United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Las Vegas, NV 89193-3478
Research and Development
EPA/600/S4-91/032 August 1992
&EPA Project Summary
Measurement of Polycyclic
Aromatic Hydrocarbons in
Soils and Sediments by
Particle-Beam/High-Performance/
Liquid Chromatography/Mass
Spectrometry
C.M. Pace, D.A. Miller, M.R. Roby, and L.D. Betowski
A draft analytical method was devel-
oped for the measurement of certain
polycyclic aromatic hydrocarbons
(PAHs) in soils and sediments by par-
ticle-beam/liquid chromatography/mass
spectrometry. The method applies to
PAHs with a molecular weight greater
than 220. Samples are prepared by SW-
846 Method 3540 with optional cleanup
using SW-846 Method 3630. The sample
extracts are then analyzed for PAHs
using a particle-beam/liquid chromatog-
raphy/mass spectrometry system.
Method detection limits are within the
range of 0.01 to 0.10 }ig/g depending
on the sample size. Mean method ac-
curacy was greater than 75 % for most
of the target analytes with relative stan-
dard deviation values between 10% and
20%. An analysis of a standard refer-
ence material using this method agreed
with certified values and with an analy-
sis performed using high performance
liquid chromatography (HPLC) with
fluorescence detection (SW-846 Method
8310). The method shows potential as
a means to measure high molecular
weight PAHs not measured by current
EPA methods.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Las Vegas, NV, 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
The (PAHs) comprise a class of poten-
tially hazardous compounds of environ-
mental concern. The PAHs were selected
for this study as part of a continuing effort
to evaluate applications of particle beam
(PB) liquid chromatography/mass spec-
trometry (LC/MS) to the measurement of
pollutants in environmental samples. Ini-
tial studies determined instrument re-
sponse characteristics to the EPA Method
610 target analytes. The analytes com-
prise 16 PAHs ranging in molecular weight
from naphthalene (MW 128) to
dibenzo(a,h)anthracene (MW 278).
The PB LC/MS was unsuitable for the
analysis of the lower molecular weight
PAHs (MW<220). Consequently, the lower
molecular weight PAHs were dropped from
further study, and four higher molecular
weight PAHs were added as potential tar-
get analytes. The additional analytes in-
cluded three MW 302 PAHs from Appen-
dix IX (51 Federal Register 5561,
February 1986): dibenzo(a,e)pyrene,
dibenzo(a,h)pyrene, and dibenzo(a.i)-
pyrene. The fourth add-on analyte was
another MW 302 PAH isomer, dibenzo(a.l)-
pyrene.
The instrument performance character-
istics of the PB LC/MS system were in-
vestigated with respect to the target PAHs.
Specific parameters considered were chro-
matography, detection limits, precision,
response range, spectral quality, and the
ability to analyze for PAHs in "real world"
Printed on Recycled Paper
-------
United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
Las Vegas, NV 89193-3478
Research and Development
EPA/600/S4-91/032 August 1992
i&EPA Project Summary
Measurement of Polycyclic
Aromatic Hydrocarbons in
Soils and Sediments by
Particle-Beam/High-Performance/
Liquid Chromatography/Mass
Spectrometry
C.M. Pace, D.A. Miller, M.R. Roby, and L.D. Betowski
A draft analytical method was devel-
oped for the measurement of certain
polycyclic aromatic hydrocarbons
(PAHs) in soils and sediments by par-
ticle-beam/liquid chromatography/mass
spectrometry. The method applies to
PAHs with a molecular weight greater
than 220. Samples are prepared by SW-
846 Method 3540 with optional cleanup
using SW-846 Method 3630. The sample
extracts are then analyzed for PAHs
using a particle-beam/liquid chromatog-
raphy/mass spectrometry system.
