Abstract: A logical NAND memory architecture comprising two-terminal, non-volatile resistive memory is disclosed. By way of example, disclosed logical NAND architectures can comprise non-volatile memory cells having approximately 4F2 area. This facilitates very high memory densities, even for advanced technology nodes. Further, the disclosed architectures are CMOS compatible, and can be constructed among back-end-of-line (BEOL) metal layers of an integrated chip. In some embodiments, subsets of two-terminal memory cells in a NAND array can be constructed between different pairs of BEOL metal layers. In other embodiments, the two-terminal memory cells can be constructed between a single pair of BEOL metal layers.

Abstract: One potential result of differing characteristics for certain two-terminal memory (TTM) is that memory management techniques, such as logical-to-physical (L2P), can differ as well. Previous memory management techniques do not adequately leverage the advantages associated with TTM. For example, by identifying and leveraging certain advantageous characteristics of TTM, L2P tables can be designed to be smaller and more efficient, which can allow the L2P table to be stored in memory that is faster and/or closer (or integrated into) an associated controller. Moreover, other memory management operations such as wear-leveling, recovery from power loss, and so forth, can be more efficient.

Abstract: A non-volatile programmable circuit configurable to perform logic functions, is provided. The programmable circuit can employ two-terminal non-volatile memory devices to store information, thereby mitigating or avoiding disturbance of programmed data in the absence of external power. Two-terminal resistive switching memory devices having high current on/off ratios and fast switching times can also be employed for high performance, and facilitating a high density array. For look-up table applications, input/output response times can be several nanoseconds or less, facilitating much faster response times than a memory array access for retrieving stored data.

Abstract: A retainer node circuit is provided that can retain state information of a volatile circuit element (e.g., a flip-flop, latch, switch, register, etc.) of an electronic device for planned or unplanned power-down events. The retainer node circuit can include a resistive-switching memory cell that is nonvolatile, having very fast read and write performance. Coupled with power management circuitry, the retainer node circuit can be activated to receive and store a signal (e.g., bit) output by the volatile circuit element, and activated to output the stored signal. Various embodiments disclose non-volatile retention of state information for planned shut-down events as well as unplanned shut-down events. With read and write speeds in the tens of nanoseconds, sleep mode can be provided for volatile circuit elements between clock cycles of longer time-frame applications, enabling intermittent power-down events between active periods. This enables reduction in power without loss of activity for an electronic device.

Abstract: Providing for a memory device having a resistive switching memory integrated within backend layers of the memory device is described herein. By way of example, the resistive switching memory can be embedded memory such as cache, random access memory, or the like, in various embodiments. The resistive memory can be fabricated between various backend metallization schemes, including backend copper metal layers and in part utilizing one or more damascene processes. In some embodiments, the resistive memory can be fabricated in part with damascene processes and in part with subtractive etch processing, utilizing four or fewer photo-resist masks. Accordingly, the disclosure provides a relatively low cost, high performance embedded memory compatible with a variety of fabrication processes of integrated circuit foundries.

Abstract: A memory cell that includes a first metal layer formed over a substrate is provided. The substrate includes one or more complementary metal-oxide semiconductor devices. The memory cell also includes a via device that connects at least a portion of the first metal layer and at least another portion of a second metal layer. The first metal layer has a first thickness having an edge thereof that serves as an electrode for a memory cell formed by the via device. The memory cell scales as a function of the first thickness and at least in part independent of a minimum feature size of the memory device.

Abstract: One potential result of differing characteristics for certain two-terminal memory (TTM) is that memory management techniques, such as logical-to-physical (L2P), can differ as well. Previous memory management techniques do not adequately leverage the advantages associated with TTM. For example, by identifying and leveraging certain advantageous characteristics of TTM, L2P tables can be designed to be smaller and more efficient. Moreover, other memory management operations such as wear-leveling, recovery from power loss, and so forth, can be more efficient.

Abstract: A resistive switching device. The device includes a first electrode comprising a first metal material overlying the first dielectric material and a switching material comprising an amorphous silicon material. The device includes a second electrode comprising at least a second metal material. In a specific embodiment, the device includes a buffer material disposed between the first electrode and the switching material. The buffer material provides a blocking region between the switching material and the first electrode so that the blocking region is substantially free from metal particles from the second metal material when a first voltage is applied to the second electrode.

Abstract: Providing for a multi-bank memory with bank-specific status feedback is described herein. By way of example, the multi-bank memory can be configured to output an availability status, pass/fail status, error correction status, or the like, for subsets of multiple memory banks. In some embodiments, the non-volatile memory can provide global status information, representing a status of all banks commonly in conjunction with bank-specific status information. Further, the subject disclosure provides addressing techniques for identifying particular banks of memory, and obtaining status information for subsets of the memory banks, or performing memory operations on targeted subsets of the memory banks.

Abstract: A first architecture for a volatile resistive-switching device with a selector layer (e.g., a highly resistive layer such as a resistive switching medium) non-planar surfaces is detailed. For example, the selector layer can have a first surface that intersects a second surface at an angle (e.g., oblique angle). The angle can be adjusted to control current-voltage response for the volatile resistive-switching device. A second architecture for volatile resistive-switching device with a first terminal having a high particle diffusivity and a second terminal having a low particle diffusivity. The second architecture can provide diode-like current-voltage responses at a sizes (e.g., sub-20 nanometers) in which conventional diodes do not scale.

