PB 198 072
CHEMICAL SPECIES IN ENGINE EXHAUST AND
THEIR CONTRIBUTIONS TO EXHAUST ODOR
Andrew Dravnieks, et al
IIT Research Institute
Chicago, Illinois
November 1970
NATIONAL 'ECHNICAL INFORMATION SERVICE
Distributed ,,, 'to foster, serve
and promote the nation's
economic development
and technological
advancement.'
U.S. DEPARTMENT OF COMMERCE
This document has been approved for public release and sale.
-------
Report No. IITRI C6183-5
Final Report
CHEMICAL SPECIES IN ENGINE EXHAUST
AND THEIR CONTRIBUTIONS
TO EXHAUST ODORS
National Air Pollution Control
Administration
and
Coordinating Research Council
NT RESEARCH INSTITUTE
-------
STANDARD TITLE PAGE
FOR TECHNICAL REPORTS
1. Report No.
3. Recipient's Catalog No.
4. Title and Subtitle
Chemical Species in Engine Exhaust and Their Contributions
to Exhaust Odor
5. Report Date
November 1970
6. Performing Organization Code
77 Authons)
Andrew Dravnieks and Anne O'Donnell
8. Performing Organization Rept. No.
9. Performing Organization Name and Address
Illinois Institute of Technology Research Institute
10 West 35th Street
Chicago, Illinois 60616
12. Sponsoring Agency Name and Address
Division of Emission Control Technology
Air Pollution Control Office, Environmental Protection Agency
5 Research Drive
Ann Arbor, Michigan 48103
10. Project/Task/Work Unit No.
TT Contract/Grant No.
CPA-22-69-98
13. TypTof Report & Period Covered
Final
1969-1970
14. Sponsoring Agency Code
15. Supplementary Notes
Work cosponsored by Coordinating Research Council - CAPE 7 Project.
at 1971 ACS Meeting at Los Angeles.
Results presented
16. Abstracts
5-This project investigated the chemical composition of the odorants in diesel
engine exhaust. Exhaust samples were collected in fluidized or packed bed collectors
filled with gas chromatographic packings. Separation of the individual compounds was
accomplished by two gas chromatograph columns in series. Final identification of the
separated compounds was accomplished by mass spectrometry. About 100 compounds were
judged to be odor relevant in diesel engine exhaust by this procedure.-
17. Key Words and Document Analysis, (a). Descriptors
Diesel Engine Exhaust Odor
Diesel Engine Emissions
Odor Component Identification
17b. Identifiers/Open-Ended Terms
17c. COSATI Field/Group
18. Distribution Statement
Public Distribution - No Restrictions
19.Security Classdhis Report)
UNCLASSIFIED
^.Security Class. (This Page)
UNCLASSIFIED
21. No. of Pages
99
227Price
3.00
FORM NB5-««7(1-70)
-------
Report No. IITRI C6183-5
CHEMICAL SPECIES IN ENGINE EXHAUST
AND THEIR CONTRIBUTIONS TO EXHAUST ODORS
Prepared by
A. O'Donnell and A. Dravnieks
of
IIT Research Institute
Technology Center
10 West 35 Street
Chicago, Illinois 60616
for
National Air Pollution Control Administration
and
Coordinating Research Council
November 1970
Copy No.
NT RESEARCH INSTITUTE
i
-------
FOREWORD
This report presents the work accomplished during the
period May 20, 1969, through January 19, 1970, on IITRI Project
No. C6183, "Chemical Species in Engine Exhaust and Their
Contributions to Exhaust Odors. " The progress achieved during
the previous year on the same program is incorporated into this
report to provide a summary of the two year effort.
Other personnel who contributed to the two year program
were Dr. R.G. Scholz, Dr. B.K. Krotoszynski , Mr. A.F. Anderson,
Mr. T.A. Burgwald, Mr. E.H. Luebcke , Mr. T.M. Rymarz, Mr. N.S.
Shaw, Mr. T.A. Stanley, and Miss J.B. Whitfield.
Respectfully submitted,
IIT RESEARCH INSTITUTE
Anne O1 Donne 11
Associate Chemist
Andrew Dravnieks
Senior Scientific Advisor
Head, Olfactronics and
Odor Science Center
Approved by:
Morton J. Klein
Director, Chemistry
Research Division
AO'D:db
II T RESEARCH INSTITUTE
ii
-------
ABSTRACT
The nature of odorous species in the .exhaust from a Detroit
Diesel Engine Division 6V-71N, operated, at constant load and
fuel conditions, was investigated using high-resolution two
column gas chromatography, mass spectrometry on resolved species,
and sensory observations. Solid adsorbent-collection devices
were used to sample from a nitrogen-diluted exhaust stream.
Although the prer.^nce of many hundreds of compounds was indicated
by the chromatograms, only a small fraction were found to exhibit
distinct odors at concentrations encountered in the exhaust. The
major concentration species, the paraffinic hydrocarbons, are
individually non-odorous. The odor-relevant species are polar,
and many exhibit low odor thresholds, occurring in the exhaust in
relatively small concentrations. From mass spectral data, a
variety of compound types was found among the more important odor
contributors including: aliphatic aldehydes, aliphatic compounds
with more than one position of unsaturation, alkyl derivatives
of benzene, indan, tetralin, and naphthalene, aldehyde and ketone
derivatives of benzene and alkylbenzenes, ancksulfur species.
Auxiliary ^ s chromatographic methods suggested\that aliphatic
acids are also odor relevant.
IT RESEARCH INSTITUTE
iii
-------
TABLE OF CONTENTS
PAGE
FOREWORD i
I. INTRODUCTION 1
II. SUMMARY AND CONCLUSIONS 2
III. EXPERIMENTAL EQUIPMENT AND PROCEDURES -
DESCRIPTION AND RATIONALE 5
A. Emission Source 5
B. Exhaust Sampling Devices 5
C. Sample Elution and Injection Techniques 9
D. Gas Chromatographic Equipment and Techniques 11
E. Gas Chromatographic - Mass Spectrometer
Interfacing 14
F. Experimental Basis for Sample Component
Characterization 15
IV. RESULTS AND DISCUSSION 22
A. Single Column Chromatograph Studies 22
B. Initial Two-Column Chromatograph Studies
on Fluidized Bed-Collected Samples 25
C. Chromatographic and Mass Spectrometric
Analyses of Diesel Exhaust Samples Collected
on Chromosorb 102 35
D. Acidic and Phenolic Components in Diesel
Exhaust 50
II T RESEARCH INSTITUTE
iv
-------
LIST OF FIGURES
PAGE
1 Calibration of Dual-Column Chromatograph 19
2 Typical Carbowax 20M Chromatogram of Diesel
Engine Exhaust 23
3 Typical Apiezon L Chromatogram of Diesel Engine
Exhaust 24
4 Typical Dual-Column Chromatogram of Diesel Engine
Exhuast 29
5 Reproduci>~ility of Dual-Column Chromatograph 30
6 Analysis of Diesel Exhaust in Two-Column
Chromatograph 32
7a- Carbowax 20M Chromatogram of Diesel Exhaust 38-
7g Collected on a High Capacity Chromosorb 102 Sampler 44
8a- Apiezon L Chromatograms of Strong Burnt Carbovax 48-
8b 20M Peaks 49
9 Standards-Kovats Index-Diesel Exhaust Samples 51
A-l Diesel Exhaust Sampling Scheme A-3
B-l Fluidized Bed Collector B-3
B-2 High-Speed Organic Vapor Collector B-4
C-l Collected Sample is Transferred to Injector C-3
C-2 Apparatus for Sample Injection into Gas
Chromatograph C-5
D-l Single Column Gas Chromatographic Flow Arrangement D-3
D-2 Sniffinj Port Design D-4
D-3 Dual-Column Gas Chromatographic Flow Arrangement D-6
F-l Mechanism for the Trapping and Release of
Components in the GC-MS Interface F-3
F-2 Interface Connection to the Two-Column Gas
Chromatograph F-4
F-3 Device for Sample Introduction into the Mass
Spectrograph F-6
F-4 Recorder Trace of m/e Peak Intensity After
Injection of Benzene Peak Trapped in Interface F-8
G-l I so-Boiling Point Curves from Two-Column
Chromatograph Data on Known Compounds G-4
G-2 I so-Molecular Weight Curves from Two-Column
Chromatograph Data on Known Compounds G-5
NT RESEARCH INSTITUTE
-------
LIST OF TABLES
PAGE
I Sampling Reproducibility - Gas Chromatographic
Analysis of Three Samples Collected from the Same
Synthetic Mixture of Vapors in Helium 10
II High Capacity Chromosorb 102 Needle Collectors -
Efficiency of Sample Recovery Procedures 10
III Two-Column Chromatograph Data from Standards 17
IV Retention Dispersion on Carbowax 20M for
Components Eluted from Apiezon L 21
V Frequency of Odor Descriptions 26
VI Odor Characterization by Kovats Indexes - Single
200-ft C20M Column 27
VII Odor Characterization by Kovats Indexes - Single
200-ft APL Column 28
VIII Summary of All Odorous Peaks on Primary Column
Carbowax 20M Chromatogram - Dual-Column
Chromatograph 33
IX Summary of All Odorous Peaks on Primary Column
Apiezon L Chromatogram - Dual-Column
Chromatograph 36
X Data Summary on Odor Relevant Exhaust Components 45
A-I Engine Operating and Sampling Conditions A-4
A-II Analysis of Diesel Fuel A-5
HT RESEARCH INSTITUTE
vi
-------
I. INTRODUCTION
The objective of the project was to investigate the nature
and identity of molecular species encountered in a diesel exhaust
reproducibly available from the same engine operated under the
same conditions with the same fuel. Emphasis was placed on those
species that by sensory assay were rated as significant odor
contributors.
The isolation and identification of the odor-relevant
components in dj.esel exhaust required the application of effective
sampling and analysis techniques suitable for handling complex
mixtures of organic substances in the presence of a large excess
of water vapor. The direct sampling from a nitrogen-diluted
exhaust progressed to solid-adsorbent-phase collectors which
completely extracted all possibly odor relevant organic vapors
from a diluted exhaust flow. Separation of components in the
collected sample was achieved by high resolution gas chromatography.
Support-coated open tubular columns with Apiezon L and Carbowax
20M stationary phases were used for the initial analyses of exhaust
samples in a single column chromatograph which provided for odor
analysis in parallel with flame ionization detection.
Furti.jr sample resolution and partial component character-
ization through polarity dispersion were achieved with a two-
column chromatograph. The dual-column system comprised two high
resolution columns of different polarity, Carbowax 20M and
Apiezon L, connected in series, with the option of selecting the
portions of the first column effluent to be injected onto the
second column. Odor analysis was performed on the effluent from
each column. With the two-column chromatograph, sufficient
isolation of the odor-relevant species from the bulk of non-odorous
components was obtained to allow the selected trapping of odorous
compounds for mass spectrometric identification with a specially
designed interfacing unit.
IT RESEARCH INSTITUTE
1
-------
II. SUMMARY AND CONCLUSIONS
High-resolution gas chromatographic analysis, sensory assays
of gas chromatographically resolved components, and mass spectro-
metric identification techniques were employed to investigate the
nature of chemical species in diesel exhaust. Emphasis was placed
on those species which appeared to carry significant odor notes at
levels corresponding to their individual concentrations in the
exhaust. The source of the exhaust was kept constant; a Detroit
Diesel Engine Division 6V-71N,operated with No. 1 diesel fuel at
54 HP and 1200 rpm,served as the source.
The collection and sample enrichment techniques consisted
of preferential absorption of organic species by non-polar organic
materials followed by recovery at higher temperatures in an inert
gas. Each analysis utilized a new sample. Three different
collecting phases and several collector geometries were used.
Chromosorb 102, a porous polymer, solid-adsorbent medium, proved
to be the most efficient collecting phase, providing the capacity
for large samples and completely extracting all species of possible
odor relevance from the exhaust sampling stream. Complete sample
extraction from a known volume of exhaust permitted the extrapolation
of odor judgements on species in the gas chromatograph effluent back
to odor relevance in the original exhaust.
Gas chromatographic separations utilized Carbowax 20M and
Apiezon L support-coated open tubular columns, singly or in tandem
in a special two-column gas chromatograph. The presence of organic
acids and phenols was explored using a Chromosorb 101 column. A
flame ionization detector and sniffing port on the effluent from
the columns were used to observe the eluted species and evaluate
their odor notes. A flame photometric sulfur detector was employed
to scan the gas chromatographic effluent for the presence of sulfur
compounds.
Mass spectrometric investigations were conducted on species
isolated after passage through the two-column gas chromatograph in
which the sample was first resolved in a Carbowax 20M column and
II T RESEARCH INSTITUTE
2
-------
individual multicomponent peaks were further separated in an
Apiezon L column. Interpretations on the nature and physico-
chemical properties of components were possible on the basis of
the positions of peaks in these two columns. Gas chromatographic
classifications were reduced to Kovats index notation for ease
of cross-referencing. The indexing was facilitated by the presence
of n-alkanes in the exhaust; these served as convenient internal
standards.
The study produced an inventory of odorous components in
diesel exhaust up to an estimated boiling point of 260°C. The
following conclusions were drawn:
1. The composition of diesel exhaust was sufficiently complex
so that even the resolution obtained with a 200' polar
gas chromatographic column with 15,000-20,000 theoretical
plates still yielded peaks which were further resolvable
into 4 to 20 species in a second non-polar column. At
sensitivities sufficient to gas chromatographically
represent species present in the exhaust at concentrations
above 10-13 g/ml, the estimated number of all components
exceeds 1000.
