Scientific analysis of Tang Dynasty Go pieces excavated from the Lafuqueke Cemetery in Xinjiang

The game of Go, known as “Weiqi” in Chinese, is a cultural treasure with a history that spans millennia and is the oldest and most globally influential board game originating from China. The earliest written reference to the game is generally recognized as the historical annal Zuo Zhuan (c. 4th century BCE)¹. Although its origins remain a topic of debate among scholars, Go has long been integral to Chinese culture. Initially a leisurely activity during the pre-Qin period, Go transitioned to an art form by the Eastern Jin dynasty (317–420 CE)². The Sui (581–618 CE) and Tang (618-907 CE) dynasties saw Go reach its peak in development³. During this period, the design of Go boards and pieces evolved, with the board’s grid lines becoming standardized to 19 × 19 after the Northern and Southern Dynasties (420–589 CE). Subsequently, Go was considered one of the four essential arts of the cultured aristocratic Chinese scholars⁴.

During the Tang Dynasty, the emperors’ profound love for Go led to the establishment of the position of “Qi Boshi” (Chess Doctor) within the imperial court, specifically responsible for teaching the game to court members⁵. In addition to the emergence of the “Qi Boshi”, the institution of the “Qi Daizhao” (Chess Player Awaiting Edict) system further exemplified the professionalization and popularization of Go activities⁵. Go became widely popular among the aristocracy and literati, eventually extending its reach to women of various social strata, from noblewomen and ladies of high birth to ordinary women in society. The widespread engagement of women in Go is well-documented, with poets such as Wang Jian extolling scenes of women playing Go in his poem “Watching the Beautiful Woman Play Go at Night”.

Moreover, Go activities gradually spread beyond China, reaching neighboring countries such as Japan, Goryeo, Silla, and Baekje. During the Tang Dynasty, Japan was particularly notable in this regard. A set of Go board and pieces used by Emperor Shōmu, currently preserved in the Shōsōin repository in Japan, indicates the Japanese imperial court’s fondness for Go during this period⁶. Additionally, Go spread westward along the Silk Road, reaching regions such as the Anxi Protectorate. Archeological findings, such as the Go board discovered in the tomb of Zhang Xiong and his wife, provide concrete evidence of the game’s format during that era (Fig. 1a). Beyond physical artifacts and written records, numerous depictions of Go scenes have been captured in various Tang Dynasty artworks. These include engravings on gold and silver vessels (Fig. 1b), carvings on stone slabs in tombs (Fig. 1c), and images cast onto bronze mirrors, as well as the most direct representations in silk paintings (Fig. 1d), offering rich visual documentation of Go being played.

a Wooden Go board excavated from Tomb 206 (Tomb of Zhang Xiong and his wife) in Astana, Turpan City (after ref. ³⁸). b Engraved figures on a gilded silver incense burner from the underground palace of Famen Temple in Fufeng County, Baoji City (after ref. ⁴⁰). c Line drawing of a Go playing and watching scene on the east panel of a stone coffin tomb in Changying Village, Yichuan County, Luoyang City (after ref. ⁴¹). d Silk Painting of a Lady Playing Go unearthed from Tomb 187 in Astana, Turpan City (after ref. ³⁹).

Full size image

Despite the discovery of many Tang Dynasty Go artifacts, including boards, pieces, boxes (jars), as well as Go-related texts, paintings, and carvings, there has been a lack of in-depth research on these materials beyond bibliographic studies. In particular, the identification of materials remains at a very rudimentary level, often categorized simply as jade, wood, ceramic, or stone⁵. For example, most reports describe Go pieces merely as “stone” without distinguishing their specific mineral types. This lack of detailed material analysis significantly hinders our ability to understand the handicraft industry, production processes, and circulation networks related to Go equipment during that period.

Given the current gaps in research, we directed our focus toward Tang Dynasty Go pieces unearthed in the Xinjiang region. By employing advanced laboratory techniques, we conducted a comprehensive analysis of the materials, design, and craftsmanship of these Go pieces. Our primary aim is to address existing research deficiencies by providing a detailed examination of the production techniques and material choices associated with Tang Dynasty Go equipment. Through rigorous material characterization and technical comparison, we aim to elucidate the technological and cultural advancements of Go during this period.

This study enhances our understanding of Tang Dynasty Go artifacts and provides valuable insights into the material culture and technological practices of the time. It offers a significant contribution to the field of historical artifact analysis and enriches our comprehension of the evolution of Go as both a cultural and technological phenomenon.

