This invention relates to a steel substrate such as steel wire or steel cord, used for reinforcing elastomer products, such as vehicle tires, hoses and conveyor belts, and to the objects thus reinforced.

It is known that for the purpose of reinforcing elastomer products, steel wire or steel cord must possess a conside- rable corrosion fatigue resistance. Numerous attempts have already been made to meet this need. British patent no. 1.352.761, for example, describes the application onto the steel substrate, of a nickel coating of a relatively high weight of more than 5 g nickel per kg of wire whether or not diffused into the steel surface. A coating of copper or brass, for example, is then applied over this nickel coating to ensure the adhesion with the vulcanized rubber, and then the wire thus coated is hardened by means of cold working. This well-known technique presents the disadvantage that a relatively great amount of nickel must be deposited which increases the production cost. If, in addition, a diffusion process is also being considered for the nickel coating in the steel surface, this will be done on the semifinished product before austenitizing, Which can complicate the produc- tion process.

It is also known that the presence of residual compressive stresses on the surface of a cold-worked steel wire substrate improves its mechanical fatigue resistance. In this connec- tion, we refer to U.S. patent no. ₍ 4.481.996 of the applicant, in which such wire and cord, brass-coated for the purpose of rubber reinforcement, is also described. Apart from the fact that this mechanical fatigue resistance does not unequivo¬ cally correlate with a corrosion fatigue resistance, it was

found that the former is more difficult to realize at very high tensile strengths, for example, above 3500 N/mm ² . The problem of mechanical fatigue and corrosion fatigue, however, has become particularly acute due to the ever increasing use of steel wire with these very high tensile strengths. More¬ over, as will further become clear from the description, it was found that with a standard brass-coated wire, residual compressive stresses alone do not guarantee adequate protec¬ tion against corrosion fatigue.

It is, therefore, an object of the invention to provide a steel substrate for reinforcing elastomers that possesses a composite cold-worked metal coating and that has a satisfac¬ tory mechanical fatigue and corrosion fatigue resistance, by means of which the above-mentioned disadvantages of a rela¬ tively thick nickel base layer and the possible diffusion treatment of this base layer are avoided. Moreover, it is an object of this invention to provide such a steel substrate with a high tensile strength, i.e. above 2800 N/mm ² , and even above 3300 N/mm ² . The substrate can possess a round, square or any other polygonal cross-section.

The invention also aims to provide such a steel substrate - in particular a steel wire - without having to fundamentally alter the manufacturing process (e.g. the total coating weight to be applied and/or the wire drawing process). The main purpose of the coating is to ensure an effective and durable adhesion with the surrounding elastomer. For economic reasons, of course, an endeavour will be made to meet this requirement with as low a coating layer weight as possible. Furthermore, thin coating layers are usually more favorable for adhesion than thicker ones.

The steel substrate that, according to the invention, meets these aims has on its surface a composite, cold-worked metal coating that contains copper, zinc, nickel and iron. The nickel content is highest in the zone under the outermost sur- face layer. It has been found, moreover, that the nickel should not be diffused by means of an explicit thermal diffu¬ sion process in the steel substrate. This is presumably because the iron/nickel alloy layer hereby produced could promote corrosion attack and hinder the realization of an effective degree of coverage by the nickel.

The weight proportion of nickel to the total weight of copper and zinc in the coating must be noticeably higher than in the conventional ternary Cu/Zn/Ni coatings such as described, for example, in British patent application 2.039.580, and Euro¬ pean patent applications 8201 and 9846. The proportion Ni/Cu+Zn must, in fact, lie between 0.50 and 1.

In order, however, to realize not only sufficient mechanical fatigue resistance, but also sufficient corrosion fatigue resistance, the invention shows that by preference, the steel wire surface should at the same time also be under mechanical compressive stress, even for wires with very high tensile strengths, e.g. above 3500 N/mm ² .

It was also found that in the zone mentioned, underneath the surface layer, the iron content on the average should by preference be lower than the nickel content.

Further, an additional metal coating, (e.g. a coating of cobalt), can still be applied on the composite coating layer for the purpose of improving the adhesion to rubbers. By preference, the coating will have a weight of at most 10 g per kg of substrate, and often even a weight of at most 6 g/kg of wire will be sufficient.

As will become apparent later on, it can be stated in summary, that according to the invention a steel substrate, in parti¬ cular a high-carbon steel wire (0.6 to 1 % weight C) with a diameter of between 0.05 mm and 1.20 mm, is produced with a metal coating of the usual thickness in which, however, a high nickel content, combined with residual compressive stresses in the surface probably reduces the tendency of the iron present to corrode. Apart from that, the presence of the nickel probably results in a relatively hard, compact, smooth surface, that holds less residual grease and oxidizes less quickly. In addition, the adhesion to elastomers is not nega¬ tively influenced and, in fact, is even more durable (better adhesion retention) than in the absence of Ni in the coating.

