Bye-bye MEPP

Recently I have discovered that biologists like MEPP (Maximum Entropy Production Principle): see Physical foundations of evolutionary theory and also the last slides in Does Entropy Play a Role in Biology?

I have not found in the papers from biologists simple applications of MEPP. As I personally like to start learning a theory with simple examples, I have decided to look at what chemists say about MEPP. Here it is:

John Ross, Alexandru D. Corlan, and Stefan C. Müller, Proposed Principles of Maximum Local Entropy Production, J. Phys. Chem. B, 2012, 116 (27), pp 7858–7865, http://dx.doi.org/10.1021/jp302088y

Below there is an abstract and a conclusion from the paper and after that my emails related to MEPP to biosemiotics list.

Proposed Principles of Maximum Local Entropy Production

Abstract: ‘Articles have appeared that rely on the application of some form of “maximum local entropy production principle” (MEPP). This is usually an optimization principle that is supposed to compensate for the lack of structural information and measurements about complex systems, even systems as complex and as little characterized as the whole biosphere or the atmosphere of the Earth or even of less known bodies in the solar system. We select a number of claims from a few well-known papers that advocate this principle and we show that they are in error with the help of simple examples of well-known chemical and physical systems. These erroneous interpretations can be attributed to ignoring well-established and verified theoretical results such as (1) entropy does not necessarily increase in nonisolated systems, such as “local” subsystems; (2) macroscopic systems, as described by classical physics, are in general intrinsically deterministic—there are no “choices” in their evolution to be selected by using supplementary principles; (3) macroscopic deterministic systems are predictable to the extent to which their state and structure is sufficiently well-known; usually they are not sufficiently known, and probabilistic methods need to be employed for their prediction; and (4) there is no causal relationship between the thermodynamic constraints and the kinetics of reaction systems. In conclusion, any predictions based on MEPP-like principles should not be considered scientifically founded.’

‘While we only examined some of the papers on the maximum entropy production principle, we found many of the statements mentioned here, as formulated, to be obviously inconsistent with the current laws [27,28] and assumptions of most physics and chemistry. The current laws provide a very effective scheme to predict quantatively the behavior of a wide variety of systems. Some supporters of the MEPP-like theories do not even consider experimental falsifiability of a hypothesis to be a requirement for a theory, although some [3] do admit, simultaneously, that the lack of this quality puts a hypothesis outside the field of science.

Thus, predictions based on MEPP-like principles should not be considered scientifically founded.’

My messages

19.04.2013 20:02

First, the main goal of thermodynamics is rather prediction of equilibria. There was an endless discussion on whether thermodynamics should be thermostatics or thermodynamics. Yet, we should start with the goal. The complete description of kinetics (some evolution equation), in my view, does not belong thermodynamics. The goal was all the time rather to predict what will happen with the system without integrating it in time. At least, this is what I understand by thermodynamics.

Hence, whether a stone will be broken up into pieces or not, in my view, is irrelevant to definition of a state. A state in thermodynamics is described by introduction of phases and then we should say what phases do we have and at what conditions. I would say that this should work in your example.

The only potential problems would be when the stone will be broken to many small pieces is the treatment of interfaces. A conventional way with the excess Gibbs energy is good for the case when there is a lot of bulk phases and a little bit of interfaces between them. When one has a lot of interfaces and a little bit of bulk phases, the Gibbs treatment of interfaces may not work any more.

20.04.2013 09:47

You can search for CALPHAD. There is not that many people working along this way, but you will find them all over the world. Hence, Russia as such is not relevant.

The difference is in the goal. I mean thermodynamics as a practical tool for solving practical problems (this is exactly what happens in CALPHAD). You seem to mean by thermodynamics some theorizing when it is impossible to use equations to solve practical problems.

Hence the difference in ideology rather could be formulated as

  • Equations that one can use to solve practical problems

vs.

  • Equations for aesthetic pleasure.

For example, you have sent a paper The 2nd Law of Thermodynamics Delineates Dispersal of Energy. It starts with

The 2nd law of thermodynamics as an equation of motion is derived from statistical physics of open systems.”

This shows the main problem. You try to replace accepted terminology (The Second Law) that has a definite meaning (take for example Atkins’ Physical Chemistry) with completely different meaning. In my view, this is a bad ideology.

20.04.2013 12:18

I do not see a problem to introduce the phase change into thermodynamics. This has been already considered for long time in modeling of phase diagrams. If you see search for CALPHAD, you will find many nice pictures (see also software Thermocalc and Factsage).

The problem would be to define what a phase is in the case for example a cell. There is a lot of interfaces as well as polymers there. Hence the definition of a phase (and thereafter a state) is a tricky problem. I should say that I am not an expert in polymer and interface thermodynamics, so I do not know the state of the art in this area.

Interestingly enough, in papers of biologists that come to my eyes, biologists to do not consider that real problem but rather they prefer to talk about THE ENTROPY in quite general terms.

20.04.2013 18:05

I believe that one finds a lot of well studied examples in chemical kinetics. Could you please point to a paper on MEPP that has worked out an example with simple chemical reactions and have demonstrated that MEPP is working there?

From what I have seen, there are still two separate disciplines: chemical thermodynamics and chemical kinetics. From what you write it seems to follow that non-equilibrium thermodynamics has already replaced chemical kinetics. I personally doubt that this is the case outside a small community working on non-equilibrium thermodynamics.

Is it again cultural differences in different scientific communities?

25.04.2013 21:45

I have worked the paper out. It is well written and I find arguments in the paper quite convincing. The authors presented several numerical examples where MEPP is not held. In my view, this means that MEPP is wrong and I am personally not going to pay attention to it.


Comments are closed.