EPA 600/3-83-106
PC84-122910
Exhaust Emissions from a Diesel Engine
Johns Hopkins Univ., Baltimore, MD
Prepared for
Environmental Sciences Research Lab,
Research Triangle Park, NC
Nov 83
Department of Commerce
Technical information Service
-------
TECHNICAL REPORT DATA
(Fleatt read Inttrucnom on iht reverie before completing)
1. REPORT NO.
EPA-600/3-83-106
2.
3. RECIPIENT'S ACCESSION NO.
PRR A 122910
4. TITLE AND SUBTITLE
EXHAUST EMISSIONS FROM A DIESEL ENGINE
5. REPORT DATE
November 1983
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
Terence H. Risby
9. PERFORMING PRGA.NIZ.ATION NAME AND ADDRESS,. . ..,..
John Hopkins University, School oFHygiene & Pub. Hlth.
Div. of Environmental Chemists
Dept. of Environmental Health Science
Baltimore, MD. 21205
10. PROGRAM ELEMENT NO.
C9YAlC/m-na/i? (FY-83)
11. CONTRACT/GRANT NO. » '
R-806558
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Sciences Research Laboratory - RTF, NC
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park. North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/09
is. SUPPLEMENTARY NOTES
t6. ABSTRACT
Studies were performed using (1) Diesel particles collected from the
undiluted exhaust of a single-cylinder engine, operated at constant speed and
load, using a binary pure hydrocarbon fuel with air or gas mixture oxidizers, and
(2) Diesel particles collected from the diluted exhaust of a multicylinder engine
operated on a commercial fuel. The physicochemical properties of the particles
were determined by static and dynamic methods. The organic adsorbate was
characterized by chromatographic and m^ss spectrometric procedures and by
microbial testing protocols. Electron paramagnetic resonance spectrometry and
high performance liquid chromatography were used to study reactivity of the
organic adsorbate.
The particles collected from the exhaust of Diesel engines operated on
binary pure hydrocarbons and on commercial fuels contained similar compounds in
the organic adsorbates. The nitrogen in these compounds is derived mainly from
the oxidizer (Air).
The microbial mutagenic activities of the organic adsorbates found on the
surface of both types of Diesel particles are comparable. These microbial
mutagenic activities can be attributed mainly to the presence of nitrated
polynuclear aromatic hydrocarbons.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Croup
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (Tllil Reporll
21. NO. OF PAGES
113
20. SECURITY CLASS (Tlii> pagel
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (R»». 4-77) PREVIOUS EDITION i> OBSOLETE .j
-------
Pildl-122910
EPA-600/3-83-106
November 1983
EXHAUST EMISSIONS FROM A DIESEL ENGINE
by
Terence H. Risby
Johns Hopkins University School of Hygiene and Public Health
Baltimore, Maryland 21205
EPA Grant Number R-806558
Project Officer
John E. Sigsby
Emissions Measurement and Characterization Division
Environmental Sciences Research Laboratory
Research Triangle Park, North Carolina 27711
HPRODUCfO 8Y
NATIONAL TECHNICAL
INFORMATION SERVICE
US. DfPARIMIM Of COMMICC!
SPRINGflilD. VA. 22161
ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
NOTICE
This document has been reviewed in accordance with U.S. Environmental
Protection Agency policy and approved for publication. Mention of trade
names or commercial prod.;~ts does not constitute endorsement or recommenda-
tion for use.
ii
-------
ABSTRACT
Studies were performed using (1) Diesel particles collected from the
undiluted exhaust of a single-cylinder engine, operated at constant speed
and load, using a binary pure hydrocarbon fuel with air or gas mixture oxi-
dizers, and (2) Diesel particles collected from the diluted exhaust of a
raulticylinder engine operated on a commercial fuel. The physicochemical
properties of the particles were determined by static and dynamic methods.
The organic adsorbate was characterized by chromatographic and mass spectro-
metric procedures and by microbial testing protocols. Electron paramagnetic
resonance spectrometry and high performance liquid chromatography were used -
to study reactivity of the organic adsorbate.
The particles collected from the exhaust of Diesel engines operated on
binary pure hydrocarbons and on commercial fuels contained similar compounds
in the organic adsorbates. The nitrogen in these compounds is derived mainly
from the oxidizer (Air).
The microbial mutagenic activities of the organic adsorbates found on the
surface of both types of Diesel particles are comparable. These microbial
mutagenic activities can be attributed mainly to the presence of nitrated
polynuclear aromatic hydrocarbons.
iii
-------
Table of Contents
Abstract • •
List of Figures vl
List of Tables vli
Acknowledgments « viii
Introduction • • • • • • 1
Conclusions and Recommendations • 2
Research Report • •' '........... 6
A. Engineering 6
Introduction • • • • 6
Experimental • • • 10
Procedures • 15
Results f... 18
B. Chemistry •• 26
li Chemical Analysis of Organic Adsorbates 26
Introduction •• 26
Experimental 27
Results 3?
2. Physicochemical Properties of Diesel Particulate Matter ... 45
a. Static Methods 45
Introduction 45
Experimental 46
Results 47
b. Dynamic Methods 65
Introduction 65
Experimental 66
Results 67
3« Transformation of Diesel Particulate Matter 74
Introduction 74
Electron Paramagnetic Resonance Spectrometry 75
Experimental 75
Results 76
Formation of 1-nitropyrene by Reaction of Nitrogen
Dioxide with Diesel Particulate Matter 83
Expe rimental 83
Results 83
C. Biology 85
Introduction • 65
Experimental 85
Results 91
Conclusions 98
References • 99
List of Publications which have resulted from this research 104
-------
List of Figures
Figure Page
1 Exhaust System 13
2 Gas Phase Instrumentation • • 1*
3 Pressurfc Traces • 21
4 Carbon Monoxide and Unburnt Hydrocarbons
as a function of Oxygen 22
5 Nitric Oxide and Nitrogen Dioxide as a function
of Oxygen • • 24
6 Ignition Delay, Particulate, and Extract as a function of
Oxygen 25
7 Block Diagram of Temperature - Programmable Platinum
Wire Probe 33
8 Capillary Gas Chromatogram of Air Oxidant Sample using
Flame lonization Detection 36
9 Capillary Gas Chromatogram of Nitrogen-free Oxidant Sample
using Flame lonization Detection 37
10 High Performance Liquid Chromatogram of Air Oxidant Sample ... 43
11 Volumetric Adsorption Apparatus 48
12 Adsorption Isotherms of Benzene on DPM-PSU (350°C) 52
13 Adsorption Isotherms of Nitrogen on DPM-EPA 54
14 Adsorption Isotherms of Benzene on DPM-EPA 55
15 Heat of Adsorption vs Surface Coverage for Benzene on
Spheron 6 58
16 Heat of Adsorption ys Surface Coverage for Benzene on
DPM-PSU (350°C) 5Q
17 Heat of Adsorption ys Surface Coverage for Hexane on
DPM-PSU (100°C) 60
18 Heat of Adsorption vs Surface Coverage for High Temperature
Benzene Adsorption on DPM-EPA (50°C) 61
19 Heat of Adsorption vs Surface Coverage for Low Temperature
Benzene Adsorption on DFM-EPA (50°C) 63
20 Heat of Adsorption vs Surface Coverage for Benzene on
DPM-EPA (400°C) 64
21 Gas Chromatographic Log Vr vs_ 1/Tc Plots for Benzene 68
22 Relative Positions of EPR Signals 77
23 Effects of Various Treatments on the EPR Signals of
DPM-PSU . 79
24 Nitration System 84
vi
-------
List of Tables
Table ' Page
1 Engine Parameters •• 11
2 Teat Conditions 18
3 Engine Data • 19
4 Chemical Characterization of Organic Adsorbates on
Diesel Particulate Matter » 38
5 Concentration of 1-nitropyrene in the Extracts of
Particulate Matter collected from Engines Operated on
Various Oxidants • 44
6 Electron Microscopy Results 4Q
7 True and Bulk Densities 50
8 6.E.T. Surface Areas 51
9 E.M. v£ B.E.T. Nitrogen Surface Areas 56
10 Isoteric Heat of Adsorption for Various Graphitized
Carbon Blacks •. 69
11 Variation of Apparent Heats of Sorption with
Volume of Adsorbate Injected 72
12 Isosteric Heats of Adsorption 73
13 Line widths and g-Values 76
14 Effects of Various Treatments on the EPR Signals of DPM-PSU ... 78
15 Effects of Selected Gases on the EPR Signals of DPM-PSU 81
16 Effects of Irradiation on the EFR Signal of DPM-PSU 82
17 Reaction of Nitrogen Dioxide with Nitrogen-free
Farticulate Matter 85
18 Genotype of tl.j TA Strains used for Kutagen Testing 86
19 Quantitative Ames Mutagenesis Results 92
20 Relative Transformation of Strain RUB 827 to Diesel
Farticulate Matter > 93
21 Percentage Survival of TA 98, TA 100 at Various Doses 94
22 Biological Activity of Used and Unused Lubricant
and Model Fuel 95
23 Biological Activity of Particulate Matter Collected from
Engine operated with Nitrogen-free Oxidant 96
24 Biological Activity of Particulate Matter Collected from
Engine operated with Different Oxidants 97
25 Biological Activity of Particulate Matter after Reactions
with Nitrogen Dioxide 98
vii
-------
ACKNOWLEDGMENTS
The principal investigator wishes to acknowledge and thank his collabora-
tors without whose help this project would not have been possible: Professor
S.S. Lestz, Professor M.R. Chedekel, Dr. J.A. Yergey, Dr. M.M. Ross, M.
Dukovich and J.D. Herr. The investigators would like to" express their thanks
to J.E. Sigsby, Jr., Dr. R.L. Bradow, Dr. S.B. Tejada, Dr. R.B. Zweidinger,
M. Walsh, Esq. and S. Blacker, Esq., of the U.S. Environmental Protection
Agency for their advice, help and continued support during the progress of
this research. Also we wish to thank Dr. S.R. Prescott and V. Lafferty for
allowing us to use their gas chromatograph-electron impact mass spectrometer
systems.
viii
-------
INTRODUCTION
A number of studies have reported that the exhaust particulate ciatter from
Diesel engine combustion represents a potential environmental health hazard.
These studies were based on the microbial activity of the organic adsorbate
which is obtained by the Soxhlet extraction of Diesel particulate matter with
dichloromethane. However these in vitro results have not been corroborated in
vivo by animal exposure studies since no measurable effects were obtained in
these latter studies. This dichotomy of results is complicated further by
other in vitro studies which have used certain physiological fluids to extract
Diesel particulate matter for microbial assays and have found the extracts to
be inactive. A possible explanation is that the positive or negative results
are caused by anomalies in the engine operation or from differences in fuels.
The chemical composition of Diesel fuel is widely variable and as a result it
is difficult to compare results which are obtained with fuels that have the
same ignition and boiling point characteristics but different chemical
compositions. An even more likely explanation is that the results are
different because they involve the use of different samples. The in vitro
analyses were performeu using collected particulate matter and the in vivo
analyses were performed with diluted Diesel exhaust. The difference may,
therefore, be the result of a sampling artifact. In this research we have
attempted to address all the variables of fuel, oxidant, lubricant and engine
operating parameters and have also made some interesting observations with
regard to sampling artifacts. We have also measured the physicochemical
properties of Diesel particulate matter and performed some preliminary studies
which have addressed the area of photochemical and chemical reactivity of
Diesel particulate matter.
-------
CONCLUSIONS AND RECOMMENDATIONS
This final report describes in detail the results that were obtained from
this research. These results have major impact on studies evaluating the
potential health risks that may be associated with exposure to Diesel
particulate matter. The major conclusions to be be drawn from this research
are as follows:
1. The compounds that have been identified and q-"ntified as sorbed on
the uurface of Diesel particulate matter are inherent to the combustion that
occurs in the Diesel engine. The same major compounds are found to be sorbed
on the particles if the Diesel engine is operated on commercial Diesel fuel or
a mixture (1:1 V/V) of n-tetradecane and 2,2,4-trimethylpentane (model) fuel
2. The combustion mechanisms, which lead to the production of the
polynuclear aromatic compounds, appear to involve short chain intermediates;
this is the only way to explain the formation of similar compounds from
different fuels.
3« The microbial activities are comparable for organic adso.rbates from
particulate matter collected from the exhaust of Diesel engines operated on
commercial Diesel fuel (5«5 Rev/ug TA98) or on the model fuel (4.3 Rev/ug
TA98).
4. Similar nitrogen containing compounds are found sorbed on the
particulate matter from engines operated on model fuel or commercial Diesel
fuel. The source of nitrogen has been thought to be the fuel or lubricant,
but since the model fuel and the synthetic lubricant (polyalkyleneglycol) are
nitrogen free the origin of th.j nitrogen in the engine operated on model fuel
must be the intake air.
-------
5«' The previous conclusion is confirmed by a study using a nitrogen free
model oxidizer intake; the acrbed molecules are found to be nitrogen free.
Also the microbial activity of the soluble organic extiact of.parviculate
matter collected frcn the engine operated on this model fuel and oxidizer is
reduced by a factor of fifty compared to the engine operated on the model fuel
and air.
6. Most of the direct micrcbial activity of the soluble organic extract
can be attributed to nitrated polynuclear aromatic hydrocarbons; including
1-nitropyrene.
7. The microbial activitj' of the soluble organic extract of particulate
matter collected from the engine operated with model fuel and oxidizer can be
restored by reacting the particuiate matter on a filter with nitrogen dioxide
(nitric acid). The microbial activities show direct correlation to the
concentration of 1-nitropyrane.
8. The direct correlation between micrcbial activity and 1-nitropyrene is
substantiated further by operating the engine with the model fuel and intake
air supplemented with various concentrations of oxygen and nitrogen.
9« The new or used lubricants used for the engine operated on model fuel
do not show anv microbial activities.
10. Preliminary studies using electron paramagnetic resonance spectronetry
show that Diesel particulate matter is photochemically active and will react
with other pollutant gases.
11. The mean particle diameter of Diesel particulate matter emitted by the
engine operated on the model fuel is found to be 36 nm by electron microscopy;
this is similar to the mean diameter of the particles which are collected from
an engine operated on commercial Diesel fuel (28 nm).
-------
12. The external surface areas of particles collected from the engine
2
operated on the model fuel are found by electron microscopy to be 86 m /g
and the external and internal surface areas of the samcj particles are found by
2
the nitrogen B.E.T. methodology to be 104 m /g. The external and total
surface area of particles collected from the engine operated on commercial
o
Diesel fuel are identical 11? m /?. The difference between these results
(if significant) can be explained by the latter particles havirg a coating of
unburnt hydrocarbons or lubricant on their surfaces which decreases the
accessibility of nitrogen to the pores of the particle.
13« The isosteric heats of adsorption for a series of model compounds were
measured for the particulate samples collected from the engine operated on the
model fuel and the engine operated on commercial Diesel fuel. The heats of
adsorption are significantly greater for the particulate matter collected from
the engine operated on the model fuel owing to the lower coverage by high
molecular weight compounds.
14. The variation of the apparent heats of sorption with increase of
surface coverage shows the expected decrease in sorption energy. This study
also shows that the particles have both high and low energy sorption sites.
15. Sampling conditions are shown to affect the quantity of adsorbates
which are present on the surface of the particles.
It is apparent from these results that a number of questions must still be
answered in order to establish conclusively whether Diesel particulate matter
represents a significant health hazard. The following recommendations
summarize the areas for future research which are obvious extensions of this
work.
1. Measure the bioavailability of sorbates from the surface of Diesel
particles using both in vitro and in vivo testing protocols.
-------
2. Determine the fate of airborne Diesel particles in typical urban
atmospheres in terms of photochemical and chemical reactions.
3- Determine the relationship between engine speed and load and the
chemical composition of the sorbate on the surface of Diesel particles.
4. Determine whether sampling artifacts play a role in the observed
biological activity of Diesel particulate matter. This study could most
easily be performed by in vivo exposure of animals to reaerosolized Diesel
particulate matter, providing that the aerosol has a similar size distribution.
-------
RESEARCH REPORT
This final report will be split, for the sake of clarity, into three areas.
(A) Engineering: The generation and collection of Diesel particulate matter.
(B) Chemistry: Determination of the physical and chemical properties of the
Diesel particulate matter, studying the reactions of the sorbates. (C)
Biology: Determination of the biological activity of the particulate matter
using various microbial assays.
A. ENGINEERING
INTRODUCTION
Diesel Combustion
The ideal Diesel cycle starts with a reversible isentropic compression:
th? piston compresses the air in the cylinder. Near top-dead-center fuel is
injected, into the hot compressed air and combustion is self-initiated. This
process is respresented in the ideal cycle as the addition of energy at
constant-pressure. At this point the actual process differs from the ideal
thermodynamic cycle. When the fuel is first injected into the cylinder, it
undergoes an ignition delay - the time between fuel vaporization and the
initiation of preflame chemical reactions. Consequently, some of the fuel
accumulates and burns at essentially constant-volume. Following ignition, the
fuel continues to spray into the combustion chamber, simulating the ideal
constant-pressure addition of heat. After the injection of fuel has ceased,
the hot gases expand to provide the power stroke of the engine. '/Then the
exhaust valve opens, the exhaust gas exits the cylinder. This blowdown
process is approximated as a constant specific-volume rejection of heat.
However, a real engine does not follow a cycle, since the exhaust gases are
expelled and atmospheric air is introduced. Furthermore, the real processes
depart from the processes of the ideal cycle in that they are neither
-------
reversible nor isentropic. Therefore, while the ideal Diesel cycle can serve
as a model to determine theoretical thermodynamic limits for an engine design,
the ideal cycle is of little use in understanding the actual Diesel combustion
system.
Combustion in the absence of nitrogen
To simulate the combustion cycle with an oxidizer other than air, several
parameters must be maintained. The compression pressure and temperature
of both the model oxidizer and air must be comparable in order to achieve
similar ignition properties. The peak combustion temperature for the model
oxidizer must be close to the peak temperature obtained with air, in order to
produce similar combustion products. The temperature and pressure of the
oxidizer at the engine inlet should be maintained as close to ambient
atmospheric conditions as possible. Constant fuel-to-oxygen ratios must be
maintained in the air and the nitrogen-free oxidizer systems, and a constant
mass flow rate of oxygen into the engine must be produced. These parameters
are governed by a system of thermodynamic, equations and properties of various
inert gases and oxygen can be substituted to calculate the composition of a
nitrogen-free oxidizer that will simulate air in the combustion cycle.
