WO2023118929A1

WO2023118929A1 - Method of inspecting a carrier transport system, vacuum processing apparatus, computer program, and computer-readable storage medium

Info

Publication number: WO2023118929A1
Application number: PCT/IB2021/062095
Authority: WO
Inventors: Christian Wolfgang Ehmann
Original assignee: Applied Materials, Inc.
Priority date: 2021-12-21
Filing date: 2021-12-21
Publication date: 2023-06-29
Also published as: CN118435334A

Abstract

A method of inspecting a carrier transport system that includes drive units for transporting a carrier through a vacuum processing apparatus along a transport path and levitation magnets for magnetically holding at least a first part of a carrier weight during the carrier transport is provided. The method includes: (i) measuring, with one or more sensors provided at the carrier, sensor data characteristic of the carrier transport as a function of time or carrier position during the carrier transport; and (ii) determining, in a computer implemented process, root causes of anomalies in the sensor data. Positions of the root causes in the vacuum processing apparatus can be localized based on occurrence times of the anomalies in the sensor data. The root causes of the anomalies may be determined utilizing a computer implemented model of the carrier transport system, particularly a digital twin of the carrier transport system. Alternatively or additionally, the root causes of the anomalies may be determined utilizing a trained machine learning model.

Description

METHOD OF INSPECTING A CARRIER TRANSPORT SYSTEM, VACUUM PROCESSING APPARATUS, COMPUTER PROGRAM, AND COMPUTER- READABLE STORAGE MEDIUM

FIELD

[0001] Embodiments of the present disclosure relate to the inspection of carrier transport systems for transporting carriers through a vacuum processing apparatus, for example through a vacuum deposition system. The inspection can diagnose problems in carrier transport. Embodiments further relate to a vacuum processing apparatus with a carrier transport system, to a computer program configured to run in a data processing unit of a vacuum processing apparatus, as well as to a computer-readable storage medium with such a computer program.

BACKGROUND

[0002] A vacuum processing apparatus is an apparatus that includes at least one vacuum processing chamber with a processing area, in which a substrate can be processed, e.g. coated. The vacuum processing apparatus can be a vacuum deposition system for depositing a material on the substrate with a deposition source. Several methods are known for the deposition of a material on a substrate. For example, a substrate may be coated by using a physical vapor deposition (PVD) process, such as a sputtering process or an evaporation process, a spraying process, or a chemical vapor deposition (CVD) process.

[0003] Coating processes, i.e. material deposition processes, may be considered for large area substrates, e.g. in display manufacturing technology. The tendency towards larger substrates, e.g. in manufacturing larger displays, results in larger vacuum processing apparatuses. Coating processes can also be considered for semiconductor processing, e.g. deposition on wafers, specifically for manufacturing electronic devices.

[0004] A substrate is typically transported through the vacuum processing apparatus while being held on a carrier, which is a carrying device for carrying the substrate. The carrier that carries the substrate is typically transported through the vacuum processing apparatus along a transport path T, for example into a vacuum processing chamber that houses a deposition source. A carrier transport system may be provided for transporting the carrier through the vacuum processing apparatus, particularly under a sub-atmospheric pressure.

[0005] Transporting a substrate held on a carrier through a vacuum processing apparatus is challenging. Specifically, the carrier transport may lead to small particles due to frictional forces between the carrier and stationary structures of the carrier transport system which can negatively affect the substrate processing. Therefore, it is beneficial to keep frictional forces low during the carrier transport. Further, extensive carrier vibrations or other unwanted carrier movements during the carrier transport may lead to substrate damage or even substrate breakage, e.g. if the substrate includes a thin glass plate.

[0006] In view of the above, it would be beneficial to improve the carrier transport through vacuum processing apparatuses to ensure a smooth and reliable carrier transport without undesirable anomalies in carrier movement. Specifically, it would be beneficial to provide a diagnosis method suitable for inspecting and improving carrier transport systems in vacuum processing apparatuses.

SUMMARY

[0007] In light of the above, a method of inspecting a carrier transport system and a vacuum processing apparatus with a carrier transport system are provided according to the independent claims. Further, computer programs and computer-readable storage mediums to be used in a vacuum processing apparatus for inspecting a carrier transport system are provided. Further features, details, aspects, implementation and embodiments are shown in the dependent claims, the description and the drawings.

[0008] According to one aspect, a method of inspecting a vacuum transportation system is provided. The vacuum transportation system includes drive units arranged along a transport path for transporting a carrier through a vacuum processing apparatus and levitation magnets for magnetically holding at least a first part of a carrier weight during the carrier transport. The method includes: measuring, with one or more sensors provided at the carrier, sensor data characteristic of the carrier transport as a function of time or as a function of carrier position during the carrier transport; and determining, in a computer implemented process, root causes of anomalies in the sensor data. In some embodiments, the positions of the root causes in the vacuum processing apparatus can be localized based on occurrence times of the anomalies in the sensor data in the computer implemented process.

[0009] In some embodiments, a computer implemented model, particularly a digital twin, of the carrier transport system is used for at least one of determining the root causes of the anomalies and localizing the positions of the root causes in the vacuum processing apparatus. Alternatively or additionally, a trained machine learning model is used for determining the root causes of the anomalies, and optionally for localizing the positions of the root causes in the vacuum processing apparatus.

[0010] According to one aspect, a vacuum processing apparatus that includes a carrier transport system is provided. The carrier transport system includes drive units arranged along a transport path for transporting a carrier through the vacuum processing apparatus, levitation magnets arranged along the transport path for magnetically holding at least a first part of a carrier weight during the carrier transport, and the carrier that is provided with one or more sensors for measuring sensor data that are characteristic of the carrier transport during the carrier transport. The vacuum processing apparatus further includes a data processing unit with a memory including instructions and with a processor, wherein the instructions, when executed by the processor, cause the data processing unit to determine root causes of anomalies in the sensor data and optionally to localize positions of the root causes in the vacuum processing apparatus based on occurrence times of the anomalies in the sensor data.

[0011] According to some embodiments, the instructions may cause the data processing unit to generate a computer implemented model, particularly a digital twin, of the carrier transport system. The digital twin can be used for determining the root causes of the anomalies. Alternatively or additionally, the memory stores a trained machine learning model that is configured to determine the root causes of the anomalies and optionally to localize the positions of the root causes in the vacuum processing apparatus.

[0012] According to one aspect, a vacuum processing apparatus that includes a carrier transport system is provided. The carrier transport system includes drive units arranged along a transport path for transporting a carrier through the vacuum processing, levitation magnets arranged along the transport path for magnetically holding at least a first part of a carrier weight during the carrier transport, and the carrier that is provided with one or more sensors for measuring sensor data that are characteristic of the carrier transport during the carrier transport. The vacuum processing apparatus further includes a memory including instructions and a processor, wherein the instructions, when executed by the processor, cause the data processing unit to generate a computer implemented model, particularly a digital twin, of the carrier transport system based on transport system data. The transport system data describe the carrier transport system and may include positional information of the drive units and of the levitation magnets and dimensional information about the carrier. The transport system data may optionally include positional information of support rollers that are arranged along the transport path and are configured to mechanically support at least a second part of the carrier weight during the carrier transport.

[0013] According to one aspect, a data processing unit including a memory that stores a digital twin of a carrier transport system of a vacuum processing apparatus as described herein is provided.

[0014] According to one aspect, a computer program is provided. The computer program includes instructions which, when the program is executed by a processor, cause the generation of a computer implemented model of a carrier transport system based on transport system data that include positional information of drive units and of levitation magnets of the carrier transport system arranged along a transport path as well as dimensional information about a carrier to be transported with the carrier transport system.

[0015] The computer implemented model may be configured to calculate forces locally or temporarily acting between the carrier and one or more stationary structures of the carrier transport system during the carrier transport for determining root causes of anomalies in the carrier transport. The stationary structures of the carrier transport system may include levitation magnets, e.g. top levitation magnets and/or bottom levitation magnets, drive units, support rollers, and/or other stationary units arranged at the transport track along which the carrier is transported by the carrier transport system.

