Archaeological textiles preserved by copper mineralization

Abstract

The mineralization mechanism responsible for the fossilization of archaeological textiles in close proximity to metal artifacts presents a sophisticated preservation process at both macro and micro levels. This study examines archaeological textiles dating from 2200 BC to AD 1900, sourced from three distinct archaeological sites. The focus is on understanding the microstructural degradation of fibers within a specific burial environment and the preservation achieved through mineralization. These archaeological fibers of archaeological textiles exhibit morphological preservation in the immediate vicinity of copper-based objects. Utilizing tools such as a digital camera, scanning electron microscopy with energy dispersive spectroscopy (SEM–EDS), high-resolution synchrotron-based microtomography (μCT), and enzyme-linked immunosorbent assay (ELISA), we examined fiber morphology, conducted elemental analysis, identified fiber types, and analyzed fiber characteristics. Our findings reveal the presence of smooth-surfaced wools and silks, fibers covered with calculi, and fiber impressions—all subjected to mineralization. These mineralized fibers can be categorized into three distinct stages of mineralization, each exhibiting varying carbon content. We inferred a correlation between mineralization rate and carbon content while also identifying mineralization density distribution on these textiles. Lastly, this study provides insights into the preservation states of textiles across three different mineralization stages, enriching our understanding of the deterioration of organic archaeological material.

Introduction

The mechanism of mineralization holds significant importance in the analysis of early organic fabrics across various disciplines, such as archaeology, material science, geoscience, and environmental science [1]. Mineralization, commonly defined as the replacement of physical shapes and constituent materials of fibers with minerals through a fossilization mechanism [2], frequently results in the recovery of mineralized textiles from burial environments in proximity to metal objects [3, 4]. Comprehending this mechanism aids in assessing the state of preservation of these textiles [5, 6].

In recent years, there has been a steady increase of archaeologists and conservators engaged in the study of archaeological textiles [7]. This growing interest is expanding the corpus of known ancient fabrics and raising awareness of textiles in general [8]. Previous studies have provided insights into textile pseudomorphs and impressions [9,10,11,12]. These studies revealed that organic materials can be preserved as metal compounds through mineralization [13]. In this process, the physical shapes of mineralized natural fibers are retained, while the organic structures undergo degradation and replacement [14], resulting in what is termed "textile pseudomorphs." This preservation is influenced by factors such as soil pH, air, and moisture content in the soil [15]. The mineralization process involves successive steps, including the water transport of biocidal metal cations and solutes from the environment, macromolecular degradation and loss of crystallinity in fibers, and mineralization [16, 17]. Despite the extensive research on the mechanism of mineralization through advanced techniques such as scanning electron microscopy (SEM), infrared spectroscopy (FTIR), enzyme-linked immunosorbent assay (ELISA), and computed tomography (CT) scanning technology, there has been limited exploration into its impact on the preservation of archaeological textiles [10, 18, 19].

In previous studies, researchers have primarily focused on mineralization behavior of textiles, while the protective role of mineralization on textiles has not received sufficient attention. This study approaches the research from the perspective of the protective function of mineralization on textiles. The aim of this study was to figure out the preservation state of textiles from three different mineralization stages obtained from three cemeteries. This involved examining the surface of archaeological textiles and analyzing mineral element information and protein residue.

