Polycyclic Aromatic Hydrocarbons in the Air of Ten

NUMBER 6 2005

Polycyclic Aromatic Hydrocarbons in the Air of Ten Chicago Area Homes

An Li, Todd M. Schoonover, Qimeng Zou, Felice Norlock, Peter A. Scheff, and Richard A. Wadden

ABOUT THE LELAND CENTER The Mickey Leland National Urban Air Toxics Research Center (NUATRC or the Leland Center) was established in 1991 to develop and support research into potential human health effects of exposure to air toxics in urban communities. Authorized under the Clean Air Act Amendments (CAAA) of 1990, the Center released its first Request for Applications in 1993. The aim of the Leland Center since its inception has been to build a research program structured to investigate and assess the risks to public health that may be attributed to air toxics. Projects sponsored by the Leland Center are designed to provide sound scientific data useful for researchers and for those charged with formulating environmental regulations. The Leland Center is a public-private partnership, in that it receives support from government sources and from the private sector. Thus, government funding is leveraged by funds contributed by organizations and businesses, enhancing the effectiveness of the funding from both of these stakeholder groups. The U.S. Environmental Protection Agency (EPA) has provided the major portion of the Center’s government funding to date, and a number of corporate sponsors, primarily in the chemical and petrochemical fields, have also supported the program. A nine-member Board of Directors oversees the management and activities of the Leland Center. The Board also appoints the thirteen members of a Scientific Advisory Panel (SAP) who are drawn from the fields of government, academia and industry. These members represent such scientific disciplines as epidemiology, biostatistics, toxicology and medicine. The SAP provides guidance in the formulation of the Center’s research program and conducts peer review of research results of the Center’s completed projects. The Leland Center is named for the late United States Congressman George Thomas “Mickey” Leland from Texas who sponsored and supported legislation to reduce the problems of pollution, hunger, and poor housing that unduly affect residents of low-income urban communities.

This project has been funded wholly or in part by the United States Environmental Protection Agency under assistance agreement R828678. The contents of this document do not necessarily reflect the views and policies of the Environmental Protection Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use.


Principal Investigator: An Li

Co-Investigators: Peter A. Scheff and Richard A. Wadden Research Assistants: Todd M. Schoonover, Qimeng Zou, Felice Norlock

School of Public Health University of Illinois at Chicago

TABLE OF CONTENTS

PREFACE ABSTRACT INTRODUCTION PROBLEM STATEMENT BACKGROUND PAHs in the Indoor Environment Sources of Indoor PAHs Factors Affecting Indoor PAH Levels Estimation of Indoor Emission Rates

HYPOTHESES AND OBJECTIVES METHODOLOGY HOME RECRUITMENT AND SELECTION SAMPLING CHEMICAL ANALYSIS Chemicals Sample Pretreatment GC/MS Analysis Large Volume Injection PTV-GCMS QA/QC DATA MANAGEMENT AND STATISTICAL METHODS

RESULTS AND DISCUSSIONS DATA SUMMARY HOME CHARACTERISTICS AIR QUALITY MONITORING DESCRIPTION OF PAH CONCENTRATIONS Overview Seasonal Variations Short Term Variations Comparison Among Homes Indoor vs. Outdoor PAH Concentrations AIR EXCHANGE AND INDOOR EMISSION RATES Air Exchange Derived from Statistical Model Air Exchange Derived from CO2 Monitoring Data Indoor Emission Rate Estimation FACTORS AFFECTING PAH CONCENTRATIONS Surrounding Environment

NUATRC RESEARCH REPORT NO. 6

1 2 2 2 3 3 4 4 5 6 6 6 6 7 7 8 8 9 9 10 11 11 12 13 14 14 17 19 20 23 29 29 29 30 31 31

TABLE OF CONTENTS (cont.)

House Characteristics Home Occupancy and Activities Air Quality DUST SAMPLES

CONCLUSIONS IMPLICATION OF FINDINGS LITERATURE CITED ACKNOWLEDGEMENTS ABOUT THE AUTHORS PUBLICATIONS RESULTING FROM THIS STUDY ABBREVIATIONS APPENDICES A.

Household Screening Survey Questionnaire

B.

Home Status Survey During Sampling

C.

Detecyion Limits

D.

Characteristics of Participating Homes

E.

Surrogate Recoveries

F.

PAH and Air Quality Data

G.

Home Status Survey Questionnaire Responses

H.

Indoor/Outdoor Ratios of Individual PAHs by Home

I.

Air Exchange Rate and Indoor Emission Estimates

32 32 33 34 34 35 36 38 38 38 39 41 41 45 49 51 57 65 83 87 93


PREFACE The Clean Air Act Amendments of 1990 established a control program for sources of 188 “hazardous air pollutants, or air toxics,” which may pose a risk to public health. Also, with the passage of these Amendments, Congress established the Mickey Leland National Urban Air Toxics Research Center (NUATRC) to develop and direct an environmental health research program that would promote a better understanding of the risks posed to human health by the presence of these toxic chemicals in urban air. Established as a public/private research organization, the Center’s research program is developed with guidance and direction from scientific experts from academia, industry, and government and seeks to fill the gaps in scientific data. These research results are intended to assist policy makers in reaching sound environmental health decisions. The NUATRC accomplishes its research mission by sponsoring research on human health effects of air toxics in universities and research institutions and by publishing research findings in its “NUATRC Research Reports,” thereby contributing meaningful and relevant data to the peerreviewed scientific literature. The study “Polycyclic Aromatic Hydrocarbons in the Air of Ten Chicago Area Homes” was developed in response to the Mickey Leland National Urban Air Toxics Research Center (NUATRC) Request For Application 98-03 (RFA 98-03), under the NUATRC Small Grants New Investigators Award Program. This NUATRC support was developed to encourage New Investigators to develop innovative air toxics research areas. This specific RFA was developed to support shortterm research projects on exposures and/or health effects of air toxics. The projects were envisioned to be community-based pilot projects that could serve as a basis for more extended research. Preference was given to research that tested new techniques and/or were innovative or high-risk projects. Pilot projects by their nature are limited in scope and are designed to give preliminary results and new insights so that larger studies, yielding results with broader applications, might follow. Dr. An Li of the University of Illinois at Chicago was the a recipient of this NUATRC award. The underlying objective of the study was to collect data on indoor and outdoor polycyclic aromatic hydrocarbons (PAHs) over time, in individual homes and to collect some limited information on PAH-generating activities that might be taking place during the sampling period. Dr. Li used this


information to conduct an initial assessment of the contributions to outdoor sources, evaluation of indoor sources, and examination of factors that affect indoor concentrations. This report describes the nature and quality of the data and the shape of the data distributions, initial descriptive analyses, and interpretation of the results. When a NUATRC-funded study is completed, the investigators submit a draft final research report. All draft final reports resulting from NUATRC-funded research undergoes an extensive evaluation procedure which assesses the strengths and limitations of the study, comments on clarity of the presentation, data quality, appropiateness of study design, data analysis, and interpretation of the study findings. The objective of the review process is to ensure that the investigator’s report is complete, accurate, and clear. The evaluation first involves an external review of the report by a team of three external reviewers, including a biostatistician. The reviewers’ comments are then considered by members of the NUATRC Scientific Advisory Panel (SAP), and the comments of the external reviewers and the SAP are provided to the investigator. In its communication with the investigator, the SAP may suggest alternate interpretations for the results and also discuss new insights that the study may offer to the scientitfc literature. The investigor has the opportunity to exchange comments with the SAP and, if necessary, revise the draft report. In accordance with the NUATRC policy, the SAP recommends and the Board of Directors approves the publication of the revised final report. The research presented in the NUATRC Research Reports represents the work of its investigators. The NUATRC appreciates hearing comments from its readers from industry, academic institutions, government agencies, and the public about the usefulness of the information contained in these reports and about other ways that the NUATRC may effectively serve the needs of these groups. The NUATRC wishes to express its sincere appreciation to Dr. Li and her research team, the SAP, and external peer-reviewers whose expertise, diligence, and patience have facilitated the successful completion of this report.

1


ABSTRACT

INTRODUCTION

This study was developed in response to a Mickey Leland National Urban Air Toxics Research Center’s (NUATRC) New Investigators Small Grants Request for Applications. The goals of the projects are to study exposures and health effects or urban air toxics and are envisioned as pilot projects that could serve as a basis for more extended future research. Indoor and outdoor air samples at ten non-smoker homes in the Chicago area were concurrently collected once per month for a 14-month period starting June 2000. During sampling, temperature, humidity, CO2, and CO were recorded. Questionnaire surveys were used to register household activities. For each sample, 16 polycyclic aromatic hydrocarbons (PAHs) were measured using large volume injection gas chromatography and mass spectrometry (GC/MS). The total concentrations of the 16 PAHs in the indoor air ranged from 13 to 2,454 ng/m3, with a median of 208 ng/m3. The median for the outdoor air samples was 213 ng/m3, and the concentration ranged from 13 to 1865 ng/m3. The most abundant PAH in both indoor and outdoor air was naphthalene. Next to naphthalene, phenanthrene and fluorene were found to be the most abundant among all PAHs in outdoor air, while anthracene and fluorene had higher indoor concentrations than other PAHs. The correlations of indoor and outdoor concentrations appeared to be weak for light PAHs, but reasonably strong for PAHs with molecular mass 228 or higher. Indoor-to-outdoor ratios (I/O) had a mean of 2.5 and a median of 1.02 for the sum of the 16 PAHs. The medians of I/O were less than or very close to 1.0 for all PAHs except anthracene. The results suggest that, in general, the impact of indoor activities on the indoor PAH level is greater than the influence of the outdoor air. Seasonal variations were significant for indoor total PAHs, with fall having a higher median than other seasons. Ventilation rates were calculated using two models, and the indoor emission rates were estimated using a onecompartment mass balance model. The results of this work show that PAHs are present in non-smoker homes, but the concentration levels are in general lower than those reported for smoker homes. The data also show that PAH levels in urban residential environments vary considerably, reflecting the complex nature of the sources.

