 So what I would like to tell you today about is the work we've been doing over the past few years on transition metal decalcogenides It's it falls in the in the in the class of 2d Materials so so materials like graphing can be exfoliate to become perfect to decrease those The people here are people in my group. We've been doing the experiments. I'll show you Different things. I think the scope of thing that you can do is is quite vast You need materials for this work Elmwood, Belga and Ike Janini are the people who provide these materials and we've had Different collaborations depending on the specific topic that we've been working on and so These are the kind of different thing I like to tell you about I would first like to tell you about some specific property of these materials in term of just Bear semiconductors. They are Interesting because they have a good number of transport and that allows to look at many different effects both interesting from a fundamental point of view And for device applications. I will also touch upon a valuable effect, which is very closely related to Leonid's first part of the talk. I'm sure you can induce gate a superconductivity electrostatically starting from an insulator just by applying a gate voltage And if I have time I Talk about how you can use these materials to induce phenol but in graphene to announce very strongly spin orbiting graphene due to inter-official interaction and then we'll have lunch and If there is not enough time, I'm afraid we'll have to skip the last point of the Of the talk so so let me give you a little introduction about these materials. This is a Typical say crystal structure. There are different special structure, but this is the most common crystal structure for many of these transgeometallic alcogenites and This these material have been studied in the bulk for for a long time And because because there are many different phenomena that have been observed like for instance superconductivity charge density waves They can just be bare metals. They are semi conductors semi metal and and even in the bulk It's totally not exhausted the kind of scope of phenomena that that that can be seen depending on the composition because because the composition Because there are really a huge class of materials about maybe about 70 different But they have been re-iscovered recently because you can exfoliate them and get them to the monolayer level and and and not all the materials are easily stable chemical at the monolayer So so initial work has focused on these kind of compounds which are semi conducting and they are very stable in air And one thing that has been noticed in the early days, which is a few years ago is that if you take a bulk I'll say a piece of bulk of these semi conductors is an indirect gaps material but if you exfoliate down to the monolayer the gap become direct and you observe that and people have observed that by looking at the Quantum yield of photoluminescence which which explode exponentially as you as you lower the thickness So this is for the photoluminescence on a single layer and in absolute scale This is the the bilayer So if you zoom in you still see for the monolinescence for photoluminescence But on a linear scale is is almost not detectable and this is confirmed by by Benz structural calculation We show that the bulk is an indirect gap and as you lower the thickness and you go to the monolayer you end up You end up with a direct gap and so So this is this is this will these are the experiments which trigger the interest in these materials and the reason why They are particularly interesting especially coming from many years of work on graphene is because a monolayer is an hexagon and so as a Say values at a K and K point and and Contrary to graphene this valley because there is a there is no inversion asymmetry this valley are kept Also, there is a huge spin orbit interaction So this is shown here that the red and the blue are different spin direction We enter perpendicular to the plane and as you go from from the K point to the minus K point because of crumbest degeneracy The spin direction is inverted So you basically have only close to the to the deal to do this K point You have that an effect environmental that describe the system correctly is basically a Gapped Dirac fermions we've a strong spin orbit interaction meaning that the gap depends on the spin direction is just like as Leonidas Explained then you can have among other things a lot of phenomena Which are interesting which come from the fact that the electrons here have a final buried curvature Which means that basically you have They there is a whole conductivity associated to a valley and And and say the annoying thing if you wish is the opposite valley and I'm knowing from experimentalist at least is that the opposite valley as an Opposite whole conductivity So if you just measure the whole effect at zero field you see nothing because you see compensating effect So and so it would be interesting to see what you can do and and to Reveal this effect and one thing you can do is and I will come back more in detail is you can explore exploit Selection rules in optical transition which only which you can Use to only couple to one of the values and therefore to to generate a valley imbalance and and and and break this compensation between the two values So one technique that we will use quite a bit is is in part of the talk is I only liquid gating and which is a which is a way to Accumulate a very large