Method detection limits are within the
range of 0.01 to 0.10 u.g/g depending
on the sample size. Mean method ac-
curacy was greater than 75 % for most
of the target analytes with relative stan-
dard deviation values between 10% and
20%. An analysis of a standard refer-
ence material using this method agreed
with certified values and with an analy-
sis performed using high performance
liquid chromatography (HPLC) with
fluorescence detection (SW-846 Method
8310). The method shows potential as
a means to measure high molecular
weight PAHs not measured by current
EPA methods.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Las Vegas, NV, 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
The (PAHs) comprise a class of poten-
tially hazardous compounds of environ-
mental concern. The PAHs were selected
for this study as part of a continuing effort
to evaluate applications of particle beam
(PB) liquid chromatography/mass spec-
trometry (LC/MS) to the measurement of
pollutants in environmental samples. Ini-
tial studies determined instrument re-
sponse characteristics to the EPA Method
610 target analytes. The analytes com-
prise 16 PAHs ranging in molecular weight
from naphthalene (MW 128) to
dibenzo(a,h)anthracene (MW 278).
The PB LC/MS was unsuitable for the
analysis of the lower molecular weight
PAHs (MW<220). Consequently, the lower
molecular weight PAHs were dropped from
further study, and four higher molecular
weight PAHs were added as potential tar-
get analytes. The additional analytes in-
cluded three MW 302 PAHs from Appen-
dix IX (51 Federal Register 5561,
February 1986): dibenzo(a,e)pyrene,
dibenzo(a,h)pyrene, and dibenzo(a,i)-
pyrene. The fourth add-on analyte was
another MW 302 PAH isomer, dibenzo(a,l)-
pyrene.
The instrument performance character-
istics of the PB LC/MS system were in-
vestigated with respect to the target PAHs.
Specific parameters considered were chro-
matography, detection limits, precision,
response range, spectral quality, and the
ability to analyze for PAHs in "real world"
Printed on Recycled Paper
-------
samples. Following examination of instru-
ment performance characteristics, a
method was developed for the analysis of
the target PAHs in soils and sediments.
The method utilized Soxhlet extraction and
silica gel column clean-up for sample
preparation and the PB LC/MS for mea-
surement. The overall method performance
was evaluated for spiked soil samples and
a standard reference material (SRM).
Experimental
Chromatographic separations employed
a Hewlett-Packard (HP)* 1090L liquid chro-
matograph (LC) with a filter photometric
detector and a 250-mm x 4.6-mm I.D. 5-
Hm C18 column (Vydac 201TP54). The
column was at room temperature and a
flow rate of 0.4 mL/min was used. The
mobile phase program is listed in Table 1.
The LC system was coupled to an HP
5988A mass spectrometer (MS) with the
HP 59980A PB interface. An HP 59970
MS Chemstation data system controlled
the instrument.
Two separate sample preparation
schemes were used. One procedure called
for a soil (or sediment) to be sonicated in
acetonitrile. A portion of the sonication
extract was then passed through a C-18
solid phase cartridge and subsequently
concentrated. The second procedure con-
sisted of SW-846 Method 3540 followed
by solvent exchange into cyclohexane. The
cyclohexane extract was then cleaned up
using SW-846 Method 3630 followed by
solvent exchange into acetonitrile.
Two objectives were considered for the
liquid chromatographic separation of the
target PAHs. First, a mobile phase and
column were selected to effect separation
of most target analytes in 30 to 40 min.
Second, the separation had to be compat-
ible with the PB and MS systems. For
these reasons, a ternary solvent program
was employed. Acetonitrile was selected
because it gave the best selectivity for the
later-eluting target analytes. Methanol was
selected because it gave the best PB re-
sponse to the target analytes. Tetrahydro-
furan (THF) was selected because it re-
duced retention times on the last two elut-
ing analytes and reduced the overall chro-
matographic run time by 15 min.
Results and Discussion
The PB LC/MS was unsuitable for the
analysis of the lower molecular weight
PAHs (MW < 220). Presumably, these
PAHs are too volatile to pass the PB in-
terface. Figure 1 illustrates the poor PB
response to lower molecular weight PAHs
by comparison with the UV response from
a photometric detector connected in se-
ries with the PB interface. Accordingly,
only the higher molecular weight PAHs
were studied.
