06.2001 21:39 Uhr Seite 93
Mass Spectrometry
The CIS is divided into two sections: a
cathode chamber where the electrons are
emitted, and an impact chamber, where
the impact ionization of the gas particles
takes place. The two chambers are pum-
ped differentially: the pressure in the
cathode chamber comes to about
10
-5
mbar, that in the impact room about
10
-3
mbar. The gas from the vacuum
chamber is allowed to pass into the impact
chamber by way of a metal-sealed, bakea-
ble valve (pressure converter, ultrahigh
vacuum technology). There high-yield
ionization takes place at about 10
-3
mbar.
The electrons exerting the impact are emit-
ted in the cathode chamber at about
10
-5
mbar and pass through small ope-
nings from there into the impact chamber.
The signal-to-noise ratio (residual gas) via
vis the open ion source will be increased
overall by a factor of 10
+3
or more. Figure
4.8 shows the fundamental difference bet-
ween the configurations for open and clo-
sed ion sources for a typical application in
sputter technology. With the modified
design of the CIS compared with the open
ion source in regard to both the geometry
and the electron energy (open ion source
102 eV, CIS 75 or 35 eV), different frag-
ment distribution patterns may be found
where a lower electron energy level is sel-
ected. For example, the argon36
++
isotope
at mass of 18 cannot be detected at elec-
tron energy of less than 43.5 eV and can
therefore not falsify the detection of H
2
O
+
at mass 18 in the sputter processes using
argon as the working gas processes
which are of great importance in industry.
4.4.4 Aggressive gas monitor
(AGM)
In many cases the process gas to be
examined is so aggressive that the catho-
de would survive for only a short period of
time. The AGM uses the property of lami-
nar flow by way of which there is no
reverse flow of any kind. Controlled with
a separate AGM valve, a part of the wor-
king gas fed to the processes is introduced
as purging gas, ahead of the pressure
converter, to the TRANSPECTOR; this sets
up a flow toward the vacuum chamber.
Thus process gas can reach the TRANS-
PECTOR only with the AGM valve closed.
When the valve is open the TRANSPEC-
TOR sees only pure working gas. Fig. 4.9
shows the AGM principle.
4.5 Descriptive values
in mass
spectrometry
(specifications)
A partial pressure measurement unit is
characterized essentially by the following
properties (DIN 28 410):
4.5.1 Line width (resolution)
The line width is a measure of the degree
to which differentiation can be made bet-
ween two adjacent lines of the same
height. The resolution is normally indica-
ted. It is defined as: R = M / M and is con-
stant for the quadrupole spectrometer
across the entire mass range, slightly
greater than 1 or M < 1.
Often an expression such as unit resoluti-
on with 15% valley is used. This means
that the bottom of the valley between
two adjacent peaks of identical height
comes to 15 % of the height of the peak or,
put another way, at 7.5 % of its peak
height the line width DM measured across
an individual peak equals 1 amu (atomic
mass unit); see in this context the sche-
matic drawing in Fig. 4.10.
Process:
e. g. 50mbar
Impact chamber
Cathode chamber
Exit diaphragms
10
-5
10
-3
10
-5
10
-5
Pump Diaphragm
Working gas for the process (Ar)
AGM protective gas valve
Fig. 4.9 Principle behind the aggressive gas monitor (AGM)
D00 E 19.06.2001 21:39 Uhr Seite 94
Mass Spectrometry
4.5.2 Mass range
The mass range is characterized by the
atomic numbers for the lightest and hea-
viest ions with a single charge which are
detected with the unit.
4.5.3 Sensitivity
Sensitivity E is the quotient of the measu-
red ion flow and the associated partial
pressure; it is normally specified for argon
or nitrogen:
(4.1)
Typical values are:
Faraday cup:
SEM:
partial pressure
The smallest detectable partial pressure is
defined as a ratio of noise amplitude to
sensitivity:
i
+
R
= Noise amplitude
Example (from Fig. 4.11):
Sensitivity E =
Noise amplitude i
+
R
= 4 10
14
A
1 10
4
A
mbar
P
i
E
mbar
R
min
( ) =
+
E
A
mbar
=
+
1 10
2
E
A
mbar
=
1 10
4
E
i
p
A
mbar
G
=

_
,
+
partial pressure ratio
(concentration)
The definition is:
SDPPR = p
min
/ p
(ppm)
This definition, which is somewhat clum-
sy for practical use, is to be explained
using the detection of argon36 in the air as
the example: Air contains 0.93 % argon by
volume; the relative isotope frequency of
Ar
40
to Ar
36
is 99.6 % to 0.337 %. Thus
the share of Ar
36
in the air can be calcula-
ted as follows:
0.93 10
2
0.337 10
2
= 3.13 10
5
=
31.3 ppm
Figure 4.11 shows the screen print-out for
the measurement. The peak height for Ar
36
in the illustration is determined to be
1.5 10
-13
A and noise amplitude i
+
R
to
be 4 10
-14
A. The minimum detectable
concentration is that concentration at which
the height of the peak is equal to the noise
amplitude. This results in the smallest mea-
surable peak height being
1.5 10
-13
A/2.4 10
-14
A = 1.875. The
smallest detectable concentration is then
derived from this by calculation to arrive at:
31.3 10
6
/ 1.875 = 16.69 10
6
= 16.69 ppm.
4.5.6 Linearity range
The linearity range is that pressure range
for the reference gas (N
2
, Ar) in which sen-
sitivity remains constant within limits
p FC
A
A mbar
mbar
min
( )
/
=

4 10
1 10
4 10
14
4
10
which are to be specified ( 10 % for par-
tial pressure measurement devices).
In the range below 1 10
-6
mbar the rela-
tionship between the ion flow and partial
pressure is strictly linear. Between 1 10
-6
mbar and 1 10
-4
mbar there are minor
deviations from linear characteristics.
Above 1 10
-4
mbar these deviations grow
until, ultimately, in a range above 10
-2
mbar the ions for the dense gas atmos-
phere will no longer be able to reach the
ion trap. The emergency shut-down for the
cathode (at excessive pressure) is almost
always set for 5 10
-4
mbar. Depending on
the information required, there will be dif-
fering upper limits for use.
In analytical applications, 1 10
-6
mbar
should not be exceeded if at all possible.
The range from 1 10
-6
mbar to 1 10
-4
mbar is still suitable for clear depictions of
the gas composition and partial pressure
regulation (see Fig. 4.12).
4.5.7 Information on surfaces
and amenability to
bake-out
Additional information required to evaluate
a sensor includes specifications on the
bake-out temperature (during measure-
ment or with the cathode or SEMP swit-
ched off), materials used and surface
areas of the metal, glass and ceramic com-
ponents and the material and dimensions
for the cathode; data is also needed on the
electron impact energy at the ion source
(and on whether it is adjustable). These
values are critical to uninterrupted operati-
on and to any influence on the gas compo-
sition by the sensor itself.
D00
M
M
1 amu
M + 1 Atomic number
7,5%
15%
i
+
100%
Fig. 4.10 Line width 15 % valley Fig. 4.11 Detection of Argon
36
10
8
10
7
10
6
10
5
10
4
10
3 log P
Automatic shut-down:
5 10
4
log i
+
Exact measurement
range
Regulation
Fig. 4.12 Qualitative linearity curve
D00 E 19.06.2001 21:39 Uhr Seite 95
Mass Spectrometry
4.6 Evaluating spectra
4.6.1 Ionization and fundamen-
tal problems in gas ana-
lysis
Continuous change in the voltages applied
to the electrodes in the separation system
(scanning) gives rise to a relationship
between the ion flow I
+
and the atomic
number which is proportional to the m/e
ratio and expressed as:
(4.2)
(M
r
= relative molar mass,
n
e
= number of elementary charges e)
This is the so-called mass spectrum,
i
+
= i
+
(M). The spectrum thus shows the
peaks i
+
as ordinates, plotted against the
atomic number M along the abscissa. One
of the difficulties in interpreting a mass
spectrum such as this is due to the fact
that one and the same mass as per the
equation (4.2) may be associated with
various ions. Typical examples, among
many others, are: The atomic number
M = 16 corresponds to CH
4
+
and O
2
++
;
M = 28 for CO
+
, N
2
+
and C2H
+
! Particular
attention must therefore be paid to the fol-
lowing points when evaluating spectra:
1) In the case of isotopes we are dealing
with differing positron counts in the nucle-
us (mass) of the ion at identical nuclear
charge numbers (gas type). Some values
for relative isotope frequency are compiled
in Table 4.2.
2) Depending on the energy of the impac-
ting electrons (equalling the potential diffe-
rential, cathode anode), ions may be eit-
her singly or multiply ionized. For example,
one finds Ar
+
at mass of 40, Ar
++
at mass
of 20 and Ar
+++
at mass of 13.3. At mass
of 20 one will, however, also find neon,
Ne
+
. There are threshold energy levels for
the impacting electrons for all ionization
states for every type of gas, i.e., each type
of ion can be formed only above the asso-
ciated energy threshold. This is shown for
Ar in Fig. 4.13.
3) Specific ionization of the various
gases S
gas
, this being the number of ions
formed, per cm and mbar, by collisions
with electrons; this will vary from one type
of gas to the next. For most gases the ion
M
M
n
r
e
=
yield is greatest at an electron energy level
between about 80 and 110 eV; see Fig.
4.14.
In practice the differing ionization rates for
the individual gases will be taken into
account by standardization against nitro-
gen; relative ionization probabilities
(RIP) in relationship to nitrogen will be
indicated (Table 4.3).
4) Finally, gas molecules are often broken
down into fragments by ionization. The
fragment distribution patterns thus crea-
ted are the so-called characteristic spectra
(fingerprint, cracking pattern).
Important: In the tables the individual frag-
ments specified are standardized either
against the maximum peak (in % or of
the highest peak) or against the total of all
peaks (see the examples in Table 4.4).
Element Ordinal- Atomic Relative
number number frequency
H 1 1 99.985
2 0.015
He 2 3 0.00013
4 100.0
B 5 10 19.78
11 80.22
C 6 12 98.892
13 1.108
N 7 14 99.63
15 0.37
O 8 16 99.759
17 0.0374
18 0.2039
F 9 19 100.0
Ne 10 20 90.92
21 0.257
22 8.82
Na 11 23 100.0
Al 13 27 100.0
Si 14 28 92.27
29 4.68
30 3.05
P 15 31 100.0
Table 4.2 Relative frequency of isotopes
0
2
4
6
8
10
12
100 200 300 400 500
Ar
+
Electron energy (eV)
I
o
n
s

f
o
r
m
e
d

p
e
r

c
m

m
b
a
r
Fig. 4.13 Number of the various Ar ions produced, as a factor of electron energy level
Ar
4+
200 eV
Element Ordinal- Atomic Relative
number number frequency
S 16 32 95.06
33 0.74
34 4.18
36 0.016
Cl 17 35 75.4
37 24.6
Ar 18 36 0.337
38 0.063
40 99.60
Kr 36 78 0.354
80 2.27
82 11.56
83 11.55
84 56.90
86 17.37
Xe 54 124 0.096
126 0.090
128 1.919
129 26.44
130 4.08
131 21.18
132 26.89
134 10.44
136 8.87
Table 4.2 Relative frequency of isotopes
Threshold
energy
for argon ions
Ar
+
15,7 eV Ar
++
43,5 eV Ar
3+
85,0 eV
D00 E 19.06.2001 21:39 Uhr Seite 96
Mass Spectrometry
Both the nature of the fragments created
and the possibility for multiple ionization
will depend on the geometry (differing ion
number, depending on the length of the
ionization path) and on the energy of the
impacting electrons (threshold energy for
certain types of ions). Table values are
always referenced to a certain ion source
with a certain electron energy level. This is
why it is difficult to compare the results
obtained using devices made by different
manufacturers.
Often the probable partial pressure for one
of the masses involved will be estimated
through critical analysis of the spectrum.
Thus the presence of air in the vacuum
vessel (which may indicate a leak) is mani-
fested by the detection of a quantity of O
2
+
(with mass of 32) which is about one-
quarter of the share of N
2
+
with its mass of
28. If, on the other hand, no oxygen is
detected in the spectrum, then the peak at
atomic number 28 would indicate carbon
monoxide. In so far as the peak at atomic
number 28 reflects the CO
+
fragment of
CO
2
(atomic number 44), this share is
11 % of the value measured for atomic
number 44 (Table 4.5). On the other hand,
in all cases where nitrogen is present, ato-
mic number 14 (N
2
++
) will always be found
in the spectrum in addition to the atomic
number 28 (N
2
+
); in the case of carbon
monoxide, on the other hand, there will
always appear in addition to CO
+
the
fragmentary masses of 12 (C
+
) and
16 (O
2
++
)).
Figure 4.15 uses a simplified example of a
model spectrum with superimpositions
of hydrogen, nitrogen, oxygen, water
vapor, carbon monoxide, carbon dioxide,
neon and argon to demonstrate the diffi-
culties involved in evaluating spectra.
D00
0 5 10 15 20 25 30 35 40 45 50
Hydrogen
Nitrogen
Oxygen
Water
Carbon dioxide
Neon
Argon
Carbon monoxide
H
+
H
+
H
2
+
H
2
+
Ne
++
C
+
C
+
13
C
+
N
+
O
+
O
+
O
+
O
+
H
3
O
+
OH
+
H
2
O
+
Ne
+
Ar
++
22
Ne
+
N
2
+
CO
+
CO
+
13
CO
+
14
N
15
N
+
O
2
+
16
O
18
O
+ 36
Ar
+
Ar
+
CO
2
+
13
C
16
O
2
+
Fig. 4.15 Model spectrum
Evaluation problems: The peak at atomic number 16 may, for example, be due to oxygen fragments resulting from O
2
, H
2
O, CO
2
and CO; the peak at ato-
mic number 28 from contributions by N
2
as well as by CO and CO as a fragment of CO
2
; the peak at atomic number 20 could result from singly ionized
Ne and double-ionized Ar
Fig. 4.14 Specific ionization S for various gases by electrons exhibiting energy level E
Electron energy (eV)
I
o
n
s

f
o
r
m
e
d

p
e
r

c
m

m
b
a
r
D00 E 19.06.2001 21:39 Uhr Seite 97
Mass Spectrometry
Type of gas Symbol RIP Type of gas Symbol RIP
Acetone (Propanone) (CH
3
)
2
CO 3.6 Hydrogen chloride HCl 1.6
Air 1.0 Hydrogen fluoride HF 1.4
Ammonia NH
3
1.3 Hydrogen iodide HI 3.1
Argon Ar 1.2 Hydrogen sulfide H
2
S 2.2
Benzene C
6
H
6
5.9 Iodine I
2
Benzoic acid C
6
H
5
COOH 5.5 Krypton Kr 1.7
Bromine Br 3.8 Lithium Li 1.9
Butane C
4
H
10
4.9 Methane CH
4
1.6
Carbon dioxide CO
2
1.4 Methanol CH
3
OH 1.8
Carbon disulfide CS
2
4.8 Neon Ne 0.23
Carbon monoxide CO 1.05 Nitrogen N
2
1.0
Carbon tetrachloride CCl
4
6.0 Nitrogen oxide NO 1.2
Chlorobenzene C
6
H
4
Cl 7.0 Nitrogen dioxide N
2
O 1.7
Chloroethane C
2
H
3
Cl 4.0 Oxygen O
2
1.0
Chloroform CHCl
3
4.8 n-pentane C
5
H
17
6.0
Chlormethane CH
3
Cl 3.1 Phenol C
6
H
5
OH 6.2
Cyclohexene C
6
H
12
6.4 Phosphine PH
3
2.6
Deuterium D
2
0.35 Propane C
3
H
8
3.7
Dichlorodifluoromethane CCl
2
F
2
2.7 Silver perchlorate AgClO
4
3.6
Dichloromethane CH
2
Cl
2
7.8 Tin iodide Snl
4
6.7
Dinitrobenzene C
6
H
4
(NO
2
)
2
7.8 Sulfur dioxide SO
2
2.1
Ethane C
2
H
6
2.6 Sulfur hexafluoride SF
6
2.3
Ethanol C
2
H
5
OH 3.6 Toluene C
6
H
5
CH
3
6.8
Ethylene oxide (CH
2
)
2
O 2.5 Trinitrobenzene C
6
H
3
(NO
2
)
3
9.0
Helium He 0.14 Water vapor H
2
O 11.0
Hexane C
6
H
14
6.6 Xenon Xe 3.0
Hydrogen H
2
0.44 Xylols C
6
H
4
(CH
3
)
2
7.8
Table 4.3 Relative ionization probabilities (RIP) vis vis nitrogen, electron energy 102 eV
Electron energy : 75 eV (PGA 100) 102 eV (Transpector)
Gas Symbol Mass = 100 % Greatest peak = 100 % = 100 % Greatest peak = 100 %
Argon Ar 40 74.9 100 90.9 100
20 24.7 33.1 9.1 10
36 0.3
Carbon dioxide CO
2
45 0.95 1.3 0.8 1
44 72.7 100 84 100
28 8.3 11.5 9.2 11
16 11.7 16.1 7.6 9
12 6.15 8.4 5 6
Carbon monoxide CO 29 1.89 2.0 0.9 1
28 91.3 100 92.6 100
16 1.1 1.2 1.9 2
14 1.7 1.9 0.8
12 3.5 3.8 4.6 5
Neon Ne 22 9.2 10.2 0.9 11
20 89.6 100 90.1 100
10 0.84 0.93 9 4
Oxygen O
2
34 0.45 0.53
32 84.2 100 90.1 100
16 15.0 17.8 9.9 11
Nitrogen N2 29 0.7 0.8 0.9 1
28 86.3 100 92.6 100
14 12.8 15 6.5 12
Water vapor H
2
O 19 1.4 2.3
18 60 100 74.1 100
17 16.1 27 18.5 25
16 1.9 3.2 1.5 2
2 5.0 8.4 1.5 2
1 15.5 20 4.4 6
Table 4.4 Fragment distribution for certain gases at 75 eV and 102 eV
D00 E 19.06.2001 21:39 Uhr Seite 98
Mass Spectrometry
D00
No Gas Symbol 1 = 100 2 3 4 5 6
1 Acetone (CH
3
)
2
CO 43/100 15/42 58/20 14/10 27/19 42/8
2 Air 28/100 32/27 14/6 16/3 40/1 -
3 Ammonia NH
3
17/100 16/80 15/8 14/2 - -
4 Argon Ar 40/100 20/10 - - - -
5 Benzene C
6
H
6
78/100 77/22 51/18 50/17 52/15 39/10
6 Carbon dioxide CO
2
44/100 28/11 16/9 12/6 45/1 22/1
7 Carbon monoxide CO 28/100 12/5 16/2 29/1 - -
8 Carbon tetrachloride CCl
4
117/100 119/91 47/51 82/42 35/39 121/29
9 Carbon tetrafluoride CF
4
69/100 50/12 31/5 19/4 - -
10 Diff. pump oil, DC 705 78/100 76/83 39/73 43/59 91/32 -
11 Diff. pump oil, Fomblin 69/100 20/28 16/16 31/9 97/8 47/8
12 Diff. pump oil, PPE 50/100 77/89 63/29 62/27 64/21 38/7
13 Ethanol CH
3
CH
2
OH 31/100 45/34 27/24 29/23 46/17 26/8
14 Halocarbon 11 CCl
3
F 101/100 103/60 35/16 66/15 47/12 31/10
15 Halocarbon 12 CCl
2
F
2
85/100 87/32 50/16 35/12 - -
16 Halocarbon 13 CClF
3
69/100 85/15 50/14 31/9 35/7 87/5
17 Halocarbon 14 CF
4
69/100 12/7 19/6 31/5 50/8 -
18 Halocarbon 23 CHF
3
51/100 31/58 69/40 50/19 52/1 21/1
19 Halocarbon 113 C
2
C
13
F
3
101/100 103/62 85/55 31/50 151/41 153/25
20 Helium He 4/100 - - - - -
21 Heptane C
7
H
16
43/100 41/62 29/49 27/40 57/34 71/28
22 Hexane C
6
H
14
41/100 43/92 57/85 29/84 27/65 56/50
23 Hydrogen H
2
2/100 1/5 - - - -
24 Hydrogen sulfide H
2
S 34/100 32/44 33/42 36/4 35/2 -
25 Isopropyl alcohol C
3
H
8
O 45/100 43/16 27/16 29/10 41/7 39/6
26 Krypton Kr 84/100 86/31 83/20 82/20 80/4 -
27 Methane CH
4
16/100 15/85 14/16 13/8 1/4 12/2
28 Mehtyl alcohol CH
3
OH 31/100 29/74 32/67 15/50 28/16 2/16
29 Methyl ethyl ketone C
4
H
8
O 43/100 29/25 72/16 27/16 57/6 42/5
30 Mechanical pump oil 43/100 41/91 57/73 55/64 71/20 39/19
31 Neon Ne 20/100 22/10 10/1 - - -
32 Nitrogen N
2
28/100 14/7 29/1 - - -
33 Oxygen O
2
32/100 16/11 - - - -
34 Perfluorokerosene 69/100 119/17 51/12 131/11 100/5 31/4
35 Perfluor-tributylamine C
12
F
27
N 69/100 131/18 31/6 51/5 50/3 114/2
36 Silane SiH
4
30/100 31/80 29/31 28/28 32/8 33/2
37 Silicon tetrafluoride SiF
4
85/100 87/12 28/12 33/10 86/5 47/5
38 Toluene C
6
H
5
CH
3
91/100 92/62 39/12 65/6 45.5/4 51/4
39 Trichloroethane C
2
H
3
Cl
3
97/100 61/87 99/61 26/43 27/31 63/27
40 Trichloroethylene C
2
HCl
3
95/100 130/90 132/85 97/64 60/57 35/31
41 Trifluoromethane CHF
3
69/100 51/91 31/49 50/42 12/4 -
42 Turbomolecular pump oil 43/100 57/88 41/76 55/73 71/52 69/49
43 Water vapor H
2
O 18/100 17/25 1/6 16/2 2/2 -
44 Xenon Xe 132/100 129/98 131/79 134/39 136/33 130/15
Table 4.5 Spectrum library of the 6 highest peaks for the TRANSPECTOR
D00 E 19.06.2001 21:39 Uhr Seite 99
Mass Spectrometry
4.6.2 Partial pressure
measurement
The number of ions i
+
gas
produced from a
gas in the ion source is proportional to the
emission current i
, to the specific ioni-

zation S
gas
, to a geometry factor f repre-
senting the ionization path inside the
ionization source, to the relative ionization
probability RIP
gas
, and to the partial pres-
sure p
gas
. This number of ions produced
is, by definition, made equal to the sensi-
tivity E
gas
times the partial pressure p
gas
:
i
+
gas
(produced) = i
-
S
gas
f RIP
gas
p
gas
= Egas pgas
due to RIP
N2
= 1
EN
2
is equal to i
SN
2
f
and E
gas
is equal to EN
2
RIP
gas
Almost all gases form fragments during
ionization. To achieve quantitative evalu-
ation one must either add the ion flows at
the appropriate peaks or measure (with a
known fragment factor [FF]) one peak and
calculate the overall ion flow on that basis:
In order to maintain the number of ions
arriving at the ion trap, it is necessary to
multiply the number above with the trans-
mission factor TF(m), which will be depen-
dent on mass, in order to take into account
the permeability of the separation system
for atomic number m (analogous to this,
there is the detection factor for the SEMP;
it, however, is often already contained in
TF). The transmission factor (also: ion-
optical transmission) is thus the quotient
of the ions measured and the ions produ-
ced.
Thus
and with
E
gas
= E
N2
RIP
gas
the ultimate result is:
( )
p
i produced
BF E
gas
gas m
gas m gas
=
+
,
,
2
2
( )
p
i measured
BF E TF m
gas
gas m
gas m gas
=

+
,
,
( )
2
2
i produced i i
i
BF
i
BF
E p
gas gas m gas m
gas m
gas m
gas m
gas m
gas gas
+ + +
+
+
= + + =
= = =
( ) ....
....
, ,
,
,
,
,
1 2
1
1
2
2
(4.3)
The partial pressure is calculated from the
ion flow measured for a certain fragment
by multiplication with two factors. The first
factor will depend only on the nitrogen
sensitivity of the detector and thus is a
constant for the device. The second will
depend only on the specific ion properties.
These factors will have to be entered sepa-
rately for units with direct partial pressure
indication (at least for less common types
of ions).
4.6.3 Qualitative gas analysis
The analysis of spectra assumes certain
working hypotheses:
1. Every type of molecule produces a cer-
tain, constant mass spectrum or frag-
ment spectrum which is characteristic
for this type of molecule (fingerprint,
cracking pattern).
2. The spectrum of every mixture of gases
is the same as would be found through
linear superimposition of the spectra of
the individual gases. The height of the
peaks will depend on the gas pressure.
3. The ion flow for each peak is proportio-
nal to the partial pressure of each com-
ponent responsible for the peak. Since
the ion flow is proportional to the parti-
al pressure, the constant of proportio-
nality (sensitivity) varies from one gas
to the next.
Although these assumptions are not
always correct (see Robertson: Mass
Spectrometry) they do represent a useful
working hypothesis.
In qualitative analysis, the unknown spec-
trum is compared with a known spectrum
in a library. Each gas is definitively deter-
mined by its spectrum. The comparison
with library data is a simple pattern reco-
gnition process. Depending on the availa-
bility, the comparison may be made using
any of a number ancillary aids. So, for
example, in accordance with the position,
size and sequence of the five or ten highest
peaks. Naturally, comparison is possible
( ) p i measured
E BF RIP TF m
gas gas m
N gas m gas
=

+
,
,
( )
2
2 2
1 1
only after the spectrum has been standar-
dized, by setting the height of the highest
line equal to 100 or 1000 (see Table 4.5 as
an example).
The comparison can be made manually on
the basis of collections of tables (for
example, A. Cornu & R. Massot: Compil-
ation of Mass Spectral Data) or may be
effected with computer assistance; large
databases can be used (e.g. Mass Spectral
Data Base, Royal Society of Chemistry,
Cambridge).
When making comparisons with library
information, it is necessary to pay attenti-
on to whether identical ion sources or at
least identical electron impact energies
were used.
All these capabilities are, however, gene-
rally too elaborate for the problems
encountered in vacuum technology. Many
commercial mass spectrometers can show
a number of library spectra in the screen
so that the user can see immediately
whether the library substance might be
contained in the substance measured.
Usually the measured spectrum was the
result of a mix of gases and it is particu-
larly convenient if the screen offers the
capacity for subtracting (by way of trial)
the spectra of individual (or several) gases
from the measured spectrum. The gas can
be present only when the subtraction does
not yield any negative values for the major
peaks. Figure 4.16 shows such a step-by-
step subtraction procedure using the
Transpector-Ware software.
Regardless of how the qualitative analysis
is prepared, the result is always just a
suggestion, i.e. an assumption as to
which gases the mixture might contain.
This suggestion will have still to be exami-
ned, e.g. by considering the likelihood that
a certain substance would be contained in
the spectrum. In addition, recording a new
spectrum for this substance can help to
achieve clarity.
D00 E 19.06.2001 21:39 Uhr Seite 100
Mass Spectrometry
4.6.4 Quantitative gas analysis
Particular difficulties are encountered
when interpreting the spectrum of an
unknown mixture of gases. The proporti-
ons of ion flow from various sources can
be offset one against the other only after
all the sources have been identified. In
many applications in vacuum technology
one will be dealing with mixtures of a few
simple gases of known identity, with ato-
mic numbers less of than 50 (whereby the
process-related gases can represent
exceptions). In the normal, more compli-
cated case there will be a spectrum with a
multitude of superimpositions in a com-
pletely unknown mixture of many gas
components; here a qualitative analysis
will have to be made before attempting
quantitative analysis. The degree of diffi-
culty encountered will depend on the num-
ber of superimpositions (individual/a few/
many).
In the case of individual superimpositions,
mutual, balancing of the ion flows during
measurement of one and the same type of
gas for several atomic numbers can often
be productive.
Where there is a larger number of super-
impositions and a limited number of gases
overall, tabular evaluation using correction
factors vis vis the spectrum of a calibra-
tion gas of known composition can often
be helpful.
In the most general case a plurality of
gases will make a greater or lesser contri-
bution to the ion flow for all the masses.
The share of a gas g in each case for the
atomic number m will be expressed by the
fragment factor Ff
m,g
. In order to simplify
calculation, the fragment factor Ff
m,g
will
also contain the transmission factor TF
and the detection factor DF. Then the ion
current to mass m, as a function of the
overall ion currents of all the gases invol-
ved, in matrix notation, is:
The ion current vector for the atomic num-
bers m (resulting from the contributions
by the fragments of the individual gases) is
i
i
i
BF BF
FF
BF BF
I
I
I
j
m
u
j k j o
m g
u k uo
k
g
o
+
+
+
+
+
+
1
]
1
1
1
1
1
1
1
1
1
]
1
1
1
1
1
1
1
1
1
]
1
1
1
1
1
1
1
1
, ,
,
, ,
D00
A Parent
range
A
Parent spectrum
Parent spectrum Assumption:
Groups
1 Kr =
=
=
ypton
+
2 Krypton
++
2 1
Library spectrum:
Krypton
Parent spectrum without krypton
Assumption:
3 Argon
+
4 Argon
++
Library spectrum:
Argon
Parent spectrum without argon Assumption:
5 Neon
+
Library spectrum:
Neon
Parent spectrum after detection
of krypton, argon and neon
3
4
5
Fig. 4.16 Subtracting spectra contained in libraries
D00 E 19.06.2001 21:39 Uhr Seite 101
Mass Spectrometry
LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002
equal to the fragment matrix times the vec-
tor of the sum of the flows for the indivi-
dual gases.
or:
(in simplified notation: i = FF I)
where i
m
+
= ion flow vector for the atomic
numbers, resulting from contributions of
fragments of various individual gases
= fragment matrix
I
g
+
= Vector of the sum of the flows for the
individual gases or:
One sees that the ion flow caused by a gas
is proportional to the partial pressure. The
linear equation system can be solved only
for the special instance where m = g
(square matrix); it is over-identified for
m > g. Due to unavoidable measurement
error (noise, etc.) there is no set of overall
ion flow I
+
g
(partial pressures or concentra-
tions) which satisfies the equation system
exactly. Among all the conceivable soluti-
ons it is now necessary to identify set I
+
*
g
which after inverse calculation to the parti-
al ion flows I
+
*
m
will exhibit the smallest
squared deviation from the partial ion cur-
rents i
m
+
actually measured. Thus:
This minimization problem is mathematic-
ally identical to the solution of another
equation system
FFT i = FFT FF I
which can be evaluated direct by the com-
puter. The ion current vector for the indivi-
dual gases is then:
[ ] [ ]
[ ]
I
FF i FF BF
FF BF
T T
T
=

1
det
( ) i i
m m
=
*
+ +
min
2
Ff
m, g
6444474448
i p E RIP FF TF
m g N g m m
+

2
Transmission factor
for the mass m
Fragment factor
for the gas to mass m
Relative ionization probability
for the gas
Nitrogen sensitivity (equipment constant)
Partial pressure of the gas
Ion current for atomic number m
BF
m g
g k
,
=
0
i BF I
m m g g
g k
+ +
=
=
,
0
4.7 Software
4.7.1 Standard SQX software
(DOS) for stand-alone
operation (1 MS plus
1 PC, RS 232)
The conventional software package (SQX)
contains the standard routines for the
operation of the mass spectrometer
(MS) various spectra depictions, queries
of individual channels with the correspon-
ding screen displays as tables or bar
charts, partial pressure conversion, trend
displays, comparison with spectra libra-
ries (with the capability for trial subtrac-
tion of library spectra), leak testing mode
etc. and for sensor balancing, as well.
Using PCs as the computer and display
unit naturally makes available all the usual
functions including storing and retrieving
data, printing, etc. Characteristic of the
conventional software package is that spe-
cific individual spectra will be measured,
even though the measurement is fully
automated and takes place at a point in
time which is specified in advance. A
spectrum of this type can thus be only a
snapshot of a process in progress.
4.7.2 Multiplex/DOS software
MQX (1 to 8 MS plus
1 PC, RS 485)
The first step toward process-oriented
software by INFICON is the MQX. It makes
possible simultaneous monitoring of a
maximum of eight sensors and you can
apply all the SQX functions at each sensor.
4.7.3 Process-oriented soft-
ware Transpector-Ware
for Windows
Transpector-Ware is based on an entirely
new philosophy. During the course of the
process (and using settings the recipe
determined beforehand) data will be
recorded continuously like the individual
frames in a video. These data can be sto-
red or otherwise evaluated. It is possible
in particular to analyze interesting process
sections exactly, both during the process
and retroactively, once the process has run
to completion, without having to interrupt
the measurement operations which are
running in the background. Where
ongoing monitoring of identical processes
is undertaken the program can generate
statistics (calculating mean values and
standard deviations) from which a band-
width for favorable process operation
can be derived. Error reports are issued
where limit values are exceeded.
4.7.4 Development software
TranspectorView
This software used to for develop custom
software versions for special situations. It
is based on the LabView development
package and includes the drivers required
to operate the Transpector.
4.8 Partial pressure
regulation
Some processes, such as reactive sputter
processes, require the most constant pos-
sible incidence rates for the reacting gas
molecules on the substrate being coated.
The incidence rate is the same as the
impingement rate discussed in Chapter
1; it is directly proportional to the partial
pressure. The simplest attempt to keep the
partial pressure for a gas component con-
stant is throughput by regulating with a
flow controller; it does have the disadvan-
tage that the regulator cannot determine
whether, when and where the gas con-
sumption or the composition of the gas in
the vacuum chamber changes. The far
superior and more effective option is parti-
al pressure control using a mass spectro-
meter via gas inlet valves. Here the signifi-
cant peaks of the gases being considered
are assigned to channels in the mass spec-
trometer. Suitable regulators compare the
analog output signals for these channels
with set-point values and derive from the
difference between the target and actual
values for each channel the appropriate
actuation signal for the gas inlet valve for
D00.102
D00 E 19.06.2001 21:39 Uhr Seite 102
Mass Spectrometry
the channel. A configuration of this kind
has been realized to control six channels in
the QUADREX PPC. Gas inlet valves mat-
ching the unit can also be delivered.
The gas used to measure the impingement
rate (partial pressure) must naturally be
drawn from a representative point in the
vacuum chamber. When evaluating the
time constant for a regulation circuit of this
type it is important to take into account all
the time aspects and not just the electrical
signal propagation and the processing in
the mass spectrometer, but also the vacu-
um-technology time constants and flow
velocities, as illustrated in Figure 4.17.
Pressure converters or unfavorably instal-
led gas inlet lines joining the control valve
and the vacuum vessel will make particu-
larly large contributions to the overall time
constant. It is generally better to establish
a favorable S/N ratio with a large signal (i.e.
through an inlet diaphragm with a large
opening) rather than with long integration
periods at the individual channels. Contra-
sted in Figure 4.18 are the effects of boo-
sting pressure and lengthening the integra-
tion time on signal detectability. In depic-
tions a, b and c only the integration period
was raised, from 0.1 to 1.0 and 10 seconds
(thus by an overall factor of 100), respec-
tively. By comparison, in the sequence a-d-
e-f, at constant integration time, the total
pressure was raised in three steps, from
7.2 10
-6
mbar to 7.2 10
-5
mbar (or by a
factor of just 10 overall).
4.9 Maintenance
(Cathode service life, sensor balancing,
cleaning the ion source and rod system)
The service life of the cathode will depend
greatly on the nature of the loading. Expe-
rience has shown that the product of ope-
rating period multiplied by the operating
pressure can serve as a measure for the
loading. Higher operating pressures (in a
range of 1 10
-4
to 1 10
-3
mbar) have a
particularly deleterious effect on service
life, as do certain chemical influences such
as refrigerants, for example. Changing out
the cathode is quite easy, thanks to the
simple design of the sensor. It is advisab-
le, however, to take this opportunity to
change out or at least clean the entire ion
source.
Sensor balancing at the mass axis (often
erroneously referred to as calibration) is
done today in a very easy fashion with the
software (e.g. SQX, Transpector-Ware)
and can be observed directly in the screen.
Naturally, not only the arrangement along
the mass axis will be determined here, but
also the shape of the lines, i.e. resolution
and sensitivity (see Section 4.5).
It will be necessary to clean the sensor
only in exceptional cases where it is heavi-
ly soiled. It is usually entirely sufficient to
clean the ion source, which can be easily
dismantled and cleaned. The rod system
can be cleaned in an ultrasonic bath once
it has been removed from the config-
uration. If dismantling the system is una-
voidable due to particularly stubborn
grime, then the adjustment of the rods
which will be required afterwards will have
to be carried out at the factory.
D00
TMP50CF
Sensor
Pressure stage
Vacuum vessel
Mass spectrometer
Regulation valve
t
3
t
2
t
1
t
6
t
5
t
4
Fig. 4.17 Partial shares for overall time constants

Fig. 4.18 Improving the signal-to-noise ratio by increasing the pressure or extending the integration time
a
b c
d f e
D00 E 19.06.2001 21:39 Uhr Seite 103
Leak Detection
5. Leaks and their
detection
Apart from the vacuum systems them-
selves and the individual components used
in their construction (vacuum chambers,
piping, valves, detachable [flange] connec-
tions, measurement instruments, etc.),
there are large numbers of other systems
and products found in industry and rese-
arch which must meet stringent require-
ments in regard to leaks or creating a so-
called hermetic seal. Among these are
many assemblies and processes in the
automotive and refrigeration industries in
particular but also in many other branches
of industry. Working pressure in this case
is often above ambient pressure. Here
hermetically sealed is defined only as a
relative absence of leaks. Generalized
statements often made, such as no detec-
table leaks or leak rate zero, do not
represent an adequate basis for accep-
tance testing. Every experienced engineer
knows that properly formulated accep-
tance specifications will indicate a certain
leak rate (see Section 5.2) under defined
conditions. Which leak rate is acceptable is
also determined by the application itself.
5.1 Types of leaks
Differentiation is made among the follo-
wing leaks, depending on the nature of the
material or joining fault:
Leaks in detachable connections: Flan-
ges, ground mating surfaces, covers
Leaks in permanent connections:Solder
and welding seams, glued joints
Leaks due to porosity: particularly follo-
wing mechanical deformation (ben-
ding!) or thermal processing of poly-
crystalline materials and cast compo-
nents
Thermal leaks (reversible): opening up
at extreme temperature loading (heat/
cold), above all at solder joints
Apparent (virtual) leaks: quantities of
gas will be liberated from hollows and
cavities inside cast parts, blind holes
and joints (also due to the evaporation
of liquids)
Indirect leaks: leaking supply lines in
vacuum systems or furnaces (water,
compressed air, brine)
Serial leaks: this is the leak at the end
of several spaces connected in series,
e.g. a leak in the oil-filled section of the
oil pan in a rotary vane pump
One-way leaks: these will allow gas to
pass in one direction but are tight in the
other direction (very seldom)
An area which is not gas-tight but which is
not leaky in the sense that a defect is pre-
sent would be the
Permeation (naturally permeability) of
gas through materials such as rubber
hoses, elastomer seals, etc. (unless
these parts have become brittle and
thus leaky).
5.2 Leak rate, leak
size, mass flow
No vacuum device or system can ever be
absolutely vacuum-tight and it does not
actually need to be. The simple essential is
that the leak rate be low enough that the
required operating pressure, gas balance
and ultimate pressure in the vacuum con-
tainer are not influenced. It follows that the
requirements in regard to the gas-tightn-
ess of an apparatus are the more stringent
the lower the required pressure level is. In
order to be able to register leaks quantita-
tively, the concept of the leak rate with
the symbol Q
L
was introduced; it is mea-
sured with mbarls
-1
or cm
3
/s (STP) as
the unit of measure. A leak rate of
Q
L
= 1 mbarls
-1
is present when in an
enclosed, evacuated vessel with a volume
of 1 l the pressure rises by 1 mbar per
second or, where there is positive pressu-
re in the container, pressure drops by 1
mbar. The leak rate Q
L
defined as a mea-
sure of leakiness is normally specified in
the unit of measure mbarls
-1
. With the
assistance of the status equation (1.7) one
can calculate Q
L
when giving the tempera-
ture T and the type of gas M, registering
this quantitatively as mass flow, e.g. in the
g/s unit of measure. The appropriate
relationship is then:
(5.1)
( )
Q
p V
t
R T
M
m
t
L
=

=

where R = 83.14 mbar l/mol K, T = tem-

perature in K; M = molar mass in g/mole;
m for the mass in g; t is the time period
in seconds. Equation 5.1 is then used
a) to determine the mass flow m / t at a
known pV gas flow of p V/t (see in
this context the example at 5.4.1) or
b) to determine the pV leak gas flow where
the mass flow is known (see the follo-
wing example).
Example for case b) above:
A refrigeration system using Freon (R 12)
exhibits refrigerant loss of 1 g of Freon per
year (at 25 C). How large is the leak gas
flow Q
L
? According to equation 5.1 for
M(R12) = 121 g/mole:
Thus the Freon loss comes to
Q
L
= 6.5 10
6
mbarls
-1
. According to the
rule of thumb for high vacuum systems
given below, the refrigeration system men-
tioned in this example may be deemed to
be very tight. Additional conversions for
Q
L
are shown in Tables VIIa and VIIb in
Chapter 9.
The following rule of thumb for quantita-
tive characterization of high vacuum
equipment may be applied:
Total leak rate < 10
-6
mbarls
-1
:
Equipment is very tight
Total leak rate 10
-5
mbarls
-1
:
Equipment is sufficiently tight
Total leak rate > 10
-4
mbarls
-1
:
Equipment is leaky
A leak can in fact be overcome by a
pump of sufficient capacity because it is
true that (for example at ultimate pressure
p
end
and disregarding the gas liberated
from the interior surfaces):
(5.2)
(Q
L
Leak rate, S
eff
the effective pumping
speed at the pressure vessel)
p
end
Q
L
S
eff
=
( )
Q
p V
t
mbar K g
mol K g mol year
mbar
s
L

8314 298 1
121 1
8314 2 10 1
1211
315 10
1
2
7
.
. .98
.
8314 2 10
1 10 315
10
65 10
2
2
7
7
. .98
.21 .
mbar
s
mbar
s
D00 E 19.06.2001 21:39 Uhr Seite 104

Concept / criterion Comment Q
L
[mbar l/s] Relevant particle size
Water-tight *
)
Droplets Q
L
< 10
2
Vapor-tight Sweating Q
L
< 10
3
Bacteria-tight *
)
(cocci) Q
L
< 10
4
Avg. 1 m
(rod-shaped) Avg. 0.5
-1
m, 2
10
m long
Oil-tight Q
L
< 10
5
Virus-tight *
)
(vaccines such as pox) Q
L
< 10
6
3 10
7
m
(smallest viruses,
bacteriophages) Q
L
< 10
8
3 10
8
m
(viroids, RNA) Q
L
< 10
10
1 10
9
m (thread-like)
Gas-tight Q
L
< 10
7
Absolutely tight Technical Q
L
< 10
10
*
)
As opposed to vapor, it is necessary to differentiate between hydrophilic and hydrophobic solids. This
also applies to bacteria and viruses since they are transported primarily in solutions.
Leak Detection
Where S
eff
is sufficiently great it is possi-
ble regardless of the value for the leak
rate Q
L
always to achieve a pre-determi-
ned ultimate pressure of p
end
. In practice,
however, an infinite increase of S
eff
will run
up against economic and engineering limi-
tations (such as the space required by the
system).
Whenever it is not possible to achieve the
desired ultimate pressure in an apparatus
there are usually two causes which can be
cited: The presence of leaks and/or gas
being liberated from the container walls
and sealants.
Partial pressure analysis using a mass
spectrometer or the pressure rise method
may be used to differentiate between these
two causes. Since the pressure rise
method will only prove the presence of a
leak without indicating its location in the
apparatus, it is advisable to use a helium
leak detector with which leaks can, in
general, also be located much more quick-
ly.
In order to achieve an overview of the cor-
relation between the geometric size of the
hole and the associated leak rate it is pos-
sible to operate on the basis of the follo-
wing, rough estimate: A circular hole 1 cm
in diameter in the wall of a vacuum vessel
is closed with a gate valve. Atmospheric
pressure prevails outside, a vacuum insi-
de. When the valve is suddenly opened all
the air molecules in a cylinder 1 cm in dia-
meter and 330 m high would within a
1-second period of time fall into the hole
at the speed of sound (330 m/s). The
quantity flowing into the vessel each
second will be 10
13
mbar times the cylin-
der volume (see Fig. 5.1). The result is that
for a hole 1 cm in diameter Q
L
(air) will be
2.6 10
4
mbarls
-1
. If all other conditions
are kept identical and helium is allowed to
flow into the hole at its speed of sound of
970 m/s, then in analogous fashion the Q
L
(helium) will come to 7.7 10
+4
mbarls
-1
,
or a pV leaking gas current which is larger
by a factor of 970 / 330 = 2.94. This grea-
ter sensitivity for helium is used in leak
detection practice and has resulted in the
development and mass production of hig-
hly sensitive helium-based leak detectors
(see Section 5.5.2).
Shown in Figure 5.1 is the correlation bet-
ween the leak rate and hole size for air, with
the approximation value of Q
L
(air) of
10
+4
mbarls
-1
for the 1 cm hole. The
table shows that when the hole diameter is
reduced to 1 m (= 0.001 mm) the leak rate
will come to 10
-4
mbarls
-1
, a value which
in vacuum technology already represents a
major leak (see the rule of thumb above). A
leak rate of 10
-12
mbarls
-1
corresponds to
hole diameter of 1 ; this is the lower
detection limit for modern helium leak
detectors. Since the grid constants for
many solids amount to several and the
diameter of smaller molecules and atoms
(H
2
, He) are about 1 , inherent permeati-
on by solids can be registered metrologi-
cally using helium leak detectors. This has
led to the development of calibrated refe-
rence leaks with very small leak rates (see
Section 5.5.2.3). This is a measurable lack
of tightness but not a leak in the sense
of being a defect in the material or joint.
Estimates or measurements of the sizes of
atoms, molecules, viruses, bacteria, etc.
have often given rise to everyday terms
such as watertight or bacteria-tight;
see Table 5.1.
Compiled in Figure 5.2 are the nature and
detection limits of frequently used leak
detection methods.
D00
p = 1013 mbar, Hole diameter d = 1 cm
Gas speed = Speed of sound = 330
m
s
mbar
s

mbar
s

s
s
cm
s
3
m
s
1
4
2
Volume/second:
Quantity/second:
Diameter cm Leak rate
1013 mbar 25.95 = 2.63 10 10
+4 +4
330 cm = 25.95 10 = 25.95

