← Benchmarks | Electronic Transport

Carrier Mobility BENCHMARK

Hole and electron mobility at 300 K predicted from composition alone, scored on15 semiconductors selected by agreement within1% between two hole-mobility references. Electron mobility is evaluated against the single available electron reference. Median error 6.6% for holes and 9.3% for electrons; MAPE 21.7% and 22.2%, driven by a few outliers shown below. Run 22 October 2026.

6.6%
Hole median error
MAPE 21.7%
9.3%
Electron median error
MAPE 22.2%
11/15
Holes within ±15%
electrons 9/15
15
Agreed-reference materials
elemental, III-V, II-VI, IV-VI

Why this cohort

Two reference sets, scored only where they agree

The hole-mobility references disagree on 8 of 23 materials — for example SiC, where one lists 120 cm²/V·s and the other 40. The 15-material cohort is selected where these two hole references agree within 1%. Both carrier types are evaluated on those same 15 materials; electron mobility is scored against its single available reference. Mean absolute percentage errors are 21.7% for holes and 22.2% for electrons; the 6.6% and 9.3% headlines are medians.

Family Scorecard

Mean absolute percentage error by material family

Family N Hole MAPE Electron MAPE Largest error
Elemental (Si, Ge) 2 0.6% 0.8% Ge
III-V 9 23.5% 23.2% AlP
II-VI 3 36.3% 36.9% ZnTe
IV-VI 1 3.9% 12.2% PbS
ALL 15 21.7% 22.2% AlP, AlAs, ZnS

Per-Material Predictions

300 K mobilities in cm²/V·s; signed error = (predicted − experimental) / experimental

Material μh exp μh FluxMateria Error μe exp μe FluxMateria Error
Si 450 449 -0.2% 1400 1401 +0.0%
Ge 1900 1880 -1.0% 3900 3837 -1.6%
AlAs 200 353 +76.5% 200 378 +88.9%
AlN 14 14 +0.3% 135 136 +0.8%
AlP 450 30 -93.3% 80 31 -61.8%
GaAs 400 392 -2.1% 8000 8269 +3.4%
GaN 200 184 -7.9% 1000 907 -9.3%
GaP 150 169 +12.4% 250 209 -16.6%
GaSb 1000 971 -2.9% 3000 2912 -2.9%
InAs 500 467 -6.6% 40000 37977 -5.1%
InSb 850 773 -9.0% 77000 61520 -20.1%
CdTe 100 103 +3.3% 1050 1145 +9.1%
ZnS 5 8 +65.8% 165 221 +34.1%
ZnTe 100 140 +39.7% 340 569 +67.4%
PbS 700 673 -3.9% 600 673 +12.2%

All predictions from composition only. Download: carrier_mobility_agreed_cohort_2026-10-22.json.

Error Distribution

Share of the 15 materials within each error band

7 / 5
holes / electrons within 5%
11 / 9
within 15%
11 / 11
within 25%

The median material is predicted within 6.6% (holes) and 9.3% (electrons). The mean is pulled up by a few hole outliers (AlP 93%, AlAs 77%, ZnS 66%, ZnTe 40%); they are reported, not excluded.

Comparison with DFT and ML

Carrier mobility prediction trade-offs: accuracy, speed, and data dependence

Metric FluxMateria DFT (Boltzmann transport) ML (CGCNN / GNN)
μe error Median 9.3%, MAPE 22.2% (15 materials)
holes: median 6.6%
20–100% typical (per material) 30–200% MAPE on out-of-domain
Required input Composition only Full crystal structure + scattering parameters Composition + featurized graph + training set
Runtime Sub-millisecond Hours per material Milliseconds inference, hours training
Training data None None (calibrated functional) Thousands of labeled mobility values
Fitted parameters Fixed calibration, no per-material terms Pseudopotential + scattering rate cutoffs Millions
Novel compositions Physics-grounded extrapolation Recompute required Often degrades beyond training domain

Key takeaway: From composition alone and at sub-millisecond runtime, FluxMateria predicts 300 K hole and electron mobility with a median error of 6.6% and 9.3% on the 15 semiconductors where the experimental references agree. A few outliers (AlP, AlAs, ZnS, ZnTe holes) dominate the mean; DFT-based Boltzmann transport remains the ab initio reference but requires the full crystal structure and hours per material.

What is carrier mobility — and why does it matter?

The single number that decides how fast a semiconductor can switch, sense, or convert energy

The property in plain language

Carrier mobility (μ) measures how quickly electrons (or holes) drift through a material when an electric field is applied. It is the proportionality constant between drift velocity and field, reported in cm²/V·s. A higher mobility means current flows more easily, switches turn on and off faster, and less energy is lost to heat for the same delivered current.

Mobility is intrinsic to the material: it is set by composition, bonding, and how strongly carriers scatter off the lattice. It is what separates a 1,400 cm²/V·s silicon wafer from an 8,500 cm²/V·s GaAs wafer — and is the reason GaAs dominates RF and Si dominates logic.

