Testing 22K Indian Wedding Gold: Step-by-Step XRF Guide

Indian wedding gold is one of the most demanding categories a buying counter encounters. The metal is high karat and genuinely valuable, the pieces are often heavy, and the emotional weight attached to them means every reading is scrutinized. At the same time, the construction techniques traditional to this jewelry break several assumptions that XRF quietly depends on.

An operator who tests a kundan necklace the way they test a plain chain will get a number, and that number may overstate the metal by a significant margin. This guide covers what makes 22K bridal gold different, where the standard procedure fails, and the specific protocol that produces defensible valuations.

What 22K actually means

Karat is a fraction of 24. 22K is 22 parts gold in 24, or 91.6% by mass. The remaining 8.4% is typically copper and silver, added because pure gold is too soft to hold a setting or survive daily wear.

India’s hallmarking system stamps this as 916, referring to 916 parts per thousand. Since the mandatory hallmarking regime administered by the Bureau of Indian Standards was extended, a compliant piece carries a BIS mark, the purity grade, and a six-digit alphanumeric HUID.

Why 22K rather than 18K

Western jewelry gravitates to 14K and 18K for durability. Indian bridal gold stays at 22K for reasons that are simultaneously cultural and financial: the jewelry functions as a store of household wealth, and higher purity means a higher proportion of the purchase price is recoverable metal rather than alloy and labor.

This matters at the counter because the seller usually knows the karat with confidence. Unlike a customer bringing in inherited costume pieces, a family selling wedding gold generally knows what they bought and roughly what it should return. A reading that comes in low will be challenged, and the operator needs to explain it credibly.

The tolerance question

A 22K piece will rarely assay at exactly 91.6%. Manufacturing tolerance, surface condition, and measurement uncertainty all contribute. Readings in the 90.5% to 92.5% band are normally consistent with genuine 22K.

What deserves attention is a reading well outside that band. Above roughly 93% suggests surface enrichment rather than genuinely higher purity. Below 89% suggests either a lower-karat alloy sold as 22K, or that the tested spot is not representative of the piece.

Construction features that distort readings

This is where Indian bridal gold departs sharply from generic jewelry, and where most valuation errors originate.

Lac filling

A great deal of traditional Indian jewelry — particularly bangles, larger pendants, and temple-style work — is filled with lac, a natural resin. The filling gives a lightweight hollow shell the heft and rigidity of a solid piece, and it is entirely legitimate traditional construction rather than a deception.

The problem for XRF is that lac is organic. Carbon, hydrogen and oxygen fluoresce far too weakly for a benchtop instrument to detect, so the filling is invisible. The instrument measures the gold shell, correctly reports 91.6%, and gives no indication whatever that a large fraction of the item’s weight is resin.

An operator who multiplies total weight by 91.6% on a lac-filled bangle will overvalue it substantially. On a heavy bangle the resin can account for a third or more of the mass.

Kundan settings

Kundan work sets stones using thin foils of highly refined gold, typically close to 24K, pressed into a lac base. The visible gold at the setting is therefore purer than the body of the piece.

Testing on or near a kundan setting reads high — sometimes markedly so — because the analyzer is measuring foil that genuinely is near-pure gold. The reading is accurate for the spot and misleading for the item.

Meenakari enamel

Meenakari applies vitreous enamel to the gold surface for color. Enamel is a glass, composed largely of silicon, sodium, and metal oxides. It attenuates the gold signal without contributing a detectable fluorescence of its own.

A reading taken through enamel comes out low, and unlike most low readings this one has nothing to do with alloy content. Enamelled areas must simply be avoided.

Hollow and thin-walled construction

Much bridal jewelry is deliberately hollow to achieve visual scale without prohibitive metal cost. Walls of 0.3–0.5 mm are common and comfortably exceed the critical depth XRF needs, so composition readings on the shell are valid.

The valuation risk is not the reading but the assumption behind it. Hollow items feel substantial while containing far less metal than their volume suggests.

Solder joints

Chains, clasps, and multi-component pieces are joined with solder alloys that melt below the parent metal. Gold solder is typically lower karat and richer in silver and zinc.

Measuring across a joint pulls the reading down and introduces zinc that is not present in the body alloy. A zinc signature on a piece that should be a simple gold-copper-silver alloy is a reliable indication that the measurement spot included solder.

The measurement protocol

The following sequence addresses each of the distortions above. It takes slightly longer than a single-spot reading and eliminates the great majority of valuation errors.

