 Well, good afternoon. I'm Sang Kim, the Island-Feld head of the Davidson School of Chemical Engineering and today we have a very special event and a very special speaker and I would like to call our interim dean of the College of Engineering Mark Lundstrom to do the introductions. Mark. Thank you Sang. Good afternoon everyone. So welcome to our very first engineering frontiers lecture. So you know our academic units, they frequently host distinguished visitors who come and interact with our students and our faculty. This new series was created to provide a college-wide venue for distinguished speakers addressing topics that are both important and timely and that's certainly the case for our topic today, semiconductors. So you all know about the chip shortage. Chips are being called the fuel of the 21st century. Not only are we not manufacturing enough chips, we're losing our lead in some parts of this critical technology that we invented and we've realized that we don't have the engineering and the scientific workforce to regain our lead. So it reminds me of Sputnik and the race to the moon. We rose to the challenge at that time and the nation is now rising to this new challenge. These are really exciting and important times for semiconductors. Now I first met our speaker David Lee shortly before I stepped in as interim dean. Our previous dean, Meng Cheng, asked me to summarize our semiconductor initiatives to the Engineering Leadership Council. So this is a group of CEOs and retired CEOs, strong supporters of the college and people who provide us with strategic advice. David is a member of that council. So I gave my presentation on semiconductors and we had a discussion and I quickly learned that David knows an awful lot about semiconductors. We actually had a follow-up phone call for an hour or more and I've been in semiconductors my entire career, you know, following it. Brief time in industry, but most of the time following it from academia. And it was fascinating to me to get an insider's view of this of this industry and it's going to be a real treat to hear him talk today. So we're lucky to have him with us today and to talk about this important topic. After his talk and after the Q&A, there will be a panel discussion which is not directly related to semiconductors, but it's important for any students who are preparing for a career in semiconductors or any other industry. So I would encourage you to stay around for the panel discussion. So David Lee is recently retired as the president and CEO and director of Cabot Microelectronics Corporation, CMC Materials, which is a global material supplier to the semiconductor industry. Before assuming that role in 2015, Mr. Lee served as vice president of the Asia-Pacific region for the company for seven years. And prior to that, he held various senior leadership and management positions since he joined the company in 1997. Including among those positions were regional and product line management, operations, supply chain and investor relations. So he really has a very broad set of experiences and perspective on this industry. Mr. Lee earned a bachelor of science degree from our chemical engineering department at here at Purdue and an MBA from Northwestern University. Please join me in welcoming David Lee. Thank you, thank you. All right, so get my slides in order here. All right, well, thank you for that kind introduction and it's great to be here. It's an honor and I'm here to talk a little bit about semiconductors, the industry that I spent around three decades of my career working in and obviously a very exciting industry to be involved in, especially today. I just want to introduce myself and my family a little bit more. As Dean Luntram mentioned, I'm a proud boiler maker. I graduated in 1995 and I can honestly say I would not have been here if it wasn't for Purdue. So I'm always grateful for the great education, the connections and wisdom that I got from Purdue. And again, I wouldn't be here today if it wasn't for Purdue. How I came to choose chemical engineering is an interesting topic. You see my mom and dad in the middle, my dad is a PhD chemical engineer, my mom is a PhD statistician and my sister is also a PhD chemical engineer. So the idea of free will may have been a bit flexible in my household, but anyway, I think I chose the right path. And as Mark mentioned, I was the CEO of CMC Materials, one of the largest semiconductor materials companies for the last eight years and currently serve on a number of boards and advisory committees. And I'm based in Singapore. So this weather, although I grew up in Chicago, is a little bit colder than I'm used to. But again, it's great to be here. Okay, so today's focus, I just want to give everyone first an overall feeling of the importance of semiconductors. And we're going to do a little audience participation to get us started with that. I'll talk a little bit about this, the fab, the semiconductor fabrication process, and there's a great video from our friends at Infineon, an important semiconductor producer. I'm going to show that really I think brings to life the semiconductor fabrication process. Semiconductors have also been in headline news quite a bit recently, so I'm going to talk a little bit about the global dynamics and the implications of why semiconductors and where they're produced is so important and then end with trends and opportunities. The one thing I would say is and simple message of this lecture is I want to convince all of you that there's never been a better time to get involved in the semiconductor industry. I've known the semiconductor industry and been part of it for almost 30 years. I can tell you right now is one of the most exciting times for all of you to consider exploring a career in semiconductor. And I would say there's no better place to explore that than Purdue University. If you think about what has been announced recently, whether it's the partnership with Sky Water or the partnership with Media Tech, I hope that many of you have an opportunity to actually step foot in the fab once it's constructed or be part of chip design. Fantastic opportunities here at Purdue University and I think it's a really exciting time to be part of the industry overall. Okay, so the one thing I would say before I go on to this next part which involves all of you taking out your devices is as I was getting ready to give this lecture, I was studying up on what has been taught and said about semiconductors before. And one of the first