Upload
independent
View
0
Download
0
Embed Size (px)
Citation preview
Pergamon 0045-6535(95)00275-S
Chemosphere, Vol. 31, No. 9, pp. 414H155, 1995 Copyright Q 1995 Elsevier Science Ltd
Printed in Great Britain. All rights reserved
004%6535195 $9X+0.00
METHODOLOGY FOR THE DETERMINATION OF PRIORITY POLLUTANT
POLYCYCLIC AROMATIC HYDROCARBONS IN MARINE SEDIMENTS
Christopher D. Simpsona, William R. Cullena, Kristine B. Quinlana, Kenneth J. Reimerb
a Chemistry Department, University of British Columbia, 2036 Main Ma& Vancouver, BC, Canada, V6T 1Zl
b Environmental Sciences Group, Royal Roads Military College, PM0 Victoria, BC, Canada, VOS 1BO
(Received in USA 10 February 1995; accepted 6 September 1995)
ABSTRACT
The development and application of a simple method for the isomer specific determination of
polycyclic aromatic hydrocarbons (PAHs) in marine sediments is described. The method involves
Soxhlet extraction of the sample with methylene chloride, a one-step cleanup on a Plorisil column,
and analysis by using capilliary gas chromatography with flame ionisation detection (GC-HD).
Using this method, quantitative recoveries from spiked sediment samples were obtained for all of
the 16 USEPA priority pollutant PAHs, with the exception of naphthalene. This method allows for
substantial time and cost savings when compared to current isomer specific analytical
methodologies, and will find application where availability of resources and instrumentation is
limited.
INTRODUCTION
Polycyclic aromatic hydrocarbons (PAUs) are important environmental contaminants because
many of these compounds are potential or proven carcinogens [l-3]. Unfortunately their low
aqueous solubility, limited volatility, and recalcitrance towards degradation allow PAHs to
accumulate to levels at which they may exert toxic effects upon the environment [4-61. The United
4143
4144
States Environmental Protection Agency (USEPA) has included 16 unsubstituted PAHs in their
list of 129 priority pollutants [7], including the fourteen individual compounds listed in Tables 1
and 2, and two benzofluoranthene isomers (benzo(b)fluoranthene, benzo(k)fluoranthene).
PAHs are present in significant amounts in coals and oils, and they are formed as by-products from
many combustion sources [8-l 11. Thus, PAH contamination is widespread and much attention is
being focused on the analysis of PAHs in many environmental compartments. A number of
analytical procedures involving multiple-stage extract cleanup and GCMS analysis have been
reported that provide low detection limits and good reproducibility [ 12-151, however these
methods are expensive and time consuming. Simpler, more cost effective methodologies are
required by industry to ensure their compliance with permissable discharge regulations, and by
government or academic agencies concerned with pollution monitoring and effects on the
environment.
For many sites where contaminant levels are high, low detection limits are not required and
features such as high sample throughput and reduced analysis cost are more important. We report
here a simple cost efficient procedure for the analysis of PAHs in sediment samples. The
procedure will be suitable for laboartories with limited resources, or who do not have access to
GCMS facilities. High sample throughput and reduced analysis time is achieved by the use of a
single-stage extract cleanup step, while the use of GC-FIJI rather than GCMS allows significant
cost savings on instrumentation and expensive isotopically labelled standards.
EXPERIMENTAL
SAMPLE PREPARATION
Surface sediment samples were collected from several contaminated and pristine sites by using
Smith-McIntyre or petite ponar grab samplers. The samples were mixed thoroughly in the grab
sampler, stored in 1L pre-cleaned and oven baked amber glass bottles and frozen at -20°C for
transport back to our laboratory.
The sediments were freeze dried, and after removing any stones and pebbles with metal tweezers,
the resulting powder was ground and re-mixed using a mortar and pestle to ensure sample
4145
homogeneity. Sub-samples (ca. lo-20g) were then weighed out into glass extraction thimbles,
covered with ca. 2g anhydrous sodium sulfate (BDH assurance grade; pre-cleaned by baking at
600°C for 6 hr), and Soxhlet extracted with methylene chloride (130 mL) for 8 hours. Prior to
extraction a quantification standard (QSTD), hexamethylbenzene (Eastman-Kodak), was spiked
onto the sample. The extract, in a round bottom flask, was then concentrated under reduced
pressure (300 mtorr) to ca. 0.5 mL in a rotary evaporator equipped with a dry ice/acetone
condenser. The water bath was maintained at 3O’C.
