JP2019032290A - Drift correction method of scan type probe microscope, and scan type probe microscope with drift correction function - Google Patents

Abstract

To achieve a highly accurate shape measurement by a scan type probe microscope even with respect to a fine semiconductor device.SOLUTION: A scan type probe microscope is configured to scan a probe along a main scan line on a sample surface in an X-axis direction to conduct a measurement-purpose main scan; move the probe in a sub scan direction (a Y-axis direction); scan the probe along a next main scan line; repeat the measurement-purpose main scan by Lto L; after completing a specific measurement-purpose main scan, cause the probe to be scanned along a reference shape measurement-purpose scan line preset to the sample surface, and measure a reference shape; compare a current reference shape measurement result with a previously measured reference shape measurement result to thereby compute an amount of drift occurring due to a time elapse between the current reference shape measurement and previously measured reference shape measurement; and correct a currently obtained measurement result of the measurement-purpose main scan.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a drift correction method for a scanning probe microscope and a scanning probe microscope having a drift correction function, and in particular, a height that occurs in a sub-scanning direction with respect to a measurement result of a measurement apparatus that acquires measurement data by raster scanning. The present invention relates to a method for correcting drift deformation about two axes in the direction and the main scanning direction, and a scanning probe microscope having a correction function for correcting such drift deformation.

In the scanning probe microscope (SPM), the shape measurement data of the sample surface is acquired by measurement based on a probe scanning operation (raster scanning) including main scanning and sub scanning.
In raster scanning, when main scanning is in the X-axis direction and sub-scanning is in the Y-axis direction, when main scanning in the X-axis direction ends, the main scanning in the X-axis direction is performed again by shifting by ΔY in the sub-scanning direction. By repeating this, measurement data of the sample surface is obtained.
One main scan is completed in a relatively short time. However, if the sub-scan is repeated by ΔY, the time gradually elapses from the first main scan, particularly between immediately after the start of main scan and immediately before the end. As time elapses, the influence of drift in the sub-scanning direction cannot be ignored on the measurement result, and the relative positional relationship between the main scanning lines becomes larger in error from the actual sample shape.
Therefore, in order to correct such drift deformation for each main scanning line, it is necessary to know the drift amount during measurement of each main scanning line.

In Patent Document 1, first, the inclination in the height direction of each main scanning line with respect to the main scanning direction is corrected, then the average height of each main scanning line is obtained, and the average height of all main scanning lines is equal. It is described that the correction is made so that
Non-Patent Document 1 obtains a plurality of SPM images and confirms in which direction the common feature points in those images are moving, thereby obtaining the drift velocity of the apparatus and performing a plurality of measurements. It describes that correction is performed on each main scanning line of an image to be corrected by interpolating the drift velocity obtained from the image.
Non-Patent Document 2 proposes that drift correction in the height direction is performed by combining a normal raster scanning measurement result and a drift correction measurement result. In the experiment shown here, first, normal raster scanning measurement is performed, and then drift correction measurement is performed. In the first normal measurement, the effect of drift is negligible because the main scan is completed in a short time, but since the sub scan takes a long time, drift deformation occurs in the sub scan direction, and the relative position between each main scan line is the actual position. It will be different from the shape. Next, after the main scanning axis is rotated by 90 °, measurement for drift correction is performed.
At this time, a narrow range in the sub-scanning direction is selected from the measurement range in normal measurement, and measurement is performed with a small number of main scanning lines. The main scanning direction of drift correction measurement is the sub-scanning direction in normal measurement. Taking advantage of the small amount of drift in the sub-scanning direction when the number of lines is small, this result is the height of the normal measurement result in the sub-scanning direction. Treat as a standard. Assuming that drift does not occur in the horizontal direction only in the height direction, the measurement data can be expressed in the form of an xy matrix, and the component in each xy can be handled as information on the height z. Since the two measurement results have the same xy coordinate result, the drift in the sub-scanning direction in the normal measurement can be corrected by comparing the height z of the point of those coordinates.

