Errors in alkylated polycyclic aromatic hydrocarbon and sulfurheterocycle concentrations caused by currently employedstandardized methods
Nicholas M. Wilton a, Stephen A. Wise b, Albert Robbat Jr. a, *
a Tufts University, Department of Chemistry, 62 Talbot Ave, Medford, MA, 02155, United Statesb National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, MD, 20899, United States
Alkylated polycyclic aromatic hydrocarbon (PAH) and polycyclic aromatic sulfur heterocycle (PASH)standardized methods often rely on gas chromatography/mass spectrometry operated in the selected ionmonitoring mode (GC/MS-SIM). The objective of this study is to develop a method that produces accuratedata while minimizing sample preparation and achieving low levels of detection. Most standardizedmethods are based on acquiring a given homologue's molecular ion (1-ion). Some methods include asecond, confirming ion (2-ion) in the hopes of excluding non-target ion signals from the total homologuepeak area. Although all methods use homologue-specific retention windows, these windows differgreatly among the methods. In this paper we evaluate, for the first time, errors in quantitation caused byusing different windows as well as common ion effects when target and/or matrix compounds coelute.Two NIST-certified Standard Reference Materials (SRMs), crude oil SRM 1582 and marine sediment SRM1941b, were analyzed by five standardized methods and by the new method we developed, which relieson spectral deconvolution of three to five ions per PAH/PASH and as many fragmentation patterns asneeded to correctly identify the C1 to C4 homologues (MFPPH). All of the standardized methods over-estimated the concentrations of the majority of alkylated homologues whereas MFPPH obtained valuesmuch closer to NIST-certified concentrations. Rather than straight-line integration of all peaks in theretention window or recognizing peak patterns, the MFPPH data analysis software integrates only thosepeaks that meet the compound identity criteria.
Polycyclic aromatic hydrocarbons (PAH) and polycyclic aromaticsulfur heterocycles (PASH) are the most often analyzed pollutantsin petroleum [1], coal tar [2], creosote [3], and associated wasteproducts [4]. Much attention is given to PAH and PASH because theyare toxic [5,6] and known carcinogens [7,8], mutagens [9e11], andteratogens [12]. Scientists, engineers, and regulators worldwiderely on selective and accurate quantitation of PAH/PASH to supporthazardous waste site investigations and environmental forensics[13,14], weathering [15,16] and toxicity studies [17,18].
To obtain defensible data, numerous governmental and non-governmental organizations have published standardizedmethods. For instance, the U.S. Environmental Protection Agency(EPA) promotes SW846 methods for the detection of PAH in soil/sediment [19,20], air [21], and water [22]. Of particular interest isthe EPA's equilibrium partition sediment benchmark model [23],which predicts benthic organism survivability in contaminatedsamples bymeasuring 18 parent and 16 alkylated PAH homologues,commonly known as PAH34. In the European Union, the OSPARCommission measures both parent and alkylated PAH and PASH insediments [24], while the European Committee for Standardization[25] monitors these compounds when conducting oil spill forensicinvestigations. Likewise, the American Society for Testing andMaterials (ASTM) targets 14 parent and alkylated PAH/PASH in itsforensic method [26] and 23 PAH and homologues for pore watertoxicity [27]. Other organizations such as the International Councilfor the Exploration of the Sea (ICES) [28], the National Oceano-graphic and Atmospheric Administration (NOAA) [29e31] and theUS Geological Survey (USGS) [32] employ methods to quantify 50parent and alkylated PAH and PASH. All of these methods rely ongas chromatography/mass spectrometry (GC/MS) for analysis.
Most of these standardizedmethods state that theMS should beoperated in the selected ionmonitoring (SIM) mode. The purportedgoal is to increase sensitivity by acquiring data for themolecular ionof each homologue (1-ion). Integration of 1-ion signals is typicallyperformed by manual integration of peaks after drawing a straightbaseline from the first- to the last-eluting isomers for each homo-logue (the retention window). Some methods require co-maximization of the molecular ion to confirm homologue identity(2-ion detection method).
Although high quality data are desired, retentionwindows differfrom method-to-method such that some of the standardizedmethodsmiss compoundswhile others include non-target PAH andPASH from the sample. For instance, the results of two NationalEnvironmental Laboratory Accreditation (NELAC) certified labora-tories, employing 1-ion detection methods, showed concentrationdifferences were the result of peak integration across differentretentionwindows [33]. In another example, a round-robin study ofthe Nordtest Methodology for Oil Spill Identification [34,35], whichlater became method CEN/TR 15522-2 [25], indicated laboratoriesexperienced “problems in identifying and integrating the correct[homologue] peaks”. Similarly, a different round-robin study of 26NELAC-certified laboratories found analysts produced, on average,27% inaccuracy for the concentrations of the parent PAH versus 53%for the alkylated homologues in marine sediment QA10OIL01 [36].