The game pieces in this study were excavated from a Tang Dynasty tomb in the Lafuqueke Cemetery. The cemetery is located in Bostan Village, Wubao Town, Yizhou District, Hami City, Xinjiang Uygur Autonomous Region, situated on an earthen terrace on the east bank of the Poplar River, ~70 m north of the ancient city of Lafuqueke (Fig. 2). The rectangular terrace measures 150 m in length from north to south, 73 m in width from east to west, and stands about 3.5 m high, covering a total area of around 10,000 m². Rescue archeological excavations were conducted from September to December 2019 and from June to July 2020.

The location of Lafuqueke Cemetery.

Full size image

Established around the 3rd century BCE, Loulan became an important kingdom along the Silk Road. In the 1st century BCE, Loulan was renamed Shanshan. Throughout the Northern and Southern Dynasties, Shanshan endured repeated invasions, prompting many inhabitants to migrate. Some Shanshan people resettled in what is now Hami City, where they established fortified communities and lived in close-knit groups. Their presence led to the area becoming known as Nazhi. The Lafuqueke site, identified as the ancient Shanshan people’s Nazhi City built during the Tang Dynasty, has revealed significant historical artifacts. So far, a total of 102 ancient burials, one kiln, and one well have been discovered. The burials include 21 tombs with sloping passages, 20 tombs with vertical caves, 58 above-ground tombs, and three infant burials with unclear architectural forms⁷. The tombs with sloping passages fill a historical gap in the westward transmission of tomb styles from the Central Plains⁸.

Specifically, the Go pieces were unearthed from tomb M35. This tomb also yielded burial objects exhibiting both Chinese and Central Asian cultural characteristics, directly showcasing distinct cultural influences. These artifacts include a Sasanian Persian silver coin, a tiger-and-dragon mirror, a necklace made of agate and gold-inlaid glass beads, an iron hairpin, an iron brooch, iron tweezers, iron scissors, a wooden comb, and a white porcelain makeup box (Fig. 3). The bronze mirror exemplifies typical Han Dynasty style, while the silver coin, minted during the reign of Sasanian ruler Jamasp (496–498 CE), features an encircling beaded border, Pahlavi inscriptions, the date “Year 2” on the left, and the minting location, “Ahvaz,” on the right, establishing the tomb’s earliest possible date. Meanwhile, the white porcelain makeup box, a popular item in the Tang Dynasty, further confirms that the tomb dates to this period.

Some fine artifacts from the tomb M35.

Full size image

Seventeen white Go pieces and two black Go pieces were placed inside a clam shell with dimensions of 6.2 cm in length, 5.7 cm in width, and 1.7 cm in height. This clam shell was positioned next to the head of the tomb occupant (Fig. 4). Certainly, the discovery of the Go pieces underscores the significant presence of Central Plain culture in this region. It demonstrates how Go, as a form of leisure and entertainment, reached the Western regions and even extended further along the Silk Road.

A Close-up image; B Overall view.

Full size image

For detailed study, with the generous permission of the archeological sample custodians, one black piece and one white piece from the set were selected for comprehensive laboratory analysis (Fig. 5).

Front, back, and side views of the black and white Go pieces.

Full size image

Digital microscopy

Initial observations were conducted using the HIROX RH-2000 3D digital LM, which offers a wide depth of field, large working distance, and high-intensity LED lighting. The microscope, features ultra-fast autofocus, a multifocal system, a Z-axis stepping motor (50 nm per step), and a motorized rotary head with 360° rotation.

Raman spectroscopy

The Raman spectrometer (XploRA Plus, Horiba) featured a CCD detector (−60 °C, 1024 × 256 pixels), a 600 grooves mm⁻¹ grating, a confocal microscope, and an automated X, Y stage. Laser power was adjusted with a density filter to prevent thermal effects. Spectral data and images were acquired using a 10× objective lens (N.A. = 0.25). A 532 nm solid-state laser with 1 mW output power for 1 s was used to avoid mineral transformation.

XRD

The experiment was conducted using a Bruker D8 Discover micro X-ray diffractometry (µ-XRD) equipped with a Cu Kα radiation source (λ = 1.54184 Å) and a two-dimensional VANTEC-500 detector. The system utilized a micro-focus beam with a spot diameter of 400 μm, operating at a tube voltage of 50 kV and a tube current of 1 mA. The diffraction data were collected over a 2θ range of 18–75 degrees using step-scanning mode with three steps, each step taking 100 s. The goniometer radius for the measurements was 198 mm.