No single inert diluent was found to satisfy the thermodynaraic equations,
but they were satisfied by a mixture of carbon dioxide and argon. The actual
composition of the nitrogen-free oxidizer was calculated to be as follows:
21% oxygen, 23^ argon, and 56$ carbon dioxide.
Combustion with Nitrogen- and Oxygen-Enriched Air
The mixture of argon, oxygen and carbon dioxide eliminates nitrogen from
the combustion process, but the effects of excess nitrogen and oxygen in the
oxidant were also of interest. Therefore, small amounts of excess oxygen or
nitrogen were added to the intake charge. Although these additions altered
some of the engine parameters the thennodynanic and transport properties of
the engine changed vary little(l).
-------
Nitrogen added to the intake charge, simulates exhaust-gas-recirculation
(EGR) and this process is known to produce the following effects: reduction
in the emission of nitric oxide; increase in the exhaust temperature, and
increases in the emissions of particulate matter(l,2,3,4). The reduction in
nitric oxide is the reason for interest in Diesel EGR, bat the increase in the
emission of particulate matter is undesirable. However, the level of the
organic adsorbate is not increased by the use of EGR(5).
An increase in the oxygen content of the intake charge has both beneficial
and detrimental effects upon the engine. As oxygen is added to the air, the
ignition delay becomes considerably shorter, and the rate of pressure rise
increases(l,4). Also, the peak cylinder pressure increases until a maximum
value is reached at about 38$ oxygen(l). In direct contrast to nitrogen
addition, excess oxygen decreases the emission of particulate matter and
increases the emission of the oxides of nitrogen. Both these emissions are
dependent upon the oxygen radical concentration as indicated by the Zeldovich
equations:
0 + N2 ss=£ NO + N
N + 02 SP* NO + 0
0 + NO ^r N02
NO + 02s?= NO- + 0
Diesel Exhaust Emissions
Even though all engines emit both gaseous and aerosol products, Diesel
engines emit 30 to 50 times more particulate matter than comparable SI
engine(5)-
-------
Psrtifile Formation
For many years particle emissions have been indicators of Diesel engine
performance and condition and have been a main consideration in engine
design. Particulate matter, which is an aerosol of carbonaceous particles
suspended in the exhaust gas, is a product of incomplete combustion and is
commonly called soot or smoke. Although equilibrium equations fail to predict
the production of these particles at typical combustion pressures and
temperatures, it is always present in the exhaust(6).
Particulate matter is defined as any dispersed matter in the solid or
liquid phase present in diluted exhaust gases near ambient conditions(6).
These particles may vary from very small, single particles of 10 nm to long
clusters in the 10 to 30 ym range(6). The mass of the particulate matter in
the exhaust stream is affected by the type of fuel and fuel additives.
Increases in the fuel density, the sulfur content or aromatic content increase
the emission of particles(2). The ignition delay also affects the particulate
emissions since fuels with low cetane numbers yield lower masses of particulate
matter. This indicates that longer ignition delays are beneficial, giving the
fuel and air more time to mix. However, the gaseous products show little
fuel-to-fuel variation. Analysis of the major components of particulate
matter shows few differences between the fuels tested(2). The major
constituents of the particulate are carbon, hydrocarbons, sulfates, metals,
and water(6).
Precombustion chamber engines also emit up to 20^ less particles than a
comparable direct injection engine(3)« Also a difference in the nature of the
particulate matter has been observed. The indirect-injection engine produces
drier particles which contain more carbon, whereas the direct injection engine
produces oily particles characteristic of higher hydrogen-to-carbon ratiosf?).
-------
Dilution
. As the exhaust products travel down the exhaust pipe the temperature drops
and organic molecules sorb and condense on the carbon particles(7,8).
Initially, when the temperature is above 200°C, most organic molecules are
in the gas phase. This is evident when particulate matter is collected at
elevated temperatures, since only \% of the particulate matter collected at
200°C can be extracted into dichloromethane(s). Other results have shown a
drop in the amount extractable from 35$ to 205? when the collection temperature
of diluted exhaust was increased from ?5°C to 100°C (9). The U.S.
Environmental Protection Agency has set a maximum of 52°C for the collection
of particulate matter. This temperature is sufficient to condense or sorb
most of the high molecular weight organic molecules onto the carbonaceous
particles(7)• Approximately one-third of the total organic molecules in the
gas phase are thought to be sorbed onto the particles. These compounds
include polynuclear aromatic compounds which may be the crigin of ti.e
biological activity(lO).
The transformation of the collected particulate matter varies widely.
Factors such as sedimentation, diffusion to the walls, evaporation,
condensation, collision, gain or loss by surface charge, liquid and gas phase
reactions, catalysis, -surface adsorption, and photochemical reactions will
change the chemical and physical properties of the particulate matter(6).
EXPERIMENTAL.
Engine and Dynamometer
The engine used in this study was an AVCO-Lycoming Bernard W-51 industrial
Diesel engine. It is a small-single cylinder, four-stroke cycle, air-cooled,
direct-injected Diesel engine capable of 3000 revolutions per minute (RPM).
Table 1 summarizes some of the engine parameters.
10
-------
Table 1. Engine Parameters
Engine Speed
Rack
Air Inlet Pressure
Lubricant Temperature
Injection Timing
Filter Temperature
Displacement
Compression Ratio
Bore
Stroke
Rated Power
Frictional Power
2400 +_ 25 RPM
Full
760 _+ 5 torr
61 +_ 2°C
27° BTDC
51 _+ 2°C
355.6 mL
18:1
7.62 cm
7.78 cm
4.47 bktf at 3000 RPM
2.14 fkW at 2400 RPM
This engine is directly coupled to a General Electric type TLC-6 cradled
electric dynamometer which loads and motors the engine. The dynamometer is
used to start the engine and to measure the frictional power. Engine torque
output was calculated from the dynamometer balance-beam which has a scale with
0.05 kg subdivisions. The dynamometer equation is as follows:
1.0 ckW = (Scale Unit. RPM)/8046.
Pressure Measurements
A Kristal 601A piezoelectric pressure transducer, mounted in the cylinder
head of the engine, monitors the pressure and the output is amplified by a
Kiag Swiss Type 5002 charge amplifier. The data are stored on a Nicolet
Series 2090 model 206 Explorer digital oscilloscope with a model III magnetic
disc memory unit.
11
-------
Fuel
Since Diesel fuel is manufactured from various crude oils by complex
refining processes, a 1:1 mixture by volume of 2,2,4-trimethyipentane and
n-tetradecane has been substituted as a model fuel. This mixture has a
measured cetane number of 53.4 and this mixture of two pure hydrocarbons was
expected to reduce the number of species in the exhaust and thereby simplify
its analysis.
To prevent'possible contamination of the combustion products with
lubricant, polyalkyleneglycol (Ucon LB525 Union Carbide) was used to lubricate
the engine. Since this synthetic oil starts to decompose above 71°C, a
retrofitted oil cooler was used to maintain an oil operating temperature of
60° C.
Exhaust Emission
In this study both the exhaust particles and gases were of interest and
this section describes the hardware and instruments used to monitor and
collect the exhaust.
Gas Phase
The gases and particles were continuously sampled via a stainless steel
sample probe in the exhaust pipe. The sample probe is shown in Figure 1, and
the gas-phase instrumentation is shown in Figure 2. The following exhaust
gases were measured: carbon dioxide, carbon monoxide, oxygen, oxides of
nitrogen, and unburned hydrocarbons. These gases were measured as described
below.
The oxides of nitrogen were measured by a heated chemiluminescent analyzer
(Bectean model 955) calibrated with a 550 ppm NO span gas. The oxygen was
monitored with an amperometric sensor, (Beckman model 741) calibrated with
air. The carbon monoxide and carbon dioxide were
12
-------
Figure 1. Exhaust System
Porticuiaie -
Fitter
Vacuum
Pump
Engine
Exhaust
Manifold
Heal Exchanger-
J L.
] —Temperature Sensor
To Gas
Phase
Exhaust Stack
Exhaust Probe-
Exhaust
Mixing Tank
Inclined Tube
Manometer
13
-------
Figure 2. Gas Phase Instrumentation
Row Engine "I
Exhoust
Cold Trop
Filter
Zero and Calibrating
Gases
-------
monitored by. infrared spectrometers (Beckman type 864 and IR-15A) calibrated
with 415 ppm CO and 14.9$ CO^ calibration gases. The unburned hydrocarbons
are analyzed with an unheated flame ionization detector (Beckman type 109)
calibrated with 530 ppm methane in nitrogen. The line to this latter
in3tmment was heated.
•Particulate Matter
Particles were collected on Teflon-coated, glass-fiber filters (Pallflex
type T60A20). These filters are expected to collect at least 90* of Diesel
particles and were contained in a stainless steel sample holder. The stainless
steel sample probe (10.7mm id) is located in the exhaust stack (40.9nm id).
The flow is turbulent at all locations in the exhaust stack and therefore a
representative sample is ensured(ll).
Since the raw exhaust is not diluted with air, the heat exchanger was used
to cool the exhaust to 51«7°C. The exhaust gas temperature drops
approximately 100°C in the heat exchanger, thereby condensing high molecular
weight compounds onto the carbon particles.
The difference in total pressure between the sample probe and the exhaust
stack was monitored continuously. The static pressure difference indicates
the differential dynamic pressure. The valve on the vacuum pump was adjusted
to ensure that a constant pressure drop (and thus velocity and flow rate) was
maintained in the sample line during filter loading. Approximately 17$ of the
combustion products were sampled through the prob.=>.
PROCEDURES
The object of this study is to discern the effect of various inlet
oxidizers upon the particulate matter. For all conditions the engine was
operated at full rack, 2400 RPM, with the 1:1 fuel mix of n-tetradecane and
2,2,4-triae thyIpentane.
15
-------
Nitrogen-Free Oxidant
The nitrogen-free oxidizer consisted of a mixture of gases (21% oxygen,
argon, and ^6% carbon dioxide) which simulated air in the Diesel
combustion cycle. It was assumed that the volumetric efficiency of the engine
is independent of gas composition and a calibrated intake rotameter was used
to ensure that a flow equal to that of air entered the engine.
One problem with this gas mixture is that under pressure the carbon
dioxide liquifies and settles to the bottom of the cylinders. As a result, a
gas mixture which is rich in argon and oxygen is produced that does not have
the desired thermodynamic properties. This problem was solved by agitating
the cylinders, wrapping them with heating tapes, and heating above >1°C, the
critical temperature of carbon dioxide.
Another drawback of this mixture is that the high concentration of carbon
dioxide increases the ignition delay to such an extent that the engine did not
run. When the injection timing was advanced it caused the engine to knock and
to operate roughly. This problem was rectified by the addition of oxygen into
the intake manifold until the engine ran smoothly. The oxygen which was added
was approximately 10^ of the baseline-air flow and the flow of the
nitrogen-free gas was 90$ of the baseline-air flow. Therefore, the actual
gas, which entered the engine, was maintained at standard conditions, and had
a composition of 28.92 02, 20.7$ Ar, and 50.4? C02-
Air Supplemented with Nitrogen or Oxygen
The effects of excess'Oxygen and nitrogen upon the chemistry of the
particulate natter were also of interest and these studies were performed by
supplementing the engine with excess oxygen or nitrogen. Since only 5% and
10^ of additional nitrogen or oxygen were used, negligible effects on the
engine or its operating conditions were observed.
16
-------
In supplementing the air with nitroger the amount of oxygen was held
constant, so the fuel to oxygen ratio was the same as the baseline condition,
but, the mass of the diluent (nitrogen) increases. Supercharging the engine
with 5£ and 10# nitrogen results in the intake charges being composed of 20.0£
and 19-1^ oxygen by volume, respectively. Since the ratio of specific heats
for nitrogen is 1.4, the compression process remains constant although
increases in the intake manifold pressure of 0.031 and 0.062 bar, respectively,
were measured.
When the air was supplemented with 5% or 10? oxygen the flow rate of
nitrogen remains constant and the composition intake charge is 24.8$ and 28.2?
oxygen, respectively. The ratio of specific heats for oxygen is also about
1.4. The test conditions are outlined in Table 2.
Filters
Exhaust gas was sampled for 10 to 25 minutes, depending upon the loading
rate. The optimum mass of particulate matter on each filter was found to be
in the range of 0.1 to 0.2g since after a loading of 0.2g the pressure drop
across the filter becomes too large. The filters were placed in a desiccator
for 24 hours before they were weighed. Also the filters were protected from
light tc prevent photochemical reactions and refrigerated to prevent any
reactions until they were extracted and analyzed.
The organic adsorbates were removed by the Scxhlet extraction with
dichloromethane for 24 hours at a rate of two cycles per hour. The majority
of the solvent was removed by a rotary evaporator and the extract was blown to
final dryness with dry nitrogen or argon. Finally, the extract was sealed in
a vial, protected from light, and refrigerated.
17
-------
Table 2. Test Conditions
Test RPM
1
2*
3
4
5
6
* Test 2
The
2400
2400
2400
2400
2400
2400
Load
Full
Pull
Pull
Full
Full
Full
was later changed to
composition was then
Inlet Oxidizer Composition
Air
N2-Free
Air+5? N2
Air+10? N2
Air+5? 02
Air+10? 02
90? N2-free
28.9? 02, 20.
21.0? 02,
21.0? 02,
20.0? 02,
19.1? 02,
24.8? 02,
28.2? 02,
+ 10? 02
7? Ar, 50.4?
79.0? N2
23-0? Ar 56.0? C02
80.0? N2
80.9? N2
75.2? N2
71.8? N2
C02
RESULTS
INTRODUCTION
The experimental results of this part of the study are presented in the
following sections. Table 3 lists the various paraneters for the different
intake oxidants
Engine Performance
The reduced power output for the nitrogen-free oxidizer compare^ to tha*
for air was attributed to the delayed ignition. The large concentration of
carbon dioxide delays the ignition by interfering with the chain propagating
radicals that initiate the combustion process.
No attempt was made to adjust the injection timing to compensate for the
increased ignition delay experienced with the nitrogen-free oxidizer. As a
18
-------
Table 3. Engine Data
Oxidizer 00
Parameter 19
NO (g/IkW-hr)
NO (g/IkW-hr)
A
CO(g/IkW-hr)
HC(g/IkW-hr)
Particuiate
(g/IkW-hr)
Extract
(mg/IkW-hr)
SOP '(%}
Fuel Hate
(kg/hr)
BSFC (kg/IkW-hr)
Ignition Pelay
Peak Pressure
(Bar)
Exhaust (°C)
Temperature
Head (°C)
Temperature
Power (BkW)
Inlet (Har)
PreiiMiiri;
•11
0.72
1.03
2.74
2.78
1.69
86.5
5-12
0.96
0.37
20.7
61.2
444.4
168.3
2.58
1.07
80.0% N2
20.0$ 02
1.48
1-93
1.98
1.65
1.49
49.6
3-33
0.95
0-37
20.0
59-4
472.2
175.6
2.54
1.04
79.0? N2
21.0* 02
2.13
2.73
1.96
1.85
1.31
44.6
3-43
0.95
0.39
19.6
54.5
491.1
182.8
2.42
1.01
75.2* N2
24.8?; o2
4.17
5.27
0.42
0.56
0.61
51.7
8.48
0.95
0.38
18.6
62.7
49^-3
184.4
2.53
1.04
71.4* N2
28.65? 0?
7.05
8.06
0.15
0.23
0.30
57.7
19.22
0.96
0.38
17.0
59.7
470.0
186.7
2.^3
1.07
28. <&, 0?
20.7* Ar
0
0
1.8R
0.72
0.41
14.2
. 34.68
1.0
0.47
24.5
49.?
5R7.2
196.1
?.14
1.01
-------
result a large concentration of supplemental oxygen was required to start the
engine and it fired several seconds after being motored at the operating
speed. This permitted a large quantity of fuel to accumulate in the piston
before violent ignition occurred. Two pistons and ring sets failed prior to
developing a safe starting procedure.
Pressure traces for some of the different inlet oxidizers are shown in
Figure 3t and peak pressures are listed in Table 3. Some of the traces in
which the oxygen concentration is below 215? show high peak pressures. Since
the fuel and oxidant have a longer time to mix and accumulate during the
increased ignition-delay period, the combustion process approaches constant-
volume combustion, and the resulting maximum pressures may be high.
Gas Phase Emissions
The gases which were analyzed in the exhaust are as follows: oxides of
nitrogen, carbon monoxide, carbon dioxide, oxygen, and unburr.ed hydrocarbons.
The pollutants regulated by the U.S. Environmental Protection Agency are the
oxides of nitrogen, carbon monoxide and unburned hydrocarbons. The data are
presented on an indicated power-specific mass-flow basis (grams per indicated
kilowatt-hour). This emission rate was chosen in order to standardize the
mass of pollutant; with respect to the power. Since this engine is snail, a
large portion of the fuel energy is used for frictional work. Indicated power
was used to allow a closer correlation with other engines.
Carbon Monoxide and Unburned Hydrocarbons
Unburned hydrocarbon and carbon monoxide emissions are shown in Figure A.
These two pollutants result from incomplete combustion, and the trends were
the same for each of them. As the oxygen concentration increases, the two
pollutants decrease. Lean combustion accounts for this effect, since th
-------
Figure 3- Pressure Traces
.2390, , .207Ar , .SOiCO
40 20 TDC 20
Oei>r»es Crank Ansle fC,\)
21
-------
Figure 4. Carbon Monoxide and Unburnt Hydrocarbons as a function of Oxygen
80
Percent Nitrogen in the Intake Charge
78 . 76 74 72
j:
o o
- o
.'0 22 2i Jo
Percent Oxygen in the Int.ikc Charge
22
-------
Oxides of Nitrogen
Figure 5 shows the trend in nitric oxide and nitrogen dioxide emissions
with increasing oxygen concentrations. As the percentage of oxygen in the
intake charge increases, nitric oxide increases linearly, and nitrogen dioxide
reaches a maximum value near 25% oxygen. The Zeldovich mechanism as well as
simple chemical kinetics predict these increases.
When excessive oxygen concentrations are used the combustion resembles a
constant-pressure process and the resulting peak pressures and temperatures
decrease so that the oxidation of nitrogen is discouraged. However, the high
oxygen concentrations favor the oxidation of nitrogen since there is more
oxygen available for the reaction. The result is two opposing phenomena.
