National Gallery Technical Bulletin Volume 26, 2005
The Technology of Red Lake Pigment Manufacture: Study of the Dyestuff Substrate
Jo Kirby, Marika Spring and Catherine Higgitt
If recipes for the red lake pigments used in western European easel painting from the twelfth century or earlier until the end of the eighteenth century are examined, it is clear that, apart from the dyestuff, by far the most common ingredient was alum, generally potash alum, potassium aluminium sulphate, AlK(SO4)2·12H2O. The use of alum as a mordant in textile dyeing is well known; in the case of lake pigments, it was used as a reagent to form a substrate for the dyestuff, to make a pigment. The reaction between this and an alkali forms a type of hydrated alumina which precipitates together with the dyestuff, in solution with one or other of the reagents. The alkali was commonly lye prepared from wood ash, but it could equally well be made from lime, or varieties of calcium carbonate, such as chalk, marble dust, egg shells or cuttlefish bone could have been used. These last are more usually found in recipes for yellow lake pigments, but also occur in recipes for red or rose-pink lake pigments from brazilwood, Caesalpinia spp (note 1).
The analysis of red lake pigments has lagged behind that of other pigments used in easel painting. For many years examination of the dyestuff component was generally unfeasible as the size of sample required for spectroscopic or chromatographic analysis was unacceptably large. As these methods became more sophisticated and the sample size could be reduced, examination of the dyestuff extracted from paint samples became possible following methods similar to those already used for textile dyestuff samples (note 2). Until recently, surprisingly little interest has been shown in examination of the substrate, in spite of the fact that it influences the colour, transparency and working properties of the pigment, as well as its permanence (note 3). It is usually assumed that in a painting dating from before the nineteenth century the substrate is hydrated alumina. Before suitable instrumental methods of analysis were available, only the presence of aluminium could be detected, paradoxically often by means of a microchemical test using a dyestuff such as morin or alizarin as reagents (note 4). The nature of the aluminium compound present could not be determined.
The use of energy dispersive X-ray microanalysis in the scanning electron microscope (SEM–EDX) has made the examination of lake substrates much easier, but at the same time has shown that, while aluminium is indeed often the major component detected, the elemental composition of the pigment is more complicated than might be expected. Small, but significant, amounts of other elements, such as sulphur, phosphorus, silicon, magnesium, potassium and calcium, are often also present, and these are not due to interference from other pigments. In some paintings the red lake pigment particles contain relatively little aluminium and the major element detected by EDX is sulphur. The evidence of the recipes suggests that the presence of a significant proportion of calcium as well as aluminium might be expected in some brazilwood lakes, or in certain eighteenth-century cochineal pigments, since calcium-containing materials are listed as ingredients (note 5). The variation in the elements detected by EDX, and the presence of elements other than aluminium, may be explained by the way in which the pigment was made. The sources of the dyestuff, the alkali, the alum – even the recipe itself and the vessels used – all may affect the composition.
In order to investigate how these factors affect the composition of the red lake pigment, an analysis was made of the dyestuffs and other raw materials used to make the pigments, and of a range of lake pigments made in the laboratory following historical recipes, as well as of lake pigments in samples taken from paintings during routine cataloguing work or conservation treatment. The methods used for investigation of inorganic constituents were SEM–EDX and Fourier transform infrared microscopy (FTIR) and spectroscopy. Highperformance liquid chromatography (HPLC) was used to analyse the dyestuffs; the medium and other organic constituents of the paint samples were identified by gas chromatography–mass spectrometry (GC–MS) and FTIR. In certain cases microchemical tests were used. Wherever possible, the paint samples chosen for analysis contained layers consisting only or mainly of red lake, to minimise interference from other pigments. Occasionally more than one red lake pigment was present in the sample, often in different layers and sometimes with different substrates. The results are summarised in Table 1 (see p. 86).
In The Virgin and Child with a Pomegranate (NG 2906), from the workshop of Botticelli and painted around 1480–1500, an intense cherry-red lake pigment has been used in the Virgin’s red dress (plate 1). It contains the dyestuff extracted from the kermes insect, Kermes vermilio Planchon. In the EDX spectrum of the red lake, the largest peak is from aluminium (Al). The FTIR spectrum shows that the substrate is essentially hydrated alumina; it shows a feature at 600 cm-1 characteristic of Al–O lattice vibrations, and other bands (broad c.3400 and c.1650cm-1), from coordinated water (note 6). However, small peaks for sulphur (S), phosphorus (P), silicon (Si), calcium (Ca) and potassium (K) can also be seen in the EDX spectrum (fig. 1), and within the lake glaze are small regions containing less Al and rather more Ca, P and S, suggesting that small amounts of calcium phosphate and sulphate are present.
As the Table shows, elements other than aluminium are frequently detected in lake pigments, even if only in small amounts. Although they could derive from the alkali or alum used, EDX analysis of the scale insects and plant material from which dyestuffs were extracted shows that they all contain these elements, in varying amounts. To explore the contribution of the dyestuff raw material to lake pigments, specimens of each dyestuff from several different sources, together with pigments prepared from them where available, were examined: kermes, Kermes vermilio Planchon (five sources); Polish and Armenian (or Ararat) cochineals, Porphyrophora polonica L. and P. hameli Brandt (one each); the Mexican cochineal insect, Dactylopius coccus Costa (six sources); lac, Kerria lacca Kerr, as sticklac, the form in which it was imported into Europe before the eighteenth century (three sources); and madder root, Rubia tinctorum L. (four sources). In the case of brazilwood, the evaporated aqueous extract from the Old World variety sappanwood, Caesalpinia sappan L., was examined as it was difficult to detect anything in the wood itself.
