A novel QSPR model for prediction the dispersibility of graphene in various solvents
کد مقاله : 1118-INZC6
نویسندگان:
ابوذر خواجه *
گروه مهندسی شیمی، دانشگاه صنعتی بیرجند، بیرجند، ایران
چکیده مقاله:
The dispersibility of graphene in various solvents is an important factor for production and application of graphene. In this work, the dispersibility of graphene in 40 organic compounds were taken from Ref. [1], which was used as the data set. In order to variable selection and building a linear QSPR model, for predicting the dispersibility of graphene, the combining modified particle swarm optimization (MPSO) and multiple linear regression (MLR) [2,3] was used. The optimal linear QSPR model for prediction the dispersibility of graphene in organic solvents by using MPSO-MLR is as follows:

Log DG = - 0.7699 × Ke - 0.6657 × X2A - 0.5297 × MATS1m + 1.1567 × R3v - 0.2797 × G1m + 0.6702
N=40; R2= 0.904; R2adj =0.890; Q2LOO=0.875.

The analyses of selected descriptors indicated that the MATS1m is the most relevant descriptor in the model. MATS1m is a 2D autocorrelation descriptor, which describe when distributed atomic mass increases along topological molecular structure of solvents the dispersibility of graphene decreases.
کلیدواژه ها:
QSPR; Graphene; Dispersibility; MPSO-MLR.
وضعیت : مقاله پذیرفته شده است