 Good afternoon, everybody. This doesn't make any sound. It's mostly for the camera. So even though I'm holding it, you're not going to feel any effect. So I'm Claude Cochini. I'm a professor of mechanical engineering and the associate dean for academic affairs. So welcome to our next celebration of faculty career colloquium. Just to give you a very brief background, this is something that was instituted by the college a little bit more than three years ago to give the opportunity to senior faculty, those who have become full professor and have been at that rank for more than seven years to give a talk like this where they get to share their careers and their experiences with faculty, students, staff, et cetera. And then they get a chance. They're supposed to do this once every seven years. And then they get a chance to talk to the dean and the head to plan for the next seven years. And so today, we have the distinct pleasure of having Professor Paul Soika. And he's been here since 1983. That's 33 years. So that's actually more than me. So I always feel great when we talk about faculty who've been here longer than I have. And he's going to talk to us about, well, you see the title there. So all yours, Paul. For those of you who are students, you have just been exposed to what is termed as admin speak. So when Professor Kokini says opportunity translated into plain language, that means you will do this. It's only a question of when. And by the way, it's going to be in 2016. So anyway, Claude, thank you for the opportunity. What I'm going to show you, hopefully it's going to take more than 45 minutes, is me going backwards in the field of science and engineering opposite to what everybody else has been doing since I actually got here to Purdue. So we're going to start off with who I am. It seems that lots of people believe that old white guys like me had ancestors that came over on the Mayflower. As you can see, that's not the case to my most. Come back. So part of the family came from right there. I have no idea how that's pronounced, but let's call it Glaris. There is a new Glaris in Wisconsin, so I'm assuming it's the same place, or was named after them. Come on back. Another branch of the family came from Luang. Well, actually, that's where they embarked for their trip to the US. The next group. Yeah, OK. Now we're starting to head a little more east, Koblenz, which at the time was the Austro-Hungarian Empire. And rumor has it that the family bolted for a better way of life because the Austro-Hungarian emperor was conscripting all of the men of about age 18. And they wanted to sneak one of the older brothers of my great-grandmother out of the country. And then my father just tried to teach me today how to pronounce those names. I wasn't a very good student. I can tell you that in the first one, the last word is tarnaw. And what you see is a little farming village in Poland. And there's one more. I have no better idea how to pronounce that one. But here, what you're seeing is, oops, wait a minute. Let's go back. We need to be out in this area right out in here. And that's where my grandmother came from. And it's another little farming community in Poland. OK, so that's where everybody comes from. So what about me? Oops, going to give us some more family background first. All right, so this is the original family settlement in northern Michigan. If you're wondering why there are no trees over here, well, there was a lot of lumber cut off in northern Michigan in the late 1800s. That's what made Grand Rapids the furniture capital of the world. They would just send it down by river. And all the busy people in Grand Rapids were turning out beds and tables and chairs and dressers and desks and so on and so forth. The original family home is here. That's where my grandmother was born. And my mom and dad now live, can I find that tree? About right there. They were the first people there. And so they had a bridge named after them because it went across the river right there. And in keeping with their anglicization, they changed their family name, which in French would be Stéphane, into Stéphane. Maybe they had bill collectors who were looking for them. I don't know. All right, so there is my mother's side of the family. That is my grandmother right there. They were not rich people. You can see that my great grandmother was not a happy person in this picture. The reason for that was that this fairly large building is much bigger than it needs to be for a husband, a wife, and four kids. And they ran actually a fishing camp throughout the entire summer months, from late April until early October. And my grandmother had to cook for all of the people who were there, breakfast and dinner. And there could sometimes be 25 or 30. So I'm guessing she just got done with the morning meal. How many people know what this thing is over here on the right? Anybody know? That's an outhouse. This picture was taken 60, 70, 75 years after that one. The family accommodations had not improved much, as you could see in that intervening time. And in fact, that's me right there. Check out those cool knickers and that ultra cool hat. And so actually