Background

Phthalate esters (PAEs) are a class of manufactured organic chemicals that have been used as plasticizers/additives, emollients, antifoaming agents, humectants, or carriers in various industrial and consumer products [1,2,3,4]. Phthalate esters were dominated plasticizers, accounting for 70–80% of the market share of plastic plasticizers, and the domestic PVC production capacity was 22.82 million tons in 2017 [5]. From January to September 2019, the total output of national plastic products industry enterprises has been reached 59.94 million tons [6].

The employment of various phthalates differs widely with chemical and physical properties [7]. Low molecular weight PAEs, such as dimethyl phthalate (DMP), diethyl phthalate (DEP), and di-n-butyl phthalate (DBP), are mainly used as solvents or carriers in personal care products and coatings. DBP is additionally used in cellulose esters, epoxy resins, and special adhesive formulations [1]. Meanwhile, PAEs with long/branching alkyl chain, such as di(2-ethylhexyl) phthalate (DEHP), di-n-octyl phthalate (DnOP), and benzyl butyl phthalate (BBzP), are mainly used as plasticizers in the polymers to improve flexibility, workability, and general handling properties [1, 8].

Most PAEs are physically mixed to the polymeric materials, thereby facilitating their easy entrance to the environment during their manufacture, usage, or disposal [4, 9]. The widespread application of products containing PAEs in daily life resulted in their universality in environments and frequent detection in various environmental matrices, including drinking water [10], surface water [11], indoor/outdoor air [12,13,14], dust [7, 15], sediment [16, 17], soil [18, 19], food [20], and urine [21, 22]. Meanwhile, some PAEs [e.g., DEHP, DBP, BBzP, DEP, and di-n-hexyl phosphoric acid (DHXP)] have been proven to be endocrine-disrupting compounds, and exposure to phthalates may result in reproductive effects [23], hypertension, and childhood obesity problems [8, 24,25,26]. Six PAEs, including DMP, DEP, BBzP, DBP, DEHP, and DnOP, have been listed as priority pollutants by the USEPA and the European Union due to their widespread occurrence in the biosphere and potential adverse effects on the environment and humans [27, 28, 29, 29]. DEP, DBP, and DnOP have also been identified as priority pollutants by the National Environmental Monitoring Center in China [30].

Persistent organic pollutants (POPs) tend to accumulate in the dust and soils [31]. Humans spend most of their time indoors; thus, indoor dust is a considerable daily exposure source of PAEs for humans. To date, an increasing number of studies reported the contamination of PAEs in indoor dust in different building environments. However, limited information of PAEs in dust is available. For instance, a few studies investigated the PAEs in different types of indoor dust in the USA [32], Germany [33, 34], Denmark [35], Norway [35, 36], Sweden [9, 37], and Japan [38]. As one of the largest PAE producers and users in the world, China has also conducted a few surveys on the occurrence of PAEs in indoor dusts. However, these surveys are limited to localized regions, such as Xi’an [15], Beijing [3], Nanjing [39], Xinjiang [40], and the Pearl River Delta [41]. Children are physiologically more vulnerable to environmental pollution when compared with adults because children are considerably exposed due to their frequent skin contact with PAE-containing products [42]. Therefore, assessing the health risk from exposure to PAEs in residential dust in China is urgent. Till date, only one study reported a nationwide survey on the characteristics of PAEs in indoor dust collected from bedrooms or saloons in China. However, this study only estimated the daily intake (DI) via various routes, and the exposure risks were not assessed [43]. In this study, eight PAEs were analyzed in 94 indoor dust samples collected across China. This study aims to investigate the concentration, spatial distribution characteristics, sources, and health risks to human (including noncancer and carcinogenic risks) exposure to PAEs in indoor dusts from residences and dormitories.

Materials and methods

Standards

Eight PAE standards (DMP, DBP, DEP, DIBP, BBzP, DEHP, DnOP, and dinonyl phthalate (DNP)) were obtained from AccuStand (New Haven, CT, USA), and their structures and selected physicochemical properties were summarized in Additional file 1: Table S1. The deuterated internal standards of diethyl phthalate-3,4,5,6-d4 (DMP-d4, 0.10 g·mL−1), di-n-butyl phthalate-d4 (DBP-d4), and bis (2-ethylhexyl) phthalate-3,4,5,6-d4 (DEHP-d4) were purchased from AccuStand. Acetone and ethyl acetate used were chromatographically pure and purchased from Burdick & Jackson.

