A study on ancient casting method through microstructures of impurities and copper found in bronze excavated from the Unified Silla Period

DOI: https://doi.org/10.21203/rs.3.rs-2611343/v1

Abstract

In Hwangnyongsa temple site in Dongdaebong mountain in Korea, a large number of gilt-bronze artifacts were excavated. Optical microscope, X-ray fluorescence, scanning electron microscope, energy dispersive X-ray spectroscopy, focused ion beam, transmission electron microscopy are used to identify the structure of the specimen and impurities. This article investigates the impurities and copper grains in bronze Buddha’s robe from Hwangnyongsa temple site to suggest ancient bronze craft technology in the unified Silla period. XRF results of specimen indicate that were made using alloy of copper and tin, and gold plating was done on the front side only. The microstructure of specimen was confirmed as a recrystallized polygonal structure, with twins and impurities. The results of impurities’ EDS indicated the molar ratio of Cu : S as 2 : 1 and electron diffraction pattern corroborated with that of Cu2S (JCPDF 33-0490). A layer of specimen surface consisted of high lead layer and copper grains. It is speculated that the copper grains were formed by reaction of Pb with matte (Cu2S) during casting. Lead with black gas would help to make high quality bronze. The copper grains used to exist in Silla, as detected in Koryo bronze artifacts. Hence, it was possible that the ancient artisans knew that lead make bronze quality good when they were casting bronze.

Introduction

Bronze began to be used in the early stages of human development. In the early Bronze Age, there were lots of impurities inside the bronze, however, iron comprised lower than 0.1%, while in the late Bronze Age, the impurities decreased, but the iron comprised higher than 0.3% [1]. This may be because copper ores containing iron were used without smelting technology developed. If the smelting technology did not develop, the iron remained unfiltered from the artifacts. Approximately 1% of copper oxide ore, near the outcrop, can be easily found on the ground; the 99% of copper sulfide-based (chalcopyrite, pyrite etc.) ores containing iron are underground. After copper ores which existed on the surface were all used up, people dug up the ground to get copper ore. The process of removing sulfur from iron-free copper is similar to removing oxygen from copper oxide. However, in the case of iron-containing chalcopyrite smelting, the most important process is removing iron from chalcopyrite. 

In Korea Peninsula, 298,548 kg of metal was used to cast the Hwangnyongsa bell in the 13th year of King Gyeongdeokn (754), 184,020 kg of metal and iron on Bunhwangsa bronze statue the following year, and 72,000 kg of brass was used to make bell of king Sungdeok in the 6th year of King Hyegong (770) [2]. It indicates that the Unified Silla (7th–10th) significantly progressed in mining, and changed the smelting copper technology compared with previous periods. As a result of copper slag’s component analysis [3] which was excavated from the ruins of the Unified Silla Period, sulfur and iron were detected in large amounts. It was indicated that copper sulfide ore was already used during this period. It is speculated that ore was already mined using the shaft method [4]. According to the excavated copper crucible from the ruins of 376 Hwangnam-dong and 681–1 Dongcheon-dong, Gyeongju analyses, copper oxide and charcoal were put into crucible together and heated to high temperature [5]. The temperature would be approximately 1,000 ℃ because the XRD result of slag could not detect effectively fayalite and crystalline substances [4].

A study of this excavated artifacts is considered useful in understanding the bronze casting technologies, particularly of the Unified Silla period (7th–10th), because no detailed studies of the microstructures of impurities and copper grains remains unpublished. This article investigates the impurities and copper grains that were fabricated during the unified Silla period and suggests ancient bronze casting method. 

Background

Hwangnyongsa temple, located in Dongdaebong mountain in Hwangnyong-dong, Gyeongju-si, Gyeongbuk, South Korea, was found in 633 at 『Bulguksagogeumyeokdaegi』 [6]. Hwangnyongsa temple was maintained until the Choseon Dynasty (14th–19th) Buddhism was suppressed due to the policy of promoting Confucianism, and this temple was abolished. Hwangnyongsa temple was one of the twin pagoda Buddhist temples that intensively appeared from the late 7th century to the mid–8th century. It is presumed that it was one of the royal temples, which was similar to Gameunsa site, Hwangboksa temple and Bulguksa temple, which are closely related to the royal family. 

