Aromatic VOCs at Major Road Junctions of a Metropolis in India : Measurements Using TD-GC-FID and PTR-TOF-MS Instruments

Ambient mass concentrations of benzene and toluene were measured at 12 different road junctions of Ahmedabad city in India during the pre-monsoon season of year 2015. A Thermal Desorption-Gas Chromatography-Flame Ionization Detector (TD-GC-FID) technique was used for the analysis of aromatic volatile organic compounds (VOCs) in air samples. In each of both inner and outer ring roads, air samples were collected at 6 sites to investigate the spatial variation of benzene and toluene. The mass concentrations of benzene and toluene show strong site-to-site and day-to-day variations. The average mass concentration of benzene varied in the ranges of 11–35 μg m and 4–12 μg m along the inner and outer roads, respectively. The mass concentration of toluene varied in the ranges of 43–142 μg m and 11–28 μg m along the inner and outer roads, respectively. Overall, the mass concentrations of VOCs along the inner road were 3–5 times higher than those measured along the outer road.The mass concentrations of benzene and toluene show good correlation suggesting their common emission sources (mostly vehicular). However, the enhancement ratios of ∆Toluene/∆Benzene (~4.0 μg μg) along both the roads were higher than the typical ratios (1.5–3.5 μg μg) reported for vehicular emissions. The higher values of ∆Toluene/∆Benzene are due to the emissions of VOCs also from industrial and other non-traffic sources. During the daytime, the lower mass concentrations of VOCs and lower ∆Toluene/∆Benzene (~2 μg μg) indicate the role of photochemical aging. The combined diurnal trend of ∆Toluene/∆Benzene agrees well with that measured at central Ahmedabad using the proton-transfer-reaction time of flight mass spectrometer (PTR-TOF-MS). However, compared to weekdays, the mass concentrations of VOCs show reduction and increase during the Sunday and Saturday, respectively. The mass concentration of VOCs and their ratio were towards the higher side of data reported for different urban sites of the world.


INTRODUCTION
In the global atmosphere, volatile organic compounds (VOCs) are ubiquitous and play important role in the earth's environment and climate change (Velasco et al., 2008).VOCs are emitted from a variety of both natural and anthropogenic sources, however their predominant contribution varies from the region to region (Goldstein and Galbally, 2007).The major anthropogenic sectors include emissions from chemical and petroleum industries, combustion of fossil fuels in automobile engines and power plants (Sahu, 2012).However, particularly relevant to the rural regions of developing countries, the incomplete combustion of biomass burning and biofuel are still important source of VOCs (Sahu andLal, 2006a, Sahu et al., 2015).The major loss of ambient VOCs is due to rapid reaction with hydroxyl (OH) radicals.Therefore, the increasing level of VOCs and relatively faster consumption of OH can significantly reduce the oxidation capacity of the atmosphere.Consequently, the atmospheric lifetime of other gaseous constituents such as methane (CH 4 ) and carbon monoxide (CO) may increase in future (Shindell et al., 2007).Most VOCs that are very reactive can significantly affect the radical chemistry including the heterogeneous interactions.For example, VOCs are important precursors of ground-level ozone (O 3 ) production in the presence of NO x (= NO + NO 2 ) and sunlight (Sahu et al., 2013;Yadav et al., 2014;Yadav et al., 2016).Moreover, photo-oxidation of VOCs leads to the production of simple and multifunctional oxygenated-VOCs (OVOCs) which are transformed to secondary organic aerosols (SOA) through gas-particle partitioning (Seinfeld and Pandis, 2006).The United States Environmental Protection Agency (USEPA) has identified several VOCs as toxic or carcinogen (Sanchez et al., 2008).Therefore, longterm exposure to the high concentrations of such VOCs above the permissible exposure limits can cause acute and chronic health effects.For example, exposure to high concentrations of benzene (C 6 H 6 ) is a well-established cause of cancer in humans (Baan et al., 2009).Similarly, exposure to nonpermissible levels of toluene (C 7 H 8 ) has been known to cause eye, nose, and throat irritation as well as headaches, dizziness, and feelings of intoxication.However indirectly, mostly in the downwind regions of urban and industrial centers, the elevated concentrations of both O 3 and SOA are key pollutants which are produced by the oxidation of VOCs.Therefore, irrespective of whether the effect is direct or indirect, the higher levels of various VOCs particularly in many urban and industrial areas of Asia is as major cause for concern (Louie et al., 2013).
Depending on the region, emissions from both natural and anthropogenic sources make dominant contribution.The contributions of VOCs from anthropogenic (mainly fossil fuel combustion) and biomass burning sources were estimated to be 10.8 Tg yr -1 and 2.2 Tg yr -1 , respectively for India (Streets et al., 2003a, b;Pandey and Sahu, 2014).Several urbanized regions of India are experiencing serious air pollution mainly due to ever increasing number of motor vehicles (Gurjar et al., 2010).Nonetheless, the major source of VOCs in most urban areas is vehicles including tailpipe and evaporative emissions (Parrish et al., 2009).The lack of VOCs measurements in India has been highlighted by the global modeling community (Srivastava and Majumdar, 2010).Thus far, the measurements of VOCs in India have been limited to a few sites (e.g., Srivastava et al., 2005;Sahu and Lal, 2006a, b;Sahu et al., 2010;Lal et al., 2012).A recent review paper (Chauhan et al., 2014) has reported the progress and status of the measurements of harmful VOCs in urban sites of India.
The present study is based on ambient measurements of two aromatic VOCs (benzene and toluene) using TD-GC-FID instrument at different urban sites of Ahmedabad from March to May, 2015.The primary objective of this study is to characterize the site-to-site variation of aromatic VOCs and to identify their local sources.A case study highlighting the "weekend effect" and impact of photochemical aging in the distribution of benzene and toluene has been discussed.However briefly, we have also compared the TD-GC-FID data of aromatic VOCs with our earlier measurements using a high time and mass resolved PTR-TOF-MS instrument at Physical Research Laboratory (PRL), Ahmedabad during a winter month of year 2013 (Sahu and Saxena, 2015;Sahu et al., 2016).

