måndag 7 november 2011

The Essence of Planck's Radiation Law and Its Proof

Planck trying to convince a disbelieving Bohr that his proof of Planck's Law is correct. On the blackboard: Maxwell's equations.

The derivation of Planck's Law (of blackbody radiation) given by Planck is based on particle statistics borrowed from thermodynamics with little connection to the real physics of radiation as interaction of matter and electromagnetic wave motion.

Planck was forced into particle statistics because a radiation law based on wave motion seemed to lead a Rayleigh-Jeans law with an "ultraviolet catastrophe" with energy exploding to infinity like $\nu^2$ with frequency $\nu$ without upper bound. By a statistical assumption that highly energetic waves are rare Planck gave physics a way to avoid the catastrophe which earned almost infinite fame and a Nobel Prize.

But Planck left a question without answer: Is it possible to derive a correct radiation law without resort to statistics, using instead classical electrodynamics described by Maxwell's wave equations?

In Mathematical Physics of Blackbody Radiation and Computational Blackbody Radiation I show that this is indeed possible, by replacing statistics by a much more basic physical assumption of finite precision computation. Let me here presents the essence of my argument:

I consider mathematical wave model of the form:
  • $U_{tt} - U_{xx} - \gamma U_{ttt} - \delta^2U_{xxt} = f$
where the subindices indicate differentiation with respect to space $x$ and time $t$, and
  1. $U_{tt} - U_{xx}$: material force from vibrating string with U displacement
  2. $- \gamma U_{ttt}$ is Abraham-Lorentz (radiation reaction) force
  3. $- \delta^2U_{xxt}$ is a viscous force
  4. $f$ is exterior forcing,
and $\gamma$ and $\delta^2$ are (small) positive coefficients subject to a frequency dependent switch from outgoing radiation with to absorption/internal heating defined below.

The finite precision computation is represented by the $\delta^2$ viscosity which sets a smallest scale in space of size $\delta$ by damping frequencies higher than $\frac{1}{\delta}$.

The coefficients $\gamma$ and $\delta$ are chosen so that only one is positive for a certain frequency $\nu$ in a spectral decomposition, with a switch from $\gamma >0$ to $\delta >0$ at $\nu = \frac{T}{h}$, where $T$ is temperature and $h$ is a fixed precision parameter representing atomic dimension in the string of atoms modeled by the wave equation.

The switch thus defines a temperature dependent cut-off frequency $\frac{T}{h}$ with the effect that only frequencies below the cut-off are emitted as radiation while frequencies above cut-off are stored as internal heat energy of the string.

The essence of the proof is the following energy balance established by a spectral analysis assuming periodicity in space and time:
  • $\int \gamma U_{tt}^2dxdt + \int \delta^2U_{xt}^2dxdt = \epsilon \int f^2 dxdt$
with $\epsilon\approx 1$ is a coefficient of emissivity (= absorptivity), which we write in condensed form as
  • $R + A = F$, or in a spectral decomposition $R_\nu + A_\nu = F_\nu$, or in words
  • Radiation + Absorption = Forcing,
where $R = \int \gamma U_{tt}^2dxdt$, $A=\int \delta^2U_{xt}^2dxdt$ and $F=\epsilon \int f^2 dxdt$.

As just said, the switch effectively means that
  • $R_\nu = F_\nu$ for $\nu <\frac{T}{h}$
  • $A_\nu = F_\nu$ for $\nu >\frac{T}{h}$,
which expresses that the forcing $F_\nu$ is remitted as outgoing radiation for $\nu<\frac{T}{h}$, and is absorbed and stored as internal heat energy for $\nu > \frac{T}{h}$.

The coefficient $\epsilon$ represents emissivity = absorptivity, in accordance with Kirchhoff's Law. An essential aspect is that $\epsilon$ is independent of $\gamma$ and $\delta$, which expresses that radiation (to up to the constant $\epsilon$) is independent of the nature of the radiating/absorbing body.

Planck's law in the form $R+A = F$ with only one of $R$ and $A$ non-zero for each frequency
is a variant of Planck's classical law with a sharp switch instead of a continuous transition from radiation to absorption/heating.

Planck's law in the form $R + A = F$ thus contains the classical law for radiation as $R=F$ but also the further information that absorption into internal heat energy $A=R$ only occurs for frequencies above cut-off.

