NEHA November 2022 Journal of Environmental Health

12 Volume 85 • Number 4 A D VANC EME N T O F T H E SCIENCE cent), and thus we found that flame retardant bands were not consistently distinguishable by ATR-FTIR of intact consumer products. Therefore, we routinely used extraction-IR to screen for flame retardants. We identified phosphorus-based f lame retardants in 36 samples taken from 18 child car seats. Samples included fabrics, polyurethane foams, and fabric–foam composites. An overview of these findings is presented in Table 1 with details in Supplemental Table 2. FTIR allowed discovery of a little-known flame retardant chemical that is a mixture of two cyclic phosphonates: 5-ethyl2-methyl-2-oxido-1,3,2-dioxaphosphinan5-yl)methyl methyl methylphosphonate and bis[(5-ethyl-2-methyl-1,3,2-dioxaphosphorinan-5-yl)methyl] methyl phosphonate p,p’-dioxide (PMMMPs). FTIR spectra of extracts from several car seat samples closely matched a spectrum in the HR Polymer Additives and Plasticizers Library called “phosphonate ester (cyclic)” or “Antiblaze 1045.” A literature search led to a CAS number and structure, revealing the mixture to be PMMMPs (Wu et al., 2019). To validate the finding, LC/MS/MS was carried out as described in Wu et al. (2019). An authentic standard for PMMMPs was not available, but a technical mixture (Hans TEX-3) was obtained from a supplier. Using this mixture as a standard, LC/MS/MS testing confirmed the presence of PMMMPs in the car seat fabrics, which was the first report of PMMMPs in consumer products in North America. This flame retardant had previously been reported in window curtains purchased in Japan (Miyake et al., 2018). Figure 4B shows spectra from car seat fabric and its isopropanol extract revealing PMMMPs. The intact fabric has a characteristic polyethylene terephthalate spectrum (“polyester”) with indications of an additive, but the matrix bands and subpercent level of the flame retardant make further identification di¢cult. Performing a multicomponent search using Omnic Specta software did not correctly identify PMMMPs in the mixture. Obtaining a drop of extract, however, led to the correct identification. Data in Supplemental Table 2 show we correctly identified PMMMPs with simple extraction-IR down to a concentration of slightly >400 mg/kg or 0.04%. The method did not produce false positives. PMMMPs were visible in the intact FTIR spectra for many samples, although extraction made the bands clearer. Phosphate esters—used as flame retardants in fabrics, polyurethane foams, and PVC articles—were also assessed. Supplemental Table 2 shows detection of TBOEP, TEP, and a small number of other flame retardants by extraction-IR compared with LC/MS/MS. Car seat samples with TBOEP ranging from 356 to 3,461 mg/kg were correctly identified by extraction-IR. The method did not detect TBOEP at 206 mg/kg. The method did not produce false positives. Extraction-IR performed poorly for TEP detection, failing in samples that concurrently contained higher levels of PMMMPs, presumably due to PMMMPs bands obscuring TEP. A total of four samples showed apparent false positives; the reason is unknown but could be due to nonhomoAttenuated Total Reflectance–Fourier-Transform Infrared (ATR-FTIR) Spectra Illustrating How Extraction-IR Reveals Additives Note. A: Cow-milking inflation liner intact sample and evaporated ethanol extract. Labeled peaks are consistent with styrene-butadiene rubber (intact) and with ortho-phthalates (extract). The six key phthalate peaks are labeled in bold. B: Child car seat fabric intact sample and evaporated isopropanol extract. Labeled peaks are characteristic of poly(ethylene terephthalate) (intact) and PMMMPs (extract). PMMMPs = 5-ethyl-2-methyl-2-oxido-1,3,2-dioxaphosphinan-5-yl)methyl methyl methylphosphonate and bis[(5-ethyl-2-methyl-1,3,2-dioxaphosphorinan-5-yl)methyl] methyl phosphonate p, p’-dioxide). 500 700 900 1,300 1,500 1,700 Absorbance 1,100 1,728 1,712 1,408 1,463 1,315 1,242 1,075 1,062 1,027 1,339 1,240 1,094 1,016 1,464 1,380 1,286 1,435 1,069 1,600 1,580 1,124 1,073 743 961 739 908 697 Intact Extract 500 700 900 1,300 1,100 1,500 1,700 Absorbance Wavenumber (cm-1) Wavenumber (cm-1) 969 827 909 871 722 846 Intact Extract A B FIGURE 4