Atmospheric Distribution of PAHs and Quinones in the Gas and PM 1 Phases in the Guadalajara Metropolitan Area , Mexico : Sources and Health Risk

Polycyclic aromatic hydrocarbons (PAHs) and quinones in the gas phase and as submicron particles raise concerns due to their potentially carcinogenic and mutagenic properties. The majority of existing studies have investigated the formation of quinones, but it is also important to consider both the primary and secondary sources to estimate their contributions. The objectives of this study were to characterize PAHs and quinones in the gas and particulate matter (PM1) phases in order to identify phase distributions, sources, and cancer risk at two urban monitoring sites in the Guadalajara Metropolitan Area (GMA) in Mexico. The simultaneous gas and PM1 phases samples were analyzed using a gas chromatography–mass spectrometer. The lifetime lung cancer risk (LCR) due to PAH exposure was calculated to be 1.7× 10−3, higher than the recommended risk value of 10−6, indicating a potential health hazard. Correlations between parent PAHs, criteria pollutants, and meteorological parameters suggest that primary sources are the main contributors to the Σ8 Quinones concentrations in PM1, while the secondary formation of 5,12-naphthacenequinone and 9,10-anthraquinone may contribute less to the observed concentration of quinones. Additionally, naphthalene, acenaphthene, fluorene, phenanthrene, and anthracene in PM1, suggest photochemical degradation into unidentified species. Further research is needed to determine how these compounds are formed.


Introduction
There is growing concern over breathable particle-bound polycyclic aromatic hydrocarbons (PAHs) due to their potential carcinogenic, mutagenic and immunosuppressant effects on human health [1,2].The PAHs are produced by the incomplete combustion and thermal alteration of organic matter [3].The ambient levels, phase, and molecular structures of the PAHs depend directly on the type of fuel and combustion technology [4].For instance, PAHs of low molecular weight (LMW) of two to four rings predominate in the gas phase and may condense in the particle phase after emission [5].In contrast, PAHs of high molecular weight (HMW) of more than five rings in the structure are typically emitted and found in the particle phase [1,6].
The PAHs may undergo photochemical reactions producing more toxic compounds, such as oxygenated PAHs, which are direct mutagens [7,8].The major derivative compounds observed in the ambient air samples include quinones.Quinones are currently not considered in the international regulations despite their mutagenic properties, which highlights the importance of monitoring their ambient levels [7].The quinones are potent redox active compounds upon deposition within the lung and can undergo enzymatically (e.g., P450/P450 reductase) and non-enzymatically redox cycling, generating superoxide anion radicals.Under biological conditions, they may be converted to hydroxyl radicals ( • OH), which are potent oxidizing agents that may damage essential macromolecules, cause oxidative stress and allergic diseases [9,10].Furthermore, PM 1 bound-PAHs+quinones (BPQ) represent a significant inhalation health hazard to humans as they can penetrate into the bronchial and pulmonary regions of the respiratory system.The PM 1 BPQ deposition has been correlated positively with major damage in mitochondrial DNA replication, protein synthesis and cellular metabolism, which eventually may lead to mutations and cancer [11][12][13].
In urban areas, typical high ambient levels of quinones have been reported for 1,2-naphthoquinone (1,2-NQ), 1,4-naphthoquinone (1,4-NQ), 9,10-phenanthrenequinone (9,10-PQ), and 9,10-anthraquinone (9,10-AQ) [14][15][16][17].Quinones can be either emitted from gasoline and diesel vehicle combustion and formed as a product of gas phase LMW photochemical reactions with atmospheric oxidant species, such as hydroxyl radicals ( • OH), nitrate radicals ( • NO 3 ) and heterogeneous reactions between particulate PAHs and ozone [18].A detailed characterization of quinones and PAHs may help to identify emission sources and quantify population exposure in view of reducing public health risks.However, to date, source apportionment of quinones has been frequently hindered by the lack of simultaneous measurements of quinones and their parents in the gas and particulate matter (PM) phases.Combining monitoring of such phases can help to identify their primary emissions and secondary formation processes.For instance, ultrafine and submicron breathable particles have been used as robust indicators for identifying sources of PAHs and quinones, and pathways for secondary aerosol formation [19,20].
Gas and particle phase PAHs have been studied extensively, while quinones in the gas phase or as PM have received less consideration.Moreover, few studies have monitored simultaneously quinones and PAHs ambient levels in the gas and particle phases [16,21,22], while most of the existing studies have focused on quinones formation excluding the source apportionment.To date, no study has addressed PAHs and quinones sources by monitoring simultaneously the gas and PM 1 phases.Therefore, the identification of PAHs secondary oxidation processes and their behavior in the atmosphere under different meteorological can help to uncover possible reaction pathways and clarify the mechanisms for PAHs reactions and their quinones products.In the Guadalajara Metropolitan Area (GMA), Mexico, only the presence of quinones in PM 2.5 has been addressed to date [23], which highlights the importance of generating information that contributes to the knowledge of the PAHs' and quinones' ambient levels and behavior in the gas and submicron phases [16].In order to establish phase distributions, identify the influence of primary and secondary sources, and estimate cancer risk from the occurrence of PAHs and quinones, we characterized simultaneously PAHs and quinones in the gas and PM 1 phases at two urban monitoring sites in the GMA.