PROBLEM STATEMENT

2

Polycyclic aromatic hydrocarbons (PAHs) are a group of compounds that consist of two or more fused benzene rings. They are the by-products of incomplete combustion or pyrolysis of virtually all organic matter (Harvey, 1991). Since 1775 when the carcinogenic effects of PAH-rich combustion chimney soot were discovered (see Harvey, 1991), a tremendous amount of research has been done to understand their sources, physicochemical properties, environmental behavior, fate, carcinogenecity, and other toxicological health effects. PAHs are considered to be the most widely distributed class of potent carcinogens present in the human environment, and many of them are listed as known, probable, or possible carcinogens by various international and national agencies. In addition to cancer, PAHs have also been found to cause morphological, physiological, and developmental abnormalities in test animals, increase allergic immune responses in humans at low levels, and act synergistically with other air toxics such as ozone in causing adverse health effects (Harvey, 1997, 1991; Seymour et al., 1997; Diaz-Sanchez et al., 1996; USEPA, 1991). Since most people spend more than half of their time at home, human exposure to PAHs in the indoor environment has received increasing attention in recent years (Naumova et al., 2002; Ohura et al., 2002; Levy et al., 2002; Chao et al., 2002; Van Winkle and Scheff, 2001; Zhu and Liu, 2001; Wallace, 2000; Sugiyama et al., 2000; Li and Ro, 1999; Prince et al., 1999; Dubowsky et al., 1999; Van Winkle, 1996a, 1996b; ATSDR, 1996; Chuang et al., 1991; Otson et al., 1991). Studies conducted so far have shown that (1) the concentration of PAHs in residential indoor air is generally higher than outdoors, indicating the existence of indoor sources; (2) cigarette smoking is the predominant source of PAHs in homes with smokers; and (3) outdoor air may contribute significantly to the indoor PAHs. In addition, most indoor PAHs are associated with particles with aerodynamic diameters less than 2.5 µm (Koyano et al, 2001; Sugiyama et al., 2000). These small particles are able to reach the lower respiratory tract in humans, and thus, raise concerns about their effects on public health. Although much has been learned, knowledge of the behavior of indoor PAHs is still very limited. Few data are available on the indoor PAH emission rates. Little research has been conducted on the movement and indoor-outdoor exchange or the attenuation and fate of PAHs in residential


An Li

settings. Information about PAH partitioning between gas and air-borne particles and between air and solid surfaces in the indoor environment is also lacking. Reported patterns of seasonal variations of PAH levels in urban environments were inconsistent (Dimashki, et al., 2001). It is also not clear how indoor PAH levels correlate with or are affected by factors such as geographic locations, house characteristics, residents’ occupations, lifestyles, and daily activities, etc. BACKGROUND PAHs in the indoor environment Air pollution is one of the major pathways of human exposure to PAHs. Over the past decades, numerous studies have investigated PAHs in ambient air worldwide. However, other than cigarette smoking, there has been relatively little research on indoor exposures to PAHs until recent years. In an 8-home pilot study in Columbus, OH in 1986-1987 by Chuang et al. (1991), the most abundant indoor PAH was found to be naphthalene. This finding is supported by another study conducted by the Illinois Department of Public Health (IDPH) and Agency for Toxic Substances and Disease Registry (ATSDR) in 1994-1995, which revealed indoor mean naphthalene concentrations of 851 ng/m3 (Van Winkle and Scheff, 2001). The reported

naphthalene concentrations ranged from 550 to 1,800 ng/m3 in kitchens and from 800 to 4,200 ng/m3 in living rooms in the Ohio study (Chuang et al., 1991). Other representative mean naphthalene concentrations are 2,300 ng/m3 (Krause et al., 1987), 11,000 ng/m3 (DeBortoli et al., 1986), and 12,000 ng/m3 (Otson et al., 1994). Kostiainen (1995) measured the indoor volatile organic compounds (VOCs) in 50 homes where no Sick Building Syndrome was found among residents and reported a median naphthalene concentration value of 310 ng/m3, which is significantly lower than that found in the sick building homes. Besides naphthalene, fluorene was also found to be ubiquitous in 45 indoor air samples collected in 10 homes in southeast Chicago in 1994-1995 (Van Winkle, 1996a). Other PAHs often found in these homes included fluoranthene, pyrene, acenaphthene, phenanthrene, and acenaphthylene. Higher molecular weight PAHs such as benzo[ghi]perylene and coronene were also found (CARB, 1994; Chuang et al., 1991). Recently, Naumova et al. (2002) investigated PAHs in 55 nonsmoker residence homes in heavily industrialized areas of three US cities. With five most volatile PAHs including naphthalene being excluded, the sum of 30 PAHs ranged from 16 to 350 ng/m3 for indoor and 4.0 to 160 ng/m3 in outdoor air. Concentrations of several PAHs reported from literature sources are summarized in Table 1. Some PAH

3)a Table 1. for PAH concentrations in indoorinairIndoor of residence (ng/mHomes Table 1. Literature Literaturereferences References for PAH Concentrations Air of homes Residence (ng/m3) a

Naphthalene Acenaphthylene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benz[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene Indeno[123-cd]pyrene Dibenz[a,h]anthracene Benzo[ghi]perylene

Van Winkle, 1996a b

Mitra and Ray (1995) c

USEPA, 1991d

851 (nd-5000) 12.8 (nd-122) 79.9 (nd-1040) 59.0 (nd-600) 60.0 (nd-820) 2.05 (nd-40) 6.64 (nd-26) 6.91 (nd-47) 0.56 (nd-14) 0.20 (nd-3) 0.07 (nd-3) 0.07 (nd-3) 0.09 (nd-3)

1060 13.2

1000 10

84.5 2.41 11.3 7.01 0.42 1.17

59 2.0 7.2 5.6 0.24 0.93

0.44 0.49

0.31

0.64

CARB, 1994 e

19

Naumova, et al, 2002 f

27 (9.1-330)

2.3

0.52

0.051 (0.0027-0.57)

2.1

a. Average (minimum - maximum). b. Samples collected in nonsmoker homes. Some data were also reported in ATSDR (1996) nd=non-detectible. c. Samples collected in nonsmoker homes. d. Samples collected in the living room of nonsmoker homes. e. Samples collected during daytime. f. Average of geometric means from three US cities.


3


derivatives, including nitro-PAHs and oxygenated PAHs, were also found in indoor air, with concentrations lower than those of their parent compounds (Chuang et al., 1991). Sources of Indoor PAHs PAHs originate predominantly from anthropogenic processes, although their natural formation in the environment can be substantial in the case of forest fires or volcanic eruptions. In indoor environments, studies have indicated that smoking is a major source variable influencing non-industrial indoor PAH exposure (Chuang et al., 1991). Tobacco smoke is a complex mixture containing more than 150 compounds in the gas phase and more than 2,000 components in the particulate phase, which includes numerous PAHs (Hoffmann et al., 1978). However, other sources of indoor PAHs cannot be ignored, especially for non-smoker homes. These nontobacco-related sources include, for example, indoor penetration of outdoor PAHs that are released from local industrial, transportation, and recreational sources, and indoor activities such as cooking, operation of heating and cooling systems, chemical storage and usage, and garage activities. In addition, characteristics of homes such as occupancy, the air exchange rate, type of furnishings, garage attachment, and humidity may also affect the levels and distributions of PAHs in homes. Outdoor air may contribute significantly to the indoor PAHs. In a heavily industrialized and densely populated area in south Chicago, the total concentration of 15 PAHs was found to be 507 ng/m3, compared with 19.0 ng/m3 in a remote location in Wisconsin (Cotham and Bidleman, 1995). Comparable PAH concentrations were also reported by Harner and Bidleman (1998) and USEPA (1996) in the same area. Naumova et al. (2002) reported strong correlations (R2 = 0.64 - 0.99) between indoor and outdoor concentrations for the sum of 5-7 ring PAHs in selected homes of three US cities. Based on findings by Sweet and Vermette (1992), outdoor concentrations of airborne pollutants may be influenced significantly by location, wind direction, periodic release episodes, time, and other factors. However, the outdoor samples collected in closest proximity to the sampling house were found to pair with indoor observation results (Van Winkle and Scheff, 2001). Higher indoor concentrations were observed when compared to the outdoor levels for most PAH compounds (Van Winkle, 1996a; Chuang et al., 1991), indicating the

4

presence of indoor sources of PAHs. The mean indoor/outdoor concentration ratios of PAHs in Van Winkle’s study (1996a, 1996b) were 4.1 for 3-ring PAHs, 1.3 for 4-ring PAHs except chrysene, but less than 1.0 for 5-ring benzo(a)pyrene and benzo(e)pyrene as well as chrysene. These ratios ranged from about 0.5 to 2.0 in a southern California study and were found to be higher in smoker homes (CARB, 1994). Naumova et al. (2002) reported indoor/outdoor ratios less than 1.0 for PAHs with five to seven rings in most studied homes. Based on the indoor/outdoor ratio, PAHs with two or three rings seem to have the strongest indoor sources, while heavy PAHs in indoor air are primarily from outdoor sources. However, since PAHs with more than four rings have a strong tendency to attach to solid surfaces, their amounts in the indoor environment may be underestimated from the air concentration data alone. Of indoor PAHs bound to airborne particulates, a high percentage were found in the fine fraction (less than 2.5 µm diameter) of the particles (Sugiyama et al., 2000). Factors Affecting Indoor PAH Levels Few studies have evaluated factors which affect the levels of indoor PAHs. Chuang et al. (1991) evaluated the contributions of different heating and cooking systems to indoor PAH exposure. Homes with gas heating systems appeared to have higher indoor PAH concentrations than those with electric systems. However, it was concluded that estimated effects of gas/electric heating and cooking systems were not well characterized due to the small sample sizes and the lack of statistical significance of observed differences in their study. Van Winkle (1996a) used linear regression analysis to examine the association between several parameters (predictor variables) and the measured indoor concentrations. The emission factor, which is defined as the difference between the mean emission rate with and without the source variable was then estimated for each of the 16 PAHs as well as selected VOCs and metals. The highest emission factors were estimated to be 796 µg/hr from electric heaters and 675 µg/hr from mothball storage for naphthalene. Of the housing characteristic predictor variables observed in more than one home, mothball storage, washer/drier-in-utility-room, and recentremodel-activity were found to have the highest number of significant positive associations with increased PAH emissions. Homes with mothball storage had significantly higher emission rates of naphthalene, acenaphthene, fluorene, and phenanthrene (Van Winkle,


