31762100056686

Download as pdf or txt
Download as pdf or txt
You are on page 1of 74

An evaluation of the hydrogen-chlorine flame as an excitation source for flame photometry

by William Kemp Johnston


A thesis submitted to the Graduate Faculty in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY in Chemistry
Montana State University
© Copyright by William Kemp Johnston (1969)
Abstract:
Emission spectra were obtained in the ultraviolet and visible regions of the spectrum, 2500Å to 6500Å,
for silver, calcium, cadmium, cobalt, chromium, copper, iron, magnesium, manganese, molybdenum,
sodium, nickel and strontium by aspirating ethanolic solutions of salts of these elements into a
hydrogen-chlorine flame. The emission intensities were compared to those obtained in a hydrogen-air
flame in order to evaluate the potential of a hydrogen-chlorine excitation source. Of these elements
only silver, copper and magnesium gave emission intensities comparable to those obtained in the air
supported source. Molecular chloride bands were observed in the case of calcium, copper and
magnesium. The copper band was unusually strong compared to emission from this element in the air
flame.
In addition to the elements already mentioned aluminum, gold, potassium, lithium and zinc were fed to
this unusual flame to determine if they would emit in the halogen atmosphere. Emission was not
detected in the 2500Å to 6500Å region.
In the hydrogen-air flame emission from copper, CuH, C2, CH and OH was maximized by varying the
total flow of gas to the burner, the ratio of oxidant flow to total gas flow and the aspiration rate. This
was also done for copper, CuCl, C2, CH and OH in the hydrogen-chlorine flame. Emission intensities
from these species were also determined as a function of the height of observation above the tip of the
burner.
Elements known for their formation of refractory oxides were fed to the hydrogen-chlorine source in an
organic media. While band or line emission was not observed for vanadium, zirconium, niobium,
tantalum, tungsten and the rare-earth elements, a weak band was observed for molybdenum when a
sheathed hydrogen-chlorine flame was used.
AN EVALUATION OF THE HYDROGEN- CHLORINE FlAM-E AS AN
EXCITATION SOURCE FOR FLAME PHOTOMETRY

by
WILLIAM KEMP JOHNSTON

A thesis submitted to the Graduate Faculty in partial


fulfillment of the requirements for the degree

of

DOCTOR OF PHILOSOPHY

in

Chemistry

Approved

Chairman, Examining Committee

Gradiiate Dean [/

MONTANA STATE UNIVERSITY


Bozeman, Montana

March, 1969
I

-iii-

ACKNOWLEDGEMENT

I- would like to thank the Research Corporation, the.National Science

Foundation, and the Montana Agricultural Experiment Station for their

support of this investigation.

I am grateful for the assistance and suggestions I received from Drs.

GoL. Baker, E.E. Frahm and R.A. Woodriff.' A thank you is also extended to
*

Mr. HoD. Via and the Instrument Service for the fabrication of apparatus

vital to my research. ■ ’

The moral support I received from my wife Pat, and my children Anne

Marie and Susan Lynn, is especially appreciated.


-iv~

TABLE OF CONTENTS

• ■ Page
LIST OF T A B L E S ................................ '......... v

LIST OF FIGURES. .'...................................... vi

ABSTRACT ...................... viii

I. INTRODUCTION . « ............................. I

II. STATEMENT OF P R O B L E M ...................................." 3

III. EXPERIMENTAL ............................................ 8


A. Emission Apparatus .............................. 8
B. Burner System. . .........................■ . . . . 10
C. Choice of Solvent. 10
D. Cursory Investigation........... '............. '. 12
E. Maximization of Copper Emission. . . . . . . . . . 18
F. Atomic Absorption Measurements............ .. . . 25
G. Emission from Cg, CH and OH. . .......... .. 25
. H. The Hydrogen-Chlorine Flame and
Refractory Elements ............................. 30

IV. DISCUSSION...................... 42

V. CONCLUSION................................................... 50

VI. ■■ SUGGESTIONS FOR FUTURE WORK....................... 50


. • - •

VII. LITERATURE C I T E D ...................■..................... 52

VIII. APPENDIX ................................................. 55


• ;>
-V-

LIST OF TABLES

Tablg . Page
I. Dissolution Procedures for Cursory Investigation........ 13

II. Emission Intensities in the Hydrogen-Chlorine and


Hydrogen-Air Flames ................ .. 14

III. Response Surface Methods Leading to Maximization of


Emission for Copper, CuH and C u C l .................... 20

IV. Levels of Flame Parameters Giving Optimum Emission


for Copper, CuCl, CuH, Cg, CH and OH. ................. 21

V0 Response Surface Methods Leading to Maximization of.


Emission for Cg, CH and OH............................ 28

VI. Maximization of Emission from Molybdenum in the Sheathed


• Hydrogen-Chlorine F l a m e .............. ' .............. 40

VII. Ratio of Emission Intensities for Silver, Calcium,


Cadmium, Cobalt, Chromium, Copper, Iron, Magnesium,
Manganese, Sodium, Nickel and Strontium in the
Hydrogen-Air and Hydrogen-Chlorine F l a m e s ........ .. . 43
-vi-

LIST OF FIGURES

Figure ' Page


I. Flame Emission Apparatus . -.............................. 9

■ 2. Glass .Burner for the Hydrogen-Chlorine Flame ............ 11


O
3. Emission. Intensity for Copper at 3247.5A as a
Function of the Height Above the Burner Tip in the
Hydrogen-Chlorine and Hydrogen-Air Flames... 23

4. Emission Intensity for CuH at 4280A and CuCl at


4359A as a Function of the Height Above the .Burner
Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames . . . 24

5. Modifications in the Optical Arrangement for


Atomic Absorption...................... 26

6. Percent Absorption for Copper at 3247.5A as a


Function of the Height Above the Burner Tip in the
Hydrogen-Chlorine and Hydrogen-Air Flames......... 27

7. Relative Emission Intensity of the Cg Molecule at


5165A as a Function of the Height Above the Burner
Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames . . . 31

8p Relative Emission Intensity of the CH Radical at


4315A as a Function of the Height Above the Burner
Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames . . . 32

9o Relative Emission Intensity of the OH Radical at


. 3067.8A as a Function of the Height Above the Burner
Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames . . . 33

10. Emission Spectra for 1000 j^g/ml Molybdenum Solution


and for'the Ethanol-Mater Solvent from 4300A to 4500A
in the Sheathed Hydrogen-Chlorine Flame.................. 37

11. Emission Intensity for Molybdenum Band at 4366A as a


Function of the Molybdenum Concentration . . . . . . . . . 38

12. Copper Chloride Band Emission Between 4000A and 430.0A


in the Hydrogen-Chlorine F l a m e ............ .............. 56

13. Background Emission from the Hydrogen-Chlorine Flame


Between 2500A and 4500A. ................................. 57
-vii-

Figure , Page

14. ■ Background Emission from theoHydrogen-Chlorine


Flame Between 4500A and 6 5 0 0 A .......................... 58

15. Background Emission from Absolute Ethanol in the


Hydrogen-Chlorine Flame Between.25OOA and 45OOA . . . . . 59

16. Background Emission from Absolute "Ethanol in the


Hydrogen-Chlorine Flame Between 4500A and 6500A ........ 60

17. Background Emission from Ethyl Acetate in the Hydrogen-


Chlorine Flame Between' 2500A and 4500A.................. 61

18. Background Emission from Ethyl Acetate in the Hydrogen-


Chlorine Flame Between 4500A and 6500A.............. '. . 62

19. Background Emission from Acetonitrileoin the Hydrogen-


Chlorine Flame Between 2500^ and 4500A.................. 63

20. Background Emission from Acetonitrileoin the Hydrogen-


Chlorine Flame Between 4500A and 6500A.................. 64
-viii-

ABSTRACT

Emission spectra were obtained in the ultraviolet and visible regions


■O
of the spectrum, 2500A to 6500A, for silver, calcium, cadmium, cobalt,

chromium, copper, iron, magnesium, manganese, molybdenum, sodium, nickel

and strontium by aspirating ethanolic solutions of salts of these elements

into a hydrogen-chlorine flame. The emission intensities were compared

to those obtained in a hydrogen-air. flame in order to evaluate the poten­

tial of a hydrogen-chlorine excitation .source. Of these elements only

silver, copper and magnesium gave emission intensities comparable to"those

obtained in the air supported source. Molecular chloride bands were

observed in the case of calcium, copper and magnesium. The copper band

was unusually strong compared to emission from this element in the air

flame.

In addition to the elements already mentioned aluminum, gold, pot­

assium, lithium and zinc were fed to this unusual flame to determine if

they would emit in the halogen atmosphere. Emission was riot detected in
O
the 2500A to 6500A region. ■ •

In the hydrogen-air flame emission from copper, CuH, Cg, CH and QH

was maximized by varying the total flow -of gas to the burner, the ratio

of oxidant flow to total gas flow and the aspiration rate. This was also

done for copper, CuCl, Cg, CH and OH in the hydrogen-chlorine flame.

Emission intensities from these species were also determined as a function

of the height of observation above the tip of the burner.

•Elements known for their formation of refractory oxides were fed to

the hydrogen-chlorine source in an organic media. While band or line


emission .was not observed for vanadium, zirconium, niobium, tantalum,

tungsten and the rare-earth elements, a weak band was observed for moIyb

denum when a sheathed.hydrogen-chlorine flame was used.