Sample collection and preparation

The collection of 94 dust samples, including 72 bedroom or drawing room dust samples from houses and 22 dust samples from dormitories, was conducted from seven regions in China during August and October 2019. Detailed information on all sampling sites is presented in Additional file 1: Figure S1. Before sampling, the floor to be sampled is cleaned by owners. Dust samples were collected by gently sweeping the top of furniture and the floor with disposable brushes. All brushes used before sampling were precleaned with methanol. After sampling, the samples were wrapped with aluminum foils and stored in a polyethylene zip-lock bag. Then, the samples were transported to the laboratory within three days. Subsequently, dusts were sieved through a 200-mesh sieve, homogenized thoroughly, and then stored in a brown glass jar at 4 °C until analysis. All samples were analyzed within 30 d.

Sample analysis

Indoor dust samples (0.10 g, dry weight) were put into a 15 mL glass tube, and 200 ng internal standards were spiked to the samples. PAEs in dust samples were extracted with 2 mL ethyl acetate by mechanical oscillation at 180 rpm for 1 h and left at room temperature for 1 h. Approximately 1 mL of extract was transferred to a GC vial. The extraction method consulted the standard “Soil quality -Determination of selected phthalates using capillary gas chromatography with mass spectrometric detection (GC/MS)” (ISO 13913-2014) [44].

Instrumental analysis

The extracts were analyzed with a Shimadzu gas chromatography–mass spectrometry (GCMS-QP2010Plus, Shimadzu, Japan) equipped with an electron ionization (EI) source. Quantification of each target compound was performed in the single ion monitoring (SIM) mode. The analytes were separated using a capillary column (DB-5MS, 30 m × 0.25 mm × 0.25 μm). The flow rate of the ultra-high purity helium carrier gas (purity 99.999%) was maintained at 1.0 mL·min−1. The column temperature program started at 60 °C (held for 1.0 min), ramped to 220 °C at a rate of 20 °C·min−1 and held for 1.0 min, and finally increased at 5 °C·min−1 to 290 °C and held for 9.0 min. The ion source, injection port, and quadrupole, and transmission line temperatures were maintained at 300, 300, and 250 °C, respectively. The injection volume was 1 μL. Quantification was conducted by the external calibration method based on a six-point calibration curve for each compound.

Quality assurance and quality control (QA/QC)

All kinds of plastic containers were strictly avoided to reduce background contamination during the entire sample processing. All glassware was precleaned three times with acetone before use. The calibration curves were measured for all eight PAEs with concentrations ranging from 20 ng·mL−1 to 1000 ng·mL−1, and the regression coefficients of calibration curves were > 0.99. The LODs of PAEs ranged from 0.0008 μg·g−1 for DBP to 0.0048 μg·g−1 for DnOP. The LOQs for DMP, DEP, DIBP, DBP, BBzP, DEHP, DnOP, and DNP were 0.0038, 0.0098, 0.0027, 0.0024, 0.0128, 0.0036, 0.0143, and 0.0095 μg·g−1, respectively (Additional file 1: Table S2). Blank samples were processed with every batch of 20 samples. No target compound was detected in procedural blanks; therefore, the background concentration of these chemicals was neglected. The relative recoveries of target PAEs in spiked matrix samples ranged from 75.95 to 125.36%, while those of internal standards in spiked blank mats, spiked matrices, and dust samples were 80.34–131.61%, 76.54–116.86%, and 76.89–120.98%, respectively. The precision of the analytical method was evaluated by randomly selecting 10 samples analyzed in duplicate, with the coefficient variation below 15%.

Statistical analysis

Statistical analysis was conducted with IBM SPSS Statistics 25.0, EPA PMF 5.0, and Origin 2019. A spatial distribution map of PAEs was created using ArcGIS version 10.2. Concentrations lower than the LOQ were assigned a value of LOQ/2 for statistical analysis. Pearson’s correlation was employed to analyze correlations between PAE concentrations in various environmental matrices. PCA and PMF receptor models used in this study are widely used factor analysis receptor models which do not need source profiles. The explicit descriptions of these models can be found in our previous studies [45,46,47]. They are also provided in the supplementary information. Generally, they can be expressed by the following basic Eq. (1):

$${x}_{ik}=\sum_{j=1}^{p}{g}_{ip}{f}_{pk}+{e}_{ik}$$
(1)

where \({x}_{ik}\) is the concentration of ith species for the kth sample; \({f}_{pk}\) is the contribution of the pth source to the kth sample; \({g}_{ip}\) is the ith species concentration from the pth source; and \({e}_{ik}\) is the error. Factors extracted from the ambient concentrations can be linked to potential source categories.

