Patents Assigned to Texas Instruments - Acer Incorporated
  • Patent number: 6649308
    Abstract: The ultra-short channel transistor in a semiconductor substrate includes a gate structure that is formed on the substrate. Side-wall spacers are formed on the side walls of the gate structure as an impurities-diffusive source. Source and drain regions are formed in the substrate. A metal silicide contact is formed on the top surface of the gate structure, and on the surface of the source and drain regions. Extended source and drain regions are formed beneath the side-wall spacers and connect next to the source and drain regions.
    Type: Grant
    Filed: March 30, 1998
    Date of Patent: November 18, 2003
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6548362
    Abstract: A method of forming MOSFET with buried contacts and air-gap gate structure is disclosed. The method comprises following steps firstly, a gate is formed of pad oxide layer and a nitride layer sequentially on a silicon substrate, which has trench isolations. Then, a polysilicon layer and an oxide layer are deposited in order on all areas. Subsequently, an etched-back using the nitride layer a stopping layer is achieved. After that the nitride layer is removed thereby, forming a gate hollow region. After the pad oxide layer is removed, an oxynitride layer is regrown to be as the gate oxide. Thereafter, a silicon is deposited on all areas and refills in the gate hollow region. A planarization process is again performed using the oxide layer as an etch-stopping layer. Subsequently, the oxide layer is removed. S/D/G ion implanted into the polysilicon layer and the silicon layer. Then, the nitride spacers are removed to form dual recessed spaces.
    Type: Grant
    Filed: June 4, 1999
    Date of Patent: April 15, 2003
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6432785
    Abstract: The proposed method of the present invention forms MOSFETs with improved short channel effects and operating speeds over conventional devices. The method for fabricating MOSFETs includes the following steps. At first, isolation regions are formed on a semiconductor substrate and a gate insulating layer is formed on the substrate. A first conductive layer is then formed on the gate insulating layer and a first dielectric layer is formed on the first conductive layer. A removing process is performed to remove portions of the gate insulating layer, the first conductive layer and the first dielectric layer to define a gate structure. A layer formation step is carried out to form a thermal oxide layer on the substrate and on sidewalls the first conductive layer. Doped dielectric sidewall spacers are then formed on sidewalls of the gate structure. A removing step is carried out to remove portions of the thermal oxide layer uncovered by the doped dieletric sidewall spacers.
    Type: Grant
    Filed: March 23, 1999
    Date of Patent: August 13, 2002
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6331456
    Abstract: The present invention discloses a method to form CMOS transistors for high speed and lower power applications. A high energy and low dose phosphorous is implanted in a silicon substrate to fabricate an N-well after a pad oxide layer and a silicon nitride layer is formed. After a thick field oxide is formed by using a high temperature steam oxidation process, another high energy and low dose multiple boron implantation is then performed to fabricate a buried heavily boron doped region. A rapid thermal processing (RTP) system is following used to activate the boron dopant to form buried p+ layer and to recover the implanted damages. All the field oxide films are then removed by using a diluted HF or BOE solution. After porous silicon is obtained via anodic electrochemical dissolution in the HF solution, the porous silicon is then thermally oxidized to form the separate n-type silicon islands. Next, a thick CVD oxide film is deposited and then etched back to planarize device surface.
    Type: Grant
    Filed: May 4, 1998
    Date of Patent: December 18, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6316316
    Abstract: The method for forming flash memory includes the following steps. At first, a semiconductor substrate with an isolation region formed upon is provided. The semiconductor substrate has a pad oxide layer and a first nitride layer formed over. A portion of the first nitride layer and of the pad oxide layer are removed to define a gate region. A first oxide layer is formed and then a sidewall structure is formed. The semiconductor substrate is doped with first type dopants. A first thermal process is performed to form a second oxide layer and drive in the first type dopants. The sidewall structure and the first nitride layer are then removed, and the first oxide layer is removed to expose a portion of the substrate under the first oxide layer. Silicon grains are formed on the pad oxide layer, the exposed portion of substrate, and the second oxide layer. The exposed portion of the substrate is then etched to leave a rugged surface on the exposed portion of the substrate.