Materials and methods

Materials

Archaeological fibers from three cemeteries were examined in this study, dating back to different periods: AD 1800–1900 from the Barkol Kazak Autonomous County cemetery in Hami, Xinjiang, China; AD 1300–1500 from the Yishigan Qinghe County cemetery in Altay Prefecture, Xinjiang, China; and approximately 2200 BC from the Qiemu'erqieke cemetery in Altay City, Altay Prefecture, Xinjiang, China. These fibers were morphologically preserved in close proximity to copper-based objects. All three excavation sites are located in Xinjiang, China, representing about one-third of all saline land in the country. Notably, around one-third of the cultivated land in Xinjiang has experienced secondary salinization [20]. The soils in Xinjiang are characterized by a high calcium content, with abundant elemental contents of Si, Al, Fe, Na, Ka, Mg, and S [21]. The soils in this region are frequently alkaline due to the relatively dry climate [22]. The relationship between soil dryness and alkalinity is that typically, drier soils tend to be more alkaline. This is because in drier soils, water evaporates, leaving behind alkaline substances such as sodium carbonate and calcium carbonate. As a result, the alkalinity of the soil increases. However, this relationship is influenced by various factors including soil type, climatic conditions, vegetation cover, and others. Despite these commonalities, there exist notable dissimilarities between the environments of the Barkol Kazak Autonomous County Cemetery, Yishigan Qinghe County Cemetery, and Qiemu'erqieke Cemetery, as detailed in Table 1.

Table 1 Details of the samples [23,24,25,26,27,28,29,30,31,32]

S1 represents an archaeological textile fragment situated in proximity to a copper buckle, displaying a fully preserved textile structure, as illustrated in Fig. 1A S2 consists of archaeological textile fragments associated with a copper mirror, showcasing a relatively intact textile structure. Additionally, S3 and S4 comprise powders scraped from the surface of a copper mirror, with S3 representing residue from the front and S4 from the back

Fig. 1

Photographs of the archaeological textile samples: A S1 from Barkol Kazak Autonomous County Cemetery; B S2 from Yishigan Qinghe County Cemetery; C S3 and (D) S4 from Qiemu'erqieke Altay City Cemetery

Methods

Scanning electron microscopy with energy dispersive spectroscopy (SEM–EDS)

Photographs of the samples were captured using a digital single-lens reflex camera (CANON EOS 30D). For the observation of longitudinal and cross-sectional morphologies, as well as the analysis of metal elements, SEM–EDS (JEOL JSM 5610LV) was employed. The appropriate portions of the samples were selected for cutting and pasting onto electrically conductive adhesive. They were handled with gloves, knives and tweezers. To enhance conductivity, the samples were coated with the thickness of the platinum coating 15 nm and examined at an accelerating voltage of 20 kV. The fibers retained their physical morphology, and copper elements were found on the fiber surface. We aimed to investigate whether the copper elements on the fibers would disappear after cleaning. Therefore, we cleaned the fibers using hydrogen peroxide. Notably, a portion of the sample from S2 underwent cleaning with 15% hydrogen peroxide (H₂O₂, soaking for 45 s) and was subsequently dried before sputtering.

Enzyme-linked immunosorbent assay (ELISA)

The advantages of using ELISA for fiber identification include its high sensitivity and specificity, allowing for the accurate detection of target proteins characteristic of different fiber types. In the ELISA test, a blank control, a negative control, and a positive control were established. phosphate buffer saline(PBS) with a pH of 7.4 served as the blank control. For the negative control, the fibroin or wool keratin polyclonal antibody was replaced by PBS with a pH of 7.4. The positive control utilized an extracting solution of wool keratin or fibroin. Detailed methods are outlined in our previous articles [33, 34] (The detailed experimental materials and methods are also provided in the appendix).

Synchrotron-based X-ray microtomography (μCT)

All samples were analyzed without any preliminary preparation. They were handled with gloves, knives and tweezers. Synchrotron-based X-ray microtomography (Xradia 610 Versa) was employed to observe the archaeological samples. The configuration was similar to the one described by Li et al. [35] (beam: 80 keV).

