Verifying Gold-Coated Jewelry: When XRF Readings Mislead

Verifying gold-coated jewelry across a retail counter is a different task from assaying scrap in a back room. The customer is watching, the transaction has to move, and the conclusion has to be explainable in plain language to someone who may disagree with it.

This tutorial sets out a sixty-second counter workflow: which three points to test, how to read the result while the customer waits, and how to communicate an unwelcome finding without turning a valuation into an argument. The analytical basis is covered in our plated versus solid karat tutorial; this is the operational version.

The sixty-second workflow

Speed matters at a counter, but speed comes from sequence rather than from rushing any single step. The following runs comfortably inside a minute once practised.

Seconds 0–10: handle and inspect

Before the analyzer is involved, pick the piece up. Three things register immediately:

• Weight relative to size. Gold is dense. A piece that looks substantial and feels light is telling you something before any measurement.
• Wear at contact points. Clasp tongues, ring shank undersides, and chain link contact faces show plating wear as a colour change to white or pink.
• Marks. Look for GP, GF, RGP, HGE, or a karat stamp. Their presence is informative; their absence proves nothing.

Experienced buyers form a hypothesis here and use the analyzer to confirm it rather than to discover it from scratch.

Seconds 10–20: clean and position spot one

Wipe the intended test area with isopropyl alcohol. Skin oil, polish residue, and tarnish all attenuate signal and depress readings — this is a genuine source of false low results, not a formality.

Position the largest available flat area over the aperture, confirming the piece fully covers it. A partially covered aperture lets the beam strike the stage and contaminates the reading.

Seconds 20–35: the three readings

Take readings at three deliberately chosen locations:

1. A flat, protected face — the piece at its most favorable
2. A high-wear edge or contact point — where any coating is thinnest
3. A second body location distant from the first — testing consistency across the piece

The second reading is the one that does the work. Skipping it is the single most common reason plated material passes as solid.

Seconds 35–45: read the result

Three questions, in order:

• Are there elements that do not belong? Iron, high zinc, or high nickel ends the analysis immediately.
• Does the gold percentage match a karat standard? 41.7, 58.5, 75, 91.6 — or something in between that matches nothing.
• Do the three readings agree within two points? Spread beyond that indicates non-uniform material.

Seconds 45–60: conclude and communicate

State the finding and the basis for it in the same breath. What that sounds like is covered below.

Choosing the three spots on common item types

“A worn edge” is only useful guidance if you know where it is on the item in front of you.

Rings

Test the outer face, the inner shank, and the shoulder beside the setting. The inner shank is in constant skin contact and wears fastest, and it is also where the karat mark usually sits.

A ring that reads well on the outside and poorly on the inner shank is plated. A ring that reads consistently on both but differs at the shoulder has probably been repaired or had a stone reset with lower-karat solder.

Chains and necklaces

Test the clasp, a link near the clasp, and a link from the middle of the length. The clasp is a separate manufactured component and is frequently a different material — replaced clasps are extremely common.

For fine chains, coil the chain to fully cover the aperture. A single thin link will not, and the reading will include the stage.

Bracelets and bangles

Test the outer face, the inner surface where it contacts the wrist, and a hinge or clasp component. The inner surface is the wear point.

On rigid bangles, also weigh and estimate volume — hollow construction is common and the analyzer cannot see it.

Earrings

Test the decorative front and the post or hook separately. These are frequently different metals joined together, and a reading on one says nothing about the other.

Findings are often surgical steel or a hypoallergenic base even on genuine gold earrings, so a base-metal reading on a post is not by itself evidence that the front is plated.

Watches

Test the case back, a lug underside, and an outer case edge. Lugs and case backs are protected; outer edges wear. Bracelet links are frequently finished differently from the case and should be treated as separate items.

Reading the result at the counter

Five patterns cover nearly every outcome, and each has a clear next action.

Consistent, standard karat, no base metals

Solid gold. Proceed on the average of the three readings.

Flat face high, worn edge markedly lower

Plated. The coating survives where protected and has thinned where worn. This is the textbook signature and needs no further testing.

Base metals present at any spot

Plated or base metal throughout. Iron, chromium, or zinc above roughly 15 percent settles it.

