OK so today Speaker is for each of the F. Scott Fitzgerald. So yeah. Mark has bachelor's and master's degrees and mechanical engineering from it was and then came to Georgia Tech also mechanical engineering focusing on droplet generation. He then went and did and oversee post-doc it knew this worked on some microfluidic problems and has no combined microfluidics and some of the men's type of work in the work that he's doing that was C.E.O. for open cell technologies and so we're going to hear about its products and the technology behind it. Thanks a lot. First I'd like to thank everybody for having me here today to talk about our past research efforts a lot of what I'll be talking about we did during my Ph D. work Georgia Tech and actually a couple post-docs and then and then also about our commercialization efforts which More recently started with an open cell technologies so about the first I would say half of the talk is going to be focused on droplet generation but the generation of the droplet how we study the droplets and then how the droplets can be used in various applications. After that I'll move move on and talk about the actual cell treatment application which is utilizing the same device in order to deliver or enable delivery of drugs and D.N.A. into biological cells the first portion of the talk again was mostly performed in the lab of Dr Andre Federer here at Georgia Tech that was during the Ph D. and then two thousand and eight two thousand and nine I did some post-op work there and then the open cell technologies work started around two thousand and nine and it's carried through what we're doing today. So to start out. I wanted to. A little bit of historical perspective on the study a droplet generation so droplet generation by the breakup of continuous Jets due to capillary instability has actually been studied since about the sixteen hundreds the most complete analysis at least to the time this work was done in the eight hundred seventy S. through the eighteen nineties by Lord Reilley. This carried through into the early one thousand nine hundred with a significant increase in research into this area and then it kind of exploded in the one nine hundred fifty S. sixty's and seventy's with the introduction of continuous and drop on demand in Jet printing and that's sort of carried through to more contemporary research and development which is typically focused on a particular applications a lot of them having to do with inkjet printing. So the major I guess changes that have occurred throughout the history of this study of droplet generation. Are in the link scales which as you can see have gone from about five millimeters back in one thousand nine hundred down to less than fifty microns. Which again is not nano scale. We're starting to get there. We're down to about five microns. Maybe a little bit less than that now. And the other major change that has happened was actually in the analysis of the droplet formation process. So as you can see back in eighteen thirty this is actually a sketch that was drawn from a visual observation by survive our move forward to Lord Reilley who actually took pictures on film using a strobe scopic illumination method with a spark gap that he used to. To flash and produce kind of frozen droplets in space and then if you advance forward quite a ways one thousand nine hundred and then to today there's much more advanced technologies and ultra high speed imaging systems they can capture resolutions down to around Micron scale things that occur. Twenty million frames per second is about the maximum There are now this is just a table that shows sort of the evolution of inkjet printing over the last twenty or thirty years. And mostly what I want to stress is the actual size of the nozzle radio starting at around twenty five microns so this is a fifty micron diameter hole moving forward through the ninety's and then sort of what we have today down to about the smallest is around a micron diameter and what I want to note here is that the nozzle radio is really the only way to reliably control the droplet diameter. So if you want smaller and smaller droplets you need to go to smaller and smaller orifice sizes and typically the advancements in the inkjet printing process is and the reduction in droplets size has been achieved through more sophisticated fabrication techniques that allow these smaller hole sizes to be made. As more background I wanted to go through some of the applications that rely on micro scale droplets these range anywhere from biomedicine biotechnology the life sciences down through fuel process. And material printing which is really similar to printing. I'm just briefly going to talk about what the requirements are for some of these applications I'm not really going to get into a description of the application themselves for vaccine and drug aerosolization you want it to deposit in a certain location within the airway and in order to achieve that you want relatively small droplets they need to be micron in scale and they need to be very uniform because you'd like them all to be deposited in the same area of the airway. You also need to have a fairly well controlled flow rate whether this can vary depending on the age of your patient the mass of the patient things like that. In electro spray ionization what you want is to have dry ons and. The liquid droplets that you produce here shown by Taylor cone production and explosion. You want you want them to be very small in this case you actually are down on the nano scale the droplets you want are on the order of one hundred nanometers or so and that really concludes the Nano portion of my talk right there. But that application does get down into the scale. For atomization of fuels the size of the droplets on the droplet uniformity is not quite as important as in some of the applications. Although moving to smaller droplets more uniform droplets can be beneficial for flame propagation during the combustion process. And also you want to be able to achieve a wide range of flow rates with your device. Anywhere from low flow rates for potential low power applications to higher flow rates say in an internal combustion engine and then the