More on the data

Visual inspection of each print was carried out, as well as examination under ultraviolet illumination and several non-invasive analytical techniques were used to identify the colorants. These included: X-ray fluorescence spectroscopy (XRF), excitation emission matrix fluorescence spectrophotometry (EEM), fiber optics reflectance spectroscopy (FORS), and in some instances reflectance imaging spectroscopy (RIS). To serve as references, colorant samples were formulated and printed using organic and inorganic materials traditionally believed to have been used during the Edo period (1603–1868). The non-invasive techniques used to compile this database permit hypothetical identifications of the colorants in different areas of a given print, but strictly speaking cannot provide certain identifications, which would require analysis of samples. This database is meant to be a beginning point for continuing targeted research on the specific types of colorants found in Edo period Japanese prints.

Analytical Methods

This section provides an overview of the individual analytical techniques, followed by an “In More Detail” section with additional information on interpretation of the data and potential limitations in the accuracies of the identifications shown in the tables in the database.

----------------------------------------What is XRF?-------------------------------

XRF analysis using an open architecture Bruker ARTAX

XRF spectrum of yellow area exhibiting major Arsenic peak

X-ray fluorescence (XRF) spectroscopy is a nondestructive analytical method used to determine the elemental composition of materials. When excited by X-rays, each element emits secondary or fluorescent X-rays that have energies unique to the element. The XRF instrument used in this research, a Bruker Artax micro-XRF spectrometer, cannot detect elements lighter than silicon, thus organic colorants (which mainly consist of carbon, oxygen, and hydrogen) cannot be identified.

For this database, XRF was used to tentatively identify inorganic colorants and metals based on the element(s) detected. For example, if XRF detects the elements arsenic and sulfur in a yellow area, the pigment is most likely to be orpiment (arsenic trisulfide), the most common natural yellow colorant that contains arsenic.

Expand

XRF in more detail:

While chemical elements can suggest specific pigments, from a scientific point of view other techniques not used in this study are required to identify pigments with certainty (such as Raman spectroscopy or X-ray diffraction). In the case of yellows that contain arsenic, orpiment is the most likely pigment, but there are two other possibilities. In the later part of the Edo Period, a synthetic pigment of the same chemical composition was manufactured, and this artificial pigment cannot be distinguished from the natural mineral by XRF alone. There is also another yellow-colored arsenic sulfide, pararealgar, but to date this compound has not been specifically identified in Japanese prints. Thus identifications of orpiment, based on detection of arsenic, while probably correct, should be considered tentative.

Other examples:

• Detection of iron in a red or red-brown area likely indicates the presence of red earth or ochre;
• Detection of copper and zinc in a metallic-looking area likely indicates the use of ground brass metal
• Detection of lead in an orange-red area probably indicates red lead, while detection of lead in a white area probably indicates lead white. }}

----------------------------------------What is EEM?------------------------------

EEM analysis using an Agilent Cary Eclipse Spectrometer with a fiber optic probe

Excitation and emission maxima of the three red colorants

3D EEM plots of reds

Some organic colorants used in Japanese woodblock prints can be characterize based on their ability to fluoresce when exposed to ultraviolet (and even visible) illumination. Excitation emission matrix (EEM) fluorescence spectrophotometry measures emission (fluorescence) of a material (in the visible range) as a function of the wavelength to which the material is exposed (in the ultraviolet as well as visible range). The typical display of results is a color contour map with excitation wavelength on the y-axis and emission wavelength on the x-axis. These maps can allow for (at least tentative) identification of some materials if the contour maps are visibly different for each of the potential materials. Even the lack of fluorescence, which indicates the material is non-fluorescent, may provide a key piece of information for identification of certain colorants. For this research, an Agilent Cary Eclipse Spectrometer with a fiber optic probe was used.

The excitation and emission maxima of the red colorants, safflower, madder, and sappanwood are similar but not identical (see left). The emission maxima for safflower is in the yellow-orange region, madder is in the orange region, and sappanwood is in the orange-red region. (Viewing these colorants side by side under ultraviolet illumination would show the subtle differences in tone.) The relative intensity of the absorption or emission is indicated by the color with dark red representing highest intensity. The oblique bar on the plot is where the excitation and emission wavelengths are identical, thus the detector is saturated at these points.

Some of the yellows in ukiyo-e prints also fluoresce. For example, the strongly-fluorescing turmeric is easily identifiable by EEM. However, other organic yellow colorants such as Japanese pagoda tree (enju) and gardenia (kuchinashi), which belong to a general chemical class known as flavonoids, exhibit relatively little fluorescence, and while their presence can be hypothesized by EEM, they are impossible to differentiate from one another.