[0016] Further, a computer-readable medium storing thereon the computer program described herein is provided. [0017] According to one aspect, a computer-readable storage medium is provided storing a trained machine learning model. The trained machine learning model is configured to receive input data based on sensor data of one or more sensors provided at a carrier and measured as a function of time or carrier position during carrier transport through a vacuum processing apparatus, the sensor data being characteristic of the carrier transport. The trained machine learning model is configured to provide, as an output, root causes of anomalies in the sensor data and optionally positions of the root causes in the substrate processing apparatus.

[0018] Embodiments are also directed at apparatuses for carrying out the disclosed methods and include apparatus parts for performing each described method aspect. These method aspects may be performed by way of hardware components, a computer programmed by appropriate software, by any combination of the two or in any other manner. Furthermore, embodiments according to the disclosure are also directed at methods for operating the described apparatuses. The methods for operating the described apparatus include method aspects for carrying out the functions of the apparatus. Embodiments are also directed at methods of manufacturing processed substrates, particularly coated substrates, in a vacuum processing apparatus described herein and/or using a carrier transport system described herein.

BRIEF DESCRIPTION OF THE DRAWINGS

[0019] So that the manner in which the above recited features of the present disclosure can be understood in detail, a more particular description of the disclosure, briefly summarized above, may be had by reference to embodiments, some of which are illustrated in the appended drawings. It is to be noted that the appended drawings illustrate only typical embodiments and are therefore not to be considered limiting of the scope, for the disclosure may admit to other equally effective embodiments.

[0020] FIG. 1 shows a schematic top view of a vacuum processing apparatus with a carrier transport system according to embodiments of the present disclosure;

[0021] FIG. 2 shows a schematic side view of a carrier transport system to be inspected according to embodiments described herein;

[0022] FIG. 3 shows a schematic sectional view of the carrier transport system of FIG. 2; [0023] FIG. 4 schematically illustrates the use of a digital twin of the carrier transport system for determining root causes of anomalies;

[0024] FIG. 5 shows a screenshot taken during use of the digital twin for determining and localizing root causes of anomalies in carrier transport;

[0025] FIG. 6 is a flow diagram illustrating a method of inspecting a carrier transport system utilizing a computer implemented model of the carrier transport system;

[0026] FIG. 7 is a block diagram illustrating a method of inspecting a carrier transport system utilizing a trained machine learning model; and

[0027] FIGS. 8A-C show examples to be used as training data for a machine learning model described herein.

DETAILED DESCRIPTION OF EMBODIMENTS

[0028] Reference will now be made in detail to the various embodiments of the disclosure, one or more examples of which are illustrated in the figures. Within the following description of the drawings, same reference numbers refer to same components. Only the differences with respect to individual embodiments are described. Each example is provided by way of explanation of the disclosure and is not meant as a limitation of the disclosure. Further, features illustrated or described as part of one embodiment can be used on or in conjunction with other embodiments to yield yet a further embodiment. It is intended that the description includes such modifications and variations.

[0029] FIG. 1 shows a vacuum processing apparatus 1000 with a carrier transport system 100 according to embodiments described herein in a schematic top view. The carrier transport system 100 is configured for transporting a carrier 110 that carries a substrate S or another object in a vacuum environment, particularly through one or more vacuum chambers of the vacuum processing apparatus. For example, the vacuum processing apparatus 1000 may include a plurality of vacuum chambers arranged next to each other, e.g., in a linear arrangement. FIG. 1 exemplarily shows the vacuum processing apparatus 1000 with a vacuum processing chamber 101 (also referred to herein as a “processing module (PM)”) that houses a processing device, particularly a deposition source 105 for coating the substrate S. The vacuum processing chamber 101 may be arranged next to a second vacuum chamber 102 (for example, a high-vacuum loading module (LM-HV)). A third vacuum chamber 103 (for example, a low-vacuum loading module (LM-PV)) may be arranged next to the second vacuum chamber 102.

[0030] The carrier 110 that carries the substrate S can be moved or conveyed along a transport path T through the vacuum processing apparatus 1000, for example in a transport direction along a first track T1 from the third vacuum chamber 103 through the second vacuum chamber 102 into the vacuum processing chamber 101, where the substrate S can be processed, e.g. coated with a material. After the processing, the substrate can be transported back into the third vacuum chamber 103 for unloading, for example along a second track T2 that may extend next to and parallel to the first track T1. A track switch can be provided for transferring the carrier from the first track T1 to the second track T2 and/or vice versa in a lateral direction L transverse to the transport direction. However, the arrangement of vacuum chambers of the vacuum processing apparatus 1000 and the transport sequence are not limited to such a setup, and FIG. 1 merely shows an exemplary setup of a vacuum processing apparatus. In some embodiments, the vacuum processing apparatus is an in-line deposition system.

[0031] In some embodiments, a vacuum processing chamber 101 of the vacuum processing apparatus houses a deposition source 105 configured to deposit a material on the substrate S that is carried by the carrier 110. The carrier transport system 100 may be configured to transport the carrier 110 to a position in which the substrate S faces toward the deposition source 105 for being coated.

[0032] FIG. 2 is a schematic side view and FIG. 3 is a schematic sectional view of the carrier transport system 100 of FIG. 1. A short portion of the carrier transport system along the transport path T is depicted in FIG. 2. FIG. 3 shows a sectional view in a sectional plane X (see FIG. 2) that is perpendicular to the transport path T.

[0033] Referring to FIGS. 1-3, the carrier transport system 100 includes drive units 113 for transporting the carrier through the vacuum processing apparatus and levitation magnets 111, 112 for magnetically holding at least a first part of the carrier weight during the carrier transport. Specifically, the drive units 113 may be configured for exerting a drive force on the carrier 110 that drives the carrier 110 along the transport path T in the transport direction. The levitation magnets 111, 112 may exert a magnetic holding force on the carrier that acts in an essentially vertical direction on the carrier and counteracts at least a part of the weight force of the carrier. In other words, a part of the carrier weight can be magnetically held during the carrier transport by the levitation magnets 111, 112. In some embodiments, the levitation magnets may also provide a lateral stabilization (or lateral guiding) of the carrier, particularly by ensuring that the carrier remains essentially at a predefined lateral position relative to the levitation magnets during the transport. Alternatively or additionally, a separate side stabilization device, e.g. one or more side stabilization magnets may be provided.

[0034] For example, the levitation magnets 111, 112 may magnetically hold 50% or more and 120% or less of a weight force of the carrier during the carrier transport. A remaining weight force of the carrier (if present) can be mechanically supported on a carrier support, such as on (optional) support rollers 114. A partial or total magnetic levitation of the carrier 110 reduces the friction during the carrier transport, such that the generation of small particles that may negatively affect the substrate processing in the vacuum processing apparatus can be reduced or avoided. The quality of the substrate processing, particularly the layer deposition quality, can be improved.

[0035] As is schematically depicted in FIGS. 1 and 2, the drive units 113 may be arranged along the transport path T, for example at predetermined distances from each other, particularly at essentially regular intervals in the transport direction. A plurality of drive units may be provided along the transport path T, for example five, ten, twenty or more drive units, each drive unit being arranged at a specific position along the transport path. For example, at least two drive units may be arranged in each vacuum chamber through which the transport path T extends, enabling a carrier transport into and/or out of the respective vacuum chamber.

[0036] In some embodiments, the drive units 113 can be arranged below the carrier 110 during the carrier transport. The drive units 113 may interact with a first counterpart 121 of the drive units 113 that is provided at a bottom portion of the carrier 110 for propelling the carrier 110 in the transport direction (see FIG. 2).

