Plasma supermodel integrated with feature scaled profile model for accelerated etch process development

Abstract

Plasma parameters at the surface of the wafer are determined based on plasma processing conditions using a plasma supermodel. A feature scaled profile model may be utilized to predict a post-processing profile of the surface of the wafer. Correlations in the plasma hypermodel may be recalibrated if the post-processing profile is outside of an experimentally referenced convergence criterion.

Description

[0001] CROSS REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to provisional patent application filed on October 10, 2022 and assigned as U.S. application No. 63 / 414,603, the disclosure of which is hereby incorporated by reference. Technical Field

[0003] The present disclosure relates to semiconductor manufacturing. Background Art

[0004] Fabrication of semiconductor devices such as logic and memory devices typically includes processing semiconductor wafers using a number of fabrication processes to form various features and multiple levels of semiconductor devices. For example, photolithography is a semiconductor fabrication process that involves transferring a pattern from a mask to a photoresist disposed on a semiconductor wafer. Additional examples of semiconductor fabrication processes include, but are not limited to, chemical mechanical polishing (CMP), etching, deposition, and ion implantation. An arrangement of multiple semiconductor devices fabricated on a single semiconductor wafer may be separated into individual semiconductor devices.

[0005] Many of these fabrication processes use plasma. For example, plasma etching, ion implantation, and plasma enhanced chemical vapor deposition (PECVD) all generate plasma as part of their operation. The operational settings for these processes can be based on simulations of plasma parameters.

[0006] In the previous method, a rigorous physics-based plasma model is used to simulate the plasma in the etching reactor to obtain the plasma parameters at the wafer surface, such as free radical flux and ion energy. Then, these plasma parameters are merged into the feature scaling profile model to predict the resulting etching profile. This previous method uses two separate models (a rigorous physics-based plasma model and a feature scaling profile model), which are executed in series. Each model is calibrated separately, and the data is manually transferred from the plasma model to the profile model. In addition to this manual transmission, there is almost no communication or interaction between the rigorous physics-based plasma model and the feature scaling profile model. The surface chemistry in the feature scaling profile model and the plasma chemistry in the rigorous physics-based plasma model cannot be jointly optimized. It is also impossible to implement a numerical algorithm to calibrate the two separate models. Therefore, the iterative method of looping back and forth between the two models to perform calibration is time-consuming and inefficient.

[0007] Therefore, new technologies and systems are needed. Summary of the invention

[0008] In a first embodiment, a method is provided. The method includes receiving plasma processing conditions at a processor and using the processor to determine plasma parameters at a surface of a wafer based on the plasma processing conditions using a plasma hypermodel. The plasma processing conditions may include one or more of: pressure, gas chemistry, temperature, flow rate, source power, bias power, source power, or pulse conditions.

[0009] The method may further include inputting the plasma parameters into a feature scaling profile model, and using the processor to predict a post-processing profile of the surface of the wafer using the feature scaling profile model. The processor may be used to compare the post-processing profile to an experimental reference. If the post-processing profile is outside of a convergence criterion of the experimental reference, the processor may be used to recalibrate correlations in the plasma supermodel. The experimental reference may be a TEM image or an XSEM image.

[0010] The plasma hypermodel may include matrices that are multiplied by the plasma processing conditions.

[0011] The plasma conditions may be used in an etching process, a deposition process, or an ion implantation process.

[0012] The method may further include exposing the wafer to plasma using the plasma processing conditions in a plasma processing tool.

[0013] A non-transitory computer-readable medium storing a program may be configured to instruct a processor to execute the method of the first embodiment.

[0014] In a second embodiment, a system is provided. The system includes a plasma processing tool having: a plasma chamber; a stage disposed in the plasma chamber and configured to hold a wafer; and a plasma generation system. The plasma processing tool can be an etching tool, a deposition tool, or an ion implantation tool. A processor is in electronic communication with the plasma processing tool. The processor is configured to receive plasma processing conditions and determine, based on the plasma processing conditions, a plasma parameter at a surface of the wafer using a plasma hypermodel. The plasma processing conditions can include one or more of: pressure, gas chemistry, temperature, flow rate, source power, bias power, source power, or pulse conditions.

