ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Regular Article
An Experimental Archaeological Study on Iron Sand Smelting in the Korean Peninsular
Hyunkyung ChoNamchul ChoJeongwook HanTeacheon Rho
Author information
JOURNALS OPEN ACCESS FULL-TEXT HTML

2014 Volume 54 Issue 5 Pages 1038-1043

Details
Abstract

The possibility of the use of iron sand smelting in the ancient Korean Peninsula has suggested in recent iron artifacts analysis study. In this study, an iron sand smelting experiment is conducted and a sword is reproduced based on known archeological materials. 22 kg of iron ingot was produced after smelting using 70 kg of iron sand from Gampo. Iron ingot folded and forged 15 times for removing impurities from it and then 2.52 kg of steel ingot was produced. Steel ingot was hammered, heated, annealed, quenched and tempered to make reproduced sword. To obtain the experimental archaeological data, the products of each stage of iron manufacture were analyzed. Gampo iron sand whose total Fe content was 66.10% contained titanium and vanadium. Iron ingot confirmed carbon steel and impurities could be removed by folding and forging. Steel ingot was hypoeutectoid steel that differ depending on the number of folding. The remained non-metallic inclusions as a form of slag were gradually removed over the repeated process of folding and forging. Non-metallic inclusions remaining in the sword were of a glass phase and found to contain titanium and vanadium. This supports the idea that when titanium and vanadium are found in non-metallic inclusions of a steel product, they are the result of iron sand smelting.

1. Introduction

The iron culture in the Korean Peninsula continued to develop in close relation to the growth of the ancient country. In the ancient iron manufacture, the first stage of iron manufacture technology is to secure the raw materials. Raw materials for iron production include iron ore and iron sand. In such regions in the Korean Peninsula as Changwon Daho-ri site,1) Jincheon Seokjang-ri, Milyang Sachon, and Chungju Chilguem-dong, iron ore has been excavated and there were many reports that iron ore was used in iron artifacts manufacturing.2,3) In a results of recent iron artifacts analyses, however, the existence of titanium and vanadium, two of the indicating elements of iron sand smelting, was confirmed,4,5) Also a research that Chilji-do stored in Isonokamijingu, Japan, was made of iron sand was released.6) According to these researches, it is drawing attention the possibility of the use of iron sand smelting in the Ancient Korean Peninsula.

In this study, an iron sand smelting experiment is conducted and a sword is reproduced based on known archeological materials. Specifically based on reference that imply the possibility of 21 swords stored in Shosoin, Japan, being made in Silla7) and historical articles of Sejongsillok Jiriji on iron sand deposits,8) the iron sand smelting and sword producing process of Silla were represented. The raw materials and products in each stage of the process including smelting and refining were scientifically analyzed to examine the iron sand smelting process and iron swords of Silla in terms of archeological experiment.

2. Experimental Method

2.1. From Iron Sand Smelting to Reproduction of Sword

As recorded in Sejongsillok Jiriji, iron sand was collected from a seashore, Gampo, Gyeongju, Gyeongsangbuk-do (Fig. 1(a)). The fuel was pine charcoal, which was cut (3×3×3 cm) and put into the funarce (Fig. 1(b)). A round shaped furnace was designed, and the three elements - siliceous earth, yellow ochre, and limestone - were mixed in water in order to minimize the time wasting of crushing after smelting. When each of the three layers is separated after smelting, it is easy to grasp the shape and location of the iron ingot produced in the furnace (Fig. 1(c)).

Fig. 1.

Instruments of smelting; (a) cross section of iron sand at Gampo seashore, (b) pine charcoal, (c) Drawing of reproduction furnace.

20 kg of pine charcoal was put into the furnace and made a fire to dry furnace. Then 1 kg of iron sand and 2.5 kg of pine charcoal were put in the furnace at intervals of 10 minutes. Slag piled up at the bottom of the furnace was discharged four times in total. 70 kg of iron sand and 175 kg of charcoal were used in total. After the fourth discharge of slag, the furnace was taken apart. The produced iron ingot amounted to 22 kg, and the recovery rate was 30.9%. It took 17 hours in total from heating to checking the produced iron ingot.

Iron ingot was put into the smithery hearth filled with pine charcoal and heated at 1500°C. The heated iron ingot was hammered to produce rectangular steel ingot. Meanwhile, impurities inside were melted down and flew out because of the pressure of the hammer. Pine charcoal powder for carburizing and yellow ochre water for prevention of decarburizing coated on rectangular steel ingot, and then it was heated again. The re-heated steel ingot was cut by an axe, which was followed by folding and forging. This folding and forging process was repeated 15 times, and then the steel ingot without impurities amounted to 2.52 kg.

