JP2008112889A - Focus measuring method, method of manufacturing semiconductor device, and exposure system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a focus measuring method which can accurately measure a focus value and an amount of focus deviation, and to provide a method of manufacturing a semiconductor device and an exposure system using the same. <P>SOLUTION: The focus value is found by using the dependency of an LER value of a resist pattern on a focus value. The dependency of an LER value of each pattern on a focus value is acquired in advance. In finding the focus value, an LER value of a resist pattern for measuring a focus which is formed by exposing an object for which the focus value is to be measured is measured. The focus value for the LER value can be found from the already acquired dependency of an LER value on a focus value. The amount of focus deviation is a difference between the focus value and the best focus value. By employing a line pattern and a space pattern as the resist pattern for measuring a focus value, the direction of the focus deviation can also be found. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method for measuring an exposure focus in a lithography process, a method for manufacturing a semiconductor device using the method, and an exposure system.

  Conventionally, an exposure apparatus (stepper or the like) is used to form a fine element pattern of a semiconductor integrated circuit. The exposure apparatus performs projection exposure for forming an image of a mask pattern on a reticle on a photosensitive resist film formed on a semiconductor substrate. By developing the exposed resist film, the projected mask pattern is patterned. In order to obtain a resist pattern having a predetermined cross-sectional shape and dimensions, it is necessary to project the mask pattern on the reticle onto the photoresist film without defocusing. The inventors of the present application have proposed a method of measuring a focus shift described in Patent Document 1 as a method for confirming the presence or absence of a focus shift during the manufacturing process of the semiconductor device.

  In this method, an isolated line pattern and an isolated space pattern made of a resist film are formed, and a focus shift amount is detected by using the focus dependency of these pattern dimensions. FIG. 8 is a cross-sectional view showing the isolated line pattern and the isolated space pattern. In this method, the top dimension 101 of the isolated line pattern 100, the bottom dimension 102 of the isolated line pattern 100, the top dimension 103 of the isolated space pattern 110, and the bottom dimension 104 of the isolated space pattern 110 are measured. The dimension measured in the isolated line pattern 100 is the line width, and the dimension measured in the isolated space pattern 110 is the space width.

  FIG. 9 is a diagram illustrating the focus dependency of each of the dimensions 101 to 104. FIG. 9A shows the focus dependency of the isolated line pattern 100, and FIG. 9B shows the focus dependency of the isolated space pattern 110. The horizontal axis of FIG. 9A indicates the focus value, and the vertical axis indicates the normalized inclination amount ΔLn of the isolated line pattern 100. The normalized inclination amount ΔLn is a difference ΔLn = ΔL−ΔLo between the inclination amount ΔL, which is a value obtained by subtracting the bottom dimension 102 from the top dimension 101, and the inclination amount ΔLo at the time of best focus. In FIG. 9B, the horizontal axis indicates the focus value, and the vertical axis indicates the normalized inclination amount ΔSn of the isolated space pattern 110. The normalized inclination amount ΔSn is a difference ΔSn = ΔS−ΔSo between the inclination amount ΔS, which is a value obtained by subtracting the bottom dimension 104 from the top dimension 103, and the inclination amount ΔSo at the time of best focus. In the graphs shown in FIGS. 9A and 9B, the focus value at which the vertical axis is zero may be considered as the best focus value in the patterns 100 and 110, respectively.

  FIG. 10 is a diagram showing the focus dependency of the sum of the normalized inclination amounts ΔLn and ΔSn (hereinafter referred to as a shift index) shown in FIG. Here, the horizontal axis of FIG. 10 indicates the focus value, and the vertical axis indicates the shift index.

As shown in FIG. 10, the shift index usually shows a constant change with respect to the focus value. In the relationship shown in FIG. 10, the focus value at which the shift index is zero can be regarded as the best focus value that satisfies both the isolated line pattern 100 and the isolated space pattern 110. Therefore, before the pattern is exposed in the exposure process of the semiconductor device, first, the top dimensions 101 and 103 and the bottom dimensions 102 and 104 of the isolated line pattern 100 and the isolated space pattern 110 are measured, and the results are calculated. The focus dependence of the shift index shown in FIG. 10 is acquired in advance. Then, the top dimensions 101 and 103 and the bottom dimensions 102 and 104 of the isolated line pattern 100 and the isolated space pattern 110 formed during actual pattern exposure are measured and compared with this graph. Thereby, it is possible to easily calculate how much the actual focus value at the time of exposure deviates from the best focus value. This method can detect the presence / absence of focus deviation and the amount of focus deviation easily and accurately, and is very useful as a focus measurement method.
JP 2005-12158 A

  However, in recent years, semiconductor integrated circuit element patterns and resist patterns have been miniaturized. In the formation of such a fine resist pattern (for example, a resist pattern formed by a process technology node of 90 nm or less), LER (Line Edge Roughness) becomes a size that cannot be ignored relative to the dimension of the resist pattern. . Under such circumstances, it is difficult to perform accurate resist dimension measurement, and the focus value and the amount of focus deviation cannot be evaluated with high accuracy even using the method of Patent Document 1. Here, LER refers to the degree of unevenness of the pattern edge in plan view of the formed resist pattern. In the LER, a single molecule or a lump of molecules (molecular cluster) of the resist material constituting the resist pattern appears as irregularities on the end face of the pattern. LER has become a problem since it has become apparent in the manufacturing process of semiconductor devices that use KrF excimer laser light or ArF excimer laser light in the far ultraviolet region as a light source of an exposure apparatus.