2. The majority of the organic species are individually
non-odorous at concentrations encountered in the exhaust.
Less than 100 species exhibit marked odoros at their
respective concentrations. Among the non-odorous
jomponents, aliphatic hydrocarbons with at most one
d'^'hle bond seem to account for most of the species.
The largest concentrations occur in the range between
CIQ aiid Ci4 alkane elution times.
3. Aldehydes of the n-series, from ethanal through octanal,
were identified mass spectrometrically, and the presence
of higher members of the series through undecanal was
suggested from sensory and gas chromatographic data.
Aldehydic derivatives of benzene, alkyl substituted
benzenes, and furan were also identified. The odor notes
of the various aldehydes are highly dependent on the
nature of the aldehyde, and,therefore, they cannot be
considered as an homogeneous sensory class.
4. Alkyl derivatives of benzene, indan, tetralin, and
naphthalene were classes of identified components
contributing a considerable number of strongly odorous
species. These odors were classified as "burnt-pungent."
5. Mass spectral data indicated the presence of aliphatic
and cyclic species with more than one position of
II T RESEARCH INSTITUTE
3
-------
unsaturation. These components carried "burnt" odor
notes.
6. Specifically foul odors were observed for six species.
One was identified as trimethylthiophene. The identity
of the others was not clarified but all were strongly
polar, indicating oxygenates or highly unsaturated non-
oxygenates.
7. Approximately 36 species with burnt (but not exclusively
burnt) odor notes were observed. They eluted from the
Carbowax 20M column beginning approximately with the Cn
alkane position and generally tended to be strongly
polar. Five species with particularly intense burnt
odors were observed in the 160°-180° boiling point range.
They possessed very low odor thresholds, occurring in
concentrations insufficient for mass spectrometric
identification.
8. The presence and odor relevance of the free n-alkanoic
acids was confirmed by sensory assay of the gas
chromato graphic effluent at calibrated retention times.
Pure acids with more than nine carbon atoms were judged
sufficiently non-odorous to be excluded as contributors
to exhaust odor.
9. Since the number of significantly odorous species in
diesel exhaust is limited, it appears feasible that valid
correlations can be found between the analytical
composition and the overall odor intensity and character.
However, since the strongly odorous components belong
selectively to certain members of a variety of chemical
classes of species and diverge in odor character and
intensity, an orderly relation between the presences of
certain chemical classes (e.g., total aldehydes) and odors
cannot be expected. Rather, the correlations should deal
with the presence and concentration of certain members of
certain chemical classes.
NT RESEARCH INSTITUTE
4
-------
III. EXPERIMENTAL EQUIPMENT AND PROCEDURES - DESCRIPTION AND
RATIONALE
A. Emission Source
A Detroit Diesel Engine Division 6V-71N with N-60 injectors
served as the exhaust gas generator for this study. The engine,
rated at 210 HP at 2100 rpm, was operated on #1 Diesel Fuel at
54 HP and 1200 rpm. It was mounted on a bed plate and loaded
with a Midwest eddy current dynamometer capable of absorbing
300 HP.
Because the exhaust sampling devices must be operated near
ambient temperature, it was necessary to cool the engine exhaust
without condensing its water content. By diluting a portion of
the exhaust with pure, dry, non-odorous nitrogen,the exhaust
temperature was reduced without passing through the dew point.
At typical operating conditions, overall fuel to air ratio of
0.172 and intake air at 78°F and 23% relative humidity, and
assuming complete combustion of a Ci4H3Q fuel, the dew poinc of
the exhaust was calculated to be 87°F. At the 11:1 nitrogen to
exhaust dilution ratio used, the calculated dew point of the
exhaust sampling stream was below 30°F. Descriptions of the
exhaust SQ. pling system, engine and sampling operating conditions,
and #1 Diese.1. ^uel analysis are included in Appendix A.
B. Exhaust Sampling Devices
Three types of sampling devices were used during the project
to collect exhaust samples for analysis: fluidized bed samplers,
packed bed samplers, and high capacity packed bed samplers.
Collection media employed in the samplers were Apiezon L, SF-96,
and Chromosorb 102. Characteristics of the sampler types varied,
but all permitted a collection and concentration of vapor
components suitable for gas chromatographic and odor analysis
without water vapor collection. The low affinity of the non-
polar collection phases for water eliminated the necessity of
additional sample pretreatment,and the small amounts of water
IIT RESEARCH INSTITUTE
5
-------
collected did not interfere with later sample transfer and
analysis procedures.
The fluidized bed sampler is a glass column containing lOg
of 40/60 mesh Fluoropak 80 with a 15% coating of Apiezon L held
loosely packed. Diluted diesel exhaust was pulled through the
sampler at a rate of 45 1/min. At this sampling rate the bed
material is in constant motion (hence the name, fluidized bed),
providing good contact between the exhaust and the collector
surfaces. During a 30-min sampling period about 1400 liters of
diluted exhaust passed through the bed.
The Apiezon L fluidized bed sampler collects each vapor
component until an equilibrium is attained between the component
concentration in the vapor and in the Apiezon L solution. The
concentration factor is dependent upon the partition coefficient
for each vapor component. Since the Apiezon L phase is a non-
polar material, it exhibits little affinity for water vapor.
The effective concentration of less volatile vapor components was
the useful property of this sampler.
However, in order to make a judgement on the odor relevance
of any particular component in the exhaust, a sampler was needed
which would collect all components from a known volume of exhaust.
The first samplers of this type employed were packed bed collectors
containing 4.2g of 15% SF-96 (a methylsilicone fluid) on 80/100
mesh Chromosorb HP. Two liters of the diluted exhaust were
pulled through the collector at a rate of 100 ml/min. For the
more polar and more volatile exhaust components,this bed, like
the fluidized bed, collects until equilibrium is attained between
the vapor phase and solution phase concentrations. For the
heavier components, the equilibration concentrations will be
greater than the amount present in the gas sampled,so that the
entire amount in the sample will be collected. As with the
Apiezon L collector phase, water has a limited solubility in
SF-96.
Two drawbacks of the SF-96 collector developed during the
course of exhaust analysis procedures. First, interpretation of
(IT RESEARCH INSTITUTE
6
-------
odor relevance data was difficult because of the fractional
collection of some components. Only exact knowledge of the
partition coefficients of these components in the SF-96 phase
would allow calculation of the original component concentration
in the exhaust. Secondly, it became necessary to collect a
larger sample because of the sample splittings involved in the
gas chromatographic procedures and for mass spectrometric
analyses. Increasing the collector capacity by increasing the
amount of SF-96 -..as not a suitable means for increasing sample
size, since then water collection began to reach a level where
it interfered in the gas chromatographic analysis. Also,
larger amounts of water temporarily change the polarity
characteristics of the stationary phase with a marked loss in
resolution and shift in Kovats Indexes of polar components.
A collector phase of Chromosorb 102 (C-102) , a styrene-
divinylbenzene copolymer, was adapted for collection of diesel
exhaust samples in the packed bed configuration. The material
is a solid, porous, non-polar adsorbent with a high collection
capacity due to an extremely expanded surface area, 300-400
The following comparison of an SF-96 collector with a C-102
collector * the same physical size (16 cm-^ of packed column)
emphasizes ti.e superior collection efficiency of C-102:
Volume of Vapor Retained
at 25°C, ml
Vapor SF-96 C-102
Pentane 300 14,000
Methanol 240 2,000
Ethanol 400 12,000
Water vapor retention volume in the C-102 collector is
200 ml at room temperature.
Data available indicated that retentions of various organic
compounds in C-102 follow the order of retention times on non-
polar silicones, such as Dow Corning 550, in which the Kovats
Index of ethanol is approximately 500, similar to n-pentane.
NT RESEARCH INSTITUTE
7
-------
Thus, all compounds with K.I. £ 500 would be fully retained from
the first 10 liters of sample gas, Tabulations of Kovats Indexes
for oxygenates in DC 550 column at 120°C list only the following
compounds with Kovats Index below 500: formaldehyde, methyl ether,
methanol, acetaldehyde, methyl formate, ethanol (K.I. = 498,
already fully retained from 10-1), and ethylene oxide.
The partially collected components from a 10 liter exhaust
sample would now include only those polar species more volatile
than ethanol and non-polar species more volatile than pentane.
Such species have low odor contribution potential,, Therefore, the
C-102 sampler was assumed to be a satisfactory total collector
for all organic species of possible odor interest in the exhaust.
Thus, a better judgement of the odor relevance of particular
components could be made,since extrapolations from component
concentrations at the gas chromatograph sniffing ports back to
actual concentrations and odor relevance in the original exhaust
were possible
In addition to the superior sample enrichment factor, the
C-102 adsorbent is not limited by one inherent defect occurring
with absorbent-coated supports. To accumulate in the film of
liquid or grease coating the support, the vapor molecules must
diffuse through the film material. The rate of this process
depends on the diffusion coefficient of the component in the film,
It is slow for a film of grease-like Apiezon L and faster for the
fluid methyl-silicone, SF-96., However, for an adsorbent such as
C-102, this rate limitation is absent Every molecule impacting
on the surface reaches its equilibration site almost instantaneously
except for residual effects caused by a slight diffusion into the
structural matrix of the polymer. Thus, faster collection is
feasible with the C-102 material. When mass spectrometric analysis
of gas-chromatographically-separated exhaust components necessitated
W.O. McReynolds, "Gas Chromatographic Retention Data," Preston
Technical Abstracts Co., Evanston, Illinois, 1966.
NT RESEARCH INSTITUTE
8
-------
the larger samples, a wider diameter C-102 packed-bed sampler was
fabricated to permit the collection of a greater amount of
exhaust in a reasonable amount of time.. A total sample of 100
liters of dilute exhaust (9.1 liters actual exhaust) was
collected at a rate of 3 1/min,
Details on the construction of the various collectors used
and associated sampling parameters are included in Appendix B,
C. Sample elution and Injection Techniques
The collected diesel exhaust sample was removed from the
various sampler types by reverse flushing with cryogenicall/
purified helium while heating the sampler* Eluted sample was
collected in a specially designed injector needle held at liquid
nitrogen temperatures during the transfer. The cold injector
needle containing the sample was then attached to the chromatograph.
The sample was subsequently flash vaporized and injected into the
chromatograph with carrier gas* Details of the elution and
injection apparatus are included in Appendix Co
The sequence of the collection and transfer steps was found
to be reasonably reproducible0 Reproducibility was tested by
preparing a synthetic mixture of organic vapors in air, including
both polar and non-polar species to represent a portion of the
range of component types in diesel exhaust, Samples of the mixture
were collected in high capacity C-102 samplers, then eluted,
injected,, and analyzed under the conditions used for exhaust
samples. Table I indicates the reproducibility obtained. Using
the same synthetic mixture, the efficiency of sample recovery in
the 60-min elution period used for exhaust samples is shown to be
90% or better in Table 11= Injections of diesel samples after
collection in the injector needle consistently proved to be
95-98% complete.
NT RESEARCH INSTITUTE
9
-------
Table I
SAMPLING REPRODUCIBILITY
GAS CHROMATOGRAPHIC ANALYSIS OF THREE SAMPLES
COLLECTED FROM THE SAME SYNTHETIC MIXTURE OF VAPORS
IN HELIUM
Column: 200-ft Carbowax 20M support-coated, open tubular
Analyzed at 2°/min temperature rise
Amounts Found
(in peak areas
Sample
Vapor
Heptane
Dodecane
Tetradecane
1-Butanol
1-Hexanol
2-Octanone
Propanal
Butanal
Ethyl Butyrate
1
3560
22
49
56
591
115
52
11
696
2
3710
21
50
57
602
112
50
10
737
3
3700
20
46
59
584
120
50
Flow Stopped
690
Table II
HIGH CAPACITY CHROMOSORB 102 NEEDLE COLLECTORS
EFFICIENCY OF SAMPLE RECOVERY PROCEDURES
Compound
Heptane
Dodecane
Tetradecane
1-Butanol
1-Hexanol
2-Octanone
Propanal
Butanal
Ethylbutyrate
1-Octanol +
Hexadecane
Consecutive 30-Min,
Sample Transfers
Peak Areas, cm^
3564,0
22*4
49.0
56.3
591.4
115 = 2
53,6
10,9
696,3
294,4
7,7
0
0
44.8
19,2
0
0
56,0
18,6
4,2
0
0
50.0
16.0
0
0
12.0
133,4 19,2 15.7
% Recovery
in 60-Min.
99.1
88.0
100.0
100,0
92.7
89.4
100.0
100.0
98.4
98.0
NT RESEARCH INSTITUTE
10
-------
D. Gas Chromatoqraphic Equipment and Techniques
The study of diesel exhaust for the objective of elucidating
odorous components was conducted principally through the
application of high resolution gas chromatographic techniques,
coupled with provisions for odor analyses and mass spectrometric
investigations. Stationary phases of Apiezon L and Carbowax 20M
in high resolution columns were used individually in a single
column instrument and in unique combination for a two-column
chromatograph. The chromatographic techniques provided the
capability for sample component separation, thus eliminating the
need for additional sample pretreatment procedures.
The single column gas chromatograph was equipped with a
sniffing port and a flame ionization detector in parallel to
monitor split portions of the column effluent. Two support-
coated open tubular (SCOT) columns, 200' by 0.020" i.d., with
Apiezon L and Carbowax 20M stationary phases and a packed column,
4' by 3 mm o.d., containing 60/80 mesh Chromosorb 101 were
applied to diesel exhaust samples.
