&EPA
United States
Environmental Protection
Agency
Environmental Sciences Research EPA-600 2-80-068
Laboratory April 1980
Reseaich Triangle Park NC 27711
Research and Development
Analytical
Procedures for
Characterizing
Unregulated
Emissions from
Vehicles Usi xg
Middle-Distillate
Fuels
Interim Report
-------
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-600/2-80-068
April 1980
ANALYTICAL PROCEDURES FOR CHARACTERIZING UNREGULATED
EMISSIONS FROM VEHICLES USING MIDDLE-DISTILLATE FUELS
Interim Report
by
Lawrence Smith
Mary Ann Parness
E. Robert Fanick
Harry E. Dietzmann
Southwest Research Institute
San Antonio, Texas 78284
Contract No. 68-02-248?-
Project Officer
Ronald L. Bradow
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 endorse-
ment or recommendation for use.
-------
FOREWARD
The Clean Air Act as amended in 1977 requires manufacturers of
automobiles to certify that the emissions of automobiles represent no
unreasonable risk to the public health and safety, EPA's role in enforcing
this section of the Act has been to formulate methods and procedures by
which toxic pollutants which might be emitted from various kinds of
automobile engines could be measured and assessed. The Environmental
Sciences Research Laboratory contributes to this overall agency effort
through programs engaged in
i-
i studies to identify and measure toxic pollutants in source
emissions and in the ambient air.
development of methods and procedures to measure air pollutants
development of modeling procedures which permit prediction
of ambient air quality impacts from source emissions data.
This report is the second of two similar documents relating the
development of analytical methods for measuring trace toxic pollutants in
mobile source exhaust gas. The first of these reports provided fully tested
procedures for 10 toxic gases in the exhaust of gasoline engines. The
current report deals with many of the same compounds and methods now
fully qualified for use with distillate-fueled engines such as diesel
or gas-turbine powerplants. It is intended that this report serve as a
working guide to the automotive industry and to a variety of government
and academic research institutions, providing well-tested analytical
methods for studying hazardous pollutant emissions from automotive
powerplants.
A.H. Ellison
Director
Environmental Sciences Research Laboratory
-------
ABSTRACT
This research program was initiated with the objective of developing,
codifying and testing a group of chemical analytical methods for measuring
toxic compounds in the exhaust of distillate-fueled engines (i.e. diesel,
gas turbine, Stirling, or Rankin cycle powerplants). It is a part of a
larger effort to characterize these components from a number of prototype
powerplants and, thus, represents a logical first step in the process.
Methods of collection and analysis for aldehydes and ketones, for* hydro-
gen cyanide and cyanogen, for hydrogen sulfide, carbonyl sulfide and
organic sulfides, for ammonia and amines, for nitrous oxide, sulfur
dioxide, individual hydrocarbons, for soluble sulfate and N-nitrosodi-
methylamine, benzo-a-pyrene, and phenols were studied in detail. Ten
analytical procedures were developed and codified. Interference studies
and proof-tests in diesel engine exhaust were conducted with every
procedure and the results of these experiments are reported in detail.
All of the procedures were found to be suitable for use in exhaust
emissions characterization studies. The sampling parameters were found
to be adequate for the collection of trace levels of exhaust components
using standard CVS sampling techniques. Interferences were, in general,
minimal although there were two significant problem areas. Phthalate
ester interferes with crotorialdehyde determinations and this contaminant
must be avoided in the procedure. In the hydrogen sulfide method, S02
decreases the apparent sulfide, and its presence must be corrected for.
While other interferences were noted, all could be avoided with the
appropriate precautions noted in the final procedure.
Qualification tests were conducted by introducing known quantities of
these pollutants into the exhaust of a diesel engine operating on a
standard emissions test CVS tunnel. The results of these experiments
indicated completely quantitative recovery for aldehydes and ketones,
S02, nitrous oxide, total cyanide and phenols. Hydrogen sulfide is lost
to the extent of ten percent at normal exhaust levels. Amines, ammonia
and organic sulfides can be lost in sampling in significant amounts in
the CVS apparatus. These losses must be taken into account when calcu-
lating exhaust contributions.
IV
-------
CONTENTS
Foreword xii
Abstract iv
Figures vi
Tables X
1. Introduction 1
2. Aldehyde and Ketone Procedure 3
3. Total Cyanide Procedure 20
4. Individual Hydrocarbon Procedure 43
5. Organic Amine Procedure 53
6. Sulfur Dioxide Procedure 74
7. Nitrous Oxide Procedure 89
8. Hydrogen Sulfide Procedure 98
9. Ammonia Procedure 112
10. Organic Sulfide Procedure 127
11. Phenol Procedure 158
12. The Qualification Experiment 178
13. Results and Conclusions 184
References 187
Appendices
A. Aldehyde and Ketone Procedure 197
B. Total Cyanide Procedure 227
C. Individual Hydrocarbon Procedure .251
D. Organic Amine Procefure 273
E. Sulfur Dioxide Procedure 294
F. Nitrous Oxide Procedure 317
G. Hydrogen Sulfide Procedure 334
H. Ammonia Procedure 359
I. Organic Sulfide Procedure 382
J. Phenol Procedure 415
K. Sulfate Procedure 440
L. DMNA Samplnig Procedure 460
M. DMNA Analysis Procedure 470
N. BaP Sampling and Analysis 484
-------
FIGURES
Number Page
1 Plot of the formaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 8
2 Plot of the acetaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 9
3 Plot of the acetone-DNPH derivative concentration determined
by prpcedure vs actual concentration 10
4 Plot of the isobutyraldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 11
5 Plot of the methylethylketone-DNPH derivative concentration
determined by procedure vs actual concentration 12
r
6 Plot of the crotonaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 13
7 Plot of the hexanaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 14
8 Plot of benzaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration 15
9 Effect of elapsed time on hydrogen cyanide in clear and dark
bags 23
10 The effect of elapsed time on cyanogen in clear and dark
bags *..... 24
11 The effect of elapsed time on hydrogen cyanide in a blend of
hydrogen cyanide in clear and dark bags 25
12 The effect of elapsed time on cyanogen in a blend of hydrogen
cyanide and cyanogen in clear and dark bags ...'.' 26
13 The effect of elapsed time on hydrogen cyanide and cyanogen
in a dark bag with humid nitrogen 27
14 Total cyanide calibration curve at low concentrations
(0-2 ppm) 33
15 Total cyanide calibration curve at low concentrations
(0-10 ppm) 34
VI
-------
FIGURES (Cont'd)
Number Page
16 The effect of elapsed time on sample development 36
17 Time-sample decay curve 38
18 Time-sample decay curve (exhaust only) 48
19 Time-sample decay curve (standard only) 49
20 Time-sample decay curve (exhaust + standard) 50
21 GC peak areas of pentafluorobenzoyl amine derivatives vs
time 59
22 Linearity of monomethylamine GC response (plot on log-log
scale) 63
23 Linearity of dimethylamine GC response (plot on log-log
scale) 64
24 Linearity of trimethylamine GC response (plot on log-log
scale) 65
25 Linearity of monoethylamine GC response (plot on log-log
scale) 66
26 Linearity of diethylamine GC response (plot on log-log
scale) 67
27 Linearity of triethylamine GC response (plot on log-log
scale) 68
28 SC>2 calibration curve 84
29 Detector linearity curve 91
30 Sample decay curve (short term) 94
31 Sample decay curve (long term) 95
32 Time-Light exposure study (low concentration) 101
33 Time-Light exposure study (high concentration) 102
34 Beer's Law plot for methylene blue 103
35 Ammonia calibration curve 120
vn
-------
FIGURES (Cont'd)
Number Page
36 Proposed GC flow schematic for analysis of organic sulfides
(Step 1) 129
37 Proposed GC flow schematic for analysis of organic sulfides
(Step 2) 130
38 Proposed GC flow schematic for analysis of organic sulfides
(Step 3). . . / 131
39 Gas chromatograph separation of several organic sulfides in
prepared blend 133
40 Cold trap experiment flow schematic 135
41 Typical gas chromatograph trace of organic sulfides 141
42 Typical GC separation of organic sulfides on acetone-washed
Porapak QS column 143
43 Typical organic sulfide separation with Tenax-GC column .... 144
44 Organic sulfide permeation blend with secondary dilution,
near detection limit of GC FPD system 145
45 Organic sulfide permeation blend with secondary dilution,
concentrated on Tenax-GC trap and thermally desorbed into
GC FPD system 146
46 Carbonyl sulfide linearity plot 151
47 Methyl sulfide linearity plot 152
48 Ethyl sulfide linearity plot. 153
49 Methyl disulfide linearity plot 154
50 Linearity of o-chlorophenol GC response 168
51 Linearity of phenol GC response 169
52 Linearity of, salicylaldehyde GC response 170
53 Linearity of m-cresol and p-cresol GC response 171
54 Linearity of p-ethylphenol, 2-isopropylphenol, 2,3-xylenol,
3,5-xylenol and 2,4,6-trimethylphenol GC response 172
viii
-------
FIGURES (Cont'd)
Number Page
55 Linearity of 2,3,5-trimethylphenol GC response 173
56 Linearity of 2,3,5 ,6-tetramethyIphenol GC response 174
57 Dilution tunnel-CVS system used in qualification experiments. . 179
58 Apparatus for injection of pollutant into dilution tunnel
without exhaust 180
59 Apparatus for injection of pollutant into dilution tunnel
with exhaust 181
60 Modified mist generator 182
IX
-------
TABLES
Number Page
1 Physical Properties of the Aldehydes and Ketones (4,5) 4
2 Injection Repeatability 7
3 Multiple Extractions of DNPH Solutions (All units are mg DNPH
derivative/mS, of toluene) 16
4 Percent Recovery of Propionaldehyde 18
5 Experiments Conducted for HCN and C2N2 Bag Stability 22
6 The Effect of Stopper Tip and Absorbing Reagent Concentration
on Collection Efficiency at Room Temperature 29
7 The Effect of Absorbing Reagent Temperature on HCN Collection
Efficiency 31
8 Calibration Curve Linearity at Several Cyanide
Concentrations 32
9 Sample Injection Repeatability for Two Cyanide
Concentrations 35
10 Total Cyanide Gaseous Recovery by Direct CVS Injection 39
11 Total Cyanide Recovery from Dilute Exhaust Without Filter or
With Non-Heated Filter 39
12 Total Cyanide Recovery from Dilute Exhaust With Heater
Filter 40
13 Important Facts on Individual Hydrocarbons 45
14 Injection Repeatability on Two Separate Occasions 46
15 List of Individual Organic Amines Included in the Emissions
Characterization Inventory 54
16 Mixing Procedure for Preparation of Pentafluorabenzoylamine
Derivatives 58
17 Injection Repeatability Experiments 62
18 Organic Amine Recovery from the CVS Dilution Tunnel Only. ... 70
X
-------
TABLES (Cont'd)
Number Page
19 Organic Amine Recovery from the CVS Dilution Tunnel With
Exhaust 72
20 Interferences to SC>2 Analysis 79
21 Sulfate Standard Stability 79
22 S02 Collection Efficiency as a Function of Flowrate and
Temperature 80
23 Injection Repeatability for Ion Chromatograph 82
24 Calibration Curve for Sulfur Dioxide 83
25 Sulfur Dioxide Recovery from CVS-Tunnel Injection 86
26 Sulfur Dioxide Recovery from Dilute Exhaust by CVS-Tunnel
Injection During Hot FTP Driving Cycle 87
27 Injection Repeatability Over the Range of Detector Linearity. . 92
28 Nitrous Oxide Qualification Experiments - No Vehicle 96
29 Nitrous Oxide Qualification Experiment With Vehicle Exhaust . . 96
30 The Effect of Sample Flow Rate and Absorbing Reagent
Temperature on the Collection Efficiency 104
31 The Effect of Individual Exhaust Components on the
Development of Methylene Blue 105
32 The Effect of Anions on the Development of Methylene Blue . . . 106
33 The Effect of Sulfur Dioxide Interference on the Development
of Methylene Blue 107
34 Hydrogen Sulfide Recovery - No Exhaust Present 108
35 Effect of Ferric Ion Solution on Hydrogen Sulfide Recovery
from Dilute Exhaust 109
36 NH, Collection Efficiency as a Function of Flowrate and
Temperature 115
37 Injection Repeatability for Ion Chromatograph 117
38 Repeatability of Ammonia Standard 116
XI
-------
TABLES (Cont'd)
Number Page
39 Calibration Curve for Ammonia 119
40 Sample and Standard Stability as a Function of Time 121
41 Ammonia Recovery from CVS-Tunnel, No Exhaust 122
42 Ammonia Recovery from Dilute Exhaust (No Heated Filter) .... 124
43 Ammonia Recovery from Dilute Exhaust, Heated Filter 125
44 List of Sulfur Compounds Included in the Analysis of Organic
Sulfides 127
45 List of Chemical and Physical Characteristics of Various
Sulfur Compounds Potentially Present in Automotive Exhaust. . 134
46 The Effect of Cold Trapping at -78°C on Carbonyl Sulfide and
Methyl Sulfide at Various Concentrations, Flow Rates and
Trap Sizes 136
47 The Effect of Cold Trapping at -196°C on Carbonyl Sulfide
and Methyl Sulfide with Various Trap Sizes 138
48 The Efficiency of Various Materials Trapping Sulfides at
Several Temperatures 139
49 Injection Repeatability for the Organic Sulfides 148
50 Trap Repeatability for Organic Sulfide Collection 149
51 Trap-to-Trap Repeatability for Organic Sulfide Collection . . . 149
52 Percent Recoveries of the Organic Sulfides from the CVS
Tunnel Only 156
53 Percent Recoveries of the Organic Sulfides from the CVS
Tunnel and Exhaust 157
54 Physical Properties of Phenols Possible in Exhaust 158
55 Extractions with Methylene Chloride 161
56 Extractions with Ether 162
57 Extraction Efficiency as a Function of pH of Aqueous Solution . 160
58 Effect of Reducing Sample Volume by Kuderna Danish Concentrator
on Phenol Recovery 163
xii
-------
TABLES (Cont'd)
Number Page
59 Effect of Diesel Fuel on Recovery of Phenol 164
60 Interferences to Phenol Recovery or Analysis 165
61 Linearity Ranges of Internal Standard and of Phenols in
Exhaust 167
62 Injection Variability of Phenol 175
63 Baseline Phenol Emission Levels from Mercedes 240D Diesel . . . 175
64 Percent Recoveries from Injection of Phenol into Exhaust of
Mercedes 240D Diesel 176
65 Analytical Procedures of Emissions Characterization 185
xm
-------
SECTION 1
INTRODUCTION
The objective of this project is to evaluate the emissions of regulated
and nonregulated pollutants in the exhaust of vehicles having advanced-con-
cept powerplants. Examples of engine types which are being considered for
testing in this project include gas turbines, Stirling cycle, turbocharged
Diesel, Rankine cycle, stratified charge, and advanced Otto-cycle. The
first phase of this project includes the development of analytical techno-
logy to provide qualitative and quantitative measurements of unregulated
exhaust products of the engines to be tested. This report is a summary of
the results of this phase of the project.
Candidate analytical procedures were selected for each of the following
compounds or groups of compounds.
aldehydes and ketones nitrous oxide
hydrogen cyanide + cyanogen sulfur dioxide
hydrogen sulfide individual hydrocarbons
organic sulfides + carbonyl sulfide phenols
ammonia N-nitrosodimethylamine
organic amines benzo-a-pyrene
soluble sulfate
The procedures selected represent an assessment of the optimum proce-
dures available at the time of this report and with the approval of the
project officer will be used to measure the appropriate unregualted pollu-
tants. Reviews of the literature, procedural development work, validation
experiments, and qualification experiments are discussed for ten of these
analytical procedures.
These ten analytical procedures are listed in the following paragraphs
along with the appropriate section of the report in which they are discussed
as well as a brief description of the procedure.
-------
Aldehydes and Ketones (Section 2) - The collection of aldehdyes
(formaldehyde, acetaldehyde, isobutyraldehyde and hexanaldelyde) and ketones
(acetone and methylethyIketone) is accomplished by bubbling CVS diluted
exhaust through glass impingers containing 2,4-dinitrophenylhydrazine (DNPH)
in dilute hydrochloric acid. The aldehydes and ketones (also knwon as
carbonyl compounds) react with the DNPH to form their respective phenyl-
hydrazone derivatives. These derivatives are 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 then the pentane is evaporated in the vacuum oven. The remain-
ing 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 dissolved extract
is injected into a gas chromatograph and analyzed using a flame ionization
detector.
Total Cyanide (Hydrogen Cyanide plus Cyanogen) (Section 3) - The
collection of total cyanide is accomplished by bubbling CVS diluted exhaust
through glass impingers containing a 1.0 N potassium hydroxide absorbing
solution. This solution is maintained at ice bath temperature. An aliquot
of the absorbing reagent is then treated with KH2P(>4 and Chloramine-T. A
portion of the resulting cyanogen chloride is injected into a gas chromato-
graph equipped with an electron capture detector (BCD). External CN~ stan-
datds are used to quantify the results.
Individual Hydrocarbons (Section 4) - For measurement of selected
individual hydrocarbons, methane (CH4) , ethane ^2^), ethylene (C2H4) ,
acetylene (C2H2), propane (C^EQ), propylene (C-fl^), benzene (CgHg) , and
toluene (CjUg), a sample of CVS diluted exhaust is collected in a Tedlar
bag. This bagged sample is then analyzed for individual hydrocarbons using
a gas chromatographic system containing four separate columns and a flame
ionization detector. The peak areas are compared to an external calibration
blend and the individual hydrocarbon concentrations are obtained using a
Hewlett-Packard 3354 computer system.
Organic Amines (Section 5) - The collection of organic amines (mono-
methylamine, monoethylamine and dimethylamine, trimethylamine, diethylamine,
and triethylamine) is accomplished by bubbling CVS diluted exhaust through
glass impingers containing dilute sulfuric acid. The amines are complexed
by the acid to form stable sulfate salts which remain in solution. A
portion of this solution is then 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 quantify the
results.
Sulfur Dioxide (Section 6) - The concentration of sulfur dioxide in
dilute exhaust is determined as sulfate using a ion chromatograph. Sulfur
dioxide is collected and converted to sulfate by bubbling dilute exhaust
through two glass impingers containing a 3 percent hydrogen peroxide ab-
sorbing solution. The samples are analyzed on the ion chromatograph and
compared to standards of known sulfate concentrations.
-------
Nitrous Oxide (Section 7) - For measurement of nitrous oxide, a sample
of the CVS diluted exhaust is collected in a Tedlar bag. This bagged sample
is then analyzed for nitrous oxide using a gas chromatograph equipped with
an electron capture detector. Calibration blends are used to quantify the
results. Gas chromatograph peak areas are obtained using a Hewlett-Packard
3354 computer system.
Hydrogen Sulfide (Section 8) - The collection of hydrogen sulfide is
accomplished by bubbling CVS diluted exhaust through glass impingers con-
taining a buffered zinc acetate solution which traps the sulfide ion as
zinc sulfide. The absorbing solution is then treated with N,N-dimethyl-
paraphenylene diamine sulfate and ferric ammonium 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 nm in a 1-cm or 4-cm pathlength
cell depending upon the concentration.
Ammonia (Section 9) - Ammonia in CVS diluted automotive exhaust is
measured in the protonated form, NH4+, after collection in dilute 112804.
The acidification is carried out in a glass impinger maintained at ice bath
temperature. A sample from the impinger is analyzed for ammonia in an
Ion Chromatograph and the concentration in the exhaust is calculated by
comparison to an ammonium sulfate standard solution.
Organic Sulfides (Section 10) - The collection of carbonyl sulfide (COS)
and the organic sulfides, methyl sulfide (dimethylsulfide, (CH3)2S), ethyl
sulfide (diethylsulfide, (C2H5)2S) and methyl disulfide (dimethyldisulfide,
(^3)282) i is accomplished by passing CVS diluted exhaust through Tenax
GC traps at -76°C. At this temperature the traps remove the organic sul-
fides from the dilute exhaust. The organic sulfides are thermally desorbed
from the traps into a gas chromatograph sampling system and injected into
a gas chromatograph equipped with a flame photometric detector for analysis.
External organic sulfide standards generated from permeation tubes are used
to quantify the results.
Phenols (Section 11) - The collection of phenols (phenol; salicyl-
aldehyde; m-cresol and p-cresol; p-ethylphenol, 2-isopropylphenol, 2,3-
xylenol, 3,5-xylenol and 2,4,6,-trimethylphenol; 2,3,5,-trimethylphenol;
and 2,3,5,6,-tetramethylphenol) is accomplished by bubbling CVS diluted
exhaust through two Greenburg-Smith impingers containing 200 m& of 1 N KOH.
The phenols react with the KOH and remain in solution. The contents of each
impinger are acidified and extracted with ethyl ether. The samples are
partially concentrated, combined and then further concentrated to about 1 m£.
An internal standard is added and the volume is adjusted to 2 mil. The final
sample is analyzed by the use of a gas chromatograph and concentrations of
individual phenols are determined by comparison to external and internal
standards.
These ten analytical procedures underwent a series of validation and
qualification experiments. The validation experiments were carried out to
determine if the sampling and instrument parameters were appropriate for the
-------
quantitative analysis of dilute exhaust. The qualification experiments were
carried out to determine if the compounds of interest could be quantitatively
recovered from the Constant Volume Sampler (CVS)-dilution tunnel with and
without the presence of exhaust in the tunnel.
Validation experiments included checks for sample stability, sample
collection efficiency, detector linearity, interferences, extraction effi-
ciency and repeatability, and analysis repeatability.
Sample stability checks were performed using repeated analyses of the
same sample at intervals over a specified period of time and comparing the
results to the initial analysis. Aldehydes and ketones (after extraction),
total cyanide, individual hydrocarbons, organic amines, sulfur dioxide,
nitrous oxide, ammonia, and phenols (after extraction) were found to be
stable for several days. The organic sulfides and hydrogen sulfide samples
were found to be stable for approximately one day.
Sample collection efficiency experiments were performed by passing a
known concentrations of sample through a series of impingers or traps and
analyzing each impinger or trap individually for the compound of interest.
All the procedures discussed in this report have a collection efficiency of
98% or better. Detector linearity experiments were performed by preparing
several samples of various known concentrations and plotting resulting
peak areas (or heights) versus the concentrations. All instruments demon-
strated 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 flame photometric detector. The sample flow rate
can be lowered to prevent overloading the collecting Tenax trap.
To determine the interferences for each procedure, known exhaust com-
ponents were introduced into the sample to determine their effect on the
resultant measurements. 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 and benzalde-
hyde. In the hydrogen sulfide procedure, sulfur dioxide decreases the
apparent hydrogen sulfide concentration, and its presence or absence must
be recorded. Thiophene and ethyl sulfide can not be effectively separated
with the normal gas chromatographic operating conditions and therefore,
thiophene must be included as a possible source of error in the analysis
for ethyl sulfide. The other procedures have interference that can be
avoided if care is taken.
To determine extraction efficiency and repeatability for the aldehyde
and ketone and the phenol procedures, several samples of known concentra-
tions were prepared and a number of analyses were performed. The extrac-
tion efficiency is approximately 100 percent for the aldehyde and ketone
procedure, however the overall repeatability varies up to 15 percent at
concentrations of 0.2-2.0 mg derivative per m£ toluene. The results of
extraction repeatability experiments for aldehyde and ketone DNPH derivative
-------
concentrations below 0.025 mg DNPH derivative per m£ toluene indicate that
the variability in the extraction process can be very significant (i.e.,
0.94 percent for benzaldehyde at 0.016 mg/m£). This variability needs to
be taken into account when evaluating data obtained using this procedure.
The extraction efficiency for the phenol procedure is only about 68 percent
due to unavoidable problems in the drying down process. This value is
repeatable if the extraction procedure is followed closely. These losses
must be taken into account when analyzing data obtained from the phenol
procedure.
To determine analysis repeatability, several samples of known concen-
trations were prepared and a number of complete analyses were performed at
each concentration. The results of these tests were then compared to
determine analyses repeatability. The test-to-test repeatabilities are
documented for all procedures in this report. In most cases, repeatability
is difficult to obtain at the lower concentrations, while the repeatability
at high concentrations is easily obtained.
The qualification experiments were performed to determine if the com-
pounds of interest could travel the length of the dilution tunnel in the
presence of dilute exhaust without significant loss by reaction with exhaust
or the tunnel itself. The compounds were introduced at the same point at
which the exhaust enters the tunnel and were sampled at the normal sampling
point (see Section 12).
Qualification experiments were carried out on the aldehyde and ketone,
organic amine, sulfur dioxide, nitrous oxide, hydrogen sulfide, total
cyanide, organic sulfide, ammonia, and phenol procedures to determine the
recovery of known amounts of each pollutant from the CVS tunnel with and
without exhaust (phenols CVS dilution tunnel with exhaust only). Aldehydes
and ketones,: sulfur dioxide, nitrous oxide, total cyanide and phenols can
be recovered quantitatively from the CVS dilution tunnel with and without
(not done for phenols) exhaust. There is a 10 percent loss of hydrogen
sulfide with and without exhaust present. The organic amines, ammonia,
and the organic sulfides experience significant losses in the CVS dilution
tunnel with and without exhaust present.
Despite the fact that the analytical procedures for the organic amines
and the organic sulfides have procedural detection limits of 2 and 0.2 ppb
respectively, the losses in the dilution tunnel could prevent the detection
of organic amines at levels lower than 20 ppb and the detection of organic
sulfides at levels lower than 10 ppb in dilute exhaust. At ammonia levels
of 5-10 ppm there is a 25 percent loss of ammonia to the dilution tunnel and
an additional fifteen percent loss to exhaust.
The procedures discussed in this report have been found to be the
optimum procedures at the time of .this report for collecting and analyzing
dilute exhaust samples and are recommended for use in this project.
-------
A finalized copy of the analytical procedures discussed in Section 2-11,
the BCA sulfate procedure, and DMNA procedure, sampling conditions for DMNA,
and an outline for BaP collection and analysis are included as an appendix.
The literature search, procedural development work, and validation experi-
ments for some of the compounds were carried out under another EPA Contract,
68-02-2497 (1). The procedures discussed in this report were developed for
the measurement of pollutants in dilute exhaust. The use of these procedures
for the measurement of pollutants in raw exhaust is not recommended without
additional validation and qualification work to document the acceptability
of the procedures.
-------
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 derivatives, and the melting points of the
2,4 dinitrophenylhydrazone derivatives 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 (2). 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 methylethylketone.
PROCEDURAL DEVELOPEMENT
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 (3). 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 aldehyde 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 pen-
tane 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 in to a gas chromatograph
and analysed 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 OP THE ALDEHYDES AND KETONES (4,5)
Molecular
Weight
CO
Aldehyde or Ketone
Formaldehyde
Acetaldehyde
Acetone
Propionaldehyde
Acrolein
Isobutyraldehyde
Methylethylketone
Crotonaldehyde
Hexanaldehyde
Benzaldehyde
ICU Name
Methanal
Ethanal
2-Propanone
Propanal
Propenal
2-Methylpropanal
2-Butanone
transZ-Butenal
Hexanal
Benzenecarbonal
Chemical
Formvila
CH2O
CH3CHO
CH3COCH3
CH3CH2CIIO
CII2 :CHCHO
CH3CH(CH3)CHO
CH3COCH2CH3
CH3CH:CHCHO
CH3(CII2>4CHO
C6H5CHO
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
0.805
0.850
0.814
1.042
DNPH
Derivative
210.15
224.19
238.21
238.21
236.20
252.23
252.23
250.21
279.28
286.25
Melting
Point Der
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 experi-
ments 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/HC1 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 effi-
ciency 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 com-
bined. Since the analysis of the two impingers combined is less manpower
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 signifi-
cant 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 amounts of sample to
flow through the absorbing reagent without loss in absorbing efficiency or
the physical loss of any absorbing reagent. A heated filter is used to pre-
vent diesel particulate from contaminating the sampling system. The filter
and the line connecting the filter to the dilution tunnel are heated to
375°F to prevent the aldehydes and ketones from being retained on the removed
particulate. A Teflon line connecting the filter to impingers is heated to
175°F 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 inadver-
tant 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 repeata-
bility 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/mil 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, amd 0.002 mg of each
derivative/mi 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, methylethyIketone, and crotonaldehyde give linear plots
throughout tfie region of interest. Formaldehyde, acetaldehyde, isobutyr-
aldehyde and hexanaldehyde give linear plots except at the lower concentra-
tions (<0.02 mg derivative/m£ toluene). Benzaldehyde gives a plot which is
not linear above 2.0 mg/m£ toluene. The benzaledhyde-DNPH derivative is not
soluble in toluene at concentrations greater than 2.0 mg/mA. This fact
should be taken into account if high concentrations of benzaldehyde are ex-
pected (>5 ppm for a 23 minute sampling period at 4 A/minute) .
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, methylethyIketone, 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. Tha 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
10
-------
TABLE 2. INJECTION REPEATABILITY
DNPH Aldehyde
or Ketone
Derivative
Formaldehyde
Acetaldehyde
Acetone
Isobutyraldehyde
Methyl ethy Iketone
Crotonaldehyde
Hexanaldehyde
Renzaldehyde
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
1.3
1.1
1.1
1.8
3.0
6.2
9.6
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 derivative/ml
Standard
Avg. for Std. %
5 Inject. Dev. Dev.
0.022 0.002 9.1
0.024 0.002 8.3
0.027 0.001 3.7
0.020 0.001 5.6
0.024 0.001 4.2
0.015 0.001 6.7
0.015 0.001 6.7
0.025 0.008 32
0.002 mg derivative/ml
Standard
Avg. for Std. %
5 Inject. Ugv. 2.ev_-
0.007 0.0007 JO
0.002 0.0007 35
0.003
0.002
0.001
0.001
0.0007 23
0.0007 35
0.0007 70
o.ooii no
-------
10.0 -
0.001
0.002
0.005 0.01 0.02
Actual concentration
0.05 0.1 0.2 0.5 i.o 2.0
(rag formaldehyde-DNPH derivative/ml toluene)
5.0
10.0
Figure 1. Plot of the formaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
12
-------
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 (rag acetaldehyde-DNPH derivative/ml toluene)
10.0
Figure 2. Plot of the acetaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
13
-------
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 acetone-DNPH derivative/ml toluene)
5.0
I
10.0
Figure 3. Plot of the acetone-DNPH derivative concentration
determined by procedure vs actual concentration.
14
-------
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 isobutyraldehyde-DNPH derivative/ml toluene)
10.0
Figure 4. Plot of the isobutyraldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
15
-------
0.001
0.002 0.005 0.01 0.02 0.05 0.1 0.2 0.5 l.Q 2.0 5.0
Actual concentration (mg methylethylketone-DNPH derivative/ml toluene)
10.0
Figure 5. Plot of the methylethylketone-DNPH derivative concentration
determined by procedure vs actual concentration.
16
-------
§
I
-H
I
-ri
U
10.0
5.0
2.0
1.0
I
IB
I
U
I
01
I
8
8
a
1
0.5
0.2
0.1
0.05
0.02
0.01
C 0.005
o
•H
-P
ID
g
U
0.002 _
0.001
I
I
I
I
I
I
I
I
1
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 derivative/ml toluene)
10.0
Figure 6. Plot of the crotonaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
17
-------
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 hexanaldehyde-DNPH derivative/ml toluene)
10.0
Figure 7. Plot of the hexanaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
18
-------
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 benzaldehyde-DNPH derivative/ml toluene)
5.0 10.0
Figure 8. Plot of benzaldehyde-DNPH derivative concentration
determined by procedure vs actual concentration.
19
-------
TABLE 3. MULTIPLE EXTRACTIONS OF DNPH SOLUTIONS
(All units are mg DNPH derivative/m£ of toluene)
ro
o
Extraction
First
Second
Third
Fourth
Fifth
Sixth
Seventh
Average
Standard
Deviation
Form-
aldehyde
0.018
0.023
0.020
0.009
0.014
0.020
0.014
0.017
±0.005
Acet-
aldehyde
0.011
0.037
0.032
0.013
0.019
0.031
0.027
0.024
±0.010
Acetone
0.005
0.015
0.015
0.005
0.007
0.014
0.043
0.015
±0.013
MEK
0.000
0.003
0.000
0.000
0.000
0.004
0.014
0.003
±0.005
Cron ton-
aldehyde
0.003
0.004
0.003
0.002
0.003
0.005
0.004
0.003
±0.001
Hexan-
aldehyde
0.001
0.002
0.001
0.000
0.002
0.002
0.000
0.001
±0.001
Benz-
aldehyde
0.014
0.027
0.044
0.000
0.014
0.004
0.007
0.016
±0.015
-------
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/m£ toluene (ex-
cept for two acetone DNPH 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
extraction process for concentrations of aldehyde DNPH and ketone DNPH
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 15 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 occassions.
A gas chromatography-mass spectroscopy study was carried out on three samples
obtained from a gasoline powered vehicle. The three samples either contained
abnormally high concentrations of crotonaldehyde-DNPH derivative or benz-
aldehyde-DNPH derivative or both. The results from this study revealed that
neither crotonaldehyde nor benzaldehyde was present in the samples. Further
gas chormatography-mass spectroscopy studies were carried out on two of 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 (orignating 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 Mercedes 240D vehicle.
Hot FTP (23 minute test) driving cycles were followed to generate exhaust
for the vehicle baseline emissions and for the tunnel plus vehicle experi-
ments. Aluminum cylinders containing 1350 and 436 ppm propionaldehyde in
balance nitrogen were used as the source for aldehydes in the experiments.
The cylinders were named using the aldehyde-DNPH procedure. The flow of
propionaldehyde into the tunnel was regulated to give concentrations of 0.5-
2 ppm propionaldehyde in the dilution tunnel. Injections of propionaldehyde
into the tunnel without exhaust gave recoveries that ranged from 85 to 115
21
-------
percent with an average of 102 percent (Table 4). The recovery of propion-
aldehyde in the presence of vehicle exhaust and without a heated filter
ranged from 59 percent to 89 percent with an average of 76 percent (Table 4)
The recovery of propionaldehyde from the dilution tunnel in the presence of
exhaust while using a heated filter ranged from 82 to 120 percent for an ,
average of 99 percent (Table 4). The injections with the vehicle were cor-
rected for the vehicle baseline emission of propionaldehyde. If a heated
line and filter is used to remove particulate and if propionaldehyde is re-
presentative of the aldehydes, then it appears that there is little or no
loss of aldehyde in the dilution tunnel with or without vehicle exhaust.
TABLE 4. PERCENT RECOVERY OF PROPIONALDEHYDE
Tunnel Only
Tunnel + Vehicle
No Heated Filter
Tunnel + Vehicle
Heated Filter
Run
1
2
3
4
5
6
Avg
Recovery %
85
115
99
96
110
106
102 ± 11
Run
1
2
3
4
5
6
Avg
Recovery
83
86
89
76
64
59
76 ± 12
Run
1
2
3
4
5
6
7
8
Avg
Recovery
86
82
120
106
85
107
103
103
99 ± 13
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 Vminute. The procedure has a minimum detection limit
of approximately 5 ppb. This carbonyl concentration in the exhaust gives a
corrsponding concentration of 0.002 mg/m£ in toluene.
22
-------
The accuracy of the procedure in the 0.5-20 ppm concentration range for
the aldehydes and ketones in dilute exhaust is approximately 10-15 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/ltd toluene and
gives a linear response for formaldehyde, acetaldehyde, isobutyraldehyde and
hexanaldehyde DNPH derivative concentrations between 0.02 and 8 mg derivative/
m£ toluene. The benzaldehyde-DNPH derivative gives 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-ethyhexyladipate 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 usually
broad and the benzaldehyde peak, if present, can be observed on top of this
interference. If care is taken, a reliable value can be determined for the
benzaldehyde. 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 contami-
nation after collecting the sample and before analysis.
Propionaldehyde can be recovered quantitatively from the dilution tunnel
with or without diesel exhaust present if a heated filter is used. If pro-
pionaldehyde is representative 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.
23
-------
SECTION 3
TOTAL CYANIDE PROCEDURE
LITERATURE SEARCH
Hydrogen cyanide is a flammable, toxic, and colorless liquid at room
temperature and has 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 dissociates in an aqueous solu-
tion as do the hydrogen halides. Hydrogen cyanide (HCN) has a molecular
weight of 27.03, a boiling point of 24.70°C, and a melting point of -13.42°C.
It is a linear molecule with C-H and C.HN bond distances of 1.06 and 1.15 A,
respectively. It is a weak monoprotic acid with a dissociation constant of
2.1 X 10~9. This highly poisonous compound is a respiratory inhibitor and
irreversibly combines with the iron complex in the blood, stopping the oxi-
dation processes in tissue cells and causing death by asphyxiation. Com-
mercially, 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, and 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 oxaloni-
trile. Pure cyanogen is stable, although the impure gas may polymerize to
paracyanogen between 300° and 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 dis-
proportionation reaction in basic solution:
(CN) 2 + 20H~ —>• CN~ + OCN~ + H2O
Cyanogen (C2N2> has a molecular weight of 52.04, a freezing point of -27.9°C
and a boiling point of -21.17°C. Cyanogen is a symmetrical and linear mole-
cule with a C-C bond distance of 1.37 A and a C=N bond distance of 1.13 A.
Its physiological 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,
any many others. In all cases above, cyanogen is produced from hydrogen
cyanide. Although none of these are exactly applicable for an automotive
24
-------
system, a 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, colorimetry,
specific ion electrode, and gas chromatography. The Liebig determination 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 man-
power 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 (6) 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 (7) was used
with biological samples (blood, urine, and gastric contents). The samples
were collected in sodium of 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 mVmin
were the column conditions. A glass lined injector and interface were also
used to preserve sample integrity.
Bag stability experiments with hydrogen cyanide and cyanogen were con-
ducted 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 is
collected. Clear and aluminum foil tape covered Tedlar plastic bags were used
25
-------
to conduct bag stability experiments. Dark bags (aluminum foil tape covered)
were tested to determine the effect of photochemical decomposition on hydro-
gen cyanide and cyanogen.
A list of bag stability experiments which were conducted is shown in
Table 5. Each bag contained approximately one cubic foot of the dilute gas.
TABLE 5. EXPERIMENTS CONDUCTED FOR HCN AND C-N BAG STABILITY
Compound
HCN
C2N2
HCN & C N
HCN & C2N2
Clear Bags
Nitrogen Air
X X
X X
2
*
Exhaust
X
X
X
Dark
Nitrogen
X
X
X
X
Bags
Air
X
X
X
Exhaust
X
X
X
* Blend of hydrogen cyanide and cyanogen in humid nitrogen.
Experiments were conducted with nitrogen, air, dilute exhaust and humid nitro-
gen. 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 mH 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
9 through 13 show the effect of elapsed time on the stability of hydrogen
cyanide and cyanogen. Figure 9 demonstrates the stability of hydrogen cyanide
in clear and dark bags with a variety of atmospheres. Peak areas for hydro-
gen 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 ± 7 percent and is indicated in all
figures by a dotted line). On the other hand, cyanogen showed a considerable
percent loss in the clear bags with both nitrogen and exhaust (Figure 10) .
In the dark bag, cyanogen remained stable except in the presence of exhaust.
Figures 11 and 12 show the effect of a blend of hydrogen cyanide and
cyanogen in clear and dark bags with the various atmospheres. Again, hydro-
gen 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 similarly 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 13) 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
26
-------
0)
(Jl
c
(0
u
C
0)
O
t-i
0)
CM
is r-
10
Legend
Nitrogen
Air
Exhaust
Nominal Range for Injection
Repeatability .
-5
-10
-15
_L
_L
20 40 60 80
Time, Minutes
Hydrogen cyanide in clear bags
100
120
15
10
cu
0i
O
-P
C
0)
O
^
0)
Hi
Nominal Range for Injection
Repeatability
-5
-10
-15'
n.
_L
J_
20 40 60 80 100
Time, Minutes
Hydrogen cyanide in dark bags
120
Figure 9. Effect of elapsed time on hydrogen
cyanide in clear and dark bags.
27
-------
30
20
10
Legend
Nitrogen
• Air
A Exhaust
Nominal Range for
Injection Repeatability
40 60
Time, Minutes
Cyanogen in clear bags
100
120
„Nominal Range for
Injection Repeatability
-10
-20 -
-30
40 60 80 100
Time, Minutes
Cyanogen in dark bags
120
Figure 10. The effect of elapsed time on
cyanogen in clear and dark bags.
28
-------
30 r-
20 _
10
c
(0
o
-p
c
0)
Q
oJ -10
d.
-20
-30
30
20
d) 10
0
-p
c
o
o
0 -10
-20
-30
Legend
Nitrogen
Air
Exhaust
Nominal Range for
Injection Repeatability
I
I
20
40
80
60
Time/ Minutes
Hydrogen cyanide in clear bags
100
120
Nominal Range for
Injection Repeatability
J_
20
40
80
60
Time, Minutes
Hydrogen cyanide in dark bags
100
120
Figure 11. The effect of elapsed time on hydrogen cyanide in a blend
of hydrogen cyanide and cyanogen in clear and drak bags.
29
-------
a
A
u
•P
0)
u
i-l
0)
c
JS
u
c
0)
o
so r~
20
10
-10
-20
-30
Legend
Nitrogen
Air
Exhaust
Nominal Range for
Injection Repeatability
I
20
30
20
, n
1C
40 60 80
Time, Minutes
Cyanogen in clear bags
100
120
-10
-20
-30
Nominal Range for
Injection Repeatability
T.
x
_L
_L
0 20 40 60 80 100 120
Time, Minutes
Cyanogen in dark bags
Figure 12. The effect of elapsed time on cyanogen in a blend
of hydrogen cyanide and cyanogen in clear and dark bags.
30
-------
CO
20
10
0
- 10
- 20
C
rfl
£
0
-40
0)
Z -50
-------
showed only a slight decrease in concentration.
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.
t
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 0.1 M 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 concen-
tration. Instability of the potentiometrie measurement was observed in all
concentration ranges, especially in the low concentration range. Attempts
to improve the electrode stability and potential drift were unsuccessful.
Efforts using the specific ion electrode were abandoned for another pro-
cedure 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 in-
sensitivity to hydrocarbons. 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" O.D. glass column packed with 100/120 mesh Porapak QS
B. 6' X 1/4" O.D. stainless steel column packed with 50/80 mesh
Porapak Q
C. 6* X 1/4" O.D. 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&
32
-------
head space and a septum cap. A sample development period of 5 minutes was
required. After vigorously shaking the vial for 5 seconds, 100 y£ 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 of this report.
VALIDATION EXPERIMENTS
After selecting an analytical method, validation experiments were con-
ducted to determine detector linearity, detection limits, injection repeat-
ability, stability of reagents and sample, sampling parameters, etc. Once
the validation experiments were complete, the procedure was considered
ready for testing.
Collection parameters were determined with a series of experiments de-
signed to check sample flow rates, absorbing reagent concentration, absorbing
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, the reagent concentration on the collection efficiency.
The results of these experiments are shown in Table 6. A set of three of the
TABLE 6. THE EFFECT OF STOPPER TIP AND ABSORBING
REAGENT CONCENTRATION ON COLLECTION EFFICIENCY AT ROOM TEMPERATURE
yg CN /ft sample
Collection Sample
Device Flow
impinger 1.0
impinger 1.0
impinger 1.0
bubbler 1.0
bubbler 1.0
bubbler 1.0
impinger 4.0
impinger 4.0
impinger 4.0
bubbler 4.0
bubbler 4.0
bubbler 4.0
1.0 N KOH
Run
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
1
2
3
Avg
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
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
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
0.1 N KOH
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
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
33
-------
same type collection devices (impinger or fritted glass tipped bubblers) were
filled with 1.0 N or 0.1 N potassium hydroxide. A hydrogen cyanide calibra-
tion blend that contained a nominal 2 ppm concentration in a balance of nitro-
gen was passed through the absorbing reagent at 1.0 and 4.0 £/min. 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 accoutre the results of the first
set plus the effect of reagent temperature. Five sets of three impingers
filled with 25 mi each of 1.0 N potassium hydroxide absorbing reagent were
used. The sample flow rate was set at 4.0 Jl/min. The first set of impingers
was sampled at ambient room temperature (16-29°C) and a second set of im-
pingers was sampled at ice bath temperatures. The sample collection effici-
ency for the ambient temperature experiments showed a high degree of vari-
ability. 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 7.
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~/m£ ranges. Table 8 and Figures 14 and 15 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 an gas chromatography
technique which does not involve the use of internal standard. To establish
sample injection reproducibility, two nominal cyanide ion concentrations,
2.0 and 0.2 yg/m£ , were used. Five separate samples of each concentration
were developed and injected. The results are shown in Table 9.
Three separate experiments involving the sample storage and sample
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 y£ of the remaining head space was also injected. The
decay of the peak areas for a two-hour period is shown in Figure 16. 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
34
-------
TABLE 7. THE EFFECT OF ABSORBING REAGENT TEMPERATURE
ON HCN COLLECTION EFFICIENCY
Absorbing Reagent
Temperature
Date
10/1.1/77
10/13/77
10/17/77
10/J7/77
10/17/77
GO
cn
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
op
72
61
63.
73
U4
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
A/min
4.0
4.0
4.0
4.0
4.0
Avg
4.0
4.0
4.0
4.0
4.0
Avg
pg CN~/ft3 pg CN~/m3
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
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
0.0
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
1.68
2.07
1.46
1.43
1.13
1.55
1.77
1.67
1.86
1.92
1.65
1.77
-------
TABLE 8. CALIBRATION CURVE LINEARITY AT SEVERAL CYANIDE CONCENTRATIONS
Background
CN Cone .
Test Date yg/m&
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
Q
area
4600
2594
1132
533
0
4102
2448
953
518
0
7816
4023
1824
694
0
6813
3538
2Q27
643
0
36
-------
LEGEND
CO
-•J
10.-
o
o
•s
(V
10/03/77
10/04/77
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Cyanide Ion Concentration, |Jg CN~/ml
1.6
l.H
Figure 14. Total cyanide calibration curve at low concentrations (0-2 ppm).
-------
co
CO
34567
Cyanide Ion Concentration, |ig/CN~/ni£
10
Figure 15. Total cyanide calibration curve at low concentrations (0-10 ppm).
-------
TABLE 9. SAMPLE INJECTION REPEATABILITY FOR
TWO CYANIDE CONCENTRATIONS
Nominal
Sample
1
2
3
4
5
x
sx
Cv
1
2
3
4
5
x
sx
Cv
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
Peak
X32
X32
X32
X32
X32
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
2
6621
7095
6838
6711
6544
6762
216
3
.5
.0
.0
.2
39
-------
0 i-
C
(S
u
C
CL)
U
M
0)
CM
(U
cn
•P
d>
U
M
20
40
60
30 60
Time, Minutes
Sample decay with time
90
120
10
15
20
25 -
30
60
90'
120
Time, Minutes
Five samples with varying development time
Figure 16. The effect of elapsed time on sample development.
40
-------
fifth after 120 minutes. The sample decay, as a function of time,is also
shown in Figure 16. 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 17. As a result,
samples can be stored for a period of several weeks 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.
Cyanate and thiocyanate ions produced a positive interference at both concen-
trations. 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 Mercedes 240D. Hot FTP (23 minute test) driving cycles were followed
to generate exhaust for the vehicle baseline emissions and for the tunnel
(18 inch diameter) injection + vehicle experiments. A cylinder containing
485 ppm hydrogen cyanide in balance nitrogen was used as the source for hydro-
gen cyanide. The flow of cyanide into the tunnel was regulated to give a
concentration of 0.5 to 1 ppm hydrogen cyanide in the dilution tunnel.
The baseline emission rate for the Mercedes 240D was-0.01 ppm. Injec-
tion of hydrogen cyanide into the tunnel without exhaust gave recoveries
that ranged from 82 percent to 108 percent with an average of 98 percent
(Table 10). The recovery of hydrogen cyanide in the presence of vehicle
exhaust without a filter to remove particulate from the sampled exhaust gave
recoveries that ranged from 68 to 84 percent with an average of 76 percent
(Table 11). The recoveries ranged from 75-85 percent (Table 11) when a non-
heated 0.5 y Fluoropore filter was used to remove particulate from the sampled
41
-------
O.SO
0.70
Q.60
I
a
c
0 0.50
§ 0.40
0. 30;
0)
T3
•H
C'
u
0.20.
0.10
0.
I
50
10
20
30
40
Time, Days
Figure 17. Time-sample decay curve.
42
-------
TABLE 10. TOTAL CYANIDE GASEOUS RECOVERY
BY DIRECT CVS INJECTION
Actual ppm
Injected
485
485
485
485
485
485
485
485
.Nominal Flow
Rate, ft /min
HCN Blend
0.35
0.35
0.35
0.35
0.35
0.35
0.35
0.35
CVS
270
270
270
270
270
270
270
270
Run
1
1
2
2
3
3
3
Sample
1
2
3
1
2
1
2
3
Calculated
ppm HCN
dilute
0.60
0.60
0.60
0.61
0.61
0.60
0.60
0.60
Observed
ppm*
0.60
0.61
0.62
0.60
0.66
0.57
0.58
0.49
Average
Percent
Recovery
HCN
100
102
103
98
108
95
97
82
98 ± !
Background subtracted from observed concentration (0.03 ppm)
TABLE 11. TOTAL CYANIDE RECOVERY FROM DILUTE EXHAUST
WITHOUT FILTER OR WITH NON-HEATED FILTER
Total Cyanide Cone. as HCN, ppm
Actual ppm
Injected Run
Calculated
Sample Amount Observed
non-filtered
485
485
485
485
485
1
1
2
3
3
1
2
1
1
2
0.60
0.60
0.61
0.62
0.62
0.42
0.47
0.50
0.53
0.44
Corrected*
0.41
0.46
0.49
0.52
0.43
Percent
Recovery
68
77
80
84
69
Average 76 ± 7%
filtered/non-heated**
485
485
485
4
5
6
1
1
1
0.60
0.61
0.62
0.46
0.50
0.54
0.45
0.49
0.53
75**
80**
85**
Average 80 - 5%
Concentration corrected for background levels and
vehicle baseline emissions
**
Particulate removed from exhaust stream with
non-heated 0.5 y Fluoropore filter
-------
exhaust stream. Higher recoveries of 88-113 percent (average 99 percent)
were obtained when a heated glass fiber filter (375°F) was used to remove
particulate from the sampled exhaust stream. (Table 12).
TABLE 12. TOTAL CYANIDE RECOVERY FROM DILUTE EXHAUST
WITH HEATED FILTER
Actual ppm
Injected Run
Total Cyanide Cone, as HCN, ppm
Calculated
485
485
485
485
485
485
485
485
1
1
1
1
1
1
1
1
Sample
1
2
3
1
2
3
1
2
Amount
Observed Corrected*
0.54
0.54
0.54
0.55
0.55
0.55
0.57
0.57
0.61
0.54
0.64
0.54
0.54
0.61
0.53
0.58
Percent
Recovery
0.58 107
0.51 94
0.61 113
0.51 93
0.51 93
0.58 105
0.50 88
0.55 96
Average 99 ± 9%
A
Concentration corrected for background levels and for
vehicle baseline emissions
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 con-
centration 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.
Sulfate, 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 existence and interference of these ions.
44
-------
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 than hydrogen cyanide in potassium hydroxide. This
difference in trapping efficiency was discovered while naming high concentra-
tion 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 m£ of 1.0 N potassium hydroxide absorbing reagent.
2. Absorbing reagent held at ice bath temperature (0°C-5°C).
3. Sample flow rate of 4.0 £/min.
4. Impingers rather than fritted glass bubblers.
5. Two impingers in series.
These parameters were sufficient to collect a sample from dilute exhaust
within the detection limits of the procedure.
The measurement of hydrogen cyanide in the presence of cyanogen is dif-
ficult 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 m£ of the various solutions is added to
this vial. When the vial is tightly capped, a 1 m£ 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
45
-------
for only a short time after complete development. The sample may be stored
undeveloped in the potassium hydroxide absorbing reagent for at least three
weeks.
When a heated filter is used to remove particulate from the sampled ex-
haust stream, 99 percent of the hydrogen cyanide injected into the dilution
tunnel can be recovered. When a non-heated filter or no filter is used, only
76-80 percent of the cyanide can be recovered. From these experiments, it is
recommended that a heated filter be used in the sampling system to increase
recoveries and to prevent contamination of the sampling system.
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.
46
-------
SECTION 4
INDIVIDUAL HYDROCARBON PROCEDURE
LITERATURE SEARCH
The eight individual hydrocarbons (methane, ethane, ehtylene, 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 techniques.
Hydrocarbons are of interest as exhaust components because of their po-
tential for photochemical smog formation. Hydrocarbons are placed into four
classes according to their participation in atmospheric reactions. Methane,
ethane , acetylene , propane , and benzene are placed in Class I , the non-
reactive category. The Class II reactive category includes the C^, 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
0 + H C =CH2 •*• H C = 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 (8) proposed a three-column system capable of
analyzing at least 22 hydrocarbons . Two packed columns were required to re-
solve the GI and C2 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.
Papa et al (9) presented a procedure for the analysis of Cj_ through
hydrocarbons in automotive exhaust. This dual column system consisted of a
packed column with a mixture of stationary phases for the resolution of C-j_
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 (10) proposed a simple analytical system for the
determination of hydrocarbons according to their potential for photochemical
47
-------
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
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 (11) . 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 13 lists the compounds
of interest, along with chemical formulas, boiling and melting points, syn-
onyms, and molecular weights.
PROCEDURAL DEVELOPMENT
The gas chromatogarph procedure that will be used for the determination
of the individual hydrocarbons is similar to the procedure used by Black et
al (11) and consists of a four column system that is capable of resolving
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.
Columnlii consists of 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 mer-
cury sulfate (HgSO4) and 20 percent sulfuric acid (H2SO4) on Chromosorb W.
Columns II, III, and IV are used isothermally and Column I undergoes a tem-
perature 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 resolve benzene and toluene from the other aromatics, paraffins,
olefins, acetylenes, and oxygenated hydrocarbons. Three timers, four sole-
noid valves, and five six-port gas sampling valves are required to accomplish
the complicated sample flow through the columns. When exhaust from diesel
powered vehicles is analyzed, higher molecular weight hydrocarbons have been
found to interfere with the analysis. The compounds can be effectively re-
moved by simply passing the exhaust sample through an ice trap before it
enters the analytical system. The actual analytical procedure is included
as an attachment to this report.
VALIDATION
This gas chromatographic procedure has been used with much success on
a variety of projects. The validation of this procedure consists of the
injection repeatability for all eight components of the calibration blend
and bag sample stability. All other parameters were determined from previous
experience with this analytical procedure.