Results and discussion

Preservation via metal permeation

As illustrated in. Fig. 1, both samples from the Barkol Kazak Autonomous County cemetery (S1) and Yishigan Qinghe County cemetery (S2) remarkably maintain the structural integrity of their fabrics. The physical structure of the S2 is highly fragile, which poses a significant challenge to the sampling process. Initially, we employ optical microscopy to identify appropriate regions within the samples. Subsequently, these regions are excised and ultimately affixed onto conductive adhesive. Throughout this procedure, there is a considerable risk of disrupting the fabric structure of the samples. We also conducted multiple samplings to mitigate the potential for fortuitous outcomes. Despite the fragility of S2 during sample preparation, it exhibits a relatively complete surface morphology at the micro level, as depicted in Fig. 2. Notably, the fibers from these two cemeteries are highly preserved despite a burial time difference exceeding 1000 years.

Fig. 2

SEM images of archaeological textiles. A fibers from S1, B fibers from S2

In S1, certain fibers exhibit scales on their surfaces and possess medullary cavities, as illustrated in Fig. 3A. These fibers have a diameter ranging from 20 to 30 microns, aligning with the characteristics of wool fibers [36, 37]. Conversely, S2 displays fibers with a roughly triangular cross-section and a diameter of about 10 microns, reminiscent of the characteristics of silk fibers [34, 38]. To further identify the fibers in S1 and S2, we applied the ELISA method for immunoassay analysis. Immunoassays, known for their high sensitivity and specificity, hold significant potential as a diagnostic tool in cultural heritage research [39,40,41]. The results are presented in Fig. 4.

Fig. 3

Cross-sectional SEM images of archaeological fibers. A fiber from S1 and B fiber from S2 (Fig. 4B)

Fig. 4

ELISA results for S1 and S2; A detection of wool keratin(Positive: Wool keratin); B detection of silk fibroin(Positive: Silk fibroin protein)

ELISA Detection for S1 and S2 was conducted using PBS, with silk fibroin or wool keratin serving as positive controls(Refer to the appendix). The dashed line in the graph represents the test criterion (VAR_{negative control} × 3 × OD_{negative control}; VAR: variance; OD (Optical Density) refers to the absorbance of a detected substance at a specific wavelength, which is usually proportional to its concentration). Results were interpreted based on whether the OD_samples exceeded the dashed line, indicating a positive result identifying the characteristic component, and vice versa. As depicted in \* MERGEFORMAT Fig. 4, both OD_S1 and OD_S2 surpass the dashed line, confirming that S1 and S2 are wool and silk, respectively. Natural organic fibers, such as silk and wool fibers, are generally not expected to endure over extended periods [42,43,44]. The intriguing discovery here is the high preservation of these protein fibers in burial environments over millennia [45, 46], given that protein fibers, especially silk, are relatively resistant to acids but not to alkalis and are easily damaged by substances containing alkali. Notably, the pH of the excavation soils is higher than 8.0, as indicated in Table 1. To gain further insights into this exceptional preservation, we conducted an elemental analysis of the samples using SEM–EDS.

All spectra have been normalized. Elemental analysis of the fibers from S2 is presented in Fig. 5 and Table 2. Spectrum 1 corresponds to a "clean fiber" devoid of obvious contaminants, while Spectrum 2 and Spectrum 3 pertain to a fiber with some attached contaminants. In Spectrum 2, the "contaminant" is closely attached and exhibits a contrast similar to that of the fiber, whereas in Spectrum 3, the "contaminant" covers part of the fiber's surface, displaying a much stronger contrast. In EDS test results, wt% represents the mass percentage of elements, while at% represents the atomic percentage of elements. Wt% indicates the proportion of an element's mass in the sample, whereas at% indicates the proportion of an element's atomic quantity in the sample. The carbon content of the "clean" fiber (53.42 wt%, Spectrum 1) surpasses that of the contaminated fiber (36.60 wt% in Spectrum 2 and 34.88 wt% in Spectrum 3), with Spectrum 2 and Spectrum 3 showing marginal differences in carbon content.

