Identification and quantification of methylated PAHs in sediment by two-dimensional gas chromatography/mass spectrometry

Abstract

Alkylated polycyclic aromatic hydrocarbons (alkyl-PAHs) are ubiquitously present in the environment and they are recognized as a toxicologically hazardous group. The biggest obstacle in the assessment of environmental risks of alkyl-PAHs is identification and quantification; the complete (chromatographic) separation of alkylated homologues is difficult if not impossible. Therefore, alkyl-PAHs are usually identified as a group of isomers with the same degree of alkylation and quantified as one group using one chromatographic response factor. In this study we demonstrate that the relative response factors of twenty-three methylated PAHs with the same molecular weight of 242 (six methyl-chrysenes, twelve benz[a]anthracenes and five benzo[c]phenanthrenes) range from 0.1 for 12-methylbenz[a]anthracene and 4-methylbenzo[c]phenanthrene to 1.7 for 6-methylbenz[a]anthracene. Quantification of methylated PAHs with equal molecular weights as a group using the same relative response factor can thus overestimate or underestimate their concentrations and, therefore, the toxicological risk of an environmental sample. A two-dimensional gas chromatography method was developed with which fourteen methylated PAHs (M_w = 242) could be separated. Twelve of them were identified and quantified in Elbe River sediment.

---

Introduction

Alkylated polycyclic aromatic hydrocarbons (alkyl-PAHs) are ubiquitously present in the environment; they were identified and partially quantified in sediments, crude oil, automobile exhaust, wood combustion products and food.^1–10 Anthropogenic sources of alkyl-PAHs are mostly of petrogenic origin;¹¹ the ratios of the alkyl-homologues to the parent PAHs can be used as an indicator of petrogenic/pyrogenic PAH distribution¹² and as a weathering indicator.⁶

We were able to identify alkyl-PAHs with different parent PAHs and different alkylation levels in River Elbe sediment.¹³ The contribution of alkyl-PAHs to the toxic activity of environmental PAH mixtures has already been acknowledged^2,7–9 and the detailed toxicological profiling of some individual homologues is also available.^14–18 It has been often concluded that alkyl-PAHs form a toxicologically hazardous group although they are not fully recognized as dangerous pollutants. Brack and Schirmer postulated that the hazard environmental assessment should focus more on methylated PAHs than on their parent compounds.⁷

Identification and quantification form the biggest obstacle in assessing environmental risks of alkyl-PAHs. They can neither be fully separated by gas nor by liquid chromatography and their standards are poorly available. Individual isomers coelute and mass spectrometric detection does not improve the situation, because the spectra of the isomers are identical. Therefore, alkyl-PAHs are usually identified and quantified as a whole group of isomers with the same degree of alkylation.^{2–5,12,19,20} Because of that insufficient separation, the toxic potency of the sample may be significantly over- or underestimated.

Different alkyl-PAHs can cause different toxic effects and, additionally, chromatographic response factors (different for every congener) may influence quantification. In order to assess environmental risks, congener-specific analysis is required. In this study, a two-dimensional gas chromatography (GC×GC) method with mass spectrometric (MS) detection is applied to improve the separation of the methylated homologues (M_w = 242) of chrysenes (1–6MC), benz[c]phenanthrenes (1–5MBP) and benz[a]anthracenes (1–12MBA) from each other and from other possible interferences. These homologues were chosen in this study because of their abundant occurrence in environmental samples and toxic responses that were confirmed in several Effect Directed Analysis (EDA) studies.^2,7,8 In in vivo studies 5MC, 7MBA and 12MBA appear to be strong carcinogens¹⁴ and the tumor initiating ability was reported for 6MBA, 9MBA, 6MC and 3MC.^15,17,21

Experimental section

Materials

The methyl-PAH standard solutions of 1MC, 2MC, 3MC, 4MC, 6MC, and 1–12MBA in isooctane were obtained from Dr M. Machala (Veterinary Research Institute, Brno, Czech Republic). Chrysene, 5MC, 1MBP, 2MBP, 3MBP, 4MBP, 5MBP and deuterated benz[a]anthracene (neat products) were purchased from Sigma Aldrich. All solvents used (isooctane, hexane, acetone) were obtained in picograde quality from Promochem, Wesel, Germany.

