JPH01202603A - Fine dimensions measuring apparatus using laser beam - Google Patents

Abstract

PURPOSE:To enable stable measurement of fine dimensions in a submicron range at a high accuracy, by a method wherein an object to be measured is irradiated with condensed dual beam lights to prepare a pattern of intensity of reflected light with the deflection thereof to perform a computation from the intensity of the pattern. CONSTITUTION:Laser light 100 emitted from a laser light source 1 is separated into dual beam lights 101 and 102 and further converted into condensed dual beam lights 105 and 106 having a fine spot diameter with a parallel optical path. Here, a peak-to-peak distance setting means 6 controls a dual beam light separating means 2 through a dual beam light control driver 5 to vary a dis tance (CW) between light intensity peaks while a light deflection control means 7 controls the means 2 through the driver 5 to deflect the light by a range according to a voltage value. Then, the dual beam lights 105 and 106 are made to irradiate an object 8 to be measured, the reflected light thereof are over lapped through a condenser lens 104 and received with a photoelectric converter section 9 to obtain an intensity pattern of the reflected pattern varying accord ing to a distance CW. Then, proper dimension values can be calculated from an intensity level of the pattern.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to an apparatus for measuring minute dimensions in the submicrometer region using laser light.

[Conventional technology]

Conventional micro-dimensional measurement involves illuminating the object to be measured with a white light source, magnifying it at high magnification with a microscope, and then receiving the magnified image with a CCD image sensor and performing signal processing. used.

For signal processing, the video signal output from the CCD image sensor is binarized by determining the slice level, and the number of pixels included in the binarized pattern is counted, and the number of pixels is converted to the size. Dimensions are being measured.

[Problem to be solved by the invention]

Illumination light from a white light source has low spatial coherency;
The light cannot be focused to a minute spot diameter of about 2 μm in diameter, and the light intensity distribution at that time is also unstable. When the dimensions of the object to be measured reach the submicron region of 1 μm or less,
The amount of light reflected from the object to be measured is small compared to the amount of illumination light, and when the reflected light is detected, the difference between the intensity level of the illumination light and the intensity level due to the object to be measured becomes small.

Binarization processing using such a signal with a poor S/N ratio is
It becomes difficult to set the slice level that is determined by the housing, and even if it is set, the binarization pattern becomes thick (changes) even with a slight change in the slice level, resulting in low dimensional measurement accuracy and unreliable measurements. There is a drawback that it becomes

SUMMARY OF THE INVENTION An object of the present invention is to provide a dimension measuring device capable of solving the above-mentioned conventional problems and stably measuring minute dimensions in the submicron region with high precision.

[Means to solve the problem]

In order to achieve the above object, the present invention has the following configuration.

a two-beam light separating means for separating laser light emitted from a laser light source into two light beams that approach each other and travel in different directions; and a light condenser that condenses the two light beams and travels along optical paths parallel to each other. a condensed two-beam light converting means for converting into a 2' beam light, and first and second condensed beams having different peak-to-peak distances between the points where the light intensity of the individual beams constituting the condensed two-beam light is maximum; a peak-to-peak distance setting means for setting the distance to a distance of , and separating the two beams by irradiating the condensed two-beam light having the first and second peak-to-peak distances onto the surface of the object whose dimension is to be measured. a light deflection control means for controlling light deflection when the first and second light deflections are made in a preset area by the means; The reflected light from the object to be measured is detected, and the first
reflected light intensity pattern storage means for creating a second reflected light intensity pattern and storing the first and second reflected light intensity patterns; and at least one of the first and second reflected light intensity patterns. pattern shape determining means for determining the shape of one reflected light intensity pattern; and pattern intensity detecting means for detecting a reference intensity and an extreme value intensity of a preset area for each of the first and second reflected light intensity patterns. , dimension calculation means for calculating first and second dimensions of the object to be measured from the reference intensity and extreme intensity of the first and second reflected light intensity patterns; This configuration is provided with a dimension comparing means for comparing the dimensions of and a dimension determining means for determining the dimension according to the result of the dimension comparison.