[0015] The processor may be further configured to input the plasma parameters into a feature scaled profile model and utilize the feature scaled profile model to predict a post-processing profile of the surface of the wafer. The processor may be configured to compare the post-processing profile to an experimental reference. If the post-processing profile is outside of a convergence criterion of the experimental reference, the processor may also be configured to recalibrate correlations in the plasma supermodel. The experimental reference may be a TEM image or an XSEM image.

[0016] The plasma hypermodel may include matrices that are multiplied by the plasma processing conditions.

[0017] The processor may be configured to send instructions to the plasma processing tool to generate plasma utilizing the plasma processing conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] For a more complete understanding of the nature and purpose of the present disclosure, reference should be made to the following detailed description taken in conjunction with the accompanying drawings, in which:

[0019] Figure 1 is a flow chart of an embodiment of a calibration process based on integration of a plasma supermodel and a feature scaled profile model according to the present disclosure;

[0020] Figure 2 is a block diagram of an exemplary system according to the present disclosure; and

[0021] Figure 3 Contour lines showing the correlation between process conditions (power and pressure) and plasma parameters at the wafer surface in an Ar / Cl 2 plasma supermodel, calibrated based on XSEM images from different process conditions. DETAILED DESCRIPTION

[0022] Although the claimed subject matter will be described in terms of specific embodiments, other embodiments, including embodiments that do not provide all of the benefits and features listed herein, are also within the scope of the present disclosure. Various structural, logical, process step, and electronic changes may be made without departing from the scope of the present disclosure. Therefore, the scope of the present disclosure is defined solely by reference to the appended claims.

[0023] In embodiments disclosed herein, a plasma super model is integrated with a feature scaling profile model for process development. Based on process conditions, the plasma super model generates plasma parameters at the wafer surface, which are then used as inputs to the feature scaling profile model to predict feature profiles. Thus, the plasma super model and the feature scaling profile model can be jointly optimized. The integration of the plasma super model and the feature scaling profile model can enable model calibration and process optimization from macro process conditions (e.g., power and pressure) to micro feature metrics (e.g., critical dimensions and sidewall angles).

[0024] The plasma supermodel has tunable parameters rather than several large fixed parameter sets in a rigorous physics-based plasma model. Where the plasma supermodel is integrated with the feature scaling profile model, metrological references can be used to calibrate both the plasma supermodel and the feature scaling profile model through numerical algorithms based on experimental references such as transmission electron microscope (TEM) and cross-sectional scanning electron microscope (XSEM) images. TEM images and XSEM images are more easily accessed during plasma reactor operation than plasma parameters at the surface. The plasma supermodel and the feature scaling profile model, once calibrated, can be used to predict profiles for other processing conditions. The plasma supermodel can capture both linear and nonlinear relationships between reactor knobs (e.g., power and pressure) and plasma parameters (e.g., energy and flux) at the wafer surface in different forms (e.g., first order, second order, or cross terms).

[0025] Process development of computational models depends on the accuracy and efficiency of the models. High-fidelity and efficient models that take as input the process conditions tuned by knobs on the process reactor and as output the predicted feature profiles can improve high-volume manufacturing.

[0026] In previous approaches to achieve high fidelity and efficient models, both reactor scaled modeling and feature scaled modeling are used in tandem. For reactor scaled plasma modeling, which is used to model the plasma in an etch reactor using a rigorous physics-based approach, process conditions (e.g., power and pressure) tuned by knobs on the reactor are inputs. Physics equations for transport, dynamics, and electromagnetics, as well as chemical reactions for vapor phase reactions and surface reactions are solved. The outputs are plasma parameters (e.g., flux and energy) at the wafer surface. For feature scaled profile modeling, which is used to model the evolution of feature profiles using a rigorous physics-based approach, the outputs of reactor scaled modeling, which are plasma parameters, are used as inputs. The transport and surface reactions of plasma particles are modeled, with the resulting etch feature profiles as outputs.