These additional advantages thus make the steel wire very suitable for processing into steel cord, for which resistance against corrosion fatigue and against fretting are extremely important. In order to obtain good rubber penetration, the steel cord - which comprises a number of intertwined steel wires - will preferably contain helicoidal intermediate spaces running between the neighbouring wires.

With reference to the accompanying figures, one embodiment of the invention will now be described in more detail. Moreover, the additional advantages will be clarified.

Figure 1 presents schematically a wire cross-section with the composite coating layer.

Figures 2 and 3 represent graphically the corrosion fatigue values for a number of wires after ageing.

Figure 4 represents graphically the course (the gradient) of the nickel content over the thickness of the coating for a nickel layer that is not diffused in the steel substrate in comparison with a nickel layer that is diffused in it.

For comparison, figures 5, 6 and 7 represent similar gradient profiles for Cu, Zn, Fe and Ni .

Figure 8 refers to the degree of oxidation near the wire sur- face.

Figure 9 illustrates via the carbon gradient the presence of grease remnants' in the surface of the wire.

The surface 4 of the steel wire 1 shown in Figure 1, is covered with a composite metal coating 2 which contains copper, zinc, iron and nickel. The nickel content in this coating is highest in zone 6 under the surface layer 3. On the surface 7 of the coating an additional metal layer 5 can be present ; e.g. a thin layer of cobalt applied by means of plasma sputtering. Apart from this, on the outside surface of the coated steel substrate (with or without cobalt top layer) a protective phosphate layer can also be applied, in accor¬ dance with the specifications given in the European published patent application 257.667. A quantity of from 4 to 30 mg of phosphate ions (PO4) per ² of substrate surface area turned out to be appropriate. The application of a protective layer of oil, (instead of a phosphate layer) - e.g. a naphthenic oil as reported in DE 2.824.173 or a low-sulphur oil or an emulsion of rust inhibitors in oils - is also use¬ ful as additional protection against corrosion and corrosion fatigue. By preference, these oils should be applied during the process of twisting the wires into cord. This oiling is especially useful for wires with high tensile strengths.

Among other things, the phenomenon of corrosion fatigue has to do with the sensitivity of the steel substrate surface to hydrogen diffusion, which leads to the formation of brittle cracks (so-called hydrogen embrittlement) . This hydrogen diffusion is a result of the environmental conditions to which the reinforced rubber products (e.g. pneumatic tires) are subjected during use: mainly humidity and corrosive environments, possibly combined with an increased operating temperature. These operating conditions are simulated by means of specific ageing tests for the purpose of determining the corrosion fatigue behaviour of the steel substrates. In particular, thin brass coatings that cannot ensure complete covering and wire damage (fretting) make the steel substrate sensitive to hydrogen embrittlement and to sudden, unpredic- table wire ruptures. It is indeed at the location of the ruptures or of the inclusions near the surface of the wire that the attack by hydrogen initiates a degradation of the steel and further attacks into the body of the wire, resulting in a weakening of the wire cross section.

Example

In the following example, with the aid of corrosion fatigue tests, the behavior of three series of steel cords having the same structure are compared. The first series (I) related to steel cords with a nickel coating which, according to the invention, is not diffused in the wire surfaces. In the second series (II) the nickel coating was diffused in the steel surface before the brass layer was applied. The third series (III) related to steel cord without a nickel layer applied under the brass. Each steel cord was built up of twelve wire filaments twisted together in one operation: (compact configuration: core 3 x 0.21 + wrapping 9 x 0.19). Before being twisted together, the filaments had a high tensile strength (R ) of between something like 3600 and 3700

N/mm ² for the 0.19 mm diameters and between 3250 and 3350 N/mm ² for the 0.21 mm diameters. In each case, the cable dia¬ meter was 0.83 to 0.04 mm. The lay length was 12.5 mm.

The conventional electrolytic nickel coating (sulphamate bath) with a weight of 2 g per kg wire was applied with the first series (I) after austenitizing and with the second series (II) before austenitizing of the semifinished product (diameter 1.40 mm). The diffusion time (for Ni in the steel surface) during austenitization was 60 seconds at 900°C. Then a brass layer (with 63.5 % Cu) was applied, with a weight of 3 g per kg wire for the first and second series, and of 5 g per kg wire for the third series. The brass was formed by the deposition of successive layers of copper and zinc, followed by the usual ther odiffusion treatment. During this thermodif- fusion treatment a part of the nickel is also diffused in the brass layer, as will become clear from the discussion of Figure 4.

