Characterization of corrosion products formed on bronze ornaments excavated from the Chu tombs of the Spring and Autumn Periods in Dangyang, China

DOI: https://doi.org/10.21203/rs.3.rs-2106574/v1

Abstract

130 corrosion products on bronze ornaments excavated from the Chu tombs of the Spring and Autumn periods in Dangyang, Hubei Province, China, were examined by X-ray diffraction, and statistically analyzed. The results showed that the main corrosion products were cassiterite (SnO2), copper sulfates, and sulfides, and the rare tenorite (CuO). And corrosion products of different colors such as blue-green and black-gray were specifically detected by Raman spectroscopy. The Raman spectra of blue-green corrosion products were identified as Cu hydroxide sulfates and Cu hydroxide carbonate such as brochantite, posnjakite, antlerite and malachite, the black-gray corrosions were confirmed as copper sulfides, only the good quality Raman spectrum of covellite was obtained. The shifts of certain band positions in different Raman spectra of the black-gray corrosion products, as well as the causes of D and G peaks in the spectra were discussed. And the advantages and disadvantages of using Raman spectroscopy for the identification of corrosion products were presented and discussed. Finally, with reference to previous literature, it was proposed that the SO 42 - ions and sulfate-reducing bacteria in the soil, along with the closed and anoxic burial environment, may explain the sulfate and sulfide corrosion products on the bronze ornaments. The presence of tenorite and charcoal indicates that the tombs might have undergone fire.

Introduction

From March 2017 to December 2019, The National Museum of China carried out the conservation and restoration of 655 pieces of Chu metal ornaments in the Yichang Museum dated to the late Spring and Autumn period, which were excavated from the Caojiagang #5 and Zhaoxiang #12 Chu tombs in Danyang, Hubei Province, China. The Caojiagang #5 tomb was excavated in October 1984. The tomb had been robbed, and only nearly 1,000 pieces of small bronze artifacts, metal armor ornaments, leather armor, musical instruments, bones and shells, and bamboo and wood artifacts were found, including 196 pieces of metal ornaments (Yichang Prefectural Museum 1988). The Zhaoxiang #12 tomb was excavated in December 1996. The tomb had also been robbed, but the inner and outer coffins were well kept, with a total of 459 pieces of identifiable metal ornaments as well as more than a hundred scattered gold foil and tin fragments unearthed. The two tombs are only three kilometers apart.

The material types and forms of the metal ornaments excavated from the two tombs are largely the same. Materials include tin–lead alloy and bronze, mainly in the swallow-tailed, three-legged, tiger, crab, and bi forms (a flat jade disc with a circular hole in the center). Most of the tin–lead ornaments are covered with decorated tin foil (with a few covered with gold foil), whereas the bronze ornaments are covered with decorated gold foil. The metal ornaments excavated from the two tombs are large in number, rich in type, diverse in material, and well made. They are of a relatively high rank among similar artifacts from Chu tombs and have high historical value. Scholars have published the scientific analysis of the production process of the excavated metal ornaments (Cheng et al. 2019). The present work focuses on the analysis of the corrosion products of the ornaments, and further explores the causes of their formation.

The corrosion situation of the bronze ornaments from the two tombs is similar, characterized mainly by corrosion products in blue-green and black-gray colors. The blue-green corrosion products are common on bronzes excavated from tombs, but the main characteristic of corrosion of these bronze ornaments is the black-gray tint of the surface of most artifacts. Even with gold foil covering them, the black-gray corrosion products are still closely bound to the gold foil, some black as carbon.

Materials And Methods

In this study, Raman spectroscopy combined with microscope were used to investigate the chemical and structural composition of the corrosion products with different colors. And X-ray diffraction (XRD) was also performed to conduct macroscopic examination of 130 types of corrosion products of different colors and characteristics. The results were statistically analyzed to determine the pattern of distribution and characteristics of the corrosion products. In addition, we used ion chromatography (IC) to analyze the soluble salts in the soil adhering to the artifacts to provide a reference for exploring the causes of corrosion.

Microscope: The surface profiles of samples were obtained by SmartZoom5 digital microscope produced by Zeiss.

X-ray diffractometer: A D/max 12kW X-ray powder diffractometer was performed at room temperature and copper X-ray tube ( Kα = 0.15418 nm) with bent crystal graphite monochromator; The tube voltage was 40 kV and the tube current was 100 mA.

