Open Access
Add to BibSonomyAbstract
The frequency of a distributed feedback diode laser (DFB-LD) is stabilized on Cesium (133Cs) D2 saturated absorption lines by the polarization rotated optical feedback method (PROF). Different from the conventional frequency stabilization methods by adjusting the LD pump current, no extra electrical feedback is needed with the PROF. The self-homodyne beat spectra FWHM linewidth of the DFB laser is measured to be 1.1 MHz, greatly reduced by a factor of about 40 from its free-running linewidth of 44 MHz; and the optical frequency drift is reduced from 96 MHz down to 6.6 MHz.
© 2014 Optical Society of America
1. Introduction
High frequency stability and narrow linewidth semiconductor lasers play very important roles in many applications, such as laser cooling of atoms, coherent optical communications and high resolution spectroscopy [1–3]. However, the spectral purity of free-running semiconductor lasers is usually insufficient for such applications, even if the distributed feedback (DFB) lasers or distributed Bragg reflector (DBR) lasers seem relatively better. Many frequency stabilization and linewidth narrowing techniques have been proposed and adopted. Most of them use electrical negative feedback, i.e., taking error signals from some frequency discriminator and adjusting the pump current of the laser diode (LD) by electrical feedback [4,5]. In those methods modulation and demodulation and servo electronics are usually needed. In addition, extra intensity-modulated (IM) noise and linewidth broadening effects would be accompanied with the modulation. To avoid such shortcomings, modulation-free schemes showed their advantages [6,7]. References [8–11] proposed and demonstrated a different method, termed the “incoherent optical negative feedback”, which is based on the factor that the external optical injection will decrease the carrier density in LD active region and thus change the refractive index and the lasing frequency consequently. In the method the frequency discriminator should provide a positive slope feature (optical feedback power increases with increasing optical frequency) to lock the lasing frequency. Such an optical feedback should not interfere with the lasing mode, therefore the polarization of optical feedback is rotated to the orthogonal direction. The method is given a shortened term of PROF for conciseness [11]. Actually the method takes the LD as both a servo feedback unit and the controlled object and thus gets rid of all electrical feedback circuits. It is expected the optical feedback has merit of faster response than electrical feedback and good also for the frequency stabilization and linewidth narrowing.
In this paper, a simple and robust optical configuration which combines PROF and saturated absorption spectrum (SAS) is used to stabilize the frequency of a DFB laser on Cs-D2 hyperfine absorption line [12]. The self-homodyne beat spectra full width half maximum (FWHM) linewidth of the DFB laser is measured to be 1.1 MHz, greatly reduced from its free-running linewidth of 44 MHz by a factor of about 40, indicating the effect of linewidth reduction by PROF. Meanwhile, the optical frequency drift is reduced from 96 MHz down to 6.6 MHz.
2. Experiment setup
The experimental setup is shown in Fig. 1. A DFB diode laser emitting at about 852 nm (EYP-DFB-0852-00150-1500-TOC03-0002) is used in the experiment, which is driven by a commercial laser diode controller (ILX Lightwave, LDC-3714C) with the current resolution of 2 μA and noise of <1.5 μA. The laser is operated to output 10 mW power; the polarization extinction ratio (PER) of laser beam is measured to be 125:1 with a Glan lens, and its polarization direction is shown in Fig. 1. A Cs cell is inserted as the frequency discriminator, whose temperature is kept at about 35°C. Both the polarization rotated optical feedback and saturated absorption signal are simultaneously provided by a reflective mirror (about 50% reflectivity) and a quartz low-order λ/4 waveplate. The distance between the DFB laser and the reflective mirror is 12 cm. The feedback intensity can be optimized by adjusting the reflective mirror mechanically. A small portion of the feedback beam is split by a 45° glass plate between Cs cell and the laser to be used as a monitor of Cs absorption spectrum.
