United State*
Environmental Protection
Agency
Environmental Science* Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-79-041
February 1979
Research and Development
Chemical
Composition of
Exhaust
Particles from
Gas Turbine
Engines
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR) .
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-041
February 1979
CHEMICAL COMPOSITION
OF EXHAUST PARTICLES FROM
GAS TURBINE ENGINES
by
D. J. Robertson, 0. H. Elwood, and
R. H. Groth
UNITED TECHNOLOGIES CORPORATION
Pratt & Whitney Aircraft Group
Commercial Products Division
East Hartford, Connecticut 06108
EPA Contract 68-02-2458
Project Officer
J. N. Braddock
Emissions Measurement and Characterization Division
Environmental Science Research Laboratory
Research Triangle Park, N.C. 27711
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, N.C. 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Sciences Research
Laboratory, U. S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products contain endorsement or recom-
mendation for use.
-------
PREFACE
In order to assess accurately the risks involved in the emission of
participate matter from aircraft gas turbine engines, the U. S. Environmen-
tal Protection Agency must, in addition to quantifying the mass emissions of
particulate matter from such sources, determine the chemical composition of
these particulates. It has been known for some time that fossil fuel-fired
combustion sources emit a number of substances which exhibit varying degrees
of toxicity; some of these substances such as certain polycyclic organic
compounds and selected nitrosamines are thought to be carcinogenic. Al-
though there have been no extensive studies performed to date, there is rea-
son to believe that aircraft gas turbine engines burning conventional avia-
tion fuels also produce these substances.
Limited testing to date, both at Pratt & Whitney Aircraft and through
other agencies, however, indicates very strongly that concentrations of tox-
ic substances in turbine particulates are extremely low. This necessarily
imposes a requirement for trapping large amounts of sample in order to en-
sure that an adequate amount of material is available to perform a reliable
qualitative analysis. The amount of total sample required must be deter-
mined from the sensitivity of the analytical methods used, as well as from
the concentration, known or estimated, of the compounds of interest. In
addition, the particulate collection apparatus design must take into con-
sideration not only the collection of the material of interest but its pre-
servation as well, both in character and quantity. Another consideration
must also be the efficiency of the collection system and its ability to col-
lect sufficient sample for chemical characterization within a reasonable
amount of testing time. Any attempt to meaningfully characterize particu-
lates from aircraft gas turbine engines must necessarily employ the use of
high efficiency, high-flow rate, filtration techniques. It is clear that a
simple filtration scheme employing the use of a device such as the EPA/SAE
smoke meter will not be sufficient. While no such schemes have been shown
to be completely satisfactory for sampling gas turbine engines, there are a
number of promising approaches available.
Under this contract an appropriate high volume sampling system was
designed which was used to collect particulate samples from the exhaust of a
Pratt & Whitney Aircraft PT6A-45 Gas Generator. A series of comprehensive
chemical analyses were performed to broaden our knowledge of the chemical
nature of the organic material entrained on the particles.
111
-------
ABSTRACT
Solid particulate matter, mainly carbon, emitted into the air from the
combustion of fossil fuels contains a variety of organic species adsorbed on
it. In order to assess the hazards associated with such emissions from small
aircraft gas turbine engines burning conventional kerosene type fuels, a stu-
dy was undertaken to collect and analyze exhaust particulates; in particular,
polycyclic organic compounds and nitrosamines, some of which may be carcino-
genic. As part of this effort, a high volume sampling collection system was
developed to obtain an adequate amount of sample within a reasonable period
of engine operating time due to the low concentrations of particulate and de-
letrious materials in the exhaust stream. The sampling system satisfactorily
filtered up to 45 m-* of exhaust gas. Although moisture and temperature
problems interfered with the efficiency of the sampling system, it provided a
qualitative analysis of the particulate. Collection of the particulates was
made over a range of engine power settings at idle, approach, climb and take-
off, using low suIfur(0.00655% S)and high sulfur (0.25% S) fuels. Extraction
of the organic matter from the sample was done in a Soxhlet extractor, usual-
ly using hexane, then analyzed by HPLC, GCMS, NMR and other procedures to de-
termine the total organics adsorbed, the PAH content, and the presence of ni-
trosamines and phenols.
Total organics were determined by a backflush chromatographic procedure.
This analysis showed that the organic material entrained on particulates
emitted from gas turbines is a small fraction of the total organics emitted
(less than 1%). Although this amount is a small fraction of the total organ-
ics emitted, it is significant because of the respirable nature of the parti-
culates. Polynuclear aromatic hydrocarbons (PAH) were determined by GC/MS and
high performance liquid chromatography (HPLC) techniques. Most of the PAH
were non-carcinogens and were composed of the 3 to 4 fused ring compounds.
The GC/MS technique identified specific compounds and the HPLC gave a good
indication of the relative amounts of compounds in the 3 to 4 fused ring
types versus the 5 to 6 fused ring types. The larger fused ring compounds ex-
isted in low concentrations. Phenols and nitrosamines were isolated and then
measured by gas chromatography using,a flame ionization detector and nitrogen
detector. Nitrosamines were not found and the presence of phenols was detec-
ted at low concentrations. PNA and total organic levels decreased with in-
crease in power setting and were higher in the exhaust from low sulfur fuels.
Sulfur oxides measured by wet chemical techniques showed that a good material
balance was obtained between fuel bound sulfur and the S02/S03 in the ex-
haust gases.
IV
-------
Results of this effort indicate that the sampling system shows good po-
tential for the collection of participates but that further development is
needed for application of the system to larger gas turbine engines such as
the JT8D. The program also identified the chemical analysis techniques and
the type of future measurements which would yield meaningful data in the
assessment of particulate emissions.
This report is submitted in fuIfiIIment of EPA contract 68-02-2458 by
United Technologies Corporation under the sponsorship of the Environmental
Protection Agency. This report covers the period November 5, 1976 through
March 31, 1978. The technical effort was completed in February 1978.
-------
CONTENTS
Preface iii
Abstract iv
Abbreviations ix
Acknowledgment xi
Section 1. Introduction 1
Section 2. Conclusions 3
Discussion of Conclusions 4
Section 3. Recommendations 6
Discussion of Recommendations 6
Section 4. Technical Discussion 7
Sampling System 7
Nature of the particulates to be sampled 7
Sampling methods 9
Design criteria 9
System hardware 9
Probe 11
System cooling 15
Fi Iter materials 15
Sample degradation 16
Temperature control 18
Flow measurement 18
System operation 22
Packed bed fi Iters 22
Test Vehicle 23
Engine 23
Combustor 24
Test stand 24
Gas generator instrumentation 26
Emission instrumentation 27
Trial Runs 29
Exit pipe mapping 29
Test Procedure 35
Phase I 38
Phase II 39
Sample Identification 40
Mass Emissions Measurement Technology 42
Filters 43
Balance 43
Mass emissions testing 43
Smoke Measurement Methodology 45
Conclusions 47
Analytical Procedures 49
Sample treatment 49
vii
-------
CONTENTS (Cont'd)
Organic analysis 50
Total organics 50
High performance liquid chromatographic analysis 55.
Gas chromatograph-mass spectrometer analyses 68
Packed Bed Filter Studies 76
Nitrosamine analyses 89
Phenol analyses 91
Spectral data 94
Nuclear magnetic resonance analyses 94
Ultraviolet analyses 100
Infrared analyses 106
Fuel analysis 107
Boiling point distribution 110
Sulfur oxides emissions 117
Proton activation analysis/x-ray analysis 139
Elemental analysis 142
References 144
Bibliography 147
Appendices
A. GC/MS Analysis of Polynuclear Mixes and Typical Turbine 152
Combustor Exhaust
B. PNA Contribution from Filters and Solvents 158
; vui
-------
ABBREVIATIONS
BAP Benz (a) Pyrene
BP Boiling Point
El Emission Index (lbs/1000 Ibs fuel)
EPAP EPA Emission Parameter (lbs/100 SHP/cycle)
ESFC Effective Specific Fuel Consumption
ESHP Equivalent Shaft Horse Power
F/A Fuel to Air Ratio
GC/MS Gas Chromatograph Coupled to Mass Spectrometer
HC/THC Hydrocarbon Emissions
HP Engine Horsepower
HPLC High Performance Liquid Chromatograph
IR Infrared
JT8D P&WA Jet Turbine Engine
JT9D P&WA Jet Turbine Engine
M Molecular Weight
NG Gas Generator Speed
NMR Nuclear Magnetic Resonance
NO (NO + N0?) Emissions
J\ C-
P Pressure
PAH Polyaromatic Hydrocarbons
PNA Polynuclear Aromatics
POM Polynuclear Organic Material
PT6A-45 P&WAC Turbo-Prop Engine
SHP Output Shaft Horsepower
SLS Sea Level Static
T Temperature
'S Gas Temperature at First Turbine Stage Exit
UV Ultraviolet
IX
-------
w Mass Flow
. Gas Genere
P Pressure Drop
w Gas Generatbr Fuel Flow
u Atomic Hydrogen - Carbon Ratio of Fuel
P
8 Ambient Pressure Ratio
29792
6 Ambient Temperature Ratio
59.0°F
Subscripts
0 Ambient
1 Engine/Gas Generator Inlet
3 . Compressor Inlet
4 Compressor Exit
5 First Turbine Stage Exit
7 Engine Exit
CB Cabin Bleed
-------
ACKNOWLEDGMENT
The cooperation of Pratt & Whitney Aircraft of Canada, Ltd (PWACL) and
Dr. R. H. Groth, Chairman of the Department of Chemistry, Central Connecticut
State College, New Britain, Connecticut is gratefully acknowledged. PWACL
participated in the design of the high volume sampling system and in the col-
lection of particulate samples from a PT6A-45 gas generator. Comprehensive
chemical analyses and interpretation of the data were performed by Messrs. T.
J. Blasko, A. G. Glastris and M. D. Kahn of P&WA in conjunction with Dr. R.
H. Groth. Further acknowledgments are given to Mr. J. H. Elwood, P&WA Program
Manager, and D. J. Robertson who assisted in management of the program, and
to Mr. J. N. Braddock, EPA Program Manager, who guided and monitored the per-
formance of the program.
-------
SECTION 1
INTRODUCTION
The overall intent of this work was to aid the EPA and the industry in
assessing the risk associated with the emission of particulates from gas
turbine engines on which are adsorbed complex organic species. This work
was accomplished during a 14 month program in three phases:
Phase I: Engine emission demonstration
Phase II: Exhaust particulate collection
Phase III: Chemical analysis :and interpretation
The test vehicle selected for this program was the Pratt & Whitney
Aircraft PT6A-45 gas generator which is representative of current produc-
tion, high population small gas turbine engines.
An additional requirement of the contract called for the design and
development of a high volume particulate collection scheme specifically ad-
apted for gas turbine engine testing. This requirement is critical because
experience has shown that the collection system used often defines the na-
ture of the particulates collected. This can be especially true when work-
ing with volatile species such as polynuclear aromatic compounds (PNA).
The engine emissions demonstration phase provided data which demon-
strated P&WA's ability to operate the test vehicle in a controlled and re-
peatable fashion with respect to power, gaseous emissions, smoke and partic-
ulate mass emissions. This was accomplished during a series of five trial
runs over the usual power ranges (idle, approach, climb and takeoff).
More than 100 particulate samples were obtained for a wide variety of
chemical analyses. These samples encompassed the whole range of engine
power settings using the standard Jet A-l fuel as well as Jet A-l doped with
0.26% sulfur to evaluate the effects of fuel bound sulfur on emission char-
acteristics. The effectiveness of the high volume sample system was limited
only by the occasional high ambient dewpoint and temperature and by the need
to control sample filter temperature in order to preserve the integrity of
the volatile organic species.
A comprehensive chemical analysis of the organic material extracted
from the particulate matter was undertaken with the primary emphasis on
polynuclear aromatic hydrocarbons, especially those considered possibly car-
cinogenic. The analyses ranged from simple infrared and ultraviolet absorp-
tion spectroscopy to sophisticated nuclear magnetic resonance (NMR), compu-
ter aided combined gas chromatograph/mass spectrometer (GC/MS) and high
1
-------
performance liquid chromatography (HPLC). In addition, specific tests were
performed to detect the presence of phenols and nitrosamines. The extremely
low concentrations of significant organic species taxed the detection limits
of many of the state-of-the-art analytical procedures.
Following the summary of major conclusions is a detailed description of
the analyses performed and the results obtained.
-------
SECTION 2
CONCLUSIONS
1. A high volume sampling system for collection of participates from gas
turbine engines was designed, fabricated and adapted to a Pratt &
Whitney Aircraft PT6A-45 gas generator. The system collected a suffi-
ciently large sample for the chemical analyses planned considering the
low concentration (approx. 10 mg/m^) of particulates in turbine ex-
haust.
2. Control of the sample collection system temperature and flow rate is
essential due to the volatility of the organic species under investi-
gation. Though this volatility is a known sampling problem, the high
temperatures encountered in gas turbine sampling necessitate precise
control and monitoring at the filter surface.
3. The purity of solvents and filters is critical at low levels and
therefore purity must be established and maintained. Many spectro-
quality solvents and filters evaluated contained interfering sub-
stances which would severely bias the analytical results if used in
sampling the small concentrations of organic species found in gas tur-
bine exhaust.
4. Due to the extremely low concentrations of organic species found and
the wide variations in sample humidity, temperature and flow condi-
tions found in gas turbine exhaust, interpretation of the data should
be primarily on a qualitative basis with little emphasis on the abso-
lute numbers.
5. The organic material entrained on particulates emitted from gas tur-
bine engines is only a small fraction of the total organics emitted.
However, due to the respirable nature of the particulates, their anal-
ysis is of considerable significance.
6. The multitude of chemical analyses performed revealed the presence of
numerous polynuclear aromatic compounds. Aromatic compounds with one
ring or two fused rings were in an order of magnitude more abundant
than the PAH having three or more fused rings. The vast majority of
these compounds were the small, 3 to 4 fused ring compounds, with very
few 5 to 6 fused ring compounds present. The concentrations were ex-
tremely low and very few of the compounds are known carcinogens. The
maximum amount of polynuclear material in any one sample was less than
2 ppb. Total amount of carcinogens such as benz(a)pyrene and benzo-
phenanthrene were an order of magnitude less.
-------
7. No nitrosamines were found.
8. The presence of phenol was noted but at a very low concentration (part
per tri I lion).
9. The concentrations of polynuclear aromatic hydrocarbons in the exhaust
samples follow the overall hydrocarbon trend which decreases with in-
creasing power setting.
10. Results obtained from very diverse analytical techniques, e.g., NMR,
HPLC, GC/MS and total organic measurements were consistent.
11. A good material balance (within _+ 6%) was obtained between fuel bound
sulfur and the S02/S03 in the exhaust gases.
12. There was some indication that the levels of oxygenates and polynu-
clear aromatic hydrocarbons are higher in the low sulfur fuel exhaust
samples.
13. The only known (2) carcinogenic PAH identified were benzofluoranthene,
benzophenanthrene, and benz(a)pyrene. All of these compounds were be-
low 0.1 parts per billion concentration.
DISCUSSION OF CONCLUSIONS
During the sampling operations in Canada, the ambient temperature and
relative humidity varied considerably. As a result the temperature and mois-
ture content of the exhaust also showed wide variations. Engine power set-
ting also contributed to these variations. At lower power settings (idle)
moisture condensed on the filters and seriously affected the flow character-
istics of the filter. Consequently the amounts of particulate matter and
adsorbed materials were lowered substantially. The total flows were thus
only an approximation in some cases. The presence of variable amounts of
moisture also affected the quantity of adsorbed matter. The results obtained
were therefore a qualitative indication only and not an absolute quantita-
tive assay.
The high volume sampling system was found to be satisfactory for fil-
tering up to 45 m3 of exhaust gas and to yield an adequate size sample in
a reasonable time. Moisture and temperature problems with the sampling sys-
tem represent areas of future development if quantitative data is needed.
The system was adequate to provide a qualitative picture of the chemical na-
ture of the particulate.
Measurement of the total organics and the PNA by gas chromatography-
mass spectroscopy from packed bed filters (Chromosorb 102) showed that less
than 1% of the organic material is adsorbed on the particulate matter and
over 99% passed through the Mitex filter. This small amount however, could
be carried along with the partiallates and become lodged in the lungs.
Thus, it could be of great significance from a health standpoint.
-------
Polynuclear aromatic hydrocarbons were found by GO/MS and HPLC tech-
niques. Mostly these PAH were non-carcinogens and of the 3 to 4 fused ring
compounds. The GC/MS permitted specific identifications but the HPLC, under
the conditions employed, did not fully resolve the complex mixtures. The
HPLC did however give a good indication of the relative amounts of compounds
in the 3 to 4 fused ring types versus the 5 to 6 fused ring types. Very few
of the larger fused ring compounds were found and these were in very low
concentrations.
Nitrosamines were not found but at the temperature occurring in the ex-
haust stream they would likely be unstable even if formed in the engine.
Phenol analyses were limited to the several compounds for which the EPA;pro-
cedure (EPA-650/2-75/056) was developed. This does not mean that other phen-
ols or oxygenates are absent. The levels found and the occurrence in actual
exhaust samples of these few phenols were low.The concentrations of PAH in
the exhaust decreased with increasing engine power setting. This result was
indicated by the data in the HPLC and GC/MS analyses. Because of sampling
variations, this result should be considered qualitative. The general agree-
ment between the two methods support the qualitative generalization. Sim-
ilarly, a correlation between, a) oxygenate level and PAH level, and b) the
sulfur levels in the fuels used is also supported by these two measurement
techniques. Total organics measurements further corroborate the trend of
higher organics with low sulfur fuel and with lower engine power setting.
Both gas flow and temperature elevation reduce the collecting efficien-
cy for benz(a)pyrene. An even more serious loss would occur with lower mol-
ecular weight (fewer fused rings) compounds. Therefore, the temperature of
collection is very critical.
Sulfur oxides measured in the exhaust gases by wet chemical procedures
agree well with the sulfur analyses of both the high sulfur and low sulfur
fuels. This suggests that virtually all of the sulfur is emitted as S02/
-------
SECTION 3
RECOMMENDATIONS
1. Advanced design work on the sampling system should be carried out to
improve flow measuring characteristics, temperature regulation and col-
lection efficiency to obtain more quantitative and reproducible data.
2. The present sampling system should be adapted to measure the mass emis-
sions and chemical characteristics of partiallates emitted from a high
population, large gas turbine engine such as the JT8D and JT9D.
3. Future measurements should be extended to include materials collected
on packed bed filters, such as Chromosorb 102 followed by cryogenic
trapping to evaluate the efficiency of collection.
4. The analytical technique for organic materials measurement should be
limited to gas chromatography, high performance liquid chromatography and
gas chromatograph/mass spectrometry.
5. Analysis such as boiling point determination, NMR, UV, IR and elemental
should be omitted since they yield information of limited value.
DISCUSSION OF RECOMMENDATIONS
The current sampling system was found to have problems associated with
flow measurement, humidity and temperature control. For a sampling system
to be adapted to large engines such as the JT8D, specific parameters must be
considered, such as time available for sampling and temperatures associated
with the exhaust stream.
The high population engine such as JT8D is more likely to be subject to
regulation and for this reason, as well as for its greater usage, the nature
of its effluent both adsorbed on the particulate matter and also that por-
tion collected on the packed bed filter must be determined. Some early
6C/MS analyses of samples from an JT8D style experimental combustor showed
the presence of some of the same PNA compounds and should be investigated
further. Some details of this work are given in Appendix A. Preliminary
studies have shown that under 1% of the total sample, organics and PNA are
adsorbed on the particulates. Additional material may pass through the
packed bed filter and hence cryogenic trapping is suggested to recover it.
The analyses which yielded the most significant information in this
study were, phenol-nitrosamine, HPLC, total organics and 6C/MS. Other anal-
ysis specifically: boiling point determination , NMR, UV, IR, and elemental
gave little useful information for these complex mixtures.
-------
SECTION 4
TECHNICAL DISCUSSION
SAMPLING SYSTEMS
Nature of the Particulates to be Sampled
In spite of the considerable amount of work done by Pratt & Whitn
Aircraft Group, Division of United Technologies Corporation and others
there is little agreement as to what is considered particulate matter. If
particulate matter is collected simultaneously using existing techniques,
there is little likelihood of agreement in terms of the absolute amounts and
composition of the material collected. Therefore, it has become the practice
to define particulates in terms of the method of collection and analysis.
Considerable work is being done in government and private agencies to stand-
ardize a method of measurement and to interpret what the method actually
does. However, this current program has contributed to and enhanced our un-
derstanding of particulate emissions.
Particulate matter emitted in the exhaust of gas turbine engines is
known to consist of aerosols, finely divided carbon and other particles.
Aerosols are typically made up of unburned and partially burned fuel, sul-
fates formed from the sulfur in the fuel, trace elements normally found in
fuel, water droplets containing combustion byproducts, material ingested
into the engine inlet, and materials attributable to normal wear processes
in the engine. All of these particles may have possible toxicological or
carcinogenic effects. For the various classes of organic species likely to
be present, the anticipated variability in toxicity and perhaps smog forming
capability makes it desirable to obtain specific qualitative and quantita-
tive detail. Many of the particulates mentioned have polycyclic organic mat-
ter (POM) associated with them. These POM compounds are made up largely of
complex organic hydrocarbons whose structure includes three or more fused
rings, possibly aromatic. Some of these compounds have shown some evidence
of carcinogenic effects when applied to rats, and there is some thought that
similar effects might be obtained in humans^2).
Polycyclic organic matter is highly reactive, and considerable care
must be taken in handling to preclude or minimize sample degradation. Sul-
fur trioxide, along with other atmospheric oxidants, and photo-oxidation
will degrade these POM compounds. Degradation reactions are particularly
accelerated when the compounds are adsorbed on carbonaceous material such as
is found in gas turbine engine exhaust. The collection and preservation of
POM compounds for analysis requires special attention, particularly to pre-
vent the loss of volatile organic compounds'2).
-------
A considerable body of evidence has been accumulated in recent years
suggesting that POM compounds are found as normal by-products of fossil-fuel
systems. It is anticipated that gas turbine engines are no exception. For
example, studies at Moscow Airport resulted in the finding of benz(a)pyrene
(BAP) which was attributed to jet aircraft. Similarly, earlier work per-
formed under sponsorship of the Air Force School of Aerospace Medicine and
at Pratt & Whitney Aircraft, resulted in the identification of a number of
POM compounds in gas turbine combustor exhaust (3,4,5,6,7,8,9).
Evaluation of airborne particulate matter has resulted in the identifi-
cation and classification of numerous compounds, of which several are sus-
pected of being strongly to mildly carcinogenic. Investigation of these com-
pounds has resulted in a number of analytical methods for their measurement
and in an understanding of the requirements for sampling, sample handling,
and the quantity of material required for analysis (10,11,12,13,14,15,16).
The many studies of diesel and automotive exhaust sources have resulted
in the identification of a number of POM compounds, and it is logical to ex-
tend these investigations to gas turbine aircraft powerplants.
It is also known that the amounts and types of POM compounds present in
automotive exhaust are dependent upon the fuel used and the fuel-to-air-
ratio. Tests have shown that fuel that is rich in aromatics, produces more
POM compounds, particularly polynuclear aromatic (PNA) compounds, than does
fuel having less amounts of aromatic compounds. In addition, certain
amounts of nitric oxide in the exhaust will lower the PNA content. It is
reasonable to assume that the same phenomena will hold for gas turbine en-
gines.
In addition to organic compounds such as POM, there are other sub-
stances of interest in the exhaust. The presence of nitrosamines (known car-
cinogens) has been reported in food, air, water, and diesel exhaust. It
seems likely that they would also occur in the combustion by-products of gas
turbine engines. As in POM compounds, it is anticipated that the nitrosa-
mines will be present in very small quantities, necessitating large volume
sampling. However, like POM compounds, although present in very small quan-
tities, nitrosamines may still have environmental impact due to high toxi-
city or carcinogenicity. Analytical techniques have been developed recently
which permit the separation and measurement of the various nitrosamine com-
pounds (17,18,19,20,21).
Other materials found in aircraft gas turbine exhaust are more well
known and do not pose any particular problems in either collecting or analy-
zing samples; however, problems can be encountered in obtaining samples for
sample weighing. A significant portion of the sulfate fraction collected on
a sample can be attributed to suIfuric acid, which is extremely hydroscopic.
Extreme care, therefore, must be taken in handling and weighing of the
filters (22,23).
-------
Sampling Methods
Our experience has been that particulate materials in gas turbine en-
gine exhaust are found in concentrations on the order of 5 to 10 milligrams
per cubic meter. In collecting particulate samples for the separation and
identification of organic compounds, using glass fiber filters, we have de-
termined that a minimum of thirty-five cubic meters of exhaust gas should be
filtered to allow for quantitative as well as qualitative analysis for or-
ganics obtained by Soxhlet extraction of the glass fiber filters (24).
•'i
Filtering this much exhaust gas using ordinary EPA filtering techniques
requires a large amount of time, necessitating long engine operation times
that would result in making sample collection prohibitively expensive. How-
ever, sampling time could be decreased by using a large filter (293 mm dia.)
with a large capacity vacuum pump. A system capable of collecting enough
sample material in less than 30 minutes of running time per sample point was
considered a reasonable objective (25,26,27). Other factors were considered
such as the possibility that volatile organic compounds, including some PNA
and N-nitrosamines, would not be collected on the filter.
To investigate collection efficiency for the organic compounds, P&WA
evaluated packed bed filters packed with polymeric beads. The polymeric
beads were packed in a chromatographic type column capable of handling suf-
ficient sample flow (28,29). The polymer columns were returned to the P&WA
Physics and Chemistry laboratory and the organics were extracted using the
standard Soxhlet apparatus.
Design Criteria
The High Volume Sampling System was designed on the assumption that
5-10 mg/m3 exhaust particulates would be found at the exhaust plane of the
PT6A-45. It was also assumed that to accomplish sampling in a reasonable
time period (approximately 1/2 hour) and to achieve the estimated 0.5 grams
sample considered desirable for organic analysis, a sampling rate of about
3.3 nH/min would be necessary.
It was initially considered necessary to reduce the sample gas tempera-
ture from a 1580°F maximum at takeoff to no more than 250°F at the fil-
ter surface. Sample degradation studies conducted after the sampling system
was constructed, indicated that the filter temperature should be further re-
duced to 160°F. To reduce the sample temperature, a significant degree of
water cooling was considered necessary. However, at the same time sample re-
sidence was kept below 5 seconds in the sample lines to minimize sample loss
on the walls of the sampling system.
System Hardware
A sampling system shown schematically in Figure 1 was designed in which
the pressure at the exhaust plane provided a portion of the sample flow.
This flow was augmented by use of a Roots* 3514J vacuum blower (Figure 2).
*0resser Industries, ConnersviIle, Ind.
-------
LINE \R RAKE
SAMPLE PLENUM
S. S. HEAT EXCHAMGER
TUBE
30 =74 RESISTOFLEX 'TEFLON CORE'
~|
_
A A A293 0/JTEF
2" ORIFICE METER TUBES
1.6" ORIFICE
FILTER HOLDER
316 S.S.
LON f ILTER
10 =24 AEROQUIP
+ IV BALL VALVE
5' ^=24 AEROQUIP
PUMP MANIFOLD
VAC RELIEF VALVE
f (SET 8" HGI
ROOTS 3514J PUMP
800 SCFM »>>4" HC,
Figure 1. High volume sampling system block diagram.
Figure 2. Prototype high volume sample system components,
10
-------
The blower was capable of 800 ft3/min @ 4" Hg vacuum and
the primary driving force at idle, where ram pressure in
was minimal. Flow was monitored by
1.6" orifices coupled to the exits
3, 4, 5, 6) sized to accept 293 mm
were 1" 316 stainless steel to the
lines were 1-1/2" Resistoflex (Tef
was considered
the exhaust plane
a system of 5 orifice meter tubes with
of 5 cone shaped filter holders (Figures
diameter circular filters. Sample lines
final heat exchanger. After this point
on core).
Probe
The probe designed was a five point linear rake (Figure 7, 8) mounted
in an 18-inch section of exhaust duct (Figure 9, 10) immediately behind the
engine. The probe elements were 3/4-inch I.D. 316 stainless steel tube set
at centroids of equal area within the duct and reinforced at critical stress
points with Hastalloy. It was calculated that 3/4-inch orifices would be
necessary to avoid a choked flow condition at idle at 3.3 m-^/min per fil-
ter element. The 316 stainless steel proved to have sufficient temperature
tolerance to avoid high temperature oxidation throughout the test program.
The five tubes of the linear rake were coupled to a 12-inch mixing ple-
num outside the exhaust duct. This plenum was designed to average out any
differences in sample composition between each of the five probe elements.
Some radiational cooling of the exhaust gas was also expected.
Figure 3. Orifice meter flow measuring tubes
l l
-------
r. MPT
Figure 4. 293 mm filter holder
assemb led.
CONSTRUCTION 3I6S S
FILTFR DIAMETER 293mm
FILTER TVPE MILLJPORE tQJJMITEX
ft f LOW 100-1 7b SCFM
Figure 5. High volume filter holder
assembly.
Figure 6. 293 mm filter holder disassembled.
12
-------
Figure 7. Sample rake and plenum chamber,
I
i
o
o
o
o
PLENUM <6" X 1?")
6" S.S.PIPE FITTED W 5 ^?4
SITTINGS AND 4 ^FITTINGS
I INE AH HAKE TUBING SIZE 0.777 I.I) 0 R?6 O D
TAILPIPE BLOCKAGE AHEA APPROX 83 IN
Figure 8. Linear rake and sample plenum.
13
-------
Figure 9. Rake installed in exhaust duct,
Figure 10. Sample probe in exhaust duct,
14
-------
System Cooling
It was initially planned to cool the sample gas using a 4 foot long
water cooled heat exchanger (Figure 11). During the testing, this proved to
be insufficient for cooling. Additionally, the degree of radiational cooling
anticipated to occur before the heat exchanger was below expectations.
Prior to the test program, the heat exchanger was enlarged by adding ap-
proximately 20 feet additional stainless steel tubing placed in a water
filled trough. The water flow in this trough and consequently the tempera-
ture was continuously variable and controllable. The entire distance from
the sample plenum at the probes to just ahead of the filter housings was
water cooled and this cooling was found to be sufficient at all power set-
tings.
Figure 11. Heat exchanger,
Fi Iter Materials
A number of filter materials were examined for stability at tempera-
tures up to 250°F, solvent compatibility and interfering substances when
subjected to HPLC analysis for PNA. Filters considered were standard Mi Mi-
pore, PVC, Mitex, Fluoropore, Gelman type A glass fiber, type E, type A-E,
and Nuclepore. Mitex, a pure Teflon filter, was ultimately chosen for its
total absence of interfering contaminants, its high temperature and solvent
compatibility and its high strength. Use of a Gelman type A-E filter, fired
at 500°C for 1 hour to combust contaminants, was considered. It was free
of contaminants and was compatible with high temperature, however, its mech-
anical strength was so reduced as to make use of these filters undesirable.
15
-------
Details of a preliminary study on solvent and filter selection for PNA
analysis are given in Appendix B.
Sample Degradation
To best determine the maximum desirable filter temperature during test,
a series of experiments was conducted to measure percent sample recovery of
BAP after exposure to elevated temperature and airflow. Table 1 shows the
results of placing 0.0050 mg BAP in a pyrex disk in an oven for 30 minutes
at temperatures ranging from 72°F to 230°F. Sample loss is apparent
above 160°F. Table 2 shows the results of placing 47 mm Mitex filters
doped with 0.0050, mg BAP in an oven for 30 minutes at temperatures ranging
from 72°F to 230°F. Considerable loss is seen to occur at some point be-
tween 130°F and 160°F. This is a "worst case" situation since the pres-
ence of carbon (typical of turbine exhaust) would reduce the losses substan-
tially. Table 3 shows the results of similar samples exposed to tempera-
ture, but with the addition of a 40 l/min airflow. The degree of sample loss
is shown to be further aggravated by airflow.
TABLE 1. BAP DEGRADATION, IN PYREX DISHES
Temperature Percent Recovery
72<>F 100
130 102
160 . 77
200 72
230 18
TABLE 2. BAP DEGRADATION ON MITEX FILTERS
Temperature Percent Recovery
Baseline Blank
72°F
130
160
200
230
100
94
97
66
62
38
16
-------
TABLE 3. BAP DEGRADATION ON MITEX WITH AIRFLOW*
Temperature
Percent
Baseline Blank
720F
175
218
100
93
36
22
*40 liters/min
Figure 12 graphically illustrates the effect of temperature on sample
recovery under the conditions described above. BAP on Mitex and with airflow
suffers the greatest sample loss at any given temperature when compared to
the effects of temperature on BAP alone and BAP on Mitex.
Based on the above findings, it is theorized that lower molecular
weight substances may volatilize more readily than BAP and at a lower tem-
perature, thus contributing to sample loss at temperatures well below the
160°F taken as a maximum sampling limit.
01
o.
e/j
Z
5
a.
o
o.
O
I-
01
U
cc
100
80
60
40
,n
20
& ^
O BAP IN PYREX DISHES
A BAP ON MITEX FILTERS
D BAPONMITEXW/
46 L/MIN AIRFLOW
70
90
110
130
150
170
190
210
230
250
TEMPERATURE
Figure 12. Effect of temperature on sample recovery.