A similar range of elements was detected in both the plant and insect specimens: S, P, Si, Ca and K, as seen in the painting from the workshop of Botticelli, and also chlorine (Cl), magnesium (Mg), sodium (Na) and copper (Cu). The presence of these elements is not surprising since all living things require a range of minerals to sustain life, usually in very small amounts. In the samples of madder root needle-shaped silicates of sodium, magnesium and aluminium could be observed, which probably derive from aluminosilicates in the soil and, in two specimens, calcium-containing crystals were found: madder will grow in calcium-rich soils and, indeed, it has been observed that the colour of textile dyeings obtained is redder if the plant has been grown on calcareous soils (note 7). Madder also prefers a moist soil, rich in organic matter, with a reasonable phosphorus content (note 8).
The same elements were present in the scale insects, the most significant being P, S, K and Cl, followed by Mg. It is notable that a larger amount of these elements was detected in the New World cochineal insects, which are also the richest in dyestuff, containing at least 10% – and potentially as much as 19% – by weight (note 9). All the insects were markedly richer than the plant sources in phosphorus, an element involved in metabolic processes fundamental to the life of all animals. Scale insects feed on plant sap, the adults remaining attached, immobile, to the plant host and it may be that, like other phytophagous insects, they require a relatively large amount of potassium, but little sodium: in general very little sodium was detected. Other elements likely to be necessary for life include the magnesium and copper detected in all the insects; possibly also manganese and zinc, although these were not found (note 10). Mg was sometimes detected in combination with P (specimens of kermes and cochineal), or with Si (cochineal). Particles containing K and P (perhaps potassium phosphate) were frequently found. Ca was only detected in small amounts, sometimes (in the bodies of the cochineal and kermes insects) as discrete particles in combination with P (so presumably calcium phosphate) or occasionally S (presumably calcium sulphate).
There was some difference between the relative quantity of elements detected within the dyestuffrich bodies of the insects, and in the waxy surface coating they secrete. In most soft scale insects this coating consists largely of waxy, lipid material, and may be quite substantial in those that are immobile in the adult stage of their life cycle: in the case of the lac insect it is very massive indeed and is the source of shellac resin (note 11). As it is secreted by the insect the coating is likely to include excess elements excreted in order to maintain the insect’s physiological equilibrium; in general EDX analysis of the white coatings showed higher levels of calcium salts and silicates than the bodies of the insects. It is also possible that elements could be introduced by surface treatments or spraying, or from the environment in which the animal lives. Armenian cochineal, which has a pronounced white waxy coating, spends its adult life just below the soil surface, attached to the roots of a grass growing in saline marshes; it is therefore no surprise to find sodium and chlorine in the coating.
Examination of lakes prepared from the dyestuff raw materials discussed above showed that some of the elements detected in the insects are also present in the pigments. Hydrated alumina substrates prepared in a way that is analogous to the method used for the pigments, but without dyestuff, contained none of the additional elements with the exception of sulphur: this is discussed below. A reference lake made directly from kermes insects, following a seventeenth-century Italian recipe (note 12), was found to contain a significant amount of P (the peak height in the EDX spectrum is about one-third that of the principal, Al, peak), in addition to trace amounts of S, Ca, Mg and Cu. The only possible source of the phosphorus was the insect and these particular specimens were relatively phosphorus-rich. Apart from aluminium and sulphur (discussed below), the element most frequently detected in the laboratory lake pigments was phosphorus, most conspicuously in the lakes from insect sources. When the dyestuff is not obtained directly from the insect, the phosphorus is present at much lower levels. In a kermes lake prepared by extracting the colorant from silk according to a fifteenth-century Italian recipe, using potassium carbonate solution and precipitating with potash alum (note 13), only a little P, S, Si, Ca and Cu (the elements present in the insect source) were detected in addition to Al. Other elements detected in the insect, including K, Na, Mg and Cl, presumably remained in solution or were washed away and were thus not incorporated into either the dyed textile or the lake pigment derived from it.
A very similar pattern of elements is quite often found in lake pigments from fifteenth- and sixteenth-century paintings, such as the Virgin and Child with a Pomegranate lake pigment described above, and a similar kermes lake in Francesco Bissolo’s Virgin and Child with Saints Michael and Veronica and Two Donors (NG 3083), dating from 1500–25 (plate 2). In a cross-section from Saint Veronica’s red cloak in the latter painting (plate 3), much aluminium was detected in the red lake particles, together with some S, K, Ca, P and Si. In addition, tiny particles of calcium phosphate (Ca, P) were identified within the cherry-red lake particles (figs 2 and 3). Similar particles were seen in the red lake glaze in the Botticelli workshop painting. It is, of course, possible that a little ground bone was added to the lake by design or by chance, but one would then expect the particles to be larger and more irregularly sized: these particles are extremely small. Also, in several of the dyestuff raw materials that were analysed, small particles of calcium phosphate were found, so it is likely that the calcium phosphate in the paint samples originates from the dyestuff source.
Since the same series of elements was seen in all the raw dyestuffs, EDX analysis does not provide precise information on the origin of the dyestuff. However, where a lake pigment is particularly rich in phosphorus it is likely to contain an insect dyestuff. One such lake is that used in the uppermost glaze layer of the kneeling king’s red brocade cloak in Veronese’s Adoration of the Kings (NG 268), painted in 1573. The dyestuff in this case has been identified as perhaps having been extracted from Polish cochineal (note 14). A significant amount of phosphorous was also seen in several of the kermes red lake pigments in the paintings in the Table, for example that used for the Virgin’s dress in Marco Marziale’s Virgin and Child with Saints (NG 804, 1507), and in Matteo di Giovanni’s Saint Sebastian (NG 1461, probably 1480–95).