Automatic Featurization Aided Data-Driven Method for Estimating the Presence of Intermetallic Phase in Multi-Principal Element Alloys

Description:

There are two folders in the dataset:

(A. ) AutomaticFeaturizerMPEA.zip : 

AutomaticFeaturizerMPEA.pynb is an interactive computational environment developed in python programming language to compute the properties of Multi Principal-component Element Alloy (MPEA) system .

The code present in the .ipynb file is capable of firstly extracting composition feature from the name of MPEA and then compute thermodynamic/physical and chemical properties using these composition features. The properties that can be automatically calculated and featurized by the code are: 
1. Valence Electron Concentration
2. Average Melting Temperature
3. Enthalpy of Mixing
4. Entropy of Mixing
5. Electronegativity
6. Atomic size difference
7. Omega parameter  
8. 41 elemental composition

The primary source used for the calculation of these alloy properties are taken from two python libraries: "pymatgen" and "matminer". 

The AutomaticFeaturizerMPEA.pynb is not only capable of calculating an alloy's properties but also featurizes them along with the composition and export it as a .csv file for further machine learning training process. The AutomaticFeaturizerMPEA.pynb file consists of two parts:

1. "Part 1": A sample set of 25 MPEAs is used to demonstarte how to calculate & featurize the porperties and composition of alloy, save the calculated values in a .csv file.
2. "Part 2": The alloy property of a single MPEA "Ti0.8CoCrFeNiCu" is calculated using the same code, if one needs to calculate them for a single MPEA at a time (as in our GUI application IMCATHEA https://github.com/subediupadesh/IMCATHEA) 

The repository also consists two "CSV" files:

1. "MPEA_data.csv" file is the sample set of 25 collected MEPA observations with their source and phase information in one hot encoding method for which we want to calculate their properties and featurize them.
2. "Featurized_MPEA.csv" file is the processed ready to train file which consists of MPEA, their properties and composition featurized in special order. It is the final file generated by the AutomaticFeaturizerMPEA.pynb python file.

(B.) IMCATHEA.zip: 

IMCATHEA is a python based Graphical User Interface application capable of detecting the IM phase in a High Entropy Alloy or Multi Principal Element Alloy system using a Neural Network Algorithm.

## For Linux User:
Python with following libraries are needed to run the GUI application: pymatgen (v2022.0.16), matminer (v0.7.4), tensorflow (v2.2.0), pandas (v1.3.4), numpy (v1.21.2), tkinter (v8.6), sklearn (v1.0.1).

"IMCATHEA_GUI.py" also needs 4 extra supporitng files (icon, standardizations, and model) to operate, which will be made available from the author upon reasonable request. 

## For Windows User:
Executable (.exe) file: 

User can run "IMCATHEA_GUI.exe" wihout any softwares to be installed.

## Procedure for using "IMCATHEA" GUI Application:

Step 0. Open / Run "IMCATHEA_GUI.py" (for Linux users) & "IMCATHEA_GUI.exe" (for Windows users) file to open the GUI application

Step 1. User need to select the No. of Elements/Components in the HEA where IM phase is to be detected, from the dropdown menu at the top left corner.
            After selection of element size/number (from 2 to 10),

Step 2. User can select each element of HEA, one at a time from the drop down menu generated just below "No. of Component" tab

Step 3. After selection of each element a blank space is provided just at the right side of selected element tab where user need to enter the corresponding composition/elemental           fraction of the element.

Step 4. Repeat Step 2-3 until the last element of the HEA and it's composition/elemental fraction is entered.

Step 5. Press "Detect IMC" tab to get the prediction if IM is present in the provided HEA or not along with the physical properties of the entered HEA displayed in the right side of the GUI application.

Step 6. If user wants to detect the presence or absence of IM phase for another HEA, user can click on "Restart" tab at the right top side of GUI application to restart the application instantly.

If the user makes errors while selecting any options from dropdown menu as suggested in Step 1,2,3,4 user can click on "Restart" tab at the right top side of GUI application to restart the application instantly and start the process for prediction again from Step 1.

## Video (YouTube) Tutorial for using IMCATHEA:
The stepwise process to predict the phase for an example case of "CrNbTiZr" MPEA is shown in the YouTube video tutorial at: https://youtu.be/G3aHE1Wsmgk