Characteristics of PAHs in Particulates in the Atmospheric Environment of Hat Yai City , Thailand , and Relationship with RubberWood Burning in Rubber Sheet Production

Natural rubber is a main commodity of Thailand, half of which is formed into ribbed smoked sheets (RSS). Fuel wood is burned during RSS production to heat and dry the rubber sheets. Smoke from the burning wood contains chemical species in particles and gases, notably polycyclic aromatic hydrocarbons (PAHs). In this study, the influence of smoke from fuel-wood burning by rubbersmoking cooperatives on atmospheric air in the neighboring town of Hat Yai, Songkhla Province in southern Thailand, was evaluated for the year 2005. The particle size distributions of ambient air in downtown Hat Yai and Prince of Songkla University (PSU) clearly showed bi-modal behavior, indicating multiple sources of particles. TSP was generally low, averaging 45.6 and 46.2 μg/m for downtown Hat Yai and PSU, respectively. The TSP depended on precipitation, particularly when precipitation was high. The correlation between TSP and rubber sheet production from all rubber cooperatives in Songkhla Province is clear. PAH size distributions of ambient air particles clearly showed single-modal behavior, suggesting that most of the PAHs are associated with small particles. The relationship between wood burning and PAHs in ambient air in Hat Yai and PSU is also clear. Low PAH concentrations during the dry season could result partly from wind direction (E to NE), which blows inland from the Gulf of Thailand. During the rainy season, PAH concentration at both locations was low; however, it was inversely proportional to the precipitation. Hence, rubber sheet production, wind direction and precipitation are important determinants of PAH concentration in Hat Yai area.


INTRODUCTION
Thailand is the world leader in natural rubber (Hevea brasiliensis) production.Total rubber production in 2004 for Thailand was 2.9 million metric tons (Thailand Rubber Research Institute, 2005).Among this, 43% was formed into ribbed smoked sheets (RSS) and 80% of that was exported.China, Japan and United States are the three largest natural rubber consuming countries.
In the RSS production, fuel wood (usually rubber-wood) is burned to supply heat (and smoke) to rubber sheets in the rubber smoke rooms.Smoke from the wood burning contains fine particles and other gas species.
Incomplete combustion results in the formation of polycyclic aromatic hydrocarbons (PAHs) (Furuuchi et al., 2006) and other chemical components.PAHs include hundreds of compounds that are carcinogenic, especially those containing 4-6 aromatic rings.
Two main problems arise from the PAHs generated from the wood burning; workplace pollution affected by the leakage of smoke particles into the working area, and atmospheric pollution to nearby surroundings.
Most PAHs are associated with fine airborne particles, typically 0.5 micron or less (DeMartins et al., 2002).These particles can travel over long ranges and may cause health problems for humans in areas surrounding the emission sources.
PAH emission characteristics from burning of various kinds of wood were investigated by Venkataraman et al. (2002), Hedberg et al. (2002), Hays et al. (2003).Furuuchi et al. (2006) studied the characteristics of smoke particles from rubber-wood combustion and evaluated the influences on workplace environments and the surrounding atmosphere.
The PAH concentration inside the rubber smoking workplace was found to be extremely high, particularly of those with a larger number of aromatic rings in the fine fraction of particles.PAH concentration was about ten times higher than ambient conditions which could lead to serious health problems of the workers.
Since PAHs in particulate phase are airborne, transport of these compounds affects the quality of air in general.PAHs in atmospheric air has then become a topic of interest throughout the world.Many researchers have tried to quantify amounts of PAHs present in atmospheric air in many cities, as well as traced the sources of these compounds (Khalili et al., 1995;Bi et al., 2002;Chetwittayachan et al., 2002;De Martinis et al., 2002;Miguel et al., 2004;Duan et al., 2005).
Effects of PAHs from biofuel burning on atmospheric air were obvious.This work is then aimed at the characterization of the PAHs in the particulate phase in the atmospheric air in the city of Hat Yai, Songkhla Province in southern Thailand (Fig. 1) and how they are related to rubber-wood burning in RSS production.The city is situated in the middle of an area of factories using fuel wood, including several RSS rubber cooperatives.