[Effect]

Laser light with a Gaussian intensity distribution is separated into two beams, focused on a minute spot diameter, irradiated onto the surface of the object to be measured, deflected, and the reflected light detected. A reflected light intensity pattern corresponding to the combined light intensity distribution of the two beams of light can be obtained, but the distance between the peaks between the two points where the light intensity of the individual beams making up the two beams is maximum is different from each other. If at least two values are set, the light intensity distribution of the two beams changes depending on the distance between the peaks, so different reflected light intensity patterns are detected depending on the distance between the peaks. The shape of the reflected light intensity pattern is unique to the surface shape of the object to be measured - the surface condition. Therefore, by determining the shape of the reflected light intensity pattern, it is possible to distinguish the object to be measured from dust, scratches, etc. Determine whether there is a shape defect, etc.

Furthermore, since the intensity level of the reflected light intensity pattern and the intensity level of the special extreme value position change uniquely depending on the dimensions of the object to be measured, by changing the distance between the peaks of the two beams mentioned above, it is possible to A plurality of reflected light intensity patterns having shapes and intensity levels are detected, and a unique dimension value can be calculated for each reflected light intensity pattern. When calculating this dimension, calculating the dimension only from the extreme intensity level will result in an incorrect measurement because when the intensity level of the two beams of light to irradiate changes, the intensity level of the reflected light will also change. The reference intensity is always detected at the same time, and the dimensions are calculated from the ratio of the extreme intensity to the reference intensity.

A plurality of dimension values obtained for one object to be measured are compared, and if the deviation is within a certain range, the measured value is considered correct and, for example, the average value is determined as the dimension.

〔Example〕

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

FIG. 1 shows a system block diagram of the minute dimension measuring device of the present invention.

Reference numeral 1 denotes a laser light source, for example, a He-Ne laser oscillation tube, which emits laser light 100. 2 is a two-beam light separating means that separates the incident laser light 100 into two light beams 101 and 102 that approach each other and travel in different directions. The separation angle of the two beams 101 and 102 is θm
shall be. Reference numeral 6 denotes an optical system for setting the beam shape and traveling direction of the 2' beam lights 101 and 102 to desired values, which includes the above-mentioned two-beam light separation means 2 and the beam splitter 106.
, a condenser lens 104, and various other lenses and optical elements (not shown).

Reference numeral 4 denotes a condensing two-beam light conversion means, which is specifically composed of an objective lens, and converts the two-beam light 101 traveling at a separation angle θm,
The beam 102 is caused to travel along a parallel optical path, and is converted into two condensed beams 105 and 106 condensed into a minute spot diameter. Reference numeral 5 denotes a two-beam light control driver that generates an electrical signal for electrically controlling the optical operation of the two-beam light separation means 2, and applies the obtained electrical signal to the two-beam light separation means 2. Reference numeral 6 denotes a peak-to-peak distance setting means, which is specifically composed of an AC generator that generates various frequencies with a sine wave, and applies an output signal to the two-beam light control driver 5, and the two-beam light separation means 2. The optical operation of the beam separation is controlled to change the separation angle θm to an angle according to the frequency of the AC signal, so that the light intensity of each beam 105 and 106 constituting the two condensed beams is maximized. The peak-to-peak distance CW between the two peaks is changed. Since the light intensity distribution of the laser beam 100 shows a Gaussian distribution, the two beams of condensed beam 105,
The light intensity distribution of each of the beams 106 is also a Gaussian distribution, and the combined light intensity distribution is also a superposition of Gaussian distributions. The frequency of the AC signal emitted from the peak-to-peak distance setting means 6 is set to a first frequency fm1 and a second frequency fm1, which are different from each other.
The frequencies f''29...fmn are set, and the separation angles of the two beams are set to different angles θmu, θr112...
θmn, and therefore the two focused beams 105, 10
6 inter-peak distances CW are mutually different first distances CW,
, second distance CW2, . . . CW.

