Fast analysis of 29 polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs with ultra-high performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry

Polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs are ubiquitous in the environment. Some of them are probable carcinogens and some are source markers. This work presents an ultra-high performance liquid chromatography-atmospheric pressure photoionization-tandem mass spectrometry (UHPLC-APPI-MS/MS) method for simultaneous analysis of 20 PAHs and nine nitro-PAHs. These compounds are separated in 15 minutes in the positive mode and 11 minutes in the negative mode, one half of GC/MS analysis time. Two pairs of precursor/product ions are offered, which is essential for confirmation. This method separates and quantifies benzo[a]pyrene (the most toxic PAHs) and non-priority benzo[e]pyrene (isomers, little toxicity) to avoid overestimation of toxin levels, demonstrating its importance for health-related researches. With 0.5% 2,4-difluoroanisole in chlorobenzene as the dopant, limits of detection of PAHs except acenaphthylene and those of nitro-PAHs except 2-nitrofluoranthene are below 10 pg and 3 pg, respectively, mostly lower than or comparable to those reported using LC-related systems. The responses were linear over two orders of magnitude with fairly good accuracy and precision. Certified reference materials and real aerosol samples were analyzed to demonstrate its applicability. This fast, sensitive, and reliable method is the first UHPLC-APPI-MS/MS method capable of simultaneously analyzing 29 environmentally and toxicologically important PAHs and nitro-PAHs.

dant in the ambient air, or with toxicological significance. Besides, feasibility of LC-APPI-MS/MS analysis is considered. All priority PAHs were selected except naphthalene, due to its high vapor pressure. It was shown that GC-MS is a better analytical choice for naphthalene 19,30,31 [1,2,3-cd]pyrene (IND). In addition to 15 priority PAHs, five other PAHs were chosen for reasons given below. Benzo[e]pyrene (BEP) is used as a reference PAH to assess temporal variability and degradation patterns of other PAHs in environmental media due to its stability 3 ; while cyclopenta[cd]pyrene (CPP) is classified as a probable human carcinogen 8 . Both were frequently analyzed in environmental studies. Moreover, coronene (COR) and benzo [b]naphtho [1,2-d] thiophene (BNT) were proposed as source markers for gasoline emissions and traffic emissions, respectively [32][33][34] . Lastly, retene (RET) is a known marker for biomass burning 33,35 . These markers are important in source apportionment studies 34 .
The MS/MS parameters were optimized by manual tuning to obtain the best response signals via ramping various electric potentials. Standard solutions with 0.5 μ g/mL concentrations in acetonitrile were infused into the MS at 20 μ L/min. The scan type was "Multiple Reaction Monitoring" (MRM). The polarity was positive for 20 PAHs and six nitro-PAHs and negative for the other three nitro-PAHs ( Table 1). The Q1 and Q3 resolutions were "unit" (0.7 ± 0.1 amu). The dwell time for each mass was 50 ms. For the infusion experiments, all scan parameters were the same except that the dwell time was 200 ms.
Column Separations and Analysis. For determination of PAHs and nitro-PAHs, good chromatographic separation is essential to differentiate isomeric compounds in the MS owing to their nearly identical fragmentation. Different analytical columns and separation conditions were investigated with the aim of achieving a short separation time and high selectivity and sensitivity. Sufficient separation of the target analytes was finally achieved with the conditions presented below.
Chromatographic separation was performed using a Pinnacle DB PAH 100 mm × 2.1 mm × 1.9 μ m UHPLC column (Restek, Bellefonte, PA, USA) connected to an Acquity UPLC BEH C18 VanGuard Pre-column (2.1 mm × 1.7 μ m, Waters). The column oven was maintained at 30 °C. The mobile phase solvents were 100% water (A) and 100% acetonitrile (B) with a flow rate of 300 μ L/min. Dopant was delivered at one tenth of that flow rate. The elution gradient was 50%(A)/50%(B) initially, 100%(B) at 8-15 minutes, and 50%(A)/50%(B) at 15.1-19 minutes in the positive mode with curve 6; while in the negative mode, the elution gradient was 50%(A)/50%(B) initially, 40%(A)/60%(B) at 5 minute, 30%(A)/70%(B) at 10 minute, 100%(B) at 12-15 minutes, and 50%(A)/50%(B) at 15.1-19 minutes with curve 6 except at 10-15 minutes with curve 9. Sample injection volume was 5 μ L. An amount of 5 μ L of standard solutions was injected on column using a 10 μ L loop with the "partial loop with needle overfill" method. LOD is defined as the level with "signal to noise" ratio equal to 3; and the "signal to noise" ratio is calculated based on peak height to peak height comparison.
