Personal exposure to airborne particulate matter due ...

6 downloads 4107 Views 675KB Size Report
Jan 11, 2016 - ments), (ii) Samsung Korea dryer (6 experiments), (iii) LG dryer .... (70e130 μg/m3), involving a much shorter activity period. (10e30 s), fall ...
Building and Environment 98 (2016) 145e149

Contents lists available at ScienceDirect

Building and Environment journal homepage: www.elsevier.com/locate/buildenv

Technical note

Personal exposure to airborne particulate matter due to residential dryer lint cleaning Kai-Chung Cheng a, *, Daisy Zheng a, Afua O. Tetteh b, Hye-Kyung Park b, c, Kari C. Nadeau b, Lynn M. Hildemann a a b c

Civil and Environmental Engineering Department, Stanford University, Stanford, CA 94305, USA Division of Immunology and Allergy, Stanford University School of Medicine, Stanford, CA 94305, USA Department of Internal Medicine, Pusan National University School of Medicine, Busan, Republic of Korea

a r t i c l e i n f o

a b s t r a c t

Article history: Received 27 November 2015 Received in revised form 25 December 2015 Accepted 8 January 2016 Available online 11 January 2016

Exposure to airborne particles during and after cleaning dryer lint was examined via 30 experiments involving 4 dryers in a laundry room of a Northern California home. Gravimetric and real-time air samplers measured mass and size-resolved number concentrations in close proximity to the cleaning activity. The size distributions varied greatly between loads of clothing, with particle diameters > 10 mm contributing the bulk of the airborne lint dust volume. Average 5-min exposures to PM10 varied from < 10 to > 300 mg/m3. Cumulative frequency distributions of 1-min-averaged PM10 measurements were used to characterize the probabilities of different short-term exposure levels during and at different elapsed times after lint cleaning. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Indoor personal exposure Particle size distribution Airborne lint dust Household cleaning Particle resuspension PM10

1. Introduction The average American spends 69% of the time in their homes [1], where people are frequently in close proximity to air pollution sources such as cooking, smoking, and household cleaning. Near an indoor emission source, pollutant levels are substantially higher than further awaye this “proximity effect” is why personal exposure levels (using a monitor worn by a person) are consistently higher than indoor background levels (from a stationary indoor monitor) [2e5]. For example, Ferro et al. [6,7] found that household vacuuming can cause 2.5e10 mm particle exposure levels to be ~2  as high as background levels in the same room. Acevedo-Bolton et al. [8] showed that fine particle (diameters  2.5 mm; PM2.5) exposures when sitting next to a smoker averaged ~4  as high as the background PM2.5 in two houses. Previous studies have examined a range of indoor activities (i.e., folding clothes/rugs [7], vacuuming [9,10], walking [11,12], wiping the blackboard [13], and even body movements during sleep [14]) that suspend particles indoors, leading to elevated exposures due to

* Corresponding author. E-mail address: [email protected] (K.-C. Cheng). http://dx.doi.org/10.1016/j.buildenv.2016.01.008 0360-1323/© 2016 Elsevier Ltd. All rights reserved.

the proximity effect. However, cleaning dryer lint has not been examined, even though it can quickly suspend substantial amounts of particles, in close proximity to the person performing the cleaning. A few past studies have chemically characterized bulk lint samples for trace metals [15,16], polybrominated diphenyl ethers (PBDEs) [17,18], or poly- and perfluorinated compounds (PFCs) [19]. These studies proposed that dryer lint could serve as an indicator of the amount of chemical present in the household [15,17], as a surrogate for dermal [18,19] or hand-to-mouth [18] exposure, or as a more general indicator of exposure in the home [16]. None of these studies investigated airborne particle concentrations or inhalation exposures when cleaning the lint trap. One clinical study [20], investigating sensitization to detergent enzymes, measured exposure during dryer lint cleaning in the lab, reporting airborne levels of 0.04e1.2 ng protein/m3. This is the only study, to our knowledge, that directly examined human exposure to airborne lint. However, this study did not examine levels of exposures in a real residential laundry room, nor did it characterize the mass concentration or size distribution of airborne dust produced by dryer lint cleaning. Our first goal is to investigate exposure to airborne particles