The Guadalajara Metropolitan Area (GMA)
The GMA is the second most populous City in Mexico with about 4.5 million inhabitants and covers an approximate surface area of 2239 km 2 .It is located around 500 km W of Mexico City (20 • 39 54 N, 103 • 18 42 W), and lies at an average altitude of 1540 m a.s.l. in the Atemajac Valley and the Tonala Plain, surrounded to the north-east by the Rio Grande Santiago Canyon.A volcanic range located south of the GMA constitutes a natural physical barrier preventing wind circulation, which causes air masses' stagnation [24,25].Due to the GMA latitude, solar radiation (SR) generates a highly photoreactive atmosphere [26].

Sampling Sites
PAHs and quinones in the gas and PM 1 phases were sampled in a densely populated area at the Centro (CEN) and Tlaquepaque (TLA) monitoring sites, which form part of the air-quality monitoring network of the Jalisco State Government (SIMAJ) (Figure 1a).The CEN site is located in downtown GMA (Figure 1b

Samples Collection
Samples of ambient air were collected at the CEN and TLA monitoring sites between April and June 2015 during the warm-dry season, which is characterized by high temperatures and SR due to clear sky conditions [27,28].Air was sampled during 24 h (from 00:00 to 23:59 CDT) of each sampling day (Supplementary Material, Table S1) using partisol speciation 2300 samplers (Rupprecht and Patashnick Co., Albany, NY, USA), that were placed 6 m above ground level near to the air inlets of the SIMAJ instrumentation and operated at a flow rate of 16.7 liters per minute (lpm).In total, 15 duplicated samples of gas phase and 15 duplicated samples of PM 1 were collected at each site.The sampling train was integrated by an anodized inlet with a PM 1 impactor (ChemComb 3500, Franklin, MA, USA), equipped with a quartz filter (Ø 47 mm, GE Whatman, Amersham, Buckinghamshire, UK) for PM 1 collection.For the gas-phase sampling, a cartridge of 17 mm polyurethane foam (PUF)/2 g of XAD-4 resin was used (Sigma-Aldrich Co., St. Louis, MI, USA)/PUF sandwich (PXP).Each quartz filter was baked for 12 h at 550 • C before use.The PUF and XAD-4 were pre-cleaned by sonication with a mixture of n-hexane: methylene chloride (1:1 v/v) over two periods of 30 min, the PUF was dried inside an oven at 40 • C and the XAD-4 resin was dried under a stream of nitrogen.The collected samples were transported at 4 • C and stored in a freezer at −20 • C until extraction and analysis.

Chemical Analyses
Four PAHs (fluorene-d 10 , fluoranthene-d 10 , pyrene-d 10 , benzo[a]pyrene-d 10 ) and two quinones (1,4-naphthoquinone-d 6 , anthraquinone-d 8 ) were added to the samples prior to extraction as surrogates [23].Then, the PXP and filters were ultrasonically extracted using methylene chloride for 30 min twice at 45 • C; the organic extracts were concentrated on a rotary evaporator to approximately 1 mL, followed by filtration through 0.45 µm PTFE membrane filters.The samples were reduced to dryness under a stream of nitrogen.Five internal standards (naphthalene-d 8 , acenaphthene-d 10 , phenanthrene-d 10 , chrysene-d 12 , perylene-d 12 ) were added to the concentrate extracted from the samples and the volume was adjusted to 90 µL using methylene chloride.The extracts were analyzed for PAHs and quinones using a gas chromatography-mass spectrometer (Agilent Technologies, Santa Clara, CA, USA, 6890N GC, 5975 MS), equipped with a 5% phenyl methyl siloxane column (30 m, 0.25 mm × 0.25 µm; HP5MS, Agilent J&W), with helium as carrier gas (1.1 mL min −1 , constant flow).The initial oven temperature was 40 • C and was raised to 110 and then maintained for 10 min.One µL of extract was injected in splitless mode and quantified using single ion monitoring mode (SIM).The mass spectrums were obtained using electron impact (EI) mode (70 eV).

