XRF Tutorial: Distinguishing Gold Plating from Solid Karat

The single most expensive mistake at a buying counter is paying solid gold prices for plated material. It happens not because the analyzer fails, but because the analyzer succeeds — it accurately measures the surface it is pointed at, and on a plated item that surface genuinely is gold.

This tutorial covers the reading patterns that distinguish the two, the multi-spot method that exposes plating, and the specific situations where XRF alone cannot settle the question. It assumes you already know how to operate the instrument; the subject here is interpretation.

Why a single reading cannot answer the question

XRF interrogates only the top few tens of microns of any sample. For gold L lines in a gold matrix, meaningful signal originates within roughly 20 to 60 microns of the surface.

A typical decorative plating is 0.05 to 0.5 microns — thin enough that the substrate underneath contributes strongly to the spectrum. That is the detectable case. But heavier coatings change the picture entirely:

• Flash plating, 0.05–0.25 µm: substrate signal dominates, easily detected
• Standard plating, 0.25–1 µm: mixed signal, clearly detectable
• Heavy plating, 2.5–10 µm: substrate signal weak or absent
• Gold filled, 20–50 µm: no substrate signal whatsoever

On the last category, the spectrum is indistinguishable from solid gold. No detector quality, no extended counting time, and no software correction changes that — the information never reaches the instrument. Recognizing which category you are in is therefore the first analytical step, and it is covered quantitatively in our plating thickness limits guide.

The signature of plated material

When the coating is thin enough for substrate signal to escape, three characteristic patterns appear. Recognizing them turns a composition table into a diagnosis.

Base metal elements that do not belong

Solid karat gold is alloyed with a predictable and limited set of elements: copper, silver, zinc in small amounts, nickel or palladium in white gold. The proportions vary by karat and color, but the element list is short.

Plated items sit on base metal, and that base metal announces itself:

• Copper with substantial zinc — a brass substrate, by far the most common in costume jewelry
• Nickel at high concentration — either a barrier layer or a nickel-silver substrate
• Iron or chromium — stainless steel base, common in watch cases and men’s jewelry
• Zinc at high concentration — zinc alloy casting, typical of inexpensive fashion pieces

The distinction that matters is proportion. A trace of zinc in a gold alloy is normal. Zinc at 15 or 20 percent alongside copper at 60 percent is brass, and the gold reading is a coating.

Gold percentages that match no standard

Karat gold comes in defined grades: 10K at 41.7%, 14K at 58.5%, 18K at 75%, 22K at 91.6%, 24K near 99.9%. A genuine piece reads close to one of these figures.

A plated item produces whatever mixture of coating and substrate the geometry happens to yield — 34%, or 47%, or 62%. These are not karat grades. A gold percentage that sits between recognized standards, and matches none of them, is a strong indicator of a coating.

The exception worth noting is worn plating, where readings drift downward over time as the layer thins. A piece reading 22% gold with heavy brass is not a low-karat alloy; it is plating near the end of its life.

Variance between measurement spots

This is the most reliable indicator and the one that requires the least interpretation.

Solid gold is homogeneous. Readings taken at different points on a solid piece agree closely — typically within 1 to 2 percentage points, with the variation attributable to counting statistics and surface condition.

Plating is not uniform. Electrodeposition builds up more heavily on edges, points, and outward-facing surfaces, and more thinly in recesses and on interior faces. Wear then removes material preferentially from contact points. The result is genuine spatial variation in the measured gold percentage.

A spread of more than 2 percentage points between spots on a piece that should be homogeneous is the signal to investigate further.

The three-spot method

The procedure below takes under two minutes and resolves the great majority of cases.

Spot one: a flat, unworn face

Choose the largest flat area available, away from edges, joints, and settings. Clean it with isopropyl alcohol first — polish residue and skin oil both depress readings and produce false variance.

This reading represents the piece at its most favorable. On a plated item it will be the highest of the three.

Spot two: a high-wear edge or contact point

This is the diagnostic reading. Choose the clasp tongue, the underside of a ring shank, a chain link that rubs against its neighbors, or any edge showing visible wear.

Plating is thinnest exactly where wear is greatest. On a plated item this spot frequently reveals substrate that the flat face conceals entirely. Operators who skip this step miss the majority of detectable plating.

Spot three: a second location distant from the first

Test the opposite end of the piece, or the reverse face. This checks homogeneity across the item rather than within one small region.

