Abstract
A statistical study between ‘Mesophase Lower Transition Temperatures’ (MLTTs) and their structural properties is carried out to understand the effect of structural behaviour on mesomorphic property.
Introduction: To establish a “Quantitative Structure and Property Relationship (QSPR) model” a set of randomly selected thirty-nine mesomorphic compounds is constructed. The backward stepwise regression analysis method is used to find out the good correlation between the “Mesophase Lower Transition Temperatures (MLTTs)” data set and “physical descriptors” like AMR, bpol, ASP-0, DELS, SdssC, etc. Physical descriptors are selected based on their good r2-values and p-values with respective MLTTs. The derived QSPR equation shows a good correlation between structural properties and mesomorphic properties of compounds.
Methods: Validation of the derived QSPR equation is carried out on the test series of eight compounds. The MLTTs of these compounds are predicted through the statistically derived QSPR equation and then compared with experimentally measured MLTTs. The average percentage error observed between predicted MLTTs and experimentally measured MLTTs is 10.95 % for all the thirty-nine compounds of the trial set and 10.64% for 8 compounds of the test series, respectively.
Results & Discussion: A low average percentage error suggests a reasonably acceptable degree of accuracy of the generated QSPR model to predict MLTTs of the compounds having a similar type of structure. In the present study, not only MLTTs are predicted, but an effort is also made to predict “Latent Transition Temperatures” (LTTs) of some nonmesomorphic compounds from the derived QSPR equation.
Conclusion: This computational study gives an insight into developing new QSPR models for the different type of liquid crystals homologous series, through which various types of mesomorphic properties, like mesomorphic thermal stability, mesomorphic upper transition temperature, mesophase length, phase behaviour, etc. can study and predict.
Keywords: Physical descriptors, prediction, QSPR, regression analysis, isotropic, turbid, transition temperatures.
Graphical Abstract
5