Substrate deformation detection and correction

Abstract

Disclosed are methods and apparatus for detecting and correcting deformation of an incoming substrate. A substrate is positioned in a first process chamber, and the presence and type of substrate bow is detected in the first process chamber. Based on the substrate bow detection and a determination of whether the substrate exhibits compressive or tensile bow, a substrate processing routine is selected for execution. The substrate processing routine can be executed in the first process chamber or in a second process chamber to correct or mitigate the bow before or during further processing of the substrate.

Description

Background Art Technical Field

[0002] Embodiments of the present disclosure relate generally to process chambers for fabricating substrates, and more particularly, to methods and apparatus for substrate deformation detection and correction in a process chamber.

[0003] Related technical notes

[0004] Substrates may be received from suppliers with either deformed or flat profiles. In some cases, after various processing operations, the substrates may become deformed or further deformed (including bowing). This deformation can reduce processing accuracy and cause damage to the substrate. Incoming substrates can be processed to reduce or remove this bowing. However, these processes often result in overcorrection or undercorrection of the deformation, failing to adequately address the bowing issue.

[0005] Therefore, there is a need to be able to detect and correct incoming substrate deformations. Summary of the Invention

[0006] The present disclosure generally relates to detecting and correcting deformation of an incoming substrate. In one example, a method for substrate processing includes: generating a plasma in a first process chamber while a substrate is positioned in the first process chamber; and generating a fingerprint of the substrate based on a plurality of sensors in the first process chamber. The method may further include comparing the fingerprint to a plurality of stored fingerprint models to determine whether the substrate is deformed; and based on determining that the substrate is deformed, selecting a substrate processing program for the substrate to correct the deformation of the substrate. In another example, a method for substrate processing includes: generating a fingerprint of a substrate positioned in a first process chamber; comparing the fingerprint to a plurality of stored fingerprint models, wherein each fingerprint model is associated with compressive bending or tensile bending; and based on the comparison, selecting a substrate processing program for the substrate.

[0007] In another example, a computer-readable storage medium is configured to execute instructions to cause a system to: generate a plasma in a first process chamber, with a substrate positioned in the first process chamber; and generate a fingerprint of the substrate based on a plurality of sensors in the first process chamber, the plurality of sensors being configured to detect low-frequency reflected power or high-frequency reflected power. The system may be further configured to compare the fingerprint with a plurality of stored fingerprint models to determine whether the substrate is deformed; and, based on determining that the substrate is deformed, select a substrate processing program for the substrate to correct the deformation of the substrate. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order that the above-described features of the present disclosure may be understood in detail, a more particular description of the present disclosure, briefly summarized above, may be had by reference to embodiments, some of which are illustrated in the accompanying drawings. It should be noted, however, that the drawings illustrate only exemplary embodiments and are therefore not to be considered limiting of the scope, as the disclosure may admit to other equally effective embodiments.

[0009] Figure 1A and Figure 1B Detectable substrate deformation according to an embodiment of the present disclosure is shown.

[0010] Figure 2 is a partial schematic diagram of a system for detecting and correcting deformation of an incoming substrate according to an embodiment of the present disclosure.

[0011] Figure 3 A method of analyzing a plurality of low frequency (LF) refractions and classifying the plurality of LF refractions in a first process chamber according to an embodiment of the present disclosure.

[0012] Figure 4 is a graph of percentage of first process chamber output power versus time for a plurality of substrates having both tensile and compressive bows according to an embodiment of the present disclosure.

[0013] Figure 5 is a graph of peak low frequency power reflected for multiple substrates according to an embodiment of the present disclosure.

[0014] Figure 6 A method for detecting deformation of an incoming substrate according to an embodiment of the present disclosure is provided. DETAILED DESCRIPTION

[0015] The present disclosure relates to detecting and correcting deformation (bow) of incoming substrates. Substrates used to manufacture electronic devices, including semiconductors and solid-state memory devices, undergo multiple processing operations. Substrates received from substrate manufacturing suppliers or from internal suppliers or upstream operations may be received as flat substrates. However, in some examples, substrates may deform during upstream operations, resulting in inferior devices and / or scrapping of substrates in downstream operations (including inspection). Such deformation (including bow) may result in damage or scrapping due to the challenges of placing the curved substrate on a substrate support pedestal. If the substrate is not properly placed on the pedestal, subsequent processing of the substrate (including material deposition and patterning) may be challenging. This is particularly true in atomic layer deposition (ALD), chemical vapor deposition (CVD) and physical vapor deposition (PVD) systems, as well as other systems designed for forming thin metal, dielectric or combination layers, including those using plasma.

