REVIEW OF DIESEL PARTICULATE MATTER SAMPLING

METHODS

FINAL REPORT

Prepared by

David B. Kittelson, Ph.D., Megan Arnold and

Winthrop F. Watts, Jr., Ph.D.

University of Minnesota
Department of Mechanical Engineering
Center for Diesel Research
Minneapolis, MN

January 14, 1999

TABLE OF CONTENTS

ABSTRACT..........................................................................................................................3
INTRODUCTION ................................................................................................................5
DIESEL ENGINE TECHNOLOGY AND EMISSION REGULATIONS .............................7
PHYSICAL AND CHEMICAL NATURE OF DIESEL AEROSOL .....................................9
AEROSOL SIZE DISTRIBUTION.....................................................................................11
DIESEL AEROSOL DYNAMICS ......................................................................................13
Diesel Particle Dynamics and Behavior ............................................................................13
Particle-Wall Interactions - Losses of Particles to Sampling Surfaces ...............................13
Particle Dynamics and Transformations............................................................................15
Effects Of Test Conditions on Diesel Aerosol ..................................................................16
Dilution and Saturation Ratios .........................................................................................17
Engine Emission Characteristics.......................................................................................21
Residence Time ...............................................................................................................22
On-Road Exhaust Dilution...........................................................................................23
Laboratory Exhaust Sampling Dilution.........................................................................23
INSTRUMENTS FOR MEASUREMENT OF DIESEL PARTICLE MATTER..................24
Critical Aerosol Parameters .............................................................................................24
Instruments Used to Measure Diesel Aerosol Size Distributions.......................................25
Summary of Particle Measurement Instruments................................................................27
EVALUATION OF ROADWAY, TUNNEL, AND OCCUPATIONAL TEST METHODS 29
EVALUATION OF DIESEL TEST LABORATORY METHODS......................................40
Exhaust Constituents .......................................................................................................41
Emission of Carbonaceous Agglomerates and Volatile Material ...................................41
Ash Constituents in Exhaust.........................................................................................46
Exhaust Sampling Conditions ..........................................................................................48
Relative Humidity and Temperature of Dilution Air .....................................................48
Dilution Ratio and Residence Time ..............................................................................49
Conclusions .....................................................................................................................50
RECOMMENDATIONS ....................................................................................................50
SPARK IGNITION (SI) MEASUREMENTS .....................................................................51
FUTURE RESEARCH........................................................................................................53
ACKNOWLEDGEMENTS.................................................................................................54
BIBLIOGRAPHY ...............................................................................................................54

01/14/99 Page 2
ABSTRACT

Federal government regulations limit emissions from heavy-duty, on-highway diesel engines.
These regulations have reduced diesel particulate matter emissions (DPM) in the past few
years. To meet these regulations diesel engine manufacturers have developed low-emission
diesel engines. Recent laboratory tests conducted at Michigan Technological University by
Bagley, et al. (1996) showed that a prototype, low-emission engine produced increased
numbers of nanoparticles when measured under typical laboratory conditions, however it is
unclear whether this increase was due to the engine or to the laboratory test conditions.
Similar increases have been observed at the University of Minnesota and at other institutions
but the test conditions were not the same.

Many factors influence the size distribution of DPM and the production of nanoparticles.
However, comparing data from one laboratory to another is very difficult because there is no
standard method for diluting, sampling and measuring the size distribution of diesel exhaust
aerosol. Engine type is but one factor affecting the DPM size distribution, others include:
• the instrumentation used to measure the size distribution
• the method of dilution and the dilution ratio
• temperature and composition of dilution air
• particle residence time in the sampling system
• particle losses in the sampling system
• the chemical composition of DPM and exhaust gases
• adsorption and condensation characteristics of diluted exhaust constituents.

These factors must be considered when interpreting size distribution data gathered by different
methods.

Understanding the fundamental science underlying particle formation and measurement is

critical to research organizations such as universities, health and environmental organizations,
and local, state and federal governments. Future low emission engines must be designed with
an understanding of particle formation and measurement. Sound science dictates that future
environmental decisions be based an understanding of the causes and effects of pollution.
Further, it is important that data collected in the diesel test laboratory be representative of
actual ambient exposure conditions. Environmental decisions should be based upon data that
are not compromised by possible artifacts resulting from laboratory conditions.

With this in mind, the U.S. Environmental Protection Agency sponsored a one-year project at
the University of Minnesota Center for Diesel Research to review DPM sampling methods.
Project objectives were:
• to critically evaluate methods used in diesel test laboratories to determine diesel particulate
matter size distributions under steady-state and transient operating conditions;
• to critically evaluate methods and data resulting from recent roadway studies to determine
diesel particulate matter size distributions, and compare the results to determine how well
laboratory size distributions mimic real-world conditions;

01/14/99 Page 3
• to examine the relationship between methods used for laboratory dilution and the actual
atmospheric dilution process;
• to recommend the best current method(s);
• to identify areas requiring further research;
• to publish the results in the refereed literature and make it available on the Internet.

Three reports were prepared under the EPA grant. The first report describes instruments used
to characterize diesel aerosol, summarizes information obtained from engine laboratory visits
and present results from a diesel aerosol sampling questionnaire circulated to major engine
labs. The second report reviews fundamental aerosol dynamics and non-laboratory studies
that have focused on determining the size distribution of diesel aerosol. This report
summarizes the findings of the project including recently published information not included in
the first two reports, presents general recommendations for diesel aerosol measurement and
recommendations for future research.

The major findings from the literature review are:

• Because conditions vary so greatly in diesel exhaust testing, it is critical to understand how
sampling and testing conditions may affect the nature of the diesel particulate matter being
measured.
• The strongest driving force for gas to particle conversion occurs in the critical dilution
ratio range (5 - 50:1), roughly the same range produced by the typical laboratory dilution
tunnel system 3 - 20:1. These conditions are not representative of real-world conditions.
• The balance between carbonaceous agglomerates and volatiles (hydrocarbons and sulfuric
acid) in diesel exhaust is critical to the understanding of particle dynamics.
• To correctly interpret diesel aerosol measurements and to correlate roadway and
laboratory experiments, the relationships between exhaust aerosol and testing parameters
must be well understood.
• Selection of aerosol instrumentation should be based on particle size and aerosol
characteristics, the goals of the study, and the respective limitations of each instrument.

Laboratory techniques that result in diesel aerosol measurements similar to ones obtained
under real-world conditions need to be developed and evaluated. This will allow future
regulatory decisions to be based on laboratory data that differentiate solid and liquid
(hydrocarbon and sulfate) aerosols and dilution conditions that are representative of actual
tailpipe and on-highway conditions.

01/14/99 Page 4
INTRODUCTION

Federal government regulations limiting emissions from heavy-duty, on-highway diesel

engines have driven DPM emissions measured on the Federal Transient Cycle from 0.6
g/bhp-hr in 1988 to less than 0.1 g/bhp-hr in 1994 and 0.05 g/bhp-hr for urban buses in 1996.
Diesel engine manufacturers have developed low-emission diesel engines to meet these
regulations. Many of the engine modifications that reduce DPM emissions from on-highway
engines are also applied to off-road engines.

Recently, Michigan Technological University (MTU) conducted a laboratory study to

characterize the physical and chemical composition and the mutagenicity of emissions from
two heavy-duty diesel engines operated with low-sulfur diesel fuel at a constant speed under
two different steady-state load conditions, with and without an exhaust aftertreatment devices.
One interesting result reported by Bagley, et al. (1996) was that the newer engine, meeting
1991 emissions levels1, significantly reduced emissions, but dramatically increased the number
and volume of very small nuclei mode particles2 when compared to a similar engine meeting
1988 emissions levels3. The research was conducted using a prototype 1991 Cummins
LTA10 engine, with a high pressure, mechanically controlled, unit fuel injection system and
other design features now commonly used in heavy-duty, high-speed, on-highway diesel
engines. The particular engine configuration used for this research was a pre-production
prototype developed to meet 1991 Federal on-highway emissions limits. Up to 40 % of the
volume of particles in the exhaust was found to be present within the nuclei-mode range of
particle-size. It is important to note that the MTU data were generated under steady-state
conditions. Outside of the laboratory, most DPM is generated under transient operating
conditions. Dilution conditions used in the Federal Heavy Duty Test Procedure are quite
different from on-highway dilution condition, thus transient particle size measurements based
on this test procedure are also not entirely representative of real-world conditions. Therefore
there is a need to gather field data representative of real-world conditions to confirm the MTU
findings.

The large concentration of nuclei mode particles, accounting for a large percentage of the
total volume of particles in the exhaust, differs from most past reported laboratory results.
Previous research measuring the size distribution of DPM from diesel engines found
distributions which were bimodal and lognormal (Abdul-Khalek, et al., 1995, Baumgard, et
al., 1985, Kittelson et al., 1991). Typically, the majority of the mass or volume

1 0.25 g/hp-hr particulate matter emissions and 5.0 g/hp-hr NOx emissions.
2 The nuclei mode particles present in diesel exhaust primarily consist of condensed organic compounds,
condensed sulfuric acid, and primary soot particles. Nuclei mode particles typically have diameters of
approximately 5 nm to 50 nm, with a volume mean diameter of approximately 20 nm.
3 0.60 g/hp-hr particulate matter emissions and 6.0 g/hp-hr NOx emissions.

01/14/99 Page 5
concentration4 of the DPM was found within the accumulation mode5 of the particle size
distribution as opposed to the nuclei mode. However, the engines and sampling conditions
used in these studies were not of the same type as used by MTU. On the other hand, roadway
studies done in the late 1970s and early 1980s reported nearly as large fractions of aerosol in
the nuclei mode as MTU (Whitby, et al., 1975 and Kittelson, et al., 1988).

Many factors influence the size distribution of DPM, and comparing data from one laboratory
to another is very difficult because, as illustrated by the lab survey presented in EPA
Supplemental Report No. 1 (Kittelson, et al., 1998a), there is no standard method for diluting,
sampling and measuring the size distribution of diesel exhaust aerosol. Engine type is but one
factor affecting the DPM size distribution, others include:
• the instrumentation used to measure the size distribution
• the method of dilution and the dilution ratio
• temperature and composition of dilution air
• particle residence time in the sampling system
• particle losses in the sampling system
• the chemical composition of DPM and exhaust gases
• adsorption and condensation characteristics of diluted exhaust constituents.

These factors must be considered when interpreting size distribution data gathered by different
methods.

Understanding the fundamental science underlying particle formation and measurement is

critical to research organizations such as universities, health and environmental organizations,
and local, state, and federal governments. Future low emission engines must be designed with
an understanding of particle formation and measurement. Sound science dictates that future
environmental decisions be based on understanding the causes and effects of pollution.
Further, it is important that data collected in the diesel test laboratory be representative of
actual ambient exposure conditions. Environmental decisions should be based upon data that
are not compromised by possible artifacts resulting from laboratory conditions.

With this in mind, the U.S. Environmental Protection Agency sponsored a one-year project at
the University of Minnesota Center for Diesel Research to review DPM sampling methods.
Project objectives were:
• to critically evaluate methods used in diesel test laboratories to determine DPM size
distributions under steady-state and transient operating conditions;

4 For particles of a known density, particle mass concentrations can be determined from particle volume
concentrations.
5 The accumulation mode particles typically have diameters ranging from approximately 50 nm to 1.0 µm,
with a volume mean diameter of approximately 100 to 300 nm. Accumulation mode particles in diesel
exhaust are thought to primarily consist of agglomerated carbon soot particles with adsorbed volatile
compounds.

01/14/99 Page 6
• to critically evaluate methods and data resulting from recent roadway studies to determine
DPM size distributions, and compare the results to determine how well laboratory size
distributions mimic real-world conditions;
• to examine the relationship between methods used for laboratory dilution and the actual
atmospheric dilution process;
• to recommend the best current method(s);
• to identify areas requiring further research;
• to publish the results in the refereed literature and make it available on the Internet.

Under this program, the Center for Diesel Research reviewed the literature, conducted an
engine laboratory survey, visited a number of engine labs and prepared three reports. The
first report included:
• a description and characterization of the aerosol instruments used to measure diesel
aerosol, and
• reported findings from the survey of engine laboratories.

The second report included:

• an overview of diesel aerosol chemical and physical characteristics,
• a discussion of the processes that affect measurement of diesel aerosol,
• illustrated how various parameters influence the size distribution of aerosol particles
obtained under different experimental and real-world conditions.

This report summarizes the findings from the project, including recently published information
not included in the other reports, presents general recommendations for diesel aerosol
sampling and recommendations for future research. Major sections include:
• Diesel engine technology and regulations
• Physical and chemical nature of diesel aerosol
• Aerosol size distribution
• Diesel aerosol dynamics
• Instruments for measurement of diesel particulate matter
• Evaluation of roadway, tunnel and occupational test methods
• Evaluation of diesel test laboratory methods
• Recommendations
• Spark ignition measurements
• Future research.

DIESEL ENGINE TECHNOLOGY AND EMISSION REGULATIONS

Interpreting and comparing diesel aerosol studies conducted over the last 30 years is
confounded by changes that have taken place in engine and aftertreatment technology and fuel
quality. Many of these changes took place because of increasingly stringent State and Federal
regulations limiting diesel emissions. Early diesel engines maximized fuel economy and
performance, but Federal regulations required engine manufacturers to reduce emissions. In
general manufacturers have met the standards by reducing emissions without decreasing

01/14/99 Page 7
engine performance. Engine manufacturers improved engine design by modifying the fuel
injection system, the combustion chamber and improving engine controls. These
improvements enhanced combustion. Improved seals and rings reduced oil consumption,
which reduced the organic component of diesel particulate emission. Electronic control
systems were introduced to precisely control fuel injection. These changes resulted in reduced
DPM emissions.

In 1993, the EPA established a fuel sulfur level of 0.05 wt % for fuel used on-highway. This
represented a dramatic decrease in fuel sulfur content, which was typically at or above
0.3 wt % prior to the regulation. As a result of this regulation, the sulfur content of DPM was
also reduced. As shown in the lab survey, all of the U.S. labs surveyed report using low sulfur
fuel.

Table 1 lists changes in Federal heavy-duty diesel emission standards and table 2 summarizes
changes in diesel engine technology and fuel over the last three decades. For the most part
engine manufacturers have tried to avoid the use of expensive aftertreatment technology while
focusing on improving engine technology. A review of aftertreatment devices and their
impact on emissions can be found elsewhere (Johnson, et al., 1994).

Table 1. Federal emission standards for oxides of nitrogen and

particulate matter for vehicles powered by heavy-duty diesel engines.
Oxides of Nitrogen, Particulate Matter,
g/bhp-hr g/bhp-hr
Engine Year Truck and Bus Truck Bus
1985 10.7 None None
1988 10.7 0.6 0.6
1990 6.0 0.6 0.6
1991 5.0 0.25 0.25
1993 5.0 0.25 0.10
1994 5.0 0.10 0.07
1996 5.0 0.10 0.05 (0.07 in use)
1998 4.0 0.10 0.05/0.07
Source: Nauss, K. M. et al. (1995)

01/14/99 Page 8
Table 2. Changes in diesel engine technology and fuel over time
Technology 1970's 1980's 1990's
Mostly naturally Mostly turbo-charged /
aspirated Mostly turbo-charged air to air aftercooling
Heavy-duty diesel Some turbo-charged / /water to air aftercooling 4 stroke cycle Direct
engine aftercooling 4 stroke cycle injection
Both 4 and 2 stroke Direct injection
cycles in use
Direct injection
predominates
Light-duty diesel Indirect injection Indirect injection Indirect injection, small
engine direct injection
introduced
Fuel sulfur content >= 0.3 wt % >= 0.3 wt % <= 0.05%wt % (1993)
Source: Sawyer and Johnson, (1995)

PHYSICAL AND CHEMICAL NATURE OF DIESEL AEROSOL

Diesel aerosol consists mainly of highly agglomerated solid carbonaceous material and ash,
and volatile organic and sulfur compounds. The structure is illustrated schematically in
figure 1. Solid carbon is formed during combustion in locally fuel rich regions. Much of the
carbon is subsequently oxidized with the residue exhausted in the form of solid agglomerates.
A tiny fraction of the fuel and atomized and evaporated lube oil escape oxidation and appear
as volatile or soluble organic compounds (generally described as the soluble organic fraction,
SOF) in the exhaust. The SOF contains polycyclic aromatic compounds containing oxygen,
nitrogen, and sulfur. Most of the sulfur in the fuel is oxidized to SO2, but a small fraction is
oxidized to SO3 that leads to sulfuric acid and sulfate aerosol. Metal compounds in the fuel
and lube oil lead to a small amount of inorganic ash.

