How to tell if a diamond is real

Three different questions hide inside “is it real”

A stone can fail to be real in three unrelated ways, and a test that answers one of them says nothing about the other two.

• A simulant is a different material cut to imitate diamond: cubic zirconia, moissanite, white sapphire, white topaz or plain glass. Almost every test described below is aimed at this question.
• A lab-grown diamond is diamond. Same carbon, same hardness, same refractive index of 2.42, same behaviour in every home test. It is not an imitation, and it is worth a fraction of an equivalent mined stone.
• A treated stone is natural diamond whose colour or clarity has been improved by fracture filling, laser drilling or irradiation. Real, and worth considerably less than an untreated equal.

Decide which question you are asking before you start, because the last two cannot be answered at a kitchen table by anyone.

The fog test, and its limits

Breathe on the stone from close range, as you would on a cold window. Diamond conducts heat better than almost any material on earth – roughly 2,000 watts per metre-kelvin, against about 2 for cubic zirconia – so it draws the warmth out of the condensation and clears in a second or less. Zirconia holds the mist for two or three seconds, and glass longer still.

The caveats matter more than the test. The stone has to be clean: a film of hand cream, soap or skin oil holds fog on anything. A ring worn all afternoon is already at skin temperature and will shed fog quickly whatever it is made of. And moissanite conducts heat well enough to clear almost as fast as diamond, so a pass tells you “not glass, probably not zirconia”, never “diamond”.

Why the newspaper test fails on a mounted stone

The read-through test works like this: place a loose stone table-down on a line of newsprint and try to read the letters through it. A well-cut round brilliant bends and returns light so completely that you see no text and no sharp line. Glass and shallow, poorly cut imitations often let the print through.

It only works on a loose stone with a flat table that can sit flush on the page. Once a stone is set you cannot perform it at all: the mount holds the pavilion off the paper, the setting lets light in from the side, and prongs sit in the way. It also passes well-cut zirconia and moissanite, which scramble text as thoroughly as diamond does, and it can wrongly condemn a genuine but shallow old-cut stone. Treat a failed read-through on a loose stone as a reason to look further, and treat the test as unavailable on anything mounted.

What a 10x loupe shows that the eye does not

A loupe costs less than a takeaway and answers more than every other home test combined. Hold it close to your eye, bring the stone to the loupe, and use a strong light from the side.

• Facet junctions. Diamond is 10 on the Mohs scale and its facet edges stay razor-sharp for decades. Zirconia and glass abrade; under magnification their junctions look softened and rounded, and the table shows fine scratching.
• Doubling. Moissanite is doubly refractive. Look through the crown at an angle rather than straight down and the back facet edges appear as two lines slightly offset from one another. Diamond is singly refractive and never doubles.
• Inclusions. Small crystals, feathers and pinpoints are evidence of a natural stone. Perfectly round gas bubbles are evidence of glass or paste.
• A girdle inscription. Many graded stones carry a report number laser-inscribed on the girdle, and lab-grown stones are frequently inscribed LG or LAB GROWN. It is tiny; 10x with a bright light may show it, 20x is better.

Fire, weight and the tester on the counter

Under a single point of light, diamond returns mostly white flashes with a restrained amount of colour – its dispersion is 0.044. Moissanite disperses 0.104, more than twice as much, and throws obvious rainbows that people describe as disco-ball fire. Zirconia sits between the two and reads as glassy rather than crisp. This is a trained comparison, not a verdict, and it is unreliable on stones under about half a carat.

Weight is more decisive on a loose stone. Zirconia is far denser than diamond, roughly 5.7 against 3.52, so a zirconia cut to the millimetre dimensions of a one-carat diamond weighs close to 1.7 carats. Any jeweller with a gauge and a carat scale settles that in a minute.

The handheld probe on a jeweller’s counter measures thermal conductivity, which is why older probes call moissanite a diamond. Dual testers add an electrical-conductivity channel that flags moissanite, and even those misread the rare boron-bearing type IIb diamonds, which conduct electricity, as moissanite. Ask which kind of tester is being used.

The mount is evidence too

Nobody sets a two-carat mined diamond in base metal. Look inside the band for a fineness mark: 750, 585 or 375 for gold, 950 for platinum, 925 for sterling silver. A 925 stamp beneath a large white stone is a strong hint that the stone is zirconia, since sterling is the standard mount for costume and fashion stones. Our guide to reading gold hallmarks covers what each mark means.

Then look at wear. A green or grey line where plating has broken at the shank, a yellow tone showing through white metal at the shoulders, thin or splayed prongs, glue anywhere near a stone – each is an observation worth more than a hunch about sparkle.

What a photograph can and cannot settle

A photograph carries more than people expect: the colour and wear of the metal, the outline and proportions of the cut, how the stone handles light, how it is held, and, in a macro shot of the inside of the band, the hallmark itself. That is enough to reason towards likely genuine, inconclusive or likely simulant with the observations set out on both sides. It is what Carat does from a photograph, and every carat, cut, colour and clarity figure it returns is labelled an estimate rather than a grade.

What a photograph cannot do: measure thermal or electrical conductivity, weigh the stone, show doubling reliably at phone resolution, or separate lab-grown from mined diamond. That last one is not a limit of photography. It is a limit of everything except laboratory instruments.

What would settle each question

• Simulant or diamond: ten minutes at a jeweller with a dual thermal and electrical tester, a loupe and a scale. Usually free while you wait.
• Lab-grown or mined: a report from a grading laboratory such as GIA or IGI. This is spectroscopy, not a counter test, and no shopfront device does it.
• What it would grade: the same laboratory, working on an unmounted stone. Weight, cut, colour and clarity measured rather than estimated.
• Treated or untreated: the laboratory again. Fracture filling shows a flash effect under magnification that a gemmologist knows to look for.

Three things not to do: no scratch test, because zirconia, moissanite and quartz all scratch glass and a diamond can chip against a hard edge; no flame; no household bleach, which attacks alloyed metal and any glue in the setting.