Raman spectroscopy: The spectra were recorded by a Renishaw inVia Raman microscope system under a 50× objective lens with a 532 nm laser. The laser power on the sample surface was 2.5–5 mW. The acquisition time was 10 seconds, and acquisition was performed one to two times. The spectral resolution of the instrument was 1–2 cm− 1, and the spatial resolution was 0.5 cm− 1.

Ion chromatography: A DX-600 ion chromatograph (DIONEX Corporation, USA) was applied with AS14 separation column, 3.5–1.0 mmol/L NaCO3 isocratic elution and ECD ASRS-ULTRA automatic electrochemical suppression cycling mode, flow rate was 1.2 mL/min and suppression current was 40 mA.

Results

Raman spectral analysis

Raman spectroscopy was used to analyze the corrosion products of 17 pieces of bronze ornaments numbered Y6109 (Fig. 1), Y6023(Fig. 2), Y6031, Y6640, Y6625, Y6110, Y6001, Y6011, Y6032, Y6037, Y6042, Y6112 (Fig. 4), Y6083 (Fig. 5), Y6105, Y6123, Y6445 and Y6662, focusing mainly on the corrosion products of blue-green and black-gray colors, and 36 Raman spectra were obtained. The results are shown in Table 1. The Raman spectra are illustrated in Fig. 3, Fig. 6 & Fig. 7. Table 2 lists the phases present in a patina and their main Raman fingerprint, together with references found in literature for the same compounds. The relative intensities of the bands are characterized as being very strong (vs), strong(s), medium(m), weak(w), very weak(vw), broad (br) and shoulder(sh).

Blue-green corrosion products

The corrosion products of bronze ornament Y6109 are rich in color and morphology, needle-like, blue-green (Fig. 1a) & round granular, green (Fig. 1b) corrosion products were observed under microscope. The needle-like, blue-green corrosion products contain two phases. The Raman spectrum of phase 1 is shown in Fig. 3a with characteristic peaks at 122(vw), 145(w), 176(m), 230(vw), 336(w), 415(sh), 442(s), 504(m), 619(w), 972(vs), 1109(vw), 1581(vw), 3251(w), 3398(w), and 3556(w) cm − 1. The Raman spectrum of phase 2 is shown in Fig. 3b with characteristic peaks at 118(w), 138(m), 156(m), 196(s), 239(w), 317(w), 389(m), 423(m), 448(w), 482(m), 506(w), 595(m), 607(w), 620(w), 907(w), 972(vs), 1075(w), 1097(w), 1122(w), 3259(w), 3370(w), 3402(w), 3566(s) and 3590(w) cm− 1. The Raman spectrum of the round granular, green corrosion product is shown in Fig. 3c, with characteristic peaks at 154(w), 181(m), 220(m), 270(m), 359(w), 432(s), 534(m), 720(vw), 1064(m), 1374(w), 1490(s) and 3380(m)cm− 1. Compared the spectra with the RRUFF database (http://rruff.info/about/about general.php) and references (Hayez et al. 2004; Gilbert et al. 2003; McCann et al. 1999; Bouchard et al. 2003), Phase 1 and phase 2 are respectively attributed to the posnjakite [Cu4(SO4)(OH)6(H2O)] and the brochantite [Cu4(SO4)(OH)6], and the round granular, green corrosion product are identified as malachite [CuCO3·Cu(OH)2].

Bronze ornament Y6023 (Fig. 2) has an overall dark gray color with spotted, green corrosion products in the lower right corner of the artifact, which consist of antlerite [Cu3(SO4)(OH)4] and brochantite according to Raman analysis. The Raman spectrum of antlerite is shown in Fig. 3d with characteristic peaks at 122(m), 145(w), 170(vw), 216(vw), 247(w), 266(m), 343(w), 415(s), 441(vw), 484(m), 599(m), 629(w), 649(vw), 986(vs), 1075(m), 1136(w), 1172(m), 3487(s), and 3579(s) (Hayez et al. 2004; Gilbert et al. 2003). 

