In the last couple of decades, terahertz (THz) technologies, which lie in the frequency gap between the infrared and microwaves, have been greatly enhanced and investigated due to possible opportunities in a plethora of THz applications, such as imaging, security, and wireless communications. Photonics has led the way to the generation, modulation, and detection of THz waves such as the photomixing technique. In tandem with these investigations, researchers have been exploring ways to use silicon photonics technologies for THz applications to leverage the cost-effective large-scale fabrication and integration opportunities that it would enable. Although silicon photonics has enabled the implementation of a large number of optical components for practical use, for THz integrated systems, we still face several challenges associated with high-quality hybrid silicon lasers, conversion efficiency, device integration, and fabrication. This paper provides an overview of recent progress in THz technologies based on silicon photonics or hybrid silicon photonics, including THz generation, detection, phase modulation, intensity modulation, and passive components. As silicon-based electronic and photonic circuits are further approaching THz frequencies, one single chip with electronics, photonics, and THz functions seems inevitable, resulting in the ultimate dream of a THz electronic–photonic integrated circuit.

Deep learning has been an important topic in fault diagnosis of motor bearings, which can avoid the need for extensive domain expertise and cumbersome artificial feature extraction. However, existing neural networks have low fault recognition rates and low adaptability under variable load conditions. In order to solve these problems, we propose a one-dimensional fusion neural network (OFNN), which combines Adaptive one-dimensional Convolution Neural Networks with Wide Kernel (ACNN-W) and Dempster-Shafer (D-S) evidence theory. Firstly, the original vibration time-domain signals of a motor bearing acquired by two sensors are resampled. Then, four frameworks of ACNN-W optimized by RMSprop are utilized to learn features adaptively and pre-classify them with Softmax classifiers. Finally, the D-S evidence theory is used to comprehensively determine the class vector output by the Softmax classifiers to achieve fault detection of the bearing. The proposed method adapts to different load conditions by incorporating complementary or conflicting evidences from different sensors through experiments on the Case Western Reserve University (CWRU) motor bearing database. Experimental results show that the proposed method can effectively enhance the cross-domain adaptive ability of the model and has a better diagnostic accuracy than other existing experimental methods.

The hot-carrier photothermoelectric (PTE) effect in two-dimensional materials can be used to develop room-temperature, fast, and sensitive detectors for microwave, terahertz, and far-infrared radiations. Here, we present a van der Waals semimetal PtSe ₂ terahertz detector and find a giant terahertz PTE response mediated by hot carriers at the metal-PtSe ₂ interfaces. The detector demonstrates an excellent zero-bias responsivity of 0.62 ampere per watt (A/W), a low-noise equivalent power of 19.6 picowatt per square root of hertz (pW/Hz ^1/2 ), and a fast response time of 4.5 nanosecond (ns) at 0.1 terahertz (THz). The metal doping effect and asymmetric terahertz heating play vital roles in the hot-carrier PTE response at the metal-semimetal interfaces. Moreover, scanning photocurrent mapping shows that the short-wave zero-bias photoresponse is closely localized to the metal-semimetal interfaces, further revealing the metal doping effect. The bias-dependent photocurrent indicates that the short-wave photoresponse still originates from the PTE effect and the photovoltaic response is negligible. Our results provide important guidance for developing high performance semimetal detectors.

In this letter, a broadband bias-tunable terahertz (THz) quantum well infrared photodetector (QWIP) with a GaAs/AlGaAs asymmetric step multi-quantum well using the intersubband transition was proposed, designed, fabricated, and characterized. It is found that the THz QWIP has two photon response peaks at 4.5 THz and 6.5 THz and the bias-tunable photon response in the 4.0–6.5 THz frequency range of the device is attributed to the field-assisted tunneling of electrons in the second excitation subbands to continuum states.