 You can follow along with this presentation using printed slides from the Nanohub. Visit www.nanohub.org and download the PDF file containing the slides for this presentation. Print them out and turn each page when you hear the following sound. Enjoy the show. We'll be reviewing what we have learned very briefly. I have just picked up a few slides from before that summarizes or that highlights the important problems from before and also probably a personal perspective that what I see ahead going forward. Let me begin by a quick review of what we have learned. We have learned many things but the essence of what we have learned is not really too many. I will talk about that going forward many of the things that we have learned the principles will remain the same the MOSFET, the bipolar, the diodes, quantum mechanics, statistical mechanics the basic things are not going to change but as we go for new devices how we apply them, apply the basic concepts that probably requires a good understanding of what we have learned and finally I will conclude. Now if you remember in the beginning of the semester when I introduced 606 I said that we learned something about the foundation of physics which is on quantum mechanics and statistical mechanics and they will lead to the transport equation, the drift diffusion equation, the thermionic emission equation and all those and that would lead to an understanding of registers where sometimes you have seen we have shine light on it, we have put various boundary conditions and applied battery across it. We have done all sorts of things and we have talked about diode minority carrier transport bipolar junction transistor various nonlinearity or non-ideal effects associated with it and finally for last 10 lectures or so we had been talking about MOSFET, MOSFET and MOS capacitors. And I said what we didn't discuss or wouldn't this would be discussing in this course are that eventual applications of these devices in variety of configurations for example laser is just a diode but we didn't discuss it but essentially it's just a diode. Thin film transistor I will talk a little bit about that today it's a MOSFET, it's a pure simple MOSFET. If somebody says I have a thin film circuit you shouldn't be worried you should just say okay I know it is a MOSFET I will have to just do the square law, the channel length dependence and everything is exactly the same thing. MEMS, MEMS is exactly again a MOSFET, MOSFET that has a variable the gate capacitance is variable, everything that you have done in fact that you can easily do, easily do the whole analysis, the MEMS analysis is you can copy essentially line by line from the MOSFET analysis almost not everything. So the idea is that although we haven't discussed all the semi-classical devices for various applications many of them if you know these three it's like an alphabet then you can read or apply for any other and finally the circuit systems that is of course a combination of all those things together and that we haven't discussed and the 6.06 was this part. So what is it in this part that you should remember in general? Well the thing that you need to remember is that how we arrived at the transport equation and what were the essential and the transport equation always are so complicated remember three-dimensional equation partial differential equation of the diffusion drift and all those electron holes going through that many times you can solve them numerically no problem and you have done that right you have done that in homework but many times simplifications are necessary and what I've tried to do throughout the course to explain the various simplifications that goes with registers diode and bipolar so you know you have learned about that. So what is it from this quantum mechanics statistical mechanics devices that we tried to understand the thing was that in the beginning the entire quantum mechanics idea of the quantum mechanics was encapsulated in the idea of an effective mass that there were periodic potential I'll come to that and when you solve the Schrodinger equation in the periodic potential we got the band diagram or I'm sorry the EK diagram and from the EK diagram we argued that electrons mostly sit near the bottom and when they sit near the bottom then the curvature of how they move can be described by something called an effective mass and that also gave us this idea about this band gap the separation between conduction and valence band the highest occupied level and the lowest unoccupied yes highest occupied level and lowest unoccupied level and so those two were from quantum mechanics what did you learn from statistical mechanics well farming function that's the only thing we got from the statistical mechanics and that we derived in equilibrium and we combined these two two ideas the two notions into this into this effective parameters one was this effective density of state well effective density of state looks like density of state but actually is very different in some way right because remember effective density of state depends on temperature if you change the temperature there is this kt in 3d it has t to the power 3 halves because effective density of state are not the density of state total density of state but only the fraction that would be occupied at a given temperature by electrons so you multiply by the farming function and then integrate and effectively put them at the conduction bandage or the valence bandage so NC and NV both of them they know about band gaps and band structure as well as the farming functions we talked about mobility and the scattering time which is related to the scattering time essentially that shows any deviation from perfect periodicity that was encapsulated in mobility and scattering time discussion okay once we had that then there is one other thing that we derived that was very important was that at any point in the non-equilibrium conditions under all non-equilibrium conditions the n multiplied by p is always equal to ni squared the intrinsic carrier concentration and the quasi-farm level separation