Overlay error measurement system and method based on angle resolution scatterometer

Abstract

The invention discloses an overlay error measurement system and method based on an angle resolution scatterometer, and the system comprises a computer control and processing system, a light source and a to-be-measured sample, and the axis of the light source is perpendicular to the axis of the to-be-measured sample. A collimating lens, a polarizing film, a first diaphragm, a first relay lens group and a first beam splitter prism are sequentially arranged along the irradiation direction of the light source, and the first diaphragm is placed at the conjugate plane position of the rear focal plane of the microscope objective; the alignment error measuring system based on the angle resolution scatterometer has the beneficial effects that the alignment error measuring system based on the angle resolution scatterometer is designed by utilizing the advantages of low cost, no damage, high efficiency and the like of an angle resolution scatterometer measuring technology. And on the basis of the basic principle that the positive and negative primary light intensity difference and the overlay error are approximately in a linear relationship within a certain range, a neural network model is trained for data fitting. And obtaining a predicted overlay error through the measured scattered light intensity information and the trained neural network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of semiconductor manufacturing and detection technology, in particular to a system and method for overlay error measurement based on an angle-resolved scatterometer. BACKGROUND

[0002] With the feature size of photolithography process in integrated circuit manufacturing being continuously reduced, the requirement for overlay error index is gradually improved, and there is an urgent need for an overlay error measurement system and method with sub-nanometer precision. Overlay error refers to the alignment deviation between adjacent circuit layers in a chip during exposure. Rapid measurement and accurate evaluation of overlay error are the key to optimization of lithography machine operating parameters and process yield management.

[0003] Currently, common overlay error measurement methods include an image-based overlay error measurement method (IBO) and a diffraction-based overlay error measurement method (DBO). The IBO measurement technology uses a high-resolution bright-field optical microscope with image recognition and measurement functions to measure the deviation of the pattern position in a specially designed overlay mark to achieve overlay error measurement. The overlay mark in the DBO measurement technology is a specially designed nanometer grating structure, and the overlay error is extracted by measuring the diffraction signals of the overlay mark, such as spectrum or angle-resolved spectrum, through certain methods. Compared with the IBO technology, the DBO technology is gradually becoming the main means of overlay error measurement in advanced nodes, without being limited by the diffraction limit and tool-induced deviation. The optical scattering measurement method can obtain information such as nanometer structure morphology parameters by matching or fitting the light intensity or reflectivity of the sample scattering field with the scattering field information generated by the model. This technology has the advantages of low cost, non-destructive, high efficiency, etc. SUMMARY

[0004] The present application aims to provide an overlay error measurement system and method based on an angle-resolved scatterometer to solve the problems raised in the background.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solution: an overlay error measurement system based on an angle-resolved scatterometer, comprising a computer control and processing system, a light source and a sample to be measured, the axis of the light source is perpendicular to the axis of the sample to be measured,

[0006] A collimating lens, a polarizer, a first aperture, a first relay lens group and a first light splitting prism are sequentially arranged along the irradiation direction of the light source, and the first aperture is placed at the conjugate position of the back focal plane of the microscope objective;

[0007] The sample to be measured is arranged at a position coaxial with the first light splitting prism, and a microscope objective is arranged between the first light splitting prism and the sample to be measured;

[0008] The second number of relay lens group, the second number of diaphragm, the second number of light splitting prism are sequentially arranged in the direction away from the microscope objective.

[0009] The second number of light splitting prism divides the emergent light containing sample information into two paths, which are received by the first number of area array CCD camera directly and the second number of area array CCD camera through imaging lens; the first number of area array CCD camera and the second number of area array CCD camera are in communication connection with the computer control and processing system, wherein the first number of area array CCD camera corresponds to the main light path and collects pupil image; the second number of area array CCD camera corresponds to the auxiliary light path and collects spatial image, two cameras are arranged to realize real-time regulation and control of relative spatial position of the measured overlay mark and the detection spot; the positive and negative first-order light intensity difference is collected for analyzing overlay error.

[0010] As preferred, the light source is a supercontinuum laser white light source.

[0011] As preferred, the polarizer is located between the collimating lens and the first number of diaphragm, for polarization state regulation of the incident parallel light beam.

[0012] As preferred, the aperture size and orientation of the first number of diaphragm and the second number of diaphragm are dynamically adjustable, the first number of diaphragm is used to control the size and angle of the light spot irradiated to the sample, and the second number of diaphragm is used to control the transmission of specific order diffraction light.

[0013] As preferred, the first number of relay lens group and the second number of relay lens group are both composed of two groups of double-cemented spherical lenses, the focal length ratio of the first number of relay lens group is used to change the scaling ratio of the incident light path, combined with the fiber core diameter, the collimating lens and the microscope objective magnification, to realize the control of the detection spot size.