    Type: Grant
    Filed: June 18, 1999
    Date of Patent: November 13, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6294416
    Abstract: The present invention discloses a method of forming CMOS transistors with self-aligned planarization twin-well by using fewer mask counts. After a silicon nitride layer is formed over a first pad oxide layer on a semiconductor substrate, an N-well region is defined by first implanting in the semiconductor substrate. After removing the first photoresist layer, a second ion implantation is performed to define a P-well region. Next, both the silicon nitride layer and the first pad oxide layer are removed. A high temperature long time anneal is done to form a deep twin-well. A plurality of trench isolation regions is formed to define an active area region. A second pad oxide layer is formed on the substrate. A high energy and low dose blanket phosphorous is implanted in a semiconductor substrate for forming a punch-through stopping layer of the PMOSFET device. A low energy and low dose blanket BF2 implant then adjust both the threshold voltages of the PMOSFET and NMOSFET.
    Type: Grant
    Filed: May 7, 1999
    Date of Patent: September 25, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6294797
    Abstract: A gate insulator layer is formed over the semiconductor substrate and a first silicon layer is then formed over the gate insulator layer. An first dielectric layer is formed over the first silicon layer. A gate region is defined by removing a portion of the gate insulator layer, of the first silicon layer, and of the first dielectric layer. A doping step using low energy implantation or plasma immersion is carried out to dope the substrate to form an extended source/drain junction in the substrate under a region uncovered by the gate region. An undoped spacer structure is formed on sidewalls of the gate region and a second silicon layer is formed on the semiconductor substrate. The first silicon layer is then removed and another doping step is performed to dope the first silicon layer and the second silicon layer. A series of process is then performed to form a metal silicide layer on the first silicon layer and the second silicon layer and also to diffuse and activate the doped dopants.
    Type: Grant
    Filed: November 15, 1999
    Date of Patent: September 25, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6284612
    Abstract: The method of the present invention includes the following steps. First, a gate oxide layer is formed on the substrate. An undoped polysilicon layer is formed over the gate oxide layer. Then, a first dielectric layer is formed over the undoped polysilicon layer. A photoresist layer is formed over the first dielectric layer. Next, the photoresist layer is patterned to define a gate region. An etching process is performed to the photoresist layer to narrow the gate region. Portions of the first dielectric layer are etched using the residual photoresist layer as a mask. The undoped polysilicon layer is etched using the residual photoresist layer and the residual first dielectric layer as a mask. Then, a PSG is layer deposited over the residual first dielectric layer and the substrate. Subsequently, the PSG layer is etched back to form side-wall spacers to serve as ion diffusion source. A noble or refractory metal layer is deposited on all areas of the substrate.
    Type: Grant
    Filed: April 14, 1999
    Date of Patent: September 4, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6265263
    Abstract: The method for forming a DRAM capacitor can include the following steps. First, a first dielectric layer is formed on a semiconductor substrate, followed by the formation of a second dielectric layer on the first dielectric layer, and the formation of a third dielectric layer on the second dielectric layer. Next, the first, second, and third dielectric layers are patterned to form a contact hole therein. A doped polysilicon layer is then formed within the contact hole and over the third dielectric layer, followed by the formation of a fourth dielectric layer over the doped polysilicon layer. A patterning step patterns the fourth dielectric layer and the doped polysilicon layer to define a storage node. A hemispherical grained silicon layer is then formed on the fourth dielectric layer, on sidewalls of the storage node, and on the third dielectric layer.
    Type: Grant
    Filed: April 16, 1999
    Date of Patent: July 24, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6265259
    Abstract: The method of the present invention is to fabricate a CMOS device without boron penetration. Firstly, a gate oxide layer is formed on a semiconductor substrate. A first silicon layer is formed upon the gate oxide layer. Thereafter, a second silicon layer is stacked on the first silicon substrate, and N type dopant is in situ doped into the second silicon layer, and then a third silicon layer is stacked upon the second silicon layer. A gate structure is formed by patterning the stacked silicon layers, and source/drain structures with LDD regions are subsequently formed in the substrate by ion implantation processes. Finally, a thermal treatment is performed to form shallow source and drain junction in the substrate, thereby achieving the structure of the CMOS device. Meanwhile, the N type dopant is driven to the boundaries of stacked silicon layers of gate structure so as to act as diffusion barriers for suppressing boron penetration.