An Li

1996a). Due to the small data set, however, only associations between a specific pollutant and a single predictor variable were modeled in this study; regression models with multiple predictor variables were not developed. Little information is available regarding other factors, such as house structure and location, occupation, age, and race of the residents on the indoor PAH level. Some source variables were found to be correlated with such factors. For example, in 10 Chicago homes, Van Winkle (1996a) found that the mothballs-stored-in-home variable was correlated with Hispanic occupants (100%) and the location Torrence, and the washer/drier-ventedindoors variable was positive for 80% of the homes in Altgeld area. He also reported that a higher number of employed residents had an association with increased concentrations of a few halogenated compounds and decreased concentrations of naphthalene, acenaphthene, and phenanthrene (p ≤ 0.05). These results, however, should be interpreted with caution due to the small number of observations. Estimation of Indoor Emission Rates Mathematical models are often used to characterize chemical or physical processes that affect indoor pollutant concentrations and to predict indoor concentrations or exposure under different circumstances. A one-compartment mass balance model, as shown in Equation 1, has been used to estimate expected indoor pollutant concentrations (Wadden and Scheff, 1982):

[1]

where C is the concentration indoors (Ci) and outdoors (Co); t is time; q is volumetric flow rate for make-up air (q0), recirculation (q1), and infiltration (q2); F is filter efficiency for make-up air (F0) and recirculation air (F1); V is room volume; S is indoor source emission rate; R is indoor sink removal rate; and k is a mixing factor which varies from 0 to 1 (Wadden and Scheff, 1982). A steady state solution for Equation 1 can be arrived at by setting the left hand side of Equation 1 to zero. This results in Equation 2: [2]


where Ci,ss is the steady-state indoor concentration and E is a proportionality constant for the particular pollutant of interest, such that R = ECi. For particulate pollutants, E can be estimated as the product of the surface area and the particle deposition velocity. A modified form of the steady-state model that incorporates an estimate of pollutant penetration and deposition loss can be represented by: [3]

Assumptions for Equation 3 include: (1) no resuspension of the settled pollutants; and (2) a constant emission rate over the measurement or modeling interval. This simplified model specifies that the indoor pollutant concentration is dependent on the outside penetration (PCo), the indoor chemical emission (S) and removal (E) rates, and the effective ventilation rate (kq). In Equation 3, P is the penetration factor that describes the amount that the infiltration of chemicals is reduced by the filtering effect of the building shell. P greater than unity are physically unrealistic, but the change in outdoor concentration or indoor source activities can cause an apparent penetration factor greater than 1.0 due to unsatisfied equilibrium conditions or particle resuspension (Thatcher and Layton, 1995). Koutrakis and Briggs (1992) estimated P values ranging from 0.58 to 1.01 for 17 elements, Dockery and Spengler (1981) reported P of 0.65 for sulfate particles, and Wallace (1996) and Thatcher and Layton (1995) calculated P of approximately 1.0. Van Winkle (1996a) obtained average P of 0.89 for seven elements and for PAHs. Equations 1, 2, and 3 can be used to transform measured concentration to emission rate, S. Specifically, the non-steady-state solution of Equation 1 can be used with changes in occupancy to calculate the flow of ventilation air through the home, while Equation 3 can be used to convert indoor air concentrations to indoor emission rates. This and similar one-compartment models have been used intensively in indoor air quality studies (Chuang et al., 1991; Fisk et al., 1987; Grimrud et al., 1984). In our previous studies, the model has been applied to indoor bioaerosol and VOC emissions (Wadden et al., 1997; Rizzo, 1996; Van Winkle, 1996a, 1996b; Wadden et al., 1995a, 1995b; Wadden et al., 1994; Scheff et al., 1992; Wadden et al., 1991). Van Winkle (1996a) used Equation 3 to estimate indoor emission

5


rates for PAHs in the 10 non-smoker homes in southeast Chicago. The resulting mean emission rate is 245 µg/hr for naphthalene, 36.7 µg/hr for acenaphthene, 22.8 µg/hr for fluorene, 20.0 µg/hr for phenanthrene, 4.36 µg/hr for acenaphthylene, and 1.55 µg/hr for pyrene.

HYPOTHESES AND OBJECTIVES The basic hypotheses for this project were that PAHs in outdoor air contribute significantly to the levels of indoor PAHs in urban homes and that they are also generated from indoor activities and sources. The overall objectives of this study are to (1) assess the contributions from outdoor sources to the polycyclic aromatic hydrocarbons (PAHs) in urban homes; (2) to evaluate the sources of indoor PAHs; and (3) to examine the factors that control or modify the indoor air concentrations of PAHs.

Chicago area; (3) living in his/her present house for the next 14 months after the initial introductory meeting; (4) having no smokers in the family; and (5) allowing the air in his/her home to be sampled for a 48-hour period once a month for at least 12 months beginning summer 2000. Homes with smokers were excluded in order to avoid the potential confounding effect of other sources from more predominant cigarette smoke. Only single family homes were chosen in an attempt to further standardize the data for comparison. A Household Screening Survey Questionnaire (Appendix A) was used to select participants from all volunteers who responded. The purpose of this initial survey was to select and characterize the homes. Question categories were: 1. 2.

The specific aims of this project included: 3. 1.

2.

3.

To mesurement of the concentrations of 16 selected PAHs in 10 homes. Each home would be sampled once per month for a year, with simultaneous outdoor sampling. Measurement of indoor/outdoor air exchange or ventilation associated with each indoor air quality (IAQ) measurement, and estimation of the PAH indoor emission rates. Identification of the characteristics of homes such as occupancy, the air exchange rate, garage attachment, presence of indoor sources such as smoking and cooking, type of furnishings, operation of heating and cooling systems, humidity, and mothball usage, which may affect the levels of PAHs in urban residential homes.

METHODOLOGY HOME RECRUITMENT AND SELECTION Searching for project participants started with the distribution of advertising flyers. Town meetings were held at different neighborhoods to explain the objectives, and participants’ responsibilities. procedures, Requirements as a volunteer participant included (1) being 18 years of age or older, homeowner or the head of household; (2) living in a single family house in the

6

4. 5.

House characteristics (age, size, layout, recent remodeling) Occupancy and general activities (number of people, occupations, ages, pets) Ventilation and appliances (heating and cooling systems, laundry location and venting, stove, fireplace, and heaters) Routine chemical storage and usage Miscellaneous

Copies of the study protocol and participant consent form were submitted to and approved by the appropriate University of Illinois at Chicago Institutional Review Board (IRB). Consent forms were signed by individual participants and the Principal Investigator before sampling began. SAMPLING Air sampling was conducted from June 2000 to August 2001. Indoor and outdoor air samples were simultaneously collected once per month at each of the participating homes. Sampling in HOME 09 did not start until November 2000. Only four homes were sampled in December 2000, because of heavy snowstorms and the holidays. AIRCON-2 high volume air samplers (Gilian Instrument Corp., Caldwell, NJ) were used. A Model AirChek HV30 pump made by SKC (Eighty Four, PA), which produces little noise, was used for indoor sampling after December, 2000. A rotameter was used for field calibration of the pumps. The rotameter was calibrated using a bubble airflow meter in the laboratory.


An Li

Sampling time was 46 hours and generally ran from morning to morning. The airflow rate was 10 Lpm, resulting in a total air volume of around 28 m3. The air was drawn by the pump through a filtercartridge assembly. The pump was connected to the assembly through Tygon tubing. The filter had a diameter of 3.2 mm and was made of quartz (Supelco, Bellefonte, PA). Two types of sampling cartridges were tested and compared in a preliminary study. The modified Supelco ORBO-1000 cartridge (Supelco, Bellefonte, PA) was chosen because of the difficulties encountered with assembling and re-using the other one, which was a noncommercial glass cartridge originally designed by the Research Triangle Institute (Research Triangle Park, NC). The modified ORBO-1000 cartridge consists of a 2.2 cm outside diameter by 7.6 cm long glass tube containing 3.0 cm of polyurethane foam (PUF), about 2.0 g XAD2 resin, and another 3.0 cm PUF. The entire assembly has a length of 15 cm including a 6.0 mm outside diameter “tail” for tubing connection. The PUF/XAD2/PUF is the method of choice for ambient air sampling of PAHs, as described in EPA Compendium Method TO-13A (USEPA, 1999). It has also been evaluated for sampling PAHs in indoor air and proved to have higher collection efficiencies than the PUF-only cartridges for light molecular mass PAHs (Chuang et al., 1987). Exactly 2.0 µL of a field surrogate solution was added to the cartridge using a syringe before the pump started. It was determined during preliminary sampling tests that no lag time should be allowed between exposure of the spiked cartridge to the air and starting the pump, to minimize the loss of spiked field surrogates. The cartridge was spiked on site, and the sampling pumps were initiated directly thereafter. The filter-cartridge assembly was held at a height of five feet, which represents the human adult breathing zone. The inlet of the cartridge was kept away from objects such as walls and furniture, which might block or interfere with the airflow. Indoor air samples were collected in or close to the kitchen. This sampling location was selected because, in most homes in the Chicago area, (1) the kitchen is much closer than other rooms to the potential PAH sources, including stoves, and is often close to the family room where fireplace would be and to the garage, (2) the kitchen is often located at the center of the house, and (3) the kitchen is often the place where most family activities occur. In addition, since sampling lasted for more than 46 continuous hours, sampling in bedrooms would bother the residents at night due to pump noise. Outdoor air samples were simultaneously collected at