Applying the decision rule

Compare the three:

• All three within 2 points, no unexpected base metals, gold percentage near a karat standard — consistent with solid karat gold. Proceed on the average.
• Spread exceeds 2 points, or base metals appear at any spot, or the percentage matches no standard — treat as plated or uncertain until resolved.
• All three read as clean high-purity gold but the item is unusually light for its size — this is the heavy-plated or gold-filled case, and XRF alone will not settle it.

Gold filled: the case XRF cannot decide alone

Gold filled is not heavy plating. It is a mechanically bonded layer, and the US standard defines it by weight rather than thickness — commonly 1/20 gold filled, meaning the gold layer constitutes at least 1/20 of the item’s total weight.

In practice this produces layers of 20 to 50 microns, comfortably beyond the depth XRF can see through. Every reading returns solid gold values because every reading is measuring solid gold.

Marks to look for

Gold filled items are usually marked, and the marks are specific:

• GF — gold filled, often with a fraction such as 1/20 12K GF
• RGP — rolled gold plate, a thinner variant
• HGE or HGP — heavy gold electroplate
• GP — gold plated, thinner still

These marks are frequently worn, tiny, or located on a clasp that has been replaced. Their absence proves nothing, but their presence is decisive and should be checked before the analyzer comes out.

Density as the practical test

The reliable non-destructive check is mass against volume. Approximate densities:

• 24K gold: 19.3 g/cm³
• 22K gold: 17.3 g/cm³
• 18K yellow gold: 15.6 g/cm³
• 14K yellow gold: 13.1 g/cm³
• Brass: 8.4–8.7 g/cm³
• Sterling silver: 10.4 g/cm³

A gold-filled item over brass weighs roughly half what solid 18K of the same dimensions would. That gap is large and unambiguous. Water displacement gives volume for irregular shapes; for simple forms, dimensional measurement is adequate.

Substrate identification: reading the base metal

When substrate signal does escape the coating, it identifies what the item is actually made of. That identification is worth more than the gold percentage, because it tells you what you are buying rather than what it is wearing.

Brass

Copper 60 to 70 percent with zinc 30 to 40 percent. By far the most common base for costume and fashion jewelry, and the substrate behind the great majority of gold-plated items at a buying counter. Brass takes plating well and machines easily, which is why it dominates.

Diagnostic marker: zinc above roughly 15 percent. No legitimate karat gold alloy carries zinc at that level.

Nickel silver

Copper with nickel and zinc, containing no silver whatsoever despite the name. Used where a whiter base is wanted under thin gold, or as the body of costume pieces sold as silver-toned.

Diagnostic marker: nickel at 10 to 20 percent alongside copper and zinc. Also a regulatory flag, since nickel release is restricted in the EU for items in prolonged skin contact.

Stainless steel

Iron with chromium, typically 16 to 18 percent, often with nickel. Common in watch cases, men’s bracelets, and modern minimalist designs.

Diagnostic marker: iron at high concentration. Iron has no place in any precious metal alloy, and its presence is unambiguous.

Zinc alloy castings

Zinc-dominant with aluminium and small copper additions, used for inexpensive cast components. Extremely soft and low density.

Diagnostic marker: zinc as the dominant element. These items also feel conspicuously light for their bulk, which is often noticed before the analyzer confirms it.

Sterling silver

Silver 92.5 percent with copper. This is the vermeil case, and the only substrate that is itself a precious metal with real recovery value.

Diagnostic marker: silver at or near 92.5 percent under the gold layer. Worth distinguishing carefully, because a vermeil piece has genuine silver value that a brass-based plated piece does not.

Tungsten

Rare in jewelry but the substrate of concern in bar and coin fraud, chosen because its density of 19.25 g/cm³ is nearly identical to gold’s 19.32. This defeats the density check that catches every other substrate.

Diagnostic marker: tungsten Lα at 8.40 keV, detectable only when tungsten is at or near the surface. Deep cores require the layered methods set out in our tungsten detection tutorial.

Training the reading habit

The difference between an operator who catches plated material and one who does not is rarely equipment. It is a small number of habits applied consistently.

Look at the element list before the gold number

The instinct is to read the gold percentage first. The more informative field is the element list, because an unexpected element settles the question faster than any percentage does. Iron, high zinc, or high nickel ends the analysis immediately.

Check the spectrum, not only the table

The composition table is the software’s interpretation. The spectrum is the measurement. Peak shape anomalies, unexplained small peaks, and unusual background all indicate something the automatic quantification may have smoothed over.

Handle the piece before measuring

Weight relative to apparent size, edge wear patterns, the sound it makes when set down, and the presence of a worn-through area at a contact point are all information available before the analyzer is switched on. Experienced buyers often suspect plating before they measure and use the instrument to confirm rather than to discover.

