validation_LiAl.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "25d5b0d5-f330-4dcb-9b7c-f57c4bea9596",
   "metadata": {},
   "source": [
    "# **Workshop: From electrons to phase diagrams**\n",
    "\n",
    "# Day 2: Validation of the potentials\n",
    "\n",
    "Once we have the fitted potentials, it is necessary to validate them in order to assess their quality with respect to applications.\n",
    "\n",
    "In this exercise, we use the fitted potentials and perform some basic calculations."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4756d4c9-304a-4ccc-b772-ba67d008c5a4",
   "metadata": {},
   "source": [
    "## Import the fitted potentials for Li-Al (from earlier excercise)\n",
    "\n",
    "The same directory contains a `helper.py` file which among other things, also contains the necessary specifications of each of the potentials that we will use today. Individual potentials are descrbed in the LAMMPS format as:\n",
    "```\n",
    "pot_eam = pd.DataFrame({\n",
    "    'Name': ['LiAl_eam'],\n",
    "    'Filename': [[\"../potentials/AlLi.eam.fs\")]],\n",
    "    'Model': [\"EAM\"],\n",
    "    'Species': [['Li', 'Al']],\n",
    "    'Config': [['pair_style eam/fs\\n', 'pair_coeff * * AlLi.eam.fs Li Al\\n']]\n",
    "})\n",
    "\n",
    "```\n",
    "A list of such DataFrames describing the potentials is saved in a list called `potentials_list`. We import the list as:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b90e0ac0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Filename</th>\n",
       "      <th>Model</th>\n",
       "      <th>Species</th>\n",
       "      <th>Config</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>[/home/minaam.qamar/workshop_POTENTIALS/new_no...</td>\n",
       "      <td>EAM</td>\n",
       "      <td>[Li, Al]</td>\n",
       "      <td>[pair_style eam/fs\\n, pair_coeff * * AlLi.eam....</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Name  ...                                             Config\n",
       "0  LiAl_eam  ...  [pair_style eam/fs\\n, pair_coeff * * AlLi.eam....\n",
       "\n",
       "[1 rows x 5 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from helper import potentials_list\n",
    "\n",
    "potentials_list = [potentials_list[0]]\n",
    "\n",
    "# display the first element in the list\n",
    "# which is an EAM potential\n",
    "potentials_list[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c84560c",
   "metadata": {},
   "source": [
    "### Import other important modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "83f7a2c9-d45a-4987-9e35-59badd754d4f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6.366608268"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pylab as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import time\n",
    "from helper import get_clean_project_name\n",
    "from pyiron_atomistics import Project\n",
    "from pyiron import pyiron_to_ase\n",
    "import pyiron_gpl\n",
    "\n",
    "# save start time to record runtime of the notebook\n",
    "time_start =  time.process_time()\n",
    "time_start"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "acc0ee8f",
   "metadata": {},
   "source": [
    "### Create a new project to perform validation calculations\n",
    "\n",
    "It is useful to create a new project directory for every kind of calculation. Pyiron will automatically create subdirectories for each potential and property we calculate. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "706be2a9-5f94-4eb5-8e4f-6c349fe216b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "pr = Project(\"validation_LiAl\")\n",
    "\n",
    "# remove earlier jobs\n",
    "# pr.remove_jobs(silently=True, recursive=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b84ed62-e841-4526-893e-dc4f61477c88",
   "metadata": {},
   "source": [
    "### Define the important pases to consider for validation\n",
    "\n",
    "We construct a python dictionary `struct_dict` which contains a description of all the important phases that we want to consider for this exercise. The descriptions given in the dictionary will be later used by Pyiron to generate or read the structural configurations for the respective phases.\n",
