JP2013200180A - Pattern shape measurement instrument and pattern shape measurement method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow measurement with high reliability and accuracy.SOLUTION: In a pattern shape measurement method, a plurality of shape models are created with respect to a measurement object pattern, a first optical condition for shape model determination and a second optical condition for shape measurement are determined, a shape model most suitable for shape measurement of the measurement object pattern is selected from among the plurality of shape models, and a shape of the measurement object pattern is calculated by fitting of second and third waveforms. The first and second optical conditions are determined in accordance with first waveforms obtained under a plurality of optical conditions with respect to the plurality of shape models. The most suitable shape model is selected using feature quantities of the first waveforms. The second waveforms are acquired correspondingly to dimensions of the selected shape model while successively changing the dimensions of the selected shape model. The third waveform is acquired by irradiating the measurement object pattern with light under the second optical condition.

Description

  Embodiments described herein relate generally to a pattern shape measuring apparatus and a pattern shape measuring method.

  As a shape measuring method in the manufacturing process of a semiconductor device, there is a technique called scatterometry. In this measurement method, for a measurement target pattern having a repetition period, its shape change is predicted by simulation to obtain a theoretical waveform in advance, and a signal waveform obtained by actually irradiating the measurement target pattern with light, This is a technique for obtaining a three-dimensional shape of a measurement target pattern by fitting a theoretical waveform. Since this method can perform cross-sectional measurement with non-destructive, high throughput, and high accuracy, it has attracted particular attention in recent years as a method for measuring shapes such as wiring width and height.

  Scatterometry requires the creation of a shape model before performing measurements. The shape model is created mainly by combining squares, trapezoids, etc. with reference to an actual image obtained from a cross-sectional SEM (Scanning Electron Microscope) or the like (see FIG. 2). Based on the model, a theoretical waveform is created by solving the optical behavior from the Maxwell equation. When measuring, gradually change the parameters such as dimensions and height (hereinafter referred to as “configuration parameters”) of each part of the shape model so that the matching rate between the measured waveform and the theoretical waveform is the highest. Then, each numerical value when the difference between the measured waveform and the theoretical waveform becomes the smallest is output as a measured value.

  On the other hand, there is a library method as a method for simplifying the fitting. The point that a theoretical waveform is required is the same as the method for individually creating the shape model described above, but the library method is different in that all theoretical waveforms assumed in advance are created and held as a library. Then, by extracting an optimal waveform from the library at the time of measurement, the simulation time required for fitting can be shortened.

  Thus, the first shape model creation is important in scatterometry. In the manufacturing process of a semiconductor device, there are many changes in a manufacturing apparatus such as a film forming apparatus and an etching apparatus, and changes in manufacturing conditions, and a change in pattern shape can naturally occur accordingly. In particular, there are many problems in which the measurement target part is affected by the shape change of a part different from that and gives an incorrect numerical value. If this shape change is within the range of the scatterometry shape model, it can be measured without problems, but if it changes more than expected, the shape model will not be able to follow and an incorrect value will be output. It will be Shimadzu.

  As a countermeasure, it is necessary to reassemble the shape model based on the cross-sectional SEM image or the like.

  However, in this case, there are problems that the number of variable parameters increases, and there are a plurality of solution candidates, and an incorrect numerical value is output, which may cause a decrease in reproducibility.

JP 2003-344029 A

  The problem to be solved by the present invention is a pattern shape measuring apparatus and pattern shape measuring device that enable measurement with high reliability and accuracy even when the shape of the pattern changes more than expected due to a change in the manufacturing process. Is to provide a method.

  The pattern shape measurement method of the embodiment includes a step of creating a plurality of shape models for a measurement target pattern having a repetition period, a first optical condition used for shape model discrimination, and a second optical condition used for shape measurement. Determining the optimum shape model for shape measurement of the measurement target pattern from the plurality of shape models, obtaining the second and third waveforms, and the second and third waveforms And calculating the shape of the measurement target pattern by fitting. The first and second optical conditions are determined from first waveforms obtained under a plurality of optical conditions for the plurality of shape models. The optimum shape model is selected using the feature quantity of the first waveform. The second waveform is acquired corresponding to each dimension while sequentially changing the dimension of the selected shape model. The third waveform is acquired by irradiating the measurement target pattern with light under the second optical condition.