Figure 2 shows the composition of the particulate matter for a current technology diesel
engine tested using the U.S. Heavy Duty Transient Test. The sulfuric acid/sulfate fraction is
roughly proportional to the fuel sulfur content. The fraction associated with unburned fuel
and lube oil (SOF) varies with engine design and operating condition. It can range from less
than 10 % to more than 90 % by mass. SOF values are highest at light engine loads when
exhaust temperatures are low.

01/14/99 Page 9
 Solid Carbonaceous / Ash Particles with Adsorbed
Hydrocarbon / Sulfate Layer

Sulfuric Acid Particles

0.3 µm
Hydrocarbon / Sulfate Particles

Figure 1. - Typical structure of engine exhaust particles (Kittelson, 1998a).

Particle Composition For A

Heavy-Duty Diesel Engine
Tested in Heavy-Duty Transient Cycle

Unburnt Fuel
7%

Unburnt Oil Carbon

25%
41%

Sulfate and Water

14% Ash and Other
13%

Figure 2. - Particle composition for a heavy-duty diesel engine tested in a

heavy-duty transient cycle (Kittelson, 1998a). [Composition varies widely with
engine design.]

01/14/99 Page 10
AEROSOL SIZE DISTRIBUTION

Figure 3 shows the idealized diesel aerosol number and mass weighted size distributions
(Kittelson, 1998a). The distributions are trimodal and lognormal in form. The concentration
of particles in any size range is proportional to the area under the corresponding curve in that
range. Most of the particle mass exists in the so-called accumulation mode in the 0.05 to 1.0
µm diameter range. This is where the carbonaceous agglomerates and associated adsorbed
materials reside. The nuclei mode typically consists of particles in the 0.005 to 0.05 µm
diameter range. This mode usually consists of volatile organic and sulfur compounds that
form during exhaust dilution and cooling, and may also contain solid carbon and metal
compounds. The nuclei mode typically contains 1-20 % of the particle mass and more than
90 % of the particle number. The coarse mode contains 5-20 % of the particle mass. It
consists of accumulation mode particles that have been deposited on cylinder and exhaust
system surfaces and later reentrained. Also shown in Figure 3 are definitions used to define
atmospheric particles: PM10, D (diameter) < 10 µm; fine particles, D < 2.5 µm; ultrafine
particles, D < 0.10 µm; and nanoparticles D < 0.05 µm or 50 nm. Definitions of ultrafine and
nanoparticles are not universally agreed upon. Note that by number, nearly all of the particles
emitted by a diesel engine (the same is true for a spark ignition engine) are nanoparticles.

0.2 1
Nanoparticles Fine
0.18 Dp < 50 nm Particles 0.9
dlogDp

Dp < 2.5µm
Ultrafine Particles
0.16 0.8

Alveolar Deposition Fraction

Dp < 100 nm PM10
Dp < 10 µm
0.14 0.7

0.12 Fractional deposition of particle 0.6

3
with density of 1 g/cm
0.1 0.5

0.08 0.4

0.06 0.3

0.04 0.2
Nuclei Accumulation Coarse
0.02 Mode Mode Mode 0.1

0 0

0.001 0.010 0.100 1.000 10.000

µm)
Diameter (µ
Mass Weighting Number Weighting Alveolar Deposition Fraction

Figure 3. - Typical engine exhaust size distribution both mass and number
weightings are shown (Kittelson, 1998a).

For comparison, figure 4 shows the idealized atmospheric aerosol size distribution, which is
also generally trimodal and lognormal in form (Whitby and Cantrell, 1975). The figure shows
the mechanisms such as condensation and coagulation that transfer aerosol mass from one size

01/14/99 Page 11
range to another. The smallest of the three modes, 0.001 to 0.08 µm, is the Aitken nuclei
range, which consists of primary aerosol from combustion sources such as diesel engines and
secondary aerosol formed from coagulation of primary aerosols to form chain agglomerates.
The next size range, 0.08 to 1.0 µm is the accumulation range, that contains emissions in this
size range plus aerosol accumulated by mass transfer through the coagulation and
condensation processes from the nuclei range. The last range, 1.0 to approximately 40 µm is
referred to as the coarse aerosol size range. Aerosols within this range generally originate
from mechanical processes such as grinding, mechanical fracture and bulk material handling.
There is very little exchange of mass from the nuclei and accumulation modes to the coarse
particle mode under most conditions.

Figure 4. - Idealization of an atmospheric surface area distribution showing

the principal modes, sources of mass, and the processes involved in mass
transfer and removal (Whitby and Cantrell, 1975).

Figure 5 illustrates the fate of exhaust particles in the atmosphere and how the processes of
nucleation, condensation and adsorption affect the formation, dispersion and deposition of
exhaust aerosols.

01/14/99 Page 12
Figure 5. - Schematic illustrating the fate of exhaust particles in the
atmosphere and how the processes of nucleation, condensation, adsorption
affect the formation, dispersion and deposition of exhaust aerosols.

DIESEL AEROSOL DYNAMICS

Diesel Particle Dynamics and Behavior

Measurement of diesel aerosol is affected by three main parameters: the environmental

conditions experienced by the emissions, the sampling/measurement system used to
characterize the emissions, and the chemical and physical composition of the engine emissions.
An understanding of how exhaust conditions interact with exhaust constituents is critical if the
aerosol size distribution and composition are to be determined.

Many mechanisms affect aerosol behavior during sampling and measurement. These are
described in numerous textbooks such as Hinds, (1982). These mechanisms can be divided
into two categories: (1) wall interactions with the aerosol and (2) particle dynamics and
transformations during sampling. The various mechanisms in the two categories are listed
below and a more in-depth description of each of these mechanisms is presented in EPA
Supplemental Report No. 2 (Kittelson, et al., 1998b).

Particle-Wall Interactions - Losses of Particles to Sampling Surfaces

Table 3 summarizes the impact of the various particle-wall interactions that affect particle
measurement and recommends techniques to minimize their impact.

01/14/99 Page 13
Table 3. Particle – Wall Interactions

Mechanism Impact Recommendations

• Removes small number of the • Avoid sharp changes in sample
Inertial largest particles (represents a line diameter or direction
impaction larger fraction of the total • Isokinetic sampling not as
particulate mass) important for submicron
• Introduces variability through particles
unpredictable reentrainment of
deposited mass
• Inertial and gravitational losses
are quite small for particles in
0.1-1.0 µm range
• Electric charge build-up on • Use metal sample lines to avoid
Electrostatic sample lines attracts charged electrostatic build-up
deposition diesel particulate matter, • Avoid Teflon sample lines,
causing it to deposit on line prone to large build-up of
walls electrostatic charge

• Large temperature differences • Insulate / heat sample lines to

Thermophoretic between exhaust gas and sample avoid large differences in
deposition lines causes thermal driving temperature between lines and
force depositing particles on line exhaust gas
walls
• Most significant losses occur
during transient duty cycles
• Diffusion for particles in the size • Difficult to eliminate diffusional
Diffusional range that represents the deposition
deposition majority of the mass is negligible • Use short sampling lines
(> 0.1 µm)
• Diffusion for particles (< 0.03
µm) in the size range that
represents the majority of
number is also small, but should
be considered for the smallest
size ranges measured
• Minimal impact on diesel • Avoid unnecessarily long
Gravitational particulate losses horizontal sections of sample
deposition • May introduce variability lines where large particles may
through unpredictable settle
reentrainment of small deposited
mass
Source: Kittelson, D. B., and J.H. Johnson (1991)

01/14/99 Page 14
Heat transfer during sampling and dilution cools the exhaust and helps in meeting maximum
temperature requirements. However it results in thermophoretic deposition of particles on
sampling and dilution system surfaces. Eventually, these deposits are reentrained in the
exhaust stream. Reentrainment is unpredictable and increases variability in mass
measurements because of the increase in the number of coarse (> 1,000 nm) particles. These
particles are not necessarily representative of diesel aerosol and make aerosol size distribution
measurement more difficult. Another problem related to temperature is the condensation of
volatile matter (hydrocarbons and sulfates) on cooler dilution tunnel walls. These volatiles are
precursors of new nanoparticles that form when the walls are heated volatizing the material.

In their review of variability in particle emission measurements during heavy duty transient
tests, Kittelson and Johnson (1991) discussed the impact of exhaust system temperatures on
particle measurement and provided recommendations to minimize the effects on aerosol
sampling. They calculated losses during the heavy-duty transient test for a typical test facility
(Caterpillar). Calculated mechanical losses in the accumulation mode size range were less
than 0.2%, but total losses of about 5% were predicted when thermophoretic deposition was
included.

A recent study by Ahlvik, et al. (1998) calculated that sampling losses due to mechanical
deposition (diffusive, gravitational, inertial, aspirational) were less than 5 % for tests
conducted in a light-duty chassis dynamometer test facility and a heavy-duty test cell.

Particle Dynamics and Transformations

Particle dynamics and transformations also impact aerosol measurement and are well
understood qualitatively. These processes include coagulation, nucleation,
adsorption/desorption, and condensation/evaporation. However, the application of those
theories to explain or predict diesel exhaust behavior during sampling and measurement is not
easy. Table 4 summarizes these particle processes and their impact on particle measurement.
Further discussion of how these particle processes can affect diesel exhaust aerosol
characteristics is presented in subsequent sections. Included for each particle process is a brief
statement of the impact of the mechanism on the aerosol being measured.

01/14/99 Page 15
Table 4. Particle Dynamics and Transformations

Process Impact
• Dependent on particle size and concentration
• Does not affect total particle mass
• Causes decrease in particle number concentration and increase
Particle coagulation in particle size
• Increase in particle size may affect loss mechanisms
• May affect diesel aerosols if dilution is delayed, not critical
after typical diesel exhaust dilutions
• Typical time constant, τ = 1/kNo ~ 109/No (s) for diesel size
particles, No = initial particle concentration (1/cm3) (Fuchs,
1964)
• Adsorption / desorption of volatile components will affect size
and mass of measured particulate matter
Adsorption/desorption • Availability of particulate surface will affect degree of
adsorption / desorption
• Driven by saturation ratio
Nucleation • Homogeneous nucleation may create large numbers of new
particles
• Nucleation rates are highly nonlinear functions of saturation
ratio
• Heterogeneous nucleation leads to the growth of existing
particles
• Condensation / evaporation of volatile constituents will affect
Condensation/ size and mass of measured particulate matter
evaporation • Affected by saturation ratio, testing conditions such as:
temperature, pressure, humidity
• Particles formed by nucleation may grow by condensation

Effects Of Test Conditions on Diesel Aerosol

The environmental conditions experienced by diesel exhaust after leaving the engine tailpipe
coupled with the characteristics of the engine emissions will influence the nature of the
measured aerosol. It is critical to understand this interaction if sampling conditions in the
laboratory are going to produce results similar to those generated under real-world conditions.

Because conditions vary so greatly in diesel exhaust testing, it is critical to understand how
testing conditions may affect the nature of the diesel particulate matter being measured.

The key environmental or test parameters that have been identified include:
• dilution ratio and the resulting saturation ratio
• dilution rate
• time the particles spend at varying dilution ratios or sampling conditions (residence time)

01/14/99 Page 16
• humidity and temperature
• background particle and gas concentrations

Emission characteristics that have a role in affecting the measured aerosol include:
• size, number, and composition of particles emitted from the engine
• composition and quantity of volatile particle precursors in the exhaust

Dilution and Saturation Ratios

To interpret diesel aerosol measurements and to reproduce environmental aerosol conditions

in the laboratory, it is necessary to understand how the dilution ratio affects the saturation
ratio, the exhaust temperature and the volatile fraction of the exhaust. As shown in the
laboratory survey, the type of dilution, dilution ratio and the environmental conditions vary
between facilities. Further, these conditions can be vastly different from the dilution and
environmental conditions experienced on-highway, resulting in an aerosol with markedly
different chemical and physical characteristics. The parameters that affect the gas-to-particle
processes of adsorption/desorption, condensation/evaporation, and nucleation are summarized
in this report. EPA Supplemental Report No. 2 (Kittelson, et al., 1998b) presents a more
detailed discussion of these processes. This section will relate the dilution ratio to the
resulting particle processes that may occur.

Exhaust Sampling System Dilution

Regulations for the Federal Test Procedure define DPM by specifying the maximum
temperature in the primary mixing zone (190° C, 375° F), and the maximum temperature of
collection at the filter face (52° C, 125° F) for DPM samples. The filter face temperature may
be achieved by single or two-step dilution. Some test facilities use additional dilution for
aerosol instruments that require lower particle concentrations. There are no set guidelines for
the amount of dilution at each step or the residence times of exhaust at the various dilution
ratio values. This wide latitude in achieving the desired conditions for an exhaust sample
introduces a wide variance in DPM measurements especially those measurements to determine
the aerosol size distribution.

Volatile Particle Precursors in the Exhaust

One of the important parameters to consider in characterizing diesel aerosol is the amount and
fate of volatile organic matter. The fate of the hydrocarbons is affected by the dilution
process, laboratory or atmospheric. The resulting dilution air/exhaust mixture temperatures
and organic concentrations during dilution will affect the saturation ratios and aerosol
behavior during sampling or during particle lifetimes in the atmosphere. Volatile organic
matter is associated with the smallest diesel exhaust particles (nanoparticles) and
understanding processes that affect the fate of these exhaust hydrocarbons is becoming
increasingly important.

01/14/99 Page 17
Volatile organic matter may remain gaseous (disregarding the use of a catalytic converter) or
may undergo gas-to-particle conversion to form the SOF, the extractable fraction of DPM.
As engine exhaust is diluted and cooled, SOF may move from the gas phase to the particle
phase by two paths: adsorption on or absorption into existing particles or nucleation to form
new particles. All of the hydrocarbons that are collected on a filter are called the SOF.

Adsorption of hydrocarbons from the gas phase to existing aerosol particles will increase the
size and mass of nuclei and accumulation mode diesel aerosol while decreasing the gaseous
hydrocarbon concentration. Nucleation of new particles formed mostly of hydrocarbons from
the gas phase will leave the accumulation mode particles unaffected, but may cause a dramatic
increase in the number nanoparticles in the nuclei mode.

Adsorption or Nucleation of Volatile Particle Precursors

Adsorption of hydrocarbons on the surface of the carbonaceous core of diesel particles may
be modeled using the Langmuir, BET, or other theories of isothermal adsorption behavior.
The saturation ratio, S, is the ratio of the partial pressure of a condensable species to its local
vapor pressure. For both models, increasing S (at low values of saturation ratio) causes a
corresponding increase in organic adsorption to available surface area. For intermediate
values of S, there is a slow increase in adsorption for increasing saturation ratio. As S
approaches 1.0 adsorption gives way to condensation. Supplemental Report No. 2 (Kittelson,
1998b) provides more information on the modeling of adsorption of hydrocarbons on soot
particles and includes illustrations of the relation of adsorption to saturation ratio.

If saturation ratios are sufficiently high, new particles may form by homogeneous nucleation
of gaseous species. High saturation ratios are more likely if there is insufficient particle
surface available in the exhaust for extensive adsorption. Nucleation can be responsible for
the formation of nuclei-mode particles in diesel aerosol. The nuclei mode is frequently
observed when diesel exhaust is rapidly diluted under either atmospheric or dilution tunnel
conditions (Dolan, et al., 1975; Kittelson et al., 1988). Nuclei-mode particles consist
primarily of volatile materials and will be affected by changing saturation conditions that may
cause continued condensation or evaporation back into gaseous form.

It has been observed in diesel exhaust that volatile organic matter can nucleate and grow to
form nuclei mode particles at saturation values below supersaturation. There are
heterogeneous nuclei present in diesel exhaust in the form of sulfuric acid and possibly
metallic ash. It has been hypothesized that these heterogeneous nuclei are acting as the
nucleation sites for the SOF and that the particles are growing by adsorption onto the metallic
ash or absorption into the sulfuric acid. This would allow the growth of SOF nuclei mode
particles at saturation values below 1.0. Calculations of saturation ratios associated with
hydrocarbons comprising the SOF, suggest that they rarely approach 1.0, and never reach
values of 3 - 4 necessary for homogeneous nucleation (Friedlander, 1977). Thus, sulfuric acid
nuclei are likely precursors to the formation hydrocarbon containing nanoparticles.

