Purcell-enhanced solid-state laser cooling

TL;DR

The paper investigates how the Purcell effect can enhance solid-state laser cooling by increasing radiative decay rates and saturation intensity, enabling higher cooling power in optical refrigeration. It introduces a patterned slot-waveguide structure around a Yb-doped silica active layer, showing an average Purcell factor of about 18 and near-unity escape efficiency, leading to a maximum cooling-power density roughly 40 times larger than a bare layer under realistic parasitic absorption. The analysis combines theory with RCWA simulations to quantify the dependence on pump wavelength, intensity, cladding loss, and ion density, revealing that saturation-limited cooling and external quantum efficiency control the enhancements. The results indicate a practical design pathway for higher cooling power and potential applications across wavelengths and materials, including other rare-earth ions and higher-index claddings.

Abstract

We show that Purcell effect can lead to a substantial enhancement in the maximum cooling power for solid-state laser cooling. We numerically demonstrate such enhancement in a patterned slot-waveguide structure using ytterbium-doped silica as the active material. The enhancement arises primarily from the increase of saturation power density and the escape efficiency, and can persist in spite of the presence of parasitic absorption in the structure surrounding the active material. Our results point to a new opportunity in photonic structure design for optical refrigeration.

Figures (3)

• Figure 1: Enhanced laser cooling of a patterned slot waveguide nanostructure (a) as compared to a bare active layer (b). The active layer (ytterbium-doped silica with $10nm$ thickness) is surrounded by a patterned higher-index cladding layer (with refractive index 4 and thickness $86nm$). The surface pattern is a periodic square lattice of air holes with period $0.91\mu m$, diameter $0.39 \mu m$, and depth $13nm$ on each side of the structure. (c) Purcell factor $F$ defined as ratio of emitted power compared to a homogenous medium. (d) Absorption spectrum in both the active layer (black curve) and cladding layers (grey curve) normalized to the bare active slab absorption. All absorption coefficients are assumed to be equal in this plot.
• Figure 2: (a) Cooling power density $P_c$ spectrum as a function of pump wavelength $\lambda_p$ and intensity $I_p$ for the slot structure. (b,c) Maximum cooling power density $P_{c,max}$ as a function of the pump wavelength and temperature for the slot structure (blue curve) and for the bare active layer (orange curve). Inset shows saturation of $P_c$ as a function of the total absorbed power $P_{abs}$ at optimal pump wavelength ($\lambda_p \approx 1.036\mu m$ for the slot structure and $\lambda_p \approx 1.046 \mu m$ for the bare active layer). The cooling efficiency ($= P_c/P_{abs}$) is around 2% at low pump intensities. Only a small fraction ($\sim 10^{-4}$ at resonance for the slot structure) of incident power is absorbed.
• Figure 3: Maximum cooling power density (optimized over both pump intensity and pump wavelength) as a function of the cladding absorption $\alpha_{b,clad}$ (a) and rare-earth ion density $N_0$ (b) [orange curve in (b) is scaled $\times 5$ for clarity]. Blue curves correspond to the slot structure and orange curves correspond to the bare active layer.