 Good afternoon, everyone. It's not a good afternoon, right? It's my great pleasure to introduce Professor Leon Pan. He's an associate professor in the School of Mechanical Engineering. He got his master's and PhD from Berkeley. I find I got that correctly. In 2009 and 10, respectively. And then shortly thereafter joined us. He also got his BS from the University of Science and Technology of China. Before he came to Berkeley and before he came to Purdue, he worked as a postdoc in the NSF Nanoscale Science and Engineering Center, INSEC, for scalable and integrated nanomanufacturing, CNAM, if you want the acronyms. Since he's at Purdue, he studies light matter interaction with an emphasis on developing novel, micro, and nanomanufacturing processes, products, and systems for lithography, data storage, communication, and thermal and energy applications. Dr. Pan is also a recipient of the NSF Career Award in 2016. And I hope to hear a little bit about that today. So with that, I will hand it over to you, and I'm looking forward to your presentation. Thank you. Thank you. Thank you for the introduction. So today I'm going to talk about some technical things but relate to how, what's the methodology, what's my motivation to work in this field. So this is a very brief overview. One thing I want to explain is that I didn't get my PhD in one year. So that's actually an intermediate point. I just saw a report that gave me a master degree. Actually, I did spend almost five and a half years on my PhD. So please don't take it. So speaking of scalable nanomanufacturing, the strict definition is saying you need to do mass production with sophisticated features at sub-100 nanometer scale. So, and if you've backdated nanotechnology, you can go back to the 1959 where Professor Richard Feynman gave the speech. But the arrow for nanomanufacturing really starts in 2003 and where the first sub-100 nanometer feature hits the mass production. And in 2015, the first 22 nanometer really hit the production in the semiconductor chips. And that don't be confused. So this is really the device physical size and people talking about 10 nanometer, no technology. That is really brand name. That's now really geometry. And now these people can do best. It's about 16 nanometer in the lithography size where they call 10 nanometer plus a technical node. And the people project the near-term market for creating nanoscale features is about $100 billion. This is a very expensive process at this moment. So many of the applications can use this nano manufacturing. However, the main customer will be the semiconductor where they can afford a huge amount of money. If you look at fabrication in the semiconductor industry, they actually go through a very simple cycle. They do lithography and do itching. They go through this cycle like hundreds of times creating some semiconductor chip for it to use either in your computer or in your cell phones. If you break down the cost, actually lithography nowadays contributes more than half of the manufacturing cost of the semiconductor devices. Also, that determines the ultimate performance of your device. Today I'm just going to talk about a little bit of history and what type of work I have to do with this field. If you talk about lithography, you have to look at one of the roadmap. The horizontal axis is a time of year. The vertical axis is a size of feature. This is caribatting the smaller line width that could build your device. This line, at time goes, you want to make the feature size as small as possible. This is driven by the Morse law. Every time you have to make the device better and give people motivation to swap out their old device, meanwhile you have to boom your productivity. It's very important. If you really position yourself underneath the line, you can make a good fortune. On the contrary, if you really step a little bit above, you're going to be a limited auto business. When I was a kid, there were so many different semiconductor companies I heard of, no longer exist. That's because at some point, they actually like behind and go through this region. This is what my people really predict in the year of 2011. But what really happened? After that, nobody really were able to follow this line. Everybody started stepping into the dead zone. What problem is people start to make the device and the customer do not have the desire to buy. Meanwhile, you sustain the very last workforce to producing those type of devices. Even in the newspaper, people started to talk about the death of Morse law. If you look at the performance device, this is a long, long plot. Look at this. Over the years, the improvement has been 50% per year and then slowed down to 23%. Over the years, they predict something about 3%. But almost saying that it goes to a stop. Since it comes to the end, people just start to polish the corners and make it slightly better. What happened after that? If you look at the new roadmap, actually, there's a delay outside over there, but they will say, there's a new way for people to scale down. That's really because people get something to switch to a new technology. What has happened here? This is where the industry is. At least, they were able to move down. They feel that things are possible as far as 7 nanometers. To understand that in general, we can look at the notion of ice cream. Now, the horizontal axis is the miser of applied effort and the vertical axis is saying, how effectively you can use this effort. In the last plot, we were trying to go that counter. Now, we're going the other way around. This is a roadmap we call scaling. You have to be able to scale up the efficiency and the product better. Of course, if you get a slope like this, that means it's a little bit slower process. If you go a slope like this, there's a better efficiency. If you look at this, look at developing history, most of the technology falls into this type of curve called ice curve. At the beginning, you do a lot of fundamental work just to learn what's the basics there. Eventually, you build up enough potential. A