August 22, 2016
Eric Duoss of Lawrence Livermore National Laboratory explains how advancements in materials science are expanding material choices.
PwC: Eric, can you please tell us about your research?
Eric Duoss: Sure. Our research focuses on developing new materials and methods for additive manufacturing. We are creating novel 3D printing processes that can achieve feature sizes that are smaller than what most commercially available methods can produce-so microscale and, in some cases, nanoscale feature sizes. The processes we’re creating have the ability to incorporate different materials in the same printed part. Also, because of the geometric complexity that additive manufacturing enables, you can tune or control material properties such as density, stiffness, strength, and thermal expansion by designing complex geometries, or microarchitectures. We are creating materials via microarchitected design that have properties or property combinations that were previously unobtainable.
PwC: Is your research focused only on materials, or are you studying the printing processes as well?
ED: It’s both, and common between the two is the materials science. We strive to understand structure-process-propertyperformance relationships. That means understanding how the feedstock materials, printing processes, and post-processing interplay to affect the final part properties and performance. We do this with a combination of experimentation and simulation.
PwC: How do you control material properties
Eric Duoss: In the case of architected design, we do that by controlling the arrangement of material at microscale or nanoscale. For example, we have printed microarchitectures that are extremely stiff and strong yet lightweight.
“With post-processing, the finest features we’ve demonstrated are down to 200 nanometers.”
We have also designed architectures with programmed coefficients of thermal expansion [CTE] and Young’s modulus. We do this by precisely placing disparate materials in the same unit cell along with empty space or void. The unit cells are then configured into a lattice that has the programmed CTE. The CTE values can range from positive to negative or can even be neutral. In the case of a negative CTE, the material will actually shrink as you heat it. The common theme for architected materials is that these structures would be difficult or impossible to fabricate via any other way besides additive methods.
PwC: What resolutions are you working with?
Eric Duoss: The resolution of printing depends upon the specific process. In one process called direct ink writing [DIW], we use an extrusion technique where we can employ very small diameter nozzles. The nozzle size sets the ultimate feature size. The smallest nozzles we’ve printed with are 500 nanometers in diameter. That’s pretty small. With post-processing, the finest features we’ve demonstrated are down to 200 nanometers. In practice, we most often work with feature sizes between 1 and 500 micrometers. I should point out that if you’re increasing the resolution, you might also be increasing the build time. Generally there is a tradeoff there.
PwC: How does your research fit within the broader trends in 3D printing?
Eric Duoss: In 3D printing [also known as additive manufacturing] so far, hardware designers and mechanical engineers have developed printers and processes.I see the field expanding its focus to materials. Designers love what they can do with 3D printing processes and the complex geometries they are able to make. But a continuing challenge is that these designers cannot use the materials they desire or get the properties they want with existing 3D printing processes.A lot of our work focuses on addressing these challenges. We ask ourselves, what new materials are possible for existing and emerging processes? We try to design new feedstock materials so they’re compatible with existing printing processes and new processes.
We develop probably 10 to 20 new materials for our in-house 3D printing processes per year. What we create is early stage R&D that’s not necessarily ready for commercialization. Other research efforts at LLNL [Lawrence Livermore National Laboratory] are focused on methods to qualify and certify many of these new materials for manufacturing and application purposes.
PwC: Can you provide an example of a new material for an existing process?
Eric Duoss: We are designing new feedstocks that are elastomeric, or highly flexible and stretchable materials, for DIW. This method is akin to fused deposition modeling [FDM], the technology in many of the desktop 3D printers, only it does not use a heated nozzle. Both have an extruder, but in FDM, the filament [the feedstock material] is a thermoplastic. When you heat the plastic, it melts, which allows it to flow. When it exits the nozzle, it rapidly cools and solidifies.
For DIW, we design complex fluids, which we call inks, that also can be extruded at room temperature. Complex fluids are a mixture of particles and a suspension media, and they are often viscoelastic- that is, they behave like solids under standard conditions, but undergo a change to more liquid-like behavior under shear.
“Our long-term hope is that they [models] will be predictive, so that during the design of the component, someone could analyze the performance characteristics [of material] and adjust the design accordingly.”
In the case of DIW, when you flow the ink out of a nozzle, it is designed to be shear thinning- that is, its viscosity [fluid resistance to flow] is lowered in the presence of shear. As the fluid exits the nozzle, it essentially will gel or solidify, so it maintains its shape even at room temperature. This is really great, because now extrusion-based printing is opened beyond the subset of thermoplastic materials that FDM was limited to. It opens up a much larger material space, including many of these elastomeric materials that we are currently pursuing. The innovation here is developing the complex fluid to have particular flow properties, or what is also called rheological properties.
