The present invention relates generally to the monitoring of advanced process control systems and more particularly to a system, method, and medium of monitoring the performance of a process output during semiconductor manufacture in relation to specification limits and monitoring the accuracy of a model that predicts the performance of the process output.
Semiconductor manufacture is becoming an increasingly automated process requiring precise methods of process control to ensure a quality output. Since the process is automated, safeguards are required to protect the manufacturing system and acknowledge when the manufacturing system, or tool, is unstable and is performing poorly.
Typically, there are only several factors that are measured during the semiconductor manufacturing process, for example, thickness of a film after the film has been deposited, polished, and/or etched. Because of this, occasionally there will be situations in which the tool performance changes due to factors that are not directly measured. For example, one factor that is not directly monitored that can cause failures in a semiconductor device is an increased amount of particles on the wafer, where the increase in particles is caused by an increase in the pressure in a chamber where the manufacturing process is being performed. As the manufacturing process is designed, an experiment may be conducted to determine how many particles are introduced based on various levels of pressure. Since particles cannot be measured while the manufacturing process is executing, the designer must assume that the model is correct.
For situations in which there is no automated control, changes in the process performance as a result of errors in the model may be directly observed in the wafer properties. In the particles example, when the pressure in the chamber increases, the increase in the amount of particles on the wafer may be directly observed as a change in the thickness of the wafer. A human controller would notice the change in the wafer thickness and, in examining the process to determine the source of the thickness increase, would notice that the pressure had changed. The human controller would also perhaps notice that the change in pressure had caused the increase in the number of damaging particles on the wafer.
When advanced process control (“APC”) techniques are applied, however, the APC methodology attempts to compensate for any changes in the manufacturing process and such changes may not be as easily observed. In the particles example, the thickness of the wafer is regulated, such that, when the model has predicted perfectly the required pressure in the chamber, as the pressure changes during execution of the process the thickness of the wafer does not change. However, when the model is not correctly predicting the behavior of the process, these pressure variations may cause an increase in particles to occur. However, although particles are being introduced and are damaging the wafer, the APC will not automatically detect these variations in pressure (i.e. the APC only detects an increase in the thickness of the wafer).
Thus, the use of advanced process control methods demonstrates a need for examining the behavior of the process in the context of a process that is being controlled. Two types of monitoring techniques, for example, process health monitoring and model health monitor, are often used to fulfill this need.
Process health monitoring may be used to effectively monitor, for example, an automated process that is under computer control. Process health monitoring detects deviation of controlled outputs of the process, or tool, away from some predetermined target area. Process health monitoring may, itself, be an automated procedure. Process health monitoring methods provide high-level information for, for example, each controlled output of a process. For example, process health monitoring may be applied to chemical mechanical planarization (“CMP”), chemical vapor deposition (“CVD”), etching, electrochemical plating processes, (“ECP”), physical vapor deposition (“PVD”), etc. Such monitoring is accomplished by taking measurements of the process parameters that are of concern, then, to perform statistical analysis of those measurements, and, finally, to compare the statistical analysis to desired limits. Thus, a determination is made as to whether any specified controlled output(s) has strayed too far from a predetermined target.
Model health monitoring, which may be used to monitor each run-to-run (“R2R”) control model for CMP, CVD, ECP, PVD, etc., detects deviation of, for example, the R2R model from the actual behavior of the process, or tool. Model health monitoring also may be an automated procedure. In the case of model health monitoring, the statistical analyses may include such pertinent information as model predictions and necessary previous data to perform these model predictions. Health monitoring may, itself, be an automated procedure.
Prior methods of process and model health monitoring employed indices relating to such monitoring. However, prior methods of monitoring were used for continuous processes such as, for example, controller performance monitoring. Controller performance monitoring looks at a desired, best controller performance based on specific data, which are calculated using time series analysis, and takes a ratio of a current variance to the minimum variance controller performance. However, unlike with semiconductor manufacturing processes, controller performance monitoring takes into account only the continuous process, rather than monitoring distinct points in the process.
A continuous process, in general, refers to a process that is run in a mode where things are constantly coming in and constantly going out. A simple example is a tank that has fluid coming in and fluid going out. In a continuous process, the goal is to continually maintain the process in a certain state. For example, in the case of the tank, the goal would be to control the rate at which fluid is being pumped into, or out of, the tank such that the level of fluid in the tank is maintained at a constant level.
Controller performance monitoring is performed using minimum variance control theory for systems that have dynamics. In other words, the dynamic process is monitored only to determine what factors are affecting the maintenance of the continuous, on-going process. Prior methods of process and model health monitoring made use of the dynamic equations that are used to do control of such continuous processes.
In contrast, semiconductor processes are usually modeled as static processes for the purposes of run-to-run control. Rather than the manufacturing of a wafer being a continuous process, once a wafer is completed, the process is over. The process, itself, is repeated for subsequent wafers without being altered. A static, or discrete, process such as manufacturing a wafer can only be monitored in terms of how the process performed for prior, discrete manufacturing occurrences. An action in a static process (for example, a deposition time change or change in polish time), which occurred on the previous three wafers, may not have much of an effect on the processing of the subsequent wafer. Such static processes lack the dynamic equations used to model continuous process and, therefore, the model and process health monitoring techniques utilized for continuous process cannot be employed in monitoring static processes, e.g., semiconductor manufacturing.
What is desired is a method and system that allows a controller to monitor the performance of a static manufacture process during the entire cycle of the process such as to maintain the performance of the process as the process is repeated.
The present invention addresses the lack in the prior art described above by providing techniques to monitor static processes and to quantize the results of the monitoring with one or more indices. In the case of model health monitoring, the one or more indices can be used to monitor the performance of the process controller. In the case of process health monitoring, the one or more indices can be used to monitor the performance of the process, itself. An index, for example, a number, that characterizes the performance of either the controller or the process provides an “at-a-glance” metric that provides information as to whether or not the controller or the process is performing within acceptable limits. The purpose of the one or more indices is to provide some notification to, for example, a human operator that something in the on-going manufacturing process requires attention.
In general, embodiments of the present invention contemplate that model health monitoring for a static process, such as semiconductor manufacturing, estimates a variance of a specific controlled output over time and, then, benchmarks, or compares, this variance with an expected variance. Based upon this comparison of actual variance to the expected variance, an estimate is provided of how well the process is being controlled or how well the model is able to predict the behavior of the process and thus able to control the process. The result of this estimate is then calculated as a single, model health index. In one or more embodiments of the present invention, the model health index may also be used to trigger some sort of notification function if the controller and/or the model is not operating within acceptable limits or seems in danger or doing operating outside of these acceptable limits.
In general, embodiments of the present invention contemplate that process health monitoring for a static process, such as semiconductor manufacturing, estimates not only the variance of a specific controlled output over time, but also a bias, the difference between the actual specific controlled output and a target output. The estimated bias and the estimated variance is then used to construct a probability distribution of how likely it is that the controlled output will be within some desired performance limits. Based upon this probability distribution, a single, process health index can be calculated that represents this likelihood. In other embodiments of the present invention, the process health index may also be used to trigger some sort of notification function if the process is not operating within acceptable limits or seems in danger or doing operating outside of these acceptable limits.
