Additive Manufacturing (3D Printing): Successes, Challenges, and the Path Forward
Release date: November 23, 2020


Chip Blankenship and Ji Ma, professors in the University of Virginia’s Materials Science and Engineering Department, join the CORROSION journal Interview Series to discuss why we’re drawn to additive manufacturing. Topics discussed on this episode include successful applications of the technique, current challenges, and future applications.
Transcript
[introductory comments]
Sammy Miles: Thank you for joining me.
Chip Blankenship: Good to be here with you, Sammy.
Ji Ma: Thanks, Sammy.
SM: We’ve previously touched on the basics of additive manufacturing in another podcast episode. Today’s episode is going to build on that, focusing on why we’re drawn to additive manufacturing, some current obstacles, and ongoing work to evolve this process. Before we dive in, would you both like to share a little about yourselves?
CB: Sure, Sammy. This is Chip. I’m a metallurgist by training, and most of that experience is in aerospace structural alloys: aluminum, titanium, and nickel-based superalloys. I spent most of my 25-year GE career in the aircraft engines group, leading commercial engine programs through design, test, and entry into service. I also held senior executive positions there, in other GE divisions, and at other companies. Now I’m back in the academic world, supporting research and preparing our next generation of leaders to take their turn.
JM: Thanks, Sammy. This is Ji. Unlike Chip, I never left the safety bubble of the academic world. My background, I am also a metallurgist by training. I spent many years working on smart materials, phase transformation in different alloys. Before that, I was in structural design. I’ve gone through a few different iterations of different areas in academia.
SM: Great. To start us off, I guess we want to know what has spurred the interest in additive manufacturing?
CB: First of all, designs engineers were thrilled with the idea that they could make assemblies as one part, reducing part count and complexity in the structures they were designing. Also designing specifically for the additive manufacturing pathway, really opening up the design space to make a single part. They love the lack of tooling required and the resulting speed to get a part in their hands. Then again, meeting dimensions requirements really isn’t the end of the design story. The material scientist is simultaneously worried that the printed material is not the same material that’s in the mechanical property handbook, yet intrigued with how this thermal processing path might be used to tailor mechanical and maybe even other properties. Everyone’s worried about the as-printed surface and specific manufacturing defects, such as porosity. Everyone then gets aggravated with all the post-processing that’s required to relieve some of these issues that show up on this processing path.
JM: As a materials designer, we’re trying to create new materials. Additive, beyond just delivering very complex geometries, also gives us a lot of new tools to play with. As a materials designer, we’re control freaks in many ways. We want different ways to tailor then change the material properties. With additive, you have the ability to change material properties on the fly at different locations of the material. So it’s a very exciting platform, even for us, to create materials with properties that you cannot get from any other technique, for example.
CB: And you know, it’s not just the technology folks that are super excited about additive. As business leaders, we are intrigued by the idea that the supply chain could be disrupted, and our company can be more vertically integrated because we make our own parts instead of source from the experts out there in forging or casting.
SM: With that, that includes a few things that are really good about additive manufacturing. Where have we been really successful with it? What can we do now? What have we been really good at, moving forward?
CB: Some of the highlights of what’s been done so far, if you look at the medical field, customized parts for implants into people, where different size and customization is really important, where flexibility in matching the stiffness of a part to a bone or a joint in an orthopedic kind of implant, as well as in dental implants. They’ve been very successful in this field so far.
JM: I think later on in this discussion we’ll pour a little bit of cold water on many aspects of additive, but I do want to say that, beyond printing final parts, additive manufacturing and 3D printing in general has been extremely successful in prototyping, which is really where a lot of the current effort has been going to already. They have been doing very well in that area.
CB: That was really the start of it, wasn’t it, Ji? I mean, with SLA and other things, where designers could hold in their hand a piece of plastic, something they wanted somebody else to make out of metal, right?
JM: Exactly. It’s not — I wouldn’t say additive — it’s a completely new and unproven technology. It is something that’s already embedded into many areas in our design process and production process. I think now the question is, How do we complete the whole entire cycle and turn this into something that is going to be making the part that’s going on the next airplane, putting into the next component into the engine? That’s the challenge.
CB: Speaking of engines, GE, with their LEAP engine for the A320neo and the Boeing 737 Max, at the end of 2018, they passed a milestone of making the 30,000th fuel nozzle for that engine. The really attractive thing about additive for that advanced cobalt alloy fuel nozzle was, you used to have to have 20 parts and weld and braze these things together, and additive you can just skip all that and print the part. So even though it takes a long time to print a single part, you cut out so many other assembly activities that it’s really worth it. I think that’s the thing that we face right now, is right now you have to choose a really good part where additive adds value to making the part versus just any part.