JP 2009-58479 A

C. A. Clifford and M. P. Seah, Meas. Sci. Technol. 20, 095103 (2009). F. Marinello et al., Meas. Sci. Technol. 18, 689 (2007).

However, in the correction described in Patent Document 1, when the measurement target has unevenness and the average height of each main scanning line is not uniform, the average height of each main scanning line will fluctuate. There is a problem that high-precision correction cannot be performed.
Further, in the correction described in Non-Patent Document 1, since the time-varying drift deformation occurring in one image is obtained by linear interpolation of the drift amounts obtained from a plurality of measurement images, an error from the actual drift amount is obtained. Will become bigger. In particular, when nonlinear drift deformation occurs between the main scanning lines, the error tends to increase.

The drift correction shown in Non-Patent Document 2 does not consider convolution derived from the probe shape of the probe. In general, since the tip of a probe (probe) of a scanning probe microscope has a finite size, when the sample is scanned by tracing this, the probe shape is convoluted with the measurement result. The convolved shape depends on the probe shape (the apparent probe shape when the probe is vibrated), and therefore the convolved shape varies depending on the scanning direction of the probe.
That is, since the main scanning direction is rotated by 90 ° between the initial measurement as the shape result and the measurement used as the reference in the sub-scanning direction, the probe shape convoluted in the measurement result is also different, and the corresponding amount is an error. Will be corrected. For example, when using the amplitude modulation mode that is currently used most in atomic force microscopes, even if the probe tip shape has a radius of curvature of several nanometers, the apparent probe is caused by the vibration of the probe. The tip shape is several tens of nm. When the uneven surface of the sample surface is scanned with this probe, the probe contact point, which is the reference for data acquisition, is different from the actual contact with the sample surface. Error occurs. An error similar to this also occurs in modes other than the amplitude modulation mode such as the frequency modulation mode and the contact mode.

  Furthermore, even if it is assumed that there is no drift in the horizontal direction, in the actual measurement, there is measurement noise such as noise derived from the apparatus and environmental noise, and therefore data with exactly the same coordinates cannot be acquired. For this reason, the accuracy of the combined calculation of the normal raster scan measurement result and the drift correction measurement result becomes low when the number of points used for the combination is small. In this drift correction method, in order to correct a normal raster scan measurement result, a combination process is performed for each main scan line of the measurement result. The number of points that can be used for calculation in this combination processing is equal to the number of main scanning lines in the drift correction measurement result, and it is necessary to reduce the number of points to prevent the lift correction measurement result from being affected by drift. Calculations are sensitive to measurement noise.

Scanning probe microscopes are widely used for shape measurement and physical property measurement in combination with shape measurement in the manufacturing process and research and development of semiconductor devices. With the recent miniaturization of semiconductor devices, the shape of a measurement target is also reduced, and the size thereof is often 100 nm or less, and the accuracy of shape measurement required for quality control is 1 nm or less.
Thus, highly accurate dimension measurement is indispensable for manufacturing management of semiconductor devices, and the measurement accuracy greatly affects the yield of products.

  Therefore, an object of the present invention is to realize highly accurate shape measurement even for a semiconductor device having a small measurement dimension by correcting a drift generated during measurement with high accuracy.

  In order to solve the above problems, a drift correction method for a scanning probe microscope according to the present invention is a drift correction method for a scanning probe microscope that scans a probe to measure the shape of the sample surface, and scans the sample surface. Scanning the probe along the line to perform shape measurement along the scan line, moving the probe, scanning the probe along the next scan line, and repeating the shape measurement; After the shape measurement along the specific scanning line is completed in the first step, the probe is scanned along the reference shape measurement scanning line set in advance on the sample surface, and the reference shape along the scanning line is measured. By comparing the current reference shape measurement result of step 2 and the second step with the previously measured reference shape measurement result, the time elapsed between them A third step of calculating the drift amount generated by the third step, and a fourth step of correcting the current shape measurement result obtained in the first step based on the drift amount calculated in the third step. Composed.