In monitoring static processes, such as semiconductor wafer manufacture, the model and/or process health index also gives an indication of the entire manufacturing process for a number of wafers by indicating whether an error in the manufacture of a particular wafer, for example, a wafer whose thickness is effected by a build up of particles, is due to an actual defect in the model and/or the process, for example, a change in the pressure of the manufacturing chamber, or whether the error is an isolated, non-representative fluke, such as a bad wafer. The calculation of the model and/or process health index ideally provides a filtering mechanism by which isolated errors are not reflected in the index by determining whether a particular controlled output is non-representative outlier value.
It is one feature and advantage of the present invention to monitor the performance of a process model using one or more indices.
It is another feature and advantage of the present invention to monitor the performance of a process output using one or more indices.
It is another feature and advantage of the present invention to monitor the performance of multiple process models using one or more indices.
It is another feature and advantage of the present invention to monitor the performance of multiple process outputs using one or more indices.
These and other features and advantages of the present invention are achieved in a method for monitoring performance of an advanced process control system for at least one static process output. One or more embodiments of the present invention includes a method for monitoring performance of an advance process control system for at least one process output, which includes receiving process performance data for the at least one static process output and comparing the process performance data to a predicted value for the process performance and/or a target value for the process performance. The method also includes calculating at least one index that reflects comparison of the process performance data to the predicted value and/or the target value for the process performance. The method further includes indicating the results of the calculation on the at least one index. Indicating the results includes sending an indication to a controller that the at least one index is beyond an acceptable point, halting processing of the at least one process output if the at least one index is beyond an acceptable point, and/or storing the at least one index as an indication of the processing performance of the at least one process output.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for at least one process output that includes receiving process performance data for the at least one process output and then calculating a model health index and/or a process health index. The model health index indicates an estimate of an ability of a model to predict the behavior of the at least one process output as compared to an expected output. The process health index indicates an estimated probability of violation by the at least one process output of predefined specification limits. The method also includes indicating the results of the calculation based on the at least one of the model health index and the process health index.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for at least one process output that includes calculating a variance of a prediction error for a processing performance of the at least one process output and/or a probability for violating specification limits of the processing performance of the at least one process output. The variance and probability are based on an exponentially weighted moving average.
If the variance of the prediction error is calculated, the method also includes calculating a model health index. The model health index is a ratio of an exponentially weighted moving average-based estimate of a standard deviation of the prediction error to an expected estimate of the prediction error. The exponentially weighted moving average-based estimate of the standard deviation of the prediction error is derived from the variance of the prediction error.
If the probability for violating specification limits is calculated, the method further includes calculating a process health index. The process health index is a ratio of the probability for violating the specification limits to a specified probability limit. The method also includes indicating the results of the calculation based on the model health index and/or the process health index.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for at least one process output that includes receiving process performance data for the at least one process output and calculating a current model health index or a current process health index. The current model health index indicates an estimate of an ability of a model to predict the behavior of a current one of the at least one process output as compared to an expected output. The current process health index indicates an estimated probability of violation by a current one of the at least one process output of predefined specification limits.
If the current model health index is calculated, the method also includes calculating a subsequent model health index, which indicates an estimate of an ability of a model to predict the behavior of a subsequent one of the at least one process output as compared to an expected output. If the subsequent model health index is calculated, the method further includes storing the current model health index and the subsequent model health index, such that comparing the current model health index and the subsequent model health index give an indication of a processing performance of the at least one process output.
If the current process health index is calculated, the method also includes calculating a subsequent process health index, which indicates an estimated probability of violation by a subsequent one of the at least one process output of predefined specification limits. If the subsequent process health index is calculated, the method further includes storing the current process health index and the subsequent process health index, such that comparing the current process health index and the current process health index gives an indication of the processing performance of the at least one process output.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for at least one process output that includes calculating a current variance of a prediction error for a processing performance of the at least one process output and/or a current probability for violating specification limits of the processing performance of the at least one process output. The current variance and the current probability are based on an exponentially weighted moving average.
If the current variance of the prediction error is calculated, the method also includes calculating a current model health index. The current model health index is a ratio of a current exponentially weighted moving average-based estimate of a standard deviation of the prediction error to an expected estimate of the prediction error. The current exponentially weighted moving average-based estimate of the standard deviation of the prediction error is derived from the current variance of the prediction error.
If the current model health index is calculated, the method further includes calculating a subsequent model health index, which is calculated in a substantially similar manner to the current model health index. If the subsequent model health index is calculated, the method also includes storing the current model health index and the subsequent model health index, such that comparing the current model health index and the subsequent model health index gives an indication of the processing performance of the at least one process output.
If the current probability for violating specification limits is calculated, the method further includes calculating a current process health index. The current process health index is a ratio of the probability for violating the specification limits to a probability limit.
If the current process health index is calculated, the method also includes calculating a subsequent process health index, which is calculated in a substantially similar manner to the current process health index. If the subsequent process health index is calculated, the method further includes storing the current process health index and the subsequent process health index, such that comparing the current process health index and the subsequent process health index give an indication of the processing performance of the at least one process output.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for a plurality of process outputs that includes calculating a first model health index of a process performance of a first one of the plurality of process outputs and/or a first process health index of the process performance of the first one of the plurality of process outputs. The method also includes calculating a second model health index of the process performance of a second one of the plurality of process outputs and/or a second process health index of the process performance of the second one of the plurality of process outputs.
If the first model health index and the second model health index are calculated, the method further includes calculating an aggregate model health index of the process performance of the plurality of process outputs. If the first process health index and the second process health index are calculated, the method also includes calculating an aggregate process health index of the process performance of the plurality of process outputs.
One or more embodiments of the present invention also includes a method for monitoring performance of an advanced process control system for at least one process output that includes estimating a process deviation. The process deviation indicates deviation of a process performance from a target process performance and/or a model of the process performance. The method also includes characterizing a current estimate of the process performance using a first index that represents the deviation of the process performance from the target process performance and/or a second index that represents the deviation of the model performance from a specified model performance. The method further includes performing a notification function based on the value of the first index and/or the second index.
The features and advantages of the present invention are also achieved in a system for monitoring performance of an advanced process control system for at least one process output. The system includes a first memory that stores a predicted value for process performance of the at least one process output and/or a target value for process performance of the at least one process output. The system also includes a second memory that stores process performance data of the at least one process output and a third memory that stores at least one of a model health algorithm and a process health algorithm. The model health algorithm is used to calculate a model health index of the process performance and the process health algorithm is used to calculate a process health index of the process performance. The system further includes a processor that calculates the model health index using the model health algorithm and/or the process health index using the process health algorithm. The model health index is calculated based on a comparison of the predicted value and the process performance data of the at least one process output. The process health index is calculated based on a comparison of the target value and the process performance data of the at least one process output.