XRF

The μ-EDXRF spectrometer M4 Tornado (Bruker Nano GmbH, Germany) was used for high-resolution elemental data acquisition, operating with a tungsten tube (50 kV, 600 μA, 30 W). The polychromatic beam is focused to a spot size of 17 μm at 17.48 keV, increasing to 32 μm at 2.29 keV. Two XFlash 430 Silicon Drift detectors detect signals under vacuum conditions (20 mbar), with short acquisition times (1 ms per 25 µm spot). This setup ensures accurate concentration determination, light element identification, and reduced background noise. The working distance was controlled to avoid applying pressure to the sample. Based on its advantages, this approach can be used to obtain elemental data for samples at large scales.

SEM-EDS

Scanning electron microscopy (SEM) was conducted using a Phenom Prox coupled with an energy dispersive X-ray spectrometer (EDS) system. For SEM analysis, specimens were adhered to conductive tape to improve observation quality and enhance data acquisition. The SEM observation and EDS analysis were carried out at 15 kV, an acquisition time of 90 s, and a consistent working distance maintained between 12 and 15 mm. It is more suitable for observing and testing elemental features in localized areas.

Craftsmanship details

The Go pieces have undergone weathering or erosion due to the burial environment. Notably, the white piece has become more porous on one side, with some areas developing a brown crust and others exhibiting yellowish-white speckled crystals that appear to have grown from the surface. The black piece has a small chip on the edge, possibly due to use or inadequate craftsmanship, and its surface is scattered with white impurity minerals. Overall, the black piece is smoother and exhibits a waxy luster compared to the white piece. The visible grinding marks indicate that sanding and polishing were integral steps in their production. Additionally, the color of both black and white Go pieces closely matches that of modern standard ones.

Both pieces display a convex oblate shape on both sides, with slight differences in exact shape, but relatively close in size and weight. The black Go piece has a diameter ranging from ~11.47 mm to 11.89 mm, a thickness of about 4.20 mm, and weigh 0.82 grams. In contrast, the white Go piece has a diameter ranging from ~11.53 mm to 11.73 mm, a thickness of about 5.56 mm, and weigh 0.90 grams (Fig. 5). The superior roundness of the pieces clearly indicates that they were produced under a stringent quality control process.

For comparative study, information on well-documented official archeological finds of Go pieces from the Tang Dynasty and earlier periods was collected^2,5,9. While most records include dimensions, almost none provide weight measurements (Table 1). Generally, Tang Dynasty Go pieces range from 1.1 to 2 cm in diameter, with thicknesses varying from 0.3–0.4 cm for thinner pieces to up to 0.8 cm for thicker ones. The Go pieces from the Lafuqueke site are smaller than many other Tang Dynasty pieces, closely resembling those excavated from the Shouchang Ancient City Site in Dunhuang, as well as the black Go pieces from the Zheng Shaofang Tomb in Yanshi, Henan.

According to unearthed artifacts and historical records, the shape of Go pieces had become standardized by the Tang Dynasty, with constraints on their dimensions. Prior to the Sui and Tang dynasties, most Go pieces were square or nearly square in shape and relatively small in size, with no evidence of regular circular pieces. This shape might have been influenced by other board games of the time, such as Liubo².

The use of square or polygonal Go pieces requires a specific orientation when placed on the board to avoid a cluttered appearance. This shape not only diminishes the visual esthetics but also hinders the players’ ability to observe the game clearly. Unlike other board games, Go is characterized by a large number of pieces, which continually increase on the board during gameplay. Using polygonal pieces necessitates attention to their orientation, potentially distracting players and affecting their focus and strategic thinking. However, the existence of these polygonal pieces also highlights the early developmental stage and the initial lack of refinement in Go’s evolution⁵.

In terms of the colors of the Go pieces, black and white were predominant, aligning with Eastern Han scholar Ban Gu’s description in the “Yi Zhi” (Purpose of Go) that “the pieces are black and white, representing yin and yang”¹⁰. However, there were also instances where pieces were differentiated by varying shades of color, primarily due to the natural properties of the materials used to make them. However, from the Tang Dynasty onwards, the color of Go pieces became standardized to black and white, establishing a fixed convention.

