TW201526130A - Integrated use of model-based metrology and a process model - Google Patents

Abstract

Methods and systems for performing measurements based on a measurement model integrating a metrology-based target model with a process-based target model. Systems employing integrated measurement models may be used to measure structural and material characteristics of one or more targets and may also be used to measure process parameter values. A process-based target model may be integrated with a metrology-based target model in a number of different ways. In some examples, constraints on ranges of values of metrology model parameters are determined based on the process-based target model. In some other examples, the integrated measurement model includes the metrology-based target model constrained by the process-based target model. In some other examples, one or more metrology model parameters are expressed in terms of other metrology model parameters based on the process model. In some other examples, process parameters are substituted into the metrology model.

Description

Model-based measurement and integration of a process model Cross-reference to related applications

The present patent application claims priority to U.S. Provisional Patent Application Serial No. 61/738,760, filed on Dec. The subject matter of this patent is incorporated herein by reference in its entirety.

The described embodiments relate to metrology systems and methods, and more particularly to methods and systems for improved parametric measurements.

Semiconductor devices such as logic and memory devices are typically fabricated by applying a sequence processing step to a sample. The various features of the semiconductor device and the plurality of structural layers are formed by such processing steps. For example, lithography relates in particular to a semiconductor fabrication process that produces a pattern on a semiconductor wafer. Additional examples of semiconductor fabrication processes include, but are not limited to, chemical mechanical polishing, etching, deposition, and ion implantation. A plurality of semiconductor devices can be fabricated on a single semiconductor wafer and then separated into individual semiconductor devices.

A metrology process is used in various steps during a semiconductor fabrication process to detect defects on the wafer to promote higher yields. Optical metrology technology offers the potential for high throughput without the risk of sample damage. Several optical-based techniques and associated analysis algorithms, including scatterometry and reflectance measurements, are typically used to characterize critical dimensions, membranes Thickness, composition and other parameters of the nanostructure.

Traditionally, optical metrology is performed on targets consisting of thin films and/or repetitive periodic structures. Such film and periodic structures generally represent actual device geometry and material structure or an intermediate design during device fabrication. As devices (eg, logic and memory devices) move toward smaller nanoscale sizes, characterization becomes more difficult. Devices incorporating complex three-dimensional geometries and materials with a variety of physical properties create characterization difficulties.

For example, modern memory structures are typically three-dimensional structures that make it difficult for optical radiation to penetrate to the underlying high aspect ratio. In addition, the gradual increase in the number of parameters required to characterize a complex structure (e.g., FinFET) results in a gradual increase in parameter correlation. As a result, the measurement model parameters of the characterization target are often not reliably decoupled.

In response to these challenges, more complex optical tools have been developed. The measurement is performed above (and usually simultaneously) a number of machine parameters (eg, wavelength, incident orientation and angle, etc.) over a wide range. As a result, the measurement time, the calculation time, and the total time to produce reliable results (including the measurement recipe) increased significantly. In addition, the spread of light intensity over a large wavelength range reduces the illumination intensity at any particular wavelength and increases the single non-determinism of the measurements performed at that wavelength.

Due to the increasing small resolution requirements, multi-parameter correlations, increasingly complex geometries, and the increasing use of non-transparent materials, further measurement applications present challenges for measurement. Therefore, there is a need for methods and systems for improved metrology.

The present invention presents methods and systems for integrating a process-based target model and a measurement-based target model optimization measurement model. By integrating a measurement-based target model and a process-based target model to improve the predictive outcome of one or both of the measurement model and the process model.

In one aspect, a system employing an integrated metrology model is used to measure the structural and material characteristics of one or more targets (eg, material composition, structure, and dimensional characteristics of the film, etc.). in In another aspect, a measurement system employing an integrated measurement model is used to directly measure process parameter values.

A process-based target model can be integrated with a measurement-based target model to produce an integrated measurement model in a number of different ways.

In some examples, a limit to the range of values of the measurement model parameters is determined based on the process-based target model. More specifically, the limits are determined based on the range of achievable values of the process model parameters.

In some other examples, the integrated measurement model includes a measurement-based target model that is limited by a process-based target model. This reduces the size of the solution space associated with the integrated measurement model. In this way, the process-based restriction set of the measurement-based target model parameters is defined by the process-based target model and applied to the measurement-based target model.

In some other examples, one or more measurement model parameters are represented based on other measurement model parameters based on the process model. This reduces the total number of floating parameters of the integrated measurement model and reduces the parameter correlation. This also increases the efficiency and robustness of the adaptation engine (eg, regression engine) due to the smaller search space limited to one of the target changes allowed by the manufacturing process.

In some other examples, process parameters are substituted into the metrology model. These process parameters are then parsed into part of the analysis of the measured data using an integrated measurement model. In this way, the parameterization of the integrated measurement model includes process-based variables and directly determines the process parameter values from the measurement signals.

In some examples, an integrated measurement model is used to sequentially measure process parameter values. The integrated measurement model includes one of the components used as a measurement analysis. The measurement-based target model determines the geometric parameter values from the measurement data. The integrated measurement model also includes a process-based model to determine process parameter values from geometric parameter values.

In another aspect, the process model can be improved based on the measurement model. In some examples, the information obtained using the self-measurement model improves the calibration of a process model. In an instance The pre-characteristic relationship between geometric contours and process variations can be used to process recipe generation. In addition, the process model can be calibrated in one of the optimized measurement models.

In yet another aspect, the integrated measurement model can be fully or partially assembled from a process-based target model.

In yet another aspect, the disclosed methods and systems can be employed in the context of multi-objective modeling. In some examples, the integrated metrology model allows for a combined analysis in which multiple targets are resolved using the metrology model parameters while parsing some of the targets using process parameters. In addition, restrictions derived from a process model can be used to link parameters for different targets.

In yet another aspect, a cross-wafer process variation model can be combined with a process-based model of the target structure.

