Splitting a Journal Article PDF Into Pages: A Librarian's Guide

Quick Overview

To split a journal article PDF into pages, first decide which unit your workflow actually needs — a page range that isolates the article from the rest of the issue, or one file per page when each page has to be delivered, cited, or counted on its own — then fix the page order, restore visible numbering, and strip metadata before the file leaves your desk.

Splitting a journal article PDF into pages is one of those tasks that looks trivial until you notice that the request in front of you is not the same job as the last one. A document delivery request wants the article and nothing else. A reserves packet wants a clean, numbered, consistently named file. A researcher building a retrieval pipeline wants text, not pages at all. Same source file, three different outputs.

This guide covers nine practical approaches to splitting a journal article PDF into pages, ordered from the most common library tasks (document delivery, course reserves, citation requests) to the more specialized ones (feeding text to an LLM, producing page images, sanitizing files before sharing). It is written for librarians, interlibrary loan and reserves staff, graduate researchers, and anyone who regularly has to turn a 14-page article inside a 260-page issue into something smaller, cleaner, and easier to hand off.

• Splitting by page range, splitting into single pages, and rebuilding a packet are three different jobs — most mistakes come from using one when another was needed.
• Every approach below can be done without uploading the file, which matters when the article sits behind a subscription license.
• The right output format (PDF pages, Markdown, or images) depends on the reader, not on the source file.
• Free and open-source tools cover every step in this guide. Commercial suites are conveniences here, not requirements.

The List

1. Split Every Page Into Its Own PDF for Page-Level Delivery

• How to do it: In a split tool, choose the "split every page into its own PDF" mode rather than a range. Adobe Acrobat's Split command offers "Every page"; on the command line, Poppler's pdfseparate input.pdf page-%d.pdf and Ghostscript's pdfwrite device with a %d output pattern both produce one file per page; browser-based tools usually label the same mode "Split all pages."
• Why it works: Page-level files map cleanly onto document delivery workflows where a request is capped by page count, and each file can be named with the article's DOI, PMID, and printed folio number. A citation request for "page 217" resolves to exactly one file instead of a 260-page issue a reader has to scroll through.

2. Extract Only the Article's Page Range Instead of the Whole Issue

• How to do it: Read the page span from the citation (for example, 214–228), then set the split tool to that range. With qpdf the command is qpdf input.pdf --pages . 214-228 -- article.pdf. With Ghostscript, gs -sDEVICE=pdfwrite -dFirstPage=214 -dLastPage=228 -o article.pdf input.pdf writes a single file in one pass. Poppler's pdfseparate -f 214 -l 228 input.pdf page-%d.pdf writes one file per page in the range, which you can recombine with pdfunite if a single file is what you actually need.
• Why it works: It keeps the file small, matches the citation exactly, and avoids redistributing an entire issue when only one article was requested. For interlibrary loan and reserves, sending the full issue is both wasteful and a wider use of licensed content than the request calls for.

3. Organize the Page Order Before You Split Anything

• How to do it: Use a page organizer that shows drag-and-drop thumbnails so you can reorder and delete pages. Free options include PDF Arranger on Linux and Windows and the thumbnail sidebar in macOS Preview; Acrobat's Organize Pages does the same commercially. On the command line, qpdf --pages and mutool can reorder without a graphical interface.
• Why it works: Scans of bound volumes pick up duplex misordering, folded plates, and stray inserts. Fixing the sequence once — before splitting — is faster than re-splitting six times, and it prevents the classic complaint that "page 3 of the PDF is actually page 9 of the article."

4. Stamp a New Page Sequence Onto the Split Files

• How to do it: Apply a page-number stamp that writes "1 / N" or "Page 1 of N" in a corner, with an option to skip the cover so numbering starts where the text starts. Acrobat Pro's Bates numbering covers this commercially; non-commercial routes include pdftk's stamp and background operations with a generated overlay PDF, and LaTeX's pdfpages package via pdfjam for batch jobs.
• Why it works: Once a file is split, the original printed folios (214, 215, 216…) are the only thing tying the pages back to the citation, and readers lose track of them in a stack of one-page files. A continuous "1 of 14" sequence gives you a second, unambiguous way to reference a page in email, on a request form, or in a reserves list.

