Afternoon, everyone. My name is Guha Manogharan, I’m an associate professor of mechanical engineering at Penn State, and I also co-direct over 3D printing center called Sim 3D. I’m on the other end of the spectrum from most of you, I’m here with a problem, not with a solution, and this a, a problem that I think a lot of you might be interested in helping solve, which is on multimaterial laser powder bed fusion. So this my partner in crime is Ajit Piser from Imperial College London, who couldn’t make it here. So what we’re going to talk about today is a new processing method, and all our stumbling blocks in terms of design, a different lens scapes and opportunities.

0:37 So, with that so the process snapshot is on the left, a cool looking micro CT of TPMS gyroid with copper, bronze, and steel, on the right and a snapshot of multimaterial powder deposition on a single layer, as an example. So before I go further, I want to acknowledge Ajit and his student J Wank, and my own student Jaclyn Griffi, who have been fantastic in taking on a new manufacturing technology, and, you know, going through all the struggling pains, which is not which is always fun.

1:08 So just taking a step back to give you a quick summary of what SIM 3D is. So we are we are an AR manufacturing center, originally started with Dara funding at Penn State. We have upla research lab, as well as College of Engineering, and Earth and Mineral Science. We are about 25 to 30 staff and 45 faculty who are using the facility, and we do all kinds of work in narrative manufacturing, all the way from design to processing and testing, and so forth, across the MRL and TRL spectrum. And if any of you are ever going to be in the Penn State region, you have you have a, you know, open invitation to reach out to me, I allow to have you come and see our arity manufacturing center. And of interest of these days, we are particularly interested in working in refractory alloys, you know, high temperature ceramics, high temperature polymers, and new designs and methods, and so forth.

2:00 So, with that, here is another quick snapshot of, you know, our capabilities. I think everything we can we would need, in terms of laser powder bed, electron we used to have an electron beam powder bed as well. Now we also have coold spray, binder cheting, and and the list goes on and on. So, if quite a comprehensive setup for both the academic aspect of lower TRL research, as well as having full-time applied research lab staff who take that to transition to TRL. So that’s been that’s been a lot of fun, and we do have some that could be potentially classified.

2:34 So I also wear a different ad, which is a PI of my own lab, in addition to co-directing the center. So we do leverage a lot of the design, and we do predominantly experimental manufacturing work, but with with an emphasis on biomedical applications, for direct metal AM, DLP, for a lot of the orthotics. And if time permits, I will get into 3D sand printing. If not if any of you are interested in leveraging design, designed for sand casting, you might wonder why, because now we can 3D print sand moldes. So, so a lot of the gating, and all of the, you know, design challenges that foundaries have known for decades, could be completely eliminated.

3:11 So the one that you see at the back, you know, the city that you see, we’ve shown consistently that we reduce casting defects by over 90% as cast, and flexual strength improvement of somewhere between 15 to 20%, depending on whether it’s cast ion or nickel aluminum bronze, just by reducing the turbulence of the casting, you don’t have to have a straight sprew. So anyways, not to digress there, so that’s like a quick snapshot.

3:35 And I Al also, I was talking to some of you know, few of you yesterday during the networking, and we’re talking about how do we make sure what we teach in the teach in the course, or teach in the, you know, in a university, is relevant to what they will experience in industry. So we have a masters in artive manufacturing and design, I particularly want to highlight the Anson lab portion, which is which is the fun part, in my opinion, because I’m an AM guy. So students take all the design, processing, engineering course, and in my course, if typically we have a case study, where these are grad courses, you will only get a grade if you successfully finish the project, or you get a deferred grade to buy more time.

4:09 And I added this slide because somebody asked about how do you make sure those who are learning DfAM in school is also able to understand its industrial applications. So here is one example, so we always choose a project this, like, a, you know, a $40 Amazon air engine, but you have to redesign it 400, am incorporate additional fun functionality, and you get a grade only if you have the same RPM or better. So, so I think I so we, we, we do these kinds of stuff, when it comes to, you know, DfAM and stuff, it’s been quite, quite, quite entertaining, and quite U rewarding, I should say.