In addition to ill-defined integration windows, we and othershave shown that PAH and PASH interfere with one another[13,37e40]. These interferences primarily stem from two sources:
1) Two or more homologues from the same parent compoundelutewithin the same retentionwindowand have common ions.One example is C3 and C4 benzothiophene; m/z 176 is the mo-lecular and minor ion for both homologues.
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
2) PAH and PASH elute within the same retentionwindow and alsohave common molecular and/or confirming ions [13,41,42]. Forexample, m/z 234 is the molecular ion of C4 phenanthrene andC0 benzo[b]naphtho[2,3-d]thiophene. Similarly, C1 benzo[b]naphtho[2,3-d]thiophene and C5 phenanthrene have commonmolecular and fragment ions.
While PAH and PASH can be fractionated [41,42], the vast ma-jority of peer-reviewed research, standardized methods, andStandard Reference Material (SRM) certification programs quantifythese homologues without the benefit of separating one from theother.
Our aim is to evaluate the extent of inherent error within 1- and2-ion standardized methods. In addition, we evaluate the dataquality of using GC/MS-SIM based on detecting multiple fragmen-tation patterns per homologue (MFPPH) to reduce or eliminateerrors. Unlike 1- or 2-ion detection methods, MFPPH acquires 3e5ions for each PAH and PASH fragmentation pattern. New spectraldeconvolution software correctly assigns peaks to each homologueindependent of matrix interferences and without sample frac-tionation. To evaluate method performances, parent and alkylatedPAH and PASH concentrations were measured in unfractionatedcrude oil, SRM 1582, and marine sediment, SRM 1941b, using thehomologue retention windows found in OSPAR, CEN, ASTM, ICES,USGS, and MFPPH. For each of these methods, accuracy, precision,sensitivity, and selectivity were compared to the National Instituteof Standards and Technology (NIST) certified SRM concentrations.
2. Experimental
2.1. Materials and sample preparation
Sources for solvents, solid PAH/PASH and standards are providedin the Supporting Information. Prior to analysis, SRM 1582 wasdiluted with methylene chloride to produce a 5% solution of crudeoil. Pressurized liquid extraction of marine sediment SRM 1941bwas used to obtain extracts for analysis. Amore detailed descriptioncan be found in the Supporting Information [33]. Because recentround-robin studies from NIST indicated 1/3 to ~2/3 of laboratoriesperformed no sample fractionation [36,43], the SRM samples werenot fractionated prior to 1-ion, 2-ion and MFPPH analyses.
2.2. Instrumentation
The GC/MS (models 6890/5977A, Agilent Technologies, SantaClara, CA) was equipped with an autosampler and cooled injectioninlet (models MPS2/CIS6, Gerstel, Mülheim an der Ruhr, Germany).Table S2 in the Supporting Information provides instrument oper-ating conditions. Table 1 lists SIM/MFPPH ion groups. Ion acquisi-tion window start/stop times were set 20 s before and after elutionof the first and last compound in each group. Some alkylated ho-mologues elute in more than one ion group. The Ion Analytics(Andover, MA) data analysis software automatically identifies eachC0 to C4 PAH and PASH and correctly assigns each compound.
Ion group retention indices (RI) are calculated from the reten-
tion times (RT), Index ¼ 100*�Nþ ðtr;X�tr;1Þ
ðtr;2�tr;1Þ
!, where N is the num-
ber of aromatic rings in the first bracketing compound. tr,1 and tr,2are the retention times of the first and second bracketing com-pounds, respectively. tr,X is the start/stop time of the target iongroup. By rearranging the equation for tr,X, method-specific start/stop timeswere calculated to determine theMS acquisition time foreach ion group. The equation is based on the Lee index, wherenaphthalene, phenanthrene, chrysene and benzo[ghi]perylene are
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
Table 1MFPPH selected ion monitoring ions and acquisition windows (or retention index, RI).
the bracketing compounds. This approach is independent of stan-dardized method GC operating conditions as long as RXi-5 orequivalent stationary phase is used and all analytes elute during thelinear time domain of the temperature program employed.