Expand

EEM in more detail:

As noted above, specific red or yellow dyestuffs may display characteristic EEM graphs that can be used to compare with graphs from prints to identify different dyestuffs. However, the fluorescence of specific materials can be affected by concentration of the dye, mixture with other dyestuffs or pigments, or unknown differences in the preparation of the dye for use as an ink on the print and the preparation of the reference dyes. Thus not every example of a specific dyestuff will produce exactly the same EEM contour map. In addition, while analyses of reference samples prepared in the lab suggest that the major reds – safflower, madder, and sappanwood, should be able to be distinguished, this is not always the case. As an aid in characterization, excitation and emission spectra can be extracted from the EEM data file. These better show the excitation and emission of a material than the 3D contour maps. Analyses sometimes show spectra that do not very well match reference spectra from any of the three known organic reds that could have been used in the prints. In some instances, reds bleed through the paper during printing, and EEM data from front and back of a red area can be acquired. When these differ from one another (usually in a subtle fashion), it seems possible that a mixture of two main reds was probably used, one of which bleeds through the paper more easily than the other.

One example illustrates the complexities, and potential uncertainties, in interpretations of fluorescence data from organic reds in Japanese prints.

----------------------------------------What is FORS?----------------------------

Ocean Optics FORS components for reflective analysis

FORS spectra for indigo and dayflower

Fiber Optics Reflectance spectroscopy (FORS) measures the reflectance of a point on an object across the visible and near infrared region (about 400-1000 nm). FORS uses a fiber optic probe and a high intensity halogen light source to examine very small spots, minimizing the interference from ambient light. For this research, an Ocean Optics miniature spectrometer was used. Reflectance spectra can distinguish between some colorants. For example, the three common blue colorants used in Japanese woodblock prints (dayflower, indigo, and Prussian blue) can be easily identified by FORS. Greens were usually made from a combination of a blue and a yellow colorant, while purples were usually made from a combination of a blue and a red colorant. The blue colorants in these mixtures can usually be determined by FORS analysis. FORS is not usually useful for identifying yellow colorants, but can sometimes be useful for the organic reds.

Expand

FORS in more detail:

Reflectance spectra for the three common blues are show below, along with spectra of each blue from a print. The paper used in the prints typically has a slight yellowish tint, thus for the FORS analysis, an area of bare paper is used to acquire a background spectrum, which is then subtracted from the spectrum of the colored area. In principal, this gives spectra which should be more directly comparable between standards prepared in the lab and areas on prints.

----------------------------------------What is RIS?--------------------------------

While FORS measures reflectance at a spot, Reflected Image Spectroscopy (RIS) provides a reflectance map across a selected region, which could include an entire print. The RIS data consists of individual spectra recorded at adjoining points that are usually well under 0.1 mm in diameter. The instrument used here is a Nerios HERA visible/near infrared hyperspectral camera (measuring in the 400-1000 nm range). RIS data are typically evaluated by highlighting areas of the object that was scanned which show identical or very close to identical reflectance spectra. Thus, the data can be a means of indicating where a specific colorant is present over an entire print. For example, the images below show where different colors are distributed in one print.

• 

  RGB projection of hyperspectral data
  MFA# 11.30331

• 

  RIS points similar to yellow selection
  MFA# 11.30331

• 

  RIS points spectrally similar to blue 1 selection
  MFA# 11.30331

• 

  RIS points spectrally similar to blue 2 selection
  MFA# 11.30331

• 

  RIS points spectrally similar to pink selection (safflower)
  MFA# 11.30331

• 

  RIS points spectrally similar to red selection (madder)
  MFA# 11.30331

Expand

RIS in more detail:

Each individual spectrum from a RSI scan can be viewed. One of these or a group of adjoining spectra from a print can be selected as a “reference” spectrum. An algorithm (Spectral Angle Mapper, or SAM) is used to examine all spectra in the scan and determine which ones most closely match the reference spectrum. With appropriate further data processing, the map of the results will accurately show areas of the print that contain the same colorant (or mixture of colorants) as the “reference” spectrum (or group of spectra). The specific colorant or mixture of colorants in the “reference” spectrum or spectra are identified by comparison with spectra from reference materials or from data obtained by other analyses (such as EEM).

The value of this technique is that it can indicate where a specific colorant, which may have only been identified or hypothesized in one small area by one of the point analysis techniques, is present within the entire print. The points analyzed by XRF, FORS and EEM spectroscopy on a print are chosen to include all distinct colorants that appear by close visual examination to be present on the print, but the points may not always have included all colorants, and the points do not show the distribution of a colorant over the entire print.

