Can you actually tell if a diamond is real at home?

No, not with confidence. Home tests can sometimes rule a stone out, but none of them can rule one in. The reason is that the two materials most likely to be sitting in a ring instead of a diamond, moissanite and cubic zirconia, each defeat a different set of the popular tests, and a laboratory-grown diamond passes all of them because it is chemically a diamond. The only definitive answers are a grading report you verify yourself on the laboratory's own website with the girdle inscription matched to it, or a qualified gemmologist with a dual thermal and electrical tester.

Does the fog test work on a diamond?

Barely. Breathing on a stone tests heat conduction, and diamond does clear fast, but so does moissanite, which conducts heat almost as well. There is no defined number of seconds that separates a pass from a fail, and the result shifts with room temperature, humidity, how close you hold the stone and whether there is skin oil on it. At best a stone that stays fogged for several seconds is probably glass or something similarly poor at conducting heat. A stone that clears instantly has told you nothing you can act on.

Will moissanite pass a diamond tester?

Yes, on a thermal-only tester it will read as diamond. GIA researchers reported this when synthetic moissanite reached the jewellery market in the late 1990s, noting that the thermal probes then on sale reacted to it as though it were diamond, because moissanite conducts heat almost as well as diamond does. That failure is the entire reason dual testers exist. A dual tester adds an electrical conductivity probe, and moissanite is electrically conductive while ordinary diamond is not, so the second reading separates them.

Can any home test tell a lab-grown diamond from a natural one?

No, and this is the most important thing on the page. A laboratory-grown diamond is a diamond. Same carbon, same crystal structure, essentially the same optical and physical properties. GIA states plainly that because the two are essentially chemically and optically the same, traditional gemological observations and instrumentation cannot tell them apart. Separating them takes spectroscopy and fluorescence imaging in a laboratory. So a stone can pass every test on this page and still be lab-grown. Origin is a paperwork question, not a test-at-home question.

Does the water test prove a diamond is real?

No. Dropping a loose stone in water tests density, and diamond is dense enough to sink immediately, but so is almost every stone it might be confused with. Cubic zirconia is considerably denser than diamond and sinks faster. Moissanite, white sapphire, zircon, topaz and spinel all sink. The only thing this test reliably identifies is a stone light enough to hesitate on the way down, which in practice means glass or quartz, and even then only sometimes. It is the most widely repeated diamond test and the least useful one.

Should I do the scratch test on my diamond?

No. Do not do it. It is a destructive test, it can chip or abrade the stone, it can mark the setting, and a positive result proves nothing anyway. Diamond is the hardest gem material at 10 on the Mohs scale, but moissanite sits at 9.25 and white sapphire at 9, and both will scratch almost everything a diamond will. Diamond also has perfect cleavage in four directions, so it is hard but not indestructible. You risk damaging something valuable in exchange for a reading you could not trust.

How do I check a GIA report number myself?

Ask the seller or your insurer for the report number, then type it into GIA Report Check at gia.edu/report-check on your own device. Compare every line on screen against the paperwork and the stone: carat weight, colour, clarity, cut grade where one is given, the measurements in millimetres, fluorescence and whether the report is for a natural or a laboratory-grown diamond. Then ask to see the girdle under 10x magnification and check that any laser inscription reads as the same number. Report Check covers reports dated from 1 January 2000 onwards.

Where can I get a diamond tested in South Africa?

GIA runs a laboratory in Bedfordview, Johannesburg, which accepts submissions by transport or walk-in, so a stone can be graded without leaving the country. That is the route for a definitive answer on identity and origin. For a faster and cheaper first pass, a qualified gemmologist or a registered jewellery valuer with a dual thermal and electrical tester and a microscope can separate a diamond from a simulant in minutes. Ask what instruments they will use before you go, and ask for the finding in writing.

How to Tell If a Diamond Is Real

Bottom line up front

Short answer: you cannot confirm a diamond at home. Home tests can sometimes rule a stone out. None of them can rule one in.

Two things give a definitive answer. A grading report from a major laboratory whose number you verify yourself on that laboratory’s own website, with the girdle inscription matched against it, and a qualified gemmologist with a dual thermal and electrical tester, a loupe and a microscope.
Moissanite defeats the classic tests, including a handheld thermal tester. When synthetic moissanite reached the jewellery market, GIA researchers reported that its thermal properties are so close to diamond that the probes then on sale read it as diamond. That single failure is why dual thermal and electrical testers exist, and it quietly invalidates the fog test, which is the same measurement done worse.
A laboratory-grown diamond is a real diamond. Same carbon, same crystal structure, essentially the same optical and physical properties. GIA states that because the two are essentially chemically and optically the same, traditional gemological observations and instrumentation cannot tell them apart. It takes laboratory spectroscopy, so a stone can pass every test on this page and still be lab-grown.
Skip the laboratory report only for a piece you will never insure, sell or leave to anyone. A gemmologist with a dual tester settles identity in minutes for a fraction of the cost. The full report earns its money when the stone goes onto an insurance schedule, into an estate or into a sale, because those are the moments somebody else decides what it is.