online courses I came across was from Dean Lunchstrom who had a series of semiconductor lectures. It's available on YouTube. Fantastic set of online classes. I would highly encourage all of you to check them out. So that's the first thing I ran into, but different than Dean Lunchstrom's courses, there will not be any tests or quizzes associated with this lecture. Okay, but there will be some audience participation. So if you'll indulge me, I just want to make sure the technology works. So all of you have a device either phone or laptop or iPad. If you go to the site menti.com and use that code or you can scan this QR code. And what I'd like you to do is think about two or three products in your life that are enabled by semiconductors. Okay, so let's take a moment to do that. So here is a word cloud real time of the input that all of you are putting in. Let's see what we have. This is always interesting to see. I did this last week at a panel I was involved in. The word cloud, the larger words are the ones that are represented more frequently. And then there are obviously the smaller ones as well. Okay, great. I think we're settling in on 96, almost 100 participants. Phone, car, laptop. Okay, yep. I think those are the ones. Xbox. All right, I see that. I hear that. So these are just some of the things that we think about when we think about semiconductors. Let me see if I can go back to my presentation. And so I think we caught, this is a smart group, right? Not surprisingly, you have most of those key devices. But I would also say semiconductors are extremely important for a number of other industries. For example, last week I was in a panel with, I think I was the only private sector person there. Everyone else was from the government and they were focused on semiconductors. Why? Because they see semiconductors as enabling the next generation of military technology, defense technology. So hypersonics or next generation jet fighters. So that's really important. I was in a panel of other CEOs that are really excited about the next generation of industrial automation. So AI and robotics. You can see a picture of that there as well. I think automotive was on there. That's a key driver for next generation semiconductors. If you get in a car today, the level of semiconductor content is only going up. And as we approach the EV generation, that's only going to increase exponentially. So those are some examples. Some of you may have seen the most recent release from Tesla, their first foray into robotics. That's their robot. Personally, I wasn't that impressed, but they have a lot of other inventions. Some of you may be sporting the new iPhone, the iPhone 14. If you notice, Apple's actually actively, when they launch a new device, they're highlighting the semiconductor content or semiconductor chip because it's customized. So they're using marketing like the A15 chip or the M1 chip, just showing how important semiconductors are to enabling their next generation devices. Not as popular as it was last year, but cryptocurrency, also a huge driver. And I would say, especially based in Singapore, there's a lot of activity, not just in cryptocurrency, but in blockchain technology, which I think has a lot of interesting applications for future use and, of course, driven by semiconductors. And then the last one I think maybe we take for granted is all of the applications, whether it's Amazon or Netflix, that we use all the time, are very reliant on semiconductor content and the cloud. So memory is really, really driving a tremendous amount of semiconductor content. In fact, I saw an interesting statistic the other day to say that we have now crossed over the point where there's more data being produced by machines than there are humans. So if you think about AI and supercomputing, there's more data being generated today by machines and stored in the cloud and processed and analyzed, and there are actually by humans. So fantastic breakthroughs enabled by semiconductor content. Just to give a little bit more detail, today, this is a slide by McKinsey that I really like. It breaks down the semiconductor industry into the different segments. So today, semiconductor, total value of the industry is around $600 billion. It's expected to increase to over $1 trillion over the next eight years and driven by those industries that I mentioned. So automotive, really key focus area for semiconductors, data storage, all this memory that we're, you know, all our pictures, videos, as well as machine generated data, as well as wireless industries. So 5G, all the infrastructure needed to support 5G and the next generation and all the related devices. So it's an extremely healthy, vibrant and growing industry expected to grow to $1 trillion by 2030. Shifting gears a little bit, one of the most famous, I would say, it's not really a law, it's more of an observation, it's called Moore's Law. Many of you may be familiar with this. This was actually an observation coined by one of the original founders of Intel, and this is a slide from Intel, our good friends from Intel. And this person, Gordon Moore, theorized back in, I think it was the 60s, that the number of transistors per integrated circuit would roughly double every two years. Okay, so that prediction has proved to be true over the next, over the last 50, 60 years, and although the industry has, you know, continued to find ways to pack more transistors in, that's not the only function or feature that we think about for semiconductors. There's also power consumption, cost, but I think, nevertheless, the industry continues to find ways to innovate and push forward. And there's all kinds of new technologies. There's a few listed there, advanced packaging and so on. But I think the industry continues to be extremely vibrant from an innovation standpoint. So let's think about Moore's Law in a different context. If you were to apply Moore's Law to a different industry, say automotive, and you would just apply the improvements that have been made in speed, power and cost to the automotive industry over the last 30 years, it would result in a Rolls-Royce that costs $40, could circle the globe eight times on one gallon of gas, and have a top speed of 2.4 million miles per hour. That's the kind of car I would like to drive. But that gives you a little bit of context about the innovation and improvements that have been made in the semiconductor industry that we, I think, take for granted when we have these devices. These devices are more powerful