TY - JOUR
T1 - Ultracompact multimode parallel all-optical logic gates by inverse design
AU - Shi, Kangle
AU - He, Lu
AU - Zhang, Xiangdong
N1 - Publisher Copyright:
© 2026 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.
PY - 2026/8/10
Y1 - 2026/8/10
N2 - Integrated all-optical logic processing is a promising pathway toward ultrafast and high-throughput photonic computing. To enable large-scale on-chip optical systems, logic gates must be extremely compact while supporting high information capacity. Here, we report the design of an ultracompact multimode parallel all-optical logic gate platform on a silicon photonics chip using inverse design. Specifically, we demonstrate a 10 × 8 µm2 multimode 2 × 2 50:50 power splitter that simultaneously supports five transverse electric (TE) modes and enables fundamental logic operations—including XOR, OR, and NOT—with a single device. By cascading two identical splitters, more complex functions such as AND, NAND, NOR, XNOR, and even a half-adder are realized, all within a footprint order of magnitude smaller than conventional approaches. Our architecture achieves mode-division multiplexed logic processing, significantly enhancing data throughput without sacrificing integration density. This work provides a scalable and highly integrated solution for future optical computing, on-chip interconnects, and advanced photonic signal processing systems.
AB - Integrated all-optical logic processing is a promising pathway toward ultrafast and high-throughput photonic computing. To enable large-scale on-chip optical systems, logic gates must be extremely compact while supporting high information capacity. Here, we report the design of an ultracompact multimode parallel all-optical logic gate platform on a silicon photonics chip using inverse design. Specifically, we demonstrate a 10 × 8 µm2 multimode 2 × 2 50:50 power splitter that simultaneously supports five transverse electric (TE) modes and enables fundamental logic operations—including XOR, OR, and NOT—with a single device. By cascading two identical splitters, more complex functions such as AND, NAND, NOR, XNOR, and even a half-adder are realized, all within a footprint order of magnitude smaller than conventional approaches. Our architecture achieves mode-division multiplexed logic processing, significantly enhancing data throughput without sacrificing integration density. This work provides a scalable and highly integrated solution for future optical computing, on-chip interconnects, and advanced photonic signal processing systems.
UR - https://www.scopus.com/pages/publications/105047782803
U2 - 10.1364/OE.596685
DO - 10.1364/OE.596685
M3 - Article
AN - SCOPUS:105047782803
SN - 1094-4087
VL - 34
SP - 28858
EP - 28873
JO - Optics Express
JF - Optics Express
IS - 16
ER -