Server, semiconductor device manufacturing system and manufacturing method

Abstract

The purpose of the present invention is to provide a technique for detecting shape defects that cannot be obtained only by means of plane size data obtained from a top view, and accurately controlling etching parameters. Therefore, a server of the present invention controls an etching parameter so that a desired processing result of a semiconductor manufacturing apparatus can be obtained on the basis of a characteristic amount of an etching shape formed in a sample, a change amount of a target value, and data relating to the etching parameter of the semiconductor manufacturing apparatus. The feature amount uses a value obtained from secondary electron data from the surface of the sample or interference light data from the surface of the sample. The feature quantity sometimes is a value obtained from the size of the etched shape of the sample on the basis of secondary electron data or interference light data. Further, there is a spatially or temporally changing value as the amount of change in the feature amount.

Description

Technical Field

[0001] The present invention relates to a server, a semiconductor device manufacturing system and a manufacturing method. Background Art

[0002] At present, plasma etching systems are part of semiconductor manufacturing tools used to manufacture semiconductor devices close to the atomic level. One of the requirements in device manufacturing equipment is the improvement of uniformity. It is known that in order to manufacture semiconductor devices close to the atomic level, it is not sufficient to rely solely on the following: observe from the upper surface of the wafer to measure the lateral dimensions of the semiconductor device, and perform uniformization based on the measured data. For example, there is a CD-SEM in the above-mentioned measuring device, but only the geometric plane dimension data of the top view obtained from the CD-SEM can only obtain the surface dimensions of the pattern, so it is impossible to measure the bow shape and notch shape of the part slightly deeper from the surface of the pattern. Therefore, even with the same CD (Critical Dimension) value, it is impossible to detect the three-dimensional geometric defects of the pattern. If it cannot be detected like this, it is clear that defects will occur in the ion implantation process and CVD process after etching. In addition, these defective shapes are not only related to the temperature of the electrostatic chuck during etching, but also to the plasma density, so complex etching parameter control is required.

[0003] As shown in Patent Document 1, the prior art is the following technology: based on the geometric dimension data of the top view, i.e., the CD value, and according to the deviation of the CD value within the chip surface, the temperature value of the electrostatic chuck is fed back, thereby achieving a uniform CD value within the chip surface and improving the chip surface yield.

[0004] Prior Art Literature

[0005] Patent Literature

[0006] Patent Document 1: JP Patent No. 5925943 Summary of the invention

[0007] -Problems to be solved by the invention-

[0008] However, since the prior art is a technology for making CD values ​​uniform, it cannot cope with the problem that a bow or notch shape occurs in the depth direction despite showing the same CD value. In addition, since it is a technology for providing feedback on the temperature of the electrostatic chuck, it cannot cope with the problem of pattern shape caused by plasma density.

[0009] Therefore, the present invention aims to provide a technique for detecting shape defects that cannot be obtained only from plane dimension data obtained from a top view and for accurately controlling etching parameters.

[0010] -Methods for solving problems-

[0011] In order to solve the above-mentioned problems, one of the representative embodiments of the present invention is a server that controls etching parameters so as to obtain the desired processing result of the semiconductor manufacturing device based on the characteristic quantity of the etching shape formed on the sample and the change in the target value, and the correlation data of the etching parameters of the semiconductor manufacturing device, wherein the characteristic quantity is a value obtained from secondary electron data from the surface of the sample or interference light data from the surface of the sample.

[0012] -Effects of the Invention-

[0013] According to the present invention, it is possible to detect shape defects that cannot be obtained only by plane dimension data obtained from a top view, and to accurately control etching parameters. In addition, the user can optimize the etching process recipe that does not cause device defects without observing the cross section of the wafer when or before the device defects occur, and can non-destructively optimize the etching process recipe that does not cause device defects.

[0014] The problems, structures, and effects other than those described above will become clear from the description of the embodiments for carrying out the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the semiconductor device manufacturing system involved in the first embodiment.

[0016] Figure 2A This is a correlation diagram between the pattern shape and the secondary electron distribution when the acceleration voltage is high.

[0017] Figure 2B This is a correlation diagram between the pattern shape and the secondary electron distribution when the acceleration voltage is low.

[0018] Figure 3 It is a correlation diagram between pattern shape and secondary electron distribution under given etching parameters.

[0019] Figure 4 This is a flowchart of an example of a method for performing etching parameter control in the first embodiment.