4:47 Here is a quick snapshot of we are just doing some testing on hydrogen fuel based turbine engine component that is defamed. I cannot show the exact geometry, but I thought I’ll show you a much cooler looking experimental results, and so forth.

7:37 So, with that, I want to get into why we got into this old multimaterial laser powder. But we originally started with looking at meta biomaterials, so can we achieve the biom- biomechanical performance, as well as the surface topology of human bone, that’s what got us started in this in this path.

7:56 So we started looking into different bone sites, you know, looking at different geometries. We started looking into, you know, which of the self-supporting geometry should be most efficient in being able to achieve all properties of structural importance, surface importance for bone growth, as well as patient conformity, right? So we started looking into that, and we, we, we started learning quite a bit very quickly, depending on which anatomy sites, you can prefer one type of TPMS over the other, because you do get very distinct ductile versus brittle transition behavior, just under static. And now we, we have published this also on fatigue, where you can see, you know, different sections have sorry, different types of sheets have linear behavior with, you know, normal yield versus, you know, abrupt britt- brittle behavior, and so forth.

And again, I want to go back to the, the, the, the motivation for all of this, right? So what you see there is just a, you know, a schematic of what a cortical bone and, you know, tabular or cancellous bone would be. So can we get the best of both, when you when we go from, you know, weing, you know, different PPMS, which people, others have also done, since we since we did this.

7:56 So we started looking into what is the effect of po size of certain types of TPMS sheets, and its and its porosity, and it’s it’s its mechanical properties. So it was it was quite interesting to see, also, I mean, I’m I am not presenting it here, but we also saw quite interesting results when it when we did the wet lab test in all of this, and we wanted to go further into looking into actual fatigue behavior as well.

19:49 And what we found was the distinct fracture mechanism. These, these are all Ti64, by the way, we pretty, pretty similar to what we would observe in natural occurring bone. This is signature paper in this field from, you know, 18 years ago. But I think the idea of being able to manage your eeld behavior, and also your fracture behavior, was was was quite was quite interesting to see how we can emulate that kind of behavior, also for biomedical implants, because stress shielding is a is the number one problem, as, as, as a lot of you would know.

So, with that, we started thinking of, you know, what, what, what are the alternatives, because, you know, to avoid s shielding and so forth. And the next time you walk down, you know, the supplemental section of your grocery, local grocery store, you’ll see a lot of zinc and magnesium bone supplements, they’re really, really good for bone growth. So we started looking into zinc and magnesium, and we notice that they’re not strong enough, like, you know, stainless steel or TI 64, so you cannot make them out of the whole implant, but they are really, really good for o generation.

So that got us into looking into what if we can do multimaterial, as this is one of the application domains that, that, that, that drove us. And so we, we, we got we, you know, we got into work with AOS out of Belgium, who originally developed this for powder die filling applications, but then we were looking at this from laser powder bed standpoint. And I’ll briefly explain, in a few minutes, how it works, kind of like a different thought process of multimaterial deposition in a single layer, and not just across layer. And we have been able to achieve somewhere between 150 to 300 micron pixel resolution, and being able to use it for a wide range of application, including eat exchangers, where you can have a really hot side, really cold side, or corrosion resistance, U marine gade alloys, with, you know, structural steel.

So we, we, we looked at this as more as a multimaterial AM that enables local functionality, because what people have been doing for a while with material jetting, with the polymers, and even with some, you know, with the applications you can see, several of them here, we were kind of looking at it from metal. And that’s what got us in U working with Duann, in terms of exploring I’m really tired of doing STL files for all of this because we would like to be able to assign local processing conditions to doing multimaterial laser powder bed.

So there are wide range of applications you can think of, you know, air electrodes, biomedical implants, eat exchangers, and and the list goes on and on. So, but that’s all like down the line, because we have quickly learned there are lot of, lot of problems that we need to solve, in order to be printing with one material and qualifying it in LPBF. For those of you have done it, you know, is is a pain, trying to doing it for multiple materials is even worse, but I think that that’s what makes it fun, and that’s where I think, again, Sim 3D having the balance of fundamental research and uplate research has been very helpful.

