ADHERENCE OF VITREOUS ENAMEL ON CAST IRON- A NEVER ENDING STORY
Dr. Jörg C. Wendel
Wendel Email, Postfach 1951, D-35668 Dillenburg, Germany

Introduction
The adherence of vitreous enamel on its substrate is the most basic need of any product. There is no other item a producer can loose his good name in a faster way, than with enamel chipping on his products.
The adherence of vitreous enamel on cast iron is caused by a good adhesive intermediate oxide layer [1, 2] and on mechanical interlocking of the materials [3, 4]. This can be enhanced by the use of electrolytic corrosion due to the influence of adhesive oxides [5] or with a good preparation of the part by the means of shot blasting. The formation of tinder on cast iron is important and its growth must be controlled [3, 6].

The nature of the oxide layer
To understand the formation of the oxide layer on cast iron during the firing procedure we have to view the phase diagram of Iron and Iron oxides depending of the oxygen pressure (Fig. 1).
Starting at a temperature of approximately 560°C the iron starts to form oxides. Depending on the amount of oxygen it can form three different oxides: Wüstite (Fe_1-xO), Magnetite (Fe₃O₄) or Hematite (Fe₂O₃). The β-Iron nowadays is seldom named as it has still the same body centred cubic structure of α-Iron but due to the fact of exceeding the Curie temperature (768 °C) the magnetic properties are changing. Concerning the adherence of the oxide layer to the cast iron it is very important which one of the oxides is formed. We know, that Hematite is reducing the adherence very much and Magnetite is increasing the adherence. The Wüstite is mainly a transition phases and will later, when cooling down the part, decompose to Magnetite and α-Iron (Eq. 1)

4 FeO = Fe₃O₄+ Fe (1)

The free reaction enthalpy of the reverse reaction is shown in Figure 2.

We also know that the thickness of the iron oxide layer is important. Well in the case the layer is very thick we will have mostly a mixture of Magnetite; Hematite and some Wüstite. As the thermal expansion rate of these oxides is different the layer will from cracks during the cooling procedure, which will decrease the adherence on the cast iron.

Fig. 1 Phase diagram Fe- Fe2O3, calculated with data from [7, 8].

Fig. 2 Free reaction enthalpies for the most important reactions of the iron.

There are mainly five reactions determining the result of the oxidation reaction. The results of the calculation are shown in the Fig. 2. All thermodynamically data have been calculated using data from Barin, Knacke, Kubaschewski [9]. The first reaction which is the formation of Wüstite (Eq. 2) is very fast

(1-x) Fe + 1/2 O₂ = Fe_1-xO (2)

The next two reactions (Eq. 3, Eq. 4) are concurring. The differences in enthalpy between Wüstite and FeO have been respected, but for the notation it is easier to work with the FeO. Regarding these concurring reactions the temperature range between 880 °C and 980 °C is good for the formation of Magnetite

3 FeO + ½ O2 = Fe₃O₄ (3)
2 FeO + ½ O2 = Fe₂O₃ (4)

The worst reaction would be the reaction of Magnetite to Hematite (Eq. 5) which is thermodynamically unfavourable above 700 °C

2 Fe₂O₄+ ½ O₂ = 3 Fe₂O₃ (5)

The reaction of Wüstite with Hematite forming Magnetite (Eq. 6) is very helpful, but unfortunately very slow

FeO + Fe₂O₃= Fe₃O₄ (6)

Mechanical interlocking
The first aim is to control the oxygen diffusion in order to control the formation of Magnetite. The second aim is to increase the mechanical interlocking of vitreous enamel and cast iron. This can be done mainly by two means. The first one and most important is the preparation of the part with shot blasting. We will focus on that point later. The other possibility is the electrolytic corrosion of the iron with adhesive oxides like CoO (Eq. 7)