    "\n",
    "For unary phases, we provide an initial guess for the lattice parameter and use pyiron to generate the structural prototype.\n",
    "\n",
    "For binary phases, we provide a phase name and an additional dictionary `fl_dict` which maps the phase name to a `.cif` file saved in a subdirectory. Pyiron will use this information to read the respective configurations from the file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "28778cef-2a07-4794-888f-7239500e7b5a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Al': {'s_murn': ['fcc', 'bcc'], 'a': 4.04},\n",
       " 'Li': {'s_murn': ['bcc', 'fcc'], 'a': 3.5},\n",
       " 'Li2Al2': {'s_murn': ['Li2Al2_cubic']},\n",
       " 'LiAl3': {'s_murn': ['LiAl3_cubic']},\n",
       " 'Li9Al4': {'s_murn': ['Li9Al4_monoclinic']},\n",
       " 'Li3Al2': {'s_murn': ['Li3Al2_trigonal']},\n",
       " 'Li4Al4': {'s_murn': ['Li4Al4_cubic']}}"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "struct_dict = dict()\n",
    "struct_dict[\"Al\"] = dict()\n",
    "struct_dict[\"Al\"][\"s_murn\"] = [\"fcc\",\"bcc\"]\n",
    "struct_dict[\"Al\"][\"a\"] = 4.04\n",
    "\n",
    "struct_dict[\"Li\"] = dict()\n",
    "struct_dict[\"Li\"][\"s_murn\"] = [\"bcc\",\"fcc\"]\n",
    "struct_dict[\"Li\"][\"a\"] = 3.5\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "struct_dict[\"Li2Al2\"] = dict()\n",
    "struct_dict[\"Li2Al2\"][\"s_murn\"] = [\"Li2Al2_cubic\"]\n",
    "# struct_dict[\"Li2Al2\"][\"a\"] = 3.7\n",
    "\n",
    "struct_dict[\"LiAl3\"] = dict()\n",
    "struct_dict[\"LiAl3\"][\"s_murn\"] = [\"LiAl3_tetragonal\"]\n",
    "# struct_dict[\"LiAl3\"][\"a\"] = 3.7\n",
    "\n",
    "struct_dict[\"LiAl3\"] = dict()\n",
    "struct_dict[\"LiAl3\"][\"s_murn\"] = [\"LiAl3_cubic\"]\n",
    "# struct_dict[\"LiAl3\"][\"a\"] = 3.7\n",
    "\n",
    "struct_dict[\"Li9Al4\"] = dict()\n",
    "struct_dict[\"Li9Al4\"][\"s_murn\"] = [\"Li9Al4_monoclinic\"]\n",
    "# struct_dict[\"Li9Al4\"][\"a\"] = 3.7\n",
    "\n",
    "struct_dict[\"Li3Al2\"] = dict()\n",
    "struct_dict[\"Li3Al2\"][\"s_murn\"] = [\"Li3Al2_trigonal\"]\n",
    "# struct_dict[\"Li3Al2\"][\"a\"] = 3.7\n",
    "\n",
    "struct_dict[\"Li4Al4\"] = dict()\n",
    "struct_dict[\"Li4Al4\"][\"s_murn\"] = [\"Li4Al4_cubic\"]\n",
    "\n",
    "struct_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23b2e6d9",
   "metadata": {},
   "source": [
    "a dictionary is described to map the binary phases to their file locations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "c1820db7",
   "metadata": {},
   "outputs": [],
   "source": [
    "fl_dict = {\"Li2Al2_cubic\": \"mp_structures/LiAl_mp-1067_primitive.cif\",\n",
    "           \"LiAl3_tetragonal\":\"mp_structures/LiAl3_mp-975906_primitive.cif\",\n",
    "           \"LiAl3_cubic\":\"mp_structures/LiAl3_mp-10890_primitive.cif\",\n",
    "           \"Li9Al4_monoclinic\":\"mp_structures/Li9Al4_mp-568404_primitive.cif\",\n",
    "           \"Li3Al2_trigonal\":\"mp_structures/Al2Li3-6021.cif\",\n",
    "           \"Li4Al4_cubic\":\"mp_structures/LiAl_mp-1079240_primitive.cif\"}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "198e9745-734a-4502-8f1b-0330ba8c8fca",
   "metadata": {},
   "source": [
    "### (a) Ground state: E-V curves\n",
    "\n",
    "Using a series of nested `for` loops, we calculate the murnaghan EV-curves using all three potentials for all the defined structures.\n",
    "\n",
    "We loop over:\n",
    " - All the potentials defined in `potentials_list` and name the project according to the potential\n",
    "   - All the chemical formulae defined in the keys of `struct_dict`\n",
    "     - All phases defined for a given chemical formula\n",
    "     \n",
    "Within the loops, the first step is to get the structure basis on which we will perform the calculations. \n",
    "\n",
    "- For unary phases, we use the pyiron function `pr_pot.create_ase_bulk(compound, crys_structure, a=compound_dict[\"a\"])` \n",
    "- For binary structures, we read the basis using `pr.create.structure.ase.read(fl_path)` with the `fl_path` given by `fl_dict` defined earlier.\n",
