Abstract:
A processing tool (1) for manufacturing semiconductor devices (2), e.g. a lithography cluster, comprises a device transfer area (8) with an optical sensor (10), preferably a CCD-camera, and an illumination system (11) mounted within, such that a substrate (2) being transferred to or from one of its processing chambers (1a, 1b, 1c) can be scanned during its movement at low resolution. The substrate (2) may be either a semiconductor wafer to be manufactured or a reticle or mask used to perform an exposure on said wafer. The scanning is performed twice, prior and after processing in at least one the processing chambers (1a, 1b, 1c) of the processing tool (1). Both images are compared and optionally subtracted from each other. Defects imposed to the substrate due to contaminating particles only during the present processes with sizes larger than 10 microm are visible on the subtracted image, while defects imposed earlier are diminished as well as structures formed from e.g. a mask pattern below 10 microm. Pattern recognition allows an efficient classification of the defects being just detected in a processing tool (1). Thus, semiconductor device yield and metrology capacity are advantageously increased.
Abstract:
While a first leading semiconductor wafer (11) already processed in a process appliance (1) and belonging to a batch is being measured in a microscope measuring instrument (2) in relation to values for the structure parameters 30, a second or further semiconductor wafer (12) belonging to the batch is processed in the process appliance (1). An event signal (100) reports, for example, an inspection carried out successfully of the first wafer, so that the following wafers (12) no longer need to be inspected. Using the measured results, the process parameters (31) of the process appliance (1) are automatically readjusted. Events such as maintenance work or parameter drifts in trend maps etc. are detected in control units (8 or 9) and, via the output of an event signal (102), for example in an event database (40), lead to the event-based selection of structure parameters (30') to be measured and/or to the initiation of a leading wafer (11). Limiting-value violations (21) of at least one process parameter (31), detected by a control unit (8), are responded to by a warning signal (101) and likewise fed into the event database (40).
Abstract:
While the first forward processed semiconductor wafer (11) of a batch in a process device (1) is measured for values for structural parameters (30), a second or other semiconductor wafers (12) of the batch is/are processed in said process device (1). A result signal (100) indicates a successfully conducted inspection of the first wafer for example, whereupon the subsequent wafers (12) do not need to be examined. The process parameters (31) of the process device (1) are automatically adjusted according to the measuring results. Events such as maintenance or parameter drift in trend cards etc. are detected in control units (8 or 9) and lead to event-related selection of the structural parameters (30') and/or initiation of a forward wafer (11) by outputting an event signal (102) e.g. in an event data bank (40).A warning signal (101) is issued by a control unit (8) in response to detected threshold value infringements (21) for at least one process parameter (31) and is also fed into the event data base (40).
Abstract:
While a first leading semiconductor wafer (11) already processed in a process appliance (1) and belonging to a batch is being measured in a microscope measuring instrument (2) in relation to values for the structure parameters 30, a second or further semiconductor wafer (12) belonging to the batch is processed in the process appliance (1). An event signal (100) reports, for example, an inspection carried out successfully of the first wafer, so that the following wafers (12) no longer need to be inspected. Using the measured results, the process parameters (31) of the process appliance (1) are automatically readjusted. Events such as maintenance work or parameter drifts in trend maps etc. are detected in control units (8 or 9) and, via the output of an event signal (102), for example in an event database (40), lead to the event-based selection of structure parameters (30') to be measured and/or to the initiation of a leading wafer (11). Limiting-value violations (21) of at least one process parameter (31), detected by a control unit (8), are responded to by a warning signal (101) and likewise fed into the event database (40).
Abstract:
The plasma processing chamber (2) is coupled to a gas feed/vacuum generator (3) and the process parameters are controlled by an advanced process tool, APC, (1). The end point of the plasma cleaning operation is determined by monitoring the DC voltage bias signal (4) using a suitable transducer (5). The output signal is fed to a processor (6).