Abstract: Training machine learning systems using a training data set, gradient descent, and a loss function. The machine learning system includes memory and reads and writes to memory according to read and write profiles. The loss function is associated with machine learning system memory read and write profile gradients. The loss function includes a loss function penalty term, the loss function penalty term being associated with the read and write profile gradient differences. Trained machine learning systems are then provided.

Abstract: A semiconductor radiation monitor is provided that includes a charge storage region composed of a dielectric material nanosheet, such as, for example an epitaxial oxide nanosheet, which is sandwiched between a top semiconductor nanosheet and a bottom semiconductor nanosheet. A functional gate structure is located above the top semiconductor nanosheet and beneath the bottom semiconductor nanosheet.

Abstract: An apparatus (e.g., an imaging system) includes a circuit, including: a p-i-n diode having a cathode coupled to a cathode bias voltage or ground; a charge transistor having a first source/drain terminal coupled to an anode of the diode; a storage capacitor having a first terminal coupled to a second source/drain terminal of the charge transistor and a second terminal coupled to the cathode; an amplification transistor having a gate terminal coupled to the first terminal of the storage capacitor and a first source/drain terminal coupled to a reference voltage; a read transistor having a first source/drain terminal coupled to a second source/drain terminal of the amplification transistor; a data line having a first terminal coupled to a second source/drain terminal of the read transistor; and a readout circuit coupled to a second terminal of the data line, providing an output voltage corresponding to charge on the storage capacitor.

Abstract: A two-transistor-two-resistor (2T2R) resistive random access memory (ReRAM) structure, and a method for forming the same includes two vertical field effect transistors (VFETs) formed on a substrate, each VFET includes an epitaxial region located above a channel region and below a dielectric cap. The epitaxial region includes two opposing protruding regions of triangular shape that extend horizontally beyond the channel region. A metal gate material is disposed on and around the channel region. A portion of the metal gate material is located between the two VFETs. A ReRAM stack is deposited within two openings adjacent to a side of each VFET that is opposing the portion of the metal gate material located between the two VFETs. A portion of the epitaxial region in direct contact with the ReRAM stack acts as a bottom electrode for the ReRAM structure.

Abstract: A one-transistor-two-resistor (1T2R) resistive random access memory (ReRAM) structure, and a method for forming the same, includes forming a vertical field effect transistor (VFET) including an epitaxial region located above a channel region and below a dielectric cap. The epitaxial region includes two opposing protruding regions of triangular shape bounded by <111> planes that extend horizontally beyond the channel region. A ReRAM stack is conformally deposited on the VFET. The ReRAM stack includes an oxide layer located directly above the epitaxial region, a top electrode layer directly above the oxide layer and a metal fill above the top electrode layer. Each of the two opposing protruding regions of the epitaxial region acts as a bottom electrode for the ReRAM stack.

Abstract: A semiconductor structure may include a vertical field effect transistor, the vertical field effect transistor may include a top source drain, a bottom source drain, and an epitaxial channel and a resistive random access memory below the vertical field effect transistor. The resistive random access memory may include an epitaxial oxide layer, a top electrode, and a bottom electrode. The top electrode, which may function as the bottom source drain of the vertical field effect transistor, may be in direct contact with the epitaxial channel of the vertical field effect transistor. The epitaxial oxide layer may separate the top electrode from the bottom electrode. The top source drain may be arranged between a dielectric material and the epitaxial channel. The dielectric material may be in direct contact with a top surface of the epitaxial channel. The epitaxial oxide layer may be composed of a rare earth oxide.

Abstract: A first vertical T-FET has a source heavily doped with a source concentration of a source-type dopant, a drain doped with a drain concentration of a drain-type dopant, and a channel between the source and drain. The source, channel, and drain are stacked vertically in a fin or pillar perpendicular to a substrate. A gate stack encompasses the channel sides and has a drain overlap amount overlapping the drain sides and a source overlap amount overlapping the source sides. External contacts electrically connect the gate and source and/or drain. The source-type dopant and the drain-type dopant are opposite dopant types. In some embodiments, a second vertical T-FET is stacked on the first vertical T-FET. Different VT-FET devices are made by changing the materials, doping types and levels, and connections to the sources, channels, and drains. Device characteristics are designed/changed by changing the amount of source and drain overlaps of the gate stack(s).

Abstract: A memory network can be constructed with at least memory write weightings, memory read weightings and at least one read vector, the memory write weightings parameterizing memory write operations of a neural network to the memory matrix, the memory read weightings parameterizing memory read operations of the neural network from the memory matrix. At least one of the write weightings, the read weightings, or elements of the at least one read vector, can be initialized to have sparsity and/or low discrepancy sampling pattern. The memory network can be trained to perform a task.

Abstract: A semiconductor structure may include two vertical transport field effect transistors comprising a top source drain, a bottom source drain, and an epitaxial channel and a resistive random access memory between the two vertical transport field effect transistors, the resistive random access memory may include an oxide layer, a top electrode, and a bottom electrode, wherein the oxide layer may contact the top source drain of the two vertical field effect transistor. The top source drain may function as the bottom electrode of the resistive random access memory. The semiconductor structure may include a shallow trench isolation between the two vertical transport field effect transistors, the shallow trench isolation may be embedded in a first spacer, a doped source, and a portion of a substrate.

