Electronic Property Prediction for Materials Decisions
Electronic functionality depends on more than a single band-gap value. Crystal phase, composition, spin state, defects, interfaces, strain, temperature, and the chosen electronic-structure approximation can all change the quantities that determine device behavior.
CD ComputaBio combines curated customer and public data, structure-aware machine learning, and fit-for-purpose first-principles calculations to predict and interpret electronic properties before costly synthesis or device fabrication. Each engagement starts from an application-specific target profile and ends with traceable results, uncertainty statements, and recommended validation. This service sits within our AI for Semiconductor and Electronic Materials portfolio.
Core Electronic Property Services
Band Structure and Density of States
We calculate spin-resolved band dispersion, total and projected density of states, band-gap character, orbital contributions, Fermi-level position, and symmetry-resolved features using an agreed DFT level and convergence protocol.
Band Alignment and Work Function
Vacuum-aligned slab calculations and interface-aware analyses estimate ionization potential, electron affinity, work function, contact offsets, and alignment trends. Surface orientation, termination, dipoles, and finite-size convergence are documented.
Carrier and Transport Descriptors
We extract effective-mass tensors, band velocities, valley degeneracy, and anisotropy, with optional Boltzmann-transport descriptors under stated relaxation-time assumptions. Predicted mobility is not reported without an appropriate scattering model.
Band-Gap and Relativistic Refinement
For shortlisted systems, we can assess spin–orbit coupling, DFT+U sensitivity, hybrid-functional calculations, or other agreed corrections where semilocal DFT is insufficient for the decision.
AI Property Prediction and Screening
Composition- and structure-aware models predict agreed endpoints across larger search spaces. Splits are grouped to limit leakage, baselines are retained, and uncertainty or out-of-domain flags accompany candidate scores.
Composition, Strain, and Interface Trends
Controlled studies map how alloying, polymorph, lattice strain, dimensionality, magnetic order, or selected interfaces shift electronic descriptors, separating calculated trends from experimentally confirmed performance.
Integrated Project Workflow
| Stage | Key Activities | Decision Gate |
|---|---|---|
| 1. Define the target | Specify material class, device context, phase, environment, property windows, fidelity, budget, and acceptance criteria. | Approved target profile and calculation matrix. |
| 2. Curate structures and data | Validate stoichiometry, symmetry, magnetic state, provenance, units, duplicates, and calculation/measurement conditions. | Analysis-ready structures and auditable dataset. |
| 3. Build the baseline | Relax representative structures, test convergence, select pseudopotentials and exchange-correlation treatment, and establish simple ML baselines. | Method justified for the intended comparison. |
| 4. Predict and verify | Run AI screening or parameter sweeps; recalculate priority and near-threshold cases with harmonized physics-based settings. | Uncertainty-aware property set and qualified shortlist. |
| 5. Interpret and transfer | Resolve orbital/structural drivers, rank trade-offs, document limitations, and design confirmatory spectroscopy, transport, or device measurements. | Decision-ready report and next-step validation plan. |
Decision-Ready Deliverables
Electronic Property Package
Relaxed structures, calculation settings, convergence records, band/DOS files, orbital projections, aligned energy levels, and agreed transport descriptors.
Model and Candidate Scorecard
Curated data, model card, validation splits, error metrics, applicability flags, uncertainty intervals where supported, ranked candidates, and rejection rationale.
Technical Interpretation
Publication-ready plots, structure–property analysis, method limitations, reproducible tabular outputs, and experimentally actionable validation recommendations.
Representative Applications
Transistors and Power Electronics
Assess band gaps, effective masses, anisotropy, contact alignment, and wide-band-gap candidates for channel, barrier, or power-device roles.
Photovoltaics and Photodetectors
Evaluate absorber and transport-layer band edges, direct/indirect gaps, orbital character, and interface alignment across target spectral windows.
Transparent and Conducting Materials
Balance optical-gap requirements with dispersive carrier bands, dopability hypotheses, work function, stability, and elemental constraints.
2D, Magnetic, and Quantum Materials
Resolve dimensionality, symmetry, spin polarization, spin–orbit sensitivity, valley structure, and substrate- or strain-dependent electronic trends.
Scientific Basis and Open-Access Evidence
Community benchmarks such as Matbench demonstrate the value of fixed tasks, consistent cross-validation, and transparent baselines when comparing materials-property models.1 The NIST JARVIS infrastructure illustrates how standardized high-throughput calculations can connect crystal structures with electronic and related properties.2
Official Materials Project guidance emphasizes that Kohn–Sham gaps from semilocal DFT require method-aware interpretation and may systematically underestimate experimental fundamental gaps.3 We therefore select computational fidelity according to the decision and report the approximation behind every value.
1 Dunn, A.; Wang, Q.; Ganose, A.; Dopp, D.; Jain, A. Benchmarking materials property prediction methods: the Matbench test set and Automatminer reference algorithm. npj Computational Materials 2020, 6, 138. https://doi.org/10.1038/s41524-020-00406-3. Open Access, CC BY 4.0.
2 Choudhary, K.; Garrity, K. F.; Reid, A. C. E.; et al. The joint automated repository for various integrated simulations (JARVIS) for data-driven materials design. npj Computational Materials 2020, 6, 173. https://doi.org/10.1038/s41524-020-00440-1. Open Access, CC BY 4.0.
3 Materials Project. Electronic Structure Methodology. https://docs.materialsproject.org/methodology/materials-methodology/electronic-structure. Accessed July 23, 2026. Publicly available official documentation.
Why Work with CD ComputaBio?
Our projects connect electronic-structure physics, data engineering, and AI without obscuring the assumptions that control the answer. Methods are modular, calculation settings are traceable, and each recommendation is linked to its evidence, uncertainty, and next validation step.
To define the right fidelity for your material system and device question, please contact us or submit the inquiry form below.
