Introduction

Polycyclic aromatic hydrocarbons (PAHs) of the C26H16 group include 37 benzenoid cata-condensed isomers with molecular mass 328 Da listed in Fig. 1 [1]. Reports about the separation and unquestionable identification of the PAHs in this group are scarce. Bemgard et al. [2] reported the gas chromatographic (GC) analysis of the C26H16 cata-condensed PAHs focusing on the evaluation of the performance of four different chromatographic columns for the separation of fourteen isomers and the respective identification of these isomers in a coal tar sample, i.e., Standard Reference Material (SRM) 1597. In that report, the calculation of the chromatographic retention indices allowed the identification of benzo[c]picene and the tentative identifications of benzo[c]pentaphene, dibenzo[f,k]tetraphene, dibenzo[c,k]tetraphene, naphtho[2,3-c]tetraphene, and benzo[b]picene. The retention indices calculations were performed using two series of PAHs as index markers. The first series composed of phenanthrene, chrysene, picene, benzo[c]picene, and dinaphtho[2,1-a:2,1-h]anthracene, and the second series composed of coronene and benzo[a]coronene. In a more comprehensive analysis, Bergvall et al. [3], reported the determination of the PAHs from 252 to 302 Da and the tentative identification of PAHs from 316 to 376 Da in urban dust (SRM 1649a) and diesel particulate matter (SRM 1650 and SRM 2975). An accelerated extraction method followed by qualitative and quantitative analysis using gas chromatography with mass spectrometry (GC/MS) was applied for the characterization of the PAHs. In the studies of Bemgard et al. [2] and Bergvall et al. [3], normal-phase liquid chromatography (NPLC) was used for the isolation of the fraction of interest and GC/MS was used for the identification of the PAHs. Although the two studies provided interesting results regarding the tentative presence of some C26H16 cata-condensed PAHs in a variety of samples, the lack of reference standards and the broad NPLC fractions from the samples precluded unquestionable identification of a significant number of C26H16 cata-condensed PAHs.

Fig. 1
figure 1

The 37 C26H16 cata-condensed benzenoid PAHs ordered with respect to the increase in steric hindrance irregularities. I with no steric hindrance, II to XVI with one or more bay regions, XVII to XXXII with one or more cove regions, XXXIII to XXXVI with one fjord region, and XXXVII with one helical region

Full size image

With the availability of authentic standards, one strategy for the identification of the C26H16 cata-condensed PAHs is use of a combination of NPLC with ultraviolet–visible spectroscopy (NPLC/UV–vis), and GC/MS. Initially, NPLC/UV–vis using a polar stationary phase, e.g., aminopropyl (NH2CH2CH2CH2-) with a nonpolar mobile phase such as hexane-dichloromethane mixtures, allows separation of the PAH isomer groups based on the number of aromatic carbons [4–8]. GC/MS then provides the separation and molecular mass determination for identification of the isomers through comparison of the elution times of the unknown PAHs with elution times of authentic standards of C26H16 cata-condensed PAHs. This method is analogous to the approach used in the characterization of PAHs with molecular mass up to 302 Da in two SRM air particulate samples [8]. In that report, NPLC/UV–vis was used for the separation of eight fractions and the combination of GC/MS and reversed-phase liquid chromatography with UV–vis was used for the separation, identification, and quantification of the PAHs. Using this approach, 55 unsubstituted PAHs and polycyclic aromatic sulfur heterocycles (PASHs) were identified and 75 unsubstituted PAHs and PASHs were characterized based on molecular mass only [8].

The sample under investigation in this report is a complex mixture of PAHs from coal tar for which mass fractions of selected PAHs have been assigned and which is disseminated as SRM 1597a [9–11] (an updated version of the above mentioned SRM 1597). For SRM 1597a, certified values for 34 PAHs, reference values for 36 PAHs, and reference values for 10 PASHs are reported in the Certificate of Analysis [9]. In spite of the comprehensive characterization of SRM 1597a, there are still numerous PAHs that have not been identified or quantified in this material. Thus, the interest in the application of new experimental methods for the qualitative and quantitative elucidation of additional PAHs in this mixture remains active.

