JP5460479B2 - Pattern dimension measuring apparatus and contour line forming apparatus - Google Patents

Description

  The present invention relates to a pattern dimension measuring apparatus and contour line extracting apparatus, and more particularly to a pattern dimension measuring apparatus and contour line extraction capable of performing measurement and contour line formation with high accuracy even for a pattern that shrinks by irradiation with an electron beam or the like. Relates to the device.

  In recent years, in photolithography, a fine circuit pattern is processed using a photoresist (hereinafter referred to as “ArF resist”) that reacts with argon fluoride (ArF) excimer laser light. This ArF resist has a known phenomenon that acrylic resin or the like undergoes a condensation reaction by electron beam irradiation to reduce the volume (hereinafter referred to as “shrink”).

  In Patent Document 1 and Patent Document 2, a curve indicating the relationship between the number of times of measurement and the amount of shrinkage is used as a method for appropriately obtaining the actual dimension value (that is, the dimension value before shrinkage) of a pattern that causes shrinkage. A method is described in which a dimension value (zero time value) before shrinking is estimated by preparing and fitting the curve to the first and subsequent measurement values.

WO 03/021186 (corresponding US Pat. No. 7,659,508) WO03 / 098149 (corresponding US Pat. No. 7,285,777)

  When performing processing pattern observation and length measurement using an electron microscope, there is a problem that shrinkage occurs due to electron beam irradiation at the time of image acquisition, and original dimension measurement cannot be specified. Patent Documents 1 and 2 are methods for acquiring an image several times continuously at a measurement point, creating a fitting curve from a dimensional change, and estimating an original dimension that is a dimension before shrinking. However, the present inventors have clarified that, in the current fine circuit pattern, the shrink depends on the pattern size, and the finer the pattern, the faster the reaction. That is, it has been found that there is a possibility that high measurement accuracy cannot be maintained by an estimation method using a simple fitting curve depending on the pattern size.

  Hereinafter, a description will be given of a pattern dimension measuring apparatus and an outline extracting apparatus for the purpose of appropriately measuring a dimension or extracting an outline of a pattern edge even when the degree of shrinking varies depending on the pattern size. To do.

  As an aspect for achieving the above object, in the following, in the pattern dimension measuring apparatus for measuring the dimension of the pattern formed on the sample based on a signal obtained by scanning the sample with an electron beam, A pattern dimension measuring apparatus comprising: an arithmetic unit that corrects the pattern dimension value by adding a correction amount corresponding to the dimension of the pattern to the dimension value of the pattern obtained based on a signal. suggest.

  As another mode for achieving the above object, a contour line for extracting a contour line of a pattern edge formed on the sample based on a signal obtained by scanning the sample with an electron beam is described below. In the contour line extraction device provided with the forming unit, the contour line forming unit is configured to generate the contour based on a plurality of pieces constituting the contour line obtained based on the signal or a correction amount registered for each part. A contour extraction apparatus is proposed in which the position of each point constituting a line is corrected and a new contour is formed based on the corrected point.

  According to the above-described aspect, it is possible to derive the pattern dimension value with high accuracy regardless of the amount of shrink that changes according to the size of the pattern. Further, according to the other aspect, it is possible to extract a contour line that accurately reproduces the shape before shrinking even if the pattern has a complicated shape.

The figure which shows the relationship between the length measurement value of the pattern shrunk by electron beam irradiation, and the frequency | count of a measurement. The figure explaining the example which estimates the length measurement value before shrink based on the curve (shrink curve) which shows the relationship between length measurement value and the frequency | count of a measurement. The figure explaining an example of the shrink curve of a fine pattern. The figure explaining the example which estimates the length measurement value before shrinking based on the shrink curve of a fine pattern. The figure explaining the relationship between the estimated value of the dimension before shrink of a fine pattern, and an actual dimension value. The flowchart explaining a (DELTA) S curve creation process. Explanatory drawing explaining (DELTA) S curve. The flowchart explaining the calculation process of standard shrink amount. The flowchart explaining the length measurement value estimation process using a shrink curve. The flowchart explaining the length measurement value estimation process of a hole pattern. The figure explaining the shrink example of an OPC pattern. The figure explaining the shrink example of an OPC pattern. The flowchart explaining the correction process to the outline before shrink. The figure explaining the search of a resist width. Schematic explanatory drawing of a scanning electron microscope. Schematic explanatory drawing of the control apparatus of a scanning electron microscope. The figure explaining an example of the outline extraction method. The flowchart explaining an outline extraction process. The figure explaining the correction method to the outline before shrink. The figure explaining the estimation process of the dimension value before shrink.

