Distribution of Volatile Organic Compounds over a Semiconductor

and Department of Radiological Technology, Yuan-Pei Institute of Science and Technology, ... The results entail that hi-tech semiconductor industr...
0 downloads 0 Views 1MB Size
Environ. Sci. Technol. 2005, 39, 973-983

Distribution of Volatile Organic Compounds over a Semiconductor Industrial Park in Taiwan KONG-HWA CHIU,† BEN-ZEN WU,‡ CHIH-CHUNG CHANG,§ U S H A S R E E , ‡ A N D J I U N N - G U A N G L O * ,‡,| Department of Natural Sciences, National Science Council, Taipei, Taiwan, Republic of China, Environmental Chemistry Laboratory, Department of Atomic Science, National Tsing Hua University, 101, Section 2, Kuang Fu Road, Hsinchu 300, Taiwan, Republic of China, Research Center for Environmental Changes, Academia Sinica, Taipei, 11529, Taiwan, Republic of China, and Department of Radiological Technology, Yuan-Pei Institute of Science and Technology, Hsinchu 300, Taiwan, Republic of China

This study examined volatile organic compounds (VOC) concentration in ambient air collected during the years 2000-2003 at several different locations of Hsinchu Sciencebased Industrial Park (HSIP) in Taiwan. A canister automated GC-MS system analyzed the volatile organics in ambient air grasp samples according to TO-15 method. Oxygenated volatiles were the most abundant VOC detected in HSIP followed by aromatics that are commonly used as solvents in the semiconductor industries. The major components measured in the ambient air are 2-propanol (29-135 ppbv), acetone (12-164 ppbv), benzene (0.7-1.7 ppbv), and toluene (13-20 ppbv). At some of the sampling locations, odorous compounds such as carbon disulfide and dimethyl sulfide levels exceed threshold values. The estimated toluene/benzene ratio is very high at most of the sites. However, the total amount of VOC is reduced over the years from 2000 to 2003 due to strict implementation on use and discharge of solvents in industries. There exists no definite seasonal pattern for sporadic occurrence of high levels of some of the volatile organics. Stagnant weather conditions with low wind speeds aid accumulation of toxic species at ground level. The results entail that hitech semiconductor industries are still a potential source for harmful organic substances to surrounding microenvironment.

Introduction In view of growing concern over global warming, it has become essential to acquire the emission profile of volatile organic compounds (VOC) in Taiwan to understand the industrial air pollution problems. Although the emission profiles of the chemical and petroleum industries in Taiwan have been developed, little is known about ambient air in * Corresponding author telephone: +886-3-5731175; fax: +8863-5727305; e-mail: [email protected]. † National Science Council. ‡ National Tsing Hua University. § Academia Sinica. | Yuan-Pei Institute of Science and Technology. 10.1021/es049110m CCC: $30.25 Published on Web 01/14/2005

