Abstract: The disclosure relates to a photovoltaic solar array which is provided with a matrix having spherical photovoltaic diode particles embedded therein in an ordered array, the P-type region of each particle extending to one matrix surface and the N-type region of each particle extending to an opposed matrix surface. Backside metallization is disposed on the matrix backside surface to interconnect the particles extending thereto and frontside conductors are provided on the opposing matrix surface to interconnect the particles extending thereto. The matrix includes two portions, the first portion being a layer extending to the frontside formed of a clear glass. The second portion of the matrix is, in effect, two layers, one disposed at the P-N junctions of the particles being a lead base glass for junction passivation, this layer being overcoated with a reflective layer to provide additional reflectivity of light entering the matrix onto the particles. This increases the amount of light impinging on the particles.

Abstract: A high-voltage CMOS process, providing (for 5 micron geometries) both field thresholds and junction breakdowns in excess of 20 volts, wherein only one channel stop implant is used. A double-well process in an epitaxial structure is used. Phosphorus is preferably used as the dopant for the N-tank, and boron is used for the blanket channel stop implant. The boron tends to leach into oxide, and the phosphorus tends to accumulate at the surface, and a high field threshold is achieved over both PMOS and NMOS regions.

Abstract: A memory decoder wherein a power-up device is interposed between a NOR decoder and ground (VSS), rather than between the decoder and VDD. Preferably the signal to the power-up transistor is itself decoded, so that the power-consuming NOR circuits are inactive over a majority of the chip, even during power-up conditions.

Abstract: In stacked CMOS, a single gate in first level polycrystalline silicon is used to address both an N-channel device in the substrate and an overlaid p-channel device. The p-channel device has self-aligned source and drain regions formed by diffusing a dopant from doped regions underlying them. The doped regions are formed by planarizing a doped insulating layer, and etching the doped layer back to the upper level of the gate prior to deposition of a second polysilicon layer.

Abstract: A memory including various selectively configurable peripherals which provide on-chip low-level control features and a configuration RAM storing bits which both provide unclocked full logic-level outputs to control the selectively configurable peripherals and can also be accessed and read out. That is, each cell in the configuration RAM has two output modes: a digital continuous output, which is provided as a continuous control signal to various peripheral circuits and a selectable analog output which is used to read the information stored in the configuration RAM.

Abstract: A byte-wide memory with column redundancy. The redundant columns can each be substituted for any column in the half-array, without regard to which bit position the defective column relates to. Fuses store the address information of the defective columns, and when a match between the externally received column address and the stored defective-column-address is found, the sense amplifier for the bit position which contains that defective column is disabled, and the output of the redundant column (selected by whichever word line is activated) is multiplexed into the I-O buss. Thus, before the row address signal has even been decoded, the defective column has been disabled and one of the redundant columns has effectively been substituted. This configuration means that it is not necessary to have one redundant column for every bit position, but each redundant column can substitute for a defective column in any bit position, and more than one defective column in a single bit position can each be replaced.

Abstract: Using the present invention, a gate for a MESFET may be fabricated having a minimum gate length while having a low resistance gate. In addition, the present invention provides a method for forming a gate and gate recess which are perfectly aligned which is the optimal structure for high frequency power MESFETs. A two layer masking layer is fabricated having a first layer which may be etched uniformly and a second layer of lithographic material which may be photolithographic material such as AZ resist. A gate opening is patterned in the photoresist material and a metal such as gold is deposited by evaporation from acute angles on opposite sides of the gate opening in the resist. The deposited metal serves as a mask which covers all but a very small portion of the opening in the photoresist. The silicon nitride layer is then etched to form a gate opening and gate recess.

Abstract: Using a method according to one embodiment of the present invention, an EPROM array may be fabricated providing a dense EPROM array.First the polycrystalline silicon floating gates are formed and partially patterned on the surface of a substrate. A thin thermally grown oxide layer is then formed over the entire array. The source/drain regions are then implanted through the thin silicon dioxide layer into the substrate. Next a thick silicon dioxide layer is deposited by chemical vapor deposition on the surface of the array. The surface of the array is then coated with photoresist which, because of its nature, provides a planarized surface on the top layer of photoresist. The photoresist and the silicon dioxide layer are then etched using an etching process which provides an etching ratio of 1 to 1 between photoresist and silicon dioxide. The photoresist is completely etched away thus leaving the planarized silicon dioxide surface.

Abstract: N-channel devices are fabricated with lightly doped drain/source extensions in a CMOS process, without the requirement of an extra mask level. A merged mask technique uses an oversized version of the N-channel gates, expanded by two alignment tolerances per side, combined with the regular N+ source/drain mask. The oversized gate photoresist prevents the heavy N+ source/drain implant from counterdoping the previously introduced lightly doped drain blanket implant. In the P-channel regions the N-type LDD extensions are counterdoped by the regular P+ source/drain implant. This high-voltage process provides 20 V parts with 4 micron geometries, scalable to other voltages.

