Introduction

Weapon played an important role in the development of human society. They were also an important part of the splendid culture of ancient China, which reflected the culture, history, science, technology, and skills of a certain period. From archeological excavation, it was found that some weapons were made of different types of metals or alloys, or the same alloy with different ratios. Some of them were gifts or ritual objects, such as a bronze sword with jade handle and iron blade (Jiang 1999) and a knife with gold handle (Tian and Lei 1993). Most weapons were for equipping army, such as bronze battle-axes with iron blade (Gettens et al. 1971; Yuan and Zhang 1977; Li 1976), bronze paring knifes with iron blade, bronze dagger-axes with iron blade (Jiang 1999), and even the bi-metallic bronze swords (Lian and Tan 2002; He 1990; Chen 1981; Ding et al. 2012).

It is well known that bronze weapons have been widely used during Chinese dynasties Xia (2146 BC–1675 BC), Shang (1675 BC–1029 BC), and Zhou (involving 1029 BC–771 BC of the West Zhou: and 770 BC–221 BC of the Spring-Autumn and Warring States Periods). In addition to regular bronze swords, up to now, several bi-metallic bronze swords have been unearthed or collected in the most area of China, which achieved the highest performance and was considered to be one representative type of the advanced technologies. The bi-metallic bronze sword was invented based on the following principle: the spine was made of low-tin (Sn) bronze with good flexibility, and the blade was made of high-Sn bronze with high mechanical strength and hardness, which consequently resulted in a stronger lethality and longer service life. This sword was usually manufactured by using a two-step casing technology, and connected by mortise and tenon joint in cross-section, which bore the rigidity and flexibility simultaneously (Lian and Tan 2002).

In general, ancient Chinese bronze Ge was produced in a one-step casting by using piece-mold casting technique with two pieces of mold (Chen 2010). The manufacturing process was as follows: (1) prepare the molds of Ge; (2) pour the molten bronze into the molds; (3) after solidification, move the Ge out of the mold and polish it, and sharpen the blade. Many researches (Zhao et al. 2012; Yang et al. 2013; Jia et al. 2011; Sun et al. 2011) revealed that the Chinese bronze Ge was made of Cu–Sn alloy or lead (Pb) added Cu–Sn alloys (simplified as Cu–Sn+Pb) with high Sn contents in a range of 10–20 wt%.

Due to the difference in function and usage between Ge and sword, few bi-metallic bronze Ges were discovered hitherto. And study on its manufacturing process has not been reported. In this work, two bi-metallic bronze Ges were occasionally discovered, which was discovered near the Longwang Mountain in Huanggang City, Hubei Province, China. According to the systematical study on the chemical compositions and microstructures, we get more acquainted with the making process of bi-metallic bronze Ge, and its historical background and the motive in making these objects were also discussed. It is expected to give deep understanding of the manufacturing process of bronze weapons during the Warring States Period, and provides new scientific and technical evidences for archeology and metallurgy history.

Archeological backgrounds

According to the record of the museum, these two bi-metallic bronze Ges, as shown in Fig. 1, were discovered from a brick factory located around Longwang Mountain at Huangzhou District of Huanggang City, Hubei province, China, in 1995, when the workers were digging soil, and now collected in the Huangzhou Museum with the record no. 74 and no. 76. Actually, they are not unearthed from a known cemetery.

Fig. 1
figure 1

Images of the bronze Ges. a No. 74; b no. 76

Full size image

Huangzhou District is located in the eastern part of Hubei Province, on the north bank of the middle reaches of the Yangtze River. It is a historical and cultural area, which belonged to the Chu State during the Warring States Period. There are many ancient cemeteries around the Longwang Mountain, such as Longwang Mountain Cemetery, Wangjiachong Cemetery, Guoerchong Cemetery, and Caojiagang No. 5 Tomb. Most of them are identified as the Chu Tombs in the middle and late times of the Warring States period (Wu and Hong 2000; Huang 2001; Wang and Wu 1983; Hubei Provincial Institute of Cultural Relics and Archaeology 1993).