divided by kt right this is something we learned now of course fn and fp you have to calculate depending on various situation but if you know that you can always find out what n and p is this is non-equilibrium statistical mechanics in some way because it happens when fn and fp are different from each other of course when they are equal you you get the original relationship so that is all that we learned from quantum and statistical mechanics now a very quick reminder that how we went there first of all we made that argument that electron in a crystal essentially goes through a periodic potential and do remember that there are different types of crystals we talked about in 1d there is one type in 2d there are five types right five types of periodic things and then 3d there were 14 Bravais lattices there are 14 Bravais lattices and that argument was very important because that allowed us to get the periodic potential and from that periodic potential we solved the Schrodinger equation the age of the age of the deline zone was pi over a a being the smallest lattice spacing and each dot each solution was separated by each other by how much pi divided by na n being the number of atoms right you remember that and so we counted that each band each band if you don't include spin each band has n state n is the number of atoms that you began the discussion with and if you include spin it will be 2n always even and remember how that was used to explain the difference between metal and the semiconductor right that even states will always remain fully filled if you have even number of electrons then it will remain fully filled so then you can have a semiconductor if you have odd number of electrons then one of the bands must be half filled and that gives you a metal so you remember all those discussion about how this information was was used and from there we got the effective density of state and this is let's say like a three-dimensional if a sorry density of state which is counting how many places electrons can set and depending on whether it's a one-dimensional crystal 2d or 3d then you will have different number of states to begin with what do you think this particular one is this is a three-dimensional one because in 1d you'd have a very large density of state in the beginning and then it will go down 2d it will be flat and 3d is start small and then gradually builds up with three halves of the energy that as you go up in the energy it goes up at three halves okay so we got that and eventually we got the effective mass with the second derivative of the energy band diagram you remember that this is for the conduction band goes with the green and for the valence band there was this negative effective mass but then we had negative charges for the holes and then that's how we and caps got a positive effective mass for the holes so that is how we encapsulated information now the reason I bring it back is to make you aware that this is you learn this good most of the things you will be encountering let's say in an air future or the things that you'd encounter in work this recipe and everything will find will be fine so if you have a new device you can just go ahead get what the effective mass is and solve the problem no problem but you have to be careful that all periodic potential need not give rise to an effective mass that's very important anytime you have a new device you should always examine back that your whether your foundations on which you are doing will be doing the calculation whether your foundation is still okay or not right and this is what most of the students don't do or not I should say students most of the scientists don't do they see a new device they bring in the MOSFET open up the book look at the MOSFET equation start from the MOSFET it's not where you should start you should always check back whether your band structure and other information which lead to MOSFET equations whether the foundations are still okay or not that may take 10 minutes if you satisfy yourself it's okay then you go ahead and use whatever semi-classical equations you want to use if it is not then you take a step back and root you then other so one example is is this this graphene the material of graphene and you have done homework on this right do you remember you did the periodic cell and you had to make that argument that the unit cell there is a special way you had to define the unit cell remember you did the homework you also did the homework in which I tried to show you that this is a two-dimensional thing right two dimensional lattice so you know therefore the corresponding ek diagram will be two-dimensional one axis will be kx another axis will be ky and the ages of the building zone it will be pi u over a and pi u over b do you remember that do you also remember that this is a direct band gap material but the direct band gap doesn't occur doesn't occur in the gamma point it occurs near one of the ages you remember this is the homework you did and essentially if you blow it up you will see that the ek relationship is actually linear you see that as the energy is going up this is a cone opening up whose sort of radius which is the k is directly proportional to e there is no parabolic relationship anymore right so in this case you obviously have a density of state by the way so the region that you see with the cone shape region which essentially just accounts for the middle of this region where the density of state is flat this density of state is flat you knew that right in two-dimensional the density of state is flat anyway but in this case a for a constant ek relationship in a state of density of state going up you know three halves or one over square root of e this is essentially flat what happens if you go a little bit above that you can see from the left hand plot you know the e band diagram plot that very close to the conduction and valence band age sort of then it is like a cone little cone coming up but if you go up a little bit more then it's sort of deviating from a pure cone shape region and so therefore the density of state way