PVC Shower Curtain Study
Phase 1: Concentrations of Potentially Hazardous Chemicals in PVC Shower Curtains
David E. Camann Michelle M. Zuniga Alice Y. Yau
Report for Phase 1
May 17, 2007 Southwest Research Institute Project 12662.01 Center for Health, Environment and Justice Michael Schade, PVC Campaign Coordinator Objective To measure the concentrations of selected potentially hazardous chemicals in five polyvinyl chloride (PVC) shower curtains provided by the Center for Health, Environment and Justice. Introduction The Center for Health, Environment and Justice CHEJ) intends to sponsor a chamber study to measure the emission of one or more potentially hazardous chemicals from a PVC shower curtain. A preliminary study was performed by Southwest Research Institute (SwRI) as Phase 1 in order to measure the concentrations of selected classes of potentially hazardous chemicals in five popular PVC shower curtains. The selected classes were chlorine, diester phthalates, organotins, and volatile organic compounds (VOCs). An analysis for metals was substituted for the VOC analysis after one VOC sample was analyzed, because such high concentrations of some VOCs were present that they saturated the column and required analysis of more diluted sample leachates to avoid instrument damage, thus raising the planned VOC detection limits. The Phase 1 measurements are intended to permit CHEJ to select the curtain to be evaluated in the chamber study and SwRI to determine the most appropriate chemical(s) to measure in the chamber study. Methods and Materials CHEJ purchased five unopened PVC shower curtains on September 27, 2006 at Bed Bath and Beyond, K-Mart, Sears, Target, and Wal-Mart and shipped them to SwRI. A description of each curtain is provided in the transmittal letter from Michael Schade (see 1
Appendix A). The five curtains were logged-in by SwRI on September 28, 2006 and instructions for each analysis were issued (see Sample Receipt and Task Order in Appendix A). The five curtain sample IDs are BBB 01, K Mart 02, Sears 03, Target 04, and Wal-Mart 05. Each shower curtain was opened, cut it into multiple pieces, and the pieces were weighed to achieve the following aliquot weights of each curtain for each analysis: • Chlorine: 1 gram (but 3 aliquots of 1 g of one curtain for a duplicate and matrix spike). • VOC: 2 aliquots of 5 grams, each in a 40 mL vial. • Phthalates: 1 gram. • Organotins: 1 gram. • Metals: 1 gram If a curtain consisted of two sheets, pieces of equal size were used from each sheet to achieve the specified weights. Chlorine was determined in bomb combustates of the 1 g aliquots by ion chromatography by a modification of ASTM D808-05. VOC sample preparation of a 5 g sample aliquot of Wal-Mart 05 was achieved by leaching the VOCs into heated water, with purge and trap by EPA Method 5035. The 65 target VOCs were determined by GC/MS by EPA Method 8260. Total VOCs were estimated by integrating the area under the entire chromatogram from the Method 8260 analysis, assuming the same response for all compounds as for the internal standard. An analysis for metals was substituted for the VOC analysis after the Wal-Mart 05 sample was analyzed because high concentrations of some VOCs which were present saturated the column and required raising the planned VOC detection limits for the other samples to avoid instrument damage. Three preparations of separate 1 g portions of each sample were used for the metals analyses. One 1 g aliquot was prepared and analyzed for mercury by CVAA according to SW-846 Method 7471A. Two other acid digestions were used for the remaining metals analytes. Copper and silver were determined from an open vessel digestion of another 1 g aliquot with concentrated nitric acid. The remaining metals (aluminum, arsenic, barium, beryllium, calcium, cadmium, cobalt, chromium, iron, potassium, magnesium, manganese, sodium, nickel, lead, antimony, selenium, thallium, vanadium and zinc) were determined from an open vessel digestion of another 1 g aliquot with concentrated nitric acid and aqua regia. All metals except mercury were determined by ICP by SW-846 Method 6010B. The phthalates extraction was based on the extraction method described by Hao-Yu Shen (2005). A 1 g aliquot for phthalates analysis was spiked with di-n-pentyl phthalate-d4 as an extraction surrogate and then extracted by soaking in 10 mL of hexane for 30 minutes, followed by sonication for 10 minutes. The solvent fraction was decanted off, fresh solvent added and the process repeated. The two solvent fractions were combined and the extract was reduced to 10 mL final volume. Diethyl phthalate (DEP), di-n-butyl phthalate (DBP), butyl benzyl phthalate (BBzP), di(2-ethylhexyl) phthalate (DEHP), 2
diisononyl phthalate (DINP), and di-n-nonyl phthalate (DNP) concentrations in the five sample extracts and in a solvent blank were determined on an Agilent 6890 GC equipped with a 5973 Mass Selective detector in full scan mode. A ZB-5ms 30 meter x 0.25 mm x 0.25 um analytical column was used. Di-(2-ethylhexyl) phthalate-d4 was used as the internal standard for quantitation. The extraction and derivatization of the organotins were based on the methods described by Dirks et al. (1994). A 1 g aliquot for organotins analysis was spiked with tri-n-propyl tin as an extraction surrogate and then extracted by adding 4 mL deionized water, 1 mL of glacial acetic acid, 1 mL of Diethyldithiocarbamic acid (DDTC) solution in pentane and 25 mL of hexane. The mixture was sonicated in an ultrasonication bath for 30 min. After phase separation, the organic layer was decanted into a beaker and a fresh 25 mL portion of hexane was added. The mixture was extracted for 30 minutes with a stirring bar. The sample was then centrifuged at 2500 rpm for 5 minutes. The combined hexane extracts were dried over Na2SO4. The Na2SO4 was rinsed twice with 10 mL portions of hexane. The combined organic phase was then evaporated to dryness and brought to 1 mL with n-octane. The extracts were cleaned with a Florisil/silica gel Restek SPE tube. An aliquot of each cleaned extract and the solvent blank were derivatized with npentylmagnesium bromide (Grignard reagent). Concentrations of monobutyltin, dibutyltin, tributyltin, tetrabutyltin, tricyclyohexyltin, triphenyltin, and di-n-octyltin in the derivatized extracts and solvent blank were determined on an Agilent 6890 GC equipped with a 5973 Mass Selective detector in selected ion monitoring mode. A HP-5 30 meter x 0.32 mm x 0.25 um analytical column was used. Tributyl phenyltin was used as the internal standard for quantitation. Results and Discussion The sample analysis data sheet for chlorine is provided in Appendix B and the chlorine concentrations measured in each shower curtain are presented in Table 1. Chlorine was a major chemical component of each shower curtain, with chlorine concentrations by weight ranging from 30% in Sears 03 to 35% in BBB 01 and K Mart 02. The chlorine concentration measurements were repeatable, with results of 34.6% and 35.3% obtained for the two replicate preparations and analyses of BBB 01. The high chlorine concentration demonstrates that polyvinyl chloride (PVC) was a primary constituent of each tested shower curtain. A narrative and the volatile organic analysis data sheets from the analysis of the undiluted Wal-Mart 05 shower curtain are provided in Appendix C. The estimated concentrations of detected target VOCs and of probable VOCs measured in the Wal-Mart 05 shower curtain are summarized in Table 2. Toluene, 2-butanone, and 4-methyl-2-pentanone (also called methyl isobutyl ketone (MIBK), were found to be present at very high concentrations in the undiluted analysis of the Wal-Mart 05 shower curtain, with responses which substantially exceeded the linear calibration range. Ethylbenzene, m/pxylene, and o-xylene were present at lower concentrations, (~ 300 to 150 ng/g), but still above the linear calibration range. Ten other targeted VOCs were present in Wal-Mart