final application I wanted to mention briefly is the manufacture of multilayer parts and printed circuits via the printing of conductive solder and again the smaller the droplets the better the resolution in your and your printing. Typically these droplets are around the fifteen to twenty micron range. What's more important for these applications is that usually the material that you're trying to print may be more difficult to print because it's it's properties are not similar to water which for most of these other applications you do have more similar to water's properties typically they have a very high viscosity which is the most difficult property overcome it can be up to one hundred two thousand times the water. So what we've been doing in Dr FEDOROFF lab here at Georgia Tech and then carrying through into Mike into the company that we have started so originally what we wanted to do is to develop a good understanding of the ultrasonic droplet generation process from the particular droplet generator the. That I'll show you in a few slides and then after developing this understanding of how the device works with then I apply that knowledge to solve real world problems and I'll show you a few of those today as well. So the main body of my talk is the outline is shown here. Start with a discussion of the device design some modeling of how the system works and then fabrication move into hydrodynamic investigations which include high resolution visualization So some experimental work and then also simulations of the droplet generation process briefly talk about a few applications that we've worked on here at Georgia Tech that rely on the ability of the device to produce small uniform droplets and then I'll move into what I've been working on more recently which is application of the device to delivery and transfer D.N.A. into cells and I'll mention the company at the end as well. We've just been notified that will receive we're receiving a decent amount of funding starting in January. So we are hiring two positions and I'll show both of those a kind of a last slide so don't leave if you're looking for a job. So the device. I'll be talking mostly about today is shown here consists of three main components. Piers electric transducer shown here in grey which actually weights the droplet generation process. We have an inner polymer partition which basically makes up a reservoir of whatever sample you want to print and then we have a silicon fabricated nozzle microarray And these are put together in this sandwich structure here. If you look at a cross-section of the device you again you see the electric transistor the fluid sample resit war and then the shape of the nozzle array which is a number of peer metal shape nozzles when the piers electric transducer is driven at a resident. Frequency of the chamber you get a standing pressure field that develops inside of the fluid chamber with a maximum pressure gradient near the nozzle tips and that's actually what leads to the formation and injection of droplets I've shown here just a couple of the different device geometries that we've looked at I'll get into that more when I talk about the fabrication. So what we have is a sonic actuator actuation of the device that so that allows us to achieve low flow rate atomization which originally this was used as a fuel processor for small scale power generation and so that's important for that application area. What's also important is that we achieve low power operation through the resonant operation so cavity resonant operation here. And then also through the acoustic wave focusing within these pure metal nozzles and the final thing we have is a simple memes batch fabrication pyramid all nozzles are created by an isotropic potassium hydroxide. Before we actually fabricated and put the devices together we did some preliminary modeling of the acoustic response of the device used commercial software this is these are answers results you see here this is a one millimeter thick trick half millimeter spacer and five hundred micron thick or half millimeter. Silicon Chip that the nozzles would be formed in shown here are the first three resonant modes that correspond to droplet injection within the cavity. So that the first resonant know it's a half wave resonance. You see there is a fairly good planar structure of the way with a minimum at the surface which goes through to a maximum at the focal point of the nozzle where Jackson would be at the second cavity resonant you have a one wavelength that goes from a maximum to a minimum and then to another maximum near the nozzle tip and then at the third fluid cavity resonance two point four meg Hertz. You start to lose a little bit of the planar structure of your pressure field but you can see there's a minimum maximum minimum and then a maximum near the tip. Now if you zoom in at the second resonant frequency it's a little easier to see that focusing behavior. What we found with this simulation is that we get between a two three and three and a half times again between the pressure profile starting at the Ph to electric through to the focal point which is very near the nozzle tip and a large pressure gradient here at the tip again for droplet ejection. If you move from the kind of a space domain into a frequency domain which is shown here and this the electrical him put impedance to the electric as a function of your operating frequency between point three and two point. You see. Kind of a lot of other behavior of the device through this plot. It does highlight the three resonant frequencies where you would get droplet injections so here is one megahertz about one point six and then two point four megahertz you also see that the strongest resonance of the device is actually the natural longitudinal resonance of the. Which is around two point one two point two megahertz in this case you see some other resonances due to different I guess flat walls of the internal chamber structure and then the final thing that we saw from these experiments which is not necessarily shown here but you know we've done a lot of these with various different fluids and geometries is that these ejection frequencies really only depend on the