Treat a surprising result as a question

A reading that does not fit — too high, too low, or matching no standard — is a prompt to test again elsewhere, not a number to record and move past. Most valuation errors survive because nobody asked a second question.

Rhodium, vermeil and the cases that read backwards

Not every coating lowers the reading. Several common treatments push it in the opposite direction, and an operator expecting plating to always depress gold content will misread them.

Rhodium plating over white gold

Almost all commercial white gold carries a rhodium flash, typically 0.05 to 0.2 microns, applied because the underlying alloy has a faintly yellow or grey cast. Rhodium is a platinum-group metal and reads as such.

On an instrument that resolves it, rhodium appears as a distinct element with no legitimate place in a gold alloy — an immediate signal of surface treatment. On a lower-resolution instrument the rhodium L line at 2.70 keV sits unresolved among palladium at 2.84 and silver at 2.98, and the whole region reads as an indistinct rise.

The practical point is that rhodium detection on white gold is expected and normal, not evidence of counterfeiting. It becomes significant only when the underlying alloy fails to match the marked karat once the flash is accounted for.

Vermeil

Vermeil is gold over sterling silver, requiring at least 2.5 microns of gold and a silver base of at least 92.5% under the US definition. It is legitimate and legally defined, not a counterfeit category.

Vermeil is diagnostically awkward because the substrate is itself a precious metal. A reading showing gold with substantial silver could be vermeil, or could be a genuine low-karat gold-silver alloy. Distinguishing them requires the wear-edge test: on vermeil, a worn area shifts sharply toward pure silver, whereas an alloy stays proportionally consistent.

Surface enrichment on genuine gold

This is the inverse trap. Pickling, polishing, and long wear preferentially remove copper and silver from the outermost microns of a genuine karat alloy, leaving a gold-enriched surface layer.

The consequence is that a genuine 18K piece can read 78 or 79 percent on a heavily polished face, while the bulk is 75%. An operator interpreting a high reading as evidence of higher karat will overpay.

The wear-edge reading corrects this too. Enrichment is a surface phenomenon; a worn edge exposes bulk material and reads closer to true composition. This is one of several reasons the same three-spot method serves multiple diagnostic purposes at once.

Interpreting the spread

The pattern of disagreement between spots carries more information than the average does.

High flat face, low worn edge

Classic plating. The coating survives on the protected face and has thinned or worn through at the contact point. The magnitude of the gap indicates roughly how worn the piece is.

Uniformly moderate with strong zinc and copper

Thin plating over brass across the whole piece, or a very low-karat gold alloy. The zinc proportion distinguishes them: gold alloys use zinc sparingly, brass is 30 to 40 percent zinc.

One anomalous spot, two consistent

Usually a localized feature rather than a plating question — a solder joint, a repair, a stone setting, or a contaminated area. Re-clean and re-measure before drawing conclusions.

All spots high and consistent, weight low

Heavy plating, gold filled, or hollow construction. XRF has reached its limit and the density check governs.

All spots consistent, weight consistent, percentage matches a standard

Solid karat gold. Proceed with confidence.

When to escalate beyond XRF

Three situations justify moving to a destructive or semi-destructive method.

High value with unresolved ambiguity

If the item is worth enough that an error exceeds the cost of testing, escalate. Coulometric stripping removes coating from a small defined area and measures the charge required — accurate well beyond XRF’s saturation ceiling and only minimally destructive.

A dispute requiring documentary evidence

Cross-sectional metallography — cut, mount, polish, measure under magnification — is definitive and reveals structure no indirect method can. It is the arbitration standard when a transaction is formally contested.

A file test as a practical middle ground

Filing a small notch in a concealed area and re-testing inside the notch exposes subsurface material directly. It is destructive in a minor, usually invisible way, and it is fast. Many counters use this as the standard escalation before committing to laboratory methods, and it settles the gold-filled question immediately.

Whichever route is chosen, the decision to escalate should follow a pre-agreed threshold rather than an in-the-moment judgment. Deciding with a valuable lot waiting tends to produce optimistic conclusions.

Frequently asked questions

Can XRF tell if gold is plated?

For coatings up to roughly 5 microns, yes — substrate elements appear in the spectrum and readings vary between spots. For gold-filled material at 20 to 50 microns, no. The coating exceeds XRF’s measurement depth and the spectrum matches solid gold.

What gold percentage indicates plating?

There is no single threshold. The stronger indicators are a percentage that matches no karat standard, the presence of base metals such as zinc or nickel at significant concentration, and variation exceeding 2 percentage points between measurement spots.