I

INTRODUCTION

The history of flame photometry dates back to the middle of the 18th

century when chemists described the colors imparted to alcohol and candle

flames by the introduction of metallic salts. However, chemists did not

recognize the qualitative value of flame photometry until 1848. At that

time Herschel in referring to metallic salts, concluded that "the colors

thus communicated by the different bases of flames, afford in many cases

a ready and neat way of detecting extremely minute quantities of them".

In 1870 Janssen first suggested the use of flame excitation sources for

quantitative determinations. Three years later the first real instrument

called a "spectronatromet.re" was developed by Champion, Pellet, and

Grenier. . Capable of determining sodium with an accuracy of 2 to 5% this

instrument was a significant advancement to the field of analytical chem­

istry (14, 19, 33, 39).

In this "spectronatrometre" sample was introduced into the flame by

mean's of a platinum wire. The observed emission was compared to that

obtained from excess sodium chloride placed .in a platinum cone above a

reference flame. Neutral wedges were used to reduce radiation from the

reference flame and thus provide a reproducible means of matching inten-


■ .V , ,

sities from the two sources (33, 39).

Modifications in the method of introducing sample into the flame by

Gouy spectacularly .improved the "spectronatrometre". Gouy aspirated

solution into a city gas-air flame by means of a pneumatic atomizer. The

advantage of this technique was that excitation conditions could.be held

constant thereby improving reproducibility (33,39).


-2-

Further improvements on the instrument designed by Champion, Pellet

and Grenier were made by Lundegarh in 1928. In place of-the insensitive

method of matching intensities visually, emission from an acetylene-air

flame was- analyzed by a spectrograph and recorded on a photographic plate.

The densities of the lines were then measured with a photoelectric dens­

itometer (14,19,33,39).

Flame photometers as they are known today differ from Lundgarh1S

instrument in that they employ sophisticated photomultiplier-amplifier

detector systems in conjunction with strip chart recorders for displaying

the output. The burner systems used in current instruments have not

changed a great deal from.those used by Gouy in 1879. Pneumatic atomizers'

are relatively common although they are considerably more efficient than

those employed by Gouy one hundred years ago.

Because of the advantages of flame photometric analyses (the amount

.of material required, the time per determination and the accuracy of the

method) flame photometry is presently used in a variety of industrial

applications ranging from the determination of lutetium formed by trans­

mutation in irradiated hafnium to the determination of trace amounts of

copper and iron in sugar syrup (17,33).

Because of its wide application in industry a great deal of effort

has been put forth in attempts to improve the technique and make it even

more useful. Research in flame photometry has quite naturally been con­

cerned with increasing both the.number of elements which may be determined

and detection limits.- The following dissertation is likewise directed

toward these goals.


STATEMENT OF THE PROBLEM

The excitation source is obviously the heart of any determination by

flame photometric techniques. In this respect a variety of flames, (hydro

gen-air, hydrogen-oxygen, acetylene-air, acetylene-oxygen, natural gas-air

etc.) have been employed in an attempt to get elements to emit more in­

tense and characteristic radiation. A brief consideration of the theor­

etical aspects involved in flame excitation explains why some unusual gas.

combinations have been investigated. -

A flame is merely a means of providing energy to an atom so as to

thermally excite it to emission. In the process of emitting a photon an

atom moves from a high energy electronic state to a more stable lower

energy state. The intensity of this emission is given by the following

expression (33).

I = L S A N h vN(®a/gQ) e- ^ erg/sec ■ (I)

* where L = flame thickness,


S = area of observation,
A = Einstein transition probability,
h = Plank1S constant,
v = the frequency of the spectral line emitted, ■
3
N = the number of atoms/cm in the flame,
g and g ’= statistical weights of the excited and ground
states respectively,
E^ = excitation energy,
k = Boltzman1s constant,
O -
T = temperature in K.

This is defined as the .probability, that the excited atom will make
the transition in unit time with the emission of a photon. .
-4-

Of these variables the number of atoms and the temperature are of

immediate concern. The others remain constant for a given element with

the exception of the flame thickness and the area of observation which are

functions of the experimental arrangement. From the expression given above

it is apparent that the intensity of emission for a given element can be

increased by either increasing the temperature or the number of atoms or

both. Thus any adjustment in the excitation source which affects either

of these parameters will affect the intensity of emission.

With this theoretical basis investigators have looked for flame

sources which would give higher temperatures. Hence, measurements of

flame temperatures and calculations by thermodynamically minded chemists

have led to the investigation of unusually high energy flames such as

oxy-cyanogen (4,5,6) , fluorine-hydrogen (17.,21,37,50) , ozone-cyanogen (54),

ozone-hydrogen (53), perchlorylfluoride-hydrogen (45,46), and even atomic

hydrogen (35,36). Although this approach has been moderately successful,

it nevertheless has left something to be desired. In equation (I) relating

emission intensity to temperature one cannot overlook the dependence on

free atom concentration. As one increases the temperature of a source

the degree of ionization likewise increases. In the case of alkali and

alkaline-earth elements this effect is quite noticable.

. Chemical reactions involving free, atoms and combustion products or

reactants may decrease the free atom concentration and.also adversely

affect emission intensities. Since the analyte is generally introduced

as an aqueous or organic solution, possible reactions involving the sol­

vent must also be considered.


i

-5-

Because of these many complications a detailed explanation of flame

processes has been difficult. Although somewhat simplified, the following

describes in sequence what is known to take place when a solution of a

metallic salt is aspirated into a flame.

As a sample droplet enters the reaction zone of a flame it acquires

energy primarily through heat transfer so as to evaporate solvent assoc-

iated with the sample. Depending on the energy available the resulting

particle may then vaporize and dissociate to give free atoms, molecules

or radicals. These in turn can react with combustion products, reactants

or solvent to give additional species (1,2,14,19,31,33,39,42). . The part­

icular path which is followed quite naturally depends on the element

involved and the characteristics of the excitation source. In the case of

titanium, vanadium, zirconium, niobium, molybdenum, hafnium, tantalum,

tungsten and the rare-earths, in the usual oxygen supported flames, stable

oxides are produced which do not emit useful radiation. Hence the deter­

mination of these elements by flame photometric techniques is troublesome.

Many of the high temperature flames mentioned earlier were used in

an attempt to dissociate refractory compounds. However, Fassel and assoc­

iates were more successful in approaching the problem with the intention of

creating a reducing atmosphere to promote dissociation and thus obtain

free atoms and atomic emission. To accomplish this they used fuel-rich

oxy-acetylene flames along with organic solvents. Their success added


2
support to the concept known as chemiluminescence and provided the

_ 2
Chemical processes leading directly to the formation of an abnormally
high population in the excited state (29).
-6-

analytical chemist with an improved method for determining rare earths,

vanadium, niobium, rhenium, titanium, molybdenum and tungsten (18,23,24,

25,26,43).

As a substitute to the fuel, rich oxy-acetylene flame the hydrogen-

chlorine flame was suggested and is the subject of this dissertation. Use

of chlorine in conjunction-with organic solvents was expected to produce

a reducing atmosphere (47) suitable for the determination of many refrac­

tory elements. By substituting chlorine for oxygen one would not expect

formation of stable oxides. However, one might expect metal chlorides to

form in place of oxides (13). A comparison of the limited data on free

energies and enthalpies of formation for metal oxides and chlorides showed

that in a number of cases the chlorides were less stable (10,11,27,32,34,

44,55,58). Hence, from a thermodynamic point of view use of a chlorine

supported flame merited investigation.

In the' event that the hydrogen-chlorine source lacked sufficient

energy to dissociate even chlorides it was felt that molecular emission

bands from these compounds might be of interest. Hydroxide and oxide

bands for calcium and magnesium are intense enough and sufficiently narrow

in band width so that they are commonly used in analytical determinations

(14,33,39). If chloride bands shared these characteristics, they could

also be useful analytically.

At this point in the discussion one may wonder why chlorine was

chosen rather than one of the other halogens. ' Fluorine, bromine and

iodine all had disadvantages which eliminated them from consideration.

Bromine and iodine were known to produce low energy flames. Fluorine, on
-7-

the other hand, produces a very hot flame when combined with hydrogen

(17,21,37,50). However, this particular halogen was not-used because of

problems involved in handling it.safely.

Due to thermodynamic considerations hydrogen, acetylene and cyanogen

were the only fuels which seemed worthwhile investigating. Acetylene,

when combined-with chlorine, produced an extremely bright flame due to

incandesence from carbon particles. Although this type of atmosphere was

desirable, the high background emission was a disadvantage. As an added

complication chlorine and acetylene were known to ignite spontaneously

at room temperature. References to a cyanogen-chlorine flame could not

be found in the literature. Therefore, cyanogen was set aside for poss­

ible investigation at a later date.

As a result the hydrogen-chlorine flame was chosen for evaluation

as an excitation source in flame photometry. The low temperature of this

flame was recognized as a distinct disadvantage. Nevertheless it was

anticipated that elements troubled by refractory compound formation in

the conventional oxygen supported flames would emit useful radiation in

this particular source.


EXPERIMENTAL

Emission Apparatus

The basic instrument used in this investigation was assembled from

components from several manufacturers and as such was considered more

versatile and far superior to any commercial instrument available. The

heart of the instrument was a model 1700-111 3/4 meter Spex Industries Inc.

monochrometer with a 1200 grooves/mm, 102x102 mm Bausch & Lomb grating

blazed for 3000 A, a continuously variable, servo-controlled, automatic

scanning control and bilaterally opening to 3 mm wide by 50 mm high

straight slits. . The entrance slit also included a stepped fishtail and

Hartman diaphram shutter. The detector was a R106 photomultiplier tube

having an S-19 spectral response from the Hamamatsu TV Company Ltd. The

power supply for the photomultiplier was a model 412B from the John Fluke

Manufacturing Company, Inc. The output from the photomultiplier was fed

to a Princeton Applied Research Corporation type A preamplifier and from

there to a model HR-8 lock-in amplifier from this same company. A model

BZ-I mechanical light chopper also from the Princeton Applied Research

Corp. was used to modulate the light beam coming from the flame. The out­

put from the amplifier was finally displayed on a Honeywell Lab 19 re­

corder operated on the 0-10 millivolt range. This experimental set up is

illustrated schematically in Figure I. as a matter of convenience.