 2005 American Chemical Society

the semiconductor and electronic industries (1, 2). Most of the electronic and communication devices are composed of integrated circuit chips. These chips are produced through a series of processes such as crystal growth and preparation, wafer fabrication, integrated circuit packaging, and testing. The wafer fabrication processes are the most complicated, and many toxic chemicals maybe used. Also, fluorinated compounds (FCs) and several acids are essential to the semiconductor manufacturing process, which contribute to air emissions from semiconductor industry. To respond to international commitment (RIO Earth Summit ’92 and Kyoto Protocol-97) over reduction of greenhouse gas emissions, several reports concerning reduction and reuse of FCs have been reviewed (3-5). Apart from fluorinated compounds and acids, diverse ranges of volatile organics are employed in the process of manufacturing in semiconductor industries. Volatile organics such as aromatics, halogenated hydrocarbons, and polar volatile organics such as aldehydes and ketones are used in large quantities and are emitted into atmosphere by volatilization or are discharged into wastewater. Thus discharged volatiles are reported to cause potential mutagenic effects and are toxic to the surrounding microenvironment, besides ending up as harmful byproducts as a result of light-induced photochemical reactions (6-8). Hsinchu Science-based Industrial Park (HSIP), located in Hsinchu of northwestern Taiwan, accommodated 136 IC firms, 50 computer and peripherals companies, 60 telecommunication companies, 56 optoelectronics companies, 15 precision machinery companies, and 18 biotechnology companies at the end of 2002. Following the successful experience of the HSIP and trends in global hi-tech industry, the government is keenly involved in transforming Taiwan into a “Green Silicon Island”. Although this industry is a success from an economic point of view, the burden it creates on the environment resulting from fugitive emissions also has become a serious threat to the environment and public heath. The residents living near the factories protest about the stinking odor, suspected to be released from the factories. In a sampling campaign conducted in 2000 at HSIP, continuous FTIR monitoring identified compounds such as ammonia, methanol, carbon tetrafluoride, and sulfur hexafluoride in ppbv and sub-ppbv ranges (9). Besides, GC-MS analysis of grab samples detected over 40 compounds consistent in three seasons that included acetone, dichloromethane, propylene glycol monomethyl ether acetate, and trimethylsilanol. A recent work on large concentration variability of chlorofluorocarbons (CFCs) and periodic emission episode observations were correlated to the usage in this industry, wind direction, and topography (10). Besides the release of greenhouse gases such as CFCs, this industry has become a potential source for release of other toxic volatiles. For most of the studies on semiconductor industrial air composition for toxic substances, CFCs and inorganic gases were evaluated and other volatile organics are seldom reported. Measurement of inorganic pollutants was conducted using a sensitive porous metal sampler during January 2001 to February 2003 to understand whether the inorganic pollutants emission standard is met and investigated the emission rates, emission profiles, emission factors, and control efficiencies of various gaseous/aerosol species (11). The emission factor of VOC from semiconductor manufacturing in Taiwan was estimated based on data collected over a time period October 2000 to September 2001 to initiate the collection of and to determine the air pollution fee from the factories (12). Indoor air analysis of a semiVOL. 39, NO. 4, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

973

FIGURE 1. Geographical location of (a) Hsinchu city in Taiwan and sampling sites in HSIP during the years (b) 2000, (c) 2002, and (d) 2003. For 2000: A1, front door of a high school; A2, information company; A3, a company; A4, semiconductor factory; A5, semiconductor factory; A6, semiconductor factory; A7, electrical company; A8, semiconductor corporation; A9, semiconductor factory under construction; B1, south end of road; B2, CD-RW, DVD ROM factory; B3, semiconductor factory; B4, road junction; B5, disk paint factory; B6, semiconductor factory; B7, semiconductor factory; B8, semiconductor factory; B9, chipset factory; C1, road junction; C2, semiconductor factory; C3, road junction; C4, semiconductor factory; C5, notebook factory; C6, road junction; C7, road near radiation center; D1, road junction (near semiconductor factory); D2, road junction (near semiconductor factory); D3, communication server factory; D4, road junction; D5, road junction (near IC factory); D6, show center of the science park; D7, road junction; D8, front door of sewage treatment plant; D9, front door of National Space Program Office; F1, housing community; F2, housing community; F3, housing area; F4, basketball court; F5, housing area. For 2002: A1, semiconductor factory; A2, semiconductor factory; A3, semiconductor factory; A4, network communication factory; B1, semiconductor factory; B2, biotechnology factory; C1, semiconductor factory; D1, semiconductor factory; D2, computer monitor factory; D3, semiconductor factory; D4, semiconductor factory; D5, semiconductor factory; D6, TFT-LCD factory; D7, biotechnology factory; D8, semiconductor factory; D9, TFT-LCD factory; D10, semiconductor factory; E1, Institute of Micro-electro mechanical system factory. For 2003: A1, integrated circuit company; A2, photoelectric company; A3, computer industry; A4, integrated circuit company; B1, integrated circuit company; B2, integrated circuit company; B3, computer; B4, integrated circuit company; B5, biotechnology industry; B6, photoelectric industry; B7, integrated circuit company; B8, photoelectric company; C1, computer industry; C2, biotechnology industry; C3, sewage plant; C4, communication industry; C5, integrated circuit company; D1, integrated circuit company; D2, photoelectric company; D3, communication industry; D4, integrated circuit company; D5, photoelectric company; E1, housing area; E2, housing area; E3, housing area. conductor industry measured VOC such as acetone, IPA, 2-heptanone, and toluene (13). Temporal variations of the acetone and IPA concentrations show that these were the two highest among those of VOC detected at any time and were attributed to the improper use of organic solvents during operation. The wastewater discharge from high-tech industries measured acetone as high as 4 ppm that is released into surrounding atmosphere directly (14-16). Most of the studies used sampling methods. Nonsampling methods are seldom reported for analysis of industrial organic solvents in ambient air. In a previous study using a non974