[0014] As preferred, the light splitting ratio of the first number of light splitting prism and the second number of light splitting prism is:.

[0015] As preferred, the microscope objective is a high numerical aperture microscope objective, used to collect scattered light field in a large incident angle range.

[0016] As preferred, the sample to be measured is placed on a displacement stage that can move in X-Y-Z three axes, and is ensured to be within the detection spot range and at the focal point position of the microscope objective under the premise of being perpendicular to the optical axis.

[0017] An overlay error measurement method based on an angle-resolved scatterometer, comprising the following steps:

[0018] S1, controlling two area array CCD cameras to collect images through a computer control and processing system;

[0019] S2, after the pupil image and the spatial image are collected by the first area array CCD camera and the second area array CCD camera, the pupil image and the spatial image are fed back to a computer control and processing system;

[0020] S3, the computer control and processing system establishes a system simulation model according to the measured scattering light information;

[0021] S4, after the rationality of the simulation model is verified, the overlay error is simulated and calculated.

[0022] As preferred, the system simulation model is based on the basic principle that the positive and negative first-order light intensity differences and the overlay error are approximately linearly related within a certain range, the relationship between the positive and negative first-order light intensity differences and the overlay error is trained by a neural network, and the specific operation steps of the model include;

[0023] 1) data set is made, a certain number of data samples of the positive and negative first-order light intensity differences and the overlay error are collected through experiments;

[0024] 2) data preprocessing is performed, including data cleaning, removing outliers, normalization and the like, so as to ensure the accuracy and consistency of the data;

[0025] 3) in the network training stage, the preprocessed data samples are input into the neural network for training;

[0026] 4) after the training is completed, the mean square error is used to evaluate the neural network model;

[0027] 5) according to the positive and negative first-order light intensity differences measured by the actual system and the trained neural network model, the value of the overlay error is predicted.

[0028] Compared with the prior art, the beneficial effects of the present application are that a kind of overlay error measurement system based on angle resolved scatterometer is designed by using the advantages of low cost, non-destructive and high efficiency of angle resolved scatterometer measurement technology.And based on the basic principle that the positive and negative first-order light intensity differences and the overlay error are approximately linearly related within a certain range, a neural network model is trained for data fitting.Through the measured scattering light intensity information and the trained neural network, the predicted overlay error is obtained.Compared with the prior art, the method and system proposed in the present application can realize the requirements of high efficiency, high precision and low cost overlay error measurement. BRIEF DESCRIPTION OF DRAWINGS

[0029] Fig. 1 It is a kind of overlay error measurement system based on angle resolved scatterometer schematic diagram;

[0030] Fig. 2 It is the light path diagram of angle resolved scatterometer measuring scattering light intensity at different angles;

[0031] Fig. 3This is a flowchart of the overlay error measurement method of the present invention.

[0032] The diagram is labeled as follows: 1. Light source; 2. Collimating lens; 3. Polarizer; 4. Aperture No. 1; 10. Aperture No. 2; 5. Relay lens group No. 1; 9. Relay lens group No. 2; 6. Beam splitter No. 1; 11. Beam splitter No. 2; 7. Microscope objective; 8. Sample to be tested; 12. Area array CCD camera No. 1; 14. Area array CCD camera No. 2; 13. Imaging lens; 15. Computer control and processing system. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Please see Figs. 1-3 The present invention provides a technical solution comprising a computer control and processing system 15, a light source 1, and a sample 88 to be tested. The axis of the light source 1 is perpendicular to the axis of the sample 88 to be tested. A collimating lens 2, a polarizer 3, a first aperture 4, a first relay lens group 5, and a first beam splitter 6 are arranged sequentially along the illumination direction of the light source 1.

[0035] The first CCD camera 12, the first beam splitter 6, the second beam splitter 11, the second aperture 10, the first relay lens group 9, the second relay lens group 9, the microscope objective 7, and the sample to be tested 8 are placed coaxially.

