 It's great. Thank you so much. That's even better. So I come from a slightly different community, but it's a good opportunity to discuss with all of you the dissipation phenomena that occur in these nanoporous materials. But before I go into the subject of the matter, I would like to point out other research lines that we are covering, because that might give rise to, say, to other interactions at lunchtime or so. So we also deal with superhydrophobic surfaces, metal nanoclaster with the final aim of, so to say, studying the tribological properties at surfaces, and interfacial slip, both liquid solid and liquid liquid. But today I will talk about dissipation phenomena in nanoporous materials immersed in water. And I would like to spend some time in explaining what the system looks like. And it is really a sort of granules of nanoporous materials. And inside of each of these granules, there are billions of pores. And they're made hydrophobic by some tricks and then immersed in water. You can also use another known wetting liquid. And then you can do pressure volume cycles in this assembly, so in this sealed container, and plot out curves similar to this. And what you see is that because of hydrophobicity, water enters into the pores only at relatively large pressure. So here in this special case, it's on the order of 50 megapascals. So you have this intrusion pressure. And on the other hand, once all the nanopores are wet, then in order to extrude water, you have to form a new phase. And this generally occurs at lower pressures. And so these two intrusion extrusion pressure define an hysteresis cycle, which of course gives rise to dissipation. And here dissipation is key because you can have an entirely different range of application depending on how much energy you dissipate per cycle. In particular, you can have vibration dampers when you have a large but repeatable cycle. You can have single use shock absorbers, like in car bumpers, when you don't have hysteresis, when you don't have extrusion. So you just dissipate energy once, and then you throw away your system. So you dissipate once, but then you don't go back. Or you can tailor your system to have intrusion extrusion pressure very close by such that you can minimize energy dissipation. And you can store energy in the form of interfacial energy and take advantage of the enormous surface area of these pores, which can go from hundreds of square meters per gram to thousands of square meters per gram. And then I added to the slide after the talk of Alessandro Siria also the fact that more recent application take into account electro-tribology happening in this system because you can charge and discharge by intrusion extrusion that is the solid. But what is most striking actually is the extrusion process because it's like seeing boiling at extreme pressures or at two or three megapascals. And this can be visualized, for instance, if you add hydrophobic solutes. And the reason is that because of confinement and because it's hydrophobic, water fluctuations or liquid fluctuation, if you wish, inside of these pores is enhanced. And let me repeat the video once again. And so what happens is that really you shift the pressure at which you can see boiling or, if you wish, trying by a great amount. And you can tune it. So the main question for today is, how do we control wetting and drying dissipation coming from wetting and drying hysteresis? And in order to, which in the end, boils down to how you control boiling and confinement. And we have different ingredients we can play with. And today, on the menu, we will play with all of them. So you can change the hydrophobicity, the size of the pores, the geometry, the topology, for instance, how much they are connected. And then you can have different types of pores. So molecularly defined or mesopores. And then you can play with pore elasticity. So the computational, but even experimental challenge of these systems is that you're really dealing with a confined phase transition. So in order to go from the field states to the empty states, you have to overcome a barrier, which is typically larger than KBT. So you have two timescales, one of the microscopic dynamics, and one which is a useless. So to say, waiting time before you overcome the barrier. So typically, you have to use rare events techniques in order to accelerate sampling. But in this case, the collective variables, so the descriptors for the wetting or drying process are not trivial at all. So along the years, we have used a version of umbrella sampling, or we have used the string method in collective variables, which allow you to, so to say, use much more general collective variables, the continuum equivalence. Now we're moving towards machine-learned collective variables in which, so to say, you don't have to guess what the variables are, but they are automatically learned. So let's try to have first a macroscopic idea of why drying occurs at all in nanoscale confinement. And this is relatively easy, because if you write down the free energy for a confined fluid, you have a volume term, minus PV, and a surface term, gamma A, where A is the area of this slit. And if you compare the free energies for the capillary vapor and the capillary liquid, this gives rise to the coexistence conditions. And this is really Kelvin Laplace, and also known as Kelvin Laplace law, which tells you that if you have a hydrophobic contact angle, so theta young larger than 90, you can see drying at positive pressures, so basically at pressures larger than the ambient pressure. And this has to do with hydrophobicity, but also with confinement. So this effect is larger, the more you confined your system. OK, so these were the two first ingredients of our menu. And the other thing one has to keep in mind is that once again, drying is really a nucleation phenomenon in which you have to generate a bubble, and thus you have the usual competition between bulk and surface terms, which gives rise to a free energy barrier. You can see this as a signature, because if you do constant frequency experiments, so you cycle over intrusion and extrusion, you see that the extrusion pressure depends on the frequency, or if you wish, on the time of your experiment, and also on the temperature. So you really have a thermally activated signature of this process. So since we