United StMee
Environmental Protection
Agency
Environmental Sconces Research
Laboratory
Research Trienjjle Park NC 2771 1
EPA -600 2 79017
February 1979
Reeearch end Development
Analytical
Procedures for
Characterizing
Unregulated
Pollutant
Emissions From
Motor Vehicles
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR) ,
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-017
February 1979
ANALYTICAL PROCEDURES FOR CHARACTERIZING UNREGULATED
POLLUTANT EMISSIONS FROM MOTOR VEHICLES
by
Harry E. Dietzmann
Lawrence R. Smj.th
Mary Ann Parness
E. Robert Fanick
Southwest Research Institute
San Antonio, Texas 78284
Contract No. 68-02-2497
Project Officer
Frank Black
Emissions Measurement and Characterization Division
Environmental Sciences Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, N. C. 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
-------
FOREWARD
The Clean Air Act, in the August 1977 amendments, calls for assurance
that no emission control device, system, or element of design used in a
new motor vehicle or new motor vehicle engine cause or contribute to an
unreasonable risk to public health, welfare, or safety. This mandate
implies the measurement of a variety of presently non-regulated compounds
in the exhaust and/or evaporative emissions from such vehicle. The
Environmental Protection Agency, in Advisory Circular 76, and in other
documents currently in review, calls for measurement of many potentially
hazardous compounds.
The Environmental Sciences Research Laboratory in EPA's Office of
Research and Development is currently sponsoring a contractual effort at
Southwest Research Institute, San Antonio, Texas wherein the emission
rates of several non-regulated pollutants from three-way catalyst equipped
motor vehicles are being examined. This document describes in detail the
analytical procedures being used in this study. These procedures are not
necessarily the best or only means of determining the emission rates of
the compounds of interest. They represent an assessment of the optimum
procedures available at the time of the contract.
Comment is herewith being solicited from any laboratory attempting
to apply the defined procedures. If difficulties are experienced,
improvements defined, or new procedures developed, please communicate
same to the EPA Project Officer and at appropriate time intervals a
document will be issued transmitting this information to all holders of the
base document.
Dr. Ronald L. Bradow, Chief
Mobile Source Emissions Research Branch
Emissions Measurement and Character-
ization Division
111
-------
ABSTRACT
Analytical procedures are described that may be used to assess motor vehicle
emission rates of several unregulated pollutants including aldehydes, organic
amines, sulfur dioxide, nitrous oxide, several individual hydrocarbons
including benzene, hydrogen sulfide, total cyanide, organic sulfides, nickel
carbonyl, ammonia, sulfate, and N-nitrosodimethylamine (sampling conditions
only). A series of validation experiments involving motor vehicle exhaust
with injects of known quantities of the compounds of interest and the Constant
Volume Sampling system commonly used in emissions certification are described
for several of the analytical procedures. 'The Clean Air Act as amended
August 1977 requires in section 202(a) 4 that unregulated pollutants emitted
from motor vehicles be measured to assure that no unreasonable risk to public
health and welfare exists.
IV
-------
CONTENTS
Foreward iii
Abstract iv
Figures vi
Tables ix
1. Introduction 1
2. Aldehyde and Ketone Procedure 2
3. Organic Amine Procedure. . . . : 19
4. Sulfur Dioxide Procedure 38
5. Nitrous Oxide Procedure 53
6. Individual Hydrocarbon Procedure 63
7. Hydrogen Sulfide Procedure 71
8. Total Cyanide Procedure 84
9. Organic Sulfide Procedure 108
10. Nickel Carbonyl Procedure 140
11. Ammonia Procedure 147
12. Results and Conclusions 164
References 167
Appendices
A. Interim Report, Task I 176
B. Aldehyde and Ketone Procedure 222
C. Organic Amine Procedure 251
D. Sulfur Dioxide 273
E. Nitrous Oxide Procedure 295
F. Individual Hydrocarbon Procedure 313
G. Hydrogen Sulfide Procedure 335
H. Total Cyanide Procedure 359
I. Organic Sulfide Procedure 382
J. Nickel Carbonyl Procedure 417
K. Ammonia Procedure 429
L. Sulfate Procedure 452
M. DMNA Sampling Procedure. 472
v
-------
FIGURES
Number
Page
1 Plot of the formaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 7
2 Plot of the acetaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 8
3 Plot of the acetone-DNPH derivative concentration determined
by procedure vs actual concentration . 9
4 Plot of the isobutyraldehyde-DNPH derivative concentration
determined by procedure vs actual concentration. ...... 10
5 Plot of the methylethylketone-DNPH derivative concentration
determined by procedure vs actual concentration 11
6 Plot of the crotonaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 12
7 Plot of the hexanaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 13
8 Plot of benzaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 14
9 GC peak areas of pentafluorobenzoyl amine derivatives
vs time 25
10 Linearity of monomethylamine GC response (plot on log-log
scale) •. . . 29
11 Linearity of dimethylamine GC response (plot on log-log
scale) 30
12 Linearity of trimethylamine GC response (plot on log-log
scale) 31
13 Linearity of monoethylamine GC response (plot on log-log
scale) 32
14 Linearity of diethylamine GC response (plot on log-log
scale) 33
vi
-------
FIGURES (Cont'd).
15 Linearity of triethylamine GC response (plot on log-log
scale) 34
16 SO2 calibration curve 49
17 Detector linearity curve 55
18 Sample decay curve (short term) 57
19 Sample decay curve (long term) 58
20 Time-sample decay curve (exhaust only) 68
21 Time-sample decay curve (standard only) 69
22 Time-sample decay curve (exhaust and standard) 70
23 Time-Light exposure study (low concentration) 74
24 Time-Light exposure study (high concentration) 75
25 Beer's Law plot for methylene blue 76
26 Effect of elapsed time on hydrogen cyanide in clear and
dark bags 88
27 The effect of elapsed time on cyanogen in clear and
dark bags 89
28 The effect of elapsed time on hydrogen cyanide in a blend
of hydrogen cyanide and cyanogen in clear and dark bags... 90
29 The effect of elapsed time on cyanogen in a blend of
hydrogen cyanide and cyanogen in clear and dark bags .... 91
30 The effect of elapsed time on hydrogen cyanide and cyanogen
in a dark bag with humid nitrogen 92
31 Total cyanide calibration curve at low concentrations
(0-2 ppm) 98
32 Total cyanide calibration curve at low concentrations
(0-10 ppm) 99
33 The effect of elapsed time on sample development 102
34 Time-sample decay curve 103
35 Proposed GC flow schematic for analysis of organic
sulfides (Step 1) 110
vii
-------
FIGURES (Cont'd) .
36 Proposed GC flow schematic for analysis of organic
sulf ides (Step 2) ...................... 11;L
37 Proposed GC flow schematic for analysis of organic
sulfides (Step 3) ...................... 112
38 Gas chromatograph separation of several organic sulfides
in prepared blend ......................
39 Cold trap experiment flow schematic .............. 116
40 Typical gas chromatograph trace of organic sulfides ...... 122
41 Typical GC separation of organic sulfides on acetone-washed
Poropak QS column ...................... 124
42 Typical organic sulfide separation with Tenax-GC column. . . . 125
43 Organic sulfide permeation blend with secondary dilution,
near detection limit of GC FPD system ............ 126
44 Organic sulfide permeation blend with secondary dilution,
concentrated on Tenax-GC trap and thermally desorbed into
GC FPD system ........................ 127
45 Carbonyl sulfide linearity plot ................ 132
46 Methyl sulfide linearity plot ................. 133
47 Ethyl sulfide linearity plot ................. 134
48 Methyl disulfide linearity plot ................ 135
49 Nickel carbonyl analysis and dilution system flow
schematic .......................... 142
50 CL detector linearity for nickel carbonyl with a No. 74
Wratten filter ....................... 144
51 CL linearity for nickel carbonyl with No. 54 Wratten
filter ........................... 145
52 Ammonia calibration curve ................... 155
vnx
-------
TABLES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Physical Properties of the Aldehydes and Ketones (3,4) . . . .
Injection Repeatability
Multiple Extractions of DNPH Solutions (all Units are mg
DNPH derivative/m£ of Toluene)
Percent Recovery of Propionaldehyde
List of Individual Organic Amines Included in the
Emissions Characterization Inventory
Mixing Procedure for Preparation of Pentafluorabenzoylamine
Derivatives*
Injection Repeatability Experiments
Summary of Amine Recoveries (Tunnel Only) and Tunnel
History Effects with Methylamines
Analysis of the Contents of Bubblers 1 and 2 Separately
and Combined
Sulfate Standard Stability
S02 Collection Efficiency as a Function of Flowrate
and Temperature
Injection Repeatability for Ion Chromatograph
Calibration Curve for Sulfur Dioxide
Sulfur Dioxide Baseline from Plymouth Fury Test Car
Sulfur Dioxide Recovery from Direct CVS Injection
Sulfur Dioxide Recovery from Dilute Exhaust by Direct CVS
Injection During the Hot FTP Driving Cycle
Page
3
6
15
17
20
24
28
36
43
43
44
45
47
48
50
51
51
IX
-------
TABLES (Cont'd).
18 Injection Repeatability Over the Range of Detector
..... ...... 56
Linearity
19 Typical Nitrous Oxide Exhaust Concentrations and Emission
Rates on Several Driving Cycles with Two Vehicles used
During Qualification Experiments 60
20 Nitrous Oxide Injection Recoveries with 1977 260 V8
Oldsmobile Cutlass with Oxidation Catalyst 61
21 Nitrous Oxide Injection Recoveries with 1978 305 V8
Oldsmobile Cutlass with Oxidation Catalyst 61
22 Nitrous Oxide CVS Qualification Experiments - No Vehicle ... 62
23 Important Facts on Individual Hydrocarbons 65
24 Injection Repeatability on Two Separate Occasions 66
25 The Effect of Sample Flow Rate and Absorbing Reagent
Temperature on the Collection Efficiency 77
26 The Effect of Individual Exhaust Components on the
Development of Methylene Blue 78
27 The Effect of Anions on the Development of Methylene Blue. . . 79
28 The Effect of Sulfur Dioxide Interference on the
Development of Methylene Blue 80
29 Hydrogen Sulfide Gaseous Recovery by Direct CVS Injection. . . 81
30 Hydrogen Sulfide Gaseous Recovery from Dilute Exhaust by
Direct CVS Injection 81
31 Experiments Conducted for HCN and C_N Bag Stability 87
£* £,
32 The Effect of Stopper Tip and Absorbing Reagent Concen-
tration on Collection Efficiency at Room Temperature .... 95
33 The Effect of Absorbing Reagent Temperature on HCN
Collection Efficiency 96
34 Calibration Curve Linearity at Several Cyanide Concen-
trations 97
35 Sample Injection Repeatability for Two Cyanide Concen-
trations , 100
-------
TABLES (Cont'd).
36 Total Cyanide Gaseous Recovery'from Dilute Exhaust by Direct
CVS Injection During a Hot FTP Driving Cycle 105
37 Total Cyanide Gaseous Recovery by Direct CVS Injection .... 106
38 List of Sulfur Compounds Included in the Analysis of
Organic Sulfides 108
39 List of Chemical and Physical Characteristics of Various Sul-
fur Compounds Potentially Present in Automotive Exhaust. . . 115
40 The Effect of Cold Trapping at -78 "C on Carbonyl Sulfide
and Methyl Sulfide at Various Concentrations, Flow Rates
and Trap Sizes 117
41 The Effect of Cold Trapping at -196°C on Carbonyl Sulfide
and Methyl Sulfide with Various Trap Sizes 119
42 The Efficiency of Various Materials Trapping Sulfides at
Several Temperatures 120
43 Injection Repeatability for the Organic Sulfides 129
44 Trap Repeatability for Organic Sulfide Collection 130
45 Trap to Trap Repeatability for Organic Sulfide Collection'. . . 130
46 Percent Recoveries of the Organic Sulfides from the CVS
Tunnel Only 137
47 Percent Recoveries of the Organic Sulfides from the CVS
Tunnel and Exhaust 138
48 NH^ Collection Efficiency as a Function of Flowrate and
Temperature 150
49 Retention of Ammonia in CVS (constant volume sampler) 152
50 Injection Repeatability for Ion Chromatograph 153
51 Repeatability of Ammonia Standard Preparation 154
52 Calibration Curve for Ammonia 156
53 Sample and Standard Stability as a Function of Time 157
54 Ammonia Baseline from Plymouth Fury Car 158
55 Ammonia Recovery from Direct CVS Injection 159
XI
-------
TABLES (Cont'd).
56 Ammonia Recovery from Dilute Exhaust by Direct CVS Injection
of 0.1 Percent NH3 During the Hot FTP Driving Cycle 16,0
57 Ammonia Recovery from Dilute Exhaust by Direct CVS Injection
of'1.0 Percent NH3 During the Hot FTP Driving Cycle 161
58 Analytical Procedures for Emissions Characterization 165
xii
-------
ACKNOWLEDGEMENT
The project, "Characterization of Emissions from Motor Vehicles Designed for
Low NOX Emissions", for which this report is a part was initiated by the
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
The scientific and engineering effort on which this report is based was
accomplished by the Department of Emissions Research, Southwest Research
Institute, 6220 Culebra Road, San Antonio, Texas, under EPA Contract Number
68-03-2497. This report, designated as Interim Report II, encompasses the
effort to accomplish Task II of this project, Qualification of Methodology.
Task II of this project began September 1977 and was completed November 1978.
The EPA Project Officer during this period was Frank Black. The project was
initially under the supervision of Sherrill F. Martin, Project Leader, and
Harry Dietzmann, Senior Research Chemist. Lawrence R. Smith, Senior Research
Scientist, and Harry Dietzmann directed the project beginning August 1978.
This project was identified within Southwest Research Institute as Project
11-4840-001.
Xlll
-------
SECTION 1
INTRODUCTION
The objective of this project is to select, qualify, and then utilize
suitable measurement instrumentation and techniques for the characterization
of regulated and a variety of currently non-regulated exhaust emissions from
gasoline automobiles using various catalytic methods to achieve low oxides
of nitrogen. Task I of this project includes a detailed literature search
to determine candidate analytical methods for each of the following compounds
or group of compounds.
Nitrous Oxide Hydrogen Sulfide
Hydrogen Cyanide Sulfuric Acid
Ammonia Sulfur Dioxide
Cyanogen Carbonyl Sulfide
Organic Amines Organic Sulfides
N-nitrosodimethylamine Nickel Carbonyl
The analytical methods chosen for Task II were selected by the Project
Officer. Task II includes the validation and qualification of the analytical
methods selected in Task I. Task III involves the use of the selected
analytical methods to measure exhaust emissions from gasoline automobiles
which use catalytic methods to achieve low oxides of nitrogen.
This report is a summary of the results of Task II in which the selected
analytical methods of Task I are validated. A review of the literature
search of Task I, procedural development work, validation experiments, and
qualification experiments for ten analytical procedures are discussed. These
procedures include: aldehyde and ketone procedure (Section 2), organic amine
procedure (Section 3), sulfur dioxide procedure (Section 4), nitrous oxide
procedure (Section 5), individual hydrocarbon procedure (Section 6), hydrogen
sulfide procedure (Section 7), total cyanide procedure (Section 8), organic
sulfide procedure (Section 9) , nickel carbonyl procedure "(Section 10) , and
ammonia procedure (Section 11). Interim Report I, a finalized copy of the
analytical procedures discussed in Section 2-11, the BCA sulfate procedure,
and sampling conditions for DMNA are included as an appendix.
-------
SECTION 2
ALDEHYDE AND KETONE PROCEDURE
LITERATURE SEARCH
The individual aldehydes and ketones that are included in this analysis
are: formaldehyde, acetaldehyde, acetone, acrolein, propionaldehyde, iso-
butyraldehyde, methylethylketone, crotonaldehyde, hexanaldehyde, and benz-
aldehyde. Acetone, acrolein, and propionaldehyde are not resolved from each
other under normal gas chromatographic operating conditions and all three are
reported together as acetone. The common names, the International Union of
Chemists approved names, the chemical formulas, the molecular weights, the
melting points, the boiling points, the densities, the molecular weights of
the 2,4 dinitrophenylhydrazone derivative, and the melting points of the
2,4 dinitrophenylhydrazone derivative are presented in Table 1. The alde-
hydes and ketones have a characteristicly pungent odor, are flammable, are
photochemically reactive, can cause respiratory problems, and are severe eye
irritants. The 1976 American Conference of Government Industrial Hygienists
has recommended threshold limit values for several of the aldehydes and ke-
tones (1) . These values range from 0.1 ppm for acrolein to 1000 ppm for
acetone. Other values listed were 2 ppm for formaldehyde and crotonaldehyde,
100 ppm for acetaldehyde and 200 ppm for methlethylketone.
PROCEDURAL DEVELOPMENT
A procedure, which is already in use at Southwest Research Institute,
developed by the Mobile Source Emissions Research Branch of the ESRL-EPA at
Research Triangle Park, North Carolina, was selected for the analysis of the
aldehydes and ketones (2) . This procedure involves bubbling exhaust through
glass impingers containing 2,4 dinitrophenylhydrazine (DNPH) in dilute hydro-
chloric acid. The exhaust sample is collected continuously during a test
cycle. The aldehydes and ketones (also known as carbonyl compounds) react
with the DNPH to form their respective phenylhydrazone derivatives. These
derivatives are either insoluble or only slightly soluble in the DNPH/HC1
solution and are removed by filtration followed by pentane extractions. The
filtered precipitate and the pentane extracts are combined, and the pentane
is evaporated in a vacuum oven. The remaining dried extract contains the
phenylhydrazone derivatives. The extract is dissolved in a quantitative
volume of toluene containing a known amount of anthracene as an internal
standard. A portion of this extract is injected into a gas chromatograph
and analyzed using a flame ionization detector. A copy of this procedure as
used by the Department of Emissions Research at Southwest Research Institute
will be included as an attachment to this report.
-------
TABLE 1. PHYSICAL PROPERTIES OF THE ALDEHYDES AND KETONRS (3,4)
Ul
Molecular
Weight
Aldehyde or Ketone
Formaldehyde
Acetaldehyde
Acetone
Propionaldehyde
Acrolein
I sobu tyr aldehyde
Methylethylketone
Crotonaldehyde
Hexanaldehyde
Benzaldehyde
ICU Name
Methanal
Ethanal
2 -Propa none
Propanal
Propenal
2-Methylpropanal
2-Butanone
trans2*-Butenal
Hexanal
Benzenecarbonal
Chemical
Formula
CII20
CH3CHO
CH3COCH3
CH3CH2CHO
CH2:CHCHO
CH3CH(CH3)CHO
CH3COCH2CH3
CH3CH :CHCHO
CH3(CH2)4CHO
CgH^CHO
Molecular
Weight
30.03
44.05
58.08
58.08
56.07
72.11
72.11
70.09
100.16
106.13
Melting
Point
- 92
-121
- 95
- 81
- 87
- 65
- 84
- 74
- 56
- 26
Boiling
Point
- 21
21
56
49
53
63
80
105
128
178
Density
0.815
0.783
0.790
0.806
0.841
0.794
O.B05
0.850
0.814
1.042
DNPH
Derivative
210.15
224.19
238.21
238.21
236.20
252.23
252.23
250.21
279. 2R
286.25-
Melting
Point Her
167
168
128
156
165
182
___
190
104
237
-------
VALIDATION EXPERIMENTS
Several experiments were carried out to determine the validity of the
DNPH procedure for the analysis of the aldehydes and ketones. These
experiments included checks for: GC injection variability, linearity of
detector response, sample stability in the DNPH absorbing solution and in
toluene, trapping efficiency of the DNPH/HCl solution, interferences, and
extraction plus injection repeatability.
The finalized sampling conditions used to collect the aldehydes and
ketones are listed below as is a discussion on their selection. Two glass
impingers in series, each containing 40 m& of 2 N HCl/2-4 dinitrophenyl-
hydrazine, are used to collect the aldehydes and ketones. The two impingers
together trap 98+ percent of the carbonyl compounds. This collection
efficiency was determined by bubbling known amounts of the aldehydes and
ketones through a series of impingers and analyzing each impinger separately.
No advantage was found in using more than two impingers. There was no
observed difference in analyzing the contents of the two impingers separately
or combined. Since the analysis of the two impingers combined is less man-
power intensive, the two impingers are analyzed together. During sampling,
the two impingers are kept in a 0°C ice bath. The ice bath offers no sig-
nificant advantage in collection efficiency over room temperature, but does
provide a stable sampling temperature during the test. The 0°C temperature
also lowers the vapor pressure of the aqueous absorbing solution and thus
prevents loss of any significant amount of water from the absorbing solution
during sampling. The sample flow rate through the impingers is maintained
at 4 liters a minute. This flow provides the largest amount of sample to
flow through the absorbing reagent without loss in absorbing efficiency or
the physical loss of any absorbing reagent. A teflon sample line connecting
the CVS to the impingers is heated to 77°C in order to prevent water from
condensing in the sample line. Some of the aldehydes and ketones are water
soluble, and the condensation of water in the sample line could cause a
significant loss of sample in the sample line.
The HCl/DNPH absorbing reagent has been found to be stable over several
days; however, to prevent the possibility of contamination or the inad-
vertant use of "old " absorbing reagent, the solution is prepared daily as
needed.
The samples have been found to be stable for at least two days in the
absorbing reagent. However, to prevent the possibility of contamination of
the samples by their standing in the lab for prolonged periods, the samples
are extracted, dried, and dissolved in toluene all in the same day. Once
the sample is dissolved in toluene it is stable for relatively long periods
of time. Samples run and re-run over a period of two weeks showed no signi-
ficant change in concentrations.
To determine the GC injection repeatability for the procedure over a
wide range of concentrations, four standards containing 1.6, 0.2, 0.02 and
0.002 mg of each aldehyde and ketone DNPH derivative per m£ of toluene were
prepared. These are the concentration ranges expected when sampling dilute
exhaust. Each standard was injected into the GC five consecutive times.
-------
The concentration determined by the procedure for each of the derivatives was
averaged over the 5 runs, and a standard deviation as well as a percent
standard deviation was calculated. The results of these injection repeat-
ability experiments are presented in Table 2. The injection repeatability is
good for the 1.6 mg derivative /m£ toluene standard (percent deviation ranges
from 1.1 percent for formaldehyde to 9.6 percent for benzaldehyde) and the
0.2 mg derivative/m£ toluene standard (percent deviation ranges from 0.5
percent for acetaldehyde to 5.9 percent for benzaldehyde). At the two lower
concentrations the standard deviation in the injection repeatability was
found to be much larger. The 0.02 mg derivative/m£ toluene standard gave
percent deviations ranging from 3.7 percent for acetone to 32 percent for
benzaldehyde. The 0.002 mg derivative/m& toluene standard gave percent
deviations which ranged from 10 percent for formaldehyde to 110 percent for
benzaldehyde. It appears from the data that the injection repeatability is
good at higher derivative concentrations, but much more erratic at very low
concentrations.
To determine the linearity of the detector for the concentration ranges
of interest for each of the derivatives, seven standard solutions were pre-
pared which contained 8.0, 4.0, 1.6, 0.8, 0.2, 0.02, and 0.002 mg of each
derivative/m£ toluene. These standards were made by weighing out required
amounts of each derivative and dissolving them in the appropriate amount of
toluene to give the required concentrations. The solution containing 0.2 mg
of each derivative/m£ toluene was used as the standard and the other six
solutions were compared to this standard. Figures 1-8 show plots of the pro-
cedure determined concentration vs the actual concentration on a log-log
scale. Acetone, methylethylketone, and crotonaldehyde give linear plots
throughout the region of interest. Formaldehyde, acetaldehyde, isobutyr-
aldehyde and hexanaldehyde give linear plots except at the lower concen-
trations (<0.02 mg der/m£ toluene). Benzaldehyde gives a plot which is not
linear above 2.0 mg/m£ toluene. The benzaldehyde-DNPH derivative is not
soluble in toluene at concentrations greater than 2.0 mg/m£. This fact should
be taken into account if high concentrations of benzaldehyde are expected
(>5 ppm for a 23 minute sampling period at 4 £/min),
An experiment was carried out to determine the extraction repeatability
for the DNPH procedure at low concentrations of DNPH-aldehyde derivatives.
One liter of DNPH absorbing solution containing small amounts-of pure for-
maldehyde, acetaldehyde, acetone, methylethylketone, crotonaldehyde, hexan-
aldehyde, and benzaldehyde DNPH derivatives was prepared. Seven extractions
(80 m£ for each extraction) were carried out over a period of two weeks.
The results from the extractions are presented in Table 3. These results
were determined in units of mg DNPH derivative/m£ of toluene. The values
for each of the seven extractions, the average, and the standard deviation
are listed for each of the aldehydes and ketones. Multiple injections were
also carried out on each sample over the two week period. The values ob-
tained generally did not vary more than 0.002 mg derivative/mil toluene
(except for two DNPH-acetone values). This finding indicates, for several
of the compounds, that a large part of the variation in values was due to the
extraction process and not the lack of injection repeatability.
The results of these experiments indicate that the variability in the
-------
TABLE 2. INJECTION REPEATABILITY
DNPH Aldehyde
or Ketone
Derivative
Formaldehyde
Acetaldehyde
Acetone
Isobutyraldehyde
Methylethylketone
Crotonaldehyde
Hexanaldehyde
Benzaldehyde
1.6 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
1.567
1.727
1.617
1.561
1.575
1.781
1.682
1.710
0.018
0.022
0.018
0.017
0.029
0.053
0.105
0.165
1
1
1
1
1
3
6
9
.1
.3
.1
.1
.8
.0
.2
.6
0.2 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
0.188
0.201
0.210
0.208
0.206
0.206
0.204
0.222
0.007
0.001
0.002
0.002
0.003
0.003
0.003
0.013
3.7
0.5
0.9
1.0
1.5
1.5
1.5
5.9
0.020 mg
Avg . for
5 Inject.
0.022
0.024
0.027
0.020
0.024
0.015
0.015
0.025
derivative/ml
Standard
Std. %
Dev. Dev.
0.002
0.002
0.001
0.001
0.001
0.001
0.001
0.008
9.1
8.3
3.7
5.6
4.2
6.7
6.7
32
0.002 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
0.007 0.0007 10
0.002 0.0007 35
0.003 0.0007 23
0.002 0.0007 35
0.001 0.0007 70
0.001 O.OO/I 110
-------
10.0 -
0.001
0.002
0.005 0.01 0.02 0.05 0.1 0.2 0.5 1-0 2.0
Actual concentration (mg formaldehyde-DNPH derivative/ml toluene)
5.0
10.0
Figure 1. Plot of the formaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
-------
10.0
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0
Actual concentration (mg acetaldehyde-DNPH derivative/ml toluene)
10.0
Figure 2. Plot of the acetaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
-------
I
4>
a
tc
o.
§
M
3
•O
§
10.0
5.0
2,0
1,0
0.5
0.2
0.1
0.05
•O
g 0.02
O.01
0.005
4J
C
I
0.002
0.001
0.002 0.005 0,01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
Actual concentration (rag acetone-DNPH derivative/ml toluene)
5.0
10.0
Figure 3. Plot of the acetone-DNPH derivative concentration
determined by procedure vs actual concentration.
-------
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0
Actual concentration (mg isobut.yraldehyde-DNPH derivative/ml toluene)
10.0
Figure 4. Plot of the isobutyraldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
10
-------
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0
Actual concentration (mg methyl.ethylketone-DNPH derivative/ml toluene)
10.0
Figure 5. Plot of the methylethylketone-DNPH derivative concentration
determined by procedure vs actual concentration.
11
-------
10.0
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0 5.0
Actual concentration (mg crotonaldehyde-DNPH d&rivative/ml toluene)
10.0
Figure 6. Plot of the crotonaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
12
-------
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1,0 2.0 5.0
Actual concentration (mq hexanaldehyrde-DNPH derivative/ml toluene)
10.0
Figure 7. Plot of the hexanaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
13
-------
10.0
0.00
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
Act-.ual concentration (mq benzaldehyde-DNPH derivative/ml toluene)
5.0 10.0
Figure 8. Plot of benzaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
14
-------
TABLE 3. MULTIPLE EXTRACTIONS OF DNPH SOLUTIONS
(All Units are mg DNPH derivative/m £ of Toluene)
Form- Acet- Cronton- Hexan- Benz-
Extraction aldehyde aldehyde Acetone MEK aldehyde aldehyde aldehyde
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Average
Standard
Deviation
0.018
0.023
0.020
0.009
0.014
0.020
0.014
0.017
±0.005
0.011
0.037
0.032
0.013
0.019
0.031
0.027
0.024
±0.010
0.005
0.015
0.015
0.005
0.007
0.014
0.043
0.015
±0.013
0.000
0.003
0.000
0.000
0.000
0.004
0.014
0.003
±0.005
0.003
0.004
0.003
0.002
0.003
0.005
0.004
0.003
±0.001
0.001
0.002
0.001
0.000
0.002
0.002
0.000
0.001
±0.001
0.014
0.027
0.044
0.000
0.014
0.004
0.007
0.016
±0.015
15
-------
extraction process for concentrations of DNPH aldehyde and DNPH ketone
derivatives below 0.025 mg DNPH derivative/m£ toluene can be very signifi-
cant (i.e., 94 percent for benzaldehyde at 0.016 mg/m£). This variability
needs to be taken into account when evaluating data obtained using this
procedure. At higher DNPH derivative concentrations (0.2-2.0 mg derivative
per m£ toluene) the overall test variability (trapping, extraction and
injection) is approximately 10 percent. This value was obtained from the
standard' deviation of tunnel recovery and trapping efficiency experiments.
The DNPH analysis for the aldehydes and ketones has given abnormally
high concentrations of crotonaldehyde and benzaldehyde in isolated occasions.
A gas chromatography-mass spectroscopy study was carried out on three samples
obtained from EPA contract 68-03-2499. The three samples either contained
abnormally high concentrations of crotonaldehyde DNPH derivative or benz-
aldehyde-DNPH deriavtive or both. The results from this study revealed that
neither crotonaldehyde nor benzaldehyde was present in the samples. Further
gas chromatography-mass spectroscopy studies were carried out on two ot the
samples to determine what compounds were present. In both samples, the
crotonaldehyde peaks were due to a phthalate, and the benzaldehyde peaks
could not be identified. This study revealed that the samples contained
several other phthalates as well as di-2-ethylhexyladipate (a fuel stabi-
lizer) . Many phthalate esters (e.g., dioctyl, dibutyl, dimethyl, etc.,)
are found in lubricants and plastics. It is possible that the phthalate
peaks found in the above samples were due to contamination in the extraction
process (e.g., from a pipette bulb, etc.). In subsequent testing, extreme
care will be taken to assure the samples do not come into contact with
plastics and other materials which could cause contamination. It is also
possible that some of the phthalates which are found in small quantities in
the samples are from the exhaust (originating from lubricants) and are
possible interferences in the procedure. Also, the di-2-ethylhexyladipate
appears to produce a minor interference.
QUALIFICATION EXPERIMENTS
Qualification Experiments were carried out using a 1972 Plymouth Fury
with a 360V-8 engine (no catalyst). Hot FTP (23 minute test) driving cycles
were followed to generate exhaust for the vehicle baseline emissions and for
the tunnel injection + vehicle experiments. An aluminum cylinder containing
1350 ppm propionaldehyde in balance nitrogen was used as the source for
aldehydes in the experiments. The cylinder was named using the aldehyde-
DNPH procedure. The flow of propionaldehyde into the tunnel was regulated
to give a concentration of 1-2 ppm propionaldehyde in the dilution tunnel.
Injections of propionaldehyde into the tunnel without exhaust gave recoveries
that ranged from 85 percent to 121 percent with an average of 103 percent
(Table 1). The recovery of propionaldehyde with real vehicle exhaust ranged
from 78 percent to 115 percent with an average of 100 percent (Table 4) . The
injections with the vehicle were corrected for the vehicle baseline emission
of propionaldehyde. If propionaldehyde is representative of the aldehydes,
then it appears that there is little or no loss of the aldehydes in the
dilution tunnel with or without exhaust present.
16 .
-------
TABLE 4. PERCENT RECOVERY OF PROPIONALDEHYDE
Tunnel Only Tunnel + Vehicle
Run Recovery % Run Recovery %
1 117 1 90
2 - 116 2 115
3 85 3 106
4 89 4 106
5 121 5 85
6 89 6 103
Avg 103 7 78
8 110
Avg 100
RESULTS AND CONCLUSIONS
The concentration of aldehydes and ketones in dilute exhaust can be
determined by (1) trapping the aldehydes and ketones in a DNPH/HC1 absorbing
solution, (2) removing the resulting derivative from the absorbing solution
by filtration and extraction with pentane, (3) evaporating off the pentane
(4) dissolving the dried extract in toluene, and (5) analyzing the resulting
solution with a gas chromatograph equipped with a flame ionization detector.
The aldehydes and ketones are effectively trapped in the absorbing solution
at a flow rate of 4 £/min. The procedure has a minimum detection limit of
approximately 5 ppb. This carbonyl concentration in the exhaust gives a
corresponding concentration of 0.002 mg/m£ in toluene.
The accuracy of the procedure in the 0.5-20 ppm concentration range for
the aldehydes or ketones in dilute exhaust is approximately 10 percent. The
accuracy of the procedure in the 0-0.05 ppm range is not as good and values
can vary as much' as 100 percent. The gas chromatograph system gives a linear
response for acetone, methylethylketone, and crotonaldehyde DNPH-derivative
concentrations between 0.002 and 8 mg derivative/mil toluene and gives a
linear response for formaldehyde, acetaldehyde, isobutyraldehyde and hexan-
aldehyde DNPH derivative concentrations between 0.02 and 8 mg derivative/m&
toluene. The benzaldehyde-DNPH derivative give a linear response in the 0.02
to 2 mg derivative/m£ toluene concentration range. The benzaldehyde
derivative is not soluble at concentrations greater than 2 mg/m£ toluene.
Phthalates and di-2-ethylhexyladipate were found by mass spectroscopy
to be interferences in the procedure. Many phthalate esters (e.g., dioctyl»
dibutyl, dimethyl, etc.) are found in lubricants and plastics, and di-2-
ethylhexyladipate is used as a fuel stabilizer. Contamination from phtha-
lates could occur in the extraction process or in sample storage if the
sample is allowed to come into contact with plastics, a pipette bulb, a
lubricant, etc. It is also possible that some phthalates originate from the
exhaust (from lubricants) and are possible interferences in the procedure.
The benzaldehyde and crotonaldehyde values can be affected by these inter-
ferences. The interfering peak in the region of benzaldehyde is usally broad
17
-------
and the benzaldehyde peak, if present, can be observed on top of this inter-
ference. If care it taken, a reliable value can be determined for the benz-
aldehyde. Any value reported for crotonaldehyde may be artifically high due
to possible phthalate contamination. Extreme care must be taken when
handling the sample in order to eliminate any possibility of contamination
after collecting the sample and before analysis.
Propionaldehyde can be recovered quantitatively from the dilution
tunnel with or without exhaust present. If propionaldehyde is representa-
tive of the aldehydes and ketones, there is little or no loss of the
aldehydes in the dilution tunnel with or without exhaust present.
Overall the DNPH procedure should provide a relatively accurate method
for determining the concentration of aldehydes and ketones in dilute exhaust,
and its use is recommended for this project.
18
-------
SECTION 3
ORGANIC AMINE PROCEDURE
LITERATURE SEARCH
The individual amines that are included in this analysis are monome-
thylamine, dimethylamine, monoethylamine, trimethylamine, diethylamine and
triethylamine. The chemical formulas, molecular weights, boiling points,
freezing points, and synonyms for these low molecular weight aliphatic amines
are presented in Table 5. In general, these amines have a fish-type odor at
lower concentrations, but more of an ammoniacal odor at higher levels. The
1968 American Conference of Governmental Industrial Hygientists has recom-
mended a threshold limit value of 10 ppm.
The measurement of individual low molecular weight amines has been con-
ducted using a variety of gas chromatograph techniques. Hoshike (6,7) re-
ported gas chromatographic separation of lower aliphatic amines in the free
form and as their Schiff base derivatives. A glass column was employed to
provide a separation of 11 amines using temperature programming and a ther-
mal conductivity detector. This work was directed toward achieving a satis-
factory separation rather than being concerned with minimum detection limits.
Sze (8), et al reported separation of methyl amines, ammonia, and methanol
using a mixture of tetrahydorxyethylethlenediamine and tetraethylenepenta-
mine. O'Donnel and Mann (9) used Dowfax 9N9, Carbowax 400, and Carbowax 20M
to separate mixtures of aliphatic amines, aromatic amines, and aliphatic
amines. This work was performed using synthetic blends on a gas chromato-
graph with a thermal conductivity detector. McCurdy and Meiser (10) used a
gas chromatograph with a flame ionization detector to determine fatty amines
in trace quantities. The fatty amines were converted to trifluoracetyl
derivatives, providing a sensitivity of 0.05 ppm fatty amine in water.
Smith and Waddington (11) used aromatic polymer beads to separate a
wide range of aliphatic amines. Peak tailing was found to exist because of
two types of active sites on the polymer: simple acidic sites which can be
neutralized by treatment with base, and metal ions which must be deactivated
by addition of an involatile complexing agent. Glass columns were used in a
gas chromatograph with a flame ionization detector. Synthetic blends rang-
ing from CI-CQ were separated and analyzed using this approach. In another
study Carbopak B/4 percent Carbowax 20 M/0.8 percent KOH (12) and 28 percent
Pennwalt 223/4 percent KOH (13) have been reported to give satisfactory sep-
arations of lower aliphatic amines.
Analysis of amines as derivatives has been shown to be a valuable
analytical tool to determine trace quantities (14). Thirteen different
19
-------
TABLE 5. LIST OP INDIVIDUAL ORGANIC AMINES
INCLUDED IN THE EMISSIONS CHARACTERIZATION INVENTORY
Name
Monomethylamine
Monoethylamine
Dimethylamine
Trimethylamine
Diethylamine
Triethylamine
(5)
(5)
(5)
(4)
(4)
(4)
Carbon
No.
1
2
2
3
4
6
Chemical
Formula
CH3NH2
C2H5NH2
(CH3) 2NH
(CH3)3N
(C2H5)2NH
(C HC1N
Molecular
Weight
31.058
45.085
45.085
59.112
73.14
101.19
Boiling
Point, °C
- 6.32
16.58
6.88
2.87
56.3
89.3
Freezing
Point, °C
- 93.5
- 81.0
- 92.19
-117.08
- 50.
-114.7
Synonyms
Methylamine ,
aminome thane
Ethylamine ,
aminoethane
None
None
None
None
-------
derivatives were evaluated in terms of FID and BCD response characteristics.
This work was limited to primary amines, and under optimum condidtions
amines down to 10 picograms could easilty be quantified using an BCD detector.
Clark and Wilk (15-) used an BCD to evaluate the properties of halogenated
amine derivatives. No increase in the sensitivity for the trifluoroacetyl
amine derivatives using BCD was observed.
Hosier (16), et al quantitatively measured aliphatic amines volatilized
from cattle feedyards. Direct gas chromatograph injection of acid solutions
and GC separation of the pentafluorobenzoyl derivatives of the malodorous
volatiles were used in identification. The derivatized amines were analyzed
using a gas chromatograph equipped with an electron capture detector. The
two columns used in this work were 6 percent DECS and 10 percent Igepal
CO 880.
Methylamine and ethylamine were detected in irradiated beef by Burks
(17) , et al. Several techniques, including colorimetric paper chromato-
graphy and gas chromatography, were used in quantifying results. Gas
chromatogrphic determination of free mono-, di-, and trimethylamines in
biological fluids was performed by Dunn (18), et al. A flame ionization
detector was used to quantitatively separate the lower aliphatic amines.
Separation of mono-, di-, and trimethylamine from extracts of fish tissue
was achieved by Gruger (19).
Andrea (20), et al developed a precolumn inlet system for the gas
chromatographic analysis of trace quantities of short-chain aliphatic
amines. Losses inherent in the collection and direct gas chromatograph
analysis of field air samples containing volatile amines necessitated an
indirect analytical scheme. A teflon tube (3" x 5/16" OD) was filled with
20/30 mesh Ascarite and placed in the injector inlet of the gas chromato-
graph. Samples were collected in dilute sulfuric acid and aliquots were
injected into the pre-column of the GC. Release of the free amines was
found to be sufficiently reproducible for quantification of results. This
technique avoided the problems encountered by Umbreit (21), et al, and
Hardy (22) when using base loaded columns to analyze acidified aqueous
solutions of amines from fish. The in situ release of the free amines from
their salts produced a chromatographic column that changed with every in-
jection. In addition, the column had a very short usable lifetime and
lacked reproducibility after extended use.
Bowen (23) described a gas chromatograph procedure for the analysis of
aromatic amines using an adsorption technique. Quantitative adsorption and
desorption of aromatic amines using Tenax GC was demonstrated at the nanogram
level. Samples were pulled through the Tenax GC trap for specific sampling
periods, thermally desorbed at 250°C, and analyzed in a GC with a FID. The
author recommended use of a NPD to increase sensitivity for aromatic amines
on the tail of hydrocarbon solvents and eliminate venting the solvent.
PROCEDURAL DEVELOPMENT
From the results of the literature search it was determined that the
analysis of the amines should be conducted by the use of gas chromatography.
21
-------
A Perkin-Elmer Model 3920B gas chromatograph was dedicated for_this purpose.
This instrument has a dual/differential electrometer and has linear tempera-
ture programming capabilities along with a sub-ambient oven accessory. The
instrument has been equipped with a flame ionization detector (FID), a ni-
trogen phosphorus detector (NPD), and an electron capture detector (BCD) and
can be connected to a chemiluminescent detector. Of the specialty detectors
available for the analysis of the amines, the NPD appeared to be the prime
candidate and initial work was carried out using this detector. Because the
amines are notorious for tailing and reacting with metal sites, a glass
lined heated injector port and a glass interface were installed so that with
the use of a glass column, the system would be glass throughout.
Lecture bottles of methylamine, dimethylamine, trimethylamine, and
ethylamine, along with pure liquids of diethylamine and triethylamine, and
a Tracor Model 412 Permeation Calibration System containing permeation tubes
of all six amines were used as sources of the organic amines in the pro-
cedural development experiments. These sources allowed a method for pre-
paring blends of varing concentrations of the organic amines.
Several column packing materials (all were developed especially for the
analysis of amines) and column lengths were evaluated to determine which
could provide the best peak separation with the shortest analysis time. The
columns evaluated included: 12" x 1/4" glass columns packed with 28 percent
Pennwalt 223 amine packing, a 6' x 4mm (id), 6' x 2 mm (id) and a 12' x 1/4"
glass column packed with 4 percent Carbowax 20M and 0.8 percent KOH on
Carbopak B, and a 6' x 4 mm (id) glass column packed with 2 percent KOH on
Chromosorb 103. Several column temperatures, programming rates, and carrier
flow rates (helium) were tried for each of the columns.
The best separation was accomplished using the 4 percent Carbowax 20M
and 0.8 percent KOH on Carbopak B packing material in the 6' x 4 mm (id)
glass column. This column would only partially separate dimethylamine and
ethylamine under conditions which gave very broad peaks at long retention
time. In order to increase sensitivity and shorten the analysis time an
initial temperature of 130°C was chosen. At this temperature, the dimeth-
ylamine and ethylamine coalesce into a single sharp peak, and the analysis
time is under 30 minutes.
The FID, NPD, and chemiluminescent detector were evaluated as detectors
for the organic amines. The NPD was more sensitive than either the FID or
the chemiluminescent detector. Also the NPD is only sensitive to compounds
containing both carbon and nitrogen, eliminating many interferences that
would be present using the FID.
In order to analyze automobile exhaust for the organic amines, it was
found necessary to concentrate the amines in a trap or an absorbing reagent
to obtain enough sample for a satisfactory analysis. Two collection pro-
cedures were evaluated, one in which the amines are collected in a tran
filled with 1 gram Tenax-GC packing material, and another in which the
amines are collected by bubbling the amines thr-nnr^ = -j
was some breakthrough of the amLes through SrSLS^^^Tit *""
liquid nitrogen temperatures. Bubbling the amines through an acid solution
22
-------
proved to be the superior method for collecting the amines.
A sulfuric acid solution was found to effectively trap the amines at
room temperature. To release the amines from the sulfuric acid solution into
the GC column, an Ascarite loaded precolumn was installed into the injector
block of the GC. This precolumn was found to work very well in releasing the
amines into the GC column; however, the lifetime of the precolumn was found
to vary from one injection to several hundred. The 4 percent Carbowax 20M
and 0.8 percent KOH on Carbowax B packing material in the GC column was de-
signed to be used with aqueous solutions and proved to be satisfactroy when
used with aqueous sulfuric acid solutions.
The usefulness of the precolumn was usually terminated by the aqueous
injections temporarily dissolving the Ascarite and the Ascarite redrying to
form a plug, thereby preventing the sample from entering the column. Some
time was spent on trying to determine why some precolumns lasted for only
one injection while others for several hundred, but the results were incon-
clusive.
Poor injection repeatability resulted from a variety of problems. These
problems included previous GC injection histroy, glass syringe purging tech-
nique, precolumn conditioning, and columneffects . Because of the problems
mentioned above, and alternate method of analysis was evaluated. This
method consisted of collecting the organic amines in glass impingers using
dilute sufuric acid, converting the trapped amines to their pentafluoro-
benzayl chloride derivatives, and analyzing for these derivatives using a
gas chromatography equipped with an electron capture detector. It was hoped
that this method would (1) convert the amines to stable derivatives which
would improve sample injection repeatability, and (2) provide improved de-
tection limits with an electron capture detector.
This alternate procedure was found to be unsuitable for the detection of
amines at the ppb levels. Tertiary amines (trimethyl- and triethylamine) can-
not be detected by this procedure, and the secondary amines (dimethyl- and di-
ethylamine) had a low sensitivity that made detection in the ppb range almost
impossible. The peak areas of the primary amines (methyl- and ethylamine)
were found to be time dependent. The GC peak areas of the methyl- and ethyl-
amine derivatives were plotted against the time they were allowed to stand
after initial mixing of the reagents to produce the derivatives. The mixing
procedure is included in Table 6. The standing time includes 2 minutes of
vigorous shaking (1 minute for the 1 minute test) plus any remaining time
the mixture was allowed to stand at room temperature before injection into
the GC. The effect of elapsed time on peak area is shown in Figure 9. This
figure shows a rapid increase in peak area followed by a rapid decrease in
the area. In order to obtain reproducible data at ppb levels, the injection
time after mixing can vary by only seconds. Under normal operating condi-
tions this would not be possible. Because of the time limitation, the
procedure was abandoned for the quantitative analysis of the organic amines.
The GC-NPD procedure using the Ascarite precolumn has been reevaluated
and most of the problems involved with its use have been solved. The
inconsistent lifetime of the precolumn remains a problem in the procedure.
23
-------
TABLE 6. MIXING PROCEDURE FOR PREPARATION OF
PENTAFLUORABENZOYLAMINE DERIVATIVES *
1. Pipette 1.0 m£of 0.01 N sulfuric acid containing 94 ppb
methylamine and 220 ppb ethylamine into a 10 m & reacti-vial.
2. Pipette 3 m£,of toluene into reacti-vial.
3. Pipette 1 m£ of pentafluorobenzoyl chloride (PFBC) solution
(50 vg PFBC in 100 m.£ toluene) into reacti-vial.
4. Pipette 1 mJJ. of 10 percent aqueous potassium hydroxide solution
into reacti-vial.
5. Shake and allow to stand for X minutes.
* General procedure from private communication with Arvin R.
Hosier, USDA, ARS, P. O. Box E, Fort Collins, Colorado 80521
24
-------
35
30
25
o
o
20
fl)
0)
8
15
10
I
Muthylamine Derivativ
Ethylamine Derivative
I
0 123456789 10
Time elapsed (in minutes) after mixing and before injection
11
Figure 9. GC peak areas of pentafluorobenzoyl amine derivatives vs. time.
25
-------
The syringes can be cleaned by purging several times with the next sample to
analyzed Memory effects in the precolumn and GC column are not a problem
as long as a blank is injected into the system after injection of a sample
with an amine concentration greater than one ppm. This blank purges out
the system for the-next sample. The repeatability of the system is also
improved if a series of 4-5 injections of a solution containing 1 ppm of
each of the amines is made into the system. This must be done each time the
instrument has been unused for periods of greater than one hour.
The procedure chosen for the analysis of the organic amines consists of
trapping the amines in dilute sulfuric acid solution and analyzing the
solution using a GC equipped with an Ascarite precolumn, a 6' x 4 mm glass
column packed with 4 percent Carbowax 20M and 0.8 percent KOH on Carbopax B,
and a nitrogen-phosphorus detector. A finalized copy of the procedure is
included as an appendix to the interim report.
VALIDATION EXPERIMENTS
Several experiments were carried out to determine the validity of the
amine procedure for the analysis of the organic amines. The experiments
included checks for GC injection variability, linearity of detector response,
sample stability in the absorbing solution, and trapping efficiency of the
0.01 N sulfuric acid solution.
The finalized sampling conditions used to collect the organic amines are
listed below, as is a discussion of their selection. A single glass impinger
containing 25 m & of 0.01 N sulfuric acid is used to collect the organic
amines. This single impinger traps 99+ percent of the organic amines at low
ppm and ppb amine concentrations. This collection efficiency was determined
by bubbling known amounts of organic amines through a series of impingers
and analyzing each impinger separately. No advantage was found in using
more than one impinger or higher concentrations of sulfuric acid except when
the concentration of the amines exceeded 5 ppm. The 0.01 N sulfuric acid
concentration was selected over higher acid concentrations (0.1 and 0.1 N)
in order to prevent the neutralization of the ascarite in the precolumn any
sooner than necessary. The concentration of the organic amines in exhaust
should never approach the 5 ppm concentration; therefore, the single glass
impinger containing 0.01 N sulfuric acid should be sufficient to trap the
organic amines. Sulfuric acid was chosen over hydrochloric acid as the
absorbing acid because of its higher boiling point. Hydrochloric acid is
more volatile and could vaporize into the analytical column during analysis.
During sampling, the impinger is kept in a 0°C ice bath. The ice bath offers
no significant advantage in collection efficiency over room temperature, but
does provide a stable sampling temperature during the test. The 0°C tem-
perature also lowers the vapor pressure of the aqueous absorbing solution
and thus prevents loss of any significant amount of water from the absorbing
solution during sampling. The sample flow rate through the impingers is
maintained at 4 liters a minute. This flow rate provides the largest
amount of sample flow through the absorbing reagent without loss in absorbing
efficiency or the physical loss of any absorbing reagent. A teflon sample
line connecting the CVS to the impingers is heated to 77 °C in order to prevent
water from condensing in the sample line. The organic amines are water
26
-------
soluble, and the condensation of water in the sample line could cause a
significant loss of sample in the sample line.
Samples have been found to be stable in the sulfuric acid absorbing
solution for months. A two month standard containing 0.1 ppm of mono-, di-,
and trimethylamine showed no significant decrease in concentration when com-
pared to a freshly prepared standard. The absorbing solution is also stable
over long periods of time with the only worry being contamination from any
amines which might be present in the laboratory environment.
To determine the GC injection repeatability for the procedure over a
wide range of concentrations, three standards containing 0.01, 0.1, and 1
ppm of mono-, di-, and triethylamine were prepared. Each standard was
injected into the GC ten consecutive times. The area of each resulting peak
was averaged over the ten runs and a standard deviation was calculated. The
results of the injection repeatability experiments are presented in Table 7.
Injections of the 0.01 N sulfuric absorbing solution were also made into the
GC system. Peaks for monomethylamine and ddjnethylamine/monomethylamine (the
two compounds give one peak in the procedure and are analyzed together as
Cof^N) were detected in the absorbing solution and gave areas which corre-
sponded to 50 percent of the area for the monomethylamine and dimethylamine
in the 0.01 ppm standard and 20 percent of the monoethylamine in the 0.05
ppm standard. The procedure is not as sensitive to the ethylamines as it
is to the methylamines; therefore, higher concentrations of the ethylamines
(10, 1, 0.1 ppm) were used in the repeatability experiments. For the meth-
ylamines the injection repeatability improves with increasing concentration
of the methylamines. The standard deviation for the 1 ppm standard con-
taining mono-, di-, and trimethylamine is 5-6 percent, while the deviation
for the 0.1 ppm standard is slightly higher at 7-8 percent. The standard
deviation for the 0.01 ppm standard is even larger at 12-21 percent. Con-
centrations at or below 0.01 ppm of the methylamines are difficult to
determine due to the poor injection repeatability and the interference from
the absorbing solution. The injection repeatability follows no definite
trend for the ethylamines. The standard deviations for mono-, di-, and tri-
ethylamine remain relatively constant at the three concentrations studied
(0.1, 1, and 10): 7-8 percent for monoethylamine, 4-7 percent for dimeth-
ylamine, and 7-10 percent for triethylamine. Concentrations below 0.05 ppm
of the ethylamines are difficult to determine due to the broadness of the
diethylamine and triethylamine peaks and to the interference from the
absorbing solution.
To determine the linearity of the nitrogen-phosphorus detector for each
of the amines at the concentration ranges of interest, standard solutions
containing 0.01, 0.05, 0.1, 0.5, and 1 ppm of mono-, di-, and trimethylamine,
and 0.05, 0.1, 1, 5, and 10 ppm of mono-, di-, and triethylamine were pre-
pared. These were made by weighing out required amounts of each of the
organic amines-hydrochloric acid salts and dissolving them in the proper
amount of sulfuric acid absorbing solution. Figures 10-15 show plots of the
GC peak areas versus the concentration for each of the organic amines on a
log-log scale. The relative GC areas for monomethyl-, dimethyl- and mono-
ethylamine were corrected for the background peaks found in the absorbing
solution. Mono-, di-, and trimethylamine give linear GC responses from
27
-------
TABLE 7. INJECTION REPEATABILITY EXPERIMENTS
Amine
Monome thylamine
Monome thylamine
Monome thy 1 ami ne
Dime thylamine
Dime thylamine
Dime thylamine
Trime thylamine
Trime thylamine
Trime thylamine
Monoe thylamine
Monoe thylamine
Monoe thylamine
Diethylamine
Diethylamine
Diethylamine
Triethylamine
Trie thylamine
Triethylamine
Concen-
tration (ppm)
1
0.1
0.01
1
0.1
0.01
1
0.1
0.01
10
1
0.1
10
1
0.1
10
1
0.1
Average
Area
8803
2004
1305
7081
1385
1006
5778
1044
344
10,943
8189
2748
7025
3460
724
10,921
5446
1481
Standard
Deviation
533
134
276
372
96
117
276
87
46
977
626
178
483
134
32
1014
564
110
Percent
Deviation
6.1
6.7
21.1
5.3
6.9
11.6
4.8
8.3
14.2
8.4
7.6
6.5
6.9
3.9
4.4
9.3
10.4
7.4
28
-------
10.0
5.0
2.0
~ 1.0
•rt
rt
u
U H
(U O
!>
•H -M
4J (d
td
H It!
(U (U
0.5
0.2
0.1
0.05
0.02
0.01
I
I
I
0.001 0.002 0.005
0.01 0.02 0.05 0.1
Concentration (ppm)
0.2
0.5 1.0
Figure 10. Linearity of monomethylamine GC Response
(plot on log-log scale).
29
-------
10.0
0.02 r-
0.01
0.001 0-002
0.005 0.01 0.02 0.05 0.1
Concentration (ppm)
0.2
0.5 1.0
Figure 11. Linearity of dimethylamine GC response
(plot on log-log scale) •
30
-------
10.or
5.0 -
•H
-------
100.0 -
0,
0.01 0.02
0.05
0.1 0.2
Concentrate on
0.5
(ppm)
1.0
2.0
5.0 10.0
Figure 13, Linearity of monoethylamine GC response
(plot on log-log scale).
32
-------
ioo.or-
so.o -
0.01 0.02 0.05 0.1 0.2 0.5
Concentration (ppm)
1.0
2.0
5.0 10.0
Figure 14. Linearity of diethylamine GC response
(plot on log-log scale).
33
-------
100.0
50.0 —
0.1
0.01 0.02 0.05 0.1 0.2 0.5 1.0 3.0
Concentration (ppm)
5.0
10.0
Figure 15. Linearity-of triethylamine GC response
(plot on log-log scale).
34
-------
0.01 to 1 ppm and di-, and triethylamine give linear responses in the 0.05
to 10 ppm region. Monoethylamine gives a linear response from 0.05 ppm to
1 ppm, but shows some deviation from linearity in the 1.0 to 10 ppm range.
Ammonia at concentrations between 10-100 ppm in sulfuric acid does not
give as large as a peak as does a 0.01 ppm solution of monomethylamine, and
therefore does not present any major problems as an interference. However,
the retention time (0.6 min) is close to that of monomethylamine (0.85 min)
and care must be taken not to confuse one peak for the other. Acetonitrile
(CH3CN) has been found in exhaust at concentrations near the 0.1 ppm level.
The NPD is sensitive to this compound and gives a peak in the chromatogram
at a retention time of 1.8 minutes. This retention time is near that of
trimethylamine (2.0 min) and care must be taken not to confuse the two com-
pounds. No other compounds in exhaust have been found to be interferences
in the procedure.
QUALIFICATION EXPERIMENTS
Injection experiments were carried out with three amines with the CVS
tunnel. The amines included in this study were methylamine, dimethylamine
and trimethylamine. A blend of the three amines was prepared in a spectra-
seal aluminum cylinder, and this cylinder was used to inject the amines into
the dilution tunnel.
A test sequence was developed to determine the injection recovery for
the three methylamines. Three tests were conducted for the amines at ppm
levels ranging from 0.12-0.19 ppm. Each test was conducted on a sequence
basis with a 10 minute soak with the CVS off between each 30 minute collec-
tion interval. After the 10 minute soak of the third test, samples were
collected to determine if the organic amines could be purged from the sys-
tem. After the fifth test, this was extended to a two hour soak.
The results of the organic amine recovery experiments are presented in
Table 8. As expected, the recovery of the methylamine was very low, 6-13
percent for the amine injections. Dimethylamine recovery increased from 18
to 36 percent recovery after three consecutive injections. Recoveries from
trimethylamine were more reasonable with 76-93 percent recovery for three
consecutive tests. Methylamine, dimethylamine, and trimethylamine could be
detected in the tunnel even after three hours with the CVS blower on.
At this time, it is uncertain as to the precise reasons for the losses,
but all possible steps have been made to preserve the integrity of the
sampling system and the sample handling prior to injection into the gas
chromatograph. It is doubtful that any substantial improvement could be
made to the system without going to heating the tunnel, etc. The losses
of the low molecular weight amines were not unexpected and these experiments
confirmed those fears.
In summary, methyl- and dimethylamine had low recoveries in the tunnel
without exhaust, and after CVS air purging of the tunnel, both amines were
detected in measurable quantities. Trimethylamine recoveries improved with
the number of consecutive injections and could be essentially purged after
35
-------
TABLE 8. SUMMARY OF AMINE RECOVERIES (TUNNEL ONLY)
AND TUNNEL HISTORY EFFECTS WITH METHYLAMINES
Percent Recovery
Test
1 (amines injected)
10 min soak-CVS off
2 (amines injected)
10 min soak-CVS off
3 (amines injected)
10 min soak-CVS off
4 (no amines injected)
1 hour soak-CVS on
5 (no amines injected),
2 hour soak-CVS on
6 (no amines injected)
0.19 ppm
Methylamine
6.2
10.1
13.1
2.0%*
2.5%*
1.1%*
0.14 ppm 0.12 ppm
Dimethylamine Trimethylamine
18
29
36
4.8**
3.7**
2.6**
76
89
93
5,1%***
1.7%***
-"0.3%***
Note: All sample collections were over a 30 minute test period. Tests
1-3 had amines injected whereas Tests 4-6 had only the CVS on.
* based on amount of methylamine injected during Test 3
** based on amount of dimethylamine injected during Test 3
*** based on amount of trimethylamine injected during Test 3
36
-------
three hours with the CVS on. Based on the results of these experiments,
injection experiments with actual vehicle exhaust were not conducted.
RESULTS AND CONCLUSIONS
The concentration of organic amines in dilute exhaust can be determined
by collecting the amines in 0.01 N sulfuric acid and analyzing the solution
with a GC equipped with an ascarite precolumn and a nitrogen phosphorus
detector. The amines are effectively trapped in 25 ml of 0,01 N sulfuric
acid absorbing solution at a flow rate of 4 &/min. For a twenty-three minute
test and a sample flow rate of 4 £/min, the procedure has a minimum detec-
tion limit of 2 ppb for each organic amine.
The accuracy of the procedure decreases as the concentration of the
amine in the absorbing solution decreases. At a 0.01 ppm concentration of
the organic amines in the absorbing solution, the percent standard deviation
for the GC is 12-21 percent. The absorbing solution itself gives peaks
equal to 0.005 ppm monomethylamine and dimethylamine/ethylamine. At the
0.01 ppm level or lower it is difficult or impossible to determine the con-
centration of amines. This concentration is equivalent to 2 ppb of the
amines in dilute exhaust (23 minute test, sampling at 4 £/min).
Acetonitrile and ammonia have been found in exhaust samples and give
peaks in the chromatograms. Ammonia at concentrations of 10-100 ppm gives
a peak approximately the size of a 0.01 ppm methylamine peak. The separation
of the ammonia and the methylamine peak is 0.25 minutes, but the two can
easily be distinguished if care is taken. Acetonitrile has a retention time
that differs from trimethylamine by only 0.2 minutes, but the two can also
be easily distinguished.
i
The amines are notorious for sticking to metal sites. The qualifica-
tion experiments represent another example of this problem. The amines had
low percent recoveries from the CVS dilution tunnel. The precent recovery
increased directly as the number of injections into the tunnel increased.
This phenomenon is probably due to the gradual coating of the tunnel with
the amines, thus neutralizing the number of metal sites in the tunnel. With
the CVS blower on the amines were slowly released from the tunnel up to
three hours after the amine injections were made into the tunnel. It is
possible that it the amines are present in concentrations of less than 0.1
ppm, the percent recovery may be very low or essentially zero. With this in
mind, raw exhaust was sampled for organic amines on EPA Contract 68-03-2499
and the results were compared to the corresponding results for CVS diluted
samples. The raw exhaust samples did not show a significant increase in the
concentration of the organic amines when compared to the dilute exhaust
samples. In most cases, no amines were detected in either raw or CVS
diluted exhaust.
The organic amine procedure should provide a relatively accurate method
for determining the concentration of the organic amines exiting the CVS
tunnel; however, amine losses in the CVS tunnel must be taken into account
when reporting these concentrations.
37
-------
SECTION 4
SULFUR DIOXIDE PROCEDURE
LITERATURE SEARCH
At room temperature and atmospheric pressure sulfur dioxide is a highly
irritating, nonflammable, and colorless gas. The gas is readily detectable
at concentrations of 3-5 ppm by the human sense of smell. Physical proper-
ties of sulfur dioxide, SO2 (sulfurous acid anhydride) include a freezing
point of -75.5°C (1 atm) , a boiling point of -10.0°C (1 atm) , and a molecular
weight of 64.063 (5).
The bulk of published literature regarding the analysis of sulfur di-
oxide has dealt with ambient air sampling. With the development of instru-
mental methods of analysis, the ability to measure sulfur dioxide in sta-
tionary and mobile source exhausts now exists. The following review of
references reveals a wide variety of analytical techniques used in the mea-
surement of sulfur dioxide concentrations.
A frequently used method for the analysis of sulfur dioxide is a color-
imetric method. The most commonly employed colorimetric technique is the
West-Gaeke method (24-28). This method has been collaboratively tested, with
the lowest concentration range studied being well above the levels most fre-
quently found in rural and global background air (29).
A modified version of the West-Gaeke method involves the collection of
sulfur dioxide in 0.1M sodium tetrachloromercurate(II) (TCM). Sulfur di-
oxide reacts with the TMC to form a dichlorosulfitomercurate complex (DCSM).
In this modified version, the DCSM resists oxidation by oxygen in the air
and oxygen dissolved in the absorbing solution. Ethylenediamine tetracetic
acid disodium salt (EDTA) is added to the TCM absorbing solution to complex
any heavy metals that could oxidize sulfur dioxide before the DCSM is formed
(30,31), and sulfamic acid is added to the absorbing solution to destroy any
interfering nitrite ion which might be present (32).
The colorimetric determination of sulfur dioxide is based upon the mea-
surement of the red-violet color produced by the reaction of DCSM with hy-
drochloric acid, pararosaniline and formaldehyde. The effect of the para-
rosaniline dye purity on the colorimetric procedure has been repbrted by
several researchers (33,34). Since the dye purity does effect the results
of the colorimetric procedure, various techniques for the purification of
commercial grade pararosaniline have been published (25,27,35), and para-
rosaniline purified especially for the colorimetric analysis of sulfur di-
oxide is commercially available (35).
38
-------
A major potential source of error associated with the West-Gaeke colori-
metric method for measuring sulfur dioxide is the widely differing collection
efficiency reported for Greenburg-Smith and midget impingers at low sulfur
dioxide concentrations (36). Urone, et al, investigated the collection ef-
ficiency of a TCM solution by the use of microgram quantities of sulfur
dioxide tagged with 35S (37). In this investigation, it was found that a
series of bubblers cannot be used to determine absorber collection efficiency.
Bostrom observed a 99 percent collection efficiency for a concentration range
100-1000 ppb sulfur dioxide in a TCM solution (38).
Work has been conducted in the development of other colorimetric methods
for the analysis of sulfur dioxide. Attari developed a procedure whereby
sulfur dioxide is absorbed into a solution of ferric ammonium chloride, per-
chloric acid, and phenanthroline dye (39) . A color complex with an absorb-
ance of 510 mm was formed, and although the color developed within 10 minutes,
it tended to fade with time. Hydrogen sulfide was found to be an interfer-
ence in the procedure.
Kawai used the reaction of barium chloranilate with sulfate as an in-
direct measurement of sulfur dioxide (40). Sulfur dioxide was absorbed in
a solution containing hydrogen peroxide and barium chloranilate. Barium
chloranilate reacts with the sulfate ion producing a red-violet chloranilic
acid ion. Although this method may be satisfactory for flue gas analysis, it
lacks the sensitivity required for ambient air analyses.
Conductivity methods have been used for continuously monitoring sulfur
dioxide in air (41). The conductivity of a dilute sulfuric acid-hydrogen
peroxide reagent changes due to the absorption of pollutants. This change
in conductivity is assumed to result primarily from sulfur dioxide absorbed
from the sampled air and oxidized to sulfuric acid. In many cases, sulfur
dioxide is the major pollutant present; however, if other pollutants are
present, their collection efficiency and solubility may be significantly
different than for sulfur dioxide. Several field comparisons of conductivity
with other sulfur dioxide procedures indicate a fair agreement (42-47).
Hydrochloric acid gas, ammonia, and chlorine substantially increase con-
ductivity. Shikiya and McPhee found two- to fourfold differences between
different conductivity analyzers and between conductivity and colorimetric
analyzers (44). Although the conductivity procedure may be acceptable for
point sources of sulfur dioxide in isolated areasf its high potential for
positive and negative interferences limits its application.
lodometric methods were among the first adapted for air pollution
analysis from the industrial hygiene litereature. With this method, the
sulfur dioxide is collected in an impinger containing a standard NaOH. The
absorbing solution is acifified and the liberated sulfurous acid is titrated
with standard iodine solution (45). Another method employs a standard
iodine-potassium absorbing solution (46). lodometric methods of analysis
for sulfur dioxide generally suffer from a lack of sensitivity and inter-
ferences from hydrogen sulfide-
Adsorption sampling methods have also been developed for the measure-
ment of sulfur dioxide (48). Sulfur dioxide is adsorbed on silica gel,
39
-------
desorbed, and reduced to hydrogen sulfide at 700-900°C over a platinum
catalyst. The hydrogen sulfide is then absorbed in a 2 percent ammonium
molybdate solution and determined colorimetrically. Although this tech-
nique is relatively specific for sulfur dioxide, the final colorimetric
determination by the molybdenum complex does not utilize the most sensitive
method available.
In addition to the aforementioned techniques, sulfur dioxide has been
measured by filtration (49-53) and static collectors (54-60) . Air samples
are passed through potassium bicarbonate impregnated filters and analyzed
for sulfate. The collection efficiency of these filters is dependent upon
humidity, temperature,- and the atmospheric concentration of sulfur dioxide.
The lead peroxide candle static collector was developed by Wilson and Mc-
Connell as an inexpensive method for measuring relative "sulfation" of the
atmosphere (54). The sulfur dioxide collection efficiency is dependent
upon temperature, relative humidity, wind speed, atmospheric concentration
of sulfur dioxide, and the length of exposure period (55). Ikeda determined
ambient sulfur dioxide levels by collecting samples on active carbon filters,
washing the filters with distilled water, and titrating with barium chlorani-
late.
With the advent of modern instrumental methods of analysis, specifically
gas and ion chromatography, a substantial amount of data has been published.
Most trace gas analysis for sulfur dioxide has been conducted using gas
chromatographs with flame photometric detectors (FPD) (61-70). The PPD is
highly selective for sulfur compounds and has low minimum detection limits.
Analysis for sulfur dioxide is generally performed using all teflon or glass
systems. Sulfur dioxide will react with active sites in the gas chromato-
graph system, making the use of inert materials essential for trace quanti-
tative analysis. Gas chromatographs with FPD and linearizing circuitry pro-
vide a wide dynamic range for ambient and source sulfur dioxide levels. In
some instances, the collection technique precludes the use of GC-FPD
techniques; i.e., bag sampling from dilute automotive exhaust or source
sampling. The use of gas chromatography would be a prime candidate if the
sample integrity could be assured in the sample acquisition and subsequent
analysis.
A more resent development in methods of analysis for sulfur dioxide
involves the use of ion chromatograph (71) . This technique involves col-
lection in a hydrogen peroxide absorbing reagent and measurement of the
resulting sulfate ion using ion chromatography. Ion chromatography is a
specialized area of liquid chromatography which will separate and quantify
the individual cations or anions. This technique has been applied to the
measurement of sulfur dioxide in ambient air, and adaptation to dilute auto-
motive exhaust should not be difficult.
Other instruments are commercially available that are reported to mea-
sure sulfur dioxide in ambient or dilute automotive exhaust. Such instru-
ments include continuous detection by pulsed fluorescent UV and second
derivative UV analyzers. A pulsed fluorescent UV analyzer for sulfur di-
oxide was found to give recoveries on the order of 115-125 percent, indi-
cating that a positive interference is present (72). The second derivative
40
-------
UV sulfur dioxide analyzer has inherent problems when being used on contin-
uous samples. The mirrors are located in the actual cell and become coated
with various exhaust components even after filtration of the sample. The
mirrors will become etched and need resurfacing if the unit is used in the
presence of sulfur dioxide, sulfate ion, or other corrosive exhaust com-
ponents. The inherent noise level, along with the consistent mirror problem,
.preclude the use of a second derivative UV analyzer for measuring sulfur di-
oxide on a continuous basis.
A variation of the GC method for measuring sulfur dioxide is the use of
a continuous analyzer using an FPD detector. Although this approach is good
in theory, it has several problems associated with the performance of the
FPD detector. These units were originally designed to monitor sulfur dioxide
levels in the ambient air and adaptation to automotive exhaust was not
straightforward. Air samples had essentially the same oxygen and nitrogen
levels all of the time; however, dilute exhaust samples have variable car-
bon dioxide, oxygen arid nitrogen concentrations. The species have been
found to cause quenching effects on a FPD detector. With the constantly
changing carbon dioxide, oxygen, and nitrogen it would be impossible to
correct for any quenching effect. The use of a continuous FPD analyzer for
measuring sulfur dioxide in automotive exhaust would not be acceptable un-
less the quenching effects could be eliminated.
PROCEDURAL DEVELOPMENT
From the results of the literature search it was determined that the
analysis of sulfur dioxide should be conducted by the use of ion chromato-
graphy. An ion chromatograph built at Southwest Research Institute was
dedicated for this purpose. This instrument utilizes a modified Swagelok
reducing union for a conductivity cell, a Hall conductivity detector, a
Milton Roy mini-pump, a Soltec multivoltage recorder, a Glenco Scientific
pulse dampener, and polyethylene cubitainers from Cole Farmer Instrument
Company for the analysis of sulfur dioxide. A minimal amount of procedural
development work was necessary for this procedure; however, several instru-
ment and sampling parameters did have to be determined. The selection of
these parameters are discussed in detail in the Validation Experiments
section.
In order to analyze automotile exhaust for sulfur dioxide, a trap or
an absorbing reagent must be used to concentrate the sulfur dioxide. A
method which has been previously used at Southwest Research Institute for
collecting and concentrating sulfur dioxide was selected and validated for
use in this project. This method consists of bubbling dilute exhaust through
a dilute aqueous hydrogen peroxide solution. The hydrogen peroxide reacts
with the sulfur dioxide to give sulfate ion which remains in the absorbing
solution.
The parameters selected for the analysis of sulfur dioxide are listed
below. The sulfur dioxide from the exhaust is bubbled through two impingers
(maintained at ice bath temperatures) in series with each impinger containing
25 mil of a 3 percent hydrogen peroxide solution. The exhaust flows through
the impingers at a rate of 4£/min. The two impingers together trap 99
41
-------
percent of the sulfur dioxide present in exhaust. A 0.5 micron Fluoropore
filter is installed in the sampling line prior to the bubblers to remove
particulate which could contaminate the separator column during analysis. A
portion of the absorbing solution is loaded into the sample loop and injected
into the ion chromatograph. For analysis the ion chromatograph utilized
three columns and an eluent composed of 0.003M NaHCO3 plus 0.0024M Na2CC>3.
The eluent flows at 30 percent of full pump capacity through a 3 x 150 mm
precolumn (this column helps prevent contamination of the separator column),
a 3 x 500 mm separator column and a 6 x 250 mm suppressor column packed with
AG-l-xlO anion suppressor resin (this neutralized the ionic effect of the
eluent while increasing that of the sample ion). A finalized copy of the
procedure is included as an appendix to the interim report.
VALIDATION EXPERIMENTS
Sulfur dioxide validation experiments were performed to verify the
sampling and instrument parameters. These experiments involved the deter-
mination of sampling flowrate, sampling temperature, kind and concentration
of absorbant, and number of bubblers required to collect 100 percent of the
sulfur dioxide. The variables associated with the ion chromatograph that
were determined included type and concentration of eluent, injection loop
size, flowrate, injection variability, and linearity of response. In
addition to determining sampling and instrument parameters, validation ex-
periments were performed to verify certain portions of the ion chromato-
graphic procedure for sulfur dioxide analysis. Tests for interferences,
sample stability, and standard stability were among those conducted. Also,
the method of washing glassware and the effect of combining the contents of
the two bubblers for analysis were studied.
A number of possible interferences were tested by bubbling the suspected
gas at 4 £/min through three bubblers in series. Each bubbler contained 25
mJl of 3 percent hydrogen peroxide and was maintained at ice bath temperatures
(0-5°C). The tests lasted approximately twenty minutes each. The results
are shown in Table 9. In the zero air, zero nitrogen, 3 percent CC>2 and
100 ppmc HC tests, no detectable amount (less than 0.01 ppm SO2) was found.
A positive interference of 0.01 ppm SO2 was produced in the 100 ppm NOX
test. The greatest interference, 0.02 ppm SO2, was found in the 100 ppm CO
test. Another source of interference was caused by the sulfuric acid-
chromic acid bath in which the bubblers were washed. The sulfate ion from
the sulfuric acid could not be sufficiently rinsed from the bubblers, even
with repeated deionized water rinses. For this reason, a 1:1 (v:v) nitric
acid and water solution was used to wash the bubblers used in the sulfur
dioxide procedure. The sulfate present in the hydrogen peroxide absorbant
caused a positive interference that could be corrected by subtracting the
sulfate value of a background sample and from the sulfate value of an exhaust
sample.
Another test which was conducted to verify the SO2 procedure was the
combining of the contents of both bubblers for analysis on the ion chroma-
tograph. Table 10 shows the results of this experiment and compares these
values to those obtained by separate analysis of the contents of each bub-
bler. The S02 concentrations calculated at the 1 x 30 attenuation setting
42
-------
Suspected
Interference
Zero air
Zero N2
3% CO2-run 1
3% C02-run 2
100 ppmc HC
100.Ppm NOX
100 ppm CO
TABLE 9. INTERFERENCES TO SO2 ANALYSIS
Concentration SO2(ppm SO2)
1
0.00
0.00
0.00
0.00
0.00
0.01
0.01
Bubbler
2
0.00
0.00
0.00
0.00
0.00
0.00
0.01
3
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Total SO? (ppm)
0.00
0.00
0.00
0.00
0.00
0.01
0.02
TABLE 10.. ANALYSIS OF THE CONTENTS OF BUBBLERS
1 AND 2 SEPARATELY AND COMBINED
Separate Combined
Sample Attenuation Analysis (ppmSO?) Analysis (ppmSO2) %^ Diffexence
1
2
3
1
1
1
X
X
X
30
30
30
0
0
0
.60
.55
.52
0
0
0
.94
.64
.51
57
16
2
.4
.1
.1
AVERAGE 0.56 0.70 25.0
4
5
6
100
100
100
2.00
2.03
2.02
2.40
2.44
2.42
19.9
19.8
20.1
AVERAGE 2.02 2.43 20.3
43
-------
were from an SO2 baseline produced by a noncatalyst 1972 Plymouth Fury
equipped with a 360 V8 engine. The concentrations at the 1 x 100 attenuation
setting were obtained from sulfur dioxide recovery experiments using the same
test car. The combined baseline values in Table 10 are between 2 percent and
57 percent higher than the separately analyzed samples, with the average
being 25 percent. The recovery values for the combined bubblers in the re-
covery tests also ran about 20 percent higher than the separately analyzed
bubbler samples. Because of the dilution problems encountered when combining
the contents of both bubblers, each bubbler is analyzed separately. The last
validation experiment for the ion chromatographic method of sulfur dioxide
analysis involved determining the stability of samples and standards over a
period of time. The sulfuric acid standards made up in filtered deionized
water remained stable for at least fourteen weeks. Sulfate standards made
up from stock solutions prepared on 11/30/77 and 3/13/78 were analyzed on
3/13/78 and the peak areas were compared. The results, shown in Table 11,
indicate that the fourteen week old standards repeated within 10 percent of
the freshly prepared standards. A study was also conducted with a variety
TABLE 1L SULFATE STANDARD STABILITY
Standard
Concentration
,
M,
0.96
0.96
4.80
4.80
9.60
9.60
38.40
38.40
96.00
96.00
Standard
Preparation
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
(13/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
Height %
Attenuation (in) Amplitude Difference
3
3
10
10
30
30
100
100
300
300
4.73
4.77
7.72
7.62
5.41
5.45
7.93
7.81
6.09
6.08
125,844
125,080
148,259
147,300
130,852
131,036
149,864
148,668
135,666
135,638
0.6
0.6
0.1
0.8
0.0
of samples of different ages to determine sulfate longevity in the 3 percent
hydrogen peroxide absorbant. One week after collection and initial analysis
a sample obtained from an SO2 exhaust recovery experiment remained at 0.13
ppm S02. A two week old 0.06 ppm S02 baseline sample produced similar re-
sults. There was no change observed in the sulfur dioxide level. A tenth
44
-------
week collection efficienty sample, however, did vary from its initial con-
centration of 0.13 ppm by decreasing 7.9 percent to 0.12 ppm SC>2. This is
greater than the injection repeatability of 1.2 percent, however, within the
minimum detection limit of 0.01 ppm SO2- The samples appear to be stable
at least two weeks but less than ten weeks, the break-off point probably
lying between four and six weeks.
The second portion of validation testing included the determination of
SC>2 sampling parameters. Nominal,concentrations of 5 and 12 ppm SC>2 were
collected for 20 minutes in three bubblers, each containing 25 m£ of 3 per-
cent hydrogen peroxide. The results of these tests are shown in Table 12.
TABLE 12. S02 COLLECTION EFFICIENCY AS A
FUNCTION OF FLOWRATE AND TEMPERATURE
% SO-
Test
Flow-
rate Temp ,
(Vmin) (°F)
SO2 Concentration (ppm) and
(%) in bubbler
1
2
3
Bubbler
1+2+3
(ppm)
trapped in
bubblers
1 and 2
Nominal 5 ppm SO2
1
2
3
4
4
4
2
2
72
32
32
75
4.31(95
6.73(96
6.48(94
6.71(98
-3)
.7)
.6)
.8)
0.16(3.4)
0.17(2.4)
0.21(3.1)
0.12(1.8)
Nominal 12 ppm
5
6
7
8
9
10
11
12
4
4
4
4
4
4
4
4
32
32
32
32
32
32
32
32
22.4(92
24.8(99
28.5(99
29.3(97
30.0(98
21.2(97
29.9(98
28.9(98
.5)
.1)
.0)
.8)
.5)
.9)
.5)
.5)
1.25(5.2)
0.13(0.5)
0.22(0.8)
0.35(1.2)
0.24(0.8)
0.20(0.9)
0.23(0.8)
0.32(1.1)
0
0
0
0
S02
0
0
0
0
0
0
0
0
.06(1
.06(0
.16(2
.09(1
.54(2
.09(0
.07(0
.30(4
.21(0
.25(1
.23(0
.12(0
-3)
.1)
.3)
-4)
.2)
.4)
.2)
.0)
.7)
.2)
.8)
.4)
4
6
6
6
24
25
28
29
30
21
30
29
.52
.96
.85
.93
.2
.0
.8
.9
.4
.7
.4
.3
98.7
99.1
97.7
98.6
97.7
99.6
99.8
99.0
99.3
98.8
99.3
99.6
The largest quantity of sulfur dioxide was retained at a flowrate of 4 &/min
at ice bath temperature. 99.1 percent SO2 was collected in the first two
bubblers under those conditions. It is desirable to prevent small particu-
late debris in the exhaust from entering the samples and, thus, the columns
of the ion chromatograph. If this form of contamination is allowed to col-
lect in the columns the liquid flow becomes hampered causing increased back-
pressure. To avoid this problem each day a fresh 0.5 micron Fluoropore
filter is inserted in the sampling line prior to the bubblers. Tests were
45
-------
conducted with and without filters in the sampling line to see if any sulfur
dioxide was retained on the filter. An unfiltered baseline sample produced
by the Plymouth Fury test car was compared with two filtered baseline samples
from the same car. The filtered samples, averaging 0.44 ppm SO2 ran 6.8
percent higher than the unfiltered sample (0.41 ppm), a difference of 0.03
ppm SO2. Sulfur dioxide recovery from the exhaust of the same test car ran
1.5 percent lower for the filtered sample (1.35 ppm SO2) compared to the un-
filtered sample (1.37 ppm SC>2). The difference for these exhaust recovery
samples was 0.02 ppm SO2-
The third portion of validation testing involved the determination of
instrument parameters; eluent concentration and flowrate, columns, sample
loop size, linearity of response and injection variability. Choice of
eluent concentration and flowrate will depend on the columns chosen and the
species present in the samples. Exhaust samples contain a variety of amines,
fluoride, chloride, nitrite, phosphate, nitrate and sulfate. Nitrate elutes
just prior to sulfate necessitating the use of an efficient separator column.
A 3 x 500 ram glass column packed with patented resin is the separator column
chosen for sulfate analysis. The suppressor column is a 6 x 250 mm glass
column with AG 50W-X16 resin. A 0.003M NaHC03 and 0.0024M Na2CO2 eluent
solution flowing at 30 percent of pump capacity gives good baseline resolu-
tion when a 500 y& sample loop is used. Another instrument factor which
needed to be determined was the injection variability.
The ion chromatograph has an injection repeatability of 1.1 or 1.2
percent as shown in Table 13. This is represented by Cv (coefficient of
variation) which is the standard deviation divided by the mean and multi-
plied by 100. The mean or average is represented by x and standard devi-
ation of Sx. For these calculations peak heights instead of peak areas were
used since the heights repeated much better and with greater precision than
the areas.