Human exposure and health risk assessment

The formulas with slight modification from the US EPA [48], which have been widely employed in previous studies [27, 43, 49], are defined as Eqs. (24) to assess the average daily dosage (ADD, mg·mg·kg−1·d−1) of PAEs for adults and children via different exposure pathways of PAEs, namely ingestion, dermal absorption, and inhalation. ADDing, ADDder, and ADDinh are the ADD (mg·kg−1·d−1) through dust ingestion, dermal absorption, and inhalation, respectively.

$${\mathrm{ADD}}_{\mathrm{ing}}=\frac{{\mathrm{C}}_{\mathrm{dust}}\times {\mathrm{IR}}_{\mathrm{ingestion}}\times \mathrm{EF}\times \mathrm{ED}\times \mathrm{CF}}{\mathrm{BW}\times \mathrm{AT}}$$
(2)
$${\mathrm{ADD}}_{\mathrm{der}}=\frac{{\mathrm{C}}_{\mathrm{dust}}\times \mathrm{SA}\times \mathrm{AF}\times \mathrm{ABF}\times \mathrm{EF}\times \mathrm{ED}\times \mathrm{CF}}{\mathrm{BW}\times \mathrm{AT}}$$
(3)
$${\mathrm{ADD}}_{\mathrm{inh}}=\frac{{\mathrm{C}}_{\mathrm{dust}}\times {\mathrm{IR}}_{\mathrm{inhalation}}\times \mathrm{EF}\times \mathrm{ED}}{\mathrm{BW}\times \mathrm{AT}\times \mathrm{PEF}}$$
(4)

where \({\mathrm{C}}_{\mathrm{dust}}\) is the PAE concentration measured in dust (mg·kg−1). \({\mathrm{IR}}_{\mathrm{ingestion}}\) is the dust ingestion rate (mg·d−1). \({\mathrm{IR}}_{\mathrm{inhalation}}\) is the dust inhalation rate (m3·d−1). SA is the dermal exposure area (cm2). ABF is the dermal adsorption fraction. AF is the dermal adherence factor for dust (mg·cm−2·d−1). BW is the body weight (kg). AT is the averaging time (d): for noncancer risks, AT (d) = ED × 365; for carcinogens risks, AT (d) = average lifetime × 365. ED is the exposure duration (y). EF is the exposure frequency (d·y−1). CF is the conversion factor (1.0 × 10−6 kg·mg−1). PEF is the particle emission factor (1.36 × 109 m3·kg−1).

Carcinogenic and noncancer risks of PAEs were assessed following the procedures by the US EPA (2001) [48]. Noncancer risks of PAE exposure via dust were quantified by the hazard quotient (HQ) and hazard index (HI) using Eqs. (56):

$$ {\text{HQ}}_{{\text{i}}} = \frac{{{\text{ADD}}_{{\text{i}}} }}{{{\text{RfD}}}} $$
(5)
$$ {\text{HI}} = \sum {\text{HQ}}_{{\text{i}}} $$
(6)

where \(\mathrm{RfD}\) is the reference dose value of each PAE (mg·kg−1·d−1), HQ represents the health risks of the individual PAEs to human health via different exposure routes, and i represents different exposure routes. The values of HI > 1 × 10−6 indicate that the exposure of inhabitants to PAEs might induce the noncancer risk.

As for the carcinogenic PAEs, the lifetime average daily exposure doses (LADD, mg·kg−1·d−1) were estimated on the basis of the following expressions (79):

