We propose and demonstrate experimentally a laser source whose linewidth is adjustable independently of its other characteristics. This source can be used to test whether a particular laser would be suitable in a system, without the need to purchase several different lasers. It also has the advantage that the linewidth is generated digitally so it is extremely stable over time. We demonstrate a dialed-linewidth emulator between 256 kHz to 150 MHz. The narrowest linewidth shown by this technique is the original linewidth of the semiconductor laser source used in the setup. We also investigate the effect of driving our modulator into its nonlinear range.
©2010 Optical Society of America
The performance of many communication systems depends critically on the characteristics of the semiconductor laser at the transmitter [1–5]. To prototype a system and assess the effect of linewidth on the system, it would be useful to have an optical source whose linewidth could be “dialed up” without affecting its output power, intensity noise or modal structure. A typical way to vary the linewidth of a semiconductor laser is by using the dependence of linewidth, Δv, on the inverse of laser output power, Δv = F(1/P) . However, creating a wide linewidth requires operation at very low powers and this causes a decrease in spectral purity such as the appearance of additional modes and a very high Relative Intensity Noise (RIN).
In this paper, we propose a method of emulating a semiconductor laser with a variable linewidth. With this method, the dialed-up linewidth is independent from the laser’s other characteristics. For example, a wide linewidth can be achieved even when the laser operates at a high output power giving a low Relative Intensity Noise (RIN). We demonstrate the emulator experimentally over a linewidth range of 256 kHz to 150 MHz full-width half-maximum (FWHM). The narrowest linewidth in this range depends on the original linewidth of the laser source. We also show that the dialed-linewidth laser is stable over many hours because the linewidth is generated digitally. An application of this laser linewidth emulator is to allow the effects of linewidth to be investigated in a transmission system. For example, the effect of phase-noise conversion into intensity noise in a dispersive fiber . Previously, it was difficult to separate the effects of laser linewidth and laser RIN .
The principle of the emulator is to phase-modulate the output of a narrow-linewidth laser to broaden its linewidth, Δv, using digitally-generated phase modulation. That is, the phase θn at sample time n.ΔT will be updated every timestep, ΔT. From Eq. (4).99  the change in phase per timestep, θn - θn-1, is a Gaussian random variable with a variance σ2 = 2πΔvΔT. This phase could be applied by using a lithium niobate (LiNbO3) phase modulator as illustrated in Fig. 1a ; however, this modulator has a limited modulation range, set by the limits of its drive voltage. Thus the phase has to be “reset” by adding or subtracting 2π radians whenever a voltage limit is approached. This reset produces a large frequency chirp, which is undesirable. Thus, we instead use a complex optical modulator (C-MZI) based on nested Mach-Zehnder Interferometers (MZIs), shown in Fig. 1b . This allows the phase to be driven indefinitely around the complex plane, without need for reset events.
The output field, Eout, of a complex optical modulator, when its upper and lower MZIs are combined in quadrature by adjusting Vc (see Fig. 1b) is related to its input field, Ein, by
In the experimental setup, VI and VQ will be generated digitally then converted to an analog drive for the modulator using an arbitrary waveform generator (AWG). Because the AWG repeats the same sequence periodically (every TAWG), the optical spectrum comprises delta functions spaced at 1/TAWG. However, if the primary laser source’s linewidth is significant compared to 1/TAWG; each delta function will be convolved with this linewidth, causing the spectrum to be continuous. This produces some phase correlation between the successive periods. However, the correlation will reduce when TAWG.Δv >> 1. This characteristic must be considered when using the emulator in a system or measurement where a long-scale correlation (>TAWG) may be important.
3. Experimental setup and linewidth measurements
Figure 2 shows the experimental setup. The laser source was a Photonetics Tunics External Cavity Laser (ECL) tuned to 193.0 THz with a 256-kHz linewidth. The green inset shows the laser emulator which includes a Tektronix 7102 2 × 10GS/s AWG driving a Sumitomo T-SBXI.5-20P 40-Gbit/s complex optical modulator. The modulating signals, VI and VQ were generated using VPItransmissionMakerTM using its “LaserCW” model and a 10-GHz optical simulation bandwidth (10 GS/s sample rate). The waveforms were 52.4288-µs long and the AWG repeated them continuously. 5-GHz low pass filters (LPF) were used at the RF inputs of the C-MZI to filter out the image spectrum. The DC biases (Va, Vb in Fig. 1) were adjusted to ensure each MZI was biased at null power and Vc was adjusted so the upper and lower MZI’s outputs are in quadrature. In the experiment, k was 1.35V and Vπ was 5.6V.
Figure 3 shows linewidth measurement obtained using an Agilent High Resolution Spectrophotometer (HRS). The dialed-linewidth was 20 MHz. The HRS fitted a Lorentzian curve to its spectrum, giving a linewidth estimate of 20.119 MHz.
The second linewidth measurement used a self-heterodyne coherent receiver [10,11] as shown in the right-hand part of Fig. 2. This technique requires a frequency offset, foffset, between the reference laser and the measured laser . A frequency shift of 2.5 GHz can be generated by adding a linearly-increasing phase to the modulating signals. The modulated signal after the EDFA is connected via a polarization controller (PC) into the signal input of a Kylia MINT optical hybrid. The local oscillator (LO) input of the hybrid is directly connected to the ECL laser. The outputs of the hybrid were detected with two U2T 30-GHz balanced photodiodes (PDs) giving inphase (I) and quadrature (Q) components of the downconverted signal. These signals were then amplified with MiniCircuits ZX60-14012L-st RF amplifiers before being input into a Tektronix 72004 20-GHz 50-GSa/s digital sampling oscilloscope (DSO). To obtain the power spectral density, a Fast Fourier Transform (FFT) was applied to the I waveform (the Q waveform would give identical results). This produced a Lorentzian-shaped spectrum, centered at foffset (2.5 GHz) as shown in Fig. 4 . A width measurement 20-dB down from the spectrum’s peak gives √99 times the FWHM linewidth . The figure shows that the measured linewidth is in good agreement with the 20-MHz dialed laser linewidth.