One or more embodiments of the present invention also includes a system for monitoring performance of an advanced process control system for at least one process output that includes one or more tools, which measure the at least one process output. The system also includes a controller, coupled to tool(s), which provides for central control of the tool(s).
The controller implements instructions for controlling the tool(s), including: estimating a process deviation, which indicates deviation of a process performance from a target process performance and/or a model of the process performance; characterizing a current estimate of the process performance using a first index that represents the deviation of the process performance from the target process performance and/or a second index that represents the deviation of the model performance from a specified model performance; and performing a notification function based on the value of the first index and/or the second index.
The features and advantages of the present invention are also achieved in a computer-readable medium of instruction for monitoring performance of an advanced process control system for at least one static process output. The instruction includes, receiving process performance data for the at least one process output and comparing the process performance data a predicted value for the process performance and/or a target value for the process performance. The instruction also includes calculating at least one parameter that reflects comparison of the process performance data to the predicted value for the process performance and/or the target value for the process performance. The method also includes indicating the results of the calculation based on the at least one parameter.
One or more embodiments of the present invention also includes a computer-readable medium of instruction for monitoring performance of an advanced process control system for at least one static process output. The instruction includes, receiving process performance data for the at least one process output and calculating a model health index and/or a process health index. The model health index indicates an estimate of an ability of a model to predict the behavior of the at least one process output as compared to an expected output. The process health index indicates an estimated probability of violation by the at least one process output of predefined specification limits.
Reference now will be made in detail to various embodiments of the present invention. Such embodiments are provided by way of explanation of the invention and are not intended to be limited thereto. In fact, those of ordinary skill in the art may appreciate upon reading the present specification and viewing the present drawings that various modifications and variations can be made.
Monitoring the process and model health of the tool allows a controller, whether a human controller or some automated controller, to evaluate the performance of the tool. For example, alarms and warnings, which can be triggered by a decrease in either the process or model health of the tool, can be configured such that the monitoring system can stop the tool, for example, if the process and/or model health goes beyond a certain limit. Alternatively, if the process and/or model health goes beyond, for example, some less severe limit, the controller may be notified by, for example, an e-mail, a page, or by a message send to a personal data assistant (“PDA”).
I. Model Health Monitoring
A methodology for obtaining σactual is discussed below.
The length of the process history, which is involved in the calculation of the model health, is determined by a specific EWMA coefficient, λ, which is a given value (step 202). It should be noted that “given” values, as described herein, may be determined by experience or through direct measure, as generally known to those skilled in the art. A predicted value for the controlled output of the semiconductor wafer is determined based on a process model (step 204), as described below in greater detail.
The calculation of the model health index is an iterative process that is performed by considering the difference between an actual value of a controlled output of, for example, a semiconductor wafer k+1,
and a predicted value for the controlled output of the semiconductor wafer k+1,
where “k” indicates the wafer number (step 206). If this difference is larger than an estimate of the standard deviation of the prediction error for the previous wafer k, σkactual (step 208), multiplied by a factor, K, indicating that the model is significantly different from the actual controlled output of the tool for that wafer, then that controlled output is ignored as a flier, or a non-representative outlier value, which is not representative of the controller output of the tool. The initial value of σkactual, σ0actual, is given (step 202).
Since the process for determining the model health relies upon the most recently received information, more accurate values for σkactual are “learned” by continuously gathering information from the process. Because the initial value of the standard deviation is assumed, the first several values of σkactual will most probably not accurately reflect the actual standard deviation of the process, and thus, the difference between
will almost always be greater than σkactual. Therefore, screening for outliers, or flier, which involves a value, K, and the current estimate of the standard deviation of the prediction error, σkactual, does not occur until the wafer number is greater than some specified wafer number, NW (step 210). The initial estimation of the standard deviation of the prediction error, thus, always is used for the first several wafers, up to some wafer number Nw, which is a given value (step 202). The previous considerations can be summarized by the following conditional equation:
where K is a given coefficient (step 202).
If the condition holds true, an EWMA estimate of prediction error variance for the wafer k+1, Dk+1, is calculated (step 218). In general, D is an estimate of the predication error variance, or in other words, an estimate of the difference in an expected controlled output for a wafer and the actual controlled output for that wafer, that is calculated for a number of different wafers that undergo the process. Dk+1 is based on the difference between
and
the length of the process history, and the estimate of the prediction error variance for the previous wafer k, Dk. In one or more embodiments of the present invention, this can be calculated as follows:
where D0 is given (step 202). Of course, it should be understood that other specific ways to calculate Dk+1 are also contemplated.
The estimate of the standard deviation of the prediction error for the wafer k+1,
is then calculated (step 220) as follows:
The model health index, IM, for the wafer k+1, as contemplated by one or more embodiments of the present invention, can then be calculated (step 222) as described previously:
In the next iteration, wafer k+1 becomes wafer k (step 224) and the values Dk+1 and
become Dk and
respectively.
If the condition does not hold true (from Equation 1), the controlled output for wafer k+1,
is considered a non-representative outlier value, or a flier. Therefore, the values of Dk+1 and
are not calculated (step 212) but are taken as the values of Dk and σkactual, respectively. Thus, the model health index for wafer k+1, Ik+1M, remains the same as the model health index for the previous wafer k, IkM (step 214). Wafer k+1 then becomes wafer k (step 216). It should be understood that other specific ways to calculate IM are also contemplated by one or more embodiments of the present invention.
The complete dynamic calculation of Ik+1M can be summarized as follows:
In developing the model, for example for a sub-atmospheric chemical deposition (“SACVD”) process, film properties of interest 412 are identified and outcome determinative processing model variables 414 are selected for the model, as illustrated schematically in
Regardless of the type of film substance for which a model is created, to obtain DOE data, an experiment is run which perturbs or varies the values of the processing variables of interest about a center point (or median value). One or more processing variables can be varied. The film properties of interest in the resultant film are measured for each combination of inputs. Data can be acquired empirically, by carrying out a series of experiments over a range of values of the processing variables. The data is fit to the appropriate curve (linear or non-linear) to define the model.
II. Process Health Monitoring
versus a probability limit, Prlimit, where Prlimit is a given value, for example, 95%. If the tool is performing beyond the predetermined specification limits, indicating that the process is performing poorly,
will be greater than Prlimit. If, however, the tool is performing within the predetermined specification limits, indicating that the process is performing within acceptable limits,
will be less than Prlimit. The value of Ik+1P is, for example, limited between 0 and 1, with a higher value indicating acceptable performance of the process. Thus, Ik+1P is calculated as follows:
A methodology for obtaining
is discussed below.
The length of the process history, which is involved in the calculation of the process health, is determined by a specific EWMA coefficient, λ, which is a given value (step 502). As stated previously, it should be noted that “given” values, as described herein, may be determined by experience or through direct measure, as generally known to those skilled in the art. A set of specification limits for the controlled output of the processed semiconductor wafer is determined based on a desired performance of the tool (step 504).
The probability that a controlled output of the tool will be within with certain specification limits can be modeled using a standard normal distribution bell curve, as illustrated in
where Prk+1 is a value for Prcalc for a wafer k+1.
