 So we're about ready to start the last session of the afternoon. And the title of this session is Simulating Materials and Process Across the Length and Time Scales. These are the two speakers we have. Just a word of introduction, the speakers are there. This is really multi-scale modeling, but with some particular aspects. So multi-scaling modeling is a much misused term as everybody knows. But here we're looking at some time and length scale problems. And as you see on the length scales I've put along the bottom there, there is a really difficult length scale which covers an enormous range of numbers of atoms where there are several simulation and modeling methods that get used. And if we look on the time scale, a molecular dynamics covers a fairly big range of time scales these days, especially if you can use some sort of classical force fields. So we are often very demanding on time scales and we want to do thermodynamics. And one of the difficult subjects in thermodynamics is electrochemistry. So we've got people now giving talks who are experts in certain aspects of the intermediate length scales and the difficult bits of electrochemistry who are going to talk about their work. So without more ado I shall hand over to the first speaker who is Caetano Melando Rodriguez. I'd like to thank the organizers for this opportunity to be back to ICTP. The idea is that... So first of all the total energy was one of the first conferences that I have attended and it also opens the doors to be part of ICTP as a postdoc. I am very grateful for the ICTP particularly the things that I learned here, the way to pursue excellence, tolerance and also free thinking without frontiers. It's something that in my country nowadays you have to think about it again. And definitely the experience here was very good for scientific and also for personal point of view. The challenge that I am going to talk today actually started also at ICTP when I was here on the top. One of the questions I was mainly involved with to work on materials under extreme conditions, high pressure, high temperature. And one of the questions that I have addressed, I mean, puzzle me, is actually how can you use the knowledge of atomic and molecular interactions in order to use for developing countries? I mean particularly the industry and also the resources that you have there. And some of these questions actually guide me what I am going to show to you today, that is the use of a multi-scale automatic simulations, where you are going to try to use for something that is interesting for a developing country. Particularly the natural resources exploration, here there will be the oil and gas, and also the infrastructure, the cement and asphalt materials. And to do that you are going to use it multi-scale molecular simulations. And not only can it be interesting for the industry, but also can help us to understand better some phenomena like your wetability and also fluid dynamics. And this can be very challenging. Now, the point that the system that you are going to try to address is the fluid flows through nano-porous media. And this can be very interesting for several kinds of applications, like as I mentioned the oil and gas, and also other porous medias like cement, batteries, fuel cells, and so on. The challenge here is that mainly for the space scale you have to deal with a lot of different odd of magnitudes between those pores. In particularly for the nano-porous media, those are the ones that experimentally they can be, they are invisible, so it means that even nowadays you don't have the resolution to see such nano-porous, but they are extremely important because thus control the permeability of the fluid through that media. Now, the point is from the nano-point of view, actually also the physics of the fluid that is confined can be very interesting because under the confinement the surface can play a role and can change a lot of the properties of the fluid that are confined, particularly the thermodynamics, your phase diagram can change, and also with that you need it really to go to the atomic scale. It means that the continuum model does not work at that scale. Now, one of the points that you have is actually how can you address this fluid dynamics through the nano-porous where either the molecular dynamics are not able to completely take care about the time, and also the continuum model is not able to complete the physics that you have on the systems. And to do that is very challenging for the point of view of two modeling complex systems, and here we are talking about realistic conditions over scales, means there's not only real materials, but real materials under operation conditions. And to do that you have several ways in terms of methodology. The chairman have already introduced some of the difficulties that you have to cross in terms of space and time, and also you have some of the physics, the laws of physics that can deal with each kind of region in terms of scales that you have it here. Now, for the phenomena that I want to address, the fluid dynamics, you actually have to go for all these scales. If you want to try to understand how the fluid confines on a scale can be described. Now, what you have done was actually a hierarchical way where we're going to start with quantum mechanics using first principle calculations in order to derive force fields that can capture the dynamics that you see at the first principle. Then you're going to use the molecular dynamics that should determine the thermodynamic properties of those systems. And finally do the so-called Let's Boltzmann method where you're going to carry on now the fluid dynamics in the right scale of space and also time. Now, this can be very challenging in the sense of you have to build some bridges between those methods, and this is more or less what I'm going to tell you today with one example that is how you're going to displace oil or gas through a power media that is the rock. Now, for the first step, you actually derive the assumption that I