Abstract: A MOS read only memory, or ROM, is formed by a process compatible with standard P-channel or N-channel metal or silicon gate manufacturing methods. The ROM is programmed either after the protective nitride layer has been applied and patterned, usually the last step in the slice processing method before electrical testing of the devices, or after the electrical testing of the devices. All potential MOS transistors in the ROM array are initially at a logic "0" or a logic "1". An electron beam slice printing machine is used to program the selected transistors in the ROM array to change their logic state by exposing the gates of the selected transistors to an electron beam. The gates to be exposed are predetermined by a coding on a magnetic tape which corresponds to the desired ROM code. No electron beam mask is necessary since the beam only exposes in selected areas.

Abstract: Using a process in accordance with the teachings of this invention, an integrated circuit may be fabricated providing refractory metal silicide layers, such as TiSi.sub.2, of differing thicknesses to provide optimal reductions in the sheet resistances of the regions in which refractory metal silicide layers are formed. In one embodiment of the present invention a field effect transistor having a polycrystalline silicon gate is fabricated to provide a gate having optimally minimized sheet resistance and source and drain regions having TiSi.sub.2 layers of the appropriate thickness to avoid punch-through leakage problems.

Abstract: A static random access memory wherein positive feedback is used in the bit line loads. The output of the first sense amplifier stage is fedback to the gates of depletion-made bit line load transistors, to provide positive feedback during the read or write operation. That is, since one of the complementary bit lines which the accessed memory cell is attempting to pull down sees a load impedance which gradually becomes higher and higher, the memory cell can pull down this bit line more rapidly. To accomplish this with stability, the first sense amplifier stage has less than unity open loop gain, and a succeeding sense amplifier stage is therefore used for further amplification.

Abstract: A sidewall-nitride isolation technology refines process control over lateral oxide encroachment by preventing any thinning of the nitride moat-masking layer during the nitride etch step which clears the sidewall nitride layer from the bottom of the etched recesses in silicon. This is done by initially patterning the moat regions in an oxide/nitride/oxide stack, rather than the nitride/oxide stack of the prior art.

Abstract: A stacked CMOS structure is disclosed which uses buried N++ source and drain for the non-self-aligned bulk N-channel driver devices together with an oversized polygate on which a non-self aligned P-channel load device is made from a second layer of poly or recrystallized poly. The non-self aligned pair of stacked devices provides increased density of devices per unit area with a simple process at the cost of increased gate to source and gate to drain parasitic capicitances.

Abstract: A CMOS device configuration in which a complete CMOS inverter is contained in the space normally required for a single NMOS transistor of equivalent geometry. A first polysilicon layer of normal thickness and N+ doping is used for the N channel gate, and a second polysilicon layer is deposited conformally over the oxide which encapsulates the first polysilicon layer. The second polysilicon layer is thin and doped p-type. The second layer is only lightly doped initially, and is then doped more heavily by a low-energy implantation. The portions of the second poly layer which are adjacent to the sidewalls of the gate level in first poly will be shielded from the heavy implantation, and will therefore provide relatively lightly doped p-type channel regions, to form a pair of PMOS polysilicon transistors addressed by the N+ first poly gate electrode. Preferably the channel doping of these polysilicon transistors is at least 10.sup.17.

Abstract: A MESFET is fabricated using a self-aligned gate process. This process uses a vertical (anisotropic) etch to self-align the gate and source/drain. The vertical etch, in conjunction with a two-level insulator, creates a barrier between the gate and source/drain, so that when metal is deposited and reacted, and any excess removed, the gate is selfaligned with the source/drain, and contacts to the source/drain and gate are well isolated. The alignment obtained by this process is advantageous in that series channel resistance is reduced, and a more compact structure is attained for improvement in packing density.

Abstract: Pole encoding of a linear predictive all-pole model of speech is accomplished by first finding poles up to the number required for good prediction (e.g., ten). These poles are extracted from the LPC predictor polynomial, using, e.g., a slightly modified Bairstow method. Those poles having a sufficiently narrow bandwidth (i.e., those sufficiently near the unit circle) are separately encoded, since these poles generally correspond to perceptually important formants. The remaining poles are lumped together to form a residual polynomial. The residual polynomial is then transformed to produce reflection coefficients, and all reflection coefficients above the first two are discarded. This provides an efficient spectral-shaping polynomial of a reduced degree. Thus, pole encoding is made possible using a reduced and adaptively varied bit rate.

Abstract: Hydrogen annealing permits deposition of good quality polysilicon atop TiO.sub.2. Hydrogen annealing of TiO.sub.2 prevents the tremendous hydrogen affinity of as-deposited TiO.sub.2 from disrupting process reactions during deposition of polysilicon.