4. Results for a range of dialed linewidths
The linewidths from the measurement with HRS and the coherent receiver for various dialed linewidths are plotted in Fig. 5 . The two measurements show that the linewidth follows the desired dialed linewidth accurately over a wide range of linewidths.
Figure 6 shows the stability of the laser emulator compared with a commercial DFB laser (FLD5F6CX-E36) operated at 30 mA and 40 mA, measured over 2 hours, using the HRS. The measured linewidth of the DFB laser fluctuates by up to a factor of two, whereas the emulated linewidth laser remains stable over time because it is generated digitally.
5. Effects of complex modulator driving voltage and electrical bandwidth limitation
Figure 7 shows the simulated effect of changing the drive amplitude, k. Figure 7a shows that for low drive levels (k = 0.1 × Vπ), there is only phase modulation indicated by a circle in the complex plane. When the modulator is driven into its nonlinear region (k = Vπ), the output amplitude depends on with phase, as indicted by the square of Fig. 7b. Interestingly, the measured linewidth was not affected by overdriving the modulator. Figure 7c shows the RF spectrum the modulator’s output for linear and nonlinear drive levels. Driving the modulator in a nonlinear region causes a large increase in the intensity noise of the emulator, which would itself affect the performance of most systems when tested using the emulator.
Figure 8 shows the simulated effects of limiting the RF bandwidth, to simulated the posssible effects of the combined frequency response of the digital-to-analog converters, drive electronics and optical modulator. Without any limit to the bandwidth, the spectrum is Lorentzian. With 2.5-GHz 4th-order Bessel filters limiting the electrical bandwidth, the higher-frequencies of the tails of the Lorentzian spectrum are reduced in power. This would change the statistics of the phase noise; for example, the variance of the frequency fluctuations (the time-derivative of the phase) will be reduced.
We have demonstrated a method that emulates the linewidth of a semiconductor laser over the range of 256 kHz to 150 MHz. We have also shown that the emulated linewidth is stable over many hours because it is generated digitally. This emulator can be used in systems to investigate the effects of linewidth on system performance. For successful operation, a source laser with low intensity noise and a narrow linewidth should be chosen. The complex modulator should be driven in its linear region otherwise intensity noise will be added.
We should like to thank VPIphotonics (www.vpiphotonics.com ) for the use of their simulator, VPItransmissionMakerWDM V8.3. This work is supported under the Australian Research Council’s Discovery funding scheme (DP1096782).
References and links
1. Z. Zan, A. J. Lowery, and M. Premaratne, “Laser RIN and linewidth requirements for direct detection optical OFDM,” in Conference on Lasers and Electro-Optics CLEO 2008 (San Jose, 2008), paper CWN2.
2. J. A. P. Morgado and A. V. T. Cartaxo, “Assessment of laser noise influence on direct-detection transmission system performance,” J. Lightwave Technol. 21(3), 759–768 (2003). [CrossRef]
3. W. K. Marshall, B. Crosignani, and A. Yariv, “Laser phase noise to intensity noise conversion by lowest-order group velocity dispersion in optical fiber: exact theory,” Opt. Express 23, 165–167 (2000).
4. P. Laurencio, S. O. Simőes, and M. C. R. Medeiros, “Impact of the combined effect of RIN and intermodulation distortion on OSSB/SCM systems,” J. Lightwave Technol. 24(11), 4250–4262 (2006). [CrossRef]
5. M. Ahmed and M. Yamada, “Effect of intensity noise of semiconductor lasers on the digital modulation characteristics and the bit error rate of optical communication systems,” J. Appl. Phys. 104(1), 013104 (2008). [CrossRef]
6. C. H. Henry, “Theory of the linewidth of semiconductor lasers,” IEEE J. Quantum Electron. 18(2), 256–259 (1982). [CrossRef]
7. S. Yamamoto, N. Edagawa, H. Taga, and Y. Y. H. Wakabayashi, “Analysis of laser phase noise to intensity noise conversion by chromatic dispersion in intensity modulation and direct detection optical fiber transmission,” J. Lightwave Technol. 8(11), 1716–1722 (1990). [CrossRef]
8. L. Kazovsky, S. Benedetto, and A. Willner, “Laser phase noise model,” in Optical Fiber Communication Systems (Artech House, Inc., Norwood, 1996).
9. D. Fonseca, A. V. T. Cartaxo, and P. Monteiro, “Optical single-sideband transmitter for various electrical signaling formats,” J. Lightwave Technol. 24(5), 2059–2069 (2006). [CrossRef]
10. D. Derickson, “Laser linewidth characterization,” in Fiber Optic Test and Measurement (Prentice Hall, New Jersey, 1998).
11. B. J. C. Schmidt, Z. Zan, L. B. Du, and A. J. Lowery, “120 Gbit/s over 500-km using single-band polarization-multiplexed self-coherent optical OFDM,” J. Lightwave Technol. 28(4), 328–335 (2010). [CrossRef]
12. Z. Zan, L. B. Du, and A. J. Lowery, “Experimental demonstration on the reduction of linewidth impact in a self-heterodyne optical OFDM system,” in Optical Fiber Communication Conference (OFC/NFOEC), (OSA), (San Diego 2010), paper JThA8.