So, so this is the system that we have in our Pro X 320. The reason we did that was, we have already used the system with a wide range of insute monitoring capabilities for single material, so we didn’t want to rein the will we know that we’ll be giving up a little bit on the build volume, but we still decided to go with that, and I think we are the first one, as far as we know, in the US, to have this capabilities, third one worldwide, and probably the second one worldwide with three materials. And you might wonder why three materials, because the powder recycling is going to be challenging, so we chose a third material that could basically be a filler material that you can basically, you know, melt a wall around, you know, that that U follows the contour of your part.

So what you see there, at the bottom right, is basically the, the, the video of the system working. Basically, you have the oper on top of the rollers, or drums, as they call it, and you have mesh that are around the drum, and each drum is associated with, you know, one of the materials, and the mesh of the M of the drum refers to your pixel resolution. So what you have in figure A, at the top left, is basically cross-section of the drum, and what you have in figure B is the slicing of, for example, in this case, PSU, and the student who did this went to Florida for undergrad, and clearly sneaked in the Florida colors, as you can, because there is still pen shed color of white, and white and blue. But so you can see the slice information for materials A, B, and C, corresponding to the three drums.

And on top of the drums you have opers, you basically you’re using negative pressure, like, think of it like a clock, at 12:00 you’re using negative pressure, where the circumference of the Dr of, of the each drum is equal to your Y, and the height of your drum is equal to your X of your bill plate. So you’re using negative pressure to hold on to the powder with all of the mesh, and right when your particular pixel is at 6:00, is tangential to your bill plate, and you synchronize it based on your RPM of your drum and your travel speed, to depositing specific powder in specific walk.

This is really, really exciting, and it still is really exciting to get get into this, because I think there are a lot of opportunities here. We have taken this system, and we are building our own jetting system on top of it for ey temperature ceramics, where one of the materials could be dissolvable, or you could just burn it during curing. And we also exploring this system for using nano composits, for where one of this material could be ceramic that you can embed nanoceramics in, you know, in your melt poool. So, again, low TRL, low MRL, admittedly, and I have no quals about it, because we have to be honest with the technology and figure out what where the challenges are.

So if any of this is interesting to you, please do reach out to us, because we would we are quickly learning there are a lot of unresolved, and there are a lot of interesting challenges. And I also want to admit there are other methods of others that have also come up with different methods of doing multiple materials in a laser powder bed. Some, can you can only do it across layers, some, you know, which are based on, like, suctioning of powders and depositing second powder, can do it within the layer.

So the applications is again, I, I think it’s quite interesting, but what I would like for this particular audience is to to get in mind, and again, looking at the really nice breadth and depth of presentation for the last two days, is we need to solve problems at multiple lens scales. So, for example, if you’re doing multimaterial testing, and again we are going back to functionally graded materials, DED and welding, and all their test standards, which we know does not apply directly to laser powder bed, so we have to reinvent a lot of there.

And u, I, I I’m, I’m mechanical manufacturing engineer by trade, and I did okay in metallurgy, but it’s coming back and biting me, where I really need to go back and tackle a lot of the fundamentals of phase transformation issues, so that we avoid inter metallics. So, so far, I think leaning a lot on laser cladding and functionally graded materials in DED world has helped us, but but what I think is, again, this is kind of an open invitation, for, for, for, for, for CDFAM community, to look at different lens scales.

So, whether, you know, you do calfed or ICME for multimaterial, not only looking at material compatibility, but looking at what is should be the volumetric energy density across the transition, like, from the transition boundary into your bulk part, is going to have a major influence on your structure and property.

For example, in functionally graded materials, you can gradually grade it. Here, we do not have grad- gradual gradation, unless we, we are trying to trick the system into having, like, you know, five different layers of different powders, and then going with the meltpool. But it’s it’s a bandid to a bandid to the overall problem.