The steric hindrance or nonplanar geometry of PAHs is linked to the mutagenic/carcinogenic potential of PAHs [12]. Among the 37 C26H16 cata-condensed PAHs in Fig. 1, 16 include a cove region (XVII, XVIII, XIX, XX, XXI, XXII, XXIII, XXIV, XXV, XXVI, XXVII, XXVIII, XXIX, XXX, XXXI, XXXII), 4 include a fjord region (XXXIII, XXXIV, XXXV, XXXVI), and 1 includes a helical region (XXXVII). However, only 14 out of these 21 nonplanar structures have procedures of synthesis reported in the literature (i.e., XVIII, XX, and XXXVII [13]; XVII [14]; XIX in unreported results from W Schmidt; XXI and XXV [15]; XXII [16]; XXIII and XXXIII [17]; XXIV, XXVI, and XXXIV [18]; and XXXV [19]), and only 1 of these 21 nonplanar structures, XVII, has the carcinogenic potential evaluated and demonstrated [14, 20]. Consequently, it is evident that only the identification of C26H16 cata-condensed PAHs in actual samples will encourage the syntheses of the other 7 remaining nonplanar PAHs (XXVII, XXVIII, XXIX, XXX, XXXI, XXXII, XXXVI) and the evaluation of the biological behavior of the 20 untested nonplanar structures. For this study, reference standards of the following nineteen C26H16 cata-condensed PAHs were available: I, II, III, IV, V, VI, VII, VIII, IX, X, XI, XII, XIII, XIV, XVII, XVIII, XIX, XX, and XXXVII. In parallel theoretical research efforts, the characterization of the aromatic sextets [21] and the prediction of the locations of maximum absorbance in the UV–vis spectra [22] of cata-condensed PAHs, i.e., C26H16, were analyzed so as to have a comprehensive view of the behavior of these PAHs.

In this study we identified and quantified the presence of 13 C26H16 cata-condensed isomers in a coal tar extract (SRM 1597a): benzo[c]pentaphene, II; benzo[c]picene, III; dibenzo[f,k]tetraphene, IV; benzo[h]pentaphene, V; dibenzo[a,l]tetracene, VI; dibenzo[c,k]tetraphene, VII; naphtho[2,3-c]tetraphene, VIII; dibenzo[a,c]tetracene, IX; benzo[b]picene, XI; dibenzo[a,j]tetracene, XII; naphtho[2,1-a]tetracene, XIII; dibenzo[c,p]chrysene, XVII; and dibenzo[a,f]tetraphene, XVIII. As far as we know, this is the first time that III has been quantified and the first time that II, IV, V, VI, VII, VIII, IX, XI, XII, XIII, XVII, and XVIII have been identified and quantified in an environmental combustion sample.

Materials and methods

Materials

Chemicals

SRM 1597a was obtained from the Office of Reference Materials, National Institute of Standards and Technology, Gaithersburg, Maryland. The C26H16 cata-condensed PAH standards were obtained from the following sources: I from W Schmidt, Institut für PAH-Forschung, Greifenberg, Germany and TJ Chow, Department of Chemistry, National Taiwan Normal University, Taipei, Taiwan; II, III, IV, V, VII, VIII, IX, X, XI, XIII, XIV, XVIII, XIX, XX, and XXXVII from W Schmidt, Institut fur PAH-Forschung, Greifenberg, Germany; and VI, XII, and XVII from AK Sharma, Penn State University, College of Medicine, Department of Pharmacology, Hershey, Pennsylvania. The purity of these compounds was not assessed. The C24H14 peri-condensed PAH standards used in this study were from the same sources as reported by Schubert et al. [23]. Hexanes (99.9 %) was purchased from Fisher Scientific (Fair Lawn, NJ, USA). Dichloromethane (99.8 %) was purchased from JT Baker (Center Valley, PA, USA).