  Hereinafter, a method for forming an appropriate shrink curve will be described with reference to the drawings. FIG. 15 is a diagram illustrating one embodiment of a scanning electron microscope (SEM).

  An electron beam 1503 extracted from an electron source 1501 by an extraction electrode 1502 and accelerated by an acceleration electrode (not shown) is focused by a condenser lens 1504 which is a form of a converging lens, and then is scanned on a sample 1509 by a scanning deflector 1505. Are scanned one-dimensionally or two-dimensionally. The electron beam 1503 is decelerated by the negative voltage applied to the electrode built in the sample stage 1508, and is focused by the lens action of the objective lens 1506 and irradiated onto the sample 1509.

  When the sample 1509 is irradiated with the electron beam 1503, secondary electrons and electrons 1510 such as backscattered electrons are emitted from the irradiated portion. The emitted electrons 1510 are accelerated in the direction of the electron source by the acceleration action based on the negative voltage applied to the sample, and collide with the conversion electrode 1512 to generate secondary electrons 1511. The secondary electrons 1511 emitted from the conversion electrode 1512 are captured by the detector 1513, and the output I of the detector 1513 changes depending on the amount of captured secondary electrons. Depending on the output I, the brightness of a display device (not shown) changes. For example, when forming a two-dimensional image, an image of the scanning region is formed by synchronizing the deflection signal to the scanning deflector 1505 with the output I of the detector 1513.

  In addition, although the example of FIG. 15 demonstrates the example which detects the electron emitted from the sample by converting once with a conversion electrode, of course, it is not restricted to such a structure, For example, it is the acceleration | stimulation of the accelerated electron. It is possible to adopt a configuration in which the detection surface of the electron multiplier tube or the detector is arranged on the orbit.

  The control device 1514 controls each component of the scanning electron microscope, and forms a pattern on the sample based on a function of forming an image based on detected electrons and an intensity distribution of detected electrons called a line profile. It has a function to measure the pattern width.

  FIG. 16 is a diagram for explaining the details of the control device 1514 of the scanning electron microscope. The control device 1514 includes a calculation unit 1603 and a storage unit 1602. The apparatus condition control unit 1603 stores the apparatus conditions (electron beam acceleration voltage, magnification (field size), beam current, focusing condition of each lens, and pattern measurement) stored in the apparatus condition storage unit 1610. Control signal is generated on the basis of the conditions. The measurement unit 1604 measures the pattern width based on the detection signal obtained by SEM. The measurement unit 1604 forms a line profile based on the detection signal, and measures the dimension value of the desired pattern based on the peak width of the line profile. The measurement result in the measurement unit 1604 is stored in the measurement result storage unit 1611. The S-curve forming unit 1605 forms a curve (S-curve) having the horizontal axis as the number of measurements and the vertical axis as the dimension value based on the measurement result obtained by the measuring unit 1604. Information about the formed S curve is stored in the measurement result storage unit 1611. In the measurement value difference calculation unit 1606, the difference between the measurement values between different measurement times is calculated. The calculation result is stored in the measurement value difference storage unit 1612. The dimension estimation unit 1606 estimates a pattern dimension before irradiation with an electron beam based on the S curve formed by the S curve forming unit 1605. A specific method will be described later. The estimation result is stored in the dimension estimation result storage unit 1613.

  The ΔS curve forming unit 1608 forms a curve (ΔS curve) with the horizontal axis representing the pattern size (or design size) and the vertical axis representing the measured value difference, based on the calculation result by the measurement value difference calculating unit 1606. A specific selection method will be described later.

