 Un'opportunità per presentare il lavoro che dobbiamo fare nella nostra laboratoria. Ho lavorato in un piccolo gruppo computativo che è legato da Daniele Passerone e questo gruppo computativo è embeddato in una realtà esperimentale che è legato da Daniele Passerone. Uno dei cori attiviti di questa laboratoria dove le simulazioni rilettano in parallel per l'esperimento è la fabbricazione di materiali nanomateriali di base di carbon iniziano dai moleculi all'atomica precisione, i nanoribboni in questo caso, i graffini nanoribboni, all'opzione di dimostrare i devices. In il mio talk I will give you a long introduction that also summarizes some of the milestone we achieved in the past in these fields and then I will move to more recent results on how we can tune the band gap of these graphine nanoribbon. Everything starts from graphine, there are charred carrier mobilities in a graphine, makes it appealing for applications in information technologies, if you want to do, for example, feed effect transistor, but there is a problem, graphine does not have a gap. Now the question is can we open a gap in graphine and then use it for transistors? Of course yes, the answer is yes. There are several ways to open a gap in graphine. One of these ways is quantum confinement. If we cut graphine in stripes, one dimension stripes that are the so-called nanoribbons, due to quantum confinement we will have a quantization of the states and we will have an opening of the gap. Now there are different ways in which we can cut graphine into nanoribbons. There are two main ways, this red one and this dark one. That create ribbons that are named after the shape of their edges. The red is an armchair ribbon and the black is zigzag ribbon and the zigzag ribbons are predicted to have interesting spin polarized edge states. Now if we focus on armchair ribbons, their electronic properties can be easily traced down derived from graphine and depending on the width, the number of rows of carbon that we have in the ribbon, we can have that in the back structure of the ribbon. We get close or not or we even eat the direct cone of the graphine and in this way basically increasing the width of the nanoribbon, the band gap will decrease and with superimposed annihilation due to different families of the nanoribbons. Now if we want to think about devices that we could fabricate, given the reduced size of these stripes that are just nanometer wide, it's crucial to avoid the defects in the fabrication of these objects. Is it possible to create atomic precise graphine nanoribbons? Yes. In our laboratory, Roman Fazel and Pascal Ruffier followed the initial work of Leonhard Grille for the surface-assisted coupling of molecular precursors and in these directions, our first studies, computational and theoretical, where investigate the unrelated coupling of this class of molecules. In this case, it was the formation of a 2D net of graphine with regular holes. And with this study, we basically demonstrated that everything in the coupling of these molecules is driven by the Ullman coupling on noble metal substrates. Our next step was to study cyclic data generation reaction. So how we can go from this molecule that seems like a patch of graphine but they are missing carbon-carbon bonds. How we can go from this molecule to this patch of graphine and basically here the formation of internal carbon-carbon bonds and removal of hydrogen atoms that were passivating the carbons. We demonstrated it is driven both by van der Waals reactions with the sub-state, metalsub-state, and by a moderate catalytic activity of the sub-state. Now, wrapping up all these ideas, Fazel and Ruffier succeeded in this picture. So the idea of getting a molecular precursor that has a shape that is cleverly designed and that has anchoring points that are saturated with boron or soalogen atoms. When you deposit these atoms on a noble metal substrate, the molecules will assemble in polymers and then increase the temperature, the cyclic data generation will start and we will end up with our nanostructures. The first successful example is this one starting from a dibromo beyond a trill molecule. We fabricated a seven arm-chair nano-ribbon that in this community is kind of the test system for both for theory and for reproducibility of the experiments. And from the simulation point of view, one of the things that we have to do most frequently is compare the scanning probe images that we have from the experiment, STM or non-contact AFM with simulation. And not always we are so lucky to have images that are so clear where there is basically nothing to understand from the completion point. But already in these cases to understand the origin of this feature in STM, simulations are fundamental. After these first results demonstrated by our laboratory, a huge amount of this class of always arm-chair nano-ribbons appeared in the literature from different laboratories, even including precise atomic doping. Here this work, if I remember words, was by Mike Cromney and Steven Louis, inclusion of Boron in seven arm-chair nano-ribbons. And also arm-chair nano-ribbons with the fancy shapes, Chevron shapes and again with atomic precise doping in nitrogen. This also from our laboratory and also from other laboratories up here. Now the question is can we out of these created devices to some extent, yes, we will see in one slide later. Now what do we do from the computational point of view? Standard, basically standard DFT calculations, always keep in mind that these objects are fabricated and characterized on a metallic substance. Whenever we can, so basically when we simulate reactions, chemical reaction on the substance or when we have to compare STM images in the experiment and simulation or atomic force microscopy images, we do DFT calculation with full substance, including in DFT, the substance and the adsorbate. So we can easily reach systems of the size of 2,000 atoms. Now in the case of atomic force microscopy, on top of this we use an empirical model that mimic the tip of the cantilever of the FM which is attached to the CO molecule, is modeled with a simple empirical model. That works pretty well in many cases. When you have to go beyond DFT, like for example GW, where we cannot include the substance and the only thing we can do is, as I was also discussed yesterday by Ian Willem, include image charge correction to our GW calculation that are done in gas phase. Now, if we focus, as I said, on our same nanoribbons, everything works smoothly. Everything is simple and the simple calculation match very well experiments. Now in this case, for example, we have the seven-arm chair that is predicted in theory by GW to have a gap of 3.7 volt. If we take these ribbons and we go along the axis of these ribbons