    Type: Grant
    Filed: July 13, 1999
    Date of Patent: July 24, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6259130
    Abstract: The device includes a gate oxide formed on a semiconductor substrate. Oxide regions are respectively formed on the substrate and adjacent to the gate oxide. Textured oxides are formed on the substrate, between the gate oxide and the oxide regions. A floating gate consists of a first polysilicon portion, second polysilicon portions and a third portion that is composed of hemisperical grained silicon (HSG-Si). The first polysilicon portion is formed on the gate oxide. Isolations are formed on the side walls of the first polysilicon portion. The second polysilicon portions are respectively formed next to the isolations and over a portion of the oxide regions. The HSG-Si is formed on the upper surface of the first polysilicon portion and the second polysilicon portions. A dielectric layer is formed on the HSG-Si of the floating gate. A control gate is formed on the dielectric layer. The doped regions are formed in the substrate and under the textured oxides and the oxide regions.
    Type: Grant
    Filed: March 9, 1999
    Date of Patent: July 10, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6255167
    Abstract: A method of fabricating buried bit line flash EEROM cells with shallow trench floating gates for suppressing the short channel effect is disclosed. The method includes the following steps. First, a first polysilicon layer with conductive impurities and a nitride capping layer are sequentially formed on a silicon substrate. The nitride cap layer serves as an anti-reflection coating (ARC) layer for improving the resolution of lithography. Then a photo-mask pattern on the ARC layer is formed to define trench regions, an anisotropic etching is performed to etch away unmasked portions of the nitride cap layer through the first polysilicon layer and slightly recess the silicon substrate using the patterned mask as a mask. After removing the patterned mask, a thermal annealed process is performed to grow a polyoxide layer on the sidewall of the first polysilicon layer and an thin oxynitride layer on the surface of the recessed silicon substrate.
    Type: Grant
    Filed: June 4, 1999
    Date of Patent: July 3, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6214696
    Abstract: The method includes forming a pad oxide, a polysilicon layer over a substrate. Next, an oxide layer is formed over the polysilicon layer. An opening is formed in the oxide layer, the polysilicon layer, and the pad layer. A trench is formed by etching the substrate using the oxide layer as a mask. A sidewall structure is then formed on the opening. Next, an exposed portion of the substrate is etched by using the sidewall structure as a mask. The sidewall structure and the oxide layer are then removed. An oxide and an oxynitride layer are then formed on the aforesaid feature. A semiconductor layer is then formed over the oxynitride layer. A portion of the semiconductor layer is oxidized for forming an insulating layer. Finally, a refilling layer is formed over the insulating layer and the substrate is planarized for having a planar surface.
    Type: Grant
    Filed: September 10, 1999
    Date of Patent: April 10, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6211016
    Abstract: A method for fabricating a high speed and high density nonvolatile memory cell is disclosed. First, a semiconductor substrate with defined field oxide and active region is prepared. A stacked silicon oxide/silicon nitride layer is deposited and then the tunnel oxide region is defined. A thick thermal oxide is grown on the non-tunnel region. After removing the masking silicon nitride layer, the source and drain are formed by an ion implantation and a thermal annealing. The pad oxide film is etched back, and a metal silicide film is formed and then stripped. A topography of the doped substrate region is then made rugged. Thereafter, a thin oxide is grown on the rugged doped substrate region to form a rugged tunnel oxide. Finally, the floating gate, the interpoly dielectric, and the control gate are sequentially formed, and the memory cell is finished.
    Type: Grant
    Filed: April 1, 1999
    Date of Patent: April 3, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6211556
    Abstract: A MOSFET device with buried contact structure on a semiconductor substrate has the following major elements with their relative locations. A gate insulator is on a portion of the substrate and a gate electrode is on the gate insulator. A gate sidewall structure is located on sidewalls of the gate electrode. Inside the substrate, a lightly doped source/drain region is under the gate sidewall structure, and a doped source/drain region is abutting the lightly doped source/drain region and located aside from a region under the gate sidewall structure. In addition, a doped buried contact region is also in the substrate next to the doped source/drain region. On the substrate, a silicon connection is located on a portion of the doped buried contact region, and a shielding block is on the doped buried contact region covering only a region uncovered by the silicon connection.
