A size-segregated gas/particle partitioning coefficient K_Pi was proposed and evaluated in the predicting models on the basis of atmospheric polybrominated diphenyl ether (PBDE) field data comparing with the bulk coefficient K_P. Results revealed that the characteristics of atmospheric PBDEs in southeast Shanghai rural area were generally consistent with previous investigations, suggesting that this investigation was representative to the present pollution status of atmospheric PBDEs. K_Pi was generally greater than bulk K_P, indicating an overestimate of TSP (the mass concentration of total suspended particles) in the expression of bulk K_P. In predicting models, K_Pi led to a significant shift in regression lines as compared to K_P, thus it should be more cautious to investigate sorption mechanisms using the regression lines. The differences between the performances of K_Pi and K_P were helpful to explain some phenomenon in predicting investigations, such as Po Land K_OA models overestimate the particle fractions of PBDEs and the models work better at high temperature than at low temperature. Our findings are important because they enabled an insight into the influence of particle size on predicting models.

在大气多溴二苯醚（PBDE）现场数据与体积系数K _P的比较基础上，提出了尺寸分离的气体/颗粒分配系数K _{P i}并在预测模型中进行了评估。结果表明，上海东南农村地区大气多溴二苯醚的特征与以前的调查基本一致，这表明该调查代表了大气多溴二苯醚的当前污染状况。K _{P i}通常大于体K _P，表明TSP被高估了在散装的表达（总悬浮颗粒的质量浓度）ķ _P。在预测模型中，与K _P相比，K _{P i}导致回归线发生重大变化，因此使用回归线研究吸附机理应更加谨慎。K _{P i}和K _P的性能差异有助于解释某些预测研究中的现象，例如P o Land K _OA。这些模型高估了多溴二苯醚的颗粒分数，并且该模型在高温下的效果要好于在低温下的效果。我们的发现很重要，因为它们有助于洞察粒径对预测模型的影响。