Two different standards were analyzed: 0.5 ^ SO4"2 and 4 Q^gSO4~j.
m£ " m£
The final instrument parameter determined was the linearity of response
of sulfate standards at different attenuations. The sulfate standards, made
up from sulfuric acid and filtered deionized water, maintained linearity at
each attenuation but the slopes became steeper as the sensitivity decreased.
Table 14 shows heights corresponding to each standard used and Figure 16
shows the graphical representation of the data. At the 1 x 10 scale setting,
the relative slope was 1.7, at the 1 x 30 setting it was 2.5 and at the 1 x
100 setting the slope was 2.6. it was not necessary to cary the curve any
further, since no samples have been obtained that fall in the higher con-
centration range. However, it was found that linearity was maintained at
concentrations from 40 to 100 V^SO^-2^
~mZ
QUALIFICATION EXPERIMENTS
Qualification tests were conducted on the same test car from which
validation data was obtained. This test car was a noncatalyst 1972 Plymouth
Fury equipped with a 360 V8 engine. A baseline was run on the car to
46
-------
TABLE 13. INJECTION REPEATABILITY FOR ION CHROMATOGRAPH
Sample Concentration (
Attenuation Height (in)
nU6
I 3.84
2 3.84
3 3.84
4 3.84
5 3.84
6 3.84
7 0.48
8 0.48
9 0.48
10 0.48
11 0.48
12 0.48
1 x 10 5.37
1 x 10 5.43
1 x 10 5.51
1 x 10 5.45
1 x 10 5.50
1 x 10 5.53
x 5 . 46 in
Sx 0.06 in
Cv 1.1%
1 x 10 0.81
1 x 10 0.80
1 x 10 0.79
1 x 10 0.81
1 x 10 0.80
1 x 10 0.81
x 0.80 in
S,, 0.01 in
Cv 1.2%
47
-------
TABLE 14. CALIBRATION CURVE FOR SULFUR DIOXIDE
Standard + Heights Corrected
Concentration (HH i) Attenuation Height (in) to 1 x 10 scale (in)
0.48
0.96
1.48
1.92
2.88
3.84
4.80
4.80
7.68
9.60
19.20
28.80
38.40
1 x 10
1 x 10
1 x 10
1 x 10
1 x 10
1 x 10
1 x 10
1 x 30
1 x 30
1 x 30
1 x 100
1 x 100
1 x 100
0.85
1.87
2.38
3.23
4.89
6.62
8.16
2.60
4.77
6.49
3.37
6.15
8.32
7.80
14.31
19.47
33.70
61.50
83.20
48
-------
90
30
70
60
•H
2 calibration curve.
49
-------
determine S02 emissions of the car by itself. First, a concentration of
500 ppm S02 was injected into the constant volume sampler (CVS) without
exhaust for 31 minutes and then into the CVS with the exhaust of the test
car during a hot FTP driving cycle. From these injections, the recovery of
SO, was determined from each test. The baseline data in Table 15 shows that
an average of 0.49 ppm SO2 is produced from the car when the hot FTP driving
cycle is run.
TABLE 15. SULFUR DIOXIDE BASELINE FROM PLYMOUTH FURY TEST CAR
S02 Cone.
Test
Date (ppm SO2)
1 2/27/78 0.66
2 3/06/78 0.45
3 3/06/78 0.42
4 3/06/78 0.43
Average = 0.49 ppm SO2
The SO2 recovery tests show that 101 percent of the sulfur dioxide injected
into the CVS alone survives the trip to the collection point (Table 16) and
99.7 percent of the SO2 injected into the exhaust of the test car (Table 17)
is recovered. These values essentially represent 100 percent recovery in
both instances.
RESULTS AND DISCUSSION
The ion chromatographic method of sulfur dioxide analysis is a simple,
sensitive and relatively rapid procedure with a minimal number of inter-
ferences. Zero air, nitrogen, 3 percent CO2 and 100 ppmc HC did not inter-
fere within the minimum detection limit of 0.01 ppm S02. However, 100 ppm
NOX and 100 ppm CO produced positive interferences of 0,01 and 0.02 ppm S02,
respectively. The sulfuric acid-chromic acid bath which had been previously
used to wash the bubblers also provided positive interference for samples
collected in bubblers washed in this bath. The problem was averted by re-
placing the sulfuric acid-chromic acid with 1:1 (v:v) nitric acid. The
manufacturer has also stated that persulfite will interfere with sulfate
analysis and that oxylate ion will interfere if the separator column capa-
city is reduced. No problem has been noted with these two species. The
decision to analyze the samples in each of the two bubblers separately
rather than in combination was due to the results of a study comparing the
two methods. It was discovered that a 20-25 percent increase in sulfate
concentration occurs when the samples from a set of bubblers are combined.
This is apparently a dilution problem that can. be avoided by analyzing
each sample separately.
The effect of age on sulfate standards and samples was investigated
50
-------
TABLE 16. SULFUR DIOXIDE RECOVERY FROM DIRECT CVS INJECTION
SOj Recovered
Nominal S02 Injected Total
Flowrate Vol. S02 Cone. S02 Diluted Sample
(ftVmin) Injected Injected Volume Cone.
Test SO, CVS (ft3)S> (ppra) (ft3)f (ppm)
1 0.89 298 26.612 472 9239 1.37
2 0.89 298 26.612 472 9239 1.38
3 0.89 298 26.612 472 9239 1.36
aThis concentration, representing 100% recovery, was
obtained by back calculating from recovered sample
concentration and percent recovery.
"Volume corrected to 1 atin pressure and 63°F.
TABLE 17. SULFUR DIOXIDE RECOVERY FROM
DIRECT CVS INJECTION DURING THE HOT FTP
S02 Recovered
Nominal SO2 Injected Total
Flowrate Vol. SO2 Cone. SO2 Diluted Sample
(ft -^min) Injected Injected Volume Cone.
Test S02 CVS {ft3)b (ppm) (ftj)b (ppm) C
1 0.87 299 19.773 472 6886 1.35
2 0.87 299 19.773 472 6886 1.35
3 0.90 299 20.226 472 6869 1.34
4 0.90 299 20.226 472 6869 1.42
5 0.90 299 20.226 472 6869 1.41
•
Calc. Amt.
of SO2
Recovered Percent
(ppm)a Recovery
1.36 101
1.36 102
1.36 100
Average = 101%
DILUTE EXHAUST BY
DRIVING CYCLE
Calc. Amt.
of S02
Recovered Percent
(ppm) a Recovery
1.36 99.6
1.36 99.6
1.39 96.4
1.39 102
1.40 101
Average = 99.7%
aThis concentration, representing 100% recovery, was
obtained by back calculating from recovered sample
concentration and percent recovery.
"Volume corrected to 1 atm pressure and 68°F.
clncludes baseline correction.
51
-------
and it was found that sulfuric acid standards remained stable for at least
fourteen weeks and the exhaust samples for at least two weeks but less than
ten weeks. The actual lifetime probably lies between four and six weeks.
This relatively long period of sample stability allows for some leeway in
case the samples can not be analyzed immediately. The best collection ef-
ficiency was obtained when the dilute exhaust flowed at 4 H/iain through two
bubblers in series, each containing 25 m£ of 3 percent hydrogen peroxide
maintained at ice bath temperature (0-5°C). Fluoropore filters (0.5 micron)
are inserted in the sampling line each day to prevent contamination of the
samples and subsequent column poisoning in the ion chromatograph. A pos-
sible 1-4 percent loss in sulfur dioxide recovery may be expected as a
result of the filter. The loss might also be due to the filter trapping
sulfuric acid from the exhaust and, thus, excluding it from the sample. The
linearity of response of the ion chromatograph is maintained in the sulfate
concentration range 0.5 to 100 ug SO4~2 (100 ppm) . However, changing the
attenuation on the ion chromatograpS causes a discontinuity in the calibra-
tion curve. This discontinuity is seen as a slope change in Table 14. The
standards analyzed at each attenuation obviously fall into a linear pattern
even though the slopes differ. A different set of standards must therefore
be run for each sensitivity setting.
The results of the qualification tests indicate that 100 percent SO? is
recovered from both the injection of sulfur dioxide into the CVS alone and
the injection into dilute exhaust. The baseline SC>2 emission from the test
car averaged at 0.49 ppm sulfur dioxide while the car was run on the hot FTP
driving cycle.
The ion chromatographic method of sulfur dioxide analysis is a simple,
sensitive, specific, and relatively rapid procedure with few interferences.
No intermediate steps are involved, lessening the change of sample loss or
contamination. The ion chromatograph is sensitive to 0.01 ppm SC>2 and
samples can be analyzed in 10 to 15 minutes. The difference in retention
times between the various ions in the sample allows for definite peak iden-
tification. Sulfate analysis on the ion chromatograph is also unaffected by
most interferences plaguing a number of other sulfur dioxide procedures.
52
-------
SECTION 5 .
NITROUS OXIDE PROCEDURE
LITERATURE SEARCH
There are six common oxides of nitrogen: nitrous oxide (N2O), nitric ox-
ide (NO), nitrogen dioxide (NO2), dinitrogen trioxide (N2O3), dinitrogen te-
traoxide (N204), and dinitrogen pentoxide (N-Og). In addition to these, there
are two different oxides that have the empirical formula N03. Both are very
reactive and have only been identified by spectroscopy as transient species.
Nitrous oxide is a colorless, nonflammable gas at room temperature with
a slightly sweet taste and odor. Some synonyms are dinitrogen oxide or nitro-
gen monoxide, hyponitrous acid anhydride, factitious air and laughing gas.
Nitrous oxide is the least reactive and noxious of the oxides of nitrogen. At
room temperature it is relatively inert; but at 500°C, it decomposes to nitro-
gen, oxygen, and nitric oxide and at elevated temperatures, it will support
combustion and oxidizes certain organic compounds and alkali metals. Nitrous
oxide, N2O, has a molecular weight of 44.01, a melting point of -90.8°C, and
a boiling point of -88.5°C. It is a linear molecule with a N-N bond distance
of 1.128 A and a N-O bond distance of 11.84 & and is isoelectrpnic with carbon
dioxide. When inhaled, nitrous oxide may cause hysteria, insensibility to
pain, or unconsciousness and therefore used as an anesthetic for minor oper-
ations, including dentistry. It is also used as a nontoxic dispersing agent
in commercial whipped cream. Commercially, nitrous oxide is prepared by the
thermal decomposition of ammonium nitrate, the controlled reduction of ni-
trites or nitrates, the slow decomposition of hyponitrites, and by the thermal
decomposition of hydroxylamine.
The analysis of nitrous oxide has been conducted using mass spectrometry,
infrared spectroscopy, and gas chromatography. Of these, the most sensitive
method is gas chromatography.
There are three gas chromatography methods that have been used to ana-
lyze for nitrous oxide. Two of these require cold traps to collect and con-
centrate the sample. With the third technique, grap samples are collected in
Tedlar plastic bags and analyzed with an electron capture detector.
PROCEDURAL DEVELOPMENT
The gas chromatograph operating conditions and sampling system specifi-
cations were obtained from the Project Officer. A two column system with
column backflush and isothermal termperature operation was constructed for the
analysis. The stripper column is a 2' x 1/8" stainless steel tube filled with
10 percent OV-17 on 80/100 mesh Gas Chrom Q. The analytical column consists
of a 6' x 18" stainless steel column packed with 120/150 mesh Porapak Q. A
53
-------
series of six port valves and timers is used to direct the sample flow through
the columns. The samples are then analyzed with an electron capture detector.
Since this method has been successfully applied to automotive exhaust, no other
significant effort was applied to the procedural development. A description
of the analytical system and the adapted procedure is presented as an appendix
to this report.
VALIDATION EXPERIMENTS
Several experiments were conducted to validate the system for detector
linearity, injection repeatability, and bag sample stability. Also, a means
of calibrating the system using permeation tubes and calibration gases was
investigated. The results are reported below.
Initially, a Tracer Model 412 Permeation System with an Ecocal permeation
assembly was used to calibrate the instrument. The concentration of nitrous
oxide could be set by changing the diluent gas flow over the permeation tube.
This means of calibration can be used between the dynamic concentration range
of 0.23 ppm to 6.31 ppm. However, a ghost peak was also generated with this
permeation system. Efforts to eliminate this extraneous peak were unsuccess-
ful. The permeation rate from this tube is not solely dependent on the con-
centration of nitrous oxide and would make this means of calibration difficult.
Because of the problems associated with the permeation system, a static method
of calibration was pursued. Four cylinders of calibration gas were obtained.
The nitrous oxide concentrations of the cylinders ranged from 1.31 ppm to
9.90 ppm. No ghost peaks were observed and quantitative results were obtained.
Detector linearity over a wide range of concentrations is helpful and
sometimes necessary when a variety of widely differing samples are to be ana-
lyzed. Such is the case with gaseous bag samples. No easy method to dilute
the bag concentrations into the linear range of the instrument is available.
Therefore, the detector must be linear in all of the concentration ranges ex-
pected. The detector linearity for the electron capture detector was deter-
mined with calibration gases from 1 to 10 ppm (Figure 17). Sample concen-
trations within this range are linear with respect to the detector.
With a gas sample loop and electrical/pneumatic sample flow control,
sample injections are not as subject to human error and are more reproducible
than syringe sample injections. The injection repeatability for the four
calibration gas standards is shown in Table 18. This table demonstrates that
sample injection reproducibility is reliable with the present analytical
system.
Due to the length of time required for sample collection and subsequent
gas chromatograph analysis, nitrous oxide must be stable for at least several
hours. The bag stability was determined by taking a random Tedlar bag filled
with exhaust from an emissions test. The two samples were reprocessed peri-
odically. The first bag was processed hourly for five hours and the second
bag was processed periodically for several weeks. The time sample decay curve
for both samples is shown in Figures 18 and 19. Samples may be stored for a
period of about two weeks without adverse effects to the sample concentration.
After this time, it is difficult to preserve sample integrity.
54
-------
10
ui
ui
I
04
c
o
•rl
•P
«J
M
-p
fl
(U
o
0
u
QJ
T!
•H
M
3
O
•H
a
6 -
4 -
6 8 10
Peak Area (counts x 10,000)
12
14
16
Figure 17. Detector linearity curve.
-------
TABLE 18. INJECTION REPEATABILITY OVER THE RANGE
OF DETECTOR LINEARITY
N2O Concentration, ppm
1.31
2.16
4.95
Area
Average
Standard Deviation
Coefficient of
Variation
9.90
17935
17811
18346
18149
18319
18112
235
1.30
28020
28759
28974
28931
28671
444
1.55
61146
61057
61004
61325
61448
61196
186
0.30
118085
118935
118687
119448
119810
119005
118995
599
0.50
56
-------
30
20
10
Ui
Cn
C
(d
u
-P
c
-------
10
(Jl
03
-------
QUALIFICATION EXPERIMENTS
Qualification recovery experiments were conducted for nitrous oxide with
two vehicles over several driving cycles and with several concentrations of
nitrous oxide injected into the system. The three driving cycles employed in
these experiments were the Federal Test Procedure (FTP), Sulfate Emissions
Test, Cycle 7 (SET-7), and the Highway Fuel Economy Test (HFET). The vehicles
tested were two Oldsmobile Cutlass Supremes with oxidation catalysts. The
first was a 1977 model with a 260 V8 engine and the second was a 1978 model
with a 305 V8 engine. Three concentrations of nitrous oxide were used to
generate dilute concentrations ranging from 0.20 ppm to 7.19 ppm.
The baseline emission rates for each car were determined for the three
driving cycles. These values are shown in Table 19. These baseline exhaust
emission values were used to correct the observed concentrations with exhaust.
These values were also corrected for the concentration of nitrous oxide in
the air. The injection recoveries for the two vehicles are presented in
Tables 20 and 21. The injection recoveries were also determined without the
vehicle. Table 22 lists the results from these tests.
The combined recoveries over the four concentration ranges is essentially
100 percent. With each vehicle, the average percent recovery for the three
different driving cycles is also 100 percent. Therefore, total sample recovery
can be expected for the analysis and complete sample integrity is preserved
in the CVS.
RESULTS AND CONCLUSIONS
The measurement of nitrous oxide in dilute exhaust can be conducted with
a gas chromatography technique. Dilute exhaust is collected in a Tedlar bag
as a grab sample. Sample analysis of the bag sample with an electron capture
detector and comparison to a set of calibration blends determines the concen-
tration in dilute exhaust. The minimum detectable limit of this procedure is
0.01 ppm.
Detector linearity, injection repeatability, bag sample stability, and
a static means of calibration were demonstrated for the system. The electron
capture detector employed is linear over the range of sample concentrations
expected. This enables direct sample analysis without secondary dilution.
The injection repeatability for an automated sampling system with a gas sam-
pling loop is excellent for the analytical procedure. The sample integrity
is also maintained if the samples cannot be analyzed immediately. This en-
ables minor system repairs without holding up testing.
The average CVS percent recoveries is essentially 100 percent in dilute
exhaust for nitrous oxide over the concentration ranges expected. No losses
were observed with or without the vehicle. Even under a variety of emissions
testing conditions (driving cycles and vehicle), the average percent recovery
is consistent. This is expected due to the inertness and stability of nitrous
oxide. Sample integrity can be expected throughout the entire testing pro-
cedure and sample concentrations are not subject to the instability of the
compound tested.
59
-------
TABLE 19. TYPICAL NITROUS OXIDE EXHAUST CONCENTRATIONS
AND EMISSION RATES ON SEVERAL DRIVING CYCLES
WITH TWO VEHICLES USED DURING QUALIFICATION EXPERIMENTS
1977
1977
Driving
Cycle
FTP
(cold)
FTP
(cold)
FTP
(hot)
FTP
(hot)
SET- 7
SET- 7
SET- 7
HFET
HFET
HFET
260 V8 Olds Cutlass
Run Bag
1 1
2
3
2 1
2
3
3 1
2
3
4 1
2
3
5 only
6 only
7 only
Avg
8 only
9 only
10 only
Avg
ppm N2O^a)
1.33
0.42
1.68
— __
1.24
0.68
2.36
1.09
0.46
1.41
1.26
1.26
1.40
1.31
0.77
0.68
0.47
0.64
ing/km^
29.2
38.9
26.6
24.6
24.6
27.5
25.6
10.8
9.5
6.7
9.0
305 V8 Olds Cutlass
ppm N2O^a)
0.52
0.27
0.77
0.58
0.25
Q.86
0.59
0.23
0.72
0.57
0.18
0.99
0.98
0.91
0.98
0.96
1.23
0.96
1.20
1.13
mg/km ^T
14.5
15.3
13.7
14.4
19.2
17.9
19.3
18.8
17.4
13.6
17.0
16.0
values are background corrected
FTP emission rates are weighted
60
-------
FET
SET-7
TABLE 20. NITROUS OXIDE INJECTION RECOVERIES WITH 1977
260 V8 OLDSMOBILE CUTLASS WITH OXIDATION CATALYST
Nitrous Oxide, ppm
Driving
Cycle
FET
FET
FET
SET-7
Calculated
Amount
Run Injected
1 7.04
2 7.22
3 , 7.41
4 8.11
Observed
(Exhaust + Air
+ Injected)
8.88
9.10
9.21
10.18
TABLE 21. NITROUS OXIDE INJECTION
305 V8 OLDSMOBILE CUTLASS WITH
Driving
Cycle
FTPcs
FTPs
FTPhs
FTPcs
FTPs
FTPhs
Calculated
Amount
Run Injected
1 8.73
1 8.80
1 8.89
2 8.35
2 8.59
2 7.88
Nitrous Oxide
Observed
(Exhaust + Air
+ Injected)
8.85
9.73
10.12
9.21
9.83
11.24
Exhaust
Only
0.64
0.64
0.64
1.31
RECOVERIE
OXIDATION
, ppm
Exhaust
Only
0.55
0.26
0.82
0.55
0.26
0.82
Background
0.37
0.37
0.37
0.36
S WITH 1978
CATALYST
Background
0.33
0.33
0.33
0.40
0.40
0.40
Percent
Recovery
111.8
112.1
110.6
105.0
-
Percent
Recovery
91.3
103.8
100.9
98.9
106.7
116.4
3
4
9.24
9.28
11.58
11.72
1.13
0.96
0.33
0.33
109.6
112.4
61
-------
TABLE 22. NITROUS OXIDE CVS QUALIFICATION EXPERIMENTS - NO VEHICLE
Injection Blend
Concentration,
ppm N?O
95
95
831
831
831
831
8075
8075
Nominal Flow
Rate, ft3/min
N?O Blend CVS
0.75
0.75
0.40
0.40
0.75
0.75
0.35
0.35
350
350
350
350
350
350
350
350
Calculated
ppm NoO
Dilute^)
0.23
0.22
0.80
0.77
1.97
2.03
2.00
6.64
6.27
Observed
Run ppm N2O
(a)
1
2
Avg
1
2
Avg
1
2
3
Avg
1
2
Avg
0.27
0.26
0.27
0.96
0.64
0.80
1.98
2.28
2.28
2.18
6.90
7.12
7.01
Percent
Recovery
118.7
115.9
117.3
119.5
83.7
101.6
103.9
110.0
107.0
This procedure provides a rapid and sensitive method for the analysis of
nitrous oxide in dilute exhaust. A single bag sample requires about two
minutes for sample loop purging and seven minutes for the automated analysis.
The total analysis time is about nine minutes per sample. The automated
system provides simplicity and ease of operation and makes this procedure
ideal for routine analysis.
62
-------
SECTION 6
INDIVIDUAL HYDROCARBON PROCEDURE
LITERATURE SEARCH
The eight individual hydrocarbons (methane, ethane, ethylene, acetylene,
propane, propylene, benzene, and toluene) have been measured by innumerable
techniques. One of the most efficient techniques for the individual determi-
nation of all of these compounds in a single analysis is with gas chromato-
graphy. Because of its efficiency, this means of analysis was selected over
any of the other available.
Hydrocarbons are of interest as exhaust components because of their po-
tential for photochemical smog formation. Hydrocarbons are palced into four
classes according to their participation in atmospheric reactions. Methane,
ethane, acetylene, propane, and benzene are placed in the Class I, the non-
reactive category. The Class II reactive category includes the C4 and higher
paraffins, while the Class III reactive category encompasses all of the aro-
matics except benzene. The olefins are placed in the Class IV reactive cate-
gory. When olefins such as ethylene react with ozone
03 + H2C + HC2 •*• H2C = 0 + HO + HCO
the precursors of photochemical smog are formed. These free radicals then
participate in other atmospheric reactions that result in oxidant formation.
Dimitriades and Seizinger(73) proposed a three-column system capable of
analyzing at least 22 hydrocarbons. Two packed columns were required to re-
solve the GI and Cj hydrocarbon components and an open tubular column was
used to resolve the other components. The complete analysis consisted of two
different sample loop sizes. This procedure was considered time consuming
and the number of compounds to be analyzed was excessive for the project.
Papa et al(74) presented a procedure for the analysis of C± and C12 ny~
drocarbons in automotive exhaust. This dual column system consisted of a
packed column with a mixture of stationary phases for the resolution of C±
and C2 hydrocarbons and an open tubular column. About 200 individual peaks
were obtained from the method. This procedure also required two sample loop
injections. Excessively low temperatures were required for resolution of C^
and C2 hydrocarbons with this analytical system.
Klosterman and Sigsby(75) proposed a simple analytical system for the
determination of hydrocarbons according to their potential for photochemical
smog formation. A flame ionization analyzer was used in their work, though
this technique did not employ the use of a gas chromatograph. A column
63
-------
similar to that used by Klosterman and Sigsby was used to scrub the oxygenated
hydrocarbons and olefins from benzene and toluene by Black et al(76). This
method utilizes four packed analytical columns for the resolution of the de-
sired compounds. Methane, ethane, ethylene, acetylene, propane, and propylene
are resolved with the first two columns; and benzene and toluene are resolved
with the other two. As with the other procedures, two sample loops are re-
quired for the combined analysis of paraffins, olefins, and^aromatic hydro-
carbons. This procedure was also designed as a simple and inexpensive method
for the determination of smog related compounds. Table 23 lists the compounds
of interest, along with chemical formulas, boiling and melting points, syn-
onyms, and molecular weights.
PROCEDURAL DEVELOPMENT
The gas chromatographic procedure for the determination of the individual
hydrocarbons in dilute exhaust is currently "in hand" and has been used on a
variety of projects. The four column system is capable of resolving the eight
individual hydrocarbons. Columns I and II in the system consist of an 8" x
18" stainless steel tube packed with 80/100 mesh Porapak Q and a 4' x 1/8"
teflon column packed with 35/60 mesh type 58 silica gel, respectively. Column
III consists of a 15' x 1/8" stainless steel tube packed with 15 percent, 1,
2, 3-tris (2-cyanoethoxy)propane on 60/80 mesh chromosorb PAW; and column IV
consists of a 2' x 1/8" stainless steel tube packed with 40 percent mercury
sulfate (hgS04) and 20 percent sulfuric acid (H2SO4) on chromosorb W. Columns
II, III, and IV are used isothermally and column I undergoes a temperature
program sequence. The primary purpose of Column I is to resolve methane from
air, while Column II resolves C2 and C3 hydrocarbons. Columns III and IV re-
solve benzene and toluene from the other aromatics, paraffins, olefins, acety-
lenes, and oxygenated hydrocarbons. Three timers, four solenoid valves, and
five si-port gas sampling valves are required to accomplish the complicated
sample flow through the columns. No procedural development was required for
this analysis. The actual analytical procedure is included as an attachment
to this report.
VALIDATION
Since this gas chromatographic procedure has been used with much success
on a variety of projects, procedural validation was limited to injection re-
peatability and bag sample stability experiments. These were determined dur-
ing routine analysis and demonstrate the stability of the instrumentation.
All other parameters were determined from previous experience with this ana-
lytical procedure.
The injection repeatability for the individual hydrocarbon procedure was
conducted on two separate occasions. Table 24 shows the data accumulated on
each occasion. The injection repeatability for the two 10 ml sample loops is
not greater than +_ 2 percent.
The bag sample stability experiment was conducted on a random" sample from
an emissions test. The sample was collected during the driving cycle and ana-
lyzed immediately afterward. This sample was then reprocessed periodically
for several weeks. A bag sample of the calibration standard and a bag sample
64
-------
TABLE 23. IMPORTANT FACTS ON INDIVIDUAL HYDROCARBONS
171
Compound Formula
Methane CH4
Ethylene C2&4
Ethane C0HC
z b
Acetylene £2^-2
Propane C3H8
Propylene ^3*%
Benzene C&H6
Toluene C^Hg
Molecular
Weight
16.04
28.05
30.07
26.04
41.11
42.08
78.12
92.15
Melting
Point
-182.48
-169.15
-183.3
- 80.8
-189.69
-185.25
5.5
- 95
Boiling
Point
-164
-103.71
- 88.63
- 75
- 42.07
- 47.4
80.1
110.6
Synonyms
marsh gas, methyl hydride
ethene, elayl, olefiant gas
bimethyl, dimethyl, methylethane ,
ethyl hydride
ethyne , ethine
dimethylmethane, propyl hydride
propene, methylethylene , methyl-
ethene
benzol, phene, cyclohexatriene
methylbenzene , phenylmethane ,
toluol, methacide
-------
TABLE 24. INJECTION REPEATABILITY ON
TWO SEPARATE OCCASIONS
]
2
3
4
5
6
7
8
9
10
Methane
72H3
7527
7792
7566
7585
7387
7700
7455
7080
7611
Ethylene
10097
10467
10621
10479
10495
10300
10575
10307
10493
10481
Ethane
9597
9873
10039
9892
9900
9690
9994
9724
9878
9085
Acetylene
12034
12544
12803
12628
12675
12398
12753
12460
12624
12714
Propane
14809
15092
153r,2
14825
15016
14925
15147
14702
14933
14996
Propylene
15569
15724
16152
15910
15899
15441
15931
15629
15842
15857
Benzene
13697
13815
13877
13977
13824
14041
14042
14066
14049
13943
Toluene
15348
15570
15466
15478
15576
15780
15799
15459
15783
Average
Standard
Deviation,.
sx
Coefficient
of variation,
7559.10
152.98
2.02
10431.50 9847.20
154.55
1.48
136.90
1.39
12563.30
224.85
1.79
14979.70
187.42
1.25
15795.40
206.83
1.31
13933.10
125.33
0.90
15584.33
166.20
1.07
1
2
3
4
5
6
Average
Standard
Deviation,
Coefficient
of variation.
C.
R593
86"
8f.02
8640
8702
8561
8622
48.59
0.56
11893
11937
11993
11885
12022
11900
11S38
57.20
0.48
10973
11011
11067
10977
11113
10946
11015
63.70
0.58
14803
14992
15139
15170
15371
15252
15121
200.07
1.32
15705
15912
15722
15775
15825
15879
15803
83.77
0.53
17867
18020
17818
17926
17850
17955
17906
75.10
0.42
16466
16604
16553
16409
16267
16433
16455
118.21
0.72
18546
18537
18735
18479
18530
18562
18565
87.95
0.47
66
-------
of exhaust doped with the calibration standard were also processed period-
ically- The time-sample decay curve for each compound is shown in Figures 20,
21, and 22. The sample integrity can be preserved for approximately five days.
QUALIFICATION EXPERIMENTS
The analysis of individual hydrocarbons in dilute exhaust has previously
been conducted on many projects. On the request of the Project Officer, no
qualification experiments were conducted with the CVS for this procedure.
Also, long term experience with this procedure has given an insight into the
sample integrity for the complete analytical system.
RESULTS AND CONCLUSIONS
The measurement of individual hydrocarbons in dilute exhaust has been
conducted with a gas chromatography technique. Tedlar film bags are filled
with dilute exhaust during each driving cycle. Analysis of the bag sample re-
quires a complicated system of four analytical columns with backflush and tem-
perature program capability. Sample concentrations are determined by compari-
son to a calibration blend of all eight hydrocarbons. The minimum detectable
limit is 0.1 ppmC to 0.2 ppmC. The higher molecular weight compounds approach
the higher minimum detectable limit.
Injection repeatability and bag sample stability were demonstrated for
the system. The largest injection variability was with methane and acetylene
and the smallest was with benzene and toluene. A 2 percent variability can
therefore be expected for the six components of the first sample loop and
first two columns (Ci - C3> , and a 1 percent variability can be expected for
the second sample loop and second two columns (benzene and toluene). This
agreement is much better than can be expected of a syringe sample injection.
The bag sample stability shows that dilute exhaust samples will be stable for
about five days. Only propylene and toluene were shown to have a large de-
crease in concentration over a period of time in exhaust. With the standard
only sample, all compounds showed the same decrease in concentration. A leak
in the bag is suspected as a cause of this drastic change in concentration.
However, even with a leak in the bag, the sample concentration is stable for
about five days. Samples must be analyzed before this time to maintain con-
fidence in the sample concentrations obtained. Otherwise, the sample inte-
grity is lost due to sample decay, bag leakage, and/or permeation through the
walls of the bag.
This procedure provides an effective means for the analysis of individual
hydrocarbons in dilute exhaust. A single bag requires about four minutes to
purge the sample loops, 23 minutes for total peak elution, and five minutes to
cool the oven temperature back to room temperature and reset the instrumen-
tation. The total analysis time per sample is about 32 minutes. The auto-
mated system provides a simplified operation for an otherwise complicated pro-
cedure and enables routine analysis for a large quantity of samples.
67
-------
20
0
HI
o<
c
a
u
u -20
c
"j
0
^
-40
-60
Nominal Range ~jf
Injection Repeatabilj
A -/.
* «• * i J 111*
Legend
• Methane
_ • Ethylene
A Ethane
T 1 1 1 1 1 1
10
20 30
Time , Days
40
50
60
20
0
0)
?
!3
I -20
a;
o
'M
-------
50
o> 0
a
1
w
4J
C
0)
o
-
Nominal Range of
Injection Repeatability
t 1
-*O V .a *-
* *
1
Legend
9 Methane
• £ thy lane
*
• Acetylene
•
-50 _
-100
10
15 20
Time, Days
30
35
Laaend
SO
-------
20
2
O -20
-40
-60
20 r-
-20
-40
-60
nominal range of
injection repeatability
• Methane
• Ethylene
A Ethane
« Acetylene
10
20
30
Time, Days
40
• Propane
• Propylene
A Benzene
• Toluene
»
a
Time, Days
50
60
t
nominal range of
injection repeatability
T^
0
1
10
1
20
1
30
1
40
1
50
,
60
Figure 22. Time-sample decay curve (exhaust and standard)
70
-------
SECTION 7
HYDROGEN SULFIDE PROCEDURE
LITERATURE SEARCH
Hydrogen sulfide is a very flammable and toxic gas at room temperature
and has the characteristic odor of rotten eggs. The chemical formula for
hydrogen sulfide (hydrosulfuric acid or sulfureted hydrogen) is ^S. Hydro-
gen sulfide is a bent molecule with a H-S-H bond angle of 93.3° and an S-H
bond distance of 1.3455 A. Hydrogen sulfide has a boiling point of -60.33°C,
a melting point of -85.49°C, and a molecular weight of 34.08. It is a very
weak diprotic acid with dissociation constants:
H2S + H2O •-•= H3O+ + HS~ KI = 5.7 x 10~8
HS- + H20 ;_zz H30+ + S= KZ = 1.2 x 10~15
Hydrogen sulfide may be detected by its odor at about 1 ppb; however,
olfactory fatigue soon results and higher concentrations may not have an
unduly objectionable odor. Death is caused by systemic poisoning and res-
piratory paralysis from exposure to high concentrations.
Hydrogen sulfide is prepared commercially as a by-product from many
chemical processes and by the treatment of metallic sulfides with a mineral
acid such as hydrochloric or sulfuric acid. Hydrogen sulfide produced in
exhaust is probably formed by the reduction of sulfur compounds in the fuel.
With an excess of oxygen, it burns to form sulfur dioxide and water:
2H2S + 302 —>• 2H20 + 2S02
and with insufficient oxygen to form free sulfur and water:
2H2S + 02 —>> 2H20 + 2S
Hydrogen sulfide also reacts with sulfur dioxide to form free sulfur and
water:
2H2S + SO2 —>• 2H O + 3S
This reaction may be significant if high levels of sulfur dioxide are pro-
duced in exhaust.
The analysis of hydrogen sulfide has been conducted with an entire
71
-------
spectrum of analytical methods. Some of these methods include: surface
reactions on plates, tiles, tapes or filters, wet chemical, fluorimetry,
infrared spectroscopy, sulfur ion selective electrode, coulometry, gas chro-
matography, and colorimetry. Most of these are not applicable to dilute
exhaust sampling but are applicable for ambient air sampling and "on line"
systems. The best applicable means of analysis for dilute exhaust is the
colorimetric technique.
There are two colorimetric methods available for the analysis of hydro-
gen sulfide. These are the sodium nitroprusside method and the methylene
blue method. The sodium nitroprusside method has a lower detection limit of
about 1 ppm. This method was not considered sensitive enough for the concen-
trations expected in dilute exhaust. The methylene blue method, on the other
hand, has a reported lower detection limit of 1-3 ppb.
The absorbing reagent is the key to successful analysis with this pro-
cedure. Hydrogen sulfide is precipitated as the sulfide in the presence of
metal ions. Cadmium and zinc hydroxide, cadmium sulfate, and zinc acetate
have been used as the absorbing media. However, several authors have re-
ported the oxidation of cadimium and the photochemical decomposition of
cadmium sulfide. Bamesberger and Adams (77) suggested the use of 1 percent
STRactan 10 as a stabilizer for cadmium absorbing solutions. On the other
hand, zinc solutions do not appear to have these inherent problems. Flairan
and James (78) tested all of the above absorbing reagents and found zinc
acetate to be the most efficient absorbent. For these reasons, zinc acetate
was selected as the best absorbing reagent,
PROCEDURAL DEVELOPMENT
A procedure for the analysis of hydrogen sulfide by the methylene blue
method was obtained from the Project Officer under EPA Contract 68-03-2499.
This procedure is a modification of the technique used by Gustafsson (79).
A buffered zinc acetate solution is used as the absorbing reagent. This pro-
cedure was compared to the one recommended for ambient air sampling by Adams
et al (80) which used cadmium hydroxide as the absorbing reagent. The
selected analytical procedure is included as an attachment to the interim
report.
The cadmium hydroxide method presented several problems. First, sulfides
in alkaline solutions are easily oxidized by air. Second, cadmium sulfide is
photosensitive and solutions must be protected at all times from exposure to
light. The use of special glassware or aluminum foil wrappings are necessary
to prevent exposure to light. The addition of a stabilizer such as STRactan
10 helps to minimize the effect of photochemical decomposition, but special
handling precautions are still necessary. Because of these reasons, cadmium
solutions are hard to work with, and some cadmium compounds are toxic. Cad-
mium, cadmium oxide, cadmium sulfate, and cadmium sulfide were included on a
tentative carcinogen list issued by OSHA in July, 1978. Zinc sulfide, on the
other hand, is not photosensitive, the solutions are much easier to work
with, and zinc compounds are not as toxic. The problems with the cadmium
hydroxide method were demonstrated by comparing the two methods.
72
-------
Several authors have reported that methylene blue may be bleached by
exposure to light. In order to determine what effect this might have on a
developed sample, two high (18-19 yg/100 m£) and two low (2-3 yg/100 mi)
concentration standards were prepared and developed for fifteen minutes.
One of each concentration was then exposed to light. After developing for
15 minutes in the dark, the other two were wrapped with aluminum and stored
in the dark. The absorbance of each was determined periodically for several
weeks. The time-light exposure decay curves are shown in Figures 23 and 24.
The results are discussed in the Results and Discussion Section.
Hydrogen sulfide is readily volatilized from acidic aqueous solutions.
In alkaline solutions sulfide ion may be oxidized by dissolved oxygen. The
pH of the buffered zinc acetate absorbing reagent is 7.0. This reagent re-
mains at a pH of 7 even after bubbling with dilute exhaust. Oxidation of
dissolved sulfide ion does not occur rapidly at this pH. After addition of
the amine solution and ferric ion to the absorbing reagent, the pH is below
2.0. At this pH, the trapped sulfide ion reacts to form methylene blue.
Buffering of the absorbing reagent and subsequent change of pH in the pre-
sence of the amine solution and ferric ion minimizes the losses due to oxi-
dation or volatilization.
There are two possible methods available for generating a Beer's Law
plot for calibration. The first technique requires extensive reagent pre-
paration and tedious titrations, Thiosulfate solution is standardized
against potassium dichromate. This standardized thiosulfate solution is used
to standardize a dilute iodine solution. The standard sulfide solution con-
centration is then determined with an iodometric method. Aliquots of the
standardized sulfide solution are used to generate a Beer's Law curve.
The other method requires the use of a hydrogen sulfide permeation tube.
The calibration curve is generated by bubbling a known concentration of
hydrogen sulfide for varying lengths of time. Generation of a calibration
curve in this manner takes into account the collection efficiency of the
bubblers. This method is quick, efficient and more consistent with the way
the samples are actually taken. It also enables the generation of a daily
calibration curve without being manpower intensive.
The calibration curve for methylene blue is shown in Figure 25. This
curve was determined for a standardized sulfide ion solution on two separate
occasions and follows Beer's Law at low concentrations. After about
70 ygS=/100 m£, the curve begins to deviate from Beer's Law- Concentrations
of hydrogen sulfide in dilute exhaust are expected to stay well within the
linear range of the calibration curve.
VALIDATION EXPERIMENTS
After selecting an analytical method, validation experiments were con-
ducted to determine the necessary sampling and procedural parameters. These
experiments included trapping efficiency, calibration curve linearity, and
interferences from dilute exhaust. Since the methylene blue procedure is
a well documented analytical technique, only simple experiments were con-
ducted to verify any other procedural parameters. Collection efficiency and
73
-------
8 r
C T.
Dark
Light
10
20
30
time, min.
Short term
40
50
60
10 r-
• Dark
Ligh'
20 30 40
time, days
Long term
Figure 23. Time-Light exposure study (low concentration)
74
-------
o
c
X
0
'J
25
20 -
15 -
10
• Light
• Dark
•
T 1 1
012
. , _ ^
~ ^ - —
\ Projected from
long term study
fill
345S
time, hours
Short term
o
o
20 r-
15
10
0 Light
Dark
30
time, days
Lona term
Figure 24. Time-Light exposure study (high concentration).
75
-------
1.6
1.5
1.4
1.3
1.2
1.1
1.0
o
d
I 0.8
0.6 -
0.5 -
0.4 -
0.1 _
• Day 1 (11/23/77)
• Day 2 (12/1/77)
K*
J I
1 i ll
j I
j I
0 10 20 30 40 50 60 70 80 90 100 110 _ 120 130 140 150 160 170 180 190 200
Sulfide concentration, |jy S=/100 tni
Figure 25. Beer's Law plot for methylene blue.
-------
other sampling parameters were determined with a series of experiments. The
first experiment determined the collection efficiency at room temperature
(23° to 26°C). A 5 ppm concentration of hydrogen sulfide was passed through
the absorbing reagent at 1.0 and 4.0 Vmin. The experiment was then repeated
with a sample flow of 4.0 H/min and an absorbing reagent temperature of 0° to
5°C. This temperature was achieved by immersing the impingers in an ice bath.
The data for this study is shown in Table 25.. Sample flow rate and absorbing
reagent temperature did not have a measurable effect on the collection ef-
ficiency.
TABLE 25. THE EFFECT OF SAMPLE FLOW RATE AND ABSORBING
REAGENT TEMPERATURE ON THE COLLECTION EFFICIENCY
Absorbing
Test Reagent Sample Percent H^S Collected per Bubbler
Number Temp., °C Flow Rate 1 2 3
1 23 1.0 98,0 1.7 0.3
2 23 1.0 98.2 1.3 0.5
Avg. 98.1 1.5 0.4
1 25 4.0 96.4 3.2 0.4
2 25 4.0 96.4 3.0 0.5
3 25 4.0 95.7 3.8 0.5
4 25 4.0 95.9 2.5 1.6
Avg. 96.1 3.1 0.8
1 0 4.0 95.0 5.0
2 0 4.0 93.9 5.3 0.8
3 0 4.0 92.8 6.2 0.9
4 0 4.0 92.7 6.3 0.9
Avg. 93.6 5.7 0.9
The interferences for this procedure are also well documented. In an
attempt to find the sources of these interferences, several experiments
were conducted. The first experiment involved the interferences produced
with only the absorbing reagent. A series of calibration gases were passed
through the absorbing reagent. These gases were 495 ppm carbon monoxide,
2.0 percent carbon dioxide, nitrogen dioxide, and a 5 ppmC hydrocarbon blend.
These were compared to background air and a blank with no gas bubbled through
it. Each gas was bubbled at 4.0 Vmin for twenty minutes and developed for
methylene blue. No interference from these gases was observed.
A second experiment investigated the interferences from individual
exhaust gas components on sulfide ion doped absorbing reagent. The gases
used were carbon monoxide, carbon dioxide, compressed air, a hydrocarbon
blend, sulfur dioxide and NOX. Each of these gases were passed through a
separate impinger filled with the doped absorbing reagent for twenty minutes
at 4.0 Vmin. These were then compared to the doped absorbing reagent after
development for methylene blue. Table 26 shows the results of this experi-
ment. NOX at 3000 ppm and sulfur dioxide at 5 ppm were found to quench the
production of methylene blue.
77
-------
TABLE 26. THE EFFECT OF INDIVIDUAL EXHAUST COMPONENTS
ON THE DEVELOPMENT OF METHYLENE BLUE
Methylene Blue
Apparent
Sulfide Ion
Gas
Doped absorbing
reagent
Carbon dioxide
Carbon monoxide
Sample
1
2
1
1
29,900
2,709
Absorbance
0.716
0.708
0.711
0.704
Cone., yg/m £
0.646
0.638
0.641
0.634
Doped absorbing
reagent
Air
Hydrocarbon
Sulfur dioxide
NOX
1
2
1
1
1
1
168
5
3,460
0.552
0.560
0.574
0.554
0.461
0.247
0.492
0.500
0.513
0.494
0.408
0.213
Doped absorbing
reagent
NO,
x
1
2
315
315
0.648
0.654
0.666
0.621
0.582
0.587
0.599
0.557
Sulfur dioxide
1
2
5
5
0.484
0.533
0.429
0.475
A third experiment was designed to check the interference of anions in
the development of methylene blue. Sodium salts of sulfate, thiosulfate, and
bisulfite ions were used. Each of these anions was added to separate solu-
tions of sulfide ion doped and undoped absorbing reagent. The solutions
were then developed for methylene blue and compared to the doped absorbing
reagent. Thiosulfate and bisulfite was investigated to determine the speci-
fic source of the sulfur dioxide interference. The results are shown in
Table 27. Only bisulfite ion and thiosulfate ion caused a negative inter-
ference .
Finally, an additional experiment was conducted to help determine the
fource of sulfur dioxide interference. Approximately 2.5 ft^ of 5 ppm sulfur
dioxide was passed through an impinger filled with the zinc acetate absorbing
reagent. This operation was then repeated six times. A 1 m£ aliquot of the
78
-------
standard sulfide ion solution was added to two of these impingers. To two
others, 5 mi were added. All six were developed for methylene blue. These
were then compared to similar concentrations of sulfide ion solution doped
absorbing reagent that did not undergo sulfur dioxide bubbling. These values
are shown in Table 28. Again, the absorbance for methylene blue was de-
creased by the presence of sulfur dioxide.
TABLE 27. THE EFFECT OF ANIONS ON THE DEVELOPMENT
OF METHYLENE BLUE
Anion
Sample
Absorbance
Apparent
Sulfide Ion
Cone., yg/m£
Sulfate ion
Thiosulfate ion
Bisulfite ion
Doped absorbing
reagent
Sulfate ion
Thiosulfate ion
Bisulfite ion
1
2
1
2
1
2
1
2
1
2
1
2
doped with hydrogen sulfide
0.238 0.205
0.241 0.207
0.207 0.177
0.211 0.181
0.239 0.206
0.231 0.198
0.241
undoped
0.001
0.000
0.009
0.004
0.007
0.005
0.207
79
-------
TABLE 28. THE EFFECT OF SULFUR DIOXIDE INTERFERENCE
ON THE DEVELOPMENT OF METHYLENE BLUE
Apparent
Sulfide ion Sulfide ion
added, m£ Absorbance cone., yg/m£
Sulfur dioxide passed
through absorbing reagent
1 0.067 0.055
1 0.070 0.057
5 0.463 0.410
5 0.326 0.284
0 0.002 0.000
0 0.006 0.000
No sulfur dioxide passed
through absorbing reagent
1 0.082 0.067
5 0.583 0.521
QUALIFICATION EXPERIMENTS
A 1972 Plymouth Fury with a 360 V8 engine was used in the qualification
experiments for hydrogen sulfide. No catalyst was present on this vehicle.
The baseline emission rate for this vehicle was below the detection limits
for the analytical procedure. This baseline was established with three
separate hot FTP driving cycles. Hydrogen sulfide was injected into the
CVS with and without the vehicle. The concentration of hydrogen sulfide
injected was 515 ppm. With the vehicle present, the hydrogen sulfide was
injected into the raw exhaust stream just before entering the CVS. Samples
were taken from the dilute stream and passed through a buffered zinc acetate
absorbing reagent. The samples were then analyzed with a Beckman spectro-
photometer. The data for this experiment is given in Tables 29 and 30.
Essentially 100 percent recovery was obtained without exhuast. However, a
decrease of almost 10 percent was observed with exhaust. These results
include corrections for baseline exhaust and background air concentrations
of hydrogen sulfide.
RESULTS AND CONCLUSIONS
The measurement of hydrogen sulfide in dilute exhaust can be conducted
with a colorimetric technique. Hydrogen sulfide is trapped in a buffered
zinc acetate solution. Upon treatment with N, N dimethyl p-phenylene
80
-------
TABLE 29. HYDROGEN SULFIDE GASEOUS RECOVERY
BY DIRECT CVS INJECTION
Nominal Flow
Rate, ftVmin
H2S blend
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
CVS
320
320
320
320
320
320
320
320
320
Run
1
1
2
2
2
3
3
3
Sample
1
2
3
1
2
3
1
2
3
Calculated
ppra H2S.
Dilute
0.62
0.62
0.62
0.65
0.65
0.65
0.69
0.69
0.69
Observed
ppm *
0.64
0.67
0.66
0.73
0.74
0.72
0.74
0.74
0.73
Percent
Recovery
H2S
102.7
107.5
105.9
111.9
113.4
110.4
107.1
107.1
105.6
Average 108.0
*This value is corrected for the background.
TABLE 30. HYDROGEN SULFIDE GASEOUS RECOVERY
FROM DILUTE EXHAUST BY DIRECT CVS INJECTION
Hydrogen Sulfide Concentration, ppm
Run
1
1
1
2
2
2
3
3
3
Sample
1
2
3
1
2
3
1
2
3
Calculated
Amount Injected
0.66
0.66
0.66
0.68
0,68
0.68
0.65
0.65
0.65
Observed
(exhaust+air) *
0.58
0.60
0.60
0.61
0.64
0.64
0.59
0.54
0.61
(air only)
0.01
0.01
0.01
0.01
0.01
0.01
0.00
0.00
0.00
Percent
Recovery
87.0
90.1
90.1
87.7
92.0
92.0
90.5
82.8
93.6
Average 89.5
*This value is corrected for the baseline
concentration in exhaust.
81
-------
diamine sulfate, and ferric ammonium sulfate, cyclization occurs to form
methylene blue. The analysis is conducted spectrophotometrically at 667 nm.
The minimum detectable concentration is 0.01 ppm,
Several experiments were conducted to determine the interferences and
their sources in the analysis of hydrogen sulfide. Such individual exhaust
gas components as carbon dioxide, carbon monoxide, and hydrocarbons show no
effect on the absorbance of methylene blue. NOX showed a negative inter-
ference only at concentrations ten times higher than that expected in dilute
exhaust. Sulfur dioxide also shows a negative interference at 5 ppm. To
determine the source of the sulfur dioxide interference, several experiments
were conducted. These experiments were discussed in a previous section. In
all cases with sulfur dioxide present, the absorbance for methylene blue was
decreased. Also, a broad peak was observed from 525 nm to 675 nm in the
visible region of the spectrum. The presence of bisulfite ion and thio-
sulfate ion produces the same effect. Bisfulfite ion can be produced from
sulfur dioxide by the simplified reaction:
SO + HO —> HSO3~ + H+
Thiosulfate ion forms bisulfite ion in strongly acidic solutions:
S2O3= + H+ —>• HSO3~ + S
Sulfate ion shows no interference.
The presence of sulfur dioxide in the absorbing reagent apparently
quenches the production of methylene blue from hydrogen sulfide. Sulfur
dioxide acts as an oxidizing agent toward hydrogen sulfide(free sulfur is
formed) and as a reducing agent toward methylene blue, an oxidation-
reduction indicator. Sulfur dioxide dissolves in water to form sulfurous
acid. Since the presence of bisulfite ion produces a similar interference,
the decrease in the absorbance for methylene blue is probably due to this
reaction. The elimination of the sulfur dioxide interference is not an easy
task. It should be recognized, however, that an apparent decrease in
concentration of hydrogen sulfide is observed when sulfur dioxide is present.
The sampling parameters used for the collection of hydrogen sulfide in
dilute exhaust were determined partly by necessity and partly by consistency.
The sample flow rate of 4.0 Vmin was selected to insure a sufficient sample
for analysis although the lower flow rate showed a slightly better collection
efficiency in the first bubbler. The use of an ice bath to cool the absorb-
ing reagent was selected for simplicity and consistency with the other
analytical procedures which require an ice bath. The absorbing reagent
temperature produces little or no effect on the collection efficiency for
ambient temperature sampling, but sample breakthrough is possible at
exhaust gas sampling temperatures greater than ambient conditions. Two
impingers filled with buffered zinc acetate absorbing reagent are necessary
for complete sample recovery. These parameters are sufficient to collect
sample concentrations within the detection limits of the procedure from
dilute exhaust.
82
-------
The qualification experiments revealed an average CVS recovery in dilute
exhaust of 89.5 percent. No losses were observed without the vehicle (no
exhaust); therefore, a 10 percent loss of hydrogen sulfide can be expected
in the presence of dilute exhaust. Since hydrogen sulfide reacts with sul-
fur dioxide under the proper conditions, the most important source of
hydrogen sulfide lost in the CVS is probably from the sulfur dioxide present
in dilute exhaust.
The effect of light on the stability of methylene blue was determined
for four samples over a period of weeks. Both concentrations were stable
for about two days whether exposed to the light or kept in the dark. The
high concentration samples required a slightly longer time to develop (30
minutes) than the low concentration samples. Both concentrations exhibited
a steady decay with time after the initial development. The sample decay was
independent of the exposure to light. At 10 days for the low concentrations
and 20 days for the high concentrations, an increase in the apparent concen-
tration was observed. Inspection of the entire wavelength extinction curve
showed that the absorbance was no longer due to methylene blue but some other
constituent in the solution. No attempts were made to specifically define
the source of this absorbance. However, the solutions were found to be
stable in the light for at least several hours after development in the dark.
If samples cannot be processed by this time, it is recommended that they be
discarded because of the difficulty in preserving the sample integrity.
Two possible absorbing reagents were compared to determine the best one
for the analysis. Zinc acetate was selected rather than cadmium hydroxide.
Ease of use, reduced toxicity, and photochemical stability were the criteria
for this selection.
This procedure provides a sensitive method for the analysis of hydrogen
sulfide in dilute exhaust. A single sample requires five to ten minutes to
add the reagents, thirty minutes to develop, and three to five minutes to
analyze the sample. Absorbing reagent stability helps to simplify the
analysis and makes this procedure ideal for analyzing a large number of
samples.
83
-------
SECTION 8
TOTAL CYANIDE PROCEDURE
LITERATURE SEARCH
Hydrogen cyanide is a flammable, toxic, colorless liquid at room tempera-
ture, with the characteristic odor of bitter almonds. Some synonyms for
hydrogen cyanide are hydrocyanic acid, prussic acid, and formonitrile.
Hydrogen cyanide is a covalent molecule and is capable of dissociation in an
aqueous solution, as are the hydrogen halides. Hydrogen cyanide (HCN) has a
molecular weight of 27.03, a boiling point of 25.70°C, and a melting point of
-13.24°C. It is a linear molecule with C-H and C^N bond distances of 1.06
and 1.15 A, respectively. It is a weak monoprotic acid with a dissociation
constate of 2.1 X 10~9. This highly poisonous compound is a respiratory
inhibitor and irreversible combines with the iron complex in the blood,
stopping the oxidation processes, in tissue cells and causing death by
asphyxiation. Commercially, hydrogen cyanide is prepared by reacting methane,
ammonia, and air over a platinum catalyst at 1000-1200°C, by the reaction of
nitric oxide and gasoline at 1400°C, the reaction of hydrocarbons, ammonia,
and oxygen at 600-1500°C, and many other methods. Reactions similar to these
may be responsible for the hydrogen cyanide produced in exhaust.
Cyanogen is a flammable, toxic, colorless gas at room temperature and
like hydrogen cyanide has the characteristic odor of bitter almonds. Some
synonyms for cyanogen are dicyan, oxalic acid, dinitrile, and oxalonitrile.
Pure cyanogen is stable, although the impure gas may polymerize to paracya-
nogen at 300° to 500°C or by exposure to ultraviolet light. Cyanogen
dissociates into CN radicals and can oxidatively add to lower valent metal
atoms, giving dicyano complexes. It resembles halogens in the disproportion-
ation reaction in basic solution:
(CN) 2 + 20H
Cyanogen (C^) has a molecular weight of 52.04, a freezing point of -27.9°C,
and a boiling point of -21.17°c. It is symmetrical and linear molecule with
a C-C bond distance of 1.37 A and a C=N bond distance of 1.13 A. Its physiol-
ogical effect on living tissue is similar to that of hydrogen cyanide.
Cyanogen is prepared by many methods: air oxidation of hydrogen cyanide over
a silver catalyst at 300-600°C, passage of hydrogen cyanide over cuprous
oxide at ambient temperatures, reaction of hydrogen cyanide and chlorine over
a surface-active material such as activated charcoal at >700°C, and many
others. In all cases above, cyanogen is produced from hydrogen cyanide.
Although none of these are exactly applicable for an automotive system, a
84
-------
similar process may be responsible for any cyanogen that is produced.
The analyses for hydrogen cyanide, cyanogen, and/or cyanide ion has
been performed by several basic analytical techniques: titration, colori-
metry, specific ion electrode, and gas chromatography. The Liebig determin-
ation of cyanide ion by titration with silver ion was discarded as a means
of analysis because of the low concentrations that were expected from exhaust
samples. Colorimetry has previously been used by SwRI and has been found to
be manpower intensive. An alternative procedure was sought with this factor
in mind. The best means of analysis was with either a specific ion electrode
or a gas chromatograph.
Three acceptable procedures were selected from the literature. Sekerka
and Lechner(Sl) reported the use of a cyanide ion-selective electrode for the
analysis of cyanide ion in waste water. The specific ion electrode was used
in conjunction with a colorimetric technique to determine the reliability of
the procedure. The samples were collected in sodium or potassium hydroxide
and analyzed potentiometrically. The minimum detectable limit reported was
about 2 ppb. The second technique required the use of Tedlar bag samples
and subsequent analysis with a gas chromatograph using a nitrogen phosphorus
detector (NPD). The third technique reported by Valentour et al (82) was
used with biological samples (blood, urine, and gastric contents). The
samples were collected in sodium or potassium hydroxide and the trapped
cyanide ion was reacted with chloramine-T to produce cyanogen chloride.
The cyanogen chloride was then analyzed with a gas chromatograph using an
electron capture detector (BCD). After preliminary experiments, the final
analytical procedure selected was a significant modification of the Valentour
et al procedure.
PROCEDURAL DEVELOPMENT
Attempts to analyze hydrogen cyanide and cyanogen separately were un-
successful and the details are reported below. The inability to analyze
hydrogen cyanide and cyanogen separately led to consideration of several
specific procedures for the analysis of hydrogen cyanide and cyanogen in the
form of cyanide ion.
Initially, it was decided to determine the concentration of hydrogen
cyanide and cyanogen by collecting a bag sample of the dilute exhaust and
analyzing it with a gas chromatograph using a nitrogen phosphorous detector.
This detector was selected because of its specificity to carbon-nitrogen
compounds. Hydrogen cyanide and cyanogen can be resolved with a 6' X 1/4"
O.D. glass column packed with 100/120 mesh Porapak QS. Isothermal column
temperature operation at 50 °C and a helium carrier gas flow rate of 60 ml/min
were the column conditions. A glass lined injector and interface was also
used to preserve sample integrity.
Bag stability experiments with hydrogen cyanide and cyanogen were
conducted to determine if sample integrity could be maintained over a short
period of time. Bag stability is necessary due to time required to collect
the sample and the subsequent waiting period before the sample can be
analyzed. The bag sample lifetime should be at least two hours after a sample
85
-------
is collected. Clear and aluminum foil tape covered Tedlar plastic bags were
used to conduct bag stability experiments. Dark bags (aluminum foil tape
covered) were tested to determine the effect of photochemical decomposition
on hydrogen cyanide and cyanogen.
A list of bag stability experiments which were conducted is shown in
Table 31. Each bag contained approximately one cubic foot of the diluent
gas. Experiments were conducted with nitrogen, air, dilute exhaust and humid
nitrogen. Dilute exhaust was selected at random from bag samples generated
during other tests, and humid nitrogen was generated by passing nitrogen
through an impinger containing deionized water. Hydrogen cyanide and/or
cyanogen were then added to each bag to give a nominal concentration of
about 2 ppm. At twenty to thirty minute intervals, 5 mil of the gas was
removed with a glass gas-tight syringe and injected into the gas chromatograph.
The percent change in the concentration was then calculated using the initial
injection. Figures 26 through 30 show the effect of elapsed time on the
stability of hydrogen cyanide and cyanogen. Figure 26 demonstrates the
stability of hydrogen cyanide in clear and dark bags with a variety of atmos-
pheres. Peak areas for hydrogen cyanide remained within the nominal range of
injection variability for at least 80 minutes, and no definite trends were
observed. (The nominal range of injection variability was set at +_ 1 percent
and is indicated in all Figures by a dotted line.) On the other hand, cyano-
gen showed a considerable percent loss in the clear bags with both nitrogen
and exhaust (Figure 27). In the dark bag, cyanogen remained stable except in
the presence of exhaust.
Figures 28 and 29 show the effect of a blend of hydrogen cyanide and
cyanogen in clear and dark bags with the various atmospheres. Again, hydrogen
cyanide was stable within the limits of injection variability for a short
period of time. In both cases, hydrogen cyanide was stable on the order of
about 60 minutes. Cyanogen behaved similarily to hydrogen cyanide in the
clear bag, but a steady decrease in concentration was observed in all atmos-
pheres with the dark bag.
In humid nitrogen (Figure 30) there was a 70 percent loss of hydrogen
cyanide after only 20 minutes. After this initial loss, the level of hydro-
gen cyanide remained relatively constant. Cyanogen under the same conditions
showed only a slight decrease in concentrations.
The short bag lifetime of hydrogen cyanide and cyanogen prevent the use
of grab samples of exhaust. At this point, the alternative procedures were
investigated for the analysis of hydrogen cyanide and cyanogen. These tech-
niques required samples to be collected in an aqueous solution.
At the same time the work with bag samples was underway, efforts to
develop a procedure using the specific ion electrode were being conducted.
Potassium cyanide standard solutions were prepared with 0.1M potassium
hydroxide. A calibration curve was to be determined by plotting the
measured potential in millivolts as a function of the log of the cyanide
ion concentration. Instability of the potentiometric measurement
was observed in all concentration ranges, especially in the low concentration
range. Attempts to improve the electrode stability and potential drift were
86
-------
TABLE 31. EXPERIMENTS CONDUCTED FOR HCN AND C N BAG STABILITY
Dark Bag
Nitrogen Air Exhaust
XXX
XXX
XXX
X
* Blend of hydrogen cyanide and cyanogen in humid nitrogen
Compound
HCN
CJN,,
2 2
HCN & CJSL
2 2
HCN & C_N *
2 2
Clear Bags
Nitrogen Air Exhaust
XXX
XXX
X X
-
87
-------
0)
tn
C
U
•P
a
o
o
a.
15
10
-5
Nitrogen
Air
Exhaust
Nominal Range for Injection
Repeatability ^
-••A
-10
20 40 60 80
Time, Minutes
Hydrogen cyanide in clear bags
100
120
-------
-------
(U
en
a
a
o
c
Q)
o
30
20
10
Q) -10
-20
-30
a;
1
,c
u
C
0)
o
M
a)
04
30
20
10
-10
-20
-30
Nitrogen
Air
Exhaust
Nominal Range for
Injection Repeatability
I
20 40 60 80
Time, Mi nute s
Hydrogen cyanide in clear bags
100
120
Nominal Range for
Injection Repeatability
20
40
80
100
120
60
Time, Minutes
Hydrogen cyanide in dark bags
Figure 28. The effect of elapsed time on hydrogen cyanide in a blend
of hydrogen cyanide and cyanogen in clear and dark bags.
90
-------
c
V
0
30
20
0) 10
cr
Hi
-20
-30
Nitrogen
Air
Exhaust
Nominal Range for
Injection Repeatability
30
20
10
C
<3
O
4-1
8
!M
01
(V
-10
-20
20 40 60 80
Time, Minutes
Cyanogen in clear bags
100
120
Nominal Range for
Injection Repeatability
I
I
20
40 60 80
Time, Minutes
Cyanogen in dark bags
100
120
Figure 29. The effect of elapsed time on cyanogen in a blend
of hydrogen cyanide and cyanogen in clear and dark bags.
91
-------
to
20
10
0
10
20
c 30
(0
c
0
u
M
0)
40
50
60
70
80
90
100
20
Nominal Range for Injection
Repeatability
Legend
Cyanogen
Hydrogen Cyanide
40 60
Time, Minutes
80
100
120
Figure 30.
The effect of elapsed time on hydrogen cyanide and cyanogen
in a dark bag with humid nitrogen.
-------
unsuccessful. Efforts using the specific ion electrode were abondoned for
another procedure using gas chromatography. A gas chromatograph procedure
which did not require bag samples for the collection of hydrogen cyanide
and cyanogen was investigated. This procedure used a chemical collection of
cyanide ion in sodium or potassium hydroxide. Initially, the analysis was
to be conducted by reacting chloramine-T with the trapped cyanide ion in an
acid buffered solution to produce cyanogen chloride. Cyanogen chloride was
then extracted by hexane and analyzed with an electron capture detector. The
electron capture detector was chosen because of its high sensitivity and
selectivity to halogenated compounds and relative insensitivity to hydro-
carbons. Problems with impurities in the hexane caused broad peaks with an
excessive analysis time. To eliminate the problem, the following items were
tried:
1. Temperature program sequences
2. Column backflush
3. Column changes
A. 6" X 1/4" OD glass column packed with 100/120 mesh Porapak QS
B. 6' X 1/4" OD stainless steel column packed with 50/80 mesh
Porapak Q
C. 6' X 1/4" OD stainless steel packed with 7 percent Hallcomid
M-18 on 90/100 mesh Anakrom ABS
4. Hexane purification with charcoal
5. Other extracting solvents (i.e., cyclohexane, etc.)
None of these proved to be satisfactory and long analysis times were the
result.
A modification of the above procedure was tried by eliminating the hexane
layer and conducting the analysis in the same manner as described above,
except that the sample was placed in an air tight reaction vial with a 1 m£
head space and a septum cap. A sample development period of 5 minutes was
required. After vigorously shaking the vial for 5 seconds, 100 yt of the
head space was injected into the gas chromatograph. An electron capture
detector was used for the analysis. As a result of this modification, a
rapid analysis time was achieved. The finalized analytical procedure is
included as an attachment to this report.
VALIDATION EXPERIMENTS
After selecting an analytical method, validation experiments were
conducted to determine detector linearity, detection limits, injection
repeatability, stability of reagents and sample, sampling parameters, etc.
Once the validations experiments were complete, the procedure was considered
ready for testing.
Collection parameters were determined with a series of experiments
93
-------
designed to check sample flow rates, absorbing reagent concentration, absorb-
ing reagent temperature, impingers or fritted glass bubblers, and collection
efficiency. All of these experiments were conducted with hydrogen cyanide.
The first experiments conducted were to determine the effects of stopper
tip, sample flow rate, and reagent concentration on the collection efficiency.
The results of these experiments are shown in Table 32. A set of three of the
same type collection devices (impinger or fritted glass tipped bubblers)
were filled with l.ON or 0.1N potassium hydroxide. A hydrogen cyanide cali-
bration blend that contained a nominal 2 ppm concentration in a balance of
nitrogen was passed through the absorbing reagent at 1.0 and 4.0 Vmin. Each
experiment was repeated three times. All of these experiments were conducted
in a special blending building, which was external to the main building. This
building did not have the normal temperature controls within the building and
the ambient temperature fluctuated with the weather. The room temperature
ranged from about 15 to 30°C during the experiments.
After careful examination of the data, several trends can be observed.
First, in all cases, more cyanide ion was collected with the stronger absorb-
ing reagent. Secondly, more cyanide ion was trapped with the impinger
than the fritted glass tipped bubbler. The possible reason for this was a
flow restriction due to the fritted glass tip. Finally, the higher flow rate
produced, more consistent results with both concentrations of the absorbing
reagent.
The next set of experiments took into account the results of the first
set plus the effect of reagent temperature. Five sets of three impingers
filled with 25 m& each of l.ON potassium hydroxide absorbing reagent were
used. The sample flow rate was set at 4.0 £/min. The first set of impingers
was sampled at ambient room temperature (16-29°C) and a second set of
impingers was sampled at ice bath temperatures. The sample collection effici-
ency for the ambient temperature experiments showed a high degree of varia-
bility. The collection efficiency for the first impinger was between 70 and
100 percent. At this temperature three impingers would be necessary to
collect the entire sample even at low concentrations. With the ice bath, the
first bubbler was sufficient to collect the entire sample as well as giving
more consistent results. The data for these experiments is shown in Table 33.
Detector linearity was demonstrated for two cyanide ion concentration
ranges. A linear response was observed in the 0 to 2 and the 0 to 10 yg
CN~/mi ranges. Table 34 and Figures 31 and 32 show the detector linearity.
All samples are expected to be within this concentration range. If samples
are obtained that are not in these regions, the samples will be diluted to
a concentration which falls within the linear response of the detector.
Sample injection reproducibility is essential for any gas chromato-
graphy technique which does not involve the use of internal standard. To
establish sample injection reproducibility, two nominal cyanide ion concen-
trations, 2.0 and 0.2 ;ig/m£, were used. Five separate samples of each
concentration were developed and injected. The results are shown in Table 35.
Three separate experiments involving the sample storage and sample
94
-------
TABLE 32. THE EFFECT OF STOPPER TIP AND ABSORBING
REAGENT CONCENTRATION ON COLLECTION EFFICIENCY AT ROOM TEMPERATURE
Collection
Device
impinger
impinger
impinger
bubbler
bubbler
bubbler
impinger
impinger
impinger
bubbler
bubbler
bubbler
Sample
Flow
1.0
1.0
1.0
1.0
1.0
1.0
4.0
4.0'
4.0
4.0
4.0
4.0
Run
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
1
93.47
65.71
74.28
77.82
11.44
0.00
0.00
3.81
37.15
29-18
28.96
31.76
20.20
25.30
18.94
21.48
pg
.0 N KOH
2
7.71
9.26
6.38
7.78
0.00
0.00
0.00
0.00
5.35
3.18
2.90
3.81
1.41
2.96
2.10
2.16
CN /ft"
3
1.55
2.61
0.00
1.39
0.00
0.00
0.00
0.00
0.67
0.60
0.35
0.54
0.00
3.52
0.32
1.28
sample
0.1
1
50.02
60.13
49.92
53.36
4.64
1.11
2.57
2.77
29.60
27.20
27.24
28.01
15.14
7.42
16.21
12.92
N KOH
2
7.08
6.06
6.16
6.43
1.40
0.00
0.00
0.47
3.53
3.24
2.69
3.15
1.85
0.85
1.78
1.49
3
6.21
0.00
5.48
3.90
2.73
0.00
0.00
1.37
0.77
1.08
0.62
0.82
0.00
0.00
0.00
0.00
95
-------
TABLE 33. THE EFFECT OF ABSORBING REAGENT TEMPERATURE ON HCN COLLECTION EFFICIENCY
Absorbing Reagent
Temperature
Date
10/11/77
10/13/77
10/17/77
10/17/77
10/17/77
10/12/77
10/12/77
10/12/77
10/12/77
10/12/77
Run
1
2
3
4
5
1
2
3
4
5
°F
72
61
63
73
84
32
32
32
32
32
°C
22
16
17
23
29
0
0
0
0
0
Time
min
20
20
20
20
20
20
20
20
20
20
Flow
fc/min
4.0
4.0
4.0
4.0
4.0
Avg
4.0
4.0
4.0
4.0
4.0
Avq
pg or/ft3
1
48.61
64.98
35.27
30.06
24.46
40.68
55.76
51.96
57.86
54.00
51.05
54.13
2
4.39
0.00
7.02
10.33
8.09
5.97
0.00
0.00
0.00
0.00
0.00
0.00
3
0.00
0.00
3.76
3.95
2.26
1.99
0.00
0.00
0.00
0.00
0.00
0.00
Total
53.00
64.98
46.05
44.34
34.81
48.64
55.76
51.96
57.86
54.00
51.05
54.13
1
1716.6
2294.7
1245.5
1061.6
863.8
1436.7
1968.1
1834.9
2043.3
1907.0
1802.8
1911.6
pg CN~/m3
2
155.0
0.0
247.9
364.8
285.7
210.7
0.0
0.0
0.0
0.0
0.0
0.0
3
0.0
0.0
132.8
139.5
79.8
70.3
0.0
0.0
0.0
0.0
0.0
O.O
Total
1871.7
2294.7
1626.2
1565.8
1229.3
1717.4
2147.1
1834.9
2043.3
1907.0
1802.8
1947.2
1
1.54
2.07
1.12
0.97
0.80
1.30
1.77
1.67
1.86
1.92
1.65
1.77
ppm CN~
2
0.14
0.00
0.22
0.33
0.26
0.19
0.00
0.00
0.00
0.00
0.00
0.00
3
0.00
0.00
0.12
0.13
0.07
0.06
0.00
0.00
0.00
0.00
0.00
0.00
Total
l.r,8
2.07
1,46
1.43
1.13
i . sr.
1.77
1.67
1.86
1.92
1.65
1. 77
-------
TABLE 34. CALIBRATION CURVE LINEARITY AT SEVERAL CYANIDE CONCENTRATIONS
Background
CN Cone .
Test Date yg/n&
1 10/11/77 9.64
4.82
1.93
0.96
0.00
2 10/12/77 9.64
4.82
1.93
0.96
0.00
3 10/03/77 1.93
0.96
0.48
0.19
0.00
4 10/04/77 1.93
0.96
0.48
0.19
:; 0.00
GC
Attn
X256
X256
X256
X256
X256
X256
X256
X256
X256
X256
X64
X64
X64
X64
X64
X64
X64
X64
X64
X64
Sample
height
45
26
11
5
1
40
24
9
5
0
80
42
19
8
1
68
36
17
6
0
area
4650
2644
1182
583
50
4102
2448
953
518
0
7939
4146
1947
817
123
6813
3538
2027
643
0
Corrected
height
44
25
10
4
0
40
24
9
5
0
79
41
18
7
0
68
36
17
6
0
area
4600
2594
1132
533
0
4102
2448
953
518
0
7816
4023
1824
694
0
6813
3538
2027
643
0
97
-------
10
10/0 V 7
• 10/04/77
00
o
M
X
3
o
0.2
I
0.4
0.6 O.R 1.0 1.2
Cyanidp Ion Concentration, (jq C
1.4
1.6
1.8
2.0
Figure 31. Total cyanide calibration curve at low concentrations (0-2 ppm).
-------
10
Cyanide Ion Concentration, |ty CN~/ml
Figure 32. Total cyanide calibration curve at low concentrations (0-10 ppra)
-------
TABLE 35. SAMPLE INJECTION REPEATABILITY FOR TWO CYANIDE CONCENTRATIONS
Sample
1
2
3
4
5
x
sx
Cv
1
2
3
4
5
x
Sx
Cv
Nominal
ppm
2.0
2.0
2.0
2.0
2.0
0.2
0.2
0.2
0.2
0.2
GC
Attn
X256
X256
X256
X256
X256
X32
X32
X32
X32
X32
Peak
Height
62
60
61
61
63
61.4
1.1
1.9
66
70
68
65
64
66.6
2.4
3.6
Area
6807
6550
6675
6703
6913
6730
137.5
2.0
6621
7095
6838
6711
6544
6762
216.0
3.2
lO'O
-------
stability were also conducted. Three separate samples of known concentration
were developed for the required time and injected as usual. At thirty minute
intervals, 100 yl of the remaining head space was also injected. The decay
of the peak areas for a two hour period is shown in Figure 33. Five separate
samples of equal concentration were developed for varying lengths of time.
The first sample was injected immediately, the second after 30 minutes, the
third after 60 minutes, the fourth after 90 minutes, and the fifth after 120
minutes. The sample decay was a function of time is also shown in Figure 33.
In both cases, the concentration of cyanogen chloride in the head space is
dependent on the length of time in which the sample was developed. The third
experiment involves the effect of real exhaust samples that have been stored
over a period of time. Sample storage stability is necessary when samples
cannot be processed immediately or if confusing data is to be checked at a
later date. A random sample was chosen and reprocessed periodically for
50 days. The results are shown in Figure 34. As a result, samples can be
stored for a period of several months without adverse effects.
The freshness and stability of the reagents is also very important for
the quantitative analysis of total cyanide. Solutions of both chloramine-T
and the buffer were stored for various lengths of time. Samples developed
with these stored solutions were found to be inferior to freshly prepared
reagents. For these reasons, the reagents should be prepared daily.
Several ions were tested for interference with the production of cyanogen
chloride or the production of other compounds with similar retention times in
the column. Those ions tested were sulfate, phosphate, permanganate, nitrate,
carbonate, chloride, bromide, cyanate, thiocyanate, and ammonium ions. The
potassium salts of each of these ions were prepared in 100 ppm and 1 ppm
concentrations in the presence of 4 ppm cyanide ion. The sulfate and nitrate
salts of ammonium ion were then tried after the potassium salts of the sulfate
and nitrate ions were found not to interfere. Aliquots of each were then
developed for cyanogen chloride. Sulfate, phosphate, nitrate, carbonate, and
ammonium ions showed no effect on the development of cyanogen chloride in the
100 ppm or 1 ppm ranges. Chloride, bromide, and permanganate ions produced
little or no effect at low concentrations. At high concentrations, both
bromide and permanganate ions decreased the concentration of cyanogen chloride
produced. On the other hand, chloride ion increased the concentration. Cya-
nate and thiocyanate ions produced a positive interference at both concentra-
tions. Apparently, these two ions also form a halide in the presence of
chloramine-T with the same retention times as cyanogen chloride.
QUALIFICATION EXPERIMENTS
Qualification experiments for the total cyanide procedure were conducted
with a 1978 Oldsmobile Cutlass Supreme with a 305 V8 engine. The Oldsmobile
was equipped with an oxidation catalyst. The baseline emission rate for the
Oldsmobile was 0.028 ppm total cyanide for a hot FTP driving cycle.
Two concentrations with a nominal concentration of 200 ppm and 500 ppm
hydrogen cyanide were injected into the raw exhaust stream. Samples were
passed through l.ON potassium hydroxide absorbing reagent and were analyzed
according to the gas chromatography procedure previously reported. Hydrogen
101
-------
I 20
(0
J5
CJ
-P
c
<1J
o
M
40
60
•••A
• Run 1
B Run 2
A Run 3
I
30 60
Time, Minutes
Sample decay with time
90
120
0
5
0)
en
1 10
u
c
°> 15
0 x;>
V
20
25
^
. \
- \
•
^^s.
*^— —
"" — ~"0^
^^r ^^^^^
^*"^*«
^V
T_ i i i i
0 30 60 90 120
Time, Minutes
Five samples with varying development time
Figure 33. The effect of elapsed time on sample development.
102
-------
ft-
c
0
•H
4J
-------
cyanide was also injected into the CVS without exhaust. Tables 36 and 37
represent the data from these experiments.
Without exhaust, the average percent recovery for both concentrations of
hydrogen cyanide is 103.0 percent. With exhaust, the average percent re-
covery for both concentrations is 92.9 percent. This calculation includes
a correction for the baseline emissions from the exhaust. There is a 6 to 7
percent loss of hydrogen cyanide in the presence of exhaust.
RESULTS AND CONCLUSIONS
The measurement of hydrogen cyanide and cyanogen in dilute exhaust can
be conducted with a gas chromatography technique. Cyanide ion is trapped in
a potassium hydroxide solution and reacted with chloramine-T to produce
cyanogen chloride. Injection of the cyanogen chloride determines the concen-
tration of cyanide ion in the sample. This procedure has a minimum detection
limit of 0.01 ppm cyanide ion.
The effect of interfering ions in the absorbing reagent was investigated.
The ions investigated included sulfate, phosphate, permanganate, nitrate,
carbonate, chloride, bromide, cyanate, thiocyanate, and ammonium ions. Sul-
fate, phosphate, nitrate, carbonate and ammonium ions exhibited no effect on
the cyanide ion concentration while chloride, bromide, and permanganate ions
interfered only at high concentrations. High concentrations of chloride,
bromide, and permanganate ions are not expected in dilute exhaust and the
cyanide ion concentrations should not be affected by these ions. The
presence of cyanate and thiocyanate ions affect the cyanide ion concentration,
and therefore, the definition of total cyanide must take into account the
possible existance and interference of these ions.
As a result of preliminary testing with real exhaust, it was discovered
that two bubblers were necessary to efficiently trap the cyanide ion that
was present in exhaust. Two factors which might necessitate the use of two
bubblers instead of one are presented below. First, cyanogen has a much
lower trapping efficiency that hydrogen cyanide in potassium hydroxide.
This difference in trapping efficiency was discovered while naming high
concentration cylinders which were to be used in the qualification experiments.
Secondly, the temperature of the exhaust stream is somewhat higher than the
temperature of the gases used in the determination of the sampling parameters.
The same breakthrough can be expected as with ambient conditions because the
sample gas is not cooled effectively by only one impinger in the ice bath.
Two impingers are therefore necessary for complete sample recovery. The final
sampling parameters are listed below:
1. 25 ml of 1.ON potassium hydroxide absorbing reagent.
2. Absorbing reagent held at ice bath temperature (Q°C - 5°C).
3. Sample flow rate of 4.0 Z/min.
4. Impingers rather than fritted glass bubblers.
104
-------
TABLE 36. TOTAL CYANIDE GASEOUS RECOVERY PROM DILUTE EXHAUST BY DIRECT
CVS INJECTION DURING A HOT FTP DRIVING CYCLE
Total Cyanide Concentration
as HCN, ppm
Actual ppm
Injected
210
210
210
485
485
485
Run
1
2
2
1
2
2
Sample
1
1
2
1
1
2
Calculated
Amount
0.208
0.232
0.232
0.542
0.562
0.562
Observed
(exhaust
+ air) *
0.229
0.239
0.232
0.553
0.553
0.543
Exhaust
Correction
0.028
0.028
0.038
Average
0.028
0.028
0.028
Average
Percent
Recovery
96.7
90.8
87.8
91.8
96.8
93.4
9J.il
94.0
* Values corrected for two bubblers. Only one was used at this time.
105
-------
TABLE 37. TOTAL CYANIDE GASEOUS RECOVERY BY DIRECT CVS INJECTION
Nominal
Actual ppm
Injected
210
210
210
210
485
405
485
485
Rate, ft
HCN Blend
0
0
0
0
0
0
0
0
.35
.35
.35
.35
.35
.35
.35
.35
Flow
3/min
CVS
350
350
350
350
350
350
350
350
Calculated
Run
1
2
3
3
1
2
3
3
Sample
1
1
1
2
1
1
1
2
ppm HCN
dilute
0
0
0
0
0
0
0
0
.213
.208
.218
.218
.471
.493
.510
.510
Percent
Observed Recovery
ppm* HCN
0
0
0
0
0
0
0
0
.214
.226
.225
.227
Average
.472
.465
.561
.528
Average
100
108
103
103
104
100
94
110
103
102
.4
.5
.0
.9
.0
.1
.3
.0
.5
.0
* Values corrected for two bubblers. Only one was used at this time.
106
-------
5. Two impingers in series.
These parameters were sufficient to collect a sample from dilute exhaust with-
in the detection limits of the procedure.
The measurement of hydrogen cyanide in the presence of cyanogen is
difficult if wet chemical techniques are used. In clear or dark Tedlar bags,
hydrogen cyanide is stable for at least 60 minutes, if the humidity within
the bag is not too high. High humidity increases the possibility of hydrogen
cyanide condensation on the walls of the bag. Cyanogen, on the other hand,
cannot be quantitively stored in the presence of exhaust. Therefore, bag
samples for the measurement of cyanogen is only a qualitative tool which can
determine if cyanogen is actually produced in exhaust.
Injection repeatability, sample stability, and sample storage are three
basic requirements for most analytical methods. The injection repeatability
is well within the expected nominal 5 percent limit for a gaseous syringe
injection. The concentration of cyanogen chloride within the head space is
dependent upon the volume of the head space, the room temperature, and con-
centration of cyanide ion present. A 5 m£ reaction vial with a septum cap is
used in the analysis. A total of 4 mJl of the various solutions is added to
this vial. When the vial is tightly capped, a 1 mJl head space remains above
the solution. This head space remains constant unless the vial is not tightly
capped or the wrong volumes of reagents are pipetted into the vial. Cyanogen
chloride obeys Henry's law in the head space. Henry's laws states that the
mass of a slightly soluble gas that dissolves in a definite mass of a
liquid at a given temperature is very nearly proportional to the partial
pressure of that gas. Henry's law holds for gases which do not chemically
unite with the solvent and is obeyed by a variety of gases in dilute solutions
and all gaseous solutions at the limit of extreme dilution. The sample
stability is maintained for only a short time after complete development. The
sample must be injected five minutes after adding the reagents. After that
time, the sample begins to decay and the sample integrity is not maintained.