smart guy came along saying, okay, I see it. I take the potential, I take a very rapid growth over there, but eventually the potential is gone and it goes back to improving and materialize the technology. Many technology actually stays in that state, such as the internal combustion engine, like each bike. Many of them actually stay on that, and they are trying to look for future improvement. If you go back to the curve of the same ladder, it's a similar thing. If you go slow and then very fast, eventually it gets slower. This is one of the ice curves. What happened in the recent couple of years is because something else happened. There was another ice curve being developed. This time, people call it South ice ray. It's a new thing. People feel like this is really a junk. It's not going to work at that state. At the start of all the years, people are getting more hope, and then suddenly they make a breakthrough and take over. Now these people actually start to look at the new generation of technology that's enabled people to scale up further down to the better feature size. If you look at those type of breakthroughs here, there are many aspects you could do, including the new physics and the higher throughput and lower cost. You can have more functionality, lower power consumption need to be flexible. There are so many things you could have. Here, what we're going to talk about is focusing on the three aspects. You can do new physics, the higher throughput and lower cost. We'll talk about a couple of projects that have been working on in the past. First of all, in the new physics. This actually has been established for many, many years. People said that the best feature you can get from this projection type lithography is roughly half the wavelength. In the past, people have been scaling down the wavelength from 400 nm all the way to 193. At this point, that's where the first S-curve set of set trade. People say, what can we do? Usually what people do is if I don't find a solution, I'll look at the history of what has been done in the past. I'll try to reproduce. We found that nice usable wavelengths of 157. They tried really hard and didn't work. Eventually, they found that you need to push even further to somewhere as short as 13.5 nm in the extreme S-curve range in order to make it work. In this case, this is really a sub-transition of technology. If you look at the current tool, they have DPUv2 at 193. It costs like $15 million per tool. This is only the two costs. In the end, if you really consider a cost ownership, you actually need to pay four to eight times over two in order to sustain that too. You need to handle those two in order to build a fundry. For the new generation, they do that for part of the process. The two costs are somewhere around $100 to $200 million depending on what specification you really want that to. It's a very expensive process, but the bad thing is that there are the risk-physical engineering limits, and here they're facing a lot of engineering challenges. Now there's the skill-down how to look at the new technology at even shorter wavelengths. This is the background. What do we have? What can we do? Provide an alternative to see if there's another way to scale up a new technology. By looking at the phase area, we found that the distraction limit only applies to the far field. That means if I put the two holes very close to each other, you cannot use projection to reconstruct that into the image two holes. But if you direct your word in the near field, you still can dissolve these two points. It sounds very straightforward, but it wasn't that easy. We actually went through quite complicated analysis, tried to couple the light through vibration electron. In that case, we were able to focus the light. In the past, we were able to focus the light from quarter wavelengths out of the way to one over 36 for wavelength. You want to use that. We also put everything in a ray. For this focusing device, we only have one micron. We can put a thousand of them in a single device and put on the same head and scan over the surface to produce lithography pattern. It sounds quite straightforward. We spent almost six years to produce something, but eventually, we got something at 22 nanometers. This tracker was broken by the lithography tool just for a couple of years ago. The very interesting is this transformation where diffusion is now. This knowledge we developed is also related to another technology for data storage. There's a system of law similar to the Morse law called the Crater's law, saying that how the density of the storage can scale up. They also fixed the same brick wall called the superparamagnetic limit. This actually used similar technology to break that law. Now, second, I want to talk about our efforts to increase the throughput. By looking at a wavelength thing, if I don't use the light, I could use something with shorter wavelengths than the electron. It can be as small as a thing smaller than one nanometer. This is a scanning electron being lithography tool. This is very slow and roughly costs about $2 million. The footprint is reasonable size of the office. What happens to this? If you really want to make this to be prototype tool to make some research development, you roughly need $100 to operate it in parallel. But if you say I want to equip a small army, like a thousand people to do that, then you need $10,000 to operate in parallel. Then I say, okay, I really want to produce things for everyone on the planet. Then you need $1 million to implement in parallel. However, people say, okay, now we want to make money. $10 million for this tool. Is that possible? Actually, we have collaboration with Prashu and also we have a US Payton Award. What we found is the size of the tool really determined by the source of the electron beam. It's about this size. If you can make this through micro, everything is scaled down into micro scale. What we did is eventually we can make an integrated key feature of the electron source and optics in a footprint of about 10 microns by 10 microns