Next Article in Journal
Keeping Antibiotics at Home Promotes Self-Medication with Antibiotics among Chinese University Students
Next Article in Special Issue
A System Based on the Internet of Things for Real-Time Particle Monitoring in Buildings
Previous Article in Journal
Development, Implementation and Evaluation of an Educational Intervention to Prevent Low Speed Vehicle Run-Over Events: Lessons Learned
Previous Article in Special Issue
Association between Four-Level Categorisation of Indoor Exposure and Perceived Indoor Air Quality

Int. J. Environ. Res. Public Health 2018, 15(4), 686; https://doi.org/10.3390/ijerph15040686

Article
Source Apportionment and Influencing Factor Analysis of Residential Indoor PM2.5 in Beijing
National Institute of Environmental Health, Chinese Center for Disease Control and Prevention, Beijing 100021, China
*
Author to whom correspondence should be addressed.
Received: 28 February 2018 / Accepted: 3 April 2018 / Published: 5 April 2018

Abstract

:
In order to identify the sources of indoor PM2.5 and to check which factors influence the concentration of indoor PM2.5 and chemical elements, indoor concentrations of PM2.5 and its related elements in residential houses in Beijing were explored. Indoor and outdoor PM2.5 samples that were monitored continuously for one week were collected. Indoor and outdoor concentrations of PM2.5 and 15 elements (Al, As, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, Pb, Se, Tl, V, Zn) were calculated and compared. The median indoor concentration of PM2.5 was 57.64 μg/m3. For elements in indoor PM2.5, Cd and As may be sensitive to indoor smoking, Zn, Ca and Al may be related to indoor sources other than smoking, Pb, V and Se may mainly come from outdoor. Five factors were extracted for indoor PM2.5 by factor analysis, explained 76.8% of total variance, outdoor sources contributed more than indoor sources. Multiple linear regression analysis for indoor PM2.5, Cd and Pb was performed. Indoor PM2.5 was influenced by factors including outdoor PM2.5, smoking during sampling, outdoor temperature and time of air conditioner use. Indoor Cd was affected by factors including smoking during sampling, outdoor Cd and building age. Indoor Pb concentration was associated with factors including outdoor Pb and time of window open per day, building age and RH. In conclusion, indoor PM2.5 mainly comes from outdoor sources, and the contributions of indoor sources also cannot be ignored. Factors associated indoor and outdoor air exchange can influence the concentrations of indoor PM2.5 and its constituents.
Keywords:
indoor PM2.5; residential houses; chemical elements; source apportionment; influencing factor analysis

1. Introduction

The majority of adults spend more than 80% of the day in a variety of indoor environments, mainly in their houses [1,2,3]. The contribution of indoor air is very important for accurate estimation of individual air pollution exposure. A vast number of epidemiological studies have shown that both short and long-term exposure to fine particulate matter (PM2.5) are harmful to the human respiratory and cardiovascular systems [4,5,6,7,8,9,10]. However, the PM2.5 exposure assessment in most of these studies is usually based on ambient PM2.5 data, regardless of the difference between indoor and outdoor PM2.5 concentration. And failure to account for the contribution of indoor PM2.5 concentration may lead to exposure misclassifications, which have become source of measurement bias in the PM2.5 epidemiologic studies linking health effects to PM2.5 exposure [11,12,13,14].
Indoor PM2.5 comes from both outdoor and indoor sources. Outdoor PM2.5 typically comes from sources such as fossil fuels combustion, motor vehicle exhaust, industrial emissions, soil and dust, and particles from these sources can penetrate into indoor through building gap, door, window and mechanical ventilation [12,15,16,17,18]. A survey conducted in Beijing showed that 54–63% of indoor PM2.5 originated from the outdoors with all windows closed [15]. Indoor sources such as smoking, cooking and indoor activities, also can generate particles to increase indoor PM2.5 concentration, those sources have been recognized as another important contributor of indoor PM2.5 [19,20,21,22,23].
The source apportionment of outdoor PM2.5 have been explored in many studies across the world [16,17,18,24,25,26,27,28], but information about source apportionment of indoor PM2.5, especially for residential houses, is relatively scarce. Chemical elements, including metal and metalloid elements, are common constituents of PM2.5 in urban cities, and the use of elements as tracers to identify PM2.5 from different sources has been applied in many studies [16,25,29,30,31,32,33,34,35,36]. Chemical elements in outdoor PM2.5 can be divided into several tracer groups for different sources, such as, Al, Mg and Ca which are usually used as trace elements for soil dust, V and Se are usually used as trace elements for fossil fuel combustion; Pb and Cu are usually used as trace elements of motor vehicle exhaust [16,29,30,31,32,33,34,35,36]. Indoor sources can also provide chemical elements for PM2.5. Elements such as Cd, As, Cu, Ni, Hg, Mn and Al have been detected in cigarette smoke [37,38,39,40], and Zn, Al, Mg and Ca can be found in cooking fumes [41,42,43].
Metal and metalloid elements attached to particles can enter the body through breathing. Previous studies have also addressed that those elements in PM2.5 are linked to diseases such as stroke, pneumonia and asthma [44,45,46,47]. Besides, the toxicity of elements on human health has been reported in many studies [48,49,50,51,52,53], especially for Cd and Pb, whereby exposure to low concentrations of Cd was associated with renal toxicity, osteoporosis and bone fractures [49,50,51], and low Pb exposure could lead to neurotoxicity and immune system damage [52,53], so metal and metalloid elements have been suspected to play an important role in the health effects of PM2.5 [16,47,54,55]. Therefore, understanding concentrations of metal and metalloid elements is essential not only for source apportionment of indoor PM2.5, but also for controlling those sources to reduce the health hazards.
Additionally, there are factors can affect the exchange of indoor and outdoor air, such as window opening behaviors, building characteristics, air conditioner use, and outdoor meteorological variables, etc., and these factors may also have an impact on the concentration of indoor PM2.5 and its constituents [12,13,22,56,57,58,59,60,61,62,63,64].
Concentrations of indoor PM2.5 and related chemical elements in residential house have been explored in several individual cities, such as in Kocaeli City [65], Saint Paul [55], Antwerp [66] and other cities [67,68,69], but the levels and sources of PM2.5 pollution, meteorological conditions, human behaviors and building characteristics are not always the same in different cities. In Beijing, ambient PM2.5 pollution has been serious in recent years, especially in autumn and winter [18,30,70]. Ambient PM2.5 pollution in Beijing is dominated by vehicle emissions [71,72], but fossil fuels combustion from heating activities has become another important contributor to ambient PM2.5 during the heating season [30,31,73]. Several studies have investigated indoor PM2.5 concentrations and their relationship with outdoor PM2.5 in residential houses in Beijing [58,59,60,61,62,63], but the information about concentrations of metal and metalloid elements in indoor PM2.5 is quite rare [74,75,76]. In this study, indoor and outdoor PM2.5 of residential houses were continuously monitored for one week during the non-heating season (NHS) and heating season (HS) in Beijing. The concentrations of PM2.5 and 15 kinds of related elements (Al, As, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, Pb, Se, Tl, V, Zn) were calculated, indoor and outdoor concentrations of PM2.5 and its constituents were compared. The possible sources of indoor PM2.5 and potential influencing factors of indoor PM2.5 and its constituents’ concentrations were analyzed.

2. Methods

2.1. Participant Recruitment

The participants were recruited by telephone interview to determine the building type and floor of the houses. Since the majority of citizens in Beijing live in apartment buildings, and a certain percentage of citizens live in villas and courtyards, all three types of building were surveyed. Apartment buildings were divided into three groups based on the building height: 1–3 floors, 4–9 floors and 10 floors or above. Stratified sampling was used to select houses based on their building type and floor group, and there were at least five houses in each type of building and each group of apartments.

2.2. Indoor and Outdoor PM2.5 Sampling

The indoor and outdoor sampling was conducted in two seasons: August 2013–October 2013 was defined as the non-heating season, and November 2013–March 2014 was defined as the heating season. Each house was sampled for a continuous seven-day during the non-heating season and heating season. Indoor samplers were placed in the living room (if the living room was not available, then placed in a bedroom). The intake port of the outdoor sampler was extended at least 1 meter outside the available windows that were hardly used, and away from the outlet of the air conditioner (AC) or extractor hood. The PM2.5 samples of the indoor and outdoor areas were collected at the same time by a PCXR8-Universal Sampling Pump (SKC, Eighty Four, PA, USA). The sampling target flow rate was 0.5 L/min. Samples with flow rates more than 20% different from the target rate were considered invalid. The indoor and outdoor PM2.5 was collected on 37 mm Teflon PTFE filter and their mass concentrations were measured by the gravimetric method. Daily outdoor atmospheric pressure, temperature, relative humidity (RH) and wind speed during sampling periods were obtained from the National Meteorological Information Center (Beijing, China). More information about indoor/outdoor sampling had been previously described [58].

2.3. Questionnaires and Recoding Sheets

Questionnaires and recoding sheets were investigated during the sampling to obtain the information of potential influencing variables of indoor PM2.5. A baseline questionnaire was completed by each house at the beginning of indoor/outdoor sampling in non-heating season, to collect variables of building characteristics and general inhabitant behaviors. At the end of each sampling in two seasons, a sampling questionnaire was completed to observe the consistency and diversity of resident’s behavior during sampling period, compared to information from baseline questionnaire. The main information of these two questionnaires was shown in Table S1 in the Supplementary Materials. The residents also recorded information on cooking, window opening and air conditioner (AC) use every day on a recording sheet.

