A novel method for generating a dark hollow beam (DHB) is proposed and studied both theoretically and experimentally. A coherent combination technique for laser arrays is implemented based on adaptive optics (AO). A beam arraying structure and an active segmented mirror are designed and described. Piston errors are extracted by a zero-order interference detection system with the help of a custom-made photo-detectors array. An algorithm called the extremum approach is adopted to calculate feedback control signals. A dynamic piston error is imported by LiNbO3 to test the capability of the AO servo. In a closed loop the stable and clear DHB is obtained. The experimental results confirm the feasibility of the concept.
©2010 Optical Society of America
In recent years, coherent beam combination (CBC) techniques of laser arrays have attracted much attention because of their success in saving beam quality where dynamic phase noise exists, such as in high-power systems with thermal-effect aberrations [1,2] or in long-distance propagation with atmospheric turbulence [3,4]. CBC techniques with an adaptive optics (AO) servo, such as a master oscillator power amplifier (MOPA) [1,5] and an optical phased array , have a powerful ability to correct phase noise in unit beams. Thus, an output laser could achieve high coherent efficiency and keep good beam quality.
The propagation properties of a laser array have been investigated comprehensively. The results indicate that some special beams, such as a dark hollow beam (DHB) , a flat-topped beam (FTB) , and a vortex beam [9,10] can be obtained by using a laser array with a suitable initial phase and amplitude distributions. Beams with different intensity profiles are useful in various fields––for example, in optical trapping and manipulations, particularly in biophysical science for trapping living cells and organelles [11,12], and in atomic physics for manipulating neutral atoms [13,14]. For such applications, Gaussian beams and FTBs are suited to trap micron-sized particles with a refraction index higher than that of the ambient index, whereas doughnut-shaped beams, such as DHBs and Laguerre–Gaussian beams, have to be adopted to trap particles with a refraction index lower than that of the ambient index. A potential application of DHBs carrying optical vortices propagating through atmospheric turbulence would be in optical communication, where the topological charge could be used as the information carrier . Actually, methods for generating different types of beams and beam shaping have already been proposed and studied extensively in areas such as holograms , diffractive optics , and specific waveguides . Most of these methods are often implemented with a single beam and/or fixed phase plate. The motivation for using AO and a laser array scheme is mainly because of the significant advantages in solving the problem of beam quality in a dynamic phase noise environment, which is unavoidable in some applications such as optical communications [10,15] and high-power systems [19,20]. A laser array can be regarded as multiple sub-apertures divided from a large single beam, and the correction of phase noise, including not only piston and small-tilt errors but also high-order phase aberrations, can be accomplished via the piston error correction of each unit beam by the AO servo [3,4].
In this paper, a novel method for generating a DHB is proposed. The concept of using a laser array and an AO servo is studied theoretically and experimentally in detail. In Section 2, the theoretical formulation is presented. The experimental configuration and results are given in Section 3 and Section 4, respectively. Finally, Section 5 summarizes the main results obtained in this paper.
2. Theoretical formulation
Consider the laser array shown in Fig. 1 . The unit beams are located symmetrically on rings, and the radius difference between two adjacent rings is equal to R. Each beam is a single-mode Gaussian laser. The complex amplitude of N unit beams on input plane is
The propagation of the beam passing through a first-order optical system parameterized by an ABCD transfer matrix is characterized by the Huygens–Fresnel diffraction integral . The field distribution of the unit beam on the output plane could be expressed asEq. (1) into Eq. (2) and after tedious integral calculations, the field distribution could be described asEquation (3) indicates that the Gaussian property of the unit beam will not be changed after propagating through an ABCD system without an aperture. The combined field is
The transfer matrix of the beam path shown in Fig. 2 could be calculated as
Consider the Fraunhofer diffraction in which the length parameters become . Equation (6) could be simplified to
If the field on the output plane is the DHB, the intensity of the center spot should be zero:Eqs. (4), (7), and (8), we can obtain the formula of the DHB field:
There are infinite solutions to this formula. Here we consider a special case in which seven unit beams are hexagonally distributed, as in the red part in Fig. 1. According to the symmetrical characteristics, the amplitude and phase of the six surrounding beams should be equal and are assumed to be and . Then Eq. (9) is simplified to
One solution is
It should be clarified that we only are concerned with one period, owing to the periodicity of phase. Now we can substitute Eq. (11) and (7) into Eq. (4) and compute the far field via numerical simulation. The related parameters refer to the experimental values: wavelength , focus length , waist diameter , and space between two rings . Figure 3 shows the simulation results:
In the far field the DHB appears, and the obvious dark area exists in the center. Furthermore, the influences of the amplitude and phase distributions are analyzed. Figure 4 shows the far fields deriving from different near fields. Respectively, Fig. 4(a) is in different amplitude distributions while phases keep , and Fig. 4(b) is in different phase distributions while amplitudes keep . The curves are profiles of the corresponding patterns.
