How to find the DOI of a PDF
A DOI (Digital Object Identifier) is the most reliable way to cite an article, and reference managers can build a complete citation from it. Most journal articles published in the last two decades have one, but it is not always obvious where it is in the PDF. Work through these steps in order.
1. Look where publishers print it
- The first page, in the header or footer, often next to the journal name, volume and pages, or in the "Cite this article" box.
- Near the copyright line or the received/accepted dates.
- As a link:
https://doi.org/10.xxxx/…or the olderdx.doi.orgform, or written asdoi:10.xxxx/…. - Sometimes only on the last page or in the running header of every page.
A DOI always starts with 10., followed by a number, a slash, and a suffix. If you see something like 10.1038/s41586-020-2649-2, that is it.
2. Let a tool scan the PDF
The in-browser extractor searches the first pages of the PDF for a DOI pattern, and also reports arXiv IDs, PubMed IDs and ISBNs. The file is processed on your own device and is not uploaded. Reference managers do something similar: Zotero's "Retrieve Metadata for PDF" looks for a DOI or ISBN in the file and fetches the full record.
3. Check the PDF properties
Some publishers embed the DOI in the PDF's metadata rather than (or as well as) the visible text. In a PDF reader, open File → Properties (or Document Properties) and look at the subject, keywords or the advanced/XMP metadata. Be careful: these fields are sometimes empty or copied from a template.
4. Look it up from the title
- Crossref search (search.crossref.org): paste the title, or the whole reference, and pick the matching result. Crossref registers DOIs for most journal articles, books, chapters and conference papers.
- Crossref Simple Text Query: paste a whole reference list and get DOIs back for the entries it can match.
- DataCite search (commons.datacite.org): for datasets, software, theses in some repositories, and preprints (including arXiv).
- PubMed: for biomedical articles, the record shows the DOI next to the PMID.
- The publisher's article page almost always shows the DOI near the title.
Match on more than the title: check authors, year and journal, since different articles can share very similar titles, and corrections or errata have their own DOIs.
5. Verify it
Open https://doi.org/ followed by the DOI. It should land on the article you have. When copying a DOI from a PDF, watch for line breaks inside it, trailing full stops or brackets that are not part of it, and ligatures or odd characters introduced by text extraction.
When there is no DOI
- arXiv preprints have an arXiv ID and an automatically assigned arXiv DOI (
10.48550/arXiv.…). See arXiv identifiers explained. - Books are identified by an ISBN, usually on the copyright page; individual chapters may still have DOIs.
- Older articles, theses, reports and grey literature often have no DOI at all. Cite them with full bibliographic details and a stable URL (a repository handle, for example) instead. Do not invent or guess a DOI from a publisher URL.