Apiezon L is a non-polar phase and retardation of sample
components in the column is determined primarily by their boiling
points. Retention time in the polar Carbowax 20M column is
determined by uoth the boiling point and the polarity of the
components, that is, the interaction of their functional groups
with the column material. Thus, on the Carbowax 20M column, polar
species are retained longer than non-polar species having similar
boiling points. In general, a polar phase such as Carbowax 20M
2
is better suited to samples containing oxygenated species.
The Chromosorb 101 is a non-polar solid adsorbent, a styrene-
divinylbenzene polymer resin. It will chromatograph well the
lower free fatty acid and phenols. The free acids cannot be
chromatographed 0:1 either Carbowax 20M or Apiezon L,and phenols will
2
A.B. Littlewood, "Gas Chromatography, " Academic Press, New York,
N.Y., 1962, p. 427.
IIT RESEARCH INSTITUTE
11
-------
not chromatograph on Carbowax 20M. ' Thus, the Chromosorb 101
column was applied to exhaust samples to specifically investigate
the presence and odor relevance of these two component types.
A diesel exhaust sample was also analyzed on a single
column chromatograph equipped with a dual detector system. A
fl-^.me ionization detector operated in the normal mode and a
flame photometric detector specific to sulfur compounds monitored
split portions of the effluent from a 20', 0.125" i.d. packed
column of 20% Carbowax 20M on 60/80 mesh Chromosorb P.
The two-column gas chromntograph, which was used for the
major portion of the exhaust study, consists basically of two
high resolution columns of different polarity connected in series,
with the option of selecting the particular portions of the first
column effluent to be injected onto the second column. This
instrumental technique affords two interrelated advantages. The
first is the increased resolution, particularly useful for the
diesel exhaust samples under study that produced multicomponent
peaks even when chromatographed through a high resolution column.
The second advantage is the information provided on the nature of
individual components from the resultant retention times on the
polar - non-polar column set. The unique advantage is that each
component can be rechromatographed as it is eluted from one
column. In addition, the system is designed to include odor
analysis of the effluent from each column for the specific purpose
of locating and isolating the odor relevant constituents of the
ibid., p. 414 ff, p. 436 ff.
4
H.P. Burchfield and Eleanor E. Storrs, "Biochemical Applications
of Gas Chromatography," Academic Press, New York, N.Y., 1962,
p. 271 ff, p. 345 ff, p. 527 ff.
Sam S. Brody and John F. Chaney, J. of Gas Chromatoaraphv.
Feb., 1966, 42-46. ~ tL~JL
NT RESEARCH INSTITUTE
12
-------
sample. Dual sets of parallel sniffing ports and flame
ionization detectors are provided.
Most of the diesel exhaust samples were analyzed in the
two-column instrument using a 200', 0.020"i.d., Carbowax 20M
SCOT column as the primary column with a 50', 0.020" i.d.,
Apiezon L SCOT column as the second column. Several samples
were analyzed with a reverse column set. The primary column
was replaced with a 200', 0.020"i.d., Apiezon L SCOT column and
the second column changed to a 50' Carbowax 20M SCOT column,
0.020" i.d.
Specific descriptions of the gas chromatographic equipment
and techniques are covered in Appendix D. For all chromatographic
analyses of diesel exhaust samples on Carbowax 20M, Apiezon L,
and the two column sets, the column temperature was programmed
at 2°C/min from 40° to 180°C and isothermal at 180°C thereafter.
In order to correlate the resultant data on diesel exhaust
samples for the entire period of study from all the chromaLographs
used, and to eliminate the effects of slight variations in
chromatographic conditions from one analysis to another, the Kovats
retention index system was applied to the chromatographic data.
The Kovaf~ Index system, described in Appendix E, relates the
retention tin,-? of a component to a scale determined by the
retention times of the n-alkanes. This method of dealing with
chromatographic data was particularly suited to diesel exhaust
samples since the n-alkanes from hexane through hexadecane were
confirmed present in the samples. Thus, the reference points of
the Index Scale were available as internal standards. For indexing
above hexadecane, n-alkane standards were added to the samples.
From the n-alkane reference points, under the linear temperature
program conditions used, a calibration curve was obtained. The
Kovats Indexes of all other sample peaks were interpolated from
this curve using computer programs also described in Appendix E.
JIT RESEARCH INSTITUTE
13
-------
E. Gas Chromatoqraphic - Mass Spectrometer Interfacing
For the task of identifying odor relevant components in the
exhaust by mass spectrometry, a GC-MS interfacing device was
developed for the particular requirements of the diesel exhaust
sample. Because of the complexity of the sample, the two-column
chromatograph was needed to provide sufficient resolution of
components by rechromatographing odorous Carbowax 20M peaks on the
Apiezon L column. The resultant Apiezon L chromatograms of
individual Carbowax 20M peaks generally showed several components,
particularly many aliphatic hydrocarbons, but sufficient separation
of the one or more odorous species in the peak was achieved. Only
these odor-relevant components were required for mass spectrometric
analysis.
The entire interfacing device was designed to incorporate the
following features:
1. The capability of rejecting peaks in the Apiezon L
chromatograms, trapping only the odorous peaks of
interest.
2. Discrete stations for cryogenically trapping individual
peaks and later releasing the peak for mass spectrometric
analyses by heating the same trapping area.
3. A small volume system, relying only on the pumping speed
of the mass spectrometer to transfer trapped components
from the interface to the mass spectrometer. Because
the distance travelled at the low leak rate diffused the
trapped components during transfer, the peaks were
retrapped at the entrance to the ion chamber in order to
inject them into the MS in good peak form,
4. A closed system when attached to either the gas chrom-
atograph or the mass spectrometer, avoiding any
contamination of sample during the process of detachment
from the former and connection to the latter.
A detailed description of the design and operation of the
interfacing assembly is presented in Appendix F. The mass
spectrometer used in the diesel exhaust study was the Hitachi
Perkin-Elmer magnetic scanning Model RMU-6D.
IIT RESEARCH INSTITUTE
14
-------
F. Experimental Basis for Sample Component Characterization
The purpose of this diesel exhaust study was to elucidate and
characterize those components in the exhaust which are odor
relevant. The question of odor relevancy hinged on the effectiveness
of the entire sampling procedure, so that the odor judgements made on
sample components by monitoring the chromatograph sniffing ports
could be related back to odor relevancy in the original exhaust.
Exhaust sampling required the total collection sampler and efficient
transfer and injection techniques described in the above sections.
The following calculation demonstrates the relation between sniffing
port concentration and original exhaust concentration for a
component using a 10 liter diluted exhaust sample collected on the
Chromosorb 102 sampler and analyzed on the two-column gas
chromatograph.
A component in the original exhaust at a concentration of
w g/ml results in a sample concentration of 910w g when 10 liters
of 11:1 diluted exhaust are collected. Assuming a 90% efficiency
for sample transfer and injection and considering the sample
splitting in the chromatograph, which delivers 25% of the injected
sample to each sniffing port, then approximately 200w g of the
component reaches the sniffing port. This concentration elutes
from the chromatograph in 30-40 seconds (average peak width) with
30-40 ml of the helium-humidified air mixture (60 ml/min flow).
Thus, the odor of the component is monitored at a concentration of
5 w g/ml, or about five times the concentration in the original
exhaust. Frcru detector sensitivity calibrations, any component
occurring in the exhaust at concentrations greater than
1 x 10-13 g/ml will produce a noticeable peak on the chromatograms
from a 10 liter sample of diluted exhaust.
With the two-column instrument, equipped with a sniffing port
at each column exit, components in the exhaust sample that were
odor relevant could be isolated, Odorous areas in the Carbowax 20M
chromatogram were rechromatographed on the second column and the
odorous components identified from among the Apiezon chromatogram
peaks. In addition, inherent in this isolation process is a
partial characterization of the unknown sample components.
Obtaining retention data for a component on two different
stationary phases is the basis of the characterization technique.
IIT RESEARCH INSTITUTE
15
-------
The retention time of a component on Carbowax 20M is determined
by the combination of a component's boiling point and polarity,
whereas the Apiezon L retention time is determined primarily by
boiling point only. In effect, the amount of retardation in the
second Apiezon column will define the degree of polarity of a
component. When a multicomponent peak in the Carbowax 20M effluent,
which may be composed of a range of polar, lower boiling components
to non-polar, higher boiling components, is injected onto the
second column, the Apiezon L dispersion will be a functional group
separation. The two-column data obtained permits an estimation
of the retention index difference, Al, of Kovats Index notation.
By the thermodynamically based Kovats Index rules, the £1 is
characteristic of the structure of a substance, the combined
effect of the functional group and skeletal structure of the
molecule.
In order to calibrate the polarity dispersion resulting from
the two-column technique, several types of standards were run under
the same experimental conditions used for sample analysis. Table
III lists the two-column data for the calibration standards. The
retention time data on Carbowax 20M is converted to Kovats Indexes.
Conversion of the Apiezon L data to Kovats Indexes is not possible,
so the corrected retention times are used. The homologous series
of n-alkanes define the maximum retention times possible in
Apiezon L chromatograms throughout the range of the Carbowax 20M
chromatogram since no species could be more non-polar and thus
could not experience greater retardation. These n-alkane
retention limits define the time interval that must be maintained
between successive injections of Carbowax 20M effluent onto the
second column. After this delay time from the peak injection
point, all possible components have eluted from the Apiezon L
column. The homologous series of 1-alcohols represent the species,
among single functional group types, that would experience minimum
retardation. The presence of more than one functional group, of
course, will add to the polarity and thus reduce the retardation.
NT RESEARCH INSTITUTE
16
-------
Table III
TWO-COLUMN CHROMATOGRAPH CALIBRATION DATA FROM STANDARDS
Compound
Pentar.c
1-Pentene
3-Methylpentane
2-Pentene
Hexane
i-Hexene
2, 1, 4-TrimeHiyl ppnfnnfi
2-Hexene
Heptane
Cyclohexane
DimethyIhexane
1-Heptene
3-Heptene
2,3,4-Trimethylpentane
2,3,5-TrimethyIhexane
Propanol
Octane
1,2-Epoxybutane
1-Octene
2-Methyltetrahydrofuran
2-Ethyl-l-hexene
Methanol
2-Octene
Butanal
2-Propanol
Die thyIsulfide
2-Wutanone
2,5-Dihydrofuran
Nonane
Tetrahydropyran
n thanol
2-Methyltetrahydropyran
Benzene
Dihydropyran
1-Nontne
L>_ nyl -n-propanoate
i-Pentanone
Pentanal
Thiophene
Decane
'itiiyl-n-butan <-ite
Toluene
2-Butenol
n-Propy1-n-propanoc c
2-Oci.yne
1 -Det:ene
Di-n-propylsulfide
H-xanal
P^raldehyde
Undecane
n-Propy1-n-butanoate
Ethylbenzene
1-Butanol
Ethyl-n-pentanoate
p-Xylene
Nitromethane
m-Xylene
2-Heptanone
Heptanal
Kovats Retention
Index Time8
Carbowax Apiezon L
20M (mini
500
536
570
571
600
642
65.1
670
700
719
723
740
743
756
764
792
800
826
830
858
860
862
870
871
883
889
893
897
900
910
912
918
920
925
931
946
966
970
984
1000
1023
1025
1027
1033
1034
1035
1060
1070
1075
1100
1110
1112
1118
1123
1124
1130
1130
1170
1175
0.37
0.30
0.62
0.43
1.03
0.85
2.11
1.07
2.50
2.41
3.15
2.10
2.33
3.72
4.04
0.40
4.97
0.64
3.95
1.22
4.09
0/34
4.10
0.60
0.29
1.35
0.44
0.69
.40
.53
.43
7.
1.
0.
2.00
1.12
1.21
5.43
0.76
0.70
0.79
0.96
8.86
1.16
1.72
0.48
1.30
2.50
6.38
2.74
1.18
0.91
9.40
1.69
2.14
0.55
1.53
2.32
0.09
2.27
1.30
Compound
Kovats
Index
Carbowax
20M
1.41
Methyl-n-hexanoate 1175
Cyclopentanone 1176
o-Xylene 1176
Dodecane 1200
n-Propyl-n-pentanoate 1209
Ethyl-n-hexanoate 1223
1-Pentanol 1225
1,3,5-Trimethylbenzeneb 12 27
1-methyl-2-ethylbenzeneb 1242
1-methyl-3-isopropylbenzene" 1248
1-methyl-4-isopropylbenzeneb 1248
Di-n-butylsulfide . 1257
1,2,4-trimethylbenzene 1262
2-Octanone 1275
Methyl-n-heptanoate 1276
Octanal , 1280
l,3-DiethylbenzeneD 1281
1-Methyl-3-n-propylbenzeneb 1284
Cyclohexanone . 1288
l-Methyl-4-n-propylbenzene 1289
l,4-Diethylbenzeneb 1297
Tridecane 1300
l,3-Dimethyl-5-ethylbenzeneb 1300
n-Propyl-n-hexanoate 1308
1,2,3-Trimethylbenzene 1320
Ethyl-n-heptanoate 1323
1-Hexanol 1331
4-Methylcyclohexanone 1338
l,l-dimethylindanb 1363
Cyclohexanol 1380
Methyl-n-octanoate 1380
Nonanal 1387
Tetradecane 1400
n-Propyl-n-heptanoate 1409
Ethyl-n-Octanoate 1425
1-Heptanol 1438
2-Furaldehyde 1467
Methyl-n-nonanoate 1483
Decanal 1494
Pentadecane 1500
n-Propyl-n-octanoate 1511
Furfurylacetate 1520
Benzaldehyde 1527
Ethyl-n-nonanoate 1527
1-Octanol 1547
Methyl-2-furoate 1573
Tetrahydrofurfurylacetate 1573
Methyl-n-decanoate 1587
l,l,4,6-Tetramethylindanb 1587
Undecanal . 1600
Hexadecane 1600C
Tetrahydrofurfurylpropionate 1643
p-Tolualdehyde 1658
Acetophenone 1659
2,6-dimethyltetralinb 1669
Heptadecane 1700
Octadecane 1800
2,2/5,7-Tetramelhyltetralinb 1811
2,5,0-Trimethyltetralinb 1884
Nonadecane 1900
Retention
Time0
Apiezon L
(min)
1.46
0.75
2.20
9.61
2.05
1.88
0.80
2.60
2.40
2.77
2.92
3.72
2.57
.58
1.76
1.72
2.85
2.87
1.02
2.92
2.90
9.83
2.80
2.35
2.39
2.18
1.03
1.22
2.90
0.81
2.02
2.00
10.03
2.63
2.49
1.23
0.63
2.29
2.27
10.28
2.91
0.65
0.74
2.75
1.44
0.55
1.05
2.53
3.41
2.57
11.74
1.13
0.96
0.83
3.61
15.8
24.0
6.10
6.69
37.8
Apiezon L retention time is corrected for column dead volume.