[0037] In some implementations, the drive units 113 may be linear motors configured to contactlessly drive the carrier 110 along the transport path. The first counterpart 121 arranged at the carrier may include a countermagnet and/or a ferro- or permanentmagnetic component of the carrier configured to interact with the drive units 113 for driving the carrier in the transport direction. A contactless electromagnetic driving force can be exerted by the drive units 113 on the carrier. It is to be noted that linear motors typically not only generate a force in a transport direction, but also a force in a direction perpendicular thereto (here: a downwardly directed force on the carrier in the vertical direction V), as it is schematically depicted by vertical arrows in FIG. 2. Alternatively or additionally, the drive units 113 may include a mechanical drive, such as one or more drive rollers that are driven in rotation by a motor for moving the carrier in the transport direction on the one or more drive rollers.

[0038] As is further schematically depicted in FIG 2, the levitation magnets 111, 112 may be arranged along the transport path T, for example at predetermined distances from each other, particularly at essentially regular intervals in the transport direction. A plurality of levitation magnets 111, 112 may be arranged along the transport path T, for example five, ten, twenty or more levitation magnets, each levitation magnet being arranged at a specific position along the transport path. For example, at least two levitation magnets may be arranged along the transport path T in each vacuum chamber through which the transport path T extends, enabling an at least partial carrier levitation during the carrier transport through the respective vacuum chamber. The levitation magnets 111, 112 may magnetically interact with respective counterparts provided at the carrier, particularly with ferromagnetic or permanentmagnetic components that are provided at the carrier.

[0039] In some embodiments, the levitation magnets 111, 112 may include top levitation magnets 111 that are arranged at an upper portion of a carrier transportation space, e.g. above the carrier. The top levitation magnets 111 may be configured for magnetically holding at least a part of the carrier weight during the carrier transport. The top levitation magnets 111 may magnetically interact with a second counterpart 122 that is arranged at the carrier 110. The second counterpart 122 may be a ferromagnetic or a permanentmagnetic component provided at the carrier that can be attracted by the levitation magnets (see FIGS. 2 and 3).

[0040] In some embodiments, the levitation magnets 111, 112 may include bottom levitation magnets 112 that may be provided at a lower portion of a carrier transportation space, particularly below the substrate S and/or below the top levitation magnets 111 (if present). The bottom levitation magnets 112 may be configured to magnetically hold at least a part of the carrier weight during the carrier transport. The bottom levitation magnets 112 may magnetically interact with a third counterpart 123 that is arranged at the carrier 110, particularly below the second counterpart 122. The third counterpart 123 may be a ferromagnetic or a permanentmagnetic component provided at the carrier that can be attracted by the levitation magnets (see FIGS. 2 and 3).

[0041] In some embodiments, 10% or more and 60% or less of the carrier weight may be magnetically held by the top levitation magnets 111 during the carrier transport, and/or 10% or more and 60% or less of the carrier weight may be magnetically held by the bottom levitation magnets 112 during the carrier transport. A full or partial carrier levitation may be provided. If the levitation magnets hold only a part of the carrier weight during the carrier transport, a remaining part of the carrier weight (such as 10% or more and 30% or less) may be supported on support rollers 114 during the carrier transport. Alternatively, if the carrier is transported completely contactlessly, the levitation magnets may carry the full carrier weight and no mechanical support may be provided.

[0042] In some embodiments, which can be combined with other embodiments described herein, some or all of the levitation magnets 111, 112 are passive levitation magnets, particularly permanent magnets. A passive levitation magnet may attract the carrier upwardly, e.g. by exerting an attractive magnetic force on a respective magnetic counterpart of the carrier in an upward direction, as it is schematically depicted in FIGS. 2 and 3.

[0043] In some embodiments, an active levitation magnet may be provided in alternative or in addition to one or more passive levitation magnets. An active levitation magnet is a levitation magnet, whose strength is actively controlled to maintain a predetermined gap distance between the carrier and the stationary track of the carrier transport system. However, the levitation magnets 111, 112 in the embodiment depicted in FIGS. 2 and 3 are purely passive. A passive levitation magnet is a levitation magnet whose strength is not actively controlled, such as a permanent magnet.

[0044] In some embodiments, which can be combined with other embodiments described herein, the carrier transport system further includes support rollers 114 provided along the transport path T that mechanically support at least a second part of the carrier weight during the carrier transport. A plurality of support rollers 114 may be provided along the transport path T at predetermined intervals, particularly at regular spacings, as it is schematically depicted in FIG. 1 and FIG. 2. For example, ten, twenty or more support rollers 114 may be arranged along the transport path T, each support roller being arranged at a specific position along the transport path. At least two support rollers may support the carrier at each carrier position during the carrier transport. At the carrier position depicted in FIG. 2, four support rollers 114 support the carrier simultaneously (see four arrows in FIG. 2 illustrating resulting weight forces R carried by the support rollers that are “active” at the respective carrier position, i.e. that are in contact with the carrier). There may also be carrier positions along the transport path at which the carrier is supported by less or more than four support rollers.

[0045] Accordingly, in the depicted embodiment, a first part of the carrier weight is magnetically held by the levitation magnets 111, 112, and a second part of the carrier weight is mechanically supported on the support rollers 114 during the carrier transport. The (partial) mechanical support of the carrier on the support rollers 114 can facilitate a stabilization of the carrier in the vertical direction V, the lateral direction L and/or in the transport direction T. A completely contactless levitation of a carrier by magnetic forces is challenging and typically uses an active control of the levitation magnets for maintaining a stable position of the levitated carrier, which can be more complex. On the other hand, by magnetically holding only a part of the carrier weight and mechanically supporting the remaining part of the carrier weight, the carrier can be held in a stable position via the mechanical support while at the same time reducing problems associated with frictional forces by the partial magnetic levitation. Accordingly, the carrier transport system 100 shown in FIGS. 1-3 allows for a reliable carrier transport along the transport path T while reducing the generation of small particles that would have a negative effect on the substrate processing without an unnecessary complexity.

[0046] It will be apparent that a plurality of different forces may act at each carrier position along the transport path T between different parts of the carrier and different stationary structures of the carrier transport system, including forces between the drive units 113 and the first counterpart 121 of the carrier, between the top levitation magnets 111 and the second counterpart 122 of the carrier (if present), between the bottom levitation magnets 112 and the third counterpart 123 of the carrier (if present), and between the support rollers 114 and a fourth counterpart 124 of the carrier, such as a support rail, that may be provided at the carrier for being supported on the support rollers 114 (if present). During the movement of the carrier along the transport path T, the values of said forces and the components that exert a force on the carrier continuously vary. For example, when the carrier moves out of the range of a specific levitation magnet, the holding force of said specific levitation magnet may drop to zero, but a resulting force R supported on one or more support rollers and/or a holding force generated by another levitation magnet may increase.

[0047] Providing a smooth and stable carrier transport may therefore be challenging. Specifically, it may be challenging to provide a correct alignment, positioning, and parameter setting for the plurality of components of the carrier transport system 100 that leads to a smooth and low-friction carrier transport. Further, if undesired events, such as particle generation and/or substrate damage, happen during the carrier transport, it may be difficult to diagnose the root causes for such undesired events in view of the plurality of components and interactions that happen simultaneously or in direct succession between the components of the carrier transport system, and further in view of the fact that the vacuum processing system is typically closed, i.e. not easily accessible for inspection or review. Flooding the vacuum processing system with air for inspection would cause enormous efforts and costs.

[0048] Root causes for such undesired events may, for example, include a bad alignment between specific components of the carrier transport system, a temporal overcompensation of gravity by magnetic levitation forces, scratching at loose parts, bent parts due to temperature variations, vacuum deformations, defective components, and others. For example, a temporal overcompensation of gravity by magnetic levitation forces may lead to a “lift-off’ of the carrier from the support rollers 114, whereupon a lower rail 125 of the carrier may hit the support rollers 114 (so-called “skipping”), which entails the risk of considerable carrier vibrations and even substrate damage.

[0049] In view of the above, the embodiments described herein provide computer- implemented diagnosis methods for inspecting a carrier transport system 100 of a vacuum processing apparatus 1000 as described above. Specifically, root causes of undesired events happening during carrier transport can be determined in the computer-implemented process, such that the root causes can be removed or at least reduced in a more targeted way.

