Site News  |   Monitor Keywords  |   Monitor Archive  |   Organizer  |   Account Info  |

Design structure and system for identification of defects on circuits or other arrayed products

Design structure and system for identification of defects on circuits or other arrayed products description/claims

This application is a divisional of U.S. patent application Ser. No. 10/459,132, filed on Jun. 10, 2003, and claims priority thereto through U.S. patent application Ser. Nos. 11/493,092 (filed Jul. 25, 2006) and 11/926,605 (filed Oct. 29, 2007).

This invention relates generally to a design structure and a system and method for manufacturing products having an array of sub-components, such as wafers and semiconductor circuits. More particularly, the present invention relates to efficiently identifying and characterizing defects in such products, improving yield and identifying yield trends, and a design structure therefore. The present invention is particularly valuable in the fabrication of wafers and semiconductor circuits.

In the semiconductor industry, there is a continuing movement towards higher integration, density and production yield, all without sacrificing throughput or processing speed. The fabrication of integrated circuits (ICs) requires a complex process to ensure the proper balance between throughput, processing speed and yield. Inspections and tests are designed to detect unwanted variations in the wafers produced, as well as in the equipment and masks used in the fabrication processes. One small defect in either the devices produced or in the process itself can render a finished device inoperable.

Manufacturing ICs is a complex process that may involve hundreds of individual operations of fabrication, inspection and testing steps that are interwoven throughout the entire process. The fabrication process includes the diffusion of predetermined amounts of a dopant material into predetermined areas of a wafer, which is typically silicon, to produce active devices such as transistors. This is usually accomplished by forming a layer of silicon dioxide on the wafer, then utilizing a photomask and photoresist to define a pattern of areas into which diffusion occurs through a silicon dioxide mask. Openings are then etched through the silicon dioxide layer to define the pattern of precisely sized and located openings through which diffusion will take place. After a predetermined number of diffusion operations have been carried out to produce the desired number of transistors in the wafer, they are interconnected. Interconnection lines, or interconnects, are typically formed by deposition of an electrically conductive material that is formed into the desired interconnect pattern by a photomask, photoresist and/or etching process. Some high-performance ICs like the 1.3 GHz IBM Power4 microprocessor have hundreds of millions of transistors on a chip measuring 2 cm by 2 cm. Such chips include various devices disposed among 6 or 8 vertical layers and copper interconnects that total over a mile in length. Both devices and interconnects are measured in submicron dimensions.

In view of the device and interconnect densities required in current ICs, it is desirable that the manufacturing processes be carried out with utmost precision and in a manner that minimizes defects. In order to achieve reliable operation, the electrical characteristics of the circuits must be kept within carefully controlled limits, which require a high degree of control over all of the formation operations and fabrication processes. For example, in the photoresist and photomask operations, the presence of contaminants such as dust, minute scratches and other imperfections in the patterns on the photomasks can produce defective patterns on the semiconductor wafers, resulting in defective integrated circuits. Furthermore, defects can be introduced in the circuits during the diffusion operations themselves. Defective circuits may be identified by, for example, optical tools and electron-based tools. Various inspection tools offer different advantages (e.g., different resolution, different magnification, different wafer throughput). Typically, the smallest defects are inspected with SEM, or scanning electron microscopy. Optical diagnostic tools such as pico-second imaging circuit analysis, laser voltage probing, light-induced voltage alteration, optical beam induced current, Seebeck effect imaging, thermally-induced voltage alteration, and soft defect localization are becoming more common in IC chip fabrication. Once defective ICs have been identified, it is desirable to take steps to minimize the number of defective ICs produced in the manufacturing process, thus increasing the yield of non-defective ICs.

Defects are the primary killers of devices, wafers and circuits formed during manufacturing processes, resulting in yield losses. Increases in device density require smaller devices and interconnects, which appear to be approaching a physical limit of operability for certain such devices (i.e., field-effect transistors with channel layers several atomic layers thick). Defect detection in such atomic level devices becomes increasingly challenging. Specifically, it is more difficult to detect defects in individual devices or interconnects due to their smaller size, yet higher device density on a single chip means fewer total defects per chip are acceptable without causing chip failure. Many of the defects that cause poor yield in ICs were caused by particulate contaminants or other random sources. However, many of the defects seen in modem IC processes are not a result of particulates or random contaminants, but rather stem from systematic sources. Examples of systematic defect sources include printability problems from using aggressive lithography tools, poly stringers from improperly formed silicides, gate length variation from density driven and optical proximity effects. Other examples of defects include bubbles and particles in the photoresist layer of the IC. The diagnosis of defects in the photoresist layer can only be accomplished after the photoresist is developed. Furthermore, it is typically the case that the bubbles and particle defects in the photoresist do not appear as bubbles or particle defects after the photoresist is developed, but may take on some other distorted shape or appearance, further complicating the diagnosis.

In attempting to decrease the number of defective ICs produced in the manufacturing process, thereby increasing the yield, it is necessary to consider that any one or more of possibly several hundred processing steps may have caused a particular circuit to be defective. With such a large number of variables, it can be extremely difficult to determine the exact cause or causes of a defect or defects in a particular circuit thereby making it extraordinarily difficult to identify and correct yield reductions. While inspection of the completed ICs may provide some indication of which process operation may have caused the circuits to be defective, inspection equipment often does not capture many of the sources of systematic defects and/or the tools can be difficult to use effectively and reliably. Furthermore, inspection equipment may detect false defects, false alarms or nuisance defects that frustrate attempts to reliably detect true defects or sources of defects.