In a modification of the procedure used by Maher and Brown [16], sodium sulfate (2g) was
poured onto the concentrated extract in a round bottom flask. The extract was adsorbed onto the
sodium sulfate and the solvent was removed under a gentle stream of nitrogen. This step is
necessary to completely remove the methylene chloride, as this solvent does not allow an adequate
separation of the aliphatic fraction from the aromatic fraction on the Florisil columns. Evaporation
of the methylene chloride without the presence of an adsorbent results in substantial and
uncontrolled losses of the more volatile PAH. The adsorbed extract was added to a 1.2 cm
diameter glass chromatography column that had been slurry packed with 3 g Florisil in hexane
(Fisher Scientific; 60-100 mesh, activated at 350°C for six hr, 2% deactivated with distilled
water). Two 1 mL aliquots of hexane were used to rinse the round bottom flask, and these rinsings
were used to wash the extract onto the Florisil. This allows quantitative transfer of the extract
onto the column. It also ensures that the PAHs will be adsorbed in a narrow band beneath the
sodium sulfate on top of the Florisil bed. The PAHs were sequentially eluted with 15 mL of
hexane and 25 mL of a 1:l (v:v) hexane:methylene chloride mixture. The first 8 mL of eluent
which contained a large amount of the aliphatic material was discarded. If the extract is not
washed onto the Florisil with the rinsings from the round bottom flask, the PAHs desorb from the
sodium sulfate into the eluent reservoir when the eluent is added, and elute in a broad band which
would not be resolved from the hydrocarbon fraction. Careful separation of the bulk of the
aliphatic material from the analytes using this procedure is critical if a non-selective GC detector
such as an FID is to be used for quantification.
The eluent containing the analytes was then concentrated under reduced pressure (ca. 95 mtorr) to
ca. 0.5 mL on the rotary evaporator. Toluene (2 mL) was added as keeper, and the remainder of
4146
the hexane and methylene chloride was evaporated. The final extract was pipetted into a glass vial;
the round bottom flask was rinsed twice with 0.5 mL washings of toluene; the rinsings were added
to the sample vial, and a recovery standard (RSTD), octacosane (Eastman-Kodak), was added to
the vial. All solvents used were HPLC grade (Fisher Scientific).
Each sample was injected splitless into a Hewlett Packard 5890 gas chromatograph (GC)
equipped with a flame ionization detector (FID). The injector was held at 300°C and purged with
He after 1 minute to minimize tailing of the solvent peak. The analysis was carried out by using a
30m FSOT PTE-5 capillary GC column (Supelco Ltd.; 0.32 mm internal diameter, 0.25 pm
stationary phase), He carrier gas (45 cm.sec-l flow rate), and the oven was programmed as
follows: initial temperature 100°C for 4 min.; 5”CXnin. to 2OO”C, held for 3 min.; 8Wmin. to 250
‘C, held for 5 min.; 12°C/min. to 3OO”C, held for 20 min. The detector was held at 320°C. The
data were collected and analyzed by using Shimadzu EZChromTM software (version 3.1), running
on a DellTM 466/m computer. Compound identification on the basis of EI mass spectra was
undertaken for selected samples by using a Varian Vista 6000 GC interfaced to a Nermag RlO-
1OC quadrupole mass spectrometer operated in the EI mode (70 eV). The transfer line between
the GC and the mass spectrometer was held at 300°C. The GC column and temperature program
used were identical with that used for the FID analyses. Because of poor sensitivity and drifting
instrumental response, the data from the GCMS was not sufficiently reproducible for quantitative
analysis.