Since nitric oxide increases linearly with the oxygen concentrations the
dominant factor in forming nitric oxide must be the oxygen concentration. It
was found that nitrogen dioxide reaches a maximum value at the highest exhaust
temperature and peak pressure. As the oxygen concentration further increases,
the exhaust temperature and peak pressure decline and so does the nitrogen
dioxide concentration. Thus, the temperature and pressure determine the
formation of nitrogen dioxide at high oxygen concentrations.
Particulate Emissions
As the oxygen concentrations in the intake charge increase, the particle
mass-emission-rate decreases sharply as shown in Figure 6. Since more oxygen
is available to react with the carbon, fewer fuel-rich areas exist when the
fuel is pyrolyzed. Also, the ignition delay is shortened by higher oxygen
concentrations so that the fuel has more tine to be oxidized and less time to
pyrolyze.
The mass of organic adsorbate on the particulate matter was of concern
because it contains biologically active compounds. The mass of the adsorbate
increased with increasing oxygen in the intake charge. The largo quantity of
organic adsorbate that was extracted from the nitrogen-free particulate matter
was formed from lubricant which was pumped into the upper cylinder during the
long starting cycle.
23
-------
Figure 5« Nitric oxide and Nitrogen dioxide as a function of Oxygen
Percent Nitrogen in Che Intake Charge
78 76 74
72
i
22 ?4 26
Percent Oxygen in the Intake Charge
24
-------
Figure 6. Ignition Delay, Particulate, and Extract as a function of Oxygen
Percent Mttrogen In the Intake Charge
80 78 76 74
72
to
•—' r-t
II
U
a c,
** •
D «-t
I
I
."0 22 2i 26
Percent Oxygen it\ the Intake Charge
-------
B. CHEMISTRY
1. CHEMICAL ANALYSIS OF ORGANIC ADSORBATES.
INTRODUCTION
Polynuclear aromatic compounds (PAC), and unburned fuel and lubricants
constitute the major components of the soluble organic fraction of Diesel
particulate matter. PAC include polynuclear aromatic hydrocarbons (PAH), as
well as substituted and heterocyclic derivatives of PAH's. An understanding
of the overall reactions leading to the formation of the PAC, and their
environmental fate following release from the engine, is a requisite for a
complete understanding of the health implications of Diesel engines.
The major objective of this part of the research was to provide a better
understanding of the combustion process leading to the formation of the
particle-adsorbed polynuclear aromatic compounds. Chemical characterization
of the simplified soluble organic fraction, generated from a model aliphatic
hydrocarbon fuel, was used to attain this goal.
Unlike the analysis of the soluble organic fraction generated from a full-
boiling range fuel, the fuel simplifications introduced in this project have
reduced the complexity to an extent that permits the separation and analysis
r^
of the entire SOF in a single pass. Capillary gas chromatography (GC ) and
high performance liquid chromatography (HPLC) were chosen as complementary
separation techniques for the analyses. The effluent from the chromatographic
columns were monitored by positive and negative chemical ionization mass
spectrometry (CIMS), in addition to the standard detection methods for each
separation technique. These methods have proven to be extremely useful for
the separation and identification of trace organic mixtures.
-------
EXPERIMENTAL
Capillary Gas Chromatography
Gas chromatographic separations were carried out on a Varian Aerograph
Model 3700 Gas Chromatograph equipped with the requisite injector, detector
and pneumatic options for capillary column operation. Separations utilizing
<2
flame ionization detection served as the basis for the subsequent GC /MS
experiments, and the retention data were used for identifying the solutes.
Nitrogen selective detection was employed for comparison of the nitrogen
containing compounds from the air and the nitrogen-free oxidant.
The columns used in this study were 15pm or 30pm wail-coated ope'h-tubular
(VCOT) (0.9mm o.d. x 0.25 mm i.d.) borosilicate glass capillary columns, coated
with a 30 un film thickness of SE-54 stationary phase. The column was butt-end
coupled to the injector insert with a 0.8 mm i.d. Vespel ferrule, providing a
zero dead-volume fitting. Care was exercised to ensure that the column was
cut at a 90 angle, and that pieces of septa were removed from the injector,
in order to eliminate possible adsorption losses for higher boiling components
of the samples.
The same flame tip was utilized for the flame ionization detector and
thermionic specific detector experiments. However, the latter detector
required changing the tower so that a ceramic-bead containing an alkaline
earth salt could be positioned over the flame tip.
Ultra nigh purity.nitrogen (99-999^) was initially used as the carrier gas
2 2
to obtain adequate separations for the GC -FID and GC -TSD experiments,
2
and as a charge-exchange gas for the GC /CIMS experiments. In subsequent
studies high purity helium was used instead of nitrogen, in order to reduce
analyses tines. Injector and detector temperatures for all analyses were
270°C and 320°C, respectively. The column oven was temperature programmed
from 100°C, with an initial hold of 1 min, to 300°C, with a final holT of
5 nin. The temperature programming rate was 2°C/min for the helium
27
-------
carrier gas. Injections of 1.0 yL aliquots of the samples were made, and the
detector responses were recorded simultaneously on a strip-chart recorder and
by computer data acquisition.
The computer used throughout these studies was a MODCOMP 11/25
mini-computeri equipped with 64 kilo bytes of memory, direct memory processor,
interval timer, three 1.3 mega byte moving-head disc drives, and a 9-track
magnetic tape unit. Additional peripherals include a Tektronix Model 4662
digital plotter, Texas Instruments,Omni 800 line printer, and Tektronix Model
1006 graphics display unit. The computer program gas chromatography (GASCHB)
was used for the gas chromatographic analysis, and for the acquisition,
manipulation, and presentation of the chromatographic data. Chromatographic
data were stored on magnetic tape.
Capillary Gas Chromatography/Mass Spectrometry
Positive and negative ion chemical ionization mass spectrometry were used
for the.-identification of the capillary gas chromatographic effluents and
these results were confirmed by capillary gas chromatography/electron impact
mass spectrometry.
The capillary gas chromatograph, which was used for the previous studies,
was interfaced to a chemical ionization mass spectrometer via a heated fused
silica interface. A Scientific Research Instruments Corporation BIOSPECT
chemical ionization mass spectrometer was used for subsequent investigations.
This system includes a chemical ionization source, quadrupole mass filter and
a continuous dynode electron multiplier mounted off-axis to the quadrupole
filter. The multiplier was modified to allow negative ion detection by
grounding the entrance plate, and biasing the ion beam deflector so that it
operated as a conversion dynode. In this manner, the multiplier could be
operated with the same configuration in both positive and negative ion modes.
The bias voltage used for negative ion detection was supplied by a Keithley
Model 246 high voltage supply. The electrometer output was coupled to a
Tektronix D10 single beam oscilloscope for real time display, and to a
28
-------
computer interface for data acquisition. This study utilized the extensive
computer software package, entitled program CAD, which was developed by our
research group(l2). The program is constructed in overlay format to conserve
computer core, and allows for real time data acqusition and display of mass
spectra, as well as manipulation, presentation, storage, and retrieval of
collected data. Data for the GC /MS experiments were collected in the CP.MS
mode, in which the quadrupole mass filter is scanned at a software designated
rate, with a chosen delay between each scan, and each mass spectrum is saved
separately. The greatest number of mass spectra that can be accommodated by
the disc storage with the present software system is 400 scans. Because
capillary GC peaks are typically less than 10 sec in duration, and the
chromatographic analysis takes at least 30 min, the limitation of 4-00
scans/experiment would seriously reduce the chromatographic resolution of the
GC /CIMS experiments. A modification of the data acquisition overlay COLECT
has alleviated this problem, by use of a threshold test. The operator
specifies a threshold at the outset of each experiment, and only those mass
spectral scans which contain at least one point whose intensity exceeds the
threshold are saved. In this manner, only those scans which occur during the
elution of a chromatographic peak should be saved. Appropriate modifications
were made to the plotting, storage, and retrieval overlays to accommodate the
data from the threshold experiments.
9
Identical chromatographic conditions for the GC"/CIMS experiments were
used in order that the data cuuld be compared. The mass scale was calibrated
with the ions from methane and methyl stearate. The collection of mass
spectral data was initiated following elution of the solvent and reatcration
of the normal operating pressure (2 x 10"' torr). Mass spectra were
collected in the range from m/z 110 to m/3 310 at a rate of 75 scans/min for
the duration of the chromatographic run, using an empirically derived
threshold. The mass spectral data were stored on magnetic tape.
GC2/E;>jrS data were obtained using a Finnigan Model QWA-20/30B GC?/Ko
System and a Hewlett-Packard Model 5985 GC /MS System. Both systems were
equipped with splitless injectors, and 30um fused-silica columns coated with
29
-------
SE-54- Column effluents were directed into the MS source through fused-silica
interfaces. Both mass spectrometers used quadrupole mass filters and the mass
spectra were collected in the range 1 to 500 m/z. Associated computer
hardware and software were used to acquire the data and to perform library
searches based on the National Bureau of Standards library of 25,000 mass
spectra.
High Performance Liquid Chromatography
During the initial stages of this research, soluble organic extract samples
\:ere separated using a solvent extraction scheme to remove the acidic and basic
fractions from the neutral portion of the extract. The neutral fraction, which
comprises greater than 90 percent of the mass of the extract, was separated by
column chromatography on silica gel into paraffinic, aromatic, transitional
and oxygenated fractions. Iwo difficulties which became apparent, were poor
separation reproducibility and the requirement that the entire sample be used
for the separation. Because the number of sample transfers was large, and the
sample size typically less than 1 mg, contamination of the samples often
resulted. As a result the impurities often dominated the subsequent mass
spectrometric identifications. The fractions that were collected were
characterized by chemical ionization mass spectrometry via a solids probe
inlet. These early experiments indicated that the standard solids probe was
unable to generate repeatable, rapid evolution of the samples. These
difficulties led to the evaluation of HPLC methods for the separation of the
extract, and prompted the development of a temperature-programmable,
platinum-wire probe for the MS characterizations.
HPLC separations of the extract were accomplished using a gradient elution
system, and UV absorption detection, followed by collection of the effluent for
subsequent mass spectrometric identifications. A microprocessor-controlled
Varian Model 5000 Liquid Chromatograph equipped with a 6-port Valco Model
AH-60 sample injector and a 254 nm UV absorption detector, was used for the
separations. The separation was performed on a 4.6 mm x 25 cm column, pact-red
with 5 ;;ta silica (Spherisorb-S5W) preceded by a 4 mm x 5 cm guard column,
packed with HC Pellosil. All the solvents used in the HPLC separations were
"Distilled in Glass".
50
-------
The gradient program was initiated from the microprocessor following
injection of 25 uL aliquots of the extract. Solvent flow rates were
maintained at 1.0 mL/min. The initial mobile phase of 100^ n-hexane was held
for 10 min, followed by a linear introduction of dichloromethane over the next
10 min. The mobile phase was held for 10 min at 100$ dichloromethane.
Acetonitrile was then introduced over the next 5 min, and held at 10C$ for the
remaining 5 cin of the separation run. This 40 minute program was followed by
a reverse gradient to 100/£ n-hexane, which was held for 30 Ein in order to
condition the column before the next sample. Sample effluents were monitored
with a 254 nm, fixed-wavelength UV absorption detector, and the signal WES
recorded on a strip-chart recorder. Sixteen fractions were collected manually
during the course of each 40 min experiment and were allowed to evaporate
overnight at room temperature in the dark.
Mass Spectral Analysis of _HPLC Fractions
Positive and negative ion, chemical ionization mass spectrometry was used
to identify the components contained in the collected fractions. It has been
demonstrated by other workers that rapid heating of the sample in the source
of the mass spectrometer favors evaporation ever competitive decomposition
reactions, and they have been able to obtain mass spectra of thermally labile
species(l3~15)« In light of these findings, a platinum-wire prota was
developed for the introduction and rapid evaporation of the sample.
This temperature-programmable probe is capable of scanning linearly a wide
temperature range under computer control. The present design allows the
temperature of the probe to be raised from source-block temperatures to
700°C, at rates of up to 500°C/sec. The temperature of the probe tip is
increased step-wise and the mass spectral and temperature data are collected
at each step. The probe consists of a 0.255 nm o.d. x 2 cm platinum wire
attached to 1-5 mm o.d. nickel leads, which are fed through a modified 6 mm
o.d. solids probe. The platinum filament serves as both a semple heater and
as a platinum resistance thermometer, allowing accurate sample temperatures to
be calculated.
31
-------
Figure 7 is a block diagram of the electronics which enable the probe to
be operated under on-line computer control. This controller provides accurate
measurements of the voltage across the filament and the current through the
filament, allowing the resistance to be calculated at each sampling point by
the application of Ohm's law. The cross-sectional area of the platinum wire
is known, and the temperature coefficient of resistivity and resistivity at
0°C have been determined by other workers(l6). Provided that the length.of
the filament is known, these data can be used to calculate the temperature of
the filament at each step.
The length of the platinum filament is exceedingly difficult to measure
mechanically, since it is dependent upon the point of contact with the nickel
leads. The length was therefore calibrated by repeatedly scanning the input
voltage from 0.000 V to 0.125 V, and measuring the current, while the probe
tip was immersed in an ice bath. The thermal capacity of the ice bath was
sufficient to maintain a constant temperature for the filament over the range
of input voltages. Ohm's law was used to calculate the resistance at each
point, and a linear least-squares fit of the calculated resistances yielded
the resistance at 0°C. This was used, in conjunction with the resistivity
at 0°C, to calculate the length of the wire. Calibration data were stored
on magnetic tape. Overlays £alibrate ^temperature _p_robe (CTPROB) and
manipulate Jjemperature £iie (MTFILE) have been added to program CAD to
accomplish these tasks. Overlay y_iew _t_emperature data (VTDATA) was also added
for the presentation of temperature profiles and for the calculation of
linear-least-squares slopes for the temperature data. The calibration was
checked on a daily basis over a one month period and found to be constant.
Evaporated HPLC fractions were dissolved in 20 uL of dichloromethane, and
a 5 vL aliquot placed upon the probe tip. The sample was allowed to air dry
for 2 min before inserting the probe into the mass spectrometer inlet. Upon
insertion, mass spectral scanning commenced from m/z 110 to m/z 310, and 75
scans were collected during a.1 min experiment. Mass spectral data were
collected using program CAD. Modifications were made to overlay COLECT which
allow the operator to input the starting voltage, and up to 5 linear ranges of
-------
Figure 7. Block Diagram of Temperature - Programmable Platinum Wire Probe
Inout Voltage
Multiplication
and Regulation
Probe Current
Multiplication
Precision
Scaling
Potentiometer
Digltal-Analoz
Converter
Sample-and-hoL
Amplifier
SI Wire
Pt Filament
\nalog-Digital
Input Output Interface
Central
Process!n^
fnit
Direct
Memo ry
Processor
-------
voltages for the platinum filament, for each experiment. The probe filament
was maintained at 0.001 V for the first 10 scans in order to minimize
evolution of the sample before the probe was seated in the source block. The
voltage of the probe was then linearly increased from 0.001 V to 0.500 V over
the remaining 65 scans. Since the source block was maintained at 150°C, a
temperature ramp of 150°C to approximately 600°C was used in these
experiments. This ramp allows the highly polar and involatile components of
•' *
the samples to be rapidly vaporized, usually yielding intact molecular ions.
Both positive and negative chemical ionization spectra were collected for each
sample. The mass spectral and temperature data were stored on magnetic tape.
High performance Liquid Chromatographic Procedure for the Selective
Determination of 1-Hitropyrene
A selective ultratrace analytical procedure which was developed by Dr. S.
Tejada of the U.S. Environmental Protection Agency was used to quantify
l-nitropyrene(l7)• This procedure is based on isocratic reverse phase (water:
methanol 80:20) high performance liquid chromatography with fluorescence
detection (360 nm excitation, -130 run emission). The essential feature of this
method involved placing a reduction column (2 -cm, piatinuni-rhodiun catalyst on
silica gel), maintained at 60-80°C, between two reverse phase
octadecylsilane (ODS, 15 cm) columns. The first ODS column produced the
initial separation and after 12 min the colunn effluent was passed onto the
reduction column which reduced the nitro-aromatic compounds to
amino-aronatics. The effluent from the reduction column passed onto the
second ODS column. After 22 min. the reduction column and the first ODS
column were switched out of the system and the separation was continued on the
second ODS colunn. The 1-nitropyrene eluted after 27 min and was quantified
by the fluorescence detector. After A5 min the entire system was programmed
through a clean-up cycle.
34
-------
RESULTS
Analysis of Organic Adsorbate by gas or liquid Chromatography -
Mass Spectrometry
The extract from four air oxidizer and four nitrogen-free oxidizer
collection runs were analyzed in these experiments. The gas chromatographic
results, which were found to be the most sensitive means of fingerprinting the
samples, demonstrated that there was excellent intrasample repeatability
(correlation >0.95) and intersample similarity for the two sets of samples.
These results show the reproduciMlity of the engine and fuel system used in
this study.
Gas chromatographic retention data for the air oxidizer camples, based
upon flame ionization detection fesults (Figure 8), and chromatographic data
for standard compounds generated under the same conditions, were extremely
useful for differentiating between various isomers of the polynuclear aromatic
compounds. Retention indices presented in Table 4 were calculated using the
system developed by Lee and coworkers(lS). Peak areas were integrated after
computer-assisted baseline subtraction had been performed. Average emissions
presented in Table 4 in terms of yg/g particle and mg/kg fuel were calculated
using a response factor for multiple injection of a standard solution of
phenanthrene, calculated particle emission rates, and fuel consumption rates.
Some of the minor peaks, which were not observed by capillary gas
chromatography-electron impact mass spectrometry, could not be identified on
the basis of their retention indices and their chemical ionization mass
spectra. However, for completeness these compounds have been included in
Table 4 with their nominal molecular ions.
The results for gas chromatographic separation of the nitrogen-free
oxidizer samples, using flame ionization detection, are presented in Figure
9. The striking difference between these samples and the air oxidizer samples
was the presence of relatively broad peaks throughout the chronatograms for
-------
Figure 8. Capillary Gas Chromatogram of Air Oxidizer Sample using Flame
lonization Detection
90.-'
80.-
co
£70.-
§60.-
HI
5 50.-
J2 40 .-
UJ
x 30.-
V
LJ 2Q . -
CL
1. 0.-
Q-
1
j
1
u
2
1 1
n
jjjj,
1 2
JUUjluAJLj-
i
^
3
3
LjuJJU
1 1 1
.