Research institute of Buddhist cultural heritage conducted a survey of Hwangnyongsa temple site in Dongdaebong mountain, which was supported by the Cultural heritage administration. Up to now, excavation has been carried out two times, and many relics, for example corridors, building site, stonework, stone rows, and entrances, have been found on the west pagoda. A large number of gilt-bronze artifacts, such ghost faces, Danggans like flagpole, stylobates, and lion statues, were excavated from the western section from the Central. It was identified that Hwangnyongsa temple was a very important temple in the Unified Silla period, through the excavated gilt-bronze artifacts, sculpted Buddha statues, stone Buddha statues, and lower pedestal stones [6].

Method

X-Ray Fluorescence(XRF)

At the non-destructive scientific analysis of the excavated artifacts, Bruker S1 titan 600, an XRF analyzer was used, and component analysis of the artifact’s surface was performed. To know exact surface component, impurities were eliminated from artifact’s surface before analyze XRF. Each of five points or more was analyzed, and the average value of the results was set as the analysis result value.

Optical Microscope(Om)

It was used for samples which contained metal corrosive substances eliminated during conservation of excavated artifacts. Samples were removed from archaeological artifacts either using careful drilling to minimize the visual impact of sampling. Therefore, samples that were extremely small, were mounted with epoxy resin and polished to a mirror surface. Mounted cross-sections were polished with #800, #1000 and #2000 sandpapers, and 3 µm and 0.5 µm aluminum pastes. Etching with 10% ammonium persulfate was performed. Mounted cross-sections were observed using a Nikon Eclips LV100Npol optical microscope.

Scanning Electron Microscope(Sem) And Energy Dispersive X-ray Spectroscopy(Eds)

Mounted cross-sections were observed using a FEI Inspect F SEM. Inspect F enables high-resolution inspection and characterization capabilities using a high brightness, high current Schottky field emission source providing clear, sharp and noise free imaging. The samples were observed after surface carbon coating because the corrosion layer does not have electrical conductivity. This equipment also includes and Oxford Instruments EDS spectrometer. Etching specimen may affect the EDS results, therefore unetched sample were analyzed.

Focused Ion Beam(Fib)

To observe a specific part of the specimen, a nano-sized specimen was produced with a JEOL JIB-4601F focused ion beam. JIB-460F is a multi-beam system equipped with FE-SEM and high-output FIB column. FIB can surface mill in the area of tens of microns to tens of nanometers, and simultaneously, it is possible to observe the inside using FE-SEM.

Transmission Electron Microscopy(Tem)

Nano-sized specimen was observed in a JEOL JEM ARM 200F transmission electron microscopy(TEM) with EDS. The TEM obtains an enlarged image by passing an electron beam through a thin specimen, and is a high-resolution device that can see the arrangement of atoms.

Results

Sample

Among the excavated artifacts, seated Buddha’s robes were selected as a specimen. A pedestal which was covered by the Buddha’s robes, supporting the Buddhist image, is called as Sanghyunza in Korean. The Buddha’s robe was excavated in two sizes; photos in Figure 2 (a) to (c) are big robes and (d) to (e) are small robes. The width x height of the big and the small ones are 31.6 x 21.9 cm, 21.1 x 18.1 cm, respectively. It is indicated that it was a part of big Buddha statue or Buddha’s pedestal. Both of robes are bent in the shape of L, and there are grooves and holes that appear to be used for fixating it to somewhere. It is presumed that each part was made of cast bronze and fixed into the pedestal frame. When both of Buddha’s robes were excavated, they were covered with soils and green corrosions. After removed covered soils, surface of the Buddha’s robes was bumpy with non-uniform attachments. Gold foil existed on the front side, however, it could not be seen inside. The big Buddha’s robe had crack from left side, but it is not too unstable.

