Abstract: An improved multilevel interconnect structure is provided. The interconnect structure includes several levels of conductors, wherein conductors on one level are staggered with respect to conductors on another level. In densely spaced interconnect areas, interposed conductors are drawn to dissimilar elevational levels to lessen the capacitive coupling between the interconnects. By staggering every other interconnect line in the densely patterned areas, the interconnects are capable of carrying a larger amount of current with minimal capacitive coupling therebetween.

Abstract: The present invention generally provides a semiconductor substrate having an extended test structure and a method of fabricating such a substrate. A method of forming an extended test structure on a semiconductor substrate, consistent with one embodiment of the invention, includes forming a first test structure pattern over a first portion of the substrate and forming a second test structure pattern of the second portion of the substrate which partially overlaps the first portion of the substrate such that the first test structure pattern and the second test structure overlap. The first test structure pattern may be formed using, for example, reticle and a second test structure pattern may be formed using the same reticle. The first and second test structure patterns may, for example, be formed in a scribe line of the substrate.

Abstract: An improved multilevel interconnect structure is provided. The interconnect structure includes pillars spaced from each other across a wafer. The pillars are placed between levels of interconnect or between an interconnect level and a semiconductor substrate. The pillars are spaced from each other by an air gap, such that each conductor within a level of interconnect is spaced by air from one another. Furthermore, each conductor within one level of interconnect is spaced by air from each conductor within another level of interconnect. Air gaps afford a smaller interlevel and intralevel capacitance within the multilevel interconnect structure, and a smaller parasitic capacitance value affords minimal propagation delay and cross-coupling noise of signals sent through the conductors.

Abstract: Dopant atoms have coefficients of diffusion that vary due to implant damage. Damaged regions are selected and created by implanting silicon atoms into a silicon substrate prior to formation of a gate electrode. The silicon atoms act as a getter for attracting selected dopants that are trapped in the silicon substrate. Dopants are implanted in the vicinity of the damaged regions and diffused by transient-enhanced diffusion (TED) into the damaged regions by thermal cycling to accumulate dopant atoms. Transient-enhanced diffusion improves the doping of a substrate by enhancing the diffusion of dopants at relatively low anneal temperatures. Dopant accumulation sets particular selected electrical properties without placing an excessive amount of dopant in regions adjacent to junctions for purposes including threshold control for a field device, threshold setting for a transistor, and prevention of device punchthrough.

Abstract: An improved method for planarizing an interlevel dielectric comprising two chemical mechanical polish steps. After an interlevel dielectric containing a topographical valley between a pair of topographical peaks is formed, the dielectric is chemically-mechanically polished in a first polish step at a first force using a first polish pad having a first rigidity to round the sharp dielectric corners or edges that exist at the transition between the peaks and valleys. After the first polish step has rounded the edges, a second polish step is performed with a second polish pad of second rigidity. The second polish pad is more rigid than the first polish pad and the second force is greater than the first. The second polish steps uses a high viscosity slurry to reduce slurry turnover in the regions proximate to the dielectric valleys thereby reducing the chemical etching in the valleys and improving the planarization efficiency.

Abstract: An IGFET with a gate electrode and metal spacers in a trench is disclosed. The IGFET includes a trench with opposing sidewalls and a bottom surface in a semiconductor substrate, metal spacers adjacent to the sidewalls and the bottom surface, a gate insulator on the bottom surface between the metal spacers, protective insulators on the metal spacers, a gate electrode on the gate insulator and protective insulators, and a source and drain adjacent to the bottom surface.

Abstract: A method of making N-channel and P-channel IGFETs with different gate thicknesses and spacer widths is disclosed. The method includes providing a semiconductor substrate with a first active region of a first conductivity type and a second active region of a second conductivity type, forming a first gate over the first active region and a second gate over the second active region, wherein the second gate has a substantially greater thickness than the first gate, forming first spacers in close proximity to opposing sidewalls of the first gate and second spacers in close proximity to opposing sidewalls of the second gate, wherein the second spacers have a substantially greater width than the first spacers due to the second gate having a substantially greater thickness than the first gate, and forming a first source and a first drain of the second conductivity type in the first active region and a second source and a second drain of the first conductivity type in the second active region.