the first two summers that we lived in, what's my parents' house now, that was the restroom. Now my father's side of the family. This is a family home in a little place called Turner's Falls, Massachusetts. Actually, it was three stories. Yeah, you can see them all. And there is my father's family, good Polish Catholics. And this fellow right over there is my dad, the seventh youngest, seventh oldest, so the third youngest of all nine kids. Oh man, I forgot that. I had a souvenir. So despite what Colicott says, I am proud of being born in Detroit. And in fact, there is the hospital that I was born in, Harper Hospital. It's still there. Let's get rid of some of these. And we're going to skip ahead. So this is the physics astronomy building at Michigan State. It was the physics astronomy building when I was there. It's now the psychology building. I don't know if all the physicists went crazy or if they just moved all the staff. But that's what that is. Let's see what else we got here. Oh, and I got a dose of reality because I discovered about three years into my undergraduate education that the only way that you were going to get a job in the physics community after you graduated was to become a particle physicist. And to be quite honest, I did not want to become a particle physicist. I quite enjoyed some aspects of the field, optics, electricity and magnetism. I'm trying to find the engineering building here. Electricity and magnetism, mechanics, and so on and so forth. But there wasn't a whole lot of, ah, here we go. There wasn't a whole lot of market for those sorts of folks. So I changed my major to, sorry, started my graduate program in mechanical engineering. And our lab, where is our lab? That's Red Cedar. Ah, here we go. OK, our lab was right back here in this corner. And in fact, one of my colleagues from my graduate student days, when they poured some fresh cement out here in the back, got the great idea to take a nail out and write Kerber's Laser Lab in the cement. And that was fine until our advisor came walking down the stairs right here, looked through a big plate glass window, saw that, and noticed that the dean who was walking with him saw it as well. Yeah, times were tough after that in the lab. He was not very happy. OK, let's see. I know I have. Can I find? Oh, there we go. So that's me. Eight days after I was bored, you'll notice two things. I don't have any more hair on the side of my head then than I do now, but I have a lot more then than I do have on top. OK, here we go. So now my molecular level of life. It turns out that the Department of Defense has long been interested in lasers as a weapon system. TRW developed something called the Ballistic, sorry, the baseline development laser 43 years ago now. And then something called Miracle in 1985, so 31. And then Alpha, which was part of the SDI program, which many of you may, if you remember that time of year, was what was euphemistically termed Star Wars. So that was the motivator for my thesis research and that of my colleagues. There was actually a proposed weapons platform. They still have it. The Air Force still has it. You can see that it's a 747. And the turret is a little bit easier to see on the right hand view. So all the turret, which is where all the nasty stuff came out is right here. It's got pitch, yaw, and I don't know, maybe not, maybe roll. You tell me. I don't know about those arrow terms. But basically, you can steer it around. Inside the plane, instead of people, you have an enormous amount of fuel. In our case, as you'll see in a minute, hydrogen and fluorine gas bottles. You have some illumination lasers so that the gunner can track the target. You have a ranging system, which uses a CO2 laser, sort of a LiDAR system. There is a bulkhead here, which is really, to be honest, unnecessary. Because if this stuff back here goes boom, all the bulkheads in the world aren't going to save anybody. But in any case, all of the energy is produced back here. There should be some, I guess they don't show it, the optics to get the beam out of the laser itself and out through the nose are running under here. And then up here, we have a pilot, co-pilot, engineer, and whoever was probably the weapon specialist is what they call them. OK, so at this point, Bob could go to sleep, because he knows all this stuff. But if the rest of you don't know exactly how a laser works, and you hopefully do not take Austin Powers, Dr. Evil, as the world's expert on lasers, this is what they actually are. So what you have is what's called a gain medium. So a laser is nothing more than an amplifier. It's an optical amplifier, but it's an amplifier. You have to have a feedback mechanism. And in fact, for an optical signal, there's a mirror over here at number three, and there's a mirror over there at number four. This one usually reflects as close to 100% as one could make. This one is somewhat less, so that some of the signal could be coupled out. And there it is, moving off to the right. Now, the whole idea of the laser is to generate a population inversion. All right, so let's look over here on the left. This is what's called a three level system, because