Standard solutions

The separation method was developed using individually prepared 10 μg mL⁻¹ standards and standard solutions of methyl-PAHs.

Relative response factors were determined using standard solutions of methyl-PAHs prepared within the concentration range of 1.2–24.0 μg mL⁻¹ and spiked with 1.8 μg mL⁻¹ stock solution of chrysene.

Methyl-PAHs in the River Elbe sediment samples were quantified using calibration mixtures containing benzo[c]phenanthrene, benz[a]anthracene, triphenylene, chrysene, naphthacene, 2MBP, 3MBP, 4MBP, 5MBP, 2MBA, 7MBA, 9MBA, 10MBA, 1MC, 2MC, 3MC and 5MC within the concentration range of 1.2–50.0 μg mL⁻¹ and spiked with 50 μL of a 22 μg mL⁻¹ stock solution of deuterated benz[a]anthracene (internal standard).

Sediment samples

River Elbe surface sediment samples were collected on April 2005 at Prelouč in the Czech Republic. Sediment from this location was analyzed previously within the scope of the MODELKEY project.¹³ The top 5 cm of sediment was sampled with a Van Veen grab, sieved (<63 μm), homogenized, and freeze-dried prior to the analysis.

Extraction

Two grams of sediment was spiked with deuterated benz[a]anthracene (internal standard) and mixed with ca. 2 g Na₂SO₄ and small glass pellets in a glass extraction thimble. The thimble was placed in a Soxhlet extractor and the sediment was extracted with 150 mL hexane–acetone (1 : 1 v/v) for 12 h. The extract was carefully reduced to about 2 mL on a rotary evaporator (Heidolph) with iso-octane as a keeper. The final solvent was iso-octane.

Fractionation

The extracted compounds were separated into two fractions, depending on their polarity, on an open glass column filled from bottom-up with glass wool, 2 g Cu–silica (1 : 1 w/w; both silica and Cu activated), 5 g aluminium oxide (activated), 4 g silica (activated) and ca. 0.5 g Na₂SO₄. Compounds were eluted with solvents of increasing polarity into two fractions: 5 mL hexane (discharge) and 100 mL dichloromethane–hexane (3 : 1). The aliquots were gently evaporated on a rotary evaporator to ca. 1 mL and transferred to GC-vials. They were further evaporated to almost dryness and re-dissolved in 1 mL of isooctane.

Analysis with GC×GC-FID

The GC×GC system was built from a HP 6890 (Hewlett-Packard, USA) GC with a standard FID detector. The GC was equipped with an air modulator assembly consisting of two air jets (custom-made at the VU University, Amsterdam, Netherlands). Pressurized air was used for cooling.

The GC columns used in the first and second dimensions are listed in Table 1. All column connections were made by means of press-fit connectors (Techrom, Purmerend, The Netherlands).

Table 1 Overview of first- and second-dimension columns used

Commercial code	Stationary phase	Temperature limit^a (°C)	Dimensions (m × mm × μm)	Producer^b
a Maximum isothermal temperature/maximum programmed temperature. b J&K Scientific, Folsom, USA; SGE International, Rinwood, Australia; Quadrex, New Haven, USA; J&K Environmental, Milton, Canada.
First-dimension column
DB-5	5% Phenyl-methylpolysiloxane	325/350	30 × 0.25 × 0.25	J&K Scientific
DB-5	5% Phenyl-methylpolysiloxane	325/350	60 × 0.25 × 0.25	J&K Scientific
DB-XLB	Second generation arylene	340/360	30 × 0.25 × 0.25	J&K Scientific
HT-8	8% Phenyl-methylpolysiloxane	360/370	50 × 0.25 × 0.25	SGE International
Second-dimension column
007-65HT	65% Phenyl-methylpolysiloxane	360/360	1 × 0.10 × 0.10	Quadrex
LC-50	50% Liquid crystalline-methylpolysiloxane	270/270	1.5 × 0.10 × 0.10	J&K Scientific
LC-50	50% Liquid crystalline-methylpolysiloxane	270/270	2 × 0.10 × 0.10	J&K Scientific

---

Modulation was performed at the beginning of the second column with a modulation period of 5 s. Helium (99.999%) was used as a carrier gas with a velocity of 1.1 mL min⁻¹. 1 μL injections were made using an auto sampler with the injector operated in the splitless mode at 280 °C; the purge time was 2 min. The GC-oven program was as follows: 90 °C for 2 min, 20 °C min⁻¹ to 200 °C and then 1.5 °C min⁻¹ to the final temperature of 320 °C (10 min).