01/14/99 Page 18
Doyle (1961) predicted that homogenous nucleation between sulfuric acid (H2SO4) and H2O
may be triggered in the atmosphere with a minute partial pressure of H2SO4 of 10-9 Torr at
50 % relative humidity and a temperature of 298 K. This was also supported in the work of
Mirabel and Katz (1974) and Hamill (1976). They also predicted a critical cluster size of
about 0.7 nm. In a typical diesel exhaust without any after-treatment device, sulfuric acid
concentration may range from 5-20 PPM (this is based on 0.04 % sulfur mass in fuel). This
concentration of H2SO4 is sufficiently high to trigger nucleation during dilution and cooling.
Baumgard (1996) used a similar approach to Mirabel and Katz and predicted rapid nucleation
of sulfuric acid in a dilution tunnel with a dilution ratio of 10. In the current work of Abdul-
Khalek (1999), it is also shown that sulfuric acid may nucleate in mini-dilution systems at
dilution ratios between 10-50.

Saturation ratios of both hydrocarbons and sulfuric acid behave in essentially the same manner
during exhaust dilution. As the saturation ratios of a diesel exhaust increase, the favored
process changes from adsorption to nucleation. While the rate of adsorption increases with
increasing S, the rate of nucleation is extremely nonlinear in S. Nucleation can be described as
nearly an “on - off “ phenomenon. Once supersaturation is sufficiently high and the nuclei are
present, large numbers of small particles form very quickly.

Figure 6 shows the relation of saturation ratio and the competing gas-to-particle processes of
adsorption and nucleation with the onset of nucleation occurring at a critical value of
saturation ratio above which the nucleation rate rapidly increases. Thus, engine and dilution
conditions that lead to sufficiently high saturation ratios may cause nucleation to increase
dramatically.

Relative Rates of Gas to Particle Conversion

50 250

40 200

Nucleation
30 150
Rate

20 100

10 50

Adsorption
0 0
0 0.5 1 1.5 2 2.5 3

Saturation Ratio
Figure 6. - Figurative diagram showing the effect of saturation ratio on
competing gas-to-particle conversion processes.

01/14/99 Page 19
Figure 7 shows the relation of dilution ratio and the resulting saturation ratio for two
representative condensable hydrocarbons of a diesel exhaust. The two hydrocarbons have
different boiling points and are shown at exhaust temperatures of 480o K (207o C) and 600o K
(327o C). The dilution air temperature is 300o K (27o C) and the dilution process is assumed
adiabatic. Using these hydrocarbons as representatives of typical diesel exhaust volatiles, the
saturation ratio has the highest values for dilution ratios of about 5 to 50 for both
hydrocarbons.

Figure 7. - Effect Of Dilution Ratio and S For Two Condensable Species At

Two Exhaust Temperatures (Kittelson and Dolan, 1980)

Thus, the strongest driving force for condensation, adsorption and nucleation occurs in
the critical dilution ratio range (5 - 50:1), roughly the same range produced by a dilution
tunnel system (3 - 20:1).

01/14/99 Page 20
Engine Emission Characteristics

Engines are characterized based upon the particulate matter emissions in g/bhp-hr and by the
type of engine such as heavy-duty on or off- highway and light duty. Prior to 1988, EPA did
not regulate emissions but from 1990 to 1994 EPA required engine manufacturers to reduce
emission for heavy-duty on-highway emissions from 0.60 - 0.10 g/bhp-hr. As engine
technology improved, some engine types experienced greater reductions in solid carbonaceous
particles than in volatile materials. This was the case for the engine family tested in the HEI
study (Bagley, et al., 1996). The 1988 engine gave SOF values in the range 0.25 - .30 while
the newer 1991 engine gave values in the 0.60 – 0.75 range.

While dilution ratio is a critical parameter of the atmosphere or laboratory conditions, the
amount of particulate surface available on which volatile organic matter may adsorb is a
critical parameter of the exhaust aerosol. The amount of material adsorbed onto the
carbonaceous agglomerates is dependent on the particulate surface area available as well as
the saturation ratio.

Older engine designs that emit higher concentrations of carbonaceous agglomerates have large
surface areas available for adsorption of volatile materials. This will tend to prevent saturation
ratios from going high enough to trigger nucleation. On the other hand, with newer designs in
which the carbonaceous component has been removed more effectively than volatiles, there is
little surface area available for the adsorption of these volatiles. This drives saturation ratios
higher and makes nucleation more likely.

Figure 8 shows how the presence of carbonaceous agglomerates keeps the organic saturation
ratio low by adsorbing the SOF. The depletion of organic vapor will prevent the saturation
ratio from becoming high enough during dilution to promote nucleation.

Even for diesel engines emitting relatively low concentrations of carbonaceous agglomerates
and high hydrocarbon fractions, saturation ratios high enough for homogeneous nucleation are
not usually reached during exhaust dilution. Thus, other types of particles might serve as
heterogeneous nucleation or absorption sites. It is possible that even the reduced sulfuric acid
levels found in the exhaust of diesel engines fueled with low sulfur fuel may be sufficient to
form nuclei that absorb hydrocarbons. It is also possible that other exhaust constituents such
as metal contaminants from lubricating oil may be acting as heterogeneous nucleation sites for
nanoparticle formation. At present details of these processes are not understood.

In summary, an engine producing a large amount of volatile material but a low amount of
carbonaceous soot is likely to have more particle nucleation and growth than an engine
producing an equal amount of volatile material with more carbonaceous agglomerates.

01/14/99 Page 21
 Influence of Exhaust Soot on Saturation Ratio
Adiabatic Dilution, Equilibrium Adsorption
3

High saturation ratios may lead to nucleation of new particles

2.5
Saturation Ratio

1.5

0.5

0
1 10 100 1000

Dilution Ratio
Low Soot High Soot
Figure 8.- Saturation vs. dilution ratio for engines emitting different amounts
of DPM.

Residence Time

The kinetics of adsorption become important when the exhaust is going through rapid changes
in temperature or adsorbate concentration. For laboratory measurements, filters used to
measure DPM are usually in the exhaust stream long enough that it is assumed that the SOF
will reach equilibrium with the solids on the filter. For laboratory studies attempting to
simulate on-highway dilution conditions using near real-time instruments, the rate at which
adsorption and condensation occurs must be considered.

It is expected that the tendency to produce nuclei during dilution will be maximized at an
intermediate rate of dilution. Higher saturation ratios will be produced and nucleation will be
more likely if dilution takes place on a time scale faster than characteristic times for
adsorption. Under these circumstances, there is little time for adsorption to take place and the
exhaust behaves as if there is a much smaller surface area available for adsorption. On the
other hand, if the rate of dilution is too high, the exhaust will spend little time in the
intermediate dilution ratio range where saturation ratios and nucleation rates are highest and
nucleation will be minimized.

Calculations were made for adsorption of a representative organic at dilution conditions that
led to the highest saturation ratio. A typical SOF of 10% on DPM was assumed and an

01/14/99 Page 22
accumulation mode particle diameter of 0.1 µm was used. These assumptions gave a time
constant for adsorption/condensation mass transfer of 0.25 s. (Kittelson and Dolan, 1980)
This value provides a time scale when considering the roadway and laboratory dilution values
and residence times provided below.

On-Road Exhaust Dilution

Figure 9 shows measured roadway dilution ratios with respect to time from emission from the
engine stack for a heavy-duty diesel engine. For this study, exhaust measurements were taken
at varying distances behind a flatbed diesel truck traveling at about 55 mph. Figure 9 shows
the dilution ratio as a function of atmospheric residence time. For this study, the exhaust
passed through the range of critical dilution (5 and 50:1) in less than 0.25 s. The exhaust was
diluted about 100:1 in approximately 0.45 s and above 1000:1 in approximately 1 s.

Laboratory Exhaust Sampling Dilution

Table 5 shows typical laboratory values at primary and secondary dilution stages and the
typical residence times at each stage. The values in the table are taken from the survey of
diesel engine test facilities performed and from the literature.

Table 5. - Primary and secondary dilution values and typical residence times
Primary Dilution Ratio Range 2 – 50
Residence Time at Primary Dilution 0.5 - 2 (s)
Total Dilution Ratio Range (including 5 – 1000
secondary dilution)
Residence Time at Secondary Dilution 0.5 - 3 (s)

Laboratory aerosol measurement conditions are within the critical range of dilution ratio and
corresponding saturation ratios that will greatly affect the adsorption, nucleation and
condensation processes of the particles.

To correctly interpret diesel aerosol measurements and to correlate roadway and

laboratory experiments, the relationships between exhaust aerosol and testing parameters
must be well understood.

01/14/99 Page 23
 Roadway Study of Diesel Exhaust
Time Exhaust Spends at Critical Dilution Ratios
100

80
Dilution Ratio

60

CRITICAL DILUTION RATIOS

40

20

0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45

Time From Stack [sec]

Figure 9. - On-road diesel exhaust sampling time exhaust aerosol spent at

critical dilution ratios for a heavy-duty diesel engine in a flatbed truck
traveling at 55 mph (Kittelson, et al., 1988 adapted from figure 4.46).

INSTRUMENTS FOR MEASUREMENT OF DIESEL PARTICLE MATTER

This section summarizes real-time or near-real-time instruments used to measure DPM.

Critical aerosol parameters are listed and discussed to provide context for the instrument
discussion. The principle of operation of each instrument is listed and related to the
corresponding particle property. Further detail is provided in Supplemental Report No. 1
(Kittelson, et al., 1998a).

Critical Aerosol Parameters

The U.S. has historically regulated DPM emissions based on a particulate mass standardized
to engine output. Gravimetric analysis of DPM collected on filters placed in engine exhaust
lines has given an indication of engine output of total mass of particulate matter.

However, there is a growing recognition that other particle characteristics may have a
more significant impact on health than total particulate mass.

Historically, PM10 (particles smaller than 10 µm) has been used to define the respirable
aerosol fraction. Now regulations include standards for PM2.5 for particles smaller than 2.5
µm. With the recognition that over 90 % of the number of diesel particles reside in the range
below 0.1 µm while only 1-20 % of the mass typically resides in the same ultrafine particle
range, the parameters used to characterize diesel particles need to reflect these new concerns.

01/14/99 Page 24
Adverse health effects of respirable aerosol are a function of particle size, deposition rate,
composition and other factors. The smaller the particle the greater the penetration. Particle
toxicity is related to particle composition and physical characteristics. The smaller the
particles, the greater the surface area for the same overall particle mass respired and the higher
the potential activity of the particles in the lung. Figure 3 illustrated relationships between
idealized DPM number and mass weighted size distributions and the alveolar deposition curve
(Kittelson, 1998a, Morrow, et al., 1964, Raabe, 1982). On a number basis, nanoparticle
deposition efficiency is much greater than accumulation mode deposition efficiency, although
on a mass basis it is less. Nanoparticles are typically hydrocarbons or sulfate and form by
nucleation during dilution and cooling of the exhaust, while accumulation mode particles are
mainly carbonaceous soot agglomerates formed directly by combustion.

To determine the mass and number size distribution of diesel aerosol over the entire size range
requires more than one instrument. Depending upon the principle of operation these
instruments can determine the size distribution in a span of time ranging from seconds to
hours.

Table 6. Critical Particle Measurement Parameters

Parameter Comments
Particle mass • total particle mass may not be the best indicator for possible
health impact
Particle size • instruments measure different properties that are translated
to different particle diameters (i.e. aerodynamic diameter,
electrical mobility diameter, etc.)
Particle number • majority of number of diesel particles resides in ultra-fine
concentration size range that penetrate deepest into the lung
Particle surface area • may be a more accurate indication of degree of adverse
impact on health
Particle composition • compounds adsorbed to particle surfaces may play a more
important role in interactions with lung tissue

Instruments Used to Measure Diesel Aerosol Size Distributions

Frequently used instruments to measure diesel aerosol size distributions are briefly described
below and table 7 provides a summary. The listing of instrumentation is not an exhaustive list,
however the instruments most commonly mentioned in the literature and those instruments
that have unique capabilities for measuring the diesel aerosol size distribution or chemical
composition below 100 nm are discussed.

Micro-Orifice Uniform Deposit Impactor (MOUDI): is a cascade impactor that provides a

mass weighted size distribution of an aerosol. Particles are separated by aerodynamic
diameter on aluminum foil stages or substrates for subsequent gravimetric analysis.
Depending upon the amount of mass to be collected and the nature of the particles, the
substrates may be used with or without grease. The MOUDI sizes particles from 0.56 to 10

01/14/99 Page 25
µm using 4 stages per decade, although the number of stages can be altered depending upon
experimental conditions. Measurements are susceptible to measurement artifacts and are not
near to real-time. (Marple, et al. 1991)

Nano-Micro-Orifice Uniform Deposit Impactor (Nano-MOUDI): is a modified MOUDI.

The final four stages of the nano-MOUDI use low-pressure operation to size particles down
to 10 nm. The nano-MOUDI also provides mass-weighted size distributions. Measurements
are susceptible to measurement artifacts and are not near to real-time. (Marple, et al. 1994)

Electrostatic Low Pressure Impactor (ELPI): is a cascade impactor used in Europe that
collects particles separated by aerodynamic diameter for gravimetric analysis and provides
real-time number-weighted size distributions. The particles are charged and each stage uses
an electrometer to measure the charge of the particles deposited on the collection plate.
Measurements are near real-time with a scan time as low as 2 s. (Keskinen, et al., 1992;
Dickens, et al., 1997; Ahlvik, et al., 1998)

Electrical Aerosol Analyzer (EAA): has historically been the most common instrument used
in the U.S. for diesel particulate measurement. The EAA provides a number-weighted particle
size distribution based on electrical mobility. Diesel exhaust particles are electrically charged
and then routed through an electric mobility analyzer that only allows particles of a specified
electric mobility to pass through and be measured by a downstream electrometer. (Liu, et al.,
1974; Liu and Pui, 1974a; Dolan, et al., 1980; Groblicki and Begeman, 1979, Fang and
Kittelson, 1984; Baumgard and Kittelson, 1985; Baumgard and Johnson, 1992; Bagley, et
al.,1996)

Differential Mobility Particle Sizer (DMPS): measures an electrically charged aerosol with
an electric mobility analyzer that only allows particles of a specified electric mobility to pass
through and be measured downstream. A DMPS uses a condensation nucleus counter to
count the particles within an electric mobility range because of its increased sensitivity over an
electrometer. This allows the size interval of particles passed by the mobility analyzer and
counted to be smaller than for an EAA with an electrometer. (Pui and Swift, 1995)

Scanning Mobility Particle Sizer (SMPS): is similar to a DMPS except that while the electric
mobility analyzer of the DMPS makes discrete changes in the voltage used to control what
interval of electric mobility of particles passes through, the voltage of the SMPS is ramped
exponentially. This allows a scan of exhaust size distribution in a time interval as short as 30 s.
A SMPS also uses a condensation nucleus counter to count particles passing through the
mobility analyzer. (Wang and Flagan, 1990; Keady, et al. 1983; Greenwood, et al. 1996)

Photoelectric Aerosol Sensor (PAS): provides a surface area-weighted particle concentration

and is related to the quantity of particle-bound PAH in the exhaust. Particles are irradiated
with an UV light with enough intensity to cause the particles to emit electrons and become
positively charged. The positive particles are then collected and the charge measured by an
electrometer. The intensity of the particle photoemission is directly related to the PAH

01/14/99 Page 26
content of the particle. (Burtscher and Siegmann, 1993; Burtscher, et al., 1993; Hart, et al.
1993)

Aerosol Time-of-Flight Mass Spectrometer (ATOFMS): measures both aerodynamic size

(~0.2 to <10 µm,) and chemical composition of individual particles up to about 600 particles
per minute. A particle enters the ATOFMS and is accelerated to its terminal velocity. Smaller
particles have greater aerodynamic mobility and reach higher terminal velocities. The particles
are sized based on time of flight between two lasers. Each particle is then pulsed by a laser
and the resultant desorbed and ionized material is characterized for chemical composition by a
dual-polarity mass spectrometer. (Silva and Prather, 1997; Noble and Prather, 1996, 1998;
Gard, et al. 1997)

Epiphaniometer: measures surface concentration of an aerosol by passing the particles

through a chamber with radioactive lead isotopes that attach to the particles. The level of
radioactivity of the particles is measured and is directly proportional to the particle surface
area. (Gaggeler, et al., 1989a,b)

Summary of Particle Measurement Instruments

Table 7 summarizes current aerosol instrumentation used to measure diesel aerosol size
distributions and lists some of the limitations associated with each instrument. The particle
mass collected by the MOUDI can be analyzed for chemical composition and the ATOFMS
can provide chemical composition by particle size over a limited range. However, as shown in
the figure 10, selection of such instruments as the MOUDI and the ATOFMS will preclude
measurement of the nuclei mode of a diesel exhaust aerosol. While the nano-MOUDI
measures down into the nuclei mode particle sizes, it does not provide real or near real-time
measurements. Although the ELPI measures only the larger nuclei particles, its fast response
and ability to measure down to ambient particle concentrations account for its use in
European DPM studies6. The EAA covers both the nuclei and accumulation mode particle
size ranges, but is susceptible to error introduced by changes in the aerosol during its
relatively long sampling time. The DMPS and the SMPS both cover more of the nuclei mode
size range than the ELPI but require at least 30s to obtain a size distribution, compared to
about 2s for the ELPI. Figure 10 summarizes the typical dynamic particle size range covered
by these instruments.