PwC: You’ve talked a lot about understanding the flow and microarchitecture and so on. How do you capture and communicate this understanding?
Eric Duoss: To tackle that challenge, LLNL is developing sophisticated models that capture-at multiple length and time scales- the behavior of materials during printing. For example, LLNL researcher Wayne King is leading an effort to model the laser melting process for metal-based additive manufacturing. The model captures the understanding of how the laser interacts with the material, what thermal gradients form during the printing process, how the microstructure changes during this melting process, the residual stresses that build up from the rapid heating or cooling, and so on.
You can scale up the model to the finished component and really understand how those process conditions affect a component’s overall performance. Such multi-scale models, enabled by LLNL’s high-performance computing capability, really are an effective way of capturing and communicating our understanding of additive manufacturing processes. “Our long-term hope is that they [models] will be predictive, so that during the design of the component, someone could analyze the performance characteristics [of material] and adjust the design accordingly.”
PwC: Could these models be predictive of properties of the final product?
Eric Duoss: We are in the early days in building and using these models. Our long-term hope is that they will be predictive, so that during the design of the component, someone could analyze the performance characteristics and adjust the design accordingly. We also hope that these models can accelerate qualification and certification of the end products. It’s for these reasons that such models have gained a lot of attention. At many of the industry events we attend, manufacturers and end users of these additive manufacturing platforms really want to see these models evolve and become more sophisticated. In many cases, you simply cannot achieve this sophistication without the significant high-performance computing resources available at places like LLNL.
In addition, one of the really exciting aspects of additive manufacturing is that it’s a layer-bylayer process, so there is potential to capture data at each step in the build process via in situ [real-time] characterization. This capability starts to become a big data problem and you have to look for ways to make use of all the data that you collect, preferably in real time. For example, we’re looking at ways to characterize the material as it is being built up. If you can do it well enough and react quickly, then if a defect occurs during manufacturing, you can discard the item or you can go back and correct that defect on the fly if there is a way to do so. Of course, we are also looking at ways to use the captured data to refine and validate our models and push them to become more predictive.
PwC: How are you able to print fully functional, complete systems?
Eric Duoss: To build systems, we must deal with multiple materials. For us, something that has worked out very well is to build structures with one process and add some functionality with another process, and then add more functionality again with a third process. Essentially we are combining multiple steps in a fabrication cycle. We call such a multistep process deterministic deposition.
Deterministic deposition is a bit different from traditional 3D printing, because now you’re placing material onto a preexisting 3D structure. With most 3D printing technologies, you start from a flat substrate and the part grows layer by layer. Depositing material on a 3D structure requires a lot more sophistication, accuracy, and referencing.
From what I can tell, 3D printing onto a 3D shape is very new. There will be a lot of cases where a customer or an end user will say, “I have this existing device, and it has some form factor. How can I add to it?” In many cases, that device is not a nice rectilinear device. Often it’s curvilinear, and it has some complexity. That’s one issue we needed to solve before we could combine multiple processes.
We’ve used this technique to build antennas and different RF [radio frequency] devices. We are also working on things like metamaterials, batteries, electrodes, and resistors. We are building a toolset that in the end could get you to a fully functional, active smart system.
PwC: What are you learning about using multiple materials in your build cycle? What is the future potential here?
Eric Duoss: As long as you are dealing with materials of the same type-all organic material, for example-this is an easier problem. Some printers already combine hard and soft plastic. If you start talking about multi-material in the sense that you want to put plastic with metal, then that brings more challenges because the materials are so very different.
Using the understanding we are gaining at the micro and nano levels, we have developed a microfluidic control system that allows us to pattern multiple materials in the same part. It’s all in the same tool, but it’s pretty neat because we can flow one material into our chamber, shoot an image [in the case of a photosensitive resin], cure in selected locations, flow that material out, and flow a second material in. These layers are very thin-they’re micronscale thickness. In some cases, you also could create smooth gradients between materials, depending on how you cure it.
We can accomplish such control because we have good knowledge of the photosensitive resin; understand the way that resin interacts with light; and understand issues like a cure profile, the kinetics, the depth, and how the light is absorbed. We have a process model to help us accelerate from the build process, but also to help us know how to move and change the resin itself.
However, more research and development are required to fully realize the potential of multi-material 3D printing-basically patterning plastics, metals, and ceramics in the same part using the same process. In the future, you will also see different active systems-sensors, electronics, combined with structural elements and so on-built into the same component using 3D printing.