 Good morning everybody, I'm Emiliano, Miliani Politi, and I just want to remind you that what is our plan, in the sense that as you have seen the lecture that you have just seen, is an introduction, a very short recap about quantum mechanics or better, quantum computational chemistry. The concepts that are really essential to understand the CP2K tutorial that we are going to do immediately after, in the configuration just quantum mechanics, let us say, not QMM yet. In the afternoon, instead we will focus more on the QMM aspects of the course. And of course, already in the tutorial when you will start playing with CP2K, you will be able to see some keywords that maybe not all of you are familiar with, pseudo-potential, basis set, etc. But of course the reason that was not explained immediately is just that at the beginning we want to give you a recap of the main concept, and then for all these advanced, let us say, topics we dedicated the session of tomorrow, where also in the lecture, in the third one, we will focus on these, let us say, more advanced topics. So I think, Emiliano, somebody has already asked a question, so if you are to ask. Yeah, yeah, yeah. So the basis set is intimidating to a non-QMM trained experimentalist. Any advice? Okay. Advice we have a lot, and of course, also doing the tutorials, much more will be explained, because when you will, let us say, deal with the specific system, we can give you specific advice. In general, and moreover, I will introduce a bit more concept or things in the lecture of tomorrow regarding the basis sets. In particular, the ones, for example, why we use the one in the tutorials, the practicals, and which are the, let us say, different possibilities offered in CP2K, in particular regarding the calculation with molecules. We will see a bit more in tomorrow in details. In any case, consider that CP2K is even a bit more complex from this point of view in terms of basis set with respect to other quantum mechanical codes, because you have to deal with both, let us say, the localized basis set and plain way basis sets. The reason is because with this strategy, with this approach, with the GPW approach, you can have very important advantages, in particular for people like us that normally have to deal with the large systems, and in particular the fact that CP2K has the ability to scale very well with respect to the size of the system. Larger is the system, of course, more computational expensive is the calculation, but how the computational effort increased with the size of the system is not so dramatic, let us say, as it happened, for example, for many other codes. In particular, for example, the typical comparison with another code which was popular before CP2K, which was CPME. Also pure Gaussian codes, Gaussian-based codes, Gaussian-based using codes like Gaussian, they are scaling in DFT usually as n to the power of 3 in the number of the molecular orbitals, and but the CP2K itself, it scales almost linearly with the number of molecular orbitals. So the third-order dependence will go in the size of your QM box, but inside the QM box you can have really a lot of atoms, and this there will be almost linear scaling with the number of atoms inside your QM box. This is a huge advantage of the CP2K. Yes, and also in addition for any trained experiment, so I think tomorrow the Holly will show you how you can choose wisely the basis set, how to check whether your basis set is good enough for the calculation. This will be tomorrow in the Holly's presentation. Yeah. Okay, there are other questions. Yes, I can read them. So good morning, Colin. I'm using the basis and real advantage of computational long-range interactions and users with QM on the basis set, second presentation. So I think that will be covered mostly in the second lecture, the advantage and disadvantage with QM. So one of the major advantage of using plane waves is because they are periodic, so you can do a periodic MMM, which with Gaussian basis set is now kind of questionable, so now appearing some very preliminary things how we can do this. But in plane waves it's very natural, so it will be covered in the second lecture. Yes. Then what are the differences between Cart-Parinello molecular dynamics and Born-Oppenheimer molecular dynamics? Okay, let me see if I can share with you these, oops, let's see if I manage. Can you see the slides that I have now on my desktop? This slide that I didn't include in this presentation was a bit more advanced regarding molecular dynamics and published molecular dynamics didn't focus specifically on this one in any case. Here I collected, let's say, three different schemes of abynition molecular dynamics, which let's say in general are rather popular when you deal with these topics and with different kind of complexity and in particular I have tried to make them to give this a similar appearance in order to, that it's a bit easier to compare them. Let's start with the Born-Oppenheimer molecular dynamics. You can see here what does it mean. Capital R means how are the coordinates of the nuclei, so how should I move the nuclei? The nuclei are moved more or less with something very similar to the Newtonian equation in the sense there is M acceleration equal to a force, but this force, of course, the force that we have to use to move the nuclei comes from an optimization. Sorry, Miliana, probably something happened to you. Can you share again your slide because it's now a black screen. And so I was mentioning Born-Oppenheimer, so Newtonian equation like to move the nuclei, but of course the forces come from a calculation, a quantum calculation, because it depends on the wave function of the electron, the position of the electron and that evolve according to the Schrodinger equations. And so this is the typical way I have to describe the Born-Oppenheimer molecular dynamics. Before you minimize the wave function of your system, you calculate the forces and then you move, you make one step of the molecular dynamics with the Newtonian equation, you move the nuclei with the Newtonian equations. Similarly, let us say in this sense, is the carl Parinello molecular dynamics, but what is the real difference? You can see here that still there is a Newtonian-like equation to move the nuclei, but the force is calculated in a rather different way. Instead to solve directly the time-independent Schrodinger equations, you have a different equation that is much similar somehow to another Newtonian-like equation. Of course it is not so straightforward, I did not enter in the details, but the idea here is that I have a way to calculate how the electrons that evolve, electrons for which I will need them in order to get the force to move the nuclei later. But the electrons dynamically evolve with an equation with the Newtonian-like, which is much easier to calculate with respect to the Schrodinger equation. This of course in the past was a very large advantage with respect to having a Newtonian equation like for the electrons with respect to the Schrodinger equation. Nowadays I have to say that improving all the algorithm to solve the Schrodinger equation and improving more and more the Born-O'Penheimer scheme, more or less in terms of performance the two approaches, Born-O'Penheimer and Carparinello, are very similar in terms of performance. And of course there is still the advantage of Born-O'Penheimer that in principle is a bit easier for a user to deal with to control with respect to Carparinello molecular dynamic where you have one additional parameter, let us say, to tune it every time, which is not necessary for Born-O'Penheimer. In case you are asking that what is this other Herrenfest molecular dynamics that I mentioned here, is a scheme, let us say, still a scheme of ab initium molecular dynamics where still the nuclei are moved with the Newtonian-like equation. But this time the, let us say, the wave function, the electronic wave function that you have to use in order to find the force to move the nuclei is not obtained by a time-independent-shaded equation, considering that, let us say, that everything is freezed, but you have to use the time, you have to use, you want to use the time-dependent-shaded equation, the real, full-time-dependent-shaded equation to calculate the electronic wave function or your problem. And of course, this molecular dynamics, ab initium molecular dynamics approach is a bit more general, allows you to, let us say, deal with problems where you cannot assume that you can use a time-independent-shaded equation at each time step. And so there are cases where this is possible, but this is necessary. But of course, this approach is computational, more expensive. I do not even know if currently this approach is implemented, this C2K. This really do not know yet. I know that it's available in other code, but I'm not sure that it's currently implemented. I can add here that RNFest dynamics is probably out of scope of this course, but it's usually for the excited state dynamics. So when you have several states, then probably you need to use the RNFest or fewer situations for surface-hopic dynamics or something like that. So it's another option, which is not for this course. Okay, so any other questions?