up in energy and way down in energy will be different right and in here and here at higher energies will be different but for very close to this band age this is the density of state is essentially constant but you also realize that this is not going to work there is no second derivative if you try to compute it you'll be in trouble and I also told you that when you don't have an effective mass only thing you care about is the position of the electron velocity of the electron for that to calculate those two quantities to follow the electrons around where they are you don't need effective mass and please review that part of the discussion because if you know that part then all devices all materials will be fine you either can define effective mass or you cannot both cases you can handle the problem so these would be the two cases to remember now about Fermi functions there are several things we discussed we looked at the density of state of course and from the density of state we we looked into the number of states per energy so for example in a three-dimensional crystal you have at energy you want fewer states e2 you have a little bit more e3 little bit more had it been one dimensional then the e1 would have a lot more than e2 and e3 so the idea is that once you have the density of state you know where they like where to put electrons let's say you have four electrons now in doing this derivation the most important assumption the derivation was simple right just bookkeeping you see you want to conserve energy you make the argument that electron can either be created or destroyed right and every state one electron if you don't account for spin if you account for spin you can put two so we did a very simple derivation you know it's no problem algebra was not a problem but the important point I wanted to make that this way of doing things only occurs if is only valid if it is a extended crystal where two electrons actually although they are in the same energy they are not physically close to each other because if they are physically close to each other there will be strong Coulomb repulsion and we haven't accounted for that Coulomb repulsion here do you remember and the example I gave yes for example there are two students who are not in good terms one could be in the second floor of electrical engineering and one could be in the second floor of chemical engineering they are perfectly happy they are both in the second floor second level energy but they are not physically close to each other so everything is fine this is that argument you make when you do this this type of derivation of the farming function but I also told you about when you put an electron in a localized state like a donor and acceptor in that case of course the farming function gets modified by this degeneracy factor gd that accounts for the correlation now this is very important because although you know dopant densities and other things that is classically treated everybody knows you or now know how to handle this this correlation effect and sort of repulsion effects are very important for example this is for example a molecule right the people you have you heard about molecular electronics right where you put instead of this huge MOSFET you put one molecule and this could be a C60 a fuller in molecule a and this between two contacts two metal contacts so the two sides are these two tips are two metals you have once a molecule and the electrons can go from one to another and the tip shape region that is coming up that could be like a gate if I give you this problem and ask you to do the band diagram you should be able to do it it's a metal semiconductor metal sort of thing with a gate at one point but the farming function for this because the electrons are so close to each other there is no way that they will be far separated so in this case you have to worry about farming function and apply it properly right so that's for that you have to do so you cannot just simply take in that case well I have read in 606 what effective density of state is now let me put it in a molecule and start calculating well that will be dangerous so it's good to if you see something this different from from the things that you have learned in this course then take a step back and ask yourself that what type of density of state should I put here what type of farming function should I put here those things you should think about by the way there are lots of papers people are already working on it so once you recognize there is a trouble then if you go in google or you know search for papers you'll find many papers so it's not a problem but the thing is to identify that you have a problem and then there are lots of solutions so that's something you have to be careful about okay so I have my quantum mechanics density of state if a band gap and all and farming function appropriately modified for the particular situation then you have the transport equation first one is a Poisson equation all the time Poisson equation to me is a single most important equation because everything at the end is governed by the Poisson equation if you do not have good control over the band diagram then no devices that you make are ever going to work you know current comes later first of all you must be able to control the current then everything else sort of follows so that's and we saw graphically how to solve that by band diagram right that's the solution of the Poisson equation and then there were continuity equation then this is a drift diffusion equation of course in many cases we have also applied thermionic emission what is the difference between thermionic emission and drift diffusion in one case scattering is very important and scattering is hiding in the mobility and in the diffusion coefficient right that's where the scattering is is sort of encapsulated if you don't have any scattering let's say across a hetero barrier and the electrons going on two sides without scattering in that case you cannot use this theory then you use thermionic emission theory now many times they will give you the same results