Semi-empirical quantum chemistry methods are based on the Hartree-Fock formal­ism, but make many approximations and obtain some parameters from empirical data. They are very important in computational chemistry for treating large mole­cules where the full Hartree-Fock method without the approximations is too expen­sive. The use of empirical parameters appears to allow some inclusion of electron correlation effects into the methods. Within the framework of Hartree-Fock calcu­lations, some pieces of information (such as two-electron integrals) are sometimes approximated or completely omitted. As with empirical methods, we can distinguish if: These methods exist for the calculation of electronically excited states of poly­enes, both cyclic and linear.

AM1 is basically a modification to and a reparameterization of the general theo­retical model found in MNDO. Its major difference is the addition of Gaussian func­tions to the description of core repulsion function to overcome MNDO’s hydrogen bond problem. Additionally, since the computer resources were limited in 1970s, in MNDO parameterization methodology, the overlap terms, Ps and Pp, and Slater orbital exponent’s Zs and Zp for 5- and p — atomic orbitals were fixed. That means they are not parameterized separately just considered as Ps = Pp, and Zs = Zp in MNDO. Due to the greatly increasing computer resources in 1985 comparing to 1970 s, these inflexible conditions were relaxed in AM1 and then likely better parameters were obtained.

Optimization of the original AM1 elements was performed manually by Dewar using chemical knowledge and intuition. He also kept the size of the reference pa­rameterization data at a minimum by very carefully selecting necessary data to be used as reference. Over the following years many of the main-group elements have been parameterized keeping the original AM1 parameters for H, C, N and O un­changed.

Of course, a sequential parameterization scheme caused every new parameter­ization to depend on previous ones, which directly affects the quality of the results. AM1 represented a very considerable improvement over MNDO without any in­crease in the computing time needed.

AM1 has been used very widely because of its performance and robustness com­pared to previous methods. This method has retained its popularity for modeling organic compounds and results from AM1 calculations continue to be reported in the chemical literature for many different applications.