Article pubs.acs.org/est
Measuring and Predicting the Emission Rate of Phthalate Plasticizer from Vinyl Flooring in a Specially-Designed Chamber Ying Xu,† Zhe Liu,‡ Jinsoo Park,‡ Per A. Clausen,§ Jennifer L. Benning,∥ and John C. Little‡,* †
Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, Austin, Texas, United States Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University (Virginia Tech), Blacksburg, Virginia, United States § New Technologies Group, National Research Centre for the Working Environment, Lersø Parkallé 105, DK-2100 Copenhagen Ø, Denmark ∥ Department of Civil and Environmental Engineering, South Dakota School of Mines and Technology, Rapid City, South Dakota, United States ‡
S Supporting Information *
ABSTRACT: The emission of di-2-ethylhexyl phthalate (DEHP) from vinyl flooring (VF) was measured in specially designed stainless steel chambers. In duplicate chamber studies, the gas-phase concentration in the chamber increased slowly and reached a steady state level of 0.8−0.9 μg/m3 after about 20 days. By increasing the area of vinyl flooring and decreasing that of the stainless steel surface within the chamber, the time to reach steady state was significantly reduced, compared to a previous study (1 month versus 5 months). The adsorption isotherm of DEHP on the stainless steel chamber surfaces was explicitly measured using solvent extraction and thermal desorption. The strong partitioning of DEHP onto the stainless steel surface was found to follow a simple linear relationship. Thermal desorption resulted in higher recovery than solvent extraction. Investigation of sorption kinetics showed that it takes several weeks for the sorption of DEHP onto the stainless steel surface to reach equilibrium. The content of DEHP in VF was measured at about 15% (w/w) using pressurized liquid extraction. The independently measured or calculated parameters were used to validate an SVOC emission model, with excellent agreement between model prediction and the observed gas-phase DEHP chamber concentrations.
■
INTRODUCTION
associated with exposure to certain phthalates. In addition, epidemiologic studies in children suggest associations between phthalates and the risk of asthma and allergies.1,11−15 Concentrations of phthalate metabolites measured in the general population using biomonitoring methods (blood and urine) provide evidence of widespread human exposure.5,16 Biomonitoring data based on blood suggest that over 75% of the U.S. population is exposed to phthalates.17 When urinary concentrations of secondary metabolites are measured, the estimate increases to 95%.18 Despite these health concerns and the ubiquitous exposures, few studies have been carried out to characterize emissions of phthalates, probably due to the difficulties associated with sampling and analysis of SVOCs.19 Uhde et al.20 measured phthalate concentrations emitted from 14 PVC-coated wall
Phthalate esters, used as plasticizers to enhance the flexibility of polyvinylchloride (PVC) products, are recognized as major indoor pollutants.1−3 These semivolatile organic compounds (SVOCs) are found in a wide range of consumer products including floor and wall coverings, car interior trim, floor tiles, gloves, footwear, and artificial leather.1 Following the restrictions on phthalates in toys and child care articles in the United States and Europe, phthalates used in PVC products are changing rapidly, with a shift from low to high molecular weight phthalates.4 Because the phthalate additives are not chemically bound to the polymer, slow emission from the products to the air or other media usually occurs. The potential adverse health effects of phthalate esters and their metabolites are detailed in several recent reviews.5−10 Collectively, these reviews suggest that exposure to some phthalates may result in irreversible changes in the development of the reproductive tract, especially in males. Effects such as increases in prenatal mortality, reduced growth and birth weight, and skeletal, visceral, and external malformations, are © 2012 American Chemical Society
Received: Revised: Accepted: Published: 12534
June 9, 2012 September 15, 2012 October 23, 2012 October 23, 2012 dx.doi.org/10.1021/es302319m | Environ. Sci. Technol. 2012, 46, 12534−12541
Environmental Science & Technology
Article
emission model using independently measured or calculated model parameters.
