We propose a composite waveguide configuration based on an inverted polymer channel structure with upper nematic liquid crystal cladding. This configuration can achieve a more homogenous liquid crystal molecular alignment between the core and the liquid crystal material by minimizing the rubbing damage during preparation of the alignment layer. We demonstrated our idea with a variable optical attenuator which exhibited a 24 dB of attenuation range over a tuning peak voltage of 10 V at 1550 nm.
© 2009 OSA
Integrated optics offers the solution to miniaturize optical devices of high functionality on a common substrate. Among them, planar lightwave circuits (PLCs) provide low-loss characteristics and high productivity, demonstrating good potential of key integrated optical devices for modulation, switching, and multiplexing [1–3]. Large number of materials can be used to fabricate PLC devices, such as polymer, silica (both fiber and silica-on-silicon), lithium niobate, and III-V semiconductor compounds [4–7]. Each type of these materials has its own pros and cons. However, compared with other materials, polymers permit the mass production of low-cost high-port count optical integrated circuits . They can be easily spin-coated and integrated on any surface of interest with good planarity. Moreover, polymers offer low material cost for high volume fabrication using UV or embossing techniques . They also exhibit large thermo-optic (TO) and electro-optic (EO) coefficients, making it possible to be implemented as highly efficient active optical devices, such as variable optical attenuators (VOAs) and switches .
In polymer PLC devices, TO effect is mostly used for tuning optical characteristic of waveguide devices. However, the main concern for this method includes power consumption, heat dissipation and potential geometric deformation of waveguide structure under high heater temperature. These problems can be solved by using EO effect as an alternative. EO polymers have been investigated because they offer advantages such as large EO coefficient, high speed and low driving voltage. However the stability of EO polymer remains an issue to be tackled.
Liquid crystal (LC) is also considered a promising material for PLC devices [10,11]. In particular, the effective EO exhibited in a LC optical waveguide is at least a few orders of magnitude larger than that of the best EO bulk material available. Moreover, there is huge interest in the material development and fabrication technologies in the LC display industry. On the other hand, LC applications in optical communication sectors are receiving increasing attention at a rapid pace. Various liquid crystal based optical devices have been developed, including optical switches, attenuators, and filters. Advantages of LC materials such as fast switching time with microseconds, low controlling voltages and power consumption with sub-milliwatt driving power, and high reliability have been demonstrated with advanced LC technologies [12,13].
In applying LC to integrated optics, there is a major drawback due to their large material loss. The nematic LC has a loss of 18 dB/cm , making them undesirable to be used as a core propagation medium of PLC devices. Later, d’Alessandro et al reported the propagation loss of 5 dB/cm by improving the alignment and reducing the scattering loss at higher wavelength in the near infrared . Another alternative method to circumvent this problem is to use LC as the cladding media to tune the optical properties in polymer waveguide devices. As a result, the overall insertion loss in such a device can be significantly reduced. Some successful devices have been demonstrated using LC as an upper cladding [16–18] placed mostly on inorganic materials rather than on polymers. There are reported works employing LC as cladding materials on polymer waveguides, but these are only demonstrated experimentally in slab waveguide structures , or in theoretical studies [20–22]. In fact, there are difficulties in integrating LC with polymer channel waveguide structures using conventional waveguide fabrication techniques. It is difficult to deposit and to pattern the LC films because they are mostly in liquid form. Moreover, LC is an anisotropic material and its refractive index depends on the molecule orientation. The homogeneous molecule orientation is critically determined by the alignment qualities between the LC and the waveguide surface concerned . It is highly desirable to have a flat waveguide surface for quality molecule alignment. Unfortunately most polymer waveguide devices are based on either strip or rib structure which does not have a flat surface for alignment. Such channels are normally damaged during the preparation of the alignment layer by mechanical scratching as shown in Fig. 1 . For channel waveguides with only a few micrometer dimensions, it is also difficult to make LC molecules alignment near the edge.