are so confined, we can give a look to the atomistic details, and that this was an all atom simulation with tip for P water. And in order to drive the event and to measure the free energy, we used a coarse-grained, as a collective variable, the coarse-grained density of the system. And basically what we obtained is what is the most probable way in which a bubble forms inside a cylindrical pore, which occurs with an asymmetric formation of a bubble, which then forms like a pillory neck, and this is the transition state, and then two symmetric menisci that move away towards the cavity mouth. But the most important bit is that we have the free energy profile. From this free energy profile, we can see that there are strong deviations from microscopic observables. But moreover, we can basically repeat the calculation different pressures and build virtual intrusion-extrusion cycle to be compared with experiments. And here is the most interesting part for this community. So of course, depending on the frequency at which you drive your system, you have different dissipation energy per cycle. And if you go very, very slow, of course, you dissipate less energy. And if you kick it hard, instead, you have stronger hysteresis. The main point is that intrusion, per se, is as well an activated event. But you don't see it in experiments because it's basically the nucleation bubble is so small, sorry, because it has a very large critical volume. And so very small differences in pressure are sufficient to accommodate for very different frequency. On the other hand, since the critical bubble for extrusion, so for the formation is very small, you have an extreme dependence of the extrusion pressure on the frequency. And this gives rise overall to dissipated energy, which scales with the logarithm of time. And this, so to say, was known, wow, one sec, was known also explains experimental data for a variety of different pores materials. So now we come to a different constructive parameter that we can use in order to tune this dissipated energy, which is how pores are connected. So an experimental friend of ours came to us saying, look, we have two materials. The one on the left has semi-independent pores of six nanometers. The one on the right has roughly the same size, but the pores are interconnected. One exhibits extrusion, and one does not exhibit extrusion at all. What is the explanation for that? So the nominal size is the same, but the behavior is entirely different. And so we made a spherical cow model for these pores, one with independent pores, and one with lateral cavities, which mimic, so to say, interconnection between pores. And what you see is that, indeed, one of the two is able to close the cycle, so to exhibit extrusion, and the other one not. And if you want to understand this in, so to say, hand-waving fashion, the reason is that the case on the right has very small hydrophobic pores, which are never wet. So even at high pressures, you cannot wet them. So it's like the main pore has a super hydrophobic surface, rather than only a hydrophobic surface. And these favors the formation of the first bubble. So it's able to decrease the free energy barrier for nucleating the first bubble from 150 kBt to 5 kBt. So it really accelerates nucleation. And, well, our friend did not believe that this is an entirely general mechanism, so has nothing to do with water. So he repeated the experiments with mercury. Mercury is advantageous because you don't even need to silanize the surfaces in order to have a non-wetting liquid. And so he used these model pores, which are cylindrical, MCM41, and these are the pores which are random. And you see that quite unexpectedly, the regular pores have very large hysteresis cycle because they lack these pore connections. Whereas the random material, because of these pore connectivity, has an extremely low hysteresis. So then we move to microporous materials, which in this community means that that one was silica. And this one, instead, is really molecularly defined. So it's a system in which you have a crystal structure in which you have embedded pores. So you're at scales that range from one nanometer below. And the main point is studying what happens if you have connections between these pores. So you have the main pores of 1.5 nanometers and then smaller pores of 0.7 nanometers. And so you can compute the pressure of intrusion and extrusion. And here comes the surprise. Based on the previous result on mesopores, we expected these pores to be more hydrophobic because of this super hydrophobic argument. Here instead, we noticed that the presence of lateral cavities can make these pores all the way from more hydrophilic to more hydrophobic. So more hydrophobic would be the case of mesopores. The two lines are the reference pores without connections. But you can also think of making them more hydrophilic. And the main difference is the length of the lateral cavities. So for long lateral cavities, the behavior is more hydrophobic. And for short lateral cavities, the behavior is more hydrophilic. And the reason for disqualitatively different behavior as compared to the mesopores is that water inside the supernanometric cavities connecting the main pores forms a single file. So basically what happens is that forming hydrogen bonds across these connections is energetically favored for short pores, whereas it is progressively less favored for longer pores. So this is very, very different from microscopic prediction based on Kelvin Laplace law, which would say that there is no dependence whatsoever on the length of the pore, but just on the pore radius. And by the way, at these supernanometric pore radius, the pressure at which you expect intrusion is extremely high. So in the end, you have a finite probability of wetting the pores, even for subnanometric pores of one nanometer, which is entirely unexpected based on microscopic grounds. So very recently, we started playing also with pore elasticity. The only thing that I want to tell here is that if the pores materials is also elastic, there can be counterintuitive effects. Because you would expect that since the pore shrinks, the intrusion pressure increases. But these pores materials are quite complex. And so for instance, in ZIF 8, we noticed that the opposite happens. So if the material is stiff, you have a higher