EPA/600/A-96/087
Imaging Mass Spectrometry of Particulate-Associated
Polynuclear Aromatic Hydrocarbons
Michael J. Holland, Richard W. Linton
Department of Chemistry, CB# 3290,
University of North Carolina, Chapel Hill, NC 27599-3290
Joette L. Steger, Raymoix! G. Merrill
Radian Corporation, P.O. Box 13000, Research Triangle Park, NC 27709
Greg S. Strossman, Thomas F. Fister, Robert W. Odom
Charles Evans & Associates, 301 Chesapeake Dr., Redwood City, CA, 94063
Time-of-flight secondary ion mass spectrometry (TOF-SIMS) was employed for die in situ investigation of PAHs
on individual environmental particulates. TOF-SIMS is a microprobe mass spectrometry technique capable of
providing surface elemental and chemical information with lateral spatial resolution < 1 **m. TOF-SIMS analysis
allows for the correlation of microscopic characteristics of different particulate, filter, and sorbent substrates with
adsorbed organic species. Silver membrane and glass fiber filters loaded with coal flyash or silica gel were placed
in a Method 0010 spiking train and exposed to vapor phase benzo[a]anthracene, beozo[g,h,i] peryleoe or pyrei*.
TOF-SIMS allowed surface chemical analysis of single particles on Ag filter media taken direcdy from the
Method 0010 sampling train, without additional sample preparation. Secondary ion images indicated that
beozo[a]anthracene is found primarily on the filter media and not on the flyash surfaces, and is of interest with
respect to the elucidation of sampling artifacts. Particulate matter was also transferred to silicon wafers where
native levels of diesel soot POM were imaged using TOF-SIMS.
INTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs), present in the effluent streams of combustion sources in both the vapor
phase and as adsorbed species on particulate matter, are of regulatory interest because many are known or
suspected carcinogens. Measured particle-size distributions of PAHs in rural and urban environments reveal that
PAHs are most often associated with particles < 1.4 nm, the size fraction with die greatest retention time in the
lung.1 Furthermore, recent epidemiological studies present consistent evidence that chronic exposure to particulate
matter, particularly the respirable fraction with aerodynamic diameter < 2.S^m, is responsible for increased
morbidity and mortality in populations of susceptible individuals (e.g., those with chronic obstructive pulmonary or
cardiovascular diseases) as well as the general populace.2*3
Environmental particles are complex mixtures of particle sizes and compositions. Coal combustion plume
flyash is a heterogeneous mixture of particles which can be broadly categorized into three major fractions: mineral
particles, mainly oxides of Si, Al, Ca and Mg with trace metal impurities; magnetic particles, which are mixtures
of various Fe oxides; and carbonaceous particles representing (he organic coal fraction. Electrostatic precipitator
(ESP) hopper ash is often used as a surrogate for plume cm- stack flyash in studies of environmental particles.
Another common anthropogenic particle, diesel soot, consists of agglomerations of sub-micron, spherical,
carbonaceous particles with substantial native levels of nitro-PAH.4 Studies using model particle surfaces and
different coal flyash types have shown that the particle substrate to which die organic molecules are adsorbed can
affect the photochemical reactivity of the adsorbed species.5,6,7
Currently, studies of PAH adsorption and transformation primarily rely upon bulk analysis schemes whereby large
quantities of particles, > 100 mg, are extracted using Soxhlet or supercritical fluid extraction techniques.* The
extracts are then concentrated and analyzed by HPLC, GC or GC/MS. Bulk extraction methods have a number of
disadvantages. Extraction into an organic matrix and subsequent sample work-up can lead to non-quantitative
recovery. Irreversible adsorption of some POMs to substrates may also limit die accuracy of extraction-based
1
-------
techniques.9,10 In addition, extractions average results over a large number of particles and recover POM not only
from the surface but also from within particle pores. Important information regarding organic surface comporents
and their distribution among the various fractions of a heterogeneous particulate matter like flyash is lost
Correlation of the organic species with surface elemental composition, which can differ significantly from the bulk
particle composition, is difficult or impossible.11
The surface chemistry of organic toxics including photochemical reactions, heterogeneous reactions with gaseous
pollutants, and the effects of particle or filter substrates on sorbed organics are not well understood.