As illustrated in the instrument descriptions and the instrument summary table, instruments
measure different particle characteristics based on different particle dynamics. This wide array
of instruments can be used to measure different particle size ranges and parameters.
Supplemental Report No. 1 provides a more in-depth discussion of these instruments.

Which instrument is used is dependent on the particle sizes and characteristics of interest,
the goals of the study, and the respective limitations of each instrument.

6 As illustrated by the lab survey, many of the European labs are using both the ELPI and the SMPS.

01/14/99 Page 27
Table 7. Instrument summary
Instrument Parameter Range Comments
Inertial mobility instuments (cascade impactors)
Micro-orifice Aerodynamic 0.056 - 10 um * Not real-time, gravimetric and chemical
uniform deposit diameter analysis performed after collection.
impactor (MOUDI)
Nano- MOUDI Aerodynamic 0.010 - 18 um * Same as MOUDI
* Low pressure may cause volatile particles
diameter evaporate.
Electrical low Aerodynamic 0.030 - 10 um * Near realtime with time resolution from 2 - 15 s.
pressure impactor diameter * Can measure ambient concentrations, but low
(ELPI) pressure may cause volatile particles to evaporate.
Electrical mobility instruments
Electrical aerosol Electrical Mobility 0.003 - 1 um * Time resolution between 2-3 min, 11 size intervals.
analyzer Equivalent * Strongly distorted results if concentration and size
Diameter change during scan.
* More sensitive to these changes than SMPS.
* Requires high dilution (> 200:1) to prevent
electrometer saturation.
* Results for diameters < 10 nm are questionable
unless number concentrations are high, due to
diffusional losses.
* Pre-classification stage must be used to remove
particles larger than 1.0 um for diesel aerosol
measurements
Differential mobility Electrical Mobility 0.01 - 0.5 um * Discrete changes in electric field of DMPS makes
particle sizer Equivalent sampling for entire range slower than for SMPS.
(DMPS) Diameter * Biased results if concentration and size change
during scan.
* Pre-classification stage must be used to remove
particles larger than 1.0 um for diesel aerosol
measurements
Scanning mobility Electrical Mobility 0.010 - 0.7 um * 2 - 4 min scan time for 100 size intervals - can ramp
particle sizer Equivalent DMA electric field to reduce scan time.
(SMPS) Diameter * Time resolution of a few seconds for a single
channel.
* Scan times as short as 30 s may be used with lost
resolution
* Biased results if concentration and size change
during scan.
* Pre-classification stage must be used to remove
particles larger than 1.0 um for diesel aerosol
measurements
nano-Differential Electrical Mobility 3 - 50 nm * Time resolution of about 10 s required to scan
mobility analyzer Equivalent all intervals.
(nano-DMA) Diameter * Not fully characterized and developed

01/14/99 Page 28
Table 7. Instrument summary
Miscellaneous instruments
Photoelectric aerosol Surface area - < 1 um * Material condensing on particle surface can
sensor (PAS) senses PAH bound lower limit not interfer with measurement.
to carbon surface * Does not measure particle size
* Response time < 1 s
reported * Particles > 1 um are not charged efficiently.
Aerosol time-of-flight Aerodynamic 0.3 - 50 um * Real-time measurement of up to 600 part./min.
mass spectrometer diameter and
(ATOFMS) chemical
composition
Epiphaniometer Fuchs surface < 7.0 um * Time resolution is between 5 - 30 min depending
area upon the analysis code used.
* For particles between 0.10 - 3 um the instrument
must be calibrated to determine response.
Diffusion battery Diffusion 0.01 - 0.5 um * Slow sampling time requires steady-state
mobililty conditions.
* Strongly distorted results if concentration and size
change during scan.

Particle Sizing Instrum ents

M O U D I (m a n u a l )

A e r o d y n a m ic S izing ELPI

nano-M O U D I
(m a n u a l )

M iscellaneous ATOFMS

EAA

E lectrical Mobility
SMPS
S izing

nano-SMPS

1 10 100 1000 10000 100000

S izing Range (nm )

Figure 10. - Dynamic range of particle sizing instruments

EVALUATION OF ROADWAY, TUNNEL, AND OCCUPATIONAL TEST

METHODS

A number of studies have determined the size distribution of diesel aerosol in different
environments such as mines, tunnels, roadways and ambient atmosphere. In most cases these
studies have evaluated a mixed combustion aerosol consisting of emissions from both gas and
diesel fueled vehicles. In some cases the investigators determined the mass, number and

01/14/99 Page 29
volume size distributions, and in a few cases the nanoparticle size distribution was determined.
The methods and procedures used to determine these size distributions varied as did the
conditions under which the aerosol was generated. However, despite these differences the
characteristic multi-modal size distribution is frequently observed. The results from these
studies are summarized in table 8.

01/14/99 Page 30
 Table 8. Studies Measuring Diesel Aerosol Size Distributions Outside of The Laboratory.

Type of Study and Dilution Ratio1 Measurement Results Reference

Aerosol Method(s)
Mining/Heavy Duty 100 - 150:1 MOUDI Accumulation mode by mass: 0.15 - Bureau of Mines,
Diesel 0.2 µm 1992
Mining/Heavy Duty < 100:1 CNC Aitken nuclei count: Gillies and
Diesel • No diesels 78 - 175 x 103 Schimmelpfennig,
part./cm3 1983
• With diesels 3.1 - 17.0 x 105
part./cm3
Roadway/Mixed > 500 - 1,000:1 CNC, prototype Nuclei V mode peak at 0.02 µm Whitby, et al., 1975
EAA Mean Aitken nuclei count:
• 1.75 x 106 N/cm3
N/Va ratio: 2 x 103 - 5 x 104
Smog chamber samples showed
effects of coagulation on particle
size.
Roadway/Auto ~500 - 1,000:1 EAA, OPC, CNC • Background -Three distinct Wilson, 1977, et al.
Spark ignition, modes with mean size by V of
Catalyst equipped 0.03, 0.24, 6.0 µm
• Test - 0.02, 0.24, 6.0 µm (V)
• Nuclei mode shift due to
coagulation effects - longest
aerosol aging times transferred
about 1/3 nuclei mode V to
accumulation mode.
Roadway/Light Duty 100 - 1,000:1 Mobile lab - EAA DR of 100:1 within 20 m of exhaust Dolan and Kittelson,
Diesel primary pipe at 20 Km/hr in 0.4 s; DR of 1979
instrument 1,000:1 after 1 s
N/V range 7.0x104 - 2.1 x 105
VMD 0.2 µm
Roadway 1,000:1 NA Summary paper: Kittelson and Dolan,
• Roadway dilution < 1 s dilution 1980
ratio > 1,0000:1
• Lab dilution 1 - 5 s dilution
ratio 5 - 20 :1
• Results from roadway and lab
differ due to coagulation,
condensation, adsorption and
DR effects.
Tunnel/Mixed ~ 500 - 1,000:1 Impactor • Principal exhaust components Pierson and
EC and OC Brachaczek, 1983
• MMD 0.15 µm
Tunnel/Mixed ~ 500:1 Hering low After correction for background a Venkataraman, et al.
pressure impactor unimodal distribution was observed 1994
with 85 % of the mass < 0.12 µm.

01/14/99 Page 31
Tunnel/Mixed ~100-500:1 TEOM, LAS-X, Tunnel entrance: Weingartner, et al.,
DMA, CNC, • Nuclei count between 1.4 - 5.1 1997
Epiphaniometer, x 104
Aethalometer, • Nuclei mode peak between 0.02
Photoemission - 0.03 µm
sensor Tunnel exit:
• Accumulation mode peak N and
V between 0.12 - 0.20 µm
• Mass shift between nuclei and
accumulation mode occurring
during 6- 20 min residence
time.
Atmospheric/Mixed ~ > 1,000:1 DMA clone Influence of heterogeneous reactions Ulevicius, et al.,
on particle growth observed: 1994
• Nuclei mode peak 0.02 - 0.04
µm
• Accumulation mode peak 0.05 -
0.3 µm
Atmospheric/Mixed ~ > 1,000:1 10 stage impactor Size distribution and composition Horvath, et al., 1996
analysis:
• Accumulation mode peak 0.5
µm in urban samples
• Coarse mode peat 3.0 µm
• Accumulation mode peak for
black carbon 0.25 µm
• For particles < 0.06 µm
elemental carbon from
combustion is the main
contributor.
Atmospheric/Mixed ~ > 1,000:1 SMPS, CNC. Particle number concentration Tuch, et al., 1997
LAS-X determined primarily by particles <
0.1 µm, N/Ma mean 1.6 x 104

Mass concentration was determined

by particles 0.1 - 0.5 µm, N/M mean
8.4 x 101
Atmospheric/Mixed ~ > 1,000:1 HE-NE laser Observed particle concentrations in Trier, 1997
spectrometer the 65 - 900 nm range were 2 - 16 x
103 part./cm3; did not detect < 50
nm particles

GMD 0.10 - 0.16 µm

Nuclei mode peak 0.084 µm

Accumulation mode peak 0.181 µm
1
In some cases the dilution ratio was not given in the paper and is estimated. In these cases a ~ is shown.
A
N/V = number of particles/volume [units]
B
N/M = number of particles/mass [units]

Kittelson, et al. (1988), conducted a very comprehensive study for the Coordinating Research
Council. Five field studies were conducted: three studies of particles in the exhaust plumes of
diesel highway tractors operating on roadways, a study of particles upwind and downwind of

01/14/99 Page 32
slowly moving diesel trucks, and a study of particles upwind and downwind of a diesel engine
test center. These field studies provide the most comprehensive diesel aerosol information
and are discussed in detail.

The objectives were to examine the physical, chemical and biological characteristics of diesel
aerosol as a function of atmospheric dilution and aging. Atmospheric dilution ratios of the
samples ranged from 2 to 24,000. Atmospheric residence times ranged from about 0.01 to
120 s, however, the effective aging times associated with some of the measurements were
longer due to long residence times in the sampling bags and longer filtration times.

The principal findings of this study are summarized below.

• Typically about 85% of the mass of particles emitted by diesel engines was in the
submicron diameter range. The submicron size distribution was typically bimodal with
"nuclei" and "accumulation" modes at about 0.025 and 0.25 µm volume geometric mean
diameters, respectively.

Figure 11 illustrates an average size distribution based on a series of measurements made over
a roadway in the in the plume of a large heavy-duty highway tractor. For comparison figure
12 illustrates a laboratory size distribution obtained from a Volkswagen rabbit operating at
simulated 50-mph road load conditions. The measurements were made with an EAA and
rapid dilution system providing 1000:1 dilution to prevent coagulation effects. The size
distribution is quite different as the nuclei mode is totally absent. The absence of the nuclei
mode is attributed to the fact that the Rabbit engine produced little volatile material, which
would normally nucleate during the dilution process. The heavy-duty engine on the other
hand produced a substantial amount of HC, which was available for nucleation.

• The volume fraction in the nuclei mode was highly source dependent. It was found to
range from 0 to 35% in these experiments. Consequently, the number-to-volume (or
number-to-mass) ratio was highly variable, ranging from 2,300 to 62,000 part/µm3.
Number concentration was not a good indicator of DPM volume or mass concentration.
[Note: the similarity to the Whitby, et al. (1975) roadway data]

• For most of the cases studied, the volume fraction in the nuclei mode was nearly equal to
the solvent extractable fraction. This suggests that much of the material in the nuclei
mode was extractable. This is illustrated in figure 13.

01/14/99 Page 33
Figure 11. - Average size distribution obtained from chase experiments,
heavy-duty diesel (From Kittelson, et al. 1988)

Figure 12. - Average size distribution obtained from laboratory experiments

using a Volkswagen Rabbit light-duty engine (From Kittelson, et al. 1988)

01/14/99 Page 34
Figure 13. - Volume fraction versus extractable fraction in the nuclei mode
(From Kittelson, et al. 1988)

• Coagulation increases the mean diameter of the nuclei mode and transfers mass from the
nuclei mode into the accumulation mode, thus moving the two modes closer together and
reducing the mass in the nuclei mode. It also increases the mean diameter of the
accumulation mode, but much more slowly. The decrease in the mass and number of
nuclei mode particles takes place on a time scale ranging from minutes for highly polluted
urban conditions to hours for clean rural conditions. Thus, coagulation may significantly
alter the size distributions of diesel particles as they exist in the atmosphere.

• For the highway cruise conditions the dilution ratio increased nearly linearly with time and
typically took only 1-2 s to reach 1,000. Under these dilution conditions little coagulation
occurs. The time scale of dilution for particles emitted from the engine test cell stacks was
much longer than that of roadway dilution. In the worst cases over 100 s were required to
reach a dilution ratio of 1,000. With this relatively slow dilution, coagulation during
dilution was significant and reductions in the number of nuclei mode particles and the
number to volume ratio by as much as a factor of 3 were observed. Although the dilution
conditions of the test cell study may appear to be somewhat non-representative, the time
scale of dilution is not very different from that associated with slow moving congested
urban traffic, where coagulation during dilution will probably also be important.

01/14/99 Page 35
• Light absorption
2
cross sections for diesel particles observed in this study ranged from 5.7
to 6.1 m /g and did not seem2
related to particle size. Light scattering cross sections
ranged from 1.5 to 3.9 m /g and increased with particle size. Thus, coagulation growth
should not have a strong influence on light absorption cross sections but may increase the
light scattering cross sections of diesel particles as they age in the atmosphere.

Table 9 summarizes the properties of the aerosol size distributions measured in this study.
The range of N/VT (number of particles divided by EAA volume concentration) ratio (2,300
to 62,000) is similar to the one reported by Whitby, et al. (1975). Kittelson (1998a)
summarizes the data from the Kittelson, et al.(1988) study and additional data from the
University of Minnesota and others in table 10. It is clear from these tables that N/VT ratio
and thus the number of particles emitted per unit mass of particulate matter are highly
variable. The results suggest that N/VT ratios in the range of 50,000 to 60,000 part./µm3 (or,
assuming spherical, unit density particles, 5 to 6 x 1016 particles per gram of emitted particle
mass) are probably not uncommon for older diesel and SI engines under highway cruise
conditions. This is consistent with the N/VT ratio of 26,000 (Kittelson, et al., 1988) observed
over a rural freeway under mixed traffic conditions. This ratio is lower than source levels
because of dilution with aged ambient aerosol and coagulation. The values of 120,000 to
360,000 observed by Bagley, et al. (1996) are significantly higher than reported for other
engine types. The only way such a high N/VT ratio was produced at the University of
Minnesota was when a diesel engine equipped with a catalyst in an off-design was operated in
the laboratory under high load condition that led to significant sulfate formation. This led to
an N/VT ratio of 400,000.

Table 9. Properties of diesel exhaust size distributions (Kittelson, et al. 1988)

Property Ccp/Ct m/VT N/VT 3 VN/VT Ec DGVn DGVa
g/cm3 part/µm µm µm
Range 0.12 - 0.2 1.2 - 2.2 2,300 - 0 - 0.35 0 - 0.38 0.022 - 0.19 - 0.34
62,000 0.032
Typical 0.15 1.2 NA NA NA 0.025 0.25
value

Ccp/Ct Mass fraction of particles with aerodynamic diameters greater than 0.9 um.
m/VT Submicron filter mass concentration divided by EAA volume concentration.
N/VT Number of particles divided by EAA volume concentration.
VN/VT Volume fraction in the nuclei mode.
Ec Solvent extractable fraction.
DGVn Geometric volume mean diameter of nuclei mode.
DGVa Geometric volume mean diameter of accumulation mode.
NA Highly variable so that defining a typical value was not appropriate.

