--- title: "Experimental structures in Ouro's crystal database" description: "The Crystallography Open Database is now a public dataset on Ouro, and novelty checks now say whether a match was observed experimentally." url: "https://ouro.foundation/posts/mmoderwell/experimental-structures-in-ouros-crystal-database" id: "01a15ea0-63fd-7daf-b183-3a5ba6ac5298" type: "post" --- # Experimental structures in Ouro's crystal database Post by @mmoderwell in #materials-science (all) · 2026-10-21 - Asset id: `01a15ea0-63fd-7daf-b183-3a5ba6ac5298` - Page: https://ouro.foundation/posts/mmoderwell/experimental-structures-in-ouros-crystal-database Until now, "known" in Ouro's [Crystal DB](https://ouro.foundation/services/mmoderwell/crystal-db) meant known to a DFT database. Now the novelty check also tells you whether a structure has been **observed experimentally**, and against which records. ## The Crystallography Open Database, on Ouro [COD](https://www.crystallography.net/cod/) is the largest open collection of experimentally determined crystal structures. All of it is now a public dataset on Ouro: - **534,739 entries.** Every released entry, with its formula, cell, space group, temperature, R factor and the paper it came from. - **76,065 CIFs.** We keep the CIF for every inorganic entry, as COD publishes it. Organic and metal-organic entries keep their metadata and link to COD. - **CC0.** Use it however you like, and please credit each structure's authors. > Embedded dataset: [Crystallography Open Database](https://ouro.foundation/datasets/mmoderwell/crystallography-open-database) Most of COD's inorganic structures were published before 2000; organic structures took over after that. The dataset's Views tab also breaks the inorganic ones down by crystal system, number of elements, and most common elements. ## Experimental flags on Materials Project Materials Project records whether each material has been observed, and which ICSD entries it was matched to. Our MP records now carry both: - **48,892** of our 152,620 MP materials are flagged experimental. - **51,225** carry ICSD collection codes. > [!NOTE] > Why not ICSD itself? Its license doesn't allow redistribution, so Ouro stores only the ICSD codes Materials Project links to, never ICSD data. COD is the open alternative. ## What changed in the routes (all free) **Novelty** now matches against COD alongside Materials Project, Alexandria and GNoME. Each match says whether it was observed experimentally, and the response has an `experimentally_known` flag. Here it is on COD's rock-salt NaCl: > Embedded route: [Check whether a crystal structure is already known](https://ouro.foundation/routes/mmoderwell/check-whether-a-crystal-structure-is-already-known) > NaCl (Fm-3m) matches mp-22862 (Materials Project) and 9 other published structures. It was observed experimentally (ICSD 18189, 28948, 41411). **Materialize** takes COD ids too: "COD 1000041", the bare number, or a COD link. The CIF lands on Ouro under CC0, with the structure's own paper as its citation. > Embedded route: [Get the CIF of a Materials Project, Alexandria, GNoME or COD material](https://ouro.foundation/routes/mmoderwell/get-the-cif-of-a-materials-project-alexandria-gnome-or-cod-material) Two pulled in that way: NaCl from 1965, and α-quartz from 1935. > Embedded file: [NaCl (Fm-3m): Crystallography Open Database 1000041](https://ouro.foundation/files/mmoderwell/nacl-fm-3m-crystallography-open-database-1000041) > Embedded file: [SiO2 (P3121): Crystallography Open Database 1011097](https://ouro.foundation/files/mmoderwell/sio2-p3121-crystallography-open-database-1011097) ## Good to know - **COD has no energies,** so it's prior art for novelty, not part of the hull. - **Matching uses the CIF as written.** Partial occupancies and missing hydrogen positions match less often, and COD's formula field can differ from what its coordinates describe. `reduced_formula` and `num_atoms` in the dataset describe the coordinates. - **"New" still means new to these databases,** not to the literature. - **The code is open:** [crystal-db](https://github.com/ourofoundation/crystal-db) has the COD build script (`scripts/cod.py`) and the matching code. Please cite COD when you use it: [Gražulis et al. 2009](https://doi.org/10.1107/S0021889809016690). ## Assets in this post - [Crystallography Open Database](https://ouro.foundation/datasets/mmoderwell/crystallography-open-database) (dataset, id `01a15c82-afe9-73cc-b0de-759388e72f42`) - [Check whether a crystal structure is already known](https://ouro.foundation/routes/mmoderwell/check-whether-a-crystal-structure-is-already-known) (route, id `3dced825-cfea-4711-89a9-1d7998242001`) - [Get the CIF of a Materials Project, Alexandria, GNoME or COD material](https://ouro.foundation/routes/mmoderwell/get-the-cif-of-a-materials-project-alexandria-gnome-or-cod-material) (route, id `e4812d4a-caae-4c11-a1a1-ff92af687d8d`) - [NaCl (Fm-3m): Crystallography Open Database 1000041](https://ouro.foundation/files/mmoderwell/nacl-fm-3m-crystallography-open-database-1000041) (file, id `ba38da65-9aa8-4fe8-a818-d1f66d27a04c`) - [SiO2 (P3121): Crystallography Open Database 1011097](https://ouro.foundation/files/mmoderwell/sio2-p3121-crystallography-open-database-1011097) (file, id `059a41f2-723e-488f-80bc-f4e761f2acf8`) - [Crystal DB](https://ouro.foundation/services/mmoderwell/crystal-db) (service, id `9c71b718-32fa-4fc1-855f-688aa363b663`)