Invisibility Cloak Printed on a Photonic Chip

Abstract

Invisibility cloak capable of hiding an object can be achieved by properly manipulating electromagnetic field. Such a remarkable ability has been shown in transformation and ray optics. Alternatively, it may be realistic to create a spatial cloak by means of confining electromagnetic field in three-dimensional arrayed waveguides and introducing appropriate collective curvature surrounding an object. We realize the artificial structure in borosilicate by femtosecond laser direct writing, where we prototype up to 5,000 waveguides to conceal millimeter-scale volume. We characterize the performance of the cloak by normalized cross correlation, tomography analysis and continuous three-dimensional viewing angle scan. Our results show invisibility cloak can be achieved in waveguide optics. Furthermore, directly printed invisibility cloak on a photonic chip may enable controllable study and novel applications in classical and quantum integrated photonics, such as invisualising a coupling or swapping operation with on-chip circuits of their own.

Invisibility Cloak Printed on a Photonic Chip

Zhen Feng,, Bing-HongWu, Yu-Xi Zhao, Jun Gao,, Lu-Feng Qiao,, Ai-LinYang,, Xiao-Feng Lin, & Xian-Min Jin,

Invisibility cloak capable of hiding an object can be achieved by properly manipulating electromagnetic printed invisibility cloak on a photonic chip may enable controllable study and novel applications in

Transformation optics is a prevalent approach to modify material parameters, carefully constructing articial material and manipulating electromagnetic wave1,2. Due to the enormous difficulties in the realization of cloaking, scientists have been seeking various simplied models38 to approach the prefect cloaking. Among those methods, Schuig et al.3 use articially structured metamaterial to fulll the dimension-reduced cloaking scheme which could only cloak the light at some specic frequencies and polarizations. To broaden the operating frequency range57, the carpet cloak is introduced to operate at microwave9,10 and optical1113 frequency bands. Metamaterials are also applied to achieve a skin cloak which can hide volumetric objects at optical frequencies14.

In diusive light scattering medium where light does not conform to the ballistic propagation, broadband macroscopic cloak made of thin shells of metamaterial is realized in three dimensions, according to Ficks equation15.

Ray optics provides a new way of concealing macroscopic objects by using calcite crystal16,17 and standard optical devices such as lens18 and prism19, which can guide light though a predened path by Snells Law. We introduce a distinctive broadband cloaking scheme in waveguide optics that can be applied on a photonic chip. To be specic, light can be conned in the core of waveguide with higher refractive index associated with a modest curve, which shows its capability of guiding and manipulating light eld in an arrayed conguration. Within the tolerance of curvature, a waveguide is bended to circumvent a central region. Optical information through the waveguide is acquired at the output while the exact propagation path remains condential, which allows hidden objects to exist in the areas rounded by waveguides. The articial structure requires a waveguide fabrication technique in three-dimensional space, which can be met by the intrinsic processing capacity of femtosecond laser writing2022. A refractive index contrast can be generated via nonlinear absorption of femtosecond laser in transparent materials, where a trajectory of laser focus transforms to a waveguide. Elaborate control of the trajectory enables complex optical circuits2224 as well as three-dimensional conguration21,24.

Results

As a cloaking scheme, two prerequisites should be met: a clear image transfer and a hidden inner structure25.

Foreground mask featured 0.7mm0.7mm is set before the entrance window of the photonic chip and this image is expected to be soundly transported through the chip without any interference with an inner structure . In order to cloak the 0.44mm0.35mm structure inside a chip, which is fabricated with a high power laser, we

State Key Laboratory of Advanced Optical Communication Systems and Networks, Institute of Natural Sciences Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China,

R

A P

SCIENTIFIC REPORTS

1

www.nature.com/scientificreports/

Figure 1. (a) Principle of on-chip cloaking. Directional light through a mask of is incident onto the window where thousands of waveguides embedded in the chip, propagating along the waveguides to circumvent the inner structure and re-assembling to restore the mask image without interference of . (bd) Characterization of cloaking structures with dierent pitches: (b) Cross sections of cloaking windows under the microscope with scale bars and duty cycles. Output images of the cloaking structures with (d)/without (c) the inner structure with pitches varying from 13m to 31m.

prototype up to 5,000 waveguides in three dimensions along a spindle path opening up the central space for the to be hidden in (Fig.1a).