SPEX 1700-111 PAR KR-8 HONEYWELL

Pd > hj
3/4 meter f/6.8 R106 Amplifier
> Czerny Turner
Spectrometer
PM with Type A
Preamplifier Lab-19
Recorder

Burner

FLUKE
Model 412B
Power Supply

Figure I. Flame Emission Apparatus


— 3.0—

Burner System

Finding a burner which would withstand the corrosive atmosphere '

created by the chlorine and yet maintain a stable flame was somewhat of a

problem. The ordinary Beckman burners commonly used in flame photometry

did not last long when hydrogen and chlorine gas mixtures were used in

themo An all glass burner shown in Figure 2 was finally used to support

the chlorine flame. An adjustable slide assembly with metric scale and

vernier from Velmex, Inc. was adapted to the burner mount to permit ad­

justment of the horizontal and vertical position of the flame with res­

pect to the optical axis of the instrument. A Sage model 249-2 constant

speed syringe pump was used to feed solution to the flame while gas flow

rates were monitored by Gilmont flow meters calibrated with a wet test

flow meter. The automatic syringe pump was found to be very useful in

that the rate at which sample was fed to the flame could.be controlled

without varying the flow of oxidant. With most commercial burners it is

not possible to do this.

Choice of Solvent

Of all the factors influencing emission intensities the nature of

the solvent has been investigated more intensively than any other (2,19,

23,31,48). Vaporization of an aersol droplet is an important step in

the sequence of processes leading to emission. Since vaporization depends

on surface tension and temperature, emission intensities are affected by

the choice of solvent. Although aqueous solvents are used a great deal,

they are nevertheless undesirable since they cool the flame. Organic
70 mm

male Luer fitting

Figure 2. Glass Burner for Hydrogen-Chlorine Flame

analyte •
-12-

solvents on the other hand do not have this effect (19,23,31,48). However,

the limited solubility of metallic salts in organic solvents severely

restricts their use. As far as organic solvents go, ethanol is generally

considered the best all around choice. This particular solvent burns

smoothly in the commonly used flames and yet many salts are moderately

soluble in it. In this investigation ethanol was used whenever possible

although acetonitrile, ethyl acetate and methyl isobutyl ketone were used

in selected cases.

Cursory Investigation

At the beginning of this investigation it was felt that a study of

the emission from a variety of elements would be worthwhile in view of t h e •

limited data reported on .a hydrogen-chlorine excitation source. Hence,

solutions of readily available compounds were prepared as indicated in

Table I. "Analysis of the emission spectra was achieved by aspirating

these solutions into a hydrogen-chlorine flame and recording the amplifier


O O

output as one scanned the spectrum from 2500A- to 6500A . A summary of

these scans is presented in Table II, pages 14-17.

The intensities given represent the potential drop across a load

resistor (470 K, 10%) connected between the anode of the photomultiplier

tube and ground. Emission above background was measured as opposed to

measuring over-all emission intensities.


I

-13-

Table I. D i s s o l u t i o n Procedures for Curs o r y Investigation.

Element 'Dissolution Procedure

Aluminum Anhydrous AlClg (Baker & Adamson reagent)'dissolved in


absolute ethanol.
Calcium CaClg'ZHgO ("Baker Analyzed" reagent) dissolved in absolute
ethanol.
Cadmium CdCl2 * ("Baker Analyzed" reagent) dissolved in absolute
ethanol.
Cobalt Metal (Matheson, Coleman & Bell) dissolved in cone. HCl,
evaporated to near dryness, followed by dissolution in
absolute ethanol.
Copper CuClg-ZHgO (Baker & Adamson reagent) dissolved in absolute
ethanol.
Chromium Metal (Fisher, 98%) dissolved in cone. HCl, evaporated to
near dryness, followed by dissolution in absolute ethanol.
Iron Metal ("Baker Analyzed" reagent) dissolved in cone. HCl,
evaporated to near dryness, followed by dissolution in
absolute ethanol.
Gold HAuCl^'SHgO ("Baker Analyzed" reagent) dissolved in absolute
ethanol.
Lithium LiCl ("Baker Analyzed" reagent) dissolved in absolute
ethanol. ' ' .
'

Magnesium Metal (Baker & Adamson reagent) dissolved in cone. HCl,


evaporated to near dryness, followed by dissolution in
. absolute ethanol.
1 ,

Manganese Metal (Baker.& Adamson reagent) dissolved in cone. HCl,


evaporated to near dryness, followed by dissolution in
} absolute ethanol.
Nickel Metal (Baker & Adamson reagent) dissolved in cone. HCl,
evaporated to near dryness, followed by dissolution in
absolute ethanol.
Silver AgNOg (Baker & Adamson reagent) dissolved in absolute
ethanol.
Strontium SrBrg-GH^O (Merck & Co. reagent) dissolved in absolute
ethanol.
Zinc' Metal (Baker & Adamson reagent) dissolved in cone. HCl, eva­
porated to near dryness, followed by dissolution in absolute
ethanol..
— 14-

Table II. Emission Intensities in the Hydrogen-Chlorine and


Hydrogen-Air Flames.

Element - Concentration Emitting Intensities Above Background


Wavelength in A Species in mv.
H 2 -CI2 H2~Air

Ag 500 Mg/ml
3281 Ag 0.96 1.90
3383 Ag 1.10 2.30
Al 1000 Mg/ml
3944 Al 0.00 0.00
3962 Al 0.00 0.00
4842 AlO 0.00 0.00
4866 AlO 0.00 0.00
4888 AlO 0.00 0.00
5102 AlO 0.00 0.00
5123 AlO 0.00 0.00
Au- 972.mg/ml
2676 Au 0.00 0.00
Ca 250 Mg/ml
4227 Ca 0.06 9.40
5539,
5542} CaOH 0.00 9.95
55b /
5810 CaCl - 0.03 0.00
5934 CaCl 0.31 0.20
6185 CaCl 0.11 0.60 .
6212 CaCl 0.08 0.80
Cd 1000 Mg/ml
3261 ^ Cd 0.12 0.94
Co 1000 .Mg/ml
3413 Co 0.48 1.12
3433 Co 0.20 0.42
3453 Co 0.50 1.40
3468 Co 0.30 0.86
3508 Co 0.41 1.04
3527 Co 0.50 1.54
3575 Co 0.28 0.52
3595 Co 0.24 0.20
3630 Co 0.19 0.06
3650 Co 0.25 0.04
-15-

Table II. (continued)

Element - Concentration Emitting Intensities Above Background


Wavelength in A Species in mv.
Hg-Clg Hg-Air

Co 1000 ng/ml
3842 Co 0.15 0.26
3871 Co 0.30 0.70
3892 Co 0.23 0.26
3992 Co 0.11 0.20
4030 Co 0.00 0.04.
4118 Co 0.09 0.20
Cr 1000 ^g/ml
3579 ' Cr 0.34 1.18
3594 Cr 0.24 1.10
3605 Cr 0.23 0.86
3891 (broad) 0.14 0.00
4254 Cr 0.57 5.00
4275 Cr 0.56 ' 4.00
4290 Cr. 0.54 3.50
Cu ' 500 ng/ml
3248 Cu 0.72 1.52
3274 Cu - 0.60 1.32
4265 CuCl 1.90 0.00
4280 CuH 0.00 0.54
4293 CuCl 4.00 0.00
4328 CuH 0.00 0.20
4334 CuCl 2.10 0.00
' 4359 CuCl 6.00 0.00
4416 ' CuCl 2.40 0.00
.4446- ■ ■ CuCl ■ 4.00 0.00
' 5240 V ‘ CuOH 0.00 0.10
5300 CuOH 0.00 0.08
5370 ' CuOH 0.00 . 0.09
Fe 1000 ng/ml
3441 V 0.07 0.65
3477 Fe 0.09 0.20
3491 • Fe 0.15 0.15
3570 Fe 0.13 0.15
3581 Fe 0.05 0.30
3648 Fe 0.13 0.15
3706 Fe 0.07 0.15
-16-

Table II. (continued)

Element - Concentration Emitting Intensities Above Background


Wavelength in A Species in m v .
H2-Cl2 Hg-Air

Fe 1000 ng/ml
3720 Fe 0.22 2.20
3737 Fe 0.22 1.85
3746 Fe 0.23 1.65
3823 Fe 0.11 0.85
' 3860 Fe 0.26 2.25
3886 Fe 0.14 0.80
3900 Fe 0.16 0.45
3930 Fe 0.09 0.60
5647 FeO 0.00 0.18
5819 FeO 0.00 0.13
5868- FeO 0.00 0.21
5903 FeO 0.00 0.26
• 5975 FeO 0.00 0.00
6095 FeO . 0.00 .' 0.01
6109 FeO 0.02 0.03
6181 • FeO 0.02 0.02
6219 FeO 0.00 0.00
K 1000 ng/ml
4044V 0.00 3.90
K
4047J
Li 1000 /ml
6104 . ' Li 0.00 0.10
Mg .200 ng/ml
2852 Mg 0.12 0.36
3625" • MgOH 0.16 0.45
' 3687 . ■ • ■ -• MgCl 1.20 .1.10
3691 , • ; MgCl 1.10 1.60
3700 MgOH 0.56 1.85
3745 MgCl 1.38 0.90
3749 MgCl - 1.42 0.90
3755 MgCl ' 1.32 0.80
3761 . ■ MgCl . 1.28 0.80
3808 MgCl 0.98 1.40
3810 ■ MgOH 0.76 1.50
3871 MgCl . 0.40 0.70
3880 MgOH 0.22 1.40
-17-