  The scanning probe microscope according to the present invention is a scanning probe microscope that measures the shape of the sample surface by raster scanning in which the probe is scanned in the main scanning direction and the sub-scanning direction. Obtained when the probe is scanned by the main scanning mechanism, a sub-scanning mechanism for positioning to a predetermined main scanning line, a control device for controlling the main scanning mechanism and the sub-scanning mechanism, and the main scanning mechanism. It has a recording device that records measurement results for each main scan line, and an arithmetic unit that calculates the measurement results recorded by the recording device, and the control mechanism ends the scanning for measurement along a specific main scan line. When it is determined that the reference shape measurement scan is performed, a command is sent to the sub-scanning mechanism and the main scanning mechanism to perform a reference shape measurement scan along a preset reference shape measurement scan line. Based on the measurement result from the recording device, the reference shape measurement result obtained this time was compared with the reference shape measurement scan obtained previously, the drift amount based on the passage of time between them was calculated, and this time obtained The measurement result by the measurement scan was corrected.

  According to the present invention, by comparing the reference shape measurement result obtained this time with the reference shape measurement scan obtained previously, the biaxial direction of the height direction and the main scanning direction based on the passage of time between the two The amount of drift can be accurately calculated, and the required measurement accuracy can be achieved even with a small semiconductor device by correcting the measurement result obtained this time based on this drift amount. .

FIG. 1 is a diagram illustrating the order of measurement main scanning and reference shape measurement main scanning. FIG. 2 is a diagram showing the order of movement of the probe (probe). FIG. 3 is a diagram showing how to obtain the sum S _{1 of the} nearest point distances between two different point groups in the process of calculating the drift amount by the ICP method. 4, in the course of the drift amount of calculation by the ICP method, the total sum S _l while translational movement in the height direction, in the calculation of the translation and the sum S _l, translational movement amount in the height direction is calculated It is a figure which shows a mode that it is performed. FIG. 5 is a diagram showing a configuration of a scanning probe microscope for correcting shape measurement data. Figure 6 is a diagram showing the amount of drift of the relationship in the case where the correction amount calculation main scanning line with the first measuring main scanning line L _11. FIG. 7 is a diagram illustrating a drift amount calculation method when the correction amount calculation main scan line is the latest measurement main scan line.

  Embodiments of the present invention will be described below with reference to the drawings.

In this embodiment, the shape of a silicon line pattern is measured by an atomic force microscope (AFM). Causes of distorting the AFM image that is the measurement result include hysteresis, creep, and thermal drift of the scanner driving element. Among these, the hysteresis and creep of the scanner driving element can be corrected by measuring the scanner with a position sensor and performing feedback control using the result.
Therefore, in this embodiment, an AFM having a feedback control function using a position sensor is used in order to separate the influence of hysteresis and creep from thermal drift.

In the present embodiment, the first main scanning line is set as a reference shape, and the time change of the drift amount is examined by alternately repeating the sample shape measurement and the reference shape measurement. That is, a probe (probe) that measures the sample shape while sequentially moving in the sub-scanning direction in the odd-numbered main scan in the time series of the main scan, and returns to the first position and measures the reference shape in the even-numbered scan. ) Control.
Here, in FIG. 1, the X-axis direction is the main scanning direction of the probe, the Y-axis direction is the sub-scanning direction of the probe, and i (i = 1, 2,..., N) is the main scanning in which the probe performs measurement main scanning. The line number, j (1,2) is a role number (measurement main scan is 1, correction amount calculation main scan, ie, reference shape measurement main scan is 2), and each main scan is L _ij . At that time, the first main scanning L ₁₁ for measurement is performed. Measurement results of the measurement main scanning L ₁₁ is set as a reference shape, the next main scan, after the main scan again return direction, along the same main scanning line as the measurement main scanning L _11, the correction amount measuring the calculation main scanning L _12. Note that L ₁₁ is also the result of the first sample shape measurement, and can also be referred to as a reference shape.

