Abstract:
A non-volatile memory device includes a substrate, an insulating layer, a fin, a number of dielectric layers and a control gate. The insulating layer is formed on the substrate and the fin is formed on the insulating layer. The dielectric layers are formed over the fin and the control gate is formed over the dielectric layers. The dielectric layers may include oxide-nitride-oxide layers that function as a charge storage structure for the memory device.
Abstract:
A method of forming a silicon-on-insulator semiconductor device including providing a substrate and forming a trench in the substrate, wherein the trench includes opposing side walls extending upwardly from a base of the trench. The method also includes depositing at least two insulating layers into the trench to form a shallow trench isolation structure, wherein an innermost of the insulating layers substantially conforms to the base and the two side walls of the trench and an outermost of the insulating layers spans the side walls of the trench so that a gap is formed between the insulating layers in the trench. The gap creates compressive forces within the shallow trench isolation structure, which in turn creates tensile stress within the surrounding substrate to enhance mobility of the device.
Abstract:
A method of forming multiple fins in a semiconductor device includes forming a structure having an upper surface and side surfaces on the semiconductor device. The semiconductor device includes a conductive layer located below the structure. The method also includes forming spacers adjacent the structure and selectively etching the spacers and the conductive layer to form the fins. The fins may be used in a FinFET device.
Abstract:
A method forming a tri-gate fin field effect transistor includes forming an oxide layer over a silicon-on-insulator wafer comprising a silicon layer, and etching the silicon and oxide layers using a rectangular mask to form a mesa. The method further includes etching a portion of the mesa using a second mask to form a fin, forming a gate dielectric layer over the fin, and forming a tri-gate over the fin and the gate dielectric layer.
Abstract:
Dopant deactivation of source/drain extensions during silicidation is reduced by forming deep source/drain regions using a disposable dummy gate as a mask, forming metal silicide layers on the deep source/drain regions, removing the dummy gate and then forming the source/drain extensions using laser thermal annealing. Embodiments include angular ion implantation, after removing the dummy gate, to form spaced apart pre-amorphized regions, ion implanting to form source/drain extension implants extending deeper into the substrate than the pre-amorphized regions, and then laser thermal annealing to activate the source/drain extensions having a higher impurity concentration at the main surface of the substrate than deeper into the substrate. Subsequent processing includes forming sidewall spacers, a gate dielectric layer and then the gate electrode.
Abstract:
The invention provides an improved well structure for electrically separating n-channel and p-channel MOSFETs. The invention first forms a shallow well in a substrate. A buried amorphous layer is then formed below the shallow well. A deep well is then formed below the buried amorphous layer. The substrate is then subjected to a rapid thermal anneal to recrystallize the buried amorphous layer. The well structure is formed by the shallow well and the deep well. A conventional semiconductor device may then be formed above the well structure. The buried amorphous layer suppresses the channeling effect during the forming of the deep well without requiring a tilt angle.
Abstract:
In many packetized communication networks, it is not feasible to obtain exact counts of traffic (OD counts) between specific origin-destination node pairs, because the link counts that are readily obtainable at router interfaces are aggregated indiscriminately over OD pairs. The best that can be done is to make a probabilistic inference concerning the OD counts from the observed link counts. Such an inference relies upon a known linear relationship between observed link counts and unknown OD counts, and a statistical model describing how the values of the OD and link counts are probabilistically distributed. Disclosed is an improved method for making such inferences. The disclosed method takes explicit account of past data when forming a current estimate of the OD counts. As a consequence, behavior that evolves in time is described with improved accuracy and smoothness.
Abstract:
A method of manufacturing a semiconductor device comprises steps of: (a) providing a semiconductor substrate comprising an upper, tensilely strained lattice semiconductor layer and a lower, unstressed semiconductor layer; and (b) forming at least one MOS transistor on or within the tensilely strained lattice semiconductor layer, wherein the forming comprises a step of regulating the drive current of the at least one MOS transistor by adjusting the thickness of the tensilely strained lattice semiconductor layer. Embodiments include CMOS devices formed in substrates including a strained Si layer lattice-matched to a graded composition Si—Ge layer, wherein the thickness of the strained Si layer of each of the PMOS and NMOS transistors is adjusted to provide each transistor type with maximum drive current.
Abstract:
A method for improving the channel doping profile of deep-submicron field effect transistors and MOSFETs. The method involves a highly localized halo implant formed in the channel region but not in the source/drain junction. The halo implant is performed through a gap formed by removal of a temporary spacer. The MOSFET is then further completed.
Abstract:
A method of manufacturing a MOSFET semiconductor device includes forming a gate electrode over a substrate and a gate oxide between the gate electrode and the substrate. Inert dopants are then implanted within the substrate to form amorphized source/drain regions in the substrate extending to a first depth significantly greater than the intended junction depth. The amorphized source/drain regions are implanted with source/drain dopants such that the dopants extend into the substrate to a second depth less than the first depth, above and spaced apart from the end-of-range defect region created at the first depth by the amorphization process. Laser thermal annealing recrystallizes the amorphous regions, activates the source/drain regions and forms source/drain junctions. Because the recrystallization front velocity towards the substrate main surface is greater than the dopant atom velocity in the liquid substrate during laser thermal annealing, the junctions are not pushed down to the amorphous/crystalline silicon interface. Thus, end-of-range defects are located in a region below and spaced apart from the junctions, and the defects are not located in the activated source/drain regions. Junction leakage as a result of the end-of-range defects is thereby reduced.