Bismuth White (Bismuth Oxychloride) and Its Use in Portrait Miniatures Painted by George Engleheart

Abstract

This article documents the discovery of ‘bismuth white’ on three late eighteenth-century portrait miniatures in the Victoria and Albert Museum collections, painted by renowned English artist George Engleheart. Metallic bismuth and bismuth-containing minerals have been known for centuries and were used on various types of artistic production, from German Wismutmalerei to medieval manuscripts and Renaissance paintings. However, until now they had never been documented on portrait miniatures, despite documentary evidence that suggests their use. The Raman analysis of the three miniatures shows that bismuth oxychloride (BiOCl, corresponding to the mineral bismoclite) is present, and XRF data prove that this material was used as a white pigment in its own right. This work is a pilot study: it represents the first step in the rediscovery of bismuth white as an artist’s pigment, and hopes to provide encouragement to other institutions to look deeper in their collections and map out the use of a relatively rare white material which until now had not been detected or documented in fine art objects.

1. Introduction

Bismuth is a relatively rare element and can be found in nature in its native metallic form [1] or in a limited number of mineral phases [2], including bismite (bismuth oxide Bi₂O₃), bismuthinite (bismuth sulfide Bi₂S₃), bismutite (bismuth carbonate (BiO)₂CO₃), beyerite (calcium bismuth carbonate Ca(BiO)₂(CO₃)₂) and bismoclite (bismuth oxychloride BiOCl) [3,4,5,6,7]. It is unclear when it was first discovered; it is not mentioned by the ancients, such as Pliny, and it is possible that it was often mistaken for other metals such as lead, tin or antimony. As discussed in other publications [8], the earliest documented use of a bismuth-containing compound as an artist’s material was in a 1384 German manuscript containing a recipe for a bismuth-containing ink. The element truly takes centre stage in the fifteenth century in modern-day Germany and Switzerland, when it gives its name to Wismutmalerei (i.e., ‘bismuth painting’), an art form where powdered bismuth is applied on small wooden boxes and burnished [9]. For an example of this type of production, see the V&A box with accession number 575-1872 on https://collections.vam.ac.uk/item/O58643/box-unknown/ (accessed on 22 June 2024). The French royal court painter Jean Bourdichon is known to have used metallic bismuth not only as a paint pigment to create a soft grey colour, but also as a tracing or sketching material for his underdrawings. This dual use was discovered during the scientific analysis of a number of illuminations from Bourdichon’s Book of Hours of Luis XII, commissioned in 1499 [8,10,11]. Other artists, including Raphael, used bismuth in their paintings [12,13], and others used it in polychrome sculpture [14].

While grey metallic bismuth has recently been ‘rediscovered’ on several works of art, the historic use of white bismuth-containing materials is not well documented. There are very few references to them over the centuries: Théodore de Mayerne (1573–1655) mentions bismuth white several times [15,16], William Wood (1768–1810) writes about ‘bizmuth white’ in the first few pages of his hand-written ledgers [17], Pierre-François Tingry (1743–1821) refers to it as Cremnitz white in his guide [18] (although this is a name usually associated with lead white), and George Field (1777?–1854) includes it in his treatise [19]. The consistent use of white bismuth pigments is only verified from the twentieth century onwards. The man-made version of bismoclite, bismuth oxychloride, was the first synthetic non-toxic nacreous pigment: it was employed as an artist’s pigment [20], and its use spread not only within the cosmetic industry (due to the lustrous quality of its appearance [21,22]) but also to other industrial sectors [23,24]. It is currently listed in the Colour Index as CI 77163.

There is some confusion about the names under which bismuth-containing white materials are recorded and referred to. While in the modern literature the name ‘bismuth white’ is associated with bismuth oxychloride [25], in others it refers to bismuth subnitrate (4BiNO₃(OH)₂.BiO(OH)) [26] (p. 42), [27] (p. 34), [28], and both compounds can unhelpfully be referred to as ‘Spanish white’ in both modern and historical references [25,29,30]. Tingry refers to it as Cremnitz white, although the name appears to encompass both the subnitrate and the oxychloride (‘made with bismuth oxidated by means of nitrous acid, or in any other manner’ [18] (p. 285)). Complicating the issue is also the fact that, occasionally, the name ‘pearl white’ makes its appearance with reference to a bismuth-containing material [19] (p. 70 and p. 186), [27] (p. 286), but more frequently—and understandably—‘pearl white’ refers to the white material prepared by crushing mother-of-pearl (for example, in the preparation of gofun in Japan) [31] (p. 207), [32,33]. Recently, databases and online resources hosted by collections and heritage institutions have recognised this confusing terminology and have clearly listed all the various designations [34,35].