[0020] Figure 5 This is a flowchart of an example of a method for performing etching parameter control in the second embodiment.

[0021] Figure 6 This is a correlation diagram between the pattern shape and the secondary electron distribution for a given number of etched wafers.

[0022] Figure 7 This is a flowchart of an example of a method for extracting a combination of feature quantities having a high correlation coefficient with respect to etching parameters.

[0023] Figure 8 It is a schematic diagram of a feature quantity distribution diagram in the fourth embodiment.

[0024] Fig. 9 This is a flowchart of an example of a method for performing etching parameter control in the fourth embodiment.

[0025] Fig.10 This is a correlation diagram between the characteristic quantity and the etching parameters. DETAILED DESCRIPTION

[0026] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In addition, the present invention is not limited to the embodiments. In the description of the accompanying drawings, the same reference numerals are used to indicate the same parts.

[0027] In one example of an embodiment of the present invention, there can be a server that calculates the temporal variation of the characteristic quantity or the spatial variation of the characteristic quantity based on the characteristic quantity obtained from the secondary electron distribution obtained from the top view of the wafer from the CD-SEM, or calculates the geometrical size difference (variation) between the characteristic quantity of the etching shape and the characteristic quantity of the target shape, and the integral value of the geometrical size difference (variation) in the etching depth direction, and performs etching parameter control in the semiconductor manufacturing device based on these related data. The following is an explanation based on the embodiment.

[0028] (First embodiment)

[0029] pass Figure 1 A first embodiment of the present invention will be described. Figure 1 1 is a schematic diagram of a semiconductor device manufacturing system according to the first embodiment. The semiconductor device manufacturing system 10 comprises: a device group including a semiconductor manufacturing device (etching device) 105, a CD-SEM PC 101, a semiconductor manufacturing device PC 102, a semiconductor inspection device 104 (for example, a CD-SEM); and a server 103. The purpose of this figure is to explain the performance improvement of the semiconductor device manufacturing system, which is important in semiconductor manufacturing.

[0030] The process recipe for performing etching processing is input to the semiconductor manufacturing apparatus PC 102, or transferred from the server 103 to the PC 102 and executed. After execution, the etching apparatus 105 transmits instruction values ​​such as the process recipe content to each device in the etching apparatus, and performs etching processing. Figure 1The various devices and components recorded are only the main parts, and the diagram shows a power supply for adjusting the temperature of the chip, a temperature adjustment electrode (heater / Peltier element), an RF power supply for applying high frequency, a microwave power supply for plasma generation, a plasma generation gas, a coil for generating a magnetic field for adjusting the plasma density, an electrostatic adsorption electrode (electrostatic chuck), and a lower electrode. The sample, i.e., the chip, is placed on the electrostatic adsorption electrode and etched by plasma treatment. The implemented processing recipe can be saved in the server 103 via the semiconductor manufacturing device PC102. Unless otherwise specified below, when it is said that the surface is inside, it means the surface of the sample (chip). In addition, it is assumed that the chip has a circular structure for explanation, and the shape of the chip has a circular shape, and the circle has a radial direction and a circumferential direction.

[0031] The etched wafer is transported to a semiconductor inspection device 104 (eg, CD-SEM). Figure 1 In the method, the recipe for CD-SEM is input into the CD-SEM PC 101, and a given measurement is performed. Here, electron beams are used in the measurement, and characteristic quantities are calculated based on the secondary electron quantity (secondary electron data) after electron beam irradiation. Here, as a representative characteristic quantity, a characteristic quantity obtained from the intensity distribution of the secondary electron quantity (secondary electron distribution) is used for explanation. As other characteristic quantities, there are various characteristic quantities such as characteristic quantities using illumination data, characteristic quantities obtained by differentiating the above-mentioned intensity distribution, etc. (sometimes characteristic quantities obtained from the secondary electron distribution of the etching shape of the sample are also referred to as distribution characteristic quantities). The measured results can be saved to the server 103 via PC 101.

[0032] Using the above-stored measurement data, the wafer yield is visualized based on the characteristic quantity distribution within the wafer surface, and if the obtained yield value is not satisfied, the processing recipe for etching processing is corrected again. It is believed that even if the CD value is the same, if there is a difference in the characteristic quantity, there will be a defect in the pattern shape of the device. The following is a further detailed description.