From Fig. 4(a) we can conclude that the amplitude distribution would influence the intensity distribution of the center area. The center becomes darker while the ratio of and increases to the proper extent. A special case of should be noticed, in which the center is flat-topped. It indicates that the FTB could also be generated via the same scheme. Besides, Fig. 4(b) indicates that the phase distribution would influence the form of the far field. Particularly, according to [9,10], the vortex beams could be obtained with a suitable phase distribution. Discussion about this is beyond our topic, so we will not go into that detail. Here, the phase shift is adverse because it will destroy the dark-hollow character. Therefore, when the phase noise is dynamic, an AO servo is necessary for phase-locking. Experimentally, the amplitudes of laser arrays are correspondingly stable; thus, the stabilization of is the main purpose and challenge. We design the zero-order interference detection method to extract real-time piston errors by using a reference beam. When the reference beam superposes the unit beam coaxially without tilt error, a thorough combination will be achieved and a zero-order interference fringe will appear. The complex amplitude of j th fringe is
Then the intensity is
Equation (13) indicates that the intensity fluctuation of the fringe corresponds with the piston error between two beams. Especially, when the piston error is zero the fringe is brightest, and when it is π the fringe is darkest. Therefore, to achieve the phase-locking of , the AO servo should stabilize the center fringe in the brightest status and the surrounding sixes in the darkest. Furthermore, the taking place of the zero-order interference phenomenon requires constraint of the tilt angle, which is
In our experiment, this value should be , which means that the tilt error is hardly acceptable. Precise adjustment of the beam path is required to eliminate the tilt errors.
3. Experimental configuration
Figure 5 shows a block diagram of the experimental configuration. All of the fiber-connected devices are polarization-maintaining so that interference efficiency could be guaranteed. The master oscillator is a semiconductor laser generator protected by an isolator. It provides the single-mode laser with a 1064 nm wavelength and a linewidth less than 1 MHz. The output power is adjustable from 0 to 150 mW. The beam is split by splitter1. One is for producing the seven unit beams via splitter2, and another is for producing the reference beam. The intensity adjuster in each channel is employed to control the amplitude distribution. Then the seven fiber lasers are linked to the AO servo with a beam-arraying structure (BAS) and an active segmented mirror (ASM). These two parts are adopted to array the unit beams into a hexagonal distribution and to provide proper phase modulation. The output beam is divided into two parts by BS1. One is for the sample system. The beam is focused by Lens1 (), and a high-speed CMOS is employed to record the far field. A 4× micro-objective lens is employed to expand the pattern size to be suitable for the detecting area. The second part is for the zero-order interference detection system, which can extract the piston errors with the help of the custom-made photo-detectors (PDs) array. The last part behind BS2, with a camera and monitor, helps to solve the difficulty in the manual adjustment of the beam expander and BAS. The tilt errors will be eliminated once the reference beam and the unit beams are confocal on the focus plane of Lens2.
3.1 The experimental AO servo
Figure 6 shows the main devices of the AO servo. The ASM, shown in Fig. 6(a), is composed of seven sub-mirrors with hexagonal distribution. Each sub-mirror is round in shape with a 16 mm diameter and is supported by three piezoceramic actuators. They are driven by a high-voltage amplifier (HVA) and can supply displacement. The sub-mirror can modulate the reflected beam with 3 degrees of freedom, including the piston and tilts. In our current experiment, three actuators are controlled by the same voltage, so the mirror moves vertically and only the piston modulation is implemented. Figure 6(b) shows a picture of the BAS. It is composed mainly of the collimators, mounting brackets, and a hexagonal pyramid. The waist diameter of the Gaussian beam emitted from the collimator is 10 mm. The collimators are held by the mounting brackets with 4 degrees of freedom. The glass hexagonal pyramid is axis-drilled and polished, with a wavelength reflective coating of 1064 nm (>99%). Six beams are reflected by the pyramid, and the 7th beam goes through the axis in the fashion of Fig. 6(c). All of the parts are assembled on an aluminum supporting plate and adjusted very carefully to assure that the seven beams are horizontal and coaxial.