geometry of the device and then the speed of sound within the fluid that you're trying to inject And so this serves as a design tool for if you wanted to switch to say methanol or some other fuel has a lower or higher speed of sound. You can use this type of I guess estimation based on these simulations to predict where. The shift in resonant frequency will occur. So having confirmed that the device operates in the way that we want we get this resonant operation and we get acoustic focusing to generate the droplets and we can move forward to actually realize this in a device the fabrication process is actually fairly simple the first four steps here show the how we prescribe the size and the arrangement of the potassium hydroxide. Nozzles we have two different techniques that we use one in the first terminates before reaching the opposite side of the way for and in the second we actually have designed the K O H edge to reach nitride membrane that's located on the back side of the way for. Them flip it over and a line or to the tips of the nozzles and so if this would be in books so it can serve the Cayo it doesn't make it all the way through the way for use in I.C.P. process. If not we use our YOU process to blow out a small hole in the middle of the membrane on the opposite side and for the membrane devices we typically sputter a thin metal layer tungsten over the surface just to strengthen those membranes. But you see here are three different geometries of the nozzle or if we have actually worked with the first was when the Terminator before reaching the opposite side of the way for and we use the I.C.P. process to form the nozzle and this particular geometry gives a high degree of control over the shape and size of the nozzle you can see in this case the scale bar is about five microns so this nozzles around four and a half microns. It's also fairly robust because the nozzle itself is in the silicon. But you can get a high pressure drop. If the church is poorly designed and it terminates a decent distance from the opposite side of the way for then the Micro Channel that you get can have some link to it. Which can increase the pressure drop for your system. The second geometry we have is a small orifice located in the nitrite membrane again this gives us a high degree of control over the shape and size and location of the nozzle the scale bar here again is about five microns so this or a fish on here is about three microns in diameter you get a fairly low pressure drop with this because the Micro Channel at the tip is very short but the membranes are fragile so if we drive the P.A.'s electric too large of an amplitude during operation we do see some blow out of the membrane which is typically beyond the operating conditions we'd like to be you do have to be careful. The final geometry is basically potassium I'd go all the way to the opposite side of the way for and then we just remove the nitrite and so in this case will get a square or dictated by the self. This gives a robust opening because again it's in bulk silicon and the lowest possible pressure drop because there is no Micro Channel. The only problem is that it's fairly difficult to brush prescribe the exact size of the orifice using it that started on the back side with a feature that's about seven hundred ten microns it's hard to go through. Through the way for to get down to a five micron orifice. And so you see the scale bar is about fifty microns this or this is around forty five microns. These are actually pretty good for the cell treatment application that I'll be talking about because for that you need the whole to be larger than the so you're pushing through so. So that pretty much summarizes the initial development of these devices at this point I guess in my Ph D. project we wanted to determine how well from a droplet generation perspective the devices perform. So we perform some visualization and also computational fluid dynamics modeling visualisation process. Here I went back to really tried and true method this technology is from about the eight hundred ninety S.. Other than the fact that our scale has decreased significantly and our time scale is also much smaller basically what we do is we have a function generator that drives the transistor through an amplifier shown here so this would be anywhere from around five hundred kilo Hertz up to around two and a half megahertz the output from the function generator to another or sync up with another function generator that sends a waveform to. For back backlight or illumination of the droplets in this way the camera that we use for collection does not have to be a high speed camera in this case it actually runs about thirty frames per second. So it's pretty standard off the shelf with Quitman the exposure time is such that each one of the images you'll see over here so each frame which are separated by different delays between these two signals each one of those represents maybe ten thousand images superimposed on top of each other so the process is extremely repeatable from cycle to cycle which is demonstrated by. How clear these images are this is around five micron droplets shown here. So with this technique we're able to get a very high spatial resolution down to around one one and a half microns. And again we've visualise a short time scale processes on the order microsecond although the actual delay between these images. Is in the nano second time around. So this is actually kind of a complicated process because in general you want the entire array to function for any of the applications that we're using but when you want to image one of them. You only want one to work and. For various experimental setups we've used we've either covered some of the nozzles so we only get one or in fact if you notice from the pressure distribution is not always going to completely uniform and so. If you if you adjust the signal that driving appears A With down enough to get to close to a threshold of where Jackson would start. You only get maybe ten or twenty of the nozzles within the array working and you can isolate them that little needle that came down to to figure out exactly which one was working so here you see examples some