hey, there's one, two, three levels. The laser transition and the signal occurs when the number of molecules, the population, in this metastable level is inverted relative to the ground state, which means essentially there are more molecules up here than there are down there. The question is, how do you get them there? Well, there are lots and lots and lots of ways. So for the moment, let's just not worry about it. We'll just say that there's some sort of excitation or pumping mechanism to get energy from the ground state up to this highly excited level. And then there's some processes in here, non-radiated processes, so collisions and so on and so forth, that gets population down to here, and then it lasers and produces a signal. And in fact, if this is a constant excitation and you don't ever run out of this particular kind of molecule, the thing can run nearly forever. And for those of you who use helium neon or argonion, those are examples of lasers where they'll just run forever, because the excitation source is just some sort of typically electron impact on the lasing species, which produces this inversion and on and on and on it goes. So the candidates, as far as the Air Force were concerned, candidate molecules were these five. Carbon dioxide, which is a linear polyatomic. Hydrogen fluoride, which is a diatomic. Deterium fluoride, which is just the deuterated version of hydrogen fluoride. Carbon monoxide, and there's a reason there's a DF slash there is, is because you use the energy that came in the formation of the DF to transfer to the CO2 by a collision. And then oxygen iodine, where excited state singlet delta oxygen was manufactured in a chemical process, and then it transferred energy to the iodine atom, which excited it, and then put out a signal at 1.3 microns, I think. Now, it turns out that it's very important to know all about the energy modes and so on and so forth, the internal energy modes of any of these. Doesn't matter which one. And so in the case of a triatomic molecule like CO2, there are actually four modes. I'll show you why in just a minute. So you have the so-called asymmetric stretching. So in that case, the two oxygen atoms are going back and forth like this. They're moving asymmetrically about the carbon atom. In the symmetric stretch, they are moving outward at the same time, inward at the same time, outward at the same time, inward at the same time. And then the bending mode, and there actually are two, because there are two non-zero moments of inertia on this molecule. So you could think of that as a flapping mode. So they're either flapping like this or like that. So four different energy storage mechanisms for vibrational motion. These things can also rotate. So there are actually two different rotational modes. You can rotate about that axis there with a non-zero moment of inertia. Or this axis here with a non-zero moment of inertia. And then finally, they have translational energy, which is just the motion of their center of mass. So the whole thing behind figuring out how these things work and designing them comes in being able to decipher ways that this energy, this vibrational energy, is transferred to that rotational energy or perhaps to a different vibrational energy. That rotational energy is transferred to another rotational energy or perhaps translational. So we have a governing equation. It doesn't look like anything we've ever seen in mechanical engineering unless you work with Professor King or Professor Luckt or Professor Meyer. Besides them, anybody know what that is? You can tell us how to pronounce it. This is Schrodinger's equation. The first term over here is the kinetic energy. This is the potential energy. This is the total energy. And each one of these things is an operator. And they operate on something called the wave function. And the wave function, when you multiply by its complex conjugate, it gives you some idea of the probability of molecules being in this, for instance, vibrational mode or that one or that one, this rotational mode or that one, and in fact, even the different translational modes. So when you solve the problem for rotation, this is called the harmonic oscillator. It looks just like the harmonic oscillator you learn about in classical mechanics, except this is an eigenvalue equation. So you get discrete solutions for this thing, which give you discrete energies. So here are various rotational energy levels. Rotation looks something like that. You could also solve this for the vibrational energies. So here is an energy diagram. Don't worry about these units. But way down here is the ground state energy. That's the first excited level, second excited level, third excited level, and so on and so forth. So we've done that. So now we know what the levels of energy are for this rotation part and this vibration part. But remember that the laser part requires this transition here. So we would like to know what the probability is in getting from here to there. And in fact, if you want to really know, hold your breath and wait for Professor Lux's book on angular momentum and molecules with publication date when? Don't hold your breath. Well, in