TWI509375B - Method and apparatus for implementing an automated in-line process control scheme during recipe execution - Google Patents

Description

Method and apparatus for implementing an automated in-line process control scheme during recipe execution

The present invention relates to apparatus and methods for implementing an automated in-line process control scheme during recipe execution.

In a competitive market, semiconductor device manufacturers must minimize waste and consistently produce high quality semiconductor devices to maintain a competitive advantage. Thus, strict control of the processing environment during substrate processing helps to achieve optimal results. Therefore, manufacturing companies use time and resources to identify methods and/or devices to improve substrate processing.

In order to provide tight control of the processing environment, it may be necessary to characterize the processing environment. In order to provide the information needed to characterize the processing environment of the processing chamber, the sensor can be used during recipe execution to retrieve processing data. These materials can be analyzed and the processing environment can be adjusted accordingly (eg, "Adjust recipes").

In general, the analysis is performed after a single substrate or substrate lot has been processed. Measurements are typically performed offline by one or more metrology tools. This method typically requires time and technology to measure and/or analyze the measurement data. If a problem is identified, additional time may be required to compare the measured data with the processed data to determine the cause of the problem. Often, the analysis can be complex and requires the interpretation of a dedicated person. Again, the analysis is typically performed after at least one (or perhaps several) of the substrates have been processed. Since the analysis is not performed in an online and immediate manner, damage and/or adverse effects may already be present on the substrate and/or processing chamber/chamber parts.

In some plasma processing tools, the sensors can be integrated into portions of the process control loop. Therefore, the sensor can collect not only processing data but also monitoring tools. In one example, a pressure gauge can be used to collect pressure data. However, for example, during recipe execution, the data collected by the pressure gauge can be used by the process module controller to adjust the pressure set point.

To facilitate the description, Figure 1 shows a simplified block diagram of the processing chamber. This figure does not represent an accurate image of the processing chamber. Instead, this diagram shows how to set up a set of sensors Implemented in the processing chamber to facilitate the execution of process recipes.

Consider, for example, the case where the substrate batch will be processed in the processing chamber 100. Prior to processing, the metrology tool 102 (which may be one or more metrology tools) may be used to perform pre-processing measurements. Pre-processed measurement data from metrology tool 102 can be uploaded to manufacturing plant master controller 106 via link 104.

To begin processing the substrate batch, the user can use the manufacturing plant master controller 106 to select a recipe for execution. In some cases, the measurement data can be used by the manufacturing plant master controller 106 to adjust recipe set points to compensate for input material differences. In one example, the pre-processing measurements of the substrate may indicate that the physical characteristics of the substrate are different than those contemplated by the formulation. Therefore, the recipe set point can be adjusted by taking into account known substrate differences.

Once the recipe is selected and the recipe is adjusted based on the pre-measurement data, the manufacturing plant master controller 106 can send the recipe to the process module (PM) controller 108 via the link 110. The substrate 112 can be loaded into the processing chamber 100. The substrate 112 can be located between the lower electrode 114 (eg, an electrostatic chuck) and the upper electrode 116. During processing, the plasma 118 can be formed to process (e.g., etch) the substrate 112.

During processing, a plurality of sensors can be used to monitor the state of process chamber 100, plasma 118, and/or substrate 112. Examples of sensors may include, but are not limited to, airflow controller (120), temperature sensors (122 and 124), pressure sensors (126), matching box controller set (128), radio frequency (RF, Radio frequency) generator controller (130), valve controller (132), turbo pump controller (134), and the like. In one example, pressure sensor 126 can capture pressure data within processing chamber 100. In another example, RF generator controller 130 and/or matching box controller group 128 can collect information about reflected power, impedance, harmonics, and the like.

The data collected by each sensor can be sent along a communication line (e.g., 140, 142, 144, 146, 148, 150, and 152) to a control data hub 136 for analysis. If any recipe set point must be adjusted based on this analysis, control data hub 136 can send the result to processing module controller 108 (via link 138), and processing module controller 108 can adjust the recipe set point accordingly . In one example, the desired pressure set point in accordance with the recipe can be set to 30 millitorres. However, according to the pressure sensor 126, the pressure measurement is actually 26 mTorr. Thus, the process module controller 108 can adjust the pressure control drive to restore pressure to the desired recipe set point.

The uni-variate orthogonal control scheme is characterized by a process control relationship between the recipe set point and the sensor. In other words, the recipe set point can be related to the data collected from a single sensor that is considered to be responsive only to a single parameter. When it is determined whether a particular recipe set point is followed, the data collected from any other sensor is generally not considered.

In the above example, the chamber pressure is adjusted based on the information provided by the pressure sensor 126. When making adjustments, the process module controller 108 can assume that the pressure sensor 126 provides accurate information and that the pressure sensor 126 is not subject to deviations and/or component wear. However, if the pressure sensor 126 has actually produced a deviation, an increase in pressure that attempts to return the chamber condition to the desired state in accordance with the process module controller 108 may result in undesirable results on the substrate 112 and create a chamber Abnormalities associated with the wall and the components within it (including the sensor itself).

