8.

33 Moisture in Gases and Liquids TO RECEIVER

A. BRODGESELL (1969, 1982) B. G. LIPTÁK % OR PPM

(1995)
MT H2O
J. F. TATERA (2003)
Flow Sheet Symbol

Types of Designs: A. Electrolytic hygrometer

B. Capacitance
C. Impedance
D. Piezoelectric
E. Heat of adsorption
F. Infrared (IR)
G. Microwave
H. Karl Fischer titrator (discussed under laboratory)
I. Drying oven (laboratory)
J. Dipole
K. Cavity Ring Down
L. Neutron
M. Calibrators

Ranges: A through G, see Table 8.33a

H. 10 ppm to 100%
I. Usually in %
J. 0.1 to 100%
K. ppt to ppm levels
L. 0 to 60%

Inaccuracy: A. 2 to 5% FS
B and C. 3%
D. 10% of actual reading or 2 ppm by volume, whichever is greater
F. 2% FS
G. For a 1 to 15% moisture range, error is within 0.5%
H and I. 0.5 to 1%

Costs: (For moisture in air and relative humidity sensors, also refer to Section 8.32.)
A. $6,000 to $15,000 with sample system
B and C. $2,000 to $10,000 for thin-film probe, $2,000 to $20,000 for flow-through
bypass analyzer
D. $5,000 to $40,000
F, G. $10,000 to $15,000
H. $10,000 to $20,000
I. $10,000 to $15,000
J. $20,000 to $27,000
K. $20.000 to $40,000
L. $15,000 to $25,000
M. $2,000 to $5,000

Partial List of Suppliers: (For moisture measurement in air and solids, also refer to Sections 8.32 and 8.34)
Ametek Inc. (www.ametek.com) (D, F)
Anarad Inc. (www.anarad.com) (F, M)
Arizona Instrument Corp. (www.azic.com) (B, I, H)
CEM Corp. (www.cem.com) (G)
Cosa Instrument Corp. (www.cosa-instrument.com) (B, H)

1434
© 2003 by Béla Lipták
 8.33 Moisture in Gases and Liquids 1435

CSC Scientific CO. Inc. (www.cscscientific.com) (B, H)

Delta-F (www.delta-f.com) (F)
Denver Instrument Co. (www.denverinstrumetn.com) (F, H)
Dewcon Instruments Inc (Division of KO-EN Technologies ab) (www.dewcon.com) (J)
Elan Technical Corp. (www.elan.thomasregister.com) (G)
Endress + Hauser Instruments (www.endress.com) (B)
Foss NIR Systems Inc. (www.foss-nirsystems.com) (F)
General Eastern Instruments (www.geinet.com) (B, C, M)
Illinois Instruments Inc. (www.illinoisinstrumetns.com) (A)
Kahn Instruments Inc. (www.kahn.com) (A, B)
Kay-Ray/Sensall/Rosemount (www.emersonprocess.com) (G)
Kin-Tek Laboratories Inc. (www.kin-tek.com) (M)
Luft Instruments Inc. (H)
MEECO Inc. (www.meeco.com) (A)
Michell Instruments (www.michell-instruments.com) (B)
Nova Analytical Systems Inc. (www.nova-gas.com) (B)
Ophir Corp. (www.ophir. com) (F)
Panametrics Inc. (www.panametrics.com) (B, C, M)
Rosemount Analytical Inc. (www.emersonprocess.com) (B)
Teledyne Analytical Instruments (www.teledyne-ai.com) (B, F)
Testo Inc. (www.testo.de or .com) (B)
Thermo Electron Corp. (www.thermo.com) (F, L)
Tiger Optics LLC (www.tigeroptics.com) (K)

Most Popular: Panametrics, Ametek, General Eastern, and Endress & Hauser

INTRODUCTION than water and the weighing procedure. Some built-in scale
oven systems even have experience issues with the weighing
There is a certain amount of overlap between the three sections procedure and air current effects on the sample holders.
that are devoted to the subject of moisture measurement in this Another widely used and accurate (sensitive to 5 ppm)
handbook. (Section 8.32 describes the measurement of mois- method of moisture analysis in gas or liquid samples is by
ture in air; Section 8.33, in liquids and gases; and Section 8.34, titration. The reagent is referred to as the Karl Fischer
in solids.) Some of the instruments, such as the chilled-mirror reagent, which consists of a solution of sulfur dioxide, iodine,
or the thin-film capacitance-type sensors discussed in Section and pyridine in methanol. The titration end points can be
8.32, can operate not only as dew point or humidity sensors found automatically (see Section 8.66) by the measurement
in air but also as moisture analyzers in other gases. Similarly, of the current flow through the solution. Recent concerns
the electrolytic hygrometer, which is discussed in this section, regarding the toxicity of pyridine have led to the use of
could also be used to measure humidity or dew point in air. substitute buffers like imidazole. The iodine water reaction
Similar overlaps exist between the coverage of this section and is the same, and it is still classified as a Karl Fischer titration.
Section 8.34, dealing with moisture in solids, particularly with
respect to microwave, neutron, and infrared analyzers. The
reader is also reminded that separate sections are devoted to PROCESS ANALYZERS
some of the analyzers that can also be used for other analytical
tasks besides moisture detection, such as infrared (Section Some of the industrial process analyzers are of the probe type
8.27) and autotitrators (Section 8.66). The reader is advised to (capacitance, fiber-optic infrared probes), others can look
refer to these sections as well for additional information con- through the process stream in the pipe (microwave), and the
cerning their suppliers and features. majority require some form of a sampling system. Table 8.33a
provides a summary of the ranges and other features of many
of the process moisture analyzers. The subject of sampling
LABORATORY ANALYZERS (which has already been covered to some extent in Sections
8.2, 8.3, and 8.12) will also be addressed in the next section.
An old but still useful method of determining the moisture of
a sample is to heat it and measure the resulting weight loss. Sampling Systems
Some drying ovens are provided with built-in weight scales and
displays that indicate the sample moisture content directly in Some sampling system features are peculiar to moisture ana-
weight percentage. The reliability of this technique is primarily lyzers, and these are covered below. The function of any sam-
a function of the presence or absence of volatile materials other pling system is to deliver a clean, appropriately representative

