SCOPE
AboutFeaturesHow it worksScreeningAccess
Perovskite cathode screening platform

Discover better cathodes with machine learning

Explore SCOPE
SCOPE logo

SCOPE

Screening Cathodes for Optimal Perovskite Efficiency

Neil Gogte Institute of Technology
About SCOPE

From a spreadsheet of candidates to a ranked shortlist

SCOPE — Screening Cathodes for Optimal Perovskite Efficiency — is a web workspace for screening perovskite cathode materials with machine learning.

Upload a dataset of compositions and descriptors. SCOPE trains and saves both an SVR and an ANN model in the background, predicts oxygen-reduction kinetics, checks every material against stability and applicability limits, and opens a ranked, exportable shortlist — switch between SVR and ANN results at any time.

2ML models
SVR · ANN
3pass / fail
checks
1ranked
shortlist
Features

Everything you need to screen cathodes

  • Automatic SVR + ANN trainingBoth models are trained and saved for you the moment you upload.
  • Pass / fail screeningRank candidates on Ehull stability and log10(k*) kinetics, with instant SVR / ANN switching.
  • Predict a single cathodeEnter descriptor values on the dashboard and get a verdict from a saved model.
  • Cost filterEstimate precursor cost and shortlist affordable materials.
  • 3D structure viewerInspect the crystal structure of any screened cathode.
  • Excel exportDownload full results as ready-to-share spreadsheets.
How it works

From dataset to results in four steps

  1. 1Upload your datasetCSV or Excel, up to 15 MB. SCOPE detects the formula, target and descriptor columns.
  2. 2Models train automaticallySVR and ANN are trained and saved in the background, with no extra clicks.
  3. 3Screening results openEvery cathode is checked and ranked. Switch between SVR and ANN results instantly.
  4. 4Predict & exportPredict a single cathode, open any cathode in 3D, see the best in each category and download Excel / CSV.
Screening

How a cathode passes

Applicability domain

The material must sit inside the range of the training data, measured by Mahalanobis distance, so predictions can be trusted.

Stability (Ehull)

Energy above hull must be at most 57 meV/atom.

Activity (ORR)

Predicted log10(k*) must be at least −6.40, the LSC benchmark.

Cost is estimated from precursor prices and shown as Cheap / Moderate / Expensive. It never changes PASS / FAIL.

Access

Two ways to sign in

User

Sign in with your roll number. Upload datasets or predict a single cathode from one dashboard; your datasets, models and results stay private to your account.

Admin

Sign in with admin credentials to review users, live activity, datasets and predictions across the platform.

Ready to screen your cathodes?

SCOPE · Screening Cathodes for Optimal Perovskite Efficiency
Neil Gogte Institute of Technology
SCOPE logo
SCOPE