Everything below is indicative only, ranked by how much it is actually worth.

”Is it real” is three different questions

Almost every argument about a stone is really an argument about which of these three is being asked. They have three different answers and three different costs.

The question	Who can answer it	What it costs you
Is it a diamond at all, or a simulant such as cubic zirconia, moissanite, glass, white sapphire, zircon or topaz	A gemmologist with a dual tester and a microscope, in minutes	A short appointment
Is it natural or laboratory-grown	A gemmological laboratory only. Nobody at a counter, and no instrument you can buy	A laboratory submission and a wait
Is it the specific stone on the paperwork, at that weight and those grades	You, free, in about two minutes	Nothing

The third is the one almost nobody runs and the one that catches the most real problems. Most trouble is not a cubic zirconia sold as a diamond. It is a quote that says G VS1 while the report says H SI1, a report that belongs to a different stone, or a stone that never had one.

The home tests, ranked by what they are worth

Test	What it actually detects	What defeats it	Weight
Facet doubling under a 10x loupe	Double refraction. Diamond is singly refractive and shows crisp single facet junctions	Cubic zirconia, glass, spinel, YAG and GGG are also singly refractive. Lab-grown diamond shows nothing	The best home test there is. Doubling means it is not a diamond. No doubling means very little
Weigh and measure	Density. Diamond sits at a specific gravity of about 3.52, cubic zirconia at 5.56 to 6.00	Moissanite at 3.17 to 3.22, spinel and topaz are all close enough to diamond to pass	Genuinely diagnostic for cubic zirconia. Needs a carat scale, calipers and a loose stone
Dot or newspaper read-through	Refractive index. A diamond bends light so hard the mark underneath is unrecognisable	Moissanite blocks read-through completely and cubic zirconia mostly does. Only works on a loose, clean stone placed table-down	A disqualifier only. If you can read the letters, it is not a diamond. If you cannot, you have learned almost nothing
Handheld thermal “diamond tester”	Thermal conductivity	Reads diamond for moissanite as well as diamond, and reads diamond for every lab-grown stone. Skin oil, dirt, a warm stone or touching the metal all skew it	Rules out cubic zirconia and glass. Nothing else. Useful, and routinely oversold
Dual thermal and electrical tester	Heat conduction plus electrical conductivity, which separates moissanite	Rare type IIb diamonds are electrically conductive and can be flagged as moissanite. Newer low-conductivity moissanite is harder to read	The best instrument answer short of a laboratory. Not a home tool
Bubbles under the loupe	Gas bubbles, swirl lines and unmelted powder, which form in cubic zirconia and glass and never in diamond	Nothing, but their absence proves nothing either	A positive tell only. Find bubbles and you are done
UV or blacklight	Fluorescence	Only 25 to 35 percent of diamonds fluoresce at all, so no glow is meaningless. Cubic zirconia and moissanite both fluoresce under long-wave UV	Close to worthless as a real-or-fake test, despite how often it is listed
Fog or breath test	Thermal conductivity, measured badly	Moissanite clears as fast as diamond. No defined pass time exists, and humidity, room temperature, distance and skin oil all move the result	Close to worthless. The most repeated test on the internet
Water or float test	Density, very crudely	Cubic zirconia, moissanite, sapphire, zircon, topaz and spinel all sink. So does most glass	Worthless. It rules out a stone light enough to hesitate, and little else
Scratch test	Hardness	Moissanite at 9.25 and white sapphire at 9 scratch nearly everything diamond does, so a pass proves nothing	Do not do it. Destructive, and see below
Heat and shatter test	Thermal shock resistance	A surviving stone could be moissanite, sapphire, topaz, spinel, YAG or GGG	Do not do it. It destroys stones and settings

Blunt version: the three tests that dominate the search results for this question, the fog test, the water test and the scratch test, are the three least worth running. The fog test and the water test cannot separate a diamond from the two materials most likely to be impersonating one. The scratch test cannot either, and it costs you money to find that out.

On the scratch test specifically

You will find pages telling you to drag the stone across glass or a mirror, or to take sandpaper to it. Do not. Mohs hardness is an ordinal ranking, not a linear one, and the top of the scale is crowded: diamond at 10, moissanite at 9.25, sapphire at 9. All three mark glass, so a pass tells you almost nothing about which one you are holding.

The cost is real. Diamond is the hardest gem material but it is not indestructible, it has perfect cleavage in four directions, and it can chip. If the stone is a simulant you have abraded something you may still want to wear, and either way you can scratch the setting, which on an older piece is where a worn claw least needs extra force. The same goes for the heat-and-shatter test, which turns an unanswered question into a destroyed one.