2 +3
10 m= 1.0 cm 10
10 m= 1.0 mm 10
10 m= 0.1 mm 10 (= 1)
10 m= 0.01mm 10
10 m= 1.0 m 10
10 m= 0.1 m 10
10 m= 0.01 m 10
10 m= 1.0 nm 10
10 m= 1.0 Angstrom 10 (Detection limit, He leak detector)
2 +4
3 +2
4 0
5 2
6 4
7 6
8 8
9 10
10 12
Fig. 5.1 Correlation between leak rate and hole size

Table 5.1 Estimating borderline leak rates
D00 E 19.06.2001 21:39 Uhr Seite 105
Range Laminar Molecular
Pressure
Gas
Leak Detection
5.2.1 The standard helium
leak rate
Required for unequivocal definition of a
leak are, first, specifications for the
pressures prevailing on either side of the
partition and, secondly, the nature of the
medium passing through that partition
(viscosity) or its molar mass. The designa-
tion helium standard leak (He Std) has
become customary to designate a situati-
on frequently found in practice, where
testing is carried out using helium at 1 bar
differential between (external) atmosphe-
ric pressure and the vacuum inside a
system (internal, p < 1 mbar), the designa-
tion helium standard leak rate has beco-
me customary. In order to indicate the
rejection rate for a test using helium under
standard helium conditions it is necessary
first to convert the real conditions of use to
helium standard conditions (see Section
5.2.2). Some examples of such conversi-
ons are shown in Figure 5.3.
5.2.2 Conversion equations
When calculating pressure relationships
and types of gas (viscosity) it is necessary
to keep in mind that different equations are
applicable to laminar and molecular flow;
the boundary between these areas is very
difficult to ascertain. As a guideline one
may assume that laminar flow is present at
leak rates where Q
L
> 10
-5
mbar l/s and
molecular flow at leak rates where
Q
L
< 10
-7
mbar l/s. In the intermediate
range the manufacturer (who is liable
under the guarantee terms) must assume
values on the safe side. The equations are
listed in Table 5.2.
Here indices I and II refer to the one or
the other pressure ratio and indices 1
and 2 reference the inside and outside of
the leak point, respectively.
5.3 Terms and
definitions
When searching for leaks one will general-
ly have to distinguish between two tasks:
1. Locating leaks and
2. Measuring the leak rate.
In addition, we distinguish, based on the
direction of flow for the fluid, between the
a. vacuum method (sometimes known as
an outside-in leak), where the direc-
tion of flow is into the test specimen
(pressure inside the specimen being
less than ambient pressure), and the
b. positive pressure method (often refer-
red to as the inside-out leak), where
the fluid passes from inside the test
specimen outward (pressure inside the
specimen being greater than ambient
pressure).
The specimens should wherever possible
be examined in a configuration corre-
sponding to their later application com-
ponents for vacuum applications using the
vacuum method and using the positive
pressure method for parts which will be
pressurized on the inside.
When measuring leak rates we differentia-
te between registering
a. individual leaks (local measurement)
sketches b and d in Figure 5.4, and regi-
stering
b. the total of all leaks in the test specimen
(integral measurement) sketches a
and c in Figure 5.4.
The leak rate which is no longer tolerable
in accordance with the acceptance specifi-
cations is known as the rejection rate. Its
calculation is based on the condition that
the test specimen may not fail during its
planned utilization period due to faults
caused by leaks, and this to a certain
V
a
c
u
u
m

m
e
t
h
o
d
O
v
e
r
p
r
e
s
s
u
r
e

m
e
t
h
o
d
Helium leak detector ULTRATEST UL 200 dry/UL 500
ULTRATEST with helium sniffer
Halogen sniffer HLD4000A
Bubble test
Pressure drop test
10
3
................10
0
10
-1
10
-2
10
-3
10
-4
10
-5
10
-6
10
-7
10
-8
10
-9
10
-10
10
-11
10
-12
mbar l s
-1
Ecotec II / Protec
Contura Z
Pressure rise
Helium leak detector ULTRATEST UL 200/UL 500 dry/Modul 200/LDS 1000
Fig. 5.2 Leak rate ranges for various leak detection processes and devices
Leak <----> Hole
Q ... Leak rate,
In short: Leak
Substance quantity hole unit of time trhough per Helium standard leak rate:
p = 1 bar, p < 1 mbar ( p = 1 bar)
Test gas = Helium
1 2

Familiar leaks: Quantity escaping: Standard He leak rate:

Dripping water faucet 34
10
3.18 10
4.19 10
430 Frigen
2
4
3 3
2
5
2
5 3
3
Water Air He Std
Air
He Std
He Std
He Std
He Std F12
Air
Air
= 6.45 0.17
= 4.24 10
0.9 10
4.3 10
1.88 10
4.33 10 = 2.8 10
mg
Ncm
g
mbar mbar
mbar
mbar
mbar
mbar
mbar mbar
mbar
mbar
s
s
a
s s
s
s
s
s
s s
s
s
4 mm diam., 1 Hz, p = 4 bar
Hair on a gasket
Bicycle tube in water
(bubble test)
2 mm diam., 1 Hz, p = 0.1 bar
Car tire loses air
25 l, 6 Mo: 1.8 --> 1.6 bar
Small refrigerant cylinder
empties in 1 year
430 g refrigerant R12, 25C
Definition: Q =
(p V)
t
Fig. 5.3 Examples for conversion into helium standard leak rates
Table 5.2 Conversion formulae for changes of pressure and gas type
( ) ( ) Q p p Q p p
I
II
II
I

1
2
2
2
1
2
2
2
( ) ( ) Q p p Q p p
I
II
II
I

1 2 1 2
Q M Q M
gas A gas A gas B gas B
Q
gas A

gas A
= Q
gas B

gas B
D00 E 19.06.2001 21:39 Uhr Seite 106
Leak Detection
degree of certainty. Often it is not the leak
rate for the test specimen under normal
operating conditions which is determined,
but rather the throughput rate of a test gas
primarily helium under test conditions.
The values thus found will have to be con-
verted to correspond to the actual applica-
tion situation in regard to the pressures
inside and outside the test specimen and
the type of gas (or liquid) being handled.
Where a vacuum is present inside the test
specimen (p < 1 mbar), atmospheric pres-
sure outside, and helium is used at the test
gas, one refers to standard helium condi-
tions. Standard helium conditions are
always present during helium leak detec-
tion for a high vacuum system when the
system is connected to a leak detector and
is sprayed with helium (spray technique).
If the specimen is evacuated solely by the
leak detector, then one would say that the
leak detector is operating in the direct-
flow mode. If the specimen is itself a com-
plete vacuum system with its own vacuum
pump and if the leak detector is operated
in parallel to the systems pumps, then one
refers to partial-flow mode. One also
refers to partial stream mode when a sepa-
rate auxiliary pump is used parallel to the
leak detector.
When using the positive pressure method
it is sometimes either impractical or in fact
impossible to measure the leakage rate
directly while it could certainly be sensed
in an envelope which encloses the test
specimen. The measurement can be made
by connecting that envelope to the leak
detector or by accumulation (increasing
the concentration) of the test gas inside
the envelope. The bombing test is a
special version of the accumulation test
(see Section 5.7.4). In the so-called sniffer
technique, another variation of the of the
positive pressure technique, the (test) gas
issuing from leaks is collected (extracted)
by a special apparatus and fed to the leak
detector. This procedure can be carried out
using either helium or refrigerants or SF6
as the test gas.
5.4 Leak detection
methods without a
leak detector unit
The most sensible differentiation between
the test methods used is differentiation as
to whether or not special leak detection
equipment is used.
In the simplest case a leak can be determi-
ned qualitatively and, when using certain
test techniques, quantitatively as well (this
being the leak rate) without the assistance
of a special leak detector. Thus the quanti-
ty of water dripping from a leaking water
faucet can be determined, through a cer-
tain period of time, using a graduated
cylinder but one would hardly refer to this
as a leak detector unit. In those cases
where the leak rate can be determined
during the course of the search for the leak
without using a leak detector (see, for
example, Sect. 5.4.1), this will often be
converted to the helium standard leak rate
(Sect. 5.2.1). This standard leak rate value
is frequently needed when issuing accept-
ance certificates but can also be of service
when comparing leak rate values deter-
mined by helium leak detector devices.
In spite of careful inspection of the indivi-
dual engineering components, leaks may
also be present in an apparatus following
its assembly be it due to poorly seated
seals or damaged sealing surfaces. The
processes used to examine an apparatus
will depend on the size of the leaks and on
the degree of tightness being targeted and
also on whether the apparatus is made of
metal or glass or other materials. Some
leak detection techniques are sketched out
below. They will be selected for use in
accordance with the particular application
situations; economic factors may play an
important part here.
5.4.1 Pressure rise test
This leak testing method capitalizes on the
fact that a leak will allow a quantity of gas
remaining uniform through a period of
time to enter a sufficiently evacuated
device (impeded gas flow, see Fig. 1.1). In
contrast, the quantity of gas liberated from
container walls and from the materials
used for sealing (if these are not suffi-
D00
Fig. 5.4 Leak test techniques and terminology
Helium
Helium
Helium
c: Integral leak detection (test gas enrichment
inside the enclosure); pressurized test gas inside
specimen
d: Local leak detection; pressurized test gas
inside the specimen
a: Integral leak detection; vacuum inside
specimen
b: Local leak detection; vacuum inside specimen
D00 E 19.06.2001 21:39 Uhr Seite 107
Leak Detection
ciently free of outgassing) will decline
through time since these will practically
always be condensable vapors for which
an equilibrium pressure is reached at
some time (see Fig. 5.5). The valve at the
pump end of the evacuated vacuum vessel
will be closed in preparation for pressure
rise measurements. Then the time is mea-
sured during which the pressure rises by a
certain amount (by one power of ten, for
example). The valve is opened again and
the pump is run again for some time, fol-
lowing which the process will be repeated.
If the time noted for this same amount of
pressure rise remains constant, then a leak
is present, assuming that the waiting peri-
od between the two pressure rise trials
was long enough. The length of an appro-
priate waiting period will depend on the
nature and size of the device. If the pres-
sure rise is more moderate during the
second phase, then the rise may be assu-
med to result from gases liberated from
the inner surfaces of the vessel. One may
also attempt to differentiate between leaks
and contamination by interpreting the
curve depicting the rise in pressure. Plot-
ted on a graph with linear scales, the curve
for the rise in pressure must be a straight
line where a leak is present, even at higher
pressures. If the pressure rise is due to
gas being liberated from the walls (owing
ultimately to contamination), then the
pressure rise will gradually taper off and
will approach a final and stable value. In
most cases both phenomena will occur
simultaneously so that separating the two
causes is often difficult if not impossible.
These relationships are shown schematic-
ally in Figure 5.5. Once it has become clear
that the rise in pressure is due solely to a
real leak, then the leak rate can be deter-
mined quantitatively from the pressure
rise, plotted against time, in accordance
with the following equation:
(5.3)
Example: Once the vacuum vessel with a
volume of 20 l has been isolated from the
pump, the pressure in the apparatus rises
from 110
-4
mbar to 110
-3
mbar in 300 s.
Thus, in accordance with equation 5.2, the
leak rate will be
The leak rate, expressed as mass flow
m / t, is derived from equation 5.1 at
Q
L
= 6 10
-5
mbar l/s, T = 20 C and the
molar mass for air (M = 29 g/mole) at
If the container is evacuated with a
TURBOVAC 50 turbomolecular pump, for
example (S = 50 l/s), which is attached to
the vacuum vessel by way of a shut-off
valve, then one may expect an effective
pumping speed of about S
eff
= 30 l/s. Thus
the ultimate pressure will be
Naturally it is possible to improve this ulti-
mate pressure, should it be insufficient, by
using a larger-capacity pump (e.g. the
TURBOVAC 151) and at the same time to
reduce the pump-down time required to
reach ultimate pressure.
Today leak tests for vacuum systems are
usually carried out with helium leak detec-
tors and the vacuum method (see Section
5.7.1). The apparatus is evacuated and a
test gas is sprayed around the outside. In
this case it must be possible to detect (on
the basis of samplings inside the appara-
p
end
Q
L
S
eff
mbar s
s
mbar
= =
6 10
5 1
30
1
2 10
6
Q
m
t
mbar
s
g
mol
mol K
mbar K
g
s
L
= =

6 10 29
8314 293 10
7 10
5
2
8
.
Q
L
mbar
s
=
_
,
1 10
3
1 10
4
20
300
9 10
4
20
300
6 10
5

Q
L
p V
t
=

tus) the test gas which has passed through

leaks and into the apparatus. Another opti-
on is to use the positive-pressure leak test.
A test gas (helium) is used to fill the appa-
ratus being inspected and to build up a
slight positive pressure; the test gas will
pass to the outside through the leaks and
will be detected outside the device. The
leaks are located with leak sprays (or soap
suds, 5.4.5) or when using He or H
2
as
the test gas with a leak detector and snif-
fer unit (5.7.2).
5.4.2 Pressure drop test
The thinking here is analogous to that for
the pressure rise method (Section 5.4.1).
The method is, however, used only rarely
to check for leaks in vacuum systems. If
this is nonetheless done, then gauge pres-
sure should not exceed 1 bar since the
flange connectors used in vacuum techno-
logy will as a rule not tolerate higher pres-
sures. Positive pressure testing is, on the
other hand, a technique commonly
employed in tank engineering. When dea-
ling with large containers and the long test
periods they require for the pressure drop
there it may under certain circumstances
be necessary to consider the effects of
temperature changes. As a consequence it
may happen, for example, that the system
cools to below the saturation pressure for
water vapor, causing water to condense;
this will have to be taken into account
when assessing the pressure decline.
5.4.3 Leak test using
vacuum gauges which
are sensitive to the type
of gas
The fact that the pressure reading at vacu-
um gauges (see Section 3.3) is sensitive to
the type of gas involved can, to a certain
extent, be utilized for leak detection purpo-
ses. Thus it is possible to brush or spray
suspected leaks with alcohol. The alcohol
vapors which flow into the device the
thermal conductivity and ionizablity of
which will vary greatly from the same pro-
perties for air will affect and change
pressure indication to a greater or lesser
Fig. 5.5 Pressure rise within a vessel after the pump is
switched off
1 Leak
2 Gas evolved from the container walls
3 Leak + gas evolution
Time
P
r
e
s
s
u
r
e
D00 E 19.06.2001 21:39 Uhr Seite 108
Leak Detection
extent. The availability of more precise,
easy-to-use helium leak detectors has,
however, rendered this method almost
completely obsolete.
5.4.4 Bubble immersion test
The pressurized test specimen is sub-
merged in a liquid bath. Rising gas bub-
bles indicate the leak. Leak detection will
depend greatly on the attentiveness of the
person conducting the test and involves
the temptation to increase the sensitivity
by using ever higher temperatures, wher-
ein the applicable safety regulations are
sometimes disregarded. This method is
very time-consuming for smaller leaks, as
Table 5.3 shows. It references leak testing
on a refrigeration system using type R12
refrigerant. Here the leak rate is specified
in grams of refrigerant loss per year (g/a).
Water is used as a test liquid (which may
be heated or to which a surfactant may be
added) or petroleum-based oils. The surfa-
ce tension should not exceed 75 dyn/cm
(1 dyn = 10
-5
N).
5.4.5 Foam-spray test
In many cases pressurized containers or
gas lines (including the gas supply lines
for vacuum systems) can be checked quite
conveniently for leaks by brushing or
spraying a surfactant solution on them.
Corresponding leak detection sprays are
also available commercially. Escaping gas
forms soap bubbles at the leak points.
Here, again, the detection of smaller leaks
is time-consuming and will depend greatly
on the attentiveness of the inspector. The
hydrogen gas cooling systems used in
power plant generators represent a special
case. These are indeed sometimes tested
in the fashion described above but they
can be examined much better and at much
higher sensitivity by sniffing the hydro-
gen escaping at leaks using a helium leak
detector which has been adjusted to res-
pond to H
2
5.4.6 Vacuum box check
bubble
As a variation on the spray technique men-
tioned above, in which the escaping gas
causes the bubbles, it is possible to place
a so-called vacuum box with a seal
(something like a divers goggles) on the
surface being examined once it has been
sprayed with a soap solution. This box is
then evacuated with a vacuum pump. Air
entering from the outside through leaks
will cause bubbles inside the box, which
can be observed through a glass window
in the box. In this way it is also possible,
for example, to examine flat sheet metal
plates for leaks. Vacuum boxes are availa-
ble for a variety of applications, made to
suit a wide range of surface contours.
5.4.7 Krypton 85 test
When dealing with small, hermetically sea-
led parts where the enclosure is leaky,
krypton 85, a gaseous, radioactive isotope,
can first be forced into the device by app-
lying pressure from the outside. Once an
exactly measured holding period has elap-
sed the pressure will be relieved, the com-
ponent flushed and the activity of the gas
charge will be measured. In the same way
it is also possible to use helium as the test
gas (see Section 5.7.4, bombing test).
5.4.8 High-frequency vacuum
test
The so-called high-frequency vacuum
tester can be used not only to check the
pressure in glass equipment but also to
locate porous areas in plastic or paint coa-
tings on metals. This comprises a hand-
held unit with a brush-like high-frequency
electrode and a power pack. The shape
and color of the electrical gas discharge
can serve as a rough indicator for the pres-
sure prevailing inside glass equipment. In
the case of the vacuum tester which
comprises primarily a tesla transformer
(which delivers a high-voltage, high-fre-
quency AC current) the corona electrode
approaching the apparatus will trigger an
electrode-free discharge inside the appara-
tus. The intensity and color of this dischar-
ge will depend on the pressure and the
type of gas. The luminous discharge phe-
nomenon allows us to draw conclusions
regarding the approximate value for the
pressure prevailing inside the apparatus.
The discharge luminosity will disappear at
high and low pressures.
When searching for leaks in glass equip-
ment the suspect sections will be scanned
or traced with the high-frequency vacuum
tester electrode. Where there is a leak an
arc will strike through to the pore in the
glass wall, tracing a brightly lit discharge
trail. Small pores can be enlarged by these
sparks! The corona discharge of the vacu-
um tester can also penetrate thin areas in
the glass particularly at weld points and
transitional areas between intermediate
components. Equipment which was ori-
ginally leak-free can become leaky in this
fashion! In contrast to the actual leak
detector units, the high-frequency vacuum
tester is highly limited in its functioning.
D00
Freon F12 loss Time taken to form Equivalent Detection time using
per year a gas bubble leak rate helium leak detector
(g/a) (s) (cm
3
[STP]/s) (s)
280 13.3 1.8 10
3
a few seconds
84 40 5.4 10
4
a few seconds
28 145 1.8 10
4
a few seconds
14 290 9.0 10
5
a few seconds
2.8 24 min 1.8 10
5
a few seconds
0.28 * 6 h 1.8 10
6
a few seconds
*) This leak rate represents the detection limit for good halogen leak detectors ( 0,1 g/a).
Table 5.3 Comparison of bubble test method (immersion technique) wit helium leak detector
D00 E 19.06.2001 21:39 Uhr Seite 109
Leak Detection
5.4.9 Test with chemical
reactions and dye
penetration
Occasionally leaks can also be located or
detected by means of chemical reactions
which result in a discoloration or by pene-
tration of a dye solution into fine openings.
The discoloration of a flame due to halo-
gen gas escaping through leaks was used
earlier to locate leaks in solder joints for
refrigeration units.
A less frequently employed example of a
chemical effect would be that of escaping
ammonia when it makes contact with oza-
lid paper (blueprint paper) or with other
materials suitably prepared and wrapped
around the outside of the specimen. Leaks
are then detected based on the discolorati-
on of the paper.
An example of a dye penetration test is the
inspection of the tightness of rubber plugs
or plungers in glass tubes, used for exam-
ple in testing materials suitability for dis-
posable syringes or pharmaceutical packa-
ges. When evaluating tiny leaks for liquids
it will be necessary to consider the wetabi-
lity of the surface of the solid and the capil-
lary action; see also Table 5.1. Some wide-
ly used leak detection methods are shown
together with the test gas, application range
and their particular features in Table 5.4.
5.5 Leak detectors and
how they work
Most leak testing today is carried out using
special leak detection devices. These can
detect far smaller leak rates than techni-
ques which do not use special equipment.
These methods are all based on using spe-
cific gases for testing purposes. The diffe-
rences in the physical properties of these
test gases and the gases used in real-life
applications or those surrounding the test
configuration will be measured by the leak
detectors. This could, for example, be the
differing thermal conductivity of the test
gas and surrounding air. The most widely
used method today, however, is the detec-
tion of helium used as the test gas.
The function of most leak detectors is
based on the fact that testing is conducted
with a special test gas, i.e. with a medium
other than the one used in normal operati-
on. The leak test may, for example, be car-
ried out using helium, which is detected
using a mass spectrometer, even though
the component being tested might, for
example, be a cardiac pacemaker whose
interior components are to be protected
against the ingress of bodily fluids during
normal operation. This example alone
makes it clear that the varying flow pro-
perties of the test and the working media
need to be taken into consideration.
5.5.1 Halogen leak detectors
(HLD4000, D-Tek)
Gaseous chemical compounds whose
molecules contain chlorine and/or fluorine
such as refrigerants R12, R22 and
R134a will influence the emissions of
alkali ions from a surface impregnated
with a mixture of KOH and Iron(III)hydro-
xide and maintained at 800 C to 900 C by
an external Pt heater. The released ions
flow to a cathode where the ion current is
measured and then amplified (halogen
diode principle). This effect is so great that
partial pressures for halogens can be mea-
sured down to 10
-7
mbar.
Whereas such devices were used in the
past for leak testing in accordance with the
vacuum method, today because of the
problems associated with the CFCs more
sniffer units are being built. The attainable
detection limit is about 1 10
-6
mbar l/s
for all the devisces. Equipment operating in
accordance with the halogen diode princi-
ple can also detect SF
6
. Consequently these
sniffer units are used to determine whether
refrigerants are escaping from a refrigera-
tion unit or from an SF
6
type switch box (fil-
led with arc suppression gas).
5.5.2 Leak detectors with mass
spectrometers (MSLD)
The detection of a test gas using mass
spectrometers is far and away the most
sensitive leak detection method and the one
most widely used in industry. The MS leak
Method Test gas Pressure range Smallest detectable
leak rate
Quantitative
measurement
mbar /s g/a R 134 a
Positive pressure
Positive pressure
Positive pressure
Vacuum
Positive pressure
Positive pressure
Positive pressure
Positive pressure
and vacuum
Positive pressure
(vacuum)
10
4
10
4
7
70
10
5
10
(10 )
6
5
10
10
12
7
10 10
3 5
10
2
10
3
10
2
70
7 10
1
7 10
1
7 10
1
7 10
(10 )
3
1
7 10
3
7 10
7 10
9
4
10 7
1
10
4
Air
Air and other
gases
Water
Air and other
gases
Helium
Refrigerants,
helium and
other gases
Substances
containing
halogens
Gases other
than air
Air and others
Air and others
Pressure
rise test
Pressure
drop test
Water pressure
test
Bubble test
Helium
leak detection
Universal
sniffer
leak detector
Halogen
leak detection
Thermal conducti-
vity leak detector
Ultrasonic
microphone
Foaming
liquids
Vacuum,
positive pressure
No
No
No
Yes
Yes
Yes
Yes
No
No
With
limitations
Table 5.4 Comparison of leak detection methods
D00 E 19.06.2001 21:39 Uhr Seite 110
Test gas
e.g. He
Test specimen
Leak detector
Exhaust
Test connection
Q = p S
He He eff
He
Leak Detection
detectors developed for this purpose make
possible quantitative measurement of leak
rates in a range extending across many
powers of ten (see Section 5.2) whereby
the lower limit 10
-12
mbar l/s, thus
making it possible to demonstrate the inhe-
rent gas permeability of solids where heli-
um is used as the test gas. It is actually
possible in principle to detect all gases
using mass spectrometry. Of all the availa-
ble options, the use of helium as a tracer
gas has proved to be especially practical.
The detection of helium using the mass
spectrometer is absolutely (!) unequivocal.
Helium is chemically inert, non-explosive,
non-toxic, is present in normal air in a con-
centration of only 5 ppm and is quite eco-
nomical. Two types of mass spectrometer
are used in commercially available MSLDs:
a) The quadrupole mass spectrometer, alt-
hough this is used less frequently due
to the more elaborate and complex
design (above all due to the electrical
supply for the sensor), or
b) the 180 magnetic sector field mass
spectrometer, primarily due to the rela-
tively simple design.
Regardless of the functional principle
employed, every mass spectrometer com-
prises three physically important sub-
systems: the ion source, separation
system and ion trap. The ions must be able
to travel along the path from the ion sour-
ce and through the separation system to
the ion trap, to the greatest possible extent
without colliding with gas molecules. This
path amounts to about 15 cm for all types
of spectrometers and thus requires a
medium free path length of at least 60 cm,
corresponding to pressure of about
1 10
-4
mbar; in other words, a mass
spectrometer will operate only in a vacu-
um. Due to the minimum vacuum level of
1 10
-4
mbar, a high vacuum will be requi-
red. Turbomolecular pumps and suitable
roughing pumps are used in modern leak
detectors. Associated with the individual
component groups are the required electri-
cal- and electronic supply systems and
software which, via a microprocessor,
allow for the greatest possible degree of
automation in the operating sequence,
including all adjustment and calibration
routines and measured value display.
5.5.2.1 The operating principle for a
MSLD
The basic function of a leak detector and
the difference between a leak detector and
mass spectrometer can be explained using
Figure 5.6. This sketch shows the most
commonly found configuration for leak
detection using the helium spray method
(see Section 5.7.1) at a vacuum compo-
nent. When the sprayed helium is drawn
into the component through a leak it is
pumped thorough the interior of the leak
detector to the exhaust, where it again lea-
ves the detector. Assuming that the detec-
tor itself is free of leaks, the amount of gas
flowing through each pipe section (at any
desired point) per unit of time will remain
constant regardless of the cross section
and the routing of the piping. The follo-
wing applies for the entry into the pumping
port at the vacuum pump:
Q = p S (5.4)
At all other points
Q = p S
eff
(5.4a)
applies, taking the line losses into account.
The equation applies to all gases which are
pumped through the piping and thus also
for helium.
Q
He
= p
He
S
eff, He
(5.4b)
In this case the gas quantity per unit of
time is the leak rate being sought; the total
pressure may not be used, but only the
share for helium or the partial pressure for
helium. This signal is delivered by the
mass spectrometer when it is set for ato-
mic number 4 (helium). The value for S
eff
is a constant for every series of leak detec-
tors, making it possible to use a micropro-
cessor to multiply the signal arriving from
the mass spectrometer by a numerical
constant and to have the leak rate dis-
played direct.
5.5.2.2 Detection limit, background,
gas storage in oil (gas
ballast), floating zero-point
suppression
The smallest detectable leak rate is dicta-
ted by the natural background level for the
gas to be detected. Even with the test
connector at the leak detector closed,
every gas will pass counter to the pum-
ping direction through the exhaust and
through the pumps (but will be reduced
accordingly by their compression) through
to the spectrometer and will be detected
there if the electronic means are adequate
to do so. The signal generated represents
the detection limit. The high vacuum
system used to evacuate the mass spec-
D00
Fig. 5.6 Basic operating principle for a leak detector
D00 E 19.06.2001 21:39 Uhr Seite 111
Leak Detection
trometer will normally comprise a turbo-
molecular pump and an oil-sealed rotary
vane pump. (Diffusion pumps were used
earlier instead of the turbomolecular
pumps.) Like every liquid, the sealing oil in
the rotary vane pump has the capability of
dissolving gases until equilibrium is rea-
ched between the gas dissolved in the oil
and the gas outside the oil. When the
pump is warm (operating temperature)
this equilibrium state represents the detec-
tion limit for the leak detector. The helium
stored in the oil thus influences the detec-
tion limit for the leak detector. It is possi-
ble for test gas to enter not only through
the test connection and into the leak detec-
tor; improper installation or inept handling
of the test gas can allow test gas to enter
through the exhaust and the airing or gas
ballast valve and into the interior of the
detector, to increase the helium level in the
oil and the elastomer seals there and thus
to induce a background signal in the mass
spectrometer which is well above the nor-
mal detection limit. When the device is
properly installed (see Fig. 5.7) the gas
ballast valve and the airing valve will be
connected to fresh air and the discharge
line (oil filter!) should at least be routed to
outside the room where the leak test takes
place.
An increased test gas (helium) background
level can be lowered by opening the gas
ballast valve and introducing gas which is
free of the test gas (helium-free gas, fresh
air). The dissolved helium will be flushed
out, so to speak. Since the effect always
affects only the part of the oil present in
the pump body at the particular moment,
the flushing procedure will have to be con-
tinued until all the oil from the pumps oil
pan has been recirculated several times.
This period of time will usually be 20 to 30
minutes.
In order to spare the user the trouble of
always having to keep an eye on the back-
ground level, what has been dubbed floa-
ting zero-point suppression has been inte-
grated into the automatic operating con-
cepts of all INFICON leak detectors (Sec-
tion 5.5.2.5). Here the background level
measured after the inlet valve has been
closed is placed in storage; when the valve
is then opened again this value will auto-
matically be deducted from subsequent
measurements. Only at a relatively high
threshold level will the display panel show
a warning indicating that the background
noise level is too high. Figure 5.8 is provi-
ded to illustrate the process followed in
zero point suppression. Chart on the left.
The signal is clearly larger than the back-
ground. Center chart: the background has
risen considerably; the signal can hardly
be discerned. Chart on the right: the back-
ground is suppressed electrically; the sig-
nal can again be clearly identified.
Independent of this floating zero-point
suppression, all the leak detectors offer
the capability for manual zero point shif-
ting. Here the display for the leak detector
at the particular moment will be reset to
zero so that only rises in the leak rate
from that point on will be shown. This ser-
ves only to facilitate the evaluation of a dis-
play but can, of course, not influence its
accuracy.
Modern leak detectors are being more fre-
quently equipped with oil-free vacuum
systems, the so-called dry leak detectors
(UL 200 dry, UL 500 dry). Here the pro-
blem of gas being dissolved in oil does not
occur but similar purging techniques will
nonetheless be employed.
5.5.2.3 Calibrating leak detectors;
test leaks
Calibrating a leak detector is to be under-
stood as matching the display at a leak
detector unit, to which a test leak is atta-
ched, with the value shown on the label
or calibration certificate. The prerequisite
for this is correct adjustment of the ion
paths in the spectrometer, also known as
tuning. Often the distinction is not made
quite so carefully and both procedures
Fig. 5.7 Correct set-up for a MSLD
Test connection
Venting
valve
Mass spectrometer
Turbomolecular pump
Roughing pump
Gas ballast
valve
Exhaust
Fig. 5.8 Example of zero-point suppression
10
6
10
7
10
8
109
10
10
10
11
caution
Equipment background level: < 2 10
10
1 10
8
1 10
10
(suppressed)
Leak: 2 10
8
2 10
8
2 10
8
Display: 2 10
8
3 10
8
2 10
8
prec
fine
D00 E 19.06.2001 21:39 Uhr Seite 112
QMA 200
external
particle
filter
flow
limiter 1
internal
particle
filter
flow divider 1
flow divider 2
flow meter
flow
limiter 2
flow
limiter 3
pv
1
2
Leak Detection
together are referred to as calibration.
In the calibration process proper the
straight-line curve representing the nume-
rically correct, linear correlation between
the gas flow per unit of time and the leak
rate is defined by two points: the zero point
(no display where no emissions are detec-
ted) and the value shown with the test leak
(correct display for a known leak).
In vacuum operations (spray technique,
see Section 5.7.1) one must differentiate
between two types of calibration: with an
internal or external test leak. When using a
test leak built into the leak detector the unit
can itself be calibrated but it can only cali-
brate itself. When using an external test
leak not just the device but also a comple-
te configuration, such as a partial flow
arrangement, can be included. Internal test
leaks are permanently installed and cannot
be misplaced. At present all the leak detec-
tors being distributed by INFICON are fit-
ted with an automatic calibration routine.
Sniffer units or configurations will as a rule
have to be calibrated with special, external
test leaks in which there is a guarantee that
on the one hand all the test gas issuing
from the test leak reaches the tip of the
probe and on the other hand that the gas
flow in the sniffer unit is not hindered by
calibration. When making measurements
using the sniffer technique (see Section
5.7.2) it is also necessary to take into
account the distance from the probe tip to
the surface of the specimen and the scan-
ning speed; these must be included as a
part of the calibration. In the special case
where helium concentration is being mea-
sured, calibration can be made using the
helium content in the air, which is a uni-
form 5 ppm world-wide.
Test leaks (also known as standard leaks
or reference leaks) normally comprise a
gas supply, a choke with a defined con-
ductance value, and a valve. The configu-
ration will be in accordance with the test
leak rate required. Figure 5.9 shows
various test leaks. Permeation leaks are
usually used for leak rates of
10
-10
< Q
L
< 10
-7
, capillaries, between
10
-8
and 10
-4
and, for very large leak rates
in a range from 10 to 1000 mbar l/s, pipe
sections or orifice plates with exactly
defined conductance values (dimensions).
Test leaks used with a refrigerant charge
represent a special situation since the ref-
rigerants are liquid at room temperature.
Such test leaks have a supply space for
liquid from which, through a shut-off
valve, the space filled only with the refrige-
rant vapor (saturation vapor pressure) can
be reached, ahead of the capillary leak.
One technological problem which is diffi-
cult to solve is posed by the fact that all
refrigerants are also very good solvents for
oil and grease and thus are often seriously
contaminated so that it is difficult to fill the
test leaks with pure refrigerant. Decisive
here is not only the chemical composition
but above all dissolved particles which can
repeatedly clog the fine capillaries.
5.5.2.4 Leak detectors with
quadrupole mass
spectrometer (Ecotec II)
INFICON builds leak detectors with qua-
drupole mass spectrometers to register
masses greater than helium. Apart from
special cases, these will be refrigerants.
These devices thus serve to examine the
tightness of refrigeration units, particular-
ly those for refrigerators and air conditio-
ning equipment.
Figure 4.2 shows a functional diagram for
a quadrupole mass spectrometer. Of the
four rods in the separation system, the two
pairs of opposing rods will have identical
potential and excite the ions passing
through along the center line so that they
oscillate transversely. Only when the
amplitude of these oscillations remains
smaller than the distance between the rods
can the appropriate ion pass through the
system of rods and ultimately reach the
ion trap, where it will discharge and thus
be counted. The flow of electrons thus
created in the line forms the measurement
signal proper. The other ions come into
contact with one of the rods and will be
neutralized there.
Figure 5.10 shows the vacuum schematic
for an Ecotec II. The mass spectrometer
D00
Fig. 5.9 Examples for the construction of test leaks
a b c d e
a Reference leak without gas supply, TL4, TL6
b Reference leak for sniffer and vacuum
applications, TL4-6
c (Internal) capillary test leak TL7
d Permeation (diffusion) reference leak, TL8
e Refrigerant calibrated leak
Fig. 5.10 Vacuum schematic for the Ecotec II
1 Diaphragm pump
2 Piezo-resistive
pressure sensor
3 Turbomolecular
pump
4 Quadrupole mass
spectrometer
5 Sniffer line
6 Gas flow limiter
7 Gas flow limiter
8 Gas flow meter
1
2
3
4
6
7
8
5
5
D00 E 19.06.2001 21:39 Uhr Seite 113
Leak Detection
(4) only operates under high vacuum con-
ditions, i.e. the pressure here must always
remain below 10
-4
mbar. This vacuum is
generated by the turbomolecular pump (3)
with the support of the diaphragm pump
(1). The pressure PV between the two
pumps is measured with a piezo resistive
measuring system (2) and this pressure
lies in the range between 1 to 4 mbar while
in the measurement mode. This pressure
must not exceed a value of 10 mbar as
otherwise the turbomolecular pump will
not be capable of maintaining the vacuum
in the mass spectrometer. The unit can
easily be switched over at the control unit
from helium to any of various refrigerants,
some of which may be selected as desired.
Naturally the unit must be calibrated sepa-
rately for each of these masses. Once set,
however, the values remain available in
storage so that after calibration has been
effected for all the gases (and a separate
reference leak is required for each gas!) it
is possible to switch directly from one gas
to another.
5.5.2.5 Helium leak detectors with
180 sector mass
spectrometer (UL 200, UL 500)
These units are the most sensitive and also
provide the greatest degree of certainty.
Here certain is intended to mean that
there is no other method with which one
can, with greater reliability and better sta-
bility, locate leaks and measure them
quantitatively. For this reason helium leak
detectors, even though the purchase price
is relatively high, are often far more eco-
nomical in the long run since much less
time is required for the leak detection pro-
cedure itself.
A helium leak detector comprises basically
two sub-systems in portable units and
three in stationary units. These are:
1. the mass spectrometer
2. the high vacuum pump and
3. the auxiliary roughing pump system in
stationary units.
The mass spectrometer (see Fig. 5.11)
comprises the ion source (14) and the
deflection system (59). The ion beam is
extracted through the orifice plate (5) and
enters the magnetic field (8) at a certain
energy level. Inside the magnetic field the
ions move along circular paths whereby
the radius for a low mass is smaller than
that for higher masses. With the correct
setting of the acceleration voltage during
tuning one can achieve a situation in which
the ions describe a circular arc with a defi-
ned curvature radius. Where mass 4 (heli-
um) is involved, they pass thorough the
aperture (9) to the ion trap (13). In some
devices the discharge current for the ions
impinging upon the total pressure electro-
des will be measured and evaluated as a
total pressure signal. Ions with masses
which are too small or too great should not
be allowed to reach the ion trap (13) at all,
but some of these ions will do so in spite
of this, either because they are deflected
by collisions with neutral gas particles or
because their initial energy deviates too far
from the required energy level. These ions
are then sorted out by the suppressor (11)
so that only ions exhibiting a mass of 4
(helium) can reach the ion detector (13).
The electron energy at the ion source is 80
eV. It is kept this low so that components
with a specific mass of 4 and higher
such as multi-ionized carbon or quadruply
ionized oxygen cannot be created. The
ion sources for the mass spectrometer are
simple, rugged and easy to replace. They
are heated continuously during operation
and are thus sensitive to contamination.
The two selectable yttrium oxide coated iri-
dium cathodes have a long service life.
These cathodes are largely insensitive to
air ingress, i.e. the quick-acting safety cut-
out will keep them from burning out even
if air enters. However, prolonged use of the
ion source may eventually lead to cathode
embrittlement and can cause the cathode
to splinter if exposed to vibrations or
shock.
Depending on the way in which the inlet is
connected to the mass spectrometer, one
can differentiate between two types of
MSLD.
1
2
3
4
5
6
7
8 9
10
11
12
13
14
Fig. 5.11 Configuration of the 180 sector mass spectrometer
1 Ion source flange
2 Cathode
(2 cathodes, Ir + Y
2
O
3
)
3 Anode
4 Shielding of the ion
source with discharge
orifice
10 Magnetic field
11 Suppressor
12 Shielding of the ion trap
13 Ion trap
14 Flange for ion trap with
preamplifier
5 Extractor
6 Ion traces for M > 4
7 Total pressure electrode
8 Ion traces for M = 4
9 Intermediate orifice
plate
D00 E 19.06.2001 21:39 Uhr Seite 114
Leak Detection
5.5.2.6 Direct-flow and counter-flow
leak detectors
Figure 5.12 shows the vacuum schematic
for the two leak detector types. In both
cases the mass spectrometer is evacuated
by the high vacuum pumping system com-
prising a turbomolecular pump and a
rotary vane pump. The diagram on the left
shows a direct-flow leak detector. Gas
from the inlet port is admitted to the spec-
trometer via a cold trap. It is actually equi-
valent to a cryopump in which all the
vapors and other contaminants condense.
(The cold trap in the past also provided
effective protection against the oil vapors
of the diffusion pumps used at that time).
The auxiliary roughing pump system ser-
ves to pre-evacuate the components to be
tested or the connector line between the
leak detector and the system to be tested.
Once the relatively low inlet pressure
(pumping time!) has been reached, the
valve between the auxiliary pumping
system and the cold trap will be opened for
the measurement. The Seff used in equati-
on 5.4b is the pumping speed of the tur-
bomolecular pump at the ion source loca-
tion:
Q
He
= p
He
S
eff,turbomolecular pump ion source
(5.5a)
In the case of direct-flow leak detectors, an
increase in the sensitivity can be achieved
by reducing the pumping speed, for exam-
ple by installing a throttle between the tur-
bomolecular pump and the cold trap. This
is also employed to achieve maximum
sensitivity. To take an example:
The smallest detectable partial pressure
for helium is
p
min,He
= 1 10
-12
mbar. The pumping
speed for helium would be S
He
= 10 l/s.
Then the smallest detectable leak rate is
Q
min
= 1 10
-12
mbar 10 l/s
= 1 10
-11
mbar l/s. If the pumping speed
is now reduced to 1/s, then one will achie-
ve the smallest detectable leak rate of
1 10
-12
mbar l/s. One must keep in
mind, however, that with the increase in
the sensitivity the time constant for achie-
ving a stable test gas pressure in the test
specimen will be correspondingly larger
(see Section 5.5.2.9).
In Figure 5.12 the right hand diagram
shows the schematic for the counter-flow
leak detector. The mass spectrometer, the
high vacuum system and also the auxiliary
roughing pump system correspond exact-
ly to the configuration for the direct-flow
arrangement. The feed of the gas to be
examined is however connected between
the roughing pump and the turbomolecu-
lar pump. Helium which reaches this
branch point after the valve is opened will
cause an increase in the helium pressure
in the turbomolecular pump and in the
mass spectrometer. The pumping speed
S
eff
inserted in equation 5.4b is the pum-
ping speed for the rotary vane pump at the
branch point. The partial helium pressure
established there, reduced by the helium
compression factor for the turbomolecular
pump, is measured at the mass spectro-
meter. The speed of the turbomolecular
pump in the counter-flow leak detectors is
regulated so that pump compression also
remains constant. Equation 5.5b is derived
from equation 5.5a:
Q
He
= p
He
S
eff
K (5.5b)
S
eff
= effective pumping speed at the
rotary vane pump at the
branching point
K = Helium compression factor at
the turbomolecular pump
The counter-flow leak detector is a particu-
lar benefit for automatic vacuum units
since there is a clearly measurable pressu-
re at which the valve can be opened, name-
ly the roughing vacuum pressure at the
turbomolecular pump. Since the turbo-
molecular pump has a very large compres-
sion capacity for high masses, heavy mo-
lecules in comparison to the light test gas,
helium (M = 4), can in practice not reach
the mass spectrometer. The turbomolecu-
lar pump thus provides ideal protection for
the mass spectrometer and thus elimina-
tes the need for an LN
2
cold tap, which is
certainly the greatest advantage for the
user. Historically, counter-flow leak detec-
tors were developed later. This was due in
part to inadequate pumping speed stabili-
ty, which for a long time was not sufficient
with the rotary vane pumps used here. For
both types of leak detector, stationary
units use a built-in auxiliary pump to assist
in the evacuation of the test port. With por-
table leak detectors, it may be necessary to
provide a separate, external pump, this
being for weight reasons.
5.5.2.7 Partial flow operation
Where the size of the vacuum vessel or the
leak makes it impossible to evacuate the
test specimen to the necessary inlet pres-
sure, or where this would simply take too
long, then supplementary pumps will have
to be used. In this case the helium leak
detector is operated in accordance with the
so-called partial flow concept. This
means that usually the larger part of the
gas extracted from the test object will be
removed by an additional, suitably dimen-
sioned pump system, so that only a part of
the gas stream reaches the helium leak
detector (see Fig. 5.13). The splitting of
the gas flow is effected in accordance with
the pumping speed prevailing at the bran-
ching point. The following then applies:
Q
Vacuum vessel
= Display
Leak detector
(5.6)
where g is characterized as the partial flow
ratio, i.e. that fraction of the overall leak
current which is displayed at the detector.
D00
Test specimen Test specimen
Test gas stream Test gas stream
p p LN
p
p
High vacuum pump
High vacuum pump
Roughing pump Roughing pump
Cold trap:
S = 6.1 /s cm
Fl 1000 cm
S = 6,100 /s
2
2
Auxiliary pump Auxiliary pump