Hydrocarbon standards supplied by Mobil Research and Development Corporation.
After Kovats Index 1600 the temperature program has ended and the temperature is maintained at 180°C.
17
-------
For example, 2-butenal, at Kovats Index 1027, with the addition of
a double bond functional group to the aldehyde functional group
shows the effect of increased polarity over n-butanal, Kovats
Index 871, by its later elution from Carbowax 20M and its shorter
retention time in Apiezon L.
Figure 1 shows some of the calibration data in a two
dimensional plot that illustrates how the information points of a
Kovats Index in the first column and a retention time in the
second column can result in structure information. "The connected
data points delineate particular functional groups areas throughout
the range of the Carbowax 20M chromatogram. Vertical columns
would represent Apiezon L chromatogratus (or the dispersion) that
would be obtained by switching onto the second column portions of
the Carbowax 20M effluent containing components having a retention
time on Carbowax 20M identified by the Kovats Index at the base of
the column.
Although polarity characterization from the two-column method
was stressed, it is also true that other information is inherent
in the data obtained and can be correlated with the two-column
data. Other molecular properties, such as boiling points, could
be represented as the third dimension in Figure 1 and also applied
to unknown sample analysis. Iso-boiling point lines are shown in
Figure 1. Appendix G describes the investigation of physicochemical
properties taken for correlation analyses using the standard data
of Table III.
A portion of the exhaust sample work on the two-column
instrument was performed with Apiezon L as the primary column and
Carbowax 20M as the second column. Although the non-polar phase
is less suitable for samples containing oxygenated species, as
cited in Section II D, the change in elution order of sample
components provided an "odor chromatogram" for comparison with the
odor notes from the analysis with Carbowax 20M as the primary
column. On Apiezon L, more polar species were eluted before the
major portion of aliphatic hydrocarbons in the sample. However,
IIT RESEARCH INSTITUTE
18
-------
Approximate column temperatures and Carbowax 20M retention times associated with the
designated Kovats Indexes:
Column temperature, °C
53
I
58
I
69
I
83
>9
I
115
130
145
158
170
Carbowax 20M retention time (min) under temperature rise of 2°C/min from 40° to 180°C
6.5 9.1 14.4 21.5 29.3 37.4 45.2 52.5 58.9 65.1 71.1
1
z
H
Z
UJ
H
Q
UJ
0
E
^b
z1
^
14
, 12
UJ
(-
i |0
z
-1 8
1
UJ 6
Q.
^
_
4
1
9
i.
i i i ~r -r
/
n-ALKANES x . -^
x ^_ - .
---"""i T '. » ..'.: ! i » '.» '
***^ !*" * J
J**" .* *»^ ^ * t» I*.
^ ' ' I ***»'* I t J
x"» « . . . H :i .
x . . _- : .
X ^-"""
/ --* "^ . lj
/ * ,**" ^v, . .
"" t. -x*
-------
the retention dispersion obtained on Carbowax 20M for peaks
injected from the Apiezon L effluent was much poorer than that
obtained with the original column sequence. Table IV lists
calibration data on some n-alkanes and n-aldehydes as examples
of the range of retardation for non-polar vs. polar species on
the secondary Carbowax 20M colunn. Comparison of this data with
the data on the same compounds in Table III demonstrates that the
dispersion possible on this column set is not sufficient either
for definitive assignments of polarity ranges for component
characterization or for the resolution needed for mass spectro-
metric studies.
HT RESEARCH INSTITUTE
20
-------
Table IV
RETENTION DISPERSION ON CARBOWAX 20M
FOR COMPONENTS ELUTED FROiM APIEZON L
Retention Time
Apiezon L, min
Pentane
Propanal
Butanal
Hexane
Pentanal
Heptane
Octane
Heptanal
Nonane
Decane
Undecane
Dodecane
Tridecan<~
Tetradecane
Pentadecane
Hexadecane
7.1
7.1
10.4
12.2
18.1
20.8
31.5
39.7
42.4
52.9
62.6
71.8
81.6
95.1
115.8
148.6
Kovats Index
Apiezon L
500
500
572
600
672
700
800
875
900
1000
1100
1200
1300
1400
1500
1600
Retention Time*
Carbowax 20M,
min
0.1
1.0
1.2
0.1
1.4
0.1
0.1
1.0
0.1
0.1
0.1
0.1
0.15
0.25
0.5
0.8
Retention time is corrected for column dead volume.
Ill RESEARCH INSTITUTE
21
-------
IV. RESULTS AND DISCUSSION
A. Single Column Chromatograph Studies
The analysis of diesel exhaust samples on the single column
chromatograph concentrated on odor characterization. During
sample analysis, an observer monitored each effluent peak for
odor. The observers were familiarized with the Turk diesel
reference odors and were asked to adhere to this classification
(aldehydic, pungent, oily, burnt), if possible, in their odor
characterizations. Since a number of observers contributed to
the diesel sample odor characterization, some odor quality ref-
erences were necessary in order to correlate the odor data. The
majority of the odor annotations were covered by the Turk clas-
sifications, although an observer often characterized an
"identifiable" odor through his own experience or frame of ref-
erence.
One observer was used for an entire sample chromatogram,
usually a 70 min analysis extending to the end of the tempera-
ture program and encompassing a Kovats Index range through 1600
(hexadecane). A series of samples collected on both Apiezon L
fluidized beds and SF-96 packed beds were chromatographed on
both the Carbowax 20M column and then the Apiezon L column in
the single column chromatograph. Figure 2 presents a typical
chromatogram from the Carbowax 20M column of an Apiezon L fluid-
ized bed sample. A typical Apiezon L chromatogram, also from a
sample collected on a fluidized bed, is shown in Figure 3.
For the Carbowax 20M chromatograms, the Turk classifica-
tions covered about 91% of all peaks for which a descriptive
term was used by the observers. For the Apiezon L column analy-
ses, about 62% of the descriptive terms were Turk annotations.
Many peaks were simply labeled "odorous" because the rapid elu-
tion of peaks did not allow much time for each individual odor
Amos Turk, "Selection and Training of Judges for Sensory Evalu-
ation of the Intensity and Character of Diesel Exhaust Odors,"
Public Health Service Publication No. 999-AP-32.
(IT RESEARCH INSTITUTE
22
-------
15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72
Retention Time, min
Engine Run No.4 Dec.20.1968. Fluidized Bed S-4
Fig. 2 TYPICAL CARBOWAX 20M CHROMATOGRAM OF DIESEL ENGINE EXHAUST
-------
12 15 18 21 24 27 30
33 36 39
Retention Time. min.
Engine Run No. 8. Feb.2,1969. Fluidized Bed No. 4
42 45 48
54 57 60 63 66 69 72
Fig. 3 TYPICAL APIEZON L CHROMATOGRAM OF DIESEL ENGINE EXHAUST
-------
judgment. The frequency with which the various descriptive
terras were used to characterize the odor quality of the odorous
peaks is presented in Table V.
The odor characterization data from the single column
Carbowax 20M and Apiezon L chromatograms are summarized in
Tables VI and VII, respectively. The summary divides the chro-
matograms into small Kovats Index ranges. The designated
odorous areas of the chromatograms served as reference points
for the further isolation and characterization of odorous com-
ponents using the two-column chromatograph.
B. Initial Two-Column Chromatoqraph Studies
on Fluidized Bed-Collected Samples
Most of the diesel exhaust samples initially analyzed with
the two-column instrument were collected on Apiezon L fluidized
beds. The primary column Carbowax 20M chromatograms showed
good reproducibility both among different bed samples from the
same engine test and among different engine tests. The repro-
ducible nature of the sample was a required condition for the
systematic investigation of the sample with the two-column sys-
tem became: (1) it was necessary to predetermine the effluent
peaks to be ^witchedonto the second column (split portions of
the first column effluent simultaneously entered the second
column and reached the first column detector); and (2) only
five to seven peaks could be switched per sample run to obtain
complete Apiezon L chromatograms of each peak, so many samples
had to be analyzed. Figure 4 shows a typical two-column chro-
matogram from a fluidized bed-collected exhaust sample. Shaded
areas on the upper Carbowax 20M chromatogram indicate the peaks
injected onto the Apiezon L column. The extent of each
Apiezon L chromatogram is marked from the point of injection to
the point of maximum permissible elution time for each injec-
tion. Odorous peaks on the Apiezon L chromatograms are also
indicated. Three sets of duplicate Apiezon L chromatograms are
shown in Figure 5, representing Kovats Index peaks at 1420,
HT RESEARCH INSTITUTE
25
-------
Table V
FREQUENCY OF ODOR DESCRIPTIONS
Single Column Data
DescriEtiveJTerrn C2QM APL
aldehydic 20.2 10.4
pungent 15.0 21.7
burnt 16.3 11.0
oily 10.2 6.5
smoky 1.6 1.2
musty 0.2 3.5
sour 1.2 0.2
sweet, pleasant 1.4 2.0
foul, unpleasant 2.0 8.0
paraffinic, naphtha, etc. 0.8 3.0
diesel 0.2 0.1
grassy 0.0 1.5
minty 0,0 0.5
misc. odors 0,0 9.5
uncharacterized odors 30.9 20.8
a conglomeration of many descriptive terms
such as paint, plastic, latex, vinyl, soap,
fruity, fishy, leather, camphor and others.
2
Frequency of term usage to describe odorous
peaks eluted from indicated chromat.ographic
partition column.
26
-------
Table VI
ODOR CHARACTERIZATION 3Y KOVATS INDEXES
SINGLE 200-ft C20M COLUMN
Kovats
700- 720
721- 740
741- 760
761- 780
781- 800
801- 820
821- 840
841- 860
861- 880
881- 900
901- 920
921- 940
941- 960
961- 980
981-1000
1001-1020
1021-1040
1041-1060
1061 -1080
1081-1100
1101-1120
1121-1140
1141-1160
1161-1180
1181-1200
1201-1220
1221-1240
1240-1260
1261-1280
1281-1300
1301-1320
1321-1340
1341-1360
1 361-1380
1331-1400
1401-1420
1421-1440
1441-1460
146] -1480
1481-15CD
1501-1520
l',2 L-1540
Ib41-1560
1561-1580
1581-1600
Number
Aid.
0
2
0
0
1
3
0
1
1
1
2
4
2
4
1
0
3
2
3
3
4
0
3
3
3
1
4
4
4
10
6
2
2
1
0
6
4
0
0
2
2
4
0
0
2
of Times
Puna.
2
1
0
0
2
1
1
0
0
0
1
0
2
2
1
1
0
2
5
n
i
0
2
1
2
1
3
1
2
2
2
3
1
4
1
2
3
0
0
1
1
1
4
1
1
Descriptive
Term
Burnt Oily
1
0
1
0
0
1
0
0
0
1
1
1
3
0
2
1
3
1
1
1
0
0
0
0
0
4
5
1
0
2
0
2
3
3
6
3
1
4
4
4
4
4
5
3
3
1
0
1
1
1
0
2
1
0
0
1
0
0
3
1
0
0
0
0
4
2
0
2
1
3
0
0
0
3
1
2
0
2
3
2
1
1
4
1
0
1
2
0
0
6
'Jsed
Foul
0
0
0
0
0
0
0
0
0
0
0
0
0
o
2
0
0
0
0
0
0
0
0
2
0
0
0
2
0
1
1
0
0
0
0
0
0
0
0
0
0
3
0
0
1
27
-------
Table VII
ODOR CHARACTERIZATION BY KOVATS INDEXES
SINGLE 200-ft APL COLUMN
Kovats
Index
700- 720
721- 740
741- 760
761- 780
781- 800
801- 820
821- 840
841- 860
861- 880
881- 900
901- 920
921- 940
941- 960
961- 980
981-1000
1001-1020
1021-1040
1041-1060
1061-1080
1081-1100
1101-1120
1121-1140
1141-1160
1161-1180
1181-1200
1201-1220
1221-1240
1241-1260
1261-1280
1281-1300
1301-1320
1321-1340
1341-1360
1361-1380
1381-1400
1401-1420
1421-1440
1441-1460
1461-1480
1481-1500
1501-1520
1521-1540
1541-1560
1561-1580
1581-1600
Number
Aid.
1
8
4
11
0
2
12
4
16
6
4
12
3
13
4
4
3
4
11
1
1
0
6
2
2
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
of Times
Punq.