In terms of appearance, the no. 74 Ge exhibits the following features: (1) the Yuan is slender, long, and slightly curved; (2) the mid-ridge is concave; (3) the Hu is narrow; the Na is wide and has hollowed-out transmuted phoenix patterns. And the no. 76 Ge shows the similar shape, but without Na. According to archeological typology, these two Ges meet the standard style of Chu Ge produced during the Warring States Period; they can be dated to the Warring States Period (475 BC–221 BC). Especially, the transmuted phoenix pattern is representative pattern prevalent during the Warring States Period. The names of different parts of Ge are shown in Fig. 2.

Fig. 2
figure 2

The names of different parts of Ge

Full size image

In fact, there were several similar Ges unearthed in the Chu tombs of the Warring States Period. For example, one Ge, called type B Ge (M5: 2), was from the Wangjiachong Tomb 5, which is adjacent to the brick factory where the Ges were found (Huang 2001). And other three Ges were also unearthed from Chu tombs in the Jiangling County of Hubei Province, which also belonged to the Chu State during the Warring States Period with about 300-km distance to Huangzhou District, such as one (maned I type Ge (M1: 288 + 295)) from the Tianxingguan Tomb 1 (The Jingzhou District Museum, Hubei Province 1982), one (M43: 21) from the Jiudian Tomb 1 (Hubei Provincial Institute of Cultural Relics and Archaeology 1995), and one (M159: 25) from the Yutaishan Tomb 1 (Hubei Provincial Institute of Cultural Relics and Archaeology 1984).

When comparing with the regular Ges from other Chu tombs, the present no. 74 Ge has an exquisite Na which shows a solid stripe molding structure, and exhibits a very high level of hollowing casting technique. In fact, this type of Ge with a hollow pattern is very rare, and only found in higher rank tombs, such as the Chaixu Ge (a king of the state of Yue) (Wu 2012) and the King Shiye Ge (Wu 2012). This suggests that the present Ge was made for a special person as the symbol of status or power, not for equipping the army in combat.

Experimental methods

The samples were cut along the edge of the broken parts, and then mounted, grounded, and polished according to standard metallographic procedures. The main processes included the following: (1) mechanical grinding to 2000 grit, successive polishing to 0.1 μm, ultrasonic cleaning in ethanol (analytical grade ethanol) and drying; (2) the polished sections were etched with a FeCl3 + HCl + C2H5OH solution (3% FeCl3) for revealing the microstructures. Microstructural observations were carried out on the polished and etched sections by using an optical microscope (Leica DM2700, Germany), a three-dimension microscope (Leica DVM6, Germany), and a scanning electron microscope (Phenom XL, Netherlands). The chemical compositions were measured on the polished and un-etched sections in the SEM by using an energy-dispersive spectrometer (EDS) for quantitative analysis, operated at 15 kV, and the unbroken parts were measured by a portable X-ray fluorescence (p-XRF) spectroscopy (Niton XL3t 950, USA). The radiograph was obtained by using X-ray radiography (XXQ-2005, China), which is vital in understanding the internal structures. The microhardness was measured by Vickers hardness tester (HXD-1000TMB/LCD, China).

Results and discussion

Surface appearances and X-ray radiographic morphologies

Figure 3 shows the surface appearances of the bronze Ges by using optical microscopes in low magnifications. Obviously, they were consisted of two different metals or alloys. For the no. 74 bronze Ge, the concave mid-ridge and Lan showed jade green color and tough surface with corrosive sediments, while the other parts were smooth with gray green color, as shown in Fig. 3 b. In addition, there was an obvious seam between Yuan and Na, as shown in Fig. 3 c. The no. 76 bronze Ge exhibited the similar surface appearance. From the cross-section of Yuan, as shown in Fig. 3 g, it was found that the internal portion was reddish brown, while the external portion was gray green, which revealed the difference of alloys. The two parts were connected via a mortise and tenon joint structure.

Fig. 3
figure 3

Macroscopic appearances of the bronze Ges. (a) Overall image of no. 74; (b) partial image of Yuan; (c) junction of body and handle; (d) top of the junction; (e) sprue trace of the junction; (f) overall image of no. 76; (g) cross-section of Yuan

Full size image

Figure 4 shows the X-ray radiographic images of the bronze Ges. They also revealed the different colors for different parts of the bronze Ges due to the different alloys. It has been known that the attenuation of X-ray is proportional to the density of the substance, and is proportional to the cube of atomic number (Jackson and Hawkes 1981). Because of the difference of attenuation of X-ray, the bronze with higher Sn content shows a bright color, while lower Sn content shows the dim color. Therefore, it could be further confirmed that the portions of the mid-ridge and Lan contained higher Cu element, while the other parts contained higher Sn element. Obviously, these belonged to a kind of the bi-metallic bronze Ge.

