Elsevier

Optics Communications

Volume 201, Issues 4–6, 15 January 2002, Pages 391-397
Optics Communications

Frequency noise characterisation of narrow linewidth diode lasers

https://doi.org/10.1016/S0030-4018(01)01689-3Get rights and content

Abstract

We examine several approaches to laser frequency noise measurement in the frequency and time domains. Commonly employed methods such as optical frequency discrimination and the Allan variance are found to be complex, expensive, time-consuming, or incomplete. We describe a practical method of demodulating a laser beat note to measure a frequency noise spectrum, using a phase-locked loop frequency discriminator based on a single low-cost integrated circuit. This method measures the frequency noise spectrum of a laser directly and in detail and is insensitive to intensity fluctuations. The advantages of this scheme are demonstrated through measurement of the frequency noise spectrum for two external cavity diode lasers (ECDL), clearly distinguishing several common noise sources. These are isolated and removed, reducing the individual laser rms linewidth from 2 MHz to 450 kHz. The spectrum is used to calculate the Allan variance, which shows almost none of the important information.

Introduction

Narrow linewidth and highly stable lasers are critically important in fields as diverse as optical communications, laser cooling and atomic frequency standards. When designing a laser system, the identification and removal of frequency noise sources is crucial to narrowing the linewidth. Accurate measurement of the laser frequency stability is also often required as an experimental parameter.

External cavity diode lasers (ECDLs) are a common example of such narrow linewidth lasers [1]. ECDLs use semiconductor diode lasers in an external cavity with dispersive feedback, often from a diffraction grating. The linewidth is greatly reduced with respect to the diode alone, and the laser can be tuned through the broad gain curve of the diode.

ECDLs exhibit varying forms of frequency noise from environmental, fundamental, and artificial sources. Environmental noise includes 50 or 60 Hz power-line-induced noise, and acoustically coupled noise, in particular at frequencies corresponding to mechanical resonances. Fundamental noise is typically dominated by white phase noise at high frequencies and flicker frequency noise (also known as 1/f noise or pink noise) at low frequencies [2]. The flicker frequency noise is often the dominant component of the linewidth, and so it is common practice to stabilise the laser to an external frequency reference, for example a Fabry–Perot etalon or sub-Doppler atomic resonance. The feedback systems usually rely on dithering the laser frequency [3] or the reference frequency [4], but in both cases it is common to find significant noise at the dither frequency, thus introducing an artificial source of frequency noise.

Section snippets

Measurement of laser frequency noise

Laser frequency noise may be measured directly at the optical frequency, or by heterodyning to a reference laser.

Analysis of radio frequency spectra

Frequency noise analysis of the beat signal can provide this information, either in the frequency domain or in the time domain. The beat note signal produced by lasers with frequency difference ν0 has voltageV(t)=V0(t)sin[2πν0+φ(t)],where V0(t) describes amplitude fluctuations of the two lasers and φ(t) is the difference of the individual phases. The instantaneous beat frequency isν(t)=ν0+12πdφ(t)dt0+Δν(t)with frequency fluctuations Δν(t)≪ν0.

Phase-locked loop frequency discriminator

The frequency noise spectrum may be recovered directly from the beat signal with a frequency discriminator which outputs a voltage proportional to Δν(t), with minimal dependence on amplitude fluctuations V0(t). The power spectral density of this voltage, for example acquired with an audio frequency spectrum analyser, is SΔν.

Previous experiments using discriminators have required complex rf electronics [12] and suffered from poor spectral resolution [13]. We describe a simple frequency

Conclusion

These results show that phase-locked loop frequency discrimination provides an unambiguous and high-resolution frequency noise spectrum using common laboratory test equipment. The frequency discriminator separates frequency noise from intensity noise, which is particularly valuable in characterising diode laser systems. In contrast with Allan variance analysis, electronic, mechanical and acoustic noise sources may be clearly identified and observed in real time. The method is applicable to

Acknowledgements

We would like to thank Zivko Jovanovski for his expert construction and testing of the radio frequency electronics. This work was supported by the Australian Research Council and conducted with the assistance of the Australian Postgraduate Award scheme (LDT, KPW, CJH).

References (16)

  • A Hemmerich et al.

    Opt. Commun.

    (1990)
  • T.P Dinneen et al.

    Opt. Commun.

    (1992)
  • M Kourogi et al.

    Opt. Commun.

    (1991)
  • A.P Willis et al.

    Opt. Commun.

    (1995)
  • H Ludvigsen et al.

    Opt. Commun.

    (1998)
  • C.E Wieman et al.

    Rev. Sci. Instrum.

    (1991)
  • Y Salvadé et al.

    J. Opt. Soc. Am. A

    (2000)
  • S Kunze et al.

    Opt. Commun.

    (1996)
There are more references available in the full text version of this article.

Cited by (0)

View full text