The rectangular steel ingot was extended as long as the sword, heated in the smithery hearth at 1350°C, and then hammered. Common swords in ancient times were about 1 m long, but two 50 cm long swords were produced in this experiment so that they could be cut later. The properly shaped swords were heated at 800°C and cooled down at room temperature. This process of annealing was repeated three times. For partial heat treatment, clay was put on the back and blade of the sword in different thickness. Then they were heated again and went through quenching to cool down in a round oil bucket. Then the tempering them at 200°C and cool them down again at room temperature were repeated three times. After the heat treatment was completed, the swords were shaped with wooden and rubber hammers, and then the blades were sharpened with whetstone. The completed sword weighed 0.40 kg. Two swords were produced from 70 kg of iron sand.

2.2. Analysis of Smelting Products

The products of each stage of the iron manufacture were analyzed (Table 1). The iron sand went through drying and powdering to examine the main component and compounds. The sample was glassified and analyzed with the Wave dispersive X-ray Fluorescence Spectroscopy (WD-XRF: S4 Pioneer, Bruker) to examine the main component. The analysis result was measured by means of the semi-quantitatively. The iron sand composition was analyzed with X-ray Diffractometer (XRD: X’PertPRO MPD, Philips). As for the analysis conditions, 2 theta was 3–70 deg, scan speed 1 sec/step, step size 0.03 deg, voltage 40 kV, and current 30 mA respectively. The target of the analysis was copper. The microstructure of the steel ingot and reproduced sword was observed with metallurgical microscope (DM 2500M, Leica). A section was cut for mounting, and went through polishing up to 1 μm. After polishing, the sample went through etching in the 3% nital (HNO3+Ethyl alcohol) solution and then observed. Chemical composition of microstructure was analyzed with a Field Emission Scanning Electron Microscope (FE-SEM: JSM-7401F, JEOL) attached an Energy Dispersive Spectroscopy (EDS: INCA, Oxford). Samples were coated with Pt and analyzed by means of an Energy Dispersive Spectroscopy (EDS: INCA, Oxford).

Table 1. The list of analyzed samples.
SamplesWeight (kg)
Iron sand70.00
Iron ingot22.00
Folded steel ingot (1 time)10.6
Folded steel ingot (3 times)6.16
Folded steel ingot (6 times)4.34
Folded steel ingot (9 times)3.74
Folded steel ingot (12 times)2.80
Folded steel ingot (15 times)2.52
Reproduced sword0.40

3. Analysis Results of Products

3.1. Iron Sand and Iron Ingot

The iron sand used in the experiment was analyzed by means of XRF. As a result, the total Fe was 66.10%. The content of titanium was 0.71 wt% and vanadium 0.24 wt% respectively (Fig. 2(a)). The XRD analysis result shows that the iron sand compound consisted of magnetite and hematite. A type of iron sand was magnetite because peaks of magnetite were stronger (Fig. 2(b)).

Fig. 2.

The analysis results of iron sand; (a) main components by XRF, (b) XRD patterns.

On the surface of iron ingot (Fig. 3(a)) produced after iron sand smelting, pearlites whose composition was almost of eutectoid, were found, and the contents of carbon decreased in the deeper section where widmannstatten ferrite and pearlite coexisted (Fig. 3(b)). The overheated structure of widmannstatten seems to be formed while iron ingot was cooled down at normal temperature before being cut down. The black spot is the pores in the iron ingot. The SEM-EDS analysis result shows the existence of non-metallic inclusion that consisted of FeO, SiO2, and Al2O3 (Fig. 3(c)). The amount of TiO2 in the non-metallic inclusion was 5.74–6.31 wt%, which was larger than that in the iron sand. This analysis result indicates that the iron ingot produced after iron sand smelting was carbon steel.

Fig. 3.

Iron ingot; (a) directly after smelting, (b) microstructure, (c) SEM image and EDS results.

3.2. Folded Steel Ingot

The refining process to remove impurities in the iron ingot involved folding and hammering. After the iron ingot was shaped as a rectangle and folded and hammered once, a corner of it as much as 3 cm was cut off as a sample. Then as it was folded 3, 6, 9, 12 and 15 times, samples were collected in the same manner. On a section of the steel ingot folded once, irregular forms of non-metallic inclusion were found. The surface included eutectoid pearlites while the inner part contained needle-shaped ferrites and grain of ferrites in the order. Figure 4 shows the microstructures (in the red box) on the sample sections of steel ingot after three times of folding and forging, 9 times of folding and forging, and 15 times of folding and forging respectively. The steel ingot folded three times contained pearlites inside unlike that folded once (Fig. 4(a)). This is probably because the surface of the steel ingot was folded inward in the folding and forging process. The pearlites of eutectoid and ferrites were shown alternatively depending on the times of folding and forging. The vertically lined non-metallic inclusion indicates the direction of folding and forging.