sharp density on top of different materials by gating and it's essentially instead of using a solid dielectric You use a liquid made of charged particles These are pretty large molecules say one minute one micron one million one nanometer in size Which electrostatic is basically like a metal meaning that there's a very short screening length of say about one nanometer And so when you apply a gate voltage in the liquid The voltage drop at the interface between the metal and the liquid there is no actual voltage drop across the liquid And so it will all drop across The the the voltage will all drop across the gate and the liquid and then between the liquid and your semiconductor If you do things properly, it will basically all drop across The liquid and a semiconductor which is a way to have a huge capacitance because then you have a capacitor With a with a dielectric stress of thickness of one nanometer and you can actually charge density up to few times and then to the 10 to the 14 plus square centimeter that you could never do with conventional Dielectrics so so so this is this is just a technical point, but it's not an important one So let's first look at the basic semiconductor property of this system and we'll focus on tanks and disulfide But if you use a different one of these semiconductors who basically be the same So the first thing you see is this this possibility to have a very large charge density accumulation You we sweep the gate voltage and so this is an example of a device You put your your your ion liquid on top you stick a wire in it and then you sweep the gate voltage And you see that you can with a fairly moderate gate voltage or a few volts you go to You can accumulate a large amount of electrons and up here you are close to 10 to the 14 or of holes So on the same system you can just by switching the gate voltage in the lab You can go from large whole density to large electron density in transistors if you look at the What is called the subtraction swing which is this quantity Normally there is an intrinsic limit essentially this is coming from how steep this can be it's coming from the fact that the finite temperature the the Fermi Dirac distribution as a as a as a finite width and in the best case if the gate capacity is huge, which is what what would be the case here you find that Say the steepest value you can have for this for this for this look is given by this number And if you actually use this ionic with gate transistors on these materials You get exactly the number that means that this capacity is basically really really large It's it's not a conventional transistor It's a trans or is a conventional transistor in which the gate capacity is you can take it in a first approximation to be infinite if you want So let's see One consequence of this is that you can already use these devices because you see Holes and electrons to to measure the gap and the idea is very simple to the band gap the idea is very simple you can sweep the gate voltage and and when you sweep the gate voltage these cause typically a shift in Fermi energy and a shift in electrostatic potential and and and The electrostatic potential is you can in a parallel paid capacitor. It's just the the charge by the By the density divided by the by the capacitance But if you're in the gap that as your state is very very small So you accumulate very small amount of charge and the capacity is very big So this term you can forget and so the if you if you compare the Gate valve the gate voltage value when you are starting to so the threshold for whole conduction a threshold for electron conduction This is just exactly equal to the gap and in fact if you do that you find it for bulk Thanks, and I saw that you find that the gap that these difference in threshold voltage is 1.4 1.4 Which is exactly what the gap is known to be in tanks and I saw fight throughout that this technique Works for even for device which are not as As ideal as this one just because the capacity is always very large of the unit liquid gate And it's actually very convenient because now you can use it for instance to measure gaps in semiconductors Which are very very small flakes which should not be you easy to measure with other experimental techniques So since you you're able to to to see and the polar transport another thing you can you can do is try to to see if you can Basically inject electrons all at the same time and and get light out and and so these are the typical curve of a transistor You see the typical saturation on when you accumulate electrons Well, I think actually holes here when you accumulate electrons They all look normal except that when you go to large drain voltages You see these upturn that you normally don't see in transistors And if you look at like you know typical book like the same book on on semiconductors You typically do not see this curve So this curve only happened as a large source of voltage because that means that locally what is happening here is that locally on one of the Contact you are inverting the effective Potential of the of the channel and that means that that you're injecting on one side one polarity of One say charge carrier one polarity in the opposite side charge charge carrying of the opposite polarity So the way you can think about it. You can think that you set the gate to zero and you go very large With the potential on opposite the contact on opposite polarities and that that means that locally the potential of the channel is different