KR102642534B1 - Apparatus and method for performing processing for improving measurement accuracy of the thickness of each layer of a multilayer thin film - Google Patents

Abstract

A processing device is provided. The processing device may include a processing unit that improves the distinguishability of each layer of the structure with respect to changes in the spectrum when measuring the spectrum of the multilayer thin film structure.

Description

Processing apparatus and method for improving the accuracy of measuring the thickness of each layer of a multilayer thin film {Apparatus and method for performing processing for improving measurement accuracy of the thickness of each layer of a multilayer thin film}

The present invention relates to an apparatus and method for measuring layer thickness of multilayer thin film structures.

When attempting to optically measure thickness or CD (Critical Dimension) in the semiconductor manufacturing process, a regression analysis method is used using the measured spectrum and theoretically calculated spectrum, or machine learning is trained using pre-prepared reference data. Use models.

However, for very thick structures of several microns or more, interference fringes appear on the spectrum not only due to the influence of the existing structure or physical properties, but also due to optical interference effects. Since these influences or interference patterns appear mixed with the information of the parameters to be obtained, it is difficult to reduce the adverse effects caused by them using existing technologies.

Korean Patent Publication No. 1005179 discloses a technology for improving measurement speed using optical interference.

Korean Patent Publication No. 1005179

The present invention is intended to provide a processing device and method that can improve parameter precision in the field of measurement of multilayer thin films.

The processing device of the present invention may include a processing unit that improves the distinguishability of each layer of the structure with respect to changes in the spectrum when measuring the spectrum of a multilayer thin film structure.

The processing method of the present invention includes a conversion step of Fourier transforming the spectrum in the spectrum measurement of the multilayer thin film structure; A processing step of removing the remainder from the real terms and imaginary terms resulting from the Fourier transform, excluding terms whose correlation satisfies a target set value; It may include a restoration step of inverse Fourier transforming the spectrum from which the rest except for the term that satisfies the target set value has been removed.

The processing device of the present invention can increase precision in determining parameters by lowering the influence of correlation between changes in parameters that are difficult to distinguish on the spectrum due to the high correlation between parameters.

The processing device of the present invention can perform Fourier analysis on the measured spectrum to reduce the influence of correlation between parameters.

For this reason, the processing device of the present invention can be used to improve the analysis accuracy or prediction accuracy of parameters by physical models or machine learning.

In particular, the processing device of the present invention can be used to obtain characteristic information such as thickness or OCD (Optical Critical Dimension) structural parameters from spectral measurement information for a semiconductor thin film or pattern structure that is very thick and has many parameters.

1 is a block diagram showing a processing device of the present invention.
Figure 2 is a flow chart showing the processing method of the present invention.
3 is a schematic diagram showing the operation of the processing device.
Figure 4 is a schematic diagram showing a multilayer thin film structure and its spectrum simulation results.
Figure 5 is a graph showing the results of adjusting the spectral correlation for parameters by adjusting the wavelength range of the existing spectrum.
Figure 6 is a graph showing the results of removing a specific Fourier term.
Figure 7 is a graph showing the source of the spectral contribution of Fourier terms that results in high correlation.
Figure 8 is a chart comparing the correlation between spectral changes due to changes in the thickness of the oxide and nitride layers in the 30th and 60th layers of the structure and changes in the thickness of other oxide and nitride layers.
Figure 9 is a graph comparing the thickness prediction degree of the 100th layer by machine learning model.
Figure 10 is a graph comparing the thickness prediction accuracy of each layer by a machine learning model before and after using the Fourier transform.
11 is a diagram illustrating a computing device according to an embodiment of the present invention.

Below, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts that are not related to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

In this specification, duplicate descriptions of the same components are omitted.

Also, in this specification, when a component is mentioned as being 'connected' or 'connected' to another component, it may be directly connected or connected to the other component, but may be connected to the other component in the middle. It should be understood that may exist. On the other hand, in this specification, when it is mentioned that a component is 'directly connected' or 'directly connected' to another component, it should be understood that there are no other components in between.

Additionally, the terms used in this specification are merely used to describe specific embodiments and are not intended to limit the present invention.

Also, in this specification, singular expressions may include plural expressions, unless the context clearly dictates otherwise.

