NEW DEVELOPMENT OF HOT-ROLLED PRODUCTS FOR DOUBLE-FACE VITREOUS ENAMELLING
Peggy Guinet, Ch. Abeloos, Ph. Harlet
RDCS, Cockerill-Sambre R&D Usinor,Belgium
P. Cantinieaux - CEDP RDT Cockerill-Sambre Usinor, Belgium

Abstract
Pickled hot rolled material has been used for long time in enameling application for boiler (one face enameled parts). Cockerill-Sambre Usinor researched double face enameling for hot rolled steels ranging from soft and formable to high strength. Hydrogen traps, mechanical properties, drawing quality and enamel adherence were evaluated.
Those steel grades were useful for several applications as sanitary market, tanks and silos.

Introduction
Since long now, enamel has been recognized for his excellent surface properties: good appearance, anti-flammability, hardness, scratch resistance, chemical resistance, graffiti-proof, easiness of cleaning. The corrosion resistance at higher temperature is currently exploited on pickled hot rolled steels for boilers.
They present now high yield strength (often higher than 300 MPa), good adherence and a perfect fish scale resistance. Unfortunately they are inappropriate for double face enameling.
The cold rolled steels on the other hand are well-equipped for this purpose.
Indeed the effect of cold rolling is to generate voids either by cracking in the bulk of the cementite (8) or by decohesion along the interface hard precipitate [or inclusion)/ferritic matrix (6). In addition the subsequent annealing provides very good deep drawing properties especially when use of Interstitial-Free steel.
However the benefits on enameling and mechanical properties of the cold rolling decreases with the cold rolled reduction (8,10) and so with the increase of the steel final thickness.
One must also always keep in mind that the supplementary steps (cold rolling, annealing, skin-pass) make the product more expensive.
The purpose of this research at Usinor Cockerill-Sambre is to be able to propose to the customers, hot rolled (low cost) products intended for double face enameling covering a wide range of mechanical properties from soft and formable to high strength steels.

Enamelling Grade Steels
Fish scale
The fish scale is the most troublesome defect in the enamel profession.
Indeed it may appear several weeks after the enameling process, what means for most of the manufacturers, at customer's home.
The origin of this defect comes from the enameling technique and the steel characteristics.
It appears when molecular Hydrogen (combination of two atomic H coming from the steel according to 1/2H₂ ⇔ Hat. [1])locally accumulates at the interface enamel/steel during cooling of the pieces. The pressure generated provokes a jump in the enamel whose shape looks like a fish scale.
Two main sources of atomic H in the steel can be noticed: the enamel itself and the furnace atmosphere (1).
The enamel is made of several compounds among which soda, borax, alumina, and sand which are melted together at high temperature, water quenched, then milled in water, added to clays for wet coating… Most of the steps may potentially enrich enamel with H₂O.
Likewise the dew point (image of the water vapor contents) of the furnace is of great importance in the apparition of the fish scale defect (7).

During firing, the H₂O generates H₂ following the reaction

At firing temperature, atomic H in solution in the steel and H₂ are in equilibrium [1].
The atomic H gives his electron up to the conductive layer 3d of the iron to become a proton.
This explain his high diffusion coefficient (~10-5 cm²/sec at 20 °C, that is to say 1012 times the one of the Carbon (2)).

figure 1 - Solubility of H in steel according to (10). Hydrogen in cc/100g steel versus temperature (°C)

As far as steel is concerned, one can see on figure 1 the strong increase of H solubility in steel with temperature.
The atom fraction c₀ of Hydrogen in equilibrium with Hydrogen gas at pressure P (in atm = 10⁵ Pa) is given by Hirth (4):

with T in K. In addition, the solubility of H in austenite jumps with respect to ferrite. So the steel get charged of atomic H during firing of enamel, all the more so because of the austenite fraction at the firing temperature.
At cooling, atomic H must escape from oversaturated steel and accumulates as H₂ beside the impermeable enamel layer, giving rise to possible fish scaling.

