Skip to content

Comparative analysis of metallographic preparation techniques and optimization of electropolishing in AISI 420 steel

FieldValue
JournalRevista de Metalurgia
Published2026-09-14
AuthorsJoan David Gutiérrez Castillo, I. Espinosa, Crystopher Paipay, Sergi Menargues, Javier Antonio Navas
DOI10.3989/revmetalm.e302.1783

The selected Test 5 protocol gave the clearest combined view of martensite, prior austenite grain boundaries, and chromium carbides. It used SiC through P4000, 3 μm polycrystalline diamond, 0.2 μm alumina, then 10% oxalic acid at 10 V in two 60 s cycles. The authors reported continuous boundaries without pitting or staining. Visual quality scored 5/5, and the full preparation took 30 min at €18.43 per sample.

Selected protocol and measured performance

Section titled “Selected protocol and measured performance”

Test 5 separated mechanical preparation from controlled anodic dissolution. The 3 μm diamond step ran for 5 min at 20 N and 150 RPM on MD-Chem cloth. The final 0.2 μm alumina step ran for 5 min at 15 N and 150 RPM. The cell used a titanium cathode and 10% aqueous oxalic acid at about 22 °C without external cooling. Each cycle lasted 60 s at 10 V, giving 120 s cumulative exposure in the same bath. The authors state that a pause between cycles limited local electrolyte heating. At 3 μm, diamond polished the martensitic matrix and carbides into one focal plane. The selected result combined that flat surface with continuous boundaries and dispersed carbide particles.

The study examined AISI 420 (X46Cr13) martensitic stainless steel. Its reported composition was approximately 85 wt% Fe, 0.44 C, 13.7 Cr, 0.86 Mn, 0.29 Si, 0.029 P, and 0.005 S. Heat treatment used austenitization at 1050 °C for 30 min, oil quenching at 25 °C, and tempering at 275 °C for 180 min. Precision cutting produced approximately 1 cm² specimens, 10 × 10 mm, with the cut plane parallel to the rolling direction. Automatic polishing controlled force and rotation speed. The general sequence used SiC paper, MD-DAC or MD-Chem cloth, 3 μm diamond suspension, and 0.2 μm alumina. The comparator was Kalling II, containing 5 g CuCl2, 100 mL HCl, and 100 mL ethanol. A Taylor Surf Plus contact profilometer measured average roughness, Ra, after each route.

Five routes changed the abrasive sequence and electrical conditions. Total time covered grinding, polishing, and etching or electropolishing.

TestAbrasive preparationAttack or electropolishingRa range (μm)Cost and total time
1SiC to P4000 + 1 μm aluminaKalling II, 20 s0.012 to 0.019€8.23; 95 min
2SiC P220 + 9, 3, and 1 μm diamond10% oxalic acid, 15 V, 60 s0.048 to 0.059€21.09; 35 min
3SiC P180 to P4000 + 3 μm diamond + 0.2 μm alumina10% oxalic acid, 6 V, 60 s0.021 to 0.026€18.43; 23 min
4SiC P180 to P4000 + 1 and 0.2 μm alumina10% oxalic acid, 10 V, 60 s0.025 to 0.035€14.19; 30 min
5SiC P180 to P4000 + 3 μm diamond + 0.2 μm alumina10% oxalic acid, 10 V, two 60 s cycles0.036 to 0.048€18.43; 30 min

The chemical route was the least expensive, but it required 95 min and received a visual quality score of 2. Test 2 used the longest diamond sequence and cost €21.09 per sample. Tests 3, 4, and 5 each received an overall score of 12. Test 5 was selected for its simultaneous grain-boundary and carbide definition.

The Kalling II route partly broke the Cr2O3-rich passive film and exposed acicular martensite. It also left granular contamination, darkened the matrix, masked grain boundaries, and made carbides difficult to identify. At 15 V, Test 2 produced dark, continuous boundaries, but dissolution thickened them and reduced morphological fidelity. Its roughness was 0.048 to 0.059 μm, the highest electropolished range. At 6 V, Test 3 yielded a specular surface without pitting, but the potential barely overcame passivity. Prior austenite boundaries therefore remained faint and slow to analyze. Test 4 raised contrast at 10 V, but alumina-only polishing produced surface relief, less-defined edges, and carbide pull-out. Diamond in Test 5 restored planarity and kept the matrix and carbides in one focal plane. Two 10 V cycles then gave the most balanced contrast in the series. The images showed a bright martensitic matrix, continuous prior austenite boundaries, and fine dark carbides within grains and along boundaries. The authors attribute the fine dark particles to chromium-rich carbides, including M23C6.

The paper’s own results make the tradeoff clear. Kalling II produced the lowest reported Ra, 0.012 to 0.019 μm, but the image was deficient in phase and boundary definition. The discussion reports 0.012 to 0.014 μm for that route, which conflicts with the earlier test table. The 6 V route gave 0.021 to 0.026 μm, yet activation was insufficient. A single 60 s cycle at 10 V gave 0.025 to 0.035 μm but retained pull-out from alumina-only preparation. Test 5 gave 0.036 to 0.048 μm, summarized as about 0.040 μm in the discussion. These results support the authors’ point that minimum Ra does not identify the best metallographic preparation. Ben Lenda et al. (2020) used equal parts of 2% nital and 4% picral to reveal cementite carbides and boundaries. The paper reports that grain boundaries remained not clearly defined with that etchant.

The evidence covers one steel composition, one heat treatment, 10 × 10 mm specimens, and laboratory conditions near 22 °C. It does not establish performance for other martensitic steels, electrolytes, cathodes, or specimen geometries. The paper reports voltage and time but no measured current density or current-time trace. It also omits the exact pause duration between the two 60 s cycles. No replicate count or uncertainty estimate is given for the visual and overall scores. Although Table 12 calls the overall result a weighted sum, the article does not state the weighting coefficients. The roughness ranges also differ between the test table and the discussion for Test 1. Dark particles are attributed to chromium-rich carbides such as M23C6, but no separate chemical confirmation method is reported. These gaps restrict quantitative process transfer beyond the tested preparation.

Process implications for AISI 420 inspection

Section titled “Process implications for AISI 420 inspection”

Engineers preparing heat-treated AISI 420 specimens should use Test 5 when both prior austenite boundaries and carbides must be resolved. It provides a controlled balance of surface flatness, phase contrast, and preparation time. Test 3 is faster at 23 min, but its visual score was 4 rather than 5. Test 4 costs less at €14.19 per sample, but the paper records carbide pull-out and lower image quality. The Kalling II route costs €8.23, yet it took 95 min and produced the rejected surface. Process selection should therefore consider phase visibility and surface artifacts, not roughness alone. For component engineering, this method supports metallographic checks of heat-treatment response. It does not measure strength, wear, corrosion, or service performance. No conclusion about diamond material selection follows because the paper uses 3 μm diamond only as a polishing abrasive.

View Original Abstract

This study presents a comparative analysis between conventional metallographic preparation methods and an optimized electrolytic attack protocol. The limitations of chemical attack with the Kalling II reagent were examined, which tended to generate oxidation stains and poor definition of grain boundaries. Through a series of systematic trials, the efficacy of a hybrid protocol that integrated polishing with 3 μm polycrystalline diamond and an electropolishing with 10% oxalic acid was evaluated. The results indicated that the application of a 10 V potential in 60-second fractional cycles prevented carbide detachment and ensured a specular surface. This methodology provided a reproducible and efficient procedure for the accurate identification of phases and grains boundaries in martensitic stainless steels, being applicable in both industrial and academic contexts.