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 4, 2005

sampling technique (open path FTIR spectrometer), organic solvents such as IPA were identified at workplaces of semiconductor industries (17). The results showed concentrations in the range of 20-500 ppbv. The ease of the technique is well-documented but is defined by its own limitations. Gas chromatography techniques in combination with extraction and adsorption techniques are well-established now for ambient air volatile organics (18). While these methods have advantages, they also must contend with issues of recovery and analysis of polar compounds. An automated sampler has been constructed for the unattended collection

TABLE 1. VOC Concentrations (ppbv) Measured Using GC-FID during May 12-16, 2000a parameter

acetone

IPA

benzene

toluene

m/p-xylene

o-xylene

average SD max time recorded at max

19.1 26.7 155.6 5/13/2000 17:00 1.5 2000/5/14 00:00

3.3 1.6 10.8 5/12/2000 22:00 nd 2000/5/13 04:00

8.2 10.5 49.5 5/13/2000 18:00 0.20 2000/5/14 22:00

11.4 14.6 140.2 5/15/2000 10:00 nd nds

1.5 2.5 12.3 5/16/2000 15:00 nd nds

1.3 1.5 7.7 5/12/2000 7:00 nd nds

min time recorded at min a

nd ) below detection limit; nds ) below detection limit several times.

TABLE 2. Sum of VOC Species (ppbv) Detected in Canister Air Samples Collected in 2000a sampling site oxygenated aromatics A1 A2 A3 A4 A5 A6 A7 A8 A9 B1 B2 B3 B4 B5 B6 B7 B8 B9 C1 C2

6.6 7.8 10.0 5.8 6.8 10.6 5.2 3.9 4.0 4.1 3.8 10.1 29.0 8.8 4.4 6.9 6.1 32.8 19.6 5.3

9.2 14.5 5.6 4.9 3.6 4.1 4.3 4.6 4.2 3.5 3.9 7.6 6.1 5.0 3.2 5.0 4.8 5.5 8.8 4.1

sampling site oxygenated aromatics C3 C4 C5 C6 C7 D1 D2 D3 D4 D5 D6 D7 D8 D9 F1 F2 F3 F4 F5

14.2 10.4 13.6 7.1 12.3 11.5 5.8 11.2 2.8 4.4 7.6 7.8 17.4 15.7 11.4 13.4 5.0 16.2 10.6

10.0 4.5 5.7 6.6 6.2 4.1 3 6.7 5.4 5.2 5 5.2 5.6 3.3 3.5 10.1 5.0 8.4 6

a Aromatics: benzene, ethylbenzene, m,p-xylene, o-xylene, styrene, toluene, 1,3,5-trimethylbenzene. Oxygenated: acetone, butanone, ethyl acetate, butyl acetate, 2-propanol.

of whole-air samples in Summa passivated stainless steel canisters to analyze both aliphatic and polar organics (19). The analytical precision values for the light oxygenated hydrocarbons such as ethanol and methanol were not satisfactory. Further research efforts were essential to analyze VOC with in-situ gas chromatographic techniques. Simultaneous trapping and analysis of both polar and nonpolar volatile organics in semiconductor operations overcome this limitation where the sampling apparatus has a multi-sorbent trap connected with an automatic sampling unit to collect and absorb volatile organic compounds at room temperature (20). In this work, ambient air samples were collected during the years 2000-2003 from different sampling points, covering the entire HSIP area, to assess VOC concentrations.

