Your videos are a lifesaver! I'm an undergrad working in a lab and the PI has just been pushing so much self-study onto me and I have no idea where to start. But your videos are giving me a good introduction. Thank you!
Thanks for the lecture professor I have two questions. 1) is it only applicable for ground state only . 2) in the term of when you said it's an exact theory but it's exchange correlation energy functional if that's the issue here so can we just “ignore” it since it's very small term and the remaining term can be treated as by average value of them as Born Oppenheimer approximation. I hope it make sense
It was great, especially the ending; the excited cat. HAHA One or two questions though, It has been observed that Minnesota functionals can show good energetics even without -D correction. Some people also say that adding -D to a Minnesota functional can lead to double correction. What is your expert opinion on that? Is there any lecture coming on Orbital Optimized MP/CC methods? Are there books/lectures already published about that? Can you share? Thanking You.
Yes, it's true, you can get good results for non-covalent interactions with some of the Minnesota functionals like M05-2X if you're "in the zone" (within maybe 4-5 Angstroms). If you go out further than that, then they don't really have a way to capture the long-range dispersion unless you add a -D to them. But if you do that, I agree that you might effectively "double count" some of those effects in practice (I've never actually tried it, but I'd be worried about it). This paper talks about how M05-2X and M06-2X can be ok for shorter ranges but not longer ranges in nucleic acids: dx.doi.org/10.1021/ct800308k
@@DavidSherrill1 I have tried with M06-2X to describe odd electron halogen bond where I have also used M062X-D3 for a sample system but the geometries didn't change. The inter-atomic distances were well within 4 A as you said. This is published in doi.org/10.1039/C9CP05374C
As Mt.Sc. undergraduate i found this insanely interesting. I'd love to know more about what we know about the exact functional. How close have we got? What are the properties we've proven it has?
I'm not familiar with a lot of the details on mathematical limits to the exact functional, but if you're interested in knowing more, check out some of the work of John Perdew
In some simulations I have run we see how close the functional is to describing reality by comparing physical properties with experiment, and some have been very close. But it is something that needs to be adjusted depending on your atomic system
Good, but very superficial. DFT and other quantum chemistry methods are about finding stationary points of the functional. All their theretical development is how to shrink the space of functions, which can compete in the variational problem. If you feel, you understood something about DFT from this lecture, I have bad news for you.
@@NeirinCedric To my opinion, such introduction creates a wrong understanding, what the DFT is about. Introductuctory and superficially are not synonyms. What is told is mostly what a user of the KSDFT codes see (that is why I used word 'superficial'). It has not been even mentioned about the variational principle (as far as I remember), which is a key to DFT, TDDFT, HF, and whatever quantum chemistry method. The new knowledge has to be connected to the old knowledge, and the introduction has to provide such connections or point out, how to establish them (i.e. mathematical tools). I cannot imagine that upon first introduction to the DFT like that I would understand, how to start learning the DFT. I would probably immediately start reading about xc functional approximations. Then, I would understand nothing, and would start using DFT codes without any knowledge on its limitations and physical meaning of the output. And this would even allow me to publish some paper, one of those thousend useless DFT papers, which are generated yearly, and bring no positive impact to the progress of science or even make it worse. DFT software has now turned into fastfood, anyone with almost no efforst can start using it and it always will provide you with some result, and you will always find a journal to publish it, because the world is full of incompetent physisics and chemists and bad journals. This lecture, to my opinion, will multiply the number of such 'scientists' who believe that they know what they do, but in fact they just multiply the amount of pseudo-scientific trash.
@@Tyomas1 I do not want to say something inaproparite about lecture. But, you have a point. I was strugling learning DFT not because the concepts was hard but the sources were not usefull and I have developed my own way to consume mathematical background of DFT. Yes, you are right it is sad that most of researchers using DFT without knowing numerical algorithms and Quantum mechanicals concepts and programming.
@@Tyomas1 If a paper helps determine the reaction kinetics of some enzyme catalysis, and propose specific inhibitors or at least classes of inhibitors... Why would such be trash? That's where I see DFT used most in form of MO:MM schemes, and the results tend to correlate well with experimental results and even have good prediction for inhibitors.
![Introduction to Density Functional Theory [Part One] Background](http://i.ytimg.com/vi/Ez_Fm4iTUeo/mqdefault.jpg)
Your videos are a lifesaver! I'm an undergrad working in a lab and the PI has just been pushing so much self-study onto me and I have no idea where to start. But your videos are giving me a good introduction. Thank you!
This is what is missing from a typical DFT introduction, thank you so much for the superb explanation (+100 points for loving black cats)
I have my research presentation tomorrow using DFT and this video helped me understand so much, thank you!!!
the cat seemed quite spooked at the prospect of DFT :D
I am very happy to find these great lectures. Greetings from Chile.
Can you help me for data of B3LYP-D3?
