JP4834706B2 - Structural inspection method - Google Patents

Description

  The present invention relates to a structure inspection method for obtaining a structure of an object to be evaluated from reflected light intensity of the object to be evaluated.

  Many methods have been reported for evaluating the size and shape from the measurement results of diffracted light obtained by making light incident on a pattern having a regular structure on the substrate to be processed and diffracting it. For example, in Patent Document 1, an ellipsometer is used to measure cos Δ and tan ψ of a pattern having a regular structure, and the dimensions are determined by comparison with cos Δ and tan ψ obtained from theoretical electromagnetic wave calculation. This method was used for the evaluation of the resist pattern after development and the pattern after etching because theoretical calculation was performed using the structure and optical constants of the film constituting the substrate to be processed. If the film is to be applied to the dimension evaluation of a resist pattern formed on a substrate to be processed, a mixed phase of a resist reaction product and a developer exists in the vicinity of the resist pattern. Could not do. Also, in Patent Document 2, as in Patent Document 1, since a mixed phase of a resist and a developer near the resist pattern is not assumed, it is difficult to accurately evaluate the pattern.

In Patent Document 1, the theoretical wavelength dispersion of the diffracted light intensity is obtained by electromagnetic wave calculation using the structure and optical constant of the film constituting the substrate to be processed, and the wavelength of the diffracted light on the actual substrate is calculated. The size and shape are calculated by comparing the measurement result of dispersion with theoretical wavelength dispersion. In this patent, it is described that the composition distribution in the depth direction of the film due to doping is taken into account, but the characteristics of the distribution are not described.
US Pat. No. 5,963,329 Japanese Patent Application No. 2000-200121

  The conventional method of evaluating the size and shape from the measurement result of the diffracted light obtained by diffracting light incident on a pattern having a regular structure on the substrate to be processed can accurately evaluate the pattern. There wasn't.

  An object of the present invention is to more accurately evaluate the size and shape of a pattern from the measurement result of diffracted light obtained by irradiating light to a pattern having a regular structure on the substrate to be processed and diffracting it. An object of the present invention is to provide a structural inspection method that can be used.

  In order to solve the above problems, the present invention adopts the following configuration.

That is, one embodiment of the present invention is an optical inspection method for a substrate on which a layer having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch is stacked. A step of preparing a substrate in which a plurality of layers having a structure characterized by a ratio and a pitch are laminated, one or more layers having a determined structure, and one or more layers having a non-determined structure; For each of the layers, a plurality of predicted values can be predicted by creating a structure within a range twice as large as the variation value of each structure expected in the process of forming each layer, with the target value of the structure of each layer as a central value. A plurality of substrate structures including the above structure, a step of creating a substrate structure library A composed of a plurality of substrate structures obtained by combining the structures of the created layers, and each substrate included in the substrate structure library A Against concrete, the case where the light enters at a certain angle, the diffraction or the light intensity obtained by reflection calculated, wherein the calculated light intensity library wavelength dispersion housed in light intensity for each substrate structure A step of creating B, a step of newly creating a plurality of structures including the structure of the layer whose structure is determined, and a substrate including a newly created structure from among the substrate structures of the light intensity library B Extracting a structure , creating a light intensity library C in which the extracted substrate structure and wavelength dispersion of light intensity corresponding to the extracted substrate structure are stored, and light at a specific angle on the prepared substrate the incident, by comparing the step of detecting the chromatic dispersion of the optical intensity obtained by diffraction or reflection wavelength dispersion of the detected light intensity and wavelength dispersion of the light intensity that is contained in the light intensity library C, The substrate Characterized in that it comprises a step of determining the structure.

  According to the present invention, it is possible to more accurately evaluate the size and shape of a pattern from the measurement result of diffracted light obtained by making light incident on a pattern having a regular structure on the processing substrate and diffracting it. .

  The details of the present invention will be described below with reference to the illustrated embodiments.

(First embodiment)
In this embodiment, an antireflection film is formed on a Si substrate, and an ArF chemically amplified resist is formed thereon with a film thickness of 300 nm. An ArF stepper is used to form an underlying film processing pattern on the resist film. After the exposure and then the heating, the development is further performed to form a resist pattern having a resist remaining size of 115 nm at a pitch of 260 nm (ratio of resist pattern to pitch = 0.442: 0.558 in the developer ratio). An application example is shown.

  FIG. 1 is a flowchart showing a structure determining method for an object to be evaluated according to the first embodiment of the present invention.

  In order to calculate the profile of the resist film being developed, a certain model is established (step S101). A reaction product of the resist and the developer is generated from the resist pattern being developed. Hereinafter, the developer containing the reaction product of the resist is referred to as a mixed phase, and the ratio of the resist in the mixed phase is defined as the mixing ratio. In general, the diffusion of the reaction product is not fast enough, and unless the method of diffusion of the reaction product, that is, the distribution of the mixing ratio in the liquid phase is taken as a model, the accuracy of pattern evaluation during development is poor. turn into. Therefore, in this embodiment, the development model is determined in the library creation process.

  FIG. 2 is a schematic diagram showing how the development proceeds. At the initial stage of development, as shown in FIG. 2, development proceeds in the film thickness direction in the exposed resist film 103. As a result, since the reaction product diffuses in the direction indicated by the arrow, the mixing ratio becomes a distribution in the film thickness direction, and the ratio decreases as the distance from the substrate increases. When development in the film thickness direction is completed, development proceeds in the horizontal direction. Therefore, a model is assumed in which the proportion of the developer in the space increases in the direction of the arrow. In FIG. 2, reference numeral 101 denotes a silicon substrate, reference numeral 102 denotes an antireflection film, reference numeral 103 denotes an unexposed resist film, and reference numeral 105 denotes a developer.

  Next, specific steps until the structure / optical constant of the substrate are given will be described. FIG. 3 shows a layer structure in the film thickness direction created based on the above-described model. Layers D2 to D1 are a mixed phase of a developer and resist reaction product, and layers R1 to R4 are a mixed phase of a developer and resist reaction product and a resist pattern. The difference between the layer D1 and the layer D2 is a difference in the mixing ratio of the reaction products. Similarly, the difference among layer R1, layer R2, layer R3, and layer R4 is the difference in the mixing ratio of reaction products. The reaction product is generated at the interface between the resist and the developer and diffuses into the developer layer. Therefore, the ratio of the reaction product is high at the bottom of the pattern.

First, a plurality of virtual component ratios assuming the ratio of each substance constituting the object to be evaluated and each measurement environment are set. Change the proportion of each substance in the space.

  Next, a database of reflectance chromatic dispersion calculation values is created based on the conditions shown in (Table 1) (step S102). The database changes the ratio of each substance in the space under the conditions shown in (Table 1), and determines an average optical constant corresponding to the ratio. Furthermore, the reflectance is obtained by performing multiple interference calculation. The wavelength dispersion of the reflectance was calculated in the range of 300 to 800 nm. Here, as a calculation method, Morham (J. Opt. Soc. Am., Vol. 12, No. 12) is used to solve the electric field, magnetic field, and intensity of diffracted light from a regular pattern using Maxwell's equations. RCWA (Rigorous coupled-wave analysis) method of 5, May 1995 1077-1086) is generally used. The optical constant of the reaction product formed by the reaction between the resist and the developer is the same as the optical constant of the resist.

The average complex refractive index is, for example, when the developer / space = 0.95 in the layer D2 of the developer layer, the complex refractive index of the developer (refractive index n, extinction coefficient k) = (1.33, 0) and the complex refractive index of the resist (1.67, 0.06) were obtained as shown in Equation (1) and Equation (2).

  After the substrate that has been subjected to the pattern exposure to the resist film and the post-exposure heating (PEB) is cooled, the substrate is moved to a developing device. A developer is supplied to the main surface of the substrate to form a developer film, and development of the resist film is started (step S103). The resist film has a device pattern area where the device pattern is exposed and a monitor pattern area where the monitor pattern is exposed.

  Next, reflectance measurement means is disposed above the portion where the pattern for dimension measurement is exposed, and the pattern under development is changed through observation of the reflectance change (step S104).

  FIG. 4 shows an outline of the reflectance measuring means. The reflectance measurement unit reflects light from the light source 211 by a mirror inside the observation optical system 212 and irradiates the monitor pattern region 203 of the resist film 201. The zero-order diffracted light reflected by the monitor pattern region 203 and generated by diffraction is collected by the observation optical system 212, and this light is transmitted to the spectroscope 213 via a quartz optical fiber to measure chromatic dispersion. ing. The chromatic dispersion measurement values detected by the spectroscope 213 are sequentially sent to the computer 214. Note that the monitor pattern region 203 is formed outside the device pattern region 202. The database 215 stores the reflectance chromatic dispersion calculated value calculated in step S102.