[0027] As disclosed herein, a plasma supermodel, rather than reactor-scaled plasma modeling, is used to correlate process conditions with plasma parameters at the wafer surface. The correlation is based on empirical assumptions. Its coefficients can be calibrated based on a cost function defined by the difference between the modeling results and the experimental reference. The plasma supermodel is integrated with a rigorous physics-based profile model for model calibration. An iterative regression method for calibration is presented in Figure 1 In the method 100 of . With the process conditions as input, the plasma supermodel generates plasma parameters at the wafer surface based on the correlations contained therein. Those plasma parameters can be used as input to a subsequent feature scaling model to predict a post-processing profile (e.g., a post-etch profile). A cost function based on the difference between the model-predicted profile and an experimental reference (e.g., a TEM or XSEM image) can be evaluated. If the evaluation results meet a preset convergence criterion, then the model calibration is complete and can be used for process optimization. If the evaluation results do not meet the convergence criterion, then a numerical algorithm can be used to recalibrate both the correlations in the plasma supermodel and the transport and surface reactions in the feature scaling profile model. This process can be repeated until the feature profile predicted by the plasma supermodel and the feature scaling profile model meets the convergence criterion.

[0028] A hypermodel involves a correlation or mapping between two data sets (i.e., input data and output data) or multiple data sets. The input data may be a control knob of a physical system (e.g., a plasma processing tool like a plasma etch tool), and the output data is the parameter that the engineer wants to control with the control knob (e.g., plasma flux, ion energy, temperature). A hypermodel may be a matrix or several matrices multiplied together.

[0029] Compared to previous methods that may only allow process engineers to iterate on a weekly basis, the embodiments disclosed herein may allow process engineers to iterate on their plasma recipes more frequently. For example, the embodiments disclosed herein may be used to iterate on a daily basis. Previous methods required both rigorous physics-based plasma modeling and rigorous physics-based etch modeling, which typically took several days to complete. Process engineers would manually pass the output of plasma modeling as input to etch modeling. The embodiments disclosed herein replace rigorous physics-based plasma modeling with a plasma hypermodel, which may take less than a second to complete. Figure 1 The process presented in automates the calibration process for both the rigorous physics-based modeling in the plasma hypermodel and the feature-scaled profile model, so efficiency can be improved and process engineers can iterate on the recipe for optimal profile control on a daily basis rather than a weekly basis.

[0030] In the plasma hypermodel, a matrix operator A is assumed to relate the process conditions x to the plasma parameters y at the wafer surface under an adjustable bias b in the form y = Ax + b. For example, these correlations can be filled in the following form.

[0031]

[0032] The above equation shows the first-order form of vector x, which may also have the following cross-term form.

[0033]

[0034] The vector x can also have the following second-order form.

[0035]

[0036] The vector x may also have a combination of first order, second order, and cross terms as shown below.

[0037]

[0038] The coefficients in matrix A and vector b can be tuned to obtain plasma parameters that produce a simulated etch profile that matches the experimental reference. After matrix A and vector b in the plasma supermodel and the transport and surface reactions in the feature scaled profile model are calibrated, these can be used to predict profiles for other process conditions. The calibrated plasma supermodel and feature scaled profile model can be used in reverse to search for optimal process conditions based on the desired feature profile.

[0039] The supermodel relates input parameters, which are usually different physical parameters (power, pressure, frequency, temperature, etc.) with different physical units and different orders of magnitude and value ranges. For example, the power can be from 1 W to 10000 W and the frequency can be from 1 kHz to 100 MHz. The output parameters are different in units, magnitudes (for example, the plasma flux can be from 1015 cm - 2 s -1 To 1018cm -2 s -1 , ion energies can be from 10 eV to 10000 eV). Given those differences in units, magnitudes, and ranges for both input and output parameters, it can be difficult to efficiently perform calculations or calibrate the hypermodel without using a processor.

