New Evidence of Traditional Japanese Dyeing Techniques: A Spectroscopic Investigation

Abstract

The Japanese textile tradition is renowned for its intricate designs achieved through a variety of dyeing techniques, including kasuri, shibori, and paste-resist dyeing. These techniques are often combined within a single textile, resulting in exceptionally elaborate creations. Our paper delves into the technical aspects and complexities of these methods, highlighting the dynamic interplay between tradition and innovation in Japanese textile production. Our scientific endeavour focused on some textiles dating between the 19th and 20th centuries and belonging to the Montgomery Collection of Japanese folk art. Employing non-invasive techniques such as visible reflectance spectroscopy and ER-FTIR spectroscopy, we uncovered key insights into the materials and methods utilized in the creation of these textiles. Our analysis revealed a diverse array of pigments and dyes, including plant-derived, inorganic, and synthetic variants. These findings illuminate the cultural syncretism between traditional Japanese practices and the adoption of new materials from the West, underscoring the dynamic nature of textile production in Japan. Furthermore, ER-FTIR spectroscopy elucidated the predominant use of cotton as the primary fibre in the textiles, aligning with historical records of Japan’s role as a major producer of cotton yarn. Analysis of white areas within the textiles revealed evidence of resist-paste dyeing techniques, particularly tsutsugaki and katazome, through the absence of dye penetration and the characteristic appearance of white lines. Confirmation of indigo dyeing techniques (aizome) was achieved through ER-FTIR spectroscopy, providing reliable identification of indigo and Prussian blue in various shades of blue present in the textiles. Additionally, the detection of Western-derived dyeing method (utsushi-yūzen) and free-hand painting (kaki-e), offers insights into the diversity of dyeing practices employed by Japanese artisans. The presence of proteinaceous materials and synthetic dyes observed in some textiles has implications for conservation practices, suggesting the need for tailored approaches to ensure the preservation of these culturally significant artifacts. Overall, these scientific results shed new light on the materials, techniques, and cultural contexts underlying Japanese textile production, advancing our understanding of this rich artistic heritage and informing future research endeavours in textile science and conservation.

1. Introduction

Textiles have long played a vital role in Japan culture and have never been considered inferior to painting or sculpture [1]. In particular, the survival of textiles from the premodern period is due to donations to temples and shrines of precious robes worn by high-ranking Buddhist monks, aristocrats, and samurai, such as kosode (small-sleeved kimono), kesa (vestment), ban (banner), uchi- shiki (altar cloth), and maku (lintel curtain). The Shōsōin treasure, kept in the Tōdaiji temple at Nara, has long contained brilliant examples of historical textiles since the 8th century [2]. Folk textiles are conserved there as well, giving evidence of the great importance attributed to handicraft objects by Japanese people [3]. The term mottainai (“don’t waste”) is often related to the field of textiles and is additional proof of the high value given to textiles. Folk was used to repurpose rags by turning them into patches, for example. The high residual value of worn-out clothes reflected the craftmanship of the brand-new textiles. The activity of weaving a textile, as well as a basket, has carried a deep cultural importance in Japanese culture. The ability of the weaver is to give an aesthetically pleasant shape to shapeless matter. The weave is also soaked with sacral significance, being a metaphorical representation of the bond with the divinities [4].

Even the dyeing of the yarns was charged with symbolic meaning. It was a common belief that dyes could imbue a fabric with special powers [3,5]. These beliefs were probably related to the medicinal properties of the plants the dyes came from; according to Japan’s indigenous Shinto religion, spirit-gods (tama) were housed in the natural compounds showing a particular power.

The choice of the colour was also charged with added significance derived from the Chinese cosmological system which has been adopted in Japan since the 7th century. According to this system, every colour should be combined with other colours according to a precise five-colour scheme (goshoku), which are a reminder of, respectively, a cardinal point, a natural element, a season, a planet, and a musical note [3,5].

A focus on the most renowned Japanese dye—indigo—is useful to frame the complexity of symbolic meanings attributed to dyes and the Japanese people’s perception of different colours [6]. Indigo (ai) was obtained from the fermented leaves of Japanese persicaria (Polygonum tinctorium Ait.), an annual plant native to Japan. It was also widely used by the lower classes thanks to the feasibility of the dyeing process. Nevertheless, obtaining a dark hue was labour and time intensive, so even in the richest samurai armours indigo was largely adopted. A lot of different hues could be obtained, and each colour had its own identity, with a precise name and significance. Dipping the yarn in the dye bath several times was believed to make the fibre itself stronger, and to keep away snakes because of the strong smell indigo has [3]. Similarly, indigo-dyed textiles were assumed to be useful in case of stomach illnesses, fever, and snake bites.

