MULTIWAVELENGTH BRILLOUIN FIBER LASER WITH MEASURABLE LINEWIDTH AND SIGNAL-TO-NOISE RATIO USING 0.16 pm RESOLUTION
OPTICAL SPECTRUM ANALYZER
Farah Diana Muhammad1, Mohd Zamani Zulkifli2 and Harith Ahmad2
1Depertment of Physics, Faculty of Science, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia
2Photonics Research Centre, Department of Physics, University of Malaya, 50603 Kuala Lumpur, Malaysia
Corresponding author: farahdiana@upm.edu.my ABSTRACT
In this work, the measurement of Brillouin Stokes is undertaken using a very high resolution optical spectrum analyzer (OSA), which reveals many new and interesting results. The Stokes lines are generated using a conventional multiwavelength brillouin fiber laser (MBFL), and measured using an OSA with a resolution of 0.16 pm, giving very distinct Stokes lines from the base level. Up to eight Stokes lines can be generated at a Brillouin Pump (BP) power of 447 mW, with the first four Stokes forming a flat comb at –15 dBm from 1550.09 to 1550.33 nm. Direct measurement of the linewidth of the Stokes lines, previously not possible, gives a value of about 0.25 pm, with a well- defined Signal-to-Noise ratio of between 57.2 dB to 9.9 dB.
INTRODUCTION
It has been of interest to develop multiwavelength fiber laser from a single laser source, which can find many attractive applications in dense wavelength division multiplexed (DWDM) optical communication systems, photonic component characterization, optical metrology, optical fiber sensors and photonics true-time-delay (TTD) beam- forming systems [1-4]. Of many approaches, multiwavelength laser generation by stimulated Brillouin scattering (SBS) process for creating Brillouin multiwavelength fiber laser (BMFL) has gained much attention and is extensively studied among the researchers [5-12]. This is attributed by the narrow linewidth of the Brillouin laser output, also known as Brillouin Stokes, which is vital for other applications such as gravitational wave detection [13] and coherent combination [14], provided that the output power of the Brillouin fiber laser is high [15]. Besides that, the small wavelength spacing of about 10 GHz (0.08 nm) and low threshold power are also the advantages of the Brillouin fiber laser [16].
Traditionally, the measurement of the linewidths of Brillouin Stokes cannot be done
self-heterodyned or heterodyned beat techniques, [17] which gives the results in frequency domain [15,18-23]. However, these techniques are complex to setup and operate, and consist of often bulky and expensive components or equipment such as the radio frequency spectrum analysers (RFSAs), high speed photodetector, acousto-optic modulators, direct current (DC) power supplies and long lengths of single mode fibers (SMFs). In addition to this, ordinary OSA also limits signal-to-noise ratio (SNR) measurements, another important characteristic of the Brillouin fiber laser outputs. As such, a very high resolution OSA that can directly measure both the linewidth and SNR will be a significant advantage, as it will allow detailed measurements.
In this paper, a new, high resolution OSA is used for the first time, to the best knowledge of the authors, to provide detailed measurements of Stokes lines such as distinct wavelength measurements, direct linewidth and SNR measurement. The OSA has a spectral resolution of up to 0.16 pm, and provides an interesting examination of the Brillouin Stokes generated from a multiwavelength Brillouin fiber laser (MBFL).
EXPERIMENTAL
Figure 1 shows the experimental setup for the multiwavelength Brillouin fiber laser, which is constructed in a linear cavity configuration by utilizing a pair of optical circulators, connected at both ends of the cavity. A 1.96 km dispersion shifted fiber (DSF) with a dispersion value of -203ps/nm is used as the nonlinear medium for generating the multiwavelength Brillouin output based on the stimulated Brillouin scattering (SBS) process. A tunable laser source (TLS) acts as the Brillouin Pump (BP), which is then amplified by an erbium doped fiber amplifier (EDFA), giving a maximum BP power of 26.5 dBm (447 mW).