you know if you use it appropriately but many other times the results could be could be very different and so the idea is that this is a complicated equation there are numerical solutions available right you can have medi t or you you saw in nano hub.org you have a set of tools many of them are actually just solving these equations numerically and I told you about that also right how do they do that solve it numerically but many times you will actually simplify it and solve it for a particular condition now that simplification is generally very powerful it has been developed over the last 30 years very powerful as a result for many modern devices also it's not only for the old many modern devices those approximations will still hold because you are just approximating a differential equation and so therefore whether it's a modern device or a old device doesn't matter so long you know the basis of the approximation. So quickly reminding you about the band diagram now you should be experts of this I'll be very sad that if students from 606 cannot draw band diagram 15 years from now but hopefully that will happen this is for example a diode very quick reminder follow the rules flat Fermi level the bulk region you put the conduction band let's say in the end region with the separation appropriate separation for the n side and then put the chi chi is the work function and similarly do the same for the p region the Fermi level to ev separation is dictated by the doping right the doping and then you again put the work function down there that's the vacuum level you have to make the vacuum level continuous you have to make the vacuum level continuous and copy down you'll have to copy these things down and whatever you get that's the answer you cannot be wrong so whatever you get no matter what device it is at least with 0th order this is how the device looks like an equilibrium why is it in equilibrium because flat quasi Fermi level is a property of the equilibrium. So equilibrium is in fact defined by flat quasi Fermi level if you have a non-homogeneity then of course there will be a gradient and as soon as you have a gradient then you have a current flow and that is not allowed in equilibrium right no current flow detailed balance every electron every pair of states equally balanced back and forth and that's different from steady state where it's balanced on average in do not individual state by state on average balance so here in equilibrium every pair of electrons you take two electrons from here and here two states from here and here they are balanced took electron two electrons from the far right side in the band diagram and here they are balanced they are balanced in every position in energy every position space they are perfectly balanced at every point so therefore this is what equilibrium how equilibrium is is defined now once the equilibrium is defined I always wanted to emphasize the importance of minority carrier equation right in minority carriers what you do is you ask the question that if I have a complicated device where is it the simplest to solve anytime you have a drift based region which is the majority carrier sites drift is always dangerous why current is small generally current you know if you have a barrier up and down going up and down the net amount of current will always be small majority carriers has a lot of electrons right 10 to the power 18 less small current lot of electrons so if you say j is q n mu e so you can see the electric field will be minus q it will be minus q right because n is very large j is small electric field will be minus q you make a little bit of mistake in the electric field your current is off by an order magnitude let us say right very dangerous therefore the easiest problem always to solve is to go and solve it in a minority carrier site where the electrons are fewer and then you can make the calculations much more easy so here in order to do the non-equilibrium part the most important thing is this part that you sort of slightly out of the out of the age of this slide which is to ground one terminal if you start drawing without grounding all potentials are related all potentials are related whether something is a plus or something is minus it always depends on what where is your reference so no matter what diagram you draw in order to equilibrium ground or one terminal you ground on terminal then all the relative potentials of all the batteries they will be specified and once you specify them then you will shift the quasi fermi levels accordingly and the quasi fermi level should always come to the other side of the junction and stop this is for the low injection condition that you understand but in general this is how you would get started the important point here are to take care of the boundary conditions in the minority carrier side i will not go into details you know that and correspondingly in the majority carriers is equal to p0 and the minority carrier you always get that separation of the quasi fermi level and that's correspondingly equal to the applied voltage va so you can calculate how many extra carriers you're injecting at the age of the p-doped region and on the other side if it's a metal then electrons goes out with infinite velocity so you will set the carrier concentration on the other side to zero if you have a surface recombination velocity as was the case in the last exam then you set a finite surface recombination velocity so first look for the two boundary conditions and solve the diffusion equation in here is there any recombination in this particular one no recombination good because then this is straight line slope at every point is the same that means current is continuous so that means you are not losing particles right so this is this is how it would work so understanding minority carrier diffusion and transport is very important not only for diode but in general for any types of series connected devices many times you will find one region where it's diffusion dominated you just go there and start solving another important one was anytime you have a