any case, the governing equation for this, for that transition rate is this thing. What this is saying is if I have one state, let's say j equals 1 here and another state, let's say j equals 2 there, what's the probability of emission from j equals 2 to j equals 1? That's due to a dipole moment, which is the thing that's written in here. So that's the other thing you need to know. And then finally, once you've done all of that, you can figure out where the emission wavelengths are going to be. So right here is HF. It has an emission band around 2.7 microns. Here is 4.3 microns CO2, which we looked at. This one is 9.6 microns CO, 9.4, I think. And this is 10.6. And then where's oxygen iodine? It's got to be over here. There it is. Down around, like I said, 1.3 microns. So now we have the energy levels and things. So let's see what makes the HF laser in particular work. And in fact, it's a combustion-driven process. You have a fuel. That's good old hydrogen. And that's a great fuel. And you have an oxidizer. You never call fluorine good old, because it's really nasty stuff, but it's a great oxidizer. So if we have, for instance, a hydrogen atom running around in a fluorine molecule, then the hydrogen will take one of the fluorines as its own, leaving another fluorine. So we end up with hydrogen fluoride and an extra fluorine. That fluorine then can run around until it finds a hydrogen, forms HF, and leaves a hydrogen alone. And back we go. And this is nothing more than the classical chain reaction. What makes hydrogen fluoride special, at least as far as lasers go, is that unlike what we tell our introductory thermodynamic students, the enthalpy of reaction doesn't just go into translational motion. It derives the temperature. What happens instead is you get this very useful pumping distribution into the upper vibrational levels. So what this chart here is telling us is that when that reaction occurs, that's the H plus F2 one, so that's this one. When that reaction occurs, about 10% of the HF molecules have a vibrational level 3. About another 14% are at 4. Now there's some disagreement, but let's take this curve. The 5s are at about 20, I don't know, 3% in vibrational level 6. It's about 32%. And then here's 7 and 8. So what we've got is we've got a chemical method now of forming the population inversion. 6 is greater than 5, 5 is greater than 4, 4 is greater than 3, 3 is greater than 2. And so in fact, what we should expect is we should expect to see laser signal out on this wavelength, that wavelength, that wavelength, and that wavelength. And over here is an example, a little cartoon of how that works. This is set up for only v equals 2 and v equals 1, but you have the squiggle lines mean a radiative transition. So you have a photon coming out at this point here. And that's for upper level v equals 2, j equals 4. And then this is 5 and 6 and 7. And in fact, there are some things called selection rules and so on and so forth that tell you how that will happen. But if you really want to know, then get your copy of Professor Luck's book. He'll probably even autograph it for you. OK, again, lasers are nothing but optical amplifiers. And so one of the things that you want to know is what the amplification is on all these different, you can think of these wavelengths as channels, the amplification on each of these channels. And my job was to go to New Mexico and work in an Air Force lab and do that. So what we're showing here are gains. It's called small signal gains because it's a very low perturbation signal. And this is for four different transitions that start in the first vibrational level. This one over here is for three different transitions that start in the second vibrational level. And all those dots that are on there were placed there by hand. And the curves were drawn with the world's best fitting tool, which is anybody know what a French curve is? Ganesh says yes. OK, those of you who know what they are know how wonderful they are. Nobody can ever question you about what polynomial order you used. But down here are computations that I ran to try to match my experimental data. And that's not bad. I mean, the scales are the same, qualitatively, they're the same ratios and so on and so forth. And these are not quite so good, but once again, not so bad. Over here on the right are signal output. So this is power coming out for different emission wavelengths. So this starts at vibrational level four and rotational level five, they're in the center. And the length of the line is how long that emission was going on for. You can see that it's not very long, a few tens of microseconds. And if you compare A with D, these are model simulations that should match up with this. It's not so bad. It was good enough to get me my degree, so I was happy. And you can compare these with those and so on and so forth. Well, what about those models? Well, the thing about those models is you have to be able to track all of these individual molecule states so you can figure out when you have your population inversion and how strong it is, and so therefore how much intensity you're getting out and so on and so forth. And in fact, there