01/14/99 Page 36
Table 10. N/VT Ratios (Number of particles divided by EAA volume concentration)
Calculated from Various Sources (1 part/µm3 corresponds to 1012 particles/g for spherical
particles with a density of 1 g/cm3) (Kittelson, 1998a)
Condition SOF N/VT Source Reference
3
part/µm
Roadway dilution 0.13 62,000 1979 DI diesel, Kittelson et al., 1988
Cummins
Roadway dilution 0.38 57,000 1979 DI diesel, Kittelson et al., 1988
Detroit Diesel
Atmospheric ~0 3,700 1979 DI diesel, Kittelson et al., 1988
dilution Caterpillar
endurance test stand
Rural freeway 0.24 26,000 Mixed traffic Kittelson et al., 1988
Rural background 0.14 4,300 Background Kittelson et al., 1988
Chassis NA 56,000 SI non-catalyst - In Hildemann et al., 1991
dynamometer - use vehicles N/V calculated from their
lab dilution data.
Chassis NA 46,000 SI catalyst - In use Hildemann et al., 1991
dynamometer - vehicles N/V calculated from their
lab dilution data.
Engine NA 400,000 1995 DI diesel, ox. Work in progress,
dynamometer catalyst in sulfate University of Minnesota
forming mode
Engine NA 25,000 4-stroke DI diesel, Work in progress,
dynamometer iron additive University of Minnesota
Engine NA 2,900 4-stroke DI diesel, Work in progress,
dynamometer baseline University of Minnesota

One of the objectives of the Kittelson, et al., (1988) roadway study was to determine the
relationship between dilution ratio and sampling distance and the impact these variables have
on particles. Figure 14 illustrates dilution of diesel aerosol as it exits the tailpipe and figure 15
summarizes the atmospheric dilution ratios measured during the roadway experiments.

01/14/99 Page 37
Figure 14. - Illustration of factors affecting diesel aerosol dilution at the
tailpipe

Data shown in figure 15 came from two different field studies (Kittelson, et al.,1988). In field
study 2 the sampling distance was determined by towing a mobile laboratory on a flatbed
trailer behind the tractor at speeds of 50 to 55 mi/hr (22-25 m/s). Atmospheric residence
times were very short, ranging from 0.01 to 0.4 s. Dilution ratios ranged from 3 to 90:1. In
field study 5 the mobile laboratory followed the tractor down the roadway at the same speeds.
Dilution ratios of 1,000:1 were obtained in approximately 1 s at a distance of about 25 m.

The relationship between the saturation ratio and dilution ratio and impact on the extractable
fraction is complex. The saturation ratio and extractable fraction peak at intermediate dilution
ratios (approximately 5 to 50:1). The opportunity for homogeneous and heterogeneous
nucleation is greatest when the saturation ratio is highest. However, the amount of time spent
at high saturation ratio is very short for on-highway conditions, typically less than 1 sec. The
importance of the extractable fraction to the nuclei mode was suggested in the discussion of
figures 11 and 12 where size distributions for engines with high and low extractable fractions
are shown. The VN/VT ratio (volume fraction in the nuclei mode) for the engine with high
extractable fraction is 53,000 part/µm3 compared to 5,100 part/µm3 for the engine with low
extractable fraction.

01/14/99 Page 38
Figure 15. - Relationship between dilution ratio and sampling distance. From
Kittelson, et al. 1988)

In most respects the diesel aerosol observed in this study were similar to those observed by
others in dilution tunnel and roadway studies. The findings confirmed the existence of the
nuclei mode reported by Whitby, et al. (1975) although the size of the mode was
unexpectedly large.

01/14/99 Page 39
These results demonstrate the subtle but important changes in the nature of diesel aerosol
occur as the aerosol dilutes and ages in the atmosphere. All the roadway studies, regardless
of engine, exhibited strong nuclei modes. In studies where it was measured, the volume
fraction in the nuclei mode was almost the same as the solvent extractable fraction, suggesting
that much of the extractable material resides in the nuclei mode. Apparently, engines which
produce a nuclei mode emit two distinct populations of particles. The nuclei mode consists of
particles composed of primarily extractable or volatile materials with a number mean diameter
falling between 0.02 - 0.04 µm. These nuclei presumably form during dilution in the
atmosphere. The accumulation mode consists of particles composed of primarily solid
carbonaceous agglomerates having a mass median diameter between 0.1 - 0.25 µm.

EVALUATION OF DIESEL TEST LABORATORY METHODS

The conditions a diesel exhaust sample stream experiences during the measurement process
may greatly affect the particle size, number and mass concentration and chemical composition
of the particulate matter. This section will evaluate current test laboratory procedures for the
sampling and measurement of diesel exhaust aerosols7. Specific examples taken from the lab
survey performed for this study and recent literature are provided below to illustrate how
diesel exhaust aerosols are affected by laboratory parameters.

In summary, the behavior and characteristics of a diesel exhaust aerosol are dependent on:
• sampling conditions of the exhaust
• chemical and physical properties of the exhaust aerosol

Issues to be considered when evaluating a diesel exhaust measurement strategy include:

• Does the competition of the processes of adsorption and nucleation achieve the same
balance as seen in an on-road environment?
• Does the coagulation in the exhaust and sampling lines of the testing environment
adequately match the real situation of the engine exhaust during on-road operation?

In addition to the two main issues of adsorption versus nucleation competition and
coagulation behavior, the following issues that affect these particle processes should also be
considered:
• Is the final dilution ratio of the exhaust above the critical range (about 5 to 50) that favors
nucleation?
• Does the exhaust spend sufficient time (~0.25 s) in the critical range of dilution ratio to
allow the volatile hydrocarbons to adsorb to nuclei and accumulation mode carbonaceous
particles? Is there sufficient time to form or grow nuclei to a detectable size?
• Is the ratio of particle surface area to volatile hydrocarbon and sulfate sufficient to adsorb
enough volatile matter to suppress nucleation ?
• Are there sampling conditions, such as dilution air relative humidity, that may affect the
characteristics of the measured exhaust aerosol?

7 The laboratory survey is not a comprehensive survey, but the results indicate that there are few if any
standard protocols for diesel aerosol measurement.

01/14/99 Page 40
Exhaust Constituents

The nature of the exhaust leaving the tailpipe of a diesel engine is dependent on the load and
speed of the engine, the fuel, lubricating oil composition8 and consumption, and the amount of
carbonaceous solid particles and volatiles (mainly hydrocarbons and sulfuric acid) emitted by
the engine. To evaluate sampling methods and interpret results it is necessary to understand
how the exhaust interacts with the environment from the point of generation to the point of
measurement.

Emission of Carbonaceous Agglomerates and Volatile Material

The current trend in diesel emissions has been a decrease of the mass of emitted DPM. This
reduction was achieved by reducing accumulation mode mass. However, recent studies have
demonstrated that there may be a sharp increase in the number of nuclei mode particles under
certain conditions. When mass concentrations are low and volatile concentrations are high the
number of nanoparticles may increase. These particles may include hydrocarbons (SOF),
carbon fragments, sulfates, and metallic ash. Hydrocarbons are probably the main source of
nuclei mode particle mass for engines running under normal conditions with low sulfur fuels
(< 0.05 %), but even in this case sulfuric acid may play a role in initial nuclei formation.
Possible reasons for the formation of large concentrations of these particles by low emission
engines include nucleation of SOF enhanced by a reduction of total carbonaceous material
mass.

As an engine exhaust dilutes and cools, the sulfuric acid and SOF may pass from the gas phase
to the particle phase by two paths: nucleation to form new particles or adsorption onto or
absorption into existing particles. The exhaust of an engine with high carbonaceous particle
mass emissions may have sufficient particle surface area to adsorb volatiles. This will prevent
the saturation ratios from rising high enough to favor nucleation and most of the volatiles will
be adsorbed by the carbonaceous particles which reside mainly in the accumulation mode
diameter range. An engine with low carbonaceous particle mass emissions may not have
sufficient particle surface area to adsorb volatiles. An exhaust with low carbonaceous material
emissions may also be achieved by the addition of aftertreatment devices, such as particle
traps, that remove solid particles. If an engine with low emission of carbonaceous material
also emits high values of volatile matter, the saturation ratios may rise high enough during
exhaust dilution and cooling for nucleation to be favored. Nucleation of volatile matter may
generate large numbers of nuclei mode particles.

If this interpretation is correct, the formation of nuclei mode particles by future engines
will depend upon the relative amounts of condensable species and solid surface on which
to condense or adsorb.

Bagley, et al. (1996): The new engine technology represented by the engine used in this
study, while markedly reducing particle mass emissions, led to sharp increases in number

8 The engine lab survey indicates that a variety of engine oils are used including synthetic oils.

01/14/99 Page 41
emissions of nuceli mode particles. The 1991 diesel engine tested showed exceptionally high
number concentrations, more than an order of magnitude higher than the 1988 engine and two
orders of magnitude higher than the 1995 engine tested by the University of Minnesota
described below. The engine studied is of very modern design with a very high-pressure (150
MPa peak injection pressure) precisely controlled fuel injection system. It has been suggested
that the very high number concentrations are due to breakup of agglomerates by oxidation
(Hunter, 1997). However, the engine produced very high values of SOF (60-75%) for the
test conditions of the HEI report. The combination of high SOF and relatively low solid
carbon mass may have caused extensive nucleation of the SOF to form a very large nuclei
mode and very large number concentrations. These results are of special concern because the
fuel injection system used on the 1991 engine may be more representative of future heavy-
duty applications than that of the 1995 engine, tested at the University of Minnesota, which
uses a moderately high pressure (90 MPa peak injection pressure) fuel injection system. The
mass of particles in the accumulation mode that could adsorb SOF has been significantly
reduced for the 1991 engine when compared to older engines.

Kittelson (1998a): The results of this study further illustrate the impact of the nature of the
diesel exhaust on the measured particles. The 1995 engine tested in this study has a similarly
low concentration of particles in the accumulation mode to the HEI engine discussed above.
However, only a small nuclei mode was formed. This engine has significantly lower SOF
values (around 30-40 %) than the engine used by Bagley, et al. Nucleation is a highly non-
linear process (Friedlander, 1977) and the concentration of condensable species in the exhaust
of this engine may not be high enough during the dilution process to cause significant
nucleation. Most of the work done at Minnesota used a two-stage dilution system that has a
very short residence time (< 100 ms) in the critical 5 to 50 dilution ratio range. The dilution
system was designed to simulate the rapid dilution that occurs in the atmosphere. This is in
contrast to the dilution system used by Bagley, et al. that had a primary dilution ratio in the
main dilution tunnel of about 10 and a residence time at this dilution ratio greater than 1
second. The test conditions used by Kittelson will tend to minimize nucleation while those
used by Bagley, et al. will favor nucleation.

More recent work at Minnesota (Abdul-Khalek, et al. 1998a) has shown that by using a
residence time in the primary dilution system on the order of a few hundred milliseconds and
varying the dilution ratios the nucleation mechanism could be “turned on or off”. This work
was done using a two-stage dilution system with residence times ranging from 40 - 6,000 ms
and a primary dilution ratio of about 18:1 and a secondary dilution ratio of about 25:1 . The
study measured particle size distributions upstream and downstream of a particle filter both
with and without a catalytic stripper (Abdul-Khalek and Kittelson, 1995). The catalytic
stripper consists of a small oxidizing catalyst mounted in an oven followed by a cooling coil.
The system is used to differentiate between volatile and non-volatile particles in a diluted
sample stream. Major conclusions from this study were:

• Emission of nanoparticles downstream of the filter was dramatically affected by residence

time in the primary dilution tunnel. Changing the residence time from 40 to 6,000 ms

01/14/99 Page 42
 increased number concentrations of nanoparticles by up to four orders of magnitude.
Volume concentration was not strongly influenced by residence time.
• Nearly all particles in the nuclei mode were volatile and < 20 nm diameter.
• The filter performed much better on a volume (or mass) than on a number basis because
nanoparticles formed by nucleation downstream of the filter influence the number
emissions appreciably but usually constitute little volume because of their very small size.

Further work by Kittelson and Abdul-Khalek (1998) examined the influence of primary
dilution temperature and residence time at a fixed dew point of 10 °C on particle size and
concentration. All tests were done with a 1995, medium-duty, direct injection engine running
on 0.03% S diesel fuel at 1600 rpm, 50% load. The concentration of particles in the
nanometer range increased with decreasing temperature and increasing residence time. An
increase in nanoparticle concentration of more than two orders of magnitude was observed as
the dilution temperature decreased from 65 to 32 °C and the residence time was increased
from 100 ms to 1 s. There was little influence on larger, accumulation mode size particles.
This illustrates the extreme sensitivity of nanoparticle concentration to dilution conditions9.

Ahlvik, et al. (1998): The effects of dilution ratio and residence time were compared over a
series of tests using a light-duty chassis dynamometer and ELPI to measure the size
distribution. The dilution ratio varied on average from 2.0 to 5.5 with a residence time of
about 2 s. In this regime it was observed that number concentration increased slightly with
higher dilution. The authors attribute this to coagulation, nucleation and condensation
occurring in the dilution tunnel. However, a strong dependence on dilution ratio was not
observed. This is probably because of the inability of the ELPI to measure particles < 30 nm,
the long residence time, and the low dilution ratios.

Kauffeldt and Schmidt-Ott (1998): The study demonstrated the effects of a particle trap on
a passenger car diesel engine exhaust. Figure 16 is the particle number concentration
measured with and without the trap for an exhaust sample dilution of 1:12. Without the trap,
the particle number count peaks in the accumulation mode size range at about 100 nm. The
same engine operated under the same conditions with a filter shows the number concentration
peak shifting to the nuclei mode size range between 10 and 30 nm. While operating with the
particulate trap, the exhaust was heated and it was found that the nuclei mode number count
decreased with increasing temperature until it disappeared at 175 oC, indicating the nuclei
mode particles are composed of volatile materials.

These findings support the theory that an engine that emits low amounts of carbonaceous
material with sufficient quantities of volatile hydrocarbons will form large numbers of volatile
nuclei mode particles under the appropriate dilution and sampling conditions.

9 The dilution air temperature varies between laboratories, generally lying between 20 and 30 °C as illustrated
by the engine lab survey. It is desirable to use filtered dilution air and to record air temperatures before and
after mixing with the exhaust. Humidity control also varies between laboratories, but it is desirable to record
the humidity after primary and secondary dilution.

01/14/99 Page 43
Figure 16. - Measured size distributions with and without particle filter (v = 120
km/h) at a dilution ratio of 1:12 for a passenger car (from Kauffeldt and
Schmidt-Ott, 1998)

Kruger, Lüders, Luers, Kaufmann, Koch and Kauffeldt (1997) and Lüders (1997): This
study also investigated the influence of exhaust gas aftertreatment and measurement
conditions on particle size characteristics of exhaust from diesel engines. (This research is
described in more detail in Supplemental Report No. 2.)

Number weighted size distributions were obtained using a differential mobility particle
spectrometer (DMPS). Sampling was done from a dilution tunnel followed by secondary
dilution.10 The authors suggest that their results are largely independent of secondary dilution
process, although they said it had some influence. The secondary dilution ratio was usually
about 10:1 and was obtained using an ejector diluter.

The engines tested were 3 l, indirect injection (IDI), naturally aspirated and 1.9 l DI,
turbocharged and intercooled passenger car diesels and a 6.9 l, DI, turbocharged and
intercooled truck engine.

• When a trap was used, the passenger car with the DI engine showed a significant decrease
in particle number for particles > 50 nm, but a significant increase in particle number for
particles < 50 nm.
• For the truck engine the particle trap again showed an increase in the number of particles
< 50 nm, but a decrease in larger particles. The use of an oxidation catalyst also showed
an increase in the in number of particles < 50 nm, but this effect was time dependent.

To determine the nature of the < 50 nm particles the DMPS was modified so the < 50 nm
particles could be analyzed separately. A heating element was integrated into the
measurement system so the aerosol could be heated before passage into the DMPS. With this

10Not reported in the published literature. Based on personal communication with the authors. (Kittelson,
1997)

01/14/99 Page 44
system, volatile particles evaporate and lose their charge, while solid or solid core particles
retain their charge. Volatiles that nucleate upon cooling have no charge and are not detected.
One potential problem is that some particles may pyrolyze in the heater.

This system was used to examine particles emitted from the different control devices. As
temperature was increased the number of 6 - 12 nm particles decreased, when the truck engine
and fiber coil were used at 2400 rpm, full load. A similar observation was made with the
truck engine and oxidation catalyst. As the temperature was increased up to a approximately
175 oC, a decrease in the < 50 nm particles was observed. The authors conclude that the
nuclei mode particles observed after the particle filter or oxidation catalyst consist entirely of
volatile material that evaporates at 180 oC. The authors suggest that the material primarily
consists of sulfuric acid droplets. The condensation effect is strongly dependent on the
temperature and the partial pressures of the single components, which are affected by the
dilution of the exhaust gas.

The effect of the dilution ratio was further characterized. For the truck engine running at
rated power without a control device while the dilution ratio was varied from 12 to 32, no
effect on the size distribution was observed. However, when the particle filter was installed
an increase in the number of < 50 nm was observed at a 12:1 dilution ratio. However, when
the ratio was increased to 19 and 32:1 the increase disappeared, as illustrated in figure 17.
The dependence of nanoparticle formation on dilution conditions is consistent with findings
from Abdul-Khalek, et al., 1998a.