with a scanning tunnel microscopy and we do spectroscopy, then we can find in the experiment the onset of electron removal and the onset of electron addition. And if we compare this experimental value of 2.4 electron volt with GW calculations of the ribbons, we see that we have a reasonable agreement when, to the GW result, we add the image charge correction. In this case, the work was a nice collaboration with Andrea Ferretti and the group of Elisa Molinari in Modena. We can do even more with experiment if I record the energy density along the ribbon axis, I obtain this kind of plot for occupied and unoccupied states. If I do the Fourier transform, I have access to the band structure of, let's call it the band structure of the ribbons, but these are finite objects. And from these I can retrieve, for example, the effective mass of the electrons on the band and these come out to be very close to the theoretical prediction. One has just to be a little bit careful doing the same experiment in theory. You can demonstrate that if you are not careful doing the experiment, so if the tip of the STM is not enough close to the substrate, you will miss some bands and this can be dangerous. It was a little bit puzzling. Sorry, I forgot to start this. It was a little bit puzzling to solve a problem that arises in ARPES measurements. When we did for this ribbon also ARPES measurements for ribbons that are perfectly parallel one each other because are grown on a steppate gold surface, we got that the effective mass from the ARPES and the effective mass from DFT was very off. There was no agreement. It took some years before in the literature it appeared the explanation for this agreement. And basically if you are not careful at which k perpendicular you are doing your ARPES experiment, you are missing the top vales band and you are erroneously assigning the top of the vales bands as we did to a band that is not the top band but is the second one. Now the only thing I can add on top of ARPES-CHAIRMAN ribbons is that indeed if we stay at the level of academic application so just a user effort to create a demonstrator devices it is possible to transfer these ribbons from the metallic substrate to insulating substrate and contact them and create a device. And they do not perform a bed. They are close to, so 10 to the 6 would be an ideal on off ratio and current bigger than one microamp. Of course what has to be improved is the device yield so one would need 90% of yield we have 10% of access here and of course one should find a way to get better aligned ribbons and log ribbons to really start having something more interesting from the device point of view. Now let's switch to a slightly different kind of ribbons so zigzag nano ribbons. As I said the success in fabricating ARPES-CHAIRMAN ribbons comes from the fact that we basically have created them via aerial coupling that works through the ULMAN coupling. If you want to fabricate ARPES-CHAIRMAN zigzag nano ribbons we should have a different kind of coupling this kind of orthofusion that doesn't work well. It's difficult to obtain. Now our experimentalists rather than pushing in this direction said why don't we play a lot with the shape of the molecules? We have excellent synthesis chemists let's try to imagine fancy molecules that then on the substance may be will alternate and bind in this fashion. And we increase the complexity of our precursor molecule we can close this pattern and finally here there would be a CH3 end up with the zigzag ribbon. I would have never believed it believe it or not it works and it works pretty well. This is the quality of the ribbons that you can obtain. Now again apparently you have a very good agreement between calculations and experiment but these ribbons are more reactive and in reality there are things that are not so well in agreement. Now let's have a look here. Here is basically the density of states measured with a scanning spectroscopy is in perfect agreement with the density of states that we obtain from GW these were also Berkeley's calculations basically the same that also Stephen Louis did and the agreement is a little bit too good should not be so good. How is obtained the experiment? This red line is measured on a portion of a ribbon that was dragged on top of the sodium chloride island to the couple of the ribbons from the substrate. Now I say that this agreement is too good because in principle if the GW would be perfect still there should be some screening from the substrate so there is something not perfectly well here. There is another aspect that I don't show you here that is we were looking we wanted to demonstrate that in zigzag ribbons we really have these edge states but and we have them if we measure on the sodium chloride but if we measure and we look for these edge states on the portion of ribbons that are on gold despite DFT predicts that these edge states would survive on gold in experiment you don't find them. Ok, that's all I wanted to tell you about this standard armchair and zigzag ribbon now I move to a different way of tuning the electronic properties of this ribbon basically not drastically changing the shape but slightly modifying it. Everything started by this work done by Oliver Groening he was digging in our AIDA database of nano ribbons. I forgot to mention that all we do from the composition point of view now for this ribbon is embedded in AIDA workflows as that were described before by Nicola. Now digging in our database of ribbons Oliver found that Stephen Louis explained that if you have a particular junction between a seven armchair and a nine armchair you can expect to have a topologically protected state at the interface if you have two junctions you can expect two states and if you play with the distance between these two junctions you can see that you can have a wide splitting of these two states up to merging them e then the ribbon basically becomes a nine armchair nano ribbon if you have many junctions then you have bends. Now what Oliver did was considering these bends considering a possible nano ribbon and saying basically I kind of have a state at each side of this functionalization of my ribbon and he was thinking about the SSH model and basically said well this is nothing but the SSH model should work like this and then just at the tight binding level he did all the homework of trying to vary the width and the position of these the effects now this case is just for the width and equal one and you can see that according to the SSH model you move from a topological insulator to trivial insulator according to the position of these functionalizing groups now this is very nice especially because these ribbons can be fabricated in reality in our laboratory and here again the tight