146

K.-C. Cheng et al. / Building and Environment 98 (2016) 145e149

when cleaning dryer lint in homes. We performed 30 experiments in a home laundry room, measuring particle mass concentrations gravimetrically and in real time, for 4 dryers. Our second goal is to examine the size distributions of airborne lint. A monitor logged number concentrations for 14 size ranges continuously in each experiment, for 0.3e20 mm particles. 2. Materials and methods We conducted experiments in a small (~12 m3) laundry room in a single-family house in Palo Alto, California. For each experiment, we used bare fingers to remove the bulk lint in the dryer lint trap produced from one load of clothing, while using gravimetric and real-time air samplers to measure exposure to airborne particles in close proximity. The gravimetric sampler consists of a 115 V AC vacuum pump connected to two aluminum filter holders in parallel: one with a PTFE membrane filter (47 mm diameter, 1-mm pore size, Pall Corp., Ann Arbor, MI, USA) and another with a PVDF membrane filter (47 mm diameter, 0.45-mm pore size, EMD Millipore Corp., Billerica, MA, USA). (The results of analyzing these two filters for trace metals and allergens, respectively, will not be discussed here.) Each filter was downstream of a cyclone separator (URG Corp., Chapel Hill, NC, USA) removing particles with aerodynamic diameters > 10 mm at 16.7 L/min. Typically, exposure studies have focused on fine particulate matter (PM2.5). Here, we chose to collect both fine and coarse respirable particles together (PM10), because in our initial tests, we were not able to accumulate sufficient PM2.5 mass. A real-time aerosol monitor (AM510 SidePak laser photometer, TSI, Shoreview, MN, USA) logged PM10 concentrations every 10 s. A size-resolved monitor (Portable Aerosol Spectrometer Model 1.100, Grimm Technologies, Inc., Douglasville, GA, USA) measured number concentrations of airborne lint particles for 14 size ranges (0.3e20 mm) every 1 min e this instrument is widely used for investigating size-specific airborne particle levels in residential settings [e.g., 21,22]. All the monitors were placed together on a laboratory hand truck (Fig. 1 (a)) - this allowed us to (i) collocate all air sampling inlets near the adult's breathing height while standing (1.5 m) during sampling and (ii) transport all the instruments in/out of the

laundry room without changing any air sampling settings between experiments, thereby measuring concentrations in a consistent manner. We carried out 30 lint cleaning experiments using 4 different dryers: (i) Samsung USA dryer (Model DV50F9A8EV; 12 experiments), (ii) Samsung Korea dryer (6 experiments), (iii) LG dryer (Model DLEX5680V; 6 experiments), and (iv) Whirlpool dryer (Model WED8900BC; 6 experiments). These dryers were purchased locally, except for the Samsung Korea dryer, a prototype with a newly designed lint trap which was provided directly from Samsung in Korea. This prototype dryer used an enclosed cylindrical container to collect bulk lint, different from the typical in-door filter screens used in the 3 commercially-available dryers. We conducted 3 lint cleaning experiments per day, with gravimetric and real-time air samplers running continuously. The airborne lint particles were collected ~0.5 m horizontally from the lint trap at ~1.3 m from the floor, to approximate the breathing location of a person performing the cleaning activity. In each experiment, we closed the laundry room door and turned off the HVAC system for the house. This gave an air change rate of ~0.1e0.6/h, estimated by the slopes of the log-linear regression lines between the measured number concentrations of the smallest size range (0.3e0.4 mm) and time during the decay periods. An investigator inside the room cleaned the lint trap for 10e30 s and then quickly exited the room after 10 min, reclosing the door. We waited ~1e2 h before the next experiment to minimize the contributions from previous cleaning activities on subsequent measurements. Each pair of PM10 filter samples represented 3 lint cleaning experiments. Dryer loads contained clothes and bedding from 3 local houses: one with 3 cats, another with 2 dogs, and the other without any pets. Each load of clothing was washed with Tide® Original liquid detergent using a top-load Whirlpool washer for 45 min (the standard washing time) and dried (with 6 dryer balls and 1 dryer sheet, to minimize static) for 50 min with the normal temperature setting. We collected and weighed the bulk lint after each drying cycle. The gravimetric PM10 concentration was calculated as the particle mass collected on each filter divided by the air volume sampled (air sampling flowrate  duration). For both filter samples,

Fig. 1. (a) Air sampling setup including the gravimetric filter sampler and two real-time monitors (SidePak and Grimm). Example time series plots for (b) PM10 and (c) size-resolved particle number concentrations measured by SidePak and Grimm monitors, respectively. Each time series shows 3 concentration increases, produced by 3 successive dryer lint cleaning experiments on one day.