coverings in regular emission chambers for 14 days and showed that the chamber concentration of phthalates with lower boiling point tends to be higher than those with higher boiling point. Afshari et al.21 tested dibutyl phthalate (DBP) and di(2ethylhexyl) phthalate (DEHP) emissions from materials such as wallpaper, PVC flooring and electric wire in the Chamber for Laboratory Investigations of Materials, Pollution and Air Quality (CLIMPAQ) as well as the Field and Laboratory Emission Cell (FLEC) and found that the chamber concentration of DEHP reached steady state after about 150 days and that sorption by chamber surfaces had a strong effect on the rate of increase of gas-phase chamber concentrations. Fujii et al.22 developed a passive-type sampler to measure the emission rate of phthalates from synthetic leather, wallpaper and vinyl flooring and found that emission rates of several phthalates increased significantly at higher temperature. Kawamura et al.23 estimated the emission rate of DEHP from building materials through microchamber studies with variable surface area. Schossler et al.24 measured air concentration for diisononyl phthalate (DINP) and DBP emitted from PVC samples in the FLEC. Clausen et al.19,25,26 performed a series of tests on emissions of DEHP from vinyl flooring in the FLEC. Their work showed that (1) the emission rate of DEHP was limited by gas-phase mass transfer, (2) strong sorption of DEHP onto the FLEC chamber surface resulted in a very long time period to reach steady state, (3) relative humidity had no impact on the emission rate, and (4) increasing the ventilation rate increased the emission rate. Although the mechanisms governing the emission of phthalates from PVC products are still not fully understood, these chamber studies provided valuable information about the emission characteristics of phthalates from PVC materials. They also collectively reveal the substantial difficulties associated with chamber tests for SVOCs due to the low gas-phase concentration, strong sorption onto surfaces, and ubiquitous contamination in laboratory facilities. Informed by these early chamber studies, Xu and Little27 developed a fundamental mass transfer model to predict emissions of SVOCs from polymeric materials (such as DEHP from vinyl flooring). The model assumed that emissions are subject to external control, including partition equilibrium at the emission surface, convective mass transfer to the bulk air, and strong partitioning onto interior sorption surfaces. However, the model ignored convective mass transfer near sorption surfaces. The model was subsequently extended to include a mass-transfer coefficient that controls that rate of transfer to sorption surfaces, and used to investigate human exposures in the residential environment.28,29 More recently, it was shown that the gas-phase concentration of DEHP in equilibrium with vinyl flooring is essentially equal to the vapor pressure of pure liquid DEHP30 rather than being controlled by partition equilibrium at the emission surface. The aim of this study is to more completely characterize the mechanisms governing the release of DEHP from vinyl flooring in chamber experiments. To reduce the time taken to reach steady-state, a specially designed chamber which maximized the surface area of the vinyl flooring (VF) source and minimized the surface area of the internal stainless steel (SS) sorption surface was used. The sorption behavior of DEHP on the SS chamber surface was investigated by introducing SS rods into the chamber, allowing the chamber surface/air partition relationship to be directly measured. The concentration of DEHP in the VF material was also measured. Finally, the new chamber studies were used to validate the current SVOC
■
MATERIALS AND METHODS Chemicals. DEHP was obtained from Absolute Standards Inc. and methanol (anhydrous, 99.9%) was obtained from VWR International Inc. Vinyl Flooring. A 2 mm thick homogeneous polyurethane reinforced vinyl flooring was used. The VF contains about 15% (w/w) DEHP as the only plasticizer. It was delivered as a roll wrapped in plastic foil from a merchant in Denmark. A few days after receipt, it was cut into two 0.45 × 0.45 m square sheets, which were placed into the emission chamber. The specific weight of the VF (∼3.9 kg/m2) was provided by the manufacturer. Emission Chamber. The emission chamber was made of type 304 stainless steel (SS) with an electro-polished internal chamber surface. As shown in Figure 1, the thin chamber “ring”
Figure 1. Configuration of the chamber; (a) side and top view, (b) photo.