Abstract: A non-volatile memory device is provided that uses one or more volatile elements. In some embodiments, the non-volatile memory device can include a resistive two-terminal selector that can be in a low resistive state or a high resistive state depending on the voltage being applied. A MOS (“metal-oxide-semiconductor”) transistor in addition to a capacitor or transistor acting as a capacitor can also be included. A first terminal of the capacitor can be connected to a voltage source, and the second terminal of the capacitor can be connected to the selector device. A floating gate of an NMOS transistor can be connected to the other side of the selector device, and a second NMOS transistor can be connected in series with the first NMOS transistor.

Abstract: Providing for a configuration cells for junction nodes of a field programmable gate array (FPGA) is described herein. By way of example, a configuration cell can comprise non-volatile resistive switching memory to facilitate programmable storage of data as an input to a control circuit of a junction node. The control circuit can activate or deactivate a junction node of the FPGA in response to a value of the data stored in the non-volatile resistive switching memory. The control circuit can comprise an SRAM circuit for fast operation of the junction node. Moreover, the non-volatile memory of the configuration cell facilitates fast power-up of the control circuit utilizing data stored in the resistive switching memory, and minimizes power consumption associated with storing the data.

Abstract: A detection circuit that can detect a two-terminal memory cell changing state. For example, in response to electrical stimuli, a memory cell will change state, e.g., to a defined higher resistance state or a defined lower resistance state. Other, techniques do not detect this state change until after the stimuli is completed and a subsequent sensing operation (e.g., read pulse) is performed. The detection circuit can detect the state change during application of the electrical stimuli that cause the state change and can do so by comparing the magnitudes or values of two particular current parameters.

Abstract: A method for forming a resistive switching device. The method includes providing a substrate having a surface region and forming a first dielectric material overlying the surface region. A first wiring structure is formed overlying the first dielectric material. The method forms one or more first structure comprising a junction material overlying the first wiring structure. A second structure comprising a stack of material is formed overlying the first structure. The second structure includes a resistive switching material, an active conductive material overlying the resistive switching material, and a second wiring material overlying the active conductive material. The second structure is configured such that the resistive switching material is free from a coincident vertical sidewall region with the junction material.

Abstract: Resistive switching memory architectures disclosed herein are capable of achieving fast read/write times and, particularly in the case of multi-bank parallel processing, executing many read or write operations per second. Because resistive switching memory is not guaranteed to be error free, resistive memory controllers can be programmed for error management when paired with such memory architectures. To reduce error management overhead, a dedicated error pin is provided to mitigate or avoid the need for a status read in conjunction with each read or write operation issued to a memory device. A status read can be implemented in response to an error signal on the dedicated error pin, but otherwise can be avoided.

Abstract: Providing for an electronic memory apparatus having high-density, non-volatile memory arrays in conjunction with a high-speed communication interface is disclosed herein. In some embodiments, the electronic memory apparatus can include multiple banks of two-terminal memory, communicatively connected to a modified dynamic random access memory bus and configured to operate according to a modified communication protocol. In one or more embodiments, the high-speed communication interface can comprise more than ten command and address pins to identify individual memory banks (or subsets of memory banks) of the multiple banks of memory, to facilitate bank-specific addressing for memory array operations. In some embodiments, the electronic memory can facilitate status information for subsets of memory banks to facilitate informed array operations, increasing duty cycle of the memory device.

Abstract: Two-terminal memory can be formed into a memory array that contains many discrete memory cells in a physical and a logical arrangement. Where each memory cell is isolated from surrounding circuitry by a single transistor, the resulting array is referred to as a 1T1R memory array. In contrast, where a group of memory cells are isolated from surrounding circuitry by a single transistor, the result is a 1TnR memory array. Because memory cells of a group are not isolated among themselves in the 1TnR case, bit disturb effects are theoretically possible when operating on a single memory cell. Read operations are disclosed for two-terminal memory devices configured to mitigate bit disturb effects, despite a lack of isolation transistors among memory cells of an array. Disclosed operations can facilitate reduced bit disturb effects even for high density two-terminal memory cell arrays.

Abstract: Providing an electrode for a two-terminal memory device is described herein. By way of example, the electrode can comprise a contact surface that comprises at least one surface discontinuity. For instance, the electrode can have a gap, break, or other discontinuous portion of a surface that makes electrical contact with another component of the two-terminal memory device. In one example, the contact surface can comprise an annulus or an approximation of an annulus, having a discontinuity within a center of the annulus, for instance. In some embodiments, a disclosed electrode can be formed from a conductive layer deposited over a non-continuous surface formed by a via or trench in an insulator, or over a pillar device formed from or on the insulator.

Abstract: Solid-state memory having a non-linear current-voltage (I-V) response is provided. By way of example, the solid-state memory can be a selector device. The selector device can be formed in series with a non-volatile memory device via a monolithic fabrication process. Further, the selector device can provide a substantially non-linear I-V response suitable to mitigate leakage current for the non-volatile memory device. In various disclosed embodiments, the series combination of the selector device and the non-volatile memory device can serve as one of a set of memory cells in a 1-transistor, many-resistor resistive memory cell array.

Abstract: A recessed high voltage metal oxide semiconductor (MOS) transistor is provided for use in a two-terminal memory cell. The two-terminal memory cell can include a resistive switching device connected to the recessed MOS transistor. The recessed MOS transistor provides for an increased channel length relative to the transistor size in comparison to a traditional MOS transistor. This allows for a decreased memory cell size while maintaining comparable electrical parameters (threshold voltage, channel length, and leakage) than would otherwise be possible. The recessed MOS transistor can be made as either a NMOS or PMOS device using n-type or p-type materials respectively, where the channel, or inversion layer, is formed by electrons (NMOS) or holes (PMOS) between the source and drain in the transistor.