This paper reports on the design, fabrication and characterization of an all-organic photonic integrated circuit working as a switching polarizer for visible light (630nm), combining organic waveguides and liquid crystals that can be electrically driven. The device was made in commercially available epoxy by laser direct writing lithography. A device with a 2dB loss and a 20dB extinction ratio for both polarizations, was simulated; the manufactured devices proved the working principle of the design. The results have led to the design of a switching polarization splitter, in which a careful choice of waveguide material and liquid crystal can lead to devices working on a wide range of wavelengths.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Photonics integrated circuits (PICs) have many potential applications [1–3] including next generation optical networks. The most employed platforms are silicon on insulator (SOI) and Indium Phosphide (InP) . InP allows for manufacturing of passive and active devices in the same platform technology . On the other hand, the SOI compatibility with CMOS technology leads to excellent processing control, low cost and high volume processing. Furthermore, SOI enables ultra- sharp bends in the optical waveguide, because of their high refractive index contrast  allowing for high device density and hence integration level.
Polymer waveguides are in some contexts an attractive alternative platform to SOI and InP due to their intrinsic low propagation losses in the visible and near infrared ranges, their relatively simple manufacturing procedures and their reasonable cost . Nowadays, polymer based PICs are mainly being used as Mode Field Diameter (MFD) adapting interconnections between conventional inorganic densely integrated SOI PICs (small MFD) [8, 9], and telecom fiber (large MFD). Polymer PICs are characterized by a relatively low index contrast, which does not allow realizing sharp bends. However, this reduced contrast also implies larger waveguide dimensions, so that the MFD can be tailored to that of an optical fiber to achieve very low coupling losses. Another possible drawback in some applications may be their limited thermal and environmental stability.
In order to make electronically active organic PICs it is necessary to combine the waveguides with electro-optics materials such as Liquid Crystals (LCs). LCs is a state of matter which is between conventional isotropic liquids and ordered solid crystals . While still being fluid the LCs are optically and dielectrically anisotropic. This means that the spatial orientation of the optical anisotropy (birefingence), also known as indicatrix, can be altered with external electromagnetic fields.
In most applications, LC devices are manufactured in such a way that the LC molecules are oriented in a given, preferred, direction in the absence of any applied field. By applying an electric field it is possible to switch a LC cell and then let it relax back to a known orientational state, when the electrical field is turned off. Typically the preferred orientation is achieved by surface treatment of the inside of the LC cell.
Interestingly, LCs have been used in integrated optics as early as 1977 , however low switching speed (ms) has limited the number of LC applications to those where speed was not an issue. Polarization manipulation is an example where speed is not relevant: infact passive devices are often employed [12, 13]. Hybrid devices combining a platform (InP, SOI or polymer) with electro-optic materials can achieve a high efficiencies as shown in various examples using slot waveguides filled with electro-optic materials [14, 15]. LC-based waveguides tuning has likewise been demonstrated [16, 17].
Polarization manipulating devices are not only important for polarization multiplexing and compensation but for polarization-transparent PICs as well . In this work, the design, simulation, manufacturing and characterization of a switching polarizer is presented. The devices is manufactured in a polymer platform employing a cost-efficient laser direct-write lithography (DWL). The active element is created by including a cavity filled with a nematic LC. The cavity is crosses through two parallel waveguides.
This basic design may lead to several devices having different functionalities. As an example, we have also included the design and simulation of a switching polarization beam-splitter.
2. Design and simulation
The device consists of two parallel polymeric waveguides made of Epocore (ncore=1.590) and Epoclad (nclad=1.578)  (Microresist Technology) which are intersected by a trench filled with LC. The trench crosses the WGs at a specific angle to allow polarization dependent switching according to the LC orientation (Fig. 1). Visible light was chosen for the application (630 nm); the device can also be designed for any other wavelength ranges, e.g. near infra-red for telecommunications (1300–1600nm). The device splits light into polarizations parallel and perpendicular to the substrate. The chosen LC (MDA-98-1602, Merck KGaA) (ne=1.78 and no=1.52) is a positive LC, i.e. the sign of the optical anisotropy, (ne − no), and the sign of the dielectric anisotropy (εe − εo) are both positive and thus the indicatrix will align parallel to an externally applied electrical field. The trench surfaces are conditioned so that the LC aligns parallel to the substrate plane and perpendicular to the longer trench walls.