discontinuity we talked about thermionic emission and anytime you have a discontinuity that's what you should start in the junction region i'll not go into the details but you just to remind you that the tunneling current is essentially oh i'm sorry the total current is essentially equal to the two fluxes different of the two fluxes whichever side is grounded which side is grounded by the way here here the metal side i have grounded so whichever side is grounded you set that quasi Fermi level the same and the other quasi Fermi level you move up with respect to it and you can then make that argument that from the metal side it is not changing with the applied voltage right metal to semiconductor side not changing with the applied voltage where from the semiconductor to metal side yes of course the barrier is low so the more electrons will flow the red electrons there'll be a number of red electrons moving from right to left is far easier or then compared to the blue ones blue ones the fraction you see above the dotted line that moves from left to right is about remains almost the same and we made this argument that all you have to do is to realize only half the electrons move to the left to right what happened to the other half they are going the other direction you know it's a random motion of electrons so only half would go and not everybody will be go or able to go only the ones above the dotted line because the ones that are below they will come there get reflected in the wall and then potential barrier and go back and so that takes care of that q5b over kt a term and vth is a thermal velocity the rate at which electrons are electrons are moving and so you make these arguments about various currents I'm just reminding you and once you have done that you can calculate calculate the net current so try to go ahead and see whether you understand these arguments and we have discussed it in rate length in the class okay bc diff diffusion theory minority carrier transport thermionic emission that's dc characteristics ac most of the time you'll bias it at a given configuration then you'll put a small signal so the bi there is always a battery that large battery va that puts the thing in a some bc biasing condition and then the small wiggle thing that you have that one that's v v0 sin omega t that one is sort of from your antenna or wherever you have the small signal coming back and forth the dc information is already incorporated in the specific values of the conductance c junction capacitance cj and the diffusion capacitance cd this is already included and the magnitude of the small signal essentially dictates how much current will pass through this if you have more small signal more current will go through this but the dc dictates the magnitude magnitude of this various various components so first thing is cj the junction capacitance associated with majority carriers right always majority carriers the junction capacitance and majority carriers can respond with dielectric relaxation time very fast right they can come in and out very fast so therefore for all majority carrier devices you must have a cj if you don't have it then you are in trouble so here i made the argument how the electrons and holes come close to the age of the depletion region in and out as you're changing your ac bias a little bit and that's how you calculate the junction capacitance how do you do it for diffusion capacitance the other one that once you inject it in the diffusion side and the minority carrier side depending on how the frequency of your wiggle of the small signal then there will be a ripple of carriers moving forward if you're doing it at a very slow speed then you will have a junction capacitance if you're doing it so fast that before the electrons can get to the other side you are asking that come back if you're sending then there'll be a phase difference of the current that is flowing through and the signal that you're applying and so there'll be a corresponding change in the capacitance that will be frequency dependent right so the diffusion capacitance correspondingly will have a value associated with this minority carrier modulation now one thing here if you don't have minority carrier you do not have you do not have minority carrier diffusion capacitance of course now do you realize why these things are in parallel why aren't the series so you should ask yourself why they are in parallel I mean looks like they should be in series you have a junction capacitance then the diffusion capacitance seems to come after the junction capacitance why aren't they in series so that is something I will ask you to think about I have answered this of course during the lectures but I will ask you to think about one more time this is many many senior people this is sort of get trips over this very simple notion large signal large signal is different from small signal large signal is different because when you change the signal suddenly then let's say from 1 to 0 then it doesn't go through slowly through this various equilibrium points to the other side but rather you you sort of give it a sudden impulse of voltages and then most of the time it will go directly it will satisfy the current immediately and satisfy the without changing the voltage the current whatever current the output circuit wants the device must give that current immediately voltage will come in little bit later why this is because of the capacitance because capacitance doesn't allow the voltage to change suddenly it allows the current to change suddenly now think about it instead we didn't have any inductance here right there is no inductance now there is something to think about why didn't we have any inductance in a diode shouldn't there be an inductance there in fact you can show that macarbon nanotube and other more modern devices does have inductance so in that case you can have inductance if you had inductance over there your current couldn't change also but for the semi classical device we are thinking about the diode right in that case only capacitance voltage doesn't change current changes