[0050] According to the embodiments described herein, the carrier is provided with one or more sensors 130 configured to measure sensor data 131 characteristic of the carrier transport. The sensor data 131 are measured as a function of time and/or as a function of carrier position during the carrier transport through the vacuum processing apparatus. The sensor data may include anomalies, for example spikes, maxima or other conspicuities in the vibrational behavior of the carrier, a conspicuous frequency spectrum of the carrier vibrations, a locally occurring carrier vibration in a specific vibration direction, a local temperature or pressure drop or rise, a local drift of the carrier away from or toward a stationary structure, and/or others.

[0051] According to the embodiments of the present disclosure, the root causes of the anomalies in the sensor data are determined in a computer implemented process. A computer implemented process for determining the root causes of the anomalies in the sensor data may be quicker and may lead to more reliable results than a purely manual analysis of the sensor data for deducing causes of the anomalies therefrom, e.g. based on the experience of an operator that analyses the sensor data. In particular, according to embodiments described herein, a trained machine learning model and/or a digital twin of the carrier transport system may be used for determining the root causes of the anomalies. In some embodiments, a digital twin and a trained machine learning model are used in combination.

[0052] As is schematically depicted in FIGS. 1 and 2, the vacuum processing apparatus 1000 may include a data processing unit 200, e.g. a computer, that includes a processor 201 and a memory 202 storing instructions, e.g., in the form of a computer program 210 stored in the memory 202. The instructions, when executed by the processor 201, may cause the data processing unit to determine the root causes of anomalies that are present in the sensor data 131, e.g., using a trained machine learning model 220 and/or a digital twin of the carrier transport system.

[0053] In some embodiments, the instructions, when executed by the processor, may cause the data processing unit 200 to generate a computer implemented model 211 of the carrier transport system, particularly a digital twin of the carrier transport system. The digital twin can then be used for determining the root causes. Alternatively or additionally, the memory 202 may store a trained machine learning model 220 that is configured to receive input data based on the sensor data 131 and to provide, as an output, the root causes of the anomalies in the sensor data.

[0054] Optionally, positions of the root causes in the vacuum processing apparatus 1000 may be localized based on occurrence times of the anomalies in the sensor data 131. Specifically, after the determination of the root cause of an anomaly in the sensor data, it is possible to localize the position of the root cause in the vacuum processing apparatus, since the occurrence time of the anomaly in the sensor data - and hence also the respective carrier position at the occurrence time - is known or can be directly deduced from the sensor data. For example, if the root cause of an anomaly in the sensor data is determined to be a misaligned support roller, the position of the misaligned support roller (for example, the respective vacuum chamber that houses the misaligned support roller) can be determined from the occurrence time of the anomaly in the sensor data and can be output.

[0055] In some embodiments, which can be combined with other embodiments described herein, the one or more sensors 130 provided at the carrier include at least one accelerometer, particularly at least one vibration sensor, more particularly at least one MEMS vibration sensor, that measures carrier vibrations as a function of time during the carrier transport. The anomalies in the sensor data may include anomalies in a vibrational behavior of the carrier during the carrier transport, e.g. vibration spikes or other temporarily high vibration amplitudes and/or a locally abnormal frequency spectrum of the measured carrier vibrations at a specific carrier position.

[0056] In some embodiments, two, three or more accelerometers may be provided at the carrier, for example an accelerometer configured to measure carrier vibrations in the vertical direction V, an accelerometer configured to measure carrier vibrations in the transport direction T, and/or an accelerometer configured to measure carrier vibrations in the lateral direction L perpendicular to the transport direction.

[0057] In some implementations, the one or more sensors 130 provided at the carrier are selected from a group including an accelerometer configured to measure vertical accelerations, an accelerometer configured to measure lateral vibrations, an accelerometer configured to measure vibrations in a transport direction, a pressure sensor, a temperature sensor, a distance sensor, a vision sensor, a camera, and/or a position sensor. A temperature sensor may measure a temperature of a vacuum environment that surrounds the carrier during the carrier transport and/or a temperature of a specific carrier part, e.g. of a substrate holding surface on which the substrate S is supported. A pressure sensor may measure a pressure in a vacuum environment that surrounds the carrier during the carrier transport as a function of time and/or carrier position. A distance sensor (or “gap sensor”) may measure a distance between the carrier and a stationary structure of the carrier transport system during the carrier transport as a function of time and/or carrier position. A position sensor may measure the position of the carrier along the transport path as a function of time. The above list of sensors is not limiting and other sensor types may be arranged at the carrier. A plurality of two, three or more sensors may be provided at the carrier, such as five, ten or more sensors. In some embodiments, the sensor data are transmitted to the data processing unit 200 in a wireless way.

[0058] In some embodiments, which can be combined with other embodiments described herein, the root causes that can be determined and/or localized with the computer implemented process include locally inappropriate magnetic forces exerted on the carrier during the carrier transport, e.g. by the levitation magnets and/or the drive units. For example, a locally inappropriate magnetic levitation force may overcompensate gravity and may hence lead to “skipping”. Alternatively or additionally, the root causes that can be determined and/or localized with the computer implemented process may include one or more misaligned or defective stationary structures of the carrier transport system, in particular one or more misaligned or defective ones of the drive units, the levitation magnets, and/or the support rollers, or a step in a transition between two adjacent vacuum chambers.

[0059] In some embodiments, which can be combined with other embodiments described herein, the method further includes removing or reducing the determined root causes of the anomalies, particularly by at least one or more of the following: locally increasing or decreasing a density of the levitation magnets and/or of the drive units; locally reducing or increasing a magnetic force exerted on the carrier by one or more of the levitation magnets and the drive units; locally reducing or increasing a density of support rollers 114 provided along the transport path; aligning or displacing one or more of the drive units 113, the levitation magnets 111, 112, the support rollers 114, or the vacuum chambers; servicing, repairing or replacing one or more of the drive units 113, the levitation magnets 111, 112, the support rollers 114, or the vacuum chambers; locally adapting the pressure and/or the temperature in the vacuum processing apparatus; repairing or modifying the carrier.

[0060] In some embodiments, which can be combined with other embodiments described herein, a computer implemented model 211 of the carrier transport system is used for at least one of determining the root causes of the anomalies and for localizing the positions of the root causes in the vacuum processing apparatus 1000. The computer implemented model may include or may be a digital twin of the carrier transport system.

[0061] A “digital twin” may be understood as a computer-implemented model of the carrier transport system that is generated by a data processing unit based on transport system data that describe components of the “real” carrier transport system. For example, a digital twin may be a “digital copy” of the actual carrier transport system that can be used for determining root causes of anomalies in carrier transport. A “digital copy” is understood herein in a broad sense as the transport system data that describe the carrier transport system in a computer memory, together with a plurality of programmed functionalities for retrieving inferences about the carrier transport from said transport system data, such as for example, forces exerted in the real system between the carrier and stationary structures of the carrier transport system during carrier transport, positions of potentially problematic components or problematic sections along the transport path, or identification of potentially defective or misaligned components. The transport system data may, for example, include information (e.g., positional information, dimensional information and/or configuration or setting information) about the drive units, the levitation magnets, the carrier, and/or the support rollers (if present).