Photochemical reactions of adsorbed POMs can alter toxicity by creating products of differing mutagenicity fan
die parent compound.12 Furthermore, die filter material used in high-volume sampling of ambient airborne
particulate matter appears to subtly impact die measured mutagenicity of PAHs on ambient airborne particles.13
Thus, in situ chemical analysis of individual particles is desirable for its ability to detect both precursors and their
breakdown products directly in their native, most environmentally relevant, chemical environment.
Microprobe techniques have been employed for the surface elemental analysis of single particles. For example,
dynamic secondary ion mass spectrometry (SIMS) and scanning electron microscopy (SEM) coupled with energy
dispersive X-ray detection have characterized trace elements on atmospheric particles.14'15 Neither of these
methods, however, are well-suited for the analysis of adsorbed organics. Laser ablation mass spectrometry, which
is capable of both organic and inorganic analyses, has been applied to the single particle analysis of coal
flyash.16'17 None of these studies attempted to correlate POM adsorption and reactivity with particle composition.
Time-of-flight secondary ion mass spectrometry, well suited for the surface analysis of organic species, has been
shown to provide a method for the in sUu analysis of organics on single particles. TOF-SIMS was employed for
the analysis of organics on polystyrene beads, a commonly used substrate in combinatorial chemistry. Recently,
TOF-SIMS was used to determine that die carbonaceous fraction of flyash particles coated with submonolayer
coverages of benzo[e]pyrene (BeP) had higher coverages of the adsorbate than did the mineral fraction.19 That
study also demonstrated the ability to monitor the photooxidation of benz[a] anthracene on individual particle
surfaces; however, the solution coating scheme employed may swell the particles or alter the levels and distribution
of native PAHs. In die present study, the feasibility of TOF-SIMS is evaluated for analysis of toxic organics on
coal flyash and diesel soot at either ambient levels or following vapor-phase spiking of PAH. The utility of TOF-
SIMS as a tool for investigating artifacts in the Method 0010 sampling train is also demonstrated.
EXPERIMENTAL METHODS
Materials
ESP hopper ash was obtained from die CP&L Cape Fear power plant (Moncure, NC). The flyash was size-
fractionated according to aerodynamic diameter, and the 3-10 ^m-diameter fraction was used in this research. No
further separation of the flyash into carbonaceous, mineral or magnetic fractions was made prior to use. The
surface area of flyash was 0.85 m2/g using BET nitrogen adsorption. Diesel soot, Standard Reference
Material 1650, was obtained from the National Institute of Standards and Technology (NIST). Table 1 lists the
certified and non-certified values for organic compound concentrations in the diesel soot NIST analyses were
based upon Soxhlet extraction of 50 to 150 mg samples and either GC/MS or HPLC analyses.4 Silver membrane
filters, 142 mm diameter with an average pore size of 1.2 txm, were obtained from Poretics Corp. (Livermore,
CA). Whatman 11.0 cm, GF/A glass microfiber filters were also used. Filters were cleaned by rinsing a
minimum of three times with HPLC-grade methylene chloride prior to use. 1,2-Benz[a]anthracene (BaA) and
pyrene were purchased at 99% purity from Aldrich Chemical Co. Benzo[g,h,i]perylene (BghiP) was purchased at
99% purity from Radian Corp. (Austin ,TX). All compounds were used as received.