The injection repeatability for the individual hydrocarbon procedure
was conducted on two separate occasions. Table 14 shows the data accumulated
on each occasion. The injection repeatability for the two 10 mil sample loops
is not greater than ±2 percent.
48
-------
TABLE 13. IMPORTANT FACTS ON INDIVIDUAL HYDROCARBONS
vo
Compound
Methane
Ethylene
Ethane
Acetylene
Propane
Propylene
Benzene
Toluene
Formula
CH4
C2H4
C2H6
C2H2
C3H8
C3H6
C6H6
"7 ft
Molecular
Weight
16.04
28.05
30.07
26.04
44.11
42.08
78.12
92.15
Melting
Point
-182.48
-169.15
-183.3
- 80.3
-189.69
-185.25
5.5
- 95
Boiling
Point
-164
-103.71
- 88.63
- 75
- 42. Q7
- 47.4
80.1
110.6
Synonyms
marsh gas, methyl hydride
ethene, elayl, olefiant gas
bimethyl, dimethyl, methylmethane ,
ethyl hydride
ethyne , ethine
dimethylmethane , propyl hydride
propene, methylethylene, methyl-
ethene
benzol, phene, cyclohexatriene
methylbenzene , phenylmethane ,
toluol, methacide
-------
TABLE 14. INJECTION REPEATABILITY ON TWO SEPARATE OCCASIONS
Indivi dual 1'uak Ai'tJJ, Keldti ve Counts (Teat 1)
Injection
1
2
3
4
5
6
7
8
9
10
Average
Standard
Deviation ,
sx
Coefficient
of variation,
cv
Methane
7283
7527
7792
7566
7585
7387
7700
7455
7680
7618
7559.30
152.98
2.02
Ethylene
10097
10467
10621
10479
10495
10300
10575
10307
10493
10481
10431.50
154.55
1.48
Ktluiuu A,:uLy luhu
•J.VJ/ 12034
•)d7, l.:'.,-n
lOu'j'J i*!HOJ
<)ti'*2 I.:.>JB
y(joo 12u7!i
9b9o 1239B
9994 12753
9724 12460
9078 I2u24
98B5 U714
9847.20 12563.30
13c.9U J24.i)5
1 . i'i 1 .79
Individual Iv-ak AI..J, KL.-IU! ivd
Injection
1
2
3
4
5
6
Average
Standard
Deviation,
Coefficient
of variation,
C
Methane
8593
8633
8602
8640
8702
8561
8622
48.59
0.56
Ethylene
11893
11937
11993
11885
12022
11900
11938
57.20
0.48
KLl^,,,, A^tyl^ut,
luO/i 14U03
11011 14992
IKK,/ 151J9
10-J/7 151 10
1111 I 15J71
109JI. 15252
11015 15121
Jr, i. 7(1 200.07
,). ,.-. 1 . >_•
Propane t'ropylune Bunzi-jit
14H09 I!)5)h9 Uny/
l'KJ')2 I'j724 1 101 'j
l'iV.,2 I£)lb2 lib//
I.1H25 Ib'JKl 1J-J7/
1'jOlfc Ibhyy 1 1H.:4
14925 15441 14041
15147 15931 14042
14702 15629 14uhb
14933 15842 14049
14996 15S57 13943
14979.70 15795.40 13933.10
187.42 206.83 125. 3i
1.25 1.31 1..-.0
Count.a (Tust 2)
l-..,L.anu 1'iopyl «,...• 1,,,.:-.,,,,
15705 178b7 I046b
15912 18020 l.>t,l)4
15722 17818 lf-,553
15775 17920 1.^409
15825 17850 li,2n/
15879 17955 164J3
15803 17906 Ifc455
83.77 75.10 llil.jl
0.53 ».4:! ,,..'.
't'oiuent; '
1534H
IV, 7u
15406
1047b
155/u
15 /BO
15799
15459
15783
15584.33
166.20
1 .07
Tuliumu
IH546
18537
18735
1B479
18530
18562
18565
07.95
.1.47
50
-------
The bag sample stability experiment was conducted on a random sample
from an emissions test. The sample was collected during the driving cycle
and analyzed immediately afterward. This sample was then reprocessed peri-
odically for several days. A bag sample of the calibration standard and a
bag sample of exhaust doped with the calibration standard were also processed
periodically. The time-sample decay curve for each compound is shown in
Figures 18, 19 and 20. The sample integrity can be preserved for approxi-
mately 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 is con-
ducted with a gas chromatography technique. Tedlar bags are filled with
dilute exhaust during each driving cycle. Analysis of the bag sample
requires a complicated system of four analytical columns with backflush and
temperature program capability. Sample concentrations are determined by com-
parison to a calibration blend of all eight hydrocarbons. The minimum detec-
table 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
be expected for the six compontents of the first sample loop and first two
columns (C^ - 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, benzene and toluene were shown to have a
large decrease 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 concen-
tration. 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 confidence in the sample concentrations obtained. Otherwise, the
sample integrity 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 indivi-
dual 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 instru-
mentation. The total analysis time per sample is about 32 minutes. The auto-
51
-------
-1-30
Legend
• Methane
+20
Ethylene
Propylene
fJotoinal Range oS
+10 h Injection Repeatability
6
3
•§ o
-10
-20
-30
6 7 8 9 10 11 12 13
Time, Days
Figure 18. Time-sample decay curve (exhaust only).
52
-------
.c
u
4J
c
I-
Nominal Range of
Injection Repeatability
;.
* * 1 t
Legend
0 Methane
• Ethylene
A Ethane
• Acetylene
-50
-100
I
10
15 20
Time, Days
25
30
35
Legend
50
„, n
1
<3
£
U
4-1
C
8
a -so
a
•
Nominal Range of
Injection Repeatability *
• t
8 1 1 •
v 9 • •
-
0
i
Propane
Propylene
Benzene
Toluene
I
-100
I
10
15 20
Time, Days
25
30
35
Figure 19. Time-sample decay curve (standard only)
53
-------
+ 30 f-
+20
+10
i
e,
-10
-20
-30
Nominal Range of
Injection Repeatability
4-
•
8
A
_L
5 6 7 8 9 10 11 12 13
Time, Days
Figure 20. Time-sample decay curve (exhaust + standard)
54
-------
+30 -
.,.20
+10
Legend
0 Propane
A Propylene
• Benzene
• Toluene
Noisinal Range for
Injection Repeatability
T
§-10
-20
-30
I
5 6 7
Time , days
10
11
12
13
Figure 20 (Cont'd) . Time-sample decay curve (exhaust + standard)
55
-------
mated system provides a simplified operation for an otherwise complicated
procedure and enables routine analysis for a large quantity of samples.
56
-------
SECTION 5
ORGANIC AMINE PROCEDURE
LITERATURE SEARCH
The individual amines that are included in this analysis are monomethyl-
amine, dimethylamine, monoethylamine, trimethylamine, diethylamine and tri-
ethylamine. The chemical formulas, molecular weights, boiling points,
freezing points, and synonyms for these low molecular weight aliphatic
amines are presented in Table 15. 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 recommended 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 (13,14) re-
ported gas chromatographie 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 (15), et al reported separation of methyl amines, ammonia, and methanol
using a mixture of tetrahydroxyethylethlenediamine and tetraethylenepenta-
mine. O'Donnel and Mann (16) 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 (17) 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 (18) used aromatic polymer beads to seperate 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 GJ-CS were separated and analyzed using this approach. In another
study Carbopak B/4 percent Carbowax 20 M/0.8 percent KOH (19) and 28 percent
Pennwalt 223/4 percent KOH (20) have been reported to give satisfactory sep-
arations of lower aliphatic amines.
Analysis of amines as derivatives has been shown to be a valuable ana-
lytical tool to determine trace quantities (21). Thirteen different deri-
57
-------
TABLE 15. LIST OF INDIVIDUAL ORGANIC AMINES
INCLUDED IN THE EMISSIONS CHARACTERIZATION INVENTORY
en
oo
Name
Monomethylamine
Monoe thy lamine
Dimethylamine
Trime thy lamine
Diethy lamine
Triethylamine
(12)
(12)
(12)
(5)
(5)
(5)
Carbon
No.
1
2
2
3
4
6
Chemical Molecular
Formula Weight
CH NH 31.058
C2H NH 45.085
(CH3) NH 45.085
(CH3)3N 59.112
(C2H_)2NH 73.14
(C0H_)W 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
Synomyms
Methylamine ,
aminome thane
Ethy lamine,
aminoethane
None
None
None
None
-------
vatives were evaluated in terms of FID and ECD response characteristics.
"This work was limited to primary amines, and under optimum conditions amines
down to 10 picograms could easily be quantified using an ECD detector.
Clark and Wilk (22) used an ECD to evaulate the properties of halogenated
amine derivatives. No increase in the sensitivity for the trifluoroacetyl
amine derivatives using ECD was observed.
Hosier (23) , 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.
Methylamine and ethylamine were detected in irradiated beef by Burks
(24), et al. Several techniques, including colorimetric paper chromato-
graphy and gas chromatography, were used in quantiflying results. Gas chro-
matographic determination of free mono-, di-, and trimethylamines in biolog-
ical fluids were performed by Dunn (25), 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 (26).
Andrea (27^, et al developed a precolumn inlet system for the gas chro-
matographic 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 ana-
lytical 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 sulfuric acid and aliquots were injected into the
pre-column of the GC. Release of the free amines was found to be sufficient-
ly reproducible for quantification of results. This technique avoided the
problems encountered by Umbreit (28), et al, and Hardy (29) 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 chroma-
tographic column that changed with every injection. In addition, the column
had a very short usable lifetime and lacked reproducibility after extended
use.
Bowen (30) 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
59
-------
analysis of the amines should be conducted by the use of gas chromatography.
A Perkin-Elmer 3920B gas chromatograph was dedicated for this purpose. This
instrument has a dual/differential electrometer and has linear temperature
programming capabilities along with a sub-ambient oven accessory. The in-
strument has been equipped with a flame ionization detector (FID), a nitrogen
phosphorus detector (NPD) , and an electron capture detector (ECD) and can be
connected to a chemiluminescent detector. Of the specialty detectors avail-
able for the analysis of the amines, the NPD appeared to be the prime candi-
date 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, dime thy lamine, 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 proce-
dural development experiments. These sources allowed a method for preparing
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 4 mm (id) , 6" x 2 mm (id) and a 12' x 1/4"
glass column packed with 4 percent Carbowax 20 M and 0.8 percent KOH on
Carbopak B, and a 6' x 4 mm (id) glass column packed with 2 percent KOH on
Chromasorb 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 20 M
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 diemthylamine 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 dimethyl-
amine 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 the satisfactory analysis. Two collection pro-
cedures were evaluated, one in which the amines are collected in a trap
filled with 1 gram Tenax-GC packing material, and another in which the amines
are collected by bubbling the amines through an acid solution. There was
some breakthrough of the amines through the Tenax-GC traps even at liquid
60
-------
nitrogen temperatures. Bubbling the amines through an acid solution 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
20 M and 0.8 percent KOH on Carbowax B packing material in the GC column was
designed to be used with aqueous solutions and proved to be satisfactory
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 history, glass syringe purging tech-
nique, precolumn conditioning, and column effects. 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 sulfuric acid, converting the trapped amines to their pentafluoro-
benzoyl chloride derivatives, and analyzing for these derivatives using a
gas chromatograph 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)
cannot be detected by this procedure, and the secondary amines (dimethyl-
and diethylamine) had a low sensitivity that made detection in the ppb range
almost impossible. The peak areas of the primary amines (methyl- and ethyl-
amine were found to be time dependent. The GC peak areas of the methyl- and
ethylamine 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 16. The standing time includes 2
minutes of vigorous shaking (1 minute for the 1 minute test) plus any re-
maining time the mixture was allowed to stand at room temperature before
injecting into the GC. The effect of elapsed time on peak area is shown in
Figure 21. 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 conditions this would not be possible. Because of the
time limitation, the procedure was abandoned for the quantitative analysis
of the organic amines.
61
-------
TABLE 16. 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 yg PFBC
in 100 m& toluene) into reacti-vial.
4. Pipette 1 m£ 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
62
-------
Methylamine Derivative
Ethylamine Derivative
1
0123456789 10
Time elapsed (in minutes) after mixing and before injection
11
Figure 21. GC peak areas of pentafluorobenzoyl amine derivatives vs time.
63
-------
The GC-NPD procedure using the Ascarite precolumn has been reevaluated
and most of the problems involved with its use have been solved. The in-
consistent lifetime of the precolumn remains a problem in the procedure.
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 20 M 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
concentrations 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
64
-------
efficiency or the physical loss of any absorbing reagent.
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 triemthylamine 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 17.
Injections of the 0.01 N sulfuric absorbing solution were also made into the
GC system. Peaks for monomethylamine and dimethylamine/monoethylamine (the
two compounds give one peak in the procedure and are analyzed together as
C2H7N) were detected in the absorbing solution and gave areas which corre-
sponded to 50 percent of the area for the monomethylamine and diemthylamine
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 trimethylamine were pre-
pared. These were made by weighing out required amounts of each of the
organic amine-hydrochloric acid salts and dissolving them in the proper
amount of sulfuric acid absorbing solution. Figures 22-27 show plots of the
GC peak areas versus the concentration for each of the organic amines on a
65
-------
TABLE 17. INJECTION REPEATABILITY EXPERIMENTS
Amine
Monomethylamine
Dime thy lamine
Trims thy lamine
Monoe thy lamine
Diethylamine
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
66
-------
10.Oi-
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 22. Linearity of monomethylamine GC response
(plot on log-log scale).
67
-------
10.0
D.02
0.01
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 23. Linearity of dimethylamine GC response
(plot on log-log scale).
68
-------
10.0
5.0
2.0
2 *
M H !-°
(IS
0) g
u
1 0.5
•H -P
-1-1 a
m
0.2
0.1
0.05
0.02
0.01
0.001 0.002
I
I
0.005 0.01 0.02 0.05 0.1
Concentration (ppm)
0.2
0.5 1.0
Figure 24. Linearity of trimethylamine GC response
(plot on log-log scale).
69
-------
100.0
50.0 -
20.0 -
•H
C
10.0 -
-------
100.0 r-
50.0
0.02
0.05
0.1 0.2 0.5 1.0
Concentration (ppm)
2.0
5.0
10.0
Figure 26. Linearity of diethylamine GC response
(plot on log-log scale).
71
-------
100.0
50.0
20.0
10.0
0) B
o
-H 4J
-P (0
(0
«
5.0
2.0
1.0
0.5
0.2
0.1
J
I
0.01 0.02 0.05
I
I
0.1 0.2 0.5 1.0
Concentration (ppm)
2.0
5.0 10.0
Figure 27. Linearity of triethylamine GC response
(plot on log-log scale).
72
-------
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
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 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) had 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
Qualification experiments were carried out using a Mercedes 240D vehicle.
Hot FTP (23 minute test) driving cycles were followed to generate exhaust
for the vehicle baseline emissions and for the tunnel plus vehicle exhaust
experiments. Two aluminum cylinders each containing three amines were used
in the experiments. One cylinder contained 174 ppm monomethylamine, 132
dimethylamine, and 107 ppm trimethylamine. The second cylinder contained
408 ppm monomethylamine, 241 ppm dimethylamine, and 123 ppm trimethylamine.
The cylinders were named by diluting the amine gas stream 300 fold with zero
air, collecting the diluted sample in 0.01 N sulfuric acid and analyzing the
sample with GC-NPD. The baseline emission values from the test vehicles
were found to be less than 0.005 ppm for all six amines investigated. A
test sequence was developed to determine the injection recovery for the
three methylamines from the CVS tunnel (18 inch diameter) without exhaust
present. Four tests were conducted for the amines at ppm levels ranging
from 0.13 to 0.22 ppm. Each test was conducted on a sequence basis with
a 10 minute soak with the CVS off between each 23 minute collection interval.
The sample lines were heated to 175°F to prevent amine losses in the sample
lines. The results of these experiments are presented in Table 18. As
expected, the recovery of the methylamine was very low (0.7 - 10.2 percent)
for the amine" injections. Dimethylamine recovery increased from 4.6 to
27.3 percent after four consecutive injections. Recoveries for trimethyl-
amine were more reasonable with 51.8 to 87.9 percent recovery for four con-
secutive tests. These results are similar to those obtained in qualification
experiments for gasoline-powered vehicles (1).
To determine the percent of amine recovery in the presence of exhaust,
a similar set of experiments was carried out. The Mercedes 240D was used
to generate exhaust during the 23 minute sampling period. All other para-
meters were the same as described above and in Table 18, with the exception
of adding a non-heated filter in the sample line. This filter was used to
73
-------
TABLE 18. ORGANIC AMINE RECOVERY FROM THE CVS DILUTION TUNNEL ONLY
Amine Injected
Monome thylamine
Monome thylamina
Monomethylamine
Monomethylamine
Dimethylamine
Dime thylamine
Dime thylamine
Dime thylamine
Trime thy1amine
Trimethylamine
Trimethylamine
Trime thylamine
ppm Nominal Flow
Amine
Injected
174
174
174
174
132
132
132
132
107
107
107
107
Rate , ft^/min
Amine Blend
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
CVS
320
320
320
320
320
320
320
320
320
320
320
320
Run
1
2
3
4
1
2
3
4
1
2
3
4
Calculated
ppm amine
dilute
0.22
0.22
0.22
0.22
0.17
0.17
0.17
0.17
0.13
0.13
0.13
0.13
Observed
ppm
0.001
0.003
0.008
0.022
0.008
0.016
0.032
0.045
0.069
0.119
0.087
0.118
Percent*
Recovery
0.7
1.4
3.7
10. 2
4.6
9.9
19.1
27.3
51.8
89.1
66.7
87.9
* All values are the average of three independent samples.
-------
prevent particulate from contaminating the sampling system. Only trace
amounts of amines were recovered in these experiments. The tests were re-
peated using a heated filter and higher concentrations of the three methyl-
amines. The heated filter and the sample line connecting the dilution tunnel
and the heated filter were heated to 375 °F. The line connecting the heated
filter and the sampling system was maintained at 175°F.
A new amine cylinder (higher in amine concentrations) and lower dilution
ratios was used to give expected ppm levels of 0.21 to 0.68 ppm in the di-
lution tunnel. The results of these experiments are presented in Table 19.
The methylamine recoveries ranged from 12.0 to 27.3 percent, dimethylamine
recoveries ranged from 18.6 to 44.2 percent and trimethylamine recoveries
ranged from 47.6 to 59.5 percent recovery. If a heated filter is used it is
possible to detect amines in exhaust at 0.2 ppm and higher levels. At levels
lower than 0.2 ppm, losses to the dilution tunnel and to the exhaust may
prevent the detection of the amines.
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 recov-
eries in the tunnel with and without exhaust present. Trimethylamine recov-
eries were generally higher and improved with the number of consecutive in-
jections.
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 m£ 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 H/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 present 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 H/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
75
-------
TABLE 19. ORGANIC AMINE RECOVERY FROM THE CVS DILUTION TUNNEL WITH EXHAUST
CTi
Ainine Injected
Mon ome t±iy 1 ami ne
Monomethylamine
Monomethylaraine
Dimethylamine
Dimethylamine
Dime thylamine
Trimethylamine
Trimethylamine
Trimethylamine
ppm
Amine
Injected
408
408
408
241
241
241
123
123
123
Nominal Flow
Rate, ft3/min
Amine Blend
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
CVS
300
300
300
300
300
300
300
300
300
Run
1
2
3
1
2
3
1
2
3
Calculated
ppm amine
dilute
0.68
0.68
0.68
0.40
0.40
0.40
0.21
0.21
0.21
Observed
ppm
0.092
0.156
0.186
0.074
0.139
0.177
0.098
0.112
0.122
Percent*
Recovery
12.0
22.9
27.3
18.6
34.7
44.2
47.6
54.9
59.5
* All values are the average ,of three independent samples
-------
that differs from trims thy lamine by only 0.2 minutes, but the two can also
be easily distinguished.
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 with and without exhaust
present. The percent 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. It is possible that if the amines are present in con-
centrations of less than 0.2 ppm, the percent recovery may be very low or
essentially zero.
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.
77
-------
SECTION 6
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 (12).
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-
surment 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 (31-35). 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 (36).
A modified version of the West-Gaeke method involves the collection of
sulfur dioxide in 0.1 M sodium tetrachloromercurate(II) (TCM) . Sulfur di-
oxide reacts with the TCM to form a dichlorosulfiromercurate 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
(37,38) , and sulfamic acid is added to the absorbing solution to destroy any
interfering nitrite ion which might be present (39) .
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 reported by
several researchers (40,41). since the dye purity does effect the results
of the colorimetric procedure, various techniques for the purification of
commercial grade pararosaniline have been published (32,34,42), and para-
rosaniline purified especially for the colorimetric analysis of sulfur di-
oxide is commercially available (42).
78
-------
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 (43). Urone, et al, investigated the collection ef-
ficiency of the TCM solution by the use of microgram quantities of sulfur
dioxide tagged with 35S (44). In this investigation, it was found that a
series of bubblers cannot be used to determine absorber collection efficiency.
Bostrom obsered a 99 percent collection efficiency for a concentration range
100-1000 ppb sulfur dioxide in a TCM solution (45) .
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 (46). 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 (47). 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 (48). 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 conduc-
tivity with other sulfur dioxide procedures indicate a fair agreement (49-54).
Hydrochloric acid gas, ammonia, and chlorine substantially increase conduc-
tivity. Shikiya and McPhee found two- to fourfold differences between
different conductivity analyzers and between conductivity and colorimetric
analyzers (51) . Although the conductivity procedure may be acceptable for
point sources of sulfur dioxide in isolated 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 standard NaOH absorbing
solution. The absorbing solution is acidified and the liberated sulfurous
acid is titrated with a standard iodine solution (52). Another method em-
ploys a standard iodine-potassium absorbing solution (53). lodometric me-
thods of analysis for sulfur dioxide generally suffer from a lack of sensi-
tivity and interferences from hydrogen sulfide.
Adsorption sampling methods have also been developed for the measure-
ment of sulfur dioxide (55) . Sulfur dioxide is absorbed on silica gel,
79
-------
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 (56-60) and static collectors (61-67). 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 (61). The sulfur dioxide collection efficiency is dependent
upon temperature, relative humidity, wind speed, atmospheric concentration
of sulfur dioxide, and the length of exposure period (62) . 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 (67).
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) (68-77). The FPD 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 recent development in methods of analysis for sulfur dioxide
involves the use of ion chromatography (78). 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.
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 (79). The second derivative
80
-------
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 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 adaption to automotive exhaust was not
staightforward. Air samples had essentially the same oxygen and nitrogen
levels all of the time; however, dilute exhaust samples have variable car-
bon dioxide, oxygen and 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 automotive 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 mJl of a 3 percent hydrogen peroxide solution. The exhaust flows through
the impingers at a rate of 4 £/min. Two impingers together trap 99
81
-------
percent of the sulfur dioxide present in exhaust. A heated glass fiber 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 0.003 M NaHCO3 plus 0.0024 M Na2CO3. 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 50W-X16
anion suppressor resin (this neutralizes 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 the 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 procedure for
sulfur dioxide analysis. Tests for interferences, sample stability, and
standard stability were among those conducted. Also, the method of washing
glassware was studied.
A number of possible interferences were tested by bubbling the suspected
interfering gas at 4 £/min through three impingers in series. Each impinger
contained 25 m£ 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 20. In the zero air, zero nitrogen, 3
percent C02 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 found 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 impingers were washed. The sulfate ion
from the sulfuric acid could not be sufficiently rinsed from the impingers,
even with repeated deionized water rinses. For this reason, a 1:1 (v:v)
nitric acid and water solution was used to wash the impingers used in the
sulfur dioxide procedure. The sulfate present in the hydrogen peroxide ab-
sorbing solution also causes a positive interference. This interference
could be corrected for by subtracting the sulfate peak area of the absorbant
from the sulfate peak area of exhaust or background samples.
Another 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 de-
ionized 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 21
indicate that the fourteen week old standards repeated within 10 percent of
82
-------
TABLE 20. INTERFERENCES TO S0? ANALYSIS
Suspended
Interference
Zero Air
Zero N
3% C02~run 1
3% CO- -run 2
100 ppmc HC
100 ppm NO
100 ppm CO
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 21. SULFATE STANDARD STABILITY
Standard
Concentration
-2
/ ny »u4 - ^
V 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)
(11/30/77)
(03/13/78)
(11/30/77)
(03/13/78)
Attenuation
3
3
10
10
30
30
100
100
300
300
Height
(in)
4.73
4.77
7.72
7.62
5.41
5.45
7.93
7.81
6.09
6.08
Amplitude
125,844
125,080
148,259
147,300
130,852
131,036
149,864
148,668
135,666
135,638
%
Difference
0.6
0.6
0.1
0.8
0.0
83
-------
the freshly prepared standards. A study was also conducted with a variety of
samples of different ages to determine sulfate longevity in the 3 percent hy-
drogen 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 SO2 baseline sample produced similar re-
sults. There was no change observed in the sulfur dioxide level. A ten
week old collection efficiency sample, however, did very from its initial
concentration of 0.13 ppm by decreasing 7.9 percent to 0.12 ppm SO2. This
is greater than the injection repeatability of 1.2 percent, however, within
the minimum detection limit of 0.01 ppm S02. 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
SO2 sampling;parameters. Nominal concentrations of 5 and 12 ppm SO2 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 22.
TABLE 22. SO2 COLLECTION EFFICIENCY AS A
FUNCTION OF FLOWRATE AND TEMPERATURE
•* SO2
trapped in
Bubbler bubblers
1+2+3(ppm) 1 and 2
Flow- SO2 Concentration (ppm) and
rate Temp, (%), in bubbler
Test (£/min) (°F)
1
2
3
.,1
2
3
4
-:4
4
2
2
72
32
32
75
4.52,
6.96
6.85
6.93
98.7
99.1
97.7
98.6
Nominal 5 ppm SOg
4.31(95.3) 0.16(3.4) 0.06(1.3),
6.73(96.7) 0.17(2.4) 0.06(0.1)
6.48(94.6) 0.21(3.1) 0.16(2.3),
6.71(98.8) 0.12(1.8) 0.09(1.4),
Nominal 12 ppm SO2
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)
The largest quantity of sulfur dioxide was retained at a flowrate of 4 £/min
at ice bath temperature. Under these conditions, 99.1 percent S02 was col-
lected in the first two bubblers. It is also desirable to prevent small par-
ti culate 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-
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(92.5)
24.8(99.1)
28.5(99.0)
29.3(97.8)
30.0(98.5)
21.2(97.9)
29.9(98.5)
28.9(98.5)
0.54(2.2)
0.09(0.4)
0.07(0.2).
0.30(4.0)
0.21(0.7)
0.25(1.2)
0.23(0.8)
0.12(0.4)
24.2
25.0
28.8
29.9
30.4
21.7
30.4
29.3
9.7.7
99.6
99.8
99.0
99.3
98.8
99.3
99.6
84
-------
lect in the columns the liquid flow becomes hampered causing increased back-
pressure. A glass fiber filter in the sampling line is used to remove a
large portion of this debris from the exhaust.
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 mm 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 0.003 M NaHCO3 and 0.0024 M Na2CO3 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 23. 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 by 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.
yg S04~2 yg ~
Two different standards were analyzed: 0.5 - „ - and 4.0
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 24 shows heights corresponding to each standard used and Figure 28
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 carry the curve any
further, since no samples have been obtained that fall in the higher con-
centration range. However, it was found that lineatity was maintained at
concentrations from 40 to 100U9S4 .
QUALIFICATION EXPERIMENTS
Qualification experiments were carried out to determine the percentage
of sulfur dioxide that could be recovered at the sampling point when known
amounts of sulfur dioxide were injected into the dilution tunnel at the point
where exhaust enters the tunnel. A Mercedes 240D vehicle was used as a
source of exhaust. Hot FTP (23-minute test) driving cycles were followed to
generate exhaust for the vehicle baseline emissions and for the tunnel plus
vehicle experiments. Aluminum cylinders containing 887 ppm and 9098 ppm
sulfur dioxide in balance air were used to inject sulfur dioxide into the
CVS dilution tunnel. The flow of sulfur dioxide into the tunnel was regu-
85
-------
TABLE 23. INJECTION REPEATABILITY FOR ION CHROMATOGRAPH
Sample
1
2
3
4
5
6
7
8
9
10
11
12
Concentration fw $°4 ] Attenuation Height (in)
\ mJl
3.84
3.84
3.84
3.841
3.84
3.84
0.48
0.48
0.48
0-.48
0.48
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
S 0.06 in
x
C 1.1%
V
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
x
C 1.2%
V
86
-------
TABLE 24. CALIBRATION CURVE FOR SULFUR DIOXIDE
Standard _2 Heights Corrected
Concentration (W S04 j Attenuation Height (in) to 1 x 10 scale (in)
\ .m& /
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
1
1
1
1
1
1
1
1
1
1
1
1
x 10
x 10
x 10
x 10
x 10
x 10
x 10
x 30
x 30
x 30
x 100
x 100
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
87
-------
90 -
30
70
60
"§> 50
40
1 x 10
30
• 1 x 30 corrected
to 1 x 10
A 1 x 100 corrected
to 1 x 10
20
10
10.00 20.00 30.00
Concentration of sulfate standard (
40.00 ,
Ug S0a~z,
nuc
Figure 28. S0? calibration curve.
88
-------
lated to give concentrations of 1 (CVS-tunnel only) to 11 ppm (CVS-tunnel
and vehicle exhaust). Injections of sulfur dioxide into the tunnel without
exhaust gave recoveries that ranged from 80 to 108 percent with an average
of 98 ± 8 percent (Table 25).
The recovery of sulfur dioxide in the presence of vehicle exhaust ranged
from 73 to 117 percent with an average of 97 ± 16 percent (Table 26) . The
sulfur dioxide recoveries were corrected for background levels and for vehicle
baseline emissions. Background levels ranged from 0.15 to 0.19 ppm while
the vehicle baseline emission levels averaged 7.73 ppm. The recoveries from
the tunnel in the presence of vehicle exhaust were carried out using a heated
filter and heated sample lines. The filter was used to prevent particulate
from contaminating the sampling system. The filter and sample lines were
heated to prevent sulfur dioxide from being retained on the removed parti-
culate. The recovery experiments indicate that sulfur dioxide can be quan-
titatively recovered from the dilution tunnel.
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 C02 and 100 ppmc HC did not inter-
fere within the minimum detection limit of 0.01 ppm SO2. 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 impingers also gave a positive interference for samples col-
lected in impingers washed in this bath. The problem was averted by replacing
the sulfuric acid-chromic acid with 1:1 (v:v) nitric acid. The manufacturer
of the ion chromatograph has stated that persulfite will interfere with sul-
fate analysis and that oxylate ion will interfere if the separatory column
capacity is reduced. No problem has been noted with these two species.
The effect of age on sulfate standards and samples was investigated
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/min through two
bubblers in series, each containing 25 m£ of 3 percent hydrogen peroxide
maintained at ice bath temperature (0-5°C). Heated glass fiber filters are
inserted in the sample line to prevent contamination of the samples and
subsequent column poisoning in the ion chromatograph. The linearity of
response of the ion chromatograph is maintained in the sulfate concentration
range 0.5 to 100 yg SO4~2 per m£ (100 ppm). However, changing the attenu-
ation on the ion chromatograph causes a discontinuity in the calibration
curve. This discontinuity is seen as a slope change in Table 14. The
standards analyzed at each attentuation obviously fall into a linear pattern
even though the slopes differ. A different set of standards must therefore
be run for each sensitivity setting.
89
-------
TABLE 25. SULFUR DIOXIDE RECOVERY FROM CVS-TUNNEL INJECTION
<£>
O
Nominal Flow
Rate ft3/min
Test
1
2
3
4
5
6
7
8
9
0.37
0.37
0,37
0.37
0.37
0.37
0.37
0.37
0.37
CVS
306
306
306
306
306
306
306
306
306
SO? Injected
Vol. S02
(ft3)a
8.547
8.547
8.547
8.493
8.493
8.493
8.446
8.446
8.446
Cone SO2
(ppm)
887
887
887
887
887
887
887
887
887
Calculated
SO2 Cone
ppm
1.08
1.08
1.08
1.07
1.07
1.07
1.06
1.06
1.06
Observed
SO2 Cone
ppm13
0.86
1.09
1.03
1.07
1.12
1.03
1.04
1.15
1.06
Percent
Recovery
80
101
95
100
105
96
98
108
100
Average 98 ± 8
Volume corrected to 1 atm pressure and 68°F
Corrected for background levels of sulfur dioxide (0.15 ppm)
-------
TABLE 26. SULFUR DIOXIDE RECOVERY FROM DILUTE EXHAUST BY CVS-TUNNEL
INJECTION DURING HOT FTP DRIVING CYCLE
Test
1
2
3
4
5
6
7
8
Nominal
Rate ft
S02
0.36
0.36
0.36
0.36
0.36
0.36
0.36
0.36
flow
•^
/rain
CVS
298
298
298
298
298
298
298
298
SO2 Injected
Vol. SO2 Cone SC>2
(ft3)a (ppm)
8.257 9098
8.257 9098
8.266 9098
8.266 9098
8.266 9098
8.372 9098
8.372 9098
8.372 9098
Calculated
SC>2 Cone
ppm
10.96
10.96
10.96
10.96
10.96
11.11
11.11
11.11
Observed
SC>2 Cone
ppm
7.98
11.34
9.68
10.88
11.91
8.65
12.09
13.04
Percent
Recovery
73
103
88
99
109
78
109
117
Average 97 ± 16
Volume corrected to 1 atm pressure and 68°F
Corrected for background levels (0.19 ppm) and vehicle baseline emissions
(7.73 ppm) of SO2
-------
The results of the qualification experiments indicate that most (97-98
percent) of the sulfur dioxide that is injected into the CVS-dilution tunnel
can be recovered with or without exhaust present.
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 chance of sample loss or
contamination. The ion chromatograph is sensitive to 0.01 ppm SO2 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.
92
-------
SECTION 7
NITROUS OXIDE PROCEDURE
LITERATURE SEARCH
There are six common oxides of nitrogen: nitrous oxide (N2O), nitric
oxide (NO) , nitrogen dioxide (NO2> / dinitrogen trioxide (^03) , dinitrogen
tetraoxide (N2O4) , and dinitrogen pentoxide (^05) . 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, 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
nitrogen, oxygen, and nitric oxide. At elevated temperatures, it will sup-
port combustion and oxidizes certain organic compounds and alkali metals.
Nitrous oxide, N20, 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 A and is isoelec-
tronic with carbon dioxide. When inhaled, nitrous oxide may cause hysteria,
insensibility to pain, or unconsciousness and therefore is used as anesthetic
for minor operations, 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 nitrites or nitrates, the slow decomposition of hyponitrites,
and by the thermal decomposition of hydroxylamine (12).
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 (80).
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 (80). With the third technique, grap samples are col-
lected in Tedlar plastic bags and analyzed with an electron capture detector
(81).
PROCEDURAL DEVELOPMENT
The gas chromatograph operating conditions and sampling system specifi-
cations were obtained from EPA-RTP (81) . A two column system with column
93
-------
backflush and isothermal temperature operation was constructed for the ana-
lysis. "She 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'xl/8" stainless steel column packed with 120/150 mesh Proapak Q. A
series of two 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 develop-
ment. 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 Tracor 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 diffi-
cult. Because of the problems associated with the permeation system, a sta-
tic 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 samples are to be analyzed. This is
the case with gaseous bag samples. No easy method to dilute the bag concen-
trations into the linear range of the instrument is available. Therefore,
the detector must be linear in all of the concentration ranges expected.
The detector linearity for the electron capture detector was determined with
calibration gases from 1 to 10 ppm (Figure 29). Sample concentrations with-
in this range are linear with respect to the detector.
With the 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 27. 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 two random Tedlarbags filled
94
-------
10 I-
en
I
CL,
.3 6
.p
4-J
q
0)
o
o
o
•8
•H
in
p
O
•H
3
I
I
6 8 10
Peak Area (counts x 10,000)
12
14
Figure 29. Detector linearity curve,
-------
TABLE 27. INJECTION REPEATABILITY OVER THE RANGE
OF DETECTOR LINEARITY
Concentration, ppm
Average
Standard Deviation
1.31
17935
17811
18346
18149
18319
18112
235
2.16
Area
28020
28759
28974
28931
28671
444
4.95
61146
61057
61004
61325
61448
61196
186
9.90
118085
118935
118687
119448
119810
119005
118995
599
Coefficient of 1.30 1.55 0.30 0.50
Variation
96
-------
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 days. The time sample decay curve
for both samples is shown in Figures 30 and 31. Samples may be stored for a
period of five days without adverse effects to the sample concentration.
QUALIFICATION EXPERIMENTS
Qualification recovery experiments were conducted for nitrous oxide with
the dilution tunnel and with real vehicle exhaust. An aluminum cylinder con-
taining 9000 ppm nitrous oxide in balance nitrogen was used as the source
for nitrous oxide- The cylinder was named by dilution with zero nitrogen
and the comparison of the diluted sample to a known standard. The exhaust
in the experiments was generated from a Mercedes 240D over hot FTP (23 minute)
driving cycles. The flow of nitrous oxide into the tunnel was regulated to
give a concentration of approximately 10 ppm nitrous oxide in the dilution
tunnel. Injections of nitrous oxide into the tunnel without exhaust gave
recoveries that ranged from 95.1 to 105.6 percent with an average of 100.5
± 3.8% (Table 28) . The recovery of nitrous oxide with real vehicle exhaust
ranged from 82.7 to 113.7 percent with an average of 99.6 ± 11.3 percent
(Table 29) . Recoveries from the injection into the tunnel without exhaust
were corrected for background levels of nitrous oxide. The injections with
the vehicle exhaust were corrected for the vehicle baseline emissions of
nitrous oxide as well as for the background levels of nitrous oxide.
RESULTS AND CONCLUSIONS
The measurement of nitrous oxide in dilute exhaust can be conducted with
gas chromatography. Dilute exhaust is collected in a Tedlar bag as a grab
sample. Sample analysis of the bag sample with an electron capture detec-
tor and comparison to a set of calibration blends determines the concentra-
tion 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 recovery is essentially 100 percent in dilute
exhaust for nitrous oxide. No losses were observed with or without the
vehicle. 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.
97
-------
20
00
10
o
•H
4J
10
c
0)
o
fi
o
o
o
OJ
2
fi
•H
0)
Cn
c
u
4J
§ -10
u
0)
-20
nominal range for
injection repeatability
I
I
3 4
Time, hours
Figure 30. Sample decay curve (short term).
-------
c
o
20
10
Q)
O
C
O
O
O
CN
nominal range for
injection repeatability
vo
to
C
•H
0)
tTl
c
-p
fl
0)
04
-20
Time, days
Figure 31. Sample decay curve (long term).
-------
TABLE 28. NITROUS OXIDE QUALIFICATION EXPERIMENTS - NO VEHICLE
Run Bag Percent Recovery*
1 1 97.2
2 102.5
2 1 105.6
2 101.2
3 1 95.1
2 101.1
Average 100.5 ± 3.8 percent
* Corrected for background levels of nitrous oxide.
TABLE 29-. NITROUS OXIDE QUALIFICATION EXPERIMENT WITH; VEHICLE EXHAUST
Run Bag Percent Recovery*
1 I 82.7
2 95.0
2 1 113.7
2 108.1
3 1 93.4
2 104.8
Average 99.6 ± 11.3 percent
* Corrected for vehicle baseline emissions and background levels of
nitrous oxide
100
-------
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.
101
-------
SECTION 8
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 H2S. 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 (12)- It is a
very weak diprqtic acid with dissociation constants:
K = 5.7 x 10~8
K = 1.2 x 10"15
Hydrogen sulfide may be. detected by its odor at about 1 ppm; 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 (>700 ppm) .
Hydrogen sulfide is prepared commercially as a by-product from many
chemical processes and by the treatment of metallic sulfides with mineral
acid such as hydrochloric or sulfuric acid (12). 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 w.ater:
and with insufficient oxygen to form free sulfur and water:
2H20 + 02 —>• 2H20 + 2S
Hydrogen sulfide also reacts with sulfur dioxide to form free sulfur and
water:
2H2S + S02 —> 2H20 + 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
spectrum of analytical methods. Some of these methods include: surface
102
-------
reactions on plates, tiles, tapes or filters, wet chemical, fluorimetry,
infrared spectroscopy, sulfur ion selective electrode, coulometry, gas chro-
matography, and colorimetric (82,83). Most of these are not applicable to
dilute exhaust sampling but are applicable for ambient air sampling or "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 (84-93). The sodium nitroprusside method has a lower detection
limit of about 1 ppm. This method was not considered sensitive enough for
the concentrations expected in dilute exhaust. The methylene blue method,
on the other hand, has a reported lower detection limit of 1-2 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, cadium sulfate, and zinc acetate
have been used as the absorbing media. However, several authors have re-
ported the oxidation of cadmium and the photochemical decomposition of
cadmium sulfide. Bamesberger and Adams (85) 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. Flamm
and James (93) tested all of the above absorbing reagents and found zinc
acetate to be the most efficient absorbant.
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 (86) .
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 (94) 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, sul-
fides in alkaline solutions are easily oxidized by air. Second, cadmium
sulfide is photosensitive and solutions must be protected at all time from
exposure to light. The use of special glassware or aluminum soil wrappings
are necessary to prevent exposure to light. The addition of a stabilizer
such as STRactan 10 helps to minimize the effect of photochemical decompo-
sition, but special handling precautions are still necessary. Cadmium solu-
tions are hard to work with and in addition some cadmium compounds are toxic.
Cadmium, cadmium oxide, cadmium sulfate, and cadmium sulfide were included
in 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. For these reasons zinc
acetate was selected as the absorbing reagent for this project.
103
-------
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 aliminum foil 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 32
and 33. The results of these experiments are discussed in the Results and
Conclusions 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. A thiosulfate solution is first standarized
against potassium dichromate. This standarized thiosulfate solution is used
to standarize a dilute iodine solution. The standard sulfide solution con-
centration is then determined with an iodimetric method. Aliquots of the
standarized 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 hy-
drogen sulfide through impingers containing the absorbing solution for vary-
ing lengths of time. Generation of a calibration curve in this.manner takes
into account the collection efficiency of the impingers. 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 34. This
curve was determined for a standarized suflide 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-
104
-------
o
a
o
CJ
• Dark
Light
10
20
30
time, rain.
Short term
50
60
10
Dark
Light
10
30
time, days
Long term
I
oO
Figure 32. Time-Light exposure study low concentration H2S.
105
-------
25
o
2 20
o
u
!-•
Ul
10 L,
Light
Dark
Projected from
long term study
time , hours
Short term
=i
E
,
S
10
20 30 40
time, days
Long term
50
60
Figure 33. Time-Light exposure study high concentration H2S,
106
-------
1.6
1.5 -
1.4 -
1.3 -
1.0 -
0.6
0.5
0.4
0.3
0.2
0.1
• Day 1 (11/23/77)
• Day 2 (12/1/77)
m^
If' .
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
Sulfide concentration, pg S=/100 mi
Figure 34. Beer's Law plot for methylene blue.
-------
ducted to verify other procedural parameters. Collection efficiency and
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 suflide was passed through
the absorbing reagent at 1.0 and 4.0 H/min. The experiment was then repeated
with a sample flow of 4.0 Vmin 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 30. Sample flow rate and absorbing
reagent temperature did not have a measurable effect on the collection ef-
ficiency.
TABLE 30. THE EFFECT OF SAMPLE FLOW RATE AND ABSORBING
REAGENT TEMPERATURE ON THE COLLECTION EFFICIENCY
Absorbing
Test Reagent Sample Percent H2S 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 absrobing 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 the background air and a blank with no gas bubbled
through it. Each gas was bubbled at 4.0 £/min for twenty minutes and de-
veloped for methylene blue. No interference from these gases was observed.
A second experiment investigated the interferences from individual ex-
haust gas components on asulfide 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 de-
velopment of methylene blue. Table 31 shows the results of this experiment.
NOX at 3000 ppm and sulfur dioxide at 5 ppm were found to quench the produc-
tion of methylene blue.
108
-------
TABLE 31. THE EFFECT OF INDIVIDUAL EXHAUST COMPONENTS
ON THE DEVELOPMENT OF METHYLENE BLUE
Gas
Doped absorbing
reagent
Carbon dioxide
Carbon monoxide
Doped absorbing
reagent
Air
Hydrocarbon
Sulfur dioxide
NOX
Doped absorbing
reagent
NO.
x
Sulfur dioxide
Sample
1
2
1
1
1
2
1
1
1
1
2
1
2
1
2
Gas
Cone.,
PPM
29,900
2,709
168
5
3,460
315
315
5
5
Absorbance
0.716
0.708
0.711
0.704
0.552
0.560
0.574
0.554
0.461
0.247
0.648
0.654
0.666
0.621
0.484
0.533
Methylene Blue
Apparent
Sulfide Ion
Cone. , yig/ml.
0.646
0.638
0.641
0.634
0.492
0.500
0.513
0.494
0.408
0.213
0.582
0.587
0.599
0.557
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 bisulfate ions were used. Each of these anions was added to separate
solutions 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 bisulfate was investigated to determine the speci-
fic source of the sulfur dioxide interference. The results are shown in
Table 32. Only bisulfate ion and thiosulfate ion caused the negative inter-
ference .
Finally, an additional experiment was conducted to help determine the
source of sulfur dioxide interference. Approximately 2.5 ft3 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
109
-------
standard sulfide ion solution was added to two of these impingers. To two
others, 5 m£ 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 33<, Again, the absorbance for methylene blue was de-
creased by the presence of sulfur dioxide.
TABLE 32. THE EFFECT OF ANIONS ON THE DEVELOPMENT
OF METHYLENE BLUE
Apparent
Sulfide Ion
An ion Sample Absorbance Cone., yg/m&
doped with hydrogen sulfide
Sulfate ion 1 0.238 0.205
2 0.241 0.207
Thiosulfate ion 1 0.207 0.177
2 0.211 0.181
Bisulfate ion 1 0.239 0.206
2 0.231 0.198
Doped absorbing reagent 0.241 0.207
un doped
Sulfate ion 1 0.001
2 0.000
Thiosulfate ion 1 0.009
2 0.004
Bisulfate ion 1 0.007
2 0.005
110
-------
TABLE 33. THE EFFECT OF SULFUR DIOXIDE INTERFERENCE
ON THE DEVELOPMENT OF METHYLENE BLUE
Sulfide ion
added, m&
1
1
5
5
0
0
1
5
Absorbance
Sulfur dioxide passed
through absorbing reagent
0.067
0.070
0.463
0.326
0.002
0.006
No sulfur dioxide passed
through absorbing reagent
0.082
0.583
Apparent
Sulfide ion
Cone., yg/m£
0.055
0.057
0.410
0.284
0.000
0.000
0.067
0.521
QUALIFICATION EXPERIMENTS
A Mercedes 240D was used in the qualification experiments for hydrogen
sulfide. The baseline emission rate for this vehicle was below the detec-
tion limits for the analytical procedure. This baseline was established
from three separate hot FTP driving cycles. Hydrogen sulfide was injected
into the .CVS-tunnel system with and without vehicle exhaust. The concentra-
tion of hydrogen sulfide injected into the tunnel was 909 ppm. The flow of
hydrogen sulfide into the tunnel was adjusted to give a diluted concentration
of approximately 1 ppm. With the vehicle present, the hydrogen sulfide was
injected into the raw exhaust stream as it entered the dilution tunnel.
Samples were taken from the dilute exhaust stream and passed through a
buffered zinc acetate absorbing reagent. The samples were treated with an
amine solution and a ferric ion solution and then analyzed with a Beckman
spectrophotome ter.
Injections of hydrogen sulfide into the tunnel without exhaust gave re-
coveries that ranged from 85.0 to 96.5 percent with an average of 90.3 per-
cent (Table 34). Initial experiments for the recovery of hydrogen suflide
in the presence of vehicle exhaust gave recoveries from 60 to 65 percent.
A second set of experiments with injections of hydrogen suflide into the
dilution tunnel with vehicle exhaust was carried out. In this experiment
five samples were collected and treated with 6 m£ of ferric ion solution
while six others were treated with 2 m£ of ferric ion solution (Table 35) .
The samples that were treated with 6 m£ of ferric ion solution gave
recoveries that ranged from 84.1 to 95.6 percent with an average of
111
-------
TABLE 34. HYDROGEN SULFIDE RECOVERY - NO EXHAUST PRESENT
Nominal Flow
Rate, ft /min
H2S
Blend
0.375
0.375
0.375
0.372
0.372
0.375
0.375
0.375
CVS
302
302
302
302
302
302
302
302
Run
1
1
1
2
2
3
3
3
Sample
1
2
3
1
2
1
2
3
Calculated
ppm H2S
Dilute
1.13
1.13
1.13
1.12
1.12
1.13
1.13
1.13
Observed
ppm*
0.96
0.99
1.09
0.98
1.02
1.02
1.07
1.03
Percent
Recovery
H2S
85.0
87.7
96.5
90.0
87.3
90.2
94.6
91.1
Average 90.3 ± 3.8
* Corrected for background levels of H2S
90.6 percent. The samples that were treated with 2 m£ of ferric ion solution
gave recoveries that ranged from 57.8 to 80.8 percent with an average of
70.7 percent. Two m£ of ferric ion solution had previously been found to
be sufficient in the production of methylene blue in the presence of exhaust
from gasoline powered vehicles. This was also the amount of ferric ion used
in the previous diesel recovery tests. The recoveries were approximately
20 percent higher when 6 m£ of ferric ion were used. The recovery of 90
percent using the 6 m£ of ferric ion is also equal to the recovery from the
tunnel when exhaust is not present.
Additional tests were performed using different amounts of ferric ion
solution. The test using 2 and 6 mJi of ferric ion solution was repeated on
eight other exhaust samples. Four were treated with 2 m£ of ferric ion while
the other four were treated with 6 m& of ferric ion. The four treated with
6 m£ gave recoveries of 19 percent higher than the four treated with 2 mi.
Recoveries averaging higher than 90 percent could not be obtained using more
than 6 m£ of ferric ion solution. The recovery experiments without vehicle
exhaust were repeated and no difference was found when 2 or 6 m£ of ferric
ion were used. Several laboratory experiments were conducted to try to de-
termine what chemical species were involved in this phenomenon, however,
no conclusive results were obtained.
112
-------
Ninety percent of the hydrogen sulfide injected into the CVS-dilution
tunnel can be recovered from the dilution tunnel with or without exhaust
present. Six m£ of ferric ion solution must be used to obtain maximum
recoveries when diesel exhaust is present.
TABLE 35. EFFECT OF FERRIC ION SOLUTION ON HYDROGEN SULFIDE
RECOVERY FROM DILUTE EXHAUST
Nominal Flow
Rate, ftj/min
H2S
Blend
0.367
0.367
0.375
0.375
0.377
0.377
CVS
296
296
297
297
296
296
Run
1
1
2
2
3
3
Sample
1
2
1
2
1
2
Calculated
ppm H2S
Dilute
1.13
1.13
1.15
1.15
1.16
1.16
Observed
ppm
0.75
0.83
0.66
0.79
0.86
0.93
m£ Ferric
Ion Added
2
2
2
2
2
2
Percent
Recovery
H2S
67.1
74.3
57.8
69.2
74.8
80.8
Average 70.7 ±
7.8
0.367
0.375
0.375
0.377
0.377
296
297
297
296
296
1
2
2
3
3
3
3
4
3
4
1.13
1.15
1.15
1.16
1.16
0.94
1.04
1.07
1.10
1.02
6
6
6
6
6
84.1
91.1
93.7
95.6
88.7
Average 90.6- ±
4.5
* Corrected for background and baseline levels of H2S
RESULTS AND CONSLUSIONS
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 dimethy1-para-phenylene
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.
113
-------
Several experiments were conducted to determine the interferences and
their sources in the analysis of hydrogen sulfide. Individual exhaust gas
components such 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 of 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 bisulfate ion and thio-
sulfate ion produced the same effect. Bisulfate ion can be produced from
sulfur dioxide by the simplified reaction:
SO2 + H20 —> HS03~ + H
Thiosulfate ion forms bisulfate ion in strongly acidic solutions:
S203= + 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 bisulfate 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 con-
centration 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 consis-
tency. The sample flow rate of 4.0 H/min 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 absorbing reagent was selected for simplicity and consistency with the
other analytical procedures which require an ice bath. The absorbing re-
agent 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.
Hydrogen sulfide qualification experiments revealed average recoveries
with or without exhaust present of 90 percent. Other experiments revealed
that in the presence of the diesel exhaust higher recoveries are obtained
114
-------
when 6 mfi- of ferric ion solution are used instead of the usual 2 m£.
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. 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 anal-
ysis and makes this procedure ideal for analyzing a large number of samples.
115
-------
SECTION 9
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 A, 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 react with protonic acids to form water
soluble ammonium salts. It also reacts to form stable metallic complexes.
Chemically, ammonia is a highly associated, stable gas with only slight dis-
sociation at 840-930°C and atmospheric pressure. The toxicity level of
ammonia for humans 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 per-
ceptible at 20-50 ppm (12, 95, 96). Commercially, ammonia is produced by
the Haber Process according to the reaction:
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 (12,
96) .
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 (97). Distillation is necessary
to remove these interferences. The indophenol method is more sensitive
than Nessler's (98), but it too suffers from contamination by formaldehyde,
SO2(10:1), Fe, Cr, Mn, and Cu (99). During color development pH must be
carefully controlled for reliable results (98). Another highly sensitive
procedure is the pyridine-pyrazalone method. It is very involved and is
susceptible to interference from some cations at high concentrations (100).
A direct colorimetric method for ammonia analysis involves collection in a
neutral solvent (dioxane) containing a quinone and subsequent absorbance
116
-------
measurement at 480 nm on a spectrophotometer. The major drawback of this
procedure is that one of the reagents, 9-(benzenesulfonamide)-p-benzoquinone,
must be synthesized, purified, extracted with benzene, and recrystalized
before use (101). 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~^ M solutions. Several instrumental optical methods
are used in ammonia analysis. These include the chloramine (102), cupri-
ammonia complex (103) , ninhydrin (104) , and electroanalytical methods.
Additionally, there is a method for direct measurement on a spectrophotometer
with a UV (105, 106) or IP, (107, 108) detector as well as a number of in-
direct colorimetric methods (109) . Gas (110, 111) and paper (112) chroma-
tography 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 (113,114). Interferences again pose a problem with these proce-
dures. 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 NH, 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 (115). Volatile amines interfere with analysis (116), and the
hydrophobic membrane of the electrode has been found to deteriorate in 2 to
3 weeks (115).