Here, n is the number of times the control is performed. Reference numeral 7 denotes an optical deflection control means, which is specifically composed of a ramp wave generator or a DC voltage generator whose DC voltage can be arbitrarily changed.
The optical deflection control means 7 controls the optical operation of the light deflection of the beam separation means 2 to deflect the light to a position according to the voltage value.
The beam is deflected by a width corresponding to the voltage range emitted from the beam.

This optical deflection is such that the above-mentioned peak-to-peak distance CW is the first distance CW.
, the first optical deflection, CW is the second distance CW2, .
...The case of CWn is defined as the second optical deflection.

Here, the optical deflection is the separation angle θ of the two beams 101 and 102.
m is kept at a constant angle, the traveling direction is changed, and, for example, the voltage emitted from the optical deflection control means 7 changes from vs to Vp.
When changing the voltage width △V up to 10
The center position of 5,106 is varied between point 107 and point 108.

Reference numeral 8 denotes an object to be measured whose dimensions are measured, including a dimension measuring section 109;
It consists of a base material part 110 and measures the dimension l of the dimension measuring part 109. The object to be measured 8 is, for example, a pattern wiring portion of an integrated circuit, a magnetic head, etc., and the dimension l of the dimension measuring section 109 is a dimension in the submicron region. The difference in level between the dimension measuring section 109 and the base material section 110 is 1 μm or less, which is within the depth of focus of the irradiated two-beam light 105 and 106.
It is assumed that the reflectance rm of 09 and the reflectance rs of the base material portion 110 are different.

The object to be measured 8 is set at the focal position of the two condensed beams 105 and 106, and a point 107 in a wider range than the dimension l is set.
The two focused beams of light are optically deflected between the point 108 and the point 108. The reflected light from the object to be measured 8 is specularly reflected and travels in the opposite direction, and its course is changed by the beam splitter 103 to form 1 reflected light 11.
1,112 is condensed by the condenser lens 104, superimposed on the - point, and received by the photoelectric conversion unit 9. The photoelectric conversion unit 9 receives light with a light receiver such as a pin photodiode, and performs current-voltage conversion to generate a voltage signal proportional to the intensity of reflected light.

In the following description, the magnitude of the photoelectrically converted voltage signal will be simply referred to as reflected light intensity.

This detection of the reflected light intensity is performed for each light deflection state for the first light deflection and the 20th light deflection. That is, when the voltage signal emitted from the optical deflection control means 7 is a voltage from a ramp wave generator that changes linearly from the voltage ■s to the voltage Vp, the minute voltage change width △Vt in that range
The reflected light intensity is detected every time.

Reference numeral 10 denotes an A/D converter which converts the reflected light intensity of an analog scene into a digital amount of reflected light intensity. Reference numeral 11 denotes a reflected light intensity pattern storage means, specifically a RAM of a semiconductor integrated circuit.
The first reflected light intensity pattern (Pi) is a set of reflected light intensity data obtained by the first light deflection.
is stored in the RAM 116, and a second reflected light intensity pattern (Pi)n, which is a set of reflected light intensity data obtained by the second optical deflection, is also stored in the RAMI 14. The reflected light intensity patterns have different shapes because the combined light intensity distributions of the two condensed beams 105 and 106 are different from each other.

Reference numeral 12 denotes a pattern shape determining means, which is specifically executed by a calculation of a microprocessor, and is used to determine whether the first reflected light intensity pattern (Pi)■, the second reflected light intensity pattern (Pi)■, or both. Determine the shape of.

If the shape of the object to be measured 8, especially the shape of the dimension measuring section 109, is abnormal, the dimension l will naturally also be abnormal, so it is necessary to determine whether or not the object to be measured is appropriate. In general, the reflected light intensity pattern is based on the surface condition, surface shape, and dimensions of the object to be measured 80 and the two focused beams 105 and 106 that are irradiated.
Since a pattern with a unique shape is shown depending on the distance between peaks, it can be determined that there is an abnormality when a pattern different from a pattern expected in advance is obtained.