The computer 214 compares the measured chromatic dispersion value of the reflectance sent from the spectroscope 213 with the calculated chromatic dispersion value stored in the database 215, and calculates a similar reflectance dispersion calculated value having substantially the same dispersion value as the measured chromatic dispersion value. Are extracted (step S105). Wherein the extraction first calculates the square sum si reflectance calculated value r _{[lambda] c} at a certain measured reflectance values at a wavelength lambda r _lambda and the wavelength lambda.

  And it searches sequentially from the reflectance calculation value with which this sum of squares becomes the minimum. In the present embodiment, a reflectance calculated value having a square sum of 110% or less is searched for a reflectance calculated value having a minimum square sum.

  A spatial ratio prediction value rpi is obtained by taking a weighted average corresponding to the sum of squares with respect to the spatial ratio which is the basis of these calculated reflectance values (step S106). The predicted spatial ratio value rpi is obtained by the following equation (4) from the spatial ratio ri from which the reflectance is calculated and the square sum si at that time.

rpi = (Σri / si) / (Σ (1 / si)) (4)
Using the equation (4), spatial ratio prediction values are obtained from the layer R1 to the layer R4D2, respectively. The spatial space ratio predicted value rpi is regarded as a pattern size ratio in the pitch.

It is determined whether the predicted spatial ratio value rp ₄ obtained by the weighted average is equal to a predetermined value (step S107). In the present embodiment, the predetermined value is 56%. In this embodiment, a resist pattern having a resist remaining size of 115 nm with a pitch of 260 nm (ratio of resist pattern to pitch = 0.442: 0.558 in the developer ratio) is formed. Therefore, it is determined whether the spatial ratio predicted value rp4 of the layer R4 is equal to 56%.

Equal to space ratio predictive value rp ₄ 56%, to stop the development by supplying a cleaning liquid to the substrate surface (step S108).

If the predicted spatial ratio rp ₄ is not equal to 56%, development is continued and steps S103 to S107 are repeated. When the predicted space ratio value rp4 is equal to 56%, the cleaning liquid is supplied to the substrate surface to stop development.

  Below, the process of step S103-S107 is demonstrated more concretely.

Table 2 shows the spatial ratios ri obtained in order from the smallest sum of squares with respect to the measured reflectance at 20 seconds after the start of development.

If you extracts the matching as in the conventional method is best, the space ratio 0.6 of the developer between the layers R1 to R4, the space ratio of the developer in the layer D _1, D ₂ is that 1.00 Result. However, looking at the sum of squares of rank 1 and rank 2, the difference is only 4%. This difference is in the range of errors when considering the ratio amplitude (in 10% increments). Accordingly, with respect to the square sum of rank 1, the sum of the squares up to + 10% is a candidate, and the space ratio predicted value rpi is obtained by a weighted average from the ratio of the developer to the space. In (Table 2), ranks 1 to 5 are targeted.

The results are shown in (Table 3).

  The bottom dimension of the resist at this stage was determined to be 114.4 nm (= 260 × (1-0.56)). Since the desired pattern size was 115 nm ± 8%, development was stopped at this stage and rinsing was performed. For confirmation, in the comparison of data-based data created under the same conditions as in (Table 1) (development / space is replaced with air / space and modeled with atmospheric optical constants), resist residual pattern dimensions Was 114.0 nm. Furthermore, when the line width of this resist pattern was measured with a length measuring machine using an electron beam, 114.1 nm was obtained, which was almost the same as the value measured during development.

  FIG. 5 shows a comparison between the reflectance calculated using the spatial ratio predicted value calculated by the above weighted average and the measured value. Here, a comparison between 500 nm and 700 nm is shown. Here, the solid line is a curve obtained by fitting measured values with a 15th-order function. It was confirmed that the reflectance chromatic dispersion calculated from the distribution shown in (Table 3) overlaps with the 15th order function very well. Therefore, it can be said that the actual measurement values are sufficiently reproduced.

  In FIG. 5, the reflectance chromatic dispersion calculated from the model with the smallest sum of squares is indicated by a wavy line. The curve drawn with a solid line and the curve drawn with a wavy line are in good agreement locally. However, it is difficult to say that they coincide in the entire wavelength region shown. Also, the resist pattern line width was recognized as 104 nm, and it was recognized as thin as 10 nm with respect to the actual value. According to this measured value, the substrate is regenerated by determining that the processing is defective. Needless to say, the previous 114 nm is a true value, and the determination of a good product and a defective product is erroneous.

According to the present embodiment, the dimensions can be predicted with higher accuracy than the conventional method. In addition, it is possible to make an appropriate determination on a non-defective product and a defective product. Table 4 shows the model conditions when the same accuracy analysis as this experiment is performed using the conventional method.

  The number of calculation values created in (Table 1) is 2.5 × 10 5 compared to that in Table 4, and the number of data can be greatly reduced. The detection time this time was 0.2 seconds, and the measurement during development could be easily performed. The conventional method takes about 14 hours. Therefore, the conventional method is extremely difficult to apply at the time of development, and is difficult to apply even for evaluation in ordinary air. Thus, the search time can be greatly shortened, and the resist pattern dimensions can be easily performed in the course of the process.

  In this embodiment, the resist pattern dimension is predicted, but the present invention is not limited to this. FIG. 6 shows the cross-sectional shape of the dimensional variation of the KrF resist being developed, using a database of reflectance chromatic dispersion calculated values calculated using the model of (Table 1). The cross-sectional shape is obtained by determining the ratio of the resist to the space for each layer and connecting it with a line in the film thickness direction. A fine dotted line is a cross-sectional shape 4 seconds after the start of development, a two-dot chain line is 6 seconds after the start of development, a dashed-dotted line is 10 seconds after the start of development, a coarse dotted line is 20 seconds after the start of development, and a solid line is a cross-sectional shape 30 seconds after the start of development. It can be seen that the skirt of the resist pattern is large after 4 seconds from the start of development. As the development progresses, the top and bottom (bottom) resist patterns gradually become thinner and the shape becomes rectangular.

  FIG. 7 shows the bottom dimension of the resist film of FIG. 6 with development time taken on the horizontal axis. It can be seen from FIG. 7 that the time required for the resist pattern to have a size of 0.44 with respect to the pitch is 20 seconds (the dotted line indicates the dimensional ratio without stopping development). As shown in FIG. 7, high-precision development control can also be achieved by sequentially obtaining the dimension with respect to time, predicting the development time to be the desired dimension, and stopping it. Since FIG. 6 is calculated by weighted average using a plurality of reflectance chromatic dispersion calculation values, high accuracy is also obtained in the relationship of FIG.

  In the conventional method, as shown in FIG. 8, it can be seen that the relationship of the size ratio with respect to the development time is reversed, and the control is greatly deteriorated. In this way, by using this method, the development control can be dramatically improved (the dotted line indicates the dimensional ratio obtained without stopping the development, which indicates that the reliability of the development stop is low).

  This embodiment relates to dimension measurement during development and dimension measurement after development, but is not limited thereto. In the case of simply measuring the dimensions, the resist pattern dimensional measurement is made up of a database of reflectance wavelength dispersion calculation values calculated by changing the spatial ratio of the material and the atmosphere, such as the oxide film etching pattern and wiring pattern, and the measured values. By analyzing with the same method, it is possible to obtain the pattern size and profile efficiently and in a short time with high accuracy.

  In this embodiment, the structure is predicted for the reflectance calculation value, but the present invention is not limited to this. The same applies to the measured values of cosδ and tanψ measured by ellipsometry technology against the database of cosδ and tanψ calculated by RCWA (Rigorous coupled-wave analysis) from the model of the membrane structure with varying spatial ratio It is also possible to find a similar waveform by a technique, predict the spatial ratio by a weighted average that increases the weight as the miscalculation is smaller, and specify the structure of the measurement substance.