2.3. Data analysis
For quality control purposes, the calibration curve is acceptablewhen the correlation coefficient, r2, was �0.990 and the relativestandard deviation of the average response factor, RF, was �20%.Continuing calibration (CC) measurements were performed at thebeginning, middle, and end of each day using a midpoint calibra-tion standard. Data collected during a given daywas acceptedwhenthe relative percent difference between RF and CC was �20%.Additional calibration information is available in the SupportingInformation.
Spectral deconvolution software was used to quantify PAH andPASH by MFPPH, 1-ion and 2-ion methods. MFPPH requires moni-toring one fragmentation pattern (molecular and two fragmenta-tion ions) for each parent PAH and PASH. In contrast, alkylatedhomologues were identified using as many fragmentation patterns
Table 2Parent compound comparison of MFPPH and NIST certified concentrations in SRM 1582
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
as needed (up to 11 for the C3 and C4 naphthalene homologues)within the expected retention windows. Each fragmentationpattern contained two to four qualifier ions to confirm homologueidentity. The compound identity criterion has been described pre-viously [33]. Briefly, the software extracts ion signal for eachspectral pattern from the data and normalizes the qualifier ionabundances to the molecular ion and expected abundances ac-cording to the following equation:
IiðtÞ ¼AiðtÞRiA1
(1)
where Ii(t) is the normalized ion intensity at scan (t), Ai(t) is theintensity of the i-th ion, and Ri is the expected ion ratio for themolecular ion (i ¼ 1) and qualifier ions i ¼ 2 to 4. The acceptable Riwas set at ± 20% for 5 consecutive scans. A histogram representingthe normalized ion ratio is generated for each scan; the flatter it is,the closer the actual ratio is to the expected ratio.
The spectral match of the average deviation of N reduced ionintensities (DI) is calculated:
Fig. 1. A) Method-specific retention windows for the C3 phenanthrene. The blue traceis the molecular ion for this homologue in marine sediment SRM 1941b. B) C3 phen-anthrene peaks meet the compound identity criterion for fragmentation patterns A, B,D, E, and F. Only those spectral patterns that meet this criterion are integrated, seeexperimental section. (For interpretation of the references to colour in this figurelegend, the reader is referred to the web version of this article.)
The criterion is met when DI � K þ D0=A1 where K is theacceptable relative percent difference and D0 is the additive errorfrom instrument noise and background signal. Scan-to-scan vari-ance criterion (DE), calculated: DE ¼ DI*log A1, is acceptable whenDI or DE are below the maximum allowable error (DEmax), whichwas set at 7 for this method.
The Q-value is determined for each scan in the peak and mea-sures the deviation of the absolute value of expected minusobserved ion ratio divided by the expected ion ratio times 100. Apeak is acceptable when Q-value is � 95. The Q-ratio compares theratio of molecular ion intensity to the intensities of each confirmingion over the whole peak. The acceptability limit for this criterion is±20%. Only MFPPH-confirmed peaks eluting within the retentionwindow are integrated; all other peaks are discarded. Baselines areautomatically generated by the software, which can be adjusted bythe analyst if necessary.
1-ion retention windows depicted in the OSPAR [24], ASTMD5739 [26] and ICES [28] methods as well as the 2-ion windowsfrom CEN [25] and the USGS [32] were compared to MFPPH incrude oil SRM 1582 and marine sediment SRM 1941b. In the 1-ionand 2-ion analyses, homologue molecular ion peak patterns wereextracted from SIM/MFPPH data files and straight-baseline inte-grated over a window corresponding to the elution of the first andlast target peaks as visually depicted in the respective method.Methods ICES and OSPAR use the same windows. Because ASTMD5739 is a non-quantitative forensic method, misestimation of theconcentration is considered equivalent to misestimating the peakarea for that homologue. 2-ion methods require an additional step,i.e., confirming the presence of a second ion, but these methods donot specify relative abundance. The USGS method only providedvisual retention windows for the phenanthrene homologues. Forthis reason, C1-C4 phenanthrenewere the only homologues studiedalthough USGS monitors additional homologue groups. Homo-logues monitored by MFPPH include C1 - C4 benzothiophene,naphthalene, dibenzothiophene, phenanthrene, 4-ring peri-condensed PASH, pyrene, 4-ring ortho-fused PASH (benzo[b]naphtho[2,3-d]thiophene) and chrysene as well as C1-C3 fluorene,see Supporting Information Table S3 for a list of all compounds,retention indices, and fragmentation patterns.