[0040] like Figure 1, at 101, plasma processing conditions are received at a processor. For example, the plasma processing conditions may include one or more of: pressure, gas chemistry, temperature, flow rate, source power, bias power, source power, pulse conditions (e.g., pulse frequency or pulse duty cycle), or other parameters. At 102, the processor may determine an output based on the plasma processing conditions using a plasma hypermodel. In an embodiment, a plasma hypermodel is used to determine plasma parameters at a surface of a wafer. The plasma hypermodel may include a matrix multiplied by the plasma processing conditions.

[0041] For example, the plasma parameters determined by the plasma supermodel may include one or more of the following: the flux of neutral radicals, the flux of ions, the energy of ions, the angle of ions, or other parameters. For a particular plasma processing tool, such as a plasma etching tool with multiple power control knobs, the plasma may give a bimodal or multimodal distribution of ions. For these complex conditions, the supermodel may also be used to determine the low energy peak of ions, the high energy peak of ions, the flux of the low energy peak of ions, and / or the flux of the high energy peak of ions.

[0042] At 103, the plasma parameters are input into a feature scaling profile model. The feature scaling profile model may be a PROETCH physics-based dry etch simulator from KLA Corporation or other model. A processor may run the feature scaling profile model. At 103, the feature scaling profile model may predict a post-processing profile of the surface of the wafer based on the plasma parameters generated by the plasma hypermodel and based on the pre-processing profile, which is shown at 105. Although disclosed with etching, the method 100 may be applied to other plasma processing techniques, such as deposition or ion implantation.

[0043] The feature scaling profile model may be used for research or development in semiconductor manufacturing. The feature scaling profile model may be used to develop new technologies, new processes, or new process flows for new structures, new materials, or new technology nodes.

[0044] The feature scaling profile model can model the evolution of features on a semiconductor wafer exposed to energetic and reactive species generated in a plasma. In an example, the feature scaling profile model can input hard mask profiles and underlying materials from a scanning electron microscope (SEM) image or design file. The feature scaling profile model can also input photoresist profiles and underlying film stacks from a lithography simulator (e.g., PROLITH developed by KLA Corporation). The incoming geometry is discretized into a collection of voxels in 3D space, where each voxel is labeled by a material identification. The evolution of this voxel-based grid over time due to interaction with the plasma is tracked to predict post-processing profiles. Other methods in the feature scaling profile model are possible and this is just one example.

[0045] like Figure 1 As shown in , at 105, the processor may compare the post-processing profile to the experimental reference. If the post-processing profile is outside the convergence criteria of the experimental reference, then a recalibration process may be initiated. The correlation of the plasma supermodel and / or the feature scaled profile model may be adjusted. If the post-processing profile is within or equal to the convergence criteria, then calibration is complete, as shown at 106.

[0046] The convergence criterion may be that the cost is less than a threshold set by the user. The cost is calculated based on the difference between the feature scaling profile model simulation result and the experimental reference. For example, for the bend critical dimension (CD), the experimental reference is 20nm. The feature scaling profile model gives 25nm. The cost is calculated as a function of the difference (i.e., 25nm-20nm). As part of the function, this difference may be multiplied by one or more factors, such as standard deviation or random noise.

[0047] For example, the adjusted parameters in the hypermodel may include linear coefficients and power coefficients for assumed correlations between input and output parameters. For example, the adjusted parameters in the feature scaling profile model may include surface reaction parameters (e.g., adhesion coefficient, sputtering yield, threshold energy, etc.) or plasma operating conditions (e.g., power, pressure, frequency, temperature, process time, process steps, etching recipes, etc.).

[0048] For example, the experimental reference may be a TEM or an XSEM. In addition to those TEM and XSEM images, other references may be, for example, critical dimensions at different heights, unique sizes (e.g., neck CD, bend CD, bottom CD, etch depth, mask remaining thickness, etc.), ideal profiles or target profiles (e.g., straight profiles or anisotropic profiles).