 
Table 1

Raman results of blue-green and black-gray corrosion products on the bronze ornaments

Number

Colors

Tenorite1

Cuprite

Posnjakite

Brochantite

Antlerite

Malachite

Atacamite

Cove*lite

Chalcocite1

Cerussite

Y6109

Blue-green,

black-gray

    

√

√

  

√

  

√

  

√

Y6023

Green

      

√

√

          

Y6031

green

    

√

    

√

        

Y6640

green

            

√

      

Y6025

green

      

√

√

    

√

    

Y6110

Green

      

√

            

Y6001

Green, black-gray

      

√

√

    

√

√

  

Y6011

green

      

√

√

    

√

    

Y6032

Green, black-gray

          

√

  

√

    

Y6037

Green, black-gray

      

√

      

√

    

Y6042

black-gray

                

√

√

Y6112

Red

Black-gray

√

√

          

√

    

Y6083

black-gray

              

√

√

  

Y6105

gray

              

√

√

  

Y6123

gray

        

√

    

√

    

Y6445

black

        

√

    

√

    

Y6662

black

                

√

  
1 means the corrosion products are possibly existed.

 
Table 2

Raman signatures of the phases (stronger peaks in bold)

Phases

Color

Formula

Main Raman peaks

cm− 1

References

Brochantite

Blue-green

Cu4(SO4)(OH)6

118, 138, 156, 196, 239, 317, 389, 423, 448, 482, 506, 595, 607, 620, 907, 972, 1075, 1097, 1122, 3259, 3370, 3402, 3566, 3590

RRUFF database

Hayez et al. 2004

Gilbert et al. 2003

Posnjakite

Blue-green

Cu4(SO4)(OH)6(H2O)

122, 145, 176, 230, 336, 415, 442, 504, 619, 972, 1109, 1581, 3251, 3398, 3556

RRUFF database

Hayez et al. 2004

Gilbert et al. 2003

Antlerite

Blue-green

Cu3(SO4)(OH)4

122, 145, 170, 216, 247, 266, 343, 415, 441, 484, 599, 629, 649, 986, 1075, 1136, 1172, 3487, 3579

RRUFF database

Hayez et al. 2004 Gilbert et al. 2003

Malachite

Green

CuCO3·Cu(OH)2

154, 181, 220, 270, 359, 432, 534, 720, 1064, 1374, 1490, 3380

RRUFF database

McCann et al. 1999 Bouchard et al. 2003

cuprite

Red

Cu2O

143, 202, 218, 408, 514, 628

RRUFF database

McCann et al. 1999

Colomban et al. 2012

Covellite

Gray

CuS

137, 263, 469, 914

RRUFF database

Bouchard et al. 2003

Smith et al. 2002

White 2009

Chalcocite

Gray

Cu2S

284, 322, 611

RRUFF database

McCann et al. 1999

Tenorite

Black

CuO

298, 345

RRUFF database

McCann et al. 1999

Black-gray corrosion products

As noted above, the most striking feature of those bronze ornaments is that the surface of most artifacts is covered with a layer of corrosion products in black-gray, regardless of whether the artifact is the metal base or covered with gold foil. For example, on the surface of the gold foil covering Y6112 (Fig. 4), Three types of corrosion products in red, black, and gray could be directly observed, and the corrosion products are relatively dense. The Raman spectrum of the corrosion products in red is shown in Fig. 6, with the characteristic peaks at 143(m), 202(sh), 218(s), 408(w), 514(br), and 628(w) cm− 1. The Raman spectrum of the corrosion products in gray is shown in Fig. 7a with the characteristic peaks at 137(w), 263(w), 470(vs), and 914(w) cm− 1. The Raman spectrum of the corrosion products in black is shown in Fig. 7b with the characteristic peaks at 137(s), 263(m), 470(s), and 607(w)cm− 1. The spectral peaks are similar for the gray and black corrosion products, albeit with some differences. In comparison to the gray corrosion products, the two peaks at 137 and 263 cm− 1 are stronger in the black corrosion products, and the new weak peak at 607 cm− 1 appeared, whereas the peak at 914 cm− 1 disappeared. In addition, the two Raman spectra share two similar peaks (around 1360 and 1580cm− 1 ) between 1300 and 1600cm− 1. Compared the spectra with the online RRUFF database and references ( McCann et al. 1999; Bouchard et al. 2003; Colomban et al. 2012; Smith et al. 2002; White et al. 2009), the red corrosion products are identified as the cuprite (Cu2O) and the gray corrosion products are consistent with the Raman spectrum of covellite (CuS) numbered R060306 in the RRUFF database, where the spectral peaks 263 and 470 cm− 1 conform with the stretching vibration of Cu-S (White et al. 2009; Xi et al. 2019; Hurma et al. 2016; Anastasia et al. 2019). The black corrosion products are mainly covellite, The weaker spectral peak at 607 cm− 1 indicates the possible existence of chalcocite (Cu2S) ( McCann et al. 1999), whereas the two weak, broad and diffuse peaks at 1360 and 1580 cm− 1 indicate the presence of humic substances (HS) or charcoal (Alon et al. 2002). This will be discussed in more detail in the next section. The XRD results (Fig. 8) indicate that the black-gray corrosion products contain a variety of corrosion products, including cassiterite (SnO2), posnjakite, cuprite, tenorite, and brochantite, with cassiterite being the dominant phase.