Material Identification

Mineral composition

Raman spectroscopy analysis

The Raman spectrum of the white Go piece displays four distinct characteristic peaks (Fig. 6). The peak at 1085 cm⁻¹ corresponds to the symmetric stretching vibration mode of C–O, the peak at 712 cm⁻¹ corresponds to the bending vibration mode of O-C-O, the peak at 284 cm⁻¹ corresponds to the translational rotation mode of CO₃²⁻ in CaCO₃, and the doubly degenerate mode at 161 cm⁻¹ is caused by translation of the CO₃²⁻ molecules¹¹. Additionally, there are two weaker peaks associated with the internal vibrations of the CO₃²⁻ group: the out-of-plane bending vibration at 1748 cm⁻¹, the antisymmetric stretching vibration at 1436 cm⁻¹¹². Calcium carbonate (CaCO₃) can occur in three polymorphs—calcite, aragonite and vaterite—each characterized by unique and non-overlapping Raman peaks. The characteristic peak near 712, 284 amd 161 cm^-1 unequivocally identifies the presence of calcite¹³.

Raman spectra of black and white Go pieces.

Full size image

The Raman spectrum of the black Go piece reveals five distinct characteristic peaks located at 1049, 672, 542, 348, and 201 cm⁻¹, consistent with the Raman spectrum of chlorite (Fig. 6). Specifically, the peak at 1049 cm⁻¹ can be attributed to the Si-O stretching vibrations within the silica tetrahedron. The peak at 672 cm⁻¹ corresponds to Si-O-Si stretching vibrations, the peak at 542 cm⁻¹ is associated with Al-O-Si vibrations, and the lower wavenumber peaks are due to lattice vibrations in the TOT layers of phyllosilicates^14,15. In Fe-rich chlorites, a notable change in the Raman spectrum is the shift of the Si-O_b-Si mode peak to near 671 cm⁻¹, whereas in Mg-rich chlorites, this peak typically appears in the 681–683 cm⁻¹ range. Additionally, Fe-rich chlorites commonly exhibit a reduced intensity of the peak near 350 cm⁻¹^16,17. These characteristics indicate that the black Go piece is iron-rich.

µ-XRD analysis

µ-XRD technology presents numerous advantages in the analysis of cultural relics¹⁸. This method enables non-destructive testing, which is essential for preserving valuable artifacts¹⁹. It is straightforward and requires no complex pre-treatment; the original sample can be analyzed directly with minimal material, as the small X-ray spot size effectively eliminates interferences²⁰. Additionally, it is particularly useful for examining special samples with varying degrees of aging, low information content, or low crystallinity, facilitating the discovery of critical information²¹.

Although the diffraction angle 2θ started only from 18 degrees, and the preceding segment of information was unavailable, clear XRD patterns were obtained within the accessible range (Fig. 7). Using software Jade 6.5 for mineral matching based on diffraction peak positions, we identified that the white piece is made of calcite, while the black piece is consistent with clinochlore, one of the Fe-rich chlorites. The XRD results not only validates the above conclusions but also provide additional detailed insights.

XRD patterns of black and white Go pieces.

Full size image

For instance, the diffraction peak (024) of the white Go piece (approximately at 2θ = 47.11 in Fig. 7) is absent. One plausible explanation is a phase transition caused by high temperatures^22,23. Simulated experiments indicate that the diffraction peak (024) gradually shifts towards higher angles, while the diffraction peak (018) (approximately at 2θ = 47.50 in Fig. 7) shifts towards lower angles. These peaks merge at temperatures between 573 and 673 K and separate again around 773 K²⁴. However, it cannot be ruled out that the disappearance of this diffraction peak may be due to a phase transition resulting from the prolonged burial of the sample.

On the other hand, chlorite has two subgroups: dioctahedral and trioctahedral, which can be effectively distinguished by their d(060) values²⁵. The dioctahedral subgroup ranges from 1.49 to 1.50 Å, while the trioctahedral subgroup ranges from 1.53 to 1.57 Ų⁶. The d(060) value of the black Go piece is 1.54 Å, indicating that the chlorite present is of the trioctahedral type²⁷. No significant diffraction peaks of other minerals were observed in the sample.

Elemental characteristics

A non-destructive elemental mapping by μ-EDXRF was performed on the clean and weathered areas of the white Go piece and the clean area of the black Go piece. To minimize interference from air, the reported results exclude elements with atomic numbers below 8. The data represents an average of the analyzed regions and has been normalized accordingly (Table 2).

