We investigate the finite power Airy beams generated by finite extent input beams such as a Gaussian beam, a uniform beam of finite extent, and an inverse Gaussian beam. Each has different propagation behavior: A finite Airy beam generated by a uniform input beam keeps its Airy profile much longer than the conventional finite Airy beam. Also, an inverse Gaussian beam generates a finite Airy beam with a good bent focusing in free space. In this paper, the analysis and experimental results of finite Airy beams are presented.
© 2013 Optical Society of America
Since Berry and Balazs predicted the existence of non-spreading wave packets  and Siviloglou and Christodoulides studied their optical version , there have been very active researches on Airy beams [3–18]. To sum up their outcomes until now, we can present three unique characteristics of Airy beams. First, Airy beams are non-diffracting ones that satisfy the Helmholtz equation, like Bessel  and parabolic beams . Second, Airy beams take up a bending trajectory in homogeneous media without any external forces. Lastly, Airy beams can recover their original shapes when a part of them is blocked by arbitrary obstacles.
To resolve the problem that ideal Airy beams are not square integrable (i.e., they carry infinite power), exponentially decaying terms are introduced to implement finite power Airy beams , which have similar features to those of the ideal one. They were generated by the optical Fourier transform of a Gaussian input beam on which cubic phase is imposed . Although they can carry finite power, these apodized finite power Airy beams cannot maintain their shapes for a long time and are gradually spreading out during propagation . Recently, based on these, various methods which tune a main lobe or side lobes of finite power Airy beams by using a flat-topped input Gaussian beam , a nonsymmetric apodization  and a sharp cutoff  were presented.
In this paper, we present some alternative methods for the apodization of Airy beams. Instead of Gaussian input beams (that involve exponentially decaying terms), we consider uniformly distributed beams of finite extent or beams having an inverse Gaussian distribution. In the case of a uniformly distributed finite-extent beam, we show that the resultant finite power Airy beam can preserve its Airy profile much farther than that generated by the Gaussian input beam. An inverse Gaussian apodization results in a unique propagation dynamics – a focused-bending beam. In each case, the solution of the finite power Airy beam is derived and verified by experiments.
2. Theoretical analysis of Airy beams
The (1 + 1)D potential-free paraxial Helmholtz equation can be written asEq. (1), the Airy beam solution can be obtained as follows :Eq. (2) at ξ = 0 is given byEq. (2) is not square integrable and thus is not physically realizable. Usually, an exponentially decaying function exp(as) is multiplied to to obtain a finite power beam:Eq. (5) and its optical Fourier transform, the first observation of finite Airy beams was conducted: the cubic phase [ ] was imposed on the Gaussian beam  using a spatial light modulator (SLM) and the resultant, cubic phase-modulated Gaussian beam was Fourier transformed using a lens system . If we ignore the higher-order terms of the relatively small constant a, in Eq. (5), the only difference between Eqs. (3) and (5) is the Gaussian function [exp(-ak2)] which is originated from the apodization, i.e., exponentially decaying term exp(as). Therefore, Eq. (3) can be understood as a plane wave (of infinite extent) with a cubic phase modulation, suggesting that the finite-extent feature of the input beam results in the finite power Airy beam.
This discussion, i.e., a finite power Airy beam can be generated by the Fourier transform of a cubic phase-modulated beam of finite extent, implies that there can be other methods for the apodization of ideal Airy beams. That is, we can use other types of finite-extent beam than the Gaussian one. Here, we consider finite power Airy beams generated by three different input beams: those having a conventional Gaussian distribution (CASE I), a uniform distribution of finite extent (CASE II), and an inverse Gaussian distribution (CASE III). Throughout this paper, we assume that the wavelength of an incident beam λ is 633 nm, x0 is 50 μm, the focal length of the lens f is 50 cm, and the SLM has 1080 × 1080 pixels with an 8 μm pixel pitch.
First, we start with the CASE I, which adopts a Gaussian beam as an input beam. Due to the finite SLM size, the incident Gaussian beam is truncated. Therefore, Eq. (5) becomesEq. (4):Eq. (7) are plotted in Fig. 1(a). In this case, the number of side lobes in the initial plane (z = 0 cm) is decreased because the Gaussian distribution of the input beam cannot fully retain the high spatial frequency components. As a result, unlike the ideal case shown in Fig. 1(d), the Airy beam is diffracted or spreads out due to insufficient power flows from the side lobes.