< 10 mbar
4
< 10 mbar
4
He He 2
TOT
TOT
MS MS
Fig. 5.12 Full-flow and counter-flow leak detector
Solution 1: Direct-flow leak detector Solution 2: Counter-flow leak detector
D00 E 19.06.2001 21:39 Uhr Seite 115
Leak Detection
Where the partial flow ratio is unknown, g
can be determined with a reference leak
attached at the vacuum vessel:
Display at the leak detector
= (5.7)
Q
L
for the reference leak
5.5.2.8 Connection to vacuum systems
The partial flow concept is usually used in
making the connection of a helium leak
detector to vacuum systems with multi-
stage vacuum pump sets. When conside-
ring where to best make the connection, it
must be kept in mind that these are usual-
ly small, portable units which have only a
low pumping speed at the connection flan-
ge (often less than 1 l/s). This makes it all
the more important to estimate based on
the partial flow ratio to be expected vis
vis a diffusion pump with pumping speed
of 12000 l/s, for example which leak
rates can be detected at all. In systems
with high vacuum- and Roots pumps, the
surest option is to connect the leak detec-
tor between the rotary vane pump and the
roots pump or between the roots pump
and the high vacuum pump. If the pressu-
re there is greater than the permissible
inlet pressure for the leak detector, then
the leak detector will have to be connected
by way of a metering (variable leak) valve.
Naturally one will have to have a suitable
connector flange available. It is also advis-
able to install a valve at this point from the
outset so that, when needed, the leak
detector can quickly be coupled (with the
system running) and leak detection can
commence immediately after opening the
valve. In order to avoid this valve being
opened inadvertently, it should be sealed
off with a blank flange during normal vacu-
um system operation.
A second method for coupling to larger
systems, for example, those used for
removing the air from the turbines in po-
wer generating stations, is to couple at the
discharge. A sniffer unit is inserted in the
system where it discharges to atmosphe-
re. One then sniffs the increase in the heli-
um concentration in the exhaust. Without
a tight coupling to the exhaust, however,
the detection limit for this application will
be limited to 5 ppm, the natural helium
content in the air. In power plants it is suf-
ficient to insert the tip of the probe at an
angle of about 45 from the top into the
discharge line (usually pointing upward) of
the (water ring) pump.
5.5.2.9 Time constants
The time constant for a vacuum system is
set by
(5.8)
= Time constant
V = Volume of the container
S
eff
= Effective pumping speed, at the
test object
Figure 5.14 shows the course of the signal
after spraying a leak in a test specimen
attached to a leak detector, for three diffe-
rent configurations:
1. Center: The specimen with volume of V
is joined directly with the leak detector
LD (effective pumping speed of S).
2. Left: In addition to 1, a partial flow
pump with the same effective pumping
speed, S
= S, is attached to the test

specimen.
3. Right: As at 1, but S is throttled down to
0.5 S.
The signals can be interpreted as follow:
1: Following a dead period (or delay
time) up to a discernible signal level, the
signal, which is proportional to the partial
pressure for helium, will rise to its full
value of p
He
= Q/S
eff
in accordance with
equation 5.9:
(5.9 )
p
Q
S
e
He
eff
t
=
_
,
1

=
V
S
eff
mbar
s

mbar
s

150
24.66
m
V
S
eff
mbar
s

mbar
s

mbar
s

8
(8 + 16.66)
16,66
(8 + 16.66)
8
mbar
s

s
m
s
3
Partial flow principle (example)
V = 150
S = 60 = 16.66 Partial flow pump (PFP)
PFP
S = S + S
eff PFP LD
S = 8 Leak detector (LD)
LD
Q = 3 10 (Leak rate)
He
5
A)
Signal to Leak detector: 3 10 = 9.73 10
5 6
Signal topartial flowpump: 3 10 = 2.02 10
5 5
Check: Overall signal Q = Q + Q = 3.00 10
He LD PFP
5
... Partial flow ratio
Overal pumping speed: S = S + S = 8 + 16.66 = 24.66
eff LD PFP
Signal amplitude:
Splitting of the gas flow (also of the test
gas!) in accordance with the effective pumping
speed at the partial flow branch point
Partial flow ratio = Fraction of the overall flow to the leak detector
B) Response time: t = 3 = 3 = 18.25 s
95%
Estimate: Value for S, V and are uncertain certain: calibrate with reference leak
= = =
= = or
Q
Q
LD
He
Q
Q + Q
LD
LD PFP
S
S + S
LD
LD PFP
1
1 + nnn
1
1 + nnn
Q
Q
PFP
LD
S
S
PFP
LD
Q = Q
LD He
Display Leak rate

Fig. 5.13 Partial flow principle
D00 E 19.06.2001 21:39 Uhr Seite 116
Leak Detection
The signal will attain a prortion of its ulti-
mate value after
t = 1 . . 63.3 % t = 2 . . 86.5 %
t = 3 . . 95.0 % t = 4 . . 98.2 %
t = 5 . . 99.3 % t = 6 . . 99.8 %
The period required to reach 95 % of the
ultimate value is normally referred to as
the response time.
2: With the installation of the partial flow
pump both the time constant and the sig-
nal amplitude will be reduced by a factor of
2; that means a quicker rise but a signal
which is only half as great. A small time
constant means quick changes and thus
quick display and, in turn, short leak detec-
tion times.
3: The throttling of the pumping speed to
0.5 S, increases both the time constant
and the signal amplitude by a factor of 2. A
large value for t thus increases the time
required appropriately. Great sensitivity,
achieved by reducing the pumping speed,
is always associated with greater time
requirements and thus by no means is
always of advantage.
An estimate of the overall time constants
for several volumes connected one behind
to another and to the associated pumps
can be made in an initial approximation by
adding the individual time constants.
5.6 Limit values /
Specifications for
the leak detector
1. The smallest detectable leak rate.
2. The effective pumping speed at the
test connection.
3. The maximum permissible pressure
inside the test specimen (also the
maximum permissible inlet pressure).
This pressure p
max
will be about 10
-1
for LDs with classical PFPs and about 2
to 10 mbar for LDs with compound
PFPs. The product of this maximum
permissible operating pressure and the
pumping speed S of the pump system
at the detectors test connection is the
maximum permissible throughput:
Q
max
= p
max
S
eff, connector
(5.10)
This equation shows that it is by no means
advantageous to attain high sensitivity by
throttling down the pumping speed. The
maximum permissible throughput would
otherwise be too small. The unit is not fun-
ctional when either due to one large leak
or several smaller leaks more gas flows
into the unit than the maximum permissi-
ble throughput rate for the leak detector.
5.7 Leak detection
techniques using
helium leak
detectors
5.7.1 Spray technique
(local leak test)
The test specimen, connected to the heli-
um leak detector, is slowly traced with a
very fine stream of helium from the spray
pistol, aimed at likely leakage points (wel-
ding seams, flange connectors, fused
joints), bearing in mind the time constant
of the system as per Equation 5.8 (see Fig.
5.14). The volume sprayed must be adju-
sted to suit the leak rate to be detected and
the size and accessibility of the object
being tested. Although helium is lighter
than air and therefore will collect beneath
the ceiling of the room, it will be so well
distributed by drafts and turbulence indu-
ced by movements within the room that
one need not assume that helium will be
found primarily (or only) at the top of the
room during search for leaks. In spite of
this, it is advisable, particularly when dea-
ling with larger components, to start the
search for leaks at the top.
In order to avoid a surge of helium when
the spray valve is opened (as this would
contaminate the entire environment) it is
advisable to install a choke valve to adjust
the helium quantity, directly before or after
the spray pistol (see Fig. 5.15). The correct
quantity can be determined easiest by sub-
merging the outlet opening in a container
of water and setting the valve on the basis
of the rising bubbles. Variable-area flow-
meters are indeed available for the requi-
red small flow quantities but are actually
too expensive. In addition, it is easy to use
the water-filled container at any time to
determine whether helium is still flowing.
The helium content of the air can also be
detected with helium leak detectors where
large leaks allow so much air to enter the
vessel that the 5 ppm share of helium in
the air is sufficient for detection purposes.
The leak rate is then:
Display (pure He) Display (atmosph. He)
=
1 5 10
-6
Q
L
= Display (pure He) (5.11)
= 2 10
+5
Display (atmospheric He)
D00
100%
100%
1,0
2,0
0,5
0
100%
95%
95%
95%
to
Dead time
3 3 3 =
( = ... Time constant)
Time
2 (3 ) (3 ) =
S
2 /
Signal
amplitude
Q
S
Q
= 2p
= p
Signal rise
Q
S
1
2
V
S
V
S
V V
S
V
S
V
S
P
2 /
Q
S + S
=
S
2 /
V
S + S
Compensation period, e.g. t = 3 = 3
95%

Slower, more sensitive Faster, less sensitive normal

S S S
S
2 /
MS MS MS
Throttle
Q
V V V
Q Q
LD LD LD
Fig. 5.14 Signal responses and pumping speed
D00 E 19.06.2001 21:39 Uhr Seite 117
Leak Detection
5.7.2 Sniffer technology
(local leak test using the
positive pressure
method)
Here the points suspected of leaking at the
pressurized test specimen (see Fig. 5.4, d)
are carefully traced with a test gas probe
which is connected with the leak detector
by way of a hose. Either helium or hydro-
gen can be detected with the INFICON heli-
um leak detectors. The sensitivity of the
method and the accuracy of locating leaky
points will depend on the nature of the
sniffer used and the response time for the
leak detector to which it is connected. In
addition, it will depend on the speed at
which the probe is passed by the leak
points and the distance between the tip of
the probe and the surface of the test spe-
cimen. The many parameters which play a
part here make it more difficult to determi-
ne the leak rates quantitatively. Using snif-
fer processes it is possible, virtually inde-
pendent of the type of gas, to detect leak
rates of about 10
-7
mbar l/s. The limitati-
on of sensitivity in the detection of helium
is due primarily to the helium in the atmos-
phere (see Chapter 9, Table VIII). In regard
to quantitative measurements, the leak
detector and sniffer unit will have to be
calibrated together. Here the distance from
the specimen and the tracing speed will
have to be included in calibration, too.
5.7.3 Vacuum envelope test
(integral leak test)
Vacuum envelope tests are integral leak
tests using helium as the test gas, in which
the test specimen is enclosed either in a
rigid (usually metal) enclosure or in a light
plastic envelope. The helium which enters
or leaves (depending on the nature of the
test) the test specimen is passed to a heli-
um leak detector, where it is measured.
Envelope tests are made either with the
test specimen pressurized with helium
(Fig. 5.4c) or with the test specimen eva-
cuated (Fig. 5.4a). In both cases it may be
necessary to convert the helium enrich-
ment figure (accumulation) to the helium
standard leak rate.
5.7.3.1 Envelope test test specimen
pressurized with helium
a) Envelope test with concentration
measurement and subsequent leak
rate calculation
To determine overall leakiness of a test
object pressurized with helium the object
shall be enclosed in an envelope which is
either rigid or deformable (plastic). The
test gas leaving the leaks accumulates so
that the helium concentration in the enve-
lope rises. Following an enrichment period
to be determined (operating period) the
change in concentration inside the envelo-
pe will be measured with a sniffer connec-
ted to the helium detection unit. The over-
all leak rate (integral leak rate) can be cal-
culated following calibration of the test
configuration with a reference concen-
tration, e.g. atmospheric air. This method
makes it possible to detect even the smal-
lest overall leakiness and is suitable in par-
ticular for automated industrial leak
testing. Due to gas accumulation, the
limits for normal sniffer techniques are
shifted toward lower leak rates and the
ambient conditions such as temperature,
air flow and sniffer tracing speed lose
influence. When using plastic envelopes it
is necessary to take into account helium
permeation through the plastic envelope
during long enrichment periods.
b) Direct measurement of the leak rate
with the leak detector (rigid envelope)
When the test specimen, pressurized with
helium, is placed in a rigid vacuum cham-
ber, connected to a helium leak detector,
the integral leak rate can be read directly at
the leak detector.
5.7.3.2 Envelope test with test
specimen evacuated
a) Envelope = plastic tent
The evacuated test specimen is surroun-
ded by a light-weight (plastic) enclosure
and this is then filled with helium once the
atmospheric air has been removed. When
using a plastic bag as the envelope, the
bag should be pressed against the test
specimen before filling it with helium in
order to expel as much air as possible and
to make the measurement with the purest
helium charge possible. The entire outside
surface of the test object is in contact with
the test gas. If test gas passes through
leaks and into the test specimen, then the
Fig. 5.15 Helium spray equipment
Avoiding the helium surge when the pistol valve is opened
a) Throttle hose or
b) Adjustable throttle valve ahead of the spray pistol
Minimum helium quantity for correct display: Changing the setting for the throttle shall not affect
indication.
The minimum quantity is always much smaller than one would set without a flowmeter (e.g. by
listening for flow or letting the helium flow across moistened lips). The simplest check without a
flowmeter: Letting gas bubble through water.
D00 E 19.06.2001 21:39 Uhr Seite 118
Leak Detection
integral leak rate will be indicated, regard-
less of the number of leaks. In addition, it
is necessary to observe when repeating
testing in enclosed areas that the helium
content of the room will rise quite rapidly
when the envelope is removed. Using pla-
stic bags is thus more advisable for one-
off testing of large plants. The plastic
envelope used here is often referred to as
a tent.
b) Rigid envelope
The use of a solid vacuum vessel as the
rigid envelope, on the other hand, is better
for repetitive testing where an integral test
is to be made. When solid envelopes are
used it is also possible to recover the heli-
um once the test has been completed.
5.7.4 Bombing test,
Storage under pressure
The bombing test is use to check the
tightness of components which are alrea-
dy hermetically sealed and which exhibit a
gas-filled, internal cavity. The components
to be examined (e.g. transistors, IC hou-
sings, dry-reed relays, reed contact swit-
ches, quartz oscillators, laser diodes and
the like) are placed in a pressure vessel
which is filled with helium. Operating with
the test gas at relatively high pressure (5
to 10 bar) and leaving the system standing
over several hours the test gas (helium)
will collect inside the leaking specimens.
This procedure is the actual bombing. To
make the leak test, then, the specimens are
placed in a vacuum chamber following
bombing, in the same way as described
for the vacuum envelope test. The overall
leak rate is then determined. Specimens
with large leaks will, however, lose their
test gas concentration even as the vacuum
chamber is being evacuated, so that they
will not be recognized as leaky during the
actual leak test using the detector. It is for
this reason that another test to register
very large leaks will have to be made prior
to the leak test in the vacuum chamber.
5.8 Industrial leak
testing
Industrial leak testing using helium as the
test gas is characterized above all by the fact
that the leak detection equipment is fully
integrated into the manufacturing line. The
design and construction of such test units
will naturally take into account the task to be
carried out in each case (e.g. leak testing
vehicle rims made of aluminum or leak
testing for metal drums). Mass-produced,
standardized component modules will be
used wherever possible. The parts to be
examined are fed to the leak testing system
(envelope test with rigid envelope and positi-
ve pressure [5.7.3.1b] or vacuum [5.7.3.2b]
inside the specimen) by way of a conveyor
system. There they will be examined indivi-
dually using the integral methods and auto-
matically moved on. Specimens found to be
leaking will be shunted to the side.
The advantages of the helium test method,
seen from the industrial point of view, may
be summarized as follows:
The leak rates which can be detected
with this process go far beyond all prac-
tical requirements.
The integral leak test, i.e. the total leak
rate for all individual leaks, facilitates
the detection of microscopic and spon-
ge-like distributed leaks which alto-
gether result in leakage losses similar
to those for a larger individual leak.
The testing procedure and sequence
can be fully automated.
The cyclical, automatic test system
check (self-monitoring) of the device
ensures great testing reliability.
Helium is non-toxic and non-hazardous
(no maximum allowable concentrations
need be observed).
Testing can be easily documented, indi-
cating the parameters and results, on a
printer.
Use of the helium test method will result in
considerable increases in efficiency (cyc-
ling times being only a matter of seconds
in length) and lead to a considerable
increase in testing reliability. As a result of
this and due to the EN/ISO 9000 require-
ments, traditional industrial test methods
(water bath, soap bubble test, etc.) will
now largely be abandoned.
D00
D00 E 19.06.2001 21:39 Uhr Seite 119
Thin Film Controllers / Control Units
6 Thin film
controllers and
control units
with quartz
oscillators
6.1 Introduction
It took a long time to go from the coating
of quartz crystals for frequency fine
tuning, which has long been in practice, to
utilization of frequency change to determi-
ne the mass per unit area as a microbalan-
ce with the present-day degree of precisi-
on. In 1880 two brothers, J. and P. Curie,
discovered the piezoelectric effect. Under
mechanical loads on certain quartz crystal
surfaces, electrical charges occur that are
caused by the asymmetrical crystalline
structure of SiO
2
. Conversely, in a piezo-
crystal deformations appear in an electri-
cal field and mechanical oscillations occur
in an alternating field. A distinction is
made between bending oscillations, thick-
ness shear mode and thickness shear
oscillations. Depending on the orientation
of the cut plane to the crystal lattice, a
number of different cuts are distinguished,
of which only the so-called AT cut with a
cut angle of 3510" is used in thin film con-
trollers because the frequency has a very
low temperature dependence in the range
between 0 and 50 C with this cut. Accor-
dingly, an attempt must be made not to
exceed this temperature range during coa-
ting (water cooling of crystal holder).
Since there is still a problem with quartz
capacity (i.e. the maximum possible coa-
ting thickness of the quartz at which it still
oscillates reliably) despite refined techno-
logy, a number of approaches have been
developed to expand this capacity:
1. The use of several crystals, one behind
the other, in a multiple crystal holder
with automatic change and data upda-
ting in the event of imminent failure of a
quartz: CrystalSix.
2. The RateWatcher function, in which the
quartz is alternately exposed to the coa-
ting beam for a short time until all mea-
surements and regulation have been
carried out and then remains covered
by a shutter for a longer period of time.
The selection of the right crystal holder
thus plays an important role in all measu-
rements with quartz oscillators. Various
crystal holder designs are recommended
for the different applications: with or with-
out shutter, bakeable for UHV, double
crystal holder or crystal six as well as spe-
cial versions for sputter applications. In
addition to these important and more
mechanical aspects, the advances in
measuring and control technology and
equipment features will be discussed in
the following.
6.2 Basic principles of
coating thickness
measurement with
quartz oscillators
The quartz oscillator coating thickness
gauge (thin film controller) utilizes the pie-
zoelectric sensitivity of a quartz oscillator
(monitor crystal) to the supplied mass.
This property is utilized to monitor the
coating rate and final thickness during
vacuum coating.
A very sharp electromechanical resonance
occurs at certain discrete frequencies of
the voltage applied. If mass is added to the
surface of the quartz crystal oscillating in
resonance, this resonance frequency is
diminished. This frequency shift is very
reproducible and is now understood preci-
sely for various oscillation modes of
quartz. Today this phenomenon, which is
easy to understand in heuristic terms, is
an indispensable measuring and process
control tool, with which a coating increase
of less than one atomic layer can be detec-
ted.
In the late 1950s Sauerbrey and Lostis dis-
covered that the frequency shift connected
with the coating of the quartz crystal is a
function of the change in mass due to the
coating material in the following way:
(6.1)
M
f
mass of the coating
M
q
mass of the quartz prior to coating
F
q
frequency prior to coating
F
c
frequency after coating
F = Ff Fc ... frequency shift due to coa-
ting
If the following are now applied:
M
f
= (M
c
M
q
) = D
f

f
A and
Mq = Dq rq A, where T = the coating
thickness, = density and A stands for
area while the index q stands for the state
of the uncoated quartz and c for the state
after frequency shift due to coating, the
following results are obtained for the coa-
ting thickness:
with
where
N = F
q
D
q
is the frequency constant (for
the AT cut N
AT
= 166100 Hz cm) and
q
= 2.649 g/cm
3
is the density of the
quartz. The coating thickness is thus pro-
portional to the frequency shift F and
inversely proportional to the density f of
the coating material. The equation
for the coating thickness was used in the
first coating thickness measuring units
with frequency measurement ever used.
According to this equation, a crystal with a
starting frequency of 6.0 MHz displays a
decline in frequency of 2.27 Hz after coa-
ting with 1 of aluminum
(de = 2.77 g/cm
3
). In this way the growth
of a fixed coating due to evaporation or
sputtering can be monitored through pre-
cise measurement of the frequency shift of
the crystal. It was only when knowledge of
the quantitative interrelationship of this
D K
F
f
f
=
K
D F
F
N
F
q q q
q
AT q
q
= =

2 2
D
F
F
D
F
F
K
F
f
q
q
q q
q f
f
= =

M
M
M
M or with
F F
F F
q
q
f
f
q q

Fig. 6.1 Natural frequency as a function of temperature
in an AT cut quartz crystal
Temperature (C)
r
e
l
.

f
r
e
q
u
e
n
c
y

c
h
a
n
g
e

(
p
p
m
)
D00 E 19.06.2001 21:40 Uhr Seite 120
effect was acquired that it became possible
to determine precisely the quantity of
material that is deposited on a substrate in
a vacuum. Previously this had been prac-
tically impossible.
6.3 The shape of
quartz oscillator
crystals
Regardless of how sophisticated the elec-
tronic environment is, the main compo-
nent for coating measurement remains the
monitor quartz crystal. Originally monitor
quartzes had a square shape. Fig. 6.4
shows the resonance spectrum of a quartz
resonator with the design used today (Fig.
6.3). The lowest resonance frequency is
initially given by a thickness shear oscilla-
tion, which is called the fundamental wave.
The characteristic motions of the thickness
shear oscillation are parallel to the main
crystal boundary surfaces. In other words:
the surfaces are shift antinodes, see Fig.
6.2. The resonance frequencies slightly
above the basic frequency are called
anharmonic and are a combination of
thickness shear and thickness rotation
oscillation forms. The resonance frequen-
cy at around three times the value for the
fundamental wave is called quasi-harmo-
nic. Near the quasi-harmonic there are
also a number of anharmonics with a
slightly higher frequency.
The design of the monitor crystals used
nowadays (see Fig. 6.3) displays a number
of significant improvements over the origi-
nal square crystals. The first improvement
was the use of round crystals. The enlar-
ged symmetry greatly reduced the number
of possible oscillation modes. A second
group of improvements involved providing
one of the surfaces with a contour and
making the excitation electrode smaller.
The two together ensure that the acoustic
energy is recorded. Reducing the electrode
diameter limits the excitation to the midd-
le area. The surface contour consumes the
energy of the moving acoustic waves befo-
re they reach the crystal edge. It is not
reflected into the center where it could
interfere with new incoming waves.
Such a small crystal behaves like an infini-
tely expanded crystal. However, if the
crystal vibrations remain restricted to the
center, one can clamp the outer edge to a
crystal holder, without engendering unde-
sired side effects. Moreover, contouring
reduces the resonance intensity of undesi-
red anharmonics. This limits the capacity
of the resonator to maintain these oscilla-
tions considerably.
Use of an adhesive coating has enhanced
the adhesion of the quartz electrode. Even
the rate spikes occurring with increasing
film stress (strain) and caused by micro-
tears in the coating were reduced. Coating
material remains at these micro-tears
without adhesion and therefore cannot
oscillate. These open areas are not registe-
red and thus an incorrect thickness is indi-
cated.
Fig. 6.4 shows the frequency behavior of a
quartz crystal shaped as in Fig. 6.3. The
ordinate represents the amplitude of the
oscillation or also the current flowing
through the crystal as a function of the fre-
quency on the abscissa.
Usually an AT cut is chosen for the coating
thickness measurement because through
the selection of the cut angle the frequen-
cy has a very small temperature coefficient
at room temperature.
Since one cannot distinguish between
D00
Fig. 6.2 Thickness shear oscillations
Node
E
Fig. 6.3 Shape of INFICON quartz crystals

Fig. 6.4 Frequency resonance spectrum
Frequency (MHz)
l
o
g

R
e
l
a
t
i
v
e

i
n
t
e
n
s
y
D00 E 19.06.2001 21:40 Uhr Seite 121
coating: frequency reduction = negati-
ve influence
temperature change:
negative or positive influence
temperature gradients on the crystal,
positive or negative
stresses caused by the coating
it is important to minimize the temperatu-
re influence. This is the only way to mea-
sure small differences in mass.
6.4 Period
measurement
Although the instruments that functioned
according to equation 6.2 were very use-
ful, it soon became obvious that for the
desired accuracy their area of application
was typically limited to F < 0.02 F
q
. Even
at a relative frequency change of
(F
q
F
c
) / F
q
< 2 %, errors of around 2 %
occurred in the coating thickness measu-
rement so that the "usable service life" of
the coating in the case of a 6-MHz monitor
crystal was about 120 kHz.
In 1961 Behrndt discovered that:
with (6.3)
T
c
= 1 / F
c
... oscillation period, coated
T
q
= 1 / F
q
... oscillation period, uncoated
The period measurement (measurement of
the oscillation duration) was the result of
the introduction of digital time measure-
ment and the discovery of the proportiona-
lity of crystal thickness D
q
and oscillation
duration T
q
. The necessary precision of
thickness measurement permits applicati-
on of equation 6.3 up to about
F < 0.05 F
q
.
In period measurement a second crystal
oscillator is essentially used as a reference
oscillator that is not coated and usually
oscillates at a much higher frequency than
the monitor crystal. The reference oscilla-
tor generates small precision time inter-
vals, with which the oscillation duration of
the monitor crystal is determined. This is
done by means of two pulse counters: the
first counts a fixed number of monitor
oscillations m. The second is started si-
M
M
T T
T
F
F
f
q
c q
q c
( )
multaneously with the first and counts the
oscillations of the reference crystal during
m oscillations of the monitor crystal.
Because the reference frequency F
r
is
known and stable, the time for m monitor
oscillations can be determined accurately
to 2/Fr. The monitor oscillation period is
then
where n is the reading of the reference
counter. The accuracy of the measurement
is determined by the frequency of the refe-
rence oscillator and the length of the coun-
ting time that is specified through the size
of m.
For low coating rates, small densities of
the coating material and fast measure-
ments (that require short counting times),
it is important to have a reference oscilla-
tor with a high frequency. All of this requi-
res great time precision so that the small
coating-related frequency shifts can be
resolved. If the frequency shift of the mo-
nitor crystal decreases between two mea-
surements on the order of magnitude of
the frequency measurement accuracy,
good rate regulation becomes impossible
(rate regulation: regulation of the energy
supply to the coating source so that a spe-
cified coating thickness growth per time
unit is maintained). The great measure-
ment uncertainty then causes more noise
in the closed loop, which can only be
countered with longer time constants. This
in turn makes the corrections due to
system deviation slow so that relatively
long deviations from the desired rate
result. This may not be important for sim-
ple coatings, but for critical coatings, as in
the case of optical filters or very thin,
slowly growing single-crystal coatings,
errors may result. In many cases, the desi-
red properties of such coatings are lost if
the rate deviations are more than one or
two percent. Finally, frequency and stabili-
ty of the reference oscillator determine the
precision of the measurement.
n
F
r
m
6.5 The Z match
technique
Miller and Bolef (1968) treated the quartz
oscillator and coating system as a single-
dimensional, coherent acoustic resonator.
Lu and Lewis (1972) developed the simpli-
fied Z match equation on that basis. Simul-
taneous advances in electronics, particu-
larly the microprocessor, made it possible
to solve the Z match equation in real time.
Most coating process control units sold
today use this sophisticated equation,
which takes into account the acoustic pro-
perties of the quartz oscillator/coating
system:
(6.4)
acoustic impedance ratio
U
q
= shear module, quartz
U
f
= shear module, film
This led to basic understanding of the con-
version of frequency shift into thickness
which enabled correct results in a practical
time frame for process control. To achieve
this high degree of accuracy, the user
must only enter an additional material
parameter Z
f
for the coating material. The
validity of the equation was confirmed for
many materials and it applies to frequency
shifts up to F < 0.4 F
q
! Note that equati-
on 6.2 was only valid up to F < 0.02 F
q
.
And equation 6.3 only up to F < 0.05 F
q
.
6.6 The active
oscillator
All units developed up to now are based on
use of an active oscillator, as shown sche-
matically in Fig. 6.5. This circuit keeps the
crystal actively in resonance so that any
type of oscillation duration or frequency
measurement can be carried out. In this
type of circuit the oscillation is maintained
as long as sufficient energy is provided by
the amplifier to compensate for losses in
the crystal oscillation circuit and the
crystal can effect the necessary phase
shift. The basic stability of the crystal
Z
d U
d U
q q
f f
( )
T
N d
d F Z
arctg Z tg
F F
F
f
AT q
f c
q c
q
_
,
1
]
1
1
_
,
D00 E 19.06.2001 21:40 Uhr Seite 122

oscillator is created through the sudden
phase change that takes place near the
series resonance point even with a small
change in crystal frequency, see Fig. 6.6.
Normally an oscillator circuit is designed
such that the crystal requires a phase shift
of 0 degrees to permit work at the series
resonance point. Long- and short-term
frequency stability are properties of crystal
oscillators because very small frequency
differences are needed to maintain the
phase shift necessary for the oscillation.
The frequency stability is ensured through
the quartz crystal, even if there are long-
term shifts in the electrical values that are
caused by phase jitter due to temperatu-
re, ageing or short-term noise. If mass is
added to the crystal, its electrical pro-
perties change.
Fig. 6.7 shows the same graph as Fig 6.6,
but for a thickly coated crystal. It has lost
the steep slope displayed in Fig. 6.6.
Because the phase rise is less steep, any
noise in the oscillator circuit leads to a lar-
ger frequency shift than would be the case
with a new crystal. In extreme cases, the
original phase/frequency curve shape is
not retained; the crystal is not able to carry
out a full 90 phase shift.
The impedance Z can increase to very
high values. If this happens, the oscillator
prefers to oscillate in resonance with an
anharmonic frequency. Sometimes this
condition is met for only a short time and
the oscillator oscillation jumps back and
forth between a basic and an anharmonic
oscillation or it remains as an anharmonic
oscillation. This phenomenon is well
known as mode hopping. In addition to
the noise of the rate signal created, this
may also lead to incorrect termination of a
coating because of the phase jump. It is
important here that, nevertheless, the con-
troller frequently continues to work under
these conditions. Whether this has occur-
red can only be ascertained by noting that
the coating thickness is suddenly signifi-
cantly smaller, i.e. by the amount of the
frequency difference between the funda-
mental wave and the anharmonic adopted
by the oscillation.
6.7 The mode-lock
oscillator
INFICON has developed a new technology
for overcoming these constraints on the
active oscillator. The new system con-
stantly analyzes the response of the crystal
to an applied frequency: not only to deter-
mine the (series) resonance frequency, but
also to ensure that the quartz oscillates in
the desired mode. The new system is
insensitive to mode hopping and the resul-
tant inaccuracy. It is fast and precise. The
crystal frequency is determined 10 times a
second with an accuracy to less than
0.0005 Hz.
The ability of the system to initially identify
and then measure a certain mode opens
up new opportunities thanks to the advan-
tages of the additional information content
of these modes. This new, intelligent
measuring device makes use of the
phase/frequency properties of the quartz
crystal to determine the resonance fre-
quency. It works by applying a synthesized
sinus wave of a certain frequency to the
crystal and measuring the phase differen-
ce between the applied signal voltage and
the current flowing through the crystal. In
the case of series resonance, this differen-
ce is exactly zero degrees; then the crystal
behaves like an ohmic resistance. By dis-
connecting the applied voltage and the
current that returns from the crystal, one
D00
Fig. 6.5 Circuit of the active oscillator
crystal
amplifier
output
Fig. 6.6 Crystal frequencies near the series resonance point
Frequency (MHz)
p
h
a
s
e

(
d
e
g
r
e
e
s
)
l
o
g

.
Z
.

(
O
h
m
)
| impedance |
series resonance
phase
Fig. 6.7 Oscillations of a thickly coated crystal
Frequency (MHz)
p
h
a
s
e

(
d
e
g
r
e
s
s
)
l
o
g

.
Z
.

(
O
h
m
)
| impedance |
series resonance
phase
D00 E 19.06.2001 21:40 Uhr Seite 123
can determine with a phase comparator
whether the applied frequency is higher or
lower than the crystal resonance point.
The crystal impedance is capacitive at fre-
quencies below the fundamental wave and
inductive at frequencies above the reso-
nance. This information is useful if the
resonance frequency of a crystal is unkno-
wn. A brief frequency sweep is carried out
until the phase comparator changes over
and thus marks the resonance. For AT
quartzes we know that the lowest usable
frequency is the fundamental wave. The
anharmonics are slightly above that. This
information is not only important for the
beginning, but also in the rare case that the
instrument loses track of the fundamen-
tal wave. Once the frequency spectrum of
the crystal is determined, the instrument
must track the shift in resonance frequen-
cy, constantly carry out frequency measu-
rements and then convert them into
thickness.
Use of the intelligent measuring system
has a number of obvious advantages over
the earlier generation of active oscillators,
primarily insensitivity to mode hopping as
well as speed and accuracy of measure-
ment. This technique also enables the
introduction of sophisticated properties
which were not even conceivable with an
active oscillator setup. The same device
that permits the new technology to identify
the fundamental wave with one sweep can
also be used to identify other oscillation
modes, such as the anharmonics or quasi-
harmonics. The unit not only has a device
for constantly tracking the fundamental
wave, but can also be employed to jump
back and forth between two or more
modes. This query of different modes can
take place for two modes with 10 Hz on the
same crystal.
6.8 Auto Z match
technique
The only catch in the use of equation 6.4 is
that the acoustic impedance must be
known. There are a number of cases where
a compromise has to be made with accu-
racy due to incomplete or restricted know-
ledge of the material constants of the coa-
ting material:
1) The Z values of the solid material often
deviate from those of a coating. Thin
coatings are very sensitive to process
parameters, especially in a sputter envi-
ronment. As a result, the existing values
for solid material are not adequate.
2) For many exotic substances, including
alloys, the Z value is not known and not
easy to determine.
3) It is repeatedly necessary to carry out a
precise coating thickness measurement
for multiple coating with the same
crystal sensor. This applies in particular
to optical multiple and semi-conductor
coatings with a high temperature coeffi-
cient T
C
. However, the effective Z value
of the mixture of multiple coatings is
unknown.
In such a case, therefore, the only effective
method is to assume a Z value of 1, i.e. to
ignore reality with respect to wave propa-
gation in multi-substance systems. This
incorrect assumption causes errors in the
prediction of thickness and rate. The
magnitude of the error depends on the
coating thickness and the amount of devia-
tion from the actual Z value.
In 1989 A. Wajid invented the mode-lock
oscillator. He presumed that a connection
existed between the fundamental wave and
one of the anharmonics, similar to that
ascertained by Benes between the funda-
mental oscillation and the third quasi-har-
monic oscillation. The frequencies of the
fundamental and the anharmonic oscillati-
ons are very similar and they solve the
problem of the capacity of long cables. He
found the necessary considerations for
establishing this connection in works by
Wilson (1954) as well as Tiersten and
Smythe (1979).
The contour of the crystal, i.e. the spheri-
cal shape of one side, has the effect of
separating the individual modes further
from each other and preventing energy
transfer from one mode to another. The
usual method of identification is to desig-
nate the fundamental oscillation as (100),
the lowest anharmonic frequency as (102)
and the next higher anharmonic as (120).
These three indices of the mode nomen-
clature are based on the number of phase
reversals in the wave motion along the
three crystal axes. The above mentioned
works by Wilson, Tiersten and Smythe
examine the properties of the modes by
studying the influence of the radius of the
cut on the position of the anharmonic in
relation to the fundamental oscillation.
If one side of the quartz is coated with
material, the spectrum of the resonances
is shifted to lower frequencies. It has been
observed that the three above mentioned
modes have a somewhat differing mass
sensitivity and thus experience somewhat
different frequency shifts. This difference
is utilized to determine the Z value of the
material. By using the equations for the
individual modes and observing the fre-
quencies for the (100) and the (102)
mode, one can calculate the ratio of the
two elastic constants C60 and C55. These
two elastic constants are based on the
shear motion. The key element in Wajids
theory is the following equation:
(6.5)
with
M ... area mass/density ratio (ratio of coa-
ting mass to quartz mass per area
unit)
Z ... Z value
It is a fortunate coincidence that the pro-
duct M Z also appears in the Lu-Lewis
equation (equation 6.4). It can be used to
assess the effective Z value from the follo-
wing equations:
(6.6)
or
Here F
q
and F
c
are the frequencies of the
non-coated or coated quartz in the (100)
mode of the fundamental wave. Because of
the ambiguity of the mathematical func-
tions used, the Z value calculated in this
way is not always a positively defined
variable. This has no consequences of any
significance because M is determined in
another way by assessing Z and measuring
the frequency shift. Therefore, the thickn-
ess and rate of the coating are calculated
one after the other from the known M.
One must be aware of the limits of this
technique. Since the assessment of Z
depends on frequency shifts of two mo-
Z
tg M Z
F
F
tg
F
F
c
q
c
q

_
,
_
,
tg M Z
F
F
Z tg
F
F
c
q
c
q

_
,
_
,
0
( )
( / )
( / )
C C
C C
M Z
coated
uncoated
55 66
55 66
1
1
+
D00 E 19.06.2001 21:40 Uhr Seite 124
des, any minimal shift leads to errors due
to substantial mechanical or thermal stres-
ses. It is not necessary to mention that
under such circumstances the Z match
technique, too, leads to similar errors.
Nevertheless, the automatic Z value deter-
mination of the Z match technique is
somewhat more reliable regarding occur-
rence of errors because the amplitude dis-
tribution of the (102) mode is asymmetric
over the active crystal surface and that of
the (100) mode is symmetric.
According to our experience, coating-rela-
ted stresses have the most unfavorable
effect on the crystal. This effect is particu-
larly pronounced in the presence of gas,
e.g. in sputter processes or reactive vacu-
um coating or sputter processes. If the Z
value for solid material is known, it is bet-
ter to use it than to carry out automatic
determination of the auto Z ratio. In
cases of parallel coating and coating
sequences, however, automatic Z determi-
nation is significantly better.
6.9 Coating thickness
regulation
The last point to be treated here is the
theory of the closed loop for coating
thickness measuring units to effect coating
growth at a controlled (constant) growth
rate. The measuring advantages of the
instruments, such as speed, precision and
reliability, would not be completely exploi-
ted if this information were not inputted
into an improved process monitoring
system. For a coating process this means
the coating rate should be kept as close
and stable as possible to a setpoint. The
purpose of the closed loop is to make use
of the information flow of the measuring
system in order to regulate the capacity for
a special evaporation source in an appro-
priately adapted way. When the system
functions correctly, the controller transla-
tes small deviations of the controlled para-
meter (the rate) from the setpoint into cor-
rection values of the re-adjusted evaporati-
on capacity parameter. The ability of the
controller to measure quickly and precise-
ly keeps the process from deviating signi-
ficantly from the setpoint.
The most widespread type of controller is
the PID controller. Here P stands for pro-
portional, I for integral and D for differen-
tial control function. In the following some
of the properties of this controller are
described in detail. Information on the
system behavior is gained through a step
response to a control fault in certain con-
troller settings.
This response is recorded, and then
improved control parameters for a new
test are estimated. This procedure is con-
tinued until a satisfactory result is achie-
ved. At the end the controller is optimized
so that its parameters exactly match the
characteristics of the evaporator source.
It is a long and frustrating process to
adjust a controller to an evaporation sour-
ce, requiring several minutes for stabiliza-
tion and hours to obtain satisfactory
results. Often the parameters selected for
a certain rate are not suitable for an altered
rate. Thus, a controller should ideally
adjust itself, as the new controllers in
INFICON coating measuring units do. At
the beginning of installation and connec-
tion the user has the unit measure the cha-
racteristics of the evaporation source.
Either a PID controller is used as the basis
for slow sources or another type of con-
troller for fast sources without significant
dead time.
In relevant literature a distinction is made
between three different ways of setting
controllers. Depending on which data are
used for the setting, a distinction is made
between the closed loop, open loop and
resonance response method.
Due to the simplicity with which the expe-
rimental data can be obtained, we prefer-
red the open loop method. Moreover,
application of this technique permits
extensive elimination of the trial and error
method.
The Auto Control Tune function developed
by Inficon characterizes a process on the
basis of its step responses. After a step-by-
step change in the power the resulting
changes in the rate as a function of time are
smoothed and stored. The important step
responses are determined, see Fig. 6.8.
In general, it is not possible to characteri-
ze all processes exactly, so several appro-
ximations have to be made. Normally one
assumes that the dynamic characteristic
can be reproduced by a process of the first
order plus dead time. The Laplace trans-
formation for this assumption (transfer to
the s plane) is approximated:
with (6.8)
K
p
= amplification in stationary state
L = dead time
= time constant
These three parameters are determined
through the response curve of the pro-
cess. An attempt has been made by means
Output
Input
K
s
p
L
s
10
1
D00
L
0
t Time t
(0.632)
point of
maximum
rise
0.0632 K
p
1.00 K
p
T = t L
1 (0.632)
K = (change in output signal)/(change in control signal)
p
Fig. 6.8 Process response to a step change with t = 0 (open loop, control signal amplified)
D00 E 19.06.2001 21:40 Uhr Seite 125
of several methods to calculate the requi-
red parameters of the system response
from curves, as shown in Fig. 6.8. This
results in a 1-point accordance at 63.2 %
of the transition (a time constant), an
exponential accordance at two points and
an exponential accordance weighted
according to the method of the smallest
squares. A process is sufficiently characte-
rized by this information so that the con-
troller algorithm can be applied. Equation
6.9 shows the Laplace transformation for
the very often used PID controller:
(6.9)
with
M(s)= controlled variable or power
K
c
= Control amplification
(the proportional term)
T
i
= integration time
T
d
= differentiation time
E(s) = process deviation
Fig. 6.9 shows the control algorithm and a
process with a phase shift of the first order
and a dead time. The dynamics of the mea-
suring device and the control elements (in
our case the evaporator and the power
supply) are implicitly contained in the pro-
cess block. R(s) represents the rate set-
point. The return mechanism is the devia-
tion created between the measured preci-
pitation rate C(s) and the rate setpoint
R(s).
The key to use of any control system is to
select the correct values for K
c
, T
d
and T
i
.
The optimum control is a somewhat
subjective term that is made clear by the
presence of different mathematical defini-
tions:
Usually the smallest square error ISE
(Integral Square Error) is used as a mea-
sure of the quality of the control:
(6.10)
Here e is the error (the deviation): e = rate
ISE e t dt =
2
( )
M s K
S
T
T S E s
c
i
d
( ) ( ) + +
_
,
1

setpoint minus measured rate. ISE is rela-
tively insensitive to small deviations, but
large deviations contribute substantially to
the value of the integral. The result is small
overshoots, but long ripple times becau-
se deviations occurring late contribute litt-
le to the integral.
The integral of the absolute value of the
deviation IAE (Integral Absolute Error)
was also proposed as a measure for con-
trol quality:
(6.11)
This is more sensitive for small deviations,
but less sensitive for large deviations than
ISE.
Graham and Lanthrop introduced the inte-
gral over time, multiplied by the absolute
error ITAE (Integral Time Absolute Error),
as a measure for control quality:
(6.12)
The ITAE is sensitive to initial and, to a
certain extent, unavoidable deviations.
Optimum control responses defined
through ITAE consequently have short
response times and larger overshoots
than in the case of the other two criteria.
However, ITAE has proven to be very use-
ful for evaluating the regulation of coating
processes.
INFICONS Auto Control Tune is based on
measurements of the system response
with an open loop. The characteristic of
the system response is calculated on the
basis of a step change in the control sig-
nal. It is determined experimentally
through two kinds of curve accordance at
two points. This can be done either quick-
ly with a random rate or more precisely
with a rate close to the desired setpoint.
Since the process response depends on
the position of the system (in our case the
coating growth rate), it is best measured
near the desired work point. The process
information measured in this way (process
ITAE t e t dt
( )
IAE e t dt =
( )
amplification K
p
, time constant T
1
and
dead time L) are used to generate the most
appropriate PID control parameters.
The best results in evaluating coating con-
trol units are achieved with ITAE. There are
overshoots, but the reaction is fast and the
ripple time short. Controller setting condi-
tions have been worked up for all integral
evaluation criteria just mentioned so as to
minimize the related deviations. With a
manual input as well as with experimental
determination of the process response
coefficients, the ideal PID coefficients for
the ITAE evaluation can easily be calcula-
ted from equations 6.13, 6.14 and 6.15:
(6.13)
(6.14)
(6.15)
For slow systems the time interval bet-
ween the forced changes in control voltage
is extended to avoid hanging the control-
ler (hanging = rapid growth of the control
signal without the system being able to
respond to the altered signal). This makes
a response to the previous change in the
controller setting and powerful control-
ler settings possible. Another advantage is
the greater insensitivity to process noise
because the data used for control do not
come from merely one measurement, but
from several, so that the mass-integrating
nature of the quartz crystal is utilized.
In processes with short response times
(short time constants) and small to
unmeasurable dead times, the PID control-
ler often has difficulties with the noise of
the coating process (beam deflection,
rapid thermal short-circuits between melt
and evaporator, etc.). In these cases a con-
trol algorithm of the integral reset type is
used with success. This controller always
integrates the deviation and presses the
system towards zero deviation. This tech-
nique works well with small or completely
imperceptible dead times. However, if it is
used with a noticeable phase shift or dead
time, the controller tends to generate
oscillations because it overcompensates
the controller signal before the system has
a chance to respond. Auto Control Tune
T T
L
T
d

