Distributions of Volatile Organic Compounds (vocs) in Indoor and Outdoor Air among Industrial, Urban and Suburban Neighborhoods

This paper presents results of a comprehensive survey conducted to better understand concentrations and spatial variations of indoor and outdoor concentrations of volatile organic compounds (VOCs) in industrial, urban and suburban neighborhoods. Three cities, representing these types of neighborhoods in Michigan, USA were selected, and 30 to 50 homes were selected in each city. Duplicate or triplicate thermal desorption tube (TDT) passive samplers were placed inside and outside of each home for 3 4 weekdays in the summer and fall of 2004 and analyzed for 90+ compounds. Related home information was collected simultaneously. Thirty six VOCs, including aromatics, chlorinated compounds, terpenes and alkanes, were frequently detected (>50% of indoor samples) with good precision (duplicate variability generally <15%). Indoor/outdoor concentration ratios ranged from 1 to 82 for the most frequently detected VOCs. Outdoor concentrations showed generally modest spatial variation, and levels depended on the city. In contrast, indoor levels varied greatly among residences, and levels showed only minor dependences by city. Most of the indoor variability was attributed to building and occupant factors. INDEX TERMS Exposure, Benzene, Terpenes, Variability, Volatile organic compounds, VOCs INTRODUCTION Substantial efforts have been made to measure indoor and outdoor concentrations of volatile organic compounds (VOCs) for the purpose of quantifying human VOC exposures. Residences are of great interest as they represent the environment where most people spend the majority of their time. A survey of 876 homes in England found that VOC concentrations were mostly related to building materials in new homes, painting and decorating events, suggesting that source emissions were more important than ventilation (Raw et al. 2004). In Finland, Kostiainen et al. (1995) found elevated VOC levels in 38 houses where occupants experienced SBS symptoms as compared to 50 “normal” houses. A study of 757 Canadian residences found VOC concentrations from <1 to 104 μg m, and many compounds were associated with tobacco smoking, consumer products and other materials (Otson et al. 1994). A wide range of indoor/outdoor (I/O) concentration ratios has been found, for example: Son et al. (2003) found I/O ratios of 0.94 to 1.51 for nine VOCs; Sarwar et al. (2002) showed typical ratios of about 3; Payne-Sturges et al. (2004) found ratios from 6 to 7 in 33 nonsmoking household; and Rehwagen et al. (2003) found average ratios of about 10 in Leipzig, Germany. I/O ratios as well as the types and levels of VOCs found in homes depend on building characteristics, human activities, and outdoor sources. Schlink et al. (2004) indicated that seasonality was the most predominant effect in indoor VOC concentrations, based on 2103 measurements in 3 different cities in Germany. Homes with attached garages have shown increased levels of benzene, xylene, ethylbenzene and toluene (Graham et al. 2004, Tsai and Weisel 2000). Several studies have reported city-to-city variation in both indoor and outdoor VOC concentrations. For example, Sakai et al. (2004) found higher indoor and outdoor levels of formaldehyde and selected chlorinated VOCs in Nagoya, Japan than in Uppsala, Sweden, attributed to the presence of more indoor sources (including construction and interior materials, unvented combustion space heaters, and moth repellents containing p-dichlorobenzene in Nagoya. Son et al. (2003) found that indoor and outdoor VOC concentrations in homes in Seoul, South Korea were significantly higher than those in Asan, a medium sized city. Despite these and many other studies, VOC exposure remains inadequately understood. Most studies have been * Corresponding author email: [email