The cesium spontaneous polarization experiment requires a source of tunable laser light at the cesium D1 transition wavelength of 894 nm, in the near infrared. The laser must produce tens of milliwatts of power at a single frequency, with a continuous tuning range of several GHz desirable. The original version of the experiment used a commercial solid-state titanium:sapphire ring laser, pumped by an argon-ion laser, both supplied by the UW Tunable Laser Facility. This produced over 1 watt of power with a very large tuning range, but cost in excess of $160,000, was shared with other research groups, and had the potential for severe maintenance headaches. Thus it was decided to switch to an external cavity diode laser system, which could produce the needed power at a cost of under $10,000, in addition to being a much simpler system with a minimum of potential problems.
The 894 nm diode laser for the spontaneous polarization experiment, along with the electronics to control it, was designed and built by Steve Lamoreaux and myself beginning in January 1996. The laser and the controller box were completed by May 1996, after which the system underwent several weeks of testing and debugging before entering service on the experiment in June 1996. This photo shows the laser during debugging, housed in the Styrofoam insulated enclosure at left and connected to the laser controller box at right. The present system produces over 20 mW of single-mode power with a linewidth less than 3 MHz. A single-mode continuous tuning range of up to 25 GHz has been achieved by simultaneously varying the diode current and the grating angle using computer D/A control. Our mechanical design allows simple alignment and optimization of the cavity with very good passive stability. This system has been in operation for almost three years, including near continuous, stable operation over the past two years.
A paper describing our 894 nm diode laser system was recently published in the November 2000 Review of Scientific Instruments [Rev. Sci. Instrum. 71, 4029-4031 (2000)]. This article may be found online at their website. Copyright © 2000 American Institute of Physics. This article may be downloaded for personal use only. Any other use requires prior permission of the author and the American Institute of Physics.
You may also download the paper directly from my site (PDF format, 305 kB).
NOTE: We had originally submitted a longer paper with much additional background information and including plots demonstrating laser performance, but the editors required that the paper be shortened to emphasize only those aspects of our design which are novel and not already well-represented in the literature. The original longer version may still be of greater utility than the published version for those interested in constructing similar laser systems. You may also download this longer paper from my site (PDF format, 291 kB). Thumbnail versions of the 5 pages:
Figure 1: Simplified 3D view of the external cavity. The outer can and thermoelectric coolers are not shown.
Figure 2: Cutaway views of the external cavity and can from the bottom (a) and side (b) [dimensions in inches].
Figure 3: Cutaway view of the laser enclosure, showing the laser diode mounting block and end plate. Outer slip-cover and lower portion of isolation mount are not shown.
Figure 4: Continuous scan of 24 GHz across the cesium D1 hyperfine spectrum, with transmission through vapor cells at 20 degrees C with zero (upper curve) and 40 Torr (lower curve) N2 buffer gas. Intensity variation is due to varying diode current.
Figure 5: Transmission through a scanning confocal Fabry-Perot etalon with FSR = 300 MHz. Solid line is a full scan across 2 FSR (scale at bottom). Dashed line is a single peak magnified 10X (scale at top), with FWHM = 3 MHz.