Lsi Logic Corporation

Lsi Corporation

Subramanian Ramesh

Ruggero Castagnetti

Ramnath Venkatraman

Yauh-Ching Liu

Gary K Giust

Carl Anthony Monzel, Iii

Gary Scott Delp

Michael Norris Dillon

Scott Allen Peterson

Robert C Armstrong

David A Gardner

Franklin Duan

Ramnath Ventatraman

Rugger Castagnetti

A basic cell circuit architecture having plurality of cells with fixed transistors configurable for the formation of logic devices and single and dual port memory devices within a structured ASIC is provided. Different configurations of ensuing integrated circuits are achieved by forming variable interconnect layers above the fixed structures. The circuit architecture can achieve interconnection of transistors within a single cell or across multiple cells. The interconnection can be configured to form basic logic gates as well as more complex digital and analog subsystems. In addition, each cell contains a layout of transistors that can be variably coupled to achieve a memory device, such as a SRAM device. By having the capability of forming a logic circuit element, a memory device, or both, the circuit architecture is both memory-centric and logic-centric, and more fully adaptable to modern-day SoCs.

Basic cell architecture for structured ASICs

An improved integrated circuit cell architecture is provided for configurability between a memory cell or logic elements. The cell architecture is configured on variable layers above a first layer of metal, with the first layer of metal and layers therebelow reserved as fixed layers. By coupling a maximum of two layout cells together, a single-port or dual-port memory cell is realized. Likewise, by interconnecting transistors within a single cell or transistors among two or more cells, a logic device is realized. Within each cell, the bit lines are arranged on a layer separate from the wordlines, and extend orthogonal to each other.

Integrated circuit cell architecture configurable for memory or logic elements

A basic cell circuit architecture having plurality of cells with fixed transistors configurable for the formation of logic devices and/or single/dual port memory devices within a structured ASIC is provided. Different configurations of ensuing integrated circuits are achieved by forming variable interconnect layers above the fixed structures. The circuit architecture can achieve interconnection of transistors within a single cell and/or across multiple cells. The interconnection can be configured to form basic logic gates as well as more complex digital and analog subsystems. In addition, each cell contains a layout of transistors that can be variably coupled to achieve a memory device, such as a SRAM device. By having the capability of forming either a logic circuit element, a memory device, or both, the circuit architecture is both memory-centric and logic-centric, and more fully adaptable to modern-day SoCs.

Basic cell architecture for structured application-specific integrated circuits

Disclosed is a method and device for providing fast-response One-Time-Programmable (OTP) memory based on SRAM memory technology and the inherent breakdown characteristics of a MOS transistor. Each memory cell of an SRAM memory cell circuit is connected to a programming circuit. The programming circuit is comprised of two groups of MOS transistors connected to the storage nodes (SN and SNB, where SNB is the complementary value of SN) of the two cross-coupled inverters of the SRAM memory circuit. A desired data set is loaded into the circuit and then is burned-in by applying and repeatedly cycling a 'burn-in' voltage across the source and drain of the MOS transistors of the programming circuit that approaches the ON STATE trigger voltage of the characteristic bipolar junction transistor contained within the MOS transistors. Upon repeated cycling of the source-to-drain voltage, the targeted MOS transistor within the programming circuit breaks down and shorts across the gate, drain, and/or source of the transistor. When the system is returned to normal operation, the programming circuits will be connected to ground, Vdd or Vss and one of the two nodes of the SRAM cell circuit will be shorted through the programming circuit to ground, Vdd or Vss, thus, forcing a retention of the programmed data state.

SRAM based one-time-programmable memory

The present invention provides a system and method for designing and modularly expanding multiport bitcells. A modular design approach is described that reduces the complexity of designing multiport bitcells while complying with DFM rules across various semiconductor fabrication providers. The modular design may be parsed into modules such as read port modules, write port modules, and pull-up modules that may be easily interconnected to build a multiport bitcell. These modules may also be independently sized and assembled to achieve desired read margins, write margins, static noise margins as well as read access times and write times.