protected] Proceedings: Indoor Air 2005 2631 small, considerable spatial and temporal variation occurs in most environments, comparisons between countries may be problematic, and conventional ambient air monitoring sites yield very limited spatial and temporal resolution. The most comprehensive studies conducted to date, including the TEAM studies in the 1980s (Wallace, 2001), NHEXAS in the 1990s (Roberson et al. 1999) and EXPOLIS in the late 1990s (Saarela et al. 2003), have provided valuable information regarding distributions and variability of indoor and outdoor VOC concentrations. Most Americans live in or near industrial, urban and suburban areas (79% in the 2000 census), but no study has characterized the distribution and variability of VOC levels in a comprehensive and validated manner. Moreover, with awareness of ETS and IAQ issues and the ensuing reductions in indoor emissions, older studies may have limited relevance to current exposures. This paper presents selected results from a comprehensive survey designed to better understand the current concentrations and the spatial variation of indoor and outdoor VOCs in industrial, urban and suburban neighborhoods. EXPERIMENTAL Three southeast Michigan cities, Ypisilanti (shortened as Y), Ann Arbor (A), and Dearborn (D), were selected to represent urban, suburban and urban/industrial neighborhoods, respectively. In each city, 30 to 50 homes were randomly selected. Duplicate and triplicate passive thermal desorption tube (TDT) type samplers were placed inside and outside of homes for 3 to 4 weekdays in summer and fall 2004, giving an effective sampling volume of ~2 L (Batterman et al. 2002). Laboratory and field blanks were collected simultaneously (10% of samples). CO2, temperature and relative humidity were continuously monitored at each sampling locations. A questionnaire and walkthrough survey was completed for each home examining basic housing, work and family characteristics, hobbies and other information potentially related to VOC exposure. Recruitment and other activities followed informed consent and other procedures approved by the University of Michigan’s institutional review board. Ninety-six VOCs were selected based on their toxicity and occurrence in indoor and outdoor air. Many of these VOCs are listed in the TO-14 and 15 methods, and they included aromatics, halogenated compounds, terpenes, alkanes, etc. Collected samples were analyzed by GC-MS (Agilent 6890/5973) following methods in Peng and Batterman (2000) using two GC-MS protocols: a full scan method with ions ranging from 26 to 250 amu, and a selective ion monitoring method using 16 time windows over the 37 min GC run time. These methods gave comparable results with the exception that SIM mode was more sensitive (Jia et al 2005). A detailed QA/QC protocol was developed and used following TO-17 (US EPA 1999). Descriptive statistics (averages, standard deviations, medians, and detection probabilities) along with the total VOC (TVOC) concentration were calculated for each city. Six VOCs were selected for further analysis, namely, benzene, toluene, chloroform, tetrachloroethene (or perchloroethene, PERC), d-limonene and naphthalene. These compounds represent different classes of compounds and different types of sources. The precision of duplicate or triplicate VOC samples measuring ≥1 μg m averaged 15% or better. Replicates were averaged, and non-detects were set to one-half of the MDL. Log-transformations were used to obtain normally distributed data used for hypothesis testing and multivariate analyses (correlations, statistical inferences and ANOVAs). Statistics were calculated using Excel and SYSTAT V.10.0 (SPSS Inc, Chicago, IL). RESULTS Distributions of VOC concentrations A total of 49 compounds were detected indoors and 44 compounds outdoors. Of these, 36 compounds were found indoors with detection probabilities over 