  In order to eliminate the influence of LER on dimension measurement, the number of measurement points of pattern width and pattern interval is increased and the average value is used as a measurement dimension, or in order to quantify LER itself, position variation with respect to a certain end face It is conceivable to use a method of calculating and feeding back to the dimension measurement result. However, these methods still have problems in measuring the focus value and the amount of focus deviation due to the fact that enormous measurement time is required or the reliability of data is not perfect.

  The present invention has been proposed in view of the above-described conventional circumstances, and a focus measurement method capable of accurately measuring a focus value and a focus shift amount even in a fine resist pattern having a relatively large LER. An object of the present invention is to provide a method of manufacturing a semiconductor device using the same and an exposure system.

  In order to achieve the above object, the present invention employs the following technical means. That is, in the focus measurement method according to the present invention, the resist film formed on the substrate is exposed with different focus values using an exposure apparatus, and the LER (Line Edge) of the focus measurement resist pattern formed by each exposure is obtained. The correspondence relationship between the (Roughness) value and the focus value is acquired in advance. Next, the exposure apparatus is used to expose the resist film formed on the substrate as a focus measurement target, and the LER value of the focus measurement resist pattern formed by the exposure is measured. Based on the correspondence, a focus value corresponding to the LER value of the focus measurement resist pattern formed by the exposure of the focus measurement target is obtained.

  In addition to the above-described configuration, it is also possible to obtain a focus shift amount in exposure of the focus measurement target from the obtained focus value and best focus value. The best focus value can be, for example, a focus value at which the LER value is minimized in the correspondence relationship.

  The focus measurement resist pattern may include a line pattern and a space pattern. In this case, the correspondence relationship is acquired for each of the line pattern and the space pattern. The focus value in the exposure of the focus measurement target is formed by the correspondence between the LER value of the line pattern and the focus value, the correspondence between the LER value of the space pattern and the focus value, and the exposure of the focus measurement target. This is obtained based on the LER value of the line pattern and the LER value of the space pattern formed by the exposure of the focus measurement target.

  As the resist pattern, it is preferable to adopt a pattern selected from a plurality of types of patterns having different shapes and having a maximum amount of change in the LER value with respect to the change in focus value of the exposure apparatus. The resist pattern preferably has a focus margin that is substantially the same as an amount of focus deviation that can occur in the exposure apparatus. Furthermore, it is preferable that the resist film has a film thickness that maximizes the intensity of the standing wave generated in the resist film in the exposed portion during the exposure of the focus measurement target. When the resist film is made of a resist material whose LER value varies depending on the temperature of baking performed after exposure and before development, the baking is preferably performed at a temperature at which the LER value is maximized.

  On the other hand, in another aspect, the present invention can also provide a method of manufacturing a semiconductor device using the focus measurement method. That is, in the method for manufacturing a semiconductor device according to the present invention, the resist film on the semiconductor substrate is exposed to form a resist pattern including a focus measurement resist pattern. Next, the LER value of the focus measurement resist pattern is measured. The focus value corresponding to the LER value is obtained based on the correspondence relationship between the LER value of the focus measurement resist pattern and the focus value of the exposure apparatus acquired in advance. From the obtained focus value and the best focus value, a focus shift amount in the exposure is obtained. It is determined whether or not the focus shift amount is within a preset allowable range. If the focus shift amount is not within the allowable range, the exposed semiconductor substrate is determined to be defective.

  The focus measurement resist pattern may include a line pattern and a space pattern. In this case, the focus value in the exposure includes the correspondence between the LER value and the focus value of the line pattern, the correspondence between the LER value and the focus value of the space pattern, and the line pattern formed by the exposure of the focus measurement target. And the LER value of the space pattern formed by the exposure of the focus measurement target. Further, it is possible to adopt a configuration in which the subsequent exposure is performed after adjusting the focus position of the exposure apparatus to the best focus position in accordance with the focus shift amount.

  In still another aspect, the present invention can also provide an exposure system that realizes the method for manufacturing the semiconductor device. The exposure system includes an exposure device that images and transfers a mask pattern onto a resist film. In the exposure system according to the present invention, the LER value calculation unit obtains the LER value from the dimension of the resist pattern for focus measurement formed by the exposure by the exposure apparatus measured by the dimension measuring apparatus. The focus value calculation unit corresponds to the LER value of the focus measurement resist pattern formed by the exposure based on the correspondence relationship between the LER value of the focus measurement resist pattern and the focus value of the exposure apparatus acquired in advance. Find the focus value. The focus deviation calculation unit obtains the focus deviation amount in the exposure from the obtained focus value and best focus value. Then, the quality determination unit determines whether or not the focus shift amount is within a preset allowable range. The exposure system may be configured to adjust the focus position of the exposure apparatus to the best focus position according to the focus shift amount.

  According to the present invention, an accurate focus value and focus shift amount can be obtained with a small amount of data even in a fine pattern with a relatively large LER. Then, by performing focus adjustment for each exposure or for each exposure unit (lot) according to the focus shift amount, it is possible to suppress an increase in focus shift and form a highly accurate resist pattern. As a result, a semiconductor device can be formed with a high manufacturing yield.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. In the present invention, the LER that is manifested in the resist pattern is digitized and the focus measurement of the exposure apparatus is performed using the digitized LER.

  First, digitization of LER will be described. The digitization of LER is performed by digitizing the unevenness formed on the edge of the resist pattern. FIG. 1 is an SEM photograph showing a state in which a fine resist pattern in which LER has become apparent is observed from above by a length measurement SEM (Secondary Electron Microscopy). The resist pattern is an isolated line pattern formed on the semiconductor substrate.

