A variety of etchants were tried to darken the matrix. Klemm’s I gave excellent results for the spheroidize-annealed W1 and the white cast iron (cementite and pearlite as Ledeburite) specimens. A Beraha-type solution (100 mL water, 0.6 mL HCl, 1 g K2S2O5) was used to color the ferritic matrix of annealed T15 and M42 and to color the martensitic matrix in the M42 specimens. Vilella’s reagent was used to darken the martensite in the D3 specimen. Beraha’s Sulfamic acid reagent (100 mL water, 3 g K2S2O5, 2 g Sulfamic acid and 1 g NH4·FHF) was used to color the ferritic annealed matrix of the T1 and M42 specimens. Other reagents were tried, but these yielded the best contrast between the unaffected carbides and the colored matrix while yielding crisp grain boundaries. The specimens were evaluated by automated image analysis to determine the volume fraction, number per unit area and size of the carbides. Next, the specimens were re-prepared and etched with the reagents listed in Table 3. Results of the etching experiments are shown in Table 4.
Most of the selective carbide etchants produced higher volume fractions and greater numbers of particles than were obtained when the matrix was darkened by etching. In a few cases, the results were close, as was the case for the M7C3 carbides in the D3 specimen etched with Murakami’s and Groesbeck’s reagents compared to the measurement made when the matrix was darkened. But, this was the exception rather than the rule.
The reagents do have reasonable selectivity and can be used to identify carbides when sophisticated electron instruments are not available. The results frequently agreed with the literature, but differences were observed. Alkaline sodium picrate colored the Fe3C and M6C carbides, and nothing else, as claimed. Murakami’s at room temperature outlined and colored Cr7C3 carbides in D3 rather than attacking them and attacked the Cr23C6 carbides in 440C instead of faintly revealing them. Groesbeck’s reagent did not attack the Cr23C6 carbides and did not reveal them. It faintly revealed the Cr7C3 carbides and it outlined the M2C carbides rather than attacking them. Electrolytic CrO3 outlined Cr7C3 carbide, as stated, but did not attack them, as claimed. It did not outline M2C carbides but attacked them. Electrolytic 10% ammonium persulfate has not been studied in the past as to its selectivity. It has mainly been used to color Cr23C6 carbide in sensitized austenitic stainless steels. In this study, it outlined Fe3C, M6C and Cr7C3 carbides and attacked Cr23C6, M2C and MC carbides. In general, they are not satisfactory for quantitative metallographic measurements as bias was obtained in the volume fractions, number of particles per unit area and their size.