Merge pull request #212 from mir-group/minorFixes

Update outputs for EPA app
phoebe-team · Apr 26, 2024 · f69a90b · f69a90b
2 parents 7f6570f + 433d544
commit f69a90b
Show file tree

Hide file tree

Showing 39 changed files with 1,847 additions and 338 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -8,6 +8,7 @@ cmake_minimum_required(VERSION 3.16)
 # set cpp standard flags
 set(CMAKE_CXX_STANDARD 17)
 set(CMAKE_CXX_EXTENSIONS OFF)
+cmake_policy(SET CMP0144 NEW)
 
 project(Phoebe LANGUAGES C CXX Fortran VERSION 1.0)
 

diff --git a/README.md b/README.md
@@ -26,21 +26,20 @@ If you feel you've found a bug or seen some unexpected behavior, please let us k
 ### Current functionalities
 #### Electronic Transport
 
-   * Electron-phonon and phonon-electron scattering by Wannier interpolation
+   * Electron-phonon and phonon-electron scattering rates by Wannier interpolation
    * Electron-phonon scattering within the electron-phonon averaged (EPA) approximation
-   * Polar correction and boundary scattering contributions to transport
    * Electronic transport coefficients (mobility, conductivity, thermal conductivity, and Seebeck coefficient)
 
 #### Phonon Transport
 
-   * 3-phonon scattering from thirdOrder.py/ShengBTE or Phono3py force constants
-   * Boundary and isotope scattering contributions to transport
-   * Phonon (lattice) thermal conductivity
-   * Lattice thermal conductivity calculations including ph-el lifetimes
+   * Phonon (lattice) thermal conductivity, including: 
+     * 3-phonon scattering from thirdOrder.py/ShengBTE or Phono3py force constants
+     * Boundary, isotope, and phonon-electron scattering contributions
+     * Lattice thermal conductivity calculations including both ph-ph and ph-el scattering 
 
 #### And more...
 
-   * BTE solutions by RTA, iterative, variational, and relaxons solvers
+   * BTE solutions by RTA, iterative, variational, and relaxon solvers
    * Calculation of electron and phonon linewidths or relaxation times on a path
-   * Wigner transport equation correction for electrons and phonons (Zener tunneling contribution to electron transport)
+   * Wigner transport equation correction for electrons and phonons 
    * Hydrodynamic transport properties (viscosity) for electrons and phonons
diff --git a/doc/sphinx/source/inputFiles.rst b/doc/sphinx/source/inputFiles.rst
@@ -135,6 +135,8 @@ Phonon BTE Solver
 
 * :ref:`dimensionality`
 
+* :ref:`thickness`
+
 * :ref:`constantRelaxationTime`
 
 * :ref:`withIsotopeScattering`
@@ -222,6 +224,8 @@ Electron BTE Solver
 
 * :ref:`dimensionality`
 
+* :ref:`thickness`
+
 * :ref:`constantRelaxationTime`
 
 * :ref:`convergenceThresholdBTE`
@@ -295,6 +299,10 @@ EPA Transport
 
 * :ref:`epaEnergyRange`
 
+* :ref:`dimensionality`
+
+* :ref:`thickness`
+
 * :ref:`kMesh`
 
 * :ref:`temperatures`
@@ -482,7 +490,7 @@ Electron Lifetimes on a Path
 
 -----------------------------------
 
-Phonon Dos
+Phonon DoS
 ----------
 
 **Functionality:** Compute the phonon density of states.
@@ -1114,7 +1122,7 @@ convergenceThresholdBTE
 dimensionality
 ^^^^^^^^^^^^^^
 
-* **Description:** Input the dimensionality of the material. As a result, transport coefficients tensors will be of size (dim x dim), and units will be suitably scaled for the desired dimensionality.
+* **Description:** Input the dimensionality of the material. As a result, transport coefficient tensors will be of size (dim x dim), and units will be suitably scaled for the desired dimensionality. For 2D materials, note that Phoebe assumes the material is oriented in x-y, with height in z. 
 
