by Jacida Sun

A cutter’s choice of diamond abrasive decides how much rough survives and how good the final surface can be. Research from gem laboratories and optical engineering explains why, and turns several workshop habits into measurable rules.

Introduction

In 2015, Rio Tinto’s Argyle mine in Western Australia recovered a 9.17-carat violet rough. Argyle master polisher Richard How Kim Kam turned it into a 2.83-carat oval that GIA graded Fancy Deep Greyish Bluish Violet. The polished stone kept 30.9% of the rough’s weight, and trade reports put the cutting time at more than 80 hours.

Each of those hours was a decision about where to remove material and where to leave it. Diamond is the extreme case, because only diamond can abrade it. The same question governs the hardest colored gems: on ruby, sapphire or chrysoberyl, the abrasive decides how much rough survives, how flat a facet stays, and how fine the final surface can be.

This article examines that question with published data rather than workshop lore. Its central argument is that the most important property of a diamond abrasive is not its average particle size but its largest particles.

Hardness is not one number

The Mohs scale ranks minerals by which one scratches which. It does not say how much harder one is than another. Indentation tests do: with a Knoop indenter, Winchell measured corundum at about 2,000 and diamond at about 8,000. The single Mohs step between them hides a roughly fourfold gap, which is why diamond is the practical abrasive for corundum.

Hardness is also only part of the problem. At the wheel, cleavage, grain structure and crystal direction often matter more.

Corundum shows that even a stone without cleavage has fast and slow directions. In a chemical mechanical polishing study of four sapphire planes, Cao and colleagues measured 5.93 nm/min on the C-plane (the basal plane) and 2.47 nm/min on the R-plane: a 2.4-fold difference under identical conditions. An earlier study at Rutgers University also found removal fastest on C-plane sapphire.

Both studies used industrial wafer processes with silica or alumina slurries, not diamond on a faceting lap, so their ranking of planes should not be carried straight to the bench. The general lesson does carry over: the same abrasive can cut one facet more than twice as fast as its neighbour.

What a micron grade actually specifies

Diamond abrasives are graded in two regimes. Coarse grits are sized with sieves under ANSI B74.16 and ISO 6106, and the finest sieve grade, 325/400 mesh, corresponds to about 44–37 µm. Below that, sieves stop working. Finer “micron” powders are graded under ANSI B74.20, by methods other than sieving.

A micron grade is therefore a distribution, not a size. ANSI B74.20 defines it with several limits. For a 2–4 µm powder they are:

• D5 at least 2.0 µm (95% of particles are coarser)
• D50 of 3.0 ± 0.3 µm (the median)
• D95 at most 4.0 µm (95% of particles are finer)
• D99.9 at most 9.0 µm (the upper limit for the largest particles)

The last line deserves attention. A powder sold as “3 micron” can, within the standard, contain particles three times that size. The next section explains why those rare particles matter more than the median.

Lapidary “grit” numbers above about 400 are not sieve sizes. One supplier’s chart, for instance, equates “3000” with 4–8 µm and “8000” with 2–4 µm. Because such labels are conventions rather than standards, the micron range is the figure to compare between products.

Why the largest particle matters most

Every grain that cuts a brittle material also cracks it. Beneath the visible scratches lies a layer of microcracks known as subsurface damage, and optical engineers treat its removal during polishing as essential.

They have also measured how deep it goes. Reviewing many glasses and grinding processes, Lambropoulos proposed that the damage depth, in micrometers, lies between 0.3d^0.68 and 2d^0.85, where d is the abrasive size in micrometers. Applied to common lapidary sizes:

These bounds were derived for optical glass, not corundum, so they indicate scale rather than predict a gemstone’s behaviour. They still make one point plainly: the upper bound is larger than the particle itself.

Researchers at Lawrence Livermore National Laboratory then asked which grains cause the deepest cracks. Measuring damage in ground fused silica, Suratwala and colleagues found that only a small fraction of grains carry enough load to fracture the surface, and that these are the larger ones. The grains that effectively cause fracture were about ten times the mean particle size. In a related test, a 9 µm slurry contaminated with just 0.1 wt% of 15 µm particles cracked deeper than the clean 9 µm slurry, and deeper even than a slurry made entirely of 15 µm particles.

Coarse grains also remove material faster: in brittle lapping, removal rate rises linearly with abrasive size. Speed and damage are two faces of the same grain.

For the lapidary, the implication is direct. A few stray coarse grains, carried on a dop, a finger or a poorly cleaned lap, can undo an entire fine stage. Controlling the tail of the size distribution, through good powder and clean practice, matters more than shaving the median.

Monocrystalline or polycrystalline?