Optimal dimensions of the LC intersection were estimated by simulation. As a first approximation, a plane-parallel slab inserted at an angle θwg, a slab, was simulated to calculate the loss caused by the refractive index gap between the polymer waveguides and the LC, and the the beam displacementd. The displacement depends on the height (h) of the plane-parallel slab, of the LC refractive index nLC and of the waveguide mode refractive index nwg as shown in Eq. (1) (Fig. 2). The nwg can be calculated using a modesolver. For simplicity, the index here is approximated to 1.58 ≈ nclad ≈ ncore, as discussed below.
This can be calculated from the waveguide dimensions and ncore and nclad.
Figure 3 shows the displacement and reflection coefficients for a simulated thickness of 20 µm and refractive indices nwg=1.58 and nLC =1.78, as a function of angle α. A minimum 2.7% reflection is found at 42°, however a trade-off between reflection losses due and the light displacement has been considered, chooseing 30° as the best compromise. At this angle the losses are 5.9% while the displacement is 5.1 µm. Losses increase significantly for angles smaller than 30°.
To complement the above analytical characterization of the initial design, a Lumerical FDTD Solutions finite element simulations were performed, employing the ordinary and extraordinary LC refractive indices no=1.52 and ne=1.78. The simulated thickness of the slab, h, was 20 µm, as above. Being optically anisotropic, both the waveguide design and the LC simulation had to be done for the two perpendicular light polarizations in-plane (TE) and out-of-plane (TM) with the substrate, as well as for the two different switching states of the LC: OFF–aligned in the substrate plane and ON–aligned in the plane perpendicular to the substrate.
Figure 4 shows the simulated light progression for either polarization, in both switching states for a waveguide separation of 5 µm. In the switched state, TM polarization propagates with an effective refractive index ne since the liquid crystal is aligned with the field applied perpendicularly to the substrate plane. The displacement can thus be simulated assuming an exact specific refractive index value i.e. ne=1.78. TE polarization propagates with no and since the no< nwg, the power in this polarization is refracted away from the output waveguide, upwards in the figure.
In the relaxed state, TM polarization propagates with no and is thus refracted away from the output waveguide. TE polarization experiences an effective refractive index neff, which is between and no since light impinges obliquely to the direction of the LC molecules, see Eq. (4). Using our design parameters (90°−θ=α=30° and alignment as in Fig. 1) neff was calculated to be 1.70. Although this value differs from ne, the numerical simulation shows that also in this case most of the light is coupled into the output waveguide (see Fig. 4). The same design parameters in the analytical approach lead to significant variation in displacement, i.e. 3.8 µm for TE polarization in the OFF state and 5.3 µm for TM polarization in the ON state.
It is worth noticing that, only 2 decimals are used in the estimation of both nwg, no, ne and neff. As seen in the Fig. 5 the displacement (d) of the light beam in the LC, is practically insensitive to any variation in the third decimal of the refractive index. The deviation is below 100 nm per thousands and is easily compensated by a taper structure in the waveguide coupling at the trench . Thus the chromatic dispersion of both the waveguide and the LC will only have marginal relevance for the final device.
LC cells were constructed using a transparent ITO coated glass as top electrode with SiO2 alignment layer and a metalized silicon wafer as bottom electrode on top of the silicon wafer multiple devices are manufactured in parallel in the polymer layer as seen in Fig. 6. The cross section of each waveguide is 3×3 µm (nTE ≈ nTM ≈ 1.585). Laser ablation is used to carve a 20 µm trench into each device on the wafer as indicated by the two red parallel lines. Prior to assembly, the polymer layer is covered by a SiO2 alignment layer. When filling the assembled cell, all trenches are filled by the LC.
Organic waveguides were manufactured with DWL. This procedure may lead to significant, systematic, differences between the design and the resulting device. The same issue is relevant to laser ablation used for digging the trench. To account for these, a priori unknown, deviations between design and final device, the same principal device design was repeated on the wafer varying the waveguide separation (d) from 3.5 to 5.5 µm. 20 different devices were designed and manufactured.
Each device consists of a pair of parallel Epocore/Epoclad (Microresist Technology)  polymeric waveguides (ncore=1.590, nclad=1.578). An excimer laser (KrF, working at 248 nm, pulse duration 5-8 ns, pulse energy 570 µJ)) was used to carve the trench. The trench was filled with MDA-98-1602 (ne=1.78 and no=1.52).
