摘要:
A split-polysilicon CMOS DRAM process incorporating self-aligned silicidation of the cell plate, transistor gates and N+ regions with a minimum of additional processing steps. By employing a light oxidation step to protect the P-channel transistor sidewall gates from silicidation during a subsequent processing step, the process avoids the problems that may be created by the double etching of the field oxide and active area regions that has heretofore been required for self-aligned silidation utilizing a split-polysilicon CMOS process. A protective nitride layer is used to prevent oxidation on those regions which are to be silicided. When this improved process is utilized for DRAM fabrication, the protective nitride layer may also be utilized as the cell dielectric. Although this process precludes the silicidation of the sources and drains of P-channel transistors, silicidation of other important regions is accomplished with very few steps required beyond those required for the basic split-polysilicon CMOS process without self-aligned silicidation of conductive regions.
摘要:
A process for creating a DRAM array having feature widths that transcend the resolution limit of the employed photolithographic process using only five photomasking steps. The process involves the following steps: creation of a half-pitch hard-material mask that is used to etch a series of equidistanty-spaced isolation trenches in a silicon substrate; filling the isolation trenches with insulative material; creation of a hard-material mask consisting of strips that are 1-1/2F in width, separated by spaces that are 1/2F in width, that is used to etch a matrix of storage trenches; angled implantation of a N-type impurity in the storage trench walls; another anisotropic etch to deepen the storage trenches; deposition of a capacitor dielectric layer; deposition of a protective polysilicon layer on top of the dielectric layer; removal of the dielectric layer and the protective polysilicon layer at the bottom of each storage trench with a further anisotropic etch; filling the storage trenches with in-situ-doped polysilicon; planarization down to the substrate level; creation of an access gate on opposite sides of each storage trench, in addition to wordlines which interconnect gates within array columns by anisotropically etching a conformal conductive layer that has been deposited on top of oxide-silicon mesas that run perpendicular to the isolation trenches and are centered between the rows of storage trenches, the oxide-silicon mesas having been created with an etch using a photoresist mask consisting of a series of parallel strips that have been laid down with minimum feature and space width, then plasma etched to 3/4F; creation of source and drains with an N-type implant; and anisotropically etching the metal layer to create bitlines along the sidewalls of the oxide mesas.
摘要:
A process for fabricating, on the more or less vertical edge of a silicon mesa, a MOS field-effect transistor which has a spacer-shaped gate and a right-angled channel path. The process involves the following steps: creating a raised region (the mesa) on a planar silicon substrate; creation of a gate oxide layer on the substrate and vertical sidewalls of the mesa; blanket deposition of a gate layer (typically polysilicon); anisotropically etching the gate layer to expose the upper surface of the mesa and leave a stringer gate around the circumference thereof; and doping the upper surface of the mesa and regions of the substrate peripheral to the circumferential polysilicon stringer to create source and drain regions. The standard process provides device density approximately double that of standard FET fabrication processes. Density can be increased even further by increasing the number of silicon mesas with a minimum pitch distance. This may be accomplished by employing the reduced-pitch masking technique disclosed in a copending U.S. patent application. Multiple transistors may be created on a single mesa by creating isolation regions within the mesa. The circumferential gate may be severed so as to provide a pair of gate inputs for transistors created on a single mesa. Enhancements common to conventional MOSFETS, such as lightly-doped source and drains, halos, etc., may be utilized for the new MOSFET process.
摘要:
A process for creating two thicknesses or gate oxide within a dynamic random memory. The process begins by thermally growing a first layer of gate oxide on a silicon substrate. This first layer is then masked with photoresist in regions where cell access transistors will ultimately be fabricated. All oxide that is not masked is then removed with an oxide etch. After the photoresist is stripped, a second layer of gate oxide is thermally grown on the substrate. The resultant oxided layer, which comprises multiple-thickness components, is used as a pad oxide layer during a conventional LOCOS operation. Peripheral driver transistors are construction on top of a thin layer of gate oxide so as to optimize their performance, whereas, cell access transistors are constructed on top of a thicker layer of gate oxide so as to minimize row line capacitance. A net increase in row line access speed is thus obtained.