[0016] For example, if a substrate is positioned on a substrate support base (or other support member) so that the substrate does not rest flat against the support surface, backside defects may occur, which may result in reduced substrate quality and which may negatively impact further processing operations. For example, if there is a gap or gaps between the substrate and the substrate support member, contamination may form on the bottom of the substrate. This contamination may spread between multiple chambers of the manufacturing system as the substrate moves between chambers. Incoming substrate curvature can be corrected, however, substrates with different types and degrees of curvature may respond differently to corrective measures. Conventional processing methods apply a single correction scheme to all incoming substrates, regardless of the presence, type, or degree of defects. The degree of curvature can be measured in units of height (such as micrometers or nanometers). Depending on the type of curvature, the degree of curvature can be positive or negative.

[0017] Figure 1A and Figure 1B FIGURE 1 illustrates substrate deformation that can be detected by the systems and methods discussed herein. Figure 1A and Figure 1B As shown, the substrate deformation can adopt a curved shape, such as Figure 1A Compression bending as shown or Figure 1B The tensile bending shown. Figure 1A In compression bending in , the bottom center surface 104 of the substrate 102 is in contact with the chamber's substrate support pedestal 106. However, the edge 108 of the substrate 102 faces upward and away from the substrate support pedestal 106 and is not in contact with the substrate support pedestal 106. This can make transport and handling of the substrate 102 challenging and can cause problems during thin film deposition. Conversely, Figure 1B The tensile bending shown in FIG results in the bottom center surface 114 of the substrate 110 not being in contact with the substrate support pedestal 106. Instead, the edge 112 of the substrate 110 is in contact with the substrate support pedestal 106. Furthermore, a cavity 116 is formed below the bottom center surface 114 of the substrate 110 due to the substrate bending.

[0018] The systems and methods discussed herein may employ a first process chamber that is optionally coupled to a second process chamber and / or additional process chambers. The process chamber has a substrate support pedestal capable of applying a voltage to the process chamber and is configured to generate a plasma within the process chamber. As discussed below, the amount of power used to generate the plasma in the first process chamber is monitored. This power monitoring is used to determine the presence, type, and extent of substrate bow. This determination is made by comparing a profile of the power used to generate the plasma in the first process chamber (which may be referred to as a fingerprint) with multiple models based on a history of fingerprint generation and analysis of multiple substrates across one or more first process chambers. The models are generated from multiple historical data regarding substrate deformation and deformation correction. In one example, the models are experimentally determined and stored in a database for later reference. The models are dynamically updated as additional substrates are processed using the systems and methods discussed herein. The fingerprints for each of compressive and tensile bow are associated with different features that facilitate model-based fingerprint characterization. The models are specific to at least the compressive and tensile bow types, and in some examples, some or all of the models may further be specific to the extent of bow within each type. Each model can be associated with (e.g., linked to) at least one substrate processing procedure configured to correct for bow. Based on the comparison, a substrate processing procedure is selected and executed to correct or mitigate substrate bow. In such an example, the substrate undergoes processing (such as a deposition or etching process) while reducing bow. Such processing improves process uniformity.

[0019] The methods discussed herein include determining: (1) when a substrate is bent, (2) the type of substrate bending (which may be tensile or compressive), and (3) the extent of the bending. A substrate processing procedure is selected based on one or more of these determinations, and the bending can be corrected in a first process chamber where the bending is measured. The first process chamber can be coupled to a second process chamber, such as a CVD, PVD, or PE-CVD process chamber, among other chambers. A transfer chamber can be used to transfer the substrate between multiple or two chambers. Alternatively, the substrate bow can be corrected in the second process chamber.

[0020] In one example that may be combined with other examples, power measurements including forward power and reflected power are obtained during operation of the first process chamber. The forward power is the power supplied to an element (e.g., a substrate support pedestal) in the first process chamber from, for example, an RF power source. The reflected power is the power lost during plasma maintenance, including those losses due to resistive losses and capacitive losses, among others. Therefore, the difference between the forward power and the reflected power is the power delivered to the load. The reflected power readings (which can be used for low-frequency or high-frequency power as discussed herein) are used to generate fingerprints that are used to generate models. These models are associated with substrate processing procedures that facilitate bow correction. The difference in the fingerprints generated between compression bow and tension bow is used to generate a model. The model indicates which substrate processing procedure to select to correct the identified substrate bow.

