SphericalHarmonicFunctions.h

Go to the documentation of this file.

#include <vector>
 
namespace dftefe
{
  namespace atoms
  {
    void
    convertCartesianToSpherical(const std::vector<double> &x,
                                double &                   r,
                                double &                   theta,
                                double &                   phi,
                                double                     polarAngleTolerance);
 
 
    //
    // We use the real form of spherical harmonics without the Condon-Shortley
    // phase (i.e., the (-1)^m prefactor) (see
    // https://en.wikipedia.org/wiki/Spherical_harmonics) NOTE: 1) The wikipedia
    // definition has the Condon-Shortley phase.
    //       2) The definition of the associated Legendre polynomial (P_lm) in
    //       Boost library also contains a Condon-Shortley phase.
    //          Thus, if you're using Boost library, multiply the P_lm
    //          evaluation with (-1)^m to remove the Condon-Shortley phase. Most
    //          Quantum Chemistry codes (e.g., QChem) do not include the
    //          Condon-Shortley phase. So to make it consistent, it is prefered
    //          to remove the Condon-Shortley phase, if there are any to begin
    //          with.
    //        3) From C++17 onwards, the <cmath> has the associated Legendre
    //        polynomial (see
    //        https://en.cppreference.com/w/cpp/numeric/special_functions/assoc_legendre)
    //           Thus, if you're using C++17 or beyond, you can use the C++
    //           standard's definition of associated Legendre polynomial instead
    //           of Boost. Note that, the C++ standard does not have the
    //           Condon-Shortley phase while Boost has it. So, we do not have to
    //           do anything special to remove it while using the C++ standard.
    //
 
    //
    // Y_lm(theta, phi) = Clm(l,m) * Dm(m) * P_lm(l,m,cos(theta)) * Qm(m,phi),
    // where theta = polar angle,
    // phi = azimuthal angle
    // P_lm is the associated Legendre polynomial of degree l and order m,
    // Qm is the real form of exp(i*m*phi),
    // C_lm is the normalization constant for P_lm,
    // D_m is the normalization constant for Q_m
    //
 
    //
    // For the definition of the associated Legendre polynomials i.e. P_lm and
    // their derivatives (as used for evaluating the real form of spherical
    // harmonics and their derivatives) refer:
    // @article{bosch2000computation,
    //     title={On the computation of derivatives of Legendre functions},
    //         author={Bosch, W},
    //        journal={Physics and Chemistry of the Earth, Part A: Solid Earth
    //        and Geodesy}, volume={25}, number={9-11}, pages={655--659},
    //        year={2000},
    //        publisher={Elsevier}
    //       }
    // We use the derivative definitions from the above reference because
    // finding the derivatives on the pole (i.e., theta = 0) is tricky. This is
    // because the azimuthal angles (phi) is undefined for a point on the pole.
    // However, the derivative is still well defined on the pole via the
    // L'Hospital's rule. However, one can avoid implementing tedious
    // L'Hospital's rule on pole and use much simpler expressions given in the
    // above reference.
    //
 
    double
    Dm(const int m);
 
    double
    Clm(const int l, const int m);
 
    double
    Qm(const int m, const double phi);
 
    double
    dQmDPhi(const int m, const double phi);
 
    double
    Plm(const int l, const int m, const double x);
 
    double
    dPlmDTheta(const int l, const int m, const double theta);
 
    double
    d2PlmDTheta2(const int l, const int m, const double theta);
  } // namespace atoms
} // namespace dftefe