Study Site
Ahmedabad (23.03°N, 72.58°E and 53 m ASL) is the largest and capital city of Gujarat State (Province) in western India.It is the 5 th largest city in India with a population of about 6.5 million (according to the census year 2011).The city is well known for "Sabarmati Ashram" which was the residence of Mahatma Gandhi located on the bank Sabarmati river (Fig. 1).In recent years, vehicular traffic in Ahmedabad city has seen a rapid increase mainly due to fast urban development and increasing population.The different categories of vehicle registered in Ahmedabad are two- wheelers (~72%), four-wheelers (~15%), three-wheelers (5%) and bus/truck (~2%).The number of registered vehicles in Ahmedabad region has been increasing with a rate of ~8.5% per year (Source: RTO, Ahmedabad).Therefore, emission from the transportation sector is a major source contributing to the rising levels of VOCs and other air pollutants in Ahmedabad.Chronologically, emission standards of Bharat Stage II (Euro 2), Bharat Stage III (Euro 3) and Bharat Stage IV (Euro 4) were implemented during the years of 2003, 2005 and 2010, respectively (https://dieselnet.com/standards/in/).Now, it has been made mandatory to obtain Pollution under Control (PUC) certificate for each and every vehicle plying on the roads of Ahmedabad.In the year 2009, the bus rapid transport system (BRTS) corridor system became operational which is designed to attract mass transit demand and improve air quality in the city.In addition to vehicular sources, emissions from thermal power plants, industrial estates and waste landfill sites located in the eastern region of the city contribute to air pollution (Giri et al., 2013).
In this study we have collected air samples at 12 different junctions along the 132 feet ring road (inner) and Sardar Patel (S P) ring road (outer) for the analysis of the mass concentration of benzene and toluene.As shown in Fig. 1, the 6 sites along the 132 feet ring road are RTO circle (west), Shivranjani (west), Juhapura (west), Narolgam (east), Hatkeshwar (east) and Navanaroda (east) while 6 sites along the SP ring road are Khodiyar (SVG circle, west), Bopal (west), Kamod (west), Geratnagar (east), Odhav (east) and Naroda (east).The traffics on the S P ring road and 132 feet road are mainly due to plying of heavy-duty vehicles (HDVs, mostly diesel-fueled) and light-duty vehicles (LDVs, mostly petrol-fueled), respectively (http://rtogujarat.gov.in/process_registration.php).The city is divided as eastern and western Ahmedabad by divided by the course of Sabarmati river consisting mainly of industrial estates and residential area, respectively.Therefore, the sampling sites in the eastern region of Ahmedabad also represent mixed emissions from industrial and vehicular sources.The description of each site and information about the samples collected during each month are given in Table 1.Most of the samples were collected during evening hours (17:00-23:00) which coincide with rush traffic on the roads of Ahmedabad and hence represent relatively fresh emissions.However only for one day, samples were collected during the afternoon hours (12:00-17:00) which coincide with the low vehicular traffic and are also subject to the photochemical aging.Air samples were also collected on Sunday and Saturday to study the "weekend effect" in the ambient concentrations of aromatic VOCs.

