*3.3. Blue Paint*

As with the yellow pennant in *Man-Eater*, the blue pennant had bleached significantly, presenting as a brittle, light blue layer of paint. In chipped areas, darker blue layers of paint were visible below the light blue layer with the most vibrant blue layers closest to the rusted metal. Interestingly, one face of the blue pennant was slightly more vivid in color, which prompted the analysis of two different cross sections. From this point onward, the lighter side will be referred to as Side A, and the darker as Side B. This could be a result of the static nature of this mobile as discussed in internal correspondence, allowing Side A to be more exposed to the sun than Side B.

Analysis of the cross section from Side A (Figure 6) showed a dramatic rusting of the steel, infiltrating through the silver layer and into the first layer of paint. As in the yellow cross section, SEM-EDS confirmed the steel rust and the aluminum flake paint. In total, 6 layers could be delineated on Side A, with colors ranging from midnight to baby blue.

**Figure 6.** Cross section imaging of a sample taken from Side A of the blue pennant in *Man-Eater*. (**a**) BSE image of the cross section and (**b**) associated EDS mapping of relevant chemical elements, of particular note are the Ti and Ca in Layers 8A and 9A, the distribution of Al, S, Na, and Si associated with ultramarine. (**c**) Light microscope image at 20× magnification showing all 9 layers of paint, aluminum anticorrosive paint (red arrow), and the rusting steel (yellow arrow). Image copyright 2021 MoMA, New York, NY, USA.

The analysis of the first two paint layers of the lighter side, 1A and 2A, indicated the use of Prussian blue (P.B. 27; C.I. 77510) as the only blue pigment. CaSO4 was also identified through Raman (1017 cm<sup>−</sup>1). Prussian blue (FeIII[FeII(CN)6]3−), was first synthesized accidentally by Dresbach in 1704 and became industrially produced by the nineteenth century [30]. It is the oldest synthetic coordination compound and has since found important use in the printing industry in addition to paints and artist materials. The dark blue color of Prussian blue is due to an intervalent transition between FeII and FeIII through a coordinated cyano group (CN) where light in the orange-red region around 700 nm is absorbed [30]. The presence of Prussian blue was identified in the Raman spectrum (Figure 7) by the 1Ag *ν*(CN) stretching vibration at 2160 cm−<sup>1</sup> and the Eg *ν*(CN) stretching vibration at 2090 cm−<sup>1</sup> [31]. The spectrum also shows peaks for *<sup>ν</sup>*(Fe−C) stretching modes at 606 and 536 cm<sup>−</sup>1, *<sup>σ</sup>*(Fe−CN−Fe) bending modes at 376 and 278, and a *<sup>σ</sup>*(Fe−C−Fe) deformation at 189 cm−1. Perhaps most interesting is the shoulder at 2123 cm−1, which corresponds to a CN− stretch related to the 1Ag *ν*(CN) mode. This shoulder is most pronounced in the "soluble," or colloidal varieties of Prussian blue, where association with a cationic species such a K+, NH4 +, or Na+ maintains charge balance, as opposed to the insoluble form that relies on a higher concentration of Fe*III* [31,32]. Chemically, Prussian blue is prone to photoreduction and fading, the same quality that made the pigment valuable for cyanotypes, an early photographic method, and soluble varieties are far more sensitive to fading than the insoluble form [33]. The presence of extenders can also exacerbate photoinduced fading, and SEM-EDS showed this paint layer to be particularly rich in magnesium silicates and other silicates. This can explain the discoloration of Layers 1A and 2A and might have prompted repainting of the blue pennant with a darker blue Layer 3A, which was shown to contain ultramarine (P.B. 29; C.I. 77007) and a smaller amount of Prussian blue (Figure 7). The darker appearance of Layer 3A might correspond to the combination of two blues rather than a single color.

**Figure 7.** Raman spectra of the pigments in all 8 layers from Side A of the blue pennant in *Man-Eater*. Ultramarine (†) is observed at 550 cm−<sup>1</sup> in Layers 3A, 4A, 5A, 6A, and 8 A; Prussian blue (‡) at 2090 and 2160 cm−<sup>1</sup> in Layers 1A, 2A, 3A, and 8A. Titanium white was observed in anatase form in layers 3A, 4A, 5A, and 6A, whereas the rutile form was observed in Layers 7A and 8A, in addition to a luminescence pattern (\*) from a Nd3+ impurity attributed to titanium white deposited on CaSO4 (•).

