Multi-technique analysis of the mural materials and techniques in the 5th cave of the five temple grottoes in Subei, China

Pigments were sampled using micro-sampling techniques. Dust was removed from the sample surfaces with cotton balls, and pigments of different colors were gently scraped using a scalpel. Samples weighing less than 0.2 mg were qualitatively analyzed using an X-ray diffractometer (XRD) [12], [13], [14], [15]. Given that these are layered murals, samples of red, white, black, and gray pigments from the Northern Zhou Dynasty bottom layer and white, blue, green, and red pigments from the Northern Song Dynasty surface layer were selected, totaling eight pigment samples. The XRD analysis results of the samples are detailed in Table 1, Table 2 and Fig. 4 The red pigment from the Northern Zhou Dynasty bottom layer shows diffraction peaks at 2θ values of 24.11°,33.15°, 35.67°, and 54.17°, corresponding to d-values of 3.7, 2.70, 2.52, and 1.69 matching the standard diffraction peaks of hematite (chemical formula Fe₂O₃). Other phases include quartz, muscovite, chlorite, and feldspar [16]. The white pigment XRD results show that the main phases are muscovite, quartz, calcite, Weddellite, and chlorite, with muscovite (chemical formula (K,Na)Al₂Si₃AlO₁₀(OH)₂) and calcite (chemical formula CaCO₃) as the coloring components. Weddellite might be a product formed under the action of water and biological activity in the collapsed fragments [17]. The gray pigment XRD results show diffraction peaks at 2θ values of 25.44°, 32°, 36.26°, and 49.04°, corresponding to d-values of 3.5, 2.79, 2.48, and 1.85, matching the standard diffraction peaks of Plattnerite (chemical formula PbO₂). Other phases include talc, quartz, feldspar, muscovite, and calcite. Based on the microscopic images of the fragments in Fig. 2, the gray paint layer consists of sporadic black pigment particles distributed over the white pigment, with Plattnerite being a product of the alteration of red lead [18]. The black pigment XRD results show that the coloring component is also Plattnerite. Other phases include quartz, muscovite, feldspar, chlorite, and calcite. Plattnerite is also a product of the alteration of red lead [19].

Fig. 5(a) shows the cross-section of the green paint layer from the surface mural. Under the microscope, the surface appears green with a thickness of approximately 33 μm. SEM-EDS [21]analysis indicates the presence of Cu and Cl elements, and combined with XRD analysis results, the coloring component is identified as atacamite (Cu₂(OH)₃Cl). Below the green paint layer is a white ground layer with a thickness of approximately 57 μm. Fig. 5(b) shows the cross-section of the red paint layer from the surface mural. Under the microscope, the surface appears red [22] with a thickness of approximately 13μm. It mainly contains Ca, S, and Fe elements, and combined with XRD analysis results, the coloring component is identified as hematite (Fe₂O₃), The strong peaks of "Ca" and "S" in the spectrum are due to the selected spectral region also containing gypsum (CaSO₄·2H₂O). Below the red paint layer is a white ground layer with a thickness of approximately 56μm. Fig. 5(c) shows the cross-section of the blue paint layer from the surface mural. Under the microscope, the surface appears blue with a thickness of approximately 54μm. The pigment particles are relatively large and mainly contain Cu elements. Combined with XRD analysis results, the coloring component is identified as azurite. Below the blue paint layer is a white ground layer with a thickness of approximately 132 μm. Fig. 5(d) and 4 (e) show the cross-sections of the black and red paint layers from the bottom mural. Under the microscope, the black paint layer appears on the surface with a thickness of approximately 38 μm, and below it is a red paint layer with a thickness of approximately 63μm. SEM-EDS analysis indicates that the black pigment contains Pb elements and the red pigment contains Fe elements. Combined with XRD analysis results, the black pigment is identified as Plattnerite, and the red pigment is identified as hematite. Fig. 5(f) shows the cross-section of the gray paint layer from the bottom mural. Under the microscope, sporadic black pigment particles are distributed on the surface, with a thickness of approximately 62 μm. SEM-EDS analysis indicates that the gray pigment contains Pb elements, and combined with XRD analysis results, the component is identified as Plattnerite (PbO₂) [23], [24].

