4.1.6. Parameters for the dynamical Green’s function

• CalcSpec

  Type : String(choose from "None", "Normal", "NoIteration", "Restart_out", "Restart_in", "Restart". "None" as default.)

  Description : The condition for the calculation of the dynamical Green’s function is specified. "None" for omitting the calculation of the dynamical Green’s function. "Normal" for calculating that function from scratch, "NoIteration" for calculating that function with the same iteration in the previous run (In this case, the Hamiltonian-vector product is not performed. Although the numerical cost is very small, the convergence is not guaranteed), "Restart_out" for calculating that function from scratch and writing the restart-file at the end, "Restart_in" for starting the calculation with the previously written restart-file, "Restart" for "Restart_out" + "Restart_in".

  The scheme for the spectrum calculation is specified by using the parameter method. If method="CG" is chosen, the shifted bi-conjugate gradient method [1] together with the seed-switch technique [2] is employed with the help of the \(K\omega\) library [3].

• SpectrumType

  Type : String (choose from "SzSz", "S+S-", "Density", "up", "down". "SzSz" as default.)

  Description : The type of the dynamical Green’s function to be computed is specified. "SzSz" for \(\langle {\hat S}_{z q} {\hat S}_{z q}\rangle\), "S+S-" for \(\langle {\hat S}^{+}_{q} {\hat S}^{-}_{q}\rangle\), "Density" for \(\langle {\hat n}_{q} {\hat n}_{q}\rangle\), "up" for \(\langle {\hat c}^{\dagger}_{q \uparrow} {\hat c}_{q \uparrow}\rangle\), "down" for \(\langle {\hat c}^{\dagger}_{q \downarrow} {\hat c}_{q \downarrow}\rangle\).

• SpectrumQW, SpectrumQL

  Type : Double (default value: 0.0)

  Description : The wave number (Fractional coordinate) of the dynamical Green’s function is specified. The reciprocal lattice vector is computed from the direct lattice vector shown in Fig. 4.1 , Fig. 4.2 , Fig. 4.4 , Fig. 4.3 .

• OmegaMin

  Type : Double (-LargeValue times the number of sites as default.)

  Description : The lower limit of the real part of the frequency.

• OmegaMax

  Type : Double (LargeValue times the number of sites as default.)

  Description : The upper limit of the real part of the frequency.

• OmegaIm

  Type : Double (0.01*LargeValue as a default.)

  Description : The imaginary part of the frequency.

• NOmega

  Type : Positive integer (200 as a default.)

  Description : The number of frequencies.

[1]	A. Frommer, Computing 70, 87{109 (2003).

[2]	S. Yamamoto, T. Sogabe, T. Hoshi, S.-L. Zhang, T. Fujiwara, Journal of the Physical Society of Japan 77, 114713 (2008).

[3]	https://github.com/issp-center-dev/Komega.