Finally, the coated wires were ^" wet drawn to their final dia¬ meter whereby residual compressive stresses were induced in the surface of the steel. The wires from the respective series were then twisted together to form the cord, thereby introducing additional residual compressive stresses, in accordance with the instructions and explanations of U.S. Patent 4.481.996.

Series by series, wires (b) were separated from a number of cords. Corrosion fatigue tests, in which the wire was repeatedly dipped in distilled water, were then carried out on these wires with a Hunter apparatus (U.S. Patent 2.435.772). Details of this testing method are also given at the bottom of page 4 of EP-A-0.220.766. The results are summa¬ rized in Table 1 and are further graphically illustrated in

Figures 2 and 3 for aged wire samples. By way of comparison, the corrosion fatigue limits (CFL - N/mm ² ) were also recorded for the wires (a) before they were processed into cord, and for the aged wires (c) that were separated from the cord. For ageing, the wires (b) were heated at 150 ^β C for 1 hour.

Table 1

From this table it is clear that wire series I according to the invention produces substantially better CFL-values than series II and III. Moreover, it turns out that the CFL-values increase as a result of the wires being twisted together into cord (see CFL-values of wires "a" in comparison to wires "b"). As will be seen from further test results, this is due mainly to the increase of residual compressive stresses on

the surface of the wire as a result of the twisting opera¬ tion.

In the ordinate of figures 2 and 3, the stress applied by the Hunter apparatus is given for aged cord wire samples, and in the abscissa the number of rotation cycles applied in the testing apparatus is given. Curves 8 and 11 refer to the CFL-values for wires according to the invention, while curves 9 and 12 refer to the wires of series II, and curves 10 and 13 refer to the still lower CFL-values of series III. Here the increased corrosion fatigue resistance is most pronounced for the thinner wrapping (or circumferential) filaments with a diameter of 0:19 mm.

The corrosion fatigue behaviour was also examined for the cords of series I, II and III, embedded in a circulating belt made of a standard rubber composition for vehicle tires. Here the rubber belt, which runs over two rollers set up parallel to one another, was subjected to tensile stress and the belt was previously conditioned in a humid atmosphere at a increased temperature. After 4xl0exp7 revolutions, during which water was constantly being sprayed on the belt, the number of wire ruptures in the cords was checked, along with the remaining adhesion of the rubber to the cords. Table 2 provides an overview of this.

The so-called porosity of the coating layer was also deter¬ mined on wires that had been separated from the naked, non-embedded cord. This porosity value is here determined relatively with respect to a 100 % (*) value for series III and is deduced from a determination of the iron content in the metal coating layer according to the well-known nitric acid test. See also Table 2.

Tabl e 2

These data allow us to conclude that the cords according to the invention rupture least, possess very good adhesion reten¬ tion (APR values) after dynamic loading under humid condi¬ tions, and have a considerably lower porosity. This latter point will, incidentally, be confirmed in the discussion of Figure 5. As is common knowledge, APR values (appearance rating) are determined by means of a visual assessment of the amount of rubber remnants that remain on the cord (degree of rubber coverage) after the peeling off of the rubber coating (in the circulating belt) from the embedded cords. Finally, it was also found that the level of the residual compressive stresses on the wire surface was very high and was comparable for the three series. This level had even risen considerably as a result of the operation of twisting the wires into cord. The presence of a thin coating layer of 5 g/kg (a coating layer that probably also is harder and more compact, due to the addition of Ni) presumably increases the counterstress during the process of twisting the wires together into cord, which increases the level of the residual compressive stresses and, probably also as a result of that, increases the CFL*-values.

As can be seen in figure 10, it is to be expected that during the austenitizing of a nickel-coated semifinished product, an iron/nickel alloy layer 23 will form, and that (possibly) on top of a remnant layer 24 of nickel. However, if the nickel layer 24 is applied after austenitizing, as outlined in figure 11, then no diffusion occurs in the underlying iron. Hence, in this second case a much thicker free almost pure nickel layer 24 remains available on the semifinished product to diffuse into the Cu/Zn layer which is later to be applied than in the case of an underlying Fe/Ni alloy layer 23. The dot-dash lines 27 in figure 10 and 11 suggest the direction of the diffusion for copper, while the dashed lines 26 illus¬ trate the diffusion of iron, and the unbroken lines 28 the diffusion of nickel. The relative differences in length of the arrows suggest herein by approximation (indicatively) the expected difference in intensity (or quantity) of diffusion that occurs during the formation of brass (by diffusion).