intrusion and extrusion pressure. If the material is flexible, instead, it has a lower intrusion and extrusion pressure. And this is because elasticity also acts on the sort of saloon doors that separate the cages of these materials. So it's not just that the material compresses and the pores shrink. So in the end, we can play with these different elements, hydrophobicity, pore size, pore connectivity. We have to take care whether we have mesopause or micropause because very different regimes. So to say, non-classical regime might apply to micropause. And pore elasticity to control dissipation and go all the way from, so to say, dissipative system to energy storage devices. This is a real system with almost no hysteresis. And micropause materials are really a nice playground to control these properties in an extremely careful way. With that, I would like to conclude and saying that the exotic phase behavior in extreme confinement provides flexible and robust dissipation mechanism, which can be driven by external parameters, so pressure, temperature, et cetera. But also that there's a large design space to control these phenomena, which I have just said. And so to say, our perspective is now to try to learn from biology in which also has hydrophobic pores at nanometric and sub-nanometric levels in order to learn different things. For instance, how to play with a complex chemistry, how to propagate movement on long directions, and how to make a comparative, so an emergent behavior because of a combination of different pores. With that, I would like to thank you for your attention and collaborators. Also for keeping time so nicely. I have air in these channels before the water comes in. And does it mean that the air gets compressed or does it get dissolved into the water? And also when you have nucleation of cavities, do they dissolve in the water, influence that process? OK, thank you. That's a very good question. So in the general case, so since you have a sealed container, you can realize this sealed container in different ways. The usual ways that you would digas both the liquid phase and the porous phase before putting them together and sealing them. So normally you have a controlled amount of air inside of it. But we're attempting to use air, so to say, as a means to favor nucleation. And at high pressures, you always dissolve air. So to say, the experimental knowledge is that above, I think, five megapascal, you dissolve all air. But this still can help drying. And we have preliminary evidence that this happens. And the reason is this, because air can act as a nucleation site which further decreases the nucleation barrier for a new bubble. So air plays a role. And it can help also controlling the extrusion pressure. So this can be controlled. And it's a good parameter to play with. That's a great question. I have sort of a clarification, maybe Naif, but just to make it clear. You have a contact angle that is larger than 90. You should expect capillary drying, right? Yeah. And it's less than 90. You should expect capillary condensation. That's a first statement. Yes, for a simple geometry, yes. I have seen some talks that people are condensating bubbles between a hydrophobic AFM tip. So that's not what you would expect, right? No, this is not what you expect. But still, a geometry plays a role, right? Because in the end, it's a competition between bulk terms. And then you have surface costs for the liquid vapor interface. And then you have an interfacial advantage in case you have a hydrophobic thing. So you really have to do the calculation on a specific geometry in order to answer it to your question. So here, everything is simple because it's a slit. And you can do it analytically. But in general, it depends on the geometry and on the curvature. The phobic and your surface is high. I guess so. I didn't look into the problem by that, I guess so. Yes. Thank you. Is there any specific length scale for the pore that the fluid could not produce a slide, I think? OK, so this is a macroscopic law that tells you that basically? For molecular dynamic simulation, have you tried different pore size to see if there is any? Yes, yes. So this is what we're currently exploring. So the microscopic law tells you that it is, so to say, more difficult as 1 over L to intrude the pore, which is hydrophobic. But then we saw some counterintuitive result, which is the one that I showed you in the last slides, that even for extremely subnanometric pores, you can put water in. And the reason is that water forms hydrogen bonds, so they can organize also in a single file. And of course, you cannot go below some threshold because you have steric hindrance. And so basically, the water molecule cannot enter. But for these subnanometric pores, you can still see water entering and that takes ambient pressure. So basically, completely define the macroscopic expectation. When you mentioned the effect of elasticity, you sort of suggested that elasticity would be reduced to some shrinkage of the pores. But why? I mean, if you increase pressure, also pores should expand rather than compress, right? OK, so it's the way to say. So this would be the image below. So as long as the pores are dry, what you're doing is a hydrostatic compression of the material. And so in this case, you expect the pores to shrink. No, no, no, that's the point, Nicole. So if you have the pores that are dry, inside you have vapor. No, no, it's not at the same pressure. It depends on the chemical potential. But basically, you can imagine that inside, you have the vapor pressure at that temperature. So this is fixed, and you keep increasing the liquid pressure. So the delta P increases. So you don't have the same pressure. And then what happens is, and this is what we publish in this paper, is that at some point, you trigger the transition, so you wet the pores. And at this point, you release all capillary energy, and this expands again. So this is why it gives rise to an enormous negative compressibility, because you compress it, and then it springs back to the initial thing. So that's why I say it compresses. Because as long as you have vapor inside, you can compress it. And then, of course, if you have the same phase in and out, nothing spectacular happens. Thank you again. Next presentation is online, so hopefully it will work. So it's on.