The gold foil on the bronze ornament Y6083 (Fig. 5) is partly covered by black-gray corrosion products. The gray corrosion products are identified to the covellite. The Raman spectrum of the black corrosion products is shown in Fig. 7c with the characteristic peaks at 284(m), 322(w), 469(m), 611(w), 1364(br), and 1586(br) cm − 1, which is similar to the spectrum of the black corrosion products on Y6112, but with some differences. The peak at 263 cm− 1 disappeared, the new weak peaks at 284, 322, 401 and 611 cm− 1 appeared, and 1360 and 1580cm− 1 peaks are sharper and more obvious compared with the Raman spectrum of gray corrosion products on bronze ornament Y6112 (Fig. 7a). The new weak peaks at 284, 322 and 611 cm− 1 indicate the possible existence of chalcocite (Cu2S) ( McCann et al. 1999; Young et al. 2010), another new weak peak at 401cm− 1 indicate the possible existence of tenorite (McCann et al. 1999).

XRD analysis

A total of 130 corrosion products on pieces of bronze ornaments were analyzed by X-ray diffraction, the results are shown in Appendix A, B and C, and those of further statistical analysis of the corrosion products detected and displayed on Fig. 9, which shows that among the 130 corrosion products analyzed, cassiterite and brochantite are the most common, which are present in 96 samples, accounting for 74% of the total. Ranking next are cuprite, djurleite (Cu1.96S), posnjakite, and antlerite. Cuprite is present in 94 samples, accounting for 72% of the total. Djurleite is present in 70 samples (54%). In addition, we observed digenite (Cu9S5), roxbyite(Cu7S4), tenorite, anilite (Cu7S4), dolerophanite ( Cu2(SO4)O), and chalcocite (Cu2S), as well as lead-bearing corrosion products, such as anglesite (PbSO4), cerussite (PbCO3 ·Pb(OH)2), and galena (PbS).

Among the bronze ornaments excavated from the Zhaoxiang and Caojiagang tombs, the corrosion products are mostly sulphates and sulfides. Malachite, a corrosion product common in soil environments, is observed in only 6 of the 130 samples. A relatively rare corrosion product, tenorite, is found in 9 out of 130 samples.

Ion chromatography analysis of soluble salts in soil adhering to artifacts

The two tombs had been backfilled long ago, so the original soil could not be collected. Therefore, we only collected residual soil adhering to the ornaments, soaking it in deionized water and then retrieving the clear liquid on the upper layer to analyze the soluble anion content in the soil by IC. The results show that the SO42− ion content of the two soil samples is as high as 320.60–339.06 mg/L, much higher than the Cl − and NO32− ion content, which is consistent with the analysis of the corrosion products where sulfates dominate (Table 3) .

 
Table 3

Results of the ion chromatography analysis

Tomb

Number

Cl− mg/L

NO32− mg/L

SO42− mg/L

Caojiagang 5# tomb

Y6646

0.31

1.59

320.60

Y6647

5.82

4.77

339.06

Discussion

Causes of corrosion

Normally, the main corrosion products formed on copper alloy artifacts in the environment of outdoor pollution are copper sulfate or alkaline copper sulfate (Martens et al. 2003; Scott 2002). However, the corrosion products of these bronze ornaments are mainly sulfates and sulfides, the most typical and unique corrosive products are the dark-colored copper sulfides. IC results of the soil adhering to the artifacts indicate that the original burial environment contained high levels of soluble sulfates. Therefore, the SO42− ions in the soil are the main cause of the sulfate corrosion products on the bronze ornaments.