In addition, in this specification, terms such as 'include' or 'have' are only intended to designate the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, and one or more It should be understood that this does not exclude in advance the presence or addition of other features, numbers, steps, operations, components, parts, or combinations thereof.

Also, in this specification, the term 'and/or' includes a combination of a plurality of listed items or any of the plurality of listed items. In this specification, 'A or B' may include 'A', 'B', or 'both A and B'.

Additionally, in this specification, detailed descriptions of well-known functions and configurations that may obscure the gist of the present invention will be omitted.

1 is a block diagram showing a processing device of the present invention.

The processing device shown in FIG. 1 may include a processing unit 130, a conversion unit 110, a restoration unit 150, and an analysis unit 170.

When measuring the spectrum of a multilayer thin film structure, for example, a semiconductor in which a plurality of thin films or layers are stacked, the processing unit 130 can improve the distinguishability of each layer of the structure with respect to changes in the spectrum.

The spectrum may contain information indicating the layers that form a structure such as a semiconductor. At this time, it is advantageous for the analysis of the specific layer that the information of a specific layer is clearly distinguished from the information of other layers. However, in the spectrum obtained using optical measurement equipment such as a spectroscopic ellipsometer, the response information of a specific layer may be distinguished from the response information of other layers in different wavelength regions, but in general, the information from each other is in the same wavelength region. may be mixed. In other words, the spectrum may contain a mixture of first elements from the first layer of the structure and second elements from the second layer of the structure. In a situation where information from different layers or materials is mixed in the spectrum, accuracy and precision may be reduced when analyzing the characteristics (for example, thickness) of the structure of the specific layer. This problem may similarly occur in other structures that invade information on a specific layer.

Therefore, in order to improve the analysis precision and accuracy of the characteristic information of each layer included in the multilayer thin film structure, such as thickness, it is better for each layer of the structure to be clearly distinguished from each other. At this time, the state, degree, and ratio in which each layer of the structure is distinguished from each other is defined as 'distinguishability'.

Improvement in distinguishability performed by the processing unit 130 may mean that differentiation is increased. Increasing, improving, or improving distinguishability may mean that the extent and ratio of other layers included in the information of a specific layer included in the spectrum decreases. Alternatively, increasing, improving, or improving the distinguishability may mean that the possibility, degree of distinction, rate, and amount of distinction that can be made to distinguish the information of a specific layer included in the spectrum from the information of other layers increases. .

The analysis unit 170 may analyze the spectrum whose distinguishability has been improved by the processing unit 130. The analysis unit 170 can determine the characteristics of each layer through spectrum analysis.

Meanwhile, the processing unit 130 can improve the distinguishability of each layer of the structure using various methods.

For example, the processing unit 130 may improve distinguishability by reducing the spectral correlation between each layer of the structure. High correlation between each layer may mean that it is difficult to distinguish changes in the physical properties or structure of each layer. Therefore, according to the processing unit 130 that reduces correlation, it becomes easier to distinguish each layer, which may mean improved distinguishability.

The processing unit 130 may reduce the correlation of each layer of the structure by narrowing the wavelength range of the spectrum. However, when narrowing the wavelength range, information on specific layers requiring analysis may be missed, so reducing correlation through adjusting the wavelength range needs to be approached cautiously. Narrowing the wavelength range may mean increasing the lower wavelength limit of the original spectrum or lowering the upper wavelength limit of the original spectrum.

Instead of adjusting the wavelength range of the spectrum, the processor 130 may reduce the correlation of each layer of the structure by removing some elements from the spectrum (which also have the form of a spectrum). For example, the processing unit 130 may reduce correlation by removing elements of other layers (eg, substrates) that affect a specific layer, some layers, or all layers from the spectrum.

In other words, the processing unit 130 can effectively distinguish and remove the second element among the first element of the first layer and the second element of the second layer that exist mixed in the spectrum. However, since the first element and the second element are mixed in the same wavelength range, it is impossible to remove only the second element while leaving the first element intact, such as by reducing the wavelength range. In other words, the processing unit may remove from the spectrum the first element of the first layer present in the spectrum and the second element of the second layer mixed with it.

For example, spectrally, components of the substrate (corresponding to the second element) may be mixed in the same wavelength range as components of a specific thin film layer (corresponding to the first element). In order to accurately analyze the properties of a specific thin film layer, it is best to remove components of the substrate that have the same wavelength range as the specific thin film layer. However, in reality, at current dimensions such as wavelength, it is difficult to distinguish only the components of the substrate in the corresponding wavelength band. The conversion unit 110 can be used to solve this problem.

