Open Access
Add to BibSonomyAbstract
An all-optical locking technique without extra electrical feedback control system for a semiconductor laser has been used in stabilizing the laser frequency to a hyperfine crossover transition of 87Rb 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 with 1 MHz level accuracy. The optical feedback signal is generated from the narrow-band Faraday anomalous dispersion optical filter (FADOF) with nonlinear saturation effect. The peak transmission of the narrow-band FADOF corresponding to 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition is 18.6 %. The bandwidth is as wide as 38.9 MHz as the laser frequency changes. After locking, the laser frequency fluctuation is reduced to 1.7 MHz. The all-optical laser locking technique can be improved to much higher accuracy with increased external cavity length. The laser we have realized can provide light exactly resonant with atomic transitions used for other atom-light interaction experiments.
© 2013 Optical Society of America
1. Introduction
Since the first semiconductor laser diode came out in 1962, the significant development has occurred in the field of semiconductor laser. Because of their high efficiency, small size, low cost and tunability, semiconductor lasers are now used in various applications, such as optical communication [1], optical sensing systems [2], laser cooling [3] and atomic frequency standard [4–5]. However, in these applications, narrow linewidth and frequency-stabilized semiconductor lasers are needed. Normally, the emission linewidth of a bare laser diode is typically between 10–100 MHz. In order to reduce the laser diode emission linewidth, an external cavity has been coupled to the diode laser. With this method, The linewidth is narrowed and the frequency tuning of the doide laser is improved. Common external cavity diode lasers (ECDLs) in either Littrow or Littman-Metcalf configurations use diffractive gratings for frequency selection [6–7]. But these lasers need precise alignment and are very sensitive to the acoustic and mechanical vibrations, so additional frequency locking techniques should be used to stabilize laser frequency fluctuations. Besides, The laser frequency need to be stabilized to the atomic resonance line in atomic spectra experiments [8–9].
Here, we demonstrate a novel narrow-band all-optical locking technique to stabilize a semiconductor laser frequency by utilizing narrow-band Faraday anomalous dispersion optical filter (FADOF) with nonlinear saturation effect of a Rubidium (87Rb) atomic vapor operating at 780 nm. In this technique, the light possessing the information of the narrow-band FADOF atomic resonance spectral profile is injected into the laser. As we know, FADOF has advantages of high transmission, high noise rejection [10] and narrow bandwidth [11] and is critically important in fields as diverse as free space optical communication, lidar remote sensing systems [12–17] and so on. Besides, FADOFs are also very potential in the frequency stabilization field of ECDLs, which makes the laser immune to fluctuations of current and temperature [18–19]. However, because the transmission band of normal FADOFs in line-center opration is so wide, the laser frequency of ECDLs with normal FADOFs stabilizing [18–19] can’t be locked to the exactly atomic transition line since there is GHz level detuning from the transmission center to atomic lines.
In order to realize the narrow-band all-optical locking to the atomic resonance line exactly, FADOFs require not only the transmission spectroscopy resonant corresponding to the atomic transition [20–36], but also a bandwidth as narrow as possible, for example, a bandwidth approaching the natural linewidth at MHz level [37]. In this paper, a narrow-band nonlinear 87Rb FADOF with a bandwidth of 38.9 MHz and its transmission peak resonant with 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition at 780 nm has been realized. Then, the narrow-band FADOF with nonlinear saturation effect is used as a frequency discriminator to lock the laser frequency to the atomic transition through all-optical feedback. The ultimate stabilizing results are presented. Finally, we obtain a compact and stabilized laser working at atomic resonance line. Without the complex electrical feedback control system, this all-optical locking technique has advantages of accuracy, reproducibility and compactness.