Spiking nf Particles
Ambient PAH levels on flyash are estimated to be in die range of 10~* monolayers, thus hindering their direct
detection on single particles by TOF-SIMS imaging.19 Using a Method 0010 sampling train for vapor phase
2
-------
spiking, particles were coated with higher levels of PAH to facilitate TOF-SIMS technique development. The
Method 0010 train consists of a probe and particulate filter, both heated, in series with a condenser, a XAD-2
sorbent module and a train containing knockout, water and silica gel impingers.20 The train employed to spike
particle-loaded filters in this project was a slight modification of the 0010 train: the heated probe and nozzle were
replaced by an atmosphere generator manifold and a heated spiking elbow through which a syringe pump
introduced a methylene chloride spiking solution. Prior to PAH spiking in the 0010 train, particles were loaded, as
received, onto filter media using a Medic-Aid aerosol generator. The temperature of the filters during spiking
averaged 55 °C.
The total amount of PAH spiked into any Method 0010 sampling train was approximately 200 Mg- The PAH
spiked onto flyash-loaded filters was determined by extracting the spiked filters in Soxhlet extractors for a
minimum of 8 hours in methylene chloride. The extracts were analyzed on a Hewlett-Packard 1050 liquid
chroma tograph equipped with a Phenomenex Envirosep-PP column (5^m, 125 x 3.2 mm) using water/ acetonitrile
gradient elution. The HPLC was equipped with dual UV/programmable fluorescence detection.
Instrumentation
TOF-SIMS analyses were performed on a Charles Evans & Associates TOF-SIMS system as shown in Figure 1.
The mass spectrometer consists of drift sections coupled to three electrostatic analyzers. Ions were detected using
a dual microchannel plate followed by a single stop time-to-digital converter (TDC). For these analyses, the ion
microprobe mode was utilized to optimize lateral resolution. A Ga+ beam operating at 25 keV with a beam
current of 640 pA was rastered across die surface sputtering off neutrals as well as positive and negative secondary
ions characteristic of the near surface composition (top 1-2 nm). The pulse repetition rate of the source was
10 kHz. Pulse widths were varied from 18-2 ns, resulting in corresponding spatial resolutions of 0.2-1.0 ^m.
Mass resolution, however, improved with decreasing pulse width from 300 (M/AM at 28 Da, FWHM) with 18 ns
pulse to 1100 with 2 ns pulse. Analysis times typically ranged from 15-35 minutes and were performed at low total
primary ion doses (< 1012 ions/cm2), known as the static SIMS domain. This dose corresponds to removal of
< 1 % of a monolayer of the surface, thereby minimizing beam damage effects on native molecular species.
Particle Analysis
Samples were prepared initially by cutting pieces approximately 1 cm2 from die particle-loaded filters and loading
them directly into the TOF-SIMS to assess die feasibility of performing single particle chemical analyses directly
upon filter media. Particles were also transferred from filter media onto pieces of silicon wafer by first cutting out
a small piece of the filter and tapping it above die wafer. Excess particles not adhering to the wafer by
electrostatic forces were removed by gently tapping the wafer prior to loading into the TOF-SIMS sample bolder.
This procedure tended to produce small clumps of a few particles spaced 1-20 Mm apart in the case of either silica
gel or coal flyash. Diesel soot tended to form agglomerates ranging from less than a micron to greater than 60 Mm
in length with 1-20 Mm spacing between agglomerates. Electron charge compensation was required when imaging
soot to prevent loss of signal in the analytical volume due to ion beam-induced electrostatic charging. Analysis
areas were typically 60 Mm x 60^m or 100 Mm x 100 Mm.
Instrumental conditions were optimized to improve detection of PAH on the particles. An 8 keV post acceleration
potential was used routinely with blanking of low mass secondary ions. Blanking masses below 35 is accomplished
by inactivating the TDC for a set time following die primary ion pulse. For samples imaged on silver membrane
filters, gating out mid-spectrum maw* (e.g., Ag+ ions) was necessary to prevent detector saturation. Gating is
accomplished by deflecting selected mass ions from die detector. The combination of blanking and gating allowed
die use of longer primary pulse widths which had the effect of increasing the intensity of higher mass peaks. TOF-
SIMS imaging of spiked flyash or native POM levels on diesel soot provided POM intensities of approximately 1 to
3 counts per pixel, for a typical pixel size of 0.055 Mm2.