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
2.4. Data quality
SIM/1-ion, 2-ion, and MFPPH method performances wereevaluated against NIST certified SRM concentrations [44,45]. Sedi-ment extraction efficiency was measured as the percent recovery of2-methylnaphthalene-d10, the surrogate, and parent compounds inSRM 1941b. Each sample was prepared and analyzed 3-times todetermine method precision. The limit of detection (LOD) was
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
based on the analysis of seven identically-prepared standardswhose concentrations were between 1- and 10-times the estimateddetection limit. The LOD was calculated according to the followingequation: LOD ¼ SD � Student
0s t, where SD is the standard devi-
ation of n ¼ 7 measurements and Student's t is the one-handedvalue at 6� of freedom and 99% confidence, or 3.707 [46]. Upperand lower confidence limits for the detection limits were calculatedby multiplying the LOD by 2.2 and 0.64, respectively.
3. Results and discussion
To our knowledge this paper is the first to evaluate standardizedmethods for the analysis of PAH/PASH and their alkylated homo-logues. We examine selectivity, precision, accuracy, and sensitivityand the extent to which MFPPH can reduce or eliminate inherenterror in these methods. We start with the premise that the processused by NIST to certify SRMs is rigorous and leads to the best es-timate of the “true” concentration of an analyte in a sample. Forexample, NIST performed extensive cleanup of SRM 1941b beforeanalyzing the sample by GC/MS and HPLC using a variety of sta-tionary phases [47]. In addition, NIST carried out a round-robininterlaboratory study, which included analytical data from 30labs. Based on these data NIST estimated the concentration of PAHand PASH in the sediment and determined the corresponding un-certainty. Although alkylated calibration standards have beenshown to improve homologue quantitation [48], the greatestcomparability with certified values is obtained by using parent RF
Fig. 2. A) MFPPH retention windows for C3, C4 and C5 naphthalene, naphthothiophene (NT),NT, DBT, and C4 naphthalene. The red trace is the molecular ion, m/z 198, for C1 dibenzothiopnaphthalene, and C1 dibenzothiophene, respectively. (For interpretation of the references t
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
since few laboratories used alkylated standards when certifyingconcentrations in each SRM. For example only 6 of 26 labs usedalkylated standards in the round-robin study used to certify SRM1582 [36].
3.1. Parent PAH and PASH
All standardized methods, including MFPPH, rely on quantifyingcompounds based on their retention time and co-maximization ofmolecular and confirming ions and their abundances. In this studyquantitation of parent compounds follows the same criteria. Sincethe same data (calibration and sample) files are used, MFPPHshould produce the same results as standardized methods for theparent compounds excluding sample preparation errors.
Table 2 compares MFPPH to NIST certified concentrations inboth the crude oil and marine sediment samples. Crude oil con-centrations differed little and since there are as many over-estimated as underestimated compounds, little bias is observed inthe data. In contrast, NIST certified sediment concentrations werehigher than MFPPH for 13 of 18 analytes, with four compoundsexceeding 20% difference. This is consistent with the recovery of thesurrogate, 2-methylnaphthalene-d10, which was 81 ± 15%.Although low, the recovery was within the acceptable QC range ofEPA's SW 846, 8270 method. Precision was very high, at 6e7% RSDon average in both matrixes. The average RSD for all PAH/PASH was7 ± 5% for SRM 1582 (n ¼ 3) and 6 ± 4% for SRM 1941b (n ¼ 3).
Table 3 shows the instrument response based on SIM/MFPPH
and C0 and C1 dibenzothiophene (DBT). The black trace is the molecular ion,m/z 184, forhene and C5 naphthalene. B, C, and D) are the spectrally deconvolved signals for DBT, C4
o colour in this figure legend, the reader is referred to the web version of this article.)
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
operating conditions, including the average response factor (RF)and relative standard deviation (RSD) for each compound as well asthe limit of detection (LOD) and upper and lower confidence limits(UCL/LCL). On-column detection limits ranged from 0.33 pg forbenzo[b]naphtho[2,1-d]thiophene to 4.2 pg for indeno[1,2,3-cd]pyrene. The average detection limit was 1.7 ± 1.0 pg. No statisticaldifference in sensitivity was observed between MFPPH and 1-iondetection methods despite the increase in the total number ofions monitored. For example, MFPPH acquires signal for 44 ions/group compared to as few as 12 ions as specified in ASTM D7363programming protocols. The on-column LOD is more than an orderof magnitude below the stated LOD required for pore water ana-lyses by solid-phasemicroextraction [27]. In our case, the LOD is thesame as the method detection limit (MDL) for MFPPH, since spec-tral deconvolution of target compounds achieves the same sensi-tivity independent of matrix complexity [39,49,50].
