Search

The deluge of sensors and data generating devices has driven a paradigm shift in modern computing from arithmetic-logic centric to data-centric processing. Data-centric processing require innovations at the device level to enable novel compute-in-memory (CIM) operations. A key challenge in the construction of CIM architectures is the conflicting trade-off between the performance and their flexibility for various essential data operations. Here, we present a transistor-free CIM architecture that permits storage, search, and neural network operations on sub-50 nm thick Aluminum Scandium Nitride ferroelectric diodes (FeDs). Our circuit designs and devices can be directly integrated on top of Silicon microprocessors in a scalable process. By leveraging the field-programmability, nonvolatility, and nonlinearity of FeDs, search operations are demonstrated with a cell footprint <0.12 μm 2 when projected onto 45 nm node technology. We further demonstrate neural network operations with 4-bit operation using FeDs. Our results highlight FeDs as candidates for efficient and multifunctional CIM platforms.

Dual-gate field-effect biosensors (bioFETs) with asymmetric gate capacitances were shown to surpass the Nernst limit of 59 mV/pH. However, previous studies have conflicting findings on the effect of the capacitive amplification scheme on the sensor detection limit, which is inversely proportional to the signal-to-noise ratio (SNR). Here, we present a systematic experimental investigation of the SNR using ultrathin silicon transistors. Our sensors operate at low voltage and feature asymmetric front and back oxide capacitances with asymmetry factors of 1.4 and 2.3. We demonstrate that in the dual-gate configuration, the response of our bioFETs to the pH change increases proportional to the asymmetry factor and indeed exceeds the Nernst limit. Further, our results reveal that the noise amplitude also increases in proportion to the asymmetry factor. We establish that the commensurate increase of the noise amplitude originates from the intrinsic low-frequency characteristic of the sensor noise, dominated by number fluctuation. These findings suggest that this capacitive signal amplification scheme does not improve the intrinsic detection limit of the dual-gate biosensors.

The motion of Au between AuSi liquid eutectic droplets, both before and during vapor-liquid-solid growth, is important in controlling tapering and diameter uniformity in Si nanowires. We measure the kinetics of coarsening of AuSi droplets on Si(001) and Si(111), quantifying the size evolution of droplets during annealing in ultrahigh vacuum using in situ transmission electron microscopy. For individual droplets, we show that coarsening kinetics are modified when disilane or oxygen is added: coarsening rates increase in the presence of disilane but decrease in oxygen. Matching droplet size measurements on Si(001) with coarsening models confirms that Au transport is driven by capillary forces and that the kinetic coefficients depend on the gas environment present. We suggest that the gas effects are qualitatively similar whether transport is attachment limited or diffusion limited. These results provide insight into manipulating nanowire morphologies for advanced device fabrication.

We use real-time observations of the growth of copper-catalyzed silicon nanowires to determine the nanowire growth mechanism directly and to quantify the growth kinetics of individual wires. Nanowires were grown in a transmission electron microscope using chemical vapor deposition on a copper-coated Si substrate. We show that the initial reaction is the formation of a silicide, eta'-Cu(3)Si, and that this solid silicide remains on the wire tips during growth so that growth is by the vapor-solid-solid mechanism. Individual wire directions and growth rates are related to the details of orientation relation and catalyst shape, leading to a rich morphology compared to vapor-liquid-solid grown nanowires. Furthermore, growth occurs by ledge propagation at the silicide/silicon interface, and the ledge propagation kinetics suggest that the solubility of precursor atoms in the catalyst is small, which is relevant to the fabrication of abrupt heterojunctions in nanowires.