Studies have shown that copper sulfides are readily produced in moist, anoxic soil environments or deoxygenated seawater (Young et al. 2010; Smith et al. 2007; Trentelman et al. 1999). Michelle Chan et al studied the characterization of the patina formed on a low tin bronze exposed to aqueous hydrogen sulfide, which represents a model for bronze corrosion in reducing conditions where sulfate-reducing bacteria in soils or deoxygenated seawater may generate H2S during respiration. The results showed that the corrosion layer formed by H2S(aq) exposure was dominated by polycrystalline Cu2S (low and high chalcocite) and smaller concentrations of CuSO4 · nH2O (Michelle et al. 2014). So we can infer that in the burial environment of the two tombs, namely, the anoxic soils rich in SO42−ions, organic matters, and sulfate-reducing bacteria, the sulfate-reducing bacteria may have reduced the sulfates according to the following reactions:

SO42− → S0 + S2− + O2

2H+ + S2− → H2S

Cu2+ + H2S → CuSx + 2xH+

This burial environment generated a variety of copper sulfides with different stoichiometric ratios, which makes the identification of these corrosion products more difficult owing to incomplete crystallization. The main copper sulfide corrosion products found in the literature include covellite, chalcocite, djurleite, digenite, anilite, bornite (Cu5FeS4), chalcopyrite (CuFeS2) and, rarely, geerite (Cu1.6S). The XRD results show that the predominant copper sulfide in 130 pieces of bronze ornaments is djurleite, which is found in more than half of the artifacts, digenite, anilite, chalcocite, and covellite are also found.

Djurleite and anilite are in the orthorhombic crystal system with stable temperatures of 72°C and 75°C, respectively. Covellite is in the hexagonal crystal system that can reach a stable temperature of 507°C. Some phases of copper sulfides are prone to thermodynamic changes and interconversion (Scott 2002). J.G. Dunn and C. Muzenda studied the thermal transformations of covellite in the temperature range 330–820°C, and the results showed that covellite underwent the following phase change process, CuS→Cu1.8S→Cu2S→CuSO4→CuO·CuSO4→CuO ( Dunn et al. 2001).

In fact, there are many reports about copper sulfides formed on buried copper alloys. For example, Duncan and Ganiaris (1987) found that most corrosion products developed on copper alloys excavated from the River Thames in London where rich in sulfate-reducing bacteria, were composed of various sulfides with stoichiometry ranging from CuS (covelite) to Cu2S (chalcocite). The corrosion products on 37 pieces of coins from the 12th to 19th centuries excavated from two archaeological sites, Largodas Olarias and Travessa do Jordão in Lisbon, Portugal, found by various means, including XRD, scanning electron microscopy with energy dispersive spectrometry (SEM-EDS), and X-ray fluorescence spectrometer, are dominated by cuprite and atacamite with a minor amount of copper sulfides, including chalcocite, covellite, digenite, and tenorite (Güncem Diktaş 2017). In studies of bronze artifacts uncovered from lake sites, Schweizer postulated that copper sulfides formed under anoxic (oxygen depleted) conditions in organic-rich soils containing SRB (Schweizer 1991).Fan et al analyzed Corrosion products on 23 bronze dagger-axes unearthed from different tombs dating to the Warring States period at Yujiaba site, Kai County, Chongqing, China by Raman spectroscopy, The results showed that covellite was present in 9 samples, and other corrosion products such as Cu hydroxide sulfate, Cu hydroxide carbonate, Cu hydroxide phosphate, Cu chloride hydroxide, Cu oxide, Pb corrosion products were also detected (Fan et al. 2020).