Both gasometric (117-119) and gravimetric (118, 120, 121, 122) tech-
niques are not sensitive enough for trace analysis, and the chemiluminescent
procedure is more involved than is practical (123, 124). The disadvantage
of an enzymatic method reacting ammonia, an a-keto ester, and reduced nico-
tinamide adenine dinucleotide (NADH) is the high cost of NADH (101) .
Ammonia has been quantitatively measured in dilute automotive exhaust
using an ion chroma to graph. 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 affect 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.
117
-------
Very little procedural development was necessary for this method of
analysis. However, instrument and sampling parameters did need to be selec-
ted. Suggested instrument variables such as type and strength of eluent,
flowrate, chartspeed, and separator column size were provided with the ion
chromatograph. These will change somewhat with each set of columns. Ultra
pure nitric acid and distilled water have been found to give the best base-
line and most rapid recovery from suppressor column regeneration. The re-
generant solution is a 0.5 N NaOH solution made from reagent grade sodium
hydroxide. Chartspeed was set at 12 in/hr, and the flowrate at about 40
percent of fullscale. Good separation was obtained with 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 resolu-
tion deteriorates.
The sampling parameters were determined as part of the validation ex-
periments. A sampling rate of 4 Vminute at ice bath temperatures was found
to be most efficient. Two bubblers containing 25 m£ of 0.01 N H2S04 capture
over 99 percent of the ammonia passing through. A filter located between
the sampling cart and the dilution tunnel, is used to prevent diesel par-
ticulate from contaminating the sampling system. The line connecting the
filter to the dilution tunnel and the line connecting the filter and the
sampling cart are heated to 175 °F in order to prevent water from condensing
in the sample line.
VALIDATION EXPERIMENTS
The first validation experiment conducted involved the selection of
sampling parameters: flowrate, collection temperature, number of bubblers,
and absorbing reagent. The data from these collection efficiency tests is
found in Table 36. Ninety-nine plus percent of the ammonia was trapped in
the first two bubblers under all test conditions. A flowrate of 4 £/minute
was selected to obtain the most sample without loss of sampling efficiency
or physical loss of absorbing solution. Sampling at ice bath temperatures
was selected to be consistent with other procedures; however, as seen in the
data, room temperature sampling is also 99+ percent efficient. Two impingers
containing 25 mi of 0.01 N H2SO4 as the absorbing solution are therefore used
to trap 99+ percent of the ammonia. .Increasing the acidity of the absorbant
(0.06 N) causes interference with the ion chromatographic analysis by
broadening the eluted peaks. Of significant importance is column contami-
nation that occurs if particulate is not filtered from the sample prior to
analysis. To prevent this contamination, a filter in the sample line is
used. Heated lines are used to prevent condensation of water and the loss
of ammonia in the sample line. The column can be poisoned by heavy metals
present in the exhaust. These compounds adhere to the column resin and will
118
-------
TABLE 36. NH3 COLLECTION EFFICIENCY AS A FUNCTION OF FLOWRATE AND TEMPERATURE
14
15
NH3 concentration (ppm) and
Test
1
2
3
4
5
6
7
8
9
10
11
12
13
Flowrate
U/min)
2
2
2
2
4
4
4
4
4
2
2
4
4
percent in bubbler
Temperature ( °F)
Ammonia
32
32
32
32
Average
32
32
32
32
32
Average
74
74
Average
75
74
1
flow diluted
17.10(99.6)
19.51(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)
3
3
1:5 with zero nitrogen
0.06(0.4)
0.10(0.5)
0.28(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
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
Bubbler
1+2+3 (ppm)
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
2
2
Average 18/44(99.7) 0.01(0.05) 0.05(0.3)
Ammonia flow diluted 1:20 with zero nitrogen
32
32
4.32(99.7)
4.15(98.9)
0
0.02(0.5)
0.01(0.3)
0.03(0.6)
18.50
Average 4.24(99.3) 0.01(0.3) 0.02(0.4)
-------
slowly elute causing broad, unidentifiable peaks to appear periodically. To
prevent contamination, a strong nitric acid solution (1 N HNC>3) is used to .
wash the precolumn weekly. If the separator column becomes contaminated 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 in-
terfere with the ammonia peak because their retention times are close to
that of ammonia. Sodium, present in the water supply, interferes when its
concentration exceeds 2 ygNa!"/m£. The tolerable limit for potassium is only
0.5 yg K+/m&. Its presence is due to the incomplete rinsing of glassware
washed in chromic acid solution. The absorbing solution, 0.01 N H2SC>4, pro-
duces a small peak with the same retention time as ammonia, but a correction
is made for this by running a blank sample each testing day. Filtered de- ,
ionized 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 required to obtain optimum results. Nitric Acid (0.0075 N)
flowing at 200 m£/hour allows good separation between peaks when a 3 x 150
mm precolumn, a 6 x 250 mm separator column, and a 9 x 250 mm suppressor
column are used. These parameters will vary between column sets, making it
necessary to check the eluent and flowrate when columns are changed. A
small loop (0.01 m£ or 0.2 m£) prevents the relatively small ammonia signal
from being overwhelmed by the large hydrogen ion peak. An attempt was made
to neutralize the acid collection medium with sodium and potassium hydroxide,
but the sodium and potassium interferences were too large to make it-practical.
Injection repeatability.figures are shown in Table 37. The mean or average
for each set of peak heights and areas is represented by x, the standard
deviation by sx, and the coefficient of variation in percent by Cv. The
coefficient of variation serves as a comparison between injections made on
the two days. This value is simply the standard deviation divided by the mean
and multiplied by 100. Calculations are done using peak areas rather than
peak heights becuase on the whole they were more reliable. The average
variation in areas on the two days ran about 3.0 percent. A comparison was
also made on the repeatability of standard preparation. Four 0.5 U9 NH4
standards were made up using the same stock solution and analyzed on the ion
chromatograph. The results are shown in Table 38.
TABLE 38. REPEATABILITY OF AMMONIA STANDARD
/yg NH4+\
Sample Concentration! 1 Attenuation Height(in) Area
1
2
3
4
0.50
0.50
0.50
0.50
3
3
3
3
0.50
0.53
0.52
0.50
x 0.50
sx 0.015
Cv 2.9%
28947
29821
29011
26819
28650
1284
4.5%
120
-------
TABLE 37. INJECTION REPEATABILITY
FOR ION CHROMATOGRAPH
Sample
Date
1 3-29-78
2 3-29-78
3 3-29-78
4 3-29-78
5 3-29-78
1 5-09-78
2 5-09-78
3 5-09-78
Concentration Attenuation Height
, /yg mA+\
\ m£ / (Miiiho) (in)
0-50 3 0.53
0.50 3 0.54
0.50 3 0.53
0-50 3 0.53
0.50 3 0.50
x 0.53
sx 0.01
Cv 2.8%
0.50 3 0.51
0.50 3 0.51
0.50 3 0.51
x 0.51
sx 0.00
Cv 0.0
Area
30990
30374
29891
29665
28947
29973
766
2.6%
24335
23205
24785
24108
814
3.4%
121
-------
The coefficient of variation for the area is 4.5 percent. Subtracting the
Cv for injection repeatability, the repeatability of standard preparation
is 1.5 percent. A 4.5 percent error then is to be expected from the instru-
ment, and standards. The particular combination of columns and the condition
of the suppressor column determines the actual repeatability.
i
The ion chormatograph gives a linear response to ammonia at the sensiti-
vity settings of 3 ymtio»andl& ymho. Table 39 lists concentrations and
corresponding heights and "arenas of points on the calibration curve. These
values are'plotted'graphically in Figure 35. (NH4)2SO4 standards ranging
1 yg. NH4+ 4.
from about 0.4 to 30 —-j (ppm NH4 ) were run at the appropriate atten-
uations, 3 or 10 ymho. The areas recorded at 10 ymho were corrected to 3
ymho.by multiplying by 10/3. Both scales show linearity Taut the slopes are
visually different with relative values of 1.6 and 1.2 for the 3 and the 10
ymho scales, respectively. The 3 ymho scale reamins linear from at least
0.4 to 8 yg NH4+/m& and the 10 umho 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+ :
comparison (0.5 r—) was prepared on May 29, 1978. It is obvious from
the data presented in Table 40 that the sample and standards 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.
QUALIFICATION EXPERIMENTS
Qualification experiments were carried out using a Mercedes 240D vehicle.
Hot FTP (23 minute test) driving cycles were followed to generate exhaust
for the vehicle baseline emissions and for the tunnel plus vehicle experi-
ments. An aluminum cylinder containing 9226 ppm ammonia in balance nitro-
gen was used as the source of ammonia in the experiments. The flow of
ammonia into the tunnel was regulated to give concentrations of 10-12 ppm
ammonia in the dilution tunnel. The baseline ammonia emission level for the
Mercedes 240D was 0.18 ppm. Injections' of ammonia into the tunnel without
exhaust gave recoveries that ranged from 74.9 to 75.5 percent with an average
of 75.2 percent (Table 41).
122
-------
TABLE 39. CALIBRATION CURVE FOR AMMONIA
Standard
Concentration
Attenuation Height Heights corrected Area corrected
\ m£ /
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
(Umho)
3
3
3
3
3
3
3
3
3
3
3
3
3
10
10
10
10
10
(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
to 3 yniho scale Area
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
9.77 170,430
13.23 228,363
15.93 299,706
21.00 409,068
to 3ymho scale
•
472,913
568,100
761,210
999,020
1,363,560
123
-------
1,400,000 [—
1,200,000 -
1,000,000 -
-------
TABLE 40. SAMPLE AND STANDARD STABILITY
AS A FUNCTION OF TIME
Date of Analysis
5-29
5-30
5-31
6-01
6-02
6-05
6-13
6-20
6-27
6-29
7-07
Age of Sample (davs)
1
2
3
4
5
8
16
23
30
32
40
Concentration (ppm NH^ )
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.00
0.02
0.03
125
-------
TABLE 41. AMMONIA RECOVERY FROM CVS-TUNNEL, NO EXHAUST
Ammonia Injected Ammonia Recovered
Nominal Flow
Rate ft3/min
Test
1
2
3
N
0.
0.
0.
13.
38
38
38
CVS
310
310
310
volume NH3a
injected
(ft3)
8.775
8.756
8.745
Concentration
NH3 injected
(ppm)
9226
9226
9226
Total diluted*
volume
(ft3)
7130
7122
7139
Sample*3 /c Calculated
Concentration Amount of NH3
(ppm) recovered (ppm)
8
8
8
.55
.49
.53
11
11
11
.35
.34
.30
Percent
Recovery
75.3
74.9
75.5
Average 75.2
ro
en a
Volume corrected to 1 atm and 68°F
Corrected for background level of ammonia (0.06 ppm)
c
Each value is the average of three samples taken during each test
-------
Two separate experiments were carried out for the recovery of ammonia in
the presence of vehicle exhaust. In the first experiment (Table 42) the
sample lines were heated to 175°P and unheated 25 mm Fluoropore filters
(0.5 y pore size) were used to remove particulate. The recoveries ranged from
53.6 to 63.5 percent with an average of 59.5 percent. Samples taken without
the filter in place gave similar results. In the second experiment (Table 43)
the filter (7 cm glass fiber filter) and the sample line between the filter
and the dilution tunnel were heated to 375°F. All other sampling conditions
remained the same. In this experiment the recoveries were lower and ranged
from 31.3 to 36.5 percent with an average of 33.1 percent. At this time,
the reasons for lower recovery using the heated filter are unknown. It
was expected that higher recoveries would be obtained, as was the case with
the organic amines (Section 5). A twenty-five percent loss of ammonia to
the dilution tunnel can be expected when sampling for ammonia at the 5-10
ppm levels. An additional fifteen percent of the sample will be lost due to
the presence of exhaust when using an unheated filter to remove particulate.
RESULTS AND DISCUSSION
The ion chromatograph was chosen as the most favorable means of measuring
ammonia because of the simple, direct, and rapid processing of samples. 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 automotive exhaust
samples.
The sampling parameters providing the most efficieny collection of am-
monia were selected. Twenty-five milliliters of the absorbing solution, 0.01 N
H2S04, 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 4 H/min is captured in these two bubblers. After sample collec-
tion it is necessary to set instrument parameters to obtain good separation
in the shortest time possible. These parameters, such as eluent concentra-
tion 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, 0.0075 N HNO3, flowing at 30 percent of pump capacity, gives good
ammonia resolution. A small sample loop (100 y&) 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 chromatogarph gives a linear response for (^4)2804 standards in the
range 0.4 to 30 V$- jp4- -, however, the attenuator is not linear between dif-
ferent sensitivitymsettings. 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 may occur.
Ammonia cannot be quantitatively recovered from the CVS-dilution tunnel with
or without exhaust present. At ammonia levels of 5-10 ppm there is a twenty-
five percent loss of ammonia to the dilution tunnel and an additional fifteen
percent loss to exhaust.
127
-------
TABLE 42. AMMONIA RECOVERY FROM DILUTE EXHAUST (NO HEATED FILTER)
Test NH-.
1 0.37
2 0.38
3 0.37
Ammonia Injected Ammonia Recovered
Nominal Flow volume NH3aConcentration Total diluted**Sample b'c Calculated
Rate ft3/min injected NH3 injected volume Concentration Amount of NH3 Percent
CVS
302
296
296
(ft3)b
8.610
8.780
8.550
(ppm)
(ft3)
9226
9226
9226
6937
6816
6797
(ppm) recovered (ppm) Recovery
7.01 11.44 61.3
7.55 11.89 63.5
6.22 11.60 53.6
Average 59.5
ro
oo
, Volume corrected to 1 atm pressure and 68°F
Corrected for background levels (0.08 ppm) and baseline (0.18 ppm) levels of ammonia
Each value is the average of three samples taken during each test
-------
10
TABLE 43. AMMONIA RECOVERY FROM DILUTE EXHAUST, HEATED FILTER
Ammonia Injected Ammonia Recovered
Nominal Flow
Rate ft3/min
Test
1
2
3
NH3
0.37
0.37
0.38
CVS
299
299
298
volume NH-ja
injected
(ft3)b
8.435
8.592
8.699
Concentration
NH3 injected
(ppm)
9226
9226
9226
Total diluted3
volume
(ft3)
6884
6872
6852
Sample D'c Calculated
Concentration Amount of NH3
(ppm) recovered (ppm)
3.55
4.21
3.65
11.31
11.53
11.71
Percent
Recovery
31.4
36.5
31.3
Average 33.1
, Volume corrected to 1 atm pressure and 68°F
Includes baseline and background correction
^"<
Each value is the average of three samples taken during each test
-------
The ion chromatograph method of 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 12-30 minute analysis
is relatively short for such a sensitive method (minimum detection limit is
O.Olppm NH3). Sample and standard stability as well as linearity of re-
sponse in the concentration range of interest are additional factors which
make this procedure the most desirable method of measuring ammonia in auto-
motive exhaust.
130
-------
SECTION 10
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 44. Carbonyl sulfide is the only sulfide of
interest that is a gas at room temperature. In general, the organic sul-
fides are malodorous compounds that produce an unpleasant odor similar to
rotton eggs. The 1968 American Conference of Governmental Industrial
Hygienists made no recommendation for threshold limit values for these
sulfides.
TABLE 44. LIST OF SULFUR COMPOUNDS INCLUDED IN
THE ANALYSIS OF ORGANIC SULFIDES
Chemical Molecular Freezing Boiling
Sulfur Compound Formula Weight Point,°C Point,°C
Synonyms
Carbonyl Sulfide COS
Methyl Sulfide
Methyl Disulfide CH SSCH
Ethyl Sulfide
C2H5SC2H5
60.075
62.13
94.20
90.19
-138.8 -50.2 Carbon oxysulfide
-98.27
37.3
Dimethylsulfide
-84.72 109.7 Dimethyldisulfide
-103.9 92.1 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 (125-129);
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 (130-132). Temperature programmed gas
chromatography was found to improve the separation of mercaptans and
sulfides (133-135). 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 (136).
Carbonyl sulfide has been quantitatively measured in natural gas (137)
and in carbonated beverages (138) by the use of gas chromatography. The
measurement of carbonyl sulfide in carbonated beverages used an electron
131
-------
capture detector and had a detection 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, desorption
under heat and vacuum, trapping at -96°C, and transferring to a gas chroma-
tograph for analysis (139).
Several columns have been used to separate sulfur compounds from nor-
mally occurring atmospheric hydrocarbons, but little success has been ob-
tained (140). A GC-microcoulometry method eliminated the interference from
the hydrocarbons and was sensitive to 1 ppm mercaptan (141) . A gas phase
chemiluminescent reaction of ozone with organic sulfides has been considered
as method of detection in monitoring low concentration of ozone and sulfur
containing pollutants (142).
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 (143) . The FPD detector has been applied to low con-
centration air pollution monitoring (72) , measurement of trace organic sul-
fides in air (144), and soil and water anlaysis (144). Permeation tubes
have been used in several cases to generate continuous samples of known
concentrations of various sulfur compounds (71,72). The use of Teflon
throughout the gas chromatograph system has been found to minimize absorp-
tive losses (144) and has increased sensitivity to 10 ppb (71).
Several columns have been evaluated at several temperatures in conjunc-
tion with the Melpar flame photometric detector (77). 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 (145-151).
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 36-38. The sample is purged
through the gas sampling valve sample loop (Figure 36, 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 37, Step 2). After all peaks of interest have eluted from the ana-
lytical column, the column is backflushed and the system is readied for the
next injection (Figure 38, Step 3).
132
-------
Control Console
CO
CO
Step 1. Sample being purged
through injection valve
GC Oven
Figure 36. Proposed GC flow schematic for analysis of organic sulfides (Step 1).
-------
Control Console
GO
Step 2. Sample injected
into GC system
Valve Oven
GC Oven
Figure 37. Proposed GC flow schematic for analysis of organic sulfides (Step 2).
-------
Control Console
CO
cn
Step 3. Column backflush
GC Oven
Figure 38. Proposed GC flow schematic for analysis of organic sulfides (Step 3).
-------
The column selected for the initial time was a 6' x 1/8" Teflon column
packed with 60/80 Chromosil 310. Several different GC operating conditions
were tried, and a preliminary set of conditions were sleected that provided
an adequate separation of the four organic sulfides of interest. The
separation of these sulfides is presented in Figure 39. The elution of
other sulfur containing compounds is also included. Table 45 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 con-
centration 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 mH.
The basic flow schematic of the sampling system is shown in Figure 40. 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 m£/min (9.50
ppm COS and 4.77 ppm 0133013) and 81.2 m£/min (1.40 pprn 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 46. 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 in this particular case,
the trapping was effective only on methyl sulfide.
136
-------
GAS CHROMATOGRAPH CONDITIONS
Perkin-Elmer 3920B w/FPD 6' x 1/8" column packed with 60/80 chromosi1
310. N2 at 20 mVmin., oven isothermal at 0°C for 8 min and programmed
to 140°C/min. at 32°C/min.
ro
K
U
to
m
ffi
U
m
35
U
CO
CO
CO
35
U
•
I
1
0)
M
0)
tn
0!
-P
3
rH
(1)
Q)
^4
0
x;
to
0
-P
3
t-H
0)
&
in
0)
-P
3
i — t
(U
CM
o
en
m co
35 t*"1
-------
TABLE 45. 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
Methyl Mereap tan
Ethyl Mercaptan
Chemical
Formula
COS
H2S
SO,,
CH3SH
Molecular
Weight Density
Boiling
Point, °C
60.075 2.5300 g/H -50.2
34.08 1.5392 g/£ -60.3
64.063 2.927 g/£ -10.0
62.13 0.848 g/mJl 37.3
94.20 1.0625 g/mi 109.7
90.19 0.836 g/mi 92.1
0.8665 g/m£
48.11 0.8665 g/m£ 6.2
62.13 0.8391 g/m£ 35
Retention
Time
2.8
4.2
10.5
17.5
15.8
12.5
13.5
138
-------
Permeation
Calibration
System
Trap
co
.Valves
System
Control
Console
Perkin-Elmer
Model 3920B
Gas
Chromatograph
Recorder
Integrator
Liquid Refrigerant
Figure 40. Cold trap experiment flow schematic.
-------
TABLE 46„ 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
mH/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
8.L.2-
81.2
81.2
81.2
in j .
1
2
1
2
1
2
1
2
1
2
1
2
1:
2
1
2
Cone
cos
9.50
9.50
9.50
9.50
9.50
9.50
9.50
9.50
1.40;
1.40
1.40.
1 .40
1.4P,
1.40
1.4Q
1.40
Trap
. , ppm
CH3SCH3
4.77
4-77
Ayg.
4.77
4.77
Avg>
4. 77,
4.77
Avg.
4.. 77
4.77-
Avg.
0.71
0.71
Avg.
0.71-
0.71
Avgv
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;
6.3-- 5,
63.2
63,0;
63.5.
63.2
63.0
62.0.
62;5
• Trap Exit
Height
CH3SCH3
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
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-
Height
CH3SCH3
4.1
13.8:
9.0
1.1
11.4
6.3
12.6
13.0
12.8
0.0
1D..O
5.0
5.2
4.2
4,. 7,
4,2
4.2
4.2-
5.0
4.5
418
0.0,
0.0,
0.0
140
-------
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 47. 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-
centration 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, Porapak 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 48. 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 Porapak 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 tem-
perature .
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-
sorption. The 300°C temperature is also low enough to prevent the destruc-
tion of the packing material in the trap. The packing material which gave
the most reproducible results in the desorption experiments was Tenax GC.
For this reason and its stability at the 300°C desorption temperature, the
Tenax GC packing material was selected for use in subsequent experiments.
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
141
-------
TABLE 47. THE EFFECT OF COLD TRAPPING AT -196°C ON
CARBONYL SULFIDE AND METHYL SULFIDE WITH VARIOUS TRAP SIZES
Trap
loop ,m£
5.0
5.0
10.0
10.0
15.0
15.0
20.0
20.0
Trap
Flow
mVmin
81.2
81.2
81.2
81.2
81.2
81.2
81.2
81.2
Trap Inlet
in j .
1
2
Avg.
1
2
Avg.
1
2
Avg.
1
2
Avg.
Cone
COS
1.40
1.40
1.40
1.40
1,40
1.40
1.40
1.40
1.40
1.40
1.40
1.40
. , ppm
CH3SCH3
0.71
0.71
0.71
0.71
0,71
0.71
0.71
0.71
0.71
0.71
0.71
0.71
Peak
COS
66.7
66.7
66.7
67.8
68.1
68.0
67.8
68.1
68.0
78.0
78.9
78.5
Height
CH3SCH3
5.3
5.3
5.3
5.2
5.2
5.2
5.2
5.2
5.2
5.5
5.5
5.5
Trap Exit
Peak
COS
15.8
15.8
15.8
16.0
15.2
15.6
8.0
10.8
9.4
8.0
14.0
11.0
Height
CH^SCH^
0.8
0.1
0.5
0.0
0.0
0.0
0.1
0.1
0.1
0.0
0.1
0.05
142
-------
TABLE 48. THE EFFICIENCY OF VARIOUS MATERIALS
TRAPPING SULFIDES AT SEVERAL TEMPERATURES
Gas chromatograph Response-Peak Ht.
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.
Trap
Temp. °C
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
Before
Trap
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Tenax-GC
Chromosorb 102
Ch romos orb 10 2
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Chromosorb 102
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Porapak Q
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chromosorb T
Chormosorb T
Chromosorb T
COS
68.2
69.5
68.9
65.0
63.5
64.3
64.8
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
CH^SCH^
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
Trap size 2" x 3/8" OD
143
-------
was thermally desorbed into the GC analysis system. The gas chromatograph
trace obtained from the desorption 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 only 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 (NaHC03) has been reported to be very effective for this purpose.
Experiments have indicated that hydrogen sulfide (I^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 hydrogem 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 of 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 41.
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 satis-
factory separation of the organic sulfides.
144
-------
16 14 12 10 86
Retention time, minutes
Figure 41. Typical gas chromatograph trace of organic sulfides.
145
-------
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 42. 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 43.
In order to determine the efficiency of the collection of 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 44. As noted, only two of the
four peaks are above the detection limits, although all four organic sul-
fides and 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 45. 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 m£/min for seven mniutes 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 m£/min when the trap is maintained at -76°C. Higher flow rates were
tried, and a breakthrough into the 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 m£/min 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.
146
-------
_-_
.
- •
—
-
—
- -J
J
.
—
oi
-U)..
SB
IN
U'
.n3
3
,.s .
to---
(M
rq
UJ
a
,V;
- -
-•'•-
. ...
.::. ..
—
. . ..
.: :.
»•*
— — tn-
••*-
P*1
— _.
_
" ^-v
N
-IrfJ
.l/>....,
|*-^
'X}
O^
o--
— .
... .
.._
.
H
I
._).
—
— . .
(A'
r*r
•^%- •
R"
...
•
.. .
.„
....
.
...
t^r,. peBMEAT/OM BL6ND Date NOV 1, 111!
Innrument r>
Column l.f
Pocked with
on 80//00
RunlSO®
*_fi*-P«
Inlet ISC
Detector; ISt
Hyd. 40
( '~H/f>
Recorder /
Inieclion /g
Sampling Dg
'.::.
—
r^>
...
—
._ —
;._.
a 3**:
.'t,^
N*
mesh
•W O upera"or fj mrnVrllfWF
'f," no I.D. Tf^1
% »t. ^/^ Li
PA8APAK O&
/V//t °C. using _3L
r. /VA Rotameter P
^CCOLOMM
1. °c PpO
_pii» N/A Rotan
Mia V/A Bour
tataAf/A Rotan
_ln/min spend / n
All. indicated /O ;
vice £/)$ SflMPl
r-'
5T
-
(T>t
••
O
...I . ..
I
1
I
I
1
I
I
•
11
•_";;:-
1 1
1 1
.S 1 ._
. .u fj 1
If 1
II 1
~.-jM-—
i
|| 1
J\\ J
_cc/r
Ida. ,
nln. ^/A
Type
j. Phase
Support
Carrier
VIA
•*a'^2lO°eX'Jmtiue*
FVP8 (Ot
neter Re
neter Re
neter Re
nV.F.S.
ml net
bar!
o. -#O cc/min
a. 31O
to. ^Xl
cc/min
cc/min
50LT££ Type
//? m\
'-W6 VALVS
_
—
::...
~^
... ..
_.. _..
j
(•
|
' f
>
...
Actual
. ._ rfVJ
j
tj f3
19 18 17 16 15 14 13 12 11 10 9 8 76 54 32 1
Retention Time, minutes
Figure 42. Typical GC separation of organic sulfides on acetone-washed Porapak QS column.
-------
sui-Pioes oat. NOV Z? 1977
Operator P. SflUAl66»S
TPff
ft. fr" O-D-
Packed with /Vd % wt. V/l
on _
Run ISO ® _M&_- C- using 3O cc/min
osia. A04 Rotameter Rdg:
Typelotfierl
Rotameter Hdg.
3&O ec/min
cc/min
Type
( >_j/fl
Recorder / in/min »eed
Rotameter Rdg.
mV.F.S.
Injection_/jf2_ ul indicated /Q ul net
Sampling Device <5/iS £AM?i-l»/6 VAU/E
11 10 9 876
Retention time, minutes
Figure 43. Typical organic sulfide separation with Tenax-GC column.
148
-------
1 — • — i : —
• ; . .__ _ _^
I : 1 1
1 i
I f— >• —
• ** ' "*
1 " — * :
I j '
1 ; 1 : :
|_ : J ; 1
Sample PE*M SvST -7 DILUTED Date NOV 2. '177
Instrument PE 3<32OB Operator P. SAUN6eR&
Column A ft. Va O.D. I.D. TF£ Type
Packed with A/A %vn. A/ A Lia. Phase
on 66/ffOmesh T£AfJt X - 6C. Sujpon
Run ISO 9 HA °C. usina ~3O cc/min. H&. Carrier
@ "7S" psia. A/A Rotameter Rda. A/A
Inlet I6O ^C 3O0-*r I8>00@ a°/fllin Type
Detsctar: /6O C FrD Type (other) nff
Hyd. ^O psig /V>9 Rotameter Rdg. 4O cc/min
Air 7^ psiq AfA Rotameter Rdq. ^£ ^ cc/min
( ) Af/t psia AVJ Rotameter Rda. A/A cc/min
Recorder / < M I indicated / «J net /O «Kl Actual
Sampling Device lS/?S •SflfHPLlNGr VALV£
— _ . -• _
4... ..
- ; :
. . ... _L
- — -. I- . -
, _
!
, _, ;
: .
— — —i-
|~--ii-j : ^ r — r--- — — — ±-~
•• * . ' ... ' ' ' • '
• • - : '. . . . ' . \
• :''': : i : • '
1 i ' . ; i : i ' '
• TJT n i ' ' -*- Jl
.,— g - ,'-'•
4-t ' i
s) . . , ; •
M
u
>.
r
0)
te
n
; _
i
.
.
•
,
: . i
. a)
20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2
Retention Time, minutes
Figure 44. Organic sulfide permeation blend with secondary
dilution, near detection limit of GC FPD system.
1 0
149
-------
Sample
.oat.<
Instrument Pg 'ZQZO S Operator £.
Column A ft. Vfl 0.0. I.D.
Packed with /yA % wt. fJA
&O/30mesh TEN A1 -
. __~^r^~ ^« > ^ a A / ^ ^
Run ISO ® V/»°C. using
@ 7S~ psio. X/XJ Rotameter Rdg.
Inlet 16,0 ~Vtr~30°-7iaQ'
Detector: 1 1* O °C
Hvd. ^O Piig
Air 7O p>ig
Mir / w L>»IU /»^r Rotametar Rda.
( I A^^ psig fit ft Rotameter Rda.
Recorder /
Uq. Phase
_ ^__ Support
cc/min. //A. Carrier
Type (other) A/A
Rotameter Rdn. 4/Q cc/min
_../min speed / ™v F a SOLT& C.
Injection /g ml indicated IO Ml net IO Al Actual
ction /£? ml indicated /O »il net
Sampling Device f?^5 SAftlfi/M
20 19 18 17 16 15 14 13 12 11 10 9 87 65 4 3 2 1
Retention Time, minutes
Figure 45. Organic sulfide permeation blend with secondary
dilution, concentrated on Tenax-GC trap and thermally desorbed
into GC FPD system.
150
-------
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 an an appendix
to this 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 m£/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 posion 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 m£/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 effiency 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 m£/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.
151
-------
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 49.
The percent deviation varies from 1 percent for methyl sulfide to 6 percent
for methyl disulfide. This deviation appears to increase with decreasing
concentration or organic sulfide.
TABLE 49. 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 repeatability 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 Tabel 50. Standard percent deviations ranged from
7 percent of 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 51 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.
152
-------
TABLE 50. TRAP REPEATABILITY FOR ORGANIC SULFIDE COLLECTION
Test 1
Test 2
Test 3
Test 4
Test 5
Average
Standard
Deviation
Percent
Deviation
TABLE 51.
Test 1
Test 2
Test 3
Test 4
Test 5
Average
Standard
Deviation
Percent
Deviation
COS
Area
20241
22052
20092
22343
17378
20421
±1985
9.7%
TRAP- TO- TRAP
COS
Area
47590
46830
44465
50590
25788
43053
±9896
23.0%
Me2S
Area
71938
66098
74326
63777
65494
68327
±4548
6.7%
REPEATABILITY
Me2S
Area
68995
43429
66874
63440
76180
65784
±8332
12.7%
Et2S
Area
34417
35156
38123
30508
32516
34144
±2864
8.4%
FOR ORGANIC
Et2S
Area
25716
20549
17283
30144
16631
22065
±5773
26.2%
Me2S2
Area
41346
39537
41683
34147
34321
38207
±3718
9.7%
SULFIDE COLLECTION
Me2S2
Area
13839
11487
10069
15901
10338
12327
±2491
20.2%
153
-------
Figures 46-49 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 gives 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 hydrogem 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 for 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 spearated 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 Mercedes 240D. Hot
FTP (23minute test) driving cycles were followed to generate exhaust for the
vehicle baseline emissions and for the tunnel injection plus vehicle experi-
ments. An aluminum cylinder containing 4-8 ppm of each of the organic sul-
fides in balance nitrogen was used as the source for the organic sulfides.
The cylinder was named by comparing GC peak areas with the GC peal; areas of
the organic sulfides generated by the permeation system. The flow of organic
sulfides into the tunnel was regulated to give a concentration of 5-10 ppb
of each of the organic sulfides in the dilution tunnel. Injections of the
154
-------
en
en
360 f—
320 -
100 200 300 400 500 600 700
GC Peak Area X0.01
800
900
1000
1100
1200
Figure 46. Carbonyl sulfide linearity plot,
-------
cn
o
at
I
360 -
320 -
280 -
240 -
200 ~
160 -
120 _
80
40
I
I
I
100 200 300 400 500 600 700
GC Peak Area X0.01
800
900
1000
1100
1200
Figure 47. Methyl sulfide linearity plot.
-------
360 I-
en
100 200 300 400 500 600 700
GC Peak Area X0.01
800
900
1000
1100
1200
Figure 48. Ethyl sulfide linearity plot%
-------
Ol
CO
100
200 300
GC Peak Area X0.01
400
500
600
Figure 49. Methyl disulfide linearity plot .
-------
organic sulfides into the tunnel without exhaust gave recoveries that varied
from approximately 93 percent for methyl disulfide to 115 percent for ethyl
sulfide (Table 52) . An interfering peak in the GC analysis for methyl
sulfide voided the tunnel recovery experiments for this compound. The per-
cent deviation of the recovery percentages ranged from 23-60 percent. This
value is higher than the expected 25% due to the trap-to-trap variations
found in the validation experiments. The recovery of the organic sulfides
with real exhaust varied from 7 percent for ethyl sulfide to 57 percent for
carbonyl sulfide (Table 53). The baseline emissions of carbonyl sulfide and
inethyl sulfide were erratic and of equal magnitude to the carbonyl sulfide
and methyl sulfide injected into the tunnel. This variation of carbonyl
sulfide and methyl sulfide from the vehicle, along with tunnel memory for
carbonyl sulfide and methyl sulfide and trap-to-trap variations, made the
percent recovery calculations very difficult and thus gave the resulting 37
and 55 standard deviations. Baseline emissions for ethyl sulfide, and methyl
disulfide were insignificant and did not affect the recovery experiment.
There is little loss of the organic sulfides in the CVS tunnel with-
out exhaust, however a 40 to 90 percent loss with exhaust in the CVS tunnel
can be expected. These losses must be taken into account in determining
organic sulfide 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 mJl/minute. The procedure has a minimum detec-
tion 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 re-
sponse for the organic sulfides in the 0.2 to 25 ppb range. If the concen-
trations of the organic sulfides exceed this range in dilute exhaust, a lower
sampling flow rate (less than 130 m£/minute) must be used to keep the detec-
tor 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
procedure. The ethyl sulfide concentration is affected by this interference.
This is a significant loss of the organic sulfides in the CVS tunnel with
exhaust. These losses must be taken into account in determining the concen-
tration of the organic sulfides when using this procedure.
Overall, the organic sulfide procedure should provide a relatively ac-
curate method for determining the concentrations of the organic sulfides in
dilute exhaust, and its use is recommended for this project.
159
-------
TABLE 52. PERCENT RECOVERIES OF THE ORGANIC SULFIDES
FROM THE CVS TUNNEL ONLY
Carbonyl Sulfide (8 ppb)
Percent Recovery
83
127
82
132
70
Methyl Sulfide (6 ppb)
Percent Recovery
Results voided
due to inter-
fering peak in
the GC analysis
Average
% Recovery
Standard
Deviation
99%
28%
Ethyl Sulfide (5 ppb)
Percent Recovery
98
139
120
132
85
Average
% Recovery
Standard
Deviation
115%
23%
Methyl Disulfide (10 ppb)
Percent Recovery
31
100
43
100
190
Average
% Recovery
Standard
Deviation
93%
60%
160
-------
TABLE 53. PERCENT RECOVERIES OP THE ORGANIC SULFIDES
FROM THE CVS TUNNEL AND EXHAUST
Carbonyl Sulfide (8 ppb)
Percent Recovery
20
42
119
52
54
Methyl Sulfide (6 ppb)
Percent Recovery
12
27
144
18
34
Average
% Recovery
57%
Average
% Recovery
47%
Standard
Deviation
37%
Standard
Deviation
55%
Ethyl Sulfide (5 ppb)
Percent Recovery
2
5
15
5
10
Methyl Disulfide (10 ppb)
Percent Recovery
5
5
25
1
17
Average
% Recovery
Standard
Deviation
7%
5%
Average
% Recovery
Standard
Deviation
11%
10%
161
-------
SECTION 11
PHENOL PROCEDURE
LITERATURE SEARCH
Phenols are compounds of the general formula ArOH, where Ar is phenyl
or substituted phenyl. Phenols differ from alcohols in having the hydroxyl
group, -OH, attached directly to an aromatic ring. Ring substitution by
alkyl, hydroxyl or carbonyl groups creates the variety of different phenols
possible in automotive exhaust. Phenols generally have high vapor pressures,
are colorless and, except for phenol, are insoluble in water. Some of the
physical properties of phenols possible in exhaust are shown in Table 54
below,
TABLE 54. PHYSICAL PROPERTIES OF PHENOLS POSSIBLE IN EXHAUST
Phenol
Phenol
Salicylaldehyde
m-cresol
p-cresol
2,3-xylenol
3,5-xylenol
p-ethyIphenol
2-i sopropyIphenol
2,4„6-trimethyIphenol
2,3,5-trimethyIphenol
2,3,5,6,-tetramethyIphenol
Internal standard-
o-chloiqphenol
Molecular
Weight
94.11
122.13
108.15
108.15
122,17
122.17
122.17
136.20
136.20
136.20
150.22
128.56
Boiling
Point,PC
182
197
202
202
218
220
219
213
221
230
247
175
Freezing
Point,°C
43
-7
12
35
75
68
47
15
72
92
118
9
Density,
1.0722
1.1674
1.0336
1.0178
0.9680
1.012
1.2634
The slightly acidic nature of phenols (Ka =< 10~ ) makes them soluble in
aqueous hydroxides yet not acidic enough to be soluble in a bicarbonate so-
lution. This property allows phenols to be separated from non-acidic com-
pounds by collection in base and from organic acids by their insolubility
in bicarbonate. The acid base equilibrium that occurs is shown below.
162
-------
The purpose of measuring phenols in exhaust is to determine if they are
present in sufficient quantities to cause health problems. A number of pro-
cedures have been published that are used for determining concentrations of
phenols. These include colorimetric or spectrophotometric methods, gas
chromatography, liquid chromatography and derivatization with subsequent
analysis by gas chromatography. Colorimetric and spectrophotometric methods
of phenol analysis (152-154) are primarily used for total phenol measurement.
This method is unacceptable because individual phenol concentrations are
desired. One liquid chromatography procedure investigated includes the
formation of fluorescent dansyl phenol derivatives which are subsequently
analyzed by a liquid chromatograph (LC) equipped with a fluorescence detec-
tor (155). The procedure is too time consuming to warrant its use. Another
liquid chromatograph technique is difficult to set up, is very involved
chemically and suffers from interferences in one of the reagents (156).
Several procedures were available in which derivatives of phenols were pre-
pared for analysis on a gas chromatograph (GC) . In one method, phenols were
alkylated over an aluminum phosphate catalyst, acetylated and analyzed on a
GC (157) . This procedure is very involved and recoveries of phenol are not
high. Several methods in which ester (158-160) and ether (161-163) deri-
vatives of phenols are produced were found in the literature search. The
production of ether derivatives of phenols and analysis by GC seemed to be
a promising method for determining the concentrations of individual phenols.
A number of GC methods not involving derivatization were also studied. Some
of these listed a variety of columns and GC instrument parameters for phenols
analysis (164-173). Phenols can be sampled from exhaust in several ways.
Activated carbon filters have been used to absorb phenol from aqueous samples
(174) and from air (175) . However, a more suitable sampling procedure for
'dilute exhaust involves collection in a hydroxide solution in impingers.
Several authors have suggested this means of removing phenols from exhaust.
Collection of phenols in aqueous hydroxide is usually followed by wet chemi-
cal workup and analysis by GC. Aqueous phenol samples are treated with a
variety of steps including acidification, extraction with an organic solvent,
distillation and extractions to remove impurities (176-179).
The procedures that appeared to be the most promising are those using
a GC for phenols analysis. Samples can be collected in impingers containing
aqueous KOH and workup can be accomplished by forming ether derivatives or by
extracting with ether (176) .
PROCEDURAL DEVELOPMENT
The procedure chosen for the collection and analysis of phenols required
a considerable amount of procedural development. Extraction, analytical and
sampling parameters needed to be determined prior to exhaust sample pro-
cessing.
The first factor investigated, extraction efficiency, was found to de-
pend on a number of variables. Type of solvent, number of solvent extrac-
tions, pH of aqueous sample and method of solvent evaporation all affected
the extraction efficiency. Two sets of spiked phenol samples were extracted
with two solvents, methylene chloride and ethyl ether. Between one and five
163
-------
consecutive extractions were performed and each set of samples using each
solvent. The amount and percent of phenol recovered by these extractions
is shown in Table 55 and 56. These figures indicate that of the phenol
recovered, most or all of it is recovered in the first two solvent extrac-
tions. However, four times as much phenol is captured in the second extrac-
tion with methylene chloride (8.0 percent) as is captured in the second
extraction with ether (1.9 percent). The average recoveries calculated for
the two solvents are probably low due to the fact that the averages include
cases when only one solvent extraction was performed. Taking into consider-
ation the large difference in extraction efficiencies between the two sol-
vents (67.6 percent with ethyl ether and 49.9 percent with methylene chlo-
ride) , ether was chosen as the organic solvent for extracting exhaust samples.
It is possible that the slightly lower boiling point of ether compared to
methylene chloride (34°C vs 40°C) allows it to be boiled off at a lower
temperature, thus preventing the evaporation of the lower boiling phenols.
Another factor influencing extraction efficiency was investigated. This
was the pH of the phenol spiked aqueous solution. The extraction efficiency
was found to be unaffected by the pH of the solution when the pH was neutral
or acidic (pH < 7) . Table 57 lists the amount and percent phenol recovered
when the spiked aqueous solution was varied from a pH of one to seven.
TABLE 57. EXTRACTION EFFICIENCY AS A FUNCTION OF pH
OF AQUEOUS SOLUTION
pH of Phenol Cone. Percent
Aqueous Recovered Spike Phenol
Solution yg/m£ yg/m£ Recovered
1 34 67 51.8
2 36 67 53.8
3 32 67 47.8
4 28 67 41.8
5 34 67 51.8
6 44 67 65.7
7 32 67 47.8
The fourth factor affecting extraction efficiency that was studied was
the means of solvent removal and sample concentration. The method producing
the highest phenol recoveries involved a two step process using a Kuderna
Danish concentrator heated by a steam bath (45°C) for initial volume reduc-
tion and a desiccating chamber modified for dry nitrogen flow for final con-
centration. Several sample concentrating techniques were tested before it
was determined that erratic phenol recoveries occurred when samples were
dried solely by heating in a Kuderna concentrator. Phenol recoveries for
several samples evaporated to 0r 1/2, 1, 2 and 5 m£ in the Kuderna concen-
trator are listed in Table 58. The trend toward increasing phenol recovery
with larger final volumes in the Kuderna concentrator is apparent from the
data presented in Table 58. Concentrating samples to a desired volume,
however, proved to be a difficult task using the Kuderna concentrator. Due
to the tapered tip on the concentrator, the solvent level changed rapidly
when the volume decreased to 5 m£ and less. It was necessary, therefore, to
164
-------
TABLE 55. EXTRACTIONS WITH METHYLENE CHLORIDE
Phenol Recovered, yg/m£
Extraction #
Sample
1
2
3
4
5
6
7
8
9
S! 10
11
12
13
14
15
16
17
18
19
20
21
1
16
24
26
24
8
23
17
36
41
37
24
_
_
—
_
-
-
_
-
_
-
2
2
2
3
-
-
-
-
-
-
-
-
9
24
37
-
-
-
—
-
-
-
345
0 - -
0 - -
0 - -
_
_
_
_
_
_
_
_ _ _
_ — _
_ _ _
_ _ _
21
21
33
24
25
32
33
Cone.
Spike
yg/m£
29
29
29
54
54
54
54
67
67
67
67
67
67
67
67
67
67
67
67
67
67
1
55.2
82.8
88.2
44.4
14.8
42.6
31.5
53.7
61.2
55.2
35.8
_
_
-
_
-
-
_
-
-
-
Percent Phenol Recovered Total
Extraction # Percent
2345 Recovered
6.9 0 - - 62.1
6.9 0 - - 89.7
10.3 0 - - 98.5
- - - - 44.4
14. 8a
42.6
- - - - 31.5
- - - - 53.7
- - - - 61.2
- - - - 55.2
- - - - 35.8
13. 4b - - - 13. 4a
35. 8b - 35.8
55. 2b - 55.2
31. 3b - - 31.3
31. 3b - - 31.3
49. 2b - 49.2
35. 8b - 35.8
37. 3b - 37.3
47. 8b 47.8
49. 2b 49.2
Average 49.9
Samples taken inadvertently to near dryness.
All extractions combined to give a total concentration.
-------
TABLE 56. EXTRACTIONS WITH ETHER
Phenol
Recovered, yg/m£
Extraction #
Sample
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
1
38
23
6
40
31
38
42
49
45
52
43
18
29
26
30
2
1
4
0
0
1
0
1
-
-
-
-
-
-
-
-
3 4
0 0
6 3
0
-
-
-
-
-
-
-
-
-
-
-
-
Cone.
Spike
yig/m£
54
54
54
54
54
54
54
54
54
54
54
54
54
54
54
Percent Phenol
Recovered
Extraction #
1
70.4
42.6
11.1
74.1
57.4
70.4
77.8
90.7
83.3
96.3
79.6
33.3
53.7
48.2
55.6
2
1.9
7.4
0
0
1.9
0
1.9
-
-
-
-
-
-
-
-
3
0
11.1
0
-
-
-
-
-
-
-
-
-
-
-
-
4
0
5.6
0
-
-
-
-
_ 'I
-
-
-
-
-
-
-
Average
Total
Percent
Recovered
72.3
50. Oa
11. lb
74.1
59.3
70.4
79.7
90.7
' 83.3
96.3
79.6
33.3
53.7
48.2
55.6
67.6
a Phenol contamination in all four extractions; total percent recovered represents
extractions 1 and 2.
Sample inadvertently taken to near dryness.
-------
TABLE 58. EFFECT OF REDUCING SAMPLE VOLUME BY KUDERNA DANISH
CONCENTRATOR ON PHENOL RECOVERY
Evaporative
Volume, m£_
5
2
1
1/2
0
Final
Volume
5
2
2
2
2
Phenol
Recovered
15
32
23
20
14
Cone.
Spike
yg/m£
21
52
52
52
52
Percent
Phenol
Recovered
71.4
61.5
44.2
38.5
26.9
find another method of sample concentration that did not require constant
attention. The second drying method attempted involved the use of the
Kuderna concentrator and a tray of heated sand equipped with a dry nitrogen
outlet. The samples were concentrated to 5 m£ in the Kuderna concentrator,
transferred to a 10 m£ beaker and then further concentrated with a stream of
dry nitrogen (while being gently heated with the sand) . This method was un-
successful due to water condensation on the beaker and nitrogen blowing sand
into the beaker. The warm sand tray was abandoned as a means of drying
phenol samples in favor of a desiccating chamber modified for the flow of
dry nitrogen. The samples were concentrated to 5 m£ in the Kuderna concen-
trator and transferred to 10 m& beakers as was done previously. The samples
were then concentrated to approximately 1 m£ in the desiccating chamber by
directing dry nitrogen into the beakers with a gas manifold. Water conden-
sation was no longer a problem because the molecular sieve/silica gel absor-
bant in the chamber absorbed any moisture that was present. The drying pro-
cess could also be easily observed through the glass window and stopped when
necessary. This last procedure was the one adopted for the concentration of
extracted phenol samples in ether due to its simplicity and lack of inter-
ferences.
A second parameter (in addition to extraction efficiency) affecting the
workup of phenol exhaust samples was investigated. This factor was chemical
interferences to phenol recovery. The source of interferences could be
contaminants in the various reagents used in sample collection or extraction
or interfering exhaust compounds trapped in hydroxide solution along with
the phenols. Several blank extractions were performed with methylene chlo-
ride and with ether using all solutions that would normally be used for
exhaust sample extraction. None of the samples produced measurable levels
of phenols. Possible interfering compounds in exhaust that may be absorbed
into the scrubber solution, 1 N KOH, are neutral hydrocarbons and organic
acids. A set of tests were performed in which 1 N KOH samples spiked with
diesel fuel were extracted in several ways. In the first experiment 1 y£
of diesel fuel was added to acidified 1 N KOH spiked with phenol and the
resulting sample was extracted and analyzed. The second test was conducted
similarly to the first except that a cyclohexane extraction was performed on
the basic solution to remove neutral hydrocarbons before acidification, ex-
traction and analysis. In the third extraction an acidified 1 N KOH solution
was spiked with 1 yfc of diesel fuel (no phenol) , extracted and analyzed.
167
-------
Several regular extractions were also performed on phenol spiked 1 N KOH
samples. The results, shown in Table 59 indicate that diesel fuel does not
interfere with phenol recovery when present by itself. However, when both
phenol and diesel fuel were present, an approximate 16 percent phenol loss
occurred. The loss increased to 30 percent when a cyclohexane extraction
was performed to remove diesel fuel. Additional extractions were performed
TABLE 59. EFFECT OF DIESEL FUEL ON RECOVERY OF PHENOL
Phenol Phenol Percent
Recovered Added Phenol
Sample Extracted (in 1 N KOH) \ig/mH yig/m£ Recovered
1. Phenol + diesel fuel 21 30 70.0
2. Phenol + diesel fuel +
cyclohexane extraction 16 29 55.2
3. Diesel fuel 0 0
4. Phenol 24 28 85.7
on an actual exhaust sample and on phenol spiked 1 N KOH samples to determine
the validity of the data obtained for Table 59. One half of an aqueous ex-
haust sample was acidified, extracted and analyzed. The remaining half of
sample was first extracted with ethyl ether to remove netural hydrocarbons.
Then the ether was extracted with 0.5 N NaOH to recover any phenol extracted
into the ether. The aqueous portions were combined, acidified, extracted
and analyzed the same as the first half. The same amount of phenol was re-
covered from each half of the sample. Also, neither of the sample halves
contained compounds that could interfere with the GC analysis of the phenols.
See Table 60. Apparently, either no neutral hydrocarbons survive the normal
extraction process or else these compounds are eluted under the solvent peak
during GC analysis of the phenol sample. The two 1 N KOH samples spiked
with phenol that were extracted for neutral hydrocarbon removal had an aver-
age phenol recovery 70 percent less than samples spiked and extracted nor-
mally. The data obtained from the experiments conducted to determine neutral
hydrocarbon interference produced conflicting results. However, since the
exhaust sample showed no evidence of interference from such compounds, it
was decided not to incorporate a neutral compound removal step into the pro-
cedure.
Interference, to phenol recovery or analysis due to the presence of
organic acids was also studied. The modification to the procedure for re-
moval of organic acids included an additional NaHCO-j extraction of the ether
containing the phenols and a back extraction of the aqueous NaHCO-, layer
with ether. Organic acids are more soluble in an alkaline aqueous solution
than in ether. However, phenols being acidic, tend to be drawn into the
NaHCO3 layer along with the organic acids. The back extraction with ether
was to recover phenols that may have been extracted into NaHCO3. Two phenol
spiked 1 N KOH samples were extracted following the modified procedure to
determine the presence of organic acids. The phenol level of the phenol
spiked 1 N KOH samples averaged 567 percent higher than samples extracted
normally. Half of each exhaust sample was extracted normally and half was
168
-------
TABLE 60. INTERFERENCES TO PHENOL RECOVERY OR ANALYSIS
Phenol Recovered Percent Phenol
Phenol Spiked 1 N KOH Samples yg/m.g, __ Recovered3
Normal extractions 3
*..'
Hydrocarbon modified samples 1 33
Organic acid modified samples-" 17 567
a
Percent phenol recovered is relative to the samples extracted normally.
Exhaust Phenol Recovered Percent Difference
Sample Extraction Method ug/m£ Between Halves
1 1/2 normal 1
1/2 HC modified 1 °
2 1/2 normal 35
1/2 org. acid modified 35 °
3 1/2 normal 41
1/2 org. acid modified 35
extracted with the modification to the procedure. The concentration of phe-
nol recovered from each sample half of the first exhaust sample was the same.
The second exhaust sample did not agree as closely as the first, though, pro-
bably due to an error in the extraction process. The sample half extracted
for removal of organic acids yielded 15 percent less phenol than the sample
half extracted normally. Since neither exhaust sample showed any evidence '
of interference from organic acids, the procedure was not modified for re-
moval of organic acids. The results of the organic acid interference tests
are shown in Table 60.
The next set of parameters that needed to be determined were those
governing the analytical portion of the phenol procedure. The instrument
parameters for the gas chormatograph (GC) and an analytical column for se-
paration of phenols needed to be selected. Also, phenols recovered from
exhaust needed to be identified and the response factors calculated. Sev-
eral different columns were installed in a Perkin-Elmer 3920 GC equipped
with a flame ionization detector (FID). These included an SE-30 WCOT glass
capillary column, a 10 percent OV-101 on 100/120 mesh Gas-Chrom Q Teflon
column, a 20 percent DECS on 80/100 mesh Chromosorb W-HP Teflon column and
a Teflon column packed with 10 percent OS 138/H3P04/SP-1200 on 100/120 mesh
Chromosorb W AW. The last column packed with 10 percent OS 138 provided
the best separation of phenols of all columns tested. A variety of tem-
perature programming sequences were experimented with on the GC. The most
efficient separation of solvent peak from phenols plus an analysis time of
169
-------
less than an hour were obtained with a temperature program of 4°/min from
70°C to 170°C. The temperature is initially held isothermally for two minutes
at 70°C. The injector and interface temperature are maintained at 200°C.