© 2003 by Béla Lipták

1436 Analytical Instrumentation

TABLE 8.33a
Summary of Moisture Analyzer Features
Type Range Sample Phase Sample System Required Remarks
Electrolytic hygrometer 0–2 to 0–2000 ppm Clean gas Yes Sample flow must be constant
Special sampling for liquids
Change of capacitance 0–10 to 0–1000 ppm Clean gas or liquids * Sample temperature must be constant
Impedance type 0–20,000 ppm Clean gas or liquids No, only for liquids Sample temperature of liquids
must be constant
Piezoelectric type 0–5 to 0–25,000 ppm Clean gas only Yes
Heat of absorption type 0–10 to 0–5000 ppm Clean gas or liquid Yes Sample flow must be constant
Special sampling for liquids
Infrared absorption Gas: 10.1–100% Gas, liquids, and slurries **
Liquid: 6–100%
Microwave absorption 0–1 to 0–70% Liquids, slurries, and No
pastes only

*Available in probe form, but can be direct pipeline-mounted only if flow velocity is under 1.6 fps (0.5 m/s).
**Fiber-optic probe (FOP) designs can be direct pipeline-mounted without sampling.

sample to the measuring element at the required pressure, TEFLON OR GLASS TUBE
temperature, and flow rate. Measurement of moisture, although
simple in principle, is complicated by the fact that most mate-
rials adsorb and desorb moisture.
At moisture levels in the percent range, moisture absorp-
tion is not normally a serious problem; however, at moisture − + − + −
P2O5 FILM
levels in the low parts per million (ppm) ranges, sampling SAMPLE GAS
system materials must be selected so as to minimize their @ 100 cc/min
+ − + − +
effect on the moisture concentration in the sample. Depending
on the levels being monitored, materials that are less likely to
interfere with the moisture content in the sample are stainless
® ® ®
steels, Teflon, Viton, Kel-F, nickel and nickel-plated mate- ELECTRODES
rials, and cadmium and cadmium-plated materials. Materials
that should almost always be avoided are copper and its alloys, FIG. 8.33b
rubber, neoprene, and elastomers in general. Operating principle of the electrolytic hygrometer.
It is important to apply the material considerations to all
parts of the sampling system, including the internals of valves
and filters.
Therefore, the electrolysis current is equal to the electron
15
Electrolytic Hygrometer flow of 2 × 10 electrons per second. Since two electrons of
charge are required to electrolyze one water molecule, the
15
The principle of measurement utilized involves the electrolysis water flow into the cell is equal to 10 molecules per second.
of water into oxygen and hydrogen. Since two electrons are If, for purposes of this example, the water present in the
required for electrolysis of each water molecule, the electrolysis sample is considered an ideal gas, the volume of water enter-
current is a measure of the water present in the sample. If the ing the sample can be calculated using the ideal gas law.
volumetric flow rate of the sample gas into the electrolysis cell
is controlled at a fixed value, then the electrolysis current is a P1 T2
function of water concentration in the sample (Figure 8.33b). V2 = V 8.33(1)
P2 T1 1
This relationship is illustrated in the following example.
Determine the water concentration in a sample when the
3
sample flow rate is 100 cc/min (0.212 ft /h) at 100°F (37.8°C) Avogadro’s Law Avogadro’s law states that equal volumes
and 10 PSIG (68.9 kPa). The electrolysis current is measured of different ideal gases at the same pressure and temperature
at 320 µA. contain the same number of molecules. The volumetric flow
rate of water vapor can be calculated as
−6
electrolysis current = 320 × 10 coulombs per second
18
1 coulomb = 6.25 × 10 electrons V1 = 379 N /A 8.33(2)

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1437

where
N = the flow of water in molecules per second POTTED PVC
26
A = Avogadro’s number (A = 2.73283 × 10 molecules/ BODY
(lb·mol)
RHODIUM
379 = volume of a pound-mole of ideal gas in cubic feet CELL ELEMENT

Subscripts 1 and 2 refer to standard and sample conditions,

respectively.
LOCKING
CLAMP
379 × 10 15
V1 = = 1.39 × 10 −9 ft 3/sec = 5.0 × 10−6 ft 3 /h
2.73 × 10 26 (−) (+)
8.33(3)
O-RING
(−) (+)
560 14.7
V2 = 5 × 10−6 × × = 3.2 × 10−6 ft 3/ h 8.33(4)
520 24.7
3
Since the sample flow is controlled at 0.212 ft /h (100 cc/min), SAMPLE
−6
the volume concentration of moisture is: 3.2 × 10 /0.212 = GAS IN
SAMPLE
GAS OUT
−6
15.1 × 10 . Expressed in parts per million (ppm), the moisture
VOLTAGE TO CELL +
content is 15.1 ppm. The example above is intended to HEATED SS OUTPUT SIGNAL
illustrate the fact that the hygrometer cell sees only a mass CELL BLOCK
flow of water—number of water molecules electrolyzed per
unit time. To obtain an output in the form of moisture content, FIG. 8.33c
the sample flow rate must be a known, constant value. Fur- Electrolytic hygrometer cell.
thermore, the accuracy of the output can never exceed the
accuracy with which the sample flow is controlled.
The commercial electrolysis hygrometer cell consists of Cell Limitations The cell will perform satisfactorily with
a small chamber containing two noble metal electrodes that a variety of samples; however, a number of sample materials
support a thin layer of desiccant. Moisture in the sample is will cause problems or should not be monitored with this
absorbed by the desiccant and electrolyzed by means of a instrument. Gases known to cause problems are hydrogen
voltage-regulated power supply connected to the electrodes. and oxygen. These gases have already been discussed in terms
Units are available for use in nonhazardous areas with the of the recombination effect, above. Some other gases for which
sample flow control, electrolysis cell, and electronics pack- this instrument is unsuitable are unsaturated monomers, alco-
aged as a single unit. When used in hazardous areas, the hols, amines, ammonia, hydrogen fluoride, and CHClF2