Fig. 1 Schematic setup of polarization rotated optical feedback experiment. Red single-arrow line: output beam; blue single-arrow line: feedback beam; Red double-arrow line: polarization direction of the output beam; Blue circle (with solid dot): polarization direction of the feedback beam.
Much attention should be paid to adjust the orientation of the λ/4 waveplate due to the inherent high sensitivity of DFB-LD to coherent optical feedback. Some kinds of instabilities would occur if the orientation of the λ/4 waveplate is not well adjusted, such as mod-hopping, square-wave oscillation phenomenon etc [13]. With the help of a Glan lens to detect the polarization state, the orientation of the λ/4 waveplate is set to ensure that the output beam of the reflective mirror is circularly polarized. Of course, high quality λ/4 waveplate, such as true zero-order λ/4 waveplate would be a better choice to improve the stability of the system.
3. Experimental results and discussion
Since the PROF is based on the dependence of refractive index on carrier density, the dependence of the lasing frequency of DFB laser on the optical injection is measured, and the polarization of the optical injection is perpendicular to the lasing mode. Another same type of DFB diode laser is used to measure the power tuning coefficient of the PROF. Figure 2 shows the experiment setup and results. In the experiment, the laser beam from DFB-LD2 with the polarization rotated 90° by the λ/2 waveplate is injected into DFB-LD1. The injection power into the DFB-LD1 is changed by gradually increasing the driving current of DFB-LD2. The frequency of DFB-LD1 output is measured by the wave-meter with resolution of 10 MHz. Figure 2(b) gives the measured frequency change with the injection power of DFB-LD2, which is measured in front of DFB-LD1. A linear fit result is obtained with a slope rate of −120 MHz/mW. Since the slope efficiency and current tuning coefficient of DFB-LD2 is about 0.9 mW/mA and −520 MHz/mA respectively, considering the reflective mirror, we can estimate that the slope rate of the injection power into DFB-LD1 is about −1.16 GHz/mW. Again, we change the injected power into the DFB-LD1 by blocking part of the injection beam (not by changing the driving current), and the frequency of DFB-LD1 will still be changed correspondingly, which indicates the tuning effect of DFB-LD1 is not related to the injection locking. It can be seen that the optical injection causes refractive index increase, implying decrease of carrier density, since the derivative of refractive index n over carrier density N is negative: .
In addition, the impact of injection power on the output power of DFB-LD1 is also quantified. We measure the optical power behind the reflective mirror (~50%), as shown in Fig. 2(a), and the output power of DFB-LD1 decreases approximately linearly with the increasing injection power. The linear fitting coefficient is about −0.028 mW/mW. However, when the DFB laser is frequency stabilized, this kind of power change would be very small.
The lasing wavelength of DFB-LD should be tuned to one edge of Cs absorption line with negative slope of , where α is the absorption coefficient and ν is the frequency, giving a positive transmission derivative. The saturated absorption spectrum of Cs cell is shown in Fig. 3(a), measured by the PD of Fig. 1 without optical feedback into the LD chip. The frequency scan is obtained by using a triangle waveform LD driven current; the curve is given by an oscilloscope and in the declined edge of scanning current; the tuning rate of lasing frequency is measured to be −550 MHz/mA beforehand. The DFB-LD is tuned around the crossover resonance peak Co35 or Co45 of the spectrum [12], and then the PROF operation is applied. It should be noted that the orientation of the low-order λ/4 waveplate need to be adjusted carefully, otherwise some kinds of instabilities would occur. Figure 3(b) gives the monitored signal by the PD when PROF works, with the laser pumped with the same triangle current as the case of Fig. 3(a). Compared with the absorption spectrum we can see that the positive slope side of the Co35 and Co45 peaks becomes gentler, implying the frequency stabilized; and very sharp down steps are found at the edge of frequency-stabilized region. Also, it can be seen that there is a current offset between the position of the SAS with and without PROF operation, which corresponds to about 130 MHz frequency offset from the figure. This frequency offset is induced by the optical feedback and basically agrees with the frequency change according to the measured power tuning coefficient.