[0049] If the post-processing profile is within or equal to the convergence criteria of the experimental reference, then calibration may be complete.The processor may send instructions to expose the wafer to a plasma utilizing plasma processing conditions in a plasma processing tool.

[0050] It should be understood that although the Figure 1 The method may be implemented in software, firmware, hardware, or a combination thereof. In one mode, the method is implemented in software as an executable program and is executed by one or more special or general-purpose digital computers, such as personal computers (e.g., IBM compatible, Apple compatible, workstation, or mainframe computers). The steps of the method may be implemented by a server or computer in which the software module resides or partially resides.

[0051] Generally speaking, with respect to hardware architecture, as will be well understood by those skilled in the art, such a computer will include a processor, a memory, and one or more input and / or output (I / O) devices (or peripherals) communicatively coupled via a local interface. For example, as is well known in the art, a local interface may be, but is not limited to, one or more buses or other wired or wireless connections. The local interface may have additional elements such as controllers, buffers (caches), drivers, repeaters, and receivers to enable communication. Further, the local interface may include address, control, and / or data connections to enable appropriate communication among the other computer components.

[0052] The processor can be programmed to perform Figure 1 The processor is a hardware device for executing software, particularly software stored in memory. The processor can be any custom or commercially available processor, a main processing unit (CPU or GPU), an auxiliary processor among several processors associated with a computer, a semiconductor-based microprocessor (in the form of a microchip or chipset), a macroprocessor, or substantially any device for executing software instructions.

[0053] The memory is associated with the processor and may include any one or combination of volatile memory elements, such as random access memory (RAM, such as DRAM, SRAM, SDRAM, etc.) and non-volatile memory elements, such as ROM, hard drive, tape, CDROM, etc. In addition, the memory may incorporate electronic, magnetic, optical and / or other types of storage media. The memory may have a distributed architecture in which various components are located remote from each other but still accessed by the processor.

[0054] The software in the memory may include one or more separate programs. The separate program includes an ordered list of executable instructions for implementing the logical functions in order to implement the functions of the module. In the examples described so far, the software in the memory includes one or more components of the method and can be executed on a suitable operating system (O / S).

[0055] The present disclosure may include components provided as source executable programs (object code), instruction codes, or any other entity including sets of instructions to be executed. When a source program, the program needs to be translated via a compiler, assembler, interpreter, etc., which may or may not be contained in a memory, so as to operate properly in conjunction with an O / S. In addition, methods implemented according to the teachings may be expressed as (a) an object-oriented programming language, which has various types of data and methods, or (b) a procedural programming language, which has routines, subroutines, and / or functions, such as, but not limited to, C, C++, Pascal, Basic, Fortran, Cobol, Ped, Java, and Ada.

[0056] Additional embodiments relate to a non-transitory computer-readable medium storing program instructions executable on a processor for performing a computer-implemented method for calibrating plasma processing conditions, as disclosed herein. In particular, an electronic data storage unit or other storage medium may contain a non-transitory computer-readable medium including program instructions executable on a processor. The computer-implemented method may include any of the steps of any of the methods described herein.

[0057] Each of the steps of the method may be performed as described herein. The method may also include any other steps that may be performed by a processor and / or computer subsystem or system described herein. The steps may be performed by one or more computer systems, which may be configured according to any of the embodiments described herein. In addition, the method described above may be performed by any of the system embodiments described herein.

[0058] Figure 2 2 is a block diagram of a system 200. The system 200 includes a plasma processing tool 206 having a plasma chamber 201, a stage 203 holding a wafer 202 in the plasma chamber 201, and a plasma generation system 204. The plasma generation system 204 may be an RF plasma generation system, an inductively coupled plasma (ICP) generation system, or another type of ion source. The plasma processing tool 206 may be an etching tool, a deposition tool, an ion implantation tool (e.g., a plasma doping tool), or another plasma processing tool. Ion implantation tools having a beamline with beam optics may also benefit from the embodiments disclosed herein.