2. Experimental schematics
Figure 1 shows relevant energy levels of 87Rb. The natural linewidth of 52P3/2 is 6 MHz. The experimental setup is shown in Fig. 2. A 5 cm long 96.5 % enriched isotope 87Rb cell is mounted in a magnetic shield box to reduce the interference of the earth magnetic field. The internal magnetic field is produced by permanent magnets and different magnetic field intensities are obtained by changing the distance of magnets. In the atom-light interaction area, the inhomogeneity of the magnetic field intensity can be ignored. The temperature of the 87Rb cell is controlled by temperature control system which consists of a heating wire with precision of 0.2 °C. ECDL represents a 780 nm external cavity diode laser which can be tuned to cover all of the 87Rb 52S1/2 → 52P3/2 transitions. We adopt EYP-RWE-0790-04000-0750-SOT01-0000 790 nm semiconductor LD. The collimation lens is 4.51 mm and the grating is 1800 lines/mm. The laser beam from 780 nm ECDL is evenly divided into two parts by a beam splitter (BS1). One is used for nonlinear saturated absorption spectra (SAS) as a standard frequency reference. The reference saturated absorption spectra of 87Rb is collected by a photodiode (PD1). The other is injected into the narrow-band FADOF system. In this system, the laser beam is divided into two parts by a polarization beam splitter (PBS). Through rotating the half-wave plate (HWP) before the PBS, we can change the intensity ratio between the two laser beams. Then, we use the strong one (1.28 mW) as a pumping laser while the weak one (0.58 mW) as a probe laser. The angle between pumping and probe beams is so small that the residual Doppler broadening is negligible.
Fig. 2 The all-optical locking experimental setup. ECDL, 780 nm external cavity diode laser; BS, beam splitter; PBS, polarization beam splitter; G, Glan-Taylor prism; R, high reflection mirror for 780 nm; HWP, half-wave plate; QWP, quarter-wave plate; M, mirror, the transmission is 20 %; PD, photodiode. SAS, saturated absorption spectra.
The narrow-band FADOF with nonlinear saturation effect are realized by the following process. The polarization of the probe beam is changed by passing through the 87Rb cell due to the circular dichroism induced in the 87Rb vapor by a nonlinear interaction with the strong circular polarization pump beam. Then, the rotated probe beam would pass through a pair of crossed Glan-Taylor prisms (G1 and G2) and is reflected by a mirror (M) with the transmission 20 %. For optical feedback, the pair of crossed Glan-Taylor prisms serves as an analyzer. The extinction ratio of the pair of crossed Glan-Taylor prisms is 1 × 10−5. Similar to the polarization spectroscopy, the Doppler-free narrow-band transmitted spectroscopy can be obtained. The transmission of FADOF is measured as the ratio of the maximum transmitted laser power when the two polarizers are perpendicular to the maximum transmitted laser power when the two polarizers are parallel. Finally, the optical feedback path is composed of M, G2, BS2, 87Rb cell, G1, R1, PBS, HWP, BS1 and ECDL.
3. Experimental results and discussion
3.1. Transmitted spectra of narrow-band FADOF with nonlinear saturation effect corresponding to atomic resonance line
Figure 3(a) shows the transmitted spectra of the narrow-band FADOF with nonlinear saturation effect at 52S1/2, F = 2 → 52P3/2, F′ = 1, 2, 3 transitions. The spectra signal is collected by PD2 by sweeping the laser frequency. The laser frequency is changed by sweeping the voltage of PZT. The magnetic field intensity is set to 11 G and the temperature of 87Rb cell is 85 °C. Here, the optimized magnetic field and temperature has been chosen to maximize the transmitted spectra. The maximum transmission is 18.4 % at the crossover transition F = 2 → F′ = 2, 3. In Fig. 3(a), the reflected probe beam is not injected into the laser, i.e., the laser is in the free-running condition.
Fig. 3 The transmitted spectral profile of 87Rb 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition with magnetic field intensity of 11 G and temperature of 85 °C. (a) The transmitted spectra of the narrow-band 87Rb FADOF with nonlinear saturation effect by using a free-running laser. The upper line is saturated absorption spectra of 87Rb and the bottom line is the transmitted spectra of the narrow-band 87Rb FADOF. (b) The profile measured by injecting the reflected probe beam into the laser. The upper line is the saturated absorption spectra of 87Rb and the bottom line is the transmitted spectra of 87Rb FADOF with nonlinear saturation effect by injecting the reflected probe beam into the laser. The optical path length between the laser diode and the mirror M is 1.2 m.
