Abstract: A novel routing rule definition for standard cells placement is disclosed. The method comprises following steps. At beginning, a statistical analysis is carried out to analyze the frequency of standard cells used in a design plane. The most frequency used types in standard cells is then used as bases for a greatest common divisor (GCD) calculation. The GCD value acquired is then as a routing pitch criterion, which is a distance of contact hole-to-contact hole, or says the standard cell width.

Abstract: A semiconductor memory device comprising a test structure is disclosed. The semiconductor device includes a plurality of memory cells, word lines, bit lines, and test pads; the word lines including a first set and a second set of word lines, connected to a first and second word line test pad, respectively; the bit lines including a first set and a second set of bit lines, connected to a first and second bit line test pad, respectively. The first set of word lines and the first set of bit lines access a first set of memory cells, the first set of word lines and the second set of bit lines access a second set of memory cells, the second set of word lines and the first set of bit lines access a third set of memory cells, and the second set of word lines and the second set of bit lines access a fourth set of memory cells.

Abstract: A high-speed built-in self-test (BIST) circuit for dynamic random access memory (DRAM) is disclosed. The circuit automatically generates a sequence of pre-defined test patterns for on-chip DRAM testing. The circuit includes two finite state machines, instead of the conventional single finite state machine. Therefore, a pipeline technique can then be applied to divide the pattern generation process into stages, leading to a higher-speed design. In addition to pipelining, protocol-based relaxation is also presented. This technique, imposing a certain protocol on the two communicating finite state machines, further relaxes the timing criticality of the design.

Abstract: A phase lock loop-based deskew buffer circuit includes a fixed delay element to delay a feedback signal and to generate a first signal from the feedback signal. A delay locked loop (DLL) generates a second signal from a reference signal and compares a phase of the feedback signal with a phase of the reference signal. The DLL is capable of keeping a relative timing of the first and second signals constant while adjusting the feedback signal to be in phase with the reference signal. This results in loop stability and cancellation of internal offset.

Abstract: A process for manufacturing flash memories is disclosed. In one embodiment, a first oxide layer is deposited over a substrate and then, a first polysilicon layer is deposited over the oxide layer. When the first polysilicon layer is etched and formed, an ONO (oxide nitride oxide) layer is deposited over the first polysilicon layer. Then, portions of the ONO layer and the first polysilicon layer are removed to form two nitride fences. A tunnel oxide layer in a conformal shape is subsequently deposited over said nitride fences, some portions of the first oxide layer, and said substrate. After depositing of tunnel oxide layer, a floating gate polysilicon layer, a second oxide layer, and a second polysilicon layer are deposited. Portions of the second polysilicon layer, the second oxide layer, the floating gate layer, and the tunnel oxide layer are, subsequently, removed. Finally, a drain well and a source well are formed in the substrate.

Abstract: A method for manufacturing a bottom storage node of a stack capacitor on a substrate is disclosed.

Abstract: A method of reading a 2-bit p-channel memory cell having a p+ drain, a p+ source, a control gate, and a floating gate formed from non-connecting hemispherical silicon grain (HSG) islands. The p+ drain and the p+ source is formed in an n-well. The method comprises: applying a positive voltage to the control gate to generate a gate induced drain leakage (GIDL) current; and measuring a drain GIDL current at the drain and a source GIDL current at the source simultaneously to determine the 2-bit data stored in the memory cell.

Abstract: A transistor structure fabricated on thin SOI is disclosed. The transistor on thin SOI has gated n+ and p+ junctions, which serve as switches turning on and off GIDL current on the surface of the junction. GIDL current will flow into the floating body and clamp its potential and can thus serve as an output node. The transistor can function as an inverter. The body (either n-well or p-well) is isolated from the n+ or P+ “GIDL switches” by a region of opposite doping type, i.e., p-base and n-base. The basic building blocks of logic circuits, e.g., NAND and NOR gates, are easily implemented with such transistors on thin SOI wafers. These new transistors on thin SOI only need contacts and metal line connections on the Vcc and Vss. The connection of fan-outs (between the output and input) can be implemented by capacitor coupling. The transistor structure and operation is useful for highperformance, low-voltage, and low-power VLSI circuits on SOI wafers.

Abstract: A two-dimensional array of single polysilicon DRAM cells is disclosed. The array comprises a plurality of DRAM cells arranged in a two-dimensional matrix, wherein each of the DRAM cells comprises: a deep n-well in a silicon substrate; a p-well within the deep n-well; a gate structure over and straddling the deep n-well and the p-well; and a n+ region within the p-well and adjacent to a sidewall of the gate structure. The array is connected together by a plurality of column bitlines, each of the column bitlines connected to the n+ regions of all of the DRAM cells that are in a common column. Further, a plurality of row wordlines are provided, each of the row wordlines connected to the gate structures of all of the DRAM cells that are in a common row.

Abstract: A current source formed in a p-type substrate is disclose. First, a deep n-well is formed within the p-type substrate and a buried n+ layer is formed within the deep n-well. Next, a p-well is formed within the deep n-well and atop the buried n+ layer. The p-well and deep n-well are then surrounded by an isolation structure that extends from the surface of the substrate to below the level of the p-well. A n+ reference structure is formed within the p-well and a gate is formed above the p-well, the gate separated from the substrate by a thin oxide layer, the gate extending over at least a portion of the n+ reference structure. Finally, a n+ output structure is formed within the p-well. An input reference current is provided to the n+ reference structure and an output current is provided by the n+ output structure.