Phenols in exhaust samples were identified by comparing the retention
times to individual standards and standard blends. The concentrations of
phenols in the blend used as the external standard need to be close to the
phenols concentrations found in exhuast due to the fact that both retention
times and response factors vary with concentration. Response factors, which
correct for the different responses of each phenol to the FID, are calculated
from the concentration and counts of each phenol in the external standard
relative to o-chlorophenol (also in the external standard). The concentra-
tions of phenols in exhaust samples are computed by comparing the area of
each phenol to the appropriate response factor and to the area and concen-
tration of the internal standard, o-chlorophenol. Using an external and an
internal standard proved to be the easiest and most accurate method of cal-
culating phenol concentrations in exhaust.
The last set of parameters that needed to be determined as part of the
procedural development were those relating to the sampling of exhaust for
phenols. The procedure chosen for phenols analysis required that phenols be
present in an aqueous solution at the start of the extraction process. Since
phenols are acidic and therefore soluble in base, it was decided that dilute
exhaust would be bubbled through 1 N KOH in glass impingers. The phenols
collected in this manner could be extracted directly. The number of impin-
gers and the flowrate of dilute exhaust passing through the impingers that
would trap the most phenols needed to be determined. Initially, experiments
were conducted with three tapered tip impingers connected in series. Each
impinger contained 25 m£ of 1 N KOH chilled to ice bath temperature. Exhaust
was pumped through the impingers at 4 £/min. The samples thus obtained were
extracted and analyzed for phenols, however, no phenols were found. In an
effort to trap more phenols dilute exhaust was passed through larger Green-
burg-Smith impingers at a higher flowrate (0.7-0.8 ft3/min). Each of the
three impingers contained 200 m£ of 1 N KOH instead of 25 mH. Measurable
levels of phenols were extracted from exhaust under the latter conditions.
Additional tests regarding the choice of sampling parameters is shown in the
Validation Experiments section.
VALIDATION EXPERIMENTS
Several experiments were performed to show that the phenol procedure is
a valid method for processing exhaust samples containing phenols. The sam-
pling parameters providing the best trapping efficiency were determined.
Dilute exhaust is allowed to flow at 0.7-0.8 ft3/min through two Greenburg-
Smith impingers in series. Each impinger contains 200 mi of 1 N KOH chilled
to ice bath temperatures (0-5°C).. From the results shown ir. Tables 55 and 56
it is obvious that no phenol is captured in bubblers two and three. However,
several other phenols (salicylaldehyde, m-cresol, p-cresol, 2,3-xylenol,
3,5-xylenol, etc.) are found in small quantities in the second impinger. For
this reason, two bubblers are used to collect phenols. Before pacsing through
the impingers the dilute exhaust flows through a heated sample line (375°F),
170
-------
a Pallflex filter and another sample line heated to 175°F. Tests were con-
ducted with and wirhout a filter. The sample line without a filter was
heated to 175°F from the CVS to the impingers. Results from the qualifica-
tion tests in Tables 62 and 63 show that from 4% to 60% more phenol is re-
covered from filtered exhaust than from unfiltered exhaust. For this reason
exhaust is filtered before sampling for phenols.
The next set of parameters that needed to be determined were those that
would give the highest recovery of phenols from the extraction process. It
was found that two ether extractions of the contents of impinger one, and
one ether extraction of the contents of impinger two gave good recoveries.
Also, better results were obtained when the final drying step was done with
dry nitrogen instead of with heat. Extraction efficiency of the phenols
procedure is approximately 68%.
Two experiments were performed to validate the analytical protion of
the phenol procedure. The first involved the analysis of phenol stnadards
in the concentration ranges expected in exhaust samples. Calibration curves
were drawn from the data and they are shown in Figures 50-56. The linearity
ranges of the internal standard and of the phenols found in exhaust vary
between 0-50 yg/m£ and 0-200 yg/m£. The range for each phenol is listed in
Table 61 below. The concentrations of phenol recovered from exhaust are well
within the linearity range of each phenol.
TABLE 61. LINEARITY RANGES OF INTERNAL STANDARD AND
OF PHENOLS IN EXHAUST
Phenol Linearity Range (yg/m£)
o-chlorophenol 0-120
phenol 0-50
salicylaldehyde 0-120
m-cresol and p-cresol 0-80
p-ethylphenol, 2-isopropylphenol,
2,3-xylenol, 3,5-xylenol and
2,4,6-trimethylphenol 0-200
2,3,5-trimethylphenol 0-100
2,3,5,6,-tetramethylphenol 0-120
Injection variability was studied as another validation test for the
phenols procedure. A 12.2 yg/m£ phenol standard in methylene chloride was
injected five consecutive times. The area of each injection is shown in
Table 62. The standard deviation is 107 and the percent variations is 2.20%
for the five injections.
171
-------
60
50
40
o
o
30
u
20
10
60 120 180
Concentration (ug/m£)
240
300
Figure 500 Linearity of o-chlorophenol GC response,
172
-------
60
50
40
o
o
•
o
X
(8
-------
60 ,-
50
40
o
o
o
X
(0
-------
60.-
o
o
(0
0)
•s
0)
n<
u
40
80 120
Concentration (yg/m£)
160
200
Figure 53. Linearity of m-cresol and p-cresol GC response,
175
-------
140 t-
130
120
110
100
90
g 80
o
•
o
* 70
QJ
a
-------
o
o
I
o
X
tfl
0)
0)
dl
u
c
60,-
50
40
30
20
10
20 40 60
Concentration
80
100
Figure 55. Linearity of 2,3,5-trimethylphenol GC response.
177
-------
60
50
40
o
o
•
o
X
10
30
-------
TABLE 62. INJECTION VARIABILITY OF PHENOL
Sample Area
1
2
3
4
5
Average
Standard Deviation
Percent Variation
QUALIFICATION
The phenol procedure was qualified by injection of an aqueous phenol
solution into the exhaust of a Mercedes 240D diesel during three successive
FTPh driving cycles. The percent recoveries from the tests represent the
amount of phenol that is expected to survive the trip through the dilution
tunnel to the sampling impingers. The test sequence consisted of FTPh
driving cycles with ten minute soaks in between. Base line phenol emission
levels were measured for three consecutive FTPh driving cycles and during
three additional FTPh driving cycles phenol was injected into the exhaust.
Two sets of impingers sampled dilute exhaust during each test. The sample
line leading to the first set was heated to 175°F and no filter was used.
The second sample line was heated to 375 °F up to a Pallflex filter and 175°F
from the filter to the second set of impingers. The average results from
the three baseline emission tests showed that filtered exhaust produced a
higher phenol concentration (24 ug/m3) than unfiltered exhaust (12
The data is found in Table 63 below.
TABLE 63. BASELINE PHENOL EMISSION LEVELS FROM
MERCEDES 240D DIESEL
yg/m3 Phenol
Test Unfiltered Filtered Difference
1 10 29 19
2 10
3 15 19 4
Avg 12 24
The difference between filtered and unfiltered exhaust was also apparent in
the results from the injection of phenol into exhaust. The data in Table 64
shows that the filtered exhaust yielded 60 percent and 17 percent more phenol
than unfiltered exhaust. No phenol was recovered from the filtered line
second phenol injection. This was probably due to sample loss during the
179
-------
TABLE 64. PERCENT RECOVERIES FROM INJECTION OF PHENOL INTO EXHAUST OF
MERCEDES 240D DIESEL
Unfiltered Filtered Difference
First Phenol Injection 52.4 112.7 60.3
Second Phenol Injection 57.7
Third Phenol Injection 76.4 93.7 17.3
Average 62.2 103.2
extraction procedure. A gradual trend towards increasing phenol recoveries
appears to occur with samples that flowed through the unfiltered sample line.
This may be due to phenols being initially absorbed onto particulate coating
the sample line. The particulate removes phenol from the gas stream until
it is saturated. Gradually less phenol is absorbed and therefore, more is
recovered in the impingers. What appears to be a trend, however, may also
be the expected variability in recoveries. The greatest difference in un-
filtered recoveries is 24 percent and in filtered recoveries it is 19 percent.
The average phenol recovery of unfiltered samples is 62.2 percent and the
average phenol recovery of filtered sample is 103.2 percent. Assuming all
phenols in exhaust can be removed with similar efficiencies, quantitative
recoveries of phenols in exhaust diluted by the CVS can be expected.
The injection of phenol into the exhaust of the Mercedes was accom-
plished by means of a Baird atomizer attached to an opening on the CVS tunnel.
An aqueous phenol solution (0.7 g/m£)was dripped into the funnel of the atom-
izer from a 50 m£ buret. Air pressure applied through the side arm of the at-
omizer sprayed the phenol solution into the tunnel where it mixed with exhaust.
Any solution that was not dispersed into the tunnel was captured in an
Erlenmeyer flask containing the mister. This remaining portion of phenol
solution was extracted and analyzed as usual. The amount of phenol injected
was calculated by subtracting the micrograms of phenol in the remaining
phenol solution from the micrograms delivered from the buret. Percent re-
coveries were computed by comparing the amount of phenol recovered to the
amount injected.
RESULTS AND CONCLUSIONS
The method chosen for measuring phenols in dilute exhaust involves col-
lection in aqueous KDH, extraction with ether and analysis on a GC equipped
with a flame ionization detector. Dilute exhaust is bubbled at 0.8 ft3/ndn
through two Greenburg-Smith impingers each containing 200 m£ of 1 N KOH
chilled to ice bath temperatures. The exhaust is heated to 375°F and is
filtered through a Pallflex filter to remove particulate. The phenol samples
are acidified, extracted two consecutive times with ethyl ether and concen-
trated. The extracts from impingers one and two are combined, further con-
centrated and spiked with the internal standard, o-chlorophenol, before
analysis with the GC. The temperature programming sequence starts with an
180
-------
isothermal hold at 70°C for two minutes followed by programming to 170°C
at 4°/min. Total GC analysis time is about 30 minutes. The injector and
interface temperatures are maintained at 200°c. A Teflon column packed
with 10% OS 138/H3PC>4/SP-1200 on Qiromosorb W AW is used for separating
phenols. One microliter of the external standard and 1 yjl of each sample
is injected into the GC. The data obtained from the GC computer system is
used to calculate concentrations of phenols.
The linear range of each phenol found in exhaust and of the internal
standard was determined. The concentrations of phenols fall well within the
linear range. Should a sample be too concentrated it can be diluted volu-
metrically to a level within the linear range.
Several factors contribute to the overall recovery of phenols from ex-
haust. These include the stability of phenols traveling from automobile to
impingers and the trapping and extraction efficiency of phenols. The results
from the qualification tests indicate that approximately 100 percent of
phenol injected into exhaust is recovered. One hundred percent phenol is
also captured in two impingers connected in series. Extraction efficiency,
however, is only about 68 percent. This low value is probably due to losses
encountered in the drying process. Injection variability of phenol into the
GC was only 2.2 percent for a series of five injections. Similar results
are expected for the other phenols found in exhaust.
Several methods for the determination of phenols in automobile exhaust
were combined and adapted to the needs of this project. The resulting pro-
cedure used to measure phenols is sensitive to about 1 yg/m£. The phenols
in order of elution are phenol; salicylaldehyde; m-cresol and p-cresol;
p-ethylphenol, 2-isopropylphenol, 2,3-xylenol, 3,5-xylenol and 2,4,6-
trimethylphenol; 2,3,5-trimethylphenol and 2,3,5,6-tetramethylphenol. Over-
all this procedure should provide a relatively accurate method for determin-
ing the concentrations of the phenols in dilute exhaust, and its use is re-
commended for this project.
181
-------
SECTION 12
THE QUALIFICATION EXPERIMENT
Qualification experiments were carried out to determine what fraction of
the unregulated pollutants entering the dilution tunnel could be recovered at
the sampling point. A constant flow of each unregulated pollutant was
injected from a pressurized cylinder into the dilution tunnel-CVS system at
the point raw exhaust normally enters the dilution tunnel (Figure 57). The
CVS diluted samples were extracted from the dilution tunnel-CVS system with
a multiport sampling probe at a point after the orifice plate on the tunnel
and before the CVS system. All qualified unregulated pollutants were sampled
at this point except for nitrous oxide which was taken as a bag sample at the
CVS (Figure 57 ).
Experiments were carried out with and without diesel exhaust present in
the dilution tunnel-CVS system. A Mercedes 240D driving over a hot FTP (23-
minute test) driving cycle was used to generate diesel exhaust for the
experiments. Baseline emission levels of each pollutant from the Mercedes
240D were measured in order to correct recovery values for pollutants present
in the exhaust.
The gaseous unregulated pollutants were injected into the dilution
tunnel by the system shown in Figures 58 and 59. The pollutant passed
through a needle valve to regulate flow, a flowmeter to monitor flowrate,
and a dry gas meter to measure the injected volume of pollutant, before
entering the dilution tunnel. A thermocouple was used to monitor the
temperature of the injected gas, and a magnehelic gauge was used to monitor
the pressure of gas passing through the injection system. This pressure was
positive and generally recorded 0-2" of water. The phenols were injected
into the dilution as a water solution using the modified mist generator shown
in Figure 60. The test sequence developed to determine pollutant recovery
consistrd of a 23-minute continuous sampling period (pollutant injected with
or without exhaust present) followed by a 10-minute soak period with the CVS
off (no pollutant injected). During this time, impingers, bags or traps
were changed to collect the next sample. After the soak period the test
sequence was repeated until three to four sampling periods were completed.
During each sampling period, three replicate impinger samples (aldehydes,
total cyanide, organic amines, sulfur dioxide, ammonia, hydrogen sulfide—
two for phenols), or two trap samples (organic sulfides) or one bag sample
(nitrous oxide) were taken.
Nominal injected pollutant flows into the tunnel were 0.35 cu ft/min
while nominal CVS flows were 300 cu ft/min. This gave an approximate 850
to 1 dilution. Percent recoveries were determined by analyzing the recovered
182
-------
Filter
Box
I
Dilution Tunnel
16" long x 18" diameter
orifice
plate
oo
co
¥
CVS
System
4" diameter
Unregulated pollutant
injected hero with or
without raw exhaust
Sample Probe
for collection of
Aldehydes & Ketones
Total Cyanide
Organic Amines
Sulfur Dioxide
Hydrogen Sulfide
Ammonia
Organic Sulf.i.des and
Phenols Samnlos
Bag sample
for
Nitrous Oxide
Figure 57» Dilution tunnel-CVS system used in qualification experiments.
-------
CO
-P.
Gas Temperature
Digital Readout
Ol8]0|
Regulating
Valve
Cylinder
Source
of
Pollutant
Magnehelic
(pressure)
1|213|4|5
Gas Volume
Digital Readout
Dilution
Tunnel
-Filter Box
Figure 58. Apparatus for injection of pollutant into dilution tunnel without exhaust.
-------
Gas Temperature
Digital Readout
0|8|D|
Flowmeter
oo
01
Regulating
Valve
Dry
Gas
Meter
o
Magnehelic
(pressure)
Cylinder
Source
of
Pollutant
Raw
Vehicle ——]
Exhaust
12 345
Gas Volume
Digital Readout
Dilution
Tunnel
Filter Box
Figure 59. Apparatus for injection of pollutant into dilution tunnel with exhaust.
-------
Phenol/water
solution in
Compressed
air in
oo
Into
Dilution
Tunnel
Modified 2000 mi
Erlenmeyer Flask
Figure 60. Modified mist generator.
-------
diluted sample, multiplying by the CVS dilution and dividing by the actual
injected pollutant concentration.
187
-------
SECTION 13
RESULTS AND CONCLUSIONS
To determine the suitability of the analytical procedures initially
selected for dilute exhaust analysis, validation and qualification experi-
ments were carried out. The validation experiments determined if the sam-
pling 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 65 along with methods of
sampling and analysis. Table 65 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 and hydrogen sulfide are stable for
several days and can be stored and rerun within hours after sampling to pre-
vent loss of sample integrity. All instruments demonstrate linearity of re-
sponse 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 documented
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 proce-
dure. Phthalates were found to interfere with the aldehyde and ketone pro-
cedure and may cause erroneous results for crotonaldehyde. In the hydrogen
sulfide procedure, sulfur dioxide decreases 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, ammonia and phenol procedures to determine the recovery of
known amounts of each pollutant from the CVS tunnel with and without exhaust
(phenols CVS tunnel with exhaust only). Aldehydes and ketones, sulfur di-
oxide, nitrous oxide, total cyanide and phenols can be recovered quantita-
tively from the CVS tunnel with and without (not done for phenols) exhaust.
There is a 10 percent loss of hydrogen sulfide with and without exhaust
present. The organic amines, ammonia, and the organic sulfides experience
significant losses in the CVS tunnel with and without exhaust. These losses
must be taken into account when determining the concentration of these com-
188
-------
TABLE 65. ANALYTICAL PROCEDURES FOR EMISSIONS CHARACTERIZATION
00
vo
Compounds
Aldehdyes and Ketones
Organic Amines
Sulfur Dioxide
Nitrous Oxide
Individual Hydrocarbons
Hydrogen Sulfide
Hydrogen Cyanide + Cyanogen
Carbonyl Sulfide + Organic Sulfides
Ammonia
Sulfate
DMNA
Phenols
BaP
Sampling
Impinge rs
Impinge rs
Impingers
Bags
Bags
Impingers
Impingers
Traps
Impingers
Filters
Traps
Impingers
Filters
Analysis
DNPH
GC-NPD
Ion Chrom.
GC-ECD
GC-FID
Meth. Blue
GC-ECD
GC-FPD
Ion Chrom.
BCA
GC-MS @ RTI
GC-FID
Fluorescence
Validation
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Not required
Not required
Yes
Not required
Qualification
Yes
Yes
Yes
Yes
Not required
Yes
Yes
Yes
Yes
Not required
Not required
Yes
Not required
@ EPA
-------
pounds in exhaust.
The procedures discussed in this report are effective in collecting and
analyzing dilute exhaust samples and are recommended for use in this project.
190
-------
REFERENCES
1. Dietzmann, H. E., Smith, L. R., Parness, M. A., and Fanick, E. R.,
"Analytical Procedures for Characterizing Unregulated Pollutant
Emissions from Motor Vehicles." Report to Environmental Protection
Agency under Contract No. 68-02-2497, February, 1979.
2. Muir, G. D. (Ed.), Hazards in the Chemical Laboratory, The Chemical
Society, London, 1977.
3. Stump, F., Oxygenated Compounds in Automobile Exhaust-GAS CHROMATO-
GRAPHIC PROCEDURE, MSERB-ESRL-EPA, Research Triangle Park, North
Carolina.
4. Roberts, R. M., Giblert, J. C., et al; An Introduction to Modern
Experimental Organic Chemistry; Holt, Rinehart and Winston, Inc.
1974, pg. 470-471.
5. Weast, R. C. (Ed.), Handbook of Chemistry and Physics, 54 Edition,
Division of the Chemical Rubber Co., 1973.
6. Sekerka, I., and Lechner, J. F., Water Res., Vol. 10, pg. 479, 1976.
7. Valentour, J. C., Aggarwal, V., and Sunshine, I., Anal. Chem., Vol. 46,
pg. 924, 1974.
8. Dimitriades, B., and Seizenger, D. E., Environ. Sci. Tech., Vol. 5,
No. 3, pg. 223, 1971.
9. Papa, L. J., Dinsel, D. L., et al; J. Gas Chromatog., Vol. 6, pg. 270,
1968.
10. Klosterman, D. L., and Sigsby, J. E., Environ. Sci. Tech., Vol. 1,
pg. 309, 1967.
11. (a) Black, F. M., High, L. E., and Sigsby, J. E., J. Chromatog. Sci.,
Vol. 14, pg. 257, May 1976. (b) Communication between Mobile Source
Emissions Research Branch, ESRL-EPA, Research Triangle Park, and
Southwest Research Institute.
12. Braker, W., and Mossman, A. L., Matheson Gas Data Book, 5th Edition,
East Rutherford, N.Y., 1971.
13. Hoshika, Y., Anal. Chem., Vol. 48, pg. 1716, 1976.
191
-------
14. Hoshika, Y., J. Chromatog., Vol. 115, pg. 596, 1975.
15. Sze, Y. L., and Borke, M. L., Anal. Chem., Vol. 35, pg. 240, 1963.
16. O'Donnell, J. F., and Mann, C. K., Anal. Chem,, Vol. 36, pg. .2097, 1964.
17. McCurdy, W. H., Jr., and Reiser, R. W., Anal. Chem., Vol. 38, pg. 795,
1966.
18. Smith, J. R. L., and Waddington, D. J., Anal. Chem., Vol. 40, pg. 522,
1977.
19. "Amine Analysis," Bulletin 737A, Supelco Co., Belefonte, Pa., 1973.
20. "Penwalt 223 Amine Packing," Alltech Associates, Arlington Heights, 111.
21. Hoffat, A. C., and Horning, E. C., Anal. Lett., Vol. 3, pg. 205, 1970.
22. Clarke, D. D., Wilk, S., et al; J. Gas Chromatog., pg. 310, 1966.
23. Hosier, A. R., Andre, C. E., et al; Environ. Sci. Tech., Vol. 7, pg.
642, 1973.
24. Burks, R. E., Jr., Baker, E. B., et al; J. Agr. Food Chem., Vol. 7,
pg. 778, 1959.
25. Dunn, S. R., Simenhoff, M. L., et al; Anal. Chem., Vol. 48, pg. 41,
1976.
26. Gruger, E. H., Jr., J. Agr. Food Chem., Vol. 20, pg. 781, 1972.
27. Andre, C. E., and Hosier, A. R., Anal. Chem., Vol. 45, pg. 1971, 1973.
28. Umbreit, G. R., Nygren, R. E., and Testa, A. J., J. Chromatog., Vol. 43,
pg. 25, 1969.
29. Keay, N. J., and Hardy, R., J. Scie. Food Agr., Vol. 23, pg. 9, 1972.
30. Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.
31. West, P. W., and Gaeke, G. C., Anal. Chem., Vol. 28, pg. 1816, 1956.
32. Adams, D. F., Corn, M., et al; "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. Publ.
Health Assn., Washington, D. C., 1972.
33. "ASTM Standards - Water, Atmospheric Analysis," ASTM Designation
D291470T, Part 23, Am. Soc. Test. Mat., Philadelphia, Pa., 1972.
34. US EPA, Fed. Register, Vol. 36, No. 158, 15492, 1971.
192
-------
35. 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.
36. Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press Inc
N.Y., 1976, Vol. 3, pg. 214.
37. Zurlo, N., and Griffini, A. M., Med. Lav.., Vol. 53, pg. 330, 1962.
38. Scaringelli, F. P., Elfers, L. , et al; Anal. Chem., Vol. 42, pg. 1818,
1970.
39. Pate, J. B., Ammons, B. E., et al; Anal. Chem., Vol. 37, pg. 942, 1965.
40. Pate, J. B., Lodge, J. P., Jr., et al; Anal. Chem., Vol. 34, pg. 1660,
1962.
41. Scaringelli, F. P., Saltzman, B. E., et al; Anal. Chem., Vol. 39,
pg. 1709, 1967.
42. Hartman-Leddon, Philadelphia, Pa.
43. Stern, A. C. (Ed.), Air Pollution, 3rd Edition, Academic Press, Inc.,
N. Y., 1976, Vol. 3, pg. 214.
44. Urone, P., Evans, J. B., and Noyes, C. H., Anal. Chem., Vol. 37,
pg. 1104, 1965.
45. Bostrom, C. E., Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.
46. Attari, A., Igielski, T. P., et al; Anal. Chem., Vol. 42, pg. 1282,
1970.
47. Kawai, T., Netsu Kanri, Vol. 22, pg. 20, 1970.
48. Thomas, M. D., Ivie, J. 0., et al; Ind. Eng. Chem., Anal. Ed., Vol. 18,
pg. 383, 1946.
49. Yocum, J. E., Richardson, R. L., et al; Proc. 49th Ann. Meet. Air
Poll. Contr. Assn., Pittsburgh, Pa., 1956.
50. Kuczynski, E. R. , Environ. Sci. Tech., Vol. 1, pg. 68, 1967.
51. Shikiya, J. M., and McPhee, R. D., 61st Annual Meeting, Paper No. 68-72,
Air Poll. Contr. Assn., Pittsburgh, Pa., 1968.
52. Jacobs, M. B., The Chemical Analysis of Air Pollutants, Chapter 8,
Wiley (Interscience), New York, N.Y., 1960.
53. Jacobs, M. B., The Analytical Chemistry of Industrial Poisons, Hazards,
and Solvents, Chapter 9, Wiley (Interscience), New York, N.Y., 1949.
193
-------
54. Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.
55. Stratman, H., Mikrochim. Acta., Vol. 6, pg. 688, 1954.
56. Pate, J. B., Lodge, J. P., Jr., et al; Anal. Chim. Acta., Vol. 28,
pg. 341, 1963.
57. Lodge, J. P., Jr., Pate, J. B., et al; Amer. Ind. Hyg. Assn. J., Vol.
24, pg. 380, 1963.
58. Forrest, J., and Newman, L., Atmos. Environ., Vol. 7, pg. 561, 1973.
59. Huygen, C., Anal. Chim. Acta., Vol. 28, pg. 349, 1963.
60. Harding, J., and Schlein, B., "Nuclear Techniques in Environmental
Pollution," SM-142a/0, IAEA, Vienna, Austria, 1971.
61. Wilson, B. H., and McConnell, F. J., J. Soc. Chem. Ind., Vol. 53,
pg. 385, 1934.
62. Huey, N. A., J. Air Poll. Cont. Assn., Vol. 18, pg. 610, 1968.
63. Thomas, F. W., and Davidson, C. M., J. Air Poll. Cont. Assn., Vol. 11,
pg. 1, 1961.
64. Stalker, W. W., Dickerson, R. C., et al; Amer. Ind. Hyg. Assn. J.,
Vol. 24, pg. 68, 1963.
65. Hickey, H. R., and Hendrickson, E. R., J. Air Poll. Cont. Assn.,
Vol. 15, pg. 409, 1965.
66. Harding, C. I., and Kelley, T. R., J. Air Poll. Cont. Assn., Vol. 17,
pg. 545, 1967-
67. Ikeda, T., J. Hyg. Chem., Vol. 13, pg. 41, 1967.
68. Stevens, R. K., Mulik, J. D., et al; Anal. Chem., Vol. 43, pg. 827,
1971.
69. Stevens, R. K., O'Keeffe, A. E., et al; Environ. Sci. Tech., Vol. 3,
pg. 652, 1969.
70. Ronkainen, P., Denslow, J., et al; J. Chromatog. Sci., Vol. 11, pg. 384,
1973.
71. Pecsar, R. E., and Hartmann, C. H., J. Chromatog. Sci., Vol. 11, pg.
492, 1973.
72. Stevens, R. K., and O'Keeffe, A. E., Anal. Chem., Vol. 42, pg. 143a,
1970.
73. Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070, 1972.
194
-------
74. Bruner, F., Ciccioli, P., et al; Anal. Chem., Vol. 47, pg. 141, 1975.
75. DeSouza, T. L. C.f Lane, D. C., et al; Anal. Chem., Vol. 47, pq 543
1975.
76. Bruner, F. , Ciccioli, P., et al; J. Chromatog., Vol. 99, pg. 661, 1974.
77. Bremner, J. M., and Banwart, W. L., Sulfur Inst. Journal, Vol. 10,
pg. 6, 1974.
78. 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.
79. 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.
80. Stern, A. C. (Ed.), Air Pollution, Third Edition, Volume III, Academic
Press, New York, 1976, pg. 275.
81. Private communication between R. B. Zweidinger and Frank Black of EPA,
Research Triangle Park, and Harry Dietzmann of Southwest Research
Institute.
82." Bethea, R. M., J. Air Poll. Cont. Assoc., Vol. 23, pg. 710, 1973.
83. Bamesberger, W. L., and Adams, D. F., Tappi, Vol. 52, pg 1302, 1969.
84. "Hydrogen Sulfide in Air Analytical Method," H.L.S., Vol. 12, pg. 362,
1975.
85. Bamesberger, W. L., and Adams, D. F., Environ. Sci. Tech., Vol. 3,
pg. 258, 1969.
86. Gustafsson, L., Talanta, Vol. 4, pg. 227, 1960.
87. Moest, R. R., Anal. Chem., Vol. 47, pg. 1204, 1975.
88. Buck, M., and Gies, H., Staub, Vol. 26, pg. 27, 1966.
89. Jacobs, M. B., Braverman, M. M., and Hochheiser, S., Anal. Chem.,
Vol. 29, pg. 1349, 1957.
90. Cave, G. C. B., Tappi, Vol. 46, pg. 1, 1963.
91. Budd, M. S., and Bewick, H. A., Anal. Chem., Vol. 24, pg. 1536, 1952.
92. Bostrom, C., Air and Water Poll. Int. J., Vol. 10, pg. 435, 1966.
93. Flamm, D. L., and James, R. E., Environ. Sci. Tech., pg. 159, 1975.
195
-------
94. Adams, D. F., Bamesberger, W. L., et al; J. Air Poll. Cont. Assn.,
Vol. 18, pg. 145, 1968.
95. Stecher, P. G. (Ed.), The Merck Index, 8th Edition, Merck and Company,
Inc., Rahway, N. J., 1968, pg. 64.
96. Masterton, W. L., and Slowinski, E. J., Chemical Principles, 3rd
Edition, W. B. Saunders Company, Philadelphia, Pa., 1973.
97. 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.
98. 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.
99. Kothny, E. L., Intersociety Committee, Health Lab. Sci., Vol. 10,
pg. 115, 1973.
100. Okita, T., and Kanomori, S., Atmos. Environ., Vol. 5, pg. 622, 1971.
101. Kramer, D. N., and Seen, J. M., Anal. Chem. , Vol. 44, pg. 395, 1972.
102. Zitomer, F., and Laumbert, J. L., Anal. Chem., Vol. 34, pg. 1738, 1962.
103. Blinn, R. C., and Gunther, F. A., Anal. Chem., Vol. 29, pg. 1882, 1957.
104. Williams, D. D., and Miller, R. R., Anal. Chem., Vol. 34, pg. 225,
1962.
105. Gunther, F. A., Barkley, J. H., et al; Anal. Chem., Vol. 28, pg 1985,
1956.
106. Kolbezen, M. J., Eckert, J. W., et al; Anal. Chem., Vol. 36, pg. 593,
1964.
107. Pierson, R. H., Fletcher, A. A., et al; Anal. Chem., Vol. 28, pg. 1218,
1956.
108. Burns, E. A., The Analysis of Exhaust Gases Trapped from Ablating
Nozzles, TRW Report No. 9840-6001-TU 000, Redondo Beach, Calif., 1964.
109. Howell, J. H., and Boltz, D. F., Anal. Chem., Vol. 36, pg. 1799, 1964.
110. Burns, E. A., Unpublished results, 1963.
111. Jenkins, R. W., Jr., Cheek, C. H., et al; Anal. Chem., Vol. 38,
pg. 38, pg. 1257, 1966.
196
-------
112. Miettinen, J. K., and Virtanen, A. I., Ann. Acad. Aci. Fennicae,
Ser. A, II, No. 41, 1951, through Chem. Abst. , Vol. 46, pg. 11500,
1952.
113. Arcand, G. M., and Swift, E. H., Anal. Chem., Vol. 28, pg. 440, 1956.
114. Krivis, A. F., Supp, G. R., et al; Anal. Chem., Vol. 35, pg. 2216,
1963.
115. Gilbert, T. R., and Clay, A. M. , Anal. Chem., Vol. 45, pg. 1758, 1973.
116. Mertens, J., Van den Winkle, P., et al; Bull. Soc. Chem. Belg. ,
Vol. 83/n° 1-2/, pg. 26, 1973.
117. Furman, N. H. (Ed.), Scott's Standard Methods of Chemical Analysis,
5th Edition, Van Nostrand, N. Y., 1958, Vol. 1, pg. 2336.
.118. Vogel, A. I., Quantitative Inorganic Analysis, 3rd Edition, Longmans,
Green, London, 1961, pg. 1092.
119. Steyermark, A., Quantitative Organic Microanalysis, Blakestone, N. Y.,
1951, pg. 56.
120. Sadek, F. S., and Reilly, C. N., Anal. Chem., Vol. 31, pg. 494, 1959.
121. Crane, F. E., and Smith, E. A., Chemist-Analyst, Vol. 49, pg. 38, 1960.
122. Crane, F. E., and Smith, E. A., Chemist-Analyst, Vol. 52, pg. 105,
1963.
v
123. Hodges, J. A., Bell, J. P., et al; Joint Conf. on Sensing Environmental
Pollutants Collected, Technical Papers, 1972, Vol. 1, 71-1067.
124. Sigsby, J. E., Jr., Black, F. M., et al; Environ. Sci. Tech., Vol. 7,
pg. 51, 1973.
125. Spencer, C. F., Baumann, F., et al; Anal. Chem., Vol. 30, No. 9,
pg. 1473, 1958.
126. Sumner, S., Karrman, K. J., et al; Mikrochim Acta, pg. 1144, 1956.
127. Liberti, A., and Cartoni, G. P., Chim. e ind., Vol. 39, pg. 821, 1957.
128. Ryce, S. A., and Bryce, W. A., Anal. Chem., Vol. 29, pg. 925, 1957.
129. Desty, D. H., and Whyman, B. H. F., Anal. Chem., Vol. 29, pg. 230,
1957.
130. Coleman, H. J., Thompson, C. C., et al; Anal. Chem., Vol. 30, pg. 1592,
1958.
197
-------
131. Desty, D. H., and Harbourn, C. L. A., Div. of Analytical and Petroleum
Chemistry, Symposium on Advances in Gas Chromatography, 132nd meeting,
ACS, New York, N. Y., September, 1957.
132. Amberg, C. H., Can. J. Chem., Vol. 36, pg. 590, 1958.
133. Sullivan, J. H., Walsh, J. T., et al; Anal. Chem., Vol. 31, pg. 1827,
1959.
134. Dal Nogare, S., and Bennett, C. E., Anal. Chem., Vol. 30, pg. 1157,
1958.
135. Harrison, G.'F., Knight, P., et al; Second Int'l Symposium on Gas
Chromatograph, Amsterdam, May, 1958.
136. Adams, D. F., and Koppe, R. K., Tappi, Vol. 42, pg. 601, 1959.
137. Schols, J. A., Anal. Chem., Vol. 33, pg. 359, 1961.
138. Hall, H. L., Anal. Chem., Vol. 34, pg. 61, 1962.
139. Adams, D. F., Kippe, R. K., et al; Tappi, Vol. 43, pg. 602, 1960.
140. Feldstein, M., Balestrieri, S., et al; J. Air Poll. Cont. Assn.,
Vol. 15, pg. 215, 1965.
141. Fredericks, E. M., and Harlow, G. A., Anal. Chem., Vol. 36, pg. 263,
1964.
142. Kummer, W. A., Pitts, J. N., Jr., et al; Environ. Sci, Tech., Vol. 5,
pg. 1045, 1971.
143. Mizany, A. I., J. Chromatog. Sci., Vol. 8, pg. 151, 1970.
144. Rasmussen, R. A., AM. Lab, Vol. 4, pg. 55, 1972.
145. Brody, S. S., and Chaney, J. E., J. Gas Chromatog., Vol. 4, pg. 42,
1966.
146. Bruner, F., Liberti, A., et al; Anal. Chem., Vol. 44, pg. 2070, 1972.
147. Elliott, L. F., and Travis, T. A., Soil Sci. Soc. Amer. Proc., Vol. 37,
pg. 700, 1973.
148. Francis, A. J., Adamson, J., et al; Bull. Ecol. Res. Comm., (Stockholm),
Vol. 17, pg. 485, 1973.
149. Lewis, J. A., and Papavizas, G. C., Soil Biol. Biochem., Vol. 2,
pg. 239, 1970.
198
-------
150. Lovelock, J. E., Maggs, R. J., et al; Nature, London, Vol 237
pg. 452, 1972.
151. Ronkainen, P., Denslow, J., et al; J. Chromatog. Sci., Vol 11
pg. 384, 1973.
152. Stanley, T. W., Sawicki, E., Johnson, H., and Pfaff, J. D.,
Mikrochim, Acta, Vol. 1, pg. 48, 1965.
153. Umbreit, G., and Houtman, R. L., Journal of Pharmaceutical Sciences,
Vol. 56, pg. 349, 1967.
154. Afghan, B. K., Belliveau, P. E., Larose, R. H., and Ryan, J. P.,
Analytica Chimica Acta, Vol. 71, pg. 355, 1974.
155. Lawrence, J. F., Journal of Chromatographic Science, Vol. 17, pg. 147,
1979.
156. Wolkoff, A. W., and Larose, R. H., Journal of Chromatography, Vol. 99,
pg. 731, 1974.
157. Sheffer, H. E., Perry, R. L., Thimineur, R. J., Adams, B. T., Simonian,
J. L., Zutty, N. L., and Clendenning, R. A., Ind. Eng. Chem. Prod.
Res. Develop., Vol. 10, pg. 362, 1971.
158. Shulgin, A., Analytical Chemistry, Vol. 36, pg. 920, 1964.
159. Argauer, R. J., Analytical Chemistry, Vol. 40, pg. 122, 1968.
160. Heenan, M. P., and McCallum, N. K., Journal of Chromatographic Science,
Vol. 12, pg. 89, 1974.
161. Kawahara, P. K., Analytical Chemistry, Vol. 40, pg. 1009, 1968.
162. Cohen, I. C., Norcup, J. , Ruzicka, H. A., and Wheals, B. B., J.
Chromatog., Vol. 44, pg. 251, 1969.
163. Seiber, J. N., Crosby, D. G., Fouda, H., and Soderquist, C. J.,
J. Chromatog., Vol. 73, pg. 89, 1972.
164. Bartle, K. D., Els tub, J., Novotny, M., and Robinson, R. J., Journal
of Chromatography, Vol. 135, pg. 351, 1977.
165. Ma, T. S., and Speigel, D. , Microchemical Journal, Vol. 10, pg. 61,
1966.
166. Mindrup, R., Jr., Industrial Research/Development, pg. 79, August,
1978.
167. Chriswell, C. D., Chang, R. C., and Fritz, J. S., Analytical Chemistry,
Vol. 47, pg. 1325, 1975.
199
-------
168. Braithwaite, B., and Penketh, G., Analytical Chemistry, Vol. 36, pg.
185, 1964.
169. Landault, C., and Guiochen, G., Analytical Chemistry, Vol. 39, pg. 713,
1967.
170. Fisher, G. E., and Neerman, J. C., I and EC Product Research and
Development, Vol. 5, pg. 288, 1966.
171. Bark, L. S., and Clarke, K. F., J. Chromatog., Vol. 48, pg. 418, 1970.
172. Di Corcia, A., Journal of Chromatography, Vol. 80, pg. 69, 1973.
173. Preston, S. T., A Guide to the Analysis of Phenols by Gas
Chromatography, PolyScience Corporation, Niles, Illinois, 1966.
174. Eichelberger, J. W., Dressman, R. C., and Longbottom, J. E.,
Environmental Science and Technology, Vol. 4, pg. 576, 1970.
175. Yrjanheikki, Erkki, Am. Ind. Hyg. Assoc. J., Vol. 39, pg. 326, 1978.
176. Brown, R. A., Searl, T. D., King, W. H., Jr., Dietz, W. A., and
Kelliher, J. M., Rapid Methods of Analysis for Trace Quantities of
Polynuclear Aromatic Hydrocarbons and Phenols in Automobile Exhaust,
Gasoline and Crankcase Oil, Final Report for CRC - APRAC Project
CAPE-12-68, Esso Research and Engineering Company, Linden, N.J., 1973.
177. Barber, E. D., Sawicki, E., and McPherson, S. P., Analytical Chemistry,
Vol. 36, pg. 2442, 1964.
178. Grouse, R. H., Garner, J. W., and O'Neill, H. J., J. of G.C., pg. 18,
February, 1963.
179. Spears, A. W., Analytical Chemistry, Vol. 35, pg. 320, 1963.
200
-------
APPENDIX A
ALDEHDYE AND KETONE PROCEDURE
201
-------
THE MEASUREMENT OF ALDEHYDES AND KETONES IN EXHAUST
The aldehydes and ketones that are included in this analysis are:
formaldehyde, acetaldehyde, acetone (acetone, acrolein, and propionaldehyde
are not resolved from each other under normal operating conditions and all
three are reproted together as acetone), isobutyraldehyde, methylethylketone,
crotonaldehyde, hexanaldehyde, and benzaldehyde. The measurement of the
aldehydes and ketones 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 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 insoluble or only slightly soluble in the
DNPH/HC1 solution and are removed by filtration followed by pentane extrac-
tions. The filtered percipitate and the pentane extracts are combined and
the pentane is removed by evaporation in the 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 dissolved extract is
injected into a gas chromatograph and analyzed using a flame ionization de-
tector. The detection limits for this procedure under normal operating con-
ditions are on the order of 0.005 ppm carbonyl compound in dilute exhaust.
SAMPLING SYSTEM
Two glass impingers in series, each containing 40 m£ of 2NHCl-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 impinger is
maintained at 0-5°C by an ice water bath, and the flow rate through the im-
pinger is maintained at 4 &/minute by the sample pump. A dry gas meter is
used to determine the total flow through the impinger during a given sampling
period. The temperature of the gas stream is monitored by a thermocouple
immediately prior to the dry gas meter. A drier is included in the system
to prevent condensation in the pump, flowmeter, dry gas meter, etc. The
flowmeter in the system allows monitoring of the sample flow to insure pro-
per flow rates during sampling. When sampling diesel fueled vehicles, a
heated filter, located between the on-off solenoid valve and the dilution
tunnel, is used to prevent diesel particulate from contaminating the sampling
system. The filter and line connecting the filter to the dilution tunnel
are heated to 375 °F in order to prevent the aldehydes and ketones from being
retained in the filter and sample line. The Teflon line connecting the
202
-------
ro
o
GO
Sample
Probe
Dilute
Exhaust
Gas Temperature
Digital Readout
Plowmeter
Regulating
Valve
Ice Bath
Temperature Readout
Dry
Gas
Meter
Gas Volume
Digital Readout
Figure 1. Aldehyde and ketone sample collection flow schematic.
-------
heated filter and the solenoid valve is heated to ~175°F in order to prevent
water from condensing in the sample line. Several views of the sampling
system are shown in Figure 2.
ANALYTICAL PROCEDURE
The analysis of the aldehydes (formaldehyde, acetaldehyde, isobutyr-
aldehyde, crotonaldehyde, hexanaldehyde, and benzaldehyde) and of the ketones
(acetone and methylethylketone) in dilute exhaust is accomplished by col-
lecting these carbonyl compounds in a hydrochloric acid (HCl)/2,4 dinitro-
phenylhydrazine (DNPH) solution as their 2,4 dinitrophenylhydrazone deri-
vatives. The derivatives are removed from the HCl/DNPH absorbing solution
by filtration and/or extractions with pentane. The filtered precipitate and
the pentane extracts are combined and the volatile solvents are removed.
The remaining extract contains the phenylhydrazone derivatives. The deri-
vatives are then dissolved in a quantitative volume of toluene containing a
known amount of anthracene as an internal standard. This solution is ana-
lyzed by injecting a small volume of the solution in to a gas chromatograph
equipped with dual flame ionization detectors. From this analysis and the
measured volume of exhaust sampled, the concentration of the carbonyl com-
pounds in exhaust can be determined. The analysis flow schematic for the
aldehydes and ketones is shown in Figure 3. A detailed description of the
procedure follows.
The aldehdyes 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 m& of a 2 N HC1 solution saturated with
DNPH. The sampling temperature and barometric pressure are recorded Curing
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 sam-
pling 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 m£ of absorbing
reagent and the water washings. The flask is then set aside for the ex-
204
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. Aldehyde and ketone sampling system
205
-------
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). Aldehyde and kstone sampling system,
206
-------
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
I
sample analyzed
in gas chromarograph
with FID
A/D converter
Recorder
Hewlett-Packard 3354
Computer System
Figure 3. Aldehyde and ketone analysis flow schematic.
207
-------
Figure 4. Impingers containing HC1/DNPH abosrbing solution.
208
-------
I >
c
c
Figure 5. Filtration of absorbing solution.
-------
traction step. The fritted glass filter containing the precipitate is con-
nected 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 the pre-
cipitate 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 furinel. 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 m£ 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 m£
separatory funnel. A second 40 mJl 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
m£ 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 derivative remain.
210
-------
Figure 6. Automatic shaker.
Figure 7. Vacuum oven,
211
-------
Each time a series of samples are collected, a blank containing 80 mJi
of HCl/DNPH solution is extracted and dried in the same manner as the samples.
This accounts for any aldehydes or interferring compounds which might be
found in the reagents used for extraction.
Two m£ of toluene which contain a quantitative amount of anthracene
C\£.05 mg/m£ toluene) as an internal standard is pipetted into the flask
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 (Figure 8).
At this point the derivative is ready for injection into the gas chromato-
graph system.
The gas chromatograph system used to analyze the toluene solution con-
taining the pherylhydrazone derivatives is shown in Figure 9- The system
consists of a Varian 1700 GC, and 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 1/8 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 flows through the columns at a rate of 40 mi/minute.
The optimum hydrogen and air flow rates are 35 m&/minute and 500 mil/minute,
respectively. The column temperature, after injection of the sample, is
programmed 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, isobutryaldehyde, methyl-
ethylketone, crotonaldehyde, hexanaldehyde, and benzaldehyde, the first peak
eluted is toluene followed by anthracene, and then the derivatives of for-
maldehyde, acetaldehyde, acetone, isobutyraldhyde, methylethylketone, cro-
tonaldehyde , hexanaldehyde, and benzaldehyde. 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 stan-
dard deviation of 0.87 percent for formaldehyde. The computer printout 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£. The concentration is
calculated by the computer from response factors which are determined daily-
Each day a standard containing known amounts of the derivatives and anthra-
cene is injected into the GC. From the anthracene and derivative areas the
computer calculates a response factor F. The F factors are used in all sub-
sequent runs during the day to determine the concentration of the derivatives.
This response is calculated from the following equation:
Response Factor (F) = Anthracene Area x mg/m£ Derivative
Derivative Area mg/m£ Anthracene
212
-------
Figure 8. 1/2 dram vials.
Figure 9. Aldehyde and ketone analytical system.
213
-------
SOTOUAldehyde & Ketone Stdjfa
Imtrunmit ynT-< »
Column 2 ft. 1/8 P.P.
1. Injection
2. Toluene
3. Anthracene
4. Formaldehyde
5. Acetaldehyde
6. Acetone
7. Isobutyraldehyde
8. Methylethylketone
9. Crotonaldehyde
10. Hexanaldehyde
LI. Benzaldehyde
m. ry>xsii ago GC
Qiromosorb G
°C mimi 40 _cc/mln. Helium
Rounwttr RMdii
• Id aUa
120 "CISOIof -
Oj_i_
/min. Htld
min. lothwl
"C.
"C
min.. Prog to "CM
Rounwtw Rdg,
Rouiratw Rdg.
Rounwur ftdg.
1 mV.F.S. so
14 12 10 8
Retention time, minutes
Fiugre 10. Chromatogram of standard.
214
-------
REPORT: 14.11 CHANNEL: 11
SAMPLE: STANDARD INJECTED AT 11:18:27 ON MAR
ISTD METHOD: DNPH11
1978
ACTUAL RUN TIME: 30*008 MINUTES
ISTD-RATIO: .050..RMG/ML STD-AMT:
0500
SAMP-AMT:
1 .0000
RT
AREA
MG/ML
ro
7.26
9.81
11.71
12.64
13.28
13.70
14.69
16.08
19.00
9638
11 159
13355
17898
16448
16469
11 167
15988
10525
BB
BB
BV
VV
VV
VV
VV
VV
BB
.203
• 202
.203
.202
.201
. 199
.202
. 198
TOTAL AREA =
122648
NAME
&ANTHRACENE
#FORMALDEHYDE
#ACETALDEHYDE
#ACETONE
#I SO-BUTYRALDEHYDE
#MEK
#CROTONALDEHYDE
#HEXANALDEHYDE
#BENZALDEHYDE
TOTAL MG/ML =
1.610
Figure 11. Computer printout of standard .
-------
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 isobutyraldehyde
2.3332 MethylethyUcetone
3.4174 Crotonaldehyde
2.342 8 Hexanaldehyde
2.9329 Benzaldehyde
When the response factor is known a concentration in mg/mH 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 Vg/m3 of
exhaust and ppm for each carbonyl compound. 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 stan-
dard 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
V = experimental volume of gas sampled in ft
VGXP = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
P = experimental barometric pressure
P P = 29.92"Hg
corr
T = experimental temperature in °F + 460
TexP = 68°F + 460 = 528°R
corr
Solving for V gives:
corr
P ("Hg) x V (ft3) x 528°R
v = e*P exp
corr T (°R) 29.92" Hg
exp
216
-------
1. Injection
2. Toluene
3. Anthracene
4. Formaldehyde
5. Acetaldehyde
6. Isobutyraldehyde
7- Methylethylketone
8. Crotonaldehyde
9. Hexanaldehyde
10. Benzaldehyde
24
22
12 10
Figure 12. Sample chroamtogram.
217
-------
REPORTS 20 CHANNEL: 11
SAMPLE: RCI IMJECTED AT 15:41:05 ON MAP. 1* 1976
ISTD METHOD: DNPHll
ACTUAL RUN TIME: 30.017 MINUTES
ISTD-RATIO: .050*R MG/ML STD-AMT:
0500
SAMP-AMT:
1.0000
AREA
MG/ML
ro
!-»
oo
7. 15
7.95
10.03
11. 88
12.83
13.76
14.85
15.95
16.77
19.20
20.23
23.52
25.03
25.48
8604
435
8877
575
463
1594
2630
146
675
648
1912
217
13
84
BV
VE
BB
BV
W
W
vv
vv
vv
VB
BV
VV
VB
BB
• 003
. 186
.010
.007
.022
.053
.002
• 004
. 012
• Oil
.001
7-6E- 5
4.9E- 4
TOTAL AREA = 26874
PROCESSED DATA FILE: *PRC11
NAME
&AMTHP.ACENE
^FORMALDEHYDE
#ACET ALDEHYDE
# I SO -BUTYRALDEHYDE
#CR0TONALDEHYDE
#HEXAN ALDEHYDE
#BENZALDEHYDE
TOTAL MG/ML = .310
RA¥ DATA FILE: *RAW11
Figure 13. Computer printout of sample
-------
The next step converts the volume from cubic feet to cubic meters by
of the conversion factor;! cubic meter is equal to 35.31 cubic feet.
use
V , Pexo ("H9)X vAvn (ft)x 528°R
ex exp _
29.92" Hg x 35.31 ft3/m3
(Equation 1)
The next step converts the mg/m£ of derivative determined by the com-
puter to mg of carbonyl collected in the two impingers. To obtain mg of
derivative, the concentration (from the computer printout) in mg/mJl is
multiplied by the volume of toluene used to dissolve the solid extract.
mg derivative = Conc_ (mg/m£) x Vol (m£)
L/63T
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&) mol. wt. carbonyl
Der *' Tol x — ; - - — - — : — f-r—
mol. wt. derivative
To obtain the number of TO of carbonyl compound the mg of carbonyl are
multiplied by the conversion factor, 1000 yg/mg.
Vig carbonyl =* Cone (mg/m£) Vol 1 (m£) mol. wt. carbonyl
Der mol. wt. derivative
x 1000 yg/mg
(Equation 2)
The concentration of the carbonyl compound in exhaust can now be found
in iug/m3 by dividing equation 2 by equation 1.
Cone (mg/m£) x Vol . (m&lxmol. wt. carbonyl
3 Der Tol _
pg carbonyl/m = - - - (1 } x y - (.ft3) x 528°
exp exp
1000 ug/mg xT (°R) x 29.92" Hg x 35.31 ft3/m3
x _ _ exp
mol. wt. derivative
(Equation 3)
219
-------
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
V = 22.4
T^ = 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
Since one mole of gas occupies 22.04& at 68°F, the density can be found in
g/H by dividing the molecular weight in g/mole by 24.04 8,/moIe.
,ori < /(M _ mol. wt. (g/mole)
den (g/Jl) -- 24.04Vmole
The density in yg/m£ can be found by converting g to yg and yg and £ to
m& as follows :
o mol. wt. g/mole 1 x 10 yg/g mol. wt. x 1000
den yg/rn* -- 24.04Vmole X 1 x 1Q3 ml/A X 24~04
(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& = — r
m
Using Equations 3 and 4 gives the ppm concentration in the form of the raw
data.
Cone (mg/m£) x Vol (m£) x mol. wt. carbonyl x 1000 yg/mJl
ppm = - B§! - — — -
P ("Hg) x V (ft ) x 528° x mol. wt. derivative
exp exp
Texp (°R) X 29'92" H9 x 35'31 ft3/m3 x 24.04
mol. wt carbonyl x 1000
220
-------
Conper (mg/m£) x VolToi (m£) * Texp (°R) x 29.92" Hg
Pexp ("Hg) X Vexp (ft3) X 528°
35.31 ft3/m3 x 24.04 A/mole
mol . wt . derivative
(Equation 5)
At this point, the concentration can be expressed in yg/m (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 67 program was developed to calculate the aldehyde and ketone concen-
tractions in yg/m3 and ppm from the raw data and phenylhydrazone derivative
concentrations (from computer printout) . This program is presented in
Figure 14.
Sample Calculations
Assume 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 15. Calculations were performed using the HP-67 programs and manual
calculations .
Manual calculation for driving cycle FTP-1 :
Cone (mg/mJl) x Vol (m£) x mol. wt. carbonyl
yg/m formaldehyde °
1000 yg/mg x T (°R) x 29.92" Hg
x 528°R
35.31 ft3/m3
X mol. wt. derivative
0.186 mg/ml, x 2m£ x 30.03 g/mole x 1000 yg/mg
29.80" Hg x 3.196 ft3 x 528°G
535°R x 29.92" Hg x 35.31 ft3/m3
X 201.15 g/mole
= 597.5 yg/m3
n
ppm formaldehyde = yg/m v density yg/mJl
mol. wt. (formaldehyde) x 1000
density yg/m£ = 24.04&~
221
-------
I S
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEYEN1RY KEY COO£ ,,,U^M^
l!OJ
010
0»
030
040
OaO
"
I LUL..& _
U
0
n
0
5
X
X
S'I'O 1
R/s
. 4 .
t
0.
RCL 1
i
ll 1/X
R/S
X
STY) •}
R/S
RCL 2
X
.
,
d
_3
X
R/S
1
2
4
q
B/e;
RCL 2
X
0
,
1
ft
Y
D /J
1
^
2
.
R/S
ur-i. •}
V
XUIii 11.
. L1JJ._
00
no
DO
2i
71
84
. 21.