TD-GC-FID and PTR-TOF-MS Instruments
The GC-FID system (7890A-G3440A, Agilent, Palo Alto, CA, USA) coupled with a TD system (UNITY2 with Air Server 3 channel module, Markes International, UK) was used for the analysis of VOCs in present study (see Fig. 2).The "Ozone Precursor" cryotrap (U-T17O3P-2S, Markes International) was used for the pre-concentration of VOCs present in air samples.The typical sorbent materials used in the "Ozone Precursor" cryotrap are Tenax TA, Carboxen 1003, Carbosieve SIII, etc.The operating parameters of the TD-GC-FID and details of several other components are given in Table 2.The TD-GC-FID system was used at the University of Crete for the measurements of (C 7 -C 11 ) compounds (Xu et al., 2003).Air samples were collected in glass bottles (800 mL) pressurized up to ~30 psi using an oil free 'Metal Below' compressor (High speed Appliances, India, mini diaphragm type, Model no.IS 4722).At each site, air samples were sucked through a 3-meter long stainless steel (SS) inlet tube so that to avoid the sampling of direct emission and ground dust.As shown in Fig. 2, the nozzle of glass flask was connected to the inlet of TD system via a silicone tube (1/4 inch diameter, length 5 cm).The mass flow control (MFC) of the TD is connected to the external pump.This ensures the transfer of air sample to the TD system at a desired constant flow rate (e.g., 40 mL min -1 ).This flow directed via a cryo-trap continues for 20 min to enrich the sample required for the detection in FID.The hardware part of TD (valve, flow controller, etc.) is controlled by the TD software.For each sample, the interface line (a silicone tube) was checked for leak before the start of transfer process.The TD-GC-FID instrument was calibrated using a dynamic dilution of the standard gas (~1 ppmv of each benzene and toluene, L5388, Ionicon Analytik GmbH Innsbruck, Austria).The dynamic dilution of the standard gas mixture was prepared using an advancedmodel "gas calibration unit" (GCU-advanced v2.0, Ionicon Analytik).As shown in Fig. 3, the scatter plot between the normalized peak area (sensitivity) and set value (volume mixing ratio in ppbv) showed excellent linear response (r 2 > 0.98) for both the VOCs.The overall uncertainties of in the analysis of benzene (~7%) and toluene (~7.9%) include the inaccuracies in the flow rate of GCU and calibration mixture (± 5%).However, to be consistent with most of the other research papers published in this field, we have converted volume mixing ratio (ppbv) to mass concentration (µg m -3 ).
The measurements of various VOCs including benzene and toluene have been performed using the PTR-TOF-MS technique since Nov 2013 at PRL, Ahmedabad.This is an online technique for the in-situ measurements of VOCs.The ambient air is introduced into the PTR-TOF-MS system via a heated (60°C) PEEK tube at a flow rate of 60 mL min -1 .The PTR-TOF-MS is fitted with a switchable reagent ion (SRI) source.The PTR-TOF-MS system has been mostly operated in the hydronium ion (H 3 O + ) mode.However, for the comparison purpose only, we have used the average data of aromatic VOCs data measured during December 2013 (winter season) in the present study.The detailed results obtained from the PTR-TOF-MS measurements are reported in a recent paper (Sahu and Saxena, 2015).