How to look at the results

The results presented in this database are based on the three (or sometimes four) methods mentioned described above. XRF was run on all data points, but FORS and EEM were utilized according to the visual color, as well as speculated (original) color if fading or other alteration is suspected to have occurred. For example, FORS was run where blue was speculated (i.e. blues, greens, purples, and faded variations), but would not have typically been run on yellows or reds. Also, at this time other analytical or photographic methods that can also aid in the identification of colorants have not been used. Therefore, the results are not all-encompassing and/or definitive.

Results

Shown below is an example of the tables that occur on eah of the print pages. These table include images of the analysis locations as well as a list of the techniques and their resultant interpretations.

Analysis point	Image	Method	Results
Pt 1		XRF	orpiment (EEM to be run)
Pt 2		XRF, EEM	inconclusive
Pt 3		XRF, EEM	orpiment, turmeric
Pt 4		XRF, EEM	safflower (FORS to be run)
Pt 5		XRF, FORS	indigo

Tabular information:

• Analysis point: Each point corresponds to its circle on the print image, which is where the data was collected using the specified analytical methods.
• Image: This is a 10x image collected by the XRF spectrometer that shows the region of analysis.
• Method: This indicates which analytical methods were used on the region of analysis.
• Results:
  -The colorants are interpreted from the collected data in conjunction with visual inspection. For example, if arsenic is detected in a yellow area, it is interpreted as orpiment (arsenic trisulfide).

-If more than one colorant is identified, they are listed in alphabetical order.

-"Inconclusive" indicates that the data collected was not conclusive enough to identify a colorant.

-The additional comment, for example in Pt 1 and Pt 4, indicates that the specified analytical method is scheduled to be conducted in the future. Additionally, Pt 4 is visually a purple color which suggests that it was most likely created from a red and blue colorant but only a red color is listed. This is because FORS which is used to identify blue colorants has not be run; hence the comment, (FORS to be run). If FORS is run in the future and a blue colorant is identified, the result will be updated, but if a blue colorant cannot be identified, the result will only reflect the identified colorant, safflower. For more detail, please see "faded colors" below.

Raw data

The below spectra and pattern are examples of the raw data collected by each analytical method. These data are interpreted by conservation scientists experienced in the particular analytical technique to identify the colorant. As such, only the interpreted results, not the raw data, are given for each print.

• 

  XRF spectrum of Orpiment
• 

  3D EEM pattern of Safflower
• 

  FORS spectrum of Indigo

What colors cannot be identified?

• Faded colors are a challenge and at times not all colors can be fully detected. Although visual observation may give a clue to the colorant or colorants present, only the results from the analytical methods are given. Therefore, if only one colorant of a mixture/overprinting was detected although visually the color is identifiable as a mixture/overprinting, only the detected color is presented in the results. In most cases, XRF, EEM, and FORS still pick up faded colors, but all three techniques are restricted by their detection limits as well as the presence of any fillers and fluorescence in the paper itself.

• Flavonoids are also a challenge. This group of organic yellow colorants containing flavonoids cannot be reliably distinguished from one another with the three analytical methods used in this research. While FTIR reflection spectroscopy has been used to differentiate the flavonoids (see Bibliography) this analytical method was not carried out for this research.
  Dayflower which can shift to a greenish blue, yellow, or tan, also has a similar EEM fluorescence to the flavonoids once it discolors. Luckily, most times the presence of dayflower can be visually distinguished from the other flavonoids and/or FORS analysis can still detect its presence.

• red lead/lead white

• Carbon black cannot be detected via the three analytical methods and was not actively pursued during this research. Therefore, carbon black used as a color (black, gray, or blue gray) or mixed with another color is not identified in this database.

• Orpiment can have different structural forms due to the way it is prepared; directly by grinding the mineral orpiment (natural orpiment) or by heat treating to create a mixture of crystalline and amorphorus form (processed orpiment), and synthetic orpiment. The mineral orpiment and synthetic orpiment both have been found in ukiyo-e prints (see Bibliography). The identification of these different structural forms of orpiment requires the use of Raman spectroscopy which was not conducted for this research.

• Mica While XRF spectroscopy and visual inspection is used to determine the presence of mica,

Japanese terms for color and materials

Throughout history, many of the colors and materials have had multiple names and still may have multiple names associated with them. For this database, we have tried to use the most common currently-used name for the raw material used to produce the color. For example, orpiment (石黄) sekiō is also commonly called kiō. It has also been called yūō (雄黄) which is actually the Chinese word for realgar, while the Chinese term for orpiment, 雌黄 (shiō in Japanese) has been also used to describe orpiment as well as gamboge.

Ukiyo-e Print Colorant Database