_
,
( . )
.
0381
1
1
0995
T
T
L
T
i

_
,

_
,
119
1 1
0738
.
.
K
K
L
T
c
p
_
,
_
,
136
1
0947
.
.
[process] [controller]
precipitation rate
C(s) K e
T1s + 1
p
aaa
K (1 + + T * s)
c d
s
T
i
( )
+
setpoint
R(s)
deviation
E(s)
L
s
Fig. 6.9 Block diagram of the PID controller
D00 E 19.06.2001 21:40 Uhr Seite 126
recognizes the properties of these fast
systems during the measurement of a step
response and utilizes the information to
calculate the control amplification for a
non-PID control algorithm.
6.10 INFICON
instrument variants
The instrument models available differ
both in hardware and software equipment:
the simplest unit, the XTM/2, is purely a
measuring or display device that cannot
control vacuum coating.
The XTC/2 and XTC/C group can control
vacuum coating sources and up to three
different coatings of a process (not to be
confused with nine different coating pro-
grams). In the case of XTM/2, XTC/2 and
XTC/C units, the AutoZero and AutoTune
functions are not available, and measure-
ment with several sensors simultaneously
as well as simultaneous control of two
vacuum coating sources are not possible.
However, the IC/5 offers all comfort func-
tions available today: measurement with
up to eight sensors with AutoZero and
AutoTune as well as capability of simulta-
neous control of two evaporator sources.
Moreover, it offers 24 material programs,
with which 250 coatings in 50 processes
can be programmed. To simplify operation
and avoid errors, the unit also has a dis-
kette drive. All types of crystal holders can
be connected here. The thickness resoluti-
on is around 1 , the rate resolution for
rates between 0 and 99.9 /s around 0.1
/s and for rates between 100 and 999 /s
around 1 /s. A particularly attractive opti-
on offered by the IC/5 is a microbalance
board with a highly stable reference
quartz. This oscillator is 50 times more
stable than the standard oscillator; long-
term stability and accuracy are then 2 ppm
over the entire temperature range. This
option is specially designed for coatings of
material with low density and at low coa-
ting rates. This is important for space con-
tamination and sorption studies, for exam-
ple.
D00
D00 E 19.06.2001 21:40 Uhr Seite 127
Applications of Vacuum Technology
7. Application of
vacuum
technology for
coating
techniques
7.1 Vacuum coating
technique
Vacuum technology has been increasingly
used in industrial production processes
during the last two decades. Some of
these processes and their typical working
pressure ranges are shown in Fig. 7.1.
Since a discussion of all processes is
beyond the scope of this brochure, this
section will be restricted to a discussion of
several examples of applications in the
important field of coating technology.
Deposition of thin films is used to change
the surface properties of the base material,
the substrate. For example, optical pro-
perties such as transmission or reflection
of lenses and other glass products, can be
adjusted by applying suitable coating layer
systems. Metal coatings on plastic web
produce conductive coatings for film capa-
citors. Polymer layers on metals enhance
the corrosion resistance of the substrate.
Through the use of vacuum it is possible
to create coatings with a high degree of
uniform thickness ranging from several
nanometers to more than 100 mm while
still achieving very good reproducibility of
the coating properties. Flat substrates,
web and strip, as well as complex molded-
plastic parts can be coated with virtually
no restrictions as to the substrate materi-
al. For example, metals, alloys, glass, cera-
mics, plastics and paper can be coated.
The variety of coating materials is also
very large. In addition to metal and alloy
coatings, layers may be produced from
various chemical compounds or layers of
different materials applied in sandwich
form. A significant advantage of vacuum
coating over other methods is that many
special coating properties desired, such as
structure, hardness, electrical conductivity
or refractive index, are obtained merely by
selecting a specific coating method and
the process paramaters for a certain coa-
ting material.
7.2 Coating sources
In all vacuum coating methods layers are
formed by deposition of material from the
gas phase. The coating material may be
formed by physical processes such as eva-
poration and sputtering, or by chemical
reaction. Therefore, a distinction is made
between physical and chemical vapor
deposition:
physical vapor deposition = PVD
chemical vapor deposition = CVD.
7.2.1 Thermal evaporators
(boats, wires etc.)
In the evaporation process the material to
be deposited is heated to a temperature
high enough to reach a sufficiently high
vapor pressure and the desired evaporati-
on or condensation rate is set. The sim-
plest sources used in evaporation consist
of wire filaments, boats of sheet metal or
electrically conductive ceramics that are
heated by passing an electrical current
through them (Fig. 7.2). However, there
are restrictions regarding the type of mate-
rial to be heated. In some cases it is not
possible to achieve the necessary evapora-
10
10
10
7
10
3
10
0
10
3
Pressure [mbar]
Annealing of metals
Degassing of melts
Electron beam melting
Electron beam welding
Evaporation
Sputtering of metals
Casting of resins and lacquers
Drying of plastics
Drying of insulating papers
Freeze-drying of bulk goods
Freeze-drying of pharmaceutical products
Fig. 7.1 Pressure ranges for various industrial processes
D00 E 19.06.2001 21:40 Uhr Seite 128
tor temperatures without significantly eva-
porating the source holder and thus con-
taminating the coating. Furthermore che-
mical reactions between the holder and the
material to be evaporated can occur resul-
ting in either a reduction of the lifetime of
the evaporator or contamination of the
coating.
7.2.2 Electron beam
evaporators
(electron guns)
To evaporate coating material using an
electron beam gun, the material, which is
kept in a water-cooled crucible, is bombar-
ded by a focused electron beam and there-
by heated. Since the crucible remains cold,
in principle, contamination of the coating
by crucible material is avoided and a high
degree of coating purity is achieved. With
the focused electron beam, very high tem-
peratures of the material to be evaporated
can be obtained and thus very high evapo-
ration rates. Consequently, high-melting
point compounds such as oxides can be
evaporated in addition to metals and
alloys. By changing the power of the elec-
tron beam the evaporation rate is easily
and rapidly controlled.
7.2.3 Cathode sputtering
In the cathode sputtering process, the tar-
get, a solid, is bombarded with high ener-
gy ions in a gas discharge (Fig. 7.3). The
impinging ions transfer their momentum
to the atoms in the target material,
knocking them off. These displaced atoms
the sputtered particles condense on
the substrate facing the target. Compared
to evaporated particles, sputtered particles
have considerably higher kinetic energy.
Therefore, the conditions for condensation
and layer growth are very different in the
two processes. Sputtered layers usually
have higher adhesive strength and a den-
ser coating structure than evaporated
ones. Sputter cathodes are available in
many different geometric shapes and sizes
as well as electrical circuits configurations.
What all sputter cathodes have in common
is a large particle source area compared to
evaporators and the capability to coat large
substrates with a high degree of uniformi-
ty. In this type of process metals, alloys of
any composition as well as oxides can be
used as coating materials.
7.2.4 Chemical vapor
deposition
In contrast to PVD methods, where the
substance to be deposited is either solid or
liquid, in chemical vapor deposition the
substance is already in the vapor phase
when admitted to the vacuum system. To
deposit it, the substance must be thermal-
ly excited, i.e. by means of appropriate
high temperatures or with a plasma. Gene-
rally, in this type of process, a large num-
ber of chemical reactions take place, some
of which are taken advantage of to control
the desired composition and properties of
the coating. For example, using silicon-
hydrogen monomers, soft Si-H polymer
coatings, hard silicon coatings or by
addition of oxygen quartz coatings can
be created by controlling process parame-
ters.
D00
Fig. 7.2 Various thermal evaporators
Hairpin-shaped
evaporator made of
twisted tungsten wire
Spiral evaporator
made of twisted tung-
sten wire
Trough-shaped
evaporator
Trough-shaped
evaporator with
ceramic coating
Basket-shaped
evaporator
Evaporator made of
electrically conductive
ceramics
Boat-shaped
evaporator
Evaporator
with ceramic
insert
Fig. 7.3 Schematic diagram of a high-performance cathode sputter arrangement
5
4
3
2
1
6
7
8
9
1 Substrates
2 Sputtered
atoms
3 Anode
4 Electrons
5 Target
6 Cathode
7 Magnetic
field lines
8 Argon ions
9 Substrate
D00 E 19.06.2001 21:40 Uhr Seite 129
7.3 Vacuum coating
technology/
coating systems
7.3.1 Coating of parts
For molded-plastic parts, vacuum coating
techniques are increasingly replacing con-
ventional coating methods, such as elec-
troplating. For example, using vacuum
coating methods, automobile reflectors
obtain a mirror-like surface, plastic articles
in the furniture, decoration, clock and
watch as well as electronics industry are
metal-coated and optical effects are crea-
ted on articles in the decoration industry.
Fig. 7.4 shows a type of vacuum system in
which large batches of molded-plastic
parts can be coated simultaneously. The
substrates are placed on a cage that rota-
tes past the coating source, a sputter
cathode in this example. In some applica-
tions, by using a glow discharge treat-
ment, the substrates are cleaned and the
surface is activated prior to the coating
process. This enhances the adhesive
strength and reproducibility of the coating
properties. A corrosion protection coating
can be applied after sputtering. In this
case, a monomer vapor is admitted into
the system and a high-frequency plasma
discharge ignited. The monomer is actived
in the plasma and deposits on the substra-
tes as a polymer coating. In this type of
system there may be plastic substrates
with a surface area of several 10 m
2
on the
cage, causing a correspondingly high
desorption gas flow. The vacuum system
must be able to attain the required pressu-
res reliably despite these high gas loads.
In the example shown, the system is eva-
cuated with a combination of a backing
and Roots pump. A diffusion pump along
with a cold surface forms the high vacuum
pump system. The cold surfaces pump a
large portion of the vapor and volatile sub-
stances emitted by the plastic parts while
the diffusion pump basically removes the
non-condensable gases as well as the
noble gas required for the sputter process.
A completely different concept for the
same process steps is shown in Fig. 7.5.
The system consists of four separate stati-
ons made up of a drum rotating around the
vertical axis with four substrate chambers
and process stations mounted in the vacu-
um chamber. During rotation, a substrate
chamber moves from the loading and
unloading station to the pretreatment sta-
tion, to the metallization station, to the
protective coating station and then back to
the initial position. Since each station has
its own pumping system, all four proces-
ses can run simultaneously with entirely
independent adjustable process parame-
ters. The vacuum system comprises of
turbomolecular pumps and backing pump
sets consisting of Roots and rotary vane
pumps.
7.3.2 Web coating
Metal-coated plastic webs and papers play
an important role in food packaging. They
preserve food longer according to storage
and transport logistics requirements and
give packaging an attractive appearence.
Another important area of application of
metal-coated web is the production of film
capacitors for electrical and electronics
applications. Metal-coating is carried out
in vacuum web coating systems. Fig. 7.6
shows a typical scheme. The unit consists
of two chambers, the winding chamber
with the roll of web to be coated and the
winding system, as well as the coating
chamber, where the evaporators are loca-
ted. The two chambers are sealed from
each other, except for two slits through
which the web runs. This makes it possible
to pump high gas loads from the web roll
using a relatively small pumping set. The
pressure in the winding chamber may be
more than a factor of 100 higher than the
pressure simultaneously established in the
coating chamber. The pump set for the
winding chamber usually consists of a
combination of Roots and rotary vane
pumps.
Fig. 7.4 Diagramm of a batch system for coating parts
Fig. 7.5 Multi-chamber parts-coating unit (rotationally symmetric in-line system
DynaMet 4V)
1 Vacuum chamber
2 High-performance cathode
3 Substrate holder
4 Substrates
5 Diffusion pump
6 Roots pump
7 Rotary piston pump
8 Cold trap
9 High vacuum valve
10 Valve for bypass line
11 Foreline valve
12 Venting valve
D00 E 19.06.2001 21:40 Uhr Seite 130
With strongly degassing rolls of paper, it
may be necessary to install a cold surface
in the winding chamber to act as a water
vapor pump. The rolls of the plastic web or
paper typically have diameters between
400 and 1000 mm and a width of 400 to
3000 mm. A precise, electronically con-
trolled winding system is required for win-
ding and unwinding as well as web gui-
dance.
During the coating process the web, at a
speed of more than 10 m/s, passes a group
of evaporators consisting of ceramic boats,
from which aluminium is evaporated. To
achieve the necessary Al-coating thickness
at these high web speeds, very high evapo-
ration rates are required. The evaporators
must be run at temperatures in excess of
1400 C. The thermal radiation of the eva-
porators, together with the heat of conden-
sation of the growing layer, yields a consi-
derable thermal load for the web. With the
help of cooled rollers, the foil is cooled
during and after coating so that it is not
damaged during coating and has cooled
down sufficiently prior to winding.
During the entire coating process the coa-
ting thickness is continuously monitored
with an optical measuring system or by
means of electrical resistance measure-
ment devices. The measured values are
compared with the coating thickness set-
points in the system and the evaporator
power is thus automatically controlled.
7.3.3 Optical coatings
Vacuum coatings have a broad range of
applications in production of ophthalmic
optics, lenses for cameras and other opti-
cal instruments as well as a wide variety of
optical filters and special mirrors. To
obtain the desired transmission or reflec-
tion properties, at least three, but someti-
mes up to 50 coatings are applied to the
glass or plastic substrates. The coating
properties, such as thickness and refrac-
tive index of the individual coatings, must
be controlled very precisely and matched
to each other. Most of these coatings are
produced using electron beam evapora-
tors in single-chamber units (Fig. 7.7). The
evaporators are installed at the bottom of
the chamber, usually with automatically
operated crucibles, in which there are
several different materials. The substrates
are mounted on a rotating calotte above
the evaporators. Application of suitable
shieldings combined with relative move-
ment between evaporators and substrates,
results in a very high degree of coating
uniformity. With the help of quartz coating
thickness monitors (see Section 6) and
direct measurement of the attained optical
properties of the coating system during
coating, the coating process is fully con-
trolled automatically.
One of the key requirements of coatings is
that they retain their properties under
usual ambient conditions over long
periods of time. This requires to produce
the densest coatings possible, into which
neither oxygen nor water can penetrate.
Using glass lenses, this is achieved by
keeping the substrates at temperatures up
to 300 C during coating by means of
radiation heaters. However, plastic lenses,
as those used in eyeglass optics, are not
allowed to be heated above 80 C. To
obtain dense, stable coatings these sub-
strates are bombarded with Ar ions from
an ion source during coating. Through the
ion bombardement the right amount of
energy is applied to the growing layer so
that the coated particles are arranged on
the energetically most favorable lattice
sites, without the substrate temperature
reaching unacceptably high values. At the
same time oxygen can be added to the
argon. The resulting oxygen ions are very
reactive and ensure that the oxygen is
included in the growing layer as desired.
The vacuum system of such a coating unit
usually consists of a backing pump set
comprising a rotary vane pump and Roots
pump as well as a high vacuum pumping
system. Depending on the requirements,
diffusion pumps, cryo pumps or turbo-
molecular pumps are used here, in most
cases in connection with large refrigerator-
cooled cold surfaces. The pumps must be
installed and protected by shieldings in a
way that no coating material can enter the
pumps and the heaters in the system do
D00
Fig. 7.6 Schematic diagram of a vacuum web coating system
1 Unwinder
2 Coating source
3 Coating roller
4 Drawing roller
5 Take-up reel
High-perfor-
mance
plasma
source
Electron
beam
evaporator
Monomer
Ar
O
2
Fig. 7.7 Coating unit for optical coating systems
D00 E 19.06.2001 21:40 Uhr Seite 131
not thermally overload them. Since shiel-
ding always reduces the effective pumping
speed, the system manufacturer must find
a suitable compromise between shielding
effect and reduction of pumping speed.
7.3.4 Glass coating
Coated glass plays a major role in a num-
ber of applications: window panes in
moderate and cold climate zones are pro-
vided with heat-reflecting coating systems
to lower heating costs; in countries with
high intensity solar radiation, solar protec-
tion coatings are used that reduce air
conditioning costs; coated car windows
reduce the heating-up of the interior and
mirrors are used both in the furniture and
the automobile industry. Most of these
coatings are produced in large in-line
vacuum systems. Fig. 7.8 shows a typical
system. The individual glass panes are
transported into a entrance chamber at
atmospheric pressure. After the entrance
valve is closed, the chamber is evacuated
with a forepump set. As soon as the pres-
sure is low enough, the valve to the eva-
cuated transfer chamber can be opened.
The glass pane is moved into the transfer
chamber and from there at constant speed
to the process chambers, where coating is
carried out by means of sputter cathodes.
On the exit side there is, in analogy to the
entrance side, a transfer chamber in which
the pane is parked until it can be transfer-
red out through the exit chamber.
Most of the coatings consist of a stack of
alternative layers of metal and oxide. Since
the metal layers may not be contaminated
with oxygen, the individual process stati-
ons have to be vacuum-isolated from each
other and from the transfer stations. Uti-
lization of valves for separating process
chambers is unsatisfactory because it
increases plant dimensions. To avoid fre-
quent and undesirable starting and stop-
ping of the glass panes, the process cham-
bers are vacuum-separated through so-
called slit locks, i.e. constantly open slits
combined with an intermediate chamber
with its own vacuum pump (Fig. 7.9). The
gaps in the slits are kept as small as tech-
nically possible to minimize clearance and
therefore conductance as the glass panes
are transported through them. The pum-
ping speed at the intermediate chamber is
kept as high as possible in order to achie-
ve a considerably lower pressure in the
intermediate chamber than in the process
chambers. This lower pressure greatly
reduces the gas flow from a process
chamber via the intermediate chamber to
the adjacent process chamber. For very
stringent separation requirements it may
be necessary to place several intermediate
chambers between two process chambers.
The glass coating process requires high
gas flows for the sputter processes as well
as low hydrocarbon concentration. The
only vacuum pump which satisfies these
requirements as well as high pumping
speed stability over time are turbo-mole-
cular pumps which are used almost exclu-
sively.
While the transfer and process chambers
are constantly evacuated, the entrance and
exit chambers must be periodically vented
and then evacuated again. Due to the large
volumes of these chambers and the short
cycle times, a high pumping speed is
required. It is provided by combinations of
rotary vane pumps and Roots pumps. For
particularly short cycle times gas cooled
Roots pumps are also used.
All major functions of a plant, such as
glass transport, control of the sputter pro-
cesses and pump control, are carried out
fully automatically. This is the only way to
ensure high productivity along with high
product quality.
7.3.5 Systems for producing
data storage disks
Coatings for magnetic- or magneto-optic
data storage media usually consist of
several functional coatings that are applied
to mechanically finished disks. If several
plates are placed on one common carrier,
the coating processes can be carried out in
a system using a similar principle to that
used for glass coating. However, most
disks must be coated on both sides and
there are substantially greater low particle
contamination requirements as compared
to glass coating. Therefore, in-line
systems for data memories use a vertical
carrier that runs through the system (Fig.
7.10). The sputter cathodes in the process
stations are mounted on both sides of the
carrier so that the front and back side of
the disk can be coated simultaneously.
An entirely different concept is applied for
coating of single disks. In this case the dif-
ferent process stations are arranged in a
circle in a vacuum chamber (Fig. 7.11).
The disks are transferred individually from
a magazine to a star-shaped transport arm.
S
Z
L
1Z
L
Z2
Process chamber
1
Intermediate chamber
Slits
Process chamber
2
S
2
S
1

Fig. 7.9 Principle of chamber separation through pressure stages
L
1Z
, L
Z2
= conductance between intermediate chamber and
process chamber 1 or 2
S
Z
= pumping speed at intermediate chamber
S
1
, S
2
= pumping speed at process chamber 1 or 2
Entrance chamber
Sputter chambers
Exit chamber
Transfer chamber 1
Transfer chamber 2
to backing pumps
Fig. 7.8 Plant for coating glass panes 3-chamber in-line system, throughput up to
3,600,000 m
2
/ year
D00 E 19.06.2001 21:40 Uhr Seite 132
The transport arm cycles one station furt-
her after each process step and in this way
transports the substrates from one pro-
cess station to the next. During cycling all
processes are switched off and the stati-
ons are vacuum-linked to each other. As
soon as the arm has reached the process
position, the individual stations are sepe-
rated from each other by closing seals.
Each station is pumped by means of its
own turbomolecular pump and the indivi-
dual processes are started. As many pro-
cess stations as there are in the system as
many processes can be performed in par-
allel. By sealing off the process stations,
excellent vacuum separation of the indivi-
dual processes can be achieved. However,
since the slowest process step determines
the cycle interval, two process stations
may have to be dedicated for particularly
timeconsuming processes.
D00
Fig. 7.10 Plant for coating data storage disks with carrier transport system
Fig. 7.11 Plant for individual coating of data storage disks
D00 E 19.06.2001 21:40 Uhr Seite 133
Instruction for Equipment Operation
8. Instructions for
vacuum equip-
ment operation
8.1 Causes of faults
where the desired
ultimate pressure
is not achieved or
is achieved too
slowly
If the desired ultimate pressure is not rea-
ched at all in vacuum equipment or if it is
attained only after an excessively long
pumping period, then the following pro-
blems may be the reason:
If the desired ultimate pressure is not reached,
then
the apparatus may be leaky or dirty,
the pump may dirty or damaged,
the vacuum gauge may be defective.
If the desired ultimate pressure is reached
only after a very long running time, then
the apparatus may be dirty,
the pumping line may be restricted,
the pump may be dirty or of too small a
capacity,
the pumping speed may be restricted
due to other causes.
In order to locate the fault, one normally
proceeds by separating the evacuated ves-
sel from the pump system (where this is
possible) and checking the vessel alone for
leaks and contamination using the pressu-
re rise method, for example. If it has been
found that the vessel is free of defects in
this regard, then the measurement system
will be checked for cleanliness (see Sec-
tion 8.38) and ultimately if required the
pump or the pumping system itself will be
examined.
8.2 Contamination of
vacuum vessels
and eliminating
contamination
In addition to the pressure rise method
(Section 6.1) there is a further method for
detecting contamination, based on the fact
that condensable vapors will generally
account for the major share of the gas mix
in dirty vessels: here the pressure reading
at a compression vacuum gauge (partial
pressure for the non-condensable gases) is
compared with that at an electrical vacuum
gauge, e.g. a thermal conductivity or
ionization vacuum gauge (measuring total
pressure). These two vacuum gauges
must, however, be clean themselves.
Where vapors are present the compression
vacuum gauge will indicate much better
pressure than the electrical vacuum gauge.
This is a sure sign that the vessel walls are
contaminated, usually with oil. Another
commonly used procedure is to compare
the pressure indication of one and the
same vacuum gauge (not a compression
vacuum gauge) with and without a cold
trap inserted in the line: Filling the cold trap
with liquid nitrogen will cause the pressure
to drop abruptly, by one power of ten or
more, if the container is contaminated
since the vapors will freeze out in the trap.
Eliminating contamination for glass
equipment
If the contaminants exhibit a high vapor
pressure, then they can be pumped out
relatively quickly. If this is not successful,
then the apparatus will have to be cleaned.
Contaminated glass components will first
be cleaned with chromic-sulfuric acid mix-
ture or if this is necessary with dilute
hydrofluoric acid (1:30). They are then rin-
sed with distilled water. If this does not
bring about the desired results, then an
organic solvent can be tried. Then the
glass components will again have to be
rinsed with methanol and distilled water.
(Do not use denatured alcohol!)
Eliminating contamination at metallic
equipment
Metal components will usually exhibit
traces of contamination by oil and greases.
If these cannot be readily removed by pum-
ping down the vessel, then an appropriate
organic solvent (denatured alcohol is
unsuitable in all cases) will have to be used
for cleaning. Maximum cleanliness can be
achieved with vapor baths such as those
commonly found in industry. If one desires
to get down to extremely low pressure ran-
ges (< 10
-7
mbar), then after cleaning
the metal components will have to be baked
out at temperatures of up to 200 C.
Seriously contaminated metal components
will first have to be cleaned by cutting away
or sandblasting the top surface. These
methods suffer the disadvantage that the
surface area for the surface thus treated
will be increased through roughening and
active centers may potentially be formed
which would readily adsorb vapor molecu-
les. Supplementary cleaning in the vapor
bath (see above) is advisable. In some
cases electrolytic pickling of the surface
may be beneficial. In the case of high vacu-
um components it is necessary to pay
attention to ensuring that the pickling does
not turn into etching, which would serious-
ly increase the surface area. Polishing sur-
faces which have been sandblasted is not
necessary when working in the rough and
medium vacuum ranges since the surface
plays only a subordinate role in these pres-
sure regimes.
8.3 General operating
information for
vacuum pumps
If no defects are found in the vacuum ves-
sels and at the measurement tubes or if
the apparatus still does not operate satis-
factorily after the faults have been rec-
tified, then one should first check the flan-
ge seals at the pump end of the system
and possibly the shut-off valve. Flange
seals are known to be places at which
leaks can appear the most easily, resulting
from slight scratches and mechanical
damage which initially appears insignifi-
cant. If no defect can be discerned here,
either, then it is advisable to check to see
whether the pumps have been maintained
in accordance with the operating instruc-
tions.
Given initially in this section are general
instructions on pump maintenance, to be
followed in order to avoid such defects
from the very outset. In addition, potential
errors and their causes are discussed.
D00 E 19.06.2001 21:40 Uhr Seite 134
8.3.1 Oil-sealed rotary vacuum
pumps
(Rotary vane pumps and
rotary piston pumps)
8.3.1.1 Oil consumption, oil
contamination, oil change
The oil serves to:
lubricate moving parts,
seal moving parts against atmospheric
pressure,
seal the valve,
fill the dead space below the valve,
to seal the various operational spaces
one from another.
In all pumps it is possible to check the oil
charge during operation using the built-in
oil level sight glass. During continuous
operation in particular it is necessary to
ensure that the oil charge never falls below
the minimum level. During a pumping pro-
cess oil-sealed rotary pumps will emit oil
vapors from the discharge port, this being
due to the high operating temperature.
This leads to oil loss to an extent which
will depend on the quantity of gas or vapor
which is drawn into the pump. Larger oil
droplets can be retained by installing a
coarse separator in the discharge line. This
will reduce oil loss considerably. The fine
oil mist filter installed in some pumps will
also retain the finest oil droplets (fine
separation) so that no oil at all will leave
the outlet of the pump and oil loss is redu-
ced practically to zero since as in coarse
separation the oil which is separated out
is returned to the pump. With pumps with-
out an integral fine separator this device is
offered as an optional extra.
If an oil-filled rotary pump is operated
without an oil separation and return devi-
ce, then it will be necessary to expect a
certain amount of oil consumption, the
extent of which will depend on the size of
the pump and the nature of the operations.
In the worst case this can amount to about
2 cm
3
for every cubic meter of air pumped
(at STP and including the gas ballast also
drawn in). Figure 8.1 makes it possible to
predict the amount of oil loss to be expec-
ted in practical situations. The example
demonstrates that greater oil losses must
be expected when operating the pump with
gas ballast. This situation, which is gene-
rally valid, is always to be taken into
account in practice.
If the pump oil has become unusabledue
to exposure to the vapors or contaminants
which are encountered in the process,
then the oil will have to be replaced. It is
impossible to formulate any hard-and-fast
rules as to when an oil change will be
required since the nature of the operations
will determine how long the oil will remain
good. Under clean conditions (e.g. backing
pumps for diffusion pumps in electro-
nuclear accelerators) rotary vacuum
pumps can run for years without an oil
change. Under extremely dirty conditi-
ons (e.g. during impregnation) it may be
necessary to change the oil daily. The oil
will have to be replaced when its original
light brown color, has turned dark brown
or black due to ageing or has become
cloudy because liquid (such as water) has
entered the pump. An oil change is also
necessary when flakes form in corrosion
protection oil, indicating that the corrosion
protection agent is exhausted.
Changing the oil
The oil change should always be carried
out with the pump switched off but at ope-
rating temperature. The oil drain (or fill)
opening provided for each pump is to be
used for this purpose. Where the pump is
more seriously contaminated, then it
should be cleaned. The applicable opera-
ting instructions are to be observed in this
case.
8.3.1.2 Selection of the pump oil
when handling aggressive
vapors
If corrosive vapors (e.g. the vapors formed
by acids) are to be pumped, then a
PROTELEN
corrosion protection oil

should be used in place of the normal
pump oil (N 62). These types of vapors will
then react with the basic (alkaline) corrosi-
on protection agent in the oil. The conti-
D00
Fig. 8.1 Oil loss for oil-sealed pumps (referenced to an approximate maximum value of 2 cm
3
oil loss per cubic meter
of air drawn in [STP])
Example: Oil losses for a TRIVAC S 16 A at pressure of 1 mbar:
a) Without gas ballast: in accordance with the pumping speed curve S = 15 m
3
/ h; according to diagram:
oil loss = 0.03 cm
3
/ h (line a)
b) With gas ballast: S = 9 m
3
/h at 1 mbar; oil loss = 0.018 cm
3
/ h (line b1), plus additional loss due to
gas ballast quantity (0.1 times the 1.6 m
3
/ h rated pumping speed); that is: Chart on the horizontal
line b2 up to atmospheric pressure: additional loss: 3 cm
3
/ h (diagonal line). Overall loss during gas
ballast operation 0.018 + 3 = 3.018 cm
3
/ h
Intake pressure [mbar]
P
u
m
p
i
n
g

