How to Clean Jewelry at Home Without Wrecking the Setting

The Method That Covers Most Jewelry - and the Word That Breaks It

For the great majority of pieces the answer is dull, and it is the same one the institutions give. GIA: "Most colored gems can be cleaned with warm water, mild dish soap (no detergents) and a soft brush." The American Gem Society is more procedural, recommending "a mixture of warm water and just a drop or two of a mild dishwashing detergent", a soak of approximately twenty minutes, a soft brush, then a rinse. And take GIA's practical detail seriously: rinse in a glass or bowl of water rather than over an open drain, because a chain around a plughole is a failure mode with no recovery.

The word that breaks the method is soak. Immersion is fine for a solid metal ring with a securely set, untreated stone, and it is wrong for four categories: porous and organic materials - GIA on coral is terse, "briefly rinse (do not soak it)"; anything strung, because of the string rather than the beads; assembled stones such as opal doublets and triplets; and anything treated, where the water is not the problem but the time is. When you are unsure which category a piece falls into, wipe it with a barely damp cloth and stop there.

Soapy water is not a fallback for difficult stones - it is frequently the only method GIA endorses for them. Opal: "The only safe way to clean opal is with warm, soapy water." Amber gets the same phrasing, and emerald, tanzanite, turquoise and lapis end up in the same place. Metal-specific routines have their own guides: cleaning silver deals with tarnish rather than dirt, and cleaning gold deals with karat, alloy and construction.

The Thing Nobody Tells You: the Setting Is the Risk

GIA's guidance on ultrasonic cleaning does not begin with the stone at all. It begins with the mounting: "Check to make sure the gemstones in your piece are securely set both before and after cleaning. The vibration generated by an ultrasonic cleaner can sometimes shake gems loose or cause damage if gems are set with their girdles touching." GIA states the shorter version elsewhere too - "Ultrasonic cleaners and steam cleaners can loosen gemstones in their settings." Two separate places, same warning, and it is about the metal rather than the mineral.

The second half of that sentence repays reading twice. The girdle is the thin band around the widest part of a faceted stone, and in pave, micropave and channel work the stones are deliberately set shoulder to shoulder with those girdles in contact. Put that piece in a bath of vibrating fluid and adjacent stones tap against one another thousands of times a second. Diamond chips diamond quite happily. So a pave halo, a diamond eternity band or a channel-set line carries a real ultrasonic risk even though the stone scores a ten on Mohs - exactly the case a stone-by-stone table marks as safe.

The consequence is a habit rather than a rule. Look at the piece before it goes near water or a machine, and press each stone gently with a fingertip or a wooden toothpick to see whether anything rocks. Then look again afterwards. GIA is worth quoting on why: "a diamond that's loose in its setting can wear through a prong over time." A stone that moves is abrading its own mounting every time you wear it.

Hardness, Toughness and Stability: Why Mohs Alone Misleads

The most useful thing GIA publishes about gem care is not a list, it is a model. Durability has three components: hardness, "resistance to scratches and abrasions"; toughness, "resistance to breaking and chipping"; and stability, "resistance to chemicals, heat, humidity and light". A gem can be excellent at one and poor at another, which is why GIA's summary is the line to remember: "No single gemstone scores a 'perfect 10.'"

The counterexample is GIA's, not ours. Emerald is a hard stone by any normal measure, and GIA notes that the oil and resin treatments most emeralds carry "can be damaged by ultrasonic cleaning or exposure to alcohol or other organic solvents." The underlying vulnerability is toughness rather than hardness: GIA reports that "some estimates state that 90 percent or more of emeralds are fracture-filled" - a figure about treatment prevalence, which GIA presents as an estimate rather than a measurement - and it is explicit that "Ultrasonic vibrations can weaken already-fractured stones." Anyone reasoning from hardness alone would have put the emerald in the machine.

Stability is where treatments live, and treatments are why two stones that look identical need opposite handling. Oiled, resin-filled, wax-impregnated, dyed and coated stones are, for care purposes, different objects from the untreated mineral. GIA's advice on lapis generalises to anything dyed: "It's best to test a small, inconspicuous area first, because some dye treatments are not stable."

Ultrasonic and Steam Are Not the Same Risk

Most articles treat the two machines as interchangeable, and they are not. GIA describes how an ultrasonic works: "These machines generate high-frequency sound waves that travel through water or a cleaning solution creating heat and pressure." The primary hazard is mechanical - cavitation and vibration, which shake stones loose, drive existing fractures further and chip stones whose girdles touch. A steam cleaner's primary hazard is thermal: it sweats fillers out of fractures, crazes heat-sensitive stones and degrades wax, oil, plastic impregnation and coatings. What they share is the setting risk, which GIA attributes to both machines equally.

