JAIC 1984, Volume 24, Number 1, Article 4 (pp. 33 to 39)
JAIC online
Journal of the American Institute for Conservation
JAIC 1984, Volume 24, Number 1, Article 4 (pp. 33 to 39)


Ann Cordy, & Kwan-nan Yeh

ABSTRACT—Laboratory identification of three major nineteenth-century blue dyes—indigo, Prussian blue, and iron-mordanted logwood—has been investigated. Acid digestion of the cellulosic fiber was followed by separation into organic and inorganic phases. The organic phase was examined with visible spectroscopy showing a maximum peak range of 602 to 605 nm for indigo and for logwood a maximum peak of 518 nm. The inorganic phase produced positive identification for Prussian blue through use of wet chemical analysis for ferrocyanide ions. Laboratory dyed samples were artificially aged with light and heat to simulate historic textile conditions. These artificially aged samples were analyzed as above and produced results similar to the non-aged dyed thread except for a loss of the visible spectroscopy peak of logwood. A flow chart for identifying these three blue dyes on small samples of cellulosic fibers is presented.


THE STUDY REPORTED HERE arose from an investigation of dyes used on mid-nineteenth-century flax sewing thread.1 A review of the nineteenth-century scientific literature narrowed the possible blue dyes used on these cellulosic fibers to indigo, Prussian blue, and logwood (mordanted with iron).2 An identification scheme was needed to discriminate among these three dyes and is presented in this paper.

Flax thread samples were prepared and dyed in the laboratory using documented nineteenth-century recipes and processes. Some of the laboratory-prepared samples were artificially aged and used for characterizing the blue dyes and their possible degradation products. A rapid acid digestion technique was used to release dyes on the fiber into solution. Both spectroscopic and wet chemical techniques were then used to systematically analyze the chemicals removed.


2.1 Materials

Two-ply, s-twist, wet-spun flax thread obtained from a commercial source was scoured and bleached prior to dyeing by nineteenth-century methods. These materials were selected after analysis of the original sewing thread indicated a flax thread of this type.3 Natural indigo and logwood were obtained from commercial sources. Prussian blue was produced on the fiber. All other chemicals used for pre-dye thread preparation and analysis of dyes were reagent-grade laboratory chemicals. Steps and chemicals involved in both pre-dye thread preparation and dyeing are reported elsewhere.4

2.2 Aging Procedures

The artificially aged samples were prepared by subjecting undyed flax thread and thread that had been dyed as above to light and heat under atmospheric conditions. This accelerated aging with light and heat was chosen to approximate the aging conditions of the nineteenth-century sewing thread. Aging was purposely severe to simulate extreme conditions so that identification techniques would have validity for historic textiles as well as for new fibers.

Lightfastness testing was done using an Atlas Fadeometer with a carbon-arc lamp. Relative humidity was kept at 32 to 36%. The dry bulb temperature was maintained at 124 to 127F. Color was noted at ten to twenty hour intervals. Testing was stopped at 215.75 hours when the thread showed signs of brittleness.

Heat-aging was done in a forced-air oven maintained at 148C. Skeins of about 0.2 grams of dyed and undyed thread were hung from the oven rack without touching any metal surfaces or each other. The maximum time in the oven was 254 hours when thread brittleness occurred.

2.3 Analytical Procedures

The analysis techniques were designed to make maximum use of limited sample size and for discrimination among the three blue dyes. The overall flow chart is presented in Figure 1. A 50% sulfuric acid solution was used to digest the flax thread within one to two hours without destroying the dyes.

Fig. 1. Dye Extraction Flow Chart

2.3.1 Ultraviolet-Visible Spectroscopy (UV-VIS).

UV-VIS analysis was done with a Beckman Model 25 Spectrophotometer with a Beckman Recorder. The methylene chloride layer of acid-digested aged and non-aged dyed flax thread was analyzed with UV-VIS spectroscopy.

2.3.2 Infrared Spectroscopy (IR).

Samples were run on a Perkin Elmer Infrared Spectrophotometer (#281-B) with an interfaced data station. The methylene chloride extraction layer was analyzed by IR.

2.3.3 Wet Chemical Analysis.

The aqueous layer from the digested samples was examined for ferricyanide and ferrocyanide ions of Prussian blue. The addition of iron salts to solutions containing ferricyanide or ferrocyanide ions gives a dark blue precipitate.5


3.1 Non-Aged Thread

3.1.1 Acid Digestion Colors.

The colors of the dyed thread solutions were noted after acid digestion and then after extraction with methylene chloride (see Table I). All indigo samples yielded a blue color on extraction. The organic indigo dye molecule was soluble in the organic layer, while some water-soluble sulfonated indigo appeared to remain in the acid layer. Logwood-dyed thread produced a pink acid solution. Prussian blue-dyed samples yielded light blue in the acid layer. Undyed flax thread produced a yellow color in the methylene chloride layer.