One embodiment of the present invention is directed to an apparatus for implementing an automated in-line process control scheme during recipe execution. The apparatus includes a control loop sensor for collecting at least a first set of sensor data to facilitate monitoring of the set point during recipe execution, wherein the control loop sensor is part of a process control loop. The device also includes an independent sensor that is used to collect at least a second set of sensor data, and the independent sensor is not part of the process control loop. The device in turn includes a hub for receiving at least one of the first set of sensor data and the second set of sensor data. The device, in turn, further includes an analysis computer communicatively coupled to the hub and configured to perform an analysis of at least one of the first set of sensor data and the second set of sensor data.

The above summary is only one of the many embodiments of the invention disclosed herein, and is not intended to limit the scope of the invention. These and other features of the present invention will be described in detail in the following detailed description of the invention.

In the following, the invention will be described in detail with reference to a number of embodiments of the invention as illustrated in the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the invention. It will be appreciated by those skilled in the art, however, that the invention may be practiced without some or all of these specific details. In other instances, well-known process steps and/or structures have not been described in detail in order not to unnecessarily obscure the invention.

Various embodiments incorporating methods and techniques are described below. It should be borne in mind that the present invention also encompasses articles of manufacture comprising computer readable media on which computer readable instructions for performing embodiments of the present technology can be stored. The computer readable medium can include, for example, a semiconductor, magnetic, optical-magnetic, optical, or other form of computer readable medium for storing computer readable codes. Further, the invention may also encompass apparatus for practicing embodiments of the invention. Such devices may include proprietary and/or programmable circuitry for performing the operations associated with embodiments of the present invention. Examples of such devices include a suitably programmed general purpose computer and/or proprietary computing device, and may include a combination of computer/computing devices and proprietary/programmable circuits for various tasks associated with embodiments of the present invention.

As described above, in order to perform substrate processing with consistent results, strict control of the processing environment is desired. However, in view of the fact that the sensor may be inaccurate, sensitive to multiple parameters, biased over time, and/or become defective, formulation adjustments based on single variable sensor data have proven to be sometimes unreliable. .

Those skilled in the art will appreciate that certain parameters may be more important than other parameters in the characterization of the substrate. In one example, the ability to control electron density (as a processing parameter) can provide tighter control over the substrate processing results than the ability to control pressure levels, which is less straightforward. However, not all parameters can be easily measured directly by a single sensor. In addition, not all parameters can be controlled by a single direct physical drive/controller. For example, the pressure level can be measured by a pressure gauge. Thus, if the pressure measurement indicates that the pressure is deviating from the desired pressure, a pressure controller can be used to adjust the pressure within the chamber for compensation. However, electron density is a parameter that cannot be directly measured by a single sensor. Instead, in order to determine the electron density, it may be necessary to perform complex The calculation, because the electron density may have to be derived from a plurality of processing data points from one or more sensors. Also, a simple direct physical drive cannot be used to control electron density during substrate processing.

In one embodiment of the invention, the inventors herein understand that the recipe can be executed by utilizing an independent data stream (which can be obtained from one or more sensors independent of the direct processing control loop). Provide confirmation before and after adjustment. Furthermore, the inventors herein have learned that by performing multivariate non-orthogonal analysis, parameters that cannot be directly measured can be derived using algorithm/model based calculations and used to perform recipe adjustments.

In accordance with an embodiment of the present invention, a method and apparatus for performing on-line process control is provided. Embodiments of the invention include an apparatus for providing an independent data stream. The independent data stream can include data collected from control loop sensors and/or independent sensors. Embodiments of the present invention also include an automatic multivariable non-orthogonal control scheme that can provide virtual sensors and/or virtual drivers to perform fault detection, fault classification, and/or recipe adjustment.

As used herein, a control loop sensor refers to a sensor that is also part of a process control loop. In other words, data from the control loop sensor is used during recipe execution to monitor the recipe set point. In the prior art, the data collected from the control loop sensor is typically used to adjust the recipe set point.

As used herein, an independent sensor refers to a sensor that is not currently part of a conventional process control loop. In one embodiment of the invention, the independent sensors can be matched and calibrated between the chambers and the chambers. In another embodiment, the independent sensor can be a redundant sensor. For example, an independent sensor can have the same style or type as the pressure gauge used for the process control loop. However, independent pressure gauges are independent of the process control loop. In an embodiment, redundant independent sensors may be located near the control loop sensor to allow for independent but identical measurements.

As described herein, a virtual sensor refers to a software-implemented sensor that is not a hardware component. In an embodiment, the virtual sensor can be a composite sensor or an analog of multiple sensors, and can provide virtual sensor measurements for parameters that are not generally directly measured. In an embodiment, the virtual parameter can be from a plurality of data Sources are calculated and/or inferred. Therefore, parameters that cannot be measured by a single sensor entity can be derived using a virtual sensor. Examples of virtual parameters may include, but are not limited to, ion flux, ion energy, electron density, etch rate versus deposition rate ratio, and the like.