Reference numeral 16 denotes a pattern intensity detecting means, which specifically performs calculation by a microprocessor, and detects in advance each of the first and second reflected light intensity patterns for the pattern determined to be appropriate by the pattern shape determining means 12. Detect the reference intensity and extreme value intensity of the set area. The details of the reflected light intensity pattern will be described in detail later, but in general, the distance CW between the peaks of the two focused beams 105 and 106
Accordingly, V-shaped, U-shaped, and W-shaped patterns are shown, and the extreme value intensity and reference intensity of the minimum (minimum) or thick portion of each pattern are detected.

Here, the reference intensity is the intensity corresponding to the base line portion of the V-shaped, U-shaped, and W-shaped patterns. The extreme value intensity of the reflected light intensity pattern changes uniquely depending on the dimension JK, but calculating the dimension only from the extreme value intensity is, for example, the irradiation focused light 2
When the intensity level of the combined light beams fluctuates, the intensity level of the reflected light also fluctuates accordingly, resulting in erroneous measurements and poor reliability.

Therefore, the reference intensity is also detected along with the extreme value intensity, but the reference intensity is a constant value regardless of the size of the distance between the peaks of the two focused beams 105 and 106, but the extreme value intensity is determined by the setting of the distance between the peaks. , the extreme value intensities of the first and second reflected light intensity patterns are different from each other.

Reference numeral 14 denotes a dimension calculation means which calculates the dimensions of the object to be measured 80 based on the relationship between the reference intensity and the extreme value intensity detected by the pattern intensity detection means 16. The reflected light intensity ratio, which is the direct ratio of the extreme value intensity obtained from the second reflected light intensity pattern and the reference intensity, is based on the dimensions of the object to be measured 80, the secondary running portion 109 and the base material portion 11.
As long as the material of each of Ru. In calculating the dimensions, a conversion coefficient is required when converting the scene related to the measured reflected light intensity into dimensions. To create this conversion coefficient, for example, the dimensions of various objects to be measured having different dimensions are measured using an electron microscope.

Next, measure the same thing using the method described above to find the ratio of the extreme value intensity to the reference intensity, and create an empirical formula that shows the correlation between the two to find the conversion coefficient. It may be stored in RAM or ROM and used as a reference for dimension calculation.

Reference numeral 15 denotes dimension comparison means, which is operated by a microprocessor and compares the first dimension and second dimension determined by the aforementioned dimension calculation means 14. Reference numeral 16 denotes dimension determining means, which is operated by the same microprocessor and finally determines the dimensions of the object to be measured 8 in accordance with the comparison result of the dimension comparing means 15.

If the difference between the summed first and second dimensions is smaller than the set range, it is assumed that the measurement has been performed correctly, and the average value of both dimensions may be determined as the dimension.

Also, if the dimensions of the two are different, it means that an incorrect measurement was made for both or one of them, and the measurement must be re-measured under the same conditions or by changing the peak-to-peak distance to a different condition. Make the measurement take place, or display NG as measurement impossible.

By changing the peak-to-peak distance CW of the two focused beams 105 and 106 into multiple distance states as described above,
By deflecting light with different combined light intensity distributions and detecting reflected light intensity patterns with different intensity levels, multiple dimensional information can be calculated, and by comparing these dimensions, high precision can be achieved. , enables highly reliable dimensional measurement.

As an optical device for performing the two-beam light separation and light deflection described above, an acousto-optic element (hereinafter abbreviated as AO) is excellent. For details of the AO and the acousto-optic element driver (corresponding to the two-beam optical control driver 5) that controls the AO, please refer to the patent application filed in 1983 by the inventor of the present application.
2-51617 and Japanese Patent Application No. 62-85545, it will be omitted in this specification.

Further, since the reflected light intensity pattern by optical deflection using two beams of light is also described in the above two patent applications, the details will be omitted in this specification.