Second, under the same conditions, we adopt a beam with a uniform intensity distribution of finite extent or a truncated plane wave as an input beam (the CASE II), In this case, we have
From the property of the Fourier transform, the (inverse) Fourier transform of Eq. (8) is a convolution of the sinc function and the Airy function as follows:Fig. 1(b). In Fig. 1(b), we can observe more side lobes at the initial plane compared with the CASE I. Figure 1(b) also shows that the incidence of a truncated plane wave and its Fourier transform after the cubic phase modulation can generate a finite power Airy beam which takes up a bending trajectory with the acceleration toward + x direction. What is interesting is that this finite Airy beam can preserve its Airy profile much farther than that generated by the Gaussian input beam. This is because more high-frequency components are retained in Eq. (8) than in Eq. (6).
Lastly, in the CASE III, we use a beam with an inverse Gaussian intensity distribution as an input beam. We haveFig. 1(c), the main lobe is suppressed at the initial plane because the inverse Gaussian distribution can retain only high spatial frequency components. However, as can be found in Fig. 1(c), the resultant finite power beam generates the main lobe after some propagation distance and is accelerated along the + x direction. That is, Eq. (11) also describes a finite power Airy beam. Actually, this case can be taken as an extreme example of the self-healing property : side lobes without the main lobe can regenerate the Airy profile during the propagation.
To compare the propagation characteristics of these three finite power Airy beams, their cross-correlations with the ideal Airy beam are calculated along the propagation direction z. In Fig. 2, we plotted the variation of the maximum value of this cross-correlation, i.e., C1(z) defined byFig. 2 are normalized by those at the initial plane (z = 0 cm). For the CASE I and II, the cross-correlation becomes maximum at the initial plane and gradually decreases. This means that the finite power Airy beams spread out or are diffracted so that they lose their initial Airy shape during propagation. However, we would like to point out that the cross-correlation value of the CASE II Airy beam is always higher than that of the CASE I at every z. This means that the CASE II Airy beam can preserve its original shape much farther than the CASE I Airy beam. In the CASE III, however, the cross-correlation increases after some propagation distance. This indicates the recovery process of the Airy profile: the regeneration of the main lobe and its acceleration along the transverse coordinate.
Calculated 2D intensity distributions of (2 + 1)D finite power Airy beams are shown in Figs. 3(a) and 3(b) (; CASE I), Figs. 3(c) and 3(d) (; CASE II), and Figs. 3(e) and 3(f) (; CASE III). Figures 3(a), 3(c), 3(e) and 3(b), 3(d), 3(f) are the results at z = 0 cm and z = 15 cm, respectively. Comparing Figs. 3(a), 3(c) and 3(e), we can find more side lobes in Fig. 3(c) than in Fig. 3(a) while the main lobe disappears in Fig. 3(e). After some propagation (z = 15 cm), the finite power Airy beam generated by a Gaussian beam (CASE I) does not preserve its initial Airy profile anymore. On the other hand, the finite power Airy beam generated by a truncated plane wave (or a uniform beam of finite extent; CASE II) maintains its initial profile although the beam is broadened due to the diffraction. Therefore, the uniform beam of the CASE II is more advantageous than the Gaussian beam of the CASE I. In the case of the inverse Gaussian beam (CASE III), although the main lobe is missing at the initial plane (Fig. 3(e)), side lobes recover the main lobe after some propagations as can be found in Fig. 3(f).
3. Experiments of the finite power Airy beams
Figures 4(a) and 4(b) show the schematic diagram and the experimental setup for the generation of finite power Airy beams. A laser beam with a 633 nm wavelength is expanded to form a collimated plane wave. To change this input wave into the Gaussian, inverse Gaussian, or uniform beam of finite extent, the SLM 1 (Mitsubishi Electric XL9U LCD projector with 1024 × 768 pixels of 11.9 μm pixel pitch) is placed between two orthogonal linear polarizers (P1 and P2) . These polarizers allow the SLM 1 to perform intensity modulations. The beams from the SLM 1 have a Gaussian distribution (CASE I), a uniform distribution of finite extent (CASE II), or an inverse Gaussian distribution (CASE III). Passing these beams through the optical Fourier transform (2-f) system, we can obtain finite power Airy beams at the initial plane (z = 0 cm). The optical Fourier transform system consists of a lens (L1; f = 50 cm) and a phase-only SLM 2 (Holoeye Pluto with 1920 × 1080 pixels of 8 μm pixel pitch) which is placed in front of the lens and imposes cubic (k3) phase. The same cubic phase mask is used to all CASES because the higher-order phase terms can be ignored due to the relatively small constant a. Another polarizer P3 is used to attenuate the output beam so that a twin image can be eliminated.