In this paper, we propose a composite waveguide configuration based on an inverted polymer channel structure using nematic liquid crystal as the cladding. Nematic LC offers many advantages compared with other types of LC materials, viz. simple chemical structure, convenient range of working temperature, commercially available at relatively low price and better mechanical properties. Using this configuration, the uneven boundary/structure over the channel waveguide surface and the roughness sidewall can be avoided. Consequently, a more homogeneous LC molecular alignment can be achieved. Besides, no rubbing damage was observed during the preparation of the alignment layer as the inverted rib waveguide is the downward direction of flat alignment layer. Using a single-mode variable optical attenuator (VOA) device, we are able to demonstrate the feasibility of such a configuration which can readily be applied to other LC based waveguide devices.
2. Design and Principle
Figure 2(a) shows a three dimensional structure of the proposed VOA device configuration. The inverted single mode benzocyclobutene (BCB) channel waveguide is fabricated on SiO2/Si wafer using conventional photolithography and reactive ion etching (RIE) technique. A nematic LC thin film is sandwiched between the waveguide and the ITO coated glass. A very thin layer of polyimide is deposited as an alignment layer on the upper surface of the waveguide structure and the ITO glass respectively. The Si substrate itself and the ITO layer act as the lower and upper electrode respectively. Figure 2(b) shows the cross-sectional dimension of the VOA which is designed to operate in single-mode region at 1550 nm wavelength.
Figure 3 shows that two of the optical axes in nematic LCs with uniaxial symmetry are equal. The indices in the principal directions x and y are equal and are referred as ordinary index no while the index in the z direction is referred as the extraordinary index ne. θ is the angle between the molecular axis and the incident light direction along OP. The refractive index, ne(θ) as seen by the extraordinary wave is given as follows:
It can be observed from Eq. (1) that ne(θ) will approach no when θ approaches 0°. For the device to operate as a VOA, the core material index ncore, must be between the maximum (ne(θ)max) and minimum (ne(θ)min) value of ne(θ). Since the refractive indices of most commonly used polymers are close to the nematic LC ordinary index, no, a more homogeneous LC molecular alignment is required to achieve a small angle θ so that ncore will be greater than ne(θ)min.
Without applying any electrical-field (E-field), the LC molecules are aligned in the light propagation direction which is parallel to the surface rubbing direction as shown in Fig. 4(a) . For transverse-electric (TE) mode, the LC index is no, however, for transverse-magnetic (TM) mode, the incoming light will see the refractive index ne(θ)min. Both are smaller than ncore, resulting in a light confinement to the core. With the applied E-field, the LC molecules re-orientate their directions perpendicular to the light propagation direction as shown in Fig. 4(b). For TE mode, the LC index is still no, so light guiding should be same. But for TM mode, the LC index gradually increases towards ne and light is increasingly absorbed by the LC cladding. Under this condition, the refractive index, ne(θ) as seen by the incident light should be close to ne and greater than ncore. As a result, the incoming light will scatter into the LC cladding, causing a drop in the output power. Therefore, the subsequent sections will deal with TM mode only.
To show the mode profile with increasing index of the LC cladding, simulations were carried out. In our simulation, the commercial propagation tool FIMMWAVE by photon design was used, providing rigorous solutions to the full Maxwell equations. A set of refractive index profiles was obtained by considering the electro-optic effect of the LC cladding. Using the calculated index profile, the light intensity profiles at each cross section for different LC indices are plotted, as shown in Fig. 5 . When the refractive index of the LC cladding (ncl) is lower than that of waveguide core (nc = 1.534), the mode is mainly confined in the channel waveguide as shown in Fig. 5(a), and only a small percentage of the optical field will be in the cladding material. When the ncl is increasing and becoming closer to nc, the mode is shifting towards the LC cladding as shown in Fig. 5(b) to 5(c). When the ncl is higher than nc , the mode mainly confined in the LC cladding as shown in Fig. 5(d).
