Dear Pratap,

Empirical evidences are the outcome of an experiment and hence the empirical result is a unified confirmation. But the, semi-empirical methods uses part of experimental results; approximations... Applying full Hartree–Fock method without any approximations is expensive due its more computing time. In applying these approximation the guidelines given in the Gaussian software will be useful.

Semiempirical Methods such as AM1, PM3, MNDO semiempirical methods replace as many calculations as possible with pre-calculated data that are an integral part of the program. This speeds up calculation to such an extent that even large molecules can be calculated.

For organic molecules, accurate structures can be calculated within minutes with the AM1 method (M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart, J. Am. Chem. Soc. 1985, 107, 3902-3909.

The PM3 method is more accurate than the AM1 method and is also parametrized for many transition metals. An even further improvement has been its successor the PM5 method.

It gives up to four times more accurate heats of formation than the other semi empirical methods AM1, PM3, and MNDO.

Many inorganic molecules can be calculated with the MNDO method 

Dear Sivakumar Sir,

                                     Thanks a lot for your time and effort. Actually I am not exposed to/ familiar with Hartree-Fork method and all. As a layman (in theoretical/ computational chemistry) what I understood from your explanation is semi empirical method needs experimental data as well as some approximations in contradiction to empirical method where only experimental data is enough. Is that what you are trying to say.

Hi Pratab

Semi-empirical method is a cheap method because the calculation is done using experimental data to get the results. For example, Hartree Fock (HF) method is purely theoretical, hence, HF calculate the bond distance and further properties. On the other hand, semi-empirical method has experimental information preloaded about bond distance and other properties, therefore it do not need to calculate them. For that rason, the semi-empirical method is very fast that you can calculate organic molecules bigger in little time. The experimental information has been preloaded by the performer of the method, hence, you do not need to have the experimental data for the using the semi-empirical method. I think the PM7 method is the best and you can use it with MOPAC software.

Best Regards.

Dear Rody Sir, Thank you very much.

Very good explanation about Semi-empirical methods

Excellent

Semi Empirical metods are considered as reference for experiential mock ups in our engineering models.

very good

From the above explanations, regarding your Good Question-

How semi empirical method is different from empirical method. Is there any guidelines/ or order of preference to use one of the method, i mean how to find the suitability of one method over other?

The second part of the Question- - - - how to find the suitability of one method over other?

Only from the Experimental part with fewer errors.

Adequate Knowledge Gained from these conversation

In addition to the ongoing question, I would like to add this to the questions: Some people tend to condemn the use of semi-empirical method in molecular modeling, is that supposed to be a good practice? Sankha Chakrabortty Bojja Rajeshwar Rao SivaKumar Krishnamoorthy Rody Soto-Rojo Pratap Chhotaray Thank you

please fallow enclosed article briefly.

Thank you for that resources. It is simple, clear, and straightforward but I would appreciate it if you share your opinion on my added question: "Some people tend to condemn the use of the semi-empirical method in molecular modeling, is that supposed to be a good practice?" Bojja Rajeshwar Rao

Toyese Oyegoke, Some people tend to condemn the use of the semi-empirical method in molecular modeling, is that supposed to be a good practice?" it is 100% correct.

So, if i get you well, Bojja Rajeshwar Rao you meant that the use of the semi-empirical method in molecular modeling should be condemned? I would like you to state your reasons either for or against it. Thank you

Semi-emperical method is recommendable. To me most emperical methodology are based on hypothetical guess and assumptions. Semi-emperical will clear doubts. Well in the field of Engineering methodology is different ball game. We need facts between laboratory and full plant experimental analysis. Thus, semi-emperical method will reduce cost of full experimentation, create a synergy and ensure clarity to checkmate experimental lapses arising from emperical studies.

In semiempirical method, the core electrons are excluded from the calculations and a minimal basis set is applied. The omitted values are obtained by fitting the results to experimental data or ab initio calculations.