As used herein, a virtual drive is a software-implemented controller that can be used to implement control of parameters that cannot be directly measured or controlled by a single physical drive. A physical drive (eg, an ion flux controller) may not exist for parameters (eg, ion flux), as this parameter may, for example, not be measured directly by a physical sensor, but must be calculated, eg, indirectly from different The source of the data is derived.

In one embodiment of the invention, a method and apparatus for an on-line process control system is provided. Traditionally, when needed, a control loop sensor can be used to retrieve processing data and provide feedback to the processing module controller to adjust recipe set points. In general, a single variable orthogonal control scheme can be used. In other words, there is a one-to-one relationship between a recipe set point and a sensor. Data from other sensors is usually not used to adjust the set point. However, data from the control loop sensor may not be sufficient to verify the chamber/plasma/substrate parameters of interest. Therefore, adjusting the recipe set point based solely on data from the control loop sensor may have side effects (eg, poor processing results, or even damage to the substrate, damage to the chamber walls, to chamber components) Damage, etc.).

In an embodiment, an independent data stream is provided to determine certain conditions associated with the chamber/plasma/substrate state. In an embodiment, the independent data stream may also include data collected only from independent sensors. As mentioned above, the independent sensor is not a sensor that is part of a conventional process control loop. In one embodiment, the independent sensor is matched and calibrated in a universal standard. In other words, an independent sensor can be used to capture the specific properties of the chamber.

In an embodiment, the independent data stream may include data collected from the control loop sensor and/or the independent sensor. In one example, pressure level related data may be collected by various control loop sensors, although only pressure data from the pressure gauge may be used, for example, to set a pressure set point. Thus, in this embodiment, the data collected by the control loop sensor can be, but is not required to be, used as part of an independent data stream to verify the data provided by a single control loop sensor.

In an embodiment, the independent data stream can be analyzed to establish a virtual sensor for determining certain conditions associated with the chamber/plasma/substrate state. As noted above, certain chamber/plasma/substrate conditions cannot be directly measured. Instead, complex calculations may have to be performed to derive parameters that can characterize these chamber/plasma/substrate states. In one embodiment, the inventors herein have learned that there is a hierarchical relationship between sensors that facilitate virtual measurements. In an example, by applying an independent data stream to a phenomenological model, we can derive virtual sensors such as ion flux distribution, electron density, etch rate, neutral density, and the like. .

In an embodiment, the independent data stream can be analyzed separately or combined with data streams from the control loop sensor to generate virtual sensor data for adjusting recipe parameters that cannot be sensed by the sensor. Direct measurement. Once the virtual sensor is generated, the process control can set points based on the virtual sensor that can be defined. During recipe execution, the sensor data provided by the virtual sensor can be compared to the virtual sensor set point and the difference can be calculated. The virtual drive can then be used to control one or more physical drives to adjust these virtual set points.

The features and advantages of the present invention are better understood by reference to the following drawings and description.

2 shows a simplified block diagram of a processing chamber having an on-line control processing device in one embodiment of the present invention. The invention is not limited by the apparatus and/or components shown. Rather, the drawings are used to facilitate the description of an embodiment of the invention as an example.

Consider, for example, the case where the substrate batch will be processed in the process chamber 200. The pre-processing measurement data (external data) can be obtained by a set of measurement tools 202 before the substrate can be processed. The measurement data from the metrology tool 202 can be uploaded to the manufacturing plant master controller 206 via the link 204. For the implementation of the invention, pre-processing measurements are not required. However, if desired, in one embodiment, the processing chamber 200 can provide a communication link (204) between the metrology tool 202 and the manufacturing plant host controller 206 to integrate the metrology data with the substrate processing. This provides the basis for compensating for variations in the input substrate and reducing undesirable variations in the output product.

For example, to initiate processing, the recipe can be selected by the manufacturing plant master controller 206. If the pre-processing measurement data is available, the formulation can be adjusted by taking into account the physical changes in the input between the substrates. Once completed, the manufacturing plant master controller 206 can send the recipe to the processing module controller 208 via the link 210. Link 210 is a two-way link that facilitates the exchange of data between manufacturing plant master controller 206 and processing module controller 208.

The substrate 212 can be loaded into the processing chamber 200. The substrate 212 can be located between the lower electrode 214 (eg, an electrostatic chuck) and the upper electrode 216. Plasma 218 may be formed to process (eg, etch) substrate 212 during processing.

A plurality of sensors can be used during recipe execution to monitor various parameters associated with process chamber 200, plasma 218, and/or substrate 212. Examples of sensors may include, but are not limited to, airflow controllers (220), temperature sensors (222 and 224), pressure sensors (226), matching box controller sets (228), radio frequency controllers (230). ), valve controller (232), turbo pump controller (234), and the like. In one example, temperature sensor 222 can collect temperature data within processing chamber 200. In another example, the turbo pump controller 234 can collect information about pump speed and flow rate.

For ease of illustration, the above sensors are grouped together and are referred to below as control loop sensors. As used herein, a control loop sensor refers to a sensor that is part of a process control loop and that has traditionally been used during recipe execution to monitor recipe set points.