    "\n",
    "Once the structure and potential is defined as part of the pr_job, we run two calculations:\n",
    "- `job_relax` to relax the structure to the ground state\n",
    "- `murn_job` to calculate the energies in a small volume range around the equilibrium\n",
    "\n",
    "As the calculations are being performed, the status(s) of each calculation is printed. If a job is already calculated, the calculations are not re-run but rather re-read from the saved data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "13f095d2-44d7-4711-b9a5-d58a95af42f6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Al_fcc_relax was saved and received the ID: 481\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:18:40,708 - pyiron_log - WARNING - The job murn_job_Al_fcc is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Al_bcc_relax was saved and received the ID: 482\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:18:44,855 - pyiron_log - WARNING - The job murn_job_Al_bcc is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li_bcc_relax was saved and received the ID: 483\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:18:49,187 - pyiron_log - WARNING - The job murn_job_Li_bcc is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li_fcc_relax was saved and received the ID: 484\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:18:55,198 - pyiron_log - WARNING - The job murn_job_Li_fcc is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li2Al2_Li2Al2_cubic_relax was saved and received the ID: 485\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:19:00,132 - pyiron_log - WARNING - The job murn_job_Li2Al2_Li2Al2_cubic is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job LiAl3_LiAl3_cubic_relax was saved and received the ID: 486\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:19:06,418 - pyiron_log - WARNING - The job murn_job_LiAl3_LiAl3_cubic is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li9Al4_Li9Al4_monoclinic_relax was saved and received the ID: 487\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:19:11,676 - pyiron_log - WARNING - The job murn_job_Li9Al4_Li9Al4_monoclinic is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li3Al2_Li3Al2_trigonal_relax was saved and received the ID: 488\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-30 19:19:14,923 - pyiron_log - WARNING - The job murn_job_Li3Al2_Li3Al2_trigonal is being loaded instead of running. To re-run use the argument 'delete_existing_job=True in create_job'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job Li4Al4_Li4Al4_cubic_relax was saved and received the ID: 489\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic was saved and received the ID: 490\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_0_9 was saved and received the ID: 491\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_0_92 was saved and received the ID: 492\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_0_94 was saved and received the ID: 493\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_0_96 was saved and received the ID: 494\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_0_98 was saved and received the ID: 495\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_0 was saved and received the ID: 496\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_02 was saved and received the ID: 497\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_04 was saved and received the ID: 498\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_06 was saved and received the ID: 499\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_08 was saved and received the ID: 500\n",
      "The job murn_job_Li4Al4_Li4Al4_cubic_1_1 was saved and received the ID: 501\n"
     ]
    }
   ],
   "source": [
    "for pot in potentials_list:\n",
    "    with pr.open(get_clean_project_name(pot)) as pr_pot:\n",
    "        for compound, compound_dict in struct_dict.items():\n",
    "            for crys_structure in compound_dict[\"s_murn\"]:\n",
    "                \n",