The conversion unit 110 may convert the first dimension spectrum into a second dimension in which the first element and the second element can be distinguished or the correlation between the first element and the second element is reduced. At this time, the processing unit 130 may remove the second element in the second dimension.

For example, a situation in which the first element and the second element are mixed may correspond to a situation in the first dimension, for example, the wavelength dimension. This situation can be alleviated by changing to a second dimension, such as frequency, that is distinct from the first dimension. In other words, in the second dimension, compared to the first dimension, the correlation between the first element and the second element is lowered or they can be easily distinguished.

The restoration unit 150 may restore the spectrum from which the second element has been removed to the original first dimension. The spectrum restored to the first dimension may be analyzed by the analysis unit 170.

Meanwhile, when the converter 110 converts the spectrum in the wavelength dimension to the frequency dimension, the converter 110 may perform Fourier transform on the spectrum in order to convert the spectrum in the wavelength dimension to the frequency dimension. At this time, the processing unit 130 may remove the remainder from the real terms and imaginary terms that are the results of the Fourier transform, excluding terms that satisfy the correlation criteria.

The processing unit 130 may remove a specific frequency band from the spectrum converted to the frequency dimension. As an example, the processing unit 130 may remove a frequency band representing the substrate of the structure. Specifically, the processing unit 130 may remove low harmonics corresponding to the initial frequency band in which waveform changes appear from the spectrum converted to the frequency domain.

Figure 2 is a flowchart showing the processing method of the present invention.

The processing method of FIG. 2 can be performed by the processing device shown in FIG. 1.

The processing method may include a conversion step (S 510), a processing step (S 520), a restoration step (S 530), and an analysis step (S 540).

The conversion step (S510) may Fourier transform the spectrum in the spectrum measurement of the multilayer thin film structure. The conversion step (S510) may be performed by the conversion unit 110.

The processing step (S 530) may remove the real part and the imaginary part, which are the results of the Fourier transform, except for the term that ensures that the correlation satisfies the set value. The processing step (S530) may be performed by the processing unit 130.

The restoration step (S 550) can inverse Fourier transform the spectrum with only terms that satisfy the set value in the processing step (S 530). The restoration step (S550) may be performed by the restoration unit 150.

In the analysis step (S 570), the spectrum restored through the restoration step (S 550) can be analyzed. In the analysis step (S 570), the characteristics of each layer can be identified through analysis of the inverse Fourier transformed spectrum. In other words, the analysis step (S 570) can determine the characteristics of the structure through spectrum analysis.

The analysis step (S 570) can determine whether the characteristics of the structure satisfy the precision set as the goal.

In the processing method of the present invention, if the characteristics of the structure identified through the analysis step (S 570) do not satisfy the target precision, a processing step (S 530), a restoration step (S 550), until the characteristics of the structure satisfy the target precision. The analysis step (S 570) can be repeated.

The analysis step (S570) may be performed by the analysis unit 170. The analysis unit 170 may be configured to analyze the spectrum in the wavelength dimension corresponding to the first dimension. Accordingly, the restoration step (S 530) may be preceded.

3 is a schematic diagram showing the operation of the processing device. Figure 4 is a schematic diagram showing a multilayer thin film structure and its spectrum simulation results.

An acquisition unit may be provided to perform simulation or measurement of the spectrum from which the parameter value desired to be measured is identified. The acquisition unit may acquire a simulated spectrum or an actual measured spectrum (S 11).

The spectrum was performed on a multilayer thin film structure and may include information representing each of a plurality of layers. At this time, it is difficult to clearly distinguish whether the spectrum has changed because of the thickness of a specific layer or because the thickness of another layer has changed. In relation to this, a correlation unit may be provided to analyze the correlation of each layer or to analyze the correlation of the spectrum according to parameter differences (S 13).

Correlation R(X,Y) can be calculated using Equation 1 below.

...Formula 1

X and Y are elements representing different layers included in the spectrum, and Cov is the covariance. is the standard deviation.

If the correlation analyzed by the correlation unit satisfies the threshold, the conversion unit 110 can operate.

The converter 110 may perform Fourier transform on the spectrum (S 10).