A sample may be stored undeveloped in the potassium hydroxide absorbing
reagent for at least three weeks.
From the qualification experiments with hydrogen cyanide the average CVS
percent recovery in exhaust is 92.9 percent. A total CVS loss of 7 percent
for hydrogen cyanide is expected with dilute exhaust. Hydrogen cyanide is
known to react with alkenes, alkynes, aldehydes, and ketones under the proper
conditions. It also forms complexes with many metals. The apparent loss
nay be explained in this manner. No steps were taken to determine the
specific sources of this anomalous behavior.
This procedure provides a rapid and sensitive method for the analysis of
total cyanide in dilute exhaust. The analysis of a single sample requires
two minutes for reagent addition, five minutes for sample development, and
five minutes for the total peak elution time. Total sample processing time
is twelve minutes per sample. The simplicity and ease of analysis makes this
procedure ideal for repetitive analysis.
107
-------
SECTION 9
ORGANIC SULFIDE PROCEDURE
LITERATURE SEARCH
The organic sulfides that are included in this analysis are carbonyl
sulfide, methyl sulfide, methyl disulfide, and ethyl sulfide. The chemical
formulas, molecular weights, freezing points, boiling points, and common
synonyms are listed in Table 38. Carbonyl sulfide is the only sulfide of
interest that is a gas at room temperature. In general, the organic
sulfides are malodorous compounds that produce an unpleasant odor similar
to rotten eggs. The 1968 American Conference of Governmental industrial
Hygienists made no recommendation for threshold limit values for these
sulfides.
TABLE 38. LIST OF SULFUR COMPOUNDS INCLUDED IN
THE ANALYSIS OF ORGANIC SULFIDES
Sulfur Compound
Chemical
Formula
Molecular Freezing Boiling
Weight Point, °C Point, °C
Carbonyl Sulfide COS
Methyl Sulfide CH SCH3
Methyl Disulfide CH3SSCH3
Ethyl Sulfide C2H5SC2H5
60.075
62.13
94.20
. 90.19
-138.8
-98.27
-84.72
-1093.9
-50.2
37.3
109.7
92.1
Synonyms
Carbon oxysulfide
Dimethylsulfide
Dime thy Idisulfide
Diethylsulfide
Several gas chromatographic methods have been used for the analysis of
organic sulfides originating from a wide variety of sources. A gas chro-
matograph with a thermal conductivity detector has been used by several
workers to analyze gas odorants for mercaptans and/or sulfides (83-87);
however, none of these works were concerned with trace gas analysis. Gas
chromatography and mass spectroscopy were used to separate and identify low
boiling sulfur compounds in crude oil (88-90). Temperature programmed gas
chromatography was found to improve the separation of mercaptans and
sulfides (91-93). The separation and identification of hydrogen sulfide,
sulfur dioxide, mercaptans, alkyl sulfides, and disulfides in Kraft pulp
digester blow gas and black liquor combustion products was accomplished
using gas chromatography (94).
Carbonyl sulfide has been quantitatively measured in natural gas (95) and
in carbonated beverages (96) by the use of gas chromatography- The measure-
108
-------
merit of carbonyl sulfide in carbonated beverages used an electron capture
detector and had a detector limit of 0.3 ppm. Improved sensitivity in the
detection of sulfur compounds in waste process gases was accomplished by
concentrating the compounds on activated silica gel at -78.5°C, desorbtion
under heat and vacuum, trapping at -196°C, and transferring to a gas chro-
matograph for analysis (97).
Several columns have been used to separate sulfur compounds from nor-
mally occurring atmospheric hydrocarbons, but little success has been ob-
tained (98). A GC-microcoulometry method eliminated the interference from
the hydrocarbons and was sensitive to 1 ppm mercaptan (99) . A gas phase
chemiluminescent reaction of ozone with organic sulfides has been con-
sidered as method of detection in monitoring low concentration of ozone
and sulfur containing pollutants (100).
The detection limits for the analysis of sulfur compounds were improved
greatly with the development of the Melpar flame photometric detector (FPD).
The characterization of the FPD response to several sulfur compounds was
carried out by Mizany (101) . The FPD detector has been applied to low
concentration air pollution monitoring (65), measurement of trace organic
sulfides in air (102), and soil and water analysis (102). Permeation tubes
have been used in several cases to generate continuous samples of known
concentrations of various sulfur compounds (64,65). The use of Teflon
throughout the gas chromatograph system has been found to minimize absorptive
losses (102) and has increased sensitivity to 10 ppb (64).
Several columns have been evaluated at several temperatures in conjunction
with the Melpar flame photometric detector (70). The columns evaluated were
Chromosorb T, Carbopak B-HT-100, Chromosil 310, and Deactigel. A number of
other sulfur compounds have been quantitatively measured from a wide variety
of sources using gas chromatography (103-109).
PROCEDURAL DEVELOPMENT
From the results of the literature search it was determined that the
analysis of the organic sulfides should be conducted by the use of a gas
chromatograph (GC) equipped with a flame photometric detector.
A Perkin Elmer Model 3920 B gas chromatograph was dedicated for this
purpose. The instrument has a linearized flame photometric detector (FPD)
and a sub-ambient oven accessory. The sub-ambient oven accessory allows
for maximum flexibility in determining GC operating conditions.
A flow schematic of the gas chromatograph analytical system used in the
procedural development work is shown in Figures 35-37. The sample is purged
through the gas sampling valve sample loop (Figure 35 Step 1). The values
are maintained isothermally at 100°C in a valve oven. The sample is
injected into the gas chromatograph after the system has been efficiently
purged (Figure 36 Step 2). After all peaks of interest have eluted from the
analytical column, the column is backflushed and the system is readied for
the next injection (Figure 37 Step 3).
109
-------
Control Console
Step 1. Sample being purged
through injection valve
Figure 35. Proposed GC flow schematic for analysis of organic sulfides (Step 1).
-------
Control Console
Step 2. Sample injected
into GC system
GC Oven
Figure 36. Proposed GC flow schematic for analysis of organic sulfides (Step 2)
-------
Control Console
Step 3. Column backflusti
GC Oven
Figure 37. Proposed GC flow schematic for analysis of organic sulfides (Step 3)
-------
The column selected for the initial work was a 6' x 1/8" Teflon column
packed with 60/80 Chromasil 310. Several different GC operating conditions
were tried, and a preliminary set of conditions were selected that provided
an adequate separation of the four organic sulfides of interest. The
separation of these sulfides is presented in Figure 38. The tlution of
other sulfur containing compounds is also included. Table 39 presents a
list of chemical and physical characteristics of various sulfur compounds
that could'be present in automotive exhaust.
A lecture bottle of carbonyl sulfide, pure liquids of methyl sulfide,
ethyl sulfide, and methyl disulfide, along with a Tracer Model 412 Permeation
Calibration System containing all four sulfides, were used as sources for
the organic sulfides in the procedural development. Permeation tubes of
methyl mercaptan, ethyl mercaptan, hydrogen sulfide, and sulfur dioxide,
blends of hydrogen sulfide and sulfur dioxide in aluminum cylinders with
balance nitrogen, lecture bottles of hydrogen sulfide, sulfur dioxide, and
methyl mercaptan, and a pure liquid of ethyl mercaptan were used in the
interference checks.
Two methods of sample acquisition were considered for the analysis of
the organic sulfides. One method would be to use sample bags obtained
during the standard CVS testing. An alternate approach would be to use a
trap packed with a material such as Tenax GC for concentrating the sample.
In this manner an exhaust sample would be pulled through the trap during
the entire test, thereby giving an effective sample volume of several liters
rather than 5-10 m£. The use of the trap would increase the limits of
detectability by a factor of over 1000. The collection by the use of
sample bags was discarded due to the expected low concentrations of organic
sulfides in exhaust and the large losses of methyl sulfide, ethyl sulfide,
and methyl disulfide onto the walls of the Tedlar bags at ppb levels.
Because the concentration of the organic sulfides is expected to be very
low in exhaust, a number of experiments were conducted to investigate various
concentration techniques that may apply to the measurement of the organic
sulfides.
The first set of experiments involved the use of a U-tube type trap and
was conducted using several trap volumes ranging in size from 5 to 20 m£.
The basic flow schematic of the sampling system is shown in Figure 39. A
permeation gas blend of carbonyl sulfide and methyl sulfide was used in
these experiments, with the actual concentration depending on the particular
experiment. Two flow rates through the traps were used: 12.0 mVmin (9.50
ppm COS and 4.77 ppm CH3SCH3) and 81.2 mJl/min (1.40 ppm COS and 0.71 ppm
CH3SCH3). The traps were maintained at -78°C during the sampling period.
The purpose of these experiments was to see if it is possible to cold trap
(at -78°C) the sulfides and then use the cold trap as a sample loop on
the gas chromatograph system. Results of these experiments are presented
in Table 40. Based on these results, it was apparent that the carbonyl
sulfide and methyl sulfide could not both be retained under any of the
trap sizes or concentrations investigated. The only condition that
indicated there may be some possibility for this method was the large
trap loop (20 m£) at the higher flow rate and lower concentration. Even
113
-------
GAS CHROMATOGRAPH CONDITIONS
Perkin-Elmer 3920B w/FPD 6' x 1/8" column packed with 60/80 chromasil 310,
N2 at 20 m£/min., oven iso at 0°C for 8 min and programmed to 140°C/min.
at 32°C/min.
Figure 38. Gas chromatograph separation of several
organic sulfides in prepared blend.
114
-------
TABLE 39. LIST OF CHEMICAL AND PHYSICAL CHARACTERISTICS OF VARIOUS
SULFUR COMPOUNDS POTENTIALLY PRESENT IN AUTOMOTIVE EXHAUST
Sulfur Compound
Carbonyl Sulfide
Hydrogen Sulfide
Sulfur Dioxide
Dimethyl Sulfide
Dimethyl Bisulfide
Diethyl Sulfide
Me'thyl Mercaptan
Ethyl Mercaptan
Chemical
Formula
COS
H2S
SO2
CH3SCH3
CH3SSCH3
C2H2SC2H
CH3SH
C2H5SH
Molecular
Weight
60.075
34.08
64.063
62.13
94.20
90.19
48.11
62.13
Density
2.5300 g/Jl
1.5392 g/i
2.927 g/£
0.848 g/m£
1.0625 g/m&
0.836 g/m&
0.8665 g/m£
0.8391 g/m&
Boiling Retention
Point, °C Time
-50.2
-60.3
-10.0
37.3
109.7
92.1
6.2
35
2.8
4.2
10.5
17.5
15.8
12.5
13.5
115
-------
PERMEATION
CALIBRATION
SYSTEM
TRAP
en
LIQUID REFRIGERANT
VALVES
SYSTEM
CONTROL
CONSOLE
RECORDER
INTEGRATOR
PERKIN-FLMER
MODEL 3920B
GAS
CHROMATOGRAPH
Figure 39. Cold trap experiment flow schematic.
-------
TABLE 40. THE EFFECT OF COLD TRAPPING AT -78°C ON
CARBONYL SULFIDE AND METHYL SULFIDE AT VARIOUS CONCENTRATIONS,
FLOW RATES AND TRAP SIZES
Trap
loop,m£
5.0
5.0
10.0
10.0
15.0
15.0
20.0
20.0
5.0
5.0
10.0
10.0
15.0
15.0
20.0
20.0
Trap
Flow
m£/min
12.0
12.0
12.0
12.0
12.0
12.0
12.0
12.0
81.2
81.2
81.2
81.2
81.2
81.2
81.2
81.2
inj.
1
2
1
2
1
2
1
2
1
2
1
2
1
2
1
2
Cone. ,
COS Cl
9.50
9.50
9.50
9.50
9.50
9.50
9.50
9.50
1.40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
Trap
ppm
rr Cf^l.7
[iOOv*fl<3
4.77
4.77
Avg.
4.77
4.77
Avg.
4.77
4.77
Avg.
4.77
4.77
Avg.
0.71
0.71
Avg.
0.71
0.71
Avg.
0.71
0.71
Avg.
0.71
0.71
Avg.
Inlet
Peak
COS
84.5
84.5
84.5
84.2
84.2
84.2
84.2
84.2
84.2
85.0
85.0
85.0
63.0
63.5
63.2
63.0
63.5
63.2
63.0
63.5
63.2
63.0
62.0
62.5
Height
CH^SCH
16.1
16.0
16.0
13.3
13.4
13.4
13.3
13.4
13.4
14.6
14.4
14.5
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.2
5.5
5.0
5.3
Trap
Peak
COS
84.2
84.2
84.2
84.5
84.5
84.5
84.9
84.9
84.9
84.9
84.9
84.9
64.9
65.0
65.0
63.0
63.5
63.2
62.0
62.1
62.1
65.0
66.0
65.5
Exit
Height
CH^SCH.,
4.1
13.8
9.0
1.1
11.4
6.3
12.6
13.0
12.8
0.0
10.0
5.0
5.2
4^2
4.7
4.2
4.2
4.2
5.0
4.5
4.8
0.0
0.0
0.0
117
-------
in this particular case, the trapping was effective only on methyl sulfide.
One additional trap temperature was investigated prior to elimination of
cold trapping as a possible concentration technique. In this experiment the
trap was cooled to liquid nitrogen temperature. The results of this experi-
ment are presented in Table 41. Only one set of concentrations (1.40 ppm
COS and 0.71 ppm CH3SCH3) was used in this experiment. All sample loop
sizes were somewhat effective in collecting the carbonyl sulfide, with a
nominal collection efficiency ranging from 76 to 86 percent. The collection
efficiency for methyl sulfide was slightly higher, but the repeatability
was less, probably due to the low peak heights for this species. From
these experiments it was apparent that cold trapping, using traps up to 20
m£ in volume, and temperatures as low as -196°C would not quantitatively
remove either carbonyl sulfide or methyl sulfide.
Efforts were then directed toward determining the feasibility of using
short stainless steel cartridges packed with an absorbing material to con-
centrate the organic sulfides. The absorption traps are lengths of stain-
less steel tubing 2 inches in length and 3/8" OD. The material is held in
the stainless cartridge by stainless micron inserts in each of the unions
on both ends. Four packing materials were selected to be evaluated at four
collection temperatures. The four packing materials that were used in this
experiment include Tenax-GC, Chromosorb 102, Poropak Q, and Chromosorb T.
The collection efficiency of these traps was evaluated at temperatures of
20°C, 0°C, -78°C, and -196°C.
A permeation calibration gas sample containing 1.40 ppm carbonyl sulfide
and 0.71 ppm methyl sulfide in a balance nitrogen gas was used for this
study. The results of this study are presented in Table 42. Three of the
four packings are essentially 100 percent efficient in removing both carbonyl
sulfide and methyl sulfide at -78°C. These three packings were Tenax-GC,
Chromosorb 102, and Poropak Q. All of the traps except Chromosorb T were
effective in removing methyl sulfide at all of the temperatures investigated.
Problems were encountered using trap temperatures of -196°C. At this tem-
perature flow, restrictions were noted in the trap as the test proceeded.
There were also problems in desorbing traps that were stored at this
temperature.
The sulfides were thermally desorbed from the traps by connecting the
traps into the gas injection system with two quick connects and immediately
injecting the sample into the GC system and placing the traps inside a Lind-
burg furnace operating at 300°C. The carrier gas upon injection flows through
the loop carrying the contents of the trap into the gas chromatograph. The
300°C temperature is the temperature needed to thermally desorb the traps
without causing broad sulfide peaks which result from gradual thermal de-
sorbtion. The 300 °C temperature is also low enough to prevent the destruction
of the packing material in the trap. The packing material which gave the
most reproducible results in the desorbtion experiments was Tenax GC. For this
this reason and its stability at the 300°C desorbtion temperature, the Tenax
GC packing material was selected for use in subsequent experiments.
118
-------
TABLE 41. THE EFFECT OF COLD TRAPPING AT -196°C ON
CARBONYL SULFIDE AND METHYL SULFIDE WITH VARIOUS TRAP SIZES
Trap Trap Inlet Trap Exit
Trap Flow Cone. ,ppm Peak Height Peak Height
loop, mfc m5/min inj. COS CH3SCH3 COS CH3SCH3 COS CH3SCH3
5.0 81.2 1 1.40 0.71 66.7 5.3 15.8 0.8
5.0 81.2 2 1.40 0.71 66.7 5.3 15.8 0.1
Avg. 1.40 0.71 66.7 5.3 15.8 0.5
10.0 81.2 1 1.40 0.71 67.8 5.2 16.0 0.0
10.0 81.2 2 1.40 0.71 68.1 5.2 15.2 0.0
Avg. 1.40 0.71 68.0 5.2 15.6 0.0
15.0 81.2 1 1.40 0.71 67.8 5.2 8.0 0.1
15.0 81.2 2 1.40 0.71 68.1 5.2 10.8 0.1
Avg. 1.40 0.71 68.0 5.2 9.4 0.1
20.0 81.2 1 1.40 0.71 78.0 5.5 8.0 0.0
20.0 81.2 2 1.40 0.71 78.9 5.5 14.0 0.1
Avg. 1.40 0.71 78.5 5.5 11.0 0.05
119
-------
TABLE 42. THE EFFICIENCY OF VARIOUS MATERIALS
TRAPPING SULFIDES AT SEVERAL TEMPERATURES
Gas Chromatograjph Response-Peak St.
Trap
Inj.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg
1
2
Avg
1
2
Avg.
Temp.
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
20
20
20
0
0
0
-76
-76
-76
°C Trap
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tsnax-GC
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chronrosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Poropak Q
Chromosorb T
Chromosorb T
Chronosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Before
COS
68.2
69.5
68.9
65.0
63.5
64.3
64.3
65.0
64.9
68.2
69.5
68.9
64.0
64.2
64.1
67.2
67.2
67.2
62.0
63.6
62.8
64.0
64.2
64.1
67.2
67.2
67.2
65.0
63.5
64.3
64.8
65.0
64.9
64.0
63.0
63.5
CH3SCH3
4.2
5.3
4.8
4.5
4.7
4.6
5.0
5.0
5.0-
4.2
5.3
4.8
4.9
4.3
4.6
4.8
4.8
4.8
5.0
5.0
5.0
4.9
4.3
4.6
4.7
4.5
4.6
4.5
4.5
4.5
5.0
3.9
4.5
4.8
4.8
4.8
After
COS
64.5
67.5
66.0
64.5
62.0
63.3
0.0
0.0
0.0
66.5
68.0
67.3
20.8
20.0
20.4
0.0
0.0
0.0
63.2
62.. 6
62.9
35.5
45.0
40.3
0.0
0.0
0.0
64.8
65.1
65.0
64.0
63.2
63.6
62.5
61.0
61.8
CH3SCH3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.9
4.5
4.2
5.0
5.0
5.0
3.0
0.0
1.5
aTrap size 2" x 3/8 OD
120
-------
A Tenax trap at -76°C was used to collect the exhaust from a 1975 Model
350 CID Chevrolet engine for the 31 minutes of an FTP. The resulting trap
was thermally desorbed into the GC analytical system. The gas chromatograph
trace obtained from the desorbtion of the trap indicated that a substantial
number of sulfur containing compounds were present. The major drawback ob-
served was the exceptionally large sulfur dioxide (802) peak.
The concentration of sulfur dioxide in the dilute exhaust will normally
range from 2 to 5 ppm, whereas the other sulfides are present o^j.y in the ppb
range. The sulfur dioxide is unstable and the proposed method is not de-
signed to quantitatively measure sulfur dioxide. Since sulfur dioxide has
no quantitative interest, efforts were directed to determine if techniques
are available that would allow sulfur dioxide removal without altering the
concentration of other sulfides. A packing material containing sodium bi-
carbonate (NaHCC>3) has been reported to be very effective for this purpose.
Experiments have indicated that hydrogen sulfide (E^S) was not stable
enough to quantify with this procedure. In order to remove the interference
from sulfur dioxide and to remove any remaining hydrogen sulfide, the dilute
exhaust is passed through a 2" x 3/8" stainless steel cartridge packed with
5 percent sodium bicarbonate on 45/60 mesh Chromosorb P-AW DMC before enter-
ing the organic sulfide collecting Tenax GC trap. The sodium bicarbonate
trap effectively removes sulfur dioxide at 10 ppm levels and hydrogen sulfide
at 1 ppm levels without affecting the organic sulfide concentrations.
Initially, a set of gas chromatograph operating parameters was developed
to provide separation of hydrogen sulfide, carbonyl sulfide, sulfur dioxide,
methyl mercaptan, methyl sulfide, methyl disulfide, and ethyl sulfide. Since
hydrogen sulfide and sulfur dioxide are not of quantitative interest, the
original GC operating parameters were modified to shorten the analysis time.
The initial GC oven temperature of 0°C was maintained for four minutes and
then temperature programmed to 140°C at 32°C/minute. The entire analysis
time was about 25 minutes. Operation at 0°C was originally selected to allow
separation of hydrogen sulfide and carbonyl sulfide. Since this separation
would no longer be necessary, the GC oven parameters were changed to provide
a compromise between separation and analysis time. This new programming rate
provides for sample injection at 80°C followed by immediate programming to
140°C at 16°C/minute. A typical trace of the organic sulfides is shown in
Figure 40.
Several recovery experiments were conducted using the Tenax GC traps.
These experiments were designed to determine the recovery of the organic
sulfides from the Tenax GC traps. The recovery from these traps was very
erratic and was not satisfactory. Initially, it was felt that the lack of
reproducibility was due to the technique employed to remove the organic sul-
fides from the Tenax GC traps. However, the GC analytical column was later
found to be suspect. Contact was made with those researchers who originally
used the GC parameters to quantitatively measure organic sulfides. Their
findings were similar to those experienced at SwRI. When using this column
packing near its maximum operating temperature, very erratic results were
experienced. After looking into this more thoroughly, it was decided to use
a different column packing that would be reproducible and still yield
121
-------
o -1
CO —
-H--
-------
satisfactory separation of the organic, sulfides.
The column packing that was selected was a specially treated Porapak QS.
Although this column is reported to be stable at higher temperatures, the
separation characteristics are not as good as the Chromosil 310 column. The
ethyl sulfide and methyl disulfide elute together. A typical calibration
blend from a permeation system is presented in Figure 41. This column was
also found to give inconsistent results after repeated use and a different
analytical column was sought for use in measuring the organic sulfides.
After reviewing the literature and conducting a brief cursory laboratory
study, it was found that the column which has the necessary qualifications
for the organic sulfide analysis is a 6' x 1/8" TFE Teflon column packed with
60/80 mesh Tenax GC. A typical gas chromatograph trace using the Tenax GC
analytical column for the analysis of the four organic sulfide is shown in
Figure 42.
In order to determine the efficiency of the collection on Tenax GC ab-
sorbing traps, a secondary dilution of the permeation calibration system was
included. The organic sulfides were diluted from a 0.1 - 3 ppm level down
to the detection limits of the FPD. A sample of the permeation blend after
secondary dilution is presented in Figure 43. As noted, only two of the
four peaks are above the detection limits, although all four organic sul-
fides arid the concentrations are listed at their elution time. The concen-
tration of the organic sulfides with Tenax-GC traps appears to have tremen-
dous potential. An example of the permeation calibration blend (with second-
ary dilution) after being sampled at a flow rate of 45 m£/min for 10 minutes
is shown in Figure 44. Only the four individual organic sulfides are ob-
served, and no extraneous peaks (reaction products, etc.) are observed.
A system was developed to re-condition the Tenax GC traps by purging
the traps with nitrogen at 500 mSl/min for seven minutes at 300°C. Several
spot checks of traps that had been conditioned under these conditions indi-
cated no trace of organic sulfide carry-over from "used" Tenax-GC traps.
The procedure that is used to desorb the organic sulfides from the traps is
also very efficient in that no organic sulfides are retained in the trap
after the thermal desorption using the GC procedure.
The Tenax GC traps have been found to effectively remove 100 percent of
the organic sulfides from a permeation calibration flow at a flow rate of
130 mJl/min when the trap is maintained at -76°C. Higher flow rates were
tried, and a breakthrough into a back-up Tenax-GC trap was observed at a
flow rate of 250 m£/min. When this occurred, it was decided to return to
130 mVmin and use this flow rate as the primary sampling flow rate.
A variety of other trap designs, temperatures, and flow rates may be
equally acceptable; but for the purpose of developing a procedure with spe-
cific goals, these conditions have been selected.
Problems have been encountered with batch to batch variation in the
Tenax-GC which have caused repeatability problems. A procedure has been
implemented to validate each Tenax-GC batch prior to sampling as well as
each individual trap.
123
-------
NJ
— -
r
Sample
Instrument PS
Column l.f
_n,.tp NOV l} /T77
V«" P.P.
Operator J
I.D.
TF£
Packed with A/A % wt. .
JVA_
gg//00mesh PORAPAK OS
Run ISO @ f/Jfi °C. using "'
cc/min.
Type
Lic|. Phase
_ Support
Carrier
Inlet
isiq. At A Rotameter Rdg^.
ISO
°C fPD Type (other)
Hyd. 4O psig
Air *?/ psiq A//A Rotameter Rda
( ) AT/^ psig //7^ Rotameter RdgL
Recorder / in/min speed / mV.F.S.
Injection 10 A*l indicated /O
-------
:r t:
- 5
ul Actual
Samoli OR6AMIC SUUPIO6S
Instnunint P.E. 391O B
Column 4» ft. Va1' Q.D.
Packed with /Vrt X«t.
Date HOV Z9 l< Hotanwer Rdg.
Air 7O pH9 MA Botametar R
I ) V/> mia ^/j Botamaw ~
Injection IP ul indicated IP u> <**
17 16
Figure 42. Typical organic sulfide separation with Tenax-GC column.
125
-------
SVST -y DILUTED D.I. MOV 21 HTT.
33208 Operator P
Tfg
Typo
Run ISO 9 HA °C.
@ 7S" p«to. MA Rotameter Rdg;
Detector; / 6O °c ffO Type totherl
Hvd. 4O psiq /V/> Sotamear Rdo..
Typ«
Air__7g. Mia /V/>
( ) A/A osio yV/^ Potameiaf Rdq. ylf^
Recorder / dWmin io«ed / mV.F.S.
Iniecrion /O Ml Pndjcated IP H net.
Sampling Device
*.! Actual
20 19 18 17 16 15 14 13 12 11 10 9 8 76
Retention time, minutes
432
Figure 43. Organic sulfide permeation blend with secondary
dilution, near detection limit of GC FPD system.
126
-------
Instrument P£ "342O S
Column it ft. Vq 0.0.
Packad with tyA % wt.
»" 60/Somesn TSMA V -
Bun ISO @ A'/*4°C. using ^O cc/min
® "7S" osiq. Xipg flQtaniBtBr Rdg.
° *~- '
_/
-------
The procedure chosen for the analysis of the organic sulfides consists
of (1) collecting the organic sulfides on a Tenax GC trap at -76°C; (2) ther-
mally desorbing the organic sulfides from the trap into the GC sampling sys-
tem; (3) injecting the organic sulfides into the GC; (4) analysis of the or-
ganic sulfides with a GC equipped with a sub-ambient oven accessory, a 6' x
1/8" Teflon analytical column packed with Tenax-GC, and a flame photometric
detector; (5) and quantifying the results with the use of permeation cali-
bration tubes. A finalized copy of the procedure is included as an appendix
to the interim report.
VALIDATION EXPERIMENTS
Several experiments were carried out to determine the validity of the
organic sulfide procedure for the analysis of carbonyl sulfide, methyl sul~
fide, ethyl sulfide, and methyl disulfide. These experiments included checks
for: GC injection variability, linearity of detector response, sample stabi-
lity in the Tenax traps, trapping efficiency of the Tenax traps, interfer-
ences, and test repeatability.
The finalized sampling conditions used to collect the organic sulfides
are listed below as is a discussion on their selection. A 2" x 3/8" OD
stainless steel trap packed with Tenax-GC is used to collect 99+ percent of
the organic sulfides. During sampling, the trap is kept at -76°C in a dry
ice-isopropyl alcohol slurry. This temperature is necessary to effectively
trap the four organic sulfides from the dilute exhaust sample. Higher tem-
peratures (greater than 0°C) allow carbonyl sulfide to break through the
Tenax-GC trap. The other three sulfides can be effectively trapped even at
temperatures as high as 20°C. The sample flow rate through the trap is main-
tained at 130 m£/min. At higher flow rates (250 mJl/min) , breakthrough of the
organic sulfides occurs. A flip top filter, a Perma Pure Drier, and a trap
containing 5 percent sodium bicarbonate on 45/60 mesh Chromosorb P-AW DMCS
precede the Tenax GC trap. The flip top filter removes particulate from the
gas stream prior to flow through the Perma Pure Drier. If particulate is
allowed to enter the Perma Pure Drier, it could poison the drier and prevent
it from functioning properly. The Perma Pure Drier removes moisture from
the gas stream which could freeze out in the -76°C Tenax-GC trap, thus re-
stricting or stopping flow through the Tenax-GC trap. The 5 percent sodium
bicarbonate trap removes sulfur dioxide from the gas stream and prevents it
from collecting in the Tenax-GC trap. The sodium bicarbonate trap will re-
move 10 ppm sulfur dioxide at a sample flow of 130 mj!/min continuously from
dilute exhaust for periods up to 30 minutes. Tenax-GC was chosen as the or-
ganic sulfide absorbing material over the other packing materials due to its
trapping reproducibility and ability to withstand desorption temperatures.
The Tenax-GC traps can be used many times without replacing the Tenax-
GC packing material. There is a large deviation in trapping efficiency from
batch to batch of the Tenax-GC packing and each batch must be validated prior
to sampling. Each trap is conditioned in an oven operating at 325° +_ 25°C
for one hour with a flow of zero nitrogen (500 mJl/min) passing through the
trap. No carry over of sulfides in the Tenax-GC traps has been found from
test to test. This lack of carry over indicates that the desorption process
removes 100 percent of the sulfides collected on the trap.
128
-------
The sample traps must be stored at -76°C before desorption and analysis
or carbonyl sulfide will be lost from the traps. The other three sulfides,
methyl sulfide, ethyl sulfide, and methyl disulfide, are stable in the traps
overnight at room temperature. In most cases, all traps are run between one
half hour and three hours after sampling. The traps are capped after sam-
pling with miniature quick connects to prevent condensation'of water and
other compounds into the trap before analysis. After the traps have been
desorbed, they are again capped to prevent contamination before they are used
in sample collection again.
To determine the GC injection repeatability for the organic sulfide
procedure, a permeation standard containing 1.95 ppm carbonyl sulfide,
3.31 ppm methyl sulfide, 0.84 ppm ethyl sulfide, and 0.20 ppm methyl disul-
fide was injected into the GC analytical system six consecutive times. The
results of this injection repeatability experiment are presented in Table 43.
The percent deviation varies from 1 percent for methyl sulfide to 6 percent
for methyl disulfide. This deviation appears to increase with decreasing
concentration of organic sulfide.
TABLE 43. INJECTION REPEATABILITY FOR THE ORGANIC SULFIDES
Average GC Standard Percent
Compound Peak Area Deviation Deviation
Carbonyl Sulfide 29596 420 1.4
Methyl Sulfide 52325 449 0.9
Ethyl Sulfide 11201 254 2.3
Methyl Disulfide 3951 243 6.2
To determine the test-to-test repeatibility for the procedure two ex-
periments were carried out. In the first experiment, organic sulfides from
a diluted permeation blend were collected on a Tenax-GC trap, desorbed into
the injection system, and injected into the analytical gas chromatograph
system. This sequence was repeated 5 times using the same Tenax-GC trap and
the resulting GC peak areas for each of the organic sulfides were averaged
over the 5 tests. Standard deviations and percent deviations were also de-
termined for the organic sulfide GC peak areas. The results of this experi-
ment are presented in Table 44. Standard percent deviations ranged from
7 percent for methyl sulfide to 10 percent for carbonyl sulfide and methyl
disulfide. The second experiment was identical to the first experiment
except that 5 different traps were used to collect the organic sulfides
instead of using the same trap 5 times. Table 45 shows the results of this
experiment. Standard percent deviations ranged from 13 percent for methyl
sulfide to 26 percent for ethyl sulfide.
To determine the linearity of the detector for the concentration ranges
of interest, a permeation system containing permeation tubes of all four
sulfides was used to generate varying concentrations of the sulfides.
129
-------
TABLE 44. TRAP REPEATABILITY FOR
ORGANIC SULFIDE COLLECTION
Test 1
Test 2
Test 3
Test 4
Test 5
Standard
Deviation
Percent
Deviation
COS
Area
20241
22052
20092
22343
17378
20421
±1985
9.7%
Me2S
Area
71938
66098
74326
63777
65494
68327
±4548
6.7%
Et2S
Area
34417
35156
38123
30508
32516
34144
±2864
8.4%
M6 nS o
Area
41346
39537
41683
34147
34321
38207
±3718
9.7<
TABLE 45. TRAP TO TRAP REPEATABILITY
FOR ORGANIC SULFIDE COLLECTION
Test 1
Test 2
Test 3
Test 4
Test 5
Standard
Deviation
Percent
Deviation
COS
Area
47,590
46,830
44,465
50,590
25,788
43,053
±9,896
23.0%
Me2S
Area
68,995
43,429
66,874
63,440
76,180
65,784
±8,332
12.7%
Et2S
Area
25,716
20,549
17,283
30,144
16,631
22,065
±5,773
26.2%
Me2S2
Area
13,839
11,487
10,069
15,901
10,338
12,327
±2,491
20.2%
130
-------
Figures 45-48 show plots of the GC peak areas vs. the nanograms of each
sulfide injected into the GC system. Carbonyl sulfide and ethyl sulfide
give linear responses in the 1-200 ng region, methyl sulfide ives a
linear response in the 1-120 ng region, and methyl disulfide gives a linear
response in the 1-55 ng (higher levels of methyl disulfide were not tried)
region. Above 120 ng of methyl sulfide and 200 ng of carbonyl sulfide, the
detector is not linear, with the peak area not increasing proportionally
with the weight of sulfide injected. The concentration range of sulfides
in dilute exhaust which would fall in this linear range with the current
sampling technique is 0.1 to 25 ppb. If the concentration of sulfides in the
dilute exhaust exceeds a concentration of 25 ppb, a lower sample flow rate
will have to be used in order to collect a smaller amount of the sulfides.
Hydrogen sulfide, sulfur dioxide, thiophene, methyl mercaptan, and
ethyl mercaptan are sulfur containing compounds that could interfere with
the organic sulfide procedure. Sulfur dioxide is present in exhaust at
levels which would obscure all other compounds in the GC procedure if it is
not removed before it enters the Tenax-GC trap. A 5 percent sodium bicar-
bonate trap preceding the Tenax-GC trap effectively removes sulfur dioxide
from the exhaust without affecting the concentration of the organic sulfides.
Hydrogen sulfide at levels of less than one ppm do not pose a problem with
the procedure as no breakthrough of hydrogen sulfide into the Tenax-GC trap
is detected by the GC-FPD. However, if hydrogen sulfide is present at
concentrations of 4 ppm or greater, some hydrogen sulfide is collected on
the Tenax-GC trap and is detected by the GC-FPD. If this higher concentra-
tion of hydrogen sulfide is present the GC parameters can be modified to
prevent hydrogen sulfide from interfering with the analysis of carbonyl
sulfide. An oven temperature program which consists of holding the oven
temperature at 0°C for 4 minutes and then programming to 140°C at 8°/minute
will separate hydrogen sulfide and carbonyl sulfide by nearly two minutes.
This program does extend the analysis time from 25 minutes to 45 minutes if
the time it takes to recool the GC oven to 0°C is included. Methyl and
ethyl mercaptan have yet to be detected in exhaust. If present, the GC
operating conditions will separate these compounds from the sulfides of
interest. With the present operating conditions, thiophene has a retention
time that differs from ethyl sulfide by only seconds. Thiophene and ethyl
sulfide have not been effectively separated by changing the GC operating
conditions and therefore, thiophene must be included as a possible source
of error in the analysis for ethyl sulfide.
QUALIFICATION EXPERIMENTS
Qualification experiments were carried out using a 1972 Plymouth Fury
with a 360 V-8 engine (no catalyst) . Hot FTP (23 minute test) driving
cycles were followed to generate exhaust for the vehicle baseline emissions
and for the tunnel injection plus vehicle experiments. An aluminum cylinder
containing 2-5 ppm of each of the organic sulfides in balance nitrogen was
used as the source for the organic sulfides. Some tests were conducted
with an aluminum cylinder containing 40 ppm carbonyl sulfide in balance
nitrogen. The cylinders were named by comparing GC peak areas with the
50 Peak areas of the organic sulfides generated by the permeation system.
131
-------
u>
to
360 r~
320 -
100 200 300 400 500 600 700 800
GC Peak Area X0.01
900
1000
1100
1200
Figure 45. Carbonyl sulfide linearity plot.
-------
to
c
•H
•o
0)
4J
01
•o
360
320
280
240 -
200
160
a 120
c
80
40
I
100 200 300 400 500 600 700
GC Peak Area XO.Oi
800
900
1000
1100
120O
Figure 46. Methyl sulfide linearity plot
-------
U)
360 r-
320 -
100 200 300 400 500 600 700
GC Peak Area X0.01
800
900
1000
1100
1200
Figure 47. Ethyl sulfide linearity plot.
-------
OJ
j?
10
_L
J_
100
200 300
GC Peak Area X0.01
400
500
600
Figure 48. Methyl disulfide linearity plot.
-------
The flow of organic sulfides into the tunnel was regulated to give a con-
centration of 2-5 ppb of each of the organic sulfides in the dilution
tunnel. Injections of the organic sulfides into the tunnel without exhaust
gave, recoveries that varied from approximately 50 percent for methyl di-
sulfide to 80 percent for carbonyl sulfide (Table 46). The percent devia-
tion of the recovery percentages ranged from 15-20 percent. This value is
expected due to the trap-to-trap variations found in the validation experi-
ments. The recovery of the organic sulfides with real exhaust (excluding
carbonyl sulfide) varied from 5 percent for methyl disulfide to 20 percent
for ethyl sulfide (Table 47).
Carbonyl sulfide recoveries ranged from 0 to 471 percent. The baseline
emissions of carbonyl sulfide were erratic and of equal magnitude to the
carbonyl sulfide injected into the tunnel. This variation of carbonyl
sulfide from the vehicle, along with tunnel memory for carbonyl sulfide,
and trap-to-trap variations, made the percent recovery calculations very
difficult and thus gave the resulting 0-471 percent recovery. Undocumented
recovery experiments with higher concentrations of carbonyl sulfide (0.1 ppm)
and vehicle exhaust gave more consistent results, however, the amount of
carbonyl sulfide collected on the Tenax trap was out of the linear range of
the FPD and values could not be determined. Baseline emissions for methyl
sulfide, ethyl sulfide, and methyl disulfide were insignificant and did not
affect the recovery experiment.
There is a 20 to 50 percent loss of the organic sulfides in the CVS
tunnel without exhaust and a 80 to 90 percent loss (excluding carbonyl sul-
fide) with exhaust in the CVS tunnel. These losses must be taken into
account in determining organic concentrations when using this procedure.
RESULTS AND DISCUSSION
The concentration of organic sulfides can be determined by: 1) trapping
the sulfides in a Tenax-GC trap at -76°C, 2) thermally desorbing the sulfides
from the Tenax GC trap into the GC injection system, 3) injecting the organic
sulfides into the analytical GC system, 4) analyzing the organic sulfides
with a gas chromatograph equipped with a flame photometric detector, and
5) quantifying the results by comparison with standards generated by a per-
meation system. The organic sulfides are effectively caught in the Tenax-
GC trap at a flow rate of 130 mil/minute. The procedure has a minimum
detection limit of approximately 0.2 ppb.
The accuracy of the procedure in the 0.2 to 25 ppb range is on the
order of 25 percent due to trap-to-trap repeatability. The FPD gives a
linear response for the organic sulfides in the 0.2 to 25 ppb range. If
the concentrations of the organic sulfides exceed this range in dilute
exhaust, a lower sampling flow rate (less than 130 mVminute) must be used
to keep the detector response in the linear range of the detector.
Sulfur dioxide, hydrogen sulfide, and thiophene are possible interfer-
ences in the procedure. Sulfur dioxide is removed by the use of a sodium
bicarbonate trap, hydrogen sulfide can be separated in the GC system by
changing oven parameters; however, thiophene remains an interference to the
136
-------
TABLE 46. PERCENT RECOVERIES OF' THE ORGANIC SULPIDES
FROM THE CVS TUNNEL ONLY
Carbonyl Sulfide (2-40 ppb)
Percent Recovery
Average
% Recovery
87.7
72.9
71.8
107.3
62.9
94.1
79.0
82.2%
Methyl Sulfide (5 ppb)
Percent Recovery
69.0
95.0
60.2
56.7
86.8
Average
% Recovery
Standard
Deviation
73.5%
±16.7%
Standard
Deviation ±15.1%
Ethyl Sulfide (2-3 ppb)
Percent Recovery
96.6
52.1
50.0
67.0
89.8
68.6
77.1
41.7
Average
% Recovery
Standard
Deviation
67.9%
±19.4%
Methyl Disulfide (2-4 ppb)
Percent Recovery
84.4
18.8
39.0
48.6
26.0
73.8
53.7
37.6
46.4
Average
% Recovery
Standard
Deviation
47.6%
±21.1%
137
-------
TABLE 47. PERCENT RECOVERIES OF THE ORGANIC SULFIDES
FROM THE CVS TUNNEL AND EXHAUST
Carbonyl Sulfide (2 ppb)
Percent Recovery
0
0
93
0
120
421
Average
% Recovery 114%
Standard
Deviation ±181%
Methyl Sulfide (5 ppb)
Percent Recovery
3
1
44
13
8
0
Average
% Recovery
Standard
Deviation
12%
±17%
Ethyl Sulfide (2-3 ppb)
Percent Recovery
25
12
14
16
35
25
Average
% Recovery 21%
Standard
Deviation ± 9%
Methyl Disulfide (4 ppb)
Percent Recovery
0
0
0
0
14
8
Average
% Recovery
Standrad
Deviation
4%
6%
138
-------
procedure. The ethyl sulfide concentration is affected by this interference.
There is a significant loss of the organic sulfides in the CVS tunnel with
and without exhaust. These losses must be taken into account in determining
the concentrations of the organic sulfides when using this procedure.
Overall, the organic sulfide procedure should provide a relatively
accurate method for determining the concentrations of the organic sulfides
in dilute exhaust, and its use is recommended for this project.
139
-------
SECTION 10
NICKEL CARBONYL PROCEDURE
LITERATURE SEARCH
Nickel carbonyl (nickel tetracarbonyl) is a colorless, flammable, highly
volatile liquid. It has a melting point of -25°C, a boiling point of 43°C
(1 atm), and a density of 1.31 g/m£ at 20°C. Nickel carbonyl has a musty
"damp cellar" or "sooty" odor (110), and can be detected by the human ol-
factory at 1-3 ppm. In 1968 the American Conference of Governmental Indus-
trial Hygienists recommended a threshold limit of 0.001 ppm for daily ex-
posure without adverse effects. Nickel carbonyl is thermodynamically unstable
and decomposes at room temperature.
The toxicity and flammability of nickel carbonyl have provided an impetus
for the development of tests to determine low level concentrations. There are
several methods available for the detection of nickel carbonyl, including
infrared spectroscopy, spark spectroscopy, flame photometry, light reflec-
tance, colorimetry, gas chromatography, and chemiluminescence.
Infrared spectroscopy is reported to have a detection limit of 4 ppm by
volume (111). Spark spectroscopy has a 0.1 ppm detection limit, but the
exposure time of the photographic plate requires considerable time. The
flame photometric method is expensive to operate, requires the use of an open
flame, and has a detection limit of 0.1 ppm (112) .
There are several colorimetric methods available for the determination
of nickel carbonyl. One method traps the nickel carbonyl with a solution of
iodine in ethanol or chloroform. The nickel in the absorbing solution is
complexed with dimethylglyoxime and measured spectrophotometrically at 400 rm
(113,114). The method has a detection limit of 2 ppb, but the collection
time is on the order of hours. A second method uses a saturated solution of
sulfur in the trifluoroethylene to trap the nickel carbonyl and a UV spectre-
photometer to analyze the sample (115). A third method reacts nickel car-
bonyl with red mercuric oxide at 200°C. This method liberates mercury which
is measured spectrographically(116). A fourth method traps the nickel car-
bonyl in dilute sulfuric acid, and sodium diethyldithiocarbonate is used to
complex the nickel. Another method uses chloramine B in ethanol as the ab~
sorber, and the nickel is complexed with dimethylgloxime (117). Potassium
iodide on silica gel has also been used to collect nickel carbonyl (118) .
In this method, iodine is released and measured colorimetrically. Still
another method uses dilute hydrochloric acid as the absorbing reagent and
alphafuridioxime is the complexing agent.
140
-------
A gas chromatograph with a Carbowax 20 ' column and an electron capture
detector was employed to determi-e nickel carbonyl in the blood and breath
of rats (119). The technique involved vacuum extraction of nickel carbonyl
from the blood and trapping it in cold ethanol (-78°C) . The electron capture
detector is insensitive to most hydrocarbons but is sensitive to organo-
metallic compounds.
A device has been developed that takes advantage of the thermodynamics
of nickel carbonyl (120). In the absence of carbon monoxide, nickel carbonyl
decomposes on contact with a hot surface, in this case borosilicate glass.
The rate of decomposition is proportional to the nickel carbonyl concentra-
tion. The intensity of plane-polarized light reflected from the surface
is calibrated in ppm of nickel carbonyl. The detection limit is 0.05 ppm,
but this method is also sensitive to iron carbonyl and possibly other metallo-
organic gases. Another method which takes advantage of the thermodynamic
properties of nickel involves the detection of nickel carbonyl by its
pyrolysis products in an ionization chamber (121).
A chemiluminescent measurement which incorporates the light emitting
reaction of nickel carbonyl, ozone, and carbon monoxide has been reported to
have a detection limit of 0.02 ppb and does not necessitate concentration of
the sample (122) .
PROCEDURAL DEVELOPMENT
In conjunction with the Project Officer, it was determined that the
analysis of nickel carbonyl should be conducted by the use of chemilumines-
cence. Two instruments are available for the analysis of nickel carbonyl,
a modified Teco 12A chemiluminescent analyzer and an EPA supplied chemilumi-
nescent analyzer.
Four aluminum cylinders which were reported to contain 0.1, 1, 5 and
10 ppm of nickel carbonyl in balance carbon monoxide were used as sources
of nickel carbonyl in the procedural development experiments.
Since the chemiluminescent detector requires a gaseous sample for
analysis, and the chemiluminescent detector has a reported lower detection
limit of 0.02 ppb nickel carbonyl, the obvious method for collecting nickel
carbonyl for analysis would entail the use of Tedlar sample bags. The
system shown in Figure 49 was fabricated to allow for the analysis of nickel
carbonyl from bag samples and for the analysis of diluted standard samples
obtained from the aluminum cylinders.
The system allows the mixing of small volumes of nickel carbonyl with
known large volumes of zero air to obtain ppb levels of nickel carbonyl. The
diluted standard (or a bag sample which can be introduced at this point)
passes through a 0.5 micron pore size Fluoropore filter to remove any nickel
particulate which could interfere with the analysis. The diluted standard or
bag sample passes through a pump and is forced under pressure through a
needle valve (to regulate flow) and a flowmeter to monitor the flow. At
this point, the sample is mixed with an equal volume of carbon monoxide. The
presence of carbon monoxide enhances the nickel carbonyl signal and must be
141
-------
Quick
Connect
Glass Dilution Tube
Teflon Mixing
Baffles
Vent to
Atmosphere
to
TECO Model 12A
W/Kodak #54
Wratten Filter
Restrictor
Plowmeter
Quartz
Furnace
100%
CO
-^ LJ LJ ^^ ^J
A Flowmeters T *—- — — »
\ ^x Pump
x Needle Valves '
Under Hood
Dry
Gas
Meter
Zero
Air
_H3
Figure 49. Nickel carbonyl analysis and dilution system flow schematic.
-------
resent for the analysis of sub ppm levels of nickel carbonyl. Before mixing
ith the sample, the carbon monoxide is passed through a quartz furnace
operating at 480°C) to remove any nickel carbonyl which might be present
n the carbon monoxide. The flow rate of carbon monoxide is also regulated
ith a needle valve and monitored with a flowmeter. The carbon monoxide and
ample (bag or diluted standard) mixture enter the chemiluminescent analyzer
hrough the sample line. The sample mixture is split in the analyzer with
5 percent of the mixture exiting the analyzer through the bypass line. This
xiting mixture passes through a quartz furnace (480°C) and is bubbled through
nitric acid scrubber to destroy the nickel carbonyl before it is vented to
he atmosphere. The remaining 5 percent of the sample-carbon monoxide enters
he jeaction chamber where nickel carbonyl reacts with ozone (generated by
he analyzer) and carbon monoxide to give a light emitting species.
The Teco Model 12A was modified for the detection of nickel carbonyl by
eplacing the standard CL optical filter with a Kodak No. 74 Wratten filter.
he No. 74 Wratten filter allows the light emitted from the nickel carbonyl
o enter the detector, but prevents the interfering light of other species
ronr entering the detector. The. instrument was evaluated with this filter
nd found to have interference from nitric oxide and an olefin blend. The
o. 74 filter was replaced with a No. 4 Wratten filter to try to remove this
nterference. This replacement removed the olefin interference and most of
he r'.tric oxide interference. A signal 2-3 times the noise level of the
L instrument still rer^ ad for a 80 ppm nitric oxide sample. There is also
. fifty fold decrease ii, the sensitivity of the detector for nickel carbonyl
'hen the No. 74 filter is replaced with the No. 54 filter. The No. 54 filter
as selected over the No. 74 filter due to its removal of a large part of the
nterference signal from nitric oxide. A finalized copy of the nickel car-
onyl procedure is included as an appendix to the interim report.
ALIDATION EXPERIMENTS
Several experiments were carried out to determine the validity of the
hemiluminescent procedure for the analysis of nickel carbonyl. These
xperiments included checks for sample stability in Tedlar bags, linearity
f detector response, and interfering compounds.
Nickel carbonyl in balance carbon monoxide has a half life of several
eeks in Tedlar bags. In the presence of ambient air or exhaust, ppb levels
f nickel carbonyl are unstable and have half lives on the order of 40-70
inntes. The half life of nickel carbonyl in air or exhaust increases with
ncreasing concentration of carbon monoxide in the Tedlar bags. The use of
ark or clear bags has little or no effect on the stability of the nickel
arbonyl.
To determine the linearity of the chemiluminescent detector, nickel
'arbonyl in balance carbon monoxide was diluted with zero air to give varying
loncentrations of nickel carbonyl. Figures 50 and 51 show dilution ratios vs
«ak height for the instrument with the No. 54 and No. 74 Wratten filters.
fce chemiluminescent detector (with either the No. 54 or the No. 74 filter)
'ives a linear response for nickel carbonyl down to the detection limits of
he instrument.
-------
28
26 ~
o
o
o
1-1
X
o
o
e
•*
•H
<
f O
o
o
•H
S3
0)
a
o
0>
N
0)
g
3
rH
O
o
•H
4J
nj
G
O
-H
-P
r-l
-H
Q
A< ~X
22
20
18
16
14
12
10
8
6
4
2
I
0 10 20
Figure 50.
30 40 50 60 70 80 90 100 110
Peak Height 0.01 Range
CL detector linearity for nickel carbonyl
with a No. 74 Wratten filter.
144
-------
2.2 r-
2.0
20
30 40 50 60
Peak Height 0.01 Range
70 80 90
Figure 51. CL linearity for nickel carbonyl with No. 54 Wratten filter.
145
-------
Interference checks were carried out on the chemiluminescent detector
with the No. 74 filter and the No. 54 filter installed. Olefins and nitric
oxide give interfering signals when the No. 74 filter is installed. A
hydrocarbon blend containing 13.7 ppm ethylene, 20.7 ppm propylene, as well
as methane, ethane,, propane, acetylene, benzene, and toluene gave a signal
equivalent to 0.002 ppb Ni(CO)4« When the No. 54 filter is installed, the
interfering signal from this blend is removed. When the No. 74 filter is in
use, the major interfering species is nitric oxide, which has a rejection
ratio of 0.002. The signal generated by nitric oxide compared to nickel
carbonyl will be large and its presence would make the analysis for nickel
carbonyl difficult. By replacing the No. 74 filter with the No. 54 filter,
most of the interfering nitric oxide signal was removed. A 3000 fold de-
crease in the response of the instrument to a 80 ppm nitric oxide sample was
observed when the No. 54 filter was substituted for the No. 74 filter. . The
remaining nitric oxide signal is on the order of 2-3 times the noise level
of the CL instrument. However, a fifty fold decrease in the sensitivity of
the instrument to nickel carbonyl is experienced when the No. 54 filter is
substituted for the No. 74 filter. Even with this loss of sensitivity, the
detection limits are less than 0.1 ppb.
QUALIFICATION EXPERIMENTS
Based on the desires of the Project Officer, no CVS recovery experi-
ments were attempted with nickel carbonyl. The potential health hazards
of nickel carbonyl injections into the CVS were not considered justified
under these circumstances.
RESULTS AND DISCUSSION
The concentration of nickel carbonyl in dilute exhaust can be deter-
mined by collecting the exhaust samples in fedlar bags and analyzing the
samples with a chemiluminescent detector which has been modified by the
use of a No. 54 Wratten filter. The procedure has a detection limit on the
order of 0.1 ppb. The exact determination of the lower detection limit, as
well as the quantification of the results of the analysis, are difficult as
the nickel carbonyl standards have been found to be unstable and give varying
results.
The exhaust bag samples must be analyzed as quickly as possible as
nickel carbonyl is unstable in exhaust and has a half life on the order of
40-70 minutes. Nitric oxide is an interference in the procedure with an
80 ppm concentration of nitric oxide giving a response of about twice to
three times the noise level of the instrument. The levels of nitric bxide
must be monitored when interpreting any response from this instrument.
This procedure will allow the detection of nickel carbonyl in exhaust at
sub ppm levels if proper documentation and precautions are undertaken when
interpreting any results using this method of detection.
146
-------
SECTION 11
AMMONIA PROCEDURE
LITERATURE SEARCH
Ammonia is a colorless, corrosive, and weakly alkaline gas with a dis-
tinctive pungent odor. It has a molecular weight of 17.03, a boiling
point of -33.35°C (1 atm) , and a freezing point of -77.7°C (1 atm) . The
ammonia molecule is pyramidal in shape with N-H and H-H bond distances
of 1.016 and 1.645 &, respectively. The H-N-H bond angle is 106.67°.
Ammonia is soluble in water, ethanol, methanol, chloroform, and ether. The
basic nature of ammonia allows it to r^act with protonic acids to form
water soluble ammonium salts. It also reacts to form stable metallic com-
plexes. Chemically, ammonia is a highly associated, stable gas with only
slig.it dissociation at 840-930°C and atmospheric pressure. The toxicity
level of ammonia for h as is about 1700 ppm with an exposure of less than
30 minutes; however, the 1968 American Conference of Governmental Industrial
Hygienists recommended a threshold limit of 50 ppm, the amount to which most
workers may be exposed to repeatedly day after day without adverse affects.
Ammonia poisoning is not necessarily a serious health hazard though its odor
is perceptible at 20-50 ppm (5, 123, 124) . Commercially, ammonia is pro-
duced by the Haber Process according to the reaction:
N2+ 3 H2 T= 2 NH3
The reaction is carried out at 400-450°C and 200-600 atm over a specially
prepared catalyst composed of iron, potassium oxide, and aluminum oxide.
The most extensive use of ammonia in industry is in soil fertilization. It
is also widely used to manufacture nitric acid via the Ostwald process (5,
124).
A number of methods for ammonia analysis are available; however, most
of these methods are subject to interferences, especially from volatile
amines. These interferences affect a number of colorimetric procedures.
Nessler's reagent is sensitive to formaldehyde, alcohols, organic compounds,
amines, sulfides, acetone, and aldehydes (125). Distillation is necessary
to remove these interferences. The indophenol method is more sensitive
than Nessler's (126), but it too suffers from contamination by formaldehyde,
S02 (10:1), Fe, Cr, Mn, and Cu (127). During color development pH must be
carefully controlled for reliable results (126) . Another highly sensitive
Procedure is the pyridine-pyrazalone method. It is very involved and is
susceptible to interference from some cations at high concentrations (128) .
A direct colorimetric method for ammonia analysis involves collection in a
neutral solvent (dioxane) containing a quinone and subsequent absorbance
147
-------
measurement at 480 nm on a spectrophotometer. The major drawback of this
procedure is that one of the reagents, c—(benzenesulfonamide)-p-benzoquinone,
must be synthesized, purified, extracted with benzene, and recrystalized
before use (129). As with the colorimetric methods, the most serious weak-
ness of titrimetry methods is the large number of interferences. The
Kjeldahl procedure includes a distillation step, but this does not eliminate
the interference from volatile amines because they, too, distill over. The
classical titrimetry methods, acidimetry, complexometry, oxidimetry, and
formal titration are also used for the determination of ammonia, but they are
generally limited to 10~4 M solutions. Several instrumental optical methods
are used in ammonia analysis. These include the chloramine (130), cupri-
ammonia complex (131), ninhydrin (132), and electroanalytical methods.
Additionally, there is a method for direct measurement on a spectrophoto-
meter with a UV (133, 134) or IR (135, 136) detector as well as a number of
indirect colorimetric methods (137). Gas (138, 139) and paper (140) chro-
matography have been employed successfully for some applications. A number
of electrochemical techniques have also been developed for ammonia analysis.
Among these are amperometry, polarography, and coulometric acidimetry, and
oxidation (141,142). Interferences again pose a problem with these pro-
cedures. The specific ion electrode for ammonia is a relatively rapid and
direct electrochemical method for determination of ammonia. However, since
a longer equilibration time (about 20 minutes) is required for low NH3 con-
centrations, the ammonia gas tends to escape from the basic solution. This
long equilibration time causes unreliable results in the concentration
range of interest (143). Volatile amines interfere with analysis (144), and
the hydrophobic membrane of the electrode has been found to deteriorate in
2 to 3 weeks (143).
Both gasometric (145-147) and gravemetric (146,148, 149, 150) techniques
are not sensitive enough for trace analysis, and the chemiluminescent pro-
cedure is more involved than is practical (151,152). The disadvantage of
an enzymatic method reacting ammonia, an a-keto ester, and reduced nicotin-
amide adenine dinucleotide (NADH) is the high cost of NADH (129).
Ammonia has been quantitatively measured in dilute automotive exhaust
using an ion chromatograph. This procedure is free from many of the common
interferences that plague the classical methods. The short analysis time
(10-15 minutes) makes it a prime candidate for ammonia measurement.
PROCEDURAL DEVELOPMENT
The procedure chosen for the analysis of ammonia involves the use of a
new type of liquid chromatograph called an ion chromatograph. Ion chromato-
graphic analysis is direct, relatively rapid (15-20 minutes), and sensitive
to 0.01 ppm NH3. Heavy metals will contaminate the system, and sodium and
potassium ions interfere with ammonia detection at 2 ppm and 0.5 ppm, re-
spectively. However, the most common and troublesome interferences, vola-
tile amines, do not affact ammonia analysis on the ion chromatograph. The
standards made up in water and the samples collected in a weak acid solution
remain stable for at least a month allowing some delay time before pro-
cessing. These advantages made the ion chromatograph the best choice as a
means of measuring ammonia.
148
-------
Very little procedural development was necessa for this method of
analysis. Basically, instrument and sampling parameters needed to be chosen.
Suggested instrument variables such as type and strength of eluent, flowrate,
chartspeed, and separator column size were provided with the ion chromato-
graph. These will change somewhat with each set of columns, ultra pure
nitric acid and distilled water have been found to give the best baseline
and most rapid recovery from suppressor column regeneration. The regenerant
solution is a 0.5N NaOH solution made from reagent grade sodium hydroxide.
Chartspeed was set at 12 in/hour, and the flowrate at about 40 percent of
fullscale. Good separation was obtained with a 6 x 250 mm separator column.
A 3 x 150 mm precolumn (packed with the same resin as the separator column)
was placed on line prior to the separator column to trap heavy metals and
particulate. If these contaminates get past the precolumn they will poison
the separator and suppressor columns. The precolumn can be cleaned weekly
with a strong acid solution, as can the separator column if resolution de-
teriorates .
The sampling parameters were determined as part of the validation ex-
periments. A sampling rate of 2 lit/min at ice bath temperatures was found
to be most efficient. Two bubblers containing 25 m£ of 0.01 N H2SO. capture
over 99 percent of the ammonia passing through. A 0.5 micron fluoropore
filter was inserted on line before the bubbler to trap exhaust particulate.
It is necessary to heat the "line from 150 to 170 °F to prevent moisture con-
densation and thus ammonia capture in the sample line.
VALIDATION EXPERIMENTS
The first validation experiment conducted involved the selction of
sampling parameters: flowrate, collection temperature, number of bubblers,
absorbing reagent, filter and the use of heated or unheated sampling lines.
The most efficient flowrate and temperature combination was determined to
be 2 H/min at ice bath temperature. The data from these collection effi-
ciency tests if found in Table 48. Ninety-five percent of the ammonia was
collected at these conditions compared with 84 percent at 4 H/min, ice bath,
90 percent at 2 £/min, room temperature, and 92 percent at 4 £/min, room
temperature. Table 48 shows that 98.4 percent of total ammonia was trapped
in the first bubbler and 99.1 percent in the first two bubblers when the
nominal 100 ppm NH3 flow was diluted 1:5. With the 1:20 flow dilution 99.3
percent of the ammonia remained in bubbler one and 99.6 percent in bubblers
one and two. Two tapered tip bubblers containing 25 m£ of 0.01N 1^304 as
the absorbing solution are therefore used to trap 99+ percent of tne NH~.
Increasing the acidity of the absorbant (0.06N) causes interference with the
ion chromatographic analysis by broadening the eluted peaks. Of significant
importance is column contamination that occurs if particulate is not filtered
from the sample prior to analysis. To prevent this contamination fluoropore
filters and heated sample lines are used. The heated lines prevent conden-
sation of water and the loss of ammonia in the sample line. A test was conduct-
ed to determine if the filter retained any ammonia. A 1978 Oldsmobile Cutlass
was run on the FTP cold driving cycle. 0.12 ppm NH3 was recovered from the
filtered sample line and 0.11 ppm NH3 from the unfiltered sample line, a dif-
ference of 0.01 ppm ammonia. A second test was performed during the recov-
ery of ammonia experiment after injection of 1 percent NH3 into the exhaust
149
-------
TABLE 48. NH3 COLLECTION EFFICIENCY AS A FUNCTION OF FLOWRATE AND TEMPERATURE
concentration (ppm) and (percent)
NH3 concentration
Ui
O
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
-
14
15
Flowrate (8,/min)
2
2
2
2
4
4
4
4
4
2
2
4
4
2
2
Temperature ( °F)
Ammonia
32
32
32
32
Average
32
32
32
32
32
Average
74
74
Average
75
74
Average
Ammonia
32
32
Average
1
Elow diluted
17.10(99.6)
39.51(99.5)
18.34(95.2)
19.59(99.5)
18.64(98.4)
16.24(99.4
17.40(100)
14.17(98.3)
17.89(99.0)
17.79(99.9)
16.70(99.3)
17.84(100)
17.87(98.5)
17.86(99.2)
18.83(99.4)
18.05(100)
18.44(99.7)
flow diluted
4.32(99.7)
4.15(98.9)
4.24(99.3)
in bubbler
2
3
Bubbler corrected for flow
1+2+3 (ppm) dilution
1:5 with zero nitrogen
0.06(0.4)
0.10(0.5)
9.28(1.5)
0.06(0.3)
0.13(0.7)
0.02(0.1)
0
0
0.11(0.6)
0.02(0.1)
0.03(0.2)
0
0.17(1.0)
0.09(0.5)
0.02(0.1)
0
0.01(0.05)
1:20 with
0
0.02(0.5)
0.01(0.3)
0
0
0.64(3.3)
0.08(0.2)
0.17(0.9)
0.07(0.4)
0
0.24(1.7)
0.07(0.4)
0
0.08(0.5)
0
0.10(0.5)
0.05(0.3)
0.10(0.6)
0
0.05(0.3)
zero nitrogen
0.01(0.3)
0.03(0.6)
0.02(0.4)
17.16
19.61
19.27
19.68
18.93
16.33
17.40
14.41
18.06
17.81
16.80
17.84
18.14
17.99
18.94
18.05
18.50
4.33
4.20
4.27
85.80
98.08
96.33
98.40
94.60
81.66
87.00
72.06
90.33
89.03
84.02
89.21
90.70
89.96
94.72
90.25
92.48
86.63
84.00
85.32
-------
of the 1972 Plymouth test car. In the first run a loss of 29 percent NH.u
(0.93 ppm) was noted in the filter in comparison to the unfiltered sample and
a loss of 55 percent (2.38 ppm NH3) in the filtered sample was observed in
the second run. Due to the great ammonia loss to the filter, fluoropore fil-
ters are not used in the ammonia sampling line. Another test was conducted
with heated and unheated sample lines to determine the effect on the amount
of ammonia recovered. Two injections of 0.1 percent NHU were made into the
CVS (constant volume sampler) for 25 minutes each. Two of the sample lines
were heated and one was unheated. The first of the heated lines had a fil-
ter. The third test allowed a 10 minute soak and then the CVS was run for
25 minutes by itself. After the third test the CVS was allowed to run for
30 irutes before sampling for 25 minutes in the fourth test. The CVS was
all >-"3d to run for 1 hour and 10 minutes before the fifth test. The sampling
periol was again 25 minutes. The results of these tests are shown in Table
49. After the ammonia injections, it is obvious that a problem is noted with
the unheated lines. The amount of ammonia captured in the second test is
double the amount caught in the first. The ammonia values in test 3-5 also
remain higher than those obatined with heated sample lines. It appears that
some ammonia sticks to the unheated sample lines and is never completely re-
moved. For these reasons heated sample lines were used. The column can be
poisoned by heavy metals present in t1 , exhaust. These compounds adhere to
the column resin and will slowly elute causing broad, unidentifiable peaks
to ?^pear periodically To prevent contamination, a strong nitric acid
solution (1 N HNO.J if i to wash the precolumn weekly. If the separtor
column becomes contami ^d it is washed similarly. The lighter metals such
as sodium and potassium elute within a reasonable length of time (less than
12 minutes) , but they can interfere with the ammonia peak beacuse their re-
tention times are close to that of ammonia. Sodium, present in the water
supply, interferes when its-concentration exceed 2 yg Na /m£. The tolerable
limit for potassium is only 0.5 yg K+/m£. Its presence is due to the in-
complete rinsing of glassware washed in chromic acid solution. The absorbing
solution, 0.01N H2SO4, produces a small peak with the same retention time as
ammonia, but a correction is made for this by taking a background sample each
testing day. Filtered deionized water interferes negligibly (<0.01 ppm) with
ammonia analysis.
Another variable for which validation tests were run is the ion chroma-
tograph. The proper combination of eluent, columns, flowrate, and sample
loop size are reuired to obtain optimum results. 0.0075N HN03 flowing at
200 mJl/hour allows good separation between peaks when a 3 x 150 mm precol-,. i,
a 6 x 250 mm separator column, and a 9 x 250 mm suppressor column are usea.
These parameters will vary between column sets, making it necessary to chock
the eluent and flowrate when columns are changed. A small loop (0.01 m& •: r
0-2 m£) prevents the relatively small ammonia signal from being overwhelm d
ty the large hydrogen ion peak. An attempt was made to neutralize the ac;d
collection medium with sodium and potassium hydroxide, but the sodium and
potassium interferences were too large to make it practical. Injection r—
peatability figures are shown in Table 50. The mean or average for each
set of peak heights and areas is represented by 5, the standard deviation by
sxi and the coefficient of variation in percent by Cv. The coefficient c
variation serves as a comparison between injections made on the two days.
This value is simply the standard deviation divided by the mean and multj
Plied by 100. Calculations are done using peak areas rather than peak
151
-------
TABLE 49. RETENTION OF AMMONIA IN
CVS (constant volume sampler)
Sample Line
1'b 2b averages of 1 and 2 3°
ld
2d
3e
4e
5e
0.60 ppm
0.71
0.06
0.05
0.05
0.77 ppm
0.55
0.14
0.03
0.05
0.68 ppm
0.63
0.10
0.04
0.05
0.62 ppm
1.35
0.41
0.15
0.18
0.5 micron Eluoropore filter inserted in sample line prior to bubblers.
Heated sample lines.
GUnheated sample lines.
d0.1 percent NH3 injected into CVS.
eCVS run with no NH3 injected.
152
-------
TABLE 50. INJECTION REPEATABILITY
FOR ION CHROMATOGRAPH
mple
1
2
3
4
5
1
2
2
•-ate Concentration Attenuation Height
t\jg NH4*%
( jpj^ ) (yrohp) (in)
3-29-78 0.50 3 0.53
3-29-78 0.50 3 0.54
3-29-78 0.50 3 0.53
3-29-78 0.50 3 0.53
3-29-78 0.5- 3 0.50
x 0.53
sx °'01
Cv 2.8%
5-09-78 0.50 3 0.51
5-09-78 0.50 3 0.51
5-39-78 0.50 3 0.51
x 0.51
sv 0.00
.Ok
Cv 0.0
Area
30990
30374
29891
29665
28947
29973
766
2.6%
24335
23205
24785
24108
814
3.4%
153
-------
heights because on the whole they were more reliable. (Peak heights tended
to vary with width even though the area counts remained the same). The aver-
age variation in areas on the two days ran about 3.0 percent. A comparison 4.
yg NH4
was also made on the repeatability of standard preparation. Four 0.5 ^
standards were made up using the same stock solution and analyzed on the ion
chromatograph. The results are shown in Table 51.
TABLE 51. REPEATABILITY OF AMMONIA STANDARD PREPARATION
yg NH4+
Sample Concentration ( ; ) Attenuation Height (in) Area
1 '• ' • " inX/ """ ~" * ~~ ———-_
1 0.50 3
2 0.50 3
3 0.50 3
4 0.50 3
sx 0.015 1284
Cv 2.9% 4.5%
The coefficient of variation for the area is 4.5 percent. Subtracting the
Cv for injection repeatability, the repeatability of standard preparation
turns out to be 1.5 percent . A 4 . 5 percent error then is to be expected
from the instrument and standards. The particular combination of columns
and the condition of the suppressor column determines the actual repeat-
ability.
The ion chromatograph gives linear response to ammonia at the sensi-
tivity settings of 3 ymho and 10 ymho. Table 52 lists concentrations and
corresponding heights and areas of points on the calibration curve. These
values are plotted graphically in Figure 52. (NHA) 0SO , standards ranging
yg ^ 2 4
0.50
0.53
0.52
0.50
x 0.50
28947
29821
29011
26819
28650
from about 0.4 to 30 — ^ - (ppm NH4+) were run at the appropriate atten-
uations, 3 or 10 ygmo. The areas recorded at 10 ymho were corrected to 3
ymho by multiplying by 10/3. Both scales show linearity but the slopes are
visually different with relative values of 1.6 and 1 . 2 for the and the 10
ymho scales, respectively. The 3 ymho scale remains linear from at least
0.4 to 8 yg NH4+/m£ and the 10 ymho scale from 8 to 30 yg NH4+/m£.
Sample and standard stability as a function of time was another factor
investigated. The sample was a background sample taken during the three bag
FTP, SET-7, and FET driving cycles on May 29, 1978, and standard used for
yg NH4"r
comparison (0.5 - T - ) was prepared on May 29, 1978. The same test car
154
-------
1,400,000 I-
rfl
-------
TABLE 52. CALIBRATION CURVE FOR AMMONIA
Standard
Concentration
,^ 4 ^
( raf. '
0.36
0.50
0.72
1.00
1.44
1.50
2.00
3.00
3.61
4.00
5.00
7.22
8.00
8.00
10.00
14.43
20.00
28.86
Attenuation
(limho)
3
3
3
3
3
3
3
3
3
3
3
3
3
10
10
10
10
10
Height
(in)
0.38
0.56
0.79
1.17
1.68
1.89
2.41
3.57
4.21
4.56
5.52
7.50
8.49
2.49
2.93
3.97
4.78
6.30
Heights corrected Area corrected
to 3 ymho scale Area to 3 ymho scale
20,987
33,412
43,798
65,530
94,950
101,103
133,785
188,785
225,901
238,585
315,582
423,980
473,131
8.30 141,874 472,913
9.77 170,430 568,100
13.23 228,363 761,210
15.93 299,706 999,020
21.00 409,068 1,363,560
156
-------
used in previous test procedures was used for this particular test. It is
obvious from the data presented in Table 53 that the sample and standard are
stable for at least three weeks (23 days) . The drop to zero ammonia and the
jump to 0.02 ppm NH3 on the thirtieth and thirty-second days, respectively,
are probably due to instrument variation rather than sample degeneration. By
the fortieth day (sixth week), however, the increase in ammonia concentra-
tions is by 0.02 ppm. At this point the sample has probably begun to lose
integrity. This is further confirmed by the fact that a ten week old FTP
sample increased from 0.47 to 0.55 ppm NH3 (17-0 percent) and that two thir-
teen week old FTP and FET samples increased from 0.08 to 0.11 ppm NH3 (37.5
percent) and from 0.17 to 0.23 ppm NH3 (35.3 percent). It appears that after
six weeks the sample concentration begins to increase sharply, indicating a
sample and standard lifetime of four to five weeks.
TABLE 53. SAMPLE AND STANDARD STABILITY
AS A FUNCTION OF TIME
Date of Analysis Age of Sample (days) Concentration (ppm NH3)
5-29 1 0.01
5-30 2 0.01
5-31 3 0.01
6-01 4 0.01
6-02 5 0.01
6-05 8 0.01
6-13 16 0.01
6-20 23 0.01
6-27 30 0.00
6-29 32 0.02
7-07 40 0.03
QUALIFICATION EXPERIMENTS
The qualification tests for the ammonia procedure included determining
a baseline ammonia emission level produced by the test car and determining
ammonia recoveries from injection of nominal 0.1 percent NH3 and 1.0 percent
NH3 into the exhaust of the test car and into the constant volume sampler
(CVS) alone. The test car for these experiments was a noncatalyst 1972
Plymouth Fury equipped with a 360 V8 engine. The baseline ammonia emission
level for this car ran about 0.07 ppm as shown in Table 54. An injection
157
-------
TABLE 54. AMMONIA BASELINE FROM
PLYMOUTH FURY CAR
Test NH3 Concentration(ppm NH3)
1 0.07
2 0.09
3 0.11
AVERAGE 0.07 ppm NH3
of 0.1 percent NH3 was made into the CVS to determine the amount of ammonia
that is lost in the CVS tunnel (Table 55) . A similar injection was made into
the exhaust of the Plymouth Fury during the hot FTP driving cycle. 64.9 per-
cent of the ammonia was recovered from the CVS injection, however, as the
data in Table 56 indicates, no constant recovery figure was obtained for the
injection of 0.1 percent ammonia into the exhaust of the test car. With
seccessive injections the recovery of ammonia increase, from 2 percent to 31
percent. This trend was confirmed by subsequent injections of 1 percent NH3
into the exhaust of the same test car. These values are shown in Table 57.
Although the recoveries are larger due to the higher concentration of ammo-
nia injected, it is obvious that the recovery of ammonia increases with each
successive test (from 41 percent to 62 percent) . An additional test was con-
ducted to determine ammonia retention in the CVS. The experiment is des-
cribed in the validation section of this report, and the results are found
in Table 49. The obvious differences in ammonia recovery due to heated or
unheated sample line are noted in the validation section. However, for both
types of sample lines, the length of time the CVS must run beofre a constant
level of ammonia is obtained is about the same. After the last injection of
ammonia, test 2, and by the end of test 3 (25-minutes with the CVS on) , an
84 percent and a 70 percent drop occurs in ammonia recovered after passing
through heated an unheated lines, repsectively. By the end of test 3 (80
minutes with CVS on) a 62 percent drop in ammonia level occurs from the sam-
ple emerging from the heated line and a 64 percent drop in the sample con-
centration from the unheated line. Thereafter, the fluctuation appears to
be due to random error since the concentration of ammonia increases slightly:
0.04 to 0.05 ppm with heated lines and 0.15 to 0.18 ppm NH3 with unheated
lines. A CVS "wash out" period of an hour to an hour and a half is required
to remove between 85 percent and 94 percent of the NH3 trapped in the CVS
after ammonia injection. Following this leveling off, no further decrease
in ammonia concentration occurs with additional running of the CVS.
RESULTS AND DISCUSSION
The ion chromatograph was chosen as the most favorable means of mea-
suring ammonia becuase of the simple, direct, and rapid processing of samp-
pies. Most compounds that interfere with alternate ammonia procedures do
not affect ammonia analysis on the ion chromatograph. The selectivity and
sensitivity of this method warrants its use for the analysis of dilute auto-
158
-------
TABLE 55. AMMONIA RECOVERY FROM DIRECT CVS INJECTION
Test
1
2
3
Nominal Flow Rate
(ft3/min)
NH3 CVS
0.36
0.36
0.42
315
315
315
Ammonia
Injected
a volume NH3
injected
(ft3)
8.416
8.485
9.947
Concentration
NH3 injected
(ppm)
908
908
908
Ammonia
Recovered
"Total diulted
volume
(ft3)
7907
7913
7859
Sample
Concentration
(ppm)
0.66
0.63
0.71
Calculated
Amount of NHj
recovered (ppm)
0.97
0.97
1.15
Percent
Recovery
68.3
64.7
61.8
ftVG.= 64.9%
10
Volume corrected to latm and 68°F
''This concentration representing 100 percent recovery, was obtained by backcalculating from recovered sample concentration
and percent recovery.
-------
TABLE 56. AMMONIA RECOVERY FROM DILUTE EXHAUST BY DIRECT
CVS INJECTION OP 0.1 PERCENT N«3 DURING THE HOT FTP DRIVING CYCLE
I—1
CT*
o
Test Nominal Flow Rate NH-, Injected NH^ Recovered Calculated Percent
(ftj/min) Volume NHj Concentration Total Diluted Sample amount of Recover
NH3 CVS injected (ft3)b NH^ injected Volume Concentration NH, recovered
(ppm) (ft3)b (ppm NH3)C (ppm) "
1 0.39 315 8.545 908 7230 0.02 1.07 1.9
2 0.40 314 8.471 908 7180 0.19 1.07 17.8
3 0.39 315 8.376 908 7218 0.33 1.05 31.3
This concentration, representing 100 percent recovery, was o-tained by back calculating from recovered sample concentration
and percent recovery.
^Volume corrected to latm pressure and 68°F.
clncludes baseline correction
-------
TABLE 57. AMMONIA RECOVERY FROM DILUTE EXHAUST BY DIRECT
CVS INJECTION OF 1.0 PERCENT NHj DURING THE HOT FTP DRIVING CYCLE
•st
1
2
3
Nominal Flow Rate
(ft3/min)
NH3 CVS
0.25 319
0.26 319
0.26 320
Volume
injected
5.789
5.918
5.924
Nil-? Injected
NH3 Concentration
(ft3)b NH3 injected
(ppm)
9791
9791
9791
NHo Recovered
Total Diluted
Volume
(ft3)b
7301
7303
7311
Sample
Concentration
(ppm NII3)C
3.19
4.33
4.93
Calculated
amount of
NH.J recovered
(ppm) a
7.76
7.93
7.93
Percent
Recovery
41.1
54.6
62.2
concentration representing 100 percent recovery, was obtained by back calculating from recovered sample concentration
and percent recovery.
^Volume corrected to latin pressure and 68 °F.
""Includes baseline correction.
-------
motive exhaust samples.
The sampling parameters providing the most efficient collection of am-
monia were selected. Twenty-five militers of the abosrobing solution, 0.01N
H2SO., is placed in each of the two bubblers in series and maintained at ice
bath temperatures. Over 99 percent of the ammonia in the dilute exhaust
flowing at 2 Ji/min is captured in these two bubblers. The heated sample line
(150-170 °F) vaporized moisture that may retain ammonia. After sample col-
lection it is necessary to set instrument parameters to obtain good separa-
tion in the shortest time possible. These parameters, such as eluent con-
centration and flowrate, will depend on the particular column set in use. A
6 x 250 mm separator column has been found to resolve ammonia adequately.
The 3 x 150 mm precolumn. removes particulate, and the 9 x 250 mm suppressor
column neutralizes the acidic eluent. With these columns installed, the
eluent, 0.0075N HN03, flowing at 30 percent of pump capacity, gives good
ammonia resolution. A small sample loop (100 yl) is necessary to prevent the
very broad H+ peak from the acidic absorbing solution from obliterating the
ammonia signal. The injection variability of the ion chromatograph is 3.0
percent, and for standard preparation the variation is 1.5 percent. The ion
chromatograph gives a linear response for (NH4)2SO4 standards in the range
yg JNn^
0.4 to 30 —, however, the attenuator is not linear between different
mx,
sensitivity settings. Therefore, a different set of standards needs to be
run at each attenuation. The study conducted on the effect of age on sample
and standard stability showed the lifetime to be between four and five weeks.
Thereafter, sharp jumps in concentration of ammonia occur. The results of
the qualification tests show that the test car, a non-catalyst 1972 Plymouth
equipped with a 360 V8 engine, yields 0.07 ppm NH^ as a baseline emission
level. A constant value was not obtained from either test in the recovery
of ammonia injected into the exhuast of the test car during the FTP hot
driving cycle. Results from both injections (0.1 percent and 1 percent NH3)
show increases in ammonia recoveries with successive testing. This trend
can be explained by the fact that ammonia, an alkaline gas, is attracted to
the metal in the CVS, the initial injection of ammonia served mainly to coat
the inside of the CVS. Large portions of the second and third tests were
also lost. Additional injections would probably show the same increasing
trend. Following a similar injection into the CVS alone, 65 percent of the
ammonia was recovered. The loss can also be attributed to the coating of
the CVS tunnel. Removal of 85 to 94 percent of the ammonia remaining in the
CVS requires running of the constant volume sampler alone for an hour to an
hour and a half. Further running of the CVS yields no additional ammonia.
The ion chromatographic method measuring ammonia is an effective and
efficient means of ammonia analysis in dilute automotive exhaust. This pro-
cedure is insensitive to most of the interferences plaguing other widely
used methods. The ion chromatograph simplifies ammonia measurements to a
one step injection, avoiding intermediate processes such as distillation,
color development, or reagent preparation. The lengthiest portion of ammo-
nia determination is the actual analysis time. This 12-20 minute analysis
is relatively short for such a sensitive method (minimum detection limit is
0.01 ppm NI^). Sample and standard stability as well as linearity of
162
-------
response in the concentration range of interest are additional factors which
make this procedure the most desirable method of measuring ammonia in auto-
motive exhaust.
163
-------
SECTION 12
RESULTS AND CONCLUSIONS
To determine the suitability of the analytical procedures selected in
Task I for dilute exhaust analysis, validation and qualification experiments
were carried out. The validation experiments determined if the sampling and
instrument parameters were appropriate for the quantitative analysis of
dilute exhaust. The qualification experiments determined if the compounds
of interest could be quantitatively recovered from the CVS tunnel with and
without the presence of exhaust in the tunnel. The analytical procedures
to be used in this project are listed in Table 58 along with methods of
sampling and analysis. Table 58 also lists the validation and qualification
experiments that were carried out.
The sampling parameters for all procedures were found to be adequate
for the collection of each of the unregulated emissions. All samples, with
the exception of the organic sulfides, nickel carbonyl, and hydrogen sul-
fide are stable for several days and can be stored and rerun at a later date.
