Open Access
Add to BibSonomy
Abstract
The use of integrated silicon photonic devices has been critical in the advancement of data communication technologies. However, the fabrication and operation of active photonic devices for these applications is often complicated and costly. By using an external 478nm laser source, we have demonstrated the ability to opto-thermally tune passive photonic devices. We have demonstrated various methods of thermal tuning using an external laser source with tuning capabilities up to 24.4 pm/mW for a passive Si3N4 chip with SiO2 cladding material. By etching the cladding layer using standard reactive ion etching techniques to better thermally isolate the individual resonators, we increased the tuning capability to 44.4 pm/mW. In this way, we have successfully tuned the resonance of a passive photonic chip without the use of electrical contacts or thermal electric devices. This poses potential alternatives to conventional thermal tuning techniques.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
The role of photonic integrated circuits (PICs) and microresonators in data communication, processing, collection, and storage is instrumental in meeting the increasingly high demands of internet traffic and data communications [1]. Although exceedingly useful in these areas, PICs can be complicated in their fabrication and operation within optical systems. The tuning of optical resonators to meet specific operational parameters is crucial to the functionality of these systems, although current methods of thermally tuning the resonance wavelength of these resonators requires either bulky thermal stages that operate with high power consumption and take a considerable time to stabilize, or intricate fabrication techniques to place heaters directly onto or within PICs [2]. By using an external short wavelength light source, we can selectively thermally tune passive PICs without the need for complex heating systems. The no-contact simplicity of this method offers a potential alternative for characterizing optical microring resonators which may be effective for quality and process control purposes within a manufacturing environment.
The thermo-optic effect describes how a change in the temperature of an optical material will alter its index of refraction. The amount of change is quantified by the thermo-optic coefficient, dn/dT [3]. When the refractive indices of the materials of a waveguide change, the effective index of that waveguide is subsequently changed [4]. For ring resonators, this causes a shift in the wavelength at which the ring resonates. This resonance shift is described by the equation
A further advantage is the ability to focus our laser onto specific areas of our integrated photonic chip, thereby allowing us to isolate which passive devices can be optically tuned in a manner not possible with a thermal stage and much more difficult with on-chip heater systems.
2. Methods
For this work we used a passive photonic chip with Si3N4 waveguides and ring resonators with width and thickness of 1.5 µm and 220nm respectively, a silicon substrate and SiO2 cladding material of 2 µm and 5.48 µm respective thicknesses. The entire chip dimensions are 8.5mm × 6mm, and the rings measure 300µm in diameter. The relevant layout of this chip, with four sets of ring resonators can be seen in Fig. 1. By fiber coupling PriTel Inc. FA model optical fiber amplifier laser source into the input facets of the chip and fiber coupling the output facets into a Yokogawa AQ6370B optical spectrum analyzer, we were able to measure the resonance peaks of the microrings as a function of tuning parameters. Although our laser source was well resolved from 1520–1580 nm, our measurements only included resonance peaks between 1546 and 1548 nm. The resonance peak measured was the most prominent extrema in the optical spectra for the four drop ports and the through port as applicable.
By securing the sample on a thermal heating stage, we measured the values of the resonant wavelength as a function of stage and chip temperature which were assumed to be the same.
Two methods of opto-thermal tuning using external 478nm laser diode sources from Nichia are explored in this paper: full-chip tuning, and targeted tuning. For the first method, the photonic chip was fiber coupled to optical fiber using standard butt-coupling techniques and a 478nm laser source was mounted and used to illuminate the photonic chip directly. The spot size of this laser source measured 6.5mm × 7mm along the principal axes of the laser beam. The resonance wavelength was measured as a function of the 478nm laser power, which is assumed to be proportional to the temperature of the chip. The laser power was measured using a Newport 818-SL power detector. The second method involved mounting the 478nm source in a microscope and directing the laser beam onto the individual microrings and measuring the shift in resonant wavelength of that ring as a function of laser power. The spot size for this method after the laser exited the microscope was roughly 300µm × 600µm along the principal axes of the elliptical beam. The beam was oriented so that the longer axis extended vertically in relation to the rings so that the beam would not be directed onto any of the neighboring rings. Additionally, we measured the shift of one resonator induced by directing our laser on each of its three neighboring resonators. From this we were able to demonstrate how much thermal isolation we were able to achieve across the chip and how the distance separation between the external beam and the resonator affected the wavelength shift in that resonator.
In an attempt to further increase the thermal isolation of each ring, the SiO2 cladding was etched away to hinder heat conduction between the resonators during the targeted tuning method. To achieve this, we etched away the oxide cladding layer using standard reactive ion etching techniques using a Plasmatherm Versaline DSE III reactive ion etcher. Post-etching, we measured the individual resonator shifts at varying laser powers, and the shift in one resonator that resulted from the illumination of each of its neighbors. From this we calculated the wavelength shift as a function of laser power, as well as the potential for thermal isolation between resonators and compared with the pre-etching results.