3.2. Alkylated PAH and PASH
An illustrative example of differences in standardized methodretention windows is shown in Fig. 1a for C3 phenanthrene in
Fig. 3. Percent differences versus NIST-certified homologue concentrations in unfractionastandardized methods.
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
marine sediment SRM 1941b. Themolecular ion for this homologueis m/z 220, see blue trace. It is evident that the concentration isdependent on retention window. For example, ASTM D5739 in-tegrates 12 more peaks compared to USGS, yielding 224 ng/g vs.175 ng/g or 29% higher concentration, see Supporting InformationTable S4. In contrast, MFPPH produced 150 ng/g. Fig. 1b makes clearall peaks within the USGS retention range are not C3 phenan-threnes. Based on spectral deconvolution of spectral patterns A-F(see Supporting Information Table S3), the correct retention rangeis CEN. The addition of a single confirming ion is insufficient toexclude matrix peaks. The NIST certified concentration for thishomologue was 165 ng/g. Notice that the USGS method excludesthe C3 phenanthrene peak at 28.9 min, which, if included, increasesthe concentration by 20 ng/g.
As stated earlier, errors in alkylated PAH and PASH concentra-tions also occur when two or more target compounds elute in thesame retentionwindow and have common ions. Using SRM 1582 asan example, Fig. 2a shows the retention widows for C3 and C5naphthalene, which overlap with C4 naphthalene, see Fig. 2c. Themolecular ion of C4 naphthalene (m/z 184, black) is also a fragmention of C5 naphthalene and is the base ion for dibenzothiophene
ted (A) crude oil SRM 1582 and (B) marine sediment SRM 1941b by MFPPH and the
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
(DBT) and naphthothiophene (NT). Manual integration of C4naphthalene peaks based solely on m/z 184 will include C5 naph-thalene, DBT, and NT peaks if not careful. Although the addition of aconfirming ion should exclude DBT and NT, C3 and C5 naphthalenealso produce C4 naphthalene confirming ions (e.g. m/z 169, 141).Another example of target compounds eluting in the same reten-tionwidow are C5 naphthalene and C1 dibenzothiophene, sincem/z198 is the molecular ion for both alkylated homologues, see Fig. 2d.
Spectral deconvolution of MFPPH ions eliminates these errors,since the software examines each peak to confirm identity. SinceDBT and NT are single compounds, the compound identity criterionexcludes all other peaks in the chromatogram from integration, seeFig. 2b. The software automatically integrates total peak area andcalculates concentration for each alkylated homologue and parentcompound. The spaces between shaded peaks in Fig. 2d makeevident the benefit of integrating individual homologue peaksversus all peaks in a retention window.
Fig. 3a (crude oil) and Fig. 3b (marine sediment) show MFPPH(blue) and standardized method (CEN (red), ICES and OSPAR (yel-low), ASTM D5739 (green) and USGS (purple)) concentration dif-ferences from the NIST certified concentrations. Actualconcentrations are in Supporting Information Table S4. Findingsindicate that the standardizedmethods overestimate alkylated PAHand PASH homologues more often than MFPPH. The degree ofoverestimation increases with the increase in alkylation and degreeof aromaticity, from C1 to C4 and from 2- to 6-rings, since thesehomologues have greater structural diversity and wider elutionwindows. The wider the elution window, the more target andmatrix compounds overlap. Because C3- and C4 homologues areoften lower in concentration than C1- and C2, these homologues areespecially affected by matrix compounds. Recall the USGS methodonly provides retention windows for alkylated phenanthrenes.ICES, OSPAR, and ASTM D5739 routinely overestimate PAH andPASH homologues more than CEN in both SRMs.
For some methods, accuracy was sample dependent. Forexample, CEN overestimated C4 benzothiophene in SRM 1941b by140% but only 20% for SRM 1582. ICES and OSPAR overestimated C1pyrene in SRM 1582 by 48% but produced concentrations within 3%of NIST for SRM 1941b. In contrast, MFPPH eliminates positive biasfrom the data as evidenced by the average difference in homologueconcentration of 2e3%. Importantly, MFPPH need not rely onretention windows since the ions selected to confirm homologueidentity minimize peak pattern recognition which is the hallmarkof the standardized methods.
