Chiral light from an emitter coupled to an achiral particle via the Purcell effect

Abstract

We demonstrate that non-chiral nanoparticles can produce chiral light when point emitters are coupled to their surface plasmon modes (SPMs) under certain conditions. Chiral emission arises from asymmetrical plasmon mode propagation from the source combined with the spin-momentum locked nature of the SPMs. The Purcell regime of cavity quantum electrodynamics (QED) ensures that radiation from the coupled mode dominates over that from the emitter itself, giving rise to photons with a circularly polarized component -- i.e. chiral light. We experimentally demonstrate this effect using electron beam-induced cathode luminescence from a gold nanorod, coupling it evanescently to a nanofiber probe which also supports spin-momentum locked light. This converts the net spin of the emission into a net directionality of propagation in the fiber modes.

Figures (12)

• Figure 1: (a) Schematic illustration of the generation of chiral light from a point emitter coupled to a GNR. (b) Purcell enhancement as a function of wavelength $\lambda$ for a dipole emitter coupled to the GNR. (c) (Upper panel) real part of the $z$-component of the electric field emitted from the GNR. (Lower panel) degree of circular polarization $P_{\rm CP}$ of the light shown in the upper panel. (d) Chirality density $\chi$ normalized by intensity. (e) $y$- component of chirality flow $\Phi_y$ normalized by energy density. (f) Degree of circular polarization $P_{\rm CP}$.
• Figure 2: (a) Degree of circular polarization for a gold nanowire fundamental mode at 600 nm wavelength, in the absence of dissipation. Thick black lines show the surface of the wire. The right hand half of the figure is for a $+x$ propagating mode, and the left hand half is for a $-x$ propagating mode, as indicated by the arrows. (b-f) Numerically calculated $P_{\rm CP}$ of the electric field in a GNR excited by an $z$-polarized point dipole source placed at (b) $(x=0,z=0)$, (c) $(x=0,z=15\,\mathrm{nm})$, (d) $(x=-50 \,\mathrm{nm},z=0)$, (e) $(x=-50\,\mathrm{nm},z=15\,\mathrm{nm})$, (f) $(x=50\,\mathrm{nm},z=15\,\mathrm{nm})$, and (g) $(x=-85\,\mathrm{nm},z=-35\,\mathrm{nm})$, as shown by a black dot in each case.
• Figure 3: (a) Principle of the experiment. An electron beam is incident on the GNR, creating an effective dipole excitation at the position where the electron stops. CL from the so-excited GNR plasmon mode couples to the optical nanofiber fundamental mode with a directionality dependent on the circular polarization component of the induced dipole moment. (b) Simulated values of $P_{\rm CP}$ (black circles) and $D$ (red circles) as a function of emitter position $x$. (c) Dependence of $P_{\rm CP}$ on $D$ inside the GNR, taken from the data shown in (b). (d) Experimental setup inside the SEM. A GNR deposited on an optical nanofiber is excited by an electron beam (diameter $\sim 5$ nm). The resulting optical emission is coupled evanescently to the fiber, and sent via a fiber feedthrough to a detector. (e) SEM image of the GNR and nanofiber from secondary electron detection. (f) CL image from light detected at one end of the fiber using a single photon counting module (SPCM) (g) Spectrum of the bare fiber (blue line) and the GNR (red line), as indicated in the legend, measured by an optical multi-channel analyzer.
• Figure 4: Experimental measurements of $D$ for the GNR shown in Fig. \ref{['fig:3']}e. (a) Experimentally measured directionality calculated from the data shown in Fig. \ref{['fig:3']}d. Vertical white lines indicate the fiber edges. (b) Numerically calculated directionality. (c) Comparison of experimentally measured directionality integrated over the GNR width (black dots with gray error bars) with the numerically predicted value (red curve). (d-f) Similar measurements showing the effect of displacement of the GNR from the ONF center. (d) SEM image of the GNR sample used. (e) The measured directionality for the sample shown in (d). (f) The averaged directionality over the region indicated by the gray dashed lines in (e). In both cases the pale blue regions in (c) and (f) indicate the GNR region.
• Figure A1: Transmission and directionality for a $145 \,\mathrm{nm} \times 50 \,\mathrm{nm}$ GNR plasmon mode excited by a point dipole source located at the upper-right corner under different mesh densities. Levels 1–8 correspond to the default automatic mesh settings in FDTD (from coarsest to finest). Level 9 corresponds to mesh size 8 with an additional refined custom mesh of $2\,\mathrm{nm}$ applied around the GNR.