than the supercomputers of just 10 years ago. So all of that is enabled by the innovation and continued improvements that have been made in the semiconductor industry. Pretty impressive. Okay, so I want to shift gears a little bit and give a quick, high-level overview of how semiconductors are produced. Okay, and again, I would highly encourage all of you that want more detail to go to Dean Lunchstrom's series of semiconductor fabrication. Excellent. But I just want to break it down into the key components and then we'll show the video. The video is around 10 minutes long and I think it does an excellent job of highlighting semiconductor fabrication processes. But I want to just break it into the key parts. The first part is if you think about semiconductors, semiconductors are primarily made of silicon. What is silicon made of? It's made of sand. So the starting component for semiconductors is sand, which is purified and made into a crystalline bowl and then sliced into wafers. So that's a key component and a starting point for producing semiconductors. Next, on the top, you'll see there's a lot of innovation and thought needed to do the actual chip design. How does the chip design get put onto the wafer? It's through a process called lithography, which is an amazing technology and one of the critical process steps into producing a semiconductor today. You'll see that in the video coming up. And then there's a series of polishing, depositing and etching steps and my personal favorite, which we'll dive into a little bit more, planarization, CMP. So my company, CMP Materials, was the world's leading provider of CMP Materials and I'm going to take a moment after the video to go through some of the innovation that was needed just to support that one step of the semiconductor process. Okay, so let's show the video, if you don't mind, our friends in the IT group doing a fantastic job showing this video. Thank you. All chips start out with a very simple raw material, sand. Sand is primarily made up of silicon dioxide or silica. Silicon is the second most abundant element in the Earth's crust but is only ever found as a compound with oxygen. Complex chemical and physical processes are required to ensure that silicon crystals meet the high production standards that apply to chips. To convert silica sand to silicon, the sand is combined with carbon and heated to an extremely high temperature to remove the oxygen. A number of other steps are required to create the finished product, namely an extremely pure monocrystalline silicon ingot called a bull with only one impurity atom for every 10 million silicon atoms. Silicon bulls are fabricated in a range of different diameters. The most common sizes are 150, 200 and 300 millimeters. Wafers with large diameters offer more space for chips. Extremely thin wafers are then cut from the silicon bulls using a special sawing technique. These wafers are the basic building blocks for subsequent chip production. Silicon is a semiconductor. This means it can conduct electricity and also act as an insulator. Its atomic structure looks like this. Every silicon atom has four outer electrons. There are no free charge carriers. As a result, the pure monocrystalline silicon is non-conductive at room temperature. To allow it to become conductive, small quantities of specific atoms are added as impurities to the wafer. These impurity atoms must have a number of outer electrons that is either one more or one less than that of silicon. Silicon is in the 14th group of the periodic table of elements. This means that elements in the 13th or 15th group have to be used in this process, referred to as doping. Boron and phosphorus atoms are the most suitable elements in these groups. They are very close to silicon on the periodic table and therefore have very similar properties. Phosphorus has five outer electrons. When it is inserted into the silicon crystal lattice, the fifth phosphorus electron can move freely. This means that the silicon phosphorus crystal is n-conductive. In contrast, boron atoms only have three outer electrons. When they are introduced into the silicon lattice, one silicon electron has nothing to bond to. This creates electron holes. The holes move through the crystal like positively charged particles, making the material p-conductive. Transistors are built on the p- and n-conductive layers that exist in a doped wafer. Transistors are the smallest control units in microchips. Their job is to control electric voltages and currents, and they are by far the most important components of electronic circuits. Every transistor on a chip contains p- and n-conductive layers made of silicon crystals. They also have an additional layer of silicon oxide, which acts as an insulator. A layer of electrically-conductive polysilicon is applied on top of this. Every transistor has three terminals. The middle one is attached to the gate, which is the electrically-conductive polysilicon. If an electrical charge is applied to only the two outer terminals, electricity cannot flow as the transistor is blocked. This changes, however, if an additional charge is applied to the middle terminal. Electrons from the p-layer are then pulled toward the middle terminal and accumulate at the area bordering the silicon crystal and the insulating gate oxide. A channel then forms underneath the gate between the islands of n-conductive material. Electrons can now flow through this channel. The electrical circuit is closed. In this way, the transistor can be switched back and forth between current-enable and disable, between zero and one, on and off. But how are these layers created on a wafer? The process to manufacture chips from a wafer starts with the layout and design phase. Highly complex chips are made up of billions of integrated and connected transistors, enabling sophisticated circuits such as microcontrollers and cryptochips to be built on a semiconductor surface, measuring just a few square millimeters in size. The sheer number of components calls for an in-depth design process. This entails defining the chip's functions, simulating its technical and physical properties, testing its functionality, and working out the individual transistor connections. Special design tools are used to draw up the plans for integrated circuits and develop a three-dimensional architecture of sandwiched layers. This blueprint is transferred to photo masks. Providing geometric images of the circuits, the photo masks are used as image templates during the subsequent