17
-------
Temperature Control
Temperature control involved the controlled metering of cooling water
through the cooling trough and heat exchanger while instantaneous control
was achieved by varying the sample flow to each filter using 1-1/2-inch ball
valves. A chrome I-aIumeI thermocouple located 1/2-inch above the center of
each filter was used for temperature measurement. At no time was the tem-
perature permitted to exceed 160°F. This maximum was considered a criti-
cal factor and was based on a balance between data obtained in the BAP deg-
radation studies and a need for obtaining a sizeable quantity of particulate
matter.
The temperature control maximum of 160°F hindered test point starts
especially at idle and approach. At start-up the filter housing temperature
was below the dew point of the sample gas, enough to condense water and wet
the filters. Under these circumstances, flows often could not be increased
sufficiently to bring the housing and filter above the dew point of the sam-
ple gas. An alternative start-up approach which reduced but did not elimin-
ate the problem was to empty the trough and heat exchanger and start the
flow of cooling water only after the gas temperature had brought the filter
housings to 160°F. In this way most of the wetting could be avoided and
flows could be maintained at diminished yet respectable levels.
Flow Measurement
Flows were measured using a system of five orifice meter tubes coupled
to the exit of each of the five filter housings. Absolute pressure was mea-
sured upstream of the orifice plate using a Wallace and Tiernan* gauge. AP
across the orifice was measured using a system of Magnehelic** gauges and
upstream and downstream temperatures were measured using chrome I-a IumeI
thermocouples connected to a Doric* digital readout. Flow data for each
filter run were calculated using the compressible flow equation:(30)
/ (Px) (AP)
Wa = (3105.44) d* AE / _J (Fa) (Fpv) (F^) (Fp)
where
Actual flow (#/hr)
orifice diameter (1.6")
upstream pressure (psia)
*Wallace and Tiernan, Belleville, New Jersey
**F. W. Dwyer Mfg. Co., Mich. City, Ind.
+Doric Scientific, San Diego, Calif.
18
-------
AP
pv
B
wv
E
A
pressure drop (psi)
upstream temperature (°R)
area factor correction
Supercompressibility correction
water vapor correction
density correction
Expansion factor
Coefficient of discharge
Water was a problem in the recording of flow data. Orifice meter tubes
at idle and at approach often operated below the dew point of the gas and as
a consequence, the pressure taps to the Magnehelic gauges (AP) filled with
water and failed to operate properly. The tubes were emptied of water when-
ever the problem ocurred; however, some uncertainity as to the actual flow
does exist for some test points at low power. The data in every case was ex-
amined for discrepancies in AP between filters in the same test run and cor-
rected to the test point average where a Magnehelic gauge was clearly inop-
erative.
Figure 13 illustrates the typical flow variations experienced from the
start of a test run. Effects encountered in the first ten minutes of every
test run were those of temperature and the flow reduction required to stay
within the 160°F maximum. Particulate loading also reduced flow with time.
Table 4 gives sampling time, tailpipe temperature, plenum temperature,
filter surface temperature and total flow for each filter sampled. Sampling
times ranged from 30 to 95 minutes, fi Iter surface temperatures ranged from
89°F to 161°F and total flow ranged from 12.9 m3 to 50.8 m3.
1.1i—
i.o
0.9
0.8
0.7
0.6
10JUMITEX FILTER INSTALLED
10
15
20
25
30
35
40
45
TIME (MINUTES)
Figure 13.
Typical flow variations while running.
19
-------
TABLE 4. SAMPLING CONDITIONS
Filter
Identification
LC/UV 1A #1
P/N 1A
IR 1A
NMR 1A
GC/MS 1A #1
BP 1A
T-ORG 1A #1
EPA 1A
EL 1A
LC/UV 2A #1
GC/MS 2A #1
BP 2A
T-ORG 2A #1
EPA 2A
EL 2A #1
LC/UV 3A 11
GC/MS 3A #1
BP 3A
T-ORG 3A #1
EL 3A #1
P/N 3A
IR 3A
NMR 3A
LC/UV 4A #1
GC/MS 4A#1
BP 4A
T-ORG 4A #1
EL 4A fl
LC/UV 1A #2
GC/MS 1A #2
T-ORG 1A #2
EL 1A #2
LC/UV 2A #2
GC/MS 2A #2
T-ORG 2A #2
EL 2A #2
LC/UV 3A # 2
GC/MS 3A #2
T-ORG 3A #2
EL 3A n
EPA 3A
LC/UV 4A #2
GC/MS 4A #2
Tai Ipipe
Temp°F
1208
1208
1208
1208
1208
1219
1219
1219
1219
1182
1182
1174
1174
1174
1174
1580
1580
1580
1580
1580
1578
1578
1578
1578
1578
1578
1578
1578
1235
1235
1235
1235
1187
1187
1187
1187
1531
1531
1531
1531
1531
1566
1557
Power
Setting
Idle
Idle
Idle
Idle
Idle
Idle
Idle
Idle
Idle
Approach
Approach
Approach
Approach
Approach
Approach
Climb
Climb
Climb
Climb
Climb
Climb
C 1 imb
C 1 imb
Take-off
Take-off
Take-off
Take-off
Take-off
Idle
Idle
Idle
Idle
Approach
Approach
Approach
Approach
Climb
Climb
Climb
Climb
Climb
Take-off
Take-off
Plenum
Temp°F
643
643
643
643
643
702
702
702
702
592
592
658
658
658
658
1086
1086
1086
1086
1086
1072
1072
1072
1050
1050
1050
1050
1050
697
697
697
697
698
698
698
698
1056
1056
1056
1056
1056
1059
1076
Filter
Temp°F
127
130
109
114
119
114
116
109
143
99
106
105
109
103
104
129
126
120
114
137
134
124
109
142
149
135
117
139
109
112
107
148
132
145
97
166
127
141
108
114
110
118
132
Sampling
Flow Time
(m3) (Minutes)
26.4
26.1
26.4
26.3
26.5
24.6
24.5
24.6
49.6
17.5
17.5
22.7
22.7
22.7
31.5
32.4
32.4
32.4
32.4
42.3
27.6
27.8
27.8
32.2
32.0
28.1
28.1
32.2
21.0
21.0
21.0
20.4
13.2
12.9
13.2
29.4
36.5
36.1
26.1
42.0
23.8
23.5
21.9
60
60
60
60
60
65
65
65
65
50
50
63
63
63
63
45
45
45
45
45
50
50
50
37
37
37
37
37
55
55
55
55
30
30
30
30
42
42
42
42
42
46
36
(Continued)
20
-------
TABLE 4 (Continued)
Filter
Identification
T-ORG 4A #2
EL 4A K
EPA 4A #2
6C/MS 4A f3
T-ORG 4A 13
EL 4A #3
LC/UV 4A 13
LC/UV IB
IR IB
NMR IB
GC/MS IB #1
BP IB
LC/UV 2B
GC/MS 2B #1
BP 2B
LC/UV 3B
GC/MS 3B #1
BP 3B
T-ORG 3B
EPA 3B
LC/UV 4B
GC/MS 4B #1
BP 4B
T-ORG 4B
EPA 4B
GC/MS IB #2
T-ORG IB
EPA IB
EL IB
GC/MS 2B #2
T-ORG 2B
EPA 2B
EL 2B
GC/MS 38 #2
EL 3B
NMR 3B
IR 3B
GC/MS 4B #2
EL 4B
Tai Ipipe
Temp°F
1566
1566
1566
1557
1557
1557
1557
1204
1204
1204
1204
1204
1148
1148
1148
1531
1521
1521
1521
1521
1564
1564
1564
1564
1564
1189
1189
1189
1189
1148
1148
1148
1148
1531
1531
1531
1531
1580
1580
Power
Setting
Takeoff
Takeoff
Takeoff
Takeoff
Takeoff
Takeoff
Takeoff
Idle
Idle
Idle
Idle
Idle
Approach
Approach
Approach
Climb
Climb
C I'imb
C 1 imb
Climb
Takeoff
Takeoff
Takeoff
Takeoff
Takeoff
Idle
Idle
Idle
Idle
Approach
Approach
Approach
Approach
Climb
Climb
C 1 imb
Climb
Takeoff
Takeoff
Plenum
Temp°F
1059
1059
1059
1072
1072
1072
1072
10T2
1042
1042
1042
1042
Filter
Temp°F
116
125
115
138
124
144
149
153
112
145
118
98
161
135
126
141
161
160
123
160
131
132
113
94
105
99
93
93
154
148
119
134
151
155
155
136
125
139
138
Flow
(n.3)
23.5
43.7
23.5
21.9
18.1
43.0
21.9
29.3
29.4
29.3
29.3
29.7
47.5
47.6
47.9
42.4
50.8
45.5
45.9
45.9
92.7
37.8
23.8
27.5
19.4
25.6
25.6
25.6
47.5
47.5
47.5
47.5
44.8
42.2
39.7
42.4
47.6
47.2
Samp 1 i ng
Time
(Minutes)
46
46
46
36
36
36
36
95
95
95
95
95
89
89
89
60
65
65
65
65
54
54
54
54
54
80
80
80
80
88
88
88
88
60
60
60
60
55
55
21
-------
System Operation
The high volume sampling system was found to operate in a manner con-
sistent with its design objectives. Sufficient material was obtained to per-
form all but the elemental analysis. In this particular case, a different
approach to the sample collection would have been necessary to obtain suffi-
cient material for meaningful results by the analytical method used.
Sample temperatures were kept within the temperature maximum of 160°F
found to be critical to BAP loss. The sampling temperatures experienced are
believed to be the best compromise between temperatures so low as to render
the system inoperable and temperatures high enough to volatilize the large
majority of organic material entrained in the particulate matter.
The orifice meter tube approach to sample flow measurement was demon-
strated to be essentially sound. However, an alternative scheme for measure-
ment of A P would be desirable to eliminate uncertainties in this measure-
ment.
Glass fiber filters would have provided considerable advantage in redu-
cing theAP across the filter. This high AP compounded problems with start-
up saturation of the filter with water and limited our use of the Roots
vacuum blower (8" Hg limit) in augmenting flow at lower power settings. The
low mechanical strength of fired glass fiber filters is, however, a consid-
erable drawback to their use.
Packed Bed Fi Iters
In addition to collecting particulates using the filter sampling system
designed for this test, a packed bed sampling device (Figure 14) was used to
sample relatively low exhaust gas volumes for both particulates and organic
vapors. These optional measurements were made because certain amounts of
organic material would be lost during any collection process designed for
particulates only.
Figure 14. Packed bed sampling device.
22
-------
The sampling device, a 1/2" O.D. x 6" long stainless steel tube was
packed with 7 to 12 grams of Chromosorb 102. The sampling rate was approxi-
mately 0.035 cubic meters per minute. The samples were obtained directly
from the sample plenum chamber shown in Figures 7 and 8.
TEST VEHICLE
The experimental version of a P&WA PT6A-45 gas generator was selected
for the program since it is representative of current production, high popu-
lation engines which are anticipated to be in production well into the
1980's. Over 10,000 engines of this type have been delivered to date. The
following discussion describes the gas generator used in this program.
Engine
The PT6A-45 represents the largest and most advanced version of the PT6
engine series. Table 5 shows the ratings and performance parameters of the
PT6A-45 engine.
TABLE 5. PT6A-45 PERFORMANCE PARAMETERS
Takeoff Rating
SLS-Std. Day ESHP/SHP
Consumption ESFC
Propeller Speed RPM
Takeoff
Cruise
Mass Flow at T.O. Ibs/sec air
Compressor Pressure Ratio
1174/1120
0.560
1620/1700
(max torque limited/max speed)
1425
8.6
9:1
The compressor of the PT6A-45 consists of three axial stages combined
with a single centrifugal stage. The combustion chamber is of the annular
reverse flow type, with 14 fuel nozzles spraying tangentially. The first
stage turbine downstream of the combustor drives the compressor. Combustor
conditions are simulated with a back pressure valve downstream of the com-
pressor turbine.
23
-------
Combustor
The combustors in use with PT6A-45 engines are small and hence highly
loaded (5.1 x 106 BTU/hr. atm. ft3). They utilize 14 simplex fuel noz-
zles of Flow Number 1.9. Other characteristics of the combustor are shown
in Table 6.
TABLE 6. PT6A-45 COMBUSTOR PARAMETERS (S.L.S.T.O.*)
Combustor mass flow (Ws) 8.6 Ibs/sec.
Pressure (Pa) 9.0 atm.
Inlet Temperature (Ts) 1071°R
Outlet Temperature (T/0 2460°R
Pressure Loss 2.3%
Mref 0.0269
Outside Diameter (D0) 15.71 in.
Inside Diameter (Dj) 11.65 in.
Length 5.41 in.
Combustor Volume 0.27 eft.
Temperature Pattern Factor _ 0.15 - 0.18
*Sea Level Static Takeoff Condition
A low emission combustor was used during the test program. Although
the combustor profile is identical to the Bill-of-Materials configuration,
the flow splits as well as cooling arrangements were modified to reduce ex-
haust emissions as well as improve combustor life. Figure 15 is an emissions
profile of the low emission combustor.
Test Stand
The gas generator was tested in a facility shown schematically in Fig-
ure 16. The intake fan supplies are to the gas generator at pressures up to
1" of water. The intake air supply ensures uniform intake temperature dis-
tribution (5°F max variation) to the gas generator.
24
-------
NORM NO (X KT>I BPM
Figure 15. Emissions profiles from low emission combustor.
AIR FROM
" BLOWER
Figure 16. Schematic of typical gas generator test facility.
f
25
-------
Combustor operating conditions on the gas generator are set up with a
remotely actuated back pressure (butterfly) valve. The exhaust from the gas
generator tail pipe is led to an exhaust duct which is kept at reasonable
temperatures through an air ejector downstream of the butterfly valve. The
test facility is also equipped with heaters for increasing inlet air temper-
ature during winter operation.
The exhaust pipe between the gas generator and the butterfly pipe was
instrumented for gas analysis and particulate sampling.
Gas Generator Instrumentation
The gas generator was instrumented extensively to monitor all para-
meters normally required to evaluate performance. These included air and
fuel flow rates into the combustor, temperatures at gas generator intake,
combustor intake, compressor turbine exit and gas generator exit. The gas
generator was also instrumented to measure combustor inlet (93) and outlet
(P4) pressures, so that determination of combustor pressure drops ( A P/P)
can be made.
Photographs of the gas generator test facility and control pane
shown in Figures 17 and 18.
are
Figure 17. Combustor rig control room.
-------
Figure 18. Combustor rig gas generator test section.
Emission Instrumentation
Exhaust gas analysis was undertaken with a Scott Model 108- Mk. Ill ex-
haust gas analysis system. The system comprises of the following instru-
ments:
Beckman Model 865-14
NDIR Analyzer for CO
Beckman Model 864-23
NDIR Analyzer for C02
Beckman Model 951H
Chemiluminescence analyzer for NO, NOX
Scott Model 415 FID
for Hydrocarbons
Accuracy
1% of fulI scale
1% of fulI scale
1% of fulI scale
1% of fulI scale
21
-------
Flow schematic of the gas analysis system is shown in Figure 19. Sample
to the HC analyzer is maintained at temperatures of 150 + 5°C and down-
stream to the other instruments at 55 ^ 5°C. All additional components
such as valves, solenoids, pumps etc. are also heated to the same tempera-
tures. The system does a wet sample analysis and no desiccants, dryers or
water traps are used in the system. !
COMPONENTS IN
• / CONTACT WITH
J SAMPLE HEATED
f\ TO 12S°F=
COMPONENTS IN CONTACT
WITH SAMPLE HEATED TO 300°F
Figure 19. Exhaust emission instrumentation.
Emission measurements were made in accordance with EPA regulations
Federal Register, 17 July 1973. The emission sampling probes are situated
between the gas generator exhaust and the butterfly valve. A total of 12
sampling points provided for collection of representative samples. This was
achieved by arranging 4 sampling probes to form a cruciform within the tail-
pipe as shown in Figure 19.
Equipment for analysis of engine intake air consisted of all of the ex-
haust emissions analyzers, plus a separate Beckman Model 400 hydrocarbon
analyzer. Inlet air humidity was measured with an EG and G Cambridge Systems
Model 880 dew point hygrometer with a 'Peltier1 cooler and optical detector
28
-------
All calibration gases used on the exhaust emissions systems were pur-
chased from Scott Research Labs and certified to 2% accuracy. Calibration
gases for the Beckman Model 400 hydrocarbon analyzer are primary standards
supplied by Matheson of Canada Limited.
TRIAL RUNS
The hardware and test instrumentation were checked out by conducting
trial runs which included a series of emissions and performance tests. The
tests were undertaken at conditions simulating ground idle, approach, climb-
out and takeoff modes. In addition, emissions mapping of the exhaust pipe
was also undertaken to determine specie distribution at the exit from the
gas generator.
The combustor test conditions simulating the four operating modes of the en-
gine are shown in Table 7.
TABLE 7. PT6A-45 GAS GENERATOR TEST CONDITIONS
Opefating
Mode
Ground Idle
Approach
Climbout
Takeoff
Gas Gen.
Speed
NG (RPM)
22,
31,
37,
37,
500
400
100
700
Fuel Flow
Wf (pph)
155
300
600
640
HP
75
336
1008
1120
Remarks
Run
Run
Run
Run
to
to
to
to
mechanical
norma 1
norma 1
norma 1
ized*
ized*
ized*
conditions.
conditions.
conditions.
conditions.
* NG (NORM) = NG (MECH) ; Wf (NORM) = Wf (MECH)
Exit Pipe Mapping
' An exhaust plane mapping was made to confirm the relative homogeneity
of exhaust samples at various points within the exhaust duct of the PT6A. A
single point probe was used to do diametral traverses along four circumfer-
ential planes and gaseous emissions were measured at nine positions in each
plane. Tests were undertaken at combustor conditions simulating ground idle
and climbout.
The emissions traverses covered two 90° sectors of the exhaust pipe.
Table 8 summarizes maximum deviations in specie concentration relative to
the mean for ground idle and climbout operating modes.
29
-------
TABLE 8. DEVIATIONS IN SPECIE CONCENTRATIONS
Avg. Concentration
Max +
Deviation
%
C02
%
2.86
3.85
3.85
GROUND
HC
ppm
161.0
18.01
22.9
IDLE
CO
ppm
410.0
7.32
3.90
CLIMBOUT
NOX CO? HC
^\ CM»
ppm % ppm
31.7 4.29 -
7.32 5.62 -
11.62 4.40 -
CO
ppm
54.2
3.41
3.97
NOX
ppm
-
_
-
Wider specie distributions were observed with hydrocarbons and NOX at
idle then at climbout. However, reasonable C02 distribution which largely
determines local fuel-air ratios indicated generally good mixing at both idle
and climbout conditions (Figures 20 and 21).
Figure 20. Fuel-air ratio distribution in exhaust pipe (PT6A-45 gas generator
at idle condition).
30
-------
Figure 21. Fuel-Air ratio distribution in exhaust pipe (PT6A-45 gas generator
at climbout condition).
The gas generator was then set up for performance and emission check
runs. Exhaust pipe instrumentation included a ten point cruciform emission
probe and a four point (Ty) temperature probe. The gas generator was run
to a matrix similar to that planned for the final (Phase II) collection
runs. This included five cycles at idle and approach followed by two climb-
out, four takeoff and again three climbout mode tests. In each case all gas
generator performance as well as exhaust emission data were collected. The
generator was run without any accessory loads or bleeds. Table 9 shows a
summary of the gas generator data and emission indices of THC, CO, C02 and
NOX. The emission index is a means of expressing the emission characteris-
tics of a combustor in relation to the fuel consumed. It is typically ex-
pressed as pounds of pollutants per thousand pounds of fuel. In addition a
carbon balance check was done at each condition comparing calculated fuel-air
ratio with fuel-air ratios from measured fuel and air flows. The conformity
of these two parameters was within 10% indicating a representative exhaust
sample. The mole fractions of THC, CO and NOX were reduced to Emission In-
dices (El) using the following relations:
31
-------
MHf (HC)
EIHr = "_; lbs/1000 Ib. fue
"L [CO](HCl
10(Mr + MH) ( 3— ) + (C0?) + T-
C H 4 *
MCQ (CO)
EIm = _~ !bs/1000 Ib. fue!
LU ^ (HC)
IO(M + M) ( 3- ) + (co) + -
MN02 (N°x}
EINn = NU^ xlbs/1000 Ib. fue
NUx (CO)
10(MC + MH) ( r) + (C02) +
t H 4 ^
where, MHC = Molecular weight of Methane
MCQ = Molecular weight of Carbon Monoxide
= Molecular weight of Nitrogen Dioxide
= Atomic weight of Carbon
= Atomic weight of Hydrogen
= Atomic Hydrogen-Carbon ratio of fuel
(HC), (CO), (NOX) = ppm concentrations of HC, CO & NOX
(C02) = * concentration of C02-
32
a
-------
TABLE 9. SUMMARY OF EMISSION DATA (PHASE I)
PT6A-45 GAS GENERATOR WITH MK VI FLAME TUBE
COND
7
3
9
10
11
12
13
14
15
16
17
13
19
20
21
22
23
24
NCRM.NG
22052.
31397.
21937.
31409.
22054.
31452.
22119.
31416.
22047.
31434.
37190.
37234.
37399.
37488.
37263.
37083.
37168.
37037.
25 ] 36868.
26
36787.
MECH.WF
154.2
308.5
154.2
307.6
154.1
3C6.5
155.2
307.3
154.1
307.4
613.4
616.8
636.6
636.4
634.3
626.7
625.6
612.1
611.5
605.3
Tl(F)
80
83
86
S4
83
80
81
81
81
77
75
74
68
69
72
74
74
75
83
84
T3(F)
276
492
200
494
278
488
276
489
275
483
625
624
616
681
620
623
621
624
632
635
EI(THC)
4.68
0.00
4.46
0.00
4.39
0.00
4.32
0.00
4.36
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
EI(CO)
25.82
4.97
26.92
4.92
25.83
4.85
25.93
4.72
26.08
5.54
1.80
1.56
2.44
2.47
2.44
2.72
2.34
2.77
2.33
2.25
EKNO2)
3.83
7.00
4.06
6.93
3.78
6.88
3.63
7.31
3.66
6.88
9.57
10.31
9.32
9.35
9.16
9.16
9.32
9.32
9.02
9.32
EFFY
.9897
.9988
.9897
.9989
.9900
.9989
.9901
.9989
.9900
.9987
.9996
.9996
.9994
.9994
.9994
.9994
.9995
.9994
.9995
.9995
CARBON BALANCE
(F/A)C*
.01558
.01461
.01508
.01476
.01558
.01476
.01533
.01476
.01543
.01486
.02133
.02153
.02104
.02079
.02104
.02104
.02104
.02105
.02119
.02104
(F/A)H+
.01436
.01477
.01453
.01475
.01439
.01458
.01432
.01468
.01432
.01462
.02014
.02019
.02060
.02055
.02071
.02056
.02052
.02108
.02042
.02030
EKC02)
3117.8
3163.4
3116.6
3163.5
3118.5
3163.6
3118.6
3163.8
3118.2
3162.5
3168.4
3168.8
3167.4
3167.3
3167.4
3166.9
3167.5
3166.9
3167.6
3167.7
%(CP2)
3.15
3.00
3.05
3.03
3.15
3.03
3.10
3.03
3.12
3.05
4.36
4.40
4.30
4.25
4.30
4.30
4.30
4.30
4.33
4.30
SIMULATED
POWER SETTING
Idle
Approach
Idle
Approach
Idle
Approach
Idle
Approach
Idle
Approach
Climbout
Clintoout
Take Off
take Off
Take Off
Take Off
Take Off
Climbout
Clinfcout
Clinfcout
REMARKS
T5 Limited
TS Limited
TS Limited
T5 Limited
TS Limited
EKC02)
3117.8
3163.4
3116.6
3163.5
3118.5
3163.6
3118.6
3163.8
3118.2
3162.5
3168.4
3168.8
3167.4
3167.3
3167.4
3166.9
3167.5
3166.9
3167.6
3167.7
%(002)
3.15
3.00
3.05
3.03
3.15
3.03
3.10
3.03
3.12
3.05
4.36
4.40
4.30
4.25
4.30
.4.30
4.30
4.30
4.33
4.30
OJ
u>
(F/A)
F/A CALCULATED FROM EMISSION DATA
(F/A)M » Wp/3600 (W3 + Wcool)
-------
Figure 22 shows a plot of average emission indices as a function of gas
generator speed. This data is fairly typical of the low emission combustor
under test.
Using the emission index data at idle, approach, climbout and takeoff
EPA emission parameters may be computed for the EPA defined (Federal Register
July 17, 1973) landing takeoff (LTO) cycle. This results in the cycle emis-
sion parameters shown in Table 10.
NORMNg (X10"J)rpm
Figure 22. Emissions profiles - phase I.
TABLE 10. LTO CYCLE EMISSIONS & EPA (1979) STANDARDS
AMBIENT EPA P***
T op S.H.** THC CO
Low Emissions 80
Combustor PT6A-45
EPA (1979) STD
(P2 Class)
59
.0083 2.61 16.23
.0063 4.9 26.8
N0x__
6.58*
12.9
* Not corrected for humidity
***lbs/1000 Ib-thrust hours/cycle
**Specific Humidity"
34
-------
The gas generator was allowed to stabilize for 10 minutes at each test
condition prior to collection of samples. Gaseous emissions were periodically
monitored while adequate number of particulate and smoke samples were collec-
ted.
Midway through the trial program samples of the low sulfur Jet A-l fuel
were collected from the fuel system and sampled for sulfur concentration.
These samples showed an average sulfur content of 0.0075 weight percent.
Table 11 is a summary of test conditions and gaseous emission data obtained
during Phase II collection runs with low sulfur Jet A-l fuel.
Prior to the second series of tests, the sulfur content of Jet A-l fuel
was increased to EPA specifications by adding ditertiary butyl disulfide to
the fuel tank. After addition of a known amount of the additive (1.12 gal-
lons/1000 Ibs.) the well mixed fuel in the tank was analyzed. These samples
gave sulfur concentrations of 0.255 weight percent on average.
The second series of tests was similar to the first, except for the
changed fuel specifications. Once again an adequate number of samples were
collected after the gas generator had stabilized at the test conditions for a
minimum period of ten minutes. In some cases takeoff modes could not be simu-
lated due to TS temperature maximum limits. This is typical of engine
operations during hot summer days (high inlet temperatures). In such cases
the gas generator was run at its TS limit (maximum temperature limit
1580°F at the first turbine stage exit).
The engine inlet air was analyzed. It showed the following constituents:
HC : 0 (i.e. none detectable)
CO : 5.5 ppm
NOX : 3.5 ppm
C02 : 0.04%
Table 12 is a summary of test conditions and gaseous emission data ob-
tained during Phase II collection runs with high sulfur Jet A-l fuel.
TEST PROCEDURE
Phases I and II of the contract called for the demonstration and docu-
mentation of the proper and consistent operation of the test vehicle with
special emphasis on consistent mass emissions and gaseous emissions.
35
-------
TABLE 11. PHASE II TESTS (LOW SULFUR FUEL)
COND
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
I
NORM
Ng
36648
36237
31285
31457
37653
37132
22067
21979
21927
31411
37634
37458
37089
31433
22293
ME CM
Wf
612.8
599.3
310.3
308.9
643.6
610.8
154.8
153.6
153.4
308.5
637.9
629.6
601.9
298.4
153.3
°F
Tl
74
78
84
80
66
73
79
85
87
87
64
72
71
69
69
op
T3
613
613
491
489
619
619
272
280
281
501
612
621
619
467
265
EMISSION INDICES
El
THC
0
0
0
0
0
0
3.93
3.78
4.36
0
0
NT
NT
NT
NT
El
CO
1.46
1.37
4.87
4.75
2.57
2.49
24.98
24.92
26.55
6.50
2.22
NT
NT
NT
NT
El
N02
9.80
9.80
6.83
6.99
9.82
9.75
3.39
3.46
3.57
7.31
8.73
NT
NT
NT
NT
Efficiency
.9997
.9997
.9989
.9989
.9994
.9994
.9906
.9908
.9899
.9985
.9995
-
-
-
-
FUEL-AIR RATIO
(F/A) c*
.02168
.02168
.01510
.01469
.02229
.02179
.01571
.01631
.01283
.01265
.02229
-
-
-
-
c ™ F/A calculated from emission data
NT = Not Taken
(F/A)M = WP/3600 (W3 + W000i)
-------
TABLE 12. PHASE II TESTS (HIGH SULFUR FUEL)
COND
42
43
44
45
46
47
48
49
50
51
52
53
54
55
••^W^MUHHHVWHIIIIIIH^**
NORM
Ng
37682
37041
31371
22193
31330
37548
37042
22188
32148
21808
37061
34963
33071
31301
^•^•••••••^••••••••••••i^^^H
MECH
Wf
637.4
599.6
300.2
154.9
303.0
646.0
609.7
154.9
314.8
164.0
608.9
475.7
377.4
297.4
^MHMMMMVMaHHH^MMBa
oF
Tl
60
64
68
73
71
64
68
74
94
93
77
92
90
76
•H^WI^VBIIIbl^MH
°F
T3
608
600
466
266
469
611
609
266
531
286
624
600
552
474
««M^HI^^BWWMH«MI
EMISSION INDICES
El
THC
0
0
0
6.09
0
0
0
4.85
NT
NT
NT
NT
NT
NT
P4M^***MIW««^H^M
El
CO
2.92
3.40
7.59
27.41
6.53
2.48
2.75
27.68
NT
NT
NT
NT
NT
NT
^•••^^••^•MMai^BVAHA^V
El
NO2
8.66
8.46
5.16
3.13
6.24
9.46
8.95
3.36
NT
NT
NT
NT
NT
NT
^^••••••••^^•••IIIII^MM
Efficiency
.9993
.9992
.9982
.9881
.9985
.9994
.9994
.9892
-
-
-
-
-
M*B^M^V^VMIII*kMIIIHHIIHH»H
CARBON BALANCE
(F/A)C*
.02230
.02145
.01290
.01282
.01289
.02229
.02154
.01560
-
-
-
-
-
•••••••••••••(••••^•^^••^•••^•MI
(F/A)M*
.02042
.01968
.01430
.01420
.01449
.02068
.01991
.01406
.01454
.01638
.02026
.01786
.01626
.01459
M^^MHUMHHHIVHMfeWIMMMIIIIBBIIBWWW
SIMULATED
POWER
SETTING
Take Off
Clirabout
Approach
Idle
Approach
Take Off
Climbout
Idle
Approach
Idle
Take Off
Approach
Approach
Idle
••••••••IIIIHMi^BiBHIW^^M^MIHB^^
REMARKS
T5 Limited at T.O.
Diaphragm failure
in sample pump
T5 Limited
TS Limited
•-•••• ----- ----- II
* (F/A)C = F/A calculated from emissions data
NT = Not Taken
(F/A)M - WF/3600 (W3 + Wcool)
-------
Phase I
Using a PT6A-45 gas generator and a typical sampling rake (see Figure
26) EPA smoke and gaseous emissions data were taken. In addition to this
data, a tailpipe mapping was performed with respect to engine emissions. The
data shows that a) the engine exhaust is reasonably uniform in terms of gas-
eous emissions, b) it is a low emissions gas turbine and c) it was operated
in a predictable and repeatable fashion.
Smoke data was taken using an EPA type smoke meter (Figure 23) that was
designed and built in conformance with Federal Register Vol. 38, No. 136,
July 17, 1973 and Aerospace Recommended Practice 1179 (5/4/70). The samples,
which consisted of a series of stained filters, were analyzed using a Photo-
volt model No. 670 reflectance meter. Replicate samples of smoke data were
taken so that smoke data from each of four power points was taken 15 times.
The filter type used.was Whatman filter paper #4.
Figure 23. SAE/EPA smoke meter.
38
-------
Mass emission data were taken using the same smoke meter described
above. In this test a hydrophobic Nuclepore, 40 mm diameter filter was used.
Each filter was preconditioned and preweighed several times to insure equili-
brium in humidity-temperature controlled atmosphere. The device used to make
these filter weight measurements was a Perkin-Elmer Model AD-2 electrobalance
located in a room where temperature was 69°F and relative humidity was 5056.
After the particulate material was collected on these filters by allowing a
0.9 to 2.1 m3 of the exhaust to pass through, they were returned to the
same room where they were allowed to equilibrate for several days then
weighed by a similar process. Replicate filter samples at each of four power
points on low sulfur fuel only were taken resulting in several loaded filters
for each power setting. Filter blanks were taken to monitor the entire
process.
Tailpipe mapping was performed on the engine using a traversing rake.
The rake was allowed to traverse a diameter of the tailpipe taking emissions
data at 9 points which represented centroids of equal areas. This process was
repeated several times with the diameter rotated 30°, 60° and 90° from
the first. The emission measurements, taken in this sequence, were NOX, CO,
C02 and total hydrocarbons.