[0062] A computer program 210 may be stored in the memory 202 of the data processing unit 200 that includes instructions that, when executed, cause the data processing unit 200 to generate the computer implemented model 211 of the carrier transport system, particularly to generate the digital twin of the carrier transport system. The computer implemented model 211 may be generated based on the transport system data that describe the “real” carrier transport system. The transport system data can be entered by a user of the data processing unit or can already be stored in the memory 202 of the data processing unit. Alternatively or additionally, transport system data of a “virtual” carrier transport system that is to be modelled can be used for the computer implemented model 211. Alternatively or additionally, the transport system data from which the computer implemented model 211 is generated can be partially data that describe the “real” carrier transport system and partially data that deviates from the “real” carrier transport system, e.g. in order to model whether root causes can be reduced or removed by a modification of the real carrier transport system. [0063] The transport system data from which the digital twin is generated by the computer program 210 may include one or more of the following: (i) positional information of the drive units 113 and/or the levitation magnets 111, 112 arranged along the transport path T; (ii) positional information of the support rollers 114 arranged along the transport path T; (iii) dimensional and/or positional information about the carrier 110 that is to be transported with the carrier transport system, such as dimensional information about counterparts of the carrier that interact with the drive units, the levitation magnets, and optionally with the support rollers during the carrier transport, particularly dimensions and/or positions of any one or more of the first counterpart 121, the second counterpart 122, the third counterpart 123, and the fourth counterpart 124 (if respectively present), (iv) details about the vacuum chambers of the vacuum processing apparatus, such as positions of transitions between adjacent vacuum chambers, or dimensions of the vacuum chambers along the transport path; (v) settings and/or configurations of any one or more of the drive units 113 and the levitation magnets 111, 112, such as generated drive forces, generated levitation forces, or generated forces exerted on the carrier in another direction. Further information about the carrier transport system may (optionally) be used for generating the digital twin, e.g., positions and details of track switches, vacuum rotation modules and/or carrier loading and unloading modules of the vacuum processing apparatus. Optionally, at least some components of the carrier transport system may be labelled with respective identifiers (for example, the support rollers of the system may be numbered), such that a correlation between real components and representations of components that are part of the digital twin is given, and an easy identification is possible.

[0064] The computer program 210 may be configured for generating the digital twin based on the transport system data. The digital twin may include a plurality of programmed functionalities, such as functions for calculating specific forces acting at specific carrier positions, functions that return root causes of anomalies in the carrier transport, functions that return positions of such root causes in the vacuum processing apparatus, and/or functions for generating a graphical representation of the carrier transport system, e.g. on a display of the data processing unit. Optionally, the sensor data 131 of the one or more sensors 130 is provided as input data to the digital twin, which may facilitate the determination of the root causes of anomalies and/or the localization of the root causes. [0065] FIG. 4 schematically illustrates the use of a digital twin of the carrier transport system for determining root causes of anomalies. The upper part of FIG. 4 shows a graphical representation of the carrier transport system 100’ generated by the digital twin, and the lower part of FIG. 4 shows calculation results of the digital twin that indicate the root cause of an anomaly in the sensor data.

[0066] In some embodiments, the computer implemented model is configured to generate a graphical representation of the carrier transport system 100’, which can for example be shown on a screen of the data processing unit. The graphical representation may include representations of the drive units 113’, of the levitation magnets 111’, 112’, and/or of the support rollers 114’, shown along a representation of the transport path T’. The graphical representation may further include a representation of the carrier 110’ shown at a settable position along the representation of the transport path T’. The graphical representation may optionally include one or more of the counterparts of the carrier, e.g., a representation of the first counterpart, a representation of the second counterpart 122’, a representation of the third counterpart 123’, and/or a representation of the fourth counterpart 124’, shown along the representation of the transport path T’. The graphical representation may depict the respective components of the carrier transport system at correct relative positions relative to each other along the transport path and/or with correct relative dimensions in the transport direction, such that the user can directly infer from the graphical representation what components are interacting with each other at a specific carrier position along the carrier transport path. The representation of the transport path T’ may be a graph with an x-axis showing the extension of the transport path, e.g. in meters.

[0067] In FIG. 4, the carrier is exemplarily depicted at a position corresponding to the 11- meter position of the real transport path. At the shown carrier position, the carrier is supported on four active support rollers, depicted as hatched diamonds, whereas the inactive support rollers are depicted as empty diamonds in the graphical representation. Further, at the shown carrier position along the transport path, the carrier is slightly tilted relative to a perfectly horizontal orientation and is arranged slightly below a target level in the vertical direction, depicted by the tilted representation of the carrier 110’ having a body center at about -0.7 mm relative to a target vertical level that is indicated as the zero-level at the left y-axis. [0068] In some embodiments, the computer implemented model may calculate a force exerted between any of the components of the carrier transport system at a specific carrier position along the transport path. For example, the computer implemented model may calculate forces exerted on the support rollers 114 during the carrier transport, e.g., at a specific carrier position or as a function of carrier position. In FIG. 4, the resulting forces R acting on the four active support rollers at the shown carrier position are depicted as black diamonds 401 (see right y-axis indicating the respective force values).

[0069] In some embodiments, the carrier implemented model may calculate a mean force on the support rollers that are respectively active as a function of carrier position and/or as a function of time during carrier transport. In the lower part of FIG. 4, the mean force exerted on the active support rollers as a function of time and as a function of carrier position is depicted as a graph 402 (the right y-axis indicating the mean force value on the active rollers).

[0070] In some embodiments, which can be combined with other embodiments described herein, the computer implemented model may calculate and output one or more forces temporarily and/or locally acting on or exerted by one or more of the carrier, the drive units, the levitation magnets, and the support rollers. One or more calculated forces may be displayed in a graph, as it is schematically depicted in FIG. 4 (see black diamonds 401 indicating forces temporarily acting on specific support rollers and the graph 402 indicating a mean force on the support rollers as a function of time and carrier position).

[0071] If the force exerted on one or more of the support rollers calculated by the computer implemented model 211 is temporarily out of a predetermined range, the computer implemented model may determine a locally inappropriate magnetic force exerted by the levitation magnets and/or the drive units on the carrier as the root cause for an anomaly in the sensor data. For example, the mean force exerted by the carrier on the support rollers at the 11 -meter position of the transport path is close to zero, as is shown at reference numeral 403. Said mean force value may be out of a predetermined range, i.e. the weight force of the carrier may be nearly overcompensated by the levitation magnets at the 11 -meter position of the transport path which may entail a risk of skipping and substrate damage. Accordingly, the root cause for an anomaly 132 in the sensor data 131 of an acceleration sensor at an occurrence time of about 20 seconds (corresponding to the 11 -meter position of the carrier along the transport path) may be determined to be an overly strong levitation force. The root cause can be removed by, for example, locally decreasing the density of the levitation magnets at the 11 -meter position.

[0072] In some embodiments, the sensor data 131 of the one or more sensors 130, particularly of one or more accelerometers arranged at the carrier, may be provided as an input to the computer implemented model of the carrier transport system. As is shown in the lower part of FIG. 4, the sensor data 131 can be depicted by the computer implemented model in a graph as a function of time and/or carrier position, and/or the sensor data 131 can be used by the computer implemented model for determining the root causes of anomalies. The sensor data 131 shown in FIG. 4 exemplarily include the anomaly 132, namely a local maximum of the carrier vibrations (here: in all three directions), at an occurrence time of about 20 seconds that corresponds to a specific carrier position along the transport path. Said specific position along the transport path can be output by the computer implemented model as the position of a root cause of the anomaly 132, together with the identified root cause. In some embodiments, the computer implemented model can be used to correlate the anomalies in the sensor data 131 with the root causes identified by the computer implemented model. For example, the computer implemented model may cause the display of the sensor data 131 as a function of carrier position, together with forces acting between the carrier and selected stationary structures of the carrier transport systems as a function of carrier position, facilitating a correlation between the anomalies and the root cause and the determination of the correct root causes and the localization of said root causes.

[0073] FIG. 5 shows a screenshot taken during use of the computer implemented model. The sensor data 131 measured by one or more accelerometers arranged at the carrier are depicted as a function of time and/or as a function of carrier position. The sensor data 131 include one or more anomalies, including the anomaly 132 (here: local maximum of carrier vibrations at an occurrence time of about 2105 seconds). Further, the computer implemented model may cause a display of a selection of components of the carrier transport system that are active as a function of time and/or carrier position during the carrier transport, facilitating the correlation between the anomaly 132 and the root cause thereof. Alternatively or additionally, the computer implemented model may directly output the components of the carrier transport system that are active at the occurrence times of an anomaly 132. In the lower part of FIG. 5, the support rollers that are active (i.e. that respectively support the carrier) as a function of time during the carrier transport are shown. The anomaly 132 can hence be correlated with the respectively active support rollers. Here, the support roller 135 (support roller No. 3 in the LM-HV) is identified as the root cause for the anomaly 132 and, at the same time, the location of said root cause is identified.

