Brazing of Cemented Tungsten Carbide to DIN 1.2343 Tool Steel

Introduction

Although cemented carbide is resistance to thermal shock, it will not withstand stresses caused by severe thermal gradients when brazed to a material such as DIN 1.2343 Tool Steel with a significantly different coefficient of thermal expansion. For comparison, thermal expansion coefficient of the Cemented Tungsten Carbide (e.g. Grade TSM33, 10 wt.% Co) is about 5.3 x 10^-6 K^-1 at 20-400 °C, while coefficient of linear thermal expansion of the DIN 1.2343 Tool Steel (0.33-0.43 C; 0.2-0.5 Mn; 0.8-1.2 Si; 4.75-5.50 Cr; 0.3 Ni; 1.10-1.50 Mo; 0.3-0.6 V; 0.25 Cu; 0.03 P; 0.03 S) in the same temperature range is 13.2 x 10^-6 K^-1. Failure can result from such stresses during cooling. Note that slow cooling affords some reduction in brazing stress during cool-down.

Accommodating the inevitable cooling stresses – General Considerations

Because of the significant mismatch in expansion coefficients that typically exist between the Tungsten Carbide and the backing material (steel shank) to which it is brazed, whichever brazing filler metal is used it needs to be able to accommodate the stresses that develop within the joints on cooling.

When cooling a joint, the backing material wants to contract faster and further than the piece of Tungsten Carbide to which it is now securely brazed. This mismatch results in the development of shear stresses within the joint. If the brazing filler metal has a high strength it will transmit these stresses directly into the relatively brittle Tungsten Carbide and cause it to crack. What is required is brazing filler metal that has a low yield point that will deform plastically allowing the stress to be dissipated. Unfortunately, the strength of brazing filler metal in a brazement is not primarily a function of the strength of that filler metal, but the strength of the component materials and the thickness of the brazing filler metal layer (seam) within the joint. Thus, the normal perception that a strong joint is required when brazing Tungsten Carbide could be questioned. In this respect, a number of points are to be emphasized:

• It can be understood that if too strong a joint is produced it will not be able to deform to dissipate the cooling stresses.
• If the filler metal is too strong it will simple transmit the stress into the component causing it to distort or the piece of Tungsten Carbide to crack to relieve the stresses that have developed.
• If the level of stress developed during cooling are extremely high, and the component sufficiently robust such that it cannot distort or crack, the brazing filler metal itself can rupture resulting in the piece of Tungsten Carbide becoming detached.
• It is also possible in circumstances where the component is sufficiently robust for the stresses developed on cooling to be contained as residual stresses within the component. The residual stresses locked up in the joint only show themselves when some additional stress is added to the component by the application of some external force, either a physical mechanical load (e.g. grinding) or thermally generated stress.

As to the quality of the joints, it is clear that if the brazing filler metal is unable to accommodate the stress that inevitable develop during the cooling of the joints, one or more of four outcomes are possible:

1. The part distorts.
2. The Tungsten Carbide cracks.
3. The joint fails completely and the Tungsten Carbide part comes off.
4. Nothing happens on initial cooling, but the Tungsten Carbide cracks or comes off in service or when the part is ground.

Outcomes (1), (2) and (3) present themselves as obvious problems that immediately point to the fact that there is something wrong with the brazing process. Outcome (4), the “nothing happen” situation is the most concerning, since unless the stresses that developed on cooling of the joint have been accommodated satisfactorily, the joints could contain a high level of residual stress. It is also important to realize that any load applied to the joint during use or grinding increases the level of stress in the joint resulting in the Tungsten Carbide cracking or coming off.

It follows from the preceding text, that when brazing Tungsten Carbide, joints with a good level of ductility are required, such that they can deform readily and allow the stresses that arise on cooling to be dissipated.

It also follows, that strong joints, where only a thin layer of brazing filler metal is present between the Tungsten Carbide and its backing material, will not allow the brazing filler metal to deform in a ductile fashion, and therefore will not allow dissipation of the cooling stresses. Joints that have a thicker layer of brazing filler metal (“seam”) between the Tungsten Carbide and the backing material will be more able to deform and therefore allow dissipation of the cooling stresses. In essence ductile joints are required, which usually means joints are going to be lower in strength than might be the case for other joints.

From the need to produce ductile joints three approaches in relation to the selection of brazing filler metal for applications involving the brazing of Tungsten Carbide have evolved:

1. The use of filler metals that possess a high level of ductility.
2. The use of filler metals that produce joints with thicker than normal layers of filler metal.
3. The use of products and methods that produce artificially thick brazed joints.

Extensive industrial experience suggests that braze strains can be relieved by, for example, a so-called a “Sandwich Braze”. The sandwich braze consists of a Copper shim between the carbide and steel part to be assembled as shown schematically in Fig. 1.