Phase II
During this phase of testing particulate material for numerous chemical
analyses were taken. In addition a sulfur collection train was used to col-
lect gaseous and aerosol sulfur products in the engine exhaust stream. Sam-
ples were obtained from the engine inlet to ascertain the quality of the in-
let air used during these tests and eliminate a possible source of error in
the final results.
Using a sampling system comprised of a linear rake and plenum chamber,
Roots model 3514J vacuum blower system, fixed orifice flow metering devices
and filter holding console (holding five 293 mm diameter filters), particu-
late material was collected for various chemical analyses. Mitex (10 micron)
and glass fiber filters were exposed to engine exhaust to collect particu-
late material in sufficient quantity for the chemical tests required. While
the sampling was being done the gas temperature in each of the filter hol-
ders were monitored. Temperature and pressure data necessary to make flow
calculations with the fixed orifice gas metering tubes was also taken.
Previous tests showed that serious degradation occurred when tempera-
tures at the filter were above 160°F. During the testing, unacceptably
high filter temperatures occurred but were resolved using two techniques.
One method was to throttle the flow so that the gases had time to cool be-
fore entering the filter holder. The other method was to use longer water
cooled heat exchangers. Using a combination of both techniques the filter
temperatures varied between 100 and 160°F, assuring minimum sample degra-
dation.
Because it was impossible to control temperature consistently the ex-
haust gas temperature at times was over-cooled, dropping the temperature be-
low the dew-point. This resulted in moisture condensing on the filter mater-
ial causing unusually high pressure drops. The Roots pump was designed to
39
-------
operate efficiently at up to 4 inches Hg vacuum. Serious damage would occur
if a vacuum of 8 inches Hg was reached. Bypass air was allowed to enter the
pump to prevent this situation. Particulate loading coupled with the conden-
sed moisture proved to be a serious obstacle in collecting what was thought
to be a reasonable weight of particulate sample. This problem was somewhat
mitigated by extending the sampling times to acquire more particulate mater-
ial.
Sulfur oxide samples were collected by drawing the exhaust gases through
a series of bubblers to extract the sulfur oxides from the exhaust stream.
These bubblers form a gas sampling train similar to the one described in
method 8 of the Federal Register, June 8, 1976. In this sampling train frit-
ted bubblers are used instead of the impingers described and the filter (and
filter holder) was omitted. The test is set up so that sulfur oxides in the
form of 503 can be differentiated from SC^. The data was taken using
four power settings and two fuels (low and high sulfur).
An attempt was made to evaluate a filter system using a packed bed of
gas chromatographic column material. Four columns of Chromsorb 102 (Figure
14) were used to collect organic material in engine exhaust. A sample line
from the sample plenum was used to conduct exhaust gases to the packed col-
umn. Sample flow through the bed was extremely slow so that relatively small
total flows were realized. The relatively slow flow rate was due primarily
to the fine mesh of the Chromsorb 102 used as the adsorbent. Samples were
obtained of two engine power settings using both high and low sulfur fuels.
All samples (filter, packed column and liquid) were packed in an ice
chest filled with dry ice immediately after they were taken. The samples
were kept continuously in this condition throughout the test in Canada, while
they were shipped back to the analytical laboratory in East Hartford, and un-
til they were eventually processed for analysis.
SAMPLE IDENTIFICATION
Due to the large number of filters that were processed in the course of
the fulfillment of the EPA contract, it was necessary to initiate an identi-
fication system. From Table 4, it can be seen that a minimum of 66 Mitex, 12
glass fiber and 20 Nuclepore filters were used. It was necessary to expose
filters to engine exhaust while the engine was at 4 power settings. In addi-
tion, fuels with two sulfur concentrations were used. In some cases repli-
cate samples were taken. These samples were identified using the following
letter/number scheme:
I. Specific Analysis (or Disposition)
1. HPLC/UV = LC/UV
2. Phenols and Nitrosamines = p/N
3. Infrared = IR
4. Nuclear Magnetic Resonance = NMR
5. Gas Chromat./Mass Spec. = GC/MS
6. Boiling Point Analysis = BP
40
-------
7. Total Organic (via GC)
8. Special for EPA (X-Ray)
9. Elemental Analysis
10. Mass Emissions
11. Sulfur Analysis
12. Proton Activation Analysis
T/ORG
EPA
EL
ME
S
PAA
II. Power Points:
1.
2.
3.
4.
Idle
Approach
Climb
Takeoff
III. Fue
A =
B =
Low sulfur
Hi sulfur
IV. Replicate Samples
First Sample = fl
Second Sample = #2
Third Sample = #3
Using this system the designation of filters for the HPLC/UV analysis
using both fuels, all power points and in some cases taking two replicate
samples were as follows:
A. HPLC/UV
1) LC/UV-1-A #1
2) LC/UV-1-A #2
3) LC/UV-2-A fl
4) LC/UV-2-A #2
5) LC/UV-3-A #1
6) LC/UV-3-A f2
7) LC/UV-4-A #1
8) LC/UV-4-A #2
9) LC/UV-1-B
10) LC/UV-2-B
11) LC/UV-3-B
12) LC/UV-4-B
D. Nuclear Magnetic Resonance
1) NMR - 1 -A
2) NMR - 3 -A
3) NMR - 1 -B
4) NMR - 3 -B
Simi lar ly:
B.
1)
2)
C.
D
2)
3)
4)
H.
D
2)
3)
4)
5)
6)
7)
8)
PhenoIs/Nitrosamines
P/N - 1 -A
P/N - 3 -A
Infrared
IR - 1 -A
IR - 3 -A
IR - 1 -B
IR - 3 -B
Special EPA Filters
EPA - 1 -A
EPA - 2 -A
EPA - 3 -A
EPA - 4 -A
EPA - 1 -B
EPA - 2 -
EPA - 3 -B
EPA - 4 -B
41
-------
E.
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
13)
14)
15)
16)
F.
1)
2)
3)
4)
5)
6)
7)
8)
6.
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
G/C - Mass Spec.
GC/MS - 1 -A #1
GC/MS - 1 -A #2
GC/MS - 2 -A #1
GC/MS - 2 -A #2
GC/MS - 3 -A #1
GC/MS - 3 -A #2
GC/MS - 4 -A #1
GC/MS - 4 -A #2
GC/MS - 1 -B #1
GC/MS - 1 -B #2
GC/MS - 2 -B #1
GC/MS - 2 -B #2
GC/MS - 3 -B #1
GC/MS - 3 -B #2
GC/MS - 4 -B f 1
GC/MS - 4 -B #2
Boi ling Point Anal.
BP - 1 -A
BP - 2 -A
BP - 3 -A
BP - 4 -A
BP - 1 -B
BP - 2 -B
BP - 3 -B
BP - 4 -B
Total Organic (via GC)
T-ORG - 1 -A #1
T-ORG - 1 -A f 2
T-ORG - 1 -A #1
T-ORG - 2 -A #2
T-ORG - 3 -A #1
T-ORG - 3 -A n
T-ORG - 4 -A #1
T-ORG - 4 -A #2
T-ORG - 1 -B
T-ORG - 2 -B
T-ORG - 3 -B
T-ORG - 4 -B
I.
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
J.
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
K.
1)
2)
3)
4)
5)
6)
7)
8)
9)
10)
11)
12)
Elemental Analysis
EL - 1 -A #1B
EL - 1 -A #2
EL - 2 -A #1
EL - 2 -A #2
EL - 3 -A #1
EL - 3 -A #2
EL - 4 -A f 1
EL - 4 -A f 2
EL - 1 -B
EL - 2 -B
EL - 3 -B
EL - 4 -B
Mass Emissions
ME - 1 -A #1
ME - 1 -A #2
ME - 1 -A #3
ME - 2 -A #1
ME - 2 -A n
ME - 2 -A #3
ME - 3 -A #1
ME - 3 -A #2
ME - 3 -A #3
ME - 4 -A #1
ME - 4 -A #2
ME - 4 -A #3
Proton Activation Analysis
PAA - 1 -A #1
PAA - 1 -A #2
PAA - 2 -A #1
PAA - 2 -A #2
PAA - 3 -A #1
PAA - 3 -A #2
PAA - 4 -A #1
PAA - 4 -A #2
PAA - 1 -B
PAA - 2 -B
PAA - 3 -B
PAA - 4 -B
MASS EMISSIONS MEASUREMENT TECHNOLOGY
Mass emissions measurements have historically been a subject of question
and controversy due to a number of variables involved with filter prepara-
tion, sampling technique, and post test weight analysis. Over the past sev-
eral years, P&WA has improved its methods in mass emissions measurement to a
point where the resultant data can be considered both repeatable and suffi-
ciently accurate to be useful as a tool in monitoring emissions in gas tur-
bine engines.
42
-------
FiIters
A number of filter materials have been examined as possible candidates
for participate collection. Among these have been Mitex (teflon), standard
Millipore (mixed esters of cellulose), PVC (polyvinyl chloride), Metricel
(mixed esters of cellulose) and Nuclepore (polycarbonate plastic). Some of
these filters have had a problem with water absorption and desorption. There
have also been problems with temperature, mechanical integrity, trapping ef-
ficiency and actual mass of the filter (since a heavier filter will absorb
more water than a lighter one of the same material).
Nuclepore filters were ordered from Nuclepore Corporation in the non-
standard 40 mm diameter used. The filters were therefore clean cut without
frayed edges and were handled only under clean room conditions. Each filter
was numbered and set in a 47 mm petri dish'half to equilibrate in a tempera-
ture and humidity controlled room (20°C, 50% RH) for a period of 72 hours.
After equilibration, each filter were passed several times over a static dis-
charge source and then placed on the weighing pan of a Perkin Elmer AD-2 el-
ectrobalance.
A minimum of three pretest and three post-test weighings were made for
each filter. Additional weighings were taken as necessary to insure that
filters were equilibrated and data were repeatable. The post-test filters
were set in petri dish halves as in the pretest preparation and were allowed
to equilibrate for a minimum of 72 hours.
Balance
Experience has shown that mechanical balances designed for microgram
weighing are not suitable for mass emissions filter analysis. The weighings
are not stable and repeatable when attempted on a marble table in an indus-
trial setting. Vibration and air movement caused by temperature and humidity
control equipment make it unlikely that they would be suitable in any setting.
The Perkin-Elmer AD-2 electrobalance was chosen for mass emissions test-
ing because it has resolution to 0.1 microgram and is electronically dampened
so that it maintains a high level of stability. Calibration was checked be-
fore, during and after each weighing period using a Class S weight set.
Mass Emissions Testing
Mass emissions were collected using a P&WA built SAE/EPA smoke meter.
P&WA Canada's multi-point emissions rake was used to deliver the sample to a
1/4" stainless steel line heated to 150°C. Immediately after testing each
filter was sealed in a petri dish and returned to the laboratory for equili-
bration. The mass of the accumulated particulate matter ranged from approxi-
mately 250 ug to 1000 ug. Sampling time ranged from 3 to 6 minutes per fil-
ter and the collected volume ranged approximately from 0.9 to 2.1 nr3.
43
-------
Mass emissions measurements from Phase I are shown in Table 13. Five
gas generator tests, each at four simulated power settings were performed.
Samples were taken in triplicate where possible. The mechanical fuel-air ra-
tio calculated from actual fuel and air flows are in parentheses in Table 13.
Mass emissions measurements from Phase II are shown in Table 14. Two
gas generator tests, each at four simulated power settings were performed.
Again, samples were taken in triplicate where possible and the mechanical
fuel-air ratio is given in parentheses.
TABLE 13. MASS EMISSIONS MEASUREMENT - PHASE I (mg/m*)
Test Number 1
3.320
Idle Z.790
4.732
Approach 4.520
4.372
7.063
Climb 7.416
7.310
7.805
Takeoff 8.123
7.840
TABLE
(F/A)H 2
3.178
(.01436) 2.896
3.567
4.874
(.01477) 4.566
4.662
6.922
(.02014) 6.886
6.745
8.158
(.02060) 7.522
7.805
(F/A)H 3 (F/A)M
2.825
(.01453) 3.108 (.01439)
3.037
4.909
(.01475) 4.308 (.01458)
4.556
7.310
(.02019) 7.381 (.02018)
7.204
7.310
(.02055) 7.875 (.02071)
8.158
4
3.108
2.790
3.289
4.379
4.238
4.308
7.275
7.169
6.922
7.981
6.922
7416
14. MASS EMISSIONS MEASUREMENT - PHASE II
(F/A)H 5
2.684
(.01432) 3.320
4.874
(.01468) 4.414
4.485
6.675
(.02042) 6.321
6.569
7.310
(.02056) 7.734
8.087
(mg/m3)
(F/A)H
(.01432)
(.01462)
(.02030)
(.02052
Test Number
Idle
Approach
Climb
Take-Off
Condition No
& Condition
1
2.418
2.110
3.217
3.143
3.098
7.079
7.806
3.098
8.478
7.946
. 53* 6.487
No. 54* 6.609
6.635
(F/A)M
(.01445)
(.01485)
(.02046)
(.02066)
(.01790)
2
2.036
1.903
1.940
3.945
3.744
7.787
3.744
9.327
9.523
5.629
5.644
5.785
(F/A)M
(.01447)
(.01482)
(.01482)
(.02055)
(.01630)
44
-------
Figure 24 shows mass emissions from Phase I and Phase II as a function
of mechanical F/A. Two additional power settings (conditions nos 53 and 54)
were run between approach and climb to help define the cente? of ihis curve
Int.tl'p nf"V ti9ht ^ ?f P°intS ab°ut the curve and * con iderld repre-
sentat ve of the mass emisS1ons of the PT6-A. The engine is seen to be oper-
ating in a repeatable manner. p
I
ol—
0.014
0.015
0.019
Figure 24. Mass emissions vs. fuel-air ratio.
SMOKE MEASUREMENT METHODOLOGY
An SAE/EPA smoke measurement system designed and buiIt by P&WA in con-
formance with CFR 40, Number 87 Part II which appears in the Federal Regis-
ter, Volume 38, Number 136, July 17, 1973 was used for the measurement of
smoke (See Figures 23 and 25). Samples were extracted from the exhaust of
the PT6-A using an emission sampling rake shown in Figure 26 and a 1/4-inch
stainless steel line maintained at 150°C.
Five gas generator engine tests, each at four simulated power settings
(idle, approach, climb and takeoff) were conducted to define smoke emissions
levels throughout the engines operating range. The simulated power setting
for take-off was TS (turbine exhaust temperature) limited and in each of
the five engine tests the highest power setting obtainable within the T§
limit was considered to be takeoff.
The smoke measuring system is a semiautomatic device which incorporates
a number of features to permit the recording of smoke data with precision and
ease of operation. The instrument features a timer-controlled, solenoid ac-
tivated main sampling valve (Valve A, Figure 25) having closed "sample" and
"bypass" positions. This system permits close control of the sample size over
relatively short sampling periods. In addition, the timing system operates a
bypass system around a positive displacement volume measurement meter to in-
sure that the meter is in the circuit only when a sample is being collected,
or during the leak-check mode. Automatic temperature control of the filter
housing is included. The silicon-rubber filter holders have support screens
for each of the filter holders.
45
-------
. .
•- •
..•-,-, I ' , I ' • , -. .
I' ' ".-•' "\il>Fl.i
(i :n HI ;jo 'HtcoMMi -.
t IL U H HOI [>E M SrHEMATlf lilAi ,HAM
Figure 25. Schematic diagram of smoke meter.
Figure 26. PT6 emission sampling rake.
46
-------
The filter holder assembly was constructed with a one-inch diameter spot
size, a diffusion angle of 7.25 degrees, and a converging angle of 27.5 deg-
rees.
A Photovolt Model 670 with a Y type search unit conforming to American
National Standard ASA Ph 2.17-1977 "Optical Reflection Measurements" was used
to determine the reflectance of the clean and stained filters. A set of
Hunter Laboratory reflectance plaques, traceable to the National Bureau of
Standards, was used to calibrate the reflectance meter. A computer program
was used to calculate W/A (mass/area) and smoke number for each filter.
Conclusions
Table 15 shows the results of five smoke tests at each of the four power
settings. Samples were taken in triplicate and the mechanical fuel/air ratio
is given as (F/A)^. Smoke numbers for idle averaged 18.2, for approach
25.9, for climb 37.2, and for take-off 43.0.
TABLE IS. SAE/EPA SMOKE NUMBERS
(F/A)M
(F/A)M
(F/A)M
(F/A)M
(F/A)M
Test Number
Idle
Approach
Climb
Takeoff
1
17.7
17.2
17.9
26.7
24.3
25.1
43.2
40.4
36.5
50.8
47.3
39.6
(.01436)
(.01477)
(.02014)
(.02060)
2
18.1
19.4
17.5
25.4
24.8
24.7
39.4
38.7
36.6
48.2
47.1
42.0
(.01453)
(.01475)
(.02019)
(.02055)
3
17.6
18.9
18.6
25.5
26.7
25.8
38.0
37.7
36.2
46.9
42.9
42.2
(.01439)
(.01458)
(.02018)
(.02071)
4
19.1
17.6
17.5
26.3
27.6
24.7
37.9
38.3
35.5
40.5
39.6
36.2
(.01432)
(.01468)
(.02042)
(.02056)
5
18.7
19.0
18.2
26.5
26.6
27.1
35.3
33.1
31.7
42.5
38.2
40.5
(.01432
(.01462)
(.02030)
(.02052)
Figure 27 plots average smoke number as a function of average (F/A)M
and Figure 28 shows the relationship between average particulate mass emis-
sions and average smoke number.
The smoke numbers are seen to be essentially repeatable with power set-
ting. Some difficulty was encountered in simulating takeoffs due to the TS
temperature limit encountered when ambient temperatures were high.
The data of Figure 27 is considered representative of this engine's
smoke emissions. The linearity of Figure 27 show a consistent and linear
relationship between smoke number and mass emissions and would be useful in
estimating mass emissions from smoke number for this engine.
47
-------
cc
UJ
DO
D
Z
UJ
y.
o
501—
45
40
35
30
25
UJ 20
£
nE 15
10
0
0.014
I
I
I
I
0.015 0.016 0.017 0.018 0.019
MECHANICAL FUEL - AIR RATIO
0.020
0.021
Figure 27. Average smoke numbers as function of average (F/A)M.
8.0 i—
o>
CO
g
CO
LU
Figure 28.
15
25 30 35
SAE/EPA SMOKE NUMBER
40
45
Reitionship between average particulate mass emissions and average
smoke number.
48
-------
ANALYTICAL PROCEDURES
Sample Treatment
In preparation for this work, a detailed computer literature search of
both NTIS and the American Chemical Society files was conducted for informa-
tion regarding the analysis of polynuclear aromatic compounds and nitrosa-
mi nes.
The teflon filters obtained for chemical analysis were removed from the
filter holder using forceps and gloves, folded, and placed in wide mouth 250
ml capacity polyethylene screw cap bottles. These bottles were stored imme-
diately in a dark container kept cold with dry ice. These conditions were
maintained until actual extraction of the filters was carried out.
All samples, except those for nitrosamine and phenol analyses, were ex-
tracted with appropriate solvents using a Soxhlet extractor. The apparatus
consisted of a 250 ml round bottom flask containing 150 ml of solvent and the
Soxhlet extractor containing the folded filter. Each extraction process was
continued for a period of at least 12 hours. General references (10, 24, 28,
29, 31) suggest that this should be a reasonable time to achieve essentially
complete extraction. The solvents used were as follows:
NMR analysis - deuterated chloroform
Total Organics, GC/MS, HPLC, UV, BP - hexane
Infrared - Carbon disulfide
After extraction was completed, the sample was concentrated by careful
evaporation of the solvent to a final volume of 1 ml.
The choice of hexane as the usual solvent was based on high performance
liquid chromatographic analyses of various solvents after concentration of
impurities in the solvents. Many of these solvents, even after redistilla-
tion, still showed a concentration of impurities which would interfere in the
analyses. Solvents considered were benzene, chloroform, methylene chloride,
cyclohexane and hexane. The hexane used in the work was triply distilled in
glass hexane obtained from Burdick and Jackson Laboratories.
The samples to be analyzed for nitrosamines and phenols were treated
with phosphoric acid and extracted manually with methylene chloride and di-
isopropyl ether respectively in accordance with the procedure described in
EPA-650/2-75/056.
The Chromosorb* 102 (a styrene - divinyl benzene polymeric material)
packed filter bed material used to explore the general magnitude of the total
organic emissions (gaseous and particulate) was put into a clean teflon fil-
ter and extracted by the Soxhlet method. The extracting solvent was 150 ml
of hexane.
*Johns-ManviIle Products Corp. Celite Division, Manville, N.J.
49
-------
Benzo (a) Pyrene and Sulfur Standards
Prior to a trial analyses of engine samples, it was considered desirable
to analyze known samples of a polynuclear aromatic hydrocarbon to establish
the sensitivity of the instrument for the substance and the linearity of the
response. Benzo (a) pyrene was chosen as a representative compound for this
work because of its previous use by others as a reference material and
because of its possible presence in exhaust gases. Samples were exchanged
with the EPA and good agreement was obtained after correction of results for
purity of BAP. BAP, as commercially available, to us had up to 30% impuri-
ties. A similar program of comparison with the EPA was carried out for sul-
fur analyses (as sulfate) with good agreement down to the level of sensitivi-
ty of the method (ASTM D-3226-73T).
Organic Analyses
The samples collected were subjected to three basic types of analyses to
characterize and semi-quantify the organic content. The total organic mea-
surements established the magnitude of organic matter in the adsorbent and
included aliphatic compounds, aromatic compounds and polycyclic organic mat-
ter. These species may or may not be oxygenated or other derivatives. The
high performance liquid chromatographic analyses were used to determine the
relative amounts of aromatic compounds (one or two fused rings) and polynu-
clear aromatic compounds. These also could include hetero atoms. The gas
chromatograph-mass spectrometer analyses specifically determined individual
PAH and PNA compounds.
Finally, to establish the relative amounts of organic matter adsorbed on
the particulate matter and the total amount emitted by the gas generator, a
packed bed filter study was carried out.
Total Organics
Samples were collected at four engine power settings using both a high
sulfur and a low sulfur fuel. Duplicate samples were taken using the low
sulfur fuel. These 12 samples were collected and extracted as described ear-
lier.
Analysis was carried out using a Hewlett-Packard Model 7620A gas chrom-
atograph with a flame ionization detector. (A photoionization detector pro-
duced by HNU, Inc., Newton, Massachusetts was reported to give much greater
sensitivity for aromatic hydrocarbons but was found to develop leaks at ele-
vated temperatures. Therefore we found it to be unusable for our purposes.)
50
-------
A valve was added to permit special backflushing. The sample was introduced
into the chromatograph and after a period of time, the valve was switched to
reverse the flow. The lower molecular weight components including the sol-
vent hexane passed through the column into the detector while the heavier
molecules remained on the column. Upon reversing the flow the heavier compo-
nents were flushed from the column into the detector giving an indication of
the total amount of heavier molecules. The column used was a 6' x 1/8" 00
stainless steel column packed with 10% UC-W98 (silicone gum) on 80-100 mesh
Diatoport S (acid washed and silanized diatomaceous earth). The column was
maintained at 190°C, the detector at 250°C and the injection port at
200°C. The carrier gas, nitrogen was set at 60 psig to give a flow of 41
ml/min. For the flame, hydrogen pressure was set at 16 psig and the air was
set at 48 psig. The valve was switched after about four minutes. The sample
injected was 1 ul of a total hexane extraction that was concentrated to 1 ml.
The instrument was calibrated using a) a composite sample of 16 polynu-
clears as shown in Table 19 plus coronene (6 fused rings) and triphenylene (4
fused rings); b) benzo (a) pyrene and c) several known compounds containing
two fused rings. In the case of the composite sample, all components except
fluorene were on the column when the flow was reversed. The two fused ring
components and fluorene were eluted before the flow was reversed.
Table 16 gives the calibration data for these standards and Table 17
gives the results for the analyses of the 12 samples in terms of retention
times and responses. The peaks shown e luted after the main hexane solvent
peak.
TABLE 16. TOTAL ORSANICS CALIBRATION
Standard
Composite Sample, *150 ng
F luorene, 10 ng
Naphthalene, 25 ng
Biphenyl, 30 ng
Acenaphthene, 20 ng
Methoxynaphthalene, 25 ng
Benzophenone, 25 ng
Benzo (a) pyrene, 108 ng
Retention
Time, Win
3.3
1.55
Response/ng
Peak Height - Peak Area
**
124.2
0.8
222.9
2.35
2.00
3.55
**
17.08
-
324
-
138.4
94.7
60.8
16.58
*10ng of each component, 150ng total plus lOng fluorene
**Backflushed out. If the flow was reversed between 3.8-4 minutes, the
composite was eluted at a retention time of 6.5-7 minutes.
51
-------
TABLE 17. TOTAL OR6ANICS ANALYSIS
Samp 1 e
1A No. 1
1A No. 2
2A No. 1
2A No. 2
3A No. 1
Retention
Time, Min.
1.25
1.35
1.7
2.05
2.45
3.00
3.6
6.55*
2.55
3.7
6.8*
0.9
1.3
1.6
2.1
3.3
4.15
6.85*
1.38
1.75
2.15
2.5
3.1
3.7
7.2*
2.0
2.25
3.2
3.95
7.15*
Response
Peak Height
420
120
372
116
1212
92
524
5952
1064
400
7008
14720
400
208
448
80
304
11152
176
294
164
1868
256
1604
7080
92
T
25
178
19440
Peak Area
«•
-
-
-
-
-
-
10679
_
-
12578
—
_
_
_
_
MI
13672
_
_
_
—
—
_
14692
_
M
—
25468
*After flow reversed
(Continued)
52
-------
TABLE 17 (Continued)
Samp I e
3A No. 2
4A No. 1
4A No. 2
4A No. 3
IB
Retention
Time, Min.
2.1
4.2
7.05*
2.05
4.05
6.65*
1.0
1.1
1.3
1.5
1.7
2.1
2.45
2.7
3.4
4.1
7.1*
1.0
1.4
1.55
1.73
2.15
2.5
3.65
6.5*
0.8
1.15
1.35
1.75
2.00
2.35
2.7
3.6
6.4*
Response
Peak Height
232
242
4390
220
180
15040
T
T
1344
320
1792
1093
1056
200
120
280
9312
584
1756
' ' 360
1120
240
3484
840
5748
784
424
T
504
127
174
72
88
7438
Peak Area
_
8543
—
-
22544
-
_
-
-
-
-
-
-
-
-
13010
_
-
-
-
-
-
-
11201
-
-
-
-
-
-
-
-
12302
*After flow reversed
(Continued)
53
-------
TABLE 17 (Continued)
Sample
28
3B
48
Retention
Time, Min.
0.8
1.2
1.65
2.0
2.4
3.25
3.95
6.35*
1.2
1.6
1.8
2.45
3.3
4.15
7.4*
0.8
1.2
1.65
1.75
2.00
2.4
2.75
3.3
6.7*
Response
Peak Height
20704
1356
120
70
103
38
23
11984
1248
108
T
142
T
96
8824
36768
2136
222
T
70
732
66
76
38080
Peak Area
_
-
-
-
-
-
-
16473
_
-
-
-
-
-
12345
_
_
_
_
_
_
_
_
31015
*After flow reversed
In order to establish the total organic content which includes all or-
ganic species containing C-C and C-H bonds, the responses must be converted
to nanograms of material. For this purpose, the sensitivities of the knowns
were used where available or estimated from sensitivities of substances with
similar retention times. For the large peak eluting after the flow was rev-
ersed, the sensitivity of benzo (a) pyrene (which is very close to that of
the composite sample) was used. These results are given in Table 18. A
range from 14.4 to 70.5 ug/m3 are shown in the twelve samples. Most of the
organic matter (92.2 - 99.6%) is in the composite peak after the flow is
reversed. No trends are apparent as a function of power setting or fuel
used. As will be shown later (See section on Packed Bed Filter Studies), a
few exploratory samples collected on Chromosorb 102 showed that the organic
matter on the particulates represent a very small percentage of the total
(0.03 - 0.29%).
54
-------
TABLE 18. TOTAL ORGANICS CONTENT
Sample
1A #1
1A #2
2A #1
2A #2
3A #1
3A #2
4A #1
4A #2
"4A #3
IB
2B
3B
4B
Flow
M3
24.5
21.0
22.7
13.2
34.4
26.1
28.1
23.5
21.9
29.3
47.5
52.0
42.7
Composite
"9* ug/m3
644.0
758.5
824.4
885.9
1535.8
515.2
1359.4
784.5
675.4
741.8
993.3
745.0
1870.3
26.3**
36.1
36.3
67.1
44.6
19.7
48.4
33.4
30.8
25.3
20.9
14.3
43.8
(96.3)
(98.4)
(93.3)
(95.2)
(99.8)
(98.5)
(99.6)
(94.9)
(92.2)
(98.3)
(93.3)
(99.3)
(95.7)
Light
ug
23.4
14.3
58.4
45.3
4.1
6.4
5.3
41.7
57.0
13.6
70.1
6.9
128.2
Ends
ug/m3
1.0
0.7
2.6
3.4
0.1
0.2
0.2
1.8
2.6
0.5
1.5
0.1
3.0
(3.7)
(1.6)
(6.7)
(4.8)
(0.2)
(1.5)
(0.4)
(5.1)
(7.8)
(1.9)
(6.7)
(0.7)
(4.3)
Total
ug/m3
27.3
36.8
38.9
70.5
44.7
20.0
48.6
35.2
33.4
25.8
22.4
14.4
46.8
*in terms of BAP sensitivity
**Numbers in parentheses are % of total of all peaks Example 43.8 (95.7%)
High Performance Liquid Chromatographic Analysis
Samples were collected at four engine power settings using both a high
sulfur and a low sulfur fuel. Duplicate samples were taken using the low
sulfur fuel. These 12 samples were collected and extracted as described
ear Ii er.
Analysis was carried out using a Du'Pont Model 830 high performance liq-
uid chromatograph with a DuPont Model 835 muItiwavelength photometer having
ultraviolet absorption and fluorescence detectors. The column was a 4.6mm ID
x 25cm stainless steel column packed with Zorbax (microparticular silica
support) octadecylsilane and was maintained at 50°C. The primary mobile
phase was 75% methanol, 25% water and the secondary mobile phase was 100%
methanol. A nonlinear gradient mode was used which averaged about 4% per
minute. The mobile phase flow was 2.5 ml/min and the pressure was 2500 psig.
The sample injected (by means of a valve) was 10ul.(32,22,34)
Calibration of the instrument was carried out using various 3, 4, 5 and
6 fused ring compounds. Table 19 gives the retention times and sensitivities
for these substances.
55
-------
TABLE 19. RETENTION TIMES AND AND SENSITIVITIES FOR HPLC KNOWNS
Retention Time
Compound min.
Fluorene*
Phenanthrene*
Anthracene*
Benzacridine*
F luoranthene*
Pyrene*
Chrysene**
Benzo (a) anthracene*
Benzo (e) pyrene*
Perylene**
Benzo (a) pyrene*
Dibenz (ah) anthracene**
Benzo (ghi) perylene**
Phenylene pyrene**
6.5
7.69
7.72
8.37
8.90
9.38
12.55
14.9
17.35
17.9
20.3
21.4
22.9
23.1
Response in Peak Height per ng
Ultraviolet Fluorescence
6 x 10-2
9.4 x 10-2
4.93 x 10-1
1.77
1.46
1.31 x 10-1
2.16 x 10-2
1.2 x 10'1
5.36 x 10-2
1.18 x 10-1
1.11
1.12 x 10-1
8.96 x 10-1
1.15
.
3.58 x 10-1
11.16
28.26
88.45
1.41
1.02 x 10-1
2.56
2.69 x 10-1
127.0
28.44
2.25
28.26
144.2
Number Fused
Rings
3
3
3
4
4
4
4
4
5
5
5
5
6
6
Molecular
Weight
166
178
178
219
202
202
228
228
252
252
252
278
276
276
* Samples used contained 1000 ng/ul
** Samples used contained 250 ng/ul
It is apparent from Table 19 that retention time, in general, increases
with an increase in the number of fused rings and also with an increase in
molecular weight e.g., compare chrysene with benzo (a) anthracene. Separa-
tion into individual compounds is not possible with the conditions used be-
cause of the closeness of the retention times. In addition many other com-
pounds may be present. In this analysis effort was directed to determine the
relative amounts of compounds containing a like number of fused rings or sim-
ilar molecular weight. Tentative identifications, based on retention time
data, are given for a few components in some samples. Absolute identifica-
tions of PAH compounds are given in the section on 6C/MS. Table 19 also
shows the differences in sensitivity from one compound to another and as a
function of detector.
Tables 20 through 34 give the results for the samples analyzed. Some
variation in retention times from the values for the knowns occurs due to
instrument flow changes. Although the pressure on the instrument's mobile
phase was easily set and controlled, the resulting flow showed some variation
from day to day. Retention time calibrations were periodically repeated.
