2.1. Sampling and sample preparation

Eight ancient coins from the Bagan Archaeological Museum in Myanmar were analyzed in this research. The eight coins exhibited difference surface colors, including silver, red, and green. Among them, eight showing differences in manufacturing technology and microstructure were selected and described in the paper (Fig. 2).

Samples were taken in the smallest possible size to ensure that no damage was caused to the original shape of the artifacts. Sampling was carried out using a diamond circular saw blade mounted on an electric motor tool. After the collected samples were packed individually in plastic bags to prevent them from mixing with each other, object information was recorded.

After checking the name and number of each artifact, samples were mounted in epoxy resin for analysis. The sample surfaces were polished using a polishing machine with silicon carbide papers in the order of grit size #800, #1000, #2000, and #4000, finishing with a 1 μm diamond suspension. The surfaces of the smoothly polished sample surfaces were etched in a solution of ethyl alcohol (120 ml), hydrochloric acid (30 ml), and ferric chloride (FeCl₃, 10 g) for several seconds. After this process the samples were cleaned with water and dried.

Fig. 2. Collected ancient coins from Bagan, Myanmar.

2.2. Method of analysis

The microstructures of the samples were observed using a reflected light microscope (DMRBE, Leica, DEU) to identify the manufacturing techniques applied to the ancient coins. The entire microstructure was observed at low magnification (x25 and x50) and specific regions were observed at high magnification (x100 and x1000). If a higher magnification (above x1000) beyond the capacity of the optical microscope was required, sample surfaces were carbon-coated and observed using a scanning electron microscope (JSM-IT300, JEOL, JPN).

The chemical compositions of the entire sample observed at low magnifications using energy dispersive X-ray spectroscopy (Oxford, UK) attached to the scanning electron microscope were analyzed to identify the overall alloy composition. The results of an area analysis performed at least five times at low magnifications (x200) using the scanning electron microscope were averaged. Inclusions, lead particles and grain boundaries providing critical clues to understanding the manufacturing techniques were observed at high magnifications. The sample surfaces were carbon-coated to obtain conductivity and analyzed at an accelerating voltage of 20 kV with a working distance of 10 mm.

3.1. Types and use of ancient coins

Coins No. 1 and 2 in Fig. 2 are silver coins produced and used during the Pyu period of Myanmar. The surfaces of these silver coins are covered with a dark gray corrosion layer, but in areas where the corrosion layer has worn off, a silver-colored surface can be observed. These coins have a typical circular shape, and both sides feature designs symbolizing currency. According to previous studies on coins excavated from Pyu ancient cities, Pyu period coins are classified into large coins with a diameter of 30–34 mm and small coins with a diameter of 17–18 mm. The two silver coins, with a diameter of 30 mm, fall into the category of large coins. Although they share the same basic form, they exhibit different morphological characteristics in the designs engraved on their front and back. Upon closer examination of the designs on these two silver coins, both display a crown-like design, yet each contains a unique internal pattern, indicating that they are distinct coins. Specifically, Coin No. 1 features a crescent-shaped pattern accompanied by circular dots, while Coin No. 2 only has circular dots. Thus, while these two silver coins are similar in size and thickness, their differing designs may suggest variations in their circulation period or intended use.

Coin No. 3 in Fig. 2 is a bronze coin produced and used during the Pyu period. The surface of this oxidized coin appears brown, and parts where the oxidation layer has worn off reveal a yellowish surface. This bronze coin shares a similar basic form and size with the silver coins. Notably, the crown-like design engraved on the front of the bronze coin resembles that of the silver coins, indicating that the bronze coin was used concurrently with the silver coins during the Pyu period. This also suggests that the currency standard varied depending on the metal used.

Coins No. 4 to No. 8 in Fig. 2 are arsenic bronze coins that are locally referred to as “silver coins” in Myanmar. The coins resemble round Go stones, and there are no signs of surface finishing after casting or inscriptions indicating a monetary unit. Therefore, although known as coins, it is unclear what purpose they served or the extent of their monetary value. However, it is known that these coins were used during the Bagan Dynasty. The surfaces of the coins are covered with a green corrosion layer, but the cross-sections cut for analysis appear silver in color.