The labels in various peaks in Fig. 3(a) is determined by the upper saturated absorption spectra of 87Rb. As we know, The six peaks from left to right of the saturated absorption spectra of 87Rb correspond to transitions of F = 2 → F′ = 1, F = 2 → F′ = 1, 2, F = 2 → F′ = 1, 3, F = 2 → F′ = 2, F = 2 → F′ = 2, 3, F = 2 → F′ = 3, respectively. Spectra lines F = 2 → F′ = 1, 2, F = 2 → F′ = 1, 3 and F = 2 → F′ = 2, 3 correspond to the crossover transitions. The bandwidth of the spectral profile is also determined through the frequency interval between transitions F = 2 → F′ = 2 and F = 2 → F′ = 3, which is 266.65 MHz. Comparing to the SAS, the bandwidth of 87Rb spectral profile is 34.9 MHz at the crossover transition F = 2 → F′ = 2, 3 in Fig. 3(a). It’s much narrower than the normal FADOF [20–34].
Since the maximum spectral peak exhibits at crossover transition F = 2 → F′ = 2, 3, stable all-optical feedback locking is achieved by injecting the reflected probe beam into the laser through a mirror (M). As a result, the laser frequency is locked to the crossover transition F = 2 → F′ = 2, 3. Figure 3(b) shows the transmitted spectra of the narrow-band FADOF with nonlinear saturation effect measured as a result of the injection, i.e., the laser is in the all-optical feedback condition. The optical path length between the laser diode and the mirror M is 1.2 m. The transmission of the peak corresponding to the crossover transition F = 2 → F′ = 2, 3 is 17.5 %. The flat segments of the spectra at F = 2 → F′ = 2, 3 transition indicate the amount of the transmission remains nearly constant. Hence, the laser’s frequency is virtually unchanged. The bandwidth is as wide as 44.4 MHz at F = 2 → F′ = 2, 3 crossover transition as the laser frequency changes. Because the high reproducible atomic resonance line, the laser frequency can be fixed to the atomic resonance line every time the laser is locked.
The same result at the crossover transition F = 2 → F′ = 1, 3 is also achieved at the same experimental conditions. As is shown in Fig. 4(b), The transmission of the peak corresponding to the crossover transition F = 2 → F′ = 1, 3 is 20 % and the bandwidth is as wide as 46.4 MHz as the laser frequency changes.
Fig. 4 The transmitted spectral profile of 87Rb 52S1/2, F = 2 → 52P3/2, F′ = 1, 3 crossover transition with magnetic field intensity of 11 G and temperature of 85 °C. (a) The transmitted spectra of the narrow-band 87Rb FADOF with nonlinear saturation effect by using a free-running laser. The upper line is saturated absorption spectra of 87Rb and the bottom line is the transmitted spectra of the narrow-band 87Rb FADOF. (b) The profile measured by injecting the reflected probe beam into the laser. The upper line is the saturated absorption spectra of 87Rb and the bottom line is the transmitted spectra of 87Rb FADOF with nonlinear saturation effect by injecting the reflected probe beam into the laser. The optical path length between the laser diode and the mirror M is 1.2 m.
3.2. Frequency locking results with all-optical locking technique at different cavity lengths
In all-optical locking scheme, the external cavity length is the optical path length between the laser diode and the mirror M. Different external cavity lengths present different distributions of cavity modes. When the laser frequency coincides with a longitudinal mode of the cavity resonator, the laser frequency is automatically locked to the longitudinal mode while exhibiting a dramatic reducing in fluctuation of the laser frequency. Because of controlling the quantity of longitudinal modes inside the laser spectra, the free spectral range (FSR) of the cavity modes extremely influences the stabilized laser frequency. Besides, the FSR becomes smaller as the extension of the external cavity. Frequency locking results with all-optical locking technique at different external cavity lengths are presented in the following.
The estimation of the laser frequency accuracy of the all-optical locking can be carried out by the fluctuation of the laser frequency after all-optical locking. Firstly, the locking results of the 1.2 m external cavity length are shown in Fig. 5. the upper line is the transmitted spectra of narrow-band 87Rb FADOF with nonlinear saturation effect by injecting the reflected probe beam into the laser. The locking result is presented as the laser frequency is scanned across the atomic resonances. The peak transmission corresponding to F = 2 → F′ = 2, 3 crossover transition is 17.5 %. The bandwidth is as wide as 44.4 MHz as the laser frequency changes. The bottom line is the frequency locking spectra via optical feedback without frequency scanning. Transmission of the feedback light is exactly 17.5 %, which indicates that we have succeeded in locking the laser frequency to a hyperfine crossover transition of the 87Rb F = 2 → F′ = 2, 3 by using an all-optical feedback locking technique for a semiconductor laser. The standard deviation of the laser frequency fluctuation is 3.3 MHz corresponding that the standard deviation of the transmission fluctuation is 1.3 % while there is 17.5 % variation in the range of 44.4 MHz bandwidth.