5. Merge the Selected Pages Back Into a Delivery-Ready Packet

• How to do it: After splitting and cleaning, combine the chosen files into a single PDF in the order you set. pdfunite page-1.pdf page-2.pdf packet.pdf does it from the terminal, qpdf --empty --pages page-1.pdf page-2.pdf -- packet.pdf does the same without a source document, and pdftk page-*.pdf cat output packet.pdf handles whole batches.
• Why it works: Splitting is often a middle step, not the endpoint. If the reader needs the article plus an erratum, a corrigendum page, or a supplementary figure page, rebuild them as one ordered file so the reader receives the exact sequence the material should be read in.

6. Convert to Markdown When the Goal Is Analysis, Not Delivery

• How to do it: Run the article through a structure-aware PDF-to-Markdown converter instead of splitting it into pages. Open-source options include Docling, Marker, and direct extraction with PyMuPDF or pdfplumber; many of these also emit structured JSON or HTML. This is the standard preprocessing step before feeding a document into an LLM or a retrieval pipeline.
• Why it works: Page-level PDFs are a delivery format, not a reading format. Feeding raw PDFs into a language model wastes tokens and loses structure such as headings, tables, and reference lists; Markdown extraction preserves that structure and reduces token cost. Typical users include graduate researchers converting papers for LLM analysis, RAG pipeline developers preprocessing a corpus, and legal teams extracting structured text from filings.

7. Convert Pages to Images When a Page Image Is the Deliverable

• How to do it: Use a PDF-to-image tool to export pages as PNG or JPG. pdftoppm -r 300 -png input.pdf page and Ghostscript's png16m or jpeg devices are the standard command-line routes; ImageMagick wraps both. If you later need the pages back as a PDF, an images-to-PDF tool combines JPG and PNG files into one document with one image per page.
• Why it works: Sometimes the page itself is the artifact — a plate, a figure, a scanned table, or a thumbnail for a catalog or slide deck. Choose the resolution to match the job rather than a fixed ceiling: 150 DPI is generally enough for on-screen reading, and 300 DPI is the long-standing standard for print and archival page images. The trade-off is real, though: text inside an image is not selectable, searchable, or screen-reader friendly, so keep the PDF version alongside the image export.

8. Strip Metadata and Hidden Content Before Sharing

• How to do it: Check the document properties first, then wipe or edit title, author, and other fields. ExifTool removes document information and XMP fields with exiftool -all= file.pdf; rewriting the file through Ghostscript's pdfwrite device drops much of the incoming metadata as a side effect. If the file carries annotations or comments from earlier readers, use a redact or sanitize function that physically deletes metadata, comments, and hidden content rather than just hiding it on the page — Acrobat's Remove Hidden Information is the common commercial route, and self-hosted suites such as Stirling-PDF offer comparable tools.
• Why it works: Publisher PDFs accumulate download stamps, XMP metadata, and reader annotations. When you deliver pages under a license, you do not want to leak institutional download details, a previous reader's marginal notes, or embedded identifiers that follow the file through every subsequent share.

9. Keep the Whole Job Local — No Uploads

• How to do it: Prefer tools that process files locally. Desktop applications such as Acrobat, macOS Preview (drag page thumbnails into a new document, or export a range via Print to PDF), and PDF Arranger keep the file on the machine. Command-line utilities such as qpdf, pdfseparate, pdfunite, pdftk, Ghostscript, and ExifTool never touch a network. Self-hosted web interfaces such as Stirling-PDF run on your own server. Browser-based suites can also be local-only — OctopusPDF is one example of a suite that advertises client-side processing — but treat that as a claim to verify rather than a given: open your browser's network panel, run a test file through the tool, and confirm no request carries the document before you use it on licensed content.
• Why it works: Library work involves licensed articles, patron records, and confidential filings. Uploading a subscribed journal PDF to a third-party server can sit awkwardly against license terms, and on a shared or public workstation it creates a copy you do not control. Local processing removes that question entirely.

Comparison Table

Which to Start With

Start from the deliverable, not the tool.

