Taiwan Semiconductor Manufacturing Comp. Ltd.

Tsu-Hui Su

Chih-Ming Chen

Chung-Yi Yu

Ming-Hsi Yeh

Kuo-Bin Huang

Szu-Yu Wang

Chun-Hsiang Fan

Chia-Shiung Tsai

Yu-Wen Wang

Ssu-Yu Liao

Ru-Liang Lee

Ta-Wei Lin

Fin and nanostructured channel structure formation techniques for three-dimensional transistors can tune device performance. For example, fin profile control can be achieved by modifying the shape of fins/nanostructured channel structures so as to reduce their line edge roughness. Consequently, current flow within the channel regions of fins and nanostructured channel structures can be improved, enhancing device performance.

Fin shape modification

Methods for improving profiles of channel regions in semiconductor devices and semiconductor devices formed by the same are disclosed. In an embodiment, a method includes forming a semiconductor fin over a semiconductor substrate, the semiconductor fin including germanium, a germanium concentration of a first portion of the semiconductor fin being greater than a germanium concentration of a second portion of the semiconductor fin, a first distance between the first portion and a major surface of the semiconductor substrate being less than a second distance between the second portion and the major surface of the semiconductor substrate; and trimming the semiconductor fin, the first portion of the semiconductor fin being trimmed at a greater rate than the second portion of the semiconductor fin.

Semiconductor device and method

A method includes forming isolation regions extending into a semiconductor substrate. A semiconductor strip is between the isolation regions. The method further includes recessing the isolation regions so that a top portion of the semiconductor strip protrudes higher than top surfaces of the isolation regions to form a semiconductor fin, measuring a fin width of the semiconductor fin, generating an etch recipe based on the fin width, and performing a thinning process on the semiconductor fin using the etching recipe.

Controlling fin-thinning through feedback

A quantum nano-tip (QNT) thin film, such as a silicon nano-tip (SiNT) thin film, for flash memory cells is provided to increase erase speed. The QNT thin film includes a first dielectric layer and a second dielectric layer arranged over the first dielectric layer. Further, the QNT thin film includes QNTs arranged over the first dielectric layer and extending into the second dielectric layer. A ratio of height to width of the QNTs is greater than 50 percent. A QNT based flash memory cell and a method for manufacture a SiNT based flash memory cell are also provided.

Silicon nano-tip thin film for flash memory cells

Some embodiments of the present disclosure relate to a method for forming flash memory. In this method, a tunnel oxide is formed over a semiconductor substrate. A layer of silicon dot nucleates is formed on the tunnel oxide. The layer of silicon dots includes silicon dot nucleates having respective initial sizes which differ according to a first size distribution. An etching process is performed to reduce the initial sizes of the silicon dot nucleates so reduced-size silicon dot nucleates have respective reduced sizes which differ according to a second size distribution. The second size distribution has a smaller spread than the first size distribution.

Uniformity control for Si dot size in flash memory

The present disclosure relates to a structure and method for reducing dangling bonds around quantum dots in a memory cell. In some embodiments, the structure has a semiconductor substrate having a tunnel dielectric layer disposed over it and a plurality of quantum dots disposed over the tunnel dielectric layer. A passivation layer is formed conformally over outer surfaces of the quantum dots and a top dielectric layer is disposed conformally around the passivation layer. The passivation layer can be formed prior to forming the top dielectric layer over the quantum dots or after forming the top dielectric layer. The passivation layer reduces the dangling bonds at an interface between the quantum dots and the top dielectric layer, thereby preventing trap sites that may hinder operations of the memory cell.

e-Flash Si dot nitrogen passivation for trap reduction

Some embodiments of the present disclosure relate to a method for forming flash memory. In this method, a first tunnel oxide is formed over a semiconductor substrate. A self-assembled monolayer (SAM) is then formed on the first tunnel oxide. The SAM includes spherical or spherical-like crystalline silicon dots having respective diameters which are less than approximately 30 nm. A second tunnel oxide is then formed over the SAM.

Well controlled conductive dot size in flash memory

The present disclosure relates to a structure and method for forming a flash memory cell with an improved erase speed and erase current. Si dots are used for charge trapping and an ONO sandwich structure is formed over the Si dots. Erase operation includes direct tunneling as well as FN tunneling which helps increase erase speed without compensating data retention.

Growing community of inventors

Taipei, Taiwan

Tsu-Hui Su