Hydrogen sulfide and organic sulfide samples must be run within hours after
sampling, and nickel carbonyl must be run as soon after sampling as possible
to prevent loss of sample integrity. All instruments demonstrate linearity
of response for expected concentration ranges (sample concentrations above
the linear range must be diluted to concentrations that fall within the
linear range of the instrument). The organic sulfides must be monitored
carefully as traps containing over 200 ng of sample fall beyond the linear
range of the FPD. The sample flow rate can be lowered to prevent overloading
the Tenax trap. Test-to-test repeatabilities for all procedures are docu-
mented in this report. In most cases, repeatability is difficult to obtain
at the lower concentrations, while the repeatability at high concentrations
is easily obtained. Interferences were checked and documented for each
procedure. Phthalates were found to interfere with the aldehyde and ketone
procedure and may cause erroneous results for crotonaldehyde. Nitric oxide
is an interference in the nickel carbonyl procedure. Nitric oxide must
be analyzed for separately and the results used to document the nickel
carbonyl results. In the hydrogen sulfide procedure, sulfur dioxide de-
creases the apparent hydrogen sulfide concentration, and its presence or
absence must be recorded. The other procedures have interferences that can
be avoided if care is taken.
Qualification experiments were carried out on the aldehyde and ketone,
organic amine, sulfur dioxide, nitrous oxide, hydrogen sulfide, total
cyanide, organic sulfide, and ammonia procedures to determine the recovery
of known amounts of each pollutant from the CVS tunnel with and without
exhaust. Aldehydes and ketones, sulfur dioxide and nitrous oxide can be
164
-------
TABLE 58. ANALYTICAL PROCEDURES FOR, EMISSIONS CHARACTERIZATION
Ui
Compounds
Aldehydes and Ketones
Organic Amines
Sulfur Dioxide
Nitrous Oxide
Individual Hydrocarbons
Hydrogen Sulfide
Hydrogen Cyanide + Cyanogen
Carbonly Sulfide + Organic Sulfides
Nickel Carbonyl
Ammonia
Sulfate
DMNA
Sampling
Impingers
Impingers
Impingers
Bags
Bags
Impingers
Impingers
Traps
Bags
Impingers
Filters
Traps
Analysis
DNPH
GC-NPD
Ion Chrom.
GC-ECD
GC-FID
Meth. Blue
GC-ECD
GC-FPD
CL
Ion Chrom.
BCA
GC-MS @ RTI
Validation
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Not required
Not required
Qualification
Yes
Yes
Yes
Yes
Not required
Yes
Yes
Yes
Not required
Yes
Not required
Hot required
-------
recovered quantitatively from the CVS tunnel with and without exhaust. There
is a 10 percent loss of hydrogen sulfide and hydrogen cyanide in the presence
of exhaust. The organic amines, ammonia, and the organic sulfides experi-
ence significant losses in the CVS tunnel with and without exhaust. These
losses must be taken into account when determining the concentration of these
compounds in exhaust.
The procedures discussed in this report are effective in collecting
and analyzing dilute exhaust samples and are recommended for use in Task in
of this report.
166
-------
ffiFEPENCES
1. Muir, G. D. (Ed.), Hazards in the Chemical Laboratory, The Chemical
Society, London, 1977.
2. Stump, F., Oxygenated Compounds in Automobile Exhaust-GAS CHROMATO-
GRAPHIC PROCEDURE, MSERB-ESRL-EPA, Research Triangle Park, North
Carolina.
3. Roberts, R. M., Giblert, J. C., et al; An Introduction to Modern
Experimental Organic Chemistry; Holt, Rinehart and Winston, Inc.
1974, pg. 470-471.
4. Weast, R. C. (Ed.), Handbook of Chemistry and Physics, 54 Edition,
Division of the Chemical Rubber Co., 1973.
5. Braker, W. and Mossinan, A. L., Matheson Gas Data Book, 5th Edition,
East Rutherford, N. Y. , 1971.
6. Hoshika, Y., Anal. Chem., Vol. 48, pg. 1716, 1976.
7. Hoshika, Y., J. Chromatog, Vol. 115, pg. 596, 1975.
8. Sze, Y. L. and Borke, M. L., Anal. Chem., Vol. 35, pg. 240, 1963.
9. O'Donnell, J. F. and Mann, C. K., Anal. Chem., Vol. 36, pg. 2097,
1964.
10. Me Curdy Jr., W. H. and Reiser, R. W., Anal. Chem., Vol. 38, pg.
795, 1966.
U. Smith, J. R. L. and Waddington, D. J., Anal. Chem., Vol. 40, pg.
522, 1977.
12. "Amine Analysis," Bulletin 737A, Supelco, Co., Belefonte, Pa.,
1973.
13. "Penwalt 223 Amine Packing," Alltech Associates, Arlington Heights,
111.
14. Moffat, A. C. and Horning, E. C., Anal. Lett., Vol. 3, pg. 205,
1970.
15. Clarke, D. D., Wilk, S., et al; J. Gas Chromatog., pg. 310, 1966.
167
-------
16. Mosier, A. R., Andre, C. E., et al; Environ. Sci. Tech., Vol. 7, pg.
642, 1973.
17. Burks Jr., R. E., Baker, E. B., et al; J. Agr. Food Chem., Vol 7,
pg. 778, 1959.
18. Dunn, S. R., Simenhoff, M. L., et al; Anal. Chem., Vol. 48, pg. 41,
1976.
19. Gruger Jr., E. H., J. Agr. Food Chem., Vol. 20, pg. 781, 1972.
20. Andre, C. E. and Mosier, A. R., Anal. Chem., Vol. 45, pg. 1971,
1973.
21. Umbreit, G. R. , Nygren, R. E., and Testa, A. J., J. Chroma tog., Vol.
43, pg. 25, 1969.
22. Keay, N. J. and Hardy, R., J. Scie. Food Agr., Vol. 23, pg. 9, 1972.
23. Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.
24. West, P. W. and Gaeke, G. C., Anal. Chem., Vol. 28, pg. 1816, 1956.
25. Adams, D. F., Corn, M., et al; "Methods of Air Sampling and Analysis,"
(Tentative Method of Analysis for Sulfur Dioxide Content of the At-
mospher (colorimetric), No. 42401-01-69T), pg. 447, Amer. Publ.
Health Assn., Washington, D. C., 1972.
26. "ASTM Standards - Water, Atmospheric Analysis," ASTM Designation
D291470T, Part 23, Am. Soc. Test. Mat., Philadelphia, Pa., 1972.
27. US EPA, Fed. Register, Vol. 36, No. 158, 15492, 1971.
28. Robinson, E. and Robbins, R. C., "Sources, Abundance, and Fate of
Gaseous Atmospheric Pollutants," Final Report, SRI Project 6755
Supplement, Stanford Research Inst., Palo Alto, Calif., 1968.
29. Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc.,
N. Y., 1976, Vol. 3, pg. 214.
30. Zurlo, N. and Griffini, A. M., Med. Lav., Vol. 53, pg. 330, 1962.
31. Scaringelli, F. P., Elfers, L., et al; Anal. Chem., Vol. 42, pg.
1818, 1970.
32. Pate, J. B., Ammons, B. E., et al; Anal. Chem., Vol. 37, pg. 942,
1965.
33. Pate, J. B., Lodge, Jr., J. P., et al; Anal. Chem., Vol. 34, pg.
1660, 1962.
168
-------
34. Scaringelli, F. P., Saltzman, B. E., et al; Anal. Chem., Vol 39
pg. 1709, 1967.
35. Hartman-Leddon, Philadelphia, Pa.
36. Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc
N. Y., 1976, Vol. 3, pg. 214. "'
37- Urone, P., Evans, J. B., and Noyes, C. H., Anal. Chem., Vol. 37,
pg. 1104, 1965.
38. Bostrom, C. E., Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.
39. Attari, A., Igielski, T. P., et al; Anal. Chem., Vol. 42, pg. 1282,
1970.
40. Kawai, T., Netsu Kanri, Vol. 22, pg. 20, 1970.
41. Thomas, M. D., Ivie, J. O., et al; Ind. Eng. Chem., Anal. Ed., Vol.
18, pg. 383, 1946.
42. Yocum, J. E. , Richardson, R. L., et al; Proc. 49th Ann. Meet. Air
Poll. Contr. Assn., Pittsburgh, Pa., 1956.
43. Kuczynski, E. R., Environ. Sci. Tech., Vol. 1, pg. 68, 1967.
44. Shikiya, J. M. and McPhee, R. D., 61st Annual Meeting, Paper No.
68-72, Air Poll. Contr. Assn., Pittsburgh, Pa., 1968.
45. Jacobs, M. B., The Chemical Analysis of Air Pollutants, Chapter 8,
Wiley (Interscience), New York, N. Y., 1960.
46. Jacobs, M. B., The Analytical Chemistry of Industrial Poisons,
Hazards, and Solvents, Chapter 9, Wiley (Interscience) , New York,
N. Y., 1949
47. Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.
48. Stratman, H. , Mikrochim. Acta., Vol. 6, pg. 688, 1954.
49. Pate, J. B. , Lodge, Jr., J. P., et al; Anal. Chim. Acta., Vol. 28,
pg. 341, 1963.
50. Lodge, Jr., J. P., Pate, J. B., et al; Amer. Ind. Hyg. Assn., J.,
Vol. 24, pg. 380, 1963.
51. Forrest, J. and Newman, L., Atmos. Environ., Vol. 7, pg. 561, 1973.
52. Huygen, C., Anal. Chim. Acta., Vol. 28, pg. 349, 1963.
53. Harding, J., and Schlein, B., "Nuclear Techniques in Environmental
Pollution," SM-142a/0, IAEA, Vienna, Austria, 1971.
169
-------
54. Wilsdon, B. H. and McConnell, F. J., J. Soc. Chem. Ind., Vol. 53,
pg. 385, 1934.
55. Huey, N. A., J. Air. Poll. Cont. Assn., Vol. 18, pg. 610, 1968.
56. Thomas, F. W. and Davidson, C. M., J. Air Poll. Cont. Ass., Vol. 11,
pg. 1, 1961.
57. Stalker, W. W., Dickerson, R. C., and et al; Amer. Ind. Hyg. Assn.,
J., Vol. 24, pg. 68, 1963.
58. Rickey, H. R. and Hendrickson, E. R., J. Air Poll. Cont. Assn.,
Vol. 15, pg. 409, 1965.
59. Harding, C. I. and Kelley, T. R., J. Air Poll. Cont. Assn., Vol. 17,
pg. 545, 1967.
60. Ikeda, T., J. Hyg. Chem., Vol. 13, pg. 41, 1967.
61. Stevens, R. K., Mulik, J. D., et al; Anal. Chem., Vol. 43, pg. 827,
1971.
62. Stevens, R. K., O'Keeffe, A. E., et al; Environ. Sci. Tech., Vol. 3,
pg. 652, 1969.
63. Ronkainen, P., Denslow, J., et al; J. Chromatog. Sci., Vol. 11, pg.
384, 1973.
64. Pecsar, R. E. and Hartmann, C. H., J. Chromatog. Sci., Vol. 11, pg.
492, 1973.
65. Stevens, R. K. and O'Keeffe, A. E., Anal. Chem., Vol. 42, pg. 143a,
1970.
66. Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070, 1972.
67. Bruner, F., Ciccioli, P., et al; Anal. Chem., Vol. 47, pg. 141, 1975.
68. DeSouza, T. L. C., Lane, D. C., et al; Anal. Chem., Vol. 47, pg.
543, 1975.
69. Bruner, F., Ciccioli, P., et al; J. Chromatog., Vol. 99, pg. 661,
1974.
70. Bremner, J. M. and Banwart, W. L., Sulfur Inst. Journal, Vol. 10,
pg. 6, 1974.
71. Mulik, J. D., Todd, G. , et al; "Ion Chromatography Determination of
Atmospheric Sulfur Dioxide," Symposium on Ion Chromatograph Analysis
of Environmental Pollutants, EPA, Research Triangle Park, North
Carolina, April, 1977.
170
-------
72. Dietzmann, H. E., "Protocol to Characterize Gaseous Emissions as a
Function of Fuel and Additive Composition," Environmental Protection
Technology Series under Contract 68-02-1275, September, 1975.
73. Dimitriades, B. and Seizenger, D. E., Environ. Sci. Tech., Vol. 5,
No. 3, pg. 223, 1971.
74. Papa, L. J., Dinsel, D. L., et al; J. Gas Chromatog., Vol. 6, pg.
270, 1968.
75. Klosterman, D. L. and Sigsby, J. E., Environ. Sci. Tech., Vol. 1,
pg. 309, 1967.
76. Communication between Mobile Source Emissions Research Branch, ESRL-
EPA, Research Triangle Park and Southwest Reseach Institute.
77. Bamesberger, W. L. and Adams, D. F., Environ. Sci. Tech., Vol. 3,
pg. 258, 1969.
78. Flatnm, D. L. and James, R. E., Environ. Sci. Tech., pg. 159, 1975.
79. Gustafsson, L., Talanta, Vol. 4, pg. 227, 1960.
80. Adams, D. F., Bamesberger, W. L., et al; J. Air Poll. Cont. Assn.,
Vol. 18, pg. 145, 1968.
81. Sekerka, I. and Lechner, J. F., Water Res., Vol. 10, pg. 479, 1976.
82. Valentour, J. C., Aggarwal, V., and Sunshine, I., Anal. Chem. Vol.
46, pg. 924, 1974.
83. Spencer, C. F., Baumann, F., et al; Anal. Chem., Vol. 30, No. 9,
pg.1473, 1958.
84. Sumner, S., Karrman, K. J., et al; Mikrochim Acta, pg. 1144, 1956.
85. Liberti, A. and Cartoni, G. P., Chim. e ind., Vol. 39, pg. 821,
1957.
86. Ryce, S. A. and Bryce, W. A., Anal. Chem., Vol. 29, pg. 925, 1957.
87. Desty, D. H. and Whyman, B. H. F., Anal. Chem., Vol. 29, pg. 230,
1957.
88. Coleman, H. J., Thompson, C. C. , et al; Anal. Chem., Vol. 30, pg.
1592, 1958.
89. Desty, D. H. and Harbourn, C. L. A., Div. of Analytical and Petro-
leum Chemistry, Symposium on Advances in Gas Chromatography, 132nd
meeting, ACS, New York, N. Y., September, 1957.
90. Amberg, C. H., Can. J. Chem., Vol. 36, pg. 590, 1958.
171
-------
91. Sullivan, J. H. , Walsh, J. T., et al; Anal. Chem., Vol. 31, pg.
1827, 1959.
92. Dal Nogare, S. and Bennett, C. E., Anal. Chem., Vol. 30, pg. 1157,
1958.
93. Harrison, G. F., Knight, P., et al; Second Int'l. Symposium on Gas
Chromatograph, Amsterdam, May, 1958.
94. Adams, D. F. and Koppe, R. K., Tappi, Vol. 42, pg. 601, 1959.
95. Schols, J. A., Anal. Chem., Vol. 33, pg. 359, 1961.
96. Hall, H. L. , Anal. Chem., Vol. 34, pg. 61, 1962.
97. Adams, D. F., Kippe, R. K., et al; Tappi, Vol. 43, pg 602, 1960.
98. Feldstein, M., Balestrieri, S., et al; J. Air Poll. Cont. Assn.,
Vol. 15, pg. 215, 1965.
99. Fredericks, E. M. and Harlow, G. A., Anal. Chem., Vol. 36, pg.
263, 1964.
100. Rummer, W. A., Pitts Jr., J. N., et al; Environ. Sci. Tech., Vol.
5, pg. 1045, 1971.
101. Mizany, A. I., J. Chromatog. Sci., Vol. 8, pg. 151, 1970.
102. Rasmussen, R. A., AM. Lab, Vol. 4, pg. 55, 1972.
103. Brody, S. S. and Chaney, J. E., J. Gas Chromatog., Vol. 4, pg. 42,
1966.
104. Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070,
1972.
105. Elliott, L. F. and Travis, T. A., Soil Sci. Soc. Amer. Proc., Vol.
37, pg. 700, 1973.
106 Francis, A. J., Adamson, J., et al; Bull. Ecol. Res. Comm., (Stockholm),
Vol. 17, pg. 485, 1973.
107. Lewis, J. A. and Papavizas, G. C., Soil Biol. Biochem., Vol. 2,
pg. 239, 1970.
108. Lovelock, J. E., Maggs, R. J., et al; Nature, London, Vol. 237,
pg. 452, 1972.
109. Ronkainen, P., Denslow, J., et al; J. Chromatog. Sci., Vol. 11,
pg. 384, 1973.
110. "Nickel Carbonyl," Amer. Ind. Hyg. Assn. J., pg. 304, 1968.
172
-------
111. Me Dowell, R. S., Amer. Ind. Hyg. Assn. J., Vol. 32, pg. 621, 1971.
112. Densham, A. B., Beale, P. A. A., et al; Journal Appi. Chem , Vol
13, pg. 576, 1963.
113. Kincaid, J. F., Stanley, E. L., et al; Amer. J. din. Path., Vol
26, pg. 207, 1956.
114. Hunold, G. A. and Pietrulla, W., Arbeitsschutz, Vol. 8, pg. 193,
1961.
115. Pitet, M. G., Arch. Mai Prof., Vol. 21, pg. 674, 1960.
116. Vol'berg, N. S., U.S.S.R., Patent No. 135,684, February, 1961.
117. Belyakov, A. A., Zavodsk Lab., Vol. 26, pg. 158, 1960.
118. Vol'berg, N. S. and Gerskhovich, E. E., Hyg. Sanit., Vol. 33, pg.
226, 1968.
119. Sunderman Jr., F. W., Roszel, N. O. , et al; Arch. Environ. Health,
Vol. 16, pg. 836, 1968.
120. Me Carley, J. E., Saltzman, R. S., et al; Anal. Chem., Vol. 28, pg.
880, 1956.
121. Ball, K. E., Bossart, C. J., et al; "Detection of Trace Amounts of
Nickel Carbonyl and Tetraethyl Lead in Air," Pittsburgh Conference
on Analytical Chemistry and Applied Spectroscopy, February, 1960.
122. Stedman, D. H., Tammaro, D. A., Analytical Letters, Vol. 9 (1),
pg. 81, 1976.
123. Stecher, P. G. (Ed.), The Merck Index, 8th Edition, Merck and Company,
Inc., Rahway, N. J., 1968, pg. 64.
124. Masterton, W. L. and Slowinski, E. J., Chemical Principles, 3rd
Edition, W. B. Saunders Company, Philadelphia, Pa., 1973.
125. Carter, R. V. , et al; Air Pollution Engineering Source Evaluation
of Ammonia Oxidation Plant Number 10, Holston Army Ammunition Plant,
Kingsport, Tennessee, Army Construction Engineering Research Lab.,
Champaign, 111., 1974, pg. 10.
126. Healy, T. V. and Pilbeam, A., Ammonia and Related Atmospheric
Pollutants at Harwell Part 1. Techniques for hourly sampling and
analysis of pollutants, Applied chemistry Division, Atomic Energy
Research Establishment, Harwell, Berkshire, 1972.
127. Kothny, E. L., Intersociety Committee, Health Lab. Sci., Vol. 10,
pg. 115, 1973.
173
-------
128. Okita, T. and Kanomori, S., Atmos. Environ., Vol. 5, pg. 622, 1971.
129. Kramer, D. N. and Sech, J. M., Anal. Chem., Vol. 44, pg. 395, 1972.
130. Zitomer, F. and Laumbert, J, L., Anal. Chem., Vol. 34, pg. 1738,
1962.
131. Blinn, R. C. and Gunther, F. A., Anal. Chem., Vol. 29, pg. 1882,
1957.
132. Williams, D. D. and Miller R. R., Anal. Chem., Vol. 34, pg. 225,
1962.
133. Gunther, F. A., Barkley, J. H., et al; Anal. Chem., Vol. 28, pg.
1985, 1956.
134. Kolbezen, M. J. , Eckert, J. W., et al; Anal. Chem., Vol. 36, pg.
593, 1964.
135. Pierson, R. H. , Fletcher, A. A., et al; Anal. Chem., Vol. 28, pg.
1218, 1956.
136. Burns, E. A., The Analysis of Exhaust Gases Trapped from Ablating
Nozzles, TRW Report No. 9840-6001-TU 000, Redondo Beach, Calif.,
1964.
137. Howell, J. H. and Boltz, D. F., Anal. Chem., Vol. 36, pg. 1799,
1964.
138. Burns, E. A., Unpublished results, 1963.
139. Jenkins Jr., R. W., Cheek, C. H., et al; Anal. Chem., Vol. 38,
pg. 38, pg. 1257, 1966.
140. Miettinen, J. K. and Virtanen, A. I., Ann. Acad. Sci. Fennicae,
Ser. A, II, No. 41, 1951, through Chem. Abst., Vol. 46, pg. 11500,
1952.
141. Arcand, G. M. and Swift, E. H., Anal. Chem., Vol. 28, pg. 440, 1956.
142. Krivis, A. F., Supp, G. R., et al; Anal. Chem., Vol. 35, pg. 2216,
1963.
143. Gilbert, T. R. and Clay, A. M., Anal. Chem., Vol. 45, pg. 1758, 1973.
144. Mertens, J., Van den Winkle, P., et al; Bull. Soc. Chem. Belg.,
Vol. 83/n° 1-2/, pg. 26, 1973.
145. Furman, N. H. (Ed.), Scott's Standard Methods of Chemical Analysis,
5th Edition, Van Nostrand, N. Y., 1958, Vol. 1, pg. 2336.
174
-------
146. Voge'l, A. I., Quantitative Inorganic Analysis. 3rd Edition T^ ™
Green, London, 1961, pg. 1092. - *— ^ Edltl°^ Longmans,
147. Steyermark, A., Quantitative Organic Microanalvsis. Blakestone, N. Y.
19 bJ./ pcj » -5o • * '
148. Sadek, F. S. and Reilly, C. N., Anal. Chem., Vol. 31, pg. 494, 1959.
149. Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 49, pg. 38, I960.
150. Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 52, pg. 105,
j.yb j •
151. Hodges, J. A., Bell, J. P., et al; Joint Conf. on Sensing Environ-
mental Pollutants Collected, Technical Papers, 1972, Vol. 1, 71-
1067.
152. Sigsby Jr., J. E., Black, F. M. , et al; Environ. Sci. Tech., Vol 7
' '
pg. 51, 1973.
175
-------
APPENDIX A
INTERIM REPORT
TASK I
176
-------
INTERIM REPORT
Task I
Selection of Analytical Methodology
This report is a summary of the results of the literature search and
communications with EPA and industry researchers. A literature search
for analytical methods was conducted on each compound (or group of simi-
lar compounds) and these results are presented in this report. A sum-
mary of the sampling and analysis procedures that are proposed for Task
II is presented in Table 1. The compounds under study are described and
the results of the literature search are detailed, in the following sec-
tions :
TABLE 1. PROPOSED SAMPLING AND ANALYSIS
METHODOLOGY FOR EMISSIONS CHARACTERIZATION
Proposed Analytical Methodology
Exhaust Species
HCN
C2N2
N2°
NH3
Organic Amines
DMNA
S02
H2S
COS
CH3SCH3
CH3SSCH3
Sampling
Nickel Carbonyl
H2S04
CVS Bag
CVS Bag
CVS Bag
Continuous into Impingers
Continuous into Impingers
Continuous onto Trap
Continuous into Impingers
Continuous into Impingers
Continuous onto Trap
Continuous onto Trap
Continuous onto Trap
Continuous onto Trap
Continuous into Impingers
Continuous onto Filter
Analysis
GC-NPD
GC-NPD
GC-ECD
Ion Chromatograph
GC-NPD
GC-CL w/MS Confirmation
Ion Chromatograph
Methylene Blue
GC-FPD
GC-FPD
GC-FPD
GC-FPD
GC-ECD
BCA
177
-------
Description Page
I. Hydrogen Cyanide 3
II. Cyanogen 6
III. Nitrous Oxide ' 8
IV. Ammonia 9
V. Organic Amines 14
VI. N-nitrosodimethylamine 18
VII. Sulfur Dioxide 24
VIII. Hydrogen Sulfide 31
IX. Organic Sulfides 35
X. Nickel Carbonyl 39
XI. Sulfuric Acid 42
The following sections cover a general description of the exhaust
species, results of the literature search, recommendations and a list
of references for those compounds included in the emissions inventory.
I. Hydrogen Cyanide
A. General Description
Hydrogen cyanide is a colorless, flammable liquid with a character-
istic odor of bitter almonds. The boiling point (1 atm) is 25.70°C and
the melting point (1 atm) is -13.24°C.^1) The density of the liquid at
20°C is 0.6871 g/ml and the specific gravity of the gas at 31°C is 0.947.
Hydrogen cyanide vapor is highly toxic and in 1968 the American
Conference of Governmental Industrial Hygienists recommended a thresh-
old limit value of 10 ppm. This is the concentration in air to which
nearly all workers may be exposed, day after day, without adverse ef-
fects.
B. Results of Literature Search
The analysis for HCN has been performed using three basic methods
of analysis, wet chemical, specific ion electrode and gas chromatography.
Prior to the advent of modern methods of instrumental analysis, a great
portion of analytical chemistry was performed using classical wet chemi-
cal techniques. This was also true for the analysis for HCN where an
analytical technique was developed for CN~.^)This technique was specific
to CN but not specific to HCN. A number of years later, the approach
presented by Epstein was applied to cigarette smoke and an automated
system was developed.^) several years later the same authors expanded
the technique to include aldehydes.^ '
178
-------
This technique is commonly referred to as the "Pyridine-Pyrazolone"
jethod and has been applied to automotive exhaust. Work has been per-
formed by EPA and its subcontractors in the characterization of automo-
tive exhaust for HCN. Although this procedure is generally considered
to be specific for HCN, it is not and this will be discussed in more de-
tail later in this section.
With the advancement of analytical technology, a technique utilizing
potentiometric determination of CN~ was developed. This technique is com-
inonly referred to as "Specific Ion Electrode" or "Ion Selective Electrode"
During the past several years a number of articles have been published
using the specific ion electrodes (SIE) . These SIE are not specific to
HCN but rather to CN~ and consequently most references are reporting CN~
rather than HCN. The SIE for measuring CN~ have been applied to water^5
a^r(12,13^n(3 biological fluids. ^14^ Direct measurement using SIE would be
extremely difficult in the concentration ranges that are anticipated. Pre
vious experience with the Metrohm siE indicate stability problems w.hen
working at continuous low levels.
This problem has been minimized by Burke"-^)by incorporating a stan-
dard addition technique to bring the concentrations to the working range
of the SIE. Although a satisfactory technique may be available that would
allow measurement of CN~, this may not be satisfactory for this research
effort. During the literature search for cyanogen (C2N2) it was learned
that CJfl, will hydrolize in an aqueous solution according to the equation
C2N2 + H2O -> HCN + HOCN
Since it is possible to form HCN from C2N,j in the presence of H^O ,
an alternative method would be desired. Existing HCN procedures require
collection in aqueous KOH or NaOH. This would assist in the artifact
formation of HCN if CoNp were present. An alternative approach would be
to develop a technique that would measure both HCN and C2N2. This will
be discussed later in this section.
The third approach to the analysis of HCN involves the use of a gas
chromatograph. There have been several articles that have been published
that are specific to the analysis of HCN. The original work was performed
using thermal conductivity detectors and involved the use of HCN in the
percent range. (16) Additional work was published by Isbell(17) regarding
the separation of HCN and C2N2, but again at levels in the percent range.
In 1975 Myerson and Chludzinski^ reported better sensitivity using a
gas chromatograph equipped with a flame ionization detector.
There are a number of important considerations that should be made
regarding the analysis of HCN using gas chromatography. Hydrogen cyanide
is reactive with any metal surfaces that it contacts, forming stable
cyanide complexes. This type of problem is not apparent when working with
high concentrations of HCN. Work is currently in Progress under EPA Con-
tract 68-03-2499 regarding the gas chromatograph analysis for HCN. Although
179
-------
this work has not been completed, a satisfactory separation of HCN
and C2N was obtained with a reasonable analysis time.
C. Recommendations for Analysis
Of the three techniques presented in the literature search, only
analysis by gas chromatography will provide an analysis that is spe-
cific for HCN. In view of the fact that cyanogen is of interest to this
research effort, it would be desirable to analyze these two species in
the same procedure. The separation of HCN and C2N2 at low levels using
gas chromatography (NPD) should be considered only a starting point. A
number of experiments regarding stability, reactivity, etc., will be
required to fully document this procedure. This technique involves di-
rect sample injection via a gas-tight glass syringe. In order to insure
sample integrity all gas chromatograph components that come in contact
with the sample should be constructed of glass or teflon. This includes
a glass-lined injector port, glass column, and glass-lined tubing in the
detector interface.
At this time it is uncertain as to the minimum detection limit of
the NPD, but it is expected to be less than 0.1 ppm on a direct injection.
During the validation of the analytical methodology this will be deter-
mined and reported. Calibration blends of HCN and C2N2 are currently
in hand to assist in any experiments that are necessary to validate the
methodology. It should be noted that although the gas chromatograph proce-
dure is specific to HCN a number of experiments will be required to in-
sure there are no interferences from other combustion products that are
known to be present in exhaust.
If the experiments designed to validate the analytical methodology
indicate that there are problems with this approach, then another method
will be recommended. If the gas chromatograph procedure is not satis-
factory and HCN and C2N2 can not be measured separately, then it is pro-
posed to combine the two and report them as CN~. However, rather than
using some of the aforementioned manpower intensive techniques , it is
proposed to use another gas chromatograph method- ' ' This involves the
collection in aqueous KOH followed by extraction with hexane and reaction
with Chloramine T to produce cyanogen chloride. An aliquot of the cyano-
gen chloride in hexane is injected into a gas chromatograph with an elec-
tron capture detector. This technique would provide adequate sensitivity
but would not separate HCN and C2N2 but rather combine the two and report
HCN and C2N2 as CN~.
D. References
1. Braker, W. and Mossman, A. L. , Matheson Gas Data Book, 5th Edition,
East Rutherford, N. J. , 1971, pg. 301.
2. Epstein, J. , Anal. Chem. , Vol. 19, pg. 272, 1947.
3. Collins, P. F. , Sarji, N. M. , and Williams, J. F. , Tobacco Sci.,
Vol. 14, pg. 12, 1970.
180
-------
4. Collins, P. F., Sarji, N. M. , and Williams, J. F. , Breitrage zur
Tabakforschung, Vol. 7, pg. 73, 1973.
5. Sekerka, I. and Lechner, J. F., Water Res., Vol. 10, pg. 479, 1975.
6. Ito, C., Sakamoto, H. , and Kashima, T., Kyoritsu Yakko Daigaku
Kenkyu Nempo, Vol. 18, pg. 29, 1973.
7. Frant, M. S., Ross Jr., J. W. , and Riseman, J. H. , Anal. Chem. ,
Vol. 44, pg. 2227, 1972.
8. Riseman, J. H., Am. Lab., Vol. 4, pg. 63, 1972.
9. Fleet, B. and von Storp, H. , Anal. Chem., Vol. 43, pg. 1575, 1971.
10. Fleet, B. and von Storp, H., Anal. Lett., Vol. 4, pg. 425, 1971
11. Conrad, F. J., Talanta, Vol. 18, pg. 952, 1971.
12. Vickroy, D. G. and Gaunt Jr., G. L., Tobacco Sci., Vol. 174, pg.
50, 1972.
13. Burke, James, "Ion Selective Electrode-Direct Technique," Ford Motor
Company procedure for measuring HCN in dilute Automotive Exhaust, 1976.
14. Llenado, R. A. and Rechnitz, G. A., Anal. Chem., Vol. 43, pg. 1437,
1971.
15. Nebergall, W. H. , Schmidt, F. C. , and Holtzclaw Jr., H. F., College
Chemistry with Qualitative Analysis, 5th Edition, D. H. Heath and
Co., Lexington, Mass., 1976, pg. 684.
16. Woolmington, K. G. , J. Appl. Chem., Vol. 11, pg. 114, 1961.
17. Isbell, R. E., Anal. Chem., Vol. 35, pg. 255, 1963.
18. Myerson, A. L. and Chludzinski Jr., J. J. , J. Chromatog. Sci., Vol.
13, pg. 554, 1975.
19. Valentour, J. C., Aggarwal. ,V., and Sunshine, I., Anal. Chem., Vol.
46, pg. 924, 1974.
II • Cyanogen
A. General Description^20^
Cyanogen is a colorless, flammable toxic gas at room temperature and
atmospheric pressure with a characteristic almond-type odor. Cyanogen is
a highly posionous gas, having a toxicity comparable to that of hydrogen
cyanide. The freezing point (1 atm) of cyanogen is -27.9°C and the boil-
ing point (1 atm) is -21.17°C. The chemical formula for cyanogen is C2N2
181
-------
and it has a molecular weight of 52.04. Synonyms for cyanogen include di-
cyan and oxalonitrile. The 1968 American Conference of Governmental
Industrial Hygienists has tentatively proposed a threshold limit value
of 10 ppm for cyanogen.
B. Results of Literature Search
Upon conducting the literature search for analytical methods for
measuring cyanogen, it became apparent that very little work has been
published in that area. The analysis of cyanogen has been performed
using gas chromatography. (17,21) jn both instances, the analysis was
carried out using laboratory blends of C2N2 with other gases. A thermal
conductivity detector was used and the analysis was performed with cyano-
gen concentrations in the percent range. Although these publications are
somewhat outdated, they will provide an initial starting point.
The past several years has seen a rapid development in gas chroma-
tograph specialty detectors. The majority of work that was performed
years ago was limited to thermal conductivity and flame ionization de-
tectors. Recently, a Nitrogen-Phosphorus Detector (NPD) was developed
by Perkin-Elmer and would be considered a prime candidate for performing
trace analysis on nitrogen compounds such as cyanogen.
Work is currently in progress on EPA Contract 68-03-2499 to in-
vestigate the use of gas chromatography as an analytical technique for
measuring HCN. ^22^ In this work, a cursory look at C-JS^ is also being
included. A Perkin-Elmer Model 3920B gas chromatograph equipped with
a NPD is used to measure EC'., and C2N2- Although this work is only in
the preliminary stages, it appears that it may be possible to determine
HCN and C2N~ in the same analysis.
Cyanogen will readily form HCN and HOCN when absorbed in an
aqueous solution so wet collection techniques will not be considered.
The only apparent alternative is to measure a direct gaseous sample if
C2N2 is to be measured separately. A number of experiments would be
required to validate the sample integrity if this approach is used.
C. Recommendations
The most promising technique for the analysis of C2N2 appears to
be the use of gas chromatography. A Perkin-Elmer 3920B with all glass
components is available for this analysis. This unit has sub-ambient
temperature capability as well as a NPD. Several preliminary experi-
ments have been conducted that indicate HCN and C2N2 can be separated
at reasonable temperatures. These experiments were conducted using nomi
nal HCN and C2N2 concentrations less than 5 ppm. Although there are a
substantial number of validation experiments that are necessary before
this technique would be acceptable, it appears to have the most promise.
Several experiments that are planned include interferences from
regulated emissions, sample stability, recovery, sample injection, etc.
All of these will be part of the methodology validation in Task II and
182
-------
reported in subsequent progress reports. During the course of the method-
ology development, if it appears that there are problems with the proposed
technique then an alternative method will be selected.
If cyanogen cannot be measured directly using gas chromatography then
it is proposed to combine the analysis of HCN and C2N2 and determine them
as CN~. A basic absorbing solution would be used to collect these exhaust
species during the emissions test. The basic solution is extracted with
hexane containing chloramine T. The CN~ reacts with the Chloramine T and
produces cyanogen chloride. The cyanogen chloride is then quantitatively
determined U|jn0 a <3as chromatograph equipped with an electron capture de-
tector (BCD).
D. References
20. Braker, W. and Mossman, A. L. , Matheson Gas Data Book, 5th Edition,
East Rutherford, N. J., 1971, pg. 155.
21. Runge, H., Z. Anal. Chem., Vol. 189, pg. Ill, 1962.
22. Valentour, J. C. , Aggarwal, V., and Sunshine, I., Anal. Chem., Vol.
46, pg. 924, 1974.
III. Nitrous Oxide
A. General Description^
Nitrous oxide is a colorless, nonflammable gas with a slightly
sweetish taste and odor. Nitrous oxide is nontoxic and nonirritating
and is extensively used as an anesthetic in medicine and dentistry.
The freezing point (1 atm) of nitrous oxide is -90.84°C and the boil-
ing point (1 atm) is -89.5°C. The chemical formula for nitrous oxide
is N20 and the molecular weight is 44.013. Common synonyms for nitrous
oxide are dinitrogen monoxide and laughing gas.
B. Results of Literature Search
The measurement of nitrous oxide has been performed by several
researchers using gas chromatography. (24-28a) Most of this data was
published during the early development of gas chromatography and con-
sequently was not concerned with trace analysis. This work was con-
ducted using a thermal conductivity detector and the minimum detection
limits were quite high at that time. The nontoxic nature of the gas
apparently precluded the development of analytical techniques for trace
quantities. During the past several months, work has been performed at
EPA-RTP in the development of a method to measure N2O directly using a
gas chromatograph equipped with an electron capture detector.(29' This
analytical procedure has been applied to dilute automotive exhaust and
Provides the necessary sensitivity for this research effort. La Hue, et al >,
described a method for the measurement of atmospheric N20 in the 250-300
PPb range. A concentration technique involving the use of molecular
5A trap is described. Upstream of the trap, water vapor and carbon
183
-------
dioxide are removed by passage of the air over anhydrous calcium sulfate
and ascarite. The ^0 is desorbed using a water saturated helium stream.
The water-displaced N2O is then trapped on a silica gel column at -76°C.
The silica gel column is heated to transfer the N2O into a gas chromato-
graph for analysis.
C. Recommendations
It is recommended that the analysis of ^0 be conducted using a gas
chromatograph equipped with an ECD. A Perkin-Elmer Model 3920B gas chroma-
tograph with an ECD is available for this analysis. The GC operating con-
ditions have been obtained from the Project Officer and a system for the
measurement of ^O has been assembled. The calibration of the instrument
will be performed using ^0 in N^ calibration blends. A Hewlett-Packard
3354 is available for data acquisition as needed. Since this method has
been successfully applied to automotive exhaust, it is not anticipated
that a significant effort will be required in the procedural development
but rather to the validation of the entire system.
D. References
23. Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition,
East Rutherford, N. J., 1971, pg. 431.
24. Jay, B. and Wilson, R., J. Appl. Phys., Vol. 15, No. 2, pg. 298, 1960.
25. De Grazio, R., J. Gas Chromatog., Vol. 3, pg. 204, 1965.
26. F & M Applications Chromatogram, 1942, F & M Scientific Division of
Hewlett-Packard Corp., Avondale, Pa.
27. Porapak © Brochure, Waters Associates, Framingham, Mass.
28. Bethea, R., and Adams, F., J. Chromatog., Vol. 10, pg. 1, 1963.
28a. Leithe, W., The Analysis of Air Pollutants, Humphrey Sci.
Pub., Ann Arbor, Michigan, 1970, pg. 176.
29. Private communication between Dr. R. B. Zweidinger and Frank Black.
30. La Hue, M. D., Axelrod, H. D., and Lodge Jr., J. P., Anal. Chem.,
Vol. 43, pg. 1113, 1974.
IV. Ammonia
A. General Description' '
Ammonia is a colorless, alkaline gas having a pungent odor. Ammonia
readily dissolves in water. The freezing point (1 atm) of ammonia is
-77.7°C and the boiling point (1 atm) is -33.35°C. The chemical formula
for ammonia is NH3 and the molecular weight is 17.031. The 1968 American
Conference of Governmental Industrial Hygienists has recommended a threshold
184
-------
limit value of 50 ppm. This is the concentration in air to which nearly
all workers may be repeatedly exposed, day after day, without adverse effects,
B. Results of Literature Search
The measurement of ammonia has been accomplished by a wide variety
of techniques. Ammonia has been determined by such classical titrimetry
methods as acidimetrity(32~41) , complexometry(42), oxidimetry(43'44>, and
formal titration. (45~47) Although these methods may be satisfactory for
specific applications, there are a number of factors which preclude their
consideration as a method for this research effort. In general, titrimetry
is limited to 10~4 M solutions and is reported to have a substantial
number of interferences.
Gravimetric methods have been successfully used for the measurement
of ammonia in several cases- v4b-bl) Generally speaking, if measurement
of trace quantities of given species is required, gravimetric techniques
are not adequately sensitive. The limited sampling time would eliminate
any further consideration of gravimetric techniques for the measurement
of ammonia. Gasometric methods have been used in determining high con
centrations of ammonia. (52-54)
A number of instrumental optical methods have been applied to the
measurement of ammonia. These methods include Nessler Reagent'55-57) f
Pyridine-Pyrazolone^58'59' , Indophenol (60,61) f chloramine ' , Cupri-
Ammonia complex(42) , Ninhydrin' , and Indirect Colorimetric Methods. (64)
The direct spectrophotometric measurement of ammonia in air has been ac-
complished using ultraviolet^65'66\ and infrared. (67f68' The use of
electroanalytical methods has been applied to the measurement of ammonia
in simple systems. The electrochemical techniques include polarography
amperometry^72'73) , coulometry (74~79) , and conductivity^80' Both gas(81'82)
and paper(°3) chromatography have been applied to the analysis of ammonia.
The GC analysis ^53^ achieved a clean separation between ammonia and the
lower molecular weight amines.
More recent advances in the measurement of NH3 in trace quantities
have involved specific ion electrode techniques. These techniques have
included a wide variety of applications. (83~91) Although the specific
ion electrode technique has distinct advantages over other methods, it
does have certain limitations, namely instability at the low concentra-
tions. With that in mind, consideration was given to several other al-
ternative methods.
In 1973, Sigsby,et at, published results of work that employed the
use of a chemiluminescent detector to continuously measure NO, NC>2 and
NH3 in dilute automotive exhaust - (92) It may be possible to modify this
approach to remove the acid gases (NO and NO2) and measure the NH3 con-
tinuously. Ammonia has been quantitatively measured in dilute automotive
exhaust(93) using an ion chromatograph. This procedure is free from
many Of the common interferences that plague the classical methods. The
analysis time required for this analysis is less than 10 minutes and makes
xt a prime candidate for ammonia analysis.
185
-------
C. Recommendation
The most promising technique for the measurement of trace quantities
of ammonia in dilute automotive exhaust is by use of the ion chromatograph.
This method involves the collection of ammonia in impingers using dilute
H2S04 as the absorbing reagent. Once the sample has been collected, an ali-
quot is injected into the ion chromatograph and the peak area is proportional
to the concentration. This measurement method is the most straightforward
of the techniques presented in the literature and probably gives the best
results.
An alternative to this technique would be a modification of the system
proposed by Sigsby, et al. An acid scrubber could be incorporated to re-
move the NO and NO2 and allow measurement of NH3 only. With the appropri-
ate modifications, this could be incorporated into a gas chromatograph
and a single bag sample could be used in lieu of continuous sampling.
Both of these approaches would require a substantial amount of effort to
validate the analytical methodology.
D. References
31. Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition,
East Rutherford, N. J., 1971, pg. 17.
32. Furman, N. H., (Ed.), Scott's Standard Methods of Chemical Analysis,
5th Edition, Van Nostrand, N.Y., 1958, Vol. 1, pg. 636.
33. Pifer, C. W. and Wollish, E. G. , Anal. Chem. , Vol. 24, pg. 519, 1952.
34. Kolthoff, I. M. and Stenger, V. A., Volumetric Analysis, inter-
science, N. Y. , 1947, Vol. 2, pg. 125.
35. Furman, N. H. , (Ed.), Scott's Standard Methods of Chemical Analysis,
5th Edition, Van Nostrand, N.Y., 1958, Vol. 1, pg. 632.
36. Pierce, W. C., Haenisch, E. L., and Sawyer, D. T. , Quantitative
Analysis, Wiley, N. Y., 1958, pg. 260.
37. Kolthoff, I. M. and Stenger, V. A., Volumetric Analysis, Interscience,
N.Y., 1954, Vol. 2, pg. 168.
38. Milner, 0. I. and Zahner, R. J., Anal. Chem., Vol. 32, pg. 294, 1960.
39. Conway, E. J., Micro-Diffusion Analysis and Volumetric Error, 2nd
Edition, Crosby-Lockwood , London, 1947.
40. Altmann, C. J. G. , de Heer, B. H. J., and Hermans, M. E. A., Anal.
Chem., Vol. 35, pg. 596, 1963.
41. Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 254.
186
-------
42. Blinn, R. C. and Gunther, F. A., Anal. Chem. , Vol. 29, pg 1882
1957.
43. Kolthoff, I. M. and Stenger, V. A., Ind. Eng. Chem., Anal. Ed
Vol. 1, pg- 79, 1935.
44. Kolthoff, I. M. and Belcher, R. , Volumetric Analysis.
N-.Y., 1957, Vol. 3, pg. 582. ~~
45. Welcher, F. J., Organic Analytical Reagents, Van Nostrand, N.Y.,
1947, Vol. 1, pg. 376."
46. Rowe, D. J., Gas Journal, Vol. 265, pg. 49, 1951.
47. Stockdale, D., Analyst, Vol. 84, pg. 667, 1959.
48. Sadek, F. S. and Reilley, C. N. , Anal. Chem., Vol. 31, pg. 494, 1959.
49. Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 49, pg. 38, 1960.
50. Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 52, pg. 105, 1963.
51. Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 566.
52. Furman, N. H.,(Ed.)* Scott's Standard Methods of Chemical Analysis.
5th Edition, Van Nostrand, N.Y., 1958, Vol. 1, pg. 2336.
53. Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 1092.
54. Steyermark, A., Quantitative Organic Microanalysis, Blakestone, N.Y.,
1951, pg. 56.
55. Taras, M. J., in Colormetric Determination of Nonmetals, Boltz, D. F.,
(Ed.), Interscience, N.Y., 1958, pg. 84.
56. Miller, G. L. and Miller, E. E., Anal. Chem., Vol. 20, pg. 481, 1948.
57. Thompson, J. F. and Morrison, G. R. , Anal. Chem., Vol. 23, pg. 1153,
1951.
58. Kruse, J. M. and Mellon, M. G., Anal. Chem., Vol. 25, pg. 1188, 1953.
59. Prochazkova, L., Anal. Chem., Vol. 36, pg. 865, 1964.
60. Scheurer, P. G. and Smith, F. , Anal. Chem., Vol. 27, pg. 1616, 1955.
61 • Bolleter, W. T. , Bushman, C. J. , and Tidwell, P. W. , Anal. Chem.,
Vol. 33, pg. 592, 1961.
62. Zitomer, F. and Lambert, J. L. , Anal. Chem., Vol. 34, pg. 1738, 1962.
187
-------
63. Williams, D. D. and Miller, R. R., Anal. Chem., Vol. 34, pg. 225,
1962.
64. Howell, J. H. and Boltz, D. F., Anal. Chem., Vol. 36, pg. 1799, 1964.
65. Gunther, F. A., Barkley, J. H., Kolbezen, M. J., Blinn, R. C., and
Staggs, E. A., Anal. Chem., Vol. 28, pg. 1985, 1956.
66. Kolbezen, M. J., Eckert, J. W., and Wilson, C. W., Anal. Chem.,
Vol. 36, pg. 593, 1964.
67. Pierson; R. H., Fletcher, A. A., and St. Clair Gantz, E., Anal.
Chem., Vol. 28, pg. 1218, 1956.
68. Burns, E. A., The Analysis of Exhaust Gases Trapped from Ablating
Nozzles, TRW Report No. 9840-6001-TU 000, Redondo Beach, Cal., 1964.
69. Nyman, C. J. and Johnson, R. A., Anal. Chem., Vol. 29, pg. 483, 1957.
70. Norton, 0. R. and Mann, C. K., Anal. Chem., Vol. 26, pg. 1180, 1954.
71. Clear, A. J. and Roth, M., in Treatise^ on Analytical Chemistry, Kol-
thoff, I. M. and Elving, P. J. (Eds.), Interscience, N.Y., 1961,
Part II, Vol. 5, pg. 284.
72. Kolthoff, I. M., Stricks, W., and Morren, L., Analyst, Vol. 78,
pg. 405, 1953.
73. Laitinen, H. A. and Woerner, D. E., Anal. Chem., Vol. 27, pg. 215,
1955.
74. Arcand, G. M. and Swift/ E. H., Anal. Chem., Vol. 28, pg. 440, 1956.
75. Krivis, A. F., Supp, G. R. , and Gazda, E. S., Anal. Chem., Vol. 35,
pg. 2216, 1963.
76. Christian, G. D., Knoblock, E. C., Purdy, W. C. , Anal. Chem., Vol.35,
pg. 2217, 1963.
77. De Ford, D. D., Johns, C. J., and Pitts, J. N., Anal. Chem., Vol. 23,
pg. 941, 1951.
78. Segal, N. S. and Wodley-Smith, R., Anal. Chem., Vol. 38, pg. 829,
1966.
79. Sambucetti, C. J., Anal. Chem., Vol. 38, pg. 105, 1966.
80. Barendrecht, E. and Janssen, N. G. L. M., Anal. Chem., Vol. 33,
pg. 199, 1961.
81. Burns, E. A., Unpublished results, 1963.
188
-------
82. Jenkins Jr., R. W., Cheek, C. H., and Linnenbom, V. S., Anal chem
Vol. 38, pg. 1257, 1966.
83. Miettinen, J. K. and Virtanen, A. I., Ann. Acad. Sci. Fennicae,
ser. A., II, No. 41, 1951; through Chem. Abst. Vol. 46, pg 11500
1952.
84. Gilbert, T. R. and Clay, A. M. , Anal. Chem., Vol. 45, pg. 1757, 1973.
85. Le Blanc, P. J. and Sliwinski, J. F., Am. Lab., Vol. 5, pg. 51,
1973.
86. Thomas, R. F. and Booth, R. L. , Environ. Sci. Tech., Vol. 7, pg. 523,
1973.
87. Woodis Jr., T. C. and Cummings Jr., J. M. , JAOAC, Vol. 56, pg. 373,
1973.
88. Vandevenne, L. and Oudewater, J. , Centre Beige d'Etude et de Documen-
tation des Eaux, Vol. 352, pg. 127, 1973.
89. Banwart, W. L. , Tabatabai, M. A., and Bremner, J. M. , Comm. Soil Sci.
Plant Anal.,Vol. 3, pg. 449, 1972.
90. Byrne, E. and Power, T., Comm. Soil Sci. Plant Anal., Vol. 5, pg. 51,
1974.
91. Me William, D. J. and Ough, C. S., Amer. J. Enol. Viticult, Vol. 25,
pg. 67, 1974.
92. Sigsby Jr., J. E., Black, F. M. , Bellar, T. A., and Klosterman, D. L.,
Environ.,Sci. Tech., Vol. 7, pg. 51, 1973.
93. Zweidinger, R. B., Tejada, S. B., Sigsby Jr., J. B., and Bradow,
R. L., "The Application of Ion Chromatography to the Analysis of
Ammonia and Alkyl Amines in Automotive Exhaust," Symposium on Ion
Chromatographic Analysis of Environmental Pollutants, EPA, Research
Triangle Park, N.C., April, 1977.
v- Organic Amines
A. General Description(94,95)
The individual amines that are included in this analysis are mono-
raethylamine, dimethylamine, monoethylamine, trimethylamine, diethylamine
and triethylamine. The chemical formulas, molecular weights, boiling
Points, freezing points and synonyms for these low molecular weight ali-
phatic amines are presented in Table 2. In general, these amines have a
ish-type odor at lower concentrations but more of an ammoniacal odor at
higher levels. The 1968 American Conference of Governmental Industrial
Hygienists has recommended a threshold limit value of 10 ppm.
189
-------
TABLE 2. LIST OF INDIVIDUAL ORGANIC AMINES
INCLUDED IN THE EMISSIONS CHARACTERIZATION INVENTORY
Name
Monome thy lamine (94)
Monoe thy lamine (94)
Dime thy lamine (94)
Trime thy lamine (95)
Diethylamine (95)
Triethylamine (95)
Carbon
No.
1
2
2
3
4
6
Chemical Molecular
Formula Weight
CH NH 31.058
•J £
C H NH 45.085
£1 O £
(CH ) NH 45.085
(CH ) N 59.112
(C2H5)2NH 73'14
(CH, )N 101.19
Boiling
Point, °C
-6.32
16.58
6.88
2.87
56.3
89.3
Freezing
Point, °C
-93.5
-81.0
-92.19
-117.08
-50.
-114.7
Synonyms
Methylamine ,
aminome thane
Ethylamine ,
aminoe thane
none
none
none
none
-------
B. Results of Literature Search
The measurement of individual low molecular weight amines has been
:onducted using a variety of gas chroma tograph techniques. Hoshika*96'97)
reported gas chromatographic separation of lower aliphatic amines in the
Iree form and as their Schiff base derivatives. A glass column was
anployed to provide a separation of 11 amines using temperature program-
ing and a thermal conductivity detector. This work was oriented toward
ichieving a satisfactory separation rather than being concerned with
linimum detection limits. Sze et al reported separation of methyl
imines, ammonia, and methanol using a mixture of tetrahydroxyethylethy-
.inediamine and tetraethylenepentamine . This work was expanded to include
i total of 10 aliphatic amines. O'Donnell and Mann^") used Dowfax 9N9,
larbowax 400 and Carbowax 20M to separate mixtures of aliphatic amines,
iromatic amines and aliphatic amides. This work was performed using
synthetic blends on a gas chromatograph with a thermal conductivity detector.
fcCurdy and Meiser^100) used a gas chromatograph with a flame ionization
ietector to determine fatty amines in trace quantities. The fatty amines
;ere converted to trifluoroacetyl derivatives providing a sensitivity of
).05 ppm fatty amine in water.
Smith and Waddington^ ' used aromatic polymer beads to separate a
tide range of aliphatic amines. Peak tailing was found to exist because
)f two types of active sites on the polymer: simple acidic sites which
:an be neutralized by treatment with base, and metal ions which must be
leactivated by addition of an involatile complexing agent. Glass columns
rere used in a gas chromatograph with a flame ionization detector. Syn-
:hetic blends ranging from CI~CQ were separated and analyzed using this
ipproach. Carbopak B/4% Carbowax 20M/0.8% KOH^102) and 28% Pennwalt 223/4%
(OH (103) have been reported to give satisfactory separations of lower
aliphatic amines.
Analysis of amines as derivatives has been shown to be a valuable
malytical tool to determine trace quantities. ^104^ Thirteen different
lerivatives were evaluated in terms of FID and BCD response characteristics.
Chis work was limited to primary amines and under optimum conditions amines
town to 10 picograms could easily be quantified using an BCD detector.
-lark and Wilk^ ^ used an BCD to evaluate the properties of halogenated
derivatives. No increase in the sensitivity for the trifluoroacetyl
derivatives using ECD was observed.
Hosier (106) et al, quantitatively measured aliphatic amines volatilized
from cattle feedyards. Direct gas chromatograph injection of the acid
solutions and GC separation of the pentafluorobenzoyl derivatives of the
malodorous volatiles were used in identification. The derivatized amines
"ere analyzed using a gas chromatograph equipped with an electron capture
detector. The two columns used in this work were 6% DECS and 10% Igepal
CO 880.
Methylamine and ethylamine were detected in irradiated beef by Burks ,
et al. Several techniques including colorimetric paper chromatography and
9as chromatography were used in quantifying results. Gas chromatographic
191
-------
determination of free mono-, di-, and trimethylamines in biological fluids
was performed by Dunn(108), et al. A flame ionization detector was used to
quantitatively separate the lower aliphatic amines. Separation of mono-, di-,
and trimethylamine from extracts of fish tissue was achieved by Gruger.(109)
Andre (H°) et al developed a precolumn inlet system for the gas
chromatographic analysis of trace quantities of short-chain aliphatic
amines. Losses inherent in the collection and direct gas chromatograph
analysis of field air samples containing volatile amines necessitated an
indirect analytical scheme. A teflon tube (3" x 5/16" OD) was filled
with 20/30 mesh Ascarite and placed in the injector inlet of the gas
chromatograph. Samples were collected in dilute H2SC>4 and aliquots were
injected into the pre-column of the GC. Release of the free amines was
found to be sufficiently reproducible for quantification of results.
This technique avoided the problems encountered by Umbreitt111) et al,
and Hardy(H2) when using base loaded columns to analyze acidified aqueous
solutions of amines from fish. The in situ release of the free amines
from their salts produced a chromatographic column that changed with every
injection. In addition, the column had a very short usable lifetime and
lacked reproducibility after extended use.
Bowen described a gas chromatograph procedure for the analysis
of aromatic amines using an adsorption technique. Quantitative adsorption
and desorption of aromatic amines using Tenax GC was demonstrated at the
nanogram level. Samples were pulled through the Tenax GC trap for specific
sampling periods, thermally desorbed at 250°C and analyzed in a GC with a
FID. The author recommended use of a NPD to increase sensitivity for
aromatic amines on the tail of hydrocarbon solvents and eliminate venting
the solvent.
C. Recommendations
The analysis for individual aliphatic amines should be conducted
using gas chromatography. Of the specialty detectors available for the
analysis of nitrogen compounds the NPD is a prime candidate as a detection
method. Sample collection in dilute H2SO4 into an all glass system with
an ascarite pre-column would provide satisfactory sample acquisition and
analysis in a reasonable analysis time. A Perkin-Elmer 3920B gas chromato-
graph equipped with a NPD and an all glass system is available for this
analysis. Subambient temperature capability is available if necessary
and a Hewlett-Packard 3354 computer system will be used for data acquisition.
An alternative approach would be to convert the amines to fluoro-
containing derivatives and measure the derivatives using a GC with an BCD.
Sample collection would be made by using impingers, as with the aforemen-
tioned approach. A third approach would be to use direct gas injection
from a bag sample collected during a driving cycle. It is uncertain as to
the feasibility of using a Tenax GC trap as a collection technique for low
molecular weight aliphatic amines. The low freezing point of these com-
pounds may be too low to quantitatively concentrate on the Tenax GC traps.
192
-------
D. References
94. Braker, W. and Mossman, A. L., Ma the sou Gas Daba Book, 5th Edition
East Rutherford, N.J., 1971, pg. 385.
95. Weast, R. C., (Ed.), Handbook of Chemistry and Physics, 54th Edition,
The Chemical Rubber Co., Cleveland, Ohio, 1973.
96. Hoshika, Y., Anal. Chem., Vol. 48, pg. 1716, 1976.
97. Hoshika, Y., J- Chromatog, Vol.115, pg. 596, 1975.
98. Sze, Y. L. and Borke, M. L. , Anal. Chem., Vol. 35, pg. 240, 1963.
99. O'Donnell J. F. and Mann, C. K., Anal. Chem., Vol. 36, pg. 2097,
1964.
100. Me Curdy Jr., W. H. and Reiser, R. W. , Anal. Chem., Vol. 38, pg. 795,
1966.
101. Smith, J. R. L. and Waddington, D. J., Anal. Chem., Vol. 40, pg. 522,
1977.
102. "Amine Analysis," Bulletin 737A, Supelco, Co., Belefonte, Pa. , 1973.
103. "Penwalt 223 Amine Packing," Alltech Associates, Arlington Heights,
111.
104. Moffat, A. C. and Horning, E. C. , Anal. Lett., Vol. 3, pg. 205, 1970.
105. Clarke, D. D. , wilk, S., and Gitlow, S. E. , J. Gas Chromatog.,
pg. 310, 1966
106. Mosier, A. R. , Andre, C. E. , and Viets Jr., F. G., Environ. Sci.
Tech., Vol. 7, pg. 642, 1973.
107. Burks Jr., R. E. , Baker, E. B. , Clark, P., Esslinger, J., Lacey Jr.,
J. C., J. Agr. Food Chem., Vol. 7, pg. 778, 19M.
108. Dunn, &. R. , Simenhoff, M. L., Wesson Jr., L. G. , Anal. Chem.,
Vol. 48, pg. 41, 1976.
109. Gruger Jr., E. H. , J. Agr. Food Chem., Vol. 20, pg. 781, 1972.
110. Andre, C. E. and Mosier, A. R. f Anal. Chem., Vol. 45, pg. 1971, 1973.
111. Umbreit, G. R. , Nygren, R. E. , and Testa, A. J., J. Chromatog.,
Vol. 43, pg. 25, 1969.
112. Keay, J. N. and Hardy, R. , J. Sci. Food Agr., Vol. 23, pg. 9,
193
-------
113. Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.
VI. N-nitrosodimethylamine( '
A. General Description
N-nitrosodimethylamine is a yellow liquid at room temperature boiling
at 152°C. It has a chemical formula of (CH3)2NNO and a molecular weight
of 74.08. It is very soluble in water, alcohol and ether. N-nitrosodi-
methylamine has caused fatal poisoning, severe liver injury and is
suspected of being a carcinogen. Common synonyms include dimethyl-
nitrosoamine, dimethyl N-nitrosoamine and nitrous dimethylamide.
B. Results of Literature Search
In 1956, Magee and Barnes(116) reported that dimethylnitrosoamine
(DMNA) caused malignant tumors in rats. Lijinsky and Epstein(117>
identified N-nitroso compounds as a major class of carcinogens that may
be related to human cancer. Magee and Barnes' ' reported that 75 of
100 N-nitroso compounds tested were shown to be carcinogenic. Koppang
and Rimslatter (H9) presented results documenting the toxic and car-
cinogenic effects of DMNA on rodents. In 1971, Magee(12°) presented
additional data to substantiate the carcinogenic effects of N-nitroso
compounds. The potency of DMNA as a carcinogen has been repeatedly
demonstrated in experimental studies on mice, hamsters , guineas pigs,
rats, rabbits and several species of fish. (121)
Standard analytical methods for N-nitroso compounds were originally
based on gas chromatography but were available for only volatile com-
pounds. (122-129) These procedures were lengthy and time consuming and
required an extensive sample clean up prior to sample concentration and
analysis by gas chromatography. Potential interferences must be removed
when using the nitrogen specific alkali flame ionization and the Coulson
electrolytic conductivity detectors. Sen(13°) and Althorped31), et al
converted the nitrosamines to a nitramine and used a gas chromatograph with
an electron capture for quantitative measurement. Allison^ used a
GC-ECD system to measure polyfluorinated amide derivatives of various
nitrosamines. Most GC methods relied on steam volatilization or vacuum
distillation as the primary purification and concentration technique.
Subsequently, all available data was limited to volatile nitrosamines.
In 1974, Fine(133) et al introduced a specific method of detection
for a variety of nitrosamines. This instrument is commercially available
from Thermo Electron Corporation and is called Thermal Energy Analysis
(TEA). The TEA uses a selective catalyst to cleave N-nitroso compounds,
splitting off the nitrosyl radical. The nitrosyl radical passes through
a cold trap (-150°C) to the reaction chamber where it reacts with ozone,
giving nitrogen dioxide in the excited state. The light emitted from
the excited state nitrogen dioxide returning to the ground state is
measured in the near infrared region of the spectrum. Although some
data is available on interfering species, a substantial effort would be
necessary before satisfactory application of the TEA to automotive ex-
haust could be considered.
194
-------
The TEA is sensitive to picogram quantities with a linear range ex-
tending over five orders of magnitude.(134) ip^e TEA ^as been interfaced
to both a gas chromatograph^135^ and more recently, a high pressure liquid
chromatograph.(136, 137) although the TEA is extremely sensitive and
selective, sample cleanup procedures are necessary.(138) once the instru-
ment had been sufficiently developed, the TEA was applied to a wide variety
of sources. <139-148b)
Pellizzarihas applied a gas chromatograph-mass spectrometer system
to identify DMNA in ambient air. d49' This technique concentrates the
organic vapors in glass cartridges filled with specially preconditioned
Tenax GC. ^ ~ DMNA recovery was accomplished by thermal desorption
and analyzed using GC-MS using capillary glass SCOT columns. Identifi-
cation of DMNA was confirmed by comparison of its mass spectrum with that
of pure N-nitrosodimethylamine. The retention times of the unknown and
pure compound were identical on three different columns. The advantage
of this approach is the identification using GC-MS; whereas the TEA does
not provide absolute identification.
In 1975, a comparison of the two techniques was possible by comparing
the results of ambient air samples in Baltimore, Md.'•*-56^ Parallel analyses
were conducted using different collection and analysis procedures. The
samples were collected over a three month period and involved a number of
industrial locations. The TEA analysis was performed by a team of researchers
from Thermo Electron Corporation. Samples were collected cryogenically and
analyzed on a TEA Model 502. The second set of researchers was headed by
Pellizzari of Research Triangle Institute. Their samples were collected on
Tenax GC cartridges, thermally desorbed and resolved by capillary gas chroma-
tography. The mass cracking patterns were continuously and automatically
obtained throughout the GC run with a Varian-MAT CH-7 GC-MS computer system.
Identification of DMNA was achieved by comparing the mass cracking pattern
to an eight major peak index d", 158)^ with supplementary confirmation by
comparing with known DMNA retention times on two different capillary columns.
At the Food Manufacturing Corporation (FMC) location, additional samples
were collected by FMC researchers.(159) These independent cold trap GC-MS
experiments conducted by the FMC analytical team also confirmed the pre-
sence of DMNA in the air on the FMC property. At the duPont location,
independent cold trap GC-MS experiments were carried out by a duPont
analytical team, and the results agreed within an order of magnitude.' '
Several experiments were conducted to validate that DMNA was not formed
either during collection or analysis. For the cryogenic procedure,
DMNA was not detected when dimethylamine and NOx(NaN02) was added to the
trap. In the field work, an increase in DMNA above background levels was
not observed, even with a thousandfold excess of dimethylamine. The pos-
sibility of DMNA formation on the Tenax cartridges was investigated both
in the laboratory and in the field. Enhancement of DMNA was not observed
above the background when dimethylamine was preloaded into the Tenax cart-
ridge. No DMNA was detected when moist air containing NO+NOX (1:1) was
drawn across a permeation tube containing dimethylamine and into the Tenax
cartridge.
195
-------
Since both approaches have demonstrated that the DMNA reported was
not as a result of artifact formation, it is of interest to compare the
results on parallel samples.(156) The agreement between the two techniques
should only be considered fair. At the high levels (32,000 Ug/m3) the agree-
ment was generally within a factor of three. At the lower levels (500 ]Jg/m3),
the agreement between the two techniques deteriorated on several samples
and was only fair on the remaining samples.
C. Recommendations
It is recommended to use an analytical system, such as the TEA or
an equivalent system for the measurement of DMNA in exhaust. In addition,
validation of the sampling and analysis methodology would be performed
with supplemental GC-MS analysis. It is anticipated that the GC-MS
analysis would be conducted by Dr. Edo Pellizzari of Research Triangle
Institute during Task II. If it is not feasible to obtain or develop a
system similar to that developed by Dr. David Fines, the alternative
would be to collect the exhaust samples on Tenax GC and send them to RTI
for DMNA analysis by GC-MS.
D. References
114. Weast, R. C. (Ed.), Handbook of Chemistry and Physics, 54th Edition,
Division of the Chemical Rubber Co., 1973, pg. C-101.
115. Stecher, P. G. , (Ed.), The Merck Index, An Encyclopedia of Chemicals
and Drugs, 8th Edition, Merck and Co., Inc., Rahway, N.J., 1968,
pg. 742.
116. Magee, P. N. and Barnes, J. M., Brit, J. Cancer, Vol. 10, pg. 114,
1956.
117. Lijinsky, W. and Epstein, S. S., Nature, Vol. 225, pg. 21, 1970
118. Magee, P. N. and Barnes, J* M., Adv. Cancer Res.. Vol. 10, M67.
119. Koppang, N. and Rimslatten, H., The toxic and carcinogenic effect of
dimethylnitrosamine in mink. Fourth meeting of the Analysis and
Formation of N-Nitroso Compounds, It'l. Agency for Research on Can-
cer, Lyon, France, Meeting held in Tallinn, U.S.S.R., October, 1975.
120. Magee, P. N., Food Cosmet. Toxicol., Vol. 9, pg. 207, 1971.
121. Christensen, H. E. and Luginbyhl, T. T., (Eds.), Suspected carcino-
gens, Subfile of NIOSH Toxic Substances List. U.S. Dept. of Health,
Education, and Welfare, Public Health Service, Center for disease
control. NIOSH, Rockville, Md., June 1975.
196
-------
122. Telling, G. M. , Bryce, T. A., and Althorpe, J., J. Agr. Food Chem.,
Vol. 19, pg. 937, 1971.
123. Gough, T. A. and Webb, K. S., J. Chromatog., Vol. 79, pg. 57, 1973.
124. Bogovski, P., Walker, E. A., and Davis, W. , "N-Nitroso Compounds in
the Environment," Proceedings of a working conference held at the
It'l. Agency for Research on Cancer held at Lyons, France, October,
1973, (IARC Scientific Publication No. 9).
125. Fazio, T. , Damico, J. N. , Howard, J. W. , White, R. H. , and Watts,
J. 0., J. Agr. Food Chem., Vol. 19, pg. 251, 1971.
126. Essigmann, J. M. and Issenberg, P., J. Food Sci. , Vol. 37, pg. 684,
1972.
127. Palframan, J. F., Macnab, J. , and Crosby, N. T. , J. Chromatog.,
Vol. 76, pg. 307, 1973.
128. Sen, N. P., Donaldson, B. , Jyengar, J. R. , and Panalaks, T. ,
Nature, Vol. 241, pg. 473, 1973.
128a. Crosby, N. T. , Foreman, J. K. , Palframan, J. F. , and Sawyer, R.,
Nature, Vol. 238, pg. 342, 1972.
129. Rhoades and Johnson, J. Chromatog. Sci., Vol. 8, 1970.
130. Sen, N. P., J. -Chromatog., Vol. 51, pg. 301. 1970.
131. Althrope, J. , Goddard, D. A., Sissons, D. J. , and Telling, G. M. ,
J. Chromatog., Vol. 53, pg. 371, 1970.
132. Alliston, T. G. , Cox, G. B. , and Kirk, R. S., Analyst, Vol. 97,
pg. 915, 1972.
133. Fine, D. H. , Rufeh, F. , and Lieb, D. , Nature, Vol. 247, pg. 309,
1974.
134. Fine, D. H. , Lieb, D. , and Rounbehler, D. P-, Anal. Chem., Vol. 47,
pg. 1188, 1975.
135. Fine, D. H. and Rounbehler, D. P., J. Chromatog, Vol. 109, pg. 271,
1975.
136. Oettinger, P. E. , Huffman, F., Fine, D. H. , and Lieb, D. , Anal.
Lett., Vol. 8, No. 6, pg. 411, 1975.
137. Fine, D. H. , Rounbehler, D. P., and Belcher, N. M., It'l. Agency
for Research on Cancer,Publ. No. 14, Lyon, France, (In Press),
presented at the 4th It'l. meeting on the analysis of N-Nitroso
compounds in the environment, Tallinn, U.S.S.R., 1975.
197
-------
138. Fine, D. H., Rounbehler, D. P. and Oettinger, P. E., Anal. Chim.
Acta, Vol. 78, pg. 383, 1975.
139. Fine, D. H. and Rounbehler, D. P., "N-Nitroso Compounds in Water,"
presented at the First Chemical Congress of the North American
Continent, Mexico City, Dec., 1975.
140. Fine, D. H. and Rounbehler, D. P., "N-Nitroso Compounds in Air
and Water," presented at the It'l. Conference on Environmental
Sensing and Assessment, Las Vegas, 1975.
141. Fine, D. H., Rounbehler, D. P., Huffman, F., Garrison, A. W., Wolfe,
N. L. , and Epstein, S. S. , Bull. Env- Contain, and Toxicology,
Vol. 14, pg. 404, 1975.
142. Fine, D. H., Rounbehler, D. P., Silvergleid, A., and Ross, R.,
Second Symposium on Nitrite in Meat Products, Central Institute
for Nutrition and Food Research TNO, Zeist, Netherlands, 1976.
143. Fine, D. H., Ross, R., Rounbehler, D. P., Silvergleid, A., and Song,
L., J- Agr. Food Chem., Vol. 24, No. 5, pg. 1069, 1976.
144. Fine, D. H. and Rounbehler, D. P., "N-Nitroso Compounds in Drinking
Water," in Identification and Analysis of Organic Pollutants in
Water, Keith, L. H., (Ed.), Ann Arbor Science, 1976, pg. 255.
145. Fine, D. H., Ross, R., Fan, S., Rounbehler, D., Silvergleid, A.,
Song, L., and Morrison, J., "Determination of N-Nitroso Pesticides
in Air, Water and Soil," presented at the 172nd American Chemical
Society, San Francisco, Calif., 1976.
146. Fine, D. H., Rounbehler, D. P., Sawicki, E., Krost, K., and De
Marrais, G. A., Anal. Lett., Vol. 9, No. 6, pg. 595, 1976.
147. Shapley, D. , Science, Vol. 191, 1976.
148. Walker, E. A. and Castegnaro, M., "New Data on Collaborative Studies
on Analysis of Volatile Nitrosamines," presented at It'l. Conference
on the Analysis of N-Nitroso Compounds, Tallinn, U.S.S.R., 1975.
148a. Fine, D. H., Rounbehler, D. P., Sawicki, E., and Krost, K., Environ.
Sci. Tech., Vol. 11, No. 6, pg. 577, 1977.
148b. Fine, D. H., Rounbehler, D. P., Rounbehler, A., Silvergleid, A.,
Sawicki, E., Krost, K., and De Marrais,G. A., Environ. Sci., Tech.,
Vol. 11, No. 6, pg 581, 1977.
149. Pellizzari, E. D., Bunch, J. E., Berkley, R. E., and Bursey, J. T.,
Biomedical Mass Spectrometry, Vol. 3, pg. 196, 1976.
198
-------
150. Pellizzari, E. D., Bunch, J. E., Berkley, R. E., and Me Crae, J.,
Anal. Chem., Vol. 48, pg. 803, 1976.
151. Pellizzari, E. D. , Development of Analytical Techniques for
Measuring Ambient Atmospheric Carcinogenic Vapors, Pub. No.
EPA-600/2-75-076, Cont. No. 68-02-1228, 1975.
152. Pellizzari, E. D. , Development of Method for Carcinogenic Vapor
Analysis in Ambient Atmosphere, Pub. No. EPA-520/2-74-121, Cont.
No. 68-02-1228, 1974.
153. Pellizzari, E.D., Bunch, J. E. , Carpenter, B. H., Sawicki, E.,
Environ. Sci. Tech., Vol. 9, pg. 552, 1975.
154. Pellizzari, E. D., Carpenter, B. H., Bunch, J. E., Sawicki, E. ,
Environ. Sci. Tech., Vol. 9, pg. 556, 1975.
155. Pellizzari, E. D., Bunch, J. E., Berkley, R. E., Me Crae, J.,
Anal. Lett., Vol. 9, pg. 45, 1976.
156. Fine, D. H., Rounbehler, D. P., Pellizzari, E. D., Bunch, J. E.,
Berkley, R. W. , Me Crae, J. , Bursey, J. T., Sawicki, E., Krost, K. ,
and De Marrais,G. A., Bull. Env. Contain, and Toxicology, Vol. 15,
No. 6, pg. 739, 1976.
157. Pellizzari, E. D., Development of Method for Carcinogenic Vapor
Analysis in Ambient Atmospheres, Pub. No. EPA-650/2-74-121,
Cont. No. 68-02-1228, pg. 148, 1974.
158. Pellizzari, E. D., Development of Analytical Techniques for Measuring
Ambient Atmospheric Carcinogenic Vapors, Pub. No. EPA-600/2-75-076,
Cont. No. 68.-02J-1288, pg. 185, 1975.
159. Wade, W. A., Report of Study of N~Nitrosodimethylamine, FMC Corp.,
Ind. Chem. Div., Baltimore, My., The Research Corp. of England
Report No. 31555, 1975.
160. Lasaske, B., Dimethylnitrosamine (DMN) Sampling Program, Belle,
W. Va. , 1975, E. I. du Pont De Nemours and Co., Inc., presented
to EPA, Jan., 1976.
VII. Sulfur Dioxide
A. General Description
Sulfur dioxide is a highly irritating, nonflammable, colorless
gas at room temperature and atmospheric pressure. The gas is readily
Detectable in concentrations of 3-5 ppm by the human nose. The freezing
P°int (1 atm) is -75.5°C and the boiling point (1 atm) is -10.0°C. The
chemical formula for sulfur dioxide is SO2 and the molecular weight is
64.063. A common synonym for SCU is sulfurous acid anhydride.
199
-------
B. Results of Literature Search
The bulk of published literature regarding the analysis of SO^ has
been in reference to ambient air sampling. During recent years, with
the development of instrumental methods of analysis, SO2 has been
measured in stationary and mobile source exhausts. A review of the
references shows a wide variety of analytical technicmes. The most com-
monly employed colorimetric technique is the West-Gaeke method.(162-166)
This method has been collaboratively tested, the lowest concentration
range studied being well above the levels most frequently present in rural
and global background air.d67) A modified version of the West-Gaeke method
involves the collection of SO2 in O.lM sodium tetrachloromercurate (II) (TCM)
to form a dichlorosulfitomercurate complex (DCSM). The DCSM resists oxida-
tion by either oxygen in the air or that dissolved in the absorbing solution.
Ethylenediamine tetracetic acid disodium salt (EDTA) is added to the TCM
absorbing solution to complex any heavy metals that could oxidize sulfur
dioxide before DCSM is complete.(168, 169) Sulfamic acid is added to
destroy any nitrite ion in the absorbing solution.(170)
The colorimetric determination of sulfur dioxide.is based upon the
measurement of the red-violet color produced by the reaction of DCSM with
hydrochloric acid, pararosaniline and formaldehyde. Investigations into
the effect of dye purity on the analytical procedure have been reported by
several researchers.(171, 172) Techniques for purification of commercial
grade dye by recrystallization have been published.(163, 165, 173) Parar-
osaniline purified especially for the colorimetric analysis of sulfur
dioxide is commercially available.(173)
The major potential source of error associated with the West-Gaeke
colorimetric method for measuring SO2 is the widely differing collection
efficiency reported for Greenburg-Smith and midget impingers at low atmos-
pheric concentrations.(174) urone et al investigated the collection
efficiency in TCM solution of microgram quantities of sulfur dioxide tagged
with 35s.(175) xt was found that series bubblers cannot be used to deter-
mine absorber collection efficiency. Bostrom observed a 99% collection
efficiency for sulfur dioxide in TCM solution within the range of 100-
1000 ppb.(17&)
Work has been conducted in the development of other colorimetric
methods for the analysis of sulfur dioxide. Attari absorbed sulfur dioxide
into a solution of ferric ammonium chloride, perchloric acid and phenanthro-
line dye. (177) & color complex with an absorbance at 510 nm was formed and
although the color developed within 10 minutes it tended to fade with time.
Hydrogen sulfide was observed to seriously interfere.
Kawai used the reaction of barium chloranilate with sulfate as an
indirect measurement of sulfur dioxide. (1?8) sulfur dioxide was absorbed
in a solution containing hydrogen peroxide and barium chloranilate.
Barium chloranilate reacts with the sulfate ion producing a red-violet
chloranilic acid ion. Although this method may be satisfactory for flue
gas analysis it lacks the sensitivity required for ambient air analyses.
200
-------
Conductivity methods have been used for continuously monitoring sul-
fur dioxide in air. (179) The conductivity of a dilute sulfuric acid-
hydrogen peroxide reagent changes due to the absorption cf pollutants.
This change in conductivity is assumed to result primarily from sulfur
dioxide absorbed from the sampled air and oxidized to sulfuric acid. In
many cases, sulfur dioxide is the major pollutant present; however, if
other pollutants are present, their collection efficiency and solubility
may be significantly different than for sulfur dioxide. Several field
comparisons of conftf§^fg'j|:y witn other sulfur dioxide procedures indicate
a fair agreement. Hydrochloric acid gas, ammonia and chlorine substan-
tially increase conductivity. Shikiya and McPhee found two- to fourfold
differences between different conductivity analyzers and between con-
ductivity and colorimetric analyzers. (182) Although the conductivity
procedure may be acceptable for point sources of sulfur dioxide in iso-
lated areas, its high potential for positive and negative interferences
limits its application.
lodometric methods were among the first adapted for air pollution
analysis from the industrial hygiene literature. With this method the
sulfur dioxide is collected in an impinger containing a standard NaOH.
The absorbing solution is acidified and the liberated sulfurous acid is
titrated with standard iodine solution. (183) Another method employs a
standard iodine-potassium absorbing solution. (184) lodometric methods of
analysis for sulfur dioxide generally suffer from a lack of sensitivity
and interferences from
Adsorption sampling methods have also been developed for the mea-
surement of sulfur dioxide. (186) Sulfur dioxide is sorbed on silica gel,
desorbed and reduced to H2S at' 700-900°C over a platinum catalyst. The
H2S is then absorbed in a 2% ammonium molybdate solution and determined
colorimetrically. Although this technique is relatively specific for SC>2,
the final colorimetric determination by the molybdenum complex does not
utilize the most sensititive method available.
In addition to the aforementioned techniques, sulfur dioxide has been
measured by filtration (18 7-191) and static collectors^192"198) Air
samples are passed through potassium bicarbonate impregnated filters and
analyzed for sulfate. The collection efficiency of these filters is de-
pendent upon humidity, temperature, and the atmospheric concentration of
sulfur dioxide. The lead peroxide candle static collector was developed
by Wilson and McConnell as an inexpensive method for measuring relative
"sulfation" of the atmosphere. (192) The sulfur dioxide collection efficiency
is dependent upon temperature, relative humidity, wind speed, atmospheric
concentration of SO2 and the length of exposure period. (193) Ikeda deter-
mined ambient SO2 levels by collecting samples on active carbon filters,
washed with distilled water and titrated with barium chloranilate.
With the advent of modern instrumental methods of analysis, spe-
cifically gas and ion chromatography , a substantial amount of data has
been published. Most trace gas analysis for SO2 has been conducted
using gas chromatographs with flame photometric detectors (FPD) . (199 " 208)
The FPD is highly selective for sulfur compounds and has low minimum
201
-------
detection limits. Analysis for SO2 is generally performed using all tef-
lon or glass systems. Sulfur dioxide will react with active sites in the
gas chromatograph system, making the use of inert materials essential for
trace quantitative analysis. Gas chromatographs with FPD and linearizing
circuitry provide ^ wide dynamic range for ambient and source SC>2 levels.
In some instances the- collection technique precludes the use of GC-FPD
techniques; i.e., bag sampling from dilute automotive exhaust or source
sampling. The use of gas chromatography would be a prime candidate if
the sample integrity could be assured in the sample acquisition and sub-
sequent analysis.
A more recent development in methods of analysis for sulfur dioxide
involves the use of ion chromatography.(209) This technique involves
collection in a hydrogen peroxide absorbing reagent and measurement of the
resulting sulfate ion using ion chromatography. Ion chromatography is a
specialized area of liquid chromatography,which will separate and quantify
the individual cations or anions. This technique has been applied to the
measurement of SO2 in ambient air and adaptation to dilute automotive exhaust
should not be difficult.
Other instruments are commercially available that are reported to
measure S02 in ambient or dilute automotive exhaust. Such instruments
include continuous detection by pulsed fluorescent'UV and second derivative
UV analyzers. A pulsed fluorescent UV analyzer for SO2 was found to give
recoveries on the order of 115-125%, indicating that a positive interference
is present.(210) The second derivative UV sulfur dioxide analyzer has
inherent problems when being used on continuous samples. The mirrors are
located in the actual cell and become coated with various exhaust components
even after filtration of the sample. The mirrors will become etched and
need resurfacing if the unit is used in the presence of SO2, SO4=, or other
corrosive exhaust components. The inherent noise level along with the
consistent mirror problem preclude the use of a second derivative UV ana-
lyzer for measuring SC>2 on a continuous basis.
A variation of the GC method for measuring SO2 is the use of a con-
tinuous analyzer using an FPD detector. Although this approach is good
in theory, it has several problems associated with the performance of
the FPD detector. These units were originally designed to monitor S02
levels in the ambient and adaptation to automotive exhaust was not straight-
forward. Air samples had essentially the same O2 and N2 levels all of the
time; however, dilute exhaust samples have variable CC>2, C>2 and N2 con-
centrations. The species have been found to cause quenching effects on a
FPD detector. With the constantly changing CC>2, O2 and N2 it would be im-
possible to correct for any quenching effect. The use of a continuous FPD
analyzer for measuring SO2 in automotive exhaust would not be acceptable
unless the quenching effects could be eliminated.