56
-------
TABLE 20. HPLC ANALYSIS
Retention
Sample Time, min
1A No. 1 0.87
0.98
1.08
1.18
1.25
1.36
1.47
1.52
1.57
1.67
1.75
1.91
2.00
2.11
2.18
2.45
2.63
2.75
2.83
3.00
3.05
3.10
3.32
3.65
3.76
3.85
4.05
4.22
4.85
5.07
5.40
5.52
5.88
6 00
w • w
6.4
6.55
6.69
7.6
7.88
8.38
8.65
8.84
9.27
1? 3
XC. . J
12.65
17.45
18.1
22.37
22.8
' ; 23.08
Total Flow
• - " ""
^
T = Trace
Response in
Identity Ultraviolet
Naphthalene
Anthracene
Benzacridine
Fluoranthene
Pyrene
Perylene
Benz (ghi)
perylene
o-phenylene
pyrene
24.6 m3
i
1862
144
416
480
272
346
208
206
T
T
T
26
T
144
64
48
T.
T
T
11
T
T
5.2
6.6
2.4
"
4
T
3.6
30
14
-
6
--.i ' "
-
Peak Height
F luorescence
.
448
6336
T
7040
T
-
806
-
2112
-
1626
T
858
-
576
3034
-
-
-
T
960
397
. 2432
-
T
T
730
-
1360
-
64
-
1366
9440
T
T
90 (21.68 ng)
—
T
T
4864 (12.82 ng)
102 (10.28 ng)
64
112
131 (0.79 ng)
128
61 (0.35 ng)
48 (0.24 ng)
• "
57
-------
TABLE 21. HPLC ANALYSIS
Retention
Sample Time, min
1A No. 2 0.87
0.93
1.06
1.12
1.2
1.26
1.33
1.37
1.44
1.48
1.53
1.62
1.70
1.86
2.08
2.25
2.38
2.54
2.74
2.87
2.98
3.16
3.70
5.13
5.42
5.53
5.88
6.05
6.90
7.83
8.00
8.40
8.5
8.63
11.53
12.58
16.45
17.1
20.23
20.65
22.33
Response
Identity Ultraviolet
5011
T
208
368
496
736
576
352
352
42
256
T
32
T
25.6
16
T
24
Anthracene
104
Fluoranthene 71.6
Pyrene 43.4
T
T
2.6
Pery 1 ene T
T
7.8
Benz (ghi)
in Peak Height
Fluorescence
.
496
T
-
5120
-
_
3328
-
572
_
3072
_
3405
973
T
1024
2867
_
_
3123
T
2432
307
_
6067
1510
1152
121 (30.84
_
18176 (52.
333 (33.55
-
—
281.6
89.6
208 (1.26
205
43.2
«
70.4 (0.41
ng)
71 ng)
og)
ng)
na)
22.62
perylene
o-phenylene
pyrene
Total Flow
4
21.0 m3
64 (0.32 ng)
58
-------
TABLE 22. HPLC ANALYSIS
Retention
Sample Time, min
2A No. 1 0.76
0.84
1.17
1.2
1.39
1.48
1.58
1.68
1.87
1.99
2.09
2.20
2.40
2.57
3.3
3.66
5.4
7.9
8.35
15.55
17.2
17.9
19.25
— • — — , — .
Response in
Identity Ultraviolet
52.4
53.6
28
12.2
15.4
T
28
T
T
T
t
T
T
Naphthalene 6
_
Fluoranthene
4.6
7
T
T
T
Total Flow 22.7 m3
•••••••^•••••••^^•^^^•i^^W^^^B^M^
Peak Height
Fluorescence
T
133
_
48
T
—
_
19
128
T
109 (0.2 ng)
.
_
_
«.
-
T • Trace
TABLE 23. HPLC ANALYSIS
Retention
Sample Time, min
2A No. 2 0.84
1.2
1.31
1.45
1.5
1.55
1.62
1.72
1.89
2.39
2.63
2.80
3.08
3.27
3.58
3.74
3.86
4.2
4.45
4.85
5.0
5.35
5.55
5.67
6.35
6.50
7.60
8.83
9.45
Response in Peak Height
Identity Ultraviolet Fluorescence
5024
.
583
512
-
608
.
T
112
-
T
44
96
144
-
100
Naphthalene 16
36
T
6
8
8
-
T
.
3.2
Anthracene T
Fluoranthene 10.8
Pyrene 3.2
Total Flow 13.2 m3
.
736
1600
.
384
.
1710
.
1920
2016
-
-
5216
-
371
-
-
3264
-
-
-
-
240
-
80
-
T
2432 (4.4ng)
25.6 (2.6 ng)
T • Trace
59
-------
TABLE 24. HPLC ANALYSIS
Sample
3A No. 1
Retention
Time, min
0.77
1.2
1.33
1.45
1.55
1.70
2.4
2.6
3.6
8.3
8.7
17.55
18.8
19.25
20.43
21.45
22.8
23.1
Identity
Fluoranthene
Benzo (ghi)
perylene
o-phenylene
pyrene
Total Flow
Response in
Ultraviolet
36
23
29
12
12.4
T
_
_
-
3.6
-
5.6
4
T
4.2
2.8
-
-
32.4 m3
Peak Height
Fluorescence
67
-
-
-
_
T
32
93
T
88 (0.56
-
-
_
-
.
T
T
ng)
T « Trace
TABLE 25.
HPLC ANALYSIS
Sample
3A No. 2
Retention
Time, min
0.75
0.80
1.17
1.27
1.38
1.48
3.32
3.75
4.75
7.5
8.05
8.54
16.5
22.4
22.85
Identity
Naphthalene
Fluoranthene
Perylene
Benzo (ghi)
perylene
o-phenylene
pyrene
Total Flow
Response in
Ultraviolet
26
32
T
11.2
T
T
_
4.4
2.4
_
-
5.8
-
-
-
26.1 m3
Peak Height
F luorescence
_
90
_
_
_
64
-
_
T
176 (0.32
20.8
16 (0.1
25.6 (0.
24 (0.12
ng)
ng)
15 ng)
ng)
T = Trace
60
-------
TABLE 26. HPLC ANALYSIS
Sample
4A No. 1
Retention
Time, min
0.75
0.8
0.88
1.25
1.33
1.45, 1.
3.6
8.37
8.82
17.4
18.1
Identity
55
Fluoranthene
Perylene
Total Flow
Response in
Ultraviolet
20
_
44.8
T
8
T
_
3.2
_
-
-
26.4 m3
Peak Height
Fluorescence
26
—
T
«,
T
64
_
144 (0.26
T
T
ng)
T * Trace
TABLE 27.
HPLC ANALYSIS
Sample
4A No. 2
Retention
Time, min
0.75
1.08
1.27
1.35
1.65
1.75
2.15
2.30
2.55
2.60
2.95
3.12
3.41
3.54
3.70
3.87
3.93
4.52'
5.13
8.08
8.50
•
Identity
Naphthalene
Fluoranthene
Pyrene
Total Flow
Response in
Ultraviolet
1320
3112
-
288
_
120
32
-
-
4
33
T
-
18
T
14.2
-
4
_
-
1.4 (1
18.1 m3
Peak Height
Fluorescence
—
-
9600
-
678
-
-
15.5
38.4
-
1312
-
122
-
-
-
1152
-
64
224 (0.4
ng) T
ng)
T • Trace
61
-------
TABLE 28. HPLC ANALYSIS
Retention
Sample Time, min
4A No. 3 0.96
1.03
1.08
1.13
1.38
1.41
1.55
1.70
1.75
2.25
2.30
2.35
2.62
2.89
3.05
3.15
3.27
3.60
3.70
3.85
3.98
4.08
4.2
4.6
4.83
8.8
9.37
16.70
Response in Peak Height
Identity Ultraviolet Fluorescence
496
-
280
T
280
-
T
-
88
28
-
40
11.2
8.4
T
44
T
T
136
Naphthalene 6
21.2
_
_
_
7.6
Fluoranthene T
13
12
Total Flow 23.5 m3
213
-
T
-
5184
-
998
-
-
112
-
58
-
906
_
T
54
_
-
_
T
406
58
_
227 (0.41 ng)
_
-
T = Trace
62
-------
TABLE 29. HPLC ANALYSIS
Retention
Sample Time, min
Response in Peak Height
Identity Ultraviolet Fluorescence
IB 0.72
0.9
1.05
1.19
1.27
1.33
1.48
1.54
1.62
1.9
1.96
2.08
2.14
2.37
2.58
2.80
3.04
3.3
3.6
3.67
4.05
4.15
4.58
4.75
5.0
5.42
5.78
6.35
6.65
6.9
7.53
7.80
7.92
8.23
8.40
8.55
8.75
9.25
9.85
12.5
13.6
15.55
17.4
18.0
20.65
21.1
22 3
L*b • v
22.8
23.1
24.75
Anthracene
Benzacridine
Fluoranthene
Pyrene
Perylene
Benzo (ghi)
perylene
o-phenylene
pyrene
Total Flow
464
_
T
56
_
304
92
52
116
_
T
_
528
T
12
12
10
T
-
T
.
T
.
2.6
13.4
3
4
6
5
-
_
T
_
-
T
_
18.4 (7.86 ng)
2.4
-
-
-
_
T
-
-
_
-
-
-
29.3 m3
1290
2752
1664
4224
_
1242
_
3904
1126
_
640
_
5632
1024
-
307
1075
3840
-
T
T
80
-
16.64
58
96
1312
-
70
144 (34
-
T
48 (0.
-
134
4128 (7
64 (6.
96
96
96
48
288 (1.
136
40
40
32
80 (0.
64 (0.
176
.7 ng)
31 ng)
.42 ng)
5 ng)
74 ng)
46 ng)
32 ng)
T = Trace
63
-------
TABLE 30. HPLC ANALYSIS
Retent i on
Sample Time, min
T • Trace
Response in Peak Height
Identity Ultraviolet Fluorescence
2B 0.76
0.80
0.97
1.10
1.20
1.30
1.44
1.48
1.53
1.63
1.75
2.49
2.60
3.05
3.32
4.12
4.41
4.62
5.52
6.45
6.75
8.8
9.3
56
_
-
6
T
3
T
_
2
-
8.6
.
2.8
_
_
_
_
_
_
_
5
Fluoranthene
Pyrene T
Total Flow 47. m3
256
968
-
-
-
.
T
.
42
-
45
-
62
78
54
T
123
46
32
_
176 (0.3 ng)
19.2 (1.94 ng)
TABLE 31. HPLC ANALYSIS
Retention
Sample Time, min
Response in Peak Height
Identity Ultraviolet Fluorescence
2B 0.75
0.8
1.0
1.21
1.33, 1.45, 1.54
1.64
1.75
2.0
2.63
3.08
3.82
4.1
4.67
5.58
6.55
8.91 Fluoranthene
Total Flow
64.8
_
-
4.2
T.T.T
-
14
.
8
3.2
_
-
-
_
_
T
47.5 m3
141
1024
1446
_
64
_
T
_
109
115
64
186
77
77
346 (0.62 ng)
T » Trace
64
-------
TABLE 32. HPLC ANALYSIS
Retention
Sample Time, min
Response in Peak Height
Identity Ultraviolet Fluorescence
3B 0.75
0.85
1.04
1.22
1.30
1.43, 1.52
1.6
1.72
2.2
3.1
3.58
4.15
4.58
6.55
8.8
68
_
_
_
10
T,T
T
T
_
4.8
_
_
_
Hexane dis-
turbance
Fluoranthene T
Total Flow 42.4 m3
256
3
2195
2790
_
—
128
_
109
128
128
64
218
77
314 (0.57 ng)
T » Trace
Retention
Sample Time, min
TABLE 33. HPLC ANALYSIS
Identity
Response in Peak Height
Ultraviolet Fluorescence
3B 0.78
0.88
0.97
1.0
1.5
1.20
1.31
1.43
1.52
1.60
1.70
1.89
2.00
2.28
2.45
3.05
3.3
3.65
4.2
4.62
5 5
V • ^
8.82
14.00
88
-
4
T
-
6
8
4
2.4
-
9.6
T
T
T
T
4
_
-
4
-
T
Fluoranthene
3.6
Total Flow 42.4 m3
.
256
3072
T
3168
-
-
-
T
186
-
T
T
-
96
96
T
80
48
144
T
112 (0.2 ng)
-
T • Trace
65
-------
TABLE 34. HPLC ANALYSIS
Retention
Sample Time, min
4B 0.78
0.97
1.06
1.15
1.21
1.30
1.55
1.62
1.67
2.13
2.33
2.78
2.95
3.23
3.75
4.45
4.78
4.85
5.73
7.5
7.82
8.23
8.43
Response in
Identity Ultraviolet
259
-
T
24
_
37
-
T
10
T
T
5.2
_
T
T
4
8
-
_
Benzacridine
Fluoranthene
Pyrene 4
9.2
Peak Height
Fluorescence
—
461
-
-
T
2624
128
-
-
-
T
294
T
32
192
-
-
64
80
76.8 (0.49 ng)
304 (0.55
64 (6.45
-
ng)
ng)
Total Flow 42.7 m3
T = Trace
In order to put this data in perspective, it is necessary to correct for
differences in total flow and to attempt to relate the number of fused rings
to relative abundances. In Table 35, the responses were totaled for UV de-
tector responses and also for fluorescence detector responses for components
having retention times up to 6 minutes, between 6 and 15 minutes and over 16
minutes. Generally, compounds having 3 or 4 fused rings elute between 6 and
16 minutes; compounds having 5 or 6 fused rings elute in 16+ minutes. Com-
pounds with fewer than 3 fused rings elute in less than 6 minutes (see Table
19). This is only a very rough approximation because as shown in Table 19,
the sensitivity in terms of response per nanogram ranges from 0.006 to 1.7
for 3 or 4 fused ring compounds with the ultraviolet absorption detector, and
from 0.36 to 88 in the case of the fluorescence detector. With 5 or 6 fused
rings, the ranges are 0.05 to 1.15 with the UV absorption detector and 0.27
to 144 with the fluorescence detector.
66
-------
TABLE 35. HPLC - TOTAL ULTRAVIOLET ABSORPTION AND FLUORESCENCE RESPONSES
Responses. Total Peak Height
Sample
IB
28
2B
3B
3B
4B
1A #1
1A #2
2A #1
2A f2
3A#1
3A#2
4A f 1
4A #2
4A #2
Under 3
UV
1669
78
94
130
83
347
3545
8254
196
7167
112
76
73
4945
1446
(98X)
(94%)
(100X)
(97%)
(100%)
(96.4%)
(98.1*)
(98.3%)
(94.2%)
(99.7%)
(87.1%)
(100%)
(95.8%)
(100%)
(98.3%)
rings
F
28961
1674
3226
7146
6019
3875
28779
60046
328
17457
192
154
90
13120
7989
1
(80.3%)
(88.1%)
(88.4*)
(98.4*)
(95*)
(89.7*)
(63.7*)
(75.4S)
(76.8*)
(87.6*)
(68.6%)
(38.9*)
(38.5*)
(98.4*)
(97.2*)
3 -
4 rings
UV
31 (2*)
5 (6*)
4 (3%)
13
61
128
12
18
3.6
-
3.2
-
13
(2.6*)
(1.7*)
(1.5*)
(5.8*)
(0.3X)
(2.8*)
(4~2*)
(0.9*)
6236
227
423
112
314
445
16038
19008
109
2458
88
176
144
224
227
5-6 rings
FL
(17.3*)
(11.9%)
(11.6*)
(1.6*)
(5*)
(10.3*)
(35.5*)
(23.9*)
(23.2?)
(12.43)
(13.4*)'
(44.4%)
(61.5*)
(1.6*)
UV
T
-
6
12
-
-
13
.
-
-
12
-
-
(0.2*)
(0.2*)
-
(10.1*)
-
-
(0.8*)
Fl
856 (2.4*)
-
368 (0.8*)
591 (0.7*)
-
-
T
66 (16.7%)
T
T
Total
Flow
29.3
47.5
47.5
42.4
42.4
42.7
24.6
21.0
22.7
13.2
32.4
26.1
26.4
18.1
23.5
In Table 35, one may note that the percent of the total response of the
sample is very high for compounds with less than 3 fused rings, with the ul-
traviolet absorption detector, generally well above 90%. The less than 3
fused ring category also commands a major portion of the fluorescence res-
ponse. Thus most of the materials collected were in this category. The num-
ber in parentheses in the table is the percent of the total.
Table 36 shows the relative reponse per cubic meter of flow. The high
sulfur samples show a large decrease in total response in all categories from
idle to higher powers. No large fused ring compounds were found except at
idle. The low sulfur samples show a similar decrease with again virtually no
5-6 fused ring compounds except at idle. The sample, asterisked on Table 36,
with very low total flow, seems to give unusually high values for responses.
Due to water in the Magnahelic gauges flow measurements are subject to inac-
curacies in any case. Values for low sulfur fuel runs are in general higher
than for high sulfur fuel runs. One other very important variable is the ef-
fect of temperature and flow on recovery of PAH such as benzo (a) pyrene (al-
ready discussed), which is a 5 fused ring compound. Even poorer recovery
would be expected with smaller, more volatile substances.
These results correspond well with those for the polynuclear aromatic
hydrocarbons measured by GC/MS which reported higher PAH values for low sul-
fur fuels than for high sulfur fuels and a lowering of PAH with increase in
power. Differences noted in the GC/MS work between samples labeled nA No. 1
and nA No. 2 again appear but concur also with lower total flows for nA No. 2
versus nA No. 1.
67
-------
TABLE 36. HPLC-TOTAL-RESPONSE/it|3 FLOW
Samp I e
Under 3 rings
UV
FL
3-4 rings
UV
FL
5-6 rings
UV FL
IB
2B
2B
3B
3B
4B
1A #1
1A 12
2A #1
2A #2
3A #1
3A f2
4A #1
4A #2*
4A #2
57.0
1.6
2.0
3.1
2.0
8.1
144.1
393.0
8.6
543.0
3.4
2.9
2.8
273.2
61.3
988.4
35.2
67.9
168.5
142.0
90.7
1169.9
2859.3
14.4
1322.5
5.9
5.9
3.4
724.9
340.0
1.1
0.1
0.1
0.3
2.5
6.1
0.5
1.4
0.1
0.1
0.6
212.8
4.8
8.9
2.6
7.4
10.4
652.0
905.1
4.8
186.2
2.7
6.7
5.4
12.4
9.6
29.2
0.2 15.0
0.6 28.1
0.4
T
2.5
T
T
0.5
*Very low total flow
GAS CHROMATOGRAPH-MASS SPECTROMETER ANALYSES
Sixteen filters from the four power points with both high sulfur and low
sulfur fuels were extracted with hexane as described previously. In addition
to these sixteen filters, representing duplicates for each power point for
each fuel, a third filter was processed for one of the points, namely takeoff
with low sulfur fuel.
The extracts were concentrated by controlled, low temperature evapora-
tion to a volume of one mi Hi liter each and delivered to Arthur D. Little,
Inc. for their analysis on a Finnegan Model 4000 mass spectrometer coupled
with a Finnegan gas chromatograph. The chromatographic column was a 20 meter
glass capillary column coated with OV-101 (methyl silicone).
One microliter samples were injected using a Grobe type splitless injec-
tion. The temperature program for the column oven was 55°C for 1*1 min.
followed by a linear gradient of 25.50C/min. to a temperature of 150°C,
and a second linear gradient of 4°C/min to a temperature of 26QOC. Fin-
ally, the 260°C temperature was maintained for 10 minutes. Mass spectrome-
tric conditions are given in Table 37.
68
-------
TABLE 37. MASS SPECTROMETRIC CONDITIONS
Finnigan Model 4000 mass spectrometer
(a) Mass range 100 - 310 amu
(b) Integration 10 ms/amu
(c) Electron multiplier 1800V
(d) Electron energy 70eV
(e) Filament emission 30 ma
(f) Scan rate 1 sec/spectrum
For the quantitative set of GC/MS analysis, the samples requiring con-
centration were evaporated under dry nitrogen from 1 ml to 100/ul for a ten-
fold effect. The other samples were simply kept as 1 ml volumes. (See Table
38 for the division of the samples). In both cases, pheny I anthracene was
added as an internal standard to give a concentration of 0.5 ng/jul in the
final analyzed sample volume. For the concentrated samples, the appropriate
amount of phenyI anthracene was added midway through the evaporation step.
Addition of internal standard to each of the samples allows correction
of the daily instrument variations, giving an accurate comparison of the PAH
levels among the samples, provided calibration data are available. The quan-
tity of internal standard added to the sample should be in large excess of
the quantity of that material originally present, while not being sufficient-
ly large to degrade GC column performance. The initial survey of the samples
indicated that the maximum signal for phenyIantracene in the samples, found
in 1A#2, was less than 100 units. A 0.5 ng sample of pheny I anthracene gives
a signal of more thand 5000 units for the conditions listed in Tables 38 and
39. Thus, the maximum error to the signal of the internal standard is less
than 2% due to residual phenyI anthracene.
To determine calibration factors for specific PAH compounds, a commerci-
ally available mixture of 10 PAH compounds, Supelco Catalog #4-9155, was dil-
uted to 1 ng/ul of each PAH compound. The individual components of the com-
mercial mixture are listed in Table 40. PhenyI anthracene (0.5 ng/ul) was
added to make up the final calibration mixture. Replicate analyses of the
calibration mixture give calibration or response factors to adjust for the
observed dependence of instrumental sensitivity to the different PAH com-
pounds. Table 41 lists the calibration factors for each component of the
standard PAH mixture obtained from the areas of the mass chromatographic
peaks of the appropriate molecular ion. This mixture covers the molecular
weight span of the observed sample species, but not all of the individual PAH
compounds.
69
-------
TABLE 38. TEST AND EXPERIMENTAL PARAMETERS FOR EACH SAMPLE
Sample
Designation
lAfl
1A#2
1B#1
1B#2
2A#1
2Af2
2B#1
2BI2
3A#1
3A#2
3B#1
3B#2
4A#1
4A#2
4A2 rep.
4B#1
4B#2
Power Setting
1 (low)
1
1
1
2
2
2
2
3
3
3
3
4 (high)
4
4
4
4
Sulfur Content
low
low
high
high
low
low
high
high
low
low
high
high
low
low
low
high
high
Analysis
Concentration
none
none
none
none
10X
none
10X
none
10X
10X
10X
10X
none
none
none
10X
10X
TABLE 39. GAS CHROMATOGRAPHIC CONDITIONS
20 meter glass capillary column coated with OV-101
Grobe type - split less injection
Multilinear temperature program
a)
b)
c)
1) 55° isothermal program for 1.1 min
2) 55QC - 150°C linear program at 25.5°C min
3) 150QC - 260°C linear program at 40c/min
4) 260°C isothermal program for 10 min
d) 1 ul sample injections
70
-------
TABLE 40. PAH STANDARD COMPOSITION
1 ng/Ml
Phenanthrene Triphenylene Perylene
Anthracene Benz(a)anthracene Benzo(a)pyrene
Pyrene Chrysene Benzo(e)pyrene
Fluoranthene
TABLE 41. CALIBRATION FACTORS FOR THE STANDARD PAH MIXTURE
Compound
Phenanthrene
Anthracene
Pyrene
Fluoranthene
Triphenylene
Bednz (a) anthracene
Chrysene
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Phenyl anthracene
Calibration Factor
2.19
3.61
2.42
2.66
1.23*
1.23*
1.23*
.734
.565
.804
1.00**
Average Factor
for Cluster
2.90
2.54
1.23
.701
*peaks not separable
**by definition (internal standard)
71
-------
For those PAH compounds for which calibrating materials were not avail-
able, response factors were computed from a least squares fit, as a function
"of molecular weight, of the response factors for the standard materials.
These response factors were then used to correct the GC peak areas specific
to the individual PAH species to give the reported quantitative data.
To identify non-PAH organic species, a wider mass range analysis was run
on the sample with the highest PAH concentration, 1A#2. The species identi-
fied from this GC/MS run were then measured in the remaining samples, rela-
tive to the phenyI anthracene internal standard for each run. All species,
PAH and other, were identified by comparison with reference MS spectra and
correlated with relative GC retention times.
In the analyses of the seventeen samples, specific identifications were
made of PAH and of oxygenate derivatives. Figure 29 shows a representative
mass spectrogram of a sample (1A #2) with specific identifications. PAH with
3 and 4 fused rings such as fluorene, fluoranthene, anthracene and pyrene are
much more abundant and represent the main components in the samples (35,36).
PAH with 5 and 6 fused rings such as perylene, and benzopyrenes are much less
abundant and indeed at power settings above idle are generally not detect-
able. The amount of oxygenated compounds and nitrogen or sulfur containing
species are greater in magnitude than found for PAH at the same power set-
tings and same fuel. This observation is especially apparent at higher en-
gine settings.
100
ISO
200
250
300
350
400
450
500
100 f-
500 550 600 650 700 750 800 850 900 950
1000
Figure 29. Representative mass spectrogram.
72
-------
The quantitative aspects of the results serve to give indications of
trends only. Absolute magnitudes are subject to several variables which
markedly affect them. These variables include uncertainties in flow measure-
ment, and stability of species as a function of temperatures and flow.
Tables 42, 43, 44, and 45 give the amounts found in nanograms/m3 for
the data. Corrections for flow differences between the samples have been
made. The concentration of PAH materials found in each sample shows consis-
tent patterns throughout the samples. The trends described previously show
the behavior of PAH concentrations as a function of test parameters. More
specifically, the distribution of PAH materials seems to be a function of the
power setting. If one considers the four sets of compounds for which there
is direct calibration, anthracene/phenanthrene (m/e 178),f luoranthene/ pyrene
(m/e 202), benzophenanthrene/chrysene/naphthacene (m/e 228), and benzo(a) py-
rene/benzo(e)pyrene/perylene (m/e 252), the level of PAH material maximizes
at the m/e 202 (fluoroanthene/pyrene) cluster for samples taken at low power
while it maximizes at m/e 178 (anthracene/phenanthrene) for the remaining
samples. That is to say that the lower power settings not only generate re-
latively higher PAH levels but relatively higher molecular weights as well.
The high mass species, m/e 228 and 252, are rapidly attenuated as the power
setting increases.
TABLE 42. 6C/HS ANALYSIS-PAH (NG/M3)
Carcinogenicity
.
-
-
to +*
•f-H-
-f
.
•H-f
.
Species
fl uorene PIS^IO
anthracene- I"14H10
phenanthrene
methyl fl uorene C14Hj2
methyl -C14H10 CMH12
fluoranthene '•le^lO
pyrene C16H1Q
aceanthra lyene ^15^12
benzof i uorene ^17^12
benzof luoranthene C^g^g
benzopnenanthrene
chrysene,
naphthacene CjgHjj
benzopyrenes C20t*12*
pervlene C-nH,,
m/e
166
178
180
192
202
202
204
216
226
228
252
252
1A«
3.85
106.0
2.57
27.85
133.6
46.79
13.96
16.19
67.17
48.68
37.62
12.11
1A K
7.52
223.8
ND
76.19
232.3
1195.2
29.10
30.62
94.76
86.66
43.19
19.67
2A «
1.89
3.43
0.06
0.06
1.20
0.69
ND
ND
ND
ND
ND
ND
2A K
13.86
659.7
11.63
136.4
278.3
90.70
79.84
7.13
13.88
6.98
ND
ND
3A 11
0.18
1.14
0.09
ND
0.18
0.12
ND
ND
ND
ND
ND
ND
3A #2
0.33
0.14
0.03
ND
0.47
0.11
ND
ND
ND
ND
ND
M
4A «
0.09
0.53
0.78
0.50
NO
ND
ND
ND
ND
ND
NO
ND
4A K
8.65
68.30
10.17
1400
2.61
1.78
ND
ND
ND
ND
NO
ND
4A #3
1.60
56.62
6.21
72.60
18.60
16.62
ND
ND
ND
ND
ND
ND
Totals
Flow.nT
517.0 2038.1 5.7 1302.3 1.8 1.1 1.9 108.7 173.5
26.5 21 17.5 12.9 32.4 36.1 32 23 21.9
* Sum of signals for both benzo (a) and benzo (e) pyrene, with benzo (e) pyrene contributing much more to the signal than benzo (a)
pyrene.
NO below instrumental detection limit of 0.010 ug/ml, or 0.001 ug/ml for 10X concentrated samples. (Total Sample)
non-carcinogenic
+ uncertain
«•+, etc.
carcinogenic
strongly carcinogenic
73
-------
TABLE 43. GC/MS ANALYSIS-PAH
Species
f luorene
anthracene-
phenanthrene
methyl f luorene
methyl -C14H10
f 1 uoranthene
pyrene
aceanthralyene
benzof luorene
benzof 1 uoranthene
benzophenanthrene
chrysene,
naphthacene
benzopyrenes
perylene
dlbenzothlophene
C13H10
C14H10
C14H12
C15H12
C16H10
. C16H10
C16H12
C17H12
C18H10
C18H12
C20H12*
C20H12
CiAS
m/e
166
178
180
192
202
202
204
216
226
228
252
252
184
IB #1
0.58
22.12
ND
5.56
36.86
3.99
2.29
3.31
33.17
24.10
41.30
5.53
2.83
18 12
5.04
34.06
1.21
9.65
42.58
9.69
3.44
6.99
42.19
25.70
22.66
7.93
5.58
28 #1
0.25
3.34
0.02
0.59
0.42
0.23
NO
ND
ND
NO
ND
ND
0.06
28 K
0.27
26.95
0.57
4.06
34.10
1.87
0.23
0.80
1.37
0.34
ND
ND
1.66
3B 11
2.66
9.25
5.71
42.5
0.77
ND
ND
NO
ND
ND
ND
ND
ND
38 #2
0.13
1.33
0.07
0.45
0.36
0.31
ND
0.18
ND
0.02
ND
ND
ND
48 f 1
0.32
3.57
0.42
4.55
0.79
0.42
ND
ND
ND
ND
ND
ND
1.82
4B 12
0.10
0.61
ND
0.50
0.61
0.40
ND
ND
ND
ND
ND
ND •
0.61
Totals 180.9 214.8 4.2 71.6 23.6 2.2 10.6 2.1
Flow, m3 2.3 29.3 25.6 47.6 47.5 50.8 44.8 37.8 47.6
* Sum of signals for both benzo (a) and benzo (e) pyrene, with benzo (e) pyrene contributing much more to
the signal than benzo (a) pyrene.
ND below Instrumental detection limit of 0.010 ug/ml, or 0.001 ug/ml for 10X concentrated samples. (Total
Samples)
TABLE 44. GC/MS ANALYSIS-OTHER COMPOUNDS
dimethyl & ethyl
napnthalenes
napnthaldehyde
phenyl phenols
f 1 uorenone
benzocinnolines
methy l-benzo-
cinno lines
xanthones
hydroxy-
benzophenone
ant hraqui none
methoxy-
phenanthrene
-cresols
-phenols
Totals
Species
C12H12
Cll¥>
C12H100
C13¥>
C12H8N2
C13H10N2
Wz
C13H10°2
C14H8°2
C15H12°2
C15H24°
C,7H280
Flow, m3
m/e
156
156
170
180
180
194
196
198
208
208
220
248
1A 11
98.49
452.8
535.8
296.6
180.8
106.8
49.81
312.5
58.49
20.94
6.87
39.25
2158.5
26.5
1A 12
107.6
804.8
1346.2
647.6
371.4
193.3
122.9
504.8
100.9
5.19
11.0
71.43
4314.3
21
2A #1
32.11
10.23
5.60
76.00
145.1
17.77
2.06
2.80
2.91
2.97
16.29
49.14
365.17
17.5
2A K
53.49
1170.5
2279.1
1426.4
528.7
554.4
362.0
1682.2
293.0
27.05
22.40
50.70
8449.6
12.9
3A #1
5.25
0.49
ND
16.33
41.36
ND
1.73
ND
ND
ND
21.85
18.81
108.0
32.4
3A 12
9.86
ND
ND
29.08
39.06
ND
ND
ND
ND
ND
44.32
25.76
146.8
36.1
4A #1
6.06
0.59
ND
16.25
ND
ND
1.75
ND
ND
ND
92.50
35.00
153.1
21
4A #2
202.2
686.9
756.5
181.7
7.78
71.30
10.17
2.96
ND
ND
41.43
72.17
2447.8
23
4A #3
20.63
110.0
66.21
80.36
ND
40.00
2.65
16.80
10.55
5.25
155.25
19.5
529.7
21.9
ND below instrumental detection limit of 0.010 ug/ml, or 0.001 ug/ml for 10X concentrated samples. (Total Sample)
74
-------
TABLE 45. GC/MS ANALYSIS-OTHER COMPOUNDS
Species "i/e IB II IB 12 28 #1 2B 12 3B fl 38 *2 4B fl ')
/
74
85
20.08
45.08
47.24
23.23
ND
ND
ND
ND
. 1.38
ND
25.59
25.00
1.14
2.23
1.92
1.94
11.80
1.32
1.36
2.25
(o.36)
\ f ™
2.03
5.94
1.64
9.52
13.68
5.05
ND
4.10
0.74
2.38
0.92
1.38
4.81
3.25
1.41
3.15
2.58
1.76
0.13
1.39
0.50
1.83
ND
ND
1.60
2.58
Totals 1341.3 1421.9 16.8 1105.3 187.0 31.2 47.6 16.8
Flow.m3 2.3 29.3 25.6 47.6 47.5 50.8 44.8 37.8 47.6
*Chronatograph1c overlap of the two components prevented Individual readings from being taken.
