Guides
Part of Bullion authentication guide: evidence, tests, and limits
How to document a non-destructive bullion examination
A non-destructive bullion examination record must show who tested it, with what equipment, and where the method stopped short of full assay.
What to take away
- Open a neutral case ID before you unwrap, weigh or move the piece.
- Photograph the item, its packaging and every test position, and keep the original files.
- Log the instrument, its settings and the reference check that ran before the unknown.
- Write "consistent," "inconsistent" or "inconclusive" per check, never one verdict for the whole item.
- Put a number on how closely the instrument could have known the answer.
A record built this way survives a second examiner. It does not turn a bench observation into an authentication opinion.
Open the case before you touch the metal
Create a neutral case ID. Note the date, the location, who holds the item and why it is being examined.
Case Opening Checklist
- Neutral case ID created
- Date, location, holder, reason noted
- Seller, invoice, advertised claim recorded
- Packaging condition on arrival
- Return window or dispute deadline
- Statusidentity under review
Add the seller, the invoice, the product claim as it was advertised, and the condition of the packaging on arrival. Note whether a return window or a dispute deadline is running, because a slow examination can outlast a refund right.
Do not open with "counterfeit." Write "identity under review" and let the checks move the status.
Photograph the received state
Shoot on a clean neutral background with diffuse light, a stable camera and a scale marker in frame. Cover the obverse and reverse, all edges and corners, and the package seams, labels and serials.
Photograph inscriptions, dates and any area that looks wrong. Add a frame of the instrument setup and the exact spot on the item where it made contact.
Keep the original files with their metadata. Cropped copies are fine for comparison; the originals stay untouched.
Then build a photograph index. Each entry carries an image number, the side shown, the lighting used and the magnification. Mark areas of interest on a duplicate, never on the original file, and include one wide frame that shows where a close detail sits on the whole object.
For a bar sealed in a card, shoot the same feature through both faces. Reflections and plastic texture can read as surface marks. If a label carries personal data, keep one unredacted evidence copy and a separate redacted working copy.
Pin down what the item is supposed to be
Record the issuer page, the product year, the size, the finish and the security-feature generation. Record what your chosen method can actually see.
A study of non-destructive analysis of a gilded object makes the point plainly: cutting a cross-section is the direct way to establish how an object is layered, and that step destroys the item. For a valuable piece it is rarely acceptable.
The same study notes that X-ray fluorescence reads surface composition. It reaches layer thickness only under favorable conditions, and a few spots on an inhomogeneous object can mislead. Put those limits in the record next to the issuer's figures. Those figures describe one product, not gold coins in general.
Log the instruments
For each device, record the manufacturer, model and serial number, plus the resolution and the last calibration or verification date. Add the reference standard used, the mode and energy or frequency setting, the contact geometry, the software version and the operator's name.
Run a check on a known reference and save it before the unknown item goes near the device. A screen reading with no method behind it is weak evidence.
Run the low-risk checks first
Work in an order that cannot damage the piece:
- Visual and package examination.
- Gross weight on a calibrated scale.
- Diameter, thickness and edge.
- Magnet response, without dragging a magnet across the face.
- Conductivity, ultrasonic, density or XRF work by a competent operator.
Stop if a step could scratch, open, heat, contaminate or demagnetize the item or its packaging. A test that changes the evidence has ended the examination.
Record results with their uncertainty
Use a table, not a verdict.
| Check | Expected | Observed | Method | Uncertainty or limit | Status |
|---|---|---|---|---|---|
| Weight | |||||
| Diameter | |||||
| Edge | |||||
| Security mark | |||||
| Conductivity | |||||
| XRF points |
Write "consistent," "inconsistent" or "inconclusive" for each row. Save a final authentication conclusion for the scope and competence you actually have.
NIST's technical note on expressing measurement uncertainty requires quantitative uncertainty statements in NIST measurement results. It adds that a result's usefulness depends largely on the quality of that statement.
Apply the same standard here. A bare number in the Observed column is not evidence until the next column says how closely the instrument could have known it.
Preserve the item and escalate
Return the piece to a clean, labeled, non-reactive holder along with its packaging components. Log every transfer of custody.
Send the next examiner the references and the raw data, not a rewritten summary. Keep the seller and payment deadlines running while the file moves.
Common questions
Why photograph the test position?
Surface composition and geometry vary across a single piece, so a second examiner needs to know exactly where a reading came from. Without that frame, the number cannot be repeated.
Can a phone camera document microtext?
Only when it resolves the feature clearly. Record the optics used and keep a wider context image beside the close-up, so the detail can be located on the item.
Should raw instrument files be kept?
Yes. They hold the settings, spectra and timestamps that a screenshot drops. The raw file is what lets someone else re-run the analysis.
What wording is safest before a conclusion?
Describe what you observed and whether each check was consistent with the expected item. Claim nothing beyond what the instruments and the record support.