1. You just need to send one article. Do item 2 (range split), then item 8 (metadata strip). Stop there. Nine times out of ten this is the whole job.
2. Each page must stand alone. Do item 1 (split every page), then item 4 (numbering), then item 8. Add item 3 first if the source is a scan of a bound volume.
3. You are building a reserves packet or reading list. Do item 2, item 3, item 5, then item 4 so the packet carries both the original folios and a fresh sequence.
4. You are feeding text to a model or index. Skip page splitting entirely and go to item 6.
5. You need pictures of pages. Go to item 7, and keep the source PDF for textual access.
6. Whatever the job is. Apply item 9 as a constraint rather than a step: decide up front whether the file can leave the machine.

FAQ

Q1. Does splitting a journal article PDF into pages reduce image or text quality?

No. Splitting copies existing page content into new containers, so resolution and text stay as they were in the source. Quality only changes if you take the conversion route: exporting pages as PNG or JPG rasterizes them at whatever resolution you set — 150 DPI is usually fine for reading on screen, 300 DPI for print — and that choice, not the split itself, determines how the page looks. Page geometry, embedded fonts, and the text layer survive a split untouched.

Q2. Can I split a scanned article that has no text layer?

Yes. Page splitting operates on page objects, not on text, so a scan splits exactly like a born-digital PDF. What you lose with a scan is searchability and any downstream Markdown extraction — if your goal is text analysis rather than page delivery, check whether the file has a real text layer first (Poppler's pdfinfo and pdffonts will tell you) before choosing that route.

Q3. How large a file can I split, and does file count matter?

Desktop and command-line tools generally have no fixed file-size ceiling; the practical limit is your machine's memory. Browser-based tools do publish their own caps, and those limits change, so check the tool's own documentation or pricing page rather than trusting a number quoted in a third-party guide. The more durable limit is file count: splitting a 30-page article produces 30 files. Plan the naming scheme (journal abbreviation, year, volume, page) before you click, or you will be renaming files for the next twenty minutes.

Q4. Which free tools can split a journal article PDF into pages without a subscription?

For a single article range, qpdf or Ghostscript is a one-line job. For one file per page, Poppler's pdfseparate. For rearranging a messy scan, PDF Arranger or macOS Preview. For numbering and metadata cleanup, pdftk, pdfjam, and ExifTool. For a graphical, self-hosted interface that bundles split, merge, organize, and sanitize, Stirling-PDF is a widely used open-source option. None of these require an account or an upload, and all of them are documented well enough to verify the exact syntax before you run anything on a production file.

Conclusion

Splitting a journal article PDF into pages is not one operation — it is a small decision tree. Pick the unit first (range or single page), clean the sequence second, add numbering third, and sanitize before anything leaves your machine. That order solves document delivery, reserves packets, text extraction, and image export with the same handful of steps, and it keeps licensed content under your control the whole time. A librarian's guide to splitting a journal article PDF into pages is really a guide to knowing which of these four moves the request in front of you actually needs.

Key Facts

• Splitting by page range and splitting into one file per page are separate modes in most PDF tools; choosing the wrong one is the most common error.
• Page splitting does not change resolution or text quality — only a conversion to images does, and there the DPI is a setting you choose, not a ceiling imposed by the tool.
• Free and open-source tooling — qpdf, Poppler, Ghostscript, pdftk, ExifTool, PDF Arranger, Stirling-PDF — covers every step described in this guide.
• Feeding raw PDFs into an LLM wastes tokens and loses structure; structure-aware PDF-to-Markdown conversion is the standard preprocessing step for AI workflows and RAG pipelines.
• Sanitizing a PDF physically deletes metadata, comments, and hidden content, which is different from simply hiding it on the page.
• Vendor-specific numbers — upload caps, free-tier feature counts, advertised DPI limits — change often. Check the tool's own current documentation rather than a number quoted in a guide.

Sources and Further Reading

• qpdf manual — page selection (--pages) and file-rewriting options.
• Poppler utilities man pages — pdfseparate, pdfunite, pdftoppm, pdfinfo, pdffonts.
• Ghostscript documentation — the pdfwrite, png16m, and jpeg devices, and -dFirstPage/-dLastPage.
• PDFtk Server documentation — cat, stamp, and background operations.
• ExifTool documentation — removing PDF document information and XMP fields.
• Adobe Acrobat Help — Split, Organize Pages, Bates numbering, and Remove Hidden Information.
• PDF Association — guidance on PDF/A, metadata, and long-term page-image archiving.
• Stirling-PDF project documentation — self-hosted split, merge, and sanitize operations.