was positioned between two VF sheets. The two VF sheets and the internal SS chamber wall form a very short cylindrical cavity. In this way, we maximized the VF emission area and minimized the SS sorption area. The air flow from the inlet passes through the chamber and exits at the outlet, and the air velocity over the VF surfaces depends on the chamber flow rate. The VF itself acts as a good gasket. PTFE sheets (from FluoroPlastics Inc.) were used in the blank chamber. The air leakage rate was less than 2% of the total flow rate. Three type 304 stainless steel precision-ground rods (3 mm diameter × 6 cm in length), having similar roughness to the interior SS chamber surface, were inserted into the chamber and then periodically removed so that the sorbed surface concentration could be measured. This allowed us to relate the surface concentration to the gas-phase concentration in the chamber. Sampling of DEHP in the Effluent Air from Chamber. A schematic of the sampling system is shown in Figure 2a. Because the sampling system was supplied with high purity air from cylinders, we assumed that no particles were present. DEHP was sampled directly on Tenax-TA tubes with a pump (SKC 224-PCXR4) calibrated to a nominal flow rate of 130 mL/min. The sampling time was 24 h. Backup tubes were connected to each primary sample tube to check for breakthrough. To reduce the loss of DEHP adsorbed onto the SS tubing and connecting parts, the shortest possible pathway was used. To evaluate the effect of DEHP loss to the SS tubing and fittings along the sampling pathway, a second sampling system with an even shorter sampling path length was used. As shown in Figure 2b, two sampling ports were drilled next to the 12535
dx.doi.org/10.1021/es302319m | Environ. Sci. Technol. 2012, 46, 12534−12541
Environmental Science & Technology
Article
humidity does not significantly influence DEHP emissions26 in stainless steel chambers, only temperature and air flow rate through the chamber were checked before and after samples were taken for analysis. Sorption Experiment. Three SS rods were inserted into the chamber as shown in Figure 1. Because the surface of the SS rods is almost identical to the surface of the internal SS chamber walls, they provide a means to measure the sorption characteristics of the chamber surface. Before use, the SS rods were cleaned with methanol. Two methods, thermal desorption and solvent extraction, were used to measure the sorbed surface concentration of DEHP on the rods. For thermal desorption, the three SS rods were taken out of the chamber and quickly put into stainless steel tubes, then sealed prior to thermal desorption analysis with a TD-GC/FID system. The stainless steel tubes were cleaned with methanol and the empty tube background concentration was checked before use. For solvent extraction, the rods were taken out of the chamber and quickly put into three glass tubes, each containing about 10 mL of methanol as solvent. The tubes were then sonicated for 15 min to help the DEHP desorb from the SS rods and dissolve into the solvent. Next, the extracts were concentrated by gentle blowing with high purity nitrogen gas to evaporate the methanol. Finally, the concentrated extracts (0.2 mL) were directly injected into a GC-MS for analysis. All glassware was cleaned with methanol prior to use. Two sorption experiments were conducted with the rods. In the first set of chamber experiments, measurements were made as the DEHP chamber concentration increased during the approach to steady state, allowing the sorption isotherm to be roughly quantified. In this case, both thermal desorption and solvent extraction were used to quantify the concentration of DEHP on the surface of the rods. In the second set of chamber experiments, measurements were made at a constant DEHP chamber concentration after steady state had been reached, allowing the sorption kinetics and equilibrium end-point to be more carefully established. In this case, only thermal desorption was used to quantify the concentration of DEHP on the surface of the rods. Measurement of DEHP Content in Vinyl Flooring. Ten samples of approximately 30 mg were cut out of the VF with a pair of scissors at randomly chosen positions. The samples were extracted with methanol using pressurized liquid extraction with a Dionex ASE 200 system. The extraction cell with the sample was preheated to 150 °C for 7 min, followed by a static extraction of 10 min at constant pressure (2000 psi). After the static extraction, the pressure was released and the extract was collected in a 40 mL glass vial. The extraction cycle was repeated three times for exhaustive extraction of both samples and blanks. 5 μL of the extract was injected onto a Tenax TA tube, and the tubes were analyzed with TD-GC-FID. Analysis of DEHP Samples. TD-GC-FID System. A thermal desorber (TD) (Perkin-Elmer ATD 400) was connected to a gas chromatograph (HP 6890 GC) with flame ionization detector (FID). The sample tubes were desorbed for 30 min at 300 °C with a He flow of 50 mL/min and a cold trap temperature of minus 20 °C. The cold trap was narrow bore (low flow trap tube) packed with a small piece of silylated glass wool. Flash heating of the cold trap to 350 °C transferred the analyte through the valves at 225 °C and the transfer line at 225 °C to the GC. The GC-FID had a constant pressure resulting in a flow of about 10 mL/min at 120 °C and was equipped with a 30 m × 0.53 mm i.d. Restek RTX-1
Figure 2. Schematic of sampling systems.