It was indeed found that, after the thermodiffusion treatment of the Cu/Zn layer 25 that was applied, and after the further wet drawing and processing into cord, the underlying iron of the Fe/Ni alloy layer 23 (series II, figure 6) was more strongly diffused through the brass to the outer surface 7 of the coating layer than was the nickel. This can be deduced from, among other things, the profiles of the concentration gradients of the metals through the cross-section of the coating layer, as shown in figures 5 and 6, in which the dotted line 17 represents the evolution of the iron content. The curves 16 in figures 5, 6 and 7 refer to the evolution of the copper content, while the dot-dash lines 18 illustrate the evolution of the zinc content. Figure 5 relates to wires with a non-diffused nickel base layer according to the inven¬ tion (series I), while figure 6 refers to filaments from series II and figure 7 to the reference samples according to

series III (without nickel). The evolution of the nickel content values through the cross-sections of the coating layers in series I and II are represented by curves 14 and 15, respectively, in figures 4, 5 and 6. From figure 4 it is clear that in the embodiment according to the invention, the nickel content (curve 14) is at highest level just under the ca. 0.04 micron thick surface layer. This leads to the conclu¬ sion that the more intense nickel diffusion according to the invention suppresses the iron content - and hence the porosi- ty - of the coating layer just underneath its surface. In this connection it is recommended that, previous to the ther- modiffusion treatment for the brass, the layered Ni/Cu/Zn wire coating be somewhat compacted onto the semifinished product, for example by means of a small drawing pass or by rolling, in order to promote a satisfactory degree of coverage (i.e. a covering layer that is evenly spread and that seals, and hence is substantially free of pores).

As a result of the suppression of this iron content, a diminished oxidation tendency of the coating layer surface is also to be expected. Indeed, this is confirmed when the oxygen profile, as represented in figure 8, is examined. Curve 19 refers to the profile of a wire from series II and curve 20 refers to the profile of a wire according to the invention from series I. The chemically more inert nature of nickel (compared to iron) probably contributes to this pheno¬ menon. Static corrosion tests (i.e. the progression of the corrosion potential as a function of time in de ineralized water) confirm this. The corrosion potential stabilizes at a level of -0.42 Volts after 16 minutes for a reference cord (series III), while for a nickel treated cord it stabilizes at -0.35 V after the same period of time.

The lower quantity of grease remnants on the wire surface can be deduced from the lower carbon content at the wire surface. Curve 22 in figure 9 refers to the C profile of a wire accor¬ ding to the invention (series I) and curve 21 to the C profile of a wire from series II. The probable explanation for this is that the relatively high nickel content near the wire surface results in a harder surface, and less of the grease remnant is then pressed into this surface during the drawing of the wire. Very good lubrication is in any case recommended in order to avoid increased wear of the drawing die. Moreover, curve 21 is also valid for the wires of series III, which are not nickel-coated.

In order to achieve satisfactory corrosion fatigue resis- tance, it will always be important - despite the low nickel quantity (e.g. 2 g/kg) - to achieve a degree of coverage as homogeneous and complete as possible (free of oxide remnants). This nickel barrier is most easily realized when diffusion in the iron surface is avoided. Moreover, wire surface damage, which would result in the local removal of the nickel barrier, should be avoided insofar as is possible by endeavouring to achieve a suitable cord geometry. In this connection, spiral wires around the cords are to be avoided as far as possible. Naturally, cord structures that enable good rubber penetration are also to be preferred.

Since adhesion, adhesion retention, corrosion and corrosion fatigue depend not only on the composition of the coating layer on the wires, but also on the composition of the elas- tomers with which these layers are to combine, it will be important to choose elastomers that, so to speak, are not "poisoned" by nickel or that do not decompose under the effect of possible undesirable reactions of the nickel with components of the elastomers. In principle, every steel cord

construction comes under consideration that, for example, can. be utilized in vehicle tires, drive belts and high-pressure hoses. By way of illustration, in the following a non-exhaustive summary is given of the constructions of the type 2x1, 2x2, 2+2, 3x1, 4x1, 5x1, 1+4, 1+6, 2+7, 3x3, 3+6, 3+9, 3x7, 4x4, 3+9+15, 7x7 and 7x19. Wires and cords treated according to the invention can, in particular, be utilized for reinforcing the tread or the carcass of vehicle tires for the purpose of increasing their lifetime as a result of the increased resistance against dynamic stresses in a corrosive or aggressive environment.