Field Sampling and Extraction. Aerosol samples were collected at two locations to assess the applicability of this analytical method, National Taiwan University (NTU) in the center of Taipei city and Hua-Lin (HL) station in the downwind mountainous area with an elevation of 400 m. Samples for particulate matters with aerodynamic diameter equal to or less than 2.5 μ m (PM 2.5 ) were collected in daytime (8 am-8 pm) and nighttime (8 pm-8 am) on July 4-11 and August 15-21, 2011. At each location, a high-volume sampler with single stage cascade for PM 2.5 (Tisch Environmental Inc., Cleves, OH, USA) was used at a flowrate of 1.13 m 3 /min with quartz filters (8 in× 10 in, Pall Life Sciences, Ann Arbor MI, USA) pre-baked at 900 °C overnight.
After sampling, one-fourth of the filter samples was spiked with 20 ng of the surrogate standards (the five-PAH isotope standards and 1NP-d9) and ultrasonically extracted with 20 mL of hexane/methylene chloride (1:4) for 30 minutes three times. The extracts were combined, concentrated to 0.5 mL with nitrogen, and purified using Waters HLB cartridges (6 mL volume, 500 mg bed mass) that were pre-conditioned with 6 mL each of methanol, methylene chloride, and hexane. The purified extracts were then filtered with 0.22 μ m porosity PTFE filters (Great Engineering Technology Corp., Taipei, Taiwan), solvent-exchanged to 0.2 mL acetonitrile, spiked with 10 ng of the internal standards (BAA-d12 and MX-d15), and analyzed with the presented UHPLC-APPI/MS/MS method. A seven-point calibration curve was prepared in acetonitrile. Seven matrix spikes were prepared; each with one-fourth of the filter Negative mode samples which were spiked with 20 ng of the target analytes and the surrogate standards. Three laboratory blanks were prepared by spiking 20 ng of the surrogate standards in one-fourth of the pre-baked filters. Matrix spikes and laboratory blanks went through the same analytical procedures as the filter samples. Two batches of CRM were extracted and analyzed in the same way; one batch (n = 7) for PAHs and another batch (n = 7) for nitro-PAHs. For PAHs, 20 mg of CRM was placed in a glass vial, spiked with 20 μ L (1 ng/μ L) of the surrogate standards, mixed by a Vortex mixer (Digisystem laboratory instrutments INC, Taiwan) for 30 seconds, then went through the same pre-treatment and analysis procedures as stated above. For nitro-PAHs, 200 mg of CRM were used following the same procedures.

Results and Discussion
Identification and Separation. The MS/MS parameters were optimized and two pairs of precursor/ product ions of these compounds were obtained (Table 1), except the second product ions of COR,  2NFL, and 3NFL. The signals of the previously-reported second product ions of six PAHs (BAA, BBF, BKF, BAP, IND, DAA) using Acquity tandem MS from Waters Corporation were not strong enough for confirmation in API 3000 22 . Thus, different second product ions are offered. Moreover, for the 15 USEPA priority PAHs and 1NP, the second product ions were also reported in earlier studies 11,23,[25][26][27] , Table 1 provides an alternative set of product ions using a widely used instrument. Furthermore, for BEP, CPP, RET, BNT, and six nitro-PAHs, the second pairs of precursor/product ions in LC-MS/MS are reported for the first time; they are essential for confirmation. Figure 1 shows MRM ion chromatograms of these target analytes and deuterated standards with on-column injection amounts of 300 pg each. The 29 PAHs and nitro-PAHs are well separated and quantified in 15 minutes in the positive mode and 11 minutes in the negative mode. Compared to 45 minutes required for separating these 29 target analytes in GC/MS (data not shown), this UHPLC-APPI-MS/MS method (totally 26 minutes) cut down the analysis time to one half.