  The contour line forming unit 1609 forms a contour line based on an image obtained by SEM. FIG. 17 is a diagram illustrating an example of a technique for extracting a contour line from a pattern image formed based on detected electrons. Note that this contour line extraction step may be performed in the control device 1514 or may be performed by another arithmetic device or the like. For contour extraction, an SEM image is first formed as illustrated in the flowchart of FIG. 18 (step 1801). Next, a first contour line (not shown) is extracted from the white band 1702 corresponding to the edge portion of the pattern 1701 on the SEM image (step 1802). As a method for extracting the first contour line, a method of extracting a pattern image composed of bitmap data from an SEM image and converting the pattern image into pattern data composed of vector data can be considered.

  Next, the layout data 1703 and the first contour are superimposed by vector data comparison or pattern matching between the formed first contour and layout data 1703 (step 1803). The layout data 1703 is line information of design data stored in the GDS format or the like, and is read from the database 1614 and applied. After such superposition, the luminance distribution information collection areas 1704 and 1705 are set to be perpendicular to the first contour line segment, and the luminance distributions 1706 and 1707 are detected (step 1804). . By extracting a pixel having a predetermined brightness in the luminance distribution formed in this way and defining the position as the second contour line position, a more accurate contour line can be formed (step 1805).

  Such an accurate contour forming method is described in Japanese Patent Application Laid-Open No. 60-169777, Japanese Patent Application Laid-Open No. 6-325176, Japanese Patent Application Laid-Open No. 8-161508, Japanese Patent Application Laid-Open No. 9-204529, and the like. Existing methods can be applied.

  As described above, the layout data and the first contour line can be associated with each other by superimposing the first contour line and the layout data. By using the line segment information of each line segment included in the layout data as each line segment information of the contour line, the contour line data can be registered in the same predetermined format as the design data.

  As described above, there is a pattern in which the volume decreases when the electron beam is irradiated. For example, the decrease in volume that occurs when an ArF resist is irradiated with an electron beam includes the acceleration voltage Vacc to the focused electron beam pattern, the electron beam current density Ipd, and the shrink amount 2S in the line-shaped pattern (shrink at one edge). Based on the amount of shrink at both edges when the amount is S, it can be obtained using equation (1).

2S = K1 · Vacc ^K2 · {1-exp (− (Ipd ^0.5 · n / K3))} (1)
Here, 2S: shrink amount (both sides), Vacc: acceleration voltage (V), K1, K2, K3: parameters determined by resist, n: number of measurements.

  Therefore, the dimensional change when images of the same point are repeatedly acquired and measured changes as shown in FIG. 1, and this dimensional change is called a shrink curve. When the shrink curve is approximated based on the above equation (1), the dimension value before the first length measurement can be estimated as shown in FIG. Further, the shrink amount generated in the first length measurement, that is, the difference between the pre-shrink dimension and the first length measurement value is called zero shrink.

  However, the inventors' experiments have shown that the actual measurement results do not follow the above formula (1) in a fine pattern with a measurement line width of 100 nm or less.

  FIG. 3 is a shrink curve when a fine pattern is measured. From this curve, the zero shrink estimated from the equation (1) becomes very small as shown in FIG. However, it has been found that this phenomenon looks like this because the shrinkage speed becomes high and the zero shrinkage is large as shown in FIG. In addition, this phenomenon appears remarkably in a fine pattern of 100 nm or less, and the above formula cannot estimate the zero shrink of the current process pattern, and cannot estimate the dimension value before shrink.

  According to the experimental results, when the shrinking speed is increased depending on the dimensions, and the observation conditions (acceleration voltage, current density) are the same, the zero shrink (ΔS) depends on the measurement dimensions, and the following equation (2) Can be expressed as

ΔS = exp (− (CD ^0.5 / K1)) (2)
Here, K1 is a parameter determined by the type of resist. In this embodiment, a procedure for estimating the dimension value before shrinking without depending on the line width is shown.

  In order to estimate the dimensions using this method, a patterned wafer is prepared using the same resist as the resist to be used, and the line width and shrinkage are measured in advance. Next, an approximate curve (ΔS curve) and a zero shrink amount are calculated from the preliminary measurement results.

  When performing measurement for estimating the pre-shrink dimension value, the correction value is calculated from the ΔS curve and the standard zero shrink amount, and the dimension value is estimated. The procedure is shown below.