A response factor mix containing known amounts of the quantification standard, the recovery
standard and the 16 PAHs (Supelco Ltd.) was made fresh each day from stock solutions, and
analyzed at least three times with every batch of samples to determine retention times and
response factors for the various analytes. Individual PAHs were identified on the basis of retention
time matches with the standard mixture.
The efficiency with which the QSTD is recovered in each sample can be calculated by comparing
the expected concentration of the QSTD, with the concentration of the QSTD calculated relative
to the RSTD. Recovery values for the QSTD were typically greater than 90%. The extraction
4147
efficiencies for the 16 PAHs were determined by spiking a standard solution of PAHs in toluene
onto the surface of a freeze dried marine sediment sample known to be uncontaminated with
PAHs [17]. The final concentration was 2 ug of each PAH per g dry sediment. Quantitative
recoveries were obtained for all PAHs including the QSTD, with the exception of naphthalene for
which the recovery was typically ca. 30%. and no increase in recovery was observed when the
extraction time was increased to 16 hours. To account for the loss of naphthalene, a correction
factor was applied when the naphthalene concentration was calculated. The incomplete recovery
of naphthalene was attributed to evaporative losses. These were minimized by carefully controlling
the adsorption of the extract onto sodium sulfate (i.e. do not overdry), and maintaining a slow
boiling rate in the rotary evaporator to minimize sample loss due to aerosol formation.
It should be noted that recoveries from a spiked sample may be substantially higher than those
from a real sample because PAHs in contaminated sediments are often tightly bound or occluded
within particles. This point was addressed by Burford et al. in a recent paper [ 181.
The limit of detection was defined for each sample as that concentration of analyte which would
give rise to the minimum peak area which could be reliably distinguished from the background by
the integration software used. This number was typically 5-200 ng.g-‘, and varied from sample to
sample as a function of the total amount of extractable material present in each sample. Trace
amounts of naphthalene (typically ca. 0.4 ng) were detected in blank extracts, but none of the
other PAHs were ever detected in blank extracts.
RESULTS AND DISCUSSION
PLORISIL COLUMN CHROMATQGRAPHY
Silica or alumina are routinely used to remove material from sediment extracts that would
otherwise interfere with analysis for PAHs, and to effect a crude separation of analytes into
compound classes. However, these sorbents have been shown to catalyze photo-oxidation of
PAHs [ 19-201, resulting in decreased recoveries. An alternative sorbent, Plorisil, has been used by
4148
that 2% deactivated Florisil may provide a better separation of the aliphatic interferences from the
PAHs. For this reason, 2% deactivated Florisil was used for the rest of this work.
ANALYSIS OF A STANDARD REFERENCE MATERIAL
TABLE 1: PAH determinations for NRCC Standard reference material HS3 (concentrations in
pg.g-’ dry sediment)
anthracene
fluoranthene
14.1 f 2.0
benzo(a)anthracene 14.6 f 2.0
benzofluoranthenes 15.1 f 1.9
benzo(a)pyrene
indeno(l,2,3-cd)pyre
dibenz(aJQanthracene
benzo(g$,i)petylene 4.8 f 0.8
.LL-_..~-~~-_-. “I ___r>.__. ._.__ _* ~DOre”liiu”“S: u = conrmence mtervill
a two independent determinations were carried out by the commercial laboratory
b sum of benzo(b)fluoranthene and benzo(k)fluomnthene only
A marine sediment standard reference material (Harbor sediment HS3, NRCC, Canada) was
analyzed in triplicate to validate the analytical procedure described in this paper. Samples of the
reference material were also analyzed by a local commercial laboratory by using USEPA protocols
and isotope dilution GCMS analysis to provide further validation [12]. The results, presented in
Table 1, ate generally in good agreement with the certified values.
4149
some workers for cleanup of PAH contaminated extracts [21-231. Florisil has the advantage of
providing good PAH separations on short columns [ 11, however Altgelt and Gouw [24] noted that
it strongly chemisorbed many aromatic hydrocarbons. This may be overcome by partially
deactivating the Florisil with water to selectively block highly active binding sites and thus
improve chromatographic performance [25]. Our data, presented below, show that very efficient
chromatography of PAHs is possible if the Florisil is adequately pretreated. In addition, no
significant photodegradation of the PAHs occurred under standard laboratory lighting conditions
(combination of sunlight and fluorescent tubes).