Instrument Performance
Detection Limits and Precision
The estimated instrument detection lim-
its and precision of the PB LC/MS system
for those PAHs investigated in this study
are shown in Table 2. The detection limits
were determined from full-scan extracted
ion chromatograms at the 25-ng level.
Detection limit values are 3 times the stan-
dard deviation of seven replicates. The
precision values were calculated from the
same set of seven injections at 25 ng.
Considerably better detection limits can
be achieved with single-ion monitoring.
Response Curves
Instrument response to PAH standard
solutions covering a 50-fold concentration
range (20 to 1000 ng) was nonlinear for
most target PAHs (response factor RSDs
> 20 percent). Response factors tended
to increase with increasing concentration.
On one occasion, however, a six point
calibration (20 to 1000 ng) exhibited es-
sentially linear response for most target
PAHs. This occurrence was the exception
and could not be reproduced. Responses
over a smaller concentration range were
also nonlinear but gave response factor
RSDs closer to 20 percent.
Retention Times
The stability of the retention times of
the target PAHs eluting from the LC col-
umn was investigated. We observed that
the retention times were susceptible to
small changes in column temperatures
under the conditions used. Upon elevat-
ing the LC oven compartment to 37.5° C
(lowest stable temperature capable by the
system) drastic losses in chromatographic
resolution were observed. Therefore, the
analyses were carried out at ambient tem-
perature.
Spectral Quality
The PB mass spectrum of dibenzo(a,e)-
pyrene displays spectral features com-
mon to all of the PAHs studied. The mo-
lecular ion is the base peak and appears
with several (M-nH)+ ions where n can be
as many as six. Another prominent fea-
ture is the presence of doubly charged
ions that appear at a mass to charge ratio
of one half the molecular ion and (M-nH)+
ions. Spectra acquired under similar con-
ditions but at different times show varia-
tions in the doubly charged ions relative
abundance. This phenomenon appears in
all the spectra of the PAHs examined but
is more pronounced in the higher molecu-
lar weight PAHs.
Performance on Soil Extracts
Figure 2 is a total ion chromatogram
(TIC) of a PAH contaminated soil from
The Dalles, OR. The soil was extracted
using acetonitrile sonication as described
in the experimental section. A stack plot
of four selected ions is illustrated in Fig-
ure 3. Note the presence of several peaks
at mass 326. Examination of mass spec-
tra from these peaks indicates that the
components associated with the peaks are
PAHs. Table 3 lists the quantities of each
target compound found by internal (d12-
perylene) and external standard calibra-
tion techniques. Also listed for compari-
son are the quantities of target compounds
measured on a separate LC system using
fluorescence detection. Examination of
Table 3 reveals agreement between PB
quantitative results and results obtained
by fluorescence detection.
Method Performance
The existing SW-846 Soxhlet extraction
procedure (Method 3540) was incorpo-
rated into a sample preparation scheme
for the PB analysis of PAHs in soils and
sediments. Because of difficulties encoun-
tered during the initial PB analysis of an
acetonitrile extract of a Canadian SRM, a
clean-up method was sought. Initial analy-
sis of the SRM suggested interference
from hydrocarbons. For this reason, the
SW-846 silica gel clean-up (Method 3630)
was employed. To evaluate overall method
performance, several spiked clean soils
and a Canadian SRM were analyzed.
A sandy loam soil was spiked in tripli-
cate at two different levels, 0.5 u,g/g and
2.5 u,g/g. The samples were prepared as
just described and the extracts were ana-
Table 1. Liquid Chromatographic Mobile Phase Program
Time (min) % Methanol % Acetonitirile
% Tetrahydrofuran
' Mention of trade names or commercial products
does not constitute endorsement or recommendation
for use.