Step one: clean the test area

Wedding gold is often stored for years and arrives with polish residue, tarnish film, and skin oil. All three attenuate signal and bias readings downward. Wipe the intended test spots with isopropyl alcohol and allow them to dry.

This step costs thirty seconds and removes a systematic error that operators frequently misattribute to alloy content.

Step two: weigh before testing

Record the total weight first. This becomes the cross-check for hollow and filled construction later, and taking it before testing prevents the reading from anchoring your expectation of what the weight should be.

Step three: select spots deliberately

Choose at least three spots, and choose them for what they represent rather than for convenience:

• A flat, plain area of the main body — the most representative of bulk alloy
• A high-wear edge or contact point — where any surface treatment has thinned
• A second body location distant from the first — testing homogeneity across the piece

Deliberately avoid kundan settings, enamelled sections, solder joints, and any area adjacent to a stone.

Step four: apply the variance rule

Compare the three readings. If they agree within about 2 percentage points, the piece is homogeneous and the average is a sound basis for valuation.

If they disagree by more than that, the piece is not uniform and a single average is not defensible. Investigate before pricing — the spread is telling you something specific about the construction. The diagnostic logic for interpreting that spread is developed in our plated versus solid gold tutorial.

Step five: reconcile weight against volume

Solid 22K gold has a density of approximately 17.3 g/cm³. Estimate the item’s volume from its dimensions and compare implied weight to actual weight.

A large shortfall indicates hollow construction, lac filling, or both. For a filled bangle, the practical approach is to price on the assessed metal content rather than on total weight, and to explain that distinction to the seller before quoting.

Regional variation in what arrives at the counter

“Indian wedding gold” covers a range of regional traditions with meaningfully different construction. Recognizing the style shortens diagnosis considerably.

South Indian temple jewelry

Typically heavy, often 22K, with substantial repoussé work depicting deities and motifs. Frequently lac filled to support the raised relief, and commonly incorporating rubies and uncut stones in kundan-style settings.

The deep relief creates a practical problem beyond composition: finding a flat area large enough to cover the measurement aperture. Recesses shadow the detector and produce depressed readings. The reverse face is usually plainer and gives a better surface.

Rajasthani and Jaipur meenakari

Distinguished by vivid enamel, frequently on the reverse of a piece whose front carries stones. This means both faces may be unsuitable — one obscured by settings, the other by enamel.

Test the edges, the connecting links, or the clasp. On a fully enamelled piece with no accessible bare metal, XRF may not be able to give a representative bulk reading at all, and that limitation should be stated rather than worked around.

Bengali and eastern filigree

Fine wire work, often lighter in weight and higher in surface-to-volume ratio. Filigree presents a problem the operator can easily miss: the piece may be smaller than the beam spot in every direction, so part of the beam passes through the gaps and strikes the sample stage.

The result includes whatever the stage is made of. Placing filigree over a clean gold-free backing, or folding the piece to present continuous metal, prevents contaminated readings.

Kerala and Malabar plain gold

Often plain, high-karat, minimal stonework, sometimes 22K or higher. This is the most straightforward category — flat surfaces, homogeneous alloy, few construction complications — and standard three-spot procedure works cleanly.

Buying versus lending against the piece

The measurement is the same; the tolerance for error is not.

Outright purchase

The valuation is final at the point of transaction. Errors are realized immediately and cannot be recovered, which justifies the full protocol including weight reconciliation on every substantial piece.

Pawn loan

Here the piece is collateral, and the exposure runs until redemption or forfeiture. Two additional considerations apply.

First, the loan-to-value ratio provides some buffer against valuation error, but it is not a substitute for accuracy — a systematically overvalued book accumulates loss quietly across many small transactions rather than in one visible event.

Second, the item must be identifiable on return. Photographing the piece alongside the reading record protects both parties, particularly with sets where individual components can be substituted.

Refining and scrap

When the destination is the melt pot, filled and hollow construction ceases to be a valuation subtlety and becomes a direct settlement issue. Lac burns off; the metal that remains is what the refiner pays on.

For lots destined for refining, assessed metal content matters more than any per-piece karat figure, and reconciliation against the refiner’s return becomes the real accuracy check. That reconciliation process is set out in our settlement audit guide.

Hallmarks: useful evidence, not proof

A BIS hallmark is meaningful and should be recorded, but treating it as a substitute for measurement is a mistake that costs money.