5
5
8
7
8
7
6
12
4
8
15
2
11
9
10
7
7
12
6
3
7
13
10
8
8
6
5
7
2
4
4
2
0
1
2
1
1
0
3
0
0
0
0
0
0
Descriptive Term
Burnt
1
1
2
0
0
1
0
4
2
1
1
2
2
2
2
1
4
3
14
5
4
8
8
6
9
2
3
4
3
3
0
0
0
0
1
0
0
1
O
0
0
0
0
0
0
Oily
0
1
8
1
2
0
0
2
2
3
1
0
1
0
1
1
2
0
1
1
2
2
0
5
2
2
6
3
1
2
1
1
1
1
2
0
0
0
0
1
0
0
0
0
0
Used
Foul
0
0
0
2
0
1
0
3
1
5
0
0
8
4
1
0
5
14
4
0
1
7
1
3
3
3
0
1
2
0
1
1
0
0
0
0
1
0
0
0
0
0
0
0
0
28
-------
659
969
VD
KOVATS INDEX
1119
1255
M34
1690
75
SA
TIME. MINUTES
ENGINE RUN No. 11. FEa 24.1969. FLUIDIZED BED NaT
A ALDEHYDIC
VA UNPLEASANT ALDEHYDIC
SA STRONG ALDEHYDIC
SPA SLIGHTLY PUNGENT ALDEHYDIC
tftf BURNT. DIESEL-LIKE
D DIESEL-LIKE
SP SLIGHTLY PUNGENT
SW SWEATY. ALDEHYDC
Fig. 4
TYPICAL DUAL-COLUMN CHROMATOGRAM OF DIESEL ENGINE EXHAUOT
-------
1-0
C20M KOVATS
INDEX PEAK 1420
EXHAUST SAMPLE A
C20M KOVATS
INDEX PEAK 1434
EXHAUST SAMPLE B
I'D
C2OM KOVATS
INDEX PEAK 1644
*- 0
0
«- 0
APIEZON L RETENTION TIME
Fig. 5 REPRODUCIBILITY OF DUAL-COLUMN CHROMATOGRAPH
SINCE THE EXHAUST SAMPLES WERE COLLECTED ON DIFFERENT
DAYS, THE DATA PRESENTED ILLUSTRATES THE EXCELLENT
REPRODUCIBILITY OF ALL PROCEDURES FROM ENGINE OPERATION
TO CHROMATOGRAPH 1C PERFORMANCE, INCLUDING THE SAMPLING,
TRANSFER AND INJECTION PROCEDURES
-------
1434, and 1646 on Carbowax 20M. Members of each pair represent
different fluidized bed samples from different engine tests to
demonstrate that reproducible exhaust samples were obtained.
The two-column analysis of odorous components from fluid-
ized bed samples concentrated on those species which appeared on
the Carbowax 20M chromatograms between Kovats Indexes 600 (hex-
ane) and 1600 (hexadecane) . The analyses aimed at finding on
the Apiezon L chromatograms the single peak producing the same
odor in quality and intensity as the Carbowax 20M peak, if such
a component w&re present, or else noting the odor-contributing
peaks which appeared. Every peak rechromatographed resulted in
a multicomponent Apiezon L chromatogram, including, in almost
all cases, several aliphatic hydrocarbon components. The sys-
tem effectively separated these hydrocarbon species, which were
determined to be individually odor irrelevant, from the odor-
ous components that occurred in low Apiezon L retention time
areas.
A partial characterization of the nature of the exhaust
sample is evident from a plot of some of the two-column data in
a polarity dispersion graph of Carbowax 20M Kovats Indexes
against Aniezon L retention times as shown in Figure 6. The
points in e^ch vertical column are the Apiezon L chromatogram
peaks resultiig from the Carbowax 20M peak identified by the
Kovats Index at the base of the column. The odor relevant
Apiezon L peaks are indicated. Dashed lines representing the
calibration data from Figure 1 identify some functional group
retention areas. The complexity of the exhaust sample is
apparent. Figure 6 also graphically illustrates the capability
of the two-column chromatograph technique for analyzing a sam-
ple of such complexity.
From the two-column studies, the total number of compo-
nents was estimated to exceed a thousand. Those components
which are odor relevant constitute a very small portion of the
diesel exhaust. Table VIII lists all the odorous Carbowax 20M
NT RESEARCH INSTITUTE
31
-------
Approximate column temperatures and Carbowax 20M retention times associated with the
designated Kovats Indexes:
Column temperature, °C
53 58 69 83 99 115 130
1 1 1 1 1 1 1
Carbowax 20M retention time (min) under temperature rise
6.5 9.1 14.4 21.5 29.3 37.4 45.2
14
i
p 12
U
0 £
2 i»
2 s.
y -J 8
^?
o o
J 1 1 1 C
1^ SB w
U 0-
Ul ^
| 2 4
O
3
2
.
:
n-ALKANES v
\
-' """ f * : : *. » ..'.: I
_ ^- * * * i * ***
si" %.. * :;*.. ;
- S" *. : * * i ~ * - ..."
/ . ,* * ; .'>.: I i
s * **" i^ " "****.
/ ^-t-*'"""^* I -;
/ *^^
-^ * *-'*' "*^\
XX
-------
Table VIII
SUMMARY OF ALL ODOROUS PEAKS ON PRIMARY COLUMN CAR BOW AX 20M CHROMATOGRAM
DUAL-COLUMN CHROHATOGRAPH
Kovato Index
Carbowax 2QM
723
746
787*
BIS
852*
857
864
875
879
895
934
953
959
963*
969
977
1024
1031
1035
1039
1066
1073
1078*
1095
1100
1116
1119
1123
1148
1152
1159
1161
1169
J..71*
118v
1186
1192
1194
1203
1214
1218
1225
1231
1237
1239
1245
1249
1255
1260
1269
1275
1281*
12B8
1295
1300
1308
Odor Note
alcohol
alcohol
sweet-alcohol
sweaty
sweaty aldehyde
sweaty aldehyde
sweaty
sweaty
unpleasant
odor
very unpleasant, "foul"
unpleasant aldehyde
unpleasant aldehyde
unpleasant aldehyde
unpleasant aldehyde
unpleasant aldehyde
pleasant
sweaty aldehyde
sweaty aldehyde
sweaty
pleasant
aldehyde
strong aldehyde
odor
odor (dlesel, oily, aldehyde)
aldehyde
aldehyde
oily aldehyde
odor
pleasant
unpleasant aldehyde
unpleasant
very "foul"
strong aldehyde
aldehyde
burnt, unpleasant
pungent
burnt
odor
pleasant
pleasant, aldehyde, pungent
burnt aldehyde
pungent aldehyde
strong aldehyde
aldehyde
odor, diesel-lilce
aldehyde
aldehyde
pungent
very "foul"
odor
strong aldehyde
sveet
unpleasant, burnt, "diesel"
odor
pungent aldehyde
Kovats Index
Carbowax 20M
1316
1319
1324
1329
1331
1339
1342
1347
1355
1357
1363
1369
1373
1379
1382
1387*
1395
1400
1403
1407
1420
1425
1434
1442
1445
1455
1457
1464
1469
1478
1484
1489
1494*
1500
1504
1510
1518
1525
1532
1535
1552
1555
1560
1567
1572
1578
1563
1568
1594
1600
1613
1620
1627
1630
1637
1644
1657
Odor Note
odor
burnt, pungent, oily, aldehyde
burnt
strong burnt
pungent burnt
sweet , pungent
aldehyde
burnt pungent
burnt, pungent, aldehyde
burnt
odor
aldehyde
burnt, pungent, aldehyde
odor
unpleasant burnt
sharp aldehyde
unpleasant burnt
unpleasant burnt
burnt "dlesel"
pungent burnt
pungent burnt
pungent
burnt, "dlesel", pungent
very pleasant
very pleasant
burnt
burnt
unpleasant, "foul"
odor
unpleasant burnt
unpleasant burnt
burnt
pungent aldehyde
burnt, "dlesel", oily
burnt, "dlesel"
odor
slight burnt
slight burnt
pungent burnt
pungent aldehyde
pungent burnt
burnt
pungent
slight burnt
sharp burnt
burnt, pungent, aldehyde
burnt pungent
burnt pungent
burnt
burnt, "diesel"
pungent aldehyde
pleasant
pungent
slight burnt
burnt pungent
burnt
very pleasant
*n-aldehyde series, propanal through decanal, identified
by comparison of their Kovats Indexes on C20M and their
retention time in APL with calibration data.
33
-------
peaks compiled from the fluidized bed samples analyzed during
this period of study.
The largest group of components in terms of concentration
were the n-alkanes. Peak sizes indicated that Cg (octane)
through C , (tetrsdecane) contributed the major concentrations,
with the maximum around C ^ and C12* In terms of odor contribu-
tion, the most important group, in number and in odor intensity,
was the aldehydic odor type. From Kovats Index rules and cali-
bration data, the homologous series of n-aldehydes from propanal
through undecanal were identified. Other aldehydic odor peaks
were found clustered around the n-aldehyde peaks, and a second
aldehydic homologous series was suggested. The samples analy-
zed with the Apiezon L-Carbowax 20M reverse column set confirmed
the presence of the n-aldehydes, which were prominent features
of the primary column Apiezon L chromatogram. Later studies on
Chromosorb 102 collected samples, which more accurately pre-
sented exhaust species in their actual relative concentrations,
showed that the Apiezon L fluidized bed collected a larger rela-
tive concentration of aldehydic species. Since the Apiezon L
phase is a partition coefficient-dependent sampling medium, spe-
cies with more favorable partition coefficients are more
effectively collected. Although present in the exhaust, some of
these aldehydic species were determined to be odor irrelevant
from the Chromosorb 102 sample analyses. Among the odorous
aldehydes, positive identification of ethanal through octanal in
n-aldehyde series was obtained from mass spectral data.
The majority of exhaust and fuel related odors comprising
the designations "burnt", "pungent", and "diesell! occurred
above Kovats Index 1200 on Carbowax 20M. The odorous Apiezon L
components of these peaks appeared at low retention times, indi-
cating polar species. For many of the strong burnt odor peaks,
no odorous Apiezon L components could be found. The larger
exhaust samples later collected on the high capacity Chromosorb
102 samplers brought these species up to detectable concentra-
tions. The two "foul" components at Kovats Index 1169 and 1269,
IIT RESEARCH INSTITUTE
34
-------
on Table VIII, constituted another example of strong odor
contribution from low odor threshold species present in small
concentrations.
The diesel exhaust samples analyzed with the reverse col-
umn set, Apiezon L-Carbowax 20M, were of interest principally
for the odor notes from the Apiezon L chromatograms compared
with the Carbowax 20M chromatograms. A summary of the odorous
peaks is presented in Table IX. The aldehydic notes con-
firmed the data from the other column set. Above Kovats Index
1200, which represents the end of the temperature program, the
Apiezon L column effluent has a continuous background odor that
made peak odor annotations difficult. The background odor,
whether from the Apiezon L phase or from diffuse sample peaks,
could be described generally as "burnt pungent."
C. Chromatoqraphic and Mass Spectrometric Analyses
of Diesel Exhaust Samples
Collected on Chromosorb 102
The use of the small diameter Chromosorb 102 packed bed
samplers, which will completely extract almost all the
organic components from the exhaust except for the very light
species,permitted the evaluation of odor relevancy of the gas
chromatograpMcally separated components. As described in sec-
tion II-F, a 10 liter sample of diluted exhaust (0.9 liter
exhaust) presents the components at the gas chromatograph snif-
fing port at a concentration about five times the concentration
in the actual exhaust. The odor annotations and two-column
data from the fluidized bed samples were correlated with the
Chromosorb 102 samples. Because of the complexity of the sam-
ple and the multi-component nature of the peaks, the shape of
the Carbowax 20M chromatograms from the two types of collected
samples, of course, differed. However, individual components
occurred at the same Kovats Index area.
The mass spectrometric analysis of peaks trapped in the
interface unit from the Apiezon L column of the two-column
NT RESEARCH INSTITUTE
35
-------
Table IX
SUMMARY OF ALL ODOROUS PKAKS ON PRIMARY COLUMN APIEZON L CHROMATOGRAM
DUAL-COL'JMN CHROMATOGRAPH
Kovats Index
Api ezon L
463
500*
544
574*
583
iOO
'i04
R27
H34
337
852
fl74*
900
90C,
910
Oil
91P
920
931
013
-------
chromatograph started on fluidized bed collected samples, but
samples from the Chromosorb collectors, both 0.375" diameter
and high capacity 0.75" diameter packed beds, were used for the
major portion of the work.
A typiqal Carbowax 20M chromatogram from a 100 liter dilu-
ted exhaust (9.1 liter exhaust) sample collected on a high
capacity Chromosorb 102 sampler is shown in Figure 7 a-g as a
reference for the diesel exhaust data summary in Table X. The
weight of total organic content for this sample size,7 obtained
by electronic integration of Carbowax 20M peak areas through
Kovats Index 1800, was 0.2 mg. Peak No. 5 on Figure 7-a, ace-
tone, was exceptionally large on this particular sample set.
It was traced to a leak at the cold trap on the engine sampling
system. It does, however, demonstrate the collection capacity
of the Chromosorb.
The compilation in Table X, consisting of the odor rele-
vant sample components, lists the two-column data of Carbcv.'ax
20M Kovats Index and Apiezon L retention time, odor characteri-
zation, concentration estimate, and mass spectral identification
or species suggested by the mass spectral data if a positive
identification could not be established. No attempt was made to
gather comp1ete quantitative information on individual compo-
nents. The prime consideration was determining whether a
component was present in the exhaust above or below its odor
threshold, regardless of its absolute concentration. Some con-
centration estimates are given to show the amounts of different
types of species which are odor relevant. The estimates are
reliable to within an order of magnitude.