Samples' Quality Control
Field blanks, laboratory blanks, and method blanks were used to monitor the sampling and analytical procedures.PAHs and quinones signals were not observed in the blanks.Quantification of each compound was performed using the relative response to the structural homolog as internal standard.All reported atmospheric concentrations for PAHs and quinones were corrected for the recovery of surrogates.Surrogates and average recovery (expressed in % ± standard deviation (SD)) were for  S2-S5).

Meteorology at the GMA
The climate of the GMA is wet/dry tropical climate, with annual averages for temperature and precipitation of 19 • C and 900 mm yr −1 , respectively [30,31]. Figure 2 shows the annual average profiles for temperature (Temp.),relative humidity (RH) and solar irradiation (SI), calculated from SIMAJ continuous records during 1996-2014 in the GMA, with the highest temperatures and SI being observed between May and July.Temperature, RH, wind speed (WS), and solar radiation during the sampling campaign were recorded every 10 min at each sampling site using a Davis Vantage Pro2 weather station.Table 1 lists the summary of meteorological parameters recorded during the sampling campaign, typical of conditions in the GMA during the warm-dry season [32].

Statistical Analysis
Statistical analyzes were carried out using Minitab 16 and OriginPro 8, with the Mann-Whitney test used for comparing two groups of data and the Kruskal-Wallis test for >2 groups of data.To perform statistical analyses, averages were selected for temperature, RH and WS, and the maximum value for SR.The Spearman rank correlation was used to test relationships between quinones and PAHs with meteorological parameters (Temp, RH, WS, and SR) and criteria pollutants (PM 10 , O 3 , NO 2 , CO, and SO 2 ).Statistical significance was defined for a threshold of p < 0.05 for all tests.

Source Apportionment
The concentrations of PAHs are usually standardized into ratios which allow a specific source to be distinguished and to retain reliably the signature of the original fossil combustion source (refractory) [33].This allows identification and assessment of multiple pollution sources in a particular monitoring site [34].The use of these ratios is based on the assumption that the isomers of PAHs have a relative thermodynamic stability or similar physical and chemical properties and, therefore, will transform and degrade at the same rate, preserving the relationship that is present in the emission; and it is expressed with Equation (1) where S is the stable isomer and U the most unstable isomer.Usually, existing studies report this relationship in the range of 0-1 [35,36]: 2.9.Health-Risk Assessment 2.9.1.Benzo[a]pyrene (BaP) Equivalency The benzo[a]pyrene equivalent (BaP eq ) has been used commonly to calculate the risk of cancer due to recurrent exposure to PAHs [37].To calculate the BaP eq for each PAH identified in the GMA, the relative potency factor (RPF) was used to denote the specific power of each PAH relative to that of benzo[a]pyrene (RPF = 1).BaP equivalents (BaP eq , ng m −3 ) were calculated as the product of each PAH concentration C PAHi with its corresponding RPF i (Equation ( 2)).Briefly, the C PAH i cancer potencies relative to BaP for each PAH were based on (i) tumor bioassay data with their associated range and relative confidence ratings; and (ii) an overview of the tumor bioassay database (total number of studies, exposure routes tested, species tested, sexes tested, and number of RPFs derived from benchmark dose (BMD) modeling) [38].Here, we calculated the BaP eq and the sum of the total BaP equivalents (BaP Teq , Equation ( 3)) considering the RPF values in Table 2 as follows [39]:  4), as proposed by the US EPA [40]: where EC (ng m −3 ) is the exposure concentration; CA (ng m −3 ) is the concentration of the individual PAH in the air; ET (h d −1 ) is the exposure time; EF (d yr −1 ) is the exposure frequency; ED (yr) is the exposure duration; and AT (lifetime in yr × 365 d yr −1 × 24 h d −1 ) is the averaging time.Assuming exposures over a lifetime and following the recommendations of the Agency for Toxic Substances and Disease Registry (ATSDR) [41], the values used were ET = 24 h, EF = 365 days, and ED = 70 years.

Cancer Risk Characterized by an Inhalation Unit Risk
The excess of lifetime lung cancer risk (LCR) for a receptor can be described by the Equation ( 5): where IUR is the inhalation unit risk of exposure to BaP; with a selected value of 8.7 cases per 100,000 people (8.7 × 10 −5 per ng m −3 ).Such a value is based on epidemiological data from studies of coke-oven workers with chronic inhalation exposure to 1 ng m −3 BaP over a lifetime of 70 years [42].
Finally, EC was defined as the exposure concentration (ng m −3 ).