ND below Instrumental detection limit of 0.010 ug/ml, or 0.001 ug/ml for 10X concentrated samples. (Total Sample)
All of the samples from high sulfur fuel showed a decrease of the levels
of PAH materials compared to the low sulfur set. This decrease was substan-
tially greater than can be accounted for by the formation of dibenzthiophene
at the observed level. The low abundance of sulfur species could be due to
preferential formation of lower molecular weight material that would not have
been trapped in the hot filter in the original collections.
The oxygenated species show a much less regular pattern than is seen in
the PAH data. The power level does not show the marked effect that is ob-
served for the PAH species, and scatter is apparent at the individual spe-
cies level. However, when the total heteroelement material is compared,
other trends very similar to those of the PAH compounds are observable.
Samples from low sulfur fuel show more total heteroelement material than do
the high sulfur samples, generally. Similarly, with the exception of two
species, anthraquinone and hydroxybenzophenone, the totals for each of the
species for all of the low sulfur samples is greater than for all of the high
sulfur samples.
The general results are as follows:
1) Low power settings yielded higher PAH levels and more PAH species
than did the higher power settings.
2) Low sulfur fuels gave generally higher PAH and aromatic oxygenate
levels than did high sulfur fuels.
75
-------
3) PAH species distribution maximizes at the m/e (mass to charge ra-
tio) of 202, CieHio cluster (fluoranthene and pyrene) for the
lowest power setting, and at the m/e of 178, C^HIQ cluster
(anthracene and phenanthrene) for the higher power settings.
4) PAH species m/e of 204 and higher fell below the detection limit
at the higher power settings.
5) Replicate samples (same power setting and fuel) showed variations
in magnitudes of species found. Variation between replicate sam-
ples seem to again follow the effect of total flow. The lower
the total flow general ly the higher the amounts found per riH.
The correlation applies also to the high sulfur fuels for the
most part and to all tables.
6) Dibenzothiophene was detected in the high sulfur samples but was
not detected in the low sulfur samples.
The total organics measurement gives an indication of the total organic
matter adsorbed on the particulates. The HPLC measurement indicates the gen-
eral magnitude of the 3-6 fused ring compounds. Tables 42 and 43 indicate
the specific 3-6 fused ring compounds and their magnitudes. Tables 44 and 45
indicate other specific compounds formed including two fused ring compounds
and oxygenated derivatives.
Packed Bed Filter Studies
This contract mandates the collection and analysis of the particulate
matter emitted by a gas turbine engine. This interest was based on health
considerations associated with particulate matter. A question remains as to
whether the major part of the organic matter is adsorbed on these particu-
lates or is emitted into the air as a vapor or aerosol and not collected on
the filter. To gain some information on this matter, a series of experiments
was carried out using a packed bed filter to collect both the particulate
matter and these other organic species not adsorbed on the particulate matter.
Packed bed filters were 1/2" O.D. x 6" to 8" long stainless steel tubes
packed with 7 to 12 grams of Chromosorb 102 (styrenedivinylbenzene polymeric
material). The Chromosorb 102 material was prewashed with ethyl alcohol,
methylene chloride and finally n-pentane, as described by Arthur 0. Little,
Inc. (29) to remove any soluble organic material before the Chromosorb 102
was placed in the tubes. Samples of engine exhaust gases were passed through
the packed bed from the engine operating at approach and climb power settings
using low sulfur fuel, and idle and climb power settings using high sulfur
fuel. The volume of gas sampled was between 0.3 and 0.7m^.
After collection of the organic material on the packed beds, the packing
material was removed and extracted with hexane as described previously. This
extract was analyzed for (1) total organics, (2) PAH by GC/MS and (3) boiling
point distribution by methods and procedures also described elsewhere.
76
-------
Table 16 gave the calibration data for various knowns in the total or-
ganics analyses. Recalibration of the instrument showed some change in sen-
sitivity, e.g., fluorene 126.8 div/ng and the composite peak 23.55 div.2/
ng. Table 46 gives the results of the total organics analyses and Table 47
gives the total ng/m3 for the samples. On comparison with the totals from
the samples of extracts of particulates only, one finds the organic level to
be much higher for the Chromosorb 102 samples at like power settings with the
same fuel. Table 48 shows the differences. The organic material on the par-
ticulates represents 0.03 to 0.29% of the total collected on the packed bed
filter.
TABLE 46. CHROMOSORB 102 SAMPLES - TOTAL ORGANIC ANALYSES
Sample
Retention Time
Minutes
Response
Peak Height Peak Area
2A
3A
•
IB
*After flow 'reversed
0.63
U.82
0.9
1.37
1.62
2.27
2.92
3.95
6.98*
0.63
0.80
0.90
1.35
1.62
2.25
2.92
3.3
3.93
6.77*
0.45
0.63
0.82
0.9
1.05
1.28
1.62
2.27
2.92
3.95
6.98*
(Continued)
77
5760
4640
27200
1840
5920
8480
1600
5840
77120
101760
6080
185600
9600
24960
41920
8320
4800
231200
207680
28480
50880
21760
180800
4800
25920
22720
38720
5120
21120
138880
—
fm.
-
_
-
_
-
175833
_
-
_
_
_
_
-
-
-
548275
_
-
-
-
-
-
-
-
-
-
361088
-------
TABLE 46. (Continued)
Retention Time
Response
Sample
3B
Minutes
0.63
0.82
0.9
1.03
1.38
1.62
1.83
2.12
2.27
2.92
4.0
7.13*
Peak Height
12960
2400
27680
37600
2880
2240
5440
18080
4640
12640
6240
178560
Peak Area
_
-
-
-
-
-
-
-
-
-
-
326765
*After flow reversed
TABLE 47. CHROMOSORB 102 SAMPLES - TOTAL ORGANIC ANALYSES
Samp 1 e
Flow
Composite
ug
Light Ends
ug
Total
2A
3A
IB
3B
0.62
0.29
0.36
0.71
7466.4
23281.3
15332.8
13875.4
12042.5
80280.3
42591.1
19542.8
320.9 517.6
527.0 18172.4
1777.5 4937.5
715.4 1007.6
12560.1
98452.7
47528.6
20550.4
78
-------
TABLE 48. TOTAL ORGANICS ANALYSES COMPARISON
Samp 1 e Source
2 A Particulates
Chromosorb 102
3A Particulates
Chromosorb 102
IB Particulates
Chromosorb 102
3B Particulates
Chromosorb 102
Total ug/m3
36.8 (0.29%)*
12560.1
average 32.35 (0.03%)
98452.7
25.8 (0.05%)
47528.6
14.4 (0.07%)
20550.4
*Using data for higher flow sample only
A. D. Little examined hexane extracts of these packed bed filters for
PAH levels. The same chromatograph-mass spectrometer system was used by A.
D. Little, Inc. as in the earlier work discussed in the section on 6C-MS.
The temperature program in the chromatograph was modified to accommodate the
higher vapor pressure species trapped on the resin compared to the filter.
The program was as follows:
1. 55°C injection with 0.8 min. hold
2. 55° to 75°C at 25.50/min.
3. 75° to 16QOC at 4°/min.
4. 1600 to 260°C at IQO/min.
5. 260° isothermal for 10 min.
The results are given in Table 49 for the amounts found in terms of
ng/m3 of exhaust gas. Table 50 compares these totals for the specific PAH
found on the Chromosorb 102 samples with those found on the particulates.
The organic material on the particulates again represents only a very small
fraction of the total, specifically 0.01 to 0.76%. Some indication exists in
the very few samples used that the total organics decreases with power set-
ting and with use of high sulfur fuel compared to low sulfur fuel.
The boiling point distribution analyses of the Chromosorb 102 samples
(see Tables 51 through 58) showed no significant differences from those of
the particulate extracts. The boiling point distribution is shown graphically
in Figure 30 for only the sample collected at the idle power point using high
sulfur fuel.
79
-------
Species
Fluorene
Anthracene/
Phenanthrene
Dibenzthiophene
Methyl Anthr./Phen
Fluor anthene
Pyrene
Aceanthrylene, etc.
Benzof 1 uorene
Benzof 1 uor anthene
Benzanthracenes,
Chrysene, etc.
Benzacridine
Benzpyrenes
Perylene
Total
ND - Not detected.
ng/m3
Composition m/e
C13H10
C14H10
C15H12
C16H10
C16H10
C16H12
C17H12
C18H10
C18H12
C1?H11N
C20H12
C20H12
Instrumenta 1
166
178
184
192
202
202
204
216
226
228
229
252
252
Limit = 0.001
Sample
IB 2A
227.8
9527.8
130.6
3444.4
2513.9
2363.9
2972.2
766.7
988.9
747.2
ND
1400.0
916.7
26000.1
ug/ml
175.8
118.7
12.9
295.2
682.3
614.5
230.6
67.7
200.0
1661.3
203.2
1308.1
408.1
7048.4
3A
96.6
3655.2
ND
1348.2
1041.4
1079.3
1279.3
300.0
324.1
420.7
ND
303.4
648.3
10496.5
3B
39.4
1260.6
16.9
457.7
398.6
390.1
408.5
115.5
271.8
171.8
ND
278.9
133.8
3943.6
TABLE 50. TOTAL PAH LEVELS-GC/MS ANALYSIS
Chromosorb Samples vs.
Sample Packed Bed
2 A 7048.4
3A 10496.5
IB 26000.1
3B 3943.6
Fi Iter Extracts
Filters
ng/rrP
5.7*
1.5
197.8
12.9
% PAH
Fi Iters
vs. Packed Bed
0.08
0.01
0.76
0.33
*Using data- only for higher flow sample
80
-------
TABLE 51. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb 2A (Paraffin Fraction)
Retention Time,
Minutes
0435
07 »1
OR 24
OP 40
0871
08 83
0903
0918
0975
1026
11T30
11 f 6
1142
11 €2
11 *2
1209
1248
1272
1292
1323
1342
1*72
1392
1417
1438
1492
1515
1533
1534'
1686
1644
1698
1752
1804
1823
1853
1905
19*8
1*95
2015
2042
20%
2131
2171
2195
Percent By
Volume
0.032
0.012
0 .0 48
8.017
0.023
5.0 50
0.011
0.142
0.042
5. 127
a. i4i
0.199
oIlSS
0.116
0 .0 74
0.452
0.0 *6
0.583
0.537
3. 9 39
3.922
!).6 US
1.'342
a. 3 79
l.~927
0.591
0.841
0.121
0.121
6.821
l."768
2 .1 51
0.931
0.768
2.039
1.651
1.2*1
1 .2 1 4
0.449
1^830
1.908
1 ."9 96
1.600
0.191
Retention Time,
Minutes
2217
2254
2274
2*00
2334
2354
2378
2411
2431
2454
2444
2500
2527
2555
2570
2597
2622
2665
2687
2732
2752
2795
2857
2917
2975
3013
3090
3UO
3339
3723
Percent By
Volume
6.422
3.735
1.960
5.027
2.708
0.670
5.596
2 ."8 32
0.593
41643
2.625
a. S7i
4.5*3
3.071
0.450
3.260
2.179
3.872
2.510
2.722
2. "073
3.297
2.268
1 .'* 09
1.220
0.838
0.461
0.152
0.003
0.003
81
-------
TABLE 52. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb 2A (Complete)
Retention Time,
Minutes
039*
0*09
0597
06 76
0695
C716
07*6
0761
^•d« ^_
fl7 7**
U i • J
o» ?2
08 09
OP 35
08 59
0875
C8«?2
0922
0932
0967
1012
1252
in 61
1073
1088
11 12
1164
11 »7
1212
1249
1281
1290
1323
1366
16 97
Ul*
* ^r
it 73
1532
t ^ ^ -*
IS 61
15<*2
16 38
16 *2
X659
Percent By Retention Time,
Volume Minutes
0.078
0.060
0.003
0.116
0.135
0.097
0.292
0.119
^1 j* ^ ^1 A
0 .2 38
0.2**
0.195
0.02*
0.069
Si-,.,-
.008
OA *m —,
.198
0.220
0.903
0.5*5
0 !"0 52
0.192
2.411
0.*28
0.153
0.068
0.080
0.*36
0.281
3.102
0.680
0.83S
0.*.67
0.737
0.378
U v vr • *J
1 .'9 52
0.8*3
1" *• A »
A C ft 4
• J *J
0.528
i' '"' !• • ^h
.1*2
0.872
0.817
2.0C2
1671
1692
1727
1769
1790
Mil
1852
m — «k«
18 93
1899
1930
1959
1998
20*3
2077
2091
2153
21*2
2219
2250
2293
2322
25*7
2*02
2**9
2*t2
2*87
2532
2560
2578
26 n
266*
2709
27*8
2« fig
2822
2865
2921
2*76
3028
3061
3092
3093
3136
3190
3*76
3708
Percent By
Vo 1 ume
1.0*8
2 .9 86
1 .*03
2.872
2.623
2.988
?.*62
2.071
0.286
3.979
2.809
2.90*
1.315
0.926
1.297
11890
3 . * t6
2.323
1.1 ft*
0.103
5 .117
0.*62
3 .1 91
2.161
1.360
1 .* 60
2.801
2.252
*.3Q6
1.057
1.7C8
1.252
0.721
0.97*
1.226
1.082
0.658
0.362
0.417
0.*S2
0.202
T
T
82
-------
TABLE 53. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb 3A (Paraffin Fraction)
Retention Time,
Minutes
0826
0979
1023
1046
1068
1096
ills
1134
1159
iltS
#13
1243
1265
12*6
13 14
13 33
1361
14 H7
14 *0
1502
1=21
1*71
15 53
16 31
1694
17*0
17 35
1« 13
1845
IB 99
1945
19 59
2039
20<57
2128
2213
2249
2270
2295
23 29
2*50
2373
Percent By
Volume
0.010
0.031
0.100
0.228
0.218
0.594
0.616
0.127
0.284
0.1*8
3.2*4
0.838
0.2 38
1.076
1.4*5
2.012
2.121
1.770
2.436
1.548
0.421
2.804
0.398
0.234
1 .6 10
3.0*9
2.332
0.974
0.709
2.320
1.431
1.161
0.668
1.023
1 .3 «1
1 .3 80
7.861
2.614
0.311
4.164
1 .624
1.552
3.277
1.970
Retention Time,
Minutes
2426
2449
2478
2495
2521
2550
2591
2617
2659
2743
2786
2840
3112
Percent By
Volume
1.575
3.490
O.I 93
3.197
3.020
3. ids
2.619
3.854
7.305
3.541
3Y023
0.254
a. 001
83
-------
TABLE 54. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromos.orb 3A (Complete)
Retention Time,
Minutes
0*5*
C369
0*10
0*28
0533
05*6
0563
0590
0675
OS 91
0726
0753
0766
07*»0
0795
08 1*
08*2
0»79
0=4.36
0971
10 16
1054
1072
10 94
11 18
1119
1171
1194
1216
1218
1255
1263
1?«8
13 ?0
It 50
1*75
1*3*
1122
1*80
ic in
A~- A.U
1532
1578
15=0
Percent By
Vo I ume
0.001
0.253
0.657
0.696
0.013
0.016
fl.lfig
8. '01 3
0.155
0.382
0.206
0.*87
0.199
0.517
3.045
0.626
0.170
5.038
0.39*
1 .'2 69
0. 61*
0.208
D.I 32
3. 0»5
0.385
0.101
0.32*
0.18*
0.1 »7
0.267
0.209
0.173
0.72*
1.270
1.302
D. 732
l."373
0.528
0.639
U • " -^ J
3.216
2.*12
1.112
1.***
Retention Time,
Minutes
1650
1646
166*
1696
1711
1773
1792
1«13
185*
1873
189*
1961
2001
20*7
2082
2111
2169
2225
2257
2297
2327
2155
2*57
2521
2517
2569
26*1
2671
2713
27*9
28 13
2869
2921
2977
3033
3069
3089
31*1
3195
32*5
3*11
Percent By
Vo 1 ume
1.801
1.102
*.338
*.095
1.661
3.186
2.071
3.173
1.370
1.128
2.02*
1 .1 7S
0.439
1.325
0.501
' n.626
0.251
2.396
0.677
1.012
0.092
*.261
5.058
D.*28
0.8*3
4.571
2.52*
3.470
0.692
3.895
3.056
0.353
2.486
1.803
0.270
0.366
0.82*
0.432
0.118
0.007
0.001
84
-------
TABLE 55. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb IB (Paraffin Fraction)
Retention Time,
Minutes
05 45
Of 17
0697
0718
0713
07*8
07 30
OP 17
CMS
0872
09 P7
09 64
C>9«4
1015
1041
1054
1090
1136
1132
1153
Ilt2
i?ao
12*1
12 52
12 82
m2
1?31
1359
13 «5
1ft 0*
IH 2*
1465
1**78
1502
1523
1636
17ft*
1797
1816
19+6
' IP «8
19*2
19 "7
20^5"
Percent By
Volume
3.00*
0.064
0.114
0.053
0.038
0.133
0.490
0.5*6
0.3S8
0.800
0.613
0.5*6
0.209
1.355
0.1*9
0.449
0.906
1.236
0.318
0.899
0.6<>8
0.356
1*423
0.416
1.639
1.895
3.606
3.048
2.604
2.946
1.868
3.140
8.559
2.252
3.546
0.950
0.979
0.542
0.438
1.679
0.692
0.454
0.097
0.308
Retention Time,
Minutes
2091
2122
21«iS
2256
2243
2263
2287
2323
2342
2366
2397
2417
2441
2471
24*7
2513
25 %2
25 S3
2fc ?9
26 «
2651
2717
2737
2762
27*3
2»42
31 3S
3592
3*12
Percent By
Vo 1 ume
0.454
0.7R7
2.500
4.213
1.S04
9,802
3.956
1.0 »9
1.029
4.1«2
1.635
1.3 "8
3.954
0.738
2.173
3.466
2.831
2.754
1.'7 25
1.542
4.380
0.869
1.872
0.826
2.064
0.363
0.001
0.023
0.021
85
-------
TABLE 56. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb IB (Complete)
Retention Time,
Minutes
03 S2
C399
0416
0486
0530
0547
0561
0579
os oo
0641
W71
0712
0729
0743
0758
0773
0786
OP 06
Oft 32
0860
08 73
0932
0967
10 19
1062
1089
1112
n4M At
P4
1186
1214
1241
1258
1280
1324
1344
1366
1398
1415
1437
1474
1504
1525
1564
15fi6
Percent By
Vo 1 ume
0.305
1.1 07
1.002
0.001
0.006
0.1 «2
0.0 SO
0.062
0.0 S3
1.290
2 .0 57
0.898
3 ;~S 31
0.908
0 .4 97
17243
1.454
2.859
3.007
0.528
2 .118
3.032
3.846
3.050
11665
8.949
li'962
0 .6 88
0.455
1 ;5 77
0.750
1 .3 80
1 .*6 90
2.005
1 .2 19
1 .811
0.360
0.9*8
3.604
1.'348
2.2*2
0.816
1.397
Retention Time,
Minutes
1612
1643
1661
1695
17 29
1773
1791
1812
1853
18 «4
1960
2000
2021
2045
2? 24
2356
2454
2472
2536
2562
2580
2638
2666
27 $2
2*12
3062
3094
3300
3446
Percent By
Vo 1 ume
1 ".'1 65
0.920
3.748
3.595
1.567
2.618
1.515
2.531
3.239
3.515
2.198
a. 70s
0.1 55
0.584
0.169
0.054
0.418
5.191
0.227
0.446
0 .9 38
0.969
1 18 84
0.039
0*052
0.014
0.080
0.013
0.001
86
-------
TABLE 57. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb 3B (Paraffin Fraction)
Retention Time,
Minutes
0*36
0*57
OS SI
0630
07 n
07 *S
0765
0789
0828
08*2
0882
0916
0933
0971
0993
1020
1059
1097
llfl
1136
11^7
1177
1197
12*2
1265
1286
IMS
1*3*
1163
I* *2
1*87
1* HI
1503
15 ?2
1572
15 9*
1631
1683
1735
1787
1803
18 36
1891
1936
1990
2032
2122
Percent By
Volume
0.115
0 .102
0 .009
0.006
0.023
0.022
0.057
0.109
0.205
0 .0 97
0.2*8
8.316
0.1 3*
0.367
0.255
07873
0.579
Q.*59
0.619
01830
O.**i
0.271
0.726
j^ ^ &
0.1 f 5
0.7*0
0.56*
0.973
0.958
0.699
3.805
0.570
0.1*9
0.96*
0 .1 70
d.o«*
0*666
1.326
17*00
0.569
3 .* 63
1.309
1.'022
1.'02*
0.818
1.338
3.269
Retention Time,
Minutes
2185
2208
22*6
2265
2290
2326
23*5
2369
2101
2*21
2**4
2*7*
2*90
2517
25*5
2559
2586
2612
2638
265*
2718
2T60
2P *1
2901
2959
3017
3073
3123
3177
3256
Percent By
Vo 1 ume
2.82*
*.776
17829
T.*785
4.782
2.031
Ii7gb
*.*77
2.780
r«2§6
4.601
r;38s
3.967
2. '3 52
0.918
1 .506
1 {8 75
5'!l34
37557
0.618
37362
2.727
2.234
17570
U19S
0.762
0.4 82
0.233
0.0*5
87
-------
TABLE 58. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Chromosorb 3B (Complete)
Retention Time,
Minutes
0367
0*07
0*23
0636
0700
0717
0762
0779
0793
0809
0876
0932
0966
1017
1058
1086
1139
1157
1191
1205
12*2
1275
1318
1340
1360
1352
1*18
1*53
1*67
1500
1524
1561
« •• dt ^
1583
1622
16*2
1660
1672
1694
1730
17 49
1773
1791
181*
1" 56
19 C3
Percent By
Volume
0.0*9
0.167
0.154
0.009
0.238
0.076
0.010
0.028
0.017
Ov.163
0.003
0.2**
Ol»655
0.5*6
0.293
1 ."3 29
0.296
0.193
0.066
0.003
1.169
0.19*
0.366
0.*83
0.2*3
0.*37
0.659
0.*6»
1.621
0.5*7
0.779
0.298
1 .2 17
0.818
0.271
1.369
0.383
1.5*4
0.797
0.* 96
1 '"."2*1
0.751
!.**!
1 .8 38
2.6*4
Retention Time,
Minutes
196*
2003
20*9
208*
2105
2129
2167
2228
2262
2300
2330
2356
2*52
2523
25*0
2570
2585
264*
2673
2717
2753
28 16
2850
2871
28 85
292*
30 C*
30*5
3068
31 02
31*2
3200
3*36
3624
Percent By
Volume
15.16*
1.311
1 .971
0.911
o.ii*
0.982
l.*079
2.906
11 ."6 6*
1.8*7
0.5*1
*.*10
5.812
0.987
1.061
1.320
2 .3 7*
1^953
?.3*6
1.0 88
179 36
1 ^S 72
0.63*
0 .1 U
1.387
0.933
tJV*12
0.*93
3.645
1.220
0.831
0.850
T
0.005
88
-------
12
10
Ul
3
§
>- 0
co
ui
o
ec
uj
o.
12
10
RETENTION TIME (WIN.) 0
BOILING POINT (°C)
COMPLETE
PARAFFIN FRACTION
,iiJiiii
i
I liilh.il
I
i
|
|
i
8
106
12
144
16
178
20
215
24
250
28
287
32
322
36
358
Figure 30. Graphic representation of boiling point distribution tables 55
and 56, idle power point using high sulfur fuel.
Nitrosamine Analyses^
Two analyses for nitrosamines were made of the extracted fraction of ex-
haust partial lates using a Perkin Elmer nitrogen-phosphorous detector. The
samples were taken at idle and climb power settings using low sulfur fuel.
The nitrosamines were extracted from the teflon filters and isolated in
the dichloromethane fraction (10 ml) using the procedure described in EPA
650/2-75/056.
89
-------
A Perkin Elmer nitrogen-phosphorous detector (Figure 31) was coupled to
a Perkin Elmer model 3920B gas chromatograph and run in the nitrogenphosphor-
ous mode. The detector uses as an alkali source, a rubidium bead, which is
heated independently of the flame with an internal wire. The flame functions
-only to ionize the sample. Due to a relatively cool flame, nitrogen contain-
ing compounds undergo a partial pyrolysis and produce intermediate cyan radi-
cals. These take up an electron from the alkali and the resulting symmetri-
cal cyanide ion migrates to the collector electrode where it liberates an
electron which can be detected by an electrometer.
VENT
COLLECTOR
ELECTRODE,
RUBIDIUM
BEAD —
r^HB
FLAME JET •
AIR
JET
POLARIZING
LEAD
COLUMN
EFFLUENT
_ MO _
MODE
-P-
MODE
Figure 31. Nitrogen phosphorous detector.
The sensitivity of the nitrogen-phosphorous detector is reported by
Perkin-Elmer to be at least 0.5 coulomb/gram for nitrogen. It was estimated,
while using calibration standards with the instrument, that the lower limit
for compounds of interest was approximately lO'*3 g. The linear range for
".he instrument was
The following conditions were observed during calibration and sample
runs.
Column: 6' x .125" 10% Carbowax 1540 (polyethylene glycol, molecular
weight 1300-1600) on ABS (acid and alcoholic base washed and silanized
diatomaceous earth), 60-70 mesh
Detector temp.: 165°C
Injector temp.: 165°C
Column temperature program: 117°C/8 min./8°C rate/165°C/16 min.
Detector bead setting: 5.40
Helium carrier: 17 ml/min., 93 psi
Hydrogen: 3 ml/min., 7.5 psi
Air: 100 ml/min., 44 psi
A 1 ul injection of dichloromethane extract was analyzed to determine if
nitrosamine interferences might be present. None were observed.
90
-------
A 0.2 ul injection of a nitrosamine standard containing 0.05 ng each of
dimethylnitrosamine, diethylnitrosamine, diisopropyInitrosamine and dibutyl-
nitrosamine was analyzed. Retention times, peak heights, and divisions per
pg are noted below.
Retention Peak Height Sensitivity.
Compound Time. Min. Divisions div/pg
DMA 5.5 15.8 316.0
DEA 7.7 12.0 240.0
DIA 12.2 9.0 180.0
DBA 16.3 5.5 110.0
Nitrosamines were not found to be present at either idle or climb power
settings using the instrumentation and detection methods described above. If
nitrosamines are present, they are below the lO"13 g detection limit of the
instrumentation.
Phenol Analyses
In addition to the analysis of the organic fraction of engine exhaust
particulate material for polynuclear aromatic compounds, the analysis for
phenols was also undertaken. Phenolic compounds, although not necessarily
carcinogenic themselves, have a synergistic effect in conjunction with cer-
tain polynuclear hydrocarbons. The phenols have a tendency to make these
polynuclear compounds much more carcinogenic than they would be alone. Two
samples of exhaust particulate were taken on teflon filters. The power con-
ditions for these samples were idle and climb out, and the fuel used was the
low sulfur type. The extraction of the phenolic compounds was performed in
accordance with the procedure given in EPA 650/2-75/056.
As a result of using the prescribed extraction techniques, the phenols
were taken up finally in diisopropy!ether (DIE). A one microliter aliquot of
this solution was injected into a gas chromatograph for analysis. Prior to
this step a calibration procedure was used to ascertain retention times and
sensitivity of six common phenol type compounds. To account for possible in-
terferences a blank was produced by using the extraction technique on an un-
exposed fiIter.
Six phenol compounds were dissolved in DIE each at a concentration in
the final solution of 17 ng/ul. The phenols used in this calibration were
phenol, o-cresol, m-cresol, p-cresol, 2, 6-dimethyl phenol and salicylalde-
hyde. The analysis was performed on a Perkin Elmer Model 3920B gas chromato-
graph using the following conditions.
Column: 6' x 0.125" stainless steel column packed with, 10% OV-3
(si Iicones with 10% phenyl) + 1% FFAP (free fatty acid phase
Carbowax 20M reacted with nitroterephthalic acid; Carbowax 20M is
91
-------
polyethylene glyco! of average MW 15000-20000) on gas Chrom. Q (acid
and alcoholic base washed and silanized diatomaceous earth), 80/100
mesh.
Carrier gas: Helium, flow 70cc/min @ 93 psi
Detector: flame ionization detector, hydrogen fuel (28 psi) air
oxidant (48 psi)
Temperatures: Oven: 105°C isothermal
Injection Port: 160°C
Detector: 150°C
Using the gas chromatograph, under the conditions described above the
retention times and sensitivities of the six phenols were obtained as shown
in Table 59.
TABLE 59. RETENTION TIME AND SENSITIVITY OF PHENOLS
Retention Sensitivities (Div/ng)
Compound Time, Min (Peak-Height)
Salicylaldehyde 5.8 4.27
2, 6-Dimethylphenol 11.8 3.22
Phenol 12.8 2.87
o-Cresol 14.6 3.34
m-Cresol 20.3* 4.09
p-Cresol 20.3*
* m and p - Cresols could not be separated.
A chromatogram of this mixture showing five peaks that represent the
six phenols (m and p cresol did not separate) is shown in Figure 32. A
chromatogram of a typical sample (IA) is shown in Figure 33.
The chromatographic analysis of the two samples (idle and climb out, us-
ing low sulfur fuel) showed various peaks but only phenol could be identified
positively from its retention time. About 4 ng of phenol was found in the
idle sample and about 1.3 ng was in the climb out sample. The concentration
of phenol in the exhaust gas sampled was calculated to be approximatly 0.15
ng/m3 anc| 0.047 ng/m^ for idle and climb out respectively.
92
-------
PHENOL MIXTURE
1. SALICYLALDEHYDE
2. DIMETHYLPHENOL
3. PHENOL
4. O - CRESOL
5. P&M- CRESOL
10
12
16
18
20
22
Figure 32. Chromatogram of mixture of phenols.
PHENOL
10
12
22
Figure 33. Chromatogram of typical exhaust sample.
93
-------
Spectral Data
Ultraviolet, infrared and nuclear magnetic resonance spectra were taken
on some of the samples. These spectra were taken without separation of the
extracts of the adsorbates. Therefore only broad generalizations can be made
for the complex mixtures analyzed. These are detailed below.
Nuclear Magnetic Resonance Analyses
Samples of particulate matter were collected at idle and at climb using
both low sulfur and high sulfur Jet A-l. The sample analyses were conducted
at the Southern New England High Field NMR Facility at Yale University's
Department of Chemistry. The support of the New England High Field NMR
Facility, made possible by a grant from the Biotechnology Resources Program
of the National Institute of Health (RR-798), is gratefully acknowledged.
Proton NMR spectra were run at 270 MHz on a Bruker HX 270 spectrometer
using the Fourier transform mode. Deuterium resonance was used for a field/
frequency lock and CDC13 was the chosen solvent. The instrument is extreme-
ly sensitive. A 0.001 M sample with sharp resonances will yield adequate
spectra in a half hour. Operations such as homonuclear decoupling and inte-
gration are available. The signal to noise ratio measured on the highest
peak of the quartet in a one pulse spectrum of 1% ethyl benzene is 120:1.
The organic fraction of the particulate samples was extracted using
CDC13 as the solvent in a Soxhlet extractor. The resulting solution was
passed through a 10 u teflon filter to remove any particulates carried over
during the extraction. The volume was reduced to 1 ml before insertion into a
5 mm O.D. NMR tube. Samples were kept under refrigeration until analysis.
Total flows, calculated particulate accumulations (based on flow and es-
timated mass emissions data determined earlier) and filter temperatures are
tabulated below:
SAMPLE
1A
3A
IB
3B
POWER
SETTING
Idle
Climb
Idle
CI imb
TOTAL
FLOW
MASS
ACCUMULATION
80.7 mg
194.9
89.9
278.3
FILTER
TEMPERATURE
113°F
135
149
139
In addition to the sample spectra, a 20 ul sample of each fuel in
CDC13 was run to document any possible differences.
94
-------
The spectra obtained appear to delineate three general regions. The
aromatic region, about 7 to 8 ppm (delta shift) is well defined. For the
purposes of general data interpretation, a delta shift of 0 to 2.5 ppm will
be defined as being largely aliphatic in nature. Those shifts lying between
the aliphatics and the aromatics will be defined as having olefinic charac-
ter. With these suppositions in mind, the following Table 60 was generated
which lists integration counts normalized to 2000 scans. Note that 0.2 mg
benzo (a) pyrene yields 90 counts when normalized to 2000 scans. This gives
a rough correlation between integration counts and the amount of material
present.
Table 60 shows the integrated reponse of groups designated aromatic,
aliphatic, and olefinic in integration counts per cubic meter of sample gas
and can be used as an approximation of the amount of material present.
TABLE 60. NMR INTEGRATED RESPONSE /m3
Sample Aromatic^;
1A
3A
IB
38
Id
Cl
Id
Cl
le
imb
le
imb
0.
0.
5.
0.
836
252
666
605
Aliphatic^;
7.
7.
7.
4.
300
554
167
811
Olefin
3.
2.
2.
0.
ilj) Total
042
806
321
957
11.
10.
15.
6.
178
612
154
373
(1) Delta shift 7-8 ppm
(2) Delta shift 0-25 2.5 ppm
(3) Delta shift 2.5-7 ppm
(4) Integrated response per mg BAP is 450
Table 61 is an adjustment of the NMR response on a hydrocarbon basis.
If CsH; is assumed a representative aromatic, CIQ^O a representative
olefin, and CioH22 a representative aliphatic, the adjusted distribution
of counts per cubic meter of sample gas is as described.