Instrumentation and chromatographic columns

NPLC/UV–vis was performed using a Varian 9012 Solvent Delivery System (Agilent, Santa Clara, CA) coupled to Jasco UV-1570 Intelligent UV/Vis Detector (Jasco, Easton, MD) using a Waters Spherisorb 5 μm NH2 10 × 250 mm semi-prep LC column (Waters, Milford, MA). GC/MS analyses were performed on an Agilent 6890 N Network Gas Chromatograph coupled to Agilent 5973 Inert Mass Selective Detector (Agilent, Santa Clara, CA) using a Restek Rxi-PAH GC Column (Restek, Bellefonte, PA), 60 m length, 0.25 mm id, 0.10 μm film thickness, maximum programmable temperature 360 °C, and minimum bleed at 350 °C, and using an Agilent DB-5 GC Column (Agilent, Santa Clara, CA), 5 % phenylmethylpolysiloxane stationary phase, 60 m length, 0.25 mm id, 0.25 μm film thickness, maximum programmable temperature 350 °C, and minimum bleed at 325 °C.

Methods

Qualitative analysis: fractionation

For the NPLC fractionation, 1 mL of the original sample of SRM 1597a was concentrated by evaporation to 0.5 mL and solvent-exchanged from toluene (the original solvent in SRM 1597a) to the solvent mixture hexane/dichloromethane (90/10 vol/vol). Next, the sample solution (0.5 mL) was added to 1 mL of the hexane/dichloromethane mixture so as to have the concentration of the sample at the maximum workable level for the fractions of interest, i.e., just below the overloading limit of the LC column during the NPLC/UV–vis analysis. The NPLC fractionation consisted of injection of five aliquots of 0.25 mL of the sample solution on the amino semi-prep column using a mobile phase of hexane/dichloromethane (80/20 vol/vol) at a flow rate of 2 mL/min. The following six fractions were collected: F0 (from 14.7 to 15.4 min), F1 (from 15.4 to 17.0 min), F2 (from 17.0 to 19.2 min), F3 (from 19.2 to 20.7 min), F4 (from 20.7 to 22.7 min), and F5 (from 22.7 to 25.7 min). Finally, the collected volumes were concentrated by evaporation to individual volumes of 0.5 mL.

Qualitative analysis: identification

The concentrated fractions were analyzed to determine the presence of C26H16 cata-condensed PAH isomers using GC/MS with on-column injection and the mass spectrometer programmed in selected-ion monitoring (SIM) mode for the m/z 328 ion. The GC was temperature programmed as follows: isothermal at 140 °C for 2 min then 20 °C/min to 325 °C, and isothermal at 325 °C for 90 min. The identification of the C26H16 cata-condensed PAHs was accomplished by matching the retention times of the reference standards with the retention times of the unknown compounds in the GC/MS analysis of the fractions.

Quantitative analysis

The quantification of the identified C26H16 cata-condensed PAHs in SRM 1597a was performed using the mass fractions reported in the Certificate of Analysis for SRM 1597a [9] for coronene, dibenzo[a,l]pyrene, and dibenzo[a,h]pyrene as reference internal standards. The detector response was assumed to be equivalent for all of the molecular mass 328 Da isomers and equivalent to the response for the reference internal standards. These results should be considered as semi-quantitative using this approach.

Results and discussion

Selection of GC column for the separation of the C26H16 cata-condensed PAHs

The Rxi-PAH and DB-5 GC columns were evaluated for this analytical study due to the high thermal limit of their stationary phases (i.e., 360 and 350 °C) which is necessary for the elution of high-molecular-mass PAHs. The Rxi-PAH column is a proprietary phase with retention behavior for PAHs similar to a 50 % phenylmethylpolysiloxane phase, and the DB-5 is a 5 % phenylmethylpolysiloxane phase. The GC/MS chromatograms for the separation of 17 C26H16 cata-condensed PAHs using the Rxi-PAH and the DB-5 columns are shown in Fig. 2a and b illustrating a similar behavior for the two columns and demonstrating the resolving power of the phenylmethylpolysiloxane stationary phases for the high-molecular-mass PAH isomers investigated in this study. The C26H16 cata-condensed PAHs hexacene (I) and benzo[a]pentacene (X) were available; however, they are not shown in Fig. 2a and b due to their low solubility in dichloromethane solution and high photo-reactivity.