    "                # Relax structure\n",
    "                if crys_structure in [\"fcc\",\"bcc\"]:\n",
    "                    basis = pr_pot.create_ase_bulk(compound, crys_structure, a=compound_dict[\"a\"])\n",
    "                else:\n",
    "                    basis = pr.create.structure.ase.read(fl_dict[crys_structure])\n",
    "                job_relax = pr_pot.create_job(pr_pot.job_type.Lammps, f\"{compound}_{crys_structure}_relax\", delete_existing_job=True)\n",
    "\n",
    "                job_relax.structure = basis\n",
    "                job_relax.potential = pot\n",
    "                job_relax.calc_minimize(pressure=0)\n",
    "                job_relax.run()\n",
    "                \n",
    "                # Murnaghan\n",
    "                job_ref = pr_pot.create_job(pr_pot.job_type.Lammps, f\"ref_job_{compound}_{crys_structure}\")\n",
    "                job_ref.structure = job_relax.get_structure(-1)\n",
    "                job_ref.potential = pot\n",
    "                job_ref.calc_minimize()\n",
    "                \n",
    "                murn_job = job_ref.create_job(pr_pot.job_type.Murnaghan, f\"murn_job_{compound}_{crys_structure}\")\n",
    "                murn_job.input[\"vol_range\"] = 0.1\n",
    "                murn_job.run()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d848f1a",
   "metadata": {},
   "source": [
    "One can display the technical details of all submitted jobs using `pr.job_table()` below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "fdc89ebb-3c2a-4315-8fe0-3ae470375223",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# pr.job_table()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "425dcaec",
   "metadata": {},
   "source": [
    "In order to get read useful results from the completed calculations (eq_energy, eq_volume, etc), it is useful to define the following functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "ef2f414b-64b8-49aa-87e9-e204950da938",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Only work with Murnaghan jobs\n",
    "def get_only_murn(job_table):\n",
    "    return (job_table.hamilton == \"Murnaghan\") & (job_table.status == \"finished\") \n",
    "\n",
    "def get_eq_vol(job_path):\n",
    "    return job_path[\"output/equilibrium_volume\"]\n",
    "\n",
    "def get_eq_lp(job_path):\n",
    "    return np.linalg.norm(job_path[\"output/structure/cell/cell\"][0]) * np.sqrt(2)\n",
    "\n",
    "def get_eq_bm(job_path):\n",
    "    return job_path[\"output/equilibrium_bulk_modulus\"]\n",
    "\n",
    "def get_potential(job_path):\n",
    "    return job_path.project.path.split(\"/\")[-3]\n",
    "\n",
    "def get_eq_energy(job_path):\n",
    "    return job_path[\"output/equilibrium_energy\"]\n",
    "\n",
    "def get_n_atoms(job_path):\n",
    "    return len(job_path[\"output/structure/positions\"])\n",
    "\n",
    "def get_ase_atoms(job_path):\n",
    "    return pyiron_to_ase(job_path.structure).copy()\n",
    "\n",
    "\n",
    "def get_potential(job_path):\n",
    "    return job_path.project.path.split(\"/\")[-2]\n",
    "\n",
    "def get_crystal_structure(job_path):\n",
    "    return job_path.job_name.split(\"_\")[-1]\n",
    "\n",
    "def get_compound(job_path):\n",
    "    return job_path.job_name.split(\"_\")[-2]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2fe57b8b",
   "metadata": {},
   "source": [
    "Using the functions defined above, one can now define a `pd.DataFrame` containing all useful results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "255c28af-e4af-48c6-ae01-e90377c94e32",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The job table_murn was saved and received the ID: 502\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a5d41b6f348b4e31b85295d74e6f93b0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Loading and filtering jobs:   0%|          | 0/9 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d4443355c4a458584cfc4343b77d69b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Processing jobs:   0%|          | 0/9 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/minaam.qamar/.conda/envs/workshop2/lib/python3.10/site-packages/pyiron_base/table/datamining.py:620: PerformanceWarning: \n",
