Chemical exposures from upholstered furniture with various flame retardant technologies

Abstract Upholstered furniture is often manufactured with polyurethane foam (PUF) containing flame retardants (FRs) to prevent the risk of a fire and/or to meet flammability regulations, however, exposure to certain FRs and other chemicals have been linked to adverse health effects. This study developed a new methodology for evaluating volatile organic compound (VOC) and FR exposures to users of upholstered furniture by simulating use of a chair in a controlled exposure chamber and assessing the health significance of measured chemical exposure. Chairs with different fire‐resistant technologies were evaluated for VOC and FR exposures via inhalation, ingestion, and dermal contact exposure routes. Data show that VOC exposure levels are lower than threshold levels defined by the US and global indoor air criteria. Brominated FRs were not detected from the studied chairs. The organophosphate FRs added to PUF were released into the surrounding air (0.4 ng/m3) and as dust (16 ng/m2). Exposure modeling showed that adults are exposed to FRs released from upholstered furniture mostly by dermal contact and children are exposed via dermal and ingestion exposure. Children are most susceptible to FR exposure/dose (2 times higher average daily dose than adults) due to their frequent hand to mouth contact.


| INTRODUC TI ON
Flame retardants (FRs) have been added to consumer products since the 1970s to reduce the risk of residential fires. The use of FRs has evolved since then in the United States (US) due to key events. These included the 2004 phase out of commercial pentabromodiphenyl ether (pentaBDE) and octabromodiphenyl ether (octaBDE), and the 2010 voluntary phase-out that additionally included decabromodiphenyl ether (decaBDE) and congeners of penta and octaBDEs after research showed that polybrominated diphenyl ethers (PBDEs) were persistent, bioaccumulative, and toxic. 1 The phase-out shifted the market to use alternative FRs such as the chlorinated FRs and eventually to organophosphorus FRs (OPFRs), shifting as research found adverse health impacts with other halogenated FRs. 2 Currently, the use of OPFRs is on the rise, while their toxicological hazard has yet to be wellcharacterized. 2 Studies have demonstrated that some OPFRs affect the endocrine system, behavioral development and preterm birth, respiratory outcomes like asthma, and allergic disease. 3 OPFRs can leach out of products over time and accumulate in settled dust in the surrounding environment. [4][5][6] Humans are exposed to OPFRs as they are found on human hands 7 and shown to metabolize in the human body. 8,9 In addition to the use of OPFRs as FRs, these chemicals are used in plasticizers, paints, glues, and for industrial processes. These organophosphates are prevalent and persistent in both indoor and outdoor environments. 10 In addition, volatile organic compounds (VOCs) are ubiquitous, and consumer products emit numerous VOCs. For this reason, indoor air has a complex mixture of VOCs many with higher concentrations than outdoor air. 11 Potential health impacts vary for specific VOCs based on exposure including cancer, reproductive harm, sensory irritation, odor annoyance, and headache. 12,13 Individuals spend 80% or greater of their day indoors and at least 66% of that in their residence. 14 This results in spending the majority of time in environments contributing to VOC and potential FR exposures.
Flammability standards for furniture, when they exist, vary by country. There is no national flammability standard in the US, but England, for example, has an open flame flammability standard. 15 The risk of fire continues to exist as domestic fires account for a large percentage of civilian fire deaths in both England 16 and the US. 17 For example, home fires caused 2720 civilian fire deaths, or 74% of the total fire deaths, in the US in 2018. 17 The furniture flammability standard in California shifted from a required open flame performance test to a smolder resistance test 18 which allowed products to be constructed and sold to the marketplace with little to no added FRs.
Studies coupling chemical exposure and fire performance risks of upholstered furniture manufactured with and without FRs are limited. This study investigated how humans may be exposed to VOCs and FRs from upholstered furniture manufactured with different FR technologies that currently exist in the market. This paper focuses on the exposure assessment component of a larger study on furniture flammability and human exposure to FRs, where chemical emissions into the surrounding environment during typical use of a lounge chair were examined using an exposure chamber and measuring inhalation, ingestion, and dermal contact exposures. The second portion of the study investigated flammability performance of the differently constructed lounge chairs and the relationship between daily exposure to FRs and fire hazards. 19

| ME THODS
The research protocol consisted of three parts: design and acquisition of upholstered lounge chairs; simulated inhalation, ingestion, and dermal contact exposure measurements; and human exposure analysis and assessment.

| Upholstered furniture
Upholstered furniture used in this study ( Figure 1) was a single person upholstered lounge chair, locally constructed for this study.
The model of the chair is commercially available, and study chair construction followed the manufacturer's typical fabrication methods and materials (Table 1).
Chairs used in this study varied only by the differing FR technologies. The four FR technologies used were as follows:

Practical Implications
• A method was introduced for evaluating potential human risks associated with flame retardants and/or other chemical exposures from a single household product, in this case study with upholstered furniture. Each household product is isolated inside an environmentally controlled exposure chamber and operated to simulate normal use as chemical emission rates are released and measured.
• The method aids in identifying sources and routes of human exposure to flame retardants and other chemicals of concern that may be associated with household products. Quantified emission rates contribute to finding the relationship between emission sources and empirical concentrations measured in indoor environments and human matrices (eg, serum, urine, breast milk, tissue, hair, nail).
• Upholstered furniture can be protected from fire hazards and chemical pollutant exposure without compromising utility using currently available technologies, for example, use of barrier textile between the cover fabric and foam of a chair. The details of the chair construction are in Supporting Information 1. In summary, the TBPP mix was identified to be 58% by weight of (4-tert-butylphenyl) diphenyl phosphate (4tB-PDPP), along with 30% by weight of TPHP, 11% by weight of (2,4-di-tert-butylphenyl) diphenyl phosphate (B4tBPPP), and 1% by weight of tris (4-tert-butylphenyl) phosphate (T4tBPP). The OPFR foam was found to contain 2.9% by weight of the TBPP mix. The RFR technology was expected to reduce leaching or migration of the FR from the product into the environment. The BNFR chair was fabricated without FRs but the PUF was wrapped and sewn with a woven fiberglass textile barrier which was placed directly under the cover textile.

| Exposure methods
All newly manufactured chairs were tested for VOC and FR chemical emissions inside a dynamic, environmentally controlled exposure chamber. Air, dust, and simulated skin transfer samples were collected during simulated chair use and analyzed for exposure levels.

| Exposure chamber
Each chair was tested in a 6 m 3 specialized exposure chamber ( Figure 2) with dynamic air flow and controlled environmental conditions. The chamber operated as a single-pass system without recirculation of chamber air and created an airtight seal with a closure mechanism with gaskets. All materials inside the chamber and chamber parts are made with low-emitting products; the chamber walls and air flow system components were constructed with electropolished stainless steel. The clean supplied air concentration remained below 10 μg/m 3 of total VOC (TVOC), 2 μg/m 3 of any individual VOC, 1 μg/m 3 of total particle mass concentration, and 2000/cm 3 of total particle count concentration. The air exchange rate was operated at 1 ± 0.05 air change per hour (ACH), maintaining temperature at 23 ± 1°C and relative humidity at 50% ± 5%. An

| Exposure testing
A typical exposure simulation and sample collection timeline is shown in Figure 3. Each test product required a total of 4 days to measure emissions of VOCs in air and FRs in air, settled dust, and simulated dermal contact. The first phase of sampling was background contaminant collections. The chair sitting mechanism was turned on in the empty chamber during the background sample collections. Air sampling was conducted for VOCs, aldehydes, and FRs in gas and particle phases. Immediately after the background airborne samples were collected, the chamber door was opened and the background dust wipe samples were collected. For the second phase of testing, the upholstered chair was introduced inside the chamber and equilibrated overnight before the sampling began.

F I G U R E 2
Chair and the automated sitting mechanism in an exposure chamber Then the sitting mechanism was turned on simulating the behavior of sitting activity on the chair for 24 h as VOCs and airborne FR samples were collected, followed by the collection of dust wipe samples. The simulated dermal FR transfer/skin absorption samples were collected from the chair seat cushion immediately after completion of air and dust sample collections. The interior surface of the chamber was wiped with deionized water and purged overnight with clean air to prepare for the next background sampling.

| Dust exposure
Settled floor dust around the test sample was collected using a wipe sample immediately after opening the chamber door after 24 h of chair agitation by the sitting mechanism, and the FRs in the dust F I G U R E 5 (A) Settled dust collection from the chair for FR analysis, and (B) simulated dermal transfer sample collection from chair with the filter patch (left) and filter placed between seat cushion and weight (right) sample were analyzed using the same method as for airborne FRs.
A fixed surface area was sampled using a defined 0.093 m 2 template and sterile gauze impregnated with n-hexane ( Figure 5a). The same surface area was wiped three times in three different directions.
Settled dust samples were collected in duplicate; one wipe sample was collected in front of the chair and the other to the side of the chair. The sampling method is based on EPA-740-R-13-001. 23 Since the sitting mechanism mechanically mimicked one sitting per minute inside the chamber, this equates to about 3 months of chair use through agitation prior to collecting dust samples.

| Dermal exposure
The simulated dermal sampling method was developed based on EPA Indoor Exposure Product Testing Protocols, 20 Thomas et al, 24 and EPA Guideline 875.2300. 25 Dermal transfer of FRs from the test product surface was simulated using a patch protocol, and FRs in the sample were analyzed as stated above. A filter paper patch im-    28 The predicted concentrations were compared with key globally accepted criteria for indoor air quality (Table 3).

| Exposure prediction
Thirty-one VOCs known as carcinogens, irritants, reproductive or developmental toxins 29 were measured from the four types of chairs.