95
-------
TABLE 61. NMR RESPONSE - HYDROCARBON BASIS
Samp
1A
3A
IB
38
le
Idle
C 1 imb
Idle
Climb
Aromatic
(as C8 H7)
12.30
3.71
83.37
8.90
Aliphatic
(as CIQ H22)
47.16
48.80
46.30
31.08
Olefin
(as CIQ H2g)
21.29
19.64
16.25
6.70
Total
80.75
72.15
145.92
46.68
Table 62 expresses adjusted counts per cubic meter of sample on a per-
cent hydrocarbon basis. Jet A-l is included for comparison purposes. This
table serves as a qualitative assessment of the hydrocarbon distribution at
each power setting sampled.
TABLE 62. PERCENT HYDROCARBON BASIS
Sample
1A
3A
IB
3B
Jet
Id
Cl
Id
Cl
le
imb
le
imb
A-l
Aromatic
15
5
57
19
12
.23
.14
.13
.07
.08
Aliphatic
58.
67.
31.
66.
82.
40
64
73
58
89
Olefin
26.
27.
11.
14.
5.
37
22
14
35
30
Figures 34 through 39 show the actual NMR spectra obtained for typical
samples: samples from low sulfur at idle and climb, samples from high sulfur
at idle and climb, and low and high sulfur Jet A-l fuels.
Both the high sulfur and low sulfur idle samples were divided into five
equal fractions each and were brought to 1/2 ml volume. To four of these
samples was added approximately 50 ug anthracene, pyrene, fluoranthene and
phenanthrene, respectively. This resulted in two groups of samples whose
only difference was a measured quantity of known contaminant.
96
-------
SAMPLE: LOW SULFUR IDLE
NUMBER OF SCANS: 1720
PPM 5 SHIFT
Figure 34. Nuclear magnetic resonance - low sulfur idle.
10
SAMPLE: LOW SULFUR CLIMB
NUMBER OF SCANS: 2000
J.
5
PPM 8
Figure 35. Nuclear magnetic resonance - low sulfur climb.
97
-------
SAMPLE: HIGH SULFUR IDLE
NUMBER OF SCANS: 1000
Figure 36. Nuclear magnetic resonance - high sulfur idle.
SAMPLE: HIGH SULFUR CLIMB
NUMBER OF SCANS: 2000
Figure 37. Nuclear magnetic resonance - high sulfur climb,
98
-------
SAMPLE: JET A-1 LOW SULFUR
NUMBER OF SCANS: 200
J_
10
5
PPM 6
Figure 38. Nuclear magnetic resonance - Jet A-1 low sulfur.
_L
10
5
PPM 6
SAMPLE: TET A-1 HIGH SULFUR
NUMBER OF SCANS: 200
Figure 39. Nuclear magnetic resonance - Jet A-1 high sulfur,
99
-------
Spectra were obtained for each of the above groups using identical run
parameters within each group so they could be overlayed and compared. If a
doped sample matched peaks with a non-doped sample and showed a significant
increase in the magnitude of the peak response, a match could be assumed.
When this was done, a probable match was obtained for fluoranthene and
phenanthrene in the high sulfur idle samples. Their presence agrees with the
GC/MS results.
Table 60 shows the expected decrease in hydrocarbons as power in-
creases. Total counts per m3 of exhaust gas decreased from 11.2 to 10.6
when going from idle to climb with the low sulfur fuel. The percentage aro-
matic material (Table 62) appears to decrease with power setting advancement
while the percentage aliphatic appears to increase. The high sulfur fuel
samples contained a larger quantity of aromatic at a given power than the low
sulfur samples.
Ultraviolet Analyses
Twelve UV scans were made of the extracted fraction of exhaust particu-
lates collected from twelve exhaust samples. Total sample flow, average fil-
ter surface temperature, and calculated mass accumulation (based on total
flow and estimated from mass emissions data) are given in Table 63.
TABLE 63. SAMPLE DATA FOR ULTRAVIOLET ANALYSIS
TotaI Average Mass
Power Flow Filter Accumulation
Sample Setting (m3) Temp. (mg)
1A#1 Idle 26.4 125°F 80.8
2A#1 Approach 17.5 - 81.4
3A#1 Climb 32.4 124 230.7
4A#1 Takeoff 32.2 139 248.3
1A#2 Idle 21.0 107 64.3
2A#2 Approach 13.2 133 61.4
3A#2 Climb 36.5 128 259.9
4A#2 Takeoff 23.5 119 181.2
IB Idle 29.3 150 89.7
28 Approach 47.5 161 220.9
3B Climb - 154
4B Takeoff 42.7 142 329.2
100
-------
The organic fraction was removed from the participates using a Soxhlet
extractor and n-hexane as the solvent. The solvent was chosen for its abili-
ty to dissolve most of the collected organic material and its freedom from
interfering peaks when injected into the liquid chromatograph. The above
samples are identical with those used for the liquid chromatograph analysis.
The UV sample runs were made after completion of the liquid chromatography
runs. All samples were brought to 3 ml volume to accommodate a 1 cm cell.
Ultraviolet and visible spectra were run on a Varian 635D spectrophoto-
meter at a slit width of 0.5 nm and a scan speed of 100 nm per minute. Scans
were made from 800 through 200 nm. Cells were Suprasil with a useable wave-
length of 165 to 2600 nm. All engine sample runs were made with n-hexane as
the reference in the double beam mode.
Figures 40, 41, 42, and 43 show the UV absorption spectra of hexane, un-
decane, fluoranthene and a mixture of 16 polynuclear aromatic hydrocarbons.
The aliphatic compounds show absorptions at around 230 and 270 nm; the poly-
nuclear compounds show absorptions at 410 and 435 nm with a broad band of
high intensity between 220 and 380 nm.
U4
o
<
00
cc
o
W
CO
UV-VIS SCAN
HEXANE
REFERENCE; Am
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
700
600
500
400
300
200
100
NM
Figure 40. Ultraviolet spectra of hexane.
101
-------
CO
CC
O
c/j
CO
UV-VIS SCAN
UNDECANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
AIR REF.
J
100
800
700
600
500
400
300
200
NM
Figure 41. Ultraviolet spectra of undecane.
HI
U
CD
CC
8
m
UV-VIS SCAN
FLUORANTHENE
SOLVENT: CHLOROFORM
REFERENCE: CHLOROFORM
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
700
100
Figure 42. Ultraviolet spectra of fluoranthene.
102
-------
UV VIS SCAN
PS101 (16 POLYNUCLEARS)
REFERENCE: N-HEXANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
700
600
500
400
300
200
100
NM
Figure 43. Ultraviolet spectra of 16 polynuclears in n-hexane.
Spectra of exhaust samples show broad absorption bands between 220 and
320 nm from idle, low sulfur samples with the general trend to lower absorp-
tion intensities and lower wavelengths as power increases and as fuel is
changed from low sulfur to high sulfur. These trends suggest a lowering of
aromatic/PAH content. Figures 44, 45, 46, 47, and 48 show representative
examples for sample 1A #1, sample 3A #1, sample IB, sample 2B, and sample
4B, respectively.
Sample 1A#1 from idle power, low sulfur, showed a much higher UV absorp-
tion and at higher wavelengths than sample 3A#1 climb power, low sulfur. This
confirms the decrease in PNA/aromatic content as power increases as evidenced
by higher UV absorptions and higher wavelengths with decrease in power. The B
samples compared to the A samples (specifically IB and 1A#1) show a similar
trend of decrease in PNA/aromatics with increase in sulfur content.
103
-------
UV-VIS SCAN
SAMPLE: 1A #1
REFERENCE: N-HEXANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
LU
O
-------
ffi
CC
O
CO
UV-VIS SCAN
SAMPLE: IB
REFERENCE: N-HEXANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
700
600
500
400
300
200
100
NM
Figure 46. Ultraviolet spectra of n-hexane - sample IB.
CO
oc
o
co
CO
UV-VIS SCAN
SAMPLE: 2B
REFERENCE: N-HEXANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
700
600
500
400
300
200
100
NM
Figure 47. Ultraviolet spectra of n-hexane - sample 2B.
105
-------
Ill
o
<
CO
DC
O
V)
m
UV-VIS SCAN
SAMPLE: 4B
REFERENCE: N HEXANE
0.5 NM SLIT
0-10 ABS. FULL SCALE
800
200
100
Figure 48. Ultraviolet spectra of n-hexane - sample 4B.
Infrared Analysis
Four exhaust samples were collected and infrared scans were made of
their extracts. Total sample flow and average filter surface temperature
are as foilows:
Sample
1A
3A
IB
38
Power Setting
Idle
Climb
Idle
CI imb
Total Flow (m3) Avg. FiIter Temp.
26.4
27.8
29.4
42.4
1090F
120
117
126
The organic fraction was removed from the partiallates using a Soxhlet
extractor and carbon disulfide as the solvent. Carbon disulfide was chosen
both for its compatability with IR techniques and its ability to act as a
suitable solvent for the majority of extractable material - including PAH.
The carbon disulfide containing the extracted organic material was reduced to
a 1 ml volume by evaporation at room temerature. A stream of dry nitrogen
was passed over the sample to aid in the evaporation process. No attempt was
made to separate the organic material into organic fractions.
106
-------
Scans were made using both a Beckman IR20A and a Perkin-Elmer model 283
spectrophotometer. The frequency scanned was 4000 to 600 (CM-*). A sealed
cell with 0.5 mm path length was used for each of the extracted particulate
samples. A 0.1 mm eel I was used in producing an IR scan of both the low sul-
fur and high sulfur Jet A-l fuel. The instruments were run double bean with
carbon disulfide as the reference for the extracts of the particulate samples
and air as the reference for the fuel samples.
The detectable limit of the IR20A was determined as 0.3 mg/ml using py-
rene. No major peaks could be discerned from baseline noise below this
level. It was concluded that a total of organic materials considerably more
than 0.3 mg/ml would be necessary to achieve sufficient response from the in-
strument to identify specific functional groups in complex mixtures.
All spectra show the expected presence of aliphatic, olefinic, and aro-
matic material. The aromatic and olefin indication of all exhaust samples
was less than present in the starting fuels. A carbonyl at approximately
1730 cnr1 is also evident in the 3A and 3B samples, especially 3B.
The IR spectra of the two fuels show no significant differences. On the
basis of IR scans, the fuels can be considered to be the same.
Based on the limited number of exhaust samples analyzed, no other corre-
lation can be made regarding effect of power setting and fuel used.
Representative IR spectra are shown in Figures 49, 50, 51, and 52 for an
aromatic, fluoranthene; an aliphatic, undecane; the starting fuel and a ex-
haust sample (3A) showing the carbonyl.
Fuel Analysis
The fuel used in this test was a common aircraft turbine engine fuel
whose designation in Canada is JP-1 (Jet A-l in the U.S.).
The fuel was subjected to various types of analyses to ascertain sulfur
content, aromaticity, boiling point distribution and PAH content. The fuel
was found to contain 0.0065& sulfur by weight. The boiling point distribu-
tion was carried out by separating the paraffins from the olefins and aroma-
tics using ASTM method D-1319-70. This paraffin fraction and the complete
fuel were analyzed for boiling point distribution using ASTM method D-2887-
73. Analysis for PAH concentration was performed by A. D. Little, Inc. and
Radian Analytical Labortories, Inc. using gas chromatograph-mass spectrome-
tric techniques.
The results of the sulfur analyses are included in the section on sulfur
analyses of the exhaust. The analyses clearly establish that we were able to
dope the standard fuel successfully to get a high sulfur content fuel as
required.
107
-------
IR SCAN
SAMPLE: FLUORANTHENE
REF: AIR
CELL THICKNESS: 0.05mm
MICROMETERS
3000
2500
2000
1800
1600
1400
1200
1000
800
600
400
200
WAVENUMBER~cnf1
Figure 49. Infrared analysis of f luoranthene.
IRSCAN
SAMPLE: UNDECANE
REF: AIR
CELL THICKNESS: 0.05mm
MICROMETERS
6 7
3000
2500
2000
1800
1600
1400
1000
800
600
400
200
WAVENUMBER~cm-1
Figure 50. Infrared analysis of undecane.
108
-------
IR SCAN
SAMPLE: LOW SULFUR FUEL
REF: AIR
CELL THICKNESS: 0.05mm
3000
2500
2000
1800 1600 1400
WAVENUMBER ~cnT1
1000
800
600
400
200
Figure 51. Infrared analysis of low sulfur fuel
MICROMETERS
IR SCAN
SAMPLE: (3A) LOW SULFUR - CLIMB
REF: CS2
CELL THICKNESS: 0.5mm
4000
3600
3200 2800 2400 2000 1800 1600
WAVENUMBER ~cm"1
1400
1200
1000
800
600
Figure 52. Infrared analysis of low sulfur climb sample.
109
-------
The PAH content of the fuel was found to be very low with levels not de-
tectable (under 500 ppb) by A. D. Little, Inc. Radian Corp. also found up to
3000 ppb. However, analytical difficulties reported by them, make their re-
sults uncertain.
The boiling point distribution determination showed no significant dif-
ferences between the low sulfur fuel and the high sulfur fuel. The data is
presented in Tables 64, 65, 66, and 67. This distribution is shown graphic-
ally in Figures 53 and 54 for both low and high sulfur fuels.The ADL report
supplement #1 confirms their similarity and reports identical aromatic and
aliphatic content. In-house measurement of aromatic content by ASTM Method
D-1319-70 showed 19.9% aromatic and 0.3% olefinic content for both fuels.
NMR studies and ultraviolet scans of the fuels also gave identical results.
In summary, except for sulfur content, fuel A (low sulfur fuel) and fuel
B (high sulfur fuel) are identical with respect to aromatic, olefin, PAH con-
tent and boiling point distribution. Any differences in characteristics of
the exhaust must be associated with the sulfur content or other variables not
considered.
The fuel analysis provided by the EPA is as follows:
low sulfur fuel 84.16%C, 14.96%H
high sulfur fuel 84.01%C, 14.97%H
Both correspond to a H:C mole ratio of 2.12
Boiling Point Distribution
Samples were collected at the four power points using both low and high
sulfur fuels. These samples were extracted with hexane in the manner de-
scribed earlier and concentrated to a volume of 1 milliter. Half of each of
these four samples were processed in accordance with a procedure to isolate
the paraffins given in ASTM D-1319-70. These paraffin portions; the other
half of the mi Mi liter concentrated samples; and samples of the starting fuel
were then analyzed in the same way in accordance with the boiling point dis-
tribution determination by gas chromatography given in ASTM D-2287-73.
The analyses were carried out on a Hewlett Packard Model 7620A Gas Chro-
matograph with a flame ionization detector. The columnn was 1/8" 0.0. x 6'
stainless steel column packed with 0V 101 (methyl silicone) on Chromosorb
W-HP (flux calcined diatomite) which was temperature programmed at 6°C/
min. from 0°C to 350°C. The 0.5 ml samples were further concentrated to
25 ul before injection and one microliter of sample was injected in all cases.
For calibration purposes several known paraffins were chromatographed
under the same conditions as the samples. Results in Table 68 show the rela-
tionship between boiling point of the paraffin and column temperature at
which elution occurred.
110
-------
TABLE 64.
SAMPLE -
BOILING POINT
Low Sulfur Fue
DISTRIBUTION ANALYSIS
1 (Paraffin Fraction)
Retention Time,
Minutes
0220
02*7
02 18
037*
0*10
0*22
0**5
0*60
0*70
0*96
0513
0539
0563
OS 07
0630
06**
OK 76
0688
072*
0758
0792
0827
08*2
0895
0921
0939
1006
r
1036
107*
1098
112*
A J* fc ^T
11 51
117*
1192
1? 13
1265
1297
1715
1ft 1*
J>^ A~
1527
Percent By
Volume
T
0*021
0.00*
T
0.001
0.007
T
0.003
ft. 00*
0.008
0.001
0.097
0 .0 52
0.1*9
0.517
0.197
O.*l8
0.868
0.611
2.198
u Ȥ *6
*.732
2.638
* .1 *1
£»* 98
4.* 37
* .*71
5. « 5*
2.990
8.107
7. '2 68
17858
6.676
2.326
5'.* 71
3.856
3 '.7 30
3.2*3
f* 4RK f» ^\
0.002
0.1 00
0.020
Retention Time,
Minutes
1638
1693
17*6
1797
18*7
18 87
19*3
2037
2125
2291
2316
2**3
Percent By
Volume
O.D11
0.006
0.009
0.001
a. Dos
T
0.002
0.001
T
T
T
0.001
111
-------
TABLE 65. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - Low Sulfur Fuel (Complete)
Retention Time,
Minutes
0176
01 «2
C216
0248
02%
03 54
03 P4
C4 00
0439
0451
0479
0*56
C524
OH, 49
0620
C643
Oc.°0
C7 16
07*8
0774
OB 19
00 39
0878
0913
0^ 31
OS 70
0*^ 35
1329
in 49
10 P4
11 IH
1113
1140
11 *3
1230
1230
12*0
1236
12 «9
1308
13 31
I7 ^2
Percent By
Volume
0.001
.004
.367
.012
.002
.002
.003
.006
.003
.003
.003
.101
.052
.IPO
.635
.215
.156
1.018
.8 83
• v> V J
3.^20
1.326
5.457
S.712
2.952
8.897
6.3P8
3.426
6.627
3.073
9.342
2.226
4.656
2.9»3
3.200
2,441
3!l27
2.145
I .*6 40
0.8*7
2.327
2.461
0.8 ^9
0.533
0.778
Retention Time,
Minutes
HP3
14 H9
14*4
1525
15 96
1637
IF 54
17*1
1? *0
1* 38
1544
2037
2056
2124
2158
21 §2
2207
2241
2288
2316
-2366
2442
2516
2588
2740
3388
34 tO
36«>3
3609
Percent By
Vo 1 ume
0.437
1.30R
0.019
0.043
0.107
0.041
0.007
0.012
0.089
0.005
0.003
a. 003
0.006
O.DQ2
0.011
0.008
O.D07
T
0.002
T
0.002
0.005
0.001
T
T
f
i
T
1
T
I
T
T
112
-------
TABLE 66. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - High Sulfur Fuel (Paraffin Fraction)
Retention Time,
Minutes
0045
00*5
0137
01*4
0202
0229
0259
0274
cftrs
03 34
03 as
0189
04 S8
0412
0446
04 59
0474
OS 03
0531
05 56
0502
0626
0640
0673
0687
07 2»
0757
07 «2
OP 45
OM7
0922
0940
0978
ID 05
1036
1073
1097
1122
ilSO
1173
11 91
Jv J. J *
1213
1247
1265
12<57
Percent By
Vo I ume
T
6.005
0.010
5.023
&I612
0.028
D.01I
0.003
0.613
0.011
0.148
0.019
0.02*
0.002
0*009
0.006
0.022
0,1 SI
fl.O 80
0.213
• 0.776
0.308
0.629
1.175
O.P48
2.833
1.020
5 .6 96
7.273
7.041
4.693
*.640
S.t40
3.108
9.433
6.237
1 .'7 90
6.0 PI
2.2«6
4.9*8
4 • 1 Z-3
2. .010
3.251
2.87*
3.1 Pi
Retention Time,
Minutes
1314
1390
1417
1490
1531
15 *3
1641
1695
1747
1796
1846
1941
24^9
3253
Percent By
Vo 1 ume
3.747
3. 918
1 .6 86
0.316
0.2*4
0.050
0.063
0.027
5.023
0.011
0.012
d.Ooi
T
T
113
-------
TABLE 67. BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - High Sulfur Fuel (Complete)
Retention Time,
Minutes
009*
0197
0287
022*
C251
0286
0339
0371
G38*
0428
044*
0*65
0478
0489
0516
0530
0556
0577
0620
0541
0662
0*84
4^^» ^^
36 88
0733
0765
0793
OP 18
0»35
0855
0*92
0925
0943
0980
10 f?4
1036
1056
1070
11 10
ill9
1146
1169
1187
1204
1232
1244
Percent By
Vo 1 ume
T
0.001
3.006
3.011
0.7*3
0.019
0.003
0.006
0.013
0.015
0.021
0.001
0.008
0.007
0.006
0 .1 57
0.079
0.257
0.958
0.352
0.238
i:379
4 .4 *8
17629
6.226
K257
5. "4 3 3
3.246
9.037
7.125
3.905
6 ."1 *4
3 .1 $7
8.9*1
2 .2 82
4.617
2 ^7 77
2,859
2.352
4.969
2*740
1 V9 *0
1.286
0,635
Retention Time,
Minutes
1259
1252
IT 09
IS 34
1356
1364
1S?6
1412
1456
1470
I486
1526
1580
1636
1692
17*4
1845
1945
1986
2040
2056
2127
2210
2293
2370
3*07
3t08
3*!*
37*0
«(A M ^
3744
Percent By
Vo 1 ume
1.804
1 .8 77
0.633
0.416
0.349
0.236
0.372
D.763
0.042
0.065
0.233
0.241
0.105
0.1 01
OV035
0.018
0.014
0.002
T
0.002
0.002
0.001
0.001
0.001
0.001
T
1
t
0,004
0.004
114
-------
12 r
10
I
o
>
co
u
oc
ut
a.
12r-
10 -
8 -
6 -
4 -
2 -
COMPLETE
-, ,
RETENTION TIME (MIN.) °
BOILING POINT (°C)
,lll ll
PARAFFIN FRACTION
1 1 1 1 1 1 1 I
0 4 8 12 16 20 24 28 32 36
gg 106 144 178 215 250 287 322 358
Figure 53. Graphic representation of boiling point distribution for tables
64 and 65.
115
-------
10 -
8 -
uj 2
5
_j
O
>•
O
K
COMPLETE
10 -
8 -
6 -
4 -
2 -
RETENTION TIME (MIN.) 0
BOILING POINT (°C)
-
, .,ull
1
0 4
68
PARAFFIN FRACTION
1 1.
8 12 16 20 24 28 32 36
106 144 178 215 250 287 322 358
Figure 54. Graphic representation of boiling point distribution for tables
66 and 67, high sulfur fuel.
116
-------
TABLE 68. B.P. OF KNOWN PARAFFINS AND COLUMN TEMPERATURE AT ELUTION
Boiling Point Column Temperature
Compound OQ at Elution, °C
Hexane
Octadecane
Eicosane
68
304
343
24
184
203
Table 68 gives the elution times which correspond to degrees Celsius at
which the component is eluted and the relative abundance in percentages.
Table 69 (a thru p) also shows the total response for the sample. These abun-
dances are shown graphically for samples collected at idle and takeoff power
points only using both high and low sulfur fuels (Figures 55 through 58).
In all cases, hexane is excluded. The bulk of the components of the
paraffin samples eluted at column temperatures between 140 and 280°C. For
samples from which the paraffins were not removed, the range was 100 to
30QOC.
The fuel components eluted between 60°C to 140<>C for both the total
sample and the paraffin portion. (See Tables 52 through 55.) This fuel was
found on analysis by the ASTM D-1319-70 method to be 19.6% aromatics and 0.3%
olefins with the balance paraffins. Such a breakdown of the samples was not
possible because of the small amount of aromatic and olefins present compared
with the large amount of hexane, a paraffin, used as a solvent.
Sulfur Oxides Emissions
A determination of sulfur oxides emissions was made in two tests using
low sulfur Jet A-l fuel (ASTM D-1655-75) and Jet A-l doped to an approximate
0.25% sulfur concentration with ditertiary butyl disulfide. The engine was
run at four power settings using low sulfur fuel (idle, approach, climb,
takeoff) and three power settings using the doped high sulfur fuel (idle, ap-
proach climb). Takeoff power was unattainable during the high sulfur fuel
tests due to engine temperature limitations brought about by an unusually
high ambient temperature level.
Sulfur oxides were collected from the exhaust stream using the high vol-
ume linear sampling rake and mixing plenum. A 1/4" O.D. stainless steel emis-
sions line delivered samples from the plenum to the sulfur oxides absorption
train. The line was heated to 150°C.
117
-------
TABLE 69 (a). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 1A (Complete)
Retention Time,
Minutes
o*ni
0*15
0511
052*
05 61
0595
077*
07%
0856
10*3
1161
1169
118*
1231
1223
1237
1258
1277
1304
1326
1351
1* *2
1402
1461
1481
1517
15 46
1598
16*9
1S55
16*2
1706
1730
1751
1829
1*50
1937
2010
2050
21 10
2136
2155
21 98
2230
227*
2322
Percent By
Volume
0.111
0.070
0.108
I.* 19
0.657
0.531
0.126
0.121
0.039
0.026
0.053
0.030
0.115
0.2*7
5.971
0.5*7
0.29*
0.707
2.* 76
1.892
1.TI9
2.41*
1.'5**
8.1*7
1.513
2 .'1 28
0.!>*2
5. 855
0.028
0.260
0.0**
0.107
0.3*5
0.719
0.213
0.256
39.5*7
0.562
0.60*
0.363
0.2*9
0.065
l.*29
l.*37
0.722
0.6*9
Retention Time,
Minutes
23*6
2392
2*20
2*66
2*92
2515
2552
2616
2650
2678
2742
2*03
2863
2920
2973
3002
3031
30**
3132
31*6
3290
3*18
33*0
3362
360*
3*8*
3618
3652
Percent By
Vo 1 ume
0 ."5 78
0.305
2.097
1.2*8
0.817
0.03*
0.*79
2.702
0.326
0.163
1.078
H821
0.162
0.7*6
1.6*9
0.123
0.201
0.967
1.534
0.680
0.239
0.032
0*061
0.151
0.25*
0.001
0 .0 36
0.001
118
-------
TABLE 69 (b). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 1A (Paraffin Fraction)
Retention Time,
Minutes
U«»28
Oft 61
0? 31
CS pg
C621
OF «6
07 «8
C»?3
p° 84
09 r>*
OM8
0975
10 27
1351
11 16
11 **
1193
12*5
0258
1293
1370
13 R7
It 1*
1* "8
15 ?8
1586
15 99
IP 37
IK 51
17 ft*
17*2
J7 Qij
|p £2
1857
1537
1992
20!?8
2116
21 78
2210
22 '7
2256
2290
2316
Percent By
Vo 1 ume
0.107
0 .09*
0.093
0.011
0.015
0.007
o.bo*
J.I 05
0.0 U
0.090
0."012
0.0*5
0.059
0.176
0.008
0.086
0.170
0.0*8
O.DS9
0.017
0.151
3. OS*
0.036
0.1»7
0.165
0.2R3
0.103
0.069
0.2 86
0.23*
0.082
0.59U
0.265
0.730
0.75S
1 .1 89
0.957
1.829
t.532
3 .9 *5
3.856
2.372
2.036
5.5 63
2.355
Retention Time,
Minutes
2335
2358
2390
2*10
2*33
2*6*
2*75
2505
253*
25*9
2*75
2* no
25*2
2654
2707
2727
2770
28 30
2850
29*9
2969
3056
30 R2
3*78
3*78
Percent By
Vo I ume
2.103
5.681
3.076
1.933
5.369
2.* 18
2.193
4.860
2.792
i.536
3.7fll
2.302
*.079
*.112
2.*38
2.892
3.969
3.717
3.027
1.159
0.8 87
1.290
0.*B3
3.3
0.036
119
-------
TABLE 69 (c). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 2A (Complete)
Retention Time,
Minutes
ION TTHE
0*18
1099
1161
1190
12*7
1260
me
1373
i*fu
11*2
U65
1*7*
IS Q5
1533
1533
16*5
1675
1696
1731
1753
1793
1853
IP 75
1918
1957
1998
20^5
2061
2081
2111
31 £5
2223
2257
22 «U
2299
2321
2*53
2378
2*29
2*33
2*99
2527
237*
2597
Percent By
Vo I ume
o;o
0.035
0.125
0*008
0.009
a. 03*
0.010
0.033
0.012
0.00*
0.013
0.0*5
0.0*7
0.073
0.032
a. ooi
0.0*7
0.325
0.079
0.007
0.091
0.025
0.177
0.263
0.1*4
11 .6*2
0.035
0.165
3,057
0.028
0.728
0.*72
3.809
0.**7
0.373
2.* 90
0.265
1.525
3.* 52
0.577
3.S71
1.155
1.5*2
0. 90S
0.506
Retention Time,
Minutes
26*3
2696
2755
2812
28*6
2871
2927
297*
3033
3086
3126
3*2*
3*3*
3640
Percent By
Vo 1 ume
36.990
1.268
1 .1 57
5.ft69
O.tte
0.873
0.792
6.989
0.6*9
0.3*3
1.833
T
0.015
T
120
-------
TABLE 69 (d). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 2A (Paraffin Fraction)
Retention Time,
Minutes
0*14
CP 66
0078
P*> 52
flQ 59
1025
1062
11 15
1153
11 «3
12f!2
12*6
1258
1273
1?93
i*73
I7 90
1*16
1*77
la <>2
1532
1"> "6
1C 05
1C *2
1673
1696
1711
1749
1775
1801
IP 16
in 47
1«03
2015
2123
21 «8
25m
22 «
2266
2292
2320
Percent By
Volume
0.070
0.010
0.020
0.002
0.0 OP
0.081
D.D03
0.B95
0.108
0.029
0.004
0 .1 $8
0.0*1
0,015
0.209
0.159
0.086
0.229
0.037
o.lfs
0.297
0.17*
0.178
3.177
0.080
0 «£ 53
a. 2 52
i.612
a. 323
1.156
0 .7 36
2.151
2. 3 3D
2.5»2
3.2*0
2.289
* .8 70
5.865
5.332
3.7*6
i.818
ft «* 82
4.653
Retention Time,
Minutes
2346
2373
2*07
2*51
2*79
2525
2554
2621
2653
2755
3296
3*39
Percent By
Volume
2. $96
8.084
1.448
12.515
0.18*
4.583
2.970
3.8 «2
^^ V ^ J 1 V ^^
2.525
3.002
1.035
0.008
0.001
121
-------
TABLE 69 (e). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 3A (Complete)
Retention Time,
Minutes
0539
05*7
0550
0655
1*72
1509
1532
1579
16*0
1570
1651
17 ns
1728
1750
1*01
W52
1*77
1907
1921
1951
1^61
20**
20%
21*2
2136
2250
2230
2278
2300
2327
2*52
237*
2*92
2*50
2*56
2518
2570
266*
26 «8
2752
2? 10
2* *6
Percent By
Vo 1 ume
0710S
0.1 51
070 f 5
3.001
HO 89
0.271
0.136
dtOft*
0.316
0.613
0.2S1
a.**7
0.126
0.419
0.091
0.669
0.596
0.190
0.365
17024
1 ."1 46
1.221
0.105
2 .* 45
0.810
3.051
0 .9 «1
07767
0.337
0.030
110 88
O.B47
0.061
10.300
0.638
O.S22
IT***
174 «6
17313
2.92*
17. 2 «3
0.958
Retention Time,
Minutes
2872
29*0
2980
30*0
309*
3r§6
32 SO
3298
33 «2
3*7*
5418
Percent By
Vo I ume
2.621
2.463
16.623
2:059
27133
*7598
2.117
079 70
1-.8 80
17***
0.959
17333
122
-------
TABLE 69 (f). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 3A (Paraffin Fraction)
Retention Time,
Minutes
0452
0*59
OS 17
OP 64
1233
1*33
1479
IS2J
1629
16 *1
1698
1735
1755
18 32
1891
1926
2018
2134
21 P8
22*4
2268
2500
2346
2378
2*20
2*52
2492
2562
2628
2694
2756
2*16
28*0
2942
3066
3124
3338
3428
Percent By Retention Time, Percent By
Volume Minutes Volume
0.112
a;o 32
8.0*9
9.030
0.027
0.009
0.003
a-;t§5
0 .3 *7
0.552
0.056
».tl2
0.142
1.810
U234
2 .'7 64
2.530
6*995
14.609
6.295
0.378
4;G57
5.533
4.925
77759
3.953
6.434
6 .1 £e
9.3§0
5. SOI
3.593
2.521
0,915
0.425
0.009
O'.OIS
0.012
0.003
123
-------
TABLE 69 (g). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 4A (Complete)
Retention Time,
Minutes
Otio
C*23
11 «?9
1366
1*67
1*97
1505
1529
15 «0
IS 19
16*1
16 tl
1653
1709
1730
1752
17 S7
1773
1798
1ft ?1
1951
1870
1903
16 58
1996
2016
20*0
2060
20 80
2033
21 32
2152
22156
2228
227ft
2360
2ft 46
2ft 90
2S 16
25 SO
26 30
2662
26 ?6
2722
27*8
Percent By
Vo 1 ume
0.065
0.0*5
0.002
0.021
0.893
0.072
0.0*3
tr.'osi
6.076
0.152
0.3 33
3.5C6
0.278
0 .5 99
0.239
truss
8 »T to
0*%7S
0 .**8
d.461
8'«i71
tfv^iik 4 M
vr^^vk j^ i*
n*n
5.511
0.985
1.315
0.697
0.67S
0.8*2
1.527
0.2*5
0.996
2.6»5
J.518
3.321
Iff. 9 59
0 ifilS
0.108
5 .1 31
8 .3 19-
0.706
1 .0 62
0.075
6.403
Retention Time,
Minutes
2«n8
2841
2868
292*
297*
3032
3086
3130
31 »8
324*
3278
3328
3*08
3*52
Percent By
Vo 1 ume
13.9*7
3.560
i.t 7*
2 .8 72
1X7206
1.19*
1.278
2.906
0.879
0.276
0.888
0.520
0.212
0.723
0.001
124
-------
TABLE 69 (h). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 4A (Paraffin Fraction)
Retention Time,
Minutes
0477
0544
OR 36
C698
P749
Oft 74
n» »7
0920
0978
0995
1034
1! ^2
1170
12*7
15 35
i607
A^-^ ^v V
-16 45
1*70
1698
1751
A * •* •
IP 52
1907
2018
20*2
2216
2298
2t^
26 18
•»* ' ft ^^
^^ -o
^r -»y
29 10
1968
29 95
3028
Percent By Retention Time, Percent By
Volume Minutes Volume
0.100 3090 1.1*2
fi257 3216 0.135
0.156 3330 0*010
0.100 3546 0.322
6.210 36 «1 0.042
0.2*9 i?!2 • 0.145
0.197
0.39S
0 «* 47
0.031
0.758
0.862
0.457
0.363
0.457
O.S02
0.1.1*
0.210
1 ;3 40
0 .1 4S
3.323
2 .fe09
0.509
3.001
3. ft 63
3 A 98
l."S92
A " 'i£- & ^
1.516
7.3.63
6.03%
6.699
8^952
12.670
9.970
2.119
4.912
3.396
1.4 23
2.P04
125
-------
TABLE 69 (i). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - IB (Complete)
Retention Time,
Minutes
1087
1465
159*
16 40
17 OQ
1732
17 S7
1777
W10
1ft 55
WIT
1963
2019
2043
2067
21 35
2167
2227
2^.05
2327
2? 53
2451
2480
25 GS
2523
2547
26*1
2941
2*73
2935
3291
3339
3*17
Percent By Retention Time, Percent By
Volume Minutes Volume
0.035
0.207
0.069
5.392
0.207
3.180
0.10*
1.5*8
0.069
2.90*
Kl 06
18.977
0.518
1.797
1.521
0.933
3.422
7 ;3 97
176 59
OT760
19.564
8.*34
0.933
2.627
4 .7 70
0.622
3.802
1:521
5 i346
2.523
1.'936
1 .2 79
b.035
126
-------
TABLE 69 (3). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - IB (Paraffin Fraction)
Retention Time,
Minutes
0723
1023
1048
1057
1112
1117
1144
1160
lite
12*1
i?«8
1*59
1**6
1ft IS
14 52
1533
15*5
1699
1751
1*03
1851
19 46
2002
2018
20%
2126
21 *8
2212
2258
2202
2328
2346
2*72
2434
2422
2446
2478
2494
2520
2550
2590
26 18
2*. SO
Percent By
Volume
0.004
o;o*7
0.020
0.045
0.167
0.006
0.004
0.169
0.842
0.261
0.149
0.12s
0.958
0.510
0.193
0.291
0.428
O.S70
0.972
0.748
1 .4 08
0.756
0.498
1 .5 38
1.4«3
1.968
3.976
4.398
3.&40
5.963
1.277
2.116
5.66W
2.435
1.439
5.665
i ;'9 79
1.775
4.298
4.074
3.299
3.993
4.928
Retention Time,
Minutes
2611
2726
2788
2846
2910
2966
30 ?6
30 «2
3114
3194
3236
3458
3688
3722
Percent By
Volume
3.096
5 .'4 25
5.023
3. 7*9
2.965
1.605
1 . ? 56
0.4*1
0.326
d.149
0.018
0.001
0.018
0.008
127
-------
TABLE 69 (k). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 2B (Complete)
Retention Time,
Minutes
0402
0*16
1377
1*13
15 QO
1532
1579
16*2
IP 72
16 91
1715
1*52
18 *3
1922
19 SO
20*4
2Q62
2150
216%
2228
22 SO
22%
2*20
2*50
2? 78
2448
2492
2522
2642
2F66
2754
27 96
2S22
2928
299*
3038
3254
3358
3J98
34 ^6
3450
3*12
3678
3706
Percent By Retention Time, Percent By
Volume Minutes Volume
0.424
0.142
0.617
0.0*5
1.029
0 .006
i;o 90
0.224
0.5R9
n.'ois
0.206
D.T841
0.436
171 20
el. 168
0.454
0.278
0 .5 93
0 .8 84
6 .562
i;i44
i ;i *s
1.816
i;si3
r;o59
t;iiBi
0.4 §4
d .0 91
2.264
0.006
0.012
0.430
1,1047
0.4.60
1 .701
i;2«is
0.266
0.073
Oi248
0.006
0.006
0.012
0.224
0.067
128
-------
TABLE 69 (I). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 2B (Paraffin Fraction)
Retention Time,
Minutes
0967
1!*11
1433
1525
15 *3
1632
1637
1732
1791
WOO
W03
1939
2032
2051
2120
2205
2261
22 *7
2320
2341
2365
23 «9
2418
2440
2472
24 «5
2514
2543
25 «
2610
2651
27 US
2ft 40
3118
3615
37 !53
Percent By Retention Time,. Percent By
Volume Minutes Volume
0.006
0.104
O.D43
0 .4 48
0.021
1.400
0.457
0 .1 10
0.274
2.062
0 .9 62
1.4P6
0.618
0.402
3.572
12.711
4.492
7.46*
3.944
2.1 f»3
7 .4 58
3. 4 S3
3.201
10.250
0.850
7.552
8.006
5.016
6.SfiS
1.270
0*691
1.880
0.292
0.012
0.012
0.015
129
-------
TABLE 69 (m). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 3B (Complete)
Retention Time,
Minutes
0766
0775
08 «1
1249
12%
1*51
1380
1480
15 96
1554
1596
1617
1624
1C 45
1*577
1707
1753
1ft 52
1880
W 18
1970
20^8
2058
2094
2128
21 §8
2?f8
2248
2270
2? 52
2*18
2^ 44
2170
2* 90
2442
2492
2516
2566
2586
2C40
2658
Percent By Retention Time, Percent By
Volume Minutes Volume
0.003 26*2 2;537
0.001 2744 ».S«6
&.D02 24 *6 1.517
0.225 2K40 2.617
0.001 3OO 0.008
0.002 3465 T
9.D19
D.001
3.019
0.012
0.016
0.001
0.002
0'.022
0*0 11
D.041
O.D29
0.049
0.025
0."2 26
73.353
0.670
0.4 AQ
0.498
9.676
0 .6 46
173 50
0.644
0.453
0 .8 42
0.547
O.t45
0^730
i ;o 59
2.292
1 .1 «0
171 62
1.469
d.894
1.700
0.637
130
-------
TABLE 69 (n). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 3B (Paraffin Fraction)
Retention Time,
Minutes
Percent By
Volume
Retention Time,
Minutes
Percent By
Vo I ume
0398
0*11
0511
11 73
1153
1*51
1539
16*7
1701
1752
IP 22
1«51
1913
IS 47
1959
20*8
21 25
2208
2246
2299
2? 68
24*3
2965
3090
0.049
0.130
0.3*7
0.091
0.046
0.011
0.103
0 .6 78
cram
0.613
0.003
17120
0.908
0.504
3.752
0.160
0.9 44
3.300
3.32*
3.4*6
1.754
78 .* 86
0.011
O.Til
131
-------
TABLE 69 (o). BOILING POINT DISTRIBUTION ANALYSIS
SAMPLE - BP - 4B (Complete)
Retention Time,
Minutes
It 63
1453
1463
1486
1627
1657
It 11
1733
1759
1788
Ifl 33
1ft 57
1« 96
1938
2016
2060
20 42
2106
21*2
22 CO
2242
22 «*6
2M4
2*42
2? 68
2446
2494
2520
2550
2578
26 ^4
26 «8
2714
2752
2*10
2*68
2924
2974
30 SO
30 «6
3126
31 «6
Percent By
Volume
0.042
0.021
0.032
0.045
0.009
0.037
0.054
0.129
a a 01
0.004
0.097
0.036
0.292
64 .8 58
0.046
0.005
0.016
0.179
0.0 ftl
0.951
o a *o
1.196
0.122
1.214
1.243
3.M9
0.361
0.566
0.357
Ii217
4 10 32
D.194
0.330
1.200
2.741
0.615
1'iSOS
4.307
0 «i» 4l
i a 19
2.702
0.663
Retention Time,
Minutes
32 Q5
32*8
3276
3292
3*58
3446
3628
3654
37flO
-1
Percent By
Volume
3.420
O.S58
0.698
0.597
0.093
0.001
0.006
0.003
0.005
132
-------
TABLE 69 (p).
