United States Environmental Protection Agency Research and Development Air and {Energy Engineering Research Laboratory Research Triangle Park, NC 27711 November 1994 EPA/600/SR-94/141 r EPA Project Summary i Characterization of Emissions from Carpet Samples Using a 10-Gallon Aquarium as the Source Chamber Zhishi Quo and Nancy Roache As part of Phase I of a carpet bioresponse study sponsored by the U.S. Environmental Protection Agency (EPA), a study was conducted to evalu- ate the emissions from carpet samples that had previously shown toxic effects on experimental mice as reported by Anderson Laboratories, Inc., Dedham, MA, in 1992. The full report describes the major findings of the chemical char- acterization work conducted at the In- door Source Characterization Labora- tory of EPA's Air and Energy Engineer- ing Research Laboratory. All other re- sults (animal testing, microbial testing, chemical analysis by sample extrac- tion, and pesticide analysis) are re- ported separately. The experimental system used in this study was first developed by Anderson Laboratories and was identical to the system that EPA's Health Effects Re- search Laboratory (HERL) used in car- pet bioresponse testing. Duplicate tests were conducted for each of three - samples received from the Consumer Product Safety Commission: two previ- ously used carpet samples plus mock (empty bags) samples. An emissions characterization team from Acurex Environmental Corpora- tion evaluated the experimental sys- tem and concluded that the test sys- tem developed by Anderson Laborato- ries was not suitable for carpet chemi- cal emisisions characterization because of poor reproducibility, nonuniform thermal conditions, and emissions from the source chamber itself. The 1-h bake cycle prior to the dynamic mode is not typical of indoor air characterization methods. This Project Summary was developed by EPA's Air and Energy Engineering Research Laboratory, Research Tri- angle Park, NC, to announce key find- ings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction In 1992, researchers at Anderson Labo- ratories, Inc., of Dedham, MA, reported toxic effects in experimental mice exposed to emissions from selected carpet samples Because of the potential public health sig- nificance of their reported findings, the U.S. Environmental Protection Agency (EPA) and the Consumer Product Safety Commission (CPSC) initiated studies in 1993 to evaluate Anderson's experimen- tal method and to replicate the reported findings. In addition to a comprehensive toxicity screen and microbial characteriza- tion, the EPA test plan (Phase I) called for a thorough chemical characterization of emissions from carpet samples collected by CPSC. This report summarizes the find- ings of the carpet emissions characteriza- Printedon Recycled Paper ------- tion performed by Acurex Environmental Corporation under EPA Contract 68-DO- 0141 at the Indoor Source Characteriza- tion Laboratory of EPA's Air and Energy Engineering Research Laboratory (AEERL). The experimental system used in this work was based on a protocol provided by Anderson Laboratories and was identi- cal to that used by EPA's Health Effects Research Laboratory (HERL) for replicat- ing Anderson Laboratories' tests. The objectives of the study were two- fold. First, emissions tests were performed to Identify potential toxic volatile organic compounds in the emissions from carpet samples and to determine the concentra- tion levels during the exposure. Second, because of the unconventional nature of the method, the experimental system was characterized to determine test conditions and the background emissions from the source chamber. Method Test System and Protocol The test system consisted of four func- tional parts: air supply system, source chamber, exposure chamber, and air flow control. During a test, the carpet sample was placed in the source chamber (a 10- gal — 38 L—glass aquarium) and heated to elevated temperatures. Humidified zero- grade air was then introduced to the cham- ber to carry the emissions to the exposure chamber, where the experimental animals are tested. Air samples were collected for * chemical analysis from the source cham- ber after the 1-h static period and from the exposure chamber during the dynamic period. Duplicate tests were conducted on each of the three carpet samples randomly re- ceived from CPSC. Samples A and C were previously used carpet samples col- lected by CPSC, and Sample B was re- ceived as an empty bag to indicate an empty chamber test. The samples were received as follows: Sample A for tests 1 and 6; Sample B for tests 2 and 3; and Sample C for tests 4 and 5. Each test consisted of four 1-h exposure periods and took two days to complete. Test Parameters The major test parameters were Size of carpet sample: 2,900 cm2 Observed sample temperature: ~50°C (at the surface of the carpet backing) ~70°C (hot spot on the fiber side) Target air temperature in source chamber: 37 ± 3°C Observed average