84"
04
ua
bl
34 01
35 52
81
84
71
1! IK:
34 u^
00
83
(id
03
71
84
(it
02
04
nq
81
84
34 02
71
00
83
01
iifi
71
H4
01
in
02
si
84
14 02
71
Input S.imj.
«- Vol . ,
rt1
Input barometer ,"H9
Input :ijn,[,l
e 'IVmp,°F
h,| ui V..I. lulurne,
lUpIlt !ii-l iiii'.
i)utput \\^/
Output pj'-ti
uut/put liq/u
Output ppm.
0 1 2.3
SO 31 S2 S3
A
3 C
, In niLj/in?
In tny/mS.
Ot>u
1
070
ObO
-
090
100
MO
-.Q. .
_ 2...
4
X
R/S
2
4
1
R/S
ECi_2_
0
.
2
8
6
X
-is/s.
0
u
£
R/S
HCI. 2
x
0
2
8
6
X
R/S
3
0
0
r>
R/S
RCL 02
X
0
.
2
8
0
X
R/S
2
9
1
6
bJ
-.- Ui
71
02
- 01
. ._ Hit
81
B4
...34 02
"OO
83
02
08
ll 1 1
.. n ..
- ii-i . .
uu .
.. ..JU
HI
84
34 U2
71
00
HI
02
08
06
71
84
03
00
00
00
81
U4
34 02
7)
00
83 j
02
08
00
71
84
02
09
01
06
output v^/ra-*
Output ppm, In
,9/mt
Output ppiu, In
Output My/m
Output ppiu/ In
Output iiy/in
REGISTERS ._ . „ . .
4
34
5 6 /
S5 36 S7
s
D
y y
Sti b4
F. I
Figure 14 (Cont'd). HP-67 program form,
223
-------
Program I
SIEP KtYENTHY KEY CODE
COMMtNIl SltP KEYtNTHY KtY COIJt
CUMMIN1&
" ~
'~
1."
-- —
14.1
150
lou
1
1 _
R/S
_RCL 02 _
X
0
3
5
9
R/S
4
tl
6
k/s
kri, 2
X
0
7
].
X
R/S
4
j
1
5
R/S
h RTN
ftl
84
3-1 02
71
00
83
03
05
09
LL
84
0
0
i
1
06
0
&
81
34 02
71
0
a
0
0
i
3
7
01
71
84
04
04
01
05
8
]
84
^S5^L_
Output. H'iii, ^" I-'J/111^
J
Output i.^/m
Output
output
Pf-iu, limy
j
Output ppni
If'U
ttJG
— • —
200
;rd —
J.'U
LABELS
A
^ t>
1
1:
c
c
^
7
11
.1
J
8
L
'•'
.1
y
-
-
-
_
.
.
FLAGS
u
1
2
J
SET STATUS
FLAGS TRIG LlISP
ON Ol [-
U II 1 i
i I i , j
i 1 ! i J
J 1 i i .
DLu 1 111 i
I'iltAl) i • SU
HAD i ! 1 r*.
n
Figure 14 (Cont'd). HP-67 program form.
224
-------
SWRI PROJECT NO. TEST NO. TEST DATE:
FUEL: CVS NO. TUNNEL SIZE: DRIVER:
SAMPLE COLLECTION BY:_
GENERAL COMMENTS:
_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
Acetaldehyde 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 mg/ml
I-Bu Aldehyde Cone yg/m3
I-Bu Aldehyde Cone ppm
MeEt Ketone Der Cone mg/ml
MeEt Ketone Cone yg/m3
MeEt Ketone Cone ppm
Cro- Aldehyde Der Cone mg/ml
Cro-Aldehyde Cone yg/m3
Cro-Aldehyde Cone ppm
Hex- Aldehyde Der Cone mg/ml
Hex-Aldehyde Cone yg/m3
Hex-Aldehyde Cone cpm
Benzaldehyde Der Cone mq/ml
Benzaldehyde Cone ya/m3
Benzaldehyde Cone ppm
1
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
2 3 4 5 fi
FTP-2
1.625
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 15. Aldehyde collection sheet,
225
-------
tnol. wt formaldehyde = 30.03 g/mole
30.03 g/mole x 1000 ,_...
density = 24.04& =
ppm = 597.5 yg/m3 4 1249 yg/m£ = 0.478 mVm3 = 0.478 ppm
The calculations for acetaldehyde, acetone, isobutyraldehyde, methylethyl-
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/m and 0.306 ppm
acetone, 663 yg/m and 0.274 ppm
isobutyraldehyde, 141 yg/m3 and 0.047 ppm
methylethyIketone, 630 yg/m3 and 0.210 ppm
crotonaldehyde, 541 yg/m3 and 0.186 ppm
hexanaldehyde, 250 yg/m and 0.060 ppm
benzaldehyde, 775 yg/m and 0.176 ppm
Note: The values used in these calculations are picked from a range of tem-
peratures, derivative concentrations, etc. to validate the calculations and
may not be representative of expected raw data. The calculations are pre-
sented to confirm the manual and HP-67 calculations give the same results.
This was confirmed for six sets of calculations.
LIST OF EQUIPMENT
The equipment required for the analysis of aldehyde 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 £/minute range.
3. Sample pump, Thomas Model 106 CA18, capable of free flow capacity
of 4 £/minute.
4. Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
capacity.
5. Regulating valve, Nupro 4MG, stainless steel.
226
-------
6. Teflon tubing, United States Plastic Corporation, 1/4" OD x
1/8" ID and 5/16" ID x 1/8" ID.
7. Teflon solenoid valve. The Fluorocarbpn 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 #2036/J/1.
13. Iron/Cons tan tan 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.
20. Stainless steel heated filter assembly - 7 cm, Scott, capable of
temperature to 204°C, includes 2 heaters, adjustable thermostat
switch, stainless steel insulated covers and sample bypass solenoid
valves.
21. Glass microfiber filter discs, Reeve Angel 934-AH, Whatman, 7 cm
diameter.
22. Flexible, heavy insulation heating tape, Briskeal***, width-1/2 inch,
length-48 inches.
23. Temperature Controller, Athena, 100-600°F.
24. Heated TFE Teflon hose, Technical Heaters Inc, 5' x 1/4", tempera-
ture limit 400°F.
227
-------
Sample Preparation
1. Fritted glass filters, Ace Glass Company, porosity D, ASTM 10-20
microns pore size, 24/40 ground glass joint, vacuum takeoff.
2. Constant temperature vacuum oven, National Appliance Company.
3. Pump for oven, Thomas Industries, Model 907CA18 2.
4. Flasks, 125 m£ capacity, 24/40 ground glass joints.
5. Separatory funnels, 125 m£.
6. Separatory funnels, 250 m£.
7. Separatory funnel shaker, Burrell Corporation, Wrist-Action ©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 m£.
11. Vials, Kimble, 1/2 dram.
12. Vacuum pump, Sargent-Welch.
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 mi, Hamilton Company, #701.
5. Dual columns, 24 x 1/8" ID, stainless tubing packed with 6.7
percent Dexsil 300 GC on Chromosorb G 60/80 mesh, DMCS treated
and acid washed.
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.
228
-------
1. Hydrochloric acid, HC1, 36.46 g/mole, concentrated (37%), analyt-
ical reagent, Mallinckrodt, Cat. #2612.
2. Pentane, C5Hi2, 72.15 g/mole, Distilled in glass (bp 35-37°C),
Burdick and Jackson Laboratories, Inc.
3. 2,4 DinitrophenyIhydrazine (2 ,4-DNPH) , (NO2) 2C6H3CH=N -NH2,
210.149 g/mole, Aldrich analyzed, Aldrick, Cat. #D19,930-3.
4. Sodium Bicarbonate, NaHC03, 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
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/ANTHRACENE SOLUTION
Toluene containing approximately 0.05 mg 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 phenyhydrazone
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 2NHCl-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.
229
-------
PREPARATION OF STANDARD SOLUTION OF PHENYLHYDRAZONE DERIVATIVES AND ANTHRACENE
A standard containing the phenylhydrazone derivatives and anthracene in
toluene is prepared to obtain a response factor of each of the derivatives
to anthracene. The solution is made by dissolving weighed amounts of an-
thracene and each of the derivatives in a quantitative volume of toluene.
These solutions contain 0.05 mg anthracene per m£ of toluene and 0.2 mg of
each derivative per m& of toluene.
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.
230
-------
APPENDIX B
TOTAL CYANIDE PROCEDURE
231
-------
THE MEASUREMENT OF TOTAL CYANIDE IN EXHAUST
The measurement of total cyanide (hydrogen cyanide and cyanogen) in
dilute exhaust is accomplished by bubbling exhaust through glass impingers
containing a 1.0 N potassium hydroxide absorbing solution. The cyanide
reacts with the potassium hydroxide to form a stable salt which remains in
solution. Upon completion of the test, an aliquot of the absorbing solution
is treated with monopotassium dihydrogen phosphate buffer and Chloramine-T.
The reaction of cyanide and Chloramine-T in the presence of the buffer re-
leases a gas, cyanogen chloride. For analysis, a portion of this cyanogen
chloride gas is injected into a gas chromatograph equipped with an electron
capture detector (ECD). External cyanide standards in 1.0 N potassium
hydroxide are used to quantify the results. The detection limit for this
procedure is less than 0.01 ppm.
SAMPLING SYSTEM
Two glass impingers in series, with each containing 25 m£ of 1.0 N
potassium hydroxide, are used to collect exhaust samples for analysis of
cyanide. A flow schematic of the sample collection system is shown in
Figure 1. The two glass impingers, when maintained at ice bath temperature
(0-5°c), collect 99+ percent of the hydrogen cyanide and cyanogen. The flow
rate through the impinger is maintained at 4 &/minute by the sample pump.
A dry gas meter is used to determine the total flow through the impinger
during a given sampling period. The temperature of the gas stream is
monitored by a thermocouple immediately prior to the dry gas meter. A
flowmeter in the system allows continuous monitoring of the sample flow.
A drier is included in the system to prevent condensation in the pump,
flowmeter, dry gas meter, etc. When sampling from diesel fueled vehicles,
a heated filter, located between the on-off solenoid valve and the dilution
tunnel, is used to prevent diesel particulate from contaminating the sampling
system. The filter and line connecting the filter to the dilution tunnel are
heated to 375 °F in order to keep hydrogen cyanide and cyanogen from being re-
tained on the removed particulate. The Teflon line connecting the heated
filter and the solenoid valve is heated to ~175°F in order to prevent water
from condensing in the sample line. Several views of the sampling system
are shown in Figure 2.
ANALYTICAL PROCEDURE
The analysis of total cyanide (hydrogen cyanide and cyanogen) in exhaust
is accomplished with the use of a gas chromatograph equipped with an :electron
capture detector (ECD). This detector is highly sensitive to halogens and
232
-------
Gas Temperature
Digital Readout
ro
CO
GO
On-Off
Solenoid
Valve
Flowmeter
Sample
Probe
Regulating
Valve
Dilute
Exhaust
Ice Bath
Temperature
Digital Readout
Dry
Gas
Meter
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
234
-------
Solenoid
Impinger
Ice Bath
Close-up of impingers (Side View)
Solenoid
Pump
Rear View
Figure 2 (Cont'd). Total cyanide sampling system,
235
-------
halogenated compounds. In this procedure the cyanide ion (CN~) is reacted
with Chloramine-T (sodium paratoluene sulfonchloramide) to form cyanogen
chloride (CNCl) which can be detected at low concentrations by the BCD.
Because of the sensitivity of the BCD to the halogenated cyanogen chloride,
cyanide can be detected at low concentrations in exhaust by this procedure.
A detailed description of this procedure follows. An analysis schematic for
the procedure is shown in Figure 3.
During each test cycle a portion of the diluted exhaust is bubbled
through two impingers in series, with each impinger containing 25 m£ of l.ON
potassium hydroxide. "The temperature of the impingers is maintained at
0-5°C by an ice water bath, and the flow rate through the impinger is main-
tained at 4 Vwinute throughout the test cycle. Upon completion of each
driving cycle, the impingers are removed and the coritent of each are trans-
ferred to a separate 30 mJi ploypropylene bottle and capped. A 1 mS, aliquot
is removed from qne of the bottles and placed in a 5 m& Glass Reacti-vial.
A 2 m£ aliquot of 1.0 Npotassium dihydrogen phosphate buffer is then added
carefully down the side of the vial. This adjusts the pH to neutral or
slightly acid. A 1 mH aliquot Qf Chloramine-T is then 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 solu-r
tion is then set aside for 5 minutes. This allows the Chloramine-T to
react completely with the trapped cyanide ion. The vial is then vibrated for
5 seconds to release cyanogen chloride into the gas phase. With a gas-tight
syringe a 100 p£ sample of the head space is removed through the septum top
and immediately injected into the gas chromatograph. This procedure is then
repeated for the second impinger. Some of the steps in this procedure are
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 used to separate the cyanogen chloride from other compounds
in the sample. The carrier gas, 95% argon-5% methane, flows through the
column at a flow rate of 40 n\£/minute. The column temperature is isothermal
and maintained at 140°C. Oxygen, carbon dioxide, and water elude from the
column before cyanogen chloride. The sample peak area is determined with a
Hewlett-Packard Model 3354 computer system with a remote teletype printout.
The peak area is compared to the peak area of a standard cyanide ion solution
which is developed in a manner similar to that of the sample. Figure 5 shows
the analytical system with gas chromatograph detector, integrator, and
recorder.
This procedure provides a rapid and sensitive method .for analyzing total
cyanide in exhaust without extensive wet chemical work up. The analysis time
is on the order of about 5 minutes after injection into the gas chromatograph.
The sensitivity of the BCD extends the minimum detectable limit to less than
0.01 ppm cyanide ion with the specified flow rates, absorbing solution vol-
ume, syringe size, vial size, and reagent quantities. This limit can possibly
be extended by changing these parameters. The simplicity and rapid data
236
-------
CVS
Glass Irapinger
Excess Reagent
saved as needed
Aliquot buffered
with KH2PO4
Chloramine-T
added
CNC1 in head gas
analyzed in GC
with BCD
A/D Converter
\
Recorder
Hewlett-Packard
3354 GC
computer system
Figure 3. Total cyanide (HCN + C H ) analysis flow schematic
237
-------
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.
238
-------
Step 3. Reagent addition
Step 4. Sample shaking
Figure 4 (Cont1 d). Various steps in sample collection and
analysis of total cyanide in exhaust.
239
-------
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.
240
-------
IVj
Fiugre 5. Total cyanide analytical system.
-------
turnover makes this procedure ideal for repetitive analysis. A gas chroma-
tograph trace for a cyanide standard is shown in Figure 6.
CALCULATIONS
This procedure has been developed to provide the user with the concen-
tration of total cyanide in exhaust. The results will be expressed in yg/m3
of exhaust and ppm. The equations for determining the concentrations in
pg/m3 and ppm are derived in the following manner.
The first step is to correct the volume of exhaust sampled to a stan-
dard temperature, 68°F, and pressure, 29.92"Hg, by use of the equation
p v P v - -'-, •,-;•
exp x exp = corr x corrt;
T T
exp corr
V = experimental volume of gas sampled in ft
V = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
P - experimental barometric pressure
PSXP = 29.92"Hg
mcorr
T = experimental temperature in °F + 460
TexP = 68°'F + 460 = 528°R
corr
Solving for V'• gives:
corr
Pg ("Hg) x Vgx (ft3) x 528°R
corr ~ T (°R) x 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°R
3OVT^ _ „__
„ GA.IJ tJAM
v vm ) = ——^ £
C0rr x 29.92"Hg x 35.31 ft3/Ht3
(Equation 1)
The next step is to find the concentration of total cyanide in
Since the gas chromatograph BCD has a linear response in the concentration
of concern, then the following equation holds.
Csam (Ug/m£) Cstd
sam Astd
242
-------
Liq. Prim
Support
C«iing_4U _ce/min. Ar/5» CH. Orr*
on 13J/12O mnn pnBAPAK
Run ISO e
» dn P*«
htld « "C ISO tor
hddfor
Inlel
Dtttctor
Hyd '•" pug
Air UA psig
( I NA P»ia NA Rolnmnr Hdo.
C. Huatad ShainU.cs Stujal T»o«
Rotwnvur Rdg.
Rotamnir Rdg. NA
cc/min
cc/min
_ cc/min
Sol tec Tvo.
R«cordtr 1 in/miniDMd 1 rnV.F.S.
65432
Detention time, min.
Figure 6. Typical trace for a standard.
243
-------
C = concentration of the sample in yg/m£
Asam = GC peak area of sample in relative units
CS , = concentration of the standard in yg/m£
AS = GC peak area of standard in relative units
std
Solving for C gives:
3 sam
C (yg/m£) A
sam = r
A . ,
std
The Csam (yg/m£) in solution is corrected for any necessary dilution by
multiplying by the dilution factor, D.F.
, , n, C ., (yg/m£) x A x D.F.
C fyg/m£) std ^ sam
sam =
std
To obtain the total amount of yg of total cyanide in the absorbing
solution, the absorbing reagent volume is multiplied by the concentration
to give:
yg sample = C (yg/m£) x Abs. Vol. (m£)
sam
c *.* (yg/m£)x A x D.F. x Abs. Vol. (m£)
= std sam
Astd
(Equation 2)
To obtain yg sample/m , Equation 2 is divided by Equation 1 to give:
-> C . , (yg/m£) x A x D.F. x Abs. Vol. (m£)
, / 3 std sam
yg sample/m =
A . , x P ("Hg) x 528°
std exp
T x 29.92"Hg x 35.31 (ft3/m3)
V (ft3!
e3!P (Equation 3)
To find the concentration of total cyanide (as HCN) in ppm, the density
of hydrogen cyanide is needed. At 29.92"Hg and 32°F, one mole of gas oc-
cupies 22.4 liters. This volume is corrected to 68°F from the equation
V Vi
T
Vj_ = 22.4£
T! = 32°F 4- 460 = 492°R
V = volume at 68°F
T = 68°F + 460 = 528°R
244
-------
Solving for V gives:
Vn v T 22.4 x 528
= 24.045,
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 22.04 £/mole
mole. wt. g/mole
density (g/£) = m^' ™~ /mo
24.04 H/mole
The density in yg/m£ can be found by converting g to yg and £ to m£ as
follows :
density yg/m£ = ao1' wt: g/mole x 1 * 106 yg/q = mol. wt. x IQQQ
24.04 £/mole x x 103 ^ 24.04
(Equation 4)
To obtain the concentration of total cyanide (as HCN) in ppm, the concen-
tration in yg CN~/m needs to first be converted to yg HCN/m . This is done
by multiplying the concentration in yg CN~/m3 by the ratio of the formula
weight of HCN to the formula weight of CN~.
T ~,, / i ™- / •* formula weight HCN (yg/y mode)
yg HCN/m-5 = yg CN /mj x - - ; - •\J_ ^,- , / - T~T
formula weight CN (yg/y mode)
, . 3 27.026 yg HCN/y mole
= yg CN'/m x 26.018 yg CN'/y mole
- yg CNVm3 x 1.039
(Equation 5)
The concentration of total cyanide (as HCN) in ppm can now be obtained by
dividing by the density in yg/m£.
ppm (HCN) = yg HCN/m * density yg/m£ = m£/m3
Using Equations 3, 4, and 5 gives the ppm concentration in the form of the
raw data.
24.04 (£) x C . . (yg CN~/m£) x A x D.F. x Abs. Vol. (mi)
StQ SoLIu
ppm (HCN) =
^ (g/mole) x 1000 x ^^ x p^ ("Hg)
T (°R) x 29.92 "Hg x 35.31 ft3/m3 x 1.039 ^ CN-
C^
-------
1T*
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEY ENTRY KEY CODE rouu^c
010
OiO
030
040
050
2
t
R/S
X
STO 1
R/S
4
6
0
+
RCL 1
f
R/S
X
STO 2
RCL 2
R/S
X
R/S
X
R/S
-'.
R/S
X
STO 3
R/S
RCL 2
X
R/S
X
R/S
V
R/S
X
R/S
RCL 3
+
R/S
1
•
0
3
8
7
X
1
1
2
4
f
R/S
h RTN
02
81
84
71
33 01
84
04
06
00
61
34 01
81
84
71
33 02 .
34 02
84
71
84
71
84
81
84
71
33 03
84
34 02
71
84
71
84
81
84
71
84
34 03
61
84
01
83
00
03
08
07
71
01
01
02
04
81
84
35 22
In Barometric "Hg
In Sample Temp. °F
In Sol. Vol. , mi!.
In Dilution Factor
In Std Cone ug/mJl
In Standard Area
In Sample Area,
Bubbler #1
Out Sara. Cc
Bubbler #1
In Di 1 utic
In Std Cone
Input Std.
In Sample t
Bubbler
Ou-t Sam. Cc
Bubbler #2
Out Cone, V
Output ppm
, pg/m3-
n Factor
, ug/mfc
Area
ire a,
82
me,
., pg/mj
ig CN~/m3
HOI
060
070
060
090
100
110
...
REGISTERS
0123
SO SI S2 S3
A i
i C
4
S4
5 6 7
S5 S6 S7
D
8 9
S8 S9
E 1
Figure 7 (Cont'd). HP-65 program form.
247
-------
Hewlett-Packard Calculations
In order to insure maximum turnaround in a minimum time period, a
Hewlett-Packard 67 program was developed to calculate the total cyanide
concentration in yg/m3 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 por-
tion of a three-bag 1975 FTP. Raw data for these tests are presented in
Fiugre 8. Calculations were performed using the HP-67 program and manual
calculations.
Manual Calculations for Driving Cycle Cold-FTP
For Bubbler #1
C . (yg/m£) x A x D.F. x Abs. Vol.
_ - , 3 std sam
yg CN /m =
AstdX Pexp ("Hg)
Texp x 29.92"Hg x 35.31 ft3/m3
X 528°R x v (ft3)
exp
= 5.0 yg/m& x ISQQ x i x 25 mfc
2000 x 29.19"Hg
x (460° + 70°) x 29.92"Hg x 35.31 ft3/m3
528o x 3.453 ftJ
986 yg CN~/m3
•Hie concentration in bubbler #2 is calculated in the same manner using the
appropriate dilution factor, standard concentrations, standard area, and
sample area:
For Bubbler #2
yaCN~/m3 - 1 ^g/m x 500 x i x 25 ml.
• * /m
x (460° + 70°) x 29.92"Hg x 35.31 ft3/m3
528° x 3.453 ft-3
131 yg CN~/m3
248
-------
SWRI PROJECT NO.11-1234 TEST NO. 001
TUNNEL SIZE: 18" DRIVER: R.R.
TEST DATE: 11-1O-79 VEHICLE: Practice
MILES: 1000
FUEL: EM-237 CVS NO. 3
SAMPLE COLLECTION BY: D'E'B- CHEMICAL ANALYSIS BY: ' ' 'CALCULATIONS BY:
GENERAL COMMENTS:
Test No.
ro
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Absorb . Rea . Vol . , m£
Dilution Factor, Bubbler #1
Std. Cone UgCN~/m& Bub. #1
Std. Area - Bubbler #1
Sample Area - Bubbler #1
Sample Cone pgasT/m3 , Bub #1
Dilution Factor, Bubbler #2
Std. Cone ygCN~/m£ Bub. #2
Std. Area - Bubbler #2
Sample Area - Bubbler #2
Sample Cone ugCN~/m^,Bub#2
Total Cone. ygCN~/m3
Total Cone, ppm HCN
Cold FTP
3.453
29.19
70
25
1
5
2000
1500
986
1
1
1000
500
131
1117
1.03
Hot FTP
3.486
28.66
75
25
5
2
1500
1800
3213
1
1
2000
1000
134
3347
3.09
SET-7
3.508
29.33
80
50
10
1
3000
2500
4374
1
2
3000
500
175
4549
4.20
HFET
1.926
29.40
85
50
2
1
5000
4500
1732
2
0.5
2000
200
96
1829
1.69
NYCC
1.525
29.10
90
25
1
2
10,000
9000
1115
I
I
3000
1900
392
1508
1.39
Backgroui
15.826
29.04
77
25
1
1
58
4000
2000
1
0.5
2000
500
7
37
0.03
d
Figure 8. Raw data sheet for total cyanide.
-------
The concentrations from the two bubblers can be added for a total concentration:
Total yg CN~/m3 = cone (bubbler #1) + cone (bubbler #2)
= 986 yg CN /m + 131 yg CN /m
= 1117 yg CN~/m3
ppm CN~(as HCN) = yg HCN/m -r density yg/m£
/ n Mol. Wt. (HCN) X 100
density yg/mJ6 = --4 ^ Q4j
24J
Mol. Wt. HCN = 27.026 g/mole
. „ 27.026 x 100 nn_. ,, . „
density yg/m£ = -. - = 1124 yg/mJ6
^4r • V/4
yg HCN/m3 = yg CN~/m3 x 1.039 yg HCN/yg CN~
= 1117 x 1.039 = 1161
ppm CN'(HCN) = 1161 T 1124 = 1.03
Note: The values used in these calculations are picked from a range of tem-
peratures, standards, dilution factors, etc. to validate the calculations
and may not be representative of expected raw data in all cases. These cal-
culations are presented to confirm that manual and HP-67 calculations give
the same results. This was confirmed on six sets of calculations.
LIST OF EQUIPMENT
The equipment required in this analysis is divided into three basic
categories: sample acquisition,, sample preparation, and sample analysis.
Manufacturer, stock number and any pertinent descriptive information are
listed.
Sample Acquisition
1. Sample pump, Thomas model 106 CA18, capable of free flow capacity
of 4 £/minute.
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 Vminute range.
4. Regulating valve, Nupro 4MG, stainless steel
250
-------
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" rn
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, connec-
tors , etc.
9. Drying tube, Nalgene Corporation, 10 cm length x 1/2 in. dianeter
10. Digital readout for dry gas meter
11. Miscellaneous electrical switches, lights, wirings, etc.
12. Six channel digital thermometer, Analog Devices, Model #2036/J/1.
13. Iron/Constantan type J single thermocouple with 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation
14. Stainless steel heated filter assembly - 7 cm; Scott, capable of
temperature to 204°C, includes 2 heaters, adjustable thermostat
switch, stainless steel insulated covers and sample bypass solenoid
valves
15. Glass microfiber filter discs, Reeve Angel 934-AH, Whatman, 7 cm
diameter
16. Flexible heavy insulation heating tape, Briskeal®, width-1/2 inch,
length-48 inches
17. Temperature Controller, Athena, 100-600°F
18. Heated TFE Teflon hose, Technical Heaters, Inc., 51 x 1/4", tem-
perature limit 400°F.
Sample Preparation
1. Glass gas syringe, Teflon tipped plunger, 100 yl, Pressure-Lok
Series A-2, Alltech Associates
2. Glass Reacti-vials, 5 m&, Pierce Chemical Company
3. Class A, 1 mJl volumetric pipets
4. Class A, 2 m£ volumetric pipets
5. Class A, 25 m& volumetric pipets
251
-------
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 mi volumetric flask
11. Vortex-Genie, Scientific Industries, Inc. Model K-550-G
Instrumental Analysis
1. Perkin-Elmer Model 3920B gas chromatograph equipped with a lin-
earized electron capture detector (BCD)
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 glass iiopingers using
a 1.0 N potassium hydroxide absorbing reagent. After collection, a buffer
potassium phosphate monobasic is added to control the pH followed by
Chlonnaine-T to convert the CN~ to cyanogen chloride. Potassium cyanide is
used as the CN~ standard in 1.0 N KDH. The reagents are listed below along
with the manufacturer and quality.
1. Potassium phosphate monobasic, formula -weight = 139.09, 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 = p-CI^Cgl^SC^NClNa'ai^O,
Assay (by titration) 96% minumum, Eastman, crystals, Eastman
Code 1022
PREPARATION OF REAGENTS
Primary Standard - the primary standard is prepared by dissolving 0.602
grams of KCN in 500 mi of 1.0 N KOH. This is equivalent to 500 ppm HCN
(500 tig HCN/m£) or a 481 ppm CN~ (481 yg CN-/m£). Additional standards are
prepared from the primary standard. A typical dilution to prepare a 0-10
pg CN~/m£ calibration curve is as follows:
252
-------
m£ of 481 yg CN /m£ Final Diluent QJ- concentration,
Primary Standard Volume, m£ yg/CN~/m£
1.000 m£ 50.0 m£ 9.62
4.000 m£ 250.0 m£ 7*.70
1.000 m£ 100.0 m£ 4.8!
3.000 m£ 500.0 m£ 2.89
1.000 m£ 250.0 m£ 1.92
1.000 m£ 500.0 m£ 0.96
Buffer Solution - A 1.0 M KH2PO4 buffer solution is prepared by dissolving
13.609 g KH2PO4 in 100 m£ of deionized H20. The buffer solution should be
prepared daily.
Absorbing Reagent - The absorbing reagent is a l.ONKOH solution. This
solution is prepared by dissolving 56.11 grams of KOH in 1000 m£ of deionized
water.
Chlormaine-T - The Chlormaine-T converts the -CN~ to CNC1. 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.
253
-------
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.P., 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 Tabakforschung,
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., Tabacco 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
Valentour,J.C., Aggarwal, V., and Sunshine, I., Anal. Chem., Vol. 46, pg.
924, 1974
Runge, H., Z. Anal. Chem., Vol. 189, pg. Ill, 1962
254
-------
APPENDIX C
INDIVIDUAL HYDROCARBON PROCEDURE
255
-------
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 (Cttg.) , ethane (C2Hs) , ethylene
(C2H4) , acetylene (C2H2) , propane (C3H8) , propylene (€3%), benzene (CQHQ) ,
and toluene (07%). 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 concen-
trations 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. C2 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,
olefins, 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 (HgSO^ and 20 percent sulfuric acid
(H2S04) 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£/minute.
The temperature program sequence is accomplished with the oven of the
256
-------
Vehicle
CVS
Tedlar Bags
Gas
Ch roma t o g r aph
A/D Converter
Recorder
HP 3354
Computer System
Teletype
Printout
Figure 1. The analysis flow schematic for individual hydrocarbons.
257
-------
I 3
01
00
Figure 2. Analytical system for individual hydrocarbons.
-------
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
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 tem-
perature operation.
Samples as well as backgrounds are collected in Tedlar bags during the
driving cycle. The sample is purged through two 10 m£ sample loops for
four (4) minutes. Samples from diesel fueled vehicles are passed through an
ice trap before entering the sample loops. The ice trap removes high mole-
cular weight compounds that can interfere with later analyses. The ice trap
consists of 8 feet of 1/4 inch stainless steel tubing submerged in an ice
bath. The initial configuration of the analytical system is shown in Figure
3. Upon injection, gas sampling valve A is activated by solenoid valve G,
the temperature program sequence is started, and the first timer begins to
count 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 analytical 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 670 seconds, the
second step begins with solenoid H activating gas sampling 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 Columns
- Ill 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 loop 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
259
-------
PRESSURE
ro
cr>
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
MifiJ COLUMN IV
Figure 3. Initial analytical system configuration.
-------
ro
PRESSURE
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
COLUMN IV
Figure 4. Step 1 solenoid G activation of gas sampling valve A.
-------
ro
CT»
ro
PRESSURE
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
*••••••••••Ja«••••••!•»•••*••
liQMfti COLUM
COLUMN IV
Figure 5. step 2 solenoid H activation of gas sampling valve B.
-------
ro
CTl
CO
PRESSURE
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
• ••••••*•••J«•••**•••••••*••• »•
[MM] COLUMN
IV
Figure 6. Step 3 solenoid E activation of gas sampling valves C and J.
-------
PRESSURE
ro
en
SOLENOID VALVE
SAMPLE IN
SAMPLE LOOP A
HELIUM
CAPILLARY RESTRICTOR
SEISCOR VALVE
IV
Figure 7. Final analytical system configuration.
-------
en
120,-
100 -
80 -
S 60
111
40
20
STEP 2 STEP 3
STEP 4
ISOTHERMAL OPERATION-
STEP 1
12 15
Time, Minutes
18
21
24
27
30
Figure 8. System operation sequence.
-------
FID DETECTOR
CT)
CT>
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 9. Sample purge position prior to sample injection.
-------
FID DETECTOR
po
SAMPLE IN
HELIUM
SAMPLE OUT
RESTRICTOR
HELIUM
Figure 10. Step 1 sample loop A injected.
-------
FID DETECTOR
ro
01
oo
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 11. Step 2 sample loop B injected.
-------
FID DETECTOR
ro
SAMPLE IN
HELIUM
SAMPLE OUT
HELIUM
RESTRICTOR
Figure 12. Step 3 simultaneous solenoid C and J activation.
-------
FID DETECTOR
po
—i
o
SAMPLE IN
HELIUM
SAMPLE OUT
RESTRICTOR
HELIUM
Figure 13. Step 4 backflushing of column III.
-------
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 backf lushing of Column III to the vent. The system remains in this
configuration until it is reset to the initial purge position. A sample
chromatogram for the calibration blend of individual hydrocarbons is shown
in Figure 14.
CONTROL SYSTEM
The control of the five Seiscor gas sampling valves is accomplished by
ATC timers and ASCO electric 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 sole-
noid valve G and gas sampling valve A. This accomplishes the complicated
column flow sequence required for this analysis. The system repeatability
is ± 1.8 percent.
CALCULATIONS
The concentration of the individual hydrocarbons aro 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:
Astdn _ Asamn
Cstdn Csamn
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:
Cstd
,
Astdn
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 ppmC. The peak area of methane
in the sample was 8593 area counts. Calculate the concentration of methane
in the exhaust.
271
-------
ro
»-sJ
ro
22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 10
Retention time, min.
Figure 14. Calibration blend for specific hydrocarbons.
-------
de-energized configuration
energized configuration
Figure 15. Solenoid valve G with gas sampling valve A
273
-------
Csamn -- Adtdn
, ^ •• % (8593) (13.616)
Csam (methane) = - (11893) -
Csam (methane) = 9.84 ppm C
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.
As am., x Cstd-
C5amn= ^
. (10973) (21.272)
Csam (toluene) =
Csam (toluene) = 12.59 ppm C
Note: "These sample calculations are presented as a example only and are not
necessarily representative of expected values in exhaust.
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
274
-------
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 Telfon 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, 36/60 mesh Type 58 Silica Gel
11. Column III, 15' x 1/8" SS, 15 percent 1, 2, 3-tris (cyanoethoxy)
Propane on 60/80 mesh Chromosorb PAW
12. Column IV, 2' x 1/8" SS, 40 percent Hg;so4 on Chromosorb W
13. Miscellaneous electrical switches, wiring, lights, etc.
275
-------
REFERENCES
Altshuller, A.P. and Bufalini, J.J. Environ. Sci. Tech., Vol. 5, pg. 39,
1971.
Federal Register, Vol. 37, No. 221, 24270-77 Nov. 1971.
Klosterman, D.L. and Sigsby, J.E., Environ. Sci0 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, E.D., Environ. Sci. Tech., Vol. 5, No. 3,
pg.*223, 1979.
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, 196,8.
Black, F.M., High, L.E. and Sigsby, J.E., J. Chromatog. Sci., Vol. 14,
pg. 257, May 1976.
Private communication between Mobile Source Emissions Research Branch, ESRL-
EPA, Research Triangle Park and Southwest Research Institute.
276
-------
APPENDIX D
ORGANIC AMINES PROCEDURE
277
-------
THE MEASUREMENT OF ORGANIC AMINES IN EXHAUST
The organic amines that are included in this analysis are: monomethyla-
mine, dimethylamine, trimethylamine, monoethylamine, diethylamine, and
triethylamine. Dimethylamine and monoethylamine are not resolved from each
other under normal operating conditions and are reported together as C2H7N.
The measurement of organic amines 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. For analysis, a portion of the sulfuric acid 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 quantify the results. Detection limits for this
procedure are on the order of 0.002 ppm in dilute exhaust.
SAMPLING SYSTEM
A glass impinger containing 25 m£ of 0.01N 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 1. The single
glass impinger is sufficient to collect 99+ percent of the organic amines.
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 4Jl/minute by the
sample pump." A dry gas meter is used to determine the total flow through!
the impinger during a given sampling period. The temperature of the gas
stream is monitored by a thermocouple immediately prior to the dry gas meter.
A drier is included in the system to prevent condensation in the pump, flow-
meter, dry gas meter, etc. The flowmeter in the system allows continuous
monitoring of the sample flow to insure proper flow rates during the
sampling. When sampling from diesel fueled vehicles, a heated filter,
located between the solenoid valve and the dilution tunnel, is used to
prevent diesel particulate from contaminating the sampling system. The
filter and the line connecting the filter to the dilution tunnel are heated
to 3,75°F in order to keep the organic amines from being retained on the
removed particulate. The Teflon line connecting the heated filter and the
solenoid valve is heated to ~175°F in order to prevent water from condensing
in the sample line. Several views of the sampling system are shown in
Figure 2.
ANALYTICAL PROCEDURE
The analysis of the organic amines (monomethylamine, dimethylamine,
trimethylamine, monoethylamine, diethylamine, and triethylamine) is
278
-------
ro
•vj
vo
Sample
Probe
Gas Temperature
Digital Readout
Heated
Filter
On-Off,
Solenoid
Valve
Plowmeter
V
Sample
Pump
Heated
Lines
Dilution
Tunnel
Dilute
Exhaust
Ice Bath
Tempe rature
Digital Readout
Regulating
•••' Valve
Dry
Gas
Meter
HTalakHBl
Gas Volume
Digital Readout
Figure 1. Organic amines sample collection flow schematic.
-------
Front View
igital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
280
-------
Solenoid
Impinger
Ice Bath
Solenoid
Filter
Ice Bath
Drier
Dry Gas Meter
Pump
Figure 2 (Cont'd). Organic amines sampling system,
281
-------
accomplished by trapping the amines in sulfuric acid and analyzing the sample
with a gas chromatograph equipped with an NPD. The NPD is highly sensitive
to organic nitrogen compounds and relatively insensitive to inorganic nitrogen
compounds. The analysis flow schematic for the organic amines is shown in
Figure 3. .A detailed description of the procedure follows.
For the analysis of the organic amines, dilute exhaust is bubbled through
a glass impinger containing 25 m£ of 0.01N sulfuric acid. Upon completion of
each driving cycle, the impinger is removed and the contents are transferred
to a 30 m& polypropylene bottle and capped. The amines, as their sulfate
salts, can be stored in solution for long periods of time without decomposition.
A Perkin-Elmer 3902B 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, Figure 4, the amines
are released 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 C2H_N. The carrier gas is helium which flows through the column
at a rate of 30 mi/minute. The column temperature is 130°G for 4 minutes
and then programmed to 170°C at a rate of 32° a minute. In a chromatogram
of a standard sample containing all six of the amines, Figure 5, the first
peak is monomethylamine, followed by the combined peak of dimethylamine-
and monoethylamine, C H N, and then by peaks of trimethylamine, diethylamine,
and triethylamine. To quantify the results, the sample peak areas are
compared to peak areas of standard solutions. Figure 6 showa the analytical
system with gas chromatograph, detector, A/D converter, and recorder.
CALCULATIONS
The procedure has been developed to provide the user with the concentra-
tion of the organic amines (monomethylamine, total dimethylamine and mono-
ethylamine as C2H7N, trimethylamine, diethylamine, and triethylamine) 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/m3
and ppm are derived in the following manner.
The first step is to correct the volume of exhaust sampled to a standard
temperature, 68°F, and pressure, 29.92"Hg, by use of the equation*
P x V P x V
exp exp _ corr corr
T T
exp corr
282
-------
CVS
\
Glass
Impinger
Unused Sample
saved as needed
Sample analysis
in gas chromatograph
with ascarite pre-columr
and NPD
A/D Converter
Recorder
Hewlett-Packard
3354
Computer System
Figure 3. Organic amines analysis flow schematic
283
-------
Figure 4. Ascarite pre-column,
284
-------
Sample injection
1.4 ppm methylamine
1.6 ppm ethylamine +
dimethylamine
1.5 ppm trimethylamine
0.9 ppm diethylamine
0.9 ppm triethylamine
Date January
Operator
4mm l.D.
Instrument
Column
%wt.Carbowax 2QM+0.8%Liq. Phase
mesh Carbopack B » KOH Support
NA °Cusina 30 cc/min. Helium Carrier
1 GO psig
held® 130
Rotameter Reading
/min. Held for
min. (other)
Heated-Glass Lined
min.. Prog to C at
held for
Inlet 20
psig NA Rotameter Rdg. 3
psig NA Rotameter Rdg. _ 10 Q
psig Rotameter Rdg.
1 mV.F.S.
ul net
Recorder 1 in/min speed
Injection 10 ul indicated
Sampling Device 10 Ul
~8 6.4 2
Retention time, min.
Figure 5. Chromatogram of amine standard.
285
-------
•
'
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"Hg
pexp = experimental barometric pressure
Pcorr = 29.92"Kg
Texp = experimental temperature in °F +• 460
Tcorr = 68°F + 460 = 528°R
Solving for Vcorr gives:
Pexp ("H9) x Vexp (ft3) x 528°R
v
corr Tex (OR) x 29.92"Hg
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.
pexp("H9)* vexp (ft3) x 528°
vcorr(m3) =—
Texp x 29.92"Hg x 35.31 ft3/m3
(Equation 1)
The next step is to find the concentration of each of the amines in
yg/m£. Since the gas chromatograph NPD has a linear response in the con-
centration of concern, then the following equation holds.
Csam (ug/mA) Cstd (ug/m£)
Astd
Csam = concentration of the sample is yg/m£
Asam = GC Peak area of sample in relative units
cstd = concentration of the standard in yg/nd
Astd = GC Peak area of standard in relative units
Solving for C gives:
cstd
The Csam(yg/irJl) in solution is corrected for any necessary dilution by
multiplying by the dilution factor, D.F.
Cstd (ug/m£) x Asam x D.F.
= - -- — -
Astd
287
-------
To obtain the total amount in ug of each amine in the absorbing solution,
the absorbing reagent volume is multiplied by the concentration to give:
Ug sample = Csam (yg/mA) x Abs. Vol. (mjl)
Cstd (yg/m&) x Asam x D.F. x Abs. Vol. (m£)
Astd
(Equation 2)
To obtain yg sample/m3, Equation 2 is divided by Equation 1 to give:
Cstd (ug/m£) x Asam x o.F. x Abs. Vol. (mfc)
Ug samp/m3 =
528°
T x 29.92"Hg x 35.31 (ft3/m3)
-exp
x -
Vexp (ft3)
(Equation 3)
To find the concentration of each amine in ppm, the densities of the
amines 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 _ Vj
T ~ T]_
Vj_ = 22.4fc
Ti = 32°F + 460 = 492°R
V = volume at 68°F
T = 68°F + 460 = 528°R
Solving for V gives:
Vj_ x T 22.4 x 528
V = = = 24.045,
Tl 492
Since one mole of gas occupies 22.045, at 68°F, the density can be found in
g/£ by dividing the molecular weight in g/mole by 22.04
mol. wt. g/mole
= 24.04
288
-------
The density in ug/m£ can be found by converting g to ug and £ to mJl as
follows :
mol. wt. g/mole 1 x 106ug/g moi. wt. x looo
den Ug/m£ = - x --- - .. __
24.04 Vmole 1 x I03m£/Jl 24.04
(Equation 4)
To obtain the concentration of each amine in ppm, the concentration in
is divided by the density in ]ag/m£
ppm = iag/m3 T yg/m£ = — =•
irr
Using Equations 3 and 4 gives the ppm concentration in the form of the raw
data.
24.04(£) x cstd (Vig/mfc) x Asam x o.F. x Abs. Vol. (m£)
- -
Mol. Wt. (g/mole) x IQOO x Astd x pexp (»Hg)
Texp(°R) x 29.92"Hg x 35.31 ft3/m3
528°R x vexp (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 67 program was developed to calculate the organic amine
concentrations in yg/m3 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 67 program and
manual calculations.
289
-------
tser I ii si rufi ions
"raanis Amines in Exhaust
STEP
°1
3?
31
1
;
3
4
5
6
7
*
3
10
LI
12
i ^
14
i ^
•
H
13
J.9 .
>C
21
22
23
24
25
;e
2"
;::
3"
NSTRUCT1ONS
Switch to on; =wii7n to run
Feed side I of cari in from right to left
Set decimal place
Input Sample Volusie
Input - Barometric Pressure
Input - Sample Temperature
Input Dilution Factor
Input - Absorbing Reagent Vol.
Input - Standard Cone. CH3NH2
Input - Standard Area CH3NH2
Input - Sample Area CHiNH?
Output Samp le ~or.c . CH }NH7
Output - Sample Cone. CI^MH?
Input - Standard zone. Co^N
Input - Standard Area CT-^N
Input - Dannie Area ~->H7N
Output Sample "one . oH-rN
Output 3'^niple '7criC . 7;H-N
._Inp_ut_ -_ 3tar.dar<3 Ocnc.. jCHjJjN
Input - ritr.adari Area <~H3)3N
Incut - Sample Area (CH-0 jN
_ -QUtpj^t _ - 5ampJ.e Cone . -,' CH^)_^N
Output - Sample Cone. (013)3^
Input - Standard, Cone. CC^H^-jNH
Input - Standard Area (C2H5)2MH
Input - Sample Area (C2H5)2NH
Ou^ut - Sample Cone. (C->H5)^NH
Output - Sample Cone- (C-jH^jj^H
Input - Standard tore. 'C^Hq) -,>J
Input_' ..Standard Area '^^S1 ^!
Input - Sample Area -C--.H-J ,N'
Output - Sample Cone. >'C-,Hri ,n
INPUT
DATA UNITS
"f^T*
"Hg
3C<
mi
ug/m?
Ug/mJl
.g/nul
yg/mf.
liri/mJ.
KEYS
9 ^Cl
A
R/'S
H/5
R/S
R/S
R/S
VS
3/S
R/S
;
R/3 •
R/S i
R/S '
R/S
R/S
R/S
R/S
R/S .
R/S !
R/S :
R/S i ,
R/S ' ;
R/S
R/3 1
R/3
R/ .3
n •:"^;
OUTPUT
DATA. UNITS
ug/m3
Dom
-g/^T"" ~
ppm
..- — ..
'^g/m
"ppn " "~"'
•*Z? "~
ppm
__ -
-cr/mj
I
i
i
Figure 7. HP-67 user instructions.
290
-------
STEP
".'-:
C20
030
KEY ENTRY
0
•
5
0
0
y
i/~
V
5 TO 1
R/3
4
6
n
a.
srr. 1
:n x" *y
h 1/x
R/S
X
R/S
X
^TT} ?
R/S
PCL 2
X
R/S
-r
R/S
X
s/s
1
2
•XG
050
^
f
"/3
RCL 2
X
R/S
=•
R/S
X
R/S
1
q
7
^
R/S
RCL 2
H
•3 / -
" "p/i
KEY CODE
OO
33
05
on
,-,o
n
84
71
"3 Tl
34
:14
Ofi
TO
Si
34 01.
35 52
81
35 62
84 •
7L
84
71
33 02
34
34 02
71
34
81
84
71
34
01
32
09
02
91
34
34 02
71
84
31
84
71
34
01
08
07
05
31
84
34 02
Jl
-14
. ...__x± ..
IP. cut-
Incut
COMMENTS STEP KEY ENTRY KEY CODE COMMENTS
ft 3
Barometer, "Ha
-Sample TemD,3F
Input Dilution Facto
Input Abs Sol Vol,
In Std Cone ug CH3NH
In St
In 5'
out :
Out '
In 34
In S
In S<
Out v
•lut I
In 3
"n .T
.d Area CH^NH2
line Area CH-NH-,
:onc ja/cn3 CH3NH
:onc ppm '7H3NH"-)
:d :onc -g/nt
:d Area C7K7N
imo Area C2H7N
:onc ug/mi
C2H7N
rone ?ptn C^^'i^
:d Cone jg/rai
tri -Xr^a CCH-j} V
^rr, \r3 ;"•!•: ;t
° [' p :i i'
r.-t -p-l—.L.
-^
«0
R/S
2
4
5
/I
34
02
34
:5
9 i 09
1 — J^ —
. =1
-•'I. 2 '4 " '
'< "1
5; •= .j^
i '• i 31
~--'S 34
X -1
I -V3 34
080
i 03
}
-I
2
f
R/S
RCL 2
X
1 00
04
02
81
84
34 02
71
R/S i 84
*
S/S
•>:
?/s
4
2
3
1
81
34
71
34
04
02
00
•>9
f 31
p; ! s/s ^4
'-. ~TM " r 22
1
!
| i
TOO
i '•;
--
1
" ".-"
5 r
r
r)
... —
Out Cone 'jg/ra3
iCH3)3M
"ut "j-nc -pm ic:?-}^';
In ftd Area
Ir. S^TO Aroa
Out "one '..g/mJ
f°2H5' 2N'H
Out Cone ppm
(C2H5)2NH
In Std Cone _ig/mi?,
(C2H5) 3N
In Std Area (C2Hs)3^
In Samp Area
^"2^5) 3N
Out Cone ug/tn^
:ut Cone r;^n
I' :' |'
e" " """ " »'
Figure 7 (Cont'd). HP-67 program :form.
291
-------
SWRI PROJECT NO.
FUEL: CVS NO.
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
Dilution Factor
Absorb. Rea. Vol. Ml
Stan. Cone. CH-,NH,yg/ml
Stan. Area CH3NH2
Sample Area CH3NH2
Sample Cone. CH-,NH9 ug/m3
Sample Cone. CH,NH, ppm
Stan. Cone. C2H7N ug/ml
Stan. Area c?H7N
Sample Area C3H-7N
Sample Cone. C7H7N ug/1"3
Sample Cone. C-jH^N ppm
Stan. Cone. (CH,)7N ug/«l
Stan. Area (CH-,),N
Sample Area (CH3) ^N
Sample Cone. (CH^) -^N ug/m3
Sample Cone. (CH-,) 7N ppm
Stan. Cone. (C2H5) 2NH yg/ml
Stan. Area (C3Hc;)7NH
Sample Area (C,Hq),NH
Sample Cone. (c,H^),NH ug/m3
Sample Cone. (C->H^) 7NH ppm
Stan. Cone. (C-jH,;) 3N ug/ml
Stan. Area (C-,HO 7N
Sample Area (C2H^) -;N
Sample Cone. (C^Hs) •$} ua/m3
Sample Cone. (C2Hs) 3N ppm
FTP-1
3.196
29.30
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
0.-098
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
0.814
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
0.0^0
NYCC
1.070
29.50
39
1
75
0.10
5000
2700
141
0.109
0.20
3000
2000
348
0.186
1.00
2000
1630
2130
0.865
0.40
3100
2810
946
0.311
0.20
5620
4020
373
0.089
Figure 8. Organic amine sample collection sheet.
292
-------
Manual calculations for driving cycle FTP-1
Cstd(yg/m£) x Asam x D.F. x Abs. Vol. (m£)
yg/m3 CH3NH2 = _—_
Astd X Pexp <"Hg)
Texp x 29.92"Hg x 35.31 ft3/m3
x
528°R x vexp (ft3)
(0.05 yg/m£) x 640 x l 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 yg/m3
PPM CH3NH2 - yg/m3 -f density yg/mJl
Mol. Wt. (CH3NH2) x 1000
density yg/mJl =
24.04£
Mol. Wt. CH3NH2 = 31.058 g/mole
31.058 x 1000
density = = 1292 yg/m£
24.045,
ppm =8.99 yg/m3 -f 1292 yg/m£ = 0.007 m£/m3 = 0.007 ppm
The calculation 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: C2HyN, 18.7 yg/m3 and 0.01 ppm; (CH3)3N, 0.05
Hg/m3 and 0.039 ppm; (C2H5)NH, 285 yg/m3 and 0.094 ppm; and (C2H5)3N,
410 yg/m3 and 0.098 ppm.
Note: The values used in these calculations are picked from a range of
temperature, standards, dilution factors, etc., to validate the calculations
and may not be representative of expected raw data. These calculations are
presented to confirm the manual and HP-67 calculations give the same results.
This was confirmed on six sets of calculations.
293
-------
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. The preparation of the absorbing solution, the ascarite pre-column
and the primary standards are also discussed.
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 &/min range.
3. Sample pump, Thomas Model 106 CA18, capable of free flow capacity of
4 5,/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.
9. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps, connectors,
etc.
10. Digital readout for dry gas meter.
11. Miscellaneous electrical switches, lights, wirings, etc.
12. Six channel digital thermometer, Analog Devices, Model #2036/J/1.
13. Iron/Constantan type J single thermocouple with 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation.
14. 30 m£ polypropylene sample storage bottles, Nalgene Labware, Catalog
#2006-0001.
15. Sulfuric Acid, H2S04, formula weight - 98.08, Certified 1N by Fisher
Scientific Company, #SO-A-212.
294
-------
16. Class A, 10 mJl volumetric pipet.
17, Class A, 1000 m& volumetric flask.
18. Stainless steel heated filter assembly - 7 cm, Scott, capable of
temperatures to 204°C, includes 2 heated, adjustable thermostat
switch, stainless steel insulated covers and sample bypass solenoid
valves.
19. Glass microfiber filter discs, Reeve Angel 934-AH, Whatman, 7 cm
diameter.
20. Flexible, heavy insulation heating tape, Briskeat®, width - 1/2 inch,
length - 48 inches.
21. Temperature Controller, Athena, 100-600°F.
22. Heated TFE Teflon hose, Technical Heaters Inc., 5' x 1/4",
temperature limit 400°F.
Instrumental Analysis
1. 10 ]il syringe, Pressure-Lok, Precision Sampling Corporation.
2. Perkin-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 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.
7. 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-HC1, formula weight = 81.55,
crystals, Eastman #94.
10. Trimethylamine Hydrochloride, (CH3)3N-HC1, formula weight = 95.57,
crystals, Eastman #265.
295
-------
11. Ethylamine Hydrochloride, C2H5NH2-HC1, formula weight = 81.55,
crystals, Eastman #731.
12. Diethylamine Hydrochloride, (C2H5)2NH'HC1, formula weight = 109.6,
crystals, Eastman #2090.
13. Triethylamine Hydrochloride, (C2H5)sN-HCl, formula weight = 137.65,
crystals, Eastman #8535.
Preparation of Absorbing Solution
The absorbing solution (0.01 N ^804) is prepared by diluting 10 mi of
1 N sulfuric acid (certified Fisher 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 4. 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 than 10 ppm are prepared by diluting higher concen-
tration standards with 0.01 N sulfuric acid.
296
-------
REFERENCES
Broker, W. and Mossman, A. L., Ma the son 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., Vol0 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.
Hosier, 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 Hosier, A. R. , Anal. Chem., Vol. 45, pg. 1971, 1973.