learned here in Trieste. Derive potentials based on a binomial dynamics and you have to do the assumption of the DFT in this case with van der Waals is correct enough as well you create an effective potential that is fitted from DFT. So in this case, I forget about the electron structure, sorry about the total energy conference, however, there are ways that you're going to carry this information where in the sense that the forces that are going to describe are the forces that have been derived by Abinish methods. The second step is also very important because you need to somehow deal with large scale. In this way, my molecular dynamics, now it can be the Abinish or classical, are going to be replaced by a distribution function where I have two main points. One is the spring part and also the collision parts. It also can include some of the forces that is involved on this phenomena, particularly the wetability or the interface problems. This is something that I'm going to discuss to you in a few slides from now. Well, one of the points is how to control the interfaces and the flow at an un-scale. To do that, there are two main ways. One is to reduce the interfacing, the other one is to reduce the viscosity. In the particular case of the oil, this is the way that it is done. There are several ways to do that. One is changing the chemical environment, means that you're going to control the wetability. The other one is either to control the confinement, so if you want to have a pressure-driving flow through your system, and finally to control the electrostatic environment. It means that there's an electrokinetic driving flow that you have in the system. What I'm going to try to show to you is that you are able, by using multi-scale molecular simulations, to know or predict how much oil you can displace over these scales. Now, just to tell you how the challenge is, so if you think in terms of a macroscopic scale, so here's a meters or kilometers, but my real system is that in a micrometer scale, the interface between the rock, the oil and the brine, this is a multi-component, multi-phasic, and under realistic conditions, means high temperature, high pressure. And what I want to try to do is to control the flow using nanoparticles that control the interface between the brine and the oil. To understand how is the effect in the flow, you have applied either to try to understand the geometry of the pores, means that you want to try to see how the shape and the size and the distribution of the pores can change, as well as the chemistry behind the wetability change. Means you're going to inject a fluid that does not have nanoparticles or can have a hydroxylated nanoparticle, hydrophilic nanoparticle, hydrophobic one or hydrophilic. So you're going to try to see if a given solution that can be displace injecting the system can allow this fluid to be displace. Now, the interesting point is to do that, you have to really start it from the first principle calculations, that you need to have a very good description about the interaction of the molecules with the rock surfaces. And this is some of the examples that I actually chose, because those are examples that I have shown in the previous Total Energy Conference. One is related to the asphaltenes, the scousite, also the solvent ion exchanging effects that I have in the system. The connective studies on mineral surface and finally the nano aggregation of asphaltenes and resins. So if you know how is this process, particularly the energy involved here, the idea that you can derive force fields that can capture the dynamics that you see at the first principle level. Now, with that, you can do something more interesting from the molecular dynamics point of view. This is an example of a table with the composition of crude oil that has been given for us for a company. You can build an oil model and also the Brian model that you can see in reservoirs. So in this way, you are able to obtain the interface properties that you have in this system and this oil model can be as complex as you want it. This is very interesting because now we are talking about the multi-component systems under really reservoir conditions, at high temperature and you are able to prove interface properties that you have there. Just to tell you, to give you to your motivation, in order to obtain one single point of the interface tension in the resource center at Petrobras, they have to spend three months in terms of measurement because you have to wait for the oil to be in equilibrium under high temperature or high pressure and this molecular dynamic can be very helpful in order to quantify those problems. Now, you also have a play to design nanoparticles in order to control this waterability and this also has been done using first principle interactions and in the end what you want is to be able to calculate by molecular dynamics using first principle forces-based simulations where you can control or quantify the interface properties between the brine, the oil, the rock and also the nanoparticles. As an example that I have chosen to show to you, this is a surface driving flow where you first try to understand how the functional groups of the nanoparticles can interact with the rock surface and this is done at a DFT level and this is mainly the force with respect to the distance so how these functional groups interact with that surface and then can be translated in a potential inter-atomic model where now you can describe the whole system, the brine, the oil, the nanoparticles and also the rocks. So up to now you can do a fully atomized calculation and even hybrid models which allow us to go for very large nanoparticles. Well what you can learn from the molecular dynamics so this is a snapshot of my movie about the molecular dynamics but the interest point here is to be able to know what is going on in the interface means you are able with the molecular dynamics to feel that you have close to the interface and not only that you are also able to calculate the interface the