CM-PATRACritical Minerals Patent and Technology Research Analytics

Patents and R&D for India's critical minerals

Patent families and funded R&D projects for India's critical minerals. Four software agents ingest, classify, resolve and audit the records; every label below carries the rule or reviewer that produced it. Data as of 2026-10-16.

Read this first. Coverage is partial. Google Patents holds few Indian-office filings for 2016–2026, so recent Indian filings are under-counted, and the R&D lists (Ministry of Mines S&T projects on SATYABHAMA, and JNARDDC's ongoing and completed projects) give titles and status, not outcomes. Counts are lower bounds, not market totals. Recent Indian filings are tracked separately from the Indian Patent Office Journal: issues 23/2026 to 42/2026 (82 mineral-technology applications); see method.

371

patent families kept after classification

210

with an Indian applicant

156

naming an India-list mineral

68/296

R&D projects naming an India-list mineral

Alerts (rule-based)

  • Niobium / Tantalum: 2 funded project(s) but no patent family found.
  • Indian-applicant filings published 2025–2026: most active India-list minerals Lithium (11), Nickel (9), Rare earths (4).

By mineral

India's list only

All minerals, including those not on India's critical-minerals list.

By processing stage

  • Recycling / secondary sources112
  • Ore preparation / agglomeration95
  • Hydrometallurgy (leaching, solvent extraction)93
  • Electrochemical recovery61
  • Beneficiation / separation46
  • Pyrometallurgy / smelting35
  • Compounds and materials26
  • Refining24
  • Other17

A family can sit in more than one stage (for example leaching and recycling), so these bars add up to 509, more than the 371 families.

By publication year

  • 20005
  • 20012
  • 200213
  • 200317
  • 200416
  • 20056
  • 20063
  • 20073
  • 20084
  • 20096
  • 20108
  • 20116
  • 20126
  • 201310
  • 20146
  • 2015141*
  • 201610
  • 201711
  • 20186
  • 20196
  • 20207
  • 20216
  • 202210
  • 202318
  • 202413
  • 202520
  • 202612

* 2015 (141): a source artefact, not a trend. Google Patents Public Data holds a dense batch of Indian-office publications dated 2015 and few for 2016–2026, so the bar reflects where the source has records, not how much patenting happened. Recent Indian filings are tracked from the IPO Journal instead.

Top applicants

  • CSIR (Council of Scientific & Industrial Research)39
  • Tata Steel22
  • Attero Recycling16
  • Hindalco Industries10
  • Outotec Finland Oy10
  • Vale Sa8
  • Cytec Tech Corp7
  • Tech Resources Pty Ltd7
  • Aditya Birla Science & Technology7
  • Lohum6
  • IIT Madras6
  • Metastable Materials6

Technology-gap matrix (India list)

Patent families against funded projects, fewest first. “0 / 0” means nothing found in these sources, not that nothing exists.

Research ahead of patents: Beryllium / Rhenium / Cadmium (1,977 papers per Indian patent); Zirconium / Hafnium (1,355 papers per Indian patent); Tungsten (890 papers per Indian patent); Selenium / Tellurium (768 papers per Indian patent); Graphite (731 papers per Indian patent). Ranked by Indian research papers (OpenAlex, 2016–2025, titles or abstracts with extraction or processing terms) per Indian-applicant patent family in CM-PATRA, patents floored at 1. A ranking, not a verdict: patent counts are floors, so a high ratio can also reflect patents missing from the source.

MineralPatentsIndian applicantFunded projectsIndian papersPapers per Indian patent
Zirconium / Hafnium1011,3551,355
Niobium / Tantalum002459459
Antimony / Bismuth2201,199600
Beryllium / Rhenium / Cadmium2101,9771,977
Selenium / Tellurium112768768
Tungsten4312,670890
Gallium / Germanium / Indium114724724
Molybdenum5211,337669
Vanadium651724145
Tin534196
Cobalt8532,539508
Graphite7543,653731
Platinum-group metals12822,450306
Potash / Strontium / Phosphate1210384284
Silicon13962,380264
Titanium221545,994400
Nickel292384,950215
Rare earths2015192,233149
Copper3318108,682482
Lithium3231122,48680

Agent run log

What each agent did on this build. The auditor is a deterministic gate; nothing is published if it fails.