      "your performance may suffer as PyTables will pickle object types that it cannot\n",
      "map directly to c-types [inferred_type->mixed,key->block2_values] [items->Index(['potential', 'ase_atoms', 'compound', 'crystal_structure'], dtype='object')]\n",
      "\n",
      "  self.pyiron_table._df.to_hdf(\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>job_id</th>\n",
       "      <th>potential</th>\n",
       "      <th>ase_atoms</th>\n",
       "      <th>compound</th>\n",
       "      <th>crystal_structure</th>\n",
       "      <th>a</th>\n",
       "      <th>eq_vol</th>\n",
       "      <th>eq_bm</th>\n",
       "      <th>eq_energy</th>\n",
       "      <th>n_atoms</th>\n",
       "      <th>phase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>298</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Al', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Al</td>\n",
       "      <td>fcc</td>\n",
       "      <td>4.039967</td>\n",
       "      <td>16.495612</td>\n",
       "      <td>85.876912</td>\n",
       "      <td>-3.483097</td>\n",
       "      <td>1</td>\n",
       "      <td>Al_fcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>311</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Al', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Al</td>\n",
       "      <td>bcc</td>\n",
       "      <td>3.898853</td>\n",
       "      <td>16.147864</td>\n",
       "      <td>48.620841</td>\n",
       "      <td>-3.415312</td>\n",
       "      <td>1</td>\n",
       "      <td>Al_bcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>324</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Li</td>\n",
       "      <td>bcc</td>\n",
       "      <td>4.195477</td>\n",
       "      <td>20.114514</td>\n",
       "      <td>13.690609</td>\n",
       "      <td>-1.757011</td>\n",
       "      <td>1</td>\n",
       "      <td>Li_bcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>337</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Li</td>\n",
       "      <td>fcc</td>\n",
       "      <td>4.253841</td>\n",
       "      <td>19.241330</td>\n",
       "      <td>13.985972</td>\n",
       "      <td>-1.758107</td>\n",
       "      <td>1</td>\n",
       "      <td>Li_fcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>350</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [4.359978178265942, 2.5172345748814804, 1.7799536377360752], index=0), Atom('Li', [6.53996726740165, 3.7758518623203585, 2.669930456604318], index=1), Atom('Al', [-3.964456982410852e-1...</td>\n",
       "      <td>Li2Al2</td>\n",
       "      <td>cubic</td>\n",
       "      <td>6.165940</td>\n",
       "      <td>58.604895</td>\n",
       "      <td>100.347240</td>\n",
       "      <td>-11.074362</td>\n",
       "      <td>4</td>\n",
       "      <td>Li2Al2_cubic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>363</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [0.0, 0.0, 0.0], index=0), Atom('Al', [1.9825515172760235, 1.9825515172760237, 2.427925369776811e-16], index=1), Atom('Al', [1.9825515172760235, 1.2139626848884054e-16, 1.9825515172760...</td>\n",
       "      <td>LiAl3</td>\n",
       "      <td>cubic</td>\n",
       "      <td>5.607502</td>\n",
       "      <td>62.227580</td>\n",
       "      <td>51.472656</td>\n",
       "      <td>-12.774590</td>\n",
       "      <td>4</td>\n",
       "      <td>LiAl3_cubic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>376</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [4.9874675377354745, 1.0099032069001204, 0.8188717268019969], index=0), Atom('Li', [3.1237856900347722, 1.4557299809760758, 2.6737242758835587], index=1), Atom('Li', [-3.44219970899484...</td>\n",
       "      <td>Li9Al4</td>\n",
       "      <td>monoclinic</td>\n",
       "      <td>13.023702</td>\n",
       "      <td>190.504371</td>\n",
       "      <td>53.125273</td>\n",
       "      <td>-28.970054</td>\n",
       "      <td>13</td>\n",
