By Sam Desai, Technical Team, Beta Diamond Products
A scratch that survives final polishing was almost never made by your final abrasive. It was made by a coarser particle that hitched a ride: contamination carried over from an earlier step, an embedded grit fragment, debris in a dirty cloth, or crystallized colloidal silica. To remove the scratch, go back one full step and repolish. To stop it from returning, find the contamination source, because repolishing without fixing the cause just schedules the same scratch for tomorrow.
First, read the scratch
Scratch width tells you the particle size that made it. A fine uniform haze of shallow scratches is normal at every stage except the last; that is the abrasive doing its job. The problem scratch is the lone deep gouge crossing an otherwise clean field. Its width corresponds to a particle many times larger than your current abrasive, which means that particle does not belong on that cloth.
Direction helps too. Scratches all running one way suggest the sample was held in one orientation too long or dragged during transfer. Random deep scratches on multiple samples from the same wheel point straight at the cloth.
The usual suspects
Carryover between steps. The classic. A 9 micron particle rides the sample, the holder, or your fingers onto the 1 micron cloth, and from then on that cloth scratches everything. Prevention is a real cleaning protocol: rinse the sample and mount under running water between every step, ultrasonic clean before the final stages, and wash your hands. Hands carry grit better than samples do.
A contaminated or worn cloth. Cloths trap debris in the nap and there is no reliable way to fully decontaminate one. If a thorough flush does not cure a scratching cloth within one sample, retire it. A cloth costs a few dollars. The samples it ruins cost more. And never share a cloth between micron sizes. Not once, not in a hurry.
Embedded abrasive breaking loose. Soft metals like aluminum and copper swallow SiC particles during grinding, then release them during polishing. Scratches that appear only on soft materials point here. Lighter grinding pressure and fresh papers help.
Crystallized colloidal silica. Silica that dries on a cloth or around a bottle cap turns crystalline and hard, and crystalline silica scratches. Give the final cloth a ten second water flush at the end of every run.
The two sources nobody suspects
When a lab has checked everything above and scratches persist, we point at two items that sound absurd until you check the chemistry.
Paper towels. A paper towel feels soft against skin, but under magnification it is a dense web of stiff, jagged wood pulp fibers, and those fibers are hard enough to abrade a polished or coated surface. This is exactly why the optics industry warns against paper towels on lenses: even a single dry wipe can leave micro scratches in a delicate coating, and a freshly polished metallographic sample is more delicate still. The towel also traps whatever grit is already on the bench and drags it across the finish. Dry samples with clean compressed air, a fresh lint free wipe, or an alcohol rinse followed by warm air. Never the roll by the sink.
Tap water. Depending on your building's plumbing, tap water can carry fine sediment, rust particles from old pipes, and mineral grit. If your final rinse deposits particles on the cloth or the sample, your rinse is a contamination step wearing a hygiene costume. For final stage rinsing, distilled or filtered water is cheap insurance, and labs on older plumbing notice the difference quickly.
The recovery procedure
Do not chase a deep scratch with your finest abrasive. The math is against you: a 0.05 micron polish removing a 5 micron deep scratch takes roughly forever and rounds the surface doing it. Step back to the abrasive one stage coarser than the scratch depth, repolish until the scratch is gone plus thirty seconds for the subsurface damage, then walk back down the sequence on clean cloths.
If scratches persist after a full restart on fresh cloths with fresh suspension, the contamination is upstream: rinse water, lubricant bottle, sample holder, or the towel you dry mounts with. Work backward until you find it.
Clean consumables are the cheapest insurance in this process. Our suspensions are filtered and tightly classified for exactly this reason, and if you suspect your current abrasive itself, we will send a sample bottle to test against it. Free, shipped fast.