s
p
e
e
d

[
m
3
1
]
D00 E 19.06.2001 21:40 Uhr Seite 135

nuous neutralization reactions will exhaust
the corrosion protection agent at a rate
depending on the quantity and acidity of
the vapors. The oil will have to be changed
more frequently, in accordance with these
factors. Corrosion protection oils are eit-
her very hygroscopic or they easily form
emulsions with water. Consequently a
pump which is filled with corrosion pro-
tection oil will absorb moisture from the
air if it is out of service for an extended
period of time. In no case should one ever
use a pump filled with corrosion protection
oil in order to pump water vapor since the
lubricating and corrosion inhibition pro-
perties of the oil would be adversely affec-
ted. Once the oil has absorbed water it will
no longer be possible for such pumps to
achieve the ultimate pressures which
would be the case with fresh corrosion
protection oil or standard pump oil (N 62).
Oil-filled pumps should, under normal
operating con-ditions, not be filled with
corrosion protection oil. N 62 oil is prefer-
red when pumping air, water vapor and
non-corrosive organic vapors in so far as
there is positive protection against the
vapors condensing inside the pump.
8.3.1.3 Measures when pumping
various chemical substances
This discussion cannot provide exhaustive
coverage of the many and varied applicati-
on fields for oil-filled vacuum pumps in the
chemicals industry. Our many years of
experience with the most difficult of che-
micals applications can be used to solve
your particular problems. Three aspects
should, however, be mentioned briefly:
pumping explosive gas mixes, condensab-
le vapors, and corrosive vapors and gases.
Explosion protection
Applicable safety and environmental pro-
tection regulations shall be observed when
planning and engineering vacuum
systems. The operator must be familiar
with the substances which the system will
be pumping and take into account not only
normal operating conditions but also
abnormal situations, operating outside
normal parameters. The most important
aids to avoiding explosive mixtures are
in addition to inertization by adding pro-
tective gases maintaining the explosion
limit values with the aid of condensers,
adsorption traps and gas scrubbers.
Protection against condensation
LEYBOLD pumps offer three options for
keeping vapors from condensing in the
pumps:
The gas ballast principle (See Fig. 2.14).
This increases considerably the amount
of vapor which the pump can tolerate.
Increased pump temperature. The rug-
ged design of our pumps makes it pos-
sible to run them at temperatures of up
to 120 C. Thus the tolerance for pure
water vapor, for example, will rise by a
factor of five when compared with nor-
mal gas ballast operation.
Using vacuum condensers (see Section
2.15). These act as selective pumps and
should be sized so that the downstream
gas ballast pump will not receive more
vapor than the amount corresponding
to the appropriate vapor tolerance.
Corrosion protection
Oil-sealed pumps are already quite satis-
factorily protected against corrosion due
to the oil film which will be present on all
the component surfaces. Corrosion is defi-
ned here as the electrochemical dis-
solution of metals, i.e. the release of elec-
trons by the metal atom and their accep-
tance by the oxidation agent (corrosive
gas). A metal atom which is susceptible to
corrosion must therefore be exposed to an
active atom of the oxidation agent.
This makes clear how the oil-sealed pump
is protected against corrosion; the concen-
tration of the oxidation agent in the oil is
negligible and thus the opportunity for the
metal to release electrons is equally small.
This also makes it clear that the use of so-
called non-rusting or stainless steels
does not make sense since oxidation is
necessary for the passivation of these
steels, in order to reach the so-called pas-
sive region for these steel compounds.
The critical passivation current density will
normally not appear in oil-sealed pumps.
a) Acids
Our pumps are fundamentally suited to
pumping acids. In special situations
problems with the oil and with auxiliary
equipment attached at the intake and/or
discharge end may occur. Our engineers in
Cologne are available to assist in solving
such problems.
b) Anhydrides
CO (carbon monoxide) is a strong redu-
cing agent. When CO is being pumped it is
therefore important that air not be used as
the gas ballast but rather that inert gases
be used at the very outside (e.g. Ar or N
2
).
Inert gas ballast should also be used when
pumping SO
2
, SO
3
, and H
2
S. A corrosion
inhibiting oil is also to be used when hand-
ling these three anhydrides. Carbon dioxi-
de (CO
2
) can be pumped without making
any special arrangements.
c) Alkaline solutions
Normal N 62 pump oil is to be used to
pump basic (alkaline) solutions. Sodium
hydroxide and caustic potash solutions
should not be pumped in their concen-
trated form. Ammonia is highly amenable
to pumping with the gas ballast valve clo-
sed. Alkaline organic media such as
methylamine and dimethylamine can also
be pumped satisfactorily, but with the gas
ballast valve open.
d) Elementary gases
Pumping nitrogen and inert gases requires
no special measures.
When handling hydrogen it is necessary to
make note of the hazard of creating an
explosive mixture. The gas ballast valve
may in no case be opened when dealing
with hydrogen. The motors driving the
pumps must be of explosion-proof design.
Oxygen: Particular caution is required
when pumping pure oxygen! Specially
formulated pump oils must be used for
this purpose. We can supply these,
accompanied by an approval certificate
issued by the German Federal Materials
Testing Authority (BAM), following consul-
tation.
e) Alkanes
The low molecular weight alkanes such as
methane and butane can be pumped with
the gas ballast valve closed or using inert
gas as the gas ballast and/or at increased
temperature of the pump. But important
Increased explosion hazard!
f) Alcohols
Once operating temperature has been rea-
ched, methanol and ethanol can be extrac-
ted without using gas ballast (N 62 pump
oil). To pump higher molecular weight
alcohols (e.g. butanol) the gas ballast
valve will have to be opened or other pro-
tective measures will have to be imple-
mented to prevent condensation.
D00 E 19.06.2001 21:40 Uhr Seite 136
g) Solvents
Acetone: Can be extracted without difficul-
ty; wait until normal operating temperatu-
re is reached.
Benzene: Caution vapors are highly
flammable. Ultimate pressure is degraded
by dilution of the pump oil. Mixtures of air
and benzene are explosive and flammable.
Work without a gas ballast! Inert gases
may possibly be used as ballast gas.
Carbon tetrachloride and trichlorethy-
lene: Amenable to pumping; non-flam-
mable and non-explosive but will be dis-
solved in oil and thus increase the ultima-
te pressure; wait until normal operating
temperature is achieved. Keep the gas bal-
last open when pumping carbon tet and
other non-flammable solvents. Use N 62
pump oil.
Toluene: Pump through the low-temper-
ature baffle and without gas ballast. Use
inert gas, not air, as the gas ballast.
8.3.1.4 Operating defects while
pumping with gas ballast
Potential sources of error
where the required ultimate
pressure is not achieved
a) The pump oil is contaminated (parti-
cularly with dissolved vapors). Check
the color and properties. Remedy:
Allow the pump to run for an extended
period of time with the vacuum vessel
isolated and the gas ballast valve open.
In case of heavy contamination an oil
change is advisable. The pump should
never be allowed to stand for a longer
period of time when it contains conta-
minated oil.
b) The sliding components in the pump
are worn or damaged. Under clean con-
ditions our pumps can run for many
years without any particular mainte-
nance effort. Where the pump has been
allowed to run for a longer period of
time with dirty oil, however, the bea-
rings and the gate valves may exhibit
mechanical damage. This must always
be assumed when the pump no longer
achieves the ultimate pressure specified
in the catalog even though the oil has
been changed. In this case the pump
should be sent in for repair or our
customer service department should be
contacted.
c) The measurement instrument is soiled
Potential sources of error when the pump
no longer turns
Check the pump electrical supply.
The pump has stood for a long time
containing contaminated or resinous
oil.
The pump is colder than 10C and the
oil is stiff. Heat the pump.
There is a mechanical error. Please con-
tact our customer service department.
Oil exits at the shaft
If oil is discharged at the shaft, then the
Seeger rotary shaft circlip in the drive bea-
ring will have to be inspected and possibly
replaced. The design of the pumps makes
it possible to replace this ring easily, follo-
wing the operating instructions provided
with the unit.
8.3.2 Roots pumps
8.3.2.1 General operating instruc-
tions, installation and com-
missioning
Roots pumps must be exactly level. When
attaching the pump it is necessary to ensu-
re that the pump is not under any tension
or strain whatsoever. Any strains in the
pump casing caused by the connection
lines shall be avoided. Any strain to which
the pump is subjected will endanger the
pump since the clearances inside the roots
pump are very small.
Roots pumps are connected to the line
power supply via the motor terminal strip;
a motor protection switch (overload/
overheating) shall be provided as required
by local codes.
The direction of motor rotation shall be
checked with the intake and outlet ports
open prior to installing the pump. The
drive shaft, seen from the motor end, must
rotate counter-clockwise. Note the arrow
on the motor indicating the direction of
rotation! If the roots pump runs in the
wrong direction, then it is reversed by
interchanging two of the phases at the
motor connection cord.
The roots pump may be switched on only
after the roughing pump has evacuated the
vacuum vessel down to the cut-in pressu-
re.
Permissible cut-in pressure P
E
will depend
on the reduction ratios of the roots pump
as against the roughing pump and is cal-
culated by dividing the permissible pressu-
re differential p
max
by the compression
ratio, reduced by the value of 1:
If the pump is protected using a dia-
phragm-type pressure switch, then the
pump will be switched on automatically. If
a combination of roots pump and roughing
pump is to convey highly volatile sub-
stances such as liquids with a low boiling
point, then it is advisable to use a roots
pump which is equipped with an integral
bypass line and a valve which will respond
to a pre-set pressure. Example: Roots
vacuum pumps RUVAC WAU/WSU.
Roots pumps from the RUVAC WAU/WSU
series, being equipped with bypass lines,
can generally be switched on together with
the forepump. The bypass protects these
roots pumps against overloading.
8.3.2.2 Oil change, maintenance work
Under clean operating conditions the oil in
the roots pump will be loaded only as a
result of the natural wear in the bearings
and in the gearing. We nevertheless
recommend making the first oil change
after about 500 hours in service in order to
remove any metal particles which might
have been created by abrasion during the
run-in period. Otherwise it will be suffi-
cient to change the oil every 3000 hours in
operation. When extracting gases contai-
ning dust or where other contaminants are
present, it will be necessary to change the
oil more frequently. If the pumps have to
run at high ambient temperatures, then the
oil in the sealing ring chamber shall also
be replaced at each oil change.
We recommend using our specially formu-
lated N 62 oil.
Under dirty operating conditions it is
possible for dust deposits to form a
crust in the pump chamber. These conta-
minants will deposit in part in the pumping
p where
E
p
k
th
max
1
k
th
Theoretical pumping speed for the roots pump
Nominal pumping speed for the roughing pump
D00
D00 E 19.06.2001 21:40 Uhr Seite 137
chamber and in part on the pumps impel-
lers. They may be removed, once the two
connection lines have been detached, eit-
her by blowing out the system with dry
compressed air or by rinsing with a suita-
ble cleaning agent, such as petroleum
ether (naphtha).
The oil in the roots pump will then have to
be changed. The rotor may be turned only
by hand during cleaning since, due to the
high speed when the motor is running, the
deposits could damage the pump as they
dislodge from the surfaces.
Grime which cannot be eliminated by
washing can be removed mechanically
using a wire brush, metallic scrubber or
scraper.
Important!
The dislodged residues may not remain in
the pump chamber. After cleaning is com-
pleted check the pump for operability by
slowly turning the impellers by hand.
There may be no resistance to rotation. It
is generally not necessary to dismantle the
roots pump. If this should nevertheless be
required due to heavy soiling, then it is
highly advisable to have this done by the
manufacturer.
8.3.2.3 Actions in case of operational
disturbances
1. Pump becomes too warm: (maximum
operating temperature 100 to 115 C)
Possible causes:
Overloading: Excessive heat of com-
pression due to an excessively high
pressure ratio. Check the pressure
value adjustments and the tightness of
the vacuum chamber!
Incorrect clearances: The distances
between the rotors and the housings
have been narrowed due to dirt or
mechanical strain.
Soiled bearings: Excessive friction in
the bearings
Improper oil level: If the oil level is too
high, the gears will touch the oil,
causing friction resistance. Where the
oil level is too low the system will not be
lubricated.
Incorrect oil type: An SAE 30 grade oil
must be used for the pump.
2. Excessive power consumption: All the
factors which can lead to elevated tem-
peratures can also cause excessisve
amounts of power to be drawn. The
motor is defective if excessive power
requirements are not accompanied by a
rise in the temperature at the pump.
3. Oiling at the pump chamber:
Possible causes:
Oil level too high: Oil is subjected to
excessive thermal loading. Oil foam is
swept along.
Oil mixed with the product: Azeotropic
degasification of the oil.
Pump leaking: Air ingress through the
oil drain or filler screw will cause a large
stream of air and conveyance of oil into
the pumping chamber.
4. Abnormal running noises:
Possible causes:
Grime at the impeller
Bearing or gearing damage
Impellers are touching the housing
In the case of bearing or gearing damage
or where the impeller scrapes the housing
the pump should be repaired only by the
manufacturer.
8.3.3 Turbomolecular pumps
8.3.3.1 General operating instructions
In spite of the relatively large gap between
the pump rotor and the stator, the turbo-
molecular pumps should be protected
against foreign objects entering through
the intake port. It is for this reason that the
pump should never be operated without
the supplied wire-mesh splinter guard. In
addition, sharp shock to the pump when
running and sudden changes in attitude
are to be avoided.
Over and above this, and particularly for
the larger types with long rotor blades,
airing the pump to atmospheric pressure
while the impellers are rotating may be
done only when observing exactly the
rules given in the operating instructions.
Under certain circumstances it is possible
to operate turbomolecular pumps under
exceptional conditions.
The turbomolecular pump may, for exam-
ple, be used unprotected inside magnetic
fields if the magnetic induction at the sur-
face of the pump casing does not exceed
B = 3 10
-3
T when radially impinging and
B = 15 10
-3
T at axial impingement.
In a radioactive environment standard tur-
bomolecular pumps can be used without
hazard at dose rates of from 10
4
to 10
5
rad. If higher dose rates are encountered,
then certain materials in the pump can be
modified in order to withstand the greater
loading. The electronic frequency conver-
ters in such cases are to be set up outside
the radioactive areas since the semicon-
ductors used inside them can tolerate a
dose rate of only about 10
3
rad. The use of
motor-driven frequency converters which
can withstand up to 10
8
rad represents
another option.
Roughing (backing) pumps are required
for the operation of turbomolecular
pumps. Depending on the size of the ves-
sel to be evacuated, the turbomolecular
pumps and forepumps may be switched
on simultaneously. If the time required to
pump the vessel down to about 1 mbar
using the particular fore-pump is longer
than the run-up time for the pump (see
operating instructions), then it is advisab-
le to delay switching on the turbomolecu-
lar pump. Bypass lines are advisable when
using turbomolecular pumps in systems
set up for batch (cyclical) operations in
order to save the run-up time for the
pump. Opening the high vacuum valve is
not dangerous at pressures of about
10
-1
mbar.
8.3.3.2 Maintenance
Turbomolecular pumps and frequency
converters are nearly maintenance-free. In
the case of oil-lubricated pumps it is
necessary to replace the bearing lubricant
at certain intervals (between 1500 and
2500 hours in operation, depending on the
type). This is not required in the case of
grease-lubricated pumps (lifetime lubri-
cation). If it should become necessary to
clean the pumps turbine unit, then this
can easily be done by the customer, obser-
ving the procedures described in the ope-
rating instructions.
D00 E 19.06.2001 21:40 Uhr Seite 138
8.3.4 Diffusion and vapor-jet
vacuum pumps
8.3.4.1 Changing the pump fluid and
cleaning the pump
Changing the pump fluid is always ne-
cessary whenever the pump no longer
achieves the required ultimate vacuum or
when its pumping speed falls off. The ser-
vice life of the pump fluid will as a rule
come to several weeks or months or
even years and will depend largely on the
operating conditions for the pump. It is
reduced by frequent pumping at high pres-
sures, by exposure to aggressive vapors
and by air ingress of longer duration (this
does not apply to silicone oil and mer-
cury).
In the case of oil diffusion and vapor-jet
pumps the danger presented to the pump
fluid where there is air ingress with the
pump hot is often overestimated. Where
air ingress (up to atmospheric pressure) is
encountered only occasionally and for
short periods of time then silicone oil will
not be attacked at all and the DIFFELEN
pump fluid will only barely be affected. The
products with considerably higher vapor
pressures which can be created through
oxidation are removed again in a short
period of time by the fractionation and
degassing equipment in the pump (see
Section 2.1.6.1). Even though the pump
fluid which was originally light in color has
been turned brown by air ingress, this
need not necessarily mean that the medi-
um has become unusable. If, on the other
hand, the pump fluid has turned cloudy
and has become more viscous, as well
(which may be the case following periods
of air ingress lasting for several minutes)
then the medium will have to be replaced.
It is possible that under certain circum-
stances cracking products from the pump
fluid may make the oil in the forepump
unserviceable so that here, too, an oil
change will have to be made.
Mercury diffusion and vapor-jet pumps are
less sensitive to air ingress than oil diffusi-
on pumps. The oxidation of the hot mer-
cury caused by the air ingress is negligible
in regard to the operating characteristics
of the pump when compared with the mer-
cury loss in the forepump line.
Changing the pump fluid: The interior sec-
tion will be extracted from the pump and
the contaminated pump fluid poured out.
The interior section and the pump body are
then cleaned with pure petroleum ether
(naphtha). The interior section and pump
body of mercury pumps should have been
cleaned beforehand with a clean brush;
use a bottle brush for the nozzle bores.
Ensure that all the nozzle orifices are pro-
perly cleaned. It is advantageous to evapo-
rate all solvent residues in a drying kiln.
Then the inside section is inserted once
again and the fresh pump fluid is installed
through the forevacuum port. It is neces-
sary to ensure that the upper nozzle cover
is not moistened with pump fluid! Do not
install too much pump fluid!
8.3.4.2 Operating errors with diffusion
and vapor-jet pumps
Potential sources of defects when the
desired ultimate pressure is not reached
Coolant temperature is too high; inade-
quate water throughput. The coolant
flow should always be monitored by a
water flowmeter in order to protect the
pump from damage. Remedy: Measure
the exit temperature of the coolant
water (it should not exceed 30 C).
Increase the coolant flow-through rate.
The cooling coils at the pump may have
to be decalcified.
Forevacuum pressure too high: This is
possible particularly where vapors
which are either evacuated from the
vessel or are created as cracking pro-
ducts from the driving medium (e.g.
following air ingress) get into the roug-
hing pump. Check the forevacuum
pressure with the oil diffusion pump
disconnected. Remedy: Run the fore-
vacuum pump for an extended period of
time with gas ballast. If this is not suffi-
cient, then the oil in the forepump will
have to be changed.
Pump fluid at the diffusion pump spent
or unserviceable: Replace the driving
medium.
Heating is incorrect: Check the heating
output and check for good thermal
contact between the heating plate and
the bottom of the boiler section. Repla-
ce the heating unit if necessary.
Leaks, contamination: Remedy: if the
pump has been contaminated by vapors
it may help to use a metering valve to
pass air through the apparatus for
some period of time; here the pressure
should not exceed 10
-2
mbar when
DIFFELEN is being used.
Measurement system old or soiled (see
Section 8.4.2).
Potential sources of error where there is
insufficient pumping speed:
Forevacuum pressure is too high:
Check the forevacuum; allow the gas
ballast pump to run for a longer period
of time with the gas ballast valve open.
It may be necessary to change the oil in
the forepump.
The pump fluid in the diffusion pump
has become unserviceable: Driving
medium will have to be replaced.
Nozzles at the diffusion pump are clog-
ged: Clean the diffusion pump.
Heating is too weak: Check heating out-
put; examine heating plate for good
thermal contact with the bottom of the
boiler chamber.
Substances are present in the vacuum
vessel which have a higher vapor pres-
sure than the driving medium being
used: among these are, for example,
mercury, which is particularly hazard-
ous because the mercury vapors will
form amalgams with the nonferrous
metals in the oil diffusion pump and
thus make it impossible to achieve per-
fect vacuums.
8.3.5 Adsorption pumps
8.3.5.1 Reduction of adsorption
capacity
A considerable reduction in pumping
speed and failure to reach the ultimate
pressure which is normally attainable in
spite of thermal regeneration having been
carried out indicates that the zeolite being
used has become contaminated by outside
substances. It does not make good sense
to attempt to rejuvenate the contaminated
zeolite with special thermal processes. The
zeolite should simply be replaced.
8.3.5.2 Changing the molecular sieve
It will be necessary to rinse the adsorption
pump thoroughly with solvents before
installing the new zeolite charge. Before
putting the adsorption pump which has
D00
D00 E 19.06.2001 21:40 Uhr Seite 139
been charged with fresh zeolite into servi-
ce it is also necessary to bake out the new
zeolite charge, under vacuum and using
the heating element associated with the
pump, for a period of several hours so that
contaminants which might have collected
during the storage period can dissipate.
8.3.6 Titanium sublimation
pumps
Each of the turns in the titanium sublimati-
on pump contains approximately 1.2 g of
useable titanium supply. At a heating cur-
rent of 50 A the surface temperature comes
to about 1850 K, the sublimation rate
approximately 0.12 g/h, i.e. a turn can be
operated continuously for about 10 hours.
Since at pressures below 1 10
-6
mbar sub-
limation is not continuous but rather only at
intervals which at low pressures (below
5 10
-8
mbar) and low gas volumes are
already more than ten times the actual sub-
limation period, one may assume a pum-
ping period of almost one month at a wor-
king pressure of 10
-10
mbar per turn.
The effective pumping speed of a titanium
sublimation pump will depend on the get-
ter screen surface and the geometry of the
suction opening. The pumping speed, refe-
renced to the surface area of the getter
surface, will be dependent on the type of
gas and the getter screen temperature.
Since inert gases, for example, cannot be
pumped at all, titanium sublimation pumps
should always be used only with an auxili-
ary pump (sputter-ion pump, turbomole-
cular pump) used to pump these gas com-
ponents. The supplementary pump can be
much smaller than the titanium sublimati-
on pump. Only in a few special cases can
one do without the additional pump.
The selection of the coolant is dictated by
the working conditions and the require-
ments in terms of ultimate pressure. At
high pressures, above 1 10
-6
mbar, more
thermal energy is applied to the getter
screen by frequent sublimation cycles. It is
for this reason that cooling with liquid
nitrogen is more favorable. At low pressu-
res water cooling may be sufficient. The
getter screen should if at all possible be
heated to room temperature before airing
the pump as otherwise the humidity in the
air would condense out on the surface.
8.3.7 Sputter-ion pumps
Sputter-ion pumps use high-voltage cur-
rent. Installation and connection should be
carried out only under the supervision of a
qualified specialist. The operating instruc-
tions shall be observed!
The service life of sputter-ion pumps
depends linearly on the pumps operating
pressure. In the case of pumps manufac-
tured by LEYBOLD, the following applies:
p T = 45 10
-3
mbar h
(p = operating pressure, T = service life)
This means that for operating pressure of
10
-3
mbar
the service life is 45 hours
10
-6
mbar
the service life is 45,000 hours
10
-9
mbar
the service life is 45,000,000 hours
If a triode pump is not needed over an
extended period of time it can either be
operated continuously at low pressure
with practically no influence on the service
life or it can be aired, removed from the
pump and packed in a dust-tight container.
The starting properties of the sputter-ion
(triode) pumps manufactured by
LEYBOLD are so good that no problems
will be encountered when returning the
units to service, even after a longer period
in storage.
When the sputter-ion pumps are installed
one should ensure that the magnetic fields
will not interfere with the operation of
other devices (ionization vacuum gauges,
partial pressure measurement units, etc.).
Mounting devices for the sputter-ion
pumps may not short circuit the induc-
tance flow and thus weaken the air gap
inductance and pumping speed.
If the ultimate pressure which can be attai-
ned is not satisfactory even though the
apparatus is properly sealed, then it will
usually be sufficient to bake out the atta-
ched equipment at about 200 to 250 C. If
the pressure here rises to about
1 10
-5
mbar when this is done, then the
sputter ion pump will become so hot after
evacuating the gases for two hours that it
will not be necessary to heat it any further
in addition. It is also possible to heat the
pump by allowing air to enter for 2 hours
at 10
-5
mbar before the other apparatus is
then subsequently baked out. If the ultima-
te pressure is still not satisfactory, then the
pump itself will have to be baked out for a
few hours at 250 to 300 C (but not higher
than 350 C!). The pump should without
fail remain in operation throughout this
period! If the pressure rises above
5 10
-5
mbar it will be necessary either to
heat more slowly or to connect an auxiliary
pump. Before airing one should allow the
hot sputter-ion pump enough time to cool
down at least to 150 C.
8.4 Information on
working with
vacuum gauges
8.4.1 Information on installing
vacuum sensors
Here both the external situation in the
immediate vicinity of the vacuum appara-
tus and the operating conditions within the
apparatus (e.g. working pressure, compo-
sition of the gas content) will be important.
It is initially necessary to ascertain whether
the measurement system being installed is
sensitive in regard to its physical attitude.
Sensors should only be installed vertically
with the vacuum flange at the bottom to
keep condensates, metal flakes and filings
from collecting in the sensor or even small
components such as tiny screws and the
like from falling into the sensorand the
measurement system. The hot incandes-
cent filaments could also bend and deform
improperly and cause electrical shorts
inside the measurement system. This is
the reason behind the following general
rule: If at all possible, install sensors
vertically and open to the bottom. It is
also very important to install measurement
systems if at all possible at those points in
the vacuum system which will remain free
of vibration during operation.
The outside temperature must be taken
into account and above all it is necessary
to avoid hot kilns, furnaces or stoves or
other sources of intense radiation which
generate an ambient temperature around
the measurement system which lies above
the specific acceptable value. Excessive
ambient temperatures will result in false
pressure indications in thermal conductivi-
ty vacuum sensors.
D00 E 19.06.2001 21:40 Uhr Seite 140
8.4.2 Contamination at the
measurement system and
its removal
The vacuum gauges used in vacuum tech-
nology for pressure measurement will cer-
tainly work under dirty conditions. This
is quite understandable since a vacuum
device or system does not serve simply to
produce low pressures but rather and pri-
marily have to run processes in chemistry,
metallurgy or nuclear physics at low pres-
sures. Here, depending on the nature of
the process, considerable quantities of
gases or vapors will be liberated either
continuously or intermittently; these can
pass into the measurement systems provi-
ded for pressure measurement and instal-
led in the vacuum system and due to
surface reactions or through simple depo-
sits can falsify the pressure measure-
ments considerably. This is true for all
types of vacuum gauges whereby, of cour-
se, high-sensitivity, high-accuracy measu-
rement systems are particularly suscep-
tible to soiling resulting from the causes
named. One can attempt to protect the
measurement systems against contam-
ination by providing suitable shielding.
This, however, will often lead to the pres-
sure registered by the measurement
system which is indeed clean deviating
considerably from the pressure actually
prevailing in the system.
It is not fundamentally possible to keep the
measurement system in a vacuum gauge
from becoming soiled. Thus it is necessary
to ensure that
the influence of the contamination on
pressure measurement remains as
small as possible and that
the measurement system can readily be
cleaned.
These two conditions are not easy to satis-
fy by most vacuum gauges in practice.
Dirt in a compression vacuum gauge will
cause an incorrect and uncontrollable
pressure indication. Dirty THERMOVAC
sensors will show a pressure which is too
high in the lower measurement range
since the surface of the hot wire has chan-
ged. In Penning vacuum gauges contami-
nation will induce pressure readings which
are far too low since the discharge cur-
rents will become smaller. In the case of
ionization vacuum gauges with hot catho-
des, electrodes and the tube walls can be
soiled which, under certain circumstances,
will result in a reduction of dielectric
strengths. Here, however, the measure-
ment systems can usually be baked out
and degassed by passing a current
through or by electron bombardment,
quite aside from the fact that ionization
vacuum gauges are often used in the ultra-
high vacuum range where it is necessary
to ensure clean operating conditions for
other reasons.
8.4.3 The influence of
magnetic and electrical
fields
In all those measurement instruments
which use the ionization of gas molecules
as the measurement principle (cold-catho-
de and hot-cathode ionization vacuum
gauges), strong magnetic leakage fields or
electrical potentials can have a major influ-
ence on the pressure indication. At low
pressures it is also possible for wall poten-
tials which deviate from the cathode
potential to influence the ion trap current.
In vacuum measurement systems used in
the high and ultrahigh regimes it is neces-
sary to ensure in particular that the requi-
red high insulation values for the high-vol-
tage electrodes and ion traps also be main-
tained during operation and sometimes
even during bake-out procedures. Insulati-
on defects may occur both in the external
feed line and inside the measurement
system itself. Insufficient insulation at the
trap (detector) lead may allow creep cur-
rents at low pressures to stimulate
overly high pressure value readings. The
very low ion trap currents make it neces-
sary for this lead to be particularly well
insulated. Inside the measurement sen-
sors, too, creep currents can occur if the
trap is not effectively shielded against the
other electrodes.
An error frequently made when connecting
measurement sensors to the vacuum
system is the use of connector piping
which is unacceptably long and narrow.
The conductance value must in all cases
be kept as large as possible. The most
favorable solution is to use integrated
measurement systems. Whenever connec-
tor lines of lower conductance values are
used the pressure indication, depending
on the cleanliness of the measurement
sensors and the connector line, may be
either too high or too low. Here measure-
ment errors by more than one complete
order of magnitude are possible! Where
systems can be baked out it is necessary
to ensure that the connector line can also
be heated.
8.4.4 Connectors, power pack,
measurement systems
The measurement cables (connector
cables between the sensor and the vacuum
gauge control unit) are normally 2 m long.
If longer measurement cables must be
used for installation in control panels, for
example then it will be necessary to
examine the situation to determine
whether the pressure reading might be fal-
sified. Information on the options for using
over-length cables can be obtained from
our technical consulting department.
D00
D00 E 19.06.2001 21:40 Uhr Seite 141
Tables, Formulas, Diagrams
9. Tables, formulas, nomograms and symbols
Unit N m
2
, Pa
2)
mbar bar Torr
1 N m
2
(= 1 Pa) 1 1 10
2
1 10
5
7.5 10
3
1 mbar 100 1 1 103 0.75
1 bar 1 10
5
1 10
3
1 750
1 Torr
3)
133 1.33 1.33 10
3
1
1) The torr is included in the table only to facilitate the transition from
this familiar unit to the statutory units N m
-2
, mbar and bar. In futu-
re the pressure units torr, mm water column, mm mercury column
(mm Hg), % vacuum, technical atmosphere (at), physicalatmosphe-
re (atm), atmosphere absolute (ata), pressure above atmospheric
and pressure below atmospheric may no longer be used. Reference
is made to DIN 1314 in this context.
2) The unit Newton divided by square meters (N m
-2
) is also
designated as
Pascal (Pa): 1 N m
-2
= 1 Pa.
Newton divided by square meters or Pascal is the SI unit for the pres-
sure of fluids.
3) 1 torr = 4/3 mbar; torr = 1 mbar.
Table I: Permissible pressure units including the torr 1) and its conversion
Abbrev. Gas C* = p
[cm mbar]
H
2
Hydrogen 12.00 10
3
He Helium 18.00 10
3
Ne Neon 12.30 10
3
Ar Argon 6.40 10
3
Kr Krypton 4.80 10
3
Xe Xenon 3.60 10
3
Hg Mercury 3.05 10
3
O
2
Oxygen 6.50 10
3
N
2
Nitrogen 6.10 10
3
HCl Hydrochloric acid 4.35 10
3
CO
2
Carbon dioxide 3.95 10
3
H
2
O Water vapor 3.95 10
3
NH
3
Ammonia 4.60 10
3
C
2
H
5
OH Ethanol 2.10 10
3
Cl
2
Chlorine 3.05 10
3
Air Air 6.67 10
3
Table III: Mean free path l
Values of the product c* of the mean free path ( and pressure p for various
gases at 20 C (see also Fig. 9.1)
1 = mbar Pa dyn cm
2
atm Torr inch Micron cm kp cm
2
lb in
2
lb ft
2
(N/m
3
) (bar) (phys.) (mm Hg) Hg () H
2
O (at tech.) (psi)
mbar 1 10
2
10
3
9.87 10
4
0.75 2.953 10
2
7.5 10
2
1.0
2
1.02 10
3
1.45 10
2
2.089
Pa 10
2
1 10 9.87 10
6
7.5 10
3
2.953 10
4
7.5 1.02 10
2
1.02 105 1.45 10
4
2.089 10
2
bar 10
3
0.1 1 9.87 10
7
7.5 10
4
2.953 10
5
7.5 10
1
1.02 10
3
1.02 10
6
1.45 10
5
2.089 10
3
atm 1013 1.01 10
5
1.01 10
6
1 760 29.92 7.6 10
5
1.03 10
3
1.033 14.697 2116.4
Torr 1.33 1.33 10
2
1.33 10
3
1.316 10
3
1 3.937 10
2
10
3
1.3595 1.36 10
3
1.934 10
2
2.7847
in Hg 33.86 33.9 10
2
33.9 10
3
3.342 10
2
25.4 1 2.54 10
4
34.53 3.453 10
2
0.48115 70.731
1.33 10
3
1.33 10
1
1.333 1.316 10
6
10
3
3.937 10
5
1 1.36 10
3
1.36 10
6
1.934 10
5
2.785 10
3
cm H
2
O 0.9807 98.07 980.7 9.678 10
4
0.7356 2.896 10
2
7.36 10
2
1 10
3
1.422 10
2
2.0483
at 9.81 10
2
9.81 10
4
9.81 10
5
0.968 7.36 10
2
28.96 7.36 10
5
10
3
1 14.22 2048.3
psi 68.95 68.95 10
2
68.95 10
3
6.804 10
2
51.71 2.036 51.71 10
3
70.31 7.03 10
2
1 1.44 10
2
lb ft
2
0.4788 47.88 478.8 4.725 10
4
0. 3591 1.414 10
2
359.1 0.488 4.88 10
4
6.94 10
3
1
Normal conditions: 0 C and sea level, i.e. p = 1013 mbar = 760 mm Hg = 760 torr = 1 atm
in Hg = inches of mercury; 1 mtorr (millitorr) = 10
-3
torr = 1 (micron m Hg column)
Pounds per square inch = lb in
-2
= lb / sqin = psi (psig = psi gauge pressure above atmospheric, pressure gauge reading; psia = psi absolute absolute
pressure)
Pounds per square foot = lb / sqft = lb / ft
2
; kgf/sqcm
2
= kg force per square cm = kp / cm
2
= at; analogously also: lbf / squin = psi
1 dyn cm-
2
(cgs) = 1 bar (microbar) = 1 barye; 1 bar = 0.1 Mpa; 1 cm water column (cm water column = g / cm
2
at 4 C) = 1 Ger (Geryk)
atm physical atmosphere at technical atmosphere; 100 - (x mbar / 10.13) = y % vacuum
Table II: Conversion of pressure units
D00 E 19.06.2001 21:40 Uhr Seite 142
D00
Table IV: Compilation of important formulas pertaining to the kinetic theory of gases
2 R T
VARIABLE
Most probable speed
of particles c
w
Mean velocity
of particles c
Mean square of velocity

of particles c
2
Gas pressure p of particles

Number density of particles n
Area-related impingement Z
A
Volume collision rate Z
V
Equation of state of ideal gas
Area-related mass flow rate q
m
, A
c* = p in cm mbar (see Tab. III)
k Boltzmann constant in mbar l K
1
mean free path in cm
M molar mass in g mol
1
General formula
p = n k T
1
p = n m
T
c
2
3
1
p = % c
2
3
1
Z
A
= p
2
c*
p = 13.80 10
20
n T [mbar]
p = 4.04 10
17
n [mbar] (applies to all gases)
p = 2.5 10
16
p [cm
3
] (applies to all gases)
Z
A
= 2.85 10
20
p [cm
2
s
1
] (see Fig. 78.2)
Z
V
= 8.6 10
22
p
2
[cm
3
s
1
] (see Fig. 78.2)
p V = 2.44 10
4
[mbar `] (for all gases)
p V = 83.14 T [mbar `]
Q
m, A
= 4.377 10
2
p [g cm
2
s
1
]
p V = R T
q
m, A
= Z
A
m
T
= p
n = p/kT
q
m, A
= 1.38 10
2
p g [cm
2
s
1
]
For easy calculation Value for air at 20 C
c
w
= c
w
= 410 [m/s]
c
=
c
2
=
c
2
= 25.16 10
4
c
= 464 [m/s]
cm
2
s
2
3 R T
M
M
T
M
2 k T N
A
N
A
2 M k T
2 N
A
M k T
p
n = 7.25 10
18
[cm
3
]
T
p
Z
A
= 2.63 10
22
p [cm
2
s
1
]
M T
p
2
Z
V
= 5.27 10
22
[cm
3
s
1
]
c*
M
T
1
Z
A
= n c
4
Z
A
= p
[ ]
8 R T
1
Z
V
=
n c
2
c
w
= 1.29 10
4
T cm
M s
[ ]
c
2
= 2.49 10
8
T cm
2
M s
2 [ ]
T cm
M s
c
= 1.46 10
4
[ ]
m
T
particle mass in g
N
A
Avogadro constant in mol
1
n number density of particles in cm
3
amount of substance in mol
p gas pressure in mbar
R molar gas constant in mbar l mol
1
K
1
T thermodynamic temperature in K
V volume in l
Table V: Important values
Designation, Symbol Value and Remarks
alphabetically unit
Atomic mass unit m
u
1.6605 10
27
kg
Avogadro constant N
A
6.0225 10
23
mol
1
Number of particles per mol,
formerly: Loschmidt number
Boltzmann constant k 1.3805 10
23
J K
1
mbar l
13.805 10
23
K
Electron rest mass m
e
9.1091 10
31
kg
Elementary charge e 1.6021 10
19
A s
Molar gas constant R 8.314 J mol
1
K
1
mbar l
= 83.14 R = N
A
k
mol K
Molar volume of 22.414 m
3
kmol
1
DIN 1343; formerly: molar volume
the ideal gas V
o
22.414 l mol
1
at 0 C and 1013 mbar
Standard acceleration of free fall g
n
9.8066 m s
2
Planck constant h 6.6256 10
34
J s
W
Stefan-Boltzmann constant s 5.669 10
8
also: unit conductance, radiation constant
m
2
K
4
e A s
Specific electron charge 1.7588 10
11
me kg
Speed of light in vacuum c 2.9979 108 m s
1
Standard reference density l
n
kg m
3
Density at = 0 C and p
n
= 1013 mbar
of a gas
Standard reference pressure p
n
101.325 Pa = 1013 mbar DIN 1343 (Nov. 75)
Standard reference temperature T
n
T
n
= 273.15 K, J = 0 C DIN 1343 (Nov. 75)
D00 E 19.06.2001 21:40 Uhr Seite 143
Unit l s
1
m
3
h
1
cm
3
s
1
cuft min
1
1 l s
1
1 3.6 1000 2.12
1 m
3
h
1
0.2778 1 277.8 0.589
1 cm
3
s
1
10
3
3.6 10
3
1 2.1 10
3
1 cuft min
1
0.4719 1.699 471.95 1
Table VI: Conversion of pumping speed (volume flow rate) units
Table VIIa: Conversion of throughput (Q
pV
) units; (leak rate) units
1 = ... mbar l/s kg h
1
(20C) kg h
1
(0C) cm
3
/h (NTP) cm
3
/s (NTP) Torr l/s g/a (F12. 20 C) g/a (F12. 25 C) cfm lusec Pa l/s slpm
mbar l/s 1 4.28 10
3
4.59 10
3
3554 0.987 0.75 1.56 105 1.54 105 1593 7.52 10
2
100 59.2 10
3
kg h
1
(20 C) 234 1 1.073 8.31 10
5
231 175 37.2 10
4
1.75 10
5
23.4 10
3
13.86
kg h
1
(0 C) 218 0.932 1 7.74 10
5
215 163 34.6 10
4
1.63 10
5
21.8 10
3
12.91
cm
3
/h (NTP) 2.81 10
4
1.20 10
6
1.29 10
6
1 2.78 10
4
2.11 10
4
44 44.7 10
2
2.11 10
1
2.81 10
2
1.66 10
5
cm
3
/s (NTP) 1.013 4.33 10
3
4.65 10
3
3600 1 0.760 1.58 10
5
1611 760 101 6 10
2
Torr l/s 1.33 5.70 10
3
6.12 10
3
4727 1.32 1 2.08 10
5
2.05 10
5
2119 1 10
3
133 78.8 10
3
g/a (F12. 20 C) 6.39 10
6
2.27 10
2
6.31 10
6
4.80 10
6
1 10.2 10
3
4.8 10
3
6.39 10
4
37.9 10
8
g/a (F12. 25 C) 6.50 10
6
4.88 10
6
1 10.4 10
3
4.89 10
3
6.5 10
4
38.5 10
8
cfm 6.28 10
4
2.69 10
6
2.89 10
6
2.24 6.21 10
4
4.72 10
4
98.16 96.58 1 0.472 6.28 10
2
37.2 10
6
lusec 1.33 10
3
5.70 10
6
6.12 10
6
4.737 1.32 10
3
1 10
3
208 205 2.12 1 13.3 10
2
78.8 10
6
Pa l/s 1 10
2
4.28 10
5
4.59 10
5
35.54 9.87 10
3
7.5 10
3
1.56 10
3
1.54 10
3
15.93 7.50 1 59.2 10
5
slpm 16.88 72.15 10
3
77.45 10
3
60.08 10
3
16.67 12.69 2.64 10
6
2.60 10
6
26.9 10
3
12.7 10
3
16.9 10
2
1
1 cm
3
(NTP) = 1 cm
3
under normal conditions (T = 273.15 K; p = 1013.25 mbar) NTP = normal temperature and pressure (1 atm; 0 C) R = 83.14 mbar l mol
1
K
1
1 cm
3
(NTP) h
1
= 1 atm cm
3
h
1
= 1 Ncm
3
h
1
= 1 std cch 1 sccm = 10
3
slpm = 10
3
N l min
1
= 60 sccs
SI coherent: 1 Pa m
3
s
1
= 10 mbar l s
1
; R = 8.314 Pa m
3
mol
1
K
1
; M in kg / mol
1 cm
3
(NTP) s
1
= 1 sccs = 60 cm
3
(NTP) min
1
60 sccm = 60 std ccm = 60 Ncm
3
min
1
1 lusec = 1 l s
1
1 = 1 micron = 10
3
Torr 1 lusec = 10
3
Torr l s
1
Freon F 12 (CCl
2
F
2
) M = 120.92 g mol
1
; air M = 28.96 g mol
1
Note: Anglo-American units are not abbreviated nonuniformly! Example: Standard cubic centimeter per minute sccm = sccpm = std ccm = std ccpm
1= ... mbar l/s cm
3
/s **) Torr l/s Pa m
3
/s g/a *) oz/yr *) lb/yr *) atm ft
3
/min l/s ft
3
/h ft
3
/min
mbar l/s 1 0.987 0.75 10
1
1.56 10
5
5.5 10
3
3.4 10
2
2.10 10
3
7.52 10
2
9.56 10
4
1593
cm
3
/s **) 1.013 1 0.76 1.01 10
1
1.58 10
5
5.6 10
3
3.44 10
2
2.12 10
3
760 96.6 10
3
1614
Torr l/s 1.33 1.32 1 1.33 10
1
2.08 10
5
7.3 10
3
4.52 10
2
2.79 10
3
10
3
1.27 10
5
2119
Pa m
3
/s 10 9.9 7.5 1 1.56 10
6
5.51 10
4
3.4 10
3
2.09 10
2
7.5 10
3
9.54 10
5
15.9 10
3
g/a *) 6.39 10
6
6.31 10
6
4.80 10
6
6.41 10
7
1 3.5 10
2
2.17 10
3
1.34 10
8
4.8 10
3
0.612 10.2 10
3
oz/yr *) 1.82 10
4
1.79 10
4
1.36 10
4
1.82 10
5
28.33 1 6.18 10
2
3.80 10
7
0.136 17.34 0.289
lb/yr *) 2.94 10
3
2.86 10
3
2.17 10
3
2.94 10
4
4.57 10
2
16 1 6.17 10
6
2.18 280 4.68
atm ft
3
/min 4.77 10
2
4.72 10
2
3.58 10
2
47.7 7.46 10
7
2.63 10
6
1.62 10
5
1 3.58 10
5
4.55 10
7
7.60 10
5
l/s 1.33 10
3
1.32 10
3
10
3
1.33 10
4
208 7.34 4.52 10
1
2.79 10
6
1 127 2.12
ft
3
/h 1.05 10
5
1.04 10
5
7.87 10
6
1.05 10
6
1.63 5.77 10
2
3.57 10
3
2.20 10
8
7.86 10
3
1 1.67 10
2
ft
3
/min 6.28 10
4
6.20 10
4
4.72 10
4
6.28 10
5
98 3.46 2.14 10
1
1.32 10
6
0.472 60 1
1 ft
3
h
1
= 1.04 10
5
stsd cc per second 1 micron cubic foot per hour = 0.0079 micron liter per second 1 kg = 2.2046 pounds (lb)
1 cm
3
s
1
(NTP) = 1 atm cm
3
s
1
= 1 scc s
1
= 1 sccss 1 micron liter per second = 0.0013 std cc per second = 1 lusec 1 cubic foot (cfut. cf) =
28.3168 dm
3
1 atm ft
3
min
1
= 1 cfm (NTP) 1 micron cubic foot per minute = 1 ft
3
min
1
= 1 cuft min
1
= 1 cfm
1 lb = 16 ounces (oz)
1 Pa m
3
/s = 1 Pa m
3
/s (anglo-am.) = 10
3
Pa l/s 1 standard cc per second = 96.600 micron cubic feet per hour 1 lusec = 1 l s
1
1 l s
1
= 127 ft
3
h
1
= 0.0013 std cc per second = 1 lusec 1 std cc/sec = 760 l s
1
*) F12 (20 C) C.Cl
2
F
2
M = 120.92 h/mol **) (NTP) normal temperature and pressure 1 atm und 0 C
Table VII b: Conversion of throughput (Q
pV
) units; (leak rate) units
D00 E 19.06.2001 21:40 Uhr Seite 144
D00
% by weight % by volume Partial pressure mbar
N
2
75.51 78.1 792
O
2
23.01 20.93 212
Ar 1.29 0.93 9.47
CO
2
0.04 0.03 0.31
Ne 1.2 10
3
1.8 10
3
1.9 10
2
He 7 10
5
7 10
5
5.3 10
3
CH
4
2 10
4
2 10
4
2 10
3
Kr 3 10
4
1.1 10
4
1.1 10
3
N
2
O 6 10
5
5 10
5
5 10
4
H
2
5 10
6
5 10
5
5 10
4
Xe 4 10
5
8.7 10
6
9 10
5
O
3
9 10
6
7 10
6
7 10
5
100 % 100 % 1013
50 % RH at 20 C 1.6 1.15 11.7
Note: In the composition of atmospheric air the relative humidity (RH) is indicated separately along
with the temperature.
At the given relative humidity, therefore, the air pressure read on the barometer is 1024 mbar
Table VIII: Composition of atmospheric air
Pressure p [mbar] 1013 1 1 10
3
10
3
10
7
< 10
7
Particle number density n [cm
3
] 10
19
10
16
10
16
10
13
10
13
10
9
< 10
9
Mean free path [cm] < 10
2
10
2
10 10 10
5
> 10
5
Impingement rate Z
a
[cm
2
s
1
] 10
23
10
20
10
20
10
17
10
17
10
13
< 10
13
Vol.-related collision rate Z
V
[cm
3
s
1
] 10
29
10
23
10
23
10
17
10
17
10
9
< 10
9
Monolayer time [s] < 10
5
10
5
10
2
10
2
100 > 100
Type of gas flow Viscous flow Knudsen flow Molecular flow Molecular flow
Other special features Convection dependent Significant change in Significant reduction- Particles on the
on pressure thermal conductivity in volume surfaces dominate
of a gas related collision rate to a great extend in
relation to particles in
gaseous space
Table IX: Pressure ranges used in vacuum technology and their characteristics (numbers rounded off to whole power of ten)
At room temperature
Standard values1 Metals Nonmetals
(mbar l s
1
cm
2
) 10
9
... 10
7
10
7
... 10
5
Outgassing rates (standard values) as a function of time
Examples: 1/2 hr. 1 hr. 3 hr. 5 hr. Examples: 1/2 hr. 1 hr. 3 hr. 5 hr.
Ag 1.5 10
8
1.1 10
8
2 10
9
Silicone 1.5 10
5
8 10
6
3.5 10
6
1.5 10
6
Al 2 10
8
6 10
9
NBR 4 10
6
3 10
6
1.5 10
6
1 10
6
Cu 4 10
8
2 10
8
6 10
9
3.5 10
9
Acrylic glass 1.5 10
6
1.2 10
6
8 10
7
5 10
7
Stainless steel 9 10
8
3.5 10
8
2.5 10
8
FPM, FKM 7 10
7
4 10
7
2 10
7
1.5 10
7
1
All values depend largely on pretreatment!
Table X: Outgassing rate of materials in mbar l s
1
cm
2
D00 E 19.06.2001 21:40 Uhr Seite 145
Nominal width (DN) Internal diam. (mm)
Series R5 R10
10 10
16 16
20 21
25 24
32 34
40 41
50 51
63 70
80 83
100 102
125 127
160 153
200 213
250 261
320 318
400 400
500 501
630 651
800 800
1000 1000
The nominal internal diameters correspond approximately to the
internal diameters of the pipeline components (DIN 2402 - Feb.
1976). The left-hand column of the nominal internal diameter series
is preferred in practice.
Table XI: Nominal internal diameters (DN) and internal diameters of tubes, pipes and
apertures with circular cross-section (according to PNEUROP)
Solvent Relative Density Melting Boiling Maximum admissible
molecular g / cm
3
point point concentration (MAC)
mass (20 C) C C cm
3
/ m
3
Acetone 58 0.798 56
Benzene (solution) 78 0.8788 5.49 80.2 25
Petrol (light) 0.68 ... 0.72 > 100
Carbon tetrachloride 153.8 1.592 22.9 76.7 25
Chloroform 119.4 1.48 63.5 61 50
Diethyl ether 46 0.7967 114.5 78 1000
Ethyl alcohol 74 0.713 116.4 34.6 400
Hexane 86 0.66 93.5 71 500
Isopropanol 60.1 0.785 89.5 82.4 400
Methanol 32 0.795 97.9 64.7 200 (toxic!)
Methylene chloride 85 1.328 41
Nitromethane 61 1.138 29.2 101.75 100
Petroleum ether mixture 0.64 40 ... 60
Trichlorethylene (Tri) 131.4 1.47 55
Water 18.02 0.998 0.00 100.0
Table XII: Important data (characteristic figures) for common solvents
D00 E 19.06.2001 21:40 Uhr Seite 146
D00
t ps %
D
C mbar g/m
3
100 1.403 10
5
1.756 10
5
99 1.719 2.139
98 2.101 2.599
97 2.561 3.150
96 3.117 3.812
95 3.784 10
5
4.602 10
5
94 4.584 5.544
93 5.542 6.665
92 6.685 7.996
91 8.049 9.574
90 9.672 10
5
11.44 10
5
89 11.60 13.65
88 13.88 16.24
87 16.58 19.30
86 19.77 22.89
85 23.53 10
5
27.10 10
5
84 27.96 32.03
83 33.16 37.78
82 39.25 44.49
81 46.38 52.30
80 0.5473 10
3
0.6138 10
3
79 0.6444 0.7191
78 0.7577 0.8413
77 0.8894 0.9824
76 1.042 1.145
75 1.220 10
3
1.334 10
3
74 1.425 1.550
73 1.662 1.799
72 1.936 2.085
71 2.252 2.414
70 2.615 10
3
2.789 10
3
69 3.032 3.218
68 3.511 3.708
67 4.060 4.267
66 4.688 4.903
65 5.406 10
3
5.627 10
3
64 6.225 6.449
63 7.159 7.381
62 8.223 8.438
61 9.432 9.633
60 10.80 10
3
10.98 10
3
59 12.36 12.51
58 14.13 14.23
57 16.12 16.16
56 18.38 18.34
55 20.92 10
3
20.78 10
3
54 23.80 23.53
53 27.03 26.60
52 30.67 30.05
51 34.76 33.90
50 39.35 10
3
38.21 10
3
49 44.49 43.01
48 50.26 48.37
47 56.71 54.33
46 63.93 60.98
45 71.98 10
3
68.36 10
3
44 80.97 76.56
43 90.98 85.65
42 102.1 95.70
41 114.5 10
3
106.9 10
3
40 0.1283 0.1192
39 0.1436 0.1329
38 0.1606 0.1480
37 0.1794 0.1646
36 0.2002 0.1829
t ps %
D
C mbar g/m
3
35 0.2233 0.2032
34 0.2488 0.2254
33 0.2769 0.2498
32 0.3079 0.2767
31 0.3421 0.3061
30 0.3798 0.3385
29 0.4213 0.3739
28 0.4669 0.4127
27 0.5170 0.4551
26 0.5720 0.5015
25 0.6323 0.5521
24 0.6985 0.6075
23 0.7709 0.6678
22 0.8502 0.7336
21 0.9370 0.8053
20 1.032 0.8835
19 1.135 0.9678
18 1.248 1.060
17 1.371 1.160
16 1.506 1.269
15 1.652 1.387
14 1.811 1.515
13 1.984 1.653
12 2.172 1.803
11 2.376 1.964
10 2.597 2.139
9 2.837 2.328
8 3.097 2.532
7 3.379 2.752
6 3.685 2.990
5 4.015 3.246
4 4.372 3.521
3 4.757 3.817
2 5.173 4.136
1 5.623 4.479
0 6.108 4.847
1 6.566 5.192
2 7.055 5.559
3 7.575 5.947
4 8.129 6.360
5 8.719 6.797
6 9.347 7.260
7 10.01 7.750
8 10.72 8.270
9 11.47 8.819
10 12.27 9.399
11 13.12 10.01
12 14.02 10.66
13 14.97 11.35
14 15.98 12.07
15 17.04 12.83
16 18.17 13.63
17 19.37 14.48
18 20.63 15.37
19 21.96 16.31
20 23.37 17.30
21 24.86 18.34
22 26.43 19.43
23 28.09 20.58
24 29.83 21.78
25 31.67 23.05
26 33.61 24.38
27 35.65 25.78
28 37.80 27.24
29 40.06 28.78
t ps %
D
C mbar g/m
3
30 42.43 30.38
31 44.93 32.07
32 47.55 33.83
33 50.31 35.68
34 53.20 37.61
35 56.24 39.63
36 59.42 41.75
37 62.76 43.96
38 66.26 46.26
39 69.93 48.67
40 73.78 51.19
41 77.80 53.82
42 82.02 56.56
43 86.42 59.41
44 91.03 62.39
45 95.86 65.50
46 100.9 68.73
47 106.2 72.10
48 111.7 75.61
49 117.4 79.26
50 123.4 83.06
51 129.7 87.01
52 136.2 91.12
53 143.0 95.39
54 150.1 99.83
55 157.5 104.4
56 165.2 109.2
57 173.2 114.2
58 181.5 119.4
59 190.2 124.7
60 199.2 130.2
61 208.6 135.9
62 218.4 141.9
63 228.5 148.1
64 293.1 154.5
65 250.1 161.2
66 261.5 168.1
67 273.3 175.2
68 285.6 182.6
69 298.4 190.2
70 311.6 198.1
71 325.3 206.3
72 339.6 214.7
73 354.3 223.5
74 369.6 232.5
75 385.5 241.8
76 401.9 251.5
77 418.9 261.4
78 436.5 271.7
79 454.7 282.3
80 473.6 293.3
81 493.1 304.6
82 513.3 316.3
83 534.2 328.3
84 555.7 340.7
85 578.0 353.5
86 601.0 366.6
87 624.9 380.2
88 649.5 394.2
89 674.9 408.6
90 701.1 423.5
91 728.2 438.8
92 756.1 454.5
93 784.9 470.7
94 814.6 487.4
t ps %
D
C mbar g/m
3
95 845.3 504.5
96 876.9 522.1
97 909.4 540.3
98 943.0 558.9
99 977.6 578.1
100 1013.2 597.8
101 1050 618.0
102 1088 638.8
103 1127 660.2
104 1167 682.2
105 1208 704.7
106 1250 727.8
107 1294 751.6
108 1339 776.0
109 1385 801.0
110 1433 826.7
111 1481 853.0
112 1532 880.0
113 1583 907.7
114 1636 936.1
115 1691 965.2
116 1746 995.0
117 1804 1026
118 1863 1057
119 1923 1089
120 1985 1122
121 2049 1156
122 2114 1190
123 2182 1225
124 2250 1262
125 2321 1299
126 2393 1337
127 2467 1375
128 2543 1415
129 2621 1456
130 2701 1497
131 2783 1540
132 2867 1583
133 2953 1627
134 3041 1673
135 3131 1719
136 3223 1767
137 3317 1815
138 3414 1865
139 3512 1915
140 3614 1967
Sources: Smithsonian Meteorological Tables 6th. ed. (1971) and VDI vapor tables 6th ed (1963).
Table XIII: Saturation pressure p
s
and vapor density %
D
of water in a temperature range from 100 C to +140 C
D00 E 19.06.2001 21:40 Uhr Seite 147
Group A
3)
Methane c
Ethane c
Propane c
Butane c
Pentane c
Hexane c
Heptane c
Octane a
Cyclohexane c
Propylene (propene) a
Styrene (s) b
Benzene (s) c
Toluene (s)
Xylene a
Naphthalene
Methanol (s) c
Ethanol (s) c
Propyl alcohol (propanol) c
Butyl alcohol (butanol) a
Phenol
Acetaldehyde (ethanal) a
Acetone (s) (propanone) c
Methyl ethyl ketone (s) (propan-2-one) c
Ethyl acetate (s) a
Butyl acetate (s) c
Amyl acetate (s)
Ethyl methacrylate
Acetic acid (ethanoic acid) b
Methyl chloride (s) a
Methylene chloride (s) (dichlormethane)
Ammonia a
acetonitrile a
Aniline
Pyridine
Group B
3)
Ethylene c
Buta-1,3-diene c
Acrylonitrile c
hydrogen cyanide a
Dethyl ether (s) c
Ethylene oxide (oxiran) c
1.4 Dioxan a
Tetrahydrofuran a
Tetrafluoroethylene a
Group C
3)
Hydrogen c
Acetylene (ethyne) c
Carbon bisulfide c
1)
Minimum Electrical Spark Gap
2)
Minimum Ignition Current
The ratio is based on the MIC value for laboratory methane
3)
Group allocation:
a according to MESG value
b according to MIC ratio
c according to both MESG value and MIC ratio
(s) solvent
Table XIV: Hazard classification of fluids according to their MESG1 and/or MIC2 values. (Extract from European Standard EN 50.014)
Legend
MESG
1)
MIC
2)
ratio
Group A
> 0.9 mm
> 0.8 mm
Group B
0.5 ... 0.9 mm
0.45 ... 0.8 mm
Group C
< 0.5 mm
< 0.45 mm
D00 E 19.06.2001 21:40 Uhr Seite 148
D00
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
= conditionally resistant
o = not resistant
Acetaldehyde o o o o x x
Acetic acid (crystalline), pure o o x
Acetic acid, industrial x x
Acetic acid vapors x x x
Acetic acid, 20 % x x x
Acetic acid, 50 % o x x x
Acetic acid, 80 % o o x
Acetic anhydride x x x
Aceto-acetic ester o x
Acetone o o x x
Acetophenone o x x
Acetylene x x x
Acrylnitrile o x
Air, clean x x x x x x
Air, oily x x x x x o
Ammonia liquid x x x x
Ammonia gas x x o x
Amyl acetate o o o x x
Amyl alcohol o o x x
Aniline o o x x x
Anthracene oil o o x x
ASTM oil No. 1 x x x x x o
ASTM oil No. 2 x x x x x o
ASTM oil No. 3 o x o
Benzaldehyde 100 % x x
Benzene o o x x o
Benzene bromide o o x o
Benzoic acid x x
Bitumen x
Blast furnace gas x x x x
Boron trifluoride x x x
Bromine o o x x
Butadiene x o
Butane x x x o
Butyl acetate o o o x
Butyl alcohol x o x x
Butyl glykol x x x
Butyraldehyde o o x x
Carbolineum o o x x x o
Carbon bisulfide o o x x o
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
o = not resistant
Carbon dioxide, dry x x x x x
Carbon dioxide, wet x x x x
Carbon tetrachloride o o o x x o
Chloracetic acid o o o x x
Chlorinated solvents o x
Clorine, dry x x
Chlorine water x o x x
Chlorine, wet o o x x x
Chlorobromomethane o x
Chlorobenzene o o o x o
Chloroform o o x x o
Chloromethyl o x x o
Citrus oils o o x
Coke furnace gas o o x o
Copra oil acid o x
Cottonseed oil x x x
Cresol x x o
Crude petroleum x x o
Cyclohexane x x x o
Cyclohexanone o o o x o
Cyclohexylamine x o
Decahydronaphtalene x x o
Desmodur T o o x o
Desmophene 2000 x x
Dibutylphthalate o o x o x x
Dichlorethylene x
Dichlorethane x x x o
Diethylamine o x x
Diethylene glycol x x x x
Diethyl ether o o x o
Diethyl ether o o x
Diethyl sebazate o o x
Dichlorbenzene o o x o
Dichlorbutylene x x o
Diesel oil o x x o
Di-isopropyl ketone o x x
Dimethyl ether o x x
Dimethylaniline o o x
Dimethyl formamide o o o x
Dioctylphthalate o o x x
Table XV: Chemical resistance of commonly used elastomer gaskets and sealing
materials
materials
D00 E 19.06.2001 21:40 Uhr Seite 149
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
o = not resistant
Dioxan o o x
Diphenyl o o x x o
Diphenyloxyd o x x
Edenol 888 x x
Essential oils o o x o
Ethyl acetate (acetic ether) o o x x
Ethane x o x x o
Ethyl acetate o o o o x o
Ethyl acrylate o x
Ethyl alcohol, denatured x
Ethyl alcohol, pure x x
Ethyl cloride x o x
Ethylene bromide o o x
Ethylene chloride x
Ethylene dichloride o o x
Ethylene glycol x x o x x x
Ethyl ether o o o x
Ethyl silicate x x x
Ethyl acrylate x
Fatty acids x
Fatty alcohol x x x x
Fir leaf oil x o x
Fluorbenzene x o x o
Hydrofluoric acid, cold, 5 % x x x
Hydrofluoric acid, cold, pure x
Formaldehyde x x x x
Formalin, 55 % x x x x
Formic acid x
Formic acid methyl ester o x
Freon 11 x x o x
Freon 12 x x x
Freon 22 o x o x x
Freon 113 x x x x
Furane o o x o
Furfurol o o o o x
Gas oill x x x o
Generator gas x x x
Glycerine x x x x x x
Glycol x x x x x x
Halowax oil o o x
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
o = not resistant
Heating fuel oil (coal base) o o x x
Heating fuel oil (petroleum crude base) x x x o
Heptane x o x x o
Hexaldehyde o o x x
Hexane x o x x o
Hydraulic fluids
Hydraulic oils DIN 51524 x x x o
Phosphoric ester HFD o o o x
Polyglycol water HFC x x x x x
Hydrobromic acid o o x x
Hydrobromic crystalline acid x x
Hydrocyanic acid x x
Hydrogen bromide x
Hydrogen gas 20 x x x
Hydrogen sulfide x x x x
Isobutyl alcohol x x x
Isopropyl acetate o o x
Isopropyl alcohol x x x
Isopropyl chloride o o x o
Isopropyl ether o x
Kerosene x x
Kerosine x x x o
Lighting gas x x o
Maleic anhydride x
Mercury x x x x
Methane x x
Methane (pit gas) x x x x
Methylene chloride o o o x o
Methyl acrylate o x o
Methyl alcohol (methanol) x o x x
Methyl ethyl ketone o o o x
Methyl isobutyl ketone o o x
Methyl methacrylate o o o x o
Methyl salicylate o o x
Naphtalene o o x o
Natural gas x x x
Nitrobenzene o o o x o
Nitrous oxide x x x
Oleic acid x x x o
Orange oil o o x x
materials
materials
D00 E 19.06.2001 21:40 Uhr Seite 150
D00
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
o = not resistant
Oxygen x x x x
Ozone o x x x x
Palmitic acid x x x o
Palm oil acid o x x
Paraffin x x x x o
Paraffin oil x x x o
Pentachlordiphenyl o x x o
Pentane x x o x x o
Perchloroethylene o o x x o
Petrol x o x o
Petrol alcohol 3:1 o x x
Petrol benzene 4:1 x o o x x o
Petrol benzene 7:3 o o o x x o
Petrol benzene 3:7 o o o x x
Petrol benzene spirit 5:3:2 o o o x o
Phenol o o x x x o
Phenyl ethyl ether o o x o
Phenylic acid (phenol) o o x
Phosphorous chloride o x x
Phthalic anhydride x x x x
Piperidine o o x
Polyglycol x x
Propane, gas x x x x x
Propylene oxyde o x
Propyl alcohol x x x x
Pydraul F-9 o o x x x
Pydraul AC o x x x
Pydraul A 150 o x x
Pydraul A 200 o x x x
Pyridine o x
Salicylic acid x x x x
Skydrol 500 x o x x
Skydrol 7000 x x x
Stearic acid x x
Styrene o o x x o
Sulfur x x x x x
Sulfur dioxide o o x x x
Sulfur trioxide, dry o x x
N
i
t
r
i
l
e
-
b
u
t
a
d
i
e
n
e

r
u
b
b
e
r

(
N
B
R
)