Of the odor relevant species identified, the aromatic
hydrocarbons appear to contribute many of the exhaust-related
odors. Although toluene, ethylbenzene, and the xylenes are
7
The total organic content of the sample is not the total
organic content of the exhaust. As explained in section II-B,
the sampler does not efficiently collect the very light hydro-
carbons, methane for example, which are assumed to be odor
irrelevant.
II T RESEARCH INSTITUTE
37
-------
X10240
OJ
00
»-
0
Kovats Index Scale
r
4
6 8
Retention Time (Min)
600 700
10
12
800
Figure 7 a
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED OH A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
u>
vO
x512
14
16
18
20 22
Retention Time (Min)
900
Kovats Index Scale
1000
24
i
26
Figure 7 b
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
x!28
x256
x64 x64
x!28
28
1100
i
30
32 34 36
Retention Time (Min)
38
40
Kovats Index Scale
1200
Figure 7 c
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
x32
1x128
47-48
42
44
1300
46 48
Retention Time (min)
50
Kovats Index Scale
52
1400
54
Figure 7 d
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
49
*>
to
x!6
x8
54
56
58
1500
60 62
Retention Time (Min)
Kovats Index Scale
64
1600
66
Figure 7 e
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
x32
*>.
U)
68
70
1700
72 74 76
Retention Time (min)
Kovats Index Scale
1800
78
80
Figure 7 f
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOS.ORB 102 SAMPLER
-------
x!6
67
68
82
84
1900
86 88
Retention Time (min)
Kovats Index Scale
90
92
2000
Figure 7 g
CARBOWAX 20M CHROMATOGRAM OF DIESEL EXHAUST COLLECTED ON A HIGH CAPACITY CHROMOSORB 102 SAMPLER
-------
Table X
DATA SUMMARY ON ODOR RELEVANT EXHAUST COMPONENTS
Peak
No. a
1
2
-'
4
5
6
7
f.'
1 j
10
i J
12
13
U
15
Jfa
17
18
19
20
21
22
23
2-1
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
KOV..J'
Index
Carbow.i.x
20M
7 JO
723
788
HOO
809
ei,9
895
91 i
934
9S7
973
1000
1015
10J1
1035
10b9
1091
1093
1109
1118
1126
Ilb2
1169
1171
1191
1J95
1200
1205
1216
1230
1230-50
1269
1273
1279
1292
1294
1300
1310
1319
1329
1332
Ketent ion
Time
Apiezon L
(min)
0.3
0.3-0.5
0.3
0.5
0.3-0.5
0.5-0.7
0.3
1.0
0.6
0.5
O.P
3.9
1.7
1.4
1 .2
1.0-1.3
1.2
1 .4
1.0
2.2
2.3
2.6
1.8
2.3
2.6
1.2
1.4
] .4
2.0
2.9
3.85
1.5
1.6
3.4
1.7
2.7
1.4
1.6
1.0
1.4
2.9
1.8
4.3
3.2
Odor Notes
sweet
sweet
sweet
foul
pleasant
unpleasant aldehyde
pleasant
foul
foul
strong foul
unpleasant aldehyde
(non-odorous)
aldehyde
( non-odorous)
unpleasant aldehyde
aldehyde
pungent aldehyde
unpleasant burnt
pungent burnt
pungent plastic
pungent
pungent
pungent burnt
pungent -fruity-aldehyde
pungent
burnt -pungent -"gar age"
. foul
aldehyde, citrus
pungent
aldehyde
burnt
pungent
strong burnt
burnt
burnt, pungent, aldehyde odors
strong foul
pungent , fuel
aldehyde
pungent
unpleasant burnt
strong burnt
unpleasant
pungent -burnt
pungent-burnt
pungent aldehyde
burnt acrid
unpleasant
strong unpleasant burnt
pungent burnt
Exhaust
Concn.
10-9 g/1
0.06
0.3
0.5
0.2
0.1
2 x 101
0.02
1
0.6
0.02
0.002
0.004
0.05
0.005
0.1
0.2
0.01
0.004
0.2
Mass Spectral Data
ethanal
n-propanal
acetone
n-butanal
2-propanol
possibly acetoin,
methylallylketone, and
1,2,3, 4-diepoxybutane
n-pentanal
decane
toluene
n-hexanal
possibly cyclic olefin
or alkyne
ethyl benzene
p-Xylene
possibly 1 ,7-octadiyne
m-Xylene
hydrocarbon with 2
positions of unsatura-
tion
n-heptanal
t r i methyl thiophene
C4 substituted benzene
n-octanal
trimethyl- or methyl-
ethyl-benzene
methallylbenzene ( 2-
methyl-3-phenyl-l-
propene) or p-ethyl
styrene
Cg substituted benzene
45
-------
Table X (cont.)
Peak
No. a
40
41
42
4 1
44
4'S
it-
'
>y
19
j t
.i
52
1 ^
' 4
''
5'-,
s"1
,H
' 'j
1)
-------
borderline cases of odor relevancy, the higher substituted
benzenes, indans, and tetralins are present in odor relevant
concentrations. High odor intensity was also characteristic of
the naphthalenes and aldehydes identified and particularly the
aldehyde and ketone derivatives of benzene, but these individual
odor qualities are submerged in the total exhaust odor. The
series of odor-important components characterized as "foul" were
not all identified, but the data obtained suggest that several
different low threshold types may contribute this type of odor,
including oxygenates, highly unsaturated hydrocarbons, and sul-
fur species.
Some burnt odor qualities can be attributed to the aromatic
species and unsaturated aliphatics, but there are also several
very strong burnt odor species with exceptionally low thresholds.
These species occurred at Kovats Indexes -1200, 1216,1294,
1329, and 1432. During earlier investigations, no odorous
components could be found on the Apiezon L chromatograms from
these Carbowax 20M peaks. The larger samples taken on the high-
capacity Chromosorb collectors increased the component
concentrations to the amount necessary to produce a detectable
peak on the Apiezon L chromatograms to signal for odor monitor-
ing. Figures 8a and 8b show sections of the Apiezon L
chromatograms produced by four of these strong burnt Carbowax
20M peaks. The area on the chromatogram producing the same odor
quality and intensity is designated. The very small concentra-
tions involved precluded any mass spectrometric investigations
of these species.
Correlations between the gas-chromatographic positions in
the two-column chromatograms and the physicochemical and func-
tional characteristics of the compounds (Figure 1 and
Appendix G) permit inferences on some properties of these uni-
dentified burnt-odor components. Thus, approximate boiling
IT RESEARCH INSTITUTE
47
-------
00
Carbowax 20M
KOVATS INDEX PEAK 1216
42
(min)
39
X8 Carbowax 20M
KOVATS INDEX PEAK ~1294
Strong Unpleasant
Burnt
x2
48
45
(min)
Figure 8 a
APIEZON L CHROMATOGRAMS OF STRONG BURNT CARBOWAX 20M PEAKS
-------
x8
Carbowax 20M
KOVATS INDEX PEAK 1329
x4
vo
54
Unpleasan
Burnt
x!6
51
(min)
48
Carbowax 20M
KOVATS INDFX PEAK 1432
57
(min)
Figure 8 b
APIEZON L CHROMATOGRAMS OF STRONG BURNT CARBOWAX 20M PEAKS
x2
54
-------
points, °C, and molecular weights of the five mentioned burnt
components are estimated: 165°, 130; 190°,170; 170°, 130; 210°,
150; and 200°, 155, respectively. The presence of -OH group is
excluded. The Kovats Index 1216 and 1329 components are rela-
tively nonpolar and cannot be simple aldehydes, other carbonyl
compounds, aryls, or derivatives of cyclic ethers. The degree
of polarity also excludes non-cyclic alkenes, but cyclic olefins
are possible. The Kovats Indexes 1200, 1294, and 1432 compo-
nents have polarity of the order of saturated carbonyl
compounds, including branched aldehydes, or aromatic derivatives
without heteroatoms but with additional unsaturation in the
side chain.
D. Acidic and Phenolic Components in Diesel Exhaust
Because acidic and phenolic compounds do not chromatograph
on the Carbowax 20M column, diesel exhaust samples were analy-
zed on a packed column of Chromosorb 101 to specifically deter-
mine the presence and odor relevance of these species in the
exhaust. Several acidic and phenolic standards were run on the
column to obtain retention data and odor references. The diesel
samples analyzed were poorly resolved, but emphasis was placed
on the odor analysis of the column effluent since the compound
types sought would elute as distinct, though perhaps unresolved,
peaks.
Figure 9 presents a summary of the calibration data and the
relevant odors from the exhaust chromatograms in Kovats Index
notation. Definite acidic odors were recognized at retention
times characteristic of 2-through 6-carbon acids with the most
intense odor for a C acid. To provide some basis for compari-
son with other chromatographic odor annotations, the acidic
odor intensities can be said to be comparable to the aldehyde
odors in the same effluent.
A definite conclusion on the presence of phenolic compounds
in the exhaust was difficult to make from the data obtained. In
II T RESEARCH INSTITUTE
50
-------
STANDARDS
KOVATS INDEX
DIESEL EXHAUST SAMPLES
Decanolc Acid
2,3,5-Trimethylphenol >
3,4-Xylenol -
2,3-Xylenol, 2,5-Xylenol-
Octanoic Acid __->
p-Ethylphenol
2,4-Xylenol,2,6-Xylenol
o-Ethylphenol ~~
m-Cresol
o-Cresol
Phenol
Hexanoic Acid
Pcnhanoic Acid
3-Methylbutanoic
Acid
Butanoic Acid
Propanoic Acid
Ethanoic Acid
1400-
1300-
1200-
1100
1000
900-
- 800-
- 700-
- 600
Burnt
^-Pungent .
Odors
«- Strong Burnt -
Pungent
Burnt Medicinal Odors
- Burnt-Medicinal Odors
- Acid Odor
Acid Odor
Acid Odor
Acid Odor
Acid Odor
\- 500
Figure 9
51
-------
the retention time areas of phenolic compounds the column
effluent odors were best described as burnt-pungent. In some
instances there was a sweet or pleasant overtone, and the pres-
ence of a slight medicinal-type odor was occasionally detected.
IIT RESEARCH INSTITUTE
52
-------
Appendix A
EMISSION SOURCE AND SAMPLING APPARATUS
NT RESEARCH INSTITUTE
A-l
-------
Appendix A
EMISSION SOURCE AND SAMPLING APPARATUS
The exhaust sampling system is presented in Figure A~l.
The diluent nitrogen is first rendered non-odorous by passage
through an activated charcoal filter immersed in a dry ice-
acetone bath. It flows through an air coil, where it is warmed
to about 50°F, and then through a jet syphon (Schutte and
Koering type 217 steam syphon) where it dilutes and cools a
portion of the engine exhaust. The diluted mixture travels
into a plenum chamber where some is collected by the sampling
devices. The major portion of the engine exhaust is discharged,
undiluted, from the engine laboratory. Generally, the volume
of engine exhaust diverted into the sampling stream was
0.3 scfm, and the volume of nitrogen diluent used was 3.2 scfm.
The entire system is fabricated from stainless steel,
glass, or Teflon, with the exception of the plenum. The plastic
plenum is lined with a polyethylene bag which is replaced each
time the engine exhaust is sampled. The volumes of extracted
exhaust and diluent nitrogen are measured with calibrated
sharp-edged orifice plate meters. Temperatures are measured
with either chromel-alumel thermocouples or bimetallic thermom-
eters. Engine operating conditions during exhaust sampling are
compiled in Table A-I.
Mobil Research and Development Corporation supplied the
data on their #1 Diesel Fuel presented in Table A-II.
NT RESEARCH INSTITUTE
A-2
-------
TO BUILDING
EXHAUST
T.C.#3
FXHAUS1 SAMPLE PROBE-
AIR MOVER
FLUIDIZED BED SAMPLING DEVICE
PLENUM
JET SYPHON
ORIFICE # I
ORIFICE #2
LLGENO
.WARMING COIL
CHARCOAL
DRY-ICE
ACETONE
BATH
J
f
THERMOMETER
^-THERMOCOUPLE
N
Figure A-I
DIESEL EXHAUST SAMPLING SCHEME
A-3
-------
Table A-I
ENGINE OPERATING AND SAMPLING CONDITIONS
Detroit V-71 Diesel Engine
Parameter
Engine speed, rpm
Engine load, ft-lb
Engine horsepower
Engine cooling water temperature,
5F
Fuel flow, Ibs/hr
Fuel pressure, psig
Oil pressure, psig
Air intake flow., scfm
Exhaust temperature, °F. T.C.#3
Room temperature, °F
Room humidity, %RH
Room pressure in, Hg
Exhaust sample line temperature
°F. T.C.#1
Nitrogen sample line
temperature °F, T,C,#2
Diluted nitrogen-exhaust sample
T#l
Typical
1200
134
54
188
25
30
55
340
450
78
23
29,2
120
50
80
Range
1200 -1250
131 - 134
52 - 54
180 -
21 -
23 -
47 -
322 -
420 -
77 -
22 -
28.4-
190
26
37
57
382
480
89
33
29.