[0074] FIG. 6 is a flow diagram illustrating a method of inspecting a carrier transport system utilizing a computer implemented model of the carrier transport system, particularly a digital twin as described herein.

[0075] In box 620, sensor data are measured with one or more sensors provided at a carrier during the transport of the carrier through a vacuum processing apparatus with a carrier transport system, the carrier transport including propelling the carrier along a transport path with drive units and magnetically holding at least a first part of a carrier weight during the carrier transport with levitation magnets.

[0076] The one or more sensors may include at least one accelerometer that measures carrier vibrations as a function of time and/or carrier position during the carrier transport. The sensor data are characteristic of the carrier transport and may include anomalies, such as local vibration maxima and/or locally unexpected frequencies in the vibration spectrum.

[0077] In box 630, root causes of the anomalies in the sensor data are determined with the computer implemented model, particularly with the digital twin, and positions of the root causes in the vacuum processing apparatus may optionally be localized. The sensor data may be given as an input to the digital twin. The digital twin may be configured to calculate and/or output forces acting between the carrier and selected stationary structures of the carrier transport system, e.g. the support rollers. The root causes of the anomalies may, for example, be deduced from the calculated forces and/or based on a correlation between the occurrence times of the anomalies and the components of the carrier transport system that are active at said occurrence times.

[0078] In box 640, the root causes may be removed or reduced, e.g. by aligning misaligned components, by repairing defective components and/or by at least locally modifying the forces that are exerted on the carrier by stationary structures of the carrier transport system. For example, the levitation force locally exerted on the carrier may be reduced, in order to reduce the risk of skipping. [0079] As is illustrated by box 610, the computer implemented model may be generated utilizing a computer program that runs on a data processing unit. The computer implemented model may be generated based on transport system data provided as input to the data processing unit, the transport system data describing at least some features of the actual carrier transport system, including, for example, positional information of the drive units, the levitation magnets and/or of the support rollers of the carrier transport system arranged along the transport path, and dimensional information about a carrier 110 that is to be transported with the carrier transport system.

[0080] According to one aspect described herein, a computer program that is configured to generate the computer implemented model, particularly the digital twin, based on transport system data as described herein is provided. The computer implemented model is typically configured to calculate forces locally and/or temporarily acting between the carrier and one or more stationary structures of the carrier transport system during carrier transport, particularly for determining root causes of anomalies in sensor data characteristic of the carrier transport. Specifically, the computer implemented model may be configured to calculate forces locally acting on or exerted by one or more of the carrier, the drive units, the levitation magnets, and the support rollers as a function of time and/or carrier position during the carrier transport.

[0081] Further, a computer-readable storage medium having the computer program stored thereon is provided according to embodiments described herein.

[0082] As described above, the computer implemented model of the carrier transport system can be used for determining the root causes of the anomalies in the sensor data. In alternative or in addition thereto, a trained machine learning model can be used for determining the root causes of the anomalies in the sensor data in some embodiments described herein. The trained machine learning model may be integrated as a functionality in the digital twin, such that the root causes can be determined in a computer implemented process based on a digital twin in combination with a trained machine learning model. The trained machine learning model may be stored in a memory of a data processing unit of the vacuum processing apparatus. Optionally, the trained machine learning model can also be used for localizing positions of the root causes in the vacuum processing apparatus. [0083] FIG. 7 is a block diagram illustrating a method of inspecting a carrier transport system utilizing a trained machine learning model 220. The trained machine learning model can be trained in a preceding training procedure.

[0084] As is schematically depicted in FIG. 7, the trained machine learning model 220 may be configured to receive input data 136 that is based on the sensor data 131 of the one or more sensors 130, particularly of one or more accelerometers provided at the carrier. The input data 136 may correspond to the sensor data 131, or the sensor data 131 may be processed before being entered as the input data 136 to the trained machine learning model 220. The processing of the sensor data 131 for providing the input data 136 may include, for example, (i) extracting one or more portions of the sensor data that include the anomalies, (ii) digitalization, (ii) combining the sensor data of several sensors for providing a set of input values from different sensors, (iii) adding additional data to the sensor data for providing the input data, such as time, carrier position, settings of processing tools, settings of components of the carrier transport system, and/or a configuration of the vacuum processing apparatus, (iv) Fourier transforming the sensor data or one or more fractions thereof for providing one or more frequency spectrums, e.g., a vibration spectrum of the carrier vibrations, or a variation of the vibration spectrum of the carrier over time. The processing of the sensor data 131 for providing the input data 136 is depicted in FIG. 7 by box 701 and may be conducted by the data processing unit, particularly in an automated way.

[0085] The trained machine learning model 220 may be configured to provide, as an output, the root causes 702 of anomalies in the sensor data 131, and optionally the positions of the root causes 702 in the vacuum processing apparatus.

[0086] In some embodiments, a machine learning model 221 may be trained in a preceding training process 705 for providing the trained machine learning model 220. The machine learning model 221 may be trained in the training process 705 with training data 137. The training data 137 may map input data 136 (based on sensor data 131 with anomalies) to corresponding known root causes of said anomalies.

[0087] Examples with respect to the provision of the training data 137 are depicted in FIGS.

8A-8C. [0088] FIG. 8A shows sensor data of an acceleration sensor measured as a function of time during the carrier transport. A periodic anomaly appears in the vibrational behavior of the carrier during the carrier transport, particularly in the form of periodic maxima (spikes) in the measured vertical carrier acceleration. The known root cause of the shown anomaly is a vertical displacement at the transitions between adjacent vacuum chambers from chamber to chamber, for example due to floor sagging or vacuum deformation. The vertical displacement between adj acent vacuum chambers that leads to the depicted anomaly is about 0.8 mm. The acceleration level of the spikes is in the range of about 0.5 g. Preventive maintenance could be done by aligning the transitions between adjacent vacuum chambers.

[0089] A set of training data 137 can be created from the sensor data depicted in FIG. 8 A and the respective known root cause of the anomaly in said sensor data (vertically misaligned transitions (0.8 mm) between adjacent vacuum chambers).

[0090] FIG. 8B shows sensor data of an acceleration sensor measured as a function of time during the carrier transport. An anomaly similar to the anomaly of FIG. 8A appears in the vibrational behavior of the carrier in the form of periodic maxima (spikes) in the measured vertical carrier acceleration, though at a higher amplitude. The known root cause for the shown anomaly is a vertical displacement at the transitions between adjacent vacuum chambers from chamber to chamber, the displacement being about 1.2 mm. Since the acceleration of the spikes is in the range of almost 1g, an immediate alignment of the transitions is beneficial for preventive substrate damage.

[0091] A set of training data can be created from the sensor data depicted in FIG. 8B and the respective known root cause of the anomaly in said sensor data (vertically misaligned transitions (1.2 mm) between adjacent vacuum chambers).

[0092] FIG. 8C shows sensor data of an acceleration sensor measured as a function of time during the carrier transport. Also here, an anomaly in the form of periodic maxima in the vibrational behavior of the carrier is visible. However, the periodicity, the amplitude, and the characteristics of the spikes are different. The known root cause of the shown anomaly is a vertical bending of the carrier rail (i.e., the fourth counterpart 124 depicted in FIG. 2) that is supported on the support rollers during the carrier transport, e.g. a bimetallic bending due to an uneven temperature of the carrier). This leads to a vibration maximum in the vertical direction caused by each of the support rollers during the carrier transport, hence the high periodicity of the anomaly.

[0093] A set of training data can be created from the sensor data depicted in FIG. 8C and the respective known root cause of the anomaly in said sensor data (bent fourth counterpart 124).