3.2 The feedback control loop of AO servo
The feedback control loop of an AO servo is mainly composed of a piston extracting system and a PC-based algorithm platform. The piston extracting system is designed according to a zero-order interference phenomenon. Figure 7 shows the main devices of it. An adjustable varifocal telescope is employed to be the beam expander, as shown in Fig. 7(a). It can expand the collimated beam into the reference beam with a large diameter (D> = 50 mm) and can be precisely adjusted via the multi-degrees-of-freedom mounting bracket. A LiNbO3 phase modulator (PM) and function generator are equipped in this channel to import phase noise. Although the unit beams are still in a stable condition, the dynamic piston error in the reference channel could completely represent the actual case in which unit beams are dynamic, whereas the reference is stable. For an experimental test of the performance of the AO servo, there is no difference between these two cases, thanks to the relativity of the piston phase. The seven zero-order fringes then irradiate on the PDs array, whose photosensitive areas are arranged to correspond with the distribution of the laser array, shown in Fig. 7(b). The size of the sub-area is . The intensity of each fringe could be detected individually and quantified in the form of voltage signals by the gain-adjustable amplifying circuit. These signals are acquired by an AD card on the PC platform. The feedback control signals are calculated by an extremum approach (EA) algorithm and exported to the HVA by a DA card.
The EA algorithm works as in the following steps:
- (1) Generate a sine wave and export it to the ASM to sweep the phase. The intensities of the fringes will fluctuate. Record the corresponding voltage curves. After adequate periods, the maximum and minimum voltage of each channel can be identified.
- (2) Set the voltage standards to and . The values are 90% of the maximum voltage and 110% of minimum voltage of j th channel, respectively.
- (3) Acquire the real-time signal from the PD arrays at the k th iteration, , which reflects the piston error between j th beam and the reference.
- (4) The feedback control voltages are calculated by:
Particularly, while ,, else ; that means the algorithm implements the maximization in the center channel and the minimization in the surrounding six channels. Therefore, once the algorithm is enabled, the phase-locking of could be accomplished. is the proportion factor that influences the control effect.
4. Experimental investigation
The first step of our experiment is the calibration of the PD arrays. We close the unit channels so the device is irradiated by the reference beam only. Then we adjust the gain knobs of the amplifying circuit to make the voltage signal from each sub-PD be equal. Then we close the reference and open the unit beams to adjust the amplitudes. As the required distribution is , the intensity of the six surrounding beams should be 2.9% of the center. Finally, we stop the intensity adjustment, open all the channels, and adjust the gain knobs once again to make the fluctuation of voltage signals be suited to the measurement range of the AD card (1~4v). This is beneficial in order to improve the precision and stability of the AO servo.
The second step is to simulate the dynamic environment via LiNbO3 PM. The function generator exports a sine wave to the PM, as shown in Fig. 8(a) . The amplitude and frequency are V and 170 Hz. As the half-wavelength voltage of the PM is 2 V, the amplitude of the piston error is .
The AO correcting effect can be recognized from Figs. 8(b) and 8(c), which record the voltage signal of the center channel () in open and closed loops, respectively. One figure is adequate because the other channels obey the same rule, thanks to the parallel character of the feedback. The figure indicates that the AO servo successfully stabilized the intensity of the zero-order fringe in the maximum (or minimum) phase. To extract the piston error from the voltage signal, we can review Eq. (13) and find
So the intensity of zero-order fringe can be expressed as
As the voltage signal is identical to the intensity, the piston extracting formula is
Depending on Eq. (18) we can obtain the evolution curves of piston errors in a π period, shown in Fig. 9 . When the control loop is open, the piston errors between the reference beam and the unit beams dither rapidly and randomly. When the loop is closed, the piston phase of the center channel is locked to be synchronized with the reference channel [Fig. 9(a)]; in the meanwhile, the six surrounding channels are locked with the π piston shift [Fig. 9(b)]. The AO servo works effectively against dynamic phase noise.
To demonstrate the capability of the AO servo, we calculate the phase spectral density (PSD), shown in Fig. 10(a) . The curves are obtained by applying the Fourier transform to the piston errors. The peak on 170 Hz is owing to the imported phase noise. The second peak on 30 Hz, which could be also detected in the other non-presented PSD curves, is ascribed to the circuit noise in the HVA. In addition, the error rejection transfer function (ERTF) in Fig. 10(b) reveals the control bandwidth. The curve is obtained by
The zero passage frequency, which is approximately 100 Hz, presents the valid correction bandwidth of the AO servo. Furthermore, the characteristic of the frequency component is similar to that of systems with power amplifiers [5,22] or atmospheric turbulence [23–25], which means that our AO servo is also competent for those usages.
Figure 11 shows the 10 s long-exposure patterns of the far field in open and closed loops. Each pattern is the average of more than 20000 real-time frames recorded by the CMOS. When the loop is open, the far field displays random interference without any particularity. When the loop is closed, the DHB appears, and the intensity distribution is similar to the theoretical simulation. The beam quality is rather good. Nevertheless, the weakened symmetry is owing to the residual errors caused by the reasons such as the limited precision of intensity adjusters and detectors, maladjustments of high-frequency components, errors in polarization maintaining, and so on.