Abstract: An erasable magnetoresistive random-access memory (MRAM) structure and a method of making the same includes an MRAM cell disposed between bit line and word line circuit elements, and a vertical-cavity surface-emitting laser (VCSEL) element disposed above the MRAM cell. A laser output of the VCSEL is directed toward the MRAM cell.

Abstract: A method for manufacturing a semiconductor memory device includes forming a first doped semiconductor layer on a conductive layer, forming a second doped semiconductor layer stacked on the first doped semiconductor layer, forming a third doped semiconductor layer stacked on the second doped semiconductor layer, and forming a memory stack layer on the third doped semiconductor layer. The memory stack layer and the first, second and third doped semiconductor layers are patterned into a plurality of pillars spaced apart from each other. In the method, a plurality of extrinsic base layers are formed adjacent the patterned second doped semiconductor layers. The patterned first, second and third doped semiconductor layers in each pillar of the plurality of pillars are components of a bipolar junction transistor device, and the plurality of pillars are parts of a memory cell array having a cross-point structure.

Abstract: A method for manufacturing a semiconductor device includes forming a first vertical transistor on a semiconductor substrate, and forming a second vertical transistor stacked on the first vertical transistor. In the method, a silicide layer is formed on a first drain region of the first vertical transistor and on a second drain region of the second vertical transistor. The silicide layer electrically connects the first and second drain regions to each other.

Abstract: Photovoltaic devices such as solar cells having one or more field-effect hole or electron inversion/accumulation layers as contact regions are configured such that the electric field required for charge inversion and/or accumulation is provided by the output voltage of the photovoltaic device or that of an integrated solar cell unit. In some embodiments, a power source may be connected between a gate electrode and a contact region on the opposite side of photovoltaic device. In other embodiments, the photovoltaic device or integrated unit is self-powering.

Abstract: A stacked FinFET mask-programmable read only memory (ROM) is provided. The stacked FinFET mask-programmable ROM includes a fin structure extending upward from an insulator layer. In accordance with the present application, the fin structure includes, from bottom to top, a lower programmable semiconductor fin portion having a first threshold voltage, an insulator fin portion, and an upper programmable semiconductor fin portion having a second threshold voltage. A lower gate structure contacts a sidewall of the lower programmable semiconductor fin portion, and an upper gate structure contacts a sidewall of the upper programmable semiconductor fin portion.

Abstract: Semiconductor structures and methods of making the same. The semiconductor structures including at least two vertically stacked nanosheet devices. In one embodiment, an integrated circuit includes a plurality of horizontal nanosheet devices (HNS devices) that are stacked vertically, on top of each other, relative to a top surface of a substrate. The plurality of HNS devices including a first HNS device and a second HNS device that each have source and drain structures.

Abstract: A method for manufacturing a semiconductor device includes forming a fin on a semiconductor substrate, and forming a bottom source/drain region adjacent a base of the fin. In the method, a dielectric layer, a work function metal layer and a first gate metal layer are sequentially deposited on the bottom source/drain region and around the fin. The dielectric layer, the work function metal layer and the first gate metal layer form a gate structure. The method also includes removing the dielectric layer, the work function metal layer and the first gate metal layer from an end portion of the fin, and depositing a second gate metal layer around the end portion of the fin in place of the removed dielectric layer, the removed work function metal layer and the removed first gate metal layer. The second gate metal layer contacts the end portion of the fin.

Abstract: A method of forming an electrical device that includes forming a multilayered fin composed of a first source/drain layer for a first transistor, a first channel layer for the first transistor, a common source/drain layer for the first transistor and a second transistor, a second channel layer for the second transistor and a second source/drain layer for the second transistor. A common spacer is formed on the common source/drain layer that separates a first opening to the first channel layer from a second opening to the second channel layer. Gate structures are then formed in the first and second openings.

Abstract: A method of forming an electrical device that includes forming an amorphous semiconductor material on a metal surface of a memory device, in which the memory device is vertically stacked atop a first transistor. The amorphous semiconductor material is annealed with a laser anneal having a nanosecond duration to convert the amorphous semiconductor material into a crystalline semiconductor material. A second transistor is formed from the semiconductor material. The second transistor vertically stacked on the memory device.

Abstract: A method of fabricating an access device in a crosspoint memory array structure during BEOL processing includes: forming at least a first doped semiconductor layer on an upper surface of a first conductive layer, the first doped semiconductor layer being in electrical connection with the first conductive layer; exposing at least a portion of the first doped semiconductor layer to a directed energy source to cause localized annealing in the first doped semiconductor layer to activate a dopant of a first conductivity type in the first doped semiconductor layer, thereby converting at least a portion of the first doped semiconductor layer into a polycrystalline layer; forming a second conductive layer over a least a portion of the first doped semiconductor layer; and etching the first doped semiconductor layer and the first and second conductive layers to form an access device that is self-aligned with the first and second conductive layers.

Abstract: A method of forming an electrical device that includes forming a multilayered fin composed of a first source/drain layer for a first transistor, a first channel layer for the first transistor, a common source/drain layer for the first transistor and a second transistor, a second channel layer for the second transistor and a second source/drain layer for the second transistor. A common spacer is formed on the common source/drain layer that separates a first opening to the first channel layer from a second opening to the second channel layer. Gate structures are then formed in the first and second openings.