In spontaneous parametric down-conversion, a single photon interacts with a nonlinear medium to produce a pair of lower energy photons. It is observed that energy, momentum, and orbital-angular-momentum are conserved in the process. However, I have seen arguments invoking Furry's theorem (see https://www.maths.tcd.ie/~fionn/qft/furry.pdf) stating that it is not possible for a photon to get destroyed to produce two photons. What is the key insight that I am missing here?


The theorem you've linked to is true - in vacuum.

SPDC happens in a non-linear medium which breaks the symmetries that underlie the hypotheses of that theorem. As such, there is no contradiction between the two.

  • $\begingroup$ I suspect there is something more because the process involves a single free photon at initial time, a finite interaction for a finite time, and then two free photons. As the process is parametric, there is no net transfer of energy, momentum, or angular momentum to the nonlinear medium. So if I look at the system only at the initial and final times as a unitary scattering process, I am not sure how the nonlinearity alone can rescue the process from the verdict of Furry's theorem. $\endgroup$ – Girish Jan 24 '19 at 9:24
  • $\begingroup$ Moreover, the nonlinear medium is also electromagnetic in nature. So although the medium does not possess spatial inversion symmetry, it does possess charge conjugation symmetry. Please correct me if I am wrong. $\endgroup$ – Girish Jan 24 '19 at 9:28
  • $\begingroup$ Unless the medium is a positronium crystal, or something similar, then it certainly does not possess charge-conjugation symmetry. (Hint: if it has electrons but it doesn't have positrons, then it's $C$-asymmetric.) But that's not really necessary - the lack of inversion symmetry is sufficient. $\endgroup$ – Emilio Pisanty Jan 24 '19 at 11:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.