old one, Fantastic Voyage from 1966. Talk about a group of scientists can take the submarine, this can zone into the blood vessel, they can move inside human body, destroy blood clot and several person's lives. Again, it's been a long time people dream one day we could do this, do drug delivery, precision surgery with precise guidance. And Richard Benman discussed the vision of swallowable surgical robots. In his famous talk in 1959, there's plenty of room at the bottom. So over time, people have been thinking, can we develop this small micro nanoscale vehicle? We can make them move and we can control them. They can find a disease site and they can do different type of treatment. And firstly, we need to think about how to power and how to make them move first, right? And actually, over time, we actually have developed a lot of different type of propulsion mechanisms. So we have this micro-nano scale object. We can power the sensor, power the reservoir using chemical reactions. We can power them using magnetic fields, using acoustic fields. Last but not least, we can leverage these biologic proportions. So in this case, we can realize high-grade proportions. not only we can make them move, we can make them functional as well. In this case, for example, if you've met a robot already at a cell size or bacteria size, how makes them be functional? This really called robot and perform certain tasks, right? One example, we can modify your surface with different bioreceptors. For example, we can modify the robotic surface with nuclear acid with like a DNA, optima or antibody receptors. In this case they could recognize the diseases as a target, for example, a pathogen or cancer cell. They can isolate and target in different type of target including nuclear acid, including protein, even cancer cells and bacteria. So this robot can be navigated into the heart-rich environment. Why we need this micro nanorobot? We have the big surgical robot already, right? Because the micro nanorobots can be very small, they have more like a cell size, bacteria size, or even smaller, they can move into the area that large robot cannot reach. Even the last few millimeters, let's say this way. So over time, we have developed a variety of types of robots so they can realize autonomous locomotion. And this actually was my PhD work, or the one of the first work in my POD, we developed, we call, I need to move with this up here. Yes, there should be a height, I know, or height. Don't make me think. Hi, okay. Okay. Very good. Thanks. Yeah. So I developed this, we call it microengine or microrobot. So based on tubular structure, how many of these microtubes, for example, we can use template in this way, using a polycarbonate membrane template, which has a contrast shape, we can polymerize polyanine, this polymeric tubular structure inside, through electrolytization. So after we deposit this polyanine, we can deposit a metallic catalyst layer. This could be a special platinum particle, platinum layer, for example. In this case, we are having a bi-layer tubular structure with an outer shell being a polymeric layer, inner shell being a catalytic layer. If we put this micro tube in the solution containing hydrogen peroxide, As we all know, platinum is a good catalyst for hydrogen peroxide decomposition. Now, platinum will catalyze hydrogen peroxide to generate oxygen and water. So, the oxygen bubble will eject from one bigger opening of the tube to make the tube go to the other direction. So, in this way, we can realize efficient propulsion inside hydrogen peroxide solution. Very efficient. They can move under magnetic guidance as well, if we incorporate the magnetic layer inside, we can deposit the ion, deposit the nickel inside using external magnet or hammerhouse coil. You can control the motion in 3D fashion. So they can move extremely fast in physiological temperature. If we give them a good amount of fuel, we optimize the catalyst film as well. So if you look at their movement speed, think about the human. The fastest human compared to their body length can only move around six body lengths per second. A cheetah, fastest animal around already, can move 20 body lengths. A fastest car can be 20 around, and space shuttle nearly 200, but these microtubes can move at a speed of 1400 body lengths per second. So think about the size of this microtube being 6-7 micrometer, like a cell size, red blood cell size. They can move at 10 millimeter size, 10 millimeter per second speed. Very efficient. So we can control the propulsion of this microtube, go through a live-on-chip device, it can go against the strong microfluidics in this case. You can realize precise control on the external magnetic field. And as I mentioned earlier, not only they can move, we can give them functionality, give them robotic arm in this case. They can recognize that capture targets. In this case, we can modify surfaces with single-stranded DNA, with optomer, with antibody, with lectin, they can realize on the flight capturing targets, including nucleic acid, protein, cancer cells, and bacteria. I just want to show you one example. We have demonstrated a lot in early years. I just want to show you one example if we modify this micro-tube surface with a receptor called lectin, which is a sugar-bonding protein. They can bond to sugar, like a minus in this case, on the bacterial surface. So they can on the fly capture bacteria E. coli, in this case, because of this bioreceptor bonding interaction. So we have a video to show you. In this case, the micro robot can contact the yeast cell, they contain different sugar conditions, they can contact but no capture, but when they reach the bacteria, they capture bacteria on the fly. You can actually capture many bacteria, transport bacteria to the destination area, and you can release the bacteria just led to micro robot moving through a low pH, a glycine-based dissociation solution. You can reuse the car again to carry more target. This is a small micro engine or micro robot we call. But again, you can argue, oh, these things rely on hydrogen peroxide, but where you can find the hydrogen peroxide in the human body, how can we make them really useful? If you rely on 10% hydrogen peroxide, there is no way you can find it in your body. Then the question comes to, what can we do? We know that our body contains hydrogen peroxide. What can we do with bioavailable hydrogen peroxide? Can we use other bioavailable fields to power this micro nanorobot? So think about the first one. We have hydrogen peroxide in our body, especially at the site of inflammation. Hydrogen peroxide is oversecreted. Actually, near tumor region, hydrogen peroxide is overexpressed as well. So invariable detection of hydrogen peroxide is one of the important applications already because it can help us to identify the inflammation site or even tumor site because hydrogen peroxide elevates in that area. And we know that our robot can catalyze hydrogen peroxide, can realize hydrogen peroxide decomposition reaction to generate a bubble. In this case, we can leverage this micro-robot based on the micro-bubble