Basic information, such as the size, surface color, excavation site, and production period of each analyzed ancient coin, is summarized in Table 1.

3.2. Elemental composition of ancient coins

The SEM-EDS elemental composition analysis results of ancient coins are shown in Table 2. According to the analysis results, the eight coins are broadly classified into silver, copper, and arsenic bronze coins based on their major constituent elements.

The two silver coins showed no detectable elements other than silver, indicating that the Pyu period possessed highly developed silver refining techniques. On the other hand, the copper-based coins revealed various metals, such as tin, zinc, lead, and iron. Notably, zinc, which was found in amounts exceeding 14 wt%, was the most prominent alloying metal after copper. The inclusion of zinc at over 14 wt% suggests that zinc was deliberately added to the alloy during casting. As the alloy ratio of zinc increases, the color of the copper alloy approaches a gold hue, indicating that zinc was likely used to alter the color of the copper coins. Although the silver and copper coins produced during the Pyu period have similar shapes, their compositions differ significantly. Therefore, it is inferred that silver and copper coins were produced using similarly shaped molds, while the circulation value of the coins varied according to the metal quality.

The arsenic bronze coins contain 3–12 wt% arsenic and a high lead content of 14–39 wt% along with copper as the base material. The alloy ratios of the arsenic bronze coins, which are cast as segregation alloys of copper-arsenic-lead-antimony, vary from coin to coin, showing no consistent alloying pattern. While the arsenic content differs significantly between arsenic bronze coins No.1 and No.5, which both have a diameter of 42.0 mm, arsenic and lead contents are similar between arsenic bronze coins No.1 and No.4, despite their large size difference. Therefore, the alloy ratio does not vary according to coin size.

3.3. Metallurgical microstructure

3.3.1. Silver coins

Fig. 3a is an optical micrograph of a small sample taken from the side of silver coin No. 1, showing grain refinement near the surface and coarser grains toward the center. Fig. 3b is an optical micrograph of the magnified upper right side in Fig. 3a, showing the difference of sizes between grains and also indicating that the internal microstructure has no corrosion layers and cracks with good conservation condition. Fig. 3c is an EDS spectrum analyzing the microstructure observed in Fig. 3b, and the analysis of the entire microstructure reveals no elements other than silver. Fig. 3d is the magnified optical micrograph of the sample taken from silver coin No. 2, showing the coin has been preserved soundly. Fig. 3e is an optical micrograph magnifying Fig. 3d, showing relatively large grains with twins formed across the grains. And the analysis of the entire microstructure using EDS mapping reveals no other elements than silver (Fig. 3f). The concentrated refinement of the microstructure near the surface of the silver coin suggests the possibility of surface finishing after casting ^[3].

Fig. 3. Microstructure of silver coins; (a) Optical micrograph of silver coin No. 1, (b) magnified region of (a), (c) EDS spectrum of silver coin No. 2, (d) optical micrograph of silver coin No. 2, (e) magnified part of (d), (f) SEM-EDS X-ray mapping of silver coin No. 2.

3.3.2. Bronze coin

Fig. 4a is an optical micrograph of the entire microstructure, revealing lead particles densely distributed between the fine grains. Fig. 4b is an electron micrograph magnifying Fig. 4a, showing the presence of three different phases. These phases were analyzed using EDS, and the results are presented in Table 3. Point 1 in Fig. 4b is the α phase, containing 81.31 wt% copper, 1.01 wt% tin and 14.86 wt% zinc. Point 2 is the δ phase and contains 25.70 wt% tin and 5.07 wt% zinc besides copper. Point 3 is lead, which contains 6.05 wt% copper. Fig. 4c is an optical micrograph magnifying the red copper oxide layer that has densely formed on the corrosion layer of the bronze coin’s surface. This oxide layer is a result of de-zincification, a typical type of brass corrosion ^[4,5]. This coin was made of a brass alloy containing copper, zinc, tin, lead and iron by casting.