H trapping
To avoid fish scaling, steel producers have developed enameling grade steels with higher H-absorption capacities (6), in other words with H-traps. Another way is to minimize the austenite fraction (and so the Hydrogen uptake) at firing temperature by use of Ti-stabilized Ultra Low Carbon steels (18), Carbon being one of the most important g-gene element.
H-traps in steel may be classified in three categories (9): Attractive Traps, Physical Traps, and Mixed Traps. In Attractive Traps, H atoms are subjected to attractive forces which are of four types: electrical fields (electron vacancy introduced by any defect), stress fields (induced by dislocation, grain boundary…), temperature gradient and "thermodynamical" force. Physical Traps are sites where it is energetically more favorable for H to stay (voids for instance). Mixed Traps present both extreme characters as most of the actual traps.
The reversible or irreversible character of H-traps is related to the probability of an eventual return from a trap to the normal lattice site. It can be seen that it is easier for a H atom to leave the attractive trap than the physical one.
The traps found in literature are numerous, from solute atoms to precipitates, passing through vacancies, grain boundaries, dislocations and voids. Pressouyre gives a classification of some of them according to size (9).

H traps in IF steels
The basic chemistry chosen by Cockerill-Sambre for the current subject is Interstitial Free steel.
For a hot rolled product, one can expect different types of H traps: precipitates (5,4,11) (TiN, Ti₄S₅ (stoechiometry of Titanium Sulfide (10)), Ti₄C₂S₂, TiC, MnS, FeTiP), grain boundaries(25) and possibly substitutional solutes which may provide traps of weaker bonding energy (and probably reversible) by means of stable hydrides formation (4).
Titanium has been determined as a reversible trap (32) and an effect on permeation has been observed (10), but no correlation was made with fish scaling.
Among those traps, precipitates are the most efficient. TiC, TiS, Ti₄C₂S₂ and TiN have been recently deep studied since the transformation reaction of TiS to Ti₄C₂S₂ according to (11)

is beneficial to remove solute carbon at elevated temperatures (about 1100 °C)(14). The precipitation sequence has been speculated to be TiN, TiS,Ti₄C₂S₂, and TiC, as the temperature decreases (11,21).
Consequently, with respect to the chemical composition and the thermomechanical process, TiN is often found in Ti stabilized Ultra Low Carbon steels with the highest size (from few tens of nm to few μm), while TiC is the smallest (few nm to few hundred nm (10)).
Lots of works have put evidence of an irreversible H trapping effect of TiC both in cold rolled and hot rolled steel sheets, especially when fine, uniformly distributed and coherent with matrix(4) (12)(15)(16)(19).
According to Hirth(4), Carbide interfaces are strong traps, so strong that once saturated with H, they should never desorb at room temperature. Such carbides are often obtain by high Slab Reheating Temperature(16)(19).
In cold rolled steel, fish scaling is more effectively prevented by the use of TiN precipitates rather than TiC (16), what is also Cockerill-Sambre's experience.
Knowing that TiN in our commercial ULC steels are of big size, rectangularshaped (NaCl type structure with lattice parameter of 0.424 nm (13)) and hard, we interpret that as follow: cold rolling generates voids of higher trapping effect along TiN incoherent precipitate than along the TiC having undergone annealing.
The effect of Titanium sulfide or carbosulfide on enameling characteristics has been investigated in the cold rolled steel sheets between others at Posco (17), CSM (5) and the University of Gand (18,10) in order to improve the fish scale resistance, and the adherence.
Yoon and Kim (17) working on ULC-Ti+S found that Titanium Sulfide was more effective as hydrogen storage (voids generated along the precipitates) than Titanium Nitride since they are more uniform and of higher average size, what might be expected according to their chemical composition. Van Cauter et Al. (18) showed that the C/S ratio, the SRT and the coiling temperature affect the fraction of Ti₄C₂S₂, that was also found to be able to precipitate during hot rolling and coiling. The size of TiS and Ti₄C₂S₂ were determined by TEM as respectively 240-310 nm and 100-240 nm, with the size of the coprecipitates situated between them.
In this study, steel with higher C/S ratio gave better permeation times. Van Cauter in (10) concludes however that for a cold rolled Ti stabilized ULC steel, TiS is better than Ti₄C₂S₂ itself better than TiC for permeation time.
According to Rizzo et Al. (5) the fish scale sensibility is correlated to the total interface area of the precipitates (TiC, TiS, or even FeTiP generated at the expense of Ti4C2S2 during coiling or batch annealing (21)) rather than their bonding energy with H. However they observe that the bonding energy decreases with the lost of coherency.
They also put evidence of the irreversible character of TiS as H traps.
In the case of IF hot rolled products for enameling, few works have already been done. Yasuda et Al. (19) have investigated the effect of precipitation of Titanium carbides on enameling performances of a hot rolled steel.
This result shows that there is a correlation between Hydrogen diffusivity and the fish scale resistance.
In figure 2, the effect of ageing temperature on apparent diffusivity and Titanium precipitates is highlighted. The drop in Hydrogen diffusivity at about 600 °C has been related to fine coherent and uniformly distributed TiC.
Those precipitates are strong H traps, strong enough to allow the double face enameling of hot rolled products.