 * **Format:** *int*
 
@@ -1123,6 +1131,20 @@ dimensionality
 * **Default:** `3`
 
 
+.. _thickness:
+
+thickness
+^^^^^^^^^^^^^^
+
+* **Description:** Input the thickness of a 2D material. This is only used if :ref:`dimensionality` = 2, and results in a scaling factor of (cell height/thickness) applied to the transport results. 
+
+* **Format:** *double*
+
+* **Required:** no
+
+* **Default:** `1.`
+
+
 .. _constantRelaxationTime:
 
 constantRelaxationTime

diff --git a/doc/sphinx/source/introduction.rst b/doc/sphinx/source/introduction.rst
@@ -3,7 +3,7 @@ Introduction
 
 Phoebe is a software for the ab-initio prediction of transport properties, such as electron-phonon limited conductivity and phonon thermal conductivity.
 
-Currently, we support integration with the ab-initio software suite Quantum ESPRESSO for electron-phonon properties and anharmonic force constants through ShengBTE. Phoebe also accepts anharmonic force constants from Phono3py, which interfaces with most mainstream DFT packages. Support for more coming soon.
+Currently, we support integration with the ab-initio software suite Quantum ESPRESSO for electron-phonon properties and anharmonic force constants through ShengBTE and Phono3py. 
 
 Additionally, Phoebe is written in C++ and designed for use on modern HPC infrastructures through hybrid MPI/OpenMP parallelization, distributed memory computation via ScaLAPACK, and support for GPU acceleration using Kokkos.
 
@@ -20,17 +20,19 @@ J. Phys. Mater. 5 035003. (2022).
 
 * Electron-phonon limited electronic transport coefficients, including the electrical conductivity, Seebeck coefficient, and electronic thermal conductivity
 
-* Lattice thermal conductivity, including effects from 3-phonon scattering, phonon-isotope and phonon-boundary scattering
+* Lattice thermal conductivity, including contributions from 3-phonon, phonon-isotope, phonon-electron, and phonon-boundary scattering
 
-* Electron-phonon scattering through Wannier interpolation, or EPA approximation
+* BTE solutions by RTA, iterative, variational, and relaxon solvers
 
-* Calculation of electron and phonon lifetimes (linewidths)
+* Electron-phonon transport via Wannier interpolation or EPA approximation
 
-* Electron and phonon viscosities
+* Wigner transport equation correction for electrons and phonons 
+
+* Calculation of electron and phonon lifetimes/linewidths (including projected onto a band path)
 
-* Plots of electronic band structure and phonon dispersion relations
+* Electron and phonon viscosities
 
-* Plots of electron or phonon density of states
+* Plots of electron band structures/phonon dispersion, electron and phonon density of states
 
 This documentation contains a brief description of the formalism used to compute these quantities. We expect that the user is already familiar with ab-initio codes and with the fundamentals of solid state physics. A good introduction to the topic is the book "Electrons and Phonons: The Theory of Transport Phenomena in Solids" by Ziman. In the theory section, we add references to contemporary literature for the latest method/algorithm developments.
 

diff --git a/doc/sphinx/source/postProcessing.rst b/doc/sphinx/source/postProcessing.rst
@@ -85,6 +85,15 @@ Plots lifetimes vs energy. Example usage::
 
 See note above under ``tau_path.py`` for more info about calcIndex (the 0 at the end).
 