In addition to the control loop sensors that are part of the process control loop, separate sensors (eg, 260, 262, and 264) may also be provided. In an embodiment, the independent sensor is not traditionally part of the process control loop. The number of independent sensors can vary. In one embodiment of the invention, the independent sensor can be matched and calibrated against the absolute standard and between itself to give a consistent result between the chamber and the chamber.

In one embodiment of the invention, an independent sensor is selected and supplied to provide at least a portion of the data overlap for some or all of the data items. In other words, data about a particular virtual sensor parameter can be retrieved by more than one sensor. In an example, the independent sensor 262 can be used to collect data (including pressure related data). Received The collected data may, for example, overlap with the pressure data collected by the pressure sensor 226.

In an embodiment, the independent sensor can be a redundant sensor. For example, an independent sensor can have the same style as a pressure gauge, which can be used in a process control loop. However, independent sensor pressure gauges are independent of traditional process control loops.

In an embodiment, the independent sensor may be comprised of a sensor that does not have a direct overlap with the control loop sensor. In one example, a voltage/current detector can be used as one of the independent sensors used in conjunction with a pressure sensor to derive virtual sensor measurements.

The data collected by the control loop sensor can be sent along the communication lines (eg, 240, 242, 244, 246, 248, 250, and 252) to the control data hub 236 for analysis (similar to prior art) ). In addition, data from independent sensors (260, 262, and 264) can also be sent to measurement sensor data hub 280 along communication lines (270, 272, and 274). In an embodiment, certain data collected by the control loop sensor may be transmitted from the control data hub 236 to the measurement sensor data hub 280 via the communication link 254. In another embodiment, all of the data collected by the control loop sensor can be sent to the measurement sensor data hub 280 via the control data hub 236.

After collecting the data and optionally performing some pre-processing work (eg, digital format conversion), the data can be sent to the analysis processor via communication line 284, which can be executed within a separate dedicated computer 282. In an embodiment, the data collected by the control loop sensor may also be transmitted from the control data hub 236 to the analysis computer 282 via the communication line 256.

From the above, we can understand that a large amount of data can be collected by the control loop sensor and the independent sensor. In an embodiment, the data collected by the independent sensors may be extremely granular. In one embodiment, the analysis computer 282 can be a fast processing module that can be used to process large amounts of data. Data can be sent directly from the sensor without first having to go through the manufacturing plant master or even the module controller. U.S. Patent Application Serial No. 12/555,674, filed on Jan. 8, 2009, the entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire entire all

In one embodiment, the analysis computer 282 can receive measurement data from the metrology tool 202 via the communication link 290 in addition to the data collected from the sensor. In an embodiment, the metrology data that may be provided to the manufacturing plant master controller 206 may also be sent to the analysis computer 282. Thus, the analysis computer 282 can be used to manipulate recipe adjustments that have previously been performed by the manufacturing plant master controller 206.

In one embodiment, the analysis computer 282 is configured to analyze the independent data stream, and the results can be sent to the processing module controller 208 via the communication link 286. Figure 3 illustrates an example of a hierarchical relationship that analysis computer 282 can use in performing its analysis. In one embodiment, a high speed communication link is used to provide immediate updates to the processing module controller 208. The results produced by analysis computer 282 may include virtual sensor set point adjustments, fault detection and classification, and multiple sensor endpoints. Based on these results, the processing module controller 208 can adjust the recipe and/or abort the processing.

Unlike prior art, a multivariate non-orthogonal control scheme can be used to define the relationship between recipe set points and sensors. A multivariable non-orthogonal control scheme can have two features: (a) there is no one-to-one relationship between recipe setpoints and virtual sensor parameters; and (b) parameters from multiple sensors are used Determine the virtual sensor parameters. In other words, the recipe set point can be related to data collected from a plurality of sensors. Unlike the prior art, the adjustment of the recipe set point is no longer based solely on the data collected by the control loop sensor. Instead, the data collected by the independent sensors (and in one embodiment, by the control loop sensor) can be used alone or in combination with a control loop sensor to determine and control certain cavities Room/plasma/substrate status.

To facilitate the description, Figure 3 shows the hierarchical relationship between the sensors/drivers in one embodiment of the invention. Consider, for example, the case where the substrate 212 is being processed within the processing chamber 200. The recipe set point can be set when the recipe is first activated. Traditionally, recipe set points are dependent on measurements from the control loop sensor. Traditionally, the process module controller 208 can adjust the recipe set point after the substrate or substrate lot has been processed using data from the control loop sensor (block 302). For ease of illustration, block 302 may be referred to as a vector S.

However, as noted above, the data from the control loop sensor may not always be accurate, and especially if a recipe set point is associated with a control loop sensor This data may not be detected when there is a single variable orthogonal relationship. Therefore, if the control loop sensor (eg, pressure sensor 226) fails, the reliance on the data provided by the control loop sensor may result in poor processing results, even damaged substrates, and even May damage the chamber components.

Additional data can be provided through other control loop sensors and independent sensors to provide independent data sources to verify pressure data, for example, prior to adjusting the recipe pressure set point. For a particular recipe set point, this data can be taken before or during recipe execution, but can be independent of the process control loop (block 304). Block 304 may be referred to as a vector V for ease of illustration.