Fig. 2
figure 2

GC/MS analyses of the 17 C26H16 cata-condensed PAH standards available with SIM for the m/z 328 ion in the following columns: a Rxi-PAH, b DB-5

Full size image

Calculation of retention indices

The retention indices for the C26H16 cata-condensed PAH standards were calculated using the original work of Lee et al. [24] based on the series naphthalene (two rings), phenanthrene (three rings), chrysene (four rings), and picene (five rings). Since the retention time of all of the C26H16 cata-condensed PAH standards available exceeds the retention time of picene in the series of Lee, an extended series using benzo[c]picene (six rings) was used. The equation for the calculation of the retention index is the following [24]:

$$ I=100\ \frac{T_{R(substance)} - {T}_{R\left({C}_z\right)}}{T_{R\left({C}_{z+1}\right)} - {T}_{R\left({C}_z\right)}}+100\ z $$

Where, I = retention index, T R(substance) = retention time of the substance under investigation, \( {T}_{R\left({C}_z\right)} \) and \( {T}_{R\left({C}_{z+1}\right)} \) = retention times of the PAH standards, e.g., picene and benzo[c]picene, bracketing the substance under investigation, and z = number of carbon atoms in the PAH standard eluting immediately before the substance under investigation. The calculated retention indices for the 17 C26H16 PAHs standards available (i.e., II, IV, V, VI, VII, VIII, IX, X, XI, XII, XIII, XIV, XVII, XVIII, XIX, XX, and XXXVII) using the Rxi-PAH and DB-5 GC columns are presented in Table 1. The study of Bemgard et al. [2] compared four GC columns, i.e., Rtx-200 with a trifluoropropylmethylpolysiloxane phase, CP-Sil 8 with a 5 % diphenyl, 95 % dimethyl phase, XTI-5 with a 5 % phenyl phase, and a silabiphenyl phase (silarylene/siloxane copolymer) with 25 % biphenyl. At the time of the Bemgard et al. [2] study, these four columns represented the typical nonpolar 5 % phenylmethylpolysiloxane-like phases used for PAH separations. The Rxi-PAH and the DB-5 columns provide elution of the C26H16 PAH over 20 and 10 min (excluding XXXVII), respectively, whereas they elute in about a 5 min range on the columns in the Bemgard study. All of the index standards elute between picene (retention index 500) and benzo[c]picene (retention index 600). Therefore, only two retention index standards (i.e., picene and benzo[c]picene) were necessary for calculation of the retention indices for the Rxi-PAH and DB-5 columns in comparison to three index standards (i.e., picene, benzo[c]picene, and dinaphtho[2,1-a:2,1-h)anthracene) used by Bemgard et al. [2] to calculate retention indices for the Rtx-200, CP-Sil 8, and XTI-5 columns. These are the first reported retention indices for the C26H16 PAH on a 50 % phenyl-like stationary phase.

Table 1 Average retention indices and standard deviations for the seventeen C26H16 cata-condensed PAHs with standards available
Full size table

Qualitative analysis

Although there is a similar chromatographic behavior between the Rxi-PAH and the DB-5 columns, the C26H16 PAHs elute over a longer time span for the Rxi-PAH column (20 min) than for the DB-5 (10 min), and the separation of the PAHs pairs VI and VII and VIII and XII is accomplished only with the Rxi-PAH column (see Fig. 2a and b); therefore, only the Rxi-PAH column was used for the analyses (qualitative and quantitative). As shown in Fig. 3, when no NPLC fractionation is involved, the GC/MS chromatogram with SIM for the m/z 328 ion indicates the presence of a very limited number of C26H16 cata-condensed PAHs with molecular mass 328 Da. Only two C26H16 cata-condensed PAHs (II and XI) were identified based on comparison of retention times with the 17 C26H16 cata-condensed PAH reference standards illustrated in Fig. 2a, and two additional isomers were characterized based only on molecular mass. The remaining peaks in the chromatogram belong to either PAHs with lower molecular mass, e.g., 326 Da, which generate a small 328 ion or to co-eluting PAHs (e.g., alkyl-PAHs of lower molecular mass) that elute in the same time range as the C26H16 cata-condensed PAHs. Consequently, the identification and subsequent quantification of C26H16 cata-condensed PAHs based on direct GC/MS analysis of SRM 1597a is not viable.

