Original scientific paper

Estimation of Properties of Triatomic Molecules from Tabulated Data Using Least-squares Fitting

Chris Carlson ; Southern College, Collegedale, TN 37315, USA
Julie Gilkeson ; Southern College, Collegedale, TN 37315, USA
Katie Linderman ; Southern College, Collegedale, TN 37315, USA
Saffron LeBlanc ; Southern College, Collegedale, TN 37315, USA
Ray Hefferlin ; Southern College, Collegedale, TN 37315, USA
Brad Davis ; Rinker-Davis Corporation, Belmont, CA 94002, USA

Abstract

This paper shows that it is feasible to make rapid forecasts of data
for large numbers of molecules by using least-squares smoothing
of tabulated data, though the forecasts are not as precise as those
from quantum-chemical computation packages which deal with one
molecule at a time. The molecules' properties were chosen to be of
value in the plasrna and astronomical physics. The work begins
with the graphical analysis of critically-analyzed data for ground
states of neutral, acyclic, main-group, row 2 to row 6, triatomic
molecules to infer a least-squares smoothing equation. The equation
is quadratic in a function (R1R2 + R2R3) of the atomic periodnumbers, quadratic in the group number of the central atom, and cubic in the total number of valence electrons. The coefficients of
the equation (some of them zero for some properties) were obtained
from high-quality tabulated data for the heat of atomization, ionization
potential, log of the partition function at 1000 K, and log of the partial-pressure equilibrium constant for the constituent atoms over the diatomic molecules at 1000 K. The equation and its coefficients
were tested by comparison with data, from the same tabulations,
for·a few molecules not in the original set. Finally, values were forecasted for 164, 145, 107, and 164 additional molecules, for
four the properties listed above and in order the same order.

Keywords

Hrčak ID:

176955

URI

https://hrcak.srce.hr/176955

Publication date:

1.5.1997.

Visits: 945 *