[0021] In the methods discussed herein, reflected power is monitored for each substrate positioned in a first process chamber. A fingerprint generated by monitoring power reflections from each substrate is compared to a model to determine the presence, type, and, in some examples, extent of bow, so that a substrate handling procedure (e.g., a model) can be selected to correct the bow, as the model is associated with the substrate handling procedure. In some embodiments, the substrate handling procedure is performed in the first process chamber to correct the bow, which can save processing costs, as transferring (robotic handling) a substrate that is at least compressively bowed from the first process chamber to a second process chamber can result in damage to the substrate. In alternative embodiments, the substrate handling procedure is performed in a process chamber other than the first process chamber to correct the bow.

[0022] Each substrate processing procedure is associated with one or more variables. The variables may include: an electrostatic voltage applied to the substrate support base, a power level and frequency for forming the plasma, a gas mixture, a gas flow rate of each gas or gas combination in the gas mixture, a pressure, a time or multiple times associated with the application of the voltage, and other appropriate parameters. The type of process parameters and the range of process parameters associated with each substrate processing procedure can be determined experimentally using data from previously processed substrates. Such data can be stored in a database in which illustrative or example fingerprints are associated with the processing procedures. According to examples herein, the type and range of these process parameters can be dynamically updated using data from substrates processed for substrate deformation. The substrate discussed herein as an incoming substrate may have one or more layers formed thereon. The layers may include silicon formed from tetraethyl orthosilicate (TEOS) and its oxide layer, and a nitride layer (such as silicon nitride), or a stack of alternating oxide-nitride layers.

[0023] Figure 2is a partial schematic diagram of a system 200 for detecting and correcting deformation of an incoming substrate. The system 200 includes a first process chamber 206 and an optional second process chamber 208. The first process chamber 206 and the second process chamber 208 can be used to form and maintain a plasma and to deposit thin film metal, dielectric, and composite layers on a substrate and / or to pattern the substrate. In one example, the system 200 includes a substrate staging apparatus 202 (e.g., a front end for receiving a substrate cassette) and a substrate transport apparatus 204 coupled to the substrate staging apparatus 202. The substrate transport apparatus 204 is configured to move substrates individually or in batches into the first process chamber 206. The first process chamber 206 includes a plurality of sensors 210 coupled to one or more impedance matching circuits 219 (the combination of which may be referred to as "automatching"). The plurality of sensors 210 can be configured to detect low-frequency or high-frequency reflected power when a plasma is generated in the first process chamber 206. The second process chamber 208 can be similarly equipped.

[0024] The first process chamber 206 is configured to receive a substrate from the substrate conveyor 204. In an embodiment, the substrate received from the substrate conveyor 204 has been subjected to a previous operation. The previous operation may include a deposition process to form one or more layers on the surface of the substrate, including Si. x O y and Si x N y In some examples, one or more layers may have been patterned in a previous operation. The incoming substrate received by the first process chamber 206 may include layers deposited to a thickness ranging from 0.1 microns to 10 microns, or other thicknesses. The methods discussed herein mitigate bowing on such substrates without negatively impacting the structure and function of the layer or stack of layers on the substrate.

[0025] In embodiments that may be combined with other embodiments, the first process chamber 206 is configured to heat the substrate one or more times to detect the presence and type of bow in the substrate, and in some configurations, may be configured to form or remove a layer on the substrate. In one example, the first process chamber 206 is configured to generate a capacitively coupled plasma therein. In this example, the first process chamber 206 includes one or more electrodes 222 (two are shown), a gas manifold 218, and a plurality of thermocouples positioned in a wall of the first process chamber 206 or in a substrate support 220. The electrodes 222 may be fully or partially embedded within the substrate support 220, coupled to the substrate support 220, or both. The gas manifold 218 facilitates the distribution of a plurality of ionizable gases into the first process chamber 206. Such gases include argon (Ar), helium (He), krypton (Kr), xenon (Xe), or other gases or combinations of gases capable of forming a plasma.