Abstract: A fabrication process that produces an air gap dielectric in which a multi-level interconnect structure is formed upon a temporary supporting material. The temporary material is subsequently dissolved away leaving behind an intralevel and an interlevel dielectric comprised of air. In one embodiment of the invention, a first interconnect level is formed on a barrier layer. A temporary support material is then formed over the first interconnect level and a second level of interconnect is formed on the temporary support material. Prior to formation of the second interconnect level, a plurality of pillar openings are formed in the temporary material and filled with a conductive material. In addition to providing a contact between the first and second level of interconnects, the pillars provide mechanical support for the second interconnect level. The temporary material is dissolved in a solution that attacks the temporary material but leaves the interconnect material and pillar material intact.

Abstract: An IGFET with source and drain contacts in close proximity to a gate with sloped sidewalls is disclosed. A method of making the IGFET includes forming a gate over a semiconductor substrate, wherein the gate includes a top surface, a bottom surface and opposing sidewalls, and the top surface has a substantially greater length than the bottom surface, forming a source and a drain that extend into the substrate, depositing a contact material over the gate, source and drain, and forming a gate contact on the gate, a source contact on the source, and a drain contact on the drain. The gate is separated from the source and drain contacts due to a retrograde slope of the gate sidewalls, and the gate contact is separated from the source and drain contacts due to a lack of step coverage in the contact material. Preferably, the contact material is a refractory metal, and the contacts are formed by converting the refractory metal into a silicide.

Abstract: A transistor with a buried insulative layer beneath a channel region is disclosed. Unlike conventional SIMOX, the buried insulative layer has a top surface beneath the channel region that is closer than bottom surfaces of the source and drain to the top surface of the substrate. Preferably, the buried insulative layer is formed by implanting oxygen into the substrate and then performing a high-temperature anneal so that the implanted oxygen reacts with silicon in the substrate to form a continuous stoichiometric layer of silicon dioxide. Advantageously, the buried insulative layer provides a diffusion barrier and an electrical isolation barrier for the channel region.

Abstract: A method of making an IGFET with a multilevel gate that includes upper and lower gate levels is disclosed. The method includes providing a semiconductor substrate with an active region, forming a gate insulator on the active region, forming a first gate material with a thickness of at most 1000 angstroms on the gate inslator and over the active region, forming a first photoresist layer over the first gate material, irradiating the first photoresist layer with a first image pattern and removing irradiated portions of the first photoresist layer to provide openings above the active region, etching the first gate material through the openings in the first photoresist layer using the first photoresist layer as an etch mask for a portion of the first gate material that forms a lower gate level, removing the first photoresist layer, forming an upper gate level on the lower gate level after removing the first photoresist layer, and forming a source and drain in the active region.

Abstract: A method for isolating a first active region from a second active region, both of which are configured within a semiconductor substrate. The method comprises forming a trench in the semiconductor substrate between said first active region and said second active region. A first dielectric layer is then formed on said trench and a polysilicon layer is deposited on said first dielectric layer. The polysilicon layer is then thermally oxidized to form a second dielectric layer. Preferably the first dielectric is a thermal oxide 40 to 500 angstroms in thickness consuming less than 200 angstroms of said first active region and said second active region. The polysilicon layer is preferably between 1000 to 2000 angstroms.

Abstract: An isolation technique is provided for improving the overall planarity of trench isolation regions relative to adjacent silicon mesas. The isolation process results in a spaced plurality of silicon risers formed in wide isolation regions. The space between silicon risers are ideally suited for optimal fill of a dielectric deposited across the semiconductor topography, i.e., across and between the silicon risers formed between active areas. The silicon risers, and optimally dimensioned trenches extending between the risers, enhance the planarity of the deposited dielectric. The deposited dielectric upper surface includes recesses of minimal elevational disparity, wherein the recesses are closely spaced in alignment directly above the trenches formed between silicon risers. The recesses can be readily removed by a chemical-mechanical polishing step with minimal deformity to the polishing pad, resulting in global planarization of the dielectric upper surface.

Abstract: An integrated circuit is provided having an improved interconnect structure. The interconnect structure includes a power-coupled local interconnect which is always retained at VDD or VSS (i.e., ground) level. The local interconnect resides a dielectric-spaced distance below critical runs of overlying interconnect. The powered local interconnect serves to sink noise transients from the critical conductors to ensure that circuits connected to the conductors do not inoperably function. Accordingly, the local interconnect extends along a substantial portion of the conductor length, and is either wider or narrower than the conductor under which it extends. The local interconnect can either be polysilicon, doped polysilicon, polycide, refractory metal silicide, or multi-level refractory metal.