[0036] The light source is a super-continuous white light source, and the light emitted by the light source forms a parallel light beam through a collimating lens 2. After the polarization state of the light beam is regulated by a polarizer, a relay lens 5 images a first aperture stop 4 to the back focal plane of a microscope objective 7. The parallel incident light beam is sequentially irradiated to the surface of a sample 8 under test through a beam splitter prism 6 and the microscope objective 7, and the scattered light emitted by the sample 8 is collected by the microscope objective 7. The light beam is sequentially regulated by the beam splitter prism 6, a second relay lens group 9 and a second aperture stop 10, and the image of the back focal plane of the microscope objective 7 is imaged to the image plane of a first area array CCD camera 12. A second beam splitter prism 11 divides the emitted light containing the information of the sample 8 under test into two paths, which are received by the first area array CCD camera 12 directly and by a second area array CCD camera 14 through an imaging lens 13 respectively. The first area array CCD camera 12 corresponds to the main light path and collects the pupil image, and the second area array CCD camera 14 corresponds to the auxiliary light path and collects the spatial image. The two cameras are arranged to realize real-time regulation of the relative spatial position of the overlay mark under test and the probe spot. Based on the basic principle that the positive and negative first-order light intensity difference and the overlay error are approximately linear, the positive and negative first-order light intensity difference is collected to analyze the overlay error.

[0037] The aperture size and orientation of the aperture stop can be dynamically adjusted. The first aperture stop 4 is used to control the size and angle of the light spot irradiated to the sample, and the second aperture stop 10 is used to control the transmission of the specific order of diffraction light.

[0038] The relay lens group is composed of two groups of double-cemented spherical lenses. The focal length ratio of the first relay lens group 5 is used to change the scaling magnification of the incident light path, which is combined with the fiber core diameter, the collimating lens and the magnification of the microscope objective to control the size of the probe spot.

[0039] The beam splitter prism includes a first beam splitter prism 6 and a second beam splitter prism 11, and the splitting ratios of the first beam splitter prism 6 and the second beam splitter prism 11 are both 50:50.

[0040] The microscope objective 7 is a high numerical aperture microscope objective 7, which is used to collect the scattered light field in a large incident angle range.

[0041] The sample 8 under test is adjusted in position by an X-Y-Z displacement stage, and is ensured to be within the probe spot range and at the focal point position of the microscope objective 7 under the premise of being perpendicular to the optical axis.

[0042] An overlay error measurement method based on an angle-resolved scatterometer, comprising

[0043] S1, controlling two area array CCD cameras to collect images through a computer control and processing system 15;

[0044] S2, after the first area array CCD camera 12 and the second area array CCD camera 14 collect the pupil image and the spatial image, the images are fed back to the computer control and processing system 15.

[0045] S3, the computer control and processing system 15 establishes a system simulation model according to the measured scattering light information;

[0046] S4, after verifying the rationality of the simulation model, the overlay error is simulated and calculated.

[0047] It should be noted that in the present application Fig. 2 The light path diagram of measuring the scattering light intensity at different angles by the angularly resolved scatterometer. The main light path is used for collecting the scattering field signal, and the pupil image is obtained through the main light path. In the exit light path, by controlling the azimuth angle and spatial position of the diaphragm, only the specific order of exit diffraction light can pass through, and other orders of diffraction light are blocked. In this way, the scattering light intensity of a specific order can be measured, and the positive and negative first-order light intensity difference is calculated. Subsequently, the pupil image and the spatial dark field or bright field image are formed through the beam splitter.

[0048] It should be noted that in the present application Fig. 3 The flow of the overlay error measurement method in the present application is shown. As shown in the figure, first, an overlay error measurement system based on an angularly resolved scatterometer is designed to collect the positive and negative first-order light intensity difference information. Then, based on the basic principle that the positive and negative first-order light intensity difference and the overlay error are approximately linearly related within a certain range, the relationship between the positive and negative first-order light intensity difference and the overlay error is more accurately learned through a neural network. The specific steps include:

[0049] Make a data set, collect a certain number of data samples of the positive and negative first-order light intensity difference and the overlay error through experiments;

[0050] Data preprocessing, including data cleaning, removing outliers, normalization and other operations, to ensure the accuracy and consistency of the data;

[0051] Determine an initial neural network model for subsequent training and prediction. Optionally, use a convolutional neural network as the initial model. The appropriate network structure and parameters can be selected.

[0052] Collect the positive and negative first-order light intensity difference information; input the positive and negative first-order light intensity difference information collected by the measurement system as input, and predict the overlay error as output. Compare the predicted value with the actual overlay error, adjust the weights and biases of the network through the back propagation algorithm, so that the network can accurately predict the overlay error. Set the loss function and threshold, and optionally set the loss function as the mean square error Error between the actual overlay error and the predicted overlay error:

[0053]

[0054] Wherein, n represents the number of samples, y represents the actual overlay error value, and y_pred represents the overlay error value predicted by the network. By minimizing the mean square error, the difference between the predicted value and the actual value can be made as small as possible.

[0055] After the training is completed, the trained network model is evaluated. Optionally, the mean square error is used to measure the accuracy and stability of the model.

[0056] Finally, according to the positive and negative first-order light intensity difference measured by the actual system and the trained neural network model, the value of the overlay error is predicted.

