The present application provides a method for improving the metal work function boundary effect in FinFET process, the method comprises steps of: depositing a first TiN layer on four fin structures. The first TiN layer has no gap between the second and the third fin structures; removing the first TiN layer up to a first distance from the midline between the second and third fin structures at the second fin structure side; depositing a second TiN layer; removing the second and first TiN layers from second fin structure. The thickness of the TiN layer at the bottom edge of the fin structure at the later structure of the ultra-low threshold voltage P-type transistor will be smaller from this process. Thus formed TiN layer is less prone to a bottom undercut during etching, thereby reducing the metal boundary effect and increasing of the threshold voltage of the device.

A semiconductor device includes a base substrate including an NMOS region and a PMOS region. The PMOS region includes a first P-type region and a second P-type region. The semiconductor device also includes an interlayer dielectric layer, a gate structure formed through the interlayer dielectric layer and including an N-type region gate structure formed in the NMOS region, a first gate structure formed in the first P-type region and connected to the N-type region gate structure, and a second gate structure formed in the second P-type region and connected to the first gate structure. The direction from the N-type region gate structure to the second gate structure is an extending direction of the N-type region opening, and along a direction perpendicular to the extending direction of the N-type region opening, the width of the first gate structure is larger than the width of the second gate structure.

Embodiments of three-dimensional (3D) memory devices having through array contacts (TACs) and methods for forming the same are disclosed. In an example, a method for forming a 3D memory device is disclosed. A dielectric stack including interleaved a plurality of dielectric layers and a plurality of sacrificial layers is formed above a substrate. A channel structure extending vertically through the dielectric stack is formed. A first opening extending vertically through the dielectric stack is formed. A spacer is formed in a plurality of shallow recesses and on a sidewall of the first opening. The plurality of shallow recesses abut the sidewall of the first opening. A TAC extending vertically through the dielectric stack is formed by depositing a conductor layer in contact with the spacer in the first opening. A slit extending vertically through the dielectric stack is formed.

A photoelectric conversion device including a plurality of substrates in a stacked state, the plurality of substrates including a first substrate and a second substrate electrically connected to each other, the photoelectric conversion device comprising: a memory cell unit including row-selection lines that are to be driven upon selection of a row of a memory cell array and column-selection lines that are to be driven upon selection of a column of the memory cell array; and a memory peripheral circuit unit that includes row-selection line connection portions and column-selection line connection portions so as to drive the row-selection lines and to drive the column-selection lines, wherein a first portion that is at least a part of the memory peripheral circuit unit is formed on the first substrate and the memory cell unit is formed on the second substrate.

In some embodiments of the method, patterning the opening includes: projecting a radiation beam toward the second dielectric layer, the radiation beam having a pattern of the opening. In some embodiments of the method, the single-patterning photolithography process is an extreme ultraviolet (EUV) lithography process. In some embodiments of the method, filling the opening with the conductive material includes: plating the conductive material in the opening; and planarizing the conductive material and the second dielectric layer to form the first metal line from remaining portions of the conductive material, top surfaces of the first metal line and the second dielectric layer being planar after the planarizing.

A method for forming a semiconductor structure includes providing a substrate, including a first region and a second region adjacent to the first region; forming a first dielectric layer on the substrate in the first region and the second region; and forming a plurality of first plug structures in the first dielectric layer. The top surface of each first plug structure is exposed by the first dielectric layer. The method further includes forming a first conductive layer on the first dielectric layer of the second region; forming a second dielectric layer on the first dielectric layer of the first region and on the first conductive layer of the second region; and forming a plurality of second plug structures in the second dielectric layer of the first region. The bottom surface of each second plug structure is in contact with the top surface of a first plug structure.

According to one embodiment, a semiconductor device includes a semiconductor layer, an element region provided on the semiconductor layer convexly, having a predetermined width in a first direction along a surface of the semiconductor layer, and extending in a second direction along the surface of the semiconductor layer and intersecting the first direction, a gate electrode arranged above the element region, a liner layer covering the gate electrode, and an element separation portion extends in the second direction on both sides of the element region in the first direction, and the liner layer continuously extends from the gate electrode to the element separation portion and the liner layer in the element separation portion lies below the element separation portion.

A method includes forming a first transistor including forming a first gate stack, epitaxially growing a first source/drain region on a side of the first gate stack, and performing a first implantation to implant the first source/drain region. The method further includes forming a second transistor including forming a second gate stack, forming a second gate spacer on a sidewall of the second gate stack, epitaxially growing a second source/drain region on a side of the second gate stack, and performing a second implantation to implant the second source/drain region. An inter-layer dielectric is formed to cover the first source/drain region and the second source/drain region. The first implantation is performed before the inter-layer dielectric is formed, and the second implantation is performed after the inter-layer dielectric is formed.

A semiconductor device that can reduce power consumption while improving the accuracy of learning and inference is provided. The semiconductor device is connected to data lines PBL, NBL, and comprises a product operation memory cell 1 for storing data of ternary value and performing a product-sum operation between a stored data and an input data INP and a data in the data lines PBL, NBL.

A semiconductor structure includes an SRAM cell that includes first and second pull-up (PU) transistors, first and second pull-down (PD) transistors, first and second pass-gate (PG) transistors, and bit line (BL) conductors. The first PU and the first PD transistors form a first inverter. The second PU and the second PD transistors form a second inverter. The first and the second inverters are cross-coupled to form two storage nodes that are coupled to the BL conductors through the first and the second PG transistors. The first and the second PU transistors are formed over an n-type active region over a frontside of the semiconductor structure. The first and the second PD transistors and the first and the second PG transistors are formed over a p-type active region over the frontside of the semiconductor structure. The BL conductors are disposed over a backside of the semiconductor structure.