1.1. Materiality and Manufacturing of Japanese Textiles

Traditionally, commoners’ garments and every-day textiles were obtained from a variety of plants including ramie (Boehmeria nivea L.), hemp (Cannabis sativa L.), wisteria (Wistaria floribunda Willd.), mulberry (Broussonetia kazinoki Sieb.), oak (Broussonetia papyrifera L.), and arrowroot (Puerarua hirsuta Thunb.) [3]. These ‘bast’ or plant-stem fibres (called asa in Japanese language) have been fundamental to Japanese tradition, as all textiles were made from these native plant fibres until the introduction of silk, cotton, wool, and later, synthetics [7]. Silk was considered a precious yarn, initially imported from China, while cotton was introduced in Japan in the 16th century, and quickly became very common as cotton textiles were softer and their production was easier.

Initially, Japanese textiles were shaped by Chinese techniques and styles, but successively, Japanese manufacturing characteristics were established. A wide array of colours was employed, including blue, green, purple, red, and yellow [8]. In particular, dyeing methods such as aizome (indigo dyeing) and benizome (safflower dyeing) were developed [1,9]. On top of indigo and safflower, a huge variety of different plants were used to obtain dyes. Secondary colours were generally obtained by dyeing one colour over another [10]. The dyeing was achieved by dipping the fibres in hot water solutions rich in the colouring principle, previously obtained by the maceration and decoction of petals, seeds, leaves, branches, and roots from a great variety of plants [11,12,13,14]. Except for vat or direct dyeing, the yarn generally underwent a prior mordanting step with tannins or inorganic salts containing aluminium, iron, or copper ions, which made the fibre–dye bond stronger and more durable, even influencing the final hue of the textile [13,14,15]. A complete discussion about dyes and mordants has been done elsewhere [12].

Japan’s textile tradition is known as one of the world’s most dynamic, especially due to the extraordinary complexity of the designs obtained, combining a huge variety of dyeing techniques. The majority of methods were based on “resisting”, that is preventing the diffusion of the dye towards specific regions of the fibres [3,16,17]. It could be done either on the threads prior to weaving, or on the surface of the woven fabric. These two design methods are referred by Japanese people by using the terms saki-zome (dyed before) or ato-zome (dyed after) [3]. A combination of different techniques could be applied, as well as the surface of the finished fabric could be decorated by painting with dyes and inks (kaki-e) [3].

Basically, two main methods can be recognized, according to the way the resisting is obtained [3,18]. By wrapping, knotting, and folding, kasuri and shibori are obtained. The first is done on the threads, while the second is performed on the woven textile. When indigo is used, the result is a white-on-blue motif, characterized by a geometrical design.

Paste-resist dyeing is used to “resist” or prevent the dye from reaching all the cloth, thereby creating a pattern and ground. White areas can be left intentionally white or receive a different colour afterwards. The Japanese tradition encompasses several paste-resist techniques, used alone or in combination, allowing to obtain figurative designs and multiple-coloured textiles [3]. Generally, sticky rice paste (furonori) is used, which prevents the diffusion of the colour in the treated parts of the textile. It can be applied multiple times in different areas to allow the use of different colours. The techniques mainly differ according to the way the design is obtained. In katazome, a paper stencil (katagami) is used to apply a furonori paste onto fabric, and then the fabric is dyed. The areas covered by the paste remain undyed, creating a pattern. Traditional katazome patterns often feature geometric shapes, nature-inspired motifs, or repeating patterns. Yūzen is a decorative dyeing technique often used in the creation of kimono [3,16]. It involves hand-painting or applying rice paste-resist to fabric using a cone-shaped tool. After the design is drawn with a water-soluble dye and left to dry, colours are applied, allowing to obtain intricate and colourful designs, often depicting nature or traditional motifs. Tsutsugaki is a freehand resist-dyeing technique where a rice-paste mixture (furonori) is applied to fabric by a paper cone ending with various shaped metal tips to obtain lines that are flat, round, thick, and thin. The technique derives from yūzen, but the cone tip is generally larger, and less refined patterns are obtained. Tsutsugaki creates bold, expressive lines and designs. After the application of dyes, steam is used to fix the colours. After dyeing, the paste is washed away, and the design emerges with white outlines. Another variation of yūzen is the katayūzen method, where the dye is directly applied to the textile using a brush through a paper stencil (katagami). This strategy allowed one to significantly reduce the time and labour associated with the manufacturing of a decorated textile, thus lowering the costs [3,16].

Many textiles are multi-coloured with an indigo ground colour. To obtain such results, the secondary colours are applied first, fixed, and then covered with a thick layer of paste. The fabric is dipped several times in the indigo dye bath and let in the air every time to oxidase the molecule. A final hot water bath removes the rice paste residue [3].

Kaki-e is not strictly a dyeing technique, but it is frequently encountered for highly decorated textiles, such as hanging scrolls (kakemono) [19], hand scrolls (emakimono) [20], sliding doors (fusuma), or folding screens (byōbu) [21]. Kaki-e is a free-hand painting technique, either monochrome or polychrome, executed on the fabric using brushes [22]. Monochrome pieces typically use sumi-e, made from soot mixed with glue derived from fishbone or animal hide. Polychrome pieces use pigments derived from natural ingredients, such as minerals, shells, corals, and semi-precious stones like malachite, azurite, and cinnabar [12]. It is probable that lake pigments, which are organic dyes extracted from plants which are turned into a powder, were once used as well. A hide glue solution, called nikawa, serves as a binder for these powdered pigments. The same painting materials were identified in a number of works dealing with ukiyo-e prints analysis [23,24,25,26,27,28].