  In FIG. 1A and FIG. 1B, a region A indicated by a broken line indicates a region where measurement for quantifying LER is performed. Further, the white point indicates the position of the edge where the measurement is performed. The edge of the resist pattern 11 is automatically detected by, for example, the edge detection function of the length measurement SEM. In addition, for each measurement position, for example, at the detected edge, locations at equal intervals along the length direction of the resist pattern 11 are automatically set by the length measurement SEM. The dimension measurement for the resist pattern is generally performed on the upper edge or the bottom edge of the resist pattern. Here, the edge of the upper surface of the resist pattern is the measurement target.

  As a method of quantifying LER, as shown in FIG. 1A, a method of quantifying the unevenness at one edge of the resist pattern 11 (here, an isolated line pattern), and as shown in FIG. There is a method of quantifying the unevenness at both edges of the resist pattern 11. For example, in FIG. 1A, the distance between each measurement position and a specific straight line L1 along the length direction of the resist pattern 11 is measured, and LER as the variation in the distance (hereinafter referred to as edge position variation). Is digitized. In FIG. 1B, the pattern width (pattern width in a direction orthogonal to a specific straight line along the length direction of the resist pattern) at each measurement position is measured, and LER is quantified as the variation in the pattern width. To do. In any method, the variation in the measured value can be quantified as a range R that is a width between the maximum value and the minimum value of the measured value, a standard deviation σ of the measured value, or the like. Since the number of measurement points increases as the length L along the length direction of the resist pattern 11 included in the region A increases, the reliability of the digitized LER (hereinafter referred to as LER value) is improved. It will be.

  FIG. 2 is a diagram showing the dependency of the LER value measured in the resist pattern 11 shown in FIG. 1 on the focus value (hereinafter referred to as focus dependency). In FIG. 2, the horizontal axis corresponds to the focus value, and the vertical axis corresponds to the LER value. As described above, the LER value of the resist pattern 11 can be expressed by edge position variations and pattern width variations. In this embodiment, 3σ (a value obtained by multiplying the standard deviation by 3) of the edge position variation on the upper surface of the resist pattern is set as the LER value. In the following, unless otherwise specified, the LER value indicates 3σ of edge position variation.

  The best focus value of the resist pattern 11 is −0.2 μm. The best focus value is a value obtained by the above-described conventional method. When obtaining the best focus value, the dimension of the upper surface of the line pattern and the dimension of the bottom surface of the line pattern are measured at a large number of measurement points in order to eliminate the influence of the unevenness of the resist pattern edge. For example, the dimension of the upper surface of the pattern is measured by measuring the edge position at each measurement point shown in FIG. 1 (b), and calculating the average value (or median value, maximum value and minimum value) of all edge positions in the X direction. The reference lines L2 and L3 in the length direction of the resist pattern determined by the intermediate value or the like are calculated and calculated as the interval between the reference lines L2 and L3. Since the focus dependence of the shift index shown in FIG. 10 is acquired by such measurement, it takes a lot of time to acquire the best focus value.

  As shown in FIG. 2, the LER value is minimal (minimum) when the best focus value obtained by the conventional method, that is, the focus value is −0.2 μm. The LER value tends to increase as the focus value shifts to the plus side or the minus side. Here, the shift to the plus side is a state in which the focus position (focal plane) of the exposure apparatus moves away from the substrate, and the shift to the minus side refers to a direction in which the focus position of the exposure apparatus approaches the substrate. It is a state to move to. The focus value is a parameter set in the exposure apparatus in order to control the amount of movement of the focus position.

  From FIG. 2, it can be understood that when the resist film is exposed with different focus values and the LER values of the formed resist patterns are obtained, the focus value at which the LER value is minimized becomes the best focus value. The fact that the LER value and the focus value have a certain correspondence is a fact newly found by the present inventors. By utilizing such a relationship between the LER value and the focus value, it is possible to easily detect the presence or absence of a focus shift in the exposure process for forming the semiconductor device.

  In this method, the focus dependency (FIG. 2) of the LER value of a resist pattern for measuring a focus shift (here, an isolated line pattern) is acquired in advance. In order to acquire the focus dependency, the resist film formed on the semiconductor substrate is exposed with different focus values using a mask including a focus deviation measurement pattern. Then, the LER value of the resist pattern corresponding to each exposure obtained by developing the exposed resist film is measured, and the focus dependency of the LER value is acquired.

  Next, the focus deviation measurement target exposure is performed on the semiconductor substrate on which the resist film is formed, and the LER value of the focus measurement resist pattern (hereinafter referred to as the measurement target resist pattern) formed by the exposure is measured. Subsequently, a focus value corresponding to the measured LER value is obtained based on the focus dependency of the LER value acquired in advance.

  At this time, if the obtained focus value matches the best focus value in the focus dependency of the LER value acquired in advance, no focus shift has occurred in the exposure of the focus shift measurement target. Further, when the obtained focus value is different from the best focus value in the focus dependence of the LER value acquired in advance, it means that a focus shift has occurred in the exposure of the focus shift measurement target.

  For example, in the example of FIG. 2, if the focus value corresponding to the LER value measured in the measurement target resist pattern is −0.2 μm, it can be determined that the exposure for forming the resist pattern was the best focus. If the focus value corresponding to the LER value measured in the measurement target resist pattern is a value different from −0.2 μm, such as −0.1 μm or −0.4 μm, exposure for forming the resist pattern is performed. It can be determined that a focus shift has occurred.

