Open Access
Add to BibSonomy
Abstract
The field of integrated photonics relies heavily on foundries to produce not only novel technologies, but also reliable ones. Examining the output of complementary metal-oxide-semiconductor (CMOS) foundries such as that affiliated with the AIM Photonics partnership provides valuable insight into the manufacturability of integrated photonic telecommunications devices when produced in large numbers. We present an analysis of the passive performance of numerous silicon microdisk resonators. At ambient temperature, the resonators exhibit on average insertion loss of ∼6 dB, a free spectral range of ∼25 nm, and quality factors of Q > 8.3 × 103. We also report a study of temperature dependence on the resonant wavelength of the devices. Our characterization of these resonators demonstrates reproducibility of qualities related to accuracy in fabrication, as well as in experimental measurement.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The chip-scale silicon photonics industry continues to see advancement as research in telecommunication wavelength regimes explores the feasibility and advantage of integrating photonic technologies with electronics [1,2]. Fabrication of silicon photonic devices relies heavily on the availability of advanced foundry technologies that enable the production of high-quality components in large volumes [3]. The American Institute for Manufacturing Integrated Photonics (AIM Photonics) aims to output products with dependable and reliable performance, while also showing the ability to scale to moderate volumes suitable for specialty applications. In this paper, we report a broad range of experimental characterization conducted on numerous silicon-on-insulator (SOI) microchip resonators [4,5]. The Albany Nanotech Complex produced large batches of chips that consisted of photonic integrated circuits (PICs) we designed using components (Fig. 1) from the standard AIM process development kit (PDK). The devices on our chips demonstrated not only competitive performance, but also intrawafer and wafer-to-wafer consistency and reproducibility.
Ample research in recent years demonstrates the development of high performance silicon-based photonic microring resonators (MRR) [6–9] and microdisk resonators (MDR) [10–12]. However, few studies [13,14] investigated the consistency and reproducibility of the functionality of such devices when produced in large batch quantities by a process akin to that implemented by AIM Photonics. Therefore, we experimentally demonstrate herein a narrow distribution in passive performance of high-speed resonant amplitude microdisk resonators (Fig. 2) containing four adjacent channels that provide distinct operating wavelengths in the optical communications region surrounding 1550 nm.
Fig. 1. GDS file depiction of resonator layout – ${\lambda _0}$, ${\lambda _1}$, ${\lambda _2}$ and ${\lambda _3}$ show the separate channels.
Single wavelength operation [15,16], high bandwidth, and low power consumption account for some of the key user benefits provided by these low loss and high Q resonators [17]. Characteristics of particular interest to us in examination of overall reliability for data communications applications include optical response properties such as free spectral range (FSR) [18], quality factor (Q factor) [19], and resonant wavelengths. We conducted an additional study of the temperature dependence of the devices found on these chips (AIM 18-01 MPW fabrication run), since the temperature dependence of silicon photonics is well-known and a challenge for some applications such as data center photonics. This work presents a comprehensive analysis of the manufacturability of silicon integrated photonic technology produced in a foundry setting.
2. Measurement methods
2.1 Loss measurement
After optimizing transmission through the input fiber for transverse electric (TE) light with a Thorlabs polarization controller – and thereby minimizing the impact of polarization on measurement procedure – we aligned a lens to the output edge coupler of the device-under-test (DUT). A CCD camera and power meter facilitated the alignment process and achievement of this loss characterization. Input laser power remained set to 0.96 mW for the entire period spent observing loss through resonators on the AIM chips. Fiber-to-lens and fiber-to-fiber measurements of insertion loss were conducted on five different rows of 4-channel devices on over 20 distinct chips.
Fig. 2. Micrograph of resonator layout. These four-channel microdisk resonator devices from the AIM PDK contain internal heaters which enable thermal tuning up to around 12 nm. The disk radius is 8 µm, and the bus waveguide width measures 0.4 µm.
These tests were motivated by a desire to statistically capture the reproducibility of devices produced at the AIM Photonics foundry. We simultaneously performed a study of the repeatability of our own process to obtain measurements manually within the laboratory. This additional work involved a technique that required removing and replacing the same chip on the sample stage, then repeating the loss measurement of a specific DUT five times (see “Repeat Chip” in Table 1).
Table 1. Resonator Insertion Loss Statistics
2.2 Optical response characterizations
To investigate the operating capability of these MDRs manufactured in large batch quantities, we measured the quality factor (Q factor), discrete resonant wavelengths, and free spectral range associated with the 4-channel devices at room temperature. With the sample chip fiber-coupled to an optical erbium-doped fiber amplifier (EDFA) source from PriTel Inc. and a Yokogawa AQ6370B optical spectrum analyzer (OSA), the optical response of these devices revealed to us the precise locations of resonant wavelengths for each channel and enabled us to determine Q, defined as
where ${\lambda _r}$ is the resonant wavelength and ${\lambda _{FWHM}}$ is the full width at half maximum of the resonance spectrum.The Yokogawa OSA resolution bandwidth of 0.02 nm was sufficient to determine Q, since on average the FWHM measures 0.18 nm. An examination of the temperature dependence of several MDRs under passive conditions completed our analysis of their operation.