In yet another aspect, an integrated metrology model can be used to provide active feedback to a processing tool (eg, lithography tools, etching tools, deposition tools, etc.). For example, the depth value and focus parameters determined using an integrated metrology model can be communicated to the lithography tool to adjust the lithography system to achieve a desired output. In a similar manner, etch parameters (eg, etch time, diffusivity, etc.) or deposition parameters (eg, time, concentration, etc.) can be included in the integrated metrology model to each provide active feedback to the etch tool or deposition tool.

The above is a summary of the present invention, and therefore, it should be understood that the invention is to be construed as illustrative and not limiting. Other aspects, features, and advantages of the devices and/or processes described herein will become apparent in the non-limiting Detailed Description.

10‧‧‧Top view

20‧‧‧ side view

30‧‧‧ side view

31‧‧‧Area

40‧‧‧Top view

41‧‧‧Area

50‧‧‧ contour map

51‧‧‧Blank area

60‧‧‧ contour map

61‧‧‧Blank area

70‧‧‧ contour map

71‧‧‧Blank area

80‧‧‧Table

90‧‧‧ schema

100‧‧‧Table

110‧‧‧Table

120‧‧‧ schema

121‧‧‧ line

122‧‧‧ line

130‧‧‧ schema

131‧‧‧ line

132‧‧‧ line

140‧‧‧ schema

Line 141‧‧

142‧‧‧ line

150‧‧‧ schema

Line 151‧‧

Line 152‧‧

160‧‧‧ schema

161‧‧‧ line

162‧‧‧ line

170‧‧‧ schema

180‧‧‧ schema

181‧‧‧ line

182‧‧‧ line

190‧‧‧ schema

Line 191‧‧

192‧‧‧ line

200‧‧‧ schema

210‧‧‧ schema

220‧‧‧ Chart

230‧‧‧ Chart

300‧‧‧System/Measuring System

301‧‧‧sample

302‧‧‧ illuminators

304‧‧‧ Spectrometer

306‧‧‧Polarized illumination beam

307‧‧‧Polarized state generator

308‧‧‧ Beam

309‧‧‧Polarized state analyzer

311‧‧‧Spectrum

330‧‧‧Computation System

331‧‧‧ processor

332‧‧‧ memory

333‧‧‧ busbar

334‧‧‧Program Instructions

340‧‧‧sample parameters/output signals

400‧‧‧ Illustrative method

500‧‧‧ Illustrative method

FIG. 1 is a diagram showing a top view of one of the schematic 3x3 contact hole arrays by lithography process simulation software modeling.

2 is a diagram showing a side view 20 of a single hole of one of the contact hole arrays illustrated in FIG. 1.

Figure 3 shows that each contact hole profile is different from one of the exposure and focus depth parameter values. One of the side views 30 of one of the contact hole profile arrays associated with one of the combinations is combined.

FIG. 4 is a diagram showing a top view of one of the contact hole arrays illustrated in FIG.

5 is a contour map 50 of critical dimension values associated with each of the results of the focus exposure matrix illustrated in FIGS. 3-4.

6 is a contour map 60 of sidewall angle values associated with each of the results of the focus exposure matrix illustrated in FIGS. 3-4.

FIG. 7 illustrates one contour map 70 of photoresist loss values associated with each of the results of the focus exposure matrix illustrated in FIGS. 3-4.

Figure 8 is a table 80 showing the correlation between focus and exposure and between critical dimension, height and sidewall angle.

Figure 9 is a diagram 90 showing one of the simplified measurement models of one of the holes in an oxide layer subjected to measurement by a two-dimensional beam profile reflectometer (2-D BPR) system.

Figure 10 is a table 100 showing the correlation between CD and SWA associated with 2-D BPR measurements.

Figure 11 is a table 110 showing the correlation between focus and exposure associated with 2-D BPR measurements.

12A through 12C each illustrate patterns 120, 130, and 140 indicating tracking performance of photoresist loss, SWA, and CD 2-D BPR measurements. The measurement model is parameterized by photoresist loss, SWA and CD.

13A to 13C each illustrate patterns 150, 160, and 170 indicating tracking performance of photoresist loss, SWA, and 2-D BPR measurement of CD. The measurement model is parameterized by photoresist loss, SWA and CD and is limited by the process model.

14A-14B each illustrate graphs 180 and 190 indicating tracking performance associated with 2-D BPR measurements of focus and exposure. The measurement model is parameterized by focus and exposure.

Figure 15 is a diagram 200 showing a structure of a shape change caused by a process variation modeled by a discrete 10-well model.

16 is a diagram 210 showing a result of a focus exposure matrix simulation of the structure illustrated in FIG.

Figure 17A is a diagram 220 showing one of the results of a primary component analysis used as one of the indications of process information presented in process variations (focus and exposure).

Figure 17B is a graph 230 showing one of the results of a primary component analysis used as one of the process information content presented in the measurement signal (e.g., alpha, beta).

18 is a diagram of one of the systems 300 for measuring a feature of a sample in accordance with the exemplary methods presented herein.

FIG. 19 is a flow chart showing one exemplary method 400 suitable for implementation by the metrology system 300 of the present invention.

20 is a flow chart of one exemplary method 500 suitable for implementation by the metrology system 300 of the present invention.

Reference will now be made in detail to the embodiments of the present invention A method and system for optimizing a measurement model based on an integration-based process-based target model and a measurement-based target model are presented. By integrating a measurement-based target model and a process-based target model to improve the predictive outcome of one or both of the measurement model and the process model.

In general, optical metrology techniques measure an indirect method of detecting the physical properties of one of the samples. In most cases, the measured optical signal is not available to directly determine the physical properties of interest. Traditionally, the metrology process consists of a measurement model that attempts to predict one of the measured optical signals based on one of the interactions between the measurement target and the particular measurement system. The measurement-based target model includes parameterization of one of the structures according to the physical properties of the measurement target (eg, film thickness, critical dimension, refractive index, grating pitch, etc.). In addition, the measurement-based target model includes parameterization of one of the measurement tools themselves (eg, wavelength, angle of incidence, polarization angle, etc.).