[0026] One or more RF power generation devices 214 are coupled to the first process chamber 206 and the second process chamber 208 and are configured to apply power to generate plasma in the chambers 206 and 208. Although two RF power generation devices 214 are shown, it is contemplated that each of the first process chamber 206 and the second process chamber 208 may share an RF power generation device 214. The one or more RF power generation devices 214 facilitate the formation of plasma within the first process chamber 206 and / or the second process chamber 208 when one or more gases, such as argon (Ar), helium (He), krypton (Kr), or xenon (Xe), are present in the first process chamber 206 and / or the second process chamber 208.

[0027] The controller 224 is coupled to the substrate staging apparatus 202, the substrate transfer apparatus 204, the first process chamber 206, and the second process chamber 208. A plurality of substrate handling programs are stored in the non-volatile memory 216 and are accessible by the controller 224, which is configured to execute the plurality of substrate handling programs. Each substrate handling program is configured to adjust one or more characteristics of conditions within the first process chamber 206 or the second process chamber 208 to mitigate substrate bow. The controller 224 can be configured to execute instructions associated with one or more applications / programs of the system 200.

[0028] A plurality of sensors 210 coupled to or disposed in the first process chamber 206 are employed to determine whether an incoming substrate has a curvature. As discussed below, the plurality of sensors 210 are further configured to determine whether a curvature is present and whether the curvature is compressive or tensile. Subsequently, based on the presence and type of curvature of the incoming substrate, a plurality of logics stored in a non-volatile memory 216 (a non-volatile computer-readable medium) are executed to select a substrate processing program to correct the determined curvature. Each substrate processing program may be associated with a substrate process (e.g., film formation, patterning, cleaning, etc.) in either or both of the first process chamber 206 or the second process chamber 208. Each of the stored substrate processing programs is associated with a type of curvature (compressive or tensile). In some examples, at least some of the substrate processing programs may be further associated with a degree of curvature. The degree of curvature may be defined as positive or negative, and / or as a numerical value or a range of values.

[0029] In one example, a substrate processing program is selected based on one or more of the type or degree of bow detected by the sensor 210. The selected program is executed by the controller 224 to process the substrate to mitigate bow by applying a voltage to the substrate support 220 (e.g., the electrostatic chuck thereof) of the first process chamber 206 and / or the substrate support pedestal 226 of the second process chamber 208. For example, a predetermined amount of voltage may be applied to the electrostatic chuck (on which the substrate is positioned). Depending on the selected program, application of the predetermined voltage reduces or eliminates substrate bow without applying excessive stress that might otherwise occur without a customized program selection. In addition to voltage, other process parameters may also be used to correct for substrate bow. Furthermore, the processing program may also direct the formation of a plasma in the first process chamber 206 or the second process chamber 208.

[0030] Figure 3 A method 300 is provided for analyzing and classifying a plurality of low-frequency (LF) reflections in a first process chamber. Low-frequency power, as discussed herein, refers to power below approximately 500 kHz, and high-frequency power, as discussed herein, is equal to or greater than approximately 13.56 MHz. In method 300, at operation 302, one or more substrates are loaded individually into a first process chamber. Optionally, the first process chamber is configured to control the one or more substrates within a temperature range of 200° C. to 500° C. while maintaining a vacuum within the first process chamber. Optional heating facilitates one or more of processing the substrates, simulating processing conditions experienced by the substrates, and reducing substrate bowing. At operation 304, a capacitive plasma is generated in the first process chamber after the substrates are received at operation 302. In one example, at operation 304, the plasma is generated by applying a current of approximately 500 kHz to the first process chamber while flowing argon (Ar), helium (He), krypton (Kr), xenon (Xe), or other gas at a pressure of approximately 1 Torr to approximately 20 Torr.

[0031] In operation 306, the amount of reflected power of the generated plasma is measured. A graph is generated to create a fingerprint of the corresponding substrate being processed. The generated graph can be referred to as a fingerprint because the graph is associated with the unique reaction of the substrate and the plasma in the chamber based on the type and degree of deformation of the substrate. In one example, the fingerprint is a graph of the percentage of reflected power (of the total output power) versus time. Other fingerprints (including heating temperature or heater output percentage versus time) are also contemplated and can be used alone or in combination with the graph of the percentage of reflected power according to various examples and combinations of examples herein. The reflected power is measured using an automatch sensor, other sensors, or impedance matching hardware equipped for such measurements.