3
05 within the linear calibration range, at concentrations between 46 ng/g for isopropylbenzene and 1 ng/g for 1,2,3-trichlorobenzene. The concentration of total VOCs in Wal-Mart 05 was estimated as 20,000 ng/g by integrating the area under the entire chromatogram in the full-scan analysis, assuming the same response for all compounds as for the internal standard. The calculation is given in Appendix C. Eleven non-targeted VOCs are presumed to be present based on the match of the sample peak with the library spectrum; their concentrations were estimated to be as high as approximately 220 ng/g of cycloheptane and 80 ng/g of decane, assuming the responses of the identified compounds and the nearest internal standard were the same. CHEJ requested that an analysis for metals be substituted for the VOC analysis after the undiluted Wal-Mart 05 sample was analyzed, because the high concentrations of some VOCs which were present had saturated the GC column and required raising the planned VOC detection limits to avoid instrument damage by only analyzing a dilution of the other samples. Since a dilution analysis of Wal-Mart 05 was not performed, more precise quantitation is unavailable for those six targeted VOCs (toluene, 2-butanone, MIBK, ethylbenzene, m/p-xylene, and o-xylene) which were present at highest concentrations in the Wal-Mart 05 sample. Chang et al. (2002) measured headspace air concentrations of 3,400 mg/m3 of toluene, 600 mg/m3 of MIBK, and 4 mg/m3 of ethylbenzene in static chamber tests of a 100% vinyl curtain with a strong “plastic” odor. These same VOCs, toluene, MIBK, and ethylbenzene, are three of the four principal VOCs detected in our analysis of the WalMart 05 curtain; and the relative air concentrations in the Chang study are consistent with the estimated curtain concentrations which we obtained for these same VOCs. However, 2-butanone was a major VOC in the Wal-Mart 05 curtain which Chang et al did not detect in the headspace air around the vinyl curtain in their study. The diester phthalate results for the undiluted solvent blank and for appropriate dilutions of the phthalates extract of each curtain sample are provided on data sheets in Appendix D. The concentrations of the six targeted diester phthalates in the five shower curtains are summarized in Table 3. Di(2-ethylhexyl) phthalate (DEHP) was the principal phthalate in three of the shower curtains: 25% by weight in the Wal-Mart 05 curtain, 24% in BBB 01, and 16% in Target 04. Diisononyl phthalate (DINP) was the principal phthalate in the other two curtains: 39% by weight in Sears 03 and 38% in K Mart 02. The Sears curtain contained a considerable concentration of DEHP (4.8%) in addition to its 39% DINP. All five curtains contained both DEHP and DINP, both at concentrations of 0.1% or higher in each curtain. Diethyl phthalate, di-n-butyl phthalate, butyl-benzyl phthalate, and di-n-nonyl phthalate were not detected in any of the shower curtains. None of the targeted diester phthalates were detected in the solvent blank, indicating no laboratory introduction of diester phthalates in any of the shower curtain extracts. The organotin results are provided on data sheets in Appendix E for the undiluted solvent blank and for the organotin extract, both undiluted and appropriate dilutions, of each curtain sample. The concentrations of the seven targeted organotin compounds measured in the derivatized extracts of the five shower curtain samples and the solvent blank are