Fig. 5

EDS profiles of fibers from S2

Table 2 Elemental analysis results of the fibers from S2

Table 2 indicates a substantial amount of Cu (7.38 wt%) on the surface of the silk fiber in Spectrum 1, suggesting that a portion of the organic phase in these fibers may have undergone replacement by an inorganic phase. Consistent with the ELISA results, where the S2 sample is lower than the positive control group, this may indicate a replacement of some silk fibroin proteins. This leads us to speculate that mineralization of the fibers has occurred. In essence, the composition of Spectrum 1 differs from that of Spectrum 2 and Spectrum 3. The differences in the elemental content between Spectrum 2 and Spectrum 3 are not significant. We support that S The disparities between the two should fall within the error margin of EDS analysis. Additionally, the SEM–EDS analysis of these fibers reveals a high content of Ca and Fe, suggesting soil as the primary source (High Ca and Fe contents have been observed in the soil of Xinjiang [47]). Ions such as calcium and iron in the soil may permeate into the interior of the fibers along with soil moisture, or deposit on the surface of the fibers, filling the gaps between them. These ions may potentially form altered minerals adhering to the surface of the fibers. Therefore, we used 15% H₂O₂ to clean the fibers.

After cleaning by 15% H₂O₂, impurities in the fibers from S2 were removed through hydrogen peroxide cleaning and subsequently tested using SEM–EDS. As depicted in Fig. 6, the fiber exhibits a smooth surface and a roughly triangular cross-section. Following cleaning, as indicated in Table 3, elements K, Al, and Fe were eliminated, and the carbon content increased. Negligible amounts of Si, S, and Ca were observed, while the copper content showed minimal change, supporting potential mineralization of the fibers.

Fig. 6

SEM cross-section of profiles of fibres from S2 cleaned by hydrogen peroxide

Table 3 Elemental analysis results for fibers from S2 cleaned by hydrogen peroxide

Figure 7 displays CT images of S1 and S2 (A and B depict textiles from S2, while C and D depict textiles from S1. A and C represent reflection imaging, and B and D represent projection imaging). As illustrated in Fig. 7B, the textile from S2 appears to consist of three superimposed layers (layer 1, layer 2, and layer 3). The diameter of a single fiber is approximately 10–15 μm, and the size of a fiber bundle ranges from 100–150 μm. Varied color intensity distributions indicate different densities of compounds present [9, 35]. Typically, inorganic mineral densities surpass those of natural fibers [48, 49]. An increased fiber mineralization rate leads to higher fiber density and corresponding increases in gray-level intensity. The silver-white layer (layer 3 in Fig. 7B) represents the fiber layer with a high level of copper mineralization. Mineralization near the copper source is intense, and different distributions of mineralization are observed for the textile. The gray level intensity changes from layer 1 to layer 3, suggesting the mineralization rates of these three layers of fibers are different—common in textile mineralization.

Fig. 7

CT images of the samples. A and B textile from S2. C and D textile from S1

Consequently, we support that S1 and S2 are textile pseudomorphs wherein part of the organic matrix of the fibers is replaced by inorganic matter. The physical shapes of the textiles are highly preserved. High levels of mineralization near the copper artefact and non-uniform mineralization for single fibers are evident. When copper artifacts corrode in burial environments, the released copper ions infiltrate into the fibers along with the soil moisture, permeating the interior of the fibers and replacing the organic tissues within. This ultimately leads to the positive mineralization of the fibers, ensuring their preservation. The infiltration of metal ions into the fibers induces mineralization, thereby preserving the fibers. This mineralization is a result of the permeation of copper ions, contributing to the well-preserved state of the fibers. We term this preservation mechanism as " preservation via metal permeation ".

Preservation via armoring

Although S3 is presented in powder form in. Fig 1C, SEM imaging (Fig. 8) clearly reveals fibers with a diameter of approximately 12 microns. Notably, particulate matter is adhering to the fiber surfaces, creating a protective layer or "armor" around the fibers. Given the fiber diameter, we hypothesize that these fibers may be silk [34].

