A semiconductor device having good characteristics without variation and a method of manufacturing the same are provided. A part of a conductive layer for a floating gate is removed by using a spacer insulating film, a first insulating film, and a second insulating film as a mask. A floating gate having a tip portion is formed from the conductive layer for the floating gate, and a part of an insulating layer for a gate insulating film is exposed from the floating gate. The tip portion of the floating gate is further exposed by selectively removing the second insulating film among the second insulating film, the insulating layer for the gate insulating film, and the spacer insulating film.

A memory cell disposed on a substrate has a first gate structure and a second gate structure. The memory cell includes a first heavily doped region adjacent to an outer side of the first gate structure. Further, a first lightly doped drain (LDD) region with a first type dopant is between the first heavily doped region and the outer side of the first gate structure. A pocket doped region with a second type dopant is overlapping with the first LDD region. The second type dopant is opposite to the first type dopant in conductive type. A second heavily doped region is adjacent to an outer side of the second gate structure, opposite to the first heavily doped region. A second LDD region with the first type dopant is disposed between the first gate structure and the second gate structure.

A memory device, a peripheral circuit thereof and a single-byte data write method thereof are provided. The peripheral circuit includes a Y decoder, a page buffer, and a write circuit. The write circuit is coupled to a memory array and the page buffer through the Y decoder and receives a byte of program data. The write circuit is based on a memory address corresponding to the program data to receive a plurality of bytes of array data stored in the memory array through the Y decoder, and the read array data is written to page buffer through the Y decoder. Next, the program data is written to the memory array through the write circuit and Y decoder, and the array data is written to the memory array by the page buffer.

Metal floating gate electrodes can be formed for a three-dimensional memory device by forming a memory opening having lateral recesses at levels of spacer material layers between insulating layers, depositing a continuous metal layer, and inducing diffusion and agglomeration of the metal into the lateral recesses to form discrete metal portions employing an anneal process. The metallic material can migrate and form the discrete metal portions due to surface tension, which operates to minimize the surface area of the metallic material. Optionally, two or more continuous metal layers can be employed to form discrete metal portions including at least two metals. Optionally, a selective metal deposition process can be performed to deposit additional metal portions including a different metallic material on the discrete metal portions. The metal floating gate electrodes can be formed without employing an etch process. A tunneling dielectric layer and a semiconductor channel can be subsequently formed.

A non-volatile memory cell includes a substrate, a select gate, a floating gate, and an assistant control gate. The substrate includes a first diffusion region, a second diffusion region, a third diffusion region, and a fourth diffusion region. The select gate is formed above the first diffusion region and the second diffusion region in a polysilicon layer. The floating gate is formed above the second diffusion region, the third diffusion region and the fourth diffusion region in the polysilicon layer. The assistant control gate is formed above the floating gate in a metal layer, wherein an area of the assistant control gate overlaps with at least half an area of the floating gate.

Devices and methods for forming a device are presented. The method includes providing a substrate prepared with a cell area separated by other active areas by isolation regions. First and second lower sub-gates of first and second transistors are formed in the cell area. A common upper sub-gate of the first and second transistors is formed. The common upper sub-gate and first and second lower sub-gates are separated by an intergate dielectric layer and the common upper sub-gate surrounds the first and second lower sub-gates.

At least one of a plurality of transistors which are highly integrated in an element is provided with a back gate without increasing the number of manufacturing steps. In an element including a plurality of transistors which are longitudinally stacked, at least a transistor in an upper portion includes a metal oxide having semiconductor characteristics, a same layer as a gate electrode of a transistor in a lower portion is provided to overlap with a channel formation region of the transistor in an upper portion, and part of the same layer as the gate electrode functions as a back gate of the transistor in an upper portion. The transistor in a lower portion which is covered with an insulating layer is subjected to planarization treatment, whereby the gate electrode is exposed and connected to a layer functioning as source and drain electrodes of the transistor in an upper portion.

A method for forming a semiconductor device structure is provided. The method includes forming a mask layer over a substrate. The method includes forming a first isolation structure and a second isolation structure passing through the mask layer and penetrating into the substrate. The method includes thinning the mask layer to expose a first portion of the first isolation structure and a second portion of the second isolation structure. The method includes partially removing the first portion, the second portion, the third portion, and the fourth portion. The method includes removing the thinned mask layer. The method includes forming a first gate over the substrate and between the first isolation structure and the second isolation structure. The method includes forming a dielectric layer over the first gate. The method includes forming a second gate over the dielectric layer and above the first gate.

A memory cell includes a read transistor, a first floating gate transistor, a program transistor, a second floating gate transistor, and a common floating gate. The common floating gate is coupled to the second floating gate transistor and the first floating gate transistor. The memory cell is programmed and erased through the common floating gate on the second floating gate transistor, and is read through the first floating gate transistor and the read transistor.

A method of forming split gate non-volatile memory cells on the same chip as logic and high voltage devices having HKMG logic gates. The method includes forming the source and drain regions, floating gates, control gates, and the poly layer for the erase gates and word line gates in the memory area of the chip. A protective insulation layer is formed over the memory area, and an HKMG layer and poly layer are formed on the chip, removed from the memory area, and patterned in the logic areas of the chip to form the logic gates having varying amounts of underlying insulation.