4
summarized in Table 4. Dibutyl tin was found in three of the curtains: at 3.5 ug/g (= 3.5 ppm) in the Wal-Mart 05 curtain, 1.4 ug/g in K Mart 02, and 0.8 ug/g in Target 04. Monobutyl tin was found at lower concentrations, ranging from 0.1 to 0.4 ug/g in these same three curtains. A low concentration of tributyl tin (0.03 to 0.04 ug/g) was detected in the extracts all five of the 1-g aliquots of the shower curtain samples. However, this level of tributyl tin is primarily a laboratory-introduced contaminant, since tributyl tin was detected at a similar level (0.04 ug/extract) in the solvent blank extracted along with the five curtain samples. Tetrabutyl tin, tricyclyohexyl tin, triphenyl tin, and di-n-octyl tin were not detected in any of the five shower curtains. None of the six other targeted organotin compounds (besides tributyl tin) were detected in the solvent blank, indicating no laboratory introduction of these six organotins in any of the shower curtain extracts. A narrative and data sheets from the metal analysis of the five shower curtain samples, a preparation blank, and a laboratory control spike are provided in Appendix F. Concentrations of the targeted metals which were detected in one or more of the five shower curtains, a duplicate of BBB 01, or the preparation blank are summarized in Table 1. The metals concentration determined in duplicate digested preparations of BBB 01 were very similar, generally agreeing within 10% of each other, except near the detection limit. Arsenic was not detected in any of the five curtains, while the concentrations of most other most toxic metals were below or near the detection limit in four or five of the five tested curtains, with more elevated concentrations only of cadmium at 0.6 ug/g in K Mart 02, lead at 18 ug/g in Sears 03, and mercury at 0.005 ug/g in BBB 01. The concentrations of calcium, aluminum, and copper varied 100-fold across the five curtains, while the concentrations of barium, sodium, and zinc showed little variability across the five curtains. Conclusions and Recommendations The two diester phthalates (DEHP and DINP) and chlorine together comprised a large percentage of the total weight of each shower curtain: 49% of the total weight of the Target 04 shower curtain, 58% of the Wal-Mart 05 curtain, 60% of the Bed Bath & Beyond 01 curtain, 74 % of the K Mart 02 curtain, and 74% of the Sears 03 curtain. Di(2-ethylhexyl) phthalate (DEHP) alone comprised 25% by weight of the Wal-Mart 05 curtain, 24% of BBB 01, 16% of Target 04, 4.8% of Sears 03, and 1.4% of K Mart 02. Diisononyl phthalate (DINP) alone comprised 39% by weight of Sears 03, 38% of K Mart 02, and about 1 % each of BBB 01, Target 04, and Wal-Mart 05. Dibutyl tin comprised about 0.0004% by weight of Wal-Mart 05 and 0.0001% of K Mart 02 and Target 04; monobutyl tin concentrations were about 1/10th the dibutyl tin concentrations in these three curtains. Since the diester phthalates and organotins are not chemically bound to the vinyl chloride in the shower curtain, they can easily migrate from within the curtain to its surface, and then evaporate into the surrounding air. Due to their great mass in the tested shower curtains, it appears that DEHP and/or DINP are the best targets to monitor in the air around one of these shower curtains in the chamber study. DEHP has lower molecular weight and higher vapor pressure than DINP, and thus the DEHP in the PVC shower curtain will have a higher emission rate to the surrounding air than will the DINP in the curtain.
5