cell and flow controller are housed in an explosion-proof (Freon ) refrigerant.
conduit, and the electronic circuitry is remote-mounted (see Alcohols are seen by the cell as water. Amines and ammo-
Figure 8.33c). nia usually react with the desiccant. Hydrogen fluoride can
corrode the internals of the cell. The data collected on CHClF2
Recombination Effect Although this instrument will operate refrigerant indicates an anomaly, although the reason for this
satisfactorily with a variety of samples, a phenomenon called is not fully understood. Unsaturated hydrocarbons such as
the recombination effect introduces large errors at low moisture butadiene or monomers with a strong tendency to polymerize
levels in hydrogen-rich or oxygen-rich samples. Recombination cannot be monitored, as the cell will be quickly coated with
is the reversion to water of the electrolysis products; it intro- polymer.
duces an error into the measurement when the recombined Generally, the instrument should not be used with sam-
oxygen and hydrogen are re-electrolyzed. Apparently, all elec- ples whose components may deposit in the cell (condensable
trodes catalyze this reaction, although some electrode materials vapors). When a cell becomes contaminated, it will show a
do so more than others. The use of rhodium as the electrode memory for polarity; that is, the outputs under forward and
material has been found to minimize recombination. reverse flow through the cell will not be equal. The electro-
When monitoring very low moisture levels in oxygen- lytic hygrometer is suitable for most inert elemental gas
rich or hydrogen-rich atmospheres, the recombination produces applications and for other gases that do not react with P2O5.
a relatively large error even with the best choice of electrode A cell will lose sensitivity when exposed to moisture
materials. For such an application, two sensors are used; one levels of a few parts per million over a period of weeks. This
measures at sample flow rate X and the other at sample flow sensitivity loss is due to the elution of desiccant with the
rate 2X. Because the error due to recombination is a constant, sample. However, this process occurs over a long period of
subtraction of the two sensor outputs yields a signal that is time and the cell can be recoated fairly easily in the field
independent of recombination. during periodic maintenance.

© 2003 by Béla Lipták

1438 Analytical Instrumentation

Liquid Samples The electrolytic hygrometer cell can be PRESSURE

FILTER
REDUCING VALVE
exposed only to gases. When the process sample is in the liquid
LIQUID ELECTROLYTIC
state, there are two ways of using the electrolytic hygrometer. SAMPLE PI
CELL
INLET SAMPLE
Recall that this process must not result in transporting to the FLOW
sensor vapors that will condense or be absorbed. CONTROL
STRIPPING
SAMPLE COLUMN
BY-PASS
1. Sample vaporization. If the liquid sample has suffi-
ciently high vapor pressure, the sample can be vapor- ELECTROLYTIC
DRYER
DRAIN
VALVE
ized, and the measurement is then made as a gas GAS
sample. The sample is vaporized by means of a vapor- NEEDLE DISCHARGE
VALVE LIQUID
izing regulator located as close as possible to the sam- STRIPPING
SAMPLE
N2 FI
ple take-off point to minimize lag. The remainder of SUPPLY
DISCHARGE

the sampling system is the same as for gas samples.

2. Sample stripping. If the sample cannot be vaporized, FIG. 8.33d
this method can be used to obtain a gas sample at Liquid sample stripping.
the cell. The moisture is stripped from the liquid sam-
ple in a falling film column. Liquid sample is contin-
uously metered into the top of the column, where it constant between capacitor plates. The value of capacitance
descends as a thin film. Dry nitrogen, metered into the is a function of plate area, plate spacing, and the dielectric
bottom of the column, ascends and removes moisture constant of the material between the plates. The dielectric
from the descending film of liquid sample. A filter constant of a material has a unique value for each substance
located at the top of the column removes any droplets and is related to the polarity characteristics of the molecules.
entrained in the nitrogen before the gas is passed This property can, therefore, be used to detect the presence
through the measuring cell. of a specific substance in a pipeline stream.
Figure 8.33e shows an air capacitor connected to a bat-
A drain valve at the bottom of the column facilitates
tery and the same capacitor with mica between the plates.
removal of stripped sample. The nitrogen is dried before
The capacitance of mica is six times that of air because the
entering the column by passing through an electrolytic dryer,
dielectric constant, K, of mica is 6, and the K of air is 1. If
similar to the measuring cell, where moisture is removed by
the mica were inserted gradually between the plates, then the
electrolysis. The supply nitrogen moisture level should be
capacitance would increase correspondingly from minimum
fairly low (less than 500 ppm) to avoid overloading the dryer
to maximum. Water, with a dielectric constant of 80, would
(Figure 8.33d).
increase the capacitance to 80 times the value of the air
capacitor.
Sampling System As pointed out previously, the electro-
The measuring cell of the hygrometer consists of two
lytic hygrometer is sensitive only to the mass flow of water
electrically insulated concentric metal cylinders to form the
into the cell; the concentration of moisture is inferred from the
measuring capacitor plates (Figure 8.33f ). The annulus
known sample flow rate. Therefore, the accuracy of the readout
is partially dependent on the accuracy of the sample flow between the cylinders is filled with alumina desiccant. Two
control and the accuracy of the moisture measurement can porous metal discs support the cylinders and retain the des-
never exceed that of the sample flow control. This fact should iccant in the annulus. The sample is allowed to flow through
be kept in mind when selecting a sample flow controller. the annulus, and process water is absorbed or desorbed by
the desiccant, which remains in equilibrium with the sample
Capacitance Hygrometer in terms of percent saturation. The desiccant thus amplifies
the moisture content of the sample because the saturation
The principle of measurement utilized is the change of capac- level of the desiccant that is in equilibrium with the sample
itance associated with a change of the sample dielectric is very much higher than that of the sample.

CAPACITOR
ELECTRODES
POSITIVE +
CHARGE +++++++++ BATTERY +
− − − −K −= 1− − − − − MICA K = 6
−
BATTERY
NEGATIVE AIR
CHARGE SPACE
Air capacitor Mica capacitor

FIG. 8.33e
The capacitance is a function of the dielectric constant of the material filling the space between the capacitor plates.

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1439

CAP will dissolve 300 ppm of water; at 70°F (21°C), it will dis-
solve 150 ppm. With a moisture level of 15 ppm, the propane
would be 5% saturated at 100°F and 10% saturated at 70°F.
O RING
SEAL
Therefore, the output signal would change 100% with no
OUT
change in moisture content if temperature were allowed to
SPRING
vary from 100 to 70°F.
The sample is maintained at constant temperature by
HANDLE MEAS. CELL
POROUS passing it through a coil immersed in a constant temperature
METAL bath immediately upstream of the measuring cell. Additional
ANNULUS FILLED DISC
WITH DESSICANTS temperature control of the sample is recommended even if
SPRING
LOADED
the process stream is temperature controlled, because ambi-
CLEAN GAS CONTACT ent conditions can affect the sample temperature.
OR LIQUID INLET

