DE102023204001A1 - Improved Run-to-Run Controller in Semiconductor Manufacturing - Google Patents

Abstract

Method (40) for setting a process control for manufacturing a component in production, comprising the following steps: Obtaining (S1) measurement results (MET, ART) characterizing the component. Determining (S2) a process step (14) with which the component is to be processed. Determining (S3) a parameterization of the process step by means of an R2R controller for the process step (14) depending on the measurement results (MET, ART). The method (40) is characterized in that either the measurement results are virtual measurement results (ART) or virtual measurement results (ART) and physical measurement results (MET).

Description

The invention relates to a method for improving a run-to-run controller (R2R controller) by taking artificial measurement data into account, as well as to a device, a computer program and a machine-readable storage medium which are configured to carry out the method.

State of the art

It is known from the state of the art that process control systems (e.g. R2R controllers, statistical process control (SPC system)) receive measurements from processed wafers. The measurements are typically physical measurements on the wafer. Based on physical measurements, product parameters relevant to process control can be determined. These measurements are, for example, so-called inline measurements, which are measured on the product using physical measuring devices. The product parameters are, for example, a thickness, depth, width of structures, specific resistance, etc. In contrast, there are process or recipe parameters that describe a manufacturing step (exposure, etching, polishing, deposition, tempering, processing time, pressure, gas flow, dose) on the wafer.

Based on the physical measurements, a recipe suggestion is determined for the next processing step of the same wafer or for the same processing step of the next wafer using the run-to-run controller (R2R), i.e. in order to make appropriate adjustments if the measurements are not within a required range.

The processing step is controlled by piece goods control; this means that the recipe for the subsequent processing step for a wafer or batch (batch of wafers) is influenced by the R2R controller. A recipe characterizes the respective manufacturing or processing step, such as exposure, etching, deposition, polishing. However, the R2R controller does not intervene during the processing of the wafer/batch. This type of control is also known as R2R process control. Run-to-run controllers for the production of semiconductors are well known, see e.g. Moyne, J. (2014). Run-to-Run Control in Semiconductor Manufacturing. In: Baillieul, J., Samad, T. (eds) Encyclopedia of Systems and Control. Springer, London. https://doi.org/10.1007/978-1-4471-5102-9 255-1 .

The R2R controller essentially serves to compensate for disturbances such as long-term drifting in production by making appropriate recipe adjustments in subsequent processing in order to keep relevant inline parameters at the desired target value.

The disadvantage, however, is that product and process parameters that arise during production cannot be used directly for process control because the data path is undefined, nonexistent or difficult to implement, e.g. quality parameters of incoming material, runtime data, process data (FDC), etc. This means that product parameters and variables derived from them from any data source cannot be used for process control tasks.

Advantages of the invention

The aim of the present invention is to provide a virtual measuring device ("artificial, virtual sensor") which, in addition to or as an alternative to the physical measuring process (MET), provides further data to the R2R controller so that the R2R controller can control more reliably and precisely.

Differentiation from Virtual Metrology (VM): with VM, artificial measurement values are generated based on process parameters. In contrast, the present virtual sensor is intended to report any data and software-based, artificial measurements - not just process parameters - back to the process control, e.g. the R2R controller.

Within the manufacturing execution system (MES), a high level of transparency is ensured by disclosing the data path from the data source to the process control system.

A further advantage is the flexible combination of measurement recipes, which can be created as desired, e.g. simple heuristics up to complex data science applications such as aggregations, classifications or estimation methods.

In summary, it can be said that the invention allows the physical measuring process to be supplemented or replaced by a software-based method for the R2R controller.

disclosure of the invention

In a first aspect, the invention relates to a method for setting a process control that is used to control a production machine during the manufacture of a component in production or manufacturing. The component is preferably a semiconductor component, such as a sensor or chip. It should be noted that the values determined according to the method of the first aspect process control can be used to manufacture one or a plurality of components. For example, semiconductor components on a wafer or in a lot/batch can be processed essentially the same according to the determined process control.

