Chip最新文献

In recent years, Internet of Things (IoT) has become more and more important owing to the rapid expansion of the number of computing devices and data sizes. The evolution of IoT requires low-power and self-operating devices to expand the coverage area of computing resources. The main components of IoT are the large-scale integration (LSI) chips, which take the function of implementing the energy harvesters, control units and applications. They exhibit different physics or phenomena, making it difficult to understand and design the entire system. The current work reviews the various methods for IoT applications by CMOS LSI chips, from the power components by energy harvesting to realistic applications with future outlooks.

A non-classical light source is essential for implementing a wide range of quantum information processing protocols, including quantum computing, networking, communication and metrology. In the microwave regime, propagating photonic qubits, which transfer quantum information between multiple superconducting quantum chips, serve as building blocks for large-scale quantum computers. In this context, spectral control of propagating single photons is crucial for interfacing different quantum nodes with varied frequencies and bandwidths. Here a deterministic microwave quantum light source was demonstrated based on superconducting quantum circuits that can generate propagating single photons, time-bin encoded photonic qubits and qudits. In particular, the frequency of the emitted photons can be tuned in situ as large as 200 MHz. Even though the internal quantum efficiency of the light source is sensitive to the working frequency, it is shown that the fidelity of the propagating photonic qubit can be well preserved with the time-bin encoding scheme. This work thus demonstrates a versatile approach to realizing a practical quantum light source for future distributed quantum computing.

Investigation into the structural and magnetic properties of organic molecules at cryogenic temperature is beneficial for reducing molecular vibration and stabilizing magnetization, and is of great importance for constructing novel spintronics devices of better performance and scaling the device size down to nanoscale. In order to explore the possibility of fabricating molecule-based memory chips of ultrahigh density, two-dimensional close-packed molecular arrays with carboxylic acid molecules were constructed in the current work and the magnetic properties in a low-temperature scanning tunneling microscope were also investigated. The results demonstrated that each nonmagnetic molecule can be controllably and independently switched into a stable spin-carrying state at 4 K by applying a voltage pulse with atomic resolution. Benefiting from the small size of a single molecule as the basic storage bit, the two-dimensional molecular arrays allowing controllable electrical manipulations on each molecule can behave as a platform of memory chip with an ultrahigh storage density of ∼320 terabytes (Tb) (or ∼2500 terabits) per square inch. This work highlights the potential and advantage of employing organic molecules in developing future cryogenic information storage techniques and devices at nanoscale.

In the current work, we explored a new knowledge amalgamation problem, termed Federated Selective Aggregation for on-device knowledge amalgamation (FedSA). FedSA aims to train an on-device student model for a new task with the help of several decentralized teachers whose pre-training tasks and data are different and agnostic. The motivation to investigate such a problem setup stems from a recent dilemma of model sharing. Due to privacy, security or intellectual property issues, the pre-trained models are, however, not able to be shared, and the resources of devices are usually limited. The proposed FedSA offers a solution to this dilemma and makes it one step further, again, the method can be employed on low-power and resource-limited devices. To this end, a dedicated strategy was proposed to handle the knowledge amalgamation. Specifically, the student-training process in the current work was driven by a novel saliency-based approach which adaptively selects teachers as the participants and integrated their representative capabilities into the student. To evaluate the effectiveness of FedSA, experiments on both single-task and multi-task settings were conducted. The experimental results demonstrate that FedSA could effectively amalgamate knowledge from decentralized models and achieve competitive performance to centralized baselines.

The design and preparation of quantum states free from environmental decohering effects is critically important for the development of on-chip quantum systems with robustness. One promising strategy is to harness quantum state superposition to construct decoherence-free subspace (DFS), which is termed dark state. Typically, the excitation of dark states relies on anti-phase-matching on two qubits and the inter-qubit distance is of wavelength scale, which limits the development of compact quantum chips. In the current work, a hybrid plasmonic quantum emitter was proposed, which was composed of strongly correlated quantum emitters intermediated by a plasmonic nanocavity. Through turning the plasmonic loss from drawback into advantage, the anti-phase-matching rule was broken by rapidly decaying the superposed bright state and preparing a sub-100 nm dark state with decay rate reduced by 3 orders of magnitudes. More interestingly, the dark state could be optically switched to a single-photon emitter with enhanced brightness through photon-blockade, with the quantum second order correlation function at zero delay showing a wide range of tunability down to 0.02.