Once a particular cause of a true or catastrophic “killer” defect has been identified after completion of the fabrication process, it can be confirmed that a problem in a particular process operation was present at the time that the particular process operation was carried out, which could have been weeks or even months earlier. Thus, the problem might be corrected only after many defective ICs have been produced. By the time the first problem has been identified, other process operations may be causing problems. Thus, after-the-fact analysis of defective ICs and identification of process operations causing these defective products are of limited value to improve the overall yield of ICs.

What is needed to advance the state of the art is a method and apparatus of adaptive filtering of wafers and chips with chip design data within a semi-conductor fabrication process for accurate identification of catastrophic defects and accurate yield trends.

Accordingly, one embodiment of the present invention is directed to a method for assessing a probability of failure to operation of a semiconductor wafer. This method includes inputting risk factor data into a memory and inputting a plurality of wafers into a semiconductor fabrication manufacturing process. Each wafer has at least one mask layer. A subset of wafers is selected to obtain a sample population and at least one region of each wafer of the sample population is inspected. Circuit design data associated with each wafer of the sample population is obtained. One or more defects that present an increased risk to the operation of a particular wafer are identified, if present on the wafer. The identification is a function of the risk factor data, the inspecting step and the circuit design data. A probability of semiconductor wafer failure is thereby established.

Another embodiment of the present invention is directed to a process for calculating a risk factor for a region of a wafer. This process includes obtaining circuit design data and tiling the circuit design data on a surface of the wafer. Location defects on the wafer are identified using design shapes that are in the region of the defect. The location risk is evaluated and a failure probability in the region is determined.

Yet another embodiment of the present invention is directed to a method of generating a data representation. This method includes generating a graph of circuit design data and identifying one or more enhanced risk regions of the circuit design data. Defects are identified and data indicative of overlap between the enhanced risk regions and the one or more defects is generated. The quantity of defective circuits is also established.

Yet another embodiment of the present invention is directed to an apparatus for assessing one or more flaws in a circuit. This is accomplished by a calculation module that selects a function to calculate a location of risk of the circuit. A location module establishes a defect location as a function of circuit design data and an evaluation module that receives data from the calculation module and the location module, and generates an output as a function of the received data; where the output identifies an inspection area of the circuit.

Yet another embodiment of the present invention is directed to a method for assessing regions of interest of a circuit, the method being stored as computer executable instructions on a computer-readable medium. The method includes the steps of providing circuit design data and providing defect data relating to one or more defects. A region of the circuit that contains a location of the defects is identified and a failure probability is generated as a function of the circuit design data, defect data and positional location.

The present invention is not limited only to integrated circuits, but also applies to any product that has an array of sub-components. The term “array” as used herein refers to an assemblage having a highly ordered arrangement, especially an assemblage of small-scale objects. Examples of products with arrayed sub-components include an array of optical devices or nano-scale mechanical devices, a flat panel display having arrayed layers and/or electro-optical components, and a micro-capillary array such as may be used in genetic sequencing. In its broader sense, then, another embodiment of the present invention is a method for evaluating defects in a product that defines an array of components. This broader method includes inputting risk factor data into a memory such as a computer readable memory, and inspecting at least one region of the product in accordance with the risk factor. The region may be defined as a particular sub-component, a type or class of sub-components, any location wherein a particular arrangement of materials is more prone to failure (such as any interface between a superconductive material and a buffer layer to which it is bound), or a particular physical location within the product. The method further includes obtaining design data associated with the product, and identifying one or more defects that present an increased risk of failure of the product based on the risk factor data, the design data, and the inspecting of the at least one region of the product. Preferably, identified defects are used to update various weighting factors that may be associated with one or more risk factors.

Yet another embodiment of the present invention is a method of enabling efficient detection of defects in a product defining an array of sub-components. This particular method includes using product design information and at least one risk factor related to manufacture of the product to populate a database, and allowing a third party access over a network to relevant information in the database. The relevant information includes at least design data tailored the detection of defects within sub-components and/or physical locations of the product, and at least one weighted risk factor that may be used to identify a specific sub-component and/or a specific physical location of the product that are prone to defects. Preferably, at least one weighted risk factor in the database is updated based on an input from the third party, the input most preferably related to actual defect detection.

Another aspect of the present invention that does not necessarily include a network includes a computer program embedded on a medium readable by a computer, such as a CD-ROM, a zip-drive, or a hard-drive to which the computer program may be downloaded via network or from an interim medium. The computer program includes several code segments that may be intertwined when composed or decompiled, but that are functionally separated into the following. A first program code segment includes design data for a product that defines an array of sub-components. A second program code segment includes risk factor data relating to likelihood of a defect in at least one of a sub-component or a physical location of the product. A third program code segment provides for inputting data related to actual defects discovered in the product. Preferably, this aspect of the invention also includes a fourth program segment that modifies the risk factor data in response to data input that relates to actual defects discovered.

• Provisional Patent
• Utility Patent

• What Is a Patent?
• What Is a Trademark or Servicemark?
• What Is a Copyright?
• Patent Laws