United States
Environmental Protection
Agency
Environmental Sciences Research
Laboratory
Research Triangle Park NC 27711
SEPA
Research and Development
EPA-600/S2-81-237 Nov. 1982
Project Summary
Techniques for the
Measurement of Aerosol
Carbon Content
Edward S. Macias
In this summary, two techniques for
total and elemental carbon analysis
are described. Both methods are
totally instrumented, automated, and
nondestructive. Total carbon is
determined using the gamma-ray
analysis of light elements (GRALE)
technique. Elemental carbon is
determined by a light reflectance
method. The extension of the methods
to Teflon filters is also described.
This Project Summary was developed
by EPA's Environmental Sciences
Research Laboratory, Research
Triangle Park, NC, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Ambient aerosols, particularly those
with diameters of less than 3 //m are a
serious pollution problem. Carbon-
aceous material is a major component
of the fine particle concentration and
has undergone extensive study in the
past few years, in large part due to
concern that these particles play an
important role in urban haze and
community health.
Particulate carbon in the atmosphere
'exists predominantly in three forms:
elemental carbon (soot) with attached
hydrocarbons, organic compounds, and
carbonates. Carbonaceous urban fine
particles are composed mainly of
elemental and organic carbon. These
particles can be emitted into the air
directly in the particulate state or
condense rapidly after introduction into
the atmosphere by chemical reactions
involving gaseous pollutant precursors
(secondary aerosol).
In this report, the development of two
techniques for total and elemental
carbon analysis is described. Both
methods are totally instrumented,
automated and nondestructive. Total
carbon is determined using the gamma-
ray analysis of light elements (GRALE)
technique. This method involves the in-
beam measurement of 7 rays emitted
during the inelastic scattering of
protons accelerated in a cyclotron. In
the second method, elemental carbon is
determined by a light reflectance
method. Much of the analysis has been
of aerosols deposited on low carbon
glass and quartz filters. Extension of the
methods to Teflon filters is also
described.
Total Carbon Analysis by the
GRALE Technique
The GRALE method is a non-destructive
technique based on the measurement
of gamma-ray emission induced by
proton bombardment of aerosol
samples. The y-ray energy is unique to a
particular nuclide and thus can be used
as a signature for the element.
Elemental concentrations are obtained
in units of //g crrr2 and are not affected
by the chemical form of the elements.
The GRALE technique has been used for
-------
Top View
Slide Projector
Proton
Beam
Havar Foil,
Aerosol Sample
Mini Computer
28K x J 6 bits
Figure 1. Schematic diagram of the sample irradiation chamber and electronics
used for carbon, sulfur, and nitrogen analysis.
total carbon analysis here but the
method can also be used to determine
the concentration of other elements
such as sulfur, nitrogen, and oxygen.
Filter samples are collected and
mounted in 5 x 5 slide mounts after
removal of the cellulose backing. The
samples are irradiated with a collimated
beam of 7 MeV protons in the external
beam facility of the Washington
University, St. Louis, 135-cm sector
focused cyclotron with an automated
sample changing chamber maintained
in 1 atm of helium. The experimental
setup is shown schematically in Figure
1.
Gamma rays produced in the proton
bombardment are detected with a 60
cm3 lithium drifted germanium Ge(Li)
detector. The 7-ray data are stored and
processed in an on-line minicomputer.
The spectra are analyzed immediately
after each irradiation with the on-line
computer. The1 intensity of each peak is
determined from the integrated peak
area after subtracting the background
and correcting for system dead-time
losses (typically 20 percent). These data
are normalized to the proton beam
intensity, and intensities of the filter
blanks are subtracted from the aerosol
results.
Carbon peak intensity is converted
into carbon mass by calibrating with
standard methionine aerosol
(CSH1102SN) deposited on the same type
of filter. The mass of methionine
deposited on the filter is determined
using a beta attenuation mass monitor.
Filter blanks, atmospheric, laboratory,
and methionine standard aerosol
samples are run under nearly identical
conditions which yield nearly equal
detector count rates. All samples are
analyzed similarly.
The GRALE total carbon technique
has been extensively compared to the
Dohrmann DC-50 carbon analyzer and
the Perkin-Elmer Model 240 Elemental
Analyser. The latter two methods are
based on the high temperature
combustion of carbon compounds with
subsequent detection of CH4 and C02,
respectively. This intercomparison
indicates that, to within 5 percent, no
systematic discrepancies exist in any of
the methods investigated. The GRALE
technique is non-destructive and
unaffected by the chemical form of the
aerosol, two advantages over
combustion methods.