 In general, Buddha statues and pedestals were made separately in unified Silla period [7]. Of the big size Buddha statues, standing Bhaisajyaguru statue of Baengnyulsa in Gyeongju is an example of separate casting. Not only the pedestal but also both hands of Bhaisajyaguru were casted separately as shown in Figure 3–(a). However, there was no example of the pedestals which were casted separately similar to excavated Buddha’s robe. In the case of ceramic pedestals, similar manufacturing methods existed. A ceramic Buddha pedestal excavated in Cheongyang Bonuiliyoji (7th) was manufactured in each parts and it has a hole to fix them. It is speculated that the Buddha’s robe similarly made with a ceramic Buddha pedestal from Bonuiliyoji (Figure 3-(b)).

Alloy composition

Table 1. Excavated Buddha’s robes XRF results (wt%)

 

Cu

Sn

Au

Pb

Big size robe front

87.7

10.1

2.0

-

Big size robe inside

92.6

5.6

-

1.8

Small size robe front

73.4

4.9

21.7

-

Small size robe inside

77.8

13.8

-

8.4

XRF results of specimen indicate that both of them were made using alloy of copper and tin, and gold plating was done only on the front side. Lead was detected only inside, and it was possible that lead and gold plating were related. Before gold plating was performed, casting of Buddha’s robe was grinded to make surface smooth; hence, the lead attachment on surface was eliminated at that time. In the Silla Dynasty, the method of gold coating used mainly mercury amalgam in the Korean Peninsula. In case of mercury amalgam, mercury with gold was applied to a bronze surface, followed by heating to a temperature at which mercury vaporized (638 K). The vaporization temperature of mercury was higher than the melting point of lead, therefore, any remaining leads of front surface would have melted. The mercury amalgam was easy to use for large and complex artifacts, therefore, it was mainly used as a gold plating method for the Silla Dynasty’s bronze [8].

Microstructure

Figurer 4-(a) shows that microstructure of specimen was confirmed to be recrystallized polygon, with twins and impurities. According to the EDS analyses as shown in figurer 4–(c), the weight % of sample was Cu – 2.89 wt% Sn. The size of polygonal structure medially increased. Polygonal structure indicated that it was annealed; twins indicate that it had been heated to the copper’s recrystallization temperature (573 K). The polygonal structure and twins indicated that the final heating of Buddha’s robe was probably performed to vaporize the mercury to coating gold. Mercury (Hg) with gold was detected on the surface of other excavated artifacts at the same time, and it was speculated that mercury amalgam would also be used for the Buddha’s robe. 

 Impurities optical microscope image is shown in Figure 4–(b). ⓐ was base metal and ⓑ were impurities. The microstructure of specimen contained polygonal impurities which were grey and 5~50 μm in size. The results of the EDS analysis as shown in Figure 4–(d) was Cu – 36.47 mol% S – 5.72 mol% Fe – 0.43 mol% Se. The molar ratio of Cu : S was 2 : 1. A trace element Se was both detected and not detected depending on the impurities. Selenium (Se) was discovered in 1817 by Jons Jacob Berzelius [9], and was mainly found in sulfuric copper ore, lead and nickel ore. It was obtained from the anode of copper refineries as electrolytic metal refining byproduct [10]. In the Silla Dynasty’s bronze, Se would be contained in Cu2S because there was no electrolytic copper refining method.

A boundary between ⓐ and ⓑ was cut with FIB and studied using TEM-EDS. As shown in Figure 5, Cu and Sn were detected in ⓐ, Cu, S and Fe were detected in ⓑ Cu–S grain. Sulfur was mainly detected but Sn was not detected in Cu–S grain. EDS mapping shows that Fe was isolated and presented in a size of 1 mm or less in the Cu–S grain. The high Fe particles could not be confirmed using SEM and OM. 

The magnified TEM image of Cu–S grains is shown in Figure 6–(a), indicating two different crystals ① and ②. EDS of the Cu–S grain inside point shows that ① was Cu – 34.01 mol% S – 1.13 mol% Fe and ② was Cu – 37.85 mol% S – 12.32 mol% Fe. ①’s and ②’s electron diffraction patterns are showed in Figure 6–(b) and 6–(c), respectively. ①’s electron diffraction pattern corroborated with that of Cu2S (JCPDF 33-0490) and that of ②’s corroborated with Fe3O(JCPDF 19-0629). Two or more crystals were mixed in ② particle, generating a structure of bright pattern, similar to magnetite (Fe3O4). These were intermediate products in the copper refining process, and it was speculated that a part of the matte was remaining. However, usually magnetite was not helpful in making copper, hence, all the iron was removed into the matte by oxidation [11].

