Fig. 2: Coherence time of resonantly scattered photons.

a Michelson interferometer experimental setup. Our fibre-based setup consists of a 50:50 beamsplitter (BSP) to separate QD emission into two arms, where one arm contains a fibre stretcher to record interference fringes and the other arm contains a coarse fibre delay. Faraday mirrors at the end of each arm ensure polarisation matching at the BSP, from where the interference signal is sent to a spectrometer for detection. Interference results were normalised to the visibility achieved by the narrowband cw laser. b Visibility of measured interference fringes as a function of the time delay τ under non-resonant (purple circles) and resonant (cyan circles) excitation. The solid cyan line represents a fit according to Eq. (2) in the main text, with the shaded area as the error bar of \({T}_{2}^{inel}\). The solid purple line represents a fit to the data as described in the main text. c Visibility of interference fringes under resonant excitation, for a range of power levels and for much longer time delays. Filled circles describe measured data points, whereas solid lines describe fits according to Eq. (2) in the main text. d \({T}_{2}^{inel}\) times extracted from the initial decays of the signals in (c) and calculated T₁ times for all powers. The deduced Fourier limit is indicated by the grey dashed line with the grey shaded area indicating the error bar. Error bars denote the 95% confidence bounds on fitting coefficients.