0
2
10
15
95
95
45
45
0
0
45
25
5
5
10
30
-------
100 ng each
UV Chromatogram of EPA Method 610 PAHs (254nm).
18000
16000 -
14000
12000 *
8 10000 "
8000 1
6000 •
4000 -
2000 -
25
Time (min.)
Particle Beam TIC of EPA Method 610 PAHs
Figure 1. Comparative chromatograms of 16 PAHs by HPLC/UV and PB LC/MS
-------
Table 2. Detection Limits and Precision of the PB LC/MS for the Analysis of PAHs
Compound
benzo(a)anthracene
chrysene
benzo(b)fluoranthene
benzo(k)fluoranthene
benzo(a)pyrene
dibenzo(a, l)pyrene
dibenzo(a, h)anthracene
benzo(g, h,i)perylene
indeno(1,2, 3-c, d)pyrene
dibenzo(a, ejpyrene
dibenzo(a, i)pyrene
dibenzo(a, h)pyrene
Quantitation
ion
228
228
252
252
252
302
278
276
276
302
302
302
Detection
limit (ng)
1.8
3.0
1.6
1.0
2.2
6.1
2.4
2.4
1.5
2.5
3.0
4.8
Precision
RSD(%)
2.4
4.1
2.1
1.4
2.9
8.1
3.1
3.2
2.0
3.4
4.0
6.3
lyzed with the PB instrument. Recoveries
were calculated by using integrated
quantitation ion abundances and a six point
external calibration. Results from one of
the low-level spikes (0.5 jig/g) were dis-
carded. Preparation of this particular spike
resulted in a two-phase extract. The two-
phase extract was probably the result of
incomplete solvent exchange. The data
from all three high-level spikes (2.5 M.g/g)
were used to determine mean recovery
and standard deviation although two of
the higher level spikes gave significantly
lower recoveries.
HPLC/UV examination of the pentane
wash from the silica gel clean-up from
one of the low recovery samples revealed
5% to 15% of the spiked amount for most
of the target analytes had washed off the
column prior to elution of the analytical
fraction. This loss may have resulted from
improper preparation of the silica column
or from nonuniform activation of the silica
gel. These results indicate the silica gel
clean-up is an area of concern and a
potential source of problems for overall
method performance. However, losses to
the column wash do not account for the
low recoveries observed for dibenzo(a,h)-
pyrene, as this target analyte was not
found in the pentane wash. This PAH was
probably not extracted efficiently with the
solvent system employed.
The recovery data were pooled and
treated as a single data set to generate
overall method precision and accuracy
values. These values are listed in Table 4
along with estimated method detection lim-
its. Method detection limits were estimated
from observed instrument detection limits.
Values were adjusted for concentration/
dilution factors imposed by the sample
preparation scheme: a 20 ul injection, a
1-mL final extract volume, and a 10 g
sample size. Final values were corrected
with the observed recoveries. The method
detection limits are estimates and have
not been experimentally verified.
Analysis of a Standard
Reference Material
An SRM was analyzed using the proce-
dures described in this report to evaluate
method performance on "real world"
samples. The SRM was a marine sedi-
ment obtained from the National Research
Council of Canada and designated as
HS-3. The material was prepared in tripli-
cate (5 g each) and taken through the
silica gel clean-up procedure. Target
i
•8
§
•Q
1.4E+5 -
1.2E+5 -
1.0E+5 -
8.0E+4 -
6.0E+4 -
4.0E+4 -
2.0E+4 -
O.OE+0
Figure 2. Panicle beam TIC of a PAH contaminated soil.
column: Sum Vydac 201TP 4.6mm x 25cm
flow: 0.4 ml/min
mobile phase:
15 20
Time (min.)
t(min)
0
2
10
18
MEOH(%)
95
95
40
0
CH3CN(%) THF(%)
0
0
55
70
5
5
5
30
-------
0)
1
•§
c
3
"^
40000 -
30000 -
20000-
10000-
0-
.