K.-C. Cheng et al. / Building and Environment 98 (2016) 145e149

air flowrates were measured before and after sampling, using a primary flow calibrator (Gillian Instrument Corp., West Caldwell, NJ, USA). Filters were equilibrated for > 24 h at controlled relative humidity (~45%) and temperature (25  C) and then weighed before and after sampling, using a Mettler M3 Microbalance (MettlerToledo, Columbus, OH, USA). The results for the two filters were averaged to give gravimetric PM10 concentration for each experimental day. SidePak monitors evaluate the PM10 mass concentration from light scattering measurements, using a calibration factor to account for the variation in light scattering with the size and composition of particles. We determined the appropriate SidePak calibration factor for airborne lint by finding the ratio of gravimetric concentration over the time-averaged SidePak measurement for each day. Calibration factors varied greatly (0.41e1.21), perhaps due to different mixtures of fibers and dust contaminants in each load of clothes. We rescaled all 10-s SidePak measurements collected on each day, using the calibration factor determined for that day. Fig. 1 (b) and (c) plot example time series of PM10 and sizeresolved number concentrations measured by SidePak and Grimm monitors, respectively, each showing 3 concentration increases, produced by 3 cleaning experiments on one day. For all experiments, the background PM10 and total particle number concentrations were 4.9e11 mg/m3 and 5300e12000 #/L (respectively) estimated by the 5-min time-averaged levels immediately before each cleaning activity. 3. Results and discussion Table 1 summarizes 5-min average PM10 exposure statistics during and immediately after lint cleaning activities (including the 10e30 s cleaning period) for the 4 dryers. For each dryer, these initial 5-min exposures varied greatly, with a maximum-tominimum ratio (Max/Min) ranging from ~4 to 36. This is not surprising, as there are variations in the amount, particle sizes, and perhaps the moisture content (not measured) of the bulk lint collected, as well as the rigor of the lint cleaning activity. The mean initial 5-min PM10 exposures (70e130 mg/m3) were roughly comparable for the 3 commercially available dryers, which use in-door filter screens to trap lint. In comparison, for 15-min activities inside a residence (i.e., folding clothes/blankets, walking/dancing on rug, and vacuuming), Ferro et al. [7] found mean PM10 exposures ranging from 30 to 420 mg/m3. Our mean 5-min PM10 exposures (70e130 mg/m3), involving a much shorter activity period (10e30 s), fall within this range. The mean for the prototype dryer experiments (5.1 mg/m3) was notably lower. We hypothesize this was because the lint trap is an enclosed cylindrical container e the ability to empty the container directly into the trash greatly reduced the resuspension of lint. Since amounts of bulk lint differed substantially (0.44e5.6 g), we divided those 5-min exposures by the bulk lint mass. However, the normalized 5-min exposures remained highly variable (Max/ Min ¼ 4-55) for each dryer. Again, the prototype dryer had a much lower mean (3.1 mg/m3-g) than the other 3 dryers (39e63 mg/m3-g).

147

A linear regression analysis found that the positive correlation between the bulk lint mass and 5-min exposures for each dryer was not significant (p > 0.1, R2 < 0.2). This could be due in part to the variation in the bulk lint moisture content, which would increase the weights of bulk lint samples but reduce the particle emissions. To assess the typical PM10 exposure levels when cleaning dryer lint, we excluded the 6 prototype dryer experiments and focused on the 24 experiments involving the other 3 dryers (with a typical design of lint trap). Fig. 2(a) shows the minute-by-minute exposures to PM10 as boxewhisker plots, from the minute before to the 8th minute after lint cleaning started. For each minute, the mean was systematically higher than the median. The 1-min exposures during the first 2 min were notably higher and more variable. The means of 1-min exposures decreased with time from the 2nd minute on, dropping below 50 mg/m3 (a 1-min benchmark concentration we chose that matches the World Health Organization (WHO) 24-h ambient PM10 guideline) in the 4th minute, and becoming comparable to the original background levels in the 8th minute (Mean ¼ 6e7 mg/m3). The characteristic settling velocity for PM10 particles (air sampling height (~1.3 m) divided by concentration decay time (8 min)) of 9.8 m/h was close to the predicted setting velocity (10 m/h) for 8 mm particles (assuming density ¼ 1 g/ cm3). We did not estimate the PM10 emission factor for each lint cleaning activity because of the difficulty of reasonably performing log-linear regressions over the short decay periods (~8 min). Although the cleaning activity lasted only 10e30 s, the maximum mean (199 mg/m3), median (111 mg/m3), and interquartile range (252 mg/m3) for 1-min exposures occurred in the 2nd minute - this could be associated with the non-instantaneous dispersion which is typical for a naturally-ventilated room [23,24]. Fig. 2(b) shows frequency distributions for 1-min PM10 exposures (24 experiments involving 3 commercially available dryers) on a log probability plot. The 1-min exposures for each 2-min period are fairly straight lines e that is, close to lognormal, and consistent with having means that were systematically higher than the medians (Fig. 2(a)). During the first 2-min period, ~5% of 1-min exposures to PM10 were > 1000 mg/m3. As time progressed from the 1st to the 5th 2-min period, both the magnitudes (y-axis) and variations (slope) of 1-min exposures decreased systematically. Comparing our measurements with the 1-min benchmark concentration (50 mg/m3), ~45% of the 1-min exposures during the first 2-min period exceeded 50 mg/m3. In contrast, all of the 1-min exposures were below the benchmark value after 6 min. To examine the airborne lint size distributions, we utilized sizeresolved measurements for the largest concentration spike produced by each lint cleaning activity (the maximum 1-min total number concentration measured by the Grimm monitor). We excluded experiments without pronounced concentration peaks (i.e., those with the prototype dryer), leaving 21 size distributions for analysis. Summing the number concentration Ni (#/L) over different size ranges i, we determined the ratio of the total number concentration for particles less than or equal to a specific size cut, Dp (mm), NDp (#/L), to that for all measurable suspended particles (up to 20 mm),

