Dr. Matthews and his team have developed a new laser-based method for 3D printing of large metal objects called Diode-Based Additive Manufacturing (DiAM). It uses high-powered lasers to flash-print an entire layer of metal powder. The process will enable large metal objects to be printed in a fraction of the time typically needed for metal 3D printers.
Tech Briefs: How did you come to consider 3D metal printing while working for the National Ignition Facility (NIF)?
Dr. Manyalibo “Ibo” Matthews: While 3D printing of metals was not explicitly considered within R&D activities supporting the National Ignition Facility (NIF), an overlap was recognized in terms of the optics that are used in both areas. Our use of NIF-related optics for 3D printing centers around a device we call an Optically Addressable Light Valve (OALV). At NIF, that device is used as a spatial light blocker.
The NIF is essentially a very large laser system with a target chamber at its core, which is used for high energy-density experiments. The laser light is very intense, 100s of terawatts (TW) of power and megajoules (MJ) of energy per total pulse. As the light passes through the optics, a smudge or defect on an optic will tend to initiate laser damage. In fact, the first half of my career here was focused on understanding that phenomenon — how optics are damaged under intense laser light, and how to repair the damage. If you keep hitting the site of the damage over and over with laser pulses, the damage site can enlarge or grow and ultimately the optic will fail. So instead of letting that happen, you can block out part of the beam.
We have an elaborate method to map out where all of those defects are on all 192 optical beamlines that face the target. We look at the optics and say, “There’s a damage site, it’s getting bigger for whatever reason.” We don’t want the next laser shot to put more light there, so we use the OALV to block out part of the beam. It’s done dynamically, meaning that we can choose at each shot, which part of the beam we want to block as we decide what makes more sense shot-to-shot. To be able to do that we need a light-blocking device (optic) that can be addressed so we can choose one set of pixels in the beam versus another. That’s where the magic of the OALV comes in. The light valve works such that shining blue light on it changes its ability to rotate the polarization of incoming laser light; light passing through will either have its polarization rotated if the valve is not addressed and not rotated if it is. Then we send that mixed-polarization-state beam through a polarization filter that will block the rotated polarization beam and allow the unrotated one through.
Tech Briefs: How is that used for 3D printing?
Dr. Matthews: The same spatial light blocker that we’ve used to take out parts of the high energy laser beam for an NIF experiment can be used with light suitable for 3D printing. It can pattern that light to print an entire layer at once, by allowing light to irradiate and melt layers of powdered metal.
Let me take a moment to explain powder-based printing of metals. The process is known as selective laser melting. It is sometimes called selective laser sintering, which is not right because we’re melting the powder. It’s also known as powder bed fusion additive manufacturing (PBFAM). The process is relatively simple: you have a build-plate of metal that you want to build a part out of. You use a powder spreader, which is like a wiper blade, that spreads a thin layer of metal powder. The powder is on the order of 30 microns thick — a third of the width of a human hair. The layers are about that thick as well. Then you focus a laser beam down to almost that same size — 50 to 100 microns — and write a desired pattern with it. You can direct the beam to write out a pattern to create an entire layer. Where you’ve written with the laser, you’ve melted the powder, so you have a melted, then solidified, trace of metal surrounded by powder that you didn’t melt. You then lower the part down, spread another layer and repeat. By doing that, you can build up a layer-by-layer 3D print of metal surrounded by unmelted powder. However, that takes a long time because you have to wait for the laser beam to go around and you still have to spread the powder and move the part down. It takes a significant amount of time just to write out all these little traces. Although we’re writing out at a meter per second, which sounds fast, the beam is only 50 microns, so if we want to do a fully solid part we have to do 50-micron scans. The writing time for a layer can range from 10s of seconds to minutes.
Our idea was to take the OALV used in the NIF so that instead of having to write each trace, we can produce an entire image at once. What we’re putting through the filter is not a focusing beam but a sheet of light — high power laser light. Wherever the filter is blocking the beam is where we have unmelted powder; where it’s allowing the beam through, we melt the powder. By doing that we take out the raster scanning, we do a whole layer at once and save a tremendous amount of time.
Tech Briefs: Has metal powder 3D printing been going on for a while?
Dr. Matthews: Yes, it goes back to the 80s in fact, but it didn’t really take off until the advent of a low-cost high-power fiber laser. Fiber lasers as a technology started in the early 2000s. They’re very efficient and have what’s called very good ‘mode quality’. That technology rejuvenated the idea of metal 3D printing. There were also a number of patents that expired, which helped spur things. Those two things coincided about 10 years ago. Since then, everyone’s been in a race to perfect the process. There are currently a number of commercial vendors of machines, mostly based in Germany. General Electric is also implementing it in production and NASA has a very active program in Inconel superalloys for rocket engines.
Tech Briefs: What sorts of parts are made using this process?
Dr. Matthews: That’s a good question, because I think that a misconception about 3D printing is that it will take over all manufacturing, but that’s almost certainly not the case. What it does well is very complex designs — things that you can’t easily or cheaply machine or assemble. The chief example of that is a metal lattice, which can offer high strength but at low weight. 3D printing can remove enough solid material to make the part light but strong, which can lead to a new generation of parts. In the NASA case, the cooling in the fuel channels for a rocket engine ideally follows along the skin of the rocket. But you can’t machine that contoured channel very well.
There is a range of metals appropriate for 3D printing, but at this moment, it depends on the industry. There’s a fairly active additive manufacturing (AM) effort for jewelry, which uses hard-to-assemble gold and silver components. For medical applications, there are titanium-based alloys for implants and magnesium-based alloys for dissolvable implants. There are also good aerospace — and potentially automotive — applications. A number of alloys are used, depending on the application: nickel superalloys for turbines, as well as titanium and aluminum alloys for structure. The most common, but probably in the long term not as useful, are stainless and tool steel.
Tech Briefs: There are two components to your design, a high-power diode array and a pulsed laser — why the combination?
Dr. Matthews: The diodes are used to deliver the majority of the heat to melt the layers, but they pulse too slowly — 20 milliseconds — so if we just use the diodes we would end up melting the entire build. All of the powder would melt because the heat diffuses away too quickly. So, we heat up the build plate with the diodes and then send in a 7-nanosecond pulse to deliver the finishing touch, melting the layer before the heat can diffuse away.
Tech Briefs: Roughly how long does it take to print a layer?
Dr. Matthews: Each layer is flashed in at 20 milliseconds – that’s how long it takes to melt the layer. The powder spreading depends on the spreader technology, typically several seconds per spread.
Tech Briefs: Is yours the only system that does the printing this way?
Dr. Matthews: Yes, as far as we know. Other technologies exist using so-called ‘direct diode’ printing where individual diodes can be addressed and imaged to the build plane, but they tend to lack the flexibility of the OALV approach used by us.
Tech Briefs: Do you have any idea when this might be commercialized?
Dr. Matthews: At this time, the technology has been licensed out to a partner.
Tech Briefs: Do you have more work to do on this project?
Dr. Matthews: Yes, we just demonstrated the printing – the engineering proof-of-principle — but there’s a whole world of material science that needs to be studied to understand how very different thermal gradients and thermal histories affect the microstructure and ultimate mechanical properties using this method. I’d say that’s an exciting world left to be explored. There’s a new project we are starting soon that will do just that.
Tech Briefs: How would you address the question of different metal properties? Would you adjust the pulse width?
Dr. Matthews: Exactly as you said. Depending on the metal and the microstructure we want, we could adjust the pulse length and the gradients of the image we’re printing with — It’s yet to be determined.