GIA's emerald page illustrates it: the ultrasonic attacks the stone's fractures, while the steam attacks what is in them - "Hot steam can cause oil or unhardened resin to sweat out of fractures." GIA's conclusion is that "Using warm, soapy water coupled with gentle scrubbing is the safest way to clean emeralds."

GIA does publish an itemised list of what must never go in an ultrasonic. Organics: "Pearls, coral, tortoise shell, ivory, shell cameos, jet, and amber should not be put in an ultrasonic cleaner." Heat and temperature-sensitive stones: "tanzanite, feldspar (sunstone and moonstone), fluorite, iolite, kunzite, lapis lazuli, malachite, opal, topaz, turquoise, zircon, and others". Anything with surface-reaching breaks that have been filled. Anything impregnated or coated with "oil, plastic or wax", where GIA names emerald, lapis lazuli, malachite and opal. And star ruby and star sapphire specifically.

Set against the safe method, that produces a table you can scan.

Material	Wipe with a soft cloth	Soak in warm soapy water	Ultrasonic
Untreated diamond, sapphire or ruby in a solid setting	Yes	Yes	GIA: not recommended for home use
Pave, micropave or channel work, any stone	Yes	Yes	Risk is the setting, not the stone - GIA on girdles touching
Fracture-filled or clarity-enhanced diamond	Yes	Briefly	Never (GIA)
Emerald, lapis lazuli, malachite, opal - oiled, waxed or plastic-impregnated	Yes	Yes, and the only method GIA endorses	Never (GIA)
Pearl, coral, tortoise shell, ivory, shell cameo, jet, amber	Yes, after every wear	Occasionally; briefly rinse coral, never soak a strung strand	Never (GIA)
Tanzanite, sunstone, moonstone, fluorite, iolite, kunzite, topaz, turquoise, zircon	Yes	Yes	Never (GIA)
Star ruby, star sapphire	Yes	Yes	Never (GIA)
Opal doublet or triplet	Yes	No - cemented assembly; our inference, not a GIA statement	Never
Gold-plated or vermeil	Yes, and nothing else	No	Never - CCI on plated silver

Here is the gap to respect: GIA publishes no equivalent standalone list for steam that we could find. Steam prohibitions have to be assembled stone page by stone page, so do not believe any article - including this one - that presents a single authoritative steam table. Where GIA flags sudden temperature changes, as it does for opal, topaz and tanzanite, steam is disqualified on the same reasoning. Beyond those, you are inferring rather than citing.

Diamonds deserve the nuance, because the trade shorthand is wrong. "Diamonds are fine in an ultrasonic" is not GIA's published position for consumers: "Harsher cleaning methods are not recommended for home use. These include powdered abrasive household cleansers, ultrasonic cleaners, and steam cleaners." The honest reading is that for an untreated, well-set diamond the risk sits in the mounting rather than the stone, and that an ultrasonic is a professional tool operated by someone who will re-tighten the setting afterwards. For fracture-filled or clarity-enhanced diamonds it is a prohibition rather than a nuance: GIA warns that "the filling can be damaged during common jewelry repairs, or if subjected to repeated cleanings with steam, acid or ultrasonics," and that "If the filler turns dark, it cannot be made colorless again."

Pearls and Opals: the Two Materials Handled Wrong Most Often

Pearls first, because the standard advice is wrong in both directions at once. GIA puts pearl at 2.5 on the Mohs scale, and adds the clause that governs everything else: "Pearl can be damaged by many chemicals and all acids. The list includes hair spray, perfume, cosmetics, and even perspiration." Pearls "should never be cleaned in an ultrasonic or steam cleaner." But GIA does not tell you to avoid soap - "It's safe to use warm, soapy water for occasional, thorough cleaning." Routine care is mechanical, "it's best to wipe cultured pearls with a very soft, clean cloth after each wearing"; soapy water is an occasional deep clean.

The reason not to submerge a strand is the silk, not the nacre. GIA's instruction is about drying - "If the pearls are strung, be sure the string is completely dry before wearing" - and wet silk stretches, wicks moisture inward at the drill holes and holds whatever it absorbed. The related rule, that pearls go on last and come off first, is sound, and it is not merely trade shorthand: GIA publishes it almost word for word, that "Pearl jewelry should be the last thing you put on and the first thing you take off." The reasoning sits in the same guidance - hair spray, perfume, cosmetics and perspiration all damage pearl, and all of them arrive before the jewellery does.