The processing unit 130 may select a specific Fourier term from the Fourier transformed value or spectrum (S30). Then, the processing unit 130 may remove the specific selected Fourier term.

The restoration unit 150 may perform inverse Fourier transform on the spectrum processed by the processing unit 130 (S 50).

The analysis unit 170 may apply the inverse Fourier transformed spectrum to the physical model or machine learning model and analyze the precision of parameter values (S 70).

The analysis unit 170 may determine whether the target precision has been obtained (S 71). If the calculated precision satisfies the set precision, the analysis unit 170 may store the inverse Fourier transformed spectrum (Fourier transform module) (S 73). If the pre-analyzed precision does not satisfy the set precision, the analysis unit 170 may request the processing unit 130 to select and remove a Fourier term different from the existing one. In this case, the processing unit 130 may restore the previously selected and removed Fourier terms and select and remove other Fourier terms. Alternatively, the processing unit 130 may additionally select and additionally remove other Fourier terms while leaving the previously selected and removed Fourier terms as they are without restoring them.

At the request of the analysis unit 170, the processing unit 130, the restoration unit 150, and the analysis unit 170 may perform the above processes S30, S50, S70, and S71 again.

Figure 5 is a graph showing the results of adjusting the spectral correlation for parameters by adjusting the wavelength range of the existing spectrum.

Figures 5 (a) and (c) show spectra according to changes in the 30th and 60th oxide layer SiO ₂ and nitride layer SiN. Figures 5 (b) and (d) show the correlation between spectra according to changes in the oxide layer and nitride layer for each layer. Looking at it, it can be seen that the correlation R ² decreased when the wavelength range was reduced from 200 to 1000 nm to 230 to 600 nm.

Figure 6 is a graph showing the results of removing a specific Fourier term.

Looking at Figure 6, it can be seen that the correlation of the spectrum is reduced by Fourier transforming the spectrum and removing the remaining terms that satisfy the correlation setting target value in the real part and imaginary part. You can.

Figure 7 is a graph showing the source of the spectral contribution of Fourier terms that results in high correlation.

For Fourier terms satisfying the set values, it was confirmed that the reflectance calculated through inverse Fourier transform was very similar to the reflectance by the silicon substrate. Through this, it was determined that the previously removed terms were related to the influence of the silicon substrate, which commonly contributes highly to different spectra, and it was confirmed that removal of the terms was effective in lowering the correlation of the spectra. Looking at it from another perspective, the spectral components of the silicon substrate can be viewed as having an overall effect on the spectral components of other layers. From this perspective, it can be seen that the correlation of the spectrum can be lowered through various methods of removing substrate elements from the spectrum.

Figure 8 is a chart comparing the correlation between spectral changes due to changes in the thickness of the oxide and nitride layers in the 30th and 60th layers of the structure and changes in the thickness of other oxide and nitride layers.

Figures 8 (a), (b), (c), and (d) show the correlation of spectral changes due to thickness changes with all other layers for the 30th and 60th oxide and nitride layers. According to the processing device and processing method of the present invention, it can be seen that the correlation can be lowered by about 10 to 30%.

Figure 8(e) shows a comparison table of correlation sizes of spectra according to changes in the 30th and 60th oxide and nitride layers.

Figure 9 is a graph comparing the thickness prediction degree of the 100th layer by machine learning model.

Compared to before using the Fourier transform, it can be seen that the thickness prediction accuracy of machine learning was greatly improved after lowering the spectral correlation using the Fourier transform (8000 out of 10000 spectral elements were used for machine learning training, 2000 The above results were obtained using dogs for testing). The improvement in accuracy here can be achieved by considering the spectral differences due to systematic differences that may exist between actual measurement equipment and the resulting influence on the analysis model, which can be achieved by considering thousands of spectra with complex differences in factors within the thin film structure. Since it is a predictive evaluation of a case, it can be interpreted as an improvement in precision that can predict the same tendency repeatedly.

Figure 10 is a graph comparing the thickness prediction accuracy of each layer by a machine learning model before and after using the Fourier transform.

Upon examination, it can be seen that the accuracy 'FFT (Fast Fourier Transform)' when the processing device is applied is higher than the accuracy 'No FFT' before the processing device is applied.

11 is a diagram illustrating a computing device according to an embodiment of the present invention. Computing device TN100 of FIG. 11 may be a device (e.g., processing device, etc.) described herein.