Table 1 Summary statistics of 5-min exposures to PM10 immediately after lint cleaning activities started, from 30 experiments with 4 different dryers. Samsung USA Number of lint cleaning experiment 12 First 5-min PM10 exposurea (mg/m3) Mean [Min-Max] 70 [7.5e150] First 5-min PM10 exposurea/bulk lint mass (mg/m3-g) Mean [Min-Max] 53 [4.9e270] a

Samsung Korea

LG

Whirlpool

6

6

6

5.1 [2.5e9.6]

100 [6.8e290]

130 [9.0e320]

3.1 [1.3e5.2]

63 [6.0e190]

39 [4.4e86]

Gravimetric filter measurements were used to determine the appropriate calibration factor for SidePak data.

148

K.-C. Cheng et al. / Building and Environment 98 (2016) 145e149

Fig. 2. (a) 1-min exposures to PM10 for 24 experiments with 3 commercially available dryers from the minute before to the 8th minute after lint cleaning started. In these box-andwhisker plots, the shaded box represents the interquartile range (25th to 75th percentile), while the solid line inside marks the median. The “whiskers”, drawn as confidence bounds, demarcate the range from the 10th to the 90th percentile. (b) Frequency distributions of 1-min exposures to PM10 on a log probability plot, representing 24 dryer experiments with 3 commercially available dryers from the 1st to the 5th 2-min periods. The 50 mg/m3 line represents the 1-min benchmark PM10 concentration.

particles (up to 20 mm), Voltotal (mm3/L):

Ntotal (#/L):

Pn NDp N ¼ Pi14¼1 i Ntotal i ¼1 Ni

(1)

Here, n is the number of size ranges less than the size cut, Dp (mm). By normalizing NDp with Ntotal, we can compare particle size distributions for different magnitudes of concentration spikes. We calculated particle volume concentrations from the number concentrations measured for 14 different size ranges, assuming spherical particles:

Voli ¼ Ni

pD3i 6

! (2)

Here, Voli (mm3/L) is the volume of airborne particles per unit volume of air for size range i. Di (mm) is the arithmetic midpoint of the upper and lower diameters for size range i. Similarly, summing the volume concentration Voli (mm3/L) over different size ranges, we determined the ratio of the total volume concentration for particles less than or equal to a specific size cut, Dp (mm), VolDp (mm3/L), to that for all measurable suspended

Pn VolDp Voli ¼ Pi¼1 14 Voltotal i¼1 Voli

(3)