Polycyclic Aromatic Hydrocarbons (PAHs)
3.1.1.Ambient Levels Overall, 16 PAHs were detected in the gas + PM 1 samples collected at the CEN and TLA monitoring sites in the GMA.Table 3 lists descriptive statistics for the PAHs observed.PAH concentrations at CEN ranged from 0.00739 ng m −3 for BaP to 103.4 ng m −3 for Nap, while at TLA ranged from 0.00163 ng m −3 for Pyr to 83.2 ng m −3 for Nap.The highest average concentrations of 61.5 ng m −3 and 50.3 ng m −3 were determined for Nap at CEN and TLA, respectively, whereas the lowest were observed for BkF at CEN (2.08 ng m −3 ) and for Ind at TLA (1.39 ng m −3 ).Although higher average concentrations were observed at CEN than at TLA for most of the PAHs, no significant differences (p > 0.05) were observed between concentrations at the two sites for all PAHs, apart from Acy and Phe.The total ambient levels of PAHs measured in the GMA (Σ 16 PAH; the sum of concentrations of all PAHs in the gas and PM 1 phases) ranged from 69.3 to 210.0 ng m −3 , with an average of 140.0 ± 39.1 ng m −3 for the two sites.The highest total concentration by PAH was 52.1 ng m −3 for Nap and the lowest of 1.59 ng m −3 for BaP. Figure 3 shows relative contributions for each PAH to the Σ 16 PAH.Overall, Nap accounted for 37.01% to the Σ 16 PAH, while BaP accounted only for 1.13%.The ΣPAH reported here for the GMA is lower than that of around 900 ng m −3 reported by Possanzini et al. [43] for the urban area of Rome, Italy.This could be due to a shorter sampling period of 6 h made in Rome, because it has been reported that prolonged sampling may result in loss of compounds [44], together with differences between sources and sites characteristics.The ΣPAH for the GMA is within the range of 60.9 to 602 ng m −3 determined by Li et al. [45] in Guangzhou, China, although their reported average of 337 ± 137 ng m −3 is around 2.4-fold that for the GMA.This could arise from differences in the population of both urban areas (around 10 million for Guangzhou) and in the local industrial activities.In contrast, Vasilakos et al. [46] reported lower ΣPAH of 5.6-127.6ng m −3 and 7.44-109 ng m −3 in Athens, Greece, for two suburban monitoring sites, with the differences observed likely due to the different surrounding environments.
The 16 PAHs observed in the gas+PM 1 phases were classified into compounds of LMW (<228) and HMW (>228).The total PAHs distribution was dominated by compounds of LMW, with an average of 128.0 ± 36.6 ng m −3 .Overall, Nap accounted for 40%, Ace 25%, and Phe 11% of the LMW species, whereas for the HMW species the BbF apportionment was 22%, Dib 20%, BghiP 17%, and BaP 13%.The contribution by PAH was similar to those reported in Strasbourg, France, of Nap 38%, Ace 18%, Phe 21% for LMW compounds, and for HMW of BbF 14%, BaP 21%, BghiP 19% [47].The differences in the contributions observed could arise from the influence of coal-combustion emissions in Strasbourg, which are negligible in the GMA.In Guangzhou, China, Li et al. [45] reported contributions for LMW compounds of 62% (Phe), 11% (Flu) and 9% (Ant), while for a BghiP (HMW) was of 16%.Although in Guangzhou and Strasbourg compounds in the gas phase and PM > 1.0 µm were considered, higher PAHs concentrations than in the GMA were reported for the gas phases.

