Abstract: A method of manufacturing a semiconductor device is disclosed, which can completely remove hard mask residues left along boundaries between a high-voltage device region and STI structures after a dry etch process, by partially reducing a thickness of each of the exposed portion of the respective STI structures adjacent to the high-voltage device region so as to sufficiently expose the residues. As a result, after a portion of an Underlying pad oxide corresponding to the high-voltage device region is removed in a subsequent process, the exposed surface of the substrate is uniform with a smooth and clear border. Therefore, no sharp corners will emerge at a border of a gate oxide subsequently grown on the exposed surface of the substrate, and the gate oxide is thus morphologically improved, thereby resulting in an improvement of the reliability of the high-voltage semiconductor device being fabricated.

Abstract: The present invention discloses a method for inhibiting the electric crosstalk of back illuminated CMOS image sensor. This invention comprises, two ion implanting layers are implanted at the different area of the backside of the pixel unit after the thickness of the backside of CMOS image sensor is reduced. The ion concentrations implanted into the two layers are controlled to decrease progressively from top to bottom. An electric field is formed from top to bottom inside the epitaxial layer. The said electric field absorbs the incident light which arrives at the substrate region outside of the space charge of the photodiode. It reduces the electron diffuses in different pixels. Consequently, it reduces the electric crosstalk of pixels, improves the manufacture process and improve the image quality of the of CMOS image sensor.

Abstract: A method of forming contact hole is disclosed, including the steps of: providing a semiconductor substrate having a first dielectric layer, a second dielectric layer and a third dielectric layer formed thereon in this order; forming a first contact hole through the third dielectric layer, the second dielectric layer and the first dielectric layer by using an etching process to expose the semiconductor substrate; removing the third dielectric layer; forming a fourth dielectric layer over the second dielectric layer, the fourth dielectric layer filling the first contact hole; forming a second contact hole through the fourth dielectric layer, the second dielectric layer and the first dielectric layer to expose the semiconductor substrate; and removing the fourth dielectric layer. The method is capable of improving the stability of the contact-hole formation process.

Abstract: A method of fabricating a metal-insulator-metal (MIM) capacitor is disclosed, wherein after capacitor trenches have been formed in a dielectric layer by dry etching, a wet etching process is further applied to the dielectric layer to etch the one or more capacitor trenches. By taking advantage of an isotropic characteristic of the wet etching process, the corners of the one or more capacitor trenches are rounded after the wet etching. Accordingly, a lower electrode, an insulator and an upper electrode formed thereafter over the dielectric layer and the surfaces of the one or more capacitor trenches will also have similar rounded corners at corresponding positions. Such design may substantially reduce the risk of occurrence of point discharge in the resulting MIM capacitor and hence may improve the operational reliability of the capacitor.

Abstract: A method for etching a polysilicon gate is disclosed, wherein the polysilicon gate includes an undoped polysilicon portion and a doped polysilicon portion that is situated on the undoped polysilicon portion. The method includes: obtaining a thickness of the undoped polysilicon portion and a thickness of the doped polysilicon portion by using an optical linewidth measurement device; and etching the undoped polysilicon portion and the doped polysilicon portion by using two respective steps with different parameters, respective etching time for the undoped polysilicon portion and the doped polysilicon portion of every wafer being adjusted in real time by using an advanced process control system. This method enables the doped and undoped polysilicon portions of each polysilicon gate on every wafer to have substantially consistent profiles between each other.

Abstract: The invention provides a measurement of lateral diffusion of implanted ions in the doped well regions of semiconductor devices comprising: designing a test model having active areas, the P-type and N-type doped well regions of the active areas are separated by STI, and the bottom width of the STI is determined; performing multiple processes on the test model comprising the ion implantation process and the tungsten interconnection process to simulate a semiconductor device structure, wherein during the ion implantation process, in the P-type or N-type doped well regions, only the first procedure of the ion implantation process is performed; scanning the test model, obtaining a light-dark pattern of the tungsten interconnects. The present invention is convenient and accessible and can provide reference to optimize the property of the doped well regions of the semiconductor devices and ensure the yield enhancement.

Abstract: A method of fabricating n-channel metal-oxide-semiconductor (NMOS) devices is disclosed, the method including: providing a substrate having a plurality of NMOS structures formed thereon; depositing a silicon nitride layer having a high tensile stress over the substrate; and sequentially exposing and dry etching a plurality of portions of the silicon nitride layer in an order of channel lengths of the plurality of NMOS structures such that each portion of the etched silicon nitride layer has a thickness proportional to the channel length of its corresponding NMOS structure. Compared to a conventional method, the above fabrication method of NMOS devices can achieve uniform performance adjustment of NMOS devices after a silicon nitride layer with a high tensile stress is deposited.