All three consistent and high, but the piece is light

Heavy plating, gold filled, or hollow construction. XRF has reached its limit and the density check governs. Solid 18K is 15.6 g/cm³; brass is around 8.5.

One spot anomalous, two consistent

Usually a localized feature — solder, a repair, a setting, or contamination. Re-clean and re-measure the anomalous spot before drawing a conclusion.

Terminology customers use, and what it means

Customers describe their jewelry using terms that carry specific legal definitions they may not know. Understanding the vocabulary prevents talking past each other.

Gold plated

An electrodeposited layer, typically 0.05 to 0.5 microns. Detectable by XRF because substrate signal escapes through such a thin coating. Under US FTC guidance, items marked as gold plated should carry a layer of at least 0.5 microns.

Recovery value is effectively nil. The gold present on a plated bracelet amounts to a fraction of a gram.

Gold filled

A mechanically bonded layer defined by weight rather than thickness — commonly 1/20, meaning gold is at least one twentieth of the item’s total weight. Typical layers run 20 to 50 microns, beyond XRF’s measurement depth.

Recovery value is real but modest. A 40 g gold-filled bracelet at 1/20 12K contains around 1 g of fine gold, which is worth calculating rather than dismissing.

Rolled gold plate

Similar bonding to gold filled but a thinner layer, typically marked RGP. Intermediate between plated and filled in both thickness and recovery value.

Vermeil

Gold over sterling silver, requiring at least 2.5 microns of gold over a base of at least 92.5 percent silver under the US definition. Distinct from other plated categories because the substrate itself has real value.

A vermeil piece should be valued on its silver content, which is frequently more than the customer expects. Missing this means underpaying on genuinely valuable material.

Gold tone or gold finish

Marketing language with no legal gold content requirement. Frequently no gold whatsoever — a brass or lacquered base metal with a gold appearance. XRF settles this immediately.

Solid gold

Means the item is gold alloy throughout, not that it is hollow or solid in construction. A hollow 18K bangle is correctly described as solid gold despite containing air, which is a genuine source of customer confusion worth clarifying explicitly.

Fitting verification into store operations

A protocol that only works when the store is quiet is not a protocol. Several practical arrangements determine whether the method survives a busy day.

Position the analyzer within reach of the counter

If verification requires walking to a back room, it will be skipped under pressure. An instrument positioned so the customer can watch and the operator does not leave the counter removes the main reason the step gets bypassed.

Power on before opening

Thermal equilibrium takes around thirty minutes. Switching the instrument on at opening means the first customer of the day receives a drifting reading. Powering on before the doors open, and running a reference standard check while preparing the floor, costs nothing and removes an entire class of error.

Keep a reference standard at the counter

A certified reference of known composition, tested at the start of each session, confirms the instrument is in control. Thirty seconds daily. When a valuation is later disputed, being able to show that the machine verified correctly that morning is a materially stronger position.

Standardize the record fields

A short printed or digital form with fixed fields — three readings with locations, elements, weight, marks, photograph — produces consistent records across staff without requiring anyone to remember what to write down.

Agree the escalation threshold

Decide in advance what value triggers a density check and what triggers refusal pending further testing. Deciding in the moment, with an attractive transaction waiting, reliably produces optimistic judgments.

Rehearse the difficult conversation

The heavy-plated piece that reads as solid gold is the hardest case, because the analyzer appears to support the customer. Staff who have practised that explanation handle it calmly; staff encountering it for the first time in front of an unhappy customer usually do not.

Talking to the customer

The technical finding is half the job. Delivering it badly turns a correct valuation into a lost customer or a public argument.

Test where they can see

Conduct the measurement in the customer’s view and narrate as you go. “I am testing three separate places so we are working from an average rather than one spot, and I am deliberately testing the clasp because that is where any surface coating wears first.”

This does two things: it pre-empts the suspicion that the machine was manipulated, and it establishes that you are following a method rather than forming an impression.

Lead with the reading, not the verdict

Compare two openings. “This is only plated” invites disagreement. “The flat face reads 71 percent gold but the clasp reads 34 percent with copper and zinc — that difference means we are looking at a gold layer over a brass base” invites a question instead.

The second version is verifiable, specific, and leaves the customer somewhere to go other than defending their belief.

