Font Size: a A A

Study On The Improvement Of Pp-LFERs For Soil Organic Carbon-water Distribution Coefficients

Posted on:2016-06-18Degree:MasterType:Thesis
Country:ChinaCandidate:H Y GuFull Text:PDF
GTID:2283330461995651Subject:Groundwater Science and Engineering
Abstract/Summary:PDF Full Text Request
The sorption behavior of organic chemicals on soil or sediment is one of the major factors affecting their environmental distribution. It is significant to measure or evaluate the sorption coefficient accurately when predicating its fate in the environment. However, the workload will be incredible if the sorption coefficient of every contaiment in the environment is measured in the lab,polyparameter linear free energy relationships(pp-LFERs) model is an efficient and economic method for evaluating the adsorption/distribution coefficient. But there are still some arguments about the molecular structural parameters in the model. Based on precious researches on the pp-LFERs,a comprehensive analysis on the molecular structure parameters(E, S, A, B, V, and L)was conducted and the EV model and V model were developed. The models are suitable for predicting Log Koc of several kinds of organic compounds, including high fluorinated and polydimethylsiloxane.The conclusions are as follows:(1)The molecular structure parameter, S, represents polarity or polarizability of compounds. There is still argument on the properties of this parameter, i.e., whether it is a parameter representing specific interaction or non-specific interaction.The analysis in the paper shows that it represnts non-specific interaction. Firstly, the term S describes orientation force and induction force which attribute to the van der Waals, while the van der Waals belongs to the non-specific interaction. Secondly, the principal componen analysis of 649 compounds shows that the term S’ physical force is the same as E, L, and V, while these three are the non-specific interaction parameters. Lastly, because different solute descriptors ought to be independent from each other, the system constants s, a, and b should not be affected substaintially. However, the fact is that the change of constant s varies from 2 to 11 times. So S is proved to be the non-specific interaction parameter.(2) The Mc Gowan volume parameter V and the hexadecane-gas partition constant L are highly inter-correlated in LV model(including S, A, B, L, and V). The system constant l and v are changed anomaly(V increases by 5.69 times and L by 31 times) when the LV model of the same partition coefficient is developed with different database. indicating that the instability of the model is caused by the inter-correlation of L and V. The inter-correlation between L and V can be relieved by adding the PFCs compounds and the organosilicons in the database, because their polarizability is low. In addition, the L values of the PFCs compounds and the organosiliconsis are lower than other organic chemicals that have the same volume V.(3) By data analysis, such as, correlation coefficient( r2), standard error(SE) and the value of Akaike Information Criterion(AIC), it e can be conclude that the V model(including S, A, B, and V) is prefer to L model(including S, A, B, and L) when it is applied to predict the condensate–liquid distribution coefficient. The data of V is easier to be obtained than that of L. So the V model is more applicable than EV model. Though the results of the EV model(including E, S, A, B, and V) is more accurate, the data of E is hard to be obtained for some compounds such as solid organics. In this case, the V model can be taken into consideration when the parameter E is absent.(4) By multiple linear regression method, the EV model and V model were developed with 132 compounds, then a test with the other 33 compounds was conducted. As a result, the EV model and V model are stable and have a good fitting ability. Therfore, they can be used to predict the Koc of unknown compounds. According to the research on the AD of the model, the high fluorinated and polydimethylsiloxane have significant influence on the model, but the forecasting error is within 2.5 times SE. It proves to be the model can be extrapolated.The model covers a much wider range of structure descriptor values:E:-1.56 to 2.6; S:-0.12 to 2.12; A: 0 to 0.73; B: 0 to 1.53; V: 0.31 to 3.52. When the compound is within the scope of structure parameters, the EV model or V model can be used to predict the Koc.At last, the adsorption mechanism was also discussed from the point of molecular interaction and it shows that cavity formation(v V) and dispersion interaction(e E) are the major driving forces for the hydrophobic compounds, while it is cavity formation and hydrogen-bonding(a A,b B) for polar chemicals.
Keywords/Search Tags:Sorption, Polyparameter linear free energy relationships model, Molecular structure parameters, Organic carbon-water distribution coefficients(Koc)
PDF Full Text Request
Related items