NEHA November 2022 Journal of Environmental Health

November 2022 • Journal of Environmental Health 9 ticizers in polyvinyl chloride (PVC) items. The infrared and Raman spectra of phthalates in particular are well-characterized (Nørbygaard & Berg, 2004; Socrates, 2004). In this article, we assess the feasibility of using ATR-FTIR along with simple sample preparation to screen for three categories of hazardous chemicals found variably in consumer products, food-contact materials, and receipt papers: 1) phthalates and nonphthalate plasticizers, 2) organophosphate flame retardants, and 3) BPA and bisphenol S (BPS) in thermal paper. The screened samples consisted of 114 consumer products and food contact materials purchased between 2014 and 2020, and >200 receipts collected from retail businesses in 2017. We also tested PVC standards containing known levels of phthalates to assess detection limits and to compare with real-world products. We discuss e’ects of co-additives and fillers on spectral identification. Finally, we highlight cases in which our FTIR approach revealed novel or unexpected chemical additives in consumer products. Methods We used a Nicolet iS5 FTIR spectrometer with a single-bounce diamond ATR accessory. Absorbance spectra were collected from 4,000–500 cm-1 with 4 cm-1 resolution averaging 12–16 scans using Omnic software. No smoothing or processing was applied to the spectra. We used a combination of visual inspection of the spectral data and match searching within FTIR libraries both purchased (i.e., Thermo Fisher Scientific in 2008) and obtained in-house. Omnic Specta software was used to help identify some multicomponent samples. Regardless of the software, to determine a positive match we required visually apparent alignment of key peaks in the experimental spectrum with a known spectrum. Chemicals purchased as FTIR standards were: 1,2-cyclohexane dicarboxylic acid diisononyl ester (DINCH; Toronto Research); decabromodiphenyl ethane (DBDPE; TCI); tris(2-butoxyethyl) phosphate (TBOEP; Wellington Laboratories); triethyl phosphate and triphenyl phosphate (TEP and TPHP, respectively; Cambridge Isotope Laboratories); and bis(2-ethylhexyl) phthalate, diisononyl phthalate, and diisodecyl phthalate (DEHP, DINP, and DIDP, respectively; Sigma Aldrich). We used two certified reference materials from SPEX CertiPrep that contained PVC with 0.8% and 7.8% total phthalates. The 7.8% certified reference material contained 30,000 mg/kg each of DINP and DIDP and 3,000 mg/kg each of DEHP, benzyl butyl phthalate (BBP), dibutyl phthalate (DBP), di-n-octyl phthalate, diethyl phthalate, and dimethyl phthalate. The 0.8% certified reference contained 1,000 mg/kg each of DEHP, BBP, DBP, di-n-hexyl phthalate, diamyl phthalate, dicyclohexyl phthalate, diisobutyl phthalate (DIBP), and DINP. Additional reference standards with phthalate levels ranging from 0.1–1.0% were prepared in our laboratory by mixing PVC powder from Millipore Sigma with the certified reference materials in appropriate mass proportions. These powders were clamped directly on the ATR stage. Other chemicals reported in our results were identified based on matches within the purchased libraries. For extraction-IR, we used isopropanolcleaned scissors or a scraping tool to remove pieces of sample to be analyzed. After placing cut pieces into a glass vial, a few drops of isopropanol or ethanol (both from Fisher Scientific) were added to cover the sample. Vial lids contained either a Teflon or polypropylene gasket that were una’ected by the solvent. After at least 10 min, a metal dipstick was used to remove a drop of solution from the vial and place it on the ATR stage. The solvent was left to evaporate; then a spectrum was collected. A method blank was prepared by placing a few drops of isopropanol or ethanol in an empty vial and analyzing it in the same way. For plasticizers and bisphenols analyzed externally, GC/MS was carried out by two laboratories, Eurofins and TUV Rheinland. Both used organic solvent extraction and GC/MS based on CPSC-CH-C1001-09.3 or CPSC-CH-C1001-09.4. For flame retardants in child car seat samples, LC/MS/MS was carried out at Indiana University as described in Wu et al. (2019). Results and Discussion Phthalates and Alternative Plasticizers We used FTIR to identify phthalates as a class, not as specific congeners (e.g., diethylhexyl phthalate), because the di’erences in their FTIR spectra are too subtle. With few exceptions, phthalate congeners used in plastic products di’er only in length and branching of the alkyl chains R and R’ (Figure 1). Most phthalates we have identified in products have R and R’ of 8–10 carbons as determined by GC/MS. Thus, their FTIR spectra are extremely similar, di’ering only slightly in CH2 and CH3 stretching (near 2,900 cm-1) and bending (near 1,400 cm-1). Distinguishing these di’erent phthalates is further complicated because a given product could include more than one phthalate congener and/or di’ering isomers. Figure 2 shows ATR-FTIR spectra of PVC powders containing di’erent levels of total phthalates. We identified six key peaks that are useful for identifying phthalate presence and distinguishing phthalates from alternative plasticizers. The six key peaks are labeled in Figure 2: “twin peaks” 1,600 and 1,580 (orthophenyl stretching); 1,124 (symmetric COC stretch) appearing as a doublet with 1,073; and 1,040; and 743 cm-1 (out-of-plane CH deformation) (Socrates, 2004). By comparing FTIR with GC/MS results, we observed that when all six key peaks are apparent in the characteristic pattern in a spectrum, phthalate presence is unequivocal. When the twin peaks are unclear but the other peaks are apparent, phthalate presence is highly likely. When only a small number of phthalate peaks are visible, such as just 1,040 and 743 cm-1, phthalate presence is suspected and can be further investigated by extraction-IR. Figure 2 suggests that the limit of detection (LOD) for phthalates in PVC by visual observation with this method is 0.3–0.4% by mass. Two of the peaks, 1,124 and 1,073 cm-1, are still weakly apparent at 0.3% and even slightly at 0.2%. Extraction-IR allows clearer detection below 0.3%. We caution, however, that this LOD is based on high-purity PVC containing only phthalates. Real-world products, discussed shortly, typically have a higher LOD due to the obscuring e’ects of fillers, additional plasticizers, and other additives. An advanced data processing technique might detect phthalates at lower levels. Such an approach has been used, for example, with food adulterants (Özen & Tokatli, 2012), but might require too much development time to be practical. Interestingly, Omnic Specta multicomponent searching did not correctly identify the