$$ {\text{LADD}}_{{{\text{der}}}} = \frac{{{\text{C}}_{{{\text{dust}}}} \times {\text{ABF}} \times {\text{CF}}}}{{{\text{AT}}}}\left( {\frac{{{\text{SA}}_{{{\text{child}}}} \times {\text{AF}}_{{{\text{child}}}} \times {\text{ED}}_{{{\text{child}}}} \times {\text{EF}}_{{{\text{child}}}} }}{{{\text{BW}}_{{{\text{child}}}} }} + \frac{{{\text{SA}}_{{{\text{adult}}}} \times {\text{AF}}_{{{\text{adult}}}} \times {\text{ED}}_{{{\text{adult}}}} \times {\text{EF}}_{{{\text{adult}}}} }}{{{\text{BW}}_{{{\text{adult}}}} }}} \right) $$
(7)
$$ \begin{array}{*{20}c} {{\text{LADD}}_{{{\text{inh}}}} = \frac{{{\text{C}}_{{{\text{dust}}}} }}{{{\text{AT}} \times {\text{PEF}}}}\left( {\frac{{{\text{IR}}_{{{\text{inhalatiionchild}}}} \times {\text{ED}}_{{{\text{child}}}} \times {\text{EF}}_{{{\text{child}}}} }}{{{\text{BW}}_{{{\text{child}}}} }} + \frac{{{\text{IR}}_{{{\text{inhalationadult}}}} \times {\text{ED}}_{{{\text{adult}}}} \times {\text{EF}}_{{{\text{adult}}}} }}{{{\text{BW}}_{{{\text{adult}}}} }}} \right)} \\ \end{array} $$
(8)
$$ {\text{LADD}}_{{{\text{ing}}}} = \frac{{{\text{C}}_{{{\text{dust}}}} \times {\text{CF}}}}{{{\text{AT}}}}\left( {\frac{{{\text{IR}}_{{{\text{ingestionchild}}}} \times {\text{ED}}_{{{\text{child}}}} \times {\text{EF}}_{{{\text{child}}}} }}{{{\text{BW}}_{{{\text{child}}}} }} + \frac{{{\text{IR}}_{{{\text{ingestionadult}}}} \times {\text{ED}}_{{{\text{adult}}}} \times {\text{EF}}_{{{\text{adult}}}} }}{{{\text{BW}}_{{{\text{adult}}}} }}} \right) $$
(9)

Carcinogenic risk (CR) was obtained from Eq. (10).

$$ {\text{CR}} = {\text{LADD}} \times {\text{CSF}} $$
(10)

where CSF represents the cancer slope factor, including ingestion cancer slope factor (\({\mathrm{CSF}}_{\mathrm{ingestion}}\)), inhalation cancer slope factor (\({\mathrm{CSF}}_{\mathrm{inhalation}}\)), and dermal adsorption cancer slope factor (\({\mathrm{CSF}}_{\mathrm{dermal}}\)). \({\mathrm{CSF}}_{\mathrm{ingestion}}\) is 0.0019 and 0.014 kg·d·mg−1 for BBzP and DEHP, respectively [27, 50]. \({\mathrm{CSF}}_{\mathrm{inhalation}}\) and \({\mathrm{CSF}}_{\mathrm{dermal}}\) were assumed to be equal to \({\mathrm{CSF}}_{\mathrm{ingestion}}\) to assess the cancer risks by dermal adsorption and inhalation [27]. A CR value below 1 × 10−6 represents the negligible or acceptable cancer risk. The relevant parameters for human exposure and health risk assessment are listed in Additional file 1: Table S3.

Results and discussion

Concentrations and spatial distribution of PAEs in indoor dust

The concentration and detection frequency of PAEs in indoor dusts from residences and dormitories are respectively shown in Fig. 1 and Additional file 1: Figure S2. All eight PAEs were detected in indoor dusts samples, demonstrating a detection frequency from 77.27 to 100%. DMP, DEP, DIBP, DEHP, and DBP were the most frequently detected compounds in all 94 samples, followed by BBzP, DnOP, and DNP in residences with the detection frequency of 81.94–95.83%. Similarly, the detection frequencies of BBzP and DnOP in the dormitories were 77.27 and 86.36%, respectively.

Fig. 1
figure 1

Concentration of PAEs in indoor dust samples (μg·g−1) from seven regions in China

The concentrations of the total PAEs (Σ8PAEs) in indoor dusts from residential buildings ranged from 8.86 μg·g−1 to 808.43 μg·g−1 with a median of 104.58 μg·g−1 and a geometric mean of 101.32 μg·g−1. The spatial distribution of Σ8PAEs is depicted in Fig. 2. Different concentrations were observed in the distribution of PAE concentration in indoor dusts among seven administrative regions of China. The highest concentration of Σ8PAEs in indoor dusts was found from Northeast China (NEC, median: 164.71 μg·g−1, range: 32.29–778.54 μg·g−1). Meanwhile, the relatively low values of Σ8PAEs were from Northwest (NWC, median: 145.11 μg·g−1, range: 16.88–534.65 μg·g−1), East China (EC, median: 107.59 μg·g−1, range: 32.91–808.43 μg·g−1), North China (NC, median: 82.60 μg·g−1, range: 38.46–416.66 μg·g−1), Central China (CC, median: 80.97 μg·g−1, range: 8.86–618.62 μg·g−1), and South China (SC, median: 71.72 μg·g−1, range: 25.04–185.32 μg·g−1). The lowest value of ΣPAEs in dust was found in Southwest China (SWC, median: 58.83 μg·g−1, range: 43.57–266.68 μg·g−1). On the whole, the concentration of PAEs in northern China was higher than that in southern China. The result was consistent with early studies [43], which was caused by the differences in usage patterns between eight PAEs among the seven geographic regions in China.