are a multitude of intermolecular energy and intramolecular energy transfer processes that are important. Here's a generic HF molecule with vibrational level V and rotational level J. And it smacks into some other species. Could be anything. Could be oxygen, nitrogen, argon, another HF, F2, H2. And what happens is there's a rearrangement in its energy that goes from this original vibrational level V to some different vibrational level, V prime, typically lower. And then the excess energy is consumed or is transferred into rotational energy, but in a larger amount than what it had originally. Now, that's not good because you want as much of the vibrational population inversion as you can. So one of the reasons you do this is try to figure out what the loss mechanisms are. I wasn't going to show that, but I will show you the detail of, all right, it's stubborn. I have here, oops, there it is. So this is, well, this is obviously highly classified information because it's black. Now, let's see if I can find this again. And just tell me when it comes back. All right, so let's get that out of the way and go down to this one. So this actually is one of the two papers that came out of my PhD thesis. And I'll show you pretty quickly why things get complicated. So here we have the two chain reactions. You might think, ah, from what we learn in thermodynamics, that's probably all you need. Nope, it's not. You have to have a dissociation of fluorine and a dissociation of hydrogen or recombination reactions because they can either accelerate or decelerate the chain, either recombination or dissociation of HF for the very same reasons. Now you get into the tricky part because now you get into what's called the vibrational to vibrational energy exchange. So here's one HF molecule hitting another HF molecule in a different state. And one comes out with one less quant of vibration. One comes out with one more. There's vibrational exchange with hydrogen. Not done yet. You can have rotation to rotation energy exchange. So two different vibrational levels, two different rotational levels. The vibrational levels stay the same, but the rotational quantum number changes. And then finally, you have to include the transformation into translational energy because that's what provides the energy, temperature rise. Here are the lazing rate equations. And that's how you write them all down. And let's see, still not done yet, not even close. So here is the generic change in the species of any vibrational rotation level in HF. We have power added, in my case, due to a flash lamp discharge. We have energy loss due to vibrational to rotational translational energy exchange. Rotational to rotational, vibrational to vibrational. This is actually lazing into, lazing out of. That's vibrational lazing. This is rotational lazing into and rotational lazing out of. Here is the photon flux equations. Now you might think, that's easy enough. Well, no not, because you have 275 or so different colors of photons that you have to track. And then we get down here. That's the gain equation, or the threshold gain equation. This is the actual gain equation. There's the first law of thermodynamics. Finally, something that we think we understand. Here's the laser energy output for another type of transition. That's the total pulse energy. There were 575 roughly species in this code, which was 2,400 lines of executable. Not common. So pretty sophisticated stuff. I think we're done with this. Yeah, we're going to skip that. So now, color scheme changes, now on to Purdue. And by the way, when I first got here, it was not known as the Maurice J. Zucrow Labs. It was TSPC, or the Thermal Sciences and Propulsion Center. So when I showed up in 1983, Claude will remember this face. This is Win Phillips, and Bob probably will as well. He was the department head at the time. And Win told me, no more laser stuff. Go to TSPC. So that's it. I stopped. Done. Never looked at it again. What's the point? The point is, is that what you learn from your PhD is not a particular field, but it's how to be conversant or how to teach yourself to work in any field. And I went from never caring at all about the Navier-Stokes equation to never caring ever again about electromagnetic field theory, molecular potential surfaces, and the like. I was fortunate enough when I got to TSPC to work with that gentleman there. His name is Arthur LeFavor. He was the first Riley professor. And his words of wisdom were combustion and sprays, sprays and combustion, repeat. So that's where my career went. So the first sort of things that I did with Arthur, and we're actually going to look at the evolution of the spray research community. Data in those days was taken using an ensemble measurement. And I'll show you examples of it in a minute. And then correlated using dimensionless groups in much the same way that you see heat transfer correlations when you study that subject. Now, there's lots of different ways that we're done to make measurements in early days. But I want to show you this one over here on the right. What those are, they are molten wax that has formed drops after being sprayed through a fuel