  In this embodiment, the spatial ratio predicted value r is calculated by using the inverse of the sum of squares as shown in Equation (4), but the present invention is not limited to this. As the spatial ratio predicted value r, any relational expression may be used as long as the spatial ratio increases as R decreases with respect to the absolute value R of the error from the measured value. For example, the predicted spatial ratio r is expressed by equation (5), and in the range that R can take, f (R) decreases as R increases (for example, a monotonically decreasing function within the range that R can take). It ’s fine.

r = (Σri × f (R)) / (Σ (f (R)) (5)
f (R) = 1 / R, 1 / Rx: x is a positive real number, eR, etc. (second embodiment)
Regarding the etching depth, a database of reflectance chromatic dispersion calculation values was created by changing the spatial ratio of the etching pattern material and the atmosphere (supply gas and reaction gas) during etching, and compared with the measured values. It is also possible to extract a plurality of calculated chromatic dispersion values similar to the above and obtain the etching depth from the weighted average based on the square sum of the spatial ratios that are the basis of them.

  FIG. 9 shows the etching depth obtained from the reflectance measured for each etching time of the oxide film. The depth (dotted line) obtained by the conventional method is a result of jerking so that the etching rate with respect to time changes. On the other hand, in the result (solid line) obtained by the weighted average according to the present invention, the etching depth with respect to the obtained etching time is obtained very well, and an approximately constant value is obtained for the etching rate after 20 seconds. (Etching depth change amount per unit etching time is constant). Even in this embodiment, the etching depth can be analyzed with almost no influence of the analysis error.

  It is also possible to control the etching stop using FIG. However, the conventional method has a problem that the fluctuation is large and the etching stop accuracy is poor, and sometimes the stopper film under the oxide film is also etched. On the other hand, according to the method of the present invention, etching can be stopped with high accuracy, processing can be realized with little damage to the stopper material under the oxide film, and the stopper film forming process can be greatly reduced. it can.

(Third embodiment)
FIG. 10 is a flowchart showing manufacturing steps of a semiconductor device according to the third embodiment of the present invention. FIG. 11 is a cross-sectional view showing a manufacturing process of a semiconductor device according to the third embodiment of the present invention.

First, as shown in FIG. 11A, a first silicon oxide film having a thickness of 100 nm is formed on an unprocessed silicon substrate (step S201).
Next, the film thickness of the silicon oxide film A of the first silicon oxide film is measured (step S202).

  Next, as shown in FIG. 11B, a second silicon oxide film having a thickness of 500 nm is formed on the first silicon oxide film (step S203). The first silicon oxide film and the second silicon oxide film are films having slightly different optical constants. The substrate on which the second silicon oxide film is formed is referred to as substrate A.

  Next, in order to measure the film thickness of the second silicon oxide film, the light intensity of the substrate A is measured (step S204). FIG. 12 shows a schematic diagram of an optical system for measuring light intensity. Light is incident on the substrate A. The reflected light from the substrate A is measured, and the chromatic dispersion of the intensity change (tan ψ) and the phase change (cos Δ) is measured. As reflected light to be measured here, polarized light such as TM polarized light, TE polarized light, other polarized light, or direct reflected light may be measured without using a polarizing plate. The incident angle may not be obliquely incident but may be perpendicularly incident. Alternatively, measurement may be performed by entering light at a single wavelength and changing the incident angle and the detection angle. Any method may be used as long as it detects light obtained by entering and reflecting light.

  From the light intensity measurement result of the substrate A, the film thickness of the second silicon oxide film is obtained (step S205).

  Next, as shown in FIG. 11C, an antireflection film having a thickness of 50 nm and a resist film having a thickness of 250 nm are applied and formed on the second silicon oxide film (step S206). The resist film is exposed through a reticle with an ArF excimer laser. Bake processing and development processing are performed to form a resist pattern of 100 nm line / space (1: 1). The resist pattern is a pattern in which line patterns and space patterns each having a width of 100 nm are alternately arranged. The substrate in this state is referred to as substrate B.

  In order to measure the film thickness of the antireflection film, the pattern dimension of the resist film, and the film thickness, the light intensity of the substrate B is measured (step S207). In FIG. 13, the schematic diagram of the optical system of the test | inspection process of the board | substrate B is shown. Light is incident on the substrate to be processed. Of the diffracted light from the substrate to be processed, zero-order light is measured, and the chromatic dispersion of intensity change (tan ψ) and phase change (cos Δ) is measured. Similar to the film thickness measurement, polarized light such as TM polarized light, TE polarized light, and other polarized light may be used, or measurement may be performed without using a polarizing plate. Further, the first-order or higher light may be used instead of the zero-order light. The incident angle may not be perpendicular incidence but may be oblique incidence. Alternatively, measurement may be performed by entering light at a single wavelength and changing the incident angle and the detection angle. Any method may be used as long as it detects light obtained by entering light and reflecting or diffracting it.

  Next, as the inspection of the antireflection film and the resist pattern, the dimension (line width) of the resist pattern is measured (step S208). As described above, in a series of processing of the semiconductor wafer, the processing step and the inspection step of the structure (film thickness, pitch, ratio, dimension, optical constant) added in the processing step are repeatedly performed.

  A specific procedure of the inspection will be described with reference to FIG. FIG. 14 is a flowchart showing the procedure of the structural inspection according to the third embodiment.

  Prior to the inspection, a light intensity library containing the substrate structure and the corresponding light intensity is created.

  First, for all the layers of the substrate, a structure group including a structure expected in each layer is created (step S301). The layer structure includes the optical constant, film thickness, pitch, ratio, dimensions, and the like of the layer. The expected structure is a structure included in a range of structural variations that can occur in the processing steps. Next, a substrate structure is created from the combination of the structure groups of each layer.

  A board structure library A is created from the board structure created in step S201 (step S302).

  Next, for each substrate structure included in the first substrate structure library, light intensity calculation obtained by reflection or diffraction when light is incident under the same conditions as in actual inspection is performed, and the wavelength dispersion of the light intensity is calculated. (Step S303). Here, as a calculation method, Morham (J. Opt. Soc. Am., Vol. 12, No. 12) is used to solve the electric field, magnetic field, and intensity of diffracted light from a regular pattern using Maxwell's equations. RCWA (Rigorous coupled-wave analysis) method of 5, May 1995 1077-1086) is generally used.

  A light intensity library B containing the substrate structures stored in the substrate structure library A and the corresponding wavelength dispersion of the light intensity is created (step S304).

Next, a layer whose layer structure is known by inspection is extracted (step S401).
Next, a new layer structure group is created for the extracted layer based on the inspection result of the layer structure (step S402). The layer structure made here is made in consideration of inspection accuracy.

  Next, the extracted layer structure group is compared with the substrate structure library A created in step S102 to extract the substrate structure including the layer structure group extracted from the first substrate structure library (step S102). S403).

  Next, a light intensity corresponding to the extracted substrate structure is selected from the light intensity library B, and a light intensity library C is created (step S404).

  A substrate on which a plurality of layers are actually formed is transported into an inspection apparatus having an optical system as shown in FIGS. 2 and 3, and the light intensity in the inspection region is measured to measure the wavelength dispersion of the light intensity. (Step S501).

  Next, the chromatic dispersion of the measured light intensity is compared with the chromatic dispersion stored in the light intensity library C (step S502).

  Then, the chromatic dispersion that best matches the chromatic dispersion of the light intensity in the light intensity library is obtained, and the substrate structure corresponding to the chromatic dispersion is determined (step S503).

  Further, as another embodiment of the present invention, there may be a procedure as shown in FIG. Steps similar to those described in FIG. 14 are denoted by the same reference numerals. In the inspection method shown in FIG. 15, the substrate structure library A does not calculate the chromatic dispersion of the light intensity. After the substrate structure library B is created, the chromatic dispersion of the light intensity is calculated (step S414), and the light intensity library C is created (step S415).

  Further, as another embodiment of the present invention, there may be a form represented by a flowchart shown in FIG. The same steps as those described in FIGS. 14 and 15 are denoted by the same reference numerals.

  In the case of the inspection method shown in FIG. 15, a layer whose structure is not determined in advance is extracted (step S421). Then, a structure group of layers extracted in step S421 is created (step S422). The layer structure includes the optical constant, film thickness, pitch, ratio, dimensions, and the like of the layer. The expected structure is a structure included in a range of structural variations that can occur in the processing steps. The layer whose structure is determined is extracted (step S423). A structure group of layers whose structure is determined is created (step S424). Then, the substrate structure library B is created based on the structure group created in step S422 and step S424 (step S403).