The effect on the PAH elution profile of varying the level of Florisil deactivation was investigated.
Eluent from the 3 g Florisil columns was collected in 2 mL fractions, blown to dryness under
nitrogen, taken up into 4 mL of acetonitrile, and analyzed for total PAHs by using UV absorbance
at 254 nm. For these experiments, the first 8 mL of eluent was not discarded. These results are
presented in Figure 1. Each curve represents the mean of three independent determinations.
m 1.0
2 1
0.8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Fraction #
FIGURE 1: Effect of Florisil deactivation on PAH elution
There was little difference in the PAH elution profile at 0% and 2% deactivation. However, at 5%
deactivation the PAH containing band was substantially broadened and there was poor separation
between the PAHs and the aliphatic compounds, most of which elute within the first 6 mL. This
trend was extended for 10% deactivated Florisil (data not shown). Although both the 0% and the
2% showed essentially the same elution profiles for the PAHs, the work of Dunn [25] suggested
4150
The data for the benzofluoranthenes are not directly comparable as the reference material was
certified for the benzo(b)fluoranthene and benzo(k)fluoranthene isomers only, whereas both we
and the commercial laboratory quoted results for the sum of three partially resolved
benzofluoranthene isomers. Complete resolution of the benzofluoranthenes is not routinely
achieved on traditional capillary GC bonded phases. Values determined in this work for anthracene
and fluorene were lower than those certified, as were the results reported by the commercial
laboratory. It is unlikely that this represents a systematic error in our methodology, as the
commercial laboratory, which also obtained low values for these compounds, routinely has no
difficulty in obtaining the certified values for anthracene and fluotene in other certified reference
materials. Therefore, the problem may be specific to this batch of sediment, or may reflect use of
HPLC in determining the certified values.
It should be noted that the 90% confidence intervals for our data in Table 1 are a measure of the
error within a single batch of samples, analyzed on the same day. Variation in instrument
performance and the laboratory environmental conditions from one day to the next contribute to a
between-batch error that is substantially higher than the values quoted in Table 1. Typically, the
between-batch error, which is a more realiitic estimate of the total error in this methodology, is
flO-20% for each analyte.
ANALYSIS OF CONTAMINATED SEDIMENTS
Four sediment samples representing varying levels and sources of PAH contamination were
analyzed by using the methods described in this paper. The results are presented in Table 2. HS is
a soil sample collected from a marsh that was subject to runoff from streets and nearby light
industrial installations. It is dominated by PAHs of intermediate molecular weight. F4 is a sediment
sample collected from a coastal Canadian fjord which is geographically isolated from major
industrial pollution sources. LS(B) is a sediment sample from the effluent pond at an ahrminum
smelter, and CD0 is a marine sediment collected near an effluent discharge from the same
aluminum smelter. The PAH levels in CD0 are substantially lower than those in LS(B), largely as a
result of dilution of the effluent stream in the harbor.
4151
TABLE 2: Concentrations of parent PAHs in several environmental samples (pLg.g~’ dry
sediment)
0.7 (0.6) 3
fluoranthene 4.1 (1.1) 18
3.7 (0.8) 16
1.9 (0.8) 8
benzo(a)anthracene 2.8 (1.2) 12
benzofluoranthenes 5.1 (2.2) 22
2.3 (1.7) 10
indeno(k2.3~cd)pyrene <O.l (-) -
dibenz(a&wbracene <0.19(-) -
standard deviation determined from three independent analyses concentration as a percent of the sum of the 15 compounds listed
The different mixtures of PAHs present in a particular sample reflects, to some extent, the diverse
sources that produced the PAJIs [8-9, 26-281. For example, the temperature at which PAHs are
produced has been shown to have a greater influence on the composition of the PAHs than does
the feedstock from which they are formed [8-91. Specifically, low temperature sources will
produce almost exclusively alkyl substituted PAHs, and when several isomeric PAHs are possible,
there is a marked predominance of the isomers with higher thermodynamic stability. This is the
case with PAHs of petrogenic origin. In contrast, PAHs formed at higher temperatures show
almost no alkyl substitution, increased levels of high molecular weight species, and the dominant
isomers are those that form most rapidly (the kinetically favored isomers). These are not
necessarily the isomers with the highest thermodynamic stability. Certain PAHs are considered to
4152
be characteristic of particular sources. For example, retene, a breakdown product of abietic acid, is
a biomarker for diagenesis of coniferous material [9], and cyclopenta(cd)pyrene is a marker for
vehicular combustion [29]. Perylene has been suggested as general indicator of plant diagenesis,
but its presence may be ambiguous as it may also be a minor component in combustion sources [9-
101.