The waveguide manufacturing is illustrated in Fig. 7 and described in . Table 1 summarizes the optimized process fabricationn parameters. Once the waveguides are manufactured, the trench is made directly in the polymer sandwich by ablation. An excimer laser was to ablate the complete clad-core-clad stack. The excimer laser beam passed a rectangular mask (2000 µm x 200 µm) and the de-magnified image (factor of 10) of the mask was projected onto the polymer coated sample. This technique has been optimized in the past to make angular cuts . 70 pulses with a ablation rate of 0.45 µm/pulse were applied to carve the desired trench in waveguide core.
The last step is making the LC cell. LC alignment in absence of any external field was conditioned by an SiO2 layer deposited by oblique thermal evaporation (50 nm with an angle of 45°) onto the polymer layer and trenches and onto the top electrode . 5 µm spacers were deposited onto the SiO2 alignment layer on the top surface, so that the LC could flow freely and fill all the trenches on the chip. A gasket seal was applied leaving a mouth open for vacuum filling as described in , A sketch of the final macroscopic structure is shown in (Fig. 6).
SEM was employed to assess the quality of the waveguides, see Fig. 8(a). These are well defined rectangular shapes. The roughness of the bottom and the top cladding has been previously measured to be 2.2 nm .
As a result of unexpected high losses, the LC alignment and the quality of the trench cut was assessed. The device was observed between crossed polarizers with the predetermined orientation aligned with one of the polarizers, (Fig. 8(b)). Light transmission inside the trench is as low as the transmission outside the trench, indicating a good LC alignment. The trench walls themselves cause some light scattering. A trench angle with respect to the waveguide of 32° was measured(Fig. 8(c)), which was 2° rotation with respect to the design. Furthermore, the cross-section of the trench (Fig. 8(d)) revealed that the cut was not a rectangular section perpendicular to the substrate surface, but rather a V-shaped cut. For wider laser-ablated trenches, typically only 5-10° deviation from 90° cuts are expected , but in our case the trenches came out with almost 25° deviation (114° instead of 90°) due to the desired high aspect ratio and thin trenches. The resulting trench is thus thinner as is the effective LC slab. At the height of the waveguide the LC slab is only 11.3 µm, 8.7 µm less than the desired value. As a result a smaller displacement is expected and thus the devices with smaller waveguide separations are expected to give the best results. Additional losses were added by the not normal ablation (∼3 dB for the TM and ∼1 dB for the TE).
4. Optical characterization, results and discussions
As mentioned before, several devices were made on the same chip to account for manufacturing tolerances. The devices differ in the distance between the waveguides at the position of the LC slab. Devices with separations ranging from 0.9µm to 2.5µm i.e. ([3.9-5.5]µm between center of the waveguides) in steps of 0.1µm were successfully manufactured.
Two chips with duplicated design were characterized. No noteworthy difference between the four repetitions was found.
A polarized laser diode emitting at 630 nm coupled to a single-mode fiber reel was used in the characterization of the device. The light was coupled into the waveguide using a nano-positioner (Thorlabs NanoMax XYZ). A fiber rotator was used to manipulate the polarization at the input of the device. The light at the device output is captured using a multi-mode fiber and a photo-detector.
Figure 9 shows the relative power transmission versus waveguides separation for the two input polarizations (TM, TE) with LC switched (Von) and unswitched (Voff). The device works as a switching polarizer. When the applied voltage is off the TE polarization is guided to the output port. When the applied voltage is on the TM mode is guided to the output port. Each of the polarizations has an optimal waveguide separation for maximum transmission to the output port. This is due to the fact that the effective refractive index of the LC for TE is smaller that for TM (neff,TE < neff,TM ≈ ne). The unswitched LC alignment is not perpendicular to the TE mode propagation (see Fig. 1), but the switched LC alignment is almost perpendicular to the TM mode propagation. The slight deviation (less than 3°) is caused by the non-vertical trench sides.
Waveguide separation of 4.9 µm provides a measured power transmission of more than 15% for both polarizations when they are guided to the output port depending on the switching state of the LC. Simulations of the structure taking the trench imperfections into account were performed and the effective refractive indices for and were refined so that the simulated maximum transmission waveguide separation corresponded to the measured maximum transmission waveguide separation in the output: and . In the case of and , a refractive index no has been employed. A reasonable fit between the simulated and the measured values was obtained (Fig. 9).
With a perfectly rectangular trench the simulated loss was around 2 dB. The actual loss is ≈8 dB, this higher loss is attributed to the reflections in non-perpendicular surfaces and mode expansion in the trench. This could be reduced making a perpendicular cut and including a taper in the coupling-in-section of the trench zone .
