Advanced Micro Devices Corporation

Vlsi Technology, Inc.

Hughes Aircraft Company

Kuang-Yeh Chang

Yowjuang William Liu

Mark I Gardner

Subhash R Nariani

Fred N Hause

Vivek Jain

Dipankar Pramanik

William J Boardman

Frederick N Hause

Calvin T Gabriel

David P Chan

Robert Louis Dawson

Charles E May

Ramachandr A Rao

Steven W Longcor

David Michael Rogers

Felix H Fujishiro

Hunter B Brugge

Chang-Ou Lee

Yi-Hen Wei

Kuan Y Liao

Hsing-Chien Ma

William W Lee

Jih-Chang Lien

Walter D Parmantie

Michael J Gatto

A metal oxide semiconductor static random access memory (SRAM) includes NMOS transistors and resistor structures implemented without multiple polysilicon layers. According to a first embodiment, the SRAM cell comprises a plurality of appropriately interconnected NMOS transistors having transistor gates formed of a polysilicon layer and resistors formed of the same polysilicon layer. In accordance with a second embodiment, the SRAM cell comprises a plurality of appropriately interconnected NMOS transistors, a dielectric layer overlying the NMOS transistors, and polysilicon resistors passing through the dielectric layer to connect the NMOS transistors to a first metal layer. The dielectric layer, deposited on the NMOS transistors, defines holes exposing drain regions in the NMOS transistors. A polysilicon layer is deposited on the dielectric layer to fill the holes, and the excess polysilicon is removed.

Method for forming an integrated circuit having improved polysilicon resistor structures

The present invention relates to a method and device for providing CMOS logic which can be operated at various operating voltages, without resulting in unbalanced operation of n-channel and p-channel CMOS transistors. In accordance with the present invention, CMOS circuitry can be provided that is operable over a range of voltages (e.g., a range from below 3 volts to a range over 5 volts) without producing unbalanced operation of n-channel and p-channel transistors. Thus, integrated circuits formed in accordance with the present invention can be operated at different voltage power sources without requiring a redesign or relay out of the integrated circuit. In accordance with the present invention, CMOS transistors can be fabricated without increased fabrication complexity to provide transistors which operate within a relatively safe region of their operating characteristics and which operate with a speed that is unaffected by the reduced voltage supply (i.e., no need to accommodate timing errors since both n-channel and p-channel transistor performance remains balanced).

Low voltage CMOS process and device with individually adjustable LDD spacers

A method of forming low resistance contact structures in vias arranged between interconnect levels is provided. The method involves interconnect lines having an anti-reflective layer formed thereupon. An interlevel dielectric layer is formed over the interconnect lines. A photoresist layer is formed over the interlevel dielectric layer and patterned to define via locations. During via etch, an organic (carbon-based) polymer layer forms upon the anti-reflective-coated interconnect lines at the bottoms of the vias. The photoresist and the etch byproduct polymer layers are then removed using a dry etch process which employs a forming gas comprising nitrogen and hydrogen. A native oxide layer subsequently forms upon the anti-reflective-coated interconnect lines when exposed to oxygen. The native oxide layer is then removed, along with any residual etch byproduct polymer, during a sputter etch procedure. Each resulting via is substantially void of polymer and oxide residue so as to present a clean via area which allows ready adherence of a plug material to the anti-reflective coating.

Method of forming low resistance contact structures in vias arranged

An isolation technique is provided for improving the overall planarity of filled isolation regions relative to adjacent silicon mesas. The isolation process results in a silicon mesa having enhanced mechanical and electrical properties. Planarity is performed by repeating the steps of filling isolation trenches, patterning large area isolation trenches, and refilling isolation trenches to present an upper surface having indents which can be readily removed by a chemical-mechanical polish. The silicon mesa upper surface is enhanced by utilizing a unique set of layers stacked upon the silicon substrate, and thereafter patterning the substrate to form raised silicon surfaces, or mesas, having the stacked layers thereon. The patterned, stacked layers include a unique combination of dissimilar compositions which, when removed, leave a silicon mesa upper surface which is recessed below the adjacent, filled trenches. The patterned stacked layers incorporate a polysilicon and/or oxide buffer which prevents deleterious migration of nitrogen from the overlying nitride layer to the underlying silicon mesa upper surface.

Semiconductor trench isolation with improved planarization methodology

A method of forming a self-aligned field oxide isolation structure without using silicon nitride. The method comprises forming a dielectric on an upper surface of a semiconductor substrate. The upper surface of the semiconductor substrate comprises an active region and an isolation region laterally adjacent to each other. A photoresist layer is patterned on top of the implant dielectric to expose regions of the implant dielectric over the active region. Nitrogen is then implanted into the active region through the implant dielectric. Nitrogen is preferably introduced into semiconductor substrate in an approximate atomic concentration of 0.5 to 2.0 percent. After the nitrogen has been implanted into a semiconductor substrate, the photoresist layer is stripped and the implant dielectric is removed. The wafer is then thermally oxidized such that a field oxide having a first thickness is grown over the isolation region and a thin oxide having a second thickness is grown over the active region. The presence of the nitrogen within the semiconductor substrate retards the oxidation rate of the silicon in the active region such that the thickness of the thin oxide is substantially less than the thickness of the thermal oxide. In a presently preferred embodiment, the field oxide has a thickness of 2,000 to 8,000 angstroms while the thin oxide has a thickness of less than 300 angstroms.