FUEL WOOD, RUBBER PRODUCTION AND WEATHER INFORMATION
The fuel wood used in RSS production is common rubber-wood abundant especially in southern Thailand.The natural rubber trees are usually cut after 25-30 years when the rubber latex productivity has reached its low point (The Thai Rubber Association, 2004).The wood used for combustion is usually in fresh condition.It is often wet from precipitation, because it is usually stored outdoors.
Combustion of fresh, wet wood results in thick smoke.Total concentration of smoke particles has been found to increase exponentially by increasing the moisture content of the wood used as fuel (Kalasee et al., 2003).Moreover, particle concentration can be dependent on the combustion process and burning time, as well as other factors.tap the rubber latex from the rubber trees because its quality significantly drops when the latex is contaminated by the rain.Small bubbles will be present throughout the contaminated rubber sheets when dried.
However, during the summer season, March to May in Thailand, the production is at its low point.This is because the rubber leaves fall during this season and latex production by the rubber trees is minimal.This explains why the rubber sheet production during March to May is very low even though there is virtually no precipitation.Rubber sheet production in 2005 peaked during January, when fuel wood usage was linearly proportional to rubber sheet production.In general, the rate of fuel wood usage was about 800-1,200 kg of fresh wood per kg of dry rubber sheets (Promtong and Tekasakul, 2007).
The weather in Songkhla Province is a typical east-coast Thailand weather pattern.
During January to April, the wind direction is from east-northeast while the wind direction during May to October is from west-southwest.
During November to December when precipitation is highest, the wind direction shifts back to the northeast direction.
Monthly-average wind speed and direction are plotted in Fig. 2. The wind pattern affects the transport of aerosol particle in atmospheric air.

EXPERIMENTAL PROCEDURE
In this study, the influence of smoke from fuel wood burning by community-level rubber

SIZE DISTRIBUTIONS AND CONCENTRATIONS
Source particle characteristics were where the accumulation-mode peaks take place at the particle size of 0.54 micron, while the coarse-mode peaks take place at 4.0 micron (see definitions of atmospheric aerosol modes in Friedlander, 2000).Concentration of accumulation-mode particles is comparable to that of the coarse-mode particles, though the size is smaller.Sources of accumulation-mode particles could be from either biomass burning (Mukherji et al., 2002), or diesel particulate matters (Vogt et al., 2003).Fuel wood, especially rubber-wood, is the most common     Board, 1995).Relationship between TSP and precipitation is shown in Fig. 6.It is obvious that TSP depends on precipitation, particularly when precipitation is high.TSP is decreased almost linearly when precipitation is increased, though there is a small period when the TSP is increased with increasing precipitation.This takes place when the precipitation is lower than 200 mm/month, the dry season (March to June).This could be an effect of long-range summer forest-fire aerosol particle transport from our neighboring countries, especially Indonesia.However, the effect is neither strong nor obvious.The correlation between TSP and rubber sheet production from all rubber cooperatives in Songkhla Province shows that increase of rubber sheet production