chamber outlet and sorbent tubes were directly inserted into the chamber through the sampling ports with fixtures. The sampling time and flow rate were the same as in the first system. Emission of DEHP into Air. The first emission test was conducted in one chamber with a separate “empty” chamber as a blank. A duplicate emission test was repeated in an identical chamber after the first test was finished, but using the same pieces of VF. The test duration was about 150 days for the first sampling system and 80 days for the second sampling system with the shorter sampling pathway. Test conditions are shown in Table 1. Before each emission test, the stainless steel plates and especially the internal chamber surfaces were cleaned with an alkali detergent, hot tap water and distilled water, and then finally rinsed several times with methanol. Background measurements were also performed before each test. Because Table 1. Test Conditions and Model Parameters parameter
value
temperature (°C) chamber volume (L), V air flow rate (mL/min), Q air exchange rate (1/h) area of test pieces (m2), A internal stainless steel surface area (m2), As chamber diameter (cm) chamber height (cm) loading (m2/m3)a concentration in equilibrium with vinyl flooring (μg/m3), y0 Convective mass-transfer coefficient (m/s), hm convective mass-transfer coefficient near sorption surface (m/s), hs sorption surface/air partition coefficient (m), Ks
22 ± 0.2 2 850 ± 20 25 ± 1 2 × 0.126 0.02 40 1.6 126 1.1
measured measured measured measured measured measured measured measured calculated calculated
4.0 × 10−4 0.01
calculated model fitted
1800
measured, Figure 5
a
comments
Loading of FLEC is 510 m2/m3 12536
dx.doi.org/10.1021/es302319m | Environ. Sci. Technol. 2012, 46, 12534−12541
Environmental Science & Technology
Article
column. The temperature program was 120 °C, held for 2 min, increased to 300 °C at 15 °C/min, held for 8 min, and finally increased to 320 °C at 20 °C/min and held for 4 min. The FID heater temperature was 275 °C. The analytical detection limit was 0.01 μg/tube estimated based on evaluation guidelines for air sampling methods using chromatographic analysis.31 The standard solutions in methanol (0.01, 0.02, 0.05, 0.1, 0.2, and 0.3 μg) were injected into the Tenax tubes. The calibration curve had an R2 value of 0.9997, with 11 standards used for each concentration (a total of six different concentrations). All tubes were analyzed in two successive desorptions to ensure complete desorption of both the tube and the TD system. The second desorption showed concentrations below the detection limit in all cases. Before use, all tubes were conditioned at 310 °C for one hour with high purity nitrogen gas at 80 mL/min flow rate, and the background concentration of the clean tubes was also checked. GC-MS System. A GC (Agilent 6890) with mass spectrometer (Agilent 5973 MS) was used to analyze the samples. The GC-MS had a constant pressure of about 20 psi with a flow rate (using He as carrier gas) of about 1 mL/min. It was equipped with 30 m × 0.25 mm i.d. DB-17 column (0.3 μm film thickness). The temperature program was 200 °C, held for 1 min, increased to 300 °C at 10 °C/min and finally held for 1 min. The MS transfer line temperature was 280 °C. The MS was operated in the electron impact ionization mode (EI+, 70 eV) with a source temperature of 230 °C using full-scan mode (m/z 45−550). SVOC Emission Model. Figure 3 provides a schematic representation of the model that predicts emissions of SVOCs
(such as DEHP) from a polymeric material slab in a chamber. For low concentrations of volatile compounds in materials, a simple partition coefficient is often used to linearly relate the concentration in the air immediately adjacent to the material surface (y0, μg/m3) and the material concentration (C0, μg/ m3).32 However, for the high weight percent of phthalate in PVC products, the simple partitioning mechanism may not be applicable.27 Recent studies found that DEHP in PVC products exists in a thermodynamically separated phase with liquid-like properties,33−36 and that y0 is essentially equal to the vapor pressure of pure DEHP.30 Considering the mass transfer within the boundary layer due to the concentration gradient, the emission rate (E, μg/m2·s) is: E(t ) = hm ·[y0 − y(t )]
(1)
where y (μg/m3) is the bulk gas-phase concentration, and hm (m/s) is the convective mass-transfer coefficient. Assuming a linear equilibrium relationship for SVOCs, the ratio of the concentration of a chemical on sorption surfaces (such as the chamber wall or the rods) to its concentration in the gas phase is equal to the surface/air partition coefficient, Ks (m), or
K s = q(t )/y0s (t )
(2)
where q is the surface concentration (μg/m2) and y0s (μg/m3) is the gas-phase concentration immediately adjacent to the surface. Assuming a boundary layer exists adjacent to the sorption surfaces, the amount of SVOC accumulated on the surface is equal to the total mass transferred through the boundary layer from the gas phase, or dq(t ) = hs ·[y(t ) − y0s (t )] dt
(3)
where hs (m/s) is the convective mass-transfer coefficient near the sorption surface. With reference to Figure 3, the accumulation of SVOCs in the chamber obeys the following mass balance: dq(t ) dy(t ) · V = E (t ) · A − ·As − y(t ) ·Q dt dt
Figure 3. Schematic representation of SVOC emissions in an experimental chamber.
