Frequency stabilization of a continuous-wave Ti:sapphire laser

T. L. Boyd, H. J. Kimble
1991 Optics Letters  
Measurements of the spectral density of frequency fluctuations are reported for an actively stabilized cw Ti:sapphire laser. For a servo loop incorporating an intracavity translatable mirror and an external-cavity acousto-optic modulator, a linewidth of 1.0 kHz rms is obtained for the fluctuations of the laser frequency as recorded within the servo loop. The modulation index associated with the frequency deviations is considerably less than one over most of the Fourier spectrum, indicating
more » ... tion in a domain of small phase noise for the fluctuations of the electric field. Because of rapid developments over the past several years, the Ti:sapphire (Ti:A1 2 0 3 ) laser is now a reliable source for broadly tunable, high-power radiation with a versatility that leads to diverse applications.' In this Letter we concentrate on the single-frequency performance of the Ti:A1 2 0 3 laser and, in particular, on the active stabilization of the laser frequency fluctuations. 2 Following the research of Schulz, who reported a free-running stability of 2 MHz for a ring Ti:A1 2 0 3 laser, 3 we employ well-established servo techniques to stabilize the frequency of a cw Ti:A1 2 0 3 laser relative to an external reference interferometer. Information about the stability of the laser's frequency is derived from spectral analysis of the error signal from the reference cavity (within the servo loop) and from an independent (out of the servo loop) monitor cavity. With regard to the in-loop error signal from the reference cavity, we observe a linewidth of 1.0 kHz rms for the fluctuations of the laser frequency. More importantly, with the exception of several discrete components (principally at harmonics of the 60-cycle line frequency), the spectral density T(f) for laser frequency fluctuations at a Fourier frequency f is such that T(f) << f, indicating a small modulation index for the residual fluctuations of the laser frequency. 4 A simple diagram of our experimental arrangement is shown in Fig. 1 . The Ti:A1 2 0 3 laser 5 is pumped by the blue-green light of a commercial Ar-ion laser operated in a power-stabilized mode. Besides the four mirrors of the folded ring cavity (one of which is mounted on a piezoelectric transducer), the intracavity elements include a Brewster-cut Ti:A1 2 0 3 crystal, a three-plate birefringent filter (BRF), an optical diode assembly (OD), and a fused-silica 6talon of 1mm thickness (E). A wavelength of 852 nm is chosen for our research, with a power from the Ti:A1 2 0 3 laser of typically 400 mW for 6-W pump power. The output beam is directed to the monitor cavity (free spectral range 3 GHz, finesse 23,000) and to the reference cavity (free spectral range 500 MHz, finesse 275). The beam path to the reference cavity includes an acousto-optic modulator (AOM) (carrier at 77 MHz) operated in double pass, which can serve as a frequency transducer outside of the laser cavity. 6 An error signal for locking the frequency of the Ti:A1 2 0 3 laser and for assessing servo performance is derived in reflection from the reference cavity by the Pound-Drever technique with phase modulation at 6 MHz applied with an electro-optic modulator (EOM). 7 Frequency deviations at the monitor cavity are detected from the fluctuations of the transmitted intensity for a detuning of approximately half of the cavity linewidth; amplitude noise is suppressed by approximately 30 dB by employing the balanced arrangement shown in Fig. 1 . Note that both the reference and monitor cavities are isolated from the laser by AOM's; a Faraday isolator is further provided for the reference cavity or the monitor cavity when the error signal from either is analyzed. Given this discussion of the detection scheme, we wish to emphasize that our measurements are directly sensitive to the spectral density T of laser frequency fluctuations and not to the spectral density b of the fluctuations of the electric field. That is, if we consider a classical field E(t) = E 0 expf-i[wot + i7(t)]} of carrier frequency wo and stochastic phase v(t), then our measurements of photocurrent fluctuations de- Fig. 1 . Outline of the principal elements for frequency stabilization of the Ti:Al 2 0 3 laser. 0146-9592/91/110808-03$5.00/0
doi:10.1364/ol.16.000808 pmid:19776792 fatcat:szrwag6jtrdyvogovxlz37biw4