In an embodiment, an empirical relationship (vector Q) may exist between blocks 302 and 304. The empirical relationship (vector Q) between vector S (302) and vector V (304) tends to be chamber specific due to the specific chamber conditions and unique sensor characteristics that can change due to manufacturing tolerances. (chamber specific).

As described above, block 304 can be used to verify the information provided by the control loop sensor in block 302. In an example, the independent sensor 264 can provide information that does not confirm the information provided by the pressure sensor 226. In other words, even though the pressure sensor 226 may give a different indication, the information provided by the independent sensor 264 still indicates that the pressure does not have to be adjusted.

However, analyzing only one parameter (eg, pressure level) or multiple directly measurable parameters may not provide all of the information needed to bring the substrate and/or plasma into a desired state. To make this process more desirable or more efficient, a virtual sensor and/or virtual drive can be provided (block 306). Block 306 may be referred to as a vector R for ease of illustration.

As used herein, a virtual sensor refers to a composite sensor or analog of a multi-sensor that can measure parameters that cannot be directly measured by a single sensor in a virtual manner. Instead, virtual sensor parameters can be calculated and/or inferred from data from a plurality of sensors. Examples of virtual parameters may include, but are not limited to, ion flux, ion energy, electron density, etch rate versus deposition rate ratio, and the like.

In an embodiment, a phenomenological relationship (vector M) may exist between the vector R and the vector V. As described herein, the phenomenon relationship refers to the parameter A relationship that can be related to or derived from another parameter, even if the relationship is non-linear or extremely complex. Therefore, in order to build a virtual sensor, the understanding of the phenomenological behavior of the formula (such as underlying physics) may be necessary, and if the basic model has validity, it is generally expected to be more pure than pure statistical analysis. Have an advantage. Therefore, the vector M tends to be specific to the type of process.

In one example, the geometry of the chamber, the state of the consumable part, the accuracy of the airflow controller, the accuracy of the pressure controller, the substrate, and the like can affect the ion flux distribution. Accurately modeling ion flux distributions with all of these effects can be extremely complex and time consuming. However, the phenomenon relationship can be defined, where RF voltage and current measurements and ion flux measurements can be used at a location, such as with an electrical model of the processing chamber, to derive virtual sensing associated with ion flux. Device.

As can be appreciated from Figure 3, traversing from block 302 to block 306 in a reliable manner may require a separate data stream (provided by block 304). Data from the independent data stream can be used to calculate the measurement of the virtual sensor in block 306. In other words, when the hierarchical relationship traverses from block 302 to block 306 via block 304, instant measurement capabilities can be provided.

In an embodiment, immediate process control capabilities may be provided when performing inverse hierarchical relationships. In other words, as the system traverses from block 302 to block 306 via block 304, a set of virtual drives can be implemented to adjust the recipe. In one example, the electron density (a virtual sensor value) can be identified as being outside of the desired paradigm. The difference between the set point electron density and the virtual electron density value can be calculated. In one embodiment, if the control loop sensor does not produce a deviation, the calculated difference can be used by the virtual drive to adjust the process to the desired set point. However, if the control loop sensor produces a slight deviation (as shown by an independent sensor), the calculated gap before adjusting the recipe may have to be modified to take into account the bias.

In an embodiment, the virtual drive can be driven in a small incremental manner. In one example, the fractional value can be used first to ensure that the virtual drive does not inadvertently exacerbate the problem, instead (in the above example) the entire computational gap is used to adjust the recipe. If the analysis after a small change indicates, for example, that the substrate is changing toward a desired state, then the formulation can be The adjustments used further adjustments. If the parameter space is well-behaved, advanced nonlinear "leap ahead" adjustments such as the steepest descent technique can be used; but if the parameter space is complex and ill conditioned, step by step The limiting method produces better results.

4 shows a simplified flow diagram in an embodiment of the present invention showing an implementation of an on-line control processing method for performing virtual measurements. As described herein, virtual measurement refers to obtaining measurement data without performing actual measurement, and the measurement data includes data that cannot be directly measured.

In an initial step 402, the recipe is downloaded to the processing module controller. In one example, manufacturing plant master controller 206 can send recipes to processing module controller 208 via communication link 210.

At next step 404, sensor calibration data (vector Q) is provided. In an embodiment, the empirical relationship between the control loop sensor and the independent sensor is provided to the analysis computer 282.

In the next step 406, the downloaded recipe is executed and the recipe is adjusted to the recipe set point (as indicated by block 302).

In the next step 408, the data is retrieved by the sensor during processing.

In the next step 410, the system checks to determine if the process is aborted.

If the process is not aborted, the system will return to step 408 to continue obtaining the data.

However, if the process is aborted, the system proceeds to step 412 to determine if the desired result is obtained. In order to make such a decision without performing actual measurement, a hierarchical relationship can be used in which a phenomenon model (vector M) is used for block 304 (vector V) to calculate a virtual measure (vector R).