In this section, we characterize the cloaking structures by dierent pitches and optimize the performance with the introduced similarity assessment. Tomography analysis is applied to unveil the evolution of light eld in the optimal one. In the end, we benchmark the robustness of the cloaking device with a continuous three-dimensional viewing angle scan.

The spindle of tightly bounded waveguides seems to be a favorable choice to obtain a high resolution image. However, considering the interactions between waveguides, the pitch is obviously a crucial parameter to determine the clarity of image transfer. In our experiment, we fabricate 10 congurations of spindles with dierent pitches ranging from 13m to 31m and the corresponding cross sections captured using a 20 objective lens with a numerical aperture of 0.04 and 16 eyepieces are exhibited in Fig.1b. A Gaussian beam (405nm) is radiated through the mask at the input of the spindle and collected at the output by a CCD camera, as shown in Fig.1c,d. Each conguration contains two sets: with (Fig.1d) and without (Fig.1c) the inner structure. As illustrated in Fig.1c,d, various outputs are observed owing to dierent pitches. In order to evaluate their performances, a uniform statistical criterion should be set up. Therefore we introduce the normalized cross correlation ()26,27, a common and convenient assessment of the similarity between two image patterns with the basic idea to compare the grey scale of each corresponding pixel between two images. Specically, when images are converted into grey matrixes, between image a and b can be represented as:

=

( )( )

( ) ( ) (1)

a a b b

a a b b

i i i

i i i i

, 2 2

where i=1, 2, represents the ith pixel of the image. a,b is the average value of all the grey scale of pixels of a and b, respectively. According to the denition, the ranges from 0 to 1; the closer to 1 the value is, the more similar two gures are to each other.

Firstly we compare the similarity of between the one only passing through the chip substrate (Fig.2a) and the one passing through the waveguides with dierent pitches (Fig.2b,c). The results (a,b; a,c) sketch that the

value increases steadily and reaches the peak at the pitch value of 21m and then decreases with the pitch getting larger. The optical coupling (crosstalk) distorts the light eld propagating in the waveguides when they get too

a b

SCIENTIFIC REPORTS

2

www.nature.com/scientificreports/

Figure 2. Area instruction of analysis and cloak characterization. (ac) Show the output images of .

Area i and ii represent the exact area of and the background, respectively: (a) Cloak o and without the inner structure . Cloak on and (b) without /(c) with . (df) Gaussian beam without mask. (d) Cloak o and with . Cloak on and (e) without /(f) with . Cloak characterization. (1) value comparison: structure (ai,bi; ai,ci)> output image (a,b; a,c)>background (aii,bii; aii,cii)hidden (d,e; d,f). (2) values peak atthe 21m pitch of the output image and structure. (3) The uctuation of values on a low level represents the numerically well cloaked .

close to each other. It conforms to a monotonous tendency, which optical coupling decreases by about 85% with every pitch increase of 2 m, according to our calculation. Due to the elliptical cross section (see Fig.1b), the crosstalk distortion is especially intensive in the vertical direction. On the other hand, a larger pitch constraints the total volume of information in virtue of the lower waveguide density. We calculate the duty cycle, the ratio of the total cross sectional area that the waveguides occupy to the CCD active area, to quantitatively show the waveguide density. (shown in Fig.1b).

As referred in the rst result (See line a,b; a,c in Fig.2), where 21 m is regarded as the optimal pitch for transmitting the image, the maximum value is still moderately large, which can be attributed to the overweight of the background (the region on the image except the ) in the calculation of . To estimate the inuence of the background, we extract region ii in Fig.2ac and calculate the , as shown in Fig.2 (aii,bii; aii,cii). The result

shows that the similarity of the background between dierent images remains at a low level (within the range of 0.180.03).

In view of the low weight of background noise visually, we clip the structure of (Region i in Fig.2ac) from

a uniform background and then re-calculated the . As expected, the values increase prominently, reaching the peak at 21 m (Fig.2 (ai,bi; ai,ci)). Furthermore, it is the consequence of the inevitable dissipation of light between waveguides that the maximum of ai,bi, peaking at 0.986, still cannot reach 1. Actually, the pitch of 21m reaches the image standard as high as a retina image.

