Car seats were purchased at retail prices from popular retailers or donated by the manufacturer for this study. Only infant and infant-to-toddler convertible seats were selected.
To test the car seats and strollers, we used a combination of techniques:
- On every component: X-ray fluorescence was used to measure potential flame retardant indicators: phosphorus, bromine, and chlorine.
- On select components: Infrared spectroscopy was used to identify material types, intentionally added phosphorus-based flame retardants, and certain other additives. Not all flame retardants are fully identified by this method.
- On outer fabric samples from most seats: Combustion with ion-selective electrode for measurement of total organic fluorine was used as a proxy for PFAS.
We isolated the components of each seat as shown in Figure 1, for a total of over 600 components from 25 seats (and 4 strollers). Seat covers and strap covers frequently consisted of several layers of fabric, foam, and interfacing, each of which we tested separately.
X-Ray Fluorescence (XRF)
XRF spectroscopy measures element levels. Our HD Mobile instrument from XOS is a high-definition XRF that uses monochromatic excitation energies of 7, 17, and 33 keV. The spot size is one millimeter. Elements heavier than aluminum are measurable. Detection limits are in the low parts-per-million (ppm) range for all elements of interest except chlorine and phosphorus, which have limits in the hundreds of ppm.
For most seats, component samples were cut with isopropanol-cleaned scissors. Seat fabric and interfacing (nonwoven fabric placed behind upholstery fabric or foam) pieces were folded several times and clamped to create a thicker and denser layer before XRF testing. This may allow a slightly more accurate measurement because the XRF penetration depth for elements of interest in most polymers is substantial, exceeding several millimeters. Thin foam layers were treated the same way.
While most seats were cut apart to obtain samples, four seats (Maxi-Cosi Romi, UPPAbaby KNOX, Clek Liing Mammoth and Clek Liing Railroad) were tested non-destructively in order to preserve the product for later donation and use. Fabrics and foams were tested by XRF directly on the seat rather than cutting and removing pieces. Soft and rigid foams were accessed underneath the seat upholstery. Our testing has shown this approach can slightly alter the levels of bromine, chlorine, and phosphorus reported by the software, but not enough to affect findings of FR presence or absence.
After collecting XRF data, we flagged bromine levels above 1000 ppm and phosphorus levels above 500 ppm as possible flame retardants. We also flagged chlorine above 1000 ppm as possibly arising from a chemical of concern such as a chlorinated flame retardant. These thresholds are consistent with those used by companies in the industry that do their own testing, and, based on our experience testing car seats, are likely to capture cases of intentional use. We tested components with those flags using FTIR spectroscopy as described in the next section.
We used a built-in XRF optimization mode for chlorine and phosphorus to improve their measurement. This optimization, however, causes the software to not report antimony, an element of concern found as a catalyst residual in some polyester fabrics. Although antimony is not a focus of this study, we collected an additional XRF scan in the default mode on all the major fabrics in order to collect antimony data.
More information on using XRF bromine, chlorine, and phosphorus as proxy indicators of FRs can be found in the references (6,17,18,19).
| Major Component Type | Minor Components | |
|---|---|---|
| Upholstery |
Seat cover fabric, headrest, fabric, seat cover foam, headrest foam |
Interfacing fabric, labels, and tags |
| Shell |
Frame, base, shade fabric and frame, rigid foam |
Clip, strap, strap cover, hook-and-loop fastener |
Fourier Transform Infrared Spectroscopy (FTIR)
A Nicolet iS5 FTIR spectrometer with a single-bounce attenuated total reflection (ATR) accessory was used to identify polymer type or fiber type as well as additives. Components with elevated bromine, chlorine, or phosphorus according to XRF were analyzed by ATR-FTIR before and after solvent extraction. We used a combination of FTIR spectral libraries from Thermo Scientific and reference standards in our own lab to match spectra and identify materials and extracted chemicals. Between the purchased libraries and our own standards we can identify many commonly used
FR chemicals, but not all possible ones.
Depending on the chemical, additives typically need to be present at levels of at least 0.1-1% in order to be detected by this method. Phosphorus-based FRs and melamine,6 also used as a FR, were identifiable by this method, while bromine-based FRs were only occasionally identifiable by this method due to poor extraction efficiency. Chlorinated organophosphate FRs are identifiable but were not found in this study. Other identifiable additives include but are not limited to plasticizers, slip agents, fillers such as talc and calcium carbonate, and some antioxidants.
To rank car seat chemical hazards, we considered the presence of phosphorus-based FRs identified by FTIR and elevated bromine measured by XRF. Samples with elevated phosphorus but without a positive flame retardant identification by FTIR were not marked as containing a chemical of concern because in some cases the source was an antioxidant containing phosphorus. Samples with elevated bromine were all marked as containing a chemical of concern because brominated chemicals, such as flame retardants and dyes, present a variety of hazards and are not necessary in children’s products.
Total Fluorine (F) Testing
Techniques that measure total organic fluorine provide a useful indicator for the presence of PFAS without the expense of targeted PFAS testing.11,20,21
From each seat and stroller tested for fluorine we cut two pieces of fabric from the main seat upholstery and the shade, if present. The fabric samples were shipped to Galbraith Labs in Tennessee. One set of the samples was subjected to oxygen flask combustion followed by ion-selective electrode measurement of fluorine. This test gives the total level of fluorine in the sample. The other set was subjected to ion-selective electrode measurement without combustion. This test gives concentration of inorganic fluorine (fluoride ion, F-).
Subtracting inorganic F from total F gave total organic F.
Six seats were obtained after the samples were shipped to Galbraith Labs and did not go through the total F test. To determine whether those seats likely contained PFAS, we used a water drop test, described in the section “How can I tell whether a seat has PFAS?” and communication with the companies about the fabrics used on their products.