chip fabrication process. To ensure that the microscopic structures of a chip are reproduced flawlessly, they have to be fabricated in a dust-free environment with stable temperature and humidity levels. In other words, they have to be made in a clean room. A clean room is a room in which no more than one particle of dust, larger than 0.5 micrometers, is permitted in around 10 liters of air. This is even cleaner than the air in an operating room. The ventilation, filtration, and supply systems in a clean room therefore have to be extremely sophisticated. Several million cubic meters of air are circulated every hour, and hundreds of air volume regulators maintain a constant air flow. Employees in these production areas have to abide by an extremely strict dress code. They are not permitted to smoke before work or wear any makeup or jewelry. Clean room production areas can only be accessed through a special airlock. Chips are built on a base wafer cut from a silicon bull. Depending on their size, several dozen or several thousand chips can be fabricated on one wafer. First of all, the surface of the wafer is oxidized in a high-temperature furnace operating at approximately 1,000 degrees Celsius to create a non-conductive layer. Then a photo-resist material is uniformly distributed on this non-conductive layer using centrifugal force. This coating process creates a light-sensitive layer. The wafer is then exposed to light through the photomask and special exposure machines known as steppers. During this process, coaster-sized areas of the chip template known as rectacles are used to transfer the complex geometric patterns of the circuit design to the silicon wafer. The exposed area of the chip pattern is developed, revealing the layer of oxide below. The unexposed part remains as is, protecting the layer of oxide. After this, the exposed layer of oxide is etched off in the areas that have been developed using wet or plasma etching. With plasma etching, special gases bond with the substrate to be removed in the reaction chamber. This enables microscopic layers to be removed in the windows that were exposed and developed in the previous step. Once the photo-resist residue has been stripped and the wafer has been cleaned, the wafer undergoes further oxidation. Electrically conductive polysilicon is deposited on this insulation layer. Then photo-resist is applied again and the wafer is exposed to light through the mask. The exposed photo-resist is stripped again. Now the polysilicon and the thin oxide layer are etched off. These two layers only remain intact in the center under the photo-resist. The next step is the doping process where impurity atoms are introduced into the exposed silicon. An ion implanter is used to shoot impurity atoms into the silicon. This changes the conductivity of the exposed silicon by fractions of a micrometer. After the photo-resist residue has been stripped, another oxide layer is applied. The wafer undergoes another cycle of applying photo-resist, exposure through the mask, and stripping. Contact holes are etched to provide access to the conductive layers, enabling the contacts and interconnections to be integrated in the wafer. This is done by depositing metal alloys onto the wafer in sputtering machines. Once again, the photo-resist and mask are applied. The unexposed strips remain as is after the etching process, providing a point of contact to the underlying layers. To give the insulation layer above the interconnections the smooth finish it requires, a chemical-mechanical process is used to polish away excess material with micrometer accuracy. These individual steps may be repeated multiple times in the fabrication process until the integrated circuit is complete. Depending on the size and type of chip, the wafer will now contain anything from several dozen to thousands of chips. Individual chips are usually sought... The end of the video, I think, is a few more minutes, but it's mostly a commercial for infinity. We'll keep going from there. I hope you gain an appreciation of the chip fabrication process and how complex and specific and specialized it is, as well as how much innovation is required in each step. My favorite step, of course, is the one that is circled, the planarization step, where they talked about using chemical and mechanical means to planarize the surface of a chip. The challenge of this process step is that you're literally removing silicon or metal selectively while not damaging the other aspects to a subatomic level. Again, I work best with using simple analogies. The way I would describe... This is our beloved football field. By the way, I was really pleased to see us prevail over Nebraska. My best friend is a Cornhusker fan, so we had an active text channel going this weekend, and it was great. But, okay, if you were to think about CMP on the scale of a football field, imagine putting two football fields side by side with one football field, one millimeter the grass, one millimeter higher than the other. The job of CMP would be to flatten that second football field within that one millimeter without damaging the surface, without touching the other football field, completely flat across the 100 yards. That's the job of CMP on a subatomic level, okay? So if I show a cross-section of an actual device, this is a cross-section of a wafer before CMP was really the mainstream way of making a chip. As you can see, there's layers of silicon and metal that are deposited on top of each other. And if you don't planarize the surface in between the next deposit layer, those deformities continue to add up. It's sort of like stacking credit cards on top of each other, right? You're not going to be able to stack many on top of each other. The bottom picture shows after CMP was introduced how much more efficient that you can pack the layers of silicon and metal on top of each other. So this has become really the standard way of making any advanced chip today. So it gives you a little bit of a feel for CMP. I thought the video was a little bit more towards the electrical engineering side of things, so I had to introduce a little bit of chemical engineering to show that I did learn something at my time at Purdue. Here's chemical mechanical planarization, the process, and there's sort of a schematic and then an actual picture. So what you see is a polishing platen that's rotating. The wafer carrier, so the wafer is put face down on a polyurethane pad, and then a CMP slurry is introduced, and that's how planarization