The locking range and feedback bandwidth are important parameters for a frequency stabilization system and practical applications. We can make an estimate of the locking range of the PROF frequency stabilization method from the positive slope side of the Co35 or Co45 peak in Fig. 3(a). According to the current tuning coefficient, the locking range for Co35 and Co45 peaks are about 30 MHz and 45 MHz respectively. In principle, the feedback bandwidth of the method is determined by the round-trip time delay of the optical feedback and the response time of carrier effect [8,9]. For our system, the distance between the DFB laser and the reflective mirror is 12 cm, as the response time of carrier effect is shorter than the round-trip time delay, then high feedback bandwidth could be predicted in principle, which could be helpful for linewidth narrowing.
Since the laser frequency is locked at the positive slope edge of Cs absorption peak by PROF, the linewidth is expected to be narrowed effectively. Figure 4 shows the self-homodyne beat spectra of the DFB laser with frequency stabilized on Co35 by PROF and in its free-running for comparison. The length of the delay line is 1.2 km, which corresponds to a time delay of 6 μs and is enough for linewidth measurement of about 300 kHz. And the beat spectra is measured by a spectrum analyzer (Agilent, N9320A) with the sweeping time set at 500 ms and resolution bandwidth set at 10 kHz. It can be seen that the FWHM linewidth of the beat spectra is dramatically reduced from 44 MHz down to 1.1 MHz by a factor of 40. Since the DFB laser is operated at 1.15 times the threshold current, the free-running linewidth seems wider than its typical value. From another point of view, the effective linewidth reduction result indicates the PROF method has a relatively broad feedback bandwidth to suppress high frequency noise.
The frequency stability of the laser is measured by heterodyning with a reference laser whose frequency is stabilized on the Cs saturated absorption spectrum by a conventional current modulation method. The frequency fluctuation of the reference laser is measured to be 2.5 MHz by heterodyning with another same frequency stabilized DFB laser beforehand. Figures 5(a) and 5(b) show the beat frequency drift for the free-running laser and frequency fluctuations for the stabilized case with PROF. It can be seen that the drift read from the figure is lowered from 96 MHz of free-running down to 7.1 MHz of PROF case. Assuming the frequency fluctuation of the two stabilized DFB laser is uncorrelated and independent, then using statistical analysis, the residual frequency fluctuation is estimated to be 6.6 MHz when the fluctuation of reference laser is deducted.
Fig. 5 Frequency stability of the DFB laser in free-running condition (a) and by PROF frequency stabilization (b).
Long frequency stability is very important for practical applications. For our system, without external intervention, the DFB laser can stay locked for about 30 min. Therefore, detailed investigations about improving the frequency stability are needed surely. One of the important issues is whether and how the intensity fluctuation of the solitary DFB laser affects the frequency stabilization; because the polarization rotated optical feedback will fluctuate accordingly. Firstly, the output power drift behind the reflective mirror when the DFB laser is operated in free-running condition was measured to be 0.04 mW. Considering the transmittance of the experiment setup, the optical feedback fluctuation is estimated to be 0.018 mW, which corresponds to a frequency drift of about 2.1 MHz by considering the measured slope rate of −120 MHz/mW. The output power stability of the laser is the ultimate factor which determines the frequency stability of this method we proposed. No automatic power control based on pump current control can be made because changing the pump current will also change the frequency. Secondly, although a simplified and robust optical configuration is designed in this work, the mechanical stability of the setup needs further improvement, especially the stability of the feedback mirror, which is operated manually in the experiment. Effective vibration isolation should be adopted. Thirdly, it is reported that the linewidth of hyperfine peaks in the saturated absorption spectrum will be broadened when the intensity of pump beam passing through the Cs cell is too high [12]. The parameters of some optical components can be optimized, especially the reflectivity of feedback mirror. Further studies are being undertaken in our group. It is believed that better performances are expected after the improvement methods adopted.