The machine parameter (P _machine ) is used to characterize the parameters of the measurement tool itself. Exemplary machine parameters comprise angle of incidence (the AOI), Analysis angle (A _0), the angle of polarization (P _o), illumination wavelength, numerical aperture (NA) and the like. The sample parameter (P _specimen ) is used to characterize the geometry and material properties of the sample. For a film sample, exemplary sample parameters include refractive index, dielectric function tensor, nominal layer thickness of all layers, layer sequence, and the like.

For measurement purposes, machine parameters are treated as known fixed parameters and a subset of sample parameters or sample parameters are treated as unknown floating parameters. The floating parameters are resolved by an adaptation process (eg, regression, database matching, etc.) that produces an optimal fit between the theoretical predictions and the measured data. The unknown sample parameter P _{specimen is} varied and the model output value is calculated until it is determined that one of the model output values closely matches one of the measured values to match a set of sample parameter values.

In many cases, sample parameters are highly correlated. This can lead to instability of the target model based on measurement. In some cases, this can be resolved by fixing certain sample parameters. However, this usually results in a significant error in the estimation of the remaining parameters. For example, an underlying layer (eg, an oxide based layer of a semiconductor material stack on a semiconductor wafer) is non-uniformly thick over the surface of a wafer. However, to reduce parameter correlation, the layers are processed to have a fixed thickness above the surface of the wafer to construct a measurement model. Unfortunately, this can lead to significant errors in the estimation of other parameters.

In one aspect, a system employing an integrated metrology model is used to measure the structural and material characteristics of one or more targets (eg, material composition, structure, and dimensional characteristics of the film, etc.). The integrated measurement model is based on the integration of a process-based target model and a measurement-based target model.

In another aspect, a measurement system employing an integrated measurement model is used to directly measure process parameter values. The integrated measurement model is based on an integrated process-based target model and a measurement-based target model.

A process-based target model predicts the structural properties of a sample based on process variables Quality (eg, geometric properties, material properties, etc.). It is suitable for predicting structural and/or material properties with a process-based target model integrated with a measurement-based target model, which is sensitive to these properties.

1 through 2 illustrate, by way of non-limiting example, simulation results of a process-based target model. 1 to 2 illustrate one of the contact hole arrays produced by the positive photoresist optical phantom (PROLITH) simulation software available from KLA-Tencor Corporation of Milpitas, California (USA). Shadow model. Although this exemplary lithography process model is generated using PROLITH software, in general, any process modeling technique or tool can be considered within the scope of this patent document. FIG. 1 illustrates a top view 10 of one of the analog 3x3 contact hole arrays modeled by the PROLITH software. Figure 2 depicts a side view 20 of one of the single holes of the array. Several geometric parameters are depicted from this side view. For example, the height (H), sidewall angle (SWA), and critical dimension (CD) of the hole are shown. In the illustrated example, the critical dimension is measured as one of the diameters of the five nanometers above the bottom of the hole.

A Focus Exposure Matrix (FEM) simulation experiment was run to produce a set of hole profiles similar to the hole profile depicted in FIG. 3 illustrates a side view 30 of one of the contact hole profile arrays associated with one different combination of exposure hole depths and focus parameter values. The results shown in Figure 3 were generated from a series of simulation experiments run by the PROLITH software. 4 is a top plan view 40 of the same contact hole array illustrated in FIG. In this way, the geometric properties of the sample (ie, CD, H, SWA) are represented by process-based target models based on process variables (ie, focus and exposure).

5 to 7 each show a contour map 50 of the CD, a contour map 60 of the SWA, and a contour map 70 of the photoresist loss. Each of the contour maps is associated with the FEM experimental results illustrated in Figures 3 through 4. The photoresist loss is measured from the nominal height of the contact hole minus the actual height and thus the height of the system. Figures 5 through 7 graphically illustrate the relationship between the process parameters derived from the process-based target model and the measurement parameters and thus present a process-based limitation of the measurement parameters. In this way, you can use these restrictions to link based on the process. The basic model and the measurement-based model to generate an integrated measurement model.

Equation (1) shows that one of the three measurement-based model parameters (ie, CD, H, and SWA) is limited by two process-based model parameters (ie, depth of focus F and exposure E). equation.

CD =-90+130 F +1.22 E -417 F ² -0.448 FE -0.0025 E ² Ht =-223-189 F +4.16 E -354 F ² +1.76 FE -0.0158 E ² SWA =56.5+37.4 F +0.431 E -230 F ² -0.1222 FE -0.0019 E ² (1)

In the example illustrated by equation (1), the process-based constraints are formulated based on a simplified adaptation function (eg, a polynomial) applied to the FEM simulation results illustrated in FIGS. 3-7. In other instances, process-based limitations may be formulated based on a process-based basis function, such as those described in U.S. Patent No. 7,009,704, issued to KLA-Tencor Technologies Corporation, on March 7, 2006. The function, the subject matter of which is incorporated herein in its entirety.

A process-based target model can be integrated with a measurement-based target model to produce an integrated measurement model in a number of different ways.

In some examples, a limit to the range of values of the measurement model parameters is determined based on the process-based target model. More specifically, the limits are determined based on the range of achievable values of the process-based parameters.

For example, as illustrated in Figures 3 and 4, some combinations of focus and exposure parameter values do not result in a function contact hole. For example, as illustrated in FIG. 3, the structure highlighted in the area 31 of the matrix of the process-based model results fails to form a hole. Similarly, as illustrated in Figure 4, the structure highlighted in the area 41 of the matrix of model-based model results fails to form a hole. Similarly, as illustrated in Figures 5-7, the combination of focus and exposure parameter values in blank areas 51, 61, and 71 each indicate that one of the functional structures is formed to fail because of certain parameters (i.e., CD, SWA). And photoresist loss) cannot be resolved. In other words, determining certain ranges of focus and exposure parameter values based on the results of the process-based target model does not produce a functional structure. Process-based model parameter values These ranges are mapped to the range of model parameter values based on the measurements to limit the range of model parameter values based on the measurements used in subsequent metrology analysis. This reduces the computation time associated with the integrated metrology model by limiting the range of measurement parameter values to the latent function structure as determined by the process-based model. By way of non-limiting example, equation (1) can be used to map the range of process-based model parameter values to the range of model parameter values based on the measurements.

