Persistent Identifier (PID)

A unique, actionable, long-term identifier assigned to a single entity in the research ecosystem -- a work, a person, an organisation, a piece of equipment, a project, or a dataset -- that is registered with a scheme-specific provider, stays associated with that entity even as its location or attributes change, resolves (typically over HTTP) to current metadata or the resource itself, and is designed to remain valid for decades, well beyond the lifetime of any one repository, publisher, or institutional system.

ByCASRAI Editorial Board

· Last updated 17 Jul 2026

Worked examples

Is an instance

A journal article assigned the DOI 10.1038/nphys1170 by Crossref remains resolvable at doi.org even after the publisher migrates platforms.

Is an instance

A researcher's ORCID iD (e.g. 0000-0002-1825-0097) stays the same identifier across every institution, funder, and publisher they interact with over a career.

Looks similar, but isn't

Not an instance

A relative link on a personal webpage is not a PID -- it identifies a location, not an entity, and has no registration authority committed to keeping it resolvable.

Not an instance

An internal database primary key (an autoincrementing row ID) is not a PID until a real PID is minted and cross-referenced to it.

A persistent identifier (PID) is a unique, actionable identifier assigned to one entity — a research output, a person, an organisation, a piece of research equipment, a grant, or a dataset — that is designed to remain valid and resolvable for the long term, independent of where the entity is currently hosted or which system last updated it. “Persistent” is a governance commitment, not a technical property of the string itself: a PID scheme only functions as advertised if a registration authority maintains resolution infrastructure and metadata over time. See PID provider and resolution service.

What makes an identifier a PID, not just an identifier

Three properties distinguish a genuine PID from an ordinary internal ID (a database row number, a URL, a local accession code):

Uniqueness within a scheme. A PID is issued once, by an authorised registration agency, and is never reissued to a different entity.
Resolvability. Given the identifier, a resolution service can return the entity’s current location or its registered metadata — typically via HTTP redirect — even after the underlying resource has moved. See PID resolution.
Governed persistence. A PID provider makes an institutional commitment to keep the identifier resolvable over decades, backed by a scheme’s technical infrastructure (a Handle server, a DOI registration agency’s systems) rather than by any single organisation’s web server staying online. See PID provider and PID consortium.

A plain URL fails the third property: it identifies a location, not an entity, and breaks the moment the resource moves (“link rot”). A PID is designed to survive exactly that kind of change by separating the stable identifier from the resource’s current address.

The major PID schemes in research

No single PID covers every entity type. Research infrastructure instead relies on a small set of schemes, each scoped to a different kind of thing, that are increasingly cross-referenced with one another:

DOI (Digital Object Identifier) — the dominant PID for scholarly works and, via DataCite, for datasets and other research outputs. Governed by the DOI Foundation under ISO 26324, with Crossref and DataCite as the two Registration Agencies most relevant to research. See persistent data identifier for the dataset-specific case.
ORCID iD — the leading PID for individual researchers, free and researcher-controlled, used to disambiguate a person across name variants, career moves, and publisher systems. See researcher identifier, which covers the person-level PID landscape (ORCID, ISNI-for-individuals, Scopus Author ID, ResearcherID) in more depth than this general PID overview does.
ROR (Research Organization Registry) — a free, open registry launched in January 2019 for identifying research-performing institutions, funders, and other organisations, built from a Crossref/DataCite/California Digital Library collaboration and now the default institutional PID referenced in DataCite and Crossref metadata.
ISNI (International Standard Name Identifier) — an ISO 27729 identifier for public identities generally, spanning both individual people and organisations across creative and research sectors; ORCID iDs and ISNI records are cross-referenced within a reserved ISNI block.
Handle System / ARK — lower-level PID infrastructure. DOIs are themselves implemented on top of the Handle System; ARKs (Archival Resource Keys) are used by some libraries and archives as an alternative persistent-identifier scheme, notably for non-DOI-scale objects.
RRID (Research Resource Identifier) — a PID scheme scoped to research resources such as antibodies, cell lines, model organisms, and software tools, distinct from the identifiers above in that it identifies a type of resource used, not a unique physical or digital instance.

Each scheme has its own registration agency, its own metadata schema, and its own resolution infrastructure, but all share the uniqueness/resolvability/governed-persistence structure above — this is what “PID” describes as a general concept, and why this general concept merits its own dictionary entry distinct from any one scheme.

Why PIDs matter for research infrastructure

PIDs are the connective layer that lets research-information systems (CRIS platforms, repositories, funder systems, publisher submission systems) exchange data about the same entity without relying on free-text name matching. A funder system that records a grant by its PID, linked to the PI’s ORCID iD and the funded institution’s ROR ID, can programmatically trace that grant’s outputs across publisher and repository systems years later — something string-matching on names and titles cannot do reliably. This underpins machine-actionable research-information exchange promoted by initiatives such as the Research Data Alliance and euroCRIS, and is a precondition for FAIR data’s “findable” and “interoperable” principles, which explicitly call for globally unique, persistent identifiers for data and metadata.

Examples

A journal article assigned the DOI 10.1038/nphys1170 by Crossref remains resolvable at doi.org even after the publisher migrates platforms or changes its URL structure.
A researcher’s ORCID iD (e.g. 0000-0002-1825-0097) stays the same identifier across every institution, funder, and publisher they interact with over a career, unlike an institutional email address or a publisher-assigned author account.
A university registered in ROR can be referenced consistently across DataCite, Crossref, and funder metadata even if the institution’s name changes through a merger.

Counter-examples

A relative link on a personal or departmental webpage is not a PID — it identifies a location, not an entity, and has no registration authority committed to keeping it resolvable.
An internal database primary key (an autoincrementing row ID in a CRIS or repository) is not a PID on its own; it only becomes part of the PID ecosystem once a real PID is minted and cross-referenced to it.
A publisher’s own proprietary article ID, valid only within that publisher’s platform, is not a PID in the interoperable sense — it isn’t resolvable by external systems or governed by a shared scheme.

Related terms

See Researcher Identifier for the person-specific PID landscape, Persistent data identifier for the dataset-specific case, and PID graph for how individual PIDs are linked together into a queryable network of research-entity relationships. For the operational mechanics of getting a PID, see PID minting, PID metadata, and PID provider.

Use in your systems

JATS XML <role> element

xml

<role vocab="credit"
      vocab-identifier="https://casrai.org/dictionary/"
      vocab-term="Persistent Identifier (PID)"
      vocab-term-identifier="https://casrai.org/dictionary/term/persistent-identifier-pid" />

Schema.org DefinedTerm (JSON-LD)

json

{
  "@context": "https://schema.org",
  "@type": "DefinedTerm",
  "@id": "https://casrai.org/dictionary/term/persistent-identifier-pid",
  "name": "Persistent Identifier (PID)",
  "identifier": "https://casrai.org/dictionary/term/persistent-identifier-pid",
  "description": "A unique, actionable, long-term identifier assigned to a single entity in the research ecosystem -- a work, a person, an organisation, a piece of equipment, a project, or a dataset -- that is registered with a scheme-specific provider, stays associated with that entity even as its location or attributes change, resolves (typically over HTTP) to current metadata or the resource itself, and is designed to remain valid for decades, well beyond the lifetime of any one repository, publisher, or institutional system.",
  "inDefinedTermSet": "https://casrai.org/dictionary/domain/persistent-identifiers#set",
  "url": "https://casrai.org/dictionary/term/persistent-identifier-pid",
  "sameAs": [],
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "publisher": {
    "@id": "https://casrai.org/#organization"
  },
  "dateModified": "2026-07-17T22:23:38",
  "inLanguage": "en"
}