Umbreit, G. R. , Nygren, R. E. , and Testa, A. J., J. Chromatog., Vol. 43,
Pg. 25, 1969.
, J. N. and Hardy, R., J. Sci. Food Agr., Vol. 23, pg. 9, 1972.
Bowen, B. E., Anal. Chem., Vol. 48, pg. 1584, 1976.
297
-------
APPENDIX E
SULFUR DIOXIDE PROCEDURE
298
-------
THE MEASUREMENT OF SULFUR DIOXIDE IN EXHAUST
The concentration of sulfur dioxide, S02, in automotive exhaust can be
determined as sulfate using an ion chromatograph. Sulfur dioxide exhaust
samples are collected in two glass bubblers, each containing 3 percent
hydrogen peroxide. The temperature of the absorbing solution is kept at 0°C
by means of an ice water bath. The bubbled samples are analyzed on the ion
chromatograph and compared to standards of known sulfate concentrations.
SAMPLING SYSTEM
Two glass impingers in series, each containing 25 m£ of a 3 percent hy-
drogen peroxide solution, are used to collect exhaust samples for the analy-
sis of sulfur dioxide. A flow schematic of the sample collection system is
shown in Figure 1. The two impingers together trap approximately 99 percent
of the sulfur dioxide. The temperature of the impinger is maintained at
0-5°C by an ice water bath, and the flow rate through the impinger is main-
tained at 4 £/minute by the sample pump. A dry gas meter is used to determine
the total flow through the impinger during a given sampling period. The
temperature of the gas stream is monitored by a thermocouple immediately prior
to the dry gas meter. A drier is included in the system to prevent condensa-
tion in the pump, flowmeter, dry gas meter, etc. The flowmeter in the system
allows continuous monitoring of the sample flow to insure proper flow rates
during sampling. When sampling diesel fueled vehicles, a heated filter,
located between the on-off solenoid valve and the dilution tunnel, is used
to prevent diesel particulate from contaminating the sampling system. The
filter and line connecting the filter to the dilution tunnel are heated to
375°F in order to prevent sulfur dioxide from being retained in the filter and
sample line. The Teflon line connecting the heated filter and the solenoid
valve is heated to 175 °F in order to prevent water from condensing in the
sample line. Several views of the sampling system are shown in Figure 2.
ANALYTICAL PROCEDURE
Sulfur dioxide in dilute exhaust is collected in two impingers connected
in series with each impinger containing 25 mi of 3 percent hydrogen peroxide.
The temperature of the impinger is maintained at 0-5°C by an ice water bath.
The flowrate through the impinger is adjusted to 4 2,/minute with a regulating
valve and the sample pump. After sampling is completed, the absorbing solu-
tion in each bubbler is transferred to a 30 m£ polypropylene bottle and capped.
The samples should be analyzed within four or five weeks after collection.
Approximately 2 m£ of the sample is loaded into the ion chromatograph sample
loop and injected. The injection inserts the sample loop volume (0.5 mfc) of
sample into the instrument. An analysis flow schematic and pictures of the
ion chromatograph are shown in Figures 3 and 4. The ion chromatograph
299
-------
CO
o
o
Sample
probe
Sample
Pump
Dilute
Exhaust
Gas Temperature
Digital Readout
Flowmeter
*
Regulating
Valve
Ice Bath
Temperature Readout
Dry
Gas
Meter
Gas Volume
Digital Readout
Figure 1. SO sample collection flow schematic.
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. SO2 sampling system.
301
-------
Solenoid
Impinger
Ice Bath
Close-up of Impingers (Side View)
Solenoid
Filter
Ice Bath
Pump
Rear View
Figure 2 (Cont'd). S02 sampling system.
302
-------
CVS
\
Glass
Impinger
Unused
sample
saved as
needed
Sample analysis in
Ion Chromatograph
with conductivity
cell
A/D Converter
I
Recorder
Hewlett-Packard
3354
Computer System
Figure 3. SO- analysis flow schematic.
303
-------
Suppressor 1
Column
Conductivity
Detector
Conductivity
Detector
Figure 4. SO^ ion chromatograph,
304
-------
utilizes two columns, the separator and the suppressor. The 3 x 500 mm
analytical column and a 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
The 6 x 250 mm glass suppressor column packed with AG 50W-X10, a strong acid'
cation exchanger, neutralizes the ionic effect of the eluent while increasing
that of the sample ion. The column packing is in the hydrogen form so that
in the presence of the eluent (NaHCO3 and Na2CO3) , H2CO3 is generated and when
sulfate is introduced, ^864 is formed.
NaHCO + Resin-H ^=^ Resin-Na + H CO
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 analy-
sis of a standard and a sample. The ion chromatograph operates at room tem-
perature at a maximum pressure of 500 psi.
CALCULATIONS
This procedure has been developed to provide the user with the concen-
tration of sulfur dioxide in exhaust. The results will be expressed in
Ug S02/m3 of exhaust and ppm. A stepwise derivation of the equations used in
calculating the concentrations is provided as well as a copy of a Hewlett-
Packard 67 calculator program (Figure 7) . This program is designed to reduce
the amount of time required to do the calculations manually. For illustra-
tion, two examples using information from the data sheet (Figure 8) are
included at the end of this section.
The first step in the calculations is to correct the volume of exhaust
sampled to a standard temperature, 68°F, and pressure, 29.92"Hg, by use of
the equation:
PxVx = Pspec x V
Tx Tspec
Solving for V gives:
PxVx x Tspec
Tx+460 Pspec
where V = volume of gas sampled at specified temperature
and pressure (ft )
Px = experimental pressure ("Hg)
Vx = experimental gas volume collected (ft )
Tx = experimental temperature (°F)
305
-------
12 10 8 6420
Retention time, minutes
Figure 5. Sample chromatogram.
306
-------
Standard Date 11-30-79
Attenuation 30x1
Eluent 0.003 M NaHCO3 + OeQQ24 M
Flowrate 167 nd/hr Chartspeed 12 in/h
Loop Size 0.5 m£ Packing AG 50W-X16
Analytical Column Pre column 3x150 mnTglass,
Analytical column 3x500 mm glass" '
Packing Strong base anion exchanger in the
bicarbonate form
Supressor Column 6x250 mm glass
15 10 5
Retention time, minutes
Figure 6. Standard Chromatogram
307
-------
I Sut Absorbing reagent volume
Input Standard Concentration
Input Standard Area
Input Sample Area
Output Safliple Concentration
Output Sample Concentration
.
-- -
INPUT
DA f A UNITS
ft3
"H.J
°F
mi
MgSO42-/i«
KtVS
1
]
1 y
; A
i R/sl
i H/S|
1 R/sl
i R/si
I R/S|
1 R/sl
1 R/sl
1 R/sj
1 1
1 a 1
1
1
|Scl
1
i
1
i
1
!
IKON
1
1
1 1
1 i
,
i
i
i
1
1
I
OUTPUT
DATA, UNITS
]iy SO-;/Ul
Pi 'in
Figure 7. HP-67 user instructions.
308
-------
STEP KEYENTRY KEY CODf COMMH,,* iltP Kty EMHY KEY CO,*
- '
-
:'.IU
it JO
OJO
U40
' V
-J'M
tI£L.A. .
3
r_i ."."
1~JL ~.~".
2
B
X
2
9
....2
•>
R/S
_Sm_iIl.
6
0
-f
KCL 1
h X ?y _
h 1/X
R/S
X
R/S
X
R/S
X
R/S
R/S
X
0
•
6
6
7
X
R/S
0
•
0
0
0
3
6
^
x
_ R/S
il ii ii..
.. .ui. .....
_. U5__
sa
.. !il. .
01
«i
05
o;>
i)iL
.._.?!
OJ
_fy. _
ca
ai
u-4
H4
ud _.
Ou
00
61
34 01
35 52
8) ~^
35 62
84
71
84
71
84
71
Si 4
81
84
71
00
81
06
06
07
84
00
B3
00
00
00
03
On
07
71
H4
In^ul ?>.iiu[.
>; v,-,l,
ft3
Input, barometer, "liq
Input. Saitii>lu Temp,
0F
Input Dilution Factc
Input Abs Suln Vol ,
Input Std
P9 SU42'
Input Std
Input Samp
Output Con
.tl ^'^II
o i a J
30 Sl S2 S3
A 1
3 C
mil
Cono
'/M
Area
le Area
c,Vig ,S02/
mj
.1
^•4
ot.o
l) 'll
UtlO
c
uao
100
_
U KB!
^ .
JS j.:
-- - --
5 6 /
S5 S6 SI
D
~ a •'
t 1
Figure 7 (Cont'd). HP-67 program form.
309
-------
SwRI Project No.
Fuel:
Test No.
Test Date:
Vehicle:
CVS No.
Tunnel Size
Driver:
Sample Collection By:
General Comments:
Chemical Analysis By:
Miles:
Calculations By:
Driving
Cycle
FTP-1
FTP-2
FTP-3
SET- 7
HFET
NYCC
BG
Sampling Conditions
Volume
Ft3
1.189
2.040
1.250
3.240
1.690
1.410
2.000
1
B.P.
"Hg
29.47
29.48
29.49
29.50
29.50
29.51
29.45
2
Temp.
OF
74.7
75.0
74.5
75.0
74.7
74.9
74.9
3
Dilution
Factor
1
1
1
1
1
1
1
4
Absorb.
Reagent
Volume m&
25
25
25
25
25
25
25
5
Standard
UgS04-2/m£
:10
30
10
40
20
15
1.0
6
Area
3200
4500
3200
4200
5700
4900
3000
7
Sample
Area
4000
3500
4300
4500
5500
5000
2700
8
ugSO2/m3
6360
6930
6500
8000
6670
6630
273
9
ppm
2.34
2.54
2.39
2.94
2.45
2.43
.0.10
10
CO
I—»
o
Figure 8. SO2 data sheet.
-------
Tspec = specified temperature = 68°F = 528 °R
Pspec = specified pressure = 29.92 " Hg
Next, the quantity of SO2 in this volume of gas is calculated. The ion
chromatograph measures the amount of sulfate (from S02) in yg S°4~2 Assuming
linearity between concentration of sulfate and peak area, a standard of known
sulfate concentration is compared to a bubbled exhaust sample.
Cst = Csamp
AREAst AREAsamp
Cst x AREAsamp
Csamp = - _____ -
_2
where Csamp = concentration of SO."2 in sample ^9 4
ml
_2
Cst = concentration of SO."2 in standard
mil
AREAsamp = area of sample (relative units)
AREAst = area of standard (relative units)
Converting to yg SO /m£:
64.06 g/mole SO
96.06 g/mole SO,"
If the sample has been diluted the dilution factor, DF, needs to be
included.
0.667 Csamp x DF
This represents the amount of SO2 in one m£ of absorbing solution. The
volume of absorbing solution is multiplied by this last quantity to give the
amount of SO collected.
0.667 Xsamp x DF x absorb, vol. (m£ ) = yg
The concentration of SO2 in Ug 3°? in the exhaust sample is obtained by
dividing yg SO by gas volume (corrected from ft3 to m3) .
311
-------
Concentration of SO ,
ug so2
yg SO x AREAsamp
yg SO ' - £. — x DF x absorb . vol . , m£
0.667 - X - -
yg SO AREAst
Px, "Hg x Vx,_ ft3 x 528°R
Tx, °F + 460 29.92" Hg
-2
yg so yg SO
0.667 ^j x cst, —JT x AREAsamp x DF x absorb, vol.,
yg so4
3
Px, "Hg x Vx, ft
3
(Tx, °F + 460) x 29.92" Hg x 35.31 ——
ft3
528°R x AREAst
To find the S02 concentration in ppm the density of the gas at the specified
conditions is needed. The density of S02 at 32°F and 29.92" Hg is 2.927
g/lit. This can be corrected to 68°F C492°R) and 29.92" Hg using Charles"
version of the ideal gas law:
V V
T
where V± is the volume at 32 °c (£) = 1.0
T-L is 32 °F = 492 °R
V is the volume at 6S°F (£)
T is 68°F - 528°R
v = (1.0 i) (528°R)
492°R
Density at 68°F and 29.92" Hg = 2'927 3
x • U /
-------
Sample Calculation
The two examples will be calculated from information recorded on the
data sheet (Figure 8) . This information does not necessarily represent
actual experimental data but serves as a means of confirming calculations
done by hand with those done with the Hewlett-Packard Calculator.
Example 1
Assume 1.189 ft of exhaust was collected in 25 nd of 3 percent H202
in the FTP-1 driving cycle at a barometric pressure of 29.47" Hg (cor-
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
in yg S04~2
of 4000 counts and the corresponding standard, - j — * — , an area of
3200 count. m£
yg so2 yg so "2
0.667 - - x Cst, - 5 - x AREAsamp
-
3
yg SO./m =
2 Px, "Hg x Vx, ft3
DF x absorb, vol., m& x (TX, °F + 460) x 29.92" Hg x 35.31 m
ft3
528°R x AREAst
0.000367 x Q.667 x 1Q x 4QQQ x i x 25 x (74.7 + 460) x 29.92 x 35.31
29.47 x 1.189 x 528 x 3200
= 6360 yg SO2/m
*3 m Q
ppm SO0 = 6360 yg SO_/m x 0.000367
•'rt *-" — w j-i -y '-"~-ijf —* — — 11/T
=2.34 ppm SO
Example 2
Assume that in the SET-7 driving cycle dilute automotive exhaust was
collected in 25 m£ of 3 percent H202 according to the procedure described
previously. The sampling conditions under which the 3.240 ft of exhaust
was collected were 75.0°F and 29.50 " Hg. An area of 4200 counts was pro-
n
duced by the 40 yg S(?4~ standard and the exhaust sample yielded an area of
313
-------
4500 counts. Inserting these values into the same equation used in Example
1 gives concentrations of 8000 yg S02/m3 and 2.94 ppm S02.
LIST OF EQUIPMENT
The equipment required for the SC>2 determination is divided into four
categories: Sampling, Analysis, Water Filtration 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. 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 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"OD 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.
12. six channel digital thermometer, Analog Devices, Model #2036/J/1.
13. 30 m£ polypropylene sample storage bottles, Nalgeie Labware,
Catalog #2006-0001.
14. Iron/Constantan type J single thermocouple with 1/4" OD stainless
steel metal sheath, Thermo Sensors Corporation.
15. Stainless steel heated filter assembly-7 cm; Scott, capable of tem-
perature to 204°C, included 2 heaters, adjustable thermostat switch,
314
-------
stainless steel insulated covers and sample bypass solenoid valves.
16. Glass raicrofiber filter discs, Reeve Angel 934-AH, Whatman, 7 cm
diameter.
17. Flexible heavy insulation heating tape, Briskeat®, width-1/2 inch
length-48 inches. '
18. Temperature Controller, Athena, 100-600°F.
19. Heated TFE Teflon hose, Technical Heaters, Inc., 5' x 1/4" tem-
perature limit 400°P.
Analysis
1. Conductivity cell, modified swagelok reducing union, Catalog
#SS-200-6-l, approximate volume, 4.5 yl.
2. Conductivity detector, Hall, Tracor 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 Partner Instrument Company, Catalog
#6100-20, 1 gallon.
Water Filtration
1. Filtration apparatus, Millipore, Catalog #XX 15 047 00.
2. Filters, Millipore, Catalog ttGSWP 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 m£ volumetric pipets.
4. Class A, 3 m£ volumetric pipets.
5. Class A, 4 m£ volumetric pipets.
6. Class A, 5 m£ volumetric pipets.
7. Class A, 10 n>Jl volumetric pipets.
315
-------
8. Class A, 20 mH volumetric pipets.
9. Class A, 25 rtd 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 mH volumetric flasks.
15. Mohr pipet, 1 mi graduated 1/10.
LIST OF REAGENTS
A list of reagents used in determination of SO2 is provided indicating
purity, manufacturer and catalog number. The function of each reagent in
the procedure is also given.
1. Water - deionized and filtered through 0.22 micron filter.
2. Primary standard - Sulfuric acid, H2SO4, certified 0. IN, formula
weight = 98.08, ACS reagent grade, Fisher Scientific Company
#SO-A-212.
3. Absorbant - Stabilized 30 percent hydrogen peroxide,
weight = 34.01, analytical reagent grade, Mallinckrodt #5239.
4. Eluent - Sodium bicarbonate, NaHCO-^, formula weight = 84.01, ACS
analytical reagent grade powder, Mallinckrodt #7412.
5. Eluent - Sodium carbonate, Na2CO3, formula weight = 105.99, ACS
analytical reagent grade anyhdrous power, Mallinckrodt #7521.
6. Regenerant - Sulfuric acid, 112804, 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 prepared by diluting 20 mJl of certified 0.1 N
f\
H2S04 (4800 yg 4 to 1000 m£ with water. The resulting solution contains
m£
316
-------
-2
96 -° yg 4 • More dilute standards are prepared by pipetting 0.5, 1, 1.5,
m£
2, 3, 4, 8, 10, 20, 30, 40, 50, 60, 70, 80, and 90 m£ of the stock solution
into 100 m£ volumetric flasks and making up the volume. -These standards re-
main stable for at least fourteen weeks. ^11 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 Solition (3 percent
100 m& of 30 percent H.2Q2 i-s diluted to 1 liter with water.
Eluent (0.003 M NaHCO^ + 0.0024 M
Concentrated stock solutions of sodium bicarbonate and sodium carbonate
are prepared in the following manner. For a 0.6 M solution of sodium bicar-
bonate, 50.41 g of solid sodium bicarbonate is dissolved in 1000 mi of water.
For a 0.48 M solution of sodium carbonate, 50.88 g of solid sodium carbonate
is dissolved in 1000 m£ of water. The carbonate solution used as the eluent
is made up by pipetting 10 m£ of each stock solution into a 2 liter volumetric
flask and diluting to mark with water.
Regenerate (1 N H2SO4) *
56 mJl 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 .
317
-------
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.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 (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. amd Rabbins, 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.
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.
Bostrom, C.E. Int. J. Air Water Poll., Vol. 9, pg. 333, 1965.
Attari, A., Igielski, R.P., and Jaselskis, B., Anal. Chem, Vol. 42,
pg. 1282, 1970.
318
-------
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. , Environm. Sci. Tech., Vol. 1, pg. 68, 1967.
Shikiya, J.M. and McPhee, R.D., 61st Annual Meeting, Paper No. 68-72, Air
Poll. Conr. 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, N.Y., 1949.
Katz, M., Anal. Chem., Vol. 22, pg. 1040, 1950.
Stratman, H. , Mikrochim, Acta. Vol. 6, pg. 688, 1954.
Pate, J.B., Lodge, Jr., J.P., and Neary, M.P., Anal. Chem. Acta., Vol. 28,
pg. 341, 1963.
Lodge, Jr., J.P., Pate, J.B., and Huitt, H.A., Amer. 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.
HueyN.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.
Hickey, H.R. and Hendrickson, E.R., J. Air Poll. Cont. Ass., Vol. 15,
Pg. 409, 1965.
Harding, C.I., and Kelley, T.R., J. Air Poll. Cont. Ass., Vol. 17, pg. 545,
1967.
319
-------
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, og. 652, 1969.
Ronkainen, P., Denslow, J., Leppanen, O., J. Chroraatog. Sci,, Vol. 11,
pg. 384, 1973.
Pescar, R.E., and Hartman, C.H., J. Chromatog. 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.
320
-------
APPENDIX F
NITROUS OXIDE PROCEDURES
321
-------
THE MEASUREMENT OF NITROUS OXIDE IN EXHAUST
This procedure was developed to measure nitrous oxide (N20) in dilute
gasoline and diesel exhaust. Standard CVS bag samples are analyzed for N20
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 N20 in exhaust is shown in Figure 1.
ANALYTICAL SYSTEM
The analysis for N20 in exhaust is conducted with a gas chromatograph
system using a Perkin-Elmer Model 3920B electron capture detector. The
system employs two pneumatically operated electrically controlled Seiscor
valves, an analytical column, and a stripper column. The gas chromatograph
separation is obtained at room temperature. A special control console was
fabricated to house the entire system except for the electron capture de-
tector.
A stripper column is included as a precautionary measure to prevent
unwanted heavier molecular weight exhaust species from entering the ana-
lytical system. Figure 2 (Step 1) illustrates the gas chromatograph flow
schematic with ttie 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 N20 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
Position Function Position Function
off purge CSV w/sample off foreflush to
analytical column
2 on sample injected off foreflush to
a-ialytical column
3 on sample injected on backflush to vent
Under normal conditions, it is not necessary to backflush the calibra-
tion standards since they are free of contaminants that would interfere with
322
-------
Vehicle
CVS
Tedlar bags
Gas
Chromatograph
HP 3354 GC
Computer System
Teletype
Printout
Figure 1. Total system flow schematic for the analysis
of nitrous oxide in exhaust.
323
-------
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, 80/100 Porapak Q)
Auxiliary
Carrier Gas
Capillary
Restricter
»«•»•••••»••*•*••»••»••••«••*••••••%
o o
Carrier
Gas
Seiscor Valve
(Gas Sampling Configuration)
Flowmeter
Female
Quick-Connect
{ | Sample or
Calibration
gas in
Figure 2. Flow schematic of nitrous oxide analytical system
(Step 1 - Purge of sample loop of GSV).
324
-------
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
I
Analytical Column
••» (61 x 1/8" SS, 80/100 Porapak Q)
Capillary
Restrictor
10 ml
sample
loop
Auxiliary
Carrier Gas
Carrier
Gas
mmmmmmmm
1—T
mmmmmmmmmmmmmmmmmmmmmmmmmt
Seiscor Valve
(Gas Sampling Configuration)
Vent
Regulating
Valve
Pump
f"l Sample or
. Calibration
Female gas in
Qui ck-Connec t
Flowmeter
Figure 3. Flow schematic of nitrous oxide analytical system
(Step 2 - Inject sample or calibration gas into system).
325
-------
Stripper Column
(21 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, 80/100 Porapak Q)
Auxiliary
Carrier Gas
Capillary
Restrictor
10 ml
sample
loop
d-
-Q
Carrier
Gas
• ••
Seiscor Valve
(Gas Sampling Configuration)
Vent
Regulating
Valve
Pump
Flowmeter
Female
Quick-Connec t
—Sample or
Calibration
gas in
Figure 4. Flow schematic of nitrous oxide analysis system
(Step 3 - Backflush OV-17 stripper column).
326
-------
the analytical column. A typical gas chromatograph trace for a calibration
blend is shown in Figure 5. A baseline separation is obtained and the N,0
peak area is obtained using a Hewlett-Packard 3354 GC computer system.
On gasoline and diesel samples it is necessary that the backflush is
inlcuded 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 along with the gas chromatograph'
operating conditions is presented in Figure 6.
CONTROL SYSTEM
The control of the two Seiscor valves if 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 7. The flow schematic for vacuum and pressure lines to the Seiscor
valve are presented in Figures 8-10.
SAMPLE CALCULATIONS
The quantification of ^0 in exhaust is based on a direct comparison
of the N20 in exhaust with a calibration blend of a known ^0 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 detec-
tor, the following relationship is true.
Let Csam = ppm concentration of ^0 in sample
Cstd = ppm concentration of ^0 in standard
Asam = area of ^0 peak in sample
Astd = area of N20 peak in standard
Astd _ Asam
Cstd ~ Csam
Solving for Csam
Csam = Asam x Cstd
Astd
327
-------
Operator B. Fanick
O.D. 0.063" I.D. S.S
Instrument PE
Column
Liq. Phase
Support
Carrier
Packed with
on 80/100
Porapak
Rotameter Reading
/min. Held for
min.
min.. Prog to C at
held for
Inlet Room
Rotameter Rdg.
Rotameter Rdg.
Rotameter Rdg.
Recorder 0.5 in/min speed
Injection 10 ul indicated _
Sampling Device Gas Sampling Valve
mV.F.S.L+N-H
ul net
L j. il
12 11 10 9 8 7 6 ~ 5 4 3
Retention time, minutes
Figure 5. Typical N2O calibration blend gas chromatograph
0
trace.
328
-------
Sample Gasoline Exhaust-CVS Date
Instrument PE 3920B Operator 5
11-10
-77
. FanicK
Column 6 ft. 1/8" O.D.O . 06 3 I.D. R . s .
Packed with - %wt. NA
on«U/lUU mesh PORAPAK Q
Run ISO @ 20 °C usina cc/min.
@ psig NA Rotameter Reading
Cii4/Ar
held @ °C ISO for min.. proq to
@ °/min. Held for min.. Proq to
held for min. (other)
Inlet Room <>c
Detector 325 °C BCD Type (other)
Hyd NA psig NA Rotameter Rdo.
Air NA psjg NA Rotameter Rdq.
( ) NA psig NA Rotameter Rdg.
Recorder 1 in/min speed 1 mV.F.S.
Injection 10 ul indicated 10 ul net
°Cat
^one
NA
HA
14A
Type
Liq. Phase
Support
Carrier
°C
°/min
Type
cc/min
cc/min
cc/min
Soltac Type
10
ul Actual
Sampling Device Gas Sampling Valve
12345678
Retention time, minutes
Figure 6. Typical gasoline-CVS exhaust sample.
329
-------
AC (-)
CO
co
o
T'l
2 SJ
3 !
4 i
12
14
15
16 j
ATC
Timer
«NC
%
— -
I
•-••"1
1 1
il
4
;
Q
— <
, ,„•/"
L
^
J
\
^ 1
1
GSV
solenoid
g S
S :
I
|re
11
12
14
15
I 16
ATC
Timer
backflush
solenoid
Figure 7. Electrical schematic for nitrous oxide analysis system.
-------
Vacuum
cap
n_.
E2
Air Pressure (30 psi)
P V
CSV
Seiscor
CSV off
Vacuum
Air Pressure (30 psi) T
cap
? V
Backflush
Seiscor
Backflush off
Figure 8. Flow schematic in electric solenoid valves
(Both valves de-energized).
331
-------
Vacuum
cap
Air Pressure (30 psi)
cap
P V
CSV
Seiscor
CSV on
Vacuum
Air Pressure (30 psi)
Backflush off
Figure 9. Flow schematic in electric solenoid valves
(CSV energized, backflush de-energized).
332
-------
Vacuum
Air Pressure (30 psi)
CSV on
Vacuum
cap
Air Pressure (30 psi) t
cap
P V
Backflush
Seiscor
Backflush on
Figure 10. Flow schematic in electric solenoid valves
(Both valves energized).
333
-------
Example 1:
A 4.95 ppm N20 (in nitrogen) calibration blend was found to give 5291
area counts for the N20 Peak- ^ exhaust sample was found to give 2674
area counts for the N20 peak. Calculate the N20 in the exhaust sample.
Asam x Cstd
Csam = Astd
2674 x 4.95
Csam =
5291
Csam =2.50 ppm N20
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 N20 in the exhaust
sample.
Asam x Cstd
534 x 1.13
Csam =
1208
Csam =0.50 ppm N20
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 house the mechanical
hardware items that are necessary for the proper operation of the N20 analy-
sis syptem. Figure 11 illustrates the complete analytical system for mea-
suring N20 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. Linerarized electron capture detector (BCD)
3. Leeds and Northrup Model W 1 mv recorder
4. Hewlett-Packard Model 3354 GC computer system
334
-------
A/D Converter
.,,
Recorder
__Control Console
Calibration
Standards
Figure 11. Nitrous oxide analytical 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 325A364A10PX (2 ea)
4. Analytical column, 6' x 1/8" SS, 120/150 Porapak Q
5. Stripper column, 21 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
336
-------
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.
De Grazio, 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 of EPA and
Harry Dietzmann of SwRI.
LaHue, M.D., Axelrod, H.D., and Lodge Jr., J.P., Anal. Chem., Vol. 43,
pg. 1113, 1974.
337
-------
APPENDIX G
HYDROGEN SULPIDE PROCEDURE
338
-------
THE MEASUREMENT OF HYDROGEN SULFIDE IN EXHAUST
The measurement of hydrogen sulfide in dilute automotive exhaust is
accomplished by bubbling dilute exhaust through glass impingers containing
a buffered zinc acetate absorbing solution. The hydrogen sulfide reacts
with the zinc acetate to form zinc sulfide which remains in solution. The
exhaust sample is collected continuously during a test cycle. Upon com-
pletion of the test, the absorbing solution is treated with N,N dimethyl-
para - phenylene diamine sulfate and ferric ammonium sulfate. This reaction
produces a highly colored heterocyclic compound, methylene blue (3,9 - bis-
dimethylaminophenazothionium sulfate). This colored solution is analyzed
with a spectrophotometer at 667 nm in a 1 cm or 4 cm pathlength cell. The
results are quantified by comparison to a standard curve. The minimum de-
tectable concentration is 0.01 ppm.
SAMPLING SYSTEM
Two glass impingers in series, each containing 50 mH (10 m£ buffered
zinc acetate solution and 40 m£ freshly vacuum boiled deionized water) of
buffered zinc acetate absorbing solution, are used to collect exhaust
samples for the analysis of hydrogen sulfide. A flow schematic of the
sample collection system is shown in Figure 1. The two impingers together
trap approximately 99+ percent of the hydrogen sulfide. 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 £/minute by the sample pump.
A dry gas meter is used to determine the total flow through the impinger
during a given sampling period. The temperature of the gas stream is
monitored by a thermocouple immediately prior to the dry gas meter. A drier
is included in the system to prevent condensation in the pump, flowmeter,
dry gas meter, etc. The flowmeter in the system allows continuous moni-
toring of the sample flow to insure proper flow rates during sampling.
When sampling diesel fueled vehicles, a filter, located between the on-off
solenoid valve and the dilution tunnel, is used to prevent diesel parti-
culate from contaminating the sampling system. The lines connecting the
filter to the dilution tunnel and the filter to the solenoid valve are
heated to 175 °F. Several views of the sampling system are shown in Figure 2.
DESCRIPTION OF METHOD
Hydrogen sulfide 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 methy-
lene blue is shown in Figure 3.
339
-------
co
CD
Sample
Probe
Gas Temperature
Digital Readout
Sample
Pump
IQIBIOI
F lowme t e r
Dry
Gas
Meter
*
Regulating
Valve
Dilute
Exhaust
Ice Bath
Temperature
Digital Readout
Gas Volume
Digital Readout
Figure 1. Hydrogen sulfide sample collection flow schematic.
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. The dilute exhaust sampling system for hydrogen sulfide.
341
-------
Close-up of Impingers (Side View)
Pump
Rear View
Figure 2 (Cont'd). The dilute exhaust sampling system for hydrogen sulfide.
342
-------
2.0
0.6
0.4
0.2
350
400
450 500 550 600
Wavelengths, nanometers
650
700
750
Figure 3. Extinction-wavelength curve of methylene blue.
343
-------
The analytical procedure for the determination of hydrogen sulfide in
dilute automotive exhaust consists of two major areas. The first is the
standarization 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 concentration of the standard sodium sulfide solution (approximately
0.03 M sodium sulfide in deionized water) is determined by an iodometric me-
thod. To three (3) Erlenmeyer flasks, 10 m£ of the absorbing solution and
50 m£ of the sulfide solution are added and the resulting solution is mixed.
Into one of the flasks 5 m£ of the 0.01 N iodine solution and 10 m£ of con-
centrated hydrochloric acid are added. The resulting solution is immediately
titrated to the starch endpoint with a standarized thiosulfate solution.
This procedure is then repeated for the remaining two flasks and for two
blanks prepared with only the absorbing reagent and 50 mi 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.
The thiosulfate solution used to titrate the sulfide solution is stan-
dardized against primary standard grade potassium dichromate. The potassium
dichromate is dreid 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 dissolved in 50 m£ of deionized water. A
freshly prepared solution of 3 g (0.02 mole) of potassium iodine, 5 mil of
6 N hydrochloric acid and 50 m& 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 flask are then washed with deionized
water followed by approximately 200 m£ of deionized water. The resulting
solution is titrated with the thiosulfate solution. As the end point is
approached, about 5 m£ of 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 standarized thiosulfate solution can then be used to
standarize the dilute iodine solution as well as determine the sulfide ion
concentration in the sulfide standard solution.
A concentrated iodine solution (~0.1 N) from which the dilute iodine so-
lution is prepared is standarized against primary standard grade arsenic tri-
oxide by an iodometric titration. A weighed portion of 0,15 to 0.20g (0.001
mole) arsenic trioxide is placed into an Erlenmeyer flask. Then 10 to 20 m£
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 re-
sulting solution is diluted to about 100 m£. About 2 m£ 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 standarization needs to be done
344
-------
CVS
I
Glass
Impinger
Permeation
Tube
Reagent
Addition
KCr2O7
Primary
Standard
AS?03
Primary
Standard
I
I
Color
Development
Thiosulfate
Standardization
Absorbance
Reading
Iodine
Standardization
Calibration
Curve of
Sulfide Standard
Figure 4. Hydrogen sulfide analysis flow schematic.
345
-------
only once for the concentrated solution.
The dilute iodine solution (-0.01 N) prepared by diluting the concentra-
ted iodine solution, is standarized against the previously standarized thio-
sulfate solution. To a beaker containing 25 m£ of the dilute iodine solution,
10 m£ of concentrated hydrochloric acid is added and the resulting solution
is immediately titrated with the standarized thiosulfate solution. Just
before the endpoint, starch is added to serve as an indicator. The dilute
iodine solution should be standarized daily.
A Beer's Law curve is determined by adding 0.1, 0.25, 0.5, 1.0, 2.0,
3.0, and 5.0 mJl of the sulfide standard solution to seven separate 100 mjl
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
not :to introduce air bubbles. Then, 2 mH of the ferric ion solution are
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 mm 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,
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 m& of vac-
uum 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 5,/min. Upon completion of each driving cycle, the impingers are re-
placed with fresh ones. To each of the collected samples, 10 m£ of the amine
solution is added through the top of the impinger and gently swirled. Then
6 m£ (2 m& for samples from gasoline powered vehicles is sufficient) 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 mJl with vacuum boiled, deionized water.
346
-------
Ifce color development is complete in 30 minutes. The proceduer 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 quan-
tity of hydrogen sulfide in dilute exhaust. A Hewlett-Packard 67 program was
developed to reduce the time required for manual calculations. The deri-
vation of the equations are given below and a copy of the steps in the pro-
gram are shown in Figure 7.
Derivation of Equation
The first step is to correct the volume of exhaust sampled to a stan-
dard temperature, 68°F and pressure, 29.92"Hg, by use of the equation
P XV P x V
exp exp _ corr corr
T T
exp corr
V = experimental volume of gas sampled in ft
vexp = volume of gas sampled in ft3 corrected to 68°F and 29.92"Hg
P = experimental barometric pressure
pexp = 29.92"Hg
TCOrr = experimental temperature in °F + 460
T6XP = 68°F + 460 = 528°R
corr
Solving for V gives:
corr
v
P ("Hg) X V (ft3) x 528°R
= exp *' exp _ _
corr T (°R) x 29.92"Hg
exp
The next step converts the volume from cubic feet to cubic meters by
use of the conversion f actor; 1 cubic meter is equal to 35.31 cubic feet.
P ("Hg) x V (ft3) x 528°R
„ , 3. exp *' exp __
corr ' ~ T x 29.92"Hg x 35.31 ft3/m3
exp
(Equation 1)
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.
-------
ls
-------
STEP KEY ENTRY KEY CODE COMMENTS STEP KEY ENTRY KEY CODE Cnuut»,<
CjlO
».M
JJO
04U
ObJ
2
-:
R/S
X
STP_]
4
A
,1
+
RCL 1
X
STO 2
BfT. •>
R/R
X
1
•
0
6
3
X
STO .J
RCL 2
X
.
o
h
,
X
R/S
RCL 3
-t-
R/S
1
4
3
4
^.
R/S
h kTO
02
81
84
J101_ ..
04
(16
O'l
fcj.
34 01
81
71
13 u2
JiUZ
„ H4
71
01
63
Oil
06
0)
71
1.) 01
.....
34 02 s
01
ULI
ilu
111
71
84
34 OJ
61
84
01
04
03
04
81
84
.35 22
ill nmni/ie VOJ, It-
Ill baiometric, "Hg
Ii. bam. le TBmp, "F
Tn Total Vol of Sol
In Sample Cone.
Bubbler itl.
-Out S
-------
The concentration of sulfide ion in ygS /m£ is obtained from the Beer's
Law Plot of the absorbance. The concentration of hydrogen sulfide in
ygH2S/m& is obtained by multiplying the concentration in ygS /m& by the
ratio of the formula weight of hydrogen sulfide to the formula weight of
sulfide ion.
/ o ~2- , n formula weight HoS(yg/y mole)
ygH9S/m£ = ygS /m£ x - - - - . . ^ A , . - r— r
Hy * ' ^ formula weight S^~(yg/y mole)
34.080 yg H2S/y mole
X 32.064 ygS2-/y mole
= ygS2~/m£ x 1.063
yg &
To obtain the total amount in yg of hydrogen sulfide, the concentration
in ygH2S/m£ is multiplied by the total volume of solution (TVS) . This is
the volume to which the absorbing solution, amine solution, and ferric ion
solution have been diluted with deionized water. In this case the volume
is 100 m£.
ygH2S = ygH2S/m£ x TVS
= ygH2S/m£ x 100 m£
= ygS2~/m£ x 1.063 Uq"§f x 100
(Equation 2)
To obtain ygH2S/m3, Equation 2 is divided by Equation 1 to give:
ygH2s/m3 =
WS x 1.063 yg*^! x 100 m£ x T x 29.92"Hg x 35.31 ft3/m3
(Equation 3)
P ("Hg) x V (ft3) x 528°
exp 3 exp
To find the concentration of hydrogen sulfide in ppm, the density of
hydrogen sulfide at 68° and 29.92"Hg is needed. The fifth edition of the
Matheson Gas Data Book lists the density of hydrogen sulfide at 0°C and 1
atmosphere at 1.5392 g/£. if the volume of l£ of gas is corrected for
temperature,
r ODOrj
(32'F + 460
it can be divided into the weight of the gas to give the density at 68°F
1.5392 g
1>073 £ = 1-434 g/£ = 1434
352
-------
To obtain the concentration of hydrogen sulfide in ppm, the concentra-
tion in yg/mj is divided by the density in yg/m£
ppra = yg H2S/m T yg/m£ =
m
(Equation 4)
At this point, the concentration can be expressed in ygH2S/m3 (Equation
3) and ppm (Equation 4) at 68°F and 29.92"Hg from the raw data.
Sample Calculations
Assume that a set of six samples was taken from dilute exhaust during
several different driving cycles. The volume of dilute exhaust sampled
during the first test was 3.185 ft3 at a barometric pressure of 29.41"Hg and
a temperature of 77°F. The analysis was conducted and Bubbler #1 was found
to contain 9.2 ygS2~/m£ while Bubbler #2 contained 0.8 ygS2~/m£
For Bubbler #1
, ygS2~/m£ x 1.063 ^"2- x 100 m£ x T x 29.92"Hg x 35.31 ft3/m3
/ 3 ygs exp
VlgH2S/m = - « - * - - -
P ("Hg) xV (ft ) x 528°R
exp exp
f\
9-2 ygS /m& x 1.063
29.41" Hg x 3.185 ft3
100 m& x (460 + 77°F) x 29.92"Hg x 35.31 ft3/m3
X 528°R
= 11218 yg/m3
The concentration for Bubbler #2 is calculated in the same manner using
n _
the appropriate concentration in ygS /m£
For Bubbler #2
,3 0.8 ygS2~/m£ x 1.063 ygHpS/ygS2"
UgH2S/m = 29.41" Hg x 3.185 ft?*
3 3
100 ml, x (460 + 77°F) x 29.92"Hg x 35.31 ft /m
X 528°R
=975 yg/m3
Total yg H2S/m3 = Cone (Bubbler #1) + Cone (Bubbler #2)
= 11218 + 975
= 12193 yg/m3
353
-------
ppm H-S = ygH S/m T density yg/m£
= 12193 ygH2S/m3 4- 1434 Ug/m3
= 8.50 ppm
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-67 program. The program will be used to calculate both ygH2S/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.
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
2. Glass impingers, Ace Glass Products, Catalog No. 7530-11, 29/42
bottler 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 £/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 DV2-144NCS.1.
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, wriings, etc.
354
-------
TEST DATE: 11-10-79 VEHICLE: Practice
SWRI PROJECT NO. 11-1234 TEST NO. 001
FUEL: EM~237 CVS NO. 3 TUNNEL SIZE: 18" DRIVER: D.A.T. MILES:
1000
SAMPLE COLLECTION BY: G-°-
GENERAL COMMENTS:
CHEMICAL ANALYSIS BY: W.M.S-CALCULATIONS BY:
CO
on
cn
Test No.
Driving Cycle
Volume, Ft3
B.P., "Hg
Temp. °F
Total Vol. of Solution, mi
o
Sample Cone-Bub. #1, ygS /mfi,
Sample Cone-Bub . #1 , ygH2S/m3
Sample Cone-Bub. #2, ygS2~/m£
Sample Cone-Bub. #3, pgH2S/m3
Total Sample Cone, ygH2S/m3
Total Cone, ppm H2S
1
Cold FTP
3.185
29.41
77
100
9.2
11200
0.8
975
12200
8.50
2
Hot FTP
3.486
28.66
75
100
7.3
8310
6.2
228
8540
5.95
3
SET-7
3.508
29.33
80
100
10.4
11600
1.1
1230
12800
8.95
4
HFET
1.926
29.40
85
100
3.1
6340
0.3
614
6960
4.85
5
NYCC
1.525
29.10
90
100
1.0
2630
0.1
263
2900
2.02
6
Backgrour
15.826
29.04
77
100
0.5
124
0.1
24.8
149
0.10
Figure 8. Raw data sheet for hydrogen sulfide analysis.
-------
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.
14. Modified 25 mm A-H Microanalysis filter holder, Millipore, Catalog
#XX50 020 00.
15. Fluoropore 25 mm filters, Millipore, Catalog #FHLP 025 00, 0.5
micron pore size.
16. Flexible heavy insulation heating tape, Briskeat , width 1/2 inch,
length 48 inches.
17. Temperature Controller, Athena, 100-600 °F
18. Heated TFE Teflon hose, Technical Heaters, Inc., 5' x 1/4", tem-
perature limit 400°F.
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 m£ 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.
11. Micro burets, .5 m£ Teflon stopcock.
12. Erlenmeyer flask, 250 m£.
13. Erlenmeyer flask, 500 m£.
14. Dropping bottle, 60 m£ ground glass pipet and rubber bulb.
15. Beaker, 100 m£.
356
-------
16. Repipet dispenser, 2 m£.
17. Repipet dispenser, 10 m£.
18. Centrifuge tube, 15 m£.
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, 40mm 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 standarize 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(C,,H 0 ) -2H O, crystal, "Baker Analyzed" reagent.
^ O ^ ^ &
2. Sodium acetate, anhydrous formula weight = 82.03; chemical formula=
NaC2H3O2, analytical reagent grade, powder, Mallinckrodt Code 7372.
3. N, N dimethyl-para-phenylene diamine sulfate, formula weight =
370.47; chemical formula = (NH2C6H4N(CH3)252*H2S°4) ' 98 Percent
minimum by titration and spectro analysis, Eastman Code 1333.
4. Sulfuric acid, formula weight = 98.08; chemical formula = H,,SO4,
ACS analytical reagent grade, Mallinckrodt Code 2876.
5. Ferric ammonium sulfate, formula weight = 482.19; chemical formula=
Fe(NH4) (S04)2*12H20, ACS analytical reagent grade, crystals, Mallin-
ckrodt Code 5064.
6. Sodium sulfide, formula weight = 240.18; chemical formula =
Na S-9H 0, ACS analytical reagent grade, crystals, Mallinckrodt
Code 8044.
7. Iodine, formula weight = 253.81; chemical formula = I,» ACS
analytical reagent grade, resublimed, Mallinckrodt Code 1008.
357
-------
8. Hydrochloric acid, formula weight = 36.46; Chemical formula =
HCl, ACS reagent, assay 36.5-38.0 percent HC1, Eastman Code 13061.
9. Sodium thiosulfate, formula weight = 248.18; chemical formula =
Na2S2O3*5H2O, ACS analytical reagent grade, crystals, Mallinckrodt
Code 8100.
10. Starch soluble; chemical formula= (C6H10O5)n, certified ACS,
powder, Fisher Code S-516.
11. Arsenic trioxide primary standard, formula weight = 197.82;
chemical formula = AS203, powder, ACS analytical reagent, Mallin-
ckrodt Code 3668.
12. Potassium iodine, formula weight = 166.01; chemical formula =
KI, compacted crystal, "Baker Analyzed" reagent.
13. Sodium hydroxide pellets, formula wieght = 40.00; chemical formula=
NaOH, caustic soda, ACS analytical reagent grade, Mallinckrodt
Code 7708.
14. Sodium bicarbonate, formula weight = 84.01; chemical formula =
NaHCO_, powder, ACS analytical reagent grade, Mallinckrodt Code 7412.
15. Sodium carbonate anahydrous 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 = K Cr O , "Baker Analyzed" reagent.
17. Litmus test paper, blue, reagent ACS, Fisher Code 14-875.
18. Glycerol, formula weight = 92.10; chemical formula =
HOCH2CH(OH)CH2OH, 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
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) gylcerol, in vacuum boiled, deionized water and di-
358
-------
luting to 1 liter. A 2 ml 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 precipating 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 m£ 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 Standarization Solutions
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 m& 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 m£ 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 pervents the formation of hydrogen
sulfite ion from thiosulfate ion in the presence of acid. The addition of
such substances as chloroform, sodium benzoate, or mercury (II) iodine in-
hibits the growth of bacteria. This solution is stable for several weeks
but should be discarded if it becomes turbid. It is standarized 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 boiling water. The resulting slurry is
slowly poured into 50 mJl of boiling deionized water and heated until clear
(about 2-3 minuted) . The solution is then cooled and centrifuged for several
minutes. The supernatant liquid is decanted into a clean, 60 mJl 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
359
-------
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 dis-
solved 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 solu-
tion is approximately 0.1 N and is used to prepare the 0.01 N solution. The
0.1 N solution is standarized against arsenic trioxide. The dilute iodine
solution is prepared by diluting 10 m& of the 0.1 N solution to 100 m&
with deionized water. This solution is standarized with a previously stan-
darized thiosulfate solution. The dilute iodine solution is prepared and
standarized daily because of the volatility of iodine and oxidation by dis-
solved oxygen.
360
-------
REFERENCES
Stecher, P.G., (Ed.), The. Merck Index. An Encyclopedia of Chemicals and Druas
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. ~ "
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, 1974.
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 Sulfue 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 Tran., 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. , Analy. Chem., Vol. 48, pg. 918. 1976.
Hseu, T.M. and Rechnitz, G.A. , Anal. Chem., Vol. 40, No. 7, pg. 1054.
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.
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.
361
-------
"Analysis of H S and SO 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. andBydalek, 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.
Buck, M. and Gies, H., Staub, Vol. 26, pg. 1966.
Jacobs, M.B., Braverman, M.M., and Hochheiser, S., Anal. Chem., Vol. 29, pg.
1349, 1957.
Cave, G.C.B., Tappi, Vol. 46, pg. 1, 1963.
Budd, M.S. and Bewick, H.A., Anal. Chem., 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.
362
-------
APPENDIX H
AMMONIA PROCEDURE
363
-------
THE MEASUREMENT OF AMMONIA IN EXHAUST
The measurement of ammonia in dilute automotive exhaust is accomplished
by bubbling dilute exhaust through glass impingers containing dilute sulfuric
acid. Ammonia is complexed by the acid to form a stable sulfate salt which
remains in solution. The exhaust sample is collected continuously during a
test cycle. A sample from the impinger is analyzed for ammonia by the use
of an Ion Chromatograph. The concentration of ammonia is calculated by
comparison to a standard.
SAMPLING SYSTEM
Two glass impingers in series, each containing 25 m£ of 0.01 N sulfuric
acid, are used to collect exhaust samples for the analysis of ammonia. A
flow schematic of the sample collection system is shown in Figure 1. The two
impingers together trap approximately 99+ percent of the ammonia. The temp-
erature 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 Vminute by the sample pxanp.
A dry gas meter is used to determine the total flow through the impingers
during a given sampling period. The temperature of the gas stream is
monitored by a thermocouple immediately prior to the dry gas meter. A
drier is included in the system to prevent condensation in the pump, flow-
meter, dry gas meter, etc. The flowmeter in the system allows continuous
monitoring of the sample flow to insure proper flow rates durinq sampling.
When sampling diesel fueled vehicles, a filter, located between the on-off
solenoid valve and the dilution tunnel, is used to prevent diesel particulate
from contaminating the sampling system. The line connecting the filter to the
dilution tunnel and the line connecting the filter to the solenoid valve are
heated to 175°F in order to prevent water from condensing in the sample lines.
Several views of the sampling system are shown in Figure 2.
PROCEDURE
Ammonia in exhaust is collected in two impingers connected in series with
each impinger containing 25 mJl of 0.01 N H2SO4. These two impingers trap 99+
percent of the ammonia. After the acidification of ammonia which takes place
at ice bath temperatures (0-5°C), the ammonia samples are poured into poly-
propylene bottles and stored. The samples are then ready for NH4+ analysis
on the ion chromatograph (Figure 3). Approximately 2 m£ of sample are used
to purge a 0.1 m£ sample loop, after which 0.1 m£ of sample is injected into
the eluent stream. Separation of ions occurs in the separator (analytical)
column. The background conductance of the eluent (0.0075 N HNO3) is neutral-
ized in the suppressor solumn. The 9 x 250 glass cation suppressor column is
packed with AG- X10, a strong base ion exchange resin in the hydroxide form.
A patented resin containing a sulfonic acid cation exchanger is packed into
364
-------
Gas Temperature
Digital Readout
CO
cr>
en
ample
Probe
Regulating
Valve
Heated Lines
Off Solenoid Valve
Flowmeter
Dilute
Exhaust
Ice Bath Temperature
Digital Readout
Dry
Gas
Meter
hla|3|q|5l
Gas Volume
Digital Readout
Figure 1. NH sample collection flow schematic.
-------
Front View
Digital
Readout
Flowmeter
Regulating
Valve
Close-up of Upper Front
Figure 2. Ammonia sampling system
366
-------
Solenoid
Impinger
Ice Bath _
Close-up of Impingers (Side View)
Dry Gas Meter
Pump
Rear View
Figure 2 (Cont'd). Ammonia sampling system.
367
-------
Polyethylene Storage
Bottle
Suppressor
Column
Ion
Chromatograph
Recorder
Figure 3. NH ion chroma.tograph-
368
-------
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 sup-
pressor column in the dilute nitric acid eluent. The hydroxide form of the
suppressor resin neutralizes the acid and then converts the cations to their
hydroxides.
HNO + Resin - OH^=± Resin - NO + H 0
Cation + NO3 + Resin - OH ^ ~- Resin - N03 + Cation"1" + OH~
The conductivity cell produces a signal for the species of interest,
NH4OH, but doesn't "see" the neutralized eluent, deionized water. Con-
ductance is interpreted as a recorder trace (chromatogram) 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 in Figures 5 and 6.
After the 12-20 minute analysis the collection conditions and areas of sample
and standard are used to compute NH3 concentration from a Hewlett-Packard 67
program.
CALCULATIONS
The purpose of this procedure is to determine the concentration of
ammonia in automotive exhaust. To do this, ammonia is converted to the
protonated form, NH4+, which is measured on the ion chromatograph. The cal-
culations 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 67 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
W y*3 and ppm NH 3 is computed. For illustration, two examples using infor-
mation from the data sheet will be included 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. A standard with peak size close to that of the sample
is selected.
PAo= pAsa x cst
sa
= c^T PAst
where PAst = peak area of the standard
PAsa = peak area of the sample
yg NH4+
C«,4. = concentration of the standard (—-~ )
• O l> iUA/
yg NH4+
Csa = concentration of the sample ( —)
369
-------
CVS
I
Glass
Impinger
Unused Sample
Saved as
Needed
Sample Analysis
in Ion Chromatograph
with Conductivity Cell
A/D
Converter
I
Recorder
Hewlett-Packard
3354
Computer System
Figure 4.
NH analysis flow schematic.
370
-------
Sample NYCC B-2 23111 6-19-79
Attenuation lOumho
Date 8-9-79
Eluent 0.01 N HNO-^ Sample Loop Size 0.2 mil
Eluent Flowrate 138 mjt/min Chartspeed 12
Pre column 3 x 150 mm, glass
Separator Column 6 x 250 mm, glass
Precolumn and Separator Column Packing Patented
resin
Suppressor Column 9 x 250 mm, glass
Suppressor Column Packing AG IX- 10
15 10
Retention time, minutes
Figure 5. Sample chromatogram.
371
-------
Sample 2.0
2 Attenuation 3yimho
Eluent 0.0012 N H->SC>4 Sample Loop Size 0.1 mi
Eluent Flowrate 167 ml/hr Chartspeed 12 in/hr
Analytical Precolumn 3x150 mm glass
Analytical Column 6x250 mm glass
Column Packing Patented resin-sulfonic acid
CO cation exchanger
Suppressor Column 9x250 mm glass
'• Packing AG 1X-10
15 10 5 0
Figure 6. Standard chromatogram
372
-------
Insii unions
8,
Aiuuoiiia In Exhaust ^*
srgp
j
*
°>
'J3
1
2
3
4
i 5
b
7
i !»
f
y
10
I
I
1
INSTRUCTIONS
Switch to on; Switch to t ui»
Peed Card in from riyht tu lull
Set Decimal place
Input Sample Volume
Input Barometric Pressure
Input Sample Temperature
Input Dilution Factor
I n£u t Absorb i ng Reagen t Vo 1 uine
Input Standard Concentration
Input Standard Area
Input Sample Area
Output Sample Concentration
Output Sample Concentration
. . -
. _ ..
-
"
INPUT
DATA UNITS
ft3
"ttq
T
ml
ligNH^/me
.
......
- -
.... . .
KtVb
1
! i
I
,| -!>,-, \
I A :
\ K/S; | ,
| R/S ! 1
1 R/S . '
i K/S' '
R/Si i
1 R/S|
R/si :
1 R/S! :
II
h i 1 RTH
i! 1
i i i
i i
i i
i i i
i ii
I i
il i
1 1 i
i
i
i I
1 M
i •!
: I i
1 ' '
1 ! i
1 ! '
1
1 1
1 !
i i
OUTPUT
M'lNllj/iu-1
I'I'M
Figure 7. HP-67 user instructions.
373
-------
STEP KEY ENTRY KEY CODE
COMMENTS
STEP KEY ENTRY KEY CODE
COMMENTS
001
010
020
030
040
050
f T.HT. A
3
5
1
1
i
5
2
a
X
?
q
.