generation. We can realize hydrogen peroxide detection. The more hydrogen peroxide, the more bubbles you can detect. How can you detect bubbles? If you have some background about medical imaging, you know that a bubble is a very good contrast for ultrasound imaging. I know actually in my later we'll talk more and more about micro bubbles. It's very interesting. So you can imagine a single bubble even at a few micrometa sites using ultrasound. So in this case, we can put micro robot inside the biological environment. You can look at hydrogen peroxide reaction, look at a micro bubble generation, we can detect hydrogen peroxide. To improve the sensitivity, we can instead of a platinum catalyst, we can use a biologic catalyst enzyme like catalyst. Catalyze is an enzyme, a good catalyst for hydrogen peroxide decomposition. We can modify this micro robot surface with catalyst, this protein enzyme layer by layer so that you can pick many nature catalysts on the surface of the micro robot. Now you can even detect biophysiologic concentration of this hydrogen peroxide inside body. We did a in vitro and in vivo test to realize this detection of hydrogen peroxide using micro robots. This is only a proof of concept application. You see that we detected this PMASE activated the neutral field, which has elevated hydrogen peroxide. We can clearly see and also clear here with the concentration. With PMASE here, you see the bubble generation in the present of catalysts and without PMASE you see a little or no bubble generation. This is one way, application of in vivo hydrogen peroxide detection. And because our body contains hydrogen peroxide, especially at disease site, which make it even more promising, how about using this for therapeutic applications? We can make the micro robots moving near the disease site. For example, tumor cells contains, tumor region contains much higher hydroproxile, up to even 100 micromolar. And we can make this micro nanorobot. They can move efficiently near the tumor region. They can target the tumor region. In this way, we present a nanorobot based on MOF. This Z67, this material, they can naturally catalyze hydroproxile deposition, decomposition. They can realize efficient propulsion, not only near the tumor region, they can even move efficiently inside the tumor cell. You know, if your micro robot or object is very small, the cell will uptake this object. Because of inside cell, this intracellular hydroproxyl level is so high, this nanoproduct will move much more efficiently in the tumor cell compared to normal cells. So in this case, not only they can move efficiently, they can carry drug, they can target metachondra. They can target metachondria because they can move much faster inside the tumor cell. They can have a higher chance to reach a metachondria, and metachondria is negatively charged in this case. And our robot is possibly charged. We modify the surface of the TPP. So due to fast motion and the charge interaction, this micro robot will have much higher chance to reach the metachondria of tumor cell, and because they also carry the drug, Dr. Rubin They show much higher efficiency to kill the tumor cell while they have minimal harm compared to normal cell. This is one of the interesting ways we realize targeted tumor therapy using this fast-moving nanorobots inside hydrogen peroxide naturally existing. So this is, I'll show you a video, some problem here, okay. Okay, you see inside the low level hydrogen peroxide, this is like tumor hydrogen peroxide level already, you can see fast displacement or enhance the diffusion in this area. And compared to healthy cell, you don't see much movement near the tumor cells, they move very efficiently because tumor cell express much higher hydrogen peroxide level, as I mentioned earlier. So now we can realize intracellular efficient propulsion and mitochondrial targeting. And we also show that using this approach, we can do the tumor treatment much more efficient compared to various controls we have tried. And this micro-robot, drug-loaded micro-robot, with this positive-charged surface, have much better efficiency. Now the tumor size, you see, compared to other approach, the smallest one. And they're fully biocompatible as well. This is one way we can realize solely using hydrogen peroxide present in vivo near the tumor to realize enhanced tumor therapy. So this is one way. We can realize propulsion using existing hydrogen peroxide in our body. And how about the other part of the body, right? We don't have hydrogen peroxide that high concentration. Maybe the propulsion is very limited. That's why we are looking at whether there is other fuel we can leverage to realize in vivo propulsion. One of the important fuel we are looking at initially is using acid. We are, in this case, our body contains a very strong acidic environment like a human stomach, pH 1 to 2, very acidic. And we know that a lot of metal, active metal like magnesium, like zinc, they are bioresolable. They are biocompatible, biodegradable. They are also reacting with strong acid to generate a bubble. In this case, we don't need hydrogen peroxide. We don't even need platinum. We can make a micro robot based on zinc or magnesium. In this case, they can move very efficiently in a very strong acidic environment. Very fast as well as millimeter speed over time. So this zinc based on micro robot can be used for the stomach application already. I will talk about application later on. So, you can argue, oh, star market is very acidic environment, that is extreme environment. How about neutral environment, like for the intestine? How can we make a micro robot move in the intestine? So we can leverage the magnesium water reaction. Magnesium could react with water. But again, this reaction is inhibited or passivated because our magnesium force is densely packed oxidation layer on the surface. So preventing further metal water reaction. So, in this case, we can coat a magnesium surface with a layer of golden nanoparticle. Golden nanoparticle will facilitate this galvanic corrosion effect. And also, our fluid contains a higher concentration of chloride. The chloride pitting effect also happens. Because of galvanic corrosion and chloride pitting corrosion, these two combine together, now magnesium can react efficiently even at a neutron pH with water continuously and realize continuous propulsion. Why we need to make this micro robot move inside GI tract, because GI tract drug delivery is very challenging. There are a lot of challenges to realize sustainable drug delivery inside GI tract. Because when we eat, swallow medication, the drug can quickly pass through the GI tract. They don't really stay inside body very long. That's why we need to frequently take medication. Then, we have different type of biological barriers. It's very hard to ********* for the mucus. Mucus layer is very hard to go through for regular nanoparticle. So and the GI tract have important problem to address. For example, for stomach infection, H. plurie, it's a very important condition. You see like the infection across all years was 55% in China. Yeah. So, people