We describe a novel approach for CMOS-compatible passively temperature insensitive silicon based optical devices using titanium oxide cladding which has a negative thermo-optic (TO) effect. We engineer the mode confinement in Si and TiO2 such that positive TO of Si is exactly cancelled out by negative TO of TiO2. We demonstrate robust operation of the resulting device over 35 degrees.
© 2013 Optical Society of America
Integrated photonic devices are extremely sensitive to ambient temperature fluctuations which limit their integration in wavelength sensitive applications. This problem is especially severe in silicon photonics where high index contrast, large thermo-optic coefficient of Si  and high quality factors make the microring resonators extremely susceptible to thermal fluctuations. For example, a Si ring resonator with a quality factor of 10,000 will tune out of resonance with only 1 °C change in temperature.
Solutions to overcome the temperature sensitivity have been proposed using polymers, external temperature compensating devices or active control of device temperature. Table 1 summarizes previously reported temperature stabilization schemes and their corresponding advantages and disadvantages.
Here we describe a novel approach for passive and CMOS-compatible temperature insensitive integrated optical devices using a metal-oxide cladding having negative thermo-optic effect. Metals oxides like TiO2 and SrTiO3 have been investigated in semiconductor industry as a gate dielectric . These metal oxides also have negative thermo-optic coefficient due to presence of a soft electronic band  (TOTiO2 ∼ −1 × 10−4K−1, TOSrTiO3 ∼ −1 × 10−5K−1). This is in contrast to commonly available dielectrics and semiconductors which have positive thermo-optic coefficients (TOSi = 1.8 × 10−4K−1, TOSiN ∼ 2 × 10−5K−1, TOSiO2 ∼ 1 × 10−5K−1). There has been some limited effort in reducing temperature sensitivity of optical devices using TiO2 overcladding [17–19]. Ref  has done a thorough analysis of TiO2 deposition and showed close to athermal performance. Here we show that completely athermal optical devices can be realized by engineering the mode-overlap between Si based materials (Si / SiN/ SiO2) and TiO2. This scheme of temperature compensation is CMOS-compatible, lossless, does not require any extra footprint and can lead to very large temperature operating range.
2. Athermal design
In order to ensure that the positive TO effect of the Si core is exactly cancelled out by negative TO effect of the cladding, we engineer the optical mode confinement in TiO2 cladding layer by tailoring the waveguide dimension and the cladding thickness. We consider both transverse electric (TE) and transverse magnetic (TM) like optical modes in a 220nm thick Si guiding layer with varying waveguide widths and TiO2 cladding thicknesses. Figure 1(a) shows the Poynting vector of the TE mode of a 250nm wide hybrid TiO2 - Si waveguide and temperature sensitivity of the resonance wavelength, when the TE mode is in resonance. Figure 1(b) shows the same for the TM mode. Optical modes are simulated using a Finite Element Method based solver (COMSOL), assuming the thermo-optic coefficients mentioned previously. The refractive indices of TiO2 was assumed to be 2.35 (measured using ellipsometry). Temperature sensitivity is characterized in terms of resonance wavelength change with temperature ( ). It is important to note that this sensitivity is independent of device structure (ring resonators or photonic crystal cavitites) and depends only on modal confinement in Si. For both TE and TM polarizations, the resonance wavelength sensitivity is ∼ 0.1nm/K for modes strongly confined in Si, and without any TiO2 cladding. This sensitivity decreases as the mode is delocalized into the TiO2 cladding and the thickness of the TiO2 cladding is increased. For achieving true athermal operation, the exact geometry of the Si waveguide and TiO2 cladding thickness needs to be chosen very carefully. For TM mode of a 220nm thick silicon waveguide, any waveguide width will have a corresponding TiO2 cladding thickness that would allow athermal operation. On the other hand, for the TE mode, if the waveguide is too wide such that the optical mode is strongly confined in Si, athermal operation will not be possible irrespective of TiO2 cladding thickness.
The athermal ring resonators were fabricated on a 220nm thick silicon-on-insulator (SOI) device layer. The waveguides were patterned using electron beam lithography and etched in chlorine chemistry in an inductively coupled reactive ion etcher. Titanium oxide was deposited on top of the waveguides using reactive sputtering of a titanium target in O2. The sputtering was performed at a pressure of 2× 10−6 Torr and 2kW power. Sputtered TiO2 films were characterized using a visible-near IR ellipsometer and Raman spectroscopy. Refractive index of ∼ 2.35 was measured at near IR wavelengths. Raman spectrum of the sputtered sample shows no visible peaks, indicating the amorphous nature of the deposited TiO2 film. AFM scan of the surface (Fig. 2(b)) indicates surfaces roughness below 2 nm RMS. Figure 2(a) shows a microscope image of a fabricated Si ring resonators with TiO2 cladding. Si waveguide width in the fabricated ring resonators was varied from 450nm to 150nm to observe the effect of mode delocalization on thermal sensitivity. Inset of Fig. 2(a) shows cross section of a waveguide with 150nm width and 200nm thick TiO2 cladding.