Variability of VOCs along 132 Feet Ring Road (Inner)
The mass concentrations of benzene and toluene at 6 different sites (road junction/crossroad) along the 132 feet ring road for five different days of sampling are given in Figs.4(a) and 4(b).The mass concentrations of both aromatic    Along the 132 feet road, the average mass concentrations of benzene in the eastern and western Ahmedabad were 23 ± 17 µg m -3 and 15 ± 11 µg m -3 , respectively.While the ambient levels of toluene show large differences as the mass concentrations in the eastern and western regions of the city were 98 ± 79 µg m -3 and 57 ± 47 µg m -3 , respectively.We have calculated the emission ratio of toluene/benzene (T/B, µg µg -1 ) for different sites along the 132 feet ring road (see Fig. 4(c)).Unlike strong site-to-site variation in the mass concentrations of benzene and toluene, their average emission ratio (T/B) shows less variability (2.5-4.3 µg µg -1 ).However, exceptionally low (0.07-1.62 µg µg -1 ) and high (2.5-5.5 µg µg -1 ) values of T/B ratio were observed during 25 March and 21 May, respectively.In spite of strong dayto-day variation, the average T/B ratio of 3.3 ± 1.4 µg µg -1 for the eastern sites was slightly higher compared to the average value (2.8 ± 1.1 µg µg -1 ) estimated for the western sites.Overall, the mass concentrations of VOCs and their ratio (T/B) were higher in east Ahmedabad compared to those in the western region of the city.Along this road, the concentrations of benzene and toluene in the eastern region were about 8.0 µg m -3 and 46.0 µg m -3 higher than those in the western region, respectively.The samples collected at Juhapura and Navanaroda on 25 April and 21 May show exceptionally high values of T/B ratio (> 4.9).It is important to mention that Juhapura and Navanaroda are located close to Table 3. Mass concentrations (average ± standard deviation) of benzene, toluene and toluene/benzene emission ratio along the inner ring road (132 feet) and outer ring road (S P) of Ahmedabad city.

Site
Concentration (µg m -3 ) Emission ratio (µg µg High T/B ratios at Narolgam on May 21 also provide an evidence of emissions from the waste landfill.Therefore, in addition to vehicular sources the emissions from waste landfill and industrial activities could have contributed to the higher T/B ratio at these two sites.

Variability of VOCs along S P Ring Road (Outer)
The mass concentrations of aromatic VOCs measured at 6 different sites along the S P ring road (outer) are given in Figs.4(d) and 4(e).Along the S P ring road, the mass concentration of benzene and toluene varied in the ranges of 0.5-32 µg m -3 and 1-88 µg m -3 , respectively.The average mass concentrations of both the compounds at different sites along the S P ring road are given in Table 3.As estimated from the composite data for different days of sampling, the mass concentration of benzene was highest at Geratnagar and lowest at Naroda.While the mass concentration of toluene shows highest at Odhav and lowest at Khodiyar.Both the sites namely Geratnagar and Odhav are located in the eastern region of Ahmedabd city.On an average, the mass concentrations of benzene in the eastern and western sectors of this road were 8.4 ± 4.0 µg m -3 and 11 ± 7 µg m -3 , respectively.While mass concentrations in the in the eastern and western regions of the city were 20 ± 19 µg m -3 and 19 ± 16 µg m -3 , respectively.Therefore, along the outer ring road, the concentrations of VOCs do not show significant differences between the eastern and western regions.The emission ratios of T/B (µg µg -1 ) for different sites and days along the S P ring road are shown in Fig. 4(f).The emission ratio of T/B show small site-to-site variability (1.5-3.2 µg µg -1 ) with highest at Odhav and lowest at Geratnagar.The Geratnagar site is located in the southeast (SE) direction from the main city centers.This site is also located in the upwind as prevailing winds were from the SE-SW sector during the study period.Therefore, relatively low abundance of toluene could be due to transport of aged air.
Overall, the values along the S P ring road are lower than those observed along 132 feet road.The mass concentrations of VOCs and their ratio (T/B) along the SP and 132 roads show distinct features.The samples collected at Naroda and Odhav on 25 April show exceptionally high values of T/B ratio (> 5.0).It is important to mention that these two sites are located very close to the major industrial estates in Ahmedabad.Therefore, in addition to vehicular sources the emissions from industrial activities also contribute to the ambient VOCs leading to higher T/B ratio at these two sites in eastern Ahmedabad.Overall, higher T/B ratios along the eastern sector of both the roads indicate the role of additional emission from non-traffic sources (e.g., industry, landfill, etc.) located in the eastern region of Ahmedabad.
The dependence of the mass concentrations of VOCs and ratios of T/B measured at all sites with wind speed is shown in Fig. 5.It can be noticed that the enhancements in the concentration of both the VOCs and their ratio were observed during strong winds from the SE-SW sector.As shown in the map (Fig. 1), the major industrial estates and a landfill area are located in this sector of the city.