The Raman spectra of many black-gray corrosion products contain 1360 (D peak) and 1580 (G peak) cm− 1. The D and G peaks in some gray corrosion products on bronze ornaments such as Y6001, and Y6105 are weak and not particularly obvious, in similar pattern as Fig. 7a, however, the D and G peaks in some black corrosion products on bronze ornaments such as Y6025, Y6001, Y6042, and Y6123 are obvious and sharp obvious, in similar pattern as Fig. 7c. Studies have shown (Alon et al. 2002) that the Raman spectra of synthetic graphite, modern charcoal and an extract of HS share two similar peaks (D and G peaks), where the G peak is characteristic of graphite and appears in all graphite materials (Yang et al.1997). The D peak is attributed to the effect of particle size, and the intensity ratio I(D)/I(G) reflects the structural properties of the material (Tuinstra et al.1970; Mennella et al.1995). It is relatively difficult to distinguish between HS and charcoal, a generally accepted fact is that the Raman spectra of HS have a fluorescence effect and a greater intensity than the Raman spectra of charcoal or synthetic graphite. Therefore, it is clear that in Fig. 7c, corrosion products contain charcoal, indicating the tombs might have undergone fire and some of the organic matters within the tombs were charred, while in Fig. 7a and 7b, the corrosion products may contain HS from the soil.

Another typical feature of the corrosion products of those bronze ornaments is the presence of a small amount of rare tenorite. The oxidation of most copper compounds in air will eventually produce tenorite on heating; the compounds will decompose to cupric oxide between 400℃and 600℃.Tenorite is a rare component in natural copper corrosion. In most terrestrial, marine, and other exposed environments, the first corrosion product to form is always a layer of red cuprite. When tenorite appears, it typically indicates that the artifact was subjected to a heating process prior to or during burial. Pourbaix diagrams suggest that tenorite should be found in many different environments. In nature, however, the paucity of tenorite occurrences is controlled by kinetic and other factors that limit the mineral's formation to a few specific conditions—for example, high-temperature oxidation and high pH (Scott 2002). Combined with the results of Raman analysis that some of the corrosion products on the bronze artifacts contain a certain amount of charcoal, it can be further inferred that the bronze ornaments in the tombs might have been subjected to fire.

Limitations of Raman spectroscopy

Raman spectroscopy is widely used in the analysis of historical artifacts and archaeological samples, because it is extremely reliable, sensitive, compound-specific, and non-destructive. XRD in combination with X-ray fluorescence spectrometer have been mostly used for the identification of corrosion products. Traditional methods of diffraction analysis all require a certain amount of sample or the grinding of the sample into powders, making the application of such methods in archaeology limited. With the application of micro-diffraction and area detectors, the amount of needed samples has been greatly reduced, and no special sample preparation is required. Because of its high sensitivity, the exemption of sample preparation, and the small amount of sample required, Raman spectroscopy is increasingly used in the detection of bronze corrosion products. Many researchers have established databases related to metal corrosion products since the 1990s. For example, Bouchard et al. collected the Raman spectra of 45 minerals related to metal corrosion products and stained glass (Bouchard et al. 2003). Clark at the University of London, UK, collected the Raman spectra of 56 inorganic pigments used before 1850 (Clark 1995). Researchers from the Department of Earth Sciences, University of Arizona, USA, established an online free RRUFF database that provides Raman and XRD spectra of various minerals, which researchers can search by mineral name or elements contained.

However, as mentioned earlier, copper sulfide corrosion products are more difficult to identify owing to incomplete crystallization. Comparatively, XRD has been used earlier in the detection of bronze corrosion products because of its more complete database of spectra, making it easier to identify copper sulfides. It is extremely difficult to obtain Raman spectra from some of metal excess groups of sulphides such as chalcocite, digenite, bornite(Cu3FeS4), and djurleite because of the metallic character and thermal sensitivity (Mernagh et al. 1993). Together with the fact that HS often attaches to the surface of excavated artifacts, makes the samples prone to producing strong fluorescence, thereby most copper sulfides can not yield good quality Raman spectra. The results reported here indicate that covellite (CuS) has a good Raman signal, meanwhile, other information, like the shifts of the band positions and changes in the ratio of peak height, reflects thermal decomposition or new phase appearance.One of the products in the thermal oxidation of covellite is chalcocite, however, it is not clear whether the appearance at 284cm− 1 band (Fig. 7c) in the black corrosion product on Y6083 is due to the new apperance of the chalcocite, or if it is due to thermal decomposition of covellite. Although cassiterite being the dominant phase in the corrosion products according the XRD result, the Raman spectrum of SnO2 has no well Raman signals. Gettens (1951, 1969) deduced that the tin oxide in bronze corrosion layers could be in a near-amorphous state, Robbiola et al (1998) shared a similar view with Gettens. Therefore, only by jointly applying multiple analytical tools, taking the strength of each, and cross-checking the results can researchers arrive at a comprehensive understanding and accurate identification of samples.