4. Experimental results
We demonstrate temperature insensitive operation of a Si microring resonator with TiO2 over-cladding. Figure 3(a) shows the temperature sensitivity of resonance wavelength of the fabricated microring resonators, for different waveguide widths and polarizations. All of these resonators were fabricated on the same chip. The temperature sensitivity was measured by collecting spectra over 10 degrees, at intervals of 2 degrees, followed by a linear fit to resonance wavelenths. Error bar corresponds to uncertainty in the linear fit. Each waveguide width corresponds to a different confinement in the Si core. Measured data is compared against numerically calculated sensitivity assuming TiO2 thickness of 200nm and λ0 = 1550nm. For the case of TE polarization (Fig. 3(a) blue line), the temperature sensitivity is around 0.09nm/K when the mode is strongly confined in Si (450nm wide waveguide). However as the optical mode is delocalized more into TiO2, temperature sensitivity decreases and becomes negative (∼ −0.03nm/K) for 150nm wide waveguide. For the case of TM polarization (Fig. 3(a) red line), the sensitivity is around 0.04nm/K for a strongly confined mode in Si and very close to zero for 150nm wide waveguide. Figure 3(b) shows the corresponding athermal transmission spectrum compared to that of a resonator without any thermal compensation. Losses in the deposited TiO2 was estimated to be around 16 dB/cm, by comparing change in quality factor of the resonances as a function of mode confinement in Si and TiO2. This loss can be reduced significantly by improving the deposition and reducing scattering at Si – TiO2 interface. Optical losses in similar material has been reported to be less than 3 dB/cm [20, 21].
We demonstrate error free operation of hybrid TiO2 – Si microring resonator based optical filter over 35 degrees. The device used was similar to the one shown in Fig. 2(a) with an extra drop waveguide coupled to the ring. We transmitted 231 − 1 PRBS (pseudo random binary sequence) data at 5 Gbps through the device, centered at resonance wavelength, and varied the stage temperature (using a thermoelectric stage and temperature controller). The data was then sent to a commercial receiver (Picometrix PT15) and a bit error rate detector. Figure 4(a) shows the bit error rate (BER) and corresponding eye diagrams for both the athermal ring resonator and a conventional Si resonator with similar quality factor (where the temperature sensitivity is around 0.09 nm/K). For the conventional uncompensated device, BER becomes greater than 1E-9 after only 2 degrees (Fig. 4(a) blue line). For our athermal device, data transmission is close to error free over 35 degrees (Fig. 4(a) red line). This operating range should extend even further. Slight variations in BER is due to fluctuation in fiber to waveguide coupling with temperature. We also characterized the power penalty of the athermal optical filter for a 1°C fluctuation in temperature, and compared it to an uncompensated Si resonator based filter . The 1°C temperature fluctuation was introduced by placing the sample on a temperature controlled stage and modulating the stage temperature at a rate of 1 Hz. This amount of temperature fluctuation is small enough such that the waveguide to fiber coupling is minimally perturbed, while modeling a realistic operating condition. The conventional resonator has a power penalty > 1dB, while the hybrid Si – TiO2 resonator has power penalty < 0.1dB (Fig. 4(b)).
In conclusion, we demonstrated a new technique for realizing athermal Si photonic devices using a TiO2 over cladding. This approach is CMOS compatible, low loss and yields large temperature operating range. Simulations show that athermal operation is relatively insensitive to slight variations in TiO2 thickness. Engineering the mode confinement is extremely critical for achieving athermal operation. It should be noted that this method is mainly applicable for resonators with radius > 10μm radius due to the need for a slightly delocalized mode. This method of passive athermalization can lead to practical monolithic integration of silicon photonic devices.
The authors would like to thank J. Drumheller for his help in developing TiO2 deposition recipe and C. Phare for his help in Raman measurements. All the fabrication work, including TiO2 deposition, was done at the Cornell NanoScale Facility, a member of the National Nanotechnology Infrastructure Network, which is supported by the National Science Foundation (Grant ECCS-0335765). The authors gratefully acknowledge support from CIAN (Center for Integrated Access Networks) for funding this work Grant EEC-0812072.