Experimental Section Sampling Location. Hsinchu Science-based Industrial Park is located in the midst of the Hsinchu County of northwestern Taiwan. Hsinchu is surrounded by two arrays of mountains that open toward the sea in the west. It is located at 120.98° E and 24.82° N and is spread over an area of 6.1 km2. As mentioned earlier, semiconductors and optoelectronics industries along with few biotechnology industries primarily constitute the industrial location along with residential areas, highways, and commercial regions within the park. The HSIP area was classified into five different sampling zones during the years 2000-2003 (Figure 1a-c). Different sampling points were selected to represent an average area of each zone. Sampling Strategy. This sampling campaign was performed in May and July 2000 and in four different seasons

TABLE 3. Sum of VOC Species (ppbv) Detected in Canister Air Samples Collected in 2002a sampling site

aliphatics

aromatics

oxygenated

sulfur compds

D1 D2 D3 D4 D7

9.7 12.7 9.8 11.8 24.1

March 14, 2002 5.2 21.2 7.7 21.6 8.7

8.4 19.4 300.2 4.0 9.8

nd nd nd 5.23 46.7

A1 A2 A3 B1 D1 D9 D8

5.2 11.7 2.8 6.2 4.8 10.0 4.1

May 22, 2002 4.6 5.2 0.6 0.8 1.2 5.3 3.5

6.7 13.0 nd 7.6 19.9 20.3 1.7

nd nd nd nd nd nd nd

A1 A2 A3 A4 B1 B2 D5 D6

25.7 11.1 12.1 13.5 12.0 20.9 12.8 12.9

September 10, 2002 10.9 5.5 5.3 5.9 7.0 8.1 3.9 5.8

9.3 6.7 6.4 5.4 6.8 52.5 10.2 5.9

0.72 nd nd nd nd 0.23 nd nd

A4 B1 B2 C1 D1 D10 E1

10.1 9.9 12.0 15.6 17.6 8.9 10.3

October 8, 2002 5.3 4.1 6.6 4.4 10.0 5.0 5.0

3.9 3.6 10.3 4.1 9.2 85.3 25.8

nd nd 0.19 0.18 0.37 0.73 0.24

a nd ) not detected. Aliphatics: ethane, ethylene, ethyne, propane, propene, n-butane, 2,2-dimethylbutane, 2,3-dimethylbutane, isobutene, isobutene, 1-butene, trans-2-butene, cis-2-butene, 2-methyl-2-butene, 2-methyl-1,3-butadiene (isoprene), n-pentane, isopentane, 1-pentene, trans-2-pentene, 2-methyl-pentane, 3-methylpentane, cyclopentane, methylcyclopentane, 2,2,4-trimethylpentane, n-hexane, 2-methylhexane, 3-methylhexane, cyclohexane, methylcyclohexane, 2-methylheptane, 3-methylheptane, n-octane, n-nonane. Aromatics: benzene, toluene, ethylbenzene, m,p-xylene, o-xylene, styrene, 1,2,4-trimethylbenzene, isopropylbenzene n-propylbenzene, 1,3,5-trimethylbenzene. Oxygenated: acetone, methyl tert-butyl ether, vinyl acetate, methyl ethyl ketone, 2-propanol (IPA), methyl isobutyl ketone. Sulfur compounds: dimethyl sulfide, carbon disulfide.

of 2002 and 2003, respectively. Continuous ambient air monitoring was carried out with GC-FID for a week (May 12-16) in 2000. Additionally, canister air samples were collected continuously every 2 h from stacks of some of the industries during May 19-24, 2000. Ambient air samples (grasp) were collected at nighttime between 10:00 and 11:15 PM with stainless steel containers from 26 and 25 different sampling points in July 2000 and 2002-2003, respectively. To avoid traffic exhaust, sampling was performed in the time period excluding rush hour. In 2002, March, May, VOL. 39, NO. 4, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

975

FIGURE 2. Sum of VOC species detected during four sampling seasons at all sampling sites in (a) 2000, (b) 2002, and (c, d) 2003.