Micron diamond is sold in two main structures: monocrystalline particles, each a single crystal, and polycrystalline particles built from many smaller crystallites. Direct comparisons on corundum are scarce, but one controlled test is public.

In a lapping trial reported in US Patent 8,182,562, 2-inch C-plane sapphire wafers were lapped on a composite copper/resin plate with 4–8 µm diamond at 10 carats per 500 ml of slurry. The results:

Polycrystalline diamond removed about one-third more material and left a slightly smoother surface. These figures need careful reading. They come from a single test, they were not peer reviewed, and the patent’s purpose was to promote a third, modified diamond that beat both. They are best read as a reason to trial polycrystalline powder at the pre-polish stage, not as a settled ranking.

Powders, pastes and lapping film

The same diamond can reach the stone in three forms. Each one handles the rogue-particle problem differently.

Commercial lapping film consists of precisely graded abrasive particles coated on a polyester film 3 mil (about 76 µm) thick. In a study of film-based laps published in CIRP Annals, Evans and colleagues showed that the abrasive film laps or polishes the work without altering the geometry of the plate beneath it, and that it can operate in either brittle or ductile two-body mode. Two consequences follow for gem cutting. The flatness of a facet is set by the plate, not by how evenly a lap was charged. And because the grains are bonded rather than loose, they cannot wander across the working surface the way loose grains can.

Film has limits. It wears and must be replaced, and on a flat plate it suits flat facets and tables rather than curved cabochon domes. For readers who want a closer look at the construction, this overview of diamond lapping film explains how graded diamond is held in a resin binder on a PET backing, and how diamond film compares with conventional abrasive films.

A sequence built on evidence

The findings above reduce to one planning rule: each stage must remove at least the damage left by the stage before it. Taking the upper bounds from the damage table as a conservative guide, a sequence for faceted corundum could run as follows. It illustrates the logic; it is not a tested recipe, and the glass-derived depths are an assumption for corundum.

1. Cut at 15 µm. Stop each facet slightly short of its final meets, because damage beneath may reach about 20 µm.
2. Pre-polish at 6 µm. Remove more than those 20 µm before changing grade. This stage leaves up to about 9 µm of its own.
3. Polish at 3 µm, then finish at 1 µm or finer. Each step clears the layer left by the one before: about 9 µm, then about 5 µm.

Why not jump from 15 µm straight to 1 µm? In brittle lapping, removal rate scales linearly with abrasive size. All else being equal, a 1 µm grade would clear the 20 µm damage layer about fifteen times more slowly than a 15 µm grade removes stock. Intermediate grades are not a luxury; they are the fastest route to a flawless surface.

Phenomenal stones add a constraint before any abrasive touches them. In star corundum, the three series of needle inclusions that create the star lie in planes parallel to the basal face, that is, perpendicular to the c-axis. To centre the star, the cutter therefore places the apex of the dome on the c-axis. No later polishing stage can correct an error made here.

Treated stones change the rules

A filled gemstone is two materials polished at once, and the filler rarely behaves like its host.

In lead-glass-filled ruby, GIA researchers found the glass to be significantly softer than corundum, so differences in polish quality can reveal filled fractures. The glass began to sweat out of fractures at 700 °C, with a melting point between 600 and 700 °C, and caustic pickling solution etched it. Filing and polishing during setting tests caused no damage to the filler, although only a few stones were tested.

Filled emerald is more delicate. In a GIA study of 33 filled emeralds, ultrasonic cleaning or soaking in solvents changed the appearance of about 29. Cedarwood oil and unhardened Opticon changed obviously in the ultrasonic bath, while surface-hardened and hard fillers proved the most durable.

This creates a real conflict at the bench. The thorough cleaning that protects a fine stage from rogue particles is exactly what can strip or cloud a filler. For treated stones, gentle manual cleaning between grades is the safer compromise.

Conclusion

The Argyle Violet lost more than two-thirds of its weight on the way to becoming a 2.83-carat gem. On hard colored stones, the margin between a good yield and a poor one is set less by the diamond itself than by how precisely it is graded and how cleanly each stage hands over to the next.

The evidence points to three habits. Read the micron distribution, and especially its upper limit, rather than the grit label. Plan each stage to remove the damage of the one before. Treat a single stray coarse grain as the most likely cause of a ruined polish, while remembering that treated stones may not tolerate the cleaning that prevents it.

About the Author:

Jacida Sun is a Sales Engineer at Grish with five years of experience in the precision polishing industry. Drawing on knowledge of Grish’s polishing materials, Jacida provides product-related technical support on abrasive selection, material performance, and application requirements. This work includes helping international customers identify suitable materials and tailored polishing solutions for a range of precision finishing application.