References and links
1. Y. Varshni, “Temperature dependence of the energy gap in semiconductors,” Physica 34, 149–154 (1967) [CrossRef]
2. P. Alipour, E. Shah Hosseini, A. A. Eftekhar, B. Momeni, and A. Adibi, “Temperature-insensitive silicon microdisk resonators using polymeric cladding layers,” in Conference on Lasers and Electro-Optics (Optical Society of America, 2009).
4. J. Teng, P. Dumon, W. Bogaerts, H. B. Zhang, X. G. Jian, X. Y. Han, M. S. Zhao, G. Morthier, and R. Baets, “Athermal Silicon-on-insulator ring resonators by overlaying a polymer cladding on narrowed waveguides,” Opt. Express 17, 14627–14633 (2009). [CrossRef] [PubMed]
5. C. T. DeRose, M. R. Watts, D. C. Trotter, D. L. Luck, G. N. Nielson, and R. W. Young, “Silicon microring modulator with integrated heater and temperature sensor for thermal control,” in Conference on Lasers and Electro-Optics (Optical Society of America, 2010).
6. P. Dong, W. Qian, H. Liang, R. Shafiiha, N.-N. Feng, D. Feng, X. Zheng, A. V. Krishnamoorthy, and M. Asghari, “Low power and compact reconfigurable multiplexing devices based on silicon microring resonators,” Opt. Express 18, 9852–9858 (2010). [CrossRef] [PubMed]
8. K. Padmaraju, D. F. Logan, X. Zhu, J. J. Ackert, A. P. Knights, and K. Bergman, “Integrated thermal stabilization of a microring modulator,” in Optical Fiber Communication Conference (Optical Society of America, 2013).
9. W. Zortman, A. Lentine, D. Trotter, and M. Watts, “Integrated CMOS compatible low power 10Gbps silicon photonic heater-modulator,” in Optical Fiber Communication Conference (Optical Society of America, 2012). [CrossRef]
10. C. Qiu and Q. Xu, “Wavelength tracking with thermally controlled silicon resonators,” in CLEO: Science and Innovations (Optical Society of America, 2011).
11. E. Timurdogan, A. Biberman, D. C. Trotter, C. Sun, M. Moresco, V. Stojanovic, and M. R. Watts, “Automated wavelength recovery for microring resonators,” in CLEO: Science and Innovations (Optical Society of America, 2012).
15. S. A. Campbell, H.-S. Kim, D. C. Gilmer, B. He, T. Ma, and W. L. Gladfelter, “Titanium dioxide (TiO2)-based gate insulators,” IBM journal of research and development 43, 383–392 (1999). [CrossRef]
16. V. Trepakov, A. Dejneka, P. Markovin, A. Lynnyk, and L. Jastrabik, “A ‘soft electronic band’ and the negative thermooptic effect in strontium titanate,” New J. Phys. 11, 083024 (2009). [CrossRef]
17. B. Guha and M. Lipson, “Athermal silicon ring resonator with bi-material cantilever for passive thermal feedback,” in CLEO: Science and Innovations (Optical Society of America, 2013).
18. S. S. Djordjevic, K. Shang, B. Guan, S. T. Cheung, L. Liao, J. Basak, H.-F. Liu, and S. Yoo, “CMOS-compatible, athermal silicon ring modulators clad with titanium dioxide,” Opt. Express 21, 13958–13968 (2013). [CrossRef] [PubMed]
19. F. Qiu, A. M. Spring, F. Yu, and S. Yokoyama, “Complementary metaloxidesemiconductor compatible athermal silicon nitride/ titanium dioxide hybrid micro-ring resonators,” Appl. Phys. Lett. 102, 051106 (2013). [CrossRef]
20. J. T. Choy, J. D. Bradley, P. B. Deotare, I. B. Burgess, C. C. Evans, E. Mazur, and M. Loncar, “Integrated TiO2resonators for visible photonics,” Opt. Lett. 37, 539–541 (2012). [CrossRef] [PubMed]
21. J. D. Bradley, C. C. Evans, J. T. Choy, O. Reshef, P. B. Deotare, F. Parsy, K. C. Phillips, M. Loncar, and E. Mazur, “Submicrometer-wide amorphous and polycrystalline anatase TiO2waveguides for microphotonic devices,” Opt. Express 20, 23821–23831 (2012). [CrossRef] [PubMed]