  In the above description, the LER value is 3σ of the edge position variation on the upper surface of the line pattern. However, the standard deviation σ and the range R of the edge position variation can be used as the LER value. The measurement position for acquiring the LER value is not limited to the upper surface of the line pattern, but may be the bottom surface of the line pattern. Further, it is not essential that the resist pattern is an isolated line pattern, and any resist pattern having an edge capable of acquiring the LER value may be used. For example, an isolated space pattern can be used. In this case, the LER value can be obtained by measuring the edge position variation on the top and bottom surfaces of the isolated space pattern and the space width variation. Further, not only a line pattern or a space pattern but also a dot pattern or a hole pattern can be used. In any case, the focus dependency of the LER value is minimized when the best focus is achieved.

  By the way, in the above-described method, the difference between the best focus value and the focus value corresponding to the LER value of the measurement target resist pattern is the focus shift amount. For this reason, it is possible to determine whether or not the focus shift amount is within a preset allowable range. However, the focus shift direction cannot be identified only by the focus dependency of the LER value shown in FIG. For this reason, the adjustment direction when performing the focus adjustment of the exposure apparatus based on the focus shift amount is unknown. Therefore, in this embodiment, the defocus direction is detected by measuring the LER values of a plurality of resist patterns having different shapes.

  FIG. 3 is a diagram showing the focus dependence of the LER values of isolated space patterns and isolated line patterns formed by the same exposure. In FIG. 3, the horizontal axis corresponds to the focus value, and the vertical axis corresponds to the LER value. In FIG. 3, a solid line 31 indicates the focus dependence of the isolated space pattern. A broken line 32 indicates the focus dependency of the isolated line pattern. The isolated line pattern is a pattern different from the isolated line pattern that has acquired the focus dependency shown in FIG.

  From the solid line 31 shown in FIG. 3, it can be understood that the best focus value of the isolated space pattern is −0.2 μm. Moreover, it can be understood from the broken line 32 shown in FIG. 3 that the best focus value of the isolated line pattern is −0.4 μm. As described above, it is well known that the best focus value varies depending on the pattern type (whether it is a space pattern or a line pattern). When the best focus value varies depending on the pattern type, generally, the focus value with the smallest focus deviation for each pattern is set as the best focus value. Here, the best focus value is −0.3 μm, which is the median value between the best focus value of the isolated space pattern and the best focus value of the isolated line pattern.

  When focus measurement is performed using both an isolated space pattern and an isolated line pattern, focus dependency of the LER values of the isolated space pattern and the isolated line pattern (as in the case of using the single resist pattern) ( FIG. 3) is acquired in advance.

  Next, the semiconductor substrate on which the resist film is formed is subjected to exposure of a focus shift measurement target, and the LER values of the formed isolated space pattern and isolated line pattern are measured. Subsequently, a focus value corresponding to the measured LER value is obtained based on the focus dependency of the LER value of these resist patterns acquired in advance.

  At this time, there are two focus values corresponding to the LER value of the isolated space pattern to be measured and the LER value of the isolated line pattern. However, as shown in FIG. 3, since the LER value of the isolated space pattern and the LER value of the isolated line pattern have different focus dependencies, one focus value can be specified.

  For example, it is assumed that the LER value of the isolated space pattern to be measured is 0.3 nm and the LER value of the isolated line pattern to be measured is 0.2 nm. At this time, the focus values corresponding to the LER value of the isolated space pattern are −0.62 μm and +0.32 μm (extrapolated value). On the other hand, the focus values corresponding to the LER value of the isolated line pattern are −0.62 μm and −0.2 μm. Since the isolated space pattern and the isolated line pattern to be measured are formed by the same exposure, the focus values of both are naturally equal. Accordingly, the focus value in this case is −0.62 μm. In this case, the amount of focus deviation from the best focus value (−0.3 μm) is −0.62 − (− 0.3) = − 0.32 μm.

  As described above, the focus value can be obtained from the LER value of any resist pattern. However, the focus value is preferably obtained based on the LER value of the resist pattern from which the focus value can be obtained with higher sensitivity.

  As shown in FIG. 3, in the focus dependence of the LER value of the isolated space pattern, the amount of fluctuation of the LER value with respect to the fluctuation of the focus value is larger on the minus side than the plus side of the best focus value of the isolated space pattern. . That is, according to the isolated space pattern, the focus value can be obtained with high sensitivity on the minus side of the best focus value of the isolated space pattern. Similarly, in the focus dependency of the LER value of the isolated line pattern, the amount of change of the LER value with respect to the change of the focus value is larger on the plus side than the minus side of the best focus value of the isolated line pattern. That is, according to the isolated line pattern, the focus value can be obtained with high sensitivity on the plus side of the best focus value of the isolated line pattern.

  Therefore, in the example of FIG. 3, it is preferable to obtain the focus value on the minus side from −0.2 μm based on the LER value of the isolated space pattern, and the focus value on the plus side from −0.4 μm It is preferable to obtain based on the LER value. The focus value of −0.4 to −0.2 μm where both ranges overlap may be obtained based on the LER value of any pattern, but based on the LER value of an isolated line pattern having a large LER value. It is preferable to obtain the focus value.

  In addition, in the example shown in FIG. 3, when the focus value is on the minus side of −0.5 μm, the LER value of the isolated space pattern is larger than the LER value of the isolated line pattern. Conversely, when the focus value is on the plus side of −0.5 μm, the LER value of the LER value pattern of the isolated line pattern is larger than the LER value of the isolated space pattern. Therefore, the focus value can also be obtained by using the magnitude relationship between the LER values of the isolated space pattern and the isolated line pattern to be measured.