Gas-particulate matter (PM 1 ) Distribution
Table 4 presents the summary of concentrations by PAH observed in the gas and PM 1 phases in the GMA.For the gas phase, the Σ 16 PAH ranged from 68.9 to 208.0 ng m −3 with an average of 137.0 ± 37.7 ng m −3 , while for PM 1 it ranged from 0.365 to 23.9 ng m −3 , with an average of 7.25 ± 8.34 ng m −3 .Individual PAHs ranged from 0.003 (BaP) to 93.3 (Nap) ng m −3 in the gas phase and from 0.000539 (Pyr) to 9.49 (BbF) ng m −3 for PM 1 .To perform statistical analyses, the PAHs were classified according to their number of aromatic rings as follows: 2-rings (Nap); 3-rings (Acy, Ace, Fl, Phe and Ant); 4-rings (Flu, Pyr, BaA and Chr); 5-rings (BbF, BkF, BaP) and 6-ring (Ind, Dib and BghiP).A Spearman analysis revealed that only PAHs with 2-and 3-rings were correlated positively (p < 0.05) with temperature, which suggests that at ambient temperature the lightest PAHs predominate in the gas phase [48].According to such classification (Figure 4), in the GMA the gas phase is dominated by PAHs of 3-rings (47%), followed by compounds of 2-rings (39%) and 4-rings (6%).In Birmingham, UK, Delgado-Saborit et al. [21] reported a similar distribution of 3-(60%) and 4-ring compounds (38%), however, they did not consider 2-ring compounds.The PM 1 showed a distribution of PAHs mainly between 3 (32%), 4 (29%) and 5 rings (22%), which is similar to that reported in Kanpur, India, by Singh and Gupta [49] for PM 1 (3 rings 28%, 4 rings 42% and 5 rings 18%).Similarly, the percentage of ∑PAH HMW reported in Kanpur of 70% is consistent with that of 65% observed in the GMA.
PAHs with HMW were expected to occur mostly as particles.However, in the GMA, compounds of HMW were also observed in the gas phase, while PAHs of LMWs were observed similarly in PM 1 .This behavior was likely caused by the adsorption of LMW PAHs upon particles during sampling, since the aging time (for fine particles) was not long enough for allowing PAHs to condense onto particles in the atmosphere.Moreover, submicron particles smaller in size than the pore size of the quartz filter used for sampling may have penetrated also the filter and entered into the PXP cartridge, contributing to the gas-PAHs mass [50].

Source Apportionment
PAH ratios may change depending on the environmental fates of these compounds.To date, most of the existing studies have focused on the calculation of ratios based on the concentrations of PAHs only in the particulate phase.However, this approach assumes that PAHs emitted in the gaseous or particulate-bound form remain in the emitted phases and does not take into account subsequent changes of phase [34].In this study, in order to reduce the overestimation or underestimation of concentrations by partitioning the compounds in phases, ratios of PAHs were calculated for the total sample (gas + PM 1 ).Because some PAHs react faster than others in atmospheric chemical processes, the ratios of PAHs in the atmosphere often will depart from those seen in source emissions [51].The lifetime of the PAHs was considered in order to select the ratios to determine a reliable source.
Table 5 lists ratios for concentrations of PAHs observed in the GMA.Overall, the observed PAH ratios were consistent with those reported in Oporto, Portugal, by Slezakova et al. [52] for vehicular sources.For instance, the ratio of Phe/Phe + Ant was 0.83, while the averages for Flu/Flu + Pyr and Ind/Ind + BghiP exhibited values of 0.53 and 0.42, respectively, which have been reported as typical for PAHs emitted from diesel combustion [5].This is supported by the BbF/BkF and BaP/BaP + Chr ratios of 1.5 and 0.37, respectively, that are within the range of PAHs originating from diesel combustion [53].Although, the calculated Phe/Phe + Ant and BaP/BaP + Chr ratios could be biased by the short atmospheric lifetimes of Ant and BaP (2.9 and 4.7 hours, respectively) [34], Nap was the most abundant species in the GMA airshed.In urban areas, Nap represents typically a significant fraction of emissions from gasoline-powered cars and light trucks [54,55].This is in agreement with the PAHs of LMW that dominate the GMA airshed and could be emitted mostly from light-duty vehicles.This also agrees with De Abrantes et al. [56] who reported that vehicles represent a significant source of naphthalene in urban areas.