Table I Acid Digestion Colors for Dyed and Undyed Flax Thread

3.1.2 UV-VIS Spectroscopy.

Only indigo colored the methylene chloride layer, producing a visible peak at 602 nm. This was close to the peak found by Whiting of 600 nm in methylene chloride.6Table II lists VIS spectroscopic maximum peaks for indigo. Prussian blue gave poor readings with this technique so no results are presented. Logwood produced only one peak as discussed below. UV spectrum for all samples produced weak peaks.

Table II Observed Absorption Maxima (NM) for Indigo in the Visible Region

Vis spectroscopic analysis of the acid layer yielded a 645 nm peak for indigo (possibly sulfonated indigo) and a 518 nm peak for logwood. Whiting had obtained a peak of 520 nm for logwood in dilute acid.7

The acid layer was neutralized and evaporated to a residue which was dissolved in water. It was hoped that a useful UV-VIS spectrum could be produced. However, the residues were not readily soluble in water and the resulting suspension gave ambiguous results. The methanol extraction of the residues produced clear peaks only for indigo at 607 nm.

3.1.3 IR Spectroscopy.

Characteristic organic functional groups were searched for from the methylene chloride layers. Due to water interference, attempts to use potassium bromide pellets of evaporated neutralized acid layers were not successful.

As expected from the color of the solutions (possible indication of dye presence), only indigo provided identifiable peaks from the methylene chloride layer, and those were for the sulphonated indigo that had been formed during the acid digestion step (sulphonated aromatic rings).

3.1.4 Wet Chemical Analysis.

Wet chemical identification gave a positive response for ferrocyanide ions only with Prussian blue. No positive responses were detected for ferricyanide ions.

Ferrous ions were found with all three dyes. The ferric ion test produced weak positive tests for all dyed samples. Ferrous sulphate was used as a reducing agent for indigo dyeing on cellulosics in the mid-nineteenth century and was also used in this study. An iron mordant was used to produce logwood blue. Iron was also present in Prussian blue. Testing for ferric or ferrous ions was, therefore, not a discriminating test between these dyes but does point to the presence of a dye since the undyed flax gave negative results for all the above ion tests.

3.2 Aged Thread

3.2.1 Color on Aging.

The Prussian blue and indigo-dyed thread remained blue throughout the light treatment. After only ten hours the logwood sample had lost its original color and had become tan.

All control thread was brittle when removed from the oven after 254 hours and broke easily when bent or gently pulled. Again, the indigo and Prussian blue dyed threads retained their blue color. The Prussian blue lost its intensity and dulled to a grey-blue similar to that of the indigo samples (a point to be aware of as some dark blue textiles of an “indigo” color may actually be dyed with Prussian blue that has faded). All logwood dyed samples became dark brown with less then ten hours of heat. The undyed thread became dark brown by the end of 86 hours.

3.2.2 Acid Digestion Colors.

Colors for aged samples in the extraction solvents are given in Table I. Undyed aged flax produced an amber color in the acid layer rather than in the organic methylene chloride layer, an indication of some fiber degradation. The indigo and Prussian blue samples still produced blue in the methylene chloride and acid solutions as had the non-aged thread. Logwood samples did not retain their blue color with aging.

3.2.3 UV-VIS Spectroscopy.

Aged dyed thread responded similarly to non-aged dyed samples in UV-VIS spectroscopy analysis except for indigo in acid and the logwood-dyed thread. The UV spectrum of aged samples did not provide positive identification of a dye.

Aged indigo samples produced VIS maxima peaks in both methylene chloride and methanol solutions that were extremely close to those of the non-aged thread (see Table II). Indigo-dyed thread produced a differing maxima peak in acid after heat treatment (670 versus 645 nm non-aged). The peak shift in acid with aging could cause unclear identification and therefore the acid layer VIS analysis might not be a good choice for aged samples even though the textile may still be blue.

Aged logwood thread lost its color within ten hours of both heat and light treatment and produced no peaks in the visible range. Prussian blue samples yielded no identifiable peaks whether aged or not.

A change in color of the dyed textile will not give the same VIS spectroscopic results due to the color shift and therefore is not a suitable test if a color change is suspected to have occurred.

