An integrated optical waveguide variable optical attenuator (VOA) made of organic/inorganic hybrid materials was fabricated. At 1550 nm, the VOA showed a very low activation power of about 13 mW, due to the large thermo-optic coefficients of the hybrid materials. The optical power attenuations achieved were more than 25 dB for both TE and TM polarization. The response time of the device was less than 4.7 ms.
© 2006 Optical Society of America
Variable optical attenuators (VOA) are essential components in optical transmission systems. They play important roles in flattening the gain spectrum of optical amplifiers, equalizing channel power levels in wavelength division multiplexed (WDM) system, and reducing power fluctuation. VOAs based on planar lightwave circuits (PLC) are very attractive for their flexibility in structure design, compact volume and the ability to hybrid integrate with other photonics devices. Several types of the integrated optical waveguide VOAs based on thermooptic (TO) effect have been reported [1–8] including the Mach-Zehnder interferometric (MZI) type. The most popular materials used to fabricate thermo-optic integrated optical waveguide VOAs are silica [1–3] and polymers [4–8]. Although silica offers excellent properties such as low propagation loss and high reliability, a VOA based on silica usually requires high driving powers due to its small thermo-optic coefficient and high thermal conductivity. On the other hand, a polymer-based thermo-optic VOA has the advantage of low power consumption because the thermo-optic coefficients of polymers are about one order of magnitude higher than that of silica. However, polymer-based VOAs usually have problems on thermal stability and long-term reliability.
Recently, sol-gel derived organic/inorganic hybrid materials are considered particularly suitable to fabricate waveguide devices, because they retain advantages of both silica and polymer. These hybrid materials usually have higher glass transition temperatures and higher deterioration temperatures than pure polymers do, and also have large thermo-optic coefficients, which are in the same order as polymers (about 1×10-4/°C) [9,10]. In this paper, we report for the first time to our knowledge, a Mach-Zehnder interferometric type thermooptic waveguide VOA fabricated with sol-gel derived organic/inorganic hybrid materials. The VOA shows several good characteristics including low power consumption of about 13 mW, large applicable wavelength range and fast response time. The polarization dependent loss (PDL), the insertion loss and thermal stability of the device are also investigated and analyzed.
2. VOA design and fabrication
Figure 1(a) shows the configuration of the proposed MZI type VOA. The VOA is comprised of input/output single-mode waveguides, S-bend splitter and combiner, and two straight phase arms with thin-film heaters. Both of the width and height of single-mode waveguide are designed to be 7 µm in order to optimize the device coupling efficiency to 1550 nm singlemode fibers. S-bend connectors are 3 mm long. Two phase arms are separated by 250 µm and are 3 mm long. The electrode heaters are located directly on the top of both phase arms with a width of 50 µm and a length of 3 mm.
Hybrid organic/inorganic materials used in this study were synthesized by using methacryloxypropyl trimethoxysilane (MAPTMS), zirconium n-propoxide (ZPO), methacrylic acid (MAA) and n-propanol as the sol-gel precursors. The detailed synthesis procedure can be found in our previous paper . Two types of hybrid materials were prepared, with ZPO molar ratios of 20% and 24.5%. They were used as the cladding layer and the core layer respectively. The refractive indices of the cladding and the core layer at 1550 nm are 1.5111 and 1.5179 respectively, measured by prism-coupling method. The relative refractive index difference Δ is 0.45%, which was chosen to ensure the single mode condition for the designed waveguide cross section. The thermo-optic coefficients of these two materials are very close and have values of-1.79×10-4/°C at the wavelength of 632.8 nm.
The VOA was fabricated on silicon wafer by the spin-coating method and followed by direct UV patterning. The under-cladding about 13 µm thick was first fabricated, then a stripe waveguide. A microscope image of the cross section of the stripe waveguide showed in Fig. 1(b) reveals a nearly square shape with a side length of about 7 µm. The surface roughness of the stripe waveguide measured by atomic force microcopy (AFM) was less than 8 nm. The well-defined shape and the smooth surface are necessary to fabricate waveguides with low propagation loss. A 7 µm thick up-cladding layer was then spin-coated on the top of the patterned device. Ni-Cr-Al alloy electrodes were evaporated by a vacuum deposition method and photo lithographed afterwards. The resistance between two contact pads was measured to be 1.6 kΩ. Finally, the fabricated device was cleaved for testing.
3. Results and discussions
The power consumption and dynamic range of the VOA device were measured by coupling 1550 nm light into the waveguide by a 40× microscope objective. A Glan prism was placed in front of the objective to change the polarization of the input light. Output light was projected directly into a single-mode fiber and detected by a power-meter.
