A memory device comprises multiple memory dice arranged vertically in a stack of memory dice and at least one thermoelectric die contacting the bulk silicon layer of at least one of the memory dice of the multiple memory dice. Each memory die of the multiple memory dice includes an active circuitry layer that includes memory cells of a memory array and a bulk silicon layer. The thermoelectric die is configured to one or both of reduce heat from the memory die when a current is applied to terminals of the thermoelectric die and generate a voltage at the terminals of the thermoelectric die when heat from the memory die is applied to the thermoelectric die.

A semiconductor device structure is provided. The semiconductor device structure includes a semiconductor substrate. The semiconductor device structure includes a gate stack over the semiconductor substrate. The gate stack includes a first insulating layer, a charge trapping structure, a second insulating layer, and a gate electrode. The first insulating layer separates the semiconductor substrate from the charge trapping structure. The charge trapping structure is between the first insulating layer and the second insulating layer. The gate electrode is over the second insulating layer. The charge trapping structure includes a first layer and a second layer. The first layer includes zinc oxide, tin dioxide, titanium oxide, zinc tin oxide, indium oxide, indium zinc oxide, indium gallium zinc oxide, zinc oxynitride, tin oxynitride, titanium oxynitride, zinc tin oxynitride, indium oxynitride, indium zinc oxynitride, or indium gallium zinc oxynitride. The second layer includes nickel oxide, tin oxide, copper oxide, nickel oxynitride, tin oxynitride, or copper oxynitride. The semiconductor device structure includes a first doped region and a second doped region in the semiconductor substrate and on two opposite sides of the gate stack.

Example embodiments relate to a method of fabricating a synapse memory device capable of being driven at a low voltage and realizing a multi-level memory. The synapse memory device includes a two-transistor structure in which a drain region of a first transistor including a memory layer and a first source region of a second transistor share a source-drain shared area. The synapse memory device is controlled by applying a voltage through the source-drain shared area. The memory layer includes a charge trap layer and a threshold switching layer, and may realize a non-volatile multi-level memory function.

The nonvolatile memory device includes a memory cell having a transistor in which an insulating isolation layer is formed in a channel region. The nonvolatile memory device includes a metal-oxide-semiconductor (MOS) transistor as a basic component. An insulating isolation layer is formed in at least a channel region, and a gate insulating layer includes an insulating layer or a variable resistor and serves as a data storage. A gate includes a metal layer formed in a lower portion thereof. First source and drain regions are lightly doped with a dopant, and second source and drain regions are heavily doped with a dopant.

A semiconductor device structure is provided. The semiconductor device structure includes a semiconductor substrate. The semiconductor device structure includes a gate stack over the semiconductor substrate. The gate stack includes a first insulating layer, a first layer, a second layer, a second insulating layer, and a gate electrode. The first insulating layer separates the semiconductor substrate from the first layer. The second layer is between the first layer and the second insulating layer. The gate electrode is over the second insulating layer. There is a P-N junction between the first layer and the second layer. The semiconductor device structure includes a first doped region and a second doped region in the semiconductor substrate. The first layer, the first doped region, and the second doped region have a first type conductivity, which is opposite to a second type conductivity of the second layer.

Example embodiments relate to a method of fabricating a synapse memory device capable of being driven at a low voltage and realizing a multi-level memory. The synapse memory device includes a two-transistor structure in which a drain region of a first transistor including a memory layer and a first source region of a second transistor share a source-drain shared area. The synapse memory device is controlled by applying a voltage through the source-drain shared area. The memory layer includes a charge trap layer and a threshold switching layer, and may realize a non-volatile multi-level memory function.

There is provided a structure of a nano memory system. The disclosed unit nano memory cell comprises a single isolated nanoparticle placed on the surface of a semiconductor substrate (301) and an adjacent nano-Schottky contact (303). The nanoparticle works as a storage site where the nano-Schottky contact (303) works as a source or a drain of electrons, in or out of the semiconductor substrate (301), at a relatively small voltage. The electric current through the nano-Schottky contact (303) can be turned on (reading 1) or off (reading 0) by charging or discharging the nanoparticle. Since the electric contact is made by a nano-Scottky contact (303) on the surface and the back contact of the substrate (301), and the charge is stored in a very small nanoparticle, this allows to attain the ultimate device down-scaling. This would also significantly increase the number of nano memory cells on a chip. Moreover, the charging and discharging (writing/erasing), as well as the reading voltages are lower than those needed for CMOS based flash memory cells, due to the small nano-Schottky contact (301) and the small size of the nanoparticle for charge storage.

A multilayer stacking wafer bonding structure is provided in the present invention, including a logic wafer with a substrate and a logic circuit layer on the substrate, multiple memory wafers bonded sequentially on the logic circuit layer to form a first multilayer stacking structure, wherein each memory wafer includes a memory layer, a silicon layer on the memory layer and multiple oxide layers in trenches of the silicon layer, and the oxide layers in the memory wafers are aligned each other in a direction vertical to the substrate, and multiple through-oxide vias (TOV) extending through the memory layers and the oxide layers in the first multilayer stacking structure into the logic circuit layer, and the TOVs do not extend through any of the silicon layers.

A method for manufacturing a semiconductor device, a first structure is formed on a first substrate. A first bonded body is formed by bonding a supporting substrate lower in rigidity than the first substrate to a first principal surface, on which the first structure is formed, of the first substrate. The first substrate is removed from the first bonded body. A second structure is formed on a second substrate. A third structure is formed on a third substrate. A second bonded body is formed by bonding a second principal surface, on which the second structure is formed, of the second substrate to a third principal surface, on which the third structure is formed, of the third substrate. The second substrate is removed from the second bonded body. A third bonded body is formed by bonding a fourth principal surface, which is exposed after the first substrate is removed, of the first bonded body to a fifth principal surface, which is exposed after the second substrate is removed, of the second bonded body. The supporting substrate is removed from the third bonded body.

A method includes: providing a plurality of memory cells arranged in rows and a columns, wherein each of the plurality of memory cells comprises a dual-gate transistor, the dual-gate transistor comprising a silicon substrate, a channel layer over the silicon substrate, a first gate structure under the channel layer, and a second gate structure over the channel layer; providing a plurality of word lines extending in a first direction and electrically connected to the rows, respectively, and wherein the first gate structure and the second gate structure of the dual-gate transistor of each of the plurality of memory cells are electrically connected to one of the word lines; providing a plurality of source lines extending in a second direction and electrically connected to the columns, respectively; and providing a plurality of bit lines extending in the second direction and electrically connected to the columns, respectively.