Fig. 4. Microstructure of bronze coin; (a) optical micrograph, (b) SEM image showing magnified lead particles and the δ phase (SEM), (c) magnified corrosion layer of de-zincification.

Table 3. SEM-EDS elemental composition of individual phases in the bronze coin

Coin	Analysis location	Elemental composition (wt%)
		Cu	Sn	As	Pb	Zn	Ni	Sb	Fe	Al	Cl	O	Total
Bronze coin	Point 1	81.31	1.01	n.d.	n.d.	14.86	0.40	n.d.	2.03	0.39	n.d.	n.d.	100.00
	Point 2	60.11	25.70	n.d.	1.51	5.07	0.72	1.84	0.78	n.d.	0.64	3.63	100.00
	Point 3	6.05	n.d.	n.d.	80.43	n.d.	n.d.	n.d.	n.d.	n.d.	9.26	4.26	100.00

3.3.3. Arsenic bronze coins

As confirmed by the compositional analysis results presented in Table 2, the five arsenic bronze coins share the common feature of being alloyed with high amounts of arsenic and lead, although the alloy ratios are not consistent. Among the five arsenic bronze coins, the microstructures of three coins (Nos. 4, 5, and 8 in Fig. 2), which show significant differences in arsenic and lead content, were analyzed.

Fig. 5a is an optical micrograph showing the entire microstructure of arsenic bronze coin 1 (No. 4 in Fig. 2), with lead appearing as dark gray dots distributed throughout the microstructure. Fig. 5b, an electron micrograph, shows the sizes and shapes of the lead particles more clearly. Fig. 5c is an electron micrograph magnifying the center of Fig. 5a, showing four phases in white and gray in addition to lead particles. These phases were analyzed using EDS and are shown in Table 4. Point 1 in Fig. 5c is the α phase, containing 88.57 wt% copper and 7.14 wt% arsenic. Point 2 contains 10.96 wt% arsenic and 2.19 wt% antimony besides copper. Point 3 contains mainly lead, and Point 4, which is an inclusion, is sulfide containing copper, iron and sulfur. The microstructure of arsenic bronze coin No.1, which contains a high amount of arsenic and lead, shows numerous lead particles of varying sizes, with a broad area of microstructure containing more than 10 wt% arsenic.

Fig. 5d is an optical micrograph showing the entire microstructure of arsenic bronze coin No. 2 (No. 5 in Fig. 2), which has the lowest lead content. Spherical lead particles of varying sizes are formed over a relatively small area. Fig. 5e is an electron micrograph observed at a similar magnification to Fig. 5d, showing the size and shape of the lead particles more clearly. Fig. 5f is a magnified electron micrograph of a portion in Fig. 5d, where, in addition to lead particles, four different phases appearing in white and gray can be observed. These phases were analyzed using EDS and are shown in Table 4. Point 5 in Fig. 5f is the α phase, containing 91.52 wt% copper and 8.48 wt% arsenic. Point 6 is the γ phase and contains 30.25 wt% arsenic and 4.84 wt% antimony besides copper. Due to the segregation of casting components, the γ phase is rarely present in the microstructure of arsenic bronze ^[6]. Point 7 in Fig. 5f, representing lead, contains 5.23 wt% copper and a small amount of 0.77 wt% iron. Point 8, which is an inclusion, is a sulfide containing copper, iron and sulfur. The microstructure of the arsenic bronze coin No. 2, which contains a high amount of arsenic over 10 wt% but has the lowest lead content, shows the formation of a γ phase containing antimony.

Fig. 5g is an optical micrograph showing the entire microstructure of the arsenic bronze coin No. 5 (No. 8 in Fig. 2), with lead distributed throughout the microstructure in a dark gray and irregular form. The electron micrograph in Fig. 5h provides a clearer view of the size and shape of the lead particles. Fig. 5i is an electron micrograph magnifying the upper left side of Fig. 5g, where four different phases are observed and analyzed using EDS (Table 4). Point 9 in Fig. 5i is the α phase, containing 91.17 wt% copper and 5.24 wt% arsenic. Point 10 in Fig. 5i is the microstructure consisting of 57.90 wt% copper, 3.27 wt% arsenic and 36.46 wt% antimony. Point 11, indicating lead, contains a trace of copper of 3.96 wt%. Point 12, which is an inclusion, is sulfide containing copper and sulfur. Arsenic bronze coin No. 5 contains a small amount of arsenic but was cast with a high lead content, reflecting its alloy characteristics.