interface station the free energy variation with the area here and in this particular case we are calculating using variations of the pressure well with that you have explored for different kinds of rocks realistic models of oil and also under the conditions of reservoir high temperature and high pressure the interface station with respect to the salinity for different kinds of nanoparticles the hydroxylated hydrophilic and hydrophobic ones also the viscosity and the contact angle that controls the wetability phenomenon in this system this is very interesting because you can carry on accurately the first principle of forces that you have derived in a proper that is very much from the interest of the industry that is the interface station however the point is what you can do next because with that you are not able to actually simulate the fluid flow in the sense of the typical size that you can have it in some post media and to do that you are going to be beyond molecular dynamics reminded that at molecular dynamics level I have derived potential that have been derived by first principles so we are moving for the scale of nanometers nanoseconds up to micrometer millimeter to microseconds in milliseconds the way that you have done is basically based on the molecular dynamics properties that you have it the viscosity, the density the interface station, the contact angle you have some parameters that have been derived for the electrical method and now you can match the molecular dynamics properties in a fluid that can be simulated at this scale that is now micrometer or millimeter so in this match you have to be worried about the relaxation time that has to do with the viscosity and also the fluid fluid and fluid mineral interactions that you can obtain directly for the interface station well this is very interesting so now what you can do is basically what I am going to show to you is a way where you are going to inject a solution that has nanoparticles or not and see how much the oil is going to displace for a given rock geometry and we started with a microphotography of the rock you build up a computational rock model and in this way you are going to try to see if I inject the solution how much oil that is trapped here can be displaced in the example I am going to show again in this case this is a Let's Boltzmann calculation that has been derived where you have embedded the fluid that you have here you have the properties that has been obtained by molecular dynamics where the force fields have been derived by force principles so this is the injection in green you see the oil that has been displaced and this is very interesting because now you can understand the fluid flow based on information that comes from the very beginning for the first peaceful point of view so typically this kind of simulation has been done in a qualitative way in the sense of they try to change the parameters in order to fit what has been seen experimentally in the end you can obtain how much oil can be displaced with respect to the time and the particles or different kind of chemical additives that you can add on the system in this way you can choose or the industry can choose the one that the best can displace your system now not only this is interesting from the point of view to optimize what chemical additives you can add in this particular system but also you are now able to see how the geometry of the system has our industry so what we have done is to create a series of geometries where you can change the shape the size and randomly distribute those particles media with the same imporosity this is something very important because the samples that you have in the rocks each sample is different for each other in this way you can control inject your fluid and see how much of oil can be displaced and you also are doing some experimental measurements to compare what you see from the last Boltzmann and also with the experimental point of view I just point out that the first time that I saw a 3D printer was here at ICTP and after that visit I come to Brazil and I want one of that and you try to do the last I mean the porous media however the resolution here is not enough so you have to really find resolution to see the geometry of the porous with respect to the system but this is something very interesting because now you can learn how the permeability change with respect to the geometry and also how is the fluid flow for those systems and not only that then you have something that I learned here the channel entropy so you can create with different entropy in the end you can have a correlation between the hydrodynamics and the porosity and this was very interesting because you can directly plug this information in a reservoir simulator Reservoir simulator means that you are dealing with many cells that are sort of meters or kilometers so in this way you can see how much oil can be displaced with the time in this tell us how is the effect of hydrodynamics or the order disorder effect that you have in this system now very much very interesting so a few years ago we started to do this project and collect a lot of information and one of them that also it was very interesting to learn here that this is called machine learning and you have been able to create a database on the interface station and the question was if you will be able to obtain a model can we describe our interface station based on molecular dynamics calculation so in the end what we have done is mainly to see how the salinity, the brine composition the oil density and the oil composition can can actually tell us something about the interface station of this system now with that we mainly have explored different models of oil completely aromatic one or alkane is also mixing of the gas gasoline that you have used in the complex oil that petrobras give to us so in this way you are able to create a database on the interface station in the end you have obtained a model that depends on the composition of the oil, the salinity and also the concentrations