If DEHP is the primary emitted chemical of interest to CHEJ, it is recommended that the shower curtain from Wal-Mart or from Bed Bath & Beyond be selected for monitoring DEHP emission in the chamber study. If emissions of both DEHP and DINP are of interest, then it is recommended that the shower curtain from Sears be selected for monitoring both DEHP and DINP emissions in the chamber study. References Chang JCS, Fortmann R, Roache N, 2002. Air Toxics Emissions from a Vinyl Shower Curtain, In: Proceedings of Indoor Air 2002, H. Levin, ed., Santa Cruz, CA. Dirkx WMR, de la Calle MB, Ceulemans M, Adams FC, 2004. J Chromatography A, 683:51-58. Shen H-Y, 2005. Talanta, 66:734-739.
6
Table 1. Concentrations (ug/g = ppm) of Inorganics in Tested Shower Curtains Inorganic Compound Chlorine Aluminum Arsenic Barium Cadmium Calcium Chromium Cobalt Copper Iron Lead Magnesium Mercury Sodium Zinc Prep Blank < 20 < 1.8 < 0.20 < 0.2 < 0.10 < 2.5 < 0.20 < 0.20 < 0.10 < 2.0 < 0.20 < 2.0 < 0.0050 < 13 < 0.25 Bed Bath & Beyond 01 Sample Duplicate 346,000 (34.6%) 1.3 < 0.13 43 0.07 9.7 < 0.13 < 0.13 < 0.10 1.5 < 0.13 < 1.3 0.0054 18 36 353,000 (35.3%) 2.0 < 0.13 46 0.08 10.4 < 0.13 < 0.13 < 0.09 1.9 < 0.13 < 1.3 0.0054 20 37 K Mart 02 Sears 03 Target 04 Wal-Mart 05 328,000 (32.8%) 104 < 0.26 77 < 0.13 36 < 0.26 < 0.26 18.0 5.3 < 0.26 5.6 < 0.0047 39 36
351,000 301,000 319,000 (35.1%) (30.1%) (31.9%) 138 7.5 64 < 0.21 < 0.13 < 0.27 42 83 56 0.59 < 0.07 < 0.13 7.4 1,440 10.1 0.22 < 0.13 < 0.27 < 0.21 0.19 < 0.27 1.6 < 0.09 21.2 3.2 3.3 2.9 1.2 17.5 < 0.27 < 2.1 6.3 < 2.7 < 0.0047 < 0.0043 < 0.0047 46 20 33 33 13 65
7
Table 2. Estimated Concentrations (ng/g = ppb) of Detected Volatile Organic Compounds (VOC) in Wal-Mart 05 Shower Curtain Volatile Organic Compound Total VOCs Targeted VOCs: Toluene 2-Butanone 4-Methyl-2-pentanone (MIBK) Ethylbenzene m/p-Xylene o-Xylene Methylene chloride Acetone Benzene Styrene Isopropylbenzene 1,3,5-Trimethylbenzene 1,2,4-Trimethylbenzene p-Isopropyltoluene Naphthalene 1,2,3-Trichlorobenzene Additional identified VOCs: 1-Butanol Cyclohexane, methylene2-Propenoic acid, 2-methyl-, methyl ester Decane Heptane, 2,2,4,6,6-pentamethylCycloheptane Undecane 1-Hexanol, 2-ethylBenzene, 1-methyl-2-(1-methylethyl)Trans-decalin, 2-methylDodecane Concentration (ng/g) in Wal-Mart 05 20,000 * 2,500 ^# 5,200 ^ 1,900 ^ 240 ^ 260 ^ 160 ^ 3 2 5 3 46 9 25 2 3 1 2* 3* 5* 82* 37* 220* 46* 24* 12* 18* 6*
* Presumptive identification and estimated value, assuming same response as internal standard. ^ Crude estimated value, since response exceeded the linear calibration range. # Probable substantial underestimate, based on peak shape in chromatogram.
8
Table 3. Concentrations (mg/g and % by weight) of Diester Phthalates in Tested Shower Curtains Diester Phthalate Compound Diethyl phthalate Di-n-butyl phthalate Butyl benzyl phthalate Di(2-ethylhexyl) phthalate Di-n-nonyl phthalate Diisononyl phthalate Solvent Blank mg/extract < 0.025 < 0.013 < 0.013 < 0.013 < 0.012 < 0.25 Bed Bath & Beyond 01 < 0.25 < 0.13 < 0.13 240 (24%) < 0.12 1.3 (0.13%) K Mart 02 < 0.25 < 0.13 < 0.13 1.4 (0.14%) < 0.12 380 (38%) Sears 03 < 0.25 < 0.13 < 0.13 48 (4.8%) < 0.12 390 (39%) Target 04 < 0.25 < 0.13 < 0.13 160 (16%) < 0.12 1.1 (0.11%) Wal-Mart 05 < 0.25 < 0.13 < 0.13 250 (25%) < 0.12 1.0 (0.10%)
Table 4. Concentrations (ug/g = ppm) of Organotins in Tested Shower Curtains Organotin Compound Monobutyl tin Dibutyl tin Tributyl tin Tetrabutyl tin Tricyclohexyl tin Triphenyl tin Di-n-octyl tin Solvent Blank (ug/extract) < 0.02 < 0.02 0.04 < 0.02 < 0.02 < 0.02 < 0.02 Bed Bath & Beyond 01 < 0.02 < 0.02 0.03* < 0.02 < 0.02 < 0.02 < 0.02 K Mart 02 0.12 1.4 0.04* < 0.02 < 0.02 < 0.02 < 0.02 Sears 03 < 0.02 < 0.02 0.03* < 0.02 < 0.02 < 0.02 < 0.02 Target 04 0.38 0.81 0.04* < 0.02 < 0.02 < 0.02 < 0.02 Wal-Mart 05 0.15 3.5 0.04* < 0.02 < 0.02 < 0.02 < 0.02
* Tributyl tin is present in the shower curtains at similar level as in the solvent blank, indicating that most of the detected tributyl tin is a laboratory-introduced contaminant.
9