  Next, an example in which the above-described inspection method is applied to the semiconductor device manufacturing process described with reference to FIGS. 10 and 11 will be described. FIG. 17 is a flowchart showing manufacturing steps of a semiconductor device according to the third embodiment of the present invention. At the time of the light intensity measurement (step S204), the silicon oxide film A and the silicon oxide film B are formed on the Si substrate. Therefore, as in the conventional example, the substrate structure library IIA creates an expected structure group of the Si substrate, an expected structure group of the silicon oxide film A, and an expected structure group of the silicon oxide film B ( Step S301). Then, a substrate structure library IIA is created using each structure group created in step S301a (step S302). In step 301a, the numbers of the structure groups of the Si substrate, the silicon oxide film A, and the silicon oxide film B were 1, 21, and 101, respectively. Therefore, the number of substrate structures included in the substrate structure library IIA is 2121.

Using the film thickness of the silicon oxide film A measured in step S203, a new structure group of the silicon oxide film A is created (step S402). Since the film thickness dA of the silicon oxide film A is required to be 102 nm and the measurement accuracy is ± 1 nm, a structure group of the silicon oxide film A as shown in Table 5 was created. The thickness increment was set to 1 nm in order to improve measurement accuracy.

  Therefore, the necessary substrate structure in the substrate structure library IIA is a combination of the structure groups shown in (Table 5). The number of structure groups of the second silicon oxide film B is three. The number of structures of the Si substrate, the silicon oxide film A, and the silicon oxide film B extracted in step S403 is 1, 3, 101. The number of chromatic dispersions of light intensity included in the optical structure library IIB created in step S404 is 303.

  At the time of the wavelength dispersion measurement of the light intensity in step S207, the silicon oxide film A, the silicon oxide film B, the antireflection film (ALL), and the resist film are formed on the Si substrate. Structure groups are respectively created from the expected structures of the silicon oxide film A, the silicon oxide film B, the antireflection film (ALL), and the resist film. Here, the structures of the silicon oxide film A and the silicon oxide film B are detected in steps S202 and S205. A structure group of the silicon oxide film A and the silicon oxide film B is created based on the detection result (step S422). In addition, a structure group of an Si substrate, an antireflection film, and a resist film, whose structure is not detected, is created from each expected structure (step S424).

Since the thickness of the silicon oxide film A is required (dA = 102 nm) and the thickness of the silicon oxide film B is required (dB = 510 nm), and the measurement accuracy is ± 1 nm, as shown in (Table 6). A group of silicon oxide films A and B was prepared. Table 6 shows a group of structures of the Si substrate, the antireflection film, and the resist film.

  As shown in (Table 6), the necessary substrate structures in the step substrate structure library IIIB are combinations of the structure groups shown in (Table 6). The number of structure groups of the Si substrate, silicon oxide film A, silicon oxide film B, antireflection film, and resist film is 1, 3, 11, 3, 231 respectively, and the number of substrate structure libraries IIIB is 22,869. . The light intensity library IIIC is created based on the substrate structure library IIIB.

  As shown in (Table 6), when the structure group of the silicon oxide film A and the silicon oxide film B was created, the structure was defined within the measurement accuracy range (± 1%). The structure may be created at ± 2%). Further, when creating the structure group of the antireflection film and the resist film, the structure group may be created in a double range (± 4%) with allowance for fluctuations in the processing steps.

  A specific inspection procedure of the conventional method will be described with reference to FIG. In FIG. 18, each step is denoted by the same reference numeral as in FIG. 14, and detailed description thereof is omitted. As shown in FIG. 18, conventionally, the light intensity library B based on the already obtained layer inspection results has not been created. The light intensity library A is compared with the result of the light intensity measurement (step S512).

  Next, the case where the method shown in FIG. 18 is applied to the flow shown in FIG. 10 will be described with reference to FIG.

At the time of step S204, the silicon oxide film A and the silicon oxide film B are formed on the Si substrate. Therefore, an expected structure group of the Si substrate, an expected structure group of the first silicon oxide film, and an expected structure group of the second silicon oxide film are respectively created (step S301a).

  In this embodiment, fluctuations in the processing steps of the silicon oxide film A and the silicon oxide film B are not in the refractive index and the extinction coefficient, and the film thickness is ± 10% with respect to the center value. The structural group was defined as in 6).

  As shown in Table 7, the number of structure groups of the Si substrate, the silicon oxide film A, and the silicon oxide film B is 1, 21, and 101, respectively, and the number of the substrate structure library IIA is 2121. (Step S302a). The light intensity library IIA was created based on the substrate structure library IIA (step S303a).

At the time of step S207, the silicon oxide film A, the silicon oxide film B, the antireflection film, and the resist are formed on the Si substrate. Therefore, an expected structure group of the Si substrate, an expected structure group of the silicon oxide film A, an expected structure group of the silicon oxide film B, an expected structure group of the antireflection film, and an expected structure group of the resist Are respectively created (step S301b). Each structural group is shown in (Table 8).

  In this embodiment, fluctuations in the processing steps are not in the refractive index and extinction coefficient, the oxide film thickness is ± 10% of the center value, the antireflection film thickness is ± 2%, and the resist film thickness is Since the resist size was ± 2% and ± 10%, a structure group was generated as shown in (Table 8). The numbers of the structures of the Si substrate, the silicon oxide film A, the silicon oxide film B, the antireflection film, and the resist film are 1, 21, 101, 3, 231 respectively. As a result, the number of substrate structures stored in the substrate structure library IIIA generated in step S302b is 453789. The light intensity library IIIA is created based on the substrate structure library IIIA (step S303b).

(The invention's effect)
(2) Improvement of measurement accuracy In the conventional method, for example, when the thickness of the silicon oxide film B is obtained in the structural inspection step (step S205), the thickness of the silicon oxide film A is d11 nm and the thickness of the silicon oxide film B is determined. When the thickness of the silicon oxide film A is d11 nm and when the thickness of the silicon oxide film A is d12 nm and the thickness of the silicon oxide film B is d22 nm, the chromatic dispersion hardly changes. Although the thickness of the oxide film B is d21 nm, there is a possibility that the thickness of the silicon oxide film A is calculated as d12 nm and the thickness of the silicon oxide film B is calculated as d22 nm. However, if the film thickness d12 nm of the silicon oxide film A is not included in d1 ± 1% in FIG. 9, the value d12 is not calculated as the film thickness. The measurement accuracy is improved.

(1) Significant reduction in the number of libraries Since the conventional method uses all the structural variations expected in each layer as a library, an enormous number of light intensity libraries are required. In this method, since the light intensity library is created by taking in the result of the inspection performed prior to the inspection in a certain structural inspection step, the number of the light intensity libraries is greatly reduced. As a result, the time for searching for similar light intensity chromatic dispersion from the measurement result of chromatic dispersion of light intensity is significantly shortened.

Table 9 shows the numbers included in the light intensity library.

  In the conventional method, the substrate structure is determined using the light intensity libraries IIA and IIIA. In this method, the substrate structure is determined using the light intensity libraries IIB and IIIB.

  Further, since the number of libraries is greatly reduced, it becomes possible to perform the inspection process without making a light intensity library in advance as shown in FIGS. For example, in the structure determination process (step S208), the substrate structure library IIIA and the light intensity library IIIA are conventionally created before the inspection process. In this method, the light intensity library IIIB is created after the substrate structure is determined without creating the light intensity library IIIA (step S205). Since the light intensity library IIIA is not formed, the capacity of the library can be greatly reduced.

  In addition, when a conventional coater / developer has a structure inspection function, it is sufficient to complete the structure inspection before the end of the series of lot processing. The inspection result may be obtained by comparing the light intensity created and measured with the light intensity library IIIB. In this way, it is possible to perform a highly accurate test while greatly reducing the capacity of the library.

(Fourth embodiment)
The present embodiment is an embodiment of a pattern evaluation method when the present invention is used.

  FIG. 20 is a diagram schematically showing pattern evaluation during development, and FIG. 21 is a flowchart showing the procedure of the pattern evaluation method. The processing procedure of the present invention will be described with reference to FIGS.

  As shown in FIG. 20, a resist is applied and baked on a specific base substrate (from the lower layer, Si 301, lower layer film 302, antireflection film 303), pattern exposure, and a developer is supplied to the baked substrate. The light is incident on the monitor area of, and diffracted light is detected. Therefore, in addition to the developer 305, a mixed phase of the resist reaction product and developer exists between the upper layer and the pattern of the resist pattern 034. The incident light is broad light having a plurality of wavelengths (for example, light from a halogen lamp). The intensity of diffracted light to be detected is TM polarization of 0th-order diffracted light in a wavelength range of 400 to 800 nm. TM polarized light is obtained by inserting a polarizing plate on the optical path of 0th-order diffracted light. The chromatic dispersion of the intensity of the diffracted light detected here is uniquely determined by the structure of the measurement region and the optical constant. Therefore, in the present invention, the diffracted light from the pattern is measured during development, and the dimension is evaluated from the result.