Modular design of multiport memory bitcells

The present invention provides a method and apparatus for reconfiguring a memory array. Aspects of the present invention include fabricating the memory array as at least one row of single-port cells up to a first metal layer. A split word line having first and second word lines is coupled to the single-port cells in each row, wherein the first word line is patterned in the first metal layer, and the second word line is patterned in a second metal layer. The split word line is further coupled to a spacer cell in the row. The method and apparatus further include programming the memory array into custom configurations based on whether the first and second word lines are connected over the spacer cell, or whether the first and second word lines are left unconnected.

Design and use of a spacer cell to support reconfigurable memories

A memory cell architecture is provided herein for increasing memory speed, performance and robustness within a highly compact memory cell layout. Though only a few embodiments are provided herein, a feature common to all embodiments includes a novel means for sharing one or more contact structures between vertically adjacent memory cells. In particular, one or more contact structures may be shared unequally between two vertically adjacent memory cells for reducing a vertical dimension, or length, of the memory cell. Other features are disclosed for producing the highly compact memory cell layout. The various features of the present invention may be combined to produce high-performance, high-density memory arrays.

Memory cell architecture

An improved memory cell architecture is provided herein for reducing, or altogether eliminating, chip-level routing congestion in System-on-Chip environments. Though only a few embodiments are provided herein, features common to the described embodiments include: the formation of bitlines in a lower-level metallization layer of the memory array, and the use of word lines and ground supply lines, both formed in inter-level metallization layer(s) of the memory array, for effective shielding of the bitlines against routing signals in the chip-level routing layer.

Memory cell architecture for reduced routing congestion

Test structures are provided for accurately quantifying shared contact resistance. The test structures are built based upon an actual memory cell, which is self-aligning to allow shared contact chains through an array of test cells. A main array of test cells is built to provide a chain of shared contact resistance. Using the main array of test cells, a resistance in the shared contact chain may be measured. Supplemental arrays of test cells is built to provide a chain of poly side resistance, a chain of island side resistance, a chain of island connection line resistance, and a chain of poly connection resistance. A tester measures resistance using the test structures and uses the values to accurately determine shared contact resistance.

Method and apparatus for characterizing shared contacts in high-density SRAM cell design

A self-aligning memory cell design is provided to allow testing of transistors in every cell of a memory circuit. A test array of these cells is fabricated with contact pads in each cell for specific components in the cell. Then, metal lines are provided to couple the contact pads in the test array. The whole test array is then probed via these metal lines. Tests may then be performed to detect random and systematic transistor degradation electrically for all cells in the circuit. Different components in the memory design may be tested by providing contact pads for the components of interest and providing metal lines coupling the contact pads.

Method and architecture for detecting random and systematic transistor degradation for transistor reliability evaluation in high-density memory

A method and system for reconfiguring a memory array is disclosed. Initially, the cells in the memory array are patterned with substantially the same structure up to a predefined layer. Thereafter, the memory array is reconfigured by patterning the cells above the predefined layer such that a first plurality of cells function as memory cells, and a second plurality of cells are patterned as a dummy row or column that function as a breakpoint for the memory array.

Reconfigurable memory arrays

The present invention provides a method and apparatus for providing dual-port capability to an SRAM array. The internal nodes of two single-port memory cells are connected to each other through metal-layer programming to form a dual-port memory cell. In a preferred embodiment, a split word line design is used for each single-port memory cell, to facilitate dual-port memory access while minimizing the need for IC layout space. An additional benefit of the present invention is that it allows &ldquo;slices&rdquo; of a memory array to be converted into dual-port memory, so as to allow both single-port and dual-port memory cells in the same memory array.

Metal-programmable single-port SRAM array for dual-port functionality

Provided are a self-aligned semiconductor fuse structure, a method of making such a fuse structure, and apparatuses incorporating such a fuse structure. The fuse break point, that point at which the electrical link of which the fuse is part is severed by a laser beam, is self-aligned by the use of photolithographically patterned anti-reflective dielectric coatings. The self-alignment allows the size location of the break point to be less sensitive to the laser beam size and alignment. This has several advantages including allowing photolithographic control and effective size reduction of the laser spot irradiating the fuse material and surrounding structure. This permits reduced fuse pitch, increasing density and the efficiency of use of chip area, and results in reduced thermal exposure, which causes less damage to chip. In addition, laser alignment is less critical and therefore less timely, which increases throughput in fabrication.

Growing community of inventors

Cupertino, CA, United States of America

Subramanian Ramesh