  For example, in the above-described case, the LER value (0.3 nm) of the isolated space pattern is larger than the LER value (0.2 nm) of the isolated line pattern. Therefore, the focus value is on the minus side of −0.5 μm. In addition, the focus values corresponding to the LER value of the isolated space pattern with which the focus value can be obtained with high sensitivity in the range are −0.62 μm and +0.32 μm. Therefore, the focus value is −0.62 μm. As described above, the focus value can be easily obtained from the magnitude relationship and one of the LER value of the isolated space pattern and the LER value of the isolated line pattern.

  In this method, the focus value to be obtained is present on the side where the LER value becomes minimum than the focus value corresponding to the focus-dependent intersection of each resist pattern, and is smaller than the LER value of the intersection. If it is in the range corresponding to the LER value, it cannot be used. For example, in FIG. 3, it is the range of -0.5 micrometer--0.3 micrometer. In this range, there are two focus values that satisfy the LER value within the focus value range defined by the magnitude relationship of the LER values in each focus dependency. For this reason, in this range, as described above, it is necessary to obtain focus values corresponding to both LER values and obtain a focus value having an equivalent size.

  By adjusting the focus value of the exposure apparatus so as to eliminate the focus shift amount according to the focus shift amount obtained as described above, the resist film can be exposed in a state where there is no focus shift. As a result, exposure for forming a fine resist pattern can be performed under the same focus condition, and a fine resist pattern can be formed very uniformly. Further, as described above, the LER value can be obtained only by measuring the dimension with respect to the upper surface of the resist pattern. For this reason, the focus dependency of the LER value can be acquired in a shorter time than the focus dependency of the shift index of the conventional method in which the dimensions of the resist pattern on the top and bottom surfaces are measured.

  When detecting the presence / absence of focus deviation and the amount of focus deviation based on the LER value, if the LER value cannot be acquired in all the fluctuation range of the focus value that can occur during exposure, there is a range in which the amount of focus deviation cannot be detected. Will occur. For this reason, it is preferable that the focus measurement resist pattern has an edge capable of measuring the LER value in the entire range of fluctuation of the focus value that may occur during exposure. From this viewpoint, for example, it is possible to use a resist pattern having a small focus pattern variation and a large focus margin even when a focus shift occurs. When such a resist pattern is used, the resist pattern has an edge capable of acquiring the LER value in the entire range of fluctuation of the focus value that can occur during exposure.

  Furthermore, in order to improve the accuracy of detecting the presence / absence of a focus shift based on the LER value, it is preferable that the LER value fluctuate greatly with respect to the focus value variation. As shown in FIG. 2, the LER value tends to increase as the focus shift increases. Therefore, the amount of fluctuation in the LER value with respect to the fluctuation in the focus value is larger in the pattern in which the fluctuation in the pattern shape is large with respect to the focus deviation than in the pattern with little fluctuation in the pattern shape with respect to the focus deviation. For this reason, it is preferable that the focus measurement resist pattern has a focus margin comparable to the amount of change in focus value that may occur during exposure.

  By adopting the above-described focus measurement method, it becomes possible to form a fine resist pattern with good reproducibility. FIG. 4 is a block diagram showing an exposure system that employs the above-described focus measurement method.

As shown in FIG. 4, the exposure system 1 of this embodiment includes an exposure device 2, a developing device 3, a dimension measuring device 4, and a focus measuring device 5. The exposure apparatus 2 uses a KrF excimer laser, an ArF excimer laser, an F ₂ excimer laser, or the like as a light source, and performs projection exposure for forming an image of a mask pattern on a resist film formed on a semiconductor substrate to be exposed (hereinafter referred to as a wafer). Device. Here, a pattern for forming a focus measurement resist pattern on a wafer is formed in a process monitor pattern formation region on a mask (reticle) for forming an actual device. The focus measurement resist pattern may be an isolated space pattern and an isolated line pattern as described above, but here, the resist pattern including the line pattern and the space pattern, which is similar to the pattern of the actual device, is used as the focus measurement pattern. Adopted as. The actual device pattern is, for example, an element isolation pattern, a gate electrode pattern, a contact pattern, a wiring pattern, or the like.

  The wafer exposed in the exposure apparatus 2 is developed in the developing apparatus 3, and the resist film is patterned. The developed wafer is measured for edge position variations of the focus measurement resist pattern by the above-described method in the dimension measuring device 4 such as the above-described length measuring SEM. The measured edge position variation data is transmitted to the focus measuring device 5 and the above-described defocus amount is calculated. The focus measurement device 5 includes a LER value calculation unit 6, a focus value calculation unit 7, a focus deviation amount calculation unit 8, a pass / fail determination unit 9, and a storage unit 10.

  FIG. 5 is a flowchart showing processing performed in the exposure system 1 shown in FIG. As shown in FIG. 5, first, the focus dependency of the focus measurement resist pattern is obtained by the method described above (S1 in FIG. 5). At this time, the exposure apparatus 2 transmits the focus value at the time of performing exposure to the dimension measuring apparatus 4. The dimension measuring device 4 transmits the measured edge position variation data to the LER value calculation unit 6 together with the received focus value. The LER value calculation unit 6 calculates the LER value from the edge position variation data, and stores the LER value together with the focus value in a storage unit 10 constituted by an HDD (Hard Disk Drive) or the like. By repeatedly executing this process, the focus dependency of the LER value is stored in the storage unit 10.

  When acquisition of the focus dependency of the LER value is completed, a wafer having a resist film formed on the surface is placed in the exposure apparatus 2 and exposure is performed. After the exposure is completed, the exposed resist film is developed by the developing device 3, and a focus measurement resist pattern is formed on the wafer (S2 in FIG. 5). Thereafter, the dimension measurement device 4 measures the space pattern of the formed focus measurement resist pattern and the edge position variation of the line pattern. The measurement of the edge position variation may be performed inline or in outline. In the case of measuring in-line, the subsequent processes are processing steps incorporated as sub-processes in the photolithography process.

