Not the document format. In materials characterization, “the PDF” refers to the Powder Diffraction File, a curated reference database of powder X-ray diffraction (XRD) patterns maintained by the International Centre for Diffraction Data (ICDD), a not-for-profit scientific organization. It has nothing to do with the Adobe document format that shares the same three-letter acronym. This guide covers what the PDF contains, how phase identification actually works, what the entry quality marks mean, how the current product line is scoped, where it fits alongside free and complementary databases, and the practical pitfalls that trip up powder XRD analysis even when the software returns a confident-looking match.
What the Powder Diffraction File Is
The PDF traces back to 1941, when Dow Chemical Company allowed X-ray powder patterns first compiled by Hanawalt, Rinn, and Frevel to be published as searchable reference cards. What began as roughly a thousand printed 3-by-5-inch cards has grown, under ICDD’s ongoing editorial process, into a database of more than one million entries covering inorganic materials, minerals, organics (including pharmaceuticals, dyes, and pigments), metals and alloys, and other crystalline and partially crystalline materials. Each entry packages a reference diffraction pattern — peak positions (as d-spacings, the spacing between crystal lattice planes) and relative intensities — together with crystallographic, chemical, and bibliographic metadata for a specific, identified phase.
What distinguishes the PDF from most other crystallographic databases is that ICDD is the entity that generates and edits reference-quality powder patterns as a matter of institutional mission, not simply an aggregator of author-submitted structures. Every entry carries an editorially assigned quality mark describing how reliable that particular pattern is, which matters enormously once you’re using the database to make an identification call rather than just browsing.
How Powder XRD Phase Identification Actually Works
Phase identification (often shortened to “phase ID”) is the process of determining which crystalline compound or compounds are present in a sample by comparing its measured diffraction pattern against reference patterns of known materials. The basic workflow:
- Measure a powder pattern. A powder or polycrystalline sample is scanned across a range of diffraction angles (2θ) with a fixed X-ray wavelength (commonly Cu Kα), producing a plot of intensity versus angle. Each crystalline phase present produces its own characteristic set of peaks.
- Extract d-spacings and relative intensities. Peak positions are converted from 2θ to d-spacing using Bragg’s law, and peak heights or integrated areas are normalized to relative intensities. This reduced peak list is the searchable fingerprint of the sample.
- Search-match against reference entries. Search-match software compares the sample’s peak list against reference patterns in the PDF (or another reference database), looking for entries whose d-spacings and relative intensities line up within tolerance.
- Rank candidates by a figure of merit. Because many entries can partially match a given pattern, especially in complex or multiphase samples, matching algorithms compute a figure of merit — a numeric score reflecting how well peak positions, intensities, and (in some algorithms) how much of the observed pattern is accounted for — to rank candidate phases. The highest-scoring candidate is a starting hypothesis, not an automatic answer; see practical pitfalls below.
Entry Quality Marks: Why a Star Pattern Should Be Preferred
Because the PDF pools reference patterns submitted and measured under widely varying conditions over more than eight decades, not every entry is equally trustworthy for identification work. ICDD editorially reviews submissions and assigns a quality designation to each entry — this is a genuinely distinguishing feature of the PDF relative to less-curated crystallographic collections, where every record is typically treated as equally authoritative. Historically, and consistent with ICDD’s own published descriptions of its editorial process, entries fall into categories along these lines: Star (high-quality) patterns, which have passed the most rigorous checks on experimental accuracy and are the preferred reference for quantitative or high-confidence identification work; indexed patterns, which have an assigned unit cell but did not meet the full experimental bar for a star rating; calculated patterns, generated computationally from a known crystal structure rather than measured experimentally; low-precision (or “poor quality”) patterns, retained for reference but flagged as less reliable; and deleted entries, kept in the database for historical traceability but superseded or withdrawn from active use. When more than one entry is available for the same phase, an analyst should default to the highest-quality mark available — a Star entry over an indexed or low-precision one — because the figure-of-merit calculation and the resulting confidence in a match are only as good as the reference pattern it’s compared against. Always confirm the current, exact quality-mark terminology and criteria against ICDD’s own documentation for the release you’re using, since ICDD periodically refines the review process and category definitions.
