Thread Content
How to optimize molecular structure, taking the benzene molecule C6H6 as an example. 1) Atoms are distinguishable; the distances between atoms in many molecular substances can be measured with a high degree of precision, and this represents very important data ； 2) For atoms of a specific type, such as carbon atoms, the number of electrons outside the nucleus is fixed, at 6. H is 1 extranuclear electron. 3) According to the principle of indistinguishability, extranuclear electrons are indistinguishable from one another. However, based on the periodic law of elements and the analytical solution for the SE of the hydrogen atom, it is believed that electrons are distributed along certain orbitals; therefore, wave functions are fitted using these orbitals. The most typical examples of such orbitals are Slater orbitals and Gaussian orbitals. The former has a clear physical meaning and includes a normalization factor (ensuring that the electron must exist within that orbital); n is the principal quantum number, *ta is the orbital exponent, and Y is the spherical harmonic function. The latter is easy to calculate; alpha is the orbital index, and L=i+j+k. Values of L equal to 0, 1, and 2 correspond to s, p, and d orbitals respectively. Generally speaking, an orbit requires at least one function to describe it; the more functions, the more accurate the description. In reality, the inner orbits are composed of three or four functions, while the outer orbits are made up of dozens to hundreds of such orbits. 4) The orbits of electrons in atoms, commonly referred to as atomic orbitals, have now been calculated. For a specified software, specified atoms, specified basis sets, and specified orbital types, the various exponential constants are fixed. This is also one of the reasons why software is expensive. 5) After the atoms have been calculated, the molecules begin to be calculated. For specific molecules, the distances between atoms within the molecule can be measured. All electrons in a molecule are indistinguishable, or in other words, all electrons in a molecule are identical. So how is the new orbit fitted? The LCAO assumption was proposed, namely that molecular orbitals are linear combinations of atomic orbitals. Fortunately, this assumption does indeed bring computational convenience and accuracy. 6) Then the optimization of the molecule becomes simple. Since for a given software, given atoms, given basis set, and given orbital type, the various exponential constants are fixed. Moreover, molecular orbitals are linear combinations of atomic orbitals. Well, the optimization of the molecular structure simply involves adjusting the distances between atoms, with each atom’s orbitals being fitted using orbitals obtained through LCAO combination ； 7) In summary, since atoms can move in three spatial directions, spherical harmonic functions are suitable for representing local coordinates; therefore, using absolute coordinates is inappropriate. Logically, for N atoms, 3^N coordinate variables are required to describe them, as the wave function is the product of the wave functions of each individual atom. However, due to LCAO, 3N coordinate variables are required to describe the total wave function. 8) It can be seen that the computational cost is closely related to the number of atoms N; if symmetry exists, the computational cost can be significantly reduced, which is why the molecular point group is extremely important ； 9) As a supplementary note on basic set knowledge, the minimum basis set is also known as the STO-3G basis set. STO is an abbreviation for Slater-type atomic orbitals, while 3G indicates that each Slater-type atomic orbital is obtained as a linear combination of three Gaussian-type functions. The STO-3G basis set is the smallest compressed Gaussian basis set. There are also cleaving basis sets and polarization basis sets. As in the basis set represented by 6-31G, each inner-shell electron orbital is a linear combination of 6 Gaussian functions, while each valence-shell electron orbital is split into two basis functions, which are respectively linear combinations of 3 and 1 Gaussian function. In the Gaussian function, the variable α has a significant impact on the shape of the function. When α takes large values, the function graph tends to gather near the origin; whereas when α takes small values, the function graph spreads out in directions away from the origin. Such Gaussian functions with very small α values are known as diffused functions. 10) After molecular optimization, it is possible to calculate energy, transition states, and so on. 11) Of course, by calculating energy, it is possible to determine the number of states, as well as calculate entropy and Gibbs free energy, etc.