1.1 Surface corrosion layer

 A layer of specimen surface was unlike any other bronze corrosion layer. Although there were multiple possibilities depending on the burial environment, generally, bronze surface layer of excavated artifacts contained CuO, Cu2O, and greenish copper corrosion matters, tin oxide and lead oxide [12, 13]. However, as shown in specimen surface, it was consisted as a thick lead layer (200 mm) as shown in Figure 7. The thick lead layer contained long needle shape crystals near surface and elliptical crystals near αCu. To confirm the components of each structure, it was analyzed using EDS mapping and the results are shown in Figure 7 (b) – (h). Pb, Si and Fe were detected using EDS mapping in the thick lead layer, and Cu was highly detected in grey and yellow elliptical grain. Because Si, Al and Fe are present in large amounts in the earth’s crust, it may indicate that these elements were related with bronze corrosion environment. In contrast, Si, Al and Fe could also be slags, which were stabilizing substances, to remove Fe from chalcopyrite. 

According to the optical microscope image as shown in Figure 8, copper grains used to exist in high lead corrosion layer in large amounts. These grains were maximum 9 mm size and were shown in ① yellow color inside, ② grey color outside and ③ black color of lead layer. The composition of these layers are listed in Table 2. Layer ① contained 93.20 mol% Cu, layer ② contained 55.47 mol% Cu, 42.31 mol% O and layer ③ contained 52.19 mol% O, 24.42 mol% Pb, 6.55 mol% S, 3.42 mol% Fe, 2.10 mol% Cu. Layer ② contained a lower copper and a higher oxygen than layer ①. Layer ③ contained a higher lead, iron and oxygen and a lower copper than layer ① and ②. To analyze the copper grain, a boundary between brown and grey was cut with FIB studied using TEM, and the results are shown in Figure 9. 

Table 2. The EDS results of corrosion layers

mol%

O

Si

S

Fe

Cu

Pb

4.14

-

1.91

0.58

93.20

0.18

42.31

-

1.58

0.34

55.47

0.18

52.19

11.31

6.55

3.42

2.10

24.42

Figure 9-(a) shows three layers’ in TEM image; layer ① yellow copper grain on the left, layer ② grey edge in the center, layer ③ black background on the right. The boundaries between the three layers were distinguished, however, as shown in Figure 9-(b), upper part of layer ① and ② breaks the boundary was confirmed. In contrast, the boundary of layer ② and ③ was clearly separated in TEM image of Figure 9-(a), hence, ③ layer seemed not to be related with layer ① and ②.

 To find out a crystal structure in each layer, we analyzed electron diffraction pattern and the results are shown in Figure 10. The electron diffraction pattern of layers ① and ② corroborated with that of Cu (JCPDF 04-0836) and layer and CuO (JCPDF 48-1548), respectively. Layer ③’s electron diffraction pattern showed rings, indicating that it was polycrystalline. In general, tenorite (CuO) is not kinetically favored and is usually found in burned burial environments or slowly heated in air [14]. It was possible that the mercury amalgam method gilded the Buddha's robes and heated to a temperature at which mercury vaporized (638 K). Hence, it seemed that layer ① and ② existed since Buddha’s robe was created. 

Discussion

Microstructure of Buddha’s robe had Cu2S grains which was intermediate of Chalcopyrite (CuFeS2). EDS of Cu2S detected Fe and Se, hence, Chalcopyrite with Se were used to make bronze alloy. It is speculated that the reasons behind the remaining Cu2S are low refining temperature, the short reaction time between copper sulfide and O2, and insufficient O2 gas. 