Reduced bird's beak field oxidation process using nitrogen implanted

An oxide layer is thermally grown over a semiconductor body, and openings are etched in the oxide layer to expose portions of the surface of the semiconductor body. Then, epitaxial regions are grown from the semiconductor body into the openings in the oxide layer, which epitaxial regions will eventually become the active regions of devices.

Advanced isolation scheme for deep submicron technology

A static metal oxide semiconductor random access memory (SRAM) having NMOS and thin film transistors (TFTs) formed from a single polysilicon layer, and a method for forming the same. The SRAM cell comprises a plurality of NMOS transistors and TFTs that are interconnected by a local interconnect structure. The single layer of poly is used to define the TFT bodies and gates of NMOS transistors in the SRAM cell. Each TFT comprises a single polysilicon layer comprising source gate and drain regions. During the fabrication process, exposed portions of the TFT polysilicon body and exposed regions of NMOS transistors react with a refractory metal silicide to form polycide and silicide regions, respectively. An amorphous silicon pattern also reacts with the refractory metal silicide to form a local interconnect structure connecting the silicided portions of the thin film transistors and the MOS transistors. This arrangement results in a TFT SRAM cell that can be implemented using simple fabrication techniques, such as single poly logic processes or ASIC processes.

Method of making SRAM cell having single layer polysilicon thin film

Semiconductor trench isolation process resulting in a silicon mesa

A novel MOS transistor structure for improving device scaling by improving short channel control includes a buried back gate beneath a channel region of the MOS transistor. A separate contact to a well that is electrically communicated to the buried back gate improves short channel controls without performance degradations. In a preferred embodiment, the back gate is grounded when turning the n-channel MOS transistor off. In alternate embodiments, the buried layer produces retrograde p wells. In some applications, multiple buried layers may be used, with one or more being planar. CMOS devices may have independent, multiple buried back gates.

Method for forming advanced transistor structures with optimum short

CMOS transistors are formed by a damascene process resulting in field oxide regions exhibiting essentially no bird's beak portions. A trench isolation is also formed in a source/drain region each transistor between adjacent junctions.

Reverse CMOS method for dual isolation semiconductor device

Integrated circuit having improved polysilicon resistor structures

A method of manufacturing a transistor having LDD regions in which the source and drain regions are formed by implanting ions through a photoresist layer at an energy of 1 MeV and greater and the LDD regions are formed by low energy ion implantation after the oxide layer is removed from the active region and the gate. In a second embodiment, the source and drain regions are formed without a photoresist layer by ion implantation and the LDD regions are formed by low energy ion implantation after the oxide layer is removed from the active region and the gate.

Method of fabricating FET or CMOS transistors using MeV implantation

A method of forming an improved isolation trench between active regions within the semiconductor substrate. The improved method incorporates a trench liner having a nitrogen content of approximately 0.5 to 2.0 percent. A pad layer is formed on a silicon substrate and a nitride layer is formed on the pad layer. Thereafter, a photoresist layer is patterned on the silicon nitride layer such that regions of the nitride layer are exposed where an isolation trench will subsequently be formed. Next, the exposed regions of the nitride layer and the pad layer situated below the exposed regions of the nitride layer are etched away to expose regions of the silicon substrate. Subsequently, isolation trenches are etched into the silicon substrate with a dry etch process. A trench liner is then formed and nitrogen incorporated into the trench liner. Incorporation of nitrogen into the trench liner can be accomplished by either forming the trench liner in the presence of a nitrogen bearing ambient or by forming a pure SiO.sub.2 trench liner and subsequently implanting the SiO.sub.2 trench liner with nitrogen. After formation of the nitrogenated trench liner, the trench is filled with a dielectric preferably comprised of a CVD oxide. Thereafter, the CVD fill dielectric is planarized and the nitride layer is stripped away.

Nitrogenated trench liner for improved shallow trench isolation

The present invention relates to a method and device for providing CMOS logic which can be operated at various operating voltages, without resulting in unbalanced operation of n-channel and p-channel CMOS transistors. In accordance with the present invention, CMOS circuitry can be provided that is operable over a range of voltages (e.g., a range from below 3 volts to a range over 5 volts) without producing unbalanced operation of n-channel and p-channel transistors. Thus, integrated circuits formed in accordance with the present invention can be operated at different voltage power sources without requiring a redesign or relayout of the integrated circuit. In accordance with the present invention, CMOS transistors can be fabricated without increased fabrication complexity to provide transistors which operate within a relatively safe region of their operating characteristics and which operate with a speed that is unaffected by the reduced voltage supply (i.e., no need to accommodate timing errors since both n-channel and p-channel transistor performance remains balanced).

Growing community of inventors

Los Gatos, CA, United States of America

Kuang-Yeh Chang