  The measured edge position variation data is transmitted to the LER value calculation unit 6 of the focus measurement device 5. The LER value calculation unit 6 that has received the data calculates the LER value from the edge position variation data, and transmits the LER value to the focus value calculation unit 7 (S3 in FIG. 5). At this time, the focus value calculation unit 7 reads the focus dependency of the LER value acquired in advance from the storage unit 10, and obtains the focus value corresponding to the received LER value by the above-described method (S4 in FIG. 5). The focus value is transmitted to the focus deviation amount calculation unit 8, and the focus deviation amount calculation unit 8 calculates the difference between the received focus value and the best focus value (S5 in FIG. 5). In the present exposure system 1, a configuration is adopted in which the focus shift amount calculation unit 8 reads the focus dependency of the LER value from the storage unit 10 and calculates the best focus value by the method described above.

  The calculated focus shift amount is transmitted to the pass / fail determination unit 9. If the received focus shift amount is within a preset allowable range, the pass / fail judgment unit 9 performs the pass quality judgment because the exposure is normally performed (S6 Yes → S7 in FIG. 5). On the other hand, if the received focus deviation amount is outside the preset allowable range, the exposure is not performed normally, and a failure is determined. (FIG. 5 S6 No → S10). The determination result of the pass / fail determination unit 9 is transmitted to the dimension measuring device 4. The dimension measuring device 4 changes the wafer transfer destination based on the received determination result. That is, a wafer that has been determined to be non-defective is advanced to the next process, and a wafer that has been determined to be defective is advanced to the regeneration process. Here, the regeneration process refers to a process in which the resist film is removed from the wafer, the resist is applied again, and re-exposure is performed.

  Further, the focus shift amount calculation unit 8 transmits the calculated focus shift amount to the exposure apparatus 2 regardless of the non-defective product determination and the defect determination. The exposure apparatus 2 adjusts the focus value, that is, adjusts the focus position to the best focus position based on the received focus shift amount (S8 in FIG. 5). For example, the exposure apparatus 2 adjusts the focus position by raising and lowering the height of the stage on which the wafer to be exposed is placed. After the focus position is adjusted, if there is a next wafer to be exposed, the next wafer is exposed (S9 Yes → S2 in FIG. 5), and if there is no next wafer, the process ends (FIG. 5). S9 No).

  Thus, by monitoring the focus shift in the lithography process for each exposure using the above-described focus measurement method, the quality of the exposure process can be significantly improved.

  The focus adjustment in the exposure apparatus 2 may be performed only when the amount of focus deviation is an amount that cannot be allowed in the photolithography process. Such a configuration can be realized by the focus shift amount calculation unit 8 receiving the determination result of the pass / fail determination unit 9 and transmitting the focus shift amount to the exposure apparatus 2 according to the received determination result. . That is, the focus deviation amount calculation unit 8 transmits the calculated focus deviation amount to the exposure apparatus 2 when the defect determination is made by the pass / fail determination unit 9, and when the non-defective product determination is made, the calculated focus deviation is calculated. Instead of the amount, data indicating no defocus is transmitted.

  In the exposure apparatus 2, it is not essential to measure the focus deviation for each exposure, and the focus deviation can be measured every time a plurality of wafers are exposed. Further, every time the exposure of the processing unit (lot) in the exposure process of the semiconductor device is completed, an average value of the focus shift amounts of the wafers belonging to the lot is calculated, and the focus position is adjusted according to the focus shift amount. A configuration may be adopted in which the exposure of the next lot is performed in the performed state. By performing feedback of the amount of focus shift to the exposure apparatus 2 as described above, it is possible to improve the stability of the focus value during exposure.

  In addition, the LER value calculation unit 6, the focus value calculation unit 7, the focus deviation calculation unit 8, and the pass / fail determination unit 9 described above include, for example, a dedicated arithmetic circuit, a processor, and a memory such as a RAM and a ROM. It can be realized as hardware and software stored in the memory and operating on the processor.

  By the way, the magnitude of the LER value varies depending not only on the pattern type such as the line pattern and the space pattern, but also on the film thickness of the resist film, the material contained in the resist film, and the like. FIG. 6 is a diagram showing the focus dependency of the LER value when three types of resist films formed with different film thicknesses are exposed using the same reticle.

  From FIG. 6, it can be understood that the focus value at which the LER value becomes the minimum, that is, the best focus value does not vary with the resist film thickness and is substantially constant. However, when the resist film thickness is different, the amount of change in the LER value when the focus value is changed to the plus side or the minus side is different. This is because when the resist film thickness varies, the intensity of the standing wave generated by the exposure light in the resist film (the intensity of the antinode of the standing wave) varies.

  FIG. 7 is a cross-sectional view schematically showing fluctuations in the intensity of the standing wave with respect to the film thickness. FIG. 7A shows a case where the intensity of the standing wave is maximum, and FIG. 7B shows a state where the intensity of the standing wave is minimum. As shown in FIG. 7A, in the state where the intensity of the standing wave in the resist film is maximized, the end surface 11a of the resist pattern 11 formed by development corresponds to the antinodes and nodes of the standing wave. Large irregularities are formed. On the contrary, as shown in FIG. 7B, in the state where the intensity of the standing wave in the resist film is minimized, the standing wave formed on the end surface 11a of the resist pattern 11 formed by development is used. The size of the unevenness corresponding to the belly and the node is reduced.