C. Recommendations
The method that appears the most appropriate for the measurement of
S02 in dilute automotive exhaust is that using the ion chromatograph.
This technique has been found to be accurate at ambient levels and could
202
-------
be readily adapted to exhaust samples. The sample is collected in hydro-
gen peroxide and an aliquot is injected into the ion chromatograph. The
resulting S04= peak is proportional to the S02 concentration and emission
rates can be determined.
An alternative procedure would be to use a modified West-Gaeke method.
Although this procedure has been used for a number of years in ambient
studies, the TCM absorbing reagent is quite dangerous and extreme care
must be exercised during its handling.
D. References
161. Braker, W. and Mossman, A. L. , Matheson Gas Data Book, 5th
Edition, East Rutherford, N. J., 1971, pg. 513
162. West, P. W. and Gaeke, G. C. , Anal. Chem., Vol. 28, pg. 1816,
1956.
163. Adams, D. F. , Corn, M. , Harding, C. I., Pate, J.P., Plumley,
A. L., Scaringelli, F. P., and Urone, P., "Methods of Air
Sampling and Analysis," [Tentative Method of Analysis for
Sulfur Dioxide Content of the Atmosphere (colorimetric), No.
42401-01-69T], pg. 447, Amer. Pub. Health Ass., Washington,
D.C., 1972.
164. "ASTM Standards - Water, Atmospheric Analysis," ASTM
Designation D291470T, Part 23, Am. Soc. Test. Mat.,
Philadelphia, Pa., 1972.
165. US EPA, Fed. Register, Vol. 36, No. 158, 15492, 1971.
166. Robinson, E. and Robbins, R. C., "Sources, Abundance, and
Fate of Gaseous Atmospheric Pollutants," Final Report, SRI
Project 6755 Supplement, Stanford Research Inst., Palo Alto,
Calif., 1968.
167. Stern,A. C. (Ed.), Air Pollution, 3rd Edition,
Academic Press, Inc., N. Y., 1976, Vol. 3, pg. 214.
168. Zurlo, N. and Griffini, A. M., Med. Lav., Vol. 53, pg. 330,
1962.
169. Scaringelli, P.p., Elfers, L., Norris, D., and Hochheiser, s,
Anal. Chem., Vol. 42, pg. 1818, 1970.
170. Pate, J. B. , Ammons, B. E. , Swanson, G. A., Lodge, Jr.,
J. P., Anal. Chem., Vol. 37, pg. 942, 1965.
171. Pate, J. B., Lodge, Jr., J. P., and Wartburg, A. F., Anal.
Chem., Vol. 34, pg. 1660, 1962.
203
-------
172. Scaringelli, F. P., Saltzman, B. E., and Frey, S. A.,
Anal. Chem., Vol. 39, pg. 1709, 1967.
173. Hartman-Leddon, Phila., Pa.
174. Stern,A. C. (Ed.), Air Pollution, 3rd Edition,
Academic Press, Inc., N. Y. , 1976, Vol. 3, pg. 214.
175. Urone, P., Evans, J. B. , and Noyes, C. H. , Anal. Chem.,
Vol. 37, pg. 1104, 1965.
176. Bostrom, C. E., Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.
177. Attari, A., Igielski, T. P., and Jaselskis, B., Anal. Chem.,
Vol. 42, pg. 1282, 1970.
178. Kawai, T., Netsu Kanri, Vol. 22, pg. 20, 1970.
179. Thomas, M. D., Ivie, J. O., and Fitt, T. C., Ind. Eng. Chem.,
Anal. Ed., Vol. 18, pg. 383, 1946.
180. Yocum, J. E., Richardson, R. L., Saslaw, I. M., and Chapman,
S. , Proc. 49th Ann. Meet. Air Poll. Contr. Ass., Pittsburgh,
Pa., 1956.
181. Kuczynski, E. R., Environ. Sci. Tech., Vol. 1, pg. 68, 1967.
182. Shikiya, J. M. and McPhee, R. D., 61st Annual Meeting, Paper
No. 68-72, Air Poll. Contr. Ass., Pittsburgh, Pa., 1968.
183. Jacobs, M. B. , The Chemical Analysis of A_ir Pollutants, Chapter 8,
Wiley (Interscience), New York, N. Y., 1960.
184. Jacobs, M. B., The Analytical Chemistry of Industrial Poisons,
Hazards, and Solvents, Chapter 9, Wiley (Interscience), New
York, N. Y., 1949.
185. Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.
186. Stratman, H., Mikrochim.Acta, Vol. 6, pg. 688, 1954.
187- Pate, J. B., Lodge, Jr., J. P., and Neary, M. P., Anal. Chim.
Acta., Vol. 28, pg. 341, 1963.
188. Lodge, Jr., J. P., Pate, J. B., and Huitt, H. A., Amer. Ind.
Hyg. Ass., J., Vol. 24, pg. 380, 1963.
189. Forrest, J. and Newman, L. , Atmos. Environ., Vol. 7, pg. 561,
1973.
190. Huygen, C., Anal. Chim. Acta, Vol. 28, pg. 349, 1963.
204
-------
191. Harding, J-, and Schlein, B., "Nuclear Techniques in Environ-
mental Pollution," SM-l-42a/9, IAEA, Vienna, Austria, 1971.
192. Wilsdon, B. H. and McConnell, F. J. , J. Soc. Chem. Ind.
Vol. 53, pg. 385, 1934.
193. Huey, N. A., J. Air. Poll. Cont. Ass., Vol. 18, pg. 610, 1968.
194. Thomas, F. W. and Davidson, C. M. , J. Air Poll. Cont. Ass.,
Vol. 11, pg. I/ 1961.
195. Stalker, W. W. , Dickerson, R. C., and Kramer, G. D., Amer.
Ind. Hyg. Ass., J., Vol. 24, pg. 68, 1963.
196. Hickey, H. R. and Hendrickson, E. R. , J. Air Poll. Cont. Ass.,
Vol. 15, pg. 409, 1965.
197. Harding, C. I. and Kelley, T. R. , J. Air Poll. Cont. Ass., Vol
17, pg. 545, 1967.
198. Ikeda, T., J. Hyg. Chem., Vol. 13, pg. 41, 1967.
199. Stevens, R. K., Mulik, J. D. , O'Keeffe, A. E. , and Krost, K. J. ,
Anal. Chem., Vol. 43, pg. 827, 1971.
200. Stevens, R. K., O'Keeffe, A. E. , and Ortman, G. C. , Environ.
Sci. Tech., Vol. 3, pg. 652, 1969.
201. Ronkainen, P., Denslow, J., Leppanen, O., J. Chromatog.
Sci., Vol. 11, pg. 384, 1973.
202. Pecsar, R. E. and Hartmann,C. H. , J. Chromatog. Sci., Vol.
11, pg. 492, 1973. -
203. Stevens, R. K. and O'Keeffe, A. E., Anal. Chem., Vol. 42,
pg. 143a, 1970.
204. Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I.,
Anal. Chem., Vol. 44, pg. 2070, 1972.
205. Bruner, F., Ciccioli, P., and DiNardo, F., Anal. Chem.,
Vol. 47, pg. 141, 1975.
206. DeSouza, T. L. C. , Lane, D. C. , and Bhatia, S. P., Anal.
Chem., Vol. 47, pg. 543, 1975.
207. Bruner, F., Ciccioli, P., and DiNardo, F., J. Chromatog.,
Vol. 99, pg. 661, 1974.
208. Bremner, J. M. and Banwart, W. L., Sulfur Inst. Journal,
Vol. 10, pg. 6, 1974.
205
-------
209. Mulik, J. D., Todd, G., Estes, E., and Sawicki, E., "Ion
Chromatography Determination of Atmospheric Sulfur Dioxide,"
Symposium on Ion Chromatographic Analysis of Environmental
Pollutants, EPA, Research Triangle Park, N. C., April,
1977.
210. Dietzmann, H. E., "Protocol to Characterize Gaseous Emissions
as a Function of Fuel and Additive Composition," Environmental
Protection Technology Series under Contract 68-02-1275,
September, 1975.
VIII. Hydrogen Sulfide
A. General Description(211'212)
Hydrogen sulfide, also called hydrosulfuric acid and sulfureted hydrogen
has the chemical formula H2S. This flammable, poisonous gas has the char-
acteristic odor of rotten eggs. The boiling point (1 atm) is -60.33°C and
the melting point (1 atm) is -85.49°C. The density of the liquid at the
boiling point is 0.993 g/ml and the specific gravity of the gas at 15°C
is 1.1895. Hydrogen sulfide is considered an extremely toxic gas. In
1968, the American Conference of Governmental Industrial Hygienists recom-
mended a threshold limit of 10 ppm for workers exposed over long periods
of time without adverse effects.
B. Results of Literature Search
Since hydrogen sulfide has such a terrible odor and is extremely poi-
sonous, several methods have been developed for its detection. Some of
these methods are plate,tiles, tapes or filters, catalysts, fluorimetry,
infrared spectroscopy, sulfur ion selective electrode, gas chromatography,
coulometry and two colorimetric methods.' '' Most of these methods
are useful for specific situations; require long sampling times; are only
useful at high concentrations; or are also sensitive to mercaptans, organic
sulfides and sulfur dioxide and trioxide. Many of these species are present
in the exhaust and would complicate the analysis.
The methods least likely to be advantageous are the tape, tile, plate
and filter methods. These surfaces are coated with lead acetate or mercuric
chloride.(215-218) A sheet of metallic silver has also been used. The
problem with this technique is one of time. The sampling times are between
10-90 minutes for the mercuric chloride tape and seven days for the metallic
silver plate. Hydrogen sulfide concentration is measured by the change in
reflectance of the plates after exposure or the optical density (percent
transmittance) of the tape. This method would be good for ambient air
samples on a daily or hourly basis.
There are two methods using a catalyst to determine hydrogen sulfide
concentration. The first method uses a nickel catalyst to reduce organic
sulfides into hydrogen sulfide, which is analyzed with the methylene blue
method discussed later.(219) The other method uses palladium.(220) This
noble metal allows hydrogen to diffuse through the metal lattice. The
206
-------
rate of hydrogen transport is constant at a given temperature and pressure
difference across the palladium foil. Hydrogen sulfide "poisons" the sur-
face of the catalyst and inhibits the diffusion rate of hydrogen. This
is used with "on line" systems and would be difficult and expensive to
adapt for use with automotive exhaust. " =>:"' -
The fluorimetry method uses fluorescein-mercuric acetate reagent.(221,222)
Hydrogen sulfide quenches the fluorescence of this reagent. This method
has a reported detection limit of one ppb. A study of the possible inter-
ferences from exhaust species has not been carried out and confirmation of
the reported sensitivity is necessary.
The infrared method uses the principle of light absorption due to
the vibrational modes of the molecule. This technique is sensitive to
about 10 ppm but any sulfide or disulfide present will interfere. This
is also a useful method for continuous monitoring of hydrogen sulfide
levels.
The sulfur-ion selective electrode contains a solid silver sulfide
membrane which separates the internal reference from the test
solution. (223-227) A voltage develops between the test solution and the
reference. This voltage depends on the activity of the silver ion.
Previous experience with ion selective electrodes has shown stability
problems at low concentrations.
Gas chromatography is a useful technique for separating and
analyzing many compounds. Many different detectors and columns have
been used to determine hydrogen sulfide. These have been used on
ambient air samples and "on line" systems. (199-208,228-238) From
preliminary tests, hydrogen sulfide is highly reactive with the species
in automotive exhaust. In this case, it is difficult to preserve sample
integrity. A method of trapping other than Tedlar bag samples would
have to be found.
A coulometric microtitration device was adapted from gas chroma-
tography to direct use on ambient air analysis and used by Adams et.
al. (239) Filters were used to eliminate some of the sulfur-containing
constituents in the sample. Coulometric detectors lack the selectivity
required. (240)
There are two colorimetric methods available.(241-250) The sodium
nitroprusside method is sensitive to about one ppm. The methylene blue
method is sensitive to 1-3 ppb. Cadmium or zinc hydroxide, cadmium sul-
fate, or zinc acetate are used as the absorbing media in a glass
impinger. An acidic solution of N,N-dimethyl-p-phenylenediamine
dihydrochloride or sulfate is then added to the absorbed sample.
Methylene blue is formed and fully developed after 15 minutes at 20°C.
The color is constant in subdued light for several hours. The ab-
sorbance at 670 nm is determined with a spectrophotometer.
207
-------
C. Recommendations for Analysis
Of the many techniques available from the literature search, the
best appears to be the methylene blue method. The proposed procedure
was obtained from the Project Officer under EPA Contract 68-03-2499 in
May, 1977. This is a modification of the technique used by Gustafsson.
Although this method is well documented, tests will be run to deter-
mine concentration limits and possible interferences.
D. References
211. Stecher, P. G. , (Ed.), The Merck Index, An Encyclopedia of
Chemicals and Drugs, 8th Edition, Merck and Co., Inc.,
Rahway, N. J., 1968, pg. 545.
212. Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th
Edition, East Rutherford, N. J. , 1971, pg. 319.
213. Bethea, R. M., J. Air Poll. Cont. Assoc., Vol. 23, pg. 710, 1973.
214. Bamesberger, W. L. and Adams, D. F., Tappi, Vol. 52, pg. 1302,
1969.
215. Falgout, D. A. and Harding, C. I., J. Air Poll. Cont. Ass.,
Vol. 18, pg. 15, 1968.
216. Biles, B., Brown, C., and Nash, T., J. Phys. E. : Scientific
Instruments, Vol. 7, pg. 309, 1974.
217. Wellinger, R. and Giever, P. M., Am. Ind. Hyg. Ass. J.,
Vol. 35, pg. 730, 1974.
218. Natusch, D. F. S., Sewell, J. R., and Tanner, R. L. , Anal.
Chem., Vol. 46, pg. 410, 1974.
219. "Trace Sulfur in Hydrocarbons by the Raney Nickel Method,"
Analytical Method - Informational, Analytical and Informa-
tional Division, Exxon Research and Engineering Co.,
September, 1975.
220. Pierce, R. W., ISA Trans., Vol. 13, pg. 291, 1974.
221. Axelrod, H. D., Gary, J. H., Bonelli, J. E., and Lodge,
J. P., Anal. Chem., Vol. 41, pg. 1956, 1969.
222. Natusch, D. F. S. , Klonis, H. B. , Axelrod, H. D., Teck, R. J.,
Lodge, Jr., J. P., Anal. Chem., Vol. 44, pg. 2067, 1972.
223. Bock, R. and Puff, H.,_Z. Anal. Chem., Vol. 240, pg. 381, 1968,
224. Nauman, R. and Weber, C., Z. Anal. Chem., Vol. 253 , pg. ill,
1971.
208
-------
225. Ehman, D. L., Anal. Chem., Vol. 48, pg. 918, 1976.
226. Hseu, T. M. and Rechnitz, G. A., Anal. Chem., Vol. 40, No. 7,
pg. 1054, 1968; See also correction: Ibid, Vol. 40, No. 11,
pg. 1661, 1968.
227. Baumann, E. W. , Anal. Chem., Vol. 46, No. 9, pg. 1345, 1974.
228. Blanchette, A. R. and Cooper, A. D. , Anal. Chem., Vol. 48,
pg. 729, 1976.
229. "Analysis of t^S and SC>2 in the ppb Range," Bulletin 722C,
Supelco, Inc., Belefonte, Pa., 1975.
230. Setter, J.R., Sedlak, J. M. , Blurton, K. F., J. Chromatog.
Sci., Vol. 15, pg. 125, 1977.
231. Greer, D. G. and Bydalek, T. J. , Environ. Sci. Tech., Vol. 7,
pg. 153, 1973.
232. Obermiller, E. L. and Charlier, G. O. , Anal. Chem., Vol. 39,
pg. 397, 1967.
233. Bethea, R. M. and Meador, M. C.r J. Chromatog. Sci., Vol. 7,
pg. 655, 1969.
234. Jones, C. N., Anal. Chem., Vol. 39, pg. 1858, 1967.
235. Wilhite, W. P. and Hollis, O.L. ,J. Gas Chromatog., Vol. 6, pg. 84,
1968.
236. Thornsberry Jr. ,w. L., Anal, Chem., Vol. 43, pg. 452, 1971
237. Adams, D. F., Jensen, G. A., Steadman, J. P., Koppe, R. K.,
and Robertson, T. J., Anal. Chem., Vol. 38, pg. 1094, 1966.
238. Martin, R. L. and Grant, J. A., Anal. Chem., Vol. 37, pg. 644,
1965.
239. Adams, D. F., Bamesberger, W, L., and Robertson, T. J., J.
Air Poll. Cont. Ass., Vol. 18, pg. 145, 1968.
240. Levaggi, D. A., Siu, W. , and Feldstein, M. , Advances in Auto-
mated Analysis, Vol. 8, pg. 65, 1972.
241. "Hydrogen Sulfide in Air Analytical Method," H.L.S., Vol. 12,
pg. 362, 1975.
242. Bamesberger, W. L. and Adams, D, F., Environ. Sci. Tech., Vol.3,
pg. 258, 1969.
243. Gustafsson, L., Talanta, Vol. 4, pg. 227, 1960.
209
-------
244. Moest, R. R. , Anal. Chem., Vol. 47, pg. 1204, 1975.
245. Buck, M. and Gies , H. , Staub, Vol. 26, pg. 27, 1966.
246. Jacobs, M. B. Braverman, M. M. , and Hochheiser, S., Anal.
Chem., Vol. 29, pg. 1349, 1957.
247. Cave, G. C. B. , Tappi, Vol. 46, pg. 1, 1963.
248. Budd, M. S. and Bewick, H. A., Anal. Chem., Vol. 24, pg. 1536,
1952.
249. Bostrom, C., Air and Water Poll. Int. J. , Vol. 10, pg. 435, 1966.
250. Flamm, D. L. and James, R. E. , Environ. Sci. Tech., pg. 159, 1975.
IX. Organic Sulfides
A. General Description
Carbonyl sulfide, dimethyl sulfide, dimethyl disulfide, and diethyl
sulfide are organic sulfides of concern on this research effort. The
chemical formulas, molecular weights, freezing points, boiling points
and common synonyms are listed in Table 3. Carbonyl sulfide is the only
sulfide of interest that is a gas at room temperature. In general, the
organic sulfides are malodorous compounds producing an unpleasant odor
similar to rotten eggs. The 1968 American Conference of Governmental
Industrial Hygienists made no recommendations for threshold limit values
for these sulfides.
B. '"'•• Results of Literature Search
-i
Spencer (251) et a± appiie(j the use of gas chromatography to analyze
gap odorants for mercaptan and sulfides. In earlier work, Sumner^2^2'
et al and Liberti (253) et al had limited their studies to mercaptans in
gas odorants. Ryce*254' et al and Desty (255) expanded the analysis of gas
odorants to include sulfides, although no retention times were given. The
aforementioned references used thermal conductivity detectors in their
analyses and were not concerned with trace gas analysis.
Coleman'25°) , et al used gas chromatography and mass spectrometry
to separate and identify eleven low-boiling sulfur compounds in crude
oil. Other researchers were performing independent studies on the re-
tention times under various gas chromatograph conditions. (257'258) im-
proved separation of mercaptans and sulfides using temperature program-
med gas chromatography was reported by Sullivan (25^) , et al. Sullivan
in effect, confirmed earlier work published by Dal Nogare^260^ et al
and Harrison (261) et al. Adams and Koppe^262) expanded the analysis to
include hydrogen sulfide, sulfur dioxide, mercaptans, alkyl sulfides
and disulfides. This work presented a gas chromatographic technique for
the separation and identification of complex mixtures in Kraft pulp di-
gester blow gas and black liquor combustion products.
210
-------
to
TABLE 3. LIST OF SULFUR COMPOUNDS INCLUDED IN
THE ANALYSIS OF ORGANIC SULFIDES
Chemical Molecular Freezing Boiling
Sulfur Compound Formula Weight Point, °C Point, °C
Carbonyl Sulfide COS 60.075 -138.8 -50.2
Dimethyl Sulfide CH3SCH3 62.13 -98.27 37.3
Dimethyl Disulfide CH3SSCH3 94.20 -84.72 109.7
Diethyl Sulfide C2H5SC2H5 90.19 -103.9 92.1
Methyl Mercaptan CH3SH 48.11 -121.0 6.2
Ethyl Mercaptan C2H5SH 62.13 -144.4 35.0
Synonyms
Carbon oxysulfide
Methyl sulfide
none
Ethyl sulfide
Methanethiol
Ethanethiol
-------
quantitatively measured carbonyl sulfide in natural gas
using gas chromatography. Although a minimum detection limit of 25 ppm
COS was reported, propylene was found to interfere with the analysis.
Hall(264) measured trace quantities of COS in carbonated beverages using
a gas chromatograph equipped with an electron capture detector. An im-
proved minimum detection limit of 0.3 ppm was a substantial improvement
over the work reported by Schols.(263) Adams1265' et al applied absorp-
tion techniques to concentrate sulfur compounds in waste process gases.
The pollutants were absorbed on activated silica gel at -78.5°C, desorbed
under heat and vacuum, trapped at -196°C and transferred onto a gas chroma-
tograph column for analysis.
In 1965, Feldstein^266^ reported separation of a variety of sulfur
compounds on a number of gas chromatograph columns. Several columns
showed promise, but did not separate organic sulfur compounds from
normally occurring atmospheric hydrocarbons. Predricks and
quantitatively measured mercaptans in sour natural gas by gas chroma-
tography and microcoulometry. Hydrocarbons did not interfere and the "
method was sensitive to as little as 1 ppm of individual mercaptan. The
method required only 30 ml of sample but required one hour to complete
the analysis.
With the development of the Melpar flame photometric detector, the
detection limits of sulfur compounds were improved considerably. In
1970, Mizany(268) characterized the FPD response to several sulfur
compounds and made comparisons to a Dohrmann microcoulometric detector.
Stevens and O'Keeffe^2^) discussed application of flame photometric
detectors to air pollution monitoring low concentrations. Permeation
calibration tubes were used to generate continuous samples of known
concentrations for various sulfur compounds. Both continuous and grab
sampling techniques were used in this study. Kummer(269) et al used
the gas phase chemiluminescent reaction of ozone with a number of or-
ganic sulfides. The possible uses of these reactions in monitoring low
concentrations of ozone and sulfur-containing pollutants were discussed.
In 1972, Rasmussen(27°) used a gas chromatograph with a flame photo-
metric detector to measure trace organic sulfur compounds in air. This
procedure was also applied to soil and water analysis and was found to
be a very sensitive and rugged instrument. The use of teflon tubing
throughout the gas chromatograph system minimized adsorptive losses.
Pecsar and Hartman(202> described an automated gas chromatographic
analysis of sulfur pollutants. By employing an all teflon system, sen-
sitives on the order of 10 ppb were achieved with confidence. Cali-
brations were made using permeation calibration tubes. Column fabri-
cation techniques were discussed with methods for achieving optimum
column and detector operation.
In 1974, Bremner and Banwart*208) used a Beckman GC-4 equipped with-
a Melpar flame photometric detector to identify volatile sulfur com-
pounds. Several columns were evaluated at several temperatures and GC
separation data is presented. Chromosorb T, Carbopak B-HT-100, Chro-
mosil 310, and Deactigel were evaluated in teflon columns. A number
212
-------
of other researchers(271-277) employed gas chromatography to quantita-
tively measure various sulfur compounds from a wide variety of sources.
Commercial analyzers are available that measure COS iu the sub-ppm
range.
C. Recommendations
Based on the results of the literature search, it is apparant that
the only technique that can detect ppb range sulfur compounds requires
the use of a gas chromatograph with a flame photometric detector. It
is proposed to use a Perkin-Elmer gas chromatograph with a flame photo-
metric detector. The organic sulfides will be concentrated on a Tenax-
GC trap, thermally desorbed into the gas chromatograph and quantitatively
measured using certified permeation calibration tubes as reference.
D. References
251. Spencer, C. F. , Baumann, F. , and Johnson, J. F. , Anal. Chem.,
Vol. 30, No. 9, pg. 1473, 1958.
252. Sumner, S. r Karrman, K. J. , and Sunden, V., Mikrochim Acta,
pg. 1144, 1956.
253. Liberti, A. and Cartoni, G. P., Chim. e ind. , Vol. 39, pg. 821,
1957.
254. Ryce, S. A. and Bryce, W. A., Anal. Chem., Vol. 29, pg. 925,
1957.
255. Desty, D. H. and Whyman, B. H. F., Anal. Chem., Vol. 29, pg. 230,
1957.
256. Coleman, H. J., Thompson, C. C. and Rail, H. T. , Anal. Chem.,
Vol. 30, pg. 1592, 1958.
257. Desty, D. H. and Harbourn, C. L. A., Div. of Analytical and
Petroleum Chemistry, Symposium on Advances in Gas Chromatog-
raphy, 132nd meeting, ACS, N.Y., N.Y., September 1957.
258. Amberg, C. H., Can. J. Chem., Vol. 36, pg. 590, 1958.
259. Sullivan, J. H. , Walsh, J. T. , and Meritt Jr., C., Anal. Chem.,
Vol. 31, pg. 1827, 1959.
260. Dal Nogare, S. and Bennett, C, E., Anal. Chem., Vol. 30, pg. 1157,
1958.
261. Harrison, G. F. , Knight, P., Kelly, R. P., and Heath, M. T. ,
Second It'l. Symposium on Gas Chromatography, Amsterdam, May,
1958.
262. Adams, D. F. and Koppe, R. K. , Tappi, Vol. 42, pg. 601, 1959-
213
-------
263. Schols, J. A., Anal. Chem., Vol. 33, pg. 359, 1961.
264. Hall, H. L., Anal. Chem., Vol. 34, pg. 61, 1962.
265. Adams, D« F , Koppe, R. K., and Jungroth, D. M. , Tappi, Vol. 43,
pg. 602, 1960.
266. Feldstein, M., Balestrieri, S., and Levaggi, D. A., J. Air
Poll. Cont. Ass., Vol. 15, pg. 215, 1965.
267. Fredericks, E. M. and Harlow, G. A., Anal. Chem., Vol. 36,
pg. 263, 1964.
268. Mizany, A. I,, J. Chromatog. Sci., Vol. 8, pg, 151, 1970.
269. Kummer, W. A., Pitts Jr., J. N., and Steer, R. P., Environ.
Sci. Tech., Vol. 5, pg. 1045, 1971.
270. Rasmussen, R. A., Am. Lab, Vol. 4, pg. 55, 1972.
271. Brody, S. S. and Chaney, J. E. , J. Gas 'Ghromatog. Vol. 4, pg. 42,
1966.
272. Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I.,
Anal. Chem., Vol. 44, pg. 2070, 1972.
273. Elliott, L. F. and Travis, T. A., Soil Sci. Soc. Amer. Proc.,
Vol. 37, pg. 700, 1973.
274. Francis, A. J,, Adamson, J., Duxbury, J. M., and Alexander, M.,
Bull. Ecol. Res. Comm., (Stockholm), Vol. 17. pg. 485, 1973.
275. Lewis, J. A. arid Papavizas, G. C. , Soil Biol. Biochem., Vol. 2,
pg.239, 1970.
276. Lovelock, J. E. , Maggs, R. J., and Rasmussen, R. A., Nature,
London, Vol. 237, pg. 452, 1972.
277. Ronkainen, P., Denslow, J., and Leppanen, O., J. Chromatog.
Sci., Vol. 11, pg. 384, 1973.
X. Nickel Carbonyl
A. General Description(278,279)
Nickel carbonyl, also called nickel tetracarbonyl, is a colorless,
flammable, highly volatile liquid. The melting point (1 atm) is -25°C
and the boiling point (1 atm) is 43.2°C. The density of the liquid at
20°C is 1.310 g/ml.
Nickel carbonyl has a musty "damp cellar" or "sooty" odor.(280)
It can be detected by the human olfactory at 1-3 ppm. In 1968, the
214
-------
American Conference of Governmental Industrial Hygienists recommended
a threshold limit for daily exposure without adverse effects of 0.001
ppm.
Carbon monoxide is a TT-acceptor ligand. (28D it bonds by donating
an electron from a sigma orbital and accepting an electron from a d TT or
dp ir metal orbital. In the case of nickel carbonyl, the valence is zero
and the structure is tetrahedral. Thermodynamic data show that it quickly
decomposes . (282>
B. Results of Literature Search
The toxicity and flammability of nickel carbonyl have provided an
impetus for the development of tests to determine concentrations. There
are several methods available. Some of these are infrared spectroscopy,
spark spectroscopy, flame photometry, light reflectance, colorimetry,
and a gas chromatography method using an electron capture detector.
Infrared spectroscopy is reported to have a detection limit of four
ppm by volume. (283) Spark spectroscopy has a 0.1 ppm detection limit
but the exposure time of the photographic plate requires considerable
time. The flame photometric method is expensive to operate and requires
the use of an open flame. (284) ^g detection limit is 0.1 ppm.
McCarley et al (285) developed a device that takes advantage of
the thermodynamics of nickel carbonyl. In the absence of carbon mon-
oxide, nickel carbonyl decomposes on contact with a hot surface, in
this case borosilicate glass. The rate of decomposition is proportional
to the nickel carbonyl concentration. The intensity of plane-polarized
light reflected from the surface is calibrated in ppm of nickel carbonyl.
The detection limit is 0.05 ppm but it is also sensitive to iron carbonyl
and possibly other metallo-organic gases. Another method used by Ball^
et al also takes advantage of the thermodynamic properties. Nickel
carbonyl is detected .by its pyro lysis products on the gaseous conduc-
tance in an ionization chamber.
There are several colorimetric methods available for the deter-
mination of nickel carbonyl. The first uses a solution of iodine in
ethanol or chloroform as the absorber. This technique was used by
Kincaid et al(287) and jjunold and Pietrulla. (288) The nickel in the
samples are complexed with dimethylglyoxime. The optical density is
measured spectrophotometrically at 440 nm. The detection limit is 2 ppb
but the collection time is on the order of hours. The second method in-
volves a saturated solution of sulfur in trifluoroethylene. ( The
samples is examined spectrographically in the ultraviolet region of
the spectrum. A third method involves the reaction of nickel carbonyl
with red mercuric oxide at 200°C. (290) The liberated mercury is de-
termined spectrographically. A fourth method uses dilute sulfuric acid
to collect nickel carbonyl. Sodium diethyldithiocarbamate is used to
complex the nickel. Belyakov(291) used chloramine B in ethanol as the
absorber and complexed the nickel with dimethylgloxime. Vol'berg and
Gerskhovich(292) used potassium iodide on silica gel as the absorber.
215
-------
Iodine ion is released and measured colorimetrically. Another method
uses dilute hydrochloric acid as the absorbing reagent and alphafuril-
dioxime as the complexirig agent. These last two seem questionable al-
though Brief et al^293^ reports a 90 percent collection efficiency with
a 3 percent hydrochloric acid solution.
A gas chromatograph with a Carbowax 20 M column and an electron
capture detector was employed by Sunderman et al (294) to determine
nickel carbonyl in the blood and breath of rats. The technique
involved vacuum extraction of nickel carbonyl from the blood and
trapping in cold ethanol (-78°C). The electron capture detector is
insensitive to most hydrocarbons but is sensitive to organo-metallic
compounds.
C. Recommendations for Analysis
An adaptation of the gas chromatography method would seem the most
likely candidate for the analysis of nickel carbonyl. Nickel carbonyl
is highly soluble in organic solvents such as ethanol, ether, benzene,
and chloroform, but is only slightly soluble in water. Any of these
solvents could be used to trap nickel carbonyl. Impingers filled with
any of these solvents would provide a satisfactory collection system
for nickel carbonyl in dilute auto exhaust. A filter would be required
to remove any particulate present in the exhaust prior to the impinger.
The advantages are the simplicity of the procedure and the speed of
determination. Experiments would have to be conducted to determine
any possible interferences.
An alternative procedure would be to use nitric acid instead of
hydrochloric or sulfuric acid in a colorimetry method. Nickel carbonyl
reacts vigorously with nitric acid but reacts slowly with hydrochloric
or sulfuric acid. Although this method is considered a good procedure,
it is manpower intensive and requires much wet chemical work.
Metal carbonyls are used to make carboxylic acids.(295) Perhaps a
technique could be developed to produce a derivative quantitatively.
This would then allow an indirect method of determination. This method
would require much development work to make it operative.
D. References
278. Stecher, P. G., (Ed.), The Merck Index, An Encyclopedia of
Chemicals and Drugs, 8th Edition, Merck and Co., Inc., Rahway,
N.J., 1968, pg. 727.
279. Braker, W. and Mossman, A. L., Ma theson Gas Data Book, 5th
Edition, East Rutherford, N.J., 1971, pg. 401.
280. "Nickel Carbonyl," Amer. Ind. Hyg. Assoc. J., pg. 304, 1968.
281. Cotton, F. A. and Wilkinson, G., Basic Inorganic Chemistry
John Wiley and Sons, Inc., New York, 1976, pg. 473.
216
-------
282. Spice, J. E. , Stavely, L. A. K,, and Harrow, G. A., J chem
Soc., pg. 100, 1955.
283. Me Dowell, R. S. , Amer. Ind. Hyg. Ass. J. , Vol. 32, pg. 621,
1971.
284. Densham, A. B. , Beale, P. A. A., and Palmer, R. , Journal Appl.
Chem.-, Vol. 13, pg. 576, 1963.
285. Me Carley, J. E. , Saltzman, R. S., and Osborn, R. H. , Anal.
Chem., Vol. 28, pg. 880, 1956.
286. Ball, K. E. , Bossart, C. J. , and Saltzman, R. S., "Detection of
Trace Amounts of Nickel Carbonyl and Tetraethyl Lead in Air,"
Pittsburgh Conference on Analytical Chemistry and Applied Spec-
troscopy, February, 1960.
287. Kincaid, J. P., Stanley, E. L. , Beckworth, C. H. , and Sunderman,
F. W. , Amer. J. Clin, Path., Vol. 26, pg. 207, 1956.
288. Hunold, G. A. and Pietrulla, W-, Arbeitsschutz, Vol. 8, pg. 193,
1961.
289. Pitet, M. G., Arch. Mai. Prof., Vol. 21, pg. 674, 1960.
290. Vol'berg, N. S., U.S.S.R., Patent No. 135,684, February 1961.
291. Belyakov, A. A., Zavodsk Lab., Vol. 26, pg. 158, 1960.
292. Vol'berg, N. S. and Gerskhovich, E. E., Hyg. Sanit. , Vol. 33,
pg. 226, 1968.
293. Brief, R. S. , Venable, F. S., and Ajemian, R. S., Ind. Hyg. J.,
Vol. 26, pg. 72, 1965.
294. Sunderman Jr., F. W., Roszel, N. O. , and Clark, R. J. , Arch.
Environ. Health., Vol. 16, pg. 836, 1968.
295. Bird, C. W., Chem. Rev., Vol. 62, pg. 283, 1962.
XI. Sulfuric Acid
A. General Description^296'297^
Sulfuric acid is a colorless, oily liquid which has a freezing point
of 10.5°C and a boiling point of 338°C. Sulfuric acid is also known as
hydrogen sulfate and the chemical formula H2S04. The molecular weight of
H2S04 is 98.08 and the density of the liquid at 20°C is 1.841 g/ml. The
strong affinity of concentrated sulfuric acid for water makes it a good
dehydrating agent. The amount of sulfuric acid us^d i
that of any other manufactured compound.
217
-------
B. Results of Literature Search
Upon initiating the literature search, it became apparent that a
number of analytical techniques for measuring sulfuric acid have been
applied to automotive exhaust. Since it was felt that a sufficient
variety of procedures had been applied to measuring sulfuric acid in
exhaust, the literature search was directed to reviewing those tech-
niques .(298-326) Methods that have been applied to measuring sulfate in
exhaust include barium perchlorate titration(323,324)t barium chloranilate
method<323'325^, iodate titration(323>, and microgravimetric analysis.(323)
The barium perchlorate titration is a general technique that has been
applied to the analysis of samples obtained on filters, by controlled con-
densation or in isopropyl alcohol bubblers. Soluble sulfates collected
on a filter are leached from the filter with nitric acid. This leach
solution is heated to boiling to remove excess acid, filtered to remove
insolubles and passed through an ion-exchange column to remove interfering
cations. After buffering the solution with methenamine to a pH of 3-4,
the resulting solution is titrated with barium perchlorate using Sulfonazo
(III) as an indicator. Samples collected in a controlled condensation
apparatus are removed by rinsing the coils and frit with deionized water.
The resulting solution is boiled down to about 5 ml., buffered to a pH
of 3-4 with methenamine and titrated in the same manner as samples from
filter systems. One variation of this method is to extract filters in
hot water, using thorin as an indicator.
If the sample was collected in isopropyl alcohol filled impingers,
the absorbing solutions are evaporated to a low volume. The solution is
then passed through an ion exchange resin to remove any cations that
would interfere by reacting with the indicator. The purified solutions
are titrated with standard barium perchlorate solution using thorin and
methylene blue as an indicator.
The barium chloranilate (BCA) method involves the analysis of water
soluble sulfate collected on Fluoropore filters during the emissions
tests. The filter is ammoniated using NH4OH, extracted in a 60% IPA
solution, and analyzed in a specially designed high pressure liquid
chromatograph system. This extract is injected into a high pressure
liquid chromatograph by means of a liquid sampling valve. The extrac-
tion solvent and LC solvent are both 60% IPA. The extract passes
through ion exchange resins to remove any interfering species and then
through a column of barium chloranilate. The BaSO4 precipitates out and
an equivalent amount of violet colored chloranilic acid ion is released
and quantitatively measured colorimetrically at 310 nm.
The iodate method is used to determine sulfate collected on glass
fiber filters. The sulfate is leached from the filter with a warm-
aqueous solution containing NaCO3 to eliminate lead ion interference.
An aliquot of this solution is added to an ethanol solution containing
formaldehyde to remove sulfite ion interference. Barium iodate is
added, along with the starch indicator, and the characteristic starch-
iodine color is measured spectrometrically at 590 nm.
218
-------
Sulfate has also been measured microgravimetrically in solutions
ranging from 1 to 100 mg. The solution is diluted to 100 ml, treated
with HCl and heated to near boiling. Barium chloride is added gradually
while stirring vigorously. The precipitate is allowed to nucleate on a
steam bath overnight at 80°C. The precipitate (and solution) is trans-
ferred to a preweighed filter paper and washed with deionized water,
ethyl alcohol and ethyl ether. The filter and precipitate are trans-
ferred to a drying oven and dried for one hour at 100°C. The filter
and precipitate are cooled in a desiccator and weighed on a semi-
roicrogram balance.
A substantial amount of data on sulfuric acid emissions has been
published using these techniques. Of the aforementioned techniques,
the BCA method is the most widely used. A number of cross checks
between EPA, EPA contractors, and industry researchers have been
conducted using a variety of sulfuric acid levels and included a
variety of analytical procedures. The general concensus was that the
BCA technique was more accurate at low sulfate levels that are antici-
pated in auto exhaust. Other techniques, such as the titration or
microgravimetric, required substantially more sample than was required
by the BCA method.
C. Recommendations
Based on the results of the literature search combined with practical
experience of Southwest Research Institute and EPA, it is recommended to
use the barium chloranilate method for measuring sulfuric acid in exhaust.
This method has been developed by EPA as a method for measuring sulfate
in automotive exhaust. Refinements to the original procedure are incor-
porated into the SwRI system and attempts will be made to keep abreast
of any additional improvements.
D. References
296. Weast, R. C. , (Ed.), Handbook of Chemistry and Physics, 54th
Edition, Division of the Chemical Rubber Co., 1973-1974, pg. B144.
297. Nebergall, W. H. , Schmidt, F. C. , and Holtzclaw Jr., H. P., College
Chemistry with Qualitative Analysis, 5th Edition, D. H. Heath and
Co., Lexington, Mass., 1960, pg. 581.
298. Gentel, J. E., et al, "Characterization of Particulates and
other Nonregulated Emissions from Mobile Sources and the
Effects of Exhaust Emissions Control Devices on These Emis-
sions," U.S. EPA-OAWP Report APTDl-1567, March, 1973.
299. Pierson, W. R., et al, "Sulfuric Acid Aerosol Emissions
from Catalyst-Equipped Engines," SAE Paper 740287, February
1974.
300. Beltzer, M. , et al, "The Conversion of SO2 Over Automotive
Oxidation Catalysts," SAE Paper 750095, February 1975.
219
-------
301. Hammerle, R. H., and Mikkor, M., "Some Phenomena Which Control
Sulfuric Acid Emissions from Oxidation Catalysts," SAE Paper
750097, February 1975.
302. Trayser, D. A., et al, "Sulfuric Acid and Nitrate Emissions
from Oxidation Catalysts," SAE Paper 750091, February 1975.
303. Bradow, R. L., et al, "Sulfate Emissions from Catalyst and
Noncatalyst Equipped Automobiles," SAE Paper 740528, October
1974.
304. Bradow, R. L. and Moran, J. B., "Sulfate Emissions from
Catalyst Cars - A Review," SAE Paper 750090, February 1975.
305. Burkart, J. B., et al, "Catalytic Converter Exhaust Emis-
sions," Paper No. 74-129, APCA 67th Annual Meeting, Denver,
Colorado, June, 1974.
306. Begeman, C. R. , et al., "Sulfate Emissions from Catalyst-
Equipped Automobiles," SAE Paper 741060, October, 1974.
307. Beltzer, M., "Particulate Emissions from Prototype Catalyst
Cars," EPA-650/2-75-054, May 1975.
308. Dietzmann, H. E. , "Protocol to Characterize Gaseous Emis-
sions as a Function of Fuel and Additives Composition,"Final
Report EPA-600/2-75-048, September, 1975.
309. Cooper, B. J., et al., "Sulfate Emissions from Monolith
Catalysts During Mileage Accumulation," SAE Paper 760035,
February 1975.
310. Crestwick, F. A., et al, "Sulfuric Acid Emissions from an
Oxidation-Catalyst Equipped Vehicle," SAE Paper 750411,
February 1975.
311. Braddock, J. N., and Bradow, R. L., "Emission Patterns of
Diesel-Powered Passenger Cars," SAE Paper 750682, June 1975.
312. Holt, E. L., et al, "Control of Automotive Sulfate Emissions,"
SAE Paper 750683, June 1975.
313. Griffing,M. E. , et al, "Exhaust Sulfur Oxide Measurement
Using Air Dilution," SAE Paper 750697, June, 1975.
314. Trayser, D. A., et al, "Effect of Catalyst Operating History
on Sulfate Emissions," SAE Paper 760036, February 1976.
315. Somers, J. H,, et al"Sulfuric Acid Emissions from Light Duty
Vehicles," SAE Paper 760034, February 1976.
220
-------
316. Trayser, D. A., et al, "Sixth and Final Annual Summary Report
on Chemical and Physical Characterization of Automotive Exhaust
Particulate Matter in the Atmosphere,"CRC Project CAPE-19-70,
September 1976.
317. Foster, J. F. , et al, "Fifth Annual Summary Report on Chemical
and Physical Characterization of Automotive Exhaust Particulate
Matter in the Atmosphere," CRC Project CAPE-19-70, March, 1976.
318. Ingalls, M. N. and Springer, K. J., "Measurement of Sulfate
and Sulfur Dioxide in Automotive Exhaust," Final Report EPA-
460/3-76-015, August 1976.
319. Environmental Protection Agency Proposed, "Recommended Practice
for Measurement of Exhaust Sulfuric Acid Emission from Light
Duty Vehicles and Trucks," November 1976.
320. Springer K. J. and Stahman, R. C., "Unregulated Emissions from
Diesels used in Trucks and Buses," SAE Report No. 770258,
February 1977.
321. Springer K. J. , and Stahman, R. C. , "Diesel Car Emissions -
Emphasis on Particulate and Sulfate," SAE Report No. 770254,
February 1977.
322. Springer K. J. , "Investigation of Diesel-Powered Vehicle Emis-
sions," EPA Report EPA-460/3-76-034, February 1977.
323. "Determination of Sulfur Compounds in Automotive Exhaust," SAE
Information Report, March, 1977.
324. "Method 8 - Determination of Sulfuric Acid Mist and Sulfur
Dioxide Emissions from Stationary Sources," Federal Register,
Vol. 36, No. 247, December 23, 1971, pg. 24893.
325. Tejada, S. B. , Sigsby, J. E., and Bradow, R. L. , "Determination
of Soluble Sulfates in Automobile Exhaust by Automated HPLC
Modification of the Barium Chloranilate Method," 1976.
326. Beltzer, M. , et al, "Measurements of Automotive Exhaust Par-
ticulate Emissions," SAE Paper 740286, February 1974.
221
-------
APPENDIX B
ALDEHYDE AND KETONE PROCEDURE
222
-------
THE MEASUREMENT OF ALDEHYDES AND KETONES IN EXHAUST
The measurement of aldehydes (formaldehyde, acetaldehyde, isobutyral-
dehyde, crotonaldehyde, hexanaldehyde, and benzaldehyde) and ketones
(acetone and methylethylketone) in exhaust is accomplished by bubbling the
exhaust through glass impingers containing 2,4 dinitrophenylhydrazine (DNPH)
in dilute hydrochloric acid. The exhaust sample is collected continuously
during the test cycle. The aldehydes and ketones (also known as carbonyl
compounds) react with the DNPH to form their respective phenylhydrazone de-
rivatives. These derivatives are insoluble or only slightly soluble in the
DNPH/HCl solution and are removed by filtration followed by pentane extrac-
tions. The filtered percipitate and the pentane extracts are combined and
then the pentane is evaporated in a vacuum oven. The remaining dried ex-
tract contains the phenylhydrazone derivatives. The extract is dissolved
in a quantitative volume of toluene containing a known amount of anthracene
as an internal standard. A portion of this dissolved extract is injected
into a gas chromatograph and analyzed using a flame ionization detector.
The detection limits for this procedure under normal operating conditions
are on the order of 0.005 ppm carbonyl compound in dilute exhaust.
LIST OP EQUIPMENT
The equipment required for the analysis of aldehydes and ketones is
divided into three groups: sample acquisition, sample preparation, and
sample analysis. Manufacturer, stock number and any pertinent descriptive
information are listed.
Sample Acquisition
1. Glass impingers, Ace Glass Products, Catalog #7530-11, plain
tapered tip stoppers with 18/7 arm joints and 29/42 bottle joints.
2. Flowmeter, Brooks Instrument Division, Model 1555, tube size
R-2-15-C, graduated 0-15, sapphire float, 0-5 £/min range.
3. Sample pump, Thomas Model 106 CA18, capable of free flow capacity
of 4 H/min.
4. Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
capacity.
5. Regulating valve, Nupro 4MG, stainless steel.
6. Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8"
ID and 5/16" OD x 1/8" ID.
223
-------
7. Teflon solenoid valve, The Fluorocarbon Company, Model DV2-144NCA1.
8. Drying tube, Analabs, Inc., Catalog #HGC-146, 6" long, 1/4" brass
fittings.
9. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, connec-
tors, etc.
10. Digital readout for dry gas meter.
11. Miscellaneous electrical switches, lights, wirings, etc.
12. Six channel digital thermometer, Analog Devices, Model S2036/J/1.
13. Iron/Constantan type J single thermocouple with 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation.
14. Variable autotransformer, Staco Inc., Type 3PN 1010.
15. Heating sleeve wrapped with insulation and insulation tape.
16. Class A, 20 m£ volumetric pipets.
17. Class A, 1000 m& volumetric flask.
18. Teflon coated stirring bar.
19. Hot plate-stirrer, Corning, PC-351.
Sample Analysis
1. Varian 1700 gas chromatograph equipped with dual flame ionization
detectors in differential operation, and a linear temperature
programmer.
2. Soltec Model B-281 1 mv recorder.
3. Hewlett-Packard Model 3354 gas chromatograph computer system with
remote teletype printout.
4. Syringe, 10 m&, Hamilton Company, #701.
5. Dual columns, 24 x 1/8" OD, stainless tubing packed with 6.7
percent Dexsil 300 GC on Chromosorb G 60/80 mesh, DMCS treated
and acid washed.
Sample Preparation
1. Fritted glass filters. Ace Glass Company, porosity D, ASTM 10 - 20
microns pore size, 24/40 ground glass joint, vacuum takeoff.
224
-------
2. Constant temperature vacuum oven, National Appliance Company.
3. Pump for oven, Thomas Industries, Model 907CA18 2.
4. Flasks, 125 mil capacity,. 24/40 ground glass joints.
5. Separatory funnels, 125 m&.
6. Separatory funnels, 250 mJl.
7. Separatory funnel shaker, Burrell Corporation, Wrist-Action(R)
type with appropriate funnel holders, Model 75. "~
8. Ring stands, labels, holders, tubing, vacuum tubing, fittings
and clamps needed for equipment manipulation.
9. Wash bottles, 500 m£.
10. Graduated cylinders, 50 mS,.
11. Vials, Kimble, 1/2 dram.
12. Vacuum pump, Sargent-Welch.
LIST OF REAGENTS
A list of the reagents used in the determination of the aldehydes and
ketones in exhaust is provided along with chemical formula, molecular weight,
purity, manufacturer, and catalog number.
1. Hydrochloric acid, HC1, 36.46 g/mole, concentrated (37%), analyt-
ical reagent, Mallinckrodt, Cat. #2612.
2. Pentane, C5H12, 72.15 g/mole, Distilled in glass (bp 35-37°C),
Burdick and Jackson Laboratories, Inc.
3. 2,4 Dinithrophenylhydrazine (2,4-DNPH) , (N02) 2C6H3CH=*I-NH2,
210.149 g/mole, Aldrich analyzed, Aldrich, Cat. #019,930-3.
4. Sodium Bicarbonate, NaHCO3, 84.00 g/mole, Mallinckrodt, Cat. #7412.
5. Anthracene, C14H10, 178.24 g/mole, K and K Laboratories, Cat.
#10714.
6. Toluene, C6H5CH3, 92.14 g/mole Baker Analyzed Reagent, Baker
Cat. #3-9460.
7. Methylene Chloride, CH2C12, 84.93 g/mole, Reagent ACS, Eastman,
Cat. #13022.
PREPARATION OF ABSORBING SOLUTION
225
-------
To prepare the absorbing solution, 163 m£ of concentrated HC1 and 2.5 g
of 2,4-DNPH crystals are added to a one liter volumetric flask containing
about 500 m& of deionized water. The flask is diluted to mark and stirred
for several hours at room temperature with an automatic stirrer/teflon
coated stirring bar to dissolve the DNPH. Fresh absorbing solution is pre-
pared daily as needed.
PREPARATION OF TOLUENE/ANTRACENE SOLUTION
Toluene containing approximately 0.05mg anthracene per m£ of toluene
is used to dissolve the dried phenylhydrazone extracts. This solution is
made by adding 100 mg of anthracene to a two liter volumetric flask and di-
luting to mark with toluene.
PREPARATION OF PHENYLHYDRAZONE DERIVATIVES
In order to obtain response factors for each of the phenylhydrazone
derivatives to anthracene, pure derivatives were prepared from their re-
spective aldehydes and ketones. These derivatives were made by adding each
of the carbonyl compounds separately to a 2N HCl-DNPH solution. The result-
ing orange to red precipitates were filtered and dried. The derivatives
were then recrystallized from hot absolute ethanal. The melting points
for each of the derivatives were compared to literature values before use.
A GC trace was also made on each of the derivatives to further check the
purity.
PREPARATION OF STANDARD SOLUTION OF PHENYLHYDRAZONE DERIVATIVES
AND ANTRACENE
A standard containing the phenylhydrazone derivatives and anthracene
in toluene is prepared to obtain a response factor of each of the deriva-
tives to anthracene. The solution is made by dissolving weighed amounts
of anthracene and each of the derivatives in a quantitative volume of
toluene. These solutions contain approximately O.OSmg anthrancene per m£
of toluene and approximately 0.2 mg of each derivative per m£ of toluene.
SAMPLING SYSTEM
Two glass impingers in series, each containing 40 mH of 2N HCl-2,4
dinitrophenylhydrazine, are used to collect exhaust samples for the analysis
of the aldehydes and ketones. A flow schematic of the sample collection
system is shown in Figure 1. The two impingers together trap approximately
98 percent of the carbonyl compounds. The temperature of the gas stream
is monitored by a thermocouple immediately prior to the dry gas meter. The
dry gas meter determines the total flow through the impinger during a given
driving cycle. The sample pump is capable of pulling a flow rate of 4 £/min.
A drier is included to prevent condensation in the pump, flowmeter, dry gas
meter, etc. The flowmeter allows continuous monitoring of the sample flow
to insure proper flow rates during the sampling. The T.eflon line connecting
the CVS and the solenoid valve is heated to ^170°F in order to prevent
water from condensing in the line. Several views of the sampling system
are shown in Figure 2.
226
-------
Gas Temperature
Digital Readout
IO
-J
On-Off
Solenoid
Valve
Flowmeter
Dry
Gas
Meter
Regulating
Valve
Dilute
Exhaust
Ice Bath
Temperature Readout
Gas Volume
Digital Readout
Figure 1. Aldehyde and Ketone sample collection flow schematic
-------
Front View
Close-up of Upper Front
Figure 2. Aldehyde and Ketone sampling system
228
Digital
Readout
Flowmeter
Regulating
Valve
-------
Solenoid
Impinger
Ice Bath
Close-up of Impingers (Side
Rear View
solenoid
Filter
Ice Bath
Drier
Dry Gas Meter
Pump
Figure 2 (Cont'd). Aldehyde and Ketone sampling system.
229
-------
ANALYTICAL PROCEDURE
The analysis of the aldehydes (formaldehyde, acetaldehyde, iso-
butyraldehyde, crotonaldehyde, hexanaldehyde, and benzaldehyde) and of the
ketones (acetone and methylethyIketone) in dilute exhaust is accomplished
by collecting these carbonyl compounds in a hydrochloric acid (HCl)/2,4
dinitrophenylhydrazine (DNPH) solution as their 2,4 dinitrophenylhydrazohe
derivatives. The derivatives are removed from the HC1/DNPH absorbing solu-
tion by filtration and/or extractions with pentane. The filtered precip-
tate and the pentane extracts are combined and the volatile solvents are
removed. The remaining extract contains the phenylhydrazone derivatives.
The derivatives are then dissolved in a quantitative volume of toluene con-
taining a knovnamount of anthracene as an internal standard. This solution
is analyzed by injecting a small volume of the solution into a gas chromat-
ograph equipped with dual flame ionization detectors. From this analysis
and the measured volume of exhaust sampled, the concentration of the
carbonyl compounds in exhaust can be determined. The analysis flow sche-
matic for the aldehydes and ketones is shown in Figure 3. A detailed
description of the procedure follows.
The aldehydes and ketones are trapped in solution by bubbling a known
volume of dilute exhaust through two glass impingers connected in series,
with each impinger containing 40 mi of a 2N HC1 solution saturated with
DNPH. The sampling temperature and barometric pressure are recorded during
this bubbling period. The carbonyl compounds in the exhaust react with the
DNPH to form slightly soluble or insoluble 2,4 dinitrophenylhydrazone deri-
vatives. The two impingers together collect 98 percent + of the carbonyls
that are present in the exhaust. The impingers are removed from the sampl-
ing cart and are allowed to stand at room temperature for at least one hour
before proceeding to the filtration and extraction steps. Figure 4 shows
two impingers containing the HCl/DNPH absorbing solution after being removed
from the sampling cart.
Under normal operating conditions the contents of the two impingers
are combined and analyzed as one sample. If either of the two impingers
contain a precipitate they are first subjected to a filtration step. If
no percipitate is present this filtration step is omitted and the extrac-
tion step, described later in the procedure, is the first step.
For the filtration step the contents of the two impingers are poured
through a fritted glass filter into a flask under vacuum (Figure 5). The
two impingers are rinsed with small portions of deionized water. This wash
water is also poured through the fritted glass filter. The precipitate in
the filter is then washed with a few m£ of deionized water. The fritted
filter is then removed from the flask containing the 80 mA of absorbing
reagent and the water washings. The flask is then set aside for the ex-
traction step. The fritted glass filter containing the precipitate is
connected to a dry flask. The two impingers that had previously contained
the filtered precipitate are then each washed with small portions of
methylene chloride. The methylene chloride dissolves any solid residue
which was not removed by the water wash. These methylene chloride washings
are poured into the fritted glass filter containing the precipitate. After
230
-------
CVS
glass impingers
derivative filtered
and extracted with
pentane from absorber
\
filtered ppt combined
with pentane extract
and solvent removed
extract dissolved in
toluene containing
anthracene as an
internal standard
sample analyzed
in gas chromatograph
with FID
A/D converter
recorder
Hewlett-Packard 3354
Computer System
Figure 3. Aldehyde and Ketone analysis flow schematic.
231
-------
Figure 4. Impingers containing HC1/DNPH
absorbing solution.
Figure 5. Filtration of absorbing solution,
232
-------
the precipitate has been dissolved by the methylene chloride, a vacuum is
applied to the flask and the methylene chloride solution is pulled through
the filter into the flask. Another small amount of methylene chloride is
poured through the filter into the flask to wash the filter. The methylene
chloride solution is now saved until the extraction step is complete.
The extraction step is carried out as follows. The contents of the
two impingers (if no precipitate is present) are transferred to a 250 m£
separatory funnel. The impingers are each washed with small portions of
deionized water which is also added to the separatory funnel. If a pre-
cipitate was found in the impingers the contents of the flask containing
the filtered absorbing reagent and the water washings from the filtration
step are transferred quantitatively to a 250 m& separatory funnel. The
flask is washed with a small portion of water, and this water is added to
the separatory funnel. Forty m£ of pentane is now added to the separatory
funnel containing the 80 mJl of absorbing reagent and water washings. The
funnel is stoppered and shaken for five minutes in an automatic shaker,
Figure 6. The shaker is stopped and the funnel is vented. After the two
phases are allowed to separate, the lower phase is collected in a second
separatory funnel. The remaining phase is transferred to a third 250 mil
separatory funnel. A second 40 mfc portion of pentane is added to the al-
ready once extracted absorbing solution. The funnel is again stoppered,
shaken for 5 minutes and vented. After the phases have separated the
lower phase is again collected in another separatory funnel. The upper
or pentane layer is combined with the pentane layer from the first extrac-
tion. A third 40 m& portion of pentane is added to the twice extracted
absorbing solution and the extraction process repeated. After the third
extraction the lower layer is discarded and the pentane layer is combined
with the pentane layers from the first two extractions. Any absorbing
solution which might have been accidently transferred with the pentane
layers is drained off. Deionized water (25-50 m£) and sodium bicarbonate
(1/4-1/2 gram) is added to the 250 m£ separatory funnel containing the 120
mi of pentane extract. The funnel is stoppered and manually shaken for
30 seconds. The phases are allowed to separate and the lower water phase
is drained off. Another 25 m£ of deionized water is added and the shaking
is repeated. After the phases have separated, the water is drained off
insuring that all traces of water are removed. The contents of the funnel
are then combined with the methylene chloride solution which was saved from
the filtration step.
The flask containing the methylene chloride solution and the pentane
extracts is then placed in a vacuum oven, Figure 7, operating at 50-60°C
and 65" water vacuum until the pentane and methylene chloride have been
removed. At this time only the dried phenylhydrazone derivatives remain.
Each time a series of samples are collected a blank containing 80 mJ,
of HC1/DNPH solution is extracted and dried in the same manner as the
samples. This accounts for any aldehydes or interfering compounds which
be found in the reagents used for extraction.
Two m£ of toluene which contain a quantitative amount of anthracene
.05 rag/hi toluene) as an internal standard is pipetted into the flask
233
-------
Figure 6. Automatic shaker.
Figure 7. Vacuum oven.
234
-------
containing the dried phenylhydrazone derivatives. The flask is then placed
in a sonic bath until all of the residue is dissolved. After the precipi-
tate has dissolved, the solution is transferred to a 1/2 dram vial. (See
Figure 8.) At this point the derivative is ready for injection into the
gas chromatograph system.
The gas chromatograph system used to analyze the toluene solution
containing the phenylhydrazone derivatives is shown in Figure 9. The sys-
tem consists of a Varian 1700 GC, an A/D converter, and a recorder. The
GC is equipped with dual columns and dual flame ionization detectors with
a single differential amplifier. The columns consist of 24 x i/e inch O.D.
stainless steel tubing packed with 6.7 percent Dexsil (polycarboranesiloxane)
300 GC on DMCS treated and acid washed, 60/80 mesh Chromosorb G. The
carrier gas is helium which flo^s through the columns at a rate of 40 mVmin.
The optimum hydrogen and air flow rates are 500 mi/min and 35 mil/min re-
spectively. The column temperature, after injection of the sample, is pro-
grammed from 120°C to 300°C at 8° a minute. In a chromatogram of a standard
sample (Figure 10) containing anthracene and the phenylhydrazone derivatives
of formaldehyde, acetaldehyde, acetone, iso-butyraldehyde, methylethyl-
ketone, crotonaldehyde, hexanaldehyde, and benzaldehyde, the first peak
eluted is toluene followed by anthracene, and then the derivatives of for-
maldehyde, acetaldehyde, acetone, iso-butyraldehyde, methylethyIketone,
crotonaldehyde, hexanaldehyde and benyaldehyde. The methylethylketone
derivative is added to the list of derivatives in order to name an unknown
peak found in some of the exhaust samples. Data obtained from the five
repetitive injections of the standard derivatives in toluene showed a max-
imum standard deviation of 4.56 percent for benzaldehyde and a minimum
standard deviation of 0.87 percent for formaldehyde. The computer print-
out of the standard, Figure 10, is shown in Figure 11. This printout gives
the retention time, area, and the name of each peak. The printout also
gives the concentration of each of the derivatives in mg/m£. This con-
centration is calculated by the computer from response factors which are
determined daily. Each day a standard containing known amounts of the de- .
rivatives and anthracene is injected into the GC. From the anthracene and
derivative areas the computer calculates a response factor F. These F
factors are used in all subsequent runs during the day to determine the
concentration of the derivatives. This response is calculated from the
following equation:
Anthracene Area x mg/m£ Derivative
Response Factor (F) = Derivative Area mg/m£ Anthracene
Typical response factors for each of the derivatives are listed below:
FACTOR NAME
1.0000 ANTHRACENE
3.1043 FORMALDEHYDE
2.7736 ACETALDEHYDE
2.2366 ACETONE
2.4160 ISO-BUTYRALDEHYDE
2.3332 METHYLETHYLKETONE
3.4174 CROTONALDEHYDE
235
-------
Figure 8. 1/2 dram vials.
Figure 9. Aldehyde and Ketone analytical system-
236
-------
- Injection
2. Toluene
3. Anthracene
4. Formaldehyde
ZZZZZZZZ 5. Acetaldehyde
6. Acetone
Isobutyraldehyde
8. Methylethylketone
9. Crotonaldehyde
-•- 10. Hexanaldehyde
ZZZZmrri" -'-•'•• Benzaldehyde
i
Figure 10. Chromatogram of standard,
237
-------
REPORT: 14.11 CHANNEL: 11
SAMPLE: RCI INJECTED AT 11:18:27 OM MAR 1» 1978
I STD METHOD: DNPH1 1
ACTUAL RU
-------
2.3428 HEXANALDEHYDE
2.9329 BENZALDEHYDE
When the response factor is known a concentration in mg/m£ for each of the
derivatives can be found. This concentration, along with the volume of
sampled exhaust is then used to calculate the concentration of the carbonyl
compounds in exhaust. Figures 12 and 13 show a typical sample chromatogram
and accompanying printout respectively.
CALCULATIONS
This procedure has been developed to provide the user with the concen-
trations of the aldehydes (formaldehyde, acetaldehyde, isobutyraldehyde,
crotonaldehyde, hexanaldehyde, and benzaldehyde) and ketones (acetone and
methylethylketone) in exhaust. The results will be expressed in yg/m3 of
exhaust and ppm for each carbonyl compound. The equations for determing
the concentrations in yg/m3 and ppm are derived in the following manner.
The first step is to correct the volume of exhaust sampled to a stand-
ard temperature, 68 °F and pressure, 29.92" Hg, by use of the equation
P v V P V
exp x exp _ corr x corr
T ~ T
exp corr
= experimental volume of gas sampled in ft3
= volume of gas sampled in ft3 corrected to 68°P and
29.92" Hg
pexp = experimental barometric pressure
Pcorr = 29.92" Hg
Texp = experimental temperature in °F + 460
Tcorr = 68°F + 460 = 528°R
Solving for VCorr gives:
P ("Hg)X V (ft3) x 528°R
V _ exp exp
corr - T (OR) 29.92" Hg
exp
The next step converts the volume from cubic feet to cubic meters by
use of the conversion factor 1 cubic meter is equal to 35.31 cubic feet.
P ("Hg) x V (ft3) x 528
v 3 exp exp
corr(m ) = - 29.92" Hg x 35.31 ftV
exp
(Equation 1)
The next step converts the mg/m£ of derivative determined by the
computer to mg of carbonyl collected in the two impingers. To obtain mg
of derivative, the concentration (from the computer printout) in mg/m2, is
multiplied by the volume of toluene used to dissolve the -did extract.
239
-------
1. Injection
2. Toluene
3. Anthracene
4. Formaldehyde
5. Acetaldehyde
6. Isobutyraldehyde
7. Methylethylketone
8. Crotonaldehyde
9. Hexanaldehyde
10. Benzaldehyde
24
22
20
18
14
Figure 12. Sample chromatogram.
240
-------
R El-OKI I 2kJ CHANNEL.: 11
SAMPLE: RCI IMJECTED AT . Jb: 41:05 3M MAR 1» 1978
I STD METHJD: DrtPHl1
ACTUAL
I SI U- RATIO:
."1HUTES
053**
STD-4MT
0500
SAMP-A^Ti
1.'302)0
RT
1 J'lAL A
AREA
•M G/ ML
7. 1 5
7. 95
0. 43
1 .88
2.83
3. 78
4. Pb
5.95
6. 77
1 9. 23
20.23
23. 52
2 S. .•? 3
25. 48
8604
435
88 77
575
463
1 594
2633
1 46
675
648
1 912
21 7
1 3
84
BV
VB
BB
BV
VV
VV
VV
VV
VV
VB
BV
VV
VB
3B
.033
. 1 86
.010
.007
.022
.053
.002
.004
. 01 2
• 01 1
.001
7. 6£- 5
4. 9E- 4
= 26874
DATA FILE: *PRC1 1
•MA^IE
#F'J R.I ALDEHYDE
#ACETALDEHYDE
#ISO-8UTYRALDEHYDE
#CRQ1J>M ALDEHYDE
#H£XANlALDEHYDE
T'JTAL MG/ML = .310
RAW DATA FILE: * RA Iwl 1
Figure 13. Computer printout of sample.
-------
mg derivative = ConcDer (mg/m£) >
To find mg of carbonyl compound per sample the mg of derivative are
multiplied by the ratio of the molecular weight of the carbonyl derivative
over the molecular weight of its phenylhydrazone derivative.
_ mol. wt. carbonyl
mg carbonyl = mg derivative x mol> wt- derivative
= Cone (mg/m£) x Vol ., (m£) x mol. wt. carbonyl
Der Tol moi. wt. derivative
To obtain the number of yg of carbonyl compound the mg of carbonyl are
multiplied by the conversion factor, 1000 yg/mg
0 Mol. Wt. carbonyl
yg carbonyl = ConcDer (mg/mtt x VolTol (mfc) x Mol> Wfc_ derivat-ive
x 1000 yg/mg
(Equation 2)
Tne concentration of the carbonyl compound in exhaust can now be found
in yg/m3 by dividing equation 2 by equation 1.
Cone (.mg/m&) x Vol (.m£) x mol. wt. carbonyl
1000 yg/mg XT (°R) x 29.2?" Hg x 35.31 ft3/m3
mol. wt. derivative
(Equation 3)
To find the concentration of each carbonyl compound in ppm, the den-
sities of carbonyls are needed. At 29.92" Hg and 32°F, one mole of gas
occupies 22.4 liters. This volume is corrected to 68 °F from the equation.
V Vx
T ~ T^
V]_ = 22.4
^ = 32°F +460 = 492°R
V = volume at 68 °F
T = 68°F + 460 = 528°R
Solving for V gives:
V x T 22.4 x 528
242
-------
Since one mole of gas occupies 22.045- at 68°F, the denisty can be found in
g/£ by dividing the molecular weight in g/mole by 24.04 £/mole
den ta/i) = mol. wt. (g/mole)
den (g/Jt) 24.04Vmole
The density in yg/m£ can be found by converting g to yg and yg and £ to
m£ as follows:
den uo/mt = mo1- wt. g/mole 1* 106 yg/g mol. wt. x IQQQ
den yg/m* 24.04£/mole 1 x 10^ m£/£ * 2^04
(Equation 4)
To obtain the concentration of each carbonyl in ppm, the concentration in
yg/m3 is divided by the density in yg/m£
ppm = yg/m3 T yg/m£ = ^
Using Equations 3 and 4 gives the ppm concentration in the form of the
raw data.
ppm =
Cone (mg/m£) x vol (m£) x mol. wt. carbonyl x iQOO yg/mi
U63T J.OJL
P ("Hg) x v (ft3) x 528° x mol. wt. derivative
exp exp
T (°R) x 29.92" Hg x 35.31 ft3/m3 x 24.04 Vmole
x -S5E -
mol. wt. carbonyl x 1000
Con_ (mg/mJl) x vol . (m£) XT (°R) x 29.92" Hg
_ DSIT TOl _ 6Xp _ . _
P ("Hg) x y (ft3) x 528°
exp exp
v 35.35 ft3/m3 x 24.04 £/mole
S\ ^ ^ ^^H^^^.^^iB^M^^i^^^BMM^fc^^H^^^M^^WIfcrtllW^Kfc*"*^^**^^-^**^"""^™^"^^^"
mol. wt. derivative
(Equation 5)
At this point, the concentration can be express in yg/m3 (Equation 3) and
ppm (Equation 5) at 68°F and 29.92" Hg f rom the raw data.
Hewlett-Packard Calculations
In order to insure maximum turnaround in a minimum time period, two
Hewlett-Packard 65 programs were developed. One calculates the aldehyde
and ketone concentrations in yg/m3 from the raw data and phenylhydrazone
derivative concentrations (from computer printout) . The other program cal-
culates the concentrations in ppm from the concentrations in yg/m3. These
programs are presented in Figures 14 and 15.
243
-------
HP-65 Program Form
Title.
Page.
.of.
SWITCH TO W/PRGM PRESS ; I I PBGM , TO CLEAR MEMORY
KEY
ENTRY
LBL
A
0
•
0
0
0
5
X
KR/S
X
ST01
R/S
4
6
0
+
RCL1
•^«
a x v
20- ~
g 1/x
R/S
X
STO2
R/S
RCL2
X
0
.
30 I
4
3
X
R/S
RCI, 2
X
0
•
1
JO 9
6
X
R/S
RCL 2
X
0
•
2
4
?•: 4
CODE
SHOWN
23
11
83
71
84
71
3301
84
61
3401
3507
81
3504
84
71
3302
84
3402
71
83
71
84
3402
71
83
71
34
3402
71
83
COMMENTS
Input sample vol., ft3
Input Barometer, "HQ
Input sample Temp °F
Input Vol. Toluene, ml
In ma /mi
Out ug/m3; In mg/ml
Out ua/m3: Tn ma /ml
KEY
ENTRY
X
R/S
RCL2
X
0
•
2
8
6
60 X
R/S
RCL2
X
0
•
2
8
6
X
7c R/S
RCL2
X
0
.
2
3
0
X
R/S
3*CL2
X
0
,
3
5
9
X
R/S
RCL 2
30 X
0
•
3
7
1
X
R/S
RTN
'CO
CODE
SHOWN
71
84
3402
71
83
71
84
3402
71
83
71
34
3402
71
83
71
84
3402
71
83
71
84
3402
71
83
71
84
24
COMMENTS
Out vig/m3; in mg/ml
Out yg/m3; in mg/ml
Out ug/m3; In mg/ml
-'•
Out |jq/m3; In mg/ml
Out ug/m3; in mg/ml
Out u g/m3
REGISTERS
Ri
R7
R3
R4
Rs
R8
R7
Ra
Ro
LABELS
A
R
C
D
F
n
1
7
3
4
5
R
7
R
Q
FLAGS
1
•)
TO RECORD PROGRAM INSERT MAGNETIC CARD WITH SWITCH SET AT W PRGM
Figure 14. Aldehyde and Ketone concentrations in
244
-------
HP-65 User Instructions
Programmer .
s
D 1.
JL
"i
i i mi
/
n \ \
___ raye nr
Datp
STEP
01
°2
01
°4
1
2
3
4
5
6
7
8
9
in
•
11
12
INSTRUCTIONS
Switch to on; switch to run
Feed card in from right to left
Initialize
Set decimal place
Input - sample' volume, ft3
Input - barometric pressure, "Hg
Input - Sample Temperature, °F
Input - Volume Tolune, ml
Input - Cone. Formaldehyde Der, mg/ml
Output - Cone. Formaldehyde Ug/m3
Innut - Cone. Acetaldehvde Der. ma/ml
Output — Cone. Acetaldehyde Ug/m
Input - Cone. Acetone Der, mg/ml
Output - Cone. Acetone, Ug/m
Input - Cone. Isobutyraldehyde Der, mg/ml
Output - Cone. Isobutyraldehyde, ug/m3
Input - Cone, methylethylketone Der, mg/ml
Output - Cone, methylethylketone, Ug/m3
Input - Cone. Crotonaldehyde Der, mg/ml
Output - Cone. Crotonaldehyde, Ug/m3
Input - Cone Hexanaldehyde Der, mg/ml
Output - Cone Hexanaldehyde , U g/m3 ___
Tnrmt- - rr\r\r- nar>* = 1^,=Vn/^o nor ma /ml
INPUT
DATA/UNITS
KEYS
1 II *
1 1
f II REG
DSP II 2 1
A || I
R/S ||
R/s II 1
H/S || |
H/S II
II
R/S II
II
R/S II
r ii i
R/S II
II
R/S II
II
„* II
II
.,» II
II
K/S J|,
II
JTM Jl
f II
OUTPUT
DATA/UNITS
(cont'd.)
Figure 14 (Cont'd). Aldehyde and Ketone concentrations in yg/m£.
245
-------
HP-65 User Instructions
! me —
Proor sm m^r
/ „ .,
(i — |_j S 1 ii
Hate
L '
0
I 1 i..__j.._«i
STEP
01
°2
°T
04
1
2
3
4
5
6
7
8
INSTRUCTIONS
Switch to on; switch to run
Feed card in from right to left
Initialize
Set decimal place
Input-cone Formaldehyde, |jg/m3
Ouput-cono Formaldehyde, ppm
Input-Cone Acetaldehyde, Ug/m
Output-Cone Acetaldehyde, ppm
Input-Cone Acetone, iig/m3
Output-Cone Acetone, ppm
Input-Cone Isobutyraldehyde, Ug/m3
Output-Cone Isobutyraldehyde, ppm
Input-Cone Methylethylketone, ua/m3
Outcut-Conc Methylethylketone, com
Incut -Cone Crotonaldehvde . Ua/m3
Output-Cone Crotonaldehyde , ppm
Input-Cone Hexanaldehyde , Ug/m3
Output-Cone Hexanaldehyde, ppm
Input-Cone Benzaldehyde , yg/m3
Output-Cone Benzaldehyde, ppm
INPUT
DATA/UNITS
KEYS
1 II
1 II
f f ][ REG ]
I DSP || 2
1 , II
1 II 1
.» II
II
_ II
II
R/S II
II
R/S ||
II
L R/S II
II
R/S II
1 H
»,« II
H
RTN ||
f II
REG ||
CLXII
II
II
OUTPUT
DATA/UNITS
Figure 15. Aldehyde and Ketone concentrations in ppm.
246
-------
HP-65 Program Form
Title.
Page of.
SWITCH TO W/PROM. PRESS I j PRGM I TO CLEAR MEMORY.
KEY
ENTRY
LBL
A
1
2
4
9
R/S
10 1
8
3
2
4
R/S
2
4
1
206
T
R/S
3
n
0
0
R/S
30
3
0
0
0
±
R/S
2
9
JOT
6
R/S
4
1
6
6
i
50 R/S
9320-06-6
CODE
SHOWN
23-
11
81
84
81
84
81
84
81
84
81
84
81
ftd
81
84
COMMENTS
„. ,..,, ._
Out cone ppm
In cone ug/m3
-•
Out cone ppm
In cone uq/m3
Out cone ppm
In cone uq/m3
Out cone ppm
In cone ug/m3
•
Out cone ppm
In cone yg/m3
In COPG U Q/rti-^
KEY
ENTRY
4
4
1
5
i
R/S
RTN
60
70
30
9C
100
CODE
SHOWN
81
84
24
COMMENTS
In cone pg/m^
Out cone ppm
REGISTERS
Ri
R?
Ra
R4
R5
Re
R7
Ra
RQ
LABELS
A
B
C
D
P
0
1
•>
3
4
S
6
7
S
a
FLAGS
1
9
TO RECORD PROGRAM INSERT MAGNETIC CARD WITH SWITCH SET AT WPOGM.
Figure 15 Cont'd). Aldehyde and Ketone concentrations in ppm.
247
-------
Sample Calculation
Assiime exhaust samples were collected in glass impingers for each por-
tion of a three bag 1975 FTP. Raw data for these tests is presented in
Figure 16. Calculations were performed using the HP-65 programs and manual
calculations .
Manual calculation for driving cycle FTP-1:
Cone (mg/m£) x Vol , (m£) x mol. wt. carbonyl
Per Tol J
. 3 , ., , , e
yg/mj formaldehyde = - - - (ll } x
exp exp
1000 yg/mg XT (°R) x 29.92" Hg
x _ exp _ _
528°R
„ 35.31 ft3/m3
mol. wt. derivative
0.186 mg/m£ x 2ml x 30.03 g/mole x IQQQ yg/mg
29.80" Hg x 3.196 ft3 x 528°R
v 535°R x 29.92" Hg x 35.31 ft3/m3
X "" '••*— II II • "I I •' -'" P^^— .^ . -I.- I— • I. I II I- !-•
201.15 g/mole
= 597.5 yg/m3
ppm formaldehyde = yg/m3 T density yg/m£
, . . , „ mol. wt. (formaldehyde) x 1000
densxty yg/m£ = - 24.04£ -
mol. wt. formaldehyde = 30.03 g/mole
, _ 30.03 g/mole x 1000 ,0.Q ,, .0
density = - ^ - = 1249 yg/mJt
ppm = 597.5 yg/m3 4- 1249 yg/mJl = 0.478 m£/m = 0.478 ppm
The calculations for acetaldehyde, acetone, isobutyraldehyde , methylethy-
ketone, crotonaldehyde , hexanaldehyde , and benzaldehyde are carried out in
the same manner by substituting the appropriate derivative concentrations
and molecular weights into the above formulas . These calculations give the
following concentrations :
acetaldehyde, 561 yg/m3 and 0.306 ppm
acetone, 663 yg/m3 and 0.274 ppm
isobutyraldehyde, 141 yg/m3 and 0.047 ppm
methyle thy Ike tone, 630 yg/m3 and 0.210 ppm
crotonaldehyde, 541 yg/m3 and 0.186 ppm
hexanaldehyde, 250 yg/m3 and 0.060 ppm
248
-------
SWRI PROJECT NO.
FUEL: CVS SO.
SAMPLE COLLECTION BY:_
GENERAL COMMENTS:
TEST NO.
TUNNEL SIZE:
_TEST DATE:_
DRIVER:
CHEMICAL ANALYSIS BY:
VEHICLE:
MILES:
CALCULATIONS BY:
Test No.
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Vol. Toluene ml
Formaldehyde Der Cone mg/ml
Formaldehyde Cone yg/m3
Formaldehyde Cone ppm
Acetaldehyda Der Cone mg/ml
Acetaldehyde Cone yg/m3
Acetaldehyde Cone ppm
Acetone Der Cone mg/ml
Acetone Cone yg/m3
Acetone Cone ppm
I-Bu Aldehyde Der Cone rag/ml
I-Bu Aldehyde Cone yg/m3
I-Bu Aldehyde Cone ppm
MeEt Ketone Der Cone mg/ml
MeEt Ketone Cone ug/m3
MeEt Ketone Cone ppm
Cro-Aldehyde Der Cone mg/ml
Cro-Aldehvde Cone yq/m3
Cro-Aldehyde Cone ppm
Hex-Aldehyde Der Cone mg/ml
Hex-Aldehyde Cone ug/m3
Hex-Aldehyde Cone ppm .
Benzaldehyde Der Cone mg/ml
Benzaldehyde Cone yg/m3
Benzaldehyde Cone ppm
FTP-1
3.196
29.80
75 •
2
0.186
598
0.479
0.127
559
0.305
0.121
663
0.274
0.022
141
0.047
0.098
630
0.210
0.086
541
0.186
0.031
250
0.060
0.093
775
0.176
FTP-2
U625
30.02
80
2
0.105
665
0.532
0.092
798
0.436
0.098
1060
0.439
0.011
139
0.046
0.084
1060
0.353
0.074
917
0.314
0.018
286
0.069
0.081
1330
0.301
FTP- 3
2.010
29.02
96
2
0.201
1100
0.881
0.157
1170
0.639
0.161
1500
0.621
0.028
305
0.102
0.097
1060
0.353
0.076
811
0.278
0.030
411
0.099
0.097
1370
0.310
SET-7
3.730
29.25
85
2
0.312
891
0.713
0.282
1100
0.600
0.285
1390
0.575
0.023
131
0.044
0.198
1130
0.377
0.105
587
0.201
0.027
194
0.047
0.121
897
0.203
HFET
8.241
29.95
83
2
0.732
921
0.737
0.612
1060
0.579
0.595
1280
0.530
0.051
128
0.043
0.252
634
0.211
0.286
705
0.242
0.078
246
0.059
0.232
757
0.171
NYCC
1.070
29.50
89
2
0.142
1410
1.130
0.102
1390
0.759
0.105
1780
0.737
0.009
179
0.060
0.075
1490
0.497
0.072
1400
0.480
0.011
275
- 0.066
0.081
2090
0.473
Figure 16. Aldehyde Collection Sheet.
249
-------
Benzaldehyde, 775 g/m3 and 0.176 ppm.
Note: The values used in these calculations are picked from a range of
temperatures, derivative concentrations, etc. to validate the calculations
and may not be representative of expected raw data. The calculations are
presented to confirm the manual and HP-65 calculations give the same re-
sults. This was confirmed for six sets of calculations.
REFERENCES
This procedure is taken from the procedure: "Oxygenated Compounds in
Automobile Exhaust-Gas Chromatograph Procedure" by Fred Stump, ESRL,
Environmental Protection Agency, Research Triangle Park, North Carolina.
250
-------
APPENDIX C
ORGANIC AMINE PROCEDURE
251
-------
THE MEASUREMENT OF ORGANIC AMINES IN EXHAUST
The measurement of organic amines (monomethylamine, dimethylamine,
trimethylamine, monoethylamine, diethylamine, and triethylamine) in exhaust
is accomplished by bubbling the exhaust through glass impingers containing
dilute sulfuric acid. The amines are complexed by the acid to form stable
sulfate salts which remain in solution. The exhaust sample is collected
continuously during the test cycle. A portion of this solution is injected
into a gas chromatograph equipped with an ascarite loaded pre-column and a
nitrogen phosphorus detector (NPD). External amine standards in dilute
sulfuric acid are used to qualify the results. Detection limits for this
procedure are on the order of 0.01 ppm-0.05 ppm.
LIST OF EQUIPMENT AND REAGENTS
The equipment and reagents for the analysis of the organic amines are
divided into two groups. The first involves the sample acquisition and the
second the instrumental analysis of the sample once it has been obtained.
