REASONS FOR CORROSION DAMAGES OF SEGMENTED, ENAMELLED TANKS
B. Radomski, G. H. Frischat, P. Hellmold
Institut für Nichtmetallische Werkstoffe, Technische Universität Clausthal, Zehntnerstr. 2a, 38678 Clausthal-Zellerfeld, Germany

Introduction
The combination of the properties of metals and glasses offers for the composite material metal/enamel varied applications in the architecture, household and sanitary areas as well asadvantages in apparatus and plant construction, especially

Owing to its low price in comparison to other materials like stainless steel, concrete and organic synthetic materials, enamel is used for segmented tanks (volume up to 20.000 m³) in the water and sewage technology.
Sheets (with a thickness between 3 –12 mm), enamelled with one, two and three coat enamels ensure a high corrosion resistance against liquid and gaseous components, which are formedduring water purification.
Occasionally damages occur above the liquid level particularly in containers for the biological purification of sewage. Till now, the causes for these selective damages have not been examined systematically. A feasible cause for the appearing damages could be a corrosive attack on thecomposite material by gaseous components (CO₂, H₂S, O₂, H₂O) in a moist atmosphere. As further corrosion reasons biologically induced initial phases of corrosion [1, 2] as well as defects ofmaterial, geometry, transport and assembly defects have to be taken into consideration. Latter fault possibilities (steel defects, strong mechanical loads of the enamel) lead to an easy direct attack of the corrosion medium on the steel, where enamel parts can be cracked away. This will advancethe corrosion of the composite material.

Aims and solution ways
The systematic laboratory examinations on the corrosion behaviour were primarily focused onindustrially used enamels (one, two, three coat enamel, see Table 1) [3, 4, 5], which were exposed to corrosion both in non-damaged and pre-damaged conditions by gas mixtures of H₂S, CO₂, H₂O and O₂, which could be varied in a wide range. Chapter 3 lists characteristic results of corrosion experiments for the following gas mixtures (see Table 2).

The corrosion tests were carried out at a temperature of 50 °C and under a relative atmospheric humidity of 50 % according to the norm DIN ISO 2733 [6]. The corrosion experiments of the three different types of enamelled sheets (size: 10 x 10 cm2) were executed in a testing equipment (see Fig. 1). The bottom (below) and the top plate (above) of the glass cylinder were inspected after each corrosion test.
A mechanical pre-damage of the enamel sheets by variable impacts with the Wegener gun was supposed to simulate the consequences of transport and assembly damages (impact, chippings at the screw-joints, etc.). Furthermore mechanical pre-damages were achieved by a fish scale formation (diffusion of hydrogen from the steel into the enamel layer).
The chemical pre-damage was obtained by etching the enamel plates with sulphuric acid (24 hours with 2 % sulphuric acid at 40°C) to simulate the influence of biologically induced initial phases of corrosion by sulphur bacilli. Sulphur bacilli are able to form sulphuric acid resulting from sulphurous educts of the sewage reprocessing. Sulphuric acid attacks the enamel at a very low pH value (<1).
The characterization of the corrosive attack of the different types of enamelled plates by the gaseous media were made by gravimetry, gloss measurement (gloss loss), optical microscopy, infrared spectroscopy and analysis of the leached components by applying ion chromatography.

Results
In the following selected results for the corrosion behaviour of the different enamels are presented.

One coat enamel
For economic reasons one-coat enamel is preferred for the production of the composite material.
As a consequence of the renunciation of the ground enamel interlayer the one coat enamel has a lower chemical resistance than the two coat enamel.
Some corrosion results of non-damaged and chemically pre-damaged enamel samples after a testing period of 7 days are given in Table 3.

After the testing period the one coat enamel showed a rough surface (a clear loss of gloss) as well as a formation of small reaction products on the enamel. The chemical treatment of the enamel (etching with sulphuric acid) led to a strong roughness on the enamel surface, which is covered for the most part with reaction products.
The time dependence of the corrosion of the non- damaged one coat enamel is presented in Table

An increase in corrosion time leads to a strong roughening up to serious changes of the enamel surface. With that a high enamel mass loss caused by the condensates of the corrosion medium is connected. This increased enamel mass loss as a result of the relatively high B₂O₃ content causes the opening of the surface near pores (larger surface), which leads to an additional increase of corrosion. The consequences of these corrosion processes are displayed schematically in Fig. 2.