air temperature in source chamber: -41 °C Volume of source chamber: Source chamber air flow rate: Target relative humidity of inlet air: 38 L , 7 L/min 45 + 5% Volatile Organic Compound Analysis The volatile organic compounds (VOCs) were collected on multisorbent traps dur- ing chamber testing and analyzed by ther- mal desorption-capillary gas chromato- graph (GC) equipped with a mass selec- tive detector (MSD) for compound identifi- cation and a flame ionization detector (FID) for compound quantification. Individually identified compounds were quantified, and the emissions of total volatile organic com- pounds (TVOCs) were estimated. The VOCs were measured using sam- pling and analysis procedures developed and implemented in previous emissions testing at AEERL. These procedures in- cluded daily tuning of the MSD for identifi- cation, five-point calibration of the GC/FID for quantification, analysis of daily check samples, analysis of field and laboratory blanks, and verification of sorbent trap background concentrations (blanks). Measurements of Particle Concentrations The instrument used for monitoring par- ticle concentration was a model 8010 PortaCount particle counter (TSI). The in- strument was operated in the "Count Mode," in which the instrument directly counts the aerosol drawn through the sample port and gives the concentration in particles per cubic centimeter (P/cm3). Particle concentrations between 0 and 5 x 10s P/cm3 can be measured with this in- strument. For comparison purposes, the particle concentrations in the laboratory air were measured before and after each exposure. Results Characterization of Physical Parameters Tracer gas measurements showed that the air in the exposure chamber was well mixed. The inlet and outlet air flow rates were measured with an electronic bubble flowmeter at the start of each test to make sure the difference between the two flow rates was within 10%. Comparison of the outlet air flow prior to each exposure and after the completion of the test indicated increased leakiness of the system during testing due to heated duct tape and poor seals in the system. The pressure differ- ence between exposure chamber and the laboratory air was negligible. The source chamber was heated with heating pads from outside, creating a poorly controlled thermal environment. Temperature data from 12 locations were collected at a frequency of one reading every minute and logged by a computer. Figure 1 shows an example of tempera- ture profiles in the source chamber. Total Volatile Organic Compounds Peak TVOC concentrations found in the source chambers at the end of the initial 1-h static heating period were approxi- mately 10 mg/m3 for each carpet sample. The background contribution to the TVOC concentration from the source chamber was <1 mg/m3 arid represented <10% of the TVOC during the carpet test. Peak TVOC concentrations averaged 2-8 mg/ m3 in subsequent samples collected from the source chamber after the 1-h static periods with the same carpet sample. Each exposure followed a pattern of an initially high concentration followed by a continu- ous decay in concentration during the 1-h exposure period 'because of the dilution by clean air (Figure 2). During subse- quent 1-h exposure periods, the source strength was lower and variable (Figure 3). Individual and Classes of Compounds in Carpet Emissions More than 200 compounds were ob- served in the carpet emissions. About 15% were identified and confirmed by interlaboratory comparison of GC/MS analysis, another 70% were tentatively identified, and the remaining 15% were not identified. The identified compounds fell into the following classes: alkanes, alkenes, cycloalkanes, cycloalkenes, oxy- genated hydrocarbons, one- or two-ring aromatic hydrocarbons, siloxanes, and phenols. Oxygenated hydrocarbons, aro- matic hydrocarbons, and siloxanes were also emitted from an empty source cham- ber. Table 1 lists compounds identified by Acurex for each test sample. Table 2 shows the classes of compounds found in the emissions from each sample. Concentrations of the predominant com- pounds, including those identified by HERL as potentially toxic for each sample, are listed in Table 3. The data are the aver- age concentration from a 60-minute sample during Exposure 2 of each test. ------- 80' 70- 60- O 50- O £ i 40 H CD Q. £> 30- 20- 10- —1— 50 100 Carpet Fiber Carpet Backing Air in Aquarium Second Exposure 150 200 250 Elapsed Time (min) 300 350 —i— 400 450 Figure 1. Typical temperature profiles in the source chamber. 