binding bends you see that on top in the middle of what would be the bends of the seven armchair bone you see the SSH bends in the experiment you see the quality of the ribbon created you see traces of these states that originate the SSH chain and can we from the experimental point of view find out whether these ribbons are topological insulator or not well the only thing we can check is if we have an end state and then if well I forgot to mention you can also see that in the experiment you can detect the width of these balance SSH band and the width of the traction band and if you are looking for states the measured scan intelligence spectroscopy images match pretty well the simple tight binding images and there is no evidence of states at the terminal of these structures meaning that from this point of view this is not a topological insulator in agreement what was predicted a table that this class of ribbon would not be a topological insulator now if instead of this class of staggered modification we considered the inverse modification from the tight binding analysis should be a topological insulator and indeed if you do the experiment you find the states at the end of the ribbon how much time do we still have no I just skip this despite this is our first demonstration of a full carbon based device is an anoribbons on top of graphene leads we are now measuring the experimentalists are measuring this and we will characterize this in a collaboration with Mathieu Luizier also this is within the CCR Malve in this because I want to move to a different idea that is another way we have another thing we can play with to tune with atomic precision the properties of these nanostructures is to introduce four, five and seven membered carbon rings in our structures the first attempt we did with this respect was creating this kind of analogous of peritotracein molecules in this case we have something close to a stone waves defect in this kind of works what is fundamental to understand what you really have on the substance because atomic force microscopy images do not help us very well to really prove that we have this kind of structure we need a tight comparison between simulated microscopy and spectroscopy images with respect to the measured ones I will not enter the details of this just another example of a molecule that was synthesized and characterized both in theory and in calculation where you have basically an inverse stone thrower waves defects in a bucket ball molecule with embedded nitrogen atoms I want instead to spend a few words and just cut me when I have to stop and I stop on perfect on these other systems basically in denofluorane that are the anti-aromatic analogous of pentacin these molecules have resonance structures where you form a radical and kill the anti-aromaticity trying to kill the anti-aromaticity of the full molecule and they are interesting at the experimental point of view because they can have they can exhibit open shell character they have a small gap they are very reactive making it difficult to fabricate them they are usually fabricated in solution with protection groups and they have strong absorption in the visible spectrum these molecules is predicted to have negligible radical character this is predicted to have this one was already fabricated on surface in solution and we also created a polymer with this pattern of in denofluorane and now I just focus on this one and what I want to say starting with this precursor we were able to create at annealing up to a even temperature a chain of these polymers where we have the pentagons that are doubly hydrogenated and here to understand what we have in the experiment really you have to carefully compare atomic force microwave simulation with the experiment and here you undoubtedly see that you have two hydrogens in the comparison with the experiment now here things do not work as well as was before if we do the STS assignment of the gap and we compare it with GW the agreement is definitely not good and this probably is due to the fact that this system now are interacting with the substance they are not as the arm chain arrangement we had before that are basically physical orbit if we increase the temperature we obtain this kind of polymer it was for us a nightmare to understand what was going on here and here again a lot of comparison with simulated atomic force microwave images and simulated spectroscopy images so here for sure we don't have two hydrogens here there remain only one hydrogen so this is really in denofloren units what's going on here at the beginning we thought that an additional hydrogen went away but if you do not do the last event calculation you see that to remove one hydrogen it cost an energy that is comparable to the temperature but to remove the second hydrogen is impossible in the end we found out that this is just a tilting of the segments when they can match a gold atom substrate they tilt to partially bind to the substrate also in this case the matching between GW plus a much charge and the measure the scan and tunnel use spectroscopy onset of balance band and the conduction band do not match now there is a puzzle here that is before obtaining these chains of in denofloren at a lower temperature in reality we obtain these ribbons that are the orthofusions of chains and for us it is still a puzzle to understand how this can happen and now magically GW and the experiment match again because this is less interact we think, we cannot demonstrate it we suppose that it is due to the again reduced interaction with the substrate and it was a puzzle because whatever mechanism we can understand how this form at an energy that is lower to the one needed to form the singularity chain this is not possible so from our reaction calculation we should have first the singularity chain what we suppose is to get this we have trapped the gold atoms that then catalyze the reaction that for some reason become relevant when forming this chain this is something that we are still investigating I let me conclude I skip the conclusions it's basically just a summary of what I said I want to stress that this work is of course a huge experimental effort and I cannot thank everybody it's really plenty of people excellent PhD students postdocs in different experimental groups that did all the work but from the computational point of view this would have not been possible without excellent PhD students in the group of Daniele Pasterone we started with Mantone Guien that is coming from ICTP it was at SATP before joining us at EMPA and then Leo Poltali and Sasha Yakutovich that are now in part with Nicola Lozane in part with Berns-Mitt in Sion and now I have Christian Emre who just started the PhD with me thank you very much for listening and