Conclusions

Combining XRD and Raman spectroscopy, the corrosion products of these bronze ornaments are identified mostly sulfates and sulfides. Malachite, a corrosion product common in soil environments, is rare present in these bronze ornaments. In addition, we found a relatively rare corrosion product, tenorite. Copper sulfides are the main reason for the overall dark-colored appearance of the bronze ornaments.

The SO42− ions in the soil may be the primary cause of the sulfate corrosion products on the bronze ornaments. The moist, anoxic burial environment combined with the presence of sulfate-reducing bacteria in the soil might have further reduced the copper sulfates to produce a variety of copper sulfides. The presence of tenorite and charcoal indicates that the bronze ornaments might have undergone fire.

Raman spectroscopy has certain limitations for the detection of bronze corrosion products. Some copper sulfide corrosion products have no obvious Raman signal. Therefore, other analytical tools need to be adopted in conjunction to make a comprehensive and accurate identification of the samples.

Declarations

Acknowledgments   The authors appreciate the assistance of Mr. Chengyun Xiao and Guanghua Xiang in providing samples. Mr. Yong Xue provided assistance in obtaining XRD spectra with the D/max 12kW X-ray powder diffractometer. The authors also thank Mr. Lu Pan for the help in the guidance of the investigation of the corrosion causes.

Funding  This project was funded by National Cultural Heritage Administration.

Authors' contributions  All authors contributed to the study conception and design; sample collection: LRT, ML, NW; picture collection: LRT, ML, XYZ; data collection: XLC, NW; analysis and interpretation of results: XLC, ML, NW; draft manuscript preparation: XLC, XYZ; final manuscript preparation: all.

Author Declarations

Conflict of interest    The authors declare no competing  interest.

Ethics approval  Not applicable. 

Consent to participate  Not applicable. 

Consent for publication  Not applicable.