In order to give a quantitative indication of that the inner structure has been well cloaked. We remove the foreground mask and compare the similarity between the output image of the chip with the uncloaked inner structure (Fig.2d), the one without the cloaked structure (Fig.2e) and the one with the cloaked structure (Fig.2f), respectively. Provided that the cloaking performance is not that good, we should be able to see the inner structure shown in Fig.2f, and a higher d,f than d,e. Apparently, the results show the values uctuate around 0.13 0.05, at a very low level which is hard to be distinguished from the background noise.

To reveal the evolution of light eld in the printed cloak, tomography analysis is applied to the mechanism of cloaking. We cut the chip transversely, record the data from the output and compare these experiment results with simulation ones.

We simulate this experiment by a commercial soware RSo (Beam type: Gaussian, wavelength: 405 nm, refractive index 1.5198) and derive the distribution of the light eld. As shown in Fig.3, aer the beam incidence, most light circumvents the cloaked area along the waveguides, while the rest is scattered by the bulk of the chip and forms the weak white noise in the background. Also we nd the light distribution in the output is quite similar with the one in the input, which means little information may have lost aer the transmission.

Inspired by the widely used tomography in biological science28,29, we perform a tomography on our cloaking chip by grinding the chip to a predened length, which takes 1.5hours for each millimeter removal on a polisher. As the chip gets shorter, by observing the cross section pattern, the process of the projection image splitting into two parts gradually apart from each other and then combining into one again is experimentally reproduced. Comparing the numerical results of the processed intensity distribution along the lateral axis (Fig.3b), we can observe that the numerical one is more uctuant, which is because the stray light smooths the uctuation in the experiment.

SCIENTIFIC REPORTS

3

www.nature.com/scientificreports/

Figure 3. Tomography of the cloaking chip. Simulative (a) and experimental (d) tomography of on-chip cloaking scheme is displayed with intensity distribution along the lateral axis in (b,c) respectively.

Three-dimensional capacity is frequently employed to benchmark the performance of a cloaking approach. In our experiment, while our cloaking approach is only implemented in one dimension to demonstrate its feasibility, the same method can be extended to three dimensions with higher complexity of waveguide arrangement and larger number of printed waveguides. We should note that cloaking in one dimension is not that trivial in certain scenarios. For instance, the chips are packaged leaving only some specic access to the public. The components placed at appropriate locations can be well hidden on the chip in the observation direction. Interestingly, even our one-dimensional conguration has its three-dimensional capacity of cloaking in a cone, which can be understood as an eect of collective numerical aperture of waveguides.

Therefore, the tolerance for viewing angle on the chip, which shows three-dimensional capacity of cloaking, is also investigated18,19. We change the view direction by adjusting the lens behind the cloaking chip (21 m pitch, with the inner structure) multi-directionally and align the output image with the corresponding direction as shown in Fig.4. The sharpness, brightness and visibility decrease as the viewing angle gets large, though the does not distort. An elliptical-like region spanning 8.0 laterally and 6.7 vertically outlines the tolerance for viewing angle. Besides, the limited angle that we derive above conforms to the numerical aperture theory of an individual multimode ber. It is remarkably interesting that, the inner structure is cloaked so well that even if the obliqueness of the viewing angle is large enough to aect the image transmission, no parts can be observed. In addition, output powers in Fig.4a are measured and plotted correspondingly (see Fig.4b) which describes the energy distribution in the output light cone.

Discussion

Length dierence of waveguides introduces phase distortion site by site. In our experiment, imperfect collective bending of waveguides may generate site-dependent length dierence and therefore dierential phases. A phase-preserved cloak3032 can be realized by setting the dierence to be integral times of 2 in the fabrication process of femtosecond direct writing. Waveguide dispersion is highly related to the bandwidth performance of a cloak. While the monochromatic laser used in our system is dispersion insensitive, a broadband wavelength can be covered for our multimode waveguides in size of 16m10m with a large V number of 13.62.

