Abstract: A method of erasing memory cells in a sector of a flash programmable memory device, the sector having a plurality of word lines and a plurality of memory cells along each of the word lines, each of the cells in the sector having a source region common to all cells in the sector, the method comprising a first step of erasing the memory cells in the sector simultaneously, then reading a first cell along a first word line to determine if the first cell is under-erased. Responsive to the first cell being erased, a second cell along the first word line is read to determine if the second cell is under-erased. Responsive to the second cell being under-erased, a negative first voltage is applied to the first word line, a positive second voltage is applied to the common source of the cells in the sector, and a positive third voltage is applied to the plurality of word lines except the first word line.

Abstract: A method of erasing a sector for a flash memory device, the sector having a plurality of word lines with each word line having a plurality of memory cells therealong, each cell having a source region, comprises various steps. First, all of the memory cells in the sector are programmed. The memory cells are then simultaneously erased by applying a first voltage to the sources and a second voltage to the word lines. Subsequently, a first cell along a first word line is read to determine if the first cell is under-erased. Responsive to the first cell being under-erased, the first voltage is applied to the source region of the first cell and the second voltage is applied to the first word line while a third voltage is applied to the plurality of word lines except the first word line, the third voltage being higher than the second voltage.

Abstract: A method and structure for a programmable read-only memory comprises a thin gate oxide over a source region and a thick gate oxide over the drain region. A semiconductor substrate is lightly doped and has regions of thin sacrificial oxide overlying what will become the transistor channel, source, and drain, and thick field oxide. The region that will become the transistor source is protected with photoresist, and the drain region and a portion of the channel is doped, for example, with boron. The resist and sacrificial oxide is stripped, and gate oxide is formed from the exposed silicon substrate. The more heavily doped drain and channel regions oxidize at a faster rate than the lightly doped source region, and thus the gate oxide formed is thicker. Floating and control gates are formed over the channel region, covering both the thicker and thinner gate oxide.

Abstract: A flash memory array comprises a primary row line and a redundant row line each having memory cells therealong. A method of accessing the flash memory array comprises preprogramming all said memory cells. Next, all memory cells are erased simultaneously. Subsequently, all memory cells along the primary row line are programmed and the cells along the redundant row line are selectively programmed. The primary row line is bypassed during any read cycle.

Abstract: A flash programmable memory device comprises first and second row lines each having memory elements therealong with the second conductive line functionally replacing the first conductive line. The memory device further includes a first program circuit for programming the memory elements along the first row line, and a second program circuit for programming memory elements along the second row line. A read circuit bypasses the first conductive line during all read cycles and reads the memory elements along the second row line.

Abstract: A programming method for flash erasable programmable memory devices (flash EPROMs) comprises a first step of erasing the array of cells, then applying a control gate voltage to access a number of control gates. Any number of control gates can be accessed, but accessing four or eight control gates may have advantages. Regardless of the number of control gates accessed, a digit line voltage is applied to access one of the digit lines, which activates a number of cells. The digit line voltage is sensed for a voltage drop, which indicates the presence of at least one over-erased activated cell. If a digit line voltage drop is detected, a sense voltage is applied to each of the activated cells to determine which is over-erased. A heal voltage is applied to the over-erased cell for an interval of time to store electrons on the floating gate of the over-erased cell.

Abstract: A semiconductor processing method of providing dopant impurity into a semiconductor substrate includes: a) providing a semiconductor substrate, the substrate comprising a first bulk region having a blanket doping of a first conductivity type dopant, the substrate comprising a second bulk region having a blanket doping of a second conductivity type dopant; b) defining field oxide regions and active area regions in each of the first and second bulk substrate regions; c) in the same masking step, masking active area regions of the first bulk substrate region while leaving field oxide regions of the first bulk substrate region unmasked and masking field oxide regions of the second bulk substrate region while leaving select active area regions of the second bulk substrate region unmasked; and d) in the same ion implanting step, ion implanting first conductivity type impurity through the unmasked portions of the first and second bulk substrate regions to simultaneously form channel stop isolation implants beneath t