figure 2 - Effect of ageing temperature on apparent hydrogen diffusivity and Ti precipitates

Fish scale detection
Several methods are used to measure the H-traps efficiency or the fish scaling resistance (22).
The first types of test are performed on enamel coated sheet specimens.
They are intended to facilitate the fish scaling either by thermal shocks or by use of a sensitive enamel. This one contains no oxide to promote adherence and emphasize the defect in the enameling conditions. The result is binary: good steel or bad steel, no middle.
Consequently, the capacity of the steel to avoid fish scaling is not quantified. In addition this test is sometimes considered as too strict.
At Cockerill-Sambre, for new developments, the results are confirmed by the others enameling methods such a way to get closer to the customer's process: 1 coat/1 firing (corrosion resistant ground coat), 2 coats/1 firing, 2 coats/2 firing.
The second type of test is based on the diffusion of Hydrogen through the steel. They can be classified in three categories:

It must be noted that the Hydrogen desorption (thermally induced) is also analyzed.
According to the method (2,15,24,25,27,28), they open the door to a permeation time, a lag time, an apparent diffusion coefficient (first and second diffusion (26)), a critical dew point, the binding energy of the traps or the concentration of trapped Hydrogen (irreversibly or reversibly) and solute into steel (26).
Ströhlein test is performed at Cockerill-Sambre. It consists in a time measurement of the saturation of the H traps.
H is chemically generated in one side and passes through the steel sheet. According to the steel grade, the inflexion point in the curve of the amount of H detected versus time, appears early or later.

figure 3 - Correlation between fish scale and permeation(22)

The permeation (TH) is determined as

TH = t₀/d₂

where t₀ is the permeation time, and d is the steel thickness.
The European norm (EN 10209) prescribe a permeation of 6.7 min/mm² for a minimal fish scale resistance.

The correlation is rarely observed between the permeation and the fish scale resistance in the Titanium bearing steels (22)(5) while its linear with other cold rolled products (figure 3 and 4).

figure 4 - Absence of correlation between permeation and fish scales in Ti stabilized ULC steels according to (10)

This may be a sign of the presence of different types of traps (reversible and irreversible) whose bonding energy could be function of the temperature.

