In a hot dip galvanising process, steel articles are dipped in a bath of molten zinc or zinc alloy. The molten zinc reacts with the steel and, through a process of diffusion, zinc-iron components are formed at the steel-zinc interface. To facilitate this reaction, the steel surface is fluxed before it is dipped in the zinc melt. This fluxing operation cleans the surface to ensure the complete coverage of the steel by a layer of zinc. The most common fluxing practice comprises a short immersion of the article in a diluted zinc chloride and ammonium chloride solution. After fluxing, the steel surface is thoroughly dried, as any water that would remain on the steel surface would evaporate explosively upon contacting the zinc melt.

In the prior art, many different flux compositions were proposed for a multitude of purposes, e. g.: fluxes with added wetting agents to obtain a more uniform flux layer on the steel, special fluxes for galvanising using zinc alloys with high aluminium concentrations, fluxes that generates less fumes, fluxes that form less ashes and fluxes that are easier to dry.

During the process of submerging an article in a molten zinc bath, the differential thermal expansion between the part already submerged and the part that is not yet submerged, creates stress in the articles being coated. This thermally induced stress, combined with any residual stress in the article, can become so high that permanent deformation and cracking occurs. This is especially true in articles with a significant residual stress, such as articles comprising sharp bends or holes, or which are made up of several welded parts. To guarantee the integrity of articles, it is in fact useful to limit the thermal stress as much as possible.

The thermal stress problem is moreover exacerbated when galvanising with recently developed highly wetting alloys. Such alloys, typically containing tin or bismuth, have been introduced to avoid so-called "black spots", i. e. small areas on the steel surface that remain uncovered after the galvanising process. These alloys are particularly useful, as they tend to stabilise the thickness of the galvanised layer for a wide range of steel types. However, their good wetting capability increases the heat transfer, which leads to accelerated warm-up of articles being dipped and to increased thermal stress as a direct consequence.

Up to now, in order to alleviate above problems, several recommendations were made to the galvanisers, including increasing the speed of immersion in the zinc melt, changing the angle of dipping, or even modifying the design of the structural elements to be coated. These techniques introduce new constraints to the process or to the articles, and tend to degrade the productivity.

It has been surprisingly been found that the thermal stress problem can be alleviated by using the invented fluxes, which provide for a significant reduction of the heat transfer between the steel and the zinc melt at the initial stage of immersing the steel articles in molten zinc.

The invention concerns a water based flux for hot-dip batch galvanising of steel articles, comprising, in total, 200 to 600 g/1 of ZnCl2 and NH4C1, with a NH4C1 to ZnClz molar ratio of 1.7 to 3.3, characterised in that the flux comprises 8 to 80 g/1 A1C13. The preferred A1C13 concentration is 10 to 50 g/l ; a still more preferred range is 10 to 25 g/1. A suitable flux can be prepared by using 250 to 600 g/1 of the double-salt ZnCl2. 2NH4Cl, or by using 200 to 500 g/l of the triple-salt ZnCl2. 3NH4Cl. It is preferred to limit the Fe concentration of the flux to less than 15 g/1, and even more preferably to less than 10 g/l.

The invention also concerns a galvanisation process, comprising the steps of: fluxing a steel article utilising the above-mentioned fluxes; drying the article; and immersing the article in a molten

bath of a zinc alloy comprising, by weight, 0.1 to less than 5 % of either one or both of Bi and Sn. The zinc bath advantageously comprises, by weight, 0.5 to 5% of either one or both of Sn and Bi, 0 to saturation of Pb, 0.025 to 0. 200% of at least one of V, Ni, Cr or Mn, 0 to 0. 05% of at least one of Al, Ca and Mg, the remainder being zinc and unavoidable impurities.

The upper limit of the A1C13 content of the flux is dictated by the increased viscosity of the flux. Too high a viscosity will indeed impair the replacement of the flux on the surface of the article by zinc alloy from the melt. The AlCl3 concentration in the flux should therefore preferably be limited to 50 g/1, or even to 25 g/1.