Fig. 3
figure 3

GC/MS chromatogram of the unfractionated SRM 1597a with SIM for the m/z 328 ion

Full size image

Due to the limitations of direct GC/MS analysis, an approach involving an NPLC fractionation of the coal tar mixture was investigated. The NPLC/UV–vis chromatogram from the fractionation of SRM 1597a is illustrated in Fig. 4. Six fractions were collected (F0, F1, F2, F3, F4, and F5) as described in the Materials and methods section. Prior to the qualitative and quantitative characterization of the unknown C26H16 cata-condensed PAHs in SRM 1597a, the elution behavior of the C26H16 cata-condensed PAH reference standards was studied using NPLC/UV–vis so as to locate the particular fraction in which each of these standards elute. The results indicated that XVII, XX, and XXXVII elute in F1; XVIII and XIX elute in F2; and II, III, IV, V, VI, VII, VIII, IX, XI, XII, XIII, and XIV elute in F4. There was no indication of the presence of C26H16 cata-condensed PAHs in F0, F3, and F5 (i.e., negligible m/z 328 response or no elution of available standards); therefore, F0, F3, and F5 were disregarded as retention time intervals of interest for the identification of C26H16 cata-condensed PAHs. The distribution of the detected ions in F0, F1, F2, F3, F4, and F5 using scan mode (from 278 to 374 m/z) is summarized in Table 2.

Fig 4
figure 4

NPLC/UV–Vis chromatogram of the fractionation of the concentrated SRM 1597a in six consecutive fractions, F0 (from 14.7 to 15.4 min), F1 (from 15.4 to 17.0 min), F2 (from 17.0 to 19.2 min), F3 (from 19.2 to 20.7 min), F4 (from 20.7 to 22.7 min), and F5 (from 22.7 to 25.7 min)

Full size image
Table 2 Distribution of the detected ions in F0, F1, F2, F3, F4, and F5 using scan mode (from 278 to 374 m/z)
Full size table

Among these fractions, Fraction 1 (F1), Fraction 2 (F2), and Fraction 3 (F3) were also analyzed as retention time intervals of interest for the identification of C24H14 (F1 and F2) and C26H14 (F3) peri-condensed PAHs, i.e., molecular mass 302 and 326 Da. GC/MS analyses of F1 with SIM for the m/z 302 ion is illustrated in Fig. 5a with peaks identified based on retention time of authentic standards as dibenzo[b,e]fluoranthene, naphtho[1,2-e]pyrene, dibenzo[a,l]pyrene, naphtho[1,2-a]pyrene, and benzo[a]perylene (See Electronic Supplementary Material (ESM) Fig. S3 for structures and names of molecular mass 302 Da PAH isomers). GC/MS analysis of F2 with SIM for the m/z 302 ion is shown in Fig. 5b with the following peaks identified based on retention time of authentic standards: naphtho[1,2-b]fluoranthene, naphtho[1,2-k]fluoranthene, naphtho[2,3-j]fluoranthene, naphtho[2,3-b]fluoranthene, dibenzo[b,k]fluoranthene, dibenzo[a,k]fluoranthene, dibenzo[j,l]fluoranthene, naphtho[2,3-k]fluoranthene, naphtho[2,3-e]pyrene, dibenzo[a,e]pyrene, naphtho[2,1-a]pyrene, dibenzo[e,l]pyrene, naphtho[2,3-a]pyrene, benzo[b]perylene, dibenzo[a,i]pyrene, and dibenzo[a,h]pyrene. Although the focus of this study was not the molecular mass 302 Da PAHs, the NPLC fractionation provides some new information on the characterization of the coal tar SRM relative to these isomers. As illustrated in Fig. 5a and b, the 302 molecular mass isomers are distributed in both F1 and F2. Even though the NPLC fractionation of PAHs is based on the number of aromatic carbons in the PAHs, isomers with nonplanar structures typically elute earlier than the planar isomers [4, 7]. Based on the NPLC retention data published previously for molecular mass 302 Da PAH isomers [7] and the characterization of the location of the various 328 molecular mass in the selected fractions mentioned above, dibenzo[a,l]pyrene, benzo[a]perylene, and dibenzo[b,e]fluoranthene have retention indices of 4.60, 5.12, and 5.36, respectively, and therefore they would elute earlier than the other molecular mass 302 Da isomers with NPLC retention indices ranging from 5.90 to 6.42. NPLC retention indices were not published for naphtho[1,2-e]pyrene and naphtho[1,2-a]pyrene since standards for these isomers were not available in the earlier study [7]. However, the five 302 Da molecular mass isomers identified in F1 are nonplanar based on the thickness values published by Sander and Wise [1] and later in Schubert et al. [23], with thickness values of 4.64 Å, 5.13 Å, 5.15 Å, 5.17 Å, and 5.47 Å for dibenzo[b,e]fluoranthene, naphtho[1,2-a]pyrene, naphtho[1,2-e]pyrene, dibenzo[a,l]pyrene, and benzo[a]perylene, respectively, compared to thickness values of typically 3.9 Å for most of the isomers in F2. The results of the qualitative and quantitative analysis of the C26H14 peri-condensed PAHs in F3 will be published elsewhere [25].