P
e
r
b
u
n
a
n
C
h
l
o
r
o
p
r
e
n
e

r
u
b
b
e
r

(
C
R
)

N
e
o
p
r
e
n
e
S
i
l
i
c
o
n
e

r
u
b
b
e
r
F
l
u
o
r
o

r
u
b
b
e
r

(
F
P
M
,

F
K
M
)

V
i
t
o
n
T
e
f
l
o
n

(
P
T
F
E
)
E
P
D
M
Medium
x = resistant
o = not resistant
Tar oil o x o
Tetrachlorethylene x x o
Tetrahydrofurane o o x o
Tetraline o o o x o
Toluene o o o x x o
Transformer oil x x x x o
Train oil o x x
Triethanolamine o o x
Tributoxyethyl phosphate o o x o
Tributyl phosphate o o o x o
Trichloroethane o o x x
Trichloroethylene o o x x o
Trichloroethyl phosphate 20 x x
Trichloroethyl phosphate 80 o x x
Trichloracetic acid 60 x x
Tricresyl phosphate o x x x
Turpentine x x o
Turpentine oil, pure x x x o
Vinyl acetate o o x
Vinylaceto-acetic acid 3:2 o o o x
Vinyl chloride, liquid x x x
Water 50 x x x x x
Water 100 x x x x
Wood oil x
Xylamon o o x x o
Xylene o o o x x o
materials
materials
D00 E 19.06.2001 21:40 Uhr Seite 151
Vacuum symbols
All symbols with the exception of those
marked with *
)
do not depend on the posi-
tion.
*
)
These symbols may only be used in the
position shown here (tip of the angle
pointing down)
The symbols for vacuum pumps should
always be arranged such that the side with
the constriction is allocated to the higher
pressure
Vacuum pumps
Piston vacuum pump
Diaphragm vacuum pump
Vacuum pump, general
Rotary positive
displacement pump *
)
Rotary plunger
vacuum pump *
)
Sliding vane rotary
vacuum pump *
)
Rotary piston vacuum pump *
)
Liquid ring vacuum pump *
)
Roots vacuum pump *
)
Turbine vacuum pump, general
Radial flow vacuum pump
Axial flow vacuum pump
Turbomolecular pump
Table XVI: Symbols used in vacuum technology (extract from DIN 28401)
Diffusion pump *
)
Adsorption pump *
)
Ejector vacuum pump *
)
Getter pump
Sputter-ion pump
Cryopump
Scroll pump *
)
Evaporation pump
Condensate trap, general
Condensate trap with heat
exchanger (e.g. cooled)
Gas filter, general
Accessories
Baffle, general
Cooled baffle
Filtering apparatus, general
Cold trap, general
Cold trap with coolant reservoir
Sorption trap
Throttling
Vacuum chamber
Vacuum bell jar
Vacuum chambers
Shut-off device, general
Shut-off devices
Shut-off valve, straight-
through valve
Right-angle valve
Stop cock
Three-way stop cock
Gate valve
Butterfly valve
Right-angle stop cock
Nonreturn valve
Safety shut-off valve
Manual operation
Variable leak valve
Electromagnetic operation
Modes of operation
D00 E 19.06.2001 21:40 Uhr Seite 152
D00
Table XVI: Symbols used in vacuum technology (extract from DIN 28401)
Hydraulic or pneumatic
operation
Electric motor drive
Weight-operated
Flange connection, general
Bolted flange connection
Small flange connection
Connections and
piping
Clamped flange connection
Threaded tube connection
Ball-and-socket joint
Spigot-and-socket joint
Taper ground joint connection
Intersection of two lines with
connection
Intersection of two lines
without connection
Branch-off point
Combination of ducts
Flexible connection
(e.g. bellows, flexible tubing)
Linear-motion leadthrough,
flange-mounted
Linear-motion leadthrough,
without flange
Leadthrough for transmission
of rotary and linear motion
Rotary transmission
leadthrough
Electric current leadthrough
General symbol for
vacuum *
)
Vacuum measurement, vacu-
um gauge head *
Vacuum gauge, operating and
display unit for vacuum gauge
head *
)
Vacuum gauge, recording *
)
Vacuum gauge with analog
measured-value display *
)
Vacuum gauge with digital
measured-value display *
)
Measurement of throughput
Measurement and
gauges
D00 E 19.06.2001 21:40 Uhr Seite 153
Kelvin Celsius Raumur Fahrenheit Rankine
Boiling point H
2
O 373 100 80 212 672
Body temperature 37 C 310 37 30 99 559
Room temperature 293 20 16 68 527
Freezing point H
2
O 273 0 0 32 492
NaCl/H
2
O 50:50 255 18 14 0 460
Freezing point Hg 34 39 31 39 422
CO
2
(dry ice) 195 78 63 109 352
Boiling point LN
2
77 196 157 321 170
Absolute zero point 0 273 219 460 0
Table XVII: Temperature comparison and conversion table (rounded off to whole degrees)
K
Kelvin
C
Celsius
C
Raumur
F
Fahrenheit
R
Rankine
K C R F R
Kelvin Celsius Raumur Fahrenheit Rankine
1 K 273 (K 273) (K 273) + 32 K = 1,8 K
C + 273 1 C C + 32 (C + 273)
R + 273 R 1 R + 32 (R + 273)
(F 32) + 273 (F 32) (F 32) 1 F + 460
(R) (R 273) (R 273) R 460 1
4
5
9
5
9
5
9
5
9
5
4
5
5
4
5
4
5
9
5
9
5
9
5
9
4
9
4
5
5
9
9
4
5
9
5
4
[ ]
[ ]
Conversion in
Fig. 9.1: Variation of mean free path (cm) with pressure for various gases
Pressure p [mbar]
M
e
a
n

f
r
e
e

p
a
t
h

[
c
m
]
Fig. 9.2: Diagram of kinetics of gases for air at 20 C
: mean free path in cm ( ~ 1/p)
n : particle number density in cm
3
(n ~ p)
Z
A
: area-related impingement rate in cm
3
s
1
(Z
A
~ p
2
)
Z
V
: volume-related collision rate in cm
3
s
1
(Z
V
~ p
2
)
1
p
Pressure p [mbar]
Z A

p
2 Z
V

p
2
D00 E 19.06.2001 21:41 Uhr Seite 154
D00
Fig. 9.3: Decrease in air pressure (1) and change in temperature (2) as a function of
altitude
P
r
e
s
s
u
r
e

[
m
b
a
r
]
Altitude [km]
T
e
m
p
e
r
a
t
u
r
e

(
K
)
Fig. 9.5: Conductance values for piping of commonly used nominal width with circu-
lar cross-section for laminar flow (p = 1 mbar) according to equation 53a.
(Thick lines refer to preferred DN) Flow medium: air (d, l in cm!)
Pipe length l [cm]
10
6
10
5
10
4
10
3
10
2
10
1
10
0
8
6
4
2
10
1
10
2
10
3
10
4 2 4 6 8
c
o
n
d
u
c
t
a
n
c
e

[
l

c
m
1
]
Fig. 9.4: Change in gas composition of the atmosphere as a function of altitude
Molecules/atoms [cm
3
]
A
l
t
i
t
u
d
e

(
k
m
)
Fig. 9.6: Conductance values for piping of commonly used nominal width with cir-
cular cross-section for molecular flow according to equation 53b. (Thick
lines refer to preferred DN) Flow medium: air (d, l in cm!)
Pipe length l [cm]
c
o
n
d
u
c
t
a
n
c
e

[
l

1
]
D00 E 19.06.2001 21:41 Uhr Seite 155
Fig. 9.7: Nomogram for determination of pump-down time tp of a vessel in the rough vacuum pressure range
p
END
p
end, p
mbar
p
START
< 1013 mbar
p
START
= 1013 mbar
p
START
p
end, P
p
END
p
end, P
R =
Column ': Vessel volume V in liters
Column ^: Maximum effective pumping speed S
eff,max
at the vessel in (left) liters per second or
(right) cubic meters per hour.
Column : Pump-down time t
p
in (top right) seconds
or (center left) minutes or (bottom right)
hours.
Column : Right:
Pressure p
END
in millibar at the END of the
pump-down time if the atmospheric
pressure p
START
( p
n
= 1013 prevailed at the
START of the pump-down time. The desired
pressure p
END
is to be reduced by the
ultimate pressure of the pump p
ult,p
and the
differential value is to be used in the co-
lumns. If there is inflow q
pV,in
, the value
p
end
p
ult,p
q
pV,in
/ S
eff,max
is to be used
in the columns.
Left:
Pressure reduction ratio R = (p
START
p
ult,p
q
pV,in
/ S
eff,max
)/(p
end
p
ult,p
q
pV,in
/
S
eff,max
), if the pressure p
START
prevails at
the beginning of the pumping operation
and the pressure is to be lowered to p
END
by pumping down.
The pressure dependence of the pumping
speed is taken into account in the nomo-
gram and is expressed in column I by
p
ult,p
. If the pump pressure p
ult,p
is small in
relation to the pressure p
end
which is
desired at the end of the pump-down
operation, this corresponds to a constant
pumping speed S or S
eff
during the entire
pumping process.
Example 1 with regard to nomogram 9.7:
A vessel with the volume V = 2000 l is to be pumped
down from a pressure of p
START
= 1000 mbar
(atmospheric pressure) to a pressure of p
END
= 10
-2
mbar by means of a rotary plunger pump with an
effective pumping speed at the vessel of
S
eff,max
= 60 m
3
/h = 16.7 l s
-1
. The pump-down time
can be obtained from the nomogram in two steps:
1) Determination of : A straight line is drawn
through V = 2000 l (column ') and S
eff
= 60 m
3
/h
-1
= 16.7 l s
-1
(column ^) and the value t = 120 s
= 2 min is read off at the intersection of these straight
lines with column (note that the uncertainty of this
procedure is around = 10 s so that the relative
uncertainty is about 10 %).
2) Determination of t
p
: The ultimate pressure of the
rotary pump is p
ult,p
= 3 10
-2
mbar, the apparatus is
clean and leakage negligible (set q
pV,in
= 0); this is
p
START
p
ult,p
= 10
-1
mbar 3 10
-2
mbar = 7 10
-2
mbar. Now a straight line is drawn through the point
found under 1) t = 120 s (column ) and the point
p
END
p
ult,p
= 7 10
-2
mbar (column I) and the
intersection of these straight lines with column
t
p
= 1100 s = 18.5 min is read off. (Again the relative
uncertainty of the procedure is around 10 % so that
the relative uncertainty of tp is about 15 %.) Taking
into account an additional safety factor of 20 %,
one can assume a pump-down time of
t
p
= 18.5 min (1 + 15 % + 20 %)
= 18.5 min 1.35 = 25 min.
Example 2 with regard to nomogram 9.7:
A clean and dry vacuum system (q
pV,in
= 0) with
V = 2000 l (as in example 1) is to be pumped down
to a pressure of p
END
= 10
-2
mbar. Since this pres-
sure is smaller than the ultimate pressure of the
rotary piston pump (S
eff,max
= 60 m
3
/h = 16.7 l
( s
-1
= 3 10
-2
mbar), a Roots pump must be used
in connection with a rotary piston pump. The former
has a starting pressure of p
1
= 20 mbar, a pumping
speed of S
eff,max
= 200 m
3
/h 55 l s
-1
as well as
p
ult,p
4 10
-3
mbar. From p
start
= 1000 mbar to p =
20 mbar one works with the rotary piston pump and
then connects the Roots pump from p
1
= 20 mbar to
p
END
= 10
-2
mbar, where the rotary piston pump acts
as a backing pump. For the first pumping step one
obtains the time constant = 120 s = 2 min from the
nomogram as in example 1 (straight line through
V = 2000 l, S
eff
= 16.7 l s
-1
). If this point in
column is connected with the point
p
1
- p
ult,p
= 20 mbar 3 10
-2
mbar = 20 mbar (p
ult,p
is ignored here, i.e. the rotary piston pump has a
constant pumping speed over the entire range from
1000 mbar to 20 mbar) in column 5, one obtains t
p,1
= 7.7 min. The Roots pump must reduce the
pressure from p
1
= 20 mbar to p
END
= 10
-2
mbar, i.e.
the pressure reduction ratio R = (20 mbar 4 10
-3
mbar) / (10
-2
mbar-4 10
-3
) = 20/6 10
-3
mbar =
3300.
The time constant is obtained (straight line V = 2000
l in column ', S
eff
= 55 l s
1
in column ^) at = 37
s (in column ). If this point in column is
connected to R = 3300 in column I, then one
obtains in column t
p, 2
= 290 s = 4.8 min. If one
takes into account t
u
= 1 minfor the changeover
time, this results in a pump-down time of
t
p
= t
p1
+ t
u
+ t
p2
= 7.7 min + 1 min + 4.8 min = 13.5 min.
D00 E 19.06.2001 21:41 Uhr Seite 156
D00
Fig. 9.8: Nomogram for determination of the conductance of tubes with a circular cross-section for air at 20 C in the region of molecular flow (according to J. DELAFOSSE and
G. MONGODIN: Les calculs de la Technique du Vide, special issue Le Vide, 1961)
T
u
b
e

l
e
n
g
t
h
C
o
r
r
e
c
t
i
o
n

f
a
c
t
o
r

f
o
r

s
h
o
r
t

t
u
b
e
s
C
o
n
d
u
c
t
a
n
c
e

f
o
r

m
o
l
e
c
u
l
a
r

f
l
o
w
T
u
b
e

d
i
a
m
e
t
e
r
C
Example: What diameter d must a 1.5-m-long pipe have so
that it has a conductance of about C = 1000 l / sec in the
region of molecular flow? The points l = 1.5 m and C = 1000
l/sec are joined by a straight line which is extended to inters-
ect the scale for the diameter d. The value d = 24 cm is obtai-
ned. The input conductance of the tube, which depends on
the ratio d / l and must not be neglected in the case of short
tubes, is taken into account by means of a correction factor
. For d / l < 0.1, can be set equal to 1. In our example
d/l = 0.16 and = 0.83 (intersection point of the straight line
with the a scale). Hence, the effective conductance of the
pipeline is reduced to C = 1000 0.83 = 830 l/sec. If d
is increased to 25 cm, one obtains a conductance of
1200 0.82 = 985 l / sec (dashed straight line).
D00 E 19.06.2001 21:41 Uhr Seite 157
Fig. 9.9: Nomogram for determination of conductance of tubes (air, 20 C) in the entire pressure range
T
u
b
e

l
e
n
g
t
h

[
m
e
t
e
r
s
]
P
r
e
s
s
u
r
e

[
m
b
a
r
]
M
o
l
.

f
l
o
w
K
n
u
d
s
e
n

f
l
o
w
L
a
m
i
n
a
r

f
l
o
w
C
l
a
u
s
i
n
g

f
a
c
t
o
r
U
n
c
o
r
r
e
c
t
e
d

c
o
n
d
u
c
t
a
n
c
e

f
o
r

m
o
l
.

f
l
o
w

[
m
3
1
]
T
u
b
e

i
n
t
e
r
n
a
l

d
i
a
m
e
t
e
r

[
c
m
]
Air 20 C
C
o
r
r
e
c
t
i
o
n

f
a
c
t
o
r

f
o
r

h
i
g
h
e
r

p
r
e
s
s
u
r
e
s
C
P
r
e
s
s
u
r
e

[
m
b
a
r
]
c
o
n
d
u
c
t
a
n
c
e

[

l

1
]
Procedure: For a given length (l) and internal dia-
meter (d), the conductance C
m
, which is indepen-
dent of pressure, must be determined in the mole-
cular flow region. To find the conductance C* in
the laminar flow or Knudsen flow region with a
given mean pressure of p in the tube, the conduc-
tance value previously calculated for Cm has to be
multiplied by the correction factor a determined in
the nomogram: C* = C
m
.
Example: A tube with a length of 1 m and an internal
diameter of 5 cm has an (uncorrected) conductance
C of around 17 l/s in the molecular flow region, as
determined using the appropriate connecting lines
between the l scale and the d scale. The conduc-
tance C found in this manner must be multiplied
by the clausing factor = 0.963 (intersection of
connecting line with the scale) to obtain the true
conductance C
m
in the molecular flow region:
C
m
= 17 0.963 = 16.37 l/s.
In a tube with a length of 1 m and an internal dia-
meter of 5 cm a molecular flow prevails if the mean
pressure p in the tube is < 2.7 10
-3
mbar.
To determine the conductance C* at higher
pressures than 2.7 10
-3
mbar, at 8 10
-2
mbar
(= 6 10
-2
torr), for example, the corresponding
point on the p scale is connected with the point
d = 5 cm on the d scale. This connecting line
intersects the a scale at the point = 5.5.
The conductance C* at p = 8 10
-2
mbar is:
C* = C
m
= 16.37 5.5 = 90 l/s.
D00 E 19.06.2001 21:41 Uhr Seite 158
D00
Fig. 9.10: Determination of pump-down time in the medium vacuum range taking into account the outgasing from the walls
P
u
m
p
-
d
o
w
n

t
i
m
e

t

[
m
i
n
]
V
o
l
u
m
e

V

[
m
i
n
3
]
[
m
3
1
]

P
u
m
p
i
n
g

s
p
e
e
d
G
a
s

e
v
o
l
u
t
i
o
n

[
m
b
a
r

2
]
Exam
ple 2
E
xam
p
le 1
weak
normal
strong
Area F [m
2
]
The nomogram indicates the relationship between
the nominal pumping speed of the pump, the cham-
ber volume, size and nature of the inner surface as
well as the time required to reduce the pressure
from 10 mbar to 10
-3
mbar.
Example 1: A given chamber has a volume of 70 m
3
and an inner surface area of 100 m
2
; a substantial
gas evolution of 2 10
-3
mbar l s
-1
m
-2
is assu-
med. The first question is to decide whether a pump
with a nominal pumping speed of 1300 m
3
/h is
generally suitable in this case. The coordinates for
the surface area concerned of 100 m
2
and a gas evo-
lution of 2 10
-3
mbar l s
-1
m
-2
result in an inters-
ection point A, which is joined to point B by an
upward sloping line and then connected via a vertical
line to the curve that is based on the pumping speed
of the pump of 1300 m
3
/h (D). If the projection to the
curve is within the marked curve area (F), the pum-
ping speed of the pump is adequate for gas evoluti-
on. The relevant pump-down time (reduction of pres-
sure from 10 mbar to 10
-3
mbar) is then given as
30 min on the basis of the line connecting the point
1300 m
3
/h on the pumping speed scale to the point
70 m
3
(C) on the volume scale: the extension results
in the intersection point at 30 min (E) on the time
scale.
In example 2 one has to determine what
pumping speed the pump must have if the vessel
(volume = approx. 3 m
3
) with a surface area
of 16 m
2
and a low gas evolution of
8 10
-5
mbar l s
-1
m
-2
is to be evacuated from
10 mbar to 10
-3
mbar within a time of 10 min. The
nomogram shows that in this case a pump with a
3
/h is appropriate.
D00 E 19.06.2001 21:41 Uhr Seite 159
Fig. 9.11: Saturation vapor pressure of various substances
Halocarbon 11
Halocarbon 12
Halocarbon 13
Halocarbon 22
Trichorethylene
Acetone
Temperature [C]
V
a
p
o
r

p
r
e
s
s
u
r
e

[
m
b
a
r
]
Temperature (C)
1,00E-12
1,00E-11
1,00E-10
1,00E-09
1,00E-08
1,00E-07
1,00E-06
1,00E-05
1,00E-04
1,00E-03
1,00E-02
1,00E-01
1,00E+00
1,00E+01
1,00E+02
1,00E+03 1000
100
10
1
10
-1
10
-2
10
-3
10
-4
10
-5
10
-6
10
-7
10
-8
10
-9
10
-10
10
-11
10
-12
0 25 50 75 100 150 200 250
Mercury
Aziepon 201
Santovac 5
(similar to Ultralen)
Diffelen
normal
Diffelen ultr a
DC 705
DC 704
Diffelen
light
V
a
p
o
r

p
r
e
s
s
u
r
e

(
m
b
a
r
)
Fig. 9.12: Saturation vapor pressure of pump fluids for oil and mercury fluid
entrainment pumps
Fig. 9.13: Saturation vapor pressure of major metals used in vacuum technology
T
e
m
p
e
r
a
t
u
r
e

[
K
]
Vapor pressure
T
e
m
p
e
r
a
t
u
r
e

[
C
]
D00 E 19.06.2001 21:41 Uhr Seite 160
D00
Fig. 9.14: Vapor pressure of nonmetallic sealing materials (the vapor pressure curve
for fluoro rubber lies between the curves for silicone rubber and Teflon)
1 NBR (Perbunan) 2 Silicone rubber 3 Teflon
V
a
p
o
r

p
r
e
s
s
u
r
e

[
m
b
a
r
]
Temperature [C]
Fig. 9.15: Saturation vapor pressure ps of various substances relevant for
cryogenic technology in a temperaturerange of T = 2 80 K
critical point
b melting point
Rough vacuum
1 to approx. 10
3
mbar
10
2
to approx. 10
5
Pa
Medium vacuum
10
3
to 1 mbar
10
1
to 10
2
Pa
High vacuum
10
7
to 10
3
mbar
10
5
to 10
1
Pa
Ultrahigh vacuum
<10
7
mbar
<10
5
Pa
Fig. 9.16: Common working ranges of vacuum pumps
10
13
10
12
10
11
10
10
10
9
10
8
10
7
10
6
10
5
10
4
10
3
10
2
10
1
10
0
10
1
10
2
Piston vacuum pump
Turbine vacuum pump
Liquid jet vacuum pump
Turbomolecular pump
Diffusion pump
Submilation pump
Sputter-ion pump
Cryopump
10
3
10
14
p in mbar
Working range for special model or special operating data Normal working range
Liquid-ring vacuum pump
Multiple-vane rotary vacuum pump
Trochoide vacuum pump
Diaphragm vacuum pump
Sliding-vane rotary vacuum pump
Roots vacuum pump
Rotary plunger vacuum pump
Gaseous-ring vacuum pump
Vapor jet vacuum pump
Diffusion ejector pump
Adsorption pump
D00 E 19.06.2001 21:41 Uhr Seite 161
Ultrahigh vacuum
<10
7
mbar
<10
5
Pa
High vacuum
10
7
to 10
3
mbar
10
5
to 10
1
Pa
Medium vacuum
10
3
to 1 mbar
10
1
to 10
2
Pa
Rough vacuum
1 to approx. 10
3
mbar
10
2
to approx. 10
5
Pa
Fig. 9.16a: Measurement ranges of common vacuum gauges
10
13
10
12
10
11
10
10
10
9
10
8
10
7
10
6
10
5
10
4
10
3
10
2
10
1
10
0
10
1
10
2
Bourdon vacuum gauge
Liquid level vacuum gauge
(McLeod vacuum gauge)
Decrement vacuum gauge
Pirani vacuum gauge
Thermal conductivity vacuum gauge
Compression vacuum gauge
Cold-cathode ionization vacuum gauge
Penning ionization vacuum gauge
Hot-cathode ionization vacuum gauge
10
3
10
14
p in mbar
Working range for special models or special operating data
Thermocouple vacuum gauge
Bimetallic vacuum gauge
Thermistor vacuum gauge
Triode ionization vacuum gauge for medium vacuum
Triode ionization vacuum gauge for high vacuum
Bayard-Alpert ionization vacuum gauge
Bayard-Alpert ionization vacuum gauge with modulator
Extractor vacuum gauge
Partial pressure vacuum gauge
The customary limits are indicated in the diagram.
Spring element vacuum gauge
Capacitance diaphragm vacuum gauge
Diaphragm vacuum gauge
Pressure balance
Piezoelectric vacuum gauge
U-tube vacuum gauge
Magnetron gauge
D00 E 19.06.2001 21:41 Uhr Seite 162
D00
Fig. 9.17: Specific volume V
sp
of saturated water vapor in m
3
/kg within a range of 0.013 to 133 mbar
Saturated vapor
Fig. 9.18: Breakdown voltage U between parallel electrodes in a homogeneous electrical field as a function of gas
pressure p distance between electrodes d (in mm) (Paschen curve), for air
D00 E 19.06.2001 21:41 Uhr Seite 163
Fig. 9.19: Phase diagram of water
liquid
Melting
Sublimation
W
a
t
e
r

v
a
p
o
r

p
r
e
s
s
u
r
e

[
m
b
a
r
]
GASEOUS
Temperature [C]
Triple point
(0.01C, 6.09 mbar)
Evaporation
SOLID
D00 E 19.06.2001 21:41 Uhr Seite 164
Statutory Units
10. The statutory
units used in
vacuum
technology
10.1 Introduction
Two federal German laws and the related
implementing provisions stipulate which
units must be used for measurements
today (generally since January 1, 1978) in
business and official documents and com-
munications. The provisions resulted in a
number of fundamental changes that also
have to be taken into account in vacuum
technology. Many of the units commonly
used in the past, such as torr, gauss, stan-
dard cubic meter, atmosphere, poise, kilo-
calorie, kilogram-force, etc., are no longer
permissible. Instead, other units are to be
used, some of which are new while others
were previously used in other fields. The
alphabetical list in Section 10.2 contains
the major variables relevant for vacuum
technology along with their symbols and
the units now to be used, including the SI
units (see below) and legally permissible
units derived from them. The list is follo-
wed by a number of remarks in Section
10.3. The purpose of the remarks is, on the
one hand, to establish a connection with
previous practice wherever this is neces-
sary and, on the other hand, to provide
explanations on practical use of the con-
tent of the alphabetical list.
The statutory units for measurements are
based on the seven basic SI units of the
Systme International (SI).
Statutory units are:
a) the basic SI units (Table 10.4.1)
b) units derived from the basic SI units, in
some cases with special names and unit
symbols (Tables 10.4.2 and 10.4.4)
c) units used in atomic physics (Table
10.4.3)
d) decimal multiples and decimal parts of
units, some with special names
Examples: 10
5
N ( m
-2
= 1 bar)
1 dm
3
= 1 l (liter)
10
3
kg = 1 t (ton)
Detailed descriptions are provided in publi-
cations by W. Haeder and E. Grtner (DIN),
by IUPAP 1987 and by S. German, P. Draht
(PTB). These should always be referred to
if the present summary tailored to vacuum
technology leaves any questions open.
10.2 Alphabetical list
1)
of variables,
symbols and units
frequently used in
vacuum technology
and its
applications (see
also DIN 28 402)
1)
The list is based on work done by Prof. Dr. I.
Lckert, for which we would like to express our
gratitude
D00
No. Variable Symbol SI- Preferred statutory No. of remark Notes
unit units in Section 10.3
1 Activity (of a radioactive substance) A s
1
(B
q
) s
1
3/1
2 (General gas constant) see no. 73
3 Work W J J, kJ, kWh, Ws
4 Atomic mass m
u
kg kg, mg see Table V in Sect. 9
5 Avogadro constant N
A
mol
1
mol
1
6 Acceleration a m s
2
m s
2
, cm s
2
7 Boltzmann constant k J K
1
j K
1
, mbar l K
1
see Table V in Sect. 9
8 Celsius temperature (theta) C 3/2
9 Vapor pressure p
v
N m
2
, Pa mbar, bar 3/3 Pa = Pascal
10 Time t s s, min, h see Table 10.4.4
11 Density (gas density) (ro) kg m
3
kg m
3
, g cm
3
3/6
12 Dielectric constant (epsilon) F m
1
F m
1
, As V
1
m
1
F = Farad
13 Diffusion coefficient D m
2
s
1
m
2
s
1
, cm
2
s
1
14 Moment of momentum L N s m N s m
15 Torque M N m N m, kN m
16 Rotational speed,
rotational frequency n, f s
1
s
1
, min
1
17 Pressure in fluids p N m
2
, Pa bar, mbar 3/3 Pa = Pascal
18 Pressure as mechanical stress p N m
2
, Pa N mm
2
3/4
19 Diameter d m cm, mm
20 Dynamic viscosity (eta) Pa s mPa s 3/5
21 Effective pressure p
e
N m
2
, Pa mbar 3/3 see also no. 126
22 Electric field strength E V m
1
V m
1
23 Electrical capacitance C F F, F, pF F = Farad
24 Electrical conductivity (sigma) S m
1
S m
1
D00 E 19.06.2001 21:41 Uhr Seite 165
Statutory Units
No. Variable Symbol SI unit Preferred statutory No. of remark Notes
units in Section 10.3
25 Electrical conductance G S S S = Siemens
26 Electrical voltage U V V, mV, kV
27 Electric current density S A m
2
a m
2
, A cm
2
28 Electric current intensity I A A, mA, A
29 Electrical resistance R (ohm) , k , M
30 Quantity of electricity (electric charge) Q C C, As C = Coulomb
31 Electron rest mass me kg kg, g see Table V in Sect. 9
32 Elementary charge e C C, As
33 Ultimate pressure p
ult
N m
2
, Pa mbar
34 Energy E J J, kJ, kWh, eV J = Joule
35 Energy dose D J k
1
3/5 a
36 Acceleration of free fall g m s
2
m s
2
37 Area A m
2
m
2
, cm
2
38 Area-related impingement rate Z
A
m
2
s
1
m
2
s
1
; cm
2
s
1
39 Frequency f Hz Hz, kHz, MHz
40 Gas permeability Q
perm
m
3
(NTP) cm
3
(NTP) 3/19 d =day (see Tab. 10.4.4

m
2
s Pa m
2
d bar see no. 73 and no. 103)
41 Gas constant R
42 Velocity v m s
1
m s
1
, mm s
1
, km h
1
43 Weight (mass) m kg kg, g, mg 3/6
44 Weight (force) G N N, kN 3/7
45 Height h m m, cm, mm
46 Lift s m cm see also no. 139
47 Ion dose J C kg
1
c kg
1
, C g
1
3/8
48 Pulse p
^
(b) N s N s
49 Inductance L H H, mH H = Henry
50 Isentropic exponent (kappa) = c
p
c
v
1
51 Isobaric molar heat capacity C
mp
J mol
1
K
1
J mol
1
K
1
52 Isobaric specific heat capacity c
p
J kg
1
K
1
J kg
1
K
1
53 Isochore molar heat capacity C
mv
J mol
1
K
1
54 Isochore specific heat capacity c
v
J kg
1
K
1
J kg
1
K
1
55 Kinematic viscosity (n) m
2
s
1
mm
2
s
1
, cm
2
s
1
3/9
56 Kinetic energy E
K
J J
57 Force F N N, kN, mN 3/10 N = Newton
58 Length l m m, cm, mm 3/11
59 Linear expansion coefficient (alpha) m m
; K
1
m K m K
60 Leak rate Q
L
N m s
1
mbar l ; cm
3
3/12
(NTP)
s s
61 Power P W W, kW, mW
62 Magnetic field strength H A m
1
A m
1
3/13
63 Magnetic flux density B T T 3/14 T = Tesla
64 Magnetic flux (phi) Wb, V s V s 3/15 Wb = Weber
65 Magnetic induction B T T see no. 63
66 Mass m kg kg, g, mg 3/6
67 Mass flow rate q
m
kg s
1
kg s
1
, kg h
1
, g s
1
68 Mass content w
i
kg kg
1
%,
o
/
oo
, ppm ppm = parts per million
69 Mass concentration
i
(ro-i) kg m
3
kg m
3
, g m
3
, g cm
3
70 Moment of inertia J kg m
2
kg m
2
71 Mean free path m m, cm
72 Molality b
i
mol kg
1
mol kg
1
73 Molar gas constant R J mbar l see Table V in Sect. 9

mol K mol K
74 Molar mass (quantity-related mass) M kg mol
1
kg kmol
1
, g mol
1
D00 E 19.06.2001 21:41 Uhr Seite 166
Statutory Units
75 Molar volume V
m
m
3
mol
1
m
3
mol
1
, l mol
1
76 Molar volume, standard V
mn
m
3
mol
1
m
3
mol
1
(NTP) see Table V in Sect. 9
l mol
1
(NTP)
77 Molecular mass m kg g
78 Normal stress (mech.) (sigma) N m
2
N mm
2
79 Standard density of a gas
n
(ro-en) kg m
3
kg m
3
, g cm
3
80 Standard pressure p
n
N m
2
, Pa mbar see Table V in Sect. 9
81 Standard volume V
n
m
3
m
3
(NTP), cm
3
(NTP) 3/16
82 Partial pressure p
i
N m
2
, Pa mbar 3/17
83 Period T s s, ms, s
84 Permeation coefficient P m
3
m cm
2
3/18

s m
2
bar s mbar
85 Planck constant h J s J s see Table V in Sect. 9
86 pV throughput q
pV
N m s
1
mbar l s
1
3/19
87 pV value pV N m mbar l 3/19
88 Radius (also molecular radius) r m cm, mm, m
89 Space charge density (ro) C m
3
C m
3
, As m
3
90 Solid angle (omega) sr sr sr = steradian
91 Relative atomic mass A
T
3/20 nondimensional variab.
92 Relative molecular mass M
r
3/21 nondimensional variab.
93 Relative particle mass M
r
nondimensional variab.
94 Residual vapor pressure p
rd
N m
2
, Pa mbar
95 Residual gas pressure p
rg
N m
2
, Pa mbar
96 Residual (total) pressure p
r
N m
2
, Pa mbar
97 Reynold number nondimensional Re nondimensional variab.
variable
98 Saturation vapor pressure p
s
N m
2
, Pa mbar
99 Throughput (of a pump) q
pV
, Q N m s
1
mbar l s
1
100 Pumping speed S m
3
s
1
m
3
h
1
, l s
1
see no. 132
101 Stress (mech.) , , N m
2
N m
2
, N mm
2
3/4 see no. 18
(ro,
sigma, tau)
102 Specific electron charge e m
e
1
C kg
1
C kg
1
, As kg
1
103 Specific gas constant R
i
J kg
1
K
1
mbar l 3/22
kg K
104 Specific ion charge e m
1
C kg
1
C kg
1
, As kg
1
105 Specific electrical resistance (ro) m cm, mm
2
m
1
106 Specific volume v m
3
kg
1
m
3
kg
1
; cm
3
g
1
107 Specific heat capacity c J kg
1
K
1
J kg
1
K
1
, J g
1
K
1
3/23
108 Stefan-Boltzmann constant s (sigma) W W see Table V in Sect. 9

m
2
K
4
m
2
K
4
109 Quantity of substance (n) mol mol, kmol
110 Throughput of substance q
v
mol s
1
mol s
1
111 Concentration of substance c
i
mol m
3
mol m
3
, mol l
1
for substance i
112 Collision rate Z s
1
s
1
113 Conductance C, German: L m
3
s
1
m
3
s
1
, l s
1
114 Flow resistance R s m
3
s m
3
, s l
1
115 Number of particles N nondimensional variab.
116 Particle number density (volume-related) n m
3
cm
3
117 Particle number density (time-related) q
N
s
1
s
1
see no. 120
118 Particle throughput density j
N
m
2
s
1
m
2
s
1
, cm
2
s
1
see no. 121
119 Particle mass m kg kg, g
120 Particle flux q
N
s
1
s
1
see no. 117
D00
D00 E 19.06.2001 21:41 Uhr Seite 167
Statutory Units
121 Particle flux density j
N
m
2
s
1
m
2
s
1
, cm
2
s
1
see no. 118
122 Thermodyn. temperature T K K, mK
123 Temperature difference T, K K, C 3/24
124 Temperature conductivity a m
2
s
1
a =
1
c
p
125 Total pressure p
t
N m
2
, Pa mbar 3/3
126 Overpressure p
e
N m
2
, Pa mbar 3/3
127 Ambient pressure P
amb
N m
2
, Pa mbar, bar 3/3
128 Speed of light in vacuum c m s
1
m s
1
, km s
1
129 Evaporation heat L
d
J kJ
130 Viscosity, dynamic (eta) Pa s mPa s see no. 20
131 Volume V m
3
m
3
, l, cm
3
132 Volume throughput (volumetric flow) q
v
m
3
s
1
m
3
h
1
, l s
1
133 Volume concentration
i
m
3
m
3
l l
1
, %,
o
/
oo
, ppm ppm = parts per million
(sigma-i)
134 Volume-related collision rate Z
v
s
1
m
3
s
1
m
3
, s
1
cm
3
135 Quantity of heat Q J J, kJ, kWh, Ws 3/25
136 Heat capacity C J K
1
J K
1
, kJ K
1
137 Thermal conductivity W W

(lambda) K m K m
138 Heat transfer coefficient (alpha) W
K m
2
139 Path length s m m, cm
140 Wave length (lambda) m nm 3/11
141 Angle (plane) , , rad rad rad, , , 3/26 rad = radian
(alpha,
beta, gamma)
142 Angular acceleration (alpha) rad s
2
rad s
2
143 Angular velocity (omega) rad s
1
rad s
1
144 Efficiency (eta) nondimensional variab.
145 Time t s s, min, h, nn, mn see Table 10.44
146 Period of time t, t s s, min, h see Table 10.44
10.3 Remarks on
alphabetical list
in Section 10.2
3/1: Activity
The unit previously used was curie (Ci).
1 Ci = 3.7 10
10
s
-1
= 37 ns
-1
3/2: (C) Celsius temperature
The term degrees Celsius is a special name
for the SI unit kelvin (K) [see no. 122] for
indicating Celsius temperatures. The term
degrees Celsius is legally approved.
3/3: Pressure
The revised version of DIN 1314 must
be complied with. The specifications of
this standard primarily apply to fluids
(liquids, gases, vapors). In DIN 1314,
bar (1 bar = 0.1 MPA = 10
5
Pa) is stated
in addition to the (derived) SI unit,
1 Pa = 1 N m
-2
, as a special name for one
tenth of a megapascal (Mpa). This is in
accordance with ISO/1000 (11/92), p. 7.
Accordingly the millibar (mbar), a very
useful unit for vacuum technology, is also
permissible: 1 mbar = 102 Pa = 0.75 torr.
The unit torr is no longer permissible.
Special note
Exclusively absolute pressures are measu-
red and used for calculations in vacuum
technology.
In applications involving high pressures,
frequently pressures are used that are
based on the respective atmospheric pres-
sure (ambient pressure) pamb. According
to DIN 1314, the difference between a
pressure p and the respective atmospheric
pressure (ambient pressure) pamb is desi-
gnated as overpressure p
e
: p
e
= p p
amb
.
The overpressure can have positive or
negative values.
Conversions
1 kg cm
-2
= 980.665 mbar = 981 mbar
1 at (technical atmosphere) =
980.665 mbar = 981 mbar
1 atm (physical atmosphere) =
1013.25 mbar = 1013 mbar
1 atmosphere above atmospheric pressure
(atmospheric overpressure) =
2026.50 mbar = 2 bar
1 atm
1 torr = 1 mm Hg = = 1
760
33.322 Pa= 1.333 mbar
1 meter head of water = 9806.65 Pa =
98 mbar
D00 E 19.06.2001 21:41 Uhr Seite 168
Statutory Units
1 mm Hg = 133.332 Pa = 1.333 mbar =
4/3 mbar
The pressure as mechanical stress
(strength) is generally given in pascal (Pa)
and in N nm
2
.
Conversions:
1 Pa = 1 N m
2
= 10
6
N mm
2
1 kg cm
2
= 98,100 Pa = 0.981 N mm
2
= 0,1 N mm
2
1 kg mm
2
= 9,810,000 Pa =
9.81 N mm
2
= 10 N mm
2
3/5: Dynamic viscosity
The unit previously used was poise (P).
1 P = 0.1 Pa s = 1 g cm
1
s
1
3/5a: Energy dose
Rad (rd) is no longer permissible.
1
1 rd = J kg
1
100
3/6: Weight
DIN 1305 is to be complied with in this
context. Because of its previous ambiva-
lence, the word weight should only be
used to designate a variable of the nature
of a mass as a weighing result for indica-
ting quantities of goods.
The designations specific weight and
specific gravity should no longer be
used. Instead, one should say density.
3/7: Weight force
See DIN 1305. The previous units pond (p)
and kilopond, i.e. kilogram-force, (kp) as
well as other decimal multiples of p are no
longer used.
1 kp = 9.81 N
3/8: Ion dose
The previously used unit was the Rntgen
(R). 1 R = 2,58 10
4
C kg
1
3/9: Kinematic viscosity
The previously used unit was stokes (St).
1 St = 1 cm
2
s
1
; 1 cSt = 1 mm
2
s
1
3/10: Force
The dyne, the CGS unit for force, is no lon-
ger used.
1 dyne = 10
-5
N
3/11: Length/wavelength
The unit ngstrm () (e.g. for wave-
length) will no longer be used in future
(see Table 4.6).
1 = 10
-8
cm = 0.1 nm
3/12: Leak rate
In DIN 40.046 sheet 102 (draft of August
1973 issue), the unit mbar dm
3
s
-1
(= mbar l s
-1
) is used for the leak rate.
Note that the leak rate corresponding
to the unit 1 mbar l s
-1
at 20 C is
practically the same as the leak rate
1 cm
3
s
-1
(NTP). (See also 3/17)
3/13: Magnetic field strength
The previously used unit was the oersted
(Oe).
1 Oe = 79.577 A m
-1
3/14: Magnetic flux density
The previously used unit was the gauss
(G).
1 G = 10
-4
Vs m
-2
= 10
-4
T (T = Tesla)
3/15: Magnetic flux
The previously used unit was the maxwell
(M).
1 M = 10
-8
Wb (Weber)
3/16: Standard volume
DIN 1343 must be complied with.
The designation m
3
(NTP) or m
3
(p
n
, T
n
)
is proposed, though the expression in
parentheses does not belong to the unit
symbol m3 but points out that it refers to
the volume of a gas in its normal state
(T
n
= 273 K, p
n
= 1013 mbar).
3/17: Partial pressure
The index i indicates that it is the partial
pressure of the i-th gas that is contained
in a gas mixture.
3/18: Gas permeability
The permeation coefficient is defined as
the gas flow m
3
s
-1
(volumetric flow p
V
)
that goes through a fixed test unit of a
given area (m
2
) and thickness (m) at a
given pressure difference (bar).
According to DIN 53.380 and DIN 7740,
Sheet 1, supplement, the gas permeability
(see no. 40) is defined as the volume of a
gas, converted to 0 C and 760 torr, which
goes through 1 m
2
of the product to be
tested at a certain temperature and a cer-
tain pressure differential during a day
(= 24 hours).
3/19: pV throughput/pV value
DIN 28.400, Sheet 1 is to be taken into
account here. No. 86 and no. 87 have a
quantitative physical significance only if
the temperature is indicated in each case.
3/20: Relative atomic mass
Misleadingly called atomic weight in the
past!
3/21: Relative molecular mass
Misleadingly called molecular weight in
the past!
3/22: Specific gas constant
As mass-related gas constant of the
substance i. Ri = Rm ( Mi
-1
; Mi molar
mass (no. 74) of the substance i. See
also DIN 1345.
3/23: Specific heat capacity
Also called specific heat:
Specific heat (capacity) at constant
pressure: c
p
.
Specific heat (capacity) at constant
volume: c
V
.
3/24: Temperature difference
Temperature differences are given in K,
but can also be expressed in C. The
designation degrees (deg) is no longer
permissible.
D00
D00 E 19.06.2001 21:41 Uhr Seite 169
Statutory Units
3/25: Quantity of heat
The units calorie (cal) and kilocalorie (kcal)
are no longer be used.
1 kcal = 4.2 kJ
3/26: Angle
1 radian (rad) is equal to the plane angle
which, as the central angle of a circle, cuts
out an arc having a length of 1 m from the
circle. See also DIN 1315 (8/82).
1 = rad: 1 = 1/60; 1 = 1/60.