In this paper, a novel method for generating a DHB from a laser array has been proposed and studied, both theoretically and experimentally, for the first time to our knowledge. It also has been demonstrated that a CBC with an AO servo is a powerful technique for beam shaping to eliminate the influence of dynamic phase noise in some applications. A variety of beam profiles such as FTB and vortex beams could also be realized by using such a concept, which deserves further study.
References and links
1. G. D. Goodno, H. Komine, S. J. McNaught, S. B. Weiss, S. Redmond, W. Long, R. Simpson, E. C. Cheung, D. Howland, P. Epp, M. Weber, M. McClellan, J. Sollee, and H. Injeyan, “Coherent combination of high-power, zigzag slab lasers,” Opt. Lett. 31(9), 1247–1249 (2006). [CrossRef] [PubMed]
3. L. A. Beresnev, T. Weyrauch, M. A. Vorontsov, L. Liu, and G. W. Carhart, “Development of adaptive fiber collimators for conformal fiber-based beam projection systems,” Proc. SPIE 7090, 709008 (2008). [CrossRef]
4. H. Bruesselbach, S. Q. Wang, M. Minden, D. C. Jones, and M. Mangir, “Power-scalable phase-compensating fiber-array transceiver for laser communications through the atmosphere,” J. Opt. Soc. Am. B 22(2), 347–353 (2005). [CrossRef]
6. W. M. Neubert, K. H. Kudielka, W. R. Leeb, and A. L. Scholtz, “Experimental demonstration of an optical phased array antenna for laser space communications,” Appl. Opt. 33(18), 3820–3830 (1994). [CrossRef] [PubMed]
7. B. Lü and H. Ma, “Beam propagation properties of radial laser array,” J. Opt. Soc. Am. A 17(11), 2005–2009 (2000). [CrossRef]
8. K. Gao, L. Xu, R. Zheng, G. Chen, H. Zheng, and H. Ming, “Flat-top laser beam generated by coherent beam combining of Gaussian laser,” Chin. Opt. Lett. 8(1), 45–47 (2010). [CrossRef]
9. L. G. Wang, L. Q. Wang, and S. Y. Zhu, “Formation of optical vortices using coherent laser beam arrays,” Opt. Commun. 282(6), 1088–1094 (2009). [CrossRef]
10. J. Li, H. Zhang, and B. Lü, “Composite coherence vortices in a radial beam array propagating through atmospheric turbulence along a slant path,” J. Opt. A, Pure Appl. Opt. 12(6), 065401 (2010).
11. K. T. Gahagan and G. A. Swartzlander, “Trapping of low-index microparticles in an optical vortex,” J. Opt. Soc. Am. B 15(2), 524–534 (1998). [CrossRef]
13. J. P. Yin, Y. F. Zhu, W. J. He, and Z. Z. Wang, “Atom guiding and cooling in dark hollow laser beam,” Phys. Rev. A 58(1), 509–513 (1998). [CrossRef]
14. Z. Y. Wang, Y. M. Dong, and Q. Lin, “Atomic trapping and guiding by quasi-dark hollow beam,” J. Opt. A, Pure Appl. Opt. 7(3), 147–153 (2005). [CrossRef]
15. G. Gbur and R. K. Tyson, “Vortex beam propagation through atmospheric turbulence and topological charge conservation,” J. Opt. Soc. Am. A 25(1), 225–230 (2008). [CrossRef]
16. J. Arlt, K. Dholakia, L. Allen, and M. J. Padgett, “The production of multiringed Laguerre–Gaussian modes by computer-generated holograms,” J. Mod. Opt. 45(6), 1231–1237 (1998). [CrossRef]
21. S. A. Collins Jr., “Lens-system diffraction integral written in terms of matrix optics,” J. Opt. Soc. Am. 60(9), 1168–1177 (1970). [CrossRef]
22. R. Yang, P. Yang, L. Dong, M. Ao, and B. Xu, “A strip extracting algorithm for phase noise measurement and coherent beam combining of fiber amplifiers,” Appl. Phys. B 99(1-2), 19–22 (2010). [CrossRef]
23. X. L. Wang, Y. X. Ma, P. Zhou, H. T. Ma, X. Li, X. X. Xu, and Z. J. Liu, “Coherent beam combining of two W-level fiber amplifiers in turbulence atmospheric environment based on stochastic parallel gradient descent algorithm,” Laser Phys. 19(5), 984–988 (2009). [CrossRef]
25. L. Liu, M. A. Vorontsov, E. Polnau, T. Weyrauch, and L. A. Beresnev, “Adaptive phase-locked fiber array with wavefront phase tip-tilt compensation using piezoelectric fiber positioners,” Proc. SPIE 6708, 67080K (2007). [CrossRef]