Abstract:
Middle-of-line (MOL) metal resistor temperature sensors for localized temperature sensing of active semiconductor areas in integrated circuits (ICs) are disclosed. One or more metal resistors are fabricated in a MOL layer in the IC adjacent to an active semiconductor area to sense ambient temperature in the adjacent active semiconductor area. Voltage of the metal resistor will change as a function of ambient temperature of the metal resistor, which can be sensed to measure the ambient temperature around devices in the active semiconductor layer adjacent to the metal resistor. By fabricating a metal resistor in the MOL layer, the metal resistor can be localized adjacent and close to semiconductor devices to more accurately sense ambient temperature of the semiconductor devices. The same fabrication processes used to create contacts in the MOL layer can be used to fabricate the metal resistor.
Abstract:
Aspects disclosed in the detailed description include nanowire channel structures of continuously stacked heterogeneous nanowires for complementary metal oxide semiconductor (CMOS) devices. Each of the nanowires has a top end portion and a bottom end portion that are narrower than a central portion. Furthermore, vertically adjacent nanowires are interconnected at the narrower top end portions and bottom end portions. This allows for connectivity between stacked nanowires and for having separation areas between vertically adjacent heterogeneous nanowires. Having the separation areas allows for gate material to be disposed over a large area of the heterogeneous nanowires and, therefore, provides strong gate control, a shorter nanowire channel structure, low parallel plate parasitic capacitance, and low parasitic channel capacitance. Having the nanowires be heterogeneous, i.e., fabricated using materials of different etching sensitivity, facilitates forming the particular cross section of the nanowires, thus eliminating the use of sacrificial masks/layers to form the heterogeneous nanowires.
Abstract:
A semiconductor device includes a diffusion area, a gate structure coupled to the diffusion area, and a dummy gate structure coupled to the diffusion area. The gate structure extends a first distance beyond the diffusion area, and the dummy gate structure extends a second distance beyond the diffusion area.
Abstract:
Methods for designing fin-based field effect transistors (FinFETs) are disclosed. In one embodiment, an initial FinFET design is evaluated to ascertain the space between fins (i.e., the "fin pitch"). Additionally, the spacing between interconnect metal modules (i.e., the "metal pitch") is ascertained. A ratio of metal pitch to fin pitch is established. From this initial ratio, isotropically scaled sizes are considered along with anisotropically scaled sizes. The variously scaled sizes are compared to design criteria to see what new size best fits the design criteria.
Abstract:
An integrated circuit (IC) package is described. The IC package includes a first die having a first power delivery network on the first die. The IC package also includes a second die having a second power delivery network on the second die. The first die is stacked on the second die. The IC package further includes package voltage regulators integrated with and coupled to the first die and/or the second die within a package core of the integrated circuit package.
Abstract:
Aspects disclosed include static random access memory (SRAM) arrays having substantially constant operational yields across multiple modes of operation. In one aspect, a method of designing SRAM arrays with multiple modes operation is provided. The method includes determining performance characteristics associated with each mode of operation. SRAM bit cells configured to operate in each mode of operation are provided to the SRAM array. SRAM bit cells are biased to operate in a mode of operation using dynamic adaptive assist techniques, wherein the SRAM bit cells achieve a substantially constant operational yield across the modes. The SRAM bit cells have a corresponding type, wherein the number of SRAM bit cell types in the method is less than the number of modes of operation. Thus, each SRAM array may achieve a particular mode of operation without requiring a separate SRAM bit cell type for each mode, thereby reducing costs.
Abstract:
Static random access memory (SRAM) bit cells with wordlines on separate metal layers for increased performance are disclosed. In one aspect, an SRAM bit cell is disclosed employing a write wordline in a second metal layer, a first read wordline in a third metal layer, and a second read wordline in a fourth metal layer. Employing wordlines in separate metal layers allows wordlines to have increased widths, which decrease wordline resistance, decrease access time, and increase performance of the SRAM bit cell. To employ wordlines in separate metal layers, multiple tracks in a first metal layer are employed. To couple read wordlines to the tracks to communicate with SRAM bit cell transistors, landing pads are disposed on corresponding tracks disposed in the first metal layer. Landing pads corresponding to the write wordline are placed on corresponding tracks disposed in the first metal layer.
Abstract:
Aspects describing modified self-aligned quadruple patterning (SAQP) processes using cut pattern masks to fabricate integrated circuit (IC) cells with reduced area are disclosed. In one aspect, a modified SAQP process includes disposing multiple mandrels. First spacers are disposed on either side of each mandrel, and second spacers are disposed on either side of each first spacer. A cut pattern mask is disposed over the second spacers and includes openings that expose second spacers corresponding to locations in which voltage rails are to be disposed. The voltage rails are formed by removing the second spacers exposed by the openings in the cut pattern mask, and disposing the voltage rails in the corresponding locations left vacant by removing the second spacers. Routing lines are disposed over routing tracks formed between each set of the remaining second spacers to allow for interconnecting of active devices formed in the IC cell.
Abstract:
Tie-off structures for middle-of-line (MOL) manufactured integrated circuits, and related methods are disclosed. As a non-limiting example, the tie-off structure may be used to tie-off a drain or source of a transistor to the gate of the transistor, such as provided in a dummy gate used for isolation purposes. In this regard in one aspect, a MOL stack is provided that includes a metal gate connection that is coupled to a metal layer through metal structure disposed in and above a dielectric layer above a gate associated with the metal gate connection. By coupling the metal gate connection to the metal layer, the gate of a transistor may be coupled or "tied-off" to a source or drain element of the transistor. This may avoid the need to etch the metal gate connection provided below the dielectric layer to provide sufficient connectivity between the metal layer and the metal gate connection.
Abstract:
Middle-of-line (MOL) manufactured integrated circuits (ICs) employing local interconnects of metal lines using an elongated via are disclosed. Related methods are also disclosed. In particular, different metal lines in a metal layer may need to be electrically interconnected during a MOL process for an IC. In this regard, to allow for metal lines to be interconnected without providing such interconnections above the metal lines that may be difficult to provide in a printing process for example, in an exemplary aspect, an elongated or expanded via(s) is provided in a MOL layer in an IC. The elongated via is provided in the MOL layer below the metal layer in the MOL layer and extended across two or more adjacent metal layers in the metal layer of the MOL layer. Moving the interconnections above the MOL layer can simplify the manufacturing of ICs, particularly at low nanometer (nm) node sizes.