Manufacturer, stock number and any pertinent descriptive information are
listed.
Sampling
1. Glass impingers, Ace Glass Products, Catalog #7530-11, plain tapered
tip stoppers with 18/7 arm joints and 29/42 bottle joints.
2. Flowmeter, Brooks Instrument Division, Model 1555, tube size R-2-15-C,
graduated 0-15, sapphire float, 0-5 1/min range.
3. Sample pump, Thomas Model 106 CA18, capable of free flow capacity of
4 1/min.
4. Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
capacity.
5. Regulating valve, Nupro 4MG, stainless steel.
6. Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8" ID
and 5/16: OD x 1/8" ID.
7. Teflon solenoid valve, The Fluorocarbon Company, Model DV2-144NCA1.
8. Drying tube, Analabs Inc., Catalog #HGC-146, 6" long, 1/4" brass
fittings.
252
-------
9. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, connectors,
etc.
10. Digital readout for dry gas meter.
LI. Miscellaneous electrical switches, lights, wirings, etc.
12. Six channel digital thermometer, Analog Devices, Model #2036/J/1.
L3. iron/Constantan type J single thermocouple qith 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation.
L4. 30 ml polypropylene sample storage bottles, Nalgene Labware, Catalog
#2006-0001.
L5. Sulfuric Acid, H2SO4,•formula weight - 98.08, Certified IN by Fisher
Scientific Company, #SO-A-212.
16. Class A, 10 IP £ volumetric pipet.
L7. Class A, 1000 ml volumetric flask.
Instrumental Analysis.
1. 10 yl syringe, Pressure-Lok, Precision Sampling Corporation.
2. Parkin-Elmer Model 3920 B gas chromatograph equipped with an ascarite
loaded pre-column, a teflon interface, and a nitrogen phosphorus
detector (NPD) .
3. Soltec Model B-281 1 mv recorder.
4. Hewlett-Packard Model 3354 gas chromatograph computer system with
remote teletype printout.
5. Glass insert for 1/8" and 1/4" heated injectors (used as pre-column) ,
Perkin-Elmer, #009-1958.
6. Ascarite, 20-30 mesh, Arthur H. Thomas Company, Catalog #C049-U86.
"I. Teflon tubing (used for interface in gas chromatograph) , Analabs
Inc., 1/6" OD x 0.03" ID, #HGC-024.
8. Methylamine Hydrochloride, CH3NH2-HC1, formula weight = 67.52,
crystals, Eastman #116.
9- Dimethylamine Hydrochloride, (CH3) 2NH-HCl, formula weight = 81.55,
crystals, Eastman #94.
iO- Trimethylamine Hydrochloride, (CH3)3N-HCl, formula weight =95.57,
crystals, Eastman #265
253
-------
11. Ethylamine Hydrochloride, C2H5NH2iHCl, formula weight =-- 81.55,
crystals, Eastman #731
12. Diethylamine Hydrochloride, (C2H5)2NH-HC1, formula weight = 109,6,
crystals, Eastman #2090
13. Triethylamine Hydrochloride, (C2H5)3N-HC1, formula weight = 137.65,
crystals, Eastman #8535.
Preparation of Absorbing Solution
The absorbing solution (0.01 N H2S04) is prepared by diluting 10 ml of
1 N sulfuric acid (certified Fishcer Scientific Company) to 1 liter with
deionized water.
Preparation of Ascarite Pre-Column
The ascarite pre-column is prepared by packing a Perkin-Elmer glass
insert for a heated injector block with 20-30 mesh ascarite and plugging
both ends with 1/4" of glass wool. The packed pre-column is shown in Figure
1. This packed pre-column is then inserted into the heated injector and
held in place against the column with a spring loaded metal tube and the
septum cap.
Preparation of Primary Standards
The primary standards for the organic amines are prepared by dissolving
a weighed amount of the amine-hydrochloric acid salt in 0.01 N sulfuric acid
and diluting the resulting mixture to the proper volume with 0.01 N sulfuric
acid. Standards less that 10 ppm are prepared by diluting higher concen-
traion standards with 0.01 N sulfuric acid.
SAMPLING SYSTEM
A glass impinger containing 25 ml of 0.01 N sulfuric acid is used to
collect exhaust samples for the analysis of the organic amines. A flow
schematic of the sample collection system is shown in Figure 2. The single
glass impinger is sufficient to collect 99 percent+ of the organic amines.
The temperature of the gas stream is monitored by a thermocouple immediately
prior to the dry gas meter. The dry gas meter determines the total flow
through the impinger during a given driving cycle. The sample pump is
capable of pulling a flow rate of 4 liters/minute. A drier is included to
prevent condensation in the pump, flowmeter, dry gas meter, etc. The flow-
meter allows continuous monitoring of the sample flow to insure proper flow
rates during the sampling. Several views of the sampling system are shown
in Figure 3.
ANALYTICAL PROCEDURE
The analysis of the organic amines (monomethylamine, dime thy lamine,
trimethylamine, monoethylamine, diethylamine, and triethylamine) is ac-
complished by trapping the amines in sulfuric acid and analyzing the sample
254
-------
,.; <. o '*•**
Figure 1. Ascarite pre-column.
255
-------
Gas Temperature
Digital Readout
tvj
Cn
IQIBIOI
Flowmeter
On-Off
Solenoid
Valve
Sample
X
Sample \ Heated
Probe Line
Dilute
Exhaust
Dry
Gas
Meter
#
Regulating
Valve
Ice Bath
Temperature
Digital Readout
Gas Volum©
Digital Readout
Figure 2. Organic amines sample collection flow schematic,
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 3. Organic Amines sampling system.
257
-------
Solenoid
Impinger
Ice Bath
Close-up of Impingers (Side View)
olenoid
'"ilter
Ice Bath
Pump
Rear View
Figure 3 (Cont'd). Organic Amines sampling system.
258
-------
with a gas chromatograph equipped with a NPD. The NPD is highly sensitive
to organic nitrogen compounds and relatively insensitive to inorganic ni-
trogen compounds. The analysis flow schematic for the organic amines is
shown in Figure 4. A detailed description of the procedure follows.
For the analysis of the organic amines, dilute exhaust is bubbled
through a single glass impinger containing 25 m£ of 0.01 N sulfuric acid.
fhis sulfuric acid solution traps 99 percent+ of the organic amines as their
sulfate salts. The temperature of the impinger is maintained at 0-5°c by an
ice water bath, and the flow rate through the impinger is maintained at 4
&/min. Upon completion of each driving cycle, the impinger is removed and
the contents are transferred to a 30 mA polypropylene bottle and capped.
The amines, as their sulfate salts, can be stored in solution for long peri-
ods of time without decomposition.
A Perkin-Elmer 3920B gas chromatograph equipped with an ascarite loaded
pre-column, a teflon interface, and a nitrogen phosphorus detector (NPD) is
used to analyze the sample. A 10 y£ portion of the sample is injected into
the gas chromatograph (GC) . In the ascarite pre-column the amines are re-
leased from their sulfate salts into the GC column. The column is a 6" X
4 mm glass column containing Carbopack B coated with 4 percent carbowax
20 M and 0.8 percent KOH. The column effectively separates the amines, with
the exception of ethylamine and dimethylamine which are reported together as
total C2H7N. The carrier gas is helium which flows through the column at a
rate of 30 m£/min. The column temperature is 130°C for 4 minutes and then
programed to 170°C at a rate of 32° a minute. In a chromatogram of a stand-
ard sample containing all six of the amines, Figure 5, the first peak is
momomethylamine, followed by the combined peak of dimethylamine and mono-
ethylamine, C2H7N, and then by trimethylamine, diethylamine, and triethy-
lamine. The peak areas are compared to standard solutions. Figure 6 shows
the analytical system with gas chromatograph, detector, integrator, and
recorder.
CALCULATIONS
This procedure has been developed to provide the user with the con-
centration of the organic amines (monomethylamine, total dimethylamine and
monoethy lamine as 02%^ trimethy lamine, diethylamine and trie thy lamine) in
exhaust. The results will be expressed in yg/m3 of exhaust and ppm for each
of the amines. The equations for determining the concentrations in yg/m
and ppm are derived in the following manner.
The first step is to correct the volume of exhaust sampled to a stand-
ard temperature, 68°F, and pressure, 29.92"Hg, by use of the equation
P V P V
exp x vexp _ corr x corr
T T
exp corr
259
-------
CVS
Glass
Impinger
Unused sample
saved as needed
Sample analysis
in gas chromatograph
with ascarite pre-column
and NPD
A/D
converter
Recorder
Hewlett-Packard
3354
computer system
Figure 4. Organic Amines Analysis Flow Schematic.
260
-------
Sample injection
1.4 ppm methylamine
1.6 ppm ethylamine +
dimethylamine
1.5 ppm trimethylamine
0.9 ppm diethylamine
0.9 ppm triethylamine
t i i—
L2 io
Figure 5.
~~&~ 6 4
Retention time, rain.
Chromatogram of amine standard.
261
-------
' •
•
'
A/D
NPD Detector Control UnitJ Converter " Recorder
9
Figure 6. Total amine analytical system.
-------
Vexp = experimental volume of gas sampled in ft3
Vcorr = volume of gas sampled in ft3 corrected to 68°F and 29 92"Hq
pexp = experimental barometric pressure
Pcorr = 29.92"Hg
Texp = experimental temperature in °F + 460
Tcorr = 68°F + 460 = 528°R
Solving for Vcorr gives:
pexp ("H9> x Vexp (ft3) x 528°R
Vcorr ~ ~~ ~~—
Texp (°R> x 29.92"Hg
The next step converts the volume from cubic feet to cubic meters by
use of the converstion factor 1 cubic meter is equal to 35.31 cubic feet.
vexp (ffc3) x 528°
Vcorr (m3) =
TeXp x 29.92"Hg x 35.31
(Equation 1)
The next step is to find the concentration of each of the amines in
Ug/mJl. Since the gas chromatograph NPD has a linear response in the con-
centration of concern, then the following equation holds.
csam d-'g/m^) cstd (W/i^)
Asam Astd
Csam = concentration of the sample in yg/m&
Asam = GC Peafc area of sample in relative units
Cgt(j = concentration of the standard in iag/m£
Astd = GC- Pea^ area of standard in relative units
Solving for Csam gives:
cstd (V9/m£) x Asam
Astd
-ocuii ..-=,,-.-, in solution is corrected for any necessary dilution by
multiplying by the dilution factor, D.F.
cstd (yg/m£)
Astd
263
-------
To obtain the total amount in yg of each amine in the absorbing
solution, the absorbing reagent volume is multiplied by the concentration
to give:
yg sample = Csam (yg/m£) x Abs. Vol. (m£)
Cstd (yg/m£) x Asam x D.F. x Abs. Vol. (m£)
(Equation 2)
To obtain yg sample/m3, Equation 2 is divided by Equation 1 to give:
Cstd (^/m£) X ASam x D-F- x Abs. Vol. (m£)
yg samp/m3=
Astd>
-------
The density in yg/m£ can be found by converting g to yg and £ to m£ as
follows:
mol. wt. g/mole 1 x I06ug/g mol. wt. x
den yg/m£ =
24.04 Jl/mole 1 x 103m£/£ 24.04
(Equation 4)
To obtain the concentration of each amine in ppm, the concentration in
yg/m3 is divided by the density in yg/m£
, i . . „ m£
ppm = yg/mj - yg/m£= ^-j
Using Equations 3 and 4 gives the ppm concentration in the form of the raw
r?a+-a
24.04(£) x cstd (yg/m£) x Asam x D.F. x Abs. Vol. (m£)
Mol. Wt. (g/mole) x iQOO x Astd x Pexp ("Hg)
Te (°R) x 29.92"Hg x 35.31 ft3/m3
x
data.
528°R x yexp (ft3)
(Equation 5)
At this point, the concentration can be expressed in yg/m3 (Equation 3)
and ppm (Equation 5) at 68°F and 29.92"Hg from the raw data.
Hewlett-Packard Calculations
In order to insure maximum turnaround in a minimum time period a
Hewlett-Packard 65 program was developed to calculate the organic amine
concentrations in yg/irr and ppm from the raw data. This program is
presented in Figure 7.
Sample Calculation
Assume exhaust samples were collected in glass impingers for each
portion of a three-bag 1975 FTP. Raw data for these tests are presented
in Figure 8. Calculations were performed using the HP-65 program and
manual calculations.
265
-------
HP-65 Program Form
Title.
ORGANIC AMINES IN EXHAUST
Page.
.of.
SWITCH TO W,PBGM. P«£S5 H7N
Out Cone }$• T2H7M
KEY
ENTRY
5
•f
R/S
RCL 2
X
R/S
*
R/S
60 X
R/S
2
4
5
9
T
R/S
RCL 2
70 X
R/S
i
R/S
X
R/S
3
0
4
2
80 T
R/S
RCL 2
X
R/S
T
R/S
X
R/S
90 4
2
0
9
f
R/S
RTN
100
CODE
SHOWN
81
84
3402
71
84
81
84
71
84
81
84
3402
71
84
81
84
• 71
84
81
84
3402
71
84
31
84
71
84
81
84
24
COMMENTS
Out Cone ppm C?H7N
In Std Cone yg/ml (CH^) •}]
In Std Area (CH3)3N
In Samp Area (CH3) 3N
Out Cone Uq/m3(CH3)3N
Out Cone ppm (CH-^N
In Std Cone lag/ml (C2Hs)
-
In Std Area (C-HUKNH
* ~* *"
In Samp Area (C2H5)2NH
Out Cone wg/nr* (C7HS) 2NH
Out Cone ppm (C2Hs) 2NH
In Std Cone Ug/ml (C2Hs
In Std Area (C,H5) 3N
In Samp Area (C?Hs) 3N
Out Cone ua/m3 (C2H5)3N
Out Cone ppm (C^HO 3N
REGISTERS
Rl.
R2
R3
R4
Rs
R6
m
R7
Ra
RQ
LABELS
A
B
C
3*
E
o
1
2
3
A
S
6
7
8
Q
FLAGS
1
9
TO RECORD PROGRAM INSERT MAGNETIC CARD WITH SWITCH SET AT W ROOM
Figure 7. HP-65 program form.
266
-------
HP-65 User Instructions
Title _
Programmer.
Page_
Date.
.of.
STEP
Ol
02
°3
04
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IS
16
INSTRUCTIONS
Switch to on; Switch to run
Feed card in from right to left
Initialize
Set decimal place
Input Sample Volume, ft3
Input - Barometric Pressure, "Hg
Input - Sample Temperature, °F
Input - Dilution Factor
Input - Absorbing Reagent Vol . , ml
Input - Standard Cone, ug/ml CH3NH2
Input - Standard area CH3NH2
Input - Sample Area CH3NH2
Output - Sample Cone, yg/m3 CH3NH2
Output - Sample Cone, ppm CH3NH2
Input - Standard Cone, ug/ml C2H7N
Input - Standard Area C2H7N
Input - Sample Area C2H7N
Output - Sample Cone, yg/m3 C2H7N
Output - Sample Cone, ppm C2H7N
Input - Standard Cone, yg/ml (013 )3N
Input - Standard Area (CH,) ,N
Input - Sample Area (CH-,) ,N
Output - Sample Cone, yg/m3 (CH3) 3N
Output - Sample Cone, ppm (CHg^N
Input - Standard Cone, yg/ml (C-^Hs) ;NH
Input - Standard Area (C2Hs) ?NH ____
INPUT
DATA/UNITS
KEYS
1 II 1
1 11 1
1 f il REG 1
1 DspHfT"!
A ||
R/S^1|
R/s II
R/S II
R/S II
R/S II
R/S II
R/S II
R/S II
II
.« II
*» II
1 R/S II 1
rani i
! II I
1 R/S II
i „* ii i
i ,/, ii
„„ II
i ii
I,™ ir ,.,
1 ,/c II
OUTPUT
DATA/UNITS
Figure 7 (Cont'd). HP-65 user instructions.
267
-------
HP-65 User Instructions
/
u — L
i
/
in i i
rute
1
STEP
17
18
19
20
INSTRUCTIONS
Input - .Sample Area (C2Hs) 2^8
Output - Sample Cone, ug/m3 (C2H5) 2NH
Output - Sample Cone, ppm (C2H5)2NH
Input - Standard Cone, pg/ml (C2H5) 3N
Input - Standard Area (C2Hs) 3N
Input - Sample Area (C^j) 3N
Output - Sample Cone. Ug/m3 (C2H5) 3N
Output - Sample Cone, ppm (€2*1$) jN
INPUT
DATA/UNITS
KEYS
i vs~r
1 R/S II 1
1 II
U/s 1!
U/s II
IR/S II
IH/S ||
1 II 1
In. II
1 f II 1
1 REG II 1
1 C.X II 1
1 1
1 II 1
II
1 1
II 1
H
| II
H
II
H
H
II
II 1
1 II !
OUTPUT
DATA/UNITS
-
•
Figure 7 (Cont'd). HP-65 User instructions .continued,
268
-------
SWRI PROJECT NO.
FUEL: CVS NO.
SAMPLE COLLECTION BY:_
GENERAL COMMENTS:
TEST NO.
TUNNEL SIZE:
JTEST DATE:
DRIVER:
_CHEMICAL ANALYSIS BY:
VEHICLE:
MILES:
CALCULATIONS BY:
Test No.
Driving Cycle
Volume Ft3
B.P." Hg
Temp°F
Dilution Factor
Absorb. Rea. Vol. Ml
Stan. Cone. CH-,NH,yg/ml
Stan. Area CH3NH2
Sample Area CH3NH2
Sample Cone. CH^NHo ug/m3
Sample Cone. CH^NH, ppm
Stan. Cone. C 2*178 ug/ml
Stan. Area C2H7*1
Sample Area CjHitt
Sample Cone. C7H7N ug/m3
Sample Cone. C-,H7N ppm
Stan. Cone. (CHi) ^N ug/ml
Stan. Area (CH7) ,N
Sample Area (CH,) ,tj
Sample Cone. (CH-i)-iN Ug/m3
Sample Cone. (CH,) ,N ppm .
Stan. Cone. (C2H^) jNH ug/ml
Stan. Area (C,Hq) ,NH
Sample Area (C,He)-,NH
Sample Cone. (C-jHcKNH U9/m
Sample Cone. (C->Hc) ->NH ppm
Stan, cone. (C^HcKN ug/ml
Stan. Area (C?H<;) -»N
Sample Area (CoHc) ->N
Sample Cone. (C2Hs) 3N ug/m3
Sample Cone. (C2Hs) 3N ppm
1
FTP-1
3.196
29.80
75
1
25.
0.05
1000
640
8.99
0.007
0.10
1500
1000
18.7
0.010
0.50
1500
1020
95.5
0.039
0.90
1560
1760
285
0.094
1.60
3120
2850
410
23456
FTP-2
1.625
30.02
80
5
25
0.10
2000
1000
138
0.107
0.20
3000
1800
332
0.177
1.00
3000
2800
2580
1.050
1.50
3000
4321
5980
1.970
2.00
1260
780
3430
O.J14
FTP-3
2.010
29.02
96
10
50
0.02
3000
1880
119
0.092
0.30
6000
5000
2380
1.270
0.20
4000
2100
1000
0.407
0.10
1870
2000
1020
0.335
0.10
4000
3160
753
0.179
SET- 7
3.730
29.25
85
2
50
1.00
10000
800
79.9
0.062
2.00
15000
8500
1130
0.603
2.00
2000
1600
1600
0.650
1.00
1420
1070
753
0.247
0.50
6210
3100
249
0.059
HFET
8.241
29.95
83
1
75
0.10
4000
2000
16.5
0.013
0.20
2000
1500
49.5
0.026
1.50
1000
875
433
0.176
0.50
1900
1870
162
0.053
0.60
3110
2000
,.127
0.030
NYCC
1.070
29.50
89
1
75
0 . 10'
5000
2700
141
0.109
0.20
3000
2000
348
0.186
1.00
2000
IfiSO
2130
0.865
0.40
3100
2810
946
0.311
0.20
5620
4020
373
0.089 ..
Figure 8. Organic Amine sample collection sheet.
269
-------
Manual calculations for driving cycle FTP-1
Csta(yg/m£) x Asam x D.F. x &bs. Vol. (ml)
Ug/m3 CH3NH2 =
Astd x pexp <"H9>
Texp x 29.92"Hg x 35.31 ft3/m3
x .
COQ°T> X \7 /-FH-"^
528 R x vexp (ft )
(0.05 ug/m&) x 640 x 1 x '25
1000 x 29.80"Hg
(460 + 75) x 29.92"Hg x 35.31 ft3/m3
x
528°R x 3.196 ft3
=8.99 U9/m3
PPM ai'3NH2 = Ug/m3 -r density yg/m£
Mol. Wt. (CH3NH2) x 1000
density yg/m£ =
24.045,
Mol. Wt. CH3NH2 = 31.058 g/mole
31.058 x 1000
density = 2AM = 1292 yg/m£
ppm = 8.99 Ug/m3 4- 1292 yg/mS, = 0.007 mS,/m3 = 0.007 ppm
The calculations for C2H7N, (CH3)3N, and (C2H5)3N are carried out in the
same manner by substituting the appropriate standard concentrations, areas
and molecular weights into the above formulas. These calculations give the
following concentrations: C2H7N, 18.7 ug/m3 and 0.01 ppm; (CH3)3N,
0.05 Ug/m3 and 0.039 ppm; (C2H5) NH, 285 Ug/m and 0.094 ppm; and (C2H5) 3N,
410 ug/nr and 0.098 ppm.
Note: the values used in these calculations are picked from a range of
temperatures, standards, dilution factors, etc. to validate the calculations
and may not be representative of expected raw data. These calculations are
presented to confirm that manual and HP-65 calculations give the same results.
This was confirmed on six sets of calculations .
270
-------
REFERENCES
Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N. J., 1971, pg. 385.
Weast, R. C., (Ed.), Handbook of Chemistry and Physics, 54th Edition, The
Chemical Rubber Co., Cleveland, Ohio, 1973.
Hoshika, Y., Anal. Chem., Vol. 48, pg. 1716, 1976.
Hoshika, Y., J. Chromatog., Vol. 115, pg. 596, 1975.
Sze, Y. L. and Borke, M. L. , Anal. Chem., Vol. 35, pg. 240, 1963.
O'Donnell, J. F. , and Mann, C. K., Anal. Chem., Vol. 36, pg. 2097, 1964.
McCurdy, W. H. Jr., and Reiser, R. W., Anal. Chem., Vol. 38, pg. 795, 1966.
Smith, J. R. L. and Waddington, D. J., Anal. Chem., Vol. 40, pg. 522, 1977.
"Amine Analysis," Bulletin 737A, Supelco Co., Belefonte, Pa., 1973.
"Penwalt 223 Amine Packing," Alltech Assoc., Arlington Heights, Illinois.
Moffat, A. C. and Horning, E. C., Anal. Lett., Vol. 3, pg. 205, 1970.
Clarke, D. D., Wilk, S., and Gitlow, S. E., J. Gas Chromatog., pg. 310,
1966.
Mosier, A. R., Andre, C. E., and Viets, F. G. Jr., Environ. Sci. Tech.,
Vol. 7, pg, 642, 1973.
Burks, R. E. Jr., Baker, E. B., Clark, P., Esslinger, J., Lacey, J. C. Jr.,
J. Agr. Food Chem., Vol. 7, pg. 778, 1959.
Dunn, S. R., simehoff, M. L., Wesson, L. G. Jr., Anal. Chem., Vol. 48,
Pg. 41, 1976.
Gruger, E. H. Jr., J. Agr. Food Chem., Vol. 20, pg. 781, 1972.
Andre, c. E. and Mosier, A. R. Anal. Chem., Vol. 45, pg. 1971, 1973.
Ombreit, G. R., Nygren, R. E., and Testa, A. J., J. Chromatog. Vol. 43,
P9- 25, 1969.
Kfiay, j. N. and Hardy, R. , J. Sci. Food Agr., Vol. 23, pg. 9, 1972.
271
-------
Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.
272
-------
APPENDIX D
SULFUR DIOXIDE PROCEDURE
273
-------
THE MEASUREMENT OF SULFUR DIOXIDE IN EXHAUST
The concentration of SO2 in automotive exhaust can be determined as
sulfate using the ion chromatograph. SO2 exhaust samples are collected in
two glass bubblers, each containing 3 percent hydrogen peroxide. The tem-
perature of the absorbing solution is kept at 0°C by means of an ice water
bath. The bubbled samples are then directly analyzed on the ion chromato-
graph and compared to standards of known sulfate concentrations.
LIST OF EQUIPMENT
The equipment required for the SO^ determination is divided into four
categories: Sampling, Analysis, Water Filteration and Sample preparation.
Manufacturer, stock number and any pertinent descriptive information are
listed.
Sampling
1. Glass impingers, Ace Glass Products, Catalog #7530-11, plain tapered
tip stoppers with 18/7 arm joints and 29/42 bottle joints.
2. Flowmeteis, Brooks Instrument Division, Model 1555, R-2-15-C, sapphire
ball, 0-5 lit/min range, graduated 0-15.
3. Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
capacity.
4. Digital readout xor dry gas meter.
5. Sample pump, Thomas, Model #106 CA18 3, 4 lit/min.
6. Drying tube, Analabs Inc., Catalog #HGC-146, 6" long, 1/4" brass
fittings.
7. Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8" ID
and 5/16" x 3/16" ID.
8. Teflon solenoid valve, The Fluorocarbon Company, Model #DV2-144N Cal.
9. Miscellaneous Teflon nuts, ferrules, unions, tees, connectors and
clamps.
10. Miscellaneous electrical switches, lights, wiring, etc.
11. Regulating valve, Nupro 4M6, stainless steel.
274
-------
12. Six channel digital thermometer, Analog Devices, Model #2036/j/l.
13. 30 m£ polypropylene sample storage bottles, Nalgene Labware,
Catalog #2006-0001.
14. iron/Constantan type J single thermocouple with 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation.
15. Modified 25 mm A-H Microanalysis filter holder, Millipore, Catalog
#XX50 025 00.
16. Fluoropore 25 mm filters, Millipore, Catalog #FHLP 025 00, 0.5
micron pore size.
1. Conductivity cell, modified Swagelok reducing union, Catalog
#SS-200-6-l, approximate volume, 4.5 \il.
2. Conductivity detector, Hall, Tracer 700.
3. Multivoltage recorder, Soltec, Model #B-281 H.
4. Mini-pump, Milton Roy, series 196-0066-033, 46/460 m£/hr capacity.
5. Pulse dampener, Glenco Scientific, Catalog #PD 1000.
6. Polyethylene cubitainers, Cole Farmer Instrument Company, Catalog
#6100-20, 1 gallon.
Water Filtration
1. Filtration apparatus, Millipore, Catalog #XX 15 047 00.
2. Filters, Millipore, Catalog #GSWP 047 00, 0.22 micron pore size.
Sample preparation
1. 3 cc disposable syringes, Becton-Dickson, Catalog #5585.
2. Class A, 1 m& volumetric pipets.
3. Class A, 2 mH volumetric pipets.
4. Class A, 3 mJl volumetric pipets.
5- Class A, 4 m£ volumetric pipets.
5- Class A, 5 m£ volumetric pipets.
7- Class A, 10 mH volumetric pipets.
275
-------
8. Class A, 20 m£ volumetric pipets.
9. Class A, 25 m£ volumetric pipets.
10- Class A, 50 m£ volumetric pipets.
11. Class A, 100 m£ volumetric pipets.
12. Class A, 100 m£ volumetric flasks.
13. Class A, 1000 m£ volumetric flasks.
14. Class A, 2000 m£ volumetric flasks.
15. Mohr pipet, 1 m& graduated 1/10.
LIST OF REAGENTS
A list of reagents used in determination of SC>2 is provided indi-
cating purity, manufacturer and catalog number. The function of each rea-
gent in the procedure is also given.
1. Water - deionized and filtered through 0.22 micron filter.
2. Primary standard - Sulfuric acid, 112804, certified 0.1N, formula weight
= 98.08, ACS reagent grade, Fisher Scientific Company #SO-A-212.
3. Absorbant - Stabilized 30 percent hydrogen peroxide, 11502, formula
weight = 34.01, analytical reagent grade, Mallinckrodt #5239.
4. Eluent - Sodium bicarbonate, NaHC03, formula weight = 84.01, ACS
analytical reagent grade powder, Mallinckrodt #7412.
5. Eluent - Sodium carbonate, Na2CO3, formula weight = 105.99, ACS
analytical reagent grade anhydrous power, Mallinckrodt #7521.
6. Regenerant - Sulfuric acid, H2SO4, formula weight = 98.08, ACS
analytical reagent grade, Mallinckrodt #2876.
PREPARATION OF REAGENTS
Water is prepared by filtering deionized water through a 0.22 micron
Millipore filter and storing in polyethylene bottles. All solutions and
dilutions are made up to volume with water prepared in the above manner.
Primary Standard
The stock solution is prepard by diluting 20 m£ of certified 0.1N
H2S04(4800 yg S°4 ) to 1000 m£ with water. The resulting solution contains
mx,
276
-------
ug S04 2
96.0 £ • More dilute standards are prepared by pipetting 0.5, 1, 1.5,
2, 3, 4,'ST, 10, 20, 30, 40, 50, 60, 70, 80, and 90 mJl of the stock solution
into 100 m& volumetric flasks and making up to volume. These standards re-
main stable for at least fourteen weeks. All glassware used in the prepara-
tion of standards should be washed with 1:1 (v:v) nitric acid and then rinsed
copiously with tap water with a final rinse of filtered deionized water.
Absorbing Solution (3 percent H2O2)
100 mJl of 30 percent H2O2 is diulted to 1 liter with water.
Eluent (0.003 M NaHC03 + 0.0024 M Na2CO3)
Concentrated stock solutions of sodium bicarbonate and sodium carbonate
are made up. For 0.6 M NaHCO3, 50.41g is diluted to 1000 mil with water.
0.48 M NaHC03/is prepared by diluting 50.88g of the solid to 1 liter with
water. The carbonate solution used as the eluent is made up by pipetting
10 m& of each stock solution into a 2 lit volumetric flask and diluting to
the mark with water.
Regenerant (1 N H2SC>4) *
56 m£ of 95 percent H2SO4 is diluted to 2 liters with water.
*A total of 4 liters of each of these solutions are prepared to fill
the 4 liter reservoirs in the chromatograph.
SAMPLING SYSTEM
The dilute automotive .exhaust sample first passes through a fluoropore
filter after which its flow is controlled by an on/off Teflon solenoid
valve. See Figure 1. When in the "on" position the gas continues on
through the impingers where SO2 is oxidized to SO4~2 in a 3 percent hydro-
gen peroxide solution maintained at O°C by an ice water bath. Two sets of
bubblers are required to retain 99 percent of total SO4~2 collected. After
reaction the gas enters a molecular sieve-silica gel dryer where moisture
is- removed. The gas is then drawn by a Thomas sample pump through a Singer
dry gas meter. The gas flow through the meter is controlled by a Nupro
regulating valve and the flowrate is measured by a Brooks flowmeter. The
dry gas meter is equipped with a thermocouple and two digital readouts for
temperature (°F) and volume (ft3) measurement. Several views of the sam-
pling system are illustrated in Figure 2.
PROCEDURE
Dilute exhaust is collected in two bubblers, each containing 25 m£ of
3 percent H202. The temperature is maintained at 0-5°C by an ice water bath
and the flowrate is adjusted to 4 lit/min. After sampling is completed, the
absorbing solution in each bubbler is transferred to a 30 m& polypropylene
277
-------
Gas Temperature
Digital Readout
NJ
~j
oo
Sample
Probe On-Off
' Solenoid
Valve
Flowmeter
Sample
Pump
Dry
Gas
Meter
Regulating
Valve
Dilute
Exhaust
Ice Bath
Temperature
Readout
Gas Volume
Digital Readout
Figure 1. SC>2 sample collection flow schematic.
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. S02 sampling system.
279
-------
Solenoid
Impinger
Ice Bath
Close-up of Impingers (Side View)
Solenoid
Filter
Ice Bath
Drier
Dry Gas Meter
Pump
Rear View
Figure 2 (Cont'd). SC>2 sampling system.
280
-------
bottle and capped. These samples should be analyzed within four or five
weeks after collection to prevent sample deterioration. Approximately 2 ml
of the solution is loaded into the sample loop and injected. This inserts
the sample loop volume (0.5 m£) of sample into the instrument. An analysis
flow schematic and pictures of the ion chromatograph are shown in Figures 3
and 4. The ion chromat9graph utilizes two columns, the separator and the
suppressor. The 3 x 500 mm analytical column and the 3 x 150 mm precolumn
are packed with a patented resin composed of a strong base anion exchanger in
the bicarbonate form. The analytical column separates the anions before
entering the suppressor column. This 6 x 250 mm glass suppressor column
packed with AG 50W-X10, a strong acid cation exchanger neutralized the
ionic effect of the eluent while increasing that of the sample ion. The
column packing is chosen in the hydrogen form so that in the presence of
the eluent (NaHCC>3 and Na2CO3) H2CC>3 is generated and when sulfate is intro-
duced H2S04 is formed.
NaHCO3 + Resin-H ^=^Resin-Na + H2CO3
Na+ Anion" + Resin-H ^==?Resin-Na + H+ Anion"
The acid being more conductive than the hydrogen carbonate produces a signal
on the conductivity meter. This can be interpreted as peak height from a
trace or as peak area measured by the Hewlett-Packard 3354 computer system.
Figures 5 and 6 show two representative chromatograms produced in the anal-
ysis of a standard and a sample. The ion chromatograph operates at room
temperature at a maximum pressure of 500 psi.
CALCULATIONS
This procedure provides a method of determining the concentration of
S02 in automotive exhaust. When the calculations described below are car-
ried out the concentration is expressed in m, ^ and ppm S02- A stepwise
derivation of the equation descirbing the analysis will be provided as well
as a copy of the Hewlett-Packard 65 program which includes all calculations.
See Figure 7. This program was designed to reduce the amount of time re-
quired to do the calculations manually- To illustrate the process two sample
calculations will be performed from the information on the data sheet
(Figure 8) . The first step in the calculations is the assumption that S02
is an ideal gas, i.e., the ideal gas law, PV = nrt, is valid. This equation
gives the volume of sampled gas corrected to specified temperature and
pressure if the experimental conditions are known.
PxVx Tspec
Tx+460 Pspec
where V = volume of gas sampled at specified temperature and
pressure
Px = experimental pressure (in Hg)
Vx = experimental gas volume collected (ft )
281
-------
CVS
glass
impinger
unused
sample
saved as
needed
sample analysis in
ion chromatograph
with conductivity
cell
A/D
converter
recorder
Hewlet-Packard
3354
computer system
Figure 3. S02 analysis flow schematic.
282
-------
Recorder
Conductivity
Detector
Eluent
Pulse
Dampener
Sample
Pump
Analytical
'olumn
Suppressor
Column
Conductivity
Cell
Conductivity
Detector
Figure 4. SO Ion chromatograph.
283
-------
Analytical Column Precolumn 3x150 mm glass,
Analytical column 3x500 mm glass
Packing strong base anion exchanger in the
biocarbonate
Suppressor Column 6x250 mm glass
Packing AG 50W-X16
Eluent 0.003M NaHCOs + 0.0024M
Flowrate 167 ml/nr
Chart speed 12 in/hr
Loop Size 0.5 ml
Sample Set 7 Run 3 Bubbler 1
Date 1/6/78
Collection 12/22/77 Attenuation 30x1
15 10 5
Retention time, minutes
Figure 5. Sample chromatogram.
284
-------
_J_. ' ' *
4-K Analytical Column -Precolumn 3x150 Ttun gi.*^
Analytical column 3x500 mm glass
Packing Strong base anion' exchanger in the
biocarbonate form ~~~ ~~~—
Suppressor Column 6x250 mm glass
Packing AG 50W-X16
h Eluent 0.003M NaHCOs + Q'.0024M Na2CO3
- Flowrate 167 ml/hr
i Chart speed 12 in/hr
1 Loop Size 0.5 ml
TH Sample 7.68>ig S04~2 standard Date 1/6/78
rf —m±—r_.
GollectiOn attSauation 30x1
Ti;
._,..
^f!444V
- -- -i— 4- -i- '-
..,,
t Pi --;-
i ' , i '
r^-i-M-rt-
. -f ..r-. -. -. _,
. . ;_J4._.^_»_,^_
•-41 •::!••••-.. h-Mt
..,.., • ; . ' ;.. !
+-H-
^J4..aii;u4.Tt::z
! 1
i
20 15 10 5
Retention time, minutes
Figure 6. Standard chromatogram.
285
-------
HP-65 User Instructions
Tifls
PITT.FUP r>T^"TE lii ^XI'AUS'T (LINEAR)
HARRY E. DIETZMANN
rpmmer
/
ft ! _!
1 «
i !
Paqe nf
natB 11/11/77
£
/ '
b i i i i >;B
STEP
°1
02
°T
04
1
INSTRUCTIONS
Switch to on; switch to run
Feed card in from right to left
Initialize
Set decimal nlace
Input-sample volume, ft-3
2 i Input-Barometric pressure, "Hq
3
4
5
6
7
Insut-Samcle Teirirarature , °F
Incut-Dilution factor
Input-Absorbing reagent volume r ml
Ug S04~2
Input-Standard concentration. — Bd
Input-Standaz 5 -rea
8 Inp"t-Sair,riie Aiea
ug so2
Output-Saispi- ccncS'itration, 3 —
9- '|
Output-Sample corctntration, ppm
INPUT
DATA/UNITS
KEYS
1 II
II I
f HREG
IDSP \\2
A II
R/S ||
IB/S II
R/s II
R/S ]f
Ip/fi II
IR/S II
|R/s ||
II
„« II
1
11
PTN '
If II
|REG ||
|CLX ||
1 II
1 II
1 II
1 II
II
r ii
II
OUTPUT
DAT A/ UNITS
- — ^- ..
•<
Figure 7. HP-65 program form.
286
-------
HP-65 Program Form
]f SULFUR DIOXIDE IN EXHAUST (LINEAR)
SWITCH TO W/PRGM. PRESS [7] | PROM j TO CLEAR MEMORY.
Page.
.of.
KEY
ENTRY
LBL
A
3
5
•
3
1
i
5
B 2
ft
x
2
9
•
9
2
•f
R/S
20 X
STO 1
R/S
4
6
0
+
SCL 1
2
Ra
R4
Rs
Re
R7
Ra
Ra
LABELS
A
B
c
D
F
0
1
2
3
4
5
6
7
a
«
FLAGS
1
2
Figure 7 (Cont'd.) HP-65 program form.
287
-------
SWRI PROJECT NO,
FUEL!£/i|-M7-oo' CVS NO. 3
TEST NO. OP/ TEST DATE! 12.-12- 17 VEHICLE ! /W?C7/C£
TUNNEL SIZE a /K, DRIVER i fiill f MILES: /OOP
SAMPLE COLLECTION BY ! naryflnn P. CHEMICAL ANALYSIS BY :^/&a£_CALCULATIONS
GENERAL COMMENTS:
00
03
NO.
1.
2
3
4
5
6
7
8
X
DRIVING-
CYCLE
FTP-1
FTP- 2
FTP- 3
SET- 7
HFET
NYCC
BG
X
SAMPLING CONDITIONS
VOLUME
FT3
1.139
2.040
I.2SO
3.1HO
/.t90
1.410
2.000
©
B.P,
"HG
29.97
29. 4 a
29.^9
Z9.&0
2.9. SO
Z9.5/
29- VS
©
TEMP ,
op
7^.7
7S.O
7^.S
7S.O
7V. V
7^.9
7^.9
©
DILU-
TION
FACTOR
1
/
/
/
)
1
1
®
ABSORB
JEAGENT
VOLUME
ML
25
Z5
2.S
2.S
•us
2_5
2S
©
STANDARD
jg NH4+
i'l
/ o
3O
10
to
20
/S
/-O
©
AREA
31OO
ysoo
32.00
42.00
57OO
^900
3OOO
©
SAMPLE
AREA
^iOOO
3SOO
^300
V5OO
ssoo
500O
2700
®
oiaJiH3_
foSfcO
69 3O
(bSOO
aooo
6(b70
££30
273
PPM
23M
2-SV
2,37
2,9¥
2.ys
2.^3
O./O
-
Figure 8. SO2 data sheet.
-------
Tx = experimental temperature ( °F)
Tspec = specified temperature = 68 °F = 528 °R
Pspec = specified pressure = 29.92 in Hg
Next, the quantity of SO2 in this volume of gas is calculated. The ion
_^
chromatograph measures the amount of sulfate (from SO9) in ^g SO4 . Assum-
^ mZ
ing linearity between concentration of sulfate and peak area, a standard of
known sulfate concentration is compared to a bubbled exhaust sample.
Cst _ _ Csamp
AREAS t AREAsamp
Csamp _ Cst x AREAsamp
AREAst
-2 yg S°~2
where Csamp = concentration of 864 in sample
in Xt
Cst = concentration of SO* in standard
mx/
AREAsamp = area of sample (relative units)
AREAst = area of standard (relative units)
Converting to yg SO2/m£:
64.06 g/mole SO9
Csamp x _ =—j - 0.667 Csamp
96.06 g/mole 804"
If the sample has been diluted the dilution factor, DF, needs to be
included
0.667 Csamp x DF
This represent the amount of S02 in one m& of -absorbing solution. The
volume of absorbing solution is multiplied by this last quantity to give
the amount of SO2 collected.
0.667 Csamp x DF x absorb, vol. (m£) = yg SO2
yg SOo • j i,
The concentration of SO2 in - 3-* in the exhaust sample is obtained by
dividing yg S02 by gas volume (corrected from ft3 to m3) .
289
-------
Concentration of S02 ,
m°
ua SO Cst' ^ S°4 X AREAsamp
0.667 MSJ ? ;; rJl x DF x absorb, vol., m£
Ug SO.~2 AREAst
4
Px, in Hg x Vx, ft3 x 528°R
Tx, °F + 460 29.92 in Hg
—
0.667 ^g S°2 x Cst, ^g S°4 x AREAsamp x DF x absorb, vol., m£
pg 804 2 m£
I mj-ui--:- :i|- --- - . " --- — . . _ - - - -. --- .-mi, - _. ____ 1 -- - .T - j ----- --- -
Px, in Hg x vx, ft3
(Tx, °F + 460) x 29.92 in Hg x 35.31
X
528°R x AREAst
To find the SO2 concentration in ppm the density of the gas at the specified
conditions is needed. The density of S02 at 0°C and 29.92 in Hg is 2.927
g/lit. This can be corrected to 68 "F (20°C = 293*K) and 29.92 in Hg using
Charles' version or. the ideal gas law:
V V
T
where V± is the volume at 0°C (lit) =1.0 lit
T^ is 0°C = 273°K
V is the volume at 20°C (lit)
T is 20°C = 293°K
v - ' , m,
V - S - = 1.073 lit
Density at 68°F (20°C) and 29.92" Hg = 2'927g -
1.073 lit
= 2.728 g/lit = 2728
rax/
= 0.000367 —
ug
0. 000367 ==
290
-------
Calculation
Example 1
Assume 1.189 ft of exhaust was collected in 25 m£ of 3 percent H O
in the FTP-1 driving cycle at a barometric pressure of 29.47 in Hg (cor-2
rected) and temperature of 74.7°F. The sampling and analysis were both
performed according to the procedure outlined in a previous section. An
excess of sample injected into the 0.5 m£ sample loop produced a peak area
ir\ yg so.. 2
01 HUUU uvJUHwo am-t L»JJLC= v*i/j. A. ^o^wjutaj.1
3200 counts.
Onnmm 2.™, x n f,^i
yg SOo
ppm SO, = ±
U 2 Px,
iy s L-CHIU.O..LU ,
Ug S02
v (^*i-i ^
yg so4-2 " CoL'
in Hg x vx, ft3
n -, an area or
yg S04~2
j x AREAsdltlp
DF x absorb, vol., m£ x (TX, °F + 460) x 29.92 in Hg x 35.31
m3
ft-
528°R x AREAst
_ 0.000367 x Q.66"7 x 10 x 4QQQ x l x 25 x (74.7 + 460) x 29.92 x 35.31
29.47 x 1.189 x 528 x 3200
Sample ppm SO2 = 2.34 ppm
Example 2
Assume that in the SET-7 driving cycle dilute automotive exhaust was
collected in 25 m£ of 3 percent H2O2 according to the procedure described
previously. The sampling conditions under which the 3.240 ft3 of exhaust
was collected were 75.0°F and 29.50 in Hg. 4200 counts was the area pro-
— *?
duced by the 40 ^g S04 standard and the exhaust sample yielded an area
of 4500 counts. Inserting these values into the same equation used in
Example 1 gives a concentration of 2.94 ppm S02- Note that the information
from the data sheet used in these calculations does not necessarily repre-
sent actual values obtained from exhaust sampling.
291
-------
REFERENCES
Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N. J., 1971, pg. 513.
West, P. W. and Gaeke, G. C., Anal, Chem., Vol.. 28, pg. 1816, 1956.
Adams, D. P., Corn, M., Harding, C. I., Pate, J. P., Plumley, A. L.,
Scaringelli, F. P., and Urone, P., "Methods of Air Sampling and Analysis,"
[Tentative Method of Analysis for Sulfur Dioxide Content of the Atmosphere
(colormetric), No. 42401-01-69T], pg. 447, Amer. Pub. Health Ass., Washington,
D. C., 1972.
"ASTM Standards - Water, Atmospheric Analysis," ASTM Designation D291470T,
Part 23, Am. Soc. Test. Mat., Philadelphia, Pa., 1972.
US EPA, Fed. Register, Vol. 36, No. 158, 15492, 1971.
Robinson, E. and Robbins, R. C., "Sources, Abundance, and Fate of Gaseous
Atmospheric Pollutants," Final Report, SRI Project 6755 Supplement, Stanford
Research Inst., Palo Alto, Calif., 1968.
Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc., N. Y.,
1976, Vol. 3, pg. 214.
Zurlo, N. and Griffini, A. M., Med. Lav., Vol. 53, pg. 330, 1962.
Scaringelli, F. P., Elfers, L., Norris, D., and Hochheiser, S. Anal. Chem.,
Vol. 42, pg. 1818, 1970.
Pate, J. B., Ammons, B. E., Swanson, G. A., Lodge, Jr., J. P.,
Anal. Chem., Vol. 37, pg. 942, 1965.
Pate, J. B., Lodge, Jr., J. P., and Wartburg, A. F., Anal. Chem., Vol. 34,
pg. 1660, 1962.
Scaringelli, F. P., Saltzman, B. E., and Frey, S. A., Anal. Chem., Vol. 39,
pg. 1709, 1967.
Hartman-Leddon, Phila., Pa.
Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc., N. Y.,
1976, Vol. 3, pg. 214.
Urone, P., Evans, J. B., and Noyes, C. H., Anal. Chem., Vol. 37, pg. 1104,
1965.
292
-------
Bostrom, C. E. Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.
Attari, A-, Igielski, T. P., and Jaselskis, B., Ar.al. Cl^r,., Voi 42
pg. 1282, 1970.
Kawai, T., Netsu Kanri, Vol. 22, pg. 20, 1970.
Thomas, M. D,, Ivie, J. O., and Fitt, T. C., Ind. Eng. Chem., Anal. Ed.,
Vol. 18, pg. 383, 1946.
Yocum, J. E., Richardson, R. L., Saslaw, I. M., and Chapman, S., Proc, 49th
Ann. Meet. Air Poll. Contr. Ass., Pittsburgh, Pa., 1956.
Kuczynski, E. R., Environ. Sci. Tech., Vol. 1, pg. 68, 1967.
Shikiya, J. M. and McPhee, R. D., 61st Annual Meeting, Paper No. 68-72, Air
Poll. Contr. Ass., Pittsburgh, Pa., 1968.
Jacobs, M. B., The Chemical Analysis of Air Pollutants, Chapter 8, Wiley
(Interscience), New York, N. Y., 1960.
Jacobs, M. B., The Analytical Chemistry of Industrial Poisons, Hazards, and
Solvents, Chapter 9, Wiley (Interscience), New York, ':>;. ¥., 1949.
Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.
Stratman, H., Mikrochint. Acta, Vol. 6, pg. 688, 1954.
Pate, J. B., Lodge, Jr., J. P., and Neary, M. P., Anal. Chim. Acta., Vol. 28,
pg. 341, 1963.
Lodge, Jr., J. P., Pate, J. B., and Huitt, H. A., AtT.er. Ind. Hyg. Ass., J.,
Vol. 24, pg. 380, 1963.
Forrest, J. and Newman, L., Atmos. Environ., Vol. 7, pg. 561, 1973.
Huygen, C., Anal. Chim. Acta, Vol. 28, pg. 349, 1963.
Harding, J., and Schlein, B., "Nuclear Techniques in Environmental Pollution,"
SM-142a/9, IAEA, Vienna, Austria, 1971.
Wilsdon, B. H. and McConnell, F. J., J. Soc. Chem. Ind, Vol. 53, pg. 385, 1934.
Huey, N. A., J. Air. Poll. Cont. Ass., Vol. 18, pg. 610, 1968.
Thomas, F. W. and Davidson, C. M., J. Air Poll. Cont. Ass., Vol. 11, pg. 1,
1961.
Stalker, W. W., Dickerson, R. C., and Kramer, G. D., Amer. Ind. Hyg. Ass.,
J-, Vol. 24, pg. 68, 1963.
293
-------
Hickey, H. R. and Hendrickson, E. R., J. Air Poll. Cont. Ass., Vol. 15,
pg. 409, 1965.
Harding, C, I. anrt K^lley, T. R., J. Air Poll. Cont. Ass., Vol. 17, pg. 545,
1967.
Ikeda, T., J. Hyg. Chem., Vol. 13, pg. 41, 1967.
Stevens, R. K., Mulik, J. D., O'Keeffe, A. E., and Krost, K. J., Anal. Chem.,
Vol. 43, pg. 827, 1971.
Stevens, R. K., O'Keefe, A. E., and Ortman, G. C., Environ. Sci. Tech.,
Vol. 3, pg. 652, 1969.
Ronkainen, P., Denslow, J., Leppanen, O., J. Chroma tog. Sci., Vol. 11,
pg. 384, 1973.
Pescar, R. E. and Hartmann, C. H., J. Chroma tog. Sci., Vol. 11, pg. 492, 1973.
Stevens, R. K. and O'Keefe, A. E., Anal. Chem., Vol. 42, pg. 143a, 1970.
Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I., Anal. Chem.,
Vol. 44, pg. 2070, 1972.
Bruner, F., Ciccioli, P., and DiNardo, F., Anal. Chem., Vol. 47, pg. 141, 1975.
Bruner, F., Ciccioli, P., and DiNardo, F. , J. Chromatog., Vol. 99, pg. 661,
1974.
Bremner, J. M. and Banwart, W. L., Sulfur Inst. Journal, Vol. 10, pg. 6,
1974.
Mulik, J. D., Todd, G. , Estes, E., and Sawicki, E., "Ion Chromatography
Determination of Atmospheric Sulfur Dioxide," Symposium on Ion Chromato-
graphic Analysis of Environmental Pollutants, EPA, Research Triangle Park,
N. C., April, 1977.
Dietzmann, H. E., "Protocol to Characterize Gaseous Emissions as A Function
of Fuel and Additive Composition," Environmental Protection Technology Series
under Contract 68-02-1275, September, 1975.
DeSouza, T. L. C., Lane, D. C., and Bhatia, S. P., Anal. Chem., Vol. 47,
pg. 543, 1975.
294
-------
APPENDIX E
NITROUS OXIDE PROCEDURE
295
-------
THE MEASUREMENT OF NITROUS OXIDE IN EXHAUST
This procedure was developed to measure nitrous oxide (N-O) in dilute
gasoline and diesel exhaust. Standard CVS bag samples are analyzed for NO
using calibration blends to quantify the results. Gas chromatograph peak
areas are obtained using a Hewlett-Packard 3354 computer system. This tech-
nique has a minimum detection limit of less than 0.01 ppm. The total system
schematic for the analysis of N2O in exhaust is shown in Figure 1.
ANALYTICAL SYSTEM
The analysis for N2O in exhaust is conducted with a gas chromatograph
system using a Perkin-Elmer Model 3920B electron capture detector. The sys-
tem employs two pheumatically operated electrically controlled Seiscor valves,
an analytical column, and a stripper column. The gas chromatograph separa-
tion is obtained at room temperature. A special control console was fabri-
cated to house the entire system except for the electron capture detector.
A stripper column is included as a precautionary measure to prevent
unwanted heavier molecular weight exhaust species from the analytical system
from entering. Figure 2 (Step 1) illustrates the gas chromatograph flow
schematic with the gas sampling valve in the purge position and the backflush
valve is foreflushing to the analytical column. Figure 3 (Step 2) illu-
strates the flow schematic when the gas sampling valve is actuated and the
backflush valve still in the foreflush position to the analytical column.
Once the N»O peak foreflushing has eluted, the backflush valve is activated
and the heavier molecular weight species retained on the stripper column are
backflushed to vent, as shown in Figure 4. A summary of the individual steps
is presented below:
Gas Sampling Valve Backflush Valve
Step Position Function Position Function
1 off purge GSV w/sample off foreflush to
analytical column
2 on sample injected off foreflush to
analytical column
3 on sample injected on backflush to vent
Under normal conditions, it is not necessary to backflush the cali-
bration standards since they are free of contaminants that would interfere
296
-------
Vehicle
CVS
Tedlar bags
Electrometer
Gas
Chromatograph
Recorder
A/D
Converter
HP 3354 GC
Computer Systen;
Teletype
Printout
Figure 1. Total system flow schematic for the
analysis of nitrous oxide in exhaust.
297
-------
Stripper Column
(2' x 1/8" SS, 10% OV-17
on 80/100 Gas Chrom Q)
Seiscor Valve
(normal configuration)
Perkin-Elmer
3920 B
Electron Capture
Detector
Analytical Column
(61 x 1/8" SS, 120/150 Porapak Q)
Auxiliary
Carrier Gas
Capillary
Restrictor
10 ml
sample
loop
••i»•**«•*••••»•••»»•
0—-—o
Carrier
Gas
n
-X.
Seiscor Valve
(Gas Sampling Configuration)
Vent -^-
Regulating
Valve
Pump
Si—I
Flowmeter
Figure 2.
Female
Quick-Connect
Flow schematic of nitrous oxide analytical system
(Step 1 - Purge of sample loop of CSV).
Sample 01
Calibration
gas in
298
-------
Stripper Column
(21 x 1/8" SS, 10% OV-17
on 80/100 Gas Chrom Q)
Seiscor Valve
(normal configuration)
Perkin-Elmer
3920 B
Electron Capture
Detector
Analytical Column
(61 x 1/8" SS, 120/15OPorapak Q)
Capillary
Restrictor
10 ml
sample
loop
o
Auxiliary
Carrier Gas
Carrier
Gas
Seiscor Valve
(Sample Inject Configuration)
Vent
Flowmeter
Female
Quick-Connect
Sample or
Calibration
gas in
Figure 3. Flow schematic of nitrous oxide analytical system
(Step 2 - Inject sample or calibration gas into system).
299
-------
Stripper Column
(2' x 1/8" SS, 10% OV-17
on 80/100 Gas Chrom Q)
Seiscor Valve
(backflush configuration)
Perkin-Elmer
3920 B
Electron Capture
Detector
Analytical Column
(61 x 1/8" SS, 120/150 Porapak Q)
Auxiliary
Carrier Gas
Capillary
Restrictor
10 ml
sample
loop
Carrier
Gas
mmmmmmm*
••^•
Seiscor Valve
(Purging Configuration)
Vent
Regulating
Valve
Female
Quick-Connect
ample or
Calibration
gas in
Flowmeter
Figure 4. Flow schematic of nitrous oxide analysis system
(Step 3 - Backflush OV-17 stripper column).
300
-------
with tha analytical column. A typical gas chromatograph trace for a cali-
bration blend is shown if Figure 5. A baseline separation is obtained and
the $2° Peak area is obtained using a Hewlett-Packard 3354 GC computer system,
On gasoline and diesel samples it is necessary that the backflush is
included in the analysis to prevent contamination of the analytical column.
Using the system described, 60 seconds are allocated to the foreflush posi-
tion and 360.seconds are allowed for backflushing the stripper column. A
typical gasoline-CVS sample GC trace is presented in Figure 6, while a die-
sel CVS-sample gas chromatograph trace is illustrated in Figure 7. The gas
chromatograph operating conditions are listed on.Figures 6 and 7.
CONTROL SYSTEM
The control of the two Seiscor valves is accomplished by ATC timers and
ASCO electric solenoid valves. The electrical schematic for the control of
the Seiscor valves using these timers and electric solenoid valves is shown
in Figure 8. The flow schematic for vacuum and pressure lines to the Seiscor
valve are presented in Figures 9-11.
EQUIPMENT
This analysis is performed using a gas chromatograph equipped with an
electron capture detector. The detector, detector heater controls, electro-
meter, recorder and GC integrator are major electronic components in the
detection system. A control console was fabricated to incuse the mechanical
hardware items that are necessary for the proper operation of the ^0 analy-
sis system. Figure 12 illustrates the complete analytical system for mea-
suring ^O in exhaust. The major items that are included in each of these
systems is listed below: '_.
Gas Chromatograph
1. Perkin-Elmer Model 3920B gas chromatograph
2. Linearized electron capture detector (BCD)
3. Leeds and Northrup Model W 1 mv recorder
4. Hewlett-Packard Model 3354 GC computer system
5. Hewlett-Packard Model 1865A A/D converter
Control Console System
1. Seiscor valve - gas sampling configuration
2. Seiscor valve - backflush configuration
3. ATC timers, Model 325A346A10PX (2 ea)
4. Analytical column, 6' x 1/8" SS, 120/150 Poropak Q
301
-------
ft. PANICS.
Instrument Pg SM32..O p Operator
Column £ ft.
Packed with
Liq. Phase
Support
Carrier
on O/IO&mesb
ISO
Run ISO® gQ
@ 44 psig. AM Rotameter Rdg.
O« A *• • ^* ^B .K. ^^
Hyd. MA psig
Air N A psig
) A/4 psig
Rotameter Rdg.
Rotameter Rdg
Rotameter Rdg
RecorderjO.i IN/min speed
Injection IO (Wl indicated
Sampling Device &AS SAMPLIM&
mV.F.S. _
IQ «l net
0
Figure 5.
1 2 3 4 5 6
Retention time, minutes
Typical N?0 calibration blend gas chromatograph trace,
302
-------
1
1
1-
•
I-
I
•
I
•
•:.
• —
I
I
1
l~. -~
• -:• -
-
-
~
••
30-4— J
-- - t- -
. . t..
— r* — ~
— r -
. .. .._...
-70-q— |
- _-p'
j
_.
-
_ .
r-
—
L~_-_I1-.
60— ' — P~
?
i
"it-
1 A I
4U 1
t "I
.. . _£...
i —
F
L_~
. . —
aT J
J _C
J
H)
^ ' "
r 4,
4- -+
fO-j--
(1 ---r —
10--*—
#•
f
- - —4-
r
8
-
; ...:
|
' ~
;
i
t
1 J
...
--.;.f -"-•
. — . _.
1
^9
Sample GAfcOLlME EYH AU VT.CVOate II- 15-77
Instrument PE 3940 & Operator p SAUNDFftS
Column
6 ft. «/fl.V O.D.O.OA« I.D.
-------
Sample DIESEL g*tiAUST-CVS Date
Instrument PC 333.Q R Operator
Column
P.P.0.043 I.D.
A/A
Liq. Phase
Support
Packed with —*
on SO I/OO mesh
. «.w>">
cc/min. GHj /fffCarrier
Run
- 44 psig.
Inlet ROOM.
Detector: 32iS
Rotameter Rdg.
Hyd. ft/A psig
Air //4 psig
( > klA psig
Recorder AS" ff»/min speed / mV.F.S/Jc.* Wj:
Rotameter Rdg.
cc/min
cc/min
Type
Rotameter Rdg.
Rotameter Rdg.
Injection 10 Ml indicated lO Ml net_
Sampling Device GAB SAIHP2.IM6 VALVE.
IO Ml Actual
34567
Retention time, minutes
Figure 7. Typical diesel-CVS exhaust sample.
10
304
-------
AC (-)
ATC
Timer
backflush
solenoid
Figure 8. Electrical schematic for nitrous oxide analysis system.
-------
Vacuum
Air Pressure (30 psi)
Vacuum
Air Pressure (30 psi) f
Backflush off
Figure 9. Flow schematic in electric solenoid valves
(Both valves de-energized).
306
-------
Vacuum
LJ
Air Pressure (30 psi) |
CSV on
Vacuum
U
Air Pressure (30 psi) [
cap
P V
Backflush
Seiscor
Backflush off
Figure 10. Plow schematic in electric solenoid valves
(CSV energized, backflush de-energized).
307
-------
Vacuum
n
cap
Air Pressure (30 psi)
u
cap
p v
CSV
Seiscor
CSV on
Vacuum
cap
Air Pressure (30 psi)
Cap
P V
Backflush
Backflush on
Figure 11. Flow schematic in electric solenoid valves
(Both valves energized).
308
-------
A/D Convnarter
'
.'
Figure 12. Nitrous oxide analytical system.
-------
5. Stripper column, 2' x 1/8" SS, 10% OV-17 on 80/100 Gas Chrom Q
6. ASCO solenoid valve, Model 834501 (2 ea)
7. Brook flowmeter, R-2-15-A w/SS float, 0-150 scale
8. Metal Bellows MB-155 pump
9. Female quick-connect, stainless steel
10. Nupro Model 2M stainless steel regulating valve
11. Stainless steel tubing (0.01"ID) for capillary restrictor.
12. Miscellaneous stainless steel, copper and Teflon tubing
(1/8" and 1/16")
13. Miscellaneous stainless steel and brass unions, tees, etc.
14. Bud Classic II control console cabinet
15. Miscellaneous electrical on-off switches.
SAMPLE CALCULATIONS
The quantification of ^O in exhaust is based on a direct comparison
of the N2O in exhaust with a calibration blend of a know ^O concentration.
Two basic assumptions are made in these calculations that should be con-
sidered with other systems. The first assumption is that the electron cap-
ture detector has a linearized output and that measurements are made within
the working range of the system. These two parameters were verified for this
procedure using instruments previously described. Working within the linear
range of a given gas chromatograph equipped with an electron capture detector,
the following relationship is true.
Let Csam = ppm concentration of N2O in sample
Cstd = ppm concentration of ^O in standard
Asam = area of NoO peak in sample
Astd = area of ^O peak in standard
Astd _ Asam
Cstd Csam
Solving for Csam
Csam = Asam x Cstd
Astd
310
-------
Example 1:
A 4.95 ppm N20 (in nitrogen) calibration blend was found to give 5291
area counts for the N20 peak. An exhaust sample was round to give 2674
area counts for the N2O peak. Calculate the N20 in the exhaust sample.
Csam = Asam x Cstd
Astd
Csam = 2674 x 4.95
5291
Csam =2.50 ppm N2O
Example 2:
A 1.13 ppm N20 (in nitrogen) calibration blend was found to give 1208
area counts for the N20 peak. An exhaust sample was found to give 534
area counts for the N20 peak. Calculate the concentration of N2O in the
exhaust sample.
Csam = Asam x Cstd
Astd
Csam = 534 x 1.13
1208
Csam = 0.50 ppm N2O
311
-------
REFERENCES
Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N. J., 1971, pg. 431.
Jay, B. and Wilson, R., J. Appl. Phys., Vol. 15, No. 2, pg. 298, 1960.
DeGrazio, R., J. Gas Chromatog., Vol. 3, pg. 204, 1965.
F & M Applications Chromatogram, 1942, F & M Scientific Division of
Hewlett-Packard Corp-., Avondale, Pa.
Porapak Brochure, Waters Associates, Framingham, Mass.
Bethea, R., and Adams, F., J. Chromatog., Vol. 10, pg. 1, 1963.
Leithe, W., The Analysis of Air Pollutants, Humphrey Sci. Pub., Ann Arbor,
Michigan, 1970, pg. 176.
Private communication between Dr. R. B. Zweidinger and Frank Black.
LaHue, M. D., Axelrod, H. D., and Lodge Jr., J. P., Anal. Chem, Vol. 43,
pg. 1113, 1974.
312
-------
APPENDIX F
INDIVIDUAL HYDROCARBON PROCEDURE
313
-------
THE MEASUREMENT OF INDIVIDUAL HYDROCARBONS IN EXHAUST
This procedure was developed to measure individual hydrocarbons in di-
lute automotive exhaust. The term, individual hydrocarbons (IHC) is used to
define the collection of compounds: methane (CH4) , ethane (€2^) , ethylene
(C2H4), acetylene (C2H2), propane (C3H8), propylene (C3H6), benzene (C6H6),
and toluene (CJHQ). Dilute exhaust is collected in Tedlar bags during a test
cycle and analyzed with a gas chromatographic system containing four separate
columns and a flame ionization detector (FID). The peak areas are compared
to an external calibration blend and individual hydrocarbon concentration are
analyzed with a Hewlett-Packard 3354 computer system. The analysis flow
schematic is shown in Figure 1.
ANALYTICAL SYSTEM
The analysis for individual hydrocarbons is conducted with a Varian
Aerograph Series 1400 gas chromatograph using a flame ionization detector
(FID). Four separate packed columns are used to resolve these individual
compounds. An elaborate system of timers, solenoid valves, and gas sampling
valves are used to direct the flow of the sample through the system. The
actual analytical system is shown in Figure 2.
The first two columns are used to resolve air, methane, ethylene, ethane,
acetylene, propane and propylene, respectively; while columns III and IV re-
solve benzene and toluene. Column I consists of an 8' x 1/8" stainless steel
tube packed with Porapak Q 80/100 mesh. This column is primarily used to re-
solve methane from air. It undergoes temperature programming from 25°C to
100°C at 12°/min. Column II consists of a 4' x 1/8" Teflon column packed
with 35/60 mesh type 58 Silica gel. €2 and C3 hydrocarbons are resolved with
this column. It is held isothermal at room temperature (20°C). The third
column is used to resolve benzene from the other aromatics, paraffins, ole-
fins, and acetylenes. It consists of a 15' x 1/8" stainless steel column
packed with 15 percent 1, 2, 3-tris (2-cyanoethoxy) propane on 60/80 mesh
Chromosorb PAW. This column is held isothermal at 100°C at the end of the
temperature program sequence. Column IV is a 2' x 1/8" stainless steel tube
packed with 40 percent mercury sulfate (HgSO4) and 20 percent sulfuric acid
(H2SO4) on Chromosorb W. This column resolves benzene and toluene from the
oxygenated hydrocarbons such as aldehydes and ketones. It is also held iso-
thermal at room temperature for the entire analysis sequence. All samples
pass through a 6' x 0.01" capillary restrictor before entering the detector.
Helium is the carrier gas with a column flow of 52 m£/min.
The temperature program sequence is accomplished with the oven of the
gas chromatograph. Columns I and III are in this oven although Column I is
the only one used during the temperature program. The gas sampling valves
314
-------
Vehicle
CVS
Tedlar bags
Gas
Chromatograph
A/D
Converter
Recorder
HP 3354
Computer System
Teletype
Printout
Figure 1. The analysis flow schematic for individual hydrocarbons.
315
-------
'
Figure 2. Analytical system for individual hydrocarbons.
-------
are contained in a Bendix Valve Oven. The temperature is maintained at 100°C
Columns II and IV are external to this oven for isothermal room temperature
operation.
Samples are collected in Tedlar bags during the driving cycle along with
a background bag. The sample is purged through two 10 m£ sample loops for
four (4) minutes. At this time, the initial configuration of the analytical
system is shown in Figure 3. Upon injection, gas sampling valve A is acti-
vated by solenoid valve G, the temperature program sequence is started, and
the first timer begins to count down 680 seconds. The temperature program
sequence for Columns I and II starts at 25°C and increases at 12°/min to a
final temperature of 100°C. Columns I and III are held isothermal at this
temperature for the remainder of the analysis. The configuration of the ana-
lytical system is shown in Figure 4. The sample in the first loop passes
through Columns I and II and into the detector. The peaks (in the order of
elution) are air, methane, ethylene, ethane, acetylene, propane and propylene.
After 680 seconds, the second step begins with solenoid H activating gas sam-
pling valve B. The second timer begins to count down 120 seconds. At this
time, the sample trapped in the second 10 m£ sample loop is channeled.through
Column III. The analytical system configuration is shown in Figure 5. After
120 seconds, step 3 begins. The third timer starts counting down 480 seconds
and gas sampling valves C and J are activated by solenoid valve E. Columns
I and II are backflushed through a capillary restrictor to the vent and Col-
umns III and IV are directed to the detector. The configuration is shown in
Figure 6. After 480 seconds, solenoid valve F activates gas sampling valve
D. Column III is backflushed through a capillary restrictor to the vent.
The final configuration is shown in Figure 7. The last two peaks in the order
of elution are benzene and toluene. Upon elution of the last peak, the system
is reset to the initial position.
A time/temperature system operation sequence is presented in Figure 8.
The solid line on this graph represents the gas chromatograph oven tempera-
ture during the temperature program sequence. The time at which each step
begins is also represented on the graph.
Figures 9 through 13 illustrate a simplified version of the flow of the
carrier gas and sample through the gas sampling valves. Figure 9 shows the
configuration of the gas sampling valves in the sample purge position. The
sample is pumped out of the sample bag and through the sample loops. Column
III is flushing and Column IV is backflushing to the vent and Columns I and
II are directed to the detector. At the start of an injection, the position
of gas sampling valve A changes and the trapped sample is directed to Columns
I and II. Columns III and IV remain in the flushing mode. Figure 10 illus-
trates the analytical configuration upon injection. Step 2 begins with the
trapped sample in sample loop B routed to Column III. Figure 11 demonstrates
Step 2. Gas sampling valves C and J are simultaneously switched to backflush
Columns I and II and direct the sample in sample loop B through Column III
and Column IV and then to the detector in Step 3. Finally, Step 4 results in
the backflushing of column III to the vent. The system remains in this con-
figuration until it is reset to the initial purge position. A sample chrom-
atogram for the calibration blend of individual hydrocarbons is shown in
Figure 14.
317
-------
OJ
M
CD
PRESSURE
SOLENOID VALVE
jjSAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
COLUMN IV
Figure 3. Initial analytical system configuration.
-------
PRESSURE
SAMPLE IN
SAMPLE LOOP A
SOLENOID VALVE
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
COLUMN IV
Figure 4. Step 1 solenoid G activation of gas sampling valve A.
-------
PRESSURE
w
[>G
o
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
•HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
COLUMN IV
Figure 5. Step 2 solenoid H activation of gas sampling valve B.
-------
PRESSURE
NJ
SOLENOID VALVE
{SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
!MMJ COLUMN IV
Figure 6. Step 3 solenoid E activation of gas sampling valves C and J.
-------
PRESSURE
w
to
JSAMPLE IN
« SAMPLE LOOP A
SOLENOID VALVE
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
COLUMN IV
Figure 7. Final analytical system configuration.
-------
100
80
CO
O
o
3
4J
i1
40
ISOTHERMAL OPERATION
STEP 2 STEP 3
STEP 4
SAMPLE PURGE
UJ
STEP 1
J_
JL
Figure 8.
12 15 18
Time, Minutos
System operation sequence.
21
24
27
30
-------
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
CAPILLARY
VENT
FID DETECTOR
CAPILLARY
O
WKP
COLUMN IV /
RESTRICTOR
Figure 9. Sample purge position prior to sample injection.
-------
FID DETECTOR
U)
tSJ
tn
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 10. Step 1 sample loop A injected.
-------
FID DETECTOR
to
CT>
SAMPLE IN
HELIUM
SAMPLE OUT
RESTR.CTOR
HELIUM
Q SAMPLE I I LOOP BO
\^]i\i]L^7
\ ^—A_A_^^ /
—AD d-—
Figure 11. Step 2 sample loop B injected.
-------
FID DETECTOR
CO
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 12. Step 3 simultaneous solenoid C and J activation.
-------
FID DETECTOR
U)
NJ
00
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 13. Step 4 backflushing of column III.
-------
U)
to
22 21 20 19 18 17 16 15 14 13 12 11 10 9 B 7 f> 5 4 3 2 10
Retention time, mm.
Figure 14. Calibration blend for specific hydrocarbons.
-------
CONTROL SYSTEM
The control of the five Seiscor gas sampling valves is accomplished by
ATC timers arid ASCO ^lectr.ic solenoid valves. With the solenoid valve in the
de-energized configuration, the gas sampling valve is in one position. An
electrical impulse from the timer energizes the solenoid valve. The pressure
difference created in the gas sampling valve changes the position. Figure 15
illustrates the de-energized and energized configuration for solenoid valve G
and gas sampling valve A. This accomplishes the complicated column flow se-
quence required for this analysis. The system analytical repeatability is
+_ 1.8 percent.
EQUIPMENT
This analysis Is performed using a gas chromatograph equipped with a
flame ionization detector (FID). The equipment required is divided into two
categories. The major items in each category are listed below:
Gas Chromatograph Detection
1. Varian Aerograph Series 1400 Gas Chromatograph
2. Leeds and Northrup 1 mv Recorder
3. Hewlett-Packard Model 3354 Computer System
4. Hewlett-Packard Model 1865A A/D Converter
Control Console System
1. Seiscor Model VIII Gas Sampling Valve
2. ATC Timers, Model 325A346A10PX
3. ASCO 4-way Midget Solenoid Valve, No. 8345B
4. Thomas Sample Pump, Model 106 CA18
5. Brooks Flowmeter, R-2-15-AAA with glass float, 0-15 Scale
6. Miscellaneous stainless steel and brass nuts, ferrules, unions, tees,
connectors., etc.
7. Miscellaneous stainless steel, copper, and Teflon tubing
8. Bendix Valve Oven
9. Column I, 8' x 1/8" SS, 80/100 mesh Porapak Q
10. Column II, 4' x 1/8" Teflon, 35/60 mesh Type 58 Silica Gel
330
-------
SOLENOID VALVE
P E
de-onerqizcd confiquration
VACUUM SAMPLE IN
enorai zed conf i ciuration
Figure 15. Solenoid valve G with gas sampling valve A.
331
-------
11. Column III, 15' x 1/8" SS, 15 percent 1,2,3-tris (cyanoethoxy) Pro-
pane on 60/80 Mesh Chromosorb PAW
12. Column IV, 2' x 1/8" SS, 40 percent HgSO4 on Chromosorb W
13. Miscellaneous electrical switches, wiring, lights, etc.
CALCULATIONS
The concentration of the individual hydrocarbons is compared to a cali-
bration blend with known concentrations of each of the components. The peak
areas are determined with a Hewlett-Packard Model 3354 computer system and
printed out on a remote teletype unit. The concentration of the sample is
determined with the equation:
Astd _ Asam
Cstd Csam
where Astd = the peak area of component n in the calibration blend
Cstd = the concentration of component n in the calibration blend
Asam = the peak area of component n in the sample
Csam = the unknown concentration of component n in the sample
n = the component of interest (i.e., methane, ethylene, ethane,
acetylene, propane, propylene, benzene, or toluene)
If the equation is solved of Csam, the result is:
Csam= Asam x Cstd
Astd
Example 1:
A bag sample was taken from the exhaust stream of a vehicle during a
driving cycle. The peak area of the calibration blend for methane was 11893
area counts with a concentration of 13.616 ppm C. The peak area of methane
in the sample was 8593 area counts. Calculate the concentration of methane
in the exhaust.
Asam x Cstd
Csam =
Csam =
Astd
(8593) (13.616)
(11893)
Csam =9.84 ppm C for methane
Example 2:
In the sample analysis, the peak area for toluene was 10973 area counts.
The peak area for the calibration blend was 18546 area counts for toluene
with a concentration of 21.272 ppm C. Calculate the concentration of toluene
in the sample.
332
-------
Asam x Cstd
csam =
(10973) (21.272)
Csam = - (18546) -
Csam = 12.59 ppm C for toluene
Note : These sample calculations are presented as an example only and are not
necessarily representative of expected values in exhaust.
333
-------
REFERENCES
Ultshuller, A. P. and Bufalini, J. J., Environ. Sci. Tech., Vol. 5, pg. 39,
1371.
Federal Register, Vol. 37, No. 221, 24270-77 Nov. 1971.
Klosterman, D. L. and Sigsby, J. E., Environ. Sci. Tech., Vol. 1, pg. 309,
1967.
Rasmussen, R. A. and Holdren, M. W., Chromatog. Newslett., Vol. 1, No. 2,
pg. 31, 1972.
Dimitriades, B. and Seizenger, D. E., Environ. Sci. Tech., Vol. 5, No. 3,
pg. 223, 1971.
Purcell, J. E. and Gilson, C. P., Chromatog. Newslett., Vol. 1, No. 2,
pg. 45, 1972.
Stuckey, C. L., J. Chromatog. Sci., Vol. 7, pg. 177, 1969.
Papa, L. J., Dinsel, D. L., and Harris, W. C., J. Gas Chromatog., Vol. 6,
pg. 270, 1968.
334
-------
APPENDIX G
HYDROGEN SULFIDE PROCEDURE
335
-------
THE MEASUREMENT OF HYDROGEN SULFIDE IN EXHAUST
The measurement of hydrogen sulfide in dilute automotive exhaust is
accomplished by collecting the dilute exhaust in glass impingers. The
absorbing solution is a buffered zinc acetate solution which traps the
sulfide ion as zinc sulfide. The exhaust is collected continuously during
the entire test cycle. Upon completion of the test, the absorbing solution
is treated with N,N dimethyl-paraphenylene diamine sulfate and ferric am-
monium sulfate. Cyclization occurs forming the highly colored heterocyclic
compound methylene blue (3,9-bisdimethylaminophenazothionium sulfate). The
resulting solution is analyzed with a spectrophotometer at 667 ran in a 1 cm
or 4 cm pathlength cell depending upon the concentration. The minimum de-
tectable concentration is 0.01 ppm.
SAMPLING SYSTEM
Dilute automotive exhaust samples are collected with the sampling
system shown in Figure 1. With this sampling system, the sample is re-
moved from the stream of dilute exhaust. A flow schematic of the sampling
system is shown in Figure 2. A solenoid valve is used to open and close
the sampling system from the CVS. The sample is bubbled through the ab-
sorbing solution to remove ^S. A series of two glass impingers is suf-
ficient to collect 99+ percent of H2S from the sample. The temperature of
the absorbing solution is maintained at 0-5°C with an ice bath slurry. The
gas stream is passed through a drier to remove any moisture. A pump capable
of pulling a flow rate of at least 4 £/min pulls the sample through the
system. The dry gas meter measures the volume of dried gas that has passed
through the system during each test cycle. The proper flow rate of 4 &/min
is continuously monitored with a flowmeter. The gas stream is then vented
to the atmosphere. This sampling system provides an efficient means of
trapping the dilute sample and enables rapid chemical analysis of the trapped
sample.
LIST OF EQUIPMENT
The equipment required in this analysis is divided into two basic cate-
gories. The first category involves the sample acquisition using the glass
impingers. The second category contains equipment related to the analysis
of the sample once it has been obtained. The individual items in each cate-
gory are listed below:
Sample Acquisition
1. Sample pump, Thomas Model 106 CA18, capable of free flow capacity
of 4 £/min.
336
-------
Front View
_Digital
Readout
Flowmeter
"Regulating
Valve
Close-up of Upper Front
Figure 1. The dilute exhaust sampling system for hydrogen sulfide,
337
-------
Solenoid-
Impinge
Ice Bath
Close-up of Impingers (Side View)
Pump
Figure 1 (Cont'd).
Rear View
The dilute exhaust sampling system tor hydrogen
338
-------
Gas Temperature
Digital Readout
U)
u>
On-Off
Solenoid
Flowmeter
Sample
Pump
Sample
~Probe
Dry
Gas
Meter
Regulating
Valve
hlalaiolHi
Diluta
Exhaust
Ice Bath
Temperature
Digital Readout
Gas Volume
Digital Readout
Figure 2. Hydrogen sulfide sample collection flow schematic.
-------
2. Glass impingers, Ace Glass Products, Catalog No. 7530-11, 29/42
bottle joints, 18/7 arm joints.
3. Flowmeter, Brooks Instrument Division, Model 1555, tube size
R-2-15-C, graduated 0-15, sapphire float, 0-5 H/min range.
4. Regulating valve, Nupro 4MG, stainless steel.
5. Dry gas meter, American Singer Corporation, Type Al-120, 60 CFH
capacity.
6. Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8" ID
and 5/16" OD x 3/16" ID.
7. Teflon solenoid valve, The Fluorocarbon Company, Model DV2-144NCA1.
8. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, con-
nectors, etc.
9. Drying tube, Analabs, Inc., Catalog No. HGC-146, 6" long, 1/4"
brass fittings.
10. Digital readout for dry gas meter.
11. Miscellaneous electrical switches, lights, wiring, etc.
12. Six-channel digital thermometer, Analog Devices, Model 2036/J/l.
13. Iron/constantan Type J, Thermo Sensors Corporation, single thermo-
couple, 1/4" OD stainless steel metal sheath.
Sample Preparation
1. Pipet-aid, Order No. JX-7290.
2. Class A, 10 m£ volumetric pipets.
3. Class A, 15 m& volumetric pipets.
4. Class A, 20 m& volumetric pipets.
5. Class A, 25 m£ volumetric pipets.
6. Class A, 100 ml volumetric flask.
7. Class A, 500 m£ volumetric flask.
8. Class A, 1000 m£ volumetric flask.
9- Class A, 25 m£ burets with Teflon stopcock.
10. Class A, 100 m& graduated cylinder.
340
-------
11. Micro burets, 5 m£ with Teflon stopcock.
12. Erlenmeyer flask, 250 m£.
13. Erlenmeyer flask, 500 m£.
14. Dropping bottle, 60 m£ with ground glass pipet and rubber bulb.
15. Beaker, 100 m£.
16. Repipet dispenser, 2 m£.
17. Repipet dispenser, 10 mil.
18. Centrifuge tube, 15 md.
19. Watch glass.
Instrumental Analysis
1. Beckman Model 25 spectrophotometer with recorder.
2. One cm pathlength disposable curvettes, CI regular.
3. Beckman UV silica cell, 40 mm pathlength, No. 580016.
LIST OF REAGENTS
This procedure requires the sample collection in glass impingers with
a buffered zinc acetate absorbing solution (0.25 M zinc acetate and 0.10 M
sodium acetate). An amine reagent (0.005 M N,N dimethyl-para-phenylene
diamine sulfate and 3.50 M sulfuric acid) and a ferric ion solution (0.25 M
ferric ammonium sulfate and 0.5 M sulfuric acid) are mixed with the absorb-
ing solution. Sodium sulfide is used as the standard. An iodometric
titration is used to standardize the sodium sulfide. All of the reagents
used in preparing these solutions are listed below along with the manu-
facturer and the quality.
1. Zinc acetate, dihydrate formula weight = 219.49; chemical formula
Zn{C2H3C>2) 2*2H2°» crystal, "Baker Analyzed" reagent.
2. Sodium acetate, anhydrous formula weight = 82.03; chemical formula1
NaC2H3O2, analytical reagent grade, powder, Mallinckrodt Code 7372.
3. N,N-dimethyl-pphenylene diamine sulfate, formula weight = 370.47;
chemical formula = (NH2CeH4N(CH3)2^H2S04, 98 percent minimum by
titration and spectro analysis, Eastman Code 1333.
4. Sulfuric acid, formula weight = 98.08; chemical formula = H2SO4,
ACS analytical reagent grade, Mallinckrodt Code 2876.
341
-------
5. Ferric ammonium sulfate, formula weight = 482.19; chemical formula =
Fe(NH4)(304)2*12H2°'ACS analytical reagent grade, crystals, Mallin-
ckrodt Code 5064.
6. Sodium sulfide, formula weight = 240.18; chemical formula =
Na,S*9H O, ACS analytical reagent grade, crystals, Mallinckrodt
Code 8044.
7. Iodine, formula weight = 253.81; chemical formula = 12, ACS
analytical reagent grade, resublimed, Mallinckrodt Code 1008.
