Basis set construction

Review articles covering basis set construction and performance are... 

The original VB method has been difficult to use in practice because of the nonorthogonality of the atomic orbital basis, but there has been a revival of interest in it recently. Gallup and co-workers116-118 have described a new technique for carrying out such calculations, and the results of applications to the first-row hydrides. Calculations were carried out using a minimal basis set constructed from gaussian lobe orbitals. The orbitals were scaled to their best atom value and also optimally scaled in the molecule. Atomic populations were also computed.118... 

All calculations employed extended basis sets constructed from contracted Gaussian type orbitals. Basis sets were, in general, taken from the work of Huzlnaga (26) and Dunning and Hay (27). 

ANO Atomic Natural Orbitals. A basis set constructed from maximization of the... 

When calculating the wavefunction it is important to make a choice of basis set J. t that is suitable for the available computing power and the accuracy desired. A straightforward early approach to basis set construction was to fit an accurate Slater-type atomic orbital (STO) with n gaussians,called STO- G. The quality of STO- G wavefunctions increases as n increases. It was determined that = 3 was a good starting point, and the STO-3G basis set has been widely used, particularly where computing resources were limited or for lai er molecules. 

Exercise 3.4. Orthonormality in doubie-zeta basis sets—construction of the Lithium 2s radial orbital from the doubie-zeta Clementi basis. 

The purpose of this chapter is to review the progress which has been made over the past few years on the problem of reducing the error associated with basis set truncation in molecular calculations and in atomic calculations using the algebraic approximation or basis set expansion technique. It is clearly not possible, within the space available, to give a completely comprehensive account of all developments which have recently been made in the field of basis set construction, a field that forms the foundation upon which the vast majority of contemporary atomic and molecular electronic structure studies are based. The review is, therefore, necessarily selective but, nevertheless, should provide an up-to-date account of the most important aspects of current thinking on the basis set expansion method or algebraic approximation. 

In principle, the number of basis spinors m needs to be infinitely large for a complete basis and hence for an exact representation of a molecular spinor in this basis. For practical reasons, however, it must be as small as possible in order to keep the computational effort as low as possible. The basis set size should therefore be small but still allow for sufficiently accurate calculations. In order to achieve this, we need to exploit the physics of the problem to the largest extent a procedure in which the LCAO idea is the first step. Hence, we emphasize that the expansion of Eq. (10.2) provides an optimum description of an atomic spinor. Keeping in mind the idea of a minimal basis set constructed of atomic spinors, we may well freeze the coefficients and reduce the number m back to the smaller number m. Then, only the m coefficients djfc are to be determined rather than the 4m coefficients. To use fixed in a molecular calculation is known as using a contracted basis set. Various variants of such contractions are known but we shall not delve deeper into such purely technical issues. Basis functions that have not been contracted are called primitives or primitive basis functions. 

The choice and generation of basis sets has been addressed by many authors [190,192,528,554-563]. While we consider here only the basic principles of basis-set construction, we should note that this is a delicate issue as it determines the accuracy of a calculation. Therefore, we refer the reader to the references just given and to the review in Ref. [564]. In Ref. [559] it is stressed that the selection of the number of basis functions used for the representation of a shell riiKi should not be made on the grounds of the nonrelativistic shell classification nj/j but on the natural basis of j quantum numbers resulting in basis sets of similar size for, e.g., Si/2 and pi/2 shells, while the p /2 basis may be chosen to be smaller. As a consequence, if, for instance, pi/2 and p /2 shells are treated on the tijli footing, the number of contracted basis functions may be doubled (at least in principle). The ansatz which has been used most frequently for the representation of molecular one-electron spinors is a basis expansion into Gauss-type spinors. 

Linear dependencies of Gaussian-type orbital basis sets employed in the framework of the HF SCF method for periodic structures, which occur when diffuse basis functions are included in a basis set in an uncontrolled manner, were investigated [468]. The basis sets constructed avoid numerical linear dependences and were optimized for a number of periodic structures. The numerical AO basis sets for solids were generated in [469] by confining atoms within spheres and smoothing the orbitals so that the first and second derivatives go to zero at the boundary. This forms small atomic-like basis sets that can be applied to solid-state problems and are efficient for treating large systems. 

Other functions maybe more relevant for the other contributions. Various basis sets constructed in this way have been successfully used in ab initio and DFT calculations further improvement of convergence with the basis set extension is achieved reoptimizing the exponents and contraction coefficients (Benedikt et al. 2008 Jensen 2006). 

Formally, the system of equations (10) is the same as with the Roothaan-Hartree-Fock method, and also in its solution we find the same three steps selection of the basis set, construction of the matrix elements, and diagonalization of the Hamiltonian or Fock matrix. In practical implementation, however, the two methods differ considerably in all the three steps. Below we will comment on each step separately. 

In the long run, the most important products of this Nd study were the tools that were developed to help deal with basis set construction and jls restrictions placed on the 4f electron subgroup. We had reached that limit where necessity had forced us to invent a new approach in order to move further into the row. Having gone through the pains of developing... 

---