Fig. 5
figure 5

GC/MS analyses with SIM for the m/z 302 ion of the following samples: a Fraction 1 (F1), b Fraction 2 (F2). The names of the structures are given in Figure S3 of the supporting information

Full size image

The GC/MS analyses of F1 with SIM for the m/z 300 ion showing the reference internal standard of coronene and F1 with SIM for the m/z 302 ion are shown in Fig. 6b and c, respectively. Also, the GC/MS analysis of F1 with SIM for the m/z 328 ion is shown in Fig. 6d and it is compared with the GC/MS analysis of the 17 C26H16 cata-condensed benzenoid PAH reference standards illustrated in Fig. 6a. Furthermore, the GC/MS analyses of F2 with SIM for the m/z 302 ion, F2 with SIM for the m/z 328 ion, and F4 with SIM for the m/z 328 ion are shown in Fig. 7b, c, and d and are compared with the GC/MS analysis of the 17 C26H16 cata-condensed benzenoid PAH reference standards illustrated in Fig. 7a.

Fig. 6
figure 6

GC/MS analyses of the following samples: a The 17 C26H16 cata-condensed PAH standards available with SIM for the m/z 328 ion, b Fraction 1 (F1) with SIM for the m/z 300 ion, c Fraction 1 (F1) with SIM for the m/z 302 ion, d Fraction 1 (F1) with SIM for the m/z 328 ion. The names of the structures are given in Figure S3 of the supporting information

Full size image
Fig. 7
figure 7

GC/MS analyses of the following samples: a The 17 C26H16 cata-condensed PAH standards available, b Fraction 2 (F2) with SIM for the m/z 302 ion, c Fraction 2 (F2) with SIM for the m/z 328 ion, d Fraction 4 (F4) with SIM for the m/z 328 ion. The names of the structures are given in Figure S3 of the supporting information

Full size image

Fraction 1, F1

As shown in Fig. 6d, the GC/MS chromatogram with SIM for the m/z 328 ion of F1 is composed of 10 unknown peaks A, B, C, D, E, F, G, H, I, and J. Of the 10 unknown peaks observed in F1, peak F is identified by matching its observed retention time with the retention time of XVII. The peak J belongs to a C26H14 peri-condensed PAHs with molecular mass 326 Da, which generates a small 328 ion, and the unknown peaks A, B, C, D, E, G, H and I could not be identified based on retention times compared with standards. A retention time match between G and H, with XVIII is disregarded because XVIII elutes in NPLC in the retention time interval of fraction F2 and not in the retention time interval of fraction F1. Most of the C26H16 isomers for which authentic standards were not available (i.e., XXII through XXXVI) are nonplanar structures with thickness values of between 5.23 Å and 6.20 Å, [1] these nonplanar isomers would be expected to elute in F1 or F2, if they are present in the coal tar. Therefore, it is highly likely that peaks A-I (except F) in F1 could be some of these nonplanar isomers.