Paintings on paper and textiles are fragile artworks, whose integrity has to be preserved using non-invasive techniques of analysis. When possible, a multi-analytical approach is applied, including different techniques such as X-ray Fluorescence, reflection Fourier Transform Infrared (FTIR) spectroscopy, Raman, Ultraviolet-Visible (UV-Vis) reflectance spectroscopy, and hyperspectral imaging (HSI) [20,23,25,27,29,30,31,32]. In particular, FTIR spectroscopy is a promising method for fibre recognition alternative to microscopic observation [33,34]. External Reflection (ER) FTIR spectroscopy has been successfully used to non-invasively analyse a wide variety of artistic materials and different supports, such as illuminated manuscripts [30,35], easel paintings [36,37], textiles [38,39], and photographs [12]. Thus, ER-FTIR spectroscopy is adept at providing information about the chemical composition of fibres, binders, and pigments.

It is obviously advantageous that these techniques can be applied non-invasively, as sampling is not always feasible. However, due to the lack of separation of the colorants, the results are not always conclusive. In some cases, only High-Performance Liquid Chromatography-Mass Spectrometry (HPLC-MS) is able to reveal the exact composition of the dye. This ability is particularly important when identifying synthetic dyes, due to the high variety of similar compounds within a class [23,40]. Nevertheless, identifying the dye class rather than the specific dye molecule is sometimes sufficient, for example, when determining whether synthetic or natural dyes are present and understanding how objects were created and treated [32,41]. For instance, UV-Vis reflectance spectroscopy is a widely used technique that can provide preliminary, if not conclusive, information about the identity of pigments and dyes [24,42,43,44,45,46]. Specific literature demonstrated that the technique is trustworthy as far as (a) pure colours are considered or (b) a large set of reference spectral data is collected from several mock-up samples dyed with known raw materials. The main limitations to the direct identification of dyestuffs arise from the concentration of dyes on the fibres and the mixtures of different colorants. When synthetic dyes are investigated, the exact identification cannot generally be obtained, due to the high variability within a class [32,41]; yet, identifying the dye class could be sufficient to confirm the proposed dating of the textiles.

1.2. Aim of the Study

This study aimed to comprehensively investigate the dyeing techniques and materials used in Japanese folk textiles, with a focus on providing scientific insights into their composition and production methods. The research utilized a significant collection of historical textiles, with the constraints of on-site and non-invasive analyses. By applying non-invasive spectroscopic techniques, such as visible reflectance spectroscopy and ER-FTIR spectroscopy, we aimed to identify the diverse array of pigments and dyes utilized in Japanese folk textiles, as well as the primary fibre materials employed in their construction. Particular attention was given to the technical aspects and materiality of the dyeing techniques. By analysing the application of these methods within individual textiles, we aimed to bridge the gap between traditional textile scholarship and modern scientific analysis, providing a holistic understanding of Japanese textile heritage.

2. Materials and Methods

2.1. Reference Materials

A custom-made database of Japanese dyes and pigments applied to paper was built and analysed by visible reflectance spectroscopy as a reference. In total, 31 traditional materials, 11 synthetic materials, and 25 mixtures of two dyes and/or pigments were chosen and tested based on a previous thorough bibliographical search [12]. Dyes were bought as powdered or raw materials from Pigment Tokyo (Japan) and extracted according to a recipe tested in a previous work [12]. Extracted dyes and pigments were subsequently mixed with a 3% w/w solution of animal glue and alum (KAl(SO₄)₂) and applied with different concentrations on white drawing paper (Fabriano, weight 180 g/m²). As for the mordant, only alum was considered as it is definitely the most cited one in the specific literature. For each dye, pigment, or mixture, 3 to 5 different concentrations were considered to account for concentration-induced variability.

2.2. The Montgomery Collection of Japanese Folk Art

The seven objects examined in this study belong to the Montgomery Collection of Japanese folk art, collected over more than fifty years by collector Jeffrey Montgomery. It stands as one of the most significant private collections of Japanese folk art, encompassing textiles, furniture, ceramics, sculptures, and paintings [3,47]. Displayed in numerous exhibitions worldwide, the Collection was hosted at the MUSEC in 2021–2022 and analysed during that period.

Various types of textiles are represented in the Collection, dating predominantly from the late Edo to the Taisho era (from the second half of the 19th century to the 1920s). A common textile type is the futonji, a decorated cover traditionally included in a bride’s trousseau. Part of the traditional Japanese bedding set, known as a futon, it typically comprises a padded mattress and a quilted bedcover. A futonji could be repurposed as a smaller wrapping cloth (furoshiki), usually square, utilized for wrapping and/or transporting goods. Traditional materials include silk and cotton, with attention given to the aesthetics of furoshiki and how it is folded to enhance the design pattern.