Separate the finding from blame

Most people presenting plated jewelry believe it is solid, often because they were told so or inherited it as such. Framing matters: the piece is gold plated, which was extremely common in the period it was made, and it does not mean anyone deliberately misled them.

Where a piece was recently purchased as solid, the finding may genuinely matter to the customer beyond your transaction. Offering the reading in writing gives them something to take back to the seller.

Offer the demonstration

When a customer disputes the result, the strongest response is to test again and let them choose the spot. If the piece is genuinely plated, a customer-chosen worn edge will show it. If it is solid, you have removed the objection at no cost.

Refusing to re-test reads as defensiveness even when the original reading was correct.

Explain the density check when it applies

The hardest conversation is the heavy-plated or gold-filled piece that reads as solid gold at every spot. Here the analyzer supports the customer’s position and the weight contradicts it.

Show the arithmetic. Solid 18K at 15.6 g/cm³ against the item’s measured dimensions gives an expected weight; the actual weight is substantially below it. This is concrete and checkable, and it moves the conversation from opinion to measurement.

What to record

Counter transactions get disputed later, and recollection is a weak position. A workable record fits in a few fields:

• Each reading with its location — “outer face 71.2, clasp 34.1, mid-link 68.9”, not “average 58”
• Elements detected at each spot, since base metals are usually the decisive evidence
• Weight, and the density conclusion where the check was performed
• Marks observed and their location
• The reference standard check from that session
• A photograph of the piece, which resolves identity disputes on returned pawn collateral

Recording the individual readings rather than the average is the detail that matters most. The pattern is the evidence; the average conceals it.

Common counter mistakes

Testing only the easiest surface

The flat outer face is the fastest spot to test and the least diagnostic. An operator under time pressure tests it alone and misses the majority of detectable plating.

Skipping the clean

Ten seconds with isopropyl alcohol removes a systematic downward bias. Operators who skip it periodically underpay on genuine gold and cannot explain why the reading looked low.

Accepting a high reading uncritically

A reading above the marked karat is not good news. Surface enrichment from polishing depletes copper and silver from the top microns, and a genuine 18K piece can read 78 or 79 percent on a polished face. Paying on that figure overpays.

Testing through a coating

Lacquer, enamel, and heavy tarnish all attenuate signal without appearing in the results. The reading comes out low for reasons unrelated to alloy content.

Testing an item smaller than the aperture

Thin chains, small findings, and earring posts frequently fail to cover the measurement window. Part of the beam strikes the stage and the result includes whatever the stage is made of.

Using a cold instrument

Readings taken before the analyzer has reached thermal equilibrium drift. Thirty minutes from power-on before the first valuation of the day, confirmed with a reference standard, is covered in the setup and deployment guide.

Frequently asked questions

How long should a counter verification take?

About sixty seconds once the sequence is practised: ten seconds handling and inspection, ten cleaning and positioning, fifteen for three readings, and the remainder interpreting and communicating. Rushing the three-spot requirement is where errors enter.

Which single spot matters most?

The high-wear contact point — clasp tongue, inner ring shank, or worn edge. Plating is thinnest there, so it reveals substrate that a protected flat face conceals. If only one reading is possible, take it there rather than on the easy surface.

What gold percentage indicates plating?

No single number. The indicators are a percentage matching no karat standard, base metals such as zinc or nickel at significant concentration, and more than two percentage points of variation between spots.

Can I verify without the customer seeing?

You can, but testing in view is better practice. Visible testing pre-empts suspicion of manipulation and makes a disputed result far easier to defend, because the customer watched the method being applied.

What if the customer disputes the reading?

Offer to re-test and let them select the spot. A genuinely plated piece will show substrate at a customer-chosen worn edge, and a genuinely solid piece will confirm itself. Declining to re-test looks defensive regardless of who is right.

Does a karat stamp settle the question?

No. Stamps are worn, transferred between pieces, applied before later repairs, and sometimes counterfeited. Treat a stamp as supporting evidence and the measurement as the basis for price.

How do I handle a piece that reads solid but feels light?

This is the heavy-plated, gold-filled, or hollow case, and XRF cannot resolve it. Weigh the piece, estimate volume from dimensions, and compare against solid density — 15.6 g/cm³ for 18K, 13.1 for 14K. A large shortfall is decisive.