[0033] (Secondary electron distribution)

[0034] Figure 2A This is a correlation diagram between the pattern shape and secondary electron distribution when the acceleration voltage is high. Figure 2B This is a correlation diagram between the pattern shape and the secondary electron distribution when the acceleration voltage is low. Figure 2A and Figure 2BCollectively referred to as Figure 2. In Figure 2, a schematic diagram of the pattern shape of the device is recorded on the left, and a schematic diagram of the distribution of the secondary electron intensity (secondary electron distribution) is recorded on the right. The TopCD values ​​of the pattern shapes in Figure 2 are all the same. Since the difference in the secondary electron amount occurs due to the high and low acceleration voltages, there is an optimal acceleration voltage that makes it easy to extract the undesirable shape. Here, a bow shape is cited as an example of an undesirable shape to illustrate how the secondary electron distribution changes when the acceleration voltage is high and low, or which one is optimal. In addition, with respect to the pattern shape of the device, when it is a shape formed by etching, it is sometimes also referred to as an etched shape.

[0035] Figure 2A This is an explanation when the acceleration voltage is high (acceleration voltage: Vh1). (a-1) is the case of a bow shape (bad condition shape), and (a-2) is the case of a target shape. Since the penetration depth of the primary electrons of the acceleration voltage is deep, the amount of secondary electrons that seep out from the side walls of the pattern at the initial etching depth DL is small, and accordingly, the amount of secondary electrons is also small. For this reason, the characteristic value of the pattern shape when there is a bow shape with a reduced lateral width (the intensity of the secondary electrons at the center of the top surface of the pattern) and the characteristic value of the target shape with a normal lateral width (hereinafter, the characteristic value of the target shape is simply referred to as the target value) are both a, and there is no difference (variation), making it difficult to detect the bow shape.

[0036] When the etching is advanced to the etching depth DH of the final shape, as shown in (a-1), the amount of secondary electrons that seep out to the side wall of the pattern increases, and the value of the characteristic amount (secondary electron intensity at the center of the top surface of the pattern) increases to b. Therefore, since a value larger than the target value, i.e., c, is shown, it is easy to detect the bow shape. That is, in the case of ba>ca (i.e., b>c), it can be determined that there is a bow shape.

[0037] On the other hand, Figure 2B As shown, when the acceleration voltage is low (acceleration voltage: Vl1), the penetration depth of the primary electrons is shallow, so it is the depth DL at the beginning of etching. Even if the bow shape is small as shown in (b-1), the secondary electrons seep out more from the side wall of the pattern, so the value of the characteristic quantity (the secondary electron quantity intensity at the center of the top surface of the pattern) becomes larger to d. For this reason, the difference (variation) from the target value e, i.e., de, is large, and even at the beginning of etching, it becomes easy to detect the bow shape. Of course, when it is the final shape of the etching process, i.e., the etching depth DH, the difference (variation) between the characteristic quantity and the target value, i.e., fg, is also large, and at this time, it becomes easy to detect the bow shape.

[0038] As described above, since the ease of detecting a pattern shape defect changes by adjusting the acceleration voltage applied to primary electrons during electron beam irradiation, it is important to adjust the acceleration voltage to the optimum value.

[0039] As described above, since the distribution feature quantity also includes information on the etching depth direction of the etching shape, it is possible to obtain defects such as a bow shape in the etching depth direction that cannot be obtained from the plane dimension data obtained from the top view by simply calculating the difference in the distribution feature quantity in the wafer surface. In addition, the change in the feature quantity can be calculated by tilting the wafer using the tilt function in the CD-SEM, thereby calculating a more detailed change in the feature quantity.

[0040] (Feature quantities and etching parameters)

[0041] Next, the correlation between the characteristic amount and the etching parameter will be described. Figure 3 It is a correlation diagram between the pattern shape and the secondary electron distribution under given etching parameters. As examples, the pattern shape and feature values ​​at the edge of the wafer, and the pattern shape and feature values ​​at the center of the wafer are shown. In addition, the temperature environment and the plasma density environment are shown as etching parameters at each location. Here, as an example, the device at the center of the wafer is the pattern target shape, the temperature at the center of the wafer is T2, and the plasma density is P2. The device at the edge of the wafer is a bowed shape, which is a location that causes adverse conditions, and the temperature at the edge of the wafer is T1, and the plasma density is P1. Figure 3 In FIG. 1 , the vertical axis is set to z and the horizontal axis is set to B to show a diagram schematically showing the pattern shape of the edge of the wafer and the center of the wafer. L1 to L4 are film type switching lines. The oblique line portion represents the shape difference between the edge of the wafer and the center of the wafer. The area of ​​the shape difference portion (area difference) becomes ∫ΔBdz. Here, since there are two methods of optimizing the etching parameters by obtaining ∫ΔBdz or optimizing without obtaining ∫ΔBdz, the following is an explanation.