[0059] The processor 205 is in electronic communication with other components of the plasma generation system 204 or plasma processing tool 206. For ease of illustration, the processor 205 is illustrated as a single functional block, but in practice the processor 205 may include multiple interconnected processors with suitable interfaces for receiving and outputting the signals illustrated in the figure and described in the text.

[0060] The processor 205 can be part of the plasma processing tool 206 or separate from the plasma processing tool 206. For example, the processor 205 can be on a system or other server within a semiconductor manufacturing facility. The processor 205 can be used offline relative to the plasma processing tool 206.

[0061] The processor 205 may be configured to execute Figure 1 Thus, the processor 205 may receive the plasma processing conditions and utilize the plasma hypermodel to determine the plasma parameters at the surface of the wafer 202. The processor 205 may also be configured to input the plasma parameters into a feature scaling profile model and utilize the feature scaling profile model to predict a post-processing profile of the surface of the wafer 202.

[0062] The processor 205 may be configured to compare the post-processing profile to an experimental reference (eg, a TEM image or an XSEM image). If the post-processing profile is outside of the convergence criteria of the experimental reference, the processor 205 may be configured to recalibrate correlations in the plasma supermodel.

[0063] If the post-processing profile is within or equal to the convergence criteria, the processor 205 may optionally be configured to send instructions to the plasma processing tool 206 to generate plasma within the plasma chamber 201 using plasma processing conditions.

[0064] Although described in conjunction with a plasma processing tool, the embodiments disclosed herein may operate on or with a metrology tool, an inspection tool, or an inspection tool. Such tools may include light-based, ion beam-based, x-ray-based, or electron beam-based systems. The ion beam-based imaging subsystem may be a focused ion beam (FIB) system, a helium ion microscope (HIM) system, or a secondary ion mass spectrometry (SIMS) system. Such a system may be used to generate an image of a test wafer, such as an image with a pre-processing profile or a reference for comparison with a post-processing profile (e.g., a pre-etch profile versus a post-etch profile, a pre-deposition profile versus a post-deposition profile, or a pre-implantation profile versus a post-implantation profile).

[0065] The following examples are presented to illustrate the present disclosure. The examples are not intended to be limiting.

[0066] Example 1

[0067] For silicon etching processes using Ar / Cl2 plasma, process conditions include pressure, power, pulse frequency, and duty cycle. The plasma parameters at the wafer surface that directly affect the etching profile include at least flux, energy, and angle. The empirical correlation between plasma parameters and process conditions is assumed as follows.

[0068] Ion flux: flux 离子 =a×pressure+b

[0069] Fraction of high energy ion flux: Fraction 高能量离子 =c×power+d

[0070] Fraction of low energy ion flux: Fraction 低能量离子 =1-score 高能量离子

[0071] Radical Flux: Flux Cl =e×pressure+f

[0072] High ion energy: energy 高能量离子 =g×power+h×pressure+i

[0073] Low ion energy: Energy 高能量离子 =j×power+k×pressure+l

[0074] The above equation can be written in the following matrix format.

[0075]

[0076] The coefficients in matrix A and vector b can be tuned to match experimental references. Figure 3 The contours of the correlations based on the plasma supermodel are shown in . Figure 1 The iterative approach presented in is based on XSEM images from different processing conditions, and these correlations in the Ar / Cl2 plasma supermodel are calibrated together with the silicon etching mechanism developed in the feature scaling profile model (PROETCH from KLA).

[0077] Example 2

[0078] Self-aligned double patterning (SADP) is characterized by performing photolithography at a pitch twice larger than the desired pitch followed by spacer deposition, spacer etch back and core etch processes. Double and multiple patterning techniques reduce technology nodes, where the challenge is the non-uniformity of the spacers.

[0079] An example of a feature scaling profile model is PROETCH, developed by KLA. PROETCH is a three-dimensional, physics-based plasma dry etch simulator that describes the processes that occur during a plasma etch process, including energetic and thermal species propagation through features, physical and chemical sputtering of materials, thermal etching, and surface passivation. The input may include an etch prescription and incoming geometry. The output may include a 3D etch profile, feature-level local flux distribution, and surface composition. PROETCH may be integrated with PROLITH, a lithography simulator developed by KLA, for process control.