Next, it moves in the sub-scanning direction by −ΔY, and the second measurement main scan L ₂₁ is performed. Thereafter, in parallel with the main scan in the return direction, the control returns to the sub-scan direction again by ΔY, and is used for calculating a correction amount for correcting the measurement result obtained in the main scan L ₂₁ along the same main scan line as L ₁₁ . performing the main scanning L _22.
Next, it moves in the sub-scanning direction by −2ΔY in parallel with the main scanning in the return direction, performs the third measurement main scan L ₃₁ , and again in the sub-scanning direction by 2ΔY in parallel with the main scanning in the return direction. back, along the same main scanning line as L _11, performs correction amount calculation main scanning L ₃₂ for correcting the measurement results obtained in the main scanning L _31. FIG. 2 shows the measurement sequence up to the main scans L ₁₁ , L ₁₂ , L ₂₁ , L ₂₂ , L ₃₁ and L ₃₂ .

In this way, the correction amount calculation main scan L _i2 occurs between the time when the first measurement main scan L ₁₁ is performed and the time when the measurement main scan L _i1 is performed with the main scan line number i. for detecting the drift amount, it is performed by the same main scanning line as the measurement main scanning L _11.
In this embodiment as well, as in a normal scanning probe microscope, the main scanning in the return direction after one main scanning is completed is performed in parallel with the movement in the sub-scanning direction. Even if it increases, the time from the end of the measurement main scan to the start of the correction amount calculation main scan does not increase much.

In this way, after the measurement main scan L _n1 is performed on the nth main scan line, the measurement scan returns to the same main scan line as L ₁₁ each time, and the measurement result obtained by this measurement main scan L _n1 is corrected. The correction amount calculation main scan L _n2 for this is repeated along the same main scan line as L ₁₁ .
That is, as shown in FIG. 1, when measured by the probe scan, the chronological _{order, L 11, L 12, L} 21, L 22, L 31, L 32, becomes ··· L _n1, L _n2. The data obtained by each main scan L _ij is composed of a three-dimensional point group.

As described above, in particular, as n increases, that is, as the main scanning line number on which the probe performs the measurement main scan increases, the time elapses from the time when the first measurement main scan L ₁₁ is performed. Deviation occurs in the sample standard. Therefore, after the measurement main scan L _i1 is performed at a specific main scan line number i, the measurement returns to the first main scan line number 1 each time to correct the measurement result obtained by the measurement main scan L _i1. The correction amount calculation main scan L _i2 is performed along the reference shape measurement scan line, and a correction value is obtained from the measurement result obtained thereby, so that each reference shape measurement scan line is measured. It is possible to accurately correct the deviation of the sample reference from the measurement start time.

In order to obtain the correction amount, matching calculation is performed so that the measurement result of L _i2 matches L ₁₁ as much as possible, and the drift amount is obtained. Since L _i2 and L ₁₁ are the results of measuring the shape with the same position sensor coordinates, when the centroids G _i2 and G ₁₁ are obtained simply, the difference ΔG = G _i2 −G ₁₁ is the drift amount. As required.
On the other hand, due to the instability of the probe position control, environmental noise during measurement, and drift in the main scanning direction in the AFM measurement result, strictly speaking, L _i2 and L ₁₁ have different shape results. It has become. In that case, the above-described matching method using the center of gravity may not be optimal, and a more general matching method is required for higher accuracy.
Specifically, one of the points constituting L ₁₁ using arithmetic in the ICP (Iterative Closest Point) method, which is a representative method of point cloud matching, a method derived therefrom, or other point cloud matching methods The calculation is performed so that the group P ₁₁ and the other point group P _i2 constituting L _i2 match each other as much as possible. From the result, the drift in the biaxial direction in the height direction and the main scanning direction during the measurement of L _i2 is calculated. The amount is obtained and this is the correction value.

