Open Access
Add to BibSonomyAbstract
We present a method to generate an optical reference comb signal with subpicometer wavelength accuracy. XPM nonlinear effect in a fiber is used to get an optical frequency comb signal, free of frequency chirp and wavelength instabilities, from a pulsed gain-switched laser diode. Principles of such comb generation are presented as well as the application of this comb as a ruler to measure frequency differences in high resolution optical spectrum measurements. To confirm this method, Brillouin filtering optical spectrometric technique is used to characterize a modulated optical source. Typical resolution of this technique allows 0.08 pm wavelength accuracy.
©2008 Optical Society of America
1. Introduction
Previously demonstrated high resolution optical spectrum measurement techniques [1], achieving 0.08 pm resolution and up to 80 dB dynamic range, are of great relevance for the diagnostic and monitoring of signals in optical communication systems [2-4]. Once resolution is achieved, highly accurate wavelength measurements can be of interest for particular cases in which high precision is required. Spectrometric systems using scanning tunable laser source (TLS) inherit wavelength accuracy provided by TLS specification [5]. Hence, wavelength accuracy measurements independent of TLS characteristics will require simultaneous measurement of the signal under test (SUT) with wavelength reference combs.
Absolute wavelength accuracy is usually provided by interferometry of SUT with reference wavelengths obtained from lasers stabilized in the molecular or atomic absorption lines of some gases such as acetylene and similar [6]. On the other hand, pulsed lasers have been proposed as optical sources for calibration purposes and spectroscopy applications [7-10]. They generate a laser frequency comb consisting of hundreds of equally-spaced laser modes over a bandwidth from few GHz up to several THz, depending on pulse repetition rate and pulse width. The shorter the laser pulse, the broader the range of frequencies of the comb. The mode spacing, which is constant in the frequency space, is given by the pulse repetition frequency that can be readily given with very high precision value.
There are many methods to generate optical pulses. Some of the most common commercially available optical pulsed sources are mode-locked solid state or fiber ring lasers. Overcoming their complexity and cost, gain-switching laser diodes have been widely accepted as one of the simplest and most reliable optical pulsed sources [11]. Its major drawbacks are phase mismatch between consecutive pulses and direct modulation of the laser diode which causes a time-varying carrier density in the active region of the device, which in turn causes a variation in the output wavelength from the laser during the emission of the optical pulse. These result in degradation of their optical spectra [11,12].
In general, the use of frequency combs for spectrometric purposes provides relative wavelength accuracy calibration once mode spacing is assured to be stable and well defined. But absolute wavelength accuracy requires to introduce and additional optical reference, such as a TLS, which marks an absolute wavelength origin for the comb. The TLS wavelength can be readily provided, for example, by means of a wavelength meter whose absolute accuracy is then inherited. In reference [13], D. Jones et al. present a technique for generating optical frequency combs at an arbitrary central frequency by using the nonlinear Cross Phase Modulation (XPM) effect between coherent femtosecond pulses from Kerr-lens mode-locked Ti:sapphire laser and stabilized cw laser beams.
Firstly, we propose in this paper to generate optical comb signals from pulsed gain-switched laser diodes, by using the nonlinear XPM effect induced on a TLS providing the absolute wavelength reference. The fundamental feature of these laser diodes is that the optical comb is hidden inside their optical spectrum due to the incoherent character of the emitted pulses: on one hand, phase mismatch correlation between consecutive pulses results in a continuous optical spectrum which corresponds to a unique averaged pulse, on the other hand, chirp effects associated with the direct modulation results in strong asymmetries in the optical spectrum. The XPM nonlinear interaction occurring between the pulsed laser diode and the TLS, along a suitable fiber length, converts the variable intensity of the pulsed source into proportional phase modulation of optical field amplitude [4, 14]. The undesirable optical chirp effects degrading spectrum of the pulsed laser diode are fully removed in the optical spectrum of the converted field. Experimental results demonstrate that a well defined uniform comb is readily obtained, showing equally spaced frequency peaks in the optical spectrum. This method provides a suitable way to use the very compact and potentially cheap gain-switched laser diodes for comb generation purposes. Our use of diode lasers offers a considerable improvement in cost, size and complexity.