The capacity of the cloaking structure depends on the arrangement of waveguides. In our conguration, we implement the cloaking structure in the curvature radius 85.9mm (see Methods) with refractive index contrast 0.0015. We obtain an optimal cloaking performance in a hidden region in size of 0.44 mm 0.35 mm. With the

SCIENTIFIC REPORTS

4

www.nature.com/scientificreports/

Figure 4. (a) Image quality on the cloaking chip with dierent viewing angles. We get 13 12 output images by adjusting the lens behind the chip for continuously multi-directional angles and aligning the output image with the corresponding directions. (b) Normalized energy angular distribution. An elliptical-like region spanning 8.0 laterally and 6.7 vertically outlines the tolerance for viewing angle in energy distribution.

decrease of curvature radius, larger bending loss will be introduced, however, the relationship between cloaking area and curvature radius is not always monotonous.

In conclusion, we have presented the rst experimental demonstration of invisibility cloak in waveguide optics that can hide millimeter-scale structure inside a photonic chip. The printed waveguides conne and guide light eld in a three-dimensional array, and evolve with collective curvature in the propagating direction to mimic the Poynting vector in transformation-optics-based cloak1. A larger invisible volume is attainable with tightly bending cloak spindle of high refractive index contrast33 and rearrangement of waveguides. The quality of image transfer that benchmarks the performance of the cloak can be upgraded by increasing the waveguide density and suppressing crosstalk-induced distortion simultaneously, which is achievable if we can manipulate all the phases of evanescent light interference between waveguides24. The on-chip image transfer itself may stimulate novel applications in deep space exploration34 and biological sensing35. Furthermore, the printed cloak allows backdoor operations like swap or coupling isolated from observers, representing an emerging problem of information security and hacking on a photopic chip.

Methods

A femtosecond laser (10 W, 1026 nm) with 290 fs pulse duration and 1MHz repetition rate is frequency doubled to 513nm and directed into an spatial light modulator (SLM) to create burst trains which is focused on a borosilicate substrate (20 mm 20 mm 1 mm) with a 50 objective lens with a numerical aperture of 0.55 to fabricate several thousand spindle waveguides (85.9 mm-radius curves with 1.5mm-long straight design at both ends; refractive index increase of 0.0015 in the borosilicate substrate 1.5198) at a constant velocity of 10mm/s. Array is formed with a cross section size of 0.7mm0.7mm and great eorts have been made to process depth independent waveguides through power and SLM compensation. An inner structure of reduced Planck constant whose visibility is guaranteed by a high power laser direct writing back and forth hundreds of times is located in cloaking area with the size of 0.44mm0.35mm (See Fig.5 for overall design parameters).

Since we only concern about the in images a and b (See Eq.1), the exact CCD image clipping is necessary. To eliminate the error caused by the slight shi of , we make a clip of image a with a resolution matrix of MN, and then crop a smaller box of image b with the matrix of mn (m M, n N) in the same way. Transforming the RGB matrix into grey scale, we compare the m n matrix extracted from b with every m n submatrix of a. The nal normalized correlation coefficient value between a and b, a,b, is chosen to be the maximum of the (Mm)(Nn) results. Specically in our demonstration, we

SCIENTIFIC REPORTS

5

www.nature.com/scientificreports/

Figure 5. Overall design parameters. Design parameters of (a) transverse cross section at both ends, (b) the cloaking structure and (c) mask and (d) inner structure with their corresponding images.

Figure 6. Schematic of viewing angle measurement. Viewing angle scan in (a) vertical view, (b) lateral view and (c) 3D view. ui - object distance, vi - image distance. i can be calculated by the measurement of angle i (i=1, 2).

crop image a of 240 320pixels and image a of 204236pixels. Thea,b value is chosen to be the maximum of (240204)(320236)=3024 results.

A commercial Beam Propagation Method soware is used to test the cloaking scheme qualitatively in advance. We build a glass block with volume of 20mm1.6mm0.8mm, transparent boundary condition, refractive index 1.5198 as a photonic chip where 38 38 waveguides (the size of each cross section is 16m10m) are imbedded with index increase of 0.0015. All other geometry conforms to the fabrication structure. A 0.8mm0.8mm beam, cut from a non-polarized Gaussian beam of 405nm wavelength by a square mask, is launched on the cloaking chip. All simulation results are illustrated in Fig.3a,b.