Fig. 1: Structure of the “Sandwich Braze” (schematically)

The Copper is not melted because only a low or “medium” temperature (e.g. Ag-based) brazing (filler) metal is used. The Cu shim is malleable enough to deform under the brazing strain without losing its bond to the steel or carbide parts. Using the Copper shim sandwiched between two shims of Ag-based filler material within the brazing assembly is useful only for light or medium duty since it will tend to “mush” and be squeezed out if the application involves heavy loading or high impact. It will not provide the uniform support required to prevent breakage of the carbide. A potential alternative is a Nickel shim, which will withstand more impact. However, Nickel does not have the malleability of copper and will not relieve the braze strains as effectively as Copper.

Another possibility is to conduct the stresses to the generally ductile brazing alloy by widening the brazing gap. The gap width can be adjusted by a Ni-network incorporated into the brazing alloy (Fig. 2).

a)	b)

Fig. 2: Back-scattered Electron Images (BEIs) of the Cemented Carbide/Steel joint fabricated by brazing using Ag-based filler with the incorporated Ni-network: a) general view and b) zoomed image of the brazing seam. Note. Width of the brazing seam is indicated in (b).

Specifics of the project

1. The use vacuum furnace brazing

 In the present project, (low-) vacuum (~10^-2 mbar) furnace brazing, and not a R.F. (induction heating) will be employed. Induction heating is inherently non-uniform process: namely, the thin surface layer is overheated relative to the sample core, whereas each local microstructural constituent is exposed to a different temperature. This happen mainly because the depth of field penetration into cemented carbide sample is thin and because its constituents Cobalt and Tungsten Carbide (WC) have (very) different electrical resistivity and thermal conductivity. As a consequence of this overheating, there may be a partial depletion of cobalt at the surface because Cobalt melts and evaporates at a substantially lower temperature than Tungsten Carbide. This leads to formation of grooves surrounded cemented carbide crystals, thus causing a partial loss of bonding between constituent phases. These grooves assist crack initiation and propagation resulting in a brittle mode of fracture. In addition, a fast heating/cooling cycle results in deleterious thermal stresses which are strong at sample surfaces, edges and corners.

Another advantage of the vacuum furnace brazing is that although the DIN 1.2343 tool steel contains substantial percentage of Chromium, the joining to the Cemented Tungsten Carbide can be accomplished without using any flux. However, the use of vacuum brazing imposes serious restrictions on the type of filler metals that can be used. Obviously, materials having high vapour pressure must be avoided.

2. Some aspects of post-braze heat-treatment

In general, brazing of toll steel is best done prior or in conjunction with the hardening (or “toughening”) operation. The steel being brazed should be studied carefully to determine:

1. The proper heat treating cycle.
2. The kind of quench necessary (vacuum or protective gas).
3. The optimum brazing filler metal.
4. The proper technique for the combined heat treating and brazing operation to achieve desired properties and service life.

It should be stressed here that the quenching can set up severe temperature gradients which produce differential expansions and contractions that may rupture brazed joints. Also, austenite to martensite transformations (in certain steels), that can be take place during the brazing cycle or heat-treatment, or both, may result in the parts contracting, expanding, finally contracting again. (Note that the specific volumes of ferrite + carbide, of austenite, and of martensite are all somewhat different.) The annealed structure of alloyed ferrite and carbide contracts upon austenization, since the face-centered cubic (FCC) lattice has a greater packing density of atoms. There is an expansion when austenite changes to martensite on cooling but even when this transformation is complete, the expansion does not exactly coincide with the prior contraction upon heating and formation of austenite. The atomic dispersion of Carbon in martensite and perhaps the presence of microstresses result in a greater volume that corresponds to the original ferrite and carbide. Therefore, tempering of martensite is accompanied by a slight contraction.)

3. Approach in the present project

It has already been highlighted, that obtaining good wetting and bonding of a brazing filler metal onto Tungsten Carbide is not as a straightforward a proposition as it can be for many other materials. It is important to keep in mind that too high processing temperature leads to an excessive grain growth, which will have a detrimental effect on the mechanical properties of the steel. On the other hand, it is also clear that in order to provide a sufficient fluidity of the liquid brazing alloy, the processing temperature should be, at least 40 °C higher than the corresponding liquidus temperature. It is to be mentioned here that capillary flow is enhanced greatly when the surface is roughened. (Grit blasting or shot peening of the brazing surface of Cemented Carbide prior to joining can be recommended.) However, too high fluidity of a filler metal is not always a good thing: sometimes, as explained earlier, we need a relatively thick “brazing seam”.

Based on the above considerations it would be prudent to use a Mn-bearing filler metals. Some of the Cu-Ni-Mn brazing alloys wet carbides at temperatures compatible with steel heat-treatment temperatures. They also have the reported ability to fill “wide-gap” joints and provide good elevated temperature properties.

             The first tempering of the DIN 1.2343 tool steel shank was carried out immediately after the brazing cycle (“hardening”) in the same vacuum furnace at 580 °C for 2 hours followed by slow (“furnace”) cooling. The second tempering to the specified hardness HRC can be performed in air at 620 °C for 2 hours.