SM: Building on that, then, when you're printing out that many parts, how do you ensure consistency? If you use the same design, can you print it on a different machine? Are you having to worry about the powders that you're using?
CB: Some of that depends on whether we’re talking about printing a polymer, a polymer composite, or a metal. Back to where we’ve been really successful, there’s this company called Local Motors that are making an electric, low-speed vehicle for things like shuttling people around in an urban environment or a college campus or the like. They print 80% of this vehicle. It’s a polymer printing process. They’ve done a lot of R&D on what the material properties are for a chassis that’s printed that way. It takes nine hours to print it, but there’s 90% fewer parts than a vehicle like that. They’ve done their homework and they understand what good looks like for their process, but it’s a lot of in-house expertise used to do that. The concern, like you bring up, is if you want to cut and paste an operation to another location or another business or another company, how do you know that what’s done in that machine is going to produce a good part or a good assembly if you don’t have the technical know-how on site?
JM: That knowledge is expensive. It takes time. It takes a lot of effort to develop in house. Of course, the goal is to say, Maybe the company that produces the printers can deliver that for the companies who are going to buy it. But that’s not always the case. Of course, the company that tried to do this, they’ve tried to produce standard building parameters, standard geometry, standard powder, but we still have, in some sense, issues dealing with when you’re going to different machines, you don’t get exactly the same behavior coming back even though the part and the material is all the same.
SM: I guess, with that, what all can be different? What translates machine to machine and part to part, and what’s different?
JM: It’s estimated, for better or worse, there are more than 100 different knobs you can turn on a metal printer. For somebody like me, I get excited because now I have 100 different things I could change. But maybe for somebody who’s in a company, this gives them a heart attack because it would be so difficult to make sure the consistency is there. That’s an open question. That is really the question of the day. We know there are major parameters that have very large effects.
For example: How powerful is the laser? How fast is it moving? What pattern on the surface is it making when it’s moving? But there are other parameters that come in. For example, a powder bed system where the laser is coming down and essentially melting all the powders together, there’s a gas that’s flowing at the same time that blows some of the excess powder and the vapors away. How fast is that gas flowing? Then you have other parameters. Where on the build plate are you? Or even parameters that deal with, for example, feeding speeds in some bed, that powder that’s coming in and how the powder gets spread, every single layer. A lot of these questions we do not have answers for. We know there are —
CB: It sounds a lot like welding, Ji. Why can’t we have speeds and feeds and voltage and current, which has power input to the weld zone, and we have a heat affected zone. We have tons of experience with this sort of thing in making fabrications. Why do we wring our hands on once we put it in this sort of system?
JM: We do know there are some of these parameters that are absolutely important. I think the question then becomes, as we start to push technology toward something that’s there to deliver final parts and there are now very stringent tolerances and requirement on what those parts can or cannot be. Even small changes or even small unpredictabilities in some of these parameters become very important. We have understanding now of what the so-called big parameters, what they do. But we are still seeing that, even with very good control of those, we still see some variations. Now these secondary parameters are no longer secondary because now some of these can begin to control what the final processing, whether you have a part that fails or whether you have a part that passes.
Some of these are things that the community’s working on. For example, there is a big effort from the National Institute of Standards and Technology, which is called the AM-Bench 2022. It used to be called 2021 because COVID delayed the entire thing by another year. What they were trying to do is they’d built a very well-controlled benchmark machine that allows you to tune every single little thing that can happen in its process. They have an organization of about 30 different universities and companies trying various combinations of these parameters and seeing what the effect on the final build is. These are things that are currently going on in the community. The effort is not finished, so we do not know what the final answer is, but they’re starting to push closer to understanding what are some of these parameters.
CB: Sammy and Ji, some of the things that Ji was just mentioning, all these different parameters, once we solve some of this and really figure out what’s important and find a way to make that transportable from machine to machine, from powder lot to powder lot, from company to company, and even from part to part, we still have to get the community, that is, the designer, materials engineer, the product owner for the product that the part’s going to go into, the regulating authority in the case, maybe, of the FAA for an aerospace application.