SEAL &
Thin-Film Capacitance Probes The design and operation
INSULATOR of thin-film capacitance probes have been described in the
previous section in connection with Figure 8.32j and 8.32k.
These units are also suitable for explosion-proof area classi-
fications and can be used in measuring the moisture content
FIG. 8.33f of many gas and liquid samples (many liquid hydrocarbons,
Capacitance-type measuring cell. for example). The probe units are suitable for installation of
high-pressure processes and are provided with microproces-
The measuring capacitor is part of an electrical circuit sor-based electronics capable of self-diagnostics and of mul-
that includes a reference capacitor. This circuit is usually tiple measurements or unit conversions (Figure 8.33g). These
powered by a 15-kHz fixed-amplitude sinewave. The mea- probes are available with a variety of covers/shields that can
suring and reference capacitors are switched alternately into be used for a variety of sampling situations (particulate, bub-
the circuit such that its output voltage is a function of the bles, various viscosity liquids, etc.).
connected capacitance. The output signal is a sinewave of A unique configuration of a thin-film capacitance ceramic
the same frequency as the power signal and whose amplitude sensor is one with an integral calibration system. The in situ
varies with the measuring and reference capacitors as they and automatic calibration is performed against a reference
are switched into the circuit. This difference in amplitude is gas. The precision of this system is the result of the use of a
related to the measured capacitance, which, in turn, is a func- chilled-mirror dew-point hygrometer to measure the mois-
tion of the moisture content of the sample. ture content of the reference gas used in the calibration
The electronic chassis is normally mounted near the mea- (Figure 8.33h).
suring cell and sampling system, which should be close to
the sample take-off. In addition to an integral indicator and
output signals of standard serial, milliampere, and millivolt Limitations The capacitance instruments are not suitable
ranges, the electronic chassis can be provided with multiple for polar materials such as alcohols, because these become
range switching and alarm contacts. conductive at the 15-kHz operating frequency and short the
measuring capacitor. Data on instruments operating at higher
Sampling Systems It was pointed out previously that the frequencies were not available.
desiccant in the measuring cell is the equilibrium with the Free sulfur and iron oxide concentrate on the desiccant
sample in terms of percent saturation. For this reason, it is will short the element.
necessary to maintain the sample and measuring cell at con- At sample viscosities above 500 Saybolt Seconds
stant temperatures. For example, propane at 100°F (38°C) Universal (SSU), moisture should not contact the desiccant

1/2" (12 mm)

DIAMETER
PROBE TIP

FIG. 8.33g
The gold/aluminum oxide probe can measure the moisture content of both gases and liquids. (Courtesy of Endress + Hauser Inc.)

© 2003 by Béla Lipták

1440 Analytical Instrumentation

ANALYSER SYSTEM RACK

CERMET HYGROMETER
ANALOGUE/ALARM OUTPUTS
SENSOR CABLE

SAFETY
EARTH

SENSOR SAMPLING SYSTEM

VENT
AMBIENT TEMPERATURE
CONTROL
CALIBRATION
SAMPLE
TUBING

REFRIGERATION
SYSTEM

MAIN POWER SUPPLY

CERAMIC SENSOR
AUTOMATIC DEWPOINT GENERATOR

E L SL N

HUMIDITY
SELECTORS
SATURATOR POWER SUPPLY

t
PROCESS SAMPLE INLET
TEMPERATURE CONTROL
CALIBRATION CASE SAFETY PRESSURE
SAMPLE INLET RELIEF VENT
FAST LOOP SAMPLE VENT (OPTIONAL)
SYSTEM
VENT

AIR OR INERT
PRESSURE SWING DRYER GAS SUPPLY

NON-HAZARDOUS AREA HAZARDOUS AREA

FIG. 8.33h
Thin-film ceramic sensor provided with integral and automatic means for recalibration. (Courtesy of Michell Instruments Ltd.).

at reasonable flow rates, and the measurement becomes is applied over the aluminum oxide. Leads from the gold and
meaningless. aluminum electrodes of the probes connect the sensing ele-
The life of the desiccant is partially a function of solids ment to the measuring circuitry (Figure 8.33i).
in the stream, which pass through the filter. While the desic- Water vapor penetrates the gold layer and equilibrates on
cant can be replaced very readily, the short life of the desic- the aluminum oxide. The number of molecules adsorbed to
cant excludes the use of this instrument on sample streams the aluminum oxide is a function of the water vapor pressure
containing large quantities of fines. in the sample. Each water molecule adsorbed contributes a
distinct increment to the total conductivity of the aluminum
Impedance Hygrometer oxide. The total probe impedance, the reciprocal of probe
conductivity, is thus a measure of water vapor pressure in the
This instrument measures the water content of a sample by sample. Water vapor pressure of a gas sample uniquely deter-
means of a probe whose electrical impedance is a function mines the dew-point temperature and moisture content of the
of the vapor pressure of moisture in the fluid. Impedance is sample. The output is normally calibrated to these units, as
the apparent opposition to the flow of alternating current. they are more convenient to use than vapor pressure. In the
The probe consists of an aluminum strip that is anodized to case of certain liquid samples, the moisture content can be
form a porous layer of aluminum oxide. A thin coat of gold measured through the application of Henry’s law.

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1441

OUTER GOLD APPROX. 3" (76 mm)

ELECTRODE CABLE TO
(METAL EVAPORATED ELECTRICAL READOUT 3/4" (19 MM) SENSITIVE
OVER ANODIZE) CONNECTIONS
FLANGE ELEMENT

ANODIZED SURFACE ALUMINUM

BASE

OUTER GOLD
ELECTRODE TEFLON SIGNAL
ALUMINUM CABLE CONNECTOR SUPPORT LEAD
OXIDE
FIG. 8.33j
Flanged probe.
ALUMINUM
BASE
PROBE CROSSECTION
MEASURING
CELL
FIG. 8.33i EXPLOSION
WITH PROBE
Impedance-measuring sensor. PROOF
CONDUIT SAMPLE IN
FITTING FOR
In terms pertinent to this measurement, Henry’s law WIRING
states that, at constant temperature, the mass of water vapor TERMINALS