Fig. 1 .
Fig. 1.Map of southern Thailand indicating the city of Hat Yai where samples were collected.
smoking cooperatives on atmospheric air is evaluated.There are about 700 such cooperatives throughout Thailand.Most of them are located in the south.Their main product is RSS.Air samplings were conducted at downtown Hat Yai and Prince of Songkla University (PSU), both in Songkhla Province next to the Malaysia border, in the year 2005.Downtown Hat Yai is a congested, heavytraffic area, while PSU is in the suburbs of Hat Yai where the influence of traffic is low.Both locations are surrounded by several rubber smoking cooperatives.Size fractionated particles at both locations were measured simultaneously using two 8stage cascade (Andersen) impactors (Tokyo Dylec, AN-200).The cut-off diameters of the impaction stages are 0.43, 0.65, 1.1, 2.1, 3.3, 4.7, 7.0, and 11.0 micron.80-mm-diameter quartz fiber filters (Advantec, QR-100) were used for smoke particle collection and for subsequent PAH analysis.They were pretreated in a controlled environment (room temperature, 50% RH) for 72 hours.Sampling flow rate was 28.3 L/min and sampling time was 3 weeks for each sample to ensure sufficient amount of particles collected for PAH analysis.A total number of 5 samples were collected at the downtown Hat Yai station from March to September, and 2 samples were collected at PSU in March and September.Note that only one sample per month was conducted.Total suspended particulates (TSP) were measured by the commercial portable high-volume sampler (Shibata, HV500F).Sampling flow rate was set at a constant 500 L/min and sampling time was 24 hours.The filters used were identicaltype quartz fiber filters (Advantec, QR-100) with 110-mm-diameters.High-volume samplings were conducted for 8 months from March to December at downtown Hat Yai, and 5 months from February to December at PSU.One to four samples per month were collected.Only one sample in each month was analyzed for PAHs.Both the Andersen and highvolume samplers were placed at the open-roof top floors of four-story (PSU) and three-story (downtown) buildings.All collected filter samples were then treated in the identical environment as the pretreatment for 72 hours and then analyzed to obtain the amount of 15 different PAH components]anthracene (BaA), Chrysene (Chr), Benzo[a]pyrene (BaP), Benzo[b]fluoranthene (BbF), Benzo[k]fluoranthene (BkF), Dibenz[a,h]anthracene (DbA), Indeno[1,2,3cd]pyrene (IDP) and Benzo[ghi]perylene (BghiPe).A piece of each filter sample containing 4-5 mg of particulates was cut into small pieces.PAHs were extracted ultrasonically twice using ethanol:benzene (1:3, v/v) for 15 minutes.Then the solution was filtered and the filtrate was evaporated in a rotary evaporator to remove ethanol and benzene.The residue was re-dissolved in acetonitrile and kept in a refrigerator (-20 °C) for analysis.The analysis was conducted using an HPLC (Hitachi, L-2130, 2200, 2300, 2485) with a fluorescence detector along with an Inertsil ODS-P column (5 µm, 4.6 mm diameter, 250 mm length).The mobile phase was a mixture of acetonitrile and water with gradient elution from 55% acetonitrile to 100% acetonitrile at 1 mL/min and 20°C.The recovery efficiency was confirmed to be 0.82 ± 0.12 (n = 3) by spiking known amounts of external PAH standards (Accustandard, 0.2 mg/mL of each 15 PAH in 1:1 of CH 2 Cl 2 :MeOH) onto the pre-treated filters and then processing by the same experimental procedure used for the samples.Travel blank values of PAHs were subtracted from analyzed values.The method used in this study was modified from Toriba et al. (2003), Tang et al. (2005) and Furuuchi et al. (2007).
Fig. 3.It clearly displays bi-modal behavior and cheapest fuel used in industries around the city, including rubber smoking cooperatives and factories, and in the households outside of the town.Note that diesel-engine automobile is one of the dominant modes of transport in Hat Yai, as well as many other cities in Thailand.Most of the diesel-engine vehicles are old and emit black smoke, which is likely to contribute to the particles in the accumulation mode in the downtown area.The sources of coarse-mode particles could be from re-suspension of ground dust or other atmospheric aerosol particles.

Fig. 3 .
Fig. 3. Size distribution of particulate matters in Hat Yai downtown.

Fig. 10 .Fig. 11 .
Fig. 10.Mass fraction of PAH compounds from high-volume sampling in dry season in (a) Hat Yai downtown, and (b) PSU.

Fig. 12 .
Fig. 12. Mass fraction of PAH compounds from high-volume sampling in rainy season in (a) Hat Yai downtown, and (b) PSU.

Fig. 13
Fig. 13 shows the PAH concentration profiles from both downtown Hat Yai and PSU along with the rubber sheet production in Songkhla Province.PAH concentration is defined as the summation of 15 components analyzed.During the dry season (January -

Fig. 15
Fig. 15 shows the relationship of the PAH concentration and precipitation.During the dry season (January-April), the PAH concentrations at both locations are low as mentioned and correlations with precipitation are not clear.During the moderate-rain (May-August) and heavy-rain periods (September-December), however, PAH concentration is

Table 1 .
Parameters for the fitted equations for PAH (4-6 rings) concentration using relation

Table 2 .
Parameters for the fitted equations for PAH (4-6 rings) mass fraction using relation upstream sources than Hat Yai's downtown sampling station about 3 km away.