8. Hydrochloric acid, formula weight = 36.46; chemical formula =
HC1, ACS reagent, assay 36.5-38.0 percent HC1, Eastman Code 13061.
9. Sodium thiosulfate, formula weight = 248.18; chemical formula =
Na2S2O3*5H20, ACS analytical reagent grade, crystals, Mallinckrodt
Code 8100.
10. Starch soluble; chemical formula = (CQE^QO^)n, certified ACS,
powder, Fisher Code S-516.
11. Arsenic trioxide primary standard, formula weight = 197.82;
chemical formula = AS2O3, powder, ACS analytical reagent, Mallin-
ckrodt Code 3668.
12. Potassium iodide, formula weight = 166.01; chemical formula =
KI, compacted crystal, "Baker Analyzed" reagent.
13. Sodium hydroxide pellets, formula weight = 40.00; chemical formula=
NaOH, caustic soda, ACS analytical reagent grade, Mallinckrodt
Code 7708.
14. Sodium bicarbonate, formula weight = 84.01; chemical formula =
NaHC03, powder, ACS analytical reagent grade, Mallinckrodt Code 7412.
15. Sodium carbonate anhydrous powder, formula weight = 105.99;
chemical formula = Na2CO3, ACS analytical reagent grade, Mallin-
ckrodt Code 7521.
16. Potassium dichromate, crystal, primary standard, formula weight =
294.22; chemical formula = K2Cr2O7, "Baker Analyzed" reagent.
17. Litmus test paper, blue, reagent ACS, Fisher Code 14-875.
18. Glycerol, formula weight = 92.10; chemical formula =
HOCH2CH(OH)CH20H, ACS analytical reagent, Mallinckrodt Code 5092.
PREPARATION OF REAGENTS
The various reagents needed for the analysis of hydrogen sulfide can be
divided into two separate categories. The first category includes the
342
-------
reagents for sample acquisition and color development. The other category
encompasses the standard solutions and solutions used for standardization
The chemicals used to make these solutions are ACS analytical reagent grade
with the exception of potassium dichromate and arsenic trioxide which are
a primary reagent grade.
Sample Acquisition and Color Development
Absorbing Reagent — The absorbing reagent is prepared by dissolving
54.9 g (0.25 mole) of zinc acetate, 8.2 g (0.10 mole) of sodium acetate,
and 40 m& (0.56 mole) glycerol, in vacuum boiled, deionized water and diluted
to 1 liter. A 2 mi portion of 0.05 M sodium sulfide solution is added
dropwise to the diluted solution with vigorous shaking. This removes traces
of heavy, metals by precipitating them as their insoluble sulfides. This
solution is then set aside overnight. The resulting solution is filtered
through a fine textured paper with the first 50 mfc portion being discarded.
This solution is stable for at least one (1) week. A gradually developing
cloudiness is of no consequence.
Amine Solution — A 0.005 M solution is prepared by dissolving 0.93 g
of N,N dimethyl-para-phenylene diamine sulfate in 75 m& of deionized water.
Then, 197 m£ of concentrated sulfuric acid is slowly added and the mixture
is allowed to cool. The resulting solution is diluted to 1 liter with
deionized water. This solution is stable for about six (6) months.
Ferric Ion Solution — The ferric ion solution is prepared by dissolving
120.6 g (0.25 mole) of ferric ammonium sulfate in 750 m£ of deionized water.
A 27 m£ portion of concentrated sulfuric acid is added and the solution is
allowed to cool before diluting to 1 liter with deionized water. The solu-
tion is stable for one (1) month.
Standards and Standardization Sojlutions
Sulfide Standard Solution — Approximately 8 g (0.03 mole) of sodium
sulfide is rinsed with vacuum boiled, deionized water to remove traces of
sulfide from the surface of the crystals. The crystals are then dissolved
in vacuum boiled, deionized water and diluted to 1 liter to give a concen-
trated solution. A more dilute solution of sulfide ion is prepared by
diluting 10 ml of the concentrated solution to 1 liter with vacuum boiled,
deionized water. This solution is standardized by iodometric titration.
The approximate concentration is 10 ppm sulfide ion by weight. This solu-
tion should be prepared immediately before use.
Thiosulfate Standard Solution ~ Approximately 2.5 g (0.01 mole) of
sodium thiosulfate is dissolved in 500 mSL of freshly vacuum boiled, de-
ionized water in a 1 liter volumetric flask. A 0.1 g (0.001 mole) portion
of sodium carbonate is added and the solution is stirred until dissolved.
The solution is then diluted to 1 liter and stored in the dark. Addition
of a small amount of sodium carbonate prevents the formation of hydrogen
sulfite ion from thiosulfate ion in the presence of acid. The addition r^
such substances as chloroform, sodium benzoate, or mercury (II) iodide in
hibits the growth of bacteria. This solution is stable for several weeks
343
-------
but should be discarded if it becomes turbid. It is standardized with a
potassium dichromate primary standard.
Starch Indicator — The indicator is prepared by making a paste of 0.5 g
of soluble starch in 2 or 3 m£ of deionized water. The resulting slurry is
slowly poured into 50 m£ of boiling deionized water and heated until clear
(about 2-3 minutes) . The solution is then cooled and centrifuged for several
minutes. The supernatant liquid is decanted into a clean, 60 m£ reagent
bottle equipped with a pipet. This aqueous starch solution will decompose
due to bacterial action in several days. This can be prevented by storing
the indicator under sterile conditions, the addition of mercury (II) iodide
to inhibit the bacterial action, or by preparing fresh daily.
Iodine Solution — About 25 g (0.15 mole) of potassium iodide is
dissolved in 10 m& of deionized water. Then, 12.7 g (0.05 mole) of iodine
crystals are added and dissolved with occasional stirring. The solution is
filtered and diluted to 1 liter with deionized water. The resulting solution
is approximately 0.1 N and is used to prepare the 0.01 N solution. The 0.1 N
solution is standardized against arsenic trioxide. The dilute iodine solution
is prepared by diluting 10 mJl of the 0.1 N solution to 100 m£ with deionized
water. This solution is standardized with a previously standardized thio-
sulfate solution. The dilute iodine solution is prepared and standardized
daily because of the volatility of iodine and oxidation by dissolved oxygen.
DESCRIPTION OF METHOD
Hydrogen sulfiSe is collected in a buffered zinc acetate absorbing re-
agent from dilute automotive exhaust. The highly colored species, methylene
blue, is generated by the addition of N,N dimethyl-para-phenylene diamine
sulfate and ferric ion. Methylene blue has an absorbance maxima in the red
region of the visible spectrum. The extinction wavelength curve for methylene
blue is shown in Figure 3.
The analytical procedure for the determination of hydrogen sulfide in
dilute automotive exhaust consists of two major areas. The first is the
standardization and calibration of the standard sulfide ion solution. The
second area includes the sample acquisition and color development of the
hydrogen sulfide sample. Each will be discussed in detail below. The
analysis flow schematic for this procedure is shown in Figure 4.
Standardization and Calibration
The thiosulfate solution is standardized against primary standard grade
potassiun\ dichromate. The potassium dichromate is dried in an oven at 150°
to 200°C for 1 to 2 hours. A weighed 0.10 to 0.15 g (0.001 mole) portion of
dried potassium dichromate is placed in a 500 m£ Erlenmeyer flask and dis-
solved in 50 m£ of deionized water. A freshly prepared solution-of 3 g
(0.02 mole) of potassium iodide, 5 m£ of 6 N hydrochloric acid and 50 Bl£ of
deionized water is then added. This solution is gently swirled, covered
with a watch glass, and allowed to stand for five (5) minutes. The sides of
the flask are then washed with deionized water followed by approximately 200
m£ of deionized water. The resulting solution is titrated with the standardized
344
-------
400
450
500 550 600
Wavelengths, nanometers
650
700
750
Figure 3. Extinction-wavelength curve of methylene blue.
345
-------
CVS
Class
Impinger
Permeation
Tube-
Reagent
Addition
KCr207
Primary
Standard
AS203
Primary
Standard
Color
Development
Thiosulfate
Standardization
Iodine
S t andardi zation
Absorbance
Reading
Calibration
Curve of
Sulfide Standard
Figure 4. Hydrogen sulfide analysis flow schematic.
346
-------
thiosulfate solution. As the end point is approached, about 5 m£ of the
starch indicator is added. The solution is blue from the starch-iodine
complex before the end point is reached. At the end point, there is a
change in color from blue to green due to the production of Cr (III) ion
This standardized thiosulfate solution is used to standardize the dilute"
iodine solution and also determine the sulfide ion concentration in the
sulfide standard solution.
The concentrated iodine solution is standardized against primary stand-
ard grade arsenic trioxide by an iodimetric titration. A weighed portion of
0.15 to 0.20 g (0.001 mole) arsenic trioxide is placed into an Erlenmeyer
flask. Then, 10 to 20 mfc of 1.0 M sodium hydroxide are added to dissolve
the solid. With a small piece of blue litmus paper as the indicator, 1.0 M
hydrochloric acid is added dropwise until the arsenic trioxide solution is
slightly acidic. About 1.0 g (0.01 mole) of sodium bicarbonate is slowly
added to prevent the loss of solution due to the effervescence of carbon
dioxide. The resulting solution is diluted to about 100 m£. . About 2 mil
of the starch indicator is added and the solution is then titrated with the
0.1 N iodine solution. During the titration, the top of the buret should be
covered to minimize the volatilization of iodine. The standardization needs
tfli.be done only once for the concentrated solution.
The dilute iodine solution (0.01 N) is standardized against a previously
standardized thiosulfate solution. To a beaker containing 25 m£ of the
dilute iodine solution, 10 mH of concentrated hydrochloric acid is added and
the resulting solution is immediately titrated with the standardized thio-
sulfate solution. Just before the endpoint, starch is added to serve as an
indicator. The filute iodine solution should be standardized daily.
The concentration of the standard sulfide solution is determined by an
iodometric method. To three (3) Erlenmeyer flasks, 10 mJl of the absorbing
solution and 50 mJl of the sulfide standard solution are added and the re-
sulting solution is mixed. Into one of the flasks 5 mH of the 0.01 N
iodine solution and 19 m£ of concentrated hydrochloric acid are added. The
resulting solution is immediately titrated to the starch endpoint with stan-
dardized thiosulfate solution. This procedure is then repeated for the re-
maining two flasks and for two blanks prepared with only the absorbing reagent
and 50 mA of vacuum boiled, deionized water. The excess iodine in the solution
is reacted with the thiosulfate and the amount of sulfide ion present can be
back-calculated.
A Beer's Law curve is determined by adding 0.1, 0.25, 0.5, 1.0, 2.0,
3.0, and 5.0 mA of the sulfide standard solution to seven separate 100 nd
volumetric flasks with each containing 10 m£ of the absorbing reagent. A
blank of the absorbing reagent is also prepared. A 10 m& portion of the
amine solution is carefully poured down the side of each flask taking care
n°t to introduce air bubbles. Then, 2 m£ of the ferric ion solution is
added and gently swirled. The contents are then diluted to 100 m£ with
vacuum boiled, deionized water and placed in the dark to develop for 30
minutes. This procedure should be repeated for the remaining flasks and the
blank. The absorbance is read at 667 nm in a spectrophotometer against the
reagent blank. The concentration of the samples are used to determine a
best fit plot of the Beer's Law curve. In the higher concentration ranges,
347
-------
One and four cm pathlength curvettes
Reading the absorbance
Figure 5. The analysis of hydrogen sulfide in exhaust.
348
-------
• ••
,!
"
Figure 6. Beckman Model 25 spectrophotometer.
-------
the curve is nonlinear and does not necessarily follow Beer's Law-
Another more efficient means of generating a Beer's Law curve requires
the use of permeation tubes as a calibration standard. The calibration
curve is generated by passing a diluent gas (preferably nitrogen) over the
permeation tube and through a set of two impingers containing the buffered
zinc acetate absorbing reagent. The length of time sampled is proportional
to the concentration in the set of impingers. Not only is this a quick and
efficient means of calibration, but it also takes into account the collection
efficiency of the sampling technique. The samples are developed by the tech-
nique described above. The absorbance is read at 667 nm and this absorbance
is used to determine the calibration curve. This best fit plot of the Beer's
Law curve is used to determine the concentration of the samples. A quick
and easy method of providing a daily calibration for the instrument can be
obtained with this method.
Sample Acquisition and Color Development
To two clean glass impingers, 10 m& .of acetate buffer and 40 mH of vacuum
boiled, deionized water are added. These impingers are connected in series
in the sampling system. During each test cycle, a portion of the diluted
exhaust is bubbled through the absorbing reagent at a flow rate of 4.0 £/min.
Upon completion of each driving cycle, the impingers are replaced with fresh
ones. To each of the collected samples, 10 mJl of the amine solution is
added through the top of the impinger and gently swirled. Then 2 m!L of the
ferric ion solution is added in the same manner and mixed for 30 seconds.
The solution is quantitatively transfered from the impingers to a 100 m£
volumetric flask and diluted to 100 m£ with vacuum boiled, deionized water.
The color development is complete in 30 minutes. The procedure is repeated
with the remaining samples. After 30 minutes, the absorbance is measured on
a spectrophotometer at 667 nm against a reagent blank and the concentration
determined from the calibration curve. The one and four cm curvettes are
shown in Figure 5 and the spectrophotometer is shown in Figure 6.
This procedure is well documented and commonly used for the analysis of
hydrogen sulfide. With a 4 cm cell, the minimum detectable quantity is on
the order of 0.01 ppm. Extensive wet chemistry is involved to establish the
calibration curve but the sample analysis time is minimal after color develop-
ment.
CALCULATIONS
This procedure has been selected and developed to determine the quantity
of hydrogen sulfide in dilute exhaust. A Hewlett-Packard 65 program was de-
signed to reduce the time required for manual calculations. The derivation
of the equations are given below and a copy of the steps in the program are
shown in Figure 7.
Derivation of Equation
The concentration of sulfide ion in yg/mJl is obtained from the Beer's
Law plot of the absorbance. This is the actual concentration in the impinger
350
-------
HP-65 User Instructions
rammer
fu
IUL
E. Robert Fanick
HYDROGEN SULFIDE IN Uq/m3
— L 1 — i i !*
Page of
— DatP 8/25/78
/— ^
u i i
J — i — m
STEP
°1
0-,
°1
04
1
2
3
4
5
6
•J
8
9
10
tl
12
INSTRUCTIONS
Switch To On; Switch To Run
Feed Card From Left To Right
Initialize
Set Decimal Place
Input Sample Volume
Input Barometric Pressure
Input Sample Temperature
Input Dilution Factor
Input Solution Volume
input Bubbler #1 Concentration
Output Bubbler SI Concentration
Input Bubbler S2 Concentration
Output Bubbler #2 Concentration
Output- Sum Total Concentration
nn-t-nnt- Tnt-sl fonrpntration
INPUT
DATA/UNITS
FT3
"Hg
°F
n&
Ug/mi
Ug/m£
->
KEYS
I!
1!
f j[ REG
DSP || 2
A II
R/S ||
R/S ||
R/S ||
K/S j|
R/S
1 II
R/S i!
R/S II
II
R/S ||
R/S ||
i II 1
II
H
II
II
1
Jl
II 1
L_JI 1
1 J
OUTPUT
DATA/UNITS
°Sm3
"Hd
°Rm3
1
m3
1
nj
mi.
^
Ug/m3
ug/m3
Ug/m3
pg/m3
pg/m3
ppm
Figure 7. HP-65 User Instructions
351
-------
HP-85 Program Form
Title HYDROGEN SULFIDE IN EXHAUST (2 BUBBLER CALCULATION)
Page of
SWITCH TO W/PflGM. PBfSS ' I ' PRGM 1 TO CLEAfl MEMORY.
KEY
ENTRY
LBL
A
2
T
R/S
X
STO 1
R/S
4
10 6
0
+
RCL 1
T
' R/S
X
R/S
X
STO 2
2&CL 2
R/S
X
1
.
0
6
3
X
STO 3
30 R/S
RCL 2
R/S
X
1
.
0
fi
3
X
<0 R/S
RCL 3
+
R/S
0
.
0
0
0
7
50 5
CODE
SHOWN
23
11
02
81
84
71
3301
84
04
06
00
61
3401
81
84
71
84
71
3302
3402
84
71
01
83
00
06
03
71
3303
84
3402
84
71
01
83
00
nfi
03
71
RA
3403
61
84
00
83
00
00
00
07
05
COMMENTS
Input-Sample Vol., Ft3
Input-Barometer, "Hg
Input-Sample Temp. , 'F
Input Dilution Factor
Input Abs. Soln. Vol.mJl
Input Sample Cone. , U9
Bubbler #1
•
Output Sample Cone. , Ug
Bubbler #1
Input Sample Cone., yg/
Bubbler #2
nnt-piif- ^flmnl & Pnnr1 pg
Bubbler #2
Output Total Cone. , U<3
KEY
ENTRY
0
2
X
R/S
RTK
60
70
"/mi
3oS/m3
=/mi
S0,<5/m3
25/m3
XX!
CODE
SHOWN
00
02
71
84
24
COMMENTS
Output-PPM H2S
-
REGISTERS
Ri
R2
R3
R4
Rs
R6
R7
Ra
Ro
LABELS
A
B
C
D
F
n
1
9
3
4
' 5
8
7
R
P
FLAGS
1
?
-Figure 7 (Cont'd). HP-65 Program Form.
352
-------
and does not take into account any of the experimental parameters The first
correction is for the dilution factor and the absorbing solution volume in
•I:
Csam = Cex x D.F, x A.S.V.
(y
-------
The concentration in ppm is:
16289 yg 5= v 0.0007502 ml, S= 34.08 yg H2S
, _H2__ x _____ x 32>Q6 ^ sg
ppm H2S = 13.0
The values for the six sets of sample data are given in Figure 8 along with
the data sheet. The calculations were carried out manually and with the
HP-65 program. The program will be used to calculate both yg S=/m3 and ppm
H2S and to insure rapid data turnaround. The values used in these calcula-
tions were picked at random and may not represent the expected values in
exhaust.
354
-------
HYDROGEN SULFIDE SAMPLE COLLECTION DATA SHEET
SWRI PROJECT NO, / / - OOOO -OOP TEST NO. OOP TEST DATE! 2/6/78 VEHICLE; SAMPLE.
FUEL'. EM-237 CVS NO. & TUNNEL SIZE 8" DRIVER! ABC MILES! OOOO
SAMPLE COLLECTION BY! BKF CHEMICAL ANALYSIS BY!_££F_CALCULATIONS BYi
GENERAL COMMENTS:.
OJ
Un
NO,
1
I
3
4
5
%
1
8
X
DRI-VI-NG
CYCLE
•
X
SAMPLING CONDITIONS
VOLUME'
FT3
3J85
2.813
2.8OO
2.805
2.75^
2.32O
©
B.P,
"HG
21.W
2^1
2?-38
2^.^-3
2^.92
•?'*? 35
*-«. » • --«>^ .^.^
©
TEMP,
°F
77
80
85
77
7k>
^0
©
DILU-
TifftM •
t 1 UIN
FACTOR
2
2
2
2
2
2
©
ABSORB
REAGEN
VOLUME
ML
50
50
50
50
so
50
®
SAMPLE
ABSORB ,p s=£
O.ltf j Jf.2
0.723 1 C6. 5
0.5/1 U-7.O
/.O67 \JOO.O
0.875 JS/.&
0.772 \7l.l
I
^>H ff\
s^^J W
^Wl
PPM
/3.0
67.3
W.7
103, 8
8^7
S^.^
®
Figure 8. Raw data sheet for hydrogen sulfide analysis
-------
REFERENCES
Stecher, P. G., (Ed.), The Merck Index, An Encyclopedia of Chemicals and
Drugs, 8th Edition, Merck and Co., Inc., Rahway, N. Y., 1968, pg. 545.
Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N. J., 1971, pg. 319.
Bethea, R. M., J. Air Poll. Cont. Assoc., Vol. 23, pg. 710, 1973.
Bamesberger, W. L. and Adams, D. F., Tappi, Vol. 52, pg. 1302, 1969.
Falgout, D. A. and Harding, C. I., J. Air Poll. Cont. Assoc., Vol. 18,
pg. 15, 1968.
Biles, B., Brown, C., Nash, T., J. Phys. E.: Scientific Instruments,
Vol. 7, pg. 309, 19"?4.
Wellinger, R. and Fiever, P. M., Am. Ind. Hyg. Assoc. J., Vol. 35, pg.
730, 1974.
Natusch, D. F. S., Sewell, J. R., and Tanner, R. L., Anal. Chem., Vol.
46, pg. 410, 1974.
"Trace Sulfur in Hydrocarbons by the Raney Nickel Method," Analytical Method
Information, Analytical and Informational Division, Exxon Research and
Engineering Co., September 1975.
Pierce, R. W., ISA Trans., Vol. 13, pg. 291, 1974.
Axelrod, H. D., Gary, J. H., Bonelli, J. E., and Lodge, J. P., Anal. Chem.,
Vol. 41, pg. 1959, 1969.
Natusch, D. F. S., Klonis, H. B., Axelrod, H. D., Teck, R. J., Lodge, Jr.,
J. P., Anal. Chem., Vol. 44, pg. 2067, 1972.
Bock, R. and Puff, H., Z. Anal. Chem., Vol. 240, pg. 381, 1968.
Nauman, R. and Weber, C., Z. Anal. Chem., Vol. 253, pg. Ill, 1971.
Ehman, D. L., Anal. Chem., Vol. 48, pg. 918, 1976.
Hseu, T. M. and Rechnitz, G. A., Anal. Chem., Vol. 40, No. 7, pg. 1054,
1968; see also correction: Ibid, Vol. 40, No. 11, pg. 1661, 1968.
356
-------
Baumann, E. W., Anal. Chem., Vol. 46, No. 9, pg. 1345, 1974.
Blanchette, A. R. and Copper, A. D., Anal. Chem., Vol. 48, pg. 729, 1976,
"Analysis of H2S and SO2 in the ppb Range," Bulletin 722C, Supelco, Inc
Belefonte, Pa., 1975.
Setter, J. R., Sedlak, J. M., Blurton, K. F., J. Chromatog. Sci., Vol 15
pg. 125, 1977.
Greer, D. G. and Bydalek, T. J., Environ. Sci. Tech., Vol. 7, pg. 153,
1973.
Obermiller, E. L. and Charlier, G. O., Anal. Chem., Vol. 39, pg. 397,
1967.
Bethea, R. M. and Meador, M. C., J. Chromatog. Sci., Vol. 7, pg. 655,
1969.
Jones, C. N., Anal. Chem., Vol. 39, pg. 1858, 1967.
Wilhite, W. F. and Hollis, O. L., J. Gas Chromatog., Vol. 6, pg. 84, 1968.
Thornsberry Jr., W. L., Anal. Chem., Vol. 43, pg. 452, 1971.
Adams, D. F., Jensen, G. A., Steadman, J. P., Koppe, R. K., and Robertson,
T. J., Anal. Chem., Vol. 38, pg. 1094, 1966.
Martin, R. L. and Grant, J. A., Anal. Chem., Vol. 37, pg. 644, 1965.
Adams, D. F., Bamesberger, W. L., and Robertson, T. J., J. Air Poll. Cont.
Assoc., Vol. 18, pg. 145, 1968.
levaggi, D. A., Siu, W., and Fedlstein, M., Advances in Automated Analysis,
Vol. 8, pg. 65, 1972.
"Hydrogen Sulfide in Air Analytical Method," H. L. S., Vol. 12, pg. 362,
1975.
Bamesberger, W. L. and Adams, D. F., Environ. Sci. Tech., Vol. 3, pg. 258,
1969.
Gustafsson, L., Talanta, Vol. 4, pg. 227, 1960.
Moest, R. R. , Anal. Chem., Vol. 47, pg. 1204, 1975.
B"ck, M. and Gies, H., Staub, Vol. 26, pg, 1966.
Jacobs, M. B., Braverman, M. M., and Hochheiser, S., Anal. Chem., Vol. 29,
P9. 1349, 1957.
G. c. B., Tappi, Vol. 46, pg. 1, 1963.
357
-------
Budd, M. S. and Bewick, H. A., Anal. Cheni. , Vol. 24, pg. 1536, 1952.
Bostrom, C., Air and Water Poll. Int. J., Vol. 10, pg. 435, 1966.
Flamm, D. L. and James, R. E., Environ. Sci. Tech., pg. 159, 1975.
358
-------
APPENDIX H
TOTAL CYANIDE PROCEDURE
359
-------
THE MEASUREMENT OF TOTAL CYANIDE IN EXHAUST
The measurement of total cyanide (hydrogen cyanide + cyanogen) in exhaust
is accomplished by collecting the exhaust in glass impingers. The 1.0 N
potassium hydroxide absorbing solution is maintained at ice bath temperature
(0°C) . An exhaust sample is collected continuously during the test cycle.
Upon completion of the test, an aliquot of the absorbing reagent is treated
with KH PO. and Chloramine-T. A portion of the resulting cyanogen chloride
is injected into a gas chromatograph equipped with an electron capture
detector (BCD). External CN- standards are used to quantify results. The
minimum detection limit is less than 0.01 ppm for this procedure.
SAMPLING SYSTEM
Glass impingers are used to collect exhaust samples for subsequent chemi-
cal reaction with chloramine-T in a buffered solution. A flow schematic of
the sample collection system is presented in Figure 1. Two glass impingers
are sufficient to collect 99+% of the HOST and C2N2, provided the impingers
are maintained at 0-5°C with an ice bath. The temperature of the gas stream
is monitored by a thermocouple immediately prior to the dry gas meter. The
dry gas meter determines the total flow through the impingers during a given
driving cycle. The sample pump should be capable of pulling a flow rate of
4 liters/minute. A drier is included to prevent condensation in the pump,
flowmeter, dry gas meter, etc. The flowmeter allows continuous monitoring
of the sample flow to insure proper flow rates during the sampling. Figure 2
shows the actual sampling system used in the analysis.
LIST OF EQUIPMENT
The equipment required in this analysis is divided into two basic cate-
gories. The first category involves the sample acquisition using the glass
impingers. The second category contains equipment related to the analysis
of the sample once it has been obtained. The individual items in each cate-
gory are listed below:
Sample Acquisition
1. Sample pump, Thomas model 106 CA18, capable of free flow
capacity of 4 H/m±n
2. Glass impingers, Ace Glass Products, catalog no. 7530-11
29/42 bottle joints, 18/7 arm joints
3. Flowmeter, Brooks Instrument Division, Model 1555, Tube size
R-2-J.5-C, graduated 0-15, sapphire float, 0-5 %,/min range
360
-------
u>
o>
H
On-Off
Solenoid
Valve
Sample
Probe
Gas Temperature
Digital Readout
Flowmeter
Impinger
Sample
Pump
Dilute
Exhaust
Dry
Gas
Meter
Regulating
Valve
Ice Bath
Temperature
Digital Readout
Gas Volume
Digital Readout
Figure 1. Total cyanide sample collection flow schematic.
-------
Front View
Digital
Readout
Flowmeter
Regulating
" Valve
Close-up of Upper Front
Figure 2. Total cyanide sampling system.
362
-------
Solenoid
Close-up of Impingers (Side View)
Dry Gas Meter
limp
Rear View
Figure 2 (Cont'd). Total cyanide sampling system.
363
-------
4. Regulating valve, Nupro 4MG, stainless steel
5. Dry gas meter, American Singer Corporation, Type AL-120,
60 CFH capacity
6. Teflon tubing, United States Plastic Corporation, 1/4" OD x
1/8" ID and 5/16" OD x 1/8" ID
7. Teflon Solenoid Valve, The Fluorocarbon Company, Model DV2-
144NCA1
8. Miscellaneous teflon nuts, ferrules, unions, tees, clamps,
connectors, etc.
9. Drying tube, Nalgene Corporation, 10 cm length x 1/2 in. dia.
10. Digital readout for dry gas meter
11. Miscellaneous electrical switches, lights, wirings, etc.
12. Model 2036/J/l, Analog Device, 6 channel, digital thermometer
13. Iron/Constantan, Thermo Sensors Corporation, type J single
thermocouple, 1/4" OD stainless steel metal sheath
Sample Preparation
1. Glass gas syringe, teflon tipped plunger, 100 yl, Pressure-Lok
Series A-2, Alltech Associates
2. Glass Reacti-vials, 5 mi, Pierce Chemical Company
3. Class A, I mi volumetric pipets
4. Class A, 2 mi volumetric pipets
5. Class A, 25 mi volumetric pipets
6. Class A, 50 mi volumetric flask
7. Class A, 100 mi volumetric flask
8. Class A, 250 mi volumetric flask
9. Class A, 500 mi volumetric flask
10. Class A, 1000 mil volumetric flask
Instrumental Analysis
1. Perkin-Elmer Model 3920B gas chromatograph equipped with a
linearized electron capture detector (BCD)
364
-------
2. Soltec Model B-281 1 mv recorder
3. Hewlett-Packard Model 3354 gas chromatograph computer system
with remote teletype printout
LIST OF REAGENTS
This procedure requires the sample collection in a glass impinger using
1.0 N potassium hydroxide absorbing reagent. A buffer is added to control
the pH and Chloramine T is added to convert the CN- to cyanogen chloride.
Potassium cyanide is used as the CN- standard in l.ON KOH. The reagents are
listed below along with the manufacturer and quality.
1. Potassium phosphate monobasic, formula weight = 136.09, chemical
formula = KH2PO4, ACS Analytical Reagent Grade, crystals, Mallinc-
krodt Code 7100.
2. Potassium hydroxide, formula weight = 56.11, chemical formula =
KOH, ACS Analytical Reagent Grade, pellets, Mallinckrodt Code 6984
3. Potassium cyanide, formula weight = 65.12, chemical formula KCN,
ACS Analytical Reagent Grade, granular, Mallinckrodt Code 6881
4. Chloramine-T (sodium para-toluene sulfonchloramide trihydrate) ,
formula weight = 282.70, chemical formula = l-O^CeH^SC^NClNa-S^O
Assay (by titration) 96% minimum, Eastman, crystals, Eastman
Code 1022
PREPARATION OF REAGENTS
Primary standard - the primary standard is prepared by dissolving 0.602 grams
of KCN in 500 ml of 1.0 N KOH. This is equivalent to 500 ppm HCN ( 500 yg
HCN/mi) or a 481 ppm CN- (481 yg CN-/mJl) . Additional standards are prepared
from the primary standard. A typical dilution to prepare a 0-10 yg CN-/m&
calibration curve is as follows:
m£ of 481 yg CN-/m£ Final Diluent CN- concentration,
Volume, mi yg/CN-/m£
1.000 mi
4.000 mi
1.000 mi
3.00 mi
1.000 mi
1.000 mi
Buffer Solution - A 1.0 N
J-J.609 g KH2P04 to 100 mi
50.0 mi
250.0 m£
100.0 m£
500.0 mi
250.0 m£
500.0 mi
KH2PO4 buffer solution
of deionized H2O. The
9.62
7.70
4.81
2.89
1.92
0.96
is prepared by adding
buffer solution should be
^5°£bing_Reagent - The absorbing reagent is a l.ON KOH solution. This
sol^i5n~is prepared by dissolving 56.11 grams of KOH in 1000 mi of dexonized
365
-------
water. The absorbing reagent is stable for one week.
Chloramine-T - The Chloramine-T is the primary reagent responsible for
converting the CM- to CNC1 that is measured using a GC-ECD system. This
reagent is prepared by dissolving 250 mg in 100 m£ of deionized water.
This reagent is the most critical in this procedure and should be prepared
daily.
ANALYTICAL PROCEDURE
The analysis of total cyanide (HCN + C2N2) in automotive exhaust is
accomplished with the use of an electron capture detector (BCD). This
particular detector is highly sensitive to halogens and halogenated com-
pounds. Chloramine-T (sodium para-toluene sulfonchloramide) oxidizes the
cyanide ion to CN+ which in turn combines with chloride ion to form cyanogen
chloride (CNC1). Total cyanide can then be analyzed by the following pro-
cedure. The analysis flow schematic for the total cyanide procedure is shown
in Figure 3.
During each test cycle a portion of the diluted exhaust is bubbled
through 25 m£ of l.ON potassium hydroxide absorbing reagent in two impingers
at a flow rate of 4.0 Vmin. Upon completion of each driving cycle the im-
pingers are replaced with fresh ones. A 1 m& aliquot is removed from one
of the impingers and placed in a 5 m£ Glass Reacti-vial. A 2 m& aliquot
of l.ON potassium dihydrogen phosphate buffer is then added carefully down
the side of the vial. This adjusts the pH to neutral or slightly acid. A
1 m£ aliquot of Chloramine-T is carefully added down the side of the vial
to the buffered solution. Turbulent addition of this reagent can cause
premature release of cyanogen chloride. The cap with a septum top is
immediately screwed tightly into place. The resulting solution is then set
aside for 5 minutes. This allows the Chloramine-T to react completely with
the trapped cyanide ion. The vial is vibrated for 5 seconds to release
cyanogen chloride into the gas phase. A 100 u& sample of the head space
is removed through the septum top of the vial using a gas-tight glass
syringe and immediately injected as a single slug into the gas chromatograph.
This procedure is then repeated for the second impinger. The step-wise
procedure for total cyanide analysis is shown in Figure 4.
A Perkin-Elmer 3920B gas chromatograph with an BCD is used to analyze
the sample. A 6' x 1/8" stainless steel column packed with 100/120 mesh
Porapak Q is employed to analyze the sample. The carrier gas is argon/5 per-
cent methane at a flow rate of 40 mjt/min. and the column temperature is iso-
thermal at 140°C. In the order of elution, the peaks are oxygen, carbon di-
oxide, water, and cyanogen chloride. The peak area is determined with a
Hewlett-Packard Model 3354 computer system with a remote teletype printout.
The peak area is compared to a standard cyanide ion solution which is deve-
loped in a manner similar to that of the sample. Figure 5 shows the actual
analytical system with gas chromatograph, detector, integrator, and recorder.
This procedure provides a rapid and sensitive method of analyzing HCN
and C2N2 in exhaust without extensive wet chemical work up. The analysis
time is on the order of about 5 minutes after injection into the gas
366
-------
CVS
Glass Impinger
I
Excess Reagent
saved as needed
Aliquot buffered
with KH2P04
Chloramine-T
added
CNCl in head gas
analyzed in GC
with BCD
A/D Converter
i
Recorder
Hewlett-Packard
3354 GC
computer system
Figure 3. Total Cyanide (HCN +
Analysis Flow Schematic.
367
-------
Step 1. Glass Reacti-vial with septum cap
Step 2. Aliquot removal
Figure 4. Various steps in sample collection
and analysis of total cyanide in exhaust.
368
-------
Step 3. Reagent addition
Step 4. Sample shaking
Figure 4 (Cont'd). Various steps in sample collection
and analysis of total cyanide in exhaust.
369
-------
Step 5. Head gas removal
Step 6. Sample injection
Figure 4 (Cont'd). Various steps in sample collection
and analysis of total cyanide in exhaust.
370
-------
Figure 5. Total cyanide analytical system.
-------
chromatograph. The sensitivity of the BCD extends the minimum detectable
limit to less than 0.01 ppm cyanide ion with the specified flow rates, ab-
sorbing solution volume, syringe size, vial size, and reagent quantities.
This limit can possibly be extended by changing these parameters. The sim-
plicity and rapid data turnover makes this procedure ideal for repetitive
analysis. A gas chromatograph trace for total cyanide in dilute gasoline ex-
haust is shown in Figure 6 and a chromatogram for the cyanide standard is
shown in Figure 7.
CALCULATIONS
This procedure has been developed to provide the user with the as-
measured total CN- in the exhaust. Calculation of the results are expressed
in ug CN-/m3 of exhaust and ppm CN-. In order to allow implementation of
this analytical procedure, several items related to the final calculation of
total cyanide are included in this section. The derivation of the equation
used in calculating emission concentrations is included for record. The basic
calculations are straightforward and a Hewlett-Packard 65 program was devel-
oped to provide rapid data turnaround during a given day's run.
DERIVATION OF EQUATION
Let Csam = concentration of sample, yg/m£
Cstd = concentration of standard, yg/m&
Asam = GC peak area of sample, relative units
Astd = GC peak area of standard, relative units
Since the gas chromatograph electron capture detector (BCD) has demonstrated
a linear response in the CN- concentration range of concern, the following
equation holds true:
Csam _ Cstd
Asam Astd
Cstd x Asam
Csam =
Astd
The Csam in the solution is corrected for any necessary dilution by multi-
plying by the dilution factor. At this point, Csam is still expressed in yg
CN-/m£ absorbing reagent.
raam - Cstd, yg/mjx Asam x D.F.
Csam = ——
Astd
To obtain the total cyanide in the absorbing solution, the absorbing reagent
volume (expressed in mJl) is multiplied to get the equation:
„ _ _ Cstd, yg/mJl x Asam x D.F. x Abs. vol, m&
L>SCUU ~— INI iJTt :IIL _ii _ . __._ _, _,_ .
(ug/ft3) Astd
372
-------
ll~2~7
Sample GASOUtle.
Instrument PE.
Column
Operator BOB FANICK
D. S.S. Type
Liq. Phase
Support
cc/min. Air/SZCfjCamer
on fOOX/aOmesh PORAPAK
Run ISO @ IfrO °C. using
fO psig. A/A Rotameter Rdg.
Inlet 2QQ °C HEATED 5T4/«K.eSS ST-ESL Type
Detector: 325 °C £Cg^ Type (other)
Hyd. A/> _ psig f/
Air A/A psig H
( ) NA psig
/
Rotameter Rdg.
Rotameter Rdg.
mV.F.S..SOL.T'gC
/OQ ul net /<9O ul Actual
in/min speed
Iniection/QQ ul indicated
Sampling Device GLASS GAS
SYRINGE.
76543
Retention time, iran.
Figure 6. Typical gasoline CVS-exhaust sample.
373
-------
Samole2.8fj.pm CAT Date I2.-I1 -77
Instrument P£ 342.O B Operator J.C. CAIN
Packed with ft A * ««• MA
on /00//10 mesh POfM PA M O
Liq. Phase
Support .
Run ISO @ HEp °C. using fO cc/min.,4r/S%f Carrier - -
@ *^O psiq. MA Rotameter Rda. A/A
... Inlet 200 6cHfATE.DSTUULE.SS *TS£t- Type
Detector: 32 S °CSCD Type (other) ^ 'A
Hyd. ArU psifl A/f( Rotameter Hdg. Mi cc/min
Air At A psia A^ Rotameter Rdg. _AfJ cc/min
( ) Af A Dsia /M Rotameter Rdg. ALA cc/min
Recorder_J__in/min speed / mV.F.S. SOLT&C Type
InjoMiBn/QQ ul indicated /0Q ul "8t 'OO Ul Actual
- Samolino Device &LASS &AS SYR\NGE
: ', '
'. I '• j ' '
i i : : ! • ; '
.-:••-.. ! ;.' i j. . •
r r i , r J
. ... - j . : i
, - , . .1. , .
-; - ; T7 -;;." ' ^i'^ .r '.T'1::'.
L: ;-.-•-...•?-•- .-- , 7.".v~ "-^-
-
• •— •-- j •— -i - ;- - -
-• ; ••; .r!' — T::- ; .;".i"
- -- - — -! i-- •-)-•- i- 1
; --^ - i f -1 I -•)
i i ' f
j ..—!.-.- [ - j -.--j -f
| ... J . i .- ; - -- i
1 --I i - - - i -
----- • t - - - ; ;-
-"- ~t - ;• f. " .r
:. i -•..-[. •:- ..-I-.:-:-.- jr.
"" ; { j ; ^j ' rvs " i
1. .... .f ,. . | . ; j . . |
... . .. : ..j , .. i
i '• ' i . ~.
:-.:• I.---JT: 77: ~^—.-;- -.:| -•]•---+-
i ' i ' i • s
1 t _ i nv
:.;; .. ; •-;;-- ]-; - :-;;;- j- ;;;: j
i .. ! - • _LJ
a
H
T"I:
"?__ ..L-, ..
'" m
Its
...;... 4 ...
..-L.. -
---'- " -\
--_ ~ — 1 - -
" \
'. ' ' u
: - 'W
-- P ;- --°
* 1 J^
-zZ"l.i::|
It .
• i •'
L • ill 1
•z -- --
. 61 - - -
_ (J _^~__
.> • ^^
.:S. L -::-.:-
r ro
- J ',
•__LL
i t-
i- L
;„,..!
.. ,
j .j, _ ^
}••- * •-• — — •
.'. " " I.".. '
i-Vi;:>:; :
:;„ j ...
•i - * - j
I
4 --
! t
: - - - -i — - -
i •;
: :: - j 1: .
.: H. j^ .....
' ^ ! ;
5 ;
_3 i
T 04 j" " ""^-1
Ij.- < ; - -
: . . «. 1 . . : .-.
!~ -
S 1
654 3 21 0
Retention time, min.
Figure 1. Typical Trace for a Standard.
374
-------
The next step is to convert the sample concentration to yg per volume ex-
pressed in cubic meters. The sample acquisition is effected using dry gas
meters that have units expressed in cubic feet. To make this conversion, the
following steps are observed.
Csam _ Cstd, yg/m£ x Asam x D.F. x Abs. vol., m£
(yg/ft3) Astd x Sam. Vol., ft3
Csam _ Cstd, g/m£ x Asam x D.F. x Abs. vol., mJl x 35.31 ft3
(yg/m3) Astd x Sam. Vol., ft3 x 1 m3
The sample volume was measured at the actual conditions of the test and is
corrected to 20°C (68°F) and 760 mm Hg (29.92 Hg)
Csam = Cstd, yg/m& x Asam x p.p. x &bs. voi., m£
(yg/m3) Astd x sam. vol., ft3 x l m3 x B.P.,"Hg
x 35.31 ft3 x 29.92"Hg x sample temp °F + 460°R
528°R
At this point, the concentration of the sample is expressed in yg/m3 at 20°C
(68°F) and 760 mm Hg (29.92" Hg) . This concentration is then transferred
to a master computer sheet for the final calculations to express the total
CN- emission rate in mg/km.
Hewlett-Packard 65 Calculations
In order to insure maximum data turnaround in a minimum time period, a
Hewlett-Packard 65 program was prepared to calculate CN- concentrations from
the raw data. This program is presented in Figure 8 and will be used to
calculate yg CN-/m3 and ppm CN- (as HCN) . The data sheet used in the sample
acquisition is also used in calculating the total CN- and ppm CN- (as HCN)
concentrations.
Sample Calculations
Assume exhaust samples were collected in glass impingers for each
portion of a three-bag 1975 FTP. The raw data for these tests is presented
in Figure 9. Calculations were performed using the HP-65 program and
manual calculations. The results of the first bag calculations are sum-
marized below:
Cold start (505 seconds) Manual Calculations
pg CN-/m3 = Cstd, yg/m£ x Asam x D.F. x Abs. vol., ml x 35.31 ft3
Astd x sam. vol., ft^
, 29.92"Hg x (samp.Temp., °F + 460°Rl
X 1 m3 x B.P., "Hg x 528°R
375
-------
yg CN-/m3 = (5.0) (1500) (1) (25) (35.31) (29.92) (75 + 460) _
(2000) (3.286) (1) (29.52) (528)
1035 yg CN-/m3
ug CM- Y 0.000859 v 27.026 ug HCN
ppm L-n— = -j * x
mj 1 yg CN- 26.018 Ug CN-
(1035) (0.00859) (27.026) n n
ppm CN- = (26.018) = °'9
Cold Start (505 seconds) HP-65 Results
pg CN-/m3 = 1035 yg CN-/m3
ppm CN- =0.92 ppm
(as NCH)
Note: The values used in these calculations are picked from a range of
temperatures, standards, dilution factors, etc. to validate the calcu-
lations and may not be representative of expected raw data in all cases.
These calculations are presented to confirm that manual and HP-65 calcu-
lations give the same results. This was confirmed on six sets of calcu-
lations .
376
-------
HP-65 User Instructions
Title
rammer E- *>bert Fanick
raqe nr
n=to 4/7/78
f Total Cyanide in yg/m3 ERF
booa i
Ift/7/78
.J t
1 01
-L
JL
1
1 ! H
STEP
°1
°2
°3
04
1
2
3
4
5
6
8
9
10
11
12
13
14
INSTRUCTIONS
Switch to On; Switch to Run
Feed Card from Left to Right
Initialize
Set Decimal Place
Input Sample Volume
Input Barometric
Input Sample Temperature
Input Dilution Factor
Input Solution Volume
Input Standard Concentration
Input Sample Area Bubbler SI
Output Bubbler #1 Cone.
Input Standard Area
Input Sample Area Bubbler #2
Output Bubbler #2 Cone.
Output Total Cone.
Output Total ppm HCN
INPUT
DATA/UNITS
^
" Hg
OF
mi
Ug/mA
counts
counts
counts
KEYS
1 II
H
f 11 Rea
Dsp II 2
R/S ||
R/S ||
R/S||
R/S H
R/S II
0 /O II
R/S ||
R/S ||
R/S ||
R/S 11
iVb |i
1 *»\\
V* \\
nvs"H i
I II I
1 CL XJ|
RTN ||
f || Reg
II
II
II
11
II
OUTPUT
DAT A/ UNITS
"R m3
"Hg
°R m3
" 1
m3
1
^
mi
m3
M£
m3
yg
in3 counts
Hf
mj
ug
mj counts
^
mj
H|
m3
ppm HCN
Figure 8. HP-65 program form.
377
-------
HP-65 Program Form
Title.
Total Cyanide in yg/m-3
Page.
.of.
SWITCH TO W/PRGM. PRESS I ( | I PRGM ' TO CLEAR MEMORY.
KEY
ENTRY
LBL
A
2
f
R/S
X
STO 1
R/S
4
,06
0
+
RCL 1
T
R/S
X
R/S
X
R/S
2d*
STO 2
RCL 2
R/S
T
R/S
X
STO 3
R/S
RCL 2
S/S
V
R/S '
X
R/S
RCL 3
+
K/S
0
.
4.P
0
0
9
2
5
X
1
.
0
5(3
CODE
SHOWN
23
11
02
81
84
71
3301
84
04
06
00
61
3401
81
84
71
84
71
84
71
3302
3402
84
81
84
71
3303
84
3402
84
81
84
-71
84
3403
.61
84
00
83
00
00
00
09
02
05
71
01
83
00
03
COMMENTS
Input Sample Volume
Input Barometric
Input Temperature
Input Dilution Factor"
Input Solution Vol.
Input Stand. Cone. wg/mJ
Input Std. Area
Input Samp. Area
Bubbler #1
Output Samp. Cone.
Bubbler 41 yg/m-i
Input Std. Area
Input Samp . Area
Bubbler #2
Output Samp. Cone.
Bubbler #2 ug/mj
r-^,, —
KEY
ENTRY
8
7
X
R/S
RTN
g MOP
g NOP
60
73
80
9'^
•oo
CODE
SHOWN
08
07
71
84
24
3501
3501
COMMENTS
Output ppm HCN
REGISTERS
R, Samp.Vol
X Baro.
(°R m3).
R2 Constant
for Calc
R3Samp.
Conc.ug/i
Bubbler #:
R4
Rs
R6
RT
Ra
Ro
LABELS
A Samp. VoJ
R
e
D
E
n
1
2
3
a.
5
6
7
R
Q
FLAGS
1
.9
T0 RECORD PROGRAM INSERT MAGNETIC CARD WITH SWITCH SET AT WPRGM.
Figure 8 (Cont'd). HP-65 Program Form.
378
-------
TOTAL CYANIDE SAMPLE COLLECTION DATA SHEET
SWRI PROJECT NO, ll-ia--54.-<>7p TEST NO. OOl TEST DATE ! M -*V77. VEH1 CLE !
FUEL! gM-ayr-001 CVS NO. 3 TUNNEL SIZE Bin DRIVER'. But- P MILES! /OOP
SAMPLE COLLECTION BY! BOB F.
GENERAL COMMENTS'.
CHEMICAL ANALYSIS BY! Ja»i-£_CALCULATIONS BY!
NO,
1
2
3
4
5
8
7
8
X
DRI-VI-NG
CYCLE
FTP-1
«
FTP- 2
FTP- 3
SET- 7
HFET
NYCC
BG
X
SAMPLING CONDITIONS
VOLUME
FT3
3.28fc
l.W
1312
3.$OZ
£2/7
O.W*
©
B.P,
"HG
29.52.
30.75"
lW
13.50
ntf
lioo
®
TEMP,
'F
75"
9ii
^
$6
90
81
©
DILU-
TION •
FACTOR
/
5"
10
2
I
I
©
ABSORB
"?EAGEN
VOLUME
ML
2fT
25
SO
£0
z
-------
REFERENCES
Braker, W. and Mossman, A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N.J., 1971, pg. 301
Epstein, J., Anal. Chem., Vol. 19, pg. 272, 1947
Collins, P. F., Sarji, N. M., and Williams, J. F., Tobacco Sci., Vol. 14,
pg. 12, 1971
Collins, P. F., Sarji, N. M., and Williams, J. F., Breitrage zur Tabakfor-
schung, Vol. 7, pg. 73, 1973
Sekerka, I. and Lechner, J. F., Water Res., Vol. 10, pg. 479, 1976
Ito, C., Sakamoto, H., and Kashima, T., Kyoritsu Yakko Daigaku Kenkyu Nempo,
Vol. 18, pg. 29, 1973.
Frant, M. S., Ross Jr., J. W., and Riseman, J. H., Anal. Chem., Vol. 44
pg. 2227, 1972
Riseman, J. H., Am. Lab., Vol. 4, pg. 63, 1972
Fleet, B., and von Storp, H., Anal. Chem., Vol. 43, pg. 1575, 1971
Fleet, B., and von Storp, H., Anal. Lett., Vol. 4, pg. 425, 1971
Conrad, F. J., Talanta, Vol. 18, pg. 952, 1971
Vickroy, D. G. and Gaunt Jr., G. L., Tobacco, Sci., Vol. 174, pg. 50, 1972
Burke, James, "Ion Selective Electrode-Direct Technique," Ford Motor Company
procedure for measuring HCN in dilute automotive exhaust, 1976
Llenado, R. A. and Rechnitz, G. A., Anal. Chem., Vol. 43, pg. 1437, 1971
Nebergal, W. H., Schmidt, F. C., and Holtzclaw Jr., H. F., College Chemistry
with Qualitative Analysis, 5th edition, D. H. Heath and Co., Lexington, Mass.,
1976, pg. 684
Woolmington, K. G., J. Appl. Chem., Vol. 11, pg. 114, 1961
Isbell, R. E., Anal. Chem., Vol. 35, pg. 255, 1963
Myerson, A. L. and Chludzinski Jr., J. J., J. Chromatog. Sci., Vol. 13, pg. 554»
1975
380
-------
Valentour^J. c., Aggarwal, V., and Sunshine, I., Anal. Chern., Vol. 46,
Runge, H., Z. Anal. Chem., Vol. 189, pg. Ill, 1962
381
-------
APPENDIX I
ORGANIC SULFIDE PROCEDURE
382
-------
THE MEASUREMENT OF ORGANIC SULFIDES IN EXHAUST
The measurement of organic sulfides; carbonyl sulfide (COS), methyl
sulfide (dimethylsulfide, (CH3)2S), ethyl sulfide (diethylsulfide,
(C2H5)2S) and methyl disulfide (dimethyldisulfide, (CH^S^) in exhaust is
accomplished by passing the exhaust through Tenax GC traps at -76 °C. The
organic sulfides are removed from the exhaust by the traps at this tempera-
ture. The exhaust sample is collected continuously during the test cycle.
Thr organic sulfides are thermally desorbed from the traps into a gas
chromatograph sampling system and injected into a gas chromatograph equip-
ped with a flame photometric detector for analysis. External organic
sulfide standards generated from permeation tubes are used to quantify the
results. Detection limits are on the order of 0.1 ppb.
LIST OF EQUIPMENT
The analysis for the organic sulfides is performed using a gas chroroa-
tograph equipped with a flame photometric detector. The gas chromato-
graph, control console, sample collection, trap conditioning and trap
preparation are the basic functions in the analysis. The major equipment
required for each function is listed below.
Gas chromatograph and control console
1. Perkin-Elmer Model 3920B gas chromatograph equipped with a lin-
earized flame photometric detector (FPD) and subambient tempera-
ture programmer.
2. Soltec dual channel recorder, Model B-281, 1 mv recorder.
3. Hewlett-Packard Model 3354 GC computer system with remote tele-
type printout.
4. Hewlett-Packard Model 1865A A/D converter.
5. Metronix Dynacalibrator Model 220-R for generation of organic
sulfide standards.
6. Bendix valve oven.
7. Lindberg Furnace/Heavy Duty Model 55035.
8. Seiscor valve - gas sampling configuration.
9- Seiscor valve - backflush configuration.
383
-------
10. ATC timers, Model 325A346A10PX (2 ea.).
11. Analytical column, 6' x 1/8" Teflon, 60/80 Tenax-GC.
12. ASCO solenoid valve, Model 834501 (2 ea.).
13. Brook flowmeter, R-2-15-A w/ss float, 0-150 scale.
14. Metal Bellows MB-158 pump.
15. Female quick-connect, stainless steel.
16. Nupro Model 2M stainless steel regulating valve.
17. Miscellaneous stainless steel, copper and Teflon tubing (1/8"
and 1/6").
18. Miscellaneous stainless steel and brass unions, tees, etc.
19. Bud Classic II control console cabinet.
20. Miscellaneous electrical on-off switches.
Sampling
1. Perma Pure drier, Model PD 625 12S (17").
2. Brooks flowmeter, R-2-15-AA, sapphire float, 0-150 scale.
3. Metal Bellows MB-158 pump.
4. Tenax-GC trap.
5. Sodium bicarbonate trap.
6. Miscellaneous stainless steel and Teflon tubing (1/16", 1/8"
and 1/4").
7. Miniature stainless steel Swagelok female quick-connects.
8. Miniature stainless steel Swagelok male quick-connects. '
9. Miscellaneous stainless steel and brass unions, tees, etc.
10. Stainless steel 7.0 cm flip-top filter.
11. Reeve Angel Type AH 7.0 cm fiber glass filters.
12. Nupro Model 4M stainless steel regulating valve.
13. Dewar flask, 1 quart capacity.
384
-------
Trap Preparation and Conditioning System
1. Lindberg Furnace/Heavy Duty Model 55035.
2. Brooks flowmeter, R-2-15-A, glass float, 0-150 scale.
3. Nupro Model 4M brass regulating valve.
4. Swagelok 3/8" x 1/8" stainless steel reducing unions (2 per trap).
5. Stainless steel fritted discs, 50y, 3/8" OD (2 per trap).
6. Stainless steel tubing, 2" x 3/8" OD (0.028" wall) (1 per trap).
7. Tenax-GC, 60/80 mesh (about 1 gram per trap).
8. Miscellaneous stainless steel and Teflon tubing (1/8" and 1/4").
9. Miscellaneous stainless steel and brass unions, tees, etc.
10. Asbestos gloves, pair.
11. Refillable plexiglass gas drier, 6".
12. Nupro in-line filter, brass, Model 4.
Expendables
1. Zero hydrogen gas (GC).
2. Zero air (GC).
3. Zero helium carrier gas (GC).
4. Nitrogen (Perma Pure drier).
5. Isopropyl alcohol, CH3CHOHCH3.
6. Sodium bicarbonate.
7. Dry ice.
8. zero nitrogen (permeation system).
CONTROL SYSTEM
A control system was developed to systematically control the flow of
the two Seiscor valves. This control is accomplished by ATC electric
timers and ASCO electric solenoid valves. This system employs one Seiscor
valve in a gas sampling valve (CSV) configuration with the second Seiscor
valve in a backflush configuration. These valves are pneumatically
operated and electrically controlled. The electrical schematic for the
385
-------
control of the Geiscor valves usiirj the ATC timers and ASCO electric sole-
noid valves is shown in Figure 1. The flow schematic for vacuum and pres-
sure lines to the Seiscor valve are presented in Figures 2-6. The Seiscor
valves have been found to operate much more dependably if a vacuum assist
is inducted in the valve actuation controls.
SAMPLING SYSTEM
A Tenax GC trap is used to collect exhaust samples for the analysis
of the organic sulfides. A flow schematic of the sample collection system
is shown in Figure 7. The trap collects 99+ percent of the sulfides at
flows up to 130 mVmin. Several views of the sampling system are shown in
Figure 8. The various components of the sampling system and their functions
are listed below.
Item
Component
NaHCO3 trap
Tenax-GC trap
Perma-Pure Drier
Sample Pump
Low temp bath
Flip-top filter
7. Regulating- valve
Description
2" x 3/8" OD x 0.035" wall stainless steel
cartridge packed with 5 percent NaHCC>3 on
45/60 mesh Chromosorb P (this trap removes
any t^S or SC>2 in the exhaust sample) .
8.
Flowmeter
2" x 3/8" OD x 0.035" eall stainless steel
cartridge packed with preconditioned 60/80
mesh Tenax-GC (this trap collects and con-
centrates the organic sulfides) .
Model PD-625 12S Perma-Pure Drier (this dryer
removes the moisture in the exhaust without
jeopardizing the sample integrity) .
Model MB-158 Metal Bellows Vacuum/Compressor
Pump. The sample pump pulls the exhaust sample
through flip- top filter and Perma-Pure Drier
and forces the sample under pressure through
the remainder of the system.
A constant low temperature bath is obtained by
using a CC^-isopropyl alcohol slurry. A bath
temperature of -76 to -78 °C is obtained with
this bath.
A 7.0 cm stainless steel flip-top filter is
included to remove all particulate from the
exhaust sample prior to its entry into the
Perma-Pure Drier.
A Nupro SS-4MG regulating needle valve is used
to control the exhaust flow through the NaHCOs
and Tenax-GC traps.
A Brooks Model 1550 flowmeter with R-2-15-C
386
-------
OJ
OD
I
2
4
12
14
15
in
lu
.~-_J
ATC
Timer
_
'—
f
I
j
i
i
,*••••••• • !!•• III
t _r-t _, ^
1,
1
2
d
5
12
14
15
1R
lu
ATC
Timer
,
~~j
— '
^
J
s
),
*
GSV ^ gH
solenoid gj <3 r<
r^ °J-
I _
• " "' — D
9
i
^ backflus
gj solenoii
?lay
h
j
Aux
Figure 1. Electrical schematic for organic sulfide analysis system.
-------
Analytical Column
(61 x 1/8" TFE, 60/80 Tenax GO
Seiscor Valve
(backflush configuration)
Vent
Seiscor Valve
(Gas Sampling Configuration)
PE 3920B
Gas Chromatograph
with FPD
Carrier Gas
SS miniature
connect
10 ml sample loop
j—I Calibration
gas in
Female
quick-connect
Figure 2. Flow schematic of organic sulfide calibration system
(Step 1 - purge of sample loop of CSV).
388
-------
Analytical Column
(6' x 1/8" TFE, 60/80 Tenax GC)
Seiscor Valve
(backflush configuration)
Vent
PE 3920B
Gas Chronatograph
with FPD
Seiscor Valve
(Gas Sampling Configuration)
1 ? .
•••••••••••••••••••••••"•••^
Pump
SS miniature
connect
Carrier Gas
X
10 ml sair.ple loop
Regulating
Valve
-S
[—j Calibration
gas in
Female
quick-connect
Figure 3. Flow schematic of organic sulfide calibration system
(Step 2 - inject calibration gas into GC system).
389
-------
Analytical Column
(61 X 1/8" TFE, 60/80 Tenax-GC)
Seiscor Valve
(backflush configuration)
Vent
Perkin-Elmer-
3920B
Flame Photometric
Detector
Seiscor Valve
(Gas Sampling Configuration)
tf \
.••••••••••••••fc««»»«»*»*»^
i
Carrier
Gas
\
-CD-
Tenax Trap
** • n • • v w *^r + WVMVBWV
\
SS miniature
quick connect
300°C furnace
Regulating
Valve
Pump
•S
Female
Quick-Connect
Flowmeter
Figure 4. Flow schematic of organic sulfide analysis system
(Step 1 - connect Tenax trap in CSV).
390
-------
Analytical Column
(6* x 1/8" TFE, 60/80 Tenax-GC)
Seiscor Valve
(backflush configuration)
Vent
Perkin-Elmer
3920B
Flame Photometric
Detector
Seiscor Valve
(Gas Sampling Configuration)
300°C furnace
SS miniature
quick connect
Regulating
Valve
Female
Quick-Connect
Flowmeter
Figure 5. Flow schematic of organic sulfide analysis system
(Step 2 - inject Tenax trap contents into GC system).
391
-------
Analytical Column
(6* X 1/8" TFE, 60/80 Tenax-GC)
Seiscor Valve
(backflush configuration)
Perkin-Elmer
3920B
Flame Photometric
Detector
Seiscor Valve
(Gas Sampling Configuration)
Vent
rzL
Carrier
Gas
Tenax Trap
-CD-I i- • '
V 300°C furnace
SS miniature
quick connect
Regulating
Valve
Pump
•ft
-a
Female
Quick-Connect
Flowmeter
Figure 6. Flow schematic of organic sulfide analysis system
(Step 3 - backflush analytical column).
392
-------
Ul
to
U)
Sample
I Probe
Nitrogen In
'lip-Top Filter I
Sample Pump
E
Tunnel
Perina Pure Drier
NaHCC>3 Trap
Nitrogen Out
Vent
Flowmeter
Regulating Valve
Tenax-GC Trap
Low
Temperature
Bath
LT
Figure 7. Organic sulfide sample flow schematic.
-------
Figure 8. Several views of the organic sulfide sampling system.
394
-------
Figure 8 (Cont'd). Several views of the organic sulfide sampling system.
395
-------
(sapphire float) is used to determine the exhaust
sampling rate.
TRAP PREPARATION
The Tenax traps used for the collection of the organic sulfides are
prepared by filling a 2" x 3/8" OD (0.028" wall) stainless steel tube with
approximately 1 gram of 60/80 mesh Tenax-GC. Stainless steel fritted discs,
50y, 3/8" OD are placed at each end of the trap to hold the Tenax in the
trap while allowing a gas flow through the trap. Nut and ferrules, 3/8",
and Swagelok 3/8" x 1/8" stainless steel reducing unions hold the fritted
discs in place and allow the trap to be inserted into the sampling system.
A 1/8" stainless steel cap is placed on each end to prevent moisture and
other unwanted compounds from collecting in the trap. Figure 9 shows a
view of the completed trap and of its components. When a trap is ready to
use 1/8" Swagelok nuts and ferrules are used to connect 1/8" stainless steel
tubing to each end of the trap. At the other end of each piece of the 1/8"
stainless steel tubing a miniature male quick connect is added with 1/8"
Swagelok nuts and ferrules (Figure 10). The trap can how be connected by
the miniature quick connects to the sampling system or the desorbing sys-
tem with ease.
ORGANIC SULFIDE TENAX-GC TRAP CONDITIONG PROCEDURE
The analysis of organic sulfides in dilute automotive exhaust requires
Tenax-GC traps that have been properly conditioned. It is absolutely essen-
tial that each Tenax-GC trap undergo the identical conditioning. This will
insure that there are no residual compounds in the trap from a previous
sample, or in the case of a fresh trap, to remove any residual solvents.
Accurate quantitative data is directly dependent on performing the condi-
tioning procedure in a consistent manner according to the procedure out-
lined in section.
The Tenax-GC traps can be used repeatedly provided they are properly
conditioned. Conditioning of the Tenax-GC traps is accomplished by purging
with zero nitrogen at a selected temperature for a specified time period at
a given flow rate. A system was developed that was capable of conditioning
two Tenax-GC traps simultaneously. A flow schematic of this system is
illustrated in Figure 11. Several views of the Tenax-GC sample condition-
ing system are presented in Figure 12. This system has been shown to
reproducibly condition Tenax-GC traps to a negligible level of organic
sulfides. Traps have been successfulyy re-conditioned for as many as five
samples before replacing the Tenax-GC. It may be possible that the Tenax-
GC traps would last substantially longer, but this has not yet been estab-
lished.
The organic sulfide trap conditioning procedure is listed below in a
step-wise sequence.
1.) Turn on the furnace and adjust the temperature to 325°C
+25°c. The furnace should be allowed to stabilize for
at least 15 minutes before the first trap is conditioned.
396
-------
Figure 9. Tenax-GC trap.
Figure 10. Tenax-GC trap with quick connect.
397
-------
Drier
Zero
N--
Regulating
Valves
Flowmeters
Tenax
Traps
L
••^•••••i
mmtmmm
Furnace @ 325+25°C
Vent
(500 ml/min)
Vent
(500 ml/min)
Figure 11. Flow schematic for conditioning Tenax-GC traps
for organic sulfide analysis (dual system) .
398
-------
Figure 12. Several views of Tenax-GC trap conditioning system.
399
-------
Figure 12 (Cont'd). Several views of Tenax-GC trap conditioning system.
400
-------
2.)
The temperature readout on the furnace should be vertified
at least once a week with a digital thermocouple.
Connect the traps according to the flow schematic. Insert
the traps in the furnace (with no nitrogen) and allow the
traps to equilibriate for 5 minutes.
3.) Turn on the nitrogen and adjust the flow to 500 mJl/min.
4.) After the 60 minute conditioning period, the trap is removed
from the furnace and is allowed to cool to room temperature
with nitrogen flowing through the trap.
5.) After the trap returns to room temperature, turn off the
nitrogen and remove the trap. The system is then ready for
conditioning the next two traps.
NOTE 1: A master log should be maintained on the conditioning of each
trap. It is the responsibility of the individual who is con-
ditioning the traps to keep a record of all traps that have been
conditioned, who conditioned them and the date that they were
conditioned. All traps should be permanently identified to
enable keeping of these records.
NOTE 2: Tenax-GC traps should not be placed in the furnace if the tem-
perature is in excess of 375°C. This is the manufacturer's
maximum recommended temperature and should not be exceeded.
If this is allowed to happen, the chemical and physical pro-
perties of the Tenax-GC may be altered thereby affecting the
trapping characteristics of the Tenax-GC.
ANALYTICAL PROCEDURE
The analysis of the organic sulfides (carbonyl sulfide, methyl sul-
fide, ethyl sulfide, and methyl disulfide) in dilute exhaust is accomp-
lished by collecting the organic sulfides in Tenax traps at -76°C. The
organic sulfides are thermally desorbed from the trap into a gas chromato-
graph sampling system. The organic sulfides are analyzed by injecting the
desorbed sample into a gas chromatograph equipped with a flame photometric
detector. A standard blend containing known amounts of the four organic
sulfides is injected into the gas chromatograph to quantify the results.
Prom the GC analysis of the sample, the analysis of the standard blend,
and the measured volume of exhaust sampled, the concentration of the organ-
ic sulfides in the exhaust can be determined. The analysis flow schematic
for the organic sulfides is shown in Figure 13. A detailed description of
the procedure follows.
The organic sulfides are removed from the exhaust stream by trapping
them in Tenax GC traps. A sample pump pulls dilute exhaust from CVS
through a flip top filter, to remove particulate from the exhaust sample,
and then through a Perma Pure Drier which selectively removes moisture
from the exhaust sample. The moisture must be removed to prevent ice from
401
-------
CVS
Tenax
GC Traps
Sample Analysis
in gas chromatograph
with' FPD
A/D Converter
1
Recorder
Hewlett Packard
3354
Computer System
Figure 13. Organic sulfide analysis flow schematic.
402
-------
forming in the Tenax traps. The ice formation would plug the trap and
prevent the exhaust sample from passing through the trap. After exiting
the Perma Pure Drier the exhaust sample passes through a sodium bicar-
bonate trap which removes any interfering H2S or S02 in the exhaust sample.
The exhaust sample is then pulled through the Tenax GC trap which removes
the organic sulfides from the exhaust. The trap is held at -76°C with a
dry ice-isopropyl alcohol slurry. After the organic sulfides have been
removed by the Tenax trap the exhaust passes through the sample pump, a
regulating valve and a flow meter before exiting the sample system. The
needle valve regulates the flow through the system which is monitored by
the flow meter. A constant flow of 130 m£/min is maintained throughout the
test. The Tenax trap is disconnected from the sampling system at the two
miniature quick-connects and the two male miniature quick-connects are
capped. The trap remains in the -76°C dry ice-isopropyl slurry until it is
desorbed by the gas chromatograph injection system.
The Tenax trap is removed from the dry ice-isopropyl slurry, the
liquid is quickly wiped from the trap, and the caps are removed from the
ends of the trap. The trap is connected into the gas injection system
with the two quick connects, (Figure 4) and the sample is immediately in-
jected and placed into the Lindberg furnace operating at 300°C (Figure 5).
The carrier gas upon injection flows through the loop carrying the con-
tents into the gas chromatograph where the sulfides are separated and
identified by their retention times.
The gas chromatograph system used to analyze the organic sulfide
sample is shown in Figure 14. The system consists of a Perkin Elmer 3920B
GC, an A/D converter and a recorder. The Figure also shows the control
system, the Lindberg furnace, and the Bendix valve oven. The GC is equipped
with a linearized flame phtometric detector which has a high sensitivity to
sulfur containing compounds. The column consists of 6' x 1/8" Teflon tubing
packed with 60/80 mesh Tenax GC. The carrier gas is helium which flows
through the column at 30 mi/m.in. The optimum hydrogen and air flows are
40 m£/niin and 360 mVmin respectively. The column temperature, after in-
jection of the sample, is programmed from 30°C to 140°C at 8° a minute. In
a chromatogram of a standard sample (Figure 15) containing the four sul-
fides, the first peak eluted is carbonyl sulfide, followed by methyl sul-
fide, ethyl sulfide and methyl disulfide. The GC sulfide peaks are re-
corded on a Soltec dual channel recorder (1 mv) and peak areas and
retention times are obtained from the Hewlett Packard GC computer system.
A Metronix Dynacalibrator operating at 40°C and containing permeation tubes
of carbonyl sulfide, methyl sulfide, ethyl sulfide, and methyl disulfide is
used to supply standard concentrations of the organic sulfides. The
permeation rate of each tube is monitored by monthly weighings. Zero nitro-
gen is used to dilute the permeation gases to the concentrations desired.
The 10 m£ sample loop is purged with this permeation gas for 10 minutes
(Figure 2) and the 10 mJl of permeation gas is then injected into the GC
and analyzed. From the standard peak areas, the exhaust sample peak areas
and the volume of exhaust sampled, the concentration of the organic sul-
fides in exhaust can be determined.
CALCULATIONS
403
-------
,!
A
Figure 14. Organic sulfide analytical system.
-------
ixla. '<
-------
This procedure has been developed to provide the user with the concen-
trations of the organix sulfides (carbonyl sulfide, methyl sulfide, ethyl
sulfide, and methyl disulfide) in exhaust. The results will be expressed
in yg/m3 of exhaust and ppm for each of the sulfides. The equations for
determing the concentrations of yg/m3 and ppm are derived in the following
manner.
The first step is to find the volume of exhaust sampled from the flow
rate and the sampling time by the equation:
Vol exp (m£) = F.R.-j. (m£/min) x Ti (sec)/60 sec/min
Vol exp (m£) = volume of gas sample in m£
F.R.,. (m£/min) = flow rate of exhaust sample in m£/min
Ti (sec) = sampling time in minutes
60 sec/min = conversion of sample time in seconds to minutes
[Equation 1]
The next step is to correct the volume of exhaust sampled to a stand-
ard temperature, 68°F, and pressure, 29.92" Hg, by use of the equation
P V P V
exp x exp _ corr x corr
' Texp Tcorr
exp = experimental volume of gas sample in mi
corr = volume of gas sample in mi corrected to 68°F and
29.92" Hg
pexp = experimental barometric pressure
Pcorr = 29.92!1 Hg
Texp = experimental temperature in °F + 460
Tcorr = 68°F + 460 = 528°R
Solving for vcorr gives:
v pexp ("Hg) x vexp (ft3) x 528°R
vcorr = *-—-— *•
Texp («>R) x 29.92" Hg
[Equation 2]
Substituting Vol exp (m£) from equation 1 into equation 2 gives:
Pexp ("Hg) x F.R.X (mJl/min) x Ti (sec) x 528°R
vcorr (m£) =
rexp (°R) x 29.92" Hg x 60 sec/min
406
-------
The next step converts the volume from m£ to cubic meters by use of
the conversion factor 1 cubic meter = 10^ mA.
P
exp ("Hg) x F.R.J (mJl/min) x Ti (sec) x 528°R
vcorr (m3) =
Texp (°R) x 29-92" Hg x 60 sec/min x 106 m£/m3
[Equation 3]
The next step is to find the yg of each of the sulfides in the Tenax
trap. Since the FPD has a linear response in the region of concern, then
the following equation holds:
yg sam = yg std
Asam Astd
Ug sam = yg sample in Tenax trap
Asam = GC peak area of sample in relative units
yg std = yg standard
Astd = GC peak area of standard in relative units
Solving for yg sample gives:
yg std x Asam
ug sam = _
[Equation 4]
The yg of standard for each of the organic sulfides is determined
from the permeation rates of the permeation tubes conta.i.ning each of the
sulfides, the flow rate of the diluting gas, and the volume of gas sampled
(10 mJl) .
P.R. (rag/min) x 10 mJl
y ~
(mA/min) x 1000 ng/yg
[Equation 5]
P.R. (ng/min) = permeation rate of permeation tube at 40° in ng/min
10 m£ = volume of calibration gas injected
.R.jj (m£/min) = flow rate of diluting gas for permeation system
1000 ng/yg = converts ng to yg, one yg equals 1000 ng
Substituting yg std from Equation 5 into Equation •! gives:
P.R. (ng/min) x IQ m& x Asam
yg sam = P.R>II (mji/min) x 1000 ng/yg x Astd
[Equation 6]
407
-------
To obtain yg sample/m3, Equation 6 is divided by Equation 3 to give:
3 P.R. (ng/min) x IQ m£ x ASam x TeXp (OR)
yg sam/m - F-R-II (2^/nd.n) x 1000 ng/yg x Astd x ?exp (»Hg)
29.92" Hg x 60 sec/rain x io6 m£/m3
F.R.-j. (m£/min) x Ti (sec) x 528°R
[Equation 7]
To find the concentration of each sulfide in ppm, the densities of
the sulfides are needed. At 29.92" Hg and 32°F, one mole of gas occupies
22.4 liters. This volume is corrected to 68°F from the equation
V _ Vi
Vx = 22.4
T! = 32°F +460 = 492°R
V = volume at 68°F
T = 68°F + 460 = 528°R
Solving for V gives:
V = i XT = ' * 2 = 24.04£
Since one mole of gas occupies 22.04 at 68°F, the density can be
found in g/£ by dividing the molecular weight in g/mole by 24.04 A/mole
mol. wt. g/mole
= 24.04 £/mole
-The denisty in yg/m£ can be found by converting g to yg and £ to m£
as follows:
den yg/m£ = mo1- wt- g/mole x 1 x 106yg/g _ mol. wt. x IQQQ
24.04 £/mole 1 x !03m£/£ 24.04
[Equation 8]
To obtain the concentration of each sulfide in ppm, the concentration
in yg/m3 is divided by the density in yg/m£
ppm = yg/m3 * y g/m£ = ~
408
-------
Using Equations 7 and 8 gives the ppm concentrations in the form of
the raw data.
P.R. (ng/m£) x IQ ml, x Asam x T6xp (°R) x 29.92"Hg
F.R.ZI (m£/min) x IQOO ng/yg x Astd x Pexp ("Hg)
x 60 sec/min x ip6 m£/m3
F.R.Z (m£/min) x Ti (sec) x 528°R
v 24.04 £/mole
mol. wt. (g/mole) x ]_ooo Vg-H/g-mH
[Equation 9]
At this point, the concentration can be expressed in yg/m3 (Equation 7)
and ppm (Equation 9) at 68°F and 29.92" Hg from the raw data.
Hewlett-Packard Calculations
In order to insure maximum turnaround in a minimum time period a
Hewlett-Packard 65 program was developed to calculate the organic sul-
fide concentrations in yg/m3 and ppm from the raw data. This program is
presented in Figure 16.
Sample Calculation
Assume exhaust samples were collected in Tenax traps for each portion
of a three bag 1975 FTP. Raw data for these tests are presented in Figure
17. Calculations were performed using the HP-65 program and manual
calculations.
Manual calculations for driving cycle FTP-1
/ 3 C0c = P.R. (ng/ml,) x 1Q m£ x ASam x Texp (°R) x 29.92" Hg
yg/m cos - ^-^ ,_„ ,___, x 1000 ng/ug x hstd x pexp (»Hlg)
60 sec/min x io6 m£/m3
X F.R.i (mVmin) x Ti (sec) x 528°R
_ 667.5 ng/m£ x IQ mJl x 30200 x 535°R x 29.92" Hg
580 mVmin x 1000 ng/yg x 18514 x 29.80" Hg
60 sec/min x 1Q6 m£/m3
X 130 mVmin x 504 x 528°R
= 17.5 yg/m3
409
-------
HP-65 Program Form
Organic Sulfides in Exhaust
Page,
SWITCH TO '-V PRGM. PRESS 1' PRGM . TO CLEAR MEMORY.
KEY
ENTRY
LBL
A
3
4
0
0
0
T
R/S
tt X
R/S
X
3T01
R/S
4
6
0
+
RCL1
20 g x y
T
g i/x
R/S
T
STO2
R/S
RCL2
X n
R/S
oO *
R/S
X
R/S
2
4
9
9
f
R/S
40
RCL2
X
R/S
T
R/S
X
R/S
2
5
5C8
CODE
SHOWN
23
11
81
84
71
34
71
3301
84
61
3401
3S07
81
3504
84
81
3302
84
3402
71
84
81
84
71
84
81
84
3402
71
: 84
81
84
71
84
COMMENTS
In Samp Flow, ml/min
In Sampling Time , sec
In Barometer, "Hg
In Samp Temp, °F
In Oil Gas Flow, ml/rain
In Perm Rate COS , ng/mir
In Std Area COS
In Samp Area COS
Out ]Jg/m3 COS
Out ppm COS
In P R (CH3>2S, ng/min
In Std Area (013)28
In Samp Area (CH-?) ?S
Out yg/m3 (CH,),S
KEY
ENTRY
4
T
R/S
RCL2
X
R/S
T
R/S
60 X
R/S
3
7
5
2
T
R/S
RCL2
70 X
R/S
T
R/S
X
R/S
3
9
1
8
60 T
R/S
RTN
90
"CO
CODE
SHOWN
81
84
3402
71
84
81
84
71
34
81
84
3402
71
84
81
84
71
84
31.
84
24
COMMENTS
Out ppm (013) 2S2
In P R (C2Hs)2S, ng/niin
In Std Area (C2Hs)2S
In Samp Area (C2Hs)2.S
Out ug/m3 (C2H5)2S
Out ppm ICyB^yS
In P R (CH3)2S2, ng/min
In Std Area (0*3) 2§2
In Samp Area (CH^)7S7
Out yg/m3 (CH3)2S2
Out ppm (CH3)2S2
REGISTERS
Ri
R,
R3-
R4
R5
Rs
R7
Rs
RQ
I
LABELS
A
B
C
D
F
o
1
2
3
4
5
6
7
8
9
FLAGS
1
2
Figure 16. HP-65 program.
410
-------
HP-65 User Instructions
Title
Programmer
C \ ( ]
k? i i i i ad I
y — JL,
_i i i mi
STEP
°1
02
°3
°4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
INSTRUCTIONS
Switch to on; switch to run
Feed card in from right to left
Initialize
Set decimal place i
Input Sample Flow Rate , ml/min
Input Sampling Time, sec
Input Barometric Pressure , "Kg
Input Sample Temperature, °F
Input Dilution Gas Flow for Perm Stds, ml/min
Input Permeation Rate COS , ng/min
Input Standard Area COS
Input Sample Area COS
Output Cone lag/m3 COS
Output Cone ppm COS
Input Permeation Rate (CH3)2S, ng/min
Input Standard Area (CH3}2S
Input Sample Area (CH3)2S
Output Cone pg/m3 (013 )2S
Output Cone ppm (CH3)2S
Input Permeation Rate (C2H5)2S, ng/min
Input Standard Area (C2H5>2S
Input Sample Area (C2H5)2S
Output Cone yg/m3 (C2H5)2S
Output Cone ppm (C2H5)2S
Input Permeation Rate (CH3)2S2, ng/min
Input Standard Area (013) 2S2
INPUT
DATA/UNITS
KEYS
II
f |[ REG
DSP | 2
1 A 1!
H/S II
1 K/S ||
1 R/s II
R/S
R/S .
R/S_J|
R/S ||
R/S ||
H
R/S 1
H/S II
1 R/S II
1 R/S II
II
R/S ||
[ R/s ||
1 R/s II
' H/S ||
II
R/S |[
H
OUTPUT
DATA/UNITS
Figure 16 (Cont'd). HP-65 program.
411
-------
HP-65 User Instruction;
Title .
Page.
.of
Proorammer _
i i
_______ Date
J—L I I
STEP
17
i
j
i
INSTRUCTIONS
Input Sample Area (CH3J2S2
Output Cone Ug/n>3 (CH3) 2S2
Output Cone ppra (CH3)2S2
INPUT
DATA/UNITS
,
KEYS
! H/S |[ |
R/S ||
II
RTN ||
f II
REG 11
CLX ||
H
L "
II
II
H
H
II
II
H
H
L H
II
II
11
II
II
II
II
II
OUTPUT
DATA/UNITS
•
Figure 16 (Cont'd). HP-65 program.
412
-------
SWRI PROJECT NO.
FUEL: CVS NO.
SAMPLE COLLECTION BY: _
GENERAL COMMENTS:
TEST NO.
TEST DATE:
TUNNEL SIZE:
CHEMICAL ANALYSIS BY:
DRIVER:
_VEHICLE:
MILES:
CALCULATIONS BY:
Test No.
Driving Cycle : FTP-1 ' FTP-2 FTP-3
Sanple Flow Rate, ml/min : 139 i 110 ', 150
Sampling Time, sec i 504
B. P., "Hg
Temp, °F
Dilution Gas Flow, ml/min
Permeation Rate COS, ng/min
Standard Area COS
Sample Area COS
Sample Cone. COS, yg/m3
Sample Cone COS, ppm
Permeation Rate (CH3)2S ng/min
Standard Area (013) 2S
Sample Area (cf^JjS
Sample Cone (013)28, Ug/m3
Sample Cone {CH3>2S, ppm
Permeation Rate (C2H5>2S, ng/min
Standard Area (C2H5)2S
Sample Area (C2H5)2S
Sample Cone (C2Hs)2S, pg/m3
Sample Cone (C2H5)2S, ppm
Permeation Rate (CH3)2S2, ng/min
Standard Area (CH3)2S2
Sample Area (CH3)2S2
Sample Cone (CH3)2S2, Ug/m3
Sample Cone (CH3)2S2, ppm
867 ' 505
29.80 30.03 29.02
75 80 ' 96
580
667.5
18,514
30,200
17.5
500 i 600
667.5 'i 667.5
20,112 i 21,238
40,100 ' 33,162
17.1 ! 14.9
0.00700 0.00683 j 0.00598
1061
41,006
50,716
21.1
0.00816
445
6971
31,649
32.5
0.00865
133.5
2315
2011
1.86
0.000475
1061 | 1061
38,100 i 35,100
40,381 ! 49,162
14.4 I 21.3
0.00558 ! 0.00824
445 i 445
7017 ', 6844
34,650 1 32,111
28.2 ! 29.9
0.00751 i 0.00798
133.5 ! 133.5
2210 I 2763
3120 ! 6372
2.42 1 4.41
0.000617 i 0.00113
SET- 7 ' HFET
90 120
1397 765
29.25 ! 29.95
85 ! 83
650
450
667.5 667.5
16,542 15,962
20,122 '. 16,269
6.29
0.00252
10.2
0.00406
1061 1 1061
41,000
38,142
7.65
45,610
54,753
19.0
0.00296 0.00736
445 445
6015
7022
4.03
0.00107
133.5
1651
1561
0.978
0.00250
7113
7914
7.39
0.00197
133.5
1814
1418
1.56
0.000397
HYCC
170
1200
29.50
89
575
667.5
23,146
32,641
5.08
0.00203
1061
35,122
41,611
6.78
0.00262
445
7099
17,416
5.89
0.00157
133.5
2917
2372
0.586
M). 000 149 |
Figure 17. Organic sulfide sample collection sheet.
