One famous database for chemical substances is PubChem. If one reads the section about other identifiers than the PubChem ID for Ethanol one gets several CAS numbers (potentially deprecated ones as well), potentially several European Community (EC) Numbers and a lot of other identifiers which seem to be not really unique.

Is there a true unique identifier for chemical substances reasonable to be used in a database?

  • 1
    $\begingroup$ chemistry.stackexchange.com/questions/34563/… is relevant, although not a direct answer. The short answer is : Yes, InChI but there are alternatives $\endgroup$
    – gilleain
    Mar 23, 2020 at 16:05
  • $\begingroup$ This one is more relevant : chemistry.stackexchange.com/questions/110057/… $\endgroup$
    – gilleain
    Mar 23, 2020 at 16:07
  • 5
    $\begingroup$ The point of CAS numbers is that they label mixtures and preparations, not just pure chemicals. If you want a 1:1 mapping from a chemical to a database identifier, CAS numbers aren't even the right tool to look at in your case. $\endgroup$
    – Antimon
    Mar 23, 2020 at 16:08
  • $\begingroup$ @gilleain Seems like there is ununiqueness cause of "standard" and "non-standard" InChi as well with InChi. $\endgroup$
    – thinwybk
    Mar 23, 2020 at 16:24
  • 3
    $\begingroup$ In additon, polymorphs of the same compound may be filed under different CAS register numbers, too. Example: $\ce{TiO2}$, as anatase (1317-70-0), and as rutil (131-80-2) -- both tetragonal, but either $I4_1/amd$, packing fraction 70%; or $P4_2/mnm$, packing fraction 77%. Source: Cardelli, Materials Handbook, 3rd edition 2018, Springer Cham/Switzerland, ISBN 978-3-319-39823-3, p. 910, table 10.8. And since polymorphs may differ, e.g. in dissolution kinetics, this is relevant for pharmaceutical materials, too. $\endgroup$
    – Buttonwood
    Mar 23, 2020 at 16:33


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.