[0057] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A system for overlay error measurement based on angle-resolved scatterometer, comprising a computer control and processing system (15), a light source (1) and a sample (8) to be measured, the axis of the light source (1) is perpendicular to the axis of the sample (8) to be measured, characterized in that: a collimating lens (2), a polarizer (3), a first diaphragm (4), a first relay lens group (5) and a first light splitting prism (6) are sequentially arranged along the direction of the light source (1), the first diaphragm (4) is placed at the conjugate position of the back focal plane of the microscope objective; the sample (8) to be measured is arranged coaxially with the first light splitting prism (6), and a microscope objective (7) is arranged between the first light splitting prism (6) and the sample (8) to be measured; a second relay lens group (9), a second diaphragm (10) and a second light splitting prism (11) are sequentially arranged away from the microscope objective (7) of the first light splitting prism (6) ; the second light splitting prism (11) splits the outgoing light containing sample information into two paths, which are received by a first area array CCD camera (12) directly and a second area array CCD camera (14) through an imaging lens (13) respectively; the first area array CCD camera (12) and the second area array CCD camera (14) are in communication connection with the computer control and processing system (15), wherein the first area array CCD camera (12) corresponds to the main light path and collects pupil plane images; the second area array CCD camera (14) corresponds to the auxiliary light path and collects spatial images, and the two cameras are arranged to realize real-time regulation and control of the relative spatial position of the overlay mark to be measured and the probe spot; the positive and negative first-order light intensity difference is collected for analyzing overlay error. The light source (1) is a super-continuous laser white light source. The polarizer (3) is located between the collimating lens (2) and the first diaphragm (4), and is used for polarization state regulation of the incident parallel light beam. The aperture size and orientation of the first diaphragm (4) and the second diaphragm (10) are dynamically adjustable, the first diaphragm (4) is used for controlling the size and angle of the light spot irradiated on the sample, and the second diaphragm (10) is used for controlling the transmission of specific order diffraction light. The first relay lens group (5) and the second relay lens group (9) are both composed of two groups of double-cemented spherical lenses, the focal length ratio of the first relay lens group (5) is used to change the scaling ratio of the incident light path, combined with the fiber core diameter, the collimating lens and the microscope objective magnification, to realize the control of the probe spot size.

2. The overlay measurement system based on an angle-resolved scatterometer according to claim 1, wherein: The light splitting ratio of the first light splitting prism (6) and the second light splitting prism (11) is both 50:

50.

3. The overlay measurement system based on an angle-resolved scatterometer according to claim 2, wherein: The microscope objective (7) is a high numerical aperture microscope objective, which is used for collecting scattered light field in a large incident angle range.

4. The overlay measurement system based on an angle-resolved scatterometer according to claim 3, wherein: The sample (8) to be measured is placed on a displacement table which can move in X-Y-Z three axes, and is ensured to be in the probe spot range and at the focal point position of the microscope objective under the premise of being perpendicular to the optical axis.

5. The overlay measurement system based on an angle-resolved scatterometer according to claim 4, wherein: The system comprises the following steps:

6. The overlay measurement system based on an angle-resolved scatterometer according to claim 5, wherein: S1, controlling two area array CCD cameras to collect images through the computer control and processing system (15) ; 7. The overlay measurement system based on an angle-resolved scatterometer according to claim 6, wherein: S2, after the first area array CCD camera (12) and the second area array CCD camera (14) collect pupil plane images and spatial images, feeding back to the computer control and processing system (15).

8. The overlay measurement system based on an angle-resolved scatterometer according to claim 7, wherein: ​ 9. The method of overlay error measurement based on an angle-resolved scatterometer according to claim 8, characterized in that: ​ ​ ​ S3, the computer control and processing system (15) establishes a system simulation model according to the measured scattering light information; S4, after verifying the rationality of the simulation model, the overlay error is simulated and calculated.

10. The method of overlay measurement based on an angle-resolved scatterometer according to claim 9, wherein: The system simulation model is based on the basic principle that the positive and negative first-order light intensity difference and the overlay error are approximately linearly related within a certain range, and the relationship between the positive and negative first-order light intensity difference and the overlay error is learned through neural network training. The specific operation steps of the model include: 1) making a data set, collecting a certain number of data samples of the positive and negative first-order light intensity difference and the overlay error through experiments; 2) data preprocessing, including data cleaning, removing outliers, normalization and other operations to ensure the accuracy and consistency of the data; 3) network training stage, input the preprocessed data samples into the neural network for training; 4) after the training is completed, the mean square error is used to evaluate the neural network model; 5) according to the actual system measurement of the positive and negative first-order light intensity difference and the trained neural network model, the value of the overlay error is predicted.

Images