1 is a block diagram showing a schematic configuration of a shape measuring apparatus according to Embodiment 1. FIG. The figure explaining the creation method of a shape model. The figure which shows an example of the shape model assumed before process variation, and an approximate shape. The figure which shows an example of the shape model assumed after process fluctuation | variation, and an approximate shape. FIG. 3 is a diagram illustrating an example of a shape model used in the pattern shape measurement method according to the first embodiment. FIG. 6 is a diagram illustrating another example of a shape model used in the pattern shape measurement method according to the first embodiment. 5 is a flowchart showing a basic flow of a pattern shape measuring method according to the first embodiment. The flowchart which shows the specific procedure of the determination method of optical conditions. The flowchart which shows the specific procedure of shape measurement. The figure which shows an example which prescribed | regulated the dimension to the approximate shape shown in FIG. The figure which shows an example which prescribed | regulated the dimension to the approximate shape shown in FIG. The graph which shows an example of the theoretical waveform obtained by injecting light into a perpendicular direction about the shape model shown in FIG. 5 and FIG. The graph which shows an example of the theoretical waveform obtained by injecting light into a perpendicular direction about the shape model shown in FIG. 5 and FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The graph which shows an example of the theoretical waveform obtained by injecting light into a horizontal direction about the shape model shown in FIG. 5 and FIG. The graph which shows an example of the theoretical waveform obtained by injecting light into a horizontal direction about the shape model shown in FIG. 5 and FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG. The figure which shows an example of a theoretical waveform at the time of changing a desired measurement dimension about the shape model shown in FIG.

  Hereinafter, some embodiments will be described with reference to the drawings. In the drawings, the same parts are denoted by the same reference numerals, and repeated description thereof is omitted as appropriate. Further, in the following embodiments, a case where a fine pattern created in a semiconductor device manufacturing process such as a lithography process or an etching process is measured will be described, but the present invention is not limited to this case. However, it should be noted that the present invention is applicable to general pattern evaluation in various other industrial fields such as a liquid crystal panel manufacturing process. In the present specification, the “pattern” is not limited to what constitutes an actual product, but includes a resist pattern that is removed after being created in the manufacturing process.

(A) Embodiment of Pattern Shape Measuring Device FIG. 1 is a block diagram showing a schematic configuration of a shape measuring device according to the first embodiment. The shape measuring apparatus shown in the figure is a spectroscopic ellipsometry type scatterometry apparatus, and includes a light source 1, a polarizer 2, a stage S, an analyzer 4, a spectrometer 5, a detector 6, and a detection. A signal processing unit 7, an optical system control unit 8, a control computer 9, a shape calculation unit 10, an optical condition determination unit 11, and memories MR1 to MR3 are provided.

  The light source 1 emits white light. The stage S moves the wafer 3 by translational motion and rotational motion (φ direction) performed by an actuator (not shown) to which a control signal is given from the optical system control unit 8. A pattern TP as a measurement target obtained by actually creating an arbitrary pattern on the wafer 3 is created on the surface of the wafer 3.

  The pattern TP shown in FIG. 1 is an example of a rectangular pattern arranged at a predetermined pitch in two directions orthogonal to each other. However, the pattern TP is not limited to such a pattern. An and space pattern, a hole pattern arranged at a predetermined pitch in one direction or two directions orthogonal to each other, or a pattern in which at least one of a hole pattern, a rectangular pattern, and a line pattern is mixed is also used for shape measurement in this embodiment. Can be a target.

  White light emitted from the light source 1 enters the pattern TP via the polarizer 2 at an incident angle of angle θ. The light reflected by the pattern TP is split by the spectroscope 5 through the analyzer 4 and detected by the detector 6. The detector 6 outputs the signal of the actual spectral waveform of the pattern TP and gives it to the detection signal processing unit 7. The detection signal processing unit 7 generates measurement waveform data from the output signal of the detector 6 and stores it in the memory MR2.