[0042] (Method 1)

[0043] Method 1 is a method of controlling etching parameters without determining ∫ΔBdz.

[0044] First, ∫ΔBdz can be represented using function Q as in Equation 1.

[0045] ∫ΔBdz=Q((fa)-(ga))…Equation 1

[0046] Next, if the electrostatic chuck temperature difference ΔT=T1-T2 and the plasma density difference ΔP=P1-P2, ∫ΔBdz can also be expressed using function R as in Formula 2.

[0047] ∫ΔBdz=R(ΔT,ΔP)…Equation 2

[0048] According to formula 1 and formula 2,

[0049] Through Q(fg) = R(ΔT, ΔP)

[0050] fg can be represented using function Z as shown in Equation 3.

[0051] fg=Z(ΔT,ΔP)…Equation 3

[0052] Here, ideally, since fg=0, the values ​​of T and P that satisfy Z(ΔT, ΔP)=0 may be obtained by using Formula 3.

[0053] (Method 2)

[0054] Method 2 is a method of determining ∫ΔBdz to control etching parameters.

[0055] First, CD measurement is performed in real time at L1 to L4 using OCD (Optical Critical Dimension) to obtain ∫ΔBdz. At this time, the measurement point in the depth direction can be set as the object for stacked films of different film types. In the case of the same film, the wafer can be cut using SEM or TEM for measurement. Of course, SEM and TEM can also be used for stacked films.

[0056] Then, ∫ΔBdz can be characterized as in Equation 4 using a given function R.

[0057] ∫ΔBdz=R(ΔT, ΔP)…Equation 4

[0058] Here, since ∫ΔBdz=0 is ideal, the values ​​of T and P that satisfy R(ΔT, ΔP)=0 may be determined using equation (4).

[0059] In addition, the derivation of the function described in the present disclosure is not particularly limited, and can be obtained inductively from experiments.

[0060] Since the first embodiment adopts the method 1, it is not necessary to obtain the geometrical size difference. On the other hand, the second embodiment described later adopts the method 2.

[0061] The method of controlling the etching parameters of method 1 will be described using a flow chart. Figure 4 This is a flowchart of an example of a method for performing etching parameter control in the first embodiment.

[0062] First, after the wafer is etched, N measurement points are selected. N is a number that is 1 more than the number of parameters to be measured. For example, in this embodiment, two points are selected, namely, the center of the wafer and the edge of the wafer. In this embodiment, in order to simplify the description, the plasma density is not changed, and the parameter to be measured is only the temperature. Therefore, the above two measurement points may also be used.

[0063] Next, the wafer temperature T and plasma density P at the selected position are measured, and characteristic quantities are also measured.

[0064] Next, the wafer temperature difference ΔT and the plasma density difference ΔP between the wafer center and the wafer edge are calculated, and also the difference in characteristic quantities, that is, the amount of change fg is calculated.

[0065] Next, referring to Formula 3, in order to calculate the value of function Z, set fg as the target, T2 at the center of the wafer as the reference point, and P2. As described above, if P1 at the edge of the wafer is set to be the same as P2, T1 at the edge of the wafer is calculated according to Formula 3. This becomes the temperature at the edge of the wafer that makes the target shape. As the etching parameter Ep that satisfies this T1, the temperature near the edge of the electrostatic chuck is adjusted. Considering the heat input balance from the plasma, the temperature near the edge of the electrostatic chuck is adjusted so that the temperature at the edge of the wafer becomes T1, and etching is performed.

[0066] In this embodiment, an example of optimizing etching parameters using the change in spatial characteristic quantity is shown. Here, the characteristic quantity calculated from the secondary electron distribution in CD-SEM is described as secondary electron data, but the same process is also used for the characteristic quantity calculated from the interference light distribution in OCD as interference light data.

[0067] (Second embodiment)

[0068] pass Figure 5 The second embodiment of the present invention is described below. The difference from the first embodiment is that the formula 4 in the above-mentioned method 2 is used. The rest is the same as the first embodiment, so the description will focus on the difference. Figure 5 This is a flowchart of an example of a method for controlling etching parameters in the second embodiment. The first embodiment does not require the geometrical size difference between the target shape and the defective shape to be obtained, but the second embodiment is an example of obtaining the geometrical size difference between the target shape and the defective shape.