Using an analytical balance, at least 1 g of Plaster Layer sample was weighed, gently ground with an agate mortar, the fibrous material was removed, and the sample was sieved through a 1 mm screen before being added to the instrument for analysis [26]. The analysis results are shown in Table 3. As shown in the table, the particle size and distribution of soil samples from the coarse and fine plasters of the bottom and surface murals are different. The average particle size of the surface Plaster Layer’s coarse and fine plasters is similar, around 20 μm, but the sand and silt content distribution differs. Specifically, the fine plaster has 12.49 % less sand content than the coarse plaster and 15.03 % more silt content. A similar situation is observed in the bottom Northern Zhou Dynasty murals, where the fine plaster has 10.13 % less sand content than the coarse plaster, and 12.03 % more silt content, with similar clay content. Overall, the average particle size of the bottom Northern Zhou Dynasty Plaster Layer is smaller, and the clay content in the Plaster Layer is 14.43–9.98 % higher than that of the surface mural Plaster Layer. The particle distribution of different Plaster Layer samples is shown in Fig. 6.The Plaster Layer production techniques for the multi-layered murals of Cave 5 at Five Temple Grottoes are essentially the same as those of the Dunhuang Grottoes. Both Northern Zhou Dynasty and Song Dynasty murals use a coarse plaster with more sand as a base on the cliff, followed by a more uniform fine plaster to level the surface, making it smooth and suitable for painting. The difference lies in that field investigations found that the Northern Zhou Dynasty murals used the same fiber in both the coarse and fine plasters, while the Northern Song Dynasty murals used different fibers [27].

XRD analysis of different plaster samples reveals similar phase compositions, as shown in Table 4. The multilayered murals in Cave 5 of the Five Temples Caves exhibit comparable phase compositions in both Northern Zhou Dynasty and Northern Song Dynasty plasters, including approximately 30 %-40 % quartz, 10 %-15 % calcite, around 20 % feldspar minerals, 5 %-10 % mica, with chlorite and dolomite varying significantly from about 5 %-15 %. This is consistent with the phase compositions of plaster in the Mogao Caves murals [28]. Additionally, a small amount of gypsum was found in the fine plaster of the upper murals, which is attributed to contamination from the white ground layer during sampling.

Fibrous components were found in both surface and bottom mural samples. However, the fibers in the coarse and fine plasters of the bottom murals are similar and are suspected to be straw fibers. In contrast, the fibers in the surface fine plaster and coarse plaster differ in thickness, with the coarse plaster potentially containing straw and the fine plaster potentially containing hemp fibers. We designate the fibers in the surface fine plaster as Sample1, the coarse plaster as Sample 2, and the bottom layer fibers as Sample3 and Sample4.The results of the microscopic infrared spectroscopy analysis are shown in Fig. 7. The figure indicates that Sample1 exhibits an absorption peak at 1730 cm⁻¹ , while no peak is observed at this position for Samples 2, 3, and 4, which is a characteristic peak of hemp fibers [29]. The spectra of Samples 2, 3, and 4 are quite similar, with all showing strong and broad absorption peaks at 3335 cm⁻¹ . These peaks correspond to the stretching vibrations of hydroxyl (OH) groups. Specifically, the vibrations arise from free hydroxyl groups, silanol (Si-OH) groups associated with the straw surface, other chemically bound hydroxyls, and absorbed moisture. At 2900 cm⁻¹ , C-H stretching vibrations are observed, and an absorption peak for water hydroxyls appears near 1621 cm⁻¹ . Additionally, numerous small shoulder peaks are present in the range of 1200–900 cm⁻¹ , which are attributed to C-O-C stretching vibrations. These are characteristic absorption bands of cellulose.

The same method was applied to wheat straw fibers, with results shown in Table 5. It is known that the main component ratios in hemp fibers are: cellulose (60–70 %), lignin (5–10 %), and hemicellulose. In wheat straw, the main component ratios are: cellulose (35–45 %), lignin (20–30 %), and hemicellulose (25–30 %). In Sample 1, R₁ is 0.2975, while the R₁ values for Samples 2, 3, and 4 are around 0.8388. The higher R₁ ratio in these samples indicates that Sample 1 contains more cellulose compared to Samples 2, 3, and 4. Thus, Sample 1 is identified as hemp fiber, and Samples 2, 3, and 4 are identified as wheat straw. From these findings, it was concluded that the fibers in the fine plaster layer of the surface mural samples were flax, while the coarse plaster layer contained straw. For the bottom layer mural samples, both the coarse and fine plaster layers were determined to contain straw fibers.

Additionally, we plotted R₁ on the Y-axis and R₂ on the X-axis to create a scatter plot and compared it with other fibers from the literature. The results, shown in Fig. 8, indicate that the R₁ and R₂ values of fibers in Sample 1 are closer to those of flax. Flax fibers are known as the "queen of plant fibers" and have been cultivated and used in China for over 2000 years [33]. Flax plants are categorized by use into fiber and oil types. Flax fibers are obtained from the bast of the flax plant and are used to produce medium and high-quality linen textiles [34].