2Likes
LinkBack URL
About LinkBacksThis has never really sat well with me; does theoretical chemistry allow the order of a reaction to be derived or modeled by computer etc. or is it really only possible to obtain it by experiment?
March 12th, 2014, 06:25 PM
This has never really sat well with me; does theoretical chemistry allow the order of a reaction to be derived or modeled by computer etc. or is it really only possible to obtain it by experiment?
It is only possible to obtain them from experimental data.
Indeed, this is what I've seen wherever I've looked it up, but surely this implies that our theoretical understanding of how chemical reactions take place is incomplete?Originally Posted by PhDemon
Quantum Electrodynamics should at least allow any problem in chemistry to be posed in its entirety even if the majority can't be solved exactly. Surely at least some of the reactions can be modeled entirely? Could the order of reaction be found in this way?
Actually, am I looking at it the wrong way, say maybe the order of reaction concerns large scale dynamics and it's more a problem of statistical physics for example?
The order of a reaction is an empirical measure of how the rate of reaction varies with concentration, if you could calculate the exact mechanism from first principles and model the kinetics and thermodynamics as a function of time, concentration, temperature and pressure you may get an answer. The first part is something we can't do yet with any real accuracy.
Last edited by PhDemon; March 14th, 2014 at 05:58 AM.
Yes see PhDemon's reply. I don't see how quantum physics will help.Indeed, this is what I've seen wherever I've looked it up, but surely this implies that our theoretical understanding of how chemical reactions take place is incomplete?
Quantum Electrodynamics should at least allow any problem in chemistry to be posed in its entirety even if the majority can't be solved exactly. Surely at least some of the reactions can be modeled entirely? Could the order of reaction be found in this way?
Actually, am I looking at it the wrong way, say maybe the order of reaction concerns large scale dynamics and it's more a problem of statistical physics for example?
It strikes me that one difference between chemistry and physics is that chemistry has to cope with an almost infinite combination of arrangements of atoms. Real molecules are very complex systems, compared to most of the systems physics deals with quantitatively. We can't even solve the Schroedinger equation exactly for any system more complex than the hydrogen atom. In even a simple molecule you have numerous internal degrees of freedom among which energy can be distributed, in a huge number of quantised levels. Collisional modelling of reactions is fiendishly difficult, even in the gas phase. When you have solvation effects as well, in reactions taking place in solution, it is likely to be worse.
I see, thanks for your replies - can you give me any suggestions for where to look for attempts at this? Is that a subset of physical chemistry?
Well ultimately QM is how the world works - we have trouble calculating reaction cross-sections for collisions in the many hundreds of MeVs (the domain of other forces) but at a few eV the physics is very well understood. Quantum electrodynamics has an unparalleled agreement with experiment. As I said, I do appreciate the difficulty in solving practical problems in chemistry with molecules and many electron atoms so my query wasn't a complaint or meant to imply an obvious failure on any ones' part I was just curious why I can't find any attempt at deriving the course of a reaction and a full expression of the rate entirely from first principles.
At the simplest, 2 hydrogens colliding, the classical Hamiltonian at any point is known (and indeed the QED Lagrangian is known at every point in the process, but that's really not necessary for such low energies). Is it really true that we cannot model this sufficiently to predict the steps of the reaction in advance and find the cross section for each step?
Your point about the multiple orientations making a difference is a good one if you want to treat it like a series of colliding particles, I'm sure for things like proteins, it's vastly more complicated than my naive approach would be able to deal with!
(Also, an exact solution does exist too for the classical quantum Helium atom iirc, thanks to Richard Feynman)
I would disagree.
Not only unnecessary but also completely useless. QED is for different things.
We can. You yourself said that Hamiltonian is known. Thing is it is incredibly computationaly demanding.
I am not so sure about that. Also what you mean by classical quantum?
Well you might I suppose for H + H -> H2. But even for that, you may find you a "chaperone" molecule helps to carry off the excess energy in the newly formed H-H bond. If it does, the reaction may have a higher order, due to the greater likelihood of the bond staying intact if a molecular collision - with another molecule , that is - can be expected before it the new bond has a chance to break apart again. I have not looked up the kinetics of the atomic hydrogen combination reaction but I expect somebody has generated the data on it, at least in some pressure/temperature regimes.I see, thanks for your replies - can you give me any suggestions for where to look for attempts at this? Is that a subset of physical chemistry?
Well ultimately QM is how the world works - we have trouble calculating reaction cross-sections for collisions in the many hundreds of MeVs (the domain of other forces) but at a few eV the physics is very well understood. Quantum electrodynamics has an unparalleled agreement with experiment. As I said, I do appreciate the difficulty in solving practical problems in chemistry with molecules and many electron atoms so my query wasn't a complaint or meant to imply an obvious failure on any ones' part I was just curious why I can't find any attempt at deriving the course of a reaction and a full expression of the rate entirely from first principles.
At the simplest, 2 hydrogens colliding, the classical Hamiltonian at any point is known (and indeed the QED Lagrangian is known at every point in the process, but that's really not necessary for such low energies). Is it really true that we cannot model this sufficiently to predict the steps of the reaction in advance and find the cross section for each step?
Your point about the multiple orientations making a difference is a good one if you want to treat it like a series of colliding particles, I'm sure for things like proteins, it's vastly more complicated than my naive approach would be able to deal with!
(Also, an exact solution does exist too for the classical quantum Helium atom iirc, thanks to Richard Feynman)
You might want to read up on computational chemistry. Although there are ab initio (from first principles) techniques, they use various approximations and are very compute intensive.
https://en.wikipedia.org/wiki/Computational_chemistry
| « Grignard Reaction | Electrolysis » |
| Bookmarks |
Bookmarks |