What a current BIS mark contains

A compliant hallmark carries the BIS standard mark, the purity grade in both karat and fineness form, and a six-digit alphanumeric HUID unique to the article. The HUID can be verified through the official BIS verification channel, which returns the registered jeweller and article details.

That verification is genuinely valuable. It confirms the piece entered the hallmarking system and was registered by an identifiable jeweller.

Why it does not replace testing

Several ordinary situations produce a legitimate hallmark on a piece that does not assay as expected:

• Older pieces predate mandatory HUID and carry only a purity mark, or no mark at all. Absence of a hallmark on inherited jewelry is normal and says nothing about purity.
• The hallmark applies to the article as registered, and pieces are sometimes repaired or modified afterward with different alloy.
• Marks are transferable in practice. A stamped section can be removed from one item and attached to another.
• Counterfeit marks exist, and a convincing fake hallmark is far cheaper to produce than the gold it claims.

The sound position is that a hallmark raises confidence and directs attention, while the measurement determines the price. Where the two disagree, the measurement governs and the discrepancy is worth discussing with the seller rather than concealing.

Handling the conversation

Technical procedure is only half the job with bridal gold. The seller is frequently parting with jewelry that carries family significance, sometimes under financial pressure, and the interaction requires care that a scrap transaction does not.

Test in view and narrate

Conduct the measurement where the seller can see it, and say what you are doing as you go. “I am testing three separate places so we get an accurate average, and I am avoiding the stone settings because the foil there is purer than the rest and would read high.”

Transparency converts an opaque procedure into a shared one. It also pre-empts the suspicion that the machine has been manipulated.

Explain a low reading before quoting

If readings come in below expectation, explain the reason before the price. There is a large difference between “this is only 89%” and “the three readings averaged 89.2%, which is below the 91.6% a 22K mark indicates — this can happen when a piece has been repaired with lower-karat solder, and it means I have to price on the measured content.”

The second version is verifiable, non-accusatory, and leaves the seller able to disagree on substance rather than feeling dismissed.

Address filled pieces directly

Lac filling is traditional construction, not fraud, and sellers are frequently unaware of it. Framing matters: the piece contains less gold than its weight suggests because it was made that way, which is normal for the style and does not mean anyone was cheated.

Offering to demonstrate — pointing to the density calculation, or showing the reading at a worn edge — turns a disputed number into an explained one.

Document what you measured

Record the spot locations, individual readings, average, weight, and any hallmark details. If a transaction is later questioned, a contemporaneous record of three specific readings is a far stronger position than a recalled single number. The record-keeping structure used for bulk intake, described in our bulk scrap intake SOP, adapts directly to single high-value pieces.

Common failure patterns

Five errors account for most disputed valuations on this material.

Testing a single spot on a complex piece

A bridal set may combine plain body metal, kundan foil, enamel, and solder within a few centimetres. One reading cannot represent that. This is the most frequent and most consequential mistake.

Multiplying total weight by measured purity on a filled item

The measurement is correct and the arithmetic is correct, but the assumption connecting them is wrong. Weight includes resin the analyzer cannot see.

Accepting an elevated reading uncritically

A reading of 94% on a piece marked 916 is not good news; it is a signal that the tested area is not representative. Surface enrichment from pickling and polishing depletes copper and silver from the top few microns, and the analyzer only sees those microns.

Testing through enamel or lacquer

Both attenuate the signal and depress the reading. An operator unaware of the mechanism attributes the low result to alloy and underpays.

Testing before the instrument has stabilized

A cold instrument drifts. Thirty minutes from power-on is a reasonable minimum before the first valuation of the day, and a reference standard check afterward confirms it. The calibration routine covers this verification.

Frequently asked questions

What percentage gold should 22K read?

91.6% is the nominal figure. In practice, readings between 90.5% and 92.5% are consistent with genuine 22K once manufacturing tolerance and measurement uncertainty are accounted for. Results outside that band warrant investigation rather than immediate rejection.

Can XRF detect lac filling inside a bangle?

No. Lac is organic, composed of elements too light to fluoresce detectably, so it is invisible to the instrument. Detection relies on comparing actual weight against the weight implied by the item’s volume at 22K density, roughly 17.3 g/cm³.

Why does the reading go up near a stone setting?

Kundan settings use near-pure gold foil, typically close to 24K, pressed into a lac base. The analyzer correctly measures that foil. The reading is accurate for the spot and unrepresentative of the piece, which is why settings should be avoided as test locations.