Figure 2 shows the optical attenuation of the VOA as a function of the electrical power applied to one electrode. The maximum attenuations are -31 dB and -25 dB for TE and TM mode respectively. The activation powers, defined as the electrical power required to reach the maximum attenuation that corresponds to a π phase shift, are as low as 13.2 mW and 12.5 mW. As expected, these values are at least one order smaller than that of VOA based on silica materials [1–3] and are comparable with the polymer device . The activation power is related to the thermal properties of materials as well as the device structure. We used a 2D thermal field model  to calculate activation power and analyze the effect of cladding layer on the activation power. The results are shown in Fig. 3. The calculated value of activation power for the VOA structure (7 µm up-cladding/7 µm core/13 µm under-cladding) is around 3 mW. In the calculation, we took the thermal conductivity of the hybrid materials to be the same as that of a polymer, which is 0.17 W/mK. However, since the silica has a higher thermal conductivity (1.38 W/mK) than polymer, the real thermal conductivity of hybrid materials, consisting of organic and inorganic components, might be larger and thus the theoretical activation power should increase a little. But the measured activation power is still higher than the calculated one. The discrepancy was mainly resulted from the resistance distribution of the driving electrical circuit. After careful measurement, the effective resistance on top of the waveguide arms is found to be less than one half (0.8 KΩ) of the total resistance between connecting pads. Therefore the actual activation power should be below 6.5 mW instead. The electrical circuit can be optimized to lower the activation power further. Figure 3 also illustrates that thinner up-cladding layer and thicker under-cladding layer will dramatically reduce the activation power.
The response time of the device was measured by applying a 50 Hz square wave voltage to an electrode. The rise time of the VOA is 3.7 ms and the falling time is 4.7 ms, these values are close to the response times of the polymer device .
The wavelength dependence of attenuation of the device from 1480 nm to 1630 nm for TE polarization is plotted in Fig. 4. When the attenuation at 1550 nm is -10 dB, the spectral variation is ±2 dB. The spectral variation increases to ±5 dB when the attenuation increases to -22 dB. The spectral variation also deteriorates heavily as wavelength increases beyond 1580 nm. The wavelength dependence of attenuation comes from the 1/λ dependence of the phase-shift, which may due to the dispersion of phase imbalance and the thermo-optic coefficient dependence on wavelength. Figure 5 plots the activation power as a function of wavelength. The longer the wavelength, the more power a VOA needs to reach the maximum attenuation. The highest activation powers are about 14 mW for TE polarization and 13 mW for TM polarization at 1630 nm. The maximum attenuations are measured to be over -24 dB from 1480 nm to 1630 nm. For broad spectral usage, the wavelength dependence of the VOA devices can be compensated by a feedback system.
At a certain wavelength, the attenuation difference between the TM and the TE polarization for a specific power is defined as the PDL. The intrinsic PDL (passive component) of the VOA measured is 0.2 dB. PDL as a function of the TM polarization attenuation amplitude is plotted in Fig. 6. It can be seen that the PDL becomes large at high attenuation level. At 1550 nm, PDL is 2 dB and 6 dB at the attenuation -10 dB and -20 dB for TM polarization, respectively. Two reasons are responsible for the PDL of the VOA device; one is the different thermo-optic coefficients for the TE and TM polarization due to the anisotropic thermal expansion , and another is the asymmetric layer structure of the device. To estimate the contribution of asymmetric waveguide structure on the PDL, the TE and TM mode shapes of a straight waveguide with a cross section shown in Fig. 1(b) were simulated using the BeamPROPTM software (Rsoft). The results show that the difference of optical transmission between TE and TM polarization due to the waveguide geometrical asymmetry is negligible. So the anisotropic thermal expansion is considered to the main reason for device PDL. PDL can be reduced by using a substrate that has a thermal expansion coefficient similar to the hybrid material .
The total insertion loss of the VOA device (1cm long) measured by a fiber-waveguide-fiber system was 7.5 dB at 1550 nm and 2.6 dB at 1310 nm including reflection loss. The propagation loss of a straight waveguide was measured to be 0.5dB/cm at 1310 nm and 4.5dB/cm at 1550 nm respectively. The propagation loss of 0.5dB/cm at 1310nm is mostly from light scattering, because the absorption loss is very low at that wavelength. The scattering loss is theoretically lower at longer wavelength, thus the absorption loss at 1550 nm is at least 4 dB/cm, which mainly comes from the residual O-H absorption . Synthesis of the organic/inorganic hybrid materials by the non-hydrolytic sol-gel process is a promising method to eliminate O-H groups and is the object of the future studies. Compared with the estimated values of the straight waveguide at 1550 nm, the rest 3 dB loss of the 1 cm long VOA device might be caused by the Fresnel reflection loss and coupling loss. These losses can be substantially reduced after careful optical packaging or using the matching oil.