The same thing goes for laser powder interactions, where we need to understand the meltpool mixing, and I have some examples to talk about it. And the one that I think would be the ripe for, for, for, for, for, for this group, would be the top, you know, how do we go back to optimization, and let alone the uncertainty quantification with defect formation, and so forth.

So, again, right now we are focusing more on the multimaterial bone scaffolds, some battery electrodes, and surface fluidic applications, because what we know of single material laser powder bed, we have to now calibrate it for emissivity of different materials and understand its MP behavior. And same thing with micro C, when we are looking at grayscale and ounce F unit for different materials, how do we go back to micro C, and non-destructive testing using ultrasound or other methods, and mechanical testing with multimaterial. And so we are doing some intermittent micro CD, again, kind of like a low TRL stuff.

But I think I’ve been really inspired with a lot of the AI talks today, because, again, I think we have to have more of a different approach to what we have done classically here, because based on the alloy and the build orientation, even the slightest of part we have with multi material, when we reoriented in your bill plate, you completely, drastically change your thermal conditions, and that does not exist, as to the best of our knowledge, process simulation tools for multimaterial laser powder bed.

Because, for example, in this case, when we did copper tin and low carbon steel, we found a lot of nanocracks around 300 to 400 microns away from the transition boundaries. And I think this is all again, we did not optimize the process parameters, what we are what we postulate is, lot of the bron is eat, you know, sucking a lot of the heat from the steel, so we are getting a lot of the residual stress effect.

And we are kind of now dabbling a little bit more into how do we come up with mechanical interfaces, because, depending on the density of the mel poool, you know, you could have bones and mpol mixing issues as well. And we are looking at some of the mechanical interfaces that we could try to have both mechanical and metallurgical bonding, which is which is quite interesting. And lot of you might be quite familiar with this geometry, the MBB.

So we, we, we started looking into, okay, let’s take the classical problem and apply multi, you know, adopted for topology optimization for multimaterial. In this case it was U multiobjective, where we wanted to optimize it both for mechanical loading and for thermal dissipation. So that’s the reason why we have copper, chromium, and 316 stainless steel, that you don’t typically see public, you know, printed together.

So, so, so, so we use the alternating active phas algorithm, all of you know this, and I won’t go into the nuts and bolts, but U, what I do want to point out is, at this stage and after what, what, like, what, what, what we learned after this particular project is and again, we should have taken advantage of the third thir material, which could have been a little bit more conducive metallurgically speaking, for material A and material B. The reason we did not was, we thought we should not make the optimization routine complex to begin with.

So we started with just these two materials, where we, you know, where we used our solver to identify the optimal MBB for both steel and copper, for multiobjective. And we, again, as, as, as expected, we did we did find improvements in thermal conductivity of the overall structure, we definitely satisfied what was required for the mechanical loading conditions, and we verified this with FAA. I won’t go into all of the details, but and we also did DAC with all of this.

And I, again, I think it’s it’s a classical problem, it may or may not be very industry relevant, but again, we wanted to start at some point, and we validated the FAA results with DAC and so forth. But again, I think I think the main takeaway, as, as, for, for, for, for, for, for all of you, is at every lens scale there is opportunity to go back to think about computational design of material assignments, along with concurrent design of material structure and the processing conditions, and also thinking of how to leverage, you know, computational tools, when we trying to validate the Institute monitoring and our micro C information, and so forth.

So, again, I won’t go into a lot of this detail, but again, FAA is good. So, but I, I think there is a lot more opportunities here. For example, right now we finished a clinical study in collaboration with Upen and and University of Delaware, on textured zinc, now we would like to explore the zinc with stainless steel.

So I, I would like to leave this as the last slide, as, as more as food for thought. So everything we know for laser powder bed, and what we have learned from functionally graded materials in DED, now it is an opportunity to kind of reimagine it. And that’s why, like, you know, I, I, I reached out to Danan and said, I’m not going to come here saying this is wow, look at what I did. It’s going to be like, hey, I think we need we need collaboration from the computational design community, and we are very collaborative. So that was the reason why I’m here, and please do grab me during the breaks. I would like to talk to you if you are interested in this. And thank you.