Fig. 2
figure 2

Spatial distribution of the Σ8PAEs in residential dusts of China

When considering the individual phthalates in indoor dust among all geographical regions, the composition with the highest concentration in indoor dusts from residential buildings was DEHP, which ranged from 5.60 μg·g−1 to 754.41 μg·g−1 with a median value of 50.79 μg·g−1, followed by DBP (median: 23.56 μg·g−1) and DIBP (median: 8.84 μg·g−1). DBP and DEHP are known as the main additives used in the industrial production [51]. Moreover, DBP and DIBP showed a relatively high abundance in indoor dust, which might be related to the extensive usage of DBP and DIBP as personal care products, such as cosmetics and pharmaceutical coatings [29, 29, 41].

The median concentrations of DEP, DMP, DNP, DnOP, and BBzP were two orders of magnitude lower than DEHP. Consistent with the results by other studies, the median concentrations of DEP and DMP were generally measured at lower levels than other PAEs in indoor dusts. This measurement was probably due to the high volatility and the reported existence of DEP mainly in gas phase than that in dust phase because of its high vapor pressure and low molecular weight [9, 49, 52].

A comparison of PAE concentrations in indoor dusts in different regions worldwide is presented in Additional file 1: Table S4. DEHP, DBP, DEP, DMP, and DIBP were the most commonly reported ones. The PAE concentrations (except for DMP and DIBP) detected in this study were slightly higher than those in a previous nationwide survey on the PAE levels in the indoor dust samples in China [43]. DEHP dominated the phthalates contaminating Chinese indoor dusts consistent with the profile in many countries worldwide, but DEHP levels measured in this study were significantly lower than other studies all over the world (Additional file 1: Table S4). The sequence of concentrations of major individual PAEs with DEHP > DBP > DIBP was in agreement with previous reports in China [43], Albany/USA [32], Sweden [9], and France [53]. The DBP concentration in indoor dusts in China (23.56 μg·g−1) was lower than that in Sweden (38.00 μg·g−1) [9], Saudi Arabia (33.00 μg·g−1) [54] and at higher levels than Albany/USA (13.10 μg·g−1) [32], Delaware/USA (12.00 μg·g−1) [55], California/USA (18.20 μg·g−1) [56], France (9.10 μg·g−1) [53], Canada (16.80 μg·g−1) [57], and Kuwait (2.00 μg·g−1) [54]. The DIBP levels in indoor dusts measured in this study were similar to those in some European countries, such as Sweden with a mean value of 7.90 μg·g−1[9], Belgium at 9.40 μg·g−1[58], and Ireland at 7.00 μg·g−1[58], while lower than those from Saudi Arabia (22.00 μg·g−1) [54], Kuwait (17.00 μg·g−1) [54], and France (20.00 μg·g−1) [53]. The difference in the compositional profile of PAEs in indoor dust worldwide might be attributed to the difference in housing environment, sampling time, utilization of phthalates in consumer products, and analytical techniques.

The concentrations of Σ8PAEs in dormitories were 14.03–1877.37 μg·g−1 with a median value of 55.32 μg·g−1. The highest level of individual PAEs in dormitories was DEHP with a median value of 34.65 μg·g−1, which was 2.85 and 3.84 times higher than DBP (12.17 μg·g−1) and DIBP (9.02 μg·g−1), respectively.

A comparative analysis of the median concentration of eight phthalate esters in dust in dormitories and residences is shown in Figs. 1 and 3. Thus, the PAE concentration in residences was higher than that in dormitories, except for DIBP. The difference between phthalate concentration in dust in residences and dormitories was likely due to the excessive use of decorative materials, household appliances, and furniture in residential buildings when compared with dormitories.

Fig. 3
figure 3

Comparison of the median concentration of phthalate esters in residences and dormitories

Distribution of PAEs in regions with different administrative levels

According to the administrative levels, the sampling regions of 72 residential dusts were divided into the following three types: provincial capital, nonprovincial city, and county. Environmental pollution is closely associated with the economic development and urbanization; thus, different degrees of environmental pollution are found in the three types of regions. As shown in Fig. 4, the concentrations of Σ8PAEs descended in the order of provincial capital (114.44 μg·g−1) > nonprovincial city (105.96 μg·g−1) > county (66.83 μg·g−1). Up to now, the concentration of PAEs in urban and rural indoor dust has not been discussed. Zhu et al. [43] indicated that the high levels of Σ9PAEs were observed in dust from developed areas, and the high producing capacity and consumption amount of PAEs in developed cities may influence their geographic distribution. In addition, the levels of PAEs in urban and rural soil have been reported in previous studies, which reflected that the urbanization and industrialization commercial activities and dense population may cause the increased PAE concentrations in soil [59, 60]. The high level of PAEs in the center of urban areas was probably caused by the densely anthropogenic activities [61].

