Although there are many limitations in collecting samples of metal artefacts, iron-making by-products residues that were generated during the process of iron-making and provide a lot of valuable information on iron-making, such as the raw materials used in the process, operating temperatures, etc. In metallurgy, the residue generated when metal is extracted from raw ore is called slag (e.x., copper slag, iron slag, etc.) (Yoon, 1986). Smelting slag, which is excavated from all iron-related sites, provides a handful of information that allows researchers to identify the nature of slag and understand the details of iron-making processes and smelting techniques used in the past (Rho, 2000).

Types of iron ore, which are the raw materials used for iron-making, include Hematite, Magnetite, Iimonite, Siderite, etc. Iron ore does not exist alone but is mixed together with other types of minerals form a ‘mineral deposit,’ and among those deposits, one with a high iron content is called an iron ore. In ancient times, due to its high iron content and relatively large reserves, only Magnetites were used as the main raw material for iron-making (Jungwon National Research Institute of Cultural Heritage, 2020). Sand iron is generally known to be made of sand through the weathering of Magnetite contained in iron ore, with components such as Ti, V, and Zr contained in iron ore promoting the weathering process after being oxidized. Smelting methods are normally divided into two different types, direct smelting and indirect smelting. Direct smelting is a method in which the reduced iron mass does not melt but rather sinks to the bottom of the surface in the form of sponge iron due to the relatively low temperature (Kim, 2007). Indirect smelting is a method that uses a blast furnace to reduce iron at an extremely high temperature of 1,200℃ or higher, and at the same time, charging an excessing amount of charcoal in order to generate the carburization of iron. This lowers the melting point of iron, allowing molten iron to be obtained. The type of iron produced through this method is called cast iron (Bae and Cho, 2020). Various types of slag occur when useful metals are separated and extracted with the help of specific reducing agents, and since components from raw ores, furnace walls, and other mixtures are dissolved together into these slags, a significant amount of iron oxides are mixed into them. The mineral composition of smithing slags is not diverse compared to smelting slags, and it is difficult to distinguish between refining slags and forging slags, which are different types of smithing slags, only by mineral composition. Since refining slags are usually reheated, their total Fe content remains high, and most produce bowl-shaped iron slag at the floor level. In the case of forging slags, it has a reduced level of impurities while forming a semi-circular-shaped iron slag, along with powder-type iron slags and forging flakes. Smelting slags is mainly characterized by the detection of Fayalite, sometimes coexisting with vitreous slags or Wüstites, while smithing slags are characterized by the coexistence of Fayalite and Wüstite (Yoon, 1986).

The precise manner in which man discovered iron is unknown. Probably its discovery was the result of an accident. Folklore from the central states of India suggests that their forefathers were the first ones to make iron ‘accidentally’ in a hollow anthill. Rapid developments in iron making and its use took place around 1,400 B.C.E., and the history of early iron smelting, practised by the tribal artisans in different regions of ancient India, dates back to 1,300 B.C.E. to 1,200 B.C.E. The use of iron was relatively unknown except in areas where iron-bearing minerals were abundant. The Indus Valley civilization of around 2,500 B.C.E. to 1,800 B.C.E. formally belongs to the bronze age and a fair use of iron appears to have come in comparatively late (Vaish et al., 2000). Bihar possesses its major iron resources in the region of Hazaribagh, Singhbhum, Ranchi, and Palamau, around the plateau areas. These areas are of huge importance even in recent days. The main ore type within these areas seems to be haematite, but there are also other types of ores such as Magnetite, Iimonite, and Titaniferous ores (Chakrabarti, 1992). Also, it is believed that ancient forms of iron smelters may have flourished more in Bihar compared to other regions (Dunn, 1942). While metallurgy in India began with the dawn of the Indus Valley civilization, the Indian state of Jharkhand, which currently consists about 40% of the nation’s mineral resources, has a consistent history and tradition of metallurgy. The primitive tribes of Jharkhand, especially Asur, who are now facing the danger of extinction, are known to own and practice metallurgy of iron, copper, and gold. The Asur and Virgilya, which are the descendants of a tribe known to have first made iron in India, settled in Jharkhand about 4,000 years ago. Specifically, the Asur was one of the first people in the entire world to produce high-quality iron. It is known that the iron they have produced was of excellent quality without corrosion (Roy et al., 2021). Therefore, by analyzing iron-making by-products that were found in Bihar and Jharkhand, India, this study will find out about traditional Indian iron-making methods and techniques.

Iron-making by-products from Bihar and Jharkhand, india, have been collected and analyzed to better understand the traditional steelmaking process.

Normally, the presence and amount of TiO₂와 V₂O₅ within the slags are used as an indicator to determine whether iron ore or sand iron was used as a raw material for the iron-making process (Yoon, 1986). Typically, if the TiO₂ content is higher than 1 wt%, it is viewed that sand iron is used for the process, and if the TiO₂ content is below 1 wt%, it is assumed that iron ore is used. However, according to a reproduction experiment conducted in Korea (Cho et al., 2018), the TiO₂ content was detected as 0.73∼0.76 wt% despite using sand iron as its main ingredient. In contrast, according to another ancient iron-making furnace restoration experiment (Kim and Kim, 2017), the TiO₂ content was detected to be higher than 1.8 wt% although it used iron ore, proving that it is difficult to judge the raw material used based on the content of TiO₂. According to the result of the main component analysis, the TiO₂ content of the slags recovered from Bihar was around 1.11∼2.03 wt%, while the TiO₂ content of the slag recovered from Jharkhand was 0.72 wt%. Ulvöspinel, one of the labelling compounds that tell us whether sand iron was used as a raw material, was found in the compound analysis and microstructure analysis of BH-1 and JK-1, so it would be reasonable to say that both used sand iron as raw material. However, Ulvöspinel was not identified in both BH-2 and BH-3, and since Bihar was in area rich in Hematite, Magnetite, and Titanium ores, we cannot rule out the possibility of Titaniferous magnetite being used in the process.