injector. So some poor graduate student had to take a whole bunch of images like that and then blow them up. And with their sterret, a machinist's ruler calibrated in hundreds of an inch or 64th of an inch, either case, they had to measure each and every one of those drops. Now, that's actually digital because they were writing them down. There were no spreadsheets. So that's how things were done. And you guys thought you had it so bad. And then, of course, people were also interested in the distribution of mass throughout the spray. And here is what's called a spray patternator from about the same time. It's really nothing more than a flat piece of aluminum with a bunch of slots milled in it and a clear plate put on top. You set it up underneath your atomizer so that the apex was right at the tip. You turned it sideways, started the spray. You turned it back up. You watched until one of these tubes got mostly full. Then you tilted it gently to the side so you don't spill any. And then went back and measured the heights with your trusty sterret machinist's ruler and plotted it using graph paper. OK, after that, actually, this was now the very, very early 80s. Lasers were introduced into the spray community. And here is the schematic of the original forward light scattering drop size measurement instrument that Dobbins and Croco and Glassman discussed. And I think it was 1965. And in fact, what they did was all of this side here stayed fixed. There's your spray. This pinhole on a fold of multiply was traversed back and forth with a mechanical stage. And the signal came out onto a strip chart recorder. And then once that was done, it was all hand digitized. And then there were matrices involved before you ever got to the point where you could have data. In the middle 80s, commercial versions became available. And then it was actually automated. OK, so anyway, throughout my career, I've done a combination of instrument development, measurements, and analysis. So one of the first instrument development actually de-first for me, instrument development projects was what I did with Galen King and Brian Miles. And now one of the problems with these forward light scattering instruments was is that the beam would tend to get diverted if there was any kind of refractive index gradient in here. And that can happen immediately if you have elevated temperatures and double immediately if it also leads to evaporation. So we were bemoaning the fact. And Galen said, that's easy to fix. Just put a quadrant detector over here. And when the beam moves off the quadrant, just tell the instrument to stop taking data. So we did that. And here are examples of how it works. This is where the conditional sampling was used. This is one where it's not. All of this stuff here is garbage. Gives you a horrible result. And then again, here's a case where that conditional sampling worked. Unfortunately, I was not very wise in the way of intellectual property then, because we actually presented this at a technical conference and then got the idea that, well, maybe we should patent this too late. So one of the instrument manufacturers, Simpatec, now has that as an integral part of their systems. This was Arthur LeFavre's first venture into analytical modeling. And the student, Mark Lund, is now down at Procter & Gamble. He's an OME of, I don't know, 15 years ago or maybe more. What we did was, or what Mark did was, is he said, OK, this instability that's leading to these ligaments, we're just going to pretend that it looks like a bunch of ligaments around a circular air core. And then we're going to argue that we can apply fluid-mechanic instability to each and every one of these ligaments. And from that, we can get the wavelength that has the fastest growth rate. And we can plug that wavelength back into a very simple mass conservation expression for a drop. And that allowed us to predict straight away what the drop size was with knowing only the liquid viscosity, the liquid surface tension, the liquid density, and this thing called d sub l. Now, how are you going to get d sub l? Well, I haven't included all that detail. But I will show you that this is the kind of agreement that Mark was getting. And if you knew anything about the spray community at that time and what the situation was, this is remarkable. We get both qualitative agreement as we change, in this case, viscosity and qualitative agreement as we change surface tension. And the quantitative agreement is not very bad at all. Shortly thereafter, another student, Phil Santangelo, who's now down at the University of Georgia, I think in the biomedical school, decided to do holography. And this is nothing like the holography that Professor Chen does. Because in Phil's days, all of the recording was done on a photographic plate, which then had to be developed, wet developed. And so the frequency of measurements, in this case, was about 0.03 Hertz. No, 0.003 Hertz. Because it took about five minutes to process everything. You also had to manually reconstruct it. But even with all that, it was a fabulous piece of instrumentation for figuring out what the actual liquid breakup processes looked like. It