[0080] Silicon can be etched by Ar / Cl2 plasma. The surface sites for SADP can include hard mask (HM), Si, SiCl (lumped passivation surface states) and SiO2 (stop layer). The reaction mechanism can include the following.

[0081]

[0082]

[0083] PROETCH is used to study overlay metrology targets in DRAMs during shallow trench isolation (STI) etching using SADP for shallow trench isolation masks. Each process introduces non-uniformity, leading to variations in, for example, the hard mask used in the final etch step to form the gate. The optimization of the etch profile and etch process for specific plasma operating conditions is studied. During etching of STI, mask sputtering, bowing, front taper, and aspect ratio dependent etching can occur.

[0084] In the study, a sensitivity analysis of the model was performed using a parametric study of the mechanism parameters. Physical sputtering contributes primarily to material removal at the bottom and the etching front. Chemical sputtering affects both lateral etching at the sidewalls and vertical etching at the bottom. Chlorination favors vertical etching because the exposed sites are converted to SiCl with a lower threshold for removal. Parametric studies of plasma parameters (e.g., flux and energy) show the combined effects of multiple reactions involving specific plasma parameters. Chlorine radicals contribute to chlorination and thermal etching (enhanced undercutting). Ions contribute to vertical etching and straight profiles. Increasing ion energy causes more ions to exceed the thresholds of both chemical sputtering and physical sputtering. When etching byproducts are recaptured by the surface without escaping the feature, the redeposition of etching byproducts can affect direct deposition. When etching byproducts detach from the feature and are deposited by sputtering back into the feature by vapor phase species, the redeposition of etching byproducts can affect indirect redeposition after interacting with the vapor phase. The angular yield curve (the dependence of yield on the incident ion angle) depends on factors such as ion energy, material identification, surface roughness, etc. As the incident ion energy increases, the angular yield curve changes from chemical sputtering-like (peak at normal incidence) to physical sputtering-like (peak at 60 degree incidence).

[0085] Next, use Figure 1 The calibration is performed using the method shown in and described herein. The inputs are process conditions, pre-process profiles, and post-process profiles (e.g., pre-etch profiles and post-etch profiles). The reaction mechanism and plasma supermodel are calibrated. The cost metric is the difference between the model-determined profile and the experimental profile. The PROETCH algorithm is used as a feature scaling profile model.

[0086] The pre-etch profile may be a SEM image, which may show a hard mask formed by photolithography, etching and / or deposition. Although the pre-etch profile may be a real experimental profile, the pre-etch profile may also be a design image (desired profile).

[0087] The plasma supermodel is used to relate process conditions (e.g., power, pressure) to plasma parameters (e.g., flux, energy). Subset data is used for model calibration. Interpolation and extrapolation of the calibrated model is used as validation. The mixture of conditions is selected to cover a wide range of parameters in both dimensions. Both unique size and critical size are included in the cost with user-defined weights. Validation is performed for conditions other than the calibration set. Both interpolation and extrapolation (or mixing) are used as validation conditions. The validation results match the experimental data well in terms of both unique size and critical size.

[0088] Process optimization presents both forward and reverse problems. When looking forward, output performance (e.g., etch rate, profile, selectivity) can be predicted under specific process conditions (e.g., power, pressure, frequency) by using a calibrated model. This can reduce experimental trials and can shorten development cycles. During the reverse period, the search for process conditions for optimal etch performance achieves straight profiles, increases throughput, and minimizes aspect ratio dependent etching.

[0089] Using the calibrated model, extrapolation is performed in the time domain at a fixed etch depth. End point detection (EPD) is used to stop the simulation after the etch front reaches a preset depth. CD and profile can be visualized to obtain optimal process conditions.