The < 50 nm particles appear after the particle trap below a critical dilution ratio. The trap
drastically reduces the mass and the number of particles available as condensation adsorption
sites. This leads to the process of homogeneous nucleation of the exhaust gases and the
creation of a large number of new particles not previously observed in the exhaust stream.
The oxidation catalyst does not reduce the quantity of particulate material but can lead to an
increase in the concentration of sulfuric acid in the exhaust gas, which leads to the same effect
because of the change in the ratio of adsorption sites to condensing matter.

Compared to tunnel dilution, ambient dilution ratios may be much higher, therefore the
authors question whether the < 50 nm particle increase would be observed under ambient
driving conditions. However, it should be noted that dilution conditions found in a city are
probably quite different from those encountered on-highway (at high speed) in that
combustion particle concentrations are higher and are generated under different driving
conditions. A number of measurements on and near roadways show a nuclei mode is present
under real atmospheric conditions (Kittelson, et al., 1988, Weingartner, et al., 1997a,
Harrison, et al., 1996).

01/14/99 Page 45
Figure 17. - Effect of dilution ratio on the particle size distribution from a
heavy-duty diesel engine with a particle filter. (From Luders, 1997)

Ash Constituents in Exhaust

Recent studies suggest that ash constituents in diesel exhaust may contribute to the nuclei
mode. Ash constituents may come from contamination from lubricating oil or additives to the
fuel and typically may account for more than 10% of exhaust particle mass (Cadle, et al.,
1997; Williams, et al., 1987, 1989a and 1989b; Abbass, et al. 1987; Andrews, et al. 1993)

Abdul-Khalek, et al. (1998a) studied a 1995 DI diesel engine that exhibited both a nuclei and
an accumulation mode in the number distribution of exhaust particles. They used a catalytic
stripper to determine concentrations of solid and volatile particles in the nanometer size range.
Although the results were somewhat uncertain they suggested that although most particles in
this size range were volatile, solid particle were also present (this was later confirmed by
Abdul-Khalek, et al., 1998b). The authors attributed the formation of the nuclei mode to the
relative reduction of solid carbonaceous particle emissions compared to volatiles and the
presence of ash constituents such as calcium, zinc, and magnesium. The ash components
contribute to the nuclei mode either directly as nanoparticles themselves or by acting as
heterogeneous nuclei for condensation of volatile hydrocarbons. They performed chemical
equilibrium calculations and demonstrated that calcium compounds are gaseous at diesel
combustion temperatures but should form solid compounds within the engine as the products
expand and cool during the exhaust stroke.

01/14/99 Page 46
With low carbon emission engines, the solid carbonaceous particles will not provide enough
surface area to adsorb the ash constituents and relieve the metal supersaturation. The metals
will nucleate to form solid nanoparticles that may serve as sites for heterogeneous nucleation
of hydrocarbons and sulfuric acid.

Mayer, et al. (1998) studied metals as fuel additives to reduce particulate matter emissions
from diesel engines. The metallic fuel additives included iron, copper, and cerium and were
used as catalysts to oxidize carbon during combustion and on a passive particulate trap.
When added without a particulate trap, the metallic additives reduced the total particulate
matter mass emissions by about 10-15%. When used with a particle trap, the additives
reduced total mass emissions of particulate matter by up to 90%.

However, the investigators also found that the engine out number emissions of nanoparticles
increased, depending on the amount of metals added. Figure 18 shows the effects of metallic
fuel additives on the number of nuclei mode particles emitted from a diesel engine. There is a
dramatic increase in nanoparticle emissions with the addition of the metallic fuel additives.
Because of this effect, the authors concluded that metallic fuel additives for soot emission
reductions should not be used without a particulate trap that has the correct filtration
characteristics to remove the fine particles.

Figure 18. - Nanoparticle emissions from a diesel engine using iron (satacen)
as a fuel additive with and without a particle trap. (Mayer, 1998)

01/14/99 Page 47
Exhaust Sampling Conditions

Relative Humidity and Temperature of Dilution Air

Recent studies have investigated the impact of the relative humidity and temperature of the air
used to dilute diesel exhaust during sampling for particle measurement or in the environment
for on-highway studies.

Weingartner, Burtscher, and Baltensperger (1997b) performed a fundamental study on the

hygroscopic properties of diesel particles. The study included generating diesel particles
under different engine conditions, conditioning the particles to change their hygroscopicity,
and then varying the relative humidity and measuring the effects on particle size and number.
Results included:
• When particles were dried, exposed to higher humidity and then re-dried, they shrank with
subsequent exposure to higher humidity levels.
• When particles were not dried, they grew with increasing relative humidity.
• When particles were not dried and the fuel sulfur concentration was increased, more
growth of the exhaust particles with increasing relative humidity was seen.
• Particles generated under idle conditions shrank with increasing relative humidity even
with high sulfur fuel. Investigators attributed this to increased organic fraction of particles
generated at idle acting as a barrier to water.
• Particles generated under load that were treated with ozone and ultraviolet light formed a
bimodal number distribution with the new peak in the nuclei mode at about 30 to 40 nm.
• When a filter was used to remove the diesel particles and the remaining gas phases were
treated with ozone and ultraviolet light, a nuclei mode of particles appeared with the peak
also at about 30 to 40 nm. The investigators believed gaseous constituents were
nucleating to form nuclei particles.
• When the exhaust was not treated and no filter was used, the exhaust had a single peak at
about 90 nm in the accumulation mode size range.

These results demonstrate that particle composition including organic fraction and sulfur,
particle hygroscopic properties, and the relative humidity of the aerosol in the environment or
laboratory dilution conditions may change the measured exhaust aerosol properties. Testing
conditions for this study were not representative of dilution conditions of diesel exhaust in an
on-highway setting. However, changes in particle number and size measured with changing
relative humidity and particle hygroscopic properties indicate humidity in the environment or
dilution air in a laboratory setting could play a role in affecting diesel exhaust particle
formation and transformation.

Shi and Harrison (1997) also reported finding effects of relative humidity of dilution air on
particle number concentration and size. Increasing dilution air humidity increased the peak
diameter of the particle number concentration and total particle number and volume
concentrations. (These results are currently unpublished and were presented at a nanoparticle
measurement workshop. No further details were provided.)

01/14/99 Page 48
Kreso, et al. (1998a,b) looked at the effect of relative humidity of the dilution air on the
nucleation of condensable sulfur species in two low sulfur (0.031% wt. sulfur) fuel diesel
engines. The relative humidity in the sampling dilution tunnel of these tests was varied from
10 to 30%. No significant effect was observed on the particle number and volume
concentrations with the varying humidity levels. The authors concluded that the nuclei mode
particles from the low sulfur diesel engines tested did not consist primarily of condensable
sulfur species (H2O-H2SO4 particles) that would respond to changes in relative humidity.

Although there is evidence that the relative humidity of a diesel exhaust aerosol may affect
the particle characteristics, these effects have not been conclusively demonstrated.
Further work to determine the impact should be performed.

Dilution Ratio and Residence Time

Other issues should be considered if nanoparticle emissions are a concern. If these particles
are primarily composed of volatile materials they are formed as the exhaust dilutes and cools.
For most of the engines and fuels tested at the University of Minnesota, nuclei mode particles
appear to be mainly volatile. The only exception to this was observed when the use of
metallic fuel additives caused the formation of a large, solid nuclei mode. (Du, et al. 1998;
Kittelson, et al., 1978) The number formed depends primarily upon the dilution ratio and
residence time. Atmospheric dilution should be simulated to obtain measurements
representative of real-world conditions. Furthermore, volatile particles may have different
health impacts than solid carbonaceous particles.

Formation of nanoparticles during the dilution process has been seen to be a highly non-linear
function of dilution ratio, temperature and residence time. From figure 9 it can be seen that
the time spent in the critical dilution ratio range of 5 - 50:1 is about 0.25 s.11 This is near the
lower end of residence times used in laboratory dilution systems, so that one might expect that
laboratory conditions might often tend to overestimate nanoparticle formation. However,
many measurements made in the atmosphere over roadways show large concentrations of
nanoparticles. This suggests that roadway dilution conditions may lead to nanoparticle
formation.

Concentrations of nanoparticles over roadways are high enough that coagulation is likely to
cause nanoparticles to grow quickly into the accumulation mode size range (Harrison, et al.,
1996). The character of these agglomerates will depend upon whether the nanoparticles are
solid or volatile. Because of coagulation, the highest number concentrations and exposures to
nanoparticles will occur on or near roadways where particles are fresh.

11From the engine lab survey it can be seen that the residence time from the tailpipe to the primary dilution
tunnel varies considerably based upon the engine and duty cycle. Times range from 0.015 - 4 s with longer
rather than shorter times being more common. Reported primary dilution ratios are within the critical range
of 5 - 50:1. Long residence times within the critical dilution range favor the formation of nanoparticles.

01/14/99 Page 49
Current emission standards are mass based. Particulate matter mass tends to be conserved
during the dilution and sampling process. Some gas-to-particle mass transfer occurs during
dilution so that the mass of SOF collected depends upon dilution ratio (Kittelson and Johnson,
1991). However, for the range of dilution ratios typically used, this effect does not change the
total collected mass significantly. Temperature also changes and this affects adsorption. On
the other hand, even a small amount of mass transfer by homogeneous nucleation can have an
enormous effect on the number concentration and number-weighted size distribution.
Furthermore, coagulation, which conserves mass, may significantly change the number
concentration and size distribution. Homogeneous nucleation and coagulation are highly
nonlinear and sensitive to small changes in the dilution process. Thus, any changes in, or
additions to, current mass based emission standards to include size or number concentration
measurements will lead to significant sampling and measurement challenges.

Conclusions

It can be seen from the preceding discussion that test laboratory procedures including the
conditions of the dilution air and the amount and rate of dilution can greatly affect the nature
of the sampled exhaust. In addition, the availability of particle surface area for adsorption and
the quantity of volatile material in the emitted exhaust may also change the number and size of
the sampled particle emissions. In summary:
• dilution ratios in the critical range of 5 to 50 may promote the formation of a nanoparticle
mode as particles nucleate as the exhaust is cooled
• the role of relative humidity in dilution air is not fully understood
• the lack of carbonaceous material surface area in modern diesel engines for adsorption of
volatile compounds may allow saturation ratios to become high enough to favor
nucleation and the creation of a large nanoparticle number mode
• non-carbonaceous material in the exhaust such as metals or sulfuric acid may act as
nucleation sites for heterogeneous nucleation of volatiles, even in the presence of
carbonaceous material
• temperature residence time in the critical dilution ratio range can markedly affect the
production of nanoparticles.

RECOMMENDATIONS

Understanding diesel aerosol emissions and behavior is critical if meaningful regulations are to
be set. General recommendations for laboratory diesel aerosol measurement are made below.
These recommendations are intended to promote measurement repeatability between testing
facilities and to assist laboratories in defining critical measurement system parameters. Future
research may be required to implement these recommendations effectively. Each of these
recommendations is discussed in more detail elsewhere in the Supplemental Reports.

• Treatment of emissions sample

- carefully temperature and residence time in critical dilution ratio range (5 – 50)
- minimize residence time after final dilution
- engine to tunnel lines should be similar in temperature to exhaust pipe

01/14/99 Page 50
• Treatment of sample dilution air
- filter to remove ambient particulate matter
- control temperature
- control humidity
• Emission sampling lines
- minimize heat transfer in sampling lines
- use conducting lines to minimize electrostatic losses
- minimize sampling line length

• Aerosol measurement instruments

- avoid using instruments outside reliable measurement range (i.e. below a certain particle
size or above a particle concentration)
- be aware that instruments may have different capabilities when operated in different
modes
- errors associated with incorrect dilution and sampling may be much larger than those
associated with misapplication of an instrument

• Meticulous reporting of testing conditions

- fuel properties including sulfur content
- lubricating oil properties including metal content
- history of engine and sampling system conditioning
- dilution air treatment
- dilution ratios (including any final dilution for instrument concentration requirements)
- sampling line residence time at each dilution ratio
- order of testing at different operation modes

Specific guidelines on dilution ratios, dilution rates, residence times, temperatures, humidity,
aerosol measurement instrumentation and other details related to sampling protocol are more
difficult to provide at this time. Research is ongoing to clearly define the impact these factors
have on the diesel aerosol size distribution. However, before specific sampling protocols are
recommended sampling priorities must be established. Sampling protocols which produce
aerosol size distributions similar to those found in the ambient atmosphere may not be the
same as those needed to simply compare emissions between engines or to establish mass
emission rates. Sampling should allow characterization of the parameters of greatest
importance. These parameters include particulate matter mass, number, surface area,
chemical composition of particles and exhaust gases. Unfortunately, their relative importance
is not known. Sampling to determine one parameter may bias result for another. Clearly,
sampling priorities must be established and it is likely that as a result, multiple test protocols
will be required.

SPARK IGNITION (SI) MEASUREMENTS

Uncertainty exists regarding the relative contribution of atmospheric aerosol from SI and
diesel exhaust in real-world conditions. Diesels emit more particulate matter on a fleet
average, gram-per-vehicle mile mass basis, but because SI vehicles, at least in the U.S.,

01/14/99 Page 51
account for most of the vehicles operating on-road, the direct PM emissions from SI engines
may be more important. The Northern Front Range Air Quality Study (NFRAWQS) recently
addressed this issue.

In addition, several recent studies have determined the number size distribution from SI
exhaust. These studies are briefly described primarily to provide a comparison between
gasoline and diesel fueled engines.

Cadle, et al., (1998) and Watson et al., (1998): The goals of the NRRAWQS study were to:
• Attribute the existing air pollution in the Denver urban region to a source or source
category;
• Collect data necessary to support informed decisions leading to attainment of Federal
air quality standards.
Specific objectives included:
• Apportion carbonaceous material in airborne particles to sources;
• Determine the extent to which the Denver area is ammonia-rich (or anion-deficient)
with regard to the formation of secondary ammonium nitrate and ammonium sulfate;
• Apportion sources leading to the formation of non-carbonaceous portion of PM2.5 with
emphasis on nitrate and sulfate.

The NRRAWQS found that vehicle exhaust was the largest PM2.5 total carbon (TC)
contributor, constituting about 85 % of PM2.5 carbon at sites in the Denver metropolitan area.
Sources with emissions similar to light duty gasoline vehicles contributed about 60 % of PM2.5
TC at urban Denver sites, and these contributions were 2.5 -3 times the diesel exhaust
contributions. Emissions were particularly high during cold, cold starts and from vehicles
identified as “smokers”. However, diesels contributed the largest share of the elemental
carbon at these locations.

Maricq, Podsiadlik and Chase, (1998): This study, conducted at the Ford laboratory,
compared emissions from a variety of recent model gasoline engines and one European diesel
engine. The SMPS and ELPI were used to obtain size distribution data. Distributions were
built from repeated cycles using the ELPI or SMPS and CNC over a fixed size range. This
was necessitated due to the relatively slow response time of the SMPS when used in the
scanning mode. Size distributions derived from the two methods were consistent.
Exhaust flow rates ranged from 0.1 m3/min at idle to 2.5 m3/min during heavy acceleration
and transfer residence times ranged from 7 s to 0.4 s. The tunnel operated at constant total
flow and the instantaneous dilution ratio varied from 120:1 at idle to 2.5:1 at heavy
acceleration. For about 25 s during the FTP the dilution ratio fell below 5. The residence
time in the dilution tunnel was about 3 s. The exhaust of the diesel engine was further diluted
by a factor of 9 using an ejection pump.

Measurements were reported over the 3 phases of the FTP cycle. Mass emission rates were
low for all engines (<2 mg/mi particulate matter). The number weighted mean diameter
ranged from 35 - 65 nm with the range for mass mean diameter shifting to 100 - 300 nm. For
the diesel and direct injection spark ignition (DISI) engines the number mean diameter was

01/14/99 Page 52
somewhat larger, 60 - 70 nm. The mass and number of particles are much lower for the spark
ignition engines compared to the diesel and DISI engines. DISI emissions are about 20
mg/mi. and 1014 particles per phase compared to 1 mg/mi and 1012 particles per phase for
conventional gasoline engines. The diesel engine emissions were higher, about 80 mg/mi and
1015 particles per phase.

Greenwood, Coxon, Biddulph and Bennett (1996): Emissions from two diesel (IDI no
catalyst and turbocharged with catalyst), three gasoline and two compressed natural gas
(CNG) vehicles were evaluated using an SMPS to determine aerosol size distributions with
the focus on the number distribution. Exhaust was diluted about 20:1 in a 0.20 m diameter
tunnel with a total flow rate of 9.2 m3/min. Engines were operated on a chassis dynamometer
at constant speeds representative of the European emissions drive cycle. Speeds ranged from
idling to 120 km/h with loads ranging from 0.0 to 16.0 kW.