Secondly, we present the application of the XPM generated frequency comb as a ruler in the high-resolution optical spectrometric technique based on Brillouin filtering (BOSA) [1]. This technique provides about 0.08 pm resolution and 80 dB dynamic range, whereas the high precision value in determining the comb mode spacing assures also a 0.08 pm relative wavelength accuracy, limited by resolution, in optical spectrum measurements. The TLS tunability enables coverage of optical communications C and L-bands, and the measurement of its wavelength peak provides the absolute wavelength origin of the comb.
An experimental set-up is presented using a gain-switched laser diode generating 12 picoseconds pulses at 100 MHz pulse repetition frequency. A modulated DFB laser is measured showing subpicometric wavelength accuracy and resolution of presented method.
2. Principles of comb generation and experimental lay-out
The experimental setup for optical frequency comb generation is depicted in the squared part of Fig. 1. A commercial gain switched laser diode emitting 75 mW peak power at 1550 nm (from Advanced Laser Diode Systems, model Pilas 100 MHz), is used as pulsed source. It generates 12 ps optical pulses at Ω=100 MHz pulse repetition frequency. Jitter rms is below 4 ps. Its output optical spectrum, measured directly at point A, is shown in Fig. 2. Due to the incoherent character of the optical source, the uniform comb of laser modes corresponding to the well defined power envelope of the pulses is clearly hidden in the continuous optical spectrum, whereas the chirp effect results in a strong asymmetry. A TLS amplified by an EDFA is used as carrier wave for the XPM effect. Its output optical spectrum measured directly at point B is shown in the red curve of Fig. 3. Both optical waves are injected in a 6.6 km dispersion shifted fiber through a 90/10 optical coupler. TLS output power is about 2 mW, amplified up to 60 mW by the EDFA, to obtain 6 mW at the fiber input. Relevant parameters of the fiber are n2=2.2×10-20m2/W, D=0.01 ps/km1/2, λ0=1544 nm and S=0.07 ps/(km-nm2).
Along the fiber, the carrier wave is modulated by the XPM effect induced by the pulsed source. The magnitude of this induced phase modulation, denoted by ϕ(t), depends on the temporal intensity modulation I(t) of the pulsed laser diode optical wave by the equation [4]:
with a modulation factor m that depends on the effective fiber length Leff that takes into account the optical power attenuation, the vacuum wavelength of the carrier optical wave λ0, the core effective area Aeff and the nonlinear refractive index n2. For simplicity, we neglect XPM polarization dependences and electrostrictive contributions to XPM since their impact on the frequency comb generation is not important for our purpose.
Thereby, after propagating along the fiber, the XPM modulated carrier optical spectrum can be developed as the sum of a Dirac’s delta function δ centred at ω 0 and a term proportional to the RF spectrum of the pulsed laser diode:
where S(ω-ω 0) is the pulsed laser power spectrum. According to Eq. (2), each RF component nΩ of the pulsed laser diode electrical spectrum generates on the carrier optical spectrum, in addition to the central peak, symmetrical modulation sidebands at the optical frequencies ω 0+nΩ and ω 0-nΩ, with n an integer. As a result, the XPM broadened carrier wave represents a replica of the light intensity optical spectrum of the pulsed laser diode. By using this simple concept, a well defined optical uniform frequency comb is obtained centered at the carrier wavelength. It is independent of frequency chirp and wavelength instabilities of the optical source generating the optical pulses. It spans over a range of about 80 GHz given by the inverse Fourier transform of the 12 picoseconds pulses duration. Figure 4 is a zoom of the orange curve of Fig. 3 showing the 0.08 pm resolved laser modes of the comb. They are located at both sides of ω 0 and correspond to the equally-spaced frequencies ω 0+nΩ and ω 0-nΩ. The mode space frequency Ω can be readily given in the frequency domain by using simple and conventional RF-measuring techniques.
Fig. 2. Optical spectrum of the gain switched laser diode, 12 picoseconds pulse width at 100 MHz pulse repetition frequency, measured at point A.
Fig. 3. Obtained optical frequency comb for the TLS centered at 1544.9 nm, measured at point C (orange curve). Original TLS optical spectrum, measured at point B (red curve).
