104 Prakt. Met. Sonderband 52 (2018) 
Fig. 1: Hot mounted slab sample, normal section 
Resin 
Seal 
Stee: am 
Fig. 2: Cold mounted slab sample, taper section 10° in vertical direction 
tween scale and steel resulting from thermal stress. Incomplete infiltration facilitates oxide 
breakout, which not only introduces defects into the scale that could be misinterpreted as 
inherent porosity, but also causes scratches in the base steel. Furthermore, adhesion of 
epoxy resins was found to be superior for the samples examined. 
3. GRINDING AND POLISHING 
Due to the brittle nature of the scale, the majority of the preparation effort has to be spent 
on removing defects introduced by sectioning and previous grinding and polishing steps. In 
the present work, all of the preparation was done using a semi-automatic grind- 
ing/polishing machine. Typically, after grinding on grit 80 and grit 220 resin bonded dia- 
mond discs, at least 5 min of polishing with 9 um diamond suspension followed by around 
15 minutes with 3 pm diamond suspension were found to be necessary to remove the 
large majority of defects. In contrast, polishing with 1 ym suspension took no longer than 3 
minutes, since almost all of the defects had been removed in previous steps. 
While Chen reports that preparation using discs with napped cloth caused damage to the 
scale by the fibres pulling out oxide [5], in the present work nap-cloth discs provided a bet- 
ter phase contrast in the final 1 ym polishing step, with only minimal oxide breakout. How- 
ever, the 3 um polishing step was performed on a napless, woven acetate cloth. Ly 
En 
Fig. 3 gives an overview of the preparation of a highly porous sample taken from a lab- Mond S 
scale oxidation trial simulating the oxidation of slabs in a reheating furnace. As explained and ren 
previously, due to the fragile nature of this type of scale, the sample was cold mounted in a damage 
10° taper section. The general structure of the scale is already visible after coarse polish- Sample 
ing with 9 um (Fig. 3a), but finer details like the scale structures close to the interface are fmm 
unrecognizable. Polishing with 3 um for 3 min, which for many types of samples is suffi- 
cient, is not enough to remedy the damage done to the scale during grinding and coarse