Fig. 5. Microstructure of arsenic bronze coins; (a) optical micrograph of arsenic bronze coin No. 1, (b) magnified electron micrograph of (a), (c) magnified region of (a), (d) optical micrograph of arsenic bronze coin No. 2, (e) magnified electron micrograph of (d), (f) magnified region of (d), (g) optical micrograph of arsenic bronze coin No. 5, (h) magnified electron micrograph of (g), (i) magnified region of (g).

Table 4. SEM-EDS elemental composition of individual phases in the arsenic bronze coins

Coins	Analysis location	Elemental composition (wt%)
		Cu	Sn	As	Pb	Zn	Ni	Sb	S	Fe	Co	O	Total
Arsenic bronze coin 1	Point 1	88.92	n.d.	7.14	1.53	0.11	n.d.	1.74	0.06	0.50	n.d.	n.d.	100.00
	Point 2	83.46	n.d.	10.96	3.18	0.09	n.d.	2.19	0.04	0.08	n.d.	n.d.	100.00
	Point 3	3.55	n.d.	n.d.	91.09	0.04	n.d.	n.d.	n.d.	n.d.	n.d.	5.32	100.00
	Point 4	77.22	n.d.	n.d.	0.16	n.d.	n.d.	21.59	1.03	n.d.	n.d.	100.00
Arsenic bronze coin 2	Point 5	91.52	n.d.	8.48	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	100.00
	Point 6	64.91	n.d.	30.25	n.d.	n.d.	n.d.	4.84	n.d.	n.d.	n.d.	n.d.	100.00
	Point 7	5.23	n.d.	0.56	86.83	n.d.	n.d.	n.d.	n.d.	0.77	n.d.	6.61	100.00
	Point 8	72.49	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	23.93	3.58	n.d.	n.d.	100.00
Arsenic bronze coin 5	Point 9	91.17	n.d.	5.24	2.84	n.d.	n.d.	0.75	n.d.	n.d.	n.d.	n.d.	100.00
	Point 10	57.90	n.d.	3.27	n.d.	n.d.	1.31	36.46	n.d.	n.d.	1.06	n.d.	100.00
	Point 11	3.96	n.d.	n.d.	90.08	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	5.96	100.00
	Point 12	77.89	n.d.	n.d.	n.d.	n.d.	n.d.	n.d.	22.11	n.d.	n.d.	n.d.	100.00

4. DISCUSSION

Ancient coins from the Pyu period and the Bagan Dynasty have been excavated in the Bagan region of Myanmar. The silver and bronze coins from the Pyu city-states feature similar patterns on both sides and are uniform in size, indicating that they were produced with standardized manufacturing techniques. In contrast, the coins from the Bagan Dynasty lack patterns and vary in size, suggesting that they were not standardized. Thus, during the Bagan Dynasty, a period later than the Pyu era, non-standardized coins were produced and used. This suggests that the arsenic bronze coins from the Bagan period may have been used as intermediate materials for manufacturing metal products. Notably, arsenic bronze, which was not used during the Pyu period, was employed as an alloying material in Bagan period coins. Arsenic bronze generally refers to a binary copper-arsenic alloy, but it may contain more than 2% arsenic and other elements such as tin, antimony, and lead, resulting in ternary or quaternary alloys such as arsenic-tin bronze, antimony-arsenic bronze, lead-arsenic bronze, and lead-tin-arsenic bronze. Therefore, in a broader sense, arsenic bronze includes not only binary copper-arsenic alloys but also various copper-arsenic alloys with diverse compositions ^[4]. Alloys of copper and arsenic were produced and used across the Near East and Europe from as early as 4000 BCE through the Late Bronze Age. Arsenic copper artifacts excavated from the Susa site in Iran are considered the earliest examples of copper and arsenic alloys. It is known that ancient bronze alloying technology advanced in the order of native copper, smelted copper, arsenic bronze, and then tin bronze, with arsenic bronze being used before tin bronze in most advanced ancient civilizations ^[7,8].