often have to take antibiotics which have strong side effect to treat H-polyurase bacteria infection in the stomach. And there is also like a colon cancer, stomach cancer, both are among top three cause of death for among cancers in general. A very important problem to address, but again, very difficult to realize reliable, sustainable release of drug inside GI tract. So we could use micro robot now. Now, for example, one of the first in real study of this micro nanorobot, we can now swallow this zinc-based magnet like a micro robot that could move inside GI tract. That could move in the stomach in this case. Not only they can move, they can enhance the *********** because their fast moving will induce their *********** of the stomach lining. The micro robot now will stay on the stomach, slowly release the drug possibly, and then eventually degrade because they are made of zinc, made of magnesium, and made of polymer. So we also demonstrate that we can use this micro robot to load a different type of drug. In this case, we use silica, gold as drug carrier, as model drug carrier to show we can efficiently deliver drug to the stomach lining in this case. So really pushing further, we have to demonstrate they are fully bi-compatible. In this case, we are using probably five or six million of these micro robots. They are still showing fully bi-compatible. You don't see tissue damage in vivo in this case. They are fully compatible. That's why it's safe to swallow them. And we can demonstrate this can be used for further bacteria infection treatment. So as I mentioned earlier, H. pylori infection is affecting a lot of people. Actually my wife also suffer from this. So people have to take antibiotics. Taking antibiotics is okay, but typically people also need to take this proton pump inhibitor. Why we need this? Because drug attack effect in a relatively neutral environment. You cannot really use drug in a very acidic environment in this case. That's why people typically try to neutralize the pH of a snark first using this PPI, then the drug attack effect for bacteria targeting. But this PPI have strong side effect. So in this case, think about what if we use our micro robot? This magnesium-based micro robot, they are swallowable, eatable, fully biodegradable, and they can load the drug as well. If we eat this inside the body, firstly, they will react with gastric acid. This reaction will cause efficient motion or propulsion. Not only they can cause propulsion to improve the retention of the micro robot targeting of the stomach lining or retention, they can also neutralize the pH because magnesium acid the reaction naturally neutralize the pH of the stomach and eliminate the need of any of these PPI inhibitor use. That's why for the in vivo study we demonstrate that for this micro robot, they have dual effect, neutralize the pH and the more effective deliver drug because the micro robot can stay on the stomach lining for a longer time until they degrade, in this case with higher efficiency and also reduce side effects so they are actually working for in vivo stomach infection treatment. So looking further, we are thinking when I started my lab account tech, one of the important questions is, oh, these are micro nanorobots. They can move very efficiently inside the body, but you will lose track of them. So in vivo imaging is a very important capability we want to realize. You need to see them for many applications. For some applications, we don't really care. swallow them, put it into the stomach because they can actually move around randomly, stay in the stomach lining and deliver drugs, that's fine. But for many applications like tumor targeting, you need to see them, to target them to the tumor, right? That's why in vivo imaging is one of the grand challenge for micro robots. And the second one really, can we control them? Can we guide this micro robot deliver drug to the tumor specifically instead of random motion and do all the direction delivery, right? That's one of the advantage of micro robot if we could do that because people use nanoparticle to direct delivery. So they're not very targeted, so they pass on passive diffusion. They can diffuse to the tumor cell. They can also diffuse to the healthy cell. So it cause strong side effects. If we can do precise targeting with imaging guidance, this is one of the main advantage of this micro robot. So this is a perspective published in nature about micro nanorobots need better imaging and control. And at CarTech, luckily, I found a very good collaborator, Prof. Li Hong Wang, who is the pioneer of photoacoustic imaging. Prof. Wang is a leading expert and the lab has the deepest *********** fastest optical cameras. Here you see the video I showed here. Actually, you can see the labor-free tracking of circulating my normal tumor cells inside the mouse brain. You can even scan animal body from head to the tail, do the full body scan to the imaging. In this case, if we could make a robot with good imaging contrast, we can realize in vivo localization and control of this micro robot, right? Based on this, we developed our one new robotic platform. We call it imaging-guided micro-robotic system for targeting navigation inside the intestine. We started with the intestine because we talk about application in the gut already, inside the stomach already. In the stomach for bacterial treatment, we don't really need a precise targeting because bacterial infection near the whole stomach can make it influenced, right? But for the intestine, if one day want to treat the tumor or inflammation site, site, we need to know the specific site to trigger the drug release, right? So but that's why the intestine is very long, it's very hard to control in this case. So we developed this micro robot platform. In this case, it's essentially, first we start with a capsule. This capsule, a relatively large capsule, so that photoacoustic imaging can locate them. So one capsule contains a good amount of micro robot. The micro robot is based on the magnesium based robot I mentioned previously. The magnesium surface is coated with the golden nanoparticle. So that they can, because of galvanic reaction and chloroepetine reaction, they can realize efficient propulsion in a neutral environment. But we don't want to direct them to swallow them. If you swallow them before they reach intestine, magnesium will react with gastric acid. We get crushed or destroyed inside the stomach. We want to protect them. We want to protect them, that's why we encapsulate many of these micro-robots inside a bigger capsule. The magnesium micro-robot contains gold nanoparticle, which is also good contrast for photo-acoustic imaging. Now after putting many small micro-robots into a bigger capsule, we can see strong contrast using photo-acoustic imaging. We can track the migration of the bigger capsule through the GI tract using this real-time imaging. testing, once they reach, goes through the stomach, because the capsule shell will protect the micro robots through the harsh gastric environment, when they reach the intestine, when they reach the