immediately okay and we talked about how their corresponding responses occur and how to calculate the storage time ts and the corresponding how it goes back to the other equilibrium time constant equilibrium current let's say so you should think about it because this is sort of very important and the way we analyze this is this charge control model very quickly I'll just flip through this just to remind you where we had been we just said that solving a diffusion equation in position and in time is horrible cannot do it it's a partial differential equation we'll get rid of the spatial variation altogether and only focus on the number and therefore we integrated over the space and therefore only thing we care about is area under the curve which is let's say area under the rate curve at a particular point the total number under that without really bothering about how they are distributed over the whole region and we correspondingly got equations these are the charge control equations and you know how to solve now charge control equations hopefully not all of you could do that very easily in the second exam but in the final exam I'm sure by that time it will be crystal clear to you hopefully but the thing is that look at this equation this is almost like a if you had a rc circuit with a voltage in the beginning if if a second day student got a rc circuit with a voltage with a step voltage up and down you could solve that problem why can't you hear in the rc circuit you would have what you do have you would have a c dv dt which is like the first term and then you would have a v minus r v minus the applied voltage divided by r which is like the second term so if you can solve that equation there's no excuse why can I you cannot solve this equation exactly exactly the same equation here so just think about in terms of series connected rc circuits when you're solving this problem you'll be fine very briefly then where have you been there's this band diagrams drawing band diagrams in equilibrium that is something we have done repeatedly same rules and this will be the same rules at the end of your carrier also which is start drawing a Fermi level put the bulk regions where this supposed to be continuity of the vacuum levels and all that that's the same rule always always correct now in terms of dc current calculation if it is if you do not have any hetero barrier discontinuity most of the time this diffusion equation minority carrier diffusion equation those will be fine those will be fine but if you have a discontinuity like metal to semiconductor or between two different types of semiconductor right like in lasers and other use thermionic emission theory no rocket science here very simple and you use that many times if you have like hbt where both diffusion and thermionic emission are important in different regions you may have to put them in series and whichever is the limiting current that will govern the whole current through this structure small signal again you have to make a diffusion capacitance a junction capacitance and you calculate from the majority and the minority carriers all of them exactly the same thing but you realize in a most capacitance there were some spatial features because the oxides essentially prevented the electrons from flowing out as a result there are some what was this this is small signal i'm sorry low frequency there was a high frequency response right and then there was deep depletion right those things wouldn't have happened in a diode that happens because you have a oxide region preventing the flow of the minority carriers to the other side so you will study that a little bit deeply right and the large signal most of the time use charge control for that you will get a 0th order answer more or less immediately don't have to worry about it you always have numerical simulation but first do a few lines in the back of the envelope two pages do if i get an idea that what the solution should look like then go ahead and solve the fire it up in the nano hub or in other places and solve the problem because without the two pages you will not even know whether what's coming out of the machine is correct or you put a wrong input parameter you're supposed to put 10 you put 100 you know it's a middle at 2 a.m. in the night and you are trying to solve a problem you put 100 you get a solution you take it to your boss and the boss says that you can you are you don't have to come from tomorrow because you essentially you have no idea that how much current can flow he knows because he knows he is seeing the measurement every day so he knows how much current can flow through the structure and you the sophisticated engineer just bought email answer which is not even in the ballpark is very important to get the ballpark numbers right now very quickly i think i'll probably need five extra minutes today i'll just i'll go through about a few more extra things so i started here this is going forward looking ahead and we talked about bipolar and how the temperature increase in mosfet every every this has put each technology in jeopardy and by now you understand how that ways shaped bipolar transistor works right all of you do that this is a metal semiconductor metal type short key barrier transistors and the real bipolar didn't come around until 1952-53 manufacture of it idea was in 1948 you now know what is this transistor called the mosfet this is a fin-fed type structure electron goes from the left pillar to the right and this region which is sort of L shaped region that you see in the middle is the gate that is sort of modulating the electron flow from all three sides right so that's because of control of the short channel effect short channel effect is getting more and more difficult so you know all those given what you know the chances are high that this is not what you'll have to work with or have to finish your career with in the next next few years or next next decade two decades something like this and there are many things people are doing spin tronics biosensor displays let me give you a little bit of context so