  In the present embodiment, the light source 1, the polarizer 2, the stage S, the analyzer 4, the spectrometer 5, the detector 6, the detection signal processing unit 7, and the optical system control unit 8 correspond to, for example, a measurement waveform acquisition unit.

  The optical condition determination unit 11 extracts data on a plurality of shape models related to the inspection target pattern TP from the memory MR3, acquires a waveform for each shape model under a plurality of optical conditions by simulation, and determines the shape model from the obtained waveform. And the optical condition for shape measurement are determined and given to the control computer 9 together with the acquired waveform.

  The computer 9 selects an optimal shape model using the optical condition data given from the optical condition determination unit 11 and the waveform feature amount used in determining the optical condition, and stores each shape model stored in the memory MR1. The optimum shape model recipe is extracted from the recipe for each shape model and is given to the shape calculation unit 10. In the present embodiment, the control computer 9 corresponds to, for example, a shape model selection unit.

  The shape calculation unit 10 is provided with a measurement waveform of the inspection target pattern TP from the detection signal processing unit 7 via the control computer 9, and the theoretical waveform for each parameter by simulation while sequentially changing the configuration parameters of the optimum shape model. Alternatively, a library that is a collection of theoretical waveforms is acquired, fitting with measurement waveforms, the shape of the inspection target pattern TP is calculated, and the calculation result is output.

  Hereinafter, a pattern shape measuring method using the pattern shape measuring apparatus shown in FIG. 1 will be described more specifically with reference to the drawings.

(1) Preparation of theoretical waveform or library When measuring with scatterometry, it is necessary to create a theoretical waveform in advance. In order to create this theoretical waveform, first, it is necessary to assemble a shape model based on an image (hereinafter referred to as “real image”) obtained by capturing an actually created measurement target pattern TP. For example, as shown in FIG. 2, the shape model is created mainly by combining simple shapes such as a square and a trapezoid based on an actual image. Then, a theoretical waveform is created by solving the optical behavior when the shape model is irradiated with light from the Maxwell equation. Calculation of synthetic seismograms is susceptible in RCWA method (R igorous C oupled W ave A nalysis) or the FDTD method (F inite D ifference T ime D omain), the periodicity scatterometry required pattern, computed In many cases, RCWA with a short time is used.

(2) Preparation of Shape Model In the following, shape measurement when the process apparatus is changed during the manufacturing process and the pattern shape changes more than expected due to this change will be described.

  Prior to the change of the process apparatus, a shape model as shown in FIG. 2 was assumed, for example. However, if there is a possibility that the shape as shown in FIG. 3 or FIG. It is necessary to measure the shape including the model. In this case, the shape of FIG. 3 can be measured by performing matching by changing the size of the lower side of the trapezoid 2 in the shape model of FIG. However, in the case of FIG. 4, matching is impossible. When matching the shape model of FIG. 4 as well, it is necessary to model with trapezoids 1 to 3 and square 1. As described above, the measurement itself can be performed by reassembling the model, but it is a very laborious work.

  Therefore, there is a method of predicting the change in advance, making the shape fine, and building a shape model so as to follow any change. For example, if the shape shown in FIG. 2 is predicted in advance and the shape shown in FIG. 4 is predicted in advance and the model is built, the shapes shown in FIGS. 3 and 4 can also be measured. There is a high possibility that an incorrect numerical value is output because there are a plurality of candidates.

  In the present embodiment, for example, the shape model A was originally created assuming the shape of the shape model A of FIG. 5, but the shape of the shape model B shown in FIG. Take up the process.

  First, dimensions are defined for the approximate shape CFA of FIG. 5 as shown in FIG. Also, the dimensions are defined as shown in FIG. 11 for the approximate shape CFB of FIG. Here, in the process of this example, a dimension for which process management is to be performed is CD2. In the present embodiment, these shape models A and B and approximate shapes CFA and CFB and their dimensions are created and defined in advance and stored in the memory MR2.

(3) Determination of optical conditions The optical condition determination unit 11 takes in data relating to the shape models A and B from the memory MR2 and creates and analyzes a theoretical waveform using simulation, thereby determining the optical conditions for determining the shape model by the following method. Then, the optical conditions for shape measurement are determined.