Read more — economic stakes, target audiences, and workflow impact click to expand ↓

Why mobility is one of the most economically important material constants on Earth

  • Switching speed. Transistor cutoff frequency scales linearly with mobility. Doubling mobility roughly doubles the maximum clock speed of an analog circuit and halves the switching delay of a digital one.
  • Power efficiency. Conduction loss in a power transistor scales as 1/μ. A 10× mobility advantage translates directly into a 10× lower on-resistance for the same die area — which is the entire commercial story of GaN and SiC power electronics versus silicon.
  • Solar & detector performance. Carrier diffusion length, the distance a photo-generated carrier travels before recombining, is set by mobility × lifetime. It determines collection efficiency in solar cells and signal-to-noise in photodetectors.
  • Sensor sensitivity. Hall sensors, magnetoresistive readers, and gas sensors get their sensitivity directly from mobility — high-mobility InSb and graphene are the basis of the most sensitive room-temperature magnetic-field sensors.
  • Display brightness & refresh rate. Backplane mobility (a-Si vs. IGZO vs. LTPS) sets the maximum pixel switching rate — the difference between a 60 Hz LCD and a 120 Hz OLED smartphone display.

Who needs to predict mobility — and what they do with it

Semiconductor device engineers

Pick the channel material for a new transistor generation. Mobility decides whether a candidate clears the speed/power target before any wafer is grown.

Power electronics designers

Compare wide-bandgap candidates (GaN, SiC, Ga₂O₃, diamond) for EV inverters, data-center power, and grid converters where every percent of efficiency is billions in fuel and cooling.

Photovoltaic researchers

Screen new absorber chemistries (perovskites, chalcogenides, kesterites) for adequate carrier collection before committing months to film growth and device fabrication.

Thermoelectric & flexible-electronics teams

Maximise the power factor σS² (thermoelectrics) or organic-TFT mobility (printed displays, RFID, biosensors), where mobility is the single biggest lever on figure of merit.

Materials-discovery groups

Rank thousands of candidate compositions per day for transport quality — something DFT cannot do at that throughput, and ML surrogates cannot do reliably outside their training set.

Foundries & IP licensors

De-risk supplier qualification and process-node decisions where a wrong call on a channel material costs nine figures and 18 months of lost time-to-market.

Why a fast mobility predictor changes the workflow

Today, getting a trustworthy mobility number means either a multi-week DFT + Boltzmann-transport pipeline or an ML model that only works on materials similar to its training set. Both are bottlenecks the moment you want to screen a new chemistry, an alloyed composition, or a defect-engineered variant.

A composition-only predictor that returns μe in milliseconds, with benchmark evidence used for validation, collapses that bottleneck. It moves mobility from a downstream verification step into an upstream design variable that every materials and device team can iterate against in real time.

Methodology

How FluxMateria predicts hole and electron drift mobility

Benchmark Method Summary

Hole and electron mobilities are computed from composition by the production electronic-transport route. The engine identifies the bonding family, derives the carrier transport channel, and returns μh and μe at room temperature.

  • 15 semiconductors selected by agreement between the hole-mobility references; one independent electron reference is available
  • 4 material families: III-V, II-VI, IV-VI, elemental sp³
  • Mobility range: from single digits (ZnS holes) to thousands of cm²/V·s (InSb, GaAs electrons)
  • Conditions: 300 K, lightly-doped or intrinsic single-crystal samples
  • Metrics: median absolute percentage error and MAPE versus experiment
  • Basis: physics-based transport model with a fixed calibration; no per-material terms

Material Families Covered

Elemental sp³

Si, Ge

III-V direct-gap

GaAs, GaSb, InN…

III-V indirect-gap

AlAs, AlP, AlSb, GaP

III-V wurtzite

GaN, AlN, InN

II-VI semiconductor

ZnO, ZnS, ZnSe, ZnTe, CdS, CdSe, CdTe

IV-VI narrow-gap

PbS, PbSe, PbTe

IV-IV indirect

SiC

Scope & Limitations

Strengths

  • Median error 6.6% (holes) and 9.3% (electrons) on 15 agreed-reference semiconductors
  • Si, Ge, AlN, GaAs and GaSb hole mobilities within ±3%
  • Composition-only input: no crystal structure required
  • Sub-millisecond runtime enables mobility-aware materials screening
  • Fully reproducible benchmark artifacts and reference-source notes
  • Covers industrially dominant semiconductors (Si, GaAs, GaN, Ge, InP) plus the broader research-grade space (III-V indirect Al-X, II-VI optoelectronics, IV-VI thermoelectrics)

Known Limitations

  • Hole mobilities of AlP (93%), AlAs (77%), ZnS (66%) and ZnTe (40%) are the main open errors: indirect-gap Al-V compounds and wide-gap II-VI holes
  • Eight further materials (including SiC) are not scored because the two reference sets disagree on them
  • Effective mass (m*) currently 12.5% MAPE — sister benchmark; some heavy-fermion families remain under active development
  • Temperature dependence beyond 300 K uses a simple scaling; full T-dependent transport under development
  • Validation to date covers single-crystal intrinsic / lightly-doped samples; heavily-doped or polycrystalline regimes are not yet in scope

References

Primary data sources for experimental validation

  1. O. Madelung (ed.), Semiconductors: Data Handbook, 3rd ed., Springer, 2004.
  2. P. Y. Yu and M. Cardona, Fundamentals of Semiconductors, 4th ed., Springer, 2010.
  3. S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, 3rd ed., Wiley, 2007.
  4. NSM Semiconductor Database (Ioffe Institute), ioffe.ru/SVA/NSM/Semicond (accessed 2026).

Benchmark basis

Electron mobility is computed from Flux transport physics with fixed endpoint scaling. The semiconductor reference set is used to measure accuracy across material families.

Flux-Calibrated Physics

Try the Materials module

Enter any semiconductor formula and get electron mobility predictions alongside band gap, elastic, thermal, and optical properties — all from composition alone.

← Back to Module Request Access