In some other examples, the integrated measurement model includes a measurement-based target model that is limited by a process-based target model. In one example, the measurement analysis performed by one of the analytical measurement parameters (such as CD, H, and SWA) is limited by equation (1). In other words, only the solution for CD, H and SWA following the constraint equation (1) is considered in the measurement analysis. This reduces the size of the solution space associated with the integrated measurement model. In this way, the process-based restriction set for the measurement-based target model parameters is defined by the process-based target model and applied to the measurement-based target model.

In some other examples, one or more measurement model parameters are represented based on other measurement model parameters based on the process model. In an example, SWA is represented as a function of CD and H based on equation (1). This reduces the total number of floating parameters of the integrated measurement model and reduces the parameter correlation. This also increases the efficiency and robustness of the adaptation engine (eg, the regression engine) due to the smaller search space limited to one of the target changes allowed by the manufacturing process.

In some other examples, process parameters are substituted into the metrology model. These process parameters are then parsed into part of the analysis of the measured data using an integrated measurement model. In this manner, the parameterization of the integrated measurement model includes process-based variables (eg, focus and exposure) and the process parameter values are directly determined from the measurement signals. For example, an integrated measurement model can be formulated by replacing the measurement model parameters CD, H, and SWA with process parameters F and E using equation (1). After solving F and E based on the measurement data, the corresponding measurement parameters CD, H and SWA can be calculated from equation (1).

This method can be preferred to reduce the phase between the parameters involved in the analysis of the measured data. Relevance. For example, as depicted in the table 80 depicted in Figure 8, the correlation between focus and exposure is lower than the correlation between CD, H, and SWA.

Figure 9 is a diagram 90 showing one of the simplified measurement models of one of the holes in an oxide layer subjected to measurement by a two-dimensional beam profile reflectometer (2-D BPR) system. The experimental design (DOE) simulation results confirmed a high correlation between CD and SWA, as depicted in Table 100 of FIG. Therefore, it is expected that the 2-D BPR measurement system does not effectively distinguish between two measurement parameters. This is further confirmed in Figures 12A to 12C. FIG. 12A illustrates one of the tracking performances of the 2-D BPR measurement indicating the resistive loss of the resistive loss model using one of the photoresist loss, SWA, and CD parameterization. Line 121 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 122 represents one of the best fits between the illustrated data points. As shown in FIG. 12A, the tracking performance for resistance loss is better. FIG. 12B illustrates a graph 130 of tracking performance indicating a 2-D BPR measurement of a CD using a measurement model of photoresist loss, SWA, and CD parameterization. Line 131 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 132 represents one of the best fits between the illustrated data points. As shown in FIG. 12B, the tracking performance for the CD is poor. FIG. 12C illustrates one of the tracking performances of the 2-D BPR measurement indicating the SWA using a measurement model of photoresist loss, SWA, and CD parameterization. Line 141 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 142 represents one of the best fits between the illustrated data points. As shown in FIG. 12C, the tracking performance for the SWA is poor.

13A to 13C each illustrate tracking performance associated with 2-D BPR measurement of photoresist loss, CD, and SWA using an integrated measurement model, wherein the amount is parameterized by photoresist loss, CD, and SWA The measurement model is limited by the process model (ie, photoresist loss, CD and SWA are limited by focus and exposure). FIG. 13A illustrates a graph 150 of tracking performance indicating a 2-D BPR measurement of photoresist loss using an integrated measurement model of photoresist loss, SWA, and CD parameterization. Line 151 indicates a perfect chase in which the estimated parameter value is the same as the actual parameter value. trace. Line 152 represents one of the best fits between the illustrated data points. As shown in FIG. 13A, the tracking performance for resistance loss is better. FIG. 13B illustrates a graph 160 showing tracking performance of a 2-D BPR measurement of a CD using an integrated measurement model of photoresist loss, SWA, and CD parameterization. Line 161 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 162 represents one of the best fits between the illustrated data points. As shown in FIG. 13B, the tracking performance for the CD is improved over the tracking performance shown in FIG. 12B. FIG. 13C illustrates one of the tracking performances of the 2-D BPR measurement indicating the SWA using the integrated measurement model of photoresist loss, SWA, and CD parameterization. As depicted in Figure 13C, the tracking performance for the SWA remains poor.

14A-14B each illustrate tracking performance associated with 2-D BPR measurement of focus and exposure when the integrated parameter measurement model is used to parameterize the focus and exposure. FIG. 14A illustrates a graph 180 of tracking performance indicative of a focused 2-D BPR measurement using an integrated metrology model that is parameterized by focus and exposure. Line 181 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 182 represents one of the best fits between the illustrated data points. As shown in Figure 14A, tracking performance for focus is better. Similarly, FIG. 14B illustrates one of the tracking abilities of the 2-D BPR measurement indicating exposure using an integrated measurement model that is parameterized by focus and exposure. Line 191 indicates a perfect trace in which the estimated parameter values are the same as the actual parameter values. Line 192 represents one of the best fits between the illustrated data points. As shown in Figure 14B, the tracking performance for exposure is better. As depicted in the table 110 of Figure 11, the correlation between focus and exposure is significantly less than the correlation between SWA and CD. Therefore, 2-D BPR measurements can have more successful analytical focus and exposure than CD and SWA.

In this manner, an integrated metrology model is used to accurately measure process parameters of interest (eg, depth of focus, exposure, etch time, deposition time, etc.). This method significantly increases the information transmitted between the measurement signal and the measured process parameters, improving accuracy and measurement time.