摘要:
A CMOS integrated circuit such as a DRAM is fabricated, in which a first layer of polysilicon is used to form transistor gates, and capacitor cell plates are formed from a second polysilicon layer.N-wells are first formed, followed by initial oxide. The application of the CMOS process to the reverse poly technique provides enhanced alignment of critical transistor gates and permits the use of less mask steps in fabricating the CMOS circuit.
摘要:
This invention constitutes a 10-12 mask, split-polysilicon process for fabricating dynamic random access memories of the stacked capacitor type for the one-megabit generation and beyond. The process flow is characterized: reduced mask count due to the elimination of the N+ and p+ source-drain masking layers via the split polysilicon technique; an option to further reduce wafer processing by allowing the LOCOS stress relief (pad) oxide layer to later function as the transistor gate dielectric layer; N-channel device optimization via self-aligned punch-through and lightly-doped-drain (LDD) implants, without the addition of extra P-channel masking steps via the split poly approach; use of semi, self-aligned contact of bottom cell plate to access gate diffusion allowing tight spacing between bottom cell plate buried contact and access gate polysilicon; improved refresh characteristics achieved by avoiding reduction of isolation thickness due to the spacer oxide etch; improved refresh characteristics achieved by protecting the sensitive areas of the storage node from damage typically caused by a spacer oxide etch; improved refresh characteristics achieved by eliminating the high-dose N-channel source/drain implantation from the storage node side of the access transistor gate; and improved immunity to soft error upset achieved through the use of an optional self-aligned "Hi-C" implant that is performed without the addition of an extra masking step.
摘要:
The present invention provides a programmable structure for a programmable read-only memory (PROM) which utilizes one-sided ozone spacers constructed on the digit lines as one time programmable nodes. An oxide/nitride/oxide layer (ONO) is used as an interface between underlying parallel rows of digit lines, having one-sided ozone spacers, and overlying parallel columns of word lines in a programmable read only memory. With a each digit line passing under each word line in a row/column matrix is formed thereby providing a programmable digit/word line matrix. Each crossing point of the digit and word lines in the matrix will be permanently programmed to either a one or a zero by rupturing the thin ONO dielectric interface by applying the appropriate voltage potential between the associated digit/word line conductors.
摘要:
The present invention provides a programmable structure for a programmable read-only memory (PROM) which utilizes one-sided ozone spacers constructed on the digit lines as one time programmable nodes. An oxide/nitride/oxide layer (ONO) is used as an interface between underlying parallel rows of digit lines, having one-sided ozone spacers, and overlying parallel columns of word lines in a programmable read only memory With a each digit line passing under each word line in a row/column matrix is formed thereby providing a programmable digit/word line matrix. Each crossing point of the digit and word lines in the matrix will be permanently programmed to either a one or a zero by rupturing the thin ONO dielectric interface by applying the appropriate voltage potential between the associated digit/word line conductors.
摘要:
A process for creating and removing temporary silicon dioxide structures on an in-process integrated circuit with minimal effect on existing permanent silicon dioxide structures that are exposed. The process comprises the steps of blanket depositing an ozone-TEOS silicon dioxide layer through chemical vapor deposition on top of the in-process integrated circuit, thus covering permanent structures formed from conventional silicon dioxides (e.g. TEOS and thermal oxides), etching the ozone-TEOS layer to create said temporary structures, and removing the temporary structures using dilute hydrofluoric acid.
摘要:
A one-time, voltage-programmable, logic element has an antifuse element constructed within a trench etched in a silicon substrate. A sidewall of the trench abuts a diffusion region. The trench is lined with a nitride dielectric layer, which is in turn covered by polycrystalline silicon. The polycrystalline silicon serves as a voltage reference line. In a preferred embodiment, the diffusion region forms a first source/drain region of a field-effect transistor. In order to program the element, a voltage sufficient to rupture the nitride dielectric layer is applied between the diffusion region and the reference line. The transistor is utilized to isolate a particular logic element from other logic elements.