locations no farther than 10 meters from the house. A Household Status Questionnaire (Appendix B) was used to establish the condition of the home and the surrounding activities at the time of sampling. Specific survey variables include use of heating, cooling, air cleaning or humidification, frequency and duration of window opening, occupancy, smoker guest visit, food cooked, fireplace usage, use of pesticides, household chemical usage and storage, indoor cleaning activities, noticeable odors, and neighborhood activities. During each sampling, real-time continuous measurement of CO2, temperature, and relative humidity with Models 8550 and 8551 Q-Trak air quality monitors (TSI, St Paul, MN) were carried out both indoors and outdoors. Concentration of carbon monoxide was also recorded indoors. After sampling, the records were uploaded to a computer and retrieved for data analysis. A Sampling Data Collection Sheet was also used to log in the starting and ending air quality and meteorological conditions. It was found that Q-Trak, designed for indoor air quality monitoring, did not work well outdoors. It recorded CO2 correctly, but generated unreliable values of temperature and relative humidity on direct exposure to outdoor meteorological conditions. Therefore, the data included in this report for outdoor temperature and humidity were collected from the O’Hare Airport Weather Service of Chicago. The outdoor CO2 data were not used in any PAH-related data analysis. The Q-Traks were factory calibrated at the onset of the study. During the project, they were calibrated every two months according to the operational manual (TSI, St Paul, MN). The Q-Traks were zeroed using TSI-supplied zero air. They were spanned for CO using 50 ppmv CO span gas and for CO2 using 1000 ppm CO2 span gas. Span gas was supplied to the Q-Trak using a detector cover that was flooded with zero air and span gas. CHEMICAL ANALYSIS Chemicals Sixteen PAHs were selected based on the U.S. EPA's priority pollutants list. The abbreviations, structures, and basic properties of these PAHs are given in Table 2. Solutions of PAH mixture were purchased from Ultra Scientific (North Kingstown, RI). Internal standard solutions containing naphthalene-d8, acenaphthene-d10, phenanthrene-d10, chrysene-d12, and perylene-d12 were also purchased from Ultra Scientific. Deuterated PAHs used for field surrogates (biphenyl-d10,

7


Table 2. in this studyStudy Table 2. PAHs PAHsinvestigated Investigated in This Name

Formula

MW

CAS #

Naphthalene

NaP

C10H8

128

91-20-3

Acenaphthylene

AcNP

C12H10

152

208-96-8

AcN

C12H10

154

83-32-9

Fl

C13H10

165

86-73-7

Phenanthrene

PhA

C14H14

178

85-01-8

Anthracene

An

C14H10

178

120-12-7

Fluoranthene

FlA

C16H10

202

206-44-0

Pyrene

Py

C16H10

202

129-00-0

Benz(a)anthracene

BaA

C18H12

228

56-55-3

Chrysene

Chy

C18H12

228

218-01-9

Benzo(b)fluoranthene

BbFlA

C20H12

252

205-99-2

Benzo(k)fluoranthene

BkFlA

C20H12

252

207-08-9

BaP

C20H12

252

50-32-8

Indeno(1,2,3-cd)pyrene

IP

C20H12

252

193-39-5

Dibenz(ah)anthracene

dBahA

C22H14

278

53-70-3

Benzo(ghi)perylene

BghiP

C22H12

276

191-24-2

Acenaphthene Fluorene

Benzo(a)pyrene

Abbr’n

Structure

Sample Pretreatment

fluoranthene-d10, and benzo[a]pyrene-d12) and analytical surrogates (fluorene-d10, pyrene-d10, and benzo[g,h,i]perylene-d12) were purchased from Cambridge Isotopes (Boston, MA). Solvents used in this study were purchased from Fisher Scientific (Fair Lawn, NJ). Methylene chloride, hexane, and acetone were GC grade. Cyclohexane and pentane were HPLC (high pressure liquid chromatography) grade. Silica gel (100200 mesh, Davisil Grade 644) and anhydrous sodium sulfate were also purchased from Fisher Scientific.

8

The procedures of sample pretreatment and instrumental analysis were developed based on EPA Compendium Method TO-13A (USEPA, 1999). The quartz filter and the cartridge were analyzed together, because it was not the objective of this study to investigate the PAH partitioning between gas phase and airborne particles. Before extraction, exactly 2.0 µL of an analytical surrogate solution was added to the cartridge. The cartridge was then refluxed on Soxhlet extractors using 10% ethyl ether in hexane for 24 hours, followed by another 24-hour extraction using methylene chloride. The extracts were then combined. The volume of the extract was reduced in a Kuderna-Danish (K-D) concentrator (500 mL flask, 3-ball Snyder column, 10 mL concentrator tube) to about five mL. Subsequently, a gentle stream of nitrogen was used to bring down the volume of the extract to about two mL. Optimization of the sample cleanup procedure was carried out in the laboratory. Briefly, the cleanup was accomplished by passing the sample through a chromatographic column (11 by 300 mm). The column was filled with hexane to about 1/3, then charged with a slurry of 10 g of activated silica-gel in hexane. About 10 g of anhydrous sodium sulfate were packed on the top. Samples were transferred onto the column in about two mL hexane and eluted sequentially with 40 mL hexane, 40 mL 10% ethyl ether in hexane, and 40 mL of methylene chloride. After the sample was charged onto the column, 15 mL of the initial eluate were discarded. The next 120 mL were collected in a flask, and the volume was reduced by K-D concentrator to about four mL. It was further reduced by nitrogen to two mL. The sample was then ready for instrumental analysis. GC/MS Analysis An Agilent (Hewlett-Packard) Model 6890+ GC coupled with Model 5973 MS was used for quantitative analysis. The GC/MS was equipped with a programmable temperature vaporizing (PTV) inlet and a HP MS-5 column (30 m by 250 µm id, film thickness 0.25 µm). An injection volume of 60 µL was used for all the samples. Helium was the carrier gas with an initial pressure of 10 psi for 10 minutes, which was then ramped to 18 psi at a rate of 0.2 psi/min. After the initial holding time of five minutes, the oven temperature was increased from 5˚C to 180˚C at a rate of 10˚C/min, then to


An Li

230˚C at 6˚C/min, then to 300˚C at 3˚C/min, and maintained at 300˚C for five minutes. The GC-MS interface temperature was kept at 290˚C. The electron impact (EI) MS was tuned daily with decafluorotriphenyl phosphine (DFTPP) and operated in either the selected ion monitoring (SIM) mode or the scan mode. The scan range was 50 to 500 m/z. The ion source temperature was 230˚C, and the quadrapole temperature was 150˚C. Data acquisition was controlled by a HP Kayak ChemStation. The sixteen PAHs were divided into five groups, each of which had an internal standard and at least one surrogate (see Table 3). PAHs were identified by matching the retention time and by comparing the MS spectra of the analytes with those in the NIST spectra library. Quantification was based on the sum of the primary and two secondary ions. Concentrations of PAHs were calculated based on the relative response factor (RRF), which was obtained for each PAH using a series of five calibration standards against the internal standard Table3.3.Grouping Grouping of PAHs, surrogates, and internal standards Table of PAHs, surrogates, and internal standards Name

Abbreviation

RT** (minutes)

PI** (m/z)

Group 1 IS* naphthalene-d8 naphthalene FS* biphenyl-d10

Nap-d8 Nap BP-d10

13.74 13.78 16.67

136 128 164

68 129

137 127

Group 2 acenaphthylene IS* acenaphthene-d10 acenaphthene AS* fluorene-d10 fluorene

AcNP AcN-d10 AcN Fl-d10 Fl

17.73 18.13 18.21 19.42 19.49

152 164 154 176 166

151 162 153 88 165

153 165 152 177 167

Group 3 IS* phenanthrene-d10 phenanthrene anthracene FS* fluoranthene-d10 fluoranthene

PhA-d10 PhA An FlA-d10 FlA

21.84 21.9 22.01 25.12 25.16

188 178 178 212 202

94 179 179 106 101

189 176 176 213 203

Group 4 AS* pyrene-d10 pyrene benz[a]anthracene IS* chrysene-d12 chrysene

Py-d10 Py BaA Chy-d12 Chy

25.73 25.78 29.38 29.41 29.5

212 202 228 240 228

106 101 229 120 226

213 203 226 241 229

Group 5 benzo[b]fluoranthene benzo[k]fluoranthene FS* benzo[a]pyrene-d12 benzo[a]pyrene IS* perylene-d12 indeno[123-cd]pyrene dibenz[ah]anthracene AS* benzo[ghi]perylene-d12 benzo[ghi]perylene

BbFlA BkFlA BaP-d12 BaP Pry-d12 IP dBahA BghiP-d12 BghiP

32.5 32.57 33.36 33.43 33.62 37.54 37.71 38.54 38.67

252 252 264 252 264 276 278 288 276

253 253 132 253 260 138 139

126 126 265 126 265 277 279

138

277

2nd Ion 3rd Ion (m/z) (m/z)

