By Jay Shah, Technical Team, Beta Diamond Products
Hardness testing measures the surface you give it, and if that surface carries deformation from grinding, heat from sectioning, or rounding from polishing, the number on the readout describes your preparation, not your material. Microindentation tests, Vickers and Knoop, need a polished surface, 1 micron diamond or finer for low loads, because the indent is measured optically and its diagonals are only as crisp as the finish around them. Macro tests like Rockwell tolerate a ground surface, 400 to 600 grit, but tolerate nothing else on this list: the faces must be flat, parallel, clean, and thermally undamaged.
We supply quite a few Fortune 500 and government labs whose hardness numbers end up in certification paperwork, and the prep standard that work demands is not exotic. It is the ordinary sequence done without shortcuts, which is exactly what this article describes.
The two ways preparation corrupts the measurement
Work hardening from your own abrasives. Every grinding and polishing step deforms a layer beneath the surface, and deformed metal is harder than the same metal undisturbed. A microhardness indent at 10 or 25 grams penetrates only a few microns, which can sit entirely inside that work hardened skin. The reading comes back high, repeatably high, and looks perfectly trustworthy. The lighter the load, the worse the bias, which is why a case depth traverse on a poorly polished section drifts in ways nobody can explain. The cure is a complete polishing sequence that removes each step's damage rather than burying it, finishing at 1 micron minimum for microindentation, and for loads under about 50 grams, a final colloidal silica or fine alumina step to strip the last deformed material.
Heat from sectioning. An aggressive dry or coolant starved cut tempers, or in some steels rehardens, a zone extending well below the cut face. If your indents land in that zone, no amount of beautiful polishing rescues the number, because the material itself was altered. Section under flood coolant, feed gently, and when the measurement is critical, remove generous material in grinding before trusting the surface.
Geometry: the silent requirement
Hardness machines assume the test surface is perpendicular to the indenter and fully supported underneath. A mount with non parallel faces tilts the sample; a tilted Vickers indent comes out asymmetric, one diagonal long, one short, and the averaged number is wrong in a way that is hard to spot. Grind the back of the mount flat and parallel to the test face, and seat it on the anvil without debris underneath. For Knoop traverses near an edge, case depth, coating hardness, decarburization, edge retention becomes part of the metrology: mount in filled epoxy, polish on napless cloths, and keep the edge in the indent's focal plane.
Test standards add spacing rules, indents several diagonal widths apart and from edges, and only a flat well retained surface lets you place a tight traverse legally.
A prep recipe for microhardness work
Section wet and gently. Mount in filled epoxy if edges matter. Grind 240, 400, 600 grit SiC, one minute each, rotating 90 degrees, moderate pressure. Polish 9, 3, 1 micron diamond, adding 6 micron for hard alloys, cleaning between steps. For loads of 50 grams and below, or any case depth or thin coating work, add 60 to 90 seconds of colloidal silica. Do not etch before testing if you can avoid it; etch after, to place the indents in context, since etching relief degrades diagonal measurement.
A self check worth running once in any lab: prepare one sample twice, your routine method and the full sequence above, and run a 25 gram Vickers traverse on each. If the numbers differ beyond scatter, your routine method has been quietly biasing results, and now you know by how much.
A consumables note for labs that mix their own
One thing that surprises labs coming from the big consumable brands: we still stock loose diamond powder, both monocrystalline and polycrystalline, in the full micron range. The major brands have largely moved away from selling powder, preferring to sell the finished suspension, but labs that blend their own slurries, charge their own laps, or run legacy hardness prep procedures written around powders need it, and they currently have few places left to buy it. It is on our shelf, along with everything else in the recipe above, free samples included. Hardness numbers go into reports people act on. The prep deserves the same seriousness as the calibration sticker on the tester.