Table II. (continued)

Element - Concentration Emitting Intensities Above Background


Wavelength in A Species in mv.
Hz-Cl2 Hz-Air

Mn 1000 ng/ml
3500-4200 (board) MnOH 1.20 1.60
403 Iv
4033 r Mn • 1.80 31.50
4035
5159 MnO 0.00 0.45
5192 MnO 0.00 0.38
5360 MnO 0.00 0.55
5390 MnO 0.00 0.86
5632 MnO 0.00 0.46
5790 MnO 0.00 0.90
5810 MhO 0.00 0.62
Na 25 ng/ml
5890 Na
5896 2.00 71.00
Ni 1000 ng/ml
3371 Ni 0.12 0.45
3394 Ni 0.18 0.60
3414 Ni ' 0.44 1.40
•3438 Ni 0.22 0.45
3449 Ni - 0.26 0.65
3460 Ni 0.46 1.40
3491 Ni 0.36 0.94
3514 Ni 0.40 1.25
3522 Ni 0.69 1.95
3569 Ni 0.27 0.40
3619 Ni 0.42 0.80
3854 -Ni 0.44 0.20
Sr 250 Mg/ml
4078 Sr 0.00 0.07
4216 Sr 0.00 0.06
4607 Sr 0.05 20.00
6035 SrOH 0.00 1.00
6059 . SrOH 0.00 .1.50
6359 SrCl 0.01 0.00

Zn 1000 yg/ml
2500-6500 0.00 0.00
-18-

Instrument Settings for Table II.


Flow Rates: Oxidant 1.8 1/min Fuel 7.2 1/min Aspiration Rate: 0.746ml/faiin
Entrance Slit: 750u Exit Slit:750u Slit Height: 2mm
PM Voltage: 800V Time Constant: I sec.
Region Viewed: 2.5-7.5 cm above burner tip. Solvent: absolute ethanol.

The results of the cursory investigation summarized in table II seem

to indicate that the hydrogen-chlorine flame could be used in the determin­

ation of silver, copper and magnesium. Emission intensities for these

elements were comparable to intensities obtained' in the hydrogen-air flame.

In the case of copper the CuCl band at 4359A was four times as intense as

the strongest line or band obtained in the air supported flame.

On the other hand, although emission from calcium, cadmium, cobalt, ■

chromium, iron, manganese, sodium and nickel was observed in the unusual

source, it wasn’t equivalent in strength to that observed in the air supp­

orted flame. In the case o f lithium, potassium and strontium the hydrogen-

air flame produced very intense line emission while in the chlorine flame

emission was not observed.

The sodium doublet at 5890.CA and 5895.9%.was observed in many samples

as a result of contamination. Although the intensity of this emission was

considerably less than that observed in the corresponding hydrogen-air

flame, it is worth noting, as Slootmaekers and Van Tiggelen did not observe

emission from sodium in the hydrogen-chlorine flame (50).

Maximization of Copper Emission

The greater intensity of the copper chloride band in the chlorine

flame compared to that of the atomic copper line in the air supported

flame was impressive. However, at this point a great deal of significance


-19-

could not be placed on this comparison. For the given data the levels of

the total flow rate of gas to the burner, the oxidant to-fuel ratio and the

aspiration rate were chosen in order to give a stable flame. Since the

particular choice of flame parameters could have led to false conclusions

about the relative strength of emission for copper chloride and copper in

the hydrogen-chlorine .and hydrogen-air flames, an experiment was designed

to maximize emission from CuCl and copper independently.

In this approach, the response surface design described by Box and

Wilson (9) was adopted since it has been used successfully on many occas­

ions in related problems of optimization in both flame emission and atomic

absorption spectroscopy (15,41,48,49). In essence factorial experimenta­

tion is used to calculate linear equations relating a particular response

to experimental variables. These equations indicate the direction one must

move to approach a maximum. The levels of the variables are then shifted

in the direction indicated by the sign of the coefficients of the linear

equations and new equations are calculated from data taken at these levels.

The process is repeated until the coefficients go to zero, indicating, a

maximum has been reached, or until the sign of the coefficients changes

indicating one has passed the maximum. At this point central composite

rotatable designs have been suggested for accurately specifying the maxi­

mum (16). However, in the experiment at hand this technique could not be

used because the rate at which solution was fed to the flame could not be

varied continuously, a necessary condition for a rotatable design. As a

substitute a simple 3x3 factorial procedure was used to specify optimum

excitation conditions.
Table III. Response Surface Methods Leading to Maximization of Emission for Copper, CuH and CuCl.

Xi X2 X3
Experiment Total Oxidant Aspiration
Number Gas Flow Rate Calculated Equations
Flow Total
Flow
1 /min ml/min

Copper Emission in the Hydrogen-Chlorine Flame at: 3247.5 A


I 9.0-. 9.5 0.10-0.15 1.00 -3.00 I = 0.369 + 0.045Xi 0.495X2 +
- 1 .149X 3
2 9.5-10.0 0.05-0.08 1.00 -3.00 I = 0.668 - 0.027Xi “ 0.019X2 + 0 .097X 3

Copper Emission in the Hydrogen-Air Flame at 3247 .5 A

-OZ-
.1 . 8.8- 9.2 0.16-0.20 0.421-1.26 I = 0.505 0.085Xi - 0.OSOX2 + 0.240X3
■2 9.4- 9.8 0.06-0.10 1.00 -3.00 I = 1.225 + O.OOXi - 0.073X2 + 0.080X3
CuCl Emission in the Hydrogen-Chlorine Flame al: 4359 A
I 9.0- 9.5 0.10-0.15 1.00 -3.00 I = 7.188 + 0.213Xi 1.188X2 + 2 .313X 3
2 9.5-10.0 0.15-0.20 1.00 -3.00 I = 10.13 + 0.075Xi + I.45OX2 + 4.675X3
3 10.0-10.5 0.25-0.30 1.00 -3.00 I = 13.48 + 0.675Xi - CLZOOX2
+ 7 .800X 3

CuH Emission in the Hydrogen-Air Flame at 4280 A


I 8 .8 - 9.2 0.16-0.24 0.421-1.26 I = 0.022 + 0.002Xi - 0.OOZX2 + 0 .012X 3
2 9.4- 9.8 0.08-0.12 1.00 -3.00 I = 0.087 + O.OOIXi - 0.OOSX2
+ 0 .046X 3
3 9.8-10.0 0.05-0.07 3.06 -4.59 I = 0.240 - O.OO8X 1 - 0.006X2 + 0 .055X 3
Table IV. Levels of Flame Parameters. .Giving Optimum Emission Intensities for Copper, C ud,
CuH, Og, CH and OH.

Emitting Wavelength Flame Total Oxidant Aspiration Maximum


Species in- ' :■ Employed Gas Flow Rate Emission
A • Flow Total Intensity
Flow ■
1/min ml/min mv

Cu 3247.5 , • ; H2-Cl2 9.50 0.05 . 4.59 1.02

Cu' 3247.5 . H2-Air


9.00 0.10 4.59 1.46
8.80 0.06 3.06

CuCl 4359 H2-Cl2 11.00 0.25 3.00 25.25

CuH 4280 H 2-Air 9.00 0.04 4.59 0.38


C2 5165 H2-Cl2 11.00 0.30 3.00 19.60

C2 5165 H2-Air . 12,50 0.38 3.00 4.40

CH •4315 H2-Cl2 10.00 0.30 3.00 9.70

CH 4315 H2-Air 12.50 ■ 0.40 3.00 8.00

OH ' 3067.8 H2-Cl2 11.00 ' 0.15 3.PO 14.20


OH 3067.8 H2-Air 15.00 0.15 . '0.421 15.20
14.00
i

-22-

In table III the levels of the flame parameters and the linear equa­

tions leading to the determination of maxima are given for copper and CuH

in the hydrogen-air flame and for copper and CuCl in the hydrogen-chlorine

flame. Table IV, page 21, on the other hand, summarizes the results of ■

the final factorial experiment to accurately determine optimum excitation

conditions. In addition to giving the levels of the flame parameters

(total flow of gas to the burner, ratio of oxidant flow to total gas flow

and aspiration rate) table IV also gives the maximum emission intensities

above background in millivolts for copper and CuH in the hydrogen-air

flame and for copper and CuCl in the hydrogen-chlorine' flame.

A study was also made of the variation in emission'intensities for

copper, CuCl and CuH as the flame region being observed was varied. To.

accomplish this the burner was lowered with respect to the optical axis

of the instrument as copper solutions were fed to.the chlorine and air

supported flames while the emission intensities of the copper line and

CuH and CuCl bands were monitored. The flame parameters were adjusted so

as to give maximum emission as determined in the optimization procedures

summarized in.-.tables 111 and IV, pages 20 and 21 respectively.