Fig. 8

SEM profiles of copper mineralized archaeological textile from S3

To further ascertain the fiber type, we conducted ELISA testing. OD_S3 falls below the dashed line in Fig. 9, indicating a negative result. Thus, the amount of silk fibroin in S3 is extremely low or non-existent. Two possible explanations for this phenomenon are considered: (1) the fibers are not silk; they could possibly be linen or other types of fibers, and (2) with increasing mineralization, the protein content in the fibers decreases. Regardless of the fiber type, it is evident that the morphology of the fibers in Fig. 8 differs from that of fibers undergoing permeation preservation (Fig. 5).

Fig. 9

The result of ELISA for S3 (Positive: Silk fibroin protein)

The particulate matter is affixed to the fiber surfaces, creating an "armor-like" appearance. We selected the "fiber" (Spectrum 1) and the "armor" (Spectrum 2) for testing. The carbon content of both was similar, but the nitrogen and copper content of the former exceeded that of the latter. Additionally, phosphorus was detected on these fibers, distinguishing them from S2. Phosphogenesis is a complex natural phenomenon influenced by biogenic factors [50]. Phosphorus addition can also result in globular structures [51, 52] similar to that yielded by bio-mineralization [53,54,55].

Compared with positive mineralized fibers, the fiber in Fig. 8 is enveloped in a layer of " mineral armor ", preserving the fiber. In addition to being penetrated by copper ions, fibers can also undergo deposition of copper ions on their surface, along with other ions present in the soil. We term this preservation type as armoring preservation. Two potential reasons for the "armor" formation are considered: sedimentary mineralization (permineralization or replication of morphology in authigenic mineralization [7]) and calculus formation, involving crystal nucleation, nuclei growth, and nuclei aggregation. The presence of phosphorus in the "armor" suggests the potential involvement of microorganisms in its formation.

Preservation by impression

Figure 10 displays a fiber trace resembling a fossil embedded in the back of a bronze mirror (Spectrum 1). The diameter of the "fiber fossilization" is approximately 15 microns. Spectrum 2 represents an area at the back of the bronze mirror. SEM/EDS testing results (Table 4) indicate that C, O, and Cu are the primary elements in the fiber trace. The carbon content in Spectrum 1 (32.64 wt%) surpasses that in Spectrum 2 (22.10 wt%). Compared with other samples, the fiber carbon content in Fig. 10 is the lowest, while the copper content is the highest. Considering that fiber mineralization is a process where organic matter in fibers is replaced by inorganic substances, the originally present carbon elements in fibers may decrease as mineralization progresses, while copper elements may increase with mineralization. The fiber undergoes negative mineralization. The fibers undergo negative mineralization and are preserved in the form of imprints. We term this preservation type " preservation by impression". (Table 5).

Fig. 10

SEM–EDS analysis results for fiber from S4

Table 4 Elemental analysis results for fibers from S3

Table 5 Elemental analysis results for fiber from S4

The preservation state of archaeological textiles is influenced by the burial environment and duration [14]. These textiles are interred in alkaline soil with significant temperature variations (Table 1). S1 is buried in Barkol Kazak Autonomous County, Hami City, which experiences a temperate continental arid climate. However, being situated in the Tianshan Mountains, Hami City exhibits noticeable climate differences. The southern region is dry and hot with minimal rainfall, while the northern part is cooler with slightly more precipitation. Barkol Kazak Autonomous County, in the northern part of Hami City, has an annual average temperature of 1 ℃. Predominant soil types include chestnut soil and brown calcium (pH 8.34). This geographic setting suggests that S1 was buried in extreme conditions, which are unfavorable for bacterial growth, and benefits the protection of fibers. The favorable burial environment and short burial time (100–200 years) contribute to the robust fabric structure of S1. During the sampling process, we observed that sample S1 possesses a relatively robust physical structure. Using tweezers to extract the sample does not damage its fabric structure, and the sample cannot be torn apart by tweezers; instead, scissors or a small knife are required for sampling.