  A specific processing procedure will be described with reference to FIG. In the present embodiment, the diffracted light intensity is predicted in advance before the development processing step, and a library is created. Evaluation in this case consists of a library creation step performed before the development processing and a step of evaluating pattern dimensions in the development processing step.

  In the library creation process, first, the shape of the opening of the development area when the resist is dissolved in the developer, and the mixing ratio of the reaction product and developer in the liquid phase composed of the resist reaction product and developer. And a development model is determined (step S601).

  Next, the structure of the substrate composed of the liquid phase and the substrate to be processed and the optical constants of the developer, the substrate, and the reaction product are given (step S602). Based on the given structure and optical constant, calculation is performed when light is incident at a specific angle, and the chromatic dispersion of the diffracted light intensity is calculated (step S603). Here, as a calculation method, Morham (J. Opt. Soc. Am., Vol. 12, No. 12) is used to solve the electric field, magnetic field, and intensity of diffracted light from a regular pattern using Maxwell's equations. 5, May 1995 1077-1086) and the like. The RCWA (Rigorous coupled-wave analysis) method was used. Calculation is performed for all conditions of the given structure and optical constant, and a chromatic dispersion library of diffracted light intensity is created (S604).

  Next, the process of evaluating pattern dimensions will be described. The substrate is transported to the developing unit (step S605). Next, a developing solution is supplied onto the substrate, and development is started (step S606). Thereafter, the reflected light of the monitor area is measured (step S607). Here, it is necessary to adjust the position of the substrate and the position of the monitor unit so that light can enter the monitor region and the diffracted light can be detected by the detector. This adjustment may be performed before the supply if the substrate does not move when supplying the developer, or after the supply otherwise. Next, the measured chromatic dispersion of the diffracted light intensity is compared with the chromatic dispersion of the diffracted light intensity of the library (step S608). Then, the chromatic dispersion that best matches the chromatic dispersion of the diffracted light intensity in the library is obtained, and the pattern dimension is calculated (step S609).

  Next, the library creation process will be described in detail. As described with reference to FIG. 20, a resist reaction product is generated from the resist pattern being developed. Hereinafter, the developer containing the reaction product of the resist is referred to as a mixed phase, and the ratio of the resist in the mixed phase is defined as the mixing ratio. In general, the diffusion of reaction products is not fast enough, so the accuracy of pattern evaluation during development deteriorates unless the reaction product diffusion method, that is, the distribution of the mixing ratio in the liquid phase, is modeled. End up. Therefore, in the present invention, a development model is determined in the library creation process (step S601).

  FIG. 22A shows a schematic diagram of how development proceeds in the initial stage of development. In the initial development stage, development proceeds in the film thickness direction in the development area. As a result, since the reaction product diffuses in the direction indicated by the arrow, the mixing ratio becomes a distribution in the film thickness direction, and the ratio decreases as the distance from the substrate increases. When the development in the film thickness direction is completed, the development proceeds in the horizontal direction as shown in FIG. Therefore, the reaction product diffuses in the direction as indicated by the arrow in FIG. 23, so that the mixing ratio decreases in the lateral direction as the distance from the pattern side wall increases. Actually, since the superposition of these changes in the mixing ratio becomes the distribution of the mixing ratio, a distribution in which the mixing ratio decreases in the direction of the arrow in FIG. 23 is assumed.

Next, specific steps until the structure / optical constant of the substrate are given will be described. FIG. 24 shows a layer structure in the film thickness direction created based on the model of FIG. In FIG. 24, p is a pattern pitch, and w ₁ is a resist pattern dimension. Layer L ₀ is the developer, layer L ₀₁ is the mixed phase of the developer and resist reaction product, layers L ₁₁ to L ₁₃ are the mixed phase of the developer and resist reaction product and the resist pattern, and layer L ₂ is The antireflection film, the layer L ₃ is a lower layer film, and the layer L ₄ is Si. The difference between the layer L ₁₁ , the layer L ₁₂ , and the layer L _{13 is} the difference in the mixing ratio of the reaction products. The reaction product is generated at the interface between the resist and the developer and diffuses into the developer layer, so that the ratio is high in the lower part (L13) of the pattern.

The distribution of the mixing ratio of the layers L _{0 to} L _{1 3} is shown in FIG. Since all of the layers L0 are developers, the mixing ratio is zero. The layer L ₀₁ is a mixed phase of the developer and the reaction product, and the mixing ratio takes a constant value r ₀₁ . Since the mixed phase of the developer and the reaction product of the resist and the resist pattern are mixed in the layers L ₁₁ to L ₁₃ , the mixing ratio r becomes a function of the position, and is as shown in FIG. The mixture ratio is 1 in the resist pattern region (0 <x <w1 / 2, p-w1 / 2/2 <x <p). The mixing ratio takes a value between 0 and 1 in the mixed phase region (w1 / 2 <x <p−w1 / 2) of the developer and the reaction product of the resist. The layer L ₁₁ takes r ₁₁ (x), the layer L ₁₂ r ₁₂ (x), the layer L ₁₃ r ₁₃ (x), and takes the form of a function depending on the position. Each of the functions r ₁₁ (x) to r ₁₃ (x) has a relationship such that the mixing ratio decreases as the distance from the interface between the liquid phase and the substrate increases.

Next, based on the mixing ratio of FIG. 25, the refractive index, extinction coefficient, film thickness, and dimensions of the layers L0 to L4, which are parameters necessary for giving the structure and optical constants, are given. Table 10 shows the refractive index, extinction coefficient, film thickness, and dimensions of the layers L0 to L4 given. FIG. 26 shows the refractive index of each layer. FIG. 27 shows the extinction coefficient of each layer. The refractive index and extinction coefficient of the layers L ₀₁ and L _{11 to} L ₁₃ are both functions of the position x.

The layer L ₀ is a developer layer, and both the refractive index n ₀ and the extinction coefficient k ₀ take the values of the developer. The layer L ₀₁ is a mixed phase of the developer and the reaction product, and the refractive index n ₀₁ and the extinction coefficient k ₀₁ are both between the developer (n ₀ , k ₀ ) and the resist (n ₁ , k ₁ ). Takes a value. The film thickness is d01.

Since the mixed phase of the developer and the reaction product of the resist and the resist pattern are mixed in the layers L ₁₁ to L ₁₃ , the refractive index and the extinction coefficient are functions of the position, as shown in FIGS. 26 and 27, respectively. Become. In the resist pattern region _{(0 <x <w 1/} 2, p-w 1/2 <x <p), the refractive index, taking the resist values both extinction coefficient (n _1, k _1). In mixed phase regions of the reaction product of a developing solution and the resist _{(w 1/2 <x <} p-w 1/2), the refractive index and extinction coefficient are both take an intermediate value of the reaction product with a developer . The film thicknesses are d ₁₁ , d ₁₂ , and d ₁₃ , respectively.

The layer L ₂ is an antireflection film, and the refractive index n ₂ and the extinction coefficient k ₂ take the values of the antireflection film. The film thickness is d _2. The layer L ₃ is a lower layer film, and the refractive index n ₃ and the extinction coefficient k ₃ take the values of the lower layer film. The film thickness is d _3. The layer L4 is a Si film layer, and the refractive index n ₄ and the extinction coefficient k ₄ take the value of Si.

Parameters may vary process, the film thickness d ₁ of the layer L _01, refractive index n _01, the extinction coefficient k _01, the thickness d ₁₁ of layer L _11, refractive index n ₁₁ (x), the extinction coefficient k11 ( x), the thickness of the layer L12 d12, the refractive index n ₁₂ (x), the extinction coefficient k ₁₂ (x), the thickness d ₁₃ of layer L _13, refractive index n ₁₃ (x), the extinction coefficient k ₁₃ ( x), the dimension w ₁ of layer L _1, the thickness d ₂ of the layer L _2, and a thickness d ₃ of layer L _3. As each parameter is to take five levels, the combination of the structural and optical constants given 1,953,125 pieces 5 ⁹ (step S601, S602). Of the parameters, the refractive index and the extinction coefficient vary depending on the diffusion method of the reaction product. Calculation is performed for each structure and optical constant (step S603), and chromatic dispersion of 1953125 diffracted light intensities is created (step S604).