Various types of kimono, the iconic traditional Japanese garment, are also present. Hitatare was the most common and iconic style of clothing worn by the samurai. A hanten was a short winter coat, which started to be worn, especially by the common people, in the 18th century, during Japan’s Edo period (1603–1867).

Specialized fabrics were used for furnishing purposes, such as making byōbu (literally ‘wind wall’), folding screens composed of several panels adorned with decorative painting and calligraphy. Byōbu were employed to partition interiors and enclose private spaces, among other uses.

The analysed objects of the Montgomery Collection are depicted in Figure 1, accompanied by brief descriptions indicating object number, type, dimensions, and dating for each object. The analysed areas were visually selected to encompass the full range of colors, and analysis of the white substrate was also conducted.

Analyses were conducted using both techniques in the same area whenever possible. However, there were instances where collecting ER-FTIR spectra was not feasible due to limitations in the maximum height accessible to the instrument. Conversely, certain areas were not analysed by reflectance spectroscopy due to its non-contactless nature, and the condition of the area discouraged surface contact.

2.3. Visible Reflectance Spectroscopy

Visible reflectance spectroscopy of coloured areas was carried out using a portable Lovibond^® SP60 spectrophotometer (Lovibond, Dortmund, Germany), equipped with an integrating sphere. The standard illuminant used was standard daylight, D65, using a CIE 1964 10° standard observer. The calibration of the instrument was performed on the reference, which consisted of a disk of white ceramic for white and a trap (dispersion cone) for black. The wavelength range of measure was from 400 to 700 nm with 10 nm acquisition step, d/8°, SPIN (Specular-Included) and a circular area of 8 mm diameter was analysed (8 mm viewing/12 mm illumination). The spectra were compared with a custom-made database of Japanese dyes and pigments, including 31 traditional materials, 11 synthetic materials, and 25 mixtures of two dyes and/or pigments. For each dye, pigment, or mixture, 3 to 5 different concentrations were considered.

2.4. External Reflection Fourier Transform Infrared Spectroscopy

ER-FTIR measurements were conducted on site using an Alpha Bruker FTIR portable spectrophotometer (Bruker, Billerica, MA, USA) configured in external reflection mode, allowing for contactless measurements, equipped with a DTGS detector. Spectra were collected within the range of 7500 to 375 cm⁻¹, utilizing a resolution of 4 cm⁻¹ and averaging 200 scans. Periodic background measurements were obtained using a flat gold mirror. The instrument was positioned in front of the target area, with the optimal distance determined by focusing via the built-in camera. Subsequent fine adjustments were made via software to maximize signal intensity directly in the interferogram. On average, the working distance from the surface ranged from 1 to 1.5 cm. Spectral data were processed using pseudo-absorbance (log (1/R); where R = reflectance) as the intensity unit. The spectra of the samples were interpreted by comparison with a custom-made reference database and with the literature.

2.5. Data Treatment and Elaboration

Spectragryph optical spectroscopy software, Version 1.2.15, was utilized for visualizing and manipulating visible reflectance and ER-FTIR spectra [48]. Standard Normal Variate (SNV) processing was employed on either the entire spectrum or specific segments to correct the baseline and normalize the signal. This approach involves subtracting the mean spectrum from each individual spectrum, thereby removing information about absolute intensity while allowing for the discernment of subtle differences in the band shape [34]. Visible reflectance spectra were also subjected to second derivative when needed to facilitate the comparison of similar features. The simulation of pigment mixtures [23,49] though the linear combination of two spectra was calculated using the “Add” function in Spectragryph software, after being normalized. The y values of the two spectra were added to obtain the final spectrum. Origin Pro 2018 software (OriginLab Corporation, Northampton, MA, USA) was used for data plotting.

3. Results

3.1. Visible Reflectance Spectroscopy

Visible reflectance spectroscopy was used to study the coloured areas. Figure 2, Figure 3, Figure 4 and Figure 5 show the spectra of different parts of the samples, grouped by colour, along with the spectra of the references.

Red area spectra exhibit an inflection between 570 and 610 nm (Figure 2). Two areas of different textiles (59_2, 69_2) show a good correlation with the reference sample of Ponceau red 2R (acid red 26), so the use of this synthetic compound or a similar one belonging to the same family (aryl azo dyes) can be proposed. The inflection point was found at 590 nm. The spectrum of the other red area 59_1 can be attributed to the pigment red ochre. The spectral curve of the sample matches well with the reference spectrum; the pigment can also be easily recognized in the second derivative graph, which shows a maximum at 560 nm and a minimum at 590 nm.