Adherence
The other well known problem in Ti stabilised ULC steel is the adherence of enamel. The poor adherence is believed to be caused by the presence of TiFe₂O₄ at the surface of the steel, a compound poorly soluble in enamel (20).
To remedy this difficulty, a higher firing temperature is prescribed as well as the use of more alkaline reactive enamel frits, or the use of a flash Nickel (precipitation of pure Ni particles).
Some authors have also found a link between enamel adherence and Titanium excess. According to Yasuda et Al. (16) a Titanium excess of 0.06 wt% is the threshold above which the enamel adherence deteriorates.
The adherence of enamel on steel results, at least partially, from alloy and metallic particles precipitation at the interface (20).
At temperature lower than 600 °C, enamel is still porous to atmosphere. The surface is oxidized in Magnetite (Fe₂O₃) and Hematite (Fe3O4).
Molten enamel appearing at higher temperature is impermeable to atmosphere. Oxides are then dissolved into enamel, where iron is then present as FeII and FeIII and oxygen as O^-2.
In accordance to thermodynamic, the precipitation of FeNi and FeCo and even FeCu alloys as well as metallic particles under the shape of dendrites is observed at the interface.
They cause mechanical adherence by generating a high level of roughness, which is an important (discussed) part of the whole adherence.
The role of the flash Nickel on Ti bearing steel is to achieve the synthesis of a FeNi alloy during the temperature rise, to replace partially the one slowed down at firing temperature by the presence of TiFe₂O₄.
More generally this coating is applied when use of a less reactive enamel (direct-on-white).
In Ti stabilized ULC steels, no solute carbon is available to make the steel more reducing and so to foster the alloys precipitation, what's another reason of the poor adherence often observed.
Sulfur is believed to replace Carbon as reducing agent but is even earlier precipitated in this steel grade.
The importance and the mechanism of chemical adherence are not well known even if some believe it's predominant (5).

Weizhong et Al. (23) have studied the enamel non stoechiometric transition layer (containing FeO_1-x - SiO_2-y - AlO_3-z) in the vicinity of the interface with steel. This transition layer results from the mutual diffusion between steel and enamel. They conclude that the adhesion between the steel and this transition layer is performed by metallic bond while the adhesion between this layer and the enamel is performed by ionic and covalent bond.

Mechanical properties
H trapping in IF steel results from a sufficient amount of (appropriate) Ticompounds. This requirement is in contrast with the need of good mechanical properties (pining effect).
In cold rolled steel the precipitates and solid solution elements have strong influence on the recovery and the recristalization during annealing.
Both the grain size and the texture depends upon the precipitation states of the hot rolled product.
For instance, the excess of Ti is correlated to the grain size (21) and consequently on the yield strength but also to the r-mean value (16), and the C/S level to the rmean value (18).
The problem is simplified for hot rolled steels: the mechanical properties directly derives from the chemical composition (possible solid solution strengthening), the grain size and the amount and the size of precipitates. The last one are inherited from the rolling conditions and the coiling temperature.

Cockerill-Sambre Usinor Groupe's choice
On the basis of a bibliographic study and Cockerill-Sambre experience, we have carried out different hot rolled steel grades suitable for double face enameling based on Extra Low Carbon Ti stabilized steel.
The idea is to generate enough H traps by means of an appropriate precipitation (TiC), and to modify mechanical properties by adapting the process parameters and the level of solid solution elements and precipitates.
Industrial trials were performed with continuously casted, hot rolled and pickled steels.

Drawing quality
Due to the lack of information's on the trapping efficiency of Titanium sulfides and carbosulfides in hot rolled steel sheets, two chemical compositions were investigated (Table 1).

Precipitation simulations were performed at RDCS, using the following solubility products in austenite (29) (21) (3) at equilibrium:
Log10[Ti][N] = -15020/T + 3.82
Log10[Ti][S] = -13975/T + 5.43
Log10[Ti][C]^0.5[S]^0.5 = -17045/T + 7.9
Log10[Ti][C] = -10400/T + 4.79
Log10[Mn][S] = - 9020/T + 2.93
Log10[Ti][P] = - 4956/T + 1.36

One must keep in mind that solubility products are mainly determined in ULC steel, that is to say for lower Carbon and even Titanium levels. Results will thus only be used to discern trends.