The GC×GC-FID system was used for the development of the analytical method.

Analysis with GC×GC-MS

The GC×GC system was built from a Clarus 500 MS (Perkin Elmer, Shelton, CT, USA). The GC was equipped with an air modulator assembly consisting of two air jets (custom-made at the VU University, Amsterdam). Pressurized air was used for cooling.

A 60 m DB-5 column (J&K Scientific; 60 m × 0.25 mm × 0.25 μm) was used in the first dimension, which separates on the basis of volatility. This column was combined with a 1.5 m LC-50 column, allowing polarity-based separation and group-type separation. The columns were connected with mini press-fits (Techrom).

Modulation was performed at the beginning of the second column with a modulation period of 7 s. Helium (99.999%) was used as a carrier gas with a velocity of 1.3 mL min⁻¹. 1 μL injections were automatically made with the PTV injector operated in the splitless mode at 300 °C. The GC-oven program was as follows: 90 °C for 2 min, 20 °C min⁻¹ to 200 °C, at 1 °C min⁻¹ to 295 °C and then at 25 °C min⁻¹ to the final temperature of 320 °C (5 min). The temperature of the transfer line was 300 °C. The MS was tuned and calibrated in the electron ionization (EI) mode using heptacosafluorotributylamine (Fluka Chemie, Buchs, Switzerland) as a reference gas. The ion source temperature was 250 °C; the electron energy was 70 eV. The MS was operated in the Total Ion Current (TIC) and Selected Ion Monitoring (SIM) modes. Identification and quantification of methylated PAHs with GC×GC was performed in SIM mode (m/z = 240.1, 242.1) with a dwell time of 0.01 s. Deuterated benz[a]anthracene (Aldrich) was used as an internal standard.

Determination of relative response factors (RRFs) of twenty-three methyl-PAHs (M_w = 242) with reference to chrysene (M_w = 282) was performed in SIM mode (m/z = 228.1, 242.1).

Data acquisition was performed using TurboMass software (Perkin Elmer).

The GC×GC-MS system was used for identification and quantification of methyl-PAHs in the River Elbe sediment.

Results and discussion

Development and optimization of the chromatographic separation method of methyl-PAHs

The chromatographic separation method was developed using a GC×GC-FID system, because GC coupled to a FID detector enabled us to change the columns without a long delay for system stabilization (this is the case in a GC-MS system).

The chromatographic separation method was developed using standard solutions of twenty-three methyl-PAHs with a molecular weight of 242.

Four columns were tested to improve the separation in the first dimension: DB-5 30 m, DB-5 60 m, DB-XLB 30 m and HT-8 50 m (for details see ESI†).

The extended DB-5 column (60 m) gave the best separation; it allowed full separation of seven homologues: 1MBP, 2MBP, 3MBP, 3MC, 5MC, 1MC and 10MBA. DB-XLB allowed separation of five homologues (1MBP, 2MBP, 3MBP, 1MC and 10MBA) and HT-8 four homologues (1MBP, 2MBP, 3MBP and 10MBA).

To test the separation in the second dimension DB-5 (60 m) was coupled with 007-65HT (1.2 m) and LC-50 (1.5 m).

The DB-5 (60 m) × LC-50 (1.5 m) column combination allowed complete separation of eleven homologues and partial separation of two homologues out of twenty-three injected compounds. Nine homologues still coeluted, in three groups: 8MBA + 11MBA, 4MBA + 6MBA and 4MC + 5MBA + 3MBA + 12MBA + 6MC.