In the pristine white areas of the white Go piece, Ca is the dominant element, constituting over 96%, indicating a high purity of calcite. Based on its lower hardness and more porous nature, it can be inferred that the calcite is present in limestone rather than in marble. Other elements present in higher concentrations include Si, Mg, and Al, likely originating from dolomite and clay minerals, which are common impurities in limestone^28,29.

The classification of chlorite minerals into three subgroups relies on the Fe²⁺/R²⁺ ratio (with R representing other divalent cations) and the specific site occupied by Si in the formula³⁰. These subgroups include: (a) penninite, clinochlore, and sheridanite; (b) diabentite, brusnvigite, and ripidolite (also known as prochlorite); and (c) delessite, chamosite, and thuringite¹⁵. The cation count for the black Go piece was calculated based on 14 oxygens, resulting in a Si coefficient of 2.97, which falls within the range of 2.8–3.0. Additionally, the Fe/R²⁺ ratio is ~0.238, with Fe²⁺/R²⁺ being less than 0.25. These characteristics further confirm that the black Go piece is primarily composed of clinochlore³¹.

The elemental mapping results visually reveal several detailed features (Figs. 8, 9). First, the weathered areas on the white Go piece do not entirely coincide with the distribution of any single element, but there are significant correlations with the distributions of elements such as Si, Al, Mn, P, and Pb. Based on the elemental content, it is likely that these areas are composed of clay minerals and some associated minerals. Due to the higher surface area and negative charge of clay minerals, they more readily adsorb metal ions such as lead³². Additionally, Fe and Mn do not show a similar distribution pattern. Fe is more evenly distributed and shows higher concentrations in smaller spots. This explains why the black Go piece, composed of clinochlore, appears uniformly black, which may be attributed to its higher Fe content³³. Finally, the comparison shows that the elemental distribution of the black Go piece is more uniform, indicating less influence from weathering or other burial factors^34,35.

Element mapping of the white Go piece (Higher concentrations result in brighter colors).

Full size image

Element mapping of the black Go piece (Higher concentrations result in brighter colors).

Full size image

Based on the SEM-EDS results, various micro-regions containing impurity minerals can also be observed (Fig. 10). Iron-rich minerals are prevalent on the black Go piece, which also exhibits a relatively high carbon content. The carbon distribution closely mirrors that of iron. Although the high carbon content may be influenced by the low-vacuum conditions used during analysis, its uneven distribution suggests that the black Go piece indeed contains a significant amount of carbon, probably in the form of graphite. This graphite presence explains the uniformly black appearance of the piece. Nonetheless, other factors may contribute to clinochlore’s dark coloration. While less common, impurities such as manganese or titanium can also darken clinochlore. Furthermore, inclusions of other black or dark minerals, such as magnetite or ilmenite, may contribute to an overall darker hue, especially when these inclusions are abundant or well-distributed within the material.

a Central region; b Edge region.

Full size image

History records

During the Tang Dynasty, Go pieces were primarily made from two types of materials: fired ceramic and natural minerals. Examples of ceramic Go pieces have been unearthed from Bai Juyi’s Residence Site in Luoyang City, the Liuzi Canal Site in Huaibei, and the Huangbao Kiln in Tongchuan, Shaanxi (Table 1). Historical records from the time also mention such pieces. For instance, poet Fu Mengqiu’s work “Guoqi Fu” (An Ode to Go) includes the line: ‘The board is made of exquisite Catalpa bungei wood, and the pieces come from the kilns of Yunnan’². Specifically, these ceramic Go pieces are made from materials sourced from Yunnan Province and produced by sintering a proprietary mixture of mineral compounds derived from local stones³⁶.