4. Conclusion
All court-defensible standardized methods rely on at least threeions per compound to confirm analyte identity except whenquantifying PAH and PASH homologue concentrations where oneion is used. Consistent with the aforementioned statements madein the Nordtest study [34,35], warning is provided in most of thestandardized methods when using 1-ion analysis. For example,ICES states “this method will result in an overestimation of theconcentration, because the sum may include non-target com-pounds” [28]. EPA Method 8270D also cautions “SIM may provide alesser degree of confidence in the compound identification” [20].
Because the spectral deconvolution software eliminates additiveion current due to coeluting compounds, the MFPPH method incombination with the spectral deconvolution of homologue-specific fragmentation patterns based on 3 to 5 ions per patternyielded improved accuracy and precision based on selective targetcompound detection without sacrificing sensitivity. The results ofthis study are consistent with our earlier findings obtained fromNELAC certified labs [33]. Because MFPPH only requires additional
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
acquisition ions, MFPPH can be integrated into any standardizedmethod to improve PAH and PASH data quality.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.aca.2017.04.017.
References
[1] D.M. Pampanin, M.O. Sydnes, Polycyclic aromatic hydrocarbons a constituentof petroleum: presence and influence in the aquatic environment, in:V. Kutcherov (Ed.), Hydrocarbon, InTech, Rikeja, Croatia, 2013.
[2] I.M. Afanasov, A.V. Kepman, V.A. Morozov, A.N. Seleznev, V.V. Avdeev,Determination of polyaromatic hydrocarbons in coal tar pitch, J. Anal. Chem.64 (4) (2009), 365e365.
[3] M.R.T. Palmroth, J.H. Langwaldt, T.A. Aunola, A. Goi, U. Münster, J.A. Puhakka,T.A. Tuhkanen, Effect of modified Fenton's reaction on microbial activity andremoval of PAHs in creosote oil contaminated soil, Biodegradation 17 (2006)131e141.
[4] A.D. Wheatley, S. Sadhra, Polycyclic aromatic hydrocarbons in solid residuesfrom waste incineration, Chemosphere 55 (2004) 743e749.
[5] M. Patri, Polycyclic aromatic hydrocarbons in air and their neurotoxic potencyin association with oxidative stress: a brief perspective, Ann. Rev. Neurosci 16(1) (2009) 22e30.
[6] K.G. Kropp, P.M. Fedorak, A review of the occurrence, toxicity and biodegra-dation of condensed thiophenes found in petroleum, Can. J. Microbiol. 44 (7)(1998) 605e622.
[7] J. Jacob, The significance of polycyclic aromatic hydrocarbons as environ-mental carcinogens. 35 years research on PAH-a retrospective, Polycyc. Arom.Comp. 28 (4�5) (2008) 242e272.
[8] P. Irigaray, D. Belpomme, Basic properties and molecular mechanisms ofexogenous chemical carcinogens, Carcinogenesis 31 (2) (2010) 135e148.
[9] B. Mahadevan, L.A. Courter, W.M. Baird, Mutations induced by polycyclic ar-omatic hydrocarbons, in: C.L. Valon (Ed.), New Developments in MutationResearch, Nova Science Publishers Inc., Hauppage, New York, 2007.
[10] T. Toyooka, Y. Ibuki, DNA damage induced by coexposure to PAHs and light,Environ. Toxicol. Pharmacol 23 (2) (2007) 256e263.
[11] D. Warshawsky, Environmental sources, carcinogenicity, mutagenicity,metabolism and DNA binding of nitrogen and sulfur heterocyclic aromatics,J. Heterocyc. Chem. 21 (2) (1984) 353e359.
[12] S.M. Billiard, J.N. Meyer, D.M. Wassenberg, P.V. Hodson, R.T. Di Guilio,Nonadditive effects of PAHs on early vertebrate development: mechanismsand implications for risk assessment, Toxicol. Sci. 105 (1) (2008) 5e23.
[13] P.M. Antle, C.D. Zeigler, N.M. Wilton, A. Robbat Jr., A more accurate analysis ofalkylated PAH and PASH and its implications in environmental forensics,Intern. J. Environ. Anal. Chem. 94 (4) (2014) 332e347.
[14] S.A. Stout, S.D. Emsbo-Mattingly, G.S. Douglas, S. Gregory, A.D. Uhler,K.J. McCarthy, Beyond 16 priority pollutant PAHs: a review of PACs used inenvironmental forensic chemistry, Polycyc. Arom. Comp. 35 (2015) 285e315.