Therefore, sample LS(B) which exhibits a predominance of high molecular weight PAHs and
kinetically favored isomers (fluoranthene, benzofluoranthenes, indeno( 1,2,3-cd)pyrene,
benzo(a)pyrene) is typical of PAH contamination originating from high temperature industrial
sources such as smelters [28,30]. The amounts of the thermodynamically favored isomers relative
to their kinetically favored counterparts are higher in CD0 than in LS(B) (i.e. the thermodynamic
to kinetic isomer ratio has increased). For example, the phenanthrenelanthracene ratio has
increased from 2.7 in LS(B) to 3.8 in CDO, and the benzo(g,h,i)peryleneIindeno(l,2,3-cd)pyrene
ratio has increased from 0.63 to 1.1. This may indicate that degradation of the less stable kinetic
isomers occurs more readiiy in the receiving environment than does breakdown of the
thermodynamically favored isomers, leading to a systematic change in the PAH composition.
The PAH composition in F4 differs substantially from LS(B) and CDO, most notably in the low
levels of the benzofluoranthenes. The PAHs in this sample probably represent a combination of
deposition of atmospheric PAHs, most of which arise from anthropogenic combustion sources
[28] and natural PAHs from plant biosynthesis and diagenesis e.g. (tetene, perylene).
Unfortunately, as the scope of this work was limited to the 16 USEPA PAHs, retene and perylene
were not analyzed for in this sample.
Sample HS contained lower proportions of high molecular weight PAH than did the other
samples. In addition, a large unresolved hump was present in the chromatogram of this sample.
The latter feature is characteristic of weathered bitumens [31], however the
phenanthrenc/anthracene ratio (1.0) is much higher than would be expected from a purely
petrogenic source. Therefore we conclude the PAH suite in this sample probably represents a
comnination of inputs from urban runoff and combustion particulates.
4153
SUMMARY
The analytical procedure described in this paper has the dual advantages of lower analysis cost and
shorter analysis time over the USEPA-type protocols in general use. The method provides
compound specific, quantitative data for unsubstituted PAHs for which authentic standards are
available, with acceptable precision and limits of detection (ca. 5-200 ng.g-‘). The use of Florisil
column chromatography to remove interfering polar material was validated. In contrast to alumina
and silica, no photodegradation of the PAHs on Florisil was observed. We chose to use an FID for
quantification in order to demonstrate the application of this method on routinely available
instrumentation. The author’s recognize that, if required, better detection limits could be obtained
by using a photoionization detector (PID) or a mass selective detector (MSD). Some non-PAH
co-extractives may interfere with quantification by using GC-FID, so GCMS should be used,
where possible, for confirmation of peak identity. Alternatively, samples should be analyzed
qualitatively on a second capillary column to provide a second retention time match to aid
compound identification.
The methodology was successfully applied to the analysis of PAHs in a variety of terrestrial and
marine sediments. Different PAH mixtures were found to be present in each sample, dependent
upon the source of PAH contamination, and the degree of weathering of the sample. Work is on-
going to adapt this method to the analysis of a&y1 and hetero-atom substituted PAHs which, like
the parent compounds, are of environmental concern.
