Abstract: A block storage system includes a host and comprises a block storage module that is coupled to the host. The block storage module includes a MRAM array and a bridge controller buffer coupled to communicate with the MRAM array. The MRAM array includes a buffer widow that is moveable within the MRAM array to allow contents of the MRAM array to be read by the host through the bridge controller buffer even when the capacity of the bridge controller buffer is less than the size of the data being read from the MRAM array.

Abstract: A method of managing redundant array of independent disk (RAID) groups in a storage system includes determining wear of each of the plurality of RAID groups, computing the weight for each of RAID groups based on the wear, and striping data across at least one of the RAID groups based on the weight of each of the RAID groups.

Abstract: A method of manufacturing a magnetic memory cell, including a magnetic tunnel junction (MTJ), includes using silicon nitride layer and silicon oxide layer to form a trench for depositing copper to be employed for connecting the MTJ to other circuitry without the use of a via.

Abstract: A hybrid non-volatile memory device includes a non-volatile random access memory (NVRAM) having an array of magnetic memory elements, the NVRAM being bit-accessible. The hybrid non-volatile device further includes a non-volatile page-mode memory (PMM) made of resistive memory and organized into pages, the non-volatile PMM being page-accessible. Further included in the hybrid non-volatile memory device is a direct memory access (DMA) engine that is coupled to the NVRAM and the non-volatile PMM and transfers data between the NVRAM and the non-volatile PMM during a DMA operation.

Abstract: The present invention is directed to an MTJ memory element including a magnetic free layer structure which comprises one or more magnetic free layers that have a same variable magnetization direction substantially perpendicular to layer planes thereof; an insulating tunnel junction layer formed adjacent to the magnetic free layer structure; a magnetic reference layer structure comprising a first magnetic reference layer formed adjacent to the insulating tunnel junction layer and a second magnetic reference layer separated therefrom by a perpendicular enhancement layer with the first and second magnetic reference layers having a first fixed magnetization direction substantially perpendicular to layer planes thereof; an anti-ferromagnetic coupling layer formed adjacent to the second magnetic reference layer opposite the perpendicular enhancement layer; and a magnetic fixed layer comprising first and second magnetic fixed sublayers with the second magnetic fixed sublayer formed adjacent to the anti-ferromagnetic

Abstract: A memory device is configured to correct errors in codewords written to a memory array. Errors, if any, in a first codeword are corrected and a codeword corrector output is generated including a corrected first codeword. A data buffer receives the codeword corrector output and a first user data associated with the addressed page and generates a data buffer output including the corrected first codeword, as modified by the first user data, defined as a first codeword output. A codeword encoder receives the data buffer output and encodes the first codeword output to generate an encoded first codeword output included in a codeword encoder output. A write buffer receives the codeword encoder output and saves the same for writing to the memory array. Writing to the memory array is performed while receiving a second user data, which has a second codeword associated therewith, and correcting the second codeword.

Abstract: The present invention is directed to a spin transfer torque (STT) MRAM device having a perpendicular magnetic tunnel junction (MTJ) memory element. The memory element includes a perpendicular MTJ structure in between a non-magnetic seed layer and a non-magnetic cap layer. The MTJ structure comprises a magnetic free layer structure and a magnetic reference layer structure with an insulating tunnel junction layer interposed therebetween, an anti-ferromagnetic coupling layer formed adjacent to the magnetic reference layer structure, and a magnetic fixed layer formed adjacent to the anti-ferromagnetic coupling layer. At least one of the magnetic free and reference layer structures includes a non-magnetic perpendicular enhancement layer, which improves the perpendicular anisotropy of magnetic layers adjacent thereto.

Abstract: Chip packages are described with soft-magnetic shields that are included inside or attached externally to the package containing a MRAM chip. In one group of embodiments a single shield with vias for bonding wires is affixed to the surface of the MRAM chip having the contact pads. The limitation of shield to chip distance due to bonding wire is eliminated by VIA holes according to the invention which achieves minimal spacing between the shield and chip. A second shield without vias can be positioned on the opposite side of the chip from the first shield. In one group of embodiments a hardened ferro-fluid shield can be the only shield or the structure can include a shield with or without vias. One group of embodiments includes an external shield with vias for solder access to the package contact pads affixed to the outer surface of the package.

Abstract: A spin transfer torque magnetic random access memory (STTMRAM) element and a method of manufacturing the same is disclosed having a free sub-layer structure with enhanced internal stiffness. A first free sub-layer is deposited, the first free sub-layer being made partially of boron (B), annealing is performed of the STTMRAM element at a first temperature after depositing the first free sub-layer to reduce the B content at an interface between the first free sub-layer and the barrier layer, the annealing causing a second free sub-layer to be formed on top of the first free sub-layer and being made partially of B, the amount of B of the second free sub-layer being greater than the amount of B in the first free sub-layer.

Abstract: Embodiments of the invention are described that use a thin metallic hard mask, which can be a bi-layer film, to increase the incident IBE angle for MTJ sidewall cleaning without losing the process margin for the subsequent interconnection process. The patterned metallic hard mask pads also serve as the top electrode for the MTJ cells. Using a thin metallic hard mask is possible when the hard mask material acts as a CMP stopper without substantial loss of thickness. In the first embodiment, the single layer hard mask is preferably ruthenium. In the second embodiment, the lower layer of the bi-layer hard mask is preferably ruthenium. The wafer is preferably rotated during the IBE process for uniform etching. A capping layer under the hard mask is preferably used as the etch stopper during hard mask etch process in order not to damage or etch through the upper magnetic layer.