* IS = Internal Standard, FS = Field Surrogate, AS = Analytical Surrogate ** RT = Retention Time, PI = Primary Ion


used for that group. Each sample was analyzed twice using full scan and selected ion monitoring (SIM) modes. Large Volume Injection PTV-GCMS As with all environmental trace analysis, ensuring needed detection limits is the key to success. This had been a serious concern because of the low concentrations of many PAHs in air and the limited air volume to be collected. The analytical procedure used in this study was developed based on the EPA Compendium Method TO-13A (USEPA, 1999), which was designed for the determination of PAHs in ambient air. It requires collecting 300 m3 of ambient air over 24 hours at 225 Lpm. In this study, only 28 m3 of air was collected as stated in the Sampling section. It was not practical to increase either the sampling time or flow rate due to sampling logistics and unacceptable noise levels in the homes. At this low volume of air, the concentrations of individual PAHs must be greater than 7.0 ng/m3 to be quantitatively determined using the EPA method, in which the lowest concentration of the calibration solutions is 0.1 ng/µL. Most PAHs with 4+ rings, however, have indoor air concentrations below 1.0 ng/m3. Concern over the probable occurrence of many non-detectable concentrations led to the purchase and eventual use of the programmable temperature vaporizing (PTV) inlet, although it was not proposed in the original proposal. The PTV inlet made by Gerstel (Germany) was purchased with the support from the University of Illinois at Chicago School of Public Health. It allows introduction of up to 100 µL (compared with one or two µL using splitless injection) of samples into the GC and thus more analytes into the column and the detector. During the first six months, the operational condition of the GC/MS using the PTV large volume injection technique was optimized, and the entire analytical method was validated using the PTV technique. The results were published in the Journal of Air and Waste Management (Norlock et al., 2002). QA/QC Quality control procedures were incorporated into the entire process as an essential part of this study. Field blank and laboratory blank samples were analyzed to determine potential contamination. Instrument solvent blanks were run to check the status of the analytical system. The instrument detection limits (IDLs) and

9


method detection limits (MDLs) were determined for each of the 16 PAHs. Details of the procedure and the results can be found in Norlock et al. (2002). The IDLs and MDLs for air samples are given in Appendix C. The detection limits are far below the levels commonly found in the field. In a preliminary study, the recoveries of all 22 analytes (16 PAHs and six surrogates) were obtained for each of the sample preparation steps including extraction, concentration, and cleanup. The overall performance was good, with average recoveries in the range of 70 to 126%, and relative standard deviation for replicates ranged from 2 to 25%. Accuracy and bias were examined by the recoveries of spiked surrogates. Three deuterated aromatic hydrocarbons (biphenyl-d10, fluoranthene-d10, and benzo[a]pyrene-d12) were used as field surrogates (FS, see Table 3) to monitor the performance of the sampling and chemical analysis. Three deuterated PAHs (fluorened10, pyrene-d10, benzo[ghi]perylene-d12) were used as analytical surrogates (AS, see Table 3). Their recoveries were used to monitor the performance of laboratory analysis including extraction, concentration, cleanup, and instrumental analysis. The EPA TO-13A method requires the use of two field surrogates (fluoranthene-d10 and benzo[a]pyrene-d12) and two analytical surrogates (fluorene-d10 and pyrened10). Biphenyl-d10 was added as an additional field surrogate. Biphenyl has a comparable boiling point temperature and shorter retention time than naphthalene, and its recovery may more realistically reflect the loss of naphthalene than the other two field surrogates. Benzo[ghi]perylene-d12 was added as an additional analytical surrogate to better monitor the behavior of heavier PAHs during sample pretreatment and instrumentation. The recoveries of the surrogates are summarized in Table 4. Surrogate recoveries for individual samples are compiled in Appendix E. Excluding those of fluorened10, seven out of 1,087 recovery data were higher than 500% and excluded in obtaining the statistical summary in Table 4. Among the field surrogates, 80% of FlA-d10 recoveries and 71% of BaP-d12 recoveries fell into the acceptable range of 40 to 140%. Biphenyl-d10 may not be a good choice as an additional surrogate, since its recoveries tended to be high, probably due to interferences. For analytical surrogates, 83% of the pyrene-d10 recoveries were within 40% to 140%. The major problem with surrogates occurred with fluorene-d10 for indoor air samples collected after March 2001. A large peak co-eluted with fluorine-d10, making it

10

Table 4. 4.Summary Summarystatistics Statistics of Surrogate Recovery Table of surrogate recovery (%) Field Surrogates Measure

BP-d10

N

112

FlA-d10

Analytical Surrogates

BaP-d12

Fl-d10

Py-d10

BghiP-d12

88

113

104

Indoor Median

113

108

89.98

94.59

66.92

78.90

87.01

121.17

Mean

119.27

105.70

74.90

116.87

75.55

117.22

Stdev

106.57

70.62

60.23

110.49

37.92

53.36

Max

449.60

61.04

484.32

482.68

167.10

296.51

Min

2.75

1.25

0.03

0.00

Excluded*

1

0

1

N

108

110

Median

134.32

89.05

57.42

153.75

95.52

121.13

Mean

142.28

98.03

63.33

172.88

91.53

117.52

Stdev

88.40

46.30

37.25

115.36

28.68

47.72

Max

405.23

408.29

199.20

488.86

169.38

230.16

Min

7.58

1.63

0.00

9.96

1.63

4.34

Excluded*

3

1

0

5

0

0

0.79

0.00

24

0

1

107

111

101

Outdoor 107

* Number of recovery data which are > 500% and excluded from statistical analyses. These excluded recoveries are marked as “large” in Appendix E.

impossible to integrate even with the character ions extracted. The identity of this interfering substance was not determined. The primary ion of fluorene-d10 (m/z = 176) occurs in the MS spectra of a number of native PAHs which are likely to exist in air samples. A total of 24 out of 40 indoor samples collected after March 2001 are excluded in calculating the average and median fluorene-d10 recoveries for indoor samples. Only three out of 37 outdoor samples collected concurrently with these indoor samples need to be excluded due to the same problem. The recovery data were used to monitor the overall performance. Due to the probable interference with the surrogate compounds in GC/MS analysis, the data were not used to “correct” the measured PAH concentrations. DATA MANAGEMENT AND STATISTICAL METHODS Each sampling event generated a set of data collection documents including: 1. A completed Sampling Data Collection Sheet 2. Q-Trak 48-hour recordings of temperature, humidity, carbon dioxide, and carbon monoxide for indoor air 3. Q-Trak 48-hour recordings of temperature, humidity, and carbon dioxide for outdoor air


An Li

4. A completed Chain of Custody Form for the indoor sample 5. A completed Chain of Custody Form for the outdoorsample 6. PAH Analytical Results Summary for the indoor sample (MS Excel printout) 7. PAH Analytical Results Summary for the outdoor sample (MS Excel printout) 8. GC/MS chromatogram for the indoor sample 9. GC/MS chromatogram for the outdoor sample 10. A completed Home Status Survey 11. A completed Activity Log Sheet Both MS Excel version 2000 and the SAS System for Windows Release 8.1 were used for statistical data analyses. The results presented in this report are in most cases descriptive rather than interpretative. The PAH data were grouped to compare the indoor and outdoor data, among different homes, seasons, days of sampling, and surrounding industrial levels. For each group, the means, standard deviations, percentiles, t-statistics, and normality test parameters were obtained using the SAS Univariate Procedure for both the original and logtransformed data. Boxplots are also presented, using either normal or log scales for most data groups. The repeated-measures general linear model (GLM) was applied to both the raw data and log-transformed data to compare different levels of a class variable. The sampling design emphasizes repeated measurement at each of the ten selected homes; this makes it necessary to analyze the obtained data with repeated measure statistical techniques. GLM was used instead of ANOVA because the sample sizes of the groups involved were often different. Variations among homes and seasons were examined. Both Duncan’s multiple range test and the Scheff’s multiple comparison option were used with the repeated measures GLM to identify the pairs of groups that were significantly different at the 95% level. In cases where the distributions of the data were uncertain and the sample sizes were not large, one-way nonparametric analysis (NPAR1WAY procedures of SAS) was also performed, and the Wilcoxon rank sums scores and the χ2 were obtained. Difference in PAHs collected during different days of the week was also investigated using GLM. Regression analysis was performed to obtain emission rates and examine the dependence of PAH concentrations on house characteristics variables.

RESULTS AND DISCUSSION DATA SUMMARY Table 5 summarizes the number of samples collected and raw data generated. The data collected during the preliminary studies for the purpose of method validation are not included. The data for quality control purposes (field and laboratory blanks, surrogates, etc.) are not included. A list of individual samples by sampling date is given in Table F-1 of Appendix F. Among a total of 260 numbered samples, 13 were dust samples, and seven were field blanks or analytical blanks. Each sample was assigned an ID for easy tracking. The plan was to sample each home once a month for 12 months. However, sampling at Home 09 began two months later than at other homes. Sample collection activities were also inevitably affected by the severe snowstorms and holidays in December 2000. Therefore, the sampling period was extended to make up the missed samples. During the thirteen-month sampling period, nine Table 5. Data Data summary Summary Table 5. Indoor Air

Description

Outdoor Indoor Air Dust Total

Number of Samples Proposed samples to be collected

120

120

40

280

Attempted samplings

120

120

16

256

Failed samplings

1

9

6

16

Failed lab analyses

4

3

0

7

Successful samples

115

108

10

233

PAH Concentration Data 16

16

16

16

Number of data proposed to be generated

1,840

1,728

160

3,728

Number of data generated

1,819

1,708

160

3,687

111

79

32

222

1,708

1,629

128

3,465

Number of PAH compounds

Number of data below detection Number of non-zero data reported

Air Quality Data* 4

3

-

7

Attempted samplings

120

120

-

240

Records generated

117

115

-

232


468

345

-

813

Number of air quality parameters**

Survey Data Household Screening

Sampling Status

Number of questionnaires

10

Number of questions

33

14

330

1638


117

* Means over the sampling period, based on logged data at 1-minute intervals. ** Indoor air quality parameters: temperature, humidity, CO, and CO2 ; Outdoor air quality parameters: temperature, humidity, and CO 2