S2, S3, and S4 originate from Altay City. S2 is from Yishigan Qinghe County cemetery, Altay, while S3 and S4 are from Qiemu'erqieke Town cemetery, Altay. Despite having different morphologies, S2 retains its fabric structure, although it tears easily with tweezers. On the other hand, S3 and S4 exist in a powdered state, likely due to prolonged burial time and unique burial conditions. S2 dates back approximately 1200 years, while S3 and S4 have been buried for about 4200 years. Altay City, located in the Asian hinterland, exhibits climate zones ranging from intermediate temperature to semi-arid and arid regions. Distinct climate conditions arise from the presence of the Altai Mountains in the north and the Junggar Basin in the south, resulting in significant climate variations between the north and south. Chestnut soil, brown calcium soil, and meadow soil predominate in Altay City (pH 8.0 ~ 8.5), making the soil types in their burial environments the same. However, Qiemu'erqieke cemetery, situated south of Altay City near the Irtysh River, experiences dry conditions even adjacent to the river. The Irtysh River relies on snow-melt and mountain precipitation as its primary water sources, exhibiting limited runoff and significant seasonal variations. The high ion content and sufficient water in the soil facilitate the migration of metal ions. Consequently, S3 and S4 are more influenced by metal ion exchange compared to S2. A high metal ion exchange rate and extended burial time enhance mineral deposition and the replacement rate of mineralization. Additionally, the river's influence on ambient temperature fosters biodiversity. Phosphorus and biological mineralization are evident in S3 due to these burial conditions. This burial condition also amplifies the mineralization rate of S4, resulting in lower carbon content and copper levels comparable to the copper mirror area.

Conclusion

This study examined three archaeological mineralized textiles ranging from the early Bronze Age in China to the late Qing Dynasty, uncovering three distinct stages of mineralization that preserve archaeological textiles.

In burial environments, copper artifacts undergo corrosion, releasing metal copper ions. These ions permeate into the interior of fibers along with soil moisture, replacing the organic components of the fibers and leading to positive mineralization, thereby preserving the fibers. This phenomenon is similar to '' pre-mineralization ''. Preservation via metal permeation: archaeological fibers retain their complete physical shape, often referred to as pseudomorphs or casting. In this stage, the fiber's surface appears relatively smooth and clean, with the mineralization rate exhibiting varying distributions across the textile. The proximity to copperware results in a higher mineralization rate, and the mineralization concentration of fibers between layers is uneven. In addition to being penetrated by metal ions, metal and soil elements also deposit on the surface of fibers. Compounds deposited on the fiber surface tightly envelop the fibers, forming a layer of 'armor-like' substance. Preservation via armoring: Fibers are well-preserved within thin inorganic metal layers surrounding the organic core. After the formation of "textile pseudomorphs," active sites on the casts undergo crystal nucleation, nuclei growth, and nuclei aggregation. Subsequently, the degree of mineralization enters a new stage, where fiber surfaces are covered with "calculi." Alternatively, metal salts may deposit on fibers, forming armor that protects them from attacks by other microorganisms. As fibers undergo positive mineralization, metal ions also deposit within the gaps between fibers along with water molecules from the soil, forming a '' deposition layer ''. When the mineralized fibers corrode, imprints of the corroded fibers are left on the 'deposition layer.' It represents a negative mineralization phenomenon. These imprints can preserve the morphology of the fibers, which can be considered a form of protection: preservation by impression. Complete mineralization of fibers occurs, where the organic phase is comprehensively replaced by an inorganic phase. Fibers in this stage are preserved in the form of impressions, and the carbon content gradually decreases.

This study reveals how ancient textiles buried hundreds to thousands of years ago are preserved through three distinct states. However, due to the limited number of samples, some of the conclusions in this paper may have a degree of variability. We will investigate additional samples to supplement our findings.