6

If the energy of a photon

$E_{p}=hv$

And the energy of an electromagnetic wave is

$E_{w}\propto \hat{\mathbf B}^2$

What is the relationship between $E_{w}$ and $E_{p}$?

2 Answers2

11

You only need to rewrite $\mathbf B$ and $\mathbf E$ in terms of field $A_{\mu}$ (here $\hbar = c = 1$), $$ \tag 1 \hat{\mathbf B} = [\nabla \times \hat{\mathbf A}], \quad \hat{\mathbf E} = -\frac{\partial \hat{\mathbf A}}{\partial t} - \nabla \hat{A}_{0}, $$ which is written as infinite "sum" of photons: $$ \tag 2 A_{\mu} = \sum_{\lambda} \int \frac{d^{3}\mathbf p}{\sqrt{(2 \pi )^{3}2E_{\mathbf p}}}e_{\mu}^{\lambda}(\mathbf p )\left( \hat{a}_{\lambda}(\mathbf p )e^{-ipx} + \hat{a}_{\lambda}^{\dagger}(\mathbf p ) e^{ipx}\right). $$ After that you can easily obtain the relation between energies of sets of photons and "real" EM field: $$ \tag 3 \hat{H} = \int \hat{T}_{00}d^{3}\mathbf r = \int \frac{1}{2}\left( \hat{\mathbf B}^{2} + \hat{\mathbf E}^{2}\right)d^{3}\mathbf r. $$

If you need I'll derive it.