  In the comparison of the measured waveforms at the time of pattern evaluation (step S608), the dimensions were calculated by comparing the measured wavelength dispersion of the reflected light intensity with the wavelength dispersion of the reflected light intensity of the library.

A conventional development model is shown in FIG. In FIG. 28, as shown in FIG. 28, the distribution of the mixing ratio between the developer and the reaction product in the film thickness direction is not taken into consideration. Further, the distribution of the mixing ratio between the developer and the reaction product between the resists is not taken into consideration. The values of refractive index, extinction coefficient, film thickness, and dimensions are shown in (Table 11). The refractive index and extinction coefficient of the conventional model are shown in FIGS.

Thickness d ₁ of the layer L ₁ the parameters may vary with process, dimensions w _1, the thickness d ₂ of the layer L _2, and a thickness of the layer L _3. Assume that each parameter takes 5 levels. Then, the structure and optical constants are given 5 ⁴ 625 structure and optical constants. Calculation is performed for each structure and optical constant, and wavelength dispersion of 625 diffracted light intensities is created. Then, the measured chromatic dispersion of the reflected light intensity is compared with the calculated 626 chromatic dispersions to calculate the dimensions.

  When this method was compared with the conventional method, higher accuracy was obtained with this method assuming a mixed phase of the resist reaction product and the developer.

  In this embodiment, the chromatic dispersion is measured. However, when sufficient accuracy is obtained with a single wavelength, the measurement with a single wavelength may be performed. In this embodiment, the library is created before the development process, but may be created during the development process. Moreover, although the pattern dimension was evaluated in this embodiment, it is also possible to evaluate the pattern shape and the mixing ratio of the developer and the reaction product.

  In this embodiment, TM polarization was evaluated using a polarizing plate. However, if sufficient accuracy is obtained, detection may be performed without inserting a polarizing plate. Further, TE polarized light, other polarized light, intensity change (tan ψ), and phase change (cos Δ) may be evaluated. Further, first-order or higher-order diffracted light may be detected. Also, the angle of the incident light need not be perpendicular incidence.

  In the present embodiment, since the mixed phase model is complicated, it is compared with a library of many chromatic dispersions of 1953125. However, if the evaluation time zone is only the latter half of the development, for example, the reaction product Is considerably advanced, the processing time can be shortened unless a comparison is made with a library containing models that have not progressed. Therefore, it is also effective to determine the data to be compared according to the time zone.

(Fourth embodiment)
The present embodiment is an embodiment of a pattern forming method when the present invention is used.

  A diagram schematically showing pattern evaluation during development is FIG. 11 as in the third embodiment, and FIG. 31 is a flowchart showing a processing procedure. The processing procedure of the present invention will be described with reference to FIGS.

  As shown in FIG. 11, a resist is applied to a specific base substrate (from the lower layer to Si, the lower layer film, and the antireflection film), baked, pattern exposed, and a developer is supplied to the baked substrate. Light is incident on the region and diffracted light is detected. Incident light is broad light having a plurality of wavelengths (for example, light from a halogen lamp), and the detected diffracted light intensity is measured in the wavelength range of 400 to 800 nm without using a polarizing plate. The chromatic dispersion of the intensity of the diffracted light detected here is uniquely determined by the structure of the measurement region and the optical constant. Therefore, in the present invention, the diffracted light from the pattern is measured during development, the dimension is calculated from the result, and the development is terminated.

  A specific processing procedure will be described with reference to FIG. In the present embodiment, the diffracted light intensity is predicted in advance before the development processing step, and a library is created. Evaluation in this case consists of a library creation step and a development processing step performed before the development processing. In the library creation process, first, the shape of the opening of the development area when the resist is dissolved in the developer, and the mixing ratio of the reaction product and developer in the liquid phase composed of the resist reaction product and developer. And a development model is determined (step S701).

  Next, the structure and optical constant of the substrate composed of the liquid phase and the substrate to be processed are given (step S702).

  Based on the given structure and optical constant, calculation is performed when light is incident at a specific angle, and the chromatic dispersion of the diffracted light intensity is calculated (step S703).

  Calculation is performed for all conditions of the given structure and optical constant, and a chromatic dispersion library of diffracted light intensity is created (step S704). Here, as a calculation method, Morham (J. Opt. Soc. Am., Vol. 12, No. 12) is used to solve the electric field, magnetic field, and intensity of diffracted light from a regular pattern using Maxwell's equations. 5, May 1995 1077-1086), etc., using the RCWA (Rigorous coupled-wave analysis) method.

  Next, the development processing step will be described. The substrate is conveyed to the developing unit (step S711). Next, a developing solution is supplied onto the substrate and development is started (step S712). Thereafter, the reflected light of the monitor area is measured (step S713). For that purpose, the position of the substrate and the position of the monitor unit are adjusted so that the light enters the monitor area and the diffracted light can be detected by the detector. There is a need to. This adjustment may be performed before the supply if the substrate does not move when supplying the developer, or after the supply otherwise. Next, the measured wavelength dispersion of the diffracted light intensity is compared with the wavelength dispersion of the diffracted light intensity of the library. Then, the chromatic dispersion that best matches the chromatic dispersion of the diffracted light intensity in the library is obtained (S714), and the dimension of the pattern is calculated (step S715). The calculated dimension is compared with a desired dimension, and development completion determination is performed. If the calculated dimension is a desired value, the development is terminated (step S717). When the calculated dimension is not a desired value, the reflected light is measured again (step S713), and the waveform with the library is compared (step S714). This process is continued until the calculated pattern dimension reaches a desired value.

Since the method of giving the structure and optical constants of this embodiment (step S702) is different from that of the third embodiment, the difference will be described. A drawing incorporating a development model and a schematic diagram of a film structure are FIGS. 23 and 24, respectively, as in the third embodiment. The distribution of the optical constant of the layer 1 is also shown in FIGS. 26 and 27 as in the first embodiment. In this embodiment, the structure and optical constants are given (the same as the third embodiment so far), and then the optical constants are averaged in each layer, and the optical constants in each layer are set to constant values. deal with. Therefore,
The refractive index n _1i of the layer L _1i is

The extinction coefficient k _1i of the layer L _1i is

It becomes. Therefore, the finally obtained structure and optical constants are as shown in (Table 12).

The parameters that can be varied in the process are the layer L ₀₁ film thickness (d ₀₁ ), the reaction product diffusion method (n ₀₁ , k ₀₁ ), the layer L ₁₁ film thickness (d ₁₁ ), and the reaction product diffusion parameter. Method (n ₁₁ , k ₁₁ ), layer L ₁₂ film thickness (d ₁₂ ), reaction product diffusion method (n ₁₂ , k ₁₂ ), layer L ₁₃ film thickness (d ₁₃ ), reaction product It is assumed that the diffusion method (n ₁₃ , k ₁₃ ), the dimension of layer 1 (w ₁ ), the film thickness of layer 2 (d ₂ ), and the film thickness of layer 3 (d ₃ ) are assumed to have five levels. The optical constants are 5 × 5 × 5 × 5 × 5 × 5 × 5 × 5 × 5 × 5 and 1953125 are given (steps S701 and 702), and each is calculated (step S703). A chromatic dispersion is created (step S704).

  In the comparison of the measured waveforms at the time of pattern evaluation (step S714), the dimensions were calculated by comparing the measured wavelength dispersion of the reflected light intensity with the wavelength dispersion of the reflected light intensity of the library, and the development was terminated. A higher accuracy was obtained when the mixed phase of the resist reaction product and the developer was assumed than when the mixed phase of the resist reaction product and the developer was not used as the library.

  In this embodiment, the chromatic dispersion is measured. However, when sufficient accuracy is obtained with a single wavelength, the measurement with a single wavelength may be performed. In this embodiment, the library is created before the development process, but may be created during the development process.

  Although this embodiment was evaluated without using a polarizing plate, TM polarized light, TE polarized light, other polarized light, intensity change (tan ψ), and phase change (cos Δ) may be used. Further, first-order or higher-order diffracted light may be detected. Also, the angle of the incident light need not be perpendicular incidence.

