 Okay. Thank you very much for the introduction. Again, my name is Debbie Sineski. I'm coming from the Aeronautics and Astronautics Department. I also have a courtesy in the Electrical Engineering Department. And today I'll be talking about robust sensing technology for energy applications, primarily covering the design of high-temperature sensing technology, which we can use in a variety of power and propulsion applications. So as you may know, sensors are now a huge part of our daily lives. There are more sensors in this room than there are people, right? You guys are probably carrying about 10 sensors in your pocket or your purse if you happen to have a smartphone such as the iPhone. And sensors are useful because they actually help us provide data to our smartphone about our movements, our placements, and that allows us to interact with the phone in new ways. In addition, sensors allow this device to operate in a more energy-efficient fashion. So for example, the ambient light sensor on your smartphone actually will dim the display based on that optical input with the hopes of, again, reducing the power consumption of the phone. So there are other very complex systems that utilize sensors heavily, such as the passenger vehicles. As you can see here from this image, there are a whole slew of sensors distributed throughout the automobile. These sensors are used primarily for safety, such as our airbag deployment system. We actually have accelerometers embedded within the automobile that deploys the airbag upon high levels of deceleration. In addition, sensors tell us about the state of our engine that helps with maintenance and also energy efficiency as well. So recently there is this vision of creating an industrial internet of things and leveraging data and sensing technology in order to create industrial systems such as gas turbines using power plants and the entire operation of the power plant as well through the use of sensors. So the idea is creating instrumentation devices that can be located on these complex systems that actually haven't been updated in many, many years. So can we have this new data stream coming out and use big data analytics to process that data and then use centralized data visualization systems, maybe your cell phone, for example, in order to actually observe critical information about the way your system is operating, and then use physical and human networks, such as the internet, for example, to actually spawn a reaction based on this incoming data. And then using the cloud to securely transfer this data and store this data. And so it is approximated that using this type of internet of things or this new type of data collection and processing can actually save $270 billion over the next 15 years. So really great savings through the intelligence and the ability to operate our complex systems and power plants, etc., in new ways. So a lot of our industrial power and propulsion systems require new data sets or collection of data sets. So what I showed you was this industrial internet of things that has these complex systems. And we would like to collect all types of data streams such as pressure, temperature, gas contents, accelerations, for example, and feed that data into the cloud. Unfortunately, a lot of our power and propulsion systems found in subsurface environments or within combustion environments actually have harsh environments within them. So it's actually quite challenging to collect data within these types of conditions. And so robust sensing technology is required in order to collect these desired measurements at extreme temperatures, for example. So a harsh environment to me, and I'll be using this term throughout the talk, is extremes of pressure, temperature, shock, radiation, and chemical attack. And the ability to sense within harsh environments allows real-time monitoring of subsurface environments, combustion processes, and also critical components that are prone to deterioration as they're located in hot spots. So just an example of how sensing technology can actually help with a power generation system. Imagine co-locating sensors, electronic components, and also energy scabbing components on a single chip. And so that chip would actually collect data within a gas turbine, for example. Perhaps this chip is instrumented and located on a turbine blade and wirelessly transmitting data out to operators. This would then allow this complex system, this power generation system, to operate in a more energy efficient fashion and also provide clean emissions as well. So it is approximated that if we improve the efficiency by 1% on this industrial gas turbine, we can reduce the CO2 emissions by 400,000 tons per year per gas turbine. So even small gains in efficiency have great gains on the environment, a huge impact on the environment. Also a cost perspective, if one of these blades actually fails, it costs these power plant operators about $1 million a day to open this up and replace a blade. So the ability to obtain information about the state of health within this complex system and within this harsh environment can also produce cost savings. So how is it done today? This is an image of a turbine system undergoing interval based maintenance. And I like to call this the spaghetti soup sensing approach. So where you have these optical sensors looking inside different optical ports of the turbine in order to assess the health of the turbine and also the process as well. So I think we can do better. I think we can streamline this process. I think we can start to think about new ways of instrumenting these turbine systems and collecting data so that the maintenance process is no longer interval based, but it's predictive based. We have this data stream coming out. We can look at anomalies within that data stream and then provide input to operators in a rapid fashion and mitigate any catastrophic events before they happen. Just another example I'd like to highlight is subsurface conditions or power plants, for example, such as geothermal power plants, but also an oil and gas exploration where we have well bores. These are complex systems where data is also desirable. Specifically data about the state of health of the cement casing components and even information about the subsurface environment. So if we can illuminate that environment, provide that information to operators, it