[0069] First, after the wafer is etched, M points are selected as measurement points. M is a number that is 1 more than the number of parameters to be measured. For example, in this embodiment, two points are selected, namely, the center of the wafer and the edge of the wafer. In this embodiment, in order to simplify the description, the plasma density is not changed, and the parameter to be measured is only the temperature. Therefore, the above two points can also be used as measurement points.

[0070] Next, the chip temperature T and plasma density P at the selected position are measured. In addition, the pattern shape is also measured using cross-sectional SEM, secondary electron data from TEM, or interference light data from OCD to obtain characteristic quantities related to geometric dimensions (characteristic quantities directly obtained from the geometric dimensions of the pattern shape are sometimes referred to as shape characteristic quantities).

[0071] Next, the wafer temperature difference ΔT and the plasma density difference ΔP between the wafer center and the wafer edge are calculated, and the area difference ∫ΔBdz is calculated. In order to calculate the area difference, there are the following methods: using cross-sectional SEM or TEM to cut the wafer and measure the shape feature quantity related to the size of the pattern shape; or in order to obtain Figure 3 The average pattern width of each film type L1 to L4 described in the above is used to adjust the light source of the OCD to obtain the shape feature related to the size of the pattern shape. Then, the geometric size difference (ΔB) between the shape feature in the pattern shape and the target value, i.e., the change amount, is integrated in the etching depth direction to calculate the area difference ∫ΔBdz.

[0072] Next, referring to Formula 4, in order to calculate the value of function R, set the target ∫ΔBdz, T2 at the center of the wafer as the reference point, and P2. As described above, if P1 at the wafer edge is set to be the same as P2, T1 at the wafer edge is calculated according to Formula 4. This becomes the temperature at the wafer edge that makes the target shape. As the etching parameter Ep that satisfies this T1, the temperature near the edge of the electrostatic chuck is adjusted. Considering the heat input balance from the plasma, the temperature near the edge of the electrostatic chuck is adjusted so that the temperature at the wafer edge becomes T1, and etching is performed.

[0073] According to the above, it is possible to perform etching parameter control that detects and eliminates problems such as bowing in the etching depth direction that cannot be obtained only from planar dimensional data obtained from a top view.

[0074] In this embodiment, an example is shown in which the size of the pattern shape is directly obtained using secondary electron data or interference light data to determine the shape feature value, thereby optimizing the etching parameters.

[0075] (Third embodiment)

[0076] pass Figure 6A third embodiment of the present invention will be described. Figure 6 It is a correlation diagram between the pattern shape and the secondary electron distribution under a given number of etching process wafers. As an example, the pattern shape and feature values ​​at the center of the wafer under the etching process wafers No. 1, 100, and 1000 are shown, and the temperature environment and plasma density environment at each time are also shown. Here, as an example, the etching process wafer No. 1 is the target shape, the temperature at the center of the wafer is T1, and the plasma density is P1. The etching process wafer No. 100 is a bowed shape, causing an undesirable condition, the temperature at the center of the wafer is T100, and the plasma density is P100. In addition, the etching process wafer No. 1000 is a bowed shape, causing an undesirable condition, the temperature at the center of the wafer is T1000, and the plasma density is P1000. In Figure 6 , the vertical axis is set to z, and the horizontal axis is set to B to show a diagram schematically showing the pattern shapes of the first wafer and the 1000th wafer. L1 to L4 are film type switching lines. The oblique line portion represents the shape difference between the first wafer and the 1000th wafer. The area of ​​the shape difference portion is ∫ΔBdz. Here, as in the first embodiment, a method (method 1) for optimization without obtaining ∫ΔBdz is used.

[0077] To explain the above process, we use Figure 4 First, after the wafer is etched, N measurement points are selected. N is a number that is 1 more than the number of parameters to be measured. For example, in this embodiment, as an example, the two points are the first wafer and the 1000th wafer. In this embodiment, in order to simplify the description, the plasma density is not changed, and the parameter to be measured is only the temperature. Therefore, the above two measurement points can also be used.

[0078] Next, the wafer temperature T and plasma density P at the selected time point are measured, and characteristic quantities are also measured.