Lead, detected from inside of sample, was observed in the surface corrosion layer. In the surface, not only lead but also Si, Al, Cu and Fe were detected. The removal of Fe from chalcopyrite is inferred by the sulfur-oxygen potential plot at 1573 K in Figure 11. 

Oxygen in the air (pSO2=0.21) is used for oxidation of copper ore, and if the reaction rate of oxygen is 50%, pSO2=0.1, the reaction proceeds along the dotted line. The graph in Figure 11 shows that Fe is oxidized faster than non-ferrous metals, such as copper, lead, and zinc, and if the oxidation continues, magnetite (Fe3O4) is produced as a stable substance. Magnetite sinks to the bottom of the furnace without proceeding for the reaction, hence, it reduces the volume of the furnace. The furnace is heated up or a stabilizing substance, such as FeO must be added to prevent the production of magnetite. In general, silicon oxide is added as the stabilizing substance. In reduction smelting, if the silicon oxide is put in and the wind is strongly pushed, iron forms a slag of FeO–SiO2. After removing FeO–SiO2, blister copper can be obtained. Until a method of removing iron by adding silicon oxide to make FeO–SiO2 slag was discovered, it is speculated that copper oxide or iron free copper ore was used. To remove iron from chalcopyrite (CuFeS2), two-steps were required. First step is adding silicon oxide to make FeO–SiO2. After removal of FeO–SiOas slags, matte which contains 50%–60% copper remains. Second step is desulfurization of the matte, resulting in blister copper containing more than 98% copper [15].

The yellow grains which were detected copper over 90 mol% did not naturally occur. Data from this and previous studies lead to confirm a reason why copper grains were remaining. Figure 12 shows microstructures of detecting copper particles in a Koryo period (10th–14th) bronze casting artifacts. Figure 12–(a) is a Koryo bottle [16], (b) is Uaseosangmun mirror (bronze mirror 1) (c) is Whamunsomun mirror (bronze mirror 2), and two bronze mirrors show copper particles which were shown inside Pb, and Cu2S exist around Pb. From these microstructures, it is expected that Pb and Cu2S are closely related to the formation of copper. 

Therefore, in previous studies, we tested Cu2S and Pb powder were heated at 1273 K for 4 h and cooled at 100℃ per 1hour. The reaction results between Cu2S and Pb at 1273 K are shown in Figure 13. The experimental results had three structures which were Cu2S, Pb and Cu. The copper grains, also shown in Figure 13–(a) are similar to the ancient bronze artifacts. During Cu2S and Pb reaction at 1273 K, black gas was generated which was identified as PbS using EDS and XRD. The reaction is presented below [17]: 

Cu2S + Pb → PbS↑ + 2Cu

Copper particles were also shown in Chinese bronze mirrors, which were studied by Yokoda [18]. In Yokoda’s study, copper grains were detected near surface corrosion layer, and lead high layer also can be seen. There were no composition analyses of the Chinese mirrors, however, copper grains, were shown near lead high layer, corroborating with our finding.

According to the study on ancient copper smelting from Dongcheondong site in Gyeongju [4], slags which contain smelting tin, copper ore and lead galena were detected in crucible. We speculate that lead galena in crucible was a reaction result of Cu2S and Pb. In conclusion, to lower the smelting temperature, lead was put in copper ore. Then, to refine and make bronze alloy casting, refine tin was put into the copper ore. 

Data from these studies lead to the conclusion that copper contained Cu2S as copper smelting intermediate and Pb would be reacted with Cu2S; it helped in making high quality bronze and lower casting temperature. In fact, leaded copper was restricted to casting only because up to 2% Pb significantly increases the mobility of molten metal. Lead is insoluble in copper, and remains dispersed as minute globules in copper, which may form macroscopic lakes of lead causing serious weaknesses. Nevertheless, the Silla Dynasty’s artisans would put considerable amount of lead to refine copper, and they may empirically know this when they were casting bronze. 

Abbreviations

XRF

X-Ray Fluorescence

OM

Optical Microscope

SEM

Scanning Electron Microscope

EDS

Energy Dispersive X-ray Spectroscopy

FIB

Focused Ion Beam

TEM

Transmission Electron Microscopy

Declarations

Availability of data and materials

 Data is available within this manuscript and upon request.