  1. scout · ingest · 438 → 438
    BigQuery SQL on patents-public-data.patents.publications. One row per patent family; Indian-office publications and Indian-applicant families in the seed CPC set, publication year 2000 onward.
  2. scout · ingest · 221 → 168
    Targeted BigQuery SQL (scripts/patra/bigquery-focus-extract.sql), run 17 Oct 2026. 221 Indian families matched red mud, gallium, germanium, indium, scandium, alumina-chain or fly-ash rules; 53 were already in the seed extract.
  3. classifier · classify · 606 → 371
    CPC prefix rules, then title/abstract keyword rules; reviewed overrides applied last. 141 labelled by text where CPC gave no mineral; 6 reviewed overrides.
  4. resolver · entity-resolution · 843 → 827
    Alias rules over harmonised applicant names; cue words for type. Input = distinct published names; output = distinct canonical names.
  5. scout · ingest · 130 → 130
    Public SATYABHAMA 'Ongoing Projects' list, read from the portal's paged table. Titles, project numbers and status only. The public list does not name the institution or give outcomes.
  6. scout · ingest · 187 → 166
    JNARDDC public 'Ongoing Projects' and 'Projects Archive' pages, read in a browser on 2026-10-17. Number, title, status and (ongoing only) sponsor. Completed projects carry the year the page shows, if any.
  7. resolver · resolve · 187 → 21
    Title word overlap (Jaccard 0.6 or more) between JNARDDC and SATYABHAMA projects. Same project in both lists, kept once: SNTMOM/1445/2025 = JNARDDC S-61 (title overlap 1.00); SNTMOM/1239/2024 = JNARDDC S-57 (title overlap 1.00); SNTMOM/1167/2024 = JNARDDC S-56 (title overlap 1.00); SNTMOM/1206/2024 = JNARDDC S-55 (title overlap 0.86); SNTMOM/1049/2024 = JNARDDC S-52 (title overlap 1.00); SNTMOM/1171/2024 = JNARDDC S-51 (title overlap 1.00); SNTMOM/125/2020 = JNARDDC S-35 (title overlap 0.82); SNTMOM/529/2021 = JNARDDC S-36 (title overlap 1.00); SNTMOM/414/2021 = JNARDDC S-37 (title overlap 0.82); SNTMOM/108/2020 = JNARDDC S-38 (title overlap 0.79); SNTMOM/605/2021 = JNARDDC S-39 (title overlap 0.67); SNTMOM/156/2020 = JNARDDC S-40 (title overlap 0.63); SNTMOM/169/2020 = JNARDDC S-42 (title overlap 0.75); SNTMOM/745/2022 = JNARDDC S-43 (title overlap 1.00); SNTMOM/796/2022 = JNARDDC S-44 (title overlap 0.86); SNTMOM/847/2022 = JNARDDC S-45 (title overlap 1.00); SNTMOM/141/2020 = JNARDDC S-46 (title overlap 0.85); SNTMOM/295/2020 = JNARDDC S-47 (title overlap 1.00); SNTMOM/961/2023 = JNARDDC S-48 (title overlap 0.92); SNTMOM/991/2023 = JNARDDC S-49 (title overlap 1.00); SNTMOM/1093/2024 = JNARDDC S-50 (title overlap 0.80).
  8. classifier · classify · 296 → 68
    Title keyword rules; reviewed overrides applied last. 2 reviewed overrides. Output = projects naming an India-list critical mineral.
  9. auditor · audit-gate · 902 → 902
    Deterministic checks: identity present, date order, id format. No failures. Data-quality warnings (kept, flagged): publication precedes filing in the source: IN-2015MN00488-A.
  10. scout · ingest · 41572 → 1010
    IPO Journal 23/2026 to 42/2026 (20 issues): PDF text extraction, broad IPC/title filter. 1151 broad hits. 5 issue(s) read by hand (41/2026, 40/2026, 39/2026, 38/2026, 37/2026; 41/2026 stores only the 6 it kept); 15 issue(s) ingested automatically by scripts/patra/ipo-journal-fetch.mjs.
  11. classifier · classify · 1010 → 82
    Ontology (CPC then title words); iron/steel dropped; reviewed exclusions carry reasons; automated candidates kept only with an extraction/separation/recycling IPC class (C22B, C25C, B03, B09B, H01M10/54, H01M6/52) and a nameable mineral. 17 excluded with reasons, 824 automated candidate(s) dropped (no extraction, separation or recycling class), 87 dropped (no mineral named by IPC or title), 6 low-confidence (generic class only), 36 with an Indian applicant (name list or an Indian address). Kept separate from the patent counts.
  12. scout · ingest · 40 → 47441
    OpenAlex works API, 20 minerals x 2 requests (India-affiliated by year; worldwide count), 2016-2025. Bibliometric proxy: papers with a mineral term and a processing term in title or abstract. A paper can count under several minerals. Not merged into patent or project counts.

Sources

  • Google Patents Public Data (BigQuery) · CC BY 4.0. Indian-office publications are dense for 2015 and sparse for 2016–2026 (about 100–260 a year, mostly without CPC or abstracts); Indian-applicant filings at other offices are complete. Counts are therefore lower bounds for recent Indian-office filings. Seed extract 13 Oct; targeted red mud / gallium / germanium / indium / scandium / fly-ash extract 17 Oct (221 families, 168 new, 141 dropped as not about extraction or processing).
  • SATYABHAMA portal, Ministry of Mines — Ongoing Projects (public list) · Public government web page; used for analysis with attribution. 130 ongoing projects under the Ministry's S&T programme (the portal reports 1,681 registered and 0 completed). One funding scheme, not all Indian critical-minerals R&D.
  • JNARDDC (Jawaharlal Nehru Aluminium Research Development and Design Centre, Ministry of Mines) — Ongoing Projects and Projects Archive pages · Public government web pages; used for analysis with attribution. 20 ongoing and 167 completed projects as listed (ongoing page last updated 15 Oct 2026). Titles, numbers, sponsors for ongoing projects, years where shown; no outcomes. Projects also on SATYABHAMA are kept once.
  • OpenAlex works API (research papers) · CC0. Papers 2016-2025 with a mineral term and a processing term in title or abstract; India = at least one author affiliated with an Indian institution. Affiliation data in OpenAlex is incomplete, so counts are indicative.

JSON: /api/cmpatra/overview, /api/cmpatra/search