At the Victoria and Albert Museum (V&A), the first identification of bismuth oxychloride BiOCl occurred in early 2016 during the routine analysis of three portrait miniatures by George Engleheart, one of the most celebrated English portrait miniature artists of the Georgian period [36,37]. For a variety of logistical reasons, no further miniatures could be examined at the time, and the initial results remained unpublished. This work provides an overview of the results obtained so far and represents a pilot study that can alert other collections about the occurrence of bismuth white and provide guidelines about its characterisation. It also outlines a plan for future investigations of V&A miniatures by George Engleheart and other artists.

2. Materials and Methods

The three miniatures analysed in this study were as follows: an unknown woman, said to be Mrs Dicksee (V&A no. P.22-1937), an unknown woman (V&A no. P.20-1929) and an unknown woman (V&A no. P.26-1922), shown in Figure 1. Details about the three miniatures can be found online on the following pages: https://collections.vam.ac.uk/item/O82021/portrait-miniature-of-an-unknown-portrait-miniature-engleheart-george/; https://collections.vam.ac.uk/item/O82022/portrait-of-an-unknown-woman-portrait-miniature-engleheart-george/; https://collections.vam.ac.uk/item/O1070137/portrait-miniature-of-an-unknown-portrait-miniature-george-engleheart/ (accessed on 23 May 2024).

During the 2016 analysis campaign, the V&A miniatures were examined using a routine analysis slot, with finite time and resources. Point analysis XRF was performed first, followed by Raman microscopy in a number of areas of interest. After that, all miniatures were put back into their jewelled cases and could no longer be accessed for additional scientific investigations due to conservation precautions (opening and closing historical miniature lockets and cases can affect the miniatures themselves). A follow-up in 2024 was possible for the only miniature that was still out of its case, and scanning XRF was therefore performed. The experimental details are as follows:

Micro-X-ray fluorescence analysis (μXRF)—point analysis: The μXRF analysis was performed using a Bruker ArtTAX equipped with a molybdenum source (Bruker Nano GmbH, Berlin, Germany). The experimental conditions were set at 50 kV, 600 mA and 100 s livetime. The area examined was approximately 200 μm across.

Raman microscopy: The analyses were carried out with a Horiba XploRA (Horiba, Kyoto, Japan) equipped with two diode lasers (532 and 638 nm) and an Olympus microscope. Only the ×50 objective was used, providing an overall magnification of 500. The power at the sample was always kept below 1 mW. Total accumulation times varied between 10 s and 1 min, and no spectral manipulations were used, except for the proprietary ICS function when needed. The spectra obtained from the objects were compared to reference spectra collected in-house, to those in published databases [38,39] and to online resources [40]. A specimen of bismuth oxychloride (BM.1948,216) from the Natural History Museum, London, was used to produce a reference spectrum for comparison purposes.

Scanning X-ray fluorescence (XRF): The scans were carried out using a Bruker M6 Jetstream spectrometer equipped with a Rh-target microfocus X-ray tube and two 60 mm² XFlash silicon drift detectors (SDDs) (Bruker Nano GmbH, Berlin, Germany). The X-ray tube was operated at 50 kV and 600 μA. The elemental distribution maps were collected with a 100 μm spot size, a 120 μm step size and a dwell time of 100 ms/pixel. The X-ray fluorescence spectra were calibrated, fitted and processed using the Bruker M6 Jetstream software, version 1.6.758.0.

3. Results

The analysis of the three miniatures originated from specific questions that were unrelated to the presence or otherwise of bismuth-containing pigments (there was no reason to suspect in advance that they could have been used on the objects):

• Miniature P.22-1937 was analysed to look at the discolouration of some of the areas of the dress;
• Miniature P.20-1929 was analysed to identify the blue pigments;
• Miniature P.26-1922 was analysed to look at the discolouration on the reverse of the object.

An account of the full analysis results goes beyond the scope of this paper, but it was confirmed that none of the discolouration mentioned above was associated with the presence of bismuth-containing species.

During the course of the preliminary XRF analyses, it became apparent that bismuth was present in significant amount in many, but not all, white areas, as well as in other lightly coloured areas (an example is given in Figure 2). When individual spots in some of these areas were analysed by Raman microscopy, bismuth oxychloride was detected (Figure 3).

Table 1. Synopsis of the XRF and Raman analyses of areas containing bismuth.

Miniature	Area	Bismuth (point XRF)	Bismuth oxychloride (Raman)
P.22-1937	Pearl in headdress	✓	✓
	Pale blue dress	✓	✓
P.20-1929	Hat brim	✓	✓
	Ruff highlights	✓	✓
	Hair strands	✓	v
	Ruff (darkened)	✓	✓
P.26-1922	Shawl highlights	✓	✓
	Grey dress	✓	✓
	Dress white highlights	✓	✓

shows a synopsis of the relevant 2016 analysis results.

