What drives osmosis? (Statistics or Dynamics?)

[Swamp Thing]

This letter in PhysRevLett describes molecular dynamics simulations to investigate the underlying factors that drive osmosis. https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.97.024501

Copy on Internet Archive : https://web.archive.org/web/2019030...c9c2/427b191488a92188ce6b61edbdde21238b23.pdf

They simulate the diffusion of water through a carbon membrane that separates water volumes with different concentrations of KCl. They show that the time averaged force on a water molecule near the mouth of a pore is non zero and directed out of the pore. However, the outward force is greater on the side with higher concentration of KCl. They identify this asymmetric force as the cause of osmosis.

On the other hand, these videos on YouTube claim that osmosis can occur as a purely statistical phenomenon, as a result of the system's drift towards higher entropy:

youtu.be/2OOvMiKCp8A?si=QwexuSufLiWXIHg4

youtu.be/VCXqELB3UPg?si=OyaktNhSrLR5K9LL

My question is, are the statistical explanations valid in some cases, e.g. in non-electrolytic solutes like sugar? Can there be situations where both processes contribute simultaneously?

 

[sophiecentaur]

Maths doesn’t make things happen. It describes what happens. ‘Drifting’ to a high entropy state can only happen due to interactions (dynamics etc.) There is no conflict.

 

[Andy Resnick]

This letter in PhysRevLett describes molecular dynamics simulations to investigate the underlying factors that drive osmosis. https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.97.024501
<snip>

I'm a little confused by your question- the PRL article seems to be a simulation of the Nernst-Planck equation, the first youtube video looked like a simulation of Donnan equilibrium (Gibbs-Donnan effect), and the second video was just confusing (to me).

What exactly are you asking about- a stochastic version of the Nernst-Planck equation? I believe that's the Poisson-Nenrst-Planck model: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3122111/#:~:text=In the PNP theory, the,in terms of ion concentration.

Donnan equilibrium is a balance between osmotic and electrostatic potentials., if that helps.

 

[Swamp Thing]

‘Drifting’ to a high entropy state can only happen due to interactions (dynamics etc.) There is no conflict.

Assuming that the simulation is working correctly, there does seem to be a conflict.

If I understand correctly, in the statistical picture, the thread of water within the pore is being pushed from both ends by water atoms bumping into the pore mouth. There are fewer bumps on the high-solute side, resulting in a net flow. The time-averaged force on an individual molecule near the mouth is thus directed into the pore.

OTOH, the PRL simulation shows an outward pull near the pore mouth due to the Lennard-Jones forces (ultimately electrostatic forces). In fact, they note that the thread of water is sometimes actually broken by the negative pressure (tension) and they look at the probability of this happening. (Similar to the negative pressure in the capillaries of trees and tall plants, which also sometimes breaks / cavitates?).

the PRL article seems to be a simulation of the Nernst-Planck equation, the first youtube video looked like a simulation of Donnan equilibrium (Gibbs-Donnan effect) ...

... What exactly are you asking about- a stochastic version of the Nernst-Planck equation? I believe that's the Poisson-Nenrst-Planck model:

Thanks, I will look up the terminology you mentioned and read the linked reference.

 

[sophiecentaur]

Assuming that the simulation is working correctly,

Exactly. Do you believe the results from a simulation or what you actually measure?

 

[Swamp Thing]

Exactly. Do you believe the results from a simulation or what you actually measure?

The conflict is not about the observed result. The simulation gives reasonable predictions for the direction and rate of flow and the equilibrium pressure.

The conflict is about the microscopic description. The statistical picture has the water being pushed from both sides, with the low-solute side winning out. The paper describes it as the water being pulled from both sides with the high solute side winning out.

I don't know if someone has done an experiment that can distinguish between a pulling mechanism and a pulling mechanism.

 

[sophiecentaur]

The conflict is not about the observed result.

The title of the thread is "What drives osmosis? (Statistics or dynamics?) " so should that not be what is being discussed? The merit of a simulation can only be how accurate it is in its predictions.A simulation is surely not 'what really happens'.

I don't know if someone has done an experiment that can distinguish between a pulling mechanism and a pulling mechanism.

Until then, the question is still open, I would have thought.