11


outdoor (IDs 011, 049, 051, 076, 094, 100, 106, 116, and 134) and one indoor (ID 056) air samples were not collected properly due to power outrage and failure, broken cartridges, pump failure, and/or other reasons. Inappropriate sample handling in the laboratory resulted in failed chemical analyses for seven samples (IDs 073, 095, 098, 136, 195, 255, 256). Only 10 dust samples were successfully collected because of the improper sampling method used initially. To summarize, a total of 233 samples were successfully collected and analyzed, including 108 outdoor and 115 indoor air samples, as well as 10 dust samples. A total of 3,687 concentration values of individual PAHs were generated, including those below detection limits. These data are presented in Table F-2 (air samples) and Table F-3 (dust samples) in Appendix F. For some samples (ID 007 through 048), anthracene was not reported. Instead, the sum of phenanthrene and anthracene was given for these samples, because the GC separation of these two compounds was not successful during the early stage of the chemical analysis. This problem was solved by modifying the GC oven temperature program, and the concentrations of anthracene were obtained starting from sample ID 050. The total numbers of PAH data in indoor air, outdoor air, and dust samples are 1,819, 1,708, and 160, respectively. Of these numbers, 111 (6.1%), 79 (4.6%), and 32 (20%) were reported as “not detected.” The amount of air quality data collected by Q-Trak is large, because data were logged at one-minute intervals for each of the parameters during each of the 120 46-hour sampling events. The averages are displayed in Table 5, and a complete list appears in Appendix F, Table F-4. Two types of questionnaire surveys were conducted. The initial survey was used for home screening and contained 33 questions. The survey results are organized in Appendix D using tables and figures. There were 13 questions on the Home Status Survey Questionnaire collected during each sampling event. The responses are displayed in Appendix G.

hundred years. Types of the houses included traditional two-story, ranch, and bungalows. Six homes reported having gas forced-air heating systems, while Homes 04, 07, 08, and 10 reported using radiator heating. Only two homes (02 and 09) had wood-burning fireplaces. Seven homes used gas stoves or ovens, and Homes 02, 05, and 07 reported having electric stoves. Homes 02 through 10 had washer and dryer units in the home that were vented outdoors. All homes had garages except 05, 06, and 10, and only Home 02 had an attached garage. The ethnic characteristics of the participants include white, black, and Hispanic. Occupants less than 18 years old were counted as children in the house screening survey. No

HOME CHARACTERISTICS Locations of the participating homes are shown in Figure 1. Six homes were located within Chicago city limits, three in near suburbs, and one in Gary, Indiana. Characteristics of the houses and information about the residents at selected homes are summarized in Tables D1 and D-2, respectively, of Appendix D. The age of the houses ranged from less than ten years to over one

12

Figure 1. Location of participating homes in the Chicago area. The light gray area is the city of Chicago.


An Li


Temperature (°C)

0 Indoor

7/28

5/28

4/28

3/28

2/28

1/28

12/28

11/28

10/28

9/28

8/28

7/28

6/28

6/28

Outdoor

-20 Sampling Date

Figure 2. Mean indoor and outdoor temperatures during sampling 1800

Indoor

1600

Outdoor

1400 1200 1000

800 600 400 200

7/28

6/28

5/28

4/28

3/28

2/28

1/28

12/28

11/28

10/28

9/28

8/28

0

7/28

Seasonal variations of indoor and outdoor temperatures are shown in Figure 2. The indoor temperature appeared to slightly exceed average outdoor temperature during summer months. Indoor temperatures stabilized at around 20˚C as outdoor temperature fell in the winter. Due to the uncertainty in the collected outdoor Q-Trak data, as stated in the Methodology section, outdoor temperature and wind speed data for each sampling were obtained from the O’Hare Airport Weather Service in Chicago. These data were used in a simplified statistical model to estimate home ventilation rates for each sampling event. The CO2 concentrations in both indoor and outdoor air are shown in Figure 3. Outdoor CO2 concentration was relatively stable around 400 ppmv. Although reported here, because outdoor CO2 was measured using the Q-Trak, which was designed for indoor use, the data were not used further in data analysis and should be interpreted with caution. The indoor CO2 level was much more random and appeared to be primarily affected by indoor sources. The indoor CO profile is given in Figure 4. The indoor CO concentrations were below 4.0 ppmv, and were below

20

6/28

AIR QUALITY MONITORING

40

CO2 Concentration (ppmv)

residents reported having occupations that have been considered to be related to PAH sources or exposure. Table D-3 in Appendix D lists home chemical use and other activities that are considered to be related to indoor PAH generation. The entries in Table D-3 were from the responses to the Household Screening Survey Questionnaire (Appendix A). Therefore, they represent only the general storage and use of chemicals in each home and do not necessarily relate to the chemical usage during each sampling. All homes reported having chemicals stored in the home. The most commonly stored chemicals were nail polish and nail polish remover (10/10), paint (8/10), and drain cleaner (7/10). In addition, during the year-long study period, Homes 01 and 04 reported frequent use of candles and incense. Homes 02 and 03 reported the use of solvents within the last six months. Home 09 reported the regular use of dry cleaning services. Only Home 03 reported the storage of mothballs, which are known to be an indoor PAH source, specifically naphthalene. Table D-4 summarizes the potential indoor and outdoor sources of PAHs at each of the ten participating homes. The suspected outdoor sources, such as industries, airports, and roadways were all within a 2mile radius of each home.

Sampling Date

Figure 3. Mean indoor and outdoor CO2 during sampling

levels of concern for human exposure. Both CO and CO2 are products of incomplete combustion and therefore can be positively correlated with PAH emissions. Indoor CO2 generation and decay cycles were used to estimate the home ventilation rate during each sampling event. Indoor relative humidity, as shown in Figure 5, followed a pattern similar to outdoor temperature by decreasing in the winter and peaking in the summer around July and August.

13


CO Concentration (ppmv)

5

Indoor Only

4 3 2 1

7/28

6/28

5/28

4/28

3/28

2/28

1/28

12/28

11/28

10/28

9/28

8/28

7/28

6/28

0 Sampling Date

Figure 4. Mean indoor CO during sampling. (Note: The CO recorder reported only integers.) 80

Relative Humidity (%)

70 60

natural-log transformed data are summarized in Table 6 for the total PAHs and the sum of 15 PAHs (with naphthalene excluded). The means of individual PAHs in indoor and outdoor air samples are summarized in Tables 7 and 8, respectively. For calculations of mean and median, nondetected values are conservatively treated as zeros. This resulted in slightly underestimated means but the correct estimate of the median. The ranges of most PAH concentrations measured in Chicago are in good agreement with those previously published from similar studies. The total concentration of the 16 PAHs in indoor air ranges from 13 to 2,454 ng/m3, with an average of 336 ng/m3 and a median of 208 ng/m3. The total outdoor concentration ranges from 13 to 1,865 ng/m3, with an average of 300 ng/m3 and a median of 213 ng/m3. With the exclusion of naphthalene, the total concentrations of the 15 PAHs have a median of 30.2 ng/m3 for the indoor air and 36.0 ng/m3 for the outdoor air (Table 6). Statistical analysis found both indoor and outdoor PAH concentration data, either with or without naphthalene, were non-normally distributed based on the Shapiro-Wilk test for normality. The data Table 6. 6.Results Resultsofof SAS Univariate Procedure forPAHs Total and PAHs Table SAS univariate procedure for total 15 and PAHs 15 PAHs (ng m-3) 3) (ng/m

50 40 30

PAHs Measure

20

Indoor

N

115

Mean

Ln (PAHs) Outdoor

Indoor

Outdoor

108

115

108

336

300

5.40

5.35

Std Dev

340

288

0.94

0.88

Skewness

2.81

2.85

-0.20

-0.32

Kurtosis

12.65

11.13

0.08

0.59

Max

2454

1865

7.8

7.5

Figure 5. Mean indoor relative humidity during sampling

75% Q3

435

399

6.1

6.0

50% Median

208

213

5.3

5

25% Q1

112

123

4.7

4.8

DESCRIPTION OF PAH CONCENTRATIONS

Min

13

13

2.5

2.6

3.53

10

Total PAHs

7/28

6/28

5/28

4/28

3/28

3/28

2/28

1/28

12/28

11/28

10/28

9/28

8/28

7/28

6/28

0 Sampling Date

15 PAHs (Naphthalene Excluded)

Overview Throughout this report, the term “total PAHs” refers to the sum of the 16 PAHs studied. The data were not “corrected” based on surrogate recoveries. Such correction cannot be justified especially when the fluctuation of the surrogate recoveries was heavy and likely due to interference as discussed in the QA/QC section. Statistical analyses using the SAS Univariate procedure (SAS Institute, Cary, NC) for the original and