If only 15 priority PAHs are analyzed, it takes only 5 minutes using a shorter column (Pinnacle DB PAH 50 mm × 2.1 mm × 1.9 μ m) with 600 μ L/min at 30 °C (data not shown); it is close to 3.5 minutes reported for 16 priority PAHs with Waters Acquity tandem MS 19 . However, in order to separate BEP from the other isomers (BBF, BKF, and BAP), 100 mm long column was used instead, resulting in a slower flow rate (300 μ L/min) and longer analysis time. BEP, not a priority pollutant, is a relatively stable PAH. It is essential to separate and identify these four isomers in real samples since they are frequently found in the air and their carcinogenic potentials differ significantly. BBF and BKF exhibit only 6-14% and 3-10% of BAP's (a human carcinogen 4 ) cancer-causing potential, respectively; while BEP exhibits very little toxicity 3 . Thus, our method provides accurate concentrations for these four isomers compared to other methods concerning only priority PAHs, which may overestimate the actual concentrations of more-toxic PAHs. As a result, the subsequent health risk assessment will be overestimated as well. Table 2 and 3 shows the LODs of 29 target compounds with three different dopant solutions. In general, the best sensitivity is associated with dopant C (0.5% DFA in chlorobenzene). All the LODs of PAHs are below 10 pg except ACPY; all the LODs of nitro-PAHs are below 3 pg except 2NFL. Compared to those with dopants A and B, the LODs with dopant C are much better for ACPY and FLU and are comparable for other PAHs and nitro-PAHs. Further evaluation of linearity, accuracy, and precision were conducted with dopants A and C only.

Limits of Detection with Different Dopants.
In comparison, the LODs of PAHs with dopant C (with the best sensitivity) are all lower than those using GC-EI-MS/MS in the literatures 31 , except ACPY, the one with the highest vapor pressure among the analyzed PAHs (Table 2). Especially for some high-m.w. PAHs, our LODs are one order of magnitude lower. Compared to LODs using other LC-MS and LC-MS/MS systems with other dopants 11,19,[22][23][24][25][26]28,30,39 , the LODs with dopant C are either lower by one order of magnitude or at least comparable. The results demonstrate the advantage and applicability of our analytical method, with LODs lower than or comparable to those previously reported using GC-MS/MS, LC-MS, and LC-MS/MS systems. Furthermore,   8 pg with one exception of 11.3 pg, are lower than or comparable to those using other LC-related and GC-related methods but higher than those using GC-NICI-MS in the literatures (Table 3) 14,15,[40][41][42] . Typically, the lowest concentrations of these individual nitro-PAHs in ambient air are around 0.4-1.0 pg/m 3 8,14,20 . With a typical 24-hr sampling of 1400 m 3 of air and final sample volume of 200 μ L with 5 μ L injection, the final injection amounts are in the range of 14-35 pg. Therefore, our LODs are low enough for nitro-PAH quantification in ambient air samples. Furthermore, it took 25 minutes using GC-NICI-MS to analyze 11 nitro-PAHs including four analytes here 15 ; the last eluted nitro-PAH in their method was 6NCHRY, which eluted in 6.3 minutes using our method. The relatively shorter analytical time makes this presented method an excellent alternative for nitro-PAHs. More importantly, this method analyzes nitro-PAHs simultaneously with PAHs. Linearity, Accuracy, and Precision. The linear ranges of responses were assessed from levels close to LODs to 200 or 500 ng/mL, covering 2 orders of magnitude for most species (Table 4). The R 2 of these linear calibration curves were all greater than 0.995. Moreover, the accuracy and precision were evaluated with repeated injections of standard solutions prepared at two different concentrations (10 ng/mL and 200 ng/mL for most species, Table 4). The accuracy and precision with both dopants A and C are within 8% variability for both low and high concentrations, with few exceptions. In summary, the results show the PAH and nitro-PAH responses were linear over two orders of magnitude with fairly good precision and accuracy with both dopants A and C.

HPLC-APPI-MS/MS [25] d UHPLC-AP-PI-MS [28] e UHPLC-AP-PI-MS/MS
Certified Reference Materials. Based on the above assessment, the analytical method with dopant C has the best sensitivity and a two-order-of-magnitude linear range with good accuracy and precision. Thus, CRM in replicates (n = 7) were analyzed with this method. The obtained concentrations of 20 PAH and nine nitro-PAH species are compared with NIST certified values ( Table 5). The percent difference between the analyzed concentrations and the certified values are all less than 10.7%. The standard deviations (SD) of all analyzed species are mostly comparable with those NIST values. These results show that the combination of the presented extraction and analytical method provides sensitive, specific, and reliable results. Thus, this presented method is suitable for PAHs and nitro-PAHs analysis of urban aerosols.    at NTU and HL were 4.3 ± 2.9 and 4.2 ± 3.1 pg/m 3 , respectively; the corresponding levels in the night time were 9.1 ± 4.3 and 5.7 ± 2.0 pg/m 3 . The total nitro-PAHs at urban site (NTU) in the daytime were in the same ranges as those in the downwind mountainous site (HL); while those levels in the night time were 1.6 times higher than those at HL. In addition, the concentrations of both total PAHs and total nitro-PAHs were higher in the night time compared to those in daytime at both locations, possibly due to the lower boundary layer height in the night time.