The images of the finite power Airy beams were captured by the charge-coupled device (CCD) at z = 0 cm and z = 15 cm and shown in Figs. 5(a) - 5(f). All images are obtained under the same conditions that no adjustments in intensities were made. From these results, we can conclude that the CASE II Airy beam [Figs. 5(c) and 5(d)] retains the Airy profile much longer than the CASE I Airy beam [Figs. 5(a) and 5(b)]. Meanwhile, the CASE III Airy beam recovers the main lobe at z = 15 cm as shown in Figs. 5(e) and 5(f). These experimental results coincide well with the calculation results shown in Figs. 3(a) - 3(f).
In this paper, finite power Airy beams generated by the Fourier transform of a cubic phase-modulated beam of finite extent such as a Gaussian beam (CASE I), a uniform beam of finite extent (CASE II), and an inverse Gaussian beam (CASE III) were discussed. The propagation dynamics of resultant finite power Airy beams were analyzed and compared. We showed both theoretically and experimentally that the finite Airy beam generated by the use of a uniform input beam (CASE II) retains the Airy profile much longer than the conventional finite Airy beam (CASE I). Also, the finite Airy beam via an inverse Gaussian beam (CASE III) builds up a focused-bending beam. We expect that our works in this paper can be utilized to particle tweezing and optical trapping [4–6].
This work was supported by the National Research Foundation of Korea grant funded by the Korean government (MSIP) through the National Creative Research Initiatives Program (#2007-0054847).
References and links
1. M. V. Berry and N. L. Balazs, “Nonspreading wave packets,” Am. J. Phys. 47(3), 264–267 (1979). [CrossRef]
4. J. Baumgartl, M. Mazilu, and K. Dholakia, “Optically mediated particle clearing using Airy wavepackets,” Nat. Photonics 2(11), 675–678 (2008). [CrossRef]
5. D. N. Christodoulides, “Optical trapping: riding along an Airy beam,” Nat. Photonics 2(11), 652–653 (2008). [CrossRef]
8. P. Polynkin, M. Kolesik, J. V. Moloney, G. A. Siviloglou, and D. N. Christodoulides, “Curved plasma channel generation using ultraintense Airy beams,” Science 324(5924), 229–232 (2009). [CrossRef] [PubMed]
9. T. Ellenbogen, N. Voloch-Bloch, A. Ganany-Padowicz, and A. Arie, “Nonlinear generation and manipulation of Airy beams,” Nat. Photonics 3(7), 395–398 (2009). [CrossRef]
10. A. Chong, W. H. Renninger, D. N. Christodoulides, and F. W. Wise, “Airy–Bessel wave packets as versatile linear light bullets,” Nat. Photonics 4(2), 103–106 (2010). [CrossRef]
13. A. Minovich, A. E. Klein, N. Janunts, T. Pertsch, D. N. Neshev, and Y. S. Kivshar, “Generation and near-field imaging of Airy surface plasmons,” Phys. Rev. Lett. 107(11), 116802 (2011). [CrossRef] [PubMed]
14. Y. Lim, S.-Y. Lee, K.-Y. Kim, J. Park, and B. Lee, “Negative refraction of Airy plasmons in a metal–insulator–metal waveguide,” IEEE Photon. Technol. Lett. 23(17), 1258–1260 (2011). [CrossRef]
15. D. Choi, Y. Lim, I.-M. Lee, S. Roh, and B. Lee, “Airy beam excitation using a subwavelength metallic slit array,” IEEE Photon. Technol. Lett. 24(16), 1440–1442 (2012). [CrossRef]
18. C.-Y. Hwang, K.-Y. Kim, and B. Lee, “Dynamic control of circular Airy beams with linear optical potentials,” IEEE Photon. J. 4(1), 174–180 (2012). [CrossRef]
23. J. D. Ring, C. J. Howls, and M. R. Dennis, “Incomplete Airy beams: finite energy from a sharp spectral cutoff,” Opt. Lett. 38(10), 1639–1641 (2013). [CrossRef]