In another example, an integrated metrology model is constructed based on process simulation results. Usually, a measurement-based target model is a simple approximation of the actual goal. For example, the measurement model used for photoresist line measurement is usually a simple trapezoid in which SWA, CD, and H are measured. This is due to a significant approximation of one of the actual changes in the target geometry of process variations (eg, changes in focus and exposure). As a result, attempts to characterize the focus and exposure based on a highly simplified measurement model can be confirmed to be inconclusive because the actual geometric effects of different process parameter values are not captured by the measurements.

By way of example, Figure 16 illustrates a target profile having a complex sidewall shape in a focus exposure matrix. If a highly simplified measurement model is used, one part of the process variation information depicted in Figure 16 is lost. However, in some instances, an integrated measurement model including a more complex geometric parameterization effectively captures the shape profile caused by the process. In this way, the measurement signal can be effectively analyzed to determine process parameter values.

Figure 15 illustrates a discrete 10 well (12 DOF) model for taking shape changes due to process variations. The model contains 11 CDs at different heights. The measured CD reflection process varies and can more accurately predict the depth of focus and exposure than a single trapezoidal geometric model that is parameterized by CD, SWA, and HT. In this way, the integrated measurement model is based on the measurement model illustrated in FIG. 15 constructed by the process model simulation results illustrated in FIG. However, the sidewall shape variations are depicted in FIG. 16 and are captured by the model depicted in FIG. 15, and other shape variations (eg, photoresist line topology changes, etc.) may be considered.

In some examples, an integrated metrology model is used to sequentially measure process parameter values. The integrated measurement model includes a target model that is used as part of a measurement analysis to determine one of the geometric parameter values from the measurement data based on the measurement. The integrated measurement model also includes a process-based model to determine process parameter values from geometric parameter values.

In one example, the measurement-based target model is parameterized by a process-based principal component analysis (PCA) parameterization described in US Patent Publication No. 2013/0110477 to Stilian Pandev. The full text of the subject matter is cited Incorporated herein. A process-based PCA parameterization effectively reduces the number of degrees of freedom of the model based on the measurement, so that the self-measurement data can effectively resolve the model parameters without excessive loss of measurement information. In one example, the model reduction is performed to limit the geometric discrete model depicted in Figure 15 within the process variation space. The measured geometry (eg, CD of Figure 15) is provided as an input to a process model that predicts focus and exposure. In one example, a neural network is used as a process model. The neural network model is based on the measured CD prediction focus and exposure parameter values.

In a further aspect, the neural network model is trained by process model data. More specifically, the neural network model is trained using shape profiles generated by a PROLITH simulator, such as the simulators depicted in FIG. In this way, the focus and exposure parameter values are represented as a function of the sidewall shape parameters CD ₁ ... CDn and HT, allowing more process information to be captured by the neural network model. This reduces the loss of information from the target measurement to the focus and exposure estimate measurements compared to using a single trapezoidal model with three parameters (ie, CD, SWA, and HT).

Figure 17A illustrates the results of a primary component analysis used as one of the indications of process information presented in process variations (focus and exposure). Figure 17A shows that the process change information is transmitted to the 12 DOF geometric model and the information is effectively captured by two or three PCA components.

Figure 17B illustrates the results of one of the main component analyses used to indicate one of the process information content presented in the measurement signal (e.g., alpha, beta). The more information that is transmitted to the signal or geometric model, the better the measurement/process model can extract and separate process parameters (eg, focus and exposure).

In some examples, an integrated measurement model is used to directly measure process parameter values from the measurement signals.

For example, as discussed above, process parameters can be substituted into a target model based on measurement. In this way, the self-measurement data directly resolves the process parameter values.

In some other examples, the integrated measurement model receives the measurement signal and directly determines One of the focus and exposure parameter values of the neural network model. The neural network model is trained using a shape profile generated by a process model (eg, a shape parameter generated by PROLITH and depicted in FIG. 16) and a corresponding metrology spectrum generated by a measurement model. During neural network training, a process simulator (eg, PROLITH) is used to generate a target profile for a custom process change. A measurement spectrum corresponding to each shape profile is generated by an RCWA engine. Train the neural network using the generated spectra. The number of levels of degrees of freedom is performed by performing PCA on the generated spectrum.

During the measurement, the measured spectrum is received by an analysis engine and the measured spectrum is converted to the main component (PC) by PCA conversion used during training. The trained neural network model receives the PC and directly determines the focus and exposure parameters.

In another example, the neural network can be trained based on the measured spectra from a DOE (FEM) wafer. In this example, a process simulator or module is not required. This reduces errors from the RCWA engine and process simulator but increases the need for process variations in the DOE wafer.

By using an integrated measurement model to directly measure process parameter values from self-measured signals, information loss is reduced by eliminating intermediate models (eg, geometry, materials, or other models of the estimated measurement system). In addition, the measurement time is reduced by eliminating the regression operation.

In another aspect, the process model can be improved based on the measurement model. In some examples, the information obtained using the self-measurement model improves the calibration of a process model. In one example, a pre-characteristic relationship between geometric profile and process variation can be used for process recipe generation. In addition, the process model can be calibrated in one of the optimized measurement models.

Even more generally, the use of process and measurement models is fully integrated with a model that provides input to another model.

In yet another aspect, the disclosed methods and systems can be employed in the context of multi-objective modeling. In some examples, the integrated metrology model allows for a combined analysis in which multiple targets are resolved using the metrology model parameters while parsing some of the targets using process parameters. In addition, you can use the limits derived from a process mode to link parameters for different targets.

Although a number of examples have been described above with reference to a lithography process model and associated focus and exposure measurements, the methods and systems described herein may involve other process models (eg, etching or deposition fabrication) and other measurements ( For example, etching and deposition measurements). The methods and systems described herein may also be directed to other reference measurement techniques (eg, SEM, TEM, AFM, X-ray). Furthermore, the methods and systems described herein are discussed with reference to optical metrology systems (eg, spectral ellipsometers, reflectometers, BPR systems, etc.) but can be applied to other model-based measurements (eg, overlapping). , CD-SAXS, XRR, etc.).