[0094] Figures 8A-C show some non-limiting examples of root causes of anomalies that can be determined by the trained machine learning model, after the training of the machine learning model with the respective training data. Many further root causes can be determined with a machine learning model that is trained accordingly. For example, an anomaly that includes carrier vibrations in a lateral direction may be present in the sensor data in the event of a lateral or angular misalignment of a specific component; the anomaly may be nonperiodic, e.g. if one specific component is misaligned or defective; the anomaly may have a specific frequency that may cause vibrations, e.g. if a vacuum pump, a motor, or another rotary device of the vacuum processing apparatus is defective. It is difficult to determine the root causes of anomalies in the sensor data, if a plurality of (periodic and non-periodic) anomalies caused by several different root causes are present at the same time in the sensor data. The trained machine learning model may be capable of determining a plurality of root causes generating a complex plurality of anomalies that are simultaneously present in the sensor data of a plurality of sensors.

[0095] The trained machine learning model may associate specific features in the sensor data with corresponding specific weights, based on the training in the training process. The machine learning model may be a one-layer model (that summarizes specific features in the sensor data multiplied by the respective weight) or may be a two- or multi-layer model, specifically in the form of a neuronal network, as is typical in machine learning.

[0096] According to an aspect described herein, a computer-readable storage medium storing a trained machine learning model 220 is provided. The trained machine learning model 220 is configured to receive input data based on sensor data of one or more sensors 130 provided at a carrier and measured as a function of time or carrier position during carrier transport through a vacuum processing system. The trained machine learning model may be configured to provide, as an output, root causes of anomalies in the sensor data. The determined root causes can then be reduced or removed for improving the carrier transport and for reducing undesired events, such as carrier vibrations and skipping. [0097] Further, a method of training a machine learning model for providing a trained machine learning model as described herein is provided. The machine learning model may be trained with training data that relates input data (based on the sensor data with the anomalies) to respective known root causes of the anomalies in the sensor data.

[0098] In embodiments described herein, a computer implemented model of the carrier transport system is provided that may provide functionalities that return forces exerted during the carrier transport and locations thereof based on a mathematical model of the carrier transfer. A detailed analysis of sensor signals, particularly of vibration signals, of the carrier can be done that allows to (i) investigate certain areas for errors/mismatches from intended design, (ii) determining root causes of anomalies in carrier transport, (iii) suggestions for corrective actions.

[0099] The apparatuses and systems described herein may be configured to move and process large area substrates that may in particular have a surface of 1 m² or above. The term “substrate” may particularly embrace substrates like glass substrates, for example, a glass plate. Further, a substrate may include wafers, slices of transparent crystal such as sapphire or the like. However, the term “substrate” may embrace other substrates that can be inflexible or flexible, like e.g. a foil or a web. The substrate may be formed by any material suitable for material deposition. According to some embodiments of the present disclosure, which can be combined with other embodiments described herein, the substrate is configured for display manufacturing and may in particular be a large area substrate.

[0100] Embodiments described herein particularly relate to deposition of materials, e.g. for display manufacturing on large area substrates. According to some embodiments, large area substrates or carriers supporting one or more substrates may have a size of at least 1 m² For instance, the vacuum processing apparatus may be adapted for processing large area substrates, such as substrates of GEN 5, which corresponds to about 1.4 m² substrates, GEN 7.5, which corresponds to about 4.29 m² substrates, GEN 8.5, which corresponds to about 5.7 m² substrates, or even GEN 10, which corresponds to about 8.7 m² substrates. Even larger generations such as GEN 11 and GEN 12 can similarly be implemented. Alternatively or additionally, semiconductor wafers may be processed and coated in vacuum processing apparatuses according to the present disclosure.

[0101] In particular, the following embodiments are described herein: Embodiment 1 : A method of inspecting a carrier transport system (100) that comprises drive units (113) for transporting a carrier (110) through a vacuum processing apparatus (1000) along a transport path (T) and levitation magnets (111, 112) for magnetically holding at least a first part of a carrier weight during carrier transport, the method comprising: measuring, with one or more sensors (130) provided at the carrier, sensor data (131) characteristic of the carrier transport as a function of time or carrier position during the carrier transport; and determining, in a computer implemented process, root causes of anomalies in the sensor data (131).

Embodiment 2: The method of embodiment 1, further comprising: localizing positions of the root causes in the vacuum processing apparatus based on occurrence times of the anomalies in the sensor data.

Embodiment 3: The method of embodiment 1 or 2, wherein the one or more sensors (130) comprise at least one accelerometer, particularly at least one vibration sensor, more particularly at least one MEMS vibration sensor, that measures carrier vibrations as a function of time or carrier position during the carrier transport, and wherein the anomalies comprise anomalies in a vibrational behavior of the carrier during the carrier transport.

Embodiment 4: The method of any of embodiments 1 to 3, wherein a computer implemented model (211) of the carrier transport system is used for at least one of determining the root causes of the anomalies and for localizing the positions of the root causes in the vacuum processing apparatus.

Embodiment 5: The method of embodiment 4, wherein the computer implemented model comprises a digital twin of the carrier transport system.

Embodiment 6: The method of embodiment 4 or 5, wherein the carrier transport system (100) further comprises support rollers (114) provided along the transport path (T) that mechanically support at least a second part of the carrier weight during the carrier transport, and the computer implemented model (211) calculates a force exerted on the support rollers as a function of time or carrier position during the carrier transport.

Embodiment 7: The method of any of embodiments 4 to 6, wherein the computer implemented model is generated based on transport system data comprising positional information of the drive units (113), the levitation magnets (111, 112), and optionally support rollers (114) that are provided along the transport path, as well as dimensional information about respective counterparts (121, 122, 123, 124) of the carrier that interact with the drive units, the levitation magnets and optionally the support rollers during the carrier transport, particularly wherein the computer implemented model (211) calculates, based on the transport system data, forces exerted on the support rollers (114) as a function of time or carrier position during the carrier transport.

Embodiment 8: The method of any of embodiments 4 to 7, wherein the computer implemented model (211) is configured to generate a graphical representation of the carrier transport system on a display that includes representations of the drive units (113’), of the levitation magnets (111’, 112’), and optionally of support rollers (114’), shown along a representation of the transport path (T’), and that further includes a representation of the carrier (110’) shown at a settable position along the representation of the transport path (T’), optionally together with one or more forces temporarily and/or locally acting on or exerted by one or more of the carrier, the drive units, the levitation magnets, and the support rollers.

Embodiment 9: The method of any of embodiments 4 to 8, wherein the root causes that can be determined using the computer implemented model (211) comprise locally inappropriate magnetic forces exerted on the carrier by the levitation magnets and/or the drive units, particularly if the force exerted on one or more of the support rollers calculated by the computer implemented model (211) is temporarily out of a predetermined range.

Embodiment 10: The method of any of embodiments 4 to 9, wherein the root causes that can be determined and/or localized using the computer implemented model comprise one or more misaligned or defective ones of the drive units, the levitation magnets, and/or the support rollers.

Embodiment 11 : The method of any of embodiments 1 to 10, wherein a trained machine learning model (220) is used for at least one of determining the root causes of the anomalies in the sensor data and localizing the positions of the root causes in the vacuum processing apparatus.

Embodiment 12: The method of embodiment 11, further comprising: training a machine learning model with training data that map input data based on sensor data of the one or more sensors that include anomalies to corresponding known root causes of said anomalies to provide the trained machine learning model.

Embodiment 13: The method of embodiment 11 or 12, wherein the trained machine learning model receives input data based on the sensor data that include the anomalies and provides, as an output, the root causes of the anomalies and/or the positions of the root causes in the vacuum processing apparatus.

Embodiment 14: The method of any of embodiments 1 to 13, wherein the one or more sensors (130) provided at the carrier are selected from a group including an accelerometer configured to measure vertical accelerations, an accelerometer configured to measure lateral vibrations, an accelerometer configured to measure vibrations in a transport direction, a pressure sensor, a temperature sensor, a distance sensor, a vision sensor, a camera, and a position sensor.