Why does the reading change when I test a different spot?

Plating thickness genuinely varies across a piece, building up on edges and thinning in recesses and at wear points. Solid gold does not vary this way. The variation itself is the diagnostic signal.

Is zinc always a sign of plating?

No. Small amounts of zinc appear legitimately in some gold alloys and commonly in gold solder. Zinc at 15 percent or above alongside high copper indicates a brass substrate, which is a different matter entirely.

How do I test a very thin chain?

Thin chains often fail to cover the measurement aperture, allowing part of the beam to strike the stage and contaminate the reading. Fold or coil the chain to present continuous metal, and confirm the aperture is fully covered before measuring.

Does rhodium plating mean the item is fake?

No. Rhodium flash is standard finishing on white gold and increasingly common on yellow gold sold as tarnish resistant. It is a legitimate treatment. What matters is whether the underlying alloy matches the marked karat once the coating is accounted for.

Can I trust a hallmark instead of testing?

A hallmark is supporting evidence, not proof. Marks are worn, transferred between items, applied before later repairs, and sometimes counterfeited. Test regardless, and treat a discrepancy between mark and measurement as a subject for discussion rather than an accusation.

What is the fastest reliable check for gold filled?

Weight against volume. Gold filled over brass weighs roughly half what solid 18K of the same dimensions would. Where the density check is inconclusive and value justifies it, a small file notch in a concealed area gives an immediate answer.

The method condensed

Clean the surface. Test a flat face, a worn edge, and a distant second location. Require agreement within 2 percentage points, an element list consistent with a gold alloy, and a percentage near a recognized karat standard. Cross-check weight against volume on anything substantial.

When all readings look like solid gold but the piece is light for its size, XRF has told you everything it can and the density calculation takes over.

The underlying principle is worth restating: the analyzer is not deciding whether an item is solid — it is reporting what the surface contains, and you are deciding. Everything above is a method for making that decision on adequate evidence.

For the retail counter version of this workflow, see our gold-coated jewelry verification tutorial. For the physics behind the depth limitation, see the physics of XRF alloy composition.

Category-specific patterns

Different item types fail in characteristic ways. Knowing the pattern for the category in front of you shortens the diagnosis considerably.

Rings

The inner shank is the highest-wear surface on any ring, in constant contact with skin and frequently thinned by resizing. It is also usually where the karat stamp sits. Testing the inner shank alongside the outer face gives the wear comparison and the mark check in one movement.

Watch for rings that read differently between shank and head. Repairs and stone resettings frequently introduce lower-karat solder around the head, and a piece can be genuine at the band while the setting area is not representative.

Chains and bracelets

Chains are difficult because individual links may be smaller than the measurement aperture. Coiling the chain to present a continuous mass of metal solves this, but it introduces a second issue: the reading then averages across many links, concealing variation between them.

For a chain suspected of plating, test the clasp separately. Clasps wear faster than links and are also the component most often replaced with a different material entirely.

Earrings

Posts and hooks are frequently a different metal from the decorative element — often surgical steel or a hypoallergenic alloy attached to a gold or gold-plated body. A reading taken on the post says nothing about the front.

Test the decorative face and the finding as separate items, because in valuation terms they frequently are.

Watch cases

Gold-plated stainless steel is extremely common, and plating on watch cases is often heavier than on jewelry, sometimes reaching the saturation threshold. The case back, lugs, and bracelet links may each be finished differently.

The lug undersides and the area beneath the bracelet attachment are usually the least-worn surfaces, and the outer case edges the most worn. That contrast is the fastest diagnostic on a watch.

Coins and bars

These deserve a different mindset entirely, because the counterfeiting economics differ. Jewelry plating is usually a cost-saving in legitimate manufacture; bar and coin plating is usually deliberate fraud, and the substrate is chosen to defeat detection.

Density checks that reliably expose brass fail against tungsten. This category requires the layered verification approach rather than XRF plus weight alone.

What a defensible record contains

Plated-versus-solid determinations get challenged more than any other reading, because the price difference is large and the seller frequently disagrees. A record that holds up needs specifics.

• Each individual reading with its location, not just the average. “Outer face 74.8, inner shank 41.2, clasp 38.6” tells the story; “average 51.5” conceals it.
• The full element list at each spot, since base metal presence is often the decisive evidence.
• Weight and the density conclusion where the item was substantial enough to warrant the check.
• Any marks observed, including their location and condition.
• The reference standard result from that session, demonstrating the instrument was in control.

A record built this way converts a disputed opinion into a documented finding. It also makes the reasoning reviewable later, which is how sampling protocols actually improve.