Abstract: A method of reading a 2-bit memory cell having a drain, a source, a control gate, and a floating gate is disclosed. First, a voltage is applied to the source and drain to generate a gate induced drain leakage (GIDL) current. Next, a measurement is taken of a drain GIDL current at said drain and a source GIDL current at said source to determine the data stored in said memory cell.

Abstract: A split gate flash memory cell formed in a semiconductor substrate is disclosed. The memory cell comprises: a deep n-well formed in the substrate; a p-well formed in the deep n-well; a select gate structure formed on the p-well, the select gate structure comprising a stack of a gate oxide, a polysilicon layer, and a cap oxide; a tunnel oxide layer formed on the p-well, the tunnel oxide adjacent to the control gate structure; a floating gate formed over the select gate structure and extending over at least a portion of the tunnel oxide layer; a source formed in the p-well, the source formed adjacent to the floating gate; and a drain formed in the p-well, the drain formed adjacent to the select gate structure. The memory cell is programmed by source side channel hot electron and is erased using channel erasing to improve cycling endurance.

Abstract: A transistor formed in a semiconductor substrate having improved ESD protection. The transistor includes a gate structure formed atop of a semiconductor substrate. First and second sidewall spacers are formed on the sidewalls of the gate structure. A lightly doped source region is formed in said semiconductor substrate and substantially underneath only the first sidewall spacer. A source region is formed in said semiconductor substrate and adjacent to the first sidewall spacer and a drain region is formed in said semiconductor substrate and adjacent to the second sidewall spacer. A first ESD implant is provided that overlaps the source region and extending underneath the first sidewall spacer. A second ESD implant is formed to overlap the drain region and extending underneath the second sidewall spacer. Preferably, the ESD implants are formed using an angled ion implantation technique.

Abstract: A method for forming a T-shaped contact plug is disclosed. A first insulating layer is formed atop of a substrate. A second insulating layer is then formed atop of the first insulating layer. The first and second insulating layers are patterned and etched to form a contact opening to the substrate. A portion of the second insulating layer surrounding the contact opening is removed. Next, a barrier metal layer is formed along the walls of the contact opening and atop the second insulating layer. Then a conducting layer is formed into the contact opening and atop the barrier metal layer. Finally, a portion of the first conducting layer and barrier metal layer atop the second insulating layer is removed. This leaves a plug formed of the remaining portion of the conducting layer.

Abstract: A method of forming a planar interlayer dielectric layer over underlying structures is disclosed. First, a liner oxide layer is formed over the underlying structures. Then, a BPSG layer is formed over the liner oxide layer. The BPSG layer is polished and a cap oxide layer is formed over the BPSG layer. Finally, a nitride layer is formed over the cap oxide layer.

Abstract: A flash memory cell formed in a semiconductor substrate is disclosed. The cell includes an n-well formed within the substrate. Next, a p+ drain region is formed within the n-well. A floating gate is formed above the n-well being separated from the substrate by a thin oxide layer. The floating gate is formed adjacent to the p+ drain region. Finally, a control gate is formed above the floating gate, the floating gate and the control gate being separated by a dielectric layer. The new cell is read by measuring the GIDL current at p+/n-well junction, which is exponentially modulated by the floating gate potential (or its net charge). The new cell is programmed by band-to-band hot electron injection and is erased by F-N tunneling through the overlap area of floating gate and n-well.

Abstract: An insulating layer is formed on a semiconductor wafer. A titanium layer (Ti) is formed on the insulating layer. A titanium nitride (TiN) layer is formed on the Ti layer to act as a barrier layer. A tungsten (W) layer is deposited by using chemical vapor deposition. The tungsten layer is etched back to form a tungsten plug. While the wafer is still in the etching chamber, an in situ plans sputtering is performed to remove any fluorine contamination.

Abstract: A digitally tunable voltage reference circuit based on floating gate neuron MOSFETs and a V.sub.t referenced voltage source configuration is disclosed. The voltage reference can provide a wide range of voltage levels by biasing digital signals to the multiple inputs of the neuron MOSFET in the voltage source.

Abstract: A NOR array architecture allowing single bit, row, and column programming and erase operations is disclosed. The NOR array architecture comprises: a plurality of ETOX cells formed in a deep n-well, each of the ETOX cell having: (1) a control gate; (2) a floating gate insulated from and formed substantially underneath the control gate; (3) a p-well formed in the n-well and underneath the floating gate and the control gate; (4) a drain implant formed in the p-well adjacent to the floating gate; and (5) a source implant formed in the p-well adjacent to the floating gate. The ETOX cells are formed into a two-dimensional array including a plurality of rows and a plurality of columns. Each of the control gates of the ETOX cells in adjacent two rows sharing a common row are connected to a row wordline. Each of the source implants of the ETOX cells sharing a common row are connected to a row sourceline. Each of the drain implants of the ETOX cells sharing a common column are connected to a column bitline.

Abstract: A method of forming a bottom storage node of a DRAM capacitor over a contact plug is disclosed. The method comprises the steps of: depositing an oxide layer over the contact plug; etching the oxide layer using a first photoresist layer having with a first masking pattern, the first masking pattern allowing the removal of the oxide layer over the contact plug; depositing a polysilicon layer over the oxide layer and in electrical contact with the contact plug; forming a second photoresist layer having a second masking pattern onto the polysilicon layer, the second masking pattern being substantially similar to the first masking pattern, but rotated by a predetermined angle; and etching the polysilicon layer in accordance with the second photoresist layer until the oxide layer is reached.