14

Mean

36.0

42.3

3.30

Std Dev

27.0

28.8

0.82

0.69

Skewness

1.61

1.89

-0.70

-0.45 0.32

Kurtosis

3.27

5.62

0.85

Max

147

180

4.99

5.20

75% Q3

46.9

54.4

3.85

4.00

50% Median

30.2

36.0

3.41

3.58

25% Q1

17.7

23.1

2.88

3.14

1.8

4.2

0.58

1.44

Min


An Li

AcN

115

Fl

115

PhA*

300

330

4.16

88

177

397

2340

2.32

4.47

ND

0.36

0.78

1.66

30.42

4.90

5.56

ND

0.57

3.62

6.86

33.31

6.50

8.26

ND

0.39

4.18

9.10

52.70

94

6.07

12.25

0.01

0.27

0.74

5.05

61.8

An*

93

11.34

13.14

ND

0.15

9.54

18.10

51.54

FlA

115

2.35

2.73

0.02

0.39

1.84

3.12

17.25

Py

115

1.28

1.31

0.00

0.21

1.05

1.66

8.38

BaA

115

0.103

0.184

ND

0.015

0.056

0.129

1.50

Chy

115

0.248

0.357

ND

0.081

0.164

0.257

2.13

BbFlA

115

0.128

0.172

ND

0.012

0.074

0.177

1.36

BkFlA

115

0.126

0.267

ND

0.012

0.060

0.139

2.31

BaP

115

0.151

0.243

ND

0.027

0.103

0.198

2.22

IP

115

0.122

0.176

ND

0.014

0.031

0.214

0.94

dBahA

115

0.102

0.114

ND

0.005

0.078

0.168

0.53

BghiP

115

0.242

0.452

ND

0.052

0.190

0.276

4.57

Total PAHs

115

336

340

13

114

208

433

2454

Q1(25%) Median Q3(75%) Max

10,000

Indoor (N = 115)

336

1857

AcNP

108

1.72

2.07

0.02

0.38

0.82

2.51

10.91

AcN

108

7.04

6.62

ND

2.87

5.66

8.84

44.14

Fl

108

12.53

10.33

N

5.23

9.38

17.65

49.43

PhA*

88

10.38

10.76

0.01

3.04

9.68

14.86

78.92

An*

88

1.25

2.47

ND

0.27

0.51

1.04

14.88

FlA

108

3.93

2.81

0.21

1.91

3.48

5.20

15.42

Py

108

2.56

2.17

0.13

1.32

2.22

3.14

15.97

BaA

108

0.262

0.479

ND

0.025

0.111

0.280

2.71

Chy

108

0.437

0.451

0.007

0.156

0.274

0.579

2.89

BbFlA

108

0.303

0.575

ND

0.015

0.133

0.385

5.32

BkFlA

108

0.181

0.223

ND

0.004

0.136

0.251

1.22

BaP

108

0.269

0.520

ND

0.073

0.176

0.254

4.93

IP

108

0.312

1.160

ND

0.017

0.101

0.322

11.98

dBahA

108

0.253

1.347

ND

0.020

0.079

0.171

14.04

BghiP

108

0.393

0.712

ND

0.086

0.254

0.488

6.94

Total PAHs 108

300

289

13

123

213

392

1865

* PhA and An concentrations in samples with ID less than 50 were excluded, due to poor separations of these two compounds at the early stage of the study. See Table F-2 in Appendix F.

0.01 0.001 0.00

Chy

168

Py

96

BaA

ND

FlA

274

0.1

An

257

Q1(25%) Median Q3(75%) Max

Fl

108

Min

1.0

PhA

NaP

SD

10.0

AcN

Mean

100

NaP

N

Median 25% to 75% percentiles 10% and 90% percentiles Minimum and maximum

AcNP

PAH

Figure 6. Boxplot of individual PAH concentrations in indoor air. The absence of a circle indicates the value is out of range of y-axis. 10,000

Indoor (N = 108) 1,000 PAH Concnetration, ng/m3

3) m-3) of PAHs (ng/m(ng Table 8. Outdoor Outdoorconcentrations Concentrations by PAHs

PAH Concentration, ng/m3

1,000

* PhA and An concentrations in samples with ID less than 50 were excluded, due to poor separations of these two compounds at the early stage of the study. See Table F-2 in Appendix F.

BghiP

115

Min

IP

115

AcNP

SD

dBahA

NaP

Mean

BaP

N

BkFlA

PAH

hypothesis of normality. The skewness and kurtosis were close to zero, and the boxplot, histogram, and probability plot all showed a normal distribution pattern. This means that both indoor and outdoor PAH data fit a lognormal distribution. Boxplots of indoor and outdoor total PAHs are shown in Figures 6 and 7, respectively. In drawing the boxplots, the minimum concentrations were set to 0.001 ng/m3 when the calculated values were less than 0.001 ng/m3. The variations in the measured PAH concentrations are large. Although the PAH concentrations in air fluctuate constantly and sometimes dramatically, a few analytical difficulties may affect the accuracy of reported PAH concentration data. GC separations were poor between phenanthrene (PhA) and anthracene (An) and between benzo[b]fluoranthene (BbFlA) and benzo[k]fluoranthene (BkFlA), even after intensive modification of the GC

BbFlA

3) Table of PAHs (ng/m(ng Table 7. 7. Indoor Indoorconcentrations Concentrations by PAHs m-3)


100 10.0 1.0 0.1 0.01 0.001


BghiP

dBahA

IP

BaP

BkFlA

BbFlA

Chy

BaA

Py

FlA

An

Fl

PhA

AcN

NaP

were skewed to the right, as evidenced by the high skewness and kurtosis. After log-transformation, both indoor and outdoor groups of data failed to reject the

AcNP

0.00

Figure 7. Boxplot of individual PAH concentrations in outdoor air. The absence of a circle indicates the value is out of range of y-axis.

15


Table home-specific correlation coefficients individual and total indoor PAHs = 10) Table 9. 9.Mean Average of Home Pearson’s Specific Pearsonís Correlationamong Coefficient Among Individual and (N Total Indoor PAHs (N = 10) NaP NaP

AcNP

AcN

Fl

PhA

An

FlA

Py

BaA

Chy

BbFlA

BkFlA

BaP

dBahA BghiP

Total

1.000

AcNP

-0.077

1.000

AcN

-0.062

0.204

1.000

Fl

0.179

0.204

0.649

1.000

PhA

0.176

0.032

0.289

0.339

An

0.041 -0.024

FlA

0.203

0.129

0.443

0.485

0.253

0.003

1.000

Py

0.265

0.204

0.365

0.582

0.264

-0.125

0.571

BaA

0.238 -0.277

0.191

0.321

0.150

-0.126

0.296

0.320

1.000

Chy

0.039 -0.005

-0.006

0.228

0.117

0.079

0.139

0.142

0.327

1.000

BbFlA

0.210 -0.072

0.157

0.370

0.066

-0.121

0.378

0.366

0.677

0.512

1.000

BkFlA

0.274

0.034

0.114

0.356

0.347

-0.020

0.190

0.217

0.289

0.396

0.363

1.000

BaP

0.037 -0.048

-0.043

0.200

0.129

0.111

0.145

0.136

0.512

0.633

0.583

0.609

1.000

IP

0.200 -0.054

0.125

0.230

0.069

-0.014

0.334

dBahA

0.035

0.096

0.035

0.086 -0.024

0.168

BghiP

0.140

0.064

0.051

0.350

0.104

Total PAHs

0.980 -0.005

0.036

0.268

0.179

1.000

-0.043 -0.088 -0.136

1.000 1.000

0.089

0.487

0.278

0.613

0.484

0.487

1.000

0.079 -0.042

0.130

0.131

0.327

0.410

0.335

0.640

1.000

-0.108

0.310

0.326

0.363

0.470

0.595

0.472

0.639

0.502

0.431

1.000

0.061

0.262

0.328

0.267

0.067

0.267

0.293

0.073

0.262

0.084

0.190

operational parameters. As a result, the reported concentrations of the two compounds in each of the two pairs were evaluated together. Inaccuracy may also have resulted from the overly wide range of concentrations for individual compounds. Because naphthalene concentration was commonly much higher than other PAHs in the same sample, systematic bias in naphthalene data might occur when the integrated GC peak area exceeded the linear range of the calibration curve. To overcome this problem, two sets of calibration curves were prepared with concentrations up to 0.1 µg/mL and 20 µg/mL, respectively. The ratio of the two slopes for naphthalene (2.45) served as a correction factor, which was applied to the naphthalene concentrations calculated automatically by the ChemStation from the integrated peak area. There is no evidence that the analytical results were affected by changes in personnel. Naphthalene has the highest mean and median concentrations among the 16 PAHs measured. This is in agreement with previously published results on PAHs, and indicates that naphthalene dominates the level of total PAHs in both indoor and outdoor environments. Mean home-specific Pearson’s correlation coefficients among individual and total PAHs are given in Tables 9 and 10 for indoor and outdoor air, respectively. The correlations between naphthalene and total PAH concentrations are

16

IP

1.000

strong, and the Pearson’s coefficient is greater than 0.99 for both indoor and outdoor data at all homes except for the indoor samples at Home 09. In Home 09, although naphthalene still counts for about 60% of total PAHs, the correlation between naphthalene and total PAHs is less significant (R = 0.83, P = 0.08). Instead, the compounds showing significant correlations with the total PAH level are pyrene and benzo[a]anthacene in Home 09. Compared with two major indoor PAH studies by Van Winkle and Scheff (2001) in south Chicago and by Chuang et al (1988) in Columbus, Ohio, indoor naphthalene concentrations found in this study are lower. It was found by Van Winkle and Scheff (2001) that indoor naphthalene emission was largely associated with mothball usage, which, in turn was found to be associated with variables of race (Hispanic) and location. In this study, only one home (Home 03) reported the use of mothballs, but the naphthalene at this home is not higher than most other homes. Following naphthalene, anthracene was found to be the second highest in indoor air, while fluoranthene and phenanthrene were the second highest in outdoor air with comparable means and medians. All PAHs with three or four rings and molecular mass 152 to 202 have average concentrations higher than 1.0 ng/m3, and their medians range from 0.5 to 19 ng/m3. By comparison, PAHs with four to six rings and molecular mass 228 or higher exist in air in


An Li

Table10. 10.Mean Average of Home Pearson’s Specific Pearsonís Correlationamong Coefficient Among Individual and Total Outdoor PAHs (N = 10) Table home-specific correlation coefficients individual and total outdoor PAHs (N = 10) NaP