In the embodiment of FIG. 11 , the computing device TN100 may include at least one processor TN110, a transceiver device TN120, and a memory TN130. Additionally, the computing device TN100 may further include a storage device TN140, an input interface device TN150, an output interface device TN160, etc. Components included in the computing device TN100 may be connected by a bus TN170 and communicate with each other.

The processor TN110 may execute a program command stored in at least one of the memory TN130 and the storage device TN140. The processor TN110 may refer to a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to embodiments of the present invention are performed. Processor TN110 may be configured to implement procedures, functions, and methods described in connection with embodiments of the present invention. The processor TN110 may control each component of the computing device TN100.

Each of the memory TN130 and the storage device TN140 can store various information related to the operation of the processor TN110. Each of the memory TN130 and the storage device TN140 may be comprised of at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory TN130 may be comprised of at least one of read only memory (ROM) and random access memory (RAM).

The transceiving device TN120 can transmit or receive wired signals or wireless signals. The transmitting and receiving device (TN120) can be connected to a network and perform communication.

Meanwhile, the embodiments of the present invention are not only implemented through the apparatus and/or method described so far, but may also be implemented through a program that realizes the function corresponding to the configuration of the embodiment of the present invention or a recording medium on which the program is recorded. This implementation can be easily implemented by anyone skilled in the art from the description of the above-described embodiments.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims are also possible. It falls within the scope of invention rights.

110... conversion unit 130... processing unit
150... Restoration Department 170... Analysis Department

Claims (14)

It includes a processing unit that improves the distinction of each layer of the structure with respect to changes in the spectrum when measuring the spectrum of a multilayer thin film structure,
When the first element of the first layer and the second element of the second layer present in the spectrum are mixed,
A conversion unit is provided to convert the spectrum in the first dimension into a second dimension in which the first element and the second element are distinguished,
The processing unit removes the second element in the second dimension.
According to paragraph 1,
An analysis unit is provided to analyze the spectrum with improved discriminability by the processing unit,
The analysis unit is a processing device that determines the characteristics of each layer through analysis of the spectrum.
According to paragraph 1,
A processing device in which the processing unit improves the distinguishability by reducing the correlation of each layer of the structure.
According to paragraph 1,
The processing unit reduces the correlation of each layer of the structure by narrowing the wavelength range of the spectrum.
According to paragraph 1,
The processing unit reduces the correlation of each layer of the structure by removing some elements from the spectrum.
According to paragraph 1,
A processing device in which the spectrum is converted using Fourier transform.
delete According to paragraph 1,
A processing device provided with a restoration unit that restores the spectrum from which the second element has been removed to the original first dimension.
According to paragraph 1,
A conversion unit is provided to convert the spectrum in the wavelength dimension to the frequency dimension,
The processing unit is a processing device that removes a specific frequency band from the spectrum converted to the frequency dimension.
According to clause 9,
The processing unit is a processing device that removes a frequency band representing the substrate of the structure.
It includes a processing unit that improves the distinction of each layer of the structure with respect to changes in the spectrum when measuring the spectrum of a multilayer thin film structure,
A transformer is provided to Fourier transform the spectrum,
The processing unit is a processing device that removes the remaining terms from the real and imaginary terms that are the result of the Fourier transform except for terms whose correlation satisfies a target set value.
According to paragraph 1,
A transformer is provided to Fourier transform the spectrum,
The processing unit is a processing device that removes a low harmonic corresponding to an initial frequency band in which a waveform change appears from the spectrum converted to the frequency domain.
In a processing method performed by a processing device,
A transformation step of Fourier transforming the spectrum in spectrum measurement of a multilayer thin film structure;
A processing step of removing the remainder from the real and imaginary terms resulting from the Fourier transform except for terms whose correlation satisfies a target set value;
A restoration step of inverse Fourier transforming the spectrum from which all but the terms satisfying the target set value have been removed;
Processing method including.
According to clause 13,
Further comprising an analysis step of analyzing the inverse Fourier transformed spectrum,
In the analysis step, the characteristics of the structure are identified through analysis of the spectrum,
The analysis step determines whether the characteristics of the structure satisfy the set precision,
If the characteristics of the structure do not satisfy the set precision, the processing step, the restoration step, and the analysis step are repeated until the characteristics of the structure satisfy the set precision.

Images