Fig. 4
figure 4

Concentration of PAEs in three types of regions with different administrative levels

PAE occurrence in residential dust was mainly derived from the usage of PAE products (such as home building materials, electronic products, cosmetics, and consumer products). Previous studies have proven that phthalate contamination is related to the use of building materials and electronics. The highest concentration of Σ16PAEs in the agricultural soils of Guangzhou City was observed at the site close to the largest architecture market of South China, wherein a considerable amount of plastic wastes was discarded [62]. PAEs are also commonly used in consumer products [9], and these chemicals released from PAE products would not degrade because of the nature of the indoor environment and may persist for a long time [63]. The consumption of PAE products in large cities is higher than that in small places, which resulted in the low concentration of PAEs in the country.

Composition profiles of PAEs in indoor dust

The composition characteristics of PAEs in indoor dust are presented in Fig. 5. The composition of PAEs in residential indoor dust from different geographical areas of China was not different, indicating a similar source of PAEs in the indoor environment. Considering the composition among all geographical regions, DEHP was the most abundant PAE individual in indoor dust from residential buildings, followed by DBP and DIBP, demonstrating proportions of 60.42, 28.02, and 10.51% of Σ8PAEs, respectively. Especially, the contribution of DEHP to the total concentration of PAEs was reached 83.79% in Northeast China. Kang et al. [41] also illustrated that DBP, DEHP, and DIBP were the predominant compounds in indoor dust around the Pearl River Delta, accounting for 1.72 − 29.30%, 56.00 − 96.50%, and 0.35 − 13.40% to the total PAEs, respectively. Moreover, the predominance of DEHP and DBP was found in other environmental matrices, including soil [64, 65], water [16], sediment [16], air [13], and sludge [66]. Considering the low contribution of DMP and DEP, Orecchio et al. (2013) [67] reported that the percentages of the high volatile PAEs (e.g., DMP, DEP) in indoor dust were generally lower than those with low volatility.

Fig. 5
figure 5

Composition profiles of PAEs in indoor dusts of residences and dormitories from several regions in China

The proportions of individual PAEs in the indoor dusts in dormitories were in the order of DEHP > DBP > DIBP > DNP > DnOP > DMP > DEP > BBzP. The cumulative concentration of DEHP, DBP, and DIBP accounted for 99.00% of Σ8PAEs, which was consistent with residences. The similar pattern of PAE contamination profile in indoor dusts collected from residences and dormitories in China suggested the main sources of PAEs in different indoor dusts were similar.

Human exposure of phthalate esters via indoor dust

The ADDs of eight measured phthalates via multiple pathways in indoor dusts for children (between the age of 1–6) and adults are shown in Table 1. The results indicated that the total intakes ΣADD (ADDing + ADDinh + ADDder) of PAEs from indoor dusts for children and adults were 4.38 × 10−4 and 7.23 × 10−5 mg·kg−1·d−1 respectively indicating that children were more susceptible to the PAE intake than the adults. The ADDs of children via ingestion, dermal, and inhalation of indoor dusts were respectively 6.07-, 5.12-, and 2.14-folds higher than those of adults possibly due to the additional hand-to-mouth activities and the differences in selected parameters (such as lower body weights and high ED). Simultaneously, the calculated ΣADD, ADDing, and ADDder of women were respectively approximately 1.17, 1.17, and 1.04 times larger than men. However, the ADDinh of women was at a lower level than men, which might be attributed to different physiological characteristics, living habits, and other parameters. Dust ingestion was the main pathway of human PAE exposure from indoor dusts. The ADDs of PAEs for children and adults through dust ingestion were 2–4 orders of magnitude higher than those through dust inhalation and dermal adsorption. Similar results were also reported by other studies [27, 43].