Unlike the maximization date where a region above the burner tip

from 2.5 cm to 7.5 cm was observed, this profile data-was taken by loca­

ting the flame-exactly at the focal '.point, of the condensing lens. A light

shield with a 4 mm diameter- hole was used to. restrict emission to the

region of the flame desired. 'The results of this study are illustrated

on figures 3 and 4 for copper and GuH, CuCl respectively. Data from the

air and chlorine flames are shown on the same illustration to facilitate
comparing the two flames. '
O

O O
O 8 ° ® Q O
O
o o O O O
O
Relative Emission Intensities

O 0 O 0
O

O
O O O Hydrogen-Chlorine Flame

O Hydrogen-Air Flame

I_________ I________ 2_________ s_________I_________5_________j_________I________ r_________I


20 40 60 80 100
Height Above the Burner Tip in mm

Figure 3 Emission Intensity for Copper at 3247.5A as a Function of the Height Above the Burner Tip
in the Hydrogen-Chlorine and Hydrogen-Air Flames.
O
O
O O
O
O
Relative Emission Intensities

O
O
O
O

O
O (3 Hydrogen-Chlorine Flame

O
Q O Hydrogen-Air Flame
O
I
O M
f
O

LGoo O o O O O O O O O O o O o O o O o
E- - s— _L_______ I_______ !_______ _______ ________ A_______I____
20 40 60 80
' I
100
Height Above the Burner Tip in mm

Figure 4. Emission Intensity for CuH at 428OA and CuCl at 4359A as a Function of the Height Above
the Burner Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames0
-25-

Atomic Absorption Measurements

Since atomic absorption measurements give information on ground state

concentrations and are relatively independent of flame temperatures, it

was felt that these measurements would be useful in the evaluation of the

hydrogen-chlorine flame. If the difference between the atomic emission

from copper in the air and chlorine flames resulted from differences in

flame temperatures and not differences in copper concentration, atomic

absorption would point this out.

Hence, modifications in the apparatus used for emission studies were

made as illustrated in figure 5 and absorption data were taken at various

positions in the flame for copper. Figure 6 illustrates these data for

both the hydrogen-chlorine and hydrogen-air flames.

The hollow cathode tube employed was a Westinghouse type WL-22603*

while the power supply was a Jarrel-Ash model 82-139. Although this

particular power supply can be operated on both high intensity pulsed and

straight D.C. modes, the D.C. mode was used exclusively. The Princeton

Applied Research Corporation mechanical light chopper was placed between

the flame and the light source to modulate the light so that it could be.

picked up by the A.C, amplifier which was used.

Emission from Cg, CH and OH

Since- the relative emission intensities of the C2 molecule, the CH

radical and the OH radical, commonly observed in flame excitation using

organic solvents, reveal the chemical nature of the excitation source,

maximization procedures identical to those used for copper emission were

adopted to study the emission from these species in the chlorine supported
JARREL-ASH
Model 82-139
Power Supply

Burner

SPEX INDUSTRIES
WESTINGHOUSE
Hollow Cathode Spectrometer

26-
BZ-I

Figure 5 Modifications in the Opti c a l A r r a n g e m e n t for A t o m i c A b s o r p t i o n


1007, o o o o o o o o o o o o o O

O °
O O o
O
Percent Absorption

O O

O O

'Ll-
O Hydrogen-Chlorine Flame
O O
O Hydrogen-Air Flame
O O O

07.

I J I
40 60 80 100
Height Above the Burner Tip in mm

Figure 6 e Percent Absorption for Copper at 3247.5A as a Function of the Height Above the Burner Tip
in the Hydrogen-Chlorine and Hydrogen-Air Flames.
Table V. Response Surface Methods Leading to Maximization of Emission for Cg, CH and OH

Xi X2 X3

Experiment Total Oxidant Aspiration Calculated Equations


Number Gas Flow Rate
Total
Flow
1/min ml/min

• Cg Emission in the Hydrogen-Chlorine Flame at 5165; a


I 9.0- 9.5 0.10-0.15 1.00-3.00 I = 2.25 + O.OSX-l + 0.60X2 + I.OSX3
2 9.5-10.0 0.15-0.20 1.00-3.00 I = 4.97 + 0. 47X1 + 1.36X2 + 2.13X3
3 10.0-10.5 0.25-0.30 1.00-3.00 I 6.63 + I. 65X1 + 2.IOX2 + 4.6OX3

C2 Emission in the
Hydrogen-Air Flame at 5165 A
I . 10.0-10.5 0.20-0.25 1.00-3.00 I = 1.67 + 0.04Xi + 0.13X2 + 0.84X3
2 11.0-11.5 0.35-0.40 1.00-3.00 I = 2.58 4-0. OSX1 + o.oox2+ 1.34X3

CH Emission in the Hydrogen-Chlorine Flame at 4315- A


I 10.0-10.5 0.25-0.30 1.00-3.00 ' I = 4. 66 - 0. O2X 1 +' 0.09Xg + 2.72X3
CH Emission in the Hydrogen-Air Flame at 4315 A
I 11.5-12.5 0.35-0.40 1.00-3.00 I 5.69 + O-SlX1 + 0.39X2 + 2.09X3
Table V. (continued)

xI X2 x3

Experiment Total Oxidant Aspiration Calculated Equations


Number Gas Flow Rate
Total
Flow
1/min ml/min

OH Emission in the Hydrogen-Chlorine Flame at 3067.8 A


I 10.0-10.5 0.25-0.30 1.00-3.00 I = 6.88 + 0.28Xi - I. 02 x 2 + 2 .75X 3
2 10.0-10.5 0 .10-0.20 ■ 1.00-3.00 I = 10.5 + 0.20Xi - 1.30X2 + I.2OX3

OH Emission in the Hydrogen-Air Flame at 3067.8 A

T-I
3
O
I 11 .0 - 12.0 0.30-0.40 1.00-3.00 I = 9.88 .+ 0.42Xi - 1.19X2 +

CO
2 12.0-12.5 0.20-0.30 ' 1.00-3.00 I = 12.8 + 0.35Xi - 0.20X2 - 0 .03X 3
3 13.0-14.0 0.10-0.15 1 .00-2.00 I = 12.9, + 0.29Xi + 0.69X2 •- 0 .44X 3
-30— ■

flame. Absolute ethanol was used as the analyte in the investigation. It

was anticipated that data on Cg 5 CH and OH would show a relation to that

obtained for copper and thus aid in the evaluation of the chlorine suppor­

ted excitation source.

Table V gives the levels of the flame parameters chosen and the linear

equations relating emission intensities to these parameters. The results

of the final maximization by 3x3 factorial methods are presented in table

IV, page 21, along with similar results for copper which were described

■earlier. The relations between emission intensities and flame positions

are shown in figure 7, 8 , and 9 for Cg 5 CH and OH respectively. Once

again data taken in the air and chlorine flames were placed in the same

figure so that it would be easy to compare the chlorine flame to. the

more conventional hydrogen-air flame.

The Hydrogen-Chlorine Flame and Refractory Elements

At the beginning of this investigation little thought was given to the

use of the hydrogen-chlorine flame in the determination of copper or any

other element readily determined in more conventional flames'. The pri­

mary concern, was;the determination of elements which form refractory oxides


. •
in conventional flames. In this respect an evaluation of the emission

from titanium, vanadium, zirconium, niobium, molybdenum^ lanthanum^ tanta-•

num, tungsten, and the rare earths was sought.

Because rare earths were used by Fassel in his illustrations of the

value of fuel rich oxy-acetylene flames, they were an obvious beginning

for the evaluation- of the chlorine flame. • Following the procedure outlined

by Fassel.(23) rare-earth oxides were converted to perchlorates and


O o
O
O
Relative Emission Intensity

O Hydrogen-Chlorine Flame

O Hydrogen-Air Flame

31-
O
O

O O

8 O
O
O
O

g I I !
° Q 000000
40 60 80 100
Height Above the Burner Tip in mm

Figure 7 Relative Emission Intensity of the C2 Molecule at 5165A as a Function of the Height Above
the Burner Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames0
O
O o Q Hydrogen-Chlorine Flame
Relative Emission Intensity

O O O O Hydrogen-Air Flame
Q
O O
O
O
O
O
O U3
ND
O
O O
O
O
O O
O
O O
O O
O

40
I
60
I
O

80
O
8 89
100
Height Above the Burner Tip in mm

Figure 8 „ Relative Emission Intensity of the CH Radical at 4315A as a Function of the Height Above
the Burner Tip in the Hydrogen-Chlorine and Hydrogen-Air Flames,
Hydrogen-Chlorine Flame

Hydrogen-Air Flame

i
w

Height Above the Burner Tip in mm

Figure 9e Relative Emission Intensity of the OH Radical at 3067.SA as a Function of the Height Above
the Burner Tip in the Hydrogen-Chlorine and Hydrogen=Air Flames.
-34-

dissolved in absolute ethanol by the following procedure. Desired amounts


3
of ignited oxides were dissolved in 50 ml of 1:1 perchloric acid (72%),

and water. The resulting solutions were evaporated to near dryness and

allowed to cool whereupon the liquid solidified. After further cooling to

room temperature, the perchlorates were dissolved in absolute ethanol and

diluted to volume to give the following weight percent solutions: yttrium,

praseodymium, neodymium, holmium and lutetium - 0.1%; lanthanum - 1.0%;

and gadolinium - 0.5%. These solutions were then aspirated into chlorine

and air supported flames and the emission intensities recorded as the

spectrum was scanned from 2500A to 6500A. -

Peaks attributable to the rare earths were not observed for any of

the compounds investigated nor were bands from chlorides or oxides seen.

To further modify the atmosphere of the flame acetonitrile was used as the

solvent in place of ethanol.