Referring back to
and target value for the controlled output of the semiconductor wafer k+1,
where “k” indicates the wafer number (step 506):
If this difference is larger than an estimate of the standard deviation of a target deviation for the previous wafer k, σkactual, multiplied by a factor K (step 508), indicating that the actual controlled output of the tool for that wafer is significantly different from target controlled output, then that controlled output is ignored as a flier, or a non-representative outlier value, which is not representative of the controlled output of the tool. The initial value of σkactual, σ0actual, is given (step 502).
As described previously, more accurate values for σkactual are “learned” by continuously gathering information from the process. Since the initial value of the standard deviation is assumed, the first several values of σkactual will most probably not accurately reflect the actual standard deviation of the process, and thus, the difference between
will always be greater than σkactual. Therefore, screening for outliers, or fliers, which involves a value, K, and the current estimate of the standard deviation of the target deviation, σkactual, does not occur until the wafer number, NW, is greater than some specified wafer number (step 510). The initial estimation of the standard deviation, thus, always is used for the first several wafers, up to some wafer number NW, which is a given value (step 502). The previous considerations can be summarized by the following conditional equation:
if |Δk+1|≦K·σkactual OR k+1≦NW, (Equation 7)
where K is a given coefficient (step 502).
If the condition holds true, an EWMA estimate of target deviation variance for the wafer k+1, Dk+1, is calculated (step 512). In this case, D is an estimate of target deviation variance, or in other words, an estimate of the difference in a target controlled output for a wafer and the actual controlled output for that wafer that is calculated for a number of different wafers that undergo the process. Dk+1 is based on the difference between
the length of the process history, and the estimate of the target deviation variance for the previous wafer k, Dk+1. In one or more embodiments of the present invention, this can be calculated as follows:
Dk+1=λ(Δk+1)2+(1−λ)Dk, (Equation 8)
where D0 is given (step 502). Of course, it should be understood that other specific way to calculate Dk+1 are also contemplated.
The estimate of the standard deviation of the target deviation for the wafer k+1, σk+1, is then calculated (step 514) as follows:
σk+1=√{square root over (Dk+1)}. (Equation 9)
Next, an EWMA estimate of the target deviation mean for wafer k+1, Mk+1, is calculated (step 516) in one or more embodiments of the present invention as follows:
Mk+1=λΔk+1+(1−λ)Mk, (Equation 10)
where M0 is given (step 502).
Finally,
for wafer k+1 is calculated (step 518). In general, Prcalc represents an estimate of the probability that a controlled output for a wafer will violate some predetermined, desired specification performance limits and is calculated for a number of different wafers that undergo the process. In one or more embodiments of the present invention, this is calculated as follows:
where Pr0calc is given (step 502), such that
is based upon the estimated probability for violating specification limits for wafer k, Prkcalc, a predetermined upper specification limit (“USL”), a predetermined lower specification limit (“LSL”), and a normally distributed variable with the previously described “bell-curve” distribution (
where “x” represents either USL or LSL, as appropriate, and “erfc” is a complementary error function. Of course, it should be understood that other specific way to calculate Prk+1calc are also contemplated.
The process health index, Ik+1P, for the wafer k+1, as contemplated by one or more embodiments of the present invention, can then be calculated (step 520) as described previously:
In the next iteration, wafer k+1 becomes wafer k (step 522) and the values Dk+1, σk+1, and Mk+1, and
become Dk, σk, Mk, and Prkcalc, respectively.
If the condition does not hold true (from Equation 7), the controlled output for wafer k+1,
is considered a non-representative outlier value, or a flier. Therefore, the values of Dk+1, σk+1, and Mk+1, and
are not calculated (step 524) but are taken as the values of Dk, σk, Mk, and Prkcalc, respectively. Thus, the process health index for wafer k+1, Ik+1P, remains the same as the process health index for the previous wafer k, IkP (step 526). Wafer k+1 then becomes wafer k (step 528). It should be understood that other specific ways to calculate IP are also contemplated by one or more embodiments of the present invention.
The complete dynamic calculation of Ik+1P can be summarized as follows:
In general, both process and model health monitoring can be used to gain insight into the health of a process, although they provide different levels of analysis. As stated previously, the process health index provides an indication of how well the actual process is performing while the model health index provides an indication of whether or not the configuration of the process controller should be modified. Therefore, performing model health monitoring in addition to process health monitoring provides further information, which allows for increased refining and improvement of the control of the process.
III. Higher Level Health Monitoring
An advanced process control system may have multiple controlled outputs and also may have multiple process descriptors, which indicate different layers of a semiconductor product, as well as different products. For example, a particular tool may have multiple chambers, or resources, which are essentially places to process. On a single wafer, parameters that are typically controlled include, for example: average thickness, thickness uniformity, and dopant concentration. Although each controlled output or process descriptor can be monitored with a separate model and/or process health index, it may be more efficient to create a single, aggregate model health index and/or a single, aggregate process health index that characterizes either the health for a specific process descriptor for a specific chamber or processing station, the health of the entire processing system for a specific process descriptor, or the health of the entire system for all process descriptors.
The higher-level health-monitoring, or aggregate, index for model health and/or process health is calculated, according to one or more embodiments of the present invention, using a mean of the indices. One type of mean is a geometric mean, calculated as follows:
where N is the total number of controlled outputs being monitored in the system and yi indicates a particular controlled output. Other, alternate methods of calculating the aggregate index may be used. It should be noted that a single, aggregate index can be calculated to monitor either the process health of the multiple outputs or the model health of the multiple output but not both the process health and model health.
IV. Health Tracking and Notification
Display 850 illustrates a graphical display of model health indices for seven controlled outputs 852, 854, 857, 858, 860, 862, and 864. By contrast with display 800, in this example, display 850 shows that the models used to predict the actual values of the controlled outputs is performing poorly, as indicated by the low model health index values.
The model and/or process health index can be used to perform several notification functions of the health of the tool or system. For example, if the model and/or process health index drops below some predetermined threshold, e.g., below 0.4, a notification message may be sent to a human controller using, for example, an e-mail, a page, or a wireless PDA. Other notification methods are, of course, possible. Likewise, notification can be sent to a computerized controller, where the computerized controller may respond by raising some warning flag. If the model and/or process health index drops below some critical point, the human or computerized controller may respond by shutting down the system, and thus halting processing of the tool until the cause of the health degradation can be located and remedied.
The model and/or process health index may also be used to generally track the overall health of the system. The various iterative values of the model and/or process health index may be stored in some memory either as a single instance of health (
V. Computer Implementation
Various aspects of the present invention that can be controlled by a computer can be (and/or be controlled by) any number of control/computer entities, including the one shown in
Read only memory (ROM) 1020 and random access memory (RAM) 1018 constitute the main memory of the system. As contemplated by the present invention, a number of parameters, including, for example, the actual data from the controlled output, the model data indicating a predicted value for the controlled output as used to calculated the model health index, the specification data indicating a target value for the controlled output, and the actual model and/or process health indices as they are calculated, may be stored in the main memory of the system. Therefore, any number of ROM 1020 and/or RAM 1018 may be included in the system to accommodate storage of these parameters. Additionally, the instructions for calculating the model and/or process health indices may also be stored in these main memories.