Relative response factors of the individual homologues

Methylated PAHs are usually quantified as groups of homologues with the same degree of alkylation (e.g. C1, C2, etc.).^4,9,19,22 This is a rough approach, in which it is assumed that the chromatographic response factor of all methylated homologues is the same. However, this assumption is not correct. The relative response factors (RRFs) of twenty-three methyl-PAHs with a molecular weight of 242, with response to chrysene, were determined on GC×GC-MS using the following equation:

RRF = (A_AM_C)/(M_AA_C) (1)

where A_A = chromatographic area of the injected methyl-PAH, A_C = chromatographic area of the chrysene, M_A = mass of the injected methyl-PAH, and M_C = mass of the injected chrysene.

We established that the chromatographic response of MCs, MBAs and MBPs with reference to chrysene ranges from 0.1 for 12MBA and 4MBP to 1.7 for 6-MBA (see Table 2). Consequently, quantification of not fully separated methylated PAHs using the same response factor for all homologues gives unreliable analytical results.

Table 2 Relative response factors (RRFs) of methylated PAHs relative to chrysene. Monitored ions: 228 and 242 m/z

Compound	RRF (95% confidence interval)
1-Methylchrysene	0.34 (0.03)
2-Methylchrysene	0.60 (0.04)
3-Methylchrysene	0.46 (0.03)
4-Methylchrysene	0.32 (0.02)
5-Methylchrysene	0.19 (0.01)
6-Methylchrysene	0.23 (0.01)
1-Methylbenz[a]anthracene	0.54 (0.01)
2-Methylbenz[a]anthracene	0.73 (0.06)
3-Methylbenz[a]anthracene	0.87 (0.04)
4-Methylbenz[a]anthracene	1.17 (0.12)
5-Methylbenz[a]anthracene	1.05 (0.07)
6-Methylbenz[a]anthracene	1.72 (0.12)
7-Methylbenz[a]anthracene	0.31 (0.02)
8-Methylbenz[a]anthracene	0.57 (0.04)
9-Methylbenz[a]anthracene	0.72 (0.03)
10-Methylbenz[a]anthracene	0.96 (0.06)
11-Methylbenz[a]anthracene	0.91 (0.06)
12-Methylbenz[a]anthracene	0.12 (0.01)
1-Methylbenzo[c]phenanthrene	0.39 (0.02)
2-Methylbenzo[c]phenanthrene	0.33 (0.01)
3-Methylbenzo[c]phenanthrene	0.27 (0.01)
4-Methylbenzo[c]phenanthrene	0.14 (0.004)
5-Methylbenzo[c]phenanthrene	0.50 (0.04)

---

Identification and quantification of methylated benzo[c]phenanthrenes, chrysenes and benz[a]anthracenes

The separation method developed on GC×GC-FID was further optimized on GC×GC-MS using a River Elbe sediment extract spiked with 23 methylated PAH homologues. The best response for methyl-PAHs was obtained with an injector temperature of 300 °C and a source temperature of 250 °C. The modulation time was extended to 7 s because of the use of a slightly longer second dimension column (2 m instead of 1.5 m). The second dimension column was extended because of the wrap-around effect observed in the sediment sample. The wrap-around was most probably caused by the matrix effect (Fig. 1).

Fig. 1 GC×GC-FID chromatogram of 23 homologues of methylbenzo[c]phenanthrenes (MBP), methylchrysenes (MC) and methylbenz[a]anthrecenes (MBA) obtained on DB-5 (60 m) × LC-50 (1.2 m).

River Elbe sediment, unspiked and spiked with 23 methylated PAH homologues, was analyzed by GC×GC-MS in the SIM mode scanning for masses 242 (methyl-PAHs), 240 (deuterated benz[a]anthracene internal standard) and 228 (parent PAHs). In the unspiked Elbe River sediment five PAHs with M_w = 282 and twelve methyl-PAHs with M_w = 242 were identified and quantified (see Fig. 2 and Table 3).

Fig. 2 GC×GC-MS chromatogram of an unspiked River Elbe sediment obtained in SIM mode (masses 242 and 240) on DB-5 (60 m) × LC-50 (2.0 m). MBP: methylbenzo[c]phenanthrenes, MC: methylchrysenes, and MBA: methylbenz[a]anthrecenes.