of the salt, the ions in the brine composition so this is very much interesting because in this way you can learn a little bit about how be the facts of each one of this features and you end up if you compare with some experimental data and the other data that is available in the literature for those models you have a aero of 2% so what is very nice is the interface station point of view again experimentally you have to wait for 3 months to get one single data here and with that you don't spend so much computer time or experimental time so you can really do screen over possibilities that you have it here now what you learn here is that you can decrease the interface station because of the aromatic fraction that is most important the other one is the low salinity the diavalent cations and the sulfates are the one that if you want to change your brine because they have to inject the brine in order to displace the oil so those are the ones that you have to look for and here more or less give us a recipe how actually would be a good composition of the brine that can displace the oil now this is also interesting however to come back to the physics to try to understand how these effects what is going on from the physical point of view so what is how the interface stations change due to the composition of the oil and also to do the composition of the brine for that you have two main effects one is the you can see this peak here you have accumulation of aromatic molecules the hydrocarbons near the interface and mainly the interface tension comes from that and the second one is the change the hydrogen bond network due to the fact of the ions is going to change that and this is something that you try to look at for the different models of the ions and also increases salinity how it will affect in terms of the interface tension so basically what you learn is that the ions can change they lead to a water coordination reduction and this is a really effect that can change the miscibility of those systems and make it easier to displace now with that you want to learn a little bit more about the hydrogen bond network because in a way that I wanted to describe how is the this effect and what you have done is actually now see another interface that is the rock with the water and Brian with molecular dynamics you are able to see how the water actually you have accumulation structuring of water near the interface and more interestingly you can do something that is a Google view of the interfacial phenomena means that I going to transform my water coordination on my water network in something that is a graph theory so here the vertex is the oxygen atoms and the edge are the hydrogen bond network so with that you are then able to map how is the configurations of the water molecule near the interface far from it and how is the effect of the ions changing the hydrogen bond network of the system this is pretty much can give us a very nice insight about what is going on on the interface but with the different eyes not only looking for the water molecule itself but now from the point of view of the hydrogen bond network the point here is that if you want to visualize what happens in terms of the interface when you have the water in the oil or the water in the rock so to do that you are doing two other things that is a new way to see and to be the first one I was falling off last year I have been in a conference in PISA that to use the virtual reality to visualize our systems this is really very much interesting particularly for molecular dynamics because it's very cheap you can visualize now change your ions and see on the fly what is going on on the interface I really recommend you to try nowadays you have software that can do that and also very much interesting to use this data to transform this data in sound because if you want to try to visualize such situation for each more snapshot that is 300,000 atoms this is very difficult so what you have done is to map our hydrogen bond network in sound in the end you can listen what is going on if you have a book water or if you have water confined at one nanometer so this is going to show to you now so this is the group that happens in the music part now what you are going to do is sound is very interesting property because your brain can follow a given channel a given instrument musical instrument separately so if you try to follow the violin for instance the orchestra you follow it so if you try to follow the piano and so on so what are going to invite to you to close your eyes because as much is enough to listen and we are going to listen to two cases one is the book model for the book water and the second one will be at one nanometer so let us see the book of the water if this sounds this is how sound the water molecules the formations of the water molecules actually or the book phase now you are going to listen the same thing that is now when you are confined at one nanometer so at one nanometer you can see that this is much best and you start to see other configuration you can listen other configurations actually of course this may not be a hit parade however you can tell us a lot about the dynamics of the water particularly for hydrogen bonding network point of view and really tell us how this works so with that would you like to close I mean this is a hierarchical multi-scale approach going from first piece for molecular dynamics that Boltzmann finally machine learning and also this could be sorry for that not so noble application that the oil and gas but that is something that for Brazil for instance is very much important however the same tools can be used for batteries, fuel cells and CO2 sequestration and capture and also provide to you new ways to see and to be using either vitro reality and sonification so with this would you like to thank my team in Brazil and Emilia Tacho where is he so he telling me if I would be able to do jazz and I put the jazz part here but sorry it sounds terrible so with that I finish my talk thank you very much