Liquid crystal tunable planar Bragg Gratings produced by Direct UV Writing are capable of wavelength tuning of over 100GHz. However, such devices exhibit non-linear tuning curves with threshold points and hysteresis. We show that these effects are due to the formation of disclination structures in the liquid crystal and discuss the role of electrode defects and sample temperature on wavelength tuning.
© 2007 Optical Society of America
Wavelength Division Multiplexing (WDM) systems improve the information carrying capacity of optical telecommunications networks, with ever increasing operating speeds and channel densities . However, dense WDM systems, with channel spacings of 25GHz, require precise control of spectral features. The ability to dynamically add and drop channel information is of great importance in many multi-wavelength network architectures, and thus the development of dynamic, reconfigurable integrated optical devices in silica-on-silicon is a highly active field of research.
Integrated optical devices have the potential for efficient processing of network traffic, allowing low power consumption and efficient links between network subsystems. Such integrated optical devices often include planar Bragg gratings, which allow static wavelength filtering. However, the ability to tune the operation wavelength over several standard channel spacings could prove to be one of the key enablers in realizing an all-optical dynamic network. Bragg gratings are a key filter element in such optical networks, providing control over traffic in terms of both selectivity and routing. Finding a method of tuning these filters is of significant interest where planar devices show greater potential for integration than fibre-based technology. Such a device was proposed theoretically by Sirleto et al .
There are currently very few reports on electrically tunable planar Bragg gratings, while there is more literature on fibre gratings. This includes electromagnetically controlled stress tuning , which requires large (500mT) magnetic fields for moderate tuning of 1.3nm. Thermal tuning based on Ohmic heating of a metallic coated fiber has been demonstrated producing a shift of 4.1nm/W . Stress tuning has been shown to provide a shift of 0.185nm/µm, with a maximum tuning speed of 21nm/ms . Pure electrooptic tuning of fiber Bragg gratings has been demonstrated in thermally poled fiber, with tuning of 0.25pm/V . Finally, etched fibers using liquid crystal cladding have been shown to achieve 0.2nm tuning with an applied voltage of 1.6kV , far less than our planar equivalent.
Electrical tunability potentially allows for superior response times over the more common temperature tuned gratings. Electrically tunable devices use the principle of shifting the Bragg wavelength by modifying the effective refractive index of a waveguide in a multilayer substrate. These quantities are described by the equation: λB=2Λneff (where λB is the Bragg centre wavelength, Λ is the grating period, and neff is the effective index of the waveguide). In our work, tuning of the Bragg peak utilizes an adaptive material, a liquid crystal (LC), as an overlay on the grating .
In our previous work we determined the electrical tunability range of planar Bragg gratings with a liquid crystal overlayer. The devices were fabricated using a UV laser writing system. The samples contained waveguides and Bragg gratings that were defined using Direct Grating Writing , and then the overclad layer covering the gratings was removed using hydrofluoric acid wet etching. The Bragg response of our LC devices are then tuned via control of the bulk molecular orientation using an applied electric field. This exploits the refractive index anisotropy arising from the rod-like structure of the molecules. Modification of the LC refractive index subsequently alters the effective index seen by propagating light in the waveguide and grating structure, causing Bragg wavelength shift.
In the previous report, our Direct-UV-written Silica-on-silicon tunable LC devices exhibited a Bragg wavelength shift of 932.7pm (corresponding to 114GHz at λ=1561.8nm for TE polarization) with 170Vpp applied across the LC cell . However, it was observed that the wavelength shift is not a linear function of applied field, and indeed hysteretic behaviour was seen in the tuning curves.
The tuning curves for these devices are shown in Fig. 1 with hysteresis clearly evident between points A and B. The hysteretic behaviour was observed consistently with repeated measurements conducted over the course of several weeks. Insets (i) and (ii) show each part of the hysteresis curve separately. Inset (i) shows the tuning curve with increasing voltage whereas inset (ii) shows the curve with decreasing voltage. All measurements in the tuning curves for increasing voltage displayed two distinct points where the tuning gradient changes significantly. These points are circled in inset (i). The first lower threshold point (1) occurs at ~17Vpp, with increasing voltage, whereas the upper threshold point (2) occurs at ~57±10Vpp. The curve for decreasing voltage exhibits such a change at ~17Vpp (3).
In this paper we investigate this hysteresis by direct observation of polarisation state and related physical changes in the LC. We also provide information on environmental effects when exposing these devices to varying ambient temperatures. In particular, we investigate the possibility of sample heating when the cells are exposed to sustained high electric fields of order ~200V.
In our previous work, our Direct UV Written tunable LC devices were fabricated from silica-on-silicon substrates and were therefore opaque. To examine the LC alignment processes occurring in these devices, equivalent structures were fabricated from transparent materials for use in transmission microscopy experiments. Each LC cell consisted of an Indium Tin Oxide (ITO) coated glass substrate and a glass coverslip, and typically measured 10x20mm2 (Fig. 2). Two electrodes were prepared from the ITO layer by using a combination of spark erosion and mechanical etching to scribe a straight ~50µm wide gap into the ITO.