Embodiment 15: The method of any of embodiments 1 to 14, further comprising removing or reducing the root causes of the anomalies by at least one or more of the following: locally increasing or decreasing a density of at least one of the levitation magnets (111, 112) and the drive units (113); locally reducing or increasing a magnetic force exerted by at least one of the levitation magnets (111, 112) and the drive units (113) on the carrier; reducing or increasing a density of support rollers (114) provided along the transport path; aligning or displacing one or more of the drive units (113), the levitation magnets (111, 112) and the support rollers (114); and servicing, repairing or replacing one or more of the drive units (113), the levitation magnets (111, 112) and the support rollers (114).

[0102] While the foregoing is directed to implementations of the present disclosure, other and further implementations of the disclosure may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.

Claims

1. A method of inspecting a carrier transport system (100) that comprises drive units (113) for transporting a carrier (110) through a vacuum processing apparatus (1000) along a transport path (T) and levitation magnets (111, 112) for magnetically holding at least a first part of a carrier weight during carrier transport, the method comprising: measuring, with one or more sensors (130) provided at the carrier, sensor data (131) characteristic of the carrier transport as a function of time or carrier position during the carrier transport; and determining, in a computer implemented process, root causes of anomalies in the sensor data (131).

2. The method of claim 1, further comprising: localizing positions of the root causes in the vacuum processing apparatus based on occurrence times of the anomalies in the sensor data.

3. The method of claim 1, wherein the one or more sensors (130) comprise at least one accelerometer that measures carrier vibrations as a function of time or carrier position during the carrier transport, and wherein the anomalies comprise anomalies in a vibrational behavior of the carrier during the carrier transport.

4. The method of claim 1, wherein a computer implemented model (211) of the carrier transport system is used for at least one of determining the root causes of the anomalies and for localizing the positions of the root causes in the vacuum processing apparatus.

5. The method of claim 4, wherein the computer implemented model comprises a digital twin of the carrier transport system.

6. The method of claim 4 or 5, wherein the carrier transport system (100) further comprises support rollers (114) provided along the transport path (T) that mechanically support at least a second part of the carrier weight during the carrier transport, and the computer implemented model (211) calculates a force exerted on the support rollers as a function of time or carrier position during the carrier transport.

7. The method of claim 4 or 5, wherein the computer implemented model is generated based on transport system data comprising positional information of the drive units (113), the levitation magnets (111, 112), and optionally support rollers (114) that are provided along the transport path, as well as dimensional information about respective counterparts (121, 122, 123, 124) of the carrier that interact with the drive units, the levitation magnets and optionally the support rollers during the carrier transport, particularly wherein the computer implemented model (211) calculates, based on the transport system data, forces exerted on the support rollers (114) as a function of time or carrier position during the carrier transport.

8. The method of claim 4 or 5, wherein the computer implemented model (211) is configured to generate a graphical representation of the carrier transport system on a display that includes representations of the drive units (113’), of the levitation magnets (111’, 112’), and optionally of support rollers (114’), shown along a representation of the transport path (T’), and that further includes a representation of the carrier (110’) shown at a settable position along the representation of the transport path (T’), optionally together with one or more forces temporarily or locally acting on or exerted by one or more of the carrier, the drive units, the levitation magnets, and the support rollers.

9. The method of claim 4 or 5, wherein the root causes that can be determined using the computer implemented model (211) comprise locally inappropriate magnetic forces exerted on the carrier by at least one of the levitation magnets and the drive units, particularly if the force exerted on one or more of the support rollers calculated by the computer implemented model (211) is temporarily out of a predetermined range.

10. The method of claim 4 or 5, wherein the root causes that can be determined using the computer implemented model comprise one or more misaligned or defective ones of the drive units, the levitation magnets, and/or the support rollers.

11. The method of any of claims 1 to 5, wherein a trained machine learning model (220) is used for at least one of determining the root causes of the anomalies in the sensor data and localizing the positions of the root causes in the vacuum processing apparatus.

12. The method of claim 11, further comprising: training a machine learning model with training data that map input data based on sensor data of the one or more sensors that include anomalies to corresponding known root causes of said anomalies to provide the trained machine learning model.

13. The method of claim 11, wherein the trained machine learning model receives input data based on the sensor data that include the anomalies and provides, as an output, at least one of the root causes of the anomalies and the positions of the root causes in the vacuum processing apparatus.

14. The method of any of claims 1 to 5, wherein the one or more sensors (130) provided at the carrier are selected from a group including an accelerometer configured to measure vertical accelerations, an accelerometer configured to measure lateral vibrations, an accelerometer configured to measure vibrations in a transport direction, a pressure sensor, a temperature sensor, a distance sensor, a vision sensor, a camera, and a position sensor.

15. The method of any of claims 1 to 5, further comprising removing or reducing the root causes of the anomalies by at least one or more of the following: locally increasing or decreasing a density of at least one of the levitation magnets (111, 112) and the drive units (113); locally reducing or increasing a magnetic force exerted by at least one of the levitation magnets (111, 112) and the drive units (113) on the carrier; reducing or increasing a density of support rollers (114) provided along the transport path; aligning or displacing one or more of the drive units (113), the levitation magnets (111, 112) and the support rollers (114); and servicing, repairing or replacing one or more of the drive units (113), the levitation magnets (111, 112) and the support rollers (114).

16. A vacuum processing apparatus (1000) with a carrier transport system (100) that comprises: drive units (113) arranged along a transport path (T) for transporting a carrier (110) through the vacuum processing apparatus; levitation magnets (111, 112) arranged along the transport path (T) for magnetically holding at least a first part of a carrier weight during carrier transport; and the carrier (110) that is provided with one or more sensors (130) for measuring sensor data (131) that are characteristic of the carrier transport during the carrier transport, wherein the vacuum processing apparatus (1000) further comprises a data processing unit (200) with a memory comprising instructions and with a processor (201), wherein the instructions, when executed by the processor, cause the data processing unit to determine root causes of anomalies in the sensor data (131).

17. The vacuum processing apparatus of claim 16, wherein the memory stores a trained machine learning model (220) that is configured to receive input data based on the sensor data that include the anomalies and to provide, as an output, the root causes of the anomalies.

18. The vacuum processing apparatus of claim 16 or 17, wherein the instructions cause the data processing unit to generate a computer implemented model (211) of the carrier transport system, particularly a digital twin of the carrier transport system.

19. The vacuum processing apparatus of claim 18, wherein the computer implemented model (211) of the carrier transport system (100) is generated based on transport system data that comprises one or more of the following: positions of the drive units along the transport path; positions of the levitation magnets along the transport path; positions of support rollers for mechanically supporting at least a second part of the carrier weight along the transport path; and dimensional information about counterparts of the carrier that interact with at least one of the drive units, the levitation magnets, and the support rollers; wherein the computer implemented model is configured to calculate forces locally acting on or exerted by one or more of the carrier, the drive units, the levitation magnets, and the support rollers during the carrier transport.

20. A computer program (210) comprising instructions which, when the computer program is executed by a processor, cause a data processing unit (200) to generate a computer implemented model (211) of a carrier transport system (100) based on transport system data that comprises positional information of drive units (113) and of levitation magnets (111, 112) of the carrier transport system arranged along a transport path (T) as well as dimensional information about a carrier (110) to be transported with the carrier transport system, wherein the computer implemented model (211) is configured to calculate forces locally or temporarily acting between the carrier and one or more stationary structures of the carrier transport system during carrier transport for determining root causes of anomalies in the carrier transport.

21. A computer-readable storage medium having stored thereon the computer program (210) of claim 20.

22. A computer-readable storage medium storing a trained machine learning model (220) configured to receive input data based on sensor data (131) of one or more sensors (130) provided at a carrier (110) and measured as a function of time or carrier position during carrier transport through a vacuum processing apparatus (1000), the sensor data (131) being characteristic of the carrier transport, the trained machine learning model (220) configured to provide, as an output, root causes of anomalies in the sensor data (131).