[0079] Next, the wafer temperature difference ΔT and the plasma density difference ΔP between the first wafer and the 1000th wafer are calculated, and also the difference in characteristic quantities, that is, the variation fg is calculated.

[0080] Next, referring to Formula 3, in order to calculate the value of function Z, set fg as the target, T1 and P1 of the first wafer as the reference point. As described above, if P1000 of the 1000th wafer is set to be the same as P1, T1000 of the 1000th wafer is calculated according to Formula 3. This becomes the temperature of the 1000th wafer that makes the target shape. As the etching parameter Ep that satisfies this T1000, the temperature near the center of the electrostatic chuck is adjusted here. Considering the heat input balance from the plasma, the temperature near the center of the electrostatic chuck is adjusted in each time series so that the temperature of the center of the wafer becomes T1000, and etching is performed.

[0081] In this embodiment, an example of optimizing etching parameters using the variation of characteristic quantities over time is shown. Here, the characteristic quantities calculated from the secondary electron distribution in CD-SEM are described, but the same process is also used for the characteristic quantities calculated from the interference light distribution in OCD.

[0082] (Fourth embodiment)

[0083] pass Figure 7 A fourth embodiment of the present invention will be described. Figure 7 This is a flowchart of an example of a method for extracting a combination of feature quantities having a high correlation coefficient with respect to etching parameters.

[0084] First, etching parameters Ep are set and etching processing is performed.

[0085] Next, the shape feature is measured after etching. The shape feature takes into account various feature quantities such as etching rate, CD value, surface roughness, etc. The number of measurement points is more than one.

[0086] In addition, the distribution feature is measured after etching. The feature is not only the secondary electron intensity at the center of the top surface of the pattern obtained from the secondary electron distribution described so far, but also various feature (feature 1 / feature 2 / feature 3, etc.) are considered. There are more than one measurement point.

[0087] Next, the in-plane distribution of the feature quantity is visualized to obtain the unevenness. For this purpose, as an example, the unevenness is calculated as (maximum value - average value) / average value × 100 (%) using the maximum value of the deviation from the average data of the data related to the measured feature quantity. Then, the feature quantity whose unevenness exceeds a given ratio (%) is extracted (for example, for shape feature quantities, the number of extractions when the given unevenness is set to Q% is n, and for distribution feature quantities, the number of extractions when the given unevenness is set to R% is m).

[0088] Next, the distribution within the plane is normalized for each extracted feature quantity. Here, as an example, the maximum value of each feature quantity is normalized to 1.

[0089] Next, the normalized data of the shape feature quantity in-plane distribution and the normalized data of the distribution feature quantity in-plane distribution are compared to calculate the correlation coefficient. Figure 7 The relationship between the normalized data and the measured position (x) of the wafer is shown in FIG. Here, as an example, in the case of the etching rate and the feature value 3, the normalized data is overlapped to obtain the area (area difference) of the different part. As an example of a formula for calculating the correlation coefficient based on the area difference ratio relative to the integral value of the etching rate normalized data, the area difference ratio (%) relative to the integral value of the etching rate normalized data 100-etching rate normalized data = correlation coefficient is cited. It is calculated for the total number of combinations n×m, and the combination greater than the correlation coefficient L (%) is extracted and stored in the database (DB1) in the server.

[0090] Thus, relevant data related to high feature quantities related to etching parameter Ep can be accumulated in DB 1. Thus, control can be performed so that relevant high etching parameters are retrieved from given feature quantities from DB 1 by machine learning or other methods, and the in-plane distribution becomes uniform.

[0091] Describe actual live use cases. Figure 8 This is a schematic diagram of a feature quantity distribution diagram in the fourth embodiment. A distribution diagram of feature quantities 1 to 3 is shown as feature quantities, but among them, feature quantity 3, which has a difference in the in-plane distribution of feature quantities (high non-uniformity), is focused on. Since high etching parameters related to feature quantity 3 are stored in the above-mentioned DB1, a search is performed to accurately adjust the etching parameters so that feature quantity 3 becomes uniform (search in DB). In addition, the optimal etching parameters can be extracted from past experimental data and manually set.

[0092] The above operation is explained using a flowchart. Fig. 9 This is a flowchart of an example of a method for performing etching parameter control in the fourth embodiment.

[0093] First, the in-plane distribution of the characteristic quantity after etching the wafer 1 is normalized. That is, the maximum characteristic quantity value is set to 1.