The aerosol size distributions from all three engine types appeared to have a lognormal form;
thus the peak maximum number concentration was lower in diameter than the mean diameter.
The measured peak number concentration for the diesel vehicles under load ranged from 70 -
90 nm for the IDI engine and 50 - 60 nm for the turbocharged engine. At idle these were 32
and 46 nm, respectively. The gasoline fueled vehicles produced lower particle concentrations
at low and medium power. At higher loads a range of peak number concentration was
observed between 20 and 30 nm. At the highest load condition the total number
concentration for the IDI engine was 2.4x107 and for the turbo it was 5.32x106. The three
gasoline fueled vehicles had total number concentrations of 2.58x104, 4.23x106 and 1.61x107
at the high load condition. Much lower number concentrations were produced at the medium
and low power test conditions.

FUTURE RESEARCH

Research is recommended in the following areas:

• The diesel aerosol size distribution (mass, volume, number and surface area) must be
determined under real-world (roadway dilution) conditions using a variety of heavy and
light-duty engines operated under varying load and speed conditions in the city and on the
highway using state-of-the-art instrumentation.
• Diesel aerosol generated under conditions representative of real-world dilution must be
chemically characterized based on particle size.
• A variety of fuels must be evaluated, including low sulfur and zero sulfur fuel.
• The sensitivity of particle generation in diluting exhaust plumes to dilution ratio, dilution
rate, temperature, and humidity should be evaluated.
• Particle transformations that occur as the plume disperses downwind of the roadway
under typical urban and highway situations must be measured.
• The species and mechanisms involved in nucleation during dilution should be identified.

Appropriate engine and chassis dynamometer laboratory techniques that result in diesel
aerosol measurements similar to ones obtained under real-world dilution conditions need to be

01/14/99 Page 53
developed and evaluated so that future regulatory decisions are based upon data that are not
biased by artifacts created in the laboratory.

The fundamental goal should be to develop laboratory methods to measure diesel aerosol size
distributions and chemical characteristics in the laboratory that result in data that are similar to
those measured under roadway conditions. These methods must yield results, which are
repeatable within labs and enable inter-laboratory comparison. This entails a fundamental
understanding of aerosol formation, transformation and measurement, and the processes that
have a major effect on the aerosol size distribution such as dilution, nucleation, condensation
adsorption and coagulation.

ACKNOWLEDGEMENTS

This project was funded by grant number X825841-010 from the U.S. Environmental
Protection Agency. Mr. William Charmley from the National Vehicle and Fuels Emission
Laboratory in Ann Arbor, MI acted as the technical project officer and provided helpful
advice throughout the course of the project. Professor John H. Johnson from Michigan
Technological University, Houghton, MI served as a technical consultant and reviewer. His
experience and technical advice was of great assistance to the investigators.

BIBLIOGRAPHY12

Abbass, M. K., et al. 1987. The Aging of Lubricating Oil, the Influence of Unburnt Fuel and
Particulate SOF Contamination. SAE Paper No. 872085.
Abdul-Khalek, I.S. and D.B. Kittelson, 1995, "Real Time Measurement of Volatile and Solid
Exhaust Particles Using a Mini-Catalyst," SAE Paper No. 950236.
Abdul-Khalek, I. S., et al. 1998a, Diesel Exhaust Particle Size: Measurement Issues and
Trends, SAE Paper No. 98P-353.
Abdul-Khalek, I. S., et al. 1998b. Diesel Trap Performance: Particle Size Measurements and
Trends. SAE Paper No. 982599.
Abdul-Khalek, I. S. 1999. Influence of Dilution Conditions on Nanoparticulate Emissions
from a Diesel Engine: Experimental and Theoretical Investigation. Ph.D. Dissertation, (in
preparation).
Ahlvik, P., et al. 1998. Real Time Measurements of Diesel Particle Size Distribution with an
Electrical Low Pressure Impactor. SAE Paper No. 980410.
Ambs, J. L., and T. L. Hillman. 1992. Disposable and Reusable Diesel Exhaust Filters. In
BuMines IC 9324, pp. 67-73.
Andrews, G. E., et al. 1993. The Influence of Lubricating Oil Age on Emissions from and
IDE Diesel. SAE Paper No. 931003.
Andrews, G. 1998. Private communication with D. B. Kittelson, February, 1998.

12 Combined bibliography from Supplemental Reports 1 and 2

01/14/99 Page 54
Bagley, S. T., et al. 1996. Characterization of Fuel and Aftertreatment Device Effects on
Diesel Emissions. Health Effects Institute Research Report No. 76, 88 pp.
Baron, P. A. 1986. Calibration and Use of the Aerodynamic Particle Sizer (APS 3300).
Aerosol Sci. Technol. 5(1): 55-67.
Baron, P. A., M. K. Mazumder and Y. S. Cheng. 1993. Chapter 17 Direct Reading
Techniques Using Optical Particle Detection. In Aerosol Measurement Principles Techniques
and Applications, K. Willeke and P. Baron editors, Van Nostrand Reinhold, NY, NY., p.
381-409.
Baron, P. and K. Willeke. 1993. Chapter 2 Aerosol Fundamentals. In Aerosol Measurement
Principles Techniques and Applications, K. Willeke and P. Baron editors, Van Nostrand
Reinhold, NY, NY., p. 8-22.
Barris, M. A. 1981. Mass Transfer Processes Affecting the Nature of Diesel Exhaust
Aerosols. University of Minnesota M. S. Thesis, Minneapolis, MN.
Baumgard, K. J. and D. B. Kittelson. 1985. The Influence of a Ceramic Particle Trap on the
Size Distribution of Diesel Particles. SAE Paper No. 850009.
Baumgard, K. J. and J. H. Johnson. 1992. The Effect of Low Sulfur Fuel and a Ceramic
Particle Filter on Diesel Exhaust Particle Size Distributions. SAE Paper No. 920566.
Baumgard, K. 1995. The effect of Fuel and Engine Design on Diesel Exhaust Particle Size
Distributions, Ph.D Thesis, Michigan Technological University.
Baumgard, K. and J. H. Johnson. 1996. The Effect of Fuel and Engine Design on Diesel
Exhaust Particle Size Distributions. SAE Paper No. 960131.
Boulaud, D. and M. Diouri. 1988. A New Inertial and Diffusional Device (SDI 2000). J. of
Aerosol Sci. Vol. 19(7), pp. 927-930.
Bureau of Mines. 1992. Diesels in Underground Mines: Measurement and Control of
Particulate Emissions. Proceedings of the Bureau of Mines Information and Technology
Transfer Seminar, Minneapolis, MN, September 29-30, 1992, BuMines IC 9324, 132 pp.
Burtscher, H. and H. C. Siegmann. 1993. Photoemission for In Situ Analysis of Particulate
Combustion Emissions. Water, Air, and Soil Pollution, Vol. 68: 125-136.
Burtscher, H., D., et al. 1993. Characterization of Soot Particles by In-Situ Measurement
with Different Aerosol Analysis Tools. J. of Aerosol Sci. Vol. 23, pp. 65-67.
Cadle, S. H., et al. 1997. Particulate Emission Rates from In-use High-emitting Vehicles
Recruited in Orange County, California. Environ. Sci. Technol. Vol. 31, pp. 3405-3412.
Cadle, S. H. et al. 1998. Measurement of Exhaust Particulate Matter Emissions from In-Use
Light-Duty Motor Vehicles in the Denver, Colorado Area. Final Project Report CRC Project
E-24-1, 142 pp. (Available from the Coordinating Research Council, 219 Perimeter Center
Parkway, Suite 400, Atlanta, GA 30346-1301 and at http://charon.cira.colostate.edu)
Cantrell, B. K. and K. L. Rubow. 1992. Measurement of Diesel Exhaust Aerosol in
Underground Coal Mines. Bureau of Mines IC 9324, p. 11-17.

01/14/99 Page 55
Chen, D. R., et al. 1996. Nanometer Differential Mobility Analyzer (Nano-DMA): Design and
Numerical Modeling. J. Aerosol Sci. Vol. 27 Supplement 1 pp. S137-S138.
Chen, D. R. and D. Y. H Pui. 1997. A Novel Unipolar Aerosol Charger. Annual meeting of
the AAAR, Denver, CO, October 13-17.
Chen, D. R., et al. 1998. Design and Evaluation of a Nanometer Aerosol Differential Mobility
Analyzer (Nano-DMA). J. Aerosol Sci. Vol. 29, No. 5/6, pp. 497-509.
Cheng, Y. S. and C. H. Yeh. 1984. Analysis of Screen Diffusion Battery Data. Am. Ind. Hyg.
Assoc. J., 45(8):556-561.
Cheng, Y. S., et al. 1984. Characterization of Diesel Exhaust in a Chronic Inhalation Study.
Am. Ind. Hyg. Assoc. J., 45:547-555.
Cheng, Y. S., H. C. Yeh, and K. J. Brinsko. 1985. Use of Wire Screens as a Fan Model Filter.
Aerosol Sci. and Tech., 4:165-174.
Cheng, Y. S. 1993. Condensation Detection and Diffusion Size Separation Techniques. In
Aerosol Measurement, K. Willeke and P.A. Baron, eds. Van Nostrand Rheinhold: New York,
NY, pp. 427-451.
Clerc, J. C., and J. H. Johnson. 1982. A Computer Heat Transfer and Hydrocarbon
Adsorption Model for Predicting Diesel Particulate Emissions in Dilution Tunnels. SAE
Paper No. 821218.
Cohen, B. S. and S.V. Hering (eds.). 1995. Air Sampling Instruments for Evaluation of
Atmospheric Contaminants. American Conference of Governmental Hygienists, Inc., 651 pp.
CONCAWE, 1998. A Study of the Number, Size & Mass of Exhaust Particles Emitted from
European Diesel and Gasoline Vehicles Under Steady-state and European Driving Cycle
Conditions. Prepared for CONCAWE Automotive Emissions Management Group by Special
Task Force AE/STF-10: D. E. Hall Chairman, Brussels, Belgium, 56 pp.
Degobert, P. 1988. Emissions Diesel, Aspects Physico-Chimiques. 5 ème Symp. Sur la
Recherche en Matière de Pollution Atmosphériqe, Strasbourg, March 22-25, 13 p, (in
French).
DeMarcus, W. and J. W. Thomas. 1952. Theory of a Diffusion Battery. Report No. ORNL-
1413, Oak Ridge National Laboratories: Oak Ridge, TN.
Dickens, C. J., et al. 1997. Evaluation of Instruments for Vehicle Emission Particle Sizing.
AEA Technology Report AEAT-1180, 69 pp. (restricted distribution).
Dolan, D.F., D. B. Kittelson, and K. T. Whitby. 1975. Measurement of Diesel Exhaust
Particle Size Distributions. Paper No. 75-WA/APC-5, American Society of Mechanical
Engineers: New York, NY.
Dolan, D. F. and D. B. Kittelson. 1979. Roadway Measurements on Diesel Exhaust
Aerosols, SAE Paper No. 790492, 9 pp.
Dolan, D. F., et al. 1980. Diesel Exhaust Particle Size Distribution Measurement Techniques.
SAE Paper No. 800187, 16 pp.

01/14/99 Page 56
Doyle, G.J. 1961. Self-Nucleation in the Sulfuric Acid-Water System. J. Chem. Phys., 35,
795-799.
Du, C. J., J. Kracklauer, and D. B. Kittelson, 1998. Influence of an Iron Fuel Additive on
Diesel Combustion. SAE Paper No. 980536.
Fang, C. P., and D. B. Kittelson. 1984. The Influence of a Fibrous Diesel Particulate Trap on
the Size Distribution of Emitted Particles. SAE Paper No. 840362.
Friedlander, S. K. 1977. Smoke, Dust and haze: Fundamentals of Aerosol Behavior. John
Wiley and Sons, NY, NY.
Fuchs, N. A. 1964. The Mechanics of Aerosols. Pergamon Press, 408 p.
Gaggeler, H. W., et al. 1989a. The Epiphaniometer, A New Device for Continuous Aerosol
Monitoring. J. of Aerosol Sci., Vol. 20, No. 5, pp. 557-564.
Gaggeler, H. W., et al. 1989b. Applications of the Epiphaniometer to Environmental Aerosol
Studies. J. of Aerosol Sci., Vol. 20, No. 8, pp. 1225-1228.
Gard, E., et al. 1997. Real-time Analysis of Individual Atmospheric Aerosol Particles:
Design and Performance of a Portable ATOFMS. Anal. Chem. Vol. 69, No. 20, pp. 4083-
4091.
Gillies, A. D. S. and M. A. Schimmelpfennig. 1983. Atmospheric Fogging in Underground
Mine Airways. Mining Engineering 35(4):336-342.
Greenwood, S. J., et al. 1996. An Investigation to Determine the Exhaust Particulate Size
Distributions for Diesel, Petrol, and Compressed Natural Gas Fuelled Vehicles. SAE Paper
No. 961085.
Groblicki, P. J. and C. R. Begeman. 1979. Particle Size Variation in Diesel Car Exhaust. SAE
Paper No. 790421.
Hamill, P. 1976. The Nucleation of H2SO4-H2O Solution Aerosol Particles in the
Stratosphere. J. Atmospheric. Sci., Vol. 34, pp. 150-162.
Hammerle, R. H., et al. 1994. Emissions from Current Diesel Vehicles. SAE Paper No.
942043.
Hammerle, R. H., et al. 1995. Emissions from Current Diesel Vehicles with and without Lean
NOx and Oxidation Catalysts and Particulate Traps. SAE Paper No. 952391.
Harrison, R. M., et al. 1996. Airborne Particulate Matter in the United Kingdom. Third
Report of the Quality of Urban Air Review Group, University of Birmingham, Birmingham,
UK. 176 pp.
Hart, K. M., et al. 1993. The Correlation Between In-Situ, Real-time Aerosol Photoemission
Intensity and Particulate Polycyclic Aromatic Hydrocarbon Concentration in Combustion
Aerosols. Water, Air, and Soil Pollution, Vol. 68, pp. 75-90.
Havlicek, M. 1998. Personal Communication with M. Arnold, June, 1998. TSI, Inc.

01/14/99 Page 57
Health Effects Institute. 1995. A Critical Analysis of Emissions, Exposure, and Health
Effects. A Special Report of the Institute’s Diesel Working Group. Health Effects Institute,
April, 1995, Cambridge, MA 294 p.
Health Effects Institute. 1997. Formation and Characterization of Particles: Report of the
1996 HEI Workshop. Health Effects Institute, Cambridge, MA.
Heitbrink, W. A., P. A. Baron, and K. Willeke. 1991. Coincidence in Time-of-flight
Spectrometers: Phantom Particle Creation. Aerosol Sci. Technol. 14:112-126.
Hering, S. V. Impactors, Cyclones, and Other Inertial and Gravitational Collectors. 1995. Ch.
14 in Air Sampling Instruments for Evaluation of Atmospheric Contaminants. Ed. by B. S.
Cohen, and S.V. Hering. ACGIH, Cincinnati, OH, 8th Edition, pp. 279 – 305.
Hinds, W. C. 1982. Aerosol Technology. New York, John Wiley and Sons.
Hoppel, W. A. 1978. Determination of the Aerosol Size Distribution from the Mobility
Distribution of the Charged Fraction of Aerosols. J. Aerosol Sci. 9:41-54.
Horvath, H., et al. 1988. Diesel Emissions in Vienna. Atmospheric Environ. 22(7):1255-
1269.
Horvath, H., et al. 1996. The Size Distribution and Composition of the Atmospheric Aerosol
at a Rural and Nearby Urban Location. J. Aerosol Sci. 27(3):41.
Hueglin, Ch., et al. 1997. An Accurate, Continuously Adjustable Dilution System (1:10 to
1:104) for Submicron Aerosols. J. Aerosol Sci. 28(6):1049-1055
Hummes, D., et al. 1996. Nanometer Differential Mobility Analyzer (Nano-DMA):
Experimental Evaluation and Performance Verification. J. Aerosol Sci. Vol. 27 Supplement 1
pp. S136-S137, 1996.
Hunter, G. 1997. Formation and characterization of particles: Report of the 1996 HEI
workshop, Particle formation in combustion, HEI Communications 5, 6-7. [also see
Presentation at the Formation and Characterization of Particles Health Effects Institute
Workshop, December 3-4, 1996, Cambridge, Massachusetts. Summarized in the Workshop
Summary available from the Health Effects Institute.]
International Agency Research on Cancer. 1989. Diesel and Gasoline Engine Exhausts and
Some Nitroarenes. IARC Monograph V. 46, 458 pp.
Johnson, J. E. 1993. Hydrocarbon Oxidation in a Diesel Catalytic Converter. University of
Minnesota M. S. Thesis, Minneapolis, MN.
Johnson, J. H., et al. 1994. A Review of Diesel particulate Control Technology and Emission
Effects. SAE Paper No. 940233.
Kapadia, A. 1980. Data Reduction Techniques for Aerosol Size Distribution Measurement
Instruments. Ph. D. Dissertation, University of Minnesota.
Kauffeldt, Th. and A. Schmidt-Ott. 1998. Investigation of the Influence of Exhaust
Aftertreatment on the Particulate Phase in Diesel Exhaust Gases. J. Aerosol Sci. 29:Suppl.
1:S625-S626.