disease site, we can activate the capsule. Now by applying NIR light, continuous NIR light, we can heat up the gonadal particle to break this microcapsule. The microcapsule now will be broken near the disease site and the micro robot will be released. Now, a micro robot will move inside the intestine environment. They can do the localizer retention and the sustainable drug release to get enhanced treatment. So I have a video to show you the process. You see now this micro robot loaded with the drug, has good imaging contrast, and it contains by result of a body for propulsion. We can load them into a bigger capsule after the mice swallow this capsule. And then real-time continuously tracks the capsule location inside the body using actually photoacoustic imaging, you know, applying the post-NRR to detect the sound reflection. In this case, when they reach the disease site, we can apply continuous NRR in a very short period of time, the micro robot will get released. And autonomous motion will enhance the *********** near the disease site and release the drug and fully biodegrade and eventually we get effective treatment. That's the whole concept. And we showed that we can have precise control about how many micro robots we can trap inside one capsule. This is the one capsule trapped with one micro robot. A small group, a much larger group. Based on application, we can control this and typically we can use a much bigger one and we can realize, you know, Multiplexer drug loading as well. Now a capsule is trapping one particle and we are using FITC, in this case a green fluorescence drug model to let you see how we can effectively track one drug and Dr. Rubinstein has red fluorescence property, we can also track this anti-cancer drug. We can even load both drugs together inside the body. Now you see both fluorescence inside the capsule inside the micro robot body. So now, we show that under very thick tissue, because this photoacoustic imaging can do pretty deep ***********s at least several centimeters, it's no problem, 7 to 10 centimeter deep. We can put a very thick tissue under the thick tissue, we use chicken breast here. We can still see the fast propulsion, the migration of the microcapsule inside the tube in this case. Again, once they reach the disease site, under imaging, we can apply our light to break the tube within seconds, actually. The micro robot now will start to move and ********* the intestine. We actually showed that over 13 hours, we can continuously track this micro capsule inside the GI tract. We can select and activate the release when they're near the tumor site. So, in this case, this is actually tumor, it's very important, the tumor has very strong contrast on the photoacoustic imaging. We don't even need to label the tumor, we can actually see where are the tumor on the imaging. So, now this microcapital can get close to the tumor and we can trigger the release of the robot. And we show that using this micro robot because of active propulsion, they can stay on the mucus. I mean, they can ********* the mucus, they can stay on the intestine wall at much higher density compared to passive control. We show that in this case, because they are staying there over a longer time, they are gradually released drug, we get a sustainable drug release. And one of the reasons, not only because they are proportionate, they can ********* the mucus. And magnesium on water reaction will elevate the pH of mucus, make pH softer, in this case they can ********* deeper to the mucus. This is another mechanism for their enhanced retention inside the intestine. So we also show that because they are made of magnesium gelatin, so they are fully bi-compatible and from this histology analysis and also body weight of the mice, you see that we don't see obvious change over time indicating their good bi-compatibility. So, this is a very interesting application already, but again, we are thinking of how we can make things better, because magnesium or zinc, if they react to the water, they generate a bubble, you know, they have a limited lifetime, right? Like magnesium, after they're fully dissolved, they will not move anymore. Can we realize longer term operation and with full precise control inside the body to do a lot more things? And very recently, several months ago, we presented a new micro-robot platform. It's powered by acoustic field based on micro-bubble. So interesting structure is we can fabricate a micro-structure with a hollow core inside. We can print a shell structure and we can let this shell structure trap a gas bubble inside. If you put it in the water, a gas bubble will be trapped inside. And if you apply acoustic field, the gas bubble will oscillate or will vibrate. The vibration of the gas bubble, if we leave one opening of the tube or two opening of the tube, the gas bubble oscillation will disturb the flow around the microstructure. This will make this microstructure move very quickly on the acoustic imaging, on the acoustic field. And because of the gas bubble, as I mentioned earlier, gas bubble is very good imaging contrast. You can use acoustic field to do the real-time imaging as well. And we also, instead of using commercial resin, using two-photon polymerization, we made a resin bed of hydrogel. We can print the hydrogel structure so then they can stay inside the body. And they can eventually degrade over time after the delivery. So overall concept, we can print the entire structure using 3D printing in this case. We can load the magnetic nanoparticle ion oxide inside, fully biocompatible, allow us to precisely guide the moving direction. And because of this bubble oscillation, they can move in any environment now, in all different types of biofluids. We can real-time imaging, power them using acoustic field, image them using acoustic field, and precisely guide them to the tumor site, release drug, and fully biodegrade in the end. So, this printing process is very interesting and scalable, and we are using two-photon polymerization printing, and to prepare the racing, you see how microstructure get printed. You see the inside is hollow. Now we even print half structure, there's a hollow core, solid shell made of hydrogel, we can print many at a time. We show that when we leave some opening, after you put it into the water, a gas bubble would be trapped inside. So allow us, you can change many different structures. This printing is very versatile. You can design the structure and very quickly you can print it. So as I mentioned earlier, when you put it into water, a gas bubble will be trapped inside. And when the gas bubble vibrates on the acoustic field, the gas flow, I mean actually the water flow will be disturbed. We did a simulation, we did a passive particle tracking to look at the flow direction actually. And we found that when we have two opening, it actually works better than one opening. You see like this one opening moving speed is okay, but with two opening, they can move very quickly. Speed is very high here. And one of the challenges, you know, for this type of structure, since we try to make these things based