Automated job management

The present invention relates to the technical field of automatic job management, and more particularly to the monitoring and control of manufacturing equipment processes in semiconductor manufacturing equipment.

The semiconductor chip manufacturing industry generally does not experience the high level of automation experienced by other technical disciplines. In semiconductor chip manufacturing plants in various areas, systems and tools are often only semi-integrated or even completely independent. Furthermore, manufacturing process automation is commonly used in the sense that it not only coordinates work between manufacturing equipment, but also collects data from the manufacturing equipment for use in process improvement and other job management functions. It is often very difficult because of its proprietary communication protocols.

In a typical configuration, manufacturing equipment is grouped and loosely controlled by a monolithic software program known as a "station controller". A typical station controller communicates with either an individual or a group of manufacturing equipment using an industry standard interface known as SECS / GEM. SECS / GEM is present in most semiconductor chip manufacturing equipment that uses 300 mm silicon wafers and is the standard communication interface in factories where 200 mm wafers are used. The needs of 300 mm and 200 mm factory types are very different, but in general both use station controllers in specific parts of the process.

Current software architectures that enable station controllers are particularly limited in terms of integration between semiconductor manufacturing equipment and how data is handled regarding the operation and state of the manufacturing equipment. Where manufacturing equipment needs to provide real-time data directly from manufacturing equipment to other software applications for analysis, diagnosis and rapid correction actions of the manufacturing process, manufacturing equipment integration and There are many design obstacles to overcome in the current software architecture used to extract data from manufacturing equipment.

One limitation is that the deployed legacy solution now uses a single-client SECS / GEM communication protocol to collect data from semiconductor manufacturing equipment managed by a station controller. Therefore, only a single client can communicate with each of the manufacturing equipment, and the data availability in the SECS / GEM interface specification drives the available dataset. Since the SECS / GEM interface does not reveal the structure of the manufacturing equipment, it is impossible to determine the physical structure of the manufacturing equipment. In addition to this, SECS / GEM is not a discoverable interface, so the application cannot query the manufacturing equipment to determine its capabilities. Furthermore, since SECS / GEM does not have a security mechanism, SECS / GEM does not have the concept of client authentication and permission. Finally, the single client limit means that there is no support for simultaneous multiple client access to manufacturing equipment information.

Another fundamental problem that exists in today's station controllers is the high complexity that results from drifting from key functions of material process control. In addition to the need for core job management, multiple features have been incorporated, creating large and complex software architectures that cannot easily adapt to change. This also results in a single point of failure with multiple internal points of failure and high maintenance costs. In general, data collection has been integrated into job management, making station controllers today the only manufacturing equipment data collector that requires data consumers to interface through station controllers.

Changes in the semiconductor industry that require semiconductor manufacturers to implement effective automation integration strategies are parallel to geometric size reductions that are now focused on 45nm and below, 300mm wafers and above. Firstly attributed to the exponential increase in the resulting manufacturing data, which must be managed in response to the increase in circuit capacitance with respect to. In addition to the driving force of change described above, some other pressures increase the need for change. First, there is a need to focus on the scarcity of expert resources for problem solving and reduce the resources spent on mere data exploration. In addition, the high cost of wafer erroneous processes at 45 nm, where each wafer consists of hundreds to thousands of chips (die), has exacerbated the need for effective solutions. High cost manufacturing equipment downtime and overall equipment There are additional performance issues driving the need for efficient solutions, such as requests for effectiveness (OEE) improvements. There is a further need for real-time data that allows for faster response to process problems, and a need for improved manufacturing equipment for production time.

Current e-Diagnostics and Advanced Process Control (which requires an automated architecture to support multiple clients accessing the manufacturing equipment at the same time and be independent of the current ownership of the manufacturing equipment process control. Current solutions will not solve data access requests for applications such as Advanced Process Control (APC). The ability to implement "data on demand" is a driving factor in station controller architectures that next-generation semiconductors will focus on. As the industry moves from lot-based to wafer-level manufacturing, automated solutions are required to make efficient manufacturing and operational decisions to meet customer demands (advanced statistical process control). It would be necessary to be able to provide statistical process control) (SPC), fault detection classification, and run-to-run control applications.

For the reasons mentioned above, a typical station controller is at stake. In the past, station controllers were designed to specifically control manufacturing job management, but now station controllers have evolved into a complex, intertwined set of programs whose functionality has expanded in the same way as its complexity. did. This complexity makes maintenance or modification of the station controller, as well as its basic functions such as job management, very difficult, time consuming and expensive. In some cases, the overlapping and intertwined nature of software code, even if its changes result in the improvement of the manufacturing process required to increase the output of semiconductor chips in operation, is a factory. Make the administrator very hesitant to make any changes.

The present invention solves problems with existing station controllers primarily by separating and distributing them from legacy station controllers, and the station controllers perform the main functions of job management and control originally designed for them. , Solved by creating a new module.