Should I test every item in a mixed lot?

For anything of individual value, yes. For low-value mixed scrap the economics change and lot-level sampling is proportionate, though a spot check on the highest-value pieces remains worthwhile.

The workflow condensed

Handle the piece first and form a hypothesis. Clean the test area. Take three readings — protected face, worn contact point, distant second location. Check for foreign elements, check the percentage against karat standards, check agreement within two points. Cross-check weight against volume on anything substantial.

Then say what you found and why, using the actual numbers rather than a verdict.

The technical work is straightforward once the sequence is habitual. The part that determines whether a customer accepts the outcome is showing them the method rather than announcing the conclusion.

For the full analytical reasoning behind each pattern, see the plated versus solid karat tutorial. For coating thickness constraints, see plating thickness measurement limits.

Three transactions worth walking through

The patterns become concrete against specific cases. Each of the following is an ordinary counter situation rather than an unusual one.

The inherited bracelet

A customer brings a heavy link bracelet described as her grandmother’s gold. No visible marks. It feels reasonably substantial.

Readings: outer face 68.4 percent gold, clasp tongue 22.1 percent with copper 58 and zinc 19, mid-link 66.9 percent. The spread is enormous and the clasp reveals brass unambiguously.

The conversation started with the numbers rather than the verdict: the outer surface is gold, but where the clasp has worn through we are seeing the brass underneath, which means this is gold plated over a brass base. The customer was disappointed but not suspicious, because she watched the clasp reading appear.

Worth noting: the piece still had sentimental value and she kept it. Not every verification ends in a transaction, and handling that outcome well is what brings the customer back with something else.

The vermeil chain nearly undervalued

A fine chain reads 41 percent gold with silver at 54 percent. The initial read at the counter was low-karat gold alloy, which would have priced poorly.

The wear-edge reading changed the picture: at a worn link the gold dropped sharply and silver rose to 93 percent. That shift — proportional composition on an alloy, sharp transition on a coating — identified vermeil.

The correct valuation was on the sterling silver content plus the modest gold layer, which came out meaningfully above the low-karat-alloy estimate. This case runs in the customer’s favour, and getting it right builds more trust than any number of correct rejections.

The watch that read solid

A dress watch case reads 74.6 percent gold at three separate points with no base metals detected. Every indicator says solid 18K.

The weight told a different story: 38 g for a case whose dimensions implied closer to 70 g if solid 18K. Heavy gold plating over stainless steel, with the coating thick enough to exceed XRF’s measurement depth entirely.

This is the case where the analyzer supports the customer and the scale contradicts it. Showing the density arithmetic — dimensions, 15.6 g/cm³ for 18K, the implied weight, the actual weight — moved the discussion onto ground the customer could check for himself.

When to decline rather than discount

Not every uncertain item should be bought at a reduced price. Three situations warrant declining outright.

Readings that contradict each other without explanation

If three spots disagree and re-cleaning and re-testing does not resolve the pattern, the item is not understood. Buying it at a discount prices the uncertainty rather than eliminating it, and the discount is usually guessed.

High value with unresolved construction

Where the item is worth enough that being wrong matters and non-destructive methods have not settled the question, declining is cheaper than a discounted guess. Offering to hold the item pending destructive verification, with the customer’s consent, is a middle path worth having available.

Documentation that does not match the item

A certificate, appraisal, or hallmark that conflicts with measurement is a reason to stop rather than to negotiate. The discrepancy may be innocent, but resolving it is not a counter-side task.

Building competence across staff

Verification quality varies far more between operators than between instruments. Three practices close most of that gap.

Practise on known material

Keep a small set of known pieces — a solid karat item, a plated item, a vermeil item, a gold-filled item — and have new staff test all four. Seeing the four patterns side by side teaches interpretation faster than any written guidance.

Review the readings, not the outcomes

When a valuation is questioned, examine the recorded spot readings rather than debating the conclusion. Almost every error traces to spot selection, and reviewing selection is how the protocol improves.

Make the worn-edge reading non-negotiable

Of all the steps, this is the one that gets dropped under time pressure and the one that catches the most. Treating it as mandatory rather than advisable is the highest-value rule in the entire workflow.