3
0
32
I
Ju
4
35
J.- - .1 .1 -_ I.
j : i ( : .
0. 10. 20. 30. 40. 50. 60. 70.
RETENTION TIMECMIN)
90. 100
Peaks correspond to Table 4.
-------
Figure 9« Capillary Gas Chromatogram of Nitrogen-free Oxidizer sample using
Flame lonization Detection
10.
15. 23. 25. 30
RETENTION TIMECMIN3
35.
40.
i
45
Peaks correspond to Table 4.
37
-------
Table 4. Chemical Characterization of Organic Adsorbates on Diesel
Particulate Matter
Ref.
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Compound Name
Unknown
Naphthalene
Benzofuran,
7-methyl-
Tr.iden-1-one,
2,3-dihydro-
Methyl-
naphthalene
Methyl-
naphthalene
Phthalate-
anhydride
Unknown
Unknown
Biphenyl
n-Te tradecane
Unknown
Unknown
1-Benzopyran-
2 -one
Biphenylene, or
Acenaphthylene
Acenaphthene
Unknown
Dibenzofuran
Unknown
Retention Average Emissions
.Indices ug/g particle mg/kg fuel
M SD
198.0,0.44
200.0,0.10
204.3,0.37
212.3,0.19
214.0,0.32
216.6,0.51
217.9,0.45
222.0,0-56
224.8,0.75
227.7,0.76
230.1,0.64
232.1,0.87
235-8,0.84
238.7,0.91
239-9,0.18
248.4,0.93
250.0,0.9-1
251.4,0.97
253.4,0.99
187.22
329.10
69.49
22.41
9-89
11.12
42.39
12.66
21.76
76.38
596.78
83-04
47.71
42.77
30.29
16.97
27.70
90.03
20.20
1.361
2.394
0.505
0.163
0.007
0.081
0.308
0.092
0.158
0.555
4.340
0.604
0.347
0.311
0.220
0.123
0.201
0.655
0,147
Method
of
Identification
123-PCI
a,b,d
b,d
b,d
a,b,d
a,b
b.c
147-POI
135-PCI
a,b,d
a,b,d
147-PCI
133-PCI
b,c,d
a,b,d
a,b
197-PCI
a , b , d
15"-PCI
38
-------
Table 4. continued.
Ref.
No.
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
Compound Name
Fluorene
Unknown
9-Fluorenone
Phenanthrene
Anthracene
Methyl
9-Fluorenone
Unknown
Methyl
9-Fluorenone
Unknown
Methyl
9-Fluorenone
Benzo[ c Jcinnoline
Fluorene Quinone
Phenanthrene
Quinone
Cyclopenta-
phenanthrene-5-one
Naphtho[l,8-cd]
pyran-1 , 3-dione-,
or Fluoranthrene
Pyrene
MethyJ.pyrene
Methylpyrene
Methylpyrene.
Retention
Indices
M .SD
266.6,0.55
271.8,0.56
292.6,0.21
300.0,0.11
301.6,0.19
306.7,0.16
308.6,0.98
309.7,0.24
311.0,0.26
312.8,0.10
319-8,0.54
327.3,0.33
330.3,0.61
341.7,0.82
343-7,0.69
351.4,0.97
365-9,0.21
372.9,0-25
375.3,0.92
Average Emissions
yg/g particle mg/kg fuel
21.46
39-30
232.03
582.44
29-58
21.08
21.29
- 15-13
20.35
209-31
35-18
181.74
501.67
345.57
309.75
8.82
6.04
22.21
0.156
0.286
1.688
4.236
0.215
0.153
0.155
0.110
0.148
1.522
0.256
1.322
3-6-lQ
2.513
2.253
0.064
0.044
0.162
Method
of
Identification
a,b
171-PCI.170-NI
a,b,c,d,e
a,b,d,e
a.b
c
198-NCI
c
198-NCI
c
a, b,d,e
b,e
b,c,e
b,c,e
b,c,d,e
a,b,e
a,b,d,e
a,b,e
a,b,e
a,b,e
-------
Table 4. continued.
Ref. Compound Name
No.
Retention Average Emissions Method
Indices ug/g particle mg/kg fuel of
M SD Identification
39
40
41
42
43
Benzo[ghi j
fluoranthene
Cyclopenteno[cd]
pyrene
Chrysene
Benzofluor-
anthene
Benzofluor-
anthene
389-4,1.13
394.6,0.17
400.3,1.80
436.1,3.55
443-4,0.25
64.08
74.32
35.75
16.75
19.09
0.466
0.541
0.260
0.122
0.138
a,,b,d,e
a,b,e
a,b,e
a,b,e
a,b,e
44 Benzo[ghi] e
perylene
45 Nitropyrene e
a—Capillary gas chromatographic retention index
b--Gas chromatography/positive chenlical ionization mass spectrometry
c.—Gas chromatography/negative chemical ionization mass spectrometry
d—Gas chromatography/electron impact mass spectrometry
e—High performance liquid chromatography- Temperature-programmed
chemical ionization mass spectrometry
Unknowns are identified by molecular ion observed either by positive (PCI)
or negative (NCI) chemical ionization mass spectrometry. .
40
-------
the nitrogen-free oxidizer. These peaks are attributed to an increased level
of lubricant in the combustion chamber, as a result of the start-up procedure
for the modified oxidizer experiments. However, the major peaks iii the
nitrogen-free samples were the polynuclear aromatic compounds. These
identifications were subsequently confirmed by gas chromatography-chemical
iouization mass spectrometry.
«•*
When the flame ionization detector was replaced by a thermionic specific
detector, which is a nitrogen selective detector, and the same samples
analyzed, a lower number of nitrogen containing compounds were observed for
the nitrogen-free samples compared to the air oxidant samples. This result is
consistent with the reduction of the oxides of nitrogen in the exhaust, from
an average of 470 ppm for the air oxidant to less than 1 ppn. for the
nitrogen-free oxidizer system. This indicates that the latter oxidiier system
w"-i successful in removing nitrogen from the combustion chamber. The low
levels of nitrogen-containing compounds in all samples, in relation to the
major components of the soluble organic fraction, precluded any positive
identifications based on mass spectrometry.
Positive chemical ionization mass spectrometric total ion chromatograms
showed a good correlation with the flame ionization results, and allowed the
assignment of molecular weights for many of the solutes. With the exception
of n-tetradecane, which showed a typical aliphatic hydrocarbon base peak at
(M-l) , each peak exhibited a mass spectral base peak at (M+l) resulting
from proton transfer. Extracted ion profiling was used to deconvolute many of
the irinor peaks from the total ion chromatogram.
Negative chemical ionization mass spectrometry was especially useful as a
selective tool for substantiating the presence of oxygenated derivatives of
the polynuclear aromatic compounds, since the major non-oxygenated
constituents of the samples did not produce ions. Base peaks in the mass
spectra were observed at M , corresponding to resonance electron capture by
the compounds containing the electronegative oxygen substituents.
Electron-impact mass spectra were used to confirm many of the
identifications using searches of the National Bureau of Standards library of
25iCOO mass spectra.
41
-------
Liquid chromatographic separations of the same samples provided unique
information, which was useful for the identifications, and Figure 10 shows a
typical chromatograra. Fluoranthene, cyclopentaphenanthrene-5-one, and
naptho[l,8-c,d]pyran-l,3-dione eluted in fractions 3, 8 and 14, respectively,
while the gas chromatographic separation of the same compounds yielded
retention indices which differed only by a single unit.
The temperature programmable solids probe developed in this study proved
to be extremely useful for characterizing the involatile components of the
soluble organic fraction. Many high molecular weight polynuclear aromatic
compounds, including isomers of benzofluoranthene and benzoperylene, were
identified by liquid chromatographic separation, followed by direct-probe
chemical ionization mass spectrometry. These high molecular weight compounds
were not eluted from the gas chromatographic column.
Nitropyrene was isolated and identified in fraction 6 of each of the air
oxidant samples, and one of the argon/oxygen samples, using the temperature-
programmable probe coupled with negative chemical ionization mass spectrometrie
detection. Concentrations were too low to quantify this compound in these
experiments. The presence of this compound in one of the nitrogen-free
samples is explained by the introduction of nitrogen into the combustion
chamber as a result of cracking a piston ring during the experimental run.
With this exception, the presence of nitropyrene only in the air oxidant
samples indicates that its formation must occur as a secondary process, and is
not dependent upon fuel-bound or lubricant-bound nitrogen.
The major compounds identified in. the soluble organic fraction of Diesel
particulate matter, generated from a two-component, aliphatic hydrocarbon fuel,
are polynuclear aromatic compounds. The compounds identified are quite similar
to these observed by investigators who employed full-boiling range Diesel
fuels(l9-2l). These results indicate that the fuel used in this study is a
good model of a full-boiling range fuel. The findings also imply that
polynuclear aromatic compounds found on Diesel particles are inherent products
of the diffusion-controlled combustion process, and are r.ot only' the result of
-------
Figure 10. High Performance Liquid Chromatogram of Air Oxidizer
Sample
5 MIN
8 ' 9 io'll ' 12 '13
1 1 ' 2
5 6 7
FRACTIOH fir-s
1515
43
-------
polynuclear aromatic compounds in the Diesel fuel persisting unchanged during
combustion. In addition, the similarities between the compounds identified in
this study and other investigations, where fuels included aromatic compounds,
iriicate that the polynuclear aromatic compounds must be formed from similar
intermediates.
Selective determination of 1-nitropyrene
The extracts of the particulate samples, which were collected from engines
operated on various oxidizers, were analyzed for 1-nitropyrene and the results
are shown in Table 5«
Table 5« Concentration of 1-Nitropyrene in the Extracts of Particulate
Matter Collected from Engines Operated on Various Oxidizers.
Oxidizer
Composition (%}
1-Nitropyrer.e Nitrogen Dioxidf
(pg/IkW-hr) (g/IkW-hr)
Pyrene Particulate
(ug/IkW-hr) Matter
(g/IkW-hr)
19.1,
20.0,
21.0,
24.8,
28.6,
50.4,
°2I
°2:
02;
02;
02;
C02;
20.
80.
60.
79-
75.
71.
9,
0,
0,
2,
4,
28.9,
7,
Ar.
N2.
N2.
N2.
N2.
N2.
02!
42.
23-
4.
32.
20.
0.
1
1
6
5
1
3
0.
0.
0.
1.
1.
0.
31
45
60
10
01
0
ND
TO
406 -
ND
ND
12C*
1.
1.
1.
0.
0.
0.
69
49
31
61
30
41
ND Not Determined;
estimated emission rate, accurate quantification impossible due to the
presence of lubricant in the sample.
44
-------
These data are difficult to reconcile with our hypothesis that the
emission rates of 1-nitropyrene follow the same trends as the emission rates
of nitrogen dioxide. This hypothesis is based on reaction studies which will
be discussed subsequently in this report. The data show that there are good
correlations between the 1-nitropyrene and the particulate matter which
suggests that these parameters are dependent. A possible explanation of this
relationship is that the 1-nitropyrene is produced by reaction between the
nitrogen dioxide and the pyrene sorbed on the particle. It may be that
1-nitropyrene is a sampling artifact i.e. it is produced by a reaction
•
occurring during sampling. The nitrogen-free oxidant produces negligible
quantities of 1-nitropyrene, which is consistent with levels of the nitrogen
dioxide belov the detection limit.
2. PHYSICOCHEMICAL PROPERTIES. OF DIESEL PARTICULARS MATTER.
a. STATIC METHODS.
INTRODUCTION
In order to ascertain the relative extrac-abilities of compounds with
potential adverse health effects, the physicochemical properties of the
adsorbate/absorcent complex were determined. Although some preliminary work
has been reported on the physical properties of Diesel particulate matter
(22-24) there has been relatively little research into the thermodynamic
properties of the adsorption process compared with numerous similar studies
on carbon blacks (25-27) and coals (28}. Physicochemical properties such as
density, surface area, and heats of adsorption will to a great extent
determine the nature of the sorbed compounds/particle surface system.
i/cterraination of such properties will provide a better understanding of the
surface characteristics and potential environmental significance of airborne
particulate matter.
-------
EXPERIMENTAL
Samples
The carbon black used as a reference sample in these experiments was
graphitized Spheron 6, a channel black that had been treated at 2700°C to
give a highly uniform surface with an approximate B.E.T. nitrogen surface area
of 80 m2/g.. Two Diesel particulate samples were used, the first DPM-PSU,
was collected from the engine on the model fuel and the second Diesel sample,
DPM-EPA, was collected from a 1978 Oldsmobile 350 engine by the U.S.
Environmental Protection Agency (runs #8511-8319) operated with a commercial
Diesel fuel and lubricant. This latter sample was collected after dilution.
Electron Microscopy
Electron micrographs were taken of the two Diesel sample? with a JEOL
transmission electron microscope at magnifications ranging from 10,000 to
50,000x. The microscope sample grids were prepared by suspending the
particles in a 2% volume mixture of collodion and amyl acetate and coating
this suspension onto copper mesh grids. Particle size measurements were made
manually and the mean particle diameters were calculated from measurements of
at least 100 particles using the following equation:
dA = nd3/ nd2
where n is the number of particles counted with diameter, d. The external
surface area was calculated by the following relation:
Sext = 6/>dA
assuming high and low densities, p, of 1.85 and 2.00 g/cm , after the method
of Anderson and Smmett (29).
-------
Densities and Adsorption Isotherms
A diagram of the apparatus used to measure helium densities and adsorption
isotherms is shown in Figure 11. The capacitance manometer and associated
electronics were obtained from MKS Instruments, Inc. The tubing and pressure
sensor were maintained at temperatures higher- than that of the sample cell so
that the saturation vapor pressure waa determined by the sample temperature.
Bulk densities of the solid samples were measured by weighing the quantity
of particles occupying a known volume. True (or helium) densities were
determined by expanding helium into the sample cell without and then with a
known mass of particles.
Surface areas were calculated from adsorption isotherms with the method of
Prunauer, Emmett, and Teller (B.E.T. ) (30,31). The adsorbates used wers
o o o o
nitrogen (16.2 A /nolecule), benzene (40 A /molecule), and n-hexane (51
A /molecule) (32). The samples were degassed for 15-24 hours at the
indicated "activation" temperatures (50°-400OC) in a vacuum of
approximately 10 torr prior to each adsorption experiment to remove
adsorbed materials from previous runs and to investigate the effect of varying
the activation conditions on the available surface area. The gas buret was
used to introduce, increasing amounts of nitrogen and a 2.0 cr 10.0 uL
pressure-tight syringe was used to introduce the liquid organic adsorbates.
Heats of adsorption were calculated from two isotherms at two different
temperatures using the Clausius-Clapeyror. relationship. Adsorption heats we're
determined for benzene and n-hexane on several samples which had been heat-.
treated to varying degrees.
RESULTS
Electron Microscopy
The results of the electron microscopic measurements appear in Table 6.
The average diameters of the DPM-EPA particles are less than that of DPM-PSU
47
-------
Figure 11. Volumetric Adsorption Apparatus
nl sampl
T~ celi
I
constant
temperature
bath
2 L
mercury
reservoir
1: Capacitance manciv.ecer
'2: Electronics unit
3: Cigical readout
4: Recorder
-------
Table 6. Electron Microscopy Results.
Sample Particle Diameter Surface Area
nm P-1.85 p»2.0
DPM-PSU
DPM-EPA
Grade 6 (29)
36
26.8
410
90.1
121.0
79-1
83.3
111.1
73.2
which is similar to that of a non-porous carbon black, Grade 6 (29). The
external surface areas of the commercial Diesel samples are consequently
greater. Particle diameters increase, and surface areas decrease, the greater
the extent of graphitization of carbon blacks (22). However, external surface
area can be deceiving if the solid is pcrous or heterogeneous. For this
reason, external surface areas must be compared to adsorption surface areas
(internal + external) to accurately characterize the surface.
Densities
The results of the bulk and helium density measurements are shown in
Table 7. The bulk densities of the Spheron and of some suspended airborne
particulate matter (33) are about 5 times greater than those of the Diesel
samples. The true densities of all of the samples are similar because of the
common carbon/hydrogen-based structure. The reported true density range for
Spheron is that of amorphous carbon since it is 99? carbon (3-O• The helium
density of DFM-PSU could not be accurately measured due to a small sample
size, but the true density must be in the range of 1.5 to 2.0 g/cm" since it
is 96? carbon (34). Bulk density will have the greatest significance with
respect to atmospheric residence times. Particles which have low tendencies
to agglomerate once emitted into the air and which have low bulk densities
will remain in the air longer and can be transported longer distances than
more'dense aggregates of particles which will settle out of the atmosphere
faster.
-------
Table 7. True and Bulk Densities.
Saople Bulk Density True Density
(g/cm3) (g/cm3)
Spheron 6 0.5 1.85-2.0
DMP-PSU 0.15
DPM-EPA 0.08 1.5-1.8
Atmospheric (33) 0.49-0.64 2.0-2.6
Diesel (23) 0.10 -
Adsorption Surface Areas
The measured BET adsorption surface areas are summarized in Table 8. The
major sources of srror were in the low pressure measurements and in the
cumulative error associated with successive injected volumes. Consequently,
the uncertainty in these values is approximately —55'.
The initial benzene adsorption on Spheron experiment yielded a surface
p
area close to the assumed value of 80 in /g. The isotherm was S-shaped and
of the Type II classification according to Brunauer, Deming, Denting, and
Teller (35). This type of isotherm has a characteristic "knee" followed by an
asymptotic rise to the saturation vapor pressure which is indicative of
monolayer followed by multilayer formation. All isotherms measured for DPM-PSU
were also of this type. An example of these isotherms is shown in Figure 12.
The adsorption of benzene on DPM-PSU after the initial nitrogen experiment
resulted in a decrease in the surface area over three successive runs. The
area was increased slightly with higher temperature activation but did not
reach the initial area measured. This observation could be indicative of
-------
Table 8. B.E.T. Surface Areas
Activation
Temperature
Adsorbent (GC)
Spheron 6 400
DPM-PSU 50
100
350
DPM-EPA 50
200
400
Sample
Temperature
Adsorbate (Oc)
Benzene
Nitrogen
Benzene
Benzene
Benzene
Hexane
Nitrogen
Benzene
Hexane
Nitrogen
Benzene
Nitrogen
Nitrogen
Benzene
Hexane
Nitrogen
Benzene
20-40
-196
20
20
20-40
20
-196
-16-5
-21-0
-196
0-23
-196
-196
0-20
0-20
-196
0-20
Surface
Area
(m2/g)
73.2
103.7
111.7
83.1
65.7
62.0
77.7
74.2
77.5
83.1
77.0
41.0
69.7
65-3
74.5
112.2
106.0
51
-------
Figure 12. Adsorption Isotherms of Benzene on TPM-PSU (350°C)
Q. OPM-PSU , ACT. TEMP. -358. C. SAMPLE TEMP.- 21 .0 C
A- OPM-PSU . ACT. TEMP. -358. C. SAMPLE TEMP.- 5.0 C
s\
>—
•z
UJ
§ 2.5-
o
0
o 2.0-
V.