2.4. Determination of PM2.5 and Chemical Elements

The mass concentrations of PM2.5 were measured with an analytical balance (XP6, Mettler Toledo, Zurich, Switzerland, verification scale value of 0.001 mg). Before and after sampling, the filters were pre-conditioned for a minimum of 24 h at a constant temperature of 25 °C (±1 °C) and constant RH of 50% ± 5% before weighing.
The PM2.5 mass concentrations of indoor and outdoor areas (ρ, μg/m3) were calculated with the following equation:
ρ = w 2 w 1 V × 1000
The weight mass of the filter before sampling (w1, mg) and the weight mass of the filter after sampling (w2, mg) were measured with the same analytical balance mentioned above. The sampling volume was calculated by the recorded sampling cumulative time and the flow rate of the pump and then converted into the standard state (V, m3).
After PM2.5 weight analysis was completed, the elements concentrations of each sample were analyzed by inductively coupled plasma mass spectrometry analysis method (ICP-MS, NexION 300D, PerkinElmer, Waltham, MA, USA). Half of each Teflon filter were extracted with 7 mL of nitric acid (HNO3, 69%), 0.5 mL of hydrochloric acid (HCL, 36–38%) and 0.5 mL of hydrofluoric acid (HF, 47–51%). The purity of HNO3, HF and HCL was guaranteed reagent (GR). The digested solutions were allowed to pre-digest for 2 h in room temperature before loading into the microwave digestion system (MARs5, CEM, Matthews, NC, USA). A stepped microwave program was set: 5 min at 80 °C, followed by 5 min at 120 °C, 5 min at 160 °C, 20 min at 200 °C, and then cooled to room temperature. After the above stepped digestion process, the digestion tanks were opened and the tube wall were rinsed with ultrapure water, then the tank were set into acid-driven processor (BHW-09C, Botonyc Company, Shanghai, China) at 160 °C to clean up the acid to about 1 mL. At last, the extracts were diluted to 10 mL with ultrapure water for ICP-MS analysis. A summary of the operation conditions of the ICP-MS used in this study is shown in Table S2 in the Supplementary Materials. The detection limit of elements in this study was characterized by background equivalent concentration (BEC) value of 10 μg/L standard solution in standard mode (STD), summary of detection limits for chemical elements was shown in Table S3 in Supplementary Materials, and one-half of the detection limit of elements was used for the treatment of values below the detection limit.

2.5. Statistical Analysis

Distribution of household characteristics and meteorological variables, indoor and outdoor concentrations of PM2.5 as well as related elements was analyzed by descriptive statistical analysis. Since the concentrations of PM2.5 and its constituents were not normal distribution tested by Kolmogorov-Smirnov test, comparison of indoor and outdoor concentrations of PM2.5 and its constituents was performed by Wilcoxon signed-rank test. Mann-Whitney U test was used to compare the concentrations of PM2.5 and its constituents in different groups. Indoor/outdoor correlation coefficient of PM2.5 and its constituents was evaluated by spearman correlation analysis. A varimax rotated factor analysis (FA) was performed base on indoor elements data to identify sources of indoor PM2.5 in all residential houses. First, based on the concentrations data of elements in indoor PM2.5, correlation matrix for each element was obtained. Then, by solving the characteristic equation of the correlation matrix, eigenvalues of factors and factor loading matrix was calculated. At last, the varimax rotated method was used to maximize the sum of the variance of each factor by orthogonally rotating the factor loading matrix, which can contribute to clarify the professional meaning of each factor. The eigenvalue was set to 1.0 as a threshold to limit the number of extracted factors.
The relationship between indoor concentrations of PM2.5 and its constituents and their potential influencing variables were examined by multivariate linear regressions analysis. Indoor concentrations of PM2.5 and selected elements were log-transformed to achieve a normal distribution (confirmed by the Kolmogorov-Smirnov test), then they were set as dependent variables, potential influencing variables were set as independent variables. A stepwise variable selection procedure was used for the multivariate linear regression, with an entry level of 0.10 and an exclusion level of 0.05. Season-specific analyses were constructed under the assumption that relationship between indoor PM2.5 and potential variables would vary in two seasons. All statistical analyses were performed in the SPSS 22.0 software (SPSS Inc., Chicago, IL, USA), and the significance level was 0.05.

3. Results

3.1. Summary of Sampled Houses

A total of 52 houses in NHS and 55 houses in HS were sampled. After excluding the failed samples caused by filter damage and invalid flow rates, 47 houses in NHS and 47 houses in HS were left for further analysis. There were 28 apartments, five villas and 14 courtyards in NHS, and 26, five, 16 in HS, respectively. The descriptive household characteristics of these houses were provided in Table 1. Among the 47 houses in HS, 37 houses were successfully sampled in both seasons.

3.2. Distribution of Potential Influencing Variables for Indoor PM2.5

More than half of the houses were more than 20 years old. Most houses used liquefied petroleum gas and natural gas for cooking, and hoods were also used during cooking. The ratio of houses with indoor smoking was 10/47 in both seasons. No house used air cleaner and fresh air system during sampling. In HS, most houses used centralized heating, and no house used AC. In NHS, nine houses used an AC, and the average cumulative time of AC use during sampling was 6.8 h (range from 0.0–50.2 h). The median time of window opening per day was 14 h in NHS, but only 0.5 h in HS, respectively, 19 houses did not open window during sampling in HS. Outdoor temperature, average wind speed and atmospheric pressure were different in two seasons. The mean ± standard deviation of average outdoor temperature, average wind speed and atmospheric pressure in the NHS were 20.34 ± 6.18 °C, 8.05 ± 1.15 m/s and 101.24 ± 0.82 kPa, respectively, whereas they were 0.14 ± 2.58 °C, 11.80 ± 2.81 m/s and 102.64 ± 0.43 kPa, respectively, in the HS. The RH had no difference between two seasons.

3.3. Residential Indoor and Outdoor Concentration of PM2.5 and Elements

Table 2 provides descriptive data for PM2.5 and element concentrations of the indoor and outdoor areas. The median (Interquartile Range, IQR) of indoor and outdoor PM2.5 concentrations for all houses were 57.64(27.51) μg/m3 and 72.49(43.53) μg/m3. The median indoor concentration of 15 elements ranged from 0.53 ng/m3 to 930.24 ng/m3. For As and Se, indoor concentrations of about one-third houses were below the detection limits. The median (IQR) concentration of indoor PM2.5 was 55.44(24.12) μg/m3 in NHS and 58.23(13.28) μg/m3 in HS, and indoor PM2.5 were significantly lower than outdoor in both seasons. Outdoor concentrations of all 15 elements were higher than indoor in HS, but for Fe, As and Ca, indoor and outdoor concentrations were not different in NHS.
Spearman correlation analysis was used to evaluate the correlation of indoor and outdoor, results showed that indoor and outdoor PM2.5 concentrations were significantly correlated (p < 0.01) in both seasons, the correlation coefficient (r) was 0.56 in NHS and 0.44 in HS. Indoor Fe and Ca were not correlated with outdoor in both seasons, and indoor Al, Mn, Cu and Mg only correlated with outdoor in NHS. Indoor Pb concentration was highly correlated with outdoor, r = 0.73 in NHS and 0.85 in HS.

3.4. Elements in Indoor PM2.5 that Were Sensitive to Indoor Smoking

Houses with indoor smoking during the sampling were defined as smoking houses, and houses without indoor smoking were defined as non-smoking houses. Concentrations of PM2.5 and its constituents for smoking houses and non-smoking houses were compared in Table 3. For smoking and non-smoking houses, median indoor PM2.5 concentrations were 66.92 μg/m3 and 53.51 μg/m3 in NHS, 129.34 μg/m3 and 54.60 μg/m3 in HS, respectively. After comparing the indoor concentrations of PM2.5 and its constituents between smoking houses and non-smoking houses, indoor Cd was higher in smoking houses than in non-smoking houses in both seasons, the difference for indoor PM2.5 were only seen in HS. Indoor concentrations of PM2.5 and its constituents for smoking houses were also compared with their outdoor concentrations. Results showed that indoor and outdoor PM2.5 were no longer different in both seasons, indoor As in NHS and Cd in both seasons were higher than outside. Therefore, Cd and As in indoor PM2.5 may be the sensitive elements of indoor smoking.

3.5. Elements in Indoor PM2.5 that Were Related to Indoor Sources other than Smoking

In HS, residents of 19 houses did not open any windows during sampling, five of those 19 houses had indoor smoking. Indoor and outdoor concentrations of PM2.5 and its constituents for these houses were compared in Table 4, to find the elements in indoor PM2.5 related to indoor sources other than smoking. For the 19 houses, indoor concentrations of PM2.5, Al, Zn, As, Cd, Na and Ca were not different from outdoor, while other elements in indoor PM2.5 were significantly lower than outdoor. After excluding the five houses with smoking, the median indoor concentration of PM2.5 decreased from 69.53 μg/m3 to 61.77 μg/m3, indoor PM2.5, Cd, As and Na became significantly lower than outdoor in these 14 houses, however, the indoor concentrations of Zn, Al and Ca were still not different from outdoor. In Table 3, concentrations of all elements in indoor PM2.5, including Zn, Al and Ca, were significantly lower than outdoor for 37 non-smoking houses in HS, so the result in Table 4 that indoor and outdoor concentrations of Zn, Al and Ca were not different in these 14 houses may be due to the contribution of indoor sources other than indoor smoking.
Among the 14 houses, 13 houses had indoor cooking, and Zn was reported to be one of the most abundant metals in the cooking fumes, therefore the result of high indoor concentration of Zn may be due to indoor cooking. Al and Ca can be found in cooking fumes from previous studies, but they may also come from re-suspended indoor dust caused by human activities, so the high indoor concentrations of Al and Ca may be due to the indoor cooking and re-suspended dust.