413
-------
ppm COS = yg/m3 v density yg/m£
. mol. wt. (COS) x loop
density yg/m = - 247()4£ -
mol. wt. COS = 60.08 g/mole
60.08 g/mole x 1000 „._._
- - 24!o4 ° 24"
ppm =17.5 yg/m3 -f 2499 yg/m£ =7.00 x io~3 m£/m3 = 7.00 x 10~3 ppm
The calculations for methyl sulfide, ethyl sulfide, and methyl disulfide
are carried out in the same manner by substituting in the appropriate
permeation rates, standard areas, sample areas, and molecular weights in-
to the above formulas. These calculation give the following concentrations:
(CH3)2S, 21.1 yg/m3 and 0.00816 ppm; (C2H5)2S, 32.5 yg/m3 and 0.00865 ppm;
and (CH3)2S2, 1.86 yg/m and 0.000475 ppm.
NOTE: The values used in these calculations are picked from a range of
temperatures, pressures, etc. to validate the calculations and may not be
representative of expected raw data. These calculations are presented to
confirm that manual and HP-65 calculations give the same results. This was
confirmed for six sets of calculations.
414
-------
REFERENCES
Spencer, C. F. , Baumann, F., and Johnson, J. p., Anal. Chem. , Vol. 30, No
9, pg. 1473, 1958.
Sumner, S., Karrman, K. J., and Sunden, V., Microchim Acta, pg. 1144, 1956.
Liberti, A. and Cartoni, G. P., Chim. e ind., Vol. 39, pg. 821, 1957.
Ryce, S. A. and Bryce, W. A., Anal. Chem. , Vol. 29, pg. 925, 1957.
Dcsty, D. H. and Whyman, B. H. P., Anal. Chem., Vol. 29, pg. 230, 1957.
Coleman, H. J., Thompson, C. C. and Rail, H. T., Anal. Chem., Vol. 30, pg.
1592, 1958.
Desty, D. H. and Harbourn, C. G. A., Div. of Analytical and Petroleum
Chemistry, Symposium on Advances in Gas Chromatography, 132nd meeting, ACS,
N. Y., N. Y., September 1957.
Atnberg, C. H., Can. J. Chem., Vol. 36, pg. 590, 1958,
Sullivan, J. H., Walsh, J. T., and Meritt Jr., C., Anal. Chem., Vol. 31,
pg. 1827, 1959.
Dal Nogare, S. and Bennett, C. E., Anal. Chem., Vol. 30, pg. 1157, 1958.
Harrison, G. F. , Knight, P., Kelly, R. P., and Heath, M. ". , Second It'l.
Symposium on Gas Chromatography, Amsterdam, May 1958.
Adams, D. F. and Koppe, R. K., Tappi, Vol. 42, pg. 601, !• 559.
Schols, J. A., Anal. Chem. Vol. 33, pg. 359, 1961.
Hall, H. L., Anal. Chem., Vol. 34, pg. 61, 1962.
Adams, D. F., Koppe, R. K. , and Jungroth, D. M., Tappi, Vol. 43, pg. 602,
1960.
Feldstein, M., Balestrieri, S., and Levaggi, D. A., J. Air Poll. Cont.
Assn., Vol. 15, pg. 215, 1965.
Fredericks, E. M. and Harlow, G. A., Anal. Chem., Vol. 36, pg. 263, 1964.
Mizany, A. I., J. Chromatog. Scie., Vol. B, pg. 151, 1970.
415
-------
Kummer, W. A., Pitts Jr., J. N., and Steer, R. P., Environ. Sci. Tech.,
Vol. 5, pg. 1045, 1971.
Rasmussen, R. A., Am. Lab, Vol. 4, pg. 55, 1972.
Brody, S. S. and Chaney, J. E., J. Gas Chromatog., Vol. 4, pg. 42, 1966.
Bruner, F., Liberti, A., Possanzini, M., and Allegrini, I., Anal. Chem.,
Vol. 44, pg. 2070, 1972.
Elliott, L. F. and Travis, T. A., Soil Sci. Soc. Amer. Proc., Vol. 37, pg.
700, 1973.
Francis, A. J., Adamson, J., Duxbury, J. M., and Alexander, M., Bull. Ecol.
Res. Comm., (Stockholm), Vol. 17, pg. 485, 1973.
Lewis, J. A. and Papavizas, G. C., Soil Biol. Biochem., Vol. 2, pg. 239,
1970.
Lovelock, J. E., Maggs, R. J., and Rasmussen, R. A., Nature, London, Vol.
237, pg. 452, 1972,"
Ronkainen, P., Denslow, J., and Leppanen, O., J. Chromatog. Sci., Vol. 11,
pg. 384, 1973.
416
-------
APPENDIX J
NICKEL CARBONYL PROCEDURE
417
-------
THE MEASUREMENT OF NICKEL CARBONYL IN EXHAUST
The concentration of nickel carbonyl, Ni(CO)4, in exhaust can be de-
termined by the use of a chemiluminescent NOX analyzer modified with a
No. 54 wratten filter. Standard CVS bag samples are analyzed for nickel
carbonyl xising standard calibration blends to quantify the results. The
instrument has a linear response from 0.1 ppm down to its detection limits
(less than 1 ppb). The linearity at higher concentrations was not deter-
mined. The total system schematic for the analysis of nickel carbonyl is
shown in Figure 1.
LIST OF EQUIPMENT
The equipment for the analysis of nickel carbonyl is divided into
two parts. The first part involves the analysis of the exhaust sample and
the second, the nickel carbonyl blend dilution system. Manufacturer, stock
number, and any pertinent descriptive information are listed.
Analysis
1. Chemiluminescent analyzer, Thermo Electron Corporation Model 12A,
modified by replacing the factory installed filter with a Kodak
No. 54 Wratten filter.
2. Quartz tube furnace, Lindberg Sola Basic.
3. Quartz tubing, 5/16" OD x 42".
4. Sample pump, Thomas, Model No. 106 CA 18 3.
5. Recorder, Texas Instruments, 10 mV.
6. Modified 25 mm A-H microanalysis filter holder, Millipore,
Catalog No. XX50 025 00.
7. Fluoropore 25 mm filters, Millipore, Catalog No. FHLP 025 00,
0.5 micron pore size.
8. Female quick connects, stainless steel.
9. Miscellaneous stainless steel, brass, and Teflon nuts, ferrules,
unions, tees, connectors, etc.
10. Miscellaneous Teflon tubing (1/4", 5/16", and 3/8").
11. Flowmeter, Brooks Instrument Division, R-2-15-A, with glass float,
418
-------
Vehicle
CVS
CO
Tedlar Bags
Quartz
Furnace
Chemiluminescent
Detector
with #54 Wratten
Filter
02
Recorder
Figure 1. Nickel carbonyl. analysis flow schematic
419
-------
graduated 0-150.
12. Concentrated nitric acid.
13. Carbon monoxide, Airco, Grade 2.3.
i p
14. Oxygen, Liquid Carbonic, industrial grade.
Dilution System
1. Zero compressed air, Liquid Carbonic.
2. Dry gas meter, American Singer Corporation, Type AL-120,
60 cfm capacity.
3. Flowmeter, Brooks Instrument Division, R-16-16-A, with sapphire
float, graduated 0-100.
4. Stainless steel tubing, 0.01" ID, for capillary restrictor.
5. Line regulator, Matheson, Model No. 40-H.
6. Glass mixing bulb, modified from a 1000 m& bubble soap meter,
with Teflon mixing baffles.
7. Teflon tubing, United States Plastic Corporation, 1/4" OD x
1/8" ID,
8. Miscellaneous stainless steel, brass, and Teflon nuts, ferrules,
unions, valves, etc.
9. Copper tubing, 1/8" OD.
ANALYTICAL PROCEDURE
The analysis of nickel carbonyl in exhaust is accomplished by col-
lecting the dilute exhaust in bag samples. The exhaust sample in pumped
out of the bag through a filter to remove any nickel particulate which
might be present in the exhaust sample. The exhaust sample is then mixed
with an equal volume of^'100 percent carbon monoxide which has previously
passed through a quartz furnace. The mixture of exhaust and carbon mon-
oxide enters the chemiluminescent analyzer where the nickel carbonyl re-
acts with ozone and carbon monoxide to produce a light emitting species.
The interfering light emitted by nitric oxide (NO) and olefins is removed
by the use of a Kodak No. 54 wratten filter. The response of the detector
to the emitted light is then compared with the response to a diluted stand-
ard nickel carbonyl blend. This comparison allows the concentration of
nickel carbonyl in exhaust to be determined. Figure 2 shows a schematic
of the analysis system along with a system for diluting nickel carbonyl
blends to less than 1 ppb. A detailed description of the procedure follows.
For the analysis of nickel carbonyl, dilute exhaust is collected from
420
-------
Quick
Connect
Three-Way
Valve
Vent to
Atmosphere
to
Acid
Scrubber
Quartz
Furnace
1
°2
TECO Model 12A
W/Kodak #54
Wratten Filter
r— i
-MI
Oui
Conn
t
Glass Dilution Tube
Teflon Mixing
Baffles
Three-Way
Valve
Restrictor—>'•
Quarts
Furnace
Flowmeters
Needle Valves
Under Hood
Filter
Ni(CO)4
Calibration
Blend
Flowmeter
Pump
Dry
Gas
Meter
L88%
Zero
Air
Figure 2. Nickel carbonyl analysis and dilution system flow schematic.
-------
the CVS in Tedlar bags. The,?e Lag samples must be analyzed as quickly as
possible. Nickel carbonyl has a half life of 40-70 minutes (depending on
the level of carbon monoxide in the bag) in dilute exhaust. If the bags
sit for any prolonged length of time, there would be a significant loss
of nickel carbonyl in the bag. The exhaust sample is pumped out of the bag
with a sample pump and through a 0.5 micron pore size Fluoropore filter to
remove any nickel particulate which would interfere with the analysis. The
sample passes through the pump and is forced under pressure through a needle
valve (to regulate flow) and a flowmeter to monitor the sample flow. At
this point, the sample is mixed with an equal volume of carbon monoxide.
Before mixing, the carbon monoxide is passed through a quartz furnace
(operating at 480°C) to remove any nickel carbonyl which might be present
in the carbon monoxide. The ability of the furnace to destroy the nickel
carbonyl was demonstrated by passing a nickel carbonyl-carbon monoxide
sample through the furnace. When this was done, the signal due to the
nickel carbonyl was completely destroyed. The flow rate of the carbon mon-
oxide is also regulated with a needle valve and monitored with a flowmeter.
Under normal operating conditions, the flows for both the carbon monoxide
and the exhaust sample are regulated at 165 mJl/min before mixing. The
carbon monoxide and sample enter the chemiluminescent analyzer through the
sample line. The sample mixture is split in the analyzer with 95 percent of
the mixture exiting the analyzer through the bypass line. This exiting mix-
ture passes through a quartz furnace (480°C) and is bubbled through a nitric
acid scrubber to destroy the nickel carbonyl before it is vented to the
atmosphere. The remaining 5 percent of the sample-carbon monoxide mixture
enters the reaction chamber where nickel carbonyl reacts with ozone (gene-
rated by the analyzer) and carbon monoxide to give a light emitting species.
The light emitting species has been postulated to be formed by the following
series of reactions.
Ni (CO) 4 + 03 -»• NiO + other products
NiO + CO •* Ni + CO2
Ni + O3 •* NiO* + 02
NiO* -»• NiO + hv
The higher sensitivity of the instrument (<1 ppb) is probably due to
the ability of each nickel molecule to undergo numerous light emitting
steps before exiting the chamber. A No. 54 wratten filter is used to pre-
vent the light emitted by nitric oxide, NO, and olefins from entering the
photomultipler tube. These compounds are present in exhaust and react with
ozone to give light emitting species. Without removing the signal•due to
these species, the analysis of nickel carbonyl would be very difficult. An
80 ppm nitric oxide sample gives a signal two to three times the noise level
of the instrument and a mixture of olefins (34 ppm total) gives no signal
at all when using the No. 54 wratten filter. The instrument gives a linear
response from /X).l ppm nickel carbonyl down to its detection limits. The
linearity at higher concentration was not determined.
The results are quantified by using diluted nickel carbonyl blends.
Aluminum cylinders containing nickel carbonyl in carbon monoxide are stored
outside in a metal cage (Figure 3) and the blend is piped into the building
with copper tubing. The cylinders are named by wet chemistry. To dilute
422
-------
Figure 3. Storage of nickel carbonyl calibration blends.
423
-------
the concentration of nickel carbonyl in the blend down to the levels of
interest, the blend passes through stainless steel capillary tubing and
into a glass mixing chamber (Figure 4) where it is mixed with a large
volume of zero air. The flow of the blend through the capillary, 3-50
mj^/min, is regulated by varying the pressure on the capillary with a line
regulator. The volume of zero air that enters the mixing chamber, 0.5 to
8 liters a minute, is monitored with both a dry gas meter and a flowmeter.
A portion of this diluted blend is pumped out of the chamber and through a
0.5 micron pore size Fluoropore filter. The diluted blend then passes
through the sample pump, a needle valve, and a flowmeter before being mixed
with an equal volume of carbon monoxide. The diluted blend now enters the
detector and is analyzed as was the exhaust sample. The portion of the
diluted blend which was not pumped out of the chamber flows through a
quartz furnace (operating at 480°C) and through a nitric acid scrubber
before being vented into the room air. By comparing the signal from this
diluted blend with that of the exhaust sample, the concentration of nickel
carbonyl in exhaust can be determined. Figure 5 shows the detector, sam-
pling system, and the dilution system.
CALCULATIONS
The concentration of nickel carbonyl can be determined by comparing
the peak height of the sample with the peak height of the calibration blend.
Since the detector is linear in the region of interest, the concentration
of the sample can be found by use of the following equation:
H ,, H
std _ sam
C ~ C
std sam
where H = the peak height of nickel carbonyl in the diluted calibration
std . , ,
blend
C = the concentration of nickel carbonyl in the diluted calibration
SuQ . _ _
blend
H = the peak height of the nickel carbonyl in the sample
ScUn
C - the unknown concentration of nickel carbonyl in the exhaust
sam • .. -,
sample
Solving for C gives:
sam
TT f-l
_ sam * std
sam H .,
std
Example
A bag sample was taken from the exhaust stream of a vehicle during a
driving cycle. The peak height due to the diluted calibration blend of
nickel carbonyl was 32 units and the blend had a concentration of 0.3 ppb.
The peak height for nickel carbonyl in the sample was 19 units. The con-
centration of nickel carbonyl in the exhaust is calculate as follows:
424
-------
Figure 4. Glass dilution chamber.
Figure 5. Chemiluminescent detector, sampling and dilution system.
425
-------
H C
sam x std
TJ
Hstd
_ 19 x Q.3 ppb
Csam - 32
Csam = 0.18 ppb
Note: The sample calculation is presented as an example only and is not
necessarily representative of expected values in exhaust.
426
-------
REFERENCES
Stecher, P. G., (Ed.), The Merck Index, An Encyclopedia of Chemicals and
Drugs, 8th Edition, Merck and Company, Inc., Rahway, New Jersey, 1968,
pg. 727.
Braker, W. and Mossman, A. L. , Mathespn Gas Data Book, 5th Edition, East
Rutherford, New Jersey, 1971, pg. 401
"Nickel Carbonyl", Amer. Ind. Hyg. Assoc. J., 1968, pg. 304.
Cotton, F. A. and Wilkinson, G., Basic Inorganic Chemistry, John Wiley and
Sons, inc., New York, 1976, pg. 473.
Spice, J. E., Stavely, L. A. K., and Harrow, G. A., J. Chem. Soc., 1955,
pg. 100
McDowell, R. S., Amer. Ind. Hyg. Assoc. J., Vol. 32, 1971, pg. 621.
Densham, A. B., Beale, P. A. A., and Palmer, R., Journal Appl. Chem.,
Vol. 13, 1963, pg 576.
McCarley, J. E., Saltzman, R. S. , and Osborn, R. H. , Anal. Chem., Vol.
28, 1956, pg. 880.
Ball, K. E., Bossart, C. J., and Saltzman, R. S., "Detection of Trace
Amounts of Nickel Carbonyl' and Tetraethyl Lead in Air", Pittsburgh Confer-
ence on Analytical Chemistry and Applied Spectroscopy, February, 1960.
Kincaid, J. F., Stanley, E. L., Beckworth, C. H. , and Sunderman, F. W.,
Amer. J. Clin. Path., Vol. 26, 1956, pg. 207.
Hunold, G. A. and Pietrulla, W., Arbeitsschutz, Vol. 8, 1961, pg. 193.
Pitet, M. G., Arch. Mai. Prof., Vol. 21, 1960, pg. 674.
Vol'berg, N. S., U.S.S.R., Patent No. 135, 684, February 1961.
Belyakov, A. A., Zavodsk Lab., Vol. 26, 1960, pg. 158.
Vol'berg, N. S. and Gerskhovich, E. E., Hyg. Sanit., Vol. 33, 1968,
pg. 226.
Brief, R. S., Venable, F. S., and Ajemian, R. S., Ind. Hyg. J., Vol. 26,
1965, pg. 72.
427
-------
Sunderman Jr., F. W. , Roszel, itf. 0., and Clark, R. J., A"-b. Envix-on.
Health, Vol. 16, 1968, pg. 836.
Bird, C. W., Chem. Rev., Vol. 62, 1962, pg. 283.
Private communication with Donald H. Stdeman, Stedman, D. H. and Tammaro,
D. A.; Analytical Letters; Vol. 9 (1); 1976; pg. 81.
428
-------
APPENDIX K
AMMONIA PROCEDURE
429
-------
THE MEASUREMENT OF AMMONIA IN DILUTE EXHAUST
Ammonia in automotive exhaust can be measured in the protonated form.
NH.+ , after collection in dilute H2SO4- The acidification is carried out
in a glass irnpinger maintained at ice bath temperatures (0-5°C) . A sample
from the impinger is then analyzed for ammonia in the Ion Chromatograph and
the concentration in the exhaust is calculated by comparison to a standard.
EQUIPMENT
The equipment section lists the equipment used in the ammonia proce-
dure. It is divided into four sections corresponding to each major division
in the procedure: Sampling, Analysis, Water Filtration and Sample Prepara-
tion. For convenience the item, manufacturer, model number and any addi-
tional pertinent information are included.
Sampling
1. Glass impingers, Ace Glass Products, Catalog #7530-11, plain
tapered tip stoppers with 18/7 arm joints and 29/24 bottle joints.
2. Flowmeters, Brooks Instrument Division, Model 1555, R-2-15-C,
sapphire ball, 0-5 lit/min range, graduated 0-15.
3. Dry gas meter, American Singer Corporation, Type AL-120, 60
CFH capacity.
4. Digital readout for dry gas meter.
5. Sample pump, Thomas, Model #106 CA18, 4 lit/min'free flow
capacity.
6. Drying tube, Analabs Inc., Catalog #HGC-146, 6" long, 1/4"
brass fittings.
7. Teflon tubing, United States Plastic Corporation, 1/4" OD x 1/8"
ID and 5/16" OD x 3/16" ID.
8. Teflon solenoid valve, The Fluorocarbon Company, Model #DV2-144N
CAl.
9. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps
and connectors, etc.
10. Miscellaneous electrical switches, lights, wiring, etc.
430
-------
11. Regulating valve, Nupro 4MG, stainless steel.
12. Iron/constantan Type J, Thermo Sensors Corporation, single
thermocouple, 1/4" OD stainless steel metal sheath.
13. Six channel digital thermometer, Analog Devices, Model #.2036/J/1.
14. 30 m£ polypropylene sample storage bottles, Nalgene Labware,
Catalog #2006-0001.•
Analysis
1. Dionex Model 10 Ion Chromatograph.
2. Multivoltage recorder, Texas Instruments, Model #PS02W6A.
3. Polyethylene cubitainers, Cole Farmer Instrument Company,
Catalog #6100-20, 1 gallon.
Water Filtration
1. Filtration apparatus, Millipore, Model #XX 1504700.
2. Filters, Millipore, Model #GSWP04700, 0.22 micron pore size.
Sample Preparation
1. 3 cc disposable syringes, Becton-Dickson, Model #5585.
2. 15 m& disposable polypropylene cups, Cole-Farmer Instrument
Company, Catalog #6006-10.
3. Class A, 1 m& volumetric pipets.
4. Class A, 2 m£ volumetric pipets.
5. Class A, 3 m£ volumetric pipets.
6. Class A, 4 m£ volumetric pipets.
7. Class A, 5 mJl volumetric pipets.
8. Class A, 10 m£ volumetric pipets.
9. Class A, 20 m£ volumetric pipets.
10. Class A, 25 mJl volumetric pipets.
11. Class A, 50 mJl voluemtric pipets.
431
-------
12. Class A, 100 m& volumetric pipets.
13. Class A, 100 m£ volumetric flasks.
14. Class A, 1000 mJ6 volumetric flasks.
15. Class A, 2000 m£ volumetric flasks.
16. Mohr pipet, 1 m£ graduated 1/10.
LIST OF REAGENTS
The reagents used in the analysis of ammonia are presented in this
section. In addition to the function of each reagent, the purity, manu-
facturer and catalog number are also listed.
1. Water-deionized and filtered through a 0.22 micron filter.
2. Standard
Ammonium sulfate, (NH4)2SC>4, formula weight = 132.146,
ACS analytical reagent grade, granular, J. T. Baker
Chemical Co. #0792.
3. Absorbant
Sulfuric acid, H^SO^, formula weight = 98.08, ACS
analytic::! reagent grade, Mallinckrodt #2876.
4. Eluent
Nitric acid, HNCS, formula weight = 63.01, ACS analytical
reagent grade (Ulrrex),J.T. Baker, #1-4801.
5. Regenerant
Sodium hydroxide, NaOH, formula weight = 40.00, ACS
analytical reagent grade, pellets, Mallinckrodt #7708.
PREPARATION OF REAGENTS.
The water used ir, making all solutions and dilutions is prepared by
filtering deionized water through a 0.22 micron filter. After filtration
the water is stored in polyethylene bottles.
Standard (NH4)2SO4
The stock solution is prepared by diluting 0.3660g of (NH^)2804 to
1000 mi with water. This yields a solution with a concentration of 100
yg HN+4. Less concentrated standards are made up by diluting portions of
of the stock solution to 100 m& with water using volumetric glassware.
Absorbing Solution (0.01N 112804)
The absorbant is prepared by diluting 20.0 mil of the certified l.OOON
sulfuric acid to 2000 m£ with water.
432
-------
lluent (0.0075N HNO3*
A IN HN03 stock solution is prepared by diluting 62.5 m£ of concen-
trated nitric acid to 1000 m£ with water. The eluent is prepared by fur-
ther diluting 15 ml of the stock solution to 2000 m£ with water.
Reaenerant (Q.5N NaOH) *
40.00g of NaOH is dissolved in water in a 2 liter volumetric flask and
brought up to volume.
* 4 liters of each of these solutions are prepared and stored in labeled
polyethylene cubitainers.
SAMPLING SYSTEM
As seen in the schematic of the ammonia sampling system (Figure 1)
dilute automotive exhaust is diverted from the CVS to the sampling probe.
An on-off solenoid valve controls the gas flow to the. system. If
the valve is open a Thomas sample pump equipped with a molecular sieve-
silica gel dryer pulls the exhuast at 2 lit/min through two bubblers con-
taining 25 nd of 0.01N H2SO4- The reaction converting NH3 to NH/1" proceeds
in the 2 tapered tip bubblers at ice bath temperatures (0-5 °C). After acid-
ification the gas flow, which proceeds through the Singer dry gas meter, is
controlled by a Nupro regulating valve and is measured by a Brooks Flowmeter.
Digital readouts for the ice bath temperature ( °F) and the dry gas meter
volume (ft-^J are provided. Figure 2 shows several views of the ammonia
sampling system.
PROCEDURE
Ammonia in exhaust is collected in two bubblers connected in series each
containing 25 m& of 0.01N H2SO4. These two bubblers trap 99+ percent of
the ammonia. After the acidification of ammonia which takes place at ice
bath temperatures (0-5°C) , the samples are stored in polypropylene bottles.
The samples are then ready for NH4+ analysis on the ion chromatograph (pic-
tured in Figure 3) . Approximately 2 m£ of sample is loaded into the 0.1 mJl
sample loop and injected into the eluent stream. Separation of ions occurs
in the separator (analytical) column and the background conductance of the
eluent (0.0075N HNO3) is neutralized in the suppressor column. The 9 x 250
rom glass cation suppressor column is packed with AG 1- X10, a strong base
ion exchange resin in the hydroxide form. A patented resin containing a
sulfonic acid cation exchanger is packed into the 6 x 250 mm glass separator
column and the 3 x 150 mm precolumn. After the cations are separated in the
analytical column they pass into the suppressor column in the diulte nitric
acid eluent. The hydroxide form of the suppressor resing neutralizes the
acid and then converts the cations to their hydroxides.
HN03 + Resin - OH -^ Resin - N03 + H20
Cation + NO3~ + Resin. - OHf^ Resin - N03 + Cation+~OH
433
-------
Gas Temperature
Digital Readout
U)
Sample Regulating
Pump Valve
On-Off Solenoid Valve
Flowmeter
Dilute
Exhaust
Ice Bath
Temperature
Digital
Readout
Dry
Gas
Meter
Gas Volume
Digital Readou
Figure 1. NH^ sample collection flow schematic.
-------
Front View
"Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. Ammonia sampling system.
435
-------
Solenoid
Impinger
Ice Bath
Close-up of Impingers (Side View)
Pump
Rear View
Figure 2 (Cont'd). Ammonia sampling system.
436
-------
Separator Suppressor
Column Column
J
'
i
Recorder
Ion Chromatograph
Polyethylene Storage Bottle
Figure 3. NH3 Ion Chromatograph.
-------
The conductivity cell produces a signal for the species of interest,
NH^OH, but doesn't "see" the neutralized eluent, deionized water. Con-
ductance is interpreted as a recorder trace ( chroma togr am) or as peak area
by the Hewlett-Packard 3354 computer system. Figure 4 depicts the analysis
flow schematic of the entire ammonia procedure. Two chromatograms produced
by the analysis of a sample and a standard are shown if Figures 5 and 6.
After the 12-20 minute analysis the collection conditions and areas of sam-
ple and standard are used to compute NH3 concentration from the Hewlett
Packard 65 program. This is then recorded on a data sheet.
CALCULATIONS
The purpose of this procedure is to determine the concentration of
ammonia in automotive exhaust. To do this ammonia first needs to be con-
verted to the protonated form, NH4+, which can then be directly measured on
the ion chroma tograph . The calculations involve correcting the measured
concentration of NH^+ to NH3 at a desired temperature and pressure. These
calculations are carried out in a minimum amount of time by using a Hewlett-
Packard 65 calculator program. A copy of the program is shown in Figure 7.
Information from the data sheet (Figure 8) is entered into the calculator
and the ammonia concentration in .. 3. an(j ppm JJH is computed. For illu-
in
stration, two examples using information from the data sheet will be in-
cluded at the end of this section.
The first step in the derivation of the equation for the calculation of
ammonia concentration involves comparison of the sample to a standard of
known concentration. The standard with peak size closest to that of the
sample is the one chosen.
PAst _ PAsa _ PAsa x Cst
Cst Csa sa PAst
where PAst = peak area of the standard
PAga = peak area of the sample
yg NH4+
Cst = concentration of the standard ( - 5 - )
yg NH4+
Csa = concentration of the sample ( - p - )
This equation gives the concentration of NH."1" in the sample. To convert to
lag NH3 4 j.
to — _£ — the ratio of the formula weight of NH3 to NH^ needs to be
multiplied by Csa.
yg NH3
ygNH4+ 17.03C— ^ ygNH3
~— = 0.944 Csa, --
18.04(
y mole
438
-------
CVS
Glass
Impinger
Unused Sample
Saved As
Needed
Sample Analysis
in Ion Chromotograph
with Conductivity Cell
A/D
Converter
Recorder
Hewlett-Packard
3354
Computer System
Figure 4. NH3 analysis flow schematic.
439
-------
•r-j./-;3 Date S-z-78
Attenuation 3ymho
Eluent 0.0012 N H2SC>4 Sample Loop Size 0.1 ml
Elu-:nt Flowrate 167 ml/hr Chartspeed 12 in/hr
Analytical Precolumn 3X150 mm glass
^ Analytical Column 6X250 mm glass
Column Packing Patented resin composed of
a sulfonic acid cation exchanger
Suppressor Column 9X250 mm glass
Column Packing AG 1X-10
Figure 5. Sample chromatogram.
440
-------
Sample 1.50
Attenuation Syiriho
Eluent 0.0012 N H2S04 Sample Loop Size 0.1 ml
Eluent Flowrate 167 ml/hr Chartspeed 12 in/hr
Analytical Precolumn 3X150 mm glass
Analytical Column 6X250 mm glass
Column Packing ^Patented resin comr>sgH of
a sulfonic acid cation
Suppressor Column 9X250 mm glass
Packing AG lX-10
Figure 6. Standard Chromatogram.
441
-------
HP-65 User Instructions
Ammonia In Exhaust iLinear;
rammer Mary Ann Parr.ess
Psgp of
nate 12/15/77
/
ki i i i i
j
_jg
A
yi
1
i i i i id
STEP
°1
°2
°3
04
1
2
3
4
5
6
7
8
9
INSTRUCTIONS
Switch to on; Switch to run
Feed Card in from right to left
Initialize
Set Decimal Place
Input-Sample Volume , f t^
Input-Barometric Pressure, "Hg
Input-Sample Temperature, °F
Input-Dilution Factor
Input-Absorbing Reagent Volume, ml
Input-Standard Concentration, ///-, V'M .-t A>1
!
Input-Standard Area
Input-Sample Area
Output-Sample Concentration, yg.-lU'/m3
Output-Sample Concentration, |v'/M ^IV'la
— j
INPUT
DATA/UNITS
^
KEYS
H
1
f II PEG
DSP 1 2
A II
R/S l|
Vs II
1 R/S II 1
R/S II
R/S II
R/S II
R/S il
H
R/S II
II
RTN ][
f II
REG If
CLX II
H
H
H
II 1
1 II
1 II 1
1 II
OUTPUT
DATA/UNITS
Figure 7. HP-65 Program.
442
-------
HP-65 Program Form
Title .
Ammonia In Exhaust (Linear)
SWITCH TO W/PRGM. PRESS f , , PROM I TO CLEAR MEMORY
Page.
.of .
KEY
ENTRY
LBL
A
3
5
-
3 .
1
~
5
10 2
8
X
2
9
:
9
2
f
R/S
20 X
STO 1
R/S
4
6
0
+
RCL 1
g :Tv
30 g 1/X
R/S
X
R/S
X
R/S
X
R/S
R/S
-to x
0
g
4
4
R/S
0
^ 0
=>320-C6I6
CODE
SHOWN
23
11
83
81
71
83
81
84
71
33 01
84
61
34 01
35 07
81
35 04
84
71
84
71
84
71
84
81
84
71
83
71
84
83
COMMENTS
Input-Sample Vol. ft3
Inout-Barometer , "Ha
Input-Same le TemD. °P
Input-Dilution Factor
Input-Abs. Soln. Vol. m:
Input-Stand. Cone. ugSO
Input-Stand. Area
Input-Sample Area
,
KEY
ENTRY
0
1
f
1
X
R/S
RTN
RO
70
80
-2/ml
90
'00
CODE
L SHOWN
71
84
,
COMMENTS
i
REGISTERS
Ri
R2
Ra
84
Rs
Re
R?
Rs
Ro
LABELS
A
B
C
D
E
0
1
2
3
4
5
6
7
a
<3
FLAGS
1
2
TO RECORD PROGRAM INSERT MAGNETIC CARD WITH SWIIC- SET AT W'PRGM
Figure 7 (Cont'd). HF--65 Program.
443
-------
SWRI PROJECT NO, /
FUEL;£/9-a37-oo/_ cvs NO. ^
SAMPLE COLLECTION BY: J.c.C.
GENERAL COMMENTS:
TEST NO. oo a. TEST DATE; / /I 72 VEHICLE: PX/)c7/C£
TUNNEL SIZE a/«. DRIVER: ALto/ii MILES! /.
CHEMICAL ANALYSIS BY:^aa:2a^.CALCULAT30NS BY:.
NO.
1,
2
3
4
5
6
7
8
X
*iRT VT-Mfi
CYCLE
FTP-1
4
FTP- 2
FTP- 3
SET- 7
HFET
NYCC
BG
X
SAMPLING CONDITIONS
VOLUME
FT3
l.SHIo
2.OO
vsoo
S"000
S200
3300
©
SAMPLE
AREA
;2500
saoo
27OO
3SOO
3S 00
3SOO
2SOO
©
^toNJ^
/S9O
^300
/980
,2370
2V &O
2.WO
3W
PPM
*./?
3.1k.
2.11
3.2S
S3?
3-3^
as/
/
Figure 8. NH3 Data Sheet.
-------
The next step involves the determination of the amount of NH collected
in the bubbler . This is obtained by multiplying the volume of absorbant
(absorb, vol.) and the dilution factor (DP) by the concentration of ammonia
collected.
yg NH3
°-944 csa' x ^sorb. vol, m£ x DF = Ug NH
To find the concentration of NH3 in exhaust the volume of gas collected
first needs to be corrected to the specified temperature and pressure (68°F=
528°R, 29.92 in Hg) . The volume as read from the digital readout in cubic
feet needs to be changed to cubic meters .
VOL, ft3 x 29.92 in Hg x (TEMP, °F + 460)
• 35.31 f^S E'p" in Hg 528°F
m3
where VOL = volume of gas collected, ft3
B.P. = collection pressure, in Hg
TEMP = collection temperature, °F
The concentration of NH3 can be calculated by dividing yg NH3 by the volume
of gas.
yg NH+
0.944 x -^ x absorb, vol, m£ x DF
st
VOL, ft3 B.P., in Hg x 528*R
ft-3
35.31 ^^3— 29.92 in Hg (TEMP, °F + 460)
yg NH4+
0.944 x cst, —^ x PAsa x absorb, vol, m£ x DF
VOL, ft x PAst
35.31 ft3 x 29.92 in Hg (TEMP, °F + 460)
B.P., in Hg x 528°R
yg NH_.
m3
This concentration can also be expressed in ppm NH by taking into
consideration the density of the gas at the desired conditions (68°F, 29.92
in Hg) . The fifth edition of the Matheson Gas Data Book lists the specific
gravity of ammonia gas at 70C'F and 1 atm pressure as 1.411 Vg. rJ-1he inverse
445
-------
of th-i specinc gravity gives a density of 0.709 j~. If the volume of l£ of
gas is corrected for temperature,
v = i £ x 528C'F = 0.966 £,
V X X (7G°F + 460)
it can be divided into the weight of the gas to give the density at 68°F.
_ o 712 3. = 712 U3.
" °'712 L
0.996
= 0.00141
yg NH
yg N
When the inverse of dersity is multiplied by the concentration in — —~
in
tlie concentration is given in ppm NH3.
3 — x 0.00141 - = -r- = Ppm NH3
m pg NH3 m
Sample Calculation
The two examples will be calculated from information recorded on the
data sheet. This information does not necessarily represent actual experi-
mental data but serves as a means of confirming calcuations done by hand
with those done with the Hewlett Packard Calculator.
Example 1
Assume that in the FTP-2 driving cycle that 2.652 ft3 of dilute exhaust
is collected in 25 mil of 0.01 N H2SO4 at 74.2
-------
0.944 x 9 x 3800 * 25 x 1 x 35.31 x 29.92 x (74.2 + 460)
2.652 x 4800 x 29.42 x 528
= 2300
m3
ppm NH3 = 2300 x 0.00141 = 3. 24 ppm NH3
Example 2
Assume that 1.890 ft of dilute automotive exhaust was collected in 25
mX, of 0.01 N H^jSO^ during the NYCC driving cycle. The sampling conditions
during this test were 74.7°F and 29.49 in Hg. When the undiluted sample was
injected into the ion chrconatograph the peak produced had an area of 3500
counts. When an 8 y standard was injected it produced an area of 5200
counts.
The. same equations in Example one are used to give concentrations of
NH3 of 2400 yg , 3 and 3.34 ppm NH3.
m
447
-------
REFERENCES
Braker, W. and Mossman. A. L., Matheson Gas Data Book, 5th Edition, East
Rutherford, N. J., 1971, pg. 17.
Furman, N. H., (Ed.), Scott's Standard Methods of Chemical Analysis, 5th
Edition, Van Nostrand, N. Y., 1958, Vol. 1, pg. 636.
Pifer, C. W. and Woll-Lsh, E. G., Anal. Chem. , Vol. 24, pg. 519, 1952.
Kolthoff, I. M. and Stenger, V. A., Volumetric Analysis, Interscience,
N. Y., 1947, Vol. 2, pg. 125.
Furman, N. H., (Ed.) r Scott's Standard Methods of Chemical Analysis, 5th
Edition, Van Nostrand,, N. Y. , 1958, Vol. 1, pg. 632.
Pierce, W. C. Haenisch, E. L., and Sawyer, D. T., Quantitative Analysis,
Wiley, N. Y., 1958, pg. 260.
Kolthoff, I. M. and Stenger, V. A., Volumetric Analysis, Interscience,
N. Y., 19754, Vol. 2, pg. 168.
Milner, O. I. and iah:.2rr R. J., Anal. Chem., Vol. 32, pg. 294, 1960.
Conway, E. J., Micro-Diffusion Analysis and Volumetric Error, 2nd Edition,
Crosby-Lockwood, London, 1947,
Altmann, C. J. G., de Heer, B. H. J., and Hermans, M. E. A., Anal. Chem.,
Vol. 35, pg 596, 1963.
Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans, Green,
London, 1961, pg, 254.
Blinn, R. C. and Gunther, F. A., Anal. Chem., Vol. 29, pg. 1882, 1957.
Kolthoff, I. M. and Stenger, V. A., Ind. Eng. Chem., Anal. Ed., Vol. 1,
pg. 79, 1935.
Kolthoff, I. M. and Belcher, R., Volumetric Analysis, Interscience, N. Y.
1957, Vol. 3, pg. 582.
Welcher, F. J., Organic Analytical Reagents, Van Nostrand, N. Y. , 1947,
Vol. 1, pg. 376.
Rowe, D. J., Gas Journal, Vol. 265, pg. 49, 1951.
448
-------
Burns, E. A., The Analysis of Exhaust Gases Trapped from Ablating Nozzles,
TRW Report No. 9840-6001-TU 000, Redondo Beach, Cal., 1964.
Nyman, C. J. and Johnson, R. A., Anal. Chem., Vol. 29, pg. 483, 1957.
Norton, 0. R. and Mann, C. K., Anal. Chem., Vol. 26, pg. 1180, 1954.
Clear, A. J. and Roth, M., in Treatise on Analytical Chemistry, Kolthoff,
I. M. and Elving, P. J. (Eds.), Interscience, N. Y., 1961, Part II, Vol. 5,
pg. 284.
Kolthoff, I. M. , Stricks, W., and Morren, L., Analyst, Vol. 78, pg. 405,
1953.
Laitinen, H. A. and Woerner, D. E., Anal. Chem., Vol. 27, pg. 215, 1955.
Arcand, G. M. and Swift, E. H., Anal. Chem., Vol. 28, pg. 440, 1956.
Krivis, A. F., Supp, G. R., and Gazda, E. S. , Anal. Chem., Vol. 35, pg.
2216, 1963.
Christian, G. D., Knoblock, E. C., Purdy, W. C., Anal. Chem., Vol. 35,
pg. 2217, 1963.
De Ford, D. D., Johns, C. J., and Pitts, J. N., Anal. Chem., Vol. 23,
pg. 941, 1951.
Segal, N. S. and Wodley-Smith, R., Anal. Chem., Vol. 38, pg. 829, 1966.
Sambucetti, C. J., Anal. Chem., Vol. 38, pg. 105, 1966.
Barendrecht, E. and Janssen, N. G. L. M., Anal. Chem., Vol. 33, pg. 199,
1961.
Burns, E. A., Unpublished results, 1963.
Jenkins Jr., R. W., Cheek, C. H., and Linnenbom, V. S., Anal. Chem.,
Vol. 38, pg. 1257, 1966.
Miettinen, J. K. and Virtanen, A. I., Ann. Acad. Sci. Fennicae, ser. A.
II, No. 41, 1951; through Chem. Abst. Vol. 46, pg. 11500, 1952.
Gilbert, T. R. and Clay, A. M., Anal. Chem., Vol 45, pg. 1757, 1973.
Le Blanc, P. J. and Sliwinski, J. F., Am. Lab., Vol. 5, pg. 51, 1973.
Thomas, R. F. and Booth, R. L., Environ, Sci. Tech., Vol. 7, pg. 523,
1973.
Woodis Jr., T. C. and Cummings Jr., J. M., JAOAC, Vol. 56, pg. 373, 1973.
449
-------
Stockdale, D., Analyst, Vol. 84, pg. 667, 1959.
Sadck, F. S. and Reilley, C. N., Anal. Chem., Vol. 31, pg. 494, 1959.
Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 49, pg. 38, 1960.
Crane, F. E. and Smith, E. A., Chemist-Analyst, Vol. 52, pg. 105, 1963.
Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 566.
Furman, N. H., (Ed.), Scott's Standard Methods of Chemical Analysis.
5th Edition, Van Nostrand, N. Y., 1958, Vol. 1, pg. 2336.
Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 1092.
Steyermark, A., Quantitative Organic Microanalysis, Blakestone, N. Y.,
1951, pg. 56.
Taras, M. J., in Colorimetric Determination of Nonmetals, Boltz, D. F.,
(Ed.), Interscience, N. Y., 1958, pg. 84.
Miller, G. L. and Miller, E. E., Anal. Chem., Vol. 20, pg. 481, 1948.
Thompson, J. F. and Morrison, G. R., Anal. Chem. Vol. 23, pg. 1153, 1951.
Kruse, J. M. and Mellon, M. G., Anal. Chem., Vol. 25, pg. 1188, 1953.
Prochazkova, L., Anal. Chem., Vol. 36, pg. 865, 1964.
Scheurer, P. G. and Smith, F., Anal. Chem., Vol. 27, pg. 1616, 1955.
Bolleter, W. T., Bushman, C. J., and Tidwell, P. W., Anal. Chem., Vol. 33,
pg. 592, 1961.
Zitomer, F. and Lambert, J. L., Anal. Chem., Vol. 34, pg. 1738, 1962..
Williams, D. D. and Miller, R. R., Anal. Chem., Vol. 34, pg. 225, 1962.
Howell, J. H. and Boltz, D. F., Anal. Chem., Vol. 36, pg. 1799, 1964.
Gunther, F. A., Barkley, J. H., Kolbezen, M. J., Blinn, R. C., and Staggs,
E. A., Anal. Chem., Vol. 28, pg. 1985, 1956.
Kolbezen, M. J., Eekert, J. W., and Wilson, C. W., Anal. Chem., Vol. 36,
pg. 593, 1964.
Pierson, R. H., Fletcher, A. A., and St. Clair Gantz, E., Anal. Chem.,
Vol. 28, pg. 1218, 1956.
450
-------
Vandeveene, L. and Oudewater, J., Centre Beige D'Etude et de Documentation
des Eaux, Vol. 352, pg. 127, 1973.
Banwart, W. L., Tabatabai, M. A. and Bremner, J. M., Comm. Soil Sci.
Plant Anal., Vol. 3, pg. 449, 1972.
Byrne, E. and Power, T., Comm. Soil Sci. Plant Anal., Vol. 5, pg. 51, 1974.
Me William, D. J. and Ough, C. S., Amer. J. Enol. Viticult, Vol. 25, pg.
67, 1974.
Sigsby Jr., J. E., Black, F. M., Bellar, T. A., and Klosterman, D. L.,
Environ. Sci. Tech., Vol. 7, pg. 51, 1973.
Zweidinger, R. B., Tejada, S. B., Sigsby Jr., J. B., and Bradow, R. L.,
"The Application of Ion Chromatograpny to the Analysis of Ammonia and
Alkyl Amines in Automotive Exhaust," Symposium on Ion Chromatograhpic
Analysis of Environmental Pollutants, EPA, Research Triangle Park, N. C.
April 1977.
451
-------
APPENDIX L
SULFATE PROCEDURE
452
-------
THE MEASUREMENT OF SOLUBLE SULFATE IN EXHAUST
As used by
Department of Emissions Research
Southwest Research Institute
Developed by
Mobile Source Emissions Research Branch
Envxronmental Sciences Research Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina
January 1977
453
-------
Determination of Soluble Sulfates in Automobile
Exhaust by Automated HPLC Modification of the
Barium Chloranilate Method
by
Silvestre B. Tejada, John E. Sigsby, Jr.
and Ronald L. Bradow
Mobile Source Emissions Research Branch
Environmental Sciences Research Laboratory
1. Principle and Applicability
1.1 Principle
Automotive exhaust is vented into a dilution tunnel where it is mixed
with a flowing stream of cool filtered air. In the tunnel, the SC>3 reacts
rapidly with water in the exhaust to form sulfuric acid aerosols. The aero-
sols are allowed to grow to filterable size range and are collected on fluoro-
carbon membrane filter downstream of the tunnel via isokinetic probes mounted
in the flowing aerosol stream. Particulate sulfate salts are collected as
well.
Sulfuric acid on the filter is converted to ammonium sulfate by exposure
to ammonia vapor. The soluble sulfates are leached from the filter with a
measured volume of 60% isopropyl alcohol - 40% water solution (60% IPA). A
fixed volume of the sample extract is injected into a high pressure liquid
chromatograph (HPLC) and pumped through a column of strong cation exchange
resin in Ag+ form to scrub out the halides (Cl~, Br~), then through a column
of strong cation exchange resin in H+ form to scrub out the cations and con-
vert the sulfate to sulfuric acid and finally through a reactor column of
barium chloranilate crystals to precipitate out barium sulfate and release
the highly UV absorbing chloranilate ions. The amount of chloranilate ions
released is equivalent to the sulfate in the sample and is measured by a
sensitive liquid chromatograph UV detector at 310-313 nanometers. All the
reactions and measurement take place in a flowing stream of 60% IPA. The
scrubber and reactor columns also function as efficient filter media for any
solid reaction products formed during passage of the sample through the
column system.
1.2 Applicability
The method as specified is applicable to the determination of soluble
sulfates in automobile exhaust. It may be used for the analysis of sulfates
454
-------
in samples where the sulfates can be leached out with water or aqueous IPA
solution. Aqueous extracts must be made up to 60% IPA before they can be
analyzed .
2. Interferences
2.1 Cationic interferences are removed by the strong cation exchange resin
in H form.
2.2 Halide interferences are removed by the strong cation exchange resin in
Ag+ form. Other anions which form insoluble salts with silver are also
removed.
2.3 Sulfite is measured quantitatively as sulfate.
2.4 Anions which form strong acids after passage through the cation exchanger
in H+ form interfere positively. Nitrate at 60 ygs/m£ gives an apparent
sulfate response corresponding to 8 ugs/m£.
2.5 Presence of anionic interference is manifested by a negative peak
immediately preceding the positive apparent sulfate peak.
2.6 Organics with absorption bands at 310-313 will interfere positively.
3. Range, Sensitivity and Precision
3.1 The absolute amount of SO, normally injected into the system is between
0 - 12.5 ygs. For a typical sample injection volume of 0.5 m£, this
translates to a concentration working range of 0 - 25 ygs/m£. Working
range can be extended by using a smaller sample injection volume for
concentrated samples or conversely, by using a larger injection volume
for dilute samples .
3.2 Minimum detectable quantity of SO, is in the low nanogram range. There
are commercially available liquid chromatograph UV detectors capable of
detecting 5 nanograms of sulfate at a signal to noise ratio of about 5.
Figure 1 shows typical recorder response for a 0.5 rajj, injection of
sulfate samples at trace concentration levels (0.01 - 0.1
3.3 Precision better than 3% at 0.5 yg/m£ S0~ level and better than 2%
between 1 and 20 ygs/m£ for four repetitive runs have been attained.
Table 1 and Figure 2 show typical reproducibility obtained with the
automated BCA system.
4. Apparatus
A schematic -of the principal components of the automated BCA set up is
shown in Figure 3.
4.1 Hardware
4.1.1 Basic System
455
-------
4.1.1.1 High pressure liquid chromatograph pump (LP). The pump must be
capable of delivering liquids at flow rates of at least 3 m£/min. at
pressures as high as 1200 psi. Liquid pumps capable of delivering
pulcelis- ar.J cons tail t liquid flow are recommended for good
quantitation. Most HPLC pumps in the market are adequate.
4.1.1.2 UV detector (D) equipped with low dead volume (8 y£) flow-through
cell and a grating, prism or appropriate interference filter to
isolate, a narrow radiation band centered at 310-313 nanometers. For
low noise and long life, a UV detector equipped with low pressure
mercury lamp and a 313 nanometer narrow band pass interference filter
is recommended.
4.1.1.3 A txvo position, six port, high pressure, low dead volume sample
injection valve (SV). This must be equipped with interchangeable
exteinal loop (L). Two loop sizes are desirable: a 0.5 m£ volume
for the 0-20 yg/m£ range and a 100 y£ volume for 0 - 100 ygs/m£
range. The sample valve must be equipped with an external handle
for manual operation or an air actuator for remote and/or automatic
operation. These valves are commercially available and have pressure
ratings as high as 7000 psi.
4.1.1.4 Recorder. Must be multi-range, with chart speed as low as 10 minutes
per inch. L'_-.aI channel is preferable so that chromatogram can be
simultaneously recorded at two different sensitivities.
4.1.1.5 Strong cation exchange resin column (CX-Ag ), 4mm l.D. by 1/4 inch
O.D. by (.-> inches long stainless steel column packed xv'ith chromato-
graphic grade, strongly acidic cation exchange resin in silver (Ag )
form.
4.1.1.6 Strong cation exchange resin column (CX-H ), 4 mm l.D. by 1/4 inch
O.D. by 9 inches long stainless steel column packed with strongly
acidic cation exchange resin in hydrogen (H ) form.
4.1.1.7 Barium chloranilate column (EGA), 4 mm l.D. by 1/4 inch by 1 inch
long stainless steel column packed with crystalline barium chlor-
anilate.
4.1.1.8 1/4" to 1/16" stainless steel reducer preferably fitted with 5 micron
pore size frit - for column inlet and end fittings.
4.1.1.9 1/4" to 1/4" stainless steel unions.
4.1.1.10 1/16" to 1/16" low dead volume stainless steel couplings to inter-
connect CX-Ag+ to CX-H+ to BCA columns.
4.1.1.11 1/4" and 1/16" nuts and ferrules.
4.1.1.12 Reservoir (LR) for the solvent (60% IPA).
456
-------
4.1.2 Options - These items are needed for automating the basic system.
4.1.2.1 Integrator, for measuring peak areas. Recommended unit must have
baseline tracking capability and possibly with built-in calculation
accessory. The integrator extends useful dynamic range of detector
response for a given sample loop size beyond that of the strip chart
recorder.
4.1.2.2 Peristaltic pump, (PP), to draw sample from its container and load
it into sample loop. Silicone pump tubing is recommended.
4.1.2.3 Automatic sampler (AS), available from commercial sources. This is
needed if the number of samples to be analyzed is large and manpower
is limited.
4.1.2.4 Three timer relays to control pump, sampler, injection valve and
integrator and provide automatic reset for cyclic operations.
4.1.2.5 Prepackaged sampler systems for HPLC application are commercially
available.
5. Principle of Operation
Solvent (60% IPA) in reservoir LR (Figure 3) is continuously pumped
by an HPLC pump LP through a column of strong cation exchange resin in silver
form, CX-Ag , then through a column of strong cation exchange resin in
hydrogen form, CX-H+, then through a reactor column of barium chloranilate,
BCA, and finally through a flow-through cell of a UV detector, I), and on to
waste. CX-Ag+ removes the halides (Cl~, Br~, F~) and other anions which
precipitate with silver; CX-H+ removes metallic cations and converts the
sulfate to sulfuric acid, and the BCA reacts with the sulfate to form barium
sulfate precipitate and a soluble UV absorbing dye, chloranilic acid and its
ions. Background absorbance at 310 - 313 nanometers is continuously measured
and monitored on a strip chart recorder.
Sample is introduced into the system without flow interruption by
means of a two-way six port low dead volume sample injection valve SV. In
load position "A" (see Figure 3) the peristaltic pump, PP, draws the sample
from a cuvette in the automatic sampler AS_ and pumps it into port 4 filling
external sample loop L^ then through port 5 to waste. (Sample loop loading
may also be accomplished by pushing the sample through port 4 by means of a
syringe.) The high pressure liquid flow comes in through port 1, bypasses
the loop :L, comes out of port 2, and continues on through the three columns
and the flow-through cell of the UV detector ID and on to waste
After loop L_ is loaded with sample, injection valve SV switches to
inject position "B".~ The high pressure stream purges the loop and pushes the
sample through the cation exchangers and then through the BCA column where
the color reaction takes place. The BaS04 precipitate is retained in the
column while the acid chloranilate is carried by the flowing liquid through
the detector system for colorimetric measurement.
457
-------
6. Apparatus
6.1 Pipette, volumetric, 1, 2, 4, 5, 8, 25, 50, 100 mH
6.2 Pipette, measuring, 1, 2, 5, 10 m&
6.3 Automatic burette, 25 m£
6.4 Volumetric flasks, 10, 25, 50, 100, 500, 1000, 2000 m£
6.5 Bottles, polypropylene, with screw caps, 30, 60, 125, 250, 500, 1000
6.6 Microbalance
6.7 Vortex test tube mixer
6.8 Centrifuge
6.9 Magnetic mixer
6.10 Magnetic bars
6.11 Graduated cylinders
6.12 Automatic dispenser pipet, 5, 10, 20 m£ (optional)
6.13 Automatic burette, 10 m£ (motor driven, optional)
6.14 Ammoniation chamber (Figure 5)
7. Reagents
7.1 Isopropyl alcohol (IPA), spectro quality grade or equivalent
7.2 Water, doubly deionized, distilled.
7.3 60% IPA. Add 4 parts water to 6 parts IPA by volume. Store in
tightly capped bottles.
7.4 Barium chloranilate, suitable for sulfate analysis. Must be crystalline,
granular, preferably with average granule length of about 200 microns.
Finer particles cause excessive column pressure drop.
7.5 Cation exchange resin, chromatographic grade, strongly acidic, hydrogen
form, 100 - 200 mesh.
7.6 Hydrochloric acid (4N). Add 30 m£ concentrated hydrochloric acid to
60 m£ of deionized water.
7.7 Ammonium sulfate, primary standard
458
-------
7.8 Silver nitrate (IN). Dissolve 17 grams silver nitrate in deionized
water and make up to 100 mJL Store in the dark in an amber colored
reagent bottle.
8. Procedure
8.1 Column Preparation
8.1.1 Barium chloranilate column. In order to prepare a full column with
minimum dead volume, connect two lengths of 4 mm I.D., 1/4" O.D.
stainless steel tubing as shown in Figure 4 with a = 1", b = 2".
Connect a small funnel to open end of B with a flexible tubing sleeve.
Fill the funnel halfway with barium chloranilate and thump the tube
several times or use a vibrator (i.e., electric pencil engraver) to
pack the solid in the column. Continue the operation until B is
completely filled. Remove the funnel and cap the open end of B with
a 1/4" to 1/16" reducer fitted with 5 micron stainless steel-frit.
The 5 micron stainless steel frit in column A may be replaced with a
stainless steel wire screen with nominal porosity of 10 microns.
Connect B to the HPLC pump, connect a flexible tubing at A and direct
the tubing to waste reservoir. Fill HPLC pump reservoir with 60% IPA.
Activate HPLC pump according to manufacturer's instructions, set flow
at 3 - 4 mJl/min. and let solvent flow for at least 20 minutes. This
procedure removes the fines which can cause background drift and at
the same time compresses the barium chloranilate crystals to fill the
dead volume in column A. Deactivate the HPLC pump, disconnect -the
composite column from the pump, then column A from-the composite
column. Connect a 1/4" to 1/16" reducer fitted with 5 micron frit to
open end of A.
8.1.2 Cation exchange resin columns
8.1.2.1 Cation exchange resin, hydrogen form. Add strongly acidic cation
exchange resin, 100 - 200 mesh, to 160 mJl of 4N HC1 in a 250 mJl
Erlenmeyer flask until a wet volume equivalent to 40 m& has settled
at the bottom. Soak for at least 3 hours with occasional stirring
with a glass rod. Decant the acid, add 100 m£ deionized water, stir
and decant the liquid as soon as most of the solids have settled at
the bottom. This procedure removes most of the fines. Repeat
rinsing procedure several times until the rinse liquid gives neutral
reaction to pH paper. Transfer half of the resin to a 150 m&
Erlenmeyer flask for conversion .to the silver form.
Connect two sections of 1/4" O.D., 4 mm I.D. stainless steel tubing
as in 8.1.1 with a = 9 and b = 5. The reducer on the outlet end of
A should have a 5 micron stainless steel filter frit. These frits
-are-available commercially. -The .frit must be.able to .withstand high
pressure (1200 psi). (If the cation exchange resin breaks through
the frit and comes to contact with the barium chloranilate column,
the column gets plugged.) Connect a small funnel to open end of B
with a flexible tubing sleeve. Clamp composite column vertically
and connect open end of A to vacuum line equipped with liquid trap.
459
-------
Fill funnel with deionized water and turn vacuum slowly BO that the
column is completely filled with water. Add enough water so that
water level is in the funnel cone; stop the vacuum and add the
slurr^- of freshly washed r ;sin (H+ form). Ltt <-he resin ^etnie by
gravity until the resin top is halfway in the tunnel stem. Open the
vacuum slowly and keep adding the resin slurry until the composite
column is completely filled. Proceed as in 8.1.1 beginning with the
sentence "Remove funnel and cap open end of B ... .
8.1.2.2 Cation exchange resin, silver form. Add 60 m& of IN AgNf>3 solution
to the other half of the washed.cation exchange resin, hydrogen
form, in a 150 m£. Erlenmeyer flask. Stir with a gJass rod, cover
the flask with aluminum foil and soak the resin overnight.
Decant the AgN03 solution into a waste reservoir. Add 100 m£ de-
ionized water, ritir and decant the li.juid as soon as most of the
resins have settled at the bottom. Repeat t.he rinsing procedure
until the rinse liquid remains clear when treated with a few drops
of 4N HC1.
Connect two sections of 1/4" O.D.. 4 mm I.D. stainless steel tubing
as in 8.1.1 with a = 6" and b = 5". Load the column following the
procedure described in section 8.1.2.1.
8.2 Priming Systen for Analytical Run
Connect outlet end of cation exchange resin (Ag form) column to inlet
end of cation exchatnii resin (H+ form) column with a low dead volume 1/16"
to 1/16" stainless szeel tubing connector. Similarly connect the outlet end
of the second column to the barium chloranilate column. Plumb the composite
column to the automated set up as shown in Figure 3. Fill solvent reservoir
LR with 60% IPA, activate the HPLC pump, detector, recorder, sample injection
valve, sampler and peristaltic pump. During this initial operation dip the
sampling probe in 100 m£ of 60% IPA. Set"the liquid flow rate at 3 m£/min.
(Flow rate is conveniently measured by directing the effluent from the UV
detector to a microburette and measuring the time in seconds needed to fill
a volume of 3 m£.) Let run for at least 30 minutes. Deactivate sample
injection valve, sampler and peristaltic pump. Leave other components in
operating mode. When absorbance background is stable at the appropriate
sensitivity setting of the detector, the system is ready to analyze sulfate
samples.
8.3 Preparation of Standards
Stilfuric acid, sodium sulfate or ammonium sulfate may be used as
standards. Ammonium sulfate is preferred.
8.3.1 804 (100 ugs/m£) standard, alcoholic stock solution. Dissolve 275000
± 100 ygs of primary standard ammonium sulfate in 200 m& of deionized
water in a 2000 m£ volumetric flask. Add 300 m£ pure IPA, shake
vigorously until thoroughly mixed, and make up to volume with 60% IPA.
Store in clean polypropylene boftlas. (Note: a_. There is a volume
460
-------
decrease of about 2.7% when two .parts of water is mixed with 3 parts
of IPA. b_. Do not use detergents nor dichromate-sulfuric acid
solution for cleaning glasswares. Sulfate background from these
sources are difficult to remove. 50% (v/v) nitric acid/water is
preferable.) Prepare from this stock solution 804 calibration
standards (0.5 to 20 ygs/mJl) by dilution of appropriate aliquots with
60% IPA. Store standards in capped polypropylene bottles.
8.3.2 804 (100 ygs/m£) standard, aqueous stock solution. Dissolve 275000
± 100 ygs of primary standard ammonium sulfate in 200 m£ of deionized
water in a 2000 m£ volumetric flask and make up to volume with de-
ionized water. Store in polypropylene bottle.
8.3.3 Alternative method of preparing calibration standards. Use a repeti-
tive dispenser, burette or automatic syringe pump. Using either the
alcoholic or aqueous SO^ (100 ygs/m£) stock solution dispense appro-
priate volumes containing 10, 20, 40 ..., 180, 200 ygs of SO? into 30
or 60 m£ polypropylene bottles. Prepare 10 for each sulfate level.
Evaporate the liquids completely by placing the bottles uncapped in an
oven maintained at 80 - 100°C. Cool, cap the bottles and store until
ready to use. These solid standards can then be extracted in the same
manner as the filter samples.
8.4 Ammonia Treatment of Filter Samples
This treatment converts sulfuric acid particulate into ammonium
sulfate. Conversion to ammonium salt was observed to improve precision of
sulfate measurement. Sample losses from accidental contact of the filter
surface with another surface is minimized. An additional advantage is that
the sulfate is converted to the same form as the calibration standard.
Figure 5 shows a simple schematic of an ammoniation set-up. Filter
samples on open Petri dishes are placed face up on perforated shelves of the
ammoniation box. The box is evacuated, valve V]_ is closed, and valve V£ is
opened. Ammonia from concentrated ammonium hydroxide fills the box and
converts sulfuric acid to ammonium sulfate. One hour exposure to ammonia
vapor is adequate. ¥2 is closed, most of the ammonia is pumped out and
passed through a KH9PO, scrubber column. Vacuum is released by switching
V.. to vent.
8.5 Extraction of Soluble Sulfates.
It is important that the water/IPA ratio in the sample and in the
mobile phase be the same. A sample richer in water content than the mobile
phase will momentarily increase the solubility of barium chloranilate and
will produce a positive peak above a flat background; that richer in IPA
will produce a negative peak. Variations will occur if the solvent used in
the preparation of the standards and in the extraction of the filter samples
were taken from a different stock as that used in the mobile phase. Therefore,
it is strongly recommended that the extraction solvent be taken from the same
stock solution as the mobile phase or, if possible, directly from the liquid
461
-------
reservoir of the HPLC pump. The use of solid standards as prepared in 8.3.3
will eliminate variability due to IPA/water mismatch.
3.5.1 From fluorccarbor. membrane filtern
Place filter in appropriate size polypropylene bottle. If approximate
sulfate level is known from previous analysis of similar samples,
measure adequate volume of 60% IPA to give sulfate concentration of
about 10 Ugs/mfc. Otherwise add 10 mi 60% IPA and cap the bottle
tightly. As a rule of thumb, the sulfates constitute about 40 - 50%
of the total particulate mass if most of the particulates on the filter
is sulfuric acid. Such cases are generally encountered with filter
samples from catalyst equipped cars run on non-leaded fuels. Parti-
culate mass loading of a filter is an important piece of information
in deciding appropriate volumes of extracting solution to use.
Shake the bottle vigorously until the filter collapses and is
completely immersed in the solvent. A vortex test tube mixer is
recommended. Twenty seconds shaking with a vortex test tube mixer is
adequate to leach the soluble sulfates from the filter. If there are
no visible suspended particulates, the clear solution can be used
directly for analysis. If suspended carbon or other particles are
apparent, filter about 5 m£ of the extract by using a syringe with a
U pore size fluorocarbon in-line filter. These filter syringes are
available commercially. If the particles are sufficiently large, they
can be removed from the bulk solution by centrifugation.
8.3.2 From glass fiber filters.
This procedure is for the extraction of 47 mm diameter filters.
Adjust the volume of the extracting solvent accordingly for different
size filters.
Place the filter in appropriate size polypropylene bottle. If approxi-
mate SO^ level is known, add adequate volume of 60% IPA to give sulfate
concentration of about 10 pgs/mJt. Otherwise, add 30 m 60% IPA and
cap bottle tightly. Shake with a vortex mixer until the glass fibers
disintegrate. Where large solvent volume and inordinately long shaking
time are required, stirring with a Teflon clad magnetic stirring bar
is preferable. Glass fibers are easily separated from the bulk liquid
by using a centrifuge. If finely suspended particles are present, as
in the case of diesel and some non-catalyst exhaust samples, syringe
filtration as mentioned in 8.5.1 must be used. In some cases a 0.2 y
pore size filter is necessary.
8.6 Water extracts.
Water extracts of filter samples may also be analyzed by the automated
BCA method. One important requirement, however, is that the solution must be
made up to 60% IPA (e.g., 4 parts of the extract must be added to 6 parts of
pure IPA, v/v) before the sample can be analyzed. If the water extract is
462
-------
mixed directly with pure IPA, volume shrinkage (about 2.7%) must be taken
into account in the calculation of concentrations.
Another approach is to evaporate completely the solvent in a known
volume of the extract in polypropylene bottle similar to the preparation of
solid calibration standard in 8.3.3. The residue may be ammoniated in the
plastic bottle as in 8.4 and extracted as in 8.5.1 or 8.5.2.
8.7 Analysis
Set instrument in operating mode, remove sample probe from the holder
and dip in 100 m£ 60% IPA. Adjust the flow rate at 3 m£/min., allow the
instrument to cycle several times until a stable background is obtained, then
remount sample probe to holder. Adjust loading time or peristaltic pump rate
so that at least twice the volume of the sample as the volume of the sample
injection loop has exited the waste port of the sample injection valve.
Adjust sample injection time so that peaks from successive sample injections
do not overlap. Fill sample cuvettes with samples and rinse solution (60%
IPA) and place according to a sampling pattern, blank, sample, blank, sample,
blank. . . The rinse solution (blank) is necessary to ascertain that there is
no memory effect from previous sample injection.
A series of at least 6 calibration standards spanning the concentration
range of interest is run before samples are run. A control standard may be
placed every 10 samples as a quality check on the stability of the system.
Dilution may be necessary if the sample peak height is beyond the range of
the calibration standard. If a large number of samples needs dilution, it
may be more convenient to merely change the size of the sample loop.
Plot peak height (or area if an integrator is also used) vs concentration
in ygs/mJl of the sulfate standard. The curve is not linear. Alternatively,
the peak height or area concentration data may be fitted into a polynomial of
the form:
234
y = a + a,x + a~x + a~x + a.x + ...
J o 1 2 3 4
where: y = sulfate concentration in ygs/m&
x = peak height or area
8.8 Calculations
Calculate the concentration of sulfate as ygs S04/m£ using the cali-
bration curve or the polynomial regression equation. Total soluble sulfate
(SOp in the filter is then given by:
d x Vo x d
where: V = total volume in m£ of original sample extract
o
(SO?) = sulfate concentration of the diluted sample in ugs/m&
4 d
463
-------
d - d illation factor
- 1, if there is no dilution of the extract before
analysis,
= Vvd
V = aliquot volume in m,1 of sample diluted with 60% 7PA
o
V, = final volume in m& of the aliquot sample after dilution
a with 60% IPA
Example:
Suppose 10 mi of 60% IPA was used to extract the soluble sulfate.s in
the filter and that mi of this was further diluted with 4 m£ 60% IPA to
bring detector response within calibration range. Suppose that the concen-
tration of the diluted sample was found to be 5 ygs/m£. Then,
(SO?) - = 5 ygs/.rai,
4 a
V = 10 m£
o
V = 2 mH
a
V, - 2+4
c
--- o mfc
(SO"), = 5 x 10 x (6/2)
4 t
- 150 ygs
464
-------
ui
Fi.gure 1. Chromatogram at trace sulfate levels, 0.01, 0.02, 0.05, 0.01 ygs SO^/ml. Flow rate at
3.2 ml/min. Detector sensitivity at 0.01 absorbance Units full scale. Sample volume inje ted = 0.5ml.
-------
en
Figure 2. Reproducitility of repetitive sample injections. Flow at 3.2 ml/min. Detector sensitivity
at 0.5 AUFS. Sample volume injected = 0.5 ml. Numbers above peaks are sulfate concentrations in ygs/ml.
-------
cn
soivent out
sol.vent
in
CO
solvent out
B -
LR -
LP -
P -
SV -
L -
sample
in
solv
sample
out
in
sample in
Position A
Load
Position B
Inject
buret
liquid reservoir
PHLC pump
Pressure Monitor
Sample Injection valve
Sample loop
CX-Ag+ - Cation exchange resin, silver form
CX-H+ - Cation exchange resin, hydrogen form
BCA - Barium chloranilate
D - UV detector
FM - Flow monitor
PP - Peristaltic pump
AS - Automated sampler
Figure 3. Schematic of an automated BCA sulfate set-up,
-------
1
L/4" union
r 5 micron STAINLESS
STEEL FRIT
T
\t 1M" to 1/16" REDUCER
Figure 4. Configuration for packing column.
468
-------
V,
CTi
VACUUM
PERFORATE
SHELF
VEHT
» /1 / I I
< FILTER ON
PETRI DISH
CONCENTRATED
Figure 5. Schematic of an ammoniation set-up.
-------
I'lSCRtl'AKCIEL; BETWEEN EIA AND DEPARTMENT OF EMISSIONS RESEARCH
BCA SULFATE PROCEDURE
Bottled N2 gas is used to drive 60% IPA from a reservoir through
the HFLC system at average rate of 4.5
4.1.1.2 - The Dupon<- 837 sample cell volume is 6.3 lj£.
4.1.1.3 - A dedicated 1 mj?, sa.np.Le loop is used on a manually-operated sample
injection valve.
4.1.1.4 - Recorder is dual channel, 10 mv full scale, with variable chart
speed set at 0.75 in./min. Only one recorder channel is used »
since chrornatogram sensitivity is controlled at range switch on
spectrophotometer .
4.1.1.6 - A 6-inch long cation column is used.
4.1,1.7 -A 2- I/ 2 inch long BCA column is used,
4.1.1.10 - One-quarter inch unions are used.
4.1.2 - Non-automated system is used.
6.5 - Polypropylene bottles (30 mi) are used for extraction of filters
in 60% IPA solution.
6.8 - We don't use centrifuge.
7.2 - We use deionized water checked with AgNO,, for dissolved salts.
7.3 - We store in a 5-gallon stoppered glass bottle with dispenser.
7.7 - Ours is ACS reagent, oven dried, and stored over silica gel in a
dessicator.
8.3.1 - Stock and calibration standards are stored in glass bottles.
8.3.2 - Standard solution stored in glass bottles.
8.3.3 - We don't use alternative method of preparing calibration standards,
470
-------
8.4 - The chamber containing sample filters in open Petri dishes is
purged with ammonia from concentrated ammonium hydioxide which
is then vented out via hood. The chamber is purged for 3 minutes
and then the filters are isolated in the ammonia-filled chamber
for one hour before their extraction in 60% IPA.
8.5.2 - We do not use glass filters for sulfate analysis.
8.6 - We do not do water extracts.
8.7 Used in automated system only.
471
-------
APPENDIX M
DMNA SAMPLING PROCEDURE
472
-------
DESIGN AND CALIBRATION OF SECONDARY DILUTER
A secondary diluter, to be used in conjunction with a constant volume
sampler, has been constructed. The secondary diluter was constructed:
1. to determine the concentration of N-nitrosodimethylamine in automo-
bile exhaust
2. to determine the composition of condensable organic vapors in auto-
mobile exhaust. Condensable organic vapors are those having mole-
cular weights from about 140 to 300 atomic mass units.
Factors which had to be taken into consideration for construction of the
secondary diluter were :
1. a proportional sample of automobile exhaust is needed
2. high levels of nitric oxide must be avoided because:
a. NO + amine •> nitrosamine artifact, which would interfere with
the detection of N-nitrosodimethylamine.
b. the oxidation of NO to N©2 with materials in the system and on
the traps would cause interference .
3. the sample must be presented to the Tenax G.C. trap at an appro-
priate flow rate and temperature so that collection is efficient and
materials are not driven off the trap. (We sample at 1.6 £/min and
60°C.)
The design criterion was to avoid oxidation of NOX and have a sample of
5 ppm or less of NOX, which was accomplished by using dry nitrogen to dilute
the automobile exhaust at 25:1. (Our dilution is 16.7:1.)
The secondary diluter is shown schematically in Figure 1 and its oper-
ating principles given below:
1. diluted automobile exhaust is pulled into the secondary diluter
through a 1/8" O.D. stainless steel tube at a flow rate of approxi-
mately 2.5 Vmin; this flow is dependent upon the downstream pres-
sure drop.
2. "Dry nitrogen for secondary dilution is added through a roots meter
and rotometer, in series, at approximately 68 £/min. This flow_is
independent of the downstream pressure drop. (We run at 55 A/nan.)
473
-------
1 3
Sample In
N,
Roots
Meter
Vacuum
Excess
Vacuum
Figure 1, Secondary diluter.
-------
3. Diluted automobile exhaust and dry nitrogen are mixed at this point
and pulled through a 3/8" O.D. stainless steel tube 13" long.
4. A magnehelic gauge is used to monitor the pressure drop across the
diluter. The reference side of the magnehelic is connected to the
primary dilution tube; this allows dilution ratios to be determined
without making corrections for differences in atmospheric and pri-
mary dilution tube pressure.
5. Three simultaneous samples are pulled through; (a) glass fiber fil-
ters, (b) Tenax G.C. traps, (c) rotometers, and (d) needle valves.
The needle valves are set to maintain a flow of 2.5 Vmin through
each sample trap. (We us.e Fluoropore filters and sample at 1.6
£/min.)
6. Excess flow is discharged from the diluter through a needle valve,
which is used to maintain the proper pressure drop across the
diluter.
During the operation, it is suggested that the inlet rotometer (2) be
maintained at 80 and the magnehelic gauge, (4) be maintained at 10" water,
which will give dilution ratio of 25:1. These gauges should be checked every
five minutes to be sure that these flows do not change. (Our rotometer is
set at 65 and the magnehelic at 25" water for a dilution ratio of 16.7:1.)
CALIBRATION
The rotometers for the" secondary diluter were calibrated by using a
roots meter and dry test meter.
The secondary diluter was calibrated by several different, but similar
experiments. The purpose of these experiments was to calibrate the dilution
ratio against diluter pressure drop. Each experiment involved measuring di-
lution ratios obtained at various constant diluter, pressure drops as meas-
ured by the magnehelic gauge.
It was found early in the experiments that a constant source of .vacuum
is needed in order to maintain a steady pressure drop across the diluter.
The experiments were conducted as follows:
1. A sample of propane span gas was placed in a large empty Tedlar bag;
the concentration of propane in the bag was determined by analyzing
the contents of the bag on a hydrocarbon analyzer.
One sample rotometer (5) was disconnected from its normal vacuum
source and replaced by a diaphram pump of sufficient size to pull
2.5 £/min. The reference side of the magnehelic was left open to
atmospheric pressure, a correction for this was made for this cal-
culating the data. The rest of the system was up as it would be for
normal operation. The bag of propane span gas was connected to the
sample inlet (2) . The diluted sample was then collected in an empty
Tedlar bag downstream of the diaphragm pump. This bag was then
475
-------
analyzed on a hydrocarbon analyzer and its concentration determined.
By dividing the concentration of hydrocarbon in the span gas by the
concentration found in the bag of diluted span gas, the dilution
ratios were found for different magnehelic gauge readings.
2. The diluter was then set up as it normally will be used, except that
one rotometer (5) was still connected to the diaphragm pump. A
cylinder of span gas containing propane, carbon monoxide and carbon
dioxide was then metered into the primary dilution tube. A sample
of this diluter span gas was collected in bags on the primary di-
luter and at the same time bags were collected on the secondary di-
luter. The bags collected on both primary and secondary diluters
were then analyzed on appropriate analyzers. By dividing the con-
centration found in the primary bag by the concentration found in
the bag of diluted span gas, the dilution ratios were found for dif-
ferent magnehelic gauge readings.
3. The diluter was set up as it would be for making a normal test, with
the exception that a rotometer was attached to the sample inlet (1).
Ambient air was pulled through this rotometer while the rest of the
diluter was under normal operation. By making corrections for pres-
sure drop in the diluter, the flow of dry nitrogen into the diluter
can be divided by the flow of ambient air into the diluter, giving
dilution ratios for different magnehelic gauge readings. The ex-
pression used for correcting the pressure drop is PiV^ = V2 where:
P = initial pressure
V. = initial volume
P = final pressure
V = final volume
A correction for temperature was not made because the temperature was
stable at approximately 20°C and thought this difference would be of little
consequence .
The dilution ratios determined by these three different experiments were
plotted against differential pressures measured by the magnehelic gauge using
the general expression for flow in pipes:
(ref •
where: AP = differential pressure
f = friction
L/D = length/diameter ratio
g = gravitational constant
v = specific volume of gas, I/density
V = linear velocity
476
-------
Since the sample inlet is of constant cross section, the linear velocity
is proportional to the volumetric flow rate. In any given set of experiments,
all variables in this equation other than velocity and pressure are constant.
Because of this, the equation reduces to:
AP = CV2 = CQ2
. _ sample flow rate
where Q = &—: -—.
volumetric
The dilution ratio, R, is given by:
f Q,
R- a
0
,* « ~ ^ ^N2 ~ ^N2 Constant
If Qs a, Qw2' then R
Q ' xs R R
s
Therefore: !_ _ c,, J££
R
and the plot of the reciprocal of the dilution ratio against 1/AP should be a
straight line.
Figure 2 presents all the raw data gathered in these calibrations and
plotted on this basis. A linear least squares fit for these data is also
presented in Figure 2. A single equation was found to adequately represent
the shole date base:
1/R = 0.01858 iAP 0.02133
r2 = 0.9819
standard error of estimate = 3 percent of mean value
Reference 1: Mark's Standard Handbook for Mechanical Engineers, 7th Edition,
McGraw-Hill, N.Y. (1967), p. 4-66.
477
-------
PROCEDURE USED BY SWRI FOR OBTAINING DMNA TRAPS
PREPARATION FOR RUNNING
1. Turn on 4 rheostats in back of DMNA cart at least 30 minutes prior
to test.
2. Remove trap holder with the filter case from the cart. Insert fil-
ter (FA) into filter case so that flow goes through the dull side
of the filter (note direction of arrow). Replace o-ring. The fil-
ter and the o-ring are not to be handled with bare hands. Fasten
trap holder back onto cart and tighten nut.
3. Check nitrogen (N2> bottle to make sure it has at least 1000 psi. If
necessary, N2 flow can be set with a dummy filter with the pump on.
Check dryer on N2 bottle and replace molecular sieve and silica gel
dryer if necessary.
4. Open the green valve on the sample line.
READY TO RUN
1. Obtain trap from refrigerator. Leave it in the glass tube until
ready to use. With disposable gloves, place trap in holder. Hand
tighten both end fittings. Replace spun glass ball in glass con-
tainer and recap.
2. Just prior to start of test, turn on N2 flow. At honk of horn, turn
on sample pump and set the large N2 flowmeter with the top of the
float at 65 and the magnehelic at 25" I^O. Both knobs need to be
adjusted simultaneously since they are dependent on each other.
These two will have to be monitored throughout the test. Flowmeters
1, 2, and 3 should already be adjusted to their proper settings
(10, 5.2, 10 respectively). If not, use the large black knobs to
readjust. Temperature of the system is monitored at four points:
sample inlet from CVS, N2 inlet, mixing point and at trap. It
should be set at 60°C (140°F) by adjusting rheostats in back of
cart. These temperatures are recorded on the data sheet about once
a week.
END OF TEST
The traps are run the full cold FTP and then the hot 505. During the
soak between FTP's, the pump and the N2 bottle are shut off.
478
-------
1. When the test is over, the trap is again carefully handled with
gloves and placed back in the glass container. The container is
labeled on the cap and on the side with the trap number (from log
book) , test number and run date. The container is then placed in
the 1-gallon can in the refrigerator. Test number and run date are
recorded in the log book. Test number is also recorded on the data
sheet (same one temperatures were recorded on).
2. Close green flow valve on sample line.
3. Turn off N2 regulator.
4. Turn rheostats off.
479
-------
TABLE 1. N SECONDARY DILUTEE CALIBRATION VALUES
FEBRUARY 10, 1976
A
B
R2
(2)
(4)
VAP Carbon Monoxide
2.0 0.01712
*
27"3Q^ _ _. ,_— _
31623 0 03650
0 O1 fLCi _ __
37*417 0 046 S 1
-3 Q79QQ
41 O "3 1 1 __
1/DF = A ( V'AP~ )
0.01709 0.02030
-0.01765 -0.02630
0.99749 0.99414
0.01653 0.01430
0.05071 0.05491
1/DF
Propane
Or\ i cmQ
On^T 7 c
• Uo JL / D
«
0 * 03937
Oc\c~\ i n
. UDJLJLU
Or\ c 70 1
+ B
0.01684
-0.01622
0.97560
0.01747
0.05115
,-.
Rotameter
On 0 1 1 f-\
Or\ "3 on T
*
O O/lOfi7
On t^m 7
. UDUJLo
Overall Data
0.01858
-0.02133
0.0818
0.01584
0.05301
480
-------
TECHNICAL REPORT DATA
(nease read Instructions on the reverse before completing)
1. REPORT NO. 2. ' '
1 EPA-600/2-79-017
14. TITLE AND SUBTITLE
ANALYTICAL PROCEDURES FOR CHARACTERIZING UNREGULATED
POLLUTANT EMISSIONS FROM MOTOR VEHICLES
17. AUTHOR(S)
Harry Dietzman, Lawrence Smith, Mary Parness,
1 and Robert Fanick
J9. PERFORMING ORGANIZATION NAME AND ADDRESS
1 Southwest Research Institute
I 8500 Culebra Road
San Antonio, Texas 78284
J12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences and Research Laboratory - RTP, N(
Office of Research and Development
Environmental Protection Agency
Research Triangle Park, N.C. 27711
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
February 1979
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1AA601 CA-11 (FY-78)
11. CONTRACT/GRANT NO.
68-02-2497
13. TYPE OF REPORT AND PERIOD COVERED
Interim 3/77 - 10/78
14. SPONSORING AGENCY CODE
EPA/ 600/09
115. SUPPLEMENTARY NOTES
116. ABSTRACT
Analytical procedures are described that may be used to assess motor
vehicle emission rates of several unregulated pollutants including
aldehydes, organic amines, sulfur dioxide, nitrous oxide, several individual
hydrocarbons including benzene, hydrogen sulfide, total cyanide, organic
sulfides, nickel carbonyl, ammonia, sulfate, and N-nitrosodimethylamine
(sampling conditions only). A series of validation experiments involving
motor vehicle exhaust with injects of known quantities of the compounds
of interest and the Constant Volume Sampling system commonly used in
emissions certification are described for several of the analytical
procedures. The Clean Air Act as amended August 1977 requires in section
202(a) 4 that unregulated pollutants emitted from motor vehicles be
measured to assure that no unreasonable risk to public health and
welfare exists.
17 KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
* Air pollution
Motor vehicles
* Exhaust emissions
* Chemical analysis
* Reviews
1 i
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b.lDENTIFIERS/OPEN ENDED TERMS
19 SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. cos AT I Field/Group
13B
13F
21 B
07D
05B
495
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
481
------- |