5. Proposed advanced devices
Based on the above device a switching polarization beam splitter can be proposed.
Choosing an LC with an ordinary refractive index no=nwg=1.58 the same geometry will lead to a switching polarization beam splitter where the waveguide separations and slab height depend on ne. Fig. 10 shows a simulation of the power distribution of the two polarizations for the two switching states in a device identical to the one above but with different parameters: ne=1.78, no=1.58, α = 30° and h=20 µm. The working principle is the same as above, but in this case the both polarizations are guided for both switching states.
Alternatively, a switching polarization splitter may be manufactured using discontinuous waveguides (Fig. 11) while maintaining the LC and LC alignment as in the original design. In this case the trench positioning becomes critical since it has to be situated right at the discontinuity of the waveguides. The simulation of this device would be similar to the simulations shown in in Fig. 4(b), but with an upper waveguide located at the position where to the TE(VON ) and TM(VOFF) modes are being refracted.
A liquid crystal slab inserted in a tilted trench in the light path of two waveguides has been shown. This simple device has demonstrated several functionalities depending on the tilt angle, the refractive indices of the waveguides, the liquid crystal and the relative position of the output and input waveguides. Some examples are polarizer, optical switch and polarization splitter switch.
These functionalities have been confirmed by simulations with Lumerical FDTD. Quite moderate losses (2 dB) are predicted, making the device potentially useful for practical applications.
Actual devices based on organic waveguides have been demonstrated in the visible region. Extension to NIR wavelengths is straightforward, since no restrictions in the spectral range –other than waveguide limits–are expected. NIR range is better attainable with inorganic waveguides; the system is fully compatible with these waveguides.
The integration level of photonic integrated circuits (PICs) is far from that of electronic integrated circuits. Devices like the ones presented result in tuneable structures that may significantly increase the PIC functionality without adding much complexity, and thus contributing to the development of more integrated/sophisticated circuits.
Spanish Ministerio de Economía y Competitividad (RETOS TEC2013-47342-C2-1-R, RETOS TEC2016-77242-C3-2-R, BES-2014-070964, EEBB-I-17-12110); Comunidad de Madrid and the European Structural Funds (S2013/MIT-2790 SINFOTON-CM).
Authors are grateful for financial support. Manuel Caño-García is specially grateful for the help provided by the CMST support team.
References and links
1. T. Su, R. P. Scott, C. Ogden, S. T. Thurman, R. L. Kendrick, A. Duncan, R. Yu, and S. Yoo, “Experimental demonstration of interferometric imaging using photonic integrated circuits,” Opt. Express 25, 12653–12665 (2017). [CrossRef] [PubMed]
2. S. Feng, C. Qin, K. Shang, S. Pathak, W. Lai, B. Guan, M. Clements, T. Su, G. Liu, H. Lu, R.P. Scott, and S.J. Ben Yoo, “Rapidly reconfigurable high-fidelity optical arbitrary waveform generation in heterogeneous photonic integrated circuits,” Opt. Express 25, 8872–8885 (2017). [CrossRef] [PubMed]
3. L. Chang, M. H. Pfeiffer, N. Volet, M. Zervas, J. D. Peters, C. L. Manganelli, E. J. Stanton, Y. Li, T. J. Kippenberg, and J. E. Bowers, “Heterogeneous integration of lithium niobate and silicon nitride waveguides for wafer-scale photonic integrated circuits on silicon,” Opt. Lett. 42, 803–806 (2017). [CrossRef] [PubMed]
4. D. Liang and J. Bowers, “Photonic integration: Si or InP substrates?” Electron. Lett. 45, 578–581 (2009). [CrossRef]
5. L. A. Coldren, S. C. Nicholes, L. Johansson, S. Ristic, R. S. Guzzon, E. J. Norberg, and U. Krishnamachari, “High performance InP-based photonic ICs - A tutorial,” J. Lightwave Technol. 29, 554–570 (2011). [CrossRef]
6. Z. Sheng, D. Dai, and S. He, “Comparative study of losses in ultrasharp silicon-on-insulator nanowire bends,” IEEE J. Sel. Top. Quantum Electron. 15, 1406–1412 (2009). [CrossRef]
7. R. Dangel, J. Hofrichter, F. Horst, D. Jubin, A. La Porta, N. Meier, I. M. Soganci, J. Weiss, and B. J. Offrein, “Polymer waveguides for electro-optical integration in data centers and high-performance computers,” Opt. Express 23, 4736–4750 (2015). [CrossRef] [PubMed]
8. A. La Porta, R. Dangel, D. Jubin, N. Meier, F. Horst, and B. J. Offrein, “Scalable and broadband silicon photonics chip to fiber optical interface using polymer waveguides,” in Proceedings of the IEEE Optical Interconnects Conference (IEEE, 2017) pp. 13–14.