[15] P.M. Antle, C.D. Zeigler, D.G. Livitz, A. Robbat Jr., Two-dimensional gas chro-matography/mass spectrometry, physical property modeling and automatedproduction of component maps to assess the weathering of pollutants,J. Chrom. A 1364 (2014) 223e233.
[16] F. Yin, G.F. John, J.S. Hayworth, T.P. Clement, Long-term monitoring data todescribe the fate of polycyclic aromatic hydrocarbons in Deepwater Horizonoil submerged off Alabama's beaches, Sci. Tot. Environ. 508 (2015) 46e56.
[17] D.M. Di Toro, J.A. McGrath, Predicting the toxicity of neat and weatheredcrude oil: toxic potential and the toxicity of saturated mixtures, Environ.Toxicol. Chem. 26 (2007) 24e36.
[18] D.M. Di Toro, C.S. Zarba, D.J. Hansen, W.J. Berry, R.C. Swartz, C.E. Cowan,S.P. Pavlou, H.E. Allen, N.A. Thomas, Technical basis for establishing sedimentquality criteria for nonionic organic chemicals by using equilibrium parti-tioning, Environ. Toxicol. Chem. 10 (1991) 1541e1583.
[19] U.S. EPA. EPA-600-R-02e013, Semivolatile Organic Compounds by GasChromatography/Mass Spectrometry, Government Printing Office, Washing-ton, D.C., 2014.
[20] U.S. EPA. Method 8270D, Semivolatile Organic Compounds by Gas Chroma-tography/Mass Spectrometry, Government Printing Office, Washington, D.C.,2003.
[21] U.S. EPA. EPA-625-R-96e010b, Compendium of Methods for the Determina-tion of Toxic Organic Compounds in Ambient Air, Government Printing Office,Washington D.C., 1999.
[22] U.S. EPA. Method 8272, Parent and Alkyl Polycyclic Aromatics in SedimentPore Water by Solid-phase Microextraction and Gas Chromatography/MassSpectrometry in Selected Ion Monitoring Mode, Government Printing Office,Washington, D.C., 2007.
[23] U.S. EPA. EPA-600-R-02e012, Procedures for the Derivation of EquilibriumPartitioning Sediment Benchmarks (ESBs) for the Protection of Benthic Or-ganisms: PAH Mixtures, Government Printing Office, Washington, D.C., 2003.
[24] OSPAR, 1.5.5.16, JAMP Guidelines for Monitoring Contaminants in Biota andSediments, London, UK, 2008.
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
[25] CEN, CEN/TR 15522-2, Oil Spill Identificationdwaterborne Petroleum andPetroleum Productsdpart 2: Analytical Methodology and Interpretation ofResults Based on GC-FID and GC-MS Low Resolution Analyses, EuropeanCommittee for Standardization, Brussels, 2006.
[26] ASTM, D5739-06, Standard Practice for Oil Spill Source Identification by GasChromatography and Positive Ion Electron Impact Low Resolution MassSpectrometry, ASTM International, West Conshohocken, PA, 2012.
[27] ASTM, D7363-11, Determination of Parent and Alkyl Polycyclic Aromatics inSediment Pore Water Using Solid-phase Microextraction and Gas Chroma-tography/Mass Spectrometry in Selected Ion Monitoring Mode, ASTM Inter-national, West Conshohocken, PA, 2012.
[28] ICES, Techniques in Marine Environmental Sciences No. 45, Determination ofParent and Alkylated Polycyclic Aromatic Hydrocarbons (PAHs) in Biota andSediment, Copenhagen, Denmark, 2010.
[29] NOAA, Technical Memorandum NOS NCCOS 30, Organic ContaminantAnalytical Methods of the National Status and Trends Program: 2000-2006,Government Printing Office, Washington, D.C., 2006.
[30] NOAA, Technical Memorandum NOS ORCA 130. Sampling and AnalyticalMethods of the National Status and Trends Program, Silver Spring, MD, 1998.
[32] USGS, Water-Resources Investigations Report 03-4318, Methods of Analysisby the U.S. Geological Survey National Water Quality LaboratorydDetermi-nation of Polycyclic Aromatic Hydrocarbon Compounds in Sediment by GasChromatography/Mass Spectrometry. Government Printing Office, Washing-ton, DC.