CoO + Fe = FeO + Co (7)

and NiO. But regarding the fact that we have in cast iron enamelling always the formation of a lot of Wüstite, this Wüstite is able to react with the adhesive oxide before this could do it’s reaction with the Iron surface (Eq. 8, Eq. 9) [5, 10]

2 FeO + CoO = Fe₂O₃+ Co (8)
3 FeO + CoO = Fe₃O₄+ Co (9)

As it is shown in the figure (Fig. 3) the reaction of Wüstite with the adhesive oxide to Magnetite (Eq. 9) is exceeding 600 °C much more favourable than the electrolytic corrosion (Eq. 7). This is the reason that in cast iron enamelling adhesive oxides doesn’t have the importance as in sheet steel business. Each adhesive oxide reacting with Wüstite somewhere in the vitreous enamel layer is lost for it’s electrolytic purpose so the efficiency of this solution is going down.

Fig. 3 Free reaction enthalpy as a function of the temperature for common adhesive oxide reactions.

The influence of the cast iron composition
The components of the cast iron are having a significant influence on the adherence. Phosphorous and Silicon increase the adherence of vitreous enamel. This can be explained either by the carbon equivalent (Eq. 10)

C_eq. = C_total + (Si + P)/3 (10)

or by the good chemical relationship of Silicon and Phosphor to the chemical nature of the glass.
Sulphur and Manganese decrease the adherence of the vitreous enamel. Sulphur is supporting on the one hand the oxidation of the iron and is on the other hand forming sulphides with a very low surface tension, which decrease the wetting of the glass. Manganese is only used to combine the residual sulphur in the cast iron [11, 12]. In general it is sufficient to have approximately three times the percentage of Sulphur as Manganese content in the cast iron. Any excess of Manganese is stabilising carbides (combined carbon), which influences the adherence negatively. The content of Silicon and Phosphor are also determining the ferritic structure of the iron. On perlite the adherence is always bad (combined carbon decreases). The structure of the graphite is also important. A-Graphite is good and anything less than B-Graphite decreases the adherence very much as the ratio of surface and mass is going up. The tiny structures of graphite like D and E have less crystalline character and amorphous graphite is very reactive and burns of easily. It must be respected that the cooling rate of the iron influences very much the final structure. This is for example important for rim parts of bathtubs. If the iron is cooling down very fast in this area more perlite and some cementite will be formed which decreases the thermal expansion rate and so finally causes tensions, which will decrease the adherence of the vitreous enamel.

The influence of the furnace atmosphere
Fig. 4 is showing that the formation of iron scale (increase of weight) is supported by a higher concentration of oxygen and much more by the presence of water vapour. Water increases the oxidation of cast iron approximately ten times compared to dry air. Therefore the humidity in the furnace atmosphere should not exceed too much because as we have seen in the phase diagram (Fig. 1) there is anyway too much oxygen and too much scaling [13]. This explains the better adherence using electrically and indirectly heated furnaces instead of directly fired gas furnaces.

Fig. 4 Influence of the furnace atmosphere on the growth of iron scale. Data [14].

The spread ability of vitreous enamel on a metallic substrate
In general it is clear the better the spreading of a liquid on a substrate is the more even the coat will be – tiny defects will be bridged and a thin layer will be sufficient to cover all the surface. If the relation between specific surface energy and the viscosity of the liquid is unfavourable the liquid will contract and leave the substrate surface uncovered. To cover the surface we will need a very thick layer of liquid. This will find its indicator in the contact angle of the liquid on the substrate.
We know a liquid may spread on a surface if there is a chemical relationship between the solidstate substrate and the liquid. For example [15]:
Sodium disilicate (a glass model substance) on iron has shown a contact angle of 55° - the adherence is very bad.

