Clay used in the replication experiment was a commercially available white porcelain clay. After drying at room temperature for over 48 h, clay was bisque-fired up to 800°C, and glaze was applied to it using a single-dip method.

Glaze was formulated by varying the composition ratios of iron oxide, ash of oak, calcite, and feldspar. After glazing, clay pieces were fired at temperatures of 1,000°C, 1,200°C, and 1,300°C. Eight specimens were selected from those produced in a previous study (Choi and Han, 2020), considering changes in the microstructural characteristics based on the firing temperature, firing environment, and feldspar inclusion. The sample information and mixing ratios of glaze materials are summarized in Table 1 and Figure 1.

A portion of replicated specimens was cut, embedded in epoxy resin, and processed using a polisher (Mecatech 250 DPC, Presi, FRA) and polishing papers (#400, 800, 1200, 2400, and 4000), along with polishing suspensions (3 and 1 μL), to achieve mirror-like finish. The glaze microstructure was then observed using an optical microscope (LV100N POL, Nikon, JPN) and a scanning electron microscope (SEM, SU3800, Hitachi, JPN). Additionally, the glaze composition was analyzed using an energy dispersive spectroscopy (EDS, Ultimax, Oxford, GBR) a ttached t o SEM. The g laze l ayers of the specimens were removed using a scalpel and air grinder, powdered using a membrane bowl, and analyzed by XRD (Empyrean, Malvern PANalytical, GBR) to determine the crystal structure and by transmission Mössbauer spectroscopy (WISSEL Gmbh, DEU) to obtain quantitative information on the type of iron oxides and the oxidation state of iron in the glaze. Mössbauer spectra were measured using a transmission type of Mössbauer spectrometer with a ⁵⁷Co/Rh source in a constant-acceleration mode. The 14.4 keV gamma-rays were detected in a proportional counter filled with Xe/CO₂ gas. The Mössbauer spectra were modeled based on a Lorentzian line shape profile as a combination of quadrupole doublet and magnetic sextet components.

The observation of the cross-sectional microstructure revealed differences based on the presence or absence of the glaze material feldspar. Specimens 123 and 153, fired at the lowest temperature, showed that they did not experienced the vitrification of glaze, resulting in differently sized particles with colors of black, white, etc. No difference was found with or without the addition of feldspar.

From specimen 127, fired at 1,200°C—a temperature suitable for obtaining well-fired ceramics—the vitrification of glaze was observed. In specimen 127 without feldspar, iron oxides aggregated in the upper part of glaze, while numerous needle-like and dendritic crystals, displaying yellowish hue, formed in the lower part. This resulted in a tendency of iron oxides and crystal layers being distinguished within glaze. By contrast, specimen 157 with added feldspar showed greenish glaze, with significantly fewer crystals produced than specimen 127. The internal structure appeared relatively uniform, with some iron oxides aggregating into spherical shapes in the upper part of glaze.

Specimens 129 and 159, fired at the highest temperatures, demonstrated crystal formation concentrated in the lower region adjacent to the body. In specimen 129 without feldspar, iron oxides aggregated in the middle part of glaze, forming band-like structures. This resulted in a distinction between the brighter upper part with numerous needle-like crystals and reddish-yellow lower part. Conversely, in specimen 159 with feldspar, no aggregation of iron oxides was observed, and glaze was divided into a dark brown upper part and a greenish-yellow lower part based on color. Few needle-like crystals were formed around the greenish-yellow lower part, but they were significantly fewer compared with specimen 129.

Specimens 227 and 257, exposed to a reduction environment, exhibited different patterns depending on the presence of feldspar. In specimen 227 without feldspar, the formation of dendritic structures was observed throughout glaze, and glaze was divided into a red upper part and a black lower part based on color. By contrast, specimen 257 with feldspar generally showed needle-like structures, with some iron oxides aggregated in the upper part of glaze. Similar to specimen 227, it displayed a yellowish upper part and a reddish-yellow lower part according to color (Figure 2∼ 4).

Crystal structure analysis was conducted on a portion of glaze collected from replicated specimens. The results indicated that, regardless of the material mixing ratios, all specimens exhibited a trend of decreased peak intensity of crystalline structures as the firing temperature increased, indicating increased vitrification. Specimen 123, fired at the lowest temperature, revealed the presence of quartz, calcite, hematite and srebrodolskite. By contrast, specimen 153 with added feldspar showed the presence of alkali feldspar.

As the firing temperature increased, differences to some extent were observed in the identified crystal structures based on the presence or absence of feldspar. In specimens 127 and 129, which did not contain feldspar, the same quartz, calcite, and hematite identified in specimen 123 were present, but their peak intensities decreased in the order of specimen 123 > 127 > 129 as the firing temperature increased.

