RPA

Codebase to calculate bare susceptiblity of a lattice model (using TRIQS), and then perform general random phase approximation (RPA) on it to get instabilities of the model.

Installation

• Install the TRIQS and triqs_tprf package in a Python venv (with default name "triqs"). On Ubuntu there are precompiled packages. Otherwise this requires compilation described here.
• Add the julia repo locally in Package mode.

Usage

• Create the free Hamiltonian first. This is done using the TightBindingToolbox.jl interface. Some examples are given in ./examples/models.
• Define the interactions which are to be used for the RPA. This can also done in the same interface, with examples given in ./examples/interactions.
• Write a parent input file with details of parameters required for the calculation such as the inverse temperature, number of k-points, and number of matsubara frequencies. Refer to ./Inputs for formatting.
• Run the following command

  julia --project=../Project.toml --heap-size-hint=4G --input="../Inputs/name_of_input.yml" --run_bare=true

Input Notes

• Preferred scan input is fillings, with either explicit values:

  fillings:
    values: [0.2, 0.4, 0.6]

  or a compact range:

  fillings:
    min: 0.0
    max: 1.0
    n: 51

• Legacy mus inputs are still supported, and are converted to fillings during the bare stage.
• k_points can be either explicit coordinates or HighSymPoints labels understood by the model Brillouin zone. Example mixed format:

  k_points:
    - G
    - [0.666666666667, 0.333333333333, 0.0]
    - M2
  k_points_labels: ["\\Gamma", "K_1", "M_2"]

• If k_points_labels is omitted, labels are inferred from k_points.