Abstract: A method of writing to magnetic tunnel junctions (MTJs) of a magnetic memory array includes storing in-coming data in a cache register, reading the present logic state of a first one of a set of at least two MTJs, the set of at least two MTJs including the first MTJ and a second MTJ. The in-coming data is to be written into the second MTJ. Further steps are storing the read logic state into a data register, swapping the contents of the data register and the cache register so that the cache register stores the read logic state and the data register stores the in-coming data, applying a first predetermined voltage level to the set of MTJs thereby causing the first MTJ to be over-written, applying a second predetermined voltage level to the set of MTJs, and storing the in-coming data into the second MTJ.

Abstract: A method of configuring a computer memory system includes receiving a request from customized software driver or a BIOS extension software or a customized legacy BIOS or a customized UEFI PMM extension software or a customized UEFI BIOS, scanning memory module sockets in response to the request, recognizing memory modules in the memory module sockets, the memory modules being made of, at least in part, persistent memory modules (PMMs), configuring the PMMs to be invisible to the OS, and storing the mapping information to a designated protected persistent memory area, and presenting the PMMs as a persistent block storage to the OS.

Abstract: A spin transfer torque magnetic random access memory (STTMRAM) element and a method of manufacturing the same is disclosed having a free sub-layer structure with enhanced internal stiffness. A first free sub-layer is deposited, the first free sub-layer being made partially of boron (B), annealing is performed of the STTMRAM element at a first temperature after depositing the first free sub-layer to reduce the B content at an interface between the first free sub-layer and the barrier layer, the annealing causing a second free sub-layer to be formed on top of the first free sub-layer and being made partially of B, the amount of B of the second free sub-layer being greater than the amount of B in the first free sub-layer.

Abstract: A multi-state current-switching magnetic memory element includes a stack of magnetic tunneling junction (MTJ) separated by a non-magnetic layer for storing more than one bit of information, wherein different levels of current applied to the memory element cause switching to different states.

Abstract: A high-availability storage system includes a first storage system and a second storage system. The first storage system includes a first Central Processing Unit (CPU), a first physically-addressed solid state disk (SSD) and a first non-volatile memory module that is coupled to the first CPU. Similarly, the second storage system includes a second CPU and a second SSD. Upon failure of one of the first or second CPUs, or the storage system with the non-failing CPU continues to be operational and the storage system with the failed CPU is deemed inoperational and the first and second SSDs remain accessible.

Abstract: A storage system includes a Central Processing Unit (CPU) that has a physically-addressed solid state disk (SSD), addressable using physical addresses associated with user data and provided by a host. The user data is to be stored in or retrieved from the physically-addressed SSD in blocks. Further, a non-volatile memory module is coupled to the CPU and includes flash tables used to manage blocks in the physically addressed SSD. The flash tables have tables that are used to map logical to physical blocks for identifying the location of stored data in the physically addressed SSD. The flash tables are maintained in the non-volatile memory modules thereby avoiding reconstruction of the flash tables upon power interruption.

Abstract: A method of programming a voltage-controlled magnetoresistive tunnel junction (MTJ) includes applying a programming voltage pulse (Vp), reading the voltage-controlled MTJ, and determining if the voltage-controlled MTJ is programmed to a desired state and if not, changing the Vp and repeating the applying and reading steps until the voltage-controlled MTJ is programmed to the desired state.

Abstract: The present invention is directed to a method for reading and writing an STT-MRAM multi-level cell (MLC), which includes a plurality of memory elements coupled in series. The method detects the resistance states of individual memory elements in an MLC by sequentially writing at least one of the plurality of memory element to the low resistance state in order of ascending write current threshold. If a written element switches the resistance state thereof after the write step, then the written element was in the high resistance state prior to the write step. Otherwise, the written element was in the low resistance state prior to the write step. The switching of the resistance state can be ascertained by comparing the resistance or voltage values of the plurality of memory elements before and after writing each of the plurality of memory elements in accordance with the embodiments of the present invention.

Abstract: A memory device comprises a semiconductor substrate having a plurality of parallel trenches therein, a memory region formed in the substrate including an array of memory cells having a plurality of vertical selection transistors with respective channels formed in trench sidewalls, a plurality of buried source electrodes in trench bottoms, a plurality of paired gate electrodes formed on paired trench sidewalls, a first and second stitch region disposed adjacent the memory region along a trench direction including a first and second row of gate contacts, respectively, and a row of source contacts disposed in the first or second stitch region with each of the source contacts coupled to a respective one of the source electrodes. One of each pair of the gate electrodes is coupled to a respective one of the first row of gate contacts and the other one of each pair of gate electrodes is coupled to a respective one of the second row of gate contacts.

Abstract: Resistive memory cell array fabricated with unit areas able to be scaled down to 4 F2, where F is minimum feature size in a technology node are described. Memory cells in a pair of cells commonly include a pair of buried sources in the bottom of trenches formed in a silicon substrate. The source line is shared with an adjacent cell. A pair of gate electrodes provides a vertical channel on a sidewall of the trench. A buried word line connects the bottom of the gates on the sidewall overlying the source wherein the word line is looped at the end of the array. A drain, which is self-aligned to the gate, is formed by implantation/doping the surface of the silicon before patterning the trenches. A contact is formed on top of the drain and the resistive memory element is fabrication on the contact.