Research Articles
Materials science articles written from our data: kill reports on approaches that failed, methods, and state-of-technology landscapes. Auto-generated compound data profiles are listed separately under their own filter.
21 articles · 114 auto-generated compound data profiles are under the Compound Profile filter
Zeolites: Structure, Framework Types, and Industrial Uses
What are zeolites? How Si/Al ratio, pore size, and framework types (FAU, MFI/ZSM-5, LTA, CHA) drive their use in FCC catalysis, detergents, and NOx SCR.
Read articleWhat Is a Band Gap? Direct, Indirect, and DFT Band Gaps
A band gap sets whether a solid is a metal, semiconductor or insulator. Learn direct vs indirect gaps, how gaps are measured, and why DFT underestimates them.
Transparent Conducting Oxides: ITO and Its Alternatives
What makes a transparent conducting oxide work, why indium tin oxide dominates, and how FTO, AZO, silver nanowires, and graphene compare as ITO alternatives.
Sodium-Ion Battery vs Lithium-Ion: Chemistry, Cost, and Status
How a sodium ion battery works, its cathode and hard-carbon anode materials, and how its cost, energy density and commercial status compare with lithium-ion.
Materials Project vs OQMD vs AFLOW vs JARVIS: A Fair Comparison
How the Materials Project, OQMD, AFLOW and JARVIS-DFT differ in size, DFT functional, license and access, why their values disagree, and when to use each.
Lithium Iron Phosphate Battery vs NMC: LFP vs NMC Cathodes Compared
Lithium iron phosphate battery vs NMC: how LFP and NMC 811 cathodes compare on energy density, cycle life, thermal safety, cost, and cobalt and nickel use.
High-Temperature Superconductors: Cuprates, Iron-Based and Hydrides
What makes a superconductor high-temperature: cuprates like YBCO, iron-based compounds, MgB2, pressurized hydrides, the LK-99 claim, and real applications.
GaN vs SiC: Wide Bandgap Semiconductors Compared
Silicon carbide vs gallium nitride: how these wide bandgap semiconductors compare on band gap, breakdown field, heat, cost, and where each wins in power and RF.
Critical Minerals and Rare Earth Elements: Supply, Uses, Risks
What makes a mineral critical, the 17 rare earth elements and their uses, who mines and refines them, battery mineral concentration, and how recycling can help.
Battery Recycling: How Lithium-Ion Cathode Materials Are Recovered
How lithium-ion battery recycling works: black mass, pyrometallurgy vs hydrometallurgy vs direct recycling, LFP economics, and EU recovery targets.
63 GB of Data That Never Made It to Production
A sync-drift audit revealed 1,117 files (~63 GB) existing only locally, never uploaded to production. 32 common database tables had differing row counts with local exceeding production by 12.6M rows. A DigitalOcean key-naming bug duplicated source prefixes in every object path. Zero data corruption — but massive invisible drift.
Adjacent Is Not Preemptive: The Master Rule of Competitive Disproof
Many of our audited disproof kills did not hold up on review — the cited competitor often did something adjacent, not identical. Raw dataset does not equal scoring API; language analysis does not equal scientific analysis. Learning to distinguish adjacency from preemption saved our best candidates from premature death.
13 Ways Computational Materials Science Goes Wrong
A catalog of interpretation errors and compute failure modes from our materials discovery pipeline: thermostat overshoot (625K target / 750K actual), MACE artifacts on perovskites, PBE bandgaps presented without HSE correction, metallic DFPT results, phonon instabilities, 53 files with THz vs cm⁻¹ confusion, and more.
The Data Plane Kills Ideas That Market Analysis Can't
Four invention candidates rated 'plausible' by web-search-based competitive analysis were instantly killed when tested against real warehouse data. The patent-clear composition finder's core anti-join produced false results because the canonical crosswalk was never built, patent formulas were stored as un-reduced supercells, and 12.4% of formula values were filesystem paths.
4 Strategies to Run Quantum ESPRESSO on Modal, All Failed
We tried four image-registry strategies to run Quantum ESPRESSO on Modal's serverless GPU infrastructure. apt_install hangs at tasks=0, ephemeral detach produces the same symptom, GHCR returns 403 Forbidden, and Docker Hub returns 'manifest not found.' The only viable path is a custom Dockerfile with QE compiled from source.
NULL Should Never Mean Favorable: How Default-Optimistic Screening Corrupts Results
A data-stitching audit revealed that missing phonon data defaulted to 'stable,' missing patent counts defaulted to 'novel,' and missing PFAS flags defaulted to 'clean' across our screening pipeline. Every screen that treats NULL as favorable silently promotes unvalidated candidates. Here's what we found and how we fixed it.
When All 13 Compute Results Were Error Payloads
Every single Modal QE DFPT and HSE06 result from our April 2026 GPU compute campaign was an error payload, not usable data. 13 files landed on the volume with status ERROR, PSEUDO_ERROR, SCF_FAILED, DFPT_FAILED, PARSE_ERROR, or NOT_CONVERGED. The honest-null pattern kept failed values out of our ledger; downstream screens still had to handle nulls.
Products Survive, Compounds Die: What 120 Invention Candidates Taught Us
After evaluating 120+ invention candidates across 5 rounds of a structured generate-critique-disprove-audit cycle, we found that service/API ideas survived at ~27% while compound-level ideas survived at 0%. Every compound promotion was killed as prior art. The anti-portfolio reveals systematic patterns in what kills ideas.
Why the Most Important Problems Are the Worst Investments
Three domains scored 96-97/100 on priority and produced zero surviving candidates across 120+ evaluations. PFAS-free chrome fume suppression, rare-earth element separation, and PFAS-free immersion coolants all failed because funded incumbents captured the forcing-function budget before we arrived. The counterintuitive lesson: high priority is a warning, not a green light.
State of Solid-State Batteries in 2025: What the Data Actually Shows
Solid-state batteries promise higher energy density, faster charging, and improved safety over liquid electrolyte cells. But our reading of computational screening data on 14,000+ candidate electrolytes and of 340 patent filings (our counts; protocols and lists not shown here) suggests to us a gap between press releases and material reality. Here is what the data says about sulfide, oxide, and polymer electrolyte families — and which approaches are closest to commercialization.
Why LiCoO2 Fails as a Next-Generation EV Battery Cathode
Lithium cobalt oxide dominated consumer electronics for decades, but scaling it to electric vehicles reveals fundamental structural, thermal, and economic failures. We reviewed 47 computational studies and 12 experimental datasets (nearly 15,000 structural relaxations and 1,200+ cells by our count; sources not listed here), which indicate that LCO physics limits it for automotive applications despite engineering efforts to date.
Want to run this kind of analysis on your own materials data?