From an alloy design perspective, unlike tin, it is very difficult to predetermine or control the amount of arsenic in the alloy. However, adding arsenic creates bronze that is harder than pure copper, which highlights the significant role of arsenic in the early stages of the Bronze Age ^[9]. Generally, if the arsenic or zinc content in a bronze alloy is below 1%, it is considered an impurity from the ore without particular significance. In copper smelting, certain copper ores can naturally incorporate up to 4% arsenic and 3% tin, along with other elements such as lead, antimony, and bismuth ^[10]. However, since the arsenic bronze coins from Bagan, Myanmar, contain over 7 wt% arsenic, it is likely that arsenic was intentionally added. Additionally, the arsenic bronze coins from Bagan were produced in the 13th century when tin alloying technology had already become well established. Because arsenic is readily volatilized and oxidized at high temperatures, its controlled incorporation into molten copper is technologically challenging. Consequently, in some archaeometallurgical contexts, the use of speiss has been proposed as a method to introduce arsenic more stably into copper alloys ^[11]. In previously documented cases where the use of speiss has been directly discussed, iron consistently constitutes a major component, and Fe-As intermetallic compounds or iron-rich phases have been identified ^[12,13]. However, in the arsenic bronze coins of Myanmar, despite the relatively high arsenic contents, the iron concentration is extremely low. This compositional characteristic suggests that the use of speiss as an arsenic source in the production of these arsenical bronze alloys is relatively unlikely.

It is unclear why arsenic bronze coins, which were absent in earlier Pyu city-states sites, were produced during this period. However, since these arsenic bronze coins are still referred to as ‘silver coins’ today, it is possible that the bronze alloy was made to achieve a silver-like appearance. To investigate the possibility that arsenic bronze coins were used as intermediate materials for bronze artifact production, the compositional analysis results of bronze artifacts excavated from Bagan in the same period were compared (Table 5). No arsenic was detected in the seven small bronze statues and five bronzeware excavated from the Bagan site that have been analyzed to date ^[14,15]. Therefore, there is no direct evidence that these arsenic bronze coins were used as intermediate materials for bronze production. Nevertheless, the varying sizes of the coins and their alloy composition, which includes significant amounts of specific metals such as arsenic and lead, suggest the potential for use as intermediate materials.

5. CONCLUSIONS

This study represents an early systematic attempt to analyze ancient metal coins from the Bagan site in Myanmar, providing significant insights into the exact composition and characteristics of coins that served as major currency in the Pyu and Bagan periods. Coins from the Pyu city-states are primarily classified as silver and copper coins, while those from the Bagan Dynasty, colloquially called ’silver coins’ in Myanmar, are round and vary significantly in alloy composition. The Pyu silver coins and reddish-brown copper coins, both 3 cm in diameter, share similar front and back engravings. Analysis confirms that the Pyu silver coins are made of 100% pure silver, while the copper coins, although similar in shape, are brass alloys containing Cu (74.26 wt%), Zn (14.24 wt%), Sn (4.29 wt%), Pb (4.58 wt%), and Fe (2.35 wt%).

The Bagan Dynasty coins, taking an ingot-like form rather than traditional coin shapes, lack uniformity in size and are composed of arsenic bronze alloys containing Cu-As-Pb-Sb without Ag but with significant amounts of As and Pb. While green corrosion layers are present on the surface, the cross-sections exhibit a silver tone, likely due to arsenic in the alloy, which would have initially given the coins a silvery appearance and thus explains their designation as ‘silver coins.’ The presence of sulfur in the non-metallic inclusions of the arsenic bronze coins further suggests that the raw copper may have been derived from sulfur-containing copper ores. Overall, this research offers a foundational understanding of the materials and techniques employed in ancient Myanmar’s currency production, marking an important step in historical metallurgy studies.