« PreviousContinue »
Merged SECS and Sniffed Data (Tool A)
Process Tool A
Line 4, Bay 3
Start Time: 19:24:05
Recipe: Metal 1
300 mm ASIC
Module A Recipe: ECD Barrier
Module A Temp: 400*C
Module A Vac: 4E-9 mTORR
Module A Recipe Time: 92 Sec
Module A Gas Flow: 3.2 LPM
Module B Recipe: ECD Clean
Module B Temp: 100'C
Module B Vac: 1E-7 mTORR
Module B Recipe Time: 152 Sec
Module B Gas Flow: 6.2 LPM
Module C Recipe: ECD Seed
Module C Temp: 350*C
Module C Vac: 4E-9 mTORR
Module C Recipe Time: 230 Sec
Module C Gas Flow 1: 5.2 LPM
Module C Gas Flow 2: 8.1 LPM
Module C Gas Flow 3: 1.3 LPM
Module D Recipe: ECD Fill
Module D Temp: 300°C
Module D Vac: 4E-9 mTORR
Module D Recipe Time: 15 Sec
Module D Gas Flow: 3.2 LPM
SYSTEM AND METHOD FOR
ELECTRONICALLY COLLECTING DATA IN
A FABRICATION FACILITY
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to collecting data generated by equipment used for processing substrates used in semiconductor, data storage, and allied industries. More particularly, the present invention relates to electronically collecting data from a Semiconductor Equipment Communications Standard (SECS) data port and electronically sniffing data packets from various pieces of equipment in a semiconductor, data storage, or allied industry fabrication ("fab") facility.
2. Description of the Background Art
The semiconductor, data storage, and allied industries feature manufacturing lines that are rich in data production. Most pieces of process or metrology equipment ("tools") have a myriad of both generated and stored data. The data may include process conditions within a tool (e.g., process duration, process temperature, process gas flows, etc.), operating conditions of the tool (e.g., alarm states, input/output (I/O) signal traces, vacuum and pressure levels, etc.), or general historical data for the tool (e.g., last preventive maintenance (PM), next PM date, overall uptime, etc.). Tool data may also include various substrate feature measurements such as film thickness mapping, resistivity mapping, particle mapping, die-to-database correlations, step height values, line-width measurements, and so on. Data are typically available from both in-situ and ex-situ tools.
In the semiconductor industry, a fab tool is a piece of semiconductor fabrication equipment designed to process wafers (e.g., an ion implanter, a photolithographic stepper, a chemical vapor deposition system, etc.) or a piece of inspection equipment designed to measure or inspect wafers (e.g., a scanning surface inspection system, a critical-dimension scanning electron microscope, a spectroscopic ellipsometer, etc.). Frequently, a specialized fab tool, called a cluster tool, is also used in advanced fabrication facilities. A cluster tool is an integrated, environmentally isolated tool consisting of process, transport, and cassette modules mechanically linked together. A cluster tool module is an element of a cluster tool that performs particular functions, typically dedicated to a given process or portion of a process. A cluster tool module may also contain other modules.
Typically, there is at least one communication interface used to access data from various tools, the Semiconductor Equipment Communications Standard (SECS) protocol. SECS in one of many international consensus-based standards produced by SEMI (Semiconductor Equipment and Materials International), headquartered in San Jose, Calif. The SECS protocol communicates actively via a tool's serial port. Data available from a SECS protocol provides tool and material information such as wafer information, wafer lot number, cassette slot number being processed, recipe name, and process parameters.
A process or line engineer can use or analyze data from a tool to evaluate processing trends, view run-rules for a given process, or perform complex statistical calculations. However, to date, this type of analysis has been limited to a particular tool or a limited set of data available from a plurality of tools.
Statistical techniques are a type of data analysis for controlling a process that is well known in the
semiconductor, data storage, and allied industries. For example, statistical process control has been defined as "the use of statistical methods to analyze a process or its output to take appropriate actions to achieve and maintain a state of
5 statistical control and continuously improve the process capability" (SEMATECH Dictionary,). Appropriate actions may include monitoring or changing gas flows, temperatures, ramp rates, or the like. Other statistical techniques include using Shewhart charts (e.g., charting a group
10 mean versus standard deviation) for evaluating run-rules according to the well-known Western Electric sensitizing rules published in 1956.