Yellow areas give evidence of the use of different dyes, although their identification is considered quite problematic due the similarity of their spectra (Figure 3). In area 59_8, the presence of yamahaji can be proposed, as the spectrum of the sample clearly matches the reference spectrum. Yamahaji is a traditional Japanese dye based on the flavanonol fustin, extracted from the wood of Japanese sumac (Toxicodendron vernicifluum) [12,50]. In sample 69_5, the yellow parts were probably achieved using a synthetic dye from family of arylide, or Hansa, yellow [51]. The second derivative graphs obtained from the spectra of the sample and the reference (Hansa yellow PY3) both show a maximum at 480 nm and a minimum at 510 nm. The slight differences in the reflectance spectrum are attributed to the varying concentrations of the dye on the support [42], as proved by Figure S1.

The study of orange–red parts in objects 59 and 72 encountered some challenges. The orange area 59_7 gave a negative result when compared with realgar, one of the few known orange pigments. It was hypothesized that a mixture of yellow and red dyes might have been used, as orange dyes are uncommon. A mixture of Ponceau Red 2R and yamahaji was tested since both were identified as pure dyes in other parts of the same textile. Instead of preparing a mock-up sample of the resulting dye, a linear combination of the two spectra was calculated, as a linear mixing model has proven to be quite successful for the unmixing of pigment mixtures [23,49]. In Figure 4, the spectra of the two pure dyes are shown alongside their linear combination. The calculated spectrum closely matches the sample spectrum.

Sample 72_4 is a red area with some thin white lines across the dyed surface. It was not possible to exclude the white parts from the acquired area. Its spectrum imperfectly matches the reference spectrum for madder, as evidenced in the second derivative graph.

Figure 5 shows the spectra of purple and blue areas. The spectrum of the purple area 69_3 matches quite well with Rhodamine B (basic violet 10), as evidenced in the second-derivative graph (maximum at 570 nm and minima at 610 and 440 nm). Discrepancies can be explained based on the hypsochromic shift, as it happened with Hansa yellow [42]. The spectra of different concentrations of Rhodamine B are shown in Figure S2 as a proof. The use of this synthetic compound or a similar one belonging to same family (triarylmethane dyes) can be proposed.

The three spectra of the blue areas taken from object 65 demonstrate the variability associated with blue hues. Only the spectrum 65_2 closely matches the reference spectrum for indigo. The spectrum of 65_3 resembles that of Prussian blue, but its maximum is red-shifted relative to the reference, preventing its identification. Similarly, the spectrum of area 65_1 is similar to Prussian blue spectrum, but it differs at higher wavelengths.

3.2. ER-FTIR Spectroscopy

ER-FTIR spectroscopy was performed on both coloured and non-coloured areas of the same textile. At least one non-coloured part from each textile was analysed to characterize the bulk material. All the spectra are similar; the plots of samples 59_3 and 69_8 are shown in Figure 6, alongside the spectrum of a cotton reference. The reference spectra of hemp and viscose were evaluated as well but were excluded from the discussion as the sample spectra did not exhibit any of the peaks attributed to lignin or modified cellulose [34]. The spectral features of the cotton reference spectra precisely coincide with those of the analysed areas. By evaluating only peaks that do not show significant shifts between ATR and external reflection modes, in accordance with a previous paper [34] we assigned peaks at 1043 and 1021 cm⁻¹ to C-O stretching located in secondary and primary alcohols, respectively. Vibration at 895 cm⁻¹ is due to C-O-C symmetric stretching in plane. The C-C ring breathing band at 1155 cm⁻¹, which is characteristic of cellulose in ATR mode, is inverted at 1166 cm⁻¹ in reflection mode. The strong peak at 1105 cm⁻¹, attributed in ATR mode to C-O-C glycosidic bond, is inverted in the reflection mode. C-H bending at 1430, 1365, 1315 cm⁻¹, and C-H rocking at 1000 and 985 cm⁻¹ do not show spectral changes. We report the same for CH₂ wagging at 1335 cm⁻¹ and CH₂ twisting at 1280 cm⁻¹. A peak at 1205 cm⁻¹ is assigned to C-OH and C-CH bending. The bending vibration of OH group appears as a broad band centred at 1640 cm⁻¹ and at 3300–3100 cm⁻¹. The shoulder at 1730 cm⁻¹ (marked with an asterisk) can be related to the ageing of the fibre, as previously reported [34].

The spectra of the coloured parts closely resemble that of cotton, although a considerable number of spectra show major differences compared to the bulk material (Figure 6). The peaks appearing in the region of amides I and II (1760–1500 cm⁻¹), as well as the peaks at 1163, 1080, and 1030 cm⁻¹, indicate the presence of a proteinaceous material [38]. These signals match a reference sample of animal glue [52]. The spectral features of animal glue are evident in the spectrum of sample 59_1, giving evidence of a considerable amount of glue. The broad OH stretching band, which in cotton is dominated by the peak of the intramolecular hydrogen bonding [34], strictly resembles the same band as it appears in animal glue, centred at 3320 cm⁻¹ and attributed to amide A [38]. The same peaks exhibit a lesser intensity in sample 69_1, resulting in a spectrum that shows features of both the features of glue and cotton.

