
Self-Driving Photonic Chips: Autotune for PICs
Failed to add items
Add to basket failed.
Add to Wish List failed.
Remove from Wish List failed.
Follow podcast failed
Unfollow podcast failed
-
Narrated by:
-
By:
About this listen
A programmable photonic chip that measures its own phase and auto-calibrates in ~25 iterations—using an on-chip reference path + Kramers–Kronig to recover the full complex response over hundreds of GHz.This episode unpacks “Self-calibrating programmable photonic integrated circuits” by Xingyuan Xu, Guanghui Ren, Tim Feleppa, Xumeng Liu, Andreas Boes, Arnan Mitchell & Arthur J. Lowery (Nature Photonics, 2022). The problem: big PICs drift and vary—fabrication errors and thermal cross-talk break the link between control voltages and optical function. The fix: add a short-delay on-chip reference path, sweep a laser, and use the Kramers–Kronig relationship to recover phase from intensity, then inverse-FFT to get the complex impulse response and update each MZI/phase shifter—no prior device model needed.
They demo it on a 16-tap FIR with an 8-tap signal-processing core, FSR ≈ 160 GHz (6.25 ps delay steps), dialing diverse functions—complex sinc filters, Hilbert transformer, half-band low/high-pass, differentiator—and converging in ~25 training iterations. Platform: SiN (TriPleX) PIC with ~0.15 dB/cm loss; the reference adds <3 mm² footprint. Result: dial-a-transfer-function across the C-band with fast, stable calibration—a path to real-time reconfigurable comms, neuromorphic optics, and quantum.
Source: Xu et al., “Self-calibrating programmable photonic integrated circuits,” Nature Photonics (2022).