of the images that we've taken. I would drop injection again from a four point five micron or driven around point. Continuous a jet it continuous jet ejection from larger orifice. So this is around sixteen microns driven around point nine magnitude. And then we've also captured transition mode injection where you see what looks like kind of a continuous jet issuing from the orifice. It will go for two or three wavelengths and then it kind of immediately breaks up into droplets and so this again is a transition from being a continuous jet to discrete droplet operation. We have see each of these three ejection regimes for the three geometries that I showed earlier and to stress again the for each of these cases the ten frames that you see here represents a full cycle of ejection so I'll show you a moving in a little bit with the droplet injection but you can actually follow the interface as it creates a single droplet and then the next frame in the cycle would be kind of the first frame that shown here for each of the cases and so we did that just by changing the delay between the idea lumination and the electric signal. So we wanted to explain kind of why we get droplets sometimes and jets at some times and so what you would you need to look at is the pressure gradient that you're applying. Produce these droplets is applied in a push pull manner and so the push half of the cycle. Obviously fluid from the nozzle as the pressure gradient flips and becomes negative. It actually acts to locally slow or reverse the flow and if it's able to reverse the flow or if surface tension basically is given enough time to break the neck of the jet before the positive pressure gradient is restored. You'll get actually you'll get a droplet right at the nozzle if that's not the case and the neck is not allowed to break before the positive pressure grading is restored you get a continuation and so it's basically the relationship between the periodic forcing or the time scale associated with that periodic forcing and the action of surface tension that dictates whether or not we get droplet or Janet Jackson. This is more easy to illustrate easier to illustrate if we introduce these dimensionless parameters that are based on a ratio of the capillary time scale to the interval time scale so that's a Weber number here. And then on the ratio of the inertial timescale to the forcing time scale for this periodic force in time scale and that's the whole number. So what are explanation was was that if the time scales for the action of surface tension and the time scale related to the forcing frequency are of the same order of magnitude. Then this corresponds to a critical Weber number for the transition between droplets and Jets. It's on the order of the reciprocal of the square of the strudel number that's plotted basically as this line right here. And all of these points are the experiments that I took during my Ph D. you see the droplet ejection for well is found where we expect that droplet ejection the triangles represent transition from droplets to jets and then all the squares here are jet injection and so you see our experimental data does support the conceptual description that I showed on. Last slide is that you know this. OK So as I said we can we can actually take the individual frames and stitch them together and make sort of a movie of the ejection process this again is these are about six micron droplets issuing from a four and a half Micron nozzle operated around eight hundred kilo Hertz and so this really gives a clear image a solution of the droplet as it's growing detaching and then oscillating here as it leaves the nozzle but it doesn't give you a clear. I guess indication of what's happening within the nozzle so we did perform a computational fluid dynamics simulations which is shown here for comparison with this case they do tend to capture all of the important ejection phenomena external to the nozzle So you see the growth attachment shape oscillation of the nozzle interface evolution is captured. I guess fairly well compared to the experiments and you see this cycle the cycle uniformity. What's not shown here the Although you can see a little bit how the enter faces are attracted back into the orifice itself on each injection cycle we do we can get. The pressure gradient information velocity field within the nozzle and this is actually important for some of the later applications with this. Briefly I just want to show we have captured all of the injection modes that we saw experimentally drop which transition and Jet they match up fairly well with the experimental modes as well. They show the transition occurring approximately the same place. So that's about all say about that we have also analyzed a number of other I guess injection phenomena that we can that we can observe with the device. One of them. That's important for some of the applications is the fact that for a concert. Orifice diameter. So the orifice here is a four and a half Micron orifice. We are able to achieve a fairly large range of droplet diameters just by changing the frequency of operation. So from about four hundred killer which we get a drop of that's almost double the size of the orifice up to around two and a half to three megahertz then we get droplets that are smaller than the orifice we get a pretty good range and this is important if you want to if you have an application that requires particular droplets sizes but that droplet size might be different. Say for inhalation vaccination which will show in a minute. You can use the same chip to achieve different droplets sizes for say different patients or I guess different drug formulations that may end up in different parts of the airway. So now I'm very briefly going to talk about a few of the other device applications that we had worked on in the federal lab before I move over to a much more detailed discussion. Excuse me more detail discussion use of the device to enable drug or D.N.A. delivery into cells. So the first application I want to show is actually the aerosolization a measles vaccine for delivery. If you look at the graph over here. It shows. Basically what the droplets size distribution you want for deep