Tedious derivation

For simplicity you need Coulomb gauge $A_{0} = 0, (\nabla \cdot \mathbf A) = 0$ (eq. $(3)$ already implies that), polarization sum rule and orthogonality relations for polarization vectors, $$ \sum_{\lambda}e_{i}^{\lambda}(\mathbf p)e_{j}^{\lambda}(\mathbf p) = \delta_{ij}, \quad (\mathbf e_{\lambda}(\mathbf p) \cdot \mathbf e_{\lambda'}(\mathbf p)) = \delta_{\lambda\lambda'}. $$ and commutation relations $$ [\hat{a}_{\lambda}(\mathbf p), \hat{a}^{\dagger}_{\lambda {'}}(\mathbf k)] = \delta_{\lambda \lambda {'}}\delta (\mathbf p - \mathbf k), \quad [\hat{a}_{\lambda}(\mathbf p ), \hat{a}_{\lambda{'}}(\mathbf k)] = 0. $$ First let's calculate $(1)$ by using $(2)$ ($E_{\mathbf p} = p_{0}$): $$ \hat{\mathbf E}(x) = -\partial_{0}\hat{\mathbf A}(x) = i\sum_{\lambda}\int \frac{d^{3}\mathbf p}{\sqrt{2 (2 \pi )^{3}}}\mathbf e_{\lambda}(\mathbf p) \sqrt{E}_{\mathbf p}\left( \hat{a}_{\lambda}(\mathbf p )e^{-ipx} - \hat{a}^{\dagger}_{\lambda}(\mathbf p )e^{ipx}\right), $$ $$ \hat{\mathbf B}(x) = [\nabla \times \hat{\mathbf A}] = i\sum_{\lambda}\int \frac{d^{3}\mathbf p}{\sqrt{(2 \pi )^{3}2E_{\mathbf p}}}[\mathbf p \times \mathbf e_{\lambda}(\mathbf p)]\left( \hat{a}_{\lambda}(\mathbf p )e^{-ipx} - \hat{a}^{\dagger}_{\lambda}(\mathbf p )e^{ipx}\right). $$ Then $$ \int d^{3}\mathbf r \hat{\mathbf E}^{2} = -\sum_{\lambda , \lambda '}\int \frac{d^{3}\mathbf r d^{3}\mathbf p d^{3}\mathbf k}{(2 \pi )^{3}2}\sqrt{E_{\mathbf p}E_{\mathbf k}}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf k )) \times $$ $$ \times \left( \hat{a}_{\lambda}(\mathbf p )e^{-ipx} - \hat{a}^{\dagger}_{\lambda}(\mathbf p )e^{ipx}\right) \left( \hat{a}_{\lambda {'}}(\mathbf k )e^{-ikx} - \hat{a}^{\dagger}_{\lambda {'}}(\mathbf k )e^{ikx}\right) = \left| \frac{1}{(2\pi )^{3}}\int d^{inx}d^{3}\mathbf r = \delta (\mathbf n) e^{in_{0}x_{0}}\right| = $$ $$ = -\sum_{\lambda , \lambda {'}}\frac{1}{2}\int d^{3}\mathbf p d^{3}\mathbf k \sqrt{E_{\mathbf p}E_{\mathbf k}}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf k )) \times $$ $$ \times \delta (\mathbf p + \mathbf k)\left( e^{ix_{0}(k_{0} + p_{0})}\hat{a}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(\mathbf k) + e^{-ix_{0}(k_{0} + p_{0})}\hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(\mathbf k) \right) + $$ $$ +\sum_{\lambda , \lambda {'}}\frac{1}{2}\int d^{3}\mathbf p d^{3}\mathbf k \sqrt{E_{\mathbf p}E_{\mathbf k}}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf k )) \times $$ $$ \times \delta (\mathbf p - \mathbf k) \left( e^{ix_{0}(k_{0} - p_{0})}\hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(\mathbf k) + e^{-ix_{0}(k_{0} - p_{0})}\hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(\mathbf k)\right) = $$ $$ = -\frac{1}{2}\sum_{\lambda , \lambda {'}}\int d^{3}\mathbf p E_{\mathbf p}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(-\mathbf p )) \left( e^{2ip_{0}x_{0}}\hat{a}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(-\mathbf p) + e^{-2ix_{0}p_{0}}\hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(-\mathbf p) \right) + $$ $$ \tag 4 + \frac{1}{2}\sum_{\lambda , \lambda {'}}\int d^{3}\mathbf p E_{\mathbf p}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf p )) \left( \hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(\mathbf p) + \hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(\mathbf p) \right). $$ The same thing with $\int d^{3}\mathbf r\hat{\mathbf B}^{2}$ by using relation $$ ([\mathbf p \times \mathbf e_{\lambda}(\mathbf p)] \cdot [\mathbf k \times \mathbf e_{\lambda {'}}(\mathbf k)]) = (\mathbf p \cdot \mathbf k)(\mathbf e_{\lambda}(\mathbf p) \cdot \mathbf e_{\lambda {'}}(\mathbf k)) - (\mathbf p \cdot \mathbf e_{\lambda}(\mathbf p)) (\mathbf k \cdot \mathbf e_{\lambda {'}}(\mathbf k)) = $$ $$ = (\mathbf p \cdot \mathbf k)(\mathbf e_{\lambda}(\mathbf p) \cdot \mathbf e_{\lambda {'}}(\mathbf k)) $$ can give $$ \int d^{3}\mathbf r \hat{\mathbf B}^{2} = $$ $$ = \frac{1}{2}\sum_{\lambda , \lambda {'}}\int d^{3}\mathbf p E_{p}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(-\mathbf p )) \left( e^{2ip_{0}x_{0}}\hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(-\mathbf p) + e^{-2ix_{0}p_{0}}\hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(-\mathbf p) \right) $$ $$ \tag 5 + \frac{1}{2}\sum_{\lambda , \lambda {'}}\int d^{3}\mathbf p E_{p}(\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf p )) \left( \hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(\mathbf p) + \hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(\mathbf p) \right). $$ So after summation of $(4), (5)$ you will get that $$ \hat{H} = \frac{1}{2}\sum_{\lambda , \lambda {'}}\int d^{3}\mathbf p (\mathbf {e}_{\lambda }(\mathbf p ) \cdot \mathbf {e}_{\lambda {'}}(\mathbf p )) E_{\mathbf p} \left(\hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda {'}}(\mathbf p) + \hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda {'}}(\mathbf p) \right) = $$ $$ \tag 6 \frac{1}{2}\sum_{\lambda}\int d^{3}\mathbf p E_{\mathbf p}\left(\hat{a}_{\lambda}(\mathbf p) \hat{a}^{\dagger}_{\lambda}(\mathbf p) + \hat{a}^{\dagger}_{\lambda}(\mathbf p) \hat{a}_{\lambda}(\mathbf p) \right) = \sum_{\lambda}\int d^{3}\mathbf p E_{\mathbf p}\left( \hat{a}_{\lambda}^{\dagger}(\mathbf p)\hat{a}_{\lambda}(\mathbf p) + \delta (0)\right). $$ Eq. 6 implies "representation" of the energy of EM field as sum of energies of photons ($E_{\mathbf p} = \omega_{\mathbf p}$), because $\int d^{3}\mathbf p \hat{a}_{\lambda}^{\dagger}(\mathbf p)\hat{a}_{\lambda}(\mathbf p)$ refers to the particles number operator.