+epa_tau.py
+^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Plots lifetimes vs energy from an EPA calculation. Example usage::
+
+  python epa_tau.py epa_relaxation_times.json 0
+
+See note above under ``tau_path.py`` for more info about calcIndex (the 0 at the end).
+
 tau_bandResolved.py
 ^^^^^^^^^^^^^^^^^^^^^^^^^
 

diff --git a/doc/sphinx/source/theory/phononBTE.rst b/doc/sphinx/source/theory/phononBTE.rst
@@ -437,5 +437,5 @@ Hence, one can work with the same techniques detailed above, provided that we wo
 
 * Disadvantage 2: note that the symmetric matrix gains two Cartesian coordinate indices. As a result, in the limiting case of no symmetries in the system (only the identity), the matrix :math:`A^{ij}_{\nu^*\nu'^*}` will still be computed on the same number of wavevectors of  :math:`A_{\nu\nu'}`, but occupies 3x3 times more memory without adding any information. Therefore, for low-symmetry systems, consider disabling symmetries.
 
-* Disadvantage 3: The symmetries of the BTE are so far not applicable to the variational and relaxons solveres. This is not so much a problem with implementaiton, but instead is because of a need for a derivation of symmetries for these cases.
+* Disadvantage 3: The symmetries of the BTE are so far not applicable to the variational and relaxons solvers. This is not so much a problem with implementation, but instead is because of a need for a derivation of symmetries for these cases.
 
diff --git a/doc/sphinx/source/tutorials/elEpaTransport.rst b/doc/sphinx/source/tutorials/elEpaTransport.rst
@@ -229,7 +229,7 @@ To do this, let's have a look at the input file ``qeToPhoebeEPA.in``::
   electronFourierCutoff = 4.
   epaMinEnergy = -4. eV
   epaMaxEnergy = 10. eV
-  epaNumBins = 10
+  epaNumBins = 40
   epaSmearingEnergy = 0.05 eV
 
 The parameters in this input file are as follows:
@@ -278,6 +278,7 @@ Below is an example input file for computing electronic transport properties::
 
   appName = "transportEpa"
 
+  useSymmetries = true
   electronH0Name = "qe-elph/out/silicon.xml",
   epaFileName = "qe-elph/silicon.phoebe.epa.dat"
 
@@ -293,6 +294,8 @@ The parameters used here are:
 
 * :ref:`appName` = `"transportEPA"`: selects the app for computing electronic transport properties with EPA.
 
+* :ref:`useSymmetries`: whether or not to use symmetries of the crystal structure in the calculation. 
+
 * :ref:`electronH0Name`: points to the Quantum-ESPRESSO ``*.xml`` file created by ``pw.x``, which contains the electronic single-particle energies.
 
 * :ref:`epaFileName`: is the path to the file created at the previous step with ``elPhQeToPhoebe``.
@@ -340,16 +343,14 @@ You can learn more about how to post-process these files at :ref:`postprocessing
 
 **Files which are always output for this calculation:**
 
-* ``electron_bands.json``: contains the electron band energies used in the calculation.
-
-* ``electron_dos.json``: contains the electron density of states used in the calculation.
+* ``epa_dos.json``: contains the electron density of states used in the calculation.
 
 * ``epa_onsager_coefficients.json``: contains the electronic transport coefficients from EPA.
 
 * ``epa_relaxation_times.json``: contains the EPA relaxation times at each energy bin value.
 
 
-To understand how to parse these files in more detail, take a look at the scripts described by the :ref:`postprocessing` page. In particular, if you want to plot lifetimes vs. energy, look at ``tau.py``. If you want to plot the transport coefficients vs. doping or temperature, check out ``transport_coefficients.py``.
+To understand how to parse these files in more detail, take a look at the scripts described by the :ref:`postprocessing` page. In particular, if you want to plot lifetimes vs. energy, look at ``epa_tau.py``. If you want to plot the transport coefficients vs. doping or temperature, check out ``transport_coefficients.py``.
 