3.6. Elements in Indoor PM2.5 that Mainly Came from Outdoor

Among all houses, 37 houses were successful sampled in both seasons. A Wilcoxon signed-rank test was used to compare the concentrations of PM2.5 and its constituents between two seasons for those houses, and the results are presented in Table 5. Outdoor PM2.5 concentrations were much higher in HS, while indoor PM2.5 concentrations were not different in the two seasons. Outdoor concentrations of Pb, V, Se and Tl did not differ between the two seasons, but indoor Pb, V, Se and Tl were lower in HS, except that Pb was not significant. From the results in Table 3, indoor concentrations of Pb, Se in HS and V in NHS were significantly lower than outdoors for smoking houses, suggested that Pb, V, Se was not sensitive to indoor smoking. And from the result in Table 4, Pb, V and Se was also not related to other indoor sources. So, Pb, V and Se in indoor PM2.5 may mainly come from ambient air.

3.7. Factor Analysis of Indoor PM2.5

Results of factor analysis were showed in Table 6. Factors loadings equal to or greater than 0.4 are presented in Table 6, factors loadings smaller than 0.40 were omitted from the table, the purpose of which was to facilitate the identification of indoor PM2.5 sources. The possible source types were identified based on previous published studies and above analysis. Five factors were extracted from indoor elements data in PM2.5, which explained about 76.8% of the total variance, with the first factor (F1) accounting for 21.6%, the second (F2) 21.3%, the third (F3) 13.8%, the fourth (F4) 11.4%, and the fifth (F5) 8.8%. F1 showed high loadings for Fe, V, Mn, Cd and Cu, and these elements all related to products of fossil fuels combustion, so F1 was identified as combustion factor. F2 contained high loadings for Pb, Se, Tl and K, and moderately loadings for Cu and Mn. Elements such as Pb, Cu and Mn can be generated from motor vehicles, but Se, Pb, Tl and K related to products of fossil fuel combustion, so F2 might be factor that combined motor vehicles and combustion. F3 correlated with Zn, Ca and Mg, Zn and Ca may be related to indoor cooking from above results, therefore F3 was identified as indoor cooking factor. F4 involved with Al, Na and Mg, and these compounds have been associated with dust and soil. F5 correlated with As and Cd, these two elements may be sensitive to indoor smoking as previously mentioned, so F5 was identified as an indoor smoking factor. Besides, the factor scores of F5 in smoking houses were significantly higher than in no-smoking houses (p < 0.01), which in turn confirmed the identification of F5.

3.8. Influencing Factor Analysis of Indoor PM2.5 and Selected Elements

Multiple linear regression analysis was performed for log-transformed indoor PM2.5, Cd and Pb to identify their influencing factors. Cd was chosen to represent smoking activity because it was consistently sensitive to indoor smoking from the above analysis. Pb was chosen to represent outdoor sources because indoor Pb was not sensitive to indoor smoking and cooking, and it was highly correlated with outdoor Pb in both seasons.
The results of multiple linear regression analysis were shown in Table 7. Coefficient of determination (R2) reflects the percentage variance of indoor PM2.5 and its constituents that the influencing variables could explain. In the NHS, indoor PM2.5 concentration was mainly influenced by outdoor PM2.5 concentration, smoking during sampling and outdoor temperature and time of air conditioner use during sampling in NHS (R2 = 0.60), it was influenced by outdoor PM2.5 and smoking during sampling in HS (R2 = 0.46). Indoor Cd concentration was mainly influenced by smoking during sampling and outdoor Cd concentration in NHS (R2 = 0.51), it was influenced by smoking during sampling, outdoor Cd concentration and building age in HS (R2 = 0.63). Indoor Pb concentration was mainly influenced by outdoor Pb concentration and time of window open per day in the NHS (R2 = 0.63), it was influenced by outdoor Pb concentration, building age and RH in HS (R2 = 0.74).
The results of partial regression coefficient (β) for independent variables in Table 7 indicated that high outdoor concentration of PM2.5, Cd and Pb can increase their indoor concentration in both seasons. Smoking during the sampling period could lead to high indoor PM2.5 and Cd concentrations in both seasons. In HS, indoor concentrations of Cd and Pb may be higher for older houses. In NHS, more hours of window opening may come with higher indoor Pb concentrations, and less hours of air conditioner use may result in higher indoor PM2.5 concentration. For meteorological variables, high outdoor temperature may increase indoor PM2.5 concentration in NHS, but high RH may decrease indoor Pb in HS.