To evaluate the thermal stability of the device, waveguides made by the organic/inorganic hybrid materials were annealed at 250°C for 3 hours, no noticeable geometric deformation was observed. No apparent Tg could be found for the hybrid material from Differential Scanning Calorimetry (DSC) measurement up to 300°C. Consequently, we believe that the device can withstand much high temperature than devices made of pure polymer of the same organic moiety (PMMA), which has a Tg around 100°C.
The thermal dependence of the activation power and maximum attenuation of the VOA device were also measured in a temperature range from 25.0°C to 96.5°C (96.5°C was the highest temperature we could go in in situ optical measurement). No significant decrease of the device functionality was found. The attenuation was fluctuating within 13% for the whole temperature range; in the meantime, the fluctuation of the activation power is 15%. The results strongly indicate that VOAs with organic/inorganic hybrid materials have much better thermal stability.
A VOA based on organic/inorganic hybrid materials was fabricated and characterized. The maximum attenuations of -25 dB for TM polarization and of -31 dB for TE polarization were achieved with low power consumptions of about 13 mW at 1550nm. The power consumptions are much smaller than silica-based VOA and are comparable to the polymer-based devices. The spectral variation is ±2 dB at -10 dB attenuation and ±5 dB at -22 dB attenuation within wavelength range of 1480 nm to 1630 nm. The response time is 3.7 ms for the rising time and 4.7 ms for the falling time. The organic/inorganic hybrid materials can provide an alternatively way to fabricate thermal-active photonic devices with excellent properties.
This work was supported by National Natural Science Foundation of China under Grant #60478005, #60378034, #10474015, #10574032, #50532030, and Shanghai Science and Technology Commission under grant #04DZ14001.
References and links
1. T. Kawai, M. Koga, M. Okuno, and T. Kitoh, “PLC type compact variable optical attenuator for photonic transport network,” Electron. Lett. 34, 264–265 (1998). [CrossRef]
2. M. Svalgaard, K. Farch, and L.-U. Andersen, “Variable optical attenuator fabricated by direct UV writing,” J. Lightwave Technol. 21, 2097–2103 (2003). [CrossRef]
3. T. Hurvitz, S. Ruschin, D. Brooks, G. Hurvitz, and E. Arad, “Variable optical attenuator based on ionexchange technology in glass,” J. Lightwave Technol. 23, 1918–1922 (2005). [CrossRef]
4. T. Shioda, N. Takamatsu, K. Suzuki, and S. Shichijyo, “Polarization dependence of Mach-Zehnder interferometer switch using fluorinated polyimide waveguide,” Jpn. J. Appl. Phys. 42, L30–L32 (2003). [CrossRef]
5. S. -S. Lee, Y. -S. Jin, Y. -S. Son, and T. -K. Yoo, “Polymeric tunable optical attenuator with an optical monitoring tap for WDM transmission network,” IEEE Photon. Technol. Lett. 11, 590–592 (1999). [CrossRef]
6. Y. O. Noh, M. -S. Yang, Y. H. Won, and W. -Y. Hwang, “PLC-type variable optical attenuator operated at low electrical power,” Electron. Lett. 36, 2032–2033 (2000). [CrossRef]
7. S. M. Garner and S. Caracci, “Variable optical attenuator for large-scale integration,” IEEE Photon. Technol. Lett. 14, 1560–1562 (2002). [CrossRef]
8. G. Z. Xiao, Z. Y. Zhang, and C. P. Grover, “A variable optical attenuator based on a straight polymer-silica hybrid channel waveguide,” IEEE Photon. Technol. Lett. 16, 2511–2513 (2004). [CrossRef]
9. E. -S. Kang, T. -H. Lee, and B. -S. Bae, “Measurement of the thermo-optic coefficients in sol-gel derived inorganic-organic hybrid material films,” Appl. Phys. Lett. 81, 1438–1440 (2002). [CrossRef]
10. Xianjiang Wang, Lei Xu, Dongxiao Li, Liying Liu, and Wencheng Wang, “Thermo-optic properties of solgel-fabricated organic-inorganic hybrid waveguides,” J. Appl. Phys. 94, 4228–4230 (2003). [CrossRef]
11. W. -K. Wang, H. J. Lee, and P. J. Anthony, “Planar silica-glass optical waveguides with thermally induced lateral mode confinement,” J. Lightwave Technol. 14, 429–436 (1996). [CrossRef]
12. M. Oubaha, M. Smaihi, P. Etienne, P. Coudray, and Y. Moreau, “Spectroscopic characterization of intrinsic losses in an organic-inorganic hybrid waveguide synthesized by the sol-gel process,” J. Non-Crystal. Solids 318, 305–313 (2003). [CrossRef]