Source-labelled market dataSource, observation time and freshness are shown with every quoteData statusAll pages
GoldObserve

LOT / SAMPLE / CUPELLATION / CONTROL / SETTLEMENT

Fire Assay for Gold

Follow a gold result from lot definition and representative sampling through laboratory cupellation, quality controls, uncertainty and commercial settlement.

THE SHORT ANSWER

Fire assay measures a sample; sampling decides what the result represents

Fire assay by cupellation is a long-established laboratory method for determining gold content. ISO 11426:2021 applies to a material considered homogeneous and states a preferred gold-content range plus method modifications and interferences. The headline result can be precise while the commercial conclusion is wrong if the sample was unrepresentative, mixed lots were combined, chain of custody failed or settlement applied an undisclosed correction.

METHOD BOUNDARY

Separate the observation from the conclusion

QuestionDefensible answerUnsafe shortcut
What it measuresGold content in a prepared test portion under a defined laboratory method, controls and calculation.Calling the observation proof of authenticity.
What it can missWhether the sample fairly represents a heterogeneous lot or whole object, title, provenance, market value and internal integrity outside the sample.Treating an unobserved defect as an impossible defect.
Best roleBulk fineness determination when representative sampling, laboratory competence and decision rules are documented.Replacing identity, provenance, recourse and other tests.

CHAIN OF RESULT

Audit the path before accepting the assay certificate

01Define the lot

Exact included material, exclusions, gross weight and segregation.

02Sample

Where, how much, by whom and why the method represents the lot.

03Secure custody

Seals, identifiers, transfers, witnesses and retained material.

04Prepare

Homogenisation, splitting and contamination controls before analysis.

05Assay

Method version, replicates, reference materials, blanks and corrections.

06Settle

Reported fineness, uncertainty, payable recovery, price basis, fees and dispute process.

LABORATORY EVIDENCE

Precision, proficiency and independence answer different questions

Replicate agreement can show repeatability for prepared test portions; it cannot prove that the source sample represented the lot. Reference materials and controls test method performance; proficiency testing compares a laboratory against an external scheme. Accreditation, where applicable, defines a scope. None of these automatically removes a commercial conflict when the buyer, refiner and assayer are the same party.

Before material is cut or melted, agree who chooses the sample, who may witness, whether an independent umpire sample is retained, which method and rounding rules apply, who pays for a re-assay, and which result controls settlement. This is especially important for mixed jewelry, sweeps and scrap that may not be homogeneous before preparation.

PRIMARY AND STANDARDS SOURCES

Methods are only as strong as their scope and controls

Sources reviewed July 31, 2026. Instrument procedures, safety data, standards and acceptance criteria can change. Use the current method documentation for the exact item, instrument, laboratory and decision.

FREQUENTLY ASKED QUESTIONS

Gold fire-assay questions

What is a fire assay for gold?

It is a laboratory method that separates and measures gold from a prepared sample, commonly using cupellation. ISO 11426 specifies the method for material considered homogeneous.

Is fire assay destructive?

Yes. Representative sampling alters or destroys material, and the laboratory procedure consumes the test portion. Agree on sampling, retained material and settlement consequences first.

Is fire assay 100% accurate?

No analytical result is assumption-free. Sampling, homogeneity, weighing, method performance, interferences, corrections, controls, uncertainty and laboratory competence all matter.

Can a fire assay authenticate a whole gold bar?

It determines gold in the tested sample under the method. Extending that result to a whole bar requires a defensible sampling and homogeneity basis plus separate identity and integrity evidence.