Upon closer inspection, some spectra reveal minor peaks associated with specific pigments and dyes. Indigo and Prussian blue are easily identifiable using ER-FTIR spectroscopy, as illustrated in Figure 7. In object 63, three distinct shades of blue can be differentiated. The typical features of indigo dye, at 1634 and 1614 cm⁻¹ [53], are prominent in the area 63_3, which appears very dark, and are present with reduced intensity in sample 63_2. Other characteristic marker bands for indigo are detected at 1462 and 1482 cm⁻¹, but they coincide with absorptions due to the cellulosic substrate. Although sample 63_1 appears dark, it does not exhibit any of the diagnostic bands of indigo, suggesting that a different type of dyeing was used in this area.

In sample 65, the diagnostic features of indigo are evident in the dark background (sample 65_1). These features are less pronounced in sample 65_2 due to the less intense hue (not shown). For samples 65_3 and 65_5, a different shade of blue and different shades of green were employed. In both cases, the spectral features of Prussian blue, with its distinctive peak at 2098 cm⁻¹, appear in the spectra [12,36].

In Figure 8, additional materials related to the use of specific pigments are illustrated. The spectrum of sample 59_1 displays peaks indicative of illite at 3620, 1026, 825 cm⁻¹ and of quartz at 1161, 1145, 792 cm⁻¹ [54,55]. Illite, a type of clay, can exhibit various colours depending on the type of iron oxide it contains. Quartz is also commonly associated with these materials. Given that the analysed area is red, it is reasonable to assume the presence of iron (III) oxide [56]. The peak at 1098 cm⁻¹ is indeed characteristic of red ochre [55]. For the sample 74_3, peaks corresponding to kaolin are identifiable at 3697, 3622, 1114, 1092, 1020, 912, 796 cm⁻¹ [54,55]. Pure kaolin is white, becoming yellowish due to the presence of limonite [55]. This possibility aligns with the off-yellow colour of the analysed area. The peaks marked with asterisks (3320, 1666,1554, 870 cm⁻¹) indicate the use of animal glue as a binder.

Table 1. Summary of the main binders, pigments and dyes identified. NA means that this point of analysis was not analysed using both the techniques. Tentative identifications are marked with an asterisk (*).

Textile	Colours	Binder Identified by ER-FTIR Spectroscopy	Dyes and Pigments Identified by ER-FTIR Spectroscopy	Dyes and Pigments Identified by Reflectance Spectroscopy
kimono—59	maroon	animal glue	red ochre	red ochre
	red	animal glue		Ponceau red 2R *
	black	animal glue	Prussian blue, indigo
	orange			yamahaji, Ponceau red 2R *
	yellow	NA	NA	yamahaji
kimono—62	black
futonji—63	light black
	light blue		indigo
	blue (ground)		indigo
futonji—65	blue (ground)		indigo
	pale blue		indigo
	light blue	animal glue	Prussian blue
	red	animal glue		NA
	green	animal glue	Prussian blue	NA
futonji—69	orange	animal glue
	red	animal glue		Ponceau red 2R *
	purple	animal glue		rhodamine B *
	blue (ground)		indigo
	yellow	animal glue		hansa yellow *
	green
	grey	animal glue		NA
futonji—72	blue (ground)		indigo
	red	animal glue		akane *
	green
byobu—74	green		indigo
	yellow	animal glue	yellow ochre	yellow ochre
	red	NA	NA	yamahaji, Ponceau red 2R *

summarizes the main findings obtained from the analysis of the Montgomery Collection. An overview of the identified pigments or dyes is provided in Table S1 along with their spectral features.

4. Discussion

The results obtained from the current study enrich our knowledge about the materiality of Japanese textiles. As shown by Table 1, the majority of the materials investigated were successfully identified, either as a specific component or at least within their general class. However, microdestructive techniques would be necessary for the precise identification of each compound.

Cotton was identified as the support in all the analysed areas. Cotton occurrence is not obvious considering that the major yarn used for the manufacturing of kimono was silk. Due to its cost and to the prohibition for commoners to use this precious fibre until the Meiji era (1868–1912), folks generally used bast fibres for weaving, such as hemp and ramie, which were known collectively as asa. Cotton plant started to be cultivated in Japan in the 16th century. For a long time, it was considered a luxurious and expensive material. The widespread occurrence of cotton in the Montgomery Collection is in line with the proposed dating of the artworks to the end of 19th century. It is known that due to the industrialization of Japan coinciding with the beginning of Meiji era, Japan became one of the world’s largest producers of cotton yarn and cloth [57].