For example, in the ICP method, as shown in FIG. 3, the distance between each point P _11k (k = 1, 2,..., M) in one point group and each point P _i2k in the other point group is obtained. The minimum value is the nearest neighbor point distance D _k . Then, a calculation for obtaining the sum of squares of D _k S ₁ = Σ (D _k ² ) (l = 1, 2,..., S: s is the number of times of translation), and the translation of the point group P _i2 is arbitrarily set. Repeat s times at intervals. FIG. 4 shows an example in which the total sum S ₁ is obtained while performing translation in the height direction and plotted on a graph relating to the amount of translation. For example, when fitting is performed with a quadratic function on the total sum S ₁ regarding the translational movement amount, the minimum value is obtained. The translational movement amount at this time is obtained as a temporary correction amount in the height direction, and the point group P _i2 is moved accordingly.
The above operation is repeated until the temporary correction amount converges to 0 for the two axes in the height direction and the main scanning direction. The amount of movement in the biaxial direction from the initial position of the point group P _{i2 to} the final position after movement occurs in the height direction and the main scanning direction between the time of L ₁₁ measurement and the time of L _i2 measurement. The amount of drift in the biaxial direction is a correction value C _i .

Each L _i1 of the measurement results L ₂₁ , L ₃₁ ,..., L _n1 excluding L ₁₁ of the first shape measurement result is obtained, for example, by interpolation of correction values C _i in the time series before and after that. The corrected measurement results L ₁₁ , L ′ ₂₁ , L ′ ₃₁ ,..., L ′ _n1 are obtained. The correction amount for the two axes can be obtained by (C _i −C _(i−1) ) / 2 by performing linear interpolation from C _(i−1) and C _i , for example.

Such correction, each scanning _{L 11, L 12, L 21} , L 22, L 31, L 32, the measurement results obtained in ··· L _n1, L _n2 all stored in the memory, after completion of measurement It may be performed by off-line processing, or may be performed by real-time processing in parallel with each measurement. Real-time processing can be realized by a digital signal processing circuit such as FPGA or DSP, and the processing result can be used for correction of the position sensor signal.

Next, the configuration of the scanning probe microscope for correcting the shape measurement data described above will be specifically described with reference to FIG.
When the shape of a silicon line pattern is measured using an atomic force microscope, the AFM image that is the measurement result is generally distorted due to the hysteresis, creep, and thermal drift of the scanner driving element.
Among these, hysteresis and creep can be solved by measuring the scanner with a position sensor and performing feedback control using the result, so here we will separate the effects of hysteresis and creep from thermal drift. An AFM having a feedback control function using a position sensor is used.

In this embodiment, the first main scanning line is set as a reference shape, and the time variation of the drift amount is examined by alternately repeating the sample shape measurement and the reference shape measurement. In other words, probe control is performed such that the sample shape is measured while sequentially moving in the sub-scanning direction in the odd-numbered main scan in the time series of the main scan, and the reference shape is measured by returning to the first position in the even-numbered scan. Do.
As described above, the main scanning lines are classified into two having different roles, the first is the measurement result to be corrected, the second is the correction amount calculation in the biaxial direction (main scanning direction and height direction). Measure for use.

As shown in FIG. 5, in the atomic force microscope 1, a sample 3 such as a wafer is mounted on a sample stage 2 incorporating a coarse movement mechanism, and the tip of the cantilever 4 is placed on the surface of the sample 3. The probe 5 provided in the front is facing. The cantilever 4 is mounted on an XYZ fine movement mechanism 6 that constitutes a positioning mechanism. The XYZ fine movement mechanism 6 moves the probe 5 arranged with respect to the surface of the sample 3 in the X-axis direction (main scanning direction) and the Y-axis. The position sensor 7 has a function of moving in the direction (sub-scanning direction), the Z-axis direction (the height direction with respect to the sample surface), and a position sensor 7 that monitors the movement amounts of the X-axis, Y-axis, and Z-axis in real time. The X axis, Y axis, and Z axis form an orthogonal three-dimensional coordinate system.
Thus, the cantilever 4 and the probe 5 at the tip thereof are scanned with respect to the surface of the sample 3 in the main scanning (scanning operation in the X-axis direction) and sub-scanning (scanning operation in the Y-axis direction). Is moved at a minute distance (such as “nm” level) based on the probe scanning operation, and the amount of movement at that minute distance can also be monitored.