Table 1 Health risk of human exposure to PAEs in indoor dust

The HI of noncancer risks for children, adult female, and adult male exposure to individual PAEs in the indoor dusts was in the decreasing order of DEHP > DBP > DnOP > BBzP > DEP > DMP. The HI values of DEP and DMP from indoor dust for children and adults were all within the acceptable level. However, the HI values of DEHP and DBP for all age groups were 3–4 orders of magnitude higher than 1 × 10−6. This finding indicates the noncancer risk to children and adults from DEHP and DBP exposure through indoor dust. Moreover, the noncancer risk for PAE exposure in children in indoor dust was larger than that in adults, and the HI for children was 6.07 times larger than adults. Simultaneously, the HI values of BBzP and DnOP to children also exceeded 1 × 10−6 with the values of 2.57 × 10−6 and 2.75 × 10−6, respectively. In addition, women were found to be at heightened risks than men (Table 1). The carcinogenic risks of DEHP and BBzP in indoor dusts for adults via three exposure routes were 3.97 × 10−7 and 1.30 × 10−10, respectively. Carcinogenic risk profiles for adult males (CRDEHP: 3.97 × 10−7; CRBBzP: 1.30 × 10−10) and females (CRDEHP: 3.96 × 10–7; CRBBzP: 1.30 × 10–10) were similar. These profiles were all below the threshold value (1 × 10−6), indicating that the carcinogenic risk of human exposure to DEHP and BBzP in indoor dusts was relatively low or negligible.

Possible sources of PAEs in indoor dust

Source apportionment by PCA model

PCA and Pearson correlation analysis were performed to investigate the possible sources of PAEs in indoor dust (Fig. 6, Additional file 1: Figures. S3 and S4). The first three principal components (PC1–PC3) with an eigenvalue > 1.0 were extracted for indoor dusts in residences. PC1 explained 33.50% of the total variance with loading of 0.97 and 0.85 for DIBP and DMP, respectively. The results were in agreement with the Pearson correlation analysis in the residential dust, which inferred a significant correlation between DMP and DIBP (Fig. 6), indicating possible sharing of similar sources. PC 2 accounted for 25.82% of the total phthalate esters in these samples and was dominated by low molecular PAEs with high loading for DEHP (0.90), BBzP (0.69), and DNP (0.75). As shown in Fig. 6, DEHP was significantly correlated with BBzP and DNP, suggesting the possibility of similar sources or emission behaviors. PCA suggested high loading of DBP (0.92) and DEP (0.86) on PC3, which contributed as high as 23.39% to the total variance. Significant correlations were found between DBP and DEP (Fig. 6).

Fig. 6
figure 6

Correlation analysis of PAEs of indoor dust. The bottom left corner is the correlation analysis of phthalates in indoor dust from residences. The top right corner is the correlation analysis of phthalates in indoor dust from dormitories

PAEs were used in a wide range of applications, and individual PAE might have multiple uses. High molecular weight PAEs, such as DEHP, DnOP, and BBzP, have been widely used as plasticizers in the polymer industry to improve flexibility, workability, and general handling properties [1, 68, 69]. BBzP and DnOP are also used in construction materials and home furniture, such as wallpapers, vinyl flooring, adhesives, and synthetic leather [1, 68]. PAEs with low molecular weights, such as DMP, DBP, and DEP, are mainly used in cosmetics and personal care products. In addition, DBP and DIBP are used in applications of plasticizers and personal care products [69]. DBP is also used in special adhesive formulations, cellulose esters, and epoxy resins [1, 70]. Therefore, DMP and DIBP in PC1 and DBP and DEP in PC3 were mainly related to cosmetics and personal care products and plasticizers. DEHP, BBzP, DnOP, and DNP in PC2 were mainly associated with the use of plasticizers, construction materials, and home furniture.

For indoor dusts in dormitories, the two extracted principal components explained 48.12 and 22.31% of the total variation (Additional file 1: Figure S4). The elements of the first factor corresponded to those for indoor dust from residences. DBP, DIBP, DNP, DEHP, and BBzP were significantly correlated with each other (p < 0.01 or p < 0.05). DEP and DMP concentrations were positively correlated (r = 0.664; p < 0.05) (Fig. 6). The second factor was loaded on DMP and DEP, which may also be related to the release from cosmetics, personal care products, and surface coating materials [25].