Correlation between Benzene and Toluene
We have investigated the correlation between the mass concentrations of benzene and toluene measured along both the roads.VOC concentration ratios provide useful indicators of emission source, photochemical oxidation and transport processes.On the other hand, the influences of these processes are inextricably linked to the evolution of these ratios (Parrish et al., 2007).The ratio of different VOCs depends on several factors such as dilution and background concentration.In this study, the enhancement ratio of ∆Toluene/∆Benzene (∆T/∆B) has been determined from the slope of linear regression fit.
Fig. 6(a) shows strong correlation (r 2 = 0.89, P < 0.003) along 132 feet ring road while data along the S P ring road shows moderate correlation (r 2 = 0.69, P < 0.001).However, estimated based on the entire data, the slope (∆T/∆B) of 4.4 along the 132 feet road is higher compared to the value of 4.0 along the SP ring road.These values are slightly higher than the range of ~1.5-3.5 reported from vehicular emissions (Barletta et al., 2005;Miller et al., 2011).In addition to vehicular sources, the comparison clearly shows the influence of other sources such as industrial and landfill emissions leading to higher ∆T/∆B slopes.Further, the distinct ∆T/∆B slopes along 132 feet and S P roads with higher and lower values could be due the predominance of LDVs and HDVs, respectively.As reported in several studies the emission ratio of T/B is higher for LDVs and lower for HDVs (Lai et al., 2013).Another reason could the closer proximity of several sites along 132 feet ring road compared to the sites along the S P ring road.In the lower mass concentration regime (benzene < 10 µg m -3 and toluene < 40 µg m -3 ), the slopes of ∆T/∆B along both the roads were significantly lower (1.7 and 2.8) compared to those determined from entire data (see Fig. 6(b)).Here, we have assumed that low values are due to combined effect of reduced emission, PBL dilution and faster aging in the afternoon hours.On the other hand, the slope of ΔT/ΔB will be least influenced by the reduced emission and PBL dilution.Therefore, the cause of lower mass concentrations of VOCs can be explained by the reduction in emission and PBL dilution, but not to the lower slopes of ΔT/ΔB.In a strict sense, the lower values of ΔT/ΔB should be attributed to the sampling of photochemically aged air mass during the afternoon hours.
We have separately analyzed the correlation between two aromatic VOCs for eastern and western regions of Ahmedabad.As shown in Fig. 6(c), the mass concentrations of benzene and toluene show strong correlation (r 2 ≥ 0.90) for the both the regions of Ahmedabad.However, based on entire data in each sector, the slopes of ∆T/∆B were estimate to be 4.0 and 4.9 for the eastern and western sectors, respectively.In the lower regimes of both the VOCs, the slope of ∆T/∆B shows different comparison as slightly lower value was estimated for the western sector than that of eastern sector of the city.We have also performed the measurements of VOCs using PTR-TOF-MS at the campus of PRL in the central Ahmedabad during the month December 2013 (Sahu and Saxena, 2015).This site represents well-mixed air from different emission sources in the city.Moreover, the high time resolved measurements of VOCs using the PTR-TOF-MS provide a good opportunity to study the role of aging in the relationship between toluene and benzene.The correlation between the mass concentrations of benzene and toluene at PRL site is shown in Fig. 7.The data during the evening hours shows strong correlation (r 2 = 0.94) and ∆T/∆B value of 4.38 which falls between the slopes of eastern (3.5) and western (4.9) regions.For the daytime, the PTR-TOF-MS data shows a slope of 1.55 which is almost similar to the present result for the daytime measurements (∆T/∆B ~2) at different traffic intersection in Ahmedabad.The two major factors which determine the correlation between the mass concentrations of benzene and toluene can be the emission source and photochemical aging.The moderate to strong correlation along both the roads and regions of Ahmedabad indicate impact of common or co-located sources of these VOCs in  Ahmedabad.On the other hand, however irrespective of sampling locations (inner, outer, east, west, etc.), the lower mass concentrations of VOCs were observed in the samples collected during the afternoon hours.Therefore, the smaller ∆T/∆B values (~2) indicate the role of photochemical aging.
In a typical urban environment, both the VOCs are emitted mainly from the vehicular exhaust and removed from the ambient air by the reaction with OH radicals (Barletta et al., 2002).The reaction rate constant of toluene (5.63 × 10 -12 cm 3 molecule -1 s -1 ) is about 5 times faster than that of benzene (1.22 × 10 -12 cm 3 molecule -1 s -1 ) (Atkinson, 2000;Atkinson and Arey, 2003).Therefore, the dependence of emission ratio (T/B) or enhancement ratio (∆T/∆B) can be used to investigate the photochemical aging of air mass (Weber et al., 2007).Consequently, lower values of ∆T/∆B during the daytime indicate role of photochemical processing.The time dependence of the mass concentration of VOCs and T/B ratio measured at 12 different sites in Ahmedabad is shown in Fig. 8.The mass concentrations of both the VOCs were very low during 12:00 to 18:00 IST (Indian Standard Time).The main reasons could be the dilution due to deeper planetary boundary layer (PBL) depth, photochemical loss and lower traffic emission during the daytime.The time dependence of T/B ratios show lower (average = 1.9) and higher (average = 2.9) values during the daytime and evening hours, respectively.The daytime reduction in T/B ratio can be attributed mainly to the faster photochemical removal of toluene, while enhanced T/B values during evening hours is mainly due to sampling of freshly emitted pollutants.Although measured in winter season, the high time resolved PTR-TOF-MS data of ∆T/∆B at central Ahmedabad agrees well with the trend seen from combined data at all sites.