3.1 Device fabrication
The commercially available BCB and Epoxy Optocast 3507 were used as the core and side-cladding of our device. Their refractive indices were measured with a prism-coupler system (Metricon 2010) . The measured indices of BCB and 3507 at 1536 nm were 1.537 and 1.510 for the TE polarization, and 1.535 and 1.508 for the TM polarization respectively. The nematic LC (E7) used in this experiment was purchased from Merck. The ordinary index (no) and extraordinary index (ne) at 1536 nm and 22 °C are 1.510 and 1.689 respectively according to the manufacturer datasheet.
To fabricate the device, a layer of 3507 was first spin-coated onto a silica-on-silicon substrate and cured by a UV lamp (Novacure 2100). This was then followed by post-baking at 160°C for 1 hour. The final thickness of the 3507 film was precisely trimmed to 2 µm by etching. A channel slot with dimension 3.8 µm width by 2 µm depth was then etched away from the 3507 film using standard photo-lithography and RIE technique. To form the inverted channel waveguide, BCB was filled into the slot by spin-coating, and then cured in nitrogen at 270°C for one hour. It was necessary to spin-coat the BCB three times to achieve a flat surface. RIE was applied again to trim the BCB film to the desired thickness. The device length used in this study was 2 cm.
To prepare for the alignment layer, a 170 nm thick Nissan SE130 polyimide was spin-coated onto the waveguide surface. Unidirectional rubbing was applied to the polyimide layer in the direction parallel to the waveguide core using a velvet cloth. The purpose of rubbing is to produce micro-relieves on the layer surface to induce enough anchoring energy so that the orientation of LC molecules will align homogeneously parallel to the waveguide core. The alignment layer on the ITO glass slide was also prepared in a similar way. Moreover, the rubbing direction between the ITO glass slide and the inverted structure was anti-parallel to ensure a small LC pre-tile angle (i.e. small θ value). To complete the whole assembly, the inverted channel structure was covered by the ITO coated glass slide only in the middle region so that the end faces of the channel waveguides are without LC cladding. The LC cladding thickness was controlled to 3μm using photo-resist film SPR6112B as the spacer. Finally, the E7 LC was heated to isotropic phase and filled into the cell by capillary effect. This was then cooled down slowly to room temperature.
3.2 Loss measurements
The typical optical loss of BCB waveguide for air cladding was found about 1.5 dB/cm by cut back method . For an inverted structure with LC cladding, the insertion loss was measured by collecting the light with a single mode fiber (SMF) at the waveguide output . It was found 20 dB loss approximately. In order to measure the coupling loss, the output SMF was replaced by an objective lens using a power detector. The coupling loss was calculated from the difference between the output power collected by the objective lens and the output power coupled to the SMF. This resulted in about 6 dB/facet. By subtracting the total coupling loss of 12 dB for both facets from the insertion loss, a propagation loss of 4 dB/cm could be estimated.
3.3 Attenuation measurements
The experimental set up for measuring the attenuation characteristics is shown in Fig. 6 . A 1550 nm laser source polarized to TM mode was launched into the waveguide core of VOA via a single-mode fiber. The input polarization was controlled with a polarization controller. The output beams from the waveguide was measured with a power meter. To drive the VOA device, a 1 kHz square wave voltage was applied across the ITO glass and the Si substrate. Since at both end faces of the devices are without LC cladding, light can only be confined in the waveguide channel. There is no light input/output in the LC over layer, to or from the input/output fiber.
4. Results and Discussion
The threshold voltage, Vc, for the E7 liquid crystal is 1.42 V. When the applied voltage is smaller than the critical value Vc, there is no change in the LC’s tilt angle. If the applied voltage is beyond the critical value, the LC breaks the stable state and starts switching to make an orientation angle. The relationship between the applied voltage and the orientation angle is given as follows :
The angle θm gradually increases with the external E-field. A saturated angle of 90° is achieved when the E-field intensity is around 3.5 times Vc. Based on the calculated results, the power output of this VOA under different applied voltage can be simulated using the Beam Propagation Method (BPM). As shown in Fig. 7 , when the applied voltage is smaller than the threshold value of 3 V, the waveguide output power stays relatively constant. When the voltage is increased to above Vc, a sharp decrease in the output power can be seen. By further increasing to 7.5 V, the power is reduced by 20 dB, showing all the core power is scattered into the LC cladding.