[0032] In operation 308, the graph generated at operation 306 (e.g., a fingerprint) is analyzed. The analysis is used to determine multiple features of the fingerprint that indicate whether a bend is (1) present and (2) compressive or tensile. At operation 310, the analysis is used to generate associations between a compression designation or a tensile designation and multiple fingerprint features and combinations of features. The analysis at operation 308 is used in operation 310 to generate associations between bending features of the substrate and fingerprint features. In one example, operation 310 is a modeling operation. Each model is associated with a type of bend (compressive or tensile) and, in some embodiments, further associated with a degree of bend. In other words, operation 310 maps the bend of the substrate to a particular fingerprint. Thus, the bend of a later processed substrate can be identified based on the corresponding fingerprint of the later processed substrate.

[0033] At operation 312, the association between the substrate's bow characteristics and the fingerprint characteristics generated at operation 310 is linked to a substrate processing program. For example, the association can be determined experimentally. These linked associations are stored for later reference. The substrate processing program and the association are stored, for example, in a non-volatile memory. The association of the model with the substrate processing program facilitates correction of substrate bow. For example, each substrate processing program may include a processing recipe (and other processing details) for correcting a specific type and degree of bow based on the identified model or fingerprint. Therefore, when the substrate model and / or fingerprint is identified, the bow of the substrate can be identified, and the bow of the substrate can be mitigated by executing the corresponding substrate processing program.

[0034] At operation 314 of method 300, a plurality of inspection parameters, including coating uniformity data or substrate damage / scrap data, are received from downstream operations. Information is received at operation 314 and stored in a memory. In embodiments that may be combined with other embodiments, information from downstream operations received at operation 314 may be used to modify the associations and / or processing routines generated at operation 310 to improve processing. In other examples that may be combined with other examples herein, information from downstream operations received at operation 314 may be used to modify the associations generated at operation 312 with bow type and / or bow extent and substrate processing. Thus, the degree and extent of bow correction may be continuously refined to improve process performance.

[0035] Each of the plurality of substrate processing procedures discussed herein includes a process recipe or instructions for a voltage to be supplied to a substrate support pedestal or other substrate support member. The selected process procedure may also include other parameters, such as gas or gas mixture composition, gas flow rate, process time, temperature, and process chamber pressure, to facilitate processing of the substrate while performing bow correction. In some examples, the substrate processing procedure further includes parameters associated with transferring the substrate from a first process chamber to a second process chamber.

[0036] Figure 4 is a graph of percentage of first process chamber output power versus time for a plurality of substrates having both tensile bow and compressive bow. Figure 4 A plurality of substrate data for both compression bending and tension bending is shown. As discussed herein, the characteristics of the curves are analyzed to generate a fingerprint associated with compression bending and a fingerprint associated with tension bending. The characteristics of the curves can be associated with the type and degree of bending. Figure 4 Each fingerprint is shown and analyzed at operation 308 to obtain a plurality of features including the area under the curve, slope measured at various time periods, peak value, slope change, and other features. Such analysis facilitates determination and correction of substrate bow.

[0037] Figure 5 is a graph of peak low frequency power reflected for a number of substrates having approximately 3 microns of film deposited on the surface of the substrate. Figure 5 The peak power axis shows how much power is reflected from the plasma in the first process chamber. The more power lost during plasma maintenance, the more severe the substrate bow. Figure 5 The incoming substrate bow is the bow of the substrate under the following conditions: (1) with no voltage applied to the substrate support pedestal in the process chamber ("no ESC"), (2) with 350 volts or 600 volts applied to the electrostatic chuck of the substrate support pedestal, and (3) with 350 volts applied to the electrostatic chuck of the substrate support pedestal and the plasma stabilized in the process chamber.

[0038] In various examples, the substrate bow detected using the methods and systems herein can be -200 microns to +290 microns or greater. Figure 2 219) and a plurality of sensors 210 associated with the plurality of impedance matching circuits are coupled to the first process chamber to facilitate determination of output (forward) power and peak low frequency reflected power, respectively, as Figure 4 and Figure 5 shown.