  In the present embodiment, a recipe determination waveform is used as a determination criterion as to which shape model should be employed for shape measurement. Then, it is necessary to select under what optical conditions it is appropriate to acquire the recipe discrimination waveform. The recipe discrimination waveform will be described in detail later.

  In this embodiment, the incident direction of the light Li with respect to the inspection target pattern TP and the wavelength thereof are listed as optical conditions. Depending on the process apparatus, it is necessary to consider the incident polarization angle, the imaging polarization angle, the angle with respect to the wiring, the incident angle in the plane direction, and the like.

  The incident direction of the light Li can be an arbitrary angle with respect to the inspection target pattern TP, but here, in order to simplify the description, 90 ° (vertical) and 0 ° (parallel) with respect to the inspection target pattern TP. ) Was simulated.

  First, a theoretical waveform in the shape model A (see FIG. 5) when the incident direction of the light Li is perpendicularly incident on the inspection target pattern TP is obtained by simulation. Similarly, the theoretical waveform of the shape model B (see FIG. 6) is obtained by simulation. These theoretical waveforms are plotted on a graph. 12A and 12B are graphs of these theoretical waveforms. Since the scatterometry apparatus of this embodiment is of a spectroscopic ellipsometry type, two theoretical waveforms are obtained: a graph showing the intensity ratio with respect to the wavelength (FIG. 12A) and a graph showing the phase difference with respect to the wavelength (FIG. 12B). 12A and 12B, when the theoretical waveforms of the shape model A and the shape model B are compared with each other, there is no difference in the theoretical waveform at a wavelength of 250 nm to about 280 nm. It can be seen that the theoretical waveforms are different at wavelengths of 280 nm or more.

  Next, with respect to the desired configuration parameter, in this embodiment, CD2 (see FIGS. 10 and 11), a theoretical waveform at the time of dimensional change is created and compared. In the shape model A, FIGS. 13A and 13B show examples of simulation results when CD2 in FIG. 10 is changed by ± 5 nm from the reference value. FIG. 13A is a graph showing the intensity ratio with respect to the wavelength, and FIG. 13B is a graph showing the phase difference with respect to the wavelength. Similarly, for the shape model B, FIGS. 14A and 14B show examples of simulation results when CD2 is changed by ± 5 nm from the reference value. FIG. 14A is a graph showing the intensity ratio with respect to the wavelength, and FIG. 14B is a graph showing the phase difference with respect to the wavelength.

  From FIG. 13A to FIG. 14B, it can be seen that there is a difference in the theoretical waveform in the entire wavelength region of 250 nm to 750 nm, and there is no wavelength region in which the difference is so small that it cannot be visually confirmed between the theoretical waveforms. From this, it is understood that it is difficult to determine which of the shape model A and the shape model B should be selected in the incident direction perpendicular to the inspection target pattern TP. On the other hand, it can be said that the measurement sensitivity is high because the theoretical waveform greatly differs with respect to the CD2 dimensional change.

  Next, the theoretical waveforms of the shape models A and B were simulated when the light incident direction was incident parallel to the wiring. The results are shown in FIGS. 15A and 15B. FIG. 15A is a graph showing the intensity ratio with respect to the wavelength, and FIG. 15B is a graph showing the phase difference with respect to the wavelength. From FIG. 15A and FIG. 15B, it can be seen that the theoretical waveform of the phase difference has a larger difference between the shape models than the waveform of the intensity ratio. 16A and 16B show simulation results for the shape model A when CD2 is changed by ± 5 nm from the reference value. Similarly, for the shape model B, the simulation results are shown in FIGS. 17A and 17B when CD2 is changed by ± 5 nm from the reference value. 16A and 17A are graphs showing the intensity ratio with respect to the wavelength, and FIGS. 16B and 17B are graphs showing the phase difference with respect to the wavelength.

  Considering from FIG. 15A to FIG. 16B, while it is insensitive to the dimensional variation of CD2, there is a difference in waveform over almost all wavelength regions between the shape model A and the shape model B, that is, there is a difference in the theoretical waveform. What is not seen is a waveform in a wavelength region of 250 nm to 300 nm among the waveforms of the phase differences with respect to the wavelengths of FIGS. 16B and 17B.