can be a powerful tool to run this power plant and also extract energy in an efficient way and in a safe way. So how are we going to collect data within these harsh environments? For the last 10 years, I've been working with a material called silicon carbide. It's a wide band gap semiconductor material. So just like the silicon that you would find in your cell phone and in your laptops, it can also be doped P-type or N-type so you can make junctions, you can actually make transistors out of this material. This material platform is actually being developed for the power electronics industry. So it is a material set, again, that is readily available. Although it is relatively exotic in the semiconductor community, there is a lot of effort in developing and maturing this material platform. So I've been working with this material to make sensing devices. And so this chart shows just a comparison of silicon carbide with respect to other semiconductor materials. So what stands out to me is number one, the sublimation point at 2,800 degrees Celsius. So this is thermally a very, very robust material. It takes a lot of energy to break apart the silicon and carbon atoms. Also, the energy band gap stands out to me at 3.0 EV. That corresponds theoretically to a temperature limit of 1,000 degrees Celsius. So you can have a transistor theoretically operating at 1,000 degrees Celsius, whereas silicon, the band gap at 1.12 EV, corresponds to a temperature limit around 250 degrees Celsius. So your cell phone most likely wouldn't work at 250 degrees Celsius. Also, the chemical stability of silicon carbide stands out to me. There are no wet chemicals that etch silicon carbide at room temperature. So you can imagine it would operate and function within hot oxidizing environments. So this is some results of exposure testing of silicon carbide within supercritical environments that you would find in a geothermal reservoir where the pressure is 100 megapascal and temperature of around 422 degrees Celsius. So again, trying to mimic the geothermal wellbore environment experimentally, we took some silicon chips and before exposure you can see the state of the chip and then after exposure of 20 hours you can see the silicon just starts to degrade due to chemical erosion, whereas the silicon carbide remains intact before and after the exposure test to supercritical water. Also, we're measuring the mass change. You can see the silicon carbide has no measurable mass change, so no erosion that we can observe through this measurement. So imagine taking the silicon carbide and making a sensor chip, perhaps a multifunctional sensor chip where you have strain, pressure, temperature and acceleration all on a single chip. In addition, you would have your underlying electronics and your packaging to be complementary in order to interface well with these sensors, so the ability to take a variety of measurements at elevated temperatures using this type of integrated system. And again, this would be a new type of semiconductor platform leveraging the intrinsic temperature and chemical tolerance of silicon carbide. And so this is some work actually showing and demonstrating the ability to manufacture and test silicon carbide sensors. This is an image of a silicon carbide strain sensor. You see the tines here, so this is a suspended structure. The scale bar here is 50 microns. The diameter of your hair is 70 microns. So this is a fairly small device. You have these tines here that as you stretch the tines, you get a change in the natural frequency of the structure. And you can actually induce frequency or vibration within these times using these interdigitated comb fingers. And as you apply strain to these tines here, the change in the resonant frequency will shift, it'll shift due to the applied strain. So this is a microscale strain sensor made from silicon carbide. And we do that by using microfabrication processes to pattern the silicon carbide and also release the underlying substrate. So this is a freestanding structure here. This device was tested at 600 degrees C using an IR lamp, so locally heating the backside of the chip at this point of the project. We didn't have the electronics. We were developing them in parallel. So the interface electronics were polymer using PCB, so low temperature, but we were able to locally heat the actual sensing device using an IR lamp and then also exposing the unpackaged sensing device to a very hot oxidizing environment dry steam. This is the temperature ramp going from room temperature up to 600 degrees Celsius. The blue dots here are showing the frequency at which that device is vibrating. Remember those electrostatic interdigitated comb fingers? Those are vibrating. As we get to 600 degrees C, you can see that the device continues to operate. We're still exhibiting frequency change in the device. So this is a functioning silicon carbide sensor at 600 degrees Celsius. In addition, I told you we turned on this dry steam. So dry steam is also imparting this oxidizing environment on this device, and you can see it continues to function. We did this on multiple devices, and we thought this was a really compelling result supporting the further development of this material platform. This is an image of an all silicon carbide circuit, individual components. We have differential amplifiers, JFET components, a little bit of a digital logic as well, all made from silicon carbide. This is a cross-sectional image of an individual transistor. You have some PN junctions here and then ion implantation to create your source drain and your gates. This wafer structure is something that you can purchase commercially. We did some post-processing to create the passivation and metal interconnect as well. This is an IV response of an individual transistor going from room temperature up to 600 degrees C. So again, a functioning transistor device at elevated temperatures. In addition, you can perform energy harvesting. So if I'm trying to co-locate sensors and electronics within a gas turbine blade, there's all sorts of vibration, pressure pulsation that you can harvest and perhaps feed that back to the sensors so that you don't have to replace any batteries, for example. So creating this autonomous sensing system. This is a diaphragm here that's sensitive to pressure pulsation. It's made of silicon carbide and also aluminum nitride, which is