By the 1940s, the use of synthetic ultramarine (3Na2O·3Al2O3·6SiO2·2Na2S) was commonplace after its laboratory preparation in 1828 by Guimet in France and Gmelin in Germany [34]. The incorporation of a sulfur radical (S3 −) into the sodalite crystal lattice of sodium, aluminum, silicon, and oxygen acts as a chromophore. The broad absorption of green-yellow-orange visible light of this radical, centered around 600 nm, gives the pigment its signature blue color. This energy corresponds to an electronic transition between two singly occupied molecular orbitals [35]. Ultramarine was identified through Raman spectroscopy (Figure 7) by the S3 <sup>−</sup> radical symmetric stretch at 540 cm−<sup>1</sup> [36]. Elemental analysis by XRF confirmed the presence of Al, Si, S, and K, whereas SEM-EDS showed the presence of the lighter Na. With a relatively low refractive index of 1.5, the opacity of the ultramarine is increased by the addition of white pigments, which is seen here with the addition of anatase (TiO2), seen in the Raman spectrum at 144 cm<sup>−</sup>1, and confirmed by the presence of Ti in SEM-EDS. The Raman spectrum of Layers 4A and 5A show a higher concentration of anatase resulting in a much lighter blue than Layer 3A; the two are similar in composition and could have been applied in two successive coats that contain ultramarine and anatase. Layers 6A and 7A are even lighter in color than the previous two, but they contain the same mixture of ultramarine and anatase and could also have been applied successively as part of one campaign. The white Layer 8A shows the Nd3+ luminescence pattern associated

with rutile (TiO2) precipitated on CaSO4, which was also observed in Layer 9 of the yellow pennant. Finally, Layer 9A shows very small peaks for ultramarine and BaSO4 in addition to the dominant Nd3+ luminescence pattern of TiO2. Additionally, the detection of niobium (Nb) in the XRF spectra of the blue pennant further narrows down the type of rutile present. The presence of detectable amounts of this rare earth metal by XRF is indicative of the sulfate process for producing titanium white, where it remains after manufacture as an impurity from the ore [37]. As in the case of the pale-yellow Layer 9, TiO2 has been shown to have a catalytic effect on the degradation and fading of Prussian blue, which could explain some of the lighter blues observed in those layers containing the white pigment [28]. Interestingly, the color of the blue pennant was named "light gray" in internal correspondence from 1970, far from the original deep Prussian Blue [22].

Similar to Prussian blue, ultramarine is prone to photoinduced fading; it is also highly sensitive to acids, which are prevalent in urban settings [38]. Furthermore, it has been shown that ultramarine pigment dispersed in an alkyd binder accelerates the degradation of the paint film and results in a bleached and brittle film [26,27]. Consequently, the combination of chemical sensitivity and catalytic effect can explain the pale color of the blue pennant.

Side B (Figure 8) is darker in comparison to Side A, but it also exhibits fading. The same penetration of steel rust up through the aluminum and the first paint layers is again observed. Layers 1B and 2B contains Prussian blue and CaSO4; Layer 3B ultramarine, Prussian blue, and anatase; a thin Layer 4B followed by a thick Layer 5B contain ultramarine and anatase; a tan-colored Layer 6B with anatase; Layer 7B rutile precipitated on CaSO4; and Layer 8B ultramarine with rutile precipitated on CaSO4. After Layer 3B, there is a breakdown in the similarity between the two sides, as the lighter Layers 6A and 7A have no counterpart in the stratigraphy of Side B. The Raman spectrum of Layer 6B indicates it is composed of anatase, but the tan color can be attributed to the presence of Fe, possibly in the form of iron oxides, barites, and silicates, all detected by SEM-EDS. The purpose of this layer is unclear. Layers 7B and 8B have the same composition as Layers 7A and 8A were probably applied at the same time.

**Figure 8.** Cross section imaging of a sample taken from Side B of the blue pennant in *Man-Eater*, showing layer 6B, which is unique to this face of the pennant. Raman analysis indicates this layer to be anatase-based, and SEM-EDS shows it to be rich in iron, which gives the layer its tan color. © 2021 MoMA, New York, NY, USA.

In contrast, the paint of the blue pennant on the maquette is still vivid, and a cross section showed three layers, a white priming layer followed by two blue ones of the same compositions. The blue was identified as ultramarine, which is typical of Calder's mature palette [39]. This paint is also characterized by a ν1(C–O) symmetric stretch at 1090 cm−<sup>1</sup> characteristic of CaCO3. This further confirms that the maquette and *Man-Eater* were painted at different times and with different blue paints. The white priming layer also exhibited the Nd3+ luminescence emission pattern attributed to Nd3+ ions in rutile, and both CaSO4 (1020 cm<sup>−</sup>1) and CaCO3 (1090 cm<sup>−</sup>1), all similar to the priming layer of the yellow pennant. These paints again do no match any of those found in the blue pennant of *Man-Eater.*