Rare-earth perchlorate solutions were prepared as were the ethanol

solutions and aspirated into the hydrogen-chlorine flame. Once again when

the spectrum was scanned from 2500A to 6500A, emission from the rare-earths

was not detected. ■ However, it may be interesting to note that very intense

bands from the CN .radical-were observed arising from the solvent, aceto-

nitrile. Bands from this species are illustrated in the appendix in

recordings of the background emission obtained when acetonitrile was intro­

duced into the chlorine flame.

These compounds were purchased from Matheson, Coleman & Bell and were
reported to be 99.9% pure.
-35-

Directing attention to molybdenum, tungsten and vanadium, 1000 ^g/ml

aqueous solutions of these elements were'prepared from (NH4 )5M 07O24 • ,

(NH4)5Wy024* 6H 2O and NH3VO 3 . The solutions were aspirated into the H 2 -CI2

flame and the. spectrum scanned. Emission from molybdenum, tungsten and

vanadium was not detected. Since Fassel failed to get rare-earths to emit

from aqueous solutions even in the fuel rich oxy-acetylene flame, this

result was not entirely unexpected.

Molybdenum and vanadium were brought into organic solution by solvent

extraction techniques described in the literature (38,40,51). Ammonium

pyrrolidine dithiocarbamate,. APDG, was used as the chelating agent and

methyl isobutyl ketone (MIBK) as the solvent. Vanadium was also extracted

with 8 -hydroxyquinoline into methyl isobutyl ketone as was tungsten. A

cupferron ethyl acetate extraction was also used in the case of vanadium.

Neither of these elaborate procedures produced solutions whereby atomic

or band emission could be detected in the hydrogen-chlorine flame. MIBK

and ethyl acetate also did n o t .burn satisfactorily in the chlorine flame.

Excessive carbon formation at the tip of the burner was a constant problem

in that this caused both the flame and aspiration rate to fluctuate,

Other, elements which do not emit useful spectra in oxygen supported

flames are niobium and zirconium. In an attempt to obtain spectra from

these elements halides were dissolved in absolute ethanol which had had

dry nitrogen bubbled through it. Tungsten was also treated in a similar

manner. For this portion to the investigation samples of NbCI3 from K & K

Laboratories, Inc., ZrCl4 from Pfaltx & Bauer, Inc. and WBrg from this same

company were weighed out in a Dri Lab model HE-43-2 dry box from the
I

-36-

Vacuum/ atmospheres Corporation. WBr^ formed a precipitate when added to

the ethanol necessitating the addition of 5 ml of 50% HF'solution. The

spectrum from 2500A to 6500A was then scanned as solutions of niobium,

zirconium and tungsten were fed to the hydrogen-chlorine flame. In the

recordings obtained no evidence for emission from these elements was

detected.

Since emission from titanium hadn't been evaluated as yet, a 2000

jLtg/ml solution of this element in 1:1 ethanol-water was fed to the chlorine

flame. This solution was prepared by dissolving titanium in concentrated

HCl, evaporating to near dryness and dissolving the remaining solution in

absolute ethanol. As in the case of the other refractory elements,

emission characteristic of titanium was not observed.

At this point it was apparent that elements known for forming refrac­

tory oxides would not emit radiation in the UV-visible region of the

spectrum when aspirated into the hydrogen-chlorine flame. Although the

atmosphere of this flame was more conducive to reduction, indicated by

the increased Cg and CH emission, (see appendix for recordings of back­

ground emission in the hydrogen-chlorine flame) refractory elements never­

theless did not emit the desired atomic line spectra. It was felt that

turbulence in the hydrogen-chlorine flame may have brought oxygen in from

the surroundings and defeated the whole purpose of using chlorine as the

oxidant. For this reason the glass burner which had been employed through­

out the investigation was modified by putting a glass sheath around the

tip. Chlorine was passed through the sheath to shield the flame from the

atmosphere.
Emission Intensity in — 37“

Solvent
EtOH-ILO

4300
Wavelength in R

Figure 10«, Recording of Emission Spectra for Molybdenum in the


Sheathed Hydrogen-Chlorine Flame.
Emission Intensity in mv

38-
Molybdenum Concentration in p,g/ml

Figure 11. Emission Intensity of Molybdenum Band at 4366A as a Function of the Molybdenum Concentration
-39-

Solutions of molybdenum prepared by dissolving the metal in aqua

regia, evaporating to near dryness and dissolving the resulting solution

in a 3:1 mixture of absolute ethanol and water were now fed to the sheathed

hydrogen-chlorine flame while the spectrum was scanned from 2500A to 6500A.

A band between 4300A and. 45OOA was detected for 1000 ^g/ml solutions of

molybdenum. This region of the spectrum was also scanned when solvent

without molybdenum was fed to the flame. Recordings of this data are

illustrated in figure 10 .

The width of the•emission band observed indicated that the species

responsible for the band was molecular rather than atomic in nature. To

verify that the band was due to a molybdenum compound standard solutions

ranging in concentration from 0 to 350 ^g/ml were prepared frbm 99.8%

molybdenum metal, Fisher Scientific Company, Baker & Adamson reagent grade

acids and absolute ethanol from the U.S. Industrial Chemicals Company.

Emission intensities for the observed band were plotted as a function of

the concentration of the analyte. Figure 11, page 38, shows these data

and verifys that the band is due to a molybdenum compound. To accurately

identify the emitting species a high resolution instrument was required.

The instrument used in this investigation would have been sufficient if the

emission intensity of this particular band had been strong enough so that

slit widths on the order of 8 microns could have been used. However, since

this was not the case an accurate identification of the emitting species

could not be made.

Emission intensities from the " M o d " band appearing at 43661 were

determined for twenty-seven possible combinations of the total gas flow to


-40-

the burner, the chlorine flow to total gas flow ratio and the percentage

of the total chlorine flow passing through the sheath rather than the

burner. The aspiration rate was held constant because it s affect on the

emission was not that important. Furthermore, in the sheathed flame

maintaining a stable flame front was difficult enough without complicating

matters by varying the aspiration rate. Table VI gives the range of

flame parameters studied along with the particular combination which

produced the greatest emission intensity for molybdenum.

Table VI. Maximization of Emission from Molybdenum in the Sheathed


Hydrogen-Chlorine Flame

Total Gas Flow Chlorine Flow Chlorine Flow Emission


to Sheath Intensity
in 1 /min Total Gas Flow Chlorine Flow
to Burner

Levels of
Parameters
Investigated ■ 8.0-9.0-10.0 0.20-0.30-0.40 0.30-0.40-0.50

Parameter
Combination
Yielding Optimum
Emission 10.0 0.30 0.30 0.94 mv

Exactly why the unsheathed hydrogen-chlorine flame did not yield

atomic emission from refractory elements was not entirely understood. It

seemed as if chlorides were formed in place of oxides since oxide bands

were not observed. To be certain that these elements did not exist in the
-41-

atomic state and did not emit because of the low temperature of the flame,

atomic absorption measurements were made. Hollow cathode tubes were pur­

chased from Westinghouse Electric for molybdenum and zirconium. If these

elements were present in the atomic state, atomic absorption could be used

to verify this.

Using the optical arrangement shown in figure 5, page 26, atomic

absorption measurements were taken on solutions of both molybdenum and

zirconium aspirated into.the hydrogen-chlorine flame. For Mo a WL-22860* .

hollow cathode operated at 20 ma. was used while a WL-22937* tube operated

at 25 ma. was used for zirconium.

Absorption was not detected in the unsheathed flame for either moly­

bdenum or zirconium. Hence, this added support to our suspicions that

these elements existed as chlorides in the reaction zone of the flame and

that the desired atomic state did not exist.. .


Discussion

From the cursory investigation it was concluded that atomic emission

from silver, copper and magnesium in the hydrogen-chlorine flame was some­

what less intense than similar emission in the hydrogen-air flame. Although

intensities in the air flames are roughly twice those in the chlorine

flame, this is not surprising since the hydrogen-air flame is approximately

200° to 300° hotter than the hydrogen-chlorine flame. In equatiop (I),

page 3, it was shown that:

ICC e - E =-/ k I

where I = Emission Intensity


Ea = Excitation Energy
k = Boltzmann's constant
T = Temperature

Substituting into this expression the flame temperatures of the excitation


4
sources being compared one obtains the following range of values for the

logarithm of the ratio of emission intensities:

H2-Air
In « 0.874 - 0.912

H 2 -Gi2

In table VII the values of this ratio are given for/ elements from the

cursory investigation from which atomic spectra were obtained. .As one can

see logarithms of intensity ratios for silver, copper and ;magnesium agree

rather well with the calculated ratio. This indicates that chlorine does

not alter the concentration of atomic vapor in the flame.

The temperature of the hydrogen-chlorine flame has been reported to


be 2075 K while that of hydrogen-air flame is. reportably between 2275 K
and 2375°K (17,14).
r 43-

Tab Is VII. Comparison of Atomic Emission Intensities

Element lnlH2-Alr Element In H2~Alr


Wavelength in A Wavelength in A
1 H2-Cl2 1H 2-Cl2

Calcium
4227 5.050 3720 2.303
Cadmium
3737 2.129
3746 1.970
3261 2.058 3823 ' 2.045
Chromium 3860 . 2.158
3579 1.244 3886. 1.742
3594 1.522 3900 1.033
3605 1.319 . 3930 . 1.898
4254 2.171
Magnesium
4275 1.966
4290 . 1.869 2852 1.099
Manganese

Cobalt
3413 0.846
4033 2.862
3433 0.742 Nickel
3453 1.030 3371 1.322
3468 1.054 '3394 1.203
3508 0.932 3414 1.157
3527 1.125 3438 0.715
3575 0.621 3449 0.916
3595 -0.183 3460 0.112
3630 -1.152 3491 0.956
3650 -1.833 35.14 1.139
3842 0.548 3522 1.040
3871 0.846 3569 0.392
3892 0.122 3619 0.647-
3992 0.599 3854 -0.787
4118 0.798
Silver
Copper 3281 0.683
3248 0.747 ' 3383 • 0.737
3274 0.788
■ Sodium -
Iron 5890 3.570
3441 2.228
•Strontium '
3447 0.796
4607 ■ 5.991
3491 0.000
3570 0.140
3581 1.792
3648 1.792
3706 0.761
-44“

In attempting to explain emission intensity ratios by temperature

variations for calcium, cadmium, manganese, sodium and strontium one finds

that the greater emission in the hydrogen-air flame cannot be explained.

