.. tip::
    All input files can be downloaded: :download:`Files <https://zhjun-sci.com/qbics/input_library/keywords-grimmedisp.tar.gz>`.

grimmedisp
===========

.. contents::
   :local:

This keyword defines how to apply Grimme dispersion correction version 3, i.e. DFT-D3.

Options
------------

.. hint:: 
   
   **Always use DFT-D3 in your DFT calculations,** especially for weak interactions. Actually, it is recommended to use it in all cases.

When there is no ``grimmedisp`` in the input file, no DFT-D3 will be applied.

.. option:: type

   .. list-table::
      :stub-columns: 1
      :widths: 5 20
   
      * - Value
        - ``bj`` Becke-Johnson DFT-D3 form
      * - 
        - ``zero`` Zero-damp DFT-D3 form
      * - Default
        - None
      
   Define form of DFT-D3. There is no default value and you must assign ``bj`` or ``zero``. 
   
   In most cases, ``bj`` is better, but for some functionals, like M06, M062X, and M06HF, only ``zero`` is available.

.. option:: three_body
   
   Add three body corrections in DFT-D3. This is useful for large systems.

.. option:: tz
   
   If triple zeta basis sets are used, this keyword may bring some improvement.
      
Theoretical Background
------------------------

The DFT-D3 method is a dispersion correction to the DFT energy. It is based on the pair-wise summation of the damping functions. In modern DFT calculations, it is almost **always beneficial** to include DFT-D3 correction. So, it is recommended to use it in your **all** calculations.


Input Examples
----------------

Examples: Weak Interactions in CH\ :sub:`4`-C\ :sub:`2`\ H\ :sub:`6`\ 
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This example shows the importance of the DFT-D3 correction in the calculation of weak interactions. For CH\ :sub:`4` and C\ :sub:`2`\ H\ :sub:`6`\ , both are nonpolar, and the interaction between them is very weak. To calculate the interactions, we simply use EDA method at B3LYP+D3BJ/def2-svp level of theory:

.. code-block:: bash
    :linenos:
    :caption: disp-1.inp

    mol
        C     -0.99275967      0.12491197      0.13574919
        H     -1.00662206     -0.95659619      0.18835421
        H     -0.34007185      0.51620582      0.90571860
        H     -0.63121516      0.43418579     -0.83650623
        H     -1.99486438      0.50482451      0.28603834
        C      2.95955795     -0.58352084     -0.35430634
        H      3.53618084     -1.49857448     -0.47007336
        H      3.33520181     -0.05007297      0.51638813
        C      3.07143559      0.28355041     -1.59942759
        H      1.92316063     -0.85399112     -0.16804950
        H      4.10875722      0.55492780     -1.78655076
        H      2.69486480     -0.24574724     -2.47236370
        H      2.49506428      1.19877654     -1.48025100
    end
    
    basis
        def2-svp
    end
    
    scf
        charge  0
        spin2p1 1
        type    U  # For EDA, this must be set explicitly.
    end
    
    grimmedisp
        type bj
    end
    
    eda
        type tso
        frag 0 1 1-5   # Define CH4.
        frag 0 1 6-13  # Define C2H6.
    end
    
    task
        eda b3lyp
    end

The output is:

.. code-block:: bash
    :linenos:
    :caption: disp-1.out

    WITH BSSE correction:
    Electrostatic interaction energy:                  -0.13 kcal/mol
    Exchange-correlation interaction energy:            0.48 kcal/mol
    Polarization interaction energy:                   -0.00 kcal/mol
    Charge transfer interaction energy:                -0.11 kcal/mol
    Grimme s dispersion interaction:                   -0.87 kcal/mol
    ----------------------------------------------------------------
    Total interaction energy:                          -0.62 kcal/mol

Thus, the interaction energy between CH\ :sub:`4` and C\ :sub:`2`\ H\ :sub:`6`\ is ``-0.62 kcal/mol``. The dispersion energy is ``-0.87 kcal/mol``, which is the most important part of the interaction energy. Without DFT-D3, the interaction energy would be ``-0.62``-``-0.87`` = ``+0.25`` kcal/mol! Therefore, the complex is unbound without DFT-D3! This leads to a **quantitative error** in the interaction energy. Therefore, **always use DFT-D3 in your DFT calculations.**