What is required for the pre-measurement is a line pattern whose design dimensions are changed. For example, a line pattern whose dimensions are changed in steps of 10 nm from 50 nm to 150 nm is prepared. The ΔS curve calculation sequence is shown in FIG. Each dimension-changed pattern is continuously measured twice, and the length measurement value is recorded with CD _{1 as} the first length measurement result and CD ₂ as the _second length measurement result.

The ΔS curve obtained in this embodiment is obtained by plotting the difference between the first length measurement value and the second length measurement value for each different pattern dimension, and is a method for creating the ΔS curve. As described above, CD ₁ -CD ₂ is calculated based on CD ₁ and CD ₂ to obtain ΔS _(1-2) , and the calculation result is plotted.

After the measurement, a ΔS curve 1 is created according to the equation (2) from the pattern dimension CD ₁ or the design size and ΔS _(1-2) = CD ₁ −CD ₂ .

  FIG. 7 illustrates the ΔS curve 1 obtained at this time. Here, the X-axis is the first measurement result, but it may be the design size. Next, a representative pattern for calculating the standard zero shrink amount 2S (S) is determined. A representative pattern determination sequence is shown in FIG.

  The representative pattern for determining the standard zero shrinkage amount must be large enough to allow the shrinkage amount to be estimated by the equation (1), and is equal to or larger than the dimension at which the slope is zero in the ΔS curve. In addition, experience shows that if the line pattern is 100 nm or more, the zero shrink change due to the dimension does not appear. Therefore, a pattern having an arbitrary dimension value of 100 nm or more may be used as the representative pattern. After the representative pattern is determined, the representative pattern is continuously measured from at least 3 times to usually 5 to 10 times, and the respective dimension values are recorded.

After estimating the length measurement value CD ₀ before shrinkage from the equation (1) using the recorded dimension value, the value of CD ₀ -CD ₁ is set as the standard zero shrink amount. When the pattern used for the measurement is the same as the ΔS curve acquisition point, the dimension values used for estimation are the dimension values at the time of ΔS curve acquisition for CD ₁ and CD ₂ .

  Alternatively, a plurality of dimensional patterns that are the same as the representative pattern may be measured from the same wafer surface, and the standard zero shrink amount may be calculated using the average value.

Next, the shrink correction formula (3) is calculated from the calculated ΔS curve 1 and the standard zero shrink amount and stored.
ΔS ₁ = exp {− (CD ₁ ^0.5 / K1)} + 2S (3)
Next, FIG. 9 shows a pre-shrink measurement length estimation sequence at the time of measurement using the above equation (3).

After the length measurement, the zero shrink amount of CD ₁ can be calculated and displayed based on the previously obtained equation (3) based on the obtained dimension value. Thereby, at the time of measurement, the length measurement value before shrinking can be estimated by one length measurement, and it becomes possible to suppress damage due to the minimum electron beam irradiation.

  In the present embodiment, in order to perform dimension estimation (calculation) before shrink using the shrink curve of the representative pattern, the over shrink amount in the pattern near 40 nm is derived as a part of the correction value. The overshrink amount is the first measured value of the shrinkage amount of the measurement target pattern (for example, a pattern having a width of about 40 nm that generates a large shrinkage in the first measurement) and the shrinkage amount of the representative pattern. It is the difference between the measured values. By adding the over shrink amount to the standard zero shrink amount, a correction amount for a pattern having a large shrink is obtained.

  Therefore, by solving the equation (4), it is possible to estimate the shrink amount of the measurement target pattern in which the shrinkage is greatly generated.

2S = 2S (S) + ΔS (4)
(2S (S): standard shrink amount of representative pattern, 2S: shrink amount of measurement target pattern, ΔS: over shrink amount of measurement target pattern)
It is possible to estimate the pattern size before shrinking by adding 2S obtained as described above to the measurement value of the measurement target pattern as shown in Equation (5).

CDe = CDx + 2S (5)
(CDe: estimated pattern dimension before shrinking, CDx: measured pattern dimension)
Since ΔS changes depending on the size of the measurement target pattern as described above, it is preferable to store ΔS in a predetermined storage medium in advance according to the type of pattern and read it out during calculation. Since ΔS varies depending on the actual dimension of the pattern, it is preferable to read the value according to the size of the design data of the measurement target pattern. Alternatively, an approximate curve (function) or table indicating the relationship between ΔS and the pattern size may be created in advance, and ΔS may be derived based on the function and the pattern size.