01/14/99 Page 58
Keady, P.B., F. R. Quant, and G. J. Sem. 1983. TSI Quarterly (9) 3.
Keskinen, J., K. Pietarinen, and M. Lehtimaki. 1992. Electrical Low Pressure Impactor. J.
Aerosol Sci. 23(4):353-360.
Kittelson, D. B., et al. 1978. Diesel Exhaust Particle Size Distributions: Fuel and Additive
Effects, SAE Paper No. 780787.
Kittelson, D. B. and Dolan, D. F. 1980. Diesel Exhaust Aerosols. In Generation of Aerosols
and Facilities for Exposure Experiments, Ed. K. Willeke, Ann Arbor Science, Ann Arbor, MI,
p 337-359.
Kittelson, D. B., K. C. Moon, and D. Y. H. Pui. 1986. Diesel Particle Control with a Simple
Electrostatic Trap. SAE Paper No. 860009.
Kittelson, D. B., et al. 1988. Characterization of Diesel Particles in the Atmosphere.
Coordinating Research Council AP-2 Project Group Final Report.
Kittelson, D. B., et al. 1990. Particulate Emissions from Diesel Engines: Influence of In-
Cylinder Surface. SAE Paper No. 900645, [also in 1990 Transactions of the SAE, and J. of
Engines, 99(3), 1991].
Kittelson, D. B., and J. H. Johnson 1991. Variability in Particle Emission Measurements in
the Heavy Duty Transient Test. SAE Paper No. 910738, Published in Diesel Particulate
Emissions: Measurement Techniques, Fuel Effects and Control Technology SAE Pt-42, pp.
137-162, Edited by J. H. Johnson, et al. SAE, Warrendale, PA. [Also see An Examination of
the Sources of Variability in Particle Emission Measurements by the Transient Test Method
for Heavy-Duty Diesel Engines, Final Report prepared for the Engine Manufacturers
Association Calibration Task Force, 1990.]
Kittelson, D. B., et al. 1992. Oxidation of Soot Agglomerates in a Direct-Injection Diesel
Engine. SAE Paper No. 920111.
Kittelson, D. B., W. F. Watts, and M. J. Arnold. 1998a. Diesel Exhaust Particle Measurement
Instruments - Supplemental Report No. 1 for EPA Grant Review Of Diesel Particulate Matter
Sampling Methods.
Kittelson, D. B., M. J. Arnold, and W. F. Watts. 1998b. Aerosol Dynamics, Laboratory and
On-Road Studies - Supplemental Report No. 2 for EPA Grant Review Of Diesel Particulate
Matter Sampling Methods.
Kittelson, D. B., 1998a. Engines and Nanoparticles: A Review. J. Aerosol Sci., Vol. 29, No.
5/6, pp. 575-588, 1998.
Kittelson, D. B. 1998b. Personal communication with P. Ziemann University of California –
Riverside.
Kittelson, D. B. and I. Abdul-Khalek. 1998. Nanoparticle Formation During Exhaust Dilution.
Proceedings of 2nd ETH Workshop on Nanoparticle Measurement, Zurich, Switzerland,
August 7, 1998, 9 pp.
Kreso, A., J., et al. 1998. A Study of the Effects of Exhaust Gas Recirculation on Heavy-Duty
Diesel Engine Emissions, SAE Paper No. 981422.

01/14/99 Page 59
Kreso, A. J., et al. 1998. A Study of the Vapor- and Particle-Phase Sulfur Species in the
Heavy-Duty Diesel Engine EGR Cooler, SAE Paper No. 981423.
Kruger, M., et al. 1997. Influence of Exhaust Gas Aftertreatment on Particulate
Characteristics of Vehicle Diesel Engines. Research Report of the Forschungsvereinigung
Verbrennungskraftmaschinen e.V. (FVV), MTZ Motortechnische Zeitschrift 58:27-30.
Liu, B. Y. H. and D. Y. H. Pui. 1974a. On the Performance of the Electrical Aerosol
Analyzer. J. of Aerosol Sci., 6:249-264.
Liu, B. Y. H. and D. Y. H. Pui. 1974b. A Submicron Aerosol Standard and the Primary,
Absolute Calibration of the Condensation Nucleus Counter. J. of Colloid Interface Sci.,
47:155.
Liu, B. Y. H., et al. 1974. A Portable Electrical Analyzer for Size Distribution Measurement
of Submicron Aerosols. J. of the Air Pollution Control Assoc., 24(11):1067-1072.
Luders, H. 1997. Effect of Measurement Conditions on Ultrafine Particle Emissions. In
Proceedings of the First International ETH-Workshop on Nanoparticle Measurement, A.
Mayer, ed., Zurich, Switzerland.
Luo, L., et al. 1989. Particle Growth and Oxidation in a Direct-Injection Diesel Engine. SAE
Paper No. 890580.
Maher, E. F. and N. M. Laird. 1985. EM Algorithm Reconstruction of Particle Size
Distributions from Diffusion Battery Data. J. of Aerosol Sci., 16:557-570.
Maricq, M. 1998. Vehicle Particulate Emissions: A Comparison of ELPI, SMPS and Mass
Measurement. Included in Presentation Handouts (Presented at International Seminar on
Particle Size Distribution Measurement From Combustion Engines, Espoo, Finland, May 18-
19, 1998.)
Maricq, M. M., D. H. Podsiadlik and R. E. Case. 1998. Beyond Mass Emissions: An
Examination of the Size Resolved and Transient Nature of Particle Emissions from Motor
Vehicles. Paper submitted to Env. Sci. Tech.
Marple, V. A, K. L. Rubow and S. M. Behm. 1991. A Microorifice Uniform Deposit
Impactor (MOUDI): Description, Calibration, and Use. J. Aerosol Sci. Technol., 14:434-446.
Marple, V. A., K. L. Rubow, and B. A. Olson. 1994. Low Pressure Stages for the Micro-
orifice Uniform Deposit Impactor (MOUDI). Fourth International Aerosol Conference
Abstracts, ed. R.C. Flagan, Vol. 2, 6 pp.
Matter, U., et al. 1997. Penetration of Diesel Particle Filters Evaluated by Various Soot
Analyzing Techniques. In Proceedings of the First International ETH-Workshop on
Nanoparticle Measurement, A. Mayer, ed., Zurich, Switzerland.
Matter, U., et al. 1998a. Particle Emissions from Off-road Diesel Engines Equipped with
Particle Filters and Fuel Additives. J. Aerosol Sci. (paper submitted for publication).
Matter, U., et al. 1998b. Dynamic Characterization of Diesel Particulate Emissions Using
Two In-Situ Sensors. Included in Presentation Handouts and Presented at the International

01/14/99 Page 60
Seminar on Particle Size Distribution Measurement from Combustion Engines, Espoo,
Finland, May 18-19, 1998.
Mayer, A., J. Czerwinski, and P. Scheidegger. 1996. Trapping Efficiency Depending on
Particulate Size. SAE Paper No. 960472, also in Diesel Exhaust Aftertreatment 1996 (SAE
SP-1140), p163-174.
Mayer, A., et al. 1998. VERT: Diesel Nanoparticles Emissions: Properties and Reduction
Strategies. SAE Paper No. 980593, 12 pp.
McMurry, P. H. and J. C. Wilson. 1982. Growth Laws for the Formation of Secondary
Ambient Aerosols: Implications for Chemical Conversion Mechanisms. Atmospheric Environ.
161, p. 121-134.
Mirabel, P., J. L. Katz. 1974. Binary Homogeneous Nucleation as a Mechanism for the
Formation of Aerosols. J. Chem. Phys., Vol. 60, No. 3, pp. 1138-1144.
Moon, D. P. and J. R. Donald. 1997. UK Research Program on the Characterization of
Vehicle Particulate Emissions. (Contract Report for the Department of the Environment,
Transport and the Regions and the Society of Motor Manufacturers and Traders by ETSU – a
group within AEA). ETSU-R98, 90 pp.
Morrow, P. E., et al. 1964. Deposition and Retention Models for Internal Dosimetry of the
Human Respiratory Tract (Report of the international Commission on Radiological
Protection: ICRP: Task Group on Lung Dynamics). Health Phys. 12:173-207.
National Institute for Occupational Safety and Health. 1988. Carcinogenic Effects of
Exposure to Diesel Exhaust. CIB 50, 30 pp.
Nauss, K., et al., 1995. Critical Issues in Assessing the Carcinogenicity of Diesel Exhaust: A
Synthesis of Current Knowledge. Chapter in Diesel Exhaust: A Critical Analysis of Emissions,
Exposure, and Health Effects. A Special Report of the Institute’s Diesel Working Group.
Health Effects Institute, pp. 11-61.
Noble, C. A. and K. A. Prather. 1996. Real-time Measurement of Correlated Size and
Composition Profiles of Individual Atmospheric Aerosol Particles. Environ. Sci. Technol.,
Vol. 30, pp. 2667-2680.
Noble, C. A. and K. A. Prather. 1998. Aerosol Time-of-Flight Mass Spectrometry: A New
Method for Performing Real-time Characterization of Aerosol Particles. Appl. Occup.
Environ. Hyg. J. 13(6):439-443.
Pierson, W. R., et al. 1983. Particulate Matter Associated with Vehicles on the Road. ll.
Aerosol Sci. Technol. 2:1-40.
Pui, D. Y. H. and D. L. Swift. 1995. Direct-Reading Instruments for Airborne Particles. Ch.
16 in Air Sampling Instruments for Evaluation of Atmospheric Contaminants. Ed. by B. S.
Cohen, and S. V. Hering. ACGIH, Cincinnati, OH, 8th Edition, pp. 337 – 368.
Raabe, O. G. 1982. Deposition and Clearance of Inhaled Aerosols. Mechanism in Respiratory
Toxicology, Vol. 1, pp. 27-76. H. Witschi and P. Nettesheim, Eds. CRC Press, Boca Raton,
FL.

01/14/99 Page 61
Reichel, S., F. F. Pischinger and G. Lepperhoff. 1983. Influence on Particles in Diluted Diesel
Engine Exhaust Gas. SAE Paper No. 831333.
Rickeard, D., et al. 1996. Exhaust particulate Size Distribution: Vehicle and Fuel Influences
in Light Duty Vehicles. In Diesel Engine Combustion and Emission Control (SAE SP-1206)
SAE Paper No. 961980, p. 87-111.
Sawyer, R. F. and J. H. Johnson. 1995. Diesel Emissions and Control Technology. Chapter in
Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. A Special
Report of the Institute’s Diesel Working Group. Health Effects Institute, pp. 65-81.
Shi, J. P. and R. M. Harrison. 1997. Measurement of Fine Particle Size Distribution in Diesel
Exhaust and Ambient Air. Presented at the First International ETH-Workshop on
Nanoparticle Measurement. August 7, 1997, Niederrohrdorf, Switzerland.
Siegmann, K., L. Scherrer, and H. C. Siegmann. 1998. Included in Presentation Handouts
and Presented at International Seminar on Particle Size Distribution Measurement from
Combustion Engines, Espoo, Finland, May 18-19, 1998.
Silva, P.J. and K.A. Prather. 1997. On-line Characterization of Individual Particles from
Automobile Emissions. Environ. Sci. Tech., Vol. 31, pp. 3074-3080.
Sinclair, D., et al. 1979. Automatic Analysis of Submicron Aerosols in Aerosol Measurement
(D. A. Lundgren et al., eds.), University of Florida Presses, Gainesville, Florida, pp. 544-563.
Sinclair, D. 1986. Measurement of Nanometer Aerosols. Aerosol Sci. and Tech., 5:187-204.
Thomas, J. W. 1955. The Diffusion Battery Method for Aerosol Particle Size Determination.
J. Colloid Sci., 10:246-255.
Trier, A. 1997. Submicron Particles in an Urban Atmosphere: A Study of Optical Size
Distributions-I. Atmospheric Environ. 31(6):909-914.
Tuch, T. H., et al. 1997. Variation of Particle Number and Mass Concentration in Various
Size Ranges of Ambient Aerosols in Eastern Germany. Atmospheric Environ. 31(24):4193-
4197.
Twomey, S. 1975. Comparison of Constrained Linear Inversion and an Alternative Nonlinear
Algorithm Applied to the Indirect Estimation of Particle Size Distribution. J. of Computation
Phys., 18:188-200.
Ulevicius, V., et al. 1994. Aerosol Size Distribution Transformation in Fog. Atmospheric
Environ. 28(5):795-800.
Venkataraman, C. et al. 1994. Size Distributions of Polycylic Aromatic Hydrocarbons and
Elemental Carbon. 1. Sampling, Measurement Methods, and Source Characterization. Env.
Sci. Technol. 28(4): 555-562.
Vuk, C. T., M. A. Jones, and J. H. Johnson 1976, The Measurement and Analysis of the
Physical Character of Diesel Particle Emissions, SAE Paper No. 760131.
Wang, S. H. and R. C. Flagan. 1990. Scanning Electrical Mobility Spectrometer. Aerosol
Sci. and Tech., 13:230-240.

01/14/99 Page 62
Watson, J. G. et al. 1998. Northern Front Range Air Quality Study, final report, prepared for
Colorado State University.
Weingartner, E., et al. 1997a. Aerosol Emission in a Road Tunnel. Atmospheric Environ.
31(3):451-462.
Weingartner, E., et al. 1997b. Hydroscopic Properties of Carbon and Diesel Soot Particles.
Atmospheric Environ. 31(15):2311-2327.
Weiss, M. 1997. An On-line Mass Spectrometer for Aerosols: Development,
Characterization and Applications. Ph. D. Dissertation, Delft University of Technology,
Netherlands.
Welker, R. W., et al. 1982. Mine Particulate Size Characterization. U. S. Bureau of Mines
Open File Report, Prepared by IIT Research report number C06572-Final, p. 96.
Whitby, K. T. and B. K. Cantrell. 1975. Atmospheric Aerosols - Characteristics and
Measurement. ICESA Conference Proceedings, IEEE #75-CH 1004-1 ICESA, paper 29-1, 6
pp.
Whitby, K. T., et al. 1975. Characterization of California Aerosols - 1. Size Distributions of
Freeway Aerosol. Atmospheric Environ. 9:463-482.
Williams, P. T., et al., 1987. The Role of Lubricating Oil in Diesel Particulate and Particulate
PAH Emissions. SAE Paper No. 872084.
Williams, P. T., et al., 1989a. The Influence of PAH Contamination of Lubricating Oil on
Diesel Particulate PAH Emissions. SAE Paper No. 890825.
Williams, P. T., et al., 1989b. Diesel Fuel Dilution and Particulate Absorption Contamination
in Used Lubricating Oil. SAE Paper No. 892080.
Wilson, W. E., et al. 1977. General Motors Sulfate Dispersion Experiment: Summary of EPA
Measurements. JAPCA 27(1):46-51.
Winer, A. W. and Busby, W. F., 1995. Atmospheric Transport and Transformation of Diesel
Emissions. In Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects.
Health Effects Institute, Cambridge, MA, pp. 83-106.
Winklmayr, W., et al. 1991. A New Electromobility Spectrometer for the Measurement of
Aerosol Size Distributions in the Size Range From 1 to 1000 nm. J. of Aerosol Sci.,
22(3):289-296.
World Health Organization. 1996. Diesel Fuel and Exhaust Emissions. International Program
on Chemical Safety, Environmental Health Criteria 171, Prepared by members of the
Fraunhofer Institute of Toxicology and Aerosol Research, Germany, under the coordination
of G. Rosner. WHO, Geneva, Switzerland, 389 pp.
Yeh, H. 1993. Chapter 18 Electrical Techniques. In Aerosol Measurement Principles
Techniques and Applications, K. Willeke and P. Baron editors, Van Nostrand Reinhold, NY,
NY., p. 410-426.

01/14/99 Page 63
Ziemann, P., et al. 1995. Measurement of Particle Induced Nucleation and Growth in Low
Pressure Chemical Vapor Deposition Systems with the Particle Beam Mass Spectrometer. J.
Aerosol Sci. Vol. 26, Suppl. 1, p. S633-S634

01/14/99 Page 64