A method of forming a semiconductor structure is provided. A substrate having a memory region is provided. A plurality of fin structures are provided and each fin structure stretching along a first direction. A plurality of gate structures are formed, and each gate structure stretches along a second direction. Next, a dielectric layer is formed on the gate structures. A first patterned mask layer is formed, wherein the first patterned mask layer has a plurality of first trenches stretching along the second direction. A second patterned mask layer on the first patterned mask layer, wherein the second patterned mask layer comprises a plurality of first patterns stretching along the first direction. Subsequently, the dielectric layer is patterned by using the first patterned mask layer and the second patterned mask layer as a mask to form a plurality of contact vias. The contact holes are filled with a conductive layer.

A fin structure cutting process includes the following steps. Four fin structures are formed in a substrate, where the four fin structures including a first fin structure, a second fin structure, a third fin structure and a fourth fin structure are arranged sequentially and parallel to each other. A first fin structure cutting process is performed to remove top parts of the second fin structure and the third fin structure, thereby a first bump being formed from the second fin structure, and a second bump being formed from the third fin structure. A second fin structure cutting process is performed to remove the second bump and the fourth fin structure completely, but to preserve the first bump beside the first fin structure. Moreover, the present invention provides a fin structure formed by said process.

A method for fabricating semiconductor device includes the steps of: providing a substrate having a first region, a second region, and a third region; forming a plurality of spacers on the first region, the second region, and the third region; forming a first patterned mask to cover the spacers on the first region and the second region; and removing the spacers on the third region.

A method for manufacturing fins includes following steps. A substrate including a plurality of fins formed thereon is provided. At least an ion implantation is performed to the fins. A thermal process is performed after the ion implantation. An insulating layer is formed on the substrate, and the fins are embedded in the insulating layer. Thereafter, a portion of the insulating layer is removed to form an isolation structure on the substrate, and the fins are exposed from a top surface of the isolation structure. The insulating layer is formed after the ion implantation and the thermal process. Or, the isolation structure is formed before the ion implantation, or between the ion implantation and the thermal process.

A method for manufacturing a semiconductor device includes forming gate structures spaced apart from each other on a substrate, gate spacers covering sidewalls of the gate structures, and an interlayer insulating layer covering the gate spacers, forming a contact hole that penetrates the interlayer insulating layer to expose a sidewall of at least one of the gate spacers, forming a sacrificial gap-fill pattern filling a lower portion of the contact hole, forming a contact spacer on a sidewall of the contact hole having the sacrificial gap-fill pattern, and forming a contact filling the contact hole after removing the sacrificial gap-fill pattern.

A method for forming nanowires includes forming a plurality of epitaxial layers on a substrate, the layers including alternating material layers with high and low Ge concentration and patterning the plurality of layers to form fins. The fins are etched to form recesses in low Ge concentration layers to form pillars between high Ge concentration layers. The pillars are converted to dielectric pillars. A conformal material is formed in the recesses and on the dielectric pillars. The high Ge concentration layers are condensed to form hexagonal Ge wires with (111) facets. The (111) facets are exposed to form nanowires.

A device includes a substrate, insulation regions extending into the substrate, a first semiconductor region between the insulation regions and having a first valence band, and a second semiconductor region over and adjoining the first semiconductor region. The second semiconductor region has a compressive strain and a second valence band higher than the first valence band. The second semiconductor region includes an upper portion higher than top surfaces of the insulation regions to form a semiconductor fin, and a lower portion lower than the top surfaces of the insulation regions. The upper portion and the lower portion are intrinsic. A semiconductor cap adjoins a top surface and sidewalls of the semiconductor fin. The semiconductor cap has a third valence band lower than the second valence band.

Forming a semiconductor layer on a semiconductor substrate, a top surface of the semiconductor layer above a fin in a second region is higher than a top surface of the semiconductor layer in a first region, etching the semiconductor layer and a mask in the first region to expose a top surface of the semiconductor substrate to form a first stack, and etching the semiconductor layer and the mask in the second region to expose a top surface of the fin to form a second stack, epitaxially growing a semiconductor material on a top surface of the fin not covered by the second stack, recessing the first and second stack to expose a top surface of the semiconductor layer, a portion of the mask remains above the semiconductor layer in the first stack, top surfaces of each of the first and second stacks each are substantially flush with one another.

A static random-access memory (SRAM) cell array forming method includes the following steps. A plurality of fin structures are formed on a substrate, wherein the fin structures include a plurality of active fins and a plurality of dummy fins, each PG (pass-gate) FinFET shares at least one of the active fins with a PD (pull-down) FinFET, and at least one dummy fin is disposed between the two active fins having two adjacent pull-up FinFETs thereover in a static random-access memory cell. At least a part of the dummy fins are removed. The present invention also provides a static random-access memory (SRAM) cell array formed by said method.

A semiconductor device includes a semiconductor substrate, multiple fins formed on a front surface of the semiconductor substrate, a stress layer formed on a top surface of the fins, multiple strip-shaped gate structures formed above the stress layers, each of which extending in a direction substantially perpendicular to a direction of the fins, a contact hole etch stop layer covering the front surface of the semiconductor substrate, sidewalls of the fins, and top surfaces and sidewalls of the stress layers, a first interlayer dielectric layer over the contact hole etch stop layer, the first interlayer dielectric layer including filling voids formed therein, and a top surface of the first interlayer dielectric layer being below the top surfaces of the stress layers, a barrier liner layer over the first interlayer dielectric layer, and a second interlayer dielectric layer over the barrier liner layer and the contact hole etch stop layer.