If we saturate the glass with iron oxide the contact angle of this glass on iron will be 22° and the adherence is better.
For sodium disilicate on Magnetite we find the contact angle to be 2° - the adherence is excellent.
Now vice versa we will look for the relationship of the smallest stable liquid film and the contact angle [16].
For a contact angle of 45° we will need a liquid film of 2,48 mm thickness. If the contact angle will be 10°, we would still need 560 μm thicknesses for our liquid. In the case the contact angle is 2° the required thickness for a perfect surface coat is 110 μm.
This is precisely the situation we will find if we manage to form the iron oxides on the surface of our cast iron to Magnetite.

Shot blasting
In order to be able to build up the surface according to our needs it is necessary to have it clean in the beginning. This job is done by shot blasting. To arrive to a good shot blasting we have to take care about four things:
1) We must aim the blast stream exactly right. If we miss the target by 10 % we loose 25 % efficiency [17].
2) We have to choose the right operational mixture of the shots. Therefore we need the right size of the shots: If the shots are too big we will have too less impacts – if the shots are too small they will have too less kinetic energy to do their job [18]. The used shots should be separated when they have less than 25 % of their nominal size [19]. Control the angular shape of the shots – otherwise there will be no etching and the cleaning will only reach the top of the surface [20].
3) We must control the correct flow of the abrasive. If there is no steady state feeding and too less load the cleaning will be poor [21].
4) Take care about the dust extraction. It was detected that a residual as 2 % sand will result in the double wear of the blast wheel components [18].

Fig. 5 An example for a bad working operational mix sieve line embedded in ideal mixture curves (G 12 and G 14 Data [22]).

The figure (Fig. 5) shows the master guidelines where a correct operational mixture should be. In this operational mix we find a too fast wear of the shots resulting in too much fine particles, which haven’t been removed. The result will be an extremely rise of scrap. It is clearly shown that the necessary separation is not done. This shot blasting will be ineffective and result in a high scrap rate.

Overview of the known influences
The next figure is giving us the overview of the most influences on the adherence of vitreous enamel on cast iron.
These are regarding the cast iron: the components of the cast iron (Phosphorous and Silicon increase, Sulphur and Manganese decrease), the ferritic structure of the iron (on perlite the adherence is always bad; combined carbon decreases) and the structure of the graphite (AGraphite is good, anything less than B-Graphite decreases very much).

Fig. 6 Overview of the known influences on the adherence of vitreous enamel on cast iron.

On the side of the vitreous enamel it is the viscosity of the frit: If the frit is too soft there is a hazard of over burning, if the frit is too hard it will deny too react. The spread ability of the frit to the cast iron, the rheology of the slurry and the system used.
For the furnace atmosphere it will be the water content and the oxygen content or the carbon monoxide content respectively. Very important are the firing conditions such as the used temperature and firing time.
Regarding the shot blasting we should watch the aim of the blast stream, the correct operational work mixture of the shots, the abrasive flow and finally the effective dust extraction.
In the field of tensions we must balance the thermal expansion rate of the vitreous enamel to that of the cast iron. The final coat thickness of the vitreous enamel influences the tension very much.
And finally we should not forget that difference in the transition of perlite to ferrite (kinetics) or the need of a preheating of the cast iron also influence the thermal expansion rate of there cast iron and therefore the tensions rate very much.

Practical application

Fig. 7 Magnification 500x of ground coat without (left) and with (right) Magnetite layer.

In practical application the existence of the Magnetite layer increases the adherence notable. For the above (Fig. 7) shown ground coats the difference is approximately 2 [MPa]. To determine the magnitude of the adherence. It is used an pull-off system to find out at which force a dolly glued onto the surface will pull off the enamel from it’s substrate. The Figure below (Fig. 8) shows the increase of adherence of an armature ground coat.

Fig. 8 Example for the increase of adherence using the Magnetite layer.

Evaluation
The achieved values for the adherence of vitreous enamel on cast iron are in the range of 0 to 12 MPa (1 [Mpa] = 1 [N/ mm²] = 10.2 [kp/cm²]). Hauttmann [3] suggested a linear evaluation to name the quality of the adherence. In this work a logarithmic evaluation is suggested as practically the quality raises faster with more adherence and very high values don’t change too much any more ifthe adherence is already good.