To reach the required amounts of ZnCl2 and NH4Cl, it is particularly practical to use commercially available double-salts, being ZnCl2. 2NH4Cl, or triple-salt, being ZnCl2. 3NH4Cl. This will automatically ensure that a suitable NH4C1 to ZnCl2 molar ratio of respectively 2 and 3 is obtained. The experiments show that with mono-salts, being ZnCl2. NH4Cl, or quadruple-salts, being ZnCl2. 4NH4Cl, the heat transfer between the steel and the zinc melt becomes unsuitably high, even with proper additions of Aids. The flux may further contain well known surfactants that are typically added to improve the wetting of the steel articles in the flux tank and to enhance drainage of excess flux solution when the steel articles are extracted from the flux tank. Typically, a surface tension of less than 40 dynes/cm is recommended. It is common practice in general galvanising to check the surface tension on a regular basis, and to add surfactants as needed.

It is also recommended to limit Fe in the flux to maximum 15 g/1. It was indeed observed that Fe increases the heat transfer between the steel and the zinc melt, which is clearly undesired.

It should furthermore be noted that with'0 to saturation of Pb'is meant a concentration of 1.2 wt. % Pb at most. With'zinc and unavoidable impurities'is meant zinc with a purity according to the galvanising standard EN ISO 1461.

As explained above, the use of zinc alloys containing wetting promoters such as Sn or Bi results in accelerated warm-up of the articles upon dipping. The invented flux is particularly well suited for being combined with such alloys, as it greatly alleviates the high risk of permanent damage to most articles.

The influence of flux and alloy on the thermal stress is evaluated by measuring the heating rate inside a fluxed steel article upon its immersion into a zinc alloy bath. To standardise this measurement, use is made of a probe consisting of a small hollow steel chamber with a wall thickness of 1.3 mm, equipped with a thermocouple brazed against the inner surface of a wall. Upon immersion in the molten alloy at a descent rate of 2.5 cm/s, the temperature rise is recorded and the rate of heating is calculated (°C/s). The descent rate is sufficiently high to ensure that it is nor a critical nor a limiting parameter. The highest rate of heating reached during the immersion phase is reported. It is assumed that this rate is closely related to the maximal thermal stress endured during dipping and to the ensuing damages.

Example 1 Steel strips with a thickness of 13 mm are sharply bent at an angle of 160 degrees. This induces residual stress in the articles, which therefore will be more sensitive to the additional stress endured during dipping.

No significant deformation or cracking was evidenced, as long as the chosen combination of flux and alloy resulted in a maximum rate of heating, as measured with the above probe, of 200 °C/s or less. The same articles, processed in circumstances generating a heating rate, as measured with the above probe, of 260 °C/s, did show limited stress degradation such as small cracks along the bending line. Using conditions leading to a heating rate of 300 °C/s or more resulted in some catastrophic failures, such as the fracture of articles.

It should be noted that large articles could be even more sensitive to thermal stress, in particular when residual stress is also present. Small, unstressed articles would of course be less prone to

develop defects. In fact, any significant reduction of the heating rate is useful to better guarantee the integrity of the coated articles. A 10% reduction is deemed significant.

Example 2 In Experiment 1, the above-described probe is fluxed in a 500 g/1 aqueous solution of Florflux@, a commercial product manufactured by La Floridienne of Belgium which contains a ZnCl2. 2NH4Cl double-salt.

The probe is dried, and subsequently dipped at an immersion rate of 2.5 cm/s in a TechnigalvaS zinc alloy melt heated at 450 °C. This alloy consists of zinc with 1 wt% Pb, 0.05 wt% Ni and 0.004 wt% Al.

It thus contains no wetting promoters such as Sn or Bi.

The process according to Experiment 1 is repeated in Experiments 2 to 9, using various amounts of AlCl3 added to the commercial flux. The results are reported in Table 1.