4. Discussion

Outdoor PM2.5 concentration in heating season was much higher than in non-heating season in this study, a seasonal difference that has also been reported in several other studies conducted in Beijing [30,53,77,78], but not consistent with the study conducted in Beijing by Han [60] which may due to the short sampling periods and the frequent and rapid transition between severe pollution events and clean days. Indoor PM2.5 were significantly lower than outdoor in both seasons from Table 2, similar results showed in other studies [59,60,61], but indoor PM2.5 concentration was not different between two seasons from Table 5, which may mainly due to the low indoor and outdoor exchange in heating season and the contributions of indoor sources. The median concentration of indoor PM2.5 was 57.64 μg/m3. Median concentrations of elements in indoor PM2.5 ranged from 0.53 ng/m3 to 930.24 ng/m3, in descending order, K > Fe > Na > Al > Ca > Zn > Mg > Pb > Mn > Cu > As > Se > Cd > V > Tl. Elements such as K, Fe, Na, Al and Ca were abundant and widespread in the natural environment, their concentration was generally higher, but the contents of elements such as V, As and Se in PM2.5 were relative low, even below the detection limits.
Five factors with eigenvalue greater than 1.0 were extracted from the factor analysis based on the elements data in Table 6; they could explain about 76.8% of the total variance of indoor PM2.5. Since Pb, V and Se in indoor PM2.5 were assumed mainly to come from outdoors from the above analysis, and these three elements only showed high loading in F1 and F2, so F1 and F2 were recognized as outdoor factors, and their contribution to the variance of indoor PM2.5 was 42.9%. Elements like V, Fe, Cd, Mn, Se, Cu, Pb, Tl and K were detected from products of fossil fuel combustion in previous studies [16,29,30,31,35,36,55]. Combustion of fossil fuels was reported as the main anthropogenic V emission [16,35,55], but V only showed high loading of F1 in this study, and F1 also had high loading for Fe, Mn, Cd, therefore F1 was preliminarily identified as a combustion factor. Vehicle emissions have been identified as an important source for ambient PM2.5 pollution in Beijing [30,71,72], especially in the non-heating season, and both gasoline and diesel oil for fuels of motor vehicles originate from fossil fuels (petroleum), so the elements from motor vehicle emissions might be similar to those from fossil fuel combustion. Besides, previous studies also suggested that motor vehicle exhaust may contain Pb and Cu and motor vehicle fuel additives may contain Pb and Mn [16,29,30,36]. Pb, Tl and Se only showed high loadings in F2, so F2 might be a factor that combined motor vehicles and combustion.
F3 and F5 were recognized as indoor source (cooking and smoking) factors, where their contribution was 20.2%. Chemical elements in cooking fumes has been analyzed in several studies, and contents of Al, Zn, Mg and Ca were relatively high in PM2.5 samples from different kitchens [41,42,43], and Zn was considered to be one of the most abundant metals in cooking fumes [41,43]. In this study, data of 14 houses without smoking and window opening in heating season were analyzed in Table 4, and found that Zn and Ca may be related to indoor cooking. F3 shows high loading for Zn and Ca, and a moderate loading for Mg, so it was identified as an indoor cooking factor. Chemical elements in cigarette smoking also have been analyzed in many studies [37,38,39,40], Cd and As were detected in a majority of these studies. Of course, there are other elements released from cigarette smoking such as Al, Mn, Cu, etc. [37,38,39], but in this study, data of smoking and non-smoking houses was compared in Table 3, it was found that only Cd and As were sensitive to indoor smoking, especially Cd. A likely explanation is that for houses without smoking, the content of indoor Cd and As was relatively low and other elements like Al and Mn were much higher, proved by the above order of elements concentrations. Once indoor smoking occurred, the contribution of smoking to Cd and As can be observed, but less noticeable for Al and Mn. In addition, a comparative analysis was performed for F5, the factor score of F5 in smoking and non-smoking houses were compared, houses with indoor smoking had higher scores, so it is reasonable to identify F5 as an indoor smoking factor. A study conducted in Kocaeli City [65] also performed factor analysis for indoor PM2.5, where cooking factors and smoking factors contributed 9.7% and 10.6% of the variances, close to 11.4% and 8.8% in this study.
F4 was identified as a dust and soil factor, mainly because it had a high loading of Al, Na, Mg, and these three elements are most often found in soil and dust, which may originate from earth and soil in the ambient environment, or from suspended dust of floors and construction [26,29,30,33,49,65].
Outdoor PM2.5 is an important source of residential indoor PM2.5 [12,13,15,64], and it was significantly correlated with indoor PM2.5 concentrations in both seasons from the results of Spearman correlation analysis, and similar results also can be found in other studies [20,59,60,61,62,63]. Multiple linear regression analysis in Table 7 showed that with the increase of outdoor concentration, indoor PM2.5 concentration also increased in both seasons. Because indoor and outdoor concentrations of Pb were highly correlated in both seasons and indoor Pb was not sensitive to indoor smoking and cooking, it was selected to represent outdoor sources for multiple linear regression. Results showed that indoor Pb concentration only significantly correlated with variables including outdoor Pb concentration, time of window open per day, building age and RH. And standard regression coefficient (β’) of outdoor Pb concentration was much higher than other variables, indicating that indoor Pb in PM2.5 did mainly come from outdoors in this study.
Indoor smoking was identified as an indoor PM2.5 source, which can cause a sharp short-term increase in PM2.5 concentrations [12,61,65]. In the non-heating season, indoor PM2.5 concentrations in smoking houses was only slightly higher than in non-smoking houses, but in heating season the concentrations of indoor PM2.5 in smoking houses was more than twice as high as in non-smoking houses from Table 3. Results of multiple linear regression analysis also demonstrated that indoor smoking was an important influencing factor of indoor PM2.5 in both seasons. In the comparative analysis from Table 3, Cd was the only element that was consistently sensitive to indoor smoking in both seasons, so Cd was chosen to represent smoking activity for the multiple linear regression analysis. As expected, the results in Table 7 showed that smoking during sampling can increase indoor Cd concentrations in both seasons. However, it should be noted that outdoor Cd was also an important contributor to indoor Cd. Indoor Cd was correlated with outdoor Cd, r = 0.54 in the non-heating season, and 0.41 in the heating season from Spearman analysis. The reason that indoor Cd was sensitive to indoor smoking may due to the relatively low concentration which has already been explained above.
Apart from the above indoor and outdoor sources, factors related to the exchange of indoor and outdoor air also can affect indoor PM2.5. Window ventilation can promote the exchange of indoor and outdoor air to help outdoor PM2.5 move into a room [22,56,57,58,59,60,61,62,79]. Natural ventilation was associated with higher indoor PM2.5 [79]. In this study, a significant correlation between indoor Pb and time of window open per day was observed in non-heating season, houses with more hours of window opening had higher level of Pb from Table 7. As mentioned before, Pb was chosen as a representative element for outdoor sources, so these results indicate that window ventilation can promote the movement of outdoor PM2.5 indoors. However, indoor PM2.5 was not correlated with window ventilation, but was inversely related to time of air conditioner use in the non-heating season from Table 7. A possible explanation is that residents tend to close all the windows and reduce indoor and outdoor air exchanges when air conditioners are used to save energy, and air conditioners were also reported to increase indoor PM2.5 deposition by a filtering process [80], so indoor PM2.5, especially from indoor sources, may decrease when air conditioners are working. No significant correlation between indoor PM2.5, Pb and window ventilation in heating season was found in multiple linear regression analysis, which could be partially explained by the short window opening time and small window opening areas, as the median time of window opening per day was only 0.5 h in the heating season. Previous studies suggested that as the building age increases, the airtightness decreases, leading to high infiltration rates of outdoor particles [81,82]. Similar results can also be found in this study. In the heating season, houses with older building age had higher indoor Cd and Pb concentrations.
Meteorological variables may also affect indoor and outdoor air exchanges, thus influencing indoor PM2.5 concentrations [22,56,57,58,59,60]. In this study, the outdoor temperature was associated with indoor PM2.5 in the non-heating season from Table 7. One possible explanation is that the outdoor temperature ranged from 9.4 to 27.9 °C during sampling in the non-heating season, and residents may tend to increase window ventilation with the increase of temperature, therefore more outdoor PM2.5 enters indoors, leading to higher indoor PM2.5 concentrations. RH were correlated with indoor Pb in the heating season. High relative humidity often appeared during rainy or snowy weather in the heating season, so residents may close windows and reduce indoor/outdoor air exchange, decreasing the indoor Pb concentration.
Indoor cooking was another identified indoor source of PM2.5 [12,22,23,59,83], but a significant correlation between cooking and indoor PM2.5 was not observed in our study; this correlation was also not identified in another study conducted in Beijing [61]. One possible explanation is that the cooking time was quite short compared to the entire sampling time, and another study conducted in Beijing suggested that cooking activity can generate sharp but narrow additive pulse peaks of indoor PM2.5 concentration [59], so PM2.5 may be largely created through cooking, but its contribution to the entire sample may be low. Another explanation is that most houses used a hood and closed the kitchen doors during cooking, and our sampling pumps were usually placed in the living room or bedroom. A previous study also reported that separating the cooking place from the other rooms could reduce indoor PM2.5 [84], so, the PM2.5 from cooking could hardly be collected. Besides, we did not choose elements representing indoor cooking for multiple linear regression analysis. Though we collected some information related to indoor cooking, such as frequency of cooking, kitchen door and hoods used during sampling, etc., that was not enough because the concentrations of elements from indoor cooking can be greatly influenced by the cooking style, oils and foods used for cooking [41,42], and information about these variables were not collected in this study. Another reason is most of the houses in this study did indoor cooking during the sampling period, so the study lacked a certain number of non-cooking houses as a control to compare and analyze the results. Researchers should take these variables previously mentioned into account for future indoor cooking-related studies.
Indoor human activities such as movement and cleaning can lead to resuspension of dust and increase the concentration of indoor particles [59,83,85]. A study conducted in Sweden suggested that high concentration of particles mainly occurred during active periods of occupancy, but the variable related to indoor human activities collected in this study was population (number of residents in the house), and no significant correlation was found between population and concentrations of indoor PM2.5 and its constituents. Building type also failed to show any significant correlation with indoor PM2.5, Cd and Pb. One possible reason is that the residential ventilation behavior in different types of building was basically the same. Another reason is that indoor smoking existed in both courtyards and apartments. Besides Pb and Cd, we also tried to analyze potential influencing factors of other elements from indoor PM2.5, and an interesting result was found whereby building type was significantly associated with indoor Tl in the non-heating season, as indoor Tl in courtyards was lower than in apartments and villas. A likely explanation for this result is that Tl may exist in the new building materials which often used for apartments and villas.
However, there are also several limitations in our study. First, there were some potential variables that have not been taken into account in this study, such as indoor cleaning activities and indoor plants, which might affect the indoor PM2.5 according to previously published articles [12,22,57,59,61]. Second, houses that use coal and wood for cooking and heating were not included in this study, but the contribution of coal and wood burning to indoor PM2.5 and chemical elements is very large according to previous studies [41,69,86]. All the houses in this study did not use fresh air systems and air cleaners during the sampling periods, but they have gradually become widely used in recent years, so their impact on indoor PM2.5 should be explored in future studies. Third, the concentrations of PM2.5 and its elements in this study were calculated as the average level of the sampling week; therefore, some variables that can influence indoor PM2.5 concentrations in the short-term may not be observed and analyzed. Fourth, the data of potential influencing factors mostly were collected by questionnaire and recording sheets, so some information bias may exist in this study. At last, although the number of houses used for analysis in our study was comparable to previous studies, the sample size and the constituent data of indoor PM2.5 was still not sufficient, and they should be expanded for wider application of our findings because of the variability of residents’ activity and complexity of indoor PM2.5 sources.

5. Conclusions

In this study, targeted residential houses were separately and continuously sampled in the non-heating and heating season, and indoor and outdoor concentrations of PM2.5 and related elements of urban houses in Beijing were obtained. It was found that indoor concentrations of PM2.5 were more than twice higher in smoking houses than in non-smoking houses in the heating season, and indoor PM2.5 of smoking houses also higher in the non-heating season, but the difference was not significant. Therefore, people should avoid indoor smoking, particularly in the heating season, to reduce indoor PM2.5 concentrations and decrease potential health risks.
It was also found that among elements in indoor PM2.5, Cd and As may be sensitive to indoor smoking, Zn, Ca and Al may be related to indoor sources other than smoking, Pb, V and Se may mainly come from outdoors. Five factors were extracted for indoor PM2.5, two outdoor sources factors contributed more variance to indoor PM2.5, and the contribution of indoor source factor also cannot be ignored. The relationship between indoor concentrations of PM2.5 and two selected elements (Cd and Pb) and their potential influencing factors were explored, and it was found that outdoor concentration was an important factor for both PM2.5 and chemical elements, indoor smoking can increase indoor PM2.5 and Cd concentration, factors including window ventilation, air conditioner use, building age and meteorological variables may be associated with indoor and outdoor air exchange, and then affect indoor concentrations of PM2.5 and its constituents.
Several factors influencing indoor concentrations of PM2.5 and representative elements were found in this study which should be examined in future studies, and this study can contribute further to accurately assessing indoor PM2.5 exposure, and reducing the health impact of indoor PM2.5 and its elements on Beijing residents.

Supplementary Materials

The following are available online at https://www.mdpi.com/1660-4601/15/4/686/s1, Table S1: Main summary of the baseline questionnaire and sampling questionnaire, Table S2: Summary of detection limits for chemical elements in this study.

Acknowledgments

This research was supported by the Gong-Yi Program of the Chinese Ministry of Environmental Protection (201209008), the Gong-Yi Program of the Chinese Health and Family Planning Commission (201402022) and the Natural Science Foundation of Beijing (L150011). We appreciate the help and the cooperation of the residential families.

Author Contributions

Yibing Yang, Chunyu Xu, Na Li, Zhe Liu and Qin Wang performed the sampling work and the data collection. Yibing Yang, Liu Liu and Dongqun Xu analyzed the data and wrote the paper. Yibing Yang, Liu Liu, Chunyu Xu, Qin Wang and Dongqun Xu reviewed and edited the manuscript. Dongqun Xu also was the corresponding author of this paper. All authors read and approved the manuscript.

Conflicts of Interest

The authors declared that they have no conflicts of interest to this work.