  At the time of measuring the sample surface, an atomic force acts between the surface of the sample 3 and the tip of the probe 5, and the distance between the sample surface and the probe 5 is based on a preset reference value. It is controlled by the controller 8 so as to be kept constant. The controller 8 receives an input of a detection signal from the optical lever type optical detection device 9 that detects deformation of the cantilever 4, compares this detection signal with a reference value, and feedback-controls the Z fine movement portion of the XYZ fine movement mechanism 6. Keep the distance between the sample surface and the probe constant. Further, the controller 8 inputs a signal from the position sensor 7, and the X and Y axis position control instruction values of the XY fine movement mechanism 6 of the XYZ fine movement mechanism 6 are matched so that the values indicated by the signals from the position sensor 7 coincide. Feedback control of operation. When the probe 5 scans the surface of the sample 3 in the XY direction, these control states are always maintained. At this time, the trajectory regarding the X axis, Y axis, and Z axis of the XYZ fine movement mechanism 6 obtained by the position sensor 7 is Measurement data (shape measurement data represented by a three-dimensional point group) related to the uneven shape of the sample surface is supplied to the measuring instrument main body 10 including a computer or the like.

The measuring instrument main body 10 controls the measurement operation of the atomic force microscope 1, stores the obtained shape measurement data in a built-in memory, performs arithmetic processing as necessary, performs imaging processing, and displays it on a display or the like. It has a function to display. Therefore, a measurement program, a shape measurement data processing program, and an imaging program are stored in the memory of the measuring instrument main body 10 together with the data storage area.
In the measurement operation of the atomic force microscope 1, the operation of the XYZ fine movement portion of the XYZ fine movement mechanism 6 is feedback-controlled through the controller 8 as described above, and the probe 5 is scanned and moved with respect to the probe 5. Perform data collection.
Further, the memory of the measuring instrument main body 10 also stores a correction program for performing the above-described correction of the shape measurement data. When the arithmetic processing unit of the measuring instrument main body 10 reads out and executes the correction program stored in the memory, as described above, the first measurement is performed after the measurement main scan L _i1 is executed with the main scan line number i. The correction amount calculation main scan L _i2 is repeatedly executed on the same line as the main scan L _11, and the drift amount is corrected.

In the above embodiment, the atomic force microscope 1 is exemplified as the scanning probe microscope. However, the present invention is an atomic force microscope other than the illustrated apparatus configuration, and a scanning type other than the atomic force microscope. It can be widely applied to probe microscopes.
Further, the correction amount calculation main scan L _i2 is performed for all the measurement main scans L _i1 by returning to the same line as the first measurement main scan L ₁₁ , but various changes are possible.

That is, after the measurement main scan L _i1 is performed a plurality of times, for example, every 5 times, that is, L _{11
L 51, L 10 1, L} 15 only for ₁ ... of the measurement main scanning correction amount calculation main scanning L _12,
L _52, L _{10 2,} may be performed L _{15 2.} As the interval for performing the correction amount calculation main scan increases, the time required for the correction amount calculation main scan can be shortened. However, since the correction accuracy decreases accordingly, the correction amount calculation main scan is performed according to the required measurement accuracy. By setting the interval for performing the measurement, the measurement time can be shortened.