[0039] Figure 6 is a method 600 for detecting deformation of an incoming substrate according to an embodiment of the present disclosure. At operation 602 of method 600, a substrate is received in a first process chamber, which may be similarly configured as Figure 2 The substrate received at operation 602 may have one or more layers formed thereon. In other examples, the one or more layers may include a silicon oxide layer (Si x O y ), nickel oxide (Ni x O y ), silicon nitride (Si x N y ), an oxynitride layer, or a stack of alternating oxide-nitride layers. When the substrate is positioned in a first process chamber (e.g., Figure 2 At operation 604 , a capacitive plasma is generated in the first process chamber 206 after the substrate is received at operation 602 , when the first process chamber 206 is heated to a temperature of 200° C. to 500° C. At operation 604 , a capacitive plasma is generated in the first process chamber 206 .

[0040] During operation of the first process chamber 206, at operation 606, a plurality of sensors detects reflected power and maps the reflected power to generate a fingerprint, such as Figure 5 The fingerprint is analyzed at operation 608. The analysis includes comparing multiple features of the fingerprint with one or more established fingerprint models (e.g., Figure 3 The fingerprint models are compared to those established at operation 608. Based on the analysis at operation 608, a substrate processing procedure is selected at operation 610. At operation 612, the procedure selected at operation 610 is executed in the first process chamber 206. The selected procedure results in processing the substrate (such as deposition, etching, etc.) while clamping the substrate to remove bow. The selected procedure is selected so that substrate bow is minimized or eliminated during processing, but the substrate is not over-clamped to the extent of damaging the substrate, or the substrate is not under-clamped such that undesirable bow remains.

[0041] At operation 614, after executing the selected substrate processing procedure at operation 612, various properties of the processed substrate may be evaluated. In some examples, operation 614 may alternatively or additionally occur after subsequent downstream operations such as film deposition, patterning, or cleaning operations. At operation 616, various information based on the evaluation at operation 614 (including the substrate flatness analysis) may be stored for further use and analysis, and / or used to update existing fingerprint models, processing procedures, or other information.

[0042] In one example, after selecting a substrate processing procedure at operation 610, the substrate is transferred to a second process chamber. In such an example, operation 612 may occur in the second process chamber. The second process chamber may be similar to Figure 2 The second process chamber 208 is in the process chamber 208.

[0043] In the attached example, when Figure 6 When a compressed substrate is detected as discussed in method 600 of FIG. , various transfer operating parameters may be employed. Due to the curved nature of the substrate, the substrate may be unstable during transfer due to the curved bottom portion of the substrate causing sway or displacement. The transfer operating parameters are selected to reduce sway or displacement of the substrate. The transfer operating parameters may include transfer speed, grip pressure, and / or grip position of one or more arms of the transfer robot.

[0044] In one example, a first substrate having compressive bow is analyzed and associated with a first substrate processing procedure. The first substrate processing procedure corresponds to a first process recipe having a first voltage, a first duration, a first pressure, a first gas or gas mixture composition, and a first temperature to promote bow reduction. In one example, the first substrate processing procedure includes heating the substrate to a temperature of 200°C to 500°C. The first voltage is 500V to 1000V and may be applied for 5 seconds to 3 minutes. The first voltage may be applied at a pressure of 1 to 20 Torr in a plasma formed from He and Ar. In another example, a second substrate having tensile bow is associated with a substrate processing procedure. The second substrate processing procedure corresponds to a second process recipe having a second voltage, a second duration, a second pressure, and a second gas or gas mixture, the second process recipe being used when heating the substrate to promote bow reduction. The second substrate processing procedure instructs the process chamber to heat the substrate to a temperature of 200°C to 500°C. A second voltage of 150V to 500V is applied to the substrate support for 5 seconds to 3 minutes. The second voltage may be applied at a chamber pressure of 1 Torr to 20 Torr without generating plasma inside the process chamber to correct the substrate bow.

[0045] Thus, using the systems and methods discussed herein, incoming substrates can be analyzed to detect the presence, type, and extent of bow in the substrate. Once bow is detected, a substrate processing procedure can be selected based on the type and extent of bow and executed to mitigate the bow. Detecting and correcting bowed substrates based on the type and extent of bow improves the quality of layers formed and / or patterned downstream, reduces scrap, and improves device quality, compared to conventional methods that apply the same correction method to some or all substrates regardless of substrate condition.

[0046] To facilitate understanding, identical reference numerals have been used, where possible, to designate identical elements that are common to the figures. It is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without further recitation.