UJ
£
i l .5-
Oi DPM-PSU . ACT. TEMP. -350. C. SAMPLE TEMP.- 8.0 C
0
D
,.
0
A
* ' D
1
UJ O A O
ro .
g 1.0-J 0 A D
ID
5 0.5-
o
o
o° * °
0 ° A D °
A* E 0 ° °
$£ ° °
10.0 20.0
40.0 -50.0
30.0
PRESSURE CTORR 5
60.2 70.0
52
-------
cracks or pores on the solid surface from which benzene could not be removed
with the treatments employed here. The adsorption of benzene irreversibly
modified the surface and the original nitrogen surface area should be taken as
the actual surface area.
In contrast to DPM-PSU, the surface area of DPM-EPA was strongly dependent
on the activation temperature. Multiple nitrogen adsorption isotherms for
DPM-EPA activated to 50°, 200°, and 400°C are shown in Figure 13. The
isotherm for the 50° C activation was calculated to have a BET C-value, a
measure of the net heat of adsorption (E - EL). of 47, indicating a
relatively low energy surface. The corresponding C-value for DPM-PSU was
greater than 200. Increasing the activation temperature to 200°C resulted
in a gain of approximately 30 m /g in area. Degassing DPM-EPA at 400°C
p
increased the surface area to 112 m /g. The weight loss of this sample in
the two high temperature outgassing treatments was approximately \1% of the
original mass. It is reasonable to assume that the final outgassing
effectively cleaned the surface of presorbed material since the weight loss
due to these "thermal" extractions is_close to the percentage extractable into
dichloromethane, 17.5? (36). Since DPM-PSU was only 1% extractable into
dichloromethane, it is understandable that degassing would have relatively
little effect on the available surface area. The accessibility of more
high-energy adsorption sites to nitrogen on the highly activated DPM-EPA is
reflected in an increased C-value, 47 (50°C) to 300 (400°C), and BET
surface area, 41 (50°C) to 112 (400°C). It seems likely that loss of
presorbed material from cracks and narrow pores occurs in the high temperature
activation and it is these newly unblocked sites that cause the increase in
area and in energy of adsorption.
Benzene and hexane adsorption on DPM-EPA confirmed these changes in
surface characteristics with increasing activation. Multiple benzene
adsorption isotherms on DPM-EPA are shown in Figure 1<1. The "unactivated"
sample, 50°C, yielded a Type III isotherm and a C-value between 1.5 and 2.5.
In such situations, the benzene-benzene interactions are more significant than
53
-------
Figure 13- Adsorption Isotherms of Nitrogen on DPM-EPA
UJ
a
Of
o
o
CO
g
t/>
g
Q. DPM-EPA > ACT. TEMP.- €0. C, SAMPLE TEMP.—195.8 C
A- DPM-EPA , ACT. TEMP.-2CO. C. SAMPLE TEMP.—195.8 C
*• DPM-EPA , ACT. TEMP.—430. C. SAMPLE TEMP. —195.8 C
2.0-
1 .5-1
t . 0 —
>_ 0.5-
§
g
3.10
0.20
0.30
0.43
8.53
0.63
RELATIVE VAPOR PRESSURECP/P03
-------
Figure 14. Adsorption Isotherms of Benzene on DPM-EPA
0«
DPM-EPA , ACT. TEMP.- SB. C, SAMPLE TEMP.- 0.0 C
DPM-EPA , ACT. TEMP.-283. C. SAMPLE TEMP.- 0.3 C
DPM-EPA . ACT. TEMP.-4e8. C. SAMPLE TEMP.- 0.S C
/•N
f-
UJ
8 1.2-
o
V)
"* 1.0-
UJ
o 0.8-
*«»
ui 0 . 6 —
o:
o
Q 0.4-
h-
O
z:
•<
o
o
o
D -
* Q
A
* D
O .
« ° ^
A°
o . _ a
* ° *
o A a
A a
0.10 0.20 0.30 0.40 0.50 0.63 0.70 0.80
RELATIVE VAPOR PRESSURECP/PQ3
• 55
-------
the benzene-surface interactions, and no distinct monolayer formation (knee)
is observed in the isotherm. It is generally thought that the BET theory does
not yield reliable surface areas in these systems (37,58).
The benzene isotherm for DPM-EPA activated to 200°C was Type II with a
barely perceptible knee and an increased C-value of between 4 and 6.
Consequently, these surface areas were close to those as measured with
nitrogen. Adsorption on the highly activated sample yielded a Type II
isotherm, also, with a C-value comparable to those of the DPM-PSU isotherms,
25 to 50.
The nitrogen BET adsorption surface areas and the external surface areas
are compared in Table 9- The external areas are the averages of the ranges
from Table 6. The roughness factor is the ratio of the BET area to the
external area and is often taken as a measure of the porosity of a solid
sample. The values for DPM-PSU in parentheses are those from the 350°C
activation compared with those of the initial 50°C experiment. The roughness
factors for both of the Diesel samples indicate little difference between the
two surface areas. Both are less than the roughness factors for Grade 6, a
nonporous rubber black, and Mogul, an ink black which is an example of a
porous carbon black with a relatively large internal surface area.
Table 9- E.M. vs. B.E.T. Nitrogen Surface Areas
Sample
DPM-PSU
DPM-EPA
Grade 6 (29)
Mogul (29)
E.M. (External)
Surface Area
(m2/g)
86.7
116.1
76.9
86'. 6
B.E.T. (Internal)
Surface Area
(m2/g)
103.7 (83.1)
11?. 2
li..
ASO
Roughness
Factor
1.2 (0.
0.Q7
1.43
. '• *.2
96)
56
-------
Heats of Adsorption
The variation of the heats of adsorption with surface coverage of benzene
on Spheron is shown in Figure 15- The results are compared with those of
Pierotti and Smallwood (39) and confirm the graphitic nature of the Spheron
surface. The heats rise only slightly as the monolayer is formed and then at
about 0.8 of one monolayer they decrease rapidly as second layer formation
begins. The relative constancy of the heats while the first layer is
completed is indicative of an energetically homogeneous surface.
The benzene and hexane heats of adsorption on DPM-PSU versus surface
coverage are shown in Figures 16 and 17, respectively. The zero coverage
isosteric heats, as determined by gas chromatography, have been included as
well as literature data for graphitized carbon blacks (32,39). The initial
benzene heats are much higher than those for the carbon black but the reverse
is true for n-hexane (32). This is reasonable considering the surface groups
(such as hydroxyls) on DPM-PSU to which benzene adsorption is sensitive and
n-hexane is not (32). The other notable difference between the DPM-PSU and
graphitized carbon black heats is the more gradual decrease in the heats past
the maximum. Isirikyan and Kiselev (4-0) showed that the less homogeneous a
surface is, i.e., lower graphitization temperature, the smoother is the heat
curve and the higher is the initial heat. The benzene initial gas
chromatographic heat is probably somewhat low, and is likely to be closer to
15 kcal/mole, considering the heterogeneity at low coverage and the affinity
demonstrated by the decrease in the surface area. Although the inhomogeneity
of the DPM-PSU surface caused high initial heats and a less rapid drop in the
heats, the heat of adsorption curves, most notable for those for n-hexane,
were not drastically different from those reported for some graphitized carbon
blacks, especially those with residual heterogeneity (25).
The heat curves for benzene isotherms at 10° and 20°C on DPM-EPA
activated to 50 C, are shewn in Figure 18. The isotherms at this activation
were Type III, indicative of adsorbate-adsorbate interactions being more
significant than those between the adsorbate and the adsoi-bent. The high
57
-------
Figure 15. Heat of Adsorption vs^ Surface Coverage for Benzene on Spheron 6
Q. BENZENE /SPHERON 8
A. BENZENE /SRAPHITIZED CARBON BLACK
LU
3
n
50RPUONCKCAL/I
o
•<
U.
0
5
UJ
1 4
13
12
1 !
10
8
8
.0-
.0-
.0-
-
.0-
. 0-
.0-
9
- 1
I.
_
i o 41 T T A _
li-i i,
A . A ^ A 1 T
A A A tjl -T-
A J- O T
A1 i I T T
A! ; ° ?A
0.2 0.4 3.6 0.8 1.0 I.'2 1.4 1.6 1.8
FRACTION OF 1 MONOLAYER
58
-------
Figure 16. Heat of Adsorption vo Surface Coverage for Benzene en DPM-PSU
(350°C)
Oi BENZENE /D.P.M.-P.S.U.. ACT. T.= 350.
Ai BENZENE /GRAPHITIZEC" CARSON BLACK
LJ
_J
O
»—
^tm
_1
«t
0
^
«^y
o
H-(
o.
o
c/1
-<
u.
o
1 1 1
15
1 4
13
12
t 1
10
g
8
.0-
.0-
.0-
.0-
. 0-
_
. 0—
.0-
"e"
T
IT
GC i
? n
*^™ n — j—
1 o
» _L ^ X T
^ ^ **~ W D
^ ^ ^ J- Z T
A 1 I 5 x
* A ~ 4 5, 5 I 15
iiit
0.5 1.0 • 1 .5 2.0
FRACTION OF 1 MONOLAYER
59
-------
Figure 17. Heat of Adsorption vis Surface Coverage for n-Hexane on DPM-PSU
(10CCC)
PTIONCKCAL/MOLE)
-NO)
G> Q O
1 1 1
Q
•<
fe 9.0-
h-
S a . e-
Ar HEXANE /SRAPHITI2ED CARBON BLAC<
A
A
A I T A
K* fir
1 I *T
i T T T
-*- Ja tp T T
T 9 D T T _
1 J i i i ? 5 s
A A -^- J. J_
0.5 1.0 -1.5 2.0
FRACTION OF 1 MONOLAYER
60
-------
Figure 18. Heat of Adaorption vs Surface Coverage for High Temperature
Benzene Adaorption on DPM-EFA (?0°C)
O. BENZENE /D.P.M.-E.P.A.. ACT. T.- S3.
Ai BENZENE /GRAPHITIZED CARBON BLAC<
UJ
I
O
O
t— 1
Q_
O
to
*
UJ
13.0-
12.0-
11.0-
i
10.0-
9.0-
8.0-
GC
1
T
O
u
a-T- A A
J.OiJT _TA _-. T
~-L°l?i?^t??°
- i J. J- J. -L ^ 1A- A
iii ir;i i i
0.2 0.4 O.6 0.8 1.0 1.2 1.4 1.6 1.8
FRACTION OF 1 MONOLAYER
61
-------
initial heats are due to the few high energy sites with which the adsorbate
can interact. The heats fall very rapidly to the heat of vaporization well
before monolayer coverage. The heat curve for benzene isotherms measured at
0° and 5°C on DPM-EPA, activated to 50°C, is shown in Figure 19. The
initial heats are actually lower than the heat of vaporization of benzene, and
then increase to that heat. This unusual dependence of adsorption heats on
surface coverage is not unique. Babkin and Kiselev (41) demonstrated tha+
benzene adsorption on a 100$ methylated silica surface yielded similar heat
curves. The process of replacing the hydroxyl groups with methyl groups
resulted in changing the original h:f heat curve shown in
Figure 20. Removal of the presorbed material not only increased the available
62
-------
MMR Tl- B. C T3- S. C
BENZENE /O.P.M/E.P.A., ACT. TEMP.- SB. C
UJ
I
\i
w
a.
C*.
o
UJ
8.0-!
«
e.e-j
s . o-
4.0-
Q —
0.2 0.4 0.6 0.8 1.0 1.2 1.4
FRACTION OF 1 MONOUAYER
1 . 6
-------
Figure 20. Heat of Adsorption vs. Surface Coverage for Benzene on DPM-EPA
(4000C)
MMR Tl- 0. C T2- 21. C
BENZENE /O.P.M/E.P.A.. ACT. TEMP.-408. C
GC
VI OF ADSORPTIONCKCAL/HOLE
to o - N o>
Q (9 O O O
1 II 1 1
y 8 . 0-
! T
T 1 T
IT-
JL Q -I-
" -1 { T
1 n n * i 5 1 i
i i i i i.i i i
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.8
FRACTXON OF I MONCLAYER
-------
surface area but also gave higher heats of adsorption at higher surface
coverages. For the 50°C activation (Figure 18), the heats in the 0.4 to 0.7
monolayer coverage range are 9«7 to 9«3 kcal/mole as compared to 11.4. to 9«4
kcal/mole for the 400°C activation. This is a result of increased exposure
to carbon adsorption sites and a corresponding increase in the energy of the
surface. It is interesting to note that the haats for benzene en DPM-PSU,
activated to 350°C, in the 0.6 to 0.85 monolayer coverage range were 15-2 to
12.2 kcal/mole. Although high temperature activation of DPM-EPA increased the
energy of the surface, the heats of adsorption of benzene on DPM-PSU remained
higher. It is possible that further outgassing of DPM-EPA would bring the
heat curve closer to that of DPM-PSU, but the presence of metals or other
inorganic species may inhibit this.
b. DYNAMIC METHODS.
INTRODUCTION
Dynamic adsorption measurement techniques were also employed in order to
compare these results with those obtained using the standard static method.
The use of both measurement techniques enables the surface properties to be
deterriined under widely different conditions.
One aspect of this study was treatment of particulate samples by heating
and flushing with helium in the column and heating under vacuum prior to
adsorption measurement. The effects of the variation of the injection volume
on the measured adsorption heats were also determined. The heats of
adsorption for some basic hydrocarbons on different carbon samples after
various pretreatments should allow insight into the nature of the particulate
surfaces and into the gas-solid interactions involved in the adsorption of
vapors onto Diesel particulate natter.
The specific method used in this study has teen reported previously by
Ross, et al. (42). to measure the heats of adsorption for molecules on
65
-------
graphitized carbon blacks. These workers showed that if adsorbate surface
concentration tends to zero, then the following expression can be derived:
qst
dU/Tc
where t^ is the corrected retention time and Tc is the column temperature.
Therefore, if a plot of log t' versus 1/Tc is made t*e resulting straight
line has a slope equal to q ./2.303R, (where q ^ is the limiting isosteric
heat of adsorption). This value of the heat of adsorption was found to be
appropriate for the comparison with heats measured by the static method.
Since the flow rate changed from sample to sample as a result of
differences in the packing densities of the adsorbents, the specific retention
volume was used instead of the retention time. The specific retention volume
was calculated using the following standard expression:
V = tr? Tc ^ _3_ ( (Pi/Po)2-l)
m Tf 2. ( (Pi/Po)3-l)
where F is the flow rate, m is the mass of the adsorbent in the column, T_
is the temperature at the bubble flow meter, and Pi and Po are the pressures
at the inlet and the outlet of the column.
EXPERIMENTAL
A gas chromatograph with a thermal conductivity detector (Varian 920) was
used for all these studies; the glass column (2-nm i.d., 25-cm long) was
packed with the adsorbents (Spheron 0.3g, DPM-PSU O.lg, and DPM-EPA 0.06g).
Helium (HP, Matheson) was used as the carrier gas with a flow rate in the
range of 1.5 to 3 mL/min. The retention times of the various compounds were
measured over a range of column temperatures (t50-300°C) with the various
adsorbents used in the previous studies.
66
-------
In addition to measuring retention volumes of various compounds and using
columns packed with raw Diesel particles, chrcmatographic data were also
obtained after each of the following pretreatments: (l) 300°C with helium
flow, in the column, for 15 hrs, and (2) 500°C under vacuum (lO~ torr)
for 15 hrs followed by the first treatment.
The following compounds were used as the adsorbates? benzene, cyclohexene,
cyclohexane, 1-hexene, n-hexane, methanol, n-octane, ethylbenzene, -. "•
benzaldehyde, acetophenone, naphthalene, phenol, phenanthrene, fluorenone,
fluoranthene and anthracene (all 99+/& Aldrich reagents). The liquid compounds
*
were introduced separately onto the column in a pulse size of 0.01-10.0 uL.
The solid compounds were introduced in solution (25% in methylene chloride) in
a pulse size of 1.0 uL. The column void volume was obtained from the air peak.
RESULTS
The specific retention volumes of each adsorbate on the various adsorbents
were calculated from corrected retention times. These data were then plotted
against column temperature to obtain the apparent heats of sorption or the
isosteric heats of adsorption. Examples of these plots are shown in Figure 21.
The major source of error in the determinations was the variation of the
retention time. Consequently, the error associated with the heats is ^5%>
Isosteric Heat of Adsorption on Spheron 6
Preliminary experiments were performed with Spheron 6 to compare the values
obtained for the isosteric heats of adsorption for benzene and n-hexane with
those obtained by other workers using gas chromatography and similar
adsorbents (43,44). Injection volumes (l.O uL) of adsorbates, which
correspond to <1% coverage of the available surface area of the adsorbent,
were used since this coverage falls within the Henry's Law region of the
adsorption isotherm (45). The chromatographic peaks obtained with Spheron 6
67
-------
Figure 21. Gas Chromatographic Log Vr v£ 1/Tc Plots for Benzene
OiDPM-EPA . 0.05 JUL . 9.8 KCAL/MOLE
A>DPM-PSU . 0.05JUL , 9.9 KCAL/MOLE
2.S0H
o 2.39-1
a
t? 2. I0H
bJ
.90H
t .70-
2.39 2.50 2.60 2.70 2/80 2.90 3.00 3.10
1000/T
68
-------
were sharp and cymmstrical with very little tailing which is characteristic of
a homogeneous and relatively non-porous surface. The values of the heats
obtained in this study, compared to the previous studies are shown in Table
10. The heats for the Spheron 6 are low compared with the heats on completely
graphitized carbon blacks since graphitized Spheron 6 is known to have
residual heterogeneity (i.e., it is not 100$ C + H) in its surface.