9. S. Knauer, M. López-García, and J. G. Rarity, “Structured polymer waveguides on distributed bragg reflector coupling to solid state emitter,” J. Opt. 19, 065203 (2017). [CrossRef]
10. S. Chandrasekhar, Liquid crystals, 2nd. ed. (Cambridge University Press, 1992). [CrossRef]
11. J. Whinnery, C. Hu, and Y. Kwon, “Liquid-crystal waveguides for integrated optics,” IEEE J. Quantum Electron. 13, 262–267 (1977). [CrossRef]
12. D. Taillaert, H. Chong, P. I. Borel, L. H. Frandsen, R. M. De La Rue, and R. Baets, “A compact two-dimensional grating coupler used as a polarization splitter,” IEEE Photon. Technol. Lett. 15, 1249–1251 (2003). [CrossRef]
13. T. M. Leslie and R. G. Lindquist, “Liquid crystal planar non-blocking nxn cross-connect,” (2003). US Patent 6,559,921.
14. J. Leuthold, C. Koos, W. Freude, L. Alloatti, R. Palmer, D. Korn, J. Pfeifle, M. Lauermann, R. Dinu, S. Wehrli, M. Jazbinsek, P. Günter, M. Waldow, T. Wahlbrink, J. Bolten, H. Kurz, M. Fournier, J. M. Fedeli, H. Yu, and W. Bogaerts, “Silicon-organic hybrid electro-optical devices,” IEEE J. Sel. Top. Quantum Electron. 19, 114–126 (2013). [CrossRef]
15. R. Ding, T. Baehr-Jones, Y. Liu, R. Bojko, J. Witzens, S. Huang, J. Luo, S. Benight, P. Sullivan, J. Fedeli, M. Fournier, L. Dalton, A. Jen, and M. Hochberg, “Demonstration of a low Vπ L modulator with GHz bandwidth based on electro-optic polymer-clad silicon slot waveguides,” Opt. Express 18, 15618–15623 (2010). [CrossRef] [PubMed]
16. C.-T. Wang, Y.-C. Li, J.-H. Yu, C. Y. Wang, C.-W. Tseng, H.-C. Jau, Y.-J. Chen, and T.-H. Lin, “Electrically tunable high Q-factor micro-ring resonator based on blue phase liquid crystal cladding,” Opt. Express 22, 17776–17781 (2014). [CrossRef] [PubMed]
17. Y. Xing, T. Ako, J. P. George, D. Korn, H. Yu, P. Verheyen, M. Pantouvaki, G. Lepage, P. Absil, A. Ruocco, C. Koos, J. Leuthold, K. Neyts, J. Beeckman, and W. Bogaerts, “Digitally controlled phase shifter using an SOI slot waveguide with liquid crystal infiltration,” IEEE Photon. Technol. Lett. 27, 1269–1272 (2015). [CrossRef]
18. A. Elmogi, E. Bosman, J. Missinne, and G. Van Steenberge, “Comparison of epoxy-and siloxane-based single-mode optical waveguides defined by direct-write lithography,” Opt. Mater. 52, 26–31 (2016). [CrossRef]
19. E. Hecht, Optics2nd edition, (Addison-Wesley Publishing Company, 1987)
21. K. Kaur, A. Subramanian, P. Cardile, R. Verplancke, J. Van Kerrebrouck, S. Spiga, R. Meyer, J. Bauwelinck, R. Baets, and G. Van Steenberge, “Flip-chip assembly of vcsels to silicon grating couplers via laser fabricated SU8 prisms,” Opt. Express 23, 28264–28270 (2015). [CrossRef] [PubMed]
22. D. Cuypers, Vertically aligned nematic liquid crystal microdisplays for projection applications, (Ph.D. thesis, Ghent University2005).
23. E. Otón Martínez, Adaptive liquid-crystal phase-only passive devices, (Ph.D. thesis, E.T.S.I. Telecomunicacion (UPM)2013).