Fig. 4
figure 4

X-ray radiograph images of the bronze Ges. a No. 74; b no. 76

Full size image

Microstructures and chemical compositions

Figure 5 and Fig. 6 show the optical metallographic images and SEM morphologies of the cross-section of two bi-metallic bronze Ges. Obviously, they exhibited the similar characters, i.e., typical dendrites of casting. Table 1 lists the chemical compositions of different parts. The mid-ridges were made of pure copper. The blades were made of the Cu–Sn+Pb alloys with high Sn contents.

Fig. 5
figure 5

Cross-section metallographic microstructures of the bronze Ges. (a) Overall image of no. 74; (b) mid-ridge; (c) junction of mortise and tenon; (d) blade; (e) overall image of no. 76; (f) mid-ridge; (g) junction of mortise and tenon; (h) blade

Full size image
Fig. 6
figure 6

Cross-section SEM morphologies of the bronze Ges. (a) Low-magnification image around the junction of no. 74; (b) mid-ridge; (c) blade; (d) junction of mortise and tenon; (e) low-magnification image around the junction of no. 76; (f) mid-ridge; (g) blade; (h) junction of mortise and tenon

Full size image
Table 1 Chemical compositions of different parts of the bronze Ges (wt%)
Full size table

As to the mid-ridges, the structure mainly consisted of coarse dendrites. Some sulfate particles at the grain boundaries were observed, which were the residual impurities of the copper ore. The larger dendrite grain sizes indicated that this portion experienced an annealing process or other heat treatment, as shown in Fig. 5 b and Fig. 5 f.

However, for the edge and the blade, it was composed of fine α-Cu dendrites, (α+δ) eutectoid structures, and isolated Pb particles, as shown in Fig. 5 d and Fig. 5 h. Table 1 lists the corresponding point compositions. It is commonly known that Pb has no solid solution relationship with Cu and Cu–Sn alloys, and only exists in a separation or dissociation state. Our previous work demonstrated that the dispersed Pb particles exhibited functions of refining the casting microstructures and reducing the casting defects (Pan et al. 2007). In addition, according to the Cu–Sn phase diagram, the maximum Sn solid solubility in Cu–Sn binary alloy is 15.8 wt%, and when it exceeds this value, the intermetallic δ-Cu phase (Cu31Sn8), which contains 32.6 wt% Sn and is hard and brittle, begins to precipitate from the matrix. Therefore, these fine microstructures and numerous δ-Cu phases resulted in high strength and hardness of the bronze. The measurements showed that the hardness value of the mid-ridge of no. 74 Ge was 59.90 Hv, and that of the blade was 139.24 Hv, which revealed that the hardness of the blade was twice as hard as the mid-ridge. Similarly, the hardness variation of no. 76 was close to no. 74.

When the connected portion of the ridge and the blade was carefully examined, a transition zone was observed. Its chemical compositions were measured by using EDS with line scanning and point scanning modes, as shown in Fig. 6 d and Fig. 6 h. The results revealed that the transition zone showed a mechanical mixture with variant phases and chemical compositions, which meant that no new microstructure and phase were formed, and only had a short-distance elemental diffusion of Sn in the Cu matrix. Since the melting point (1083.4 °C) of Cu is higher than that of the Cu–Sn alloy, the molten Pb added Cu–Sn alloys of the second casting step could not melt the solid Cu at the interface.

Another interesting discovery was that the microstructure in the pure copper side of the transition zone was similar to that of welding fusion zone in the heat-affected zone (HAZ), as shown in Fig. 6 d and Fig. 6 h. In general, in welding metallurgy, the welding fusion zone refers to a transition zone between the weld metal and HAZ in the base metal in a welded joint, which is also called the semi-melting zone. It is a narrow zone, and the local melting occurs along the grain boundaries in the area adjacent to the melted weld zone. Its heating temperature is between the solidus line and liquidus line, which results in the formation of coarse grains, the heterogeneous chemical compositions, and microstructures (Huang et al. 2013; Pan 2000).