Abstract: A method and structure for a programmable read-only memory comprises a thin gate oxide over a source region and a thick gate oxide over the drain region. A semiconductor substrate is lightly doped and has regions of thin sacrificial oxide overlying what will become the transistor channel, source, and drain, and thick field oxide. The region that will become the transistor source is protected with photoresist, and the drain region and a portion of the channel is doped, for example, with boron. The resist and sacrificial oxide is stripped, and gate oxide is formed from the exposed silicon substrate. The more heavily doped drain and channel regions oxidize at a faster rate than the lightly doped source region, and thus the gate oxide formed is thicker. Floating and control gates are formed over the channel region, covering both the thicker and thinner gate oxide.

Abstract: A programming method for flash erasable programmable memory devices (flash EPROMs) comprises a first step of erasing the array of cells, then applying a control gate voltage to access a number of control gates. Any number of control gates can be accessed, but accessing four or eight control gates may have advantages. Regardless of the number of control gates accessed, a digit line voltage is applied to access one of the digit lines, which activates a number of cells. The digit line voltage is sensed for a voltage drop, which indicates the presence of at least one over-erased activated cell. If a digit line voltage drop is detected, a sense voltage is applied to each of the activated cells to determine which is over-erased. A heal voltage is applied to the over-erased cell for an interval of time to store electrons on the floating gate of the over-erased cell.

Abstract: A fabrication method for a PROM cell allows improved, lower voltage programming and reduced leakage of the charge from the floating gate to the substrate (channel) region. The inventive cell uses a thin gate oxide layer along with a floating gate which is lightly doped except on one edge. This edge, for example near the drain region, is heavily doped with an angled implant. The thin gate oxide functions as thick oxide under the lightly doped region, thereby preventing the leakage and high coupling between the substrate and floating gate of a conventional thin oxide layer. The thin oxide under the heavily doped areas of the floating gate functions as thin oxide, thereby allowing improved, lower voltage programming.

Abstract: An array of finned memory cell capacitors on a semiconductor substrate includes: a) an array of electrically insulated word lines atop a semiconductor substrate; b) first and second active regions adjacent the word lines; c) capacitor storage nodes electrically connecting with the first active regions, individual capacitor storage nodes including: i) a layer of first conductive material conductively connecting with a first active region, the layer of first conductive material having opposed outer lateral edges, and ii) a layer of conductively doped storage node polysilicon overlying and conductively connecting with the layer of first conductive material, the storage node polysilicon projecting laterally outward beyond the outer lateral edges of the first conductive material to define opposing storage node capacitor fins projecting laterally above adjacent word lines; d) a layer of capacitor dielectric electrically connecting with the storage node capacitor fins; e) a layer of electrically conductive cell poly

Abstract: A process and structure for improving the conductive capacity of a polycrystalline silicon (poly) structure, such as a bit line. The inventive process allows for the formation of a refractory metal silicide layer on the top and sidewalls of a poly structure, thereby increasing the conductive capacity. To form the titanium silicide layer over the poly feature, the refractory metal is sputtered on the poly, which reacts to form the refractory metal silicide. A second embodiment is described whereby an isotropic etch of the poly feature slopes the sidewalls; then, the refractory metal is sputtered onto the polycrystalline silicon. This allows for the formation of a thicker layer of refractory metal silicide on the sidewalls, thereby further increasing the conductive capacitance of the poly structure. Suggested refractory metals include titanium, cobalt, tungsten, and tantalum.

Abstract: The invention features a circuit wherein a serially connected transistor and anti-fuse element are biased for current to flow in a first direction or the current flows in the first direction during a programming operation and biased for a current to flow in a second direction or current is flowing in the second direction during a normal circuit operation, such as a read operation, wherein the first and second directions are opposite of one another. Thus the invention facilitates the use of a low programming potential while minimizing leakage current.