The proof that $R + A = F$ is non-trival and expresses a fundamental aspect of near-resonance
in systems of oscillators with small damping. The proof shows Planck's Law $R_\nu\sim \gamma T\nu^2$ for $\nu <\frac{T}{h}$ and Kirchhoff's Law, without resort to statistics.

The effects of Planck's statistics in modern physics are subject to an investigation in Dr Faustus of Modern Physics: Is it true that Planck's technique for avoiding the ultraviolet catastrophe led to a much bigger catastrophe of abandoning rationality in physics?



17 kommentarer:

  1. "Is it true that Planck's technique for avoiding the ultraviolet catastrophe led to a much bigger catastrophe of abandoning rationality in physics?"

    No.

    SvaraRadera
  2. To be honest, I don't really see how rationality is lost with the introduction of quanta.

    If nature have behaviours that can be modelled with quantization rationality would be abandoned by not considering it.

    Sincerely,
    Dol

    SvaraRadera
  3. To give up reality/determinism for unreality/indeterminism/statistics is
    catastrophical from scientific point of view.

    SvaraRadera
  4. Claes wrote:
    To give up reality/determinism for unreality/indeterminism/statistics is
    catastrophical from scientific point of view.


    Maybe one could argue that it is catastrophical from philosofical point of view, if one is bent that way. But come on, realism must deal with the real world. If this means indeterminism on a microskopic level so what? (släpp sargen och kom in i matchen på ren svenska).

    I don't think that most macroskopic events is computational anyway, so what is determinism?

    Sincerely,
    Dol

    SvaraRadera
  5. Determinism is that events are predictable, more or less. Science without
    prediction is empty.

    SvaraRadera
  6. Claes wrote:
    Determinism is that events are predictable, more or less. Science without
    prediction is empty.


    Have you read any books written by Nassim Taleb?

    Sincerely,
    Dol

    SvaraRadera
  7. Claes, do you know how Boltzman theoretically interpreted his law? (16 years before Planck interpreted his)

    SvaraRadera
  8. And how did he come around the ultra-violet catastrophy?

    SvaraRadera
  9. I used a wrong word: interprete shall be derive (deduce)

    SvaraRadera
  10. Claes wrote:
    Determinism is that events are predictable, more or less. Science without
    prediction is empty.


    Since quantum mechanics deals with such nice statistics it is to be considered an known unknow. One can argue that this is kind of deterministic in som broader sence.

    It seem that you simply are not confident thinking about the world in that sence, but that has no forcing on how the world behaves.

    I mean the world could evolve in the fashion of a quantum computation for all we know.

    Sincerely,
    Dol

    SvaraRadera
  11. Planck started out as a disciple of Boltzmann and and carried Bs particle statistics of thermodynamics into electromagnetics. P was well aware of the
    dark side of this venture.

    SvaraRadera
  12. I have a question concerning your model.

    How do you determine the value of $\h$? It carries dimension $Ks$, how do you connect this with atomic dimensions?

    And the magnitude of $\delta$ and $\gamma$. Can you estimate them from atomic scales or do you have to put them in manually to get the right cut off?

    Sincerely,
    Dol

    SvaraRadera
  13. How do you determine the value of ?

    That should be:
    How do you determine the value of h?

    SvaraRadera
  14. By ab initio quantum mechanics computation or by fitting to experiment.

    SvaraRadera
  15. Claes wrote:
    By ab initio quantum mechanics computation

    Can you show us the calculation?

    Sincerely,
    Dol

    SvaraRadera
  16. One more question.

    Is your ab initio approach in first or second quantization?

    Sincerely,
    Dol

    SvaraRadera
  17. Referring to the gaseous source of thermal radiation, i) the free traveling molecules have the sole kinetic energy K=½mu²; ii) during a collision the initial energy Ki is converted to potential energy P; iii) if P < P* (potential energy of the first excited state i.e. the molecule activation energy) then the molecule gets back all its initial Ki and we have Kf = Ki; iiii) if Ki > P* meanwhile the inversion of the motion the molecule also emits one or more photons, that’s Ki = Kf + n(hv).
    Then the gas does emit according to Maxwell-Boltzmann distribution and so it must have both a cut in and a cut off frequency.
    Thus, the statistical mechanics is more than sufficient to prove the BB radiation behavior.

    Michele

    SvaraRadera