What a handheld diamond tester actually detects

The pen-shaped tester sold online measures one property: how fast the stone pulls heat away from a warmed probe tip. Diamond does that extraordinarily well, so the tester beeps. Cubic zirconia is a thermal insulator, so it does not.

That is the whole capability. Read it as a cubic zirconia and glass detector, because that is what it is, and it is good at that job. What it cannot do is separate a diamond from moissanite, which is the separation that actually matters now, and it cannot separate natural from lab-grown either.

Dual testers add a second probe measuring electrical conductivity. Diamond is an electrical insulator; moissanite, a silicon carbide semiconductor, conducts. That works, with two edge cases. Rare type IIb diamonds contain boron and are electrically conductive, so a genuine blue diamond can be flagged as moissanite. And moissanite is now made with lower conductivity than the early material, narrowing the gap the instrument relies on. Treat any tester reading as a strong indication, not a verdict.

Why no test at home can tell natural from lab-grown

A laboratory-grown diamond is not an imitation. It is a diamond, grown by high pressure and high temperature or by chemical vapour deposition instead of in the earth, and GIA describes the result as having nearly the same chemical, optical and physical properties as a natural stone, adding that by eye the difference can be indistinguishable. What separates them is growth history, recorded in atomic-level defects that show up under laser-excited photoluminescence and deep-UV fluorescence imaging. That is laboratory equipment, not counter equipment.

There is one shortcut, and it is paperwork rather than a test. GIA laser-inscribes the girdle of every laboratory-grown diamond it grades with the report number and a statement identifying the stone as laboratory-grown, so reading a GIA inscription under 10x tells you which kind of report it belongs to. That tests the disclosure, not the stone, and the disclosure is what was actually at risk. The full comparison, including what each is worth on resale, is in natural versus lab grown diamonds in South Africa.

The one check worth running yourself

If you do nothing else on this page, do this. It is free, it takes about two minutes, and it answers the third question.

Get the report number in writing. From the seller before you pay, or from the insurance schedule and original paperwork on a piece you already own.
Verify it yourself, on your own device, at gia.edu/report-check. Do not let anyone type it for you and do not accept a printout as the check. Report Check covers GIA reports dated from 1 January 2000 onwards, so an older one has to go back to the laboratory instead.
Match every line. Carat weight, colour, clarity, cut grade where one is given, polish and symmetry, fluorescence, and the measurements in millimetres. The measurements line is the one people skip and the one that ties a report to a physical object. Compare it to the size guide and the grades to the colour and clarity charts.
Confirm the report says natural, in those words, if that is what you are buying.
Look at the girdle under 10x. Any inscription should read as the same report number. GIA inscriptions are invisible to the naked eye and readable under magnification.
Check the clarity plot against the stone where the report has one. The pattern of inclusions is effectively a fingerprint.

A mismatch is not automatically fraud. Reports get separated from stones over decades, pieces get reset, and centre stones get swapped during a repair without anyone recording it. But it means the paperwork no longer describes what you are holding, which is worth knowing before you insure it, sell it or hand it on. If a sale is coming, selling a diamond ring in South Africa covers what a buyer will ask for.

If there is no report at all

Inherited jewellery usually has no report, and a stone bought in the 1980s may never have had one. There are two routes.

For identity, a qualified gemmologist or a registered valuer with a dual thermal and electrical tester, a refractometer and a microscope can separate a diamond from a simulant quickly. Ask what instruments they will use before you commit, and ask for the finding in writing rather than as a verbal opinion. Diamond valuation in Johannesburg and jewellery valuation in Johannesburg cover what a valuation does and does not include, which is a separate thing from identification.

For origin, and for a report you can verify forever after, the stone has to go to a laboratory. GIA operates a laboratory in Bedfordview, Johannesburg, taking submissions by transport or walk-in, so a South African stone can be graded without leaving the country. Other laboratories including IGI grade to published standards as well. What you are buying is a number in a database that you and every future owner can check. I do not publish laboratory fees here, because they move with weight and service level and a current quote beats a stale figure on a web page.

Why you are asking this question in the first place

Here is the pattern I keep running into. People do not usually start wondering about a stone because it looks wrong. They start wondering because they cannot find anything in writing that says what it is.

That is the signal. A seller worth buying from writes the specification down before you pay: weight, colour, clarity, cut grade, laboratory and report number, natural or laboratory-grown, loose-stone price separated from the setting. A stone bought that way is never a mystery later, because the paperwork outlasts the salesperson, the shop and eventually you. A stone bought on a verbal assurance and a nice box produces exactly this question twenty years on, usually at a valuation, a divorce or an estate.