In one aspect of the invention, a software library is provided that makes it possible to implement a next generation station controller. In particular, it is possible to communicate with the manufacturing equipment server and create a thin station controller client dedicated to monitoring and controlling the manufacturing equipment process. The thin station controller client then communicates with the manufacturing equipment server that communicates with the manufacturing equipment. Industry standard communication protocols are supported and multiple client distributed architectures can be supported.

In another aspect of the invention, methods for job management and control are provided. In particular, the thin station controller client receives the job configuration information and communicates with the manufacturing equipment server. The manufacturing equipment server can implement a job factory and store job instances that provide real-time job status information to thin station controller clients. The manufacturing equipment server implements a manufacturing equipment model that separates the thin station controller client from direct communication with the manufacturing equipment and allows other clients to communicate with the manufacturing equipment.

The present invention, along with manufacturing equipment servers, provides solutions to many of the obstacles faced by those implementing station controller solutions in today's semiconductor installations. The thin station controller client according to the invention uses a different interface than SECS / GEM to connect to one or a group of manufacturing equipment, thereby a legacy station controller. Can operate without interfering with. An example of a manufacturing equipment server used with the present invention is the Equipment Interface Bridge (EIB) provided by Assist Technology. Asyst Director provided by Assist Technology is a specific example of a library for use in creating a job management client according to the present invention.

The thin station controller client realized by the present invention communicates via an advanced communication software layer that allows more flexibility in connecting to and acquiring data from the manufacturing equipment. In this case, the legacy station controller can remain in place while the thin station controller client is running, or the station controller is inefficiently packed into the legacy station controller. This means that many types of functionality can be completely disassembled into several modules that are distributed into new standalone components. The architecture of the present invention provides the advantage of using a non-entangled, open architecture solution for the realization of station controller functionality that works with the manufacturing equipment server.

In this way, thin station controller clients and manufacturing equipment servers address semiconductor factory automation by distributing some of the previously intertwined functions of legacy station controllers as independent and flexible clients. Revolutionize the method. The presence of the manufacturing equipment server allows the use of multiple distributed clients with specific functions such as job management, database management and manufacturing equipment performance tracking. The thin station controller client is a job management client whose function manages and executes jobs related to a group of semiconductor chip material processes.

The architecture of the present invention has many advantages over traditional station controllers. First, distributed client applications can be easily used. The manufacturing equipment server disconnects the manufacturing equipment for server communication from the client for server communication. Therefore, unlike the legacy station controller with multiple functions and multiple applications, it is possible to create a thin station controller client dedicated only for job management. On the other hand, other clients, such as data collection and data analysis dedicated clients, can be created in the same way and interact directly with the manufacturing equipment server without involving the thin station controller client. Data collection is effectively decoupled from job management.

This separation of functionality across multiple distributed clients eliminates the terribly detrimental dependency of non-job management applications on the data provided by the station controller. Therefore, the thin station controller client is no longer the only collection of manufacturing equipment data, thus reducing the cost of developing, implementing, and maintaining a job management solution. Only issues that directly affect job management and control need to be addressed by the thin station controller client. Another advantage of separating job management from data collection capabilities is that a single point of failure in the form of a legacy station controller is replaced by multiple potential points of failure that are not entangled in one place and are easier to diagnose. It is a point.

FIG. 1 shows a job flow according to an embodiment of the present invention. XML-style configuration 110 is used as input to the thin station controller client 120. These configurations identify behavior at the start and run time of a job. Director 130 is a software library used by client 120. Director 130 implements a full-featured application programming interface (API) for job generation and management. In a preferred job, the director 130 executes job generation and job execution services to support job state change events. Client 120 uses the job generation service to generate jobs based on XML-style configuration 110. The job execution service is used and job execution starts. Whenever the job state changes, the director 130 notifies the client 120 of the event in the form of a job state change event.

The manufacturing equipment server 160 implements the job factory 170 and includes the job instance 180. XML job requests, recipes and instructions are notified from the director 130 to the job factory 170 via the communication path 140. Job factory 170 includes job definitions and basic instructions. In an alternative embodiment, communication with Job Factory 170 supports the SEMI E40 standard for program job management. When it is time for the job to run, the job is instantiated through internal communication path 190 and individual job instances are spawned in job instance 180. The real-time job state change event is notified to the director 130 via the communication path 150. The job state change event gives the client 120 the ability to perform real-time process improvements. In an alternative embodiment, communication path 150 supports the SEMI E94 standard for job management control.

FIG. 2 shows an embodiment of the invention that could be used in a factory with other manufacturing equipment. The thin station controller client 236 incorporates a director 238, a MES handler 230, a manufacturing equipment handler 232 and a recipe handler 234 to provide an independent automated job management solution for any data consumer. Client 236 communicates with Manufacturing Execution System (MES) 210 and Recipe Management System (RMS) 220 via communication paths 215 and 225, respectively.