180
360
1 rad = 60
2
10.4 Tables
10.4.1 Basic SI units
Basic unit Symbol Variable
meter m length
kilogramm kg mass
second s time, period;
duration
ampere A electric current
kelvin K thermodyn.
temperature
mole mol quantity of
substance
candela cd luminous
intensity
10.4.2 Derived coherent
1)
SI
units with special names
and symbols
(alphabetical)
Name Symbol Variable Relationship
of unit
coulomb C quantity of
electricity
or electric
charge 1 C = 1 A s
farad F electrical
capacitance 1 F = 1 A s V
1
henry H inductance 1 H = 1 V s A
1
hertz Hz frequency 1 Hz = 1 s1
joule J energy, work, 1 J = 1 N m
quantity = Ws
of heat
lumen lm luminous flux 1 lm = cd sr
lux lx illuminance 1 lx = 1 lm m
2
newton N force 1 N = 1 kgm s
2
ohm electrical 1 = 1 V A
1
resistance
pascal Pa pressure, 1 Pa = 1 N m
2
mechanical
stress
radian rad
2)
angle, 1 rad = 1 m m
1
plane angle
siemens S electrical
conductance 1 S = 1
1
steradian sr
2)
solid angle 1 sr = 1 m
2
m
2
tesla T magnetic 1 T = 1 Wb m
2
flux density or
induction
volt V electrical 1 V = 1 W A
1
voltage
or potential
difference
watt W power, 1 W = 1 J s
1
energy flux,
heat
flux
weber Wb magnetic 1 Wb = 1 V s
flux 1 Wb = 1 V s
1)
Formed with numeric factor 1; e.g. 1 C
= 1 As, 1 Pa = 1 N m
2
2)
Additional SI unit
10.4.3 Atomic units
Basic unit Symbol Vartiable
Atomic Mass for