Fig. 9 Evaluation of the adherence: red [3] = linear, blue = logarithmic.

References
[1] A. Dietzel: Reaktionen und Haftung zwischen Glas und Metall beim Verschmelzen. Glastechn. Ber. Vol. 24 (1951), p. 263.
[2] A. Dietzel: Über den derzeitigen Stand des Haftproblems. Mitt. VDEfa Vol. 11 (1963), p. 71.
[3] A. Hauttmann: Vergleichsversuche mit dem Schlagprüfgerät und nach dem Doppel-T-Verfahren zur Beurteilung der Emailhaftung auf Gußeisen. Mitt. VDEfa Vol. 11 (1963), p. 55.
[4] Christochowitz, Gesell: Einfluß der Strahlmittel auf die Haftfähigkeit von Emaillierguß. Mitt. VDEfa Vol. 2 (1954), p. 74.
[5] A. Dietzel: Theorie der Haftung von Grundemail an Stahlblech. Mitt. VDEfa Vol. 27 (1979), p. 6.
[6] H. Moehring: Ein Beitrag zur Frage der Haftung von Email auf Gußeisen. Mitt. VDEfa Vol. 12 (1964), p. 67.
[7] L.S. Darken, R.W. Gurry: The system iron-oxygen, I. The Wustite field and related equilibria. J. Amer. Chem. Soc. Vol. 67 (1945), p. 1398.
[8] L.S. Darken, R.W. Gurry: The system iron-oxygen, II Equilibrium and thermodynamics of liquid oxide and other phases. J. Am. Chem. Soc., Vol. 68 (1946), p. 798.
[9] I. Barin, O. Knacke, O. Kubaschewski: Thermochemical properties of inorganic substances. Springer Verlag 1977 Berlin.
[10] M. Ghodsi, R. Derie, J.P. Prossnitz: Beitrag zur Theorie der Haftung oxidischer Gläser auf Metallen mit Hilfe der Elektronenmikrosonde. Mitt. VDEfa Vol. 26 (1978), p. 167 + p. 171. [11] D. Ernst: Emaillage des Fontes: Pecautiones pour assurer une bonne adherence. Ctif (1990), FO 161.
[12] A. Dietzel, E. Wegner: Emaillierfähiges Gußeisen. Mitt. VDEfa Vol. 6 (1958), p. 11.
[13] G.H. Spencer-Strong: Effect of furnace gases on physical properties of wet-process cast iron enamels. J. Am. Ceram. Soc. Vol. 22 (1939), p.225.
[14] D. Kamran: Über das Haften von Fritte- und Schmelzgrund an Gußeisen. Mitt. VDEfa Vol. 6 (1958), p. 85.
[15] R.W. Cline, R.M. Fulrath, J.A. Pask: J. Am. Ceram. Soc. Vol. 44 (1961), p. 423 + p. 430 + Vol. 45 (1962), p. 592.
[16] F. Rickmann: Mitt. VDEfa Vol. 6 (1958), p. 55.
[17] Ervin Amasteel: The Ervin Poster: Challenge No.1 Aim –Keeping your aim on target. Tipton 2003.
[18] Ervin Amasteel: The Ervin Poster: Challenge No.2 Controlling the work –mix size. Tipton 2003.
[19] P. Plantin: Guide for surface treatment before enamelling -Blasting. APEV 2003. [20] R. Schmalenbach: Grauguß, Shäroguß, Vakuumformguß – neuere Erkenntnisse der
Gußeisen-Emaillierung. VDEfa Vol. 30 (1982), p. 2 und p. 13.
[21] Ervin Amasteel: The Ervin Poster: Challenge No. ´3 Don’t short-change your blast wheel. Tipton 2003.
[22] A. Schmitz: Privatmitteilung. 2003.