figure 5 - Precipitation simulation in DQ steel grade

Results for the drawing quality grade
Table 2 gives the processing route, the mechanical properties at natural state in the rolling direction and the fish scaling results (use of a sensible enamel).One see that both the mechanical properties and the fish scale resistance are better for DQ sample.
A very good adherence was found (level 5) on both samples.

figure 6 - Precipitation simulation in DQ+S steel grade

As known, hot rolled steel sheets have a Lanckford coefficient close to 1.
In spite of this, Cockerill-Sambre has recently deeply investigated deep drawing for the sanitary ware market (30) for thick steel sheets among which hot rolled products.

The optimization of the drawing is based on modeling with finite elements and the use of multizone blankholder force.
An adequate roughness is applied also on the steel sheet, that is to say a roughness with plateaus, compromise between the quality of the lubrication (microreservoir) and the suitability for controlled movement of the metal in the blankholder.
Cockerill-Sambre has acquired a high experience in his customers profession, available to help them.
The precipitation simulation results show significant differences, mainly the TiC, Ti₄C₂S₂ and MnS levels, what was expected.
Keeping always in mind the above comment and that it gives an image of the precipitation at equilibrium, one observes that both FeTiP and Ti₄C₂S2₂ appear at higher temperature than expected (higher than 1250 °C) and that TiS are already dissolved at this temperature (figure 5 and 6).
DQ+S steel showed fish scaling on the whole width of the coil except on 15 cm near the edges. We performed permeation tests (Ströhlein test) at the edge and the axis. In agreement with the literature, we didn't find any correlation between the permeation and the effective fish scale resistance (Table 3). This confirms that the Ströhlein test appears inappropriate for IF steels.

SEM and TEM observations were performed in order to confirm the presence of the expected precipitates.
Both TiN, TiC and Ti(C,S) were found in the DQ (figure 7 and 8) while TiC were not detected in the DQ+S.
TiN, nucleated on Al₂O₃ particles, show a size of few mm, Ti(C,S) have a typical dimension of 1 to few hundred nm and the bigger one are found aligned, while TiC are smaller (around 100 nm or less), and dispersed in the ferritic matrix or on the grain boundaries.
However due to the limited resolution of the microprobe, the SEM used is unable to identify finer precipitates. TEM and precipitates dissolution's are in progress to quantify the precipitates in both steel grades.

figure 7 - SEM observations in DQ

figure 8 - SEM observations in DQ+S

Industrial trial
The DQ steel grade was tested at a sanitary manufacturer. Thanks to Cockerill-Sambre's experience, both shower basis and bathtubs were successfully drawn in the customer's factory.
The enameling needs degreasing, pickling, and a two coats/one firing technique.
Firing temperature was 850 °C. No fish scale was detected, and the adherence was very good. For the first time, a bathtub has been manufactured on a basis of a hot rolled steel sheet.

Discussion
The hypothesis of TiC as main H trap in the hot rolled steel grade followed by Cockerill-Sambre, based on a bibliographic study, has not yet been undeniably proved.
However we found several indications converging in this direction. New analysis (TEM, dissolution's) are in progress to check the reality of what has already been found and deduced.
The important result that must be kept in mind is that a hot rolled sheet steel has been used to achieve a bathtub. It means that this hot rolled steel sheet was soft enough to undergo deep drawing, and suitable for double face enameling.
The very good adherence found on all hot rolled grades, contrary to cold rolled steel sheets, could be related to the fact that Titanium oxides are removed from the steel during the pickling.

High strength steel
In parallele to the development of a soft grade steel, Cockerill-Sambre has worked on hot rolled steel designed for applications requiring a high level of mechanical properties.
The idea is to increase the yield strength by solid solution strengthening orwith other precipitates, and to avoid grain coarsening during enamel firing by pining effect of the precipitates or the solutes atoms. Two chemical compositions are also investigated to reach different yield strength levels.
The strengthening may be carried out by means of Mn, P, Cr, Nb, Si. Again the expected H traps are TiC but since the steel is most charged, other precipitates are believed to act as H traps.
It's to be observed that due to the carbon and Titanium levels, the steel grade HS+ is not Interstitial Free steel. It's confirmed by the presence of a yield point elongation. This Carbon level was chosen to both generate lots of carbides, and allow solid solution strengthening.