Next, in the first embodiment of the present invention, the peak-to-peak distance CW is 2.
An example of changing to two states will be explained. FIG. 2 shows an example of the light intensity distribution of two beam lights and the reflected light intensity pattern. Figure 2 (a) and (b) show two condensed beams of light 105 and 106.
A waveform diagram (calculation box) showing the light intensity distribution of , where the horizontal axis is the distance in the diametrical direction of the beam, and the vertical axis is the light intensity. A curve 20 shows the light intensity distribution of the focused beam 105, and a curve 21 shows the light intensity distribution of the beam 106, and each light intensity distribution is a Gaussian distribution. The beam diameter of a laser beam with a Gaussian distribution is defined as the distance between intensity points of 13.5% of the maximum intensity.

The peak-to-peak distance CW in Figure 2 (a) is 1 of the beam diameter.
The peak-to-peak distance CW2 in Figure 2 (b) is also set to 0.8D, which is 80% of the beam diameter, and C
W, is the first peak-to-peak distance, CW.

is the second inter-peak distance.

Curve 22 shows the light intensity distribution obtained by superimposing the two intensities shown by curve 20 and curve 21 in Fig. 2 (a), and curve 26 is the same (curve 20 and curve 21 is superimposed and synthesized, and is the light intensity distribution when the object to be measured 8 is actually irradiated.

In this way, if the distance between the peaks of the two beams differs, the combined light intensity distribution will differ, but the sum of the light intensities at that time is constant.

When the distance between peaks is small, it is an inverted V shape, and when it is large, it is an inverted W shape.
This results in a type of light intensity distribution. Figures 2 (c) and 2) are examples of reflected light intensity patterns, and the first light deflection was performed with the CWl peak-to-peak distance shown in Figure 2 (/Sugima). Figure 2) is an example of the reflected light intensity pattern obtained when the second light deflection is performed at the peak-to-peak distance of CW2 shown in Figure 2 (b) (all calculated values). ,
The horizontal axis of the graph is the deflection voltage value of the lamp wave voltage (corresponding to the amount of optical deflection) when light is deflected, and the vertical axis is the reflected light intensity (relative value). In the case of CW with a small distance between peaks, a V-shaped reflected light intensity pattern is obtained, and in the case of CW2 with a large distance between peaks, a W-shaped reflected light intensity pattern is obtained.

In this way, reflected light intensity patterns of different shapes are obtained depending on the magnitude of the inter-peak distance CW, and therefore the extreme value intensities at the extreme positions of each pattern are also different.

■In the case of a shaped reflected light intensity pattern, the minimum (minimum) intensity V
In the case of a p, W-type reflected light intensity pattern, there is a minimal (minimum) intensity m p and a very thick intensity Cp.

Since the extreme strengths VJ), mp, and Cp each vary uniquely depending on the dimension l, the dimensions are calculated independently from each extreme strength, and the obtained dimension values are compared.

In the above example, there were two types of cases where the distance between peaks was small and when it was large, but the setting of the distance between peaks depends on the dimensional range to be measured, the surface condition of the object to be measured, the allowable dimensional measurement accuracy, and must be set according to the measurement environment, etc.

FIG. 3 shows an example of calculation of the relationship between the dimensions and the extreme intensity of reflected light when the peak-to-peak distance CW is used as a parameter. As explained in FIG. 2, it is convenient to express the inter-peak distance based on the diameter of the laser beam.

Furthermore, in the same way, the dimensions will be standardized and expressed based on the diameter of the laser beam. The following figure will be explained by normalizing the diameter of the laser beam to 1. The horizontal axis of the graph in FIG. 3 is the normalized dimension, and the vertical axis is the reflected light intensity ratio (expressed in %) of the ratio of the extreme value intensity to the reference intensity.

Curve 61 shows the minimum intensity 791 when CW=0.1 (distance between peaks is 10% of the laser beam diameter); curve 32 shows CW=0.
．． 3 shows the change in minimum intensity Vp due to dimensions, and both have a V-shaped reflected light intensity pattern. The curve 33 shows the change in the minimum intensity vp when CW=0.5, but in this case it has a U-shaped shape.

