introduce

Boiling point is an important physical and chemical parameter, which can be used for the description and identification of compounds, the measurement of volatility (the retention behavior of compounds on chromatographic columns, the taste of some industrial products or agricultural products and the toxicity of some chemicals are closely related to the volatility of their components) and the prediction of other physical parameters (such as critical temperature, flash point, enthalpy of evaporation, etc.). ).。 Karelson et al. reviewed the early use of physical and chemical descriptors and/or. Quantitative structure-activity relationship of topological descriptors (including boiling point prediction). The thermodynamic parameters of compounds can also be obtained from the correlation analysis between their chromatographic retention behavior and physical and chemical properties. Such as boiling point of alkylbenzene, relative vapor pressure of insect information compounds, dynamic viscosity of crude oil, simulated distillation of crude oil, etc. Lu et al. studied and summarized the change law of retention value of homologues on boiling point unipolar chromatographic columns. Based on this, the boiling point of homologues can be predicted by using the retention value of compounds, but it is difficult to establish a suitable prediction model when the composition system is complex. Based on the gas chromatographic retention time data of organic reagents on two different polar chromatographic columns, a unified prediction model for boiling points of complex systems of multicomponent compounds was established. By calculating the correlation coefficient and probability value of the prediction model based on single or different polarity chromatographic column retention time, the necessity of introducing different polarity chromatographic column retention time into the prediction model to improve the boiling point prediction effect is explained.

1

experiment

1. 1

Retention time and boiling point of organic reagents

See the literature for the gas chromatographic analysis conditions of 56 organic reagents. As a standard reference, volatile organic reagents were separated and detected on nonpolar (dimethylsiloxane, film thickness 5μm) and strongly polar (polyethylene glycol, film thickness 1 micron) capillary gas chromatography columns respectively. The retention time and boiling point of organic reagents on two kinds of chromatographic columns are shown in Table L, in which the data are arranged in descending order of retention time of organic reagents on nonpolar chromatographic columns, and the measured value of boiling point (BP) is obtained by Chemfinder.

1.2

Regression analysis calculation

Using multiple linear regression

The boiling point prediction model of chromatographic retention time and boiling point was established by MLR method.

Generally speaking, when the parameters used in MLR model increase, the correlation coefficient of the model can be improved; But only when the model has high correlation coefficient and small probability value, it makes sense to introduce more parameters into the model. The probability value of the model can be calculated according to the following formula:

Where r is the correlation coefficient of MLR model; θ is the radian value of the changing angle between 0 and π/2 (corresponding to R= 1 and R = 0)； respectively); N is the experimental data point used in modeling.

There are many more.

Give an address

/medicine/pro/experimental/ chromatogram/paper/2007-05-18-370434.shtml