  From this, it is possible to determine which of the shape model A and the shape model B is appropriate for shape measurement from the peak intensity in the wavelength range of 250 nm to 300 nm in the phase difference waveform shown in FIG. 15B. However, it can be said that the measurement sensitivity is low because the change in the waveform corresponding to the change in the dimension of CD2 (FIGS. 16A to 17B) is smaller than that in the case where it is perpendicularly incident on the wiring (FIGS. 13A to 14B).

  From the above analysis, the phase difference waveform (FIG. 15B) obtained by entering parallel to the inspection target pattern TP is used as the recipe determination reference waveform, and the measurement waveform is the inspection target pattern TP. It was found that the intensity ratio / phase difference waveforms (FIGS. 13A to 14B) obtained by perpendicular incidence should be used. In this embodiment, the theoretical waveforms shown in FIGS. 12A to 17B correspond to, for example, the first waveform, and incident light parallel to the inspection target pattern TP corresponds to, for example, the first optical condition, Incident light perpendicular to the target pattern TP corresponds to, for example, the second optical condition.

(4) Selection of Shape Model The control computer 9 is supplied with optical condition data for shape model determination and optical condition data for shape measurement from the optical condition determination unit 11, and generates a control signal to generate an optical system. This is given to the control unit 8. The optical system control unit 8 controls the optical system of the shape measuring apparatus based on a control signal from the control computer 9, outputs a recipe determination signal and a measurement signal, and supplies them to the control computer 9.

  More specifically, first, light Li is incident on the inspection target pattern TP vertically from the light source 1, and a measurement waveform of the reflected light Lo is acquired. Next, the stage S is rotated by 90 ° to an actuator (not shown), and light Li is incident on the inspection target pattern TP from the light source 1 in parallel, and a measurement waveform (measurement waveform for recipe determination) of the reflected light Lo is acquired. The control computer 9 creates a waveform of a phase difference with respect to the wavelength for the actual recipe determination waveform, and selects a shape model recipe closer to the actual recipe determination waveform from the recipe determination reference waveform.

  In the present embodiment, the peak intensity in the wavelength region of 250 nm to 300 nm is calculated using the actually measured waveform for recipe discrimination. If the peak intensity is 0.5 or more, the shape model A is adopted, and if it is less than 0.5, the shape model B is adopted. When the adopted model is determined in this way, the corresponding recipe is extracted from the memory MR2 and given to the shape calculation unit 10 together with the measurement waveform data. In the present embodiment, the measurement waveform corresponds to, for example, the third waveform, and the recipe determination actual measurement waveform corresponds to, for example, the fourth waveform.

(5) Fitting and shape measurement The shape calculation unit 10 performs simulation for each parameter by simulation while sequentially changing target parameters among the configuration parameters (such as trapezoid shape and angle) of the shape model selected by the control computer 9. Obtain the theoretical waveform and perform fitting with the measurement waveform. The theoretical waveform for each parameter corresponds to, for example, the second waveform in this embodiment.

  The fitting of the measurement waveform and the theoretical waveform starts from the point where the preset initial value and the measurement data are overlapped. Normally, a difference occurs between the two from the start stage, but the configuration parameters of the shape model of the theoretical waveform are changed, and the difference between the two is reduced as compared with the measured waveform as needed. Finally, the numerical value of each configuration parameter when the difference becomes the smallest is output. In the above-described example, fitting is performed while changing the value of CD2, the shape of the inspection target pattern TP is calculated, and the calculation result is output.

  According to the shape measurement apparatus of the first embodiment described above, optical condition determination for determining an optical condition for shape model determination and an optical condition for shape measurement from waveforms obtained under a plurality of optical conditions for a plurality of shape models. Unit 11 and a control computer 9 that selects an optimal shape model for measuring the shape of the pattern to be measured from the plurality of shape models using the obtained waveform feature amount. Even when the pattern shape changes, the pattern shape can be measured with high reliability and accuracy.