10000 10 15 20 25 30 Elapsed Time (min) 35 40 45 50 Figure 2. Observed TVOC emissions from Sample A (Exposure 1, Test 1). ------- 5000- Exp1 Exp 2 Exp 3 Exposure \D Exp 4 I I Sample A RSSJ Sample B Sample C Figure 3. Average TVOC concentrations in the exposure chamber during four 1-h exposure periods of a test. Particles The particle concentrations in the cham- ber air were one to two orders of magni- tude lower than those found in laboratory air. Conclusions The objective of this study was to char- acterize the physical parameters of the test system and the chemical emissions from two specific carpet samples and the empty source chamber under test proto- col conditions. The experimental system used for the physical and chemical char- acterization was identical to the system used by HERL in their bioresponse test- ing. Although the experimental systems were identical in design and materials, the emissions generated during testing with individual systems could be different based on the following observations: • Nonuniform heating of chamber sur- faces, chamber air, and carpet samples • Development of air leakage in cham- bers during testing • Emissions of pollutants from the source chamber • Inadequate temperature control be- cause of low precision manual tem- perature controls The study results indicate that environ- mental conditions could not be precisely controlled or reproduced. Therefore, there is no assurance that identical systems would produce identical emissions. More than 200 compounds were emit- ted by the two carpet samples that were tested. Of the 200 compounds, 29 (15%) were identified by GC/MSD and confirmed, and another 70% were tentatively identi- fied. Of the 29 compounds that were con- firmed, 58% were found in both carpet samples tested, and five of the confirmed compounds were observed in all three of the test samples (two carpets and empty chamber). Most of the emissions from the empty source chamber were siloxane iso- mers with most of the emissions being less than the quantification limits of the analytical instruments. Quantitative differences of some of the individual compounds were observed dur- ing an exposure, between the four suc- cessive exposure cycles of a single test, and between replicate tests using differ- ent subsets of the same carpet sample. Although the same flow rate and tempera- ture protocols were followed throughout this study and replicate subsets of the same carpet samples were tested, no two exposures produced the same emission profile. During the exposure period, the TVOC concentration and concentrations of some individual compounds decreased with time but did not exhibit an exponen- tial decay. Some of the predominant highly volatile compounds observed in Exposure 1 were below the detectable limits of the analytical systems in subsequent expo- sures. The emissions from these tests were a function of the exposure protocol and the time during the exposure at which the samples were collected. No evidence was found to support the hypothesis that the carpet samples could generate a significant amount of particles under the experimental conditions. The data reported in this document are representative only of the two carpet samples tested during this study. The car- pet samples evaluated were not new; some of the emissions may have been of chemicals adsorbed onto the samples dur- ing previous use. , ------- Table 1. Individual Compounds Identified in the Three Samples Compound Sample A Sample B Sample C Acetone Isopropanol Benzene Acetic Acid Toluene Hexanal Ethylbenzene m,p-Xylene N,N-Dimethyl-acetamide Styrene o-Xylene a-Pinene Benzaldehyde Decane Trimethylbenzene Limonene Acetophenone Terpene Undecane n-Dodecene Camphor Naphthalene Dodecane Dodecamethylcyclohexasiloxane 4-Phenylcyclohexene Butylatedhydroxytoluene Hexadecane Butanoic acid 2,3-Dihydro-1,1,3-trimethyl-3-phenyl-1H-indene Table 2. Classes of Compounds Identified in the Three Samples Class Sample A Sample B Sample C Alkanes Alkenes Cycloalkanes Cycloalkenes Oxygenated hydrocarbons Siloxanes Substituted benzene Substituted phenol ------- Tsbte 3. Predominant Emissions by Test (Concentration unit: iig/rrf) Compound Sample'/Test A/1 A/6 B/2 B/3 C/4 C/5 Butylatedhydroxytoluene (BHTf Acetic Acid* Naphthalene1 Toluene Nonanal Tri(t-butyl) phenol Phenol Siloxane isomer (retention time 59.9 min) 386 14 19 134 49 48 ND 25 407 27 22 10 108 55 ND 32 29 ND ND 44 ND ND ND 6 18 ND ND 21 ND ND ND 4 2 ND ND 4 53 ND 15 73 21 2 3 7 1693 ND 73 39 TVOCs 1737 2198 182 115 1890 3181 'Samples: A and C = carpet; B » empty chamber used as a control. 'Identified as potentially toxic by HERL. "Coelution of nonanal and siloxane isomer. ND • not detectable by analytical system. ------- ------- Zhishi Quo and Nancy Roache are-with Acurex Environmental Corp., Research Triangle Park, NC 27709. Mark A. Mason is the EPA Project Officer (see below). t^mnloa The complete report, entitled "Characterization of E™ss™s*ro™C Using a 10-Gallon Aquarium as the Source Chamber," (Order No. Cost: $27.00, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Air and Energy Engineering Research Laboratory U.S. Environmental Protection Agency Research Triangle Park, NC 27711 United States Environmental Protection Agency Center for Environmental Research Information Cincinnati, OH 45268 Official Business Penalty for Private Use $300 EPA/600/SR-94/141 BULK RATE POSTAGE & FEES PAID EPA PERMIT NO. G-35 ------- |