Restringing is where the sources contradict themselves. GIA's care tips page states that "Pearls worn often should be restrung once a year." GIA's own pearl FAQ says only that pearl necklaces "may need to be restrung periodically as the silk string can darken, stretch, or wear thin over time" and gives no interval. We could find no study, standard or failure data behind the twelve-month figure from any body. The FAQ version is more defensible because it names observable criteria: darkening of the silk, stretching that shows as gaps between pearls, fraying at the clasp end. Condition is the trigger, not the calendar. Our guide to the types of pearls covers nacre and how thin it can be.

Opals attract a myth worth correcting precisely, because most corrections overshoot. Solid opal is Mohs 5 to 6.5 with toughness GIA rates "very poor to fair", takes warm soapy water and nothing else, and is vulnerable to heat. On storage GIA is unusually relaxed: "Some people do store their opals in water. There's no evidence that it prevents them from drying out but it can't hurt them and might be a good idea if you live in a very dry climate." So the correct shape of the correction is unnecessary, not dangerous - not "never do this". And the claim that household water contact causes crazing has no support in GIA's material, which attributes crazing to intense light and heat and to moisture loss.

That answer must not be generalised to assembled opals, and here we are reasoning rather than citing. GIA describes a doublet as "a thin layer of opal cemented to a backing" and a triplet as a thin opal layer cemented between a quartz or glass top and a backing. What GIA does not publish anywhere we could find is a statement that water damages assembled opals. The trade rule against soaking them is a reasonable inference from that word cemented - but it is an inference, and you should know which you are being given.

Metals, Plating, and How Often to Have Things Looked At

Tarnish is not dirt, and treating it as dirt is why people scrub. The Canadian Conservation Institute describes it as a film "mainly of black silver sulphide", "caused by sulphur-containing compounds such as hydrogen sulphide in the air" - a reaction with the air, which is why washing does nothing. CCI's emphasis is on prevention: "Preventing tarnish by using tarnish-inhibiting products in sealed storage containers or display cases should be a priority." That is why how you store jewelry matters more than how you clean it, and why a lined jewelry box or a closable travel case earns its place.

Every abrasive is a withdrawal from a finite account. CCI notes that abrasive polishes "leave behind a pattern of fine scratches", and that "each polishing or chemical tarnish removal takes away a certain amount of underlying silver along with the surface layer of tarnish". Its recommended mild abrasive is precipitated calcium carbonate - precipitated chalk - and the word precipitated is about particle size. On toothpaste we want to be honest about what kind of claim we are making, because no gemmological or conservation body publishes an explicit prohibition that we could find. Silver and gold sit around 2.5 to 3 on Mohs. The abrasives in toothpaste are typically hydrated silica, generally cited around 5 to 6, sized to work against tooth enamel at about 5 rather than against soft metal. Put that beside CCI on permanent scratching and cumulative metal loss and toothpaste is the wrong tool - but that is a derivation, not a citation. Baking soda is a separate case: at roughly 2.5 on Mohs it is milder and closer to CCI's chalk, though still an abrasive and still emphatically wrong on anything plated.

Commercial silver dips are a different chemistry again. CCI describes them as "composed of an acid and a complexing agent", and warns that "acids are corrosive and will damage niello, bronze, stainless steel knife blades, and organic materials such as wood." Dip must never reach a sealed hollow component - hollow beads, hollow bangles, lockets - because the liquid "becomes virtually impossible to wash... out". Being acidic, dips are also disqualifying for any piece set with pearls, coral, amber, malachite, turquoise or shell. And a dip strips an intentional oxidised finish, which is not cleaning, it is redesign.

Plated and vermeil pieces are the ones readers most often destroy while trying to help, and the decisive quote is CCI's: "If the object is silver plated, any method of tarnish removal may be very damaging." The FTC makes the scale concrete. Under 16 CFR 23.4, vermeil requires a sterling silver base coated with gold of not less than 10 karat fineness, at "a minimum thickness throughout equivalent to two and one half (2 1/2) microns (or approximately 100/1,000,000ths of an inch) of fine gold", and of "reasonable durability". That minimum is roughly one ten-thousandth of an inch, and every abrasive pass is drawn against it. You cannot re-polish a plated piece; you polish through it. The trap is that the sterling core tarnishes and shows through at wear points, and the instinctive response is to reach for silver polish or a dip - precisely what removes the remaining gold. Our explainers on gold vermeil and the 925 mark set out the standards behind both terms.