The main component of smelting slag is FeO-SiO₂, followed by the combination of Al₂O₃ and MgO from the furnace into the mix, and at times, CaO can also be added into the slag. As for the mineral phase, dendritic Wüstite exists within the mix, along with Fayalite as the columnar structure, Hercynire, etc. BH-1 is also presumed to be a smelting slag since Wüstite is observed in the mix.

The content of T⋅Fe and deoxidation agent found in the slag in this study is around 45.85∼49.25 wt% and 33.83∼ 37.69 wt%, which both fall within the range of T⋅Fe and deoxidation agent content of slags created through a direct smelting method. The direct smelting method refers to producing iron by reducing ores at a low temperature of 1,200℃ or less (Rho, 2000).

The operating temperature was estimated by converting the FeO, Al₂O₃, SiO₂ and CaO compositions that were detected through the main component analysis and illustrating it in a FAS (FeO-Al₂O₃-SiO₂) phase d iagram a nd F CS (FeO-CaO-SiO₂) phase diagram (Figure 7) . According to the FAS phase diagram, BH-1 was heated at a temperature of 1,300℃, while BH-2, BH-3, and JK-1 were heated at a temperature of 1,150℃. According to the FCS phase diagram, BH-1 was heated at a temperature of around 1,177 ∼1,200℃, while BH-2, BH-3, and JK-1 were heated at a temperature of around 1,115∼1,205℃. In the compound analysis, compounds such as Fayalite, Microcline, Hercynite, etc., which can be used to estimate temperature, were detected, allowing us to assume that the operation was performed at a temperature around 1,000∼1,200℃, which is almost consistent with the results obtained from the illustrations made in the FAS and FCS phase diagram.

The Reducible Iron Index (RII = 2.39*SiO₂/FeO + MnO) is an indicator that shows the reduction efficiency based on the chemical composition of the slag, and MnO is not reduced but binds with SiO₂. An RII value of 1.0 means that the SiO₂ in the slag is completely fluxed into FeO, while an RII value of below 1.0 means that SiO₂, remains unfluxed, and an RII value greater than 1.0 means that a surplus amount of FeO remains within the slag. The higher the RII value, the higher the reduction efficiency. The reliability of the index can be increased by considering various factors, such as the use of the same raw materials (ores) and clay components (SiO₂, etc.) that exist in the furnace wall (Charlton et al., 2010; Park, 2023). The RII values of the subjects of this study are as follows: BH-1 was 0.77, BH-2 was 1.04, BH-3 was 0.92, and JK-1 was 0.96 (Table 5). According to the microstructure analysis of each sample, the columnar structure of BH-2, BH-3, and JK-1, which all had an RII value near 1.0, was Fayalite. BH-1, the only sample with an RII value far below 1.0, was observed to have Wüstite within the mix, confirming that the reduction efficiency of BH-1 was lower than that of other samples.

It is known that the addition of limestone substances during the smelting process can lower the melting point and reduce its viscosity, improving the overall efficiency of desulfurization and removal of phosphorus (Yoon, 1986). CaO is affected not only by raw materials but also by the substances existing in the furnace walls and fuel components (Lee et al., 2017). Normally, if the ratio of CaO/SiO₂, an oxide in a non-metallic material, is higher than 0.42, it is judged that limestone substances are artificially charged into the furnace (Kim, 2014), and if it falls below 0.83 in the bivariate graph of Al₂O₃/SiO₂-CaO/SiO₂, it is viewed that a significant amount of flux is added in the process (Lee et al., 2017). By applying the subjects of this study to the formula, BH-1 and BH-3 had a value higher than 0.42, while BH-2 and JK-1 had a value lower than 0.42 in the CaO/SiO₂ graph. Meanwhile, in the Al₂O₃/SiO₂-CaO/SiO₂ bivariate graph, BH-2, BH-3, BH-4, and JK-1 had a value higher than 0.83, while BH-1 was the only sample with a value lower than 0.83 (Figure 8). Therefore, the results of the two graphs are slightly different, and the formula ‘y = 0.83x’ has been obtained through existing experimental studies where the main aim was to reproduce ancient iron-making techniques; applying them to real-world iron slags may yield different results. In conclusion, it seems that more extensive research and concrete results will be required to address such issues (Choi et al., 2022)

The analysis results shown in this study were interpreted based on the slags excavated in Korea due to insufficient data/analysis results obtained from slags collected in India. The subjects of this study are not excavated from actual iron-making ruins/sites, so it would be difficult to properly identify the traditional iron-making process of India by only using the results and analysis provided in this research. Therefore, a more accurate and better understanding of the traditional iron-making process of ancient India can be obtained by securing and analyzing various samples from furnace walls, blowpipes, etc.