How we measure

Diamond Testing Bench

The Diamond Testing Bench puts three round brilliants side by side: a natural diamond, a laboratory-grown CVD diamond and a laboratory-grown HPHT diamond. Choose a test, and each stone reacts as a typical stone of its kind does. The bench is a teaching model of three typical stones. It does not test, identify or grade a real stone. Only a gemological laboratory can find where a diamond grew.

The three stones

All three are the T73 Round Brilliant, traced on its own 3D model: the geometry we sell, with the planes that the Hearts & Arrows Scope uses. The three have one cut and one material, diamond. So each difference on the screen comes from the test, not from the cut.

Natural
Type Ia: nitrogen in clusters. In GIA’s study of its D-to-Z diamonds, 99.06% were this type (Fall 2020).
CVD
Type IIa, annealed at high pressure and high temperature after growth to remove a brown tint. About 80% of the CVD diamonds sent to GIA since 2020 had this treatment (Summer 2024).
HPHT
Weak type IIb, as grown: a trace of boron, and one small metal inclusion. More than 80% of the D-to-Z HPHT diamonds sent to GIA since 2020 contain boron that infrared can detect (Summer 2024).

Each stone shows the typical reaction. Each test also has exceptions, and the note under the test names the most important one. About 1% of natural diamonds are type II. In several tests they behave more like the laboratory-grown stones. For this reason a type II result means “test further”.

The tests

Loupe
Daylight, enlarged. Natural: a small crystal and a feather. CVD: dark pinpoints of non-diamond carbon. HPHT: a rod of metal flux. Similar marks occur in natural and laboratory-grown stones, so inclusions are only hints.
Thermal pen
All diamond conducts heat much better than glass or cubic zirconia, so all three read as diamond. Synthetic moissanite also reads as diamond on a thermal probe (GIA, Winter 1997).
UV torch, 365 nm
In a sample of 26,010 GIA reports on colorless to faint yellow diamonds, 35% fluoresced, and 97% of those with medium to very strong fluorescence were blue (Moses and others, 1997). Only 7% of near-colorless CVD stones and 2% of colorless HPHT stones showed a long-wave reaction.
Magnet
Metal inclusions can pull an HPHT stone to a strong magnet. In two GIA studies, HPHT stones of VS2 clarity or lower were weakly magnetic, and the pull was consistent only at SI and I clarity. No pull does not exclude HPHT growth.
Inscription
A laser inscription is a claim that you can compare with its report. It is not a test. GIA reported forged inscriptions on laboratory-grown diamonds that were cut to nearly match natural diamonds with a report (January 2024). Small stones often have no inscription.
Short-wave UV, 254 nm
HPHT stones usually react more to short-wave than to long-wave UV, the opposite of most natural diamonds; 88% of colorless HPHT stones reacted. With the lamp off, the HPHT afterglow lasts from less than a second to a few tens of seconds. Low-dose irradiation can reduce or remove it. A few natural type IIb diamonds and some CVD diamonds show a similar afterglow.
Deep-UV image
Diamond absorbs light below 225 nm at its surface, so the image shows the growth structure: octahedral bands in most natural diamonds, growth layers in CVD, cuboctahedral sectors in HPHT. Some recent CVD stones look much like natural type II diamonds in this image.
Crossed polarizers
HPHT stones show very weak strain or almost none, except around inclusions and cracks. Strain patterns in CVD diamonds are hard to tell from those in natural diamonds.
Infrared
Natural type Ia: the A and B nitrogen aggregates at 1282 and 1175 cm⁻¹, the platelet peak near 1360 to 1370, and hydrogen at 3107. CVD: no nitrogen. HPHT: boron at 2800. The bands from about 1700 to 2700 cm⁻¹ are the same in all diamond.
Photoluminescence
At liquid-nitrogen temperature (−196 °C). Natural: N3 at 415.2 nm. Annealed CVD: the silicon-vacancy pair at 736.6 and 736.9 nm, H3 at 503.2, weak NV lines at 575 and 637, and a trace of 415, which annealing can make. HPHT: few lines or none. GIA found the nickel pair near 883 and 884 nm in 30% of colorless HPHT stones, and no lines at all in 37% of type IIb HPHT stones. Silicon also occurs in a few natural and HPHT diamonds, and its average level in CVD fell about a hundredfold in ten years (GIA, 2024).
Screening device
Pass verifies natural. Refer means test further. In the independent ASSURE 2.0 reports that we read, devices referred from 0.2% to 13.6% of natural diamonds. No device passed a laboratory-grown stone or a simulant as natural on loose stones. One device did so for 1.9% of mounted stones (2026).

How we trace the views

The eye looks straight down the axis of the stone, with parallel rays. Each ray refracts in and reflects inside. Where a ray meets a facet below the critical angle, part of the light leaves (unpolarized Fresnel) and the rest goes on, for a maximum of 20 bounces. Daylight is a studio of soft boxes, with the head of the viewer dark overhead. We trace it one time for each color channel, each with its own refractive index (2.4135, 2.418, 2.4265: the dispersion of diamond at half strength). Under a UV lamp, the glow adds up along each path inside the stone. The deep-UV image adds glow only where a ray meets the surface, with the color of the growth structure at that point. Between crossed polarizers, the strain along each path adds up to a retardation, drawn as its interference color.

What the bench does not model

The growth patterns, strain fields, glow colors and spectra are schematic. We shaped them after the published descriptions and put each line at its published position. We did not measure them on real stones. The spectra are normalized and drawn on a true axis. There is one typical stone of each kind. The bench does not model fancy colors, melee, mounted stones, treated natural diamonds, hybrids or simulants.

Sources

We are not affiliated with these laboratories, journals or programs.