For individual species, the detectable percentages (%) are listed in Table 5. Most of the PAHs were 100% above the LODs in the real samples. The % above the LODs was 0 for ACPY, ACP, FLU, 2NFLU, 2NFL, 3NFL, and 6NCHRY. ACPY, ACP, and FLU may be predominantly present in the gaseous phase rather than the aerosol phase in summer time due to their high vapor pressures 1 . The levels of nitro-PAHs in the air are usually lower than those of PAHs with lower % of detectable 36 ; our results were consistent with the previous findings.   The concentrations of individual PAH and nitro-PAH species are shown in Fig. 2; the concentrations of non-detectable were treated as zero. The levels of individual PAH species in the aerosol phase were 0.28-785.2 pg/m 3 . The reported PAH range in ambient air in Taipei was 10-420 pg/m 3 43 ; the highest level obtained in this work is in the same order of magnitude while the lowest level is two orders of magnitude lower, possibly due to our lower LODs than the previous work. The levels of nitro-PAHs in the aerosol phase in this work were 0.19-6.8 pg/m 3 . The reported levels of 1NP, 2NP, and 7NBAA were 1.4-127, 0.8-70, and 0.7-33 pg/m 3 , respectively 20,36 ; the lowest levels in this work are in the same order of magnitude as theirs and the highest levels are one order of magnitude lower than theirs. The blank values were below the LODs except for PHEN (1.1 ± 0.7 pg/m 3 , n = 3). The recovery rates of the spiked surrogates were 85-106%. The presented results were corrected by blank values and recovery rates. The mean recovery rates of these target compounds in matrix spikes (n = 7) were 80.5-125%. No strong matrix effect was observed. The filter samples (n = 6) were also analyzed in duplicates; the precision of duplicates was mostly within 10%. In summary, the above results indicate that this analytical method is well applicable in environmental studies. Figure 2 clearly shows that the concentrations of BEP are in the same range as those of BAP. It is essential to separate and identify both BEP and BAP in the real samples in order to accurately conduct health risk assessment based on the measured concentrations since their toxicities differ substantially. With analytical methods only focusing on the EPA priority PAHs, BEP may be mistakenly taken as BAP  since they are isomers with the same pairs of precursor/product ions. The estimated health risks may be overestimated by two-folds as indicated in this work. With the presented analytical methods, accurate BEP and BAP levels are assessed; health risk assessment can be conducted and the effective control strategy can be formulated accordingly.

Conclusions
The presented method is the first UHPLC-APPI-MS/MS method capable of simultaneously analyzing 29 environmentally and toxicologically important PAH and nitro-PAH species in aerosols. With a Pinnacle DB PAH 100 mm × 2.1 mm × 1.9 μ m UHPLC column and a water/acetonitrile binary mobile phase, the 29 target analytes are separated in 15 minutes in the positive mode and 11 minutes in the negative mode, one half of GC/MS analysis time. In addition, the second pairs of precursor/product ions in LC-MS/MS are reported for these compounds, which is essential for confirmation. For ten compounds, these are reported for the first time. This method separates and quantifies four isomers (BBF, BKF, BAP, and a non-priority BEP) to avoid overestimation of toxin levels; this demonstrates its importance for health-related researches. Furthermore, the best sensitivity is associated with 0.5% DFA in chlorobenzene as the dopant; all LODs of PAHs are below 10 pg except ACPY; all the LODs of nitro-PAHs are below 3 pg except 2NFL. The responses were linear over two orders of magnitude with fairly good accuracy and precision. Certified reference materials and real samples were analyzed to demonstrate its applicability. In summary, a fast, sensitive, and reliable UHPLC-APPI-MS/MS method is presented for 29 environmentally and toxicologically important PAHs and nitro-PAHs, expanding the analysis scope beyond 16 priority PAHs. This method has wide application in health-related air pollution studies.