Elemental Carbon Analysis by
Reflectance
A laboratory reflectometer. shown
schematically in Figure 2, has been
constructed for use with fine particle
filter samples. This light-tight system
consists of two blackened aluminum
tubes (5.5-cm dia.) each mounted at 45°
to an aluminum sample chamber which
holds standard 5 x 5 cm slide mounts. A
12-V Tungsten filament lamp powered
by a stabilized 12-V supply is used as a
light source. Plano-convex lenses are
used to focus the incident light onto the
sample (0.3 cm2) and reflected light onto
a solid state photodetector. The photo-
detector output is monitored with a
digital voltmeter. Unexposed filter, taken
from the same filter roll as the samples,
is used to determine the initial, or blank,
intensity (I0).
Elemental carbon (soot) standards
were prepared by the combustion of
polystyrene, paraffin, and butane. The
aerosol generated was aspirated into a
5-L container, mixed with filtered air
and drawn through a sampler equipped
-------
Elemental Carbon Analyzer
Lamp
Solar
Collimator
Sample Holder
Figure 2. Reflectance photometer.
with the on-line reflectometer. The
mass of the deposit was determined by
beta attenuation. The reflectance of the
spot before and after loading was also
measured. Since the aerosol could
contain appreciable nonetemental
carbon (organic) compounds, a method
for their removal was developed. A set
of butane-generated samples was
mounted over a 1.3-cm dia. hole in 5 x
5 cm aluminum plates and total carbon
was determined via the GRALE
technique. The samples were then
placed in the oven of a conventional gas
chromatograph equipped with a
programmable temperature ramp.
Samples were heated at 20°C min'1 toa
fixed temperature and left at that
temperature for 20 min. The reflectance
of the sample was measured after
cooling to room temperature; the
procedure was then repeated at a
higher temperature. The reflectance of
the sample was nearly constant up to
350°C. Above 350°C, the reflectance
relative to 25°C increased, indicating
loss of elemental carbon. Therefore, a
temperature of 300°C was used to
remove organic compounds from the
soot samples with minimal loss of
elemental carbon.
Analysis of butane-generated
samples for total carbon content by
GRALE after heating to 300° shows
—25% carbon loss upon heating without
appreciable change in reflectance. The
after-heating GRALE carbon mass was
taken to be the actual elemental carbon
content of the samples which were then
considered elemental carbon
standards.
A calibration curve was prepared by
plotting the log of the ratio of the final to
initial reflectance of the elemental
carbon standards versus the carbon
mass determined by GRALE analysis
after heating to 300°C. Alternately, the
total mass of the soot standard after
heating to 300°C determined by beta
attenuation can be used in place of the
GRALE carbon mass.
Measurements with
Atmospheric Samples
Atmospheric fine particle aerosol
samples were collected on the roof of
the six-story Chemistry Building of
Washington University, located six
miles west of downtown St. Louis. The
immediate area is primarily residential
with a major expressway one mile away
which is usually upwind from the
sampling site. Samples were collected
on an automated TWOMASS sampler
equipped with the on-line reflectometer
and beta-attenuation system for soot
carbon and mass analysis. The samples
were analyzed via the GRALE technique
for total carbon content. After the
samples were heated, the carbon
content and reflectance were
remeasured. Figure 3 shows the
average decrease in total and elemental
carbon content versus the temperature
of two samples collected on preashed
Pallflex TissueQuartz (2500 QAO)
filters. A significant amount of carbon is
lost at temperatures up to 300°C
without an appreciable loss of
elemental carbon. It may be assumed
that this carbon loss at or below 300°C
is due to volatile organic compounds.
However, it is not certain that all organic
compounds have been removed at this
temperature. The amount of organic
carbon loss below 300°C is dependent
on the composition of the aerosol.
Between 300 and 450°C, the carbon
content continues to decrease with a
corresponding decrease in reflectance
due primarily to the loss of elemental
carbon.
It is instructive to compare the
amount of total and elemental carbon
determined after heating atmospheric
aerosol samples to 300°C to estimate
the amount of nonelemental carbon
remaining at that temperature. The data
show a high correlation (r = 0.96)
indicating a linear relationship between
elemental carbon and total mass. The
least squares fit to the data has a slope
of nearly unity (0.9 ± 0.2) indicating
good agreement between total and
elemental carbon. The y intercept is 3 ±
4/ug cnrT2, and indicates Jhere may be a
small amount of nonelemental carbon
remaining at 300°C in the ambient
samples.
Total and Elemental Carbon
Analysis on Teflon Filters
The work described above employed
glass or quartz filters. Teflon filters
present a different problem because
this material is 25 percent carbon
[(C2F4)n]. Previously, no carbon analysis
method has been shown to be
applicable with this high carbon filter.
However, if the carbon to fluorine mass
ratio is constant in Teflon filters, the
fluorine y-ray can be used as an internal
standard in GRALE analysis to indicate
the thickness of the filter blanks and to
determine the carbon filter blank value.
In this section, the successful extension
of the GRALE and reflectance analysis
techniques for total and elemental
carbon for use with Teflon filters is
described.
Teflon filters with 1 //m pore size and a
nominal areal density of 1mg/cm2
mounted on a 37-mm polyolefin ring.