ization into the airways and then the mouth and throat would be in this case we're actually trying to get the upper portion of the lung. So we want to around four to five micron diameter droplets here you see these are actually based on our visualization experiments it's based on a phase Doppler particle analysis so it's a laser diffraction technique to measure the distribution of the droplets sizes for three and a half six and eight or so and seven Micron droplets. You see the distribution slowly decreases into this range that we want to achieve. Now it would have been nice to form these experiments with that device at a higher frequency operation but we hadn't actually completed that characterization work after we had done this work and so it should be possible to just shift the distribution for the three nap micro nozzle down so the more of it resides in the range that we need just by operating at higher frequencies. Well so use the device. Interface for an aspect ometer as ion source. So in this case we're able to decouple the droplet formation and charging processes the devices produces the droplets we can either use the extra. Electrodes to charge the droplets or put a bias on the P.S.U. electric itself. Here you see the device coupled to of been Terry device which is basically an air flow device that focuses on a D.C. of it's the droplets so we can get dry before we get to the mass spectrometer. Detector. What you see here is an example spectrum that was actually achieved by Tom and Christina. And what's important here is that we're able to capture both dimer and the try and understand is very difficult to do with conventional Yes. And this and that has sought by ization was achieved with ninety nine point nine percent water as a solvent. So we didn't have to use fifty percent methanol which is typically used to lower the surface tension of the solution for electricity and as ation and then the final application mentioned is similar to printing although we're trying to print higher viscosity materials to demonstrate that we could use this device for material printing say three dimensional parts in typical printers that are used for this application this printing. Indicator which is given here is a function of density surface tension viscosity and then the radius of the orifice you're printing far from is limited to a range between one and ten which for practical terms limits your viscosity to about forty times that of water and your surface tension needs to be greater than about thirty percent of water without a drop of generator. We have demonstrated a consistent operation for mixtures of glass or all of water with this cost is that cause the printing indicator to be side of that range. So we've gone up to about three hundred thirty times that of water. We've also demonstrated very brief and when I say brief I mean on the order of one to two seconds with a much larger viscosity around three thousand times that of water. We are encouraged by that result that we will eventually be able to achieve more consistent operation with those fluid. And so now I'm going to switch gears quite a bit first. A little bit more than half of my talk with focused on this droplet generation how we study droplets and what we can use those droplets for. Now for the last portion of my talk I'm going to talk about kind of a totally different application of the device so this application doesn't actually rely on the ability of the device to produce smaller uniform droplets but it relies on or takes advantage of the ability of the device to produce these focused. Mechanical fields for the device I've shown you this far and also electric fields at the nozzle tip. So as a brief overview of conventional transfer action techniques. These are the most popular techniques and use your book collector operation or saw an operation. You put your entire cell population in a small bet you apply an electric or mechanical field to the cells this disrupts their membranes leading to. Aeration and then delivered by a molecule of interest. The problem with these techniques is that you are treating a large number of the cells at the same time. Typically in these book techniques the electric field or the mechanical field if it's going to clue stick field is not very uniform and so some of the cells will be treated how you would like them to be treated. So you get the desired result some really aren't treated at all and then some are treated too much. So they die. So another technique that uses a manual micro injection where you take a needle directly insert D.N.A. into the nucleus. This is extremely efficient and basically works every time but you have to do this one and a time a lot of times it's done manually and so the throughputs not very high. It's a very tedious process and then the final and what's actually probably the most popular technique used by Life Sciences researchers is a chemical technique. So you take a catatonic lip mixture in capsule the molecule that you want to deliver the cells cells to inside of that that then fuses with the cell membrane delivers the molecule into the cell. The problem with this technique is that the chemistry for each given cell line is very specific to that cell line and so what's traditionally been done is to use these cell lines that are well established within laboratories and so all your cells will be exactly the same but for applications where your sample is not as margin S. So if you're using say primary cells that are taken directly from the patient. It's chemical techniques may work well for a few cells in the population but they won't work. Well across the entire population. So most of the microfluidic technologies for transfer action attempt to address the issue with the uniformity of treatment of the book techniques. For techniques here. The first one is the most basic basic. You take the electro peroration electrodes put them close together over a small gap you push the cells between them either by pressure pressure or you like to pump them and when they go between the electro Parisian electrodes they're poor rated and you deliver your molecules a second technique shown here actually is a sawtooth structure created by the