Disk controller 1022 interfaces one or more disk drives to the system bus 1002. These disk drives are, for example, floppy disk drives 1026, or CD ROM or DVD (digital video disks) drives 1024, or internal or external hard drives 1028. These various disk drives and disk controllers are optional devices.
A display interface 1014 interfaces display 1012 and permits information from the bus 1002 to be displayed on display 1012. Display 1012 can be used in displaying a graphical user interface. Communications with external devices such as the other components of the system described above can occur utilizing, for example, communication port 1016. Optical fibers and/or electrical cables and/or conductors and/or optical communication (e.g., infrared, and the like) and/or wireless communication (e.g., radio frequency (RF), and the like) can be used as the transport medium between the external devices and communication port 1016. Peripheral interface 1010 interfaces the keyboard 1006 and mouse 1008, permitting input data to be transmitted to bus 1002. In addition to these components, system 1000 also optionally includes an infrared transmitter and/or infrared receiver. Infrared transmitters are optionally utilized when the computer system is used in conjunction with one or more of the processing components/stations that transmits/receives data via infrared signal transmission. Instead of utilizing an infrared transmitter or infrared receiver, the computer system may also optionally use a low power radio transmitter 1032 and/or a low power radio receiver 1030. The low power radio transmitter transmits the signal for reception by components of the production process, and receives signals from the components via the low power radio receiver. The low power radio transmitter and/or receiver are standard devices in industry.
Although system 1000 in
In general, it should be emphasized that various components of embodiments of the present invention can be implemented in hardware, software or a combination thereof. In such embodiments, the various components and steps are implemented in hardware and/or software to perform the functions of the present invention. Any presently available or future developed computer software language and/or hardware components can be employed in such embodiments of the present invention. For example, at least some of the functionality mentioned above could be implemented using C, C++, visual basic, Java, or any assembly language appropriate in view of the processor(s) being used. It could also be written in an interpretive environment such as Java and transported to multiple destinations to various users.
The many features and advantages of the invention are apparent from the detailed specification, and thus, it is intended by the appended claims to cover all such features and advantages of the invention, which fall within the true spirit and scope of the invention. Further, since numerous modifications and variations will readily occur to those skilled in the art, it is not desired to limit the invention to the exact construction illustrated and described, and accordingly, all suitable modifications and equivalence may be resorted to, falling within the scope of the invention.
It is to be understood that the invention is not limited in its application to the details of construction and to the arrangements of the components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced and carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein are for the purpose of description and should not be regarded as limiting.
Number | Name | Date | Kind |
---|---|---|---|
3205485 | Noltingk | Sep 1965 | A |
3229198 | Libby | Jan 1966 | A |
3767900 | Chao et al. | Oct 1973 | A |
3920965 | Sohrwardy | Nov 1975 | A |
3982440 | Groleau et al. | Sep 1976 | A |
4000458 | Miller et al. | Dec 1976 | A |
4207520 | Flora et al. | Jun 1980 | A |
4209744 | Gerasimov et al. | Jun 1980 | A |
4302721 | Urbanek et al. | Nov 1981 | A |
4368510 | Anderson | Jan 1983 | A |
4609870 | Lale et al. | Sep 1986 | A |
4616308 | Morshedi et al. | Oct 1986 | A |
4663703 | Axelby et al. | May 1987 | A |
4698766 | Entwistle et al. | Oct 1987 | A |
4750141 | Judell et al. | Jun 1988 | A |
4755753 | Chern | Jul 1988 | A |
4757259 | Charpentier | Jul 1988 | A |
4796194 | Atherton | Jan 1989 | A |
4901218 | Cornwell | Feb 1990 | A |
4938600 | Into | Jul 1990 | A |
4957605 | Hurwitt et al. | Sep 1990 | A |
4967381 | Lane et al. | Oct 1990 | A |
5089970 | Lee et al. | Feb 1992 | A |
5108570 | Wang | Apr 1992 | A |
5208765 | Turnbull | May 1993 | A |
5220517 | Sierk et al. | Jun 1993 | A |
5226118 | Baker et al. | Jul 1993 | A |
5231585 | Kobayashi et al. | Jul 1993 | A |
5236868 | Nulman | Aug 1993 | A |
5240552 | Yu et al. | Aug 1993 | A |
5260868 | Gupta et al. | Nov 1993 | A |
5270222 | Moslehi | Dec 1993 | A |
5283141 | Yoon et al. | Feb 1994 | A |
5295242 | Mashruwala et al. | Mar 1994 | A |
5309221 | Fischer et al. | May 1994 | A |
5329463 | Sierk et al. | Jul 1994 | A |
