We report a double-stage non-collinear Brillouin amplifier structure with high amplification and low noise, achieving an energy amplification of 6 × 1011 and a signal-to-noise ratio of 103 for an input signal of 5.5 × 10−14J in the regime above the pump’s stimulated Brillouin scattering threshold. The signal of the first-stage amplifier is efficiently amplified and separated from the noise output. The saturation amplification with noise suppressing is implemented in the second stage. The design principles of system parameters such as the intersection angle between the pump and signal beams, the pump energy, and the beam diameter are given.
© 2009 OSA
Brillouin amplification has been a subject of interest because of its ability to amplify weak signals with high gain [1–4] and strong signals with high efficiency . It has potential applications in lidar detection systems [6,7] and laser beam combination . In a Brillouin amplifier, spontaneous Brillouin scattering of a pump beam acts as a noise source, which will be amplified with nearly the same gain as a signal beam [3,4]. This amplified Brillouin scattering is much stronger than the Rayleigh scattering in the medium, and is considered as the main background noise of the amplifier. The noise output limits the amplifier’s performances such as signal amplification factor(SAF), signal-to-noise ratio(SNR), sensitivity and so on . To reduce the noise, the techniques by nonlinear postprocessing of the amplified signal and noise in a phase conjugate mirror(PCM) have been proposed . For a case in which the amplified signal was submerged in noise, the SNR was enhanced from 0.8 to 350 by this method, and the SAF of ~108 for an input signal of 1pJ was obtained. Based on the Ref , together with double-pass Brillouin amplification technique, an overall SAF of 3.75 × 1010 for an input signal of 0.2mW(~1pJ) was achieved . However, in these techniques, the amplified signals after processing are shifted in frequency with respect to the input signals because of introducing the PCM. On the other hand, in a collinear Brillouin amplifier, the amplified signal and noise are in the same direction, the pump has to be below or near the SBS threshold to obtain the highest SNR , but this condition limits the increase of the SAF.
In this paper, we experimentally demonstrate the high SAF and low noise achieved by the double-stage non-collinear Brillouin amplification. The aims to use non-collinear structure have three aspects. Firstly, the noise output of the first-stage amplifier can be spatially separated from the signal channel by choosing proper system parameters. This can avoid the second amplification of the first-stage noise. Secondly, the pump can exceed the SBS threshold to obtain high SAF. Thirdly, compared with the collinear structure, the experimental setup is simpler. Some optical elements such as polarizers and wave plates to extract beams are not necessary. We obtain an overall SAF of 6 × 1011 for an input signal of 5.5 × 10−14J. To our knowledge, this SAF based on Brillouin amplification is the maximum achieved among relevant experiments.
2. Experimental setup
The experiments were carried out with an injection seeded pulsed Nd:YAG laser (Continuum PowerliteTM PrecisionII 9010), which has an output at 532nm after frequency doubling, a pulse width of 5.5ns, repetition rate of 1Hz, linewidth of 90MHz, and beam diameter of 8mm. Figure 1 shows the experimental setup. S-polarized laser reflected by the beam splitter(BS) passes through a 3.5:1 compressing telescope system T1 to decrease the beam size. A half-wave plate and a polarizer P1 are used to control polarization and split the laser beam into two beams. The reflected beam is directed into the generator(BG) filled with CS2. The returning Stokes beam is used as the first-stage input signal whose energy is varied by calibrated neutral-density filters(ND). The transmitted beam from polarizer P1 is used to form the first-stage pump beam. These two p-polarized beams interact at an angle in the first-stage amplifier(BA1). The amplified signal is sent into the second amplifier(BA2) and used as the second-stage input signal. Through a half-wave plate and a polarizer P2, the beam transmitted by BS is changed into the p-polarized laser, which forms the second pump. They also non-collinearly interact in the BA2. The amplified signal energies of the first- and second-stage are detected at the positions ED1 and ED2. The SNR is determined by simultaneously measuring the amplified signal and near noise output intercepted by two apertures with the same size. Different energy detectors(Ophir PE50BB and Newport 818J-09) were used according to the values of laser energy. In the two stage amplifiers, CS2 was chosen as the nonlinear medium.
