The WXP (Wangxiangdun) site is situated 50 meters southeast of Xiashoupu Bay in Dazhuang Village, Liurenba Town, Daye County, Huangshi City, Hubei Province. According to the 3rd national official archaeological survey, the settlement site is thought to belong to the Western Zhou period. The WXP site is located 300 meters southeast of Dazhuangwu in Dongshan Village, Liurenba Town, and spans an area of approximately 12,000 square meters with an archaeological stratigraphy thickness of 1.5-2 meters. The pottery found at the site is predominantly yellow-brown temperless pottery, accompanied by red sand tempered pottery. Decorative elements on the pottery include rope patterns and string patterns featured on li-tripods, yan-steamers, dou-stemmed bowls, and guan-jars[12].
The Straitification Structure and Radiocarbon Dating Result
During this survey, a new profile was discovered at the mountain top terraces. The profile faces west and spans a width of 240cm. It can be divided into three layers:
-
The first layer ①, found below the surface soil, is the disturbance layer. It is 20-30cm thick and consists of gray-brown soil with fine sand particles. The soil in this layer is relatively loose.
-
The second layer ② is the cultural layer, which is 70-110cm thick. It contains fine sand particles within blue-gray clay. The color is lighter than that of the first layer, and it includes furnace slag, charcoal shavings, pottery fragments, and other cultural artifacts.
-
The third layer ③ is another cultural layer, mainly composed of yellow clay with a thickness of 15-30cm. This layer contains sandstone red pottery, sandstone gray pottery, furnace slag, charcoal shavings, and numerous pottery fragments. Charcoal samples were extracted from this layer for 14C dating.
To collect charcoal shavings near the pottery fragments at the bottom of the WXP site in Daye, adhere to the following standardized collection method to prevent contamination by foreign carbonaceous substances[13]:
-
Use a metal shovel to carefully extract the surrounding soil along with the charcoal shavings.
-
Wrap the collected samples in aluminum foil to avoid exposure to contaminants.
-
Ensure that the collection and extraction process strictly adheres to relevant national standards.
Once the samples have been collected, they were sent to the scientific archaeology laboratory at Peking University's School of Archaeology and Museology. The age of the charcoal shavings sample was determined to be from the late Shang Dynasty, as shown in Table 1.
The Sampling Strategy and Investment Methods
In smelting sites, various relics and artifacts can be used to study smelting technology. One of the easiest to obtain is slag, the waste product created during smelting. In copper smelting activities, slag production is a core process. Slag is discharged from the furnace in a molten state at smelting temperatures and solidifies into a dense solid, permanently sealing the metallurgical reaction information within. The properties of slag directly affect the success of the smelting process. As a result, detecting and analyzing slag can help reveal ancient copper smelting technology. At the WXP site, the metallurgical content was investigated through surface and section surveys. The smelting slag found neither accumulated nor distributed sparsely, and the quantity was small. The slag blocks came in various shapes and had mostly dark gray surfaces. The size of the slag bodies ranged from 2–5 cm in length and width, and some larger pieces were approximately 10 cm in length and width, with a thickness of 1-2.5 cm. The slag was relatively small and did not accumulate. To better understand the metallurgical content of the WXP site, 10 samples were selected for testing.
As shown in Fig. 4, there is no significant difference in color between the slag samples from the WXP site. The lighter portions of the slag may be due to the burial environment and difficulties in external cleaning. Most of the slag is small in size, with no large blocks of slag residue present at the site. The length of most slag pieces is about 10cm. The fluidity of most slag is poor, and some slag samples have many pores, The sample WXP③-02 has more pores and is lighter. Overall, the slag collected from the surface has lower viscosity than the slag unearthed from the cultural layer, and its density is also higher. The stratum slag is more irregular due to its poor fluidity, its surface is more uneven, and it adheres to more heat-treated sandy soil and the like. There are a small number of slag pieces with traces of having flowed out, such as smelting slag collected from the surface. It is not difficult to find this kind of slag in the field, as it varies in characteristics like color, appearance, texture, and specific gravity. Fortunately, almost all ancient metallurgical slags are ferrous silicates, whether they are by-products of iron or non-ferrous metal production. This fact helps in identifying and distinguishing them from non-metallurgical slags found on site.