was the first time that anybody could say definitively that the little teeny tiny champagne bubble sort of picture that some people had of this particular kind of atomization was not correct. You had bubbles like soldiers in a line coming through. And then those individual bubbles popped and formed all of this chaff as is shown there. And in fact, it showed Mark Lund's hypothesis, or his model, to be correct. OK, so then I started to get into instability analyses. Or I guess I should say my students did. And here is an expression that relates the disturbance growth rate, which is gamma, to where is the wave number for? Oh, it's buried in here. Relates to the wave number for the disturbance. This particular one is written in terms of a phenomenological unrelaxed axial tension, which is right here. That, in turn, is related to a relaxation time for a polymeric liquid and another polymeric liquid physical property, which is there. And that's the retardation time, which looks like that. All right, so now if you know the polymer molecular weight, you know it's concentration. You know it's zero shear viscosity. And some other things you can now try to predict drop sizes. So here is a map of drop size on the vertical axis. These dashed lines are the predictions. And here is Sam's experimental data. Looks pretty good. OK, a little bit after that, Mike Ullam joined our group. And we decided that we needed to come up with a way to rapidly check the quality of consumer product spray nozzles. And so Mike designed this very simple patternator. It has a laser line generator that's sent out into a sheet. The sheet goes through a spray. So the spray is heading down into the screen. It's refocused on the other side down into an array detector that's 128 pixels wide. And one then rotates the consumer product atomizer 90 degrees. And if the trace is matched, then life is pretty good. So that's a good one. And if they don't match, which is the case here, that's a bad one, miss the intellectual property thing again, because at the time that we were doing this, a small company here in town called Inurga was working on a much fancier version, which they sell today, oops, which they sell today. There we go. OK. That's not supposed to look like that. Well, all right. Anyway, at that point, it started to get interested in what's called secondary breakup. And what we have here is a time history of a drop deforming. This drop was a sphere like that, deformed into an ellipse like that, and then had this protrusion, which is called a stamen form, and push its way upstream, have bags form the bags. Oops, the bags would rupture. Well, eventually, this rim is going to rupture, and so is that statement. You can write down the second order linearized, basically harmonic oscillator problem for a spring mass damper system, where the spring is surface tension, the mass is the mass, and the damper is the viscosity. And if you do that, this is work from Selyana Lopez, her PhD. You can predict the initial breakup time. So here are her measurements, there are her predictions, and the cross stream radius as different aerodynamic loadings. And you're interested in that information because when this drop gets to the point where x over r is 1, that's when the statement starts to form, and you have to go to a whole other sub-model for your drop breakup. And the combustor designer people were pretty happy with that. So here's some more non-Newtonian spray and drop work. This is Jennifer Mallory's PhD. She's now at Rochester Institute of Technology. When you start to work with these polymer liquids, they become non-Newtonian, which means your Reynolds number gets pretty involved. There are material properties here, n, 8 to 0, 8 infinity, yeah, oh, n lambda that have to do with relaxation times in fits of the shear stress versus strain rate relationship to experimental data. Here are a couple of impinging jet breakup processes shown, flipped. This is flipped 90 degrees. And she developed a model as well. It was based on one by Dombrovsky and Johns over 50 years ago. And then it's a force balance model. So we'll zip through these. We have a pressure force, a surface tension force, an inertial force, and a viscous force, and a shear stress. And the whole thing gets put together into this ugly thing here, which is actually called a dispersion equation. And it relates the disturbance wavelength with the growth rate, or did the growth rate go up, growth rate. And from that, she can predict that maximum sheet stability wavelength. Experiments are the triangles, boxes are her predictions. And then, OK, and then now what's going to happen next is that the next student moves to the instability on those pieces there. All right, let's see. This was some more stability analysis work. I'm not going to talk about that. So here is more drop breakup stuff. This is more recent student Varun Kulkarni. You're looking at airflow here. You're looking at the Navier-Stokes equations for the liquid flow there. This is a boundary condition. And when you do a whole bunch of algebra and make some substitutions and introduce some real logical equations of state and take a first order nonlinear