3. Results and discussion
Prior to the laser tuning experiment, we measured the thermal characteristics of the photonic chip and found that the microresonator resonance position shifted by an average value of 16.1pm/°C for the five drop ports. These results can be seen in Fig. 2. This is consistent with the relatively low thermo-optic coefficient of silicon nitride (2.45 × 10−5/°C) and silicon dioxide (0.86 × 10−5/°C) [9,10].
Full-chip laser tuning demonstrated an average shift of 16.6 pm/mW as shown in Fig. 3 indicating that one milliwatt of laser power resulted in an increase in temperature of about 1°C. One limitation observed with this method came from our laser power. The optical power we were able to supply to the chip using the external laser source did not offer as high a temperature nor as large an overall shift as we were able to achieve with direct temperature control. With a more robust laser source, this method would likely show a comparable shift in wavelength as compared to traditional thermal tuning methods.
The results of the targeted laser tuning for each illuminated ring demonstrated an average resonance wavelength shift of 24.4 pm/mW and can be seen in Fig. 4. There was a 21% reduction in optical power through our microscope setup, so the amount of laser power on each ring was dramatically reduced as compared to the full chip method of exposure. This resulted in a smaller absolute shift in resonant wavelength, but this method still demonstrated a larger increase in the resonant shift as a function of laser power when compared to the full-chip exposure method.
Using a standard SiO2 etch recipe for the Plasmatherm reactive ion etcher, we etched off approximately 5.36 µm of the silica layer and left the Si3N4 waveguides and ring resonators exposed. This left us with about 100nm of partially-exposed waveguide on a 2 µm silica substrate with 120 nm of silica cladding remaining on the sides of the waveguide structures. We then measured the shift for each resonator using the targeted laser tuning method (Fig. 5). We were able to demonstrate a much larger shift of about 43.3 pm/mW in the resonant wavelength post-etch. This is likely due to an increase in temperature within the targeted area since heat will not easily dissipate without the cladding material, due to the thermal conductivity of air being much smaller than the thermal conductivity of SiO2 (0.0262 W/mK and 1.4 W/mK, respectively [11,12]). The higher thermal conductivities of Si and Si3N4 indicate that there will still be heat dissipation through the chip via the substrate material, but it will be less than was demonstrated pre-oxide-etch [13,14].
A summary of results for the various tuning methods, both pre- and post-oxide etching can be found in Table 1.
Table 1. Summary of the different tuning methods and their effectiveness
The Q factor of each resonator before and after etching the cladding layer was measured by dividing the resonant wavelength by the FWHM of the resonance transmission peak. The initial average Q factor for the four resonators was 2.05 × 104, and the average Q factor post-etch is 2.62 × 104. Overall, the change in Q factor was not significant, but the etching did slightly improve the operation of our resonators. This is likely due to a better-confined mode resulting from the increase in index contrast which subsequently decreased the interaction with the sidewalls reducing scattering and bend loss [5].
Using the same previous method of targeted laser tuning we demonstrated that before etching, the wavelength shift corresponding with the illumination of a reference ring resonator at a maximum power, as well as the illumination of its 3 resonator neighbors showed no significant variation in the shift as the laser beam was moved further away. This is what we expected to see with a SiO2 cladding material as it will more easily facilitate heat conduction across the chip than will an air cladding. The beam diameter of our 478nm source is slightly larger than the diameter of our rings, so there will be some extra heating of the chip surrounding the ring. The small separation between resonators compared to the laser focus area, and the efficiency of heat dissipation through the chip is likely the cause of this lack of thermal isolation.
After etching away the SiO2, we measured the shift in the resonant wavelength of the reference ring with the laser beam at maximum power at varying positions. Without the cladding layer we see a notable decrease in the overall laser shift as the beam is positioned farther away from the reference resonator. The normalization of this shift with the initial shift of the reference ring, as well as the normalized shift for the pre-etch data can be seen in Fig. 6.
Fig. 6. Pre- and post-etching resonance shift that occurs relative to the position of the tuning laser.
From this we can conclude that removing the oxide material reduced the ability for heat to dissipate throughout the chip. The post-etch values illustrated in Fig. 6 demonstrate a more representative trend of what we would expect for a temperature distribution curve in a solid material as compared to the pre-etched values which likely result from a much broader temperature distribution, that cannot be easily resolved in the small separation distance between resonators [15]. This further shows how the choice in cladding material affects the heat dissipation across the chip dependent on the thermal coefficients, and subsequently how much it affects the shift in resonance wavelength of neighboring rings. This increases our ability to thermally isolate and tune a single device on a chip without affecting nearby devices. These results are consistent with previously demonstrated techniques of etching thermal trenches to isolate passive components from the heat generated from active components on the same photonic chip [16–18].