References

  1. Klepeis, N.E.; Nelson, W.C.; Ott, W.R.; Robinson, J.P.; Tsang, A.M.; Switzer, P.; Behar, J.V.; Hern, S.C.; Engelmann, W.H. The National Human Activity Pattern Survey (NHAPS): A resource for assessing exposure to environmental pollutants. J. Expo. Sci. Environ. Epidemiol. 2001, 11, 231–252. [Google Scholar] [CrossRef] [PubMed]
  2. Robinson, J.; Nelson, W.C. National Human Activity Pattern Survey Data Base; USEPA: Research Triangle Park, NC, USA, 1995. [Google Scholar]
  3. Ministry of Environmental Protection. Exposure Factors Handbook of Chinese Population (Adults); China Environmental Science Press: Beijing, China, 2013. [Google Scholar]
  4. Dockery, D.W.; Pope, C.A., 3rd; Xu, X.; Spengler, J.D.; Ware, J.H.; Fay, M.E.; Ferris, B.G., Jr.; Speizer, F.E. An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 1993, 329, 1753–1759. [Google Scholar] [CrossRef] [PubMed]
  5. Laden, F.; Schwartz, J.; Speizer, F.E.; Dockery, D.W. Reduction in fine particulate air pollution and mortality: Extended follow-up of the Harvard Six Cities study. Am. J. Respir. Crit. Care 2006, 173, 667–672. [Google Scholar] [CrossRef] [PubMed]
  6. Pope, C.A.P., III; Burnett, R.T.; Thun, M.J.; Calle, E.E.; Krewski, D.; Ito, K.; Thurston, G.D. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002, 287, 1132–1141. [Google Scholar] [CrossRef] [PubMed]
  7. Miller, K.A.; Siscovick, D.S.; Sheppard, L.; Shepherd, K.; Sullivan, J.H.; Anderson, G.L.; Kaufman, J.D. Long-term exposure to air pollution and incidence of cardiovascular events in women. N. Engl. J. Med. 2007, 356, 447–458. [Google Scholar] [CrossRef] [PubMed]
  8. Eeftens, M.; Beelen, R.; De, H.K.; Bellander, T.; Cesaroni, G.; Cirach, M.; Declercq, C.; Dėdelė, A.; Dons, E.; de Nazelle, A.; et al. Development of Land Use Regression models for PM2.5, PM2.5 absorbance, PM10 and PM(coarse) in 20 European study areas; results of the ESCAPE project. Environ. Sci. Technol. 2012, 46, 11195–11205. [Google Scholar] [CrossRef] [PubMed]
  9. Atkinson, R.W.; Carey, I.M.; Kent, A.J.; Van Staa, T.P.; Anderson, H.R.; Cook, D.G. Long-term exposure to outdoor air pollution and incidence of cardiovascular diseases. Epidemiology 2013, 24, 44–53. [Google Scholar] [CrossRef] [PubMed]
  10. Wu, S.; Deng, F.; Wei, H.; Huang, J.; Wang, X.; Hao, Y.; Zheng, C.; Qin, Y.; Lv, H.; Shima, M. Association of cardiopulmonary health effects with source-appointed ambient fine particulate in Beijing, China: A combined analysis from the Healthy Volunteer Natural Relocation (HVNR) study. Environ. Sci. Technol. 2014, 48, 3438–3448. [Google Scholar] [CrossRef] [PubMed]
  11. Meng, Q.Y.; Turpin, B.J.; Polidori, A.; Lee, J.H.; Weisel, C.; Morandi, M.; Colome, S.; Stock, T.; Winer, A.; Zhang, J. PM2.5 of ambient origin: Estimates and exposure errors relevant to PM epidemiology. Environ. Sci. Technol. 2005, 39, 5105–5112. [Google Scholar] [CrossRef] [PubMed]
  12. Chen, C.; Zhao, B. Review of relationship between indoor and outdoor particles: I/O ratio, infiltration factor and penetration factor. Atmos. Environ. 2011, 45, 275–288. [Google Scholar] [CrossRef]
  13. Allen, R.W.; Adar, S.D.; Avol, E.; Martin, C.; Curl, C.L.; Timothy, L.; Liu, L.J.S.; Lianne, S.; Kaufman, J.D. Modeling the residential infiltration of outdoor PM2.5 in the multi-ethnic study of atherosclerosis and air pollution (MESA air). Environ. Health Perspect. 2012, 120, 824–830. [Google Scholar] [CrossRef] [PubMed]
  14. Shi, S.; Chen, C.; Zhao, B. Modifications of exposure to ambient particulate matter: Tackling bias in using ambient concentration as surrogate with particle infiltration factor and ambient exposure factor. Environ. Pollut. 2016, 94, 1024–1029. [Google Scholar] [CrossRef] [PubMed]
  15. Ji, W.; Zhao, B. Contribution of outdoor-originating particles, indoor-emitted particles and indoor secondary organic aerosol (SOA) to residential indoor PM2.5 concentration: A model-based estimation. Build. Environ. 2015, 90, 196–205. [Google Scholar] [CrossRef]
  16. Duan, J.; Tan, J. Atmospheric heavy metals and Arsenic in China: Situation, sources and control policies. Atmos. Environ. 2013, 74, 93–101. [Google Scholar] [CrossRef]
  17. Timmermans, R.; Kranenburg, R.; Manders, A.; Hendriks, C.; Segers, A.; Dammers, E.; Zhang, Q.; Wang, L.; Liu, Z.; Zeng, L. Source apportionment of PM2.5 across China using LOTOS-EUROS. Atmos. Environ. 2017, 164, 370–386. [Google Scholar] [CrossRef]
  18. Lv, B.; Zhang, B.; Bai, Y. A systematic analysis of PM2.5, in Beijing and its sources from 2000 to 2012. Atmos. Environ. 2016, 124, 98–108. [Google Scholar] [CrossRef]
  19. Habre, R.; Coull, B.; Moshier, E.; Godbold, J.; Grunin, A.; Nath, A.; Castro, W.; Schachter, N.; Rohr, A.; Kattan, M. Sources of indoor air pollution in New York City residences of asthmatic children. J. Expo. Sci. Environ. Epidemiol. 2014, 24, 269–278. [Google Scholar] [CrossRef] [PubMed]
  20. Massey, D.; Kulshrestha, A.; Masih, J.; Taneja, A. 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. 2012, 47, 223–231. [Google Scholar]
  21. Baxter, L.K.; Clougherty, J.E.; Paciorek, C.J.; Wright, R.J.; Levy, J.I. Predicting residential indoor concentrations of nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire and geographic information system based data. Atmos. Environ. 2007, 41, 6561–6571. [Google Scholar] [CrossRef] [PubMed]
  22. Shao, Z.; Bi, J.; Ma, Z.; Wang, J. Seasonal trends of indoor fine particulate matter and its determinants in urban residences in Nanjing, China. Build. Environ. 2017, 125, 319–325. [Google Scholar] [CrossRef]
  23. Pagel, É.C.; Costa Reis, N., Jr.; de Alvarez, C.E.; Santos, J.M.; Conti, M.M.; Boldrini, R.S.; Kerr, A.S. Characterization of the indoor particles and their sources in an Antarctic research station. Environ. Monit. Assess. 2016, 188, 167. [Google Scholar] [CrossRef] [PubMed]
  24. Jimenez-Velez, B.D.; Gioda, A.; PeréZ, U.; Rosa, Z. Particulate matter (PM10 and PM2.5) from different areas of Puerto Rico. Fresenius Environ. Bull. 2007, 16, 861–868. [Google Scholar]
  25. Eeling, O.; Mustaffa, N.I.; Amil, N.; Khan, M.F.; Latif, M.T. Source contribution of PM2.5 at different locations on the Malaysian Peninsula. Bull. Environ. Contam. Toxicol. 2015, 94, 537–542. [Google Scholar] [CrossRef] [PubMed]
  26. Mohammed, G.; Karani, G.; Mitchell, D. Trace Elemental Composition in PM10 and PM2.5 Collected in Cardiff, Wales. Energy Procedia 2017, 111, 540–547. [Google Scholar] [CrossRef]
  27. Harrison, R.M.; Yin, J.X. Chemical speciation of PM2.5 particles at urban background and rural sites in the UK atmosphere. J. Environ. Monit. 2010, 12, 1404–1414. [Google Scholar] [CrossRef] [PubMed]
  28. Zhai, Y.; Liu, X.; Chen, H.; Xu, B.; Zhu, L.; Li, C.; Zeng, G. Source identification and potential ecological risk assessment of heavy metals in PM2.5 from Changsha. Sci. Total Environ. 2014, 493, 109–115. [Google Scholar] [CrossRef] [PubMed]
  29. Zheng, M.; Zhang, Y.J.; Yan, C.Q.; Zhu, X.; Schauer, J.J.; Zhang, Y. Review of PM2.5 source apportionment methods in China. Acta Sci. Nat. Univ. Pekin. 2014, 6, 1141–1154. (In Chinese) [Google Scholar]
  30. Gao, J.; Wang, K.; Wang, Y.; Liu, S.; Zhu, C.; Hao, J.; Liu, H.; Hua, S.; Tian, H. Temporal-spatial characteristics and source apportionment of PM2.5 as well as its associated chemical species in the Beijing-Tianjin-Hebei region of China. Environ. Pollut. 2017, 233, 714–724. [Google Scholar] [CrossRef] [PubMed]