a piezoelectric material, high-temperature piezoelectric material. So as you stretch it and it displaces due to pressure pulsation, you could actually generate a voltage and couple this to a supercapacitor, for example, and power up your sensors and electronics. This is a larger scale example of a high-temperature sensor node that is wireless using a culpit circuit. So you have your silicon carbide transistor off the shelf, military grade high-temperature capacitors and resistors, and also your sensor device, all packaged on a ceramic board with an antenna so you can perform the transmission of the data. And this was able to transmit data approximately one meter at 400 degrees Celsius. So imagine further reducing this type of device using advanced microfabrication techniques where you would integrate the sensor onto the chip along with the electronics and the passive components and the antenna. So in addition to energy efficiency and the ability to use sensors to allow energy efficiency, I think it's important to take a look at our power and propulsion systems and how they impart harmful emissions on the environment. So as you can see from this chart, our energy sector is actually producing quite a bit of harmful gases, such as CO2, methane, and nitrous oxide, for example. So imagine tackling this problem through, again, instrumentation or co-locating sensors inside of your automotive engine. So the sensors would be detecting optical signatures, pressure, chemical content, feeding that information back to control circuit so that you can adjust your air to fuel ratio or your spark timing on demand in real time. That would allow your engine to operate on new fuels and also in an energy efficient and in such a way that you would have clean emissions. So just taking a look at the California regulations, they're actually regulating the standards on automobile emissions. So in order for a car to have the low emission vehicle three standard or the left three standard, it has to have very low levels of particulate matter and Mognox and CO emission levels. So this left three standard is demanding 73% reduction in emissions by 2025. And this is just a chart showing some of the, again, regulation of the levels coming out of the vehicles per mile. Just taking a look at soot particulate matter. So in this year, 2015, they're regulating .01 grams per mile of soot particulate matter. That corresponds to roughly about half the size of a grain of rice. In 2028 that is further reduced down to .001 grams per mile. So really just very, very small levels of soot particulate being allowed. So this is really demanding new types of technology in order to allow our vehicles to operate in a clean and energy efficient way. So another material set that I've been working with is gallium nitride. It's another wide band gap ceramic material. You can see here the melting point is 2500 degrees Celsius. Also the band gap is slightly higher than silicon carbide. So again, we can think about having functioning devices at temperatures as high as 1000 degrees Celsius. It is also a piezoelectric material and so you can think of new, creating new types of devices with this platform. Gallium nitride is also an emerging platform in the power electronics community due to its temperature handling and ability to handle very, very high voltages. Gallium nitride is being explored to create high power electronics, in particular the high electron mobility transistor. And so you create the high electron mobility transistor or the hemped by sandwiching a slightly doped variant of GAN or ALGAN on top of the gallium nitride. And that slightly doped variant creates a strain at the interface and that strain causes a quantum well at that interface. So a high sheet of electrons, people call that the two-dimensional electron gas. And so that two-dimensional electron gas is a way of transporting electrons. So you can actually make transistors out of this material. What's interesting is that the surface of this structure is very sensitive to ion. So if I have charged species on the surface, that will actually impact the charge characteristics of the transistor. Likewise, if I apply strain to this transistor, I can also change the transfer characteristics. So now you can have a transistor that actually serves as a sensor. And so my group has recently started to explore that. This is a cross-sectional image of a soot particulate sensor. So again, imagine having these charged soot particulates deposited on the sensor. You might remove the underlying substrate for heat transfer. And this could be used in the actual exhaust of diesel vehicles or other passenger vehicles as well to monitor the soot. Just to give you a feeling for the length scale, this is the diameter of just an image of human hair. Again, the diameters between 50 and 70 micron. The harmful soot particulates that we're concerned about are usually in the 2.5 micron scale. So that lends itself well to micro and nanoscale sensing technology. So in my laboratory we've made a very crude soot particulate generator. We use candles actually. What's interesting is that paraffin candles actually deposit positive charge and candles made from beeswax deposit negative charge. And so we're using this to create soot in a controlled environment. This is an image of a gallium nitride sensor before soot deposition and after soot deposition. You can see the dark carbon depositing on the surface. These are the sense elements, interdigitated electrodes, the sense elements covered with soot and then also a zoom in just to look at the topography of the soot. This is the 100 nanometer scale bar. This is a one micron scale bar. And so we are actually able to get particulates that are satisfactory to the 2.5 micron scale. This is just some data of our soot particulate sensor using gallium nitride. This is the current voltage response of the device with no soot deposition. The blue curve is showing the initial response of the device and then the secondary response after we heat up the device to 600 degrees C. So when you're developing high temperature sensors, it's important to actually understand the impact of the temperature on your metal contacts to your semiconductor. So we often have to burn in our devices and anneal them so that the metal contacts are stable because what's happening is your metal is starting to alloy with the semiconductor