However, beyond the basic data provided via a SECS port, the ability to collect real-time data from tools has been
15 limited. Many tools today are connected to a local area network (LAN) but currently available methods of extracting data from a tool over a LAN involve installing custom or proprietary software on the tool, sending data requests to a processor controlling the tool, waiting for acknowledg
20 ment of data requests from the tool's processor, preparing the data for spooling or streaming, and then collecting the data. This data collection scheme adds computational and timing overhead to the tool's processor in addition to added complexities and expenditures due to the custom or propri
25 etary software.
What is needed in the art is an improved means to obtain more and/or better data from a plurality of tools without affecting existing tool or fab installations.
30 SUMMARY OF THE INVENTION
The present invention is a system for electronically collecting data in a process line in a semiconductor or data storage fabrication facility. The system includes a data collection processing unit configured to sniff data packets
35 transmitted over a LAN between a plurality of tools in the fabrication facility. The data collection processing unit is further configured to determine if the sniffed data packets are valid data packets based on a configuration file established for the plurality of tools. The data collection processing unit
40 is further configured to parse one or more data fragments from the valid data packets.
The present invention is also a method for electronically collecting data from a tool in a semiconductor or data storage fabrication facility. The method includes sniffing data packets transmitted over a LAN by a plurality of tools located in the fabrication facility, determining if the sniffed data packets are valid data packets, and parsing one or more data fragments from the valid data packets.
50 BRIEF DESCRIPTION OF THE FIGURES
FIG. 1 is an exemplary overview diagram of an embodiment of the present invention for electronically collecting data through a SECS interface and a LAN interface;
55 FIG. 2 is an exemplary configuration of central processing units (CPU's) and modules within a tool showing coupling between the CPU's and the LAN interface;
FIG. 3 is a diagram showing an exemplary embodiment of an Ethernet packet containing data from tools;
go FIG. 4 is a flowchart of an exemplary method for validating and parsing electronically collected data;
FIG. 5A is a diagram showing exemplary tabular data received from the SECS interface and parsed data sniffed over the LAN; and
65 FIG. 5B is a diagram showing an exemplary embodiment of a unified data record produced by merging SECS data and parsed data.
DESCRIPTION OF PREFERRED
An electronic data collection system of the present invention collects data from one or more of a variety of tools found in a semiconductor, data storage, or allied industry 5 fabrication facility ("fab"). The present invention is a system and method for collecting operational, event, alarm, and related data from a tool's central processing unit (CPU) and/or module CPUs used for controlling the tools. Data are collected either via a SECS interface and a local-area 1° network (LAN) interface, or both. Collected data may then be distributed to a larger network (e.g., a wide-area network (WAN)) for storage, display, analysis, etc.
FIG. 1 is an exemplary embodiment of the present invention for electronically collecting tool data through a SECS 15 interface and/or a LAN interface. The exemplary embodiment is shown in the context of a fab environment 100. The fab environment 100 includes a LAN 101 and a plurality of fab tools 105A-M. Each of the plurality of fab tools 105A-M has an associated tool SECS interface 103A-«and a tool 20 LAN interface 107A-«. The FIG. 1 embodiment includes a data collection processing unit 111 and, in some embodiments, a WAN 115. In some embodiments, the data collection processing unit 111 has a processing unit SECS interface 109 and a processing unit LAN interface 113. 25
The LAN 101 is part of a communications network within the fabrication facility. In one exemplary embodiment, the LAN 101 allows for communications to occur using Ethernet packets transmitted between the plurality of fab tools 105A-«. 30
As discussed above, the plurality of fab tools 105A-M may be any instrument, machine, or device used to perform a task or measure a characteristic or quality of a substrate or the like. Any of the plurality of fab tools 105A-M may be a 3J cluster tool.
The tool SECS interface 103A-M is a physical communications port (e.g., an RS-232 serial port) running a particular communications protocol standard. The SECS protocol standard defines details for the electronic transmission of mes- 4Q sages between the plurality of fab tools 105A-M and the data collection processing unit 111. The tool SECS interface 103A-M allows a limited set of data relevant to the fab tool 105A-M to be transmitted electronically.