PDF-2, PDF-5+, and the Current Product Line
ICDD sells the Powder Diffraction File as a subscription database, not a free public resource — this is a real institutional-procurement decision for a lab or core facility, not a one-time purchase, and it’s worth budgeting for as a recurring line item alongside instrument service contracts. The product line has consolidated in recent years: PDF-2 remains a standalone, inorganics-focused product (also including common organic materials) sold on a multi-year license and containing several hundred thousand entries. PDF-4+ and PDF-4/Organics were previously sold as separate inorganic- and organic-focused comprehensive products; ICDD has since consolidated their content and features into PDF-5+, the current comprehensive release, which now covers both inorganic materials (cements, metals and alloys, batteries, minerals, solid-state devices) and organic materials (pharmaceuticals, dyes, pigments, polymers) in one database of more than one million entries, and supports laboratory, synchrotron, electron, and neutron diffraction data, not only laboratory X-ray. A specialized PDF-4/Minerals product also remains available for mineralogy-focused work. Before procuring, confirm current entry counts, license terms, and which legacy product names still apply directly with ICDD or your reseller, since release-year figures and the exact product lineup are updated regularly.
Free and Complementary Alternatives
The PDF is not the only reference source for phase identification, and in practice many labs use it alongside — or, budget permitting, instead of — other databases with different coverage and access models:
- Crystallography Open Database (COD) — free and open-access, community-curated, covering small-molecule organic, metal–organic, and inorganic structures with small to medium unit cells. COD entries are full crystal structures (in CIF format) rather than purpose-built powder search-match patterns, and its curation model (automated checks plus manual review of flagged entries) is lighter-touch than ICDD’s editorial quality-mark system, but it costs nothing to access and is a legitimate starting point, especially for structures also published in the open literature.
- Inorganic Crystal Structure Database (ICSD) — a subscription database jointly operated by FIZ Karlsruhe and (in the U.S.) NIST, specializing in fully inorganic crystal structures. Like the PDF, it is a licensed, curated product rather than a free resource, and many labs that already subscribe to ICSD for structural work use it as a cross-check against PDF phase identification results.
- Cambridge Structural Database (CSD) — maintained by the Cambridge Crystallographic Data Centre, the standard reference for small-molecule organic and metal–organic crystal structures, widely used in pharmaceutical and organic-materials phase work. It is also a subscription product with its own curation pipeline, and overlaps with PDF-5+’s organic coverage without being identical to it.
The trade-off across all of these is coverage versus curation versus cost: the PDF’s quality-mark system and purpose-built search-match tooling are what you’re paying for; COD’s openness comes with a lighter editorial process; ICSD and CSD each specialize more narrowly (inorganic and organic/small-molecule respectively) with their own licensing costs. Many core facilities maintain access to more than one, using the free options first and reserving licensed databases for cases the free coverage doesn’t resolve.
Practical Pitfalls: Why the Top Hit Isn’t Automatically the Right Phase
Search-match software will always return a ranked list of candidates, and it’s tempting to treat the top-ranked hit as the answer. Several real experimental effects can push the correct phase down the list, or make an incorrect phase look deceptively good:
- Preferred orientation. If sample particles have a plate-like or needle-like habit, they can pack with a non-random orientation during sample preparation, systematically distorting the relative intensities of the measured pattern away from the reference pattern’s ideal random-orientation intensities. This alone can demote the correct phase in a figure-of-merit ranking, even though the peak positions still match.
- Sample-height displacement. If the sample surface isn’t exactly at the diffractometer’s focusing circle, all peaks shift systematically in 2θ, which shifts the derived d-spacings and can degrade an otherwise-correct match or, less often, spuriously favor an incorrect one.
- Amorphous content. Phase identification only characterizes crystalline material. A sample that is partially amorphous will show a broad background “hump” that search-match algorithms don’t interpret as a phase at all; a purely crystalline-phase-based identification can silently ignore a substantial fraction of the sample’s actual composition.
- Peak overlap in multiphase samples. Real samples are frequently mixtures, and overlapping peaks from two or more phases can be misattributed to a single phase, or can suppress minor phases whose peaks are hidden under stronger ones from a major phase.
- Chemical plausibility beats numeric ranking. A high figure-of-merit match to a phase that makes no chemical sense for the sample — an element or compound that was never part of the synthesis, precursors, or expected reaction products — should be treated with suspicion even if the numbers look good. Reliable phase ID combines the search-match ranking with independent knowledge of what the sample plausibly contains (synthesis route, elemental analysis, known precursors).