  In FIG. 6, the dotted line 41 in which the LER value fluctuates the most with respect to the fluctuation of the focus value corresponds to the resist film thickness at which the intensity of the standing wave is maximized. Further, the solid line 43 with the smallest fluctuation of the LER value with respect to the fluctuation of the focus value corresponds to the resist film thickness at which the intensity of the standing wave is minimized. Therefore, by adopting a resist film thickness that maximizes the intensity of the standing wave generated in the resist film during exposure, the variation in the LER value can be increased with respect to the variation in the focus value. In addition, by adopting the resist film thickness, it is easy to specify the best focus value, and it is possible to detect the focus shift amount with high sensitivity. In FIG. 6, the broken line 42 corresponds to the resist film thickness when the intensity of the standing wave in the resist film is intermediate between the intensity of the dotted line 41 and the intensity of the solid line 43.

  As described above, in the case of detecting a focus shift or a focus shift amount based on the LER value, a resist pattern is formed under a situation where the LER value greatly varies with respect to the focus value variation, and the LER value is acquired. Is preferred. Therefore, when the presence or absence of a focus shift or the amount of focus shift is detected based on the LER value, the resist material included in the resist film is preferably a material that easily generates LER. In addition, the baking temperature conditions (pre-bake temperature and post-bake temperature) before and after the formation of the resist pattern related to the shape of the resist pattern after development are also preferably conditions where LER is likely to occur. By adopting such a resist material and baking temperature conditions, the amount of fluctuation of the LER value with respect to the fluctuation of the focus value can be increased.

  For example, it is possible to realize a state in which LER is likely to occur by adopting a resist material and baking temperature conditions that can more effectively reflect the difference in intensity between the antinodes and nodes of the standing wave as the resist pattern shape. .

  When a KrF excimer laser is used as the light source of the exposure apparatus, a chemically amplified resist containing an acid generator is used. In this case, it is desirable to employ a resist material and a post-baking temperature condition such that the diffusion length of the acid generated in the resist film in the exposure and post-exposure baking steps is relatively small. When the resist film is a positive resist, the acid generator in the region irradiated with light generates an acid in the resist film. The resist film in the region where the acid exists is changed into a substance soluble in the developer by a chemical reaction using the acid as a catalyst. The chemical reaction is accelerated by baking after exposure, and the resist film in the region where the acid exists is dissolved by development. For this reason, by using a resist material and baking conditions in which the diffusion length of the acid is relatively small, a region where a chemical reaction using an acid as a catalyst is caused to be a substantially exposed region. In this case, the developed resist pattern has a shape that more reflects the state of the standing wave immediately after exposure. As a result, it is possible to realize a state in which LER is likely to occur at the edge of the resist pattern.

  As described above, by forming the focus measurement resist pattern under conditions where LER is likely to occur, it is possible to more accurately determine the focus shift amount. Further, by performing the focus adjustment of the exposure apparatus based on the focus shift amount thus obtained, it is possible to realize exposure with very little focus shift.

  In the actual manufacturing process of a semiconductor device, the resist film thickness, resist material, and baking temperature conditions are optimized so that the actual device pattern can be formed satisfactorily. This optimization condition is different from the above-described condition that LER is likely to occur. Therefore, when applied to a manufacturing process of a semiconductor device, focus measurement is performed on an exposure apparatus that performs exposure of an actual device under a condition that the above-described LER is likely to occur immediately before exposure of the actual device.

  That is, a resist film made of a resist material that easily causes LER is applied to a monitor wafer different from the wafer that forms the actual device with a film thickness that easily causes LER. The monitor wafer is carried into the exposure apparatus 2, and an isolated space pattern and an isolated line pattern in which the focus dependency of the LER value is acquired in advance are exposed. After the exposure, baking is performed at a temperature at which LER is likely to occur, and the developing device 3 develops the resist film. The dimension measuring device 4 measures the edge position variation of the formed isolated space pattern and isolated line pattern, and the LER value calculation unit 6 calculates the LER value of the isolated space pattern and the LER value of the isolated line pattern. The focus value calculation unit 7 obtains a focus value corresponding to the LER value based on the focus dependency of the LER value stored in the storage unit 10. Then, the exposure apparatus 2 adjusts the focus position according to the focus shift amount calculated by the focus shift amount calculation unit 8.

  In this manner, the focus is adjusted and the exposure apparatus is used to perform pattern exposure of the actual device, so that the best focus can be obtained even in the manufacturing process of a semiconductor device including a fine resist pattern in which LER becomes remarkable. Can be exposed. As a result, a good resist pattern having a desired cross-sectional shape and dimensions can be obtained.

  As described above, according to the present invention, the accurate focus value, the best focus value, and the best focus value can be obtained with a small amount of data even in a situation where a fine resist pattern having a relatively large LER is formed. The amount of deviation can be obtained. Then, by performing focus adjustment of the exposure apparatus according to the focus shift amount, a resist pattern having a good pattern shape and dimensions can be formed. As a result, a semiconductor device can be formed with a high manufacturing yield.

  The present invention is not limited to the embodiment described above, and various modifications and applications are possible within the scope of the effects of the present invention. For example, in the above description, an example of an exposure system in which the exposure apparatus and the focus measurement apparatus are separated from each other has been described. However, the focus measurement apparatus may be integrated with the exposure apparatus, and the arrangement of each part is not particularly limited.

  The present invention has an effect that an amount of focus shift can be easily detected even in a situation where a fine resist pattern having a relatively large LER is formed, and a focus measurement method, a semiconductor device manufacturing method, and It is useful as an exposure system.