       "      <td>Li9Al4_monoclinic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>389</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Al', [2.1548001975659243, 1.2440753587819189, 1.8617841750008692], index=0), Atom('Al', [-2.154798282819334, 3.7322233132135576, 2.664676023808053], index=1), Atom('Li', [8.560563403365655e...</td>\n",
       "      <td>Li3Al2</td>\n",
       "      <td>trigonal</td>\n",
       "      <td>6.094693</td>\n",
       "      <td>72.810229</td>\n",
       "      <td>69.231669</td>\n",
       "      <td>-12.413856</td>\n",
       "      <td>5</td>\n",
       "      <td>Li3Al2_trigonal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>490</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [2.1429671662366734, 1.237242669260901, 7.662122465542487], index=0), Atom('Li', [-8.827603092953495e-10, 2.4744853400282523, 0.5913662557792814], index=1), Atom('Li', [-8.827603092953...</td>\n",
       "      <td>Li4Al4</td>\n",
       "      <td>cubic</td>\n",
       "      <td>6.061226</td>\n",
       "      <td>131.389799</td>\n",
       "      <td>71.221356</td>\n",
       "      <td>-20.506570</td>\n",
       "      <td>8</td>\n",
       "      <td>Li4Al4_cubic</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   job_id potential  ... n_atoms              phase\n",
       "0     298  LiAl_eam  ...       1             Al_fcc\n",
       "1     311  LiAl_eam  ...       1             Al_bcc\n",
       "2     324  LiAl_eam  ...       1             Li_bcc\n",
       "3     337  LiAl_eam  ...       1             Li_fcc\n",
       "4     350  LiAl_eam  ...       4       Li2Al2_cubic\n",
       "5     363  LiAl_eam  ...       4        LiAl3_cubic\n",
       "6     376  LiAl_eam  ...      13  Li9Al4_monoclinic\n",
       "7     389  LiAl_eam  ...       5    Li3Al2_trigonal\n",
       "8     490  LiAl_eam  ...       8       Li4Al4_cubic\n",
       "\n",
       "[9 rows x 11 columns]"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Compile data using pyiron tables\n",
    "table = pr.create_table(\"table_murn\", delete_existing_job=True)\n",
    "table.convert_to_object = True\n",
    "table.db_filter_function = get_only_murn\n",
    "table.add[\"potential\"] = get_potential\n",
    "table.add[\"ase_atoms\"] = get_ase_atoms\n",
    "table.add[\"compound\"] = get_compound\n",
    "table.add[\"crystal_structure\"] = get_crystal_structure\n",
    "table.add[\"a\"] = get_eq_lp\n",
    "table.add[\"eq_vol\"] = get_eq_vol\n",
    "table.add[\"eq_bm\"] = get_eq_bm\n",
    "table.add[\"eq_energy\"] = get_eq_energy\n",
    "table.add[\"n_atoms\"] = get_n_atoms\n",
    "table.run()\n",
    "\n",
    "data_murn = table.get_dataframe()\n",
    "data_murn[\"phase\"] = data_murn.compound + \"_\" + data_murn.crystal_structure\n",
    "data_murn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "30d27d75",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 480x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax_list = plt.subplots(ncols=len(potentials_list), nrows=1, sharex=\"row\", sharey=\"row\")\n",
    "\n",
    "fig.set_figwidth(20/3)\n",
    "fig.set_figheight(6)\n",
    "\n",
    "color_palette = sns.color_palette(\"tab10\", n_colors=len(data_murn.phase.unique()))\n",
    "\n",
    "\n",
    "for i, pot in enumerate(potentials_list):\n",
    "    \n",
    "    if len(potentials_list) == 1:\n",
    "        ax = ax_list\n",
    "    else:\n",
    "        ax = ax_list[i]\n",
    "        \n",
    "    data = data_murn[data_murn.potential == get_clean_project_name(pot)]\n",
    "    \n",
    "    for j,(_, row) in enumerate(data.iterrows()):\n",
    "        murn_job = pr.load(row[\"job_id\"])\n",
    "        murn_df = murn_job.output_to_pandas()\n",
    "        n_atoms = row[\"n_atoms\"]\n",
    "        \n",
    "        ax.plot(murn_df[\"volume\"]/n_atoms, murn_df[\"energy\"]/n_atoms,\"o-\",\n",
    "                lw=4,\n",
    "                label= row[\"phase\"], \n",
    "                color=color_palette[j])\n",
    "        \n",
    "    ax.set_title(f\"{get_clean_project_name(pot)}\",fontsize=22)\n",
    "    ax.set_xlabel(\"Volume per atom, $\\mathrm{\\AA^3}$\",fontsize=20)\n",