I was hoping onyx was going to present but this certainly a fantastic lecture too
Thanks for the lecture professor
I have two questions.
1) is it only applicable for ground state only .
2) in the term of when you said it's an exact theory but it's exchange correlation energy functional if that's the issue here so can we just “ignore” it since it's very small term and the remaining term can be treated as by average value of them as Born Oppenheimer approximation.
I hope it make sense
Absolutely wonderful. Very informative 👏
Brilliant man
omg I loved it, thank u so much, I learned a lot, the cat is amazing
VERY VERY helpful, thank you!!!
Good introduction!
Was kinda hoping for more of the cat
This is so good!
Wowww. Love to learn it.
The cat made the vid🐈 ❤
It was great, especially the ending; the excited cat. HAHA
One or two questions though, It has been observed that Minnesota functionals can show good energetics even without -D correction. Some people also say that adding -D to a Minnesota functional can lead to double correction. What is your expert opinion on that?
Is there any lecture coming on Orbital Optimized MP/CC methods? Are there books/lectures already published about that? Can you share?
Thanking You.
Yes, it's true, you can get good results for non-covalent interactions with some of the Minnesota functionals like M05-2X if you're "in the zone" (within maybe 4-5 Angstroms). If you go out further than that, then they don't really have a way to capture the long-range dispersion unless you add a -D to them. But if you do that, I agree that you might effectively "double count" some of those effects in practice (I've never actually tried it, but I'd be worried about it). This paper talks about how M05-2X and M06-2X can be ok for shorter ranges but not longer ranges in nucleic acids: dx.doi.org/10.1021/ct800308k
@@DavidSherrill1 I have tried with M06-2X to describe odd electron halogen bond where I have also used M062X-D3 for a sample system but the geometries didn't change. The inter-atomic distances were well within 4 A as you said. This is published in doi.org/10.1039/C9CP05374C
@@prasantabandyopadhyay2210 Interesting. Yeah, geometries are a little less sensitive than interaction energies.
As Mt.Sc. undergraduate i found this insanely interesting. I'd love to know more about what we know about the exact functional. How close have we got? What are the properties we've proven it has?
I'm not familiar with a lot of the details on mathematical limits to the exact functional, but if you're interested in knowing more, check out some of the work of John Perdew
In some simulations I have run we see how close the functional is to describing reality by comparing physical properties with experiment, and some have been very close. But it is something that needs to be adjusted depending on your atomic system
goooood
can you share your slide please?
Cute cat ;)
$\vec{r}^{\mkern3mu\prime}$
will help with overlapping of prime symbol and the arrow symbol.
This is one of those 'introductions' which assumes that you already know the topic.
Well you should know the basic at least
go jackets
Good, but very superficial. DFT and other quantum chemistry methods are about finding stationary points of the functional. All their theretical development is how to shrink the space of functions, which can compete in the variational problem. If you feel, you understood something about DFT from this lecture, I have bad news for you.
Wouldn't be an introduction if it wasn't superficial!
@@NeirinCedric To my opinion, such introduction creates a wrong understanding, what the DFT is about. Introductuctory and superficially are not synonyms. What is told is mostly what a user of the KSDFT codes see (that is why I used word 'superficial'). It has not been even mentioned about the variational principle (as far as I remember), which is a key to DFT, TDDFT, HF, and whatever quantum chemistry method. The new knowledge has to be connected to the old knowledge, and the introduction has to provide such connections or point out, how to establish them (i.e. mathematical tools). I cannot imagine that upon first introduction to the DFT like that I would understand, how to start learning the DFT. I would probably immediately start reading about xc functional approximations. Then, I would understand nothing, and would start using DFT codes without any knowledge on its limitations and physical meaning of the output. And this would even allow me to publish some paper, one of those thousend useless DFT papers, which are generated yearly, and bring no positive impact to the progress of science or even make it worse. DFT software has now turned into fastfood, anyone with almost no efforst can start using it and it always will provide you with some result, and you will always find a journal to publish it, because the world is full of incompetent physisics and chemists and bad journals. This lecture, to my opinion, will multiply the number of such 'scientists' who believe that they know what they do, but in fact they just multiply the amount of pseudo-scientific trash.
@@Tyomas1 I do not want to say something inaproparite about lecture. But, you have a point. I was strugling learning DFT not because the concepts was hard but the sources were not usefull and I have developed my own way to consume mathematical background of DFT. Yes, you are right it is sad that most of researchers using DFT without knowing numerical algorithms and Quantum mechanicals concepts and programming.
@@Tyomas1 If a paper helps determine the reaction kinetics of some enzyme catalysis, and propose specific inhibitors or at least classes of inhibitors... Why would such be trash? That's where I see DFT used most in form of MO:MM schemes, and the results tend to correlate well with experimental results and even have good prediction for inhibitors.
@@Tyomas1 can you give us a link to a more comprehensive lecture then?