5338630 | Yoon et al. | Aug 1994 | A |
5347446 | Iino et al. | Sep 1994 | A |
5367624 | Cooper | Nov 1994 | A |
5369544 | Mastrangelo | Nov 1994 | A |
5375064 | Bollinger | Dec 1994 | A |
5398336 | Tantry et al. | Mar 1995 | A |
5402367 | Sullivan et al. | Mar 1995 | A |
5408405 | Mozumder et al. | Apr 1995 | A |
5410473 | Kaneko et al. | Apr 1995 | A |
5420796 | Weling et al. | May 1995 | A |
5427878 | Corliss | Jun 1995 | A |
5444837 | Bomans et al. | Aug 1995 | A |
5469361 | Moyne | Nov 1995 | A |
5485082 | Wisspeintner et al. | Jan 1996 | A |
5490097 | Swenson et al. | Feb 1996 | A |
5495417 | Fuduka et al. | Feb 1996 | A |
5497316 | Sierk et al. | Mar 1996 | A |
5497381 | O'Donoghue et al. | Mar 1996 | A |
5503707 | Maung et al. | Apr 1996 | A |
5508947 | Sierk et al. | Apr 1996 | A |
5511005 | Abbe et al. | Apr 1996 | A |
5519605 | Cawlfield | May 1996 | A |
5525808 | Irie et al. | Jun 1996 | A |
5526293 | Mozumder et al. | Jun 1996 | A |
5534289 | Bilder et al. | Jul 1996 | A |
5541510 | Danielson | Jul 1996 | A |
5546312 | Mozumder et al. | Aug 1996 | A |
5548535 | Zvonar | Aug 1996 | A |
5553195 | Meijer | Sep 1996 | A |
5586039 | Hirsch et al. | Dec 1996 | A |
5599423 | Parker et al. | Feb 1997 | A |
5602492 | Cresswell et al. | Feb 1997 | A |
5603707 | Trombetta et al. | Feb 1997 | A |
5617023 | Skalski | Apr 1997 | A |
5627083 | Tounai | May 1997 | A |
5629216 | Wijaranakula et al. | May 1997 | A |
5642296 | Saxena | Jun 1997 | A |
5646870 | Krivokapic et al. | Jul 1997 | A |
5649169 | Berezin et al. | Jul 1997 | A |
5654903 | Reitman et al. | Aug 1997 | A |
5655951 | Meikle et al. | Aug 1997 | A |
5657254 | Sierk et al. | Aug 1997 | A |
5661669 | Mozumder et al. | Aug 1997 | A |
5663797 | Sandhu | Sep 1997 | A |
5664987 | Renteln | Sep 1997 | A |
5665199 | Sahota et al. | Sep 1997 | A |
5665214 | Iturralde | Sep 1997 | A |
5666297 | Britt et al. | Sep 1997 | A |
5667424 | Pan | Sep 1997 | A |
5674787 | Zhao et al. | Oct 1997 | A |
5694325 | Fukuda et al. | Dec 1997 | A |
5695810 | Dubin et al. | Dec 1997 | A |
5698989 | Nulman | Dec 1997 | A |
5719495 | Moslehi | Feb 1998 | A |
5719796 | Chen | Feb 1998 | A |
5735055 | Hochbein et al. | Apr 1998 | A |
5740429 | Wang et al. | Apr 1998 | A |
5751582 | Saxena et al. | May 1998 | A |
5754297 | Nulman | May 1998 | A |
5761064 | La et al. | Jun 1998 | A |
5761065 | Kittler et al. | Jun 1998 | A |
5764543 | Kennedy | Jun 1998 | A |
5777901 | Berezin et al. | Jul 1998 | A |
5787021 | Samaha | Jul 1998 | A |
5787269 | Hyodo | Jul 1998 | A |
5808303 | Schlagheck et al. | Sep 1998 | A |
5812407 | Sato et al. | Sep 1998 | A |
5823854 | Chen | Oct 1998 | A |
5824599 | Schacham-Diamand et al. | Oct 1998 | A |
5825356 | Habib et al. | Oct 1998 | A |
5825913 | Rostami et al. | Oct 1998 | A |
5828778 | Hagi et al. | Oct 1998 | A |
5831851 | Eastburn et al. | Nov 1998 | A |
5832224 | Fehskens et al. | Nov 1998 | A |
5838595 | Sullivan et al. | Nov 1998 | A |
5838951 | Song | Nov 1998 | A |
5841676 | Ali et al. | Nov 1998 | A |
5844554 | Geller et al. | Dec 1998 | A |
5857258 | Penzes et al. | Jan 1999 | A |
5859777 | Yokoyama et al. | Jan 1999 | A |
5859964 | Wang et al. | Jan 1999 | A |
5859975 | Brewer et al. | Jan 1999 | A |
5862054 | Li | Jan 1999 | A |
5863807 | Jang et al. | Jan 1999 | A |
5867389 | Hamada et al. | Feb 1999 | A |
5870306 | Harada | Feb 1999 | A |
5871805 | Lemelson | Feb 1999 | A |
5883437 | Maruyama et al. | Mar 1999 | A |
5889991 | Consolatti et al. | Mar 1999 | A |
5901313 | Wolfe et al. | May 1999 | A |
5903455 | Sharpe, Jr. et al. | May 1999 | A |
5910011 | Cruse | Jun 1999 | A |
5910846 | Sandhu | Jun 1999 | A |
5912678 | Saxena et al. | Jun 1999 | A |
5916016 | Bothra | Jun 1999 | A |
5923553 | Yi | Jul 1999 | A |
5926690 | Toprac et al. | Jul 1999 | A |
5930138 | Lin et al. | Jul 1999 | A |
5940300 | Ozaki | Aug 1999 | A |
5943237 | Van Boxem | Aug 1999 | A |
5943550 | Fulford, Jr. et al. | Aug 1999 | A |
5960185 | Nguyen | Sep 1999 | A |
5960214 | Sharpe, Jr. et al. | Sep 1999 | A |
5961369 | Bartels et al. | Oct 1999 | A |
5963881 | Kahn et al. | Oct 1999 | A |
5975994 | Sandhu et al. | Nov 1999 | A |
5978751 | Pence et al. | Nov 1999 | A |
5982920 | Tobin, Jr. et al. | Nov 1999 | A |
6002989 | Shiba et al. | Dec 1999 | A |
6012048 | Gustin et al. | Jan 2000 | A |
6017771 | Yang et al. | Jan 2000 | A |
6036349 | Gombar | Mar 2000 | A |
6037664 | Zhao et al. | Mar 2000 | A |
6041263 | Boston et al. | Mar 2000 | A |
6041270 | Steffan et al. | Mar 2000 | A |
6054379 | Yau et al. | Apr 2000 | A |
6059636 | Inaba et al. | May 2000 | A |
6064759 | Buckley et al. | May 2000 | A |
6072313 | Li et al. | Jun 2000 | A |
6074443 | Venkatesh et al. | Jun 2000 | A |
6077412 | Ting et al. | Jun 2000 | A |
6078845 | Friedman | Jun 2000 | A |
6094688 | Mellen-Garnett et al. | Jul 2000 | A |
6096649 | Jang | Aug 2000 | A |
6097887 | Hardikar et al. | Aug 2000 | A |
6100195 | Chan et al. | Aug 2000 | A |
6108092 | Sandhu | Aug 2000 | A |
6111634 | Pecen et al. | Aug 2000 | A |
6112130 | Fukuda et al. | Aug 2000 | A |