The process begins with obtaining or receiving measurement results (MET, ART) that characterize the component. The measurement results are either virtual measurement results (ART) or virtual measurement results (ART) and physical measurement results (MET). The physical measurement results can be understood to mean that these measurement results were recorded on the component using a real measuring device. The virtual measurement results can be understood to mean that these measurement results are determined based on representative (measurement) quantities.

A particular predefined process step can then be received, with which the component is to be processed. The process step can be defined using a manufacturing recipe for the component or can be specified in another way. In addition to the process step, an optimal process window can also be provided for the process step, so that the R2R controller can parameterize the process step more reliably by taking the process window into account. The process step therefore describes how the measured component is to be further processed in the immediate next step, or which processing step is to be carried out with which setting, which is given by the determined parameterization of the R2R controller. It is also conceivable that the process step is used for at least one subsequent component with regard to the measured component from a series of components to be processed.

This is followed by determining a parameterization of the process step using the R2R controller for the process step depending on the measurement results (MET, ART). The R2R controller can be set up for wafer-to-wafer or lot-to-lot or other control.

It is proposed that the virtual measurement result (ART) was recorded with a virtual measuring device, whereby the virtual measuring device carries out a measuring process by software using at least one data query. A software-based measuring process in the form of measuring recipes or computer-implemented measuring algorithms is therefore carried out by the virtual measuring device. The virtual measuring device can also be referred to as a virtual sensor. The virtual measuring device can be set up to carry out one or more different measuring recipes and thus output one or a plurality of virtual measuring results.

It is also proposed that the virtual measuring device can output the same measuring device states, in particular E10 states, as a physical measuring device. The advantage here is that it is now possible to easily integrate the virtual measuring device into established processes, e.g. as a plug-in. The rules that apply to physical measuring devices are therefore valid and applicable without fundamental changes. It should be noted that from the point of view of the R2R controller, the virtual measuring device is perceived as a measuring device. In particular, due to the communication between the R2R controller and the virtual measuring device, the virtual measuring device is like a real measuring device for the R2R controller. However, the virtual measuring device can be purely a computer program and run in the factory or on a server.

It is further proposed that after executing the process step with the determined parameterization of the R2R controller, at least one downstream measurement process is carried out on the processed component, whereby the measurement results of the downstream measurement process are provided to the R2R controller as further measurement results, whereby the measurement results of the downstream measurement process are either virtual measurement results (ART) or virtual measurement results (ART) and physical measurement results (MET). The advantage here is that feedback to the R2R controller with regard to its specification of the parameterization is more detailed.

In further aspects, the invention relates to a device and a computer program, each of which is configured to carry out the above methods, and to a machine-readable storage medium on which this computer program is stored.

• 1 schematisch einen bekannten R2R Controller;
• 2 schematisch eine Ausführungsform des erfinderischen R2R Controller mit virtuellem Sensor;
• 3 schematisch eine Kommunikation zwischen einem MES und dem virtuellen Sensor sowie dem R2R-Contoller;
• 4 schematisch ein Flussdiagramm einer Ausführungsform des erfinderischen Verfahrens.

Embodiments of the invention are explained in more detail below with reference to the accompanying drawings. In the drawings:

• 1 schematic of a well-known R2R controller;
• 2 schematically an embodiment of the inventive R2R controller with virtual sensor;
• 3 schematically a communication between an MES and the virtual sensor and the R2R controller;
• 4 schematically a flow diagram of an embodiment of the inventive method.

Description of the embodiments

1 shows schematically an information flow representation 10 of an R2R controller, which is known from the prior art.

The information flow representation 10 shows a process flow 11 for producing a semiconductor component. In a first step, one or a plurality of physical measurements 12 are carried out on a wafer. A measurement 12 can be a thickness, width, depth, etc. of the wafer. The measurement results MET are then forwarded to the R2R controller 13 and made available.

Depending on the measurement results MET, the R2R controller 13 will set a process step 14 for the wafer. In other words, the R2R controller 13 outputs a setting for the process step 14 depending on the measurement results MET, so that after executing the process step 14, the processed wafer has product parameters as close as possible to a specified value range.

Optionally, after executing the process step 14, one or a plurality of downstream physical measurements 15 can be carried out on the processed wafer in accordance with the setting output by the R2R controller 13. The post-MET measurement results obtained can be made available to the R2R controller 13 as feedback so that it can make better settings for subsequent process steps 14 depending on the subsequently obtained measurement results.