  In this embodiment, since the model of the mixed phase is complicated, it is compared with a library of many wavelength dispersions of 1953125. However, in the evaluation in the latter half of the development, the diffusion of the reaction product is considerably advanced. The processing time can be shortened unless a comparison is made with a library containing models that are not. On the contrary, in the first half of the development, the diffusion of the reaction product has not progressed so much, so that the processing time can be shortened unless the comparison with the library including the advanced model is performed. Therefore, it is also effective to determine the data to be compared according to the time zone.

  According to this embodiment, the liquid phase model composed of the reaction product of the resist and the developer in the development is clarified, and the distribution of the mixing ratio in the liquid phase increases as the distance from the interface where the liquid phase reacts with the resist increases. As a result, the diffracted light intensity of the pattern under development can be accurately predicted. As a result, the pattern evaluation accuracy was greatly improved.

(Fifth embodiment)
FIG. 32 shows a resist pattern process condition determination system according to the fifth embodiment. This system includes a coating and developing apparatus including a comparison calculation feedback unit 401, a spin coating coating unit 402, a post-coating heating unit 403, a post-exposure heating (PEB) unit 404, and a developing unit 405 including a shape measuring device 405a. The exposure apparatus 406, the carrier station 407, and a transport unit 408 capable of transporting each unit in units of 1 wafer. The coating unit 402 mixes a first resist solution and a second resist solution (partially extracted solution of the first resist solution, such as a solvent or a photosensitive agent solution) that are not shown in the figure, and then a chemical solution is applied onto the substrate to be processed. In some cases, a chemical adjustment unit that transports the liquid to the nozzle for dropping the liquid is attached.

  This system was used to determine the optimum process conditions for newly developed resists. First, 24 substrates to be processed were mounted on the carrier station 407 of this system. The mounted substrate is processed in units of several. The procedure for automatically determining process conditions will be described below with reference to FIG. FIG. 33 is a flowchart showing a procedure of a process condition determining method according to the fifth embodiment of the present invention.

First, the initial conditions recorded in the coating and developing apparatus and exposure apparatus 405 are read (step S801). In the coating unit 404, a coating film is formed on the three substrates by setting at the rotation speeds of the initial conditions shown in (Table 13) (step S802).

  Immediately after application, the shape measuring instrument 405a measures the thickness of the photosensitive resin film (step S803). The comparison calculation feedback unit 401 compares the measured film thickness with the set range (step S804).

  If the film thickness is within the set range, the rotational speed of the initial condition is set as the process condition (step S805). If the film thickness is not within the set range, the comparison calculation feedback unit 401 determines the optimum rotation speed for obtaining a desired film thickness from the relationship between the rotation speed and the film thickness (step S806). In the case of the present embodiment, a resist film having a film thickness within the set range was not obtained under the conditions shown in (Table 13). As a result of correcting the number of rotations to 2400, a resist film having a film thickness within the normal range was obtained. Therefore, 2400 rpm was set as the process condition.

Subsequently, the exposure amount and the focus position are respectively changed within the substrate surface, and a plan is set for setting two levels of conditions for three parameters of post-exposure baking (PEB) temperature and time and development time for each substrate. Based on the established experimental design, the process of coating, exposure, post-exposure heating (PEB), development, and shape measurement is sequentially performed on a total of eight sheets (step S807). Table 14 shows the setting values of each parameter.

  Then, the shape measuring instrument 405a measures the size and shape of the photosensitive resin pattern opening (region from which the photosensitive resin has been removed) (step S808). Measurement of dimensions and shape is performed by irradiating a regular pattern with light, and using Maxwell's equations from the diffracted light information from the pattern to solve the electric field, magnetic field, and intensity. Morham (J.Opt.Soc.Am., Vol.12, No.5, May 1995 1077-1086) and the like, using the RCWA (Rigorous coupled-wave analysis) method.

  It is determined whether the measured value is within an allowable range (step S809). In the case of pattern dimensions, 100 nm ± 5% is an allowable range. In the case of a pattern shape, the allowable range is a side wall angle of 88 to 90 °.

If the measured value is within an allowable range, an exposure amount-focus range (ED-margin) that can obtain a processed shape within the allowable range is obtained, and correction is performed so that this range is as wide as possible.

By comparing the eight ED-margin data and the coating / PEB / development level at that time, the effect of widening the ED-margin was seen when the baking time was longer. On the other hand, the longer time of the previous condition 2 levels was assigned. In addition, the comparison calculation feedback unit 401 indicates that there is an interaction between the heating temperature and the development time, and the baking temperature is ± 2 ° C. (3 levels) based on the temperature considered to be appropriate (first time) The development time is set to ± 10% (3 levels) based on the time considered to be appropriate (the time width used for the first experiment level). The coating unit 402, post-exposure heating (PEB) unit 404, developing unit 405, and the exposure apparatus were instructed to perform coating, exposure, PEB, development, and shape measurement processing for a total of nine sheets.

  The comparison calculation feedback unit 401 determines a condition for obtaining the maximum ED-margin, and the bake temperature corresponding to the condition is + 0.4 ° C. with respect to the reference and the optimum condition is −5% for the development time. This condition was set for each of the heating device and the developing device. A focus offset and an appropriate exposure amount were set in the exposure apparatus.

  As described above, the photosensitive resin pattern forming system automatically measures the size and shape of the photosensitive resin pattern in the photosensitive resin pattern forming system in the process of processing a total of 20 substrates, and provides feedback to related means. Thus, the coating conditions, heating conditions, exposure conditions, and development conditions could be determined.

  In addition, semiconductor devices created by creating a photosensitive resin pattern under the conditions defined and etching the substrate to be processed using this photosensitive resin pattern as a mask have dramatically improved dimensional accuracy, so device reliability I was able to dramatically improve the sex.

  These operations require 2 minutes for film thickness determination, feedback, 23 minutes for evaluation of 3 parameters and 2 levels, and 25 minutes for detailed condition determination (2 parameters and 3 levels). It was possible to optimize the patterning using the material in only 50 minutes. In the case of the conventional destructive inspection, 8 days are required, and the condition determination time can be greatly shortened.

  The selection parameter, level, correction target, and number of substrates required for condition determination in this embodiment are not limited to those in the above embodiment. Various forms can be taken depending on the required photosensitive resin material, the required items and required values for the photosensitive resin pattern, and their ranges.

  In addition, it is preferable to carry out as follows about the method of fine adjustment of a level.

1) Correction for resist pattern size of positive resist a) Thicker than desired size Correction for increasing the exposure amount of the exposure unit, or correction for increasing the heating temperature with respect to the heating unit used for heating after exposure, or The correction is performed by either a correction for increasing the developing time of the developing unit, a correction for increasing the developer concentration of the developing unit, or a correction for increasing the developer temperature of the developing unit.

b) When smaller than the desired dimension Correction for reducing the exposure amount of the exposure means, correction for lowering the heating temperature relative to the heating means used for heating after exposure, or correction for shortening the development time of the development means, or The correction is performed by either correction for lowering the developer concentration of the developing means or correction for lowering the developer temperature of the developing means.

2) Correction for opening size of positive resist a) Thicker than desired size Correction for reducing exposure amount of exposure means, correction for lowering heating temperature with respect to heating means used for heating after exposure, or development Either a correction that shortens the developing time of the means, a correction that reduces the developer concentration of the developing means, or a correction that decreases the developer temperature of the developing means. C) If the dimension is smaller than the desired dimension Increase the exposure amount of the exposure means Correction to increase the heating temperature for the heating means used for heating after exposure, or correction to increase the developing time of the developing means, or correction to increase the developer concentration of the developing means, developing means The correction is performed by either of the correction for increasing the developer temperature.

3) Correction to resist size of negative resist a) Thicker than desired size Correction to reduce exposure amount of exposure means, or correction to lower heating temperature with respect to heating means used for heating after exposure, or development Either correction for shortening the developing time of the means, correction for reducing the developer concentration of the developing means, or correction for reducing the developer temperature of the developing means is performed.

b) When it is thinner than the desired dimension Correction for increasing the exposure amount of the exposure means, correction for increasing the heating temperature relative to the heating means used for heating after exposure, or correction for increasing the development time of the developing means, Alternatively, either correction for increasing the developer concentration of the developing unit or correction for increasing the developer temperature of the developing unit is performed.

4) Correction for the opening size of the negative resist a) When the thickness is larger than the desired size Correction for increasing the exposure amount of the exposure unit, or correction for increasing the heating temperature with respect to the heating unit used for heating after exposure, or The correction is performed by either a correction for increasing the developing time of the developing unit, a correction for increasing the developer concentration of the developing unit, or a correction for increasing the developer temperature of the developing unit.

b) Correction for reducing the exposure amount of the exposure means when it is thinner than the desired dimension, correction for lowering the heating temperature relative to the heating means used for heating after exposure, or correction for shortening the development time of the development means, or The correction is performed by either correction for lowering the developer concentration of the developing means or correction for lowering the developer temperature of the developing means.