on hydrogel, the hydrogel, you know, is hydrophilic. Hydrophilic structure, how can they trap bubble, right? They will immediately loss bubble. Actually that's a problem. Even people use commercial racing, they are not that hydrophilic. The bubble, these things can work well for a very short period of time. If you put in the biological fluid, the bubble can only last few minutes, then bubble disappears. That's short life, too short to do anything, actually. That's not what we want. How we address this problem? We actually proposed this interesting dual-step surface modification strategy. We can modify the surface of this structure with long-chain soil. In this case, make the surface become very hydrophobic. So after you put into, after you put in like a dual-service modification, you put a long-chain carbon on the surface and the outer now becomes like a very hydrophobic. Hydrophobic surface is good for trapping bubbles. At least while the cell for cell monolayer still lasts, the bubble can be trapped very well. But another problem comes out. If the structure, the whole structure is hydrophobic, they will start together. They will groove together. They don't really stand alone, moving very well. That's another challenge. How to prevent aggregation. In heart dispersion, you have many individual robots that can work together. We need a hydrophilic out shell in this case. We want the ultimate goal is hydrophobic core and hydrophilic out shell. So our strategy, we make it hydrophobic first by deep coating modification, do the long-chain carbon modification. Then we do a quick oxygen plasma treatment outside to make outside hydrophilic. Now this micro robot can work really well compared to, I mean, you think about how many days in this case, they can work in PPS for over 12 days. Urine can be like many days as well. Even with the serum, they can work for three days. Not only they can, this is only like three days, still 80% of the robot remain bubble. They are still very functional. They can actually work much longer if you look at statistically working. But if you use commercial racing, this green bar is almost nothing you see. Within minutes, all the bubbles will be gone. Now after seven days, we can still see the micro robot is moving very effectively inside the PPAs. And after four days inside the serum, they still work very well. But eventually, when the self-assembled monolayer degrades, so this micro robot becomes a hydrogel shell, it gets exposed, then slowly will degrade. That's what we want to realize. They are stable enough to do certain tasks. Then eventually, fully biodegradable in the end. And this type of micro robot, they move very efficiently in a raw untreated biofluids such as urine, GI fluid, wound fluid, and even whole blood fully packed with the cell. They can still move pretty good inside the complex biofluids. They can do multi-day operation as I mentioned earlier. because we can incorporate magnetic nanoparticle inside during the printing steps. So this allow us to control them. Not only you can make one moving, you can control them writing certain type of letters, you can control many moving. That's a real application. One robot cannot do much. But a bigger swarm of robot, you can deliver many, like a large, relatively large dose of drug to the tumor site to be more effective for tumor treatment. Now, we have precise control of larger amount of robot. So come to imaging, as I mentioned earlier, so imaging is very important for in vivo application. And this micro bubble give us very good imaging contrast. Here I want to show you, yeah, if you print this full, this structure, eventually we trap full bubble, you can already see the ultrasound signal. If you trap many, you can see much stronger ultrasound signal. This is allow us to track this micro robot under the deep tissue in vivo. I have video to show you, you see a group of micro robots now make them move, we guide them to one side. We can clearly see in vivo many of these micro robot work. And if this is a tumor site and we can control many of these micro robots moving toward the tumor site, you can actually see this is the bladder tumor. Now we can guide the micro robot to deliver drug to this area. And this is in real time actually. And we did in vivo therapeutic application as well by using micro robot to treat tumor. We can precisely guide the robot to find tumor and compared to passive control, compared to random moving, we show that activated bomb, activated micro robot with imaging guidance can lead to much better therapeutic efficacy. So we get over 93% decrease in tumor size in general, so quite a successful study. So and very importantly because they are made of hydrogel, you know, they are fully biocompatible and eventually they are biodegradable. After several weeks, we see mostly they degrade in the body already because of the hydrogel hydrolysis of this PEGDA polymer structure. So, overall, it's a promising solution for available application and in new work unpublished that we can even make the micro robot to have chemotactic behavior that can, you don't even need imaging guidance. They can autonomously locate where the tumor, so they can deliver drug in a more smart way without control directed to tumor selectively. So the main part of today's talk, I still have a small session after this. On the summary of my first part of my talk, I discuss the vision of this fantastic voyage. We can develop a group of micro nanorobots that can move inside our body. They can be powered by local chemical reactions such as hydrogen peroxide, such as the acid or water metal reaction that can be powered by magnetic field or ultrasound field. And this is one of the science fiction pictures I saw when I started even my PhD many years ago. I think how future micro-robot should look like, if you have a swimming towel, you should have a micro-camera to look at where the tumor, you know, guide them, to have this payload to deliver drugs. Think about today I mentioned, we have a way to power them. We have a way to imaging them already. We actually have different ways to load drugs as well. So now it's slowly, I would say, over time, it's from more like a dream vision, move more more and more toward real biomedical applications, actually. Over the past several years, this evolution of this field is very rapid, actually. So, I think you can envision in the future there are swarm of this micro nanorobots fully precisely controlled or even in a more smarter way can enormously track the disease site, deliver drug. Yeah, and eventually no harm to the human body because they are fully biodegradable. So, as I mentioned at the beginning of my talk, my lab, actually, mostly, actually, my people focus on sensors, but today I will not talk about this wearable, flexible sensor for house monitoring applications, but I want to see a sensor is not only important to be applied to human body to look at different type of biochemical, biophysical markers for precision monitoring or different type of AI powered applications. You know, we have developed different sensors that can target a different type