Figure 1: The experimental setup for the multiwavelength Brillouin fiber laser
The pump power of the laser diode in the EDFA is set at 373 mW with output wavelength of 980 nm. The amplified BP output from the EDFA is then injected into the linear cavity through a 2x1 3dB coupler, whereby the BP output is connected to the
50% port of the 3dB coupler. The 100% port of the 3dB coupler is then connected to Port 1 of an optical circulator (designated as OC1), whereas Port 2 of the OC1 is connected to the DSF, with the other end of the DSF is connected to Port 2 of the second optical isolator (designated as OC2). Port 3 of the OC1 is then connected to Port 1 of the OC1 via the other 50% port of the 3 dB coupler, thereby forming the “mirror”
for the linear cavity. With the BP power injected at its maximum value of 26.5 dBm (447 mW), the average power at Port 2 of OC1 is measured to be approximately 24 dBm (251 mW). On the other hand, Port 3 of the OC2 is connected to Port 1 of the OC2 through a 90:10 coupler to create the second “mirror”, thus completing the linear cavity for the multiwavelength Brillouin laser. The 10% port of the optical coupler serves as the output of the Brillouin laser, which is connected to an OSA (YOKOGAWA AQ6370B) with a resolution of 0.02 nm for the measurement of the output spectrum.
The output spectrum measurement is repeated by using a resolution of 0.16 pm OSA (APEX AP2051A) for comparison purpose as well as for the performance analysis.
The schematic diagram of the measurement principle of the OSA with the resolution capability of 0.16 pm is shown in Figure 2. Such ability of this OSA with this resolution is realized by having the local oscillator (tunable laser source) to interfere with the signal under test giving beatings on the receiver section. These beatings are then filtered at the RF section of the equipment. Polarization independence is realized by splitting the signal under test into two orthogonal polarization states. These two signals are analyzed separately and are combined by the equipment software to give a polarization independent spectrum analysis. A wavelength calibrator is integrated in the OSA giving an absolute wavelength accuracy of +/- 3 pm after calibration. This consequently provides the high-resolution ability of the OSA.
Figure 2: The schematic diagram of the measurement principle of the OSA (APEX AP2051A)
In terms of the working principle of this laser cavity configuration, as the BP travels into the DSF through the 3 dB coupler and the OC1, the first Brillouin Stokes is generated in the DSF, provided that BP power exceeds the threshold power for generating the first Brillouin Stokes. The generated first Brillouin Stokes would then propagate in the opposite direction of the BP and will be reflected back by the OC1 towards the DSF to generate the second Brillouin Stokes and the similar process continues in the generation of other Brillouin Stokes until the power of the preceding Brillouin Stokes could not overcome the threshold power for generating the subsequent Brillouin Stokes.
RESULTS AND DISCUSSION
In order to distinguish the spectrum of the BP signal from the 0.02 nm resolution OSA and the 0.16 pm resolution OSA respectively, a spectrum measurement test is carried out on the BP signal output. The BP power is set at 1 dBm and fixed at output wavelength of 1500 nm. Figure 3 shows the output spectrum of the BP signal whereby the red trace is taken from the 0.02 nm resolution OSA (YOKOGAWA AQ6370B) whereas the green trace is taken from the 0.16 pm resolution OSA (APEX AP2051A).
In terms of the peak power of the output laser spectrum, the 0.16 pm resolution OSA yields a reading of -0.13 dBm whilst the 0.02 nm resolution OSA yields a reading of 0.73 dBm, which differs by 0.8 dB. The reason for the lower peak power obtained in the 0.16 pm resolution OSA is related to the power sensitivity, which would be higher with a higher resolution spectrum. The peak wavelength of the spectrum from both the 0.16 pm and the 0.02 nm resolution OSA is observed at 1500.01 nm (which indicates the output scale of the TLS is out of calibration by 0.01 nm). From the figure, it can be inferred that the output from the TLS is not really a single mode as widely being reported, due to the present of the side peaks, as can be seen from the figure, with several side modes at the wavelengths of 1550.04 nm, 1549.98 nm, 1550.07 nm, 1549.96 nm, 1550.1 nm, 1549.93 nm, 1550.13 nm, 1550.16 nm and 1550.19 nm, which could not be observed using the 0.02 nm resolution OSA.