Beyond Identification: Rietveld Refinement and Crystallite Size
Phase identification answers “which crystalline phases are present.” Two related but distinct analyses go further:
Rietveld refinement fits a calculated diffraction pattern, built from known or assumed crystal structures, against the entire measured pattern (not just peak positions) by adjusting structural, instrumental, and phase-fraction parameters until the calculated and observed patterns converge. This full-pattern fitting approach enables quantitative phase analysis (the weight or volume fraction of each phase in a multiphase mixture) and refined lattice parameters, and is a substantially more demanding analysis than search-match identification — it requires reasonably good starting structural models for each phase present and is sensitive to how well those models and the refinement strategy are chosen.
Crystallite size from peak broadening uses the fact that very small crystalline domains produce measurably broadened diffraction peaks. The Scherrer equation relates peak breadth (after correcting for instrumental broadening) to an average crystallite size. Its limits matter: the Scherrer equation estimates coherent-scattering domain size, not necessarily physical grain or particle size, which can differ if particles contain multiple crystalline domains; it becomes unreliable for domains much larger than roughly 100–200 nanometers, where instrumental broadening dominates and size-related broadening becomes too small to resolve reliably; and it assumes broadening is due to size alone, when in real samples microstrain (lattice distortion) also broadens peaks and has to be separated out, typically through more elaborate line-profile analysis, for the size estimate to be trustworthy.
Reporting and Reproducibility
Powder XRD phase-identification results are only as reproducible as the methodological detail reported alongside them — an area CASRAI’s broader interest in FAIR and reproducible research practice applies directly to. At minimum, a reproducible report of a phase identification should include: the specific PDF entry (card) number(s) used for each identified phase, so another researcher can check the exact reference pattern and its quality mark rather than just a chemical name; the instrument, radiation source and wavelength (for example, Cu Kα), and 2θ scan range and step size used to collect the pattern; and, wherever institutional and journal policy allow, deposit of the raw measured pattern (not just the final phase call) in an appropriate data repository, as called for in a project’s data management plan. None of this is optional if a phase-ID result is going to support a published claim: without the entry number, instrument details, and scan parameters, a reader has no way to independently assess whether the identification was well-supported or to reproduce the search-match against the same reference pattern.
Frequently Asked Questions
Is the Powder Diffraction File the same thing as a PDF document?
No. In materials characterization and crystallography, “the PDF” refers specifically to the ICDD’s Powder Diffraction File, a curated database of reference powder X-ray diffraction patterns. It shares an acronym with, but has no other relationship to, the Adobe Portable Document Format.
Is the Powder Diffraction File free to use?
No. Access to the PDF (PDF-2, PDF-5+, or PDF-4/Minerals) is by paid subscription from ICDD, typically licensed on a multi-year (PDF-2) or annual (PDF-5+) basis. Institutions generally procure it through a core facility, department, or library rather than as an individual purchase. Free alternatives with narrower or differently curated coverage, such as the Crystallography Open Database, exist for labs without a PDF subscription.
What’s the difference between PDF-2 and PDF-4+ or PDF-5+?
PDF-2 is a longstanding, inorganics-focused product on a multi-year license. PDF-4+ and PDF-4/Organics were previously separate comprehensive inorganic and organic products; ICDD has consolidated their content into PDF-5+, the current comprehensive release covering both inorganic and organic materials with an annual license. Confirm the current lineup directly with ICDD before procuring, since product names and terms are updated periodically.
Why did the search-match software rank the wrong phase first?
Preferred orientation, sample-height displacement, peak overlap in multiphase samples, and reference-pattern quality can all distort a figure-of-merit ranking. The top-ranked candidate is a starting hypothesis that should be checked against what’s chemically plausible for the sample, not accepted automatically.
Does phase identification tell me how much of each phase is present?
Not on its own. Basic search-match identification tells you which phases are present, not their proportions. Quantifying phase fractions in a mixture requires a full-pattern fitting approach such as Rietveld refinement.
How is the Powder Diffraction File different from the Crystallography Open Database or single-crystal X-ray crystallography databases?
The PDF is purpose-built for powder-pattern search-match phase identification and is a paid ICDD product with editorially assigned quality marks. The Crystallography Open Database is a free, open-access repository of full crystal structures, useful as a structural reference but not purpose-built as a search-match phase-ID tool. Neither should be confused with single-crystal X-ray crystallography, a different experimental technique that determines a complete atomic structure from a single crystal rather than identifying phases present in a polycrystalline powder sample.