Table 10 Isosteric Heats of Adsorption for Various Graphitized
Carbon Blacks
Isosteric Heats of Adsorption (kcal/Mole)
Adsorbate Adsorbent
Spheron Graphitic,?1} Carrot Blacks
'Benzene
n-Hexane
8.7
9-4
9.4
10.1
10.4
Pretreatment of Diesel Particulate Matter
The samples of Diesel particulate matter which were used in this and
subsequent studies have significant quantities of sorbed materials on their
surfaces (PSU 5% extractable (45) and EPA 18$ extractable into dichloromethane
(46)). These sorbed species will play significant roles in the surface
properties of the particles. Therefore, the effects of various pretreatments
were studied and the apparent heats of sorption of selected hydrocarbons were
used to monitor these pretreatments. Unfortunately, the injection volume used
(l.O L) yielded heats of adsorption less than the adsorbate heat of,
vaporization. The r?.dsorbate surface coverages were calculated (on the basis
of BET surface areas of the adsorbents) to be quite high ( 40$) which must
fall beyond the Henry's Law region of the isotherm. The heats measured,
69
-------
therefore, have very little significance and ars not reported here. However,
since ths DPM-EPA particles do have a larger quantity of presorted material
than the DPM-FSU particles, high temperature pretreatment should affect the
adsorption heats on the former to a greater extent. This has been found to he
the case with the static measurements discussed previously. The peak shapes
for the adsorbates after all pretreatments were characteristic of
heterogeneous, porous surfaces and Type II isotherm adsorbents (47). The
presence of Type II isotherms was confirmed by the observation that as the
injection volume was increased the retention volume decreased (??). The
equilibrium constant decreases as the amount of the adsorbate increases and
the isotherm is concave with respect to the pressure axis.
Effect of Surface Coverage on the Apparent Heats of Sorption
A Diesel particle is formed in the cylinder as a result of the combustion
conditions present in the turbulent, heterogeneous flame. The particles once
formed are transported from the cylinder to the ambient air in the exhaust gas
which contains high concentrations of gas-phase species. Therefore, during
the process of emission, the particles will be undergoing numerous collisions
with the gas phase molecules and as a result the molecules will condense or
adsorb onto the surface of the particle. The extent of condensation will be
dependent upon the vapor pressure of the adsorbate and the temperature of the
exhaust gases. Typical emission rates of particulate matter of 0-622g/mile
have been observed by other workers (49) for light-duty automobile Diesel
engines, of which 0.0715g/mile is extractable into organic solvents.
Similarly, total hydrocarbon emission rates of 0.50g/mile have been observed
with a hot flame ionization detector. If this hydrocarbon fraction is
separated by gas chromatography, 50? of the total hydrocarbons have sufficient
molecular weight to adsorb or condense on the particle. These figures suggest
that approximately 28% of the available gas phase (hydrocarbons) adsorbs or
condense? onto thesurface of the particles. The extent of the adsorption of
the molecules will be related to the surface properties of the absorbent and
the heat of adsorption of the adsorbate. However, cnce the particle has
adsorbed moleculeson ix.s surface, the heat of sorption cf multilayers falls
70
-------
below the heat of adsorption of the monolayer. The particles, which are
collected on the filter, contain at least 18% by weight of organic extractable
compounds indicating that the particles have multilayers of molecules sorbed
on their surfaces. It is reasonable to expect that once the particle enters
the. ambient air it has sufficient sorbed molecules on its surface for its
properties to be solely dependent upon the nature of the sorbed species.
Since the bioavailability of sorbed molecules is dependent upon the nature
and the energy of sorption of the active species, a study was performed in
which the apparent heat of sorption was measured as a function of
concentrations of the adsorbate. The following adsorbates were investigated:
acetDphenone, benzaldehyie, benzene, cyclohexane, cyclohexene, ethylbenzene,
n-hexane, 1-hexene, and n -octane. The adsorbents used in this study have been
subjected to the highest temperature activation pretreatment. Table 11 lists
some examples of the results of this study. These apparent heats of sorption
suggest that the adsorbate first sorbs on the most, energetic adsorption sites
and then sorbs on sites with lower energies. The heats continue to decrease
until there is a sufficient quantity of sorbed molecules or. the surface that
the energy of sorptior equals the differential heat of evaporation from
solution. Therefore, this study suggests that the history of collisions
between the exhaust gases and the particle has a major effect on the energy of
sorption.
Isosteric- Heats of Adsorption for Diesel Particulate Matter
The small injection volumes (0.01 or 0.05 l^L) used in the previous study
correspond to coverage in the Henry's Law region of the adsorption isotherm
and can therefore be used to quantify the isosteric heats of adsorption.
Table 12 lists these heats for various adsorbates on the two samples of Diesel
particulate matter activated at the highest temperature activation
pretreatment. The higher molecular weight adsorbates phenanthrenc,
fluorenone, fluoranthene and anthracene could not be eluted from
71
-------
Table 11. Variation Of Apparent Heats Of Sorption With Volume
of Adsorbate Injected
Adsorbate
Benzene
n-Hexane
Adsorbent Injection Volume
(PL)
DPM-EPA 0.01
0.05
0.10
1.0
10.0
DPM-PSU 0.01
0.05
0.10
1.0
10.0
DPM-EPA 0.01
0.05
0.1
1.0
10.0
DPM-PSU 0.01
0.05
0.1
1.0
10.0
Apparent Heat of Sorption
(kcal/Mole)
12.8
9-8
9-1
5-9
8.7
12.0
9-9
10.4
7.6
10.5
15.0
10.6
10.4
5.4
12.8
11.4
11.
-------
Table 12. Isosteric Heats of Adsorption
Adsorbate
Water
Methanol
Dichloromethance
n-Hexane
1-Hexene
Benzene
Cyclohexene
Cyclohsxane
Ethylbenzene
Acetophenone
Benzaldehyde
n-Octane
Phenol
Naphthalene
Anthracene
Phenanthrene
Heat of
Vaporization
9.72
8.98
-
7.63
7.79
8.15
-
7.83
9-3
11.73
11.66
9.2
11.89
12.31
16.82
14.18
Heat
DPM-PSU
10.3
9.0
8.6
11.4
10.6
12.0
10.2
8.7
15.2
16.1
17.4
15.2
16.4
16.4
-
_
of Adsorption
DPM-EPA
6.9
4.8
5-0
15.0
10.2
12.8
9-7
9-3
8.3
15.1
12.2
8.4
12.2
12.3
11.1
19.0
(kcal/mole)
Graphitized
Carbon
Blacks (bO)
5.*
5.3
-
10.4
-
9-8
9.1
8.7
12.'.
13.0
-
13.4
13-0
17.3
-
^
73
-------
the column containing DPM-PSU. These results can be rationalized on the
grounds that the DPM-EFA sample is expected to have sorbed materials on its
surface which cannot easily be desorbed. These materials will probably result
from partially combusted lubricant (49)•
The isosteric heats of absorption detarmined for benzene and n-hexane on
Spheron 6 are close to those reported in the literature for graphitized carbon
blacks. The heats for the same two compounds on both Diesel samples are
somewhat greater and this is probably due to the presence of higher energy
adsorption sites on the heterogeneous porous Diesel particles. This
heterogeniety is confirmed by the peak shapes and increasing retention volumes
with decreasing injection volumes, indicative of a Type II adsorption
isothenn. The difference between the isosteric heats of adsorption may be
explained by the increased quantity of sorbed species on the DPM-EPA particle
surface. Therefore, with the DPM-5PA samples, there is a higher probability
of the adsorbate interacting with substances condensed on the particles than
with the carbon particle itself. The variation of adsorption energies with
injection volumes indicates a similar phenomenon. With increasing surface
coverage, adsorbate molecules will be encountering fewer unoccupied adsorption
sites and more adsorbed molecules. Multilayer adsorption will then occur and
the corresponding interaction energies should be less. This has significance
when considering the bioavailability of the sorbed species on inhered
particles. The molecules adsorbed on the high energy carbon sites will be
tightly held as indicated by the isosteric heats, but additional layers of
sorbed substances may be more easily liberated.
3. TRANSFORMATION OF DIESEL PARTICIPATE MATTER.
INTRODUCTION
Particulate matter which is released into the ambient air is subject to
various conditions including widely differing levels of gases and radiation.
If the adsorbate-adsorbent system is reactive with respect to the prevailing
atmospheric conditions, the compounds sorbed on the particles may be
-------
different from those compounds identified in the organic extract of
particulate matter collected in engine exhaust lines.
Electron paramagnetic resonance spectrometry (EPR) his been widely used to
study the surface properties of carbon blacks (51-55) and coals (56,57). The
EPR signal of carbon-based solids has been found to be dependent on numerous
parameters including elemental composition and extent of graphitization. This
aspect of the research describes the potential reactions that Diesel
particulate matter may undergo with selected gases, and ultraviolet/visible
radiation.
Electron Paramagnetic Resonance Spectrometry
EXPERIMENTAL
.EPR spectra were obtained with a Varian E-104A EPH spectrometer using
2-milligram samples placed in quartz tubes. All experiments were run with a
microwave power of 5 mW and a frtquency of 9«515 GHz. The g-values were
measured using ct.a'-diphenyl-B-picrylhydrazyl (DPPH) as the external standard
(g=2.0036) and the signal peak areas were measured with a plsnin^ter.
Extracted particulate samples (-Ex) were prepared by 24-hour Soxhlet
extraction cf the particles with dichloromethane. Evacuated samples wt-re
prepared by degassing the particles at room temperature for one hour at a
pressure of 10 torr. Heat-treated samples were prepared in the same
manner but at a temperature of 150°C. The effects of the gases on the EPR
signals were determined by allowing the evacuated samples to equilibrate for
one hour with one torr of the gas prior to recording the spectrum. Irradiatior
experiments were performed in the EPR cavity using lens-colimatsd, filtered
radiation from a 150 W high pressure Xenon arc lamp. The filters used were:
1. Corning #CS3-74 (low wavelength cutoff of 4.40 nm): and 2. Corning
#CSO-56 (low wavelength cutoff of 24.0 nm). The signal maximum was monitored
for 30 min with the lamp on, then, for 30 min after.the lamp was turned off.
-------
RESULTS
The g-values and line widths of the heat-treated Diesel particulate samples
are reported in Table 13- The relative positions of the EPR signals of DPM-EPA
Spheron, ana DPPH are shown in Figure 22. The reason for the-displaced
Spheron signal and the high g-value could not be ascertained. The g-values of
the unextracted and extracted Diesel samples are close to the free electron
g-value of 2.0023, and there was little variation from this value upon gas
exposure or irradiation.
Table 13- Line Widths and g-Values
Line WidthfGauss ±0.2)
DPM-PSU
DPM-PSU-Ex
DPM-EPA
DPM-EPA-Ex
2.0029
2.0028
2.0029
2.0030
6.0
8.5
5.3
7.0
However, the line widths of PPM-EPA and DPM-EPA-Ex are slightly less than
those of the corresponding DPM-PSU samples. This could, in part, be due to
the difference in the percentage of carbon, or the carbon-hydrogen ratio,
between the two Diesel samples. In EPR studies of carbon blacks (51,55) with
increasing graphitization the line width has been shown to remain essentially
constant up to 90? carbon, and then decreases between 90-9*1? carbon due to
exchange electron narrowing. The line width then increases above 94? c.irbon
owing to the appearance of conduction electrons formed in the carbonization
process. The difference in the line widths between DPM-PSU and DPM-EPA may
also result from the presence of trace inorganic species in DPM-EPA, caused by
the combustion of the numerous additives contained in the lubricant and
commercial Diesel fuel (58,59)? The synthetic lubricant and model fuel used
eliminates the possibility of trace inorganic compounds on the DPM-PSU
^articles. The fact that the line widths of the extracted samples are greater
76
-------
Figure 22. Relative Positions of Era Signals
Evac.
77
-------
than those of the unextracted samples could be. due to radical-radical
broadening caused by the spin centers being closer to one another after the
extraction process. The g-value and its fixed nature with respect to various
treatments suggests that the EPR signal is similar to that for all carbon
blacks. The variation of the line width between particulate samples raises the
possibility of EPR being used to characterize particles emitted from different
engines and to characterize the contributions due to different fuels and oils.
Although the EPR signal arises from the carbon-hydrogen substructure common
to most carbon blacks, the signal can be affected by many surface treatments.
Table 14 shows the changes in the line width and the integrated area of the
signal for the DPM-PSU particles after a variety of pretreatments. Figure 23
shows the change in the actual DPM-PSU signal due to evacuation and subsequent
heat-treatment. Similar effects were observed for DPM-EPA particles.
Degassing the sample increases the signal intensity and narrows the line width,
and heat treatment under vacuum increases this effect significantly. These
observations probably reflect the removal of sorbed substances that physically
interact with the radical centers and thereby broaden the signal. An
evacuated sample could be exposed to the air, reevacuated, and the same signal
characteristics retrieved suggesting the interactions affecting the signal
were mostly physical in nature. By comparison of signal areas between known
amounts of particles and DPPH, the spin concentrations were calculated. All
were of the same order of magnitude, 10 spins/g. This falls in the range
of values given in the literature for carbon black spin concentrations (55).
Tablo 14- Effects of Various Treatments on the EPR Signal of DPM-PSU
Treatment
1 Attn. air
Evacuated
Heat-treated
Line Width
20 V-
10 +/-
6 +/-
(Gauss)
2
1
0.2
Change in
Signal Area
-
increase
increase
78
-------
Figure 23. Effects of Various Treatments on the EPR Signals of DPM-PSU
Treatment
1 : No
(Gain)
treatment
Evacuated
Heat-treatec (2.5x10)
(102)
-------
The effects of the selected gases on the signal of DPM-PSU-Evacuated are
summarized in Table 15• The broadening of the signal, or "poisoning action",
of paramagnetic gases such as oxygen and nitric oxide is well documented
(51.52,55). Paramagnetic adsorbates cause line broadening by a primarily
physical interaction between the gas molecules and spin centers of the solid.
The spin-lattice relaxation time is decreased and the unpaired electron energy
levels spread owing to the presence of the adsorbed species. However, this
cannot explain the observed effect of the exposure of the particles to
nitrogen dioxide. It has been shown by others (60) that the PAH,
benzo(a)pyrene, adsorbed on filters reacts with nitrogen dioxide to fora
mutagenic nitro-derivatives. This suggests a possible mechanism by which
additional radical centers could be formed upon exposure of the particles to
nitrogen dioxide: the nitrated species react with phenolic or dihydroxy-
compounds, which have been identified in the particulate extract
>X
OH OH"1
OH OH
Also, hydrocarbons, which are present on the particles, are known to undergo
hydrogen abstraction upon reaction with nitrogen dioxide (6l);
R-H + N02 = R* + HOKO
Either of these mechanisms could produce an increase in the radical
concentration in the particulate matter. The technique of monitoring the EPR
signal of the particles upon exposure to nitrogen dioxide proved to be a
convenient method of following this interaction.
80
-------
Table 15- Effects of Selected Gases on the EPR Signal of DPM-PSU
Change in Change in
Gas (l torr) Line Width (G) Signal Area
Oxygen + 3-4
Nitric Oxide + 2-3
Nitrogen dioxide - . no immediate effect,
+40-50$ upon 24-hour
exposure
Irradiation experiments were initially performed on the unevacuated samples
to determine if the signal irradiation behavior was dependent on the wavelength
of light used. The wavelength of irradiation was varied by the use of filters;
a larger initial increase (6-8? of the original signal) with a slower
subsequent decay of the signal intensity occurred as the wavelength was
decreased to the ultraviolet region. The change in the signal was greater
when more energetic radiation was allowed to pass to the sample. The effects
of irradiation on the signal intensity-of DPM-PSU with various treatments are
shown in Table 16. Filler #CS3-74 was used in all of these experiments. The
results were similar for DPM-EPA and the effects were less pronounced for the
extracted samples. The irradiation effects on samples with oxygen or nitric
oxide were similar to those of samples in air, although the net effect was an
increase for the former, possibly because of low pressure. The greatest
reactivity was observed for the degassed particles with no gases. The presence
of air or low pressures of added gases decreased the interaction of the light
with the radical centers. High temperature heat-treatment under vacuum or
exposure of the samples to nitrogen dioxide had the common effect of
inhibiting any net change in the signal intensity. It is interesting to note
that even the extracted Diesel particles displayed significant sensitivity to
irradiation while irradiation of Spheron 6 had no effect with any pretreatment.
The radiation may ionize surface sorbed compounds, such as those with hydroxy
groups, and cause them to lose electrons, thereby increasing the free radical
concentration. The Spheron particles, lacking such adsorbed species, woul<1
not be expected to react to the light in the same way as the
81
-------
Diesel samples. The sensitivity seemed to be the most pronounced when no
interfering gases were present, in vacuum, and was greater the more energetic
the radiation. When air or'other gases are present the potential.for photo-
chemical reactions is increased. Characterization of these interactions would
require much more research. These preliminary results indicate that Diesel
particular matter is sensitive to light plthough the exact nature of the
interaction is not yet known. Thorough investigation of the photochemical
reactivity of these particles would be worthwhile; there is presently a lack
of information in this area. Knowledge of the effects of atmospheric
conditions on potential biologically-active compounds sorbed onto the
particles will yield a better understanding of the ultimate health risks posed
by airborne particulate matter.
Table 16. Effects of Irradiation on the EPR Signal of DPM-PSU
Treatment
Light On
Light Off
Net Change
1 Atm. air
Evacuated
Evacuated +
02 or NO
Evacuated +
N02
sharp increase
slow decay
level off
sharp decrease
rapid increase
level off
sharp increase
level off
sharp decrease
level off
sharp decrease
slow decay
level off
sharp increase
slow increase
level off
sharp decrease
level off
sharp increase
level off
decrease
5-10?!
increase
30-502
slight
increase
none
82
-------
Formation Of 1-Nitropyrene By Reaction Cf Nitrogen Dioxide With Diesel
Particulate Matter.
EXPERIMENTAL
The EPR studies have demonstrated that Diesel particulate matter will react
with nitrogen dioxide; this is a potential route for the formation of
1-nitropyrene. Therefore a study was undertaken in which the particulate
matter collected from the nitrogen-free oxidant experiments was reacted with
known concentration of nitrogen dioxide in the presence of water vapor at
52°C (the sampling temperature). The nitration system is shown in Figure
24. By varying the flow of the nitrogen dioxide (960 ppm) and high purity
nitrogen, any concentration of nitrogen dioxide, wnich is monitored both
before and after the filter, is possible. Using this system, nitrogen dioxide
adsorption may be measured with a resolution of 1 ppm of nitrogen dioxide
(62). To obtain a trace quantity of nitric acid, the nitrogen was bubhled
through water. The particles were subsequently extracted and the
1-nitropyrene determined by the HPLC procedure already described.