Fig. 5 Frequency locking of the laser at 87Rb 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition with 1.2 m external cavity length with all-optical feedback technique. The upper line is the transmitted spectra of 87Rb FADOF with nonlinear saturation effect at 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition by injecting the reflected probe beam into the laser. The bottom line is frequency locking spectra via optical feedback.
As the FSR is 125 MHz for the 1.2 m cavity length, which is much greater than the bandwidth of the narrow-band FADOF transmission spectra 44.4 MHz, the longitudinal mode of the cavity resonator does not always coincide with the narrow-band FADOF transmission spectra, which induces the instability of the laser frequency. In order to achieve much more stable laser frequency, the external cavity length should be extended, which means the FSR of the cavity resonator becomes smaller. In principle, the stabilized laser frequency would be more stable while the FSR is less than the locking range of the FADOF transmission spectra. Hence, Considering our conditions, the external cavity length should be longer than 3.3 m at least.
Because of the limitation of the system volume, we extend the cavity length to 6 m. Figure 6 shows the frequency locking results of the 6 m cavity length. The upper line is the transmitted spectra of 87Rb FADOF with nonlinear saturation effect by injecting the reflected probe beam into the laser. The peak transmission corresponding to F = 2 → F′ = 2, 3 crossover transition is 18.6 %. The bandwidth is as wide as 38.9 MHz as the laser frequency changes. The bottom line is the frequency locking spectra via optical feedback without frequency scanning. Comparing with Fig. 5, the standard deviation of the laser frequency fluctuation is 1.7 MHz corresponding that the standard deviation of the transmission fluctuation is 0.79 %. It can be seen that the frequency locking effect of 6 m cavity length is much better than that of 1.2 m cavity length. The reason is that the FSR for 6 m cavity length is only 25 MHz, which is less than the bandwidth of the narrow-band FADOF transmission spectra, 38.9 MHz. The longer external cavity ensures that there is always a longitudinal mode of the cavity resonator coinciding with the narrow-band FADOF transmission spectra. Hence, the laser frequency can be locked within atomic resonance line by the transmitted spectra of narrow-band 87Rb FADOF under different conditions. In this case, extra electrical feedback control system, which is required for stable optical feedback by controlling the external cavity length between the laser and the mirror M, doesn’t need. Finally, the all-optical laser locking technique can be improved to much higher accuracy with increased cavity length.
Fig. 6 Frequency locking of the laser at 87Rb 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition with 6 m external cavity length with all-optical feedback technique. The upper line is the transmitted spectra of 87Rb FADOF with nonlinear saturation effect at 52S1/2, F = 2 → 52P3/2, F′ = 2, 3 crossover transition by injecting the reflected probe beam into the laser. The bottom line is frequency locking spectra via optical feedback.
4. Conclusion
In conclusion, we have succeeded in stabilizing the laser frequency to a hyperfine crossover transition of the 87Rb F = 2 → F′ = 2, 3 by using an all-optical locking technique without extra electrical feedback control system for a semiconductor laser at 1 MHz level accuracy. For the narrow-band FADOF with nonlinear saturation effect, The peak transmission corresponding to F = 2 → F′ = 2, 3 crossover transition is 18.6 %. The bandwidth is as wide as 38.9 MHz as the laser frequency changes. After locking the laser frequency, the laser frequency fluctuation is reduced to 1.7 MHz. The ECDL we have realized can provide light exactly resonant with atomic resonance transitions used for other atom-light interaction experiments.
In the future, the all-optical laser locking technique can be improved to much higher accuracy with increased external cavity length. An 100 m fiber external cavity is under developing to improve the absolute frequency accuracy of all optical laser locking to atomic resonance line. With the technique of averaging over fast scanning, kHz accuracy with all optical locking can be expected. The accuracy, reproducibility and compactness of the whole system are promising.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (Grant Nos. 10874009 and 11074011).
