技術資料195 views

Designing the Grinding-Polishing Sequence: Parameter Chain and Common Defects

Each step must remove the previous deformed layer while leaving a shallower new one; skipping steps yields artifact microstructures.

华旋光电技术部
Designing the Grinding-Polishing Sequence: Parameter Chain and Common Defects

Deformed-layer handover

The core logic of metallographic preparation: every step removes material while introducing a new, shallower deformed layer, and as long as the next step removes more than that depth, the final surface shows the true structure. Deformed depth runs about half to one abrasive size: P320 paper (46 μm grains) leaves 20-40 μm, P1200 about 10 μm, 9 μm diamond 2-3 μm, 1 μm about 0.5 μm, and final polish with 0.05 μm alumina or silica under 0.1 μm. A typical steel sequence: P320→P600→P1200→P2500→9 μm→3 μm→1 μm→final.

Parameter chain

Time each step to erase the previous scratches: once the scratch pattern is uniform in one direction, extend by 30 percent. Match speed and pressure: rough grinding at 150-300 rpm with 20-30 N per specimen; fine polishing at 100-150 rpm with 10-15 N. For soft metals like aluminium and copper, ultrasonically clean in alcohol one minute between steps against abrasive embedment, and rotate 90 degrees at each paper change so leftover scratches are obvious.

Defect quick reference

DefectCauseFix
Comet tailsDislodged grains or phase pull-out draggingFresh cloth, lower pressure
Edge roundingMounting gap, excess pressureLow-shrinkage mount, less pressure
ReliefDifferential removal of hard/soft phases, over-polishingShorten final polishing
Phantom scratchesSkipped steps, contaminated clothRepeat from the previous step

Electropolishing supplement

For materials that deform readily — austenitic stainless, pure aluminium — electropolishing replaces the final step: the specimen as anode in an electrolyte such as 10 percent perchloric acid in alcohol, at 20-40 V and below 30 °C for 30-90 s, yields a deformation-free surface. Note that electropolishing attacks inclusions and second-phase interfaces preferentially, so cross-check inclusion assessments against mechanically polished results.

⇄0比較仕様比較