And you have different kind of charge carriers. So if you if you do that you send current You have a leckos and and holes coming together and then they recombine and they should emit light Well, they should emit light if the semiconductor is a direct get semiconductor. Otherwise, there will be Things will be dominated by by no radiative or combination So in the bulk the gap is indirect and indeed here we don't see light But if you go to the monolayer level then the gap is direct and we should see light and again We can do all the all the same business here. We can measure our nice and be polar transport We can extract the gap find it is much larger as you expect for the monolayer is instead of 1.4 is 2.2 V And and again, you see the zombie polar injection regime or a high source to in bias The the current shoots up and now the question is whether when when we bias the device in this regime Whether we see light or not and the answer is yes We do so this is you can see it with your bear eyes under a microscope basically and this is the right spot here It's just like coming out of the device you can compare the spectrum so the Frequency dependence of the light that is emitted with the one with the light that is emitted not by electrical injection But by illuminating the device with a high energy light and see at the light which comes out Which is photo luminescence and they basically perfectly match So this is a this is a and and what this is is a recombination of excitons And in fact since we are able to to measure the band gap and we are able to measure the Exit on energy we can get the axiom binding energy just by directly from the measurements Which in this in for these devices is about 200 millivolt Theoretically is predicted to be even I mean there is a spread of experimental value Theoretically is expected to be even larger than that 200 millivolt is already pretty large So this is to show that this is This system in term of off electronics have quite a bit of potential They're allowed to do things which are not easy to do with other semiconductors because they are un-be-polar and and and there is there are other things which are interesting which you can do when you Combine optics and transport and one is the value of effect and since Leonid has discussed this already I will go fast but but the idea can basically from graphene and in fact from gap graphene And just like Leonid has said there is a there is a finite very curvature And that means that each valley is a conduct a whole conductivity which is equal and opposite So under equilibrium you see zero voltage But if you would be able to To to unbalance the population of the two valleys the contribution to the whole effect of the of the total contribution to the All effect would not would not cancel and you would you see a finite whole voltage at zero At zero magnetic field if you are in this situation So the question is how can you do that and the same people who did this work They also predicted that the one way to achieve this situation is to pump the system with a circular polarized light because there are selection rules that tells that that if you if you ever say one say One circular polarization you can you can only couples to one valley and the opposite circular polarization couples to the Opposite valleys therefore if you shine a shine light with this light you will according to this theory Excite electrons on the balance to the conduction band and and of course this electron can relax But under stationary condition you will have a different population of the of the two valleys So they predicted by by by measuring the whole effect a zero magnetic field Under circular polarized light you would see a whole voltage which is a manifestation of the of the of the very curvature in the in the two valleys And in graphene is not okay as I've already shown is it's it's possible to see effect related is with different techniques But not so easily with light because the gap is small and and so in the inside in these materials the gap is is To EV and so you can do it with visible light Which has a short way which are short wavelength so you can make a small device and in fact last year This has been this phenomenon has been observed by Mack in McEwing's group So what what you see here is a piece of molybdenum disulfide another one of these semiconductors again They're conceptually very similar in which they send current and under circular polarized light They measure the whole voltage and they see that indeed there is a whole voltage when light is circular polarized And if it if they if they change the polarization the whole voltage change sign and if they use linearly polarized light there is no whole voltage and And they interpret this this phenomenon in this term in terms of single particle picture but actually there is a little problem with that because They shine light at 1.9 EV and one not 1.9 EV in molybdenum disulfide is exactly the exit on energy and the exit on binding energy is large So we certainly don't have Interband transition of that energy and so the question is how is it possible to get a Whole voltage from exciting exit on which are neutral particles and they have no charge So even if you'd accumulate the exit on one side of your device It would not be possible. I mean they being neutral. They will not generate a voltage So we have been looking at this problem in tanks and disulfide and made a very similar device at first in which Again, we see the same phenomenon if we if