TABLE 4. Sum of VOC Species (ppbv) Detected in Canister Air Samples Collected in 2003a April 21, 2003 June 3, 2003 October 2, 2003 November 18, 2003 sampling site aliphatics aromatics oxygenated aliphatics aromatics oxygenated aliphatics aromatics oxygenated aliphatics aromatics oxygenated A1 A2 A3 A4 B1 B2 B3 B4 B5 B6 B7 B8 C1 C2 C3 C4 C5 D1 D2 D3 D4 D5 E1 E2 E3

0.5 0.5 0.6 0.2 0.5 nd 2.4 0.5 0.5 nd 1.3 nd nd nd 0.5 nd nd nd nd nd nd nd nd nd 2.1

2.1 4.5 4.1 2.7 8.3 6.3 6.2 4.6 4.0 6.5 3.4 5.6 3.2 6.5 3.4 3.6 5.4 5.2 5.8 6.4 5.1 4.8 1.9 3.4 11.5

14.1 12.2 6.3 7.6 3.7 2.5 2.4 4.5 3.5 2.5 1.6 0.8 6.2 9.2 4.3 3.8 6.7 8.8 7.2 52.6 7.3 5.0 4.6 5.9 9.7

0.9 1.2 1.0 1.4 0.6 nd nd nd 0.5 0.7 nd 0.7 0.5 0.9 0.5 1.0 0.4 1.3 0.4 0.9 1.7 1.5 0.4 0.8 0.3

7.0 8.2 6.6 6.9 6.0 6.2 6.9 5.8 5.1 6.2 7.1 8.2 6.9 6.4 6.3 5.6 6.2 5.8 6.5 5.7 5.9 8.3 7.6 5.0 6.8

30.1 22.9 13.6 19.5 49.6 12.6 15.3 12.2 16.4 15.1 21.7 19.3 21.1 16.1 11.6 21.5 14.9 8.8 7.2 52.6 7.3 5.0 14.1 10.4 20.7

0.8 0.6 0.8 0.7 nd 0.9 nd nd 0.8 nd 0.9 0.6 0.6 0.7 nd nd 0.8 nd nd nd nd nd nd nd nd

2.0 2.3 3.3 2.4 5.6 2.9 5.5 3.1 4.0 3.5 2.0 2.8 1.6 1.9 3.2 2.1 4.5 2.0 2.2 1.9 2.2 2.5 3.9 2.3 1.7

21.0 3.4 9.7 12.5 2.9 7.1 7.0 3.3 4.1 6.5 4.6 8.1 4.7 6.0 4.2 5.6 21.2 6.3 4.0 15.1 2.9 3.2 9.2 3.0 4.4

nd nd nd nd 4.3 nd nd 1.2 nd nd nd nd nd nd nd nd nd nd nd nd 5.5 nd 0.4 0.6 0.6

10.5 10.2 10.3 11.3 19.0 16.8 13.0 15.8 14.9 14.5 12.8 22.5 8.8 10.0 11.4 9.2 3.4 11.2 12.9 10.6 15.2 8.6 10.6 13.0 10.7

82.0 12.8 6.7 56.8 18.5 12.6 8.6 11.6 23.8 14.4 21.8 11.3 6.7 68.8 14.7 43.9 23.9 14.6 13.0 25.5 12.9 13.3 12.4 6.3 11.0

a nd ) not detected. Aliphatics: propylene, n-hexane, cyclohexane, heptane. Aromatics: benzene, toluene, ethylbenzene, m,p-xylene, styrene, o-xylene, 1,2,4-trimethylbenzene. Oxygenated: ethanol, acetone, 2-propanol (IPA), methyl tert-butyl ether, ethyl acetate, methyl ethyl ketone, methyl isobutyl ketone.

September, and October months were selected, while in 2003, April, June, October, and November were chosen. In 2002, sampling was performed only at some of the sites in every season. Most of the sampling sites concentrated at two extremes, on the west and east side of the HSIP. For some 976

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 4, 2005

of the locations close to the busy highways, wind speeds and direction along with distance from highway was considered for interpretation. Also, for samples collected after rush hours of traffic, less interference due to vehicle emissions is interpreted.