Further, the relationship between the actual measurement value of the pattern (CD ₁ (first measurement value)) and ΔS (or 2S) is made into a function or a table in advance, and based on the actual measurement value obtained at the time of actual measurement. Thus, ΔS (or 2S) may be derived.

  FIG. 20 is a flowchart illustrating a process of calculating a shrink correction amount of a fine pattern and a process of estimating a pattern dimension before shrinking by correcting a measurement value based on the correction amount.

First, CD ₁ (first measurement value) and CD ₂ (second measurement value) are measured for a plurality of patterns (for example, patterns having a dimension width of 50 nm to 150 nm) (step 2001). . Next, CD ₁ −CD ₂ is calculated for each size, and a plot (ΔS curve 1) is formed on a graph with the horizontal axis representing the pattern size (or design size is acceptable) and the vertical axis representing CD ₁ −CD ₂ (step S). 2002). Next, a pattern size corresponding to a point where the slope of the ΔS curve 1 is zero or the slope is equal to or smaller than a predetermined value (selection of a representative pattern) is performed (step 2003). This representative pattern has a first pattern size. Next, a shrink curve (S curve) is created based on n measurements for the representative pattern (step 2004). Next, “standard zero shrink amount (difference between estimated zero time value and first measured value)” is calculated based on the S curve formed for the representative pattern (step 2005). Next, a difference (ΔS: ΔS _{1 to} ΔS _n ) between the representative pattern CD ₁ -CD ₂ and a plurality of patterns CD ₁ -CD ₂ having a smaller pattern size with respect to the representative pattern is calculated (Step S ₁ ). 2006). The steps so far are the calculation step of the shrink correction amount, and the information is stored in a predetermined storage medium. Note that the above order is merely an example of a method for obtaining the shrink correction amount, and the standard zero shrink amount and ΔS may be obtained in different orders.

  Next, length measurement of the actual measurement target pattern is executed (step 2007). Next, based on the dimension value of the measurement target pattern, the standard zero shrink amount, and ΔS, the dimension value before shrinking is calculated based on Equation (5) and the like (step 2008). The pre-shrink dimension value obtained in this way is stored in a predetermined storage medium, and is output as a measured value (or estimated value) to a display device (not shown) (step 2009).

  In the present embodiment, an example has been described in which the total shrink amount is obtained by adding the standard zero shrink amount (2S (S)) and the shrink amount (ΔS) depending on the size of each pattern to the actual measurement value. Instead of calculating each time, for example, a table or a relational expression in which 2S (S) + ΔS is registered in advance for each pattern dimension (design size or measured value) is prepared, and the measured actual value (or Depending on the design size of the measurement target pattern, the total shrink amount may be derived by reading out the total shrink amount from the table or performing a calculation based on the relational expression.

  As described above, by obtaining the standard zero shrink amount and ΔS in advance even if the pattern is large in shrink due to one measurement (electron beam irradiation) and difficult to estimate the zero value by the fitting curve, Accurate estimation of the zero time value is possible. In addition, if the standard zero shrink amount and ΔS are obtained, the pre-shrink dimension can be calculated in one measurement.

  The first embodiment is a method for estimating the dimension before shrinking of the fine line pattern, but is also effective for a method of estimating the dimension value before shrinking of the fine hole pattern.

In a pattern in which a plurality of holes are densely packed, a phenomenon similar to that of a fine line pattern is observed as the hole interval is closer. In the case of a dense hole pattern, a resist region is formed between holes, and the closer the hole interval is, the finer the resist region is. Therefore, in the hole pattern, a ΔS curve can be created not by the hole size but by the hole interval. CD ₁ and CD ₂ in FIG. 6 are the measurement results of the distance between the measurement hole and the peripheral hole, respectively. The standard zero shrink amount can be calculated by the same method. FIG. 10 shows a pre-shrink measurement value estimation sequence at the time of measurement.