Data were typically acquired in the positive ion mode using an 18 ns pulse width and blanking of masses 1-35.
Flyash samples were then reanalyzed using an 8 ns pulse width and no low mass blanking to allow sufficient mass
3
-------
resolution for elemental analysis of particle surfaces. A negative ion mode analysis was performed to allow further
characterization of flyash into mineral, magnetic, and carbonaceous particles. Due to their relatively hizh vapor
pressures, PAHs tend to sublimate during extended periods of exposure to die instrument's vacuum (Iff* Torr).
Thus, total instrumental residence times were limited to roughly 30 minute
RESULTS AND DISCUSSION
Secondary Ion Spectra of Flvush
One goal of this project is to evaluate the feasibility of using TOF-SIMS to image environmental particles directly
on air sampling filter media following vapor phase spiking. In lieu of the standard glass microfiber filter, silver
membrane filters were chosen initially because they were anticipated to have a mass spectrum free of overlap with
PAHs or flyash and because silver would act as a canonization agent enhancing PAH detection via the formation of
(PAH-Ag)+ adducts. Furthermore, a conductive imaging substrate is required to prevent surface charging from
ion bombardment during TOF-SIMS analysis. A silver membrane loaded with 5 mg flyash was spiked with 200 Mg
of BaA. HPLC analysis indicated that 8.6 jig of the total BaA spiked into the sampling train was retained on the
loaded filter. This suggests that heating of the Method 0010 filter holder tends to favor deposition of vapor phase
PAH not on die particulate matter but further downstream in the sorbent module. A more efficient vapor phase
spiking system, such as a fiuidized bed reactor, needs to be investigated for these experiments.
Figure 2 shows the mass spectrum of this flyash imaged directly on the silver membrane filter media. The
spectrum is dominated by signal from the silver membrane, yet strong BaA4 (228 Da) and (BaA + Ag)+ adduct
(335, 337 Da) signals were observed. Hie contribution of just the Ag+ (107, 109 Da), Ag2+ (214, 216, 218 Da),
Ag3* (321, 323, 325, 327 Da), and Ag2Cl+ (249, 251 253, 255 Da) mass ions ranged from 24-55% for spiked
particle-free, silica gel-loaded, and fiyash-loaded membranes. Figure 3 shows (a) the Ag2Cl+ ion image, and (b)
the BaA* ion image. Dark regions in the Ag2Cl+ image indicate the location of the flyash while the BaA+ image
indicates that the PAH was detected on the membrane and not on die flyash. The signal from die silver
overwhelmed die particle signal and prevented optimization of instrumental conditions for detection of PAH cm the
particles.
Secondary Ton Spectra nf Diesel Soot
Figure 4 shows TOF-SIMS positive ion images for unspiked diesel soot on a silicon wafer support Image 4(a) is
the total ion image followed by subsequent images of the molecular ions (M+) for die native PAHs at masses (b)
228, (c) 252, and (d) 276. The mass 228 ion image does not show a high degree of correlation between particle
location and the ion counts. The outline of die largest soot agglomeration is discernible in the mass 252 image;
whereas die crescent-shaped orientation of all three soot clumps appears more clearly in die mass 276 ion image.
The mass spectrum of the diesel soot is shown in Figure 5. The top spectrum is the total positive ion spectrum for
die 55 Mm x 55 Mm image field. Note that it closely resembles the region-of-interest (ROI) background spectrum
of the silicon wafer. Given die relatively small area of the image field occupied by soot, this is not unexpected.
Peaks at 191, 193, 207 and 221 Da are associated with the silicon wafer substrate. These peaks are most likely
due to polydimethyl-siloxane (PDMS, [SiO(CH3)]a) contamination. Both the substrate PDMS contamination and
die diesel soot, as discussed below, appear to contribute to a peak at mass 189 Da.