References and links
1. E. Ip, A. Lau, D. Barros, and J. Kahn, “Coherent detection in optical fiber systems,” Opt. Express 16, 753–791 (2008). [CrossRef] [PubMed]
2. M. Ohtsu and S. Araki, “Using a 1.5 μm DFB InGaAsP laser in a passive ring cavity type fiber gyroscope,” Appl. Opt. 26, 464–470 (1987). [CrossRef] [PubMed]
3. D. Sesko, C. G. Fan, and C. E. Wieman, “Production of a cold atomic vapor using diode-laser cooling,” J. Opt. Soc. Am. B 5, 1225–1227 (1988). [CrossRef]
4. K. Huang, J. Zhang, D. Yu, Z. Chen, W. Zhuang, and J. Chen, “Application of electron-shelving detection via 423 nm transition in Calcium-beam optical frequency standard,” Chin. Phys. Lett. 23, 3198–3201 (2006). [CrossRef]
5. J. J. McFerran and A. N. Luiten, “Fractional frequency instability in the 10−14 range with a thermal beam optical frequency reference,” J. Opt. Soc. Am. B 27, 277–285 (2010). [CrossRef]
6. K. C. Harvey and C. J. Myatt, “External-cavity diode laser using a grazing-incidence diffraction grating,” Opt. Lett. 16, 910–912 (1991). [CrossRef] [PubMed]
7. S. D. Saliba and R. E. Scholten, “Linewidths below 100 KHz with external cavity diode lasers,” Appl. Opt. 48, 6961–6966 (2009). [CrossRef] [PubMed]
8. C. E. Wieman and L. Hollberg, “Using diode lasers for atomic physics,” Rev. Sci. Instrum. 62, 1–20 (1991). [CrossRef]
9. B. MacAdam, A. Steinbach, and C. Wieman, “A narrowband tunable diode laser system with grating feedback, and a saturated absorption spectrometer for Cs and Rb,” Am. J. Phys. 60, 1098–1111 (1992). [CrossRef]
10. D. J. Dick and T. M. Shay, “Ultrahigh-noise rejection optical filter,” Opt. Lett. 16, 867–869 (1991). [CrossRef] [PubMed]
11. J. Menders, K. Benson, S. H. Bloom, C. S. Liu, and E. Korevaar, “Ultranarrow line filtering using a Cs Faraday filter at 852 nm,” Opt. Lett. 11, 846–848 (1991). [CrossRef]
12. J. Tang, Q. Wang, Y. Li, L. Zhang, J. Gan, and M. Duan, “Experimental study of a model digital space optical communication system with new quantum devices,” Appl. Opt. 34, 2619–2622 (1995). [CrossRef]
13. S. D. Harrell, C. Y. She, T. Yuan, D. A. Krueger, H. Chen, S. S. Chen, and Z. L. Hu, “Sodium and potassium vapor Faraday filters revisited: Theory and applications,” J. Opt. Soc. Am. B 26, 659–670 (2010). [CrossRef]
14. A. Popescu and T. Walther, “On an ESFADOF edge-filter for a range resolved Brillouin-lidar: The high vapor density and high pump intensity regime,” Appl. Phys. B 98, 667–675 (2010). [CrossRef]
15. Y. Yang, X. Cheng, F. Li, X. Hu, X. Lin, and S. Gong, “A flat spectral Faraday filter for sodium lidar,” Opt. Lett. 36, 1302–1304 (2011). [CrossRef] [PubMed]
16. J. Wang, J. Tang, M. Duan, and J. Li, “Reserch on the mechanism and performances of anomalous dispersion atomic optical filter with off-axis incident beam,” Acta Electron. Sin. 16, 18145–18151 (2008).
