 The last talk in this session will be by Deborah Pretzi, and the title of her talk is Light Emission from Ultranarrography Narrow Ribbons, Edge and Terminia Effects. Good afternoon, everyone. Thanks for coming. And today, I will present our work on the first principle simulation of the optical properties of finite graphene nanoribbon, which was done in collaboration with an experimental group from CNRS at Strasburg within the MAX project. So as many of you know, actually, nanostructuring is one of the best strategies to obtain, to open a gap into graphene. And make it suitable for standard optoelectronic application. And this can be done either by cutting graphene into one-dimensional stripes, or by confining electrons into zero-dimensional quantum dots or molecules, if you prefer. Both as theoretical prediction have told us, actually, the quantum confinement is not the only knob that we can tune to modify the properties, the gap, and the optical properties of these type of materials. In particular, besides changing the width of the ribbon, you can also play with other properties. In particular, you can change the cutting directions, so either if you have, or the so-called chirality. So if you have zigzag or armchair ribbons, you can have very different properties. But also changing the very local conformation of the edges or adding some functional group to the edge might lead to very different properties. Here is a very simple example from a work of the group of Francesco Mauri. And they show that if you just passivate your system in a different way, you can get very different band structures. In order to exploit all this richness in terms of gap-tuneability, which is probably the most important peculiarity of this type of system, it is really crucial to get to the nanometer scale in terms of lateral confinement and, above all, to reach the full edge control. So during the last decade, we have had an intense effort to achieve both of these properties. And the most successful strategy today is based on bottom-up approaches, where the focus is basically on the design of proper molecular precursors that allow to build ribbons which have a specific width and a specific edge morphology. And actually chemists worked a lot on these strategies. And we now have a very rich library of ribbons with different width, different properties. So starting from that, in the direction of the electronic devices, they started to characterize the optical properties of these systems in different manners. First of all, the properties were probed both on substrate and in solution. There has been one of the main achievement was to achieve the role of quasi-particle and multi-particle excitation in this type of system due to the quantum, which are expected to be large in view of quantum confinement. And this was combined in both initial theoretical modeling and experiments. All of this was done basically regarding optical absorption. But if you want to go into the direction of devices, one of the most important features is also emission. So what about emission? Actually, there are a few reports about emission of these systems with very erratic features. So for instance, these seven-arm-chair graphene and ribbons, which was the first one synthesized in 2010, showed very weak, very poor emission properties. So the emission spectrum, the PL spectrum, which is this one, is weak and featureless. And you can only retrieve some luminescence if you either expose your system to blue laser for a certain time. And so you can get some sort of emission. This is the PL spectrum. Or if you hydrogenate your system. So all the emission that is observed is related to the effect of the SB2 network. So there was another report which we actually knew about through a private communication, which was PL induced from STM. Actually, the same ribbon was grown on gold. And then it was lifted with the STM tip in order to decouple it from the substrate. And they actually claims some featureless, very weak spectrums. And sometimes very broad and very bright. I mean, very bright and very strong emission. So again, also in this case, the emission feature were rather erotic. But actually, this second experiment has some very relevant properties, some very relevant feature that I want to underline. First of all, this STM induced light emission. This type of experiments allows to measure single ribbons. And the second thing is that these are the plus of this type of experiment is that you can exploit the capability of STM to manipulate the structure and to image before and after the experiment. So actually, I think that this type of experiments in this field was very valuable. And here is where the collaboration came because they wanted to better understand the origin of this light emission. So let me say some other words about the experiment. What they understood, making the experiment cleaner and cleaner, was that all this story had to do with the termini of the ribbon. So if they make the ribbon, they synthesize the ribbon and they gently lift it up. The ribbon has this type of termini with three hydrogen terminating, the zigzag termini. And what they see is this spectrum here in the bottom, which is broad, featureless, and it very much resembles the spectrum that you have when you do not have any ribbon in the junction. So it can be related to a sort of plasmonic emission. But if you remove on purpose with the tip the central hydrogen atom, so you have a C terminated benzene ring here, and then you lift it with your tip, what you see is this spectrum in the top, where you have a bright feature, which appears at about 1.6 EV plus a couple of redshifted vibronic replicas. And if you extrapolate the energy position of this peak to zero bias, what you get is a peak at about 1.15 EV. So this is much lower than the optical gap of the ribbon, either measured by RDS or resonant Raman, or what you can get from theoretical simulation. So this clearly points to the fact that the photo luminescence emission is activated by some coupling from the tip and the ribbon terminus. When you want to have some insight from ab initio simulation then you have to capture the key feature of your system, because of course you cannot describe your 20 nanometer long ribbon with the micrometer size tip. But there are a number of features that has to be there in your simulation. First of all, ribbon has to be finite. Secondly, you