The required homeotropic alignment of the liquid crystal molecules was achieved by applying a thin layer of Merck Liquicoat ZLI-3334 (0.2% solution in ethanol) surfactant to the ITO electrodes and the coverslip. The ethanol was left to evaporate and both were subsequently baked at 120°C for 2 hours to ensure good surfactant adhesion. Spacers in the form of 12µm diameter glass rods suspended in UV curing adhesive were used to create a window for the LC, and to seal the cell. Merck nematic 18523 LC was then applied to the cell under vacuum. This LC was selected as it has a refractive index close to that of silica. Transmission microscopy measurements were then performed through crossed polarizers oriented at 45° with respect to the LC cell electrode structure. This provided the best contrast to observe any polarization change in the LC. A voltage was applied to the cell using a computer controller a.c. signal generator, connected through an amplifier with variable gain, providing a voltage range up to ~200Vpp. The resultant polarization changes inside the LC cell were recorded using a camera.
In order to confirm purely electrical Bragg peak tuning of our devices, experiments were undertaken to identify any related tuning behaviour caused by thermal effects. For this measurement the UV-written samples were not etched to avoid undesired exposure of the Bragg gratings and the LC cell, as illustrated in Fig. 3, was insulated for thermal stability. A thermocouple was used to measure temperature changes in the cell and a broadband ASE source and OSA were used to record the reflection spectra from the Bragg grating within each waveguide.
4. Results and discussion
From this investigation, we believe that the hysteretic behaviour seen in our previous work  can be linked to the formation and subsequent disappearance of a specific line defect in the LC, known as a disclination line. In a nematic LC the long axes of its constituent molecules are typically oriented in one preferred direction called the ‘director’ . However, in the case of defects such as disclination lines, the director is undefined giving rise to modified optical properties. The shape of the defect in our case is governed by the restricted geometry of the etched region in our samples. Thus as the disclination line forms above our waveguides, any change in index of the LC as a result of director realignment will manifest itself as a change in centre wavelength.
The transparent cell shows an area of undefined director between the electrodes that subsequently forms a disclination at higher voltages. This effect is demonstrated in the photographs of Fig. 4, which clearly show the onset of a disclination line (from polarisation change) at 25V, the disclination line becoming confined to the weak field region at 50-75V, and retreating at 100V as the LC becomes fully-aligned with the applied field. The disclination does not reappear until 25V when decreasing the voltage. Therefore critical voltage values agree well with tuning curve steps.
For the simple case of a homeotropically aligned nematic LC having its director rotated along field lines applied perpendicular to the waveguide, variation in the tuning wavelength is a simple function of applied voltage. This is because for LC’s with positive dielectric anisotropy the director is assumed to simply rotate under torque caused by the electric field. However, if defects are present, they will disturb this reorientation process induced by the electric field during tuning.
As the wavelength tuning curves in Fig. 1 are linear at higher voltages, it can be assumed that the ratio of disclination to regions of well defined director varies with voltage. Indeed, it can be seen that, referring to the 25V image, point defects are present along the line defect . These give rise to nucleation points for discontinuity in the disclination, allowing the disclination to retreat, from the lowest point defect as shown in the 100V image. Thus the percentage of the waveguide covered by disclination reduces with increasing voltage, and so the change in effective index is not a linear function of applied voltage.
An additional result to arise from these experiments is that the LC tuning of our samples causes the Bragg wavelength to decrease with increasing voltage. While it is expected that the refractive index change of TE and TM polarizations should tune in opposite directions, with one rising as the other falls, in our sample geometry both repeatedly tune in the same direction. While this is also likely due to orientation effects of the LC, this origin of anomalous result remains the subject of further study.
In order to confirm that the 100GHz wavelength tuning achieved by our devices is purely electric-field driven, and not the result of local heating in the LC, the long-term thermal effects of applied voltage were studied in our silica-on-silicon Bragg grating samples. If an applied voltage resulted in heating of the LC, the heat would diffuse through the silica layers causing them to expand and thus altering the grating pitch. This would subsequently cause a shift in the Bragg wavelength, typically 8-12pm/°C as found in similar samples . However, no such shift was observed. Figure 5(a) illustrates the thermal stability of our devices under ambient conditions, showing less than 10pm thermal drift during the 1 hour experiment. Figure 5(b) demonstrates thermal tuning by varying the temperature of our unetched sample with the use of a Peltier heater, showing a shift of approximately 20pm/°C.
Next, a voltage of 150Vpp was applied across the electrodes for extended periods to test for any Ohmic heating. Figure 6 shows the Bragg wavelength shift for a grating in the sample. It is apparent that little or no Ohmic heating was present, and here the drift of ~5pm can be attributed to change of ambient conditions, corresponding to ~0.5°C. This demonstrates that the measured peak Bragg reflection wavelength shift in our sample (932pm) is completely voltage driven.