17. S. Liu, Y. Zhang, H. Wu, and D. Fan, “Gain assisted large-scale tunable atomic filter based on double selective optical pump induced dichroism,” Opt. Commun. 284, 4180–4184 (2011). [CrossRef]
18. K. Choi, J. Menders, P. Searcy, and E. Korevaar, “Optical feedback locking of a diode laser using a cesium Faraday filter,” Opt. Commun. 96, 240–244 (1992). [CrossRef]
19. X. Miao, L. Yin, W. Zhuang, B. Luo, A. Dang, J. Chen, and H. Guo, “Demonstration of an external-cavity diode laser system immune to current and temperature fluctuations,” Rev. Sci. Instrum. 82, 086106 (2011). [CrossRef]
20. H. Chen, C. Y. She, P. Searcy, and E. Korevaar, “Sodium-vapor dispersive Faraday filter,” Opt. Lett. 18, 1019–1021 (1993). [CrossRef] [PubMed]
21. H. Chen, M. A. White, D. A. Krueger, and C. Y. She, “Daytime mesopause temperature measurements with a sodium-vapor dispersive Faraday filter in a lidar receiver,” Opt. Lett. 21, 1093–1095 (1996). [CrossRef] [PubMed]
22. B. Yin and T. M. Shay, “A potassium Faraday anomalous dispersion optical filter,” Opt. Commun. 94, 30–32 (1992). [CrossRef]
23. E. T. Dressler, A. E. Laux, and R. I. Billmers, “Theory and experiment for the anomalous Faraday effect in potassium,” J. Opt. Soc. Am. B 13, 1849–1858 (1996). [CrossRef]
24. Y. Zhang, X. Jia, Z. Ma, and Q. Wang, “Optical filtering characteristic of potassium Faraday optical filter,” IEEE J. Quantum Electron. 37, 372–375 (2001). [CrossRef]
25. Y. Zhang, X. Jia, Z. Ma, and Q. Wang, “Potassium Faraday optical filter in line-center operation,” Opt. Commun. 194, 147–150 (2001). [CrossRef]
26. Z. Hu, X. Sun, X. Zeng, Y. Peng, J. Tang, L. Zhang, Q. Wang, and L. Zheng, “Rb 780 nm Faraday anomalous dispersion optical filter in a strong magnetic field,” Opt. Commun. 101, 175–178 (1993). [CrossRef]
27. M. Duan, Y. Li, J. Tang, and L. Zheng, “Excited state Faraday anomalous dispersion spectrum of rubidium,” Opt. Commun. 127, 210–214 (1996). [CrossRef]
28. Q. Sun, W. Zhuang, Z. Liu, and J. Chen, “Electrodeless-discharge-vapor-lamp-based Faraday anomalous-dispersion optical filter,” Opt. Lett. 36, 4611–4613 (2011). [CrossRef] [PubMed]
29. Q. Sun, Y. Hong, W. Zhuang, Z. Liu, and J. Chen, “Demonstration of an excited-state Faraday anomalous dispersion optical filter at 1529 nm by use of an electrodeless discharge rubidium vapor lamp”, Appl. Phys. Lett. 101211102 (2012). [CrossRef]
30. X. Xue, Z. Tao, Q. Sun, Y. Hong, W. Zhuang, B. Luo, J. Chen, and H. Guo, “Faraday anomalous dispersion optical filter with a single transmission peak using a buffer-gas-filled rubidium cell,” Opt. Lett. 37, 2274–2276 (2012). [CrossRef] [PubMed]
31. J. A. Zielińska, F. A. Beduini, N. Godbout, and M. W. Mitchell, “Ultranarrow Faraday rotation filter at the Rb D1 line,” Opt. Lett. 37, 524–526 (2012). [CrossRef]
32. B. Yin and T. M. Shay, “Theoretical model for a Faraday anomalous dispersion optical filter,” Opt. Lett. 16, 1617–1619 (1991). [CrossRef] [PubMed]
33. J. Menders, P. Searcy, K. Roff, and E. Korevaar, “Blue cesium Faraday and Voigt magneto-optic atomic line-filters,” Opt. Lett. 17, 1388–1390 (1992). [CrossRef]
34. B. Yin and T. M. Shay, “Faraday anomalous dispersion optical filter for the Cs-455 nm transition,” IEEE Photon. Technol. Lett. 4, 488–490 (1992). [CrossRef]
35. X. Sun, S. Wang, A. Chen, M. Zhao, and X. Zeng, “A fast efficient passive cesium ARF,” Opt. Commun. 111, 259–262 (1994). [CrossRef]
36. S. Liu, Y. Zhang, H. Wu, and P. Yuan, “Ultra-narrow bandwidth atomic filter based on optical-pumping-induced dichroism realized by selectively saturated absorption,” Opt. Commun. 285, 1181–C1184 (2012). [CrossRef]
37. Y. Wang, S. Zhang, D. Wang, Z. Tao, Y. Hong, and J. Chen, “Nonlinear optical filter with ultra-narrow bandwidth approaching the natural linewidth,” Opt. Lett. 37, 4059–4061 (2012). [CrossRef] [PubMed]