As to the two-casting process, the thermal effect on microstructure of the mid-ridge during the second casting was similar to that of the weld fusion zone on the base metal, which also produced the similar microstructure. In addition, because the thermodynamic interface could not be formed, it was impossible to have a metallurgical connection between mid-ridge and blade. In order to solve this problem, the ancient technicians wisely designed a mortise and tenon structure to form a staggered and mutually constrained stable structure and strengthened the bonding between the mid-ridge and the blade.

To further confirm connective style of the Yuan and Na of the no.74 Ge, the chemical compositions of Yuan and Na and the connected portion were measured by using a portable X-ray fluorescence (XRF) spectroscopy. Five points were measured in each portion, as listed in Table 2. Obviously, they showed the different compositions. As for the microelements, the amount of Fe content in the junction was much higher. Bi content in joint part was 0.12 wt%, while there was no Bi element in the mid-ridge and Na. Obviously, the chemical compositions of the connection part were totally different from the other parts.

Table 2 The XRF compositions of the mid-ridge, joint, and Na of no. 74 Ge (wt%)
Full size table

Possible manufacturing process

As the experimental examinations above, the bi-metallic bronze Ges exhibited the similar structures to the well-known bi-metallic bronze swords. According to the archeological discoveries, the sword-making technology was established and boomed in the area of the Wu and Yue State, and also exhibited the highest skill (Xiao 1991, 1996). It has been known that the bi-metallic bronze swords emerged in the areas of the early Warring States Period or earlier in the Wu and Yue State. The Huangzhou District was adjacent to the Wu and Yue State territory, so its culture and technological development were influenced by the Wu and Yue State. Therefore, it suggested that the manufacturing technology of the bi-metallic bronze weapons was probably introduced from the Wu and Yue State.

From observation of the Yuan cross-section, obviously, the mid-ridge and blade were connected via a mortise and tenon joint structure. And the dendrites of the mid-ridges were coarse, indicating that the ridge was subjected to a double heat treatment, while the dendrites of the blade were fine, indicating it was a one-step casting. In addition, the microstructure in the pure copper side of the transition zone was similar to that of the welding fusion zone in the HAZ. Hence, we proposed that the bi-metallic bronze Ges were fabricated by a two-step casting.

In addition, the XRF results showed the chemical compositions of connection Yuan and Na were totally different from the other parts. The presence of bismuth (Bi) in the joint revealed that the joint and the other parts of the Ge were made of different copper mines, which demonstrated that the process of connection was in a separate step. What is more, there was a seam between the junction and Na, as shown in Fig. 3 c. The metal at the joint was laminated on the protruding part of Yuan, as shown in Fig. 3 c. Besides, there was a trace of mold seam of the separated casting with composited molds on the top of the joint, while there was no seam on the adjacent Na and Yuan, as shown in Fig. 3 d. The sprue trace of this part was observed, as shown in Fig. 3 e. These phenomena implied that this was an another casting step on the previous two parts, which was the typical soldering style (Zhang 2018). Therefore, we deduced that the two parts were connected by soldering.

Based on the above experimental analysis and observations, it could be confirmed that the manufacturing process of the bi-metallic bronze Ges was as follows: firstly, cast the mid-ridge and Lan with copper; and then, put the mid-ridge and Lan into the molds for making the cutting edge and blade, and cast the edge and blade with Cu–Sn+Pb alloys; at last, solider the body together with Na by using Cu–Sn+Pb alloy.

The special choice of alloys

The above examination results revealed that the mid-ridges of Ge were made of pure copper, while the blades were made of Cu–Sn+Pb alloys. In contrast, most mid-ridges of the bi-metallic swords were made of Cu–Sn alloys with Sn contents in a range of 4~11 wt% (Lian and Tan 2002; He 1990; Chen 1981; Peng et al. 1994; Ding et al. 2012). For example, Chen et al. (Chen 1981) examined the chemical compositions of three bi-metallic bronze swords collected in Shanghai Museum. The data indicated that the mid-ridges were made of Cu–Sn alloy with average 10.48 wt% Sn, while the edges and blades were made of Cu–Sn alloy with higher Sn content 18.99 wt%. He et al.’s (He 1990) studies on the chemical compositions and microstructures of several bi-metallic bronze swords unearthed from Erzhou City, Hubei Province, China, revealed that the mid-ridges were made of Cu–Sn+Pb alloy with mean Sn content 10.247 wt% and Pb content 3.734 wt%, whereas the edges and blades were made of Cu–Sn alloys with higher Sn content 17.567 wt%. Compared to the bi-metallic bronze sword, the present Ges presented a special case by using pure Cu for the mid-ridges and Lans.