electrodes. So the electric field between the tips of the nozzles stream Lehi you push the cells through this in this case it's about fifty by thirty microns where the gap is cells from mammalian cells usually around ten to twenty five microns. So they're pushed between those electrodes electro poor rated along the path and then you get your molecule delivered to them they also have techniques that use a mobile immobilization. Here you push the. Into a small channel that's not large enough for the cell to pass through the cell gets stuck there. While it's immobilized you perform an electro pore ation and then you release the. And then this technique shows kind of a combination of the flow through of the immobilisation it's difficult to see here but actually in a small channel. There's another small channel that goes to another fluid reservoir. So the cells are pushed through this small channel they're actually trapped inside that small small hole electro poor rated and again released. So I'm not going to say any of these techniques don't work well they're actually fairly successful at what they do but they all use a ration and so in that way they're different than the approach that we're taking at least the original approach that we're taking which is to use a mechanical poor ration within our device for delivery of the drugs in D.N.A.. So stepping back a moment I want to talk a little bit about what Bio effects you might affect might expect due to mechanical deformation of the cells. So we have two parameters that are found to be important for this which is the time scale over which the cells are treated and then the maximum shear rate that you put the cells under so. For any time scale of treatment if the maximum sheer rate is low and the cell is unaffected the membrane stays intact and you don't have any delivery molecules into the cell. If you increase the maximum threshold value small poor start to open up in the cell. But as long as you don't exceed this limit shown here. Those those poorest are resealable So if you were to take away the treatment parameters the cell is able to heal. Rios and have the molecule inside of it but then if you increase the some limit these ports are no longer able to be resealed basically ripping the cell apart in the cell will die. So the conventional techniques for mechanical deformation and delivery into cells not going to get into the details of the techniques but their coverage domains are shown here in blue for four different techniques. What you see is that there is limited domain coverage. Although the domain coverage is in the desired region bio effects maps in the poor Asian region. The coverage is not very wide scale. Pretty much all of them are done on a large scale or with experiments which doesn't give you any micro environmental control over the individual cells and they're also not flow through techniques so you have issues with throughput you map our technique on to the plot. You see that for each given a race. We cover a bit more of the map than any of the other individual techniques and we could actually probably cover a little bit more if we continue to increase the orifice you would move kind of down into the right across the graph. If you look at one of these spaces say for a forty five micron orifice using a twenty two micron So you see if we. Increase the voltage applied to the P.S.U. electric It also increases the velocity with which are jacked in the cells and this kind of shifts you from this poor ration to a harsher treatment where we could achieve Lysis if we wanted to. So we had some initial proof of concept experiments were performed here at Georgia Tech just before the company was founded here you see calcium the liver into human malignant cells Lynell and four. For three we had two different orifice sizes that we use thirty six and forty five microns density of about a million cells per milliliter and the cell type is about twenty two microns in diameter. Basically the geometry of our treatment device here whereas the sandwich structure that I showed you before held together with binder clips. So it's not the most robust of systems but we were able to get successful results. So for the thirty six micron device that was kind of shown on the previous graph we pretty much killed all the cells that we treated so everything died which is not that interesting of a result to show but for the forty five micron device. We do get a significant uptick population as shown here with fluorescents microscopy we do get some cell death and we were able to show the by increasing the voltage that was driving the P.S.U. electric we were able to increase the uptick at the expense of decreasing the viability slightly. So now we kind of move into two what we're trying to develop with an open cell technologies which is to take this device which is pretty good results for mechanical poorish and buy it so and to add in the ability for Electro pore ration. So basically what we do is pattern the electro pore ation electrodes at the Celts or sorry at the nozzle tip as the cells are pushed between these electrodes and then they still experience the same pressure feel distribution poor rates the cells and and then we collect them say into a petri dish collection to. Which would then have a delivery in transfer action the wrong way. So here you see the realized device. Basically we changed the fabrication process I showed you a board to have metal the position of patterning for the electrodes we put at the nozzle tips. I showed before of the criminal nozzles the chip was flipped upside down faces down put this whole sandwich structure in this housing which is our initial prototype device. We call steam. And as you can see when the device operates it looks kind of like a clue steam mist is coming out of the device basically put a petri dish under here to collect the treated or collection tube. And here you see some of the results we've achieved with this device. Some are key results which are not necessarily a drastic increase in the taken by ability but we do get a much higher collection of fish and see. Whereas before the device was really not that robust again it was held together by a binder clips with