The middle plot in Figure 5 is an ROI spectrum of the soot agglomerate labeled as #1 in Figure 4. In this ROI, a
peak at 252 Da (C20Hu"f) corresponds to die molecular ions (M+) of benzo[a]pyrene (BaP), BeP,
benzo[k]fluoranthene (BkF), and perylene. A peak at 276 Da corresponds to the molecular ions of
BghiP and indeno[l ,2,3-cd]pyrene. A series of peaks at multiples of 13 Da below the M+ ion corresponding to the
loss of multiple CH groups are typically observed in the TOF-SIMS spectra of surface adsorbed POM [19]. Peaks
would be expected at 263 Da (CjjHji+), 250 Da (CjoHlo+), and 237 Da (Cj^Hj*) for the mass 276 PAHs; at
239 Da (C19Hu+), 226 Da (CUH1(/), and 213 Da (CnH,+) for the mass 252 PAHs; and at 215 Da (C17H„+),
202 Da (C16H10+), and 189 Da (C^H^) for the mass 228 PAHs. A series of peaks corresponding to the
4
-------
[(M-IHXCH^I* cations are expected as well [19]. Fragments are predicted at masses 226, 213,200 and 187 Da
for the mass 228 PAHs; at 250, 237, 224, and 211 Da for die mass 252 PAHs; and at 274, 261, 248, and 235 Da
for die mass 276 PAHs.
Confirmation of tbese assignments is beyond the scope of the work completed so far. Spildng soot with deuterated
compounds could substantiate the fragmentation patterns. Characteristic fragmentation patterns could prove useful
in establishing die relative coverages of isomeric PAHs co-adsorbed on a gpiven particle. Ion counts of fragment
and molecular ions from soot and the wafer were obtained from high mass resolution spectra and normalized to the
respective spectrum integrals. Hie ratio of soot signal to wafer signal for die first four [M-
-------
10. Harrison, F.L.; Bishop, D.J.; Mallon, BJ. Environ. Sci. Tech. 1985120). 186-193.
11. Linton, R.W.; Williams, P.; Evans, C.&.; Natusch, D.F.S. Anal Phem 1977 42, 1514.
12. Benson, J.M.; Brooks, A.L.; Cheng, Y.S.; Henderson, T.R.; White, J.E. Atmns Environ 1985 1QT7Y
1169-1174.
13. DeRaat, W.K.; Bakker, G.L.; DeMeijere, F.A. Atmos. Environ. 1990 24AOD, 2875-2887.
14. Cox, X.B.; Bryan, S.R.; Linton, R. W.; Griffis, D.P. Anal Chem 1987 22,2018.
15. Germani, M.S.; Buseck, R.P. Anal Chem 1991 (ft, 2232.
16. Denoyer, E; Natusch, D.F.S.; Surkyn, P.; Adams, F.C. Environ. Sci Tech. 198312, 457.
17. Wieser, P.; Wurster, R.; SeQer, H. Atmo* Environ 1980 14, 485.
18. Brummel, C.L.; Lee, I.N.W.; Zhou, Y.; Benkovic, SJ.; Winograd, N. Science 1994 264r 399.
19. Fister, T.F.; Strossman, G.S.; Willett, K.L.; Odom, R.W.; Linton, R.W. International Journal nf Ma
-------
Umm
Figure 1. Schematic of Charles Evans & Assoc. Imaging TOF-SIMS. The instrument is operated in microprobe
mode using the pulsed Ga* source for single particle analysis.
5
; UM;
2 3KB
S 2 BOD
jj iiaoi
1 w-
AiiCl
Agi
lit
a?
^lu..
¦M*
Z2S
u
241
Agi
323
XL.
a
... ji-,...
MO
2U
300
3ZI
Figure 2. Mass spectrum of BaA-spiked flyash imaged on silver membrane filter.
Figure 3. Secondary ion of BaA-spiked flyash on silver membrane, a) Ag2Cf, and b)BaA"\ Dark regions
indicate that BaA was not detected images on the flyash.