Source apportionment by PMF model

Two datasets were imported to the PMF 5.0 model in this study, and potential sources of PAEs in residences and dormitories were identified by the PMF model as shown in Fig. 7. The theory and formulas of the PMF and the rationality of the research results are stated in the supplementary material. Three sources for indoor dusts in residences were chosen as the optimal number for the PMF model and were comparable to those characterized by the PCA model with some differences. Factor 1 was responsible for 41.21% of the total variance and dominated by DIBP (83.64%), DBP (81.24%), DEHP (63.06%), and DNP (66.33%), which was associated with plasticizers employed in multifarious plastic products. Factor 2 accounted for 41.80% of the variance and obtained high concentrations for BBzP (79.82%) and DnOP (88.87%). Meanwhile, contributions to other PAEs were relatively low, and a large number of building materials and decorated furniture possibly contributed to this source. The third factor explained 17.52% of the variance and had high loads on DEP (87.70%) and DMP (61.52%), which was mainly linked to the cosmetics and personal care products.

Fig. 7
figure 7

Source profiles and contributions of PAEs in indoor dust from residences and dormitories based on the results of PMF

Two sources for dormitories were chosen as the optimal number for the PMF model. Source profiles for PAEs obtained from the PMF model are displayed in Fig. 7b. The source analysis result was consistent with that obtained by the PCA model. The contributions to the ΣPAEs of the two sources were F2 (60.38%) and F1 (36.92%). DMP (70.90%) and DEP (54.59%) were prominent in Factor 1, which is interpreted as the cosmetics and personal care products sources. For Factor 2, DBP (64.79%), DIBP (50.77%), DEHP (80.42%), BBzP (60.48%), DnOP (88.09%), and DNP (87.96%) obtained high weighting. Therefore, the second source denoted the joint contributions of plasticizers, furnishings, and construction materials.

The difference between source profile in dust in residences and dormitories was possibly related to the usage patterns and the dwelling characteristics. For example, high population density might cause a high contribution of personal care product sources in dormitories. Meanwhile, the use of additional household appliances, decoration materials, and furnishings in residential buildings might lead to the higher source contributions of the plasticizers and furnishings in residences compared with dormitories.

The fitting results were determined by linear regression between the predicted Σ8PAE concentration and the observed concentration. The fitting plots showed that the concentration predicted by the two models both fit well the observed concentration of PAEs in dormitories and residences, respectively (R2 range from 0.8197 to 0.8662) (Fig. 8). Overall, the PMF model (slopes were 0.9822 and 0.8869 for dormitory and residence, respectively) performed better with the slopes closing to “1” of the regression equations compared with the PCA model (slopes were 0.9339 and 0.8096 for dormitory and residence, respectively). The Σ8PAE concentration modeled by the PCA and PMF models presented good correlations. As shown in Additional file 1: Figure S5, R2 values were 0.8884 and 0.8060, the intercepts were 16.2120 and 1.6795, and slopes both approached unity. The comparison of the results between the two models based on the source profiles demonstrates considerable similarities between the two models, but some differences existed in source contributions. As mentioned above, 97.29% of the Σ8PAEs concentrations in dormitories were explained by the PMF model. The Σ8PAE concentrations in residences were slightly overestimated by the PMF model with 100.54%. However, the Σ8PAE concentrations in residences and dormitories were all underestimated by the PCA model with 81.70 and 70.43%, respectively. Overall, the results obtained by PMF provided practical and physical meanings. Multiple models should be conducted to reduce the weaknesses of individual approaches to obtain compelling results in future studies.

Fig. 8
figure 8

Model fit of the observed-versus-predicted Σ8PAE concentration (μg·g−1) from PCA and PMF

Conclusions

Eight PAEs, including DEP, DMP, DIBP, DBP, BBzP, DEHP, DnOP, and DNP, were detected in indoor dusts from several areas in China. Predominant contaminants in the indoor dusts were DBP, DEHP, and DIBP. The highest concentration of Σ8PAEs in indoor dust was found in Northeast China and decreased in the order of the Northwest, East China, North China, Central China, South China, and Southwest China. The median concentrations of PAEs in residential buildings were often higher than those in dormitories and the differences in source profiles between the two indoor environments. This finding suggests that the usage pattern of these chemicals was different between residences and dormitories in China. The highest PAE concentration was observed in the provincial capital, followed by nonprovincial city and country. Emissions of cosmetics and personal care products, plasticizers, and household building materials may explain the sources of PAEs in the indoor dusts. The ingestion of indoor dusts was the major pathway of human exposure to PAEs. When compared with adults, children face considerable health risks, and women are more threatened than men. Exposure to DBP and DEHP in indoor environments might cause notable noncancer risks for humans. In addition, the hazard indexes of BBzP and DnOP indicated their noncancer risk to children. The carcinogenic risk of human exposure to DEHP and BBzP in the indoor dusts was negligible. Additional attention should be provided to indoor dusts and measures should be taken to decrease daily PAE exposure of humans.