Fraction 2, F2

As illustrated in Fig. 7c, the GC/MS chromatogram with SIM for the m/z 328 ion of F2 is composed of 52 unknown peaks K-Z, AA-AZ, and BA-BJ. Of the 52 unknown peaks observed in F2, peak AJ is identified by matching its retention time with the retention time of XVIII. In addition, there is no retention time match between XIX, which elutes in F2 as it was stated, and any of the 51 remaining peaks. The peaks AP, AU, AW, AX, AY, AZ, BA, BB, BC, BD, BE, BF, BG, and BJ belong to C26H14 peri-condensed PAHs with molecular mass 326 Da that generate a small 328 ion. The other unknown peaks K, L, M, N, O, P, Q, R, S, T, U, V, W, X, Y, Z, AA, AB, AC, AD, AE, AF, AG, AH, AI, AK, AL, AM, AN, AO, AQ, AR, AS, AT, AV, BH, and BI, were not identified because their retention times do not match with any of the retention times of the other C26H16 cata-condensed PAH reference standards. The retention time matches between K with XXXVII, AC with XX, AH with XVII, and AT with IV are disregarded because XXXVII, XX, and XVII elute in NPLC in the retention time interval of fraction F1, and IV elutes in the retention time interval of fraction F4. As mentioned above, some of these unknown peaks could be the nonplanar isomers for which no authentic standards were available.

Fraction 4, F4

The GC/MS chromatogram from the analysis of F4, which contains most of the C26H16 PAH, is shown in Fig. 7d. Of the 13 unknown peaks: BK, BL, BM, BN, BO, BP, BQ, BR, BS, BT, BU, BV, and BW observed in F4, 9 peaks are identified based on retention times according to the following assignments: BM=IV, BN=V, BP=II, BQ=IX, BR=VI in co-elution with VII, BS=XII, BT=VIII, BU=XI, and BV=III in co-elution with XIII. In addition, this result is verified by the above mentioned NPLC elution analysis of the C26H16 standards that indicates IV, V, II, IX, VI, VII, XII, VIII, XI, III, and XIII elute in F4. The peak BW belongs to a C26H14 peri-condensed PAH with molecular mass 326 Da that generates a small 328 ion. The unknown peaks BK, BL, and BO are not identified because their retention times do not match any of the retention times of the other C26H16 cata-condensed PAH reference standards. However, based on planarity, [1] two of the three unknown isomers would likely be XV, thickness 3.9 Å, and XVI, thickness 3.92 Å. Fraction 4, F4, contains the largest number of C26H16 cata-condensed PAH isomers identified among the six fractions analyzed.