The analysis parts of the slag and furnace wall were selected and cut using a diamond band saw. The cut samples were fixed and embedded using TROJAN epoxy resin. The porous samples were vacuum-embedded to ensure structural stability and reduce secondary contamination of the samples during processing. The samples were then polished with waterproof sandpaper and polishing fabrics. The surface of the samples was coated via vacuum carbon sprayed before the scanning electron microscope energy spectrum analyzer (SEM-EDS) was performed for analysis. The scanning electron microscope model TESCAN VEGA3 XMU was equipped with a Bruker Nano GmbH 610M energy spectrometer to measure the composition of the matrix and metal particles, etc. The instrument working conditions were an excitation voltage of 20kV, a working distance of 15 ± 0.1mm, a collection number of > 2kcps, and a valid scanning time of 60s were set. When scanning the slag matrix, large metal particles and holes were avoided. The matrix composition was normalized by the context of the structure of the SEM observations and the results is shown in the table below:
Table 2
Matrix chemical composition of the slag from WXP(reported in wt%)
Lab No | MgO | Al2O3 | SiO2 | K2O | CaO | FeO | Cu | Sn |
wxp②01 | — | 7.6 | 34.9 | 0.9 | 4.2 | 52.4 | 1.8 | — |
wxp③01 | 0.7 | 6.9 | 30.4 | 1.0 | 5.0 | 56.0 | 0.6 | — |
wxp③02 | 1.8 | 4.7 | 27.2 | 1.3 | 12.7 | 51.8 | 1.0 | 0.3 |
wxp③03 | — | 5.8 | 26.4 | 0.8 | 3.5 | 63.5 | 1.4 | — |
wxp③04 | 0.6 | 6.6 | 19.0 | — | 1.3 | 72.5 | 0.2 | — |
wxpc01 | 1.0 | 5.8 | 38.8 | 2.0 | 7.4 | 45.1 | 0.9 | — |
wxpc02 | 0.9 | 3.9 | 24.9 | 0.3 | 1.8 | 68.3 | 1.1 | — |
wxpc03 | 2.0 | 3.5 | 34.6 | 1.1 | 7.4 | 51.4 | 0.3 | — |
wxpc04 | 2.0 | 3.5 | 34.6 | 1.1 | 7.4 | 51.4 | 0.3 | — |
wxpc05 | 0.5 | 4.8 | 29.3 | 1.3 | 5.5 | 58.6 | 1.4 | — |
The Constitution and Content Data of the Relics
In addition to iron and silicon as the main components, the slag composition contains relatively high levels of calcium and aluminum. To better determine its thermodynamic parameters, the slag matrix composition is marked on two ternary slag phase diagrams. As illustrated in these Fig. 5–6, most of the slag falls within the ideal melting area of the ternary phase diagram[15]. This indicates a high level of technical proficiency, such as ore selection processes and smelting technology at that time, ensuring stable operation of smelting activities in the furnace. Good atmosphere control during smelting is also achieved through the matrix phase of the slag. The slag matrix is primarily composed of dark gray dendritic or chain-like iron olivine and black-brown glass phase. Additionally, there are obvious capillary iron olivine crystals precipitated. This is consistent with the phases precipitated in the melting area where each component is located in the diagrams mentioned above.
While it is generally believed that the metallurgical process represented by slag composition cannot be directly determined, it is possible to infer the probable products of smelting production based on the metal particles encapsulated in the slag. As a result, the copper smelting slag collected and unearthed at the WXP site can be roughly divided into two categories. Class A is red copper smelting slag, and Class B is tin bronze alloy slag. This classification gives a basic understanding of the types of smelting processes and products that may have taken place at the site.
Five of the samples (WXP②-01, WXP③-01, WXP③-04, WXP C-01, WXP C-02, WXP C-05) were identified as Type A and represent the red copper smelting activity. Red copper metallurgical technology analysis Class A involves red copper smelting, and the resulting slag is red copper smelting slag. This slag is primarily irregular in shape, gray-black, and abundant at the site. The matrix composition of the slag reveals a high iron content, ranging from 50–70%, and a large number of metal particles. No other metal particles are highly enriched in the slag. The slag is generally dense, with some fluidity (although not ideal). The internal phase mainly consists of iron olivine and fayalite, indicating that this type of slag is copper smelting slag. The analysis area of metal particles is shown in Fig. 6. Aa1 represents pure copper particles, and Aa2, wrapped around Aa1, signifies white bronze particles. In Fig. 7, A2, B2, C1, and D1 are all red copper particles. A1 and B1 in Fig. 7 illustrate pure copper particles wrapped around white bronze particles, while D2 and D3 in Fig. 7 denote oxidized red copper particles.