we look at the if we have if we eliminate the device with Circular polarized light we see a whole voltage We change sign depending on the polarization and these appears when we have a linearly polarized voltage and The whole voltage is a black line here So the whole resistance is speaking exactly at the same position as the photoluminescence, which is the exit on energy There is a second a second exit on here that you don't see in photoluminescence But let's focus like this and these energies again quite a bit smaller than the than the back band gap in tanks and disulfide So so how does that work? To understand what is happening I mean we went we went back to look at other phenomena which which have say a similar problem And one of them is is photo current in photo current as well You can illuminate your device and you see that an announcement of the photo car and you see appear in photo current at the exit on Energy which again is strange because you generate an external which are neutral And so why should they give any photo current at all? And and and turns out that in this case though the answer is known because people have studied photo current much more And essentially if you have a device like this whether you see this photo current or not depends on the position of your laser beam So if you put your laser beam in the center, you don't see anything Because what you do you generate exit on that they diffuse and they recombine Before anything can happen, but if you put if you put your laser being close to one of the contacts What happens is that you generate an exit on these diffuses and if it diffuses towards the contact It will be split in the electric field associated to the shawty barrier The minority carrier will escape into the contact and the majority current will stay into the semiconductor And we create a valine balance population of charge carriers and so and so in fact if you now put your laser beam close to the other contact You see the same effect but with the opposite polarity because the same charge carriers escape in the other contact So opposite current So this is telling us that if we are able to map But so in photo current this phenomena is very clearly visible by doing a spatial mapping of the signal And so we do the same thing for the valley old effect and and So basically we fix our light at the frequency of the exit on so bill well below the interband transitions We apply a current and measure the value all effect and fix the bias here We see a value whole effect and then use a device like this and scan the laser beam across and see how this voltage That we measure depends on the position of the laser and the result is shown here Well, we see that the we actually see valuable effects only when the laser beam is quite close to the to the contact So what that what is happening is again you generate electros? Exitance this exit on diffuse if your laser beam easier they will recombine and nothing happens But if it is closed if they are close to the source contact where you also put your your your whole probes They they will reach this the the contacts They were splitting the the action was split into the electric field and the majority carriers will go back and give you a Valipolarized distribution which is why you see our signal and that's the signature of this phenomenon is the fact that the Magnet of the signal it becomes larger the closer is your laser beam is to the contact You could think okay. Why doesn't it happen that our contact? Well, because here you also split your excitons, but The the particle will lose its this distance too far is beyond the interval is scattering length And so you don't have valipolarization reaching the region in between the contacts so so this part this basically shows that By mapping the valial effect at different frequency you can you can this you can tell something about the origin of the process Which give which give rise to this whole voltage? Let me move on to Another part of the of the talk which is how we can still use exactly the same materials Which are semiconductors now to turn them into a semiconductor into superconductors by accumulating enough charge And a breakthrough work here is done by I was done in the group of Yoshi Vasa a couple of years ago Who took basically thick? Layers of molecular disulfide basically bulk and by using a gate voltage We usually was able to turn them from an insulator, so I'm dope semiconductor Or maybe slightly dope semiconductor due to impurities and accumulating charge at the surface Then he was able to see a to see a Transition to a superconducting state with a maximum temperature of 12 Kelvin and he Discussed these in the context of superconductivity Seen by intercalation which in the same material was was also seen earlier Which I think is a bit more Questionable, but but the observation of superconductivity with gate voltage starting from an insulator is really an impressive result So we're interested in this and see also how this varies when you go to the atomic When you bring the material to atomic thicknesses and we've been doing experiments to do this So what you see is a device again on which we put ionic liquid This is a six layer molybdenum disulfide and you see that we've applied gate voltage We are able to see a superconducting transition. However, what we find in basically all our devices There tend to be quite a bit in homogeneities because more for technical reason related to the to the property of the ionic liquid Still