In yet another aspect, a cross-wafer process variation model can be combined with a process-based model of the target structure. Typically, the process affects the entire wafer that creates a specific pattern across the wafer process. In one example, it is generally observed that a film deposition process typically results in a film thickness having one of the radially symmetric patterns across one of the wafers.

In some instances, cross-wafer process information is combined with process variation information encoded in a single target to create a precisely integrated measurement model. In one example, one or more of the underlying film modules are combined with a 10-well model based on process variation parameterization established by PROLITH. The grating-to-membrane correlation is significantly reduced and the measurement accuracy is improved by limiting the variation of the grating into the process space and limiting the underlying film to a desired symmetric film.

In yet another embodiment, an integrated metrology model can be used to provide active feedback to a process tool (eg, lithography tool, etch tool, deposition tool, etc.). For example, the depth value and focus parameters determined using an integrated metrology model can be communicated to the lithography tool to adjust the lithography system to achieve a desired output. In a similar manner, etch parameters (eg, etch time, diffusivity, etc.) or deposition parameters (eg, time, concentration, etc.) can be included in the integrated metrology model to each provide active feedback to the etch tool or deposition tool.

FIG. 18 depicts a system 300 for measuring a feature of a sample in accordance with the exemplary methods presented herein. As shown in FIG. 18, system 300 can be used to perform spectral ellipsometry measurements of one or more structures of a sample 301. In this aspect, system 300 can include a A illuminator 302 and a spectroscopic ellipsometer of one of the spectrometers 304 are provided. Illuminator 302 of system 300 is configured to produce illumination of a selected wavelength range (e.g., 150-850 nm) and direct the illumination to a structure disposed on the surface of sample 301. Next, spectrometer 304 is configured to receive illumination that is reflected from the surface of sample 301. It is further noted that light that emerges from illuminator 302 is polarized using a polarization state generator 307 to produce a polarized illumination beam 306. Radiation reflected by the structure disposed on sample 301 passes through a polarization state analyzer 309 and to spectrometer 304. Regarding the polarization state analysis, the radiation received by the spectrometer in the collection beam 308 allows for spectral analysis of the spectrometer by the radiation transmitted by the analyzer. These spectra 311 are transmitted to computing system 330 for analysis of the structure.

As depicted in Figure 18, system 300 includes a single measurement technique (i.e., SE). In general, however, system 300 can include any number of different measurement techniques. By way of non-limiting example, system 300 can be configured as a spectral ellipsometer (including a Muller matrix ellipsometer), a spectral reflectometer, a spectral scatterometer, an overlapping scatterometer, and an angular resolution beam profile reflection. A combination of a polarized beam profile reflectometer, a beam profile reflectometer, a beam profile polarimeter, any single or multiple wavelength polarimeters, or the like. Moreover, in general, measurement data collected by different measurement techniques and analyzed according to the methods described herein can be collected from multiple tools rather than integrating one of several techniques.

In a further embodiment, system 300 can include one or more computing systems 330 that are employed to perform measurements based on an integrated metrology model in accordance with the methods described herein. The one or more computing systems 330 can be communicatively coupled to the spectrometer 304. In one aspect, one or more computing systems 330 are configured to receive metrology data associated with measurements of the structure of sample 301.

In a further embodiment, one or more computing systems 330 are configured to access the model parameters using Instant Critical Size (RTCD) or to access a database of pre-computed models in accordance with the methods described herein to determine a Integrate the measurement model.

It will be appreciated that the various steps described throughout this disclosure may be by a single computing system 330 Or alternatively, a plurality of computer systems 330 are implemented. Moreover, different subsystems of system 300, such as spectral ellipsometer 304, can include a computer system suitable for implementing at least a portion of the steps described herein. Therefore, the above description of the invention should not be construed as limiting the invention. Moreover, one or more computing systems 330 can be configured to perform any other (several) steps of any of the method embodiments described herein.

Additionally, computer system 330 can be communicatively coupled to spectrometer 304 in any manner known in the art. For example, one or more computing systems 330 can be coupled to a computing system associated with spectrometer 304. In another example, spectrometer 304 can be directly controlled by a single computer system coupled to computer system 330.

Computer system 330 of measurement system 300 can be configured to receive and/or retrieve data or information by a subsystem (e.g., spectrometer 304 and the like) that can transmit media from the system, including one of wired and/or wireless portions. In this manner, the transmission medium can serve as a data link between computer system 330 and other subsystems of system 300.

The computer system 330 of the integrated measurement system 300 can be configured to receive and/or retrieve data or information from other systems (eg, measurement results, modeled inputs, etc.) by transmitting media over one of the wired and/or wireless portions. Modeling results, etc.). In this manner, the transmission medium can serve as a data link between computer system 330 and other systems (eg, on-memory measurement system 300, external memory, reference measurement source 320, or other external system). For example, computer system 330 can be configured to receive measurement data from a storage medium (ie, memory 332 or an external memory) via a data link. For example, spectral results obtained using spectrometer 304 can be stored in a permanent or semi-permanent memory device (eg, memory 332 or an external memory). In this regard, the spectral results can be input from on-board memory or from an external memory system. Moreover, computer system 330 can transmit data to other systems via a transmission medium. For example, one of the integrated measurement models or a sample parameter 340 is determined by computer system 330 to be communicative and stored in an external memory. In this regard, the measurement results can be output to another system.

Computing system 330 can include, but is not limited to, a personal computer system, a host computer system, a workstation, an imaging computer, a parallel processor, or any other device known in the art. In general, the term "computer system" is broadly defined to encompass any device having one or more processors that execute instructions from a memory medium.

Program instructions 334 for transmitting media transport implementation methods, such as those described herein, may be transmitted through one of a wired, cable or wireless transmission link. For example, as shown in FIG. 19, the program instructions 334 stored in the memory 332 are transmitted to the processor 331 through the bus bar 333. Program instructions 334 are stored in a computer readable medium (e.g., memory 332). An exemplary computer readable medium includes read only memory, a random access memory, a magnetic or optical disk, or a magnetic tape.