Values of In I .. /I » 0.874 - 0.912 indicate that for these ele-

ments chlorine has an effect on emission intensities which cannot be

accounted for by considering flame temperatures alone.


»

Since emission intensities are directly related to excited state

concentrations, it is concluded that calcium, cadmium, manganese, sodium

and strontium form chlorides in the hydrogen-chlorine flame decreasing the

concentration of metallic vapor.

Chemiluminescence (footnote. 2, page 5) has been shown to account for

excitation in flames under a variety of conditions. If chemiluminescence

is important in the hydrogen-chlorine flame, a second explanation for marked

decreases in emission in the halogen flame is proposed. Halogens, report­

edly quench emission more than nitrogen and oxygen (50). Hence, large

concentrations of chlorine could quench luminescence and thereby account

for significant decreases in emission intensities in the hydrogen-chlorine

flame.

For cobalt, chromium, iron and nickel a comparison of


t
In Itt .. /Itt to the calculated temperature ratio yields inconsistent
H 2-Air H 2-CI2 ’ ' .
results. Depending on the emission line being observed for a particular

element the logarithm of the Itt .. to Itt ratio is less than, equal -
H 2 -Axr H 2-CI2
to, or greater than the Ttt to Ttt . ratio. This variation with, the •
iV cj-2 Hg-Air
emitted radiation being considered is shown in table VII.
-45-

These apparent abnormalities were at first believed to be the result

of poor data. Emission spectra for the transition elements are very com­

plex and overlap background emission making the accurate determination of

emission intensities difficult. Although this undoubtedly explains some

of the variation, another explanation is suggested.

The reactivity of a particular atom is known to be a function of the

electronic state of the atom. Hence, the formation of metal chlorides

depopulates specific electronic states in varying degrees.

As explained in the introduction it was anticipated that refractory

elements (titanium, vanadium, zirconium, niobium, molybdenum, tantalum,

tungsten and rare earths) would emit line spectra in the hydrogen-chlorine

flame. This was not observed experimentally even when special efforts were

made to keep oxygen out of the system. Free- energy of dissociation data

for gaseous metal oxides and chlorides at high temperatures is not entirely -

reliable (11,58) and may explain these observations. Even though the

chloride's of refractory metals are generally less stable than oxides the

chlorides are relatively stable at the moderate temperature of the hydrogen-

chlorine flame (50). Chloride band emission.is seen for copper, calcium
'
and magnesium. Hence, the'"lack of emission from chlorides of titanium,

vanadium, zirconium, niobium, tantalum, tungsten, yttrium, lanthanum,

praseodymium, neodymium and holmium was not expected. Molybdenum lived

up to expectations in that ,a weak molecular molybdenum band was detected

in a sheathed hydrogen-chlorine flame.

Chloride band emission from copper, calcium and magnesium has been
-46-

reported in the literature (17,45,46). Hence, these particular bands

were expected although the strength of the copper chloride band was

delightfully surprising. When emission intensities were maximized for

CuCl emission in the hydrogen-chlorine flame and atomic copper emission

in the hydrogen-air flame by varying flame conditions, the chloride band

was found to be more than fifteen times as intense as the copper line

emission obtained in a hydrogen-air flame. Atomic absorption measurements

at numerous heights in the flame showed that for the hydrogen-air flame the

atomic copper concentration rose sharply until it reached a maximum 35 mm

above the burner tip. From 35 mm to 100 mm above the burner the concen­

tration remained virtually constant. In the chlorine flame optimized for

CuCl emission the concentration of atomic copper reached a maximum at a

point 50 mm above the burner tip. Higher in the flame the copper concen­

tration decreased steadily. This was expected since CuCl formation should

be related to the time a copper atom spends in the chlorine atmosphere and

the oxidative nature of the region being viewed. By integrating percent

absorption over the entire flame, from the burner tip to a position 100 mm

above the burner, for chlorine and air supported flames the concentration

of copper in the hydrogen-chlorine flame can be calculated relative to the

concentration in the hydrogen-air flame. In doing this it was found that

the Cu concentration in the chlorine source was 83.9% of that in the air

supported source. This value seems high, in view of the intensity of the

CuCl band. It is concluded that in the hydrogen-chlorine flame CuCl .

emits a very strong band suitable for the spectroscopic determination of

copper.
-47-

Emission from Cg, CH and OH observed when ethanol was used as the

solvent was maximized by the procedure used in the optimization of copper

emission. Comparing the optimum flame parameters for the emission from

C2, CH and OH in the hydrogen-chlorine and hydrogen-air flames, one finds '

that excitation conditions for maximum emission are not significantly

different. Total flow rates of gas to the burner were nearly the- same in

the chlorine and air flames for the emission from these species as were

oxidant flow to total gas flow ratios. For Cg and CH emission the maximum

rate at which ethanol was aspirated into the flame produced the maximum

emission in both flames. Emission from the OH radical did not follow- this

pattern. In the hydrogen-air flame emission intensities increased as the

aspiration rate decreased while in the chlorine flame an opposite trend

was observed. As the aspiration rate increased, increased intensities from

the OH radical were observed. This behavior can be explained by consider­

ing the source of oxygen for the formation of the OH radical. If one neg­

lects the possibility of drawing oxygen from the surroundings into the

reaction zone of the hydrogen-chlorine flame, the only available oxygen is

that contained in the absolute ethanol being ,fed to the flame by means of

the capillary through the center of the burner (see figure 2 , page 11).

Hence, one would expect the emission intensity to be directly related to

the rate at which ethanol was being fed to the flame.

In the hydrogen-air flame the situation is somewhat different. In

addition to oxygen from the solvent the oxidant itself can provide for the

formation of the OH radical. In fact in this particular flame the OH

radical is necessary for the propagation of the reaction. This suggests


that the aspiration rate has a negligible affect on the OH emission pro­

vided the solvent being aspirated does not react with OH radicals and

decrease their concentration. Experimentally it was observed that the

intensity of the OH band was inversely proportional to the rate at which

absolute ethanol was fed to the flame. Hence, in the hydrogen-chlorine

flame it is concluded that ethanol decreases the concentration of this

species.

A comparison of the emission intensities from Cg and CH at various

positions in the flame shows that the C^ molecule and CH radical reach a

maximum intensity 21 mm and 23 mm above the burner tip respectively in

both flames. The OH radical reaches a maximum somewhat lower in the two

flames at 19 mm. ' .

From figure 8 , page 32, it is apparent that CH emission in the air

supported flame closely resembles that in the chlorine supported flame.

This is not true of the Cg molecule. As one can plainly see in figure 7,

page 31, far greater emission from the Cg molecule was obtained from the

hydrogen-chlorine flame. Noting that emission in the air flame from the

OH radical decreases as the aspiration rate of absolute ethanol increases


' .•
and the Cg emission is significantly less in the air flame, it may be con-

eluded that the Cg molecule, originating from the organic solvent, reacts

with the OH radical and thereby decreases the observed intensities of bands

from this species. ,

Finally it has been observed that emission from atomic copper occurs

predominately when excess fuel is being furnished to the flames. In the

Hg-Air flame CuH emission is likewise increased when a fuel rich reducing
i

-49-

atmosphere is employed while the CuCl band is a maximum in a hydrogen-

chlorine flame when less fuel and more chlorine is used. These results

are what would be expected.

L
Conclusion

This investigation has attempted to determine the potential role of

a hydrogen-chlorine flame excitation source in flame phofometry. The

results have led to the conclusion that silver, copper and magnesium

emitted useful radiation although somewhat less intense than comparable

emission in the hydrogen-air flame. Lower intensities were attributable

to the lower temperature of the halogen flame. Chlorine decreased the

concentrations of atomic calcium, cadmium, manganese, sodium and strontium

in the hydrogen-chlorine flame. On the other hand, the affect of chlorine

on individual atoms of cobalt, chromium, iron and nickel was found to

depend on the particular electronic state of the atom involved.

Refractory metals (titanium, vanadium, zirconium, niobium, tantalum,

tungsten, yttrium, lanthanum, praseodymium,, neodymium and holmium) in gen­

eral did not emit spectra when introduced into the hydrogen-chlorine flame

Molybdenum was an exception in that a weak molecular band was observed

between 4300A and 45OOA when a sheathed flame was employed. Copper on the

other hand gave an intense CuCl band in the chlorine excitation source.

The strength of this emission suggested that this flame may be useful

analytically in the determination of copper.

Suggestions for Future Work

The results presented in the dissertation virtually rule out us>e of

the.hydrogen-chlorine flame in analytical determinations in the UV-visible

region of the spectrum for elements other than copper. However, the

apparent strength of the CuCl band and the conclusion that atomic line

emission from refractory metals was not seen because of stable chloride

formation suggests that molecular chloride bands may be worth further


" 5 ]_—

consideration. The red end of the visible spectrum, and the near infrared

portion of the spectrum could not be observed with the equipment available.