In specimen 157 with added feldspar, the peaks corresponding to quartz and calcite observed in specimen 153 disappeared, demonstrating only peaks ascribed to hematite and alkali feldspar. Specimen 159 exhibited only the weak peaks of alkali feldspar. Similar to specimens without feldspar, the peak intensity decreased with increasing firing temperature in the order of specimens 153 > 157 > 159.

For specimens 227 and 257, fired under an oxidation environment, quartz, alkali feldspar, hematite, and magnetite were identified. However, compared with specimen 227, specimen 257 displayed higher amorphous peaks (Table 2, Figure 5).

The glaze composition was analyzed using EDS attached to SEM. Specimens 123 and 153, fired at the lowest temperature, exhibited a higher iron oxide (Fe₂O₃) content in glaze than other specimens. The difference in the iron oxide content between specimens 123 and 153 is attributed to the variation in the mixing ratios due to the addition of feldspar.

In the remaining specimens fired at temperatures above 1,200°C, the distribution of components between the surface and interior of glaze varied depending on the presence of added feldspar. In specimens 127 and 129 without feldspar, the surface of glaze contained silica (SiO₂), ranging from 48.61 to 56.19 wt.%, iron oxide, from 7.61 to 12.30 wt.%, and calcium oxide (CaO), from 12.60 to 22.98 wt.%. Inside glaze, the silica content increased to 59.67–67.14 wt.%, while the iron oxide content decreased to 5.71–7.11 wt.% and calcium oxide to 6.56–12.72 wt.%. This indicates a trend of increasing silica content and decreasing iron oxide and calcium oxide contents from the surface toward the interior of glaze.

Specimens 157 and 159 with feldspar showed a trend similar to that of specimens 127 and 129. However, the calcium oxide content on the glaze surface was 13.26–14.86 wt.%, and in the interior, it was 8.96–11.47 wt.%, indicating a smaller variation and thus a more consistent component composition.

Specimens 227 and 257, fired under a reduction environment, displayed similar component distributions, regardless of the firing temperature. The iron oxide content ranged from 12.60 to 14.44 wt.% at the surface and 7.87 to 8.02 wt.% in the interior, showing a trend of increasing iron oxide content closer to the surface.

To examine the overall component distribution in glaze, EDS mapping was performed. Specimens 123 and 153, which did not underwent sufficient vitrification, showed that the components of production materials were evenly distributed throughout glaze.

Among specimens that experienced sufficient vitrification, specimens 127, 129, and 227 without feldspar exhibited clear stratification based on the presence of aggregated iron oxides, as observed in the cross-sectional microstructure. The aggregation of iron oxides was concentrated near the surface layer, while needle-like or dendritic crystal growth was observed near the body, suggesting the formation of anorthite crystals. For specimens 157, 159, and 257 with feldspar, while iron oxides were aggregated at the surface, there was relatively less distinct stratification (Table 3, Figure 6∼ 7).

Figure 8 shows the Mössbauer spectra of all specimens. The fitted values of magnetic hyperfine field (H_hf), quadrupole splitting (ΔE_Q), isomer shift (δ), and relative area (A) are shown in Table 4. When specific iron compounds could not be identified, the phases were represented by doublet or sextet components associated with Fe³⁺. The Mössbauer spectrum of specimen 123 shows hematite, maghemite, Fe³⁺ sextet (S5), and Fe³⁺ doublet (D1). The hematite and two Fe³⁺ doublets(D1, D2) were analyzed in specimens 127 and 129. From these results, it was observed that the hematite content decreased and the Fe³⁺ doublets (D1, D2) increased as the firing temperature increased from 1,200°C to 1,300°C under an oxidation environment. The spectrum of specimen 227 was fitted to hematite, magnetite, and two Fe³⁺ doublets (D1, D2). The spectra of specimen 127 fired in an oxidation environment and specimen 227 fired in a reduction environment were compared. It was confirmed that some of hematite was converted to magnetite in the reduction environment. The spectra of specimens 153 and 157 were analyzed as consisting of hematite and a Fe³⁺ doublet (D1). No magnetic behavior was observed in the specimen 159, which was analyzed as a Fe³⁺ doublet (D1). Based on Mössbauer spectra of specimens 153, 157, and 159, it was observed that, under oxidation environment, hematite content decreased, while Fe³⁺ doublet (D1) increased as the firing temperature increased from 1,000 to 1,300°C. The specimen 257 was analyzed as hematite, magnetite, and a Fe³⁺ doublet (D1). The Mössbauer spectrum of specimen 157 fired in an oxidation environment was compared with that of specimen 257 fired in reduction environment. In the reduction environment, some of hematite was converted to magnetite, and a new Fe³⁺ sextet (S5) was observed (Table 4).