M/Z252 ii
l\l\
III A A
5 10 15 20 25
1 1 — 1 — r~l
30
Time (min.)
1
"§
1
40000 -
30000 -
20000 -
10000 -
0-
M/Z278 1
/
A A / I
~ _y^^\»^^>^/V y V>^ / V^-X>^^N. Jv / V
I I II i^r^»y r^ >' r ^i^f^nj i^i1^!*" r N 'i > i i • i . — _- . — .
10
15 . 20 25
Time (mm.)
30
27691
M/Z302
i i i i i *i i 'i r i in
\Jk
10
400-
| 300:
•8 200-
§ yoo-
M/Z326
T~
10
75 20
Time (min.)
JL
fV I l i
25 30
I I I I I I I I I I I
15 20 25
Time (min.)
Figure 3. Selected ion chromatograms of a PAH contaminated soil.
analyte amounts were obtained by inte-
grated quantitation ion areas and external
calibration. In addition, the extracts were
analyzed by HPLC with UV diode array
detection for comparative purposes. The
results are listed in Table 5 along with the
certified SRM values and the initial PB
results on an acetonitrile extract without
clean-up.
The PB results with extract clean-up
failed to meet acceptance criteria (p ± 2s)
for only one analyte, benzo(k)fluoranthene.
Values for p and s were taken from the
experimentally determined method perfor-
mance parameters listed in Table 4. The
HPLC/UV analysis failed acceptance cri-
teria for two of the target analytes. The
PB results on the acetonitrile extract with-
out clean-up failed acceptance criteria for
all target analytes. In this particular in-
stance, extract clean-up appears to be
essential for accurate analysis. In gen-
eral, results obtained from PB analysis
with extract clean-up and HPLC/UV were
in agreement and agreed with certified
values.
Conclusions and
Recommendations
Low molecular weight PAHs (MW<220)
cannot be measured accurately with the
PB instrument. However, PAHs with mo-
lecular weight greater than 220 can be
measured with good accuracy and preci-
sion. The instrument sensitivity to these
PAHs was on the order of 1 to 10 ng in
the full scan mode. Such sensitivity allows
method detection limits comparable to or
better than those of current GC/MS-based
EPA methods.
Table 3. Comparison ofPB LC/MS Quantification Method vs. Flourescence for PAH Target Analytes
Soil Extract (\ig/g)
RT (mm)
10.94
11.75
13.43
14.69
15.75
-
17.67
18.84
19.96
20.77
-
-
m/z
228
228
252
252
252
302
278
276
276
302
302
302
Compound
benzo(a)an thracene
chrysene
benzo(b)fluoranthene
benzo(k)fluoranthene
benzo(a)pyrene
dibenzo(a, l)pyrene
dibenzofa, h)anthracene
benzo(g,h,i)perylene
indeno( 1 , 2, 3-c, d)pyrene
dibenzo(a, e)pyrene
dibenzo(a, i)pyrene
dibenzo(a, h)pyrene
IS'
6.5
26
19
5.9
6.2
-
0.8
3.7
5.6
0.9
-
-
FL»
6.4
21
16
6.8
8.8
-
-
6.4
5.2
-
-
-
EXC
5.4
19
18
7.0
6.8
—
1.1
4.8
4.5
1.0
-
-
quantitated by d12-perylene internal standard
quantitated by fluorescence detection
quantitated by external standards
-------
Instrument response to PAH standard
solutions covering a 50-fold concentration
range (20 to 1000 ng) was nonlinear for
most target PAHs (response factor RSDs
> 20 %). Nonlinear response did not ap-
pear to present particular difficulties, how-
ever, provided the response was correctly
modeled (i.e., point-to-point calibration or
polynomial curve fits). The nonlinear re-
sponse was reproducible over the course
of an analytical run (24 h) and in calibra-
tion check samples gave values within
20% of initial calibration. Further, the non-
linear PB calibration gave results in agree-
ment with HPLC/UV and HPLC/fluores-
cence analysis of "real world" samples.