happens. So it's really amazing that we're able to do this on a subatomic level. The CMP slurries are typically made of high-purity water, highly engineered particles, and then formulated chemistries. The CMP pads are highly engineered polyurethane materials with specific groove patterns to planarize the selected surface. I just want to give an example of the innovation that's needed, and this was some of the invention that happened at my company, CMC Materials, using novel chemistries to achieve breakthrough CMP performance. So here is an example of using a Syria particle. So Syria is a common particle used in CMP slurries. Why? Because Syria is very aggressive on silicon, very aggressive on glass. The problem with using Syria is that it produces a lot of scratches, and so if you have a scratch on a wafer, you might ruin the chip that you're trying to design. Through certain novel innovation, what we're able to do is coat the surface of this Syria to make it less prone to scratching the surface of the wafer. And you can see, I won't go into detail, but essentially the level of defects has been cut in half or even more through this surface modification. We call it a silane-modified shell. Similarly, using a different particle type colloidal silica, which is a good polishing particle but not as aggressive in removing by modifying the surface of that colloidal silica, we're able to achieve higher removal rates and lower erosion, which is another measure of defectivity on the particle. I just want to emphasize these just to give everyone a feel for the level of innovation needed on one particular step, of one particular material in the semiconductor process. And just to kind of wrap this all together, this is an announcement actually by one of Apple's, this was their last-generation chip. You can see they're highlighting this innovation in semiconductors as an enabling part of their next-generation devices. So this is the A14 chip. It's made with 5-nanometer technology, which is considered one of the most advanced technology nodes available today, 11.8 billion transistors, 24 billion contacts on this single chip. So you can see the level of innovation, and by the way, you can see the level of planarization that's needed in that cross-section of the chip for the A14 chip. So just to kind of wrap all these concepts together to give everyone a feel for what it takes to make one of the key chips for our iPhones today. Okay, so I want to shift gears a little bit, and again, we're going to use this app that we used before to talk about where semiconductors are produced today. And so I wanted to make this a small group activity. This is a great group of people. So maybe have groups of three or four, so get friendly with your neighbors, and talk about the following two questions. And I've listed the five major countries where semiconductor production happens today. U.S., China, Japan, Taiwan, South Korea, and then the rest of the world. And what I'd like you to do is again, go to this Menti app, you can scan the QR code or use that code on the bottom, and answer the two questions. One is, what do you think is the percent production by these five countries? And then the second question that we'll answer together is what is the percent production of advanced semiconductors that are produced by these five countries? Okay, so I'm going to let everyone gather and take a few minutes to answer these questions. I didn't want to show it first because there's a tendency like for the first mover, for everyone else to copy the first mover. So I wanted everyone to get a little bit of their own momentum going here. Okay, did everyone have a chance to get the answer or their thoughts? Okay, so here's the actual. Let's see if I can advance. So today, total semiconductor production around 20% is happening in Taiwan, 20% in South Korea. Japan is still significant, 16%. The rest of the world, about 16%. That's mostly Europe. China around 15% and U.S. at 12%. I would say if you look at the trend over time, the U.S. has declined year over year. So maybe 10 years ago, we were at 20%. Today we're at 12%. So significant decline in U.S. production of semiconductors. And we'll talk about that coming up. All right, so the second question is advanced semiconductor production by country. So again, hopefully this advances to the next screen and please input your best estimates by percent for these five different countries. Okay, how many responses? 61, all right, so we're settling at around 38%, 39%, 40% Taiwan, and then South Korea. Give everyone another few minutes to answer the question. So this is advanced semiconductor production. I call it 10 nanometers and below. That's sort of the designation used by most to talk about where advanced semiconductor production happens. Okay, all right, so this is the consensus. 37% Taiwan, 18% Korea, 13% China, and you can see the rest. Here's the actual. So 90% of advanced semiconductor production happens in Taiwan today. 9% happens in South Korea and 1% in the U.S. Everyone surprised? So there's a question, and we'll take questions at the end, but I heard that question enough. I think I heard the question was, does it account for memory, right? So great question. So this really is focused on logic. We can talk about memory, but memory production is, you typically wouldn't say 10 nanometers or something. You'd use how many layers of 3D NAND, et cetera. If you were to look at the memory split, the probably 70 plus percent is happening in South Korea. Samsung and SK Hynix are really leading. You have significant manufacturing in Japan with Toshiba and then the U.S. with Micron. Okay, so what are the implications of this? Let me switch back to my slides. And this is where I want to end. I think that we can clearly see, this is a picture of President Biden with a chip. One is you can clearly see, as we went through in the beginning, the criticality of semiconductors today. And no wonder why it is headline news and the subject of a lot of geopolitical tension. Because there's such an imbalance today where chip manufacturing happens, combined with the criticality of the technology. I had great conversations with Dean Mark and many others about the initiative that the U.S. has recently started along with Europe to really begin reshoring a lot of that capability. And I think Purdue is uniquely positioned to really help that initiative. But it's going to take a really sustained commitment to redevelop our semiconductor capabilities here in the U.S. You can see how