Absolute wavelength accuracy of the comb is determined by the accuracy in giving the TLS peak wavelength. Subpicometer accuracies can be reached for example by available commercial wavelength meters.
Relative wavelength accuracy will depend on the optical resolution of the comb measurement, and on the wavelength uniformity between adjacent modulation bands in the comb. Assuming wavelength uniformity between the 100 MHz (0.8 pm) spaced laser modes, we can assure a relative wavelength accuracy better than the 0.08 pm resolution figure of the implemented optical spectrometric technique.
The use of a TLS as carrier wave allows a convenient tuning method of the central frequency of the comb, covering optical communications C- and L-bands. The EDFA amplification power allows achieving the desirable optical power level of the comb. In this case we have selected to be more than about 10 dB over noise level in the 80 GHz comb width.
This method is applicable to different pulse parameters and any optical pulse technology, without restrictions due to frequency chirp and wavelength instabilities existing in the pulsed optical source. So it opens the way to generate many different reference combs from any power pulsed optical source.
3. Application of the comb in high-resolution optical spectrometry and results
The experimental setup for subpicometer wavelength accuracy optical spectrometric application is depicted in Fig. 1. As described above, a reference comb is generated by using XPM effect between optical wave coming from a pulsed gain switching laser diode and a carrier wave coming from a TLS. This reference comb and the SUT are measured, through an optical coupler, by means of BOSA-C equipment from Aragon Photonics Labs. Comb is used as a ruler giving equally spaced frequency peaks in the optical spectrum. The central wavelength of the comb is measured by using a ±0.3 pm accuracy wavelength meter at the TLS output, whereas the mode space frequency, which is given by the pulse repetition frequency Ω, is measured using a ±5 ppm accuracy frequency meter.
In order to test the validity of the proposed technique, a DFB laser source externally modulated with a Mach-Zehnder optical modulator has been measured. The DFB wavelength is λDFB=1554.9841±0.0003 nm and the modulation frequency Φ is set to 1 GHz ±5 kHz. Results are presented in Fig. 5. Orange curve represents the measured optical frequency comb, centered at TLS wavelength λTLS=1554.9933±0.0003 nm to cover the SUT, with the optical peaks equally spaced at frequency Ω=100 MHz ±0.5 kHz. Blue curve represents the measured SUT optical spectrum, showing the central peak and the corresponding modulation sidebands. The right second harmonic modulation sideband is also shown.
The comb provides a ruler to measure frequency differences between measured points of the SUT optical spectrum, from which absolute wavelengths can be directly calculated considering λTLS. To give the exact position of a measured point in this ruler, we use linear interpolation between adjacent peaks of the comb assuming wavelength uniformity between them.
The SUT measured optical spectrum shows that its central peak, denoted by λSUT, is located at a distance of 11.45 peaks from the origin of the comb (11 peaks counted from the origin plus 0.45 peaks calculated by linear interpolation between the wavelengths of the SUT central peak and its adjacent comb peaks, measured in the optical spectrum, which are respectively 1554.98370 nm, 1554.98414 nm and 1554.98316 nm). Hence, the measured frequency difference is 1145 MHz, which corresponds to an absolute wavelength of λSUT=1554.98407 nm, in agreement with the measured DFB wavelength λDFB.
Following the same procedure, the relative positions of the SUT modulation sidebands with respect to its central peak are measured. As it is shown in Fig. 5, they are located at distances of 10.05, 9.94 and 19.97 peaks, which correspond to frequency differences of 1005 MHz, 994 MHz and 1997 MHz. These measured frequency differences agree with the nominal values of Φ and 2Φ within the expected uncertainty interval of ±10 MHz (0.08 pm), confirming that subpicometer relative wavelength accuracy of about 0.08 pm is achieved.
4. Conclusions
A reference optical comb with subpicometer wavelength accuracy has been presented and experimentally demonstrated. It uses XPM nonlinear effect in a dispersion shifted fiber to generate a suitable optical frequency comb from a pulsed gain-switched laser diode. This optical frequency comb is extracted from the direct continuous, chirp containing, optical spectrum that exhibits the pulsed source. This technique is applicable to different pulse parameters and any optical pulse technology. The comb is used as a ruler giving equally spaced frequency peaks in the optical spectrum.