The rise of quantum computing has prompted the interest in the field of cryogenic electronics. Carbon-based materials hold great promise in the area of cryogenic electronics due to their excellent material properties and emergent quantum effects. This paper introduces the advantages of carbon-based materials for cryogenic applications and reviews recent progress in carbon nanotubes and graphene for logic devices, sensors and novel quantum devices at cryogenic temperatures. Finally, the main challenges and extensive prospects for the further development of carbon-based cryoelectronics are summarized.

A matrix-vector multiplication (MVM) optical signal processor based on mode division multiplexing (MDM) was proposed and demonstrated in the current work, which is composed of a mode multiplexer, a multimode beam splitter, a mode demultiplexer, a modulator array and combiners. In addition, the characteristics of MDM obviate the need for multiple wavelengths and therefore multiple laser light sources are unneeded, which greatly reduces the complexity and cost. A 4 × 4 MDM-MVM was realized on a standard silicon-on-insulator (SOI) platform. Combined with the off-chip light source and photodetectors (PDs), 4-level modulation has been demonstrated, and each level of the output signal could represent 2 bits of information.

With the advent of the “Big Data Era”, improving data storage density and computation speed has become more and more urgent due to the rapid growth in different types of data. Flash memory with a floating gate (FG) structure is attracting great attention owing to its advantages of miniaturization, low power consumption and reliable data storage, which is very effective in solving the problems of large data capacity and high integration density. Meanwhile, the FG memory with charge storage principle can simulate synaptic plasticity perfectly, breaking the traditional von Neumann computing architecture and can be used as an artificial synapse for neuromorphic computations inspired by the human brain. Among many candidate materials for manufacturing devices, van der Waals (vdW) materials have attracted widespread attention due to their atomic thickness, high mobility, and sustainable miniaturization properties. Owing to the arbitrary stacking ability, vdW heterostructure combines rich physics and potential 3D integration, opening up various possibilities for new functional integrated devices with low power consumption and flexible applications. This paper provides a comprehensive review of memory devices based on vdW materials with FG structure, including the working principles and typical structures of FG structure devices, with a focus on the introduction of various high-performance FG memories and their versatile applications in neuromorphic computing. Finally, the challenges of neuromorphic devices based on FG structures are also discussed. This review will shed light on the design and fabrication of vdW material-based memory devices with FG engineering, helping to promote the development of practical and promising neuromorphic computing.

Spectral imaging extends the concept of traditional color cameras to capture images across multiple spectral channels and has broad application prospects. Conventional spectral cameras based on scanning methods suffer from the drawbacks of low acquisition speed and large volume. On-chip computational spectral imaging based on metasurface filters provides a promising scheme for portable applications, but endures long computation time due to point-by-point iterative spectral reconstruction and mosaic effect in the reconstructed spectral images. In this study, on-chip rapid spectral imaging was demonstrated, which eliminated the mosaic effect in the spectral image by deep-learning-based spectral data cube reconstruction. The experimental results show that 4 orders of magnitude faster than the iterative spectral reconstruction were achieved, and the fidelity of the spectral reconstruction for the standard color plate was over 99% for a standard color board. In particular, video-rate spectral imaging was demonstrated for moving objects and outdoor driving scenes with good performance for recognizing metamerism, where the concolorous sky and white cars can be distinguished via their spectra, showing great potential for autonomous driving and other practical applications in the field of intelligent perception.

As an indispensable part to compensate for the signal crosstalk in fiber communication systems, conventional digital multi-input multi-output (MIMO) signal processor is facing the challenges of high computational complexity, high power consumption and relatively low processing speed. The optical MIMOenables the best use of light and has been proposed to remedy this limitation. However, the currently existing optical MIMO methods are all restricted to the spatial dimension, while the temporal dimension is neglected. Here, an on-chip spatial-temporal descrambler with four channels were devised and its MIMO functions were experimentally verified simultaneously in both spatial and temporal dimensions. The spatial crosstalk of single-channel descrambler and four-channel descrambler is respectively less than -21 dB and -18 dB, and the time delay is simultaneously compensated successfully. Moreover, a more universal model extended to mode-dependent loss and gain (MDL) compensation was further developed, which is capable of being cascaded for the real optical transmission system. The first attempt at photonic spatial-temporal descrambler enriched the varieties of optical MIMO, and the proposed scheme provided a new opportunity for all-optical MIMO signal processing.