SAMPLE
BOILING POINT DISTRIBUTION ANALYSIS
- BP - 4B (Paraffin Fraction)
Retention Time,
Minutes
Percent By
Volume
Retention Time,
Minutes
Percent By
VoIume
16*0
1450
1910
1950
20 56
2iB8
2130
2210
22 £6
2292
2720
2170
2»*6
2518
25 **
2596
31 H
3286
3*50*
3*3*
3378
3406
3*58
3.228
t;2l5
3.131
l."392
S.266
0.116
2.0*9
IT."025
1*392
6.T39
1 .'» 56
3.987
7.405
2.722
0.656
i;S92
D.696
2.5 95
3.038
18.101
6 .3 92
3.291
0.063
133
-------
10
i
§
>
to
Ul
u
E
10
RETENTION TIME (MIN.) 0
BOILING POINT (°C)
39.587
COMPLETE
.Illll
1.1.
PARAFFIN FRACTION
8
106
I
12
144
16
178
20
215
24
250
I
28
287
32
322
36
358
Figure 55. Graphic representation of boiling point distribution for tables
69 (a and b). idle power point (1A) using low sulfur fuel.
134
-------
12
10
13.947
ai
o
m
t- 12r-
LU
U
tr
Ul
a.
lOf-
8 -
6 -
4 -
2 -
COMPLETE
RETENTION TIME (MIN.) 0
BOILING POINT (°C)
-
) 4
68
.
PARAFFIN FRACTION
. , ,.lll ll ll I.J.
iii
8 12 16
106 144 178
12.670
i
I
4
t
9
. . I
20 24 28 32 36
215 250 287 322 358
Figure 56. Graphic representation of boiling point distribution for tables
69 (g and h) takeoff power point (4A) using low sulfur fuel.
135
-------
12
10
_!
o
18.977 19.564
COMPLETE
12
Ul
U
K
111
Q.
10
RETENTION TIME (MIN.) 0
BOILING POINT (°C)
4
68
PARAFFIN FRACTION
8
106
12
144
16
178
20
215
24
250
28
287
32
322
36
358
Figure 57. Graphic representation of boiling point distribution for tables
69 (i and j) idle power point (IB) using high sulfur fuel.
136
-------
12
10
UJ
3
64.858
ED
H 12 r-
ul
u
cc
UJ
Q.
ioH
8 -
6 -
4 -
2 -
COMPLETE
RETENTION TIME WIN.)
BOILING POINT (°C)
PARAFFIN FRACTION
—
-
-
1 1 1 I
0 4 8 12 16
68 106 144 178
' /.uia
I
gt
I
1
1
1 1
20 24
215 250
|
H
t
1 1
28 32 36
287 322 358
Figure 58. Graphic representation of boiling point distribution for tables
69 (o and p) takeoff power point (4B) using high sulfur fuel.
137
-------
The absorption train used was similar to that described in method 8 of
the Federal Register, "Standards of Performance for Stationary Sources", June
8, 1976. The absorption train consisted of three fritted absorbers in series
followed by a dry test meter to measure volume. This is an exception to the
Federal Register which recommends impingers. Fritted absorbers were used to
improve collection efficiency.
The first absorber contained 15 ml 80% isopropanol for $03 collec-
tion. The second and third absorbers contained 15 ml 3% hydrogen peroxide
each to absorb S02- The Federal Register recommended particulate filter
between the 503 and S02 absorbers was not used because it was found that
condensation and subsequent loss of sample could occur with this system.
An attempt was made to sample with a quartz sample probe with six 0.030"
holes drilled along its length at centroids of equal area. The probe was en-
cased in a stainless steel sheath. A ceramic separator was used to cushion
the quartz. It was hoped that this sampling scheme would provide a compari-
son to the use of stainless steel probes. The sample line for this system
was heated teflon. When used, however, the probe was unable to withstand the
thermal shocks encountered in rapid power setting changes. For this reason,
only the stainless steel linear rake was used.
In all sampling, a rapid bypass system for sample flow was used to de-
crease residence time and insure a more representative sample. Line adsorp-
tions were thus minimized.
The barium chloranilate method was used for analysis of sulfur oxides.
Test samples were transferred to polyethylene containers and frozen in dry
ice until analysis. These samples were reduced to approximate volumes by
evaporative heating. The method details are given in ASTM D-3226-73T.
Samples of both low sulfur and high sulfur Jet A-l fuel were analyzed
for sulfur content. A spot check of the high sulfur fuel was made in Canada
to assure that an approximate 0.3% sulfur concentration was achieved. Fuel
analysis data is presented in Table 70. Table 71 compares the calculated
fuel sulfur concentration based on emission measurements with the actual fuel
analysis shown in Table 71. Sulfur oxides are given as percent sulfur.
TABLE 70. FUEL ANALYSIS (percent S)
Sample P&WA Canada P&WA-U.S.
Low Sulfur 0.006
0.007
High Sulfur 0.2500 0.2600
0.2500
138
-------
TABLE 71. PERCENT SULFUR IN FUEL BASED ON EMISSION
MEASUREMENTS
Jet A-l Low Sulfur Fue
% Sulfur in Fuel
Based on S02
% Sulfur in Fuel
Total % Sulfur
in Fuel Based
Power Setting
Idle
Approach
Climb
Take-Off
Emission
No S02 detected
No S02 detected
0.0035
0.0179
Emission
No 503 detected
No $03 detected
No $03 detected
No 503 detected
• • • • *^«^ V V VI ^ \H» «*•
Emission
No SOX detected
No SOX detected
0.0035
0.0179
Idle
Approach
Climb
Jet A-l High Sulfur Fuel
0.2279 0.0104 0.2383
0.2769 No SOX detected 0.2769
0.2638 0.0294 0.2932
A good material balance was achieved between fuel bound sulfur and ex-
haust samples at all power settings. On an average, a relative error of 6%
exists between the fuel sulfur concentration determined in the analysis of
the high sulfur fuel and the fuel sulfur concentration calculated from mea-
sured gaseous SOX emissions. The overall limitations of the sampling meth-
od are obvious in the low sulfur runs where adsorption losses can easily ac-
count for the absence of sulfur at low power. The high sulfur sample runs do
not suffer this limitation and essentially all the sulfur expected was
detected.
Proton Activation Analysis/X-Ray Analysis
Nuclepore filters were used to collect particulates while the engine was
operated at four power settings using high and low sulfur fuels. Repeat sam-
ples were taken at four power settings using low sulfur fuel making a total
of twelve exposed filters.
These twelve Nuclepore filters were submitted to the EPA for proton ac-
tivation analysis (PAA) which was carried out at the Florida State University
(FSU) Physics Department. This type of analysis resulted in an elemental as-
say of the particulates adhering to the filter material. Results of this
139
-------
analysis are shown in Table 72. The data from FSU is reported on the basis
of concentration with respect to unit area. The data supplied by FSU was
further reduced to reconcile the concentration of the various elements with
respect to exhaust gas flow, and is given in Table 73. A clean Nuclepore fil-
ter was analyzed by PAA as a blank and the resulting data was subtracted from
the particulate data before the data was reduced as described above.
Mitex filters were used to collect particulate material while the engine
was operating at four power settings and using high and low sulfur fuel. One
repeat sample was taken at the take-off power condition using low sulfur fuel
to give a total of nine particulate laden filters. These nine filters were
sent to the EPA at Research Triangle Park where the surfaces were analyzed
using X-ray techniques. The data, reported in concentrations per unit area,
was reduced to reflect concentrations in terms of nanograms of each element
per m^ flow. These results are given in Table 74.
TABLE 72. PROTON ACTIVATION ANALYSIS DATA FROM FSU REPORT
Run #
Al
Si
Cl
Ca Cr
Fe
Ni
Cu Zn Br Pb
Climb 64
Climb 72
Approach 67
Approach 80
Take-off 70
Take-off 83
Idle 74
Idle 78
Approach 84
High-Sulfur
Take-off 85
High Sulfur
Climb 86
High Sulfur
Idle 87
High Sulfur
274
60 221
33 240
235
339
305
201
96 303
309
341
89 555
64 330
80
56
99
61
80
119
34
89
90
108
173
52
389
369
271
244
400
493
158
2083
1144
16453
10923
483
353
428
269
195
487
521
162
184
436
867
351
196
38
10
_
-
24
-
14
12
28
49
27
58
75
16
-
61
29
27
9
51
57
86
36
172
88
39
25
125
159
4
-
150
349
107
58
453
437
184
48
309
449
51
18
921
1178
327
153
397
152
94
35
210
330
29
15
306
803
328
115
56
34
17
10
39
49
9
8
52
86
30
15
83
83
39
9
38
65
12
82
109
65
25
16
16
14
33
23
24
15
23
112
35
40
27
37
39
10
25
32
34
16
14
57
56
37
16
Blank
Detection
Limits
77 49 41 31 30 25 21 11
6554 2 12
NOTE:
1. All amounts are in ng/cm2.
2. Blank values have not been subtracted.
3. Sulfur values checked with PESA (Proton Elastic Scattering Analysis)
140
-------
TABLE 73. PROTON ACTIVATION ANALYSES CONCENTRATIONS. ng/ra3
Sample
1A
1A
2A
2A
3A
3A
4A
4A
IB
2B
3B
4B
Flow
m3 Al
5.67 -
6.14 -
3.66 -
3.15 -
1.69 -
1.73 -
1.59 -
1.42 -
3.64 -
2.82 -
1.63 -
1.64 -
Si
75
39
37
186
26
405
352
164
176
1067
399
P
11
32
12
13
154
0
25
302
105
S
20
58
185
172
756
680
839
1271
481
1813
2970
4660
Cl K
63 12
77
246
167
768 114
987 29
1263 76
1535
146 38
621 22
809 152
2398 86
^^"**
-------
In the case of PAA, generally higher levels of all elements were found
in samples from the high sulfur fuels than from samples from the low sulfur
fuels at comparable engine power settings. There is some suggestion based on
PAA data that engine wear increases with the high sulfur fuel. Furthermore,
levels generally increased with' increasing power.
The trend with X-ray analyses is less regular but many elements show a
similar variation with power setting and sulfur content of fuel. The results
from PAA and X-ray analysis are not directly comparable because of fundamen-
tal differences in the analytical technique and in the sampling method.
In X-ray analysis, examination of the upper surface only is involved,
whereas in PAA, the total sample is analyzed. Uniformity between the sample
on the surface and beneath the surface of the filter cannot be assumed. ..Uni-
formity of the collected sample cannot be assumed and may explain differences
between replicate samples as well as between the PAA and X-ray analysis. Sur-
face characteristics of Nuclepore and Mitex filters are different. Non-uni-
formity of the filter material would affect the uniformity of the sample. The
X-ray analysis samples were taken on a large (293 mm dia.) Mitex filter and a
47 mm circle was cut and submitted for analysis. The PAA samples were taken
on a pre-cut 40 mm Nuclepore filter. Differences in collection efficiency of
these filter materials are probable. Possible uncertainties in sample flow
measurements would also contribute to differences between X-ray and PAA data
since the quantitative information is based on calculated mass accumulation
from flow data. There is some suggestion based on PAA data that engine wear
increases with the high sulfur fuel.
Elemental Analysis
Thirteen samples were examined for carbon, hydrogen, sulfur, oxygen, and
nitrogen using traditional combustion analyzer techniques. These particulate
samples were collected on 293 mm Gelman type A/E glass fiber filters using
the high volume sampling system. The filter surface temperature was limited
to approximately 160°F maximum and samples were taken at each of four power
settings (idle, approach, climb and take-off) using both low sulfur and high
sulfur fuel.
Samples were prepared for analysis by first separating the particulate
matter from most of the fiber backing and desiccating the material to remove
entrained water. Heat was avoided to preserve the integrity of the more vol-
atile organic fractions. The samples were homogenized using a mixer mill and
combusted in the analyzer. Table 75 lists the results of these analyses.
The data listed in Table 75 show little correlation with expected re-
sults and no correlation with power setting or fuel change. Due to the limit-
ations of this particular analytical procedure, it is probable that insuffi-
cient material was available to accomplish a successful analysis and that a
different sampling scheme would be necessary to obtain more material from the
exhaust stream. Two milligrams of organic fraction would be the minimum sam-
ple required for any such sampling system design.
142
-------
TABLE 75. ELEMENTAL ANALYSES
Sample
1A il
1A n
2A #1
2A #2
3A f 1
3A #2
4A #1
4A n
4A #3
IB
2B
3B
4B
Power
Setting
Idle
Idle
Approach
Approach
Climb
Climb
Take-off
Take-off
Take-off
Idle
Approach
Climb
Take-off
Total
Flow.m3
49.6
20.4
31.5
29.4
42.3
42.0
32.2
43.0
43.7
88.0
42.5
42.2
47.2
Filter
Temp . ,
OF
148
142
108
163
128
116
136
144
131
154
150
142
145
%C
89.3
94.0
86.9
95.2
97.2
91.6
95.9
97.5
96.8
70.2
77.5
72.3
50.3
%H
1.40
1.41
2.69
1.38
0.82
1.92
1.07
0.65
0.65
1.65
1.70
1.63
2.87
XS
0.18
0.04
0.23
1.0
0.01
0.18
0.29
0.26
0.04
0.26
0.32
0.11
0.18
%0
4.67
4.54
10.1
3.22
1.93
6.33
2.72
1.56
2.48
27.9
20.5
26.0
46.7
%N
1
1
1
1
1
1
1
1
1
1
1
1
1
143
-------
REFERENCES
1. Fenton, D. L., "Turbine Engine Partial I ate Samples: Design Study",
Report SAM-TR-76-1, 1975.
2. Particulate Polycyclic Organic Matter", edited by National Academy of
Sciences.
3. Brown, R. A., et a I, "Rapid Methods of Analysis for Trace Quantities
of Polynuclear Aromatic Hydrocarbons and Phenols in Automobile Ex-
haust, Gasoline, and Crankcare Oil", Final Report, Feb. 1969 - Dec.
1971.
4. Doran, T., and McTaggart, N. G., "The Combined Uses of High Efficiency
for the Determination of PAH in Automobile Exhaust Condensates and
Other Hydrocarbon Mixtures", J. of Gas Chrom. Sci., Vol. 12, Nov.
1974, pp. 715-721.
i-
5. Foster, J. F., et al, "Chemical and Physical Characterization, of Auto-
mobile Exhaust Particulate Matter in the Atmosphere", NTIS Report. No.
CRC-APRAC-CAPE la-70-1, Oct. 1971.
6. Long, R., "Studies on PAH in Flames", EPA R-3-72-020, July 1972.
7. Brown, R. A., et. al., "Progress in Development of Rapid Methods of
Analysis For Trace Quantities of PAH in Automobile Exhaust", CRC-APRAC
Proj. CAP-12-68, 1960.
8. Mentser, H., and Shockey, A. G.,Jr., "Chemical Characterization of
Diesel Exhaust Particu lates", PERC/RI-77/5.
9. Conkle, J. P., et al, "Hydrocarbon Constituents of T-56 Combustor Ex-
haust", SAM-TR-75-8.
10. "The Determination of PAH in Airborne Particulate Matter", Coke Re-
search Report, British Coke Research Assoc., March 1973.
11. Golden, C., and Sawicki, E., "Ultrasonic Extraction of Total Particu-
late Aromatic Hydrocarbons From Airborne Particles at Room Tempera-
ture", Intl. J. Environmental Anal. Chem., Vol. 4, 1975, pp. 9-23.
12. Sawicki, E., et al, "Tentative Method of Microanalysis for Benzo (a)
Pyrine in Airborne Particulates and Source Effluents", Health Lab.
Sci., Vol. 7, (Supp. Jan. 1972) pp. 56-59.
144
-------
13. Sawicki, E., et al, "Tentative Method of Analysis PAH Content of At-
mospheric Particulate Matter", Health Lab. Sci., Vol. 7, Supplement,
Jan. 1970, pp. 31-44.
14. Pellizzari, E. E., "Analysis of Organic Air Pollutants by Gas Chromat-
ography and Non Spectroscopy", EPA Rept. 600/2-77-100, July 1977.
15. Pellizzari, E. E., "Development of Method For Carcinogenic Vapor Anal-
ysis in Ambient Atmospheres", NTIS Report, EPA 650/2-74-121, July 1974.
16. Luedecke, E., "Analysis of Benzo(a) Pyrene in Airborne Partiallates by
Gas Chromatography", NASA TMX-71872, Feb. 1976.
17. Palfroman, J. F., McNab. J., and Crosby, N. T., "Evaluation of Alkali
Flame lonization Detector and the Coulson Electrolytic Conductivity
Detector in the Analysis of N-Nitrosamines in Foods", Journal of Chro-
matography, 76 (1973), pp. 307-319.
18. Riedmann, M., "Gas Chromatographic Detection of Nitrosamines in Foods
With Nitrogen Flame lonization Detector", Journal of Chromatography,
88 (1974) pp. 376-380.
19. Fine, D. H., et al, "N-Nitroso compounds: Detection in Ambient Air",
Science, Vol. 192, June 25, 1976, pp. 1328-1330.
20. Bogovski, P., and Walker, E. A., "N-Nitros Compounds in the Environ-
ment", 1 ARC Scientific Publications, No. 9, 1973.
21. "Food Ingredients - Nitrates and Nitrites (Including Nitrosamines)",
Battelie-Columbus Labs., NTIS Report FDABF-GRAS-20, Sept. 1972.
22. Slusher, G. R., "Sulfur Oxide Measurement in Aircraft Engine Ex-
haust", FAA-NA-75-10, (1975).
23. Bergman, F. J. et al. "Measurements of Atmospheric Sulfates",
EPA/600/4-76-015.
24. Dong, M., Locke, D. and Ferrand, E., "High Pressure Liquid Chromato-
graph Method for Routine Analysis of Major Parent PAH in Suspended
Particulate Matter", Analytical Chem., Vol. 48, pp. 368-372, Feb. 1976.
25. Ringwall, C. G., "Compact Sampling System for Collection of Particu-
lates From Stationary Source", EPA-650/2-74-029.
26. Conkle, J. P. et al, "Cryogenic Sampling of Turbine Engine Exhaust",
SAM-TR-74-54.
27. Hermann, T. S., "Development of Sampling Procedures for Polycyclic
Organic Matter and Polych lorinated Biphenyls", EPA Report: EPA
650/2-25/007, Aug. 1974.
145
-------
28. Jones, P. W. eta I, "Efficient Collection of Polycyclic Organic
Compounds From Combustion Effluents", Battelle Labs. (Columbus, Ohio),
Aug. 1975.
29. "Analysis of the Odorous Compounds in Diesel Engine Exhaust", Final
Report to CRC and EPA: A. D. Little, Inc., CRC'Proj. CAPE-7-68,
A.D.L. No. 73686-5, (1972).
30. "ASME Fluid Meters - Their Theory and Application", Report of ASME
Research Committee on Fluid Meters, 5th edition, 1959.
31. Giger and Blumer, "PAH in the Environment" Anal. Chem. Vol. 46, p.
1663, (1974).
32. Klimisch, H. J., "Separation of PAH By High Pressure Liquid Chromato-
graphy" Journal of Chromatography Vol. 83, pp. 11-14 (1973).
33. Klimisch, H. J., "Determination of PAH", Analytical Chemistry, Vol. 45
Sept. 1973, pp. 1960-1962.
34. Qazi, A. H., et al. "Identification of Carcinogenic and Noncarcinoge-
nic PAH's Through Computer Programing", Amer. Indust. Hygiene Assoc.
Journal, Dec. 1973, pp. 554-558.
35. Lao, R. C., Thomas, R. S. and Monkman, J. L., "Computerized GC-MS.
Spectrometric Analysis of PAH in Environmental Samples", J. of Chro-
matography 112 (1975), pp. 681-700.
36. Lao, R. C., Thomas, R. S., Oja, H. and Dubois, L., "Application of a
Gas Chromatograph-Mass Spectrometer - Data Processor Combination to
the Analysis of the PAH Content of Airborne Pollutants", Analytical
Chemistry, Vol. 45, May 1973, pp. 908-915.
146
-------
BIBLIOGRAPHY
1. Methods of Measuring and Monitoring Atmospheric Sulfur Dioxide U S
Department of He lath, Education and Welfare; National Center for Air*
Pollution Control; Cincinnati, Ohio; 1964.
2' «^?1?i!?9c?ol?!itant Relative Humidity by Means of Aqueaus Solutions,
ASTM E 104-51 (Reapproved 1971). '
3. Lee, M. L., Novotny, K. 0., and Bartle, K. D., "Gas Chromotography/
Mass Spectrometric and Nuclear Magnetic Resonance Spectrometric
Studies of Carcinogenic Polynuclear Aromatic Hydrocarbons in Tobacco
and Marijuana Smoke Condensates", Analytical Chemistry, Vol. 48, pp.
405-416, Feb. 1976.
4. Bhatia, I., "Gas Chromotographic Determination of Polycyclic Aromatic
Hydrocarbons", Analytical Chemistry, Vol. 43, pp 608-9, April 1971.
5. Lane, D. A., Moe, H. K., and Kstz, M., "Analysis of Polynuclear
Aromatic Hydrocarbons, Some Heterocyclics, and Hliphatics with a
Single Gas Chromotograph Column", Analytical Chemistry, Vol. 45, pp.
1776-78, August 1973.
6. Janini, G. H., Johnston, K., and Zielinski, W. L., "Use of Nematic
Liquid Crystal for Gas-Liquid Chromatographic Separation of
Polyaromatic Hydrocarbons", Analytical Chemistry, Vol. 47, pp.
670-674, April 1975.
7. Lee, M. L., Bartle, K. D., and Novotny, M. V., "Profiles of the
Polynuclear Aromatic Fraction from Engine Oils Obtained by
Capillary-Column Gas-Liquid Chromatography and Nitrogen - Selective
Detection", Analytical Chemistry, Vol. 47, pp. 540-542, March 1975.
8. Lao, R. C., Thomas, R. S., Ota, H., and Dubois, L., "Application of a
Gas Chromatography-Mass Spectrometer-Data Processor Combination to the
Analysis of the Polycyclic Aromatic Hydrocarbon Content of Airborne
Pollutants", Analytical Chemistry, Vol. 45, pp. 908-915, May 1973.
9. Tejada, S. B., Sigsby, J. E., and Bradow, R. L., "Determination of
Soluble Sulfates in Automobile Exhaust by Automated HPLC Modification
of the Barium Chloranilate Method", Mobile Source Emissions Research
Branch, Environmental Sciences Research Laboratory.
10. "Sulfur Oxides in Flue Gases (Barium ChIoranilate-Control led
Condensation Method)", ASTM D 3226-73T.
147
-------
11. Evaluation of the Reference Method for Determination of Sulfur Dioxide
in the Atmospheric (Pararosani line Method).
12. Slowik, A. A., and Sansone, E. B., "Diffusion Losses of Sulfur Dioxide
in Sampling Manifolds". Journal of the Air Pollution Control
Association, Vol. 24, pp. 245-247, March 1974.
13. Scaringelli, F. P., Elfers, L., Morris, V., and Hockheiser, S.,
"Enhanced Stability of Sulfur Dioxide Solution", Analytical Chemistry,
Vol. 42, pp. 1818-1820, Dec. 1970.
14. Cedergren, A., etal, "Comparison of High Precision Coulometer and
West-Gaeke Methods with the Gravimetric Method for Preparation of
Standard Sulfur Dioxide Gas Blends Using Permeation Tubes", Analytical
Chemistry, Vol. 47, pp. 100-106, January 1975.
15. Smith, J. P., and Urone, P., "Static Studies of sulfur Dioxide
Reactions Effects of N02, C^f,, and H^O", Environmental Science
and Technology, Vol. 8, pp. 742-746, August 1974.
16. Young, M., Driscoll, J. N., and Mahoney, K., "Potentiometric
Determination of Sulfur Dioxide in Flue Gases with an Ion Selective
Lead Electrode", Analytical Chemistry, Vol. 45, pp. 2283-2284, Nov.
1973.
17. Logsdon, 0. J., II, and Carter, M. J., "Comparison of Manual and
Automated Methods for Sulfur Dioxide in Manually Impinged Ambient Air
Samples", Environmental Science and Technology, Vol. 9, pp. 1172-74,
Dec. 1975.
18. "Analysis by Microscopical Methods for Particle-Size Distribution of
Particulate Substances of Subsieve Sizes", ASTM E20-68 (Reapproved
1974).
19. "Concentration and Particule Size Distribution of Airborne Particulates
Collected in Liquid Media Using and Electron Counter", ASTM D3365-74T.
20. Spumy, K. R., et. al., "A Note on the Sampling and Electron Microscopy
of Asbestos Aerosol in Ambient Air by Means of Nuclepore Filters".
21. Spurny, et. al, "Aerosol Filtration by Means of Nuclepore Filters,
Aerosol Sampling and Measurement", Environmental Science and
Technology, Vol. 3, pp. 464-468, May 1969.