9
110 -
45 -
77 -
132
70
85
A-4
-------
Table A-II
ANALYSIS OF DIESEL FUEL1
Test Analysis
Gravity, API 42.4
Distillation, °F - Initial 354*
10% 398
50% 459
90% 534
End 580
Flash Point, °F 150
Aniline Point, °F 151
Sulfur, % wgt 0.13
Total Mercaptan, ppm, wgt < 0.5
Mercaptan, NaOH Extract, ppm, wgt < 0.5
Viscosity, Centistokes @ 100°F 1.89
FIA Analysis, % by volume
Aromatics 15.3
Olefins 1.7
Saturates 83.0
Cetane Index 55.5
Cetane Number 56*
Nitrogen, ppm 14
Color, L 0.5
Phenols, ppm 149
Thiophenols, ppm 9
Mobil #1 Diesel Fuel.
IIT RESEARCH INSTITUTE
A-5
-------
Appendix B
EXHAUST SAMPLING DEVICES
IIT RESEARCH INSTITUTE
B-l
-------
Appendix B
EXHAUST SAMPLING DEVICES
The fluidized bed sampler, shown in Figure B-l, is a
glass column, 5 j.n. x 1,25 in., diameter., containing 10 g of
40/60 mesh Fluoropak 80 with a 15% Apiezon L coating. Stain-
less steel screens at each end of the column hold the loose
packing medium in the chamber. A small blower pulled the
diluted diesel exhaust through the bed at a rate of 45 liters
per min. The exhaust was sampled for a period of 30 min, per-
mitting about 125 liters of undiluted exhaust to pass through
the collector. Generally, eight fluidized bed samples were
collected from each engine run,
The packed bed samplers are stainless steel tubes, 10 in.
x 0.375 in. diameter; with O.L25 in, o.d. stainless steel tub-
ing silver brazed to the ends, The collector holds 4.2 g of
the collection phase, either 15% methylsilicone SF-96 on
80/LOO mesh Chromosorb HP or r.he solid adsorbent C-.102 in
60/80 mesh bize. During exhaust sampling, the bed was con-
nected through a dry ice-acetone cold trap to a leveling bulb
containing silicone oil, and 2 liters of the dilute exhaust
were pulled through the bed by displacement at a rate of
100 ml/min,
To collect larger samples and to take advantage of the
faster collection rate possible with the C-102 material, a
wider diameter packed bed sampler was fabricated, The col-
lector design is shown in Figure B-2. It consists of a 0.75 in.
o<,d. thin-wall stainless steel cylinder, 4 in, long with
0.125 in. o.d, stainless steel tubing ends. The packing mate-
rial, 5 g of 60/80 mesh C-102, is held in place by two 100
mesh stainless steel screens, one fixed and the other spring
loaded to provide reasonable compaction and eliminate channel
formation. Diluted exhaust was pulled through these samplers
JIT RESEARCH INSTITUTE
B-2
-------
GLASS
COLUMN
COATED
TEFLON-
PARTICLES
TEFLON
ADAPTORS
STAINLESS STEEL
PERFORATED DISC
STAINLESS STEEL
PERFORATED DISC
TEFLON
ADAPTORS
Figure B-I
FLUIDIZED BED COLLECTOR
B-3
-------
TO
PUMP
STAINLESS STEEL TUBING
" O.D. (0.085" LD)
SPRING
STAINLESS STEEL SCREEN
CHROMOSORB 102, 5g
STAINLESS STEEL SCREEN
STAINLESS STEEL TUBING
f
EXHAUST
INLET
HIGH-SPEED VAPOR COLLECTOR
Figure B-2
HIGH-SPEED ORGANIC VAPOR COLLECTOR
-------
at a rate of 3 liters/min for a total sample of 100 liters
(9.1 liters of undiluted exhaust).
All sampler types were cleaned and conditioned before use
by passing a purified helium flow of 60 cc/min through the sam-
plers at elevated temperatures for a minimum period of 18 hr.
Conditioning temperatures were determined by the specific tem-
peratures used for elution of collected sample. Conditioning
efficiency was checked gas chromatographically with blank elu-
tion samples. Teflon tubing was used for connections to the
samplers, and the ends were sealed with Teflon plugs.
IIT RESEARCH INSTITUTE
B-5
-------
Appendix C
SAMPLE ELUTION AND INJECTION TECHNIQUES
IIT RESEARCH INSTITUTE
C-l
-------
Appendix C
SAMPLE ELUTION AND INJECTION TECHNIQUES
Samples collected in fluidized or packed beds were
transferred to a specially designed needle for introduction
into the gas chromatograph using the assembly shown in
Figure C-l. The collector is placed in an aluminum block
thermostatted at 120°C (80°C for Apiezon L fluidized beds).
Zero quality helium is passed first through a liquid nitrogen
cooled copper coil, then through the collector (in a direction
opposite to the direction of sampling into the collector), and
into a stainless steel injector needle at a rate of 60 ml/min.
During the transfer of sample, the injector needle is held in a
rectangular cross section groove in a liquid-nitrogen-cooled
block (shown only schematically in Figure C-l). The groove is
covered with a copper plate. A copper rod from the block
extends into a Dewar of liquid nitrogen. A copper side tube
provides a path for the nitrogen to the cooling groove so that
boiling nitrogen in this side tube generates sweep gas which
purges the groove volume and prevents the formation of ice
around the injector needle. The connection between the col-
lector and injector is made with Teflon tubing. The connections
are actually heated up to the larger portion of the injector
needle to prevent any condensation of sample before this point.
A rotameter connected to the end of the injector needle moni-
tors the helium flow.
The injector needle design is shown in enlarged view in the
upper section of Figure C-l. The right end of the injector is
inserted into the injection port of the gas chromatograph, and
the left end is connected to a bellows mechanism. The
restrictor in the left end prevents backflow of sample when the
sample is later heated for injection.
After the sample has been transferred to the injector, it
is connected to the gas chromatograph through a modified
IIT RESEARCH INSTITUTE
C-2
-------
n
i
u>
n
o
r
r
w
n
H
w
o
en
t-
n
H
cn
SWAGELOK NUT
w
po
|Q
re
o
i
w
n
-jL"o.D.(0.030"l.D)
TEFLON TUBING;
FLOW - 60ml/min
SPLITTER
COLLECTOR WITH SAMPLE
^-ALUMINUM BLOCK AT 120° C
ZERO
COPPER
BLOCK
LIQUID N2
LIQUID N.
TRANSFER OF SAMPLE FROM COLLECTOR TO INJECTOR
-------
injection port that consists of a stainless steel fitting with
*
a Teflon compression plug sealing element. The plug has a
0.0625" hole through which the small end of the injector is
inserted. A seal is produced by tightening the nut surrounding
the Teflon plug. The other end of the injector is attached to
an injection assembly. The entire apparatus is shown in
Figure C-2.
After sample transfer, the cold injector is held between
two liquid nitrogen cooled copper blocks, slotted to accomo-
date the 0.125" o.d. tube. A motorized eccentric cam and
bellows mechanism pulls 1 ml of carrier gas helium from the gas
chromatograph past the cold sample, and stores the helium in
the bellows chamber. To inject the sample, the lever is
rotated to remove the cold horizontal copper blocks from the
injector and replace them with the vertical set, which are pre-
heated to 250°C. Simultaneously, the bellows movement pushes
the stored helium through the vaporizing sample and into the
chromatograph. After 10 seconds, when the helium injection is
completed, the cold blocks are again clamped around the injec-
tor. The pressure gauge on the system is necessary to deter-
mine when the pressures in the injection-chromatograph system
have stabilized after withdrawing carrier gas for injection.
The gauge also serves to detect any leaks in the system. When
not attached to the GC, the injection system is continuously
flushed with helium through the port shown capped for a sample
injection. The U-tube between the bellows and the injector
needle is immersed in liquid nitrogen to prevent sample contam-
ination from the bellows system.
*
Conax fitting.
(IT RESEARCH INSTITUTE
C-4
-------
n
>
r»
ECCENTRIC
BELLOWS
(enclosed)
PORT FOR FLUSHING
t$t-
-* n \
-MOTOR
HEATED COPPER
BLOCKS
INJECTION NEEDLE WITH SAMPLE
GAS CHROMATOGRAPH
LEVER TO CHANGE
, FROM COLD TO HOT
BLOCK
COLD VAPOR TRAP
BELLOWS
ACTUATOR SWJTCH
LIQUID NITROGEN
LIQUID NITROGEN
COOLED COPPER BLOCKS
Figure C-2
APPARATUS FOR SAMPLE INJECTION
INTO GAS-CHROMATOGRAPH
-------
Appendix D
GAS CHROMATOGRAPHIC EQUIPMENT AND TECHNIQUES
IIT RESEARCH INSTITUTE
D-l
-------
Appendix D
GAS CHROMATOGRAPHIC EQUIPMENT AND TECHNIQUES
The single column chromatograph, an Aerograph Model 1200,
is depicted in Figure D-l. Helium, carrier gas flows through the
column at a rate of 10 ml/min. At the column exit, 30 ml/min of
helium is added to the column effluent, and the mixture is split
1:1. One 20 ml/min portion is delivered to the flame ionization
detector, and the other half continues to a sniffing port where
the odor of effluent peaks is monitored. The two high resolution
columns used in the single column instrument were both 200' SCOT
columns with an i.d. of 0.020-in. The stationary phases were
Apiezon L and Carbowax 20M.
When the packed column, 4' of 3 mm o.d. pyrex tubing, of
60/80 mesh Chromosorb 101 was used, the 40 ml/min column flow
provided, after splitting, 20 ml/min flows to both the detector
and the sniffing port without the necessity of auxiliary helium.
Samples and standards were chromatographed on this column with
a programmed temperature rise of 2° or 4°/min from 100° to 300°C.
The sniffing port design is shown in Figure D-2. The portion
of the effluent flow for sniffing passes through a Teflon line,
maintained at the temperature of the detector, to the outside of
the chromatograph housing and into the center of a cylindrical
Teflon chimney (2" by 0.75" diameter). A humidified air flow of
40 ml/min is introduced at the base of the chimney and mixes with
the column effluent of 20 ml/min that is ejected upward from three
holes in the Teflon line. A 60 ml/min flow of helium and
humidified air is not disturbingly fast to the nose, and the
arrangement provides a reasonable standardization of dilution of
the odorant carried from the gas chromatograph. Because the
widths of chromatographic peaks are approximately the same when a
programmed temperature rise is used, all compounds emerge with
approximately the same dispersion in time and volume.
IIT RESEARCH INSTITUTE
D-2
-------
20 cc/min J
HEATED TEFLON
CAPILLARY
Fl DETECTOR
* 20 cc/min
- HELIUM 30 cc/min
o
o
200 ft OPEN TUBULAR
CARBOWAX 20 M COLUMN
SNIFFING PORTO
HELIUM 10 cc/min
SAMPLE INJECTION
Figure D-l
SINGLE-COLUMN GAS-CHROMATOGRAPHIC FLOW ARRANGEMENT
NT RESEARCH INSTITUTE
D-3
-------
o>
o
o
Sniffing
Port
Effluent
20ml/min
»
/]
0
e>
Teflon Tubing
-^
Low-Voltage
Insulated
Heater
.^_Z_J-
II
II
II
Teflon
<*
40ml/min
Humid Air
/
Water
Air 40 ml/min
S
Figure D-2
SNIFFING PORT DESIGN
IIT RESEARCH INSTITUTE
D-4
-------
A schematic of the two-column system is shown in Figure D-3.
For most of the exhaust sample analyses, the primary analysis
column was a 200-ft SCOT column, 0.020" i.d., with Carbowax 20M
stationary phase followed by a 50-ft Apiezon L SCOT column,
0.020" i.d. An injected sample passes through the first column
at a helium flow rate of 10 ml/min. At the exit of the column,
the effluent is split 1:1. One 5 ml/min portion is supplemented
with 35 ml/min of helium and resplit, providing 20 ml/min flows
to both a flane ionization detector and a sniffing port. The
other half of the first column effluent travels to a secondary
flow system where it is directed either onto the second analysis
column or onto a similar dummy column. The injection of a peak
from the continuous flow of primary column effluent onto the
second column is accomplished with a three-way solenoid valve
controlling an auxiliary helium input of 15 ml/min. The effluent
switching itself is valveless. When the solenoid is switched on,
5 ml/min of helium is mixed with the 5 ml/min primary column
effluent and the total directed onto the secondary analysis
column, with the remaining 10 ml/min of auxiliary helium flowing
through the dummy column. When the solenoid is off, the reverse
occurs The 10 ml/min helium-diluted primary column effluent
flows through the dummy column,and 10 ml/min auxiliary helium
passes through the analysis column. Thus, at all times, 10 ml/min
flows pass through both columns in the secondary flow system. At
the exit of the second analytical column, 30 ml/min of helium is
added to the 10 ml/min column effluent,and the total is split so
that 20 ml/min flows are provided for the second flame ionization
detector and the second sniffing port. Considering the two
portions of a particular amount of first column effluent, the one
part reaches the detector at the same time the other part reaches
the point of injection onto either of the two columns. rj;>hus, the
switching of the solenoid delineates the exact portion of first
column chromatogram rechromatographed on the second column. In
addition, since the primary column effluent is split 1:1,
NT RESEARCH INSTITUTE
D-5
-------
SNIFFING PORTS
\
d
I
cr>
n
z
m
nnr~rr~pTnp f~*\
UL 1 LL 1 UK \^J
HFI HIM ^O rr/min
10 cc/min
50ft OPEN TUBULAR
APIEZON L COLUMN
k x S~\ ci nc-TcrrrnD
5 cc/min
N
0
0
o
0
o
0
o
o
o
0
o
« HFI HIM "^S rr/min
[ 5 cc/min
i i
0
L §
0
O
o
o
o
0
o
o
o
o
o
vv
o
DUMMY!
i i O
lOcc/min °
o
o
o
o
o
o
o
o
o
J .,
200 ft OPEN TUBULAR
CARBOWAX 20 M COLUMN
SOLENOID VALVE (3-WAY)
HELIUM 10 cc/min
HELIUM 15 cc/min
Figure D-3
DUAL-COLUMN GAS-CHROMATOGRAPHIC FLOW ARRANGEMENT
-------
approximately equal concentrations reach the two sniffing ports
and two detectors.