Indigo dyeing—aizome—is a well-known Japanese technique, so it was not surprising to detect indigo in the majority of the blue areas of the Collection. ER-FTIR spectroscopy showed great performance in detecting indigotin, especially in the dark blue areas. Indigotin is equally detected when mixed with other dye to create green hues. Conversely, visible reflectance spectra are strongly influenced by the hue of the blue, preventing reliable identification. For example, ER-FTIR results showed that areas 65_1 and 65_2 are dyed with indigo, but the reflectance spectrum of area 65_1 does not match with the reference spectrum. It appears as a very dark blue. The very high concentration of indigo, which makes its detection very feasible by ER-FTIR spectroscopy, prevents identification by reflectance spectroscopy. This concentration-dependant issue is a renowned limitation of reflectance spectroscopy [42].

Another issue with reflectance spectroscopy is the identification of Prussian blue. For example, Prussian blue was easily identified by ER-FTIR spectroscopy in area 65_3, but the reflectance spectrum (Figure 5) does not agree with the FTIR attribution due to a significant shift in the absorption maximum of the sample spectrum compared to the reference. This shift is probably due to the colour coverage. The literature reports that mixing malachite green with a white pigment, such as lead white, induces a red shift effect in the absorption maximum [58]. A similar effect probably occurs in blue-dyed textiles, as the off-white cotton support alters the appearance of the blue colour [28]. This issue likely affects the spectrum of sample 72_4 as well, which imperfectly matches the reference spectrum for madder due to the presence of some white areas in the analysed part.

The detection of indigo provides insight into the skilful use of aizome techniques to achieve various hues of blue in different areas of the textile. For instance, objects 63 and 65 display indigo-dyed regions with differing shades of blue (Figure 7). This effect was likely achieved by dipping the textile in the dye bath multiple times, protecting lighter areas with furonori [3]. Indigo was also found in a green area of object 74, suggesting this hue was achieved by dipping the textile in different dye baths (Figure 8).

Another traditional technique, tsutsugaki, was identified indirectly. Objects 63 and 65 feature uneven white lines, indicative of the use of the tsutsugaki tube. Although no traces of furonori paste were found on the textile surface, the absence of any white paint in these areas confirms that they were protected from absorbing indigo rather than being coloured. The width of the white lines and the complex pattern further suggest the use of this technique.

Detecting indigo also facilitates the recognition of other colorants that could be mistaken for indigo. In objects 62 and 63, a black dye unrelated to indigo was identified, suggesting the use of a technique involving tannin and iron salts, typical of the Asiatic area [59]. Prussian blue was found in objects 59 and 65, an unexpected material for textiles. The presence of animal glue as a binder clearly appears only in the analysed areas in object 59. While Prussian blue has been used in Japan since the early 19th century, it was primarily known for colouring woodblock prints (ukiyo-e) [28]. To the best of our knowledge, there are no references about its use for textile dyeing in Japan. In the West, it became a significant artists’ pigment in the 18th century and a low-cost alternative to indigo for textiles in the first half of the 19th century, once a method to fix it directly onto fibres was discovered [60]. This result was achieved either by solubilizing the pigment in an alkaline solution before application or using binders such as gum, oil, or albumin.

The detection of proteinaceous material applied to most of the coloured areas, including those with Prussian blue, explains the use of pigments and dyes that lack affinity for cellulosic fibres. While the presence of animal glue (gelatin) has been suggested, other proteinaceous materials cannot be ruled out, especially when the spectral features of glue are not very intense. Variability in the types of proteinaceous material is evident in the spectra shown in Figure 6. It is well known from the literature that Japanese paintings on paper and silk were made using watercolours obtained by mixing animal glue (nikawa) with powdered pigments or precipitated dyes (lakes pigments) [21]. Gelatin could act as a binder, but also as a mordant. Commonly, the lack of affinity of cotton for the natural dye is improved by the mordanting technique that consists of treating the fabric with metal salts. These mordants form metallic complexes with dye and cotton improving dyeing. Because protein fibres like wool and silk were known to have good dyeability, efforts were made, as early as the half of the 19th century, to animalize cotton by depositing protein material onto it to make it more dyeable. In a dyeing and tissue-printing handbook dating to 1882 [61], albumen, animal gelatin, and wheat or rice proteins are listed as common proteinaceous materials for treating cotton before applying the dye, or to be mixed with dye and printed onto the dye. The dye would then be fixed by steaming or other methods. Protein mordanting was used, for instance, for dyeing cotton with early synthetic dyes, as they could not directly bond to cellulosic fibres [62]. Fukatsu-Fukuoka [63] reports that since 1879, a modified technique called utsushi-yūzen began to replace the traditional yūzen-zome. In this new technique, synthetic dyes or “aniline” were mixed with starch paste, traditionally used only for resisting dyes on fabric. It is likely that the rice protein in the paste facilitated the dye fixation to the cotton textile [64]. The dyed paste could be applied to the fabric through stencils, a method traditionally used for paste-resist dyeing of cloth (kata-yūzen). When the dyed fabric was steamed, the dye penetrated the fabric while the paste remained on the surface, simultaneously dyeing and resisting other dyes.