Table 3
Composition of metal prill in WXP slag(reported in wt%)
Sample code | Testing area | Composition (wt%) |
Cu | Fe | S | O |
WXP②-01 | Figure 6.Aa-1 | 100.0 | — | — | — |
Figure 6.Aa-2 | 77.3 | — | 22.0 | 0.7 |
Figure 6.Aa-3 | 79.2 | — | 20.8 | — |
WXP③-01 | Figure 6.Bb-1 | 97.9 | — | — | 2.1 |
Figure 6.Bb-2 | 75.8 | 0.3 | 21.7 | — |
Figure 6.Bb-3 | 71.8 | 3.2 | 23.4 | 1.6 |
Figure 6.Bb-4 | 75.8 | — | 23.3 | 0.8 |
WXP③-04 | Figure 7.A1 | 75.4 | 3.0 | 21.4 | 0.3 |
Figure 7.A2 | 98.9 | — | — | 1.1 |
WXPC-01 | Figure 7.B1 | 76.5 | 2.2 | 21.3 | — |
Figure 7.B2 | 99.4 | — | — | 0.6 |
WXPC-02 | Figure 7.C1 | 100.0 | — | — | — |
WXPC-05 | Figure 7.D1 | 100.0 | — | — | — |
Figure 7.D2 | 81.8 | — | 0.6 | 17.6 |
Figure 7.D3 | 86.3 | — | — | 13.7 |
Class B represents tin bronze alloy smelting slag, which is relatively limited in quantity at the site and is closely related to the early supply of tin and the amount of slag produced during alloying. The smelting technology can be deduced based on the typical phase information of the tin bronze smelting slag.
Only two furnace slags in the stratum exhibit relevant phenomena associated with tin bronze smelting activities. The tin bronze slag discovered in the late Shang Dynasty stratum primarily features a slag matrix structure composed of iron olivine, with large particles of iron ore residue remaining. The largest iron ore particle exceeds 2mm. The light gray strip area in the slag matrix represents this situation. In Fig. 8, analysis area Aa1 consists of copper oxide particles, while the surrounding Aa2 and Aa3 analysis areas show different brightness levels depending on the Sn content. However, the basic composition remains primarily Sn-Fe (Cu), and there are tiny Cu pure copper particles inside (Fig. 8.Aa4 position). This indicates that the material is at the reaction stage after adding ore, and the reaction has not yet been fully carried out.
Type B (including WXP③-02 and WXP③-03) is tin bronze alloy smelting slag, which are rare at the site, related to the early supply of tin and the amount of slag produced by alloying. Only 2 pieces of slag in the stratification were excavated from Layer ③ with plenty of pottery shards of the late Shang Dynasty. Most of the slag matrix is fayalite structure, and there are still large particles of ore remaining. The largest ore remainder over 2mm, which can be observed as light gray elongated area in the slag matrix.
The circular particles in Fig. 8.Ab are high-tin particles, and the overall tin content of the particles is above 53%. According to the different tin contents, they are divided into two phases, but they are basically metal particles formed by heating after tin ore is added to the copper smelting furnace. Ac1 in Fig. 8.Ac is iron-tin ore residue with small particle size. Ac2 and Ac3 are white ice copper particles wrapped on the outside. Ac4 has similar properties to Ac1, but has a higher copper content. They are all wrapped in high-ice copper particles. Ac5 and Ac7 are close to the slag matrix composition, mainly Fe-Si. Ac6 is high-ice copper particles after successful smelting. In Fig. 8.Ad, Fig. 8.Ad1, Ad2, and Ad3 are similar to the particles in Ac1 and Ac4. They are all wrapped in ice copper particles (Ad5) mainly wrapped in iron-tin ore residue by high-ice copper particles. The outer gray phase (Ad4, Ad6) is a floating phase.
