Colored and Timed Petri Net-based operation sequence control logic modeling of physical vapor deposition (PVD) equipment
Date Issued
2010
Date
2010
Author(s)
Chiou, Shian-Chiang
Abstract
Capital investment of a 300mm semiconductor fab has grown to 3-4 billions USD as the technology evolves. Equipment cost per fab may count up to 74% of the capital investment. Individual tool performance has therefore been even more critical to fab productivity and competitiveness than before. Analysis of Overall Equipment Effectiveness (OEE) and Intrinsic Equipment Effectiveness (IEE) of tools by semiconductor fabs indicate that un-optimized constrained cluster tool performance may contribute up to 40% of OEE loss and production time waste in IEE loss. Among the factors that affect cluster tool performance the sequencing of intra-tool operations is the most important to the throughput rate and utilization.
For the semiconductor industry, Key to the operation sequence efficiency of a cluster tool is the control of robot operations, which transfer wafers from one chamber to another for processing. By reviewing of the literature, we noticed that Petri-Net is widely used in the cluster tool modeling, many of them consist of production process modeling using Petri Net (Hu, 1999), and some focus on specific arms control logic modeling (Wu, 2006), But lack of integration between both of the those characteristics of the Petri Net modeling.
In this paper, a standard modeling process for cluster tool sequence optimization under given robot control logics is established. The process consists of firstly using Colored and Timed Petri Net (CTPN) model cluster tool and associated operations modules, then conversion of CTPN to mathematical optimization problem. A Physical Vapor Deposition (PVD) cluster tool is adopted as our study case. Our objective is to develop a standard modeling process for sequence optimization of cluster tools in general.
Modeling by using CTPN, the operations of a PVD cluster tool can be decomposed into three basic CTPN modules: Loadlock module, wafer process flow module and Transfer modules. Then, the Specific designing challenges are as following: C1) How to model control logics from the observed sequencing behaviors in the PVD tool such as Push, Pull, and first-in-first-out (FIFO) control logics? C2) How to synthesize all the modules into a complete sequencing model of the PVD tool? A two-step process is designed to address the two challenges.
In this research, we use semantic description methodology to interpret the arms control logic modules into CTPN model, and proposed a CTPN modular synthesis algorithm to generate and merge the PVD equipment operation module and arms control logic modules as a complete CTPN model, Finally using CPN tools R simulation environment to simulate the PVD cluster tool model and validate the result by comparing with the actual machine performance, in this case we analyzed with following performance index: (1) wafer output time interval (2) the transporting arm’s trigger time (3) the occupation time.
Subjects
CTPN
Petri Net
Cluster tool
Robot control logic
Synthesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-99-R97546004-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):9c018df1b229d2337fa94abaf6ae3047