22. Denee, P. B., and Stein, R. C., "An Evaluation of Dust Sampling
Membrane Filters for Use in the Scanning Electron Microscope", Powder
Technology, Vol. 5, pp. 201-204 (1971/72).
23. Novotny, M., etal, "Gas Chromatographic Column for the Viking 1975
Molecular Analysis Experiment", Science. Vol. 189, pp. 215-216,July
1975.
148
-------
24- Zlatkis, A., Bertsch, W., et. al, "Profiles of Volatile Metabolites in
25. Liu, Y. H. and Lee, K. W., "Efficiency of Membrane and Nuclepore
Filters for Submicrometer Aerosols", Environmental Science and
Technology, Vol. 10, pp. 345-350, April 1976.
25. Pierce, R. and Katz, M., "Determination of Atmospheric Isothermic
Polycyclic Arenes by Thin Layer Chromatography and Fluorescent
Spectrophotometry", Anal. Chem. Vol. 47, pp. 1743-7 (1945).
27. American Society for Testing Materials, Procedure D-2682-71 (1971).
28. Lee, M. L. and Hites, R. A., "Characterization of Sulfur-Containing
Polycyciic Aromatic Compounds in Carbon Blacks", Anal. Chem., Vol. 48,
pp. 1890-93 (1976).
29. Lee, M. L., Novotny, M. and Bartle, K. D., "Gas Chromatography/Mass
Spectrometric and Nuclear Magnetic Resonance Determination of
Polynuclear Aromatic Hydrocarbons in Airborne Particulates", Anal.
Chem. Vol. 48, p. 1566 (1976).
30. Foster, J. F. et. a I, Chemical and Physical Characterization of
Automotive Exhaust Particulate Matter in the Atmosphere, Coordinating
Research Council, CAPE 12-68-Neg. 59 and CAPE 19-70 (1972).
31. Novotny, M., Lee, M. L., Low, C. E. and Raymond, A., "Analysis of
Marijuana Samples from Different Origins by High Resolution Gas-Liquid
Chromatography for Forensic Application", Anal. Chem., Vol. 48, pp.
24-29 (1976).
3.2. Lao, R. C., Thomas, R. S. Oja, H., and Dubois, L., "Application of a Gas
Chromatography-Mass Spectrometer-Data Processor Combination to the
Analysis of Polycyclic Aromatic Hydrocarbon Content of Airborne
Pollutants", Anal. Chem., Vol. 45, pp. 908-15 (1973).
33. Golden, C., and Sawicki, E., "Ultrasonic Extraction of Total Particulate
Aromatic Hydrocarbons from Airborne Particles at Room Temperature",
Intl. J. Environ Anal. Chem., 1975, Vol. 4, pp 9-23.
34. Hare, C. T. Methodology for Determining Fuel Effects on Diesel
Particulate Emissions, EPA-650/2-75/056, Environmental Protection
Agency, Office of Research and Development, Washington, D. C. 20460,
March 1975.
35. Federal Register, Vol. 41, No. Ill, Tuesday, June 8, 1976.
36. Maserole, F. B. et. a I., "Sulfur Dioxide Interferences in the
Measurement of Ambient Particulate Sulfures", Radian Corp. Sponsored by
Electric Power Research Institute, 1976.
149
-------
37. Charleson, R. J., Ahlquist, N. C. and Horvath, H., "On the Generality of
Correlation of Atmospheric Aerosol Mass Concentration land Light
Scatter", Atmospheric Environment (Pergamon Press) Great Britain, Vol 2,
(1968) p. 455-464.
38. Chin-I Lin, Baker, M. and Charlson, R. J., "Absorption Coefficient of
Atmospheric Aersol: A Method for Measurement", Applied Optics, Vol. 12
pp. 1356-1363 (June 1973).
39. Denee, P. B. and Stein, R. L., "An Evaluation of Oust Sampling Membrane
Filters for Use in the scanning Electron Microscope", Powder Technology,
Vol. 5, pp. 201-207 (1971/72).
40. Frank, E. R., Spumy, K. R., Sheesley, D. C. and Lodge, J. P., "The Use
of Nuclepore Filters in Light and Electron Microscopy of Aerosols",
Journal de Microscopie, Vol. 9, pp. 735-740 (1970).
41. Machacona, J., Hrbek, J., Hampl, V. and Spurny, K., "Analytical Methods
for Determination of Aerosols by Means of Membrane Ultrafilters
XV-Hydrodynamical Properties of Membrane and Nuclepore Filters", Coll.
Czechoslov. Chem. Comm., Vol. 35, pp. 2087-2099 (1970).
42. Maggiore, C. J. and Rubin, I. B., "Optimization of an SEM X-Ray
Spectrometer System for the Identification and Characterization of
Ultramicroscopic Particules", Proceedings of Scanning Electron
Microscope Symposium, IIT Research Institute, Chicago, 111, (April 1973).
43. Melo, 0. T. and Phillips, C. R., "Aerosol-Size Spctra by Means of
Membrane Filters", Environmental Sci. & Tech. Vol. 8, pp. 67-71 (1974).
44. Spurny, K., Blaschke, R. and Pfefferon, G., "Analytical Mekthods for
Determination of Aerosols by Means of Membrane Ultrafilters XVI", Coll.
Czechoslov. Chem. Comm. Vol. 36, pp. 949-953 (1971).
45. Spurny, K., Hrbek, J. and Lodge, J. P., "Analytical Methods for
Determination of Aerosols by Means of Membrane Ultrafilters XZVIII
Aerosols Sampling Under Extreme Gas Conditions", Coll. Czechoslov. Chem.
Comm. Vol. 36, pp. 2749-2756 (1971).
46. Spurny, K. and Lodge, J. P., "Analytical Methods for Determination of
Aerosols by means of Membrane Ultrafilters XII - Filtration Mechanisms
of Pore Filters Studied by Means of Electron Microscopy", Coll.
Czechoslov. Chem. Comm., Vol. 33, pp. 3931-3942 (1968).
47. Spurny, K. and Lodge, J. P., "Analytical Methods for Determination of
Aerosols by Means of Membrane Ultrafilters XIII - Aerosol Sampling and
Aerosol Measurement by Means of Nuclepore Filters", Coll. Czechoslov.
Chem. Comm., Vol. 33, pp. 4385-4389 (1968).
150
-------
48. Spurny, K. R. and Lodge, J. P., "Collection Efficiency Tables for
Membrane Filters Used in the Sampling and Analysis of Aerosols and
Hydrosols", Nat'I Center for Atmospheric Research, Report No.
NCR-TN/STR-77, Vol. I, II, III (Aug. 1972).
49. Spurny, K. and Lodge, J. P., "Analytical Methods for Determination of
Aerosols by Means of Membrane UltrafiIters XI - Structural, and
Filtration Properties of Nuclepore Filters", Coll. Czechoslov. Chem.
Comm., Vol. 33, pp. 3679-3693 (1967).
50. Spurny, K. R., Lodge, J. P., Frank, E. R. and Sheesley, D. C., "Aerosol
Filtration by Means of Nuclepore Filters - Aerosol Sampling and
Measurement", Environmental Science & Technology, Vol. 3, pp. 464-468
(May 1969).
51. Spurny, K. and Madelaine, G., "Analytical Methods for Determination of
Aerosols by Means of Membrane Ultrfi Iters XIX - Efficiency Measurement
of Nuclepore Filters by Means of Latex Aersols", Coll. Czechoslov. Chem.
Comm. Vol. 36, pp. 2857-2866 (1971).
52. Spurny, K. and Pich, J., "Analytical Methods for Determination of
Aerosols by Means of Membrane UltrafiIters VII - Diffusion and Impaction
Precipitation 6f Aerosol Particles by Membrane UltrafiIters", Coll.
Czechoslov. Chem. Comm. Vol. 30, pp. 2276-2287 (1965).
53. Spurny, K. R., Lodge, J. P., Frank, E. R. and Sheesley, D. C., "Aerosol
Filtration by Means of Nuclepore Filters Structural and Filtration
Properties", Coll. Czechoslov. Chem. Comm. Vol. 30, pp. 2276-2287 (1965).
54. Spurny, K. R., Lodge, J. P., Frank, E. R. and Sheesley, D. C. and
Wilder, B., "Aerosol Filtration by Means of Nuclepore Filters, Filter
Pore Clogging", Environmental Sci. & Tech., Vol. 8, pp. 758-761 (1974).
55. Spurny, K. R., Stober, W., Ackerman, E. R., Lodge, J. P., Spurny, K., "A
Note on the Sampling and Electron Microscopy of Asbestos Aerosol in
Ambient Air by Means of Nuclepore Filters", 67th Annual Meeting of APCA,
Denver, Colorado, Paper 74-47 (June 1974).
56. Twomey, S., "Measurements of the Size of Natural Cloud Nuclei by Means
of Nuclepore Filters", Journal of the Atmospheric Sciences, Vol. 29, pp.
318-321 (1971).
57. Wedberg, G. H., Chan, K. C., and Cohen, B. L., "X-Ray Fluorescence Study
of Atmospheric Particulates in Pittsburgh", Environmental Sci. & Tech.,
Vol. 8, pp. 1090-1093 (1974).
151
-------
APPENDIX A
GCMS ANALYSIS OF POLYNUCLEAR MIXES AND
TYPICAL TURBINE COMBUSTOR EXHAUST
SAMPLES
Four samples containing polynuclear aromatic hydrocarbons were received.
Three were mixes of the first sixteen compounds of Table A-l, of graded con-
centration, while the fourth was a representative unknown sample.
Experimental Method
The initial set of runs were performed in January, 1977. A packed, OV-1,
6 ft. glass column was used, with a temperature program of 150° to 280°
at a rate of 8°/min., with a 10 minute hold at 280°. The mass spectrome-
ter was operated in the selected mass scan mode, scanning mass ranges of
166-170, 177-180, 200-204, 226-230, 250-254, 276-280, 298-302. This technique
minimizes interferences and maximizes sensitivity, and is practical when the
components are known in advance. It proved quite satisfactory for the cali-
brating samples; Figure A-l shows a chromatogram obtained from sample PS102,
which had a concentration of 100 picograms/microliter. Table A-l lists the
peak areas found for each component in this run.
This technique proved less satisfactory for the unknown sample. It was
possible to observe the anthracene/phenanthrene peaks, but little else of in-
terest. Additionally, a large amount of silicone material was present which
gave interferring peaks. An attempt was then made to fractionate and concen-
trate the sample. This allowed us to detect more PNAs, but the interferences
from the silicone peaks were still serious. The sample was run again using
chemical ionization in order to reduce the effect of the silicones. PNAs to
m/e228 were detected.
When our capillary column became operational in April, the samples were
rerun. This proved so superior that it was possible to run the unknown sample
without concentration, or separation. Additionally, many more compounds were
measured. The results of these runs are shown in Table A-l and Figures A-2
and A-3. The column is a 20 meter, OV101 coated glass type, coupled directly
to the MS without a separator. It was operated in the split less, solvent
trapping mode, with a temperature program of' 30° to 180° at a rate of
16°/min.s followed by a 4°/min. program to 260.
152
-------
TABLE A-l
m/e
166
178
202
228
229
252
276
278
300
156
170
192
Compound
Fluorene
Anthracene,
Phenanthrene
Pyrene,
Fluoranthene
Chrysene ,
Triphenylene,
Benzoanthracene
Benzacridine
Benzo (e) pyrene,
Pery 1 ene ,
Benzo (a) pyrene
Benzo (ghi) perylene
diBenz (a,h) anthracene
Coronene
Dimethyl Napthalenes
Trimethyl Napthalenes
Methyl Anthracene,
Methyl Phenanthrene
PS1021
Packed Col.
122
1000
306; 376
826
12;80
596
82; 51
56
69
OTHER3
-
-
-
PS1022
Capi 1 lary
139; 586
561;226
405; 311
248
14
54;32;53
42; 26
14
-
-
-
-
UNK #12
Capi 1 lary
110;74
1000 ;981
88; 15
19
-
4;2;2
-
-
-
234; 633;
117 ;50
121
196; 66
1.
2.
Normalized to 1000.
Capillary runs normalized
to 1000, both
to same scale.
(Packed col-
umn to independent scale.)
3. These components were looked for only in the unknown sample. They
are representative of other polycyclics usually present as combus-
tion products, and not exhaustive. Many others are almost certainly
present.
153
-------
100
228
276 .
100
300
500
I
550
r
600
650
700
Figure A-l. Chromatogram of sample PS102 using OV-1 glass column.
-------
100"
50
100
150
200
250
300
350
400
I
450
100
-
™
-
_
M
^
>00
UJ
Z
UJ
X
z
OC
0
D
u.
UJ
z
UJ
OC
^*
cu
_*/\
I 1 1
550 600 650 700
UJ
Z
5
QC
O
^f
N
Z
UJ
CO
750
UJ
US UJ
> g
ul <
X OC
OL X
OC 2
UJ* O
Z N
UJ Z
CO UJ
> 00
OC
X
u
ft
I 1
800 850 900
Z
UJ
en
OL
"w
0
N
z
UJ
00
1 1
950 1000
Figure A-2. Chromatogram of known PNA compounds using 0V-101 glass capillary
column.
-------
100
100
150
200
250
c/i
300
350
400
450
100
«
••
_
.
-
-
1 1
00 550 600
0
1-
IU
uf
z
Ul
o
CC
I
z
N
Z
Ul
m
LU
z
Ul
s.
CC
o
I
650 700
Ul
Z
Ul
CC
OL. ^
"• ?
Z
Ul
ff
^
LU _1 -2
N >-
m a
. II
1 1
750 800
N
Z
Ul
oa
I
I 1 1 i
850 900 950 1000
Figure A-3. Chromatogram of unknown PNA compounds using OV-101 glass capil-
lary column.
-------
DISCUSSION
The results listed in Table A-l compare the data obtained on the PS102
sample with the packed and capillary columns, with a considerable time sepa-
ration; and the data obtained on the PSl02 and unknown #1 samples, using the
capillary. The capillary runs are normalized to the same 1000 scale, and are
directly comparable. Both samples were run unconcentrated, through the origi-
nal value of 100 picograms/component is probably no longer valid due to the
age of the sample. The large amount and number of other components can be ob-
served in Figures A-2 and A-3. Full mass range scans were used for these
runs. A samplilng of alkylated PNAs were sought and found, as shown in Table
A-l. Others are likely present, such as the pheny! anthracenes. It. is antici-
pated that these will be searched for in future runs.
157
-------
APPENDIX B
PNA CONTRIBUTION FROM FILTERS AND SOLVENTS*
by
D. J. Robertson, R. H. Groth, D. G. Gardner, and E. G. Glastris
Pratt & Whitney Aircraft Group
East Hartford, CT 06108
ABSTRACT
The polynuclear aromatic hydrocarbon (PNA) content of participate matter
emitted into the air from combustion or other processes is receiving increas-
ing attention. The particulate matter is normally collected on a filter which
is subsequently extracted with an organic solvent and then analyzed by various
methods. The background levels (nanograms or lower) of PNA or other substances
(i.e., phenols, nitrosamines, etc.) in the filters and solvents can be signi-
ficant sources of error in analytical procedures.
In this paper we report the presence of these compounds in most of the
readily available filter media as well as in analytical grade solvents used to
extract the filters. The presence of these compounds becomes apparent only
upon concentration to a few mi Hi liters volume of about 150 mi Ililiters of the
solvent itself or after use of the solvent in extracting an unused filter. The
analysis is by means of high performance liquid chromatography using an ultra-
violet fluorescence or absorption detector. Filters used in collection or in
extraction were made of teflon, glass fiber, callulose and organic polymers
and solvents investigated were benzene, CHCI3, CH2CI2, hexane and cyclo-
hexane. The effects of heat treatment on glass fiber filters is also noted.
Finally a recommended procedure to purify and evaluate the solvent and to
choose the filter media is offered.
INTRODUCTION
The polynuclear aromatic hydrocarbon (PNA) content of particulate matter
emitted into the air from the combustion of fossil fuels or other processes is
receiving increasing attention. The particulate matter, which absorbs these
species, is normally collected on a filter which is subsequently extracted
with an organic solvent and then analyzed by various methods. The presence in
the filter or solvent of PNA or other substances (phenols, nitrosamines, etc.)
even in the nanogram concentration range and lower can be a significant source
of error in the analytical procedure.
*Presented at the 70th Annual Meeting of the Air Pollution Control Association,
Toronto, Canada, June 20-24, 1977, Paper 77-36.1.
158
-------
Filters used to collect the particulate matter may be made of teflon
glass fiber, cellulose or organic polymers. In addition, binders may be ad-
ded. Solvents used in extraction procedures include benzene^'2), methylene
chloride(3-4), chloroform^), hexane, and cyclohexane(6~8). After ex-
traction, the solvent with the extracted material is usually reduced in vol-
ume by vacuum distillation or evaporation to make a more concentrated ex-
tract. In this way impurities originally present in the solvent or extracted
from the filter material are also concentrated. The concentrated extract may
be analyzed by high performance liquid chromatography (HPLC), gas chromato-
graphy, spectral methods (NMR, UV, IR) or wet chemical procedures used, for
example, in the analysis of nitrosamines. In the nitrosamine procedure^9),
diisopropy I ether is used and hence the presence of nitrosamines or other or-
ganic bound nitrogen in this solvent is of concern. In this paper., we discuss
the presence of interfering substances in many of these filter materials and
solvents which have been established by ultraviolet and fluorescence detec-
tion methods with the HPLC and by the nitrogen-phosphorous detector (rubidium
bead) in the nitrosamine procedure. Recommendations for purification of sol-
vents and choice of filters are given.
EXPERIMENTAL METHODS
Solvents
To evaluate solvents for possible interfering contaminants, a volume of
150 ml of the solvents benzene, n-hexane, cyclohexane, chloroform or methyl-
ene chloride (all Fisher reagent grade) is reduced to a volume of 2 to 3 ml
by means of vacuum distillation. This procedure serves to concentrate the im-
purities with high boiling points and prevents decomposition of either the
solvent or impurities. The concentrated solution of the possible contaminants
is now analyzed using a Dupont Model 830 High Performance Liquid Chromato-
graph with a Dupont Model 835 Multiwavelength Photometer ultraviolet absorp-
tion and fluorescence detector. In our unit, the 254 nm wavelength was used.
The column used was packed with octadecy Isi lane (ODS) at 50°C with 80:20
methanol-water as the mobile phase at 2000 psig. This procedure was used to
detect polynuclear aromatic compounds. Known compounds containing 3, 4, 5,
and 6 fused rings were analyzed to determine retention times and sensitivi-
ties on both detectors for purpose of identification and quantification. For
those solvents containing impurities, redistillation in glass of a fresh
batch was carried out followed by concentration and analysis of impurities as
before. For nitrosamine analyses, diisopropylether is used to extract these
substances and others from a phosphoric acid-water solution^5'. Therefore,
diisopropylether was analyzed using a Perkin-Elmer Model 3920B Gas Chromato-
graph equipped with a nitrogen-phosphorous detector. The column was 6' x 1/8"
O.D. stainless steel packed with 10% KOH + 10% Carbowax 1540 on 60-80 mesh
Gas Chrom Q maintained at 125°C for four minutes and then temperature prog-
rammed to 159° at 2°/minute.
Gelman Type GF/A glass filter, Gelman Type GF/E glass fiber, Whatman GB/B
qlass fiber, Millipore Standard (mixed esters of cellulose with a triton sur-
factant ) MilHpore Teflon with polyvinyl chloride or polypropylene backing
and Millipore Mitex Teflon filters were studied. Cellulose thimbles for use
159
-------
in Soxhlet extraction were also evaluated. These filters or thimbles were ex-
tracted with 150 ml of contaminant free n-hexane in a Soxhlet extractor for
16 hours. The contaminant-free hexane was double distilled in glass. The ex-
tracting solvent was then reduced in volume to about 1.5 ml and analyzed in
the same manner as in the study of the solvents.
In addition, the glass fiber filters were evaluated for the effect of
heat to remove contaminants. The filters were placed in a muffle furnace for
two hours at 500°C prior to extraction with n-hexane as before.
RESULTS
Table B-l lists the retention times and sensitivities for known PNA com-
pounds using the HPLC with the fluorescence and ultraviolet absorption detec-
tors. Table B-2 gives the retention times and relative responses of contami-
nants found in the various solvents after concentration. In some cases, the
specific peaks projected from a broad absorption band. Redistillation of some
of the more promising solvents was carried out in glass and concentration/
analysis was then repeated. These results are also in Table B-2 for n-hexane.
Other solvents showed little improvement.
TABLE B-l. KNOWN COMPOUNDS - RETENTION TIME* AND SENSITIVITIES
Retention Time Response** (Peak Height, in.)
Compound (Min./Sec.) U.V. Abs. Fluorescence
Naphthalene
Anthracene
Fluoranthene
Pyrene
Chrysene
Bene (e) pyrene
Bene (a) pyrene
Dibenzo (ah) anthracene
Benzo (ghi) perylene
3:42
4:54
5:12
6:00
7:36
10:48
11:48
13:12
17:22
«« « •
64.8
182.4
34.8
4.1
47.2
29.6
30
256
19.2
2611
41778
190
12.8
256
15974
671
10445
*HPLC, 25 cm X 2.3 mm I.D. stainless steel column, octadecylsilane (ODS) at
50°C, 80:20 methanol-water mobile phase, pressure 2000 psig.
**10 ul injection containing 10~6 grams of compound.
160
-------
TABLE B-2. SOLVENT IMPURITIES
Solvent
Retention Time of
Impurity (Min./Sec.)
Response* (Peak Height) In.
U.V. Abs. Fluorescence
CHCI3, Reagent
CH2CL2, Reagent
Benzene, Reagent
Cyclohexane, Reagent
(large absorption band
UV between 2-9 min.,
ave. 3")
n-hexane, Reagent
(large absorption band
between 3-11 min. ave.
1.7" and 41" for UV and
f luor.)
n-hexane, Redistilled
(large absorption band
between 5.5-9 min. in
fluor. ave. 12.8")
1:15
3:20
6:30
2:40
4:15
6:15
6:50
8:30
9:20
10:00
13:30
14:40
3:20
3:50
4:05
5:15
6:30
7:10
8:00
8:30
9:35
11:00
12:00
12:30
13:00
13:30
16:50
3:00
3:54
5:30
11:15
4:00
4:15
5:00
5:20
6:30
7:30
9:00
5:20
9:00
4
38
Trace
20
3.2
Trace
Trace
Trace
Trace
Trace
Trace
Trace
34
8
2
6
0.2
0.24
Trace
Trace
0.05
0.05
____
__..
-._ __
_.__
0.5
0.5
4
2.5
0.5
0.5
2.5
2.5
1.5
0.7
0.7
....
1.2
9.6
76.8
16
64
192
16
12.8
11.5
6.4
12.8
7.7
6.4
960
1741
....
25.6
28.2
....
12.8
____
9.6
3.2
6.4
6.4
....
4.8
8
••««
19.2
____
____
70.4
....
70.4
73.6
64.0
6.4
*HPLC, 10 ul injection of concentrated solvent. Same conditions as in Table
B-l. Responses are peak heights above large bands, if any.
161
-------
Baker reagent grade diisopropyl either was found to have a significant
impurity which could be greatly reduced by twice redistilling the solvent in
glass with elimination of the last 10% of the distillate. The impurity de-
creased by a factor of 20 to a level of 10 ug/ml. Table B-3 shows the reten-
tion and sensitivities for certain nitrosamines and the solvent impurity. The
response of the nitrogen-phosphorus detector to the impurity is based on the
assumption of a similar sensitivity to N, N-nitrosamines.
Table B-4 gives the retention times and relative responses of impurities
found in the n-hexane extract of the various filters and thimble after con-
centration. In some cases, specific peaks projected from broad absorption
bands. No measurable peaks were found in the concentrated extract from the
heated Gelman GF/A Glass Fiber filter. This filter was chosen because it had
the least contaminants as shown in Table B-4. Unfortunately, the filter be-
came brittle in the heating process and not practical to use for our applica-
tion. Specific identities of contaminants in the solvents and filters are not
known. Many are likely to be PNA compounds based on their retention times. In
any case they would interfere in analyses of PNA compounds.
CONCLUSIONS
1. Glass distilled n-hexane was found to be the best solvent for use in
HPLC/fluorescence - UV detection work. Cyclohexane could be used, after puri-
fication. The chlorinated solvents and benzene were not satisfactory even on
redistillation in glass.
2. Diisopropylether may be suitable for use in nitrosamine analysis(6)
after redistillation.
3. Preheated Gelman GF/A glass fiber filters, and the Mi IMpore Mitex
Teflon 5 and 10 micron filters were found to be suitable for work as far as
PNA contamination is concerned using the HPLC/Fluorescence - UV absorption
detectors. The Teflon filter is favored because of the brittleness of the
heat treated glass fiber filter.
4. It is recommended that filter media and solvents be evaluated by the
procedures given to determine their usefulness for the specific analyses to
be carried out.
162
-------
TABLE B-3. RETENTION TIMES OF NITROSAMINES
Compound
Retention Time
Sec.
Peak Are for lu'
of 100 ppm Nitrosamine
Solvent impurity
N, N - Diethylnitrosamine
N, N - Diisopropylnitrosamine
N, N - Dibutylnitrosamine
211
400
662
1106
949881
399274
283542
207314
Instrument: Perkin Elmer Model 3920B Gas Chromatograph with nitrogen-
phosphorus detector
Column: 6' X 1/4" O.D. stainless steel packed with 10% Carbowax 1540
on Gas Chrom Q maintained @ 125°C for 4 minutes and then
programmed to 159°C @ 2°/min.
Flow: Column - H^ @ 5 psi and 2 ml/min; Air @ 44 psi and 100 ml/
min;
Carrier - He @ 15 ml/min.
Data Output: Auto labs System IV Computing Integrator
TABLE B-4. FILTER IMPURITIES
Filter
Retention Time* of Response* (Peak Height) in.
Impurity (min./sec.) U.V. Abs. Fluorescence
Gelman Type GF/A
(Large band between 3-
11 min. average 2.4"
and 64" for UV and
fluor.)
Gelman Type GF/E
(Large band between 3-
11 min. average 3" and
96" for UV and fluor.)
3:30
4:15
5:00
6:45
8:00
9:00
9:45
3:30
4:15
5:00
6:45
8:00
9:00
9:45
0.8
0.4
0.4
0.6
0.8
1.2
1.4
1.9
0.3
2.3
0.3
0.2
0.8
___ _
9.6
32
Trace
19.2
22.4
22.4
51.2
(Continued)
163
-------
TABLE B-4 (Continued)
Filter
Retention Time* of
Impurity (min./sec.)
Response* (Peak Height) in.
U.V. Abs. Fluorescence
Millipore Teflon 3
(polypropylene)
(Large band between 3-
11 min. average 2.8"
and 51.2 for UV and
fluor.)
Standard Mi 1 lipore
(Large band between 3-
11 min. average 1.6"
and 41.6" for UV and
fluor.)
Millipore PVC
Whatman GF/B
(Large band between 3-
11 min. average 8"
and 352" for UV and
fluor.)
Mi 1 lipore Mitex
Cellulose Thimble
(Large band between 3-
11 min. average 2"
and 104" for UV and
fluor.)
2:20
2:35
3:00
3:15
3:40
4:15
5:15
5:50
6:15
7:45
8:15
8:30
10:00
10:30
12:50
15:15
2:30
3:15
3:40
3:48
4:45
5:30
6:00.
No peaks but PVC soluble
and limited to 100°C.
1:12
1:50
2:24
3:12
3:50
4:30
5:05
6:30
None detected
0:42
1:04
1:50
2:25
2:40
3:00
3:24
3:34
4.8
4.8
52
3.2
2
0.4
2.4
— __
--._-
4.4
4.4
_.__
1.2
0.6
0.8
2.8
0.8
0.6
in chlorinated
2.4
0.8
0.4
1.2
0.4
--__
0.4
0.9
10.4
0.5
-___
0.1
0.3
0.5
64
6.4
6.4
44.8
102.4
6.4
12.8
6.4
12.8
12.8
Trace
102.4
320
166.4
19.2
solvents
«___
12.8
192
128
32
««••«,
....
4.8
8
4.8
----,
____
105.6
*HPLC, 10 ul injection of concentration n-hexane extract. Same condition
as in Table B-l.
Responses are peak heights above large bands, if any.
164
-------
REFERENCES
1. Pierce, R. and Katz, M. "Determination of Atmospheric Isothermic
Polycyclic Arenes by Thin Layer Chromatography and Fluorescence
Spectrophotometry" Anal. Chem. 47, 1743-7 (1975).
2. "Test for Polynuclear Aromatic Hydrocarbons in Air Participate Matter"
American Society for Testing Materials, Procedure D-2682-71 (1971).
3. Lee, M. L. and Hites, R. A., "Characterization of Sulfur-Containing
Polycyclic Aromatic Compounds in Carbon Blacks" Anal. Chem. Vol. 48, pp.
1890-93 (1976).
4. Lee, M. L., Novotny, M. and Bartle, K. D., "Gas Chromatography/Mass
Spectrometric and Nuclear Magnetic Resonance Determination of
Polynuclear Aromatic Hydrocarbons in Airborne Partiallates" Anal. Chem.
Vol. 48, pp. 1566 (1976).
5. Foster, J. F. et al, Chemical and Physical Characterization of
Automotive Exhaust Particulate Matter in the Atmosphere, Coorderating
Research Council, CAPE 12-68-Neg. 59 and CAPE 19-70 (1972).
6. Novotny, M., Lee, M. L., Low, C. E. and Raymond, A., "Analysis of
Marijuana Samples from Different Origins by High-Resolution Gas-Liquid
Chromatography for Forensic Application" Anal. Chem. Vol. 48, pp. 24-29
(1976).
7. Lao,R. C., Thomas, R. A., Oja, H., and Dubois, L., "Application of a Gas
Chromatograph-Mass Spectrometer-Data Processor Combination to the
Analysis of the Polycyclic Aromatic Hydrocarbon Content of Airborne
Pollutants" Anal. Chem. Vol. 45, pp. 908-15 (1973).
8. Golden, C. and Sawicki, E., "Ultrasonic Extraction of Total Particulate
Aromatic Hydrocarbons from Airborne Particles at Room Temperature", Intl
J. Environ. Anal. Chem. 1975, Vol. 4, pp 9-23.
9. Hare, C. T., Methodology for Determining Fuel Effects on Diesel
Particulate Emissions, EPA-650/2-75/056, Environmental Protection
Agency, Office of Research and Development, Washington, D. C. 20460,
March 1975.
10. Fenton, D. L., Turbine Engine Particulate Sampler: Design Study IITRI
Report to USAF School of Aerospace Medicine No. SAM-TR-76-1, May 1976.
11. Conkle, J. P., Lackey, W. W., Miller, R. L., Hydrocarbon Constituents of
T-56 Combustor Exhaust, USAF School of Aerospace Medicine No.
SAM-TR-75-8 April 1975.
165
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-041
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
CHEMICAL COMPOSITION OF EXHAUST PARTICLES
FROM GAS TURBINE ENGINES
5. REPORT DATE
February 1979
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
D.J. Robertson, J.H. Elwood and R.H. Groth
8. PERFORMING ORGANIZATION REPORT NO,
9. PERFORMING ORGANIZATION NAME AND ADDRESS
United Technologies Corporation
Pratt & Whitney Aircraft Group
Commercial Products Division
East Hartford, Connecticut 06108
10. PROGRAM ELEMENT NO.
1AD712 BC-42 (FY-78)
11. CONTRACT/GRANT NO.
Contract No. 68-02-2458
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory—RTP, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
EHnal 11 /7fi - "3/7R
14. SPONSORING AGENCY CODE
EPA/600/09
18. SUPPLEMENTARY NOTES
16. ABSTRACT
A program was conducted to chentically characterize particulate emissions from a
current technology, high population, gas turbine engine. Attention was focused
on polynuclear aromatic compounds, phenols, nitrosamines and total organics. Poly-
nuclear aromatic hydrocarbons (PAH) were determined by HPLC, GC/MS and NMR techniques.
Phenols and nitrosamines were isolated and then measured by gas chromatographic
methods utilizing flame ionization detection and nitrogen detection. Total
organics were determined by a backflush chromatographic procedure. The particulate
matter was collected using a high capacity pumping system incorporating 293 mm
diameter Teflon filters through which was passed up to 43 m of exhaust gas.
Extraction of the organic matter was performed in a Soxhlet extractor using hexane.
The engine was operated at idle, approach, climb and take-off powser settings with
low sulfur (0.007%S) and high sulfur (0.25%S) fuels. Most of the PAH were small
3-to-4 fused ring species. No nitrosamines were found and except in a few cases,
at low levels, no phenols. PAH and total organic levels decreased with increasing
power setting and were more concentrated in the exhaust from the low sulfur fuel.
Less than 1% of the organic matter emitted from the engine was adsorbed on the
particulate matter.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
* Nitroso compounc
* Phenols
* Air pollution
* Gas turbine engines
* Exhaust emissions
* Particles
* Chemical composition
* Chemical analysis
* Aromatic polycyclic hydrocarbons
13B
2 IE
2 IB
07D
07C
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
178
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
166
------- |