  31. Lin, Y.C.; Hsu, S.C.; Chou, C.K.; Zhang, R.; Wu, Y.; Kao, S.J.; Luo, L.; Huang, C.H.; Lin, S.H.; Huang, Y.T. Wintertime haze deterioration in Beijing by industrial pollution deduced from trace metal fingerprints and enhanced health risk by heavy metals. Environ. Pollut. 2016, 208, 284–293. [Google Scholar] [CrossRef] [PubMed]
  32. Tan, J.H.; Duan, J.C.; Ma, Y.L.; Yang, F.M.; Cheng, Y.; He, K.B.; Yu, Y.C.; Wang, J.W. Source of atmospheric heavy metals in winter in Foshan, China. Sci. Total Environ. 2014, 493, 262–270. [Google Scholar] [CrossRef] [PubMed]
  33. Duan, J.; Tan, J.; Wang, S.; Hao, J.; Chai, F. Size distributions and sources of elements in particulate matter at curbside, urban and rural sites in Beijing. J. Environ. Sci.-China 2012, 24, 87–94. [Google Scholar] [CrossRef]
  34. Guo, G.H.; Lei, M.; Chen, T.B.; Song, B.; Li, X.Y. Effect of road traffic on heavy metals in road dusts and roadside soils. Acta Sci. Circumst. 2008, 28, 1937–1945. (In Chinese) [Google Scholar]
  35. Zheng, Y.; Teng, Y. Emission characteristics of vanadium in air. Environ. Sci. Manag. 2012, 37, 20–24. (In Chinese) [Google Scholar]
  36. Tian, H.; Cheng, K.; Wang, Y.; Zhao, D.; Lu, L.; Jia, W.; Hao, J. Temporal and spatial variation characteristics of atmospheric emissions of Cd, Cr, and Pb from coal in China. Atmos. Environ. 2012, 50, 157–163. [Google Scholar] [CrossRef]
  37. Bernhard, D.; Rossmann, A.; Wick, G. Metals in cigarette smoke. IUBMB Life 2005, 57, 805–809. [Google Scholar] [CrossRef] [PubMed]
  38. Chiba, M.; Masironi, R. Toxic and trace elements in tobacco and tobacco smoke. Bull. World Health Organ. 1992, 70, 269–275. [Google Scholar] [PubMed]
  39. Armendáriz, C.R.; Garcia, T.; Soler, A.; Gutiérrez Fernández, Á.J.; Glez-Weller, D.; Luis González, G.; de la Torre, A.H.; Revert Gironés, C. Heavy metals in cigarettes for sale in Spain. Environ. Res. 2015, 143, 162–169. [Google Scholar] [CrossRef] [PubMed]
  40. Smith, C.J.; Perfetti, T.A.; Garg, R.C. IARC carcinogens reported in cigarette mainstream smoke and their calculated log P values. Food Chem. Toxicol. 2003, 41, 807–817. [Google Scholar] [CrossRef]
  41. Wang, G.; Cheng, S.; Wei, W.; Wen, W.; Wang, X.; Yao, S. Chemical characteristics of fine particles emitted from different Chinese cooking styles. Aerosol Air Qual. Res. 2015, 15, 2357–2366. [Google Scholar] [CrossRef]
  42. Abdullahi, K.L.; Delgado-Saborit, J.M.; Harrison, R.M. Emissions and indoor concentrations of particulate matter and its specific chemical components from cooking: A review. Atmos. Environ. 2013, 71, 260–294. [Google Scholar] [CrossRef]
  43. Zhang, N.; Han, B.; He, F.; Xu, J.; Zhao, R.; Zhang, Y.; Bai, Z. Chemical characteristic of PM2.5 emission and inhalational carcinogenic risk of domestic Chinese cooking. Environ. Pollut. 2017, 227, 24–30. [Google Scholar] [CrossRef] [PubMed]
  44. Chen, L.C.; Lippmann, M. Effects of metals within ambient air particulate matter (PM) on human health. Inhal. Toxicol. 2009, 21, 1–31. [Google Scholar] [CrossRef] [PubMed]
  45. Zeng, X.; Xu, X.; Zheng, X.; Reponen, T.; Chen, A.; Huo, X. Heavy metals in PM2.5 and in blood, and children’s respiratory symptoms and asthma from an e-waste recycling area. Environ. Pollut. 2016, 210, 346–353. [Google Scholar] [CrossRef] [PubMed]
  46. Liu, S.; Ganduglia, C.M.; Li, X.; Delclos, G.L.; Franzini, L.; Zhang, K. Short-term associations of fine particulate matter components and emergency hospital admissions among a privately insured population in Greater Houston. Atmos. Environ. 2016, 147, 369–375. [Google Scholar] [CrossRef]
  47. Hou, W.; Xu, X.; Lei, Y.; Cao, J.; Zhang, Y.; Chen, L.; Huo, X. The role of the PM2.5-associated metals in pathogenesis of child Mycoplasma Pneumoniae infections: A systematic review. Environ. Sci. Pollut. Res. 2016, 23, 10604–10614. [Google Scholar] [CrossRef] [PubMed]
  48. Tchounwou, P.B.; Yedjou, C.G.; Patlolla, A.K.; Sutton, D.J. Heavy Metal Toxicity and the Environment. EXS 2012, 101, 133–164. [Google Scholar] [PubMed]
  49. Åkesson, A.; Barregard, L.; Bergdahl, I.A.; Nordberg, G.F.; Nordberg, M.; Skerfving, S. Non-renal effects and the risk assessment of environmental cadmium exposure. Environ. Health Perspect. 2014, 122, 431–438. [Google Scholar] [CrossRef] [PubMed]
  50. Johri, N.; Jacquillet, G.; Unwin, R. Heavy metal poisoning: The effects of cadmium on the kidney. Biometals 2010, 23, 783–792. [Google Scholar] [CrossRef] [PubMed]
  51. Wallin, M.; Barregard, L.; Sallsten, G.; Lundh, T.; Karlsson, M.K.; Lorentzon, M.; Ohlsson, C.; Dan, M. Low-Level Cadmium Exposure Is Associated with Decreased Bone Mineral Density and Increased Risk of Incident Fractures in Elderly Men: The MrOS Sweden Study. J. Bone Miner. Res. 2016, 31, 732–741. [Google Scholar] [CrossRef] [PubMed]
  52. Caito, S.; Aschner, M. Developmental Neurotoxicity of Lead. In Neurotoxicity of Metals; Springer: Cham, Switzerland, 2017; pp. 3–12. [Google Scholar]
  53. Mishra, K.P. Lead exposure and its impact on immune system: A review. Toxicol. In Vitro 2009, 23, 969–972. [Google Scholar] [CrossRef] [PubMed]
  54. Radulescu, C.; Iordache, S.; Dunea, D.; Stihi, C.; Dulama, I.D. Risks Assessment of Heavy Metals on Public Health Associated with Atmospheric Exposure to PM2.5 in Urban Area. Roma J. Phys. 2015, 60, 1171–1182. [Google Scholar]
  55. Adgate, J.L.; Mongin, S.J.; Pratt, G.C.; Zhang, J.; Field, M.P. Relationships between personal, indoor, and outdoor exposures to trace elements in PM2.5. Sci. Total Environ. 2007, 386, 21–32. [Google Scholar] [CrossRef] [PubMed]
  56. Clark, N.A.; Allen, R.W.; Hystad, P.; Wallace, L.; Dell, S.D.; Foty, R.; Dabekzlotorzynska, E.; Evans, G.; Wheeler, A.J. Exploring variation and predictors of residential fine particulate matter infiltration. Int. J. Environ. Res. Public Health 2010, 7, 3211–3224. [Google Scholar] [CrossRef] [PubMed]
  57. Meng, Q.Y.; Spector, D.; Colome, S.; Turpin, B. Determinants of Indoor and Personal Exposure to PM2.5 of Indoor and Outdoor Origin during the RIOPA Study. Atmos. Environ. 2009, 43, 5750–5758. [Google Scholar] [CrossRef] [PubMed]
  58. Xu, C.; Na, L.; Yang, Y.; Li, Y.; Zhe, L.; Qin, W.; Zheng, T.; Civitarese, A.; Xu, D. Investigation and modeling of the residential infiltration of fine particulate matter in Beijing, China. J. Air Waste Manag. 2017, 67, 694–701. [Google Scholar] [CrossRef] [PubMed]
  59. Qi, M.; Zhu, X.; Du, W.; Chen, Y.; Chen, Y.; Huang, T.; Pan, X.; Zhong, Q.; Sun, X.; Zeng, E.Y. Exposure and health impact evaluation based on simultaneous measurement of indoor and ambient PM2.5, in Haidian, Beijing. Environ. Pollut. 2017, 220, 704–712. [Google Scholar] [CrossRef] [PubMed]
  60. Han, Y.; Qi, M.; Chen, Y.; Shen, H.; Liu, J.; Huang, Y.; Chen, H.; Liu, W.; Wang, X.; Liu, J. Influences of ambient air PM2.5, concentration and meteorological condition on the indoor PM2.5, concentrations in a residential apartment in Beijing using a new approach. Environ. Pollut. 2015, 205, 307–314. [Google Scholar] [CrossRef] [PubMed]
  61. Huang, L.; Pu, Z.; Li, M.; Sundell, J. Characterizing the Indoor-Outdoor Relationship of Fine Particulate Matter in Non-Heating Season for Urban Residences in Beijing. PLoS ONE 2015, 10, e0138559. [Google Scholar] [CrossRef] [PubMed]
  62. Deng, G.; Li, Z.; Wang, Z.; Gao, J.; Xu, Z.; Li, J.; Wang, Z. Indoor/outdoor relationship of PM2.5 concentration in typical buildings with and without air cleaning in Beijing. Indoor Built Environ. 2015, 26, 60–68. [Google Scholar] [CrossRef]
