Erratum: “Sub-megahertz linewidth single photon source” [APL Photonics 1, 096101 (2016)]

Deeper analysis of the literature^1,2 and discussions with our theory collaborators brought to light an error in our former calculations of the single photon linewidth. The stated $γs=666±16$ kHz and $γi=667±15$ kHz are the decay rates derived directly from the fit. However, the experimental data are a convolution of two comb-like spectra, and therefore the mode linewidth is actually smaller than the observed damping rate. Given $γ=γs≈γi$⁠, the FWHM linewidth Δν_s,i simplifies to¹ 

differing by a factor of 0.64. This leads to a new single photon linewidth of $429±10$ kHz.

Additionally, the experimental data for the $gs,s(2)(0)$ are missing a factor of 2 to account for possible bunching.^2–4 This results in a minimal $gs,s(2)(0)=0.032±0.003$⁠, with an observed value of $gs,s(2)(0)<0.5$ up to a heralding rate of $∼70$ kHz. We have also adjusted Fig. 4, including the doubled $gs,s(2)(0)$ values.⁵ 

All conclusions from our original article remain the same. The reduced linewidth is still, to our knowledge, the narrowest measured from SPDC to date. Although the result for $gs,s(2)(0)$ is increased by a factor of 2, the values remain far below 0.5 for low idler rates and therefore keep confirming the single photon nature of the source.