Health-Risk Assessment
The total BaP eq concentration (the sum of PAHs in the gas phase and PM 1 ) of PAHs was 17.7 ng m −3 (Table 6), with the gas phase accounting for 89.4% due to the presence of heavy PAHs with high RPFs.Dib was the major contributor of individual PAHs to the BaP TeqΣ9PAH (77.3%), followed by BaP (8.97%) and BbF (7.62%).Although Dib has an RPF 10 times higher than that for BaP [38], it was in the same order of magnitude as the reference PAH in terms of concentration.These results emphasize the importance of analyzing and evaluating this potent carcinogen as a possible biomarker, since it could be used to monitor the exposure and uptake of high-potency PAHs [57].On the other hand, Nap showed the largest contribution to the total levels of PAHs observed in the GMA.Further research is required to determine the RPF of PAHs not considered here such as naphthalene, phenanthrene and pyrene, which have been associated with chronic respiratory diseases such as asthma, severe bronchitis, lung cancer, etc. and [11,[58][59][60].
The risk value calculated here expressed as the sum of lifetime lung cancer risks was estimated in 1.7 × 10 −3 , which raises significant concerns for the public health of the GMA inhabitants.It suggests that the ambient levels of PAHs represent a potential health hazard, since the recommended level for population exposure to PAHs is lower than 10 −6 [61,62].However, further monitoring must be carried out to confirm such estimates, since risk estimates are commonly highly uncertain [63].The Jalisco Cancer Register reported in 2010 an occurrence of 5.94 cases of pulmonary cancer per 100,000 inhabitants, a ratio of 5.94 × 10 −5 [64].This value approached the tolerance limit of 10 −6 ; however, such data are not sufficient to correlate with lung cancer despite the fact that it has been well documented that high ambient levels of PAHs increase the risk of developing such a disease [57,65].The ambient levels of quinones (Σ 8 Quinones; the sum of the concentrations for all quinones in the gas phase and as PM 1 ) in the GMA ranged from 1.36 to 12.20 ng m −3 , with an average of 7.46 ± 2.88 ng m −3 .The quinones' individual levels ranged from 0.0185 (9,10-PQ) to 10.5 (1,4-PQ) ng m −3 .Table 7 shows the summary of descriptive statistics for the quinones' concentrations in the gas and PM 1 phases.Overall, the concentrations of quinones observed in the GMA were of similar magnitude, apart from the 1,4-PQ.Such concentrations are lower, except for 9,10-PQ, than those reported in Birmingham, UK, by Alam et al. [66] with averages of 5.3 (9,10-PQ), 1.4 (1,4-NQ), 0.8 (AQ), 0.4 (1,2-BAQ) and 0.5 ng m −3 (5,12-NAQ).It is also important to note that comparisons from city to city must account for seasonal variability, meteorological factors, emission sources and other factors that influence the ambient concentrations of quinones.Figure 5 shows the individual relative contributions by quinone to the total ambient concentration.Overall, the 1,4-PQ was the most abundant compound and represented about 65.9% of the total mass, followed by the 1,4-NQ (13.1%), while the other quinones exhibited contributions lower than 10%.

Gas-PM 1 Distribution
Table 8 shows the summary of the individual quinones' concentrations in the gas and PM 1 phases.The Σ 8 Quinones concentrations ranged from 1.14 to 12.20 ng m −3 , with an average of 5.44 ± 3.44 ng m −3 for compounds in the gas phase, and from 0.22 to 3.65 ng m −3 , with an average of 2.32 ± 1.16 ng m −3 for PM 1 .The quinones' individual concentrations ranged from 0.00770 (1,2-BAQ) to 10.50 (1,4-PQ) ng m −3 for compounds in the gas phase, and from 0.00349 (1,2-BAQ) to 3.16 (1,4-NQ) ng m −3 for PM 1 .No statistically significant differences (p > 0.05) were observed between sampling sites for all quinones, which may arise from similar meteorological conditions and emissions contributions due to the relatively short distance between the sampling sites.Simultaneous quantifications of quinones in the gas and particle phases remain scarce, and therefore there is little information to date regarding their phase-distribution [21].The equilibrium between the gas and particle phases is directly related to the vapor pressure of the quinones and varies from season to season [67].Here, the vapor pressure criterion (p • L, Pa at 298 K) was used for quinones, which were separated as follows: 1 × 10 −1 (1,4-NQ); 1 × 10 −3 (9,10-PQ); 1 × 10 −4 (1,4-PQ, 1,4-AQ, 9,10-AQ); 1 × 10 −7 (1,2-BAQ, 5,12-NAQ) and 1 × 10 −8 (1,4-CQ).Overall, the most abundant compounds for both phases were in the group of 1 × 10 −4 (Figure 6), where it is possible to observe quinones in both phases [17,22,68].The 1,4-NQ accounted for 52% of the total compounds in the gas phase.However, this has been commonly observed in particulate matter from vehicle emissions [69,70].Such a percentage difference of about ±25% to other existing studies is likely due to sampling artifacts and different source contributions [17,22].For instance, in Southern California, Eiguren-Fernandez et al. [16] reported quinone levels from 0.42 ± 0.36 to 1.74 ± 1.63 ng m −3 , similar to those reported here for the gas phase, and from 0.01 ± 0.02 to 0.15 ± 0.88 ng m −3 for PM 2.5 , which are in the range of those reported in this study.In contrast, in the GMA the 1,4-PQ accounted for 72% in the gas phase.Lee et al. [14] reported for the 1,4-PQ a partition range of 40-50% in the gaseous phase in a laboratory reaction chamber, although such values cannot be compared to those reported here since such a quinone was a product of a reaction and primary emissions were not considered.
The anthraquinones accounted for 100% of PM 1 , with a higher contribution for the 1,4-AQ than for the 9,10-AQ.This is consistent with the anthraquinones phases' distribution observed in Paris by Ringuet et al. [71], where the 1,4-AQ was the most abundant species in the fine and ultrafine particles, and the 9,10-AQ predominated in the coarse particles.The concentrations observed in the GMA for the 5,12-NAQ in PM 1 are similar to those reported in Boston by Allen et al. [72], where the 1,2-BAQ ranged from 0.00770 to 0.564 ng m −3 , and from 0.00349 to 0.682 ng m −3 in the gas and PM 1 phases, respectively.Overall, the quinones in the gas phase represented 58% of the samples collected, which is lower than the contributions of 69% and 84% reported by Jakober et al. [70] for quinones in the gas phase from gasoline and diesel emissions, respectively.Such a difference could be explained by the quinones' lifetime and the compounds' volatility versus sampling duration [17,73].