FIG. 19 illustrates one method 400 suitable for implementation by the metrology system 300 of the present invention. In one aspect, the cognition can implement the data processing block of method 400 via one of the pre-programmed algorithms executed by one or more processors of computer system 330. Although the following description is presented in the context of measurement system 300, the specific structural aspects of cognitive measurement system 300 herein are not meant to be limiting and should be construed as merely illustrative.

In block 401, a quantity of measurement data is received by a computer system (e.g., computer system 330). The measurement data is correlated with the measurement of a target structure by a metrology tool (eg, system 300).

In block 402, one of the quantitative measurement data based on the integrated measurement model to one of the target structures is adapted to determine a set of parameter values that characterize the target structure. The integrated measurement model is based on a process-based target model and a measurement-based target model.

In block 203, the set of parameter values is stored in memory. The second set of parameter values may be stored in onboard measurement system 300 (eg, memory 332) or the second set of parameter values (eg, via output signal 340) may be communicated to an external memory device.

FIG. 20 illustrates one method 500 suitable for implementation by the metrology system 300 of the present invention. In one aspect, the cognition can implement the data processing block of method 500 via one of the pre-programmed algorithms executed by one or more processors of computer system 330. Although the following description is presented in the context of measurement system 500, the specific structural aspects of cognitive measurement system 500 herein are not meant to be limiting and should be construed as merely illustrative.

In block 501, a quantity of measurement data is received by a computer system (e.g., computer system 330). The measurement data is correlated with the measurement of a target structure by a metrology tool (eg, system 300).

In block 502, an integrated measurement model based on the quantitative measurement data and the target structure determines that the characterization is employed to generate one or more process parameter values for one of the target structures.

In block 503, the one or more process parameter values are stored in memory. The parameter values may be stored in onboard measurement system 300 (eg, memory 332) or may be communicated to an external memory device (eg, via output signal 340).

In general, the systems and methods described herein can be implemented as part of a process for preparing an integrated measurement model for offline or instrumental measurement. Additionally, both the measurement model and any of the pre-parametric measurement models can describe one or more target structures and measurement locations.

As described herein, the term "critical dimension" encompasses any critical dimension of a structure (eg, bottom critical dimension, central critical dimension, top critical dimension, sidewall angle, grating height, etc.), any two or more structures One of the critical dimensions (eg, the distance between the two structures) and one of the two or more structures (eg, overlapping shifts between overlapping grating structures, etc.). The structure may include a three-dimensional structure, a patterned structure, an overlapping structure, and the like.

As described herein, the terms "critical dimension application" or "critical dimension measurement application" encompass any critical dimension measurement.

As described herein, the term "measurement system" includes employed at least in part to include quantities such as critical dimension measurements, overlap measurements, focus/dose measurements, and composition measurements. Any system that characterizes a sample in any aspect of the application. However, such technical terms do not limit the scope of the term "measurement system" as described herein. Additionally, measurement system 100 can be configured for measurement of patterned wafers and/or unpatterned wafers. The measurement system can be configured as an LED inspection tool, an edge inspection tool, a backside inspection tool, a giant inspection tool or a multi-mode inspection tool (involving data from one or more platforms at the same time) and a critical dimension based self-system Any other measurement or inspection tool that benefits the calibration of the parameters.

Various embodiments are described herein for a semiconductor processing system (eg, an inspection system or a lithography system) that can be used to process a sample. The term "sample" is used herein to mean a wafer, a master mask, or any other sample that can be processed (e.g., printed or inspected for defects) by methods known in the art.

As used herein, the term "wafer" generally refers to a substrate formed from a semiconductor or non-semiconductor. Examples include, but are not limited to, single crystal germanium, gallium arsenide, and indium phosphide. Such substrates are typically found and/or fabricated in semiconductor fabrication facilities. In some cases, a wafer may only comprise a substrate (ie, a bare wafer). Alternatively, a wafer may comprise one or more layers of different materials formed on a substrate. One or more of the layers formed on a wafer may be "patterned" or "unpatterned." For example, a wafer can include a plurality of dies having repeatable pattern features.

A "main reticle" can be attached to one of the main reticle manufacturing processes at one stage of the master reticle manufacturing process or can be or can not be released for use in one of the semiconductor fabrication facilities to complete the main reticle. A primary reticle or "mask" is generally defined as having substantially transparent substrate formed on one of the substantially non-transparent regions formed thereon and configured in a pattern. The substrate may comprise, for example, a glass material such as amorphous SiO ₂ . A primary reticle can be placed over a photoresist-covered wafer during one of the exposure steps of a lithography process such that the pattern on the primary reticle can be transferred to the photoresist.

One or more of the layers formed on a wafer may be patterned or unpatterned. For example, a wafer can include a plurality of dies, each having repeatable pattern features. The formation and processing of such material layers can ultimately result in a completed device. Many different types of devices can be shaped The term wafer is formed on a wafer, and the term wafer as used herein is intended to encompass a wafer on which any type of device known in the art is fabricated.

In one or more exemplary embodiments, the functions described can be implemented in any combination of hardware, software, firmware, or the like. If implemented in software, the functions may be stored as one or more instructions or code in a computer readable medium or transmitted through a computer readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transmission of a computer program from one location to another. A storage medium can be any available media that can be accessed by a general purpose or special purpose computer. By way of example and not limitation, such computer-readable media may comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device or may be used for transmitting or storing instructions or The desired program code in the form of a data structure and any other medium that can be accessed by a general purpose or special purpose computer or a general purpose or special purpose processor. In addition, any connection is appropriately referred to as a computer readable medium. For example, if a coaxial cable, fiber optic cable, twisted pair cable, digital subscriber line (DSL), or other remote technologies such as infrared, radio, and microwave technologies are used to transmit software from a website, coaxial cable, fiber optic cable. , twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of the media. Disks and optical discs as used herein include compact discs (CDs), laser discs, optical discs, digital versatile discs (DVDs), floppy discs, and Blu-ray discs, where the discs are usually magnetically replicated and the discs are laser optics. Copy the data. Combinations of the above should also be included in the scope of computer readable media.