6113462 | Yang | Sep 2000 | A |
6114238 | Liao | Sep 2000 | A |
6127263 | Parikh | Oct 2000 | A |
6128016 | Coelho et al. | Oct 2000 | A |
6136163 | Cheung et al. | Oct 2000 | A |
6141660 | Bach et al. | Oct 2000 | A |
6143646 | Wetzel | Nov 2000 | A |
6148099 | Lee et al. | Nov 2000 | A |
6148239 | Funk et al. | Nov 2000 | A |
6148246 | Kawazome | Nov 2000 | A |
6150270 | Matsuda et al. | Nov 2000 | A |
6157864 | Schwenke et al. | Dec 2000 | A |
6159075 | Zhang | Dec 2000 | A |
6159644 | Satoh et al. | Dec 2000 | A |
6161054 | Rosenthal et al. | Dec 2000 | A |
6169931 | Runnels | Jan 2001 | B1 |
6172756 | Chalmers et al. | Jan 2001 | B1 |
6173240 | Sepulveda et al. | Jan 2001 | B1 |
6175777 | Kim | Jan 2001 | B1 |
6178390 | Jun | Jan 2001 | B1 |
6181013 | Liu et al. | Jan 2001 | B1 |
6183345 | Kamono et al. | Feb 2001 | B1 |
6185324 | Ishihara et al. | Feb 2001 | B1 |
6191864 | Sandhu | Feb 2001 | B1 |
6192291 | Kwon | Feb 2001 | B1 |
6197604 | Miller et al. | Mar 2001 | B1 |
6204165 | Ghoshal | Mar 2001 | B1 |
6210983 | Atchison et al. | Apr 2001 | B1 |
6211094 | Jun et al. | Apr 2001 | B1 |
6212961 | Dvir | Apr 2001 | B1 |
6214734 | Bothra et al. | Apr 2001 | B1 |
6217412 | Campbell et al. | Apr 2001 | B1 |
6219711 | Chari | Apr 2001 | B1 |
6222936 | Phan et al. | Apr 2001 | B1 |
6226563 | Lim | May 2001 | B1 |
6226792 | Goiffon et al. | May 2001 | B1 |
6228280 | Li et al. | May 2001 | B1 |
6230069 | Campbell et al. | May 2001 | B1 |
6236903 | Kim et al. | May 2001 | B1 |
6237050 | Kim et al. | May 2001 | B1 |
6240330 | Kurtzberg et al. | May 2001 | B1 |
6240331 | Yun | May 2001 | B1 |
6245581 | Bonser et al. | Jun 2001 | B1 |
6246972 | Klimasauskas | Jun 2001 | B1 |
6248602 | Bode et al. | Jun 2001 | B1 |
6249712 | Boiquaye | Jun 2001 | B1 |
6252412 | Talbot et al. | Jun 2001 | B1 |
6253366 | Mutschler, III | Jun 2001 | B1 |
6259160 | Lopatin et al. | Jul 2001 | B1 |
6259959 | Martin | Jul 2001 | B1 |
6263255 | Tan et al. | Jul 2001 | B1 |
6268270 | Scheid et al. | Jul 2001 | B1 |
6271670 | Caffey | Aug 2001 | B1 |
6276989 | Campbell et al. | Aug 2001 | B1 |
6277014 | Chen et al. | Aug 2001 | B1 |
6278899 | Piche et al. | Aug 2001 | B1 |
6280289 | Wiswesser et al. | Aug 2001 | B1 |
6281127 | Shue | Aug 2001 | B1 |
6284622 | Campbell et al. | Sep 2001 | B1 |
6287879 | Gonzales et al. | Sep 2001 | B1 |
6290572 | Hofmann | Sep 2001 | B1 |
6291367 | Kelkar | Sep 2001 | B1 |
6292708 | Allen et al. | Sep 2001 | B1 |
6298274 | Inoue | Oct 2001 | B1 |
6298470 | Breiner et al. | Oct 2001 | B1 |
6303395 | Nulman | Oct 2001 | B1 |
6304999 | Toprac et al. | Oct 2001 | B1 |
6307628 | Lu et al. | Oct 2001 | B1 |
6314379 | Hu et al. | Nov 2001 | B1 |
6317643 | Dmochowski | Nov 2001 | B1 |
6320655 | Matsushita et al. | Nov 2001 | B1 |
6324481 | Atchison et al. | Nov 2001 | B1 |
6334807 | Lebel et al. | Jan 2002 | B1 |
6336841 | Chang | Jan 2002 | B1 |
6339727 | Ladd | Jan 2002 | B1 |
6340602 | Johnson et al. | Jan 2002 | B1 |
6345288 | Reed et al. | Feb 2002 | B1 |
6345315 | Mishra | Feb 2002 | B1 |
6346426 | Toprac et al. | Feb 2002 | B1 |
6355559 | Havemann et al. | Mar 2002 | B1 |
6360133 | Campbell et al. | Mar 2002 | B1 |
6360184 | Jacquez | Mar 2002 | B1 |
6363294 | Coronel et al. | Mar 2002 | B1 |
6366934 | Cheng et al. | Apr 2002 | B1 |
6368879 | Toprac | Apr 2002 | B1 |
6368883 | Bode et al. | Apr 2002 | B1 |
6368884 | Goodwin et al. | Apr 2002 | B1 |
6379980 | Toprac | Apr 2002 | B1 |
6381564 | David et al. | Apr 2002 | B1 |
6388253 | Su | May 2002 | B1 |
6389331 | Jensen et al. | May 2002 | B1 |
6389491 | Jacobson et al. | May 2002 | B1 |
6391780 | Shih et al. | May 2002 | B1 |
6395152 | Wang | May 2002 | B1 |
6397114 | Eryurek et al. | May 2002 | B1 |
6400162 | Mallory et al. | Jun 2002 | B1 |
6405096 | Toprac et al. | Jun 2002 | B1 |
6405144 | Toprac et al. | Jun 2002 | B1 |
6417014 | Lam et al. | Jul 2002 | B1 |
6427093 | Toprac | Jul 2002 | B1 |
6432728 | Tai et al. | Aug 2002 | B1 |
6435952 | Boyd et al. | Aug 2002 | B1 |
6438438 | Takagi et al. | Aug 2002 | B1 |
6440295 | Wang | Aug 2002 | B1 |
6442496 | Pasadyn et al. | Aug 2002 | B1 |
6449524 | Miller et al. | Sep 2002 | B1 |
6455415 | Lopatin et al. | Sep 2002 | B1 |
6455937 | Cunningham | Sep 2002 | B1 |
6465263 | Coss, Jr. et al. | Oct 2002 | B1 |
6470230 | Toprac et al. | Oct 2002 | B1 |
6479902 | Lopatin et al. | Nov 2002 | B1 |
6479990 | Mednikov et al. | Nov 2002 | B2 |
6482660 | Conchieri et al. | Nov 2002 | B2 |
6484064 | Campbell | Nov 2002 | B1 |
6486492 | Su | Nov 2002 | B1 |
6492281 | Song et al. | Dec 2002 | B1 |
6495452 | Shih | Dec 2002 | B1 |
6503839 | Gonzales et al. | Jan 2003 | B2 |
6515368 | Lopatin et al. | Feb 2003 | B1 |
6517413 | Hu et al. | Feb 2003 | B1 |
6517414 | Tobin et al. | Feb 2003 | B1 |
6528409 | Lopatin et al. | Mar 2003 | B1 |
6529789 | Campbell et al. | Mar 2003 | B1 |
6532555 | Miller et al. | Mar 2003 | B1 |
6535783 | Miller et al. | Mar 2003 | B1 |
6537912 | Agarwal | Mar 2003 | B1 |
6540591 | Pasadyn et al. | Apr 2003 | B1 |
6541401 | Herner et al. | Apr 2003 | B1 |
