Guides
Bullion authentication guide: evidence, tests, and limits
Bullion authentication guide: how to build a reference, measure with stated uncertainty, layer non-destructive tests, and escalate what stays unresolved.
What to take away
- Authentication is a conclusion from converging evidence, not one gadget reading.
- Fix the exact product, year, variant, and official specification before you judge any reading.
- Weigh and measure first, and write the tolerance next to every number.
- Match each test to the fake or alteration it can actually reveal.
- Quarantine anything unresolved. Do not clean, open, or improve it while a claim is live.
A reading means nothing until you name the item and its expected value. A one-ounce Gold Eagle is not a one-ounce fine-gold coin: different alloy, different gross weight, different diagnostics. Proof, bullion, and redesigned strikes of the same year can differ again.
The U.S. Mint's report on 2021 American Eagle design and security changes describes refreshed designs and a visible variable-reed edge on the one-ounce gold and silver coins. That feature belongs to those versions. Do not project it backward onto every Eagle.
Build the reference first
Reference data to record
- Issuer, product, metal, denomination
- Year, size, finish, mint mark
- Privy mark
- Composition, gross weight, fine content, diameter
- Published thickness, edge, inscriptions, security features
- Sourceissuer's page for exact version
- Date you pulled the data
Photograph a known genuine example in clear light if you can do it lawfully. Comparing against an image beats comparing against memory.
Examine provenance and package
Ask who sold the item, when, under what invoice description, and through whose hands it passed. On a bar, check that the serial on the metal matches the card. On a certified coin, check that the record matches both coin and holder. Look at the seals for cuts, heat marks, fresh adhesive, or replacement.
Packaging supports a conclusion. It cannot prove the metal. A genuine card can be opened, copied, and paired with a different object.
Inspect before measuring
Work under diffuse light at low magnification. Record what you see:
- relief
- lettering
- digits
- edge
- texture
- color
- luster
- tool marks Note casting seams, pits, repeated depressions, and plating wear. Also note mismatched fonts and the orientation of features relative to each other. Do not clean the item to make inspection easier.
Measure with uncertainty
Use a calibrated scale matched to the resolution you need. Zero it, and weigh the empty holder separately if the product stays packaged. Use clean caliper jaws and do not scrape a soft surface. Record the instrument, its resolution, the calibration check, and the method. Record the temperature where it matters and repeated readings.
NIST sets out the principle in its guidance on uncertainty of measurement results: a result is complete only with a quantitative statement of its uncertainty. That statement is what lets someone else judge whether the number suits its purpose and agrees with other results.
A weight quoted without a tolerance cannot be compared to a published specification, because nothing states how far apart the two are allowed to be.
A mismatch can reject an item. A match proves nothing, because another material or a hollow body can be built to match one property.
Layer physical tests
Pick non-destructive methods for the product and its package. Then read each result against what that method can and cannot show.
What it can show
- Magnet
- Ferromagnetic substitutions
- Conductivity or eddy current
- Electrical response at stated settings
- Ultrasonic
- Thickness or internal interfaces when geometry allows
- Specific gravity
- Mass against displaced volume
- XRF
- Near-surface elemental composition
Where it fails
- Magnet
- Misses most non-magnetic fakes
- Conductivity or eddy current
- Coatings and contact geometry shift the reading
- Ultrasonic
- Poor contact, odd shapes
- Specific gravity
- Air, surface tension, cavities, attachments
- XRF
- Says little about the interior
Several factors move the result:
- Instrument model
- Calibration standard
- Contact geometry
- Coatings
- Packaging
- Operator skill Keep the raw output, not a pass or fail mark.
Keep grading separate
Authentication asks whether the object is what it claims to be. Grading describes condition or production quality under one service's standard. A genuine low-grade coin is authentic. A counterfeit in a convincing holder does not become genuine because of the number printed on the label.
Escalate uncertainty
Stop transfers and sales. Isolate the item with its packaging and paperwork. Ask the seller for a remedy in writing while the deadlines in their terms are still open. For higher value or conflicting results, go to an experienced independent dealer, an assay laboratory, a mint-authorized system, or an authentication service that handles that product class.
Common questions
Does correct weight prove a coin is genuine?
No. Weight is one consistency check. It only means something alongside dimensions, design, material, and provenance, and a hollow or alloyed fake can be built to hit a published weight.
Should a sealed assay card be opened for testing?
Only after you weigh the test against certification, resale, return, and warranty consequences. Opening a card can end a return right and change what a future buyer will pay.
Is XRF a complete assay of a bar?
No. A surface reading does not necessarily establish internal construction or full-bulk composition. Treat it as one layer of evidence, not a verdict.
What should happen after two tests disagree?
Quarantine the item, verify the methods and the reference data, repeat under controlled conditions, and escalate to a qualified examiner. Do not sell or transfer it while the conflict stands.