The cheap headline is the same story in advance. A one carat diamond ring advertised at a few thousand rand is almost never a certified single stone, and the gap is specification rather than a bargain: lower colour, lower clarity, often a cluster of small stones counted as total carat weight, frequently no report from a major laboratory. In my June 2026 count of 230 GIA-certified natural stones held by six South African sellers, about 9 percent of the retail-held stock reached D to G colour with VS2 clarity or better, against 52 percent of the stock held by sellers working with trade access.

So the useful move, before you buy and when you are working out what you already own, is to search the price rather than the trade. Almost nobody in South Africa searches for a diamond wholesaler or a dealer, and those results do not repay the effort. Ask what a carat costs: the diamond price per carat in South Africa, or what a one carat diamond actually goes for. A seller who publishes a per-carat figure has already agreed to be measured against it, and the same habit tends to produce sellers who put a report number on a quote unasked. Those sellers cluster at the manufacturer, cutting house and dealer level, which is where I would send a serious natural diamond purchase.

Cheaper. A retailer buys its certified stones at exactly that level and then roughly doubles the price, about three times at a branded showroom, while the steps between one trade level and the next run at 3 to 5 percent.
Better documented. That level is where the report number tends to arrive on the quote unasked, which is the whole subject of this page, and premium specification is about five times more common there.
They do the rest too. Setting, bespoke commissions, resizing, re-tipping, insurance valuations and buy-back, usually on their own bench. If a piece needs a stone identified or reset years later, that is the same bench, not a counter that would send the job out.

Buy off a counter when the deadline is fixed, when finished rings on a hand are how you decide, or when the brand is part of the purchase.

For scale on the number itself, derive it from the published list rather than trusting anybody’s band. Rapaport’s round brilliant list dated 20 March 2026 prices G/VS1 in the 1.00 to 1.49 carat bracket at $5,400 per carat, and that list is an asking price, so a discount has to be assumed. At 10 to 30 percent off list, R16.50 to the dollar and 15 percent VAT, a loose one carat of that specification estimates to roughly R72,000 to R92,000 including VAT, with about R82,000 at 20 percent off. It is an estimate on a published asking price and a stated assumption, not a measured market price, and every input moves.

The diamond buying checklist is the version to take to an appointment, and the GIA verification guide walks through each line of a report.

What I cannot tell you

I cannot tell you what your stone is from a photograph, and neither can anyone else. Colourless simulants photograph beautifully and the properties that separate them are not visible in an image.

I do not know how a given tester model behaves, because build quality varies enormously across the cheap end of that market and I have not benchmarked them individually. Treat every one of the weightings above as a statement about the physics being measured, not about a particular device.

And I cannot give you a home method for natural versus lab-grown, because none exists. If you find a page that offers one, that is the page to stop reading.

Sources and references

GIA on identifying laboratory-grown diamonds, for the statement that traditional gemological observations and instrumentation cannot separate them: gia.edu, plus the fuller treatment of simulants, moissanite and lab-grown diamonds including girdle inscription practice: 4cs.gia.edu.
Nassau, McClure, Elen and Shigley, Synthetic Moissanite: A New Diamond Substitute, Gems & Gemology, Winter 1997, for moissanite’s refractive indices of 2.648 and 2.691, dispersion of 0.104, hardness of 9.25, specific gravity of 3.22, and the finding that thermal probes react to it as though it were diamond: gia.edu.
GIA on diamond fluorescence, for the figures that only about 25 to 35 percent of diamonds fluoresce under long-wave UV and that more than 95 percent of those fluoresce blue: 4cs.gia.edu.
International Gem Society, properties of diamond simulants, for the refractive index, birefringence, specific gravity, hardness, read-through, electrical and thermal test values in the table above: gemsociety.org, alongside its test-by-test review of at-home methods: gemsociety.org.
IGI on diamonds and simulants, for the position that no reliable do-it-yourself test reveals whether a stone is diamond or another material: igi.org.
GIA Report Check and its coverage dates: gia.edu/report-check. GIA laboratory locations, including Bedfordview, Johannesburg: gia.edu/locations.
My own 2026 South African research. A June 2026 inventory listing of 230 GIA-certified natural diamonds held by six sellers, and a July 2026 pricing study of 73 rows, 71 of them carrying a price, across 18 sellers, 17 of them South African and one an international online retailer, published as bands rather than individual stones. Method in the price index.
Rapaport round brilliant price list, 20 March 2026, G/VS1, 1.00 to 1.49 carat bracket, at $5,400 per carat, used with a stated discount assumption at R16.50 to the dollar and 15 percent VAT.

Gem properties are stable. Instruments, laboratory services and prices are not, so verify anything time-sensitive on the day. For how I research and correct these pages, see the editorial policy.

How to Tell If a Diamond Is Real (2026): What the Home Tests Actually Prove