To compare with experimental results, the device was tested with an end-fire coupling set-up. The measured output power versus the applied voltage V is shown in Fig. 7. There is a reasonable match between the experimental and simulated results. The solid line is the simulation result and the dotted line is the measured attenuation for the device. In our case, both values are normalized to 0 dB. For a peak voltage less than 3 V, there is not much change in the output power as shown. For a peak voltage larger than 3 V, the output power drops quickly with increasing peak voltage. The output power becomes constant at the peak voltage of 10 V. The attenuation range is about 24 dB for TM mode over a range of 10 V peak Voltage.
The reason for the discrepancy between the simulated and experiment results is the material loss of LC that has not been considered in our BPM simulation. The discrepancy also becomes larger at a higher value of V. For small values of V, the index contrast between the LC cladding and core is very large, and the mode field will mainly be confined in the waveguide core as shown in Fig. 5(a). The LC loss will therefore not have too much effect on the output power. However, as the value of V increases, the index contrast between the LC cladding and the core decreases, resulting in increasing penetration of the evanescent field into the cladding as shown in Fig. 5(c). This explains why the total output power has a higher loss than anticipated. The experimental results also include the losses due to the light coupling between the polymeric waveguide and LC cladding. Such coupling also depends on the refractive index difference between the LC and the polymeric waveguide.
In summary, the study is a demonstration of simple VOA. However, the idea can be implemented in even more complicated integrated optical devices. The device preserves the low loss properties of polymer devices and also makes use of the large optical anisotropy and electro-optic effect of LC. By changing the design and applying the special asymmetric attenuation characteristic of the two polarization states, a polarization dependant loss compensator can be formed. Therefore, the polarization dependence attenuation characteristic will vary from device to device.
We have proposed an inverted channel waveguide structure using nematic LC as the upper cladding. This device preserves the low loss properties of polymer devices and also makes use of the large optical anisotropy & the electro-optic effect of LC. We have demonstrated the idea with a variable optical attenuator which exhibits a 24 dB of attenuation range over a tuning peak voltage of 10 V at 1550 nm. Our proposed configuration and fabrication method used can be readily applied to other LC based waveguide devices.
The authors acknowledge the support of CERG (No. 112606) of the Research Grants Council in Hong Kong.
References and links
1. A. Liu, L. Liao, D. Rubin, H. Nguyen, B. Ciftcioglu, Y. Chetrit, N. Izhaky, and M. Paniccia, “High-speed optical modulation based on carrier depletion in a silicon waveguide,” Opt. Express 15, 660–668 (2007). [CrossRef] [PubMed]
3. H. H. Keil, H. H. Yao, and C. Zawadzki, “2×2 digital optical switch realized by low cost polymer waveguide technology,” Electron. Lett. 32(16), 1470–1471 (1996). [CrossRef]
4. L. Eldada and L. W. Shacklette, “Advances in polymer integrated optics,” IEEE J. Sel. Top. Quantum Electron. 6(1), 54–68 (2000). [CrossRef]
5. A. Szameit, J. Burghoff, T. Pertsch, S. Nolte, A. Tünnermann, and F. Lederer, “Two-dimensional soliton in cubic fs laser written waveguide arrays in fused silica,” Opt. Express 14(13), 6055–6062 (2006). [CrossRef] [PubMed]
6. Z. He, Y. Li, Y. Li, Y. Zhang, L. Liu, and L. Xu, “Low-loss channel waveguides and Y-splitter formed by ion-exchange in silica-on-silicon,” Opt. Express 16(5), 3172–3177 (2008). [CrossRef] [PubMed]
8. C. G. Choi, “Fabrication of optical waveguides in thermosetting polymers using hot embossing,” J. Micromech. Microeng. 14(7), 945–949 (2004). [CrossRef]
9. H. P. Chan, C. K. Chow, and A. K. Das, “A wide-angle X-junction polymeric thermooptic digital switch with low crosstalk,” IEEE Photon. Technol. Lett. 15(9), 1210–1212 (2003). [CrossRef]
