The workforce has developed a world-leading MWP chip able to performing ultrafast analog digital sign processing and computation utilizing optics. Credit score: Metropolis College of Hong Kong
A analysis workforce has unveiled a microwave photonic chip that considerably enhances analog digital sign processing, providing 1,000 instances the velocity and better vitality effectivity than present processors. This innovation guarantees to revolutionize numerous sectors, together with wi-fi communications and synthetic intelligence.
A analysis workforce led by Professor Wang Cheng from the Division of Electrical Engineering (EE) at Metropolis College of Hong Kong (CityUHK) has developed a world-leading microwave photonic chip that’s able to performing ultrafast analog digital sign processing and computation utilizing optics.
The chip, which is 1,000 instances quicker and consumes much less vitality than a standard digital processor, has a variety of purposes, masking 5/6G wi-fi communication programs, high-resolution radar programs, synthetic intelligence, pc imaginative and prescient, and picture/video processing.
The workforce’s analysis findings had been printed within the prestigious scientific journal Nature titled “Built-in Lithium Niobate Microwave Photonic Processing Engine.” It’s a collaborative analysis with The Chinese language College of Hong Kong (CUHK).
Addressing Trendy Communication Challenges
The speedy enlargement of wi-fi networks, the Web of Issues, and cloud-based companies has positioned important calls for on underlying radio frequency programs. Microwave photonics (MWP) know-how, which makes use of optical parts for microwave sign technology, transmission, and manipulation, presents efficient options to those challenges. Nevertheless, built-in MWP programs have struggled to concurrently obtain ultrahigh-speed analog sign processing with chip-scale integration, excessive constancy, and low energy.
“To handle these challenges, our workforce developed a MWP system that mixes ultrafast electro-optic (EO) conversion with low-loss, multifunctional sign processing on a single built-in chip, which has not been achieved earlier than,” defined Professor Wang.
Such efficiency is enabled by an built-in MWP processing engine based mostly on a thin-film lithium niobate (LN) platform able to performing multi-purpose processing and computation duties of analog indicators.
“The chip can carry out high-speed analog computation with ultrabroad processing bandwidths of 67 GHz and wonderful computation accuracies,” stated Feng Hanke, PhD pupil of EE and the primary creator of the paper.
Pioneering Lithium Niobate Photonics
The workforce has been devoted to researching the built-in LN photonic platform for a number of years. In 2018, colleagues at Harvard College and Nokia Bell labs developed the world’s first CMOS (complementary metal-oxide semiconductor)-compatible built-in electro-optic modulators on the LN platform, laying the inspiration for the present analysis breakthrough. LN is known as the “silicon of photonics” for its significance to photonics, corresponding to silicon in microelectronics.
Their work opens up a brand new analysis area, i.e., LN microwave photonics, enabling microwave photonics chips with compact sizes, excessive sign constancy, and low latency; it additionally represents a chip-scale analog digital processing and computing engine.
Reference: “Built-in lithium niobate microwave photonic processing engine” by Hanke Feng, Tong Ge, Xiaoqing Guo, Benshan Wang, Yiwen Zhang, Zhaoxi Chen, Sha Zhu, Ke Zhang, Wenzhao Solar, Chaoran Huang, Yixuan Yuan and Cheng Wang, 28 February 2024, Nature.
DOI: 10.1038/s41586-024-07078-9
The paper’s first authors are Feng Hanke and Ge Tong (EE undergraduate). Professor Wang is the corresponding creator. Different contributing authors embrace Dr. Guo Xiaoqing, PhD graduate of EE; Dr. Chen Zhaoxi, Dr. Zhang Ke, Dr. Zhu Sha (additionally at Beijing College of Know-how), Dr. Solar Wenzhao (now at CityUHK (Dongguan)), EE postdocs; and Zhang Yiwen, EE PhD pupil; and collaborators (Wang Benshan, Professor Huang Chaoran, and Professor Yuan Yixuan) from CUHK.