6546508 | Sonderman et al. | Apr 2003 | B1 |
6556881 | Miller | Apr 2003 | B1 |
6560504 | Goodwin et al. | May 2003 | B1 |
6563308 | Nagano et al. | May 2003 | B2 |
6567717 | Krivokapic et al. | May 2003 | B2 |
6580958 | Takano | Jun 2003 | B1 |
6587744 | Stoddard et al. | Jul 2003 | B1 |
6590179 | Tanaka et al. | Jul 2003 | B2 |
6604012 | Cho et al. | Aug 2003 | B1 |
6605549 | Leu et al. | Aug 2003 | B2 |
6607976 | Chen et al. | Aug 2003 | B2 |
6609946 | Tran | Aug 2003 | B1 |
6616513 | Osterheld | Sep 2003 | B1 |
6618692 | Takahashi et al. | Sep 2003 | B2 |
6624075 | Lopatin et al. | Sep 2003 | B1 |
6625497 | Fairbairn et al. | Sep 2003 | B2 |
6630741 | Lopatin et al. | Oct 2003 | B1 |
6640151 | Somekh et al. | Oct 2003 | B1 |
6652355 | Wiswesser et al. | Nov 2003 | B2 |
6660633 | Lopatin et al. | Dec 2003 | B1 |
6678570 | Pasadyn et al. | Jan 2004 | B1 |
6708074 | Chi et al. | Mar 2004 | B1 |
6708075 | Sonderman et al. | Mar 2004 | B2 |
6725402 | Coss, Jr. et al. | Apr 2004 | B1 |
6728587 | Goldman et al. | Apr 2004 | B2 |
6735492 | Conrad et al. | May 2004 | B2 |
6751518 | Sonderman et al. | Jun 2004 | B1 |
6774998 | Wright et al. | Aug 2004 | B1 |
20010001755 | Sandhu et al. | May 2001 | A1 |
20010003084 | Finarov | Jun 2001 | A1 |
20010006873 | Moore | Jul 2001 | A1 |
20010030366 | Nakano et al. | Oct 2001 | A1 |
20010039462 | Mendez et al. | Nov 2001 | A1 |
20010040997 | Tsap et al. | Nov 2001 | A1 |
20010042690 | Talieh | Nov 2001 | A1 |
20010044667 | Nakano et al. | Nov 2001 | A1 |
20020032499 | Wilson et al. | Mar 2002 | A1 |
20020058460 | Lee et al. | May 2002 | A1 |
20020070126 | Sato et al. | Jun 2002 | A1 |
20020077031 | Johannson et al. | Jun 2002 | A1 |
20020081951 | Boyd et al. | Jun 2002 | A1 |
20020089676 | Pecen et al. | Jul 2002 | A1 |
20020102853 | Li et al. | Aug 2002 | A1 |
20020107599 | Patel et al. | Aug 2002 | A1 |
20020107604 | Riley et al. | Aug 2002 | A1 |
20020113039 | Mok et al. | Aug 2002 | A1 |
20020127950 | Hirose et al. | Sep 2002 | A1 |
20020128805 | Goldman et al. | Sep 2002 | A1 |
20020149359 | Crouzen et al. | Oct 2002 | A1 |
20020165636 | Hasan | Nov 2002 | A1 |
20020183986 | Stewart et al. | Dec 2002 | A1 |
20020185658 | Inoue et al. | Dec 2002 | A1 |
20020193899 | Shanmugasundram et al. | Dec 2002 | A1 |
20020193902 | Shanmugasundram et al. | Dec 2002 | A1 |
20020197745 | Shanmugasundram et al. | Dec 2002 | A1 |
20020197934 | Paik | Dec 2002 | A1 |
20020199082 | Shanmugasundram et al. | Dec 2002 | A1 |
20030017256 | Shimane | Jan 2003 | A1 |
20030020909 | Adams et al. | Jan 2003 | A1 |
20030020928 | Ritzdorf et al. | Jan 2003 | A1 |
20030154062 | Daft et al. | Aug 2003 | A1 |
Number | Date | Country |
---|---|---|
2050247 | Aug 1991 | CA |
2165847 | Aug 1991 | CA |
2194855 | Aug 1991 | CA |
0 397 924 | Nov 1990 | EP |
0 621 522 | Oct 1994 | EP |
0 747 795 | Dec 1996 | EP |
0 869 652 | Oct 1998 | EP |
0 877 308 | Nov 1998 | EP |
0 881 040 | Dec 1998 | EP |
0 895 145 | Feb 1999 | EP |
0 910 123 | Apr 1999 | EP |
0 932 194 | Jul 1999 | EP |
0 932 195 | Jul 1999 | EP |
1 066 925 | Jan 2001 | EP |
1 067 757 | Jan 2001 | EP |
1 071 128 | Jan 2001 | EP |
1 083 470 | Mar 2001 | EP |
1 092 505 | Apr 2001 | EP |
1 072 967 | Nov 2001 | EP |
1 182 526 | Feb 2002 | EP |
2 347 885 | Sep 2000 | GB |
2 365 215 | Feb 2002 | GB |
61-66104 | Apr 1986 | JP |
61-171147 | Aug 1986 | JP |
01-283934 | Nov 1989 | JP |
3-202710 | Sep 1991 | JP |
05-151231 | Jun 1993 | JP |
05-216896 | Aug 1993 | JP |
05-266029 | Oct 1993 | JP |
06-110894 | Apr 1994 | JP |
06-176994 | Jun 1994 | JP |
06-184434 | Jul 1994 | JP |
06-252236 | Sep 1994 | JP |
06-260380 | Sep 1994 | JP |
8-23166 | Jan 1996 | JP |
08-50161 | Feb 1996 | JP |
08-149583 | Jun 1996 | JP |
08-304023 | Nov 1996 | JP |
09-34535 | Feb 1997 | JP |
9-246547 | Sep 1997 | JP |
10-34522 | Feb 1998 | JP |
10-173029 | Jun 1998 | JP |
11-67853 | Mar 1999 | JP |
11-126816 | May 1999 | JP |
11-135601 | May 1999 | JP |
2000-183001 | Jun 2000 | JP |
2001-76982 | Mar 2001 | JP |
2001-284299 | Oct 2001 | JP |
2001-305108 | Oct 2001 | JP |
2002-9030 | Jan 2002 | JP |
2002-343754 | Nov 2002 | JP |
434103 | May 2001 | TW |
436383 | May 2001 | TW |
455938 | Sep 2001 | TW |
455976 | Sep 2001 | TW |
WO 9534866 | Dec 1995 | WO |
WO 9805066 | Feb 1998 | WO |
WO 9845090 | Oct 1998 | WO |
WO 9909371 | Feb 1999 | WO |
WO 9925520 | May 1999 | WO |
WO 9959200 | Nov 1999 | WO |
WO 0000874 | Jan 2000 | WO |
WO 0005759 | Feb 2000 | WO |
WO 0035063 | Jun 2000 | WO |
WO 0054325 | Sep 2000 | WO |
WO 0079355 | Dec 2000 | WO |
WO 0111679 | Feb 2001 | WO |
WO 0115865 | Mar 2001 | WO |
WO 0118623 | Mar 2001 | WO |
WO 0125865 | Apr 2001 | WO |
WO 0133277 | May 2001 | WO |
WO 0133501 | May 2001 | WO |
WO 0152055 | Jul 2001 | WO |
WO 0152319 | Jul 2001 | WO |
WO 0157823 | Aug 2001 | WO |
WO 0180306 | Oct 2001 | WO |
WO 0217150 | Feb 2002 | WO |
WO 0231613 | Apr 2002 | WO |
WO 0231613 | Apr 2002 | WO |
WO 0233737 | Apr 2002 | WO |
WO 02074491 | Sep 2002 | WO |
Number | Date | Country | |
---|---|---|---|
20050171626 A1 | Aug 2005 | US |