Fig. 3 (a) and (b) plot NDp/Ntotal and VolDp/Voltotal  100% against different particle size cuts, Dp (mm), respectively. Each line represents the cumulative particle size distribution for the observed maximum concentration spike. Each bold solid black line (with triangles) represents the average of the 21 cumulative size distributions. They show that > 80% of the total volume of airborne lint dust is from particles with diameters > 10 mm, which represent ~1% of the total particle count. In contrast, particles < 2.5 mm contribute ~96% of the total particle count but add < 3% to the total volume. This illustrates why it was difficult to gravimetrically collect sufficient PM2.5 airborne lint dust mass (see Methodology section). Assuming a uniform particle density across different size ranges, we estimate the mean 5-min exposures to PM2.5 by multiplying the mean 5-min PM10 exposures (70e130 mg/m3; Table 1) by Vol2.5/ Vol10 (2.5/18.3; Fig. 3(b)), giving 9.6e18 mg/m3 for the 3 commercially-available dryers. The size distributions varied significantly e for example, one

Fig. 3. Cumulative size distributions for airborne lint (a) number concentration (Ni) and (b) volume concentration (Voli) for maximum 1-min concentration spikes produced by lint cleaning activities in 21 experiments. NDp/Ntotal and VolDp/Voltotal are the ratios of the total number and volume concentrations (respectively) for particles less than or equal to a specific diameter, Dp (mm), over all measurable suspended particles (up to 20 mm).

K.-C. Cheng et al. / Building and Environment 98 (2016) 145e149

distribution had Vol10/Voltotal of 49% whereas another had only 1%. This great variability could be associated with different mixtures of clothing fibers (i.e., wool, polyester, and cotton) and attached particle contaminants (i.e., soil dust, cat/dog dander, and combustion particles) that have different size ranges. In addition to the variation in the bulk lint moisture content, the highly variable partitioning (Vol10/Voltotal) at 10 mm could explain why there were no strong correlations between the mass of bulk lint and levels of PM10 exposures (p > 0.1, R2 < 0.2). 4. Summary and implications This study presents the first effort to characterize personal exposures to airborne particles generated by cleaning dryer lint. We found that lint cleaning produces an airborne particle volume concentration that is primarily coarse (> 10 mm), giving it shorter residence times in the air than PM2.5. However, the rapid release of a sizable cloud of lint particles near the person performing the cleaning can still result in 1-min exposures to PM10 > 1000 mg/m3. In addition, the coarser airborne particles emitted close to the person may deposit in the throat and be ingested, causing potential additional health risks. For 3 commercially available dryers, the mean level of 1-min PM10 decreased below 50 mg/m3 3 min after cleaning. The cumulative frequency distributions of 1-min PM10 measurements showed that all 1-min exposures dropped below this benchmark concentration (50 mg/m3) 6 min after cleaning. These results offer guidance on an appropriate wait time before entering or reentering a residential laundry room after performing everyday lint cleaning activities. They also provide insight into how long an HVAC system in a laundry room should run to reduce airborne lint exposure effectively while conserving building energy. Acknowledgments This work was funded as part of the Healthy Home Research Program sponsored by Samsung Electronics, Ltd., South Korea. The authors thank Royal Kopperud of Stanford University for advice about the assembly of gravimetric air samplers and the use of cyclones. References [1] N.E. Klepeis, W.C. Nelson, W.R. Ott, J.P. Robinson, A.M. Tsang, P. Switzer, J.V. Behar, S.C. Hern, W.H. Engelmann, The national human activity pattern survey (NHAPS): a resource for assessing exposure to environmental pollutants, J. Expo. Anal. Environ. Epidemiol. 11 (2001) 231e252. [2] C.E. Rodes, R.M. Kamens, R.W. Wiener, The significance and characteristics of the personal activity cloud on exposure assessment measurements for indoor contaminants, Indoor Air 1 (1991) 123e145. [3] W.R. Ott, Human exposure assessment: the birth of a new science, J. Expo.