-------
§
•a
ro
CJ
150
300 450
Temperature, °C
600
750
Figure 3. Average decrease in total and elemental carbon in
atmospheric aerosol samples versus increasing temperature.
and Pallflex E-70 glass fiber filter tape
with a detachable cellulose backing,
were used as filter media.
Ambient aerosols were collected with
a manual dichotomous virtual impactor
sampler and a TWO MASS automated
sequential tape sampler from the same
aerosol manifold simultaneously. Both
samplers fractionated the aerosol into
two size classes; the fine fraction
consisting of particles with
aerodynamic diameters of less than 3
//m were examined.
The dichotomous sampler employed
Teflon filters; the TWOMASS sampler
used glass fiber as a filter medium. The
dichotomous sampler operated at a flow
rate of 14 L min~1 was used to collect 6-,
12-, and 24-h samples. Six-h samples
were collected with the TWOMASS
sampler at a flow rate of 12 L min"1.
Both samplers were equipped with
automatic flow controllers.
GRALE analysis of blank Teflon filters
yielded a linear relationship between
fluorine counts and carbon counts (r =
0.99), an indication that fluorine can be
used for subtraction of carbon
background from Teflon fUters. Teflon
blanks gave a C/F count ratio of 0.128 ±
0.004. The large variation in carbon
counts on blank filters shows that
reliable calibration can be achieved only
when a correction for carbon variations
in the Teflon filters is made by using the
fluorine peak intensity as an indicator.
The average .intensity of the 1.35-
MeV fluorine peak from the blank filters
was chosen as the standard reference
peak for fluorine. The 4.43-MeV carbon
peak intensity associated only with the
deposited carbon was obtained using
the following equation:
CD = [Cs - (C/F)BFs] * NB..m
where CD is the carbon intensity from
the deposit, Cs is the total carbon
intensity from the sample, (C/F)s is the
average C/F ratio for blank filters, Fs is
the total fluorine intensity from the
sample, and NBeam is a normalization
factor which corrects for variations in
the proton beam intensity. Using this
method, a calibration curve for carbon
obtained from methionine standards
was constructed. The proportionality
between the measured x-ray counts
and deposited carbon was excellent
(r=0.99).
A calibration curve for elemental
carbon, using butane soot deposited on
Teflon filters was prepared by plotting
the log of the ratio of reflectance of the
blank filter (I0) to the reflectance after
addition of the soot (I) versus carbon
mass determined by the GRALE
technique. The deposited soot mass on
these calibration samples was also
determined gravimetrically. The
comparison of these two methods of
calibration gave an excellent linear
correlation with a slope of 1.02 ± 0.02
and a correlation coefficient of 0.99.
The total carbon detection limit is
estimated from the irradiation con-
ditions and counting statistics. The
minimum sensitivity (1tr) with Teflon
filters is 10pg/cm2 which corresponds
to 1.5 /ug/m3 for 6-h sampling at 14 L
min'1 with a 10-mm diameter sample
deposit. This finding compares to 4.5
Aig/cm2 on glass fiber filters or 0.7
//g/m3for 6-h sampling under the same
conditions as above.
The reflectance analysis of elemental
carbon on Teflon filters has a minimum
detectable limit of 1.8 Aig/cm2 or 0.3
/ug/m3 for 6-h sampling as given above.
The upper detection limit of the
reflectance method due to saturation of
the surface is ~70 jug/cm2.
Comparison of Carbon
Analysis on Teflon and
Glass Fiber Filters
Fine ambient aerosols were collected
on both Teflon and glass fiber filters
during the period of February 16-March
8, 1980, in St. Louis. Comparison of the
results on Teflon and glass fiber filters
for total carbon analyses by the GRALE
technique and for elemental carbon by
reflectance analysis was made during
-------
this period. The results from both filters
agree (total carbon slope = 0.99 ± 0.09,
r = 0.95; elemental carbon slope = 1.18
± 0.26, r = 0.93).
Conclusions and
Recommendations
Through a detailed series of
intercomparisons, two new methodsfor
the measurement of total and elemental
aerosol carbon concentrations have
been established. The methods have
since been extensively used for studies
in Charleston, WV, St. Louis, MO, the
southwestern United States, Los
Angeles, CA, and China Lake, CA, and
are recommended for future usage.
Edward S. Mac/as is with Washington University, St. Louis, MO 63130.
Charles W. Lewis is the EPA Project Officer (see below).
The complete report, entitled "Techniques for the Measurement of Aerosol
Carbon Content," (Order No. PB 82-249 152; Cost: $7.50, subject to change)
will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
. S. GOVERNMENT PRINTING OFFICE: 1982/659-095/0548
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
PS 00003^9
U S ENVIR PROTECTION AGENCY
RESlUfM 5 U8RAXY
aiO S DEARBGKN STKEET
CHICAGO It 60604
------- |