With this slogan we now come back to our roots. Only this time we don't introduce a new manufacturing technology, but a way to tailor the additive building process to our needs. We will now take a deep dive into full free form beam shaping.
With Additive Manufacturing (AM), it is possible to create complex physical objects with only one production step. A single Laser Powder Bed Fusion (LPBF) AM system can build integral metal parts which are traditionally milled and turned, cast, soldered and more. The list is long. Typical applications can range from massive injection molds, to turbine blades with medium wall thicknesses, up to heat exchangers with very filigree fins and fluid circuits. Additionally, the system must support different materials and we also want to be able to adjust surface quality and mechanical properties. All of this should be covered by an AM system.
As you can imagine this is not an easy task. There are indeed a few different exposure types which can be assigned to different features of parts, e.g. surfaces are exposed ‘smoother’ than the inside, or support material is exposed in such a way that it is easy to remove. Nevertheless, there is only one single tool that is used for all of that: the laser beam. This laser beam has fixed characteristics and of upmost importance a fixed power distribution.
This single and static tool must fit each and every application. This is typically done by modifying a set of parameters, e.g. laser power, hatch distance and scan speed. With these parameters we control the energy input in the material and how the 3D-printed part is melted. For example, if striving for productivity, we can increase laser power and scan speed. However, this mechanism has physical limits caused by the fixed power distribution of the laser beam. At a certain point we just cannot further increase productivity without sacrifying other aspects of the building process. Imagine building a house with a single hammer and the only parameters you can control is how hard and where you are hitting things. Hitting things harder probably won't finish the job faster, but you will start breaking things. What you actually want is a set of specialized tools. Each of those can perform best in the task they were made for.
And this is exactly what’s missing in LPBF. As we’ve not been able to use different tools, we’ve not been utilizing the full potential of the building process. With beam shaping we can now also change the diameter and the power distribution within the laser beam. These two new powerful parameters enable unseen processing strategies.
When we are talking about specific tools, we also want them to be as flexibly exchangeable as possible. The reason is shown in Figure 3. Having a specialized tool is nice, but if it is static, we are just as limited as with the current AM process, as every tool can only do best what it was made for. With beam shaping we can also quickly switch between tools, e.g. a special tool for surface quality or another special tool for productivity. These tools can be specifically assigned to different features of our building process, allowing us to gain advantage in many criteria.
To understand the magic behind beam shaping and how it actually improves the building process, we first have to understand how the laser interacts with the powder. For this task Jochen Philippi developed a high-speed imaging system, giving us the required insights. You will learn more about this tool in the upcoming days of our innovation week.
In these high-speed recordings three effects can be observed: heating, melting and vaporization. The usual LPBF process takes place in the vaporization regime. This is also called deep penetration welding. In this regime the material is not only melted, but also evaporated in the interaction zone. This is caused by high laser power that is focused to a small area. In this case the intensity is so high that the generated heat is not able to dissipate fast enough, causing the melt pool to increase in temperature and evaporate.
Because of the vaporization and the steep temperature gradients, this process regime is very turbulent. Strong currents in the melt pool cause spatters to be ejected and uneven surfaces of the weld track. The spatters and uneven surfaces may interfere with the laser beam and the recoating, causing the building process to become unstable. The vaporization also promotes the formation of pores in the 3D-printed part, which can lead to a decrease of the overall mechanical performance. In Figure 4 you can see for example how the scan speed and the laser power influences the appearance of the weld tracks. In the bottom left picture, you can see a gas pore and in the bottom right picture the weld track shows a very uneven surface.
The idea behind beam shaping is to modify the meltpool behavior and gain full control over the building process. This is done by redistributing the laser power in such a way that temperature gradients, and thus melt currents, are adapted. This also means that we are now not only limited to deep penetration welding anymore, but we can define specific regions that are heated or melted. We are convinced this is the key to productivity, powder efficiency and the processing of otherwise unweldable materials. Changing how the powder is melted and fused allows for new processing strategies and a wider range of optimization.