SEM photograph showing LER measurement points in one embodiment of the present invention The figure which shows the focus dependence of the LER value in one Embodiment of this invention. The figure which shows the change with respect to the pattern shape of the focus dependence of the LER value in one Embodiment of this invention. The block diagram which shows the exposure system in one Embodiment of this invention The flowchart which shows the process of the lithography process in one Embodiment of this invention The figure which shows the change with respect to the resist film thickness of the focus dependence of the LER value in one Embodiment of this invention. Schematic sectional view showing the state of standing waves Cross section of resist pattern used for focus measurement The figure which shows the focus dependence of the amount of inclination in the conventional focus measurement method Diagram showing focus dependency of shift index in conventional focus measurement method

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Exposure system 2 Exposure apparatus 4 Dimension measurement apparatus 5 Focus measurement apparatus 6 LER value calculation part 7 Focus value calculation part 8 Focus deviation | shift amount calculation part 9 Pass / fail judgment part 10 Memory | storage part 11 Resist pattern (isolated line pattern)
31 LER value of isolated space pattern 32 LER value of isolated line pattern 41 Focus dependency of LER at resist film thickness where standing wave amplitude is maximum 42 LER of resist film thickness when standing wave amplitude is medium Focus dependency 43 Focus dependency of LER at resist film thickness with minimum standing wave amplitude 100 Isolated line pattern 110 Isolated space pattern 101 Top surface line width 102 Bottom line width 103 Top surface space width 104 Bottom surface space width

Claims (13)

In the focus measurement method of the exposure apparatus,
The resist film formed on the substrate is exposed with different focus values using an exposure apparatus, and the correspondence between the LER (Line Edge Roughness) value of the focus measurement resist pattern formed by each exposure and the focus value Acquiring a relationship;
Using the exposure apparatus, exposing a focus measurement target to a resist film formed on the substrate, and measuring a LER value of a focus measurement resist pattern formed by the exposure; and
Obtaining a focus value corresponding to a LER value of a focus measurement resist pattern formed by exposure of the focus measurement target based on the correspondence relationship;
A focus measurement method comprising:   The focus measurement method according to claim 1, further comprising a step of obtaining a focus shift amount in exposure of the focus measurement target from the obtained focus value and best focus value.   The focus measurement method according to claim 2, wherein the best focus value is a focus value at which the LER value is minimized in the correspondence relationship. The focus measurement resist pattern includes a line pattern and a space pattern, and the correspondence is acquired for the line pattern and the space pattern, respectively.
The focus value in the exposure of the focus measurement target is a correspondence relationship between the LER value of the line pattern and the focus value, a correspondence relationship between the LER value of the space pattern and the focus value, and a line formed by the exposure of the focus measurement target. The focus measurement method according to claim 2, wherein the focus measurement method is obtained based on a LER value of a pattern and a LER value of a space pattern formed by exposure of the focus measurement target.   4. The pattern according to claim 1, wherein the resist pattern is a pattern that is selected from a plurality of patterns having different shapes and has a maximum amount of change in the LER value with respect to the change in the focus value of the exposure apparatus. The focus measurement method described.   5. The focus measurement method according to claim 1, wherein the resist pattern has a focus margin that is substantially the same as an amount of focus deviation that may occur in the exposure apparatus. 6.   5. The focus measurement method according to claim 1, wherein the resist film has a film thickness that maximizes the intensity of a standing wave generated in the resist film in an exposed portion during exposure of the focus measurement target. .   8. The resist film is made of a resist material whose LER value varies depending on the temperature of baking performed after exposure and before development, and the baking is performed at a temperature at which the LER value is maximized. Focus measurement method. In a method for manufacturing a semiconductor device having a lithography process,
Exposing a resist film on a semiconductor substrate to form a resist pattern including a focus measurement resist pattern;
Measuring the LER value of the focus measurement resist pattern;
Obtaining a focus value corresponding to the LER value of the focus measurement resist pattern formed by the exposure based on the correspondence relationship between the LER value of the focus measurement resist pattern and the focus value of the exposure apparatus acquired in advance; ,
Obtaining a focus shift amount in the exposure from the determined focus value and the best focus value;
Determining whether the defocus amount is within a preset allowable range;
Including
A method of manufacturing a semiconductor device, comprising: determining that the exposed semiconductor substrate is defective when the focus shift amount is not within the allowable range. The focus measurement resist pattern includes a line pattern and a space pattern,
The focus value in the exposure is the correspondence between the LER value and the focus value of the line pattern, the correspondence between the LER value of the space pattern and the focus value, and the LER value of the line pattern formed by the exposure of the focus measurement target The method for manufacturing a semiconductor device according to claim 9, wherein the method is obtained based on a LER value of a space pattern formed by exposure of the focus measurement target.   11. The method of manufacturing a semiconductor device according to claim 9, wherein the subsequent exposure is performed after the focus position of the exposure apparatus is adjusted to the best focus position in accordance with the focus shift amount. In an exposure system equipped with an exposure apparatus that images and transfers a mask pattern on a resist film,
Means for measuring a LER value of a resist pattern for focus measurement formed by exposure by the exposure apparatus;
Means for obtaining a focus value corresponding to the LER value of the focus measurement resist pattern formed by the exposure based on the correspondence relationship between the LER value of the focus measurement resist pattern and the focus value of the exposure apparatus acquired in advance; ,
Means for obtaining a focus shift amount in the exposure from the obtained focus value and best focus value;
Means for determining whether the defocus amount is within a preset allowable range;
An exposure system comprising:   13. The exposure system according to claim 12, further comprising means for adjusting a focus position of the exposure apparatus to a best focus position in accordance with the focus shift amount.