References

  1. Alon D, Mintz G, Cohen I, Weiner S, Boaretto E ( 2002) The use of Raman spectroscopy to monitor the removal of humic substances from charcoal: quality control for 14C dating of charcoal. Radiocarbon 44: 1-11
  2. Anastasia A, Antonis N, Constantinos V (2019) Bio-hydrometallurgy dynamics of copper sulfide-minerals probed by micro -FTIR mapping and Raman microspectrosco- py. Minerals Engineering 132: 39–47
  3. Bouchard M, Smith DC (2003) Catalogue of 45 reference Raman spectra of minerals concerning research in art history or archaeology, especially on corroded metals and coloured glass. Spectrochim Acta A 59: 2247-2266
  4. Chan M, Capek A, Brill DA, Garrett SJ (2014) Characterization of the patina formed on a low tin bronze exposed to aqueous hydrogen sulfide. Surf. Interface Anal 46: 433-441
  5. Cheng XL, Tang LR , Li M , Wu N ,Yang Q , Liu W, Zhang XY, Liu M, Han Y, Zhao DD(2019) Research on the composition and manufacturing techniques of gold leaves of gold -covered metal ornaments excavated from Chu tombs. Sciences of conservation and archaeology 1: 10-19 (in Chinese with English Abstract)
  6. Clark RJH (1995) Raman microscopy: application to the identification of pigments on medieval manuscripts. Chem. Soc. Rev 24: 187-196
  7. Colomban P, Tournié A, Maucuer M, Meynard P(2012) On-site Raman and XRF analysis of Japanese/Chinese bronze/brass patina – the search for specifific Raman signatures. J. Raman Spectrosc 43: 799-808
  8. Duncan SJ, Ganiaris H (1987)   Corrosion products on copper alloys and lead alloys from London waterfront sites. In: Black J (ed) Jubliee Conservation Conference: Recent Advances in the Conservation and Analysis of Artifacts, Summer Schools Press, London , pp 109-118
  9. Dunn JG, Muzenda C (2001) Thermal oxidation of covellite (CuS). Thermochimica Acta 369: 117-123 
  10. Fan XP, Wang Q, Wang YW (2020) Non-destructive in situ Raman spectroscopic investigation of corrosion products on the bronze dagger-axes from Yujiaba site in Chongqing, China. Archaeol Anthropol Sci 12: 90-99
  11. Gettens RJ (1951) The corrosion products of an ancient Chinese bronze. Journal of Chemical Education 28: 67–71.
  12. Gettens RJ (1969)  The Freer Chinese Bronzes, vol. 2: Technical studies. Smithsonian Institution, Washington, DC
  13. Gilbert B, Denoël S, Weber G, Allart D (2003)  Analysis of green copper pigments in illuminated manuscripts by micro-Raman spectroscopy. The Analyst 128:1213-1217
  14. Güncem Diktaş E (2017) A multi-analytical protocol combining EDXRF, SEM+EDS,μXRD and Monte Carlo simulations applied to analyze non-destructively Portuguese corroded copper coins. Dissertation, Unversity of  Evora
  15. Hayez V, Guillaume J, Hubin A,Terryn H (2004) Micro-Raman spectroscopy for the study of corrosion products on copper alloys: setting up of a reference database and studying works of art.  J. Raman Spectrosc 35: 732–738
  16. Hurma T, Kose S (2016) XRD Raman analysis and optical properties of CuS nanostructured film. Optik 127: 6000
  17. Martens W, Frost RL, Kloprogge JT, Williams PA(2003) Raman spectroscopic study of the basic copper sulphates— implications for copper corrosion and ‘bronze disease’.  Raman Spectrosc 34:  145–151
  18. McCann LI, Trentelman K, Possley T, Golding B (1999) Corrosion of Ancient Chinese Bronze Money Trees Studied by Raman Microscopy.  J. Raman Spectrosc 30: 121–132
  19. Mennella V, Monaco G, Colangeli L, Bussoletti E (1995) Raman spectra of carbon-based materials excited at 1064 nm. Carbon 33: 115–121
  20. Mernagh TP, Trudu AG (1993) A laser Raman microprobe study of some geologically important sulphide minerals. Chem. Geol 103: 113-127
  21. Robbiola L, Blengino JM, Fiaud C (1998) Morphology and mechanisms of formation of natural patinas on archaeological Cu–Sn alloys. Corrosion Science 40(12): 2083–111
  22. Schweizer F (1991)   In: Scott DA, Podany J, Considine BB (ed) Ancient and Historic Metals: Conservation and Scientific Research,  Proceedings of a Symposium Organized by the J. Paul Getty Museum, Getty Conservation Institute, Malibu, pp 33–50
  23. Scott DA (2002) Copper and bronze in art: corrosion, colorants and conservation. Getty Publications, Los Angeles
  24. Smith GD, Clark RJH (2002) The role of H2S in pigment blackening. J. Cult. Herit 3: 101–105
  25. Smith J, Qin Z, King F, Werme L, Shoesmith DW (2007)  Sulfide film formation on copper under electrochemical and natural corrosion conditions. Corrosion Science Section 63:135–144
  26. Trentelman K, Stodulski L, Lints R, Kim C(1999) A comparative study of the composition and corrosion of branches from Eastern Han dynasty money trees. Stud. Conserv 44:170–183
  27. Tuinstra F, Koenig JL(1970) Raman spectra of graphite.The Journal of Chemical Physics 53:1126–1130
  28. White SN (2009) Laser Raman spectroscopy as a technique for identification of seafloor hydrothermal and cold seep minerals. Chem. Geol 259: 240-252
  29. Xi Sh, Zhang X, Luan ZD, Du ZF, Li LF, LiangZW, Lian C, Yan J (2019) Micro-Raman Study of Thermal Transformations of Sulfide and Oxysalt Minerals Based on the Heat Induced by Laser   Minerals 9: 751
  30. Yang YH, Wang T (1997) Fourier transform Raman spectroscopic characterization of humic substances. Vibrational Spectroscopy 14: 105–112
  31. Yichang Prefectural Museum(1988) Chu state tomb No.5 at Caojiagang, Dangyang County. Acta Archaeologica Sinica 4: 455-500 (in Chinese with English Abstract)
  32. Young ML, Casadio F, Marvin J, Chase WT, Dunand DC (2010) An ancient chinese bronze fragment re-examined after 50 years: contributions from modern and traditional techniques. Archaeometry 52:1015–1043
  33. An integrated database of Raman spectra, X-ray diffraction and chemistry data for minerals. Available online: http: // rruff.info/ about/about general.php.