Equilibrium partition coefficients of semi-volatile organic compounds (SVOCs) between indoor surfaces and air assessed using two experimental methods

Indoor exposure to Semi-Volatile Organic Compounds (SVOCs) is affected by their sorption onto indoor surfaces such as clothes, toys, building materials and furniture. They are transferred from these contaminated surfaces to the body via various routes such as skin contact, inhalation or hand-to-mouth transfer. Knowledge of their sink surface/air partition coefficients K Surf are needed to apply indoor fate models and assess occupants’ exposure to SVOCs. Partition coefficients of triphenyl phosphate (TPP), 2-Ethylhexyl diphenyl phosphate (EHDPP) and 1,2-Cyclohexane-dicarboxylic acid, dinonyl ester (DINCH) on cotton, glass and stainless steel (SS) surfaces were experimentally determined using two different methods. In the Micro-Chamber/Thermal Extractor (M-CTE), K cotton values were close to 10 4 m. For this method, improvements were proposed to investigate the sorption on surfaces having lower sorption capacities such as glass or SS. In the sandwich-like emission cell, K glass and K SS values ranged from 20 to 465 m, and from 99 to 1570 m, respectively. In coherence with the literature, sorption on cotton was higher than on SS and glass. Moreover, based on literature data, the SVOC vapor pressure (V P ) could be a good predictor of K Surf for non-porous surfaces such as SS and glass. For porous surfaces (cotton), additional parameters related to the surface (porosity or composition) should be used with V P . Strong sorption capacity of cotton reinforces the health risk posed by contaminated clothes. This study highlights the need for further research to understand the key parameters of SVOC sorption on textiles. • New M-CTE method to experimentally determine K Surf of SVOCs • Complementary K surf data obtained with sandwich-like emission cell method • New data on K surf of SVOCs that are currently in use • Sorption on cotton textiles is higher than on stainless steel (SS) and glass surfaces • Linear relationships between LogK Surf and LogV P are satisfying for SS and glass