Table 3 Concentrations of PAHs and methylated PAHs with M_w = 242 in River Elbe sediment

Compound	Concentrations in ng g^−1 dry weight (standard deviation in parentheses)
Benzo[c]phenanthrene	116 (8)
Benz[a]anthracene	608 (22)
Triphenylene	304 (18)
Chrysene	617 (30)
Naphthacene	289 (29)
2-Methylbenzo[c]phenanthrene	15 (0.5)
3-Methylbenzo[c]phenanthrene	23 (0.4)
5-Methylbenzo[c]phenanthrene	3 (0.6)
4-Methylbenzo[c]phenanthrene	18 (0.8)
2-Methylbenz[a]anthracene	42 (3)
7-Methylbenz[a]anthracene	125 (15)
9-Methylbenz[a]anthracene	572 (29)
3-Methylchrysene	234 (25)
2-Methylchrysene	886 (23)
5-Methylchrysene	68 (2)
1-Methylchrysene	437 (8)
10-Methylbenz[a]anthracene	Below limit of quantification

---

The peaks of the unresolved homologues were also recognized; these coeluting compounds are possibly also present in River Elbe sediment: 8MBA and/or 11MBA (two coeluting homologues), 4MBA and/or 6MBA (two coeluting homologues), 4MC and/or 5MBA and/or 3MBA and/or 12MBA and/or 6MC (five coeluting homologues). 1MC and 1MBA are not present in the sediment.

The potent toxic homologues 9MBA and 7MBA^14,15 were not separated in one-dimensional GC. They coelute on all columns tested in this study. The use of the second dimension allowed us to separate and quantify them. The concentration of 9MBA is almost equal to the concentration of the parent benz[a]anthracene. The response factor of 9MBA is more than 2-fold higher than the response factor of 7MBA (see Table 2) but usually, when calculating the concentrations of methylated PAHs, these compounds are calculated as a group. Only in two EDA studies of Brack et al.^7,8 9MBA was quantified as a distinct congener in toxic fractions.

The concentration of 9MBA (591 ng g⁻¹ dry weight) found in Elbe sediment in this study was substantially higher than the concentrations in sediments of the rivers Mulde and Neckar (Germany) reported by Brack et al. (92 ng g⁻¹ and 340 ng g⁻¹ (ref. 7 and 8)).

The potentially toxic homologues 6MC, 8MBA and 12MBA¹⁴ may be present in the sediment but they are coeluting with other compounds. 12MBA and 6MC are coeluting with 4MC, 5MBA and 3MBA. The response factors of homologues coeluting in this group range from 0.1 (12MBA) to 1.0 (5MBA); quantification using one response factor for the whole group gives unreliable results.

It is theoretically possible to identify and quantify 12MBA when changing the first dimension column from 60 m DB-5 to HT-8. In that setup 12MBA would elute earlier (see ESI†) and could be separated in the second dimension from 2MC and 5MC.

When applying DB-XLB in the second dimension after DB-5 one can estimate a concentration of 6MC; it will coelute with 5MC on DB-XLB but 5MC can be quantified on a 60 m DB-5 60 column.

There are five additional homologues of methylated PAHs with mass 242: three methylated naphthacenes and two methylated triphenylenes. However, standards of these compounds are commercially not available so the identification of these compounds was not possible in this study. It is probable though that some of these compounds are present in the River Elbe sediment because of the presence of peaks with mass spectra that are similar to the spectra of MC, MBA or MBP. Furthermore naphthacene and triphenylene are present in the sediment (Fig. 2) which suggests the possible presence of their methylated homologues.

Conclusions

In this study we have demonstrated that the relative response factors of MBAs, MBPs and MCs range from 0.1 for 12MBA and 4MBP to 1.7 for 6MBA. Quantification of methylated PAHs with equal molecular weights as one group using the same relative response factor can thus overestimate or underestimate their concentrations and, therefore, the toxicological risk of an environmental sample. GC×GC-MS allows the separation and quantification of fourteen methylated homologues, most of which coelute in one dimension.

Acknowledgements

This study was supported by the European Union-funded project Models for Assessing and Forecasting the Impact of Environmental Key Pollutants on Marine and Freshwater Ecosystems and Biodiversity (MODELKEY; contract 511237-GOCE). We gratefully acknowledge the scientific help of Dr Miroslav Machala (VeterinaryResearch Institute, Brno, Czech Republic) and Dr Peter Korytár (European Commission Brussels, Belgium).