Finally, the question every cleaning article answers confidently and none should. GIA says "Jewelry should be checked every six months and cleaned frequently." Jewelers of America says to "have your piece professionally cleaned at least once a year." Those are not the same claim: GIA gives a checking interval, Jewelers of America gives a cleaning interval and no checking interval at all, so the two are only partly comparable, and neither rests on published failure data we could locate. Neither says prongs, either - the near-universal claim that prongs should be checked every six months is a trade extrapolation. Both also represent retail jewellers, for whom a six-month cadence is footfall as well as maintenance. How hard and how often you wear a piece matters far more than the calendar.

Frequently Asked Questions

What is the safest way to clean jewelry at home?

Warm water, a mild dish soap, a soft brush and a rinse. GIA gives that as the method for most coloured stones, and it is the only method it endorses for several fragile ones including opal, amber, emerald, tanzanite, turquoise and lapis. Rinse in a glass or bowl rather than over an open drain, and skip the soak entirely for anything strung, porous, treated or assembled.

Can I use an ultrasonic cleaner on my jewelry at home?

GIA does not recommend it for home use. Its diamond care guidance lists ultrasonic and steam cleaners among "harsher cleaning methods" that "are not recommended for home use", and it separately warns that both machines can loosen gemstones in their settings. The risk easiest to overlook is not the stone at all - GIA notes that vibration "can sometimes shake gems loose or cause damage if gems are set with their girdles touching", which is exactly how pave and channel settings are built.

Is toothpaste safe for cleaning jewelry?

We would not use it, but we want to be clear about what kind of answer that is. No gemmological or conservation body publishes an explicit ban that we could find, so the case is reasoning from sourced premises rather than a quoted prohibition. Silver and gold sit around 2.5 to 3 on the Mohs scale; the abrasives in toothpaste are typically hydrated silica, generally cited around 5 to 6 and sized to work on tooth enamel at about 5 rather than on soft metal; and the Canadian Conservation Institute establishes that abrasives leave permanent fine scratches and remove metal each time. On anything plated or vermeil it is not a judgement call - CCI says that on silver plate any method of tarnish removal may be very damaging.

Can pearls be cleaned with soap and water?

Yes, occasionally. GIA is explicit that "It's safe to use warm, soapy water for occasional, thorough cleaning" - the widespread advice to keep pearls away from soap entirely overstates it. Routine care is mechanical: wipe them with a very soft, clean cloth after each wearing. What you should not do is leave a strand submerged, and the reason is the silk rather than the nacre - GIA asks that the string be completely dry before wearing. Never put pearls in an ultrasonic or steam cleaner, and keep them away from anything acidic.

Should opals be stored in water?

It is unnecessary, but for a solid opal it is not dangerous, and the distinction matters. GIA addresses the myth directly: some people do store opals in water, there is no evidence it prevents them drying out, "but it can't hurt them and might be a good idea if you live in a very dry climate." Do not extend that to a doublet or triplet, which GIA describes as thin opal layers cemented to a backing. No institution we found states that water damages an assembled opal, but prolonged immersion is a reasonable thing to avoid around an adhesive layer.

How often should jewelry be professionally inspected?

There is no evidence-based answer, and the two sources usually quoted are not answering the same question. GIA says jewelry should be checked every six months. Jewelers of America says pieces should be professionally cleaned at least once a year, and states no checking interval at all. We could find no study or failure data supporting either figure, and neither body actually says prongs - the common claim that prongs need checking every six months is an extrapolation. Wear pattern beats the calendar, and any stone that rocks under a fingertip needs attention immediately.

How We Research

Our guides draw on manufacturer documentation, established industry practice, and what we learn researching products and listings for our buying guides. Where practice varies or opinion divides, we say so rather than pretending at consensus.

Before you clean anything: The cheapest improvement to any cleaning routine is doing it less often, and the second cheapest is looking before you start. Check that every stone is tight - GIA asks for that check both before and after cleaning, and notes that a stone loose in its setting can wear through a prong over time. Then ask what the piece actually is rather than what it resembles: solid or plated, treated or untreated, strung or set, one material or several. Almost every ruined piece we come across was cleaned as though it were something simpler than it was, and the commonest version of that mistake is treating a vermeil chain as though it were solid silver.