9
2
T
R/S
X
STO 1
H/S
4
6
0
•f
RCL 1
h X? v
i
h 1/X
R/S
X
R/S
X
R/S
X
R/S
R/S
X
0
.
9
4
4
X
R/S
0
•
0
0
1
4
1
X
R/S
h RTN
31 25 .IX,
03
05
83
ni
01
fll
05
02
08
71
n->
nq
83
09
02
81
84
71
33 01
84
04
06
00
61
34 01
35 52
81
35 62
84
71
84
71
84
71
84
81
84
71
00
83
09
04
04
71
84
00
83
00
00
01
04
01
71
34
35 22
Input Sample Vol, ft3
Input Baron
Input 5dir£>l
Input Dilut
Input Aba £
Input Stand
ug NH4
Input Stand
later, "Hg
e Temp.
°F
ion Factoi
oln Vol,
mi
Cone ,
VmJt
Area
Input Sample Area
Output Cone, M9 NHj/m-*
0 1 2
SO SI S2
A B
3
S3
C
060
070
oeo
090
too
10
REGISTERS
4
S4
567
SS S6 S7
D !
s a
SB S9
; r
Figure 7 (Cont'd). HP-67 program form.
374
-------
This equation gives the concentration of NH4+ in the sample. To convert
NH
to 3 the ratio of the formula weights of NH3 to NH4+ is multiplied
by
-sa*
yg
17.03(
yg NH3
y mole
sa
18.04(
4
yg NH.
mi
y mole
The next step involves the determination of the amount of NH3 collected
in the bubbler. This is obtained by multiplying the volume of absorbant
(absorb, vol.) and a dilution factor (DF) by the concentration of ammonia
collected.
' yg NH.
0.944 C
sa
mi
x absorb, vol.(mi) x DF = yg NH3
To find the concentration of NH3 in exhaust the volume of gas collected
is corrected to the specif ied temperature and pressure (68°F = 528°R and
29.92 "Hg) . The volume as read from the digital readout in cubic feet is
converted to cubic meters by dividing by 35.31 f t3/m3.
VOL, ft3 x B.P. "Hq x 528°R
3 29.92 "Hg (TEMP,
35.31 £|—
(TEMP, °F + 460)
m-
where VOL = volume of gas collected, ftj
B.P. = collection pressure, "Hg
TEMP = collection temperature, °F
The concentration of NH, is then calculated by dividing yg NH3 by the volume
of gas.
Cst
yg NH *
mi J
x PA
CZ fH
v nh^nrh vol . m'r X IYF
PAst
VOL, ft3 .. B.P. "Hq 528°R
35.31
29.92 "Hg
(TEMP, °F + 460)
rn"
0.944 x cst
\
yg NH4+
mi / X
x al^sorb. vol, w
i x DF
VOL, ft x PAst
375
-------
35.31 ft3 v 29.92 " Hg tTEMP, °F + 460)
1? _
B.P., " Hg x 528°R
yg NH3
= -3
m
The concentration of ammonia can also be expressed in ppm NH^ by taking
into consideration the density of the gas at the desired conditions (68°F,
29.92 "Hg). . The fifth edition of the Matheson Gas Data Book lists the
specific gravity of ammonia gas at 70°F and 1 atm pressure at 1.411 £/g. The
inverse of the specific gravity gives a density of 0.709 y. If the volume of
l£ of gas is corrected for temperature,
528°P
V = X
(70°F H- 460)
it can be divided into the weight of the gas to give the density at 68°F
0..996 £
= 0.00141
yg NH3
yg NH,
When the inverse of density is multiplied by the concentration in j--
the concentration is given in ppm NH3.
m yg NH3 m
Sample Calculation
The two examples will be calculated from information recorded on the
data sheet (Figure 8) . This information does not necessarily represent
actual experimental data but serves as a means of confirming calculations
done by hand with those done with the Hewlett-Packard Calculator.
Example 1
Assume that in the FTP-2 driving cycle that 2.652 ft of dilute exhaust
is collected in 25 m£ of 0.01 N H2S04 at 74.2°F and 29.42 "Hg. when the
undiluted sample is injected into the Ion Chromatograph it yields a peak area
yg NH4+
of 3800 counts. A 9 - T - standard similarly injected produces an area of
4800 counts.
376
-------
SwRI Project No.
Fuel:
Test No.
Test Date:
Vehicle:
Sample Collection By:
General Comments:
CVS No.
Tunnel Size
Driver:
Chemical Analysis By:
Miles:
Calculations By:
Driving
Cycle
FTP-1
FTP -2
FTP-3
SET-7
HFET
NYCC
BG
Sampling Conditions
Volume
Ft3
1.546
2.652
1.625
4.212
2.197
1.890
2.600
1
B.P.
"Hg
29.40
29.42
29.44
29.45
29.47
29.49
29.50
2
Temp.
oF
74.0
74.2
74.5
- 74.3
74.7
74.7
74.4
3
Dilution
Factor
1
1
1
1
1
1
1
4
Absorb.
Reagent
Volume m£
25
25
25
25
25
25
25
5
Standard
ygNH^/mfc
4
9
5
15
9
8
1.5
6
Area
3500
4800
3600
4500
5000
5200
3300
7
Sample
Area
2500
3800
2700
3500
3500
3500
2500
8
ygNH3/m3
1590
2300
1980
2370
2460
2440
374
ppm
2.23
3.25
2.79
3.35
3.46
3.44
0.53
CO
Figure 8. NH Data Sheet.
-------
The equation needed to calculate the concentration of ammonia follows:
yg NH3 _
°'944
x PAca x absorb, vol,
X
3
m
DF
VOL, ft3 x PAst
x 35.31 ft3/m3 x 29.92 " Hg x (TEMP, °F 4- 460)
B.P., "Hg x 528°R
0.944 x 9 x 3800 x 25 x l x 35.31 x 29.92 x (74.2 -I- 460)
2.652 x 4800 x 20.42 x 528
Ug NH3
= 2300 5—
ppm NH3 = 2300 x 0.00141 =3.25 ppm NH3
Example 2
Assume that 1.890 ft3 of dilute automotive exhaust was collected in 25
m£ of 0.01 N I^SC^ during the NYCC driving cycle. The sampling conditions
during this test were 74.7°F and 29.49 "Hg. When the undiluted sample was
injected into the ion chromatograph the peak produced had an area of 3500
counts. When an 8 ^g ^ standard was injected it produced an area of 5200
counts . m&
The same equations in Example one are used to give concentrations of
NH3 of 2440 and 3.44 ppm NH3
m3
EQUIPMENT
The equipment section lists the equipment used in the ammonia procedure.
It is divided into four sections corresponding to each major division in the
procedure: Sampling, Analysis, Water Filtration and Sample Preparation.
For convenience the item, manufacturer, model number and any additional
pertinent information are included.
Sampling
1. Glass impingers, Ace Glass Products, Catalog #7530-11, plain
tapered tip stoppers with 18/7 arm joints and 29/42 bottle joints.
2. Flowmeters, Brooks Instrument Division, Model 1555, R-2-15-C,
sapphire ball, 0-5 lit/min range, graduated 0-15.
378
-------
3. Dry gas meter, American Singer Corporation, Type AL-120, 60 CFH
C3.p3.cxty •
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 IDV2-144N
CAl.
9. Miscellaneous Teflon nuts, ferrules, unions, tees, clamps and con-
nectors, etc.
10. Miscellaneous electrical switches, lights, wiring, etc.
11. Regulating valve, Nupro 4MG, stainless steel.
12. Iron/Constantan type J, Thermo Sensors Corporation, single thermo-
couple, 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.
15. Modified 25 mm A-H microanalysis filter holder, Millipore, Catalog
#XX50 020 00.
16. Fluoropore 25 mm filters, Millipore, Catalog #FHLP 025 00, 0.5
micron pore size.
fn\
17. Flexible heavy insulation heating tape, BriskeaM width-1/2",
length-48".
18. Temperature Controller, Athena, 100-600°F.
19. Heated TFE Teflon hose, Technical Heaters Inc., 5' x 1/4", tem-
perature limit 400°F.
Analysis
1. Dionex Model 10 Ion Chromatograph.
379
-------
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 mH 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 mJZ, volumetric pipets.
8. Class A, 10 m£ volumetric pipets.
9- Class A, 20 m£ volumetric pipets.
10. Class A, 25 m£ volumetric pipets.
11. Class A, 50 m£ volumetric pipets.
12. Class A, 100 m£ volumetric pipets.
13. Class A, 100 m£ volumetric flasks.
14. Class A, 1000 m£ 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 sec-
tion. In addition to the function of each reagent, the purity, manufacturer
and catalog number are also listed.
380
-------
1. Water-deionized and filtered through a 0.22 micron filter.
2 . Standard
Ammonia sulfate, (NH4)2S04, formula weight = 132 146 ACS
analytical reagent grade, granular, j.T. Baker Chemical Co.
TT\J /-7^ *
3 . Absorbant
Sulfuric acid, H2SO4, formula weight = 98.08, ACS analytical
reagent grade, Mallinckrodt #2876.
4. Eluent
Nitric Acid, HN03, formula weight = 63.01, ACS analytical
reagent grade (Ultrex) , 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 in 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) 2804 )
The stock solution is prepared by diluting 0. 3660 g of (NH4)2S04 to
1000 m£ with water. This yields a solution with a concentration of 100
yig NH^ . Less concentrated standards are made up by diluting portions of
the stock solution to 100 mJl with water using volumetric glassware.
absorbing Solution (0.01 N
The absorbant is prepared by diluting 20.0 m& of the certified 1.000 N
sulfuric acid to 2000 mH with water.
Sluent (0.0075 N HNOs) *
A 1 N HNO3 stock solution is prepared by diluting 62.5 m£ of concen-
trated nitric acid to 1000 mH with water. The eluent is prepared by further
diluting 15 m£ of the stock solution to 2000 mH with water.
Regenerant (0.5 N NaOH)*
40.00g of NaOH is dissolved in water in a 2 liter volumetric flask and
diluted to volume.
*4 liters of each of these solutions are prepared and stored in labeled
polyethylene cubitainers .
381
-------
REFERENCES
Braker, W. and Mossman, A.L., Ma the son 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 Wollish, 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.), 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.,
1954, Vol. 2, pg. 168.
Milner, O.I. and Zahner, R.J., Anal. Chem., Vol. 32, pg. 294, 1960.
Conway, E.J., Micro-Diffusion Analysis and Volumetric Error, 2nd Edition,
Crosby-Lockwood, London, 1947.
Altman, 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.
Kolthofi, 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.
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.
382
-------
Norton, O.R. and Mann, C.K., Anal. Chem., Vol. 26, pg. iiso, 1954
Clear, A.J. and Roth, M., in Treatise on Analytical Chemistry. Kolthoff I M
andElving, P.J. (Eds.), Interscience, N.Y., 1961, Part Il,"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.
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., Uipublished 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.
Stockdale, D., Analyst, Vol. 84, pg. 667, 1959.
Sadek, 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 Inorgrnic Analysis, 3rd Edition, Longmans, Green,
London, 1969, pg. 566.
Furman, N.H., (Ed.), Scott's Standard Methods of Chemical Analysis, 5th
Edition, Van Nostrand, N.Y., 1958, Vol. 1, pg. 2336-
383
-------
Vogel, A.I., Quantitative Inorganic Analysis, 3rd Edition, Longmans, Green,
London, 1961, pg. 1092.
Steyermark, A., Quantitative Organic Microanalysis, Blackstone, 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., Eckert, 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.
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.
384
-------
Christian, G.D., Knoblock, E.G., Purdy, W.C., Anal. Chem., Vol. 35, pg. 2217,
1963.
Zweidinger, R.B., Tejeda, S.B., Sigsby Jr., J.B., and Bradow, R.L., "The
Application of Ion Chormatography to the Analysis of Ammonia and Alkyl Amines
in Automotive Exhaust," Symposium on Ion Chromatographic Analysis of Environ-
mental Pollutants, EPA, Research Triangle Park, N.C., April 1977.
385
-------
APPENDIX I
ORGANIC SULFIDES PROCEDURE
386
-------
THE MEASUREMENT OF ORGANIC SULFIDES IN EXHAUST
The measurement of organic sulfides; carbonyl sulfide (COS), methyl
sulfide (dimethylsulfide, (O^^S), ethyl sulfide (diethylsulfide, (C2H5)2S)
and methyl disulfide (dimethyldisulfide, (CH3)2S2) 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 temperature. The exhaust
sample is collected continuously during the test cycle. The organic sulfides
are thermally desorbed from the traps into a gas chromatograph sampling
system and injected into a gas chromatograph equipped 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.
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 1'. The trap collects 99+ percent of the sulfides at flows up
to 130 mil/min. Several views of the sampling system are shown in Figure 2.
The various components of the sampling system and their functions are listed
below.
Item Component
NaHCO trap
Tenax-GC trap
Description
Perma-Pure Drier
Sample Pump
2" x 3/8" OD x 0.035" wall stainless steel
cartridge packed with 5 percent NaHC03 on
45/60 mesh Chromosorb P (this trap removes
SO2 from the exhaust sample).
2" x 3/8" OD x 0.035" wall stainless steel
cartridge packed with preconditioned 60/80
mesh Tenax-GC (this trap collects and con-
centrates the organic sulfides).
Model PD-62512S 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.
387
-------
CO
00
00
Flip-Top
Filter
Perma Pure
Drier
Tunnel
Flowmeter
Regulating
Valve
Sample
Pump
Tenax-GC
Trap
Low Temperature
Bath
Figure 1, Organic sulfide sample flow schematic.
-------
Figure 2 . Several views of the organic sulfide sampling system.
389
-------
Figure 2. (Cont'd). Several views of the organic sulfide
sampling system.
390
-------
5 Low temp bath A. constant low temperature bath is obtained by
using a CO2-isopropyl alcohol slurry. A bath
temperature of -76 to -78°C is obtained with
this bath.
6 Flip-top filter 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.
7 Regulating valve A Nupro SS-4MG regulating needle valve is used
to control the exhaust flow through the NaHCO
and Tenax-GC traps. 3
8 Flowmeter A Brooks Model 1550 flowmeter with R-2-15-AAA
. ('SS 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 3 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 4) . The trap can now be connected by
the miniature quick connects to the sampling system or the desorbing system
with ease.
ORGANIC SULFIDE TENAX-GC TRAP CONDITIONING 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 outlined
in this section.
391
-------
Figure 3. Tenax-GC trap
Figure 4. Tenax-GC trap with quick connect.
392
-------
The Tenax-GC traps can be used repeatedly provided they are properly
conditioned. Conditioning of the Tenax-GC traps is accomplished by purqinq
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 conditioninq
two Tenax-GC traps simultaneously. A flow schematic of this system is
illustrated in Figure 5. Several views of the Tenax-GC sample conditioning
system are presented in Figure 6. This system has been shown to reproducibly
condition Tenax-GC traps to a negligible level of organic sulfides.
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.
The temperature readout on the furnace should be verified
at least once a week with a digital thermocouple.
2. Connect the traps according to the flow schematic. Insert
the traps into 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 sulfide,
ehtyl sufide, and methyl disulfide) in dilute exhaust is accomplished *Y
collecting the organic sulfides in Tenax traps at -76°C. The orgamc sulfides
393
-------
Drier
Zero
N2
Regulating
Valves
Tenax
Flowmeters
Traps
L
Furnace @ 325+ 25°C
Vent
(500 ml/min)
Vent
(500 ml/min)
Figure 5. Flow schematic for conditioning Tenax-GC traps
for organic sulfide analysis (dual system).
394
-------
Figure 6. Several views of Tenax-GC trap conditioning system
395
-------
Figure 6 (Cont'd). Several views of Tenax-GC trap
conditioning system .
396
-------
are thermally desorbed from the trap into a gas chromatograph sampling system
The organic sulfides are analyzed by injecting the desorbed sample into a qas'
chromatograph equipped with a flame photometric detector. A standard blend
containing known amounts of the four organic sulfides is injected into the qas
chromatograph to quantify the results. From the GC analysis of the sample
the analysis of the standard blend, and the measured volume of exhaust sampled
the concentration of the organic sulfides in the exhaust can be determined
The analysis flow schematic for the organic sulfides is shown in Figure 7 ' 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 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 bicarbonate trap which removes any
interfering SC>2 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 main-
tained 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, the 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 8) and the sample is immediately injected and placed into
the Lindberg furnace operating at 300°C (Figure 9) . The carrier gas upon
injection flows through the loop carrying the contents into the gas chromato-
graph where the sulfides are separated and identified by their retention
times. After the peaks of interest have passed through the column in the gas
chromatograph, the system is backflushed to remove any high molecular weight
impurities that could interfere with later analysis (Figure 10) .
The gas chromatograph system used to analyze the organic sulfide sample
is shown in Figure 11. The system consists of a Perkin-Elmer 3920B GC, and
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 photometric 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 m£'min. The optimum hydrogen and air flows are 40 m£/min and
360 mVmin, respectively. The column temperature, after injection of the
sample is programmed from 30°C to 140°C at 8° a minute. In a chromatogram
397
-------
CVS
Tenax
GC traps
I
Sample Analysis
in gas chromatograph
with FPD
A/D Converter
I
Recorder
Hewlett-Packard
3354
Computer System
Figure 7. Organic sulfide analysis flow schematic.
398
-------
Analytical Column
(6* x 1/8" TFE, 60/80 Tenax-GC)
• ••••••••••••••••a mmmm/tm m mm
. r*. /
Seiscor Valve
(backflush configuration)
Vent
Perkin-Elmer
3920B
Flame Photometric
Detector
Seiscor Valve
(Gas Sampling Configuration)
Carrier
Gas
Tenax Trap
300°C furnace
SS miniature
quick connect
Regulating
Valve
Female
Quick-Connect
Flowmeter
Figure 8 . Flow schematic of organic sulfide analysis system
(Step 1 - connect Tenax trap in CSV).
399
-------
Analytical Column
(6* x 1/8" TPE, 60/80 Tenax-GC)
Seiscor Valve
(backflush configuration)
Perkin-Elnter
3920B
Flame Photometric
Detector
Seiscor Valve
(Gas Sampling Configuration)
Vent
'—T
• •••»•••»• •«'•»% •§•'!•••• «1» * •» • *
Carrier
Gas
Tenax Trap
__ I : :
•^TjM-^M^J •«••••»••<••»•»•••••
\L 300°C furnace
SS miniature
quick connect
Regulating
Valve
Pump
Flowmeter
Female
Quick-Connect
Figure 9. Flow schematic of organic sulfide analysis system
(Step 2 - inject Tenax trap contents into GC system).
400
-------
Analytical Column
(61 x 1/8" TFE, 60/80 Tenax-GC)
mmmmmmmmmmmm mmmmm mmm/m mmmmmmmmmmmmmmm
-P I A
Seiscor Valve
(backflush configuration)
Vent
Perkin-Elmer
3920B
Flame Photometric
Detector
Seiscor Valve
(gas sampling configuration)
/
Carrier
Gas
Tenax Trap
*»•» mmmmw^mmmmmmmm |
* •
300°C furnace
SS miniature
quick connect
Regulating
Valve
Flowmeter
Female
Quick-Connect
Figure 10- Flow schematic of organic sulfide analysis system
(Step 3 - backflush analytical column).
401
-------
<
' • I
Figure 11. Organic sulfide analytical system.
-------
of a standard sample (Figure 12) containing the four sulfides, the first
peak eluted is carbonyl sulfide, followed by methyl sulfide, ethyl sulfide and
methyl disulfide. The GC sulfide peaks are recorded 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 nitrogen is used to dilute the permeation gases to
the concentrations desired. The 10 m& sample loop is purged with this per-
meation gas for 10 minutes (Figure 13) and the 10 m£ of permeation gas is then
injected into the 'GC and analyzed (Figure 14) . From the standard peak areas,
the exhaust sample peak areas and the volume of exhaust sampled, the concen-
tration of the organic sulfides in exhaust can be determined.
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 (GSV) configuration with the second Seiscor valve in a
backflush configuration. These valves are pneumatically operated and
electrically controlled. The electrical schematic for the control of the
Seiscor valves using the ATC timers and ASCO electric solenoid valves is
shown in Figure 15. The flow schematic for vacuum and pressure lines to the
Seiscor valve are presented in Figures 8-10 and 13-14. The Seiscor valves
have been found to operate much more dependably if a vacuum assist is included
in the valve actuation controls.
CALCULATIONS
This procedure has been developed to provide the user with the concen-
trations of the organic 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
determining 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..,. (mVmin) x Ti (sec)/60 sec/min
Vol exp (m£) = volume of gas sample in m£
F.R.j (mVmin) = flow rate of exhaust sample in itd/min
Ti (sec) = sampling time in minutes
60 sec/min = conversion of sample time in seconds to minutes
[Equation 13
403
-------
Sample Permeation Blend DaiJtov. 29. 1977
in.irumant PE 392OB Operator P. Saunders
Column 6 ft. 1/8 61T
Packed with NA % wt.
on 60/80 meih TENAX-GC
Run ISO ® NA °C mine 30 cc/min.
Liq. Phase
Support
Carrier
e 75 otig MA Rotameter Reading
held ® °C I SO far min., prog to
/min. Held for
held for
Inlet
Detector 160
Hvd 40 prig
Air 70 psig
( I NA Olio
Rotameter Rdg.
Rotameter
Rotameter Rdg.
- NA cc/min
Recorder 1 in/min weed 1 mV.F.S.Soltec Type
Injection 10 ul indicated 10 ul net A° ul Actual
Sampling Device Gas Sampling Valve
±=S::-4-v---i
13 12 11 10 9 8
Retention Time, minutes
Figure 12. Chromatogram of organic sulfide standard.
404
-------
Analytical Column
(61 x 1/8" TFE, 60/80 Tenax GC)
Seiscor Valve
(backflush configuration)
Seiscor Valve
(Gas Sampling Configuration)
PE 3920B
Gas Chromatograph
with FPD
Carrier Gas
SS miniature
quick connect
10 ml sample loop
Vent
Regulating
Valve
Pump
|—| Calibration
gas in
Flowmeter
Female
quick-connect
Figure 13. Flow schematic of organic sulfide calibration system
(Step 1 - purge of sample loop of CSV).
405
-------
Analytical Column
(6' x 1/8" TFE, 60/80 Tenax GC)
Seiscor Valve
(backflush configuration)
PE 3920B
Gas Chromatograph
with FPD
Seiscor Valve
(Gas Sampling Configuration)
Carrier Gas
Vent
'—T
•••••••••••••••••••••••••••^
SS miniature
connect
X
10 ml sample loop
Regulating
Valve
Pump
|—\ Calibration
gas in
Flowmeter
Female
quick-connect
Figure 14. Flow schematic of organic sulfide calibration system
(Step 2 - inject calibration gas into GC system).
406
-------
' 2
3
4
9
11
,12
"I14
Jl5
^•N
r-~T
...... _|
1 ">
I
\ A ~ *
\ *±
1 ,
A -*
1 „
^
V,
p
n^v
solenoid
|
8l "3
a sr6
,n
Aux
-111
I
[12
-*14
I
115
"U6
i
backflush
solenoid
ATC
Timer
ATC
Timer
Figure 15. Electrical schematic for organic sulfide analysis system .
-------
The next step is to correct the volume of exhaust sampled to a standard
temperature, 68°F, and pressure, 29.92" Hg, by use of the equation
p xy P x V
exp exp _ corr corr
T T
exp corr
V = experimental volume of gas sample in m£
vcorr = volume °f 9as sample i-n m^ corrected to 68°F
and 29.92" Hg
pexp = experimental barometric pressure
=29 92" Ha
«-^ «y
Texp = experimental temperature in °F + 460
Tcorr = 68°F + 460 = 528°R
Solving for Vcorr gives:
P ("Hg) xv Cft3) x 528°R
V exp exp
corr =
Texp (°R) x 29.92" Hg
[Equation 2]
Substituting Vol exp (md) from Equation 1 into Equation 2 gives:
P ("Hg) x F.R.T (rnVmin) x Ti (sec) x 528°R
corr
Tovr, (°R) x 29.92" Hg x 60 sec/min
exp
The next step converts the volume from m£ to cubic meters by use of the
conversion factor; 1 cubic meter = 10 m&.
V (m3) P ("Hg) X F-R-T (mVmin) x Ti (sec) x 528°R
corr = exP i •
Texp (°R) x 29.92" Hg x 60 sec/min x 1Q6 md/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 fol-
lowing equation holds:
A A . .
sam std
= yg sample in Tenax trap
A = GC peak area of sample in relative units
Ugstd = yg standard
= GC peak area of standard in relative units
Solving for yg sample gives :
yg . , x A
ug std
sam A
std (Equation 4)
408
-------
The yg of standard for each of the organic sulfides is determined from
the permeation rates of the permeation tubes containing each of the sulfides,
the flow rate of the diluting gas, and the volume of gas sampled (10 m£).
ua std = P.R. (ng/min) x 10
F*R'II (m^/min> x 10°°
[Equation 5]
P.R. (ng/min) = permeation rate of permeation tube at 40° in ng/min
10 m£ = volume of calibration gas injected
F.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 4 gives:
P.R. (ng/min) x 10 m£ x A
_ sam
yg sam - F.R (m£/min) x 1000 ng/yg x Agtd
[Equation 6]
To obtain yg sample/m3, Equation 6 is divided by Equation 3 to give:
P.R. (ng/min) x 10 m£ x ASS^ x T (°R)
yg sam/m3 = —
F.R.I:E (mVmin x 1000 ng/yg x Astd x Pexp ("Hg)
29.92" Hg x 60 sec/min x 1Q6 mjj,/m3
F.R. (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_
T " Tj
VT = 22.4
TI = 32°F + 460 = 492°R
V = volume at 68 °F
T = 68°F + 460 = 528°R
Solving for V gives:
V^ 22.4^528
v - T 492
Since one mole of gas occupies 22.04* at 68°F, the density can be found
in g/S, by dividing the molecular weight in g/mole by 24.04
mol. wt. g/mole
den
409
-------
The density in yg/m& can be found by converting g to yg and H to m£
as follows:
mol. wt. g/mole 1 x lp6yg/g _ mol. wt. x IQQQ
den yg/mJt = 24.04 Vmole 1 x lo3mA/£ 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&
. 3 . . „ mJl
ppm = yg/m r yg/m£ = ^3-
Using Equations 7 and 8 gives the ppm concentrations in the form of the
raw data.
P.R. (ng/mJl) x 10 mH x A x T (°R) x 29.92" Hg
sam exp
ppm =
F.R. (mVmin x IQOO ng/yg x Astd x Pexp ("Hg)
.I]:
60 sec/min x 1Q m&/m
F..R. (mVmin)xTi (sec) x 528°R
_ 24.04 Vmole _
mol. wt. (g/mole) x 1000 yg-Jl/g-m&
[Equation 9]
At this point, the concentration can be expressed in yg/m (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 67 program was developed to calculate the organic sulfide
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-67 program and manual calculations.
Manual calculations for driving cycle FTP-1
3 P.R. (ng/m£) x 10 m£ x A x T (°R) x 29.92" Hg
yg/m COS - F.R.I3. (mVmin) x 1000 ng/yg x Astd x Pex f ("Hg)
x _ 60 sec/min x IQ mJl/m _
F.R. (m£/min) x Ti (sec) x 528°R
667.5 ng/m£ 10 m& x 3Q20Q x 535°R x 29.92" Hg
580 mVmin x 1000 ng/yg x 18514 x 29.80" Hg
x 60 sec/min x ip6
130 m£/min x 504 x 528°R
410
-------
iiisiniciioiis
ukGAtllC SULFIDES IN EXHAUST
STEP
Ul
Q jj
0 i
1
i
a
4
b
i
7
8
9
)0
11
12
13
14
1 ~~>
It,
17
Ib
1'J
1 20
21
22
23
24
25.
INSTRUCTIONS
SwiJ.cH UQ >iui switcJi to. run
?'^§d card in from right to left
Set dec ima.i Piai'fi
ln£y£ Samjjle Flow B^Le
Ijjjjut Sampling Time
lnp.ut Barometric Pressyi:^
lutut Salable Teniifijaliirs.
Input OUutian Qas flow fjjr tarmeaiiun staiiaaias
Int^ut Permeation Rate COS
Input Standard Area COS
Input Srfrgpl tj Area COS
Output CojiceutxdLiiiix .COS
Output Concentration COS
Input Permeation Rate (CH^^S
ItlPUt Sta.Hd^rd Aras (QLljlaa
Input Sample Area tCHj)2^
UuLput LoucBUtratiou (Ca3)aS
QUUtUL CilULMl!.rat_LQU ICtijJjS
Input Permeation Kate (C,Hc)-,S
r _ - . .^- . , — - ,~_.-~ - X 3-4#-
?!ieut Standard. Area (C^Js^S
InjJUL Sdfitple Art;a tC^H^) jS
Output Cuiicentration (€2^1^)2^
Output Concentration {^2^5^ 2s
40BHt Permeation Rate (CH^J^S^
ioaut. .Standaxd. Arfia.. iCHj) 232
^OE^t Sample Area t^i3i2§2
Output Concentration .tCH3]2S2
Oytp^t Concentration (CJi ) -S
3 2^2
"I" ""."_""""'.". "" - -.
__ .. _ —
. - -
;'_;_ ;__" ;""
. . . —
INPUT
DATA UNITS
mi/miti
Sue
"Hu
ffli/BJ n .
. ..na/oin
nQ/min
ntj/min
nq/min
i:
KEVS
1
1
Ig sci
IA
! R/S
IR/S
IR/S
R/S
R/S
R/S
R/S
1 R/S
IR/S
IR/S
IR/S
IR/S ||
1 II
iK/s'll
IR/S
IR/S II
IB/S
1
IR/S
IR/S
IR/S 11
IR/S
t
1
|h RTN
1
1 1
1 1
[ 1
1
1 1
1
1 1
1 i
1 1
1
1
OUTPUT
OtH UNITS
M/rc3
Pl'Bl
•mj
Wffi
U'J/n.'3
ppm
H9/.«3
PiJJD
Figure 16. HP-67 user instructions.
411
-------
STEP KEY ENTRY KEY CODE
COMMtNTS
STEP KEY ENTRY KEY CODE
COMMENTS
001
010
020
030
WO
050
0
SO
A
f I.BL A
3
4
0
0
R/S
X
R/S
X
STO 1
R/S
4
0
+
RCL 1
h x 4y
h 1/x
R/S
;
STO 2
B«
prj 7
X
R/S
R/S
X
R/S
2
4
9
9
.-.
R/S
RCL 2
X
R/S
!
R/S
X
R/S
2
5
8
4
T
R/S
R[~T. 03
X
R/S
'
R/S
1
Si
31 25 11
m_
04
00
00
00
81
84
71
84
71
33 01
84
04
06
00
61
34 01
35 52
81
35 62
84
81
33 02
84
14 02
71
84
81
84
71
84
02
04
09
09
81
84
14 02
71
84
ai
84
71
R.I
n?
nq
08
04
81
84
34 02
71
84
81
84
2
S2
3
In svui'lu Flow.mlt/min
In Sampling Time, se*
In barometer, "Hg
In Sample Temp, °f
InDil Gas flow m £/ni.n
In Perm RdLe COS , ng/nii
III Std Area COS
In Sample Area COS
3ut pg/m COS
3ut ppm COS
In PR (Cn3)2S 119/mil
In Std Area (CH,)2S
In Samp Area (CHj) 2S
Ou>t 119/81 (CH>)2S
Jut ppm (CH^^S
In PR (C-jHcJ^S, nq/nu
In Std Area (C2H5)2S
tn Samu Area (CoH^l-iS
REGIE
3 4
S3 *
C
"°l)
070
080
1
090
100
11
110
>TERS
5
SS
D
X
......K/S. .,
3
7
, .. .5
2
i
R/-S
Ijf'T. ?
*
i
R/S
X
R/S
^
4
1
8
R/S
h RTN
6
S6
71
H4
LU
._ . U2. - ..
_U6
02
ai.
H4
14 112
ftil
ai
Kd
71
84
03.
09
01
08
HI
an
IS 22
•
7
S7
i
>ut HU/BI* (("• H.)-S
2 b 2
Hut t>£'IH [C ,Hs) ,h
J ^ '!' ''"'"'"
In Std Area (0)3) ^s->
"
In Std Area (CH3)2S2
Out p9/m3 (CHj)2S2
!)ut ppm (CH3)2S2
8 9
sa S9
I
Figure 16 (cont'd). HP-67 program form
412
-------
SWRI PROJECT NO.
FUEL: CVS NO.
SAMPLE COLLECTION BY:
GENERAL COMMENTS:
TEST NO.
TUNNEL SIZE:
JTEST DATE:
DRIVER:
VEHICLE:
MILES:
CHEMICAL ANALYSIS BY:
CALCULATIONS BY:
Test No.
Driving Cycle
Sample Flow Rate, m£/min
Sampling Time, sec
B.P., "Hg
Temp., °F
Dilution Gas Flow, mS/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 (CK^) 2s
Sample Area (013)23
Sample Cone . (CH3) 25 , yg/m3
Sample Cone. (013)2$, ppm
Permeation Rate (C7He;)2S, ng/min
Standard Area (C2H$)2S
Sample Area (CjHs^S
Sample Cone. (€2X5) 2s' yg/m3
Sample Cone. (C2Hs)2S, ppm
Permeation Rate (013)952, ng/min
Standard Area (013)232
Sample Area (CH3)2S2
1 2 3456
FTP-1
130
504
29.80
75
580
667.5
18,514
30,200
17.5
0.00700
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
FTP-2
110
867
30.03
80
500
667.5
20,112
40,100
17.1
0.00683
1061
38,100
40,381
14.4
0.00558
445
7017
34,650
28.2
0.00751
133.5
2210
3120
2.42
0 00061"
FTP- 3
150
505
29.02
96
600
667.5
21,238
33,162
14.9
0.00598
1061
35,100
49,162
^B^|HHMflBfla^Ba^^^M^HA
21.3
0.00824
445
6844
32,111
29.9
0.00798
133.5
2763
6372
4.41
0.00113
SET-7
90
1397
29.25
85
650
667.5
16,542
20,122
6.29
0.00252
1061
41,000
38,142
7.65
0.00296
445
6015
7022
4.03
0.00107
133.5
1651
1561
0.978
0.00250
HFET
120
765
29.95
83
450
667.5
15,962
16,269
10.2
0.00406
1061
45,610
54,753
19.0
0.00736
445
7113
7914
7.39
0.00197
133.5
1814
1418
1.56
0.000391
NYCC
170
1200
29.50
89
575
667.5
23,146
32,641
5.08
0.00203
1061
35,122
41,611
^W^^^^^HVB^^^^
6.78
0.00262
445
7099
17,416
5.89
133.5
2917
2372
0.586
0.000149
Figure 17. Organic sulfide sample collection sheet,
413
-------
=17.5 yg/m
ppm COS = yg/m * density yg/m£
mol. wt. (COS) x loop
density yg/m = 24.04£
mol. wt. COS = 60.08 g/mole
60.08 g/mole x 1QQQ
density = 24^04 ft/mole ' = 24" Ug/m
^ —3 3 —3
ppm =17.5 yg/m * 2499 yg/m£ = 7.00 x 10 m*/m =7.00X10 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
into the above formulas. These calculations give the following concentra-
tions: (CH3)2S, 21.1 yg/m3 and 0.00816 ppm; (C2H5)2S, 32.5 Ug/m3and 0.00865
ppm; and (CH3)2S2, 1.86 yg/m3 and 0.000475 ppm.
NOTE: The values used in these calculations are picked from a range of tem-
peratures, 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-67 calculations give the same results. This was
confirmed for six sets of calculations.
LIST OF EQUIPMENT
The analysis for the organic sulfides is performed using a gas chromato-
graph equipped with a flame photometric detector. The gas chromatograph,
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 linearized
flame photometric detector (FPD) and subambient temperature pro-
grammer .
2. Soltec dual channel recorder, Model B-281, 1 mv recorder.
3. Hewlett-Packard Model 3354 GC computer system with remote teletype
printout.
4. Hewlett-Packard Model 1865A A/D converter.
5. Metronix Dynacalibrator Model 220-R for generation of organic
sulfide standards.
414
-------
6. Bendix valve oven.
7. Lindberg Furnace/Heavy Duty Model 55035.
8. Seiscor valve - gas sampling configuration
9. Seiscor valve - backflush configuration.
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. Brooks 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-AAA, SS 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.
415
-------
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.
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" 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, CH CHOHCH .
J j
6. Sodium bicarbonate.
7. Dry ice.
8. Zero nitrogen (permeation system) .
416
-------
REFERENCES
Spencer, C. P., Baumann, P., and Johnson, J. P., Anal. chein., Vol. 30, No. 9
pg. 1473, 1958. '
Sumner, S., Karrman, K. J., and Sunden, V., Microchim Acta, pg. 1144, 1955.
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.
Desty, 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. L. A., Div. of Analytical and Petroleum Chem-
istry, Symposium on Advances in Gas Chromatography, 132nd meeting, ACS, N.Y,
N. Y., September 1957.
Amberg, 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. P., Knight, P., Kelly, R. P., and Heath, M. T., Second Int'l.
Symposium on Gas Chromatography, Amsterdam, May 1958.
Adams, D. F. and Koppe, R. K., Tappi, Vol. 42, pg. 601, 1959.
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. Sci • ,, Vol. 8, pg. 151, 1970
Rummer, 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.
417
-------
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.
Elliot, 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.
418
-------
APPENDIX J
PHENOLS PROCEDURE
419
-------
PHENOL PROCEDURE
Phenols (phenol; salicylaldehyde; m-cresol/p-cresol; p-ethylphenol/
2-isopropylphenol/ 2,3-xylenol/3,5-xylenol/2,4,6-trimethylphenol; 2,3,5-
triroethylphenol; and 2,3,5,6,-tetramethylphenol) in automotive exhaust can
be sampled and quantitatively analyzed with a gas chromatograph (GC) equipped
with a flame ionization detector. Dilute exhaust is passed through two
Greenburg-Smith impingers in series, each containing 200 m£ of 1 N KOH
chilled in an ice bath. The contents of each impinger are acidified and
extracted with ethyl ether. The samples are partially concentrated, com-
bined and then further concentrated to about 1 m£. An internal standard
is added and the volume is adjusted to 2 m£. The final sample is analyzed
by the use of the GC and concentrations of individual phenols are determined
by comparison to external and internal standards. The minimum detection
limit is about 1 yg/m&.
SAMPLING SYSTEM
A schematic of the phenols sampling system is shown in Figure 1. As
seen in the illustration, exhaust from the automobile is first diluted by
the constant volume sample (CVS). The dilute exhaust entering the sampling
probe is then filtered by a heated (375°F) Pallflex filter of porosity
1-100 ym to remove particulate from the gas stream. Next, a Thomas sample
pump draws the exhaust through a heated sample line at about 0.8 ft^/min.
Both the filter and sampling line are heated to prevent phenol loss to
condensation. The sample pump then pulls the warm exhaust through two
impingers, each containing 200 m£ of 1 N KOH chilled to ice bath temperatures
(0-5°C). Wet exhaust exiting from the impingers passes through a molecular
sieve/silica gel dryer before flowing through the sample pump, flowmeter and
dry gas meter. The needle valve on the flowmeter controls flow through the
sampling system and the dry gas meter measures the volume of gas in cubic
feet that passes through the impingers. Gas temperature is measured by an
iron-constantan thermocouple and can be monitored by a digital readout.
Pictures of the phenol sampling cart are shown in Figure 2.
PROCEDURE
The flow schematic for the analysis of phenols is shown in Figure 3.
This diagram describes sample treatment from collection to analysis. Diesel
exhaust is first diluted in the constant volume sampler. Phenols present
in the diluted exhaust are captured in two glass Greenburg-Smith impingers
connected in series and chilled in an ice-water bath. Once collected, the
samples are quantitatively transferred to 250 m£ polyvinylchloride storage
bottles.
420
-------
ro
Sample
Probe
Greenburg
Smith
Impingers
Filter
Dilute
Exhaust
Gas Temperature
Digital Readout
Flowmeter
Sample
Pump
Dry
Gas
Meter
Figure 1. Phenols sample collection flow schematic.
-------
Heated
Sampling
Line
Flowmeter
Ice Water
Bath
Greenburg-Smith
Impinger
Dry Gas Meter
Sampling Pump
Figure 2. Phenols sampling system.
422
-------
CVS
Greenburg-
Smith
Impingers
Extraction
with Ether
Bubblers
1 and 2
Combined
Sample
Concentrated and
Internal Standard
Added
Sample Analysis
in GC with FID
Recorder
A/D Converter
Hewlett-Packard
3354
Computer System
Figure 3. Phenols analysis flow schematic.
423
-------
The entire workup procedure for phenols is carried out under a vented
hood to prevent ether vapors from escaping into the room. Explosions can
occur when handling ether, therefore,care needs to be taken not to heat
samples to dryness. When ready for processing, the sample is poured into
a 500 m& separatory funnel with 1 N KOH washings of the storage bottle.
Thirteen milliliters of 50% H2SO4 is carefully pipeted into the funnel con-
taining the sample and the flask is swirled and very gently shaken with
venting until thoroughly mixed. Acidity is checked with litmus paper. Next,
200 m£ of ethyl ether is added, again with swirling and gentle shaking and
venting. When venting is no longer necessary the separatory funnel is shaken
for two minutes and the two layers are allowed to separate. The bottom
aqueous layer is drawn off into a second 500 m& separatory funnel and set
aside. Anyhdrous Na2SC>4 (9.4 g) is added to the phenolic ether mixture
the first funnel with swirling and very gentle shaking and venting until
venting is not needed. The contents are shaken for two minutes to remove
traces of water. The dry ether and phenol mixture are transferred to a Kuderna
Danish concentrator with washings. A boiling chip is added and the solvent
volume is reduced to about one-fourth by heating in a 45°C water bath. The
Kuderna concentrator is then set aside.
While the first ether portion is being heated, a second extraction is
performed on the aqueous layer in the second separatory funnel. First, 100
mH of ether is added to the second funnel with swirling and gentle shaking
and venting. When venting is unnecessary, the flask is shaken for two minutes.
After the two layers have separated the bottom aqueous layer is drawn off and
discarded. Anhydrous Na2SC>4 (4.7 g) is added to the ether and phenols mix-
ture and the funnel is swirled, gently shaken and vented. The flask is sha-
ken for two minutes and the contents are then transferred with washings to
the Kuderna concentrator containing the first ether extraction. A second
boiling chip is added and the sample is concentrated to approximately 5 m£
by means of the water bath. The concentrator is cooled to room temperature
and the remaining sample is transferred to a 10 m£ beaker for the final eva-
poration.
The last drying step is carried out is a desiccator box modified for
nitrogen flow. A tray of molecular sieve and silica gel absorbs moisture
that condenses on the beaker during the drying process. The nitrogen flow is
directed into the 10 m£ beakers containing the phenol sample by a six position
manifold,, The samples are placed under one of the curved needles on the mani-
fold and the solvent is evaporated to about 1 m£ by a stream of dry nitrogen.
When the sample has warmed to room temperature (about 15 minutes) the con-
centrate is transferred to a 2 m£ volumetric flask with small ether washings.
The sample is then spiked with 100 y£ of 300 yg/m£ o-chlorophenol, the inter-
nal standard., The volume is adjusted to 2 m£ with ether and the sample is
labeled, sealed with Teflon tape and refrigerated until samples are ready for
analysis. Several pictures of the equipment used in the workup of phenol
exhaust samples are shown in Figure 4.
Phenol samples and standards are analyzed by a Perkin-Elmer 3920B gas
chromatograph (GC) equipped with a flame ionization detector. The column
used to separate phenols is a 6' x 1/8" Teflon column packed with 10%
OS-138/H3PO4/SP-1200 on 100/120 mesh Chromosorb W AW. The carrier gas, zero
nitrogen, flows through the column at 50 m£/min. The temperature of the
424
-------
.
Extraction Apparatus
Sample Drying Chamber
foruorkup of
425
-------
column is programmed from 70-170°C at a rate of 4°C per minute with an initial
hold at 70 °C for 2 minutes. The purpose of temperature programming is to
prevent the solvent peak from obscuring the phenol peaks and to allow better
separation of phenols that elute at higher temperatures. The temperature of
the injector and interface is maintained at 200°C. Samples and standards
are usually analyzed at attention of XI X8 or XI X16 for 30 minutes.
Injection volume is 1 y£. A picture of the analysis system is shown in
Figure 5.
The external standard is injected first and response factors are de-
termined from the concentrations and areas of each phenol. Then samples are
injected and analyzed. Calculations are performed using sampling information
and data obtained from the GC analyses. A chromatogram of the external
standard is shown in Figure 6 and a chromatogram of a diesel exhaust sample
is shown in Figure 7.
CALCULATIONS
The information obtained from the sampling system and the GC analysis
of phenol samples is used to calculate the concentration of phenols in ex-
haust. The mathematical steps were programmed into a Hewlett-Packard 67
calculator for rapid data turnover. A copy of this program is shown in
Figure 8. The concentration of individual phenols is determined by comparing
the area of each phenol to the area of an internal "standard, JO-chlorophenol.
Different phenols do not give the same response to the flame ionization
detector. Therefore, a correction needs to be incorporated into the calcu-
lation to account for this difference in response. This correction, termed
the response factor, F, is determined by analyzing the external standard
each day before sample analysis. The area per concentration unit of each
individual phenol is compared to the per concentration unit of o-chlorophenol,
the phenol used as the internal standard.
C A
Response factor (F) = -£*• x ocpx
A C
px ocpx
where C = concentration of individual phenol in external standard,
px yg/m£
A = area produced by individual phenol in external standard,
PX
counts
A = area produced by o-chlorophenol in external standard,
counts
C = concentration of o-chlorophenol in external standard,
oqpx
Response factors provide the means of calculating the concentration of the
various phenols in the exhaust sample relative to only one phenol, the in-
ternal standard.
The concentration of any particular phenol in the exhaust sample in
yg/m£ is :
426
-------
p>
[\J
- a
Recorder A/D Converter
Gas
Chromatograph
Figure 5. Phenols analytical system.
-------
ro
co
o~chlorophencu
pnenoi
i athvde
r—ere sol an;'- j.—ere sol
r—ethylf/nenol i-ipopropy
i , 3-xylerioI, 3,5-xvienol ,
,6-trimt'thylphencI
2,3,5~trimethylpheno
7. 2,3,5,6-tetramethylpheno]
Sample External- Standarc
Initrument F -F ?9I; "'F Operato Hal Baylor
1
Tf-f Ion _ Type
id Phase
Support
nCerner
Packed with 1
100/i
Run ISO • -
_ _cc/min _ Nitro
Hotametei Readme
70 °CISO«or 2 min..proBto
/mm. Hrtd tot 3
held for ~ mm (olher I
Inlut
170 "C
.. Prog 10 _^__ C at - /mm
200 UC. Glass lined
Petectof2QO °C FID TVpe lotherl ~
Hyd PIIQ - Rotemeter HdQ- _ ~
PMO Rotemeter RdB
Rotameter Rdg
n/mm epeed 1 mV.F.S
ul indicated
injection -
Sempling Device
20
15 10
Retention time, minutes
Figure 6. Typical phenols external standard.
-------
ro
25
20 15 10
Retention time, minutes
Figure 7. Typical diesel-CVS exhaust sample.
-------
UMT Insiriiciions
PHENOLS
12-17-79
MAP
HP-6 7
STEP
Ol
"2
°3
1
2
3
5
6
7
8
9
lu
il
12
13
14
15
lib
17
1H
ly
2i>
21
22
23
24
25__
26
-
INSTRUCTIONS
Switch to on, switch to run
Peed card in from right to left
Set decimal place
Input - Sample volume
Input - Baronetric pressure
Input - Sample temperature
Input - Cone, of internal standard
Input - Area of internal standard
Input - Phenol area
Input - F (phenol)
Output - Phenol Cone. .
Input - Salicylaldehyde area
Input - P(salicylaldehyde)
Output - Salicylaldehyde cone.
Input - m-and p-cresols area
Input - P{m-and p-cresols)
Output - m-and p-cresols cone.
Input - p-ethylphenol , 2-isopropylphenol,
2,3-and 3,5-xylenols, 2,4 ,6-triroethylphenol are
Input - F (p-ethylphenol, 2-isopropylphenol ,
2,3-and 3,5-xylehols, 2,4/6-trimethylphenol
Output - p-ethylphenol, 2-isopropylphenolr
2,3-and 3,5-xylenols, 2,4,6,-trimethylphenol co
Input - 2, 3,5,-trimethylphenol area
Input - F(2,3,5,-triirethylphenol)
Output - 2, 3,5,-trimethylphenol cone.
Input - 2,3,5,6,-tetramethylphenol area
Input - F(2,3,5,6,-tetrajnethylphenolj
output - 2,3,5,6,-tetramethylphenol eonc.
Input - additional phenol area
Input - F(additional phenol)
Output - additional phenol eonc.
...
— — — ^— ^— — ' — — .
INPUT
DATA/UNITS
TitTIZ
"Hg
°F
pg/m£
counts
counts
counts
counts
counts
tc.
counts
counts
counts
.._.
KEYS
IDSP Q
A
|h/S |
IR/S |.|
|R/S
l«/s
IR/S
IR/S
IR/S
IR/S
IR/S
IR/S
IB/S
IB/S
IH/S
IR/S
IR/S
IR/S
IR/S
IR/S
[R/S
1 h OTH
1
II 1
OUTPUT
DATA/UNITS
Mg/m3
M9/J53.
IHH/5 ._
• i~
lig/m
K^/SC
M'J/m3
tiu/m3
Figure 8. HP-67 user instructions,
430
-------
STEP WtM< KEY CODt COMMENTS STEP KEY ENTflY KtY COOL MUU.MT,
-jr:
._ ^_ .,_
. _
1 U1L.A..
_ „
j
c.
...2
-X-
2.
M
2
"-'" . I 0
^
1 I K/s._._
r x
ujO
Ma
C*>
STO 1
4
6
0
HC'L 1
ll 1/X
X
STO 2
X
R/S
X
R/S
R/S
X
RCL 2
X
R/S
R/S
X
FVI. 2
X
R/S
X
RCL 2
3] 25 11
.3 US
in .. .
oa
21.....
02
_ BJL_ ..
02
81
-siQ
81
ai
71
3J 0)
04
.-.-. UU
aJ.
til
21 . .
M.1
81
_.il_02
84
H4
7 j
S4
71
34 02
71
84
84
71
_.24.._02
71
fl-1 *»_
«4 K
"x^
ni.ut-:;ami.. IL
VoJ ft^
Input h.P. "H,j
lli['US -. aluj. Iclllp "F
Jn[^uf-i oiu;
|i'
ll I-Ul - JU'd
InpuL-ph al
Input.-ph f
Outpljl -ph
iui. -.sal von
Input -pup, 2
i. I,S-X,2,4,
iJ,5-X,2,4,
0 1 2 J
SO SI S2 S3
A
B C
,... -[^
/ni«
Q<, I-
JOI5C
ire a
•
iri|»b',2.3-
6 Imp F
REGl*
4 "
S4
IUHJ
ud'i
— ._
1 00
a(-
X
.r_vs"..""
X
Rc-'l." 2
X ~
- K/S"
R/S
X
' R/s'J
X
- !"-' R'!'N
^
.
71
a-l *^.
H-J * ~~*
71 "V
1-1 i).1
'1
"""71
M OJ
)4 1)2
- -— • --
JS 22
...
.
_
.„
TEHS
5 6 '
S5 So ^ •
D
cone
'til-.' , 1 , ^,-v Hill. Jli: j
In-.;, 1,',,-ttmp F
^HI -J , j/i-t i inp cone
hi-.1 , l.'j ,t.-ti.inp acua
In-J, l,'j,,,-i,,ni|i K
| 'til-Ill '1 J I 1. ,1
-iii-a-iJi Liuual
it
_
— L
Figure 8 (Cont'd). HP-67 program form.
431
-------
C x A x F
ocps sa
A
ocps
where A = area produced by the phenol in the sample, counts
Asa = area produced by o-chlorophenol in the sample, counts
cocps _ concentration of o-chlorophenol in the sample, yg/mJl
ocps
The extraction process involves concentrating the phenols from 200 mfi, in the
impingers to 2 m£ in the final sample. The weight in Ug of the phenol in
the impingers is given by
C x A x F x V
ocps sa sa
where V = final sample volume, m&.
^t 21
Equation 1
To determine the concentration of phenols in dilute exhaust the volume
of exhaust passing through the impingers has to be measured. This volume,
measured at ambient conditions, is corrected to standard conditions (68°F
and 29.92" Hg) .
P xV Px.xV^
am am _ stp _ stp
T + 460 ~ T
am stp
where P = barometric pressure at ambient, "Hg
V = volume at ambient, ft3
am
T = temperature at ambient, °F
P = standard barometric pressure, 29.92 "Hg
Vs p = volume at 68°F and 29.92 "Hg
stp
T . = standard temperature, 68°F + 460 = 528°R.
stp
Rearranging and converting ft3 to m3,
P T ^ V
_ am stp x _ am
stp ~ P X(T + 460) " 35.31
stp am
Equation 2
Dividing the weight of the phenol captured in the impingers in Equation
1 by the volume of dilute exhaust passed through the bubblers in Equation 2
yields the concentration of the phenol in dilute exhaust in yg/m3.
C
- — — x A x F x V
A sa sa
ocps _ _
P T ^ V
am stp am
Pstp 35.31
432
-------
C Ug/m& x A counts x F x V mS,
ocps , _ sa, _ sa,
A counts
ocps ,
29.92 "Hg x(T °F + 460) x 35.31 ft3/m3
^ am,
X P "Hg x 528°R x V ft3
am, am,
Equation 3
Equation 3 yields the concentration of phenols that the Hewlett-Packard 3354
computer identifies as separate peaks. However, two peaks found in several
exhaust samples may contain more than one phenol, in which case an average
response factor is used to calculate the concentration. For example, m-
cresol and p-cresol elute at the same time. The response factor is determined
by comparing the area of the cresols peak per total concentration unit of
the two cresols to the area per concentration unit of o-chlorophenol. this
method is the same used for only one phenol except that the concentrations
of all the phenols under the peak are added together. In the case of multi-
ple phenols under a single peak, the calculated concentration represents the
total concentration of all the phenols eluting at that retention time. One
additional peak represents more than one phenol. The phenols eluting under
the same peak are p-ethylphenol , 2-isopropylphenol, 2,3-dimethylphenol, 3,5-
dimethylphenol and 2,4,6-trimethylphenol. All of these compounds are not
necessarily found in automotive exhaust, but since separation of them
is not possible using the specified equipment and ^^\^^'^
can only be assumed that they are all present individual ph enols ^ at have
been recovered from exhaust are phenol, salicylaldehyde , 2,3,5,-tnmethyl
phenol and 2,3,5,6,-tetramethylphenol.