2 shows schematically an information flow representation 20 of an embodiment of the R2R controller 16 according to the invention. The illustrated embodiment of the 2 differs, among other things, from the design of the 1 in that the R2R controller receives measurements from a virtual sensor 16 as additional input.

The virtual sensor 16 is handled as a virtual measuring device, which produces artificial measurement results ART and provides them to the R2R controller 13. The illustrated embodiment of the 2 differs from the design of the 1 in that the R2R controller 13 uses these additional measurement results ART of the virtual sensor 16 to additionally adjust the process step 14 depending on the artificial measurement results ART.

With the virtual measuring device it is thus possible to provide the factory control system (MES), in particular the R2R controller, with software-based measurements in addition to or as an alternative to the physical measurements (MET).

The advantage of a virtual measuring device is that from the perspective of the factory control system (MES), it does not necessarily have to be recognizable that no physical measurement has been carried out. The virtual measuring device ART therefore behaves like a physical measuring device in relation to the R2R controller. This can be achieved by the virtual measuring device using the same standby states (e.g. according to the E10 model) as physical measuring devices. The virtual measuring device then pretends to the outside world that it is a real measuring device. Preferably, the virtual measuring device behaves in the same way as a physical measuring device by communicating like a real measuring device. This means that the virtual measuring device can be easily integrated into the existing factory control system without fundamental changes, while maintaining the applicable rules.

3 shows schematically the communication between the MES and the virtual measuring device/virtual sensor as well as the R2R controller.

4 schematically shows a flow chart 40 of an embodiment of the method for setting a process control for producing a semiconductor device.

The process begins with step S1. In this step, measurement results (MET, ART) characterizing the component are provided. The measurement results are recorded according to 2 , i.e. the measurement results are either virtual measurement results ART or virtual measurement results ART and physical measurement results MET.

This is followed by a determination S2 of a process step 14 with which the component is to be processed. The information regarding the process step 14 can be provided via a database query.

This is followed by a determination S3 of a parameterization of the process step by means of an R2R controller for the process step 14 depending on the measurement results (MET,ART).

Optionally, after step S3, the determined parameterization can be output S4 for process step 14 for processing the component. Optionally, a system can be controlled S5 with the determined parameterization.

The concept described above according to 2 and/or procedures according to 4 can be used as follows. The different Applications are shown schematically in 2 indicated by the measurement recipes R1-R4. A measurement recipe defines how the virtual sensor performs the measurement (mathematical formation rule) and where the data comes from (data source).

In a first application example, supplier data is output by the virtual measuring device. The supplier data can be stored in a database of the factory control system (MES). For example, wafer thicknesses are retrieved from the database by the virtual measuring device as a measurement result ART. These are provided to the R2R controller as measured values, which specifies a grinding time for the processing step based on these values.

The aim of this R2R controller is to produce a specified wafer thickness and/or a wafer surface that is as uniform as possible.

The R2R controller can specify a proportional increase in the grinding time for processing by directly assigning the wafer thickness determined by the virtual measuring device relative to the desired wafer thickness.

The advantage here is that it can be guaranteed that the wafers are uniformly thick and thus each wafer can be used under the highest quality requirements.

In a second application example, FDC process data is output by the virtual measuring device.

The virtual sensor can perform anomaly detection of process parameters or measurement results, i.e. based on data as set on the production machine or as executed by the recipe.

For example, anomaly detection of process data and parameters such as temperature, pressure or gas flow can be carried out in order to detect abnormal fluctuations. These anomalies are then transmitted to the R2R controller as measurements. The R2R controller can then make appropriate recipe changes, for example initiate appropriate countermeasures if excessive fluctuations occur.

It is also conceivable that the extent to which the respective production machine can be trusted is determined based on the strength of the fluctuation. If the fluctuation is strong, it can be concluded that the machine should be trusted less. The R2 R-Controller determines a conservative recipe adjustment for the next processing if strong fluctuations have occurred. In other words, unreliability of the production machines can be detected and this unreliability can be taken into account accordingly in the process control.