dissolved in a given volume of liquid is in direct proportion FLASH BACK

to the partial pressure of water vapor in the sample. Henry’s CONDUIT ARRESTORS
SEAL
law can be restated in simpler form to read that the weight FITTING
concentration of moisture in the sample is equal to the partial SAMPLE
pressure of water vapor times a constant. However, Henry’s OUT
law holds only for liquids with moderate solubility for water
vapor, such as pure hydrocarbons, silicone oils, etc. The FIG. 8.33k
instrument cannot be used on liquid samples with a high Impedance hygrometer installation for hazardous areas.
solubility for water such as the alcohols, because the rela-
tionship expressed by Henry’s law does not hold.
the instrument is limited to those fluids with a moderate
solubility for water vapor; thus the instrument is not suitable
Installation The probe is designed for direct installation for measurement in polar liquids such as the alcohols.
into the process stream. However, where danger of explosion This sensor is suited for moisture measurement in hydro-
exists, the probe is inserted into a small sample chamber carbons that consists of only hydrogen and carbon atoms or
connected to the process line with flashback arrestors located in halogenated hydrocarbons, amides, esters, many silicone
at the sample inlet and outlet. For liquid samples, temperature oils, etc. The aluminum oxide probe is suited for moisture
must be controlled and/or monitored and compensated for. detection in cyclohexane, diethyl ether, liquefied natural gas,
Several probes can be connected to a single readout device ethane, propane, butadiene, butane, styrene, and propylene.
and each probe can be monitored directly or by means of an It is not recommended for applications where conductive
automatic scanner. A serial, milliamp, and voltage output solid particles, mercury, salts, or acids and bases are present.
signals are available for recording or other functions; single Hydrofluoric acid, for example, can damage the sensor, even
or multiple range measurements are standard (Figure 8.33j in such low concentrations as 1 ppm.
or 8.33k). The analyzer requires recalibration about twice a year
and must be protected against excessive ambient tempera-
Sampling System Often a sampling system is not required tures (158°F, or 70°C) and against excessive flow velocities.
on gas samples. On liquids, sample temperature must be held If the process stream flows at a velocity in excess of 1.7 fps
constant or monitored and compensated for. These probes are (0.5 m/s), it is advisable to place the sensor in a slow bypass.
available with a variety of covers/shields that can be used for The viscosity of the sample also needs to be taken into
a variety of sampling situations (particulate, bubbles, various consideration.
viscosity liquids, etc.). Often these shields can eliminate the
need for, or simplify the design of a full-blown sampling system. Piezoelectric Hygrometer

Limitations The instrument can be used on all gases that Piezoelectric quartz crystals have a number of uses: in com-
are not corrosive to the probe and that will not spontaneously munications, to control frequencies; in industry to measure
polymerize on contact with the probe materials. On liquids, temperature and thickness of metal films; and to generate

© 2003 by Béla Lipták

1442 Analytical Instrumentation

ultrasonic waves. In moisture measurement, advantage is

taken of the oscillating crystal’s sensitivity to deposits of
foreign material on its surface. Commercially available crys-
tals will show a frequency change of 2000 Hz per microgram DESICCANT
of materials deposited.
For moisture measurement, the quartz crystals are coated THERMOCOUPLES
WET
with a hygroscopic material and exposed to the sample. SAMPLE DRY SAMPLE
Water from the sample is absorbed by the coating, which IN IN
increases the total mass and decreases the oscillating fre-
quency of the crystals. To measure changes of decreasing SAMPLE OUT
moisture concentration and to simplify the frequency mea-
surement, two crystals are used. One crystal is exposed to a
wet sample and the other to a dry reference gas for a short
period. Then, sample and reference gas flows are switched
so that moisture is absorbed by one crystal while being
desorbed by the other. A cycle timer controls this switching.
SIGNAL LEADS
The frequency changes between the two crystals are in pro- TO READOUT
portion to their mass changes and the moisture content of
the sample gas. FIG. 8.33m
Measuring cell for the heat of an adsorption-type hygrometer.
Sampling System Dry reference gas is obtained by passing
part of the sample gas through an integral dryer, or an external
source of dry of nitrogen can be used. Two solenoid valves
cable to only those liquid samples that can be stripped or
located upstream of the measuring cell switch the sample and
completely vaporized within the operating limits of the
reference gases from one crystal to the other. Sample gas
instrument. In addition, monomers with a strong tendency to
flow to each crystal is regulated, although small variations of
polymerize, such as butadiene and styrene, may coat the
sample flow rate will not affect the measurement. For samples
crystals and prevent proper operation.
near atmospheric pressure, a vacuum pump may be used to
draw sample through the measuring cell (Figure 8.33l). Some
versions of this instrument design are being supplied with
integral permeation tube calibration/benchmark systems to Heat-of-Adsorption Hygrometer
help ensure that the measurement is still accurate.
The process of adsorption and desorption involves an
Limitations The instrument is not capable of measuring exchange of energy. During adsorption, energy is released to
moisture in liquid phase samples. For this reason, it is appli- the environment, and, during desorption, energy is removed.
When a wet gas is passed through a column of adsorbent,
which selectively adsorbs moisture, the temperature rise due
SAMPLE & WET to heat liberation is in proportion to the amount of moisture
REFERENCE REFERENCE
GAS VENT GAS SUPPLY
being adsorbed. Other factors can affect the heat of adsorp-
SAMPLE DRYER
tion. They include things like the nature of the adsorbent and
GAS FI SHUT-OFF operating temperature, and they are selected to maximize the
INLET VALVE
CHECK
heat of adsorption (Figure 8.33m).
MEASURING
CELL
VALVE The sensing element consists of two desiccant columns,
GAS
FLOW ALTERNATE each containing a number of thermocouples connected in
DURING DRY
SECOND
series. This assembly is housed in a temperature-controlled,
REFERENCE
HALF
CYCLE
GAS SUPPLY thermally insulated housing. In operation, wet sample gas is
FILTER

FILTER

SHUT-OFF passed through one column and dry reference gas through
VALVE
the other. Since adsorption and desorption occur simulta-
FLOW neously, the net thermocouple output voltage represents the
FLOW algebraic sum of the heat gained by one column and lost by
S S
ENERGIZED DE- the other. The reference gas and sample gas streams are
SOLENOID VALVES ENERGIZED switched on a time-cycle basis to maintain dynamic condi-
SAMPLE FLOW
ADJUST VALVE
REFERENCE GAS
FLOW ADJUST VALVE
tions necessary for measurement in the sensing element.
Sample flow is closely regulated, because moisture concen-
FIG. 8.33l tration is inferred from the known sample and reference flow
Piezoelectric hygrometer sampling system. rates (Figure 8.33n).