the maximum to see that we see is always is pretty comparable to what Yoshi Vasa so in his state crystal So six layers is basically is basically Bulk and if you look at the critical field actually, this is impressively large is then close to 10 Tesla already and and I think Yoshi has recently done for for tick sample as measure the critical field in the plane and it's immense can Reach up to 80 Tesla or something. So So we were interested in going to to smaller thickness and so in particular to the monolayer if you go to the monolayer Well, again, you see superconductivity, but now this is quite a bit lower and also the critical field is much smaller And so basically in the monolayer you also see superconductivity But it is much less robust. You don't see it in all the samples and and when you don't see it You can see phenomena like we can't localization which which which come from the fact that you have a strong spin order interaction So what about ticker layers like by layer trial a well as soon as you go to the by layer It resemble much more the bulk in the sense that we always see superconductivity The transition is more robust and TC is already quite a bit larger So whereas you go from 1.5 Kelvin or close to two to to basically Almost seven in the by layer and if you and larger critical field from point one to two three Tesla And if you go three layer four layers you you basically get superconductivity or vastly with with TC Which seems to to depends on thickness and become slowly higher and reaching the bulk value So it's interesting to compare all these thicknesses and see what happens So this is this is basically putting together the measurements of the of the critical temperature versus Versus thickness and critical magnetic field and you see that basically there is And again because of the enormity you have to do quite some statistics Which we have done a bit of and what it looks like is that you have a trend with TC and the magnetic field now This is log scale. So the fact is less pronounced but decreasing as you decrease the number of layers Field is movely and then going from by layer to the mono layer. There is a much steeper drop And these okay We are not since we have charged density in a virginity is difficult to tell what is a density dependence But but this doesn't seem to depend on density pen as a density pen is kind of a measure from whole effect is a bit All over the place and TC does not depend too much on that As long as it's high enough if you don't put an high enough density, you don't see superconductivity So why is that why could the TC be be much smaller in monolayers compared to ticker layers? Well, there are different reasons one of course there will be There could be I mean if you have if you have a Decumulation length or the accumulation thickness of the electron layer that you do that you that you put with the gate As a certain thickness which they mean by screening and and turns out that this ticket is about 1.5 Nanometer and the mono layer is the first layer in which this that has a thickness Which is really thinner than this 1.5 nanometer to nanometer, which is the by layer already as as a thickness comparable to the To the screening length which means that you'd be able so so it could be the superconductivity is suppressed If you if you try if you go below this screening length But what we think the most likely explanation is is that if you look at the band structure in the mono layer You accumulate electrons at the k point whereas in bilayers and ticker layers according to calculation You should be accumulating electrons Are a different part of the brilliance in a different part of the brilliant zone because it's an indirect gap and Essentially your accumulated electrons here, which is which is called the Q point So then it's not so strange that that superconductivity is different because the denser state will be different And the electron phonon coupling will be different In fact is if you want is strange that you see superconductivity for all thicknesses because in principle these materials are all different So they all have different electron phonon coupling and different denser states So there is no guarantee that if you see superconductivity for the bulk you should see superconductivity superconductivity for the bilayer or the mono layer a priori And How much time? All right, so let me go very fast to the last topic Which is how we use these materials to induce spin off the interaction in graphene This is a work we did also in collaboration with the group of Allah McDonald for theoretical calculation And there is there are clear reason why you'd be interested to try to induce spin off the interaction in graphene because we know from the work of K and a millie that if you take pristine graphene and if you if you consider the effect of spin off the interaction in principle conceptually This material is a topological insulator a gap opens But of course in graphene is by now known that this pin of interaction is way too small to see this topological insulating state experimentally Nevertheless, it'd be interesting to find a way to Amplify spin off the interaction much more to try to to be able to see these things And there have been attempts attempts typically have put graphene on metals because there are metals I don't like platinum or with very large atomic numbers And they with these in this case you can induce very strong spin off the interaction graphene But the electronic state of graphene hybridized