TA B L E 2 TVOC and top ten individual
VOCs and their average emission rate (μg/h/chair) 1 carcinogen, 30 were emitted from all chair types. The predicted con-

| FR
Material content analysis performed by independent third-party laboratories showed that no FRs were detected in the PUFs except for that used in the OPFR chairs. The presence of brominated FRs was scanned for but found none above the LOQ hence the values are not reported here. FR emission measurements inside the exposure chamber was focused on the four identified FRs from the OPFR chair, TPHP, 4tBPDPP, B4tBPPP, and T4tBPP, and these were found in the air, settled dust, and simulated dermal transfer samples ( Figure 6).
ADDs of FRs for adults, toddlers, and infants were calculated for the four simulated human exposure pathways including particle inhalation, and ultrafine particulates/volatile inhalation, ingestion, and dermal contact exposure (Figure 7 and in Supporting Information 6). The higher FR values obtained from the duplicate test samples were used to obtain the worst-case exposure levels. This exposure data was limited, but it provides an exploratory evaluation of exposure potentials by different exposure routes as presented below.

| Inhalation
TPHP was detected in both volatile and airborne particle phases, and the predicted concentration in a living room with an average ventilation rate (0.45/h) was 0.11 ng/m 3 , much lower than one-tenth TLV guideline of 300 μg/m 3 . 33 Placing the OPFR chair in the bedroom would increase the TPHP concentration by a factor of 7 (with average ventilation) to 14 (low ventilation for high performance homes) compared to the concentration in the living room. TPHP was the only FR detected in the volatile phase. In the particle phase, 4tBP-DPP was also detected at a trace level (0.026 ng/m 3 ).
For the scenario of one chair in a living room with an average ventilation, inhalation was predicted to be the route resulting in the least FR exposure, accounting for less than 1% of total ADD values (summation of all exposure routes and FRs measured) for all ages (Figure 7). Toddlers had the highest predicted inhalation ADDs with 0.015 ng/kg/day of TPHP and 0.0003 ng/kg/day of 4tBPDPP. Infants' inhalation ADDs are at 96.4% and adults' inhalation ADDs are at 37.3% of toddlers' value.

| Settled dust
TPHP and 4tBPDPP were consistently detected in settled dust from the OPFR chairs but at or below LOQ, with an average TPHP level of 15.9 ng/m 2 and 4tBPDPP level of 0.74 ng/m 2 . Ingestion exposure was predicted to be the second most significant exposure route for TPHP, and the primary exposure route for 4tBPDPP in children (Figure 7).

FR ADD via ingestion was higher for children, especially infants, due
to their frequent hand to mouth contact from touching settled dust.
Infants have the highest ingestion ADD with 18.7 ng/kg/day (23.5% of total ADD) for TPHP and 1.02 ng/kg/day (80.8%) for 4tBPDPP.
Ingestion ADDs for toddlers were 10.6 ng/kg/day (19.1% of total ADD) for TPHP and 0.58 ng/kg/day (76.4%) for 4tBPDPP. Ingestion ADD for adults accounted for less than 3% of total ADD for both TPHP and 4tBPDPP. Infant's ingestion ADDs were up to 3060 times higher than adult's ADDs due to higher absolute ingestion and lower bodyweights. The exposure modeling showed that young children, the most susceptible population, likely receive the highest FR exposure. The total ADD was higher for infants and toddlers than for adults; total ADDs from the OPFR chair were 2.32 and 1.63 times higher for infants and toddlers respectively. Infants had the largest total ADD due to them having the largest dermal and ingestion ADDs. Although measured levels were below commonly accepted indoor air quality criteria, their levels could also be influenced by materials and environmental conditions. Formaldehyde and acetaldehyde were the prevalent VOCs of concern. FR exposure modeling and assessment indicated that the primary human health concern is settled dust/dermal contact exposure, with children having greater vulnerability due to their hand to mouth transfer. The use of a barrier textile or chemically bound FRs in PUF were shown to be available technologies to minimize VOC and FR chemical exposures.

| DISCUSS ION
Applied FR technologies should be free from potentially toxic chemicals, but they should also meet their objective which is fire prevention. Harris et al. 19  other potential chemical exposures, which is relevant to protecting the health and safety of consumers and fire safety professionals.

ACK N OWLED G EM ENTS
We would like to thank the furniture manufacturers, material suppliers, and chemical suppliers who contributed their expertise and commitment in development of our test products. Furthermore, we would like to thank Dr. Heather Stapleton of Duke University for flame retardant analysis, and Scott Laughlin, graduate student at the time, for his contributions.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.