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1443

HEAT LIGHT HALF OF FILTER

EXCHANGERS SOURCE WHEEL TRANSMITS
MEASURING
SAMPLE WAVE LENGTH SAMPLE PHOTO
FLOW LENS
FI FI CELL CELL
ADJUST
VALVES
LENS ELECTRICAL
MIRROR OUTPUT
MEASURING
TO
FLOW MEASURING
CONTROL CELL
CIRCUIT
VALVES
GLASS
FLOW SYNCHRONOUS
WINDOW SAMPLE
ELECTRIC
FLOW FLOW
S SOLENOID MOTOR
SOLENOID VALVE HALF OF FILTER
VALVE (EMERGIZED) FILTER
S SAMPLE WHEEL TRANSMITS
(DE-EMERGIZED) VENT WHEEL
GAS FLOW REFERENCE WAVE LENGTH
DURING DRYER TEMPERATURE
SAMPLE CONTROLLED
SECOND
INLET ENCLOSURE
FILTER
HALF CYCLE FIG. 8.33o
Schematic representation of the infrared moisture detector.
FIG. 8.33n
Sampling system for the heat of an adsorption hygrometer. other factors such as reflection and dispersion of the radiant
energy will contribute to the attenuation. Therefore, it is
Sampling System The sampling system contains the fol- necessary to either calibrate these factors out of the measure-
lowing: ment or to use a reference wavelength that is not absorbed
by moisture but is affected by all other factors to the same
1. A dryer to remove all moisture from the sample gas extent as the measuring wavelength. The difference in the
so that the gas can be used as the reference. attenuation of the measurement and reference wavelength is
2. Solenoid valves and cycle timer to switch reference then a function of moisture content only.
and sample gas between the two desiccant columns. The sensing element consists of three groups of compo-
3. Flow regulators to hold sample and reference gas flows nents: an IR radiation source, sample cell, and radiation
constant. detector. The radiation source consists of a lamp, filters to
4. Heat exchangers within the insulated enclosure to heat pass the measuring and references wavelengths, and optics
both reference and sample gas to the operating tem- to direct the beam through the glass sample cell. The radiation
perature of the sensing elements. The sample flow to pickup consists of optics to collect the transmitted radiation
the instrument is split into two equal flow rates at the and a photocell to convert the electromagnetic energy to an
sampling system; one half is the measured sample, and electric current.
the other half is dried for use as the reference gas. The The measuring and reference wavelengths are allowed to
timer actuates solenoid valves in the sample and ref- impinge alternately on the photocell so that two sets of cur-
erence gas lines so that each desiccant column is rent pulses are produced. These pulses are converted into two
exposed alternately to reference and to sample gas. direct current (DC) signal levels whose ratio represents the
moisture content of the sample (Figure 8.33o). For more
Limitations The sensing element cannot be exposed to liq- details on IR analyzers, refer to Sections 8.23 and 8.27.
uid samples; however, the moisture content of liquid streams More complex spectrometers (IR, FTIR, NIR, and
can be measured by one of two methods. Liquids with low TDLAS) are also being used for moisture detection and mon-
boiling points can be vaporized at the sample take-off points itoring applications when the application requires or justifies
by means of a vaporizing regulator, and the measurement is it (lower levels of detection, multiple simultaneous compo-
then performed on a gas sample. Or the moisture can be nents [including moisture], complex sample matrices, etc.).
stripped from the liquid sample with dry gas, and from the One of the newer one is tunable diode laser adsorption spec-
moisture content of this gas the moisture concentration in the troscopy (TDLAS), which claims measurements down to the
liquid sample can be inferred. Sample stripping is described ppt range with an appropriate sample.
in conjunction with electrolytic hygrometers.
Sampling System There are no special sampling system
Infrared Absorption Hygrometer requirements for this instrument. Selection of hydrophobic
(lacking affinity for water) materials for sampling system
Water absorbs electromagnetic radiation in the infrared (IR) components is not critical, given that the instrument can
region of the spectrum. Specifically, infrared radiation of 1.4 measure only relatively high moisture content, where mois-
and 1.93 µm wavelength is absorbed strongly by water. By ture contributed to the sample by the sampling system is
measuring the attenuation (decrease of light intensity) of a insignificant in its effect on the measurement. When fiber-
beam of this wavelength as it passes through a sample, the optic probes (FOPs) are used, the need for sampling systems
moisture content of the sample can be determined. However, is eliminated (Section 8.23).

© 2003 by Béla Lipták

1444 Analytical Instrumentation

Limitation There are several general limitations to the field reflected by the sample. The probe of the reflectance unit
of application of this instrument. The sample fluid must not does, however, require contact with the process material.
be corrosive to the glass sample cell. It must have some Selection of the type of probe to be used is determined by
minimum transparency for the measurement and reference the reflective and absorptive properties of the sample at oper-
wavelengths. Additionally, there cannot be any other compo- ating conditions and by the physical installation. An advan-
nents present in the sample that will selectively absorb either tage of the microwave hygrometer over the infrared unit is
the measuring or reference wavelength. that there are no optical windows that may coat and effect
the measurement.
Microwave Absorption Hygrometer
Limitations The unit is not suitable for moisture measure-
In the microwave frequency band of 20 to 22 GHz (K band), ment in gases.
the wavelength is about 13 to 15 mm. In this frequency band,
moisture (free water) responds uniquely. Only the water mol-
Dipole Polarization Effect Moisture Sensor
ecule produces molecular resonance. The microwave radia-
tion can be guided by waveguides or transmitted through the The principle of this measurement relies on the fact that
process stream from its source to a detector. water molecules are formed by covalent bonds. These bonds
The operating principle of this instrument is the same as share electrons unequally between the hydrogen and oxygen
that of the infrared absorption hygrometer, namely, selective atomic nucleuses, and the result is an electrical asymmetry
absorption of electromagnetic energy by moisture in the sam- around the molecule (two slightly positively charged hydro-
ple. However, in this case, radiant energy in the microwave gen atoms and a slightly negatively charged oxygen atom).
region is used. A transmitter provides a microwave beam that When these molecules are subjected to an electric field, they
can either be transmitted through or reflected by the sample tend to align with the electric field because of their dipole
material. The receiving unit senses wave attenuation and nature. Unfortunately, other influences such as surface polar-
phase shift and provides an output. The unit senses the mass ization, ions, etc., can also create an electrical effect in the
of moisture in the beam path so that the readout is normally field.
in terms of the mass of moisture per unit volume. However, By selecting an appropriate sensor field frequency (e.g.,
with proper calibration, readout in weight percent can be 20 kHz), the influence of these other influences can be
achieved. reduced. Other polar molecules can interfere, but, because
Several probe configurations are possible, depending on water molecules are very specific, this is not considered a
the application. Figure 8.33p shows the transmitting and significant factor in most analysis. Ammonia (NH3) is the
receiving probes welded into a section of pipe. Internally, closest to water, and its interference effect is listed as about
®
Teflon contacts the sample. This type of arrangement can 3 3
1:50 (±50 gm/m NH3 equals about ±1 gm/m of water).
be advantageous on slurries and pastes, because there are
A dipole probe is depicted in Figure 8.33q. The sensor
no obstructions to flow. The units described above measure
measurement chamber is an open capacitor that is protected
the attenuation of a microwave beam in the sample by
from coating or attack by the process steam components by
moisture.
a membrane filter. The measurement output is directly pro-
The reflectance-type unit contains the source and detector
portional to the number of water molecules in the field and
in a single housing. Rather than detecting the transmitted
is reported as a mass per unit volume output. The probe can
microwave energy, this design measures the amount of energy
be made from several materials (stainless steel, inconel, etc.)
and can even be coated with nonstick materials. The result
is an available probe for most needs. The probe is usually
heated to about 10°C above the highest process temperature
to provide a constant measurement temperature and to min-
imize potential condensation and fouling. Versions of this
RECEIVER probe design have been used to measure moisture in pro-
cesses as hot as 430°C.
Because the measured signal is directly proportional to
the number of water molecules in the chamber, calibration is
TRANSMITTER
very easy. A zero-reference capacitor is hermetically sealed
in the probe and, once the zero reference is established and
because the measurement is linear, you need only one other
calibration point for the entire 0 to 100% range. The calibra-
tion will not change for the life of the instrument as long as
FIG. 8.33p the physical integrity of the measuring and reference cham-
Microwave sensor installed in pipe. bers are maintained.