At the time of hall measurement, hall measurement and hole interval measurement are performed on the same screen, and the respective dimension values are recorded. The hole interval is CD ₁ and the amount of zero shrink can be estimated from equation (3). Hall measurement results, calculates the Hall measurement results CD ₂ and the zero-shrinkage than pre-shrinking dimension CD (0), and displays.

Similarly, at the time of space measurement, it can be estimated from the space-to-space interval as in the case of hole measurement. In this case, the size before shrinking can be estimated by setting the space interval as CD ₁ and the space measurement result as CD ₂ .

  With photomasks used for transferring semiconductor patterns and flat panels, it is difficult to form resist patterns that are faithful to design data due to the effects of optical proximity effects (OPC) as the patterns become finer. It is becoming. In order to solve such a problem, an attempt has been made to realize proper pattern formation by adding an auxiliary pattern (OPC pattern). In such OPC, there is a demand for shape evaluation based on two-dimensional dimension measurement as well as one-dimensional dimension measurement as in the first and second embodiments. Based on such shape evaluation, the mask design is corrected. In this OPC measurement, it is known that a pattern contour line is extracted from an SEM image, a two-dimensional shape is evaluated based on the contour line, or a shape evaluation is performed by exposure simulation based on the contour line.

  In the case of a pattern that requires two-dimensional measurement, the measurement pattern shape and the peripheral pattern are complicated, and the amount of shrinkage may vary from pattern to pattern or pattern part. That is, the zero shrink amount may be different for each pattern or pattern portion. In the present embodiment, a description will be given of a contour line extraction method capable of performing pattern edge contour extraction with high accuracy regardless of the amount of shrinkage that differs for each pattern or pattern part.

  FIG. 11 shows an example of the OPC pattern. A description will be given taking as an example a pattern in which a T-shaped portion is formed of an ArF resist. The shrinkage of the T pattern has different zero shrinkage amounts because the resist line widths are different at points A, B, and C in FIG. Therefore, when an SEM image for contour line extraction is acquired, the shrink amount increases in the order of A> B> C, A having a smaller line width shrinks more, and C having a larger line width is less likely to shrink. When the contour line is extracted from this SEM image, the contour line enters more inside at the point A than the originally obtained contour line.

  Similarly, when the hole patterns are arranged as shown in FIG. In the case of a hole pattern, since the periphery is a resist, the shrink differs depending on the peripheral pattern density.

  The resist spreads in the upper part of the hole, whereas the resist amount is reduced because another hole is arranged in the side part of the hole. Therefore, since the zero shrink amount is larger in the vertical direction than in the horizontal direction, after shrinking, an SEM image having an elliptical shape is obtained, resulting in a contour extraction different from the original.

In the present embodiment, a method of correcting to a correct contour line using such an OPC pattern is shown. FIG. 13 shows a sequence. In order to perform contour correction, a ΔS curve using a line pattern is created and a standard zero shrink amount is measured in the same manner as in the first embodiment, and Equation (3) is obtained. Next, an SEM image for extracting a contour line is acquired, and an edge inspection contour line following the pattern edge is extracted. From the extracted edge points, as shown in FIG. 14, edge detection is performed in the resist direction in the direction perpendicular to the pattern, and the resist width CD ₁ is measured. Next, assuming that the resist width at each edge point is CD ₁ , the dimension value is estimated from equation (3). If the edge cannot be detected, the amount of resist is considered to be infinite, and the standard zero shrink amount is set as the correction value because it is not affected by the pattern shape and density. In addition, since ΔS uses the change of both edges as the shrink, the zero shrink of the single edge alone is half the value of the zero shrink.

  Next, the correction amount is converted into a pixel based on the magnification of the acquired SEM image, and the edge position is shifted from the current edge position by the correction amount with respect to the edge detection direction. This correction is performed on all edge points, and a new contour line is displayed. This makes it possible to extract the outline before shrinking without depending on the pattern shape and the peripheral pattern.