Fig. 4.

Microstructure of steel ingots; (a) folded 3 times, (b) folded 9 times, (c) folded 15 times.

The steel ingot folded and forged 9 times showed relatively even microstructure compared to that folded and forged three times. This indicates that the carburizing and decarburizing process with the yellow ochre water and charcoal powders produced steel ingot of consistent carbon contents. Both pearlites and needle-shaped widmannstatten ferrites were found, and the grains of pearlites were finer than that of steel ingot in the previous step. In the part of the steel ingot folded and forged, non-metallic inclusion was found (Fig. 4(b)). As for the steel ingot folded and forged 12 times, the carburized section on the steel ingot surface turned to be a black belt with wave patterns. The grains of ferrites became finer. A section of the steel ingot folded and forged 15 times showed a clearer black belt, where eutectoid pearlites were formed. Except the area of the black belt, widmanstatten ferrites were developed (Fig. 4(c)).

Figure 5 shows the result of analyzing the non-metallic inclusion in steel ingots summarized in Fig. 4 by means of SEM-EDS. The dark background indicates the glass phase. The gray polygon is fayalite. The white round particle on the fayalite is wustite. All of the non-metallic inclusions are slag-shaped. As the times of folding and forging increased, the shape of fayalite turned faint while the ratio of the glass phase increased. This indicates that impurities were being removed in the repeated process of folding and forging.

Fig. 5.

SEM image and EDS points of non-metallic inclusion in steel ingots; (a) folded 3 times, (b) folded 9 times, (c) folded 15 times.

3.3. Reproduced Sword

After 15 times of folding and forging, the refined steel ingot was used to make a sword. Gray wave patterns were observed on the surface of the completed sword, which was indicated with a blue line on the drawing in Fig. 6. The microstructure of a section in the middle of the sword was observed. A gray line existed inside the sword as well. The back side (Fig. 6(a)), middle (Fig. 6(b)), and blade (Fig. 6(d)) of the sword consisted of fine pearlites. The grains at the back and blade were finer than those of the internal section, and the pearlites of the blade, which went through the processing more than other parts, were much finer. The gray line looks darker at the inner part of the microstructure because of the carburizing during the folding and forging process (Figs. 6(c), 6(g), 6(e)). The bright part of the sword section consisted of pearlites and ferrites with lower contents of carbon. Widmanstatten ferrites were also found in the grain boundary of pearlites. The size of the grain might be varied depending on the extent of processing.

Fig. 6.

Drawing and microstructure of reproduced sword.

Various types of non-metallic inclusions existed within the sword. Figure 7 shows different non-metallic inclusions within the sword as observed with SEM. The EDS analysis result is presented in Table 3. The Most of the glass phases were of iron oxide while the amount of SiO2 and other components was smaller than that in the folded and forged steel ingot. At the points of No. 5 and No. 6, titanium and vanadium were the main components, which were from iron sand and not removed during the refining process.

Fig. 7.

SEM image and EDS points of non-metallic inclusion in reproduced sword.

Table 2. EDS results of non-metallic inclusion in steel ingots (Fig. 5).
Analysis PointChemical composition (wt%)
Na2OMgOAl2O3SiO2K2OCaOTiO2V2O5MnOFeOC
a-119.520.417.9525.562.101.1335.752.48
a-29.564.5011.640.870.380.580.7268.971.97
a-320.418.0024.451.611.2638.541.80
a-410.394.2910.970.980.620.1765.581.95
b-10.608.8123.452.040.8766.732.43
b-24.169.630.420.8085.823.22
b-30.6912.300.523.6988.073.48
c-18.3222.402.340.9266.723.01
c-20.603.695.931.330.4493.482.26
c-36.7519.811.730.8865.855.35
c-40.9620.0014.5167.533.61
Table 3. EDS results of non-metallic inclusion in reproduced sword (Fig. 7).
Analysis PointChemical composition (wt%)
Na2OMgOAl2O3SiO2K2OCaOTiO2V2O5MnOFeOC
11.181.421.56104.232.48
20.761.762.50110.321.97
31.271.8723.9331.812.266.213.051.4847.321.80
40.47123.531.95
56.3525.382.3126.3650.727.4423.532.43
61.352.4528.3634.635.286.747.547.993.0714.863.22
70.621.93119.333.48

4. Discussion

In general, if titanium or vanadium is found in non-metallic inclusions of a slag or iron product, there is a possibility that the product went through iron sand smelting. In the process of iron sand sword reproduction from smelting to refining, titanium and vanadium contained in iron sand were found in most of the non-metallic inclusions of the sword. It turned out that the iron ingot smelted from iron sand could be used for steel swords with no need for a separate steelmaking process since it itself was carbon steel.