  In the present embodiment, the RCWA method is used for shape measurement. However, the present invention is not limited to this, and any method can be applied to the shape measurement means as long as it can predict pattern dimensions and shapes.

  The present embodiment relates to determination of conditions for forming a resist pattern in a lithography process, but is not limited to this, and can also be applied to determination of conditions in an etching process. In addition, when a process condition determining apparatus provided with a resist adjusting means is used, it is possible to simultaneously optimize the resist solution. For the resist adjustment means, use chemicals with different types of resins, types of photosensitizers, types of dissolution inhibitors, types of solvents, etc., and those with a blend ratio changed by a mixer and applied on the substrate to be processed. As in this embodiment, it is possible to optimize the material while optimizing the coating, baking and developing conditions. When the material manufacturer implements the present invention, it becomes easy to attach and sell the optimum process conditions for the material.

In the present invention, a method of creating a sample by brute the set level is used as an experimental design, but the present invention is not limited to this. L ₁₈ is likely to be found optimal conditions with fewer samples Applying Taguchi method using orthogonal table such.

  The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the invention at the stage of implementation. Further, the above embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some constituent requirements are deleted from all the constituent requirements shown in the embodiment, the problem described in the column of the problem to be solved by the invention can be solved, and the effect described in the column of the effect of the invention Can be obtained as an invention.

The flowchart which shows the structure determination method of the to-be-evaluated object concerning 1st Embodiment. Sectional drawing which shows how to advance development of a resist film. Sectional drawing which shows the layer structure of the film thickness direction produced based on the model of how to advance development. The figure which shows schematic structure of a reflectance measurement means. The figure which shows the comparison with the calculated reflectance and a measured value. The figure which shows the result of having calculated | required the dimensional variation of the KrF resist in image development as a cross-sectional shape. The figure which shows the bottom dimension with respect to development time. The figure which shows the bottom dimension with respect to the development time calculated | required by the conventional method. The figure which shows the etching depth with respect to the etching time concerning 2nd Embodiment. 9 is a flowchart showing manufacturing steps of a semiconductor device according to the third embodiment. Sectional drawing which shows the manufacturing process of the semiconductor device concerning 3rd Embodiment. The figure which shows typically the board | substrate A and the optical system of light intensity measurement. The figure which shows typically the optical system of the board | substrate B and light intensity measurement 10 is a flowchart showing a procedure of structural inspection according to the third embodiment. 10 is a flowchart showing a procedure of structural inspection according to the third embodiment. 10 is a flowchart showing a procedure of structural inspection according to the third embodiment. 9 is a flowchart showing manufacturing steps of a semiconductor device according to the third embodiment. The flowchart which shows the procedure of the conventional structure inspection. 10 is a flowchart showing a manufacturing process of a conventional semiconductor device. The figure which showed typically the pattern evaluation during image development. 10 is a flowchart showing a procedure of a pattern evaluation method according to the third embodiment. FIG. 6 is a diagram schematically illustrating how development proceeds. The figure for demonstrating the diffusion direction of the reaction product during image development. 24 is a cross-sectional view showing a layer structure in a film thickness direction created based on the model shown in FIG. The figure which shows distribution of the mixing ratio of a developing solution and a reaction product. The figure which shows distribution of the refractive index by the model of this embodiment. The figure which shows distribution of the extinction coefficient by the model of this embodiment. The figure which shows the model of the conventional image development. The figure which shows distribution of the refractive index by the conventional model. The figure which shows distribution of the extinction coefficient by the conventional model. 10 is a flowchart showing a procedure of a pattern evaluation method according to the fourth embodiment. The figure which shows the photosensitive resin pattern formation system concerning 5th Embodiment. 10 is a flowchart showing a procedure of a process condition determination method according to the fifth embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 101 ... Silicon substrate, 102 ... Antireflection film, 103 ... Unexposed resist film, 105 ... Developer, 201 ... Resist film, 202 ... Device pattern area, 203 ... Monitor pattern area, 211 ... Light source, 212 ... Observation optics System, 213 ... spectrometer, 214 ... computer, 215 ... database

Claims (7)

An optical inspection method for a substrate on which layers having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch are laminated,
A plurality of layers having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch are laminated, and there are one or more layers whose structure is determined and one or more layers whose structure is not determined. Preparing a substrate;
For each of the layers, a plurality of predicted values can be predicted by creating a structure within a range twice as large as the variation value of each structure expected in the process of forming each layer, with the target value of the structure of each layer as a central value. A step of creating a plurality of substrate structures including the structure of
A step of creating a substrate structure library A composed of a plurality of substrate structures combining the structures of the created layers;
When light is incident on the substrate structures included in the substrate structure library A at a specific angle, light intensity obtained by diffraction or reflection is calculated, and the calculated light intensity for each substrate structure is calculated. Creating a light intensity library B containing the chromatic dispersion of
Creating a plurality of new structures including the structure of the layer whose structure has been determined;
A substrate structure including a newly created structure is extracted from the substrate structures of the light intensity library B, and the extracted substrate structure and the wavelength dispersion of the light intensity corresponding to the extracted substrate structure are stored. Creating a light intensity library C;
Incident light at a specific angle on the prepared substrate and detecting chromatic dispersion of light intensity obtained by diffraction or reflection; and
By comparing the wavelength dispersion of the light intensity that is contained in the light intensity library C with the wavelength dispersion of the detected light intensity, the step of determining the structure of the substrate,
A structure inspection method comprising: An optical inspection method for a substrate on which layers having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch are laminated,
A plurality of layers having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch are laminated, and there are one or more layers whose structure is determined and one or more layers whose structure is not determined. Preparing a substrate;
For each of the layers, a plurality of predicted values can be predicted by creating a structure within a range twice as large as the variation value of each structure expected in the process of forming each layer, with the target value of the structure of each layer as a central value. Creating the structure of
A step of creating a substrate structure library A of the substrate from a combination of the structures of the respective layers created;
Based on the structure of the layer whose structure has been determined, newly creating a structure group of the layer whose structure has been determined;
Creating a substrate structure library B in which a substrate structure including the newly created structure group is extracted from the substrate structure of the substrate structure library A;
Predicting, by calculation , wavelength dispersion of light intensity obtained by diffracting or reflecting when light is incident on the substrate structure included in the substrate structure library B at a specific angle;
Incident light at a specific angle on the prepared substrate and detecting chromatic dispersion of light intensity obtained by diffraction or reflection; and
Comparing the chromatic dispersion of the detected light intensity with the chromatic dispersion of the predicted light intensity and determining the structure of each layer;
A structure inspection method comprising: A plurality of layers having a structure characterized by a predetermined optical constant, film thickness, ratio, and pitch are laminated, and there are one or more layers whose structure is determined and one or more layers whose structure is not determined. Preparing a substrate;
For each layer for which the structure has not been determined, the target value of the structure of each layer for which the structure has not been determined is used as a central value, and the variation value of the structure expected for the process of forming the layer for which the structure has not been determined. Creating a plurality of substrate structures including a plurality of predicted structures by creating a structure within twice the range ;
Creating a plurality of structures including the structure of the layer whose structure is determined;
A step of creating a substrate structure library B composed of a plurality of substrate structures obtained by combining the structures of the created layers;
Predicting the chromatic dispersion of the light intensity obtained by making light incident on the substrate structure included in the substrate structure library B at a specific angle and diffracting or reflecting by calculation;
A step of detecting the wavelength dispersion of the light intensity obtained by making light incident on the prepared substrate at a specific angle and diffracting or reflecting;
Comparing the chromatic dispersion of the detected light intensity with the chromatic dispersion of the predicted light intensity to determine the structure of each layer;
A structure inspection method comprising: Creating the structure of a layer whose structure has been determined
The structure inspection method according to claim 1, wherein a structure within a range twice the inspection error is created with the determined structure as a center value.   The structure inspection method according to claim 1, wherein the specific angle is plural.   The structural inspection method according to claim 1, wherein the incident light includes a plurality of wavelengths. The structural inspection method according to claim 1, wherein an RCWA (Rigorous coupled-wave analysis) method is used for the calculation of the light intensity.

Images