(4)
The current version of the SVOC emissions model is obtained by combing eqs 1−4.
Figure 4. Gas-phase chamber concentration of DEHP in (a) first sampling system, and (b) second sampling system with shorter sampling path length. 12537
dx.doi.org/10.1021/es302319m | Environ. Sci. Technol. 2012, 46, 12534−12541
Environmental Science & Technology
Article
Figure 5. Sorption of DEHP on stainless steel surface; (a) sorption isotherm obtained by thermal desorption, (b) sorption kinetics.
■
Sorption Experiment. Strong sorption occurred on the stainless steel surfaces. Both solvent extraction and thermal desorption methods suggest that a simple linear relationship is sufficient to describe the stainless steel surface/air partition relationship for DEHP. However, the partition coefficient measured by the solvent extraction method was only 65% of the value measured by thermal desorption. The reason may be loss of DEHP due to sorption to glassware or when transferring the sample between vials and tubes. The results suggest that the thermal desorption method is more reliable to measure the adsorption isotherm for DEHP on SS surfaces (Figure 5a). The measurements in Figure 5a were made as the DEHP chamber concentration increased during the approach to steady state, allowing the sorption isotherm to be roughly quantified. However, the ventilated chamber is a dynamic system, and the extent to which equilibrium was actually established between the rods and the bulk chamber air needs to be evaluated. Figure 5b shows the evolution of the DEHP concentration on the SS rods over time during steady-state conditions, when the gasphase chamber concentration was constant. The surface concentration of DEHP increased progressively until about 50 days, after which the sorbed surface concentration reached equilibrium with the chamber air, corresponding to a Ks value of 1800 m. The Ks value obtained from Figure 5a is 1400 m, or about 78% of the value established under conditions where equilibrium was shown to be obtained. The likely reason for the slow kinetics is the existence of a convective boundary layer. Rearranging eq 3 to investigate the sorption kinetics of DEHP, the surface concentration on the rod can be expressed as
RESULTS Emission of DEHP into Air. There was very good agreement between the duplicate chamber tests, as shown in Figure 4a. Because the exact same pieces of VF were used in succession, this shows that the amount of DEHP emitted from the VF is negligible relative to the amount initially present. The concentrations in the blank chamber were about 10 times lower than the corresponding highest measured concentrations in the sample chamber with VF (see Figure 4a). The conditions were relatively constant over the entire test period (Table 1). The DEHP concentration increased slowly and reached steady state (0.8−0.9 μg/m3) after about 40 days. Because the ratio of VF emission surface to SS sorption surface was high, the build-up of DEHP in the gas-phase occurred much faster and the time to reach steady state was significantly reduced, compared to the previous FLEC study of DEHP emission from the same VF19 when it took ∼150 days. When the second shorter sampling path was used (Figure 2b vs Figure 2a), the mass loss of DEHP onto the SS tubing and fittings before passing through the sampling tubes was reduced. Therefore, compared to the first experiment, the time for the measured gas-phase concentration to reach steady state was even shorter (20 days (Figure 4b) versus 40 days (Figure 4a)), but it did not affect the steady state DEHP level in the chamber air. To characterize the sorption kinetics, the stainless steel rods were inserted into the chamber after steady state had been reached. The additional sorption surface area (