Although certain specific embodiments are described above for illustrative purposes, the teachings of this patent document have general applicability and are not limited to the specific embodiments described above. Therefore, various modifications, adaptations and combinations of the various features of the described embodiments can be made without departing from the scope of the invention as set forth in the appended claims.

300‧‧‧System/Measuring System

301‧‧‧sample

302‧‧‧ illuminators

304‧‧‧ Spectrometer

306‧‧‧Polarized illumination beam

307‧‧‧Polarized state generator

308‧‧‧ Beam

309‧‧‧Polarized state analyzer

311‧‧‧Spectrum

330‧‧‧Computation System

331‧‧‧ processor

332‧‧‧ memory

333‧‧‧ busbar

334‧‧‧Program Instructions

340‧‧‧sample parameters/output signals

Claims (31)

A method comprising: receiving, by a metrology tool, a quantitative measurement data associated with a measurement of a target structure; and integrating one of the quantitative measurement data of the measurement model based on one of the target structures Adapting to determine a set of parameter values that characterize the target structure, wherein the integrated measurement model is based on a process-based target model and a measurement-based target model; and storing the set of parameter values in In a memory. The method of claim 1, wherein the process-based target model characterizes the target structure based on at least one process variable. The method of claim 2, further comprising: generating the one based on one of a process-based target model or one of a plurality of parameters and one or more parameters of the measurement-based target model Integrate the measurement model. The method of claim 3, wherein the link is limited to one of a range of values of the one of the measurement model parameters based on a range of process model parameter values. The method of claim 3, wherein the link is limited to one of the one or more measurement parameters determined from the process-based target model. The method of claim 3, wherein the link is a functional relationship between two or more of the measured model parameters determined by the process-based target model. The method of claim 1, further comprising: determining one or more process parameter values based on one of the quantitative measurement data of the integrated measurement model to the target structure; and the one or more processes The parameter values are stored in a memory. The method of claim 7, wherein the one or more process parameters are parameters of the integrated measurement model of the target structure. The method of claim 7, wherein the one or more process parameters are only parameters of the integrated measurement model of the target structure. The method of claim 7, wherein the determining the one or more process parameter values relates to determining the one or more measurement parameter values based on the adaptation of the quantitative measurement data to the integrated measurement model and based on the The one or more process parameter values are determined in a functional relationship with the one or more measurement parameters to determine the one or more process parameter values. The method of claim 7, wherein the determining the one or more process parameter values involves converting the measurement data to a primary component and determining the one or more process parameter values directly from the primary components. The method of claim 7, wherein the one or more process parameter values comprise any one of a depth of focus value, an exposure value, an etch time, and a deposition time. A method comprising: receiving, by a metrology tool, a quantitative measurement data associated with a measurement of a target structure; determining one or more process parameter values of the characterization process, the process being employed to The quantitative measurement data and one of the target structures are integrated with the measurement model to generate the target structure; and the one or more process parameter values are stored in a memory. The method of claim 13, wherein the determining the one or more process parameters relates to determining at least one geometric parameter based on the one of the quantitative measurements to the target model based on the measurement, and based on the The process-based target model and the at least one geometric parameter determine the one or more process parameter values. The method of claim 14, wherein the process-based target model is a nerve Network model. The method of claim 15, further comprising: training the neural network model based at least in part on the process model data. The method of claim 13, wherein the integrated measurement model is a neural network model, and further comprising: training the neural network model based at least in part on the process model data and the measurement signals corresponding to the process model data. The method of claim 13, further comprising: generating the integrated measurement model based on a simulation result of a process-based target model. The method of claim 13, wherein the one or more process parameter values comprise any one of a depth of focus value, an exposure value, an etch time, and a deposition time. The method of claim 13, wherein the determining the one or more process parameter values characterizing a process involves converting the measurement data to a primary component and determining the one or more process parameter values directly from the primary components. A system comprising: an optical metrology tool comprising an illumination source and a detector configured to perform a measurement of a target structure; and a computer system configured to: by the amount The measuring tool receives a quantitative measurement data associated with the measurement of a target structure; determining, based on one of the quantitative measurement data of the integrated measurement model to the target structure, determining the characteristic structure of the target structure A set of parameter values, wherein the integrated measurement model is based on a process-based target model and a measurement-based target model; and storing the set of parameter values in a memory. The system of claim 21, wherein the process-based target model characterizes the target structure based on at least one process variable. The system of claim 22, wherein the integrated measurement model is based on a chain between one or more parameters of the process-based target model and one or more parameters of the measurement-based target model Knot. The system of claim 21, wherein the integrated measurement model is based on a simulation result of a process-based target model. The system of claim 21, further comprising: determining one or more process parameter values based on one of the quantitative measurement data of the integrated measurement model to the target structure; and the one or more processes The parameter values are stored in a memory. The system of claim 25, wherein the determining the one or more process parameter values involves converting the measurement data to a primary component and determining the one or more process parameter values directly from the primary components. The system of claim 25, wherein the one or more process parameter values comprise any one of a depth of focus value, an exposure value, an etch time, and a deposition time. A system comprising: a metrology tool comprising an illumination source and a detector configured to perform a measurement of a target structure; and a computer system configured to: use the measurement The tool receives a quantitative measurement data associated with the measurement of a target structure; determining one or more process parameter values of the characterization process, the process being employed to measure the data and the target structure based on the quantitative An integrated measurement model to generate the target structure; and The one or more process parameter values are stored in a memory. The system of claim 28, wherein the determining the one or more process parameters relates to determining at least one geometric parameter based on the one of the quantitative measurements to the target model based on the measurement, and based on the The process-based target model and the at least one geometric parameter determine the one or more process parameter values. The system of claim 28, wherein the determining the one or more process parameter values relates to converting the measurement data to a primary component and determining the one or more process parameter values directly from the primary components. The system of claim 28, wherein the one or more process parameter values comprise any one of a depth of focus value, an exposure value, an etch time, and a deposition time.