Abstract: The present invention discloses a method for reducing the morphological difference between N-doped and undoped poly-silicon gates after etching, comprising the following sequential steps: depositing a hard mask layer on a substrate template having N-doped poly-silicon and undoped poly-silicon to form an N-doped poly-silicon hard mask layer and an undoped poly-silicon hard mask layer respectively, and etching the undoped poly-silicon hard mask layer to make a thickness difference between the N-doped poly-silicon hard mask layer and the undoped poly-silicon hard mask layer; depositing an anti-reflection layer, and etching according to a predetermined pattern until exposing the N-doped poly-silicon, wherein when the N-doped poly-silicon is exposed, the undoped poly-silicon is etched to a certain degree; and removing residuals on the surface of the above formed structure, and etching to form an N-doped poly-silicon gate and an undoped poly-silicon gate, respectively.

Abstract: The present invention relates to the field of semiconductor integrated circuits, and particularly relates to a method which can form a contact hole in a wafer of semiconductor material. The invention has proposed a method which can form a contact hole in a wafer of semiconductor: measuring and comparing a Critical Dimension (CD) of a position corresponding to the contact hole in the hard mask with the CD required in the technology, and then, based on the measurement, adjusting the CD of the position corresponding to the contact hole in the hard mask, by conformal deposition or etching technology, to fit a requirement of the technology; the method can reduce process costs while improving production capacity.

Abstract: The invention provides a SONOS structure, a manufacturing method thereof and a semiconductor device with the SONOS structure. The SONOS structure comprises: a first tunneling oxide layer formed on a substrate, a charge storage silicon nitride layer, a second silicon oxide layer, a thin graded silicon nitride layer having graded Si/N content formed on the second silicon oxide layer, a third silicon oxide layer formed on the thin graded silicon nitride layer, and a polysilicon control gate. The Si/N content ratio of the silicon nitride of the thin graded silicon nitride layer increases gradually, wherein the silicon nitride of the graded silicon nitride layer closer to the second silicon oxide layer contains higher nitride content, and the silicon nitride of the graded silicon nitride layer closer to the third silicon oxide layer contains higher silicon content.

Abstract: The present invention provides a method of forming connection holes. The method utilizes two different gases to perform two etching processes for the interlayer dielectric layer so as to form connection holes. The etching rate of the interlayer dielectric layer in the first etching process using the first etching gas is proportional to the size of the openings which defines the connection hole while the etching rate of the interlayer dielectric layer in the second etching process using the second etching gas is inversely related with size of the openings. According to the present invention, the first etching gas and the second etching gas compensate for each other to eliminate the loading effect, thus the connection holes are formed with almost the same depth. Therefore the damage of the etching stopper layer due to the high etching rate in the larger connection holes can be avoided, which prevents the excessive variation of the connecting resistance and expands the process window.

Abstract: The present invention provides a method for improving the write margins of the SRAM cells. The method comprises: before etching a polysilicon layer to form the polysilicon gates, performing a pre-implantation process to the polysilicon layer; wherein the polysilicon layer defines SRAM NMOSFETs regions and SRAM PMOSFETs regions; wherein the pre-implantation process comprises pre-implanting the fifth-group elements to the SRAM NMOSFETs regions and the NMOSFETs regions except to the SRAM NMOSFETs regions in the polysilicon layer, and pre-implanting the third-group elements to the PMOSFETs regions excluding the SRAM PMOSFETs regions in the polysilicon layer; wherein the process of pre-implanting the third-group elements comprises forming a pre-implantation photo mask capable of covering the SRAM PMOSFETs regions and using the pre-implantation photo mask to pre-implanting the third-group elements.

Abstract: A fabrication method for improving surface planarity after tungsten chemical mechanical polishing (W-CMP) is disclosed. The method forms contact holes and dummy patterns by performing two respective photolithography-and-etching processes to ensure that the dummy patterns have a depth smaller than that of the contact holes. Then the method fills tungsten into the contact holes and dummy patterns and removes the redundant tungsten by a W-CMP process. With such a method, difference of wiring density between areas can be reduced by the dummy patterns, and hence a better surface planarity of the contact hole layer can be achieved. Besides, as the dummy patterns are formed in a pre-metal dielectric layer and their depth is well controlled, tungsten filled in the dummy patterns will not contact with the device area below the pre-metal dielectric layer, and thus will not affect the performance of the device.