RESULTS
The results of this study are shown ii. Table 17.
-------
Figure 24. Nitration System
NOX
METER
HEATER V
TWO WAY VALVE
o:
o
CO
z
LLJ
V)
UJ
01
o:
LU
Q.
Z
UJ
FILTER
EX.-: AILS'
-------
Table 17. Reaction of Nitrogen Dioxide with Nitrogen-free
Particulate Matter.
Nitrogen Dioxide 1-Nitropyrene
Concentration (ppm) Concentration (ppm)
0 2.0
.5 44.6
50 76.6
100 221
150 248
These results suggest that 1-nitropyrene may result from the process of
collecting the Diesel particulate matter since nitrogen dioxide (nitric
acid) is present in the exhaust gat, since the nitration can occur under
typical sampling conditions. Also it is reasonable to expect that other
adsortates could be similarly nitrated. .
C. -BIOLOGY
INTRODUCTION
Two microbial assays were used in this study to measure the potential
activity of Diesel particulate matter: the Ames test and the Comp test.
AMES MUTAGENICITY TEST
EXPERIMENTAL ' .
Bacterial Strains and Media
The Ames mutagenicity test system requires the use of genetically
constructed mutants of Salmonella typhimurium. The strains have been selected
for sensitivity and specificity for reversion from histidirie auxotrophy to
prototrophy. There are several bacterial tester strains containing different
mutations which render the bacteria auxotrophic for histidine. These strains
contain unique types of DNA 'it-image in the gene(s)
85
-------
coding for the enzymes essential for histidine synthesis. In this study,
TA1535, TA1537, TA1538, TA98 and TA100 were utilized, as recommended for
routine Ames mutagenicity assays (63,64,65). A description of the genotypes
of the various strains is presented in Table 18. TA1535 and TA100 have base
pair substitutions in histidine synthesizing genes and can be used to detect
mutagens causing base pair substitutions. Strains TA1537, TA1538, and TA98
contain frameshift mutations and are used to detect various frameshift
mutagens. In addition to the histidine mutation, each tester strain contains
two additional mutations: (l) the loss of the DNA excision repair capacity,
and (2) the loss of the lipopolysaccharide (LPS) barrier that coats the
bacterium. The tester strains TA98 and TA100 were developed by transferring a
resistant transfer factor (R factor) to the standard tester strains TA1535 and
TA1538, respectively (63,66). The presence of this factor makes the strains
more sensitive to pol/cyclic hydrocarbons. All of the Salmonella strains
utilized were obtained from Dr. Bruce N. Ames, Department of Biochemistry,
University of California, Berkeley, California. Upcn receiving the tester
strains, frozen permanent cultures were maintained (53). In order to ensure
that the bacteria used in the Ames test system had thos correct reversion
specificity, cultures of the various stra.ins were routinely checked to confirm
the required genotype (63-65). All five strain? were checked for histidine
and biotin auxotrophy, the presence of a "rough" outer envelope (rfa), the
absence of inducible excision repair (uvrB), and the characteristic
spontaneous reversion rate (per 10 cells) (63). In
Table 18. Genotype of the TA Strains used for Mutagen Testing
Histidine
hisG46
TA1535*
TA100
All strains
Mutation
hisC3076
^•^MOBei^v.*
TA1537
were originally
hisD3052
V^HWV^HH^
TA1538
TA98
derived from S.
LPS
rfa
rfa
Additional
Repair"1"
uvrB
uvrB
Mutations
R factor
+R
typhimurium LTp.
* The deletion (&) through uvrB also includes the nitrate reductase (chl)
and biotin (bio) genes.
86
-------
addition, strains TA98 and TA100 were tested for the presence of plasmid
pKMlOl which conveys ampicillin resistance. Master plate cultures used for
the inoculation of the nutrient broth cultures required in the test were .
prepared and maintained for no longer than one month (63)• Overnight cultures
(I2hrs.-16hrs.) of the strains to be used in testing were grown in nutrient
broth (Oxoid Nutrient Broth Media, #2) (63). The mutagenicity assay was
performed on Spizizen minimal medium supplemented with 0.5# glucose (67). Top
agar containing trace amounts of biotin an'', histidine was prepared (63)•
Quantitative Ames Mutagenicity Test
The experimental procedure for the Quantitiativs Ames Salmonella
mutagenicity Test was performed ia the following manner (63). 0.1 mL of a 12
O
to 16-hour nutrient broth culture of the tester strain (10 cells) was added
to 2.0 mL of sterile tcp agar (0.6? Difco c.^ar, 0.5? NaCl) (€3). The top agsr
contains trace amounts of histidine and biotin (0.045 mM histidine-HCl, and
0.045 mM biotin) which allows all the bacteria on th'e plain tc undergo several
divisions; this growth is necessary in many instances for the expression of
mutagenic events. Following the addition of bacteria, various concentrations
of the sample to be tested were added (usually 0.1 mL or less), followed by
0.5 mL of S-9 microsomal mix (if required) (63). The contents were mixed and
poured onto minimal glucose agar plates. The plates were in-^rte.'. and
incubated in the dark at 37°C for 48 hours. Following incubation, the
numbers 01' colonies (r^vertants to histidine prototrophy) were determined
using an Artek 080 automatic colony counter. In studies where unextracted
parti-le samples were tested directly, vhe color.ies appearing on the plate
were counted by hand. Control plates without mutagen or ?~9 mix were prepared
in the saa;e manner. Colonies appearing on these plates i^present th?
spontaneous hist i.dine revertants, whose number 13 characteristic for esch
strain. 'Che number of spontaneous histidi ,e revertarts was subtracted fror
the total number of colonies on each pi?.te to yie"..I the net number of
revortants per dose r.f sample. Unless othervise -.pecified. values presented
represent the averag- of at least three experiments incorporating several
different doses of test sample tested in triplicate. In accorclsnce "it'.
recommendations for Ames seating without microsorcal activate :m (64,65), the
87
-------
following chemicals v?re used as positive control agents: 2-nitrofluorene for
strain TA98, methyl methane sulfonate for strains TA100 and TA1535,
9-aminoacridine for strain TA1537, 4-nitro-o-phenylenediamine for TA1538- For
testing with S-9 microsomal activation, 2-aminoanthracene served as a positive
control agent for all strains (63-65). Routine transfer of histidine
prototrophs to minimal glucose agar plates was performed in order to confirm
that they were actual revertants (64,65). All tests were performed under
yellow safe lights to minimize the formation of photoproducts within the
samples of soluble organic extract.
Preparation of S-9 Microsomal Fraction
The S-9 microsomal fraction was obtained from the livers of male rats
(Sprague-Dawley) which had received a single intraperitoneal injection of
Araclor 1245 (Polychlorinated biphenyl) (63«68). All steps were carried out
at 0-4°C using sterile solutions, surgical instruments and glassware.
Arocolor 1254 at a dosage of 500 mg/kg rat weight was given intraperitoneally;
on the fifth day, the rats were sacrificed, their livers removed and weighed,
and transferred to a beaker containing 0.15 M KC1 (3 mL/g wet liver). The .
livers were minced with sterile scissors and homogenized in a Waring blender.
The homogenate was centrifuged for 10 minutes
at 9000 X g and the supernatant (S-9 fraction) decanted, and distributed into
3 dram vials. The vials were quick-frozen in a dry ice-ethanol bath and
stored at -70°C until use.
To test the requirement of the sample for metabolic activation, a new vial
of S-9 mix was prepared each day, and kept on ice until use. The S-9 mix
contains per mL: S-9 (0.1 mL), MgCT. (8 pM), KC1 (33 uH), glucose-6-
phosphate (5 uM), NADP (4 uM), and sodium phosphate, pH 7.4 (100 uM) (3). The
stock salt solution (0.4 M MgCl2, 1.65 M KCl) and phosphate buffer (0.2 M,
pH 7.4) were stored at 4°C. The S-9 mix was filter sterilized through a
0.45 vm Millex filter.
88
-------
Toxicity Testing
Two different tests were performed to determine the toxicity of the
soluble organic extract in the Salmonella tester strains. In the initial test
(Toxicity Assay I), nutrient broth was inoculated with either Salmonella
strain TA98 or TA100 and incubated with aeration at 37°C for a period of 6
hours. At this time, the extract was added to yield a final concentration of
500 yg/mL. The culture was further incubated at 37°C and at specific time
intervals removed, serially diluted in 1 X Spizizen salts and kept at 4°C
until 0.1 mL aliquots were plated onto nutrient agar. After a period of
incubation (24 hours at 37°C), the number of colonies appearing on the
nutrient agar were counted and compared to control cultures which were not
exposed to any extract sample. All counts were made on plates having between
30 and 300 colonies. A second toxicity test (Toxicity Assay II) was performed
in conjunction with the Quantitative Ames Mutagenicity Test. Overnight
cultures (12-16 hrs) of the tester strains used in routine Ames tests were
serially diluted in 1 X Spizizen salts ar.d 0.1 mL aliquots of the cells were
added to 2.0 mL of top agar, with and without the addition of varying
concentrations of the extract sample. The overlays were poured onto minimal
glucose agar and incubated for 48 hours at 37°C. After this period of
incubation, the ratio of the number of colonies appearing on control and test
plates was used to calculate the percentage survival for e?.ch of the doses
tested. All counts were made with plates having between 30 and 300 colonies.
Bacillus subtilis Comptest
Bacterial Strains and Media
The Comptest system requires the use of _B. subtilis strains RUB827
(trpC2, metBlO, and p_oU5) and RUE818 (wild type) (69). RUB827, in addition
to being auxotrophic for both tryptophan and methionine, carries two prophages
P3SX and Sp . Both of these prophages are inducible following UV-irradiation
of RUB827, a process found to be one of the "SOS" phenomena of B_. subtilis
(70). Competent cells of RUB827 were prepared during successive incubation
89
-------
periods in growth medium I and growth medium II (71). The Comptest assay was
performed on Spizizen minimal medium supplemented with 0-5% glucose and one or
both of the amino acids for which RUB827 is auxotrophic. All dilutions of
cells were carried out in 1 X Spizizen salts- The genotype of RUB827 was
routinely confirmed by demonstrating tryptophan or methionine auxotrophy, and
inhibition of growth on 0*05% MMS (polA5 mutation) (70).
Procedures for Genetic Exchange
DNA was isolated from wild type RUB818 and used to transform RU3827
.(69,71-73).
Bacillus subtilis Comptest Assay
Cells were grown in GMI for 90 min after the cessation of logarithmic,
growth, then diluted 10 fold into GMII and incubated for one hour at 37°C on
a New Brunswick gyratory shaker at 250 rpm. Following this incubation period,
DNA from RUB818 was added to RUB827 (at a final concentration of 1-5 ug/mL and
incubation was continued for 25 min. The transformation reaction was
terminated with the addition of DNase (100 ug/mL) or by rapid vortexing. The
transformed cells were divided into 0.9 mL aliquots and 0.1 mL of the test
sample at various concentrations was added. Tne cells were incubated in the
dark with aeration for 30 min. Following this exposure, the cells were
centrifuged (800 X g for 2 min), the supernatant discarded, and the pellet
resuspended in 1 mL 1 X Spizizen minimal salts. The cells were then plated
onto minimal medium, supplemented with both or only one of the amino acids for
which RUB827 is auxotrophic (tryptophan, methionine). The plates were
inverted and incubated at 37°C for 48 hrs. The colonies were scored on the
ARTEK counter after 48 hrs of growth at 37 C and duplicate determinations
were used for each dose of soluble organic extract in three experiments. The
number of transformants was determined from growth in the absence of one of
the required amino acids. The total number of viable cells was determined
from growth on minimal medium supplemented with both tryptophan and
methionine. The relative transformation efficiency (RT) was calculated bv
90
-------
dividing the percentage transformation at a particular dose of the extract by
the percentage transformation for cells not exposed to any sample. The
reference point of 0.05 RT was used to compare the potency of different
samples. Positive control agents WMS and UV-irradiation were routinely tested
to demonstrate the presence of inducible prophages.
Toxicity Testing of B. subtilis
The toxicity of the extracts towards Ii. subtilis strain RUB827 was
determined from the Comptest. The colonies appearing on minimal glucose agar
supplemented with both tryptophan and methionine are termed "viables." Any
decrease in the number of "viables" for cultures receiving doses of the sample
is an expression of the toxicity of the samples (69).
RESULTS
Determination of the Mutagenic and "SOS" Inducing Potential of Unextracted
Diesel Particulate and the extract.
Ames Mutagenicity Test
The results of the Quantitative Ames mutagenicity test for the extracts
and unextracted particle are presented in Table 19« Each sample was tested
over a wide dose range. The number of revertants per dose was determined by
.subtracting the average number of spontaneous revertarit colonies appearing on
negative control plates from the average number of revertants per plate for
each dose tested. The number of revertants per quantity of sample was
determined from the slope of the best-fit line through the linear portion of
the dose response curve, as calculated by regression analysis. The extract
had significant mutagenic activity for strains TA98 and TA100. Microsomal S-Q
activation of the extract did not significantly enhance the mutagenic activity
of the extract. In fact, in most experiments, the number of revertants per ug
of extract actually decreased slightly with the addition of the S-9 fraction.
Also the results clearly illustrate the reduced mutagenic activity of
91
-------
unextracted Diesel particles in comparison to the extract (less than 0.1
revertants per 1.0 mg of particulate matter). Microsoraal S-9 activation again
did not significantly enhance the mutagenic. activity of the sample. • - :
Table 19- Quantitative Ames Mutagenesis Results.
D.P.M from
Model Fuel/Air Oxidant
; Unextracted Particles
'
Extract
.
Unextracted Particles*
' ' Extract* "
NAD
NAD
4.3 +
3.7 *_
0.11 +
. NAD ~
5-5 +
TA 98
0.5
0.6
• 0.02
1.0
Activity
(S-9)
(S-9)
: - -
S-9)
Revertants/^g
TA 100
NAD
NAD
3.2 + 1.1
3.3 ^ 0.8
0.30 + 0.12
0.14 + 0.05
14.9 + 2.7
S-9)
(S-9)
(S-9)
For comparison purposes the data obtained for an identical engine operating
on regular Diesel fuel and lubricant and air oxidant are included.
'NAD No Mutagenic Activity Detected.
Bacillus subtilis Comptest
The results of .the JB. subtilis Comptest for the extracts and unextracted
particles are presented in Table 20 (69.73). The competent cultures of strain
RUE827 were transformed and exposed to selected doses of either unextracted
particles or their extracts. The percentage transformation and the percentage
survival, of.the total population of cells were determined at each dose. The
relative transformation efficiency (RT, the percentage transformation of the
cells treated with a specific concentration of agent/the percentage trans-
formation of cells not treated) of 0.05 is considered the reference point for
the comparison of various samples. A dose response curve for each sample was
obtained.and from these.data, the dose required to reduce the RT to 0.05 was
determined. The dataiclearly demonstrate.the: reduced DNA damaging activity of
unextracted Diesel particles in comparison to the extract. The unextracted
particles did not reduce the RT to 0.05 even at doses of 1500 ug/ml.
92
-------
Table 20. Relative Transformation of Strain RUB827 to Diesel
Particulate Matter
DPM from Average Concentration Percentage Survival
Model Fuel/air oxidant at which RT = 0.05 (ug/mL) at RT - 0.05
Unextracted
Extract
Unextracted
Extract*
Particles
Particles*
NR
170
NR
110
± 70
+ 22
97 +
68 ^
87 +
55 *
2-
10
5
9
For comparison purposes the data obtained for an identical engine
operating on regular Diesel fuel and lubricant and air oxidant are
included.
NR No reduction of RT to 0.0$
Toxicity Testing
Salmonella Toxicity Assay II
The results of Toxicity Assay II are presented in Table 21. The number of
colonies appearing on the minimal medium receiving selected doses of sample
were compared to the number of colonies appearing on minimal medium which did
not receive any sample. The extract samples had insignificant toxic effects
on strains TA90 and TA100, even at the highest concentrations tested '500 ug
per plate). It should be noted that for all of the samples tested in the Ames
assay, the mutagenic activity (revertants per yg) was determined at doses not
showing significant toxicity. The unextracted particles did not demonstrate
significant toxic effects even at a concentration of 1500 ug per plate.
Bacillus subtilis Toxicity Assay
The toxicity of the extracts and unextracted particles to RU3827 were
directly determined from the results of the Comptest (Table 20). The colonies
appearing on glucose minimal agar supplemented with both tryptophan and
methionine are termed "viables." Any decrease in the number of viables for a.,t
given dose is an expression of the toxicity of the sample. The percentage
toxicity of the extract sample at the concentration which reduces the R"£
93
-------
Table 21. Percentage Survival of TA98, TA100 at Various Doses
Doses (ug/Plate)
DPM from Model 1500 1000 500
Fuel/Air Oxidant Uneextracted UnextractedExtract
DPM DPM
TA98
TA100
TA98
TA100
79
NT
86*
NT*
NT 83
NT 81
98* 75*
NT* NT*
250
E>- tract
94
91
88*
NT*
100
Extract
NT
100
NT*
NT Not tested
For comparison purposes the data obtained for an identical engine
operating on regular Ciesel fuel and lubricant and air oxidant are
included.
to 0.05 i«i presented in Table 20. For'concentrations reducing the RT to 0.05
but which are not actual test doses, the percentage toxicity was determined at
the next highest tested dose. For samples not reducing the RT to 0.05, the
percentage toxicity was determined from the results obtained at the highest
tested dose. For any of the samples tested, no more than sixty percent of the
total cell population was killed by the extract at concentrations which are
equal to or greater than required to yield an RT of 0.05-
Biological Activities of Used, Unused Lubricant and Model Fuel
The model fuel and the synthetic lubricant (used and unuse-i) had neither
detectable mutagenic activity nor TNA damaging activity, as illustrated by the
data in Table 22. Therefore, these compounds have a limited capacity to induce
the bacterial "SOS" system. The prototype fuel was tested in increasing volume
to volume concentrations up to only 5% because of its greater than ninety-five
percent toxicity for the RUS827 at a concentration of 10*. The UCON
-------
lubricant was tested in increasing concentrations up to 20$ which displayed
optimal solubility characteristics.
Table 22. Biological Activity of Used and Unused Lubricant and Model Fuel.