    "    ax.set_ylabel(\"Energy per atom, eV/atom\",fontsize=20)\n",
    "    ax.tick_params(labelsize=18)\n",
    "    ax.legend(prop={\"size\":16})\n",
    "    #break\n",
    "fig.subplots_adjust(wspace=0.1);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fba90359-a2a5-4f83-9fa8-6dc4d87f5743",
   "metadata": {},
   "source": [
    "## (b) Elastic constants and Phonons\n",
    "\n",
    "Pyiron also has job modules to calculate elastic constants and thermal properties using the quasi-harmonic approximation given by the `phonopy` package.\n",
    "\n",
    "As in the previous task, we again loop over the defined potentials and then over the given structures.\n",
    "\n",
    "Calculating elastic constants and thermal properties is considerably more expensive than calculating EV curves. Hence, it is useful to only calculate these properties for a subset of most important structures "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7bf87f90",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>job_id</th>\n",
       "      <th>potential</th>\n",
       "      <th>ase_atoms</th>\n",
       "      <th>compound</th>\n",
       "      <th>crystal_structure</th>\n",
       "      <th>a</th>\n",
       "      <th>eq_vol</th>\n",
       "      <th>eq_bm</th>\n",
       "      <th>eq_energy</th>\n",
       "      <th>n_atoms</th>\n",
       "      <th>phase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>298</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Al', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Al</td>\n",
       "      <td>fcc</td>\n",
       "      <td>4.039967</td>\n",
       "      <td>16.495612</td>\n",
       "      <td>85.876912</td>\n",
       "      <td>-3.483097</td>\n",
       "      <td>1</td>\n",
       "      <td>Al_fcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>324</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [0.0, 0.0, 0.0], index=0))</td>\n",
       "      <td>Li</td>\n",
       "      <td>bcc</td>\n",
       "      <td>4.195477</td>\n",
       "      <td>20.114514</td>\n",
       "      <td>13.690609</td>\n",
       "      <td>-1.757011</td>\n",
       "      <td>1</td>\n",
       "      <td>Li_bcc</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>350</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [4.359978178265942, 2.5172345748814804, 1.7799536377360752], index=0), Atom('Li', [6.53996726740165, 3.7758518623203585, 2.669930456604318], index=1), Atom('Al', [-3.964456982410852e-1...</td>\n",
       "      <td>Li2Al2</td>\n",
       "      <td>cubic</td>\n",
       "      <td>6.165940</td>\n",
       "      <td>58.604895</td>\n",
       "      <td>100.347240</td>\n",
       "      <td>-11.074362</td>\n",
       "      <td>4</td>\n",
       "      <td>Li2Al2_cubic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>363</td>\n",
       "      <td>LiAl_eam</td>\n",
       "      <td>(Atom('Li', [0.0, 0.0, 0.0], index=0), Atom('Al', [1.9825515172760235, 1.9825515172760237, 2.427925369776811e-16], index=1), Atom('Al', [1.9825515172760235, 1.2139626848884054e-16, 1.9825515172760...</td>\n",
       "      <td>LiAl3</td>\n",
       "      <td>cubic</td>\n",
       "      <td>5.607502</td>\n",
       "      <td>62.227580</td>\n",
       "      <td>51.472656</td>\n",
       "      <td>-12.774590</td>\n",
       "      <td>4</td>\n",
       "      <td>LiAl3_cubic</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   job_id potential  ... n_atoms         phase\n",
       "0     298  LiAl_eam  ...       1        Al_fcc\n",
       "2     324  LiAl_eam  ...       1        Li_bcc\n",
       "4     350  LiAl_eam  ...       4  Li2Al2_cubic\n",
       "5     363  LiAl_eam  ...       4   LiAl3_cubic\n",
       "\n",
       "[4 rows x 11 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list_of_phases = [\"Al_fcc\",\"Li_bcc\",\"Li2Al2_cubic\",\"LiAl3_cubic\"]\n",
    "\n",
    "subset_murn = data_murn[data_murn[\"phase\"].isin(list_of_phases)]\n",
    "subset_murn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "0d1c799c-f10b-462d-aaea-253cee4b4b3e",
   "metadata": {
    "scrolled": true