The electron ionization (El) mass spec-
tra obtained from each of the target
analytes were consistent with structure and
comparable to reference spectra. In gen-
eral, the spectra obtained from "real"
samples were of sufficient quality to allow
tentative identification of nontarget PAHs.
However, some spectral variation was ob-
served that did not correspond to differ-
ences in tuning and mass calibration.
These variations take the form of enhanced
relative abundance of the doubly charged
molecular ion.
One of the potential applications of PB
LC/MS emerging from these studies is the
measurement of high-mass PAHs
(MW>300). Current EPA methods do not
Table 4. Method Detection Limits, Precision, and Accuracy
Compound
benzo(a)anthracene
chrysene
benzo(b)fluoranthene
benzo(k)fluoranthene
benzo(a)pyrene
dibenzo(a, l)pyrene
dibenzo(a, h)anthracene
benzo(g, h, i)perylene
indeno(1,2, 3-c,d)pyrene
dibenzo(a, e)pyrene
dibenzo(a, i)pyrene
dibenzo(a, h)pyrene
MDL
fa9/9)
0.02
0.03
0.02
0.01
0.04
0.14
0.02
0.03
0.02
0.03
0.04
0.11
Mean method
Accuracy (n=5)
(% of true value)
89
112
77
95
61
42
100
80
82
77
81
45
Standard
Deviation (%)
20
23
14
19
12
9
25
18
18
12
16
16
measure for PAHs above mass 300.
Analysis of the Canadian SRM and the
PAH-contaminated soil (from The Dalles,
OR) by PB LC/MS revealed the presence
of eight mass 302 PAHs and five mass
326 PAHs. Evidence for PAHs above
mass 326 was also obtained. These high-
mass PAHs were only present at low con-
centrations. However, the low amount ob-
served was probably due, in part, to poor
extraction efficiency with the solvents em-
ployed (methylene chloride or acetonitrile).
We recommend that work on the appli-
cation of PB LC/MS for the measurement
of high-mass PAHs be pursued. This work
would entail characterizing a PAH-con-
taminated sample for high-mass PAHs.
The work would involve investigation of
suitable extraction solvents, chromato-
graphic separation of the high-mass frac-
tion, and the identification and quantita-
tive estimation of high-mass PAHs by PB
LC/MS in combination with stop-flow fluo-
rescence spectroscopy.
Table 5. Results of SRM Analysis
Compound
benzo(a)anthracene
chrysene
benzo(b)fluoranthene
benzo(k)fluoranthene
benzo(a)pyrene
dibenzo(a, l)pyrene
dibenzo(a, h)anthracene
benzo(g, h,i)perylene
indeno(1,2, 3-c, d)pyrene
dibenzofa, e)pyrene
dibenzo(a, i)pyrene
dibenzo(a, h)pyrene
Certified
Value (\ig/g)
14.6 ±2.0
14.1 ±2.0
7.7 ± 1.2
2.8 ±2.0
7.4 ±3.6
NA
1.3±0.5
5.0 ±2.0
5.4+ 1.3
NA
NA
NA
HPLC/UV
fa9/9)
15.2± 1.5
7.0 ±0.6
4.8 ±0.6
4.8 ±0.5
4.3 ±0.5
NF
0.8 ±0.2
4.1 ±0.6
3.6 ±0.6
1.2 ±0.2
2.0 ±0.4
NF
PB with Cleanup
fag/g)
12. 1±1.1
19.4 + 2.6
4.4±0.5
5. 1 ± 0.4
3.9 ±0.4
NF
1.7 ±0.4
3.7+0.4
3.6 ±0.4
0.7 ±0.1
0.3 ±0.03
0.2 ±0.06
PB without
Clean-up
fa9/9)
5.1
3.7
2.5
1.2
1.4
NF
NF
0.8
0.8
NF
NF
NF
NA = certification not available
NF = not found
PB = particle beam
•U.S. Government Printing Office: 1992— 648-08CV60055
------- |