We have investigated the behaviour of LC cells with physical attributes similar to those of our earlier LC tunable devices . Hysteresis seen in the tuning curves of these devices can be attributed to the presence of a disclination line forming due to the confined geometry of the ITO electrode structure. This behaviour was observed using transmission microscopy with polarized light. We propose that the origin of the hysteretic response of the tuning to applied field is due to the growth and disappearance of a disclination in the LC alignment between the electrodes. The percentage of grating covered by disclination to that covered by ordered LC varies with applied voltage, and is indeed not a linear function with increasing voltage.
The LC tunable planar Bragg grating devices are also demonstrated to be substantially free of Ohmic heat induced tuning. The drift in Bragg wavelength was measured to be less than 10pm in ambient conditions. The effects of sustained a.c. voltage applied to the device are negligible. We can therefore conclude that thermal effects had no major contribution to the tuning characteristics of the cells.
Our investigations confirm purely electrical tuning of Bragg wavelength in grating structures with minimal drift in centre wavelength over 114GHz, allowing them to span over 4 channels in a DWDM system. Such structures lend themselves to applications in telecommunications where fast switching times are vital. To improve switching times in our cells, we will consider laser patterning as a tool for patterning ITO coated glass. Potentially this allows for precise periodic structures in the ITO to act as nucleation sites for point defects to increase the speed of disclination retraction.
References and links
1. C. Dragone, “Low-loss wavelength routers for WDM optical networks and high-capacity IP routers,” J. Lightwave Technol. 23, 66–79 (2005). [CrossRef]
2. L. Sirleto, L. Petti, P. Mormille, G. C. Righini, and G. Abbate, “Fast Integrated Electro-Optical Switch and Beam Deflector Based on Nematic Liquid Crystal Waveguides,” Fiber Integr. Opt. 21, 435–449 (2002). [CrossRef]
3. X.-Z. Lin, Y. Zhang, H.-L. An, and H.-D. Liu, “Electrically tunable singlemode fibre Bragg reflective filter,” Electron. Lett. 30, 887–888 (1994). [CrossRef]
4. H. G. Limberger, N. H. Ky, D. M. Costantini, R. P. Salathé, C. A. P. Muller, and G. R. Fox, “Efficient miniature fiber-optic tunable filter based on intracore Bragg grating and electrically resistive coating,” IEEE Photon. Technol. Lett. 10, 361–363 (1998). [CrossRef]
5. A. Iocco, H. G. Limberger, R. P. Salathé, L. A. Everall, K. E. Chisholm, J. A. R. Williams, and I. Bennion, “Bragg grating fast tunable filter for wavelength division multiplexing,” J. Lightwave Technol. 17, 1217–1221 (1999). [CrossRef]
6. B. Srinivasan and R. K. Jain, “First Demonstration of Thermally Poled Electrooptically Tunable Fiber Bragg Gratings,” IEEE Photon. Technol. Lett. 12, 170–172 (2000). [CrossRef]
7. S. Baek, C. Jeong, S. Y. Noh, Y. Jeong, S.-D. Lee, and B. Lee, “Electrically tunable fiber Bragg gratings using liquid crystal cladding,” in Pacific Rim Conference on Lasers and Electro-Optics, CLEO-Technical Digest, (CLEO/Pacific Rim, 2005), pp. 1078–1079.
8. I. J. G. Sparrow, D. A. Sager, C. B. E. Gawith, P. G. R. Smith, G. D. Emmerson, M. Kaczmarek, and A. Dyadyusha, “25GHz tunability of planar bragg grating using liquid crystal cladding and electric field,” Quantum Electronics and Laser Science Conference 2, 963–965 (2005). [CrossRef]
9. G. D. Emmerson, S. P. Watts, C. B. E. Gawith, V. Albanis, M. Ibsen, R. B. Williams, and P. G. R. Smith, “Fabrication of directly UV-written channel waveguides with simultaneously defined integral Bragg gratings,” Elec. Lett. 38, 1531–1532 (2002). [CrossRef]
10. F. R. M. Adikan, J. C. Gates, A. Dyadyusha, H. E. Major, C. B. E. Gawith, I. J. G. Sparrow, G. D. Emmerson, M. Kaczmarek, and P. G. R. Smith, “Demonstration of 100GHz electrically tunable liquidcrystal Bragg gratings for application in dynamic optical networks,” Opt. Lett. 32, 1542–1544 (2007). [CrossRef] [PubMed]
11. P. J. Collings and M. Hird, Introduction to Liquid Crystals, (Taylor & Francis, London, 1997). [CrossRef]
12. I. J. G. Sparrow, G. D. Emmerson, C. B. E. Gawith, P. G. R. Smith, M. Kaczmarek, and A. Dyadyusha, “First order phase change detection using planar waveguide Bragg grating refractometer,” Appl. Phys. B 81, 1–4 (2005). [CrossRef]