 .. note::
    It's a good idea to also use bands.py to plot the band structure. Fourier interpolation of the band structure can be a source of error -- you may need to go to higher k-point meshes to get a reasonable interpolation of the band structure. Check the bands at this stage to make sure they are similar to the true DFT bands -- otherwise, your results could be problematic!

diff --git a/example/Silicon-el/reference/variational_onsager_coefficients.json b/example/Silicon-el/reference/variational_onsager_coefficients.json
@@ -10,59 +10,59 @@
    "electricalConductivity": [
       [
          [
-            1067476.083423818,
-            -289451.1726509079,
-            6988.620289292474
+            1067524.755964012,
+            -289797.97319140926,
+            4576.259106139662
          ],
          [
-            -288402.6295156577,
-            1153574.4448552558,
-            -339252.03872597613
+            -290534.25409045833,
+            1153700.6483462334,
+            -337794.79923207546
          ],
          [
-            6298.688866869923,
-            -338748.84561951895,
-            999624.5190918705
+            2423.3550264169166,
+            -337602.28847748507,
+            999823.4331092425
          ]
       ]
    ],
    "electricalConductivityUnit": "S / m",
    "electronicThermalConductivity": [
       [
          [
-            1.6784141552068075,
-            -0.17095390966293136,
-            -0.1925551757224847
+            1.6759588003842663,
+            -0.1695184449742141,
+            -0.19316637320993976
          ],
          [
-            -0.16887322607943037,
-            0.967469430606753,
-            0.45201855869035085
+            -0.17307030791637715,
+            0.9671708308136298,
+            0.4527993530482056
          ],
          [
-            -0.191703686958177,
-            0.4539177794991126,
-            1.2702570365076697
+            -0.1895368529229464,
+            0.4451924672444988,
+            1.265051833070335
          ]
       ]
    ],
    "electronicThermalConductivityUnit": "W / m / K",
    "mobility": [
       [
          [
-            66.6266227276008,
-            -18.066122864709598,
-            0.43619540955680475
+            66.62966062889252,
+            -18.087768453901255,
+            0.28562765358693365
          ],
          [
-            -18.00067794376874,
-            72.00046026235661,
-            -21.1744487251345
+            -18.133723497245306,
+            72.00833726542098,
+            -21.083494981541044
          ],
          [
-            0.3931332732677861,
-            -21.14304187885133,
-            62.3916607941002
+            0.15125393775778945,
+            -21.071479404221538,
+            62.404076031683935
          ]
       ]
    ],
@@ -71,19 +71,19 @@
    "seebeckCoefficient": [
       [
          [
-            -3.834412383472706,
-            -24.797489282120214,
-            14.089965245893998
+            -3.7355359725395547,
+            -24.81297218056885,
+            14.075023704624254
          ],
          [
-            -14.256947447646306,
-            -19.694668984349708,
-            -14.015560582347534
+            -14.206137023556979,
+            -19.708154573007306,
+            -14.119749174882436
          ],
          [
-            14.752677089503676,
-            -8.432328471018659,
-            -59.49649274673085
+            15.116531799699265,
+            -8.807694880433802,
+            -59.56690767248755
          ]
       ]
    ],
Original file line number	Diff line number	Diff line change
Expand Up		@@ -437,5 +437,5 @@ Hence, one can work with the same techniques detailed above, provided that we wo

		* Disadvantage 2: note that the symmetric matrix gains two Cartesian coordinate indices. As a result, in the limiting case of no symmetries in the system (only the identity), the matrix :math:`A^{ij}_{\nu^\nu'^}` will still be computed on the same number of wavevectors of :math:`A_{\nu\nu'}`, but occupies 3x3 times more memory without adding any information. Therefore, for low-symmetry systems, consider disabling symmetries.

		* Disadvantage 3: The symmetries of the BTE are so far not applicable to the variational and relaxons solveres. This is not so much a problem with implementaiton, but instead is because of a need for a derivation of symmetries for these cases.
		* Disadvantage 3: The symmetries of the BTE are so far not applicable to the variational and relaxons solvers. This is not so much a problem with implementation, but instead is because of a need for a derivation of symmetries for these cases.