[0094] Next, the etching parameter Ep1 with a high correlation coefficient is determined from the above DB1 using the above-mentioned feature quantity in-plane distribution. Machine learning can also be used when determining. In the case where there is a problem due to intentional differences (variation) in space or time in the feature quantity, the etching parameter Ep1 is corrected to a given value, and the wafer 2 is etched again. The same process is then repeated, and when there is no problem in the feature quantity, the desired etching process recipe is obtained. Therefore, the target shape can be achieved by accurately controlling the etching parameters for the feature quantity with the problem of uneven in-plane distribution.

[0095] exist Fig.10 A method for obtaining etching parameters more efficiently is shown. Fig.10 It is a correlation diagram between the characteristic quantity and the etching parameters. It shows an example of a correlation diagram between the etching parameters and the characteristic quantity 3 (for example, the secondary electron intensity at the center of the pattern top surface) at the center of the chip and the edge of the chip. During etching, in most cases, the gas pressure at the periphery of the chip is lower, and the adhesion rate of the reaction products is lower than that at the center. For this reason, the thickness of the reaction products on the sidewalls of the pattern sometimes differs between the center and the periphery of the chip. In addition, it is believed that the difference in the characteristic quantity occurs due to the different shapes and roughness of the sidewalls. Therefore, even if the same shape is shown, the amount of secondary electrons emerging from the sidewalls will be different between the center and the periphery of the chip due to the different materials and surface roughness inside the sidewalls of the pattern. Fig.10 In the figure, when the characteristic quantity 3 is f, the etching parameters Ep1 and Ep2 are different at the center of the wafer and at the edge of the wafer. If the etching parameter Ep1 remains unchanged, the characteristic quantity difference fg will occur, which becomes a problem. In order to make the edge of the wafer closer to the center, the etching parameter Ep1 at the edge of the wafer can be set to Ep2.

[0096] As mentioned above, although embodiment of this invention is described, this invention is not limited to the said embodiment, Various changes can be made within the range which does not deviate from the summary of this invention.

[0097] For example, in the embodiment, a flow is described in which etching parameter control is performed by visualizing, normalizing, and calculating correlation coefficients of feature quantity distribution on a wafer to eliminate feature quantity differences. The visualization method, normalization method, and correlation coefficient calculation method described here are examples.

[0098] In addition, in the above embodiments, the server is described as a component for performing etching parameter control, but it may not be a server, but a virtual area, a PC terminal, or a mobile terminal, and the communication method is not limited to SMB communication, but has the same effect in FTP communication, NFS communication, etc.

[0099] In the above embodiments, an etching apparatus is described as an example of a semiconductor manufacturing apparatus. However, the same effects can be achieved with other manufacturing apparatuses such as a plasma CVD apparatus, an ashing apparatus, a surface modification apparatus, and the like.

[0100] Furthermore, in the above embodiments, CD-SEM is described as an example of a semiconductor inspection apparatus, but other inspection apparatuses such as OCD, TEM, and XRR can also have the same effects.

[0101] The following describes the embodiments of the present invention, but the present invention is not limited thereto.

[0102] (Method 1)

[0103] A server, characterized in that etching parameters are controlled so that the desired processing result of the semiconductor manufacturing device is obtained based on the characteristic quantity of the etching shape formed on the sample and the change in the target value, and the correlation data of the etching parameters of the semiconductor manufacturing device, wherein the characteristic quantity is a value obtained from secondary electron data from the surface of the sample or interference light data from the surface of the sample.

[0104] (Method 2)

[0105] In the server described in aspect 1, the feature amount is a value obtained from the size of the etching shape of the sample based on the secondary electron data or the interference light data.

[0106] (Method 3)

[0107] In the server described in method 1 or 2, it is characterized in that the change amount is a value of a spatial or temporal change.

[0108] (Method 4)

[0109] In the server according to any one of aspects 1 to 3, the correlation data includes a value obtained from an integral of the amount of change related to the dimensional difference between the etching shape and the target shape in the etching depth direction.

[0110] (Method 5)

[0111] In the server described in aspect 4, the correlation data includes a difference between an etching parameter related to the etching shape and an etching parameter related to the target shape, and the integrated value.

[0112] (Method 6)

[0113] In the server described in aspect 1, when the feature amount is obtained from the secondary electron data, the feature amount is adjusted by an accelerating voltage of an electron gun of a CD-SEM.

[0114] (Method 7)

[0115] In the server according to any one of aspects 1 to 6, the sample is formed into a plurality of films.