[33] A. Robbat Jr., N.M. Wilton, A new spectral deconvolutioneSelected ionmonitoring method for the analysis of alkylated polycyclic aromatic hydro-carbons in complex mixtures, Talanta 125 (2014) 114e124.
[35] SINTEF, STF66 A02028, Revision of the Nordtest Methodology for Oil SpillIdentification, SINTEF Applied Chemistry, Trondheim, Norway, 2002.
[36] NIST. NISTIR 7793, Interlaboratory Analytical Comparison Study to SupportDeepwater Horizon Natural Resource Damage Assessment: Description andResults for Crude Oil QA10OIL01, US Department of Commerce, Gaithersburg,MD, 2011.
[37] C. Yang, G. Zhang, Z. Wang, Z. Yang, B. Hollebone, M. Landriault, K. Shah,C.E. Brown, Development of a methodology for accurate quantitation ofalkylated polycyclic aromatic hydrocarbons in petroleum and oil
Please cite this article in press as: N.M. Wilton, et al., Errors in alkylated pocaused by currently employed standardized methods, Analytica Chimica
contaminated environmental samples, Anal. Methods 6 (2014) 7760e7771.[38] Y. Zhao, B. Hong, Y. Fan, M.M. Wen, X. Han, Accurate analysis of polycyclic
aromatic hydrocarbons (PAHs) and alkylated PAHs homologues in crude oilfor improving the gas chromatography/mass spectrometry performance,Ecotoxicol. Environ. Saf. 100 (2014) 242e250.
[39] C.D. Zeigler, A. Robbat Jr., Comprehensive profiling of coal tar and crude oil toobtain mass spectra and retention indices for alkylated PAH shows whycurrent methods err, Environ. Sci. Technol. 46 (2012) 3935e3942.
[40] C.D. Zeigler, N.M. Wilton, A. Robbat Jr., Toward the accurate analysis of C1-C4polycyclic aromatic sulfur heterocycles, Anal. Chem. 84 (2012) 2245e2252.
[42] A.H. Hegazi, J.T. Andersson, Limitations to GC-MS determination of sulfur-containing polycyclic aromatic hydrocarbons in geochemical, petroleum,and environmental investigations, Energy & Fuels 21 (2007) 3375e3384.
[43] NIST. NISTIR 7792, Interlaboratory Analytical Comparison Study to SupportDeepwater Horizon Natural Resource Damage Assessment: Description andResults for Marine Sediment QA10SED01, US Department of Commerce,Gaithersburg, MD, 2011.
[44] NIST, Certificate of Analysis: Standard Reference Material 1582, US Depart-ment of Commerce, Gaithersburg, MD, 2012. https://www-s.nist.gov/srmors/certificates/1582.pdf.
[45] NIST, Certificate of Analysis: Standard Reference Material 1941b, US Depart-ment of Commerce, Gaithersburg, MD, 2002. https://www-s.nist.gov/srmors/certificates/1941b.pdf.
[46] U.S. EPA. 51 FR 23703, Definition and Procedure for the Determination of theMethod Detection Limit, Government Printing Office, Washington, DC, 1986.
[47] S.A. Wise, D.L. Poster, M.M. Schantz, J.R. Kucklick, L.C. Sander, M. Lopez deAlda, P. Schubert, R.M. Parris, B.J. Porter, Two new marine sediment standardreference materials (SRMs) for the determination of organic contaminants,Anal. Bioanal. Chem. 378 (5) (2004) 1251e1264.
[48] C. Yang, G. Zhang, Z. Wang, Z. Yang, B. Hollebone, M. Landriault, K. Shah,C.E. Brown, Development of a methodology for accurate quantitation ofalkylated polycyclic aromatic hydrocarbons in petroleum and oil contami-nated environmental samples, Anal. Methods 6 (2014) 7760e7771.
[49] A. Robbat Jr., A. Hoffmann, K. MacNamara, Y. Huang, Quantitative identifica-tion of pesticides as target compounds and unknowns by spectral deconvo-lution of gas chromatographic/mass spectrometric data, J. AOAC Int. 91 (6)(2008) 1467e1477.
[50] C.D. Zeigler, N.M. Wilton, A. Robbat Jr., Toward the accurate analysis of C1�C4polycyclic aromatic sulfur heterocycles, Anal. Chem. 84 (5) (2012) 2245e2252.
lycyclic aromatic hydrocarbon and sulfur heterocycle concentrationsActa (2017), http://dx.doi.org/10.1016/j.aca.2017.04.017