Fig. 2 Schematic illustration of the corrosion of one-coat enamel

Two coat enamel
The largest number of corrosion tests was carried out on the two coat enamels. Ground and cover enamelled steel sheets are till now used most frequently for the corrosion prevention of tank
buildings.
Characteristic results are listed in Table 5.

In comparison to the one coat enamel the smooth and shiny surfaces of the chemically predamaged two coat enamel samples should be particularly mentioned. The analysis of the mass changes shows that the relatively high mass losses of the two coat enamels are not partially compensated by the formation of reaction products on the enamel surface in contrast to the one coat enamel.
At the mechanically pre-damaged enamel coats (impacts lead to cracks down to the metal
substrate) corrosion products of iron (rust) could be clearly recognized. The damage of the enamel layers caused a direct corrosion of the metal. The formation of voluminous corrosion products (rust) under the enamel coat led to new enamel chippings during the corrosion period.

A similar damage phenomenon has also been observed at steel-conditioned fish scale damages of the enamel layer. Apparently the corrosion speed could be correlated to the size of the fish scales.
This means that in contrast to a mechanical damage by an impact, fish scales do not cause an extended damage but the corrosion media can penetrate through narrow, interconnected cracks to the metal substrate.
The chemical pre-damage and the following gas corrosion lead to a relatively high mass loss, which causes a partial opening of bubbles and pores. The corrosion continues relatively fast down to the steel substrate if the pores (possibly over bubble chains) reach the ground enamel. This leads to a similar corrosion picture like a mechanical pre-damage by impact on the enamel. The time dependence of the corrosion of the two-coat enamel is listed in Table 6.

The visual observation of the enamel surfaces shows at first (up to 14 days test period) a roughening of the enamel surface, which could be correlated with the gloss and the mass loss.
With increasing corrosion time the roughening of the surface decreases (uniform enamel loss) and various reaction products are formed on the enamel surface, e.g. formation of a silica gel coat (see Fig. 3).

Fig. 3 Schematic depiction of the Böhmer-Hennicke model [7].

Three coat enamel
In order to increase the corrosion resistance of the composite material metal/enamel, the two-coat enamel is provided with a SiO₂-enriched transparent enamel coat. Due to the higher technology and energy requirements this three-coat enamel is mainly used for special purposes only.
Some results for a testing period of 7 days are summarized in Table 7.

Through the weathering no roughening of the enamel surface was noticed. Occasionally small crystals (reaction products) deposit on the enamel surface. The corrosion of the pre-damaged three coat enamel showed a very low mass loss, furthermore small, dark bubbles could be noticed, which do not affect the very good corrosion resistance of the enamel.
An increase of the testing period does not lead to serious changes of the enamel surface and the corrosion process.

Discussions
The investigation of the corrosion behaviour of three technically relevant container enamels towards gas and vaporous components and condensates led to very dissimilar results. The non-damaged one coat enamel showed an acceptable corrosion resistance of the enamel after short testing periods. With increasing testing time a surface damage down to the steel substrate developed. The chemical pre-damage, which can result from a biologically induced corrosion, led to a strong primary damage of the enamel surface. The consequence of this damage is the failure of the enamel coat as a corrosion prevention coat after weathering. An essential reason for the unacceptable behaviour of the one coat enamel in comparison to the outer coats of multi coat enamel is the relatively low content of network stabilizing components.
Two coat enamels show basically better results. The non-damaged two coat enamel is much more resistant against the corrosion media examined. Also after long corrosion periods no corrosion damages could be proved which is due to the high chemical resistance of the cover coat enamel (high content of resistant components). Although the chemical pre-damage of the two-coat enamel causes a clear mass loss, however this does not have any negative consequences on the long time resistance of the protective coat. Only if the chemical treatment opens bubbles and pores of the enamel, the corrosion medium can quickly reach the ground enamel and can cause there starting points for a local, progressive corrosion. The mechanical pre-damage by impact leads to strong corrosion damages (rust effloresces). Already small, mechanical damages with interconnected cracks considerably affect the corrosion resistance of the enamel sheets in a short time. The strong corrosive effect of fish scales can lead to similar causes.
The three coat enamel showed a very high corrosion resistance of non-damaged and chemically pre-damaged enamel sheets. Local corrosion appearances have not been observed on the weathered samples.
During the production and use of enamelled materials for tanks the following primary corrosion causes must be excluded:

The clearest corrosion appearances were caused by impact and tension damages which can arise during transport and assembly of the enamelled sheets as well as during the maintenance, repair and cleaning of the tanks.
The corrosion media penetrate the chippings and enamel cracks in a short time down to the metal substrate and can lead to reaction products, which can infiltrate the enamel and cause the chipping f the still intact enamel layer. These damages can be completely excluded by a professional and appropriate transport or a corresponding assembly and maintenance of the tanks. Fish scales in the enamel lead to the same corrosion appearances as if the enamels are mechanically predamaged by impact or stress cracks. Fish scale damaged enamels show the corrosion damages only after a prolonged weathering since the corrosion medium can reach the steel substrate only through narrow cracks. The corrosion speed increases strongly with the starting of the steel corrosion. Corresponding damages can be minimized or completely avoided by the choice of steel with low fish scale tendency.
Large, superficial pores or bubble chains should absolutely be avoided in the top enamel coat since they can be opened already by a small corrosive or erosive mass loss so that the advance of the corrosion medium into the ground enamel and steel substrate becomes easier. In practice superficial pores and also bubble chains cannot be diagnosed with certainty by the conductivity test (high voltage test). Therefore the formation of pores and bubbles in the cover coat during the enamelling process must be, if possible, excluded.
The chemical pre-damage of the enamel, possible by a bacterially induced corrosion (e.g. formation of sulphuric acid by sulphur bacilli), can also lead to an increased mass loss through the opened pores and bubbles. Thus the corrosion by the gaseous media and condensates can continue relatively fast down to the ground enamel and the steel substrate. In practice the chemical corrosion cannot completely be avoided but the corrosion can be limited by a change of the enamel composition (higher content of resistant components, like SiO₂ and TiO₂) or by an additional SiO₂- enriched coat (three coat enamel). A bacterial pre-corrosion of the enamel surface should be largely excluded. However decisive is the avoidance of large, superficial pores in the enamel. The chemical pre-damage of the enamel is altogether by far not as critical as mechanical pre-damages of the enamel.

Summary
The investigation of corrosion appearances on segmented, enamelled tanks used for the sewage reprocessing required systematic tests for finding possible causes. The influence of gaseous media (mixtures of H₂S, CO₂, O₂ and H₂O), present in sewage purification tanks, is especially important with regard to possible corrosion processes of the composite material metal/enamel.
The corrosion tests were executed on non-damaged and pre-damaged enamelled sheets in testing equipment based on DIN ISO 2733. The samples were characterized after corrosion by determining the changes in mass and gloss, by optical microscopy and infrared spectroscopy as well as by analysis of the leach solutions.
It could be shown that the chemical resistance of the composite material increases with increasing numbers of enamel coats and their chemical resistance. The non-damaged two-coat enamel completely fulfils the technical requirements of a high resistance of the composite material.
Mechanical (impact, fish scales) and chemical (e.g. sulphuric acid) pre-damages of the enamel coats, the latter in connection with large bubbles, lead to a high corrosion risk and have absolutely be avoided for a reliable application of the composite material.

Acknowledgements
This work was founded by the Arbeitsgemeinschaft industrieller Forschungsvereinigungen (AiF),Köln, under the auspices of the Deutsche Email Verband (DEV), Hagen, utilizing resources provided by the Bundesminister für Wirtschaft, Bonn.

[1] R. Drewello, M. Nüssler and R. Weißmann: Werkstoffe und Korrosion, Vol.45 (1994) pp. 122- 124.
[2] B. Rödicker: Mitt. DEV, Vol.49 (2001) pp. 18-22.
[3] B. Radomski, G. H. Frischat and P. Hellmold: Mitt. DEV, Vol. 48 (2000) pp. 93-101.
[4] B. Radomski: Korrosion von Emails durch gas- und dampfförmige Komponenten und Kondensate, Dissertation TU Clausthal 2001.
[5] B. Radomski, G. H. Frischat and P. Hellmold: Mitteil DEV, Vol. 50 (2002) pp. 84-90.
[6] DIN ISO 2733: Gerät für die Prüfung mit sauren und neutralen Flüssigkeiten und ihren Dämpfen, Beuth Verlag GmbH, Berlin 1984.
[7] H. H. Böhmer and H. W. Hennicke: Mitt. VDEfa, Vol. 23 (1975) pp. 42-46.