Abstract: The present invention disclosed an algorithm of Cu interconnect dummy inserting, including: divide the surface of semiconductor chip into several square windows with an area of A, each of which is non-overlap; perform a logic operation on each square window; and divide the window into two parts: {circle around (1)} the area to-be-inserted; {circle around (2)} the non-inserting area; determine the metal density of the dummy pattern that should be inserted to each square window and the line width; determine the dummy pattern that should be inserted to the windows according to the metal density, line width, the pre-set dummy pattern and the layouting rules. The beneficial effects of the present invention is: avoided the shortcomings of fill density maximization in the rule-based filling method by using reasonable metal density and line width.

Abstract: The present invention provides a method of forming a nitrogen-free dielectric anti-reflection layer comprising: introducing a reaction gas into the discharge tube until the reaction gas reaching a stable state; introducing the reaction gas into the reaction chamber and then generating a plasma, or generating a plasma and then introducing the reaction gas into the reaction chamber, wherein the time delay occurs between the two processes is utilized to perform the deposition of the nitrogen-free dielectric anti-reflection layer; finally stop introducing the reaction gas and then stop generating the plasma. The method can flexibly control the extinction coefficient and the refractive index of the nitrogen-free dielectric anti-reflection layer so as to obtain a straight photoresist pattern and greatly reduce the photoresist standing waves effect and photoresist poisoning effect.

Abstract: A method of depositing phosphosilicate glass (PSG) is disclosed. The method includes a first deposition step for depositing a first PSG layer with a sputtering deposition ratio of 0.10 to 0.16, and a second deposition step for depositing a second PSG layer with a sputtering deposition ratio of 0.18 to 0.22 after the first deposition step. The first PSG layer has a thickness smaller than that of the second PSG layer. With such two-step deposition method, flower pattern having a dramatically reduced size can be formed without occurrence of clipping or formation of sidewall voids in the resultant gate patterns. Specifically, the formed flower pattern has a height reduced by about 50% and a thickness reduced by about 30%.

Abstract: The invention provides an SONOS structure and a manufacturing method thereof The manufacturing method comprises: forming a tunneling oxide layer on a substrate; depositing a Si-rich silicon nitride layer above the tunneling oxide layer, wherein the Si/N content ratio of the Si-rich silicon nitride layer is constant; depositing a graded silicon nitride layer having graded silicon content above the Si-rich silicon nitride layer; and depositing a blocking oxide layer; wherein the silicon content of the graded silicon nitride layer is reduced in the direction from the Si-rich silicon nitride layer to the blocking oxide layer. According to the present invention, the Si-rich silicon nitride layer provides shallower trapping levels, which is beneficial to trap the charges and improve the programming and erasing speed. Furthermore, the charge retention time increases due to the constrained charges in the deep trapping levels, thus the reliability of the device enhances.

Abstract: The present invention provides a method and model for monitoring the pretreatment process of a low-k block layer. The method comprises measuring film parameters of the film formed on the silicon substrate after applying the pretreatment process for different time periods; creating a statistical process control curve according to the film parameters; setting a SPC control limit; determining the pretreatment process normal when the data point of measurement in the SPC curve is within the control limit while determining the pretreatment process abnormal when the data point of measurement in the SPC curve exceeds the control limit. According to the present invention, the failure of the pretreatment process can be prevented to improve the product reliability and stability.

Abstract: The invention provides a method for monitoring devices in semiconductor process comprising: Step a, designing a sampling plan with fixed sample size before the beginning of the semiconductor process; Step b, determining whether to sample the wafers according to the sampling plan and dispatching the wafers to be sampled to each process device before the beginning of the process step, wherein the process device is used for performing the process step; Step c, performing the process step; Step d, sampling the wafers according to the sampling plan, and performing in-line inspection to the sampled wafers according to the sampling results; Step e, repeating Step b to Step d until all the process steps are completed; Step f, performing e-test to all the wafers. According to the method, the potential risk during the semiconductor process can be minimized through the coordination of the sampling plan and the dynamic risk flag.

Abstract: The invention provides a semiconductor yield management system. The system comprises an electronic data collection module and an execution module, the execution module comprises a plurality of execution sub-modules in sequence to perform executions on an object successively, the of the execution sub-modules comprises an execution section and an inspection section; the execution section of the execution sub-module is connected with the inspection section of the preceding execution sub-module except for the first execution sub-module; the inspection section of the execution sub-module is connected with the execution section of the subsequent execution sub-module except for the last execution sub-module; the inspection module of the execution sub-module is connected with the electronic data collection module. According to the semiconductor yield management system, the potential not-good wafers can be recorded, analyzed and distributed to the corresponding execution module, which realizes the risk minimization.