The interface between Client 236 and Director 238 is an easy-to-use application programming interface (API) for job generation and management, and industry standard technologies such as XML to reduce the complexity and learning curve of system usage. Based on the use of. This allows the client 236 system to be developed based on components in an open architecture language that is not a proprietary customer language or toolkit. This API incorporates a comprehensive library of job steps such as recipe downloads, remote commands, and event waits. These job steps implement the standard business rules found in all semiconductor manufacturing equipment, allow users to quickly configure basic program management, and interface with Manufacturing Execution System (MES) 210. Minimize the time required to connect. The Director 238 API is also preferably provided with a job cascade, characterized in that jobs can be ordered before process resources are available. This automatically starts the job when resources become available, which improves the throughput of critical process equipment.

In the alternative embodiment, the job management functions exposed by the API are based on industry standards such as SEMI E40 and E94. Job recipe generation can be further enhanced by providing GUI-based utilities that facilitate job generation. Such utilities reduce complexity and required integration time by providing a toolkit that automatically generates XML-based job specifications.

The client 236 communicates with the manufacturing apparatus server 240 via the communication path 242. In a preferred embodiment, the communication path 242 utilizes Microsoft's .NET communication framework. The standardized interface on line 242 allows the development of custom components for station controllers in any .NET language that is not a proprietary customer language or toolkit.

The manufacturing apparatus server 240 communicates with the manufacturing apparatus 260 via the communication path 265. In a preferred embodiment, the manufacturing equipment server 240 is an equipment interface bridge (EIB), such as an Asyst / EIB product from Assist Technology. Communication path 265 can support any manufacturing equipment for the server communication protocol, and in preferred embodiments, Interface A, the new SEMI standard E120 (common manufacturing equipment model), E125 (manufacturing equipment self-described), and It complies with the established SEMI standards E4 / 5 (SECS), E30 (GEM) and E37 (HSMS).

In addition to this, developers can also connect to rare types of manufacturing equipment (eg, communicating using the ASCII-based ftp protocol) by creating a custom interface to the manufacturing equipment server 240.

Client 236 is dedicated to job management and control, which allows data consumers 270, 276, 282 and 288 to effectively bypass client 236 and communicate directly with manufacturing equipment server 240 for these applications. be able to. This removes the data consumer's reliance on the job management client. The data consumer 270 is a statistical process control (SPC) client that utilizes data 272 and communicates with the manufacturing equipment server 240 over a communication path 244 that utilizes a .NET framework in a preferred embodiment. is there. The data consumer 276 is capable of using data 278 to communicate with the manufacturing equipment server 240 via a communication path 244 utilizing the HTTP / SOAP standard in a preferred embodiment. Capability (EEC) client. The data consumer 282 is an advanced process control (APC) client that utilizes data 284 and communicates with the manufacturing equipment server 240 over a communication path 248 that utilizes the HTTP / SOAP standard in a preferred embodiment. is there. The data consumer 288 is a consumer of custom or proprietary data that utilizes the data 290 and communicates with the manufacturing equipment server 240 via a communication path 250 that utilizes the .NET framework in a preferred embodiment.

Since the manufacturing equipment 260 is detached from the client 236, the use of the manufacturing equipment server 240 allows the available data to extend beyond the data currently available on the SECS / GEM interface. The manufacturing equipment server 240 uses a manufacturing equipment model that blocks the client 236 from direct communication with the manufacturing equipment 260. This manufacturing equipment model provides a discoverable interface that allows client applications to query the capabilities of the manufacturing equipment. This concept of manufacturing equipment model also combines multiple manufacturing equipment into one logical unit, which is then internally grouped and managed to make it look like one manufacturing equipment to client applications. Make it possible.

Note that station controller functions are simply "canned" across various factories and cannot be duplicated. Each factory has a number of unique business rules. Rather than having a static station controller with more and more custom rules added, the architecture of the present invention provides a toolkit of the basic essential functions required by the station controller, which provides a toolkit for it. The toolkit can then be implemented as a thin client application in a distributed environment.

The connectivity of multiple clients to the manufacturing equipment server allows the core of the station controller client to focus on its specialty, job management of material process control in the factory. By separating job processing from data acquisition, the data collection client application can focus on its specific needs rather than the needs requested by the legacy station controller. A further advantage of this approach is that the personnel providing and supporting the solution in this environment no longer need to be experts in all areas supported by previous station controllers.

Although traditional station controller mounting goals have been set for semiconductor manufacturing equipment, the present invention is also suitable for assembly and testing (often referred to as back-end) mounting.

The present invention has been described with respect to several embodiments. This has been done for the purposes of illustration only, and modifications of the present invention are readily apparent to those skilled in the art and are within the scope of the present invention.

<figref num="1">The flow of job generation and management for use according to the embodiment of the present invention is shown.</figref><figref num="2">An embodiment of the present invention together with other factory manufacturing equipment is shown.</figref>