References

1. W. Jira, K. Ziegenhals and K. Speer, Food Addit. Contam., Part A, 2008, 25, 704–713.
2. S. Kaisarevic, U. Lubcke-von Varel, D. Orcic, G. Streck, T. Schulze, K. Pogrmic, I. Teodorovic, W. Brack and R. Kovacevic, Chemosphere, 2009, 77, 907–913.
3. I. Tolosa, M. Mesa-Albernas and C. M. Alonso-Hernandez, Mar. Pollut. Bull., 2009, 58, 1624–1634.
4. M. Saha, A. Togo, K. Mizukawa, M. Murakami, H. Takada, M. P. Zakaria, N. H. Chiem, B. C. Tuyen, M. Prudente, R. Boonyatumanond, S. K. Sarkar, B. Bhattacharya, P. Mishra and T. S. Tana, Mar. Pollut. Bull., 2009, 58, 189–200.
5. J. J. Lian, Y. Ren, J. M. Chen, T. Wang and T. T. Cheng, J. Environ. Monit., 2009, 11, 187–192.
6. S. Diez, J. Sabate, M. Vinas, J. M. Bayona, A. M. Solanas and J. Albaiges, Environ. Toxicol. Chem., 2005, 24, 2203–2217.
7. W. Brack and K. Schirmer, Environ. Sci. Technol., 2003, 37, 3062–3070.
8. W. Brack, K. Schirmer, L. Erdinger and H. Hollert, Environ. Toxicol. Chem., 2005, 24, 2445–2458.
9. J. Vondracek, L. Svihalkova-Sindlerova, K. Pencikova, S. Marvanova, P. Krcmar, M. Ciganek, J. Neca, J. E. Trosko, B. Upham, A. Kozubik and M. Machala, Environ. Toxicol. Chem., 2007, 26, 2308–2316.
10. S. Ahmad, S. L. Kabler, L. Rudd, S. Amin, J. Doehmer, C. S. Morrow and A. J. Townsend, Toxicol. Lett., 2008, 183, 99–104.
11. P. Garrigues, H. Budzinski, M. P. Manitz and S. A. Wise, Polycyclic Aromat. Compd., 1995, 7, 275–284.
12. M. P. Zakaria, H. Takada, S. Tsutsumi, K. Ohno, J. Yamada, E. Kouno and H. Kumata, Environ. Sci. Technol., 2002, 36, 1907–1918.
13. E. Skoczynska, P. Korytar and J. De Boer, Environ. Sci. Technol., 2008, 42, 6611–6618.
14. D. Utesch, H. Glatt and F. Oesch, Cancer Res., 1987, 47, 1509–1515.
15. J. Piskorska Pliszczynska, B. Keys, S. Safe and M. S. Newman, Toxicol. Lett., 1986, 34, 67–74.
16. S. Marvanova, J. Vondracek, K. Pencikova, L. Trilecova, P. Krcmar, J. Topinka, Z. Novakova, A. Milcova and M. Machala, Chem. Res. Toxicol., 2008, 21, 503–512.
17. M. Machala, L. Svihalkova-Sindlerova, K. Pencikova, P. Krcmar, J. Topinka, A. Milcova, Z. Novakova, A. Kozubik and J. Vondracek, Toxicology, 2008, 247, 93–101.
18. S. J. S. Baird, E. A. Bailey and D. J. Vorhees, Hum. Ecol. Risk Assess., 2007, 13, 322–338.
19. J. Tronczynski, C. Munschy, K. Heas-Moisan, N. Guiot, I. Truquet, N. Olivier, S. Men and A. Furaut, Aquat. Living Resour., 2004, 17, 243–259.
20. H. Takada, T. Onda and N. Ogura, Environ. Sci. Technol., 1990, 24, 1179–1186.
21. S. Marvanova, J. Vondracek, K. Pencikova, L. Trilecova, P. Krcmar, J. Topinka, Z. Novakova, A. Milcova and M. Machala, Chem. Res. Toxicol., 2008, 21, 503–512.
22. G. Reifferscheid, S. Buchinger, Z. Cao and E. Claus, Environ. Mol. Mutagen., 2011, 52, 397–408.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c2ay25746g

---