Figure 3: Output spectra of the TLS (the red trace is taken from the 0.02 nm resolution OSA, the green trace is taken from the 0.16 pm resolution OSA)
Figure 4 (a) shows the output spectrum of the multiwavelength Brillouin fiber laser taken at maximum BP power of 447 mW (after being amplified by the EDFA) using the 0.16 pm resolution OSA at BP wavelength of 1550 nm. There are 8 Stokes and about 5 anti-Stokes lines generated. The anti-Stokes lines observed in the figure are generated based on the four wave mixing (FWM) effect, in a way that the interaction between the BP and the first Brillouin Stokes results in the growth of the first anti-Stokes and so forth. Also can be seen, there are 2 side modes between the BP and the 1st Stokes and also between the BP and the 1st anti-Stokes, which are normally not seen using a lower resolution OSA, which comes from the output of the TLS. Four lasing wavelengths (Brilluoin Stokes) with a spacing of 0.08 nm in between them, with almost flat and stable output power at about -15 dBm are obtained, which spans from 1550.09 nm to 1550.33 nm. For comparison purposes, the output spectrum taken previously from the 0.02 nm resolution OSA is superimposed onto the trace from the 0.16 pm resolution OSA, which is shown in Figure 4 (b). The blue and red traces are for the 0.02 nm and 0.16 pm resolutions OSA respectively. As shown in the figure, the Brillouin Stokes lines as taken by the 0.02 nm resolution OSA shows small peaks on top of a broad spectrum whereas the 0.16 pm resolution OSA reveals a detail and sharper spectrum with sharp lines that start from the baselines. This allows the SNR of the Brillouin Stokes lines to be precisely measured as well as showing the noise of the laser output, which has never been seen before. As a result, this provides a better analysis of the quality or the performance of the Brillouin laser output.
Figure 4: (a) Output spectrum of the multiwavelength Brilouin fiber laser taken from the 0.16 pm resolution OSA (b) the combined output spectrum taken from both OSAs (the blue trace is taken from the 0.02 nm resolution OSA, the red trace is taken from the 0.16 pm resolution OSA)
Figure 5 (a) – (e) shows the spectrum obtained using the 16 pm resolution OSA of the Brillouin multiwavelength fiber laser output at different BP powers, after amplification by the EDFA. It is evident from the progression of Figure 5 that the system exhibits the behavior typical of a Brillouin fiber laser, with the number of Stokes lines increasing as the pump power is raised. The lasing threshold of the laser is 324 mW, whereby the Stokes lines can first be seen. The first Stokes is very distinct, while the second Stokes is also visible at a lower power. The first anti-Stokes is also visible at this point, while some side-modes arising from the BP are also observed. At a BP power of 339 mW, more Stokes are now generated, with the first and second Stokes lines having powers of about -15 dBm, while the third and fourth Stokes have smaller amplitudes. Two anti- Stokes lines are also observed. Increasing the BP power to above 363 mW, until the maximum power of 447 mW shows an increase in the Stokes lines generated, from 6 to 8 well-defined Stokes lines, as well as 3 to 5 anti-Stokes lines. It can also be seen in Figure 5 that increasing the BP power will improve the power of the Stokes lines, with the obtained comb become flatter and wider as the pump power is increased. The system is capable of generating further Stokes lines but the limited pump power of the TLS, which serves as the BP in this work, prevents this. This however can be overcome if more powerful BP sources are used as well as by enhancing the performance of the EDFA.
Figure 5: (a)-(e). Output spectrum of the Brillouin multiwavelength fiber laser at different BP power taken from the 0.16 pm resolution OSA
Figure 6 (a) and (b) shows the linewidth spectra of the first and second Brillouin Stokes, taken at a BP power of 447 mW. It can be seen that both Stokes have a well- defined laser output, with linewidths of approximately 0.25 pm for the first and second Stokes respectively. The subsequent Stokes also have about similar linewdiths, and is shown in Figure 7.
Figure 6: The zoom in view of (a) the 1st Stokes, (b) the 2nd Stokes
Figure 7: The 3 dB linewidth of the Brillouin Stokes against the number of Brillouin Stokes
Figure 8 shows signal-to-noise ratio (SNR) of the Brillouin Stokes against the number of Brillouin Stokes, taken at a BP power of 447 mW. It can be seen that the measured SNR values correspond strongly to the spectra of Figure 5, with the first four Stokes having a relatively flat power spectrum, while the power of subsequent Stokes decreases. In this manner as well, the SNR of the first four Stokes lines are almost constant, at a value of about 57.2 dB, but decreases from the fifth Stokes onwards, from 49.0 dB to its lowest value of 9.9 dB at the eight Stokes.
Figure 8: The SNR of the Brillouin Stokes against the number of Brillouin Stokes
potential for use in multiple applications that require a closely-spaced multi-wavelength comb.
