Identifying interlayer surface adhesion failure mechanisms in tinplate packaging steels

Tinplate surface morphology and chemistry is adjusted during the manufacturing process in order to meet the demands of its subsequent product use, the commonest being visual appearance and food packaging stability. A comprehensive experimental study on an industrial tinning line varied the surface roughness and the tin coating weight with the characterization through X‐ray diffraction (XRD), X‐ray photoelectron spectroscopy (XPS), white light interferometer (WLI), optical imaging, and lacquer adhesion measurement. Increasing tin weight lowers the adhesion through the production of a thicker disorganized tin oxide layer which has a greater tendency to fracture under shearing forces. There is no evidence that the substrate roughness improves the adhesion of the lacquer. Analysis of the failure location identifies fracture in the tin oxide layer below the passivation layer. The findings have impacts on the next generation of passivation materials for tinplate as it has been clearly demonstrated that growth in tin oxide thickness, particularly when unstructured, has a detrimental impact on lacquer adhesion.

Tinplate surface morphology and chemistry is adjusted during the manufacturing process in order to meet the demands of its subsequent product use, the commonest being visual appearance and food packaging stability. A comprehensive experimental study on an industrial tinning line varied the surface roughness and the tin coating weight with the characterization through X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), white light interferometer (WLI), optical imaging, and lacquer adhesion measurement. Increasing tin weight lowers the adhesion through the production of a thicker disorganized tin oxide layer which has a greater tendency to fracture under shearing forces. There is no evidence that the substrate roughness improves the adhesion of the lacquer. Analysis of the failure location identifies fracture in the tin oxide layer below the passivation layer. The findings have impacts on the next generation of passivation materials for tinplate as it has been clearly demonstrated that growth in tin oxide thickness, particularly when unstructured, has a detrimental impact on lacquer adhesion. KEYWORDS adhesion, chrome oxide, steel packaging, surface roughness, tinplate

| INTRODUCTION
Tinplate remains an important material for the metal packaging industry for the preservation of foodstuff providing long-term safe food storage capabilities. 1,2 The tinplate substrate is a multilayer system consisting of electrolytically deposited tin to a base steel substrate which can be subsequently reflown and passivated using a dichromate solution, either catholically or via dipping ( Figure 1).
The passivation fulfils four primary roles. It removes oxides present after flow melting of tin and subsequent water quench. It forms a stable and protective tin/chromium oxide film. It forms an oxide film amendable to lacquering, decorative printing, and welding, and it forms an oxide film that is stable towards sulphur staining by protein containing foodstuffs. 3 This manufacturing process produces a layered metallic system which can then be coated with a lacquer or left without a coating depending on the final product requirement. Following cutting, these coated parts are then formed into a variety of common container shapes.
The surface finish of the base steel metal is often manipulated for final product aesthetic reasons through the use of a roughening roll prior to plating. The quantity of tin applied to the surface (tin coating weight) also provides a means of aesthetic control with higher tin coating weights providing a more mirrored gloss finish at the expense of higher cost (associated with the tin) and lower scratch resistance Legislative pressures, which see the ban of the use of epoxy phenolic lacquers 6 and chrome VI 6 passivation in Europe, are restarting the interest in the area as the combined performance of these materials have been shown to perform less robustly in terms of lacquer/substrate adhesion than the incumbent technology. 7,8 These changes potentially have significant implications for the European steel packaging industry. Alternatives to chrome (VI) passivation systems have been studied including cerium, 9 TiO 2 , 10,11 sol-gel derived silanes, 12 polymers, and fluorozirconic acid. 13 While these represent advances in the pursuit of Cr (VI) free passivation, Cr (VI) has remained the dominant industry method.
Where surface characteristics and its impact on adhesion have been examined, this has focused on electrochemical characterization of adhesion 14,15 and the interaction between the packaged foodstuff and substrate corrosion. 12,[16][17][18] Where there has been an examination of the surface properties and the adhesion force, a direct correlation was derived where the adhesion was related to surface passivation constituents through 19  and lead to the exposure of base steel iron, providing potential corrosion initiation sites, 26-28 particularly at low tin coating weights.
Increasing tin coating weight is costly but can lead to a more scratch prone softer surface. A balance must therefore be met which addresses the primary issues of visual appearance, cost, and processing robustness.
In order to investigate the impact of the surface topology and tin coating weight on the adhesion of the lacquer, a comprehensive experimental study was carried out where tinplate was produced with range of base steel surface roughness and a range of tin film weight using current standard Cr (VI) passivation technologies. These substrates were characterized and then subsequently lacquered with epoxy phenolic lacquer and their adhesion measured such that any relationships between surface and adhesion characteristics could be established. The objective of the study was to ascertain the impact of the surface topology and tin coating weight on the physical and chemical characteristics of the surface and to relate these to lacquer/substrate adhesion performance.