(B) Embodiment of Pattern Shape Measuring Method FIG. 7 shows a basic flow of the pattern shape measuring method of this embodiment. First, a plurality of shape models are created by predicting shape changes caused by the manufacturing process, optical conditions for selecting an optimum shape model are determined (step S10), and then the optimum optical conditions are obtained. Shape measurement is performed by selecting the shape model and operating the corresponding measurement recipe (step S20).
FIG. 8 is a flowchart showing a more specific procedure for determining optical conditions (FIG. 7, step S10).
First, theoretical waveforms of each shape model, in this embodiment, shape model A and shape model B are created under optical condition 1 (steps S11 and S12), and the respective theoretical waveforms are compared (step S13). These theoretical waveforms may be created sequentially or simultaneously.

  Next, theoretical waveforms of the shape model A and the shape model B are newly created under the optical condition 2 (steps S14 and S15), and the respective theoretical waveforms are compared (step S16). As in the case of the first optical condition, these theoretical waveforms may be created sequentially or simultaneously.

  Next, the theoretical waveforms of the shape model A and the shape model B under the optical condition 1 and the theoretical waveforms of the shape model A and the shape model B under the optical condition 2 are mutually compared and examined (step S17). An optical condition (recipe discrimination optical condition) suitable for the recipe discrimination waveform is selected (step S18), and an optical condition (shape measurement optical condition) suitable for shape measurement is further selected (step S19). The theoretical waveforms obtained in steps S11, S12, S14 and S15 correspond to, for example, the first waveform in this embodiment.

  FIG. 9 is a flowchart showing a specific procedure of the method of operating the measurement recipe (FIG. 7, step S20).

  First, for the measurement target pattern, the scatterometry apparatus is operated under the shape measurement optical conditions selected in step S19 in FIG. 8 to obtain a measurement waveform. In the example described above, the waveform of the phase difference when the light Li is incident perpendicularly to the inspection target pattern TP is acquired. This phase difference waveform corresponds to, for example, a third waveform in the present embodiment.

  Next, the scatterometry apparatus is operated under the optical conditions for recipe determination selected in step S18 of FIG. 8, and an actual measurement waveform for recipe determination is acquired (step S22). In the above-described example, the waveform of the phase difference when the light L is incident in parallel to the inspection target pattern TP is acquired. The obtained phase difference waveform corresponds to, for example, the fourth waveform in the present embodiment.

  Next, an optimal shape model is selected using the feature quantity of the actual waveform for recipe discrimination, and a measurement recipe corresponding to this is acquired (step S23). In the example described above, the peak intensity in the wavelength region of 250 nm to 300 nm is calculated, and a shape model having a peak intensity close to the peak intensity is selected. Thereby, the shape model used for fitting is selected. When two shape models A and B are prepared as in the above-described example, a recipe for one of the shape models is selected (step S24).

  Then, using the selected recipe, the corresponding theoretical waveform is acquired while sequentially changing the configuration parameters of the shape model, and fitting is performed by sequentially comparing with the measured waveform (steps S25 to S28). And the numerical value of each constituent parameter when the difference between the measured waveform and the measured waveform is the smallest is output as the measurement result of the inspection target pattern TP (step S29). The theoretical waveform used for fitting corresponds to, for example, the second waveform in the present embodiment.

  According to the shape measurement apparatus of the first embodiment described above, an optical condition for shape model determination and an optical condition for shape measurement are determined from waveforms obtained under a plurality of optical conditions for a plurality of shape models. This is the case when the shape of the pattern changes more than expected due to a change in the manufacturing process because the optimum shape model recipe for measuring the shape of the pattern to be measured is selected from the plurality of shape models using the waveform feature values. However, it is possible to measure the pattern shape with high reliability and accuracy.

  In the above embodiment, the method of deriving a waveform using simulation when selecting each optical condition has been described. However, the present invention is not limited to this, and for example, each optical condition can also be selected from a waveform acquired from an actual wafer. Specifically, wafers having cross-sections having different shapes are prepared, and the waveforms are acquired under various optical conditions. In this case, as described in the simulation, a wafer in which items to be measured are changed is also necessary. The procedure after the stage of acquiring the waveform from the actual wafer is the same as the method using the simulation.

  In the above description, the spectroscopic ellipsometry type scatterometry is taken as an example. However, the present invention is not limited to this, and any optical system that can change the optical conditions such as the incident direction and the incident angle is used. It is also possible to use a scatterometry of a type or a single wavelength angle variable type, or a combination of these types.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and the equivalents thereof.