scale. By doing so, we can have a very large number of them to perform writing in parallel. Of course, this tool, we are working on the fundamental development of this tool. In our laboratory condition, we are able to produce features that are around 20 nanometers. This is demonstrated in parallel writing. This is a super small ESF over here because they provide funding for us to do this. This is a zoom-in ICM picture of the tool. If you look at the individual things, this is actually a very deep hole there. At the center, there's an emitting device and the money layer is electrode. To try the electron, focus it back to do writing. We made one hero device. This is the size of a U.S. quarter and this is the size of the device and carried up one million electron being in parallel. The third one is the scaling in energy. We have been looking at some applications creating nanoscale structure in 3D for a biological or any photon application. What we found is, in many cases, people refer to the two photon or multiform lithography where the key cost is the I2FAS, Femme-Sain laser itself. It's usually easily cost like half million dollar in that level. What we did is we tried to look at other phases to use just a single photon which is very much the same thing as the laser pointer you buy maybe for 10 or 20 dollars. We were able to create a similar process that can write things in 3D. This is an ongoing project and we have collaborated with Professor Xiemen Xu and Brian Bodoris and we also have a U.S. patent awarded on this part. This is a part of the thing we have worked on. In our lab we also have some of the mature technologies and other technologies cover a wide spectrum of the feature skill and the size of the build skill. That's all. Over the past number of years, I taught ten times the undergraduate course and four times the graduate course. I feel like this really helped me by teaching history I understand there are things called diffusion of knowledge. I really can appreciate those. This is my collaborators and also finding resources. I want to especially acknowledge Professor Xiemen Xu who is my faculty mentor giving me tremendous help throughout the process. Okay. We have some time for some questions. Thank you very much, Liang for the presentation. Who would like to go first? Yes, go ahead. I'll hand you the microphone. Professor Pan, I was wondering something for your electron assisted nanolithography. Is it capable of doing anything other than holes at the moment or is it able to do only hole arrays? That's a very good question. I didn't have time to show. We actually rather than writing a single dot we also create a pattern. We can print like the audits that print images directly. That's a very important thing. Also, we can use that for imaging. One of the key key challenges in the semiconductor manufacturing is try to identify the defective process at various states. You don't want possible to find out things are wrong. How to eliminate the process. Right now, they are creating feature size. Sometimes they want to see for example, they have a trench eventually like 7 nanometer and like 10 nanometer tall and there's a 1 nanometer left at the bottom. How can you see that? Those type of things are kind of challenging and we're thinking maybe those are things that we can contribute as well. Thank you. Really nice talk, Liang. So question, I think this David asked Craig the last question also. This area in particular is very ripe for technology translation. Right? So there's two challenges. One is most of the companies doing lithography are located out in the west coast. And then that's one question. How do you take what you're doing for partnerships with companies to take this? Because it seems like they ought to be really interested in this. And then second, if you're thinking also of your own innovation now that you have tenure you can take all the risk and are you thinking of starting up companies or yourself pushing this out somehow? I really appreciate the question. So I think the location of Purdue is commonly being as a disadvantage because we're far away from others. But recently what I found there's companies contacting me saying that so it seems like there's a common there's a common thing they realize it's very hard to recruit people from the coast because the living price is so high and the people really want to do finance do other things they want to don't want to work on technology. They say like okay they want to usually invite setup center or direct collaboration with the university in the mid-divise and I think we are one for the university under the spotlight. Of course there's a common thing we know who are also under the spotlight. I think we do have nowadays we do have some advantage but it has a lot of education factor in it. I think on that part I will continue to do things really fundamental that is very useful for the company. And about starting the business I really appreciate that we have PRF and the funds here. I took some workshop and learned a lot of things from them. At this moment I understand there could be some risk. As faculty regard tenure, the risk is not too high for myself. But it's a risk for the person who has put it in. So I need to better justify. Any other questions? We're pretty much on time right at 12.30. So Avin, you want to give some concluding words of wisdom to the group. I just wanted to remind everyone that we're not done with this semester's events. November 20th we have Abby Engelberg and Amy Marconet presenting. That's going to be in Potter 234 not in this building but in Potter 234-1130 so do please register to attend. It's going to be a great presentation as well. But in the meantime please join me in giving a big round of applause to Craig and Leang for their success. And to the postdocs and graduate students here, please this is the time to ask them questions and what they did to succeed. Two years down the line they'll forget about it. It's fresh in their minds. We're going to get some good sauce out now before they forget about it. Thank you all.