Abstract: Disclosed is fabricating a semiconductor wafer to form a memory array and peripheral area, the array comprising nonvolatile memory devices employing floating gate transistors and the peripheral area comprising CMOS transistors. A first layer of conductive material is applied atop insulating layers. A dielectric layer is applied atop the first conductive layer for use in floating gate transistors within the array. The dielectric layer and first conductive material are etched from the peripheral area, leaving patterned dielectric material and first conductive material in the array. A second layer of conductive material is applied atop the wafer to cover the peripheral area and dielectric layer of the array. The conductive and dielectric materials of the array are patterned and etched separately from the patterning and etching of conductive material of each of the first and second conductivity type CMOS transistors of the peripheral area.

Abstract: A programmable read-only memory device and method of fabrication are disclosed having an antifuse in the drain node of a field effect transistor. Programming is accomplished by imposing a high voltage on the transistor drain and gate which causes the antifuse to be a closed circuit; otherwise, the transistor appears as an open circuit. Locating the antifuse in the drain node as opposed to the source node avoids problems of source reverse bias.

Abstract: A process and structure for an electrically erasable programmable read-only memory is described. The PROM is manufactured with four poly layers, the poly layers forming the floating and control gates, a structure coupling isolated source areas, and the fourth layer forming the digit lines. The inventive structure allows for a self-aligned poly source line which removes the need for an etch of the field oxide which is required in conventional EEPROM designs, which are known to be difficult to control. The poly digit line is also self-aligned and has a large margin of misalignment error in an etch to expose the drain region of the substrate.

Abstract: A dynamic random access memory (DRAM) storage cell having a storage contact capacitor comprising a tungsten and TiN storage node capacitor plate and the method for fabricating the same. At least a portion of the storage node capacitor plate is formed vertically in the DRAM. The TiN is controllably etched to increase the area of the storage node capacitor plate. An upper poly layer functions as the cell plate and is insulated from the storage node capacitor plate by a dielectric layer.

Abstract: Described is an E.sup.2 PROM design comprising a channel region and a floating gate comprising P-type polycrystalline silicon. The work function difference between P-type material effectively increases the threshold voltage of the transistor. This alleviates the need for a boron V.sub.T adjust implant. Implants of material such as boron to set the threshold voltage are known to correlate with problems such as implant ionization and junction (avalanche) breakdown. These two undesired effects can be decreased or eliminated in devices comprising the invention. An optional phosphorous implant into the substrate would allow the lowering of V.sub.T to a desired level.

Abstract: The present invention comprises a method to program antifuse elements in integrated circuits, such as programmable read-only memory (PROM) or option selections/redundancy repair on dynamic random access memories (DRAMs) by utilizing the phenomenon of transistor snap-back. Multiple programming pulses are applied to an NMOS transistor which provides access to the desired antifuse element. The first pulses applied ruptures the antifuse element causing it so become a resistive short. The second programming pulses cause the access NMOS transistor to go into snap-back thus allowing a surge of current to flow through the resistively shorted antifuse thereby lowering the resistance of the shorted antifuse element substantially allowing for less power consumption and higher reliability of the permanently programmed element.

Abstract: The present invention comprises a method to program programmable structures in integrated circuits, such as programmable read-only memory (PROM), erasable programmable read-only memory family (EPROMS, EEPROMS, etc.) or option selections/redundancy repair on dynamic random access memories (DRAMs) by utilizing multiple programming pulses that vary in pulse width and amplitude. Multiple programming pulses (or a pulse train) of the present invention are applied to the desired structure to be programmed, whether that structure be a fuse in a bipolar PROM, an antifuse element in a CMOS PROM, an option selecting element (fuse or antifuse) in a DRAM, the programmable gates of the EPROM family (EEPROMs, Flash EEPROMS, etc.). The pulse train applied to the programmable structure may be as few as one pulse, having sufficient amplitude and pulse width, or the pulse train may be a plurality of pulses varying in pulse amplitude or pulse width.