  FIG. 19 is a diagram illustrating a process of forming a contour line (third contour line) that takes into account the shrinkage amount based on the contour line (second contour line) formed through the process of FIG. . First, for each point 1902 (pixel) on the second outline 1901, a correction amount assigned to each line segment is derived (FIG. 19 (2)). The attribute information of the piece to which each point belongs is added in advance to each piece on the second contour line 1901 based on the layout data 1703 when the second contour line is formed. To do. The correction amount of each piece is stored in advance in a predetermined storage medium, and is read based on the designation of each point 1902. In this example, the correction amount 1904 of the longest pattern dimension 1903 is the smallest, and the correction amount 1906 of the second longest pattern dimension 1905 is the second smallest. Further, the correction amount 1908 of the shorter dimension 1907 of the pattern dimensions 1903 and 1905 is the smallest among the three correction amounts. As described above, the smaller the pattern dimension, the more the shrinkage tends to occur. Therefore, the correction amount is registered as such. Since the correction amount of each piece also changes depending on the presence of adjacent patterns, etc., it is registered for each combination of at least the pattern dimension and the position of another pattern, and according to the situation of the pattern part forming the contour line It is better to apply.

  As the correction amount of each point, it is desirable to register the correction amount in the vertical direction of each piece of the layout data 1703. Next, a third outline 1910 is formed so as to connect the correction points 1909 whose positions are adjusted by correction (FIGS. 19 (3) and (4)). The contour correction amount is stored in advance in the contour correction amount storage unit 1615 so that it can be read out as necessary. The correction amount may be extracted from the result of actual measurement for each part of the pattern made of the same material.

  Further, if there is a region that can be regarded as constant in a portion where there is a shrinkage amount, the correction amount may be registered in units of parts of the pattern instead of each piece.

  According to the configuration as described above, it is possible to form a contour line with high accuracy even when the shrink amount differs for each pattern or pattern portion.

1501 Electron source 1502 Extraction electrode 1503 Electron beam 1504 Condenser lens 1505 Scanning deflector 1506 Objective lens 1507 Sample chamber 1508 Sample stage 1509 Sample 1510 Electron 1511 Secondary electron 1512 Conversion electrode 1513 Detector

Claims (8)

In a pattern dimension measuring apparatus for measuring the dimension of a pattern formed on the sample based on a signal obtained by scanning the sample with an electron beam,
The dimension value of the pattern is corrected by adding a correction amount based on the shrinkage amount according to the size of the pattern generated by irradiation of the electron beam to the sample to the dimension value of the pattern obtained based on the signal. An apparatus for measuring a pattern dimension, comprising: In claim 1,
The arithmetic unit responds to the correction amount based on the first shrink amount when the pattern of the first size is irradiated with the electron beam, or the size of the pattern smaller than the first size according to the first shrink amount. A pattern dimension measuring apparatus for correcting the dimension value of the pattern using a correction amount obtained by adding the second shrink amount stored in the above. In claim 2,
The arithmetic device has a difference in size between a measurement value obtained by first electron beam irradiation and a measurement value obtained by second electron beam irradiation of the first size pattern, and the first size pattern. The pattern size, wherein the correction amount is calculated based on a difference between a measurement value of the second size pattern by the first electron beam irradiation and a difference between the measurement value by the second electron beam irradiation. measuring device. In claim 1 ,
The arithmetic unit obtains a difference between a dimension value before irradiation with an electron beam and a measurement value obtained by the first electron beam irradiation for a pattern having a different size formed on the sample, and calculates a difference between the measurement values having the different size. based on the pattern dimension measuring apparatus according to claim Rukoto determined an approximate curve indicating the change in the difference according to the size of the pattern. In claim 1 ,
The arithmetic unit corrects the pattern dimension value by adding the correction amount to a dimension value obtained by first-time electron beam irradiation for a pattern formed on the sample. apparatus. In a contour line extraction device including a contour line forming unit that extracts a contour line of a pattern edge formed on the sample based on a signal obtained by scanning an electron beam on the sample.
The contour line forming unit is registered for each of a plurality of pieces or parts constituting the contour line obtained based on the signal, and based on a correction amount corresponding to a shrink amount generated by irradiation of the electron beam to the sample. Then, the position of each point constituting the contour line is corrected, and a new contour line is formed based on the corrected point . In claim 6,
The contour line extracting apparatus extracts the contour line by thinning an edge portion of an image formed based on the signal . In claim 6 ,
The contour extraction apparatus according to claim 1, wherein the correction amount increases as the width of the pattern formed by the pieces decreases.

Images