CONCLUSION
The measurement of Brillouin Stokes using an OSA with a resolution of 0.16 pm is undertaken and discussed. The Stokes lines are obtained from a typical MBFL setup, which generates up to 8 Stokes lines at a BP power of 447 mW. Using the new OSA, significant details are observed, such as the output of the BP not being a true single- mode output, as evidenced by multiple side-bands. Measurement of the Stokes lines provides distinct wavelengths, with the first four Stokes forming a flat comb ranging from 1550.09 nm to 1550.33 nm at a power of -15 dBm. Direct measurements of the linewidth and SNR of the Stokes lines are also carried out, which were not previously possible, giving the average values of 0.25 pm and 57.2 dB respectively.
REFERENCES
[1]. H. Ahmad, M. Z. Zulkifli, F. D. Muhammad, M. H. Jemangin, K. Dimyati K, B.
P. Pal and S. W. Harun, IEEE Photonics Journal 4 2050-2056 (2012) [2]. X. Dong, P. Shum and N. Q. Ngo, Optics Express 14 3288-3293 (2006)
[3]. T. Miyazaki, N. Edagawa, S. Yamamoto and S. Akiba, IEEE Photonics Technology Letters 9 910-912 (1997)
[4]. J. Chow, G. Town, B. J. Eggleton, M. Ibsen, K. Sugden and I. Bennion, IEEE Photonics Technology Letters 8 60-62 (1996)
[5]. M. R. Shirazi, M. Biglary, S. W. Harun, K. Thambiratnam and H. Ahmad, Journal of Optics A: Pure and Applied Optics 10 055101 (2008)
[6]. M. H. Mansoori, M. A. Mahdi and M. Premaratne, IEEE Journal of Selected Topics in Quantum Electronics 15 415-421 (2009)
[7]. Y. G. Shee, M. H. Mansoori, A. Ismail, S. Hitam and M. A. Mahdi, Optics Express 19 1699-1706 (2011)
[8]. T. F. S. Buttner, I. V. Kabakova, D. D. Hudson, R. Pant, E. Li and B. J.
Eggelton, Optics Express 20 26434-26440 (2012)
[9]. H. Ahmad, M. Z. Zulkifli, N. A. Hassan and S. W. Harun, Applied Optics 51 1811-1815 (2012)
[10]. J. Tang, J. Sun, L. Zhao, T. Chen, T. Huang and Y. Zhou, Optics Express 19 14683-14689 (2011)
[11]. H. Ahmad, M. Z. Zulkifli, M. H. Jemangin and S. W. Harun, Laser Physics Letter 10 055102 (2013)
[12]. R. Parvizi, H. Arof, N. M. Ali, H. Ahmad and S. W. Harun, Optics & Laser Technology 43 866-869 (2011)
[13]. Advanced LIGO homepage: http://www.ligo.caltech.edu/advLIGO
[14]. M. Wickham, J. Anderegg, S. Brosnan, D. Hammons, H. Komine and M.
Weber, in Proc. Advanced Solid State Photonics (Santa Fe, NM, USA,2004) paper MA4 pp. 1-4.
[15]. H. Ahmad, M. R. A. Moghaddam, H. Arof and S. W. Harun, IET Optoelectronics 5, 181-183 (2011).
[16]. Z. Abd Rahman, S. Hitam, M. H. Mansoori, A. F. Abas and M. A. Mahdi, Optics Express 19 21238-21245 (2011)
[17]. D. Derickson, “Fiber Optic Test and Measurement” (Ed. Upper Saddle River, NJ: Prentice-Hall, 1998), chapter 10.
[18]. J. Geng, S. Staines, Z. Wang, J. Zong, M. Blake and S. Jiang, IEEE Photonics Technology Letters 18 1813-1815 (2006)
[19]. B. Steinhausser, A. Brignon, E. Lallier, J. P. Huignard and P. Georges, Optics Express 15 6464-6469 (2007)
[20]. J. Geng, C. Spiegelberg and S. Jiang, IEEE Photonics Technology Letters 17 1827-1829 (2005)
[21]. C. Spiegelberg, J. Geng, Y. Hu, Y. Kaneda, S. Jiang and N. Peyghambarian, Journal of Lightwave Technology 22 57-62 (2004)
[22]. J. Geng, S. Staines, Z. Wang, J. Zong, M. Blake and S. Jiang, in Proc. Optical Fiber Communication Conf. (Anaheim, CA, 2006) paper OThC4.
[23]. S. Norcia, S. Tonda-Goldstein, D. Dolfi and J. P. Huignard, Optics Letters 28 1888-1890 (2003)