Precipitation simulations were also performed on these steel grades (figure 9 and 10). Again the same comments as for the drawing qualities are valid.

figure 9 - Precipitation in HS steel

figure 10 - Precipitation in HS+ steel

As Expected, TiC precipitation is higher for the HS+ grade, and C remains in solution. Again, FeTiP precipitation level is quite high, (higher than the one of TiN and Ti₄C₂S₂).
However the consequences on both mechanical properties and fish scale resistance are not investigated in this paper.

Results for the high strength steel grade
Table 5 gives the processing route, the mechanical properties at natural state transversal to the rolloing direction and the fish scaling results (use of a sensible enamel). Two levels of yield strength are reached with the different grades. Both show good enameling properties, with the absence of fish scales and again a very good adherence.
The surface treatment is the following: degreasing, flash Nickel, neutralization by means of Borax, and adapted rinses. The enameling type is a 2 coats/2 firing liquid/liquid, where the ground coat enamel used is a corrosion resistant and the cover coat may be corrosion resistant or for aesthetic purpose.
These steel grades may be used in several application, from water or chemical tanks to signpost, passing through purification stations, silos, architectural applications.
SEM observations were performed on the HS+ grade (figure 11). TiN, TiC but also Ti(C,S) were observed, with comparable sizes to those of the drawing qualities. Here again more measurements are necessary to conclude without doubts on the important role of carbide precipitation.

figure 11 - SEM observation in HS+

Discussion
The high mechanical properties derives mainly from one part (HS+ grade) from the Carbon solid solution, the precipitates and the grain size, and on the other part (HS grade) from the Manganese solid solution and the grain size. This one is determined by the level of precipitates showing pinning effect already in the hot rolling mill, and the processing route.
Those precipitates play the same pining effect during the enamel firing, allowing the grain size to remain stable.
Expected Titanium Carbides are observed but not yet quantified. Here again measurements are in progress.

Conclusion
Hot rolled steel for double face enameling based on Extra Low Carbon Tistabilized and Low Carbon chemistry are investigated at Cockerill-Sambre Usinor. The objective is to cover a wide range of mechanical properties from soft and formable to high strength steels.
Those steel grades are useful for several different application: the sanitary ware market, tanks and silos, architecture.
In the absence of voids generated during cold rolling, the expected main Hydrogen trap is TiC.
The observed enamel adherence is always excellent even without use of an appropriate surface treatment.

The mechanical properties of the drawing quality are guaranteed by the Interstitial-Free character of the steel.
For the deep drawing applications (bathtubs), an appropriate roughness is applied on the coil.
For the applications requiring higher mechanical properties, the yield strength is obtained by solid solution strengthening and grain size control.
As well the effect of increasing Carbon contents is currently evaluated at Cockerill-Sambre Usinor.

Acknowledgement
The authors are grateful for Mr. Aronica from Ferro-France for the help provided during this work.