Impurities were removed in the process of folding and forging and then refining of 22 kg of iron ingot. As the times of folding and forging increased, non-metallic inclusions turned to be of a glass phase of a slag type of iron oxide. It was also observed that the amount of non-metallic inclusions reduced, which is because the slag remaining after the forging process was destructed and discharged out of the steel ingot. As the grains turned to be finer and the contents of carbon became uniform, pearlites and widmanstatten ferrites were distributed evenly. As the times of folding and forging increased, the gray line became distinctive probably because of the pine charcoal powders on the steel ingot surface, which resulted in carburizing, followed by folding.

The gray line was observed in the sword as well. Wave patterns were shown in the surface of the sword. On the back and blade of the sword, which involved a large extent of processing, fine pearlites were observed. Pearlites of the side contained a lower content of carbon than on the back and blade as well as widmanstatten ferrites which would be formed on the grain boundary. The gray line consisted of pearlites with a higher content of carbon than the surrounding area, which indicates the trace of carburizing and folding and forging. Non-metallic inclusion remained in the sword as a form of glass phase, which shows that the excellent quality of refining. Titanium and vanadium, found in the iron sand, remained in the sword as well. This supports the idea that when titanium and vanadium are found in non-metallic inclusions of a steel product, they are the result of iron sand smelting.

5. Conclusion

This study includes an archeological experiment to represent the iron sand smelting and sword making process common in the ancient Korean Peninsula. The products of each step were analyzed, and the results are as follows:

(1) Gampo iron sand whose total Fe content was 66.10% contained titanium and vanadium. The smelting of 70 kg iron sand produced 22 kg iron ingot, which shows that the content of carbon of this carbon steel is low. The steel ingot was hypoeutectoid steel that differ depending on the number of folding.

(2) Impurities in the carbon steel iron ingot could be removed by forging. The non-metallic inclusions in a form of slag were gradually removed over the repeated process of folding and forging. The remaining inclusions turned to be of a glass phase. As the degree of perfection of the refining process is high, the non-metallic inclusion is likely to remain in a glass phase.

(3) When the process of folding and forging was repeated, a gray line was formed, which is a trace of the carburized surface in the previous step. This spot contains a higher content of carbon than surrounding areas. As the process of forging was repeated, the grain turned to be finer with the content of carbon uniform. This gray line forms wave patterns on the sword surface.

(4) Non-metallic inclusions remaining in the sword were of a glass phase and found to contain titanium and vanadium. It seems that these elements were from the iron sand. When such components are found in an iron product, therefore, they might result from iron sand smelting as supported by existing studies. It is necessary, however, to distinguish it from titan iron ore smelting, and the future experiment and study should classify the characteristics of iron sand smelting and iron ore smelting found in the Korean Peninsula.

References
  • 1)   M. J.  Son: The Iron Culture of Baekje, Juryusung Publishing, Seoul, (2008), 72.
  • 2)   J. S.  Park and  Y. D.  Jung: J. Kor. Inst. Met. Mater., 42 (2004), 307.
  • 3)   D. S.  Yoon: A Metallurgical Study of Aarly Iron Artifacts in Korea, Korea University Press, Seoul, (1984).
  • 4)   Y. S.  Jung: Master Thesis, Kongju University, Chungnam, Korea, (2009), 28.
  • 5)   E. A.  Son: Master Thesis, Kongju University, Chungnam, Korea, (2011), 21.
  • 6)   T. C.  Rho: Baekjehakbo, 6 (2011), 197.
  • 7)   J. S.  Choi: The Collections of Shosoin and Uunified Silla, Ilji publishing, Seoul, (1996), 163.
  • 8)  Sejongsillok: Jiriji, Gyeongsanng-do Gyeongju-bu, Vol. 5, ed. by Danjong, Chunchugwan, Hanyang, (1454), 647.
 
© 2014 by The Iron and Steel Institute of Japan
feedback
Top