In a third application example, geometry data is output by the virtual measuring device. A semiconductor coordinate database contains, among other things, the geometric design of the chip or the entire wafer, in particular geometries such as open area information. The background to this is that etching processes have different effects depending on the size of the etching area. This means that the time setting for the etching duration depends on the size of the open area (small/large area). Using the R2R controller, any intermediate stage of a design can be tapped depending on the geometry data and used continuously to control the etching process. This procedure can also be used for other manufacturing steps, such as implantation.

In a fourth application example, metadata is output by the virtual measuring device. Metadata can be data that is latently contained in a database. There is no independent measuring device for this data type or data sets derived based on meta information. For example, a period of time between the completion of a manufacturing step and the measurement process can be contained in the database by time stamps of the individual data points and made available to the process control system using the virtual measuring device.

In a fifth application example, physical measurements are (completely) replaced by virtual measurements. For example, certain product parameters can be predicted based on the measurement of process parameters (data source) using neural networks (or mathematical formulas), i.e. the results of virtual metrology are made accessible to the process control system (e.g. R2R controller) using the virtual measuring device.

In a sixth application example, machine states are output by the virtual measuring device. The virtual measuring device can, for example, predict wear on machine components based on physical measurement results MET. The R2R controller can then either adjust the recipe to achieve the same process results depending on the current wear or trigger a repair/maintenance. In concrete terms, the lamp runtime for epitaxy systems, the pad lifespan for polishing systems or the target lifespan for deposition systems can be determined using the virtual current measuring device and thus make it available to the process control in order to execute automated R2R controller actions (e.g. various reset scenarios or recipe corrections) based on this.

In a seventh application example, a machine learning system is used for the virtual measuring device. The execution of data-intensive AI processes (decision-making processes, classifications, cluster analysis, estimation and learning processes) often does not take place directly in the production environment, but in derived, offline environments. In order to make the results of such processes available to the production environment, the virtual measuring device can serve as a bridge between the AI algorithm and the process control system, or output the results as virtual measurement results ART. In concrete terms, distinguishable product-specific properties (e.g. speed ratings, product subgroups) can be optimally summarized using cluster analysis and then suggested to the R2R controller as a context grouping.

QUOTES INCLUDED IN THE DESCRIPTION

This list of documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions.

Cited non-patent literature

• Moyne, J. (2014). Run-to-Run Control in Semiconductor Manufacturing. In: Baillieul, J., Samad, T. (eds) Encyclopedia of Systems and Control. Springer, London. https://doi.org/10.1007/978-1-4471-5102-9 255-1 [0004]

Claims (8)

Method (40) for setting a process control for producing one or a plurality of components in production, comprising the following steps: obtaining (S1) measurement results (MET, ART) characterizing the or the plurality of components; determining (S3) a parameterization of a predetermined process step (14) for the or the plurality of components or subsequent components by means of an R2R controller depending on the measurement results (MET, ART); characterized in that either the measurement results are virtual measurement results (ART) or virtual measurement results (ART) and physical measurement results (MET). procedure according to claim 1 , wherein the virtual measurement result (ART) was recorded with a virtual measuring device (16), wherein the virtual measuring device (16) determines the measurement result by software by means of at least one data query. procedure according to claim 2 , wherein the virtual measuring device (16) is configured to output measuring device states, in particular E10 states, like a physical measuring device. procedure according to claim 2 or 3 , wherein the virtual measuring device (16) carries out a measuring process by software by means of at least one data query and by means of at least one data processing step of a result of the data query. Method according to one of the preceding claims, wherein after execution of the process step (14) with the determined parameterization at least one downstream measuring process is carried out on at least one component which was processed according to the process step (14), wherein the measurement results of the downstream measuring process are provided to the R2R controller as further measurement results, wherein the measurement results of the downstream measuring process are virtual measurement results (ART) and/or physical measurement results (MET). Device which is arranged to carry out the method according to one of the preceding claims. Computer program comprising instructions which, when executed by a computer, cause the computer to carry out the method according to claim 1 until 5 to execute. Machine-readable storage medium on which the computer program is claim 7 is stored.

Images