50% indoors, and 30 compounds outdoors. Statistics of the 18 most frequently detected VOCs (and TVOC) are summarized in Table 1. Aromatics were detected in over 90% of the samples, often with the highest concentrations, both indoors and outdoors. Naphthalene, a possible human carcinogen, had indoor concentrations averaging over 3 μg m, the US EPA reference concentration. Chlorinated compounds were detected frequently except in the ambient air in Ypsilanti (60%); average concentrations were below 2 μg m. Carbon tetrachloride, a ubiquitous compound, was consistently detected indoors and outdoors with a generally stable concentration of near 1 μg m. Terpenes (d-limonene and α-pinene) were also detected frequently indoors and out, often at concentrations exceeding 10 μg m in homes; outdoor levels were below 1 μg m. Low molecular weight aliphatics, e.g., n-nonane and methyl cyclohexane, were frequently detected indoors and outdoors. Heavier alkanes, e.g., hexadecane, were found infrequently in ambient air. Detection probabilities and concentrations of several VOCs were related to specific household activities or human behaviors. Houses with high concentrations of ethyl acetate and heavier alkanes were usually undergoing remodeling and painting. 2,5-Dimethyl furan, a known tobacco smoking indicator, was founded in the few houses sampled where residents smoked inside. Proceedings: Indoor Air 2005 2632 Concentrations of the selected compounds were log-normally distributed. Figure 1 shows the cumulative probability distributions of indoor and outdoor concentrations of d-limonene and naphthalene. For d-limonene, both indoor and outdoor concentrations could be ranked as D>Y>A (Figure 1A). For naphthalene, indoor levels in the three cities were similar; outdoor levels could be ranked as D>A>Y (Figure 1B). Indoor concentrations had much more variation than outdoor concentrations. Because the data was highly skewed, the relative interquartile range (RIQR = interquartile range/median) was calculated to represent variability. For indoor concentrations, the RIQR for all compounds ranged from 100 to 300%. In contrast, outdoor RIQRs were generally below 100. Table 1. Concentrations (μg m) of 18 VOCs and TVOC in indoor and ambient air. City Indoor / Outdoor Statistics Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Benzene 4.06 5.59 0.77 0.28 3.47 8.35 0.73 0.23 3.06 3.86 1.67 0.88 Toluene 28.00 33.37 1.56 0.84 15.61 30.49 1.14 1.51 14.82 29.61 4.25 2.75 p-Xylene,m-Xylene 9.77 16.43 0.83 0.41 13.89 51.67 1.59 5.07 9.01 15.44 3.22 1.99 1,2,4-Trimethylbenzene 3.71 5.16 0.29 0.14 2.29 5.81 0.21 0.15 3.59 7.53 1.10 0.69 Styrene 0.90 1.06 0.04 0.08 0.28 0.36 0.02 0.04 0.62 0.49 0.09 0.06 Naphthalene 6.16 15.64 0.14 0.10 1.68 2.49 0.38 0.77 3.34 7.88 0.42 0.23 1,4-dichlorobenzene 12.31 52.65 0.16 0.47 1.29 3.65 0.03 0.04 4.83 21.33 0.20 0.29 Chloroform 0.75 0.75 0.04 0.03 0.22 0.33 0.06 0.02 1.33 2.23 0.13 0.37 Carbontetrachloride 1.21 1.23 0.74 0.25 0.91 0.21 0.72 0.23 1.09 0.64 1.11 0.45 Trichloroethylene 0.04 0.20 0.01 0.04 0.02 0.04 0.01 0.03 0.06 0.11 0.04 0.03 Tetrachloroethene 0.46 0.49 0.19 0.24 0.65 0.85 0.22 0.18 1.28 3.82 0.56 0.45 a-Pinene 21.26 33.90 0.37 0.26 14.78 31.23 0.29 0.17 5.76 8.59 0.21 0.40 d-Limonene 17.63 11.45 0.21 0.11 10.63 16.09 0.15 0.12 37.46 44.65 1.05 2.18 n-Nonane 6.86 23.30 0.15 0.17 0.71 1.15 0.09 0.08 1.64 2.50 0.32 0.21 n-Hexadecane 0.80 0.96 0.02 0.04 0.28 0.16 0.00 0.00 0.62 0.74 0.02 0.08 Methyl Cyclohexane 1.46 2.04 0.07 0.04 1.12 3.71 0.06 0.05 6.31 36.42 0.31 0.55 Ethyl Acetate 6.71 19.85 0.00 0.00 7.01 19.00 0.00 0.00 2.70 5.76 0.00 0.02 2,5-Dimethyl furan 0.02 0.08 0.00 0.00 0.00 0.02 0.00 0.00 0.09 0.22 0.01 0.03 TVOC 187.21 204.92 6.95 2.81 92.64 144.26 6.88 9.18 129.83 134.62 20.24 10.86 Ypsilanti Ann Arbor Dearborn Indoor Outdoor Indoor Outdoor Indoor Outdoor A. d-Limonene B. Naphthalene 0.0 0.1 1.0 10.0 100.