  63. Zhao, L.; Chen, C.; Wang, P.; Chen, Z.; Cao, S.; Wang, Q.; Xie, G.; Wan, Y.; Wang, Y.; Lu, B. Influence of atmospheric fine particulate matter (PM2.5) pollution on indoor environment during winter in Beijing. Build. Environ. 2015, 87, 283–291. [Google Scholar] [CrossRef]
  64. Das, P.; Shrubsole, C.; Jones, B.; Hamilton, I.; Chalabi, Z.; Davies, M.; Mavrogianni, A.; Taylor, J. Using probabilistic sampling-based sensitivity analyses for indoor air quality modelling. Build. Environ. 2014, 78, 171–182. [Google Scholar] [CrossRef]
  65. Pekey, B.; Bozkurt, Z.B.; Pekey, H.; Doğan, G.; Zararsız, A.; Efe, N.; Tuncel, G. Indoor/outdoor concentrations and elemental composition of PM10/PM2.5 in urban/industrial areas of Kocaeli City, Turkey. Indoor Air 2010, 20, 112–125. [Google Scholar] [CrossRef] [PubMed]
  66. Buczyńska, A.J.; Krata, A.; Van Grieken, R.; Brown, A.; Polezer, G.; De Wael, K.; Potgieter-Vermaak, S. Composition of PM2.5 and PM1 on high and low pollution event days and its relation to indoor air quality in a home for the elderly. Sci. Total Environ. 2014, 490, 134–143. [Google Scholar] [CrossRef] [PubMed]
  67. Hassanvand, M.S.; Naddafi, K.; Faridi, S.; Nabizadeh, R.; Sowlat, M.H.; Momeniha, F.; Gholampour, A.; Arhami, M.; Kashani, H.; Zare, A. Characterization of PAHs and metals in indoor/outdoor PM10/PM2.5/PM1 in a retirement home and a school dormitory. Sci. Total Environ. 2015, 527, 100–110. [Google Scholar] [CrossRef] [PubMed]
  68. Chao, C.Y.; Wong, K.K. Residential indoor PM10 and PM2.5 in Hong Kong and the elemental composition. Atmos. Environ. 2002, 36, 265–277. [Google Scholar] [CrossRef]
  69. Zhang, Y.; Cao, S.; Xu, X.; Qiu, J.; Chen, M.; Wang, D.; Guan, D.; Wang, C.; Wang, X.; Dong, B. Metals compositions of indoor PM2.5, health risk assessment, and birth outcomes in Lanzhou, China. Environ. Monit. Assess. 2016, 188, 325. [Google Scholar] [CrossRef] [PubMed]
  70. Wang, X.; Wei, W.; Cheng, S.; Li, J.; Zhang, H.; Lv, Z. Characteristics and classification of PM2.5 pollution episodes in Beijing from 2013 to 2015. Sci. Total Environ. 2018, 612, 170–179. [Google Scholar] [CrossRef] [PubMed]
  71. Chen, Y.; Schleicher, N.; Cen, K.; Liu, X.; Yu, Y.; Zibat, V.; Dietze, V.; Fricker, M.; Kaminski, U.; Chen, Y. Evaluation of impact factors on PM2.5 based on long-term chemical components analyses in the megacity Beijing, China. Chemosphere 2016, 155, 234–242. [Google Scholar] [CrossRef] [PubMed]
  72. Ministry of Environmental Protection. Source Apportionment of PM2.5 in Beijing. 2017. Available online: http://dqhj.mep.gov.cn/dqhjzl/dqklwyjx/201709/t20170915_421691.shtml (accessed on 15 September 2017). [Google Scholar]
  73. Yuan, S.; Xu, W.; Liu, Z. A Study on the Model for Heating Influence on PM2.5 Emission in Beijing China. Procedia Eng. 2015, 121, 612–620. [Google Scholar] [CrossRef]
  74. Tao, J.; Zhao, Z.; Sheng, X.; Wang, C.M.; Tian, P.Y. Concentration level and distribution characters of rare earth elements in indoor air PM2.5 in Beijing. Capital J. Public Health 2016, 10, 202–205. (In Chinese) [Google Scholar]
  75. Tian, P.Y.; Zhao, Z.; Tao, J. Determination of 24 metals in indoor air PM2.5 by ICP-MS. Chin. J. Public Health 2016, 32, 1580–1583. (In Chinese) [Google Scholar]
  76. Han, Y.; Li, X.; Zhu, T.; Lv, D.; Chen, Y.; Hou, L.A.; Zhang, Y.; Ren, M. Characteristics and Relationships between Indoor and Outdoor PM2.5 in Beijing: A Residential Apartment Case Study. Aerosol. Air Qual. Res. 2016, 16, 2386–2395. [Google Scholar] [CrossRef]
  77. Huang, F.; Li, X.; Wang, C.; Xu, Q.; Wang, W.; Luo, Y.; Tao, L.; Gao, Q.; Guo, J.; Chen, S. PM2.5 spatiotemporal variations and the relationship with meteorological factors during 2013–2014 in Beijing, China. PLoS ONE 2015, 10, e0141642. [Google Scholar] [CrossRef] [PubMed]
  78. Zhou, B.; Shen, H.; Huang, Y.; Li, W.; Chen, H.; Zhang, Y.; Su, S.; Chen, Y.; Lin, N.; Zhuo, S. Daily variations of size-segregated ambient particulate matter in Beijing. Environ. Pollut. 2015, 197, 36–42. [Google Scholar] [CrossRef] [PubMed]
  79. Barría, R.M.; Calvo, M.; Pino, P. Indoor air pollution by fine particulate matter in the homes of newborns. Rev. Chil. Pediatr. 2016, 87, 343–350. [Google Scholar] [CrossRef] [PubMed]
  80. Howard-Reed, C.; Wallace, L.A.; Emmerich, S.J. Effect of ventilation systems and air filters on decay rates of particles produced by indoor sources in an occupied townhouse. Atmos. Environ. 2003, 37, 5295–5306. [Google Scholar] [CrossRef]
  81. Persily, A.; Musser, A.; Emmerich, S.J. Modeled infiltration rate distributions for US housing. Indoor Air 2010, 20, 473–485. [Google Scholar] [CrossRef] [PubMed]
  82. Macneill, M.; Wallace, L.; Kearney, J.; Allen, R.W.; Ryswyk, K.V.; Judek, S.; Xu, X.; Wheeler, A. Factors influencing variability in the infiltration of PM2.5 mass and its components. Atmos. Environ. 2012, 61, 518–532. [Google Scholar] [CrossRef]
  83. Isaxon, C.; Gudmundsson, A.; Nordin, E.Z.; Lönnblad, L.; Dahl, A.; Wieslander, G.; Bohgard, M.; Wierzbicka, A. Contribution of indoor-generated particles to residential exposure. Atmos. Environ. 2015, 106, 458–466. [Google Scholar] [CrossRef]
  84. Gao, X.; Yu, Q.; Gu, Q.; Chen, Y.; Ding, K.; Zhu, J.; Chen, L. Indoor air pollution from solid biomass fuels combustion in rural agricultural area of Tibet, China. Indoor Air 2009, 19, 198–205. [Google Scholar] [CrossRef] [PubMed]
  85. Qian, J.; Peccia, J.; Ferro, A.R. Walking-induced particle resuspension in indoor environments. Atmos. Environ. 2014, 89, 464–481. [Google Scholar] [CrossRef]
  86. De, G.G.; Dambruoso, P.R.; Di, G.A.; Di, P.V.; Marzocca, A.; Tutino, M. Discontinuous and Continuous Indoor Air Quality Monitoring in Homes with Fireplaces or Wood Stoves as Heating System. Int. J. Environ. Res. Public Health 2016, 13, 78. [Google Scholar]
Table 1. Household characteristics of residential houses in this study.
Table 1. Household characteristics of residential houses in this study.
Household CharacteristicsNHSHS
NN
Number of total houses4747
Building type
 apartments2826
 villas55
 courtyards1416
Building age
 <10 years106
 10–20 years1215
 >20 years2526
Cooking fuel type
 No cooking34
 Liquefied gas1211
 Coal gas04
 Nature gas2926
 Electric32
Presence of indoor cooking4140
Presence of indoor smoke1010
Table 2. Distribution of residential indoor and outdoor concentration of PM2.5 and elements (median (IQR: Interquartile Range)).
Table 2. Distribution of residential indoor and outdoor concentration of PM2.5 and elements (median (IQR: Interquartile Range)).
ConstituentAll Seasons (N = 94)NHS (N = 47)HS (N = 47)
IndoorOutdoorpd aIndoorOutdoorpd aIndoorOutdoorpd a
PM2.5 (μg/m3)57.64 (27.51)72.49 (43.53)<0.0155.44 (24.12)65.91 (28.90)<0.0158.23 (13.28)91.85 (26.61)<0.01
Al (ng/m3)400.34 (274.64)551.71 (254.69)<0.01454.95 (204.34)540.15 (222.43)<0.01306.00 (61.76)561.01 (94.42)<0.01
As (ng/m3)a9.83 (30.75)12.16 (31.62)0.116.00 (21.78)0.16 (12.18)0.1718.34 (17.02)25.81 (16.40)<0.01
Ca (ng/m3)302.85 (247.21)433.87 (319.64)<0.01288.61 (289.73)322.95 (132.04)0.85322.48 (97.38)506.56 (68.86)<0.01
Cd (ng/m3)1.84 (1.19)1.93 (1.27)<0.011.93 (1.04)1.86 (1.07)0.011.77 (0.62)2.11 (0.68)0.01
Cu (ng/m3)21.28 (12.57)35.38 (19.21)<0.0121.37 (10.54)24.89 (17.13)<0.0120.44 (5.65)38.15 (6.94)<0.01
Fe (ng/m3)675.85 (604.24)1260.48 (997.58)<0.01758.80 (582.96)1260.48 (901.70)0.05610.08 (236.62)1301.99 (486.36)<0.01
K (ng/m3)930.24 (505.95)1176.27 (838.15)<0.01892.17 (549.59)1077.51 (632.11)<0.011011.85 (181.24)1348.57 (247.78)<0.01
