None.

Metal was analysed with portable XRF machine (XRF1 in fig. 4) on a corroded area on back side of the painting where there was no visible paint flakes or other surface materials present. The analysis confirmed that the metal is an iron (Fe) alloy with high content of silicon (Si) and trace elements of aluminium, magnesium and sulphur (Table 1). We could not determine the specific type of alloy as the carbon content was not analysed. Most likely all present elements except Fe and maybe Si and S are trace elements from pigment residue or atmospheric pollution. 

Table 1: Chemical composition of iron plate. Sample XRF1 on Fig. 4. Method of analysis: handheld XRF, Dr. Eva Menart, National Museum of Slovenia.

Extensive XRF analysis was also made on various locations on both sides of the painting (see fig. 3 and 4) with a goal of identifiying inorganic pigments and composition of the ground layer. Identification of pigments such as emerald green, lead white, Naples yellow, chrome red, vermillion helped us to date the painting. 

The metal is iron alloy, but the specific type of alloy and microstructure of the metal could not be determined without metallographic analysis. 

The iron support is in satisfactory condition and is stable and solid. In places where the paint layer has fallen off, there is a fairly thick layer of passivated corrosion present, except on small local areas of fresh damage, where the paint peels to the ground layer. Corrosion is mostly inactive and stable, there are local smaller active corrosion areas. Corrosion products from the support locally penetrate the paint layer where they cause discoloration. If the paint layer is removed, thin layer of dark brown (almost black) corrosion products can be seen underneath it. Locally, underlaying  thicker deposits of corrosion products are causing paint layer to swell. 

On the front side of the painting, corrosion products are heterogenous and form multiple layers. Statigraphy of corrosion products is similar on most areas and can be divided by four main layers: outer layer of localized brittle, powdery orange products (CP1), followed by reddish brown powdery layer (CP2), followed by dark red (almost black) shiny brittle wart-like structures of different sizes that appear indvidually but cover most of the surface (CP3). The lowest layer consists of dark brown, almost black thin hard layer that covers original metal surface (CP4).(Fig. 5) Paint surface  on the bottom part of the painting was cleaned with cleaning systems that contained water (rigid agarose gels with chelators) that locally came into contact with exposed corroded metal; on those areas bright orange active corrosion (flash corrosion) formed almost immidiately after removal of rigid gels (CP1). (Fig.6).

On the back side, corrosion on the upper area of the painting are the same type as on the front side. (Fig. 7) On the lower area, corrosion type is slightly different: extensive pitting corrosion appears that forms deep individual pits across the whole surface of that area. Corrosion statigraphy is similar as on the front side of the painting: orange, brittle corrosion products (CP1) are located mainly around the pits and inside them on top of other corrosion layers, but also on top of the remaining paint layer - this is probably because of cleaning of the remaining paint layer with agarose gels which contain water (same problem as on the front side of painting). Inside the pits underneath CP1 is a layer of brownish-red brittle corrosion products (CP2), which is hard to distinguish from other layers. Next layer are dark red (almost black) shiny brittle wart-like structures (CP3) and underneath a thin, hard layer of very dark brown products (CP4). (Fig. 8).

Based on the type and colour of corrosion products we can assume that they are mostly iron oxides, with iron hydroxides appearing locally where the surface was in contact with water. This description is based solely on visual analysis. We can assume that the corrosion is atmospheric since the layer is thin and stable and the object was originally kept outside. A sample of corrosion products was taken for XRD analysis in hopes of more precise identification or confirmation of these assumptions (currently awaiting results).