Is a BIS hallmark sufficient to confirm purity?

It is strong supporting evidence and the HUID can be verified officially, but it does not replace measurement. Pieces are repaired after hallmarking, marks can be transferred between items, and counterfeit marks exist. Test regardless of marking.

My reading is above 92.5% — is the gold better than 22K?

Usually not. Surface enrichment from pickling and polishing removes copper and silver from the outermost microns, and XRF only penetrates that depth. Testing a high-wear edge where the enriched layer has worn away typically brings the reading back toward the true bulk value.

How many spots should I test on a necklace?

Three at minimum on a simple piece, and more on complex bridal sets where multiple construction techniques are combined. The rule is to test each distinct construction zone rather than to hit a fixed count.

Does testing damage the jewelry?

No. XRF is entirely non-destructive, leaves no mark, and can be repeated as often as necessary. This is a meaningful advantage over acid testing when handling pieces of sentimental value, and it is worth stating to the seller explicitly.

What does zinc in the reading indicate?

Zinc is not a normal component of a 22K gold-copper-silver alloy but is common in gold solder. Its appearance almost always means the measurement spot included a soldered joint. Move to a plain body area and re-measure.

The protocol in short

Clean the surface. Weigh first. Test at least three deliberately chosen spots, avoiding kundan foil, enamel, and solder. Require agreement within two percentage points before averaging. Reconcile weight against volume at 17.3 g/cm³ to expose hollow or filled construction. Record everything.

The single principle underneath all of it: XRF measures the surface it is pointed at, and Indian bridal gold is built from several different surfaces. Valuation accuracy comes from choosing which surfaces to measure, not from the instrument.

For the depth limitation that makes spot selection so consequential, see plating thickness measurement limits. For applying the same discipline to plated versus solid determination, see the gold-coated jewelry verification tutorial.

A worked example from the counter

The abstract rules are easier to apply against a specific case. The following combines several routine situations into one representative transaction.

What arrived

A seller brings a bridal set: one heavy bangle pair, a long haram-style necklace with kundan settings, and a plain chain. Total weight 186 g. The bangles carry a 916 stamp; the necklace has no visible mark; the chain is stamped 22K.

What the readings showed

The plain chain tested at three points along its length: 91.4%, 91.7%, 91.5%. Tight agreement, no zinc, flat profile. This is uncomplicated 22K and prices directly on weight.

The bangles tested at 91.8% on the outer face, 91.6% on a second outer position, and 92.1% near the rim. Consistent within tolerance. But the pair weighed 94 g for a volume that suggested roughly 140 g if solid. The shortfall is decisive — these are lac-filled, and pricing on 94 g × 91.6% would overstate the metal by a wide margin.

The necklace produced the widest spread: 93.8% adjacent to a kundan setting, 91.2% on a plain link, and 88.4% across a connector joint. Three readings, three different materials — foil, body alloy, and solder respectively. Averaging them would have been meaningless.

How it was resolved

Each component was priced on its own basis. The chain on full weight at measured purity. The bangles on assessed metal content after accounting for filling, with the density calculation shown to the seller. The necklace on the plain-link reading of 91.2%, with the setting and joint readings excluded and the reason explained.

The seller disputed the bangle valuation initially, which is the normal reaction. What settled it was demonstrating the arithmetic: the volume, the density of 22K gold, the implied solid weight, and the gap. The conversation moved from “your machine is wrong” to “I did not know they were filled” — which is a resolvable position.

Building consistency across staff

A protocol that only the experienced operator follows produces inconsistent valuations, and inconsistency between staff members is more damaging to trust than a single wrong number.

Write the spot-selection rules down

“Test three spots” is insufficient guidance. “Test a plain body area, a worn edge, and a second body area distant from the first; never test on or within 5 mm of a stone setting, enamel, or a visible joint” is a rule a new employee can follow on day one.

Make weight reconciliation mandatory above a threshold

Requiring a density cross-check on every item above a set weight — 20 g is a reasonable trigger — catches filled construction systematically rather than depending on an individual operator’s suspicion.

Standardize the explanation

Agree the wording used to explain a low reading, a filled piece, and a hallmark discrepancy. When every staff member gives the same explanation, the explanation reads as policy rather than improvisation, and sellers challenge it far less.

Review disputed transactions

When a valuation is contested, revisit the recorded readings rather than the outcome. Most disputes trace to spot selection, and reviewing them is the fastest way to improve the protocol. Instruments are rarely the problem; sampling almost always is.