3. Design principles of system parameters and experimental results
We first study the separation conditions for the amplified signal and noise in the first-stage amplifier. The origin of the noise is mainly ascribed to spontaneous scattering occurring in a thin region located near the end of the amplifier medium. The noise output appears as a diverging speckle pattern. Its divergence angle(half-angle) is defined as the ratio of the radius of the pump beam to the cell length . If the intersection angle between the pump and signal beams is larger than the divergence of the noise, the amplified signal and noise will be separated, and hence the noise in the signal channel will be rapidly lowered. However, the intersection angle should be less than or near the angular bandwidth of the amplifier, or else, the SAF will decrease. In our experiment, the pump beam is condensed to decrease the noise divergence so that the amplified signal is easier to be separated from the noise output. The diameters of the pump and signal beams are 2.4mm and 1.5mm, respectively. The noise divergence is inversely proportional to the cell length, so the length of BA1 cannot be too short. Comprehensively considering the noise divergence and the absorption of the medium, we choose the length of BA1 as 40cm. Thus, the divergence angle of the noise can be calculated as 3mrad, whereas the experimental divergence is somewhat larger than the theoretical one because of the backscattering of the middle and posterior portions of the medium. Next we study the angular bandwidth of the amplifier under experimental conditions above mentioned. Figure 2 shows the SAF, defined as the amplified signal energy divided by the input energy, as a function of the intersection angle θ in the medium. The average energies of the pump and signal beams are 5mJ and 0.45pJ, respectively. To obtain higher amplification, the pump pulse lags the input signal pulse by 3ns, which is approximately equal to the half-width of the pump pulse . Experimental results show that the SAF keeps beyond 108 for angles in the range 0mrad up to 10mrad. Therefore, the angular bandwidth of the amplifier is approximately ± 10mrad. According to the separation conditions, the intersection angle is set to 10mrad.
We then consider the choice of the first-stage pump energy E P1. Figure 3 shows the dependence of the SAF on E P1 at θ = 10mrad for the input of 0.45pJ. It is seen that high SAF is obtained when E P1 is above the threshold value for SBS(Measured threshold: ~3mJ). Once E P1 exceed 5mJ, the SAF approximately keeps constant. This indicates that the amplifier’s efficiency(η = E Sout/E P1, where E Sout is the amplified signal energy)decreases as a result of the stronger competition from the pump self SBS. In the collinear amplifier, the pump intensity is usually set below or near the SBS threshold to optimize SNR but sacrifice SAF. While in the non-collinear geometry, the pump intensity can exceed threshold to achieve high SAF, since the noise and signal output are not in the same channel. In addition, the signal peak also arrived by an amount equal to pulse half-width before the pump. It will interact with the pump prior to the spontaneous Brillouin scattering in the BA1. These optimizations can also reduce the noise and raise the amplification .
Therefore, the pump energy is chosen as 5mJ. The SAF is measured by decreasing the input signal energy E Sin, as is shown in Fig. 4 . The first-stage SAF of 1.1 × 109 for the input of 5.5 × 10−14J is achieved. Output spot image of the amplifier is shown in Fig. 5 . We see that the amplified signal is separated from the noise. The right spot is the amplified signal output, and the left one is the noise output. The SNR is ~102. Under the conditions of the same pump and signal energies, we experimentally observed that for collinear structure(θ = 0°), the amplified signal was submerged in the noise output with the SNR of ~0.3.
When the first-stage input signal energy is 5.5 × 10−14J, its amplified output intercepted by an aperture with the diameter of 2mm has the energy of 64μJ. For the second-stage amplification, the intersection angle and the delay time of the pump are the same as those of the first stage. The length of the BA2 is 20cm. Figure 6 shows the dependence of the SAF on the second-stage pump intensity. We design three schemes by changing the diameters of the pump and signal beams. By comparison, we see that the SAF can be raised by increasing the pump diameter D P. Moreover, the signal diameter D S should be slightly less than D P. In the BA2, the signal is strong enough to suppress the amplification of spontaneous Brillouin scattering. However, when D S is a lot less than D P, a fraction of the pump cannot fully interact with the signal, leading to uncompleted suppression of noise. In addition, the pump intensity can exceed the SBS threshold, and saturation amplification is achieved. The overall SAF is 6 × 1011, the SNR is ~103.