[0116] (Method 8)

[0117] In the server described in Mode 4, it is characterized in that, when the amount of change is a value of a spatial change, the amount of change is a value of a change in the radial direction of the sample or in the circumferential direction of the sample.

[0118] (Method 9)

[0119] In the server described in any one of aspects 1 to 8, the correlation data includes a correlation coefficient for a given etching parameter calculated from values ​​obtained by visualizing and normalizing the in-plane distribution of the sample in a plurality of feature quantities.

[0120] (Method 10)

[0121] The server according to aspect 9 is characterized in that the server includes a database storing data related to combinations of the feature quantities having high correlation coefficients and etching parameters.

[0122] (Method 11)

[0123] In the server described in mode 10, a combination of the feature quantity and etching parameter having a high correlation coefficient is determined from the database, and the etching parameter is corrected when the feature quantity is different from the target value.

[0124] (Method 12)

[0125] In the server described in mode 11, it is characterized in that the etching parameters are controlled by controlling the temperature of the sample.

[0126] (Method 13)

[0127] A semiconductor device manufacturing system, characterized by comprising a server described in any one of Modes 1 to 12.

[0128] (Method 14)

[0129] A method for manufacturing a semiconductor device, characterized in that it has the following steps: obtaining a characteristic value from secondary electron data from the surface of a sample or interference light data from the surface of the sample; obtaining data related to the characteristic value of an etched shape formed on the sample and the change in the target value and the etching parameters of a semiconductor manufacturing device; and controlling the etching parameters so that the desired processing result of the semiconductor manufacturing device can be obtained based on the related data.

[0130] -Explanation of symbols-

[0131] 10...semiconductor device manufacturing system, 101...PC for CD-SEM, 102...PC for semiconductor manufacturing apparatus, 103...server, 104...semiconductor inspection apparatus, 105...semiconductor manufacturing apparatus.

Claims

1. A server, characterized in that: The etching parameters are controlled so that the desired processing result of the semiconductor manufacturing device is obtained based on the characteristic amount of the etching shape formed on the sample and the change amount of the target value and the correlation data of the etching parameters of the semiconductor manufacturing device, The feature amount is a value obtained from secondary electron data from the surface of the sample or interference light data from the surface of the sample.

2. The server according to claim 1, wherein: The feature amount is a value obtained from the size of the etching shape of the sample based on the secondary electron data or the interference light data.

3. The server according to claim 1 or 2, wherein: The variation is a value of a spatial or temporal variation.

4. The server according to claim 2, wherein: The correlation data includes a value obtained from an integral value in an etching depth direction of the amount of change related to a dimensional difference between the etching shape and a target shape.

5. The server according to claim 4, wherein: The correlation data includes a difference between an etching parameter related to the etching shape and an etching parameter related to the target shape, and the integrated value.

6. The server according to claim 1, wherein: When the feature amount is obtained from the secondary electron data, the feature amount is adjusted by the acceleration voltage of the electron gun of the CD-SEM.

7. The server according to claim 1 or 2, wherein: The sample is formed into a plurality of films.

8. The server according to claim 3, wherein: When the amount of change is a value of a spatial change, the amount of change is a value of a change in the radial direction of the sample or in the circumferential direction of the sample.

9. The server according to claim 1, wherein: The correlation data includes a correlation coefficient for a given etching parameter calculated from values ​​obtained by visualizing and normalizing the in-plane distribution of the sample in a plurality of feature quantities.

10. The server according to claim 9, wherein: The server includes a database storing data related to combinations of feature quantities and etching parameters having high correlation coefficients.

11. The server according to claim 10, wherein: A combination of the feature quantity and etching parameter having a high correlation coefficient is determined from the database, and when the feature quantity is different from the target value, the etching parameter is corrected.

12. The server according to claim 11, wherein: The etching parameters are controlled by controlling the temperature of the sample.

13. A semiconductor device manufacturing system, characterized in that: A server according to claim 1 is provided.

14. A method for manufacturing a semiconductor device, characterized in that: The process is as follows: obtaining a characteristic quantity from secondary electron data from a surface of a sample or interference light data from a surface of the sample; Obtaining data related to the characteristic quantity of the etching shape formed on the sample, the change quantity of the target value, and the etching parameters of the semiconductor manufacturing device; and The etching parameters are controlled so that a desired processing result of the semiconductor manufacturing apparatus can be obtained based on the correlation data.

Images