Furthermore, the correction amount calculation main scan, that is, the setting of the reference shape, can be variously changed except for the first measurement main scan L ₁₁ . For example, the vicinity of the center of the sample may be used as a reference shape, or scanning in a direction orthogonal to the main scanning axis for measuring the shape of the sample may be performed. Further, non-raster scanning may be used instead of the illustrated raster scanning. However, in these cases, the reference shape measurement is also performed by the same scanning method.
The measurement main scan line may be set to a measurement main scan line at an arbitrary time in addition to the method of setting the main scan line to be measured first. As a result, a large movement operation in the sub-scanning direction performed for the correction amount calculating main scan can be reduced. For example, the operation of performing the correction amount calculation main scan by returning to the most recent measurement main scan line position is repeated.

However, in this case, as shown in FIG. 6, the value obtained by calculating the correction amount is not the drift amount from the measurement start point, but the drift amount from the latest measurement main scan line measurement point. The drift amount from the time point of the correction amount calculation main scan to the time point of the latest measurement main scan line is unknown.
Therefore, in order to know the drift amount from the measurement start time, for example, as shown in FIG. 7, the average of the drift amount obtained from the previous correction amount calculation main scan and the drift amount obtained this time is Assuming an unknown drift amount, it is possible to obtain a drift amount from a measurement start time at an arbitrary time.

  As described above, according to the present invention, the required measurement accuracy can be realized even for a minute semiconductor device, so that it can be expected to be employed secretly in the field of semiconductor manufacturing.

DESCRIPTION OF SYMBOLS 1; Atomic force microscope 2; Sample stage 3; Sample 4; Cantilever 5; Probe 6; XYZ fine movement mechanism 7; Position sensor 8; Controller 9; Optical lever type optical detector 10;

Claims (6)

A scanning probe microscope drift correction method for measuring the shape of the sample surface by scanning a probe,
By scanning the probe along the scanning line on the sample surface, the shape measurement is performed along the scanning line, the probe is moved, and the probe is scanned along the next scanning line to perform the shape measurement. A first step to repeat;
After the shape measurement along the specific scanning line is completed in the first step, the probe is scanned along the reference shape measurement scanning line set in advance on the sample surface, and the reference shape along the scanning line is measured. A second step of:
A third step of calculating the drift amount generated by the passage of time between the two by comparing the current reference shape measurement result of the second step and the previously measured reference shape measurement result;
A drift correction method for a scanning probe microscope comprising a fourth step of correcting the current shape measurement result obtained in the first step based on the drift amount calculated in the third step.   After completion of shape measurement along a specific scanning line in the first step, the probe is moved in a direction intersecting with the scanning line, and shape measurement along the next shape measurement scanning line is performed. A drift correction method for a scanning probe microscope according to claim 1.   3. The scanning probe according to claim 1, wherein the reference shape measurement scan line is any one of the measurement scan lines obtained by performing the measurement scan in the first step. 4. Microscope drift correction method.   The drift correction of the scanning probe microscope according to any one of claims 1 to 3, wherein the second step is performed every time the measurement scan in the first step is completed. Method. A scanning probe microscope that scans the probe and measures the shape of the sample surface,
A scanning mechanism for scanning the probe along a scanning line on the sample surface;
A positioning mechanism for positioning to a predetermined scanning line;
A control device for controlling the scanning mechanism and the positioning mechanism;
A recording apparatus for recording a measurement result obtained when the probe is scanned by the scanning mechanism for each scanning line;
An arithmetic device for calculating the measurement result recorded by the recording device,
When the control device determines that the measurement scan along the specific scan line is completed, the control device sends a command to the scan mechanism and the positioning mechanism, and performs a reference shape measurement along the preset reference shape measurement scan line. Let me scan,
The arithmetic device compares the reference shape measurement result obtained this time with the reference shape measurement scan obtained previously based on the measurement result from the recording device, and calculates the drift amount based on the passage of time between the two. A scanning probe microscope characterized in that the measurement result obtained by the measurement scan obtained this time is corrected. The scanning probe microscope according to claim 5, wherein the positioning mechanism moves the probe in a direction intersecting a scanning direction of the scanning mechanism and positions the probe on a next scanning line.

Images