Table 1: Maximum heating rate as a function of AlCl3 in a 500 g/1 double-salt flux, using a conventional zinc alloy bath No AlCl3 Heating rate (g/1) (°C/s) 1* 0 175 2* 172 3 8 155 4 20 142 5 40 140 6 60 135 7 70 128 8 80 140 9* 100 172 * : counter-experiment From this table and from other measurements, it appears that a significant effect of Aids on the heating rate is achieved from about 8 g/l up to about 80 g/l Aids.

Example 3 In the following series of experiments, a typical highly wetting alloy is used, namely GalvecoS, which consists of Zn with 1.1 wt% Sn, 1 wt% Pb, 0.075 wt% Bi, 0.05 wt% Ni and 0.004 wt% Al.

The influence of the composition of the flux upon the maximum rate of heating is measured according to the above-described standardised method. In Experiment 10, the same commercial flux was used as in Experiment 1. In Experiments 11 to 19, various amounts of AlCl3 are again added. The results are reported in Table 2.

Table 2: Maximum rate of heating as a function of AlCl3 in the flux, using a zinc alloy bath with wetting promoters No Aids Heating rate (g/1) (°C/s) 10* 0 345 1* 275 12 8 220 13 20 180 14 40 148 15 60 190 16 70 153 17 80 192 18* 100 265 *: counter-experiment From Experiment 10, which is a counter-example, it can be seen that using a known commercial flux together with a zinc alloy with enhanced wetting, results in a very rapid heating of 345 °C/s. This may lead to severe thermal stress damage. However, for concentrations of AlCl3 in the flux according to the invention, the rate of heating drops to 260 °C/s or less, a level that is considered as adequate to avoid permanent damage to most articles.

Example 4 In the following series of experiments, additions of AlCl3 to mono- salts, being ZnCl2. NH4Cl, double-salts, being ZnCl2. 2NH4Cl, triple- salts, being ZnCl2. 3NH4Cl, and quadruple-salts, being ZnCl2. 4NH4Cl, are tested. The same alloy is used as in Example 2.

The influence of the composition of the flux upon the maximum rate of heating is measured according to the above-described standardised method. The results are reported in Table 3.

Table 3: Influence of the type of ZnCl2. xNH4Cl salt on the maximum rate of heating, using a zinc alloy bath with wetting promoters No Salt ZnCl2 + NH4C1 total Aids Heating rate (g/1) (g/1) (°C/s) 19* Mono 300 40 300 20* Double 300 0 333 21 Double 300 10 250 22 Double 300 14 195 23* Double 400 0 390 24 Double 400 10 200 25 Double 400 14 190 26* Double 500 0 345 27 Double 500 10 255 28 Double 500 14 200 29* Triple 250 0 330 30 Triple 250 10 260 31 Triple 250 14 240 32 Triple 400 14 220 33* Quadruple 240 20 380 34* Quadruple 240 30 340

* : counter-experiment In Experiments 20 to 28, the addition of Aids to a double-salt has a significant beneficial influence on the heating rate. The same applies for experiment 29 to 32, where triple-salt is used. However,

using a mono-salt, as in Experiments 19, or a quadruple-salt, as in Experiments 33 and 34, does not lead to heating rates considered as adequate to avoid permanent damage to most articles. It is to be noted that in the industrial practice, the NH4Cl to ZnCl2 molar ratio may vary between 1.7 and 3.3, when using double-salts or triple- salts, because of exhaustion and/or adjustment of the various flux components.

Example 5 Fe is a common pollutant in industrial flux tanks. This example demonstrates that, in addition to the known consequences, extreme Fe concentrations are also to be avoided to limit thermal stress when dip-galvanising steel articles. The galvanising composition of Example 3 was used.

Table 4: Influence of dissolved Fe in a double-salt flux at 500 g/1 No Dissolved Fe Aids Heating rate (g/1) (g/1) °C/s) 14 0 40 148 35 5 40 155 36 10 40 200 37 15 40 240 38* 40 40 350 *: counter-experiment