Finally, we analyze the influence of the intersection angle on the frequency of the amplified signal. The Brillouin resonance shift ω B is a function of the intersection angle θ as 12]. For the narrow-linewidth input(i.e., the signal linewidth is much narrower than the Brillouin linewidth of the amplifier), the frequency of the amplified beam is consistent with that of the input signal, and is independent on the intersection angle. For the wide-linewidth input(the signal linewidth is near or above the Brillouin linewidth), the frequency shift between the amplified signal and the pump is equal to the Brillouin shift, and changes with the intersection angle. However, according to Eq. (1), when θ = 200mrad,the amplified beam is only shifted by 40MHz in frequency with respect to the input signal for CS2. Therefore, in the non-collinear structure, the input signal and its amplified beam have the approximately same frequency.
We have demonstrated that high amplification and low noise can be achieved by a double-stage non-collinear Brillouin amplifier, which has the advantages of simple structure, convenient adjustment, and being easy to implement. When the intersection angle between the pump and signal beams is larger than the noise divergence, and less than or near the angular bandwidth of the amplifier, the first-stage input signal is efficiently amplified and separated from the noise output. In the second-stage amplifier, the pump beam with larger diameter is used. The signal diameter is below and near the pump one, together with the stronger signal, the saturation amplification occurs and the noise is effectively suppressed. In two stages, the pump pulses should lag the signal pulses, and their intensities are set beyond the SBS threshold. For an input signal of 5.5 × 10−14J, an overall amplification of 6 × 1011 with 103 SNR is obtained. Moreover, the frequency of the amplified beam is the same as that of the input signal.
This work is supported by National Natural Science Foundation of China (Grant No. 60878005, 60778019), the Program of Science and Technology of Education the Bureau of Heilongjiang Province, China (Grant No. 11521048) and the Program of Excellent Team in Harbin Institute of Technology.
References and links
1. I. M. Bel’dyugin, V. F. Efimkov, S. I. Mikhailov, and I. G. Zubarev, “Amplification of weak Stokes signals in the transient regime of stimulated Brillouin scattering,” J. Russ. Laser Res. 26(1), 1–12 (2005). [CrossRef]
3. Y. Glick and S. Sternklar, “1010 Amplification and phase conjugation with high efficiency achieved by overcoming noise limitations in Brillouin two-beam coupling,” J. Opt. Soc. Am. B 12(6), 1074–1995 (1995). [CrossRef]
4. A. M. Scott, D. E. Watkins, and P. Tapster, “Gain and noise characteristics of a Brillouin amplifier and their dependence on the spatial structure of the pump beam,” J. Opt. Soc. Am. B 7(6), 929–935 (1990). [CrossRef]
5. Z. W. Lu, S. Y. Wang, and D. Y. Lin, “Investigation of strong signal Brillouin amplification when the intensity of Stokes beam higher than that of the pump,” Laser Part. Beams 26, 315–319 (2008).
7. W. Gao, Z. W. Lu, W. M. He, and Y. K. Dong, “High-gain amplification of weak Stokes signal of stimulated Brillouin scattering in water,” Acta Phys. Sin. 56, 2248–2252 (2008).
8. S. Y. Wang, Z. W. Lu, D. Y. Lin, D. Lei, and D. B. Jiang, “Investigation of Serial Coherent Laser Beam Combination Based on Brillouin Amplification,” Laser Part. Beams 25(01), 79–83 (2007). [CrossRef]
9. Y. Yamamoto and T. Mukai, “Fundamentals of optical amplifiers,” Opt. Quantum Electron. 21(1), S1–S14 (1989). [CrossRef]
10. S. Sternklar, Y. Glick, and S. Jackel, “Noise limitations of Brillouin two-beam coupling: theory and experiment,” J. Opt. Soc. Am. B 9(3), 391–397 (1992). [CrossRef]
11. Y. Glick and S. Sternklar, “Angular Bandwidth for Brillouin Amplication,” J. Opt. Soc. Am. B 11(9), 1539–1543 (1994). [CrossRef]
12. W. Gao, Z. W. Lu, W. M. He, Y. K. Dong, and W. L. J. Hasi, “Characteristics of amplified spectrum of a weak frequency-detuned signal in a Brillouin amplifier,” Laser Part. Beams (to be published).