4. Conclusions
The frequency stabilization and linewidth reduction of a DFB laser which combines both the polarization rotated optical feedback and saturated absorption spectrum is demonstrated. With a simple and easy aligned optical configuration, the frequency of DFB laser is stabilized on one of the hyperfine of D2-line of 133Cs saturated absorption spectrum. The self-homodyne beat spectra FWHM linewidth of the DFB laser is greatly reduced from 44 MHz to 1.1 MHz by a factor of 40, and the optical frequency drift is reduced from 96 MHz down to 6.6 MHz. The issues for further studies and improvements are discussed. Higher and longer frequency stabilities are predicted if several optimization solutions are taken.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (Grant No. 61108028, 61178031).
References and links
1. C. E. Wieman and L. Hollberg, “Using diode lasers for atomic physics,” Rev. Sci. Instrum. 62(1), 1–20 (1991). [CrossRef]
2. E. Ip, A. P. Lau, D. J. Barros, and J. M. Kahn, “Coherent detection in optical fiber systems,” Opt. Express 16(2), 753–791 (2008). [CrossRef] [PubMed]
3. S. Craft, A. Deninger, C. Trück, J. Fortágh, F. Lison, and C. Zemmermann, “Rubidium spectroscopy at 778–780 nm with a distributed feedback laser diode,” Laser Phys. Lett. 2(2), 71–76 (2005). [CrossRef]
4. M. Ohtsu and S. Kotajima, “Linewidth reduction of a semiconductor laser by electrical feedback,” IEEE J. Quantum Electron. 21(12), 1905–1912 (1985). [CrossRef]
5. M. Ohtsu and N. Tabuchi, “Electrical feedback and its network analysis for linewidth reduction of a semiconductor laser,” J. Lightwave Technol. 6(3), 357–369 (1988). [CrossRef]
6. F. Wei, D. Chen, Z. Fang, H. Cai, and R. Qu, “Modulation-free frequency stabilization of external-cavity diode laser based on a phase-difference biased Sagnac interferometer,” Opt. Lett. 35(22), 3853–3855 (2010). [CrossRef] [PubMed]
7. F. Wei, D. Chen, Y. Sun, Z. Fang, H. Cai, and R. Qu, “Modulation-free frequency stabilization based on polarization-split Sagnac loop,” IEEE Photon. Technol. Lett. 25(11), 1031–1034 (2013). [CrossRef]
8. H. Yasaka and H. Kawaguchi, “Linewidth reduction and optical frequency stabilization of a distributed feedback laser by incoherent optical negative feedback,” Appl. Phys. Lett. 53(15), 1360–1362 (1988). [CrossRef]
9. H. Yasaka, Y. Yoshikuni, and H. Kawaguchi, “FM noise and spectral linewidth reduction by incoherent optical negative feedback,” IEEE J. Quantum Electron. 27(2), 193–204 (1991). [CrossRef]
10. A. F. A. da Rocha, P. C. S. Segundo, M. Chevrollier, and M. Oriá, “Diode laser coupled to an atomic line by incoherent optical negative feedback,” Appl. Phys. Lett. 84(2), 179–181 (2004). [CrossRef]
11. K. Ying, D. Chen, H. Cai, and R. Qu, “Frequency stabilization of a DFB laser to molecular Cesium at 852 nm by polarization-rotated optical feedback,” CLEO, JW2A.87 (2012).
12. S. Ohshima, Y. Nakadan, and Y. Koga, “Spectral width of saturated absorption spectra of Cs with a laser diode,” IEEE J. Quantum Electron. 23(5), 473–475 (1987). [CrossRef]
13. D. Cheng, T. Yen, E. Liu, and K. Chuang, “Suppressing mode hopping in semiconductor lasers by orthogonal-polarization optical feedback,” IEEE Photon. Technol. Lett. 16(6), 1435–1437 (2004). [CrossRef]