    Type: Grant
    Filed: June 1, 1999
    Date of Patent: April 3, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6211002
    Abstract: This invention proposes a process to form planarized twin-wells for CMOS devices. After depositing pad oxide and a silicon nitride layers, a high-energy phosphorus ion implantation is performed to form the N-well by using a photoresist as a mask. A thick oxide layer deposited by liquid phase deposition process is then grown on the N-well region part of the silicon nitride layer, but not on the photoresist. After stripping the photoresist, a high-energy boron ion implantation is carried out to form the P-well by using the LPD-oxide layer as a mask. The thick LPD-oxide layer is removed by BOE or HF solution. Trenched isolation regions are formed to isolate and define active regions. After removing the pad oxide and the silicon nitride layer, the CMOS device is fabricated by standard processes.
    Type: Grant
    Filed: April 15, 1999
    Date of Patent: April 3, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6207505
    Abstract: A method for fabricating a high-speed and high-density nonvolatile memory cell is disclosed. First, a semiconductor substrate with defined field oxide and active region is prepared. A stacked silicon oxide/silicon nitride layer is deposited and then the tunnel oxide region is defined. A thick thermal oxide is grown on the non-tunnel region. After removing the masking silicon nitride layer, the source and drain are formed by an ion implantation and a thermal annealing. The pad oxide film is then removed. A polysilicon film is deposited over the substrate 2 and then oxidized into sacrificial oxide layer. After stripping the sacrificial oxide layer, a rugged topography is then formed on the doped substrate regions. Thereafter, a thin oxide is grown on the rugged doped substrate region to form a rugged tunnel oxide. Finally, the floating gate, the interpoly dielectric, and the control gate are sequentially formed, and the memory cell is finished.
    Type: Grant
    Filed: June 7, 1999
    Date of Patent: March 27, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6207999
    Abstract: The present invention provides a mask ROM memory to minimize band-to-band leakage. The substrate includes a normal NMOS device region and a NMOS cell region for coding. An isolation region is formed between the normal NMOS device region and the NMOS cell region. A gate oxide layer is formed on the normal NMOS device region and a coding oxide layer is formed on the NMOS cell region, respectively. In the preferred embodiments, the coding oxide layer has a thickness of about two to ten times that of the gate oxide layer. Main gates are respectively formed on the gate oxide layer and the coding oxide layer. In the present invention, the main gates comprise materials like metal and metal compounds. Spacers are formed on the side walls of the main gates. First doped regions of source and drain regions, or namely lightly doped drains (LDD) and sources, are formed under the spacers and are adjacent to the main gates.
    Type: Grant
    Filed: January 27, 1999
    Date of Patent: March 27, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6204517
    Abstract: A structure of a single-electron-transistor memory array is disclosed in the present invention. A substrate is provided. A buried oxide layer is on the substrate. A plurality of silicon wires are arranged on the buried oxide layer, wherein each of the silicon wires has a pair of ends. Oxynitride layers covers on the silicon wires. A polysilicon layer covers the oxynitride layers and the buried oxide layer. A source region and a drain region connect to a first end and a second end of each of the silicon wires, respectively.
    Type: Grant
    Filed: April 9, 1998
    Date of Patent: March 20, 2001
    Assignee: Texas Instruments-Acer Incorporated
    Inventor: Shye-Lin Wu
  • Patent number: 6204124
    Abstract: A method for fabricating a high-speed and high-density nonvolatile memory cell is disclosed. First, a semiconductor substrate with defined field oxide and active region is prepared. A stacked silicon oxide/silicon nitride layer is deposited and then the tunnel oxide region is defined. A thick thermal oxide is grown on the non-tunnel region. After removing the masking silicon nitride layer, the source and drain are formed by an ion implantation and a thermal annealing. After the pad oxide film is removed, an undoped silicon film is deposited over the substrate 2 and then etched back by a dry etching. A rugged topography is then formed on the doped substrate regions. Thereafter, a thin oxide is grown on the rugged doped substrate region to form a rugged tunnel oxide. Finally, the floating gate, the interpoly dielectric, and the control gate are sequentially formed, and the memory cell is finished.
    Type: Grant
    Filed: May 17, 1999
    Date of Patent: March 20, 2001
    Assignee: Texas Instruments - Acer Incorporated
    Inventor: Shye-Lin Wu