Characterization of systematic process variation in a silicon photonic platform

Thermal Sensing Using Micro-ring Resonators in Optical Network-on-Chip

“…A common solution for correcting PV-drifts is postfabrication calibration at the device level, which determines the fabric error through variation characterization during chip testing and then calibrate it carefully for every MR. No additional power is required after successful calibration. However, this technique is immature and less commercially practical [11], [12]. Firstly, it is difficult to do post-fabrication calibration for ONoCs where thousands and millions of MRs exist.…”

Compact 1 × N power splitters with arbitrary power ratio for integrated multimode photonics

“…We have therefore tested the system performance by adding a random deviation of ±7.5 nm to all the waveguides widths and their mutual distances. This deviation is compatible with fabrication errors in standard lithography [30,31]. When introducing this random deviation, the array symmetry with respect to the central waveguide is lost and we cannot therefore rely on Eq.…”

Monte-Carlo characterization of Si ring modulator PAM-4 eye diagram performance

“…Si photonics, with its capacity for cost-effective fabrication of large-scale photonic ICs (PIC) can provide high-performance solutions for many important applications such as data-center interconnects, 1) sensors, 2,3) microwave photonics, 4,5) and newly emerging quantum 6,7) and neuromorphic computing. 8,9) For these applications, such process variations as those in silicon-on-insulator (SOI) thickness, waveguide width, and doping concentrations can greatly affect the performance and the yield of the resulting Si PICs, [10][11][12] and, consequently, their statistical distribution and its influence on the PIC performance must be carefully investigated. In the case of Si IC technology, the performance of the target IC can be very precisely predicted in the design stage by various corner simulations 13) and Monte Carlo analysis techniques 14) using the statistical data provided by the IC manufacturer.…”

“…The influence of waveguide geometry variation such as SOI thickness and waveguide linewidth was investigated. [10][11][12] Layout-aware behavioral simulations with the wafer-scale information on spatially correlated layout dependency were carried out. 15,16) However, there is a very limited number of publications 17,18) on the statistical investigation of active Si photonic devices such as modulators, although process variation can have significant influences on the Si active device characteristics and, consequently, the total performance of Si PICs.…”

“…As the application of the Si RM is being extended from very high-speed data center interconnect applications to co-packaged optical I/Os, 21) the statistical characterization of Si RM performance is of great importance. However, previously published reports regarding the statistical studies on the Si RM are mostly focused on the resonator characteristics of RM [10][11][12]15,17) while the modulator characteristics of RMs such as electro-optic efficiency, bandwidth and output Optical Modulation Amplitude (OMA) are important. This is due to the fact that it is hard to correlate the variation of device geometry and fabrication conditions of RMs to their modulation performance variation.…”