AcNP

AcN

Fl

PhA

1.000

An

FlA

Py

BaA

Chy

BbFlA

BkFlA

BaP

IP

dBahA BghiP

NaP

1.000

AcNP

0.355

1.000

AcN

0.398

0.211

1.000

PhA

0.181

0.320

0.254

0.156

Fl

0.446

0.357

0.821

1.000

An

0.278

0.138

0.161

0.203

0.082

1.000

FlA

0.287

0.076

0.585

0.666

0.265

0.144

1.000

Py

0.440

0.362

0.556

0.585

0.416

0.300

0.711

1.000

BaA

0.153

0.420

0.088

0.091

0.315

0.133

0.040

0.152

1.000

Chy

0.438

0.672

0.174

0.402

0.271

0.296

0.324

0.495

0.521

1.000

BbFlA

0.120

0.370

0.010

0.032

0.321

0.076

0.156

0.279

0.762

0.490

1.000

BkFlA

0.165

0.181

0.017

0.143 -0.139

0.213

0.056

0.157

0.207

0.365

0.328

1.000

BaP

0.207

0.480

0.021

0.147

0.275

0.139

0.225

0.394

0.467

0.601

0.735

0.549

1.000

IP

0.010

0.154

-0.071 -0.098

0.088

0.124 -0.089

0.012

0.540

0.323

0.553

0.530

0.402

1.000

dBahA

0.056 -0.002

0.072

0.067

0.144

0.228

0.163

0.108

0.234

0.250

0.365

0.431

0.383

0.652

1.000

BghiP

0.131

0.479

0.096

0.159

0.338

0.221

0.206

0.417

0.623

0.659

0.671

0.396

0.639

0.693

0.595

1.000

Total PAHs

0.996

0.375

0.440

0.486

0.233

0.286

0.333

0.477

0.168

0.462

0.140

0.174

0.236

0.027

0.074

0.163

significantly lower amounts. No median or average for these heavy PAHs is above 0.25 ng/m3 for indoor samples and 0.5 ng/m3 for outdoor samples. Correlations among individual PAH pairs are investigated by Pearson’s correlation, and the Pearson’s correlation coefficients (R) averaged over the ten homes are given in Tables 9 and 10 for indoor and outdoor PAHs, respectively. Values of the Rs are highly home specific and mostly positive. It is also found that, in general, PAHs outdoors tend to be more correlated than those indoors, and the heavier PAHs are more likely to correlate with each other than the light PAHs. In the following sections, the PAH concentrations in indoor and outdoor air are separately examined for their seasonal as well as short-term variations and compared among different homes. A separate section is devoted to the comparison of indoor and outdoor concentrations and the indoor-to-outdoor ratio. Seasonal Variations To analyze the seasonal variations, the entire sampling period was divided into five seasons, which were Summer 1: June – August, 2000; Fall: September – November, 2000; Winter: December 2000 – February, 2001; Spring: March – May, 2001; and Summer 2: June – August, 2001. Tables 11


Total

1.000

and 12 summarize the basic statistics of the total PAH concentrations in the five seasons. Figures 8 and 9 graphically present the seasonal variations of total PAHs in both indoor and outdoor air. SAS procedures of GLM and NPAR1WAY were used to test the significance of the difference in indoor total PAH concentrations among seasons. Repeated measure GLM resulted in a F-value of 3.15 (P = 0.017) with the original data, and a F-value of 6.19 (P = 0.0002) with logtransformed data, indicating significant differences among the five seasons. Based on the results of Scheff’s multiple comparison and Duncan’s multiple range test, the difference between the means of fall and spring for indoor total PAHs is significant at the 95% confidence level. For outdoor total PAH concentrations, the results of both GLM and NPAR1WAY indicate that there is no statistically significant difference among the five seasons. For indoor PAHs, both mean and median PAHs of the fall season are the highest, and spring is the lowest among all seasons. The difference between the fall and spring medians is more than a factor of 4.0. High indoor PAH concentrations during the fall season may be explained by the fact that most homes closed windows and started the operation of heating appliances during this season. For PAHs with significant indoor sources, reduced air exchange due to activities such as window

17

Polycyclic Aromatic Hydrocarbons in the Air of Ten Chicago Area Homes Table 11. Statistical Analysis of Indoor Total PAHs by Season3 Table ) Statistical analysis of indoor total PAHs by season (ng/m ) (ng m-311. Measure

N

Summer 1

18

Mean

353

Fall

Winter

Original Data 25 500

Spring

Summer 2

24

28

20

332

186

330

Std Dev

276

335

481

220

260

Skewness

2.07

0.49

4.04

3.13

1.03

Kurtosis

5.23

-0.94

18.00

11.12

0.40

t-test

5.43

7.46

3.38

4.48

5.68

(Pr > | t |)

< 0.0001

< 0.0001

< 0.0001

< 0.0001

< 0.0001

100% (Max)

1223

1117

2454

1095

964

95%

1223

1090

697

642

868

75% (Q3)

406

701

342

193

469

50% (Median)

237

464

201

112

217

25% (Q1)

198

179

126

84

147

5%

87

91

41

27

50

0% (Min)

87

88

21

12

49

closing may result in dramatically and immediately enhanced indoor concentrations. The potential contributions from the start of heating appliances, especially furnaces with gas forced air (GFA), are discussed in “Factors Affecting PAH Concentrations” later in this report. Although PAHs were consistantly elevated in the winter, they did not reach statistical significance. Elevated PAH concentrations in outdoor air during winter may be explained by the increased house heating. In winter 2000, the Chicago area experienced sequential heavy snowstorms and cold weather throughout December and part of January. Natural gas consumption was high despite extremely high energy prices. A fireplace is a common way of providing additional home heating in Chicago, especially on weekends. The use of a fireplace might have been more frequent in winter 2000

Log-Transformed Data 5.93

5.31

4.82

5.46

Std Dev

0.68

0.83

0.98

0.90

0.90

Skewness

0.31

-0.51

-0.02

0.02

-0.31

Kurtosis

-0.05

-0.95

1.44

1.52

-0.82

t-test (Pr > | t |)

35.0

35.63

26.60

28.44

27.11

< 0.0001

< 0.0001

< 0.0001

< 0.0001

< 0.0001

Table 12. Statistical of outdoor total PAHs by season Table Statisticalanalysis Analysis of Outdoor Total PAHs by Season (ng m3-3)) (ng/m Measure

N

Summer 1

18

Mean

313

Fall

Winter

Original Data 21 282

Spring Summer 2

21

29

19

336

263

322

Std Dev

434

280

192

306

195

Skewness

3.05

2.52

0.87

3.12

0.70

Kurtosis

10.3

8.11

0.32

12.0

-0.67

t-test

3.05

4.62

8.01

4.63

7.20

(Pr > | t |)

< 0.0001

< 0.0001

< 0.0001

< 0.0001

< 0.0001

100% (Max)

1865

1288

804

1576

676

95% 75% (Q3)

1865

625

689

672

676

215

385

420

295

461

50% (Median)

179

205

320

159

253

25% (Q1)

116

105

187

99

168

5%

24

30

120

44

48

0% (Min)

24

13

113

36

48

Log-Transformed Data Mean

5.19

5.23

5.66

5.15

5.58

Std Dev

1.04

1.02

0.60

0.89

0.68

Skewness

0.21

-0.74

-0.11

0.39

0.47

Kurtosis

0.95

1.38

-1.01

-0.037

-0.6

t-test

21.1

23.5

43.5

31.0

0.62

< 0.0001

< 0.0001

< 0.0001

< 0.0001

< 0.0001

(Pr > | t |)

18

2500

Indoor Total PAHs in Indoor Air, ng/m3

5.37


2000

1500

1000

500

0 Summer 1

Fall

Winter

Spring

Summer 2

Figure 8. Boxplot of total PAHs in indoor air by season.

2000

Outdoor 1600 Total PAHs, ng/m3

Mean


1200

800

400

0 Summer 1

Fall

Winter

Spring

Summer 2

Figure 9. Boxplot of total PAHs in outdoor air by season.


An Li

than normal years, considering the abnormal price of natural gas. Added to domestic heating is workplace heating, because the heating systems at numerous businesses and industries are powered mainly by natural gas. By comparison, indoor PAHs during the winter season at our participating homes did not show an increase over the other seasons. Seasonal variation of urban atmospheric PAH concentration is a controversial issue, reflecting the complexity of PAH sources. In an urban environment, traffic can be the primary source of atmospheric PAHs, and it is non-seasonal. On the other hand, vaporization of PAHs from various solid surfaces, soil, water, and vegetation is highly seasonal because of the strong temperature dependence of vapor pressure. In residential indoor environments, cooking as a source is nonseasonal, although outdoor barbecuing occurs mostly in summer and fall. Other non-seasonal sources, including candle and incense burning, chemical storage, garage activities, etc., can be strongly home-specific. Seasonal combustion sources include fireplace wood burning and space heating using kerosene or natural gas. Both of these are winter events. In addition, indoor-outdoor air exchange can play an important role regardless of the sources. Room temperature is another important factor, because it is well known that PAHs emitted from combustion sources have a tendency to absorb onto various solid surfaces. They then become secondary emission sources with emissions being strongly temperature-dependent.

Table 13. PAH concentrations of consecutive indoor air samples for three 3 ) Concentrations of Consecutive Indoor Air Samples Table 13. PAH homes (ng/m for Three Homes (ng m-3)

PAH

HOME 05

HOME 02

HOME 09

PAH141 PAH143 02/10/01 02/12/01

PAH165 PAH167 03/17/01 03/19/01

PAH173 PAH175 03/26/01 03/28/01

Nap

362.5

178.6

51.4

86.7

33.5

81.3

AcNP

0.751

0.223

0.886

1.137

30.416

0.104

4.286