Everyone has to develop some experience and get comfortable with the process and how much margin is there once you design a part. If something goes a little bit wrong, are we still going to be okay? We have decades of experience with forging and casting and then machining or welding in terms of putting these very elaborate structures together. Once we embark upon trying to do it with additive, we just have to do the work as a team from start to finish to get comfortable with the process.
SM: Relating to that in the short term, how do you convince people, when you're trying to implement it in a company, in a — let’s say you use the example in aerospace. How do you tell them that it’s a good part when we don’t have those as many years of experience behind us?
CB: In a company that I was in, we decided to make a very mundane part with additive manufacturing. We made a titanium-6 aluminum-4 vanadium — a very garden-variety Ti-6-4 bracket — for a duct, to hold a duct in place in an airplane. There wasn’t a lot of stress, it wasn’t really subject to a lot of fatigue or aggressive environment, didn’t have to carry, like I said, a big load. But what we wanted to do was demonstrate that we could repeatably make a part that would conform to the dimensions and the strength and the ductility and the other requirements that the designers wanted and the Type Certificate holder — that is, the airplane manufacturer — as well as the regulating authority, whether that’s the FAA or EASA — that everyone could see it in action with a very low risk involved. We successfully did that, but for a metallurgist, we’re like, “We knew that would work.” But we have to keep in mind that everybody has to see the full picture.
JM: Depends on what application we are talking about. This can be a slow process. But I don't think this is a problem unique to additive. I’ve had colleagues working, let’s say, a major aerospace company. Anything new that goes onto any part of something that is critical for flight is subjected to almost a default attitude of skepticism. It takes some time for the designer to accept all of these different things in order for them to be able to trust this and incorporate this into the design. Sometimes some of that, I think, is just time, but we still have to demonstrate that the basic fundamentals are sound, so they can’t say, “Well, we’re not all of a sudden going to find accidents everywhere as we start trying to do these things.”
CB: That’s right, Ji. A lot of people grew up with the philosophy that — the sarcastic philosophy that the first thing you hear about a new material or a new process will be the best thing. The rest of the news that you get along the way will be racked with disappointment.
JM: I think we’re guilty of that as scientists. We usually publish our best results, right?
CB: Right.
SM: With that, what issues and challenges remain for additive manufacturing?
JM: There are, I think, several different parts to it, and there are obviously challenges associated with acceptance. What I would talk about instead is really on the technical challenges, our technical challenges. We’ve discussed a lot about certification. The big problem with certification is that, while you print the same stuff on different machines, you sometimes get different results, and it’s not very well explained why. I think that is — the second part of this is really where it’s a problem. If there are some variabilities, okay, we can understand where those variabilities are coming from and we can deal with it. We can design our part around it. But the fact that we don’t always understand why there are such variabilities, why are the strengths different, that creates a lot of uncertainty and risk in the whole entire process. As a lot of researchers are trying to answer this question of why are there differences, why are some of these things different, we’re starting to realize at the same time that one of the reasons is we don’t really fully understand the science behind this whole entire process.
The last three or four years, a lot of additive work has shifted toward the fundamental understanding of this process. There’s a joke that, in the early days of additive, where we’re just doing serial welding, so the whole entire process is just welding and there are no new sciences involved anywhere within this process. I think in the last couple years, with some of the work coming out of Lawrence Livermore National Laboratory and Argonne National Laboratory, we’re starting to see there are, in fact, new sciences involved with this. … New sciences in how particles, the physics of particles as it’s being energized by the laser.
Many of these new studies have focused on, do we really truly understand how this process works? Of course, not everything is going to be important. You don’t want to be simulating particle level for everything you do. But I think what comes out of this is to have a better understanding of — we talk about all these different parameters. Which ones are the ones that are most likely to be important? If we can identify that, then perhaps we can engineer a system in such a way that minimizes some of these effects. This is one of the challenges that we currently have, is to be able to figure out how to reduce some of these variabilities and understand where it’s coming from.
On the production side, there are also challenges with building parts — building parts bigger, building parts faster, building parts better. This saying where you have to hit at least two of those three things as a new technology as we continue.
CB: From a business perspective, I think we’re all cheering along the material scientists and the folks trying to get at what are the real driving forces for how to design the right composition of matter, the right process path to build a good part. Because we feel like if we get to that answer, or at least enough of a design and process space so that we can feel the system that looks a lot like CNC machines — numerically controlled machines for making a subtractive part from a forging or a casting. If we can have that technology level on the work floor, capable of making parts with additive manufacturing, then I think we could be successful on a wide scale. Until we get to that point, where it’s that much of a, let’s say, commodity or that much of a capability required to make good parts, we will struggle to get wide acceptance. We need that sort of capability so that the machine knows what to do for this alloy, for this type of part, and that the operator knows how to program it to get the part out the other side.