Ba1 in Fig. 9.Ba1 is ice copper particles, Ba3 on the inside is pure copper particles wrapped up, and there are also pure copper particles Ba2 growing. Ba4 contains about 6% Sn and is a tin bronze alloy. Bb1, Bb2, and Bb3 are high-tin unmelted mineral residues. Due to different tin contents, they have different brightness. The overall tin content is about 40%, with some characteristics of yellow tin ore (Cu2FeSnS4). The analysis area Bc is unmelted tin mineral residues (Bc2) wrapped in ice copper particles (Bc1).
Based on the phase composition of this batch of slag which is similar to other areas in the Yangtze River Basin, it is speculated that the bronze metallurgical process is similar to that of Xianglushan in Daye. It should be a two-step method: first produce red copper and discharge red copper slag; then add tin ore to the furnace containing Cu liquid to smelt bronze and discharge bronze slag. Considering that many sites have discovered copper minerals containing Sn-Fe, it is speculated that tin materials are added in the form of associated Cu stannite (SnO2, Sn-Fe minerals) rather than metallic tin.
Table 4
Composition of metal prill in WXP slag(reported in wt%)
Sample code | Testing area | Composition (wt%) | | |
O | Si | S | Fe | Cu | Sn |
WXP③-02 | Figure 8.Aa-1 | 15.4 | — | — | — | 84.6 | — |
Figure 8.Aa-2 | 40.4 | 5.4 | — | 26.0 | 2.5 | 25.6 |
Figure 8.Aa-3 | 41..2 | 5.1 | 0.5 | 38.0 | 4.8 | 10.4 |
Figure 8.Aa-4 | — | — | 19.3 | — | 80.7 | — |
Figure 8.Aa-5 | 30.4 | 4.2 | — | 8.9 | 40.0 | 16.5 |
Figure 8.Ab-1 | 28.7 | 2.2 | — | — | 15.9 | 53.2 |
Figure 8.Ab-2 | 26.4 | 2.7 | — | 1.5 | 13.4 | 55.9 |
Figure 8.Ab-3 | 28.8 | 2.9 | — | 1.1 | 11.0 | 56.3 |
Figure 8.Ac-1 | 32.7 | 4.6 | — | 20.0 | 7.7 | 35.0 |
Figure 8.Ac-2 | — | — | 22.7 | — | 77.3 | — |
Figure 8.Ac-3 | — | — | 22.0 | — | 78.0 | — |
Figure 8.Ac-4 | 27.6 | 2.4 | — | 7.0 | 42.6 | 20.4 |
Figure 8.Ac-5 | 39.8 | 8.6 | — | 43.9 | — | — |
Figure 8.Ac-6 | — | — | 20.5 | — | 79.5 | — |
Figure 8.Ac-7 | 36.7 | 16.0 | — | 26.8 | 9.0 | — |
Figure 8.Ad-1 | 27.1 | 2.1 | 0.5 | 3.9 | 40.3 | 26.1 |
Figure 8.Ad-2 | 26.5 | 1.2 | 0.5 | 10.0 | 41.8 | 20 |
Figure 8.Ad-3 | 2.6 | — | 3.9 | 4.1 | 76.7 | 12.8 |
Figure 8.Ad-4 | 27.4 | — | — | 72.1 | — | — |
Figure 8.Ad-5 | 20.7 | 1.0 | 9.7 | 32.2 | 35.1 | — |
Figure 8.Ad-6 | 26.4 | — | — | 72.6 | — | — |
WXP③-02 | Figure 9.Ba-1 | — | — | 22.7 | — | 77.3 | — |
Figure 9.Ba-2 | 16.5 | — | — | — | 83.5 | — |
Figure 9.Ba-3 | 18.5 | — | — | — | 81.5 | — |
Figure 9.Ba-4 | 30.8 | 0.8 | 0.7 | 28.7 | 32.6 | 5.9 |
Figure 9.Bb-1 | 35.5 | 2.9 | — | 15.3 | 2.1 | 40.7 |
Figure 9.Bb-2 | 33.5 | 3.3 | — | 17.0 | 0.6 | 41.9 |
Figure 9.Bb-3 | 36.9 | 2.5 | — | 27.7 | 3.1 | 25.9 |
Figure 9.Bc-1 | 1.1 | — | 20.8 | 3.5 | 74.6 | — |
Figure 9.Bc-2 | 33.0 | 2.1 | 1.8 | 14.5 | 26.1 | 16.1 |