Quantitative analysis

The preferred quantitative approach using an internal standard added to SRM 1597a prior to the analytical procedure was not feasible because of the lack of suitable internal standard PAHs (e.g., isotopically labeled PAHs). Therefore, the following alternative approaches, which are semi-quantitative, were used for the determination of C26H16 cata-condensed PAHs in F1 and F4. Coronene and dibenzo[a,l]pyrene, which were previously quantified in SRM 1597a [9–11], and which elute in fraction F1 were used as reference internal standards for quantification. For the quantification of C26H16 isomers in F2 and F4, known quantities of F1 containing coronene and dibenzo[a,l]pyrene were added to F2 and F4 to facilitate the quantification. The details of the semi-quantitative approach are presented in the ESM (Figs. S1 and S2). A second direct analysis GC/MS approach was compared using the dibenzo[a,h]pyrene (mass fraction previously determined) as an internal reference standard for the determination of XVIII in F2. The results for the calculated mass fractions (mg/kg) are summarized in Table 3 for 13 C26H16 cata-condensed PAHs. The mass fractions for the 12 PAH isomers of molecular mass 328 Da determined using dibenzo[a,l]pyrene as the internal standard are a factor of 2.2 times higher than when using coronene as the internal reference standard. Using the direct GC/MS approach (no NPLC fractionation) (see Fig. 3) with benzo[ghi]perylene (mass fraction 50.5 ± 0.6 mg/kg reported in the Certificate of Analysis for SRM 1597a [9]) as an internal reference standard to quantify II provided a mass fraction of 0.76 mg/kg compared to the value of 0.75 mg/kg ± 0.13 mg/kg obtained using coronene as the internal reference value following NPLC fractionation. This result indicates that the lower mass fraction obtained using coronene as the internal standard may be more reliable than those using dibenzo[a,l]pyrene. Determination of dibenzo[a,l]pyrene in F1 using coronene as the internal standard provides a mass fraction of 0.53 ± 0.14 mg/kg (n = 5) compared with the 1.12 ± 0.17 mg/kg reported in the Certificate of Analysis [9], which was determined by GC/MS analysis on a 50 % phenylmethylpolysiloxane stationary phase with SIM for the m/z 302 ion as described by Schubert et al. [23]. As described above, the NPLC fractionation places dibenzo[a,l]pyrene in F1 and avoids the problems associated with quantifying a small chromatographic peak eluting closely to other larger isomers.

Table 3 Calculated mass fractions (mg/kg) of the thirteen identified C26H16 cata-condensed PAHs II, III + XIII, IV, V, VI + VII, VIII, IX, XI, XII, XVII, and XVIII using coronene, dibenzo[a,l]pyrene, and dibenzo[a,h]pyrene as reference internal standards
Full size table

As shown in Table 3, the mass fractions for the 328 Da molecular mass isomers range from 0.042 mg/kg (IX) to 0.93 mg/kg (IV). The most abundant 328 molecular mass isomers are more than 10 times lower than the mass fraction of the most abundant molecular mass 302 Da PAHs, 100 times lower than benzo[a]pyrene, and 1000 times lower than naphthalene. Intrinsic in these results is the tendency of the combustion reaction to favor the formation of slightly or non-distorted molecules, e.g., IV (molecular thickness = 4.138 Å [1]), instead of highly distorted molecules, e.g., XVII (molecular thickness = 5.615 Å [1]).

Conclusions

The accurate identification and quantification of C26H16 cata-condensed PAHs requires isolation and enrichment of the isomer fraction using NPLC prior to GC/MS analysis. This is the first report of the identification and quantification of 13 C26H16 cata-condensed PAHs isomers in an environmental combustion sample (coal tar) including benzo[c]pentaphene (II), dibenzo[f,k]tetraphene (IV), benzo[h]pentapheme (V), dibenzo[a,l]tetracene (VI), dibenzo[c,k]tetraphene (VII), naphtho[2,3-c]tetraphene (VIII), dibenzo[a,c]tetracene (IX), benzo[b]picene (XI), dibenzo[a,j]tetracene (XII), naphtho[2,1-a]tetracene (XIII), dibenzo[c,p]chrysene (XVII), dibenzo[a,f]tetraphene (XVIII), and benzo[c]picene (III).

The experimental procedure shown here is adaptable for the qualitative and quantitative analysis of large PAHs, e.g., PAHs with molecular masses exceeding 300 Da. As it is shown in Table 2, GC/MS signals with up to 352 Da were detected in the NPLC separation range under analysis (14.7 min to 25.7 min). In practice, the NPLC separation range can be extended for the identification and quantification of larger compounds depending on the focus of the analysis. However, the high volatility and the low concentration of large PAHs, the general low ratio of the number of reference standards available to the number of theoretical isomers for large PAH families, and the weak stabilities of GC stationary phases above 350 °C would preclude the prolongation of the fractionation to longer NPLC elution times. In that scenario, the isomer families viable for NPLC/UV–vis with GC/MS would include the C26H14, C28H14, and C28H16. The C26H14, C28H14, and C28H16 families show a noticeable percentage of reference standards available (56, 63, and 39 %, respectively) [1] that would provide insightful information about the composition of large PAHs in complex environmental samples.