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1445

GAS MOLECULES (Including WATER)

MEMBRANE FILTER
MEASUREMENT CHAMBER
REFERENCE CHAMBER

ELECTRICAL CABLE

To TRANSMITTER
EXCITATION ROD UNIT
HEATER ELEMENT

PROCESS FLOW

FIG. 8.33q
A dipole polarization effect moisture sensor. (Courtesy of Dewcon Instruments Inc.)

Sampling System There are no special sampling system CRDS Technique

requirements for this instrument. The probe is designed to Acousto-Optical
insert directly into most processes. Extreme sample condi- Modulator Diode Laser
Mirror
tions may require some sample conditioning. Isolator
Optical
HR Mirror Fiber
Limitations The unit is not suitable for moisture measure-
ment in liquids. Cavity Ring-Down
Absorption Cell

Cavity Ring-Down Spectroscopy (CRDS) HR Mirror

Moisture Analysis
Mirror Photodiode Trigger Computer
This spectroscopic technique is different enough from those
previously discussed that it will be discussed here on its FIG. 8.33r
Cavity ring-down spectroscopy instrument. (Courtesy of Tiger Optics
own. The measurement principle is based on absorption
LLC.)
spectroscopy. The difference between this and more tradi-
tional absorbance spectroscopy is that this measurement is
exposed to the light, the sensitivity of the instrument is
based on time rather than just the magnitude of the absor-
enhanced to a level of actually achieving part-per-trillion (ppt)
bance. The actual measurement is based on a light intensity
levels of measurement on water in gases. This technology has
decay rate. By measuring the time it takes for the appropri-
been utilized primarily in the semiconductor industry, on
ate molecular absorbance light fingerprint to fade or ring-
ultrapure gases, and for bio-safety. Robust process-hardened
down, you can measure extremely low levels of the absorbing
instruments for more typical process applications are under
contaminants.
development.
A CRDS instrument is depicted in Figure 8.33r. A
continuous-wave laser emits a beam of light energy through Sampling System Special sampling systems are required
a highly reflective mirror into the absorbance cell (cavity). for this instrument. The sample must be clean, temperature
This cell is capable of reflecting the light back and forth many controlled, pressure controlled, and noncondensing in addi-
times and, by this multiple path design, will typically achieve tion to being compatible with materials that it will contact.
total path lengths in the range of 100 km (vs. typically 100 m
for typical absorbance multiple pass cells). Once the detector Limitations The unit is not suitable for moisture measure-
sees a preset level of light energy, the light source is diverted ment in liquids or condensing vapors. Gas samples should
from the cavity. On successive passes of light, a small amount be adequately cleaned and/or conditioned in an appropriate
of light is sensed and, once the light rings down, the mea- sampling system.
surement is complete.
The sample concentration is determined from the shape Neutron Backscatter Moisture On-Line Analyzer
of the ring-down decay curve. Figure 8.33s shows an example
of a ring-down decay curve. This measurement principle is essentially the same as that of
Because the measurement is time based, it is not subject the nuclear moisture gauge discussed in greater detail in
to small variations in laser intensities. Because of its very Section 8.34. It is a neutron radiation source that is focused
long path length, where molecules in the sample are frequently on a sample, and the neutron backscatter is measured and

© 2003 by Béla Lipták

1446 Analytical Instrumentation

0.5

0.4

Signal (Volts)
0.3 τ = 295.82 = 0.20 µsec

data points
0.2
fit

0.1

0.0
0 200 400 600 800 1000 1200 1400 1600
time (µsec)

FIG. 8.33s
Typical ring-down decay curve. (Courtesy of Tiger Optics LLC.)

correlated to an appropriate moisture content. High-energy required. This could be factor if you are encountering foam
neutrons are focused on the sample. Backscattered low- or slurries of significantly varying densities.
energy neutrons that have interacted with a hydrogen nucleus A sensor can typically measure a semisphere sample vol-
are detected and correlated to the hydrogen/moisture density. ume of approximately 18-in. (46 cm) radius. If one is monitor-
As a result of the advancements in this design, it is now also ing a sample of varying consistency, the overall measurement
being considered for some nonsolids applications, although accuracy can be increased by placing the sensor where it
the majority of its applications is still on solid samples. monitors a flowing sample. This effectively will give an inte-
Figure 8.33t depicts a typical installation. grated or averaged moisture value.
Because the measurement is so focused on the neutron
and hydrogen nucleus interaction, it is not an appropriate Sampling System There are usually no sampling system
measurement where hydrogen-containing species, other than requirements for this instrument. The sensor is typically
water, can vary in concentration. This essentially eliminates mounted directly on the process.
the world of hydrocarbons, but it is still suitable for inorganic
chemicals, salts, mineral applications, etc. Moisture in blast Limitations The unit is not suitable for moisture mea-
furnace coke is still the dominant application. Optional den- surement in gases or samples with varying levels of hydro-
sity compensation systems are available but are generally not gen atoms (like most hydrocarbons). If the sample density
varies significantly, a density measurement should be made
for compensation.