the microprocessors is sort of in that upper edge where sorry this didn't come out correctly but on the very high performance and you want to make the transistors as small as small as possible and there have been a lot of work right last 50 years that's what that's what we have been doing in electronics but these days most interesting or many interesting works are in the large area electronics where the cost has to be very small and at the same time it has to be very large you know in a rooftop you cannot put your microprocessor and say that this is my thermo photovoltaic cell you have to put the whole rooftop now if you are trying to buy the photovoltaic cell with the microprocessor price then you are not going to put too many photovoltaics up there therefore the fundamental change that is happening across the board is that the material this crystalline material the basic thing that I started this semester with that is probably not going to survive so you are we are talking about polysilicon transistors we are talking about polymers for batteries and other things we are talking about mesoporous type structures where where it can hold a large amount of charge in a very small volume and then there are various other types of materials and there are a lot of work lot of work in this field in terms of flexible electronics energy and biosensing lot of work now what I'm trying to tell you what you have learned this is sufficient to understand any of this if you just don't get afraid and spend just a minute about thinking about the fundamentals all of them are essentially very very simple things let me explain this is for example somebody says that I'm doing organic electronics now he shows you this device this is how organic electronics are made where you have a substrate do you see where the gate is gate is it's upside down gate is on the bottom gate is on the bottom then you have a gate insulator that's also in the bottom that's the green and the thin film semiconductor which is the yellow which is in the top and the source and drain are on the left and right so it's the upside down MOSFET that's what thin film transistors are why is it upside down because organics cannot stand high temperature so you cannot put organics first substrate first and then start putting the oxide if you put the oxide everything will go away so you put it upside down but why do you care metal semiconductor metal is a short key barrier transistor draw a short key barrier band diagram do a short key barrier current flow through this structure you are done so it is nothing new is just a different configuration you know if I took a MOSFET took my computer and flipped it around the physics of the MOSFET doesn't change right why you did change here is the same thing but the material is very different there are lots of random materials like pentasane and the transport physics the mobility effective density of state the effective masses those are different those are definitely different and this is something that sony has come out recently with a display essentially based on very similar structure and I will ask you to do this that even if I give you in the exam which I may or may not you shouldn't be you shouldn't be worried you should be able to draw a band diagram you should be able to say what transport theory to use it's a metal two metals and the semiconductor inside what transport theory do you use thermionic emission right essentially in the middle can we use the thermionic emission on the other side and if you would you be able to use a numerical simulator in nano hub of course you should be able to use the same MOSFET just set it up with the proper value of mobility effective density of state and all you should be fine so this is something that is not something that should you throw you off of course the whole research is getting the right material that has a good mobility another thing is solar cells again no problem because the solar cells is a p-n junction diode if you have to take a p-n junction diode and then shine light on it then you have what do you have shock lead all generation right you have a generation and if you have generation the electric field essentially pulls the electron and hole out and they pull the electron and hole out in through the respective contacts you get current if you know how to solve the reverse bias p-n junction diode then you shouldn't be worried about solar cells at all solar cells then essentially this is the electron hole pair coming out and then they fall through this is a band diagram they fall through this two materials and the electrons are separated out so this is this is not a big problem also in terms of what you know you should be able to able to handle this now many times what happens is electron and hole lot of them recombine before they get out so you cannot collect every photon that is coming in and so many times what people do they use the mid structure the two things sort of interpreted it steal a p-n junction steal a p-n junction but not the nice flat p-n junction that I drew in the class but you shouldn't care about that either I mean many times they do p-n junctions like this because what it allows you to do it allows you to reach the junction very quickly and then separate out the electron and hole this is a simulation of the materials that but you say again a p-n junction it's just that this is the n this is the p and it looks different looks horrible but it is a p-n junction how difficult can it be and then you correspondingly solve your minority carrier transport equation across a junction and that's about it that that's essentially about it you put the the two mirrors and the two contacts in you have the solar cell so what I'm trying to tell you that these things are very very simple every what you have learned so far if you just don't get afraid and sit down say okay this is my device and this is two pieces of white paper what did I learn in six or six let me apply it you will be ahead of most people in the field one final thing is about biosensing again a pure and in fact I gave this problem in