A Case Study of the Weekend Effect
The "weekend effect" is characterized by decrease in the ambient levels of primary pollutants due to reduction in the emissions from anthropogenic sources during the weekend (Cerveny and Balling, 1998).The specific cause of decreased emission is mainly related to the reduction in vehicular traffic during the weekend (Sahu et al., 2011).To study the "weekend effect" at different sites we have considered weekdays (Wednesday-Friday) and two nearest weekends of 22/03/2015 (Sunday) and 25/04/2015 (Saturday).The local weather conditions remained almost same for these days of measurements.The sky condition was mostly clear/sunny during all the days of sampling at Ahmedabad.In the premonsoon season (March-May), the ambient temperature is recoded to be higher than those during other season.The diurnal ranges of temperature and RH during different days of sampling at Ahmedabad are reported in Table 1.Therefore, we can assume that the relative change in the mass concentration of each compound reflects the different traffic patterns during weekdays, Saturday and Sunday.For both benzene and toluene at each site, we have estimated the mass concentration ratios of Sunday/Weekday and Saturday/Weekday to study the effect of the weekend traffic at Ahmebabad (Fig. 9).Along the 132 feet ring road, however except Shivranjani (commercial place), the ratios of both Benzene Sun /Benzene Wkd and Toluene Sun /Toluene Wkd show lower values (< 1.0) indicating the impact of weekend reduction in vehicular traffic.The highest effect of weekend reduction in ambient VOCs was observed at Hatkeshwar site (Fig. 9(a)).Similarly, except for two sites near the industrial estates (Naroda, Kamod), the ratios of both Benzene Sun / Benzene Wkd and Toluene Sun /Toluene Wkd show lower values (< 1.0) indicating impact of weekend reduction in vehicular traffic.At some sites near the major traffic intersection the higher weekend/weekday ratios (> 1.0) were observed along both the roads.These sites are close to Shivranjani, Naroda and Kamod traffic intersection which connect the central Ahmedabad to the state highway (SH) and national highway (NH).The measurements of higher Benzene Sun /Benzene Wkd and Toluene Sun /Toluene Wkd ratios at these sites are consistent with increased traffic during the weekend at these junctions.

Comparison with International Data
The mass concentrations of aromatic VOCs and their ratio at Ahmedabad have been compared with the measurements reported at different urban sites of the world (Table 4).The average mass concentrations of benzene (6.2 µg m -3 ) and toluene (19 µg m -3 ) along the outer ring road show good agreement with the measurements at Nanjing in China (Wang and Zhao, 2008), Hong Kong (Ho et al., 2002), Quito in Ecuador (Gee and Sollars, 1998), Istanbul in Turkey (Demir et al., 2011) and Algiers in Algeria (Kerbachi et al., 2006).The average mass concentrations of benzene (20 µg m -3 ) and toluene (78 µg m -3 ) along the inner ring road show good agreement with measurements at Delhi (Srivastava et al., 2005;Hoque et al., 2008) and urban road sides in China (Wang et al., 2002).However, the average emission ratio of T/B (2.4 ± 0.5 µg µg -1 ) along the outer road is about half of that at Shiohama in Japan but comparable to the values at Edmonton in Canada, Athens in Greece and Ulsan in South Korea.The T/B ratio (3.0 ± 1.4 µg µg -1 ) along the inner road is about 2 times higher than those at Algiers in Algeria but just about half the values measured at Seoul in South Korea and Tsurumi in Japan.Overall, the mass concentrations of VOCs and their ratio fall within their respective values reported for different urban sites of developed and developing countries.

