NIKON

Diffusionsschweißen mit nicht konstanter
Kraftaufbringung

Aim of the development

Up to now, the classic diffusion welding process has been implemented with constant force application. This means that the specified setpoint value of the force is recorded via corresponding load cells. A control loop with proportional and integral controllers is used to adjust the press travel as a manipulated variable to compensate for the thermally induced expansion or compression of the furnace system and charge over the temperature cycle of the welding process. This results in alternating lifting and lowering movements of the press plunger in order to approximately achieve or maintain the target force. The entire press system, including the control system, is designed and optimized to precisely control the target force in order to minimize the press movement. In contrast, the force provision in the project was to be specifically dynamic, i.e. through a continuous load change with a defined amplitude. This dynamic loading thus also alternated the stress states in the joining or boundary surfaces. This allows diffusion-impeding oxide layers to be broken down, thus further improving the formation of the composite.

Advantages and solutions

The considerations were based on structured samples with application-related requirements, dimensions and structures. These were worked out with project partners, manufactured at ifw Jena and qualified in terms of production technology. For the developed functional sample, a numerical model was first developed to simulate the resulting temperature field. Using the finite element method, a heat radiation and conduction model was created that is capable of representing the temperature field during welding. With the company MUT Advanced Heating GmbH, which specializes in the construction of furnace systems, a commercial press-temperature system was adapted so that it could be integrated into an existing furnace system at ifw Jena. By means of a software update and adjustments to the controller manipulated variables in the PLC control system, the control system could be adapted to generate controllable dynamic load cases during the running process. To check the function and performance of the press system, welding tests were carried out on simple sheet blanks and the results were evaluated metallographically. In addition, dynamic load application was demonstrated via the recorded force-time curve. For the defined shapes of the functional sample, the material for an AlMg3 foil of 100 µm thickness was first characterized in coordination with the project partners. For the further investigations, titanium and steel sheets or foils were also measured with respect to their surface profile, among other things, using a stylus instrument. The structures of the functional sample were produced by laser fusion cutting and laser sublimation cutting, which always leads to a certain amount of burr formation. In sum, these burrs at the edges lead to an uneven surface, which is therefore suitable for diffusion welding to a limited extent. For initial welding tests, the burrs on the laser-cut foils were carefully removed by grinding. A positive side effect resulting from this is the removal of the natural surface oxide layer as well as the change in the surface roughness profile and an associated further activation of the surface. The first tests on diffusion bonding under dynamic loading were carried out with a surface pressure of 3 MPa and at a welding temperature of 550 °C in order to obtain a quick estimate of the deformation differences between dynamic and static force application. However, it was found that very high shrinkage occurred, whereupon the compression was reduced to 2 MPa to reduce shrinkage. Welding experiments were then performed at 2 MPa and temperatures from 400 to 550 °C, welding both sheets (2 mm, lap joint) and specimen cubes (7.9 mm, congruent) to be able to determine any differences in shrinkage or deformation as a function of initial thickness. In the subsequent tensile shear tests on the welded lap joints, only three joints withstood the restraint load. At a welding temperature of 550 °C, the strength of the dynamically welded joint is only about half that of the statically welded joint, which makes the dynamic method appear at first glance to be the inferior variant. If the strengths are compared on the basis of shrinkage rather than welding temperature, the strength with dynamic welding and a shrinkage of 4.17% is more than twice that of static welding with a shrinkage of 4.09%. For welding tests to produce functional samples, 100 µm thick AlMg3 foils were first welded. No deviations were found in the outer dimensions, but in the channel structures. While for the 300 µm to 500 µm widths the complete channel was free, for 200 µm half or more were closed. The 100 µm wide channel was completely closed. For verification, X-ray measurements were made on statically and dynamically welded functional specimens. They showed good dimensional stability and, apart from the channel structure, no noticeable unjoined areas. Tightness was also confirmed by He leak tests. In ultrasonic tests, no irregular echo signals could be detected, which would indicate defective composite formation or internal cracks. It could be demonstrated in the process chain that a dynamic application of force during the welding process leads to lower deformations than with static process control. This made it possible to use a 50 °C higher welding temperature for a given deformation, which led to significantly increased diffusion and ultimately to almost twice the shear strength of dynamically welded components.

Target market

The joining process of diffusion bonding is also becoming increasingly important in additive manufacturing via slicing (defined layer buildup over foils or sheets). This results in new areas of application that show great application potential compared to conventional metal manufacturing. This applies to the weldability of the same or different materials through to construction and production design. One challenge, however, is that the individual process parameters have to be matched to each component to be joined in terms of component geometry and joint design, depending on its requirement profile. The performance of diffusion-welded components such as heat exchangers is significantly higher than that of components based on conventional joining technologies. Diffusion welding makes it possible to produce similar and dissimilar two-dimensional and materially bonded joints in the solid state. This means that, on the one hand, no molten phases occur, which means that precision parts with, for example, extremely fine structures, cavities or channels can also be manufactured, and, on the other hand, there is no need to use intermediate layers whose material composition differs from that of the base material to be joined.