When the areas containing bismuth were examined under a stereo microscope, they appeared to be in good condition and occasionally showed a faintly sparkly appearance. The only exceptions were seen in P.20-1929 and P.26-1922 (macrophotographs are shown in Figure 4): in the first instance, some of the highlights in the ruff (Figure 4a) may have indeed degraded and do display a marked iridescence under magnification. In the second instance, the grey appearance is deliberate and is due to the mixing of bismuth white with a finely divided black pigment (Figure 4b).

In 2021, the V&A Science Laboratory was refurbished owing to a generous grant from the Arts and Humanities Research Council (AHRC). The refurbishment included the purchase of state-of-the-art equipment, such as a Bruker M6 Jetstream, which has since been used to perform scanning XRF experiments.

Ideally, all three miniatures would have been scanned with the new equipment to obtain distribution maps of bismuth for each. However, as two of the miniatures had already been put back in their lockets and could not be reopened, scanning the miniatures through the glass proved impossible due to interference of the latter.

Only P.26-1922 was available out of its locket and could, therefore, be scanned, revealing that bismuth white had been extensively employed in the sitter’s clothing (Figure 5).

4. Discussion

All the early sources, by various degrees, suggest that bismuth white is unstable, implying that it is not the best white pigment to use. De Mayerne suggests that it blackens in the sun [15] (f10v), although elsewhere (f92r) he refers to it as ‘blanc excellent’, or ‘superb’ in the published English translation. Pierre-François Tingry also warns of the tendency of Cremnitz white to alter under the effect of light or of ‘vapours which arise from stagnant water, privies, etc.’ [18] (pp. 285 and 298). Finally, George Field also mentions the tendency of metallic whites, including that of bismuth, to change colour [19] (p. 71 and p. 186). William Wood deserves a more detailed mention: he refers to ‘bizmuth white’ as ‘bad’ on the first page of his ledgers (Figure 6a), held at the Victoria and Albert Museum’s National Art Library, and the pigment is crossed out twice in subsequent records (Figure 6b,d), perhaps showing that Wood changed his mind about using it [17].

According to the transcription of the ledgers carried out by Hark, Millunzi and Ordiway in 2017 [41], Wood mentions ‘bizmuth white’ only five times overall. However, one needs to consider that reference to the material could be hidden in plain sight: Wood often used code numbers as a shorthand for the pigments and other artists’ materials he employed in his work—it cannot be ruled out that bismuth white is one of the hitherto undeciphered numbers (a detailed study of William Wood’s ‘Memorandum of Miniatures’ manuscripts and the code therein is currently being submitted for publication elsewhere; the study also includes details of the scientific analysis and technical examination of a number of Wood’s miniatures from the V&A and other collections). It is curious that ‘bizmuth white’ is only mentioned by Wood at the very beginning of his first ledger and only in miniatures that were marked as completed between April and August 1790. It is tempting to postulate that he quickly realised that the material was not reliable nor durable and had to be replaced with an alternative white pigment. To the author’s knowledge, bismuth oxychloride has not been identified in any of the over 20 miniatures by William Wood analysed so far in different institutions [42].

Instability and unreliability are not associated with the historical pigment alone: the twentieth-century synthetic material is also recognised as relatively unstable [43] (p. 120).

As outlined in Section 3, most of the areas where bismuth oxychloride was detected do not appear to have degraded, possibly due to the limited exposure of the miniatures to light and pollutants over their lifetime. However, there are a few areas that look suspiciously grey and may have changed over time. The next step includes accessing the three miniatures again in order to perform a thorough examination of the objects both visually and using scanning XRF. Ideally, the study will be extended to the rest of the V&A miniatures attributed to George Engleheart to confirm if his use of bismuth white was as widespread as the analysis of these three miniatures suggests, and if there was a point in time when the artist stopped using it, perhaps due to its instability.

The present study suggests that the use of bismuth white by Engleheart was deliberate, especially where bold highlights were required. The way this pigment is used on the miniatures matches the use of the more conventional lead white, which is sought when a white pigment with a high covering power is needed.

That bismuth white has not been identified until now within the oeuvre of any artist indirectly confirms that its medium- and long-term behaviours were not satisfactory, and the pigment was, therefore, not a recurrent presence in late eighteenth-century artists’ palette. This hypothesis, however, should be confirmed by planning a systematic survey of late eighteenth-century and early nineteenth-century miniatures wherever possible. The V&A holds the national collection of portrait miniatures and for many decades has been at the forefront of related scholarship and research [44,45,46,47,48], thriving in an environment where curators, art historians, conservators and scientists work closely together in a cross-disciplinary setting. The museum’s plan will start with a rapid screening campaign of unmounted miniatures using scanning XRF. The miniatures which are found to contain bismuth will then be investigated by Raman microscopy to confirm the presence of bismuth oxychloride or any other bismuth-containing materials. High-resolution digital microscopy will add information as to the state of conservation of the areas painted with bismuth white, which, in turn, will provide data on the long-term stability of the pigment and on its compatibility with other materials used in this form of art.