SUMMARY AND CONCLUSIONS
Ambient air samples were collected at 12 different urban sites of Ahmedabad in India for the analysis of aromatic VOCs (benzene and toluene) during the pre-monsoon season of the year 2015.In each of 132 feet ring (inner) road and S P ring (outer) roads, air samples were collected at 6 different sites (near road junction).The mass concentrations of both benzene and toluene show significant site-to-site and day-to-day variations.However, at any given site, the mass concentration of toluene was higher than that of benzene.As expected, the mass concentrations of benzene and toluene along the inner road were significantly higher than their values along the outer road.Along the inner ring road, the average mass concentrations of benzene and toluene were 20 ± 11 µg m -3 and 78 ± 55 µg m -3 , respectively.However, significantly lowers values of benzene (6.2 ± 3.0 µg m -3 ) and toluene (19 ± 14 µg m -3 ) were measured along the outer road.The sites in the eastern Ahmedabad along both the roads are more polluted compared to the sites in the western part of the city.The most polluted sites are located very close to industrial estates in the eastern Ahmedabad.In addition to vehicular emissions, measurements in eastern Ahmedabad show the contribution of VOCs emitted from the industrial activities and waste landfill.The higher values of toluene/ benzene ratio (T/B) along the eastern region confirm the role of non-traffic sources (e.g., industry, landfill, etc.).The mass concentrations of VOCs and T/B ratio show lower and higher values during the day and evening hours, respectively.The main reason for the lower concentration could be the efficient ventilation of pollutants due to deeper PBL depth and reduced traffic emission during daytime.While the lower values of T/B during the daytime can be attributed mainly to the faster photochemical removal of toluene.The mass concentrations of VOCs show significant "weekend effect" in Ahmedabad.For most of the sites, the measurements on Sunday show substantial decrease compared to a weekday.However, at some sites near the intersections connecting central Ahmedabad to the national and state highways, both benzene and toluene were higher compared to the weekday values.Overall, this study highlights the large spatio-temporal variability of aromatic VOCs within an urban domain and also the limitations of relying on a single site unless air masses from different sources are well mixed.

Fig. 1 .
Fig. 1.The left panel shows the location of Ahmedabad city in India and right panel shows sampling sites along the SP ring road (red stars), 132 feet ring road (blue stars) and location of industries, thermal plant and Pirana landfill.Maps are taken from https://www.google.co.in/maps.

Fig. 3 .
Fig. 3. Calibration plots showing linear response in TD-GC-FID system for both benzene and toluene.

Fig. 4 .
Fig. 4. Concentrations of benzene, toluene and toluene/benzene ratio at different sites along inner ring road (132 feet) and outer ring road (S P) of Ahmedabad city.

Fig. 5 .
Fig. 5. Relationship between the mass concentrations of VOCs and ratios of T/B with wind direction for all sites during entire sampling period.

Fig. 6 .
Fig. 6.(a) Correlation between mass concentrations of benzene and toluene along 132 feet and S P ring roads, (b) same as (a) but for the lower concentration regimes.(c) Correlation between benzene and toluene in the eastern and western sectors (d) same as (c) but for lower concentration regimes.

Fig. 7 .
Fig. 7. (a) Correlation between mass concentrations of benzene and toluene measured at central Ahmedabad (PRL) using PTR-TOF-MS during winter of 2013.(b) Same as (a) but for the lower concentration regimes.

Fig. 9 .
Fig. 9.The Sunday/weekday and Saturday/weekday ratios of mass concentrations of benzene and toluene at 132 feet road and S P ring road of Ahmedabad city.

Table 1 .
The name of sites, sampling time and major emission sources along the inner (132 feet) and outer (S P) roads in Ahmedabad city.

Table 2 .
The important operating parameters of the Thermal Desorption-Gas Chromatography-Flame Ionization Detector (TD-GC-FID) for the analysis of benzene and toluene in air samples.

Table 4 .
Comparison of mass concentrations (average ± standard deviation) of aromatic VOCs and toluene/benzene (T/B) ratio at different cities of the world.