Source Attribution
Correlations between gas (g) -particle (p) quinones and PAHs, meteorological parameters and criteria pollutants (CO, SO 2 , NO X and O 3 ) were calculated with the Spearman test.Table 9 lists the Spearman coefficients for PAHs and quinones in both phases.Significant correlations were observed between Nap (p) and 1,4-NQ (p) (r = 0.83), and Phe (p) and 1,4-PQ (p) (r = 0.77), which suggests common emission sources.In Kabul and Mazar-e Sharif, Afghanistan, Wingfors et al. [74] observed that some PAHs and quinones may share emission sources such as coal-and biomass-burning, and vehicle combustion.Lee and Lane [75] suggested that vehicles can emit more NQ in the gas phase than those formed during the reaction of Nap with • OH radicals, which may lead to detection of the greatest NQ contributions in primary emissions.Barradas-Gimate et al. [23] reported that in the GMA the 1,4-PQ in PM 2.5 is emitted mostly as product of diesel combustion, and is in good agreement with the observations in Grenoble, France, for primary emissions reported by Tomaz et al. [22].Furthermore, the correlation between ΣQuinones (p) and SO 2 of r = 0.77 may confirm that quinones have a primary origin, and suggests that submicron particles can be a better indicator for primary emissions than PM 2.5 [76].The 1,4-NQ was correlated positively with the 5,12-NAQ in the particle phase (r = 0.75), which may arise from the role of the NQ as an intermediary in the formation of 5,12-NAQ through the Diels-Alder reaction [77].Additionally, the 5,12-NAQ (p) exhibited a significant relationship with O 3 (r = 0.79), likely due to the ozonation process of the parent PAH [78].Nevertheless, further information of the photochemical processes in the GMA airshed can be obtained from the statistical analysis between PM and PAH.For example, SR exhibited negative correlations with Nap (p) (r = −0.70),Ace (p) (r = −0.69),Fl (p) (r = −0.83),Phe (p) (r = −0.75)and Ant (p) (r = −0.66).This may indicate photolysis and chemical processing even if the products cannot be completely identified, as reported elsewhere in [79].A lower correlation of r = −0.47 was observed between the 1,4-NQ (g) and SR, that could be attributed to high vapor pressure, photochemical loss through photolysis (lifetime of approximately 2 h) [80], and transformation into other species such as benzoic acid and phthalic anhydride [81].The latter has been identified as a product of the oxidation of phthaldialdehyde [82], and in emissions from diesel combustion [83].
The 9,10-AQ has been detected both in primary emissions and as a secondary air pollutant [84].In this study, the 9,10-AQ showed a strong correlation with NO 2 (r = 0.82), and the latter with RH (r = 0.71).Chen and Zhu [85] suggested that the heterogeneous reaction between NO 2 and Ant adsorbed on NaCl particles may derive in the formation of 9,10-AQ and increase with RH, although the reaction mechanism is not well understood to date.No significant correlations (p > 0.05) were observed among other species, likely due to photodecomposition, different lifetimes and high solar incidence [86].Overall, the quinones in the gas phase may experience chemical processing on a scale of few hours compared with a more stable form in PM, which increases with their molecular weight [17,18,87].We calculated ratios of quinone/PAH parent for the 1,2-BAQ/BaA and 1,4-CQ/Chr, which have MW > 228 and have been reported as stable compounds [66].The 1,2-BAQ/BaA and 1,4-CQ/Chr calculated ratios of 0.10 and 0.11, respectively, suggest a primary origin.This is consistent with existing reports that have detected the 1,2-BAQ in emissions from local sources [88] and have considered it as a marker for local traffic emissions [89] because of its resistance to photodecomposition [86].However, more research is needed to determine the feasibility of using the 1,2-BAQ as a marker for traffic emissions.