This process was similar to calico printing, although albumen was preferred over wheat proteins [65]. The detection of synthetic dyes in object 69, along with signals indicating a proteinaceous material, supports the use of the utsushi-yūzen technique. The precise history of synthetic dye introduction helps date objects 69 and 59, where such dyes were found. Ponceau red 2R was found in a Japanese print dated 1889, a few years after its discovery by H. Baum in 1878 [25]. Rhodamine B was discovered in 1887 [32] and Hansa yellow was first synthetized in 1909 [51]. HPLC-MS would be the only method capable of precisely identifying synthetic dyes, given the high variety of similar compounds within a dye class. However, the identification of the dye class through reflectance spectroscopy is sufficient to confirm the proposed dating of the textiles. The detection of synthetic dyes used on textiles aligns with the use of these new “aniline” dyes, as attested in a number of previous studies on ukiyo-e prints and hand-painted photographs from the same period [12,23,25].

In objects 59, 65, and 74, the extensive use of opaque pigments (such as red and yellow ochre and Prussian blue) mixed with animal glue suggests a deliberate attempt to achieve a pictorial effect. The detection of glue-based paint aligns with the typical materials used for colouring paper screens (byōbu) [21], woodblock prints (ukiyo-e) [66], and hand scrolls (emakimono) [20]. Free-hand painting on textiles is known as kaki-e [22] and can be combined with other dyeing techniques for obtaining the desired decoration.

Object 59 exemplifies this dual identity of the textiles, midway between a cloth and a painting. This object has been found to contain a traditional dye (yamahaji), a synthetic dye (a Ponceau red), and several inorganic pigments. The absence of proteinaceous materials associated with yamahaji suggests that a traditional dyeing technique was used for the yellow parts, aligning with the pale hue typical of natural dyes, as opposed to the brighter and opaquer synthetic and inorganic ones. An aqueous solution of Japanese sumac wood (Toxicodendron vernicifluum), the source of yamahaji, was likely applied to the textile. For the red parts, a mixture of Ponceau red and proteinaceous material—hypothetically glue or rice paste—was used. The orange part is particularly interesting, as it has been probably obtained by mixing a traditional dye (yamahaji) with a synthetic one (a Ponceau red). This finding, along with the detection of the utsushi-zome method, attests to a high level of syncretism between Western and Japanese material culture. For the black part, a mixture of glue, Prussian blue, indigo, and probably carbon black (sumi) was found; for the maroon part, red ochre was mixed with sumi and glue. The design reminds one of the tsutsugaki technique, although the use of stencils (kata-zome) to speed up the dyeing process cannot be ruled out. It is probable that once applied onto the textile all the colours were fixed by steaming.

The dyeing techniques used for these textiles significantly impact their conservation practices. Loosely bound dyes as well as proteinaceous materials, such as animal glue, are vulnerable to damage from wet cleaning [67,68]. Additionally, these materials are susceptible to biological attack, as glue can become a culture medium for microorganisms following hydrolysis, serving as a nutrient source [69].

5. Conclusions

This paper presents the results of a groundbreaking study on the dyeing techniques used in the manufacturing of Japanese folk textiles. Seven textiles from the Montgomery Collection of Japanese folk art were analysed using non-invasive techniques to detect traces of dyes and other materials related to their production.

The visible reflectance spectra of several coloured areas of the textiles were compared with spectra from a custom-made database of Japanese dyes and pigments applied to paper. The comparison enabled the identification of numerous dyes and pigments, including plant-derived, inorganic, and synthetic colorants, either as specific component or at least within their general class. Their use, even within the same textile and in mixed forms, indicates a high level of cultural syncretism between traditional Japanese techniques and new materials introduced from the West.

ER-FTIR spectroscopy confirmed that cotton was the primary yarn used in these textiles. This finding is consistent with Japan’s status as one of the world’s largest producers of cotton yarn at the beginning of the 20th century. The analysis of white areas in the textiles revealed that these parts were not dyed. This evidence, along with the distinctive shape of the white lines, supports the use of resist-paste dyeing techniques based on furonori paste, specifically tsutsugaki and kata-zome.

Indigo dyeing (aizome), which was an educated guess before the analyses, was confirmed for the majority of the blue areas. Compared to reflectance spectroscopy, ER-FTIR spectroscopy provided more reliable detection of indigo, even in light and dark blue areas that are challenging to identify with reflectance spectroscopy. ER-FTIR spectra also revealed instances where blue areas were not dyed with indigo but obtained with Prussian blue, an uncommon finding for Japanese textiles.

This study found that many coloured areas contained proteinaceous material, which was used with both pigments and dyes. In some instances, the textiles were used as canvases with opaque pigments mixed with animal glue, employing a pictorial technique known as kaki-e. This study provides the first scientific confirmation of kaki-e on cotton. In other cases, the presence of proteinaceous material, which could be either animal glue or rice proteins, could be related to the application of a new technique called utsushi-yūzen. This technique involves using a paste of furonori and aniline dye mixed together, another evidence of cultural syncretism. The detection of proteinaceous materials and synthetic dyes has implications for conservation practices, as these textiles may be more sensitive to water and biological attacks.