indication of
mass m
u
particle mass;
unit 1 m
u
= 1/12 mass
of 12C
also amu
(atomic mass
unit).
Electron eV energy
volt
10.4.4 Derived noncoherent
SI units with special
names and symbols
Basic unit Symbol Definition
Day d 1 d = 86.400 s
Hour h 1 h = 3.600 s
Minute min 1 min = 60 s
Round angle 2 p rad
Degree () rad
180
Minute () rad
10.800
1
(= grad)
60
p
Second () rad
648.000
1
(= minute)
60
D00 E 19.06.2001 21:41 Uhr Seite 170
Vacuum Technology Standards
11. National and
international
standards and
recommendati-
ons particular-
ly relevant to
vacuum
technology
For around 20 years now, numerous stan-
dards and recommendations have been
drawn up at national and international level
and revised, whenever necessary, in
accordance with the state of the art. These
standards and recommendations must be
observed whenever use is made of vacu-
um equipment (pumps, gauges, valves,
etc.) and vacuum apparatus, systems and
plants are assembled. They not only con-
tain specifications applying specially to
vacuum technology, but also go beyond
this specific field and involve, for example,
physical units, formulas, noise protection
regulations, etc.
National standards are primarily DIN stan-
dards, particularly those relating to the
area of vacuum technology in the DIN
Standards Committee on Mechanical
Engineering (NAM). International stand-
ards and recommendations are drawn up
and issued
a) by the International Standardization
Organization (ISO), in particular by ISO
Committee TC 112 (vacuum technolo-
gy)
b) by the European Committee of Manu-
facturers of Compressors, vacuum
pumps and compressed air tools
(PNEUROP), in particular by PNEUROP
Subcommittee C5 (vacuum technology)
c) by the European Committee for Stan-
dardization (CEN), in particular by Tech-
nical Committee TC 138 (nondestruc-
tive testing) and Technical Committee
TC 318.
The content of the documents drawn up by
the international organizations in a) to c),
with German participation (also by
LEYBOLD), has been extensively incorpo-
rated into DIN standards, as reflected in
designations such as DIN / ISO or DIN /
EN.
The most important standards to be com-
plied with are listed in Table 11.1 below.
Abbreviations used:
D = draft
CD = Committee Draft
11.1 National and
international
standards and
recommendations
of special
relevance to
vacuum technology
A) National agreements, Part 1: DIN
DIN Title Issue
1301-8 E Units
Part 1 Names of units,
symbols 1993
Part 2 Parts and multiples
generally used 2/78
Part 3 Conversions for units
no longer used 6/79
1304 General symbols
Part 1 General symbols 3/94
Part 2 Symbols for
meteorology and
geophysics 9/89
Part 3 Symbols for electrical
energy supply 3/89
Part 5 Symbols for flow
mechanics 9/89
Part 6 Symbols for electrical
communications
technology 5/92
Part 7 Symbols for
meteorology and
geophysics
Part 7 Symbols for
electric machines 1/91
1305 Mass; weighed value, force,
weight force, weight
load, definitions 1/88
1306 Density; definitions 6/84
1313 Physical variables and
equations, definitions, spelling 4/78
1314 Pressure; basic definitions,
units 2/77
DIN Title Issue
1319 Basic definitions for
measurement technology
Part 1 Basic definitions 1/95
Part 2 Definitions for the
use of gauges 1/80
Part 3 Definitions for
measurement
uncertainty and
evaluation of gauges
and measuring
equipment 8/83
Part 4 Treatment of use
of measurements 12/85
A) National agreements, Part 1: DIN
(cont.)
DIN Title Issue
1343 Normal state, reference state 1/90
1345 Thermodynamics; basic
definitions
12/93
1952 Flow measurement with
screens, nozzles, etc. 7/82
2402 Pipelines; nominal internal
diameters, definitions,
classification 2/76
3535 Seals for gas supply Part 6 4/94
8964 Circuit parts for refrigeration
systems with hermetic and
semi-hermetic compressors
Part 1: Tests 3/96
Part 2: Requirements 9/86
(E 12/95)
16005 Overpressure gauges with
elastic measuring element for
general use Requirements
and testing 2/87
16006 Overpressure gauges with
Bourdon tube Safety-related
requirements and testing 2/87
19226 1 Control and
instrumentation
technology;
control and regulation
technology; definitions
general principles 2/94
4 Control and
instrumentation
technology;
control and regulation
technology;
definitions for control
and regulation systems 2/94
5 Control and
instrumentation
technology; control and
regulation technology;
functional definitions 2/94
25436 Integral leak rate test of safety
vessel with absolute pressure
method 7/80
D00
D00 E 19.06.2001 21:41 Uhr Seite 171
DIN Title Issue
28090 Static seals for flange
connections
Part 1 - Characteristic values
for seals and
testing methods 9/95
Part 2 - Seals made of
sealing plate; special
testing methods for
quality assurance 9/95
28400 Vacuum technology;
designations and definitions
Part 1 - Basic definitions, units,
ranges, characteristic
variables and basic
principles 5/90
Part 2 - Vacuum pumps 10/80
Part 3 - Vacuum plants;
characteristic
variables and gauges 6/92
Part 4 - Vacuum coating
technology 3/76
Part 5 - Vacuum drying and
vacuum freeze-drying 3/81
Part 6 - Analytical techniques
for surface technology 10/80
Part 7 - Vacuum metallurgy 7/78
Part 8 - Vacuum systems,
components and
accessories 10/80
(E 7/91)
symbols overview 11/76
28402 Vacuum technology: variables,
symbols, units - overview 12/76
28403 Vacuum technology; quick
connections, small flange
connections 9/86
28404 Vacuum technology: flanges,
dimensions 10/86
28410 Vacuum technology; mass
spectrometer partial pressure
gauges, definitions,
characteristic variables,
operating conditions 11/86
acceptance specifications
for mass spectrometer leak
detection devices, definitions 3/76
calibration of vacuum gauges in
a range from 10
-3
to 10
-7
mbar.
General methods; pressure
reduction through constant flow 3/76
28417 Vacuum technology; measuring
pV mass flow according to
volumetric method at constant
pressure 3/76
28418 Vacuum technology; standard
method for calibrating vacuum
gauges through direct compa-
rison with a reference device
Part 1 General principles 5/76
Part 2 Ionization vacuum
gauges 9/78
Part 3 Thermal conductivity
vacuum gauges 8/80
DIN Title Issue
acceptance specifications or
frotary piston vacuum pumps
Part 1 Rotary piston and
vane type rotary
vacuum pumps in
rough and medium
vacuum range
Part 2 Roots vacuum pumps
in medium
vacuum range
for diffusion pumps and
ejector vacuum pumps for
pump fluid vapor pressures of
less than 1 mbar 2/83
for turbo-molecular pumps 11/78
for getter-ion pumps 8/85
measuring specifications for
ejector vacuum pumps and
ejector compressors.
Pump fluid: water vapor 11/84
28431 Acceptance specifications
for liquid ring vacuum pumps 1/87
28432 Acceptance specifications for
diaphragm vacuum pumps E 5/95
53380 Testing of plastic foils,
determination of gas
permeability 6/69
(E 10/83)
45635 Noise measurement at
machines: measurement
of airborne noise, enveloping
surface methods.
Part 13 Compressors,
including vacuum
pumps, positive
displacement, turbo
and steam ejectors 2/77
55350 Definitions of quality
assurance and statistics
Part 11 Basic definitions of
quality assurance 8/95
Part 18 Definitions regarding
certification of
results of quality
tests/quality test
certificates 7/87
66038 Torr millibar; millibar torr
conversion tables 4/71
Thesaurus Vacui
(definition of terms) 1969
A) European/national agreements, EN,
DIN/EN, CEN
DIN/EN Title Issue
EN 473 Training and certification of
personnel for nondestructive
testing (including leak test) 7/93
837-1 Pressure gauges, Part 1:
Pressure gauges with
Bourdon tubes, dimensions,
measurement
technology, requirements
and testing 2/97
Selection and installation
recommendations for
pressure gauges 1/95
Pressure gauges with
plate and capsule elements,
dimensions, measurement
technology, requirements
and testing 2/97
1330-8 E Nondestructive testing -
definitions for leak test
terminology 6/94
1779 E Nondestructive
testing leak test. Instructions
for selection of a testing
method 3/95
1338-8 E Nondestructive
testing leak test.
Terminology on leak test 1994
testing determination of
characteristic variables for
mass spectrometer
leak detectors
testing bubble type
testing method 12/94
NMP 826 Calibration of gaseous
reference leaks, CD 9/95
Nr. 0995
B) International agreements, ISO, EN/ISO
DIN Title Issue
1000 SI units and recommendations
for the use of their multiples
and of certain other units 11/92
1607 / 1 Positive displacement vacuum
pumps. Measurement of
performance characteristics. 12/93
Part 1 - Measurement of
volume rate of flow
(pumping speed)
1607 / 2 Positive displacement vacuum
pumps. Measurement of
performance characteristics. 11/89
Part 2 - Measurement of
ultimate pressure
D00 E 19.06.2001 21:41 Uhr Seite 172
DIN Title Issue
1608 / 1 Vapor vacuum pumps.
Part 1 Measurement of volume
rate of flow 12/93
1608 / 2 Vapor vacuum pumps.
Part 2: Measurement of critical
backing pressure 12/89
1609 Vacuum technology. Flange
dimensions 3/86
DIN/ISO Standard atmosphere 12/79
2533
2861 / 1 Quick release couplings.
Dimensions 8/74
Part 1 - Clamped Type
2861 / 2 Quick release couplings.
Dimensions 8/80
Part 2 - Screwed type
3529 / 1 Vacuum Technology
Vocabulary 12/81
Part 1 - General terms
Vocabulary 12/81
Part 2 - Vacuum pumps and
related terms
Vocabulary 12/81
Part 3 - Vacuum gauges
3556 / 1 Measurement of performance
characteristics. 1992
Part 1 - Sputter ion pumps (E)
3567 Vacuum gauges. Calibration by
direct comparison with a
reference gauge (CD) 2/91
3568 Ionisation vacuum gauge.
Calibration by direct
comparison with a
reference gauge (CD) 2/91
3570 / 1 Vacuum gauges standard
methods for calibration 2/91
Part 1 - Pressure reduction
by continuous flow in
the pressure range
10
1
... 10
5
Pa
3669 Vacuum Technology. Bakable
flanges, dimensions. 2/86
Part 1: Clamped Type
EN/ISO Rubber and plastic hoses and
4080 hose lines determination of
gas permeability 4/95
5167 Measurement of fluid flow by
means of orifice
plates, nozzles etc. 1980
5300 Vacuum gauges of the thermal
conductivity type. 2/91
Calibration by direct
comparison with a
reference gauge (CD)
9803 Pipeline fittings-mounting,
dimensions (E) 2/93
DIN Title Issue
DIN/ISO Requirements placed on quality
10012 assurance for measuring
equipment 8/92
Part 1 Confirmation system
for measuring
equipment
C) PNEUROP/C5 (6.93)
Number Title/remark identical Issue
to DIN
5607 Vacuum pumps;
acceptance specifications 28427 1972
Part II: (Fluid
entrainment pumps)
5608 Vacuum pumps;
Part III: (Turbomolecular
pumps)
5615 Vacuum pumps;
Part IV: (Getter-ion
pumps)
6601 Measurement of per
formance of ejector
vacuum pumps and
ejector compressors 28430 5/78
6602 Vacuum pumps;
Part I: (Oil-sealed rotary
pumps and
Roots pumps) 28426 1979
6606 Vacuum flanges and
connections; 28403 1985
dimensions and
28404
PN5ASR Vacuum pumps,
CC/5 acceptance specifications
refrigerator cooled
cryopumps 7/89
D00
D00 E 19.06.2001 21:41 Uhr Seite 173
References
12. References
1. Overview,
definitions and
history
K. Diels, R. Jaekel
Leybold Vacuum Handbook
Pergamon Press
1st Ed. 1966
W. Haeder, E. Grtner
Die gesetzlichen Einheiten in der Technik
Beuth-Vertrieb GmbH, 5. Aufl. 1980,
Berlin 30, Kln, Frankfurt (Main)
H. Ebert
Vakuum-Chronik, A documentation on
works concerning vacuum that were
published before 1928
PTB-Bericht ATWD-11, September 1977
M. Dunkel
Gedenken an Wolfgang Gaede
Physikalische Bltter Nr. 34 (1978),
Heft 5, Pages 228-232 as well as
Vakuum-Technik, 27. Jahrgang, No. 4,
Pages 99-101
IUPAP (SUNANCO Commission)
Symbols, Units etc.
Document 25, 1987
Leybold AG
Vademekum, 93 pages, 1988
M. Wutz, H. Adam, W. Walcher
Theory and Practice of
Vacuum Technology
5. Aufl., 696 pages, 1992,
Friedrich Vieweg u. Sohn,
Braunschweig/ Wiesbaden
A. Guthrie and R. K. Wakerling
Vakuum Equipment and Techniques
264 pages, 1949, McGraw-Hill,
New York/London/Toronto
D. J. Hucknall
Vacuum Technology and Applications
1st Ed., 319 pages, 1991
Butterworth-Heinemann, Oxford
C. M. van Atta
Vacuum Science and Engineering
459 pages, 1965, McGraw-Hill
New York/San Francisco/Toronto/
London/Sydney
J. M. Lafferty et. al.
Foundations of Vacuum Science and
Technology
704 pages, 1998, Wiley 1998
A. Schubert
Normen und Empfehlungen fr die
Vakuum-Technik
Vakuum in der Praxis,
Vol. 3, 1991, 211-217
H. Scharmann
Vakuum Gestern und Heute
Vol. 2, 1990, 276-281
M. Auwrter
Das Vakuum und W. Gaede
Vakuum-Technik, Vol. 32, 1983, 234-247
J. F. OHanlon
A Users Guide to Vacuum Technology
3nd Ed., 402 pages, Wiley 1989,
New York
G. Reich
Wolfgang Gaede Einige Gedanken zu
seinem 50. Todestag aus heutiger Sicht
7th year, 1995, 136-140
S. German, P. Draht
Handbuch SI Einheiten
Vieweg Braunschweig/Wiesbaden,
1979, 460 pages
Gesetz ber Einheiten im Mewesen
vom 2. Juli 1969
Gesetz zur nderung des Gesetzes ber
Einheiten im Mewesen vom 6. Juli
1973
Ausfhrungsverordnungen vom 26.
Juni 1970
In Vakuum-Technik Vol. 35, 1986:
Th. Mulder
Otto von Guericke
Pages 101-110
P. Schneider
Zur Entwicklung der Luftpumpen-
Initiationen und erste Reife
bis 1730 Pages 111-123
L. Fabel
Physik in der 2. Hlfte des 19. Jahr
hunderts und die vakuumtechnische
Entwicklung bis Gaede
Pages 128-138
H.-B. Brger:
G. Ch. Lichtenberg und die Vakuum
technik
Pages 124-127
G. Reich:
Gaede und seine Zeit
Pages 139-145
H. Adam
Vakuum-Technik in der Zeit nach
Gaede (1945 to the present);
Pages 146-147
G. Reich
Die Entwicklung der Gasreibungspumpen
von Gaede, ber Holweck, Siegbahn bis
zu Pfleiderer und Becker (mit zahlreichen
Literaturangaben)
Vakuum-Technik in der Praxis, Vol. 4,
1992, 206-213
G. Reich
Carl Hoffman (1844-1910), der Erfinder
der Drehschieberpumpe
Vakuum in der Praxis, 1994, 205-208
Th. Mulder
Blaise Pascal und der Puy de Dme
Groe Mnner der Vakuum-Technik
W. Pupp und H. K. Hartmann
Vakuum-Technik, Grundlagen und Anwen-
dungen
C. Hanser, Mnchen, 1991, Wien,
2. Vacuum pumps
2.1 Positive displacement
pumps, condensers
W. Gaede
Demonstration einer rotierenden
Quecksilberpumpe
Physikalische Zeitschrift, 6, 1905, 758-760
D00 E 19.06.2001 21:41 Uhr Seite 174
W. Gaede
Gasballastpumpen
Zeitschrift fr Naturforschung,
2a, 1947, 233-238
W. Armbruster und A. Lorenz
Das maximale Kompressionsverhltnis
und der volumetrische Wirkungsgrad von
Vakuumpumpen nach dem Rootsprinzip
Vakuum-Technik, 7, 1958, 81-85
W. Armbruster und A. Lorenz
Die Kombination Rootspumpe-Wasser-
ringpumpe
H. Reylander
ber die Wasserdampfvertrglichkeit von
Gasballastpumpen
F. Fauser
Charakteristik von Pumpsystemen fr
grere Wasserdampfmengen unter
Vakuum und unter Anwendung von Kon-
densation und Kompression des
Wasserdampfes
1965 Transactions of the Third Internatio-
nal Vacuum Congress, Stuttgart, Bd. 2/II,
393-395, Pergamon Press, Oxford 1966
M. Wutz
Das Abpumpen von Dmpfen mit gekhl-
ten Kondensatoren
H. Hamacher
Kennfeldberechnung fr Rootspumpen
DLR FB 69-88, 1969
H. Hamacher
Beitrag zur Berechnung des
Saugvermgens von Rootspumpen
Vakuum-Technik, 19, 1970, 215-221
H. Hamacher
Experimentelle Untersuchungen an
Nachkhlern von Rootspumpen
Vakuum-Technik, 23, 1974, 129-135
M. Rannow
lgedichtete Vakuumpumpen in der
Chemie
Chemie-Technik, No. 7, 1978, 39-41
H. P. Berges et al.
TRIVAC-B, ein neues Vakuumpumpen-
Konzept fr universelle Anwendungen
Vakuum-Technik, 31, 1982, 168-171
References
H. Lang
Vakuumpumpen in der chemischen Indu-
strie Wlzkolbenpumpen
Vakuum-Technik, 1980, 72-82
H. F. Weber
Vakuumpumpen in der chemischen
Industrie lgedichtete
Rotationsvakuumpumpen
D. Bartels
Vakuumpumpen in der chemischen
Industrie
Flssigkeitsring-Vakuumpumpen A
R. W. Adam und C. Dahmlos
Flssigkeitsring-Vakuumpumpen B
U. Seegebrecht
Frderung trockener Luft und von
gesttigtem Luft-Wasserdampfgemisch
mit Flssigkeitsring-Vakuumpumpen
H.-D. Brger
Fortschritte beim Betrieb von
Wlzkolbenpumpen
Vakuum-Technik 1983, 140-147
U. Seegebrecht
Einflu der Temperatur des Frdermittels
auf das Saugvermgen von Flssigkeits-
ring-Vakuumpumpen bei der Frderung
von trockener Luft
P. Bachmann und H.-P. Berger
Sicherheitsaspekte beim Einsatz von
lgedichteten Drehschiebervakuumpum-
pen in CVD-Anwendungen
U. Fussel
Trockenlaufende Vakuumpumpen in der
chemischen Industrie
L. Ripper
Explosionsschutz-Manahmen an Vaku-
umpumpen (with numerous references to
relevant literature)
K. P. Mller
Trockenlaufende Drehschiebervakuum-
pumpen in einer Vielzweck-Produktions-
anlage
F. J. Eckle, W. Jorisch, R. Lachenmann
Vakuum-Technik im Chemielabor
P. Bachmann und M. Kuhn
Einsatz von Vorpumpen im Al-tzproze.
Erprobung trockenverdichtender Klauen-
pumpen und lgedichteter Drehschieber-
Vakuumpumpen im Vergleich
Vakuum in der Praxis, 1990, 15 21
U. Gottschlich
Vakuumpumpen im Chemielabor
M. H. Hablanian
Aufbau und Eigenschaften verschiedener
lfreier Vakuumpumpen fr den Grob-
und Feinvakuumbereich (wichtige
Literaturangaben)
B. W. Wenkebach und J. A. Wickhold
Vakuumerzeugung mit Flssigkeitsring-
Vakuumpumpen
U. Gottschlich und W. Jorisch
Mechanische Vakuumpumpen im
Chemieeinsatz
Vakuum in Forschung und Praxis, 1989,
113-116
W. Jorisch
Neue Wege bei der Vakuumerzeugung in
der chemischen Verfahrenstechnik
D. Lamprecht
Trockenlaufende Vakuumpumpen
P. Deckert et al.
Die Membranvakuumpumpe Entwick-
lung und technischer Stand
W. Jorisch und U. Gottschlich
Frischlschmierung Umlaufschmierung,
Gegenstze oder Ergnzung?
D00
D00 E 19.06.2001 21:41 Uhr Seite 175
References
W. Jitschin et al.
Das Saugvermgen von Pumpen: Unter-
suchung verschiedener Meverfahren im
Grobvakuumbereich
(1995) 183 -193
H.P. Berges and M. Kuhn
Handling of Particles in Forevacuum
pumps
Vacuum, Vol. 41, 1990, 1828-1832
M. H. Hablanian
The emerging technologies of oil-free
vacuum pumps
J. Vac. Sci. Technol.
A6 (3), 1988, 1177-1182
E. Zakrzewski, P. L. May and B. S. Emslie
Developments in vacuum Pumping
systems based on mechanical pumps
with an oil free swept volume
Vacuum, 38, 968, 757-760
H. Wycliffe
Mechanical high-vacuum pumps with an
oil-free swept volume
J. Vac. Sci. Technol.
A5 (4) 1987, 2608-2611
A. P. Troup and D. Turell
Dry pumps operating under harsh
condictions in the semiconductor industry
J. Vac. Sci. Technol. A7 (3),
1989, 2381-2386
P. Bachmann and M. Kuhn
Evaluation of dry pumps vs rotary vane
pumps in aluminium etching
Vacuum 41, 1990, 1825-1827
H. P. Berges and D. Gtz
Oil-free vacuum pumps of compact
design
Vacuum, Vol. 38, 1988, 761-763
2.2 Turbomolecular pumps
W. Gaede
Die Molekularluftpumpe
Annalen der Physik, 41, 1913, 337-380
W. Becker
Eine neue Molekularpumpe
W. Armbruster
Vakuumpumpenkombinationen fr Labor,
Technikum und Produktion
Chemiker-Zeitung / Chemische Apparatur,
88, 1964, 895-899
W. Becker
Die Turbo-Molekularpumpe
Vakuum-Technik, 15, 1966, 211-218 and
254-260
R. Frank et al.
Leistungsdaten von Turbo-Molekularpum-
pen des Typs TURBOVAC mit senkrecht
angeordnetem Axialkompressor
Vakuum-Technik, 24, 1975, 78 -85
W. Becker
Eine gegenberstellende Betrachtung von
Diffusionspumpen und Molekularpumpen
Ergebnisse europischer Ultrahochvaku-
umforschung
Leybold-Heraeus GmbH u. Co., in its own
publishing house, Cologne 1968, 41-48
R. Frank, E. Usselmann
Kohlenwasserstoffreier Betrieb mit Turbo-
Molekularpumpen des Typs TURBOVAC
R. Frank, E. Usselmann
Magnetgelagerte Turbo-Molekularpumpen
des Typs TURBOVAC
Vakuum-Technik, 25, 1976, 141-145
H.-H. Henning und G. Knorr
Neue luftgekhlte, lageunabhngige
Turbo-Molekularpumpen fr Industrie und
Forschung
H.-H. Henning und H. P. Caspar
Wlzlagerungen in Turbo-Molekularpum-
pen
E. Kellner et al.
Einsatz von Turbo-Molekularpumpen bei
Auspumpvorgngen im Grob- und
Feinvakuumbereich
D. E. Gtz und H.-H. Henning
Neue Turbo-Molekularpumpe fr
berwiegend industrielle Anwendungen
J. Henning
30 Jahre Turbo-Molekularpumpe
P. Duval et. al.
Die Spiromolekularpumpe
G. Reich
Berechnung und Messung der Abhngig-
keit des Saugvermgens von Turbo-Mole-
kularpumpen von der Gasart
J. Henning
Die Entwicklung der Turbo-Molekular-
pumpe
D. Urban
Moderne Bildrhrenfertigung mit
Turbo-Molekularpumpen
O. Ganschow et al.
Zuverlssigkeit von Turbo-Molekular-
pumpen
M. H. Hablanian
Konstruktion und Eigenschaften von
turbinenartigen Hochvakuumpumpen
J. H. Fremerey und H.-P. Kabelitz
Turbo-Molekularpumpe mit einer
neuartigen Magnetlagerung
H. P. Kabelitz and J.K. Fremerey
Turbomolecular vacuum pumps with a
new magnetic bearing concept
Vacuum 38, 1988, 673-676
E. Tazioukow et al.
Theoretical and experimental investigation
of rarefied gas flow in molecular pumps
1995, 53-56
D00 E 19.06.2001 21:41 Uhr Seite 176
2.3 Fluid entrainment pumps
W. Gaede
Die Diffusion der Gase durch Quecksilber-
dampf bei niederen Drcken und die Dif-
fusionspumpe
Annalen der Physik, 46, 1915, 357-392
W. Gaede
Die ldiffusionspumpe
Z. techn. Physik, 13, 1932, 210-212
R. Jaeckel, H. G. Nller und H. Kutscher
Die physikalischen Vorgnge in Diffusi-
ons- und Dampfstrahlpumpen
W. Bchler und H. G. Nller
Fraktionierung und Entgasung in
l-Diffusionspumpen
Z. angew. Physik einschl. Nukleonik, 9,
1957, 612-616
H. G. Nller
Weshalb sind systematische Fehler bei
Saugvermgensmessungen besonders
gro fr Hochvakuumpumpen groer
Leistung ?
Vakuum-Technik, 12, 1963, 291-293
W. Bchler und H.-J. Forth
Die wichtigsten Einflugren bei der
Entwicklung von Diffusionspumpen
W. Reichelt
Bemerkungen zur Arbeitsweise moderner
Diffusionpumpen
Vakuum-Technik, 13, 1964, 148-152
H. G. Nller
Theory of Vacuum Diffusion Pumps
Handbook of Physics, Vol.1, Part 6,
(pp. 323...419) Ed. A. H. Beck, Pergamon
Press Ltd., London, W.I. (1966)
G. Herklotz
Enddruckversuche mit Diffusionspumpen
hohen Saugvermgens und
Restgasspektren
Vakuum-Technik, 20, 1971, 11 14
H. G. Nller
Die Bedeutung von Knudsenzahlen und
hnlichkeitsgesetzen in Diffusions- und
Dampfstrahlpumpen
References
R. Gsling
Treibmittelpumpen
M. Wutz
Grundlagen zur Bestimmung der
charakteristischen Daten von
Dampfstrahl-Ejektorpumpen
H. Bayer
Dampfstrahlpumpen
H. Bayer
Vakuumerzeugung durch
Dampfstrahl-Vakuumpumpen
F. Hinrichs
Aufbau, Betriebsverhalten und
Regelbarkeit von
Dampfstrahl-Vakuumpumpen
2.4 Sorption pumps
G. Kienel
Zur Desorption von Gasen in Getter-
Ionenpumpen in Physik und Technik von
Sorptions- und Desorptionsvorgngen bei
niederen Drcken
Rudolf A. Lange Verlag, 1963, Esch/Tau-
nus, 266-270
W. Bchler
Ionen-Zerstuberpumen, ihre
Wirkungsweise und Anwendung
publishing house, Cologne 1966
W. Espe
Zur Adsorption von Gasen und Dmpfen
an Molekularsieben
Feinwerktechnik, 70, 1966, 269-273
G. Kienel
Vakuumerzeugung durch Kondensation
und durch Sorption
Chemikerzeitung / Chem. Apparatur 91,
1967, 83-89 und 155-161
H. Hoch
Erzeugung von kohlenwasserstoffreiem
Ultrahochvakuum
Vakuum-Technik, 16, 1967, 156-158
W. Bchler und H. Henning
Neuere Untersuchungen ber den
Edelgas-Pumpmechnismus von
Ionenzerstuberpumpen des Diodentyps
Proc. of the Forth Intern. Vacuum
Congress 1968, I. 365-368,
Inst. of Physics, Conference Series No. 5,
London
H. Henning
Der Erinnerungseffekt fr Argon bei
Trioden-Ionenzerstuberpumpen
2.5 Cryopumps and
cryoengineering
R. A. Haefer
Cryo-Pumping
456 pages, 1989 Oxford University Press,
Oxford
H. Frey und R-A. Haefer
Tieftemperaturtechnologie, 560 pages,
VDI-Verlag, Dsseldorf, 1981
G. Klipping und W. Mascher
Vakuumerzeugung durch Kondensation
an tiefgekhlten Flchen, I. Kryopumpen
W. Bchler, G. Klipping und W. Mascher
Cryopump System operating down to 2,5
K, 1962 Trans. Ninth National Vacuum
Symposium, American Vacuum Society,
216-219, The Macmillan Company,
New York
G. Klipping
Kryotechnik Experimentieren bei tiefen
Temperaturen
Chemie-Ingenieur-Technik, 36,
1964, 430-441
M. Schinkmann
Messsen und Regeln tiefer Temperaturen,
Teil I: Thermodynamische Verfahren
Metechnik, 81, 1973, 175-181
G. Schfer, M. Schinkmann
Messsen und Regeln tiefer Temperaturen,
Teil II: Elektrische Verfahren,
Metechnik, 82, 1974, 31-38
D00
D00 E 19.06.2001 21:41 Uhr Seite 177
References
R. Frank et al.
Entwicklung von Refrigeratoren fr den
Einbau in Kryopumpen
Vakuum-Technik, 30, 1981, 134-137
J. J. Scheer und J. Visser
Anwendungen von Kryopumpen in der
industriellen Vakuumtechnik
P. Duval
Diffusionspumpen, Turbo-Molekularpum-
pen oder Kryopumpen ? Auswahlkriteri-
en fr Hochvakuumpumpen
H. Henning und H.-H. Klein
Pumpen von Helium mit
Refrigerator-Kryopumpen
Vakuum-Technik, 34, 1985, 181-184
H.-H. Klein et al.
Einsatz von Kryopumpen in
Produktionsanlagen
Vakuum-Technik, 34, 1986, 203-211
D. Mller und M. Sydow
Kryopumpen im Vergleich mit anderen
Hochvakuumpumpen
Vakuum in der Praxis, 2, 1990, 270-274
G. Kiese und G. Vo
Kryopumpen mit neuartiger
Regenerationstechnik
2.6 Oil backstreaming
G. Levin
A quantitativ appraisal of the
backstreaming of forepump oil vapor
J. Vac. Sci. Technol. A 3 (6), 1985,
2212-2213
M. A. Baker and L. Laurenson
A quartz crystal microbalance holder for
low Temperature use in vacuum
Vacuum Vol. 17, (12), 647-648, 1967
(Letters to the Editor)
M. A. Baker and W. Steckelmacher
The Measurement of Contamination in
Vacuum Systems
Vuoto, scienza e technologia, Bd.3 ,
(1/2), 3-17, 1970
J. P. Deville, L. Holland and L. Laurenson
Measurement of the rate of evaporation of
Pump oils using a crystal vibrator
3rd. Internat. Vac. Congr Stuttgart 153-
160, Pergamon Press, Oxford, 1965
L. Laurenson, S. Hickman and
R. G. Livesey
Rotary pump backstreaming: An
analytical appriasal of practical results
and the factors affecting them
J. Vac. Sci. Technol. A 6 (2),
238-242, 1988
B. D. Power, A. M. I. Mech, E. Crawley
and D. J. Crawley
Sources, Measurement and Control of
Backstreaming in Oil Vapour Vacuum
Pumps
Vacuum, Vol. 4 (4), 415-437, 1957
M. A. Baker
A cooled quartz crystal microbalance
methode for measuring diffusion pump
backstreaming
Journal of Scientific Instruments (Journal
of Physics E), Series 2, Volume 1,
774-776, 1968
N. S. Harris
Diffusion pump back-streaming
Vacuum, Vol. 27 (9), 519-530, 1977
M. A. Baker
Vapour and Gas Measurements in Vacu-
um with the Quartz Crystal Microbalance
in Vol.1, Proceedings of the ninth Confe-
rence on Vacuum Microbalance Techni-
ques, Progress in Vacuum Microbalance
Techniques
Th. Gast and E. Robens ed.,
Heyden & Son Ldt., London, New York,
Rheine, 1970
M. A. Baker and L. Laurenson
The use of a quartz crystal microbalance
for measuring vapour backstreaming
from mechanical pumps
Vacuum, Volume 16 (11), 633-637, 1966
R. D. Oswald and D. J. Crawley
A method of measuring back migration of
oil through a baffle
Vacuum, Vol. 16 (11), 623-624, 1966
M. H. Hablanian
Backstreaming Measurements above
Liquid-Nitrogen Traps
Vac. Sci. Tech., Vol. 6, 265-268, 1969
Z. Hulek, Z. Cespiro, R. Salomonovic,
M. Setvak and J. Voltr
Measurement of oil deposit resulting from
backstreaming in a diffusion pump
system by proton elastic scattering
Vacuum, Vol. 41 (7-9), 1853-1855, 1990
M. H. Hablanian
Elimination of backstreaming from
mechanical vacuum pumps
J. Vac. Sci. Technol. A5 (4), 1987,
2612-2615
3. Ultrahigh vacuum
technology
G. Kienel
Probleme und neuere Entwicklungen auf
dem Ultrahochvakuum-Gebiet
VDI-Zeitschrift, 106, 1964, 777-786
G. Kienel und E. Wanetzky
Eine mehrmals verwendbare
Metalldichtung fr ausheizbare
Uktrahochvakuum-Ventile und
Flanschdichtungen
H. G. Nller
Physikalische und technische Vorausset-
zungen fr die Herstellung und Anwen-
dung von UHV-Gerten.
Ergebnisse europischer Ultrahochvaku-
um Forschung
LEYBOLD-HERAEUS GmbH u. Co., in its
own publishing house,
Cologne 1968, 49-58
W. Bchler
Probleme bei der Erzeugung von
Ultrahochvakuum mit modernen
Vakuumpumpen. Ergebnisse
europischer Ultrahochvakuum
Forschung
D00 E 19.06.2001 21:41 Uhr Seite 178
References
P. Readhead, J. P. Hobson und
E. V. Kornelsen
The Physical Basis of Ultrahigh Vacuum
Chapman and Hall, London, 1968
E. Bergandt und H. Henning
Methoden zur Erzeugung von
Ultrahochvakuum
Vakuum-Technik, 25,1970, 131-140
H. Wahl
Das Hochvakuumsystem der CERN am
450 GeV Supersynchrotron und
Speichering (SPS)
F. Grotelschen
Das UHV-System bei DESY. 1. Teil
D. Trines
Das Strahlrohrvakuumsystem des Hera-
Protonenringes
G. Schrder et al.
COSV- eine neue Forschungsanlage mit
UHV-Technologie
W. Jacobi
Das Vakuumsystem der GSI-Beschleuni-
geranlage
4. Conductances,
flanges, valves,
etc.
M. Knudsen
Gesetze der Molekularstrmung und der
inneren Reibungsstrmung der Gase
durch Rhren
Annalen der Physik, 4th issue, 28, 1909,
75-130
P. Clausing
ber die Strmung sehr verdnnter Gase
durch Rhren von beliebiger Lnge
Annalen der Physik, 5th issue, 12,
1932, 961-989
W. Rllinger
Die Verwendung von Klammerflanschen in
der Vakuumtechnik
H. Hoch
Ausheizbare Verbindungen an Hoch-
vakuum-Apparaturen
Vakuum-Technik, 10, 1961, 235-238
W. Bchler und I. Wikberg
Dual Seal Bakable Section Valves of the
CERN Intersection Storage Ring
Vacuum, 21, 1971, 457-459
K. Teutenberg
UHV-Ganzmetallventile groer Nennweite
Vakuum-Technik, 21, 1972, 169-174
H. Henning
The approximate calculation of
transmission probabilities
Vacuum, 28, 1978, No. 3, Seite 151
G. Khn
Gasstrme durch Spalte im Grobvakuum
Vakuum-Technik, 33, 1984, 171-175
R. Haberland und B. Vogt
UHV-Ventil fr extrem viele Schliezyklen
Vakuum-Technik, 34, 1985, 184-185
A. Sele
Vakuum-Ventile (VAT)
L. Fikes
Berechnung von Auspumpkurven mit
Hilfe der Analogie von Gasstrom und
elektrischem Strom
W. Herz
Zuverlssige Flanschverbindung im
Anwendungsgebiet der Tieftemperatur-
und Vakuumtechnik
5. Measurement of
low pressures
C. Meinke und G. Reich
Vermeidung von Fehlmessungen mit dem
System McLeod-Khlfalle
P. A. Readhead and J. P. Hobson
Total Pressure Measurem. below 1010
Torr with Nonmagnetic Ionisation Gauge
Brit. J. Appl. Phys., 16, 1965, 1555-1556
C. Meinke und G. Reich
Comparison of Static and Dymanic
Calibration Methods for Ionisation Gauges
J. Vac. Sci. Techn., 4, 1967, 356-359
G. Reich und W. Schulz
Probleme bei der Verwendung von
Ionisations-Vakuummetern im Druck-
bereich oberhalb 10
2
Torr
Proc. of the Fourth Intern. Vacuum
Congress, 1968,
II. Inst. of Physics Conference Series No.
6, London, 661-665
G. Reich
Probleme bei der Messung sehr niedriger
Total- und Partialdrcke
Ergebnisse europischer Ultrahoch-
vakuum Forschung
A. Barz and P. Kocian
Extractor Gauge as a Nude System
J. Vac. Sci Techn. 7, 1970, 1, 200-203
U. Beeck and G. Reich
Comparison of the Pressure Indication of a
Bayard-Alpert and an Extractor Gauge
J. Vac. Sci. and Techn. 9, 1972, 1,126-128
U. Beeck
Untersuchungen ber die Druckmessun-
gen mit Glhkathoden-Inisations-Vakuum-
metern im Bereich grer als 10
3
Torr
G. Reich
ber die Mglichkeiten der Messung sehr
niedriger Drcke
Metechnik, 2, 1973, 46-52
D00
D00 E 19.06.2001 21:41 Uhr Seite 179
References
G. Reich
Spinning rotor viscosity gauge; a transfer
standard for the laboratory or an accurate
gauge for vacuum process control
J. Vac. Sci. Technol., 20 (4),
1982, 1148-1152
G. Reich
Das Gasreibungs-Vakuummeter
VISCOVAC VM 210
Vakuum-Technik, 31, 1982, 172-178
G. Grosse and G. Messer
Calibration of Vacuum Gauges at
Pressures below 109 mbar with a
molecular beam method
Vakuum-Technik, 30, 1981, 226-231
Chr. Edelmann et al.: Mglichkeiten der
Mebereichserweiterung bei Glhkatho-
den-Ionisationsmanometern (numerous
references to relevant literature)
Chr. Edelmann
Stand und Entwicklungstendenzen der
Totaldruckmessung in der Vakuum-Tech-
nik
Vakuum-Technik, 33, 1984, 162-180
J. K. Fremerey
Das Gasreibungsvakuummeter
Vakuum-Technik, 36, 1987, 205-209
G. Messer
Kalibrierung von Vakuummetern
Vakuum-Technik, 36, 1987, 185-192
G. Messer und W. Grosse
Entwicklung der Vakuum-Metrologie in der
PTB (numerous references to relevant li-
terature)
Vakuum-Technik, 36, 1987, 173-184
G. Reich
Industrielle Vakuummetechnik
Vakuum-Technik, 36, 1987, 193-197
L. Schmidt und E. Eichler
Die Praxis einer DKD-Kalibrierstelle
C. Kndig
Vakuummegerte fr Totaldruck
Chr. Edelmann
Glhkahtoden-Ionisationsmanometer fr
hohe Drcke im Vakuumbereich
M. Ruschitzka and W. Jitschin
Physikalische Grundlagen des
Wrmeleitungsvakuummeters
T. Koopmann
Neue Trends in der Vakuum-Metechnik
Chr. Edelmann
Die Entwicklung der Totaldruckmessung
im UHV- und Extremvakuumbereich
W. Jitschin
Kalibrierung, Abnahme und Zertifizierung
(with numerous references to relevant li-
terature)
W. Jitschin
Obere Mebereichsgrenze von Glhkato-
den-Ionisationsvakuummetern
1995, 47-48
F. Mertens et al.
Einflu von Gasadsorbaten auf die Eigen-
schaften eines Glhkatoden-Ionisations-
vakuummeters mit axialer Emission nach
Chen und Suen
6. Pressure
monitoring, control
and regulation
K. G. Mller
Betriebsberwachung, Steuerung und
Automatisierung von Vakuumanlagen
Chemie-Ingenieur-Technik, 35,
1963, 73-77
G. Kienel
Elektrische Schaltgerte der Vakuumtech-
nik
Elektro-Technik, 50, 1968, 5-6
A. Bolz, H. Dohmen und H.-J. Schubert
Prozedruckregelung in der
Vakuumtechnik
Leybold Firmendruckschrift 179.54.01
H. Dohmen
Vakuumdruckmessung und -Regelung in
der chemischen Verfahrenstechnik
Vakuum in der Praxis, 6,1994, 113-115
N. Pchheim
Druckregelung in Vakuumsystemen
1995, 39-46
R. Heinen und W. Schwarz
Druckregelung bei Vakuumprozessen
durch umrichtergespeiste Rootspumpen
Vakuum-Technik, 35, 1986, 231-236
7. Mass spectrometer
gas analysis at low
pressures
H. Hoch
Total- und Partialdruckmessungen bei
Drcken zwischen 2 10
10
und
2 10
2
Torr
H. Junge
Partialdruckmessung und Partialdruck-
megerte
G-I-T May 1967, 389-394 and
June 1967, 533-538
A. Kluge
Ein neues Quadrupolmassenspektrometer
mit massenunabhngiger Empfindlichkeit
Vakuum-Technik, 23, 1974, 168-171
S. Burzynski
Microprocessor controlled quadrupole
mass spectrometer
Vacuum, 32, 1982, 163-168
W. Groe Bley
Quantitative Gasanalyse mit dem
Quadrupol Massenspektrometer
A. J. B. Robertson
Mass Spectrometry
Methuen & Co, Ltd., London, 1954
C. Brunee und H. Voshage
Massenspektrometrie
Karl Thiemig Verlag, Mnchen, 1964
D00 E 19.06.2001 21:41 Uhr Seite 180
References
A. Cornu and R. Massot
Compilation of Mass Spectral Data
Heyden and Son Ltd., London, 1966
P. Dawson
Quadrupole Mass Spectroscopy
Elsevier, Amsterdam, 1976
J. Backus
Chap. 11 in Characteristics of Electrical
Discharges in Magnetic Fields
National Nuclear Energy Series, Div. I,
Vol. 5, McGraw-Hill Book Company Inc.,
New York, 1949
J. Backus
University of California Radiation
Laboratory Report, RL 20.6.36,
Mar. 1945.
8. Leaks and Leak
detection
8.1 Mass spectrometer leak
detection
G. Kienel
Lecksuche an Vakuumanlagen auf
elektrischem Wege
Elektrotechnik, 49, 1967, 592-594
U. Beeck
Mglichkeiten und Grenzen der automati-
schen Lecksuche im Bereich unter
10
8
Torr. l/s
Lecksuche an Chemieanlagen
Dechema Monographien (Ed. H. E. Bhler
and K. Steiger), Vol. 89, Verlag Chemie,
Weinheim / New York
W. Jansen
Grundlagen der Dichtheitsprfung mit
Hilfe von Testgasen
Vakuum-Technik, 29, 1980, 105-113
K. Paasche
Lecksuche an Chemieanlagen
Vakuum-Technik, 29, 1980, 227-231
H. B. Brger
Lecksuche an Chemieanlagen mit He-
Massenspektrometer-Lecksuchern
Vakuum-Technik, 29, 1980, 232-245
Chr. Falland
Ein neuer Universal-Lecksucher mit
luftgekhlter Turbo-Molekularpumpe
Vakuum-Technik, 29, 1980, 205-208
W. Jansen
Grundlagen der Dichtheitsprfung mit
Hilfe von technischen Gasen
Vakuum-Technik, 29, 1980, 105-113
H. Mennenga
Dichtheitsprfung von Kleinteilen
Vakuum-Technik, 29, 1980, 195-200
Chr. Falland
Entwicklung von He-Lecksuchtechniken
fr UHV-Systeme groer Beschleuniger-
und Speicherringe
W. Engelhardt et al.
Lecksuchanlagen in der Industrie
Vakuum-Technik, 33, 1984, 238-241
G. Snger et al.
ber die Lecksuche bei Raumfahrzeugen
W. Jitschin et al.
He-Diffusionslecks als sekundre Normale
fr den Gasdurchflu
Vakuum-Technik, 36, 1987, 230-233
W. Groe Bley
Moderne He-Leckdetektoren
unterschiedlicher Prinzipien im prakti-
schen Einsatz
H. D. Brger
Lecksucher (with references to relevant
literature)
W. Fuhrmann
Einfhrung in die industrielle Dichtheits-
prftechnik
W. Fuhrmann
Industrielle Dichtheitsprfung ohne
Testgas nach dem Massenspektrometrie-
verfahren
1995, 179 -182
8.2 Leak detection with
halogen leak detectors
H. Moesta und P. Schuff
ber den thermionischen Halogendetektor
Berichte der Bunsengesellschaft fr
physikaische Chemie,
Bd. 69, 895-900, 1965
Verlag Chemie, GmbH, Weinheim,
Bergstrae
J. C. Leh and Chih-shun Lu
US Patent Nr. 3,751,968
Solid State Sensor
9 Film thickness
measurement and
control
G. Z. Sauerbrey
Phys. Verhandl. 8, 113, 1957
G. Z. Sauerbrey
Verwendung von Schwingquarzen zur
Wgung dnner Schichten und zur
Mikrowgung
Zeitschrift fr Physik 155, 206-222, 1959
L. Holland, L. Laurenson and J. P. Deville
Use of a Quartz Crystal Vibrator in Vacu-
um Destillation Invstigations
Nature, 206 (4987), 883-885, 1965
R. Bechmann
ber die Temperaturabhngigkeit der
Frequenz von AT- und
BT-Quarzresonatoren
Archiv fr Elektronik und bertragungs-
technik, Bd. 9, 513-518, 1955
K. H. Behrndt and R. W. Love
Automatic control of Film Deposition Rate
with the crystal oscillator for preparation
of alloy films.
Vacuum 12 ,1-9, 1962
P. Lostis
Automatic Control of Film Deposition Rate
with the Crystal Oscillator for Preparation
fo Alloy Films.
Rev. Opt. 38, 1 (1959)
D00
D00 E 19.06.2001 21:41 Uhr Seite 181
References
K. H. Behrndt
Longterm operation of crystal oscillators
in thin film deposition
J. Vac. Sci. Technol. 8, 622 (1971)
L. Wimmer, S. Hertl, J. Hemetsberger and
E. Benes
New method of measuring vibration
amplitudes of quartz crystals.
Rev. Sci. Instruments 55 (4) , 608, 1984
P. J. Cumpson and M. P. Seah
Meas. Sci. Technol., 1, 548, 1990
J. G. Miller and D. I. Bolef
Sensitivity Enhancement by the use of
Acoustic Resonators in cw Ultrasonic
Spectroscopy.
J. Appl. Phys. 39, 4589, (1968)
J. G. Miller and D. I. Bolef
Acoustic Wave Analysis of the Operation
of Quartz Crystal Film Thickness Moni-
tors.
J. Appl. Phys. 39, 5815, (1968)
C. Lu and O. Lewis
Investigation of Film thickness
determination by oscillating quartz
resonators with large mass load.
J. Appl. Phys. 43, 4385 (1972)
C. Lu
Mas determination with piezoelectric
quartz crystal resonators.
J. Vac. Sci. Technol. Vol. 12 (1),
581-582, 1975
A. Wajid
U.S. Patent No. 505,112,642
(May 12, 1992)
C. Hurd
U.S. Patent No. 5,117,192 (May 26,
1992)
E. Benes
Improved Qartz Crystal Microbalance
Technique
J. Appl. Phys. 56, (3), 608-626 (1984)
C. J. Wilson
Vibration modes of AT-cut convex quartz
resonators.
J. Phys. d 7, 2449, (1974)
H. F. Tiersten and R. C. Smythe
An analysis of contowced crystal
resonators operating in overtones of
coupled thickness shear and thickness
twist.
J. Acoustic Soc. Am. 65, (6) 1455, 1979
R. E. Bennett, C. Rutkoeski and
L. A. Taylor
Proceedings of the Thirteenth Annual
Symposium on Frequency Controll,
479, 1959
Chih-shun Lu
Improving the accuracy of Quartz csystal
monitors
Research/Development, Vol. 25, 45-50,
1974, Technical Publishing Company
A. Wajid
Improving the accuracy of a quartz crystal
microbalance with automatic determinati-
on of acoustic impedance ratio.
Rev. Sci. Instruments, Vol. 62 (8), 2026-
2033, 1991
D. Graham and R. C. Lanthrop
The Synthesis fo Optimum Transient
Response: Criteria and Standard Forms
Transactions IEEE, Vol. 72 pt. II,
Nov. 1953
A. M. Lopez, J. A. Miller, C. L. Smith and
P. W. Murrill
Tuning Controllers with Error-Integral
Criteria
Instrumentation Technology, Nov. 1969
C. L. Smith and P. W. Murril
A More Precise Method for Tuning
Controllers
ISA Journal, May 1966
G. H. Cohen and G. A. Coon
Theoretical considerations of Retarded
Control
Taylor Technical Data Sheet Taylor
Instrument Companies, Rochester,
New York
J. G. Ziegler and N. B. Nichols
Optimum Settings for Automatic
Controllers
Taylor Technical Data Sheet No. TDS
10A100, Taylor Instrument Companies,
Rochester, New York
C. Lu and A. W. Czanderna
Application of Piezoelectric Quarz Crystal
Microbalances (Vol.7 of: Methodes and
Phenomena, Their Applications in Sience
and Technology)
Elesvier, Amsterdam, Oxford, New York,
Tokio, 1984
G. Simmons and H. Wang
Single Crystal Elastic Constants and
Calculated Aggregate Properties
A Handbook
The MIT Press, Cambridge, Massachu-
setts, 1971
C. D. Stockbridge
in Vol. 5 Vacuum Microbalance
Techniques K. Behrndt, editor, Plenum
Press, Inc., New York, 1966
S. Sotier
Schwingquarz-Schichtdickenmessung
Vakuum in der Praxis 1992, 182-188
10. Materials and
material
processing
W. Espe
Werkstoffkunde der Hochvakuumtechnik
Vol. 1 1959, Vol. 2 1960, Vol. 3 1961,
VEB Deutscher Verlag der Wissenschaf-
ten, Berlin
W. Espe
Werkstoffe fr trennbare metallische Ver-
bindungen der Ultrahochvakuumtechnik
Feinwerktechnik, 68, 1964, 131-140
W. Espe
Synthetische Zeolithe und ihre Verwen-
dung in der Hochvakuumtechnik
Experimentelle Technik der Physik, XII,
1964, 293-308
H. Adam
Allgemeiner berblick ber die Werkstoffe
der Vakuumtechnik und deren Auswahl
Haus der Technik Vortragsverffentlichun-
gen Werkstoffe und Werkstoffverbindun-
gen in der Vakuumtechnik H. 172, Vul-
kan-Verlag, Dr. W. Classen, Essen, 1968,
4 13
D00 E 19.06.2001 21:41 Uhr Seite 182
References
K. Verfu
Bessere Oberflchenvergtung durch
Elektropolieren am Beispiel der Vaku-
um-Technik
VDI-Berichte, 183, 1972, 29-34
K. Verfu
Schweien und Hartlten
Haus der Technik, Vortragsverffentli-
chungen Werkstoffe und Werkstoffver-
bindungen in der Vakuumtechnik, H. 172
Vulkan-Verlag Dr. W. Classen, Essen,
1968, Seiten 39 -49
Chr. Edelmann
Gasabgabe von Festkrpern im Vakuum
Vakuum-Technik, 38, 1989, 223-243
R. Fritsch
Besonderheiten vakuumdichrter
Schweiverbindungen
H. Henning
Vakuumgerechte Werkstoffe und
Verbindungstechnik, Part 1
R. Fritsch
M. Mhlloff
11. Dictionaries
F. Weber
Elseviers Dictionary of High Vacuum
Science and Technology (German,
English, French, Spanish, Italian, Russian)
Elsevier Verlag 1968
Hurrle / Jablonski / Roth
Technical Dictionary of Vacuum Physics
and Vacuum Technology (German,
English, French, Russian)
Pergamon Press Verlag, Oxford, 1972
D00
D00 E 19.06.2001 21:41 Uhr Seite 183
Index
13. Index
Absolute pressure . . . . . . . . . . . . . . . .7
Absorption isotherms . . . . . . . . . . . . .45
Absorption pumps . . . . . . . . . . .45, 139
Absorption traps . . . . . . . . . . . . . . . .33
Accessories for rotary
displacement pumps . . . . . . . . . . . . .33
Active oscillator . . . . . . . . . . . .122, 123
Adjustment and calibration of
vacuum gauges . . . . . . . . . . . . . . . . .88
Adsorption pumps,
Instructions for operation . . . . .139, 140
Aggressive vapors . . . . . . . . . . . . . .135
AGM (aggressive gas monitor) . . . . .94
Air, atmospheric . . . . . . . . . . . . . . . . .10
ALLex pumps . . . . . . . . . . . . . . .28, 30
Ambient pressure . . . . . . . . . . . . . . . . .7
Amonton's law . . . . . . . . . . . . . . . . . .11
Anticreep barrier . . . . . . . . . . . . . . . .40
Anti-suckback valve . . . . . . . . . . . . . .19
APIEZON AP 201 . . . . . . . . . . . .39, 160
Atmospheric air . . . . . . . . . . . . . . . . .10
Atmospheric air, composition . . . . . .145
Atmospheric pressure . . . . . . . . . . . . .7
Atomic units . . . . . . . . . . . . . . . . . .170
Autoc ontrol tune . . . . . . . . . .125, 126
Automatic protection, monitoring
and control of vacuum systems . . . . .83
Auto-Z match technique . . . . . . . . . .122
Avogadro's law . . . . . . . . . . . . . . . . .11
Avogadro's number (Loschmidt
number) . . . . . . . . . . . . . . . . . . .12, 143
Backing line vessel . . . . . . . . . . . . . . .64
Baffle . . . . . . . . . . . . . . . . . . . . . . . . .36
Baffles (vapor barriers) . . . . . .36, 37, 39
Baking (degassing) . . . . . . . .61, 68, 141
Barrier gas operation . . . . . . . . . . . . .44
Basis SI units . . . . . . . . . . . . . . . . . .170
Bath cryostats . . . . . . . . . . . . . . . . . .49
Bayard-Alpert gauge . . . . . . . . . . . . . .80
Boat (thermal evaporator) . . . . . . . . .128
Boltzmann constant . . . . . . . . . .11, 143
Bombing-Test (storing under
pressure) . . . . . . . . . . . . . . . . . . . . .119
Booster (oil-jet) pump . . . . . . . . .38, 46
Bourdon vacuum gauge . . . . . . . . . . .72
Boyle-Mariotte law . . . . . . . . . . . . . . .11
Break down voltage
(Paschen curve for air) . . . . . . . . . . .163
Bubble (immersion) test . . . . . . . . .109
Calibration curves of
THERMOVAC gauges . . . . . . . . . . . . .77
Calibration inspection . . . . . . . . . . . . .81
Coating sources . . . . . . . . . . . . . . . .128
Capacitance diaphragm gauges . . . . .73
Capsule vacuum gauge . . . . . . . . . . .772
Causes of faults if/when the desired
ultimate pressure
is not achieved . . . . . . . . . . . . . . . . .134
Ceramic ball bearings (hybrid ball
bearings) . . . . . . . . . . . . . . . . . . . . . .42
CF-flange (conflathflange) . . . . . . . . .68
Changing the molecular sieve . . . . . .139
Charles' law (Gay-Lussac's law) . . . . .11
Chemical resistance of elastomer
gaskets . . . . . . . . . . . . . . .149, 150, 151
Chemical vapor deposition (CVD) . . .129
Choked flow, critical pressure
difference . . . . . . . . . . . . . . . . . . . . . .12
CIS (closed ion source) . . . . . . . . . . .93
Clamp flange . . . . . . . . . . . . . . . . . . .67
Classification of vacuum pumps . . . . .16
Clausius-Clapeyron equation . . . . . . .11
Claw pump . . . . . . . . . . . . . . . . . . . .27
Closed ion source, (CIS) . . . . . . . . . . .93
Coating of parts . . . . . . . . . . . . . . . .130
Coating thickness regulation . . . . . . .125
Cold cap baffle . . . . . . . . . . . . . . . . . .39
Cold cathode ionization vacuum
gauge . . . . . . . . . . . . . . . . . . . . . . . .77
Cold head . . . . . . . . . . . . . . . . . . . . .50
Cold surfaces, bonding of
gases to . . . . . . . . . . . . . . . . . . . . . . .51
Cold traps . . . . . . . . . . . . . . . . . . . . .39
Collision frequency . . . . . . . . . . . . . . .10
Collision rate . . . . . . . . . . . . . . . . . . .10
Common solvents . . . . . . . . . . . . . .146
Compression . . . . . . . . . . . . . . . . . . .43
Compression vacuum gauges . . . . . . .74
Condensate traps . . . . . . . . . . . . . . . .33
Condensers . . . . . . . . . . . . . . . .33, 174
Conductance . . . . . . . . . . . . . . . . .9, 13
Conductance of openings . . . . . . . . .179
Conductance of
piping . . . . . . . . . . . . .14, 155, 161, 179
Conductance, nomographic
determination . . . . . . . . . . . . . . . . . . .14
Conductances, calculation of . . . . . . .13
Connection of leak detectors to
vacuum systems . . . . . . . . . . . . . . .116
Contamination of vacuum sensors . .139
Contamination of vacuum vessels . . .134
Continous flow . . . . . . . . . . . . . . . . . .12
Continous flow cryopumps . . . . . . . . .49
Continuum theory . . . . . . . . . . . . . . .11
Conversion of leak rate units . . . . . .106
Conversion of leak rate units . . . . . .144
Conversion of pressure units . . . . . .142
Conversion of pV-throughput units . .144
Corrosion protection . . . . . . . . .135, 136
Counter-flow leak detector . . . . . . . .115
Cracking pattern . . . . . . . . . . . . . . .136
Critical pressure difference
(choked flow) . . . . . . . . . . . . . . . . . . .12
Crossover value . . . . . . . . . . . . . . . . .53
Cryocondensation . . . . . . . . . . . . . . .52
Cryopumps . . . . . . . . . . . . . . . .49, 177
Cryosorption . . . . . . . . . . . . . . . . . . .52
Cryotrapping . . . . . . . . . . . . . . . . . . .52
Crystal Six . . . . . . . . . . . . . . . . . . .120
Cut in (start) pressure . . . . . .44, 54, 55
CVD (chemical vapor
deposition) . . . . . . . . . . . . . . . .128, 129
Dalton's law . . . . . . . . . . . . . . . . . . . .11
Danger classes of fluids . . . . . . . . . .148
Data storage coating . . . . . . . . . . . .132
DC 704, DC 705 (Silicone oils) . . . . . .39
Degassing of the pump oil . . . . . . . . .37
Derived coherent and not coherent
SI units with special
names and symbols . . . . . . . . . . . . .170
Detection limit (leak detectors) . . . . .111
Determination of a suitable
backing pump . . . . . . . . . . . . . . . . . .64
Determination of pump down
time from Nomograms . . . . . . . . . . . .65
Determination of pump sizes . . . . . . .62
DI series diffusion pumps . . . . . . . . .38
Diaphragm contoller, examples
of application . . . . . . . . . . . . . . . .86, 87
Diaphragm vacuum gauges . . . . . . . .72
Diaphragm vacuum pumps . . . . . . . . .17
DIAVAC diaphragm vacuum gauge . . .72
DIFFELEN, light, normal, ultra . . .39, 160
Diffusion / vapor-jet pumps,
Instructions for operation . . . . . . . . .139
Diffusion pumps . . . . . . . . . . . . . . . .36
Diode-type sputter ion pumps . . . . . .46
Direct-flow leak detector . . . . . . . . . .115
Discharge filters . . . . . . . . . . . . . . . . .33
Displacement pumps . . . . . . . . .16, 174
DIVAC vacuum pump . . . . . . . . . . . . .17
DKD (Deutscher Kalibrierdienst)
German calibration service . . . . . . . . .81
Dry compressing rotary
displacement pumps . . . . . . . . . . . . .24
Dry processes . . . . . . . . . . . . . . . . . .57
Drying of paper . . . . . . . . . . . . . . . . .67
Drying processes . . . . . . . . . . . . . . . .60
Drying processes, selection of
pumps for . . . . . . . . . . . . . . . . . . . . .66
DRYVAC-Pumps . . . . . . . . . . . . . . . .28
D-Tek . . . . . . . . . . . . . . . . . . . . . . .110
Duo seal (sealing passage) . . .17, 18, 19
Dust separator (dust filter) . . . . . . . . .33
Dynamic expansion method . . . . . . . .82
Ecotec II . . . . . . . . . . . . . . . . . . . . .113
Effective pumping speed . . . . . . . .33, 62
Elastomer gaskets . . . .69, 149, 150, 151
Electrical break down voltage
(Paschen curve air) . . . . . . . . . . . . .163
Electron beam evaporators
(electron guns) . . . . . . . . . . . . . . . .129
Envelope test . . . . . . . . . . . . . .118, 119
D00 E 19.06.2001 21:41 Uhr Seite 184
Index
Envelope test (concentration
measurement) . . . . . . . . . . . . . . . . .118
Evacuation in the rough / medium /
high vacuum region . . . . . . . .62, 63, 64
Evacuation of gases / vapors . . . . . . .66
Evaluating spectra . . . . . . . . . . . . . . .96
Expansion method static /
dynamic . . . . . . . . . . . . . . . . . . . .81, 82
Extractor ionization vacuum
gauge . . . . . . . . . . . . . . . . . . . . . . . .80
Fast regeneration (partial
regeneration) . . . . . . . . . . . . . . . . . . .54
Fingerprint . . . . . . . . . . . . . . . . . . . . .96
Flanges and their seals . . . . . . . .67, 179
Floating zero-point suppression . . . .111
Fluid entrainment pumps . . . . . .35, 177
Foam spray leak test . . . . . . . . . . . .109
Fractionation of pump fluids . . . . . . . .37
Fragment distribution pattern . . . . . . .98
Fundamental pressure
measurement methods . . . . . . . . . . . .81
Gas analysis . . . . . . . .89, 100, 101, 180
Gas ballast . . . . . . . . . . . . . .21, 22, 111
Gas composition as a function
of altitude . . . . . . . . . . . . . . . . . . . .155
Gas constant, general (molar) .7, 11, 143
Gas density . . . . . . . . . . . . . . . . . . . . .7
Gas dependent pressure reading,
vacuum gauges with . . . . . . . . . . . . .75
Gas discharge . . . . . . . . . . . . . . .46, 78
Gas independent pressure reading,
vacuum gauges with . . . . . . . . . . . . .72
Gas laws . . . . . . . . . . . . . . . . . . . . . .11
Gas locks . . . . . . . . . . . . . . . . . . . . . .69
Gas sorption (pumping) of
vacuum gauges . . . . . . . . . . . . . .78, 79
Gas storage in the oil of rotary
vane pumps . . . . . . . . . . . . . . . . . . .111
Gaskets . . . . . . . . . . .68, 149, 150, 151
Gay-Lussac's law . . . . . . . . . . . . . . . .11
General gas constant
(Molar gas constant) . . . . . . .7, 11, 143
Getter pumps . . . . . . . . . . . . . . . . . . .45
Glass coating . . . . . . . . . . . . . . . . . .132
Halogen leak detector . . . . . . . .110, 181
Helium leak detectors with 180
sector mass spectrometer . . . . . . . .114
Helium spray equipment . . . . . . . . . .118
Helium standard leak rate . . . . . . . .106
High frequency vacuum test . . . . . . .109
High pressure ionization vacuum
gauge . . . . . . . . . . . . . . . . . . . . . . . .80
High vacuum range . . . . . . . . . . .62, 63
HLD 4000 . . . . . . . . . . . . . . . . . . . .110
HO-factor (diffusion pumps) . . . . . . .37
Hot cathode ionization vacuum
gauge . . . . . . . . . . . . . . . . . . . . . . . .78
HY.CONE pumps . . . . . . . . . . . . . . . .44
Hybrid ball bearings (Ceramic ball
bearings) . . . . . . . . . . . . . . . . . . . . .42
Hydrocarbon-free vacuum . . . . . .39, 60
IC 5 . . . . . . . . . . . . . . . . . . . . . . . .127
Ideal gas law . . . . . . . . . . . . . . . . . 7, 11
Impingement rate . . . . . . . . . . . . . . . .10
Industrial leak testing . . . . . . . . . . . .119
Influence of magnetic /
electrical fields . . . . . . . . . . . . . . . . .141
Internal compression (claw pumps) . .28
Inside-out leak . . . . . . . . . . . . . . . . .107
Integral leak rate . . . . . . . . . . . . . . .107
Internal reflow (roots pumps) . . . . . . .24
Ion desorption effect . . . . . . . . . . . . .79
Ion sputter pumps . . . . . . . . . . . .45, 46
Ionization vacuum gauge for
higher pressures up to 1 mbar . . . . . .80
Ionization vacuum gauges . . . . . . . . .77
Ionization, specific (gas analysis) . . . .96
Isotopes . . . . . . . . . . . . . . . . . . . . . .96
Kammerer compression vacuum
gauge . . . . . . . . . . . . . . . . . . . . . . . .74
Kinetic gas theory . . . . . . . . . . . . . . .11
Kinetic of gases, diagram of . . . . . . .154
Kinetic of gases, formulas . . . . . . . .143
Knudsen flow . . . . . . . . . . . . . . . . . . .13
Krypton 85 test . . . . . . . . . . . . . . . .109
Laminar flow . . . . . . . . . . . . . . . . . . .12
Langmuir-Taylor-effect . . . . . . . . . .111
Laval nozzle . . . . . . . . . . . . . . . . . . . .38
Leak detection . . . . . . . . . . . . .104, 181
Leak detection using Helium
leak detectors . . . . . . . . . . . . . . . . .117
Leak detection without leak detector .107
Leak detection, leak test . . . . . . . . . .104
Leak detectors with 180 sector
mass spectrometer . . . . . . . . . . . . .114
Leak detectors with mass
spectrometer . . . . . . . . . . . . . .110, 181
Leak detectors with quadrupole
mass spectrometer . . . . . . . . . . . . .113
Leak detectors, how they work . . . . .110
Leak rate, hole size,
conversion . . . . . . . . . .9, 104, 105, 106
Leak test (chemical reactions, dye
penetration) . . . . . . . . . . . . . . . . . . .110
Leak test, using vacuum gauges
sensitive to the type of gas . . . . . . . .108
LEYBODIFF-Pumps . . . . . . . . . . . . . .37
INFICON Quartz crystal
controllers . . . . . . . . . . . . . . . . . . . .127
Line width . . . . . . . . . . . . . . . . . . . . .94
Linearity range of quadrupole
gauges . . . . . . . . . . . . . . . . . . . . . . .102
Liquid ring pumps . . . . . . . . . . . . . . .17
Liquid sealed rotary displacement
pumps . . . . . . . . . . . . . . . . . . . . . . . .17
Liquid-filled vacuum gauges . . . . . . . .74
Literature references . . . . . . . .174 183
LN
2
cold traps . . . . . . . . . . . . . . . . . .40
Local leak rate . . . . . . . . . . . . . . . . .106
Loschmidt's number
(Avogadro constant) . . . . . . . . . .11, 143
Magnetic suspension (bearings) . . . . .42
Mass flow . . . . . . . . . . . . . . . . . .8, 104
Mass flow (leak detection) . . . . . . . .104
Mass range . . . . . . . . . . . . . . . . . . .195
Mass spectrometer, general,
historical . . . . . . . . . . . . . . . . . .89, 180
Maximum backing pressure
(critical forevacuum pressure) . . . . . .37
McLeod vacuum gauge . . . . . . . . . . .74
Mean free path . . . . . . . . . . .9, 142, 159
Measuring range of vacuum
gauges . . . . . . . . . . . . . . . . . . . . . . .161
Measuring range, favorable . . . . . . . .70
Measuring ranges of vacuum
gauges . . . . . . . . . . . . . . . . . . . . . . .162
Measuring vacuum, vacuum
gauges . . . . . . . . . . . . . . . . . . . .70, 179
Medium vacuum adsorption trap . . . .33
MEMBRANOVAC . . . . . . . . . . . . . . . .73
Mercury (pump fluid) . . . . . . .36, 39,160
Mode-lock oscillator . . . . . . . . . . . .123
Molar gas constant . . . . . . . . .7, 11, 143
Molar mass
(molecular weight) . . . . . . . . . .7, 10, 11
Molecular flow . . . . . . . . . . . . . . .12, 13
Molecular sieve . . . . . . . . . . . . .45, 139
Monolayer . . . . . . . . . . . . . . . . . . . . .10
Monolayer formation time . . . .10, 13, 61
National standards, resetting to . . . . .81
NEG pumps (non evaporable
getter pumps) . . . . . . . . . . . . . . .45, 48
Neoprene . . . . . . . . . .68, 149, 150, 151
Nitrogen equivalent . . . . . . . . . . .71, 78
Nominal internal diameter and
internal diameter of tubes . . . . . . . . .146
Nomogram . . . . . . . . . . . . . . . . . . . .65
Nomogram: conductance of
tubes / entire pressure range . . . . . .158
tubes / laminar flow range . . . . . . . .155
tubes / molecular flow range . . .155, 162
Nomogram: pump down time /
medium vacuum, taking in
account the outgasing from the
walls . . . . . . . . . . . . . . . . . . . . . . . .159
Nomogram: pump down time /
rough vacuum . . . . . . . . . . . . . . . . .156
Non evaporable getter (NEG)
pumps . . . . . . . . . . . . . . . . . . . . .45, 48
Non gas-tight area . . . . . . . . . . . . .1104
Nude gauge (nude system) . . . . . . . . .72
Oil backstreaming . . . . . . . . . . . .39, 178
Oil change . . . . . . . . . . . . . . . . . . . .137
Oil consumption . . . .135, 136, 137, 138
D00
D00 E 19.06.2001 21:41 Uhr Seite 185
Index
Oil contamination . . . . . . . . . . . . . . .135
Oil diffusion pumps . . . . . . . . . . . . . .36
Oil sealed rotary displacement
pumps . . . . . . . . . . . . . . . . . . . . . . . .18
Oil vapor ejector vacuum pumps . . . .36
Oil-free (hydrocarbon-free)
vacuum . . . . . . . . . . . . . . . . . . . .39, 60
Oils (pump fluids) . . . . . . . . . . . . . . .39
Open (normal) ion source . . . . . . . . . .90
Optical coatings . . . . . . . . . . . . . . . .131
Oscillation displacement pumps . . . . .17
Oscillator, ( active, mode-lock) .122, 123
Outgasing of materials . . . . . . . . . . .145
Outgasing rate (referred to
surface area) . . . . . . . . . . . . . . . . .9, 61
Outside-in leak . . . . . . . . . . . . . . . . .107
Overpressure . . . . . . . . . . . . . . . . . . . .7
Oxide-coated cathodes . . . . . . . . .78, 90
Partial final pressure . . . . . . . . . . . . .74
Partial flow opeartion . . . . . . . . . . . .115
Partial flow ratio . . . . . . . . . . . . . . . .116
Partial pressure . . . . . . . . . . . . . . . . . .7
Partial pressure measurement . . . . .100
Partial pressure regulation . . . . . . . .101
Particle number density . . . . . . . . . . . .7
Paschen curve . . . . . . . . . . . . . . . . .163
Penning vacuum gauges . . . . . . . . . .77
Perbunan . . . . . . . . .68, 149 - 151, 161
Period measurement . . . . . . . . . . . .122
Permissible pressure units . . . . . . . .142
Phase diagram of water . . . . . . . . . .164
Photons . . . . . . . . . . . . . . . . . . . . . . .79
PIEZOVAC . . . . . . . . . . . . . . . . . . . . .73
Pirani vacuum gauge . . . . . . . . . . . . .76
Plastic tent (envelope) . . . . . . . . . . .118
Plate baffle . . . . . . . . . . . . . . . . . . . . .39
PNEUROP . . . . . . . . . . . . . . .165 168
PNEUROP flanges . . . . . . . . . . . . . . .68
Poiseuille flow . . . . . . . . . . . . . . . . . .12
Poisson's law . . . . . . . . . . . . . . . . . .11
Positive pressure methode
(leak detection) . . . . . . . . . . . . . . . .106
Pre-admission cooling (roots
pumps) . . . . . . . . . . . . . . . . . . . . . . .27
Precision diaphragm vacuum gauge . .72
Pressure . . . . . . . . . . . . . . . . . . .7, 168
Pressure and temperature as
function of altitude . . . . . . . . . . . . . .155
Pressure converter . . . . . . . . . . . . . . .92
Pressure dependence of the
mean free path . . . . . . . . . . . . .145, 154
Pressure difference oil supply . . . . . . .19
Pressure lubrication by geared
oil pump . . . . . . . . . . . . . . . . . . . . . .19
Pressure measurement
direct / indirect . . . . . . . . . .71, 179, 180
Pressure measurement, depending on /
independent of the type of gas . . . . . .71
Pressure ranges in vacuum
technology . . . . . . . . . . .12, 56, 57, 145
Pressure regulation / control . . .84, 180
Pressure regulation / control rough
and medium vacuum systems . . . . . .84
Pressure regulation in high and
ultra high vacuum systems . . . . . . . .87
Pressure regulation, continuous /
discontinuous . . . . . . . . . . . . . . .84, 86
Pressure rise / drop (leak) test .107, 108
Pressure units . . . . . . . . . . . . . . .7, 142
PTB (Federal physical-technical
institute) . . . . . . . . . . . . . . . . . . . . . .81
Pumpdown time . . . . . . . . . . . . .62 - 66
Pump fluid . . . . . . . . . . . . . . . . . . . . .39
Pump fluid backstreaming . . . . . . . . .39
Pump fluid change cleaning
(diffusion pumps) . . . . . . . . . . . . . .139
Pump oil, selection when handling
aggressive vapors . . . . . . . . . . . . . .135
Pump throughput . . . . . . . . . . . . . . . . .9
Pumping (gas sorption) of
vacuum gauges . . . . . . . . . . . . . .77, 78
Pumping chamber . . . . . . . . . . . . . . .16
Pumping of gases . . . . . . . . . . . . . . .57
Pumping of gases and
vapors . . . . .20, 21, 34, 52, 57, 58, 135
Pumping speed . . . . . . . . . . . . . . . . . .8
Pumping speed units,
conversion of . . . . . . . . . . . . . . . . . .144
Pumping various chemical
substances . . . . . . . . . . . . . . . . . . .136
Purge gas . . . . . . . . . . . . . . . . . . . . .29
PVD (physical vapor deposition) . . .128
pV-flow . . . . . . . . . . . . . . . . . . . . . . . .9
pV-value . . . . . . . . . . . . . . . . . . . . . . .8
Quadrupole mass spectrometer . . . . .89
Quadrupole, design of the sensor . . . .90
Quadrupole, gas admission /
pressure adaptation . . . . . . . . . . . . . .93
Quadrupole, measurement system
(detector) . . . . . . . . . . . . . . . . . . . . . .92
Quadrupole, separating system . . . . .91
Quadrupole, specifications . . . . . . . . .94
Qualitative gas analysis . . . . . . . . . .100
Quantitative gas analysis . . . . . . . . .101
Quantity of gas (pV value) . . . . . . . . . .8
Quartz crystals, shape of . . . . . . . . .121
Rate watcher . . . . . . . . . . . . . . . . . .120
Reduction of adsorption capacity . . .139
Reduction ratio . . . . . . . . . . .24, 25, 137
Refrigerator cryopump . . . . . . . . .49, 51
Regeneration time . . . . . . . . . . . . . . .53
Relative ionization probability (RIP) . .98
Residual gas composition
(spectrum) . . . . . . . . . . . . . . . . . .43, 44
Response time of leak detectors . . . .117
Reynold's number . . . . . . . . . . . . . . .12
Rigid envelope . . . . . . . . . . . . .118, 119
Roots pumps . . . . . . . . . . . . . . . . . . .24
Roots pumps, Instructions for
operation . . . . . . . . . . . . . . . . . . . . .137
Rotary displacement pumps . . . . . . . .18
Rotary plunger pumps . . . . . . . . . . . .20
Rotary vane / piston pumps,
Rotary vane pumps . . . . . . . . . . . . . .18
Salt, drying of . . . . . . . . . . . . . . . . . .66
Saturation vapor pressure
(nonmetallic gaskets) . . . . . . . . . . . .161
Saturation vapor pressure . . . . . . .7, 22
(cryogenic technology) . . . . . . . . . . .161
Saturation vapor pressure (metals) . .160
(pump fluids) . . . . . . . . . . . . . . . . . .160
(solvents) . . . . . . . . . . . . . . . . . . . .160
Saturation vapor pressure and
vapor density of water . . . . . . .147, 164
Sealing passage . . . . . . . . . . . . . .18, 19
Seal-off fitting . . . . . . . . . . . . . . . . . .69
Selection of pumps . . . . . . . . . . . . . .56
Selection of pumps for drying
processes . . . . . . . . . . . . . . . . . . . . .66
Sensitivity of quadrupole sensors . . . .95
Sensitivity of vacuum gauges . . . . . . .78
Separating system of
mass spectrometers . . . . . . . . . . . . . .90
Shell baffle . . . . . . . . . . . . . . . . . . . . .39
Silicone oils, DC 704, DC 705 . .39, 160
Small flange . . . . . . . . . . . . . . . . . . . .67
Smallest detectable concentration . . . .95
Smallest detectable partial pressure . .95
Smallest detectable partial
pressure ratio . . . . . . . . . . . . . . . . . .95
Sniffer technology . . . . . . . . . . . . . .118
Software for TRANSPECTOR . . . . . .102
SOGEVAC pumps . . . . . . . . . . . . . . .18
Solvents . . . . . . . . . . . . . . . . . . . . . .146
Sorption pumps . . . . . . . . . . . . .45, 177
Specific volume of water vapor .147, 163
Spinning rotor gauge (SRG) . . . . . . .75
Spray technique (Helium) . . . . . . . . .117
Sputter ion pumps, Instructions
for operation . . . . . . . . . . . . . . . . . .140
Sputter pumps . . . . . . . . . . . . . . . . . .46
Sputtering . . . . . . . . . . . . . . . . . . . .129
Sputtering (cathode sputtering) . . . .129
Sputter-ion pumps . . . . . . . . . . . .46, 47
SRG (spinning rotor gauge),
VISCOVAC . . . . . . . . . . . . . . . . . . . . .75
Stability for noble gases (sputter ion
pumps) . . . . . . . . . . . . . . . . .46, 47, 48
Standard pressure . . . . . . . . . . . . . . . .7
Standards in
vacuum technology . . . . . . . .171 173
Static expansion method . . . . . . .81, 82
D00 E 19.06.2001 21:41 Uhr Seite 186
Index
Steam ejector pumps . . . . . . . . . . . . .40
Storage under pressure
(bombing test) . . . . . . . . . . . . . . . . .119
Stray magnetic field . . . . . . . . . . . . . .47
Stray magnetic field
(sputter ion pumps) . . . . . . . . . . . . . .47
Sublimation pumps . . . . . . . . . . .46, 47
Symbols and units,
alphabetical list . . . . . . . . . . .165 170
Symbols used in vacuum
technology . . . . . . . . . . . . . . .1572, 153
Temperature comparison and
conversion table . . . . . . . . . . . . . . . .154
Temperature in the atmosphere . . . .155
Terms and definitions
(leak detection) . . . . . . . . . . . . . . . .106
Test gas accumulation . . . . . . . . . . .119
Test leaks . . . . . . . . . . . . . . . . . . . . .113
Thermal conductivity vacuum gauge,
constant / variable resistance . . . . . . .76
Thermal conductivity vacuum
gauges . . . . . . . . . . . . . . . . . . . . . . . .76
Thermal evaporator (boat) . . . . . . . .128
THERMOVAC . . . . . . . . . . . . . . . . . .75
Thickness control with quartz
oscillators . . . . . . . . . . . . . . . . . . . .120
Thickness measurement . . . . . . . . . .120
Thin film controllers . . . . .120, 127, 181
Throtteling of pumping speed
when using condensers . . . . . . . .34, 35
Time constant . . . . . . . . . . . . . .62, 116
Titanium sublimation pump . . . . . . . .46
Titanium sublimation pumps,
Torr and its conversion . . . . . . . . . . .142
Total pressure . . . . . . . . . . . . . . . . . . .7
Transfer standard . . . . . . . . . . . . . . . .81
Transmitter . . . . . . . . . . . . . . . . . . . .75
TRANSPECTOR . . . . . . . . . . . . . . . . .89
Triode sputter ion pumps . . . . . . . . . .47
TRIVAC pumps . . . . . . . . . . . . . . . . .19
Trochoid pumps . . . . . . . . . . . . . . . .21
Tuning / adjustment and
calibration of leak detectors . . . . . . .112
Turbomolecular pumps . . . . . . . .41, 176
Turbomolecular pumps,
TURBOVAC pumps . . . . . . . . . . . . . .43
Turbulent flow . . . . . . . . . . . . . . . . . .12
Types of leak . . . . . . . . . . . . . . . . . .104
Types pV flow . . . . . . . . . . . . . . . . . .12
UL 200, UL 500 . . . . . . . . . . . . . . .114
Ultimate pressure . . . . . . . . . . . . . . . . .7
Ultra high vacuum . .11, 13, 60, 61, 178
ULTRALEN . . . . . . . . . . . . . . . .39, 161
Units, symbols . . . . . . . . . . . .165 170
U-tube vacuum gauge . . . . . . . . . . . .74
Vacumm meters, instructions on
installing . . . . . . . . . . . . . . . . . . . . .140
Vacuum coating techniques . . . . . . .128
Vacuum control . . . . . . . . . . . . . . . . .70
Vacuum equipment, Instructions for
operation . . . . . . . . . . . . . . . . . . . . .134
Vacuum gauge contant . . . . . . . . . . . .78
Vacuum method (leak detection) . . .107
Vacuum physics . . . . . . . . . . . . . . . . . .7
Vacuum pumps, literature
references . . . . . . . . . . . . . . . . . . . .174
Vacuum pumps, survey,
classification . . . . . . . . . . . . . . . .16, 17
Vacuum ranges (Pressure ranges)
. . . . . . . . . . . .13, 56, 57, 145, 161, 162
Vacuum regulation . . . . . . . . . . . . . . .70
Vacuum symbols . . . . . . . . . . .152, 153
Vacuum coating technology . . . . . . .128
Values of important physical
constants . . . . . . . . . . . . . . . . . . . . .143
Valves . . . . . . . . . . . . . . . . . . . .67, 178
Van der Waals' equation . . . . . . . . . . .11
Vapor density of water . . . . . . .147, 164
Vapor pressure . . . .7, 39, 160, 161, 164
Vapor-jet pumps . . . .38, 39, 40, 41, 139
Venturi nozzle . . . . . . . . . . . . . . . . . .38
Viscous (continuum) flow . . . . . . . . .12
VISCOVAC vacuum gauge . . . . . . . .775
Vitilan, Viton . . . . . . . . . . .68, 149, 161
Volume . . . . . . . . . . . . . . . . . . . . . . . .8
Volumetric efficiency (roots pumps) . .24
Volumetric flow . . . . . . . . . . . . . . . . . .8
Water jet pumps . . . . . . . . . . . . . . . .40
Water ring pumps . . . . . . . . . . . . . . .18
Water vapor tolerance . . . . . . . . . . . .23
Web coating . . . . . . . . . . . . . . . . . . .130
Wet processes . . . . . . . . . . . . . . . . . .58
Working pressure . . . . . . . . . . . . . . . .7
Working ranges of vacuum pumps . .161
X-ray effect . . . . . . . . . . . . . . . . . . . .78
XTC, XTM . . . . . . . . . . . . . . . . . . . .127
Zeolith . . . . . . . . . . . . . . . . . . . . . . .45
Z-Match technique . . . . . . . . . . . . .122
D00
D00 E 19.06.2001 21:41 Uhr Seite 187
D00 E 19.06.2001 21:41 Uhr Seite 188

Fundamentals PDF

Uploaded by

Copyright:

Available Formats

Fundamentals PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fundamentals PDF

Uploaded by

Copyright:

Available Formats

Fundamentals of

LEYBOLD VACUUM PRODUCTS AND REFERENCE BOOK 2001/2002

Water vapor tolerance

D00 E 19.06.2001 21:36 Uhr Seite 36

D00 E 19.06.2001 21:38 Uhr Seite 62

pressure. The relationship between the

3.3 Vacuum gauges

Fig. 3.12 Cross-section of PENNINGVAC PR 35 gauge

are negligible. The pressure corre-

Resistance of the inner surface

Separation system output

, to the specific ioni-

Fig. 4.17 Partial shares for overall time constants

where R = 83.14 mbar l/mol K, T = tem-

D00 E 19.06.2001 21:39 Uhr Seite 104

330 cm = 25.95 10 = 25.95

Fig. 5.1 Correlation between leak rate and hole size

Familiar leaks: Quantity escaping: Standard He leak rate:

tus) the test gas which has passed through

Auxiliary pump Auxiliary pump

= S, is attached to the test

Display Leak rate

Fig. 6.3 Shape of INFICON quartz crystals

D00 E 19.06.2001 21:40 Uhr Seite 122

corrosion protection oil

D00 E 19.06.2001 21:40 Uhr Seite 135

Mean square of velocity

Gas pressure p of particles

1 = rad: 1 = 1/60; 1 = 1/60.

Atomic Mass for

You might also like