Sample Calculation
trations from raw data. Experimental values are
the calculations are performed to give answers in
Example 1
^4-1-7 dfn ft3 of dilute exhaust is
During an FTPh driving cycle, ™™f*JL %l\ KOH. ^ient tern-
passed through two impingers containing 200 mx, o ^ ^ ^ respectively.
perature and pressure were ^ted ^be .^^ standard was
Before final concentration 100 y£ of the 200 ^/ ^ standard, at a
added and the volume was adjusted to 2 n*. i duced an area response
concentration of 10 pg/ml in the ^™\S^'Jl°^ °phenol are listed
of 2159 counts. The response factor and area ot eac
on the following page.
433
-------
SWRI I-ROOECT NO.
FUEL: CVS NO.
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
Final Sample Vol. (mi)
Cone. Internal Std. (ug/mH)
Area Internal Std.
Area Phenol
F Phenol
Ug/m Phenol
Area Salicylaldehyde
F Salicylaldehyde
Ug/m3 Salicylaldehyde
Area m-and p-cresols
F m- and p-cresols
Ug/m3 m- and p-cresols
. * . -- .
-X, 2,4,6-trrap
F pep, 2ispp, 2,3-S 3,5-X,
2,4,6-trmp
Ug/m3 pep, 2ispp, 2,3-53, 5
-Xr 2r4rfiJ.-1-nnp
Area 2,3,5-trmp
F 2,3,5-trmp
yg/mj 2,3,5-trmp
Area 2,3,5 ,6, -temp
F 2, 3, 5, 6, -temp
Ug/m3 2, 3, 5, 6, -temp
Area additional phenol
F Additional phenol
yg/m additional phenol
1
FTPc
17.582
29.27
77
2
10
2400
1373
0.8354
20
1457
0.5079
13
1023
0.4232
8
412
0.7778
6
876
0.5010
8
1167
0.6342
13
--
-
-
2
FTPh •
17.483
29.27
79
2
10
2159
2209
0.8354
36
6401
0.5140
64
1227
0.5105
12
2029
0.7864
31
1443
0.4766
13
2382
0.6140
29
-
-
- -
3
SET 7
18.400
29.27
79
2
10
2262
1070
0.8379
16
2078
0.4980
18
986
0.4657
3
1569
0.7680
21
2785
0.5284
26
1771
0.5865
18
-
-
-
4
FET
10.435
29.27
78
2
10
2109
879
0.3348
25
6006
0.5136
103
1500
0.5033
25
1283
0.8107
35
2235
0.4546
34
4067
0.6676
91
-
-
-
5
NYCC
8.126
29.27
80
2
10
2387
500
0.8299
16
2868
0.5200
57
3274
0.4451
55
2433
0.7776
72
2567
0.5160
50
3852
0.5744
84
-
-
-
6
85 kph
18.567
29.27
80
2
10
2182
758
0.8350
12
3543
0.5092
33
2921
0.4079
22
4155
0.8091
61
2488
0.5518
25
981
0.6020
11
-
-
—
Figure 9. Phenol data sheet.
434
-------
Phenol p
•— Area
Pheno1 0.8345 2209
Salicylaldehyde 0.5140 6401
m-cresol and p-cresol 0.5105 1227
p-ethylphenol,
2-isopropylphenol,
2,3- and 3,5-xylenol,
2,4,6-trimethylphenol 0.7864 2029
2,3,5-trimethylphenol 0.4766 1443
2,3,5,6-tetramethylphenol 0.6140 2382
The appropriate values are plugged into Equation 3 for each phenol.
Phenol
10 yg/ml, x 2209 counts x 0.8345 x2.0m£
2159 counts
29.92 "Hg x (79°F + 460) x 35.31 ft3/m3
29.27 "Hg x 528°R x 17.483 ft3
= 36 yg/m phenol
Salicylaldehyde
10 yg/m£ x 6401 counts x 0.5140 x 2.0 n&
2159 counts
29.92 "Hg x (79°F + 460) x 35.31 ft3/m3
X 29.27 "Hg x 528°R x 17.483 ft3
= 64 yg/m3 Salicylaldehyde
m-cresol and p-cresol = 12 yg/m3
E-etfaylphenol, 2-isopropylphenol, 2,3- and 3,5-xylenol and2,4,6-triinethylphenol
10 yg/m£ x 2029 counts x 0.7864 x 2.0 m&
2159 counts
x
29.92 "Hg x (79°F + 460) x 35.31 ft3/m3
29.27 "Hg x 528°R x 17.483 ft-*
= 31 yg/m3 p-ethylphenol, 2-isopropylphenol, 2,3- and 3,5-xylenol,
2,4,6-trimethylphenol
1/3,5-trimethylphenol = 13 yg/m3,
2,3,5,6-tetramethylphenol = 29 yg/m3
435
-------
Example 2
Suppose that during the FTPc driving cycle 17.582 ft of dilute exhaust
measured at 29.27 "Hg and 77°F passed through two impingers. The sample
was processed and analyzed following the phenols procedure. The 10 yg/m£
internal standard gave a responce of 2400 counts. The response factors and
areas of the other phenols are listed below.
Phenol F
Phenol 0.8354
Salicyaldehyde 0.5079
m-cresol and
p-cresol 0.4232
p-ethylphenol,
2-isopropylphenol,
2,3-and 3,5-xylenol
and 2,4,6-trimethyl-
phenol 0.7778
2,3,5-trimethylphenol 0.5010
2,3,5,6-tetramethylphenol 0.6342
Phenol
10 yg/m& x 1373 counts x 0.8354 x 2.0 m£
2400 counts
x 29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
2g>27 ,,Hg x 528oR x 17.582
= 20 yg/m3 phenol
Salicylaldehyde
x
10 yg/m£ x 1457 counts x 0.5Q79 x 2.0 m&
2400 counts
29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
29.27 "Hg x 528°R x 17.582
= 13 yg/m3 salicylaldehyde
m-cresol and p-cresol
10 yg/m£ x 1023 counts x 0.4232 x 2.0 m
2400 counts
29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
29.27 "Hg x 528°R x 17.582 ft3
= 8 yg/m3 m-cresol and p-cresol
Area
1373
1457
1023
412
876
1167
436
-------
?-ethylphenol, 2-isopropylphenol, 2,3-andJ^-xylenoj^and 2 ,4,6,-trimethylphenol
10 yg/m£ x 412 counts x 0.7778 x 2.0 m£
2400 counts
x 29.92 "Hg x (77°F + 460) x 35.31 ft3/m3
29.27 "Hg x 528°R x 17.572 ft3
= 6 yg/m3 p-ethylphenol, 2-isopropylphenol, 2,3-and 3,5-xylenol
2,4,6,-trimethylphenol
2,3,5-trimethylphenol = 8 yg/m3
2,3,5,6-tetramethylphenol = 13 yg/m3
LIST OF EQUIPMENT
The equipment needed for collection, workup and analysis of phenols in
exhaust is listed below in separate sections. The manufacturer, catalog or
model number and description are given for each entry.
Sampling
1. Greenburg-Smith glass impingers, Houston Glass Fabricating, Catalog
#310610-0028, arm joints 28/15, bottle joint 45/50.
2. Ground glass socket joint with arm modified to 5/16", Houston
Glass Fabricating, Catalog #285045, 28/12 socket.
3. Ground glass ball joint with arm modified to 5/16", Houston Glass
Fabricating, Catalog # 285040, 28/12 ball.
4. L-shaped glass connecting adapter, Houston Glass Fabricating,
Catalog #015639, male and female size 28/12.
5. U-shaped glass connecting adpater, Houston Glass Fabricating,
Catalog #160719 size 28/15 male-male socket joint, 2.25" center
to center length.
6. Thomas ball and socket joint clamp, Houston Glass Fabricating,
Catalog # 285100 size 28.
7. Class A, 2000 m£ volumetric flask.
8. Flowmeters, Brooks Instrument Division, Kynar, Sho-Rate "150" with
R-6-15-B metering tube, stainless steel ball, 1-92 CFH range,
graduated 1-100.
9. Dry gas meter, American Singer Corporation, Type Al-120, 60 CFH
capacity.
437
-------
10. Sample pump, Thomas Model #727CA39, 1 ft3/min free flow capacity.
11. Drying tube, Analabs Inc., Catalog #HGC-146, 6" long.
12. Teflon tubing, United States Plastic Corporation, 5/16" OD x 1/8"
ID and 3/8" DD x 1/4" ID.
13. Miscellaneous Teflon nuts, ferrules, unions/ tees, clamps and
connectors/ etc.
14. Miscellaneous electrical switches, lights, wiring, etc.
15. Miscellaneous Swagelok fittings.
16. Athena temperature controller, Technical Heaters Incorporated,
Model 6000, 100-600°F range, 110 volts.
17. Heated sample line. Technical Heaters Incorporated, Catalog
#LP-212-8-5, 5' length, 13/32" hose with 1/2" tube end stainless
steel fittings.
18. Pallflex Fiberfilm filters, Pallflex Products Corporation,
Catalog #T60A20, 70 mm diameter.
19. 250 m£ polyvinylchloride sample storage bottles, Nalgene Labware,
Catalog #2000-0008.
20. Iron/constantan type J single thermocouple, Thermo Sensors Corp.
Workup
1. Pear shaped 500 mJl separatory funnel, Houston Glass Fabricating,
Catalog #260145, with Teflon stopcock.
2. Class A, 3 and 10 m£ volumetric pipets.
3. Safety bulb for pipetting, Markson Science Supplies, Catalog #E-8074.
4. Miscellaneous glass beakers.
5. 250 mJl glass graduated cylinder.
6. Basic indicating litmus paper.
7. Ring stands.
8. Kuderna Danish Concentrator, Ace Glass Incorporated, Catalog
#6708-03 and 6708-35*
9. Boileezers boiling chips, Fisher Scientific Company, Catalog
#B-365.
438
-------
10. Hot plate with heat control.
11. 2000 m£ beaker for water bath.
12. 10 m£ glass beakers with pouring spout.
13. Disposable transfer pipets, Curtin Matheson Scientific Incorporated,
14. Class A, 2 m£ volumetric flasks with hexagonal base, Fisher
Scientific Company, Catalog #20814B.
15. Boekel desiccating cabinet modified for nitrogen flow, Curtin
Matheson Scientific Incorporated, Catalog #076-190.
16. 6 position gas manifold for sample concentrating, Alltech Associ-
ates, Catalog #9555.
17.. Zero grade nitrogen
18. Teflon Lab-Tape, Fisher Scientific Company, Catalog #14-831-300A,
13 mm width.
Analysis
1. Perkin-Elmer Model 3920B gas chromatograph equipped with a flame
ionization detector.
2. Soltec Model B-281 1 mv recorder.
3. Hewlett-Packard Model 3354 GC computer system with remote teletype
printout.
4. Hewlett-Packard Model 1865A A/D Converter.
5. 5 yJl Hamilton liquid syringe, Alltech Associates, Catalog #N-75.
List of Reagents
All compounds used in sample acquisition and workup are listed in this
section. Formula weights, grade of purity, manufacturer and catalog number
are listed for each reagent.
1. Potassium hydroxide, KOH, formula weight = 56.11, Mallinckrodt, 85%
analytical reagent grade pellets, Catalog #6984.
2. Ethyl ether, anhydrous, (CH3CH2)20, formula weight =74.12,
Mallinckrodt, ACS analytical reagent grade, Catalog #0848.
439
-------
3. Sodium sulfate, anhydrous, Na2S04, formula weight = 142.04,
Mallinckrodt, ACS analytical reagent grade, granular, Catalog
#8024.
4. Sulfuric acid, H2S04, formula weight = 98.08, Mallinckrodt, ACS
analytical reagent grade, Catalog #2876.
5. o'-chlorophenol, Cl (C6H4OH) , formula weight = 128.56, Eastman Kodak
Company, analytical reagent grade, Catalog #1087.
6. Phenol, C6H5OH, formula weight = 93.11, Mallinckrodt, ACS analytical
reagent grade, loose crystals, Catalog #0028.
7. Salicylaldehyde, 2-HOCgI^CHO, formula weight=122.13, Eastman Kodak
Company, analytical, reagent grade. Catalog #225.
8. m-cresol, CH3 C6H4OH, formula weight = 108.14, Aldrich Chemical
Company, 99+% Gold Label, Catalog #C8,572-7.
9. p-cresol, CH3C6H4OH, formula weight = 108.14, Aldrich Chemical
Company, 99+% Gold Label, Catalog #C8,575-1.
10. p-ethylphenol (4-ethylphenol), C2H5C6H4OH, formula weight = 122.17,
Aldrich Chemical Company, 97%, Catalog #E4,420-5.
11. 2-isopropylphenol, (CH3)2 CHC6H4OH, formula weight = 136.19,
Aldrich Chemical Company, 97%, Catalog #12,952-6.
12. 2,3-dimethylphenol (2,3-xylenol), (CH3)2 C6H3OH, formula weight=122.17,
Aldrich Chemical Company, 97%, Catalog #D17,400-9.
13. 3,5-dimethylphenol (3,5-xylenol), (CH3)2 C6H3OH, formula-weighty
122.17, Aldrich Chemical Company, 99.9+% zone refined,
Catalog #15,085-1.
14. 2,4,6-trimethylphenol, (CH3)3C5H2OH, formula weight = 136.19,
Aldrich Chemical Company, 99%, Catalog #T7,900-6.
15. 2,3,5-trimethylphenol, (CH3)3C5H2OH, formula weight = 136.19,
Aldrich Chemical Company, catalog #T7,860-:3.
16. 2,3,5,6-tetramethylphenol, (CH3)4 CgHOH, formula weight = 150.22,
Aldrich Chemical Company, 90+%, Catalog #17,877-2.
Preparation of Reagents
Absorbing solution, 1 N KOH
132.02g of KOH pellets are dissolved in deionized water in a 2000 ml
volumetric flask. When the solution cools to room temperature the volume is
adjusted to 2000 ml.
440
-------
50% H2SO4
Concentrated H2SO4 (500 mfc) is slowly added to 500 m£ of chilled
deionized water with swirling and shaking. The solution is stored in a
glass bottle.
Internal Standard - 300 yg/m£ o-chlorophenol
Approximately O.OSOOg of o-chlorophenol is added to a 100 ml volumetric
flask and filled to volume with ethyl ether. This solution is sealed with
Teflon tape and refrigerated.
External standard stock solution
The external standard is prepared by blending the following phenols in
a 1000 m& volumetric flask and diluting with CH2C12. This solution is
sealed with Teflon tape and refrigerated.
Wt. (mg)
Phenol 150
Salicylaldehyde 300
m-cresol 150
p-cresol 50
p-ethylphenol 100
2-isopropylphenol 150
2,3-xylenol 100
3,5-xylenol 50
2,4,6-trimethylphenol 100
2,3,5-trimethylphenol 100
2,3,5,6-tetramethylphenol 200
o-chlorophenol 300
Dilute external standard
10 m£ of the stock external standard to diluted to 100 m£ with CH2C12.
This solution is sealed with Teflon tape.
441
-------
REFERENCES
Brown, R. A., Searl, T. D., King Jr., W. H., Dietz, W. A. and Kelliher,
J. M.., Rapid Methods of Analysis for Trace Quantities of Polynuclear
Aromatic Hydrocarbons and Phenols in Automobile Exhaust, Gasoline and
Crankcase Oil, Final Report for CRC - APRAC Project CAPE-12-68, Esso
Research and Engineering Company, Linden, New Jersey, 1973.
Carter, M. J. and Huston, M. T., Environ., Sci. Tech., Vol. 12, pg. 309, 1978.
Preston, S. T., A Guide to the Analysis of Phenols by Gas Chromatography,
PolyScience Corporation, Niles, Illinois, 1966.
Bartie. K. D., Elstub, J., Novotny, M. and Robinson, R. J., Journal of
Chromatography, Vol. 135, pg. 351, 1977.
Ma, T. S. and Speigel, D., Microchemical Journal, Vol. 10, pg. 61, 1966.
Sheffer, H. E., Perry, R. L., Thimineur, R. J., Adams, B. T., Simonian,
J. L., Zutty, N. L. and Clendenning, R. A., Ind. Eng. Chem. Prod. Res.
Develop., Vol. 10, pg. 362, 1971.
Yrjanheikki, Erkki, Am. Ind. Hyg. Assoc. J., Vol. 39, pg. 326, 1978.
Stanley, T. W., Sawicki, E., Johnson, H. and Pfaff, J. D., Mikrochim, Acta,
Vol. 1, pg. 48, 1965.
Umbreit, G. and Houtman, R. L., Journal of Pharmaceutical Sciences, Vol. 56,
pg. 349, 1967.
Lawrence, J. F., Journal of Chromatographic Science, Vol. 17, pg. 147, 1979.
Wolkoff, A. W. and Larose, R. H., Journal of Chromatography, Vol. 99, pg. 731,
1974.
*
Afghan, B. K., Belliveau, P. E., Larose, R. H. and Ryan, J. F., Analytica
Chimica Acta, Vol. 71, pg. 355, 1974.
Mindrup, Jr., R., Industrial Research/Development, pg. 79, August, 1978.
Eichelberger, J. W., Dressman, R. C. and Longbottom, J. E., Environmental
Science and Technology, Vol. 4, pg. 576, 1970.
Barber, E. D., Sawicki, E. and McPherson, S. P., Analytical Chemistry,
Vol. 36, pg. 2442, 1964.
Heenan, M. P. and McCallum, N. K., Journal of Chromatographic Science,
Vol. 12, pg. 89, 1974.
Grouse, R. H., Garner, J. W. and O'Neill, H. J., J. of G.C., pg. 18,
February, 1963.
442
-------
Kawahara, F. K., Analytical Chemistry, Vol. 40, pg. 1009, 1968.
Kushnir, I., Barr, P. A. and Chortyk, O. T., Analytical Chemistry Vol 42
pg. 1619, 1970. *' *
Spears, A. W., Analytical Chemistry, Vol. 35, pg. 320, 1963.
Chriswell, C. D., Chang, R. C. and Fritz, J. S., Analytical Chemistry
Vol. 47, pg. 1325, 1975.
Braithwaite, B. and Penketh, G., Analytical Chemistry, Vol. 36, pg. 185, 1964,
Shulgin, A., Analytical Chemistry, Vol. 36, pg. 920, 1964.
Argauer, R. J., Analytical Chemistry, Vol. 40, pg. 122, 1968.
Landault, C. and Guiochen, G., Analytical Chemistry, Vol. 39, pg. 713, 1967.
Fisher, G. E. and Neerman, J. C., I and EC Product Research and Development,
Vol. 5, pg. 288, 1966.
Cohen, I. C. Norcup, J., Ruzicka, H. A. and Wheals, B. B., J. Chromatog.,
Vol. 44, pg. 251, 1969-
Bark, L. S. and Clarke, K. F., J. Chromatog., Vol. 48, pg. 418, 1970.
Di Corcia, A., Journal of Chromatography, Vol. 80, pg. 69, 1973.
Seiber, J. N., Crosby, D. G. Fouda, H. and Soderquist, C. J., J. Chromatogr.,
Vol. 73, pg. 89, 1972.
443
-------
APPENDIX K
SULFATE PROCEDURE
444
-------
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 SO3 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 following aerosol stream. Particulate sulfate salts are collected as
well.
Sulfuric acid on the filter is converted to ammonium sulfate by expo-
sure 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 chormatograph 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
in samples where the sulfates can be leached out with water or aqueous IPA
445
-------
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 Sulfide 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 ygs/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 SO4= 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/mJl. 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 m£ injection of
sulfate samples at trace concentration levels (0.01 - 0.1 ygs/m&.
3.3 Precision better than 3% at 0.5 yg/m£ SO^ 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
446
-------
4.1.1.1 High pressure liquid chromatoqraph pump (LP). The pump must be
capable of delivering liquids at flow rates of at least 3 mVmin at
pressures as high as 1200 psi. Liquid pumps capable of delivering
pulseless and constant liquid flow are recommended for good quanti-
tation. Most HPLC pumps in the market are adequate.
4.1.1.2 UV detector (6) equipped with low dead volume (8 y£) flow-through
cell and a grating, prism or appropriate interference filter to iso-
late 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 two position, six port, high pressure, low dead volume sample in-
jection valve (SV) . This must be equipped with interchangeable
external 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. Dual 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+), 4 mm I.D. by 1/4 inch
O.D. by 6 inches long stainless steel column packed with 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 I.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 (BCA) , 4 mm I.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) .
447
-------
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 LP through a column of strong cation exchange resin in silver
form, CX-Ag"^7 then through a column of strong cation exchange resin in hydro-
gen form, CX-H+, then through a reactor column of barium chloranilate, BCA,
and finally through a flow-through cell of a UV detector, 13, 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 sul-
furic acid, and the BCA reacts with the sulfate to form barium sulfate pre-
cipitate and a soluble UV absorbing dye, chloranilic acid and its ions. Back-
ground 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 13 and on to waste.
After loop L_ is loaded with sample, injection valve SV switches to in-
ject 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.
448
-------
6. Apparatus
6.1 Pipette, volumetric, 1, 2, 4, 5, 8, 25, 50, 100 m£
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 Centri fuge
6.9 Magnetic mixer
6.10 Magnetic bars
6.11 Graduated cylinders
6.12 Automatic dispenser pipet, 5, 10, 20 mH (optional)
6.13 Automatic burette, 10 m£ (motor driven, optional)
6.14 Ammoniation chamber (Figure 5)
7. Reagents
^•1 Isopropyl alcohol (IPA), spectro quality grade or equivalent
7.2 Water, doubly deionized, distilled.
'•^ 60% IPA. Add 4 parts water to 6 parts IPA by volume. Store in tightly
capped bottles.
7.4 Barium chlo rani late, 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 mi concentrated hydrochloric acid to
6.0 mH of deionized water.
7.7 Ammonium sulfate, primary standard
449
-------
7.8 Silver nitrate (IN). Dissolve 17 grams silver nitrate in deionized
water and make up to 100 m£. 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 com-
pletely filledo 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 stain-
less 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
m&/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 vol-
ume in column A. Deactivate the HPLC pump, disconnect the composite
column from the pump, then column A from the composite column. Con-
nect 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 ex-
change resin, 100 - 200 mesh, to 160 m£ of 4 N HC1 in a 250 m&
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 fritt 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.
450
-------
Fill funnel with deionized water and turn vacuum slowly so that the
column IB completely filled with water. Add enough water so that
water level is in the funnel cone; stop the vacuum and add the
slurry of freshly washed resin (H+ form). Let the resin settle by
gravity until the resin top is halfway in the funnel 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 1 N AgNO3 solution
to the other half of the washed cation exchange resin, hydrogen form,
in a 150 m£ Erlenmeyer flask. Stir with a glass rod, cover the
the flask with aluminum foil and soak the resin overnight.
Decant the AgNO3 solution into a waste reservoir. Add 100 m£ de-
ionized water, stir and decant the liquid as soon as most of the
resins have settled at the bottom. Repeat the rinsing procedure
until the rinse liquid remains clear when treated with a few drops
of 4 N 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 System for Analytical Run
Connect outlet end of cation exchange resin (Ag+ form) column to inlet
end of cation exchange resin (H+ form) column with a low dead volume 1/16"
to 1/16" stainless steel tubing connector. Similarily 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
samplling 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 in-
jection valve, sampler and peristaltic pump. Leave other components in opera-
ting 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
Sulfuric acid, sodium sulfate or ammonium sulfate may be used as stan-
dards. Ammonium sulfate is preferred.
8.3.1 S04= (100 ygs/m£) standard, alcoholic stock solution DissoJve 275000
± 100 pgs of primary standard ammonium sulfate in 200 m£ of ^onized
water in a 2000 m£ volumetric flask. Add 300 m£ pure IPA shake vig
orously until thoroughly mixed, and make up to volume wxth 60% IPA^
Store in clean polypropylene bottles. (Note: a. There is a volume
decrease of about 2.7% when two parts of water xs naxed wxth 3 parts
451
-------
of IPA. b. Do not use detergents nor dichromate-sulfuric acid solu-
tion 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 S04= calibration standards (0.5 to
20 ygs/m&) by dilution of appropriate aliquots with 60% IPA. Store
standards in capped polypropylene bottles.
8.3.2 S04= (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 S04= (100 ygs/m£) stock solution dispense appro-
priate volumes containing 10, 20, 40 ..., 180, 200 ygs of SO4= into 30
of 60 mS, 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 of ammonium salt was observed to improve precision of sulfate mea-
surement. Sample losses from accidental contact of the filter surface with
another sufrace 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 V2 is
opened. Ammonia from concentrated ammonium hydroxide fills the box and con-
verts sulfuric acid to ammonium sulfate. One hour exposure to ammonia vapor
is adequate. V2 is closed, most of the ammonia is pumped out and passed
through a KI^PO^ 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 monentarily increase the solubility of barium chloranilate and will pro-
duce 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 reservoir of the
HPLC pump. The use of solid standards as prepared in 8.3.3 will eliminate
variability due to IPA/water mismatch.
452
-------
8.5.1 From Fluorocarbon Membrane Filters
Place filter in appropriate size polypropylene bottle, if approximate
sulfate level is known from previous analysis of similar samples, mea-
sure adequate volume of 60% IPA to give sulfate concentration of about
10 ygs/mfi,. Otherwise add 10 m£ 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. Particulate 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 callapses and is complete-
ly 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 y pore size fluo-
rocarbon in-line filter. These filter syringes are available commer-
cially. If the particles are sufficiently large, they can be removed
from the bulk solution by centrifugation.
8.5.2 From glass fiber filters
This procedure is for the extraction of 47 mm diameter filters. Ad-
just the volume of the extracting solvent accordingly for different
size filters.
Place the filter in appropriate size polypropylene bottle. If approxi-
mate SO4= level is known, add adequate volume of 60% IPA to give sul-
fate concentration of about 10 ygs/mJl. 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 stir-
ring 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 b^he automated
BCA method. One important requirement, however, is that the •*^a*™*£
made up to 60% IPA (e.g., 4 parts of the extract must be a^ ^^fSis
pure IPA, v/v) before the sample can be analyzed, it tn
mixed directly with pure IPA, volume shrinkage (about 2.7%) must be tax
into account in the calculation of concentration.
453
-------
Another approach is to evaporate completely the solvent in a known vol-
ume of the extract in polypropylene bottle similar to the preparation of so-
lid 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 in 3 ml/win, allow the in-
strument 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. Ad-
just sample injection time so that peaks from successive sample injections do
not overlap. Fill sample cuvetted 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 concentra-
tion in ygs/m£ of the sulfate standard. The curve is not linear. Alterna-
tively, the peak height or area concentration data may be fitted into a poly-
nomial of the form:
234
y = a + a + ax + ax + ax + ...
where: y = sulfate concentration in ygs/mJl
x = peak height or area
8.8 Calculations
Calculate the concentration of sulfate as ygs SO^/mH using the calibra-
tion curve or the polynomial regression equation. Total soluble sulfate
(SO4=) in the filter is then given by:
(S04=)F = (S04=)d x VQ x d
where: V = total volume in m£ of original sample extract
= sulfate concentration of the diluted sample in Ugs/mil
454
-------
d = dilution factor
= 1, if there is no dilution of the extract before analysis
V = aliquot volume in mi of sample diluted with 60% IPA
3.
V, = final volume in m£ of the aliquot sample after dilution
with 60% IPA
Example :
Suppose 10 m£ of 60% IPA was used to extract the soluble sulfated in
the filter and that 2 m£ of this was further diluted with 4 mi 60% IPA to
bring detector response within calibration range. Suppose that the concen
tration of the diluted sample was found to be 5 ugs/m£ . Then,
(S04=)d = 5 ygs/m£
V = 10 m£
o
V = 2 m£
a
V, = 2 + 4
d
= 6 m£
(S04=)p = 5 x 10 x (6/2)
= 150 ygs
455
-------
Figure 1. Chromatogram at trace sulfate levels, 0.01, 0.02, 0.05,
0.01 ygs SC>4=/in£. Flow rate at 3.2 m£/min. Detector sensitivity at 0.01
absorbance \anits full scale. Sample volume injected = 0.5 m£
456
-------
-f=>
01
Figure 2. Reproducibility of repetitive sample injections. Flow at 3.2 m£/min. Detector sensitivity
at 0.5 AUFS. Sample volume injected = 0.5 m£. Numbers above peaks are sulfate concentrations in |jgs/m£.
-------
43.
CJ1
00
B - buret
LR - liquid reservoir
LP - PHLC pump
P - Pressure Monitor
SV - Sample Injection valve
L - Sample loop
Integrator
Position A
Inject
CX-Ag+ - Cation exchange resin, silver form
CX-H+ - Ca,tion 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.
-------
•1/4" union-
5 micron Stainless
Steel Frit
1/4" to 1/16" Reducer
Figure 4. Configuration for packing column.
459
-------
V-J
VACUUM «
*KH2P04'
Perforated
Shelf
Vent
V,
Filter on
X* Petri dish
Concentrated
NH4OH
Figure 5. Schematic pf an aitimoniation set-up.
-------
DISCREPANCIES BETWEEN EPA AND DEPARTMENT OF EMISSIONS RESEARCH
BCA SULFATE PROCEDURE
4.1.1.1 - Bottled N2 gas is used to drive 60% IPA from a reservior through
the HPLC system at average rate of 4.5 m£/min0
4.1.1.2 - The Dupont 837 sample cell volume is 6.3 yJl.
4.1.1.3 - A dedicated 1 m£ sample 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 us used since
chromatogram sensitivity is controlled at range switch on spectro-
photometer.
4.1.1.6 - A 6-inch long cation column is used.
4.1.1.7 - A 2-1/2 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 mJl) 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 AgNO3 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.
461
-------
8.4 - The chamber containing sample filters in open Petri dishes is
purged with ammonia from concentrated ammonium hydroxide 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.
462
-------
VALIDATION WORK AT SwRI
1. Diesel fuel gives a positive interference
a. 1 yg diesel fuel gives response equal to 0.02 yg SO 2~.
b. Response to diesel fuel occurs with or without BCA column in
system indicating a nonsulfate interference in the fuel.
c. Diesel fuel response can be removed by washing diesel fuel doped
filter with 25 m£ cyclohexane.
d. Sulfate response unchanged by washing sulfate doped filter with
25 m£ cyclohexane.
2. 5-20 percent of the sulfate response on some filters collected from
diesel fuel engines was not due to sulfate.
a. Filters washed with 25 m£ cyclohexane were 13-20 percent lower
in apparent sulfate response than unwashed filters. These filters
were duplicates collected from diesel powered vehicles. Cyclo-
hexane does not remove sulfate from sulfate doped filters. Cyclo-
hexane does remove diesel interference from diesel doped filters.
b. Diesel filter samples run without BCA column in analysis system
were 5-18 percent lower in apparent sulfate response than iden-
tical samples run with the BCA column in place.*
* System designed to run with and without BCA column in system was similar
to the system described by N. J. Khatri, J. H. Johnson and D. G. I^ddy
in "The Characterization of the Hydrocarbon and Sulfate Fractions of
Diesel Particulate Matter," SAE. Paper No. 780111, February-March, 1978.
In this work they found a large portion of the sulfate due to what they
attributed to hydrocarbon interferences.
463
-------
APPENDIX L
DMNA SAMPLING PROCEDURE
464
-------
DESIGN AND CALIBRATION OF SECONDARY DILUTER
A secondary diluter, to be used in conjuction with a constant volume
sampler, has been constructed. The secondary diluter was constructed:
1. to determine the concentration of N-nitrosodimethylamine in auto-
mobile 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 NO2 with materials in the system and
on the traps would cause interference.
3. the sample must be presented to the Tenax GC trap at an appropri-
ate 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 principels given below:
1. diluted automobile exhaust is pulled into the secondary diluter
through a 1/8" OD stainless steel tube at a flow rate of approxi-
mately 2.5 Vmin; this flow is dependent upon the downstream pres-
sure drop.
Dry nitrogen for secondary dilution is add ed through a^oot j .meter
and rotometer, in series, at approximately 68 £/tain. This "~ »
independent of the downstream pressure drop. (We run at 55 l/«an.)
465
-------
Vacuum
Gage
CT>
^
mm
Roots
Blower
denotes line heated to 60°C
Dilution
Flowmeter
Regulating
Valve
d
Excess
Vacuum
Tenax Trap
Sample
Flowmeter
Regulating
Valve
Vacuum
Figure 1. Secondary diluter
-------
3. Diluted automobile exhaust and
and pulled through a 3/8" 01
4. A magnehelic gauge is used to monitor the pressure drop across the
diluter The reference side of the magnehelic is connected to Se
primary dilution tube; this allows dilution ratios to be determined
without making corrections for differences in atmosoheric and™
mary dilution tube pressure. P
5. Three simultaneous samples are pulled through; (a) glass fiber fil-
ters, (b) Tenax GC traps, (c) rotometers, and (d) needle valves
The needle valves are set to maintain a flow of 2.5 Jl/min through
each sample trap. (We use 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.
nn* c,amDie rotometer (5) was disconnected from its normal vacuum
SrStS r™aV a'aiapfcra. pu*p of sufficient size to p»U
2 5 i/min The reference side of the magnehelic was left open to
opelatio V bag of propane span g,s »as connect
467
-------
to the sample inlet (3) . The diluted sample was then collected in
an emply Tedlar bag downstream of the diaphragm pump. This bag
was then 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 pri-
mary diluter and at the same time bags were collected on the se-
condary diluter. The bags collected on both primary and secondary
diluters were then analyzed on appropriate analyzers. By dividing
the concentration found in the primary bag by the concentration
found in the bag of diluted span gas, the dilution ratios were
found for different 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 pressure 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 magneh-
lie gauge readings. The expression used for correcting the pres-
sure drop is PiVi = V2 where :
P, = initial pressure
V1 = initial volume
PP = 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 magnehlic gauge using
the general expression for flow in pipe:
(ref ' 1J
where: AP = differential pressure
f ' = friction
L/D = length/diameter ratio
g = gravitational constant
v = specific volume of gas, I/density
V = linear velocity
468
-------
Since the sample inlet is of constant cross section, the linear velocity
is proportional to the volumetric flow rate, m any given set of experiments*
all variables in this equation other than velocitv anH ^
Because of this, the equation reduced to: Vei°Clty and Pressure are constant.
2 2
AP = CV = CQ
where O
sample flow rate
- * — ; - - — : — -
volumetric
The dilution ratio, R, is given by:
R •
°
N2
-Q-
N2
Constant
Therefore: I = c..
R
and the plot of the reciprocal of the dilution ratio against "YAP 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 1^ 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.
469
-------
O
Tunnel Propane Injection
n
DP " °-02030 /"SF0.0263
O Carbon Monoxj.de Injection
0.030
0.020
0.016
2.0
4.0 4.1
Figure 2, Calibration of N2 secondary dilute*, February 10, 1976
(least square lines).
-------
TABLE 1. N2 SECONDARY DILUTER CALIBRATION VALUES
FEBRUARY 10, 1976
TAP
2.0
7 O^fi
2 5495
9 TZRf,
2090/1
0 1 ftOT
j . J.O^ -J
•3 -31 CC
a • jXOO
3c; •3^'?
.D ODD
3*7 yl 1 "7
.7417
3Q"79Q Q
.0 /^yo
4-1 «-» *5 1 -i
.12311
A 0
B -0
R2 0
(2) 0
(4) 0
Carbon Monoxide
0.01712
0 02500
On^ARn
On AC. c: i
.U4D3-L
l/DF - A (fAP~)
.01709 0.02030
.01765 -0.02630
.99749 0.99414
.01653 0.01430
.05071 0.05491
l/DF
Propane
OA 1 QAO
.QlyQo
On "3 1 n^.
• UoJ. /D
On OQ 0*7
. U Jy J /
OARI i n
.UDJ.XU
Onc*7m
• UD /UX
+ B
0.01684
-0.01622
0.97560
0.01747
0.05115
Rotometer
Of\ f\ 1 ^ f
.02116
On oAm
.0300J.
OAOQQO
Ondnfi1?
n nsm ^
Overall Data
0.01858
-0.02133
0.0818
0.01584
0.05301
471
-------
PROCEDURE USED BY SwRI FOR OBTAINING DMNA TRAPS
PREPARATION OF 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 leaat 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 un*til
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" H20. Both knobs need to
be adjusted simultaneously since they are dependent on each other.
These two will have to be monitored throughout the test. Flow-
meters 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.
472
-------
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 of 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.
473
-------
APPENDIX M
DMNA ANALYSIS PROCEDURE
474
-------
THE MEASUREMENT OF DMNA IN EXHAUST
As used by
Research Triangle Institute
Research Triangle Park, N.C.
Developed by
Research Triangle Institute
Research Triangle Park, N.C.
November 1978
475
-------
RESEARCH TRIANGLE INSTITUTE
ANALYSIS FOR N-NITROSODIMETHYLAMINE IN EXHAUST GASES USING A TENAX
GS CARTRIDGE AND GAS CHROMATOGRAPHY/MASS SPECTROMETRY/COMPUTER
EPA Contract No. 68-02-2767
RTI/1514/00-01S
Special Interim Technical Report
by
E. D. Pellizzari, Project Director
Date: November 27, 1978
Project Officer
Ron Bradow
Mobile Source Emissions Reserach Branch
Mail Drop 46, ERG Annex
U.S. Environmental Protection Agency
Research Triangle Park, N, C. 27709
Prepared for the Environmental Protection Agency, Reserach Triangle
Park, N. C. 27711
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27709
476
-------
The information presented in this document is subject to the following
qualifications:
(a) The mention of a specific company does not imply the intent to
regulate that company or its activities nor that, unless specifi-
cally stated, the company is the source of a given compound;
(b) The identification of compounds were determined by mass spectro-
metric and retention index techniques and their identity are sub-
ject to the limits of this methodology.
(c) The mention of compounds in this report does not imply that they
are necessarily carcinogenic or mutagenic;
(d) The possible mutagenic or carcinogenic activity attributed to a
compound is based upon cited literature;
and (e) The experimental findings and conclusion presented in this report
should not be cited, reproduced, or included in other publications
without the expressed approval of the Project Director or Officer.
477
-------
1.0 Introduction
Because of the previous reports of the presence of N-Nitrosodimethyla-
mine in ambient air, interest in the determination of DMN in auto exhaust
heightened. N-Nitrosodimethylamine (DMN) levels in ambient air were deter-
mined for an area surrounding an industrial site in Baltimore, Maryland.
Using Tenax GC cartridge for concentrating DMN and glass capillary/glass-
liquid chromatography/mass spectrometery with specific ion (m/z 74) monitoring,
DMN was detected and quantified. On an industrial site DMN levels reached
32,000 ng/cm (10.67 parts per billion) in the ambient air.
On the basis of these and other observations, research was conducted on
determineing whether N-nitrosodimethylamine was present in auto exhaust. In
conduction with Southwest Research Institute, a study was conducted on the
detection of DMN in auto exhaust of automobiles which have been operated
under various test conditions. SwRI was responsible for generating and par-
ticipating in the collection of the auto exhaust samples during the course of
this program. The results of this research effort is described here.
478
-------
2.0 Experimental Procedures
The collection and analysis techniques given in Appendix A were modi-
fied and used for detecting DMN in auto exhaust gases.
Tenax GC sampling cartridges were prepared at the Research Triangle
Institute (RTI) and shipped by Federal Express to Southwest Research Institute
(SwRI) for the collection of auto exhaust samples and subsequent returned to
RTI for analysis. In all cases the sampling cartridges were expended or
returned within four to five weeks and replaced with a fresh batch in order
to insure a low background level. All samples were analyzed within two to
three weeks after sample collection was completed.
The sampling procedure employed by SwRI consisted of a primary and
secondary diluter and the secondary diluter and transfer lines were maintained
at 60°C. The sampling rate was 1.6 H/min and the test length was approxi-
mately 31 minutes giving a total sample volume of 50 £. The average primary
cvs dilution was 11.1 to 1 and the secondary dilution was 16.7 to 1 giving
an actual sample dilution of 185 to 1. The secondary diluter was necessary
in order to insure that the NOX and hydrocarbon values were at levels which
minimized the potential artifact formation of the N-nitrosodimethylamine on
the Tenax GC sorbent.
479
-------
(APPENDIX A)
N-NITROSODIMETHYLAMINE IN AMBIENT AIR
ANALYTICAL METHOD
Analyte: DMN Method No:
Matrix: Air Range: P.5 ppt - 10 ppb
Procedure: Adsorption on Tenax GC, Precision: ±10%
thermal desorption
with He purge, measure-
ment by capillary gas-
liquid chromatography/
mass spectrometry
Date Issued: Classification: E (Proposed)
Date Revised:
1. Principle of Method
N-nitrosodimetnylamine (DMN) is concentrated from ambient air on Tenax
GC in a short glass tube (1,2) . It is desorbed by heating and purging with
helium into a liquid nitrogen cooled nickel capillary trap and then intro-
duced onto a high resolution gas chromatographic column where is is separated
from interferences. The concentration of DMN is measured from the mass
spectrometric signal at m/e 74 (3).
2. Range and Sensitivity
2.1 The range of the mass spectrometric signal for the conditions listed
corresponds to 0.5 ppt to 10 ppb.
2.2 A concentration of 0.5 ppt of DMN can be determined in a 150-liter air
sample.
3. interferences
Interferences may result from materials having background ions of m/e_
74 (C2H8N3, C2H4N02, C2H6N20, C3H3C1, C-jHgS, C3H6O2, or C3H1C)N2) , if at the"
same retention time of DMN.
4. Precision and Accuracy
4.1 The precision of this method has been determined to be ±10% or relative
standard deviation when replicate sampling cartridges were spiked with 50 ng
(corresponding to 10 ppb in 150 £ of air). These data were obtained using
10.0 cm long glass tubes (1.5 cm i.d.)packed with 35/60 mesh of Tenax GC
(bed dimensions: 1.5 cm x 6 cm in depth).
480
-------
4 2 The accuracy of the analysis is approximately ±10% of the amount re-
ported as determined from repeated analysis of several standards.
5. Advantages and Disadvantages of the Method
5.1 The gas chromatography-mass spectrometry technique interfaced with a
Finnigan glass jet separator (Model 01512-42158 Finnigan Corp., Sunnyvale
CA) is extremely sensitive and specific for the analysis of DMN. The high
resolution gas chromatographic separation yields a retention time that is
characteristic for DMN, and relatively specific for positive assignment of
the signal as DMN. The mass spectrometer in combination with high resolution
gas chromatography yields a very high degree of specificity. The base peak
of DMN is at m/e_ 74 which is also the parent ion. In order to assign the
signal at m/e_ 74 to DMN it is absolutely necessary that the retention time
matches with the signal .
5.2 Collected samples can be stored up to 1 month with less than 10% losses.
5.3 Because DMN is a suspected carcinogen in man it is extremely important
to exercise safety precautions in the preparation and disposal of liquid and
gas standards, cleaning of used glassware, etc., and the analysis of air
samples .
5.4 Since the mass spectrometer can not be conveniently mobilized sampling
must be carried out away from the instrument.
5.5 High resolution gas chromatography /low resolution mass spectrometry is
not a convenient technique for handling a large number of samples (>100/wk) .
5.6 Efficiency of air sampling increases as the ambient air temperature
decreases (i_-e_. sensitivity increases) .
5.7 Ambient air sampling is limited to cases where the NOX levels are less
than 3 ppm when dimethylamine is also present.
6. Apparatus
6.1 Sampling Tubes
-' -
caps, and
x 150 cm) culture tubes, immediately sealed using
cooled.
•4-v, lonaer beds of sorbent may be prepared using
6.1.2 Cartridge samplers with longer oeas uj.
a proportional amount of Tenax GC.
481
-------
6.2 Gas Chromatographic Column
6.2.1 A 0.35 mm i.d. x 50 m glass SCOT capillary coated with DECS station-
ary phase and 0.1% benzyl triphenylphosphonium chloride is used. The capil-
lary column is conditioned (detector end disconnected) for 48 hr at 210°C @
1.5-2.0 m£/min helium flow.
6.3 A Finnigan type glass jet separator on a magnetic or quadropole instru-
ment is used at 200°C.
6.4 Inlet-Manifold
6.4.1 An inlet-manifold is fabricated and employed (Fiugre 1, ref, 1,2,4).
6.5 Gas Chromatograph
6.5.1 A Varian 1700 gas chromatograph or equivalent. A gas chromatograph
employing a single column oven and a temperature programmer is adequate.
6.6 Mass Spectrometer
6.6.1 A mass spectrometer with a resolution of 500-2000 equipped with
single ion monitoring capabilities must be used in conjunction with a gas
chromatograph. A Varian-MAT CH-7 has been found to be satisfactory for this
purpose (2,3).
6.7 Syringes
6.7.1 Syringes, l-m£ gas tight (Precision Sampling, Inc.) and 10 jl£
(The Hamilton Co., Inc.).
7. Reagents and Materials
All reagents must be analytical reagent grade.
7.1 N-nitrosodimethylamine
7.2 Acetone
7.3 Isoclean®
7.4 Tenax GC (35/60 mesh, Applied Science)
7.5 Two 2-liter round bottom flasks fitted with injection ports.
7.6 Soxhlet apparatus
8. Procedure
8.1 Cleaning of glassware
482
-------
GAS
METER
FLOW
METER
H3-
NEEDLE
VALVE
PUMP
CARTRIDGF 1
GLASS
FIBER
FILTER
VAPOPs COLLECTION SYSTEM
PURGE
GAS
ION
CURRENT
RECORDER
MASS
SPECTRO-
METER
GLASS
JET
SEPARATOR
TWO
POSITION
VALVE
THERMAL
DESORPTION
CHAMBER
\
CAPILLARY
TRAP
HEATED
BLOCKS
EXHAUST
ANALYTICAL SYSTEM
Figure 1.
Vapor collection and analytical systems for analysis of
hazardous vapors in ambient air.
483
-------
All glassware, glass sampling tubes, cartridge holders, etc. should be
washed in Isoclean ©/water, rinsed with doubly distilled water and acetone
and air dried. Glassware is heated to 450°F for 2 hrs.
8.2 Preparation of Tenax GC
8.2.1 Virgin Tenax GC is extracted in a Soxhlet apparatus overnight with
acetone prior to its use.
8.3 Collection of DMN in Ambient Air ,,
8.3.1 Continuous sampling of ambient air may be accomplished using a
Nutech Model 221-A portable sampler (Nutec Corp., Durham, NC) or its equi-
valent (2). Flow rates are adjusted with a metering valve through a cali-
brated rotameter. Total flow is registered by a dry gas meter.
8.3.2 For larger sample sizes it is important to realize that a larger
total volume of air may cause elution of DMN through the sampling tube. It
has been demonstrated that exceeding a total of 385, 332, 280, 242, 224, 204,
163, 156, 148, 127, 107, 93, or 79-liters of air at temperatures of 50, 55,
60, 65, 70, 75, 80, 85, 90, 95, 100, 105 or 110°F, respectively will result
in elution of DMN from the cartridge sampler. A flow of 10 cc/min to 30 £/
min may be used with the sampler described in 6.1.
8.3.3 DMN has been found to be stable and quantitatively recoverable from
cartridge samplers after 4 weeks when tightly closed in cartridge holders,
protected from light and stored at 0°C.
8.4 Analysis of Sample
8.4.1 Instrument Conditions and Set-up. The thermal desorption chamber
and six-port valve are set to 200°C. fhe glass jet separator is maintained
at 200°C. The mass spectrometer is set to monitor m/e_ 74 (Figure 2).
8.4.2 Adjust the He purge through the desorption chamber to 50 mJl/min.
Cool the Ni capillary trap at the inlet manifold with liquid nitrogen.
8.4.3 Place the cartridge sampler in the desorption chamber and desorb
for 5 min.
8.4.4 Rotate the six-port valve on the inlet-manifold to position "B",
heat the Ni capillary trap to 180°C with a wax bath.
8.4.5 Temperature program the glass capillary column from 75 to 205°C at
4°C/min and hold at upper limit for 10 min. The retention time of DMN is
approximately 26 min (Figure 3).
8.4.6 The analytical column is cooled to ambient temperature and the
next sample is processed.
484
-------
uu
90
80
70
60
50
40
30
20
10
0
.
.
•
H
m/
1 ,1 J
e 42
*"/ c 1 *±
, ,
•FIT111 'I'l'I'P^I'I'I'l1 'I11
10 20 30 40 60 70 80 90 110 120 130 140 160 170
0 bO 100 150
m/e-
Figure 2. Mass spectrum of N-nitrosodimethylamine.
485
-------
00
CTl
50 -
40 -
30 -
20 -
10 .
0
50 .
40 •
30 "
20 .
10 -
0
50 .
40 -
30 '
20
10
0
x 3
x 3
x 3
DMN (300 ng)
26 min.
xl
26 min.
xl
26 min.
TIME (MIN.)
Figure 3. Mass (m/e_ 74) Chromatograms. A = standard DMN, B,C = replicate air samples.
-------
9. Calibration and Standards
9.1 Preparation of Gas Standard
cno tW°"2 llter rOUnd b°tt0m flask with helium, warm flasks to
50 °C with heating mantels and use magnetic bar to stir vapors.
9.1.2 Inject 0.1-1 yJl of DMN into flask and let stir for 30 min Make
further dilutions into second flask by transferring milliliter gas volumes
as needed.
9.1.3 Purge air/vapor mixtures from second flask onto cartridge samplers.
9.2 Calibration
9.1.2 Prepare standard curve (with ten concentration points) by thermally
desorbing cartridge samplers loaded with 3 ng to 30 yJl of DMN. Plot m/e_ 74
response vs ng of DMN. A linear response is observed.
10. Calculations
10.1 Hie total quantity of DMN in ambient air is determined by comparting
m/e_ 74 response for samples of DMN with standard curve.
ng DMN 24.45
ppb = — - = —^~
where :
ng DMN = total ng concentration is determined in 9.2.1
V = volume of air in liters sampled at 25 °C and 760 torr
24.45 = molar volume of an ideal gas at 25°C and 760 torr
MW = moleculat weight of DMN, 74.
11. References
148 pp.
Analytical Techniques for Measuring
Va
1975, 187 pp.
3. PellaZ2ari, ... a... B««*. ..»• BerXiey ana ,. McCrae. Bio^.ica,
Mass Spec, submitted.
v T E Bunch, and E. Sawicki,
Environ. Sci.
487
-------
APPENDIX N
BaP SAMPLING AND ANALYSIS
488
-------
SAMPLING AND ANALYSIS OF BaP
Sample Collection at
Southwest Research Institute
Analysis Method Developed and used at
Environmental Protection Agency
Research Triangle Park, N.C.
June 1979
489
-------
SAMPLING AND ANALYSIS OF BENZO-CX-PYRENE
The analysis for benzo-a-pyrene (BaP) will be carried out by collecting
particulate samples on 8" x 10" glass fiber filters at Southwest Research
Institute and sending the filters to EPA-Research Triangle Park for analysis
by fluorscence spectroscopy.*
Sampling for BaP has been successfully conducted in the past at South-
west Research by collecting diesel particulate on 8" x 10" glass fiber
filters. Other filtering media and/or sizes are under consideration for
sampling and may be used in the future. Sample flow rates will depend on
filter size and loading capacity. The temperature of the dilute exhaust at
the sampling point will not exceed 125°F.
After the filters have been loaded with sample they will be weighed,
folded in half so that the particulates are inside and folded again. All
samples will be handled under yellow light as BaP is degraded in the pre-
sence of white light. The filter will then be placed in a glassive envelope
and then in a manila envelope. Several envelopes will be placed in a zip-
lock plastic bag purged with zero nitrogen and heat sealed. The samples
will be stored at -30°C until shipped in an insulated container with dry
ice via air freight to EPA-RTP for analysis.
At EPA-RTP particles on the glass fiber filters are extracted,with
methylene chloride. A portion of the methylene chloride solution is analyzed
for BaP. The methylene chloride solution is diluted with cyclohexane and
spotted on thin layer acetylated cellulose plates. The plates are developed
and the fluorescence due to BaP is measured in a spectrophotometer. The
concentration of BaP is determined by comparing the response of the sample
to the response of a series of standards.
* Private communication with EPA-RTP.
490
-------
4. TITLE AND SUBTITLE
ANALYTICAL PROCEDURES FOR CHARACTERIZING UNREGULATED
EMISSIONS FROM VEHICLES USING MIDDLE-DISTILLATE FUELS
Interim Report
. AUTHOR(S)
Lawrence R. Smith, Mary E. Parness, E. Robert Fanick
and Harry E. Dietzmann
9. PERFORMING ORGANIZATION NAME AND ADDRESS"
Southwest Research Institute
8500 Culebra Road
San Antonio, Texas 78284
REPORT NO.
EPA-60Q/2-80-068
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
April IQftn
PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT \o
10. PROGRAM ELEMENT NO.
07A1D 14-0457 (FY-80)
11. CONTRACT/GRANT NO.
68-02-2703
2. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Interim.
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEI*
ITARY NOTES
i e. ABS
This research program was initiated with the objective of developing, codifying
and testing a group of chemical analytical methods for measuring toxic compounds in
the exhaust of distillate-fueled engines (i.e. diesel, gas turbine, Stirling, or
Rankin cycle powerplants). It is a part of a larger effort to characterize these
components from a number of prototype powerplants and, thus, represents a logical
first step in the process.
Methods of collection and analysis for aldehydes and ketones, for hydrogen
cyanide and cyanogen, for hydrogen sulfide, carbonyl sulfide and organic sulfides, for
ammonia and amines, for nitrous oxide, sulfur dioxide, individual hydrocarbons, for
soluble sulfate and N-nitrosodimethylamine, benzo-a-pyrene, and phenols were studied
in detail. Ten analytical procedures were developed and codified. Interference
studies and proof-tests in diesel engine exhaust were conducted with every procedure
and the results of these experiments are reported in detail.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
*
*
* Air pollution
* Vehicles
* Distillates
Fuels
Emission
* Collecting methods
* Chemical analysis
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b.lDENTIFIERS/OPEN ENDED TERMS
1Q SECURITY CLASS tThisReportl
"UNCLASSIFIED
20 SECURiT^ CLASS i This page
UNCLASSIFIED
COSATI I ield/Group
138
116
148
21D
21 NO. OF PAGES
505
PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION .s OBSOLETE
491
------- |