The processing unit SECS interface 109 may comprise a 45 plurality of SECS ports and communicates with the plurality of fab tools 105A-M utilizing the SECS protocol through the tool SECS interface 103A-M. Alternatively, the SECS interface 109 may be a multiplexed port having an appropriate multiplexing device. 50
Data pertaining to the fab tool 105A-M are made available for the data collection processing unit 111 via SECS protocol communications between the processing unit SECS interface 109 and the tool SECS interface 103A-M. Such data includes, for example, the fab tool 105A-M identifier, a 55 module identifier, wafer information, wafer lot number, cassette slot number being processed, recipe name, and process parameters.
Frequently, fab tools 105A-M are also connected to the LAN 101 through a network interface card. The network 60 interface card may be, for example, the tool LAN interface 107A-M. Typically, a tool will have one or more CPUs for controlling various tool functions (see FIG. 2). Each of the CPUs will be coupled to communicate electronically via the LAN 101. 65
The tool LAN interface 107A-M allows the fab tool 105A-M to communicate via the LAN 101. While in com
munication via the LAN 101, any of the plurality of fab tools 105A-M, or central processing units located within the plurality of fab tools 105A-M (see FIG. 2), are capable of communicating with any other fab tool 105A-M .
The data collection processing unit 111 is able to collect data from one or more of the fab tools 105A-M through the processing unit SECS interface 109 and/or the processing unit LAN interface 113. The communication and data collection process will be discussed in more detail below. Physically, the data collection processing unit 111 may be, for example, a personal computer, workstation, or network server.
In one embodiment of the present invention, the data collection processing unit 111 passively monitors data transmitted within the fab tool 105A-M as well between the plurality of fab tools 105A-M. Passive monitoring occurs by sniffing tool network traffic on the LAN (sniffing is also referred to as "snooping"). Sniffing is passive in that neither a request for data nor an acknowledgment of receipt of data is given. The CPUs (see FIG. 2) within fab tools 105A-M do not need to be aware that any information is gleaned from network traffic by the data collection processing unit 111. For example, a function for sniffing may be embedded in a network analyzer or in data collection processing unit 111 in order to sniff and decode any relevant data packets. Once relevant packets are detected, the packets are parsed and interpreted (see FIGS. 3 and 4) based on knowledge of the tool data format. Sniffing tool network traffic eliminates the need for custom or proprietary software loaded onto a tool and does not interfere with critical timing or add to the overhead of a tool's processor.
One or more configuration files will be stored, typically on data collection processing unit 111, for comparison with parsed information received from sniffed data packets. This process will be described in more detail with respect to FIGS. 3 and 4 below. The configuration files may include information such as CPU IP addresses, module numbers associated with the IP addresses, remote procedure call (RPC) data lengths, and so on.
The WAN 115 is shown electrically coupled to the data collection processing unit 111. However, as one skilled in the art would understand, the WAN 115 may alternatively be electrically coupled to other elements, such as, for example, to the LAN 101. Additionally, the WAN 115 may be within or extend past the confines of the fab environment 100.
Data collected from both the tool SECS interface 103A-M and the tool LAN interface 107A-« may then be merged, correlated, and/or be made available for additional analysis, display, or storage. The collected data may also be made available to a larger network (e.g., the WAN 115) or multiple clients or hosts. Data may be stored on one or more databases (potentially running under multiple operating systems), displayed in a summary format to a user's display, set up as a web page so data may be displayed, accessed, and used worldwide, or otherwise accessed or analyzed.
FIG. 2 is an exemplary internal schematic of one of the fab tools 105A-«showing various modules and CPUs. As shown, each of the CPUs 201-213 interface with the tool LAN interface 107 through LAN communications links 215A-G. Each of the CPUs 201-213 may be associated with a particular module of a multi-module cluster tool or, alternatively, multiple CPUs may be required for a single module of the fab tool 105. For example, the CPU 207 associated with a module C of the fab tool 105 has two associated CPUs, a module CI CPU 211 and a module C2 CPU 213. For purposes of the present discussion, individual