In the next step 414, the system (e.g., analysis computer 282) can compare the virtual "measurement" to a predefined threshold. In this step, the system can check this processing result again to determine if the processing result is within the control limits.

If the result of the processing is within the control limits, then in the next step 416, another substrate is loaded for processing and the system returns to step 406.

However, if the virtual measurement system falls outside of the predefined threshold, then in the next step 418, the system will trigger a warning or alarm (in general, the police) The difference between the alarm and the alarm is that the warning will alert the system and the operator that adjustments, diagnostic checks and maintenance are required; and the alarm will stop processing until the corrective action is taken to prevent damage to the substrate and/or machine. In an embodiment, the triggering of a warning or alarm may cause fault detection, fault classification, and/or adjustment of the recipe.

As can be seen from Figure 4, the online control process can provide a method of virtual execution processing measurement. Unlike the prior art, the substrate does not have to be removed from the chamber and does not have to be measured using a physical metrology tool. Thus, the virtual metrology capabilities provided by the system of the present invention can reduce the cost of expensive metrology tools. Moreover, the virtual measurement capability can substantially reduce the time and resources required to perform the metrology analysis. In addition, personnel do not need to perform measurements and analysis. Instead, the system (for example, through an analysis computer) can be used to automatically collect and calculate virtual measurements. An additional advantage of the present invention is the ability to intervene during the process. Since deviations from the standard can be detected during recipe execution, a decision can be made as to whether to proceed with the process before the wafer is damaged by unrecoverable damage. In many processes, the step that most affects the critical dimension is typically the mask open step. If a deviation is detected during the mask processing step, the wafer can still be recovered by rework.

Figure 5 shows a simplified flow diagram in an embodiment of the invention showing the implementation of an on-line control process to provide instant process control capabilities.

In an initial step 502, the recipe is downloaded to the processing module controller. In one example, manufacturing plant master controller 206 can send recipes to processing module controller 208 via communication link 210.

At next step 504, sensor calibration data (vector Q) is provided. In an embodiment, the empirical relationship between the control loop sensor and the independent sensor is provided to the analysis computer 282.

In the next step 506, the recipe is executed and the recipe is adjusted to the recipe set point (as indicated by block 302).

At the next step 508, the data is retrieved during processing. Information can be obtained at different times. In one embodiment, the data is obtained, for example, at a frequency of about ten hertz.

After the first set of data settings has been obtained by the analysis computer 282, at the next step 510, a virtual measurement can be obtained. In other words, you can use the hierarchical relationship, which is now An image model (vector M) can be used for block 304 (vector V) to calculate a virtual measure (vector R).

At next step 512, the system can check to determine if the process is in a desired state.

If the process is within the desired state, then in the next step 514, the system can check to determine if the process is to end.

If the recipe is still being executed, then the system can proceed back to step 508 to obtain the next set of data.

However, if the process is aborted, then in the next step 516, the system will abort the process.

Referring back to step 512, if the process is not within the desired state, then in the next step 518, the system may perform a check to determine if a fault has been detected.

If a fault has been detected, then in the next step 520, the system will trigger an alarm and in the next step 522, the fault will be classified.

However, if no fault is detected, then in the next step 524, the adjusted recipe set point can be calculated. To determine which virtual drive can be used to adjust the recipe, a hierarchical model can be used. In one example, data has been collected from control loop sensors as well as independent sensors. In addition, the virtual sensor has been calculated based on the collected data and the phenomenal model that exists between the independent data stream and the control loop sensor. Once the virtual sensor has been determined, the virtual sensor measurement can be compared to the expected value. The difference can be used by the virtual drive to adjust the recipe.

As noted above, raw differences may not be sent to the processing module controller for actual adjustment of the recipe. Instead, any potential noise or bias (vector V) may have to be considered to derive a new recipe set point.

After the new recipe set point has been determined, in a next step 526, the system can send the new recipe set point to the process module controller.

In the next step 528, the recipe is adjusted to the new recipe set point.

Once the recipe is adjusted to the new recipe set point, the system can return to step 508 to obtain a new data set.

We can see from Figure 5 that the recipe can be fine-tuned during recipe execution. (fine-tuning) (instant). Unlike prior art, recipe adjustments can be confirmed by independent data streams. Again, the set points that can be adjusted are no longer limited to parameters that can be directly measured. Instead, parameters that depend on multiple parameters can be calculated and used for the purpose of setting points.

Also, the drive is not limited to the physical drive that is present. We can use a virtual drive that can start multiple other physical drives in sequence when it is started. In this way, process monitoring and control can be substantially de-skilled.

From the above, we can understand the method and equipment that can provide an automatic online process control solution. With an on-line process control solution, instant control is provided when each substrate is processed in the desired recipe state. Online process control also provides an online method for performing fault detection and classification in an immediate manner. Again, the on-line control process can provide a tool with virtual metrology capabilities to determine the state of the processing substrate.

While the invention has been described in terms of several preferred embodiments, modifications, substitutions, and While various examples are provided herein, it is intended that the examples are merely illustrative and not limiting.