REFERENCES
[ 11 Lee ML, Novotny MV, Bartle KD (1981) Analytical Chemistry of polycyclic aromatic
compounds. Academic press, Columbus Ohio
[2] Heidelberger C (1976) In: Freudenthal R, Jones PW (eds.) Carcinogenesis - A
comprehensive survey. Vol. 1. Polynuclear aromatic hydrocarbons: chemistry, metabolism
and carcinogenesis. Raven Press, New York
[3] Stegeman JJ, Lech JJ (1991) Environ Health perspect 90:101-109
[4] Onuska FI (1989) In: Afghan BK, Chau, ASY (eds.) Analysis of trace organics in the
aquatic environment. CRC Press Inc., Boca Raton Florida
4154
[5] Neff JM (1985) In: Rand GM, Petrocelli SR (eds.) Fundamentals of aquatic toxicology.
McGraw Hill, Washington
[6] Mix MC (1984) Rev Environ Toxic01 50:51-102
[7] Keith LH, Telliard WA (1979) Environ Sci Technoll3,4:416-423
[8] Bjorseth A, Ramdahl (eds.) (1983) In: Handbook of polycyclic aromatic hydrocarbons:
Volume 2: Emission sources and recent progress in analytical chemistry. Marcel Dekker,
New York
[9] Laflamme RE, Hites RA (1978) Geochim Cosmochim Acta 42:289-303
[lo] Venkatesan MI (1988) Mar Chem 251-27
[l l] Budxinski H, Garrigues P, Radke M, Connan J, Gudin J-L (1993) Grg Geochem 20,7:917-
926
[ 121 USEPA (1979) Federal Register 44, #233
[13] NOAA (1993) Sampling and analytical methods of the national status and trends program.:
Volume IV: Comprehensive descriptions of trace organic analytical methods. NOAA
technical memorandum NOS ORCA 7 1
[ 141 Wise SA, Sander LC, May WE (1993) J Chromatogr 642:329-349
[ 151 Cretney JR, Gee HK, Wright GJ, Swallow WH, Taylor MC (1985) Environ Sci Technol 19,
5:397-404
1161 Maher W, Brown G (1992) Org Geochem 18,5:657-668
[17] Simpson CD, Li X-F, Cullen WR, Reimer KJ (1994) paper #239 77th CSC conference.
Winnipeg, Canada
1181 Burfotd MD, Hawthorne SB, Miller DJ (1993) Anal Chem 65: 1497-1505
[ 191 Inscoe MN (1964) Anal Chem 36,13:2502-2506
[20] Katz M, Chan C, Tosine H, Sakuma T (19790 In: Jones P, Leber P (eds.) Polynuclear
aromatic hydrocarbons: Third international symposium on chemistry and biology
carcinogenesis and mutagenesis. Ann Arbor science publishers ltd, Michigan
[21] Bagg J, Smith JD (1987) In: Cooke M, Loening K, Merrit J (eds.) Polynuclear aromatic
hydrocarbons - measurements, means and metabolism. Batelle Press, Columbus Richland
[22] Howard JW, Teague RT, White RI-I, (1966) J Assoc Off Anal Chem 49,3:595-611
[23] Foster GD, Wright GA (1988) Mar Pollut Bull 19,9:459-465
4155
WI Altgelt KH, Gouw TH, (1975) In: Giddings JC, Grushka E, Keller RA, Cares J (eds.)
Advances in Chromatography: Vol. 13 Marcel Dekker Inc. New York
[25] Dunn B (1976) Environ Sci Technol 10. 10:1018-1021
[26] Yunker MB, McDonald RW, Cretney WJ, Fowler BR, McLaughlin FA (1993) Geochim
Cosmochim Acta 57:3041-3061
[27] Lipiatou E, SaJiot A (1991) Mar Chem 3251-71
[28] Sporstol S, Gjos N, Lichtenthaler RG, Gustavsen KO. Urdul K, Oreld F, Skel J (1983)
Environ Sci Technol 17:282-286
[29] Canton L, Grimalt 0 (1992) J Chromatogr 607:279-286
[30] Thrane KE (1987) Water Air Soil Pollut 33:385-393
[31] Volkmann JK, O’Leary TO, Summons RE, Bendall, MR (1992) Grg Geochem l&669-682