3. Results and discussion
3.1 Insertion loss
Loss is a critical factor to consider when determining the performance of integrated photonic technologies. As shown in Table 1, our observations reveal consistent and low off-resonance insertion losses for numerous (>100) identical MDR elements from 10 different test chips. Measurements conducted with optical fiber at both the input and output facets of the chips are generally expected to determine single-mode insertion loss more accurately than do fiber-to-lens measurements. While our results did follow this anticipated trend, we observed remarkably little increase in overall loss through the resonators after switching to the fiber-to-fiber orientation, indicating that the chip endfaces and couplers had achieved very low loss coupling to single-mode optical fiber. The “Repeat Chip” results refer to the procedure involving removing and replacing chips multiple times to quantify confidence in our experimental setup and manual alignment process.
3.2 Resonant wavelengths and free spectral range (FSR)
Preparing for efforts to pigtail and otherwise package several 18-01 chips to conduct further studies of device performance required obtaining knowledge of the precise operational wavelengths of the resonators. This information fell naturally out of the process of analyzing the optical response of these devices. The AIM PDK assigns a nominal value of 25.6 nm to the free spectral range of the four-channel resonators. Our experiments indicated an average FSR of 25.2 ± 0.4 nm, taking five separate devices on 20 distinct chips into consideration. Through careful analysis within our measurement system, we observed the resonant wavelengths for each of the four channels to be 1552.5 nm, 1558.7 nm, 1564.7 nm, and 1570.9 nm, respectively. Among dozens of chips, the locations of these resonances agreed to within 0.43 Standard Deviations (SD). We assume that the deviation from the nominal operating wavelengths is due to slight differences in silicon thickness between the “standard” and the chips that we received. As we worked towards accurately measuring the Q factors of the resonators, a handful of other operating wavelengths in the 1520–1550 nm region became associated with our device characterization process (Table 2).
Table 2. Resonator Quality Factor Statistics
3.3 Quality factor
Determination of the quality factor assists tremendously in indicating the potential performance of a resonator, in this case one with optical communications applications around 1550 nm. The results contained in Tables 2 and 3 aim to provide, in a statistical sense, an illustration of the robust and reliable silicon integrated photonics technology currently available from the AIM Photonics foundry. The specific wavelength associated with each experimentally determined resonance appears in the first column of Table 2. Careful analysis of the optical response of the 4-channel MDRs reveals not only high Q factor values, but also consistent performance throughout the batch of chips. The second and third columns in this table represent once again a comparison of multiple identical DUTs per chip, with 12 distinct chips undergoing testing.
Table 3. Manufacturing Consistency Study
We ultimately measured 227 resonators fabricated in two separate process runs at the fabrication facility, allowing for an additional investigation of manufacturing consistency over time. Our results in Table 3 contain data from 74 elements produced during the first run and 157 from the second, as well as a combined average.
3.4 Temperature dependence
As the silicon photonic resonators included on the 18-01 AIM test chips are intended for deployment in a number of different operating environments, our efforts to capture the behavior of the devices included some study of temperature dependence and its effect on certain resonator characteristics. In particular, a linear redshift in resonant wavelength with respect to increasing temperature was anticipated across all channels of the microdisk resonators. For these measurements, we added a temperature stage utilizing an external thermoelectric cooler into the setup used for previous testing. This study involved accurately controlling the heating of the entire chip with the thermal stage, and therefore did not employ the use of the internal heaters. While obtaining data, we re-optimized the fiber alignment to the edge coupler on the chip whenever necessary. Silicon photonics technologies are sensitive to variation in temperature due to the high thermo-optic coefficient of silicon [20]. Therefore, we allowed time for the chip and fibers to rest and for the OSA signal to settle before each subsequent measurement of the resonance position. Figure 3 illustrates the change in the optical response for a 4-channel device, corresponding to a gradual variation in applied temperature over a range of 15 degrees Celsius (°C).
Fig. 3. OSA signal depicting the dependence on temperature of a 4-channel microdisk resonator. The 18-01 chip was subjected to a gradual variation in temperature of 15°C.
We next considered the dependence of resonant wavelength on temperature for four separate channels within the same device (Fig. 4). In keeping with the motivation to examine the reproducibility of performance seen in the resonators from our large batch of 18-01 chips, similar temperature dependence evaluation was repeated on devices located on a few separate chips. Once again, the technology produced by the AIM foundry proved to be reliable and repeatable overall.
