Abstract:
Methods and structures for forming strained-channel FETs are described. A strain-inducing layer may be formed under stress in a silicon-on-insulator substrate below the insulator. Stress-relief cuts may be formed in the strain-inducing layer to relieve stress in the strain-inducing layer. The relief of stress can impart strain to an adjacent semiconductor layer. Strained-channel, fully-depleted SOI FETs and strained-channel finFETs may be formed from the adjacent semiconductor layer. The amount and type of strain may be controlled by etch depths and geometries of the stress-relief cuts and choice of materials for the strain-inducing layer.
Abstract:
Methods and structures for forming strained-channel finFETs are described. Fin structures for finFETs may be formed in two epitaxial layers that are grown over a bulk substrate. A first thin epitaxial layer may be cut and used to impart strain to an adjacent channel region of the finFET via elastic relaxation. The structures exhibit a preferred design range for increasing induced strain and uniformity of the strain over the fin height.
Abstract:
Methods and structures for increasing strain in fully insulated finFETs are described. The finFET structures may be formed on an insulating layer and include source, channel, and drain regions that are insulated all around. During fabrication, the source and drain regions may be formed as suspended structures. A strain-inducing material may be formed around the source and drain regions on four contiguous sides so as to impart strain to the channel region of the finFET.
Abstract:
Methods and structures for forming strained-channel finFETs are described. Fin structures for finFETs may be formed in two epitaxial layers that are grown over a bulk substrate. A first thin epitaxial layer may be cut and used to impart strain to an adjacent channel region of the finFET via elastic relaxation. The structures exhibit a preferred design range for increasing induced strain and uniformity of the strain over the fin height.
Abstract:
The invention concerns a method of forming a semiconductor layer having uniaxial stress including: forming, in a semiconductor structure having a stressed semiconductor layer, one or more first isolation trenches in a first direction for delimiting a first dimension of at least one transistor to be formed in said semiconductor structure; forming, in the semiconductor structure, one or more second isolation trenches in a second direction for delimiting a second dimension of the at least one transistor, the first and second isolation trenches being at least partially filled with an insulating material; and before or after the formation of the second isolation trenches, decreasing the viscosity of the insulating material in the first isolation trenches by implanting atoms of a first material into the first isolation trenches, wherein atoms of the first material are not implanted into the second isolation trenches.
Abstract:
The disclosure concerns a method of stressing a semiconductor layer comprising: forming, over a silicon on insulator structure having a semiconductor layer in contact with an insulating layer, one or more stressor blocks aligned with first regions of said semiconductor layer in which transistor channels are to be formed, wherein said stressor blocks are stressed such that they locally stress said semiconductor layer; and deforming second regions of said insulating layer adjacent to said first regions by temporally decreasing, by annealing, the viscosity of said insulator layer.
Abstract:
A thermal annealing flow process includes the steps of: depositing a metal or metal alloy on a silicon semiconductor structure, performing a first annealing of a rapid thermal anneal (RTA) type to produce a metal rich phase in a portion of the silicon semiconductor structure, removing unreacted metal or metal alloy and performing a second annealing as a millisecond annealing at a temperature that is below a melt temperature of the silicon material present in the silicon semiconductor structure.
Abstract:
Integrated circuits are disclosed in which the strain properties of adjacent pFETs and nFETs are independently adjustable. The pFETs include compressive-strained SiGe on a silicon substrate, while the nFETs include tensile-strained silicon on a strain-relaxed SiGe substrate. Adjacent n-type and p-type FinFETs are separated by electrically insulating regions formed by a damascene process. During formation of the insulating regions, the SiGe substrate supporting the n-type devices is permitted to relax elastically, thereby limiting defect formation in the crystal lattice of the SiGe substrate.
Abstract:
A tensile strained silicon layer is patterned to form a first group of fins in a first substrate area and a second group of fins in a second substrate area. The second group of fins is covered with a tensile strained material, and an anneal is performed to relax the tensile strained silicon semiconductor material in the second group of fins and produce relaxed silicon semiconductor fins in the second area. The first group of fins is covered with a mask, and silicon-germanium material is provided on the relaxed silicon semiconductor fins. Germanium from the silicon germanium material is then driven into the relaxed silicon semiconductor fins to produce compressive strained silicon-germanium semiconductor fins in the second substrate area (from which p-channel finFET devices are formed). The mask is removed to reveal tensile strained silicon semiconductor fins in the first substrate area (from which n-channel finFET devices are formed).
Abstract:
A method for making a semiconductor device may include forming, on a first semiconductor layer of a semiconductor-on-insulator (SOI) wafer, a second semiconductor layer comprising a second semiconductor material different than a first semiconductor material of the first semiconductor layer. The method may further include performing a thermal treatment in a non-oxidizing atmosphere to diffuse the second semiconductor material into the first semiconductor layer, and removing the second semiconductor layer.