[0090] The reversibility of the model can be demonstrated if the process conditions given by the inverse problem converge to the vicinity of the design of experiments (DOE) process conducted to produce the target profile. A straight profile (uniform CD) is used as the target / cost for optimizing the process conditions. Utilizing different time tolerances, the optimization converges to different conditions. Unique sizes (bend CD and bottom CD) are used as metrics to optimize the process recipe for the straight profile. The tolerance window is set to allow the profile to vary within the DOE tolerance range. PROETCH optimizes the process conditions that close the gap of the bottom CD while maintaining the bend CD within the tolerance range.

[0091] Therefore, a rigorous physics-based etching model (PROETCH) was developed and validated using STI etching in SADP process as a use case. The reaction mechanism was developed based on XSEM images of DOE conditions in Si etching by Ar / Cl2 plasma. A plasma supermodel based on empirical practice was implemented to relate process conditions (e.g., power, pressure) to plasma parameters (e.g., flux, energy). PROETCH was used for process development to provide etching profiles for conditions interpolated or extrapolated for DOE ranges, and to search for process conditions that can meet user-set criteria for optimal etching performance.

[0092] Although the present disclosure has been described with respect to one or more specific embodiments, it should be understood that other embodiments of the present disclosure may be made without departing from the scope of the present disclosure. Therefore, the present disclosure is considered to be limited only by the appended claims and a fair interpretation thereof.

Claims

1. A method comprising: receiving plasma processing conditions at a processor; and The processor is used to determine plasma parameters at a surface of a wafer based on the plasma processing conditions using a plasma hypermodel.

2. The method of claim 1, wherein the plasma processing conditions include one or more of: pressure, gas chemistry, temperature, flow rate, source power, bias power, source power, or pulse conditions.

3. The method according to claim 1, further comprising: inputting the plasma parameters into a feature scaled profile model; and The processor is used to predict a post-processing profile of the surface of the wafer using the feature scaling profile model. The method of claim 3 , further comprising using the processor to compare the post-processing profile to an experimental reference.

5. The method of claim 4, further comprising using the processor to recalibrate correlations in the plasma supermodel, wherein the post-processing profile is outside of a convergence criterion of the experimental reference. The method according to claim 4 , wherein the experimental reference is a TEM image or an XSEM image.

7. The method of claim 1, wherein the plasma hypermodel comprises a matrix multiplied by the plasma processing conditions.

8. The method of claim 1, wherein the plasma condition is used in an etching process, a deposition process, or an ion implantation process.

9. The method of claim 1, further comprising exposing the wafer to plasma utilizing the plasma processing conditions in a plasma processing tool.

10. A non-transitory computer-readable medium storing a program configured to instruct a processor to execute the method according to claim 1.

11. A system comprising: A plasma processing tool comprising: Plasma chamber; a carrier disposed in the plasma chamber and configured to hold a wafer; and a plasma generating system; and a processor in electronic communication with the plasma processing tool, wherein the processor is configured to: receiving plasma processing conditions; and Plasma parameters at the surface of the wafer are determined based on the plasma processing conditions using a plasma hypermodel.

12. The system of claim 11, wherein the plasma processing tool is an etching tool, a deposition tool, or an ion implantation tool.

13. The system of claim 11, wherein the plasma processing conditions include one or more of: pressure, gas chemistry, temperature, flow rate, source power, bias power, source power, or pulse conditions.

14. The system of claim 11, wherein the processor is further configured to: inputting the plasma parameters into a feature scaled profile model; and A post-processing profile of the surface of the wafer is predicted using the feature scaling profile model.

15. The system of claim 14, wherein the processor is further configured to compare the post-processing profile to an experimental reference.

16. The system of claim 15, wherein the processor is further configured to recalibrate correlations in the plasma hypermodel, wherein the post-processing profile is outside of a convergence criterion of the experimental reference.

17. The system of claim 15, wherein the experimental reference is a TEM image or an XSEM image.

18. The system of claim 11, wherein the plasma hypermodel comprises a matrix multiplied by the plasma processing conditions.

19. The system of claim 11, wherein the processor is further configured to send instructions to the plasma processing tool to generate plasma utilizing the plasma processing conditions.

Images