4. Conclusion
We have demonstrated the consistent passive performance of microdisk resonators produced by the AIM Photonics foundry. Examination of the active operation of the devices and their performance as a function of temperature will show further promise for integrated photonics communications applications such as data center photonics and supercomputing.
Funding
U.S. Department of Defense (FA8650-15-2-5220).
Acknowledgment
The authors kindly acknowledge discussions with Dr. A. L. Lentine of Sandia National Laboratories with respect to the design and layout of the photonic integrated circuits.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. B. Jalali and S. Fathpour, “Silicon photonics,” J. Lightwave Technol. 24(12), 4600–4615 (2006). [CrossRef]
2. Q. Xu, B. Schmidt, S. Pradhan, et al., “Micrometre-scale silicon electro-optic modulator,” Nature 435(7040), 325–327 (2005). [CrossRef]
3. T. Barwicz, H. Byun, F. Gan, et al., “Silicon photonics for compact, energy-efficient interconnects [Invited],” J. Opt. Netw. 6(1), 63–73 (2007). [CrossRef]
4. M. Gehl, C. Long, D. Trotter, et al., “Operation of high-speed silicon photonic micro-disk modulators at cryogenic temperatures,” Optica 4(3), 374–382 (2017). [CrossRef]
5. E. M. Fard, C. M. Long, A. L. Lentine, et al., “Cryogenic C-band wavelength division multiplexing system using an AIM Photonics Foundry process design kit,” Opt. Express 28(24), 35651–35662 (2020). [CrossRef]
6. A. Biberman, M. J. Shaw, E. Timurdogan, et al., “Ultralow-loss silicon ring resonators,” Opt. Lett. 37(20), 4236–4238 (2012). [CrossRef]
7. K. Padmaraju, J. Chan, L. Chen, et al., “Thermal stabilization of a microring modulator using feedback control,” Opt. Express 20(27), 27999–28008 (2012). [CrossRef]
8. D. Zeng, Q. Liu, C. Mei, et al., “Demonstration of ultra-high-Q silicon microring resonators for nonlinear integrated photonics,” Micromachines 13(7), 1155 (2022). [CrossRef]
9. Z. Ying, Z. Wang, Z. Zhao, et al., “Comparison of microrings and microdisks for high-speed optical modulation in silicon photonics,” Appl. Phys. Lett. 112(11), 111108 (2018). [CrossRef]
10. W. A. Zortman, M. R. Watts, D. C. Trotter, et al., “Low-power high-speed silicon microdisk modulators,” in CLEO, San Jose, CA, USA, May 16-21, 2010.
11. D. Gostimirovic and W. N. Ye, “Ultralow-power double vertical junction microdisk modulators,” IEEE J. Sel. Top. Quantum Electronics 27(3), 1–7 (2021). [CrossRef]
12. Y. Wang, N. Zhang, Z. Jiang, et al., “Chip-scale mass manufacturable high-Q silicon microdisks,” Adv. Mater. Technol. 2(6), 1600299 (2017). [CrossRef]
13. P. Sun, J. Hulme, T. Van Vaerenbergh, et al., “Statistical behavioral models of silicon ring resonators at a commercial CMOS foundry,” IEEE J. Sel. Top. Quantum Electronics 26(2), 1–10 (2020). [CrossRef]
14. A. V. Krishnamoorthy, X. Zheng, G. Li, et al., “Exploiting CMOS manufacturing to reduce tuning requirements for resonant optical devices,” IEEE Photonics J. 3(3), 567–579 (2011). [CrossRef]
15. E. Timurdogan, C. M. Sorace-Agaskar, E. S. Hosseini, et al., “Vertical junction silicon microdisk modulator with integrated thermal tuner,” in CLEO, San Jose, CA, USA, Jun. 9-14, 2013.
16. H. Shoman, H. Jayatilleka, A. H. K. Park, et al., “Compact wavelength- and bandwidth-tunable microring modulator,” Opt. Express 27(19), 26661–26675 (2019). [CrossRef]
17. W. Zhang and J. Yao, “Silicon-based single-mode on-chip ultracompact microdisk resonators with standard silicon photonics foundry process,” J. Lightwave Technol. 35(20), 4418–4424 (2017). [CrossRef]
18. W. Zhang and J. Yao, “Electrically tunable silicon-based on-chip microdisk resonator for integrated microwave photonic applications [Invited],” APL Photonics 1(8), 080801 (2016). [CrossRef]
19. M. Soltani, S. Yegnanarayanan, and A. Adibi, “Ultra-high Q planar silicon microdisk resonators for chip-scale silicon photonics,” Opt. Express 15(8), 4694–4704 (2007). [CrossRef]
20. S. Namnabat, K.-J. Kim, A. Jones, et al., “Athermal silicon optical add-drop multiplexers based on thermo-optic coefficient tuning of sol-gel material,” Opt. Express 25(18), 21471–21482 (2017). [CrossRef]