According to the phase diagram of Cu–Sn alloys, the melting point declines with the increase of Sn contents. Therefore, all mid-ridge, edge, and blade of the swords are made of Cu–Sn alloys with variant ratios; they would have a good metallurgical connection. That was to say, they had small differences in melting points, and were more likely to have mutual fusion and element diffusion at the joint of two-casting. Therefore, during the course of chop and collision, the bronze swords evaded the risks of cracking or separation, and their service life and quality would be greatly enhanced and prolonged.

In regard to why pure Cu was used for making the mid-ridges in these two bi-metallic bronze Ges, we proposed some possible reasons: (1) due to the higher melting point of pure Cu, the first-casting pure Cu mid-ridges provided a possibility to keep a rigid shape during the second-step casting, which was of advantage to have good structural integrity and the complete mortise and tenon joint; (2) as a precious metal in ancient time, to minimize use of Sn was reasonable for economic reason; (3) these two Ges might be made as a gift for a certain special person, rather than for equipping soldiers for battle, since they were produced carefully with hollowed-out decoration in Na; (4) they were probably burial objects for the dead of a special person, because lots of Pb was involved. Many researches indicated that the ancient bronze chopping and killing weapons were made of binary Cu–Sn alloys for high strength and hardness, while Pb generally reduced the strength (Chase and Thomas 1978; Liao et al. 2008).

Double-Ge halberd

According to archeological data, the halberds of the Eastern Zhou Dynasty (also called the Spring-autumn and Warring States Period) were assembled by separate parts. However, due to being poorly preserved in tombs, and the sticks that connected the Ges and spears became rotten, they were often treated as separate Ges and spears. In general, based on the way of combination, the halberds were divided into two categories: the first category was combined of a Ge and a spear; and the second type was made up of two or three Ges, in which the first one has Na, and the second and third ones have a shorter Na or no Na, and the length of Yuan decreased from top down (Jia et al. 2011).

When it comes to the two bi-metallic bronze Ges in this work, the no. 74 Ge had Na, while no. 76 Ge had no Na, and the Yuan of the no. 74 Ge was longer than that of the no. 76 Ge; they are 24.8 cm and 22.9 cm respectively. Besides, the contents of different corresponding parts of them were almost the same. These evidences showed these two Ges might be the two parts of a double-Ge halberd, as shown in Fig. 7. In addition, lots of multiple-Ge halberds were also unearthed from the nearby tombs, such as no. 5 tomb and no. 40 tomb of the Wangjiachong Cemetery, and no. 1 tomb of the Caojiagang Cemetery, which were the Chu tombs dated to the middle and late Warring States Period (Huang 2001). It disclosed that the multiple-Ge halberds were prevalent in Huangzhou District during the middle and late Warring States Period. Based on the above analysis, we confirmed that these two Ges belonged to the two portions of one halberd instead of as two separate Ges.

Fig. 7
figure 7

The double-Ge halberd picture of the Ges after restoration

Full size image

Conclusions

Two bi-metallic bronze Ges, as a kind of the Chu Ge dated to the Warring States Period (475 BC–221 BC), were firstly discovered and studied systematically by using optical microscope, 3-dimension microscope, SEM-EDS, pXRF, and X-ray radiography. It exhibits the same design philosophy with the well-recognized bi-metallic bronze sword, i.e., the mid-ridge is made of metal or alloy with good flexibility, and the blade is made of high-Sn bronze with high strength and hardness, which bring about a comprehensive performance in toughness and strength.

A possible manufacturing process is proposed as follows: (1) cast the mid-ridge and Lan with copper; (2) put the mid-ridge and Lan into the molds for making the cutting edge and blade, and cast the edge and blade with Cu–Sn+Pb alloys; (3) solider the body together with Na by using Cu–Sn+Pb alloy.

Referring to the Ges from the same unearthed site, similar alloys and similar typology, it is supposed that these two Ges were the two parts of a double-Ge halberd. And its special appearance and complex shape indicate that this halberd perhaps is made for a person of high position or power. This work provides a new evidence for the metallurgical level during the Warring States Period in China.