149

Anal. Environ. Epidemiol. 5 (1995) 449e472. [4] L. Wallace, Indoor particles: a review, J. Air Waste Manag. Assoc. 46 (1996) 98e126. [5] D. Rim, A. Novoselac, Transport of particulate and gaseous pollutants in the vicinity of a human body, Build. Environ. 44 (2009) 1840e1849. [6] A.R. Ferro, L.M. Hildemann, S.J. McBride, W.R. Ott, P. Switzer, Human exposure to particles due to indoor cleaning activities, in: C.A. Brebbia, M. Jacobson, H. Power (Eds.), Air Pollution VII, WIT Press, Southampton, UK, 1999, pp. 487e496. [7] A.R. Ferro, R.J. Kopperud, L.M. Hildemann, Elevated personal exposure to particulate matter from human activities in a residence, J. Expo. Anal. Environ. Epidemiol. 14 (2004) S34eS40. [8] V. Acevedo-Bolton, W.R. Ott, K.C. Cheng, R.T. Jiang, N.E. Klepeis, L.M. Hildemann, Controlled experiments measuring personal exposure to PM2.5 in close proximity to cigarette smoking, Indoor Air 24 (2014) 199e212. hin, O. Ramalho, S. Kirchner, Size distribution and emission rate mea[9] E. Ge surement of fine and ultrafine particle from indoor human activities, Atmos. Environ. 42 (2008) 8341e8352.  , I. Kopanakis, M. Lazaridis, Character[10] T. Glytsos, J. Ondra cek, L. D zumbova ization of particulate matter concentrations during controlled indoor activities, Atmos. Environ. 44 (2010) 1539e1549. [11] J. Qian, A.R. Ferro, K.R. Fowler, Estimating the resuspension rate and residence time of indoor particles, J. Air Waste Manag. Assoc. 58 (2008) 502e516. [12] K.C. Cheng, M.D. Goebes, L.M. Hildemann, Association of size-resolved airborne particles with foot traffic inside a carpeted hallway, Atmos. Environ. 44 (2010) 2062e2066.  s, B. So €veges, T. Weidinger, G. Kristo  f, N. Pe ter, [13] I. Salma, K. Doszt aly, T. Borso sz, Physical properties, chemical composition, sources, spatial distriZs Kerte bution and sinks of indoor aerosol particles in a university lecture hall, Atmos. Environ. 64 (2013) 219e228. [14] M.P. Spilak, B.E. Boor, A. Novoselac, R.L. Corsi, Impact of bedding arrangements, pillows, and blankets on particle resuspension in the sleep microenvironment, Build. Environ. 81 (2014) 60e68. [15] P.G. Mahaffy, N.I. Martin, K.E. Newman, B. Hohn, R.J. Mikula, V.A. Munoz, Laundry dryer lint: a novel matrix for nonintrusive environmental lead screening, Environ. Sci. Technol. 32 (1998) 2467e2473. [16] J. Ene-Parent, L. Zikovsky, Neutron activation analysis of laundry dryer lint, J. Radioanal. Nucl. Chem. 247 (2001) 197e198. [17] H.M. Stapleton, N.G. Dodder, J.H. Offenberg, M.M. Schantz, S.A. Wise, Polybrominated diphenyl ethers in house dust and clothes dryer lint, Environ. Sci. Technol. 39 (2005) 925e931. [18] A. Schecter, N. Shah, J.A. Colacino, S.I. Brummitt, V. Ramakrishnan, T.R. Harris, O. P€ apke, PDBEs in US and German clothes dryer lint: a potential source of indoor contamination and exposure, Chemosphere 75 (2009) 623e628. [19] M. Shoeib, T. Harner, G.M. Webster, S.C. Lee, Indoor sources of poly- and perfluorinated compounds (PFCS) in Vancouver, Canada: implications for human exposure, Environ. Sci. Technol. 45 (2011) 7999e8005. [20] K. Sarlo, D.B. Kirchner, E. Troyano, L.A. Smith, G.J. Carr, C. Rodriguez, Assessing the risk of type 1 allergy to enzymes present in laundry and cleaning products: evidence from the clinical data, Toxicology 271 (2010) 87e93. [21] D. Massey, A. Kulshrestha, J. Masih, M. Habil, A. Taneja, Indoor/outdoor relationship of fine particles less than 2.5 mm (PM2.5) in residential homes locations in central Indian region, Build. Environ. 44 (2009) 2037e2045. [22] D. Massey, A. Kulshrestha, J. Masih, A. Taneja, Seasonal trends of PM10, PM5.0, PM2.5 & PM1.0 in indoor and outdoor environments of residential homes located in North-Central India, Build. Environ. 47 (2012) 223e231. [23] K.C. Cheng, V. Acevedo-Bolton, R.T. Jiang, N.E. Klepeis, W.R. Ott, O.B. Fringer, L.M. Hildemann, Modeling exposure close to air pollution sources in naturally ventilated residences: association of turbulent diffusion coefficient with air change rate, Environ. Sci. Technol. 45 (2011) 4016e4022. [24] K.C. Cheng, V. Acevedo-Bolton, R.T. Jiang, N.E. Klepeis, W.R. Ott, P.K. Kitanidis, L.M. Hildemann, Stochastic modeling of short-term exposure close to an air pollution source in a naturally ventilated room: an autocorrelated random walk method, J. Expo. Sci. Environ. Epidemiol. 24 (2014) 311e318.