Conclusions
The ambient levels of PAHs and quinones were measured in the gas and PM 1 phases during the warm-dry season at two urban monitoring sites in the GMA.The cancer risk calculated for gas + PM 1 phases suggests a potential hazard to public health.Diagnostic ratios of PAHs suggest a significant contribution to the total ambient levels of emissions from vehicles and diesel combustion.The correlation between quinones and PAHs indicates that primary emissions are the major contributor to total ambient concentrations, whereas correlations between the 5,12-NAQ and 9,10-AQ with SR, O 3 and NO X show that atmospheric transformations may be a secondary source for such compounds.In addition, the correlations between SR and PAHs and quinones indicate processes of loss or degradation during sampling due to the lifetime of such compounds.Further studies are required to establish accurately the role of photochemical and chemical transformations from PAHs to quinones in the GMA airshed.The simultaneous measurements of quinones and their parents in the gas and PM 1 phases reported here permit a better understanding of the behavior of the occurrence, distribution and sources of such compounds.Additional factors such as the sampling period must be taken into account to reduce uncertainty over the ambient levels and origin of the PAHs and quinones observed.

Supplementary Materials:
The following are available online at http://www.mdpi.com/2073-4433/9/4/137/s1,Table S1: Sampling schedule, Table S2: Single ion monitoring mode (SIM) ions used in the analysis of PAHs by gas chromatography-mass spectrometry (GC-MS), Table S3: Linearity, precision, detection and quantification limits, and sensitivity from PAHs' standard solutions, Table S4: SIM ions used in the analysis of quinones by gas chromatography-mass spectrometry (GC-MS), Table S5: Linearity, precision, detection and quantification limits, and sensitivity from quinones standard solutions.
) (20 • 40 25 N, 103 • 19 59 W, 1582 m a.s.l.), and is strongly influenced by emissions from light-and heavy-duty vehicles during most of the day, with industrial emissions having a negligible influence on the sampled air.The TLA site is located some 4 km SE of the GMA downtown (Figure 1c, 20 • 38 27 N, 103 • 18 45 W, 1622 m a.s.l.), and is surrounded by schools, restaurants, and large residential areas.Measurements of CO, SO 2 , O 3 , NO x , PM 10 and PM 2.5 (criteria air pollutants for Mexican legislation), and meteorological parameters (temperature, relative humidity, wind speed, wind direction, precipitation and solar radiation) have been carried at both sites continuously since January 1996 (available online: http://siga.jalisco.gob.mx).

Figure 1 .
Figure 1.(a) Map of population density and monitoring sites located in Guadalajara metropolitan area (GMA).Blue circles show sampling sites location.(b) The Centro (CEN) and (c) Tlaquepaque (TLA) monitoring sites in the local context.

Figure 2 .
Figure 2. Annual average profile for solar irradiation, temperature and relative humidity in GMA.

Figure 3 .
Figure 3. Relative contributions by PAH to the Σ 16 PAH observed in gas + PM 1 phases in the GMA expressed in percentage.

Figure 4 .
Figure 4. Distribution of PAHs in the gas phase and as PM 1 according to the number of aromatic rings.

Figure 5 .
Figure 5. Relative individual contributions by quinone to the total determined concentration (gas + PM 1 ).

Figure 6 .
Figure 6.Distribution of quinones in the gas phase and PM 1 according to vapor pressure (Pa).

Table 1 .
Meteorological conditions in the GMA during the warm-dry season.

Table 3 .
Total ambient levels of PAHs in the gas and PM 1 phases determined in the GMA.

Table 4 .
Average ambient levels of PAHs observed in the gas and PM 1 phases at the CEN and TLA monitoring sites in the GMA.

Table 6 .
Benzo[a]pyrene toxic equivalent concentration and lung cancer risk (total sample).

Table 7 .
Total ambient levels of quinones in the gas and PM 1 phases in the GMA.

Table 8 .
Average ambient levels of quinones in the gas and PM 1 phases at the CEN and TLA sites in the GMA.

Table 9 .
Spearman coefficients for quinones and PAH correlations.