of disease, heart failure, cancer, COPD, diabetes, malnutrition, depression, fertility related problem, covenant infection, long COVID even. But I want to say sensor are also important for robots because we are robotic seminar here. But look at robotic sensing, remote robotic sensing, you have seen a probably a lot of application even now industry robot have built in sensor as well. So, so far most of sensors on robot, they are based on physical sensing. Robotic sensing are very important for agriculture, for security, environmental protection, health care. But again, so far, mostly limiting monitor pressure, how to hold the cup properly, how to sense the temperature. So far, most applications are limited to the physical sensing. And autonomous chemical sensing is actually not being demonstrated much, but still can be very important. and think about if you let a human working in a very dangerous environment with the impregnable explosives nerve agent or virus, you don't want to send a real human, you could send a robot, but how to detect those things can be very important as well. So, the challenges right now is how can we develop a robotic scheme? They can have multi-modal sensing capabilities, they can monitor physical parameters, such a temperature, such a touch. They can also monitor chemical information such as threads. Can we wirelessly control the robotic operation using human body to get real-time feedback? Can we mass-produce such sensors at high performance and low cost? And finally, can we develop an enormous robot that have automatic tracking capability to find the threat location, for example. So we demonstrated one human machine interface that can realize robotic physical chemical sensing applications. Actually, it contains one pair of e-skins, and we have on the robotic side, we build this robotic skin, you can put on the robotic hand, they can monitor pressure to sense the touch, they can monitor temperature to sense the temperature, they can monitor different type of chemicals including explosives, including nerve agent, biohazard, including even virus. So, how we can realize chemical sensing, I will talk about it in the next slide. But important, not only robots can sense these things, they can send real-time feedback when they found a threat, for example, wireless send data to the human body. And the human body can also control the robotic arm operation using a well-known BMG skin patch. You can get a signal from human body, you can realize gesture recognition, and wireless to send a comment to the robot to do different type robotic task operation. So these pairs of e-skins can communicate with each other and eventually to realize the human machine attacks and to realize robotic physical chemical sensing with real-time feedback application. And importantly, we can fabricate this robotic e-skins using printing process. In this case, we are using inject printing. So inject printing is a very common technology people use a lot. But in the office, we are using in-jet printing to print different color document, right? Using commercial color inks. But in this case, we can make our own inks. We want different material for different applications. For example, to detect nerve agent, to detect explosive, or to detect pressure sensitivity. You may want different types of nanomaterial because different material have different unique property. But how to fabricate that? So we can make our own inks made of 7 nanoparticle, gold nanoparticle, carbon nanoparticle, carbon nanoparticle, carbon nanoparticle, platinum-desecorated graphene oxide, 7 nanowire, morph, mixing, and even encapsulation layer. Now you can realize almost fully printed multi-modal sensing. It makes this process more scalable. Of course, if you think inject is not scalable enough, we can also leverage the advantage of the road-to-road printing. So, this is the way we can combining this mechanical cuttings structure. We want to make a skin that can be seamlessly attached to the robotic hand. We 3D printed robotic hand and we can put the skin onto the robotic hand. We can cut the curicamic structure so that the skin finger can really deform. They can still tolerate all this different type of strength. we can print serial interconnects, print functional nanomaterials such as gold nanoparticle, morph nanoparticle. This is for enhanced sensing performance. So the entire process can be automatic using this printing process. So eventually we get this structure. You see one sensor array contains two physical sensors and multiple chemical sensors. Chemical sensor, how can we realize chemical sensing? Because we know that if we want analyzing body fluids such as blood, such as sweat, such as saliva, we have liquid. We can dip the sensor into liquid to do detection. But the robotic hand, many places you don't have liquid to dip, right? If I want to touch the surface, whether it contains a nerve agent or bacteria, how do I do that? The way we address this problem is putting a hydrogel on the surface of the sensors. The hydrogel will sample the dry five-phase chemicals from the touch, the surface that the robot will touch. Then the hydrogel will sample the chemical and transfer or transport the chemical onto the sensor surface. The sensor interface will directly convert this targeted chemical reaction into a measurable signal. In this case, we can realize dry chemical sampling plus on-site real-time detection. We realize we can do TNT detection, we can do OP nerve agent, organophosphate, we can also do coronavirus detection. We are using a platinum graphene surface to do explosive detection. We are using morph nanoparticle to do the nerve agent detection. We are using carbon nanotube modified with antibody to realize bacteria or virus detection. So, from the human skin side, this is actually printed EMG electrodes. We can wear this at wrist. We can continuously record the EMG signal by using this skin patch. And using machine learning, a very simple machine learning, we can realize gesture recognition. In this case, you can realize robotic hand operation and control. This is a real-time demonstration of this gesture recognition and robotic hand control. Of course, there's some second delay because we get to the signal machine learning, we will analyze it mostly because of the signal recording, it takes a few seconds for us to process the data. So now we can put a relatively good amount of array. Each one contains a physical chemical sensor array. Now you can actually control the robot to touch the object. Now you can even do the mapping, see which area contains more explosive or hazard. In this case here, it contains higher concentration than others. You can see that in probably the contact area. The disease site is here. You can do the surface mapping when it touch different type of object. Of course, this is only proof of concept we can put in this manual array. You can even improve the sensor density to realize higher density touching not only mechanical feedback but also chemical feedback. Once we detect the threat actually, we also have a stimulation electrode on the human arm. In