have to consider the contact with the tip. And third of all, you have to include, at some point, many body effects, because these were demonstrated to be crucial in order to describe the optical properties. So the first part was done within a density functionality or the framework by using the quantum espresso package. And the second part was included within the GW beta-cell-peter approach by using the YAMBO code. So what we know that is special from finite ribbon, we learn actually from the EMPA group that if you consider your finite ribbon and you probe the termini of your ribbon, what you see is that you have a large gap, which is the one related to the bulk states, but you also see states in this gap, which are related to some like states localized at the zigzag termini. And if you do the calculation for that, you exactly obtain the same feature and these states can be understood considering fully spin-polarized calculation with antiferromagnetic order. So what do we have to do for that? First of all, what we want to do is to focus on one of the two termini. We want to totally avoid any termini interaction. So what we did was to make the finite ribbon asymmetric and this allows us to remove two of the states. So we are only left with two of them. And as you can see the P-DOS here, if you either include or not the spin-polarization, the only region which is changing is this one in the middle. So bulk states are not at all spin-polarized. The only states which are spin-polarized are the one in the gap. And if you do not consider spin-polarization, basically they are degenerate and lie at the Fermi level. So the next thing to do is to consider the effect of the gold cluster or the gold contact, say, with the tip. And to do this we considered the tetragonal gold cluster which could mimic the tip. And here is the P-DOS where the gray line is the contribution on carbon while the orange line is the contribution on gold. And what you can see from there is basically that the presence of gold makes the system paramagnetic. So if you compare the gray line here with the blue line there, what you see is that the system is paramagnetic and these two terminal states, so these two term states lie at the Fermi level. Actually, in addition to learning, which is the effect of the contact with the tip, we also learned something else that basically in order to study the optical properties what we could do is to avoid including the tip, which is a mess, and to consider the system isolated but starting from the spin-umpolarized ground state. So we did so. We started basically from this solution, the DFT solution for the isolated system and we included many body effects in order to describe the optical properties of this system. These are actually the results. This is the energy level and these are the conchum orbitals for the levels closest to the Fermi level. And here is the spectrum. What we observed is that in addition to states, that are related to bulk transition, so transition from bulk states in the balance to the conduction, this one, which are transition of this type that I indicated with EAA, that are also additional excitation, which actually involve transition between bulk states and the localized ones. So you involve transition between this type of states and states localized at the terminal. And these are these excitation A and B that I highlighted there. So, second thing to notice is that, of course, we need to go to finite length system to see this type of effects, but we cannot reach the limit where we would retrieve for the bulk excitation the behavior of the infinite system. So we consider the number of them of increasing length and here are the results for a number of calculations, the DFT gap, the GW gap and the main transition related to bulk states. As you see, we are very far, even in the largest, we have a peak at 2.6, which is very far from the 2EV peak that we find in absorption. But we can extrapolate, of course, and what we get is exactly that. What we get for the bulk excitation is exactly a 1.9EV limit, which is in very good agreement with the calculation for infinite systems. And if you do the same for the other type of excitation and in particular for the lowest energy excitation, you get something which is of the order of 1.1. So in very good agreement with the experimental peaks observed in STM light induced emission experiments. So it seems that we have understood the origin of the emission. So it seems that from our calculation that the emission comes from excitation which involve bulk and terminilocalized states. The third observation I want to underline is that below the bulk excitation, there are a huge number of states with very low oscillator strength. And all of them actually would have lower and lower oscillator strength as soon as you increase the length of your ribbon. So these are expected to, the oscillator strength of these states is expected to decrease because the overlap is becoming smaller and smaller as soon as you make the ribbon longer. So we point to this as one of the reason why in ensemble measurements on insulator, you do see very weak and featureless emission. So basically the reason why the emission is very poor is related to the fact that this system is intrinsically finite even if you make it longer and longer. So there are states which are there and basically kill your emission. So here I come to my conclusion. We have characterized the electronic and optical properties of graphene and ribbons and we have seen that they are governed by electron-electron interaction so we need to resort to a scheme beyond DFT in order to give an accurate description of electronic and optical properties and above all to compare with spectroscopies. In addition to this, other effects, I mean considering effects which go beyond the simple approximation of infinite system appears to be very important. So finite size, coupling with the tip, all of this becomes very important. This actually appeared recently in the nanoletters paper that I quote here, but more than that I want to thank all the people that work to this project. My co-workers at CNR Nano and Modena, Claudia Cardoso, who is now at the International Nanotechnology Lab in Braga and Andrea Ferretti and the experimental collaborators from CNRS in Strasbourg which are Michael Chong who did the measurements and Guillaume Schoen and you for your attention. Thanks.