This context is shown in Figure 5 specifically for the laser beam. With the current LPBF processes, we are limited to modifying the laser power only. This allows us to design building processes only ‘as good as possible’. Dominik Wolf from EOS Innovation Management already introduced the new parameter ‘beam diameter’, allowing even better building processes. And with beam shaping we can tailor the building process to the physical optimum.
Our current prototypes can create almost any arbitrary beam shape within a circle of 2mm on the powder bed. The beam shape can be flexibly changed within the building process and the supported laser power is over 1kW. The light engine is designed as a modular solution and can for example be easily installed as an upgrade on the M3xx-platform.
In the following we want to show you how beam shaping can be used on a real application. As we are still in development, we weren’t able to change beam shapes on the fly until now. That’s the reason the results for this application have only been simulated. Nevertheless, each of the presented beamshapes has been tested, to make sure all claims are as true as possible.
Let’s take the heat exchanger in Figure 6 as an example. This part is a good test case, as it combines different distinct features in one single part, making it ideal for multi tool processing.
This part essentially consists of only three features:
The contour for such parts usually has the requirement to be dense, i.e. no open porosity, and to have a low roughness. The surface will be in contact with different fluids and leak tightness is a mandatory requirement.
Secondly, the part obviously has a quite bulky inskin. For this feature you do not want to have too much porosity, as this would reduce the thermal conductivity and reduce the efficiency of the heat exchanger. Other than that, you want to build this as fast as possible as this is the main lever in terms of reducing cost per part. As most of the part is inskin, productivity represents the top priority here.
The last feature are the thin-walled fins. These ‘single walls’ are not as easy to build as they might appear. To ensure good integrity, they need to be processed very carefully with highly specialized exposure parameters. While they don’t have much volume, they therefore still require a good amount of time to build. For the fins the optimization target is to increase productivity while maintaining integrity.
According to these requirements the exposure tools shown in Figure 8 are proposed.
Contour: For the contour we decide to use the standard gaussian beam as it is currently used in state of the art LPBF systems. There is no smaller achieveable beam diameter within the machines optical system. For this reason we use this beam shape to produce very fine surfaces with low roughness.
Inksin: The next feature is the massive inner part. We want to build the inskin as fast as possible to lower the part costs. That is the reason why we decided to use a beam shape with a big diameter and a Top-Hat intensity distribution. Due to its size, it is capable of melting more powder than the standard gaussian laser beam. In first experiments, we also found that the amount of spatter was lower as compared to the standard process. This is a nice side effect and further reduces part costs, especially for big exposed areas.
Fins: Finally it is time to take a closer look at the fins. As learned from the contour, there is no smaller achieveable beam diameter than with the gaussian beam. Beam shaping cannot change this fact and we thus cannot create a special tool for the thin fins. To still increase productivity, we apply a simple trick. We just split our single gaussian beam into three identical gaussian beams with equal powers. Each of those beams builds their own fin. This results in a threefold of the initial productivity.
Our simulation results demonstrate that we could decrease about 48% of the building time for a single part by using specialized laser tools. In our example this would be equivalent to a cost reduction of around 30%.
Build time savings and cost reductions strongly depend on the part geometry and the boundary conditions. We showed that beam shaping can be used to optimize different aspects of an application. However, productivity is only one side of beam shaping. By controlling the melt pool behavior, we are convinced beam shaping can also create advantages with regard to powder efficiency and the processing of nowadays not processable materials in AM.
At EOS we strive for perfection in helping our customers with best in class available technology. The EOS Innovation Management team has the task to develop the products for tomorrow and the day after tomorrow. We took this task seriously and developed the first ever serial free form high power beam shaping solution for LPBF systems. As described earlier we have opened a huge new parameter space. This space is too big to be explored by ourselves. For this reason we're seeking your help. We want to keep developing this technology together in a collaboration and push AM to the next level.