this case, the human will get real-time feedback if things get detected. So this is a human control robot to do things. And of course, we also want to make things smarter, at least from proof of concept. Think about there's a leakage in the ocean or in the water, explosive or like a biohazard. How can we find the diseases site? So, we built the sensors right now on the autonomous vehicle, the autonomous swim. So, we build the sensors at different sites of this little boat, we call it a boat. So, different sensors will simultaneously sense the target of interest, let's say explosive or nerve agent. So based on the sensor response at different sites, they will autonomously have different readings of different location. So we are using a S-star algorithm to calculate where the robot will go. We can sense the gradient of this biohazard. Then eventually find where the leakage is. In this case, no human will be involved. A robot will make a decision by itself. Based on the sensing result and the algorithm built in the microcontroller. So in this case, you see we can realize wireless motion control. If you want them to turn left, you can turn left, you can turn right, you can turn right. So this is the initial external control. And then we can realize autonomic tracking. No human getting involved purely based on sensor result. You can realize this is actually the real leak disorder where we inject a little bit like a biohazard. Now you can see at this site, we do sensing, then decide where to go the next. Very simple demonstration, but this could work in a much larger scale, think about the ocean, think about the river in this case. This is at each site, you would do a sensing and then decide where to go. There are no camera, there are no human involved, fully autonomous based on decision making from the sensor without. So here again, to summarize my talk today, I showed you also the micro robot and showed you a simple interesting result we get from this human machine interface and all the way to realize autonomous robot. And there are a lot of room we can do and this year we just wrote a perspective in the nature of chemical engineering about robotic sensing as well. So more functionality we can give to robot, not only to physical sensing, chemical sensing, they can enhance more decision making, give the robot more functionality, especially when they're working in the extreme environment. And think about, we have also collaboration with NASA, think about the space exploration as well, what you can put on the robot. So, to summarize my talk, I've showed you already, from small-scale robot to larger-scale robot in our lab. Again, I want to thank my lab members that did most of the work, and we are focusing on a variety of different types of topics. As I said today, mainly focus on robotic-related research. I also want to thank our collaborators for the technology development and the in vivo evaluation as well. And finally, our funding support. Thank you very much for your attention. Thank you for the great talk. It was really a pleasure to invite you. So we have time for some questions here. And maybe Jinshuo, can you? So we'll bring the microphone around. Does anyone have any questions? For the micro or nanobots, you mentioned magnetic guidance to get them to go in the right direction. Have you considered using MRI machinery to get maybe a 3D magnetic field deep in the body? That's very interesting. So far, we haven't to do MRI control, but that's an interesting way. What we did in the lab, either use a small magnet or we have a hammerhouse coil. Of course, we didn't do human, small mice, you can put it inside. side, we can control the 3D direction as well, possibly. Like in this case, depending on the coil size, you could even build a much bigger coil. Of course, in a larger scale, it's an MRI machine. It's very doable, I think, for the control. Yeah. Quick question on the... Have you guys worked with Professor Giovanni Torverso at MIT? I know he looks on similar swallowable technologies. Yeah, Jiu is a good friend. We did not directly work on research project together, but we just had a review article just accepted recently, probably within a month it will be coming out, but we share common interests. For example, we also work on, I did not show here, microcapsule, microcapsule that can do autonomous metabolic profiling inside the gut, we can incorporate therapeutic application as well. And we are also working with another professor at Caltech, Professor Azita Imami, and we're doing IC design or, you know, magnetic localization. And Professor Imami has active collaboration with Professor Tarasso as well. There are multiple collaborative work. Yeah, overall, it's a small community. Yeah, it's great to see more people. I mean, like the leaders in the field are making good progress. Yeah. Thank you for the great talk. I know from like the picture that you showed that you said you saw during your PhD. So I'm curious, like, do the micro cameras like from that slide already exist? And then what is the bottleneck to create those cameras if they don't already exist? Right. Yeah, the real micro camera is very hard to beat. But of course, that's why it's science fiction, right? So what our strategy is, the micro camera can be two directions we can realize. One thing is the micro robots have good imaging contrast, as I said. Could be a bubble, could be a gold nanoparticle, depending which type of, could be a magnetic particle if we do like a MR imaging. We can externally localize the micro robot to control them. This is one thing. But of course, another vision is the robot can see things, right, by themselves. So called micro camera, what we could realize is we could build the biolocalized chemical reaction, chemotactic motion, they can, as I showed in the later larger scale robot, they can sense the gradient of local chemical, for example, tumor have overexpressed chemicals, many different types, peroxide is one of them. They can autonomously locate the tumor because they have the built-in sensing capability. Yeah, that's their eye in this way because biochemical gradient is the way they can find the target or with the system of external imaging technologies. Yeah, thanks. Hi, thank you for the great talk. I had a quick question about the micro nanorobots and just wanted to know if you have had any conversations with regulatory agencies about how these can be translated and what concerns they might have for getting these out in the field because of the various components involved in these micro robots. Yes. I think that's a very good question. Again, this field is relatively new. As my knowledge, they really know this product in clinical trial yet. But as you can imagine, actually what the most important part is biosafety. So they are mostly based on fully hydrogen materials or based on, you know, like a bioresolable metal material. This is widely used in the medical implant field already. And those drugs we are using, I mean, the micro robot is essentially drug carrier. The drug is already get approved. In this case, the harder part is already get approved, and