References
1. Gordon P. K. Chu, Silver Jubilee Conference Paper, The Institute of Vitreous Enamellers Bulletin, Vol. 10, N°5, 1960, pp. 235-251.
2. Mathy H., CRM, 1989.
3. Sieverts A., Zapfe G., Moritz H., Z. Phys. Chem., A183, 19 (1938/39).
4. Hirth John P., Met. Trans. A., vol. 11A, 1980, pp. 861-890.
5. Rizzo L., Fernandez Martinez B., Dürr W., ECSC Sponsored Research Project, Draft Final Report, 1996.
6. Hemmen P., Cholet V., Guillot L., La revue de métallurgie-CIT, janv. 1995, pp.131-138.
7. D'Harcour J.-C., Email Métal, N°11, 1972, pp.7-13.
8. Warnecke W., Giesel P., Schrape U., Mitteilungen des vereins Deutscher Emailfachleute E. V., 31, heft 10, 1983, pp. 145-160.
9. Pressouyre G. M., Met. Trans. A., vol 10A, 1979, pp. 1571-1573.
10. Van Couter A., "Technologisch basisonderzoek van de emailhechting op Ti-gestabiliseerde interstitiëlen-vrije stalen", Universiteit Ghent, Faculteit Toegepaste Wetenschappen, 1998.
11. Hua M., Garcia C.I., De Ardo A.J.,Met. and Mat. Trans. A, vol. 28A, 1997, pp. 1769-1780.
12. Akira Y., Kenji I., Minoru N., Junko H., Junsuke T., Trans. ISIJ, vol. 23, 1983, p.394.
13. Inoue K., Ohnuma, I., Ohtani H., Ishida K., Nishizawa T., ISIJ International, vol. 38, 1998, 9, pp. 991-997.
14. Yang X., Vanderschuren D., Dilewijns J., Standaert C., Houbaert Y., ISIJ International, vol. 36, 1996, 10, pp. 1286-1294.
15. Stevens M. F., Bernstein I. M., Met. Trans. A, vol. 20A, 1989, pp. 909-919.
16. Yasuda A, Ito K., Ikehira J., Irie T., Takasaki J., Furukawa Y., Kawasaki steel technical report, 12, 1985, pp. 45-54.
17. Yoon J.-B., Kim S.-J., 38th MWSP Conf. Proc., ISS, vol. XXXIV, 1997, pp. 425-430. 18. Van Cauter A., Dilewijns J., Hörzenberger F., De Cooman B. C., 39th MWSP Conf. Proc., ISS, vol. XXXV, 1998, pp. 315-323.
19. Yasuda A., Ito K., Matsumoto Y., Nishida M., Takahashi I., The Vitreous Enameller, 4, 1983, pp. 79-94.
20. Million E., "Etude des mécanismes d'adhérence des émaux sur l'acier", Thèse de doctorat, Université de Nancy I, 1987.
21. Copreaux J., Gaye H., Henry J., Lanteri S., "Relation précipitation-propriétés dans les aciers sans interstitiels recuits en continu", Rapport Commission Européenne, recherche technique acier, 1997.
22. Harlet Ph., Beco F., Renard L., Dawance J, The Vitreous Enameller, vol. 40, 1990, N°1, pp.10-21.
23. Weizhong C., He Hai Y., XVI International enamellers' congress, England, 1992, pp. 1F-3F.
24. Garet M., Brass A. M., Haut C., Guttierez-Solana F., Corrosion Science, vol.
40, N°7, 1998, pp. 1073-1086.
25. Ono K., Meshii M., Acta metall. Mater., vol. 40, N°6, 1992, pp. 1357-1364.
26. De Gregorio P., Valentini R., Solina A., Gastaldo F., Buonpane A., XVI International enamellers' congress, England, 1992,, pp. 1F-6F.
27. Luppo M. I., Ovejero-Garcia J., Material Characterization, 40, 1998, pp. 189-196.
28. Valentini R., Solina A., Paganini L, Journal of Materials Science, 27, 1992, pp. 6583-6589.
29. Yang X., Vanderschueren D., Dilewijns J., Standaert C., Houbaert Y., ISIJ International, vol. 36,N° 10, 1996, pp. 1286-1294.
30. De Bal J.B., Dauby B., Kergen R., Harlet P., 18th International Enamellers Congress, 1998, pp. 137-149.
32. Hamart O., Lung T., Lanteri S., Proceedings of the 40th MWSP Conference, 1998, pp. 189-194.
33. Pressouyre G. M., Bernstein I. M., Corrosion Science, Vol. 10, 1978, pp. 919-933.