Abstract: It has been discovered that different pattern densities that occur in conventional lithography produce a different final etch polysilicon gate width in high density (dense) regions of polysilicon gates as compared to low density (isolated) polysilicon gate regions. The final etch polysilicon gate width for a dense region is smaller by a predictable distance relative to the final etch polysilicon gate width for an isolated region. For example, a typical dense region has a final etch polysilicon gate width that is approximately 0.05 .mu.m smaller relative to the final etch polysilicon gate width of isolated regions having a channel length of 0.35 .mu.m. A biasing technique is employed for a polysilicon masking reticle in which the reticle is biased differently in regions of isolated polysilicon gates in comparison to regions of dense polysilicon gates.

Abstract: A method of doping an integrated circuit device channel in a semiconductor substrate laterally enclosed by an isolation structure is disclosed. The method includes steps of forming a thin oxide layer overlying the integrated circuit device channel and the isolation structure, depositing a polysilicon blanket layer overlying the thin oxide layer, patterning a photoresist mask overlying the polysilicon blanket layer and implanting dopant impurities into the polysilicon blanket layer. The method further includes steps of diffusing the dopant impurities from the polysilicon blanket layer through the thin oxide layer into the integrated circuit device channel, removing the polysilicon blanket layer, and removing the thin oxide layer.

Abstract: A region of damaged silicon is exploited as a gettering region for gettering impurities in a silicon substrate. The region of damaged silicon is formed between source and drain regions of a device by implanting silicon atoms into the silicon substrate after the formation of a gate electrode of the device. The damaged region is subsequently annealed and, during the annealing process, dopant atoms such as boron segregate to the region, locally increasing the dopant concentration in the region. The previously damaged region is in a location that determine the punchthrough characteristics of the device. The silicon implant for creating a gettering effect is performed after gate formation so that the region immediately beneath the junction is maintained at a lower dopant concentration to reduce junction capacitance.

Abstract: An isolation technique is provided for improving the overall planarity of isolation regions relative to adjacent active area silicon mesas. The isolation process results in a trench formed in field regions immediately adjacent the active regions. The trench, however, does not extend entirely across the field region. By preventing large area trenches, substantial dielectric fill material and the problems of subsequent planarization of that fill material is avoided. Accordingly, the present isolation technique does not require conventional fill dielectric normally associated with a shallow trench process. While it achieves the advantages of forming silicon mesas, the present process avoids having to rework dielectric surfaces in large area field regions using conventional sacrificial etchback, block masking and chemical-mechanical polishing. The improved isolation technique hereof utilizes trenches of minimal width etched into the silicon substrate at the periphery of field regions, leaving a field mesa.

Abstract: A photolithography mask derivation process is provided for improving the overall planarity of interlevel dielectric deposited upon conductors formed by the derived photolithography mask. The photolithography mask is derived such that non-operational conductors are spaced a minimum distance from each other and from operational conductors to present a regular spaced arrangement of conductors upon which a dielectric layer can be deposited and readily planarized using, for example, chemical-mechanical polishing techniques. The resulting interlevel dielectric upper surface is globally planarized to an even elevational level across the entire semiconductor topography. The operational conductors are dissimilar from non-operational conductors in that the operational conductors are connected within a circuit path of an operational integrated circuit. Non-operational conductors are not connected within the integrated circuit path and generally are floating or are connected to a power supply.

Abstract: A reticle provides an image pattern and compensates for a lens error in a photolithographic system. The reticle is structurally modified using image displacement data indicative of the lens error. The reticle can be structurally modified by adjusting the configuration (or layout) of radiation-transmitting regions, for instance by adjusting a chrome pattern on the top surface of a quartz base. Alternatively, the reticle can be structurally modified by adjusting the curvature of the reticle, for instance by providing a chrome pattern on the top surface of a quartz base and grinding away portions of the bottom surface of the quartz base. The image displacement data may also vary as a function of lens heating so that the reticle compensates for lens heating associated with the reticle pattern.