the lower part, we just need to show they are effective and they are safe, and they are clearly safe in this way. I can see in the next several years, we probably could see more robot in the clinic trial. I've seen several startup already in this field, and people trying to do robotic drug delivery and even like a related fertility, you know, how to like application deliver ***** even. So, those are like a very interesting applications and I can see clear pathway there, yeah. Hi, so I had a question. So as you scale up the number of micro robots, is there any dangers that come with it? So for example, like the platinum and hydrogen peroxide engine, it generates like a bunch of oxygen. Right. Like a platinum, hydrogen peroxide, that's why I talked this in the very beginning. So, it is not very practical because you rely on several percent of hydrogen peroxide to realize it's very efficient moving. That's why eventually I showed that, you know, the in vivo application, they are made of mostly biocompatible material, could be zinc, magnesium, morph, they are based on bioavailable hydrogen peroxide. We don't create, we don't add hydrogen peroxide, using naturally existing hydrogen peroxide or biofluid as a fuel. Of course, we could also external fuel as well. Overall, I've showed the biocompatibility of this micro robot, when we do that, we typically apply millions of robots into a small mice. We still don't see clear tissue damage or negative factor to the animal in this case. Hello, thanks for the very interesting talk. Just a quick thought. So you mentioned for nanorobots, you can use magnetic fields or, you know, Sonar to change the direction. I was wondering if there's utility in also doing speed control for the nanorobots. If so, how would you go about doing it? Right. Speed control is actually relatively easy to realize. For example, when we use chemical field, we can use external field, possibly ultrasound or heat to control the reaction speed, to control speed. If we use external magnetic field to power it, we can easily, I didn't show our magnetic swimmer, they can use a change of frequency, change magnetic field, you can slow down the robot. With ultrasound, it's much easier. By tuning ultrasound frequency, you can already, or power, you can control the moving speed already. You can slow them down or even stop them. Don't apply the field anymore, right? In this case, yeah. You have the on-demand control. Yeah, chemical control, chemical power robot is harder to control. I would say modulating ultrasound, modulating heat could control the reaction rate in this case. But external field power is much easier to control, yeah. Thank you. Thanks. Thanks for the amazing speech, Professor Weiko. I have two questions. The first one is, would the micro robot trigger the human body immune system? So yeah, I think that's a very good question. So depending on application, right? So many of these tumor environments, we don't want a micro robot in most of the cases enter the circulation. Because if you inject into a blood vessel, it's very hard to go against the blood flow to precisely control them. So in most of the applications, we are directly injecting, as I said, compared to bigger robots. The advantage of micro robots, it can reach the hard to reach environment. It can cross the barrier or you have a robotic knife or to place a robot close to sight but the robot can deliver the last several millimeter or centimeter, they can move precisely. And also, there are different ways to improve their immune response. If you really want an application, go to the blood vessel. We do have those applications for them inject into the body. They can cross the blood-brain barrier. So in this case, there are different ways to do surface coating. Like commercial, not really commercial, but there are a lot of nanoparticle drug carrier in the clinical trial. People use it for the PEG. People use even cell membrane coating to improve their long-term circulation to, you know, to really make them more like immune system, we're not attacking them vigorously. Those ways can be adapted to micro-robot as well. We even have work to use cell membrane to coat the micro-robot surface so that they can last longer inside the circulation, yeah. Another question is like how much does it cost to print those micro robots? Is it like very expensive or is it not? No, I think like if we think about those magnesium or zincable ones, those are based on particles. You can even say how many dollars, like $10, $20, you can get upon it even. It's very low cost. If 3D printing, I think in the matter of like a material is very low cost. I think it's just printer how scalable. And so far, still, we can do like a siren of this in a day for the last TPP printing one. It's not that crazy larger scale, but we can do several treatment sessions already. So far, I think for those 3D printing one, the cost is, I would say, time consuming is not very, very short, but it's really as the printing technology improves, they can print faster. And this tumor treatment itself, I mean, cost matters, but it's not like you have to lower will cost a wearable sensor $1 per patch in that case. I think we can still print the sirens of these in a day. But if you use deposition to prepare those polymeric particle, we can even do millions in like an hour, depending. But 3D printing is usually not a fast process, but we think it's pretty good already with the current speed. But of course, people demand a fast printing process as well for printing technology limitation, yeah. Thank you. Thanks. Do we have time for one more quick question? Here. So I echo my fellow question askers to thank you for a great talk. Thanks. So I'm just curious, when you deploy these robots, are they deployed in isolation? Like I see a lot of these videos just involve, you know, one robot moving. Or have you considered and your team considered deploying like a team of robots, maybe with heterogeneous capabilities so that they can synergize their abilities and maybe achieve. Right. So actually, most of the applications, they are not a single one. Think about, I already showed here, I believe. The swarms, yeah. Right. So think about, we have many, they are moving together. So when we do the in vivo application for the magnesium ones, we are using like millions of robots in one shot. This we are using, actually when we do this in-vivo treatment, we are doing a few syringe in one treatment in this case here. You can see many they are moving in the same direction as well, yeah. One robot cannot do much, they are too small. We need syringe of even millions of these. That's how nanoparticle research is doing, right, yeah. But it's very easy to make this many also, yeah. And do you apply the same control to all of them? Right, right, right, right. I see. As I mentioned earlier, we have the works on the way to be published. We don't need to control the chemotactic behavior that we autonomously, many, many, we will target tumor directly, yeah. Thank you very much. Thanks. Let's thank Professor Gao one more time for his great talk. Thank you very much.