Use of Local Lagrange Interpolation for Calculation of Retention Indices in Linear Temperature-Programmed Gas Chromatrography A. Baaliouamer ! L. Boudarene / B. Y. Meklati* USTHB, Institut de Chimie, Laboratoire d'Analyse Organique Fonctionnelle Syst6matique, El-Alia, BP32, Bab Ezzouar, Alger, Algeria Key Words Capillary gas chromatography Retention indices Temperature programming Local Lagrange interpolation model Summary The influence of the isothermal temperature, program rate, initial temperature and flow rate on retention indices was studied. The methods of Kov~ts, Van Den Dool and Local Lagrange Interpolation are compared. Ten experimental measurements were carried out on a capillary column coated with OV-101 stationary phase. Introduction In gas-liquid chromatography, the Kov~its retention index system [1] plays an important role in the analysis of essential oils [2-5]. Its use is considered as comple- mentary to other identification techniques such as mass Spectrometry and Fourier transform infra-red spectros- Copy, both coupled to gas chromatography. The use of retention indices can solve, for instance, the problem of the similarity of mass spectra within the same family of compounds such as esters and monoter- pene or sesquiterpene hydrocarbons [4, 6]. However, there remains a problem: index differences, sometimes important, are noticed for a similar compound, on a similar stationary phase in different laboratories, espe- cially in temperature programming. Differences of 50 index units or more have been noted on polar columns Such as poly(ethylene glycol) 20 M [4, 5]; comparison of the values obtained in our laboratory with those in the literature confirmed this difference. These differences generally arise for tile following reasons [71: - Commercial stationary phases are not ultra-pure products and the degree of polymerization is not accurately known. - Experimental conditions vary from laboratory to laboratory, such as the amount of stationary phase in the column, carrier-gas flow-rate, length and diameter of the column, the initial and final temper- ature of the oven and the temperature program rate. - Peak overloading is shown by deformation of the peak shape and modifications in retention times. - Differences may exist between the temperature measured in the oven and the effective temperature in the column. Many studies [8-23] have been interested in the repeatability and reproducibility of retention indices in temperature programming. In the present study, we show for such data a new adaptation of the local Lagrange interpolation; studied previously in our laboratory [24, 25] for the isothermal calculation of retention indices. In the calculation of retention indices either the absolute retention time or its Napierian logarithm is used. Finally, the results obtained are compared to those of Kowits [1] and of Van den Dool and Kratz [3,101. Methods of Calculations Used Kov~its indices (I1) can be obtained by linear interpola- tion on a logarithmic scale of the adjusted retention times of homologous series (usually n-alkanes) [1, 26]: In t~.x -In tk(n) 11 = 100 n + 100 In t~( n + 1) -lntR(n) n : number of carbon atoms t~.x : adjusted retention time of sample ti~.(n) = adjusted retention time of n-alkane with n carbon atoms tR(n + 1) = adjusted retention time of n-alkane with n + 1 carbon atoms, Chromatographia Vol, 35, No. 1/2, January 1993 Originals 009-5893/93/1 0067-06 $ 03,00/0 9 1993 Friedr. Vieweg & Sohn Verlagsgesellschaft mbH 67