Further, the title and the summary are provided herein for convenience, and should not be used herein to explain the scope of the claims. Further, the abstract is written in a very short form and is provided for convenience herein, and thus should not be used to explain or limit the invention as a whole. If the term "set" is used herein, it is meant to have a mathematical meaning that encompasses zero, one, or more of which is generally understood. It should also be noted that there are many alternative ways of implementing the methods and apparatus of the present invention. Therefore, it is intended that the following claims be construed as including all such modifications, permutations, and equivalents.

100‧‧‧Processing room

102‧‧‧Measurement tools

104‧‧‧ links

106‧‧‧Manufacture factory main controller

108‧‧‧Processing Module Controller

110‧‧‧ links

112‧‧‧Substrate

114‧‧‧ lower electrode

116‧‧‧Upper electrode

118‧‧‧ Plasma

120‧‧‧Airflow controller

122‧‧‧temperature sensor

124‧‧‧Temperature Sensor

126‧‧‧pressure sensor

128‧‧‧matching box controller group

130‧‧‧ Radio Frequency Controller

132‧‧‧Valve Controller

134‧‧‧ turbo pump controller

136‧‧‧Control Data Hub

138‧‧‧ links

140‧‧‧Communication lines

142‧‧‧Communication lines

144‧‧‧Communication lines

146‧‧‧Communication lines

148‧‧‧Communication lines

150‧‧‧Communication lines

152‧‧‧Communication lines

200‧‧‧Processing room

202‧‧‧Measurement tools

204‧‧‧ links

206‧‧‧Manufacture factory main controller

208‧‧‧Processing Module Controller

210‧‧‧ links

212‧‧‧Substrate

214‧‧‧ lower electrode

216‧‧‧Upper electrode

218‧‧‧ Plasma

220‧‧‧Airflow controller

222‧‧‧ Temperature Sensor

224‧‧‧temperature sensor

226‧‧‧pressure sensor

228‧‧‧Matching box controller group

230‧‧‧ Radio Frequency Controller

232‧‧‧Valve Controller

234‧‧‧Turbo pump controller

236‧‧‧Control Data Hub

240‧‧‧Communication lines

242‧‧‧Communication lines

244‧‧‧Communication lines

246‧‧‧Communication lines

248‧‧‧Communication lines

250‧‧‧Communication lines

252‧‧‧Communication lines

254‧‧‧Communication links

256‧‧‧Communication lines

260‧‧‧Independent sensor

262‧‧‧Independent Sensor

264‧‧‧Independent Sensor

270‧‧‧Communication lines

272‧‧‧Communication lines

274‧‧‧Communication lines

280‧‧‧Measurement sensor data hub

282‧‧‧Analysis computer

284‧‧‧Communication lines

286‧‧‧Communication link

290‧‧‧Communication links

The invention is illustrated by way of example and not by way of limitation, and the same reference

Figure 1 shows a simplified block diagram of the processing chamber.

2 shows a simplified block diagram of a processing chamber having an on-line control processing device in one embodiment of the present invention.

Figure 3 shows the hierarchical relationship between sensors in one embodiment of the invention.

4 shows a simplified flow diagram in an embodiment of the present invention showing an implementation of an on-line control processing method for performing virtual measurements.

Figure 5 shows a simplified flow diagram in an embodiment of the invention showing the implementation of an online control program for providing instant control capabilities.

200‧‧‧Processing room

202‧‧‧Measurement tools

204‧‧‧ links

206‧‧‧Manufacture factory main controller

208‧‧‧Processing Module Controller

210‧‧‧ links

212‧‧‧Substrate

214‧‧‧ lower electrode

216‧‧‧Upper electrode

218‧‧‧ Plasma

220‧‧‧Airflow controller

222‧‧‧ Temperature Sensor

224‧‧‧temperature sensor

226‧‧‧pressure sensor

228‧‧‧Matching box controller group

230‧‧‧ Radio Frequency Controller

232‧‧‧Valve Controller

234‧‧‧Turbo pump controller

236‧‧‧Control Data Hub

240‧‧‧Communication lines

242‧‧‧Communication lines

244‧‧‧Communication lines

246‧‧‧Communication lines

248‧‧‧Communication lines

250‧‧‧Communication lines

252‧‧‧Communication lines

254‧‧‧Communication links

256‧‧‧Communication lines

260‧‧‧Independent sensor

262‧‧‧Independent Sensor

264‧‧‧Independent Sensor

270‧‧‧Communication lines

272‧‧‧Communication lines

274‧‧‧Communication lines

280‧‧‧Measurement sensor data hub

282‧‧‧Analysis computer

284‧‧‧Communication lines

286‧‧‧Communication link

290‧‧‧Communication links

Claims (20)