  • 2
    Could you get it? – J.D'Alembert Sep 14 '14 at 14:47
  • This might be more useful to explain some of the terms here and maybe even calculate it! – Kyle Kanos Sep 14 '14 at 14:47
  • Yes, could you explain some of the terms and more importantly you state that $$ E = \int T_{00}d^{3}\mathbf r = \int \frac{1}{8 \pi}\left( \hat{\mathbf B}^{2} + \hat{\mathbf E}^{2}\right)d^{3}\mathbf r. $$ but what does $E$ denote? I'm looking for $E_{p}=f(E_{w})$ – J.D'Alembert Sep 14 '14 at 14:53
  • @J.D'Alembert : $E$ here means the full energy of EM field which is given from lagrangian formalism. Here $T_{00}$ is equivalent to hamiltonian density. $\hat{\mathbf E}, \hat{\mathbf B} $ are respectively the electric field strength and the magnetic field induction. – Andrew McAddams Sep 14 '14 at 14:58
  • @AndrewMcAddams Ok, so the question is now, how does one relate the energy of sets of photons to $E$? – J.D'Alembert Sep 14 '14 at 15:02
  • @J.D'Alembert : I'll write some calculations and thinking in a few time into the answer. – Andrew McAddams Sep 14 '14 at 15:04
  • I think one problem that you are bound to encounter is - how do you choose the limits of integration. OR rather, since you know the expression for energy density and are integrating over it, how big is the spatial extent of one photon? This question is impossible to answer IMO - this is a comparison b/w two different pictures, so I'm not sure if it will work out well. – 299792458 Sep 14 '14 at 15:18
  • @New_new_newbie : classical fields as quantum ones are localized on infinity. – Andrew McAddams Sep 14 '14 at 15:22
  • @AndrewMcAddams May I ask what $\mathbf r$ is? – J.D'Alembert Sep 14 '14 at 15:28
  • @J.D'Alembert : $\int d^{3}\mathbf r = \int dV$ here denotes integration over field lozalization. Since we discuss free case (EM field in vacuum) field localization is infinity. – Andrew McAddams Sep 14 '14 at 15:36
  • 2
    @J.D'Alembert : I afraid that in this cumbersome and non-interested math is possible to sink, but I've finished to derive the relation. – Andrew McAddams Sep 14 '14 at 16:56
  • 1
    Bravo! Nice answer, much more specific than I would have expected. +1 – Danu Sep 14 '14 at 17:22
  • 1
    @AndrewMcAddams Sorry that it took me this long to accept it my internet was disconnected and thank you for all the work that you put into your answer. – J.D'Alembert Sep 14 '14 at 17:59
  • @AndrewMcAddams Could you explain the notation used in $$\sum_{\lambda}\int d^{3}\mathbf p E_{\mathbf p}\left( \hat{a}{\lambda}^{\dagger}(\mathbf p)\hat{a}{\lambda}(\mathbf p) + \delta (0)\right)$$? – J.D'Alembert Sep 14 '14 at 18:04
  • @J.D'Alembert : here $\lambda$ means polarization (photon has 2 independent polarizations), $E_{\mathbf p} = \omega_{\mathbf p}$ denotes energy of photon with momentum $\mathbf p$, $\delta (0)$ arises because of I have made intercharge $$ \hat{a}{\lambda}(\mathbf p)\hat{a}{\lambda}^{\dagger}(\mathbf p) = \delta (\mathbf p - \mathbf p) + \hat{a}^{\dagger}{\lambda}(\mathbf p)\hat{a}{\lambda}(\mathbf p). $$ Usually we neglect by this summand because it is interpreted as vacuum energy and we can shift the energy level by this constant. – Andrew McAddams Sep 14 '14 at 18:07
  • @J.D'Alembert related to this last comment by Andrew: normal ordering – Danu Sep 14 '14 at 18:12
  • @AndrewMcAddams if $\delta (0)$ is the dirac delta function wouldn't it mean the solution is infinite as it approaches infinity? – J.D'Alembert Sep 14 '14 at 20:22
  • @J.D'Alembert : In a few words an explanation is following. QFT is (in some sense) ill, because it has singularities, which are hidden in the commutators. At the same time it means that we can modify its primary objects (lagrangian, for example) by some ways with only one requirement: these changes must correspond to physics, because this infinity is the consequence of only mathematical description. We see that infinity in expression $(6)$ is caused only by commutators. So we may throw this constant away. – Andrew McAddams Sep 14 '14 at 20:30
  • @AndrewMcAddams Does a decrease in the energy of the electromagnetic wave mean a decrease in the energy of the photon or the number of photons? – J.D'Alembert Sep 14 '14 at 20:32
  • @J.D'Alembert : it seems yes, because energy is positive definite quantity. – Andrew McAddams Sep 14 '14 at 20:40
  • @AndrewMcAddams Last questions, is it possible to rearrange equation (6) for $E_{\mathbf p}$ and if so how? – J.D'Alembert Sep 14 '14 at 20:43
  • @J.D'Alembert : what did you mean? – Andrew McAddams Sep 14 '14 at 20:46
  • Could you solve $$\hat{H}=\sum_{\lambda}\int d^{3}\mathbf p E_{\mathbf p}\left( \hat{a}{\lambda}^{\dagger}(\mathbf p)\hat{a}{\lambda}(\mathbf p) + \delta (0)\right)$$ for $E_{\mathbf p}$ (ignoring the constant)? – J.D'Alembert Sep 14 '14 at 20:49
  • @J.D'Alembert : in some sense you can solve this equation for $E_{\mathbf p}$ by acting of $\hat{H}$ on some one-particle state $| \mathbf p , \lambda\rangle$: then $$ \hat{H}| \mathbf p , \lambda\rangle = E_{p} | \mathbf p , \lambda \rangle . $$ – Andrew McAddams Sep 15 '14 at 18:17
  • @AndrewMcAddams can one simplify this? $$\sum_{\lambda} \hat{a}{\lambda}^{\dagger}(\mathbf p)\hat{a}{\lambda}(\mathbf p)$$ – J.D'Alembert Sep 15 '14 at 21:12
  • @J.D'Alembert : excuse me for the long periods between the answers. You can rewrite $\hat{a}^{\dagger}{\lambda}(\mathbf p)\hat{a}{\lambda}(\mathbf p)$ through fields $\hat{\mathbf A}$ and then to get expression similar to $(3)$. But creation/destruction operators are in some sense the most elementary quantities. – Andrew McAddams Sep 15 '14 at 21:17
  • @AndrewMcAddams I have read online what they do but could you please explain their purpose? – J.D'Alembert Sep 15 '14 at 21:19
  • @J.D'Alembert : I'll answer in a few hours, if you please. – Andrew McAddams Sep 15 '14 at 21:20
  • @AndrewMcAddams No problem, but when you do can you also explain what $$\sum_{\lambda} \hat{a}{\lambda}^{\dagger}(\mathbf p)\hat{a}{\lambda}(\mathbf p)$$ would mean? – J.D'Alembert Sep 16 '14 at 15:36
  • @J.D'Alembert : I have answered to the chat. – Andrew McAddams Sep 17 '14 at 21:28
  • @AndrewMcAddams Ok, so have I, for future refference lets continue to do so – J.D'Alembert Sep 17 '14 at 21:41
5