this prototype system we get pretty consistent operation. We get fairly minimal clogging for the cell types we've looked at this is actually for human embryonic kidney cells same cell density but the cell is a little bit smaller diameter with around fifty micron diameter chip and we've also demonstrated which is probably our most exciting result to date transfer action into the same cell line with the D.N.A. plasma green fluorescent protein. So this is twenty four hours per stream and we see this this transfer action or expression of the G.O.P. This shows the actual so. We get about ninety percent transfer action efficiency although our viability in this case has dropped a little bit. I think is around seventy five percent and we do see the operation increase is the transportation by about five to ten percent. This isn't a drastic improvement. But it's important to realize that our mechanical poor ations actually fairly efficient. We're trying to fabricate devices with larger holes now such that they won't treat the devices so harshly with the mechanical poor ration and we may be able to get a better idea of what the electoral portion parameters actually do so. That concludes basically the technical portion of the talk and I want to talk just very briefly about a start up company. So the entire research and development cycle which I've worked through the first part of the invention conception portion and demonstration of the device throughout most of the talk has been leading up to this commercialization stage within open cell technologies. So our teams given here. It's me who are professors in mechanical engineering Charlie have There's an associate professor pathology Emory Winship Cancer Institute who performs research with primary glioblastoma cells taken from patients. He's acting as a life sciences consultant right now and we also have a business development consultant. That's helped to kind of move the business forward and I'm really going to talk about this too much. I've talked quite a bit about the development of the device and what the device actually looks like I will know we're also working on a high throughput device where we can put multiple treatment arrays in parallel to do what will be designed it is designed to interface with. More of the high throughput or plate that are done. It's a little bit about the transaction market the most important part here is that instrumentation segment of this market is the fastest growing and this is because it's shown better success with primary and stem cells and finally I wanted to know. We do have funding and so our funding sources right now have been through the program and then also within the Georgia Research Alliance Venture Lab. Project to woo. Program we are a member of the across the street and finally we are hiring as I mentioned earlier we're looking for two full time people probably one temporary employee for about six months. One will be an engineer a spin electrical engineer and that's how we foresee it so be to design and implement a standalone power system right now to drive the electrical been using kind of off the shelf equipment so your function generator of power amplifier. We want to reduce that down into a standalone system. So this is the kind of experience that we're looking for in addition that work with consultants and manufacturers of some of the sub components of the device and that we're also looking to fill life sciences Operations Manager position. It's maybe a little bit beyond I guess some of your experience levels we would like some industry experience although we will consider people that have considerable experience say in a Ph D. program in the life sciences area need to be able to maintain and analyze. Knowledge about transfer action techniques and then they'll be overseeing some of our sub projects that we run with both S.P.I. our program so we work with Emory University and then also supervising and house technicians that will do some of that work. And thank people that have worked on some of the applications that are presented today. So the federal lab. And then other can contributors. Collaborators at Emory C.D.C. and then financial support. So this was a support for the U.S. research has been performed in Georgia Tech and then with an open cell technology to open it up for any questions you may have open. I want other things. So if you are interested in either of the positions that we're. I was hiring for my contact information is here you can come up after the talk as well. Or if you know of anybody that might be looking pass that along sorry. All right. Yeah. So. The device in order to operate to produce a droplets in needs operated a resident frequency but we can control where that resonant frequency falls by increasing the spacing of the fluid reservoir or changing the PH selector geometry as well. So we can shift its resonance such that it falls closer to where we would like to operate the device will operate at the resonance mostly because there's a significant amount of power that's being transferred into the chamber itself but that's not really optimal it leads to a lot of heating which is bad for some of the applications that we're looking at. And them. So the main difference between our technology and how it's a conventional inkjet printer works is that in the traditional printer. They usually operate in a low kill hurts frequency range and what they're actually doing is the if it's peers electrically driven you have a Ph D. there to forms or shears to push the droplets out and so in that way the whole liquid chambers kind of compressed. So you're increasing the pressure within this whole chamber which leads to the pressure gradient at the tip projection. If it's a bubble jet again you know you're using this. Heated fluid to produce a gas bubble which again compresses the whole fluid in the chamber to lead to eject. In our case the pressure within the chamber itself is pretty much a baseline a zero so there's maximum pressure some places minimum is at other places during the injection cycle and so the only place where this pressure gradient exists where this large pressure going to zero at the nozzle opening is that the tip for drop