A spectrometric application of this optical comb has been performed using the high resolution optical measurement technique based on Brillouin filtering to achieve 0.08 pm resolution and 80 dB dynamic range, which results in a 0.08 pm wavelength accuracy optical spectrometric technique. Measurements carried out on an externally modulated DFB laser source confirm the good performance of the proposed method.
Acknowledgments
This work was supported by the Spanish “Secretaría de Estado de Universidades e Investigación (MEC)” under Project FIS2007-64443.
References and links
1. J. M. Subias Domingo, J. Pelayo, F. Villuendas, C. D. Heras, and E. Pellejer, “Very high resolution optical spectrometry by Stimulated Brillouin Scattering,” IEEE Photon. Technol. Lett. 17, 855–857 (2005). [CrossRef]
2. S. Yang and X. Bao, “Experimental observation of excess noise in a detuned phase-modulation harmonic mode-locking laser,” Phys. Rev. A 74, 033805 (2006). [CrossRef]
3. A. Villafranca, J. Lasobras, J. A. Lazaro, and I. Garces, “Characterization of the main semiconductor laser static and dynamic working parameters from CW Optical Spectrum Measurements,” IEEE J. Quantum Electron. 43, 116–122 (2007). [CrossRef]
4. C. Heras, J. Subias, J. Pelayo, and F. Villuendas “High resolution light intensity spectrum analyzer (LISA) based on Brillouin optical filter,” Opt. Express 15, 3708–3714 (2007). [CrossRef] [PubMed]
5. T.-J. Ahn and D. Kim, “Analysis of nonlinear frequency sweep in high-speed tunable laser sources using a self-homodyne measurement and Hilbert transformation,” Appl. Opt. 46, 2394–2400 (2007). [CrossRef] [PubMed]
6. K. Nakagawa, M. de Labachelerie, Y. Awaji, and M. Kourogi, “Accurate optical frequency atlas of the 1.5-mm bands of acetylene,” J. Opt. Soc. Am. B 13, 2708–2714 (1996). [CrossRef]
7. V. Gerginov, C. E. Tanner, S. A. Diddams, A. Bartels, and L. Hollberg, “High-resolution spectroscopy with a femtosecond laser frequency comb,” Opt. Lett. 30, 1734–1736 (2005). [CrossRef] [PubMed]
8. E. V. Baklanov and P. V. Pokasov, “Optical frequency standards and femtosecond lasers,” Quantum Electron. 33, 383–400 (2003). [CrossRef]
9. R. Holzwarth, M. Zimmermann, Th. Udem, and T. W. Hänsch, “Optical Clockworks and the Measurement of Laser Frequencies with a Mode-Locked Frequency Comb,” IEEE J. Quantum Electron. 37, 1493–1501 (2001). [CrossRef]
10. J. N. Eckstein, A. I. Ferguson, and T. W. Hansch“High-resolution two-photon spectroscopy with picosecond light pulses,” Phys. Rev. Lett. 40, 847–850 (1978). [CrossRef]
11. P. M. Anandarajah, C. Guignard, A. Clarke, D. Reid, M. Rensing, L. P. Barry, G. Edvell, and J. D. Harvey, “Optimized Pulse Source Employing an Externally Injected Gain-Switched Laser Diode in Conjunction with a Nonlinearly Chirped Grating,” IEEE J. Sel. Top. Quantum Electron. 12, 255–264 (2006). [CrossRef]
12. L. Chusseau, “Propagation of Single-Mode 1.5-µm Gain-Switched semiconductor Laser Pulses in Normally Dispersive Fibers,” IEEE J. Quantum Electron. 30, 2711–2720 (1994). [CrossRef]
13. D. Jones, S. Diddams, M. Taubman, S. Cundiff, L. Ma, and J. Hall, “Frequency comb generation using femtosecond pulses and cross-phase modulation in optical fiber at arbitrary center frequencies,” Opt. Lett. 25, 308–310 (2000). [CrossRef]
14. C. Dorrer and D. N. Maywar, “RF Spectrum Analysis of Optical Signals Using Nonlinear Optics,” J. Lightwave Technol. 22, 265–274 (2004). [CrossRef]