Competing interests

 The authors declared that they have no conflicts of interest to this work.

Funding

This work was supported under the policy research program(2019R1F1A1060043) managed by National Research Foundation of Korea(NRF). The views and opinions expressed in this article are those of the authors and do not necessarily reflect the official policy or position of NRF. 

Authors' Contributions

Choi contributed to the conception of the study, performed the experiment, data analyses, wrote the manuscript.

Acknowledgements

It is an honor to have the opportunity to study the excavated artifacts of the Research Institute of Buddhist Cultural Heritage. Although it is my job to do research, I am very thankful to the colleagues who helped me in this project.

References

  1. Craddock PT, Meeks ND. Iron in Ancient copper. Archaeometry. 1987;29:2:187–204.
  2. Il yeon. Tower statue 4, Hwangnyongsa Bell, Bunhwangsa bronze statue.BongDeoksa Bell. The Heritage of the Three States. 1281;33.
  3. Yun J. Archaeometallurgical Study on Artifacts Excavated from Bronze Workshops of Unified Silla. Master degree. Hongik graduated university; 2013.
  4. Kwon H. The study on Ancient copper smelting. master degree. Kukmin graduated university; 2001.
  5. Choi Ju. The Feature of the Korean Bronze Age From the Metallurgical Viewpoint. BUMAâ…£(Proceedings of the Fourth International Conference on The Beginnings of the Use of Metals and Alloys); 1996. p. 25–46.
  6. Lee Y. Investigation results of Whangnyongdong Whangnyongsaji in Gyeongju. Academic Conference book. Research Institution of Buddhist Cultural Heritage; 2021. p.3–25.
  7. Kim D. Manufacturing techniques and development of small gilt-bronze Buddhas in Silla, Gilt-bronze Buddha of Silla. 2011. p.226–255.
  8. Lim S, Kim DS, Park D. Sunggun Kang. Study for the Film coating Techniqur of Gilt Bronze Artifacts from Mirksa Temple. Conserv Stud. 1993;14:45–76.
  9. Science History Institute. 2017. https://www.sciencehistory.org/historical-profile/jons-jakob-berzelius. Accessed 29 January 2023.
  10. Hata ryosuke. Nishimura, Tadahisa and Umetsu yoshiaki. Treatment of Selenium-bearing materials and removal of Selenium from aqueous solutions. Bull Inst Adv Mater Process Tohoku Univ. 1997;52:145–59.
  11. Shibata E, Nakamura T. Situation and problem of slag/matte separation in copper smelting. J MMIJ. 2013;129:171–6.
  12. Choi JE. Microstructure of Surface Corrosion Layer of Bronze Coin Headongtongbo Fabricated in the Koryo Period (11th Century). J Japan Inst Metals. 2011;75:8:438–44. 10.2320/jinstmet.75.438.
  13. Tada E. Nishikata, Atsushi. Corrosion of metal structure. Chem Educ. 2017;65:12.
  14. David A, Scott. Copper and bronze in Art: Corrosion, Colorants, Conservation. Getty Publications; 2002. p.95.
  15. Yamaguchi K. Utilization of Thermodynamics in Extractive Metallurgy. Netsu Sokutei. 2009;36:5.
  16. Choi JE, Kitada M. Microstructure of Ancient Koryo Bronze Bottle. J Japan Inst Met Mater. 2015;79:1:29–33. 10.2320/jinstmet.j2014032.
  17. Choi JE. A study of the Microstructure and Impurity Characteristics of cast Bronze in Koryo Period. J Conserv Sci. 2016;32:3:313–20.
  18. Yokota M. Sugaya, Fuminori. Haruhisa, Mifune, Kobori, Yoshiyuki, Shimizu, Katsuro, Nakai, Kazuo, Miyahara, Shinichi and Shimizu, Yasuji. Scientific Investigations Focusing on the Appearing Shapes of the Pure Copper Lumps Observed in or on the Ancient Bronze Mirrors. J Japan Inst Metals. 2002;66:7:708–14. 10.2320/jinstmet1952.66.7_708.