-------
.is
*&/£'>; '<£v.*
x^*': 1P^ •
0 J«»;
•] total ~ ion image
b) mass 228 ion image cj mass 252 ion image d] mass 276 ion Image
Figure 4. TOF-SIMS images of diesel soot on silicon wafer. The M* ions are shown for native PAHs of mass
228 (BaA, chrysene), 252 (BaP, BeP, BkF, perylene), and 276 (BghiP, indeno).
S
I US)
wlBD
|»!
191
IP
E £
MijU.
u
ihUllllll
T*Ul ~ Ion Spectrum: Diesel Soot on 81 Wafer
jhi.i'luiiil
ij.^liijfJ.. t - *i *
I,In,<1.1,,.
^lDD
S ®
2 03
i«i
S"
B fin
2
c
aim.
in
ROI of Large Soot Agglomerate
as
m
llllihtiJ li|l lljiiil llililiiiillliiiiJil.Jil.ijjjuulLji,j.i,liL..i.. ¦>JiJ
as uo w su a
.3D
W
w,'(D
9
^ 5dd
in
ijiiiiijJi
aWl
ROI of Background SI Wafer
Bl
Jlu
n.liU
sx
Figure 5. Mass spectrum of unspiked diesel soot on silicon wafer with ROI spectra of soot agglomerate #1 and
silicon wafer background
a) Mass 202 Ion Image
b] Mass 215 Ion Image
cj Mass 239 Ion Image
Figure 6. Secondary ion images of PAH fragments: (a) and (b) are fragments of mass 228 PAHs, and (c) is
a fragment of the mass 252 PAHs.
8
-------
TECHNICAL REPORT DATA
—
1. REPORT NO.
EPA/600/A-96/087
2 .
3.RECI
4. TITLE AND SUBTITLE
Imaging Mass Spectrometry of Particulate-Associated
Polynuclear Aromatic Hydrocarbons.
5.REPORT DATE
6.PERFORMING ORGANIZATION CODE
7. AUTHOR(S) i
Michael J. Holland, Richard W. Linton; Joette L.
Steger, Raymond G. Merrill; Greg S. Strossman,
Thomas F. Fister, and Robert W. Osom.
8.PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Chemistry, CB# 3290, University of
North Carolina, Chapel Hill, NC 27599-3290
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-D4-0022
12. SPONSORING AGENCY NAME AND ADDRESS
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13 . TYPE OF REPORT AND PERIOD COVERED
Project Report
14. SPONSORING AGENCY CODE
EPA/600/09
IS . SUPPLEMENTARY NOTES
16 . ABSTRACT
Time-of-Flight secondary ion mass spectrometry (TOF-SIMS) was employed for the in
situ investigation of PAHs on individual environmental particulates. TOF-SIMS is a
microprobe mass spectrometry technique capable of providing surface elemental and
chemical information with lateral spatial resolution <1 um. TOF-SIMS analysis
allows for the correlation of microscopic characteristics of different particulate,
filter, and sorbent substrates with adsorbed species. Silver membrane and glass
fiber filters loaded with coal flyash or silica gel were placed in a method 0010
spiking train and exposed to vapor phase benzo[a]anthracene, benzo[g,h,i]perylene or
pyxene. TOF-SIMS allowed surface chemical analysis of single particles on Ag filter
media taken directly from the Method 0010 sampling train, without additional sample
preparation. Secondary ion images indicated that benzo[a]anthracene is found
primarily on the filter media and not on the flyash surfaces, and is of interest
with respect to the elucidation of sampling artifacts. Particulate matter was also
transferred to silicon wafers where native levels of diesel soot POM were imaged
using TOF-SIMS.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/ OPEN ENDED
TERMS
C.COSATI
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This
Report)
UNCLASSIFIED
21.NO. OF PAGES
20. SECURITY CLASS (This
Page)
UNCLASSIFIED
22. PRICE
------- |