| Surface analysis
Surface and near surface chemical analysis was carried out using a Kratos Axis Ultra X-ray photoelectron spectroscopy (XPS) using a monochromatic Al Kα X-ray with a source energy of 1486 eV at TATA Steel Ijmuiden. The spectra obtained were analysed using CasaXPS. Xray diffraction (XRD) characterization was carried out on a Bruker D8 using with measurements taken in 2θ mode between 20 and 110°.
Surface roughness was measured using a Veeco NT200 white light interferometer (WLI) over a 305 μm × 232 μm area with each pixel representing 0.415 μm.

| Adhesion measurement
Although the tape test is a common industrial method for adhesion

| Physical and structural attributes
Increasing the tin coating weight reduces the surface roughness (defined by R a ) for all the surfaces (Figure 3). This impact is most  These measurements therefore provide a useful addition to the knowledge base on these substrates.
Similar characteristics are observed with the tinplate surfaces produced with the shot blasted rollers (Figure 7), with primary species of chrome hydroxide, chrome oxide, tin oxide, and a small quantity of tin.
Where the surface is rougher and the tin coating weight low, then iron is observed closer to the surface ( Figure 7A-C). The presence of iron is absent for any roughness above a 5.6 g/m 2 tin coating weight level ( Figure 6D-L).
The nature of the surface roughness, and how it is produced, does not therefore play a primary role in determining the near surface species concentration, although the rougher surfaces show the presence of iron closer to the surface.   No clear relationship exists for the relationship between the chromic species and the tin coating weight. This is to be expected as the chrome passivation is applied after the tin, and thus the tin thickness (above a minimum which is less than 2.8 g/m 2 ) does not dictate the formation of the chromic species which are above the tin.

| Adhesive attributes
The range of adhesion measured is 2.2 N (6.5 to 8.7 N), which is an order of magnitude compared with a calculated measurement accuracy of 0.2 N. Changes in the surface structure and chemistry species therefore contribute to changes in the lacquer/substrate adhesion.
What is evident from the surface/near surface chemical analysis is that there remains a considerable quantity of tin oxide and chrome hydroxide species at the upper bonding surface. This is contrast to literature 19 where these species are ignored in the prediction of adhesion. This current work shows that there is no correlation between the predicted adhesion and the measured surface adhesion With the tinplate finished with ground rolls (I, II, and III), there is a reduction in the lacquer adhesion as the tin coating weight and tin oxide level increase ( Figure 11A,B). The reduction in adhesion with increasing tin coating weight ( Figure 11A) causes an increase in the tin oxide level ( Figure 9B), which subsequently reduces the force required to produce lacquer/tinplate failure ( Figure 11B). This finding is in line with more recent literature 30 which attributed surface failure to the brittle tin oxide layer. As the thickness of the SnO x layer increases, so there is a greater opportunity for failure in this layer.
In order to confirm that the tin oxide layer was indeed the plane of failure within the structure, XPS measurements were repeated in the line of failure produced by the scratch testing for the smoothest and roughest ground roller finished substrates ( Figure 12). The ground roller finished surfaces were chosen due to their nonisotropic surface structure. This meant that scratches made in the rolling direction would be less susceptible to variations in surface species due to local topology. There is a noticeable difference in the composition of the upper surface and the composition in the valley created by the scratch. Within the scratch, there is no evidence of any chrome species (metallic, oxide, or hydroxide) indicating that these have been removed by the stylus. There is no exposed iron at the surface indicating that the failure has occurred above the iron and the iron/tin intermetallic layer. The primary difference in the surface chemistry is an increase in the oxides associated with the tin oxide formation. As  The identification of the SnO x as the point of adhesion failure has implications for next current and future developments and usage using tinplate. Any passivation technology which replaces Cr (VI) will need to prevent the growth of oxide with performance at least comparable with the incumbent Cr (VI). Thermal history 7 and prolonged storage have been shown to be contributory factors to tin oxide growth, and this imposes additional requirements on passivation. The dataset provided by this paper provides a measure against which the future passivation technologies can be compared in terms of the adhesion. An optimum production value for the tin oxide is difficult as it must consider the impact of the passivation layer and likely growth and the requirement for some amount of tin oxide in order to improve scratch performance.

| CONCLUSIONS
An experimental study of the effect of surface roughness and tin coating weight on adhesion has been carried out. There is a link between an increase in the tin coating weight and a change in the orientation of the tin on the surface indicating that higher tin coating weights have a more disorganized structure. This in turn means that the higher tin coating weights lead to thicker oxide films and these thicker tin oxide films lead to a measurable reduction in the adhesion performance. The study concluded that there is no correlation between the surface roughness and the measured adhesion and that the underlying macrostructure of the texturing roll surface (ground or shot blasted) does not play a significant role in the adhesion. The primary means of adhesion failure has proven to be the brittle tin oxide layer between the tin surface and the chrome passivation layer.