The average wind speed and wind direction was 2.1 m s-1 southwest during the sampling week of May 2000. The wind speed recorded was 0.5 m s-1 southeast direction on July 25, 2000. The average ambient air temperature was varying between 25 and 28 °C at night. The wind direction and speed recorded in different sampling zones in 2003 at each sampling location are available in Supporting Information Table 1. The sampling period is characterized by very low wind speed in the range 0-3 m s-1. The zones, A in June and C in April, June, and November, recorded still air. In October, the wind was blowing from the land in a south direction. Sometimes the wind was characterized by the peculiar smell of solvents in zone C in October and in zone D in June. Instrumentation and Analysis. The traditional sampling methods and techniques were adopted for sample collection and analysis (9-10, 13-16, 20, 32). Continuous measurement of VOC was carried out with GC-FID (Varian 3800) by preconcentrating the samples onto an adsorbent trap containing Carbopack C, Carbopack B, and Carbosieve SIII. The trapped samples were desorbed into DB-Wax column (60 m × 0.53 mm × 1 µm) at a flow rate of 4 mL min-1. The FID detector temperature was 250 °C. The oven conditions are given in Supporting Information Table 2. The analytical procedures for canister air samples, collected during 2000-2003, included the standard determination of VOC referring to the standard method TO-15, as designated by the U.S. EPA. Samples were preconcentrated onto glass beads and analyzed using GC-MS according to the conditions detailed in Supporting Information Table 2. An automated GC-MS system (GC-FID 3800, Saturn-2000 MS) with an autosampler (Nutech A-3500, USA) analyzed simultaneously with better quality data and analytical efficiency. The calibration gas standards of 1 ppmv containing 93 different gas analytes (chlorinated, aliphatic, aromatic, and oxygenated volatile organics) were obtained from SPECTRA Gases. Quality assurance was carried out according to standard procedure during every sampling period of the years 2000-2003 separately. The performance of the system was almost similar in every campaign, and some of the results are given in Supporting Information Table 3. The concentration range of calibration curve was 1-50 ppbv, which gave a regression coefficient of 0.995 or above on repeating seven times. The relative standard deviation values for most of the measurements were less than 10% with very few exceptional cases showing >20% (see Supporting Information Table 3). In cases of high values, it is very likely that values are overestimated.

Results and Discussion VOC Concentration. Several compounds were detected from all samples in various years. The concentration of oxygenated and aromatic organics monitored continuously (for one week in May 2000) with GC-FID is summarized in Table 1. The diurnal variations showed high concentrations of acetone and benzene during the day while cyclopentanone and o-xylene during the night. 2-Propanol, toluene, and m,p-xylene showed same concentration variation during both day and night. During this sampling period, the wind was in southwest direction with an average wind speed of 2.11 m s-1. In July 2000, quantification was limited to aromatics and oxygenated compounds in canister air samples (Figure 2a). The sum of VOC detected at each site is summarized in Table 2. Among all the sites monitored, B4 (road junction) and B9 (silicon chip factory) measured high levels of sum of VOC (35 and 38 ppbv). The individual components that measured high at some of the sites are described: toluene, 12.7 ppbv at A2; acetone, 12.2 ppbv at D8; butanone, 4.5 ppbv at A6; ethyl acetate, 1.6 ppbv at B7; butyl acetate, 2.2 and 3.3 ppbv at C1 and C7, respectively; IPA, 29.5 and 18.5 ppbv at B9 and

FIGURE 3. Sum of VOC species detected in different sampling zones during four sampling seasons of (a) 2000, (b) 2002, and (c) 2003. In the year 2000, only oxygenated and aromatic compounds were measured. B4, respectively; benzene, 1 ppbv at C3 and C7; ethylbenzene, 1 ppbv at C1; m-xylene, ∼1 ppbv at C1; styrene and 1,3,5trimethyl benzene,