Sample
Lubricart
Used
Unused
Model Fuel
Revertant
TA98
NAD
NAD
NAD
per ug of sample
TA100
NAD
NAD
NAD
Cone at which
equals 0.05 ( y
NR
NR
NR
RT
g/mL)
NAD No Mutagenic Activity Detected
NR No reduction of RT *o 0.05
The Effect of the use of a Nitrogen-free oxidant 01 the Mutagenic and
"SOS" Inducing Activity of Unextracted Diesel Particle? and their Extracts
In order to determine the contribution of nitrogen containing compounds to
the mutagenic and DNA damaging activities of the compounds present in Diesel
exhaust, the particulate matter from the nitrogen-free oxidant was examined.
A summary of the biological activity of the unextracted particles and extracts
collected without the presence of nitrogen is presented in Table 23.
The results clearly demonstrate the reduced mutaeenic activity and DNA
damaging activity of extract from particulate generated by a nitrogen-free
oxidant. These data illustrate an eight-fold reduction in mutagenic activity
following the removal of nitrogen from the engine system. Similarly, there is
95
-------
a twelve-fold reduction of the mutagenic activity of the extract for TA100 in
the absence of nitrogen during the operation of the engine. The mutagenic
activity of the extract toward TA98 or TA100 was not enhanced by S-9
microsomal activation. The results of the Comptest clearly demonstrate the
reduced DNA damaging activity of the extract generated from the engine
operated in the absence of nitrogen. The raw particulate demonstrated neither
mutagenic nor DNA damaging activity. The overall toxicity of the extract from-
the nitrogen-free conditions was less than the toxicity of the extract
generated from the engine operated at normal atmospheric conditions. It
should be noted, however, that the extract from either system is not
sufficiently toxic to interfere with the evaluation of the biological assays.
Table 23- Biological Activity of Particulate Matter Collected from
Engine operated with Nitrogen-free oxidant.
Sample
Unextracted
Particles
(+S-9)
Ex tract
(+S-9)
Percentage
Survival
Activity
TA98
NAD (NAD*)
NAD(NAD*)
0.5910.5(4- 3^0. 5*)
0.32^0.3(3.75). 6*)
100(83*)
Revertants/ug
TA100
NAD (NAD*)
NAD (NAD*)
0.26^0.2(3.2+1.
0.16+0. 15(3- 3^0
100(01*)
-
Cone, at which RT
equals 0.05 (yg/mL)
NR(NR*)
1*) NR( 170^20*)
.8*)
95(68*)
( ) Data obtained with air oxidant
NAD No Mutagenic Activ-ty Detected
NR No reduction of RT *,o 0.05
Mutagenic Activity of the Diesel Particulate Matter Collected with
Different Oxidants.
The biological activities of the extracts of the psrticulate samples which
were collected from the engine operated with different oxidants are shown in
Table 24.
96
-------
Table 24. Biological Activity of Particulate Matter Collected
From Engine Operated with Different Oxidants.
Oxidant % Nitrogen Dioxide Particulate 1-nitropyrene Direct Microbial
(g/IkW-hr) Matter (ug/IkW-hr) Mutagenic Activity
(g/IkW-hr) (Mega Rev/IkW-hr)
TA98 TA100
19.1,
20.0,
21.0,
24.8,
28.6,
02;
02;
02;
02;
02;
80.9,
80.0,
79.0,
75.2,
71.4,
N2
N2"
N2
N2
N2
0.
0.
o.
1.
1.
31
45
60
10
01
1.
1.
1.
0.
0.
69
49
31
61
30
42.
23.
4.
32.
20.
3
3
5
7
2
1.62
0.73
0.23
1.39
0.74
1.36
0.89
0.15
1.23
0.64
These results clearly demonstrate that the observed microbial activity
correlates with the 1-nitropyrene concentration. However, other compounds
must be contributing to the measured activity since 1-nitropyrene cannot
account for all the observed activity. The concentration of 1-nitrop/rene can
be correlated with the emission rates of particulate matter and of nitrogen
dioxide, which are the reactants for the production of 1-nitropyrene.
Mutagenic Activity of the Nitrogen-free Diesel Particulate 'Matter after
Reactions with Nitrogen Dioxide
The biological activities of the extracts of the particulate matter
samples which have been reacted with nitrogen dioxide are shown in Table 25-
Q7
-------
Table 25. Biological Activity of Particulate Matter After
Reaction with Nitrogen Dioxide
Nitrogen Dioxide 1-Nitropyrene * Revertants/pg
g/g particles ug/g particles
TA98 TA100
0
0.2
1.7
3.5
5.2
0.7
15-4
26.6
76.6
86.0
0.12
1.80
2.10
2.48
2.60
0.08
NT
1.94
2.08
2.20
NT Not Tested
These particles contained approximately 400 yg of pyrene/g of particles.
The results clearly demonstrate that the biological activities correlate
well with the 1-nitropyrene concentration. However, they also demonstrate
that other nitroaromatics must also be contributing to the observed biological
activity since the observed activities can not be explained solely on the
basis of 1-nitropyrene which has an activity of 400 to 700 revertants/ug
(TA98).
CONCLUSIONS
This research project has clearly generated as many questions as it has
solved . It has demonstrated that the microbial activity which has been
observed for the extract from Diesel particulate matter may be produced by a
.sampling artifact. Also, since mammalian systems only have nitroreductase in
the liver or intestinal tract, the presence of 1-nitropyrene may not represent
a real problem. The use of raicrobial testing systems which do not have the
enzyme nitroreductase should be investigated. Also some studies to measure
the bioavailability of organic absorbates on the surface of Diesel particulate
matter should be nade since th>3se studies will define the potential risk- of
these particles.
98
-------
References " • . .
/
1. Karim, G. and G. Ward, "The Examination of the Combustion Processes in a
Compression- Ignition Engine by Changing the Partial Pressure of Oxygen in
the Intake Charge." SAE Paper 680767, 1968.
2. Andon, J. , A. Siegel, J. Johnson, D. Leddy, S. .Smaby, J. Pitts, A. Winner
and K. VanCauwenberghe, "An Initial Assessment of the Literature on the
Measurement, Control, Transport, Transformation, and Health Effects of
Unregulated Diesel Engine Emissions." National Highway Traffic .Safety
Administration, January 1979«
3. Ullman, T. , K. Springer and T. Baines, "Effects of Six Variables on
Diesel Exhaust Particulate." ASME Paper 80-DGP-42, 1980.
4. Tsunemoto, H. and H. Ishitani, "The Role of Oxygen in Intake and Exhaust
on NO Emission, Smoke, and BMEP of a Diesel Engine with EGR System." SAE
Paper 800030, 1980.
5- Vostal, J. , "Health Aspects of Diesel Exhaust Particulate Emissions."
Symposium on Health Aspects of Automotive Emissions, May 1980.
6. Lipkea, W. , J. Johnson and C. Vuk, "The Physical and Chemical Character
of Diesel Particulate Emissions - Measurement Techniques and Fundamental
Considerations." SAE Paper SP-430, February 1978.
7. Proceedings of the Scientific Review Meeting on the USEPA Diesel Emission
Health Effects Research Program. Arlington, VA, December 1978.
8. Waters, M. , S. Nesnow, J. Huisingh, S. Sandhu and L. Claxton, Environ
Sci Res, 15, 1978.
9. MacDonald, J. , S. Plee, J. O'Arey and R. Schrenk, "Experimental
Measurements of the Independent Effects of Dilution Ratio and Filter
Temperature on Diesel Exhaust Particulate Samples." SAE Paper 800185,
1980.
10. Bradow, R. , "Diesel Particule Emissions." Symposium on Health Aspects of
Automotive Emissions, Vol. 56, No. 9, November 1980.
11. Heisey, J. and S. Lestz, "Aqueous Alcohol Fumigation of a Single-Cylinder
DI Diesel Engine." SAE Paper 811208, 1981.
12. Campana, J.E., T.H. Risby and P.C. Jurs, Anal Chim Acta, 112:321,
13. Kansen, G. and B. Munson, Anal Chem, 50:1130, 1978.
14. Baldwin, M.A. and F.W. McLafferty, Organic Mass Spectron, 7:1353, 1073.
15. Ohasi, M. , S. Yamada, H. Kudo and N. Nakayama, Biomed Mass St>ectrcm
5:578, 1978.
99
-------
16. Carter F.E., "The Platinum Metals and Their Alloys." Proceedings of the
Institute of Metals Division of the American Institute and Mining and
Metallurgical Engineers, 1928, pp. 759-85-
17. Tejada S.B., "Analytical Methods for Nitro Aromatic Compounds." In EPA
1981 Diesel Emissions Symposium, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, 1981.
18. Lee, M.L., D.L. Vassilaros, C.M. White, and M. Novotny, Anal Chem,
51:768, 1979-
19. Punkenbusch, E.F, D.G. Leddy and J.H. Johnson, "The Characterization of
the Soluble Organic Fraction of Diesel Particulate Matter." 3AE Paper
79418, 1977.
20. Schuetzle, D., F.S. Lee, T.J. Prater, and S.B. Tejeda, Int J Environ Anal
Chem, 9:93, 1981.
21. Schuetzle, D., T. Riley, T.J. Prater, T.M. Harvey and "J.F. Hunt, Anal
Chem, 54:265-272, 1982.
22. Frey J.W. and M. Corn, Nature, 216:615, 1967-
23- Frey J.W. and M. Corn, Am Ind Hyg Assoc J, 28:468, 1967.
24- Lipkea W.H., J.H. Johnson and C.T. Vuk, "The Physical and Chemical
Characterization of Diesel Particulate Emissions, Measurement Techniques
and Fundamental Considerations." SAE Paper 780108, February 1978.
25. Avgul N.1I. and A.V. Kiselev, Chemistry and Physics of Carbon, P.L.
Ualker, Jr., ed., Vol. 6, No. 1, 1970.
26. Voet A., T.G. Lamond and D. Sweigart, Carbon, 6:707, 1968.
27. Lyon, L.L., G.R. Crocker and J.S. Dahler, Proceedings of the Fourth
Carbon Conference, 71, I960.
28. Nelson, J.R. "Sorption of Organic Molecules into Caroons, Crosslinked
Polymers, and Coals of Varying Rank." Ph.D. Thesis, The Pennsylvania
State University, 1979.
29- Anderson R.B. and P.H. Emmett, J Appl Phys, 19:367, 1941.
30. Brunauer, S., P.H. Enmett and E. Teller, J Am Jhem Soc, 60:309, 1938.
31. Brunauer, S., L.E. Copeland and D.L. Kantro, The Solid-Gas Interface,
E.A. Flood, ed., New York: Dekker, 1967.
32. Isirikyan A.A. and A.V. Kiselev, J Fhys Chem, 65:<501, 1961.
100
-------
33- Corn, M., T.L. Montgomery and h.J. Reitz, Science, 159:1350, 1968.
34. Ross, M.M., M. R. Chedekel, T.H. Risby, S.S. Lestz and R.E. Yasbin,
Environ Int, in press.
35- Brunauer, S., L.S. Deming, W.E. Deming and E. Teller, J Am Chem Soc, 62:
1723, 1940.
36. Zweidinger, R., private communication, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1981.
37. Lowell, S., Introduction to Powder Surface Area: New York, Wiley-
Interscience, 1979«
38. Gregg S.J. and K.S.W. Sing, Adsorption, Surface Area, and Porosity, New
York: Academic Press, 1967.
39. Pierotti R.A. and R.E. Smallwood, J Coll Interf Sci, 22:469, 1966.
40. Isirikyan A.A. and A.V. Kiselev, J Phys Chem, 66:205, 1962.
41. Babkin, I. Yu. and A.V. Kiselev, Russ J Phys Chem. 36:1326. 1962.
42. Ross, S., J.K. Saelens and J.P. Olivier, J Phys Chem, 66:696, 1962.
43- Elkinton, P.A. and G. Curthoys, J Phys Chem, 73:2321, 196Q.
44. Kiselev, A.V., L.A. Migunova and Y.I. Yashin, Russ J Phys Chem, 42:644,
1968.
45- Risby, T.H., R.E. Yasbin and S.S. Lestz, Proceedings of the International
Symposium on the Health Effects of Diesel Engine Emission, Vol. 1, 1980,
p. 359.
46. Sigsby, J.E., personal communication, U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina, 1978.
47. Brunauer, S., L.S. Deming, W.E. Deming and E. Teller, J Am Chem Soc,
62:1723, 1940.
48. Gale, R.L., R.A. Beebe, Russ J Phys Chem, 68:555, 1964.
49. Bradow, R., personal communication, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, 1978-
50. Kiselev, A.V. and Y.I. Yashin, Gas-Ad so rptior. Chromatography, New York:
Plenum Press, 1969-
51. Ingram, D.J.E., Proceedings of the Third Carbon Conference., 93, 1959«
101
-------
52. Pastor, B.C., J.A. Weil, T.K. Brown and J. Turkevich, Phys Rev. 102:918,
1959-
53. Collins, R.L., M.D. Bell and G. Kraus, J of Appl Phys, 30:56, 1959.
54. Uebersfeld J. and E. Erb, Proceedings of the Third Carbon Conference,
103, 1959.
55. Singer, L.S., Proceedings of the Fifth Carbon Conference, 37, 1963-
56. Petrakis L. and D.W. Grandy, Anal Chem, 50:303, 1978.
57. Singer, H.L., Proceedings of the Fifth Carbon Conference, 37, 1963-
58. Environmental Protection Agency, "Health Effects Associated with Diesel
Exhaust Emissions, Literature Review anc* Evaluation." Springfield VA:
National Technical Information Services, November 1978.
59. Lipkea, W.H., J.H. Johnson and C.T. Vuk, "The Physical and Chemical
Characterization of Diesel Particulate Emissions, Measurement Techniques
and Fundamental Considerations", SAE Paper 780108, February 1978.
60. Pitts Jr., J.N., K.A.V. Cauwenberghe, D. Grosjean, J.P. Schmid, D.R.
Fitz, tf.L. Belser, Jr., G.B. Knudson and P.M. Knudson, Science 202:515,
1978.
61. Cctton F.A. and G. Wilkinson, Advanced Inorganic Chemistry, 3rd Edition,
New York: Wiley-Interscience, 1972, p. 344.
62. Bradow, R.L., Physical and Chemical Characteristics of Diesel Emissions,
personal communication. Mobil Source Emissions Resources Branch,
Research Triangle Park, North Carolina, 1979.
63. Ames, B.N., J. HcCann and E. Yamasaki, Mutat Res. 31:347, 1975-
64- DeSerres, F.J. and M.D. Shelby, Environ Mutagen, 1:87, 1979-
65- DeSerres, F.J. and M.D. Shelby, Science, 203:563, 197g.
66. McCann, J., N.E. Spengarn, J. Kobari and B.N. Ames, Proc Matl Acad Sci,
72:979, 1975-
67. Spizizen, J., Proc Natl Acad Sci. 44:1072, 1958.
68. Czygan, P., H. Green, A.J. Garro, F. Hutterer, F. Schaffner, H. Popper,
P. Rosenthal and D.Y. Cooper, Car.cer Res, 33:2983, 1973.
69. Yasbin, R.E. and R. Miehl, Appl Environ Microbiol, 39:854, 1980.
70. Yasbin, R.E., Mol Gen Genet, 153:211, 1977.
102
-------
•71. Yasbin, R.E., G.A. Wilson and F.E. Young, J Bacteriol, 121:296, 1975*
72. Dukovich, M., R.E. Yasbin, S.S. Lestz, T.H. Risby and R.B. Zweidinger,
Environ Mutagenesis, 3:253, 1981.
103
-------
List of Publications which have resulted from this research
1. Dukovich, M., "The Mutagenic and "SOS" - Inducing Activity of Diesel
Particulate." M.S. Thesis, The Pennsylvania State University, November
1981.
2. Herr, J.D., "Diesel Particulate Emissions from Synthetic Inlet
Atmosphere." M.S. Thesis, The Pennsylvania State University, March 1982.
3. Ross, M.M., "Physicochemical Properties of Diesel Particulate Matter."
Ph.D. Thesis, The Pennsylvania State Univeristy, August 1981.
4. Yergey, J.A., "Chemical Characterization of Organic Adsorbates on Diesel
Particulate Matter." Ph.D. Thesis, The Pennsylvania State University,
November 1981.
5. Risby, T.H., R.E. Yasbin and S.S. Lestz, "'Diesel Particulate Matter
Chemical and Biological Assays." International Symposium on Health
Effects of Diesel Engine Emissions, Vol. 1, No. 359, 1980.
6. Dukovich, M., R.E. Yasbin, S.S. Lestz, T.H. Risby and R.B. Zweidinger,
"The Mutagenic and Carcinogenic Potential of the Soluble Organic Fraction
Collected from Diesel Particulate Emissions." Environ Mutagenesis,
3:253, 1981.
7- Ross, M.M., T.H. Risby, S.S. Lestz and R.E. Yasbin, "Isosteric Heats of
Selected Compounds on Diesel Particulate Matter." Environ Sci Techno1,
16; 75, 1982.
8. Dukovich, M., R.E. Yasbin, S.S. Lestz, T.H. Risby and J.A. Yergey, "Role
of Diesel Fuel Additives in the Mutagenic and Carcinogenic Potential of
the Soluble Organic Fraction Extracted from Diesel Particulate Emissions.'
submitted to Environ Mutagenesis.
9- Ross, M.M., M.R. Chedekel, T.H. Risby, S.S. Lestz and R.E. Yasbin,
"Electron Paramagnetic Resonance Spectrometry of Diesel Particulate
Matter." Environ Int. in press.
10. Ross, M.M., T.H. Risby, W.A. Steele, S.S. Lestz and R.E. Yasbin,
"Physicochemical Properties of Diesel Particulate Matter." Colloids
Surfaces, in press.
11. Yergey, J.A., T.H. Risby and S.S. Lestz, "Chemical Characterization of
Organic Adsorbates on Diesel Particulate Matter." 'Anal Chem,. 54-; ^54,
1982.
12. Herr, J.D., M. Dukovich, J.A. Yergey, S.B. Tejada, S.S. Lestz and T.H.
Risby, "The Role of Nitrogen in the Observed Direct Microbial Mutagenic
Activity for Diesel Engine Combustion in a Single Cylinder DI Engine."
SAE, No. 820467,. Detroit MI, February 26, 1982.
13. Risby, T.H., R.E. Yasbin and S.S. Lestz, "Diesel Particulate Matter
Chemical and Biological Assays-" Environ Int, 5:269-279, 1981.
104
-------
------- |