An estimate of total sample concentration and the
concentrations of several individual components from the two-
column instrument were obtained by electronic integration of peak
areas using the Hewlett-Packard Model 3370 A integrator. The
flame ionization detector response, which is essentially
proportional to the weight concentration of organic material,
was calibrated with a known mixture of methane in helium.
IIT RESEARCH INSTITUTE
D-7
-------
Appendix E
CHROMATOGRAPHIC DATA PROCESSING
THE KOVATS INDEX SYSTEM
IIT RESEARCH INSTITUTE
E-l
-------
Appendix E
CHROMATOGRAPHIC DATA PROCESSING -
THE KOVATS INDEX SYSTEM
The problem of comparing absolute retention data collected
from many experiments or even from many sources is greatly assisted
by the use of the Kovats Index system. ~ Based on the fact that
the logarithms of the retention times of the n-alkanes (corrected
for the column dead volume) are a linear function of chain length,
Kovats indexing is a logarithmic transformation of the corrected
retention time of unknowns with respect to the corrected retention
times of n-alkanes. By definition, the indexes of the n-alkanes
are set equal to 100 times the number of carbon atoms (e.g.,
n-octane = 800, n-decane = 1000) and comprise the fixed points
of the linear retention index scale from which unknowns are
characterized.
The logarithmic function is applicable under isothermal
conditions. When using linear temperature programming, however,
the corrected retention times of the n-alkanes are an approximately
linear function of chain length. To use Kovats Index notation for
the data from the exhaust samples covering the range of 5 through
C16, a calibration curve was obtained from the retention times of
the internal standard n-alkanes with a computer program which
performs a least squares semi-logarithmic fit. The curve is
normalized to a straight line defined by a fifth order polynomial.
From the calibration curve, the program then determines the Kovats
Indexes of all other peaks in the chromatogram.
(1) E. Kovats, Helv. Chim. Acta. 41, 1915 (1958).
(2) A. Wehrli and E. Kovats, ibid. 12, 2709 (1959).
(3) E. Kovats, Z. Anal. Chem. 181, 351 (1961).
(4) L.S. Ettre, "The Kovats Index Retention System," Anal. Chem.
16, 8, 31A (1964).
(5) E. Kovats, An, Gas Chromatography I, 229 (1965).
NT RESEARCH INSTITUTE
E-2
-------
The Kovats retention index system also provides a number of
thermodynamically based rules that permit calculating the Kovats
Index of a particular substance and thus predicting the positions
of various homologs and derivatives. The principal rules are
summarized as follows:
1. In an homologous series, the index increases by
approximately 100 units per CH2-group (exceptions are
the lowest members of a series) , and similar substitutions
in similarly constructed compounds result in a similar
increment of the indexes.
2. Thf. index of an assymetrically substituted compound
(e.g., X-R-Y) can be calculated as the mean of the two
symmetrically substituted compounds (X-R-X and Y-R-Y).
3. The index of a non-polar substance remains almost the
same in all stationary phases.
4. The index of any substance in various non-polar phases
remains almost the same.
5. The difference in the indexes of two isomers in a non-
polar stationary phase is approximately equal to a five-
fold difference in their boiling points.
6. The difference between the indexes of a substance in a
polar and a non-polar stationary phase is characteristic
of the structure of the substance since it is an additive
function of the various functional groups in the molecule,
allowing for amount of screening of the functional groups
by the structure of the molecule.
IIT RESEARCH INSTITUTE
E-3
-------
Appendix F
GAS CHROMATOGRAPH - MASS SPECTROMETER INTERFACING
NT RESEARCH INSTITUTE
F-l
-------
Appendix F
GAS CHROMATOGRAPH - MASS SPECTROMETER INTERFACING
The apparatus for trapping specific components from the
Apiezon L effluent of the two-column chromatograph,and
subsequently releasing them for mass spectrometric analysis, was
designed with 6 mil stainless steel tubing comprising the small
volume system serving as the link between the two instruments.
Ten individual trapping stations are provided by sets of hot and
cold copper jaws in discrete positions on the 6 mil tubing.
Figure F-l shows a portion of interface assembly with two
of the trapping stations depicted. The point labeled A,
representing a small volume Swagelok connection, is the point of
attachment of the interface to either the GC or MS. The line is
6 mil tubing from point A to the liquid nitrogen cold trap.
Teflon spacers, 0.0625" wide, between trapping stations hold the
tubing securely in position. The grooves in the copper jaws
provide good contact with the tubing, and spring levers position
the jaws accurately and reproducibly in both hot and cold
settings. The hot jaws are maintained at 180°C with cartridge
heaters. The cold jaws extend into a Dewar of liquid nitrogen.
Beyond the cold trap, the line continues to a tee with solenoid
valves on each leg. Valve B, normally open, controls a helium
input of 2 cc/min. The leg through valve C, normally closed,
leads to a mechanical pump. At all times, except when actually
trapping a peak, 2 ml/min of helium flows through the line.
In Figure F-2 the interface is shown connected to the gas
chromatograph. The point of connection (A) is inside the
chromatograph oven. From the point of connection to the first
trapping station, Teflon-insulated nichrome wire is wound around
the line to maintain the temperature at 180° to 200°C. The
flow system of two-column chromatograph was slightly modified
for trapping peaks with the interface. The 10 ml/min Apiezon L
II T RESEARCH INSTITUTE
F-2
-------
"Firerod" Cartridge eoter,
:~ot Copper Ja"s,
6 nil Stainless Sttel Tubing,.
A
Cole Copper Jaws
Liquid nitrogen Dewar
Po.5i.t--0:v.--.} Lever
Hrliur, ;.cc/min
V
Liquid Nitrogen
Cold Trap
Figure F-l
MECHAl^ISM FOR THE TOAPPI1IG A1JD RELEASE OF COMPOlffillTS I.J THE GC-MS INTERFACE
-------
Auxiliary Helium Input
15cc/min
Flame lonization
Detector
Apiezon L Column
Helium, lOcc/min
Gas Chromatograph Oven
Teflon Insulated
/Nicrome Wire for
Resistance Heating
^2/1 Res
^X''
c6 mil Stainless
Steel Tubing
Solenoid Valve D, N.O.
Figure F-2
INTERFACE COUWECTIOiJ TO THE TWO-COLUMN GAS CHROMATOGRAPH
-------
column effluent is split 1:1 at point E on Figure F-2. One
5 ml/min portion is supplemented with 15 ml/min of auxiliary
helium to provide a 20 ml/min flow to the detector as usual.
The other portion of the effluent, along with the 2 ml/min flow
from the interface line, is vented at point F on Figure F-2,
except when trapping a peak. The flow through the venting line
is controlled by solenoid valve D, normally open. Solenoid
valves B, C, and D are activated by a common switch.
In the trapping operation, all hot jaws are initially on
the transfer line and solenoids B, C, and D are off. When an
Apiezon L peak is to be trapped, the cold jaws of a trapping
station are positioned on the line,starting with the station
furthest from the chromatograph. The three solenoid
valves are switched on (valves B and D closed and valve C, to
the pump, open), and the column effluent is drawn through the
transfer line for the duration of the peak elution as shown on
the Apiezon L chromatogram. The solenoids are then switched off,
and helium again flows in the reverse direction through the
transfer line and is vented along with the column effluent.
The assembly for joining the interface to the mass
spectrometer is shown in Figure F-3. The interface line is
attached at point A to a length of 6 mil stainless steel tubing
extending through a flange bolted to the MS. Inside the MS, the
tube is inserted into a pyrex tube spring positioned to deliver
the sample directly into the ion chamber of the MS. The 6 mil
tube is crimped, at point X on Figure F-3, to match the leak rate
to the pumping speed of the MS.
A set of hot and cold jaws, identical to those used for
injections into the chromatographs, provides the means for trap-
ping peaks released from the interface line and subsequently
injecting them into the MS in good peak shape. The sharp peak
injections maximize the sample concentration available for mass
spectral analysis. A 0.125 o.d. stainless steel tube sheathes
the 6 mil tube to provide good contact with the grooves in the
copper jaws.
IIT RESEARCH INSTITUTE
F-5
-------
Quartz Tube
Positioning Sprin
1/8" Stainless Steel
Tubing Jacket for 6 Mil
Tubing
i Flange Bolted Directly onto lonization
^Chamber of Mass Spectrometer
1/8" Tubing Cut Away to Show
Crimped Section of 6 Mil
/Tubing
Hot Copper Jaws
Cold Copper Jaws
Teflon Insulated Nicrome
Wire for Resistance
Heating
6 Mil Stainless
Steel Tubing
Extension from Each Cold Jaw _
is Immersed in Liquid Nitrogen
Figure F-3
DEVICE FOR SAMPLE INTRODUCTION INTO THE MASS SPECTROGRAPH
-------
In operation, the set of cold jaws are positioned on the MS
inlet tube, and the first component trapped in the interface (last
trapped from the chromatograph) is released by removing the cold
jaw at the trapping station and placing the hot jaw on the line.
When the peak has been retrapped at the MS inlet, the cold jaws
there are removed, and the hot jaws are placed around the tube to
inject the peak into the MS. The efficiency of peak injection
was tested by trapping a benzene peak from the GC and monitoring
the principlp ion peak, m/e 78, in the benzene spectrum after
injection of the benzene into the MS. The recorder trace, shown
in Figure F-4, indicated that samples are injected in good peak
form.
With the interface system attached, the mass spectrometer
operated at a vacuum of 2-3 x 10~° mm Hg. For the exhaust sample
analyses, the ionizing energy was set at 70 ev, accelerating
voltage at 1800v for a mass range of ~2-600, and coil current
usually set to scan up to 200 or 250 mass units. The resolution
(M/AM) of the Hitachi Perkin-Elmer RMU-6D is 2000.
NT RESEARCH INSTITUTE
F-7
-------
Sensitivity xl
x3
xlO
x30
Speerl 2 mm/sec
Figure F-4
RECORDER TRACE OF m/e 78 PEAK INTENSITY AFTER INJECTION
OF BEtiZEHE PEAK TRAPPED IH INTERFACE
-------
Appendix G
MOLECULAR PROPERTIES FROM TWO-COLUMN
CHROMATOGRAPHIC DATA
NT RESEARCH INSTITUTE
G-l
-------
Appendix G
MOLECULAR PROPERTIES FROM TWO-COLUMN
CHROMATOGRAPHIC DATA
The two-column analysis system results in data from which
physicochemical properties of compounds can be estimated. In
the two-dimensional plot of Carbowax 20M Kovats Index vs. cor-
rected Apiezon L delay time, other molecular properties can be
represented as the third dimension defining specific values of
this third variable. The extent of the correlation was ana-
lyzed using the two-column data compiled from the known
calibration compounds in a computerized regression program.
The computer program was BMDO22 Stepwise Regression,
Revision July 18, 1968, Health Sciences Computer Facility,
UCLA. The two independent variables were: (1) Kovats Index in
Carbowax 20M obtained under chromatographic conditions
described in Section II; and (2) the delay in minutes in the
Apiezon L column, corrected for column dead volume. The two
variables were the X and Y coordinates in a plane with the
dependent variable, Z = F (X,Y), as a curved surface above the
X-Y plane. The surface is expressed by a polynomial equation
in which X and Y, their higher powers, and cross-terms of the
various powers are used to form that Z plane for which the mean
error of estimate for Z values is a minimum and the correlation
coefficient is a maximum. A Calcomp plotter was used to plot
curves connecting the points of equal Z values in the X-Y plane.
The Z's considered, pertinent to the diesel exhaust
study, were boiling points and molecular weights. The correla-
tion obtained between these molecular properties and the two-
column data on known compounds is expressed by the following
values of the resulting mean errors of estimate and correlation
coefficients:
IIT RESEARCH INSTITUTE
G-2
-------
Mean Errors Correlation
of Estimate Coefficient
Boiling Points 5.7°C 0.995
Molecular Weights 9.1 0.972
Figures G-l and G-2 are the Calcomp plots of the iso-boiling
point and iso-molecular weight curves, respectively.
These data complement the polarity-structure information
provided by the retention difference obtained on the two-
column chromatograph. The greater the number of defining param-
eters that can be applied to an unknown sample component peak,
the more accurate are the conclusions regarding its probable
identification. In addition, partial characterization of an
unknown greatly assisted mass spectral interpretation.
IIT RESEARCH INSTITUTE
G-3
-------
7.00 9.00 H.CO 13.00 15.00
KOVRTS INDEX IN CRRBOWflX 20 M xicr2
17.00
PLOT OF 'BOILING POINT , DEGREES C1
Figure G-l
ISO-BOILING POINT CURVES FROM TWO-COLUMN CHROMATOGRAPH
DATA ON KNOWN COMPOUNDS
-------
7.00 9.00 11.00 13.00 15.00
KOVflTS INDEX IN CflRBOWflX 20 M x 10-2
17.00
PLOT OF 'MOLECULRR WEIGHT'
Figure G-2
ISO-MOLECULAR WEIGHT CURVES FROM TWO-COLUMN
CHROMATOGRAPH DATA ON KNOWN COMPOUNDS
G-5
------- |