3.2.4 IR Spectroscopy.

Artificially aged indigo-dyed samples no longer showed absorption peaks of indigo or sulphonated indigo despite the blue color of the methylene chloride layer. This was perhaps due to the decreased amount of dye in this layer with aging (lighter blue solution than non-aged). More sensitive IR detection would be needed to identify the indigo. Heat and light treated indigo, logwood, and Prussian blue samples showed only peaks for undyed flax.

3.2.5 Wet Chemical Analysis.

Even with severe artificial aging the ferrocyanide ion test still yielded positive results for Prussian blue. The ferricyanide test still produced negative results for all samples as it did for the non-aged thread. Positive tests were found for ferric and ferrous ions on all aged samples except for ferric ion with heat-treated logwood. Undyed flax yielded negative results for all tested ions.


THIS STUDY USED MODERN FLAX THREAD dyed blue with either indigo, Prussian blue or logwood (iron mordant) using nineteenth century dye recipes. Responses of severely light- and heat-treated dyed modern thread simulated aging. Severely aged thread responses to identification techniques produced results similar to those for the nonaged samples and appear reliable for this set of aging conditions.

The acid digestion technique would be good for cellulosics due to their acid sensitivity. Indigo, logwood, and Prussian blue dyes were not destroyed by this procedure. Some indigo was made water soluble with the addition of sulfonates. Logwood converted to pink in the acid when the mordant was cleaved off.

UV-VIS spectroscopy produced clear maxima peaks in the visible range for indigo and logwood. Non-aged logwood or logwood samples that are still blue can be identified in acid solution (VIS peak of 518 nm) but on aging this peak disappears entirely. Therefore, if the fiber is still blue then logwood should be identifiable with VIS spectroscopy. The UV spectrum showed interference from degraded flax fiber with all dyed samples. The UV portion is not recommended for use in discriminating among these blue dyes.

IR spectroscopy analysis of the non-aged indigo methylene chloride layer identified sulphonated groups, but with aging these identifying peaks were lost. This technique requires more sensitive sampling (e.g., microcell) to concentrate the small dye sample size. A method is needed to examine the acid layer in which non-aged logwood produced a pink and Prussian blue a blue color. The Prussian blue would not be expected to yield much data with IR due to its inorganic structure but should show the cyanide group band.

Wet chemical analysis for ferrocyanide ions identified the presence of Prussian blue both for aged and non-aged samples. If metallic ions are to be identified by wet chemical analysis, then efforts should be made to use dye recipes of the period under study. Ingredients other than dyes vary with the time period and can give false identification for certain metal ion tests. It would be best to test for the dye molecule itself or, as with Prussian blue, select a more dye-specific ion to test for (e.g., ferrocyanide ion, not ferrous or ferric ions).

Based on the results of this study and the review of literature, a suggested flow chart has been developed for identifying blue dyes on cellulosic fibers of 1850 to 1870 (see Figure 2). It is designed to be inexpensive, use a relatively small amount of sample, and test with equipment commonly available in most chemical laboratories; it is aimed at minimal toxicity exposure from solvents. It should be kept in mind that this chart is designed for cellulosic fibers and also that it does not mention certain other blue dyes of earlier periods (e.g., woad) or later synthetic blues.

Fig. 2. Flow Chart for Blue Dye Identification


A SPECIAL THANKS IS EXTENDED to Dr. Steven Spivak and Mr. Earl J. Coates whose initial questioning, encouragement, and guidance led to the development of this research. The Division of Textiles, Museum of American History, Smithsonian Institution kindly allowed access to their excellent collection of nineteenth-century dye manuals for this research.


A.Cordy, “Investigation of Thread Color Change in American Civil War Uniforms” (Ph.D. dissertation, University of Maryland, College Park, MD, 1983).

Ibid., pp. 47–51.

Ibid., pp. 156–157.

Ibid., pp. 159–167.

Moeller, Qualitative Analysis: An Introduction to Equilibrium and Solution Chemistry (N.Y.: McGraw-Hill Book Co., Inc., 1958): 236.

M.Whiting, “The Identification of Dyes in Old Oriental Textiles,” ICOM Report of the Committee for Conservation (Fifth Triennial Meeting, Zagreb, Yugoslavia, 1978).


J. HofenkdeGraaff “A Simple Method for the Identification of Indigo,” Studies in Conservation19 (1974):54–55

E.Knecht, C.Rawson, and R.Loewenthal, A Manual of Dyeing, 2 vols., 5th ed. (London: Charles Griffin and Co., Inc., 1919), 2:695.

Section Index

Copyright 1984 American Institute of Historic and Artistic Works