daunting of a challenge 90% of advanced semiconductor production happens today in Taiwan. I would say around that dependency on Taiwan, obviously there's a lot of political tension around China and Taiwan. But put aside that for a second. I think for us to be reliant on Taiwan for 90% of advanced semiconductor production is in itself risky. Because for those of you that know Taiwan, you know it's subject to natural disasters like drought, like earthquake. And so for us to really have a little bit more of a resilient supply chain, I think is extremely important. The last point I would make is that the focus is going to be on advanced semiconductor production. Why do I bring that up? The reason why I bring that up is one, obviously there's a lot of focus on that technology and innovation that's important. But as you think about this whole spectrum of semiconductors that we need in our lives today, for example, I think Mark mentioned the shortage of automobiles. That's not necessarily happening because of shortage of advanced technology chips. That's happening because of shortage of legacy chips. So you might see those types of shortages persist in the market a little longer than you might expect because there's not a lot of investment happening today in that older technology of chips that are fueling autos and things like that. So what I think are the kind of implications and why chips will continue to be a really, really significant focus of the geopolitical environment. So just to conclude, I hope you gain an appreciation, one of the criticality of the technology of semiconductors, how much it drives in all of our lives today and how much we rely on it today without even thinking about it, right? Think about a day without your phone frightening, right? But all of that is based on the cloud, semiconductors, wireless infrastructure, all the applications that are being built around this ecosystem. I think geopolitical, the situation is only going to get more tense. This is my prediction. Just because semiconductors has become such a lightning rod for different countries, whether it's China or U.S. or Europe, we all recognize we need to be a little more self-reliant on our own supply chains for semiconductors. I think that's a very complex and a thoughtful strategy needed to address that. And then, of course, underlying is that strong growth and demand. I went through a slide that said the industry is expected to grow to $1 trillion. That's because of all the devices, the connected devices that we are going to enjoy in the future. So where are the opportunities? There's a lot of investment happening today. A lot of that's happening in the Midwest, what I'm happy to say. So Intel building their fab in Columbus, the skywater facility happening at Purdue, so many different investment opportunities, a really exciting time in the industry. There's continued need for innovation. I hope I impressed upon you just the level of innovation needed in one particular step of the semiconductor fabrication process. Think about that magnified exponentially for chip design, lithography, packaging, testing. All of that needs continued innovation that we've kind of come reliant on to drive the next technology of devices. And of course, the last thing why I'm here today is the need for brilliant minds. There's never been a more exciting time to get involved in the semiconductor industry than there is today. I think you're all uniquely positioned and have a great opportunity to do that here at Purdue. I hope that many of you have an opportunity to work in the clean room or try to produce some silicon at the skywater facility or at lab scale for chip design for the MediaTek collaboration. Fantastic opportunities around the university to get involved. Thank you very much. Well, we have time for questions, I think, and I'd like to encourage the students to ask the first set of questions. And do we have the chance to move the mics around? Yes, great, thank you. Why do we need more computational power, typically? What is, like, driving the demand for even more and better chips? Yeah, great question. I think it depends on the application, right? My opinion is we probably use a fraction of the computing power that is in this phone today, but I do think there are applications. If you think about medical, all of the computing power that's needed to do generational modeling for next generation medical treatments, life science, a lot of the vaccine treatments were done with virtual modeling, that is all enabled by quantum and supercomputing, really, really important in that area. We talked about defense, so hypersonics, next generation fighter technology, all of that's relying on next generation chips, higher powered chips, and then I think there are consumer applications that really would benefit from higher powered chips. For example, are you going to be comfortable getting into a self-driving car that is reliant to make decisions locally without a super powered chip inside, right? So I think those types of applications really are quite demanding. They need extreme power, extreme energy efficiency and reasonable cost. So there are applications that require high end. We have a question in the back of the room. Yeah, hello. So I had a chance to talk to some of the representatives from TSMC's new big plant they're building on in Arizona and they were, I noticed how excited they were for the opportunities that were coming. I just had a question about what sort of advice you would give to someone who, like me, is in at Purdue University and is going to be entering the workforce soon. What sort of advice you have for people doing that? Fantastic. I think, I hope everyone caught the message. There's not a better time to get involved in the industry. So TSMC is making a huge investment in Arizona. I think if you're interested to get more involved in exploring a career in semiconductor, there's a lot of opportunity right here on campus, right? And it may not be with TSMC. It might be with Skywater or Intel or others. But I think, you know, I really have gained an even greater appreciation of how much exposure this student body can get to semiconductors here on campus. It's really, really exciting. It's unique. I've been to other campuses where there is not that opportunity to potentially work in a clean room, work in a chip design environment. All of that will be available to students here. So I'd just encourage you to lean in there. And I do think there are a tremendous amount of opportunities. Mark