Mg (ng/m3)163.35 (98.45)225.21 (130.50)<0.01163.35 (118.95)198.89 (123.37)<0.01162.82 (41.75)277.67 (52.46)<0.01
Mn (ng/m3)36.79 (20.54)54.44 (28.53)<0.0139.07 (23.49)52.67 (20.38)<0.0130.92 (5.87)71.21 (17.69)<0.01
Na (ng/m3)439.63 (273.69)566.85 (274.72)<0.01388.72 (290.57)455.47 (234.50)0.03460.08 (91.72)700.37 (181.91)<0.01
Pb (ng/m3)105.28 (64.08)145.01 (81.50)<0.01105.66 (52.73)140.52 (78.05)<0.01104.3 (36.06)162.23 (70.81)<0.01
Se (ng/m3)a3.29 (3.34)5.35 (3.14)<0.014.15 (3.97)5.46 (3.34)<0.012.14 (0.75)5.09 (2.30)<0.01
Tl (ng/m3)0.92 (0.66)1.19 (1.17)<0.011.11 (0.65)1.31 (1.30)<0.010.74 (0.25)1.11 (0.28)<0.01
V (ng/m3)a1.63 (1.77)2.40 (2.37)<0.011.71 (1.44)2.43 (1.51)<0.011.31 (0.80)2.40 (0.55)<0.01
Zn (ng/m3)217.1 (133.31)270.55 (159.02)<0.01229.3 (118.85)270.55 (153.55)<0.01207.01 (67.30)275.63 (62.21)<0.01
Note: pd is the p-value of Wilcoxon signed-rank test for comparing indoor and outdoor. There are some houses with concentrations of V, Se and As below the detection limit.
Table 3. Comparison of indoor and outdoor concentrations of PM2.5 and elements in smoking and non-smoking houses (median).
Table 3. Comparison of indoor and outdoor concentrations of PM2.5 and elements in smoking and non-smoking houses (median).
ConstituentSmoking in NHS (N = 10)Non-Smoking in NHS (N = 37)pNHS aSmoking in HS (N = 10)Non-Smoking in HS (N = 37)pHS a
IndoorOutdoorpdIndoorOutdoorpdIndoorOutdoorpdIndoorOutdoorpd
PM2.5 (μg/m3)66.9259.010.9653.5170.62<0.010.13129.3472.040.3354.6091.85<0.01<0.01
Al (ng/m3)479.09731.310.11444.88532.21<0.010.86317.14551.530.28306.00561.01<0.010.78
As (ng/m3)17.050.130.022.971.210.990.2132.5831.370.1116.3722.370.010.29
Ca (ng/m3)459.64284.330.11272.28370.990.340.15220.36459.350.07335.5630.97<0.010.06
Cd (ng/m3)2.641.340.011.802.10<0.010.023.931.550.011.342.20<0.01<0.01
Cu (ng/m3)18.6123.060.1722.7630.31<0.010.1225.2638.980.1120.4438.15<0.010.32
Fe (ng/m3)990.321578.390.28726.54905.300.110.34673.13785.050.96575.631851.86<0.010.47
K (ng/m3)923.04846.530.14892.171157.92<0.010.361277.481193.320.88962.131391.27<0.010.07
Mg (ng/m3)190.01151.400.51160.41201.22<0.010.92149.72215.950.04170.68279.87<0.010.17
Mn (ng/m3)40.3848.690.2039.0753.26<0.010.928.6642.600.2031.0971.61<0.010.79
Na (ng/m3)428.54376.370.20388.72472.35<0.010.47492.05593.780.96455.51700.67<0.010.22
Pb (ng/m3)112.39114.910.33105.66140.52<0.010.7085.1798.720.01107.92163.95<0.010.18
Se (ng/m3)5.214.210.883.865.54<0.010.111.443.910.022.485.09<0.010.36
Tl (ng/m3)1.151.120.171.111.31<0.010.790.740.880.110.721.16<0.010.90
V (ng/m3)2.102.430.011.712.43<0.010.661.802.090.441.222.40<0.010.21
Zn (ng/m3)254.30189.750.39229.30271.48<0.010.47144.57185.220.06218.21333.030.010.07
Note: pNHS is the p-value of Mann-Whitney U test for comparing smoking and non-smoking houses in non-heating season, pHS is the p-value for comparing smoking and non-smoking houses in heating season.
Table 4. Comparison of indoor and outdoor concentrations of PM2.5 and elements for houses with window close in HS (median).
Table 4. Comparison of indoor and outdoor concentrations of PM2.5 and elements for houses with window close in HS (median).
ConstituentWindows Close (N = 19)Windows Close and Non-Smoking (N = 14)
IndoorOutdoorpdIndoorOutdoorpd
PM2.5 (μg/m3)69.5384.270.4061.7784.270.04
Al (ng/m3)312.72517.090.09306482.460.29
As (ng/m3)7.8918.780.146.1614.80.04
Ca (ng/m3)226.63517.790.17241.69624.560.35
Cd (ng/m3)1.281.920.181.201.92<0.01
Cu (ng/m3)18.4538.15<0.0116.5636.86<0.01
Fe (ng/m3)622.171851.860.02610.081988.55<0.01
K (ng/m3)928.341320.130.01896.541255<0.01
Mg (ng/m3)130.37275.890.01122.18275.890.02
Mn (ng/m3)31.5571.21<0.0129.1972.78<0.01
Na (ng/m3)428.01669.730.06369.75700.37<0.01
Pb (ng/m3)105.32163.95<0.01102.75163.09<0.01
Se (ng/m3)1.474.330.011.463.960.04
Tl (ng/m3)0.691.04<0.010.521.02<0.01
V (ng/m3)2.185.04<0.012.045.16<0.01
Zn (ng/m3)175.14246.720.16183.91299.40.48
Table 5. Comparison of indoor and outdoor concentrations of PM2.5 and elements for houses that sampled in both seasons (median, N = 37).
Table 5. Comparison of indoor and outdoor concentrations of PM2.5 and elements for houses that sampled in both seasons (median, N = 37).
ConstituentNHSHSpin apout a
IndoorOutdoorpdIndoorOutdoorpd
PM2.5 (μg/m3)53.5170.62<0.0154.5491.85<0.010.710.02
Al (ng/m3)457.83609.78<0.01299.88561.01<0.010.360.51
As (ng/m3)2.990.160.4126.0731.370.010.04<0.01
Ca (ng/m3)288.61370.990.51319.11591.58<0.010.84<0.01
Cd (ng/m3)1.931.860.031.822.11<0.010.560.82
Cu (ng/m3)22.5629.50<0.0119.4738.15<0.010.32<0.01
Fe (ng/m3)749.381134.840.04610.081132.16<0.010.090.12
K (ng/m3)911.351157.92<0.01976.561391.27<0.010.510.01
Mg (ng/m3)166.34201.22<0.01162.82279.87<0.010.57<0.01
Mn (ng/m3)40.9052.67<0.0129.1971.21<0.010.040.03
Na (ng/m3)398.56511.570.01460.08700.37<0.010.44<0.01
Pb (ng/m3)105.66140.52<0.0193.16147<0.010.460.10
Se (ng/m3)4.155.46<0.012.485.09<0.01<0.010.20
Tl (ng/m3)1.161.36<0.010.741.11<0.01<0.010.21
V (ng/m3)1.712.43<0.010.772.15<0.01<0.010.24
Zn (ng/m3)229.3271.48<0.01218.21275.12<0.010.380.68
Note: pin is the p-value of Wilcoxon signed-rank test for comparing indoor concentration of the same houses in two seasons. pout is the p-value for comparing outdoor concentration of the same houses in two seasons.
Table 6. Factor analysis for indoor PM2.5 (all seasons).
Table 6. Factor analysis for indoor PM2.5 (all seasons).
ElementsF1F2F3F4F5Communality
Al 0.71 0.6
As 0.830.83
Ca 0.92 0.86
Cd0.72 0.560.88
Cu0.600.53 0.66
Fe0.89 0.86
K 0.72 0.82
Mg 0.540.61 0.74
Mn0.740.45 0.90
Na 0.74 0.73
Pb 0.78 0.72
Se 0.78 0.71
Tl 0.82 0.68
V0.78 0.75
Zn 0.86 0.78
Eigenvalue5.022.321.981.121.07
% of variance21.5521.2813.8011.378.79
Cumulative %21.5542.8256.6268.0076.78
Possible source typeCombustionMotor vehicles and combustionIndoor cookingDust and soilIndoor smoking
Table 7. Multiple linear regression of indoor PM2.5 and selected elements and their potential influencing variables.
Table 7. Multiple linear regression of indoor PM2.5 and selected elements and their potential influencing variables.
VariableCoefficient at ap a
β aSE aβ’ a
PM2.5 in NHS (R2 = 0.60)
Intercept1.1060.122 9.102<0.001
Outdoor PM2.50.0060.0010.8686.296<0.001
Smoking during sampling0.1280.0340.3863.7260.001
Outdoor temperature 0.0100.0030.4403.0570.004
Time of air conditioner use during sampling −0.0020.001−0.269−2.2510.030
PM2.5 in HS (R2 = 0.46)
Intercept1.4820.065-22.712<0.001
Outdoor PM2.50.0030.0010.4624.167<0.001
Smoking during sampling0.2660.0630.4684.225<0.001
Cd in NHS (R2 = 0.51)
Intercept−0.2640.087-−3.0470.004
Smoking during sampling0.3500.0660.6035.344<0.001
Outdoor Cd0.2050.0360.6425.692<0.001
Cd in HS (R2 = 0.63)
Intercept−1.2280.225-−5.446<0.001
Smoking during sampling0.7130.1200.5625.957<0.001
Outdoor Cd0.3210.0500.6056.379<0.001
Building age0.2190.0700.2983.1450.003
Pb in NHS (R2 = 0.63)
Intercept1.4210.069 20.673<0.001
Outdoor Pb0.0040.0010.7398.003<0.001
Time of window open per day0.0060.0020.2472.6810.010
Pb in HS (R2 = 0.74)
Intercept1.3570.124-10.973<0.001
Outdoor Pb0.0040.0011.0099.252<0.001
Building age0.0990.0300.2603.3200.002
RH−0.0080.004−0.232−2.1430.038
Note: β is the partial regression coefficient for each independent variable, SE is the standard error of β, β’ is the standard regression coefficient, t is the t-value of the regression coefficient significance test (t-test) and P is the p-value of t-test.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Int. J. Environ. Res. Public Health EISSN 1660-4601 Published by MDPI AG, Basel, Switzerland RSS E-Mail Table of Contents Alert
Back to Top