one of the final exams some time ago and in which biosensing is here here you have a x y grid of materials and x y grid of sensors so every square little square you see is the sensor you have many of them you can have let's say 10 000 of them right now each pixel in the sensor is decorated with a DNA molecule DNA has a certain amount of charge now let's say you have an unidentified solution molecule in the solution so you you have decorated all the with a known strands every pixel you have decorated with a known strand then you throw in the unknown if you throw in the unknown then the unknown is essentially going to bind wherever it finds its conjugate because DNA binding is very specific that it only binds with if it has a conjugate and in that case so I'm not going into the details of it but the main point I wanted to make is that once you know let's say this three red has lighted up therefore the three reds have found its conjugate but you already knew what the three points are this pre-pixels what the original one was like or because you put it there and so thereby you can find out what the unknown is identify what the unknown is right somebody comes to your office and says that I have this new device I want to analyze it what are you going to do you said no problem again because this is a MOSFET current going from the green source to the green drain and you see this is a thin film transistor it's upside down MOSFET right gate is on the bottom in the deep blue and you have the oxide and you have the current flow and when this charge this molecule has a certain amount of charge when the charge sits on the gate so instead of applying a gate bias you are just putting a charge on if you put a charge on can you not calculate surface band bending of course you can calculate right if you have a charge the surface must be compensated you will provide you compensating charge and therefore you can see the band bending the higher the band bending the it says that more charges have come at that point that means conjugation has occurred this is as simple as this there is nothing more and if you know the MOSFET physics this is essentially I say two hours of work to find out how current should flow in such a structure and this is something you can easily calculate that once the molecule sits in here the change in the conductors due to surface band bending you can easily you should be able to calculate now let me finally conclude on this about this vision and many times people actually work on for historically try to work on this robot that more and more sophisticated electronics and that is what electronics people thought for last 50 years that that is why the electronics is going to go there's a rapid shift and the rapid shift is because that type of robot base future base requires too many resources too many precious molecules too much heat too much energy and it requires too much energy the function is limited it can do a few things it can bring you a cup of coffee in the morning but not really too much more and it's not very reusable but on the other hand many simple devices these are processed at very low temperature right you hatch a egg at almost close to room temperature no 1100 degrees degree required and still the functions it can perform the near equilibrium the structures like this can catch this is a yellow fish it can catch its molecules so you can again do the diffusion equation to see how the molecules essentially catches food and then how it redistributes for diffusion equation carrier transport essentially you can analyze the whole thing and it does amazing amount of things so what i'm trying to tell you is a vision for the future is rapidly changing towards this more energy efficient amorphous diffusion dominated close to equilibrium things because that has to be the future of electronics and so this is this is how i think that over your carrier how this is going to change but many of the things that you have you have learned in this course will remain essentially completely unchanged so that's that's something you should think about so let me summarize so i said electronics will continue to be remain vibrant that's hopefully that will hopefully good because then that gives you a job but i have a feeling that this computing and communication that will get more specialized and out of the hand of device physicists in a circuits communication there'll be a lot of beautiful work but for device physicists we have changed many many times it was started with electrical machines branched into communication and computing and it's going to change in energy convert it energy conversion and health care and this type of other applications because electronics always solved a problem of the society it has to right devices have to do something that solves a big problem and the big problems is no longer being able to talk to your friends two hours more for no price i mean these days is so inexpensive communication that there are other areas where electronics must must branch they give you a free cell phone it is so inexpensive these days therefore in order to make money you have to go somewhere else okay let me end with a few acknowledgment one is this Muhammad Aziz Zaman he helped me with many figures drawing many figures you may have seen from time to time notes that asad we need to need to redraw this figure which is that he helped me with many figures your cycle sitting in the back this is 40 lectures in the 7 30 in the morning it's not easy so thank you these are resources and for various supports for the this videotaping as well as the for the teaching assistants and others i often discuss things with professor mark lancstrom for hbt i took many slides from him and professor priyadatta we work closely and previous e 606 classes who if you can believe it i just used to write it on the board and whatever they could take down that is what it is so at least now you can go back and listen to the lecture one more time so that i hopefully is much easier and thank you all for coming to the 7 30 morning morning class okay thank you very much