4. Conclusion
We have demonstrated the viability of using an external 478nm laser source to thermally tune integrated photonic microresonators. We have demonstrated tuning capabilities of 16.6 pm/mW and up to 44.3 pm/mW with full-chip and targeted laser tuning, respectively. By etching away the oxide layer, we were able to demonstrate an increase in the resonant shift from targeted laser tuning, as well as better thermal isolation between resonators. With these methods, we have successfully demonstrated a straight-forward, no-contact method of characterising resonator performance and thermal response. This approach may be effective for certain quality control and statistical process control purposes within manufacturing environments by offering quick and selective characterization capabilities without integrated electronics or direct-contact thermal stages. The performance and simplicity of these laser tuning methods make them potentially comparable alternatives to coarsely and selectively tune resonators as compared to conventional thermal tuning techniques.
Funding
University of Arizona; Microsoft.
Disclosures
There are no conflicts of interest related to this article.
References
1. Yu Li, Yu Zhang, Lei Zhang, and Andrew W. Poon, “Silicon and hybrid silicon photonic devices for intra-datacenter applications: state of the art and perspectives [Invited],” Photon. Res. 3(5), B10–B27 (2015). [CrossRef]
2. K. Padmaraju and K. Bergman, “Resolving the thermal challenges for silicon microring resonator devices,” Nanophotonics 3(4-5), 269–281 (2014). [CrossRef]
3. M. J. Deen and P. K Basu, Silicon photonics, 1st ed. (Wiley, 2012), Chap. 7.
4. C. L. Chen, Foundations for Guided-wave Optics (Wiley-Interscience, 2007), Chap. 2.
5. V. Raghunathan, W. N. Ye, J. Hu, T. Izuhara, M. Jurgen, and L. Kimerling, “Athermal operation of silicon waveguides: spectral, second order and footprint dependencies,” Opt. Express 18(17), 17631–17639 (2010). [CrossRef]
6. B. E. Little, S. T. Chu, H. A. Haus, J. Foresi, and J.- Laine, “Microring resonator channel dropping filters,” J. Lightwave Technol. 15(6), 998–1005 (1997). [CrossRef]
7. M. Green and M. Keevers, “Optical properties of intrinsic silicon at 300 K,” Progress in Photovoltaics: Research & Applications 3(3), 189–192 (1995). [CrossRef]
8. P. Dobrowolska, A. Krajewska, M. Gajda-Rączka, B. Bartosewicz, P. Nyga, and B. J. Jankiewicz, “Application of Turkevich Method for Gold Nanoparticles Synthesis to Fabrication of SiO2@Au and TiO2@Au Core-Shell Nanostructures,” Materials. 8(6), 2849–2862 (2015). [CrossRef]
9. S. Namnabat, K.-J. Kim, A. Jones, R. Himmelhuber, C. T. DeRose, D. C. Trotter, A. L. Starbuck, A. Pomerene, A. L. Lentine, and R. A. Norwood, “Athermal silicon optical add-drop multiplexers based on thermo-optic coefficient tuning of sol-gel material,” Opt. Express 25(18), 21471–21482 (2017). [CrossRef]
10. A. Arbabi and L. L. Goddard, “Measurements of the refractive indices and thermo-optic coefficients of Si3N4 and SiOx using microring resonances,” Opt. Lett. 38(19), 3878–3881 (2013). [CrossRef]
11. Chemical Rubber Company, CRC Handbook of chemistry and physics. (CRC Press, 1978), Chap 3.
12. E. E. Shipilrain, “AIR (PROPERTIES OF). A-to-Z Guide to Thermodynamics, Heat and Mass Transfer, and Fluids engineering: Thermopedia,” https://doi.org/10.1615/atoz.a.air_properties_of.
13. H. S. Dow, W. S. Kim, and J. W. Lee, “Thermal and electrical properties of silicon nitride substrates,” AIP Adv. 7(9), 095022 (2017). [CrossRef]
14. Robert Hull, INSPEC. Properties of Crystalline Silicon. (INSPEC, the Institution of Electrical Engineers, 1999).
15. Kroemer Kittel and Herbert Kroemer, Thermal physics. 2nd ed. (San W. H. Freeman, 1980)
16. N.-N. Feng, P. Dong, D. Feng, W. Qian, H. Liang, D. C. Lee, J. B. Luff, A. Agarwal, T. Banwell, R. Menendez, P. Toliver, T. K. Woodward, and M. Asghari, “Thermally-efficient reconfigurable narrowband RF-photonic filter,” Opt. Express 18(24), 24648 (2010). [CrossRef]
17. G. Gilardi, W. Yao, H. R. Haghighi, X. J. M. Leijtens, M. K. Smit, and M. J. Wale, “Deep Trenches for Thermal Crosstalk Reduction in InP-Based Photonic Integrated Circuits,” J. Light. Technol. 32(24), 4864–4870 (2014). [CrossRef]
18. D. Pérez, I. Gasulla, L. Crudgington, D. J. Thomson, A. Z. Khokhar, K. Li, W. Cao, G. Z. Mashanovich, and J. Capmany, “Multipurpose silicon photonics signal processor core,” Nat. Commun. 8(1), 636 (2017). [CrossRef]