so much with the metal that that you lose completely a direct fermion character of the of the like once in those cases You can also induce another interaction by creating impurities of certain types And again that works up to up to some extent, but but again you damage graphene and you you decrease the quality So what we want to do is to try to induce pin of interaction but preserve the quality of graphene And what we do we put graphene which in this scheme is the black layer on top of a tungsten disulfide Tin crystals and then on top of basically a gate So this is a conducting layer with a with an insulator And let's not go to the details But but basically on these kind of devices we still see perfect quantum all affect exactly as you expect for the art formulas in graphene And and this is a first generation of device with mobility, which are okay like 13,000 But by now we are able to get basically one order back to bigger So we have these kind of devices with mobility which are hundred thousand or bigger And then what do you see in transport? So we use weak anti-localization to detect spin out the interaction And so if you measure the conduct as a function of magnetic field and gate voltage Well in these devices, which are actually pretty small You see is difficult to do you might see a hint of something here at zero magnetic field But but UCF universal conductor for operation are too big and you don't really see a clear weak anti-localization What you can do is an ensemble averaging by changing your gate voltage By as little as possible But enough to to sort of change a microscopic character of the device of the sample and then average of the of the traces And so here you see one trace which is one cut here Then if you average nine of these traces you get the red one and if you ever 25 traces You start to see that something emerged at small magnetic field and at the same time the The the the UC at the amputee the root miss square of the UCF decrease like the square root of the number of ever So so this is what you're doing. These these are basically reproducible things And so if we have done this for different gate voltage ranges, and you see a very nice weak anti-localization signal which increases you lower the temperature and that the lowest temperature is compiled is Is a large fraction of the square of the age? We can feed this data with a theory of Edward McCann and our chairman and extract basically the spin orbit relaxation time the spin relaxation time if we do that we find that the spin relaxation time which are these black dots are Say hundred two on say hundred or more times Small shorter than then than the spin relaxation time for graphene or silicon oxide and a boronite But meaning that spin orbit interaction is much much stronger and in fact is comparable to the Intervalis scattering time on that the people have measured on a bunch of different subsets These these are data point from different groups on different materials. So it's quite remarkable And so this this measurement by themselves show that spin orbit has been has been Increased a lot just by putting graphene on on tanks of a soul fight And the fact that the interval scattering time might be playing a role in relaxing spin could also explain why neither the Diakonov perl know the leotiaffet mechanism have been have been found to really account for spin relaxation in graphene But these we are not able to tell so We can be localization does not tell us what is the kind of the functional dependence of the spin orbit that we're inducing and And and a la McDonald and his post-ocular watch and have done ab initio calculation to try to figure out what according to you We should expect and what we should expect is is is is well described by a close to the direct point is well described by a Continuum Hamiltonian given by this expression so the the dots here are result from ab initio calculation and the and the red lines have fits to the dispersion relation obtained by this Hamiltonian with these parameters and they find that that this coupling depends on the distance between graphene and and thanks at a soul fight and Basically for reasonable value experimental values or by using codes that allow to relax this distance They find that these numbers are a few millivolts So compared to what you have in pristine graphene, which is a 20 to 50 micro volts We are talking about the 200 micro the increase in spin orbit interaction now These terms are not the same don't have the same functional dependence that you would have in pristine graphene Because they are violating inversion symmetry. So in pristine graphene that would not be possible but when you have graphene on a subset that is possible and Depending on the on the relative value So if you would have these land eyes speaker than land I are then you would open a gap And these would be a topological insulator although not of the time predicted by cana mille So I guess the next step is to decrease this order and increase this This coefficient even more experimentally to try to see whether we can get ourselves into the regime or because now we have a few millivolts Spin orbit interaction a few mille volt a few mille volt gap, which is small But it's not unthinkable small in terms of being able to put your Fermi level in there and try to and try to see Edge states if this is topologically really So with this I would like to conclude and So here's my summary that topic I've been telling you about so thank you for your attention