Calibration of Moisture Analyzers

Moisture calibrators are available in two conceptual designs.

They are based on either saturated gas dilution or permeation
tube designs. Some moisture analyzers include a calibration
system integral to the instrument or as an available option.
In addition to calibration, these can be helpful to test/ensure
that the moisture sensor is still alive and responding.
The first design that will be reviewed is the saturated gas
dilution system concept. Figure 8.33u depicts a standard
range and a lower range design. Both start out with a water-
saturated vapor stream at a known temperature and, conse-
quently, known moisture content. This stream is then diluted
FIG. 8.33t in a controlled (ratio) fashion with a dry gas stream to yield
Example of a neutron backscatter moisture sensor installation. the desired value. Obviously, the larger the single-stage dilu-
(Courtesy of Thermo Electron Corp.) tion ratio, the more difficult it is to maintain precise control.

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1447

FIG. 8.33u
Example of gas dilution based moisture calibration systems. (Courtesy of Panametrics Inc.)

The lower system in the figure offers multiple opportunities If one is adding it to a system, and if the dilution gas is truly
(stages) to dilute the saturated stream with dry gas and will dry or constant, it is possible to generate an actual moisture
allow you to achieve more precise control of calibration value. Typically permeation systems operate at much lower
samples at lower values than a large-ratio single-stage dilu- flow rates than saturation dilution systems.
tion approach. A recent advancement of the permeation tube design is
The second design concept is the permeation tube design. depicted in Figure 8.33w. It depicts a permeation tube system
Figure 8.33v depicts a simple permeation tube moisture cal- that is designed to introduce moisture into gas streams that
ibration design. Permeation tube designs can be used to gen- contain reactive components. Some examples of reactive gas
erate known standards from dry gases or to make known components that were considered include ammonia, HCl,
standard additions to samples or standards. They are based HBr, chlorine, silane, and others. The reactive component
on the concept that, under controlled conditions, the rate/vol- does not come in contact with the permeation tube, and the
ume of moisture that will permeate through a permeation reactive matrix is formed by dilution after the water addition.
tube (its membrane wall) is constant. By also controlling These need to be evaluated on an individual application basis
these conditions and the dilution gas flow rates, one can at least and surely won’t work for all situations, but they are a nice
generate a reproducible and predictable moisture standard. addition to the analyst’s toolbox.

© 2003 by Béla Lipták

1448 Analytical Instrumentation

ZERO VALVE DILUTION FLOW

FLOW METER ZERO/SPAN
CONTROL
OUT
FI FC
1
PRESSURE
REGULATOR
PERMEATION OVEN

SPAN VALVE
FC
2
PERM. TUBE OVEN FLOW CONNECTION
CONTROL BLOCK
VENT

FIG. 8.33v
Example of a permeation tube based moisture calibration system. (Courtesy of Kin-Tek Laboratories Inc.)

Temperature Controlled Oven

Base Mixture
Base Mixture
To Vent

BPC 2
Oven (carrier) Flow N.O.
Permeation Tube
FC
2
Primary Mixture
C
Inert
Primary Mixture
Diluenin FI FC Overflow
1 1
Pressure BPC 1
Regulator
FC

Split Flow
4

Reactive Overflow
Diluent In FC To Safe vent
3
Pressure BPC 3
Regulator

Span/Zero
Out

FIG. 8.33w
Example of a permeation tube based moisture in reactive gases calibration system. (Courtesy of Kin-Tek Laboratories Inc.)

Limitations Generally speaking, the saturated gas dilution Cucchiara, O., The measurement of dissolved water in organic liquids using
design concept is better for generating higher concentration a hygrometer, Analysis Instrum., 15, ISA, 1977.
Dudek, J. et al., Semiconductor gases exact max, InTech, 32–33, July 2002
standards and higher standard flow rates. The permeation Gumpert, R. and Pakulis, I. E., Three moisture measurement techniques,
designs are better for lower concentration standards. InTech, August 1981.
Jutlia, J. M., Multicomponent on-stream analyzers for process monitoring
and control, InTech, July 1979.
Kohler, H. M. and Mathew, A., Continuous in-situ and elevated temperature
moisture measurement in high particulate reactive processes, ISA1996
Bibliography Conference Paper #96–002
Mator, R. J., Trace moisture analyzers and their calibration, Analysis
Bailey, S. J., Moisture sensors, Control Engineering, September 1980. Instrum., 12, ISA, 1975.
Belkim, H. M., Factor affecting response characteristics of electrolytic McKinley, J. J., Using permeation tubes to create trace concentration mois-
instruments for detecting moisture, AID/ISA Symposium, Pittsburgh, ture standards, National Conference of Standards Laboratories Con-
May 1970. ference, Atlanta, GA, July 1997.

© 2003 by Béla Lipták

 8.33 Moisture in Gases and Liquids 1449

McKinley, J. J., Using permeation tubes to prepare trace moisture standards Polta, R. C., Anderson, C. T., and Stule, D. A., Monitoring moisture in
in reactive gases, International Symp. on Humidity and Moisture, Taipei, dewatered sludge, InTech, June 1980.
Taiwan. Scelzo, M. J., The trend is toward processing and measurement at “ultralow”
Mettes, J. and Beck, T., Manufacturers today require PPB moisture detection, moisture levels, Meas. Contr., February 1992.
Meas. Contr., February 1992. Studebaker, P., Moisture analyzers balance sensitivity with price, Control,
Mettes, J. et al., Multipoints PPM moisture measurement using electrolytic November 1993.
cells, ISA Conference, Houston, October 1992. Sutherland, D., Moisture Analyzer Calibration, Meas. Contr., June 1993.
Moisture analyzer helps extend period between dryer column recharging, Yan W. B. and Mallon, T., Breakthrough technology measures parts-per-
Contr. Eng., August 1980. trillion moisture in gases, Chem. Res., February/March 2002.

© 2003 by Béla Lipták