Study On The Key Steps Of Development Of Predictive Model Of Residential Indoor Phthalate Concentrations

Phthalate esters（PAEs）are a group of endocrine disruptors that are widely present in residential indoor environments.As typical semi-volatile organic compounds（SVOCs）,PAEs can simultaneously exist in various environmental media such as indoor air,PM_2.5 and house dust.PAEs can enter the human body through multiple exposure pathways such as ingestion,inhalation and skin absorption,and pose serious threats to human health.The properties of SVOCs make exposure assessment of PAEs very complex.On the other hand,the development of predictive models for indoor PAE concentrations is one of the solutions to overcome the above difficulty in exposure assessment,given the mass transfer and partition among source materials and indoor environmental matrices.Two key steps in model development are:（1）characterization of source emission process and（2）characterization of parition of PAEs among multiple environmental media.However,there are still many uncertainties:（1）it has been widely accepted that the emission of PAEs from source material is dominated by"external control";the key characteristic emmission parameters（the gas-phase concentration of in equilibrium with the material phase（y₀）and the convective mass-transfer coefficient（h_m））were determined in chambers under controlled conditions,which are far from real environmental conditions;（2）the partition of PAEs among various indoor environmental media is generally assumed to instantaneously reach equilibrium,and PAEs adsorbed to solid phases originate from gas-phase counterpart;there is uncertainty in such assumption.The research work in this dissertation was conducted to address the above uncertainties.The research results are as follows:（1）It was found the indoor temperature,wind speed and characteristic length of emission sources along wind direction in Chinese homes are in the range of 16-28℃,0.15-0.25 m/s and3-4 m,respectively.Sensitivity analysis showed that the changes of above three built environmental conditions within the small range have negligible effects on h_m.Subsequently,the localized h_m values of Di BP,Dn BP,and DEHP were calculated to be 2.90×10^-4 m/s,2.67×10^-4 m/s and 2.32×10^-4 m/s,respectively,by using theoretical empirical equations.（2）Assuming that PAEs in source material are in solid state,the enthalpy and entropy changes of conversion of PAEs from solid to gas state were calculated using theoretical empirical equations.Using the calculations,Van’t Hoff equation was used to obtain the relationship between the content of PAEs in source material（C₀）,y₀ and temperature（T）.y₀=C₀/exp（1.22×10⁴/T-16）,y₀=C₀/exp（1.32×10⁴/T-18）and y₀=C₀/exp（1.50×10⁴/T-25）is the resultant equations of Di BP,Dn BP and DEHP,respectively.（3）The concentrations of Di BP,Dn BP and DEHP,which were measured in more than 40homes in Beijing,were used to calculate the dust-air,PM_2.5-air and skin-air partition coefficients(K:K_d,K_p,and K_lg)for 3 PAEs.The median values of K_d（m³/mg）of Di BP,Dn BP,and DEHP in the summer were 0.036-0.151,0.021-0.036 and 1.479-4.069,respectively;the counterparts in the winter were 0.027-0.065,0.022-0.245 and 0.140-3.250,respectively.The median values of K_p（m³/μg）of Di BP,Dn BP,and DEHP in the summer were 0.053,0.015 and0.025,respectively;the counterparts in the winter were 0.011,0.004 and 0.021,respectively.The median values of log(K_lg)of Di BP,Dn BP,and DEHP were 7.654,7.932,7.265（summer）and 7.902,9.419,9.015（winter）,respectively.Compared to the partition coefficients（K^*）at equilibrium state that were calculated by theoretical empirical equations,it was found that the partition of PAEs between air and solid surfaces deviated from equilibrium state.The PAEs in PM_2.5 and house dust did not completely come from gaseous counterparts.（4）By using the obtained h_m,y₀ and K derived from measurements,a steady-state model was developed to predict the concentrations of DEHP in indoor air,PM_2.5,house dust and children’s skin surfaces.The predicted values in indoor air were 335.01-413.74（375.88,median,thereafter）ng/m³（summer）and 528.40-625.85（580.01）ng/m³（winter）;the counterparts in PM_2.5 were 82.08-101.37（92.09）ng/m³（summer）and 660.50-782.31（725.01）ng/m³（winter）;the counterparts in house dust were 489.79-604.88（549.54）ng/mg（summer）and 1025.10-1214.15（1125.22）ng/mg（winter）;the predicted values on skin surfaces were 773.65-955.44（868.03）ng/m²（summer）and 480.33-570.10（528.34）ng/m²（winter）.Unlike the steady-state models developed using K^*,the model developed in this dissertation yielded predictions that were compareable to measurements.Therefore,the method has a promising future.