and I were just talking about the number of future engineering roles that will be needed to drive the next set of investments that have already been announced. It's something like 10,000 engineers that are needed to drive those fab expansions. So a ton of opportunity. What has made Taiwan such a powerhouse in the semiconductor industry? Because I was really surprised by the large dominance that they have. What pushed them to become such a leading model? Yeah, I think so. If we think about Taiwan, it's not just Taiwan. It's really one company that's producing most of the advanced chips, which is TSMC, right? So the previous question was around TSMC. I think there was a lot of it was being at the right place at the right time. I also have gotten to know them very well because they're one of our biggest customers. They have a very unique culture that's very agile, innovative, and very, very focused on manufacturing excellence, right? And so if you think about Taiwan, you think about Taiwan and you think about semiconductor, it's not a hobby for them. It is their key focus as a country to be excellent in semiconductors, right? And that's why going back to the U.S., I really think we need a sustained, thoughtful commitment in order for us to kind of redevelop our chip capabilities. Next one, I think they have this culture and focus and expectation of excellence and it's really, really focused on one company, TSMC. The second is, I would say, I think for them, having that focus for their industry and having this talent coming out of Taiwan and focused on working with customers like Apple over time has made them even better. So they have this unique business model, a focused culture, and it's a key focus for their entire country. I think those are some of the reasons why they've been successful. Go ahead. Thank you for your talk. I learned from your talk that somebody in the semiconductor company, somebody will work for IND, basic material, somebody will work in the cleaning room, somebody will work for designing chips. I'm wondering, do you have the percentage of, you know, in a semiconductor company, what is the ratio between these job opportunities? Yes. Yeah. And so I think there's a lot of data actually being compiled about that exact topic. I would say there are a significant number of engineering roles that are needed to be in the fab. And then there is an even greater number that's needed to support that whole supply chain. So for example, my company, CMC Materials, we had about 2,000 employees, 500 were in R&D, and of those 500, maybe 300 were advanced degree researchers. That's just one small niche part of the supply chain. Multiply that by all the different steps that we talked about today. So there's a lot of different opportunities there. I would say it really depends on where your interest lies, right? What engineering discipline are you? Mechanical engineering. Mechanical engineering. Okay. So you might find that you're interested in some of the automation aspects that are related to, in the fab, tremendous need for that. Okay. We also hired a ton of mechanical engineers at my company, because similarly we were focused on automating our next generation of manufacturing the polishing pad. So a lot of different applications that need brilliant mechanical engineers like yourself. So we know that semi-conductors acquired a huge super high manufacturing complexity in order to build them. So what does a country that does not have that kind of complexity, manufacturing complexity to reach that level? Like what are the prerequisites? What are the steps that can take to reach that level? Thanks for your question. So are you talking about for the U.S. to regain that manufacturing in R&D? For U.S. or maybe for any country in general? Yeah. Great question. How can you reach that complexity, manufacturing complexity? Extremely challenging. Okay. So if you just look at, and I was talking about this earlier in one of my conversations this morning, if you look at a company like Intel, super innovative, and all they do is semi-conductor. Where they find themselves today is two generations behind TSMC, right? And that's with all the access to technology, the brilliant minds. It is extremely specialized and focused discipline that's needed to innovate. What I would say is working with, I think the next generation of advanced semi-conductors are going to be done in partnership with the end-use customers. So I think we're beyond the days of, okay, I'm going to manufacture a chip, and then if you're the end user, if you're a Facebook and Apple, sorry, it's called Meta now, I guess, if you're a Meta or an Apple, you're going to use that chip. I think the future is going to be Apple and Meta and other, Amazon and others, are going to want to customize their silicon for their devices, right? And so I think there's an opportunity there if you have agile manufacturing and design to partner with these industry end users to kind of capture that attention and that share from their business. So I think that partnership with end-use device owners will be important for countries like the U.S. to catch up. I have a question over here. Do you think that semiconductor education should be incorporated into grade school at all? And if so, then what do you think is the best way to encourage people in high school and stuff to go into the semiconductor industry post-graduation? So of course I'm biased. I would say the earlier, the better. But I think in all seriousness, I think this exposure, I think today more so than ever, I think semiconductors are getting more visibility than they have been in the past. It may not necessarily mean education at junior high and grade levels, but I think just exposure and education about the importance of semiconductors and how they impact our lives, I think is really important. And then I think the traditional STEM disciplines and focus would well serve the semiconductor industry. And then I think when you get to the university level, certainly there's benefit to having some focused education and focus here, and I know Purdue is really thinking about that. So thank you. Hi. Your previous question actually tied into this one or your previous answer. Which defects of silicon does CMC specialize in and is there a specific reason why you specialize in that? And you tied that into consumer and product manufacturing. And so I just wanted to know is there a specific defect that you guys were looking into? Great question. So at CMC there's really no