  DESCRIPTION OF SYMBOLS 1 ... Light source, 2 ... Polarizer, 4 ... Analyzer, 5 ... Spectroscope, 6 ... Detector, 7 ... Detection signal processing part, 8 ... Optical system control part, 9 ... Control computer, 10 ... Shape calculation part, 11 ... Optical condition determination unit, MR1 to MR3 ... Memory (storage device), S ... Stage, TP ... Pattern to be inspected

Claims (6)

For a measurement target pattern having a repetition cycle, a step of creating a plurality of shape models by predicting a shape change caused by the manufacturing process;
A first waveform obtained for a plurality of shape models under a plurality of optical conditions is acquired, a first optical condition suitable for shape model discrimination from the first waveform, and a second optical suitable for shape measurement Determining the conditions;
Selecting an optimum shape model for shape measurement of the measurement target pattern from the plurality of shape models using the feature quantity of the first waveform;
Acquiring second waveforms corresponding to the respective dimensions while sequentially changing the dimensions of the selected shape model;
Irradiating the measurement target pattern with light under the second optical condition to obtain a third waveform;
Irradiating the measurement target pattern with light under the first optical condition to obtain a fourth waveform;
Calculating the shape of the measurement target pattern by fitting the third waveform and the second waveform;
With
The first optical condition is an optical condition in which a difference in the first waveform occurs between the plurality of shape models over the entire predetermined wavelength band.
The second optical condition is an optical condition in which a difference in the first waveform occurs for each shape model over the entire predetermined wavelength band when the dimensions of each shape model are changed.
The feature amount is a peak value of the first waveform in an arbitrary wavelength band,
The optimal shape model is selected by comparing the peak value of the fourth waveform in the arbitrary wavelength band with the peak value of the first waveform.
Pattern shape measurement method. A step of creating a plurality of shape models for a measurement target pattern having a repetition period;
A first waveform obtained under a plurality of optical conditions for the plurality of shape models is acquired, a first optical condition used for shape model discrimination from the first waveform, and a second optical used for shape measurement Determining the conditions;
Selecting an optimum shape model for shape measurement of the measurement target pattern from the plurality of shape models using the feature quantity of the first waveform;
Acquiring second waveforms corresponding to the respective dimensions while sequentially changing the dimensions of the selected shape model;
Irradiating the measurement target pattern with light under the second optical condition to obtain a third waveform;
Calculating the shape of the measurement target pattern by fitting the third waveform and the second waveform;
A pattern shape measuring method comprising: Irradiating the measurement target pattern with light under the first optical condition to obtain a fourth waveform;
The feature amount is a peak value of the first waveform in an arbitrary wavelength band,
The pattern according to claim 2, wherein the optimum shape model is selected by comparing a peak value of the fourth waveform in the arbitrary wavelength band with a peak value of the first waveform. Shape measurement method.   4. The optical condition according to claim 2, wherein the first optical condition is an optical condition in which a difference in the first waveform occurs between the plurality of shape models over a predetermined wavelength band. 5. Pattern shape measurement method.   The second optical condition is an optical condition in which a difference in the first waveform is generated for each shape model over the entire predetermined wavelength band when the dimensions of each shape model are changed. The pattern shape measuring method according to any one of claims 2 to 4. A measurement waveform acquisition means for acquiring a measurement waveform by irradiating light to a measurement target pattern having a repetition period;
A first waveform obtained under a plurality of optical conditions is obtained for a plurality of shape models obtained for the measurement target pattern, a first optical condition used for shape model discrimination from the first waveform, and shape measurement Optical condition determining means for determining a second optical condition used for
Shape model selection means for selecting an optimum shape model for shape measurement of the measurement target pattern from the plurality of shape models using the feature quantity of the first waveform;
A second waveform corresponding to each dimension is acquired while sequentially changing the dimension of the selected shape model, and a third waveform acquired by irradiating the measurement target pattern with light under the second optical condition. A shape calculating means for calculating the shape of the measurement target pattern by fitting a waveform and the second waveform;
A pattern shape measuring apparatus.

Images