A Gaussian beam (wavelength 405 nm; beam waist 1.8mm) through the 0.7mm0.7mm mask is radiated at the input of the spindle and the output light information is collected by two sequent lenses placed behind the chip for imaging, as shown in Fig.6. The key parameter, the viewing angle related to the output of the chip cannot be measured directly unless we move the lens 1 aer each measurement, which leads to vast workload of readjustment. Since the light is paraxial, we can map the output

SCIENTIFIC REPORTS

6

www.nature.com/scientificreports/

of the waveguide to the point O by lens 1 and measure the corresponding object distance (u1) and image distance (v1). Thus, can be calculated by the indirect measurement of angle for the relation between and is given by

= =

u v i

i i

1 1

where i= 1, 2 represents azimuth and elevation angles. Since images can only be formed when lens 2 is within the light cone, we shi lens 2 both in azimuth and altitude direction with a corresponding adjustment of the CCD to catch the image simultaneously. Aer recording the shiing distance Li (Not shown in Fig.5) of lens 2 and the length of u2, we can get the corresponding angle i:

= =

Lu i

i 2

Substitute (3) into (2), we get the expression of viewing angle :

=

L =

uu u i

i 1 1 2

In addition, we have also measured the output powers in every combination of azimuth and elevation angles and plotted in Fig.4b. This energy distribution demonstrates a radiational decreasing tendency of energy from the center to the edge in the output light cone, which shows a visually great agreement with the radiational decreasing of clearity in the result of Fig.4a.

References

1. Pendry, J. B. Controlling electromagnetic elds. Science 312, 17801782 (2006).2. Leonhardt, U. Optical conformal mapping. Science 312, 17771780 (2006).3. Schurig, D. et al. Metamaterial electromagnetic cloak at microwave frequencies. Science 314, 977980 (2006).4. Cai, W., Chettiar, U. K., Kildishev, A. V. & Shalaev, V. M. Optical cloaking with metamaterials. Nat. Photonics 1, 224227 (2007).5. Li, J. & Pendry, J. B. Hiding under the carpet: A new strategy for cloaking. Phys. Rev. Lett. 101, 203901 (2008).6. Kildishev, A. V., Cai, W., Chettiar, U. K. & Shalaev, V. M. Transformation optics: approaching broadband electromagnetic cloaking. New J. Phys. 10, 115029 (2008).

7. Leonhardt, U. & Tyc, T. Broadband invisibility by non-Euclidean cloaking. Science 323, 110112 (2009).8. Chen, H., Chan, C. T. & Sheng, P. Transformation optics and metamaterials. Nat. Mater. 9, 387396 (2010).9. Liu, R. et al. Broadband ground-plane cloak. Science 323, 366369 (2009).10. Ma, H. F. & Cui, T. J. Three-dimensional broadband ground-plane cloak made of metamaterials. Nat. Commun. 1, 21 (2010).11. Valentine, J., Li, J., Zentgraf, T., Bartal, G. & Zhang, X. An optical cloak made of dielectrics. Nat. Mater. 8, 568571 (2009).12. Gabrielli, L. H., Cardenas, J., Poitras, C. B. & Lipson, M. Silicon nanostructure cloak operating at optical frequencies. Nat. Photonics 3, 461463 (2009).

13. Ergin, T., Stenger, N., Brenner, P., Pendry, J. B. & Wegener, M. Three-dimensional invisibility cloak at optical wavelengths. Science 328, 337339 (2010).

14. Ni, X., Wong, Z. J., Mrejen, M., Wang, Y. & Zhang, X. An ultrathin invisibility skin cloak for visible light. Science 349, 13101314 (2015).15. Schittny, R., Kadic, M., Buckmann, T. & Wegener, M. Invisibility cloaking in a diusive light scattering medium. Science 345, 427429 (2014).