JM: One last thing that’s currently, for certain applications, is that much of the work in additive is really done on the properties side, on a very basic level. Meaning we’re looking at mechanical properties — how strong something is. Maybe the deepest we go is how good is something in fatigue? Of course, for a lot of applications, this audience will understand this very well. For example, how does a corrosion behave or [how do] properties behave in additive manufactured materials? It’s not the same as what you see in conventional. In fact, there’s a lot of discussion in these areas right now. People cannot agree with each other. Lack of knowledge in the secondary properties is still something that’s an ongoing challenge as well.
SM: With all the challenges that we’re still trying to address, once we’re successful to rein in some of those, where do you see additive manufacturing’s future state? Where would we like to get to? What is the perfect usage of it or where we see this shaping industry as we move forward?
CB: In terms of shaping industry, I think I was trying to make the point earlier that if we can get this level of technology and workforce capability required to be on par with CNC machining, and that a company can put machines on their floor next to their assembly line or their distribution center and we can really enjoy distributed manufacturing and vertical integration at that scale, I think that is going to be very interesting to companies.
As a former CEO of GE appliances, I wanted my 1100 service vans to have 3D printers in the back of them, and when they show up at a consumer’s house and an appliance needs a part, they can print that part instead of waiting a couple of days to get a part from a warehouse. That’s the sort of thing where we’re not there today. We’re not capable of that. But maybe one day we’ll have the speed and the technology and the transportability of that technology to be able to do something like that.
JM: From a materials science point of view, I think additive presents a very interesting opportunity for designing new materials. A few of the things we touched on already is the fact that you have a bunch of knobs, and each of the knobs can allow you to change the material behavior in such a way. You can change the microstructure. You can change perhaps a defect structure. That gives you a very fine-tuned way of putting in material properties in places that you normally did not have the resolution to do. We’re starting to see some of these examples. There are a lot of interesting work looking at printing parts with the same material, but by changing the way you process at different locations of the part, you can create one part that has different material properties at different locations. That’s something that’s been very difficult to do with conventional techniques.
But I think, more interestingly in some cases, we find that the material itself is improved when you subject it to the conditions of additive manufacturing. There’s been a series of papers that came out recently reporting on stainless still. When we additive manufacture stainless steel, 316, very basic grade stainless steel, the steel comes out both stronger and more ductile compared to their conventionally processed counterparts. This is sort of a holy grail in developing new materials. You want both. You want… strength and ductility, but now with this new technique, you’ve done it without doing anything. Why? And how can we harness these things? I think these are questions that the materials science community is trying to get their head around, to be able to then use this as a new way of making better materials as a whole.
CB: Yes, that’s really exciting. One of the things you’ll find different between Ji and me, Sammy, is that he likes those 100 knobs and the ability to turn all or some at his discretion. What I’d really like to see are fewer knobs, like three knobs, and maybe one of them isn’t even connected to anything, so that we can have some consistency on the production end of the process.
SM: It would really depend on the application, then, on what you're looking for with — what you just said, Chip — with more knobs versus fewer knobs, right? The assembly line approach versus the what can we do and expand the horizon?
CB: Right, absolutely. There’s R&D, and there’s new product development, and there’s the ability to serve existing customers with the products that we’re making and see if we can’t make more of the product ourselves.
SM: I think we’re about ready to wrap up. Do you all have any final thoughts before we conclude?
CB: I’d just say that I believe that there are exciting prospects for the future in additive manufacturing. There’s good science to be done. There’s great engineering yet to be performed. So I’d encourage everyone to stay tuned, and if you have any interest in the subject matter at all, jump in and help. We could use the help guarding the future path.
JM: And I think additive’s here to stay. It’s not going to be a smooth path going straight up. We’re going to have a lot of bumps, and we’re seeing some of the bumps already. But I think this is a technology for the future, and it has some really interesting things to offer for just about everyone, both the kids in the room like me, and then the adults will have to come and tell us to stop doing all these creative things, like Chip.
SM: Thank you, Ji and Chip. If you want to learn more about the research they’re doing, please go and visit University of Virginia’s Material Science and Engineering page. For listeners that want to learn more about the corrosion of additively manufactured metals and alloys, you can find a number of articles available on www.corrosionjournal.org on that topic.
[closing statements]