an ejection and the reason for this is because we're operating at the rate at the resonant frequency of the chamber but we're also operating in a much higher frequency so this megahertz frequency range that we operate is not very typical for droplet generators. There are some other techniques that utilize frequencies on that order of magnitude but they're not by any stretch of the imagine any stretch. They're the same geometry as say a conventional inkjet printer. I just want to add to that that the printer was particularly to the picture of one of his logs we actually that's the bridge or is it the organic lead in your here. So it could certainly be used for any of the types of them or talk about the same process. It was you to actually study drop what we guess that you put it it's a pretty impressive machine breaks down there when we go. And the reason that we will perform the analysis for the frequency spectrum based on the electrical in put impedance is because that was I guess a fairly straightforward very easy result to pull from and says to give to give us an indication of where the resonant frequencies of the device would be you can also get fairly similar information about the the. Pressure amplitude that you're actually getting at the at the nozzle tip which would lead to Objection. With an answer. So you can perform the same frequency sweep instead of electrical and put impedance you can plot say the maximum pressure amplitude or something like that at the focal point and you'll get kind of a similar plot which will show and actually I have one of those so this is basically which actually that's electrical improve power. I don't have the pressure but it would look something like this where you would get peaks at the resonant frequency that would show the power so it's not it's not that we needed to use the input impedance. It's just that that was a fairly straightforward way to see all of these different resonant frequencies. If that answers your question. No it's the entire structure. So you could take and actually this gets back to your question about the electrical input impedance another reason reason we use that is because without running any experiments that show ejection you can basically fill the liquid residue or the chamber attach it to a spectrum analyzer and it will generate that plot and so we can compare our experimental device which will show all the resonant frequencies for the loaded structure with the simulation so we use that also as kind of. I guess confirmation that the actual simulation we were doing actually gave us a realistic result. But yeah it's the entire structure by very little of the rate of the array. So it's mostly it's mostly so that we can get through. Volumetric throughput the air assault inhalation application which was kind of what I was focusing on for my Ph D. were also a few atomization we had very particular flow rates that we were shooting for in the air assault vaccine delivery case I think it's three cubic centimeters per minute is what they need to deliver and so you can basically calculate how many of the nozzles you need to generate that flow rate. Theoretically we could go down to basically one nozzle if we wanted to and get down to say. I think you get down to. Probably peacoat leader per droplet so about nano leader ejection from each and every leader perspective that injection for the cell treatment is actually about the highest volume metric throughput because the nozzle is pretty big so each droplet has a lot of volume we achieve. Somewhere in the neighborhood of one milliliter per minute continuous treatment. Yeah. So for me the major challenge the life sciences aspect of using that device my training was in mechanical engineering I have a pretty decent knowledge of fabrication this summer electronics mostly microfluidics but luckily the postdoc who originally done the proof of concept work is Vladimir's Arnett's and he's still here. He's still around. So I bounce a lot of ideas off of him within the development work and then our collaborators at Emory it's just it's been difficult communicating with them. Probably for about the first six months of what I was doing because you know I knew how to use a pipe. I don't know how to do any. It's just it's just a pretty steep learning curve. And since within a startup company and you don't have enough funding to hire five or ten people right away you pretty much have to balance everything and that's the main reason that we're hiring this lifesciences management person so that we have in-house capabilities that we don't currently have. But I would say that's the most difficult part fundraising is not trivial. So right now. Yeah Right now we're targeting scale researchers so pretty much any I guess lab technician that needs to perform transaction in their labs. Obviously this device. We're hoping works better with primary and stem cells which have conventionally been fairly difficult to perform transportation with but the device itself would work with any cell line I mean it will work or anything that these other techniques will work for and actually. What we're trying to do with our fabrication right now is get down to where the active component of the device. So the micro machine portion is disposable so you'd use it for one experiment and throw it away. That's really what the business model is right now for a lot of competitors that do physical techniques but it's really expensive I mean not not ours necessarily but our competitors range anywhere from five dollars to ten dollars per experiment so per transcription that you want to run and the systems can be anywhere from two thousand to ten thousand dollars which is pretty expensive compared to the chemical methods which you buy some chemicals they're expensive as well but you don't have to put out this capital outlay at the beginning. So we're hoping our system when all is said and done we have the product can also beat them on price as well. And then of course we're looking at these parallel treat manner. Scale applications as well. So like pharmaceutical developed now. Would perform these types of transfer action or drug delivery a larger scale things of that nature. Thank you.