16. Zhang, B., Luo, Y., Liu, X. & Barbastathis, G. Macroscopic invisibility cloak for visible light. Phys. Rev. Lett. 106, 033901 (2011).17. Chen, X. et al. Macroscopic invisibility cloaking of visible light. Nat. Commun. 2, 176 (2011).18. Choi, J. S. & Howell, J. C. Paraxial ray optics cloaking. Opt. Express. 22, 29465 (2014).19. Chen, H. et al. Ray-optics cloaking devices for large objects in incoherent natural light. Nat. Commun. 4 (2013).20. Davis, K. M., Miura, K., Sugimoto, N. & Hirao, K. Writing waveguides in glass with a femtosecond laser. Opt. Lett. 21, 1729 (1996).21. Nolte, S., Will, M., Burgho, J. & Tuennermann, A. Femtosecond waveguide writing: a new avenue to three-dimensional integrated optics. Appl. Phys. A Mater. Sci. Process. 77, 109111 (2003).

22. Meany, T. et al. Laser written circuits for quantum photonics. Laser Photon. Rev. 9, 363384 (2015).23. Crespi, A. et al. Measuring protein concentration with entangled photons. Appl. Phys. Lett. 100, 233704 (2012).24. Chaboyer, Z., Meany, T., Helt, L. G., Withford, M. J. & Steel, M. J. Tunable quantum interference in a 3D integrated circuit. Sci. Rep. 5, 9601 (2015).

25. Xu, L. & Chen, H. Conformal transformation optics. Nat. Photonics 9, 1523 (2014).26. Lewis, J. Fast template matching. Vis. Interface 120123 (1995).27. Halimeh, J. C., Schmied, R. & Wegener, M. Newtonian photorealistic ray tracing of grating cloaks and correlation-function-based cloaking-quality assessment. Opt. Express 19, 6078 (2011).

28. Fujimoto, J. G. et al. Optical biopsy and imaging using optical coherence tomography. Nat. Med 1, 970972 (1995).29. Low, A. F., Tearney, G. J., Bouma, B. E. & Jang, I.-K. Technology insight: optical coherence tomographycurrent status and future development. Nat. Clin. Pract. Cardiovasc. Med. 3, 154162 (2006).

30. Xu, H., Zhang, B., Yu, T., Barbastathis, G. & Sun, H. Dielectric waveguide bending adapter with ideal transmission: practical design strategy of area-preserving affine transformation optics. Journal of the Optical Society of America B 29, 1287 (2012).

31. Yu, Y. et al. Controlled scalable synthesis of uniform, high-quality monolayer and few-layer MoS2 Films. Sci. Rep. 3 (2013).32. Xu, S. et al. Broadband surface-wave transformation cloak. Proceedings of the National Academy of Sciences 112, 76357638 (2015).33. Eaton, S. M., Ng, M. L., Osellame, R. & Herman, P. R. High refractive index contrast in fused silica waveguides by tightly focused, high-repetition rate femtosecond laser. J. Non. Cryst. Solids. 357, 23872391 (2011).

34. Bland-Hawthorn, J. et al. A complex multi-notch astronomical lter to suppress the bright infrared sky. Nat. Commun. 2, 581 (2011).35. Walt, D. R. Imaging optical sensor arrays. Curr. Opin. Chem. Biol. 6, 689695 (2002).

The authors thank J.-W. Pan, I. A. Walmsley, X.-F. Chen and Y. Chen for helpful discussions. This research leading to the results reported here was supported by the National Natural Science Foundation of China under Grant No. 11374211, the Innovation Program of Shanghai Municipal Education Commission, Shanghai Science and Technology Development Funds, and the open fund from HPCL (No. 201511-01). X.-M.J. acknowledges support from the National Young 1000 Talents Plan.

, 1, 2 (2)

i

, 1, 2

(3)

i

, 1, 2

(4)

SCIENTIFIC REPORTS

7

www.nature.com/scientificreports/

Author Contributions

X.-M.J. conceived the work and supervised the project. Z.F., B.-H.W., Y.-X.Z., J.G., L.-F.Q., A.-L.Y., X.-F.L. and X.-M.J. all contributed to designing and setting up the experiment. Z.F., B.-H.W. and Y.-X.Z. analyzed the data and performed simulation. Z.F. and X.-M.J. wrote the paper with input from all authors.

Additional Information

Competing nancial interests: The authors declare no competing nancial interests.

How to cite this article: Feng, Z. et al. Invisibility Cloak Printed on a Photonic Chip. Sci. Rep. 6, 28527; doi: 10.1038/srep28527 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

SCIENTIFIC REPORTS

8