However, the majority of Go pieces encountered are crafted from natural minerals. Based on the contents of the “Tang Dynasty Gazetteer Fragment,” one of the precious documents preserved for nearly a 1000 years in the Library Cave of the Mogao Caves, it is recorded under the entry for Shazhou in Longyou Circuit: ‘Dunhuang…, tribute items: Go pieces’⁵. Similarly, recorded in the Tongdian (Comprehensive Institutions), it states, “Each prefecture in the empire must annually contribute local products as tributes according to the regulations… Dunhuang Prefecture contributes twenty sets of Go pieces.” From the 1st year of Wude (618–626 AD) to the mid-period of Kaiyuan (713–741 AD) and Tianbao (742–756 AD) during the Tang Dynasty, the court mandated that prefectures across the country offer tributes. However, only Dunhuang (later renamed Shazhou) was specifically required to contribute Go pieces, and in significant quantities—20 sets. This underscores the renowned status of Dunhuang’s Go manufacturing industry, which supplied Go pieces nationwide, including to the imperial court. The Go pieces from Dunhuang, irrespective of their color, shape, or quality, satisfied the needs of various social strata^5,9. Additionally, the New Book of Tang records, “Shazhou, Dunhuang Prefecture, subordinate to the governor-general. Local tributes: Go pieces, yellow alum, gypsum”. This further confirms the prominence of Dunhuang Go pieces as a primary local specialty tribute to the Tang court, highlighting their widespread renown during that period⁵.

In the city of Shouchang, under the jurisdiction of Dunhuang, numerous Go pieces have indeed been discovered. These pieces, small flat-round stones in black and white, are uniform in size. Characterized by a central convex area with gradually thinner edges, they are likely the tribute items from that period². If each set of Go pieces follows the 19 × 19 grid pattern used during the Tang Dynasty, 361 pieces are required per set. For twenty sets, a total of 7220 pieces would be needed. These Go pieces were crafted by hand, requiring significant effort and skill from the artisans in the workshops, especially for those pieces intended as tributes, which demanded meticulous workmanship. This was by no means an easy undertaking and necessitated a sizable industry to support such production.

The stones used for producing Go pieces were likely sourced from the Qilian Mountain region near Dunhuang, a major source of minerals such as calcite and chlorite, according to geological data³⁷. The unique geographical environment and rich mineral deposits provided excellent conditions for the development of the Go equipment manufacturing industry in Dunhuang. While current scientific evidence cannot definitively confirm that these Go pieces from Lafuqueke originated in Dunhuang, historical records and geographical proximity make this highly plausible. If validated, this finding implies that Go pieces produced in Dunhuang may have been distributed not only eastward to the Central Plains but also westward along the Silk Road, either through official channels or via private trade. Further research on Go pieces is certainly warranted, and conclusions should be based on rigorous scientific investigation rather than traditional visual inspection.

Remarkably, this discovery represents the first instance of actual Go pieces from the Tang Dynasty found in Xinjiang. While previous discoveries of Go boards and paintings depicting Go scenes have suggested the game’s spread in the region^38,39, the presence of Go pieces as burial artifacts provides definitive evidence of its popularity. This finding indicates that Go was not only valued by the elite but also embraced by the middle class and even commoners. Significantly, the burial included personal grooming items, suggesting that Go was a meaningful part of the deceased’s daily life, deeply integrated into their routine, and thus chosen to be placed carefully near the head in burial. Such discoveries enhance our understanding of daily life along the Silk Road, and we anticipate further finds that will shed light on this rich cultural landscape.

Conclusion

This study provides a comprehensive scientific analysis of the Go pieces excavated from the Tang Dynasty tomb at the Lafuqueke Cemetery in Xinjiang. By employing analytical techniques such as digital microscopy, Raman spectroscopy, XRD, XRF, and SEM, we have investigated the craftsmanship and identified the mineral compositions and of these artifacts.

The Go pieces exhibit distinctive features, such as weathering and erosion patterns, which are indicative of their prolonged burial environment. The black Go piece, with its waxy luster and smooth surface, contrasts with the more porous appearance of the white Go piece. Our comparative study with other well-documented Tang Dynasty Go pieces indicates that the Lafuqueke Go pieces are smaller but closely resemble those from the Shouchang Ancient City Site in Dunhuang and the Zheng Shaofang Tomb in Yanshi, Henan. This suggests a possible regional variation in Go piece production and standardization during the Tang Dynasty.

After multiple methods of cross-validation, the black Go piece was confirmed to be composed of clinochlore, an Fe-rich chlorite, while the white piece was identified as calcite. Historical records indicate that Dunhuang regularly contributed Go pieces as tributes to the imperial court, suggesting that the region had abundant resources. These findings support the hypothesis that the stones used for producing these Go pieces were likely sourced from the Qilian Mountain region near Dunhuang, a known major source of such minerals.

Overall, this research contributes to the broader understanding of Tang Dynasty material culture and provides a foundation for future studies on historical Go artifacts. The integration of advanced analytical techniques and historical context offers a holistic approach to the study of ancient artifacts, enhancing our appreciation of their cultural and technological significance.