[0047] While the foregoing is directed to embodiments of the present disclosure, other and further embodiments of the disclosure may be devised without departing from the basic scope thereof, and the scope of the disclosure is to be determined by the claims that follow.

Claims

1. A method for substrate processing, comprising: generating a plasma in a first process chamber while positioning a substrate in the first process chamber; generating a fingerprint of the substrate based on a plurality of sensors in the first process chamber, wherein the fingerprint is defined as a profile of power used to generate the plasma in the first process chamber, and wherein the fingerprint is generated by monitoring reflected power from the substrate; comparing the fingerprint to a plurality of stored fingerprint models to determine whether the substrate is deformed; as well as Based on determining that the substrate is deformed, a substrate processing procedure for the substrate is selected to correct for the substrate deformation.

2. The method of claim 1, further comprising: executing a selected substrate processing procedure in the first process chamber; as well as In response to executing the selected substrate processing procedure, the substrate deformation is corrected.

3. The method of claim 2, further comprising: After correcting the deformation of the substrate, transferring the substrate to a second process chamber, the second process chamber being coupled to the first process chamber via a transfer chamber; as well as A second operation is performed in the second process chamber.

4. The method of claim 1, further comprising: transferring the substrate to a second process chamber; executing a selected substrate processing procedure in the second process chamber; as well as In response to executing the selected substrate processing procedure, the substrate deformation is corrected. The method of claim 1 , wherein the plurality of sensors are configured to detect low-frequency reflected power or high-frequency reflected power.

6. The method of claim 1, wherein the substrate deformation is one of compressive bending or tensile bending, and each fingerprint model is associated with a compressive bending or a tensile bending.

7. A method for substrate processing, comprising: generating a fingerprint of a substrate positioned in a first process chamber, wherein the fingerprint is defined as a profile of power used to generate a plasma in the first process chamber, and wherein the fingerprint is generated by monitoring reflected power from the substrate; comparing the fingerprint to a plurality of stored fingerprint models, wherein each fingerprint model is associated with a type of curvature of the substrate; as well as A substrate processing procedure is selected for the substrate based on the comparison to correct for substrate bow.

8. The method of claim 7, further comprising: The fingerprint is generated based on a plurality of sensors in the first process chamber, the plurality of sensors being configured to detect low frequency reflected power or high frequency reflected power to determine the substrate bow.

9. The method of claim 7, further comprising: executing a selected substrate processing procedure in the first process chamber; as well as In response to executing the selected substrate processing procedure, the substrate bow is corrected.

10. The method of claim 7, further comprising: transferring the substrate to a second process chamber; placing the substrate on a substrate support in the second process chamber; executing a selected substrate processing procedure in the second process chamber; as well as In response to executing the selected substrate processing procedure, the substrate bow is corrected.

11. The method of claim 10, further comprising: After correcting the bow of the substrate, a second operation is performed in the second process chamber, wherein the second operation includes deposition, etching, or cleaning.

12. A computer-readable storage medium configured to execute instructions to cause a system to: generating a plasma in a first process chamber in which a substrate is positioned; generating a fingerprint of the substrate based on a plurality of sensors in the first process chamber, the plurality of sensors being configured to detect low-frequency reflected power or high-frequency reflected power, wherein the fingerprint is defined as a profile of power used to generate the plasma in the first process chamber, and wherein the fingerprint is generated by monitoring the reflected power of the substrate; comparing the fingerprint to a plurality of stored fingerprint models to determine whether the substrate is deformed; as well as Based on determining that the substrate is deformed, a substrate processing program is selected for the substrate to correct for the substrate deformation.

13. The computer-readable storage medium of claim 12, further comprising the system being configured to: performing a selected substrate processing procedure in the first process chamber; and In response to executing the selected substrate processing procedure, the substrate deformation is corrected.

14. The computer-readable storage medium of claim 12, further comprising the system being configured to: transferring the substrate to a second process chamber; placing the substrate on a substrate support in the second process chamber; performing a selected substrate processing procedure in the second process chamber; and In response to executing the selected substrate processing procedure, the substrate deformation is corrected.

15. The computer-readable storage medium of claim 12, wherein the substrate deformation is one of compressive bending or tensile bending, wherein each fingerprint model is associated with a compressive bending or a tensile bending.

Images