perturbation to this, you get two ODE's, which can be solved. And those give you predictions of the transfer size. And you can see that it fits the experimental data very well. Now, most recently, I was working with Dan Herlman and Professor Chen. Dan's at Sandia National Lands in Albuquerque. And between the two of them, they developed what's called digital inline holography. It's a big advance over the stuff that I showed you earlier that Phil Santangelo had done. Basically, we're looking at a process like this, where a drop deforms, back blows. Well, the back forms blows out. And you end up with chaff. And here is the rim right there. And these are drops, fragments, after the rim undergoes its own disintegration. Here is data straight from the digital holography system. You can see a combination of both the rim and all of the individual fragments. Because it's time resolved, you get velocities, as well as sizes, with mass closure. So now, anybody who's using these kinds of processes in their big computational models for gas turbine engines or IC engines or whatever now has a much more reliable sub-model. And let's see, you also get the drop size information out that we used to get using PDA. And then finally, some work from my sabbatical. So this is another digital inline holography system. This is drop impact. This was done when I was in Sandia last year. Here is the process. And if you look over here at the bottom row on the right-hand side, the arrows are all of the drop velocities of the various fragments that are spit off. And then they're color-coded for the sizes, with the red being the biggest. So now, you've got information on everything that you would want to know in a spray situation, which is sizes, velocities, and mass closure. Oh, yeah. And then here's some more data. I can just skip over that. OK, what next? Notice that it's now gone to gray. Well, that's because I've gone to gray, all right? No comments, Steve? All right. So it turns out that the ag spray community really needs my help. And that doesn't sound, the ag spray community doesn't sound all that high tech, but it turns out that there's an enormous amount of technology that needs to be applied because of this problem called spray drift. And spray drift is essentially any spray product that doesn't go where you want it to go. So in other words, crossing this imaginary boundary, here's stuff that's being rolled backwards, maybe because of a win or what have you. Oops. Self-correcting here for time, right? So I play to do a lot with that. Use digital inline holography to look at actually the breakup and the formation of the drops. Use PDA, particle-faced Doppler anemometry and patternation to get initial conditions, and then I hope to build up a new wind tunnel for drift measurements. And then finally, Dan Bieldenbesser got me started on, well, actually got me started on, the detailed physics of charge drops. And so now it's time to go back to that and apply DIH to charge to mass ratios, look at some synthetic plant behavior for ag sprays, and so on and so forth. And of course, thanks to all the people who did this. 67 thesis graduate students, five. Is that the way to spell engineer Timo? I hope so. 67 undergraduate researchers for undergraduate honors thesis students and lots and lots and lots of colleagues. That's it. Yes. Oh, it turns out that a lot of powder coating these days is done using charged bits of powder. And so you have a charged particle, in this case, not spherical, it's some of your regular shape. You have a charged particle, and you ground the bike. And so then there's something called an image charge, which causes the charge drop or whatever to be attracted to the metal and to bond and distribute, hopefully to distribute itself more evenly than if you just used a spray gun. It's also used in the automotive industry, but I think bikes are cooler than cars. And so I didn't want to use cars. All signal gain measurements in the HF. So there was a guy named William Jeffers at that time who had a company called Helios. And he was making small CW HF lasers that ran on electric discharge through SF6 and knocked a fluorine atom off the SF6. And then you co-flow hydrogen and then some diluences. I recall there was some argon, and maybe a little bit of oxygen, I'm not sure. But anyway, that was a CW source. And you simply adjusted the power out so that you could convince yourself that, in fact, it wasn't saturating the gain. Now because that only runs on the F plus H2 transition and you don't have a chain going, you only get pumping up to vibrational level 3. So you can't do the gain any higher than a 3 to 2 or a 2 to 1 transition. I'm just so damn happy that DIH is here that I haven't even bothered to think about it. But if you want to answer that, turn around 180 degrees and ask Terry what he thinks. Because he's a guy who's doing a lot of work on multi-phase flow diagnostics, especially at very high optical densities. And he's been doing that work for, when did, 96? Is that right? 2006. OK, 2006. But he would be the guy. I'm happy. DIH is going to see me to the end of my career. So I'm not going to think about that. Thank you.