An apparatus for implementing an automated in-line process control scheme during recipe execution, the recipe being performed on a substrate in a processing chamber of a plasma processing system, the apparatus comprising: a plurality of control loop sensors, at least for collection a first set of sensor data to facilitate monitoring of a plurality of set points of the recipe during execution of the recipe, wherein the plurality of control loop sensors operate within a process control loop, the process control loop being used Controlling the set points during execution of the recipe; a set of independent sensors for collecting at least a second set of sensor data, wherein the set of independent sensors and the process control loop for controlling the set points The path is independent of and outside the process control loop for controlling the set points; a hub for receiving at least one of the first set of sensor data and the second set of sensor data; An analysis computer communicatively coupled to the hub and configured to perform analysis of at least one of the first set of sensor data and the second set of sensor data, wherein the analysis computer package A high-speed processor to analyze large amounts of data. The apparatus for implementing an automatic on-line process control scheme during formulation execution, as described in claim 1, further comprising: a manufacturing plant main controller for selecting at least the recipe; and a processing module controller for at least Performing the recipe according to a set of established recipe set points; and a set of measurement tools for providing measurement data to at least one of the manufacturing plant main controller and the analysis computer, wherein the measurement data is used Integrate with this formula. An apparatus for implementing an automatic in-line process control scheme during formulation execution, as described in claim 1, wherein the second set of sensor data collected by the set of independent sensors includes the plurality of control loops At least a portion of the data set collected by the road sensor. An apparatus for implementing an automatic in-line process control scheme during formulation execution, as described in claim 1, wherein the second set of sensor data collected by the set of independent sensors does not include having been controlled by the plurality of The data collected by the loop sensor. An apparatus for implementing an automatic in-line process control scheme during recipe execution, as described in claim 2, wherein the analysis computer is configured to receive at least sensor calibration data, wherein the sensor calibration data includes the plurality of control loops The empirical relationship between the road sensor and the set of independent sensors. An apparatus for implementing an automated in-line process control scheme during recipe execution as described in claim 5, wherein the sensor calibration data is chamber specific. An apparatus for implementing an automated in-line process control scheme during formulation execution, as described in claim 5, wherein the analysis computer is configured to verify the first set of sensor data using the second set of sensor data. An apparatus for implementing an automatic in-line process control scheme during recipe execution, as described in claim 7, wherein the analysis computer is configured to at least create a set of virtual sensors, wherein each of the set of virtual sensors The sensor system is associated with a set of virtual parameters determined by sensor data collected by a plurality of sensors, wherein the plurality of sensors includes the set of independent sensors and the complex number a sensor that controls at least one of the loop sensors. An apparatus for implementing an automatic in-line process control scheme during formulation execution, as described in claim 8, wherein the set of dummy parameters includes at least one of ion flux, ion energy, electron density, and etch rate versus deposition rate ratio By. An apparatus for implementing an automatic in-line process control scheme during formulation execution, as described in claim 8, wherein the analysis computer is configured to establish at least a phenomenon relationship between the virtual sensor and the second set of sensor data. , where the phenomenon relationship contains At least one of the columns: a related parameter; and a parameter that can be derived from another parameter. An apparatus for implementing an automated in-line process control scheme during recipe execution as described in claim 10, wherein the analysis computer is at least used to calculate a virtual measurement to provide an instant measurement. An apparatus for implementing an automatic in-line process control scheme during formulation execution, as described in claim 11, wherein the analysis computer is configured to provide at least a set of virtual drives to provide instant process control capabilities for a set of virtual senses The recipe is adjusted when the detector value falls outside a predetermined threshold. An apparatus for implementing an automated in-line process control scheme during recipe execution, as described in claim 11, wherein the analysis computer is configured to send output from the analysis to the processing module controller, wherein the output includes a set At least one of virtual sensor set point adjustment, fault detection, classification, and multi-sensor end point. An apparatus for implementing an automated in-line process control scheme during recipe execution as described in claim 13 wherein the set of virtual sensor setpoint adjustments is used to adjust at least one recipe set point. A method of implementing an automated in-line process control scheme during formulation execution, the formulation being performed on a substrate in a processing chamber of a plasma processing system, the method comprising the steps of: drawing a recipe for substrate processing of the substrate; Providing sensor calibration data to an analysis computer, wherein the sensor calibration data includes an empirical relationship between a set of control loop sensors and a set of independent sensors; adjusting the recipe to a set of recipe settings Point; execute the recipe; Receiving a first set of sensor data from the set of control loop sensors and a second set of sensor data from the set of independent sensors; analyzing the first set of sensor data and the second set of senses At least one of the tester data to calculate a set of virtual measurements; comparing the set of virtual measurements to a predetermined threshold; and generating a warning and an alert if the set of virtual measurements is outside the predetermined threshold At least one of them. A method of implementing an automated in-line process control scheme during recipe execution as described in claim 15 wherein the analysis occurs for a predetermined period of time. A method of implementing an automated inline process control scheme during recipe execution as described in claim 16 wherein the virtual metrology is calculated using a phenomenon model. The method for implementing an automatic online process control scheme during recipe execution as described in claim 17 further includes determining the presence of a fault if the set of virtual measurements is outside the predetermined threshold. The method of implementing an automatic on-line process control scheme during recipe execution, as described in claim 18, further includes determining a set of adjusted recipe set points. A method of implementing an automated in-line process control scheme during recipe execution as described in claim 19, further comprising determining a set of virtual drives to adjust the recipe.