Silicon nitride based integrated photonic circuit to control a cold-atom source

TL;DR

This work demonstrates a silicon nitride photonic integrated circuit (PIC) that integrates cooling, pumping, and imaging functions for a rubidium-87 MOT within a compact two-chip platform (~2 cm by 2 cm). By employing piezoelectric PbZr$_x$Ti$_{1-x}$O$_3$ (PZT) actuators on the PIC, the authors achieve fast, high-extinction switching of MOT beams with extinctions up to ~50 dB and switching times around $\sim$1 $\mu$s, enabling dynamic beam control for time-of-flight measurements. The paper presents an extended Mach-Zehnder architecture (TBS-ext) that tolerates fabrication errors, a detailed comparison of PZT and thermal actuators, and the integration of cooling, pumping, and imaging into a compact, fiber-based system. They report a 3D MOT containing $7 \times 10^7$ atoms at a temperature of $\approx 270\ \mu$K, illustrating the PIC’s effectiveness for miniaturized cold-atom sensors with potential applications in inertial navigation and on-chip atomic clocks. The approach paves the way for fully on-chip, fiber-integrated optical control of atomic clouds with significant reductions in volume and complexity, while maintaining high performance through fast PZT-based switching and robust extinction control.

Abstract

We have developed a silicon nitride based photonic integrated circuit (PIC) that is responsible for the cooling, pumping and imaging of cold rubidium 87 atoms. The photonic integrated circuit consists of two chips placed next to each other and has a total area of 2x2~cm$^2$. This greatly minimizes the area needed while still having all the optical control functions to create, control and measure a magneto-optical trap (MOT). The piezo electric material Lead Zirconate Titanate (PZT) on the PIC is employed for phase shifting a Mach-Zehnder type configuration where extinction ratios up to 50 dB and switching speeds of 1 MHz are achieved. For the first time a two and three dimensional rubidium 87 MOT is realized using an active PIC. For the three-dimensional MOT, we measure $7\cdot 10^7$ atoms with a temperature of 270~$μ$K.

Figures (6)

• Figure 1: Layout of the experimental setup consisting of: a functional diagram of the fiber-based laser system for the cooler and pumper beams, a PIC for distribution of the laser light and the experimental setup for creating the 2D and 3D MOT. The photograph of the 3DMOT cell shows the atomic cloud in the center together with a sketch of the laser beams.
• Figure 2: (a) Functional diagram of one of the photonic chips showing the paths of the cooler (red) and pumper (blue) laser beam in the circuit. (b) Photograph of the PIC consisting of two identical chips. (c) Typical individual optical response of four PZT elements and one heater. (d) Measured output power (with a S121C Thorlabs powerhead) as function of time while we switch on/off the voltage applied on the 2DX1 PZT element in order to switch on/off the light. The extinction ratio of 40-50 dB is limited by the stability of the PZT element. (e) 30 $\mu$s square pulse created using a PZT switch.
• Figure 3: Experimental results for the atom number and temperature of the 3DMOT. (a) atom number in the MOT as a function of the average cooler power (per beam) for the 2D and 3DMOT beams. For each measurement the loading time is 180 s, which is sufficiently long for reaching saturation of the trapped atom number. (b) Loading curve of the atom number in the 3DMOT as a function of time for an average 3DMOT cooler power (per beam) of 3.7 mW. (c) TOF measurement indicating a 3DMOT temperature between 270 $\mu$K and 1 mK. The lines are fits. (d) 2D images of the expansion of the 3DMOT during the TOF with obtained optical density (left) and their Gaussian fit (right).
• Figure S1: Evaluation of Eq. (\ref{['TBS-extension cos']}) as a function of the effective radius $r_{eff}$. The results are plotted as red dots. The blue dots indicate the maximum output intensity that is reachable while keeping full extinction of the light when only changing $\theta_{1}$. In this case the system acts as a single switch since only $\theta_{1}$ is changed to go from zero to the maximum output transmission, which is limited by the values of $k_{1}$, $k_{2}$, $k_{3}$.
• Figure S2: (a) Simulation of the normalized transmission of an extended TBS as a function of $\theta_{1}$ and $\theta_{2}$ for the case where $k_{1} = 0.6$, $k_{2} = 0.4$ and $k_{3} = 0.5$. In this simulation the light enters and exits from the top waveguide. The horizontal line helps to visualize that the maximum and minimum are not aligned along the $\theta_{1}$ axis and one needs to change both $\theta_{1}$ and $\theta_{2}$ to switch between the global minimum and the global maximum. (b) Transmission on a logscale as a function of $\theta_{1}$ for the cross section indicated by the red horizontal line in (a). (c) Zoom in around the section close to zero transmission of (b).