A study of the emission from refractory metals in a- hydrogen-chlorine flame

In this region may be worthwhile and is suggested for a future study.

The use of chlorine in conjunction with acetylene or possibly cyanogen

may also be justified. The acetylene chorine flame was considered briefly

at the beginning of this investigation. The reducing atmosphere produced

by this combination of gases appears promising. However, the nature of

these gases and the high background emission are disadvantages which need

to be carefully considered.

Use of cyanogen will require a more cautious approach than that of

acetylene. Flames from chlorine and cyanogen have not been reported in

the literature. Hence, it is pure speculation to suggest use of this

combination for a flame photometric excitation source. One important

advantage to using this combination of gases is that higher flame temp­

eratures would be expected than those obtained from other chlorine

supported flames.
¥

Literature Cited

1. Alkemade, CeT 1J., Anal. Chem., 38, 1252 (1966).

2. Allen, J.E., Spectrochim. Acta _17, 467-473 (1961).

3. Amos, M.D., Willis, J.B., Spectrochim. Acta 22, 1325-1343 (1966).

4. Baker, M.R., Fuwa, K., Thiers, R.E., Vallee, B.L., J. Opt. Soc. Am. 48.
576 (1958).

5. Baker, M.R., Vallee, B.L., Anal. Chem. 31, 2036 (1959).

6. Baker, M.R., Vallee, B.L., J. Opt. Soc. Am. 45, 773 (1955).

7. Barrow, R.F., Endeavour VI, No. 22, (1947).

8. Barthel, C.H.,.Genie Chim. 77, 34-36 (1957).

9. Box, G.E., Wilson, K.B., Jour. Roy. Stat: Soc. B13, I (1951).

10. Brewer, L., U.S. Atomic Energy Commission UCRL - 8356 (1958).

11. Brewer, L., Somayajulu, G.R., Brackett, E., Chem. Revs. _63, 111-121
(1965).

12. Buell, B.E., Anal. Chem. 35, 372-382 (1963).

13. Bulewicz, E.M.., James, C.G., Sugden, T.M., Proc. Roy. Soc. (London)
A235, 89 (1956).

14. Burriel-Marti, F . , Ramirex-Munoz, J., "Flame Photometry", Elsevier,


Amsterdam, 1957.

15. Cellier, K.M., Stace, H 0C 0T., Appl.' Spectroscopy 20, 26-33 (1966). .

16. Cochran, W.G., Cox, G.M., "Experimental Designs", John Wiley and
Sons, New York, 1960.

17. Coller, H.E., Ph.D. Thesis, Lehigh University (1955).

18. D 1Silva, A.P., Knisely, R 0N 0, Fassel, V.A., Anal. Chem. 36, 1287-1289
(1964).

19. Dean, J.A., "Flame Photometry", McGraw-Hill, New York, 1960.

20. Dufour, R.F., U.S. Atomic Energy Commission TIP - 7605 (1960).

21. Durie, R.A., Proc. Roy. Soc. A212, 110-121 (1952).

22. Eshelman, H.C., Dean, J.A., Anal. Chem. _31, 183-187 (1959).
-53-

23. Fassel, V.A., Curry, R.H., Knisely, R 0N., Spectrochim Acta 18, 1127
(1962) .

24. Fassel, V 0A 0, Knisely, R.N., Curry, R.H., Anal. Chem. _36, 532 (1964).

25. Fassel, V 0A., Knisely, R 0N., Curry, R.H., Anal. Chem. 36, 1287 (1964).

26. Fassel, V.A., Meyers, R 0B0, Knisely, R.N., Spectrochim. Acta 19, 1187
(1963) . . —
27. Feber, R.C., Atomic Energy Commission LA - 2841 (1963).

28. Fristrom, R.M., Westberg, A oA 0, "Flame Structure", McGraw-Hill,


New York, 1965.

29. Gaydon, A.G., "The Spectroscopy of Flames," Chapman and Hall, London
(1957).

30. Gaydon, A.G., Wolfhard, H.G0, "Flames, Their Structure, Radiation and
Temperature", Chapman and Hall, London, 1960.

31. Gibson, J.H., Grossman, W 0E 0L 0, Cooke, W.D., Anal. Chem. 35, 266-277
(1963).
32. Gilbert, P.T., Proceedings of the Xth. Colloquim Spectroscopicum
Internationale, Spartan.Books, Washington, 1962.

33. Herrmann, R., Alkemade, C 0T 0J0, "Chemical Analysis by Flame Photo­


metry", Interscience, New York, 1963.

34. Hildenbrand, D 0L 0, personal correspondence, 1965.

35. Langmuir, J., Western Soc. Engrs. 3_1, 373-387 (1927).

36. Langmuir, J., Ind. Eng. Chem. _19, 667-674 (1927).

37. Mann, D 0E 0, Thrush, B 0A 0, Lide, E.R., Ball, J.J., Acquista, N.,


J. Chem. Phys. 34, 420-431 (1961).

38. Mansell, R.E., Emmel, H 0W o, Atomic Absorption Newsletter, Perkin-


Elmer Corp., 4-, 365-366 (1966).

39. Mavrodineau, R., Boiteux, H., "Flame Spectroscopy", John Wiley and
Sons, New York, 1965.

40. Miilford, C.E., Atomic Absorption Newsletter, Perkin-Elmer Corp., _5,


88-90 (1966).

41. Parsons, M 0L 0, Winefordner, J.D., Appl.. Spectroscopy 21, 368-374 (1967)


-54-

42. Pungor5 E 05 "Flame Photometry Theory", D 0 Van Nostrand Company Ltd.,


London5 1967.

43. Rarin5 C.S.5 Hambly5 A 0N 05 Anal. Chem. _37, 879-884 (1965).

44. Rossini5 F.D., Nat. Bur. Standards Circular 500, U 0S 0 Government


Printing Office, 1952.

45. • Schmauch5 G 0E0, Serfass5 E 0J 05 Appl. Spectroscopy 12, 98-102 (1958).

46. Schmauch5 G 0E 05 Serfass5 E 0J 05 Anal. Chem. JBO5 1160-1161 (1958).

47. Simmons5 R 0F., Wolfhard 5 H 0G 05 Jet Propulsion 27, 44-48 (1957).

48. Skogerboe5 R 0K 05 Ph.D. Thesis, Montana State University5 August 1963.

49. Skogerboe5 R 0K 05 Heyby 5 A 0T 05 Morrison 5 G 0H 05 Anal. Chem. _38, 1821-


1824 (1966).

50. Slootmaekers5 P 0J 05 Van Tiggelen5 A., Bull. Soc. Chim., Belg. 67,
135-146 (1958).

51. Stander5 C.M., Anal. Chem. _32, 1296-99 (1960).

52. Steel, R.G.D., Torrie5 J 0H 05 "Principles and Procedures of Statistics",


McGraw-Hill, New York 5 1960.

53. Streng5 A.G., Grosse 5 A.V., J. Am. Chem. Soc. _79, 3296-97 (1957).

54. Streng5 A 0G 05 Grosse5 A.V., J. Am. Chem. Soc. 751, 5583 (1957).

55. Stull5 D.R., "Joint Army Navy Air Force Thermochemical Tables",
Clearing House for Federal Scientific and Technical Information5 1965.

56. Tourin5 R.H., "Spectroscopic Gas Temperature Measurement", Elsevier,


Amsterdam5 1966.

57. Vallee 5 B 0L 05 Bartholomay5 A oF 05 Anal. Chem. 28., 1753-55 (1956).

58. Wicks5 C.E., Block, F 0E 05 Bureau of Mines Bulletin 605, U.S, Govern­
ment Printing Office, 1963.

59. Willis, J 0B 05 Rasmuson5 J 0O 05 Kniseley5 R 0N05 Fassel5 V.A., Spiectro-


■chim. Acta. 23, 725-730 (1968). '
APPENDIX
-56-

CuCl Band

Sensitivity: 20 mv

4OOOA 425OA 45OOA


Figure 12. Copper Chloride Band Emission Obtained in the Hydrogen-
Chlorine Flame .
OH Bands

Sensitivity: I mv

3000A 3500A 4000A

Figure 13. Background Emission from the Hydrogen-Chlorine Flame


eBUiBIJ auTaoiqo-uaScupXH aq^ uioaj uofsspug punoaSqoeg e^x BanSjj
Y0009

-58
OH Bands

I I I
3000A 3500A 4000A

Figure 15. Background Emission from Absolute Ethanol in the Hydrogen-Chlorine Flame
5OOOA 55OOA 6000A

Figure 16. Background Emission from Absolute Ethanol in the Hydrogen-Chlorine Flame.
CH

OH Bands

Sensitivity: 2 mv

3000A 3500A 4000A

Figure 17. Background Emission from Ethyl Acetate in the Hydrogen-Chlorine Flame.
Sensitivity: 2 w

5000A 5500A 6000A

Figure 18. Background Emission from Ethyl Acetate in the Hydrogen-Chlorine Flame
CN

Sensitivity: 5 mv

3000A 3500A 4000A

Figure 19. Background Emission from Acetonitrile in the Hydrogen-Chlorine Flame


2

64-
Sensitivity: 2 rav

5000A 5500A 6000A

Figure 2 0, Background Emission from Acetonitrile in the Hydrogen-Chlorine Flame


MONTANA STATE UNIVERSITY LIBRARIES

3 1762 10005668 6

J647 Johnston, pIlliam Kemp


cop.2 An evalu tion of the
hydrogen-chlorine flame
as an excitation source...

You might also like