and it's important to understand that. I haven't showed you any slides on that in this talk, but I'm happy to answer questions about that. But that's something we think about when we design these sensors. So we do a burn in process where the contact resistance increases. And then before soot deposition, we can see the IV response. And then after deposition of soot, we can see the red curve here. There's a change in the IV response. And then after regeneration or heating and burning off the soot, we can see recovery of the response. And so we can use these gallium nitride sensors to create high temperature soot particulate sensors. If we change the configuration of the gate and position nanoparticles that are sensitive to different types of gases, for example, a nanoparticle based catalyst, we can then think of creating new types of chemical sensors that can operate at high temperatures as well. And so we're exploring different types of chemical sensor, chemical sensitive gate materials to detect CO2. And currently, we're looking at copper gold nanoparticles and then other types of copper based particles. My student had a poster earlier today, giving an update on that work. But this is an image of a gallium nitride sensor. We have the source drain and then the open gate with the gold titanium catalyst. And we're in the progress of testing this device under controlled CO2 and humidity environments. And so the idea is that we can relate the concentration of CO2 to the IV characteristics of the sensor. In addition, we're creating high temperature pressure sensors. So as I mentioned before, the gallium nitride is sensitive to applied strain. So if I create a diaphragm where I have applied pressure on top of it, the diaphragm will start to displace and that will induce strain within the actual mechanical structure and change the strain profile on the transistors. So these are some transistor-like structures that are used for sensing elements. The circles here, underneath these circles, are actual diaphragms. And so when we apply pressure, we can then, again, look at the IV response in order to directly sense the pressure. And so you can think about using these pressure sensors inside of actual combustion engines, for example. This is the IV response of our gallium nitride sensor at room temperature, but under induced strain. So you can see as we increase the strain, so the blue curve is zero strain, and then as we increase the strain, we actually get a different IV response. So we get shift in the drain current and shift in the threshold voltage due to applied strain. And we've also studied the high temperature response of the sensing element at 600 degrees C under different gate voltages. So we're, again, able to have a functioning sensor or transistor device at temperatures as high as 600 degrees Celsius. So what I've shown you so far are two material platforms, silicon carbide and also gallium nitride that can be used to make transistors, sensors, energy harvesters and other electronic components such as resonators that are often used in communication. So imagine co-locating all of these devices to create this new high temperature sense node to satisfy a variety of power and propulsion systems in order to collect data inside of them, again data about the combustion processes, the state of health of critical components and the state of health of subsurface environments. There's still a lot of technical challenges left such as where to place the sensor, what are the power sources, data analytics and communications and I think these are all application specific. And just to summarize, I've shown you two ceramic semiconductor material platforms that are emerging in the power electronics communities that can be used to create harsh environment sensor technology, silicon carbide and gallium nitride and these can be used to illuminate the properties of combustion processes, subsurface conditions and structural health. In addition, there are other technologies such as metal interconnect and packaging that are required in order to realize and communication in order to realize a functioning sensor chip. So that concludes my talk. I'm happy to take any questions but before that just want to acknowledge my group, alumni, funding and also collaborators. Thank you, Professor. Thank you. Yeah, my name is Mike Chame with Baker Hughes. We would be very interested in some of the high temperature electronics and one of the parameters that we would be very interested would be density and viscosity measurement at high temperature. Have you had any work on that? I haven't done that directly but I haven't shown we've done some work with surface acoustic wave devices at high temperatures which could be used to measure fluidic viscosity at elevated temperatures as well. What about using the high temperature transistors that you mentioned in putting together sophisticated electronics such as basically computers? Is that something that you would envision can be done? Yeah, absolutely. I think that's what we're moving towards. Of course, the technology is fairly new. We're not where we are with silicon where we have memory, et cetera. I think that's eventually the goal is to create the performance, the ability to process signals, store and self power at elevated temperatures using these platforms. Yes. Thank you. Thank you very much for the interesting talk. Just I'm curious that now you developed the SSE devices or the sensor at high temperature but also in that case you need some kind of packaging materials which can stand against such high temperature. Yeah. Without that, you cannot implement this technology. Right. Absolutely. I think the packaging and the interconnect, as I said, which is really a big part of it because the semiconductors, what I've shown you, they're inherently tolerant but it's how do they interact with the metals. Again, that exterior packaging is critical as well. So there are works being done to create high temperature ceramic packaging as well. Yeah. Do you also develop such kind of the packaging materials using like ceramics or other high temperature materials? Yes, absolutely. And we design those that are application specific. So for an automotive case, you might think about using spark plugs for packaging for oil and gas. It's going to be slightly different. Okay. Thank you very much. Thank Professor Sineski again.