defect that's acceptable. And what happens when you get to lower and smaller feature sizes, a small scratch that might have been acceptable at an older technology node becomes a killer defect at a more advanced technology node. So that's why I mentioned that this drive for innovation is across the entire supply chain, across each process step. So what we were really focused is working with customers on really, really sub-atomic level scratches and how to reduce those or eliminate those. They could have been in silicon. They could be in metal. But I think the idea, the key idea here is as the circuits get smaller and smaller, the level of tolerance for any level of defect or uniformity or erosion, which is sort of a non-planarity feature, is unacceptable. So that level of innovation has to keep pace with the industry. Hi. So this is going to be a little bit of a cynical question. But if it hadn't been for the chip shortage we've seen in the past two years, and especially the media coverage on these chip shortages, do you think we'll still have this semiconductor boom? Or is it going to be delayed? Or is it just not going to happen at all? Great question. I love cynical questions. Come on. Everyone's thinking about the similar question that you're thinking about. My opinion is that there was a larger risk factor, which is that 90% of the advanced semiconductors in the world are being produced in Taiwan. That combined with this ongoing decoupling between US and China makes semiconductors a really high priority for, I think, the US, for China, for Europe, for everybody. So if it was, I think the pandemic highlighted, OK, we can't buy the cars that we need. And that may have boosted the visibility. But I think the things like the CHIPS Act were already in motion because of this reliance on one country, one company for most of the advanced semiconductor chips. I think that is really kind of the tipping point and flexion point that caused a lot of attention to be put on semiconductors. Thank you. We have a question over here. So typically, as manufacturing complexity goes up, the cost also goes up. And you're talking about creating very complex chips. So what kind of steps are needed to reduce costs while at the same time increasing complexity? Yeah. That is one of the most... Well, I think of the semiconductor industry as one of the most demanding industries that we can get involved in because the requirements for innovation are extreme. The requirements for cost reduction are similarly extreme. Part of that cost efficiency comes with more power and more energy efficient power on a single chip, right? But I think we are reaching sort of the limits of the benefits of scaling, which is that Moore's law that I showed earlier. There will be further benefits from a cost efficiency gained through things like packaging. So how you package the chips together. If you notice, Apple also talks about, like, oh, we put four of these CPUs together. I think that's another way to get at the cost efficiency aspect. But I think that is one of the realities is that the semiconductor industry recognizes the need for increased compute power along with lower costs. It's what we just come to expect. And so it is a requirement. Once you commercialize next generation semiconductor, it has to be at a reasonable price point. I know there have been a lot of excellent questions from the students, but I guess I'd like to ask one question too before we ask some more students. So, of course, as someone who teaches fluid mechanics, I'm very happy to see that your lecture presentation focused on slurries and washing and fluids and so on. But surely there must be heat transfer challenges as well. And I know over the years the question of heat dissipation and these chips has become a really big issue. And so what is the current state as you push the frontiers of semiconductor technology with the heat transfer issues? Extremely important. And so, again, I think we're reaching the limitations of miniaturization and effectiveness for things like heat transfer, energy usage. That's where packaging comes into play. And so a lot of the innovation today is thinking about how do you take away that heat in an efficient manner. I don't think we're there. One of the key areas, I think, beyond just the heat transfer for chips is also battery technology. We need more innovation in battery technology. Where's Professor Wang? She's focused on that area. Recycling of lithium, fantastic work happening in that department. So a lot of innovation needed collectively to solve the challenges of the next generation of semiconductors. I would say the heat transfer is already one of the most limiting factors for next generation chip design. And I think we need to look at those to see how to regain the leadership in manufacturing. So it's important to see where the puck is headed. I think we have time for a few more questions and then we need to wrap this up. One more question. Go ahead. What are the environmental concerns or limitations with the production and the use of the semiconductors? Yeah, so the environmental concerns and, you know, so if you just focus first on the semiconductor fabs themselves. So what do you need to make a semiconductor factory and keep it running? You need a constant supply of power. So it's very, very power-intensive, for electricity-intensive. You also need a constant supply of really high-purity water, okay? I think that from an environmental perspective, thankfully, most of our customers, like Samsung, like TSMC, have spent a lot of time thinking about how they can use and design next generation factories more efficiently from a water usage, water treatment perspective, as well as from a power consumption perspective. So they're pretty far ahead on some of the ESG requirements versus other industries, which I was really pleased to see. So, you know, as we think about the next generation of factories being built in the U.S., many of those are being built with a green energy or a carbon-neutral footprint, all the water being recycled. So there's a lot of that that's already inherent to the industry. And it goes back to... back to what I mentioned earlier, a lot of the production happening in Taiwan. Taiwan is not a naturally rich resource country, so use of water, efficient use of power has been sort of part of their DNA to begin with, which is encouraging from an environmental perspective. Thanks. Well, let's thank David, and again, thank you for attending this special lecture. Thank you.