10. V. G. Chigrinov, “Liquid crystal devices for photonics applications,” Proc. SPIE 6781, 67811 (2007).
11. A. Diaz, S. Kubo, D. H. Kwon, J. Park, D. Werner, T. Mallouk, and I. C. Khoo, “Nonlinear liquid crystal Na no-metamaterials” IEEE/LEOS Winter Topical Meeting Series, 2008, 94–95(2008).
12. I. C. Khoo, Liquid Crystal (John Wiley & Sons, 2007), Chap. 6.
13. J. L. D. Bougrenet and D. B. D. La Tocnaye, “Engineering liquid crystals for optimal uses in optical communication systems,” Liq. Cryst. 31(2), 241–269 (2004). [CrossRef]
14. J. R. Winnery, C. Hu, and Y. S. Kwon, “Liquid-crystal waveguides for integrated optics,” IEEE J. Quantum Electron. 13(4), 262–267 (1977). [CrossRef]
15. A. D’Alessandro, B. D. Donisi, R. Beccherelli, and R. Asquini, “Nematic liquid crystal optical channel waveguides on silicon,” IEEE J. Quantum Electron. 42(10), 1084–1090 (2006). [CrossRef]
16. M. Kobayashi, H. Terui, M. Kawachi, and J. Noda, “2×2 optical waveguide matrix switch using nematic liquid crystal,” IEEE J. Quantum Electron. 18(10), 1603–1610 (1982). [CrossRef]
17. S. Muto, T. Nagata, K. Asai, H. Ashizawa, and K. Arii, “Optical stabilizer and directional coupler switch using polymer thin film waveguides with liquid crystal clad,” Jpn. J. Appl. Phys. 29(Part 1, No. 9), 1724–1726 (1990). [CrossRef]
18. A. D’Alessandro, D. Donisi, L. De Sio, R. Beccherelli, R. Asquini, R. Caputo, and C. Umeton, “Tunable integrated optical filter made of a glass ion-exchanged waveguide and an electro-optic composite holographic grating,” Opt. Express 16(13), 9254–9260 (2008). [CrossRef] [PubMed]
19. W. C. Chuang, J. S. Lin, K. Y. Lee, and W. Y. Lee, “Polymer waveguide switch using liquid crystal overlayer,” SPIE‘s International Symposium on Opto-Electrics and Micro-Photonics, 1998.
20. R. Asquini and A. d’Alessandro, “BPM analysis of an integrated optical switch using polymeric optical waveguides and SSFLC at 1.55μm,” Mol. Cryst. Liq. Cryst. (Phila. Pa.) 375, 243–251 (2002). [CrossRef]
21. Q. Wang and G. Farrell, “Integrated liquid crystal switch for both TE and TM modes: proposal and design,” J. Opt. Soc. Am. A 24(10), 3303–3308 (2007). [CrossRef]
22. J. Beeckman, K. Neyts, X. Hutsebaut, C. Cambournac, and M. Haelterman, “Simulations and experiments on self-focusing conditions in nematic liquid-crystal planar cells,” Opt. Express 12(6), 1011–1018 (2004). [CrossRef] [PubMed]
23. H. Desmet, K. Neyts, and R. Baets, “Modeling nematic liquid crystals in the neighborhood of edges,” J. Appl. Phys. 98(12), 123517 (2005). [CrossRef]
25. M. Haruna, Y. Segawa, and H. Nishihara, “Nondestructive and simple method of optical-waveguide loss measurement with optimisation of end-fire coupling,” Electron. Lett. 28(17), 1612–1613 (1992). [CrossRef]