An experimentalists answer:

If you divide the energy of the electromagnetic wave by hv you will have the number of photons that are building up the electromagnetic wave.

anna v
  • 233,453
  • Isn't a macroscopic electromagnetic wave something more similar to a coherent state, rather than a Fock state? A coherent state has uncertain amount of photons. What does the number (Ew / hv) actually represent? – mpv Sep 14 '14 at 18:17
  • @mpv conservation of energy? – anna v Sep 14 '14 at 19:10
  • @mpv: the expectation value for the number of photons, I think. – Harry Johnston Sep 15 '14 at 04:21
  • @HarryJohnston from the analysis by Andrew I would think it would be the expectation of the average number of photons. But conservation of energy is an exact relation, so I expect it will be the number of photons for the energy carried by the wave. Any uncertainty will come from the classical measurement of energy, imo, not from h*nu. – anna v Sep 15 '14 at 05:48
  • I think what happens is that if you measure the number of photons, that makes the frequency of the photons uncertain, in just the same way that measuring the position of a quantum particle makes its momentum uncertain. So if you do make this measurement the answer might be that there are fewer photons than you expected, but since the frequency distribution has changed you still have the same total energy. – Harry Johnston Sep 15 '14 at 21:32
  • @HarryJohnston the Heisenberg uncertainty works for individual photons, and the average energy for individual photons will have a small spread with respect to macroscopic values.The average photon should still have nu distinctive otherwise the two slit single photon would not show interference. So within this uncertainty the count goes. – anna v Sep 16 '14 at 03:17
  • I don't think that picture is accurate anymore when you're dealing with QFT rather than plain QM. In QFT the photon count is quantized, and IIRC it doesn't commute with measurements of the electromagnetic field strength; as a consequence, the "energy of the individual photons" isn't really a well-defined concept if the system is in a coherent state. The fact that photons don't interact usually makes it OK to ignore this stuff from an experimental perspective, e.g., lasers and single photons produce similar interference patterns. But it's been a long time since I studied this stuff. – Harry Johnston Sep 16 '14 at 03:48