Curve 64 shows the minimum strength m p of the W-shaped pattern when cw=o and a, and curve 65 shows the change depending on the dimensions of the same (very thick strength Cp).As is clear from the graph in FIG. As the peak-to-peak distance CW becomes smaller, the reflected light intensity ratio has a smaller value, and the slope of the curve becomes thicker.Also, when the normalized dimension is small, the change in the reflected light intensity ratio due to the peak-to-peak distance is small (, Conversely, as the normalized dimension increases, the change in the reflected light intensity ratio due to the distance between peaks increases (becomes larger).The larger the change in the reflected light intensity ratio with respect to the change in the normalized dimension means that the detection accuracy of the dimension is higher (becomes larger). In order to cope with this, when the dimension to be measured is relatively large, high-precision measurement is generally performed by setting the peak-to-peak distance CW to be small and detecting the reflected light intensity pattern. When the dimensions are small (normalized dimensions are 0.2 or less), the reflected light intensity pattern may be measured by changing the peak-to-peak distance CW from a small state to a large state, for example, in three or more ways.

As described above, high-accuracy measurement is possible by setting the peak-to-peak distance and determining the number of times the peak-to-peak distance is changed according to the dimension region to be measured.

The calculation example shown in FIG. 3 is an example when the beam diameter of the single laser beams 105 and 106 shown in FIGS. 2(a) and 2(b) is normalized to 1. In the case of an optical system in a micro-dimensional measurement device, the beam diameter of a single focused laser beam is approximately 1.8 μm, so the actual dimension value can be determined by multiplying the standardized dimension value by 1.8. . Further, the graph shown in FIG. 3 shows that the relative reflectance, which is the ratio of the reflectance rm of the dimension measuring section 109 and the reflectance ra of the base material section 11o, is 0.25.
This is an example of the case. If the relative reflectance differs, the shape of the graph in FIG. 3 naturally changes, but the relative reflectance is small (as the relative reflectance is small, the value of the reflected light intensity ratio decreases).

Although the reflected light intensity ratio is used in the explanation of FIG. 3, the reflected light intensity ratio is the ratio between the extreme value intensity and the reference intensity.

Next, reference strength will be explained.

FIG. 4 shows an example of a reflected light intensity pattern for explaining the method of detecting the reference intensity. FIG. 4(a) is an example of a V-shaped reflected light intensity pattern, in which the position of point 40 is the minimum position. The horizontal axis of the graph is the voltage value of the lamp wave voltage generated by the optical deflection control means 7 that controls optical deflection. Let Va be the voltage value of the lamp wave voltage when the light is deflected to the minimum position of point 40. The reference intensity in the case of this reflected light intensity pattern is the intensity at a point 42 corresponding to the voltage Va at a portion 41 indicated by a dotted line. Areas 43 and 44, which have a high intensity level across the minimum area, are reflected light when the focused two-beam light 105, 106 is irradiated onto the object to be measured 80 and the base material part 110, and correspond to the baseline area. .

In addition, the reflectance rs of the base material part 1100 and the dimension measurement part 10900
90 reflectance r s ) r m.

The waveform 45 in the graph of FIG.
5.106 is a reflected light intensity pattern when the base material portion 110 of the object to be measured 8 is irradiated and the light is deflected, and is a graph showing the deflection characteristics. Generally, when using an acousto-optic element,
The deflection characteristics do not have a completely flat waveform in the entire deflection region, and the deflection efficiency decreases at both ends of the deflection region, but shows almost constant deflection efficiency in the center.

Therefore, the dimension measuring section 109 of the object to be measured 8 may be initially set so as to be located approximately at the center of the deflection area. Point 4
6 corresponds to the position where the deflection voltage is Va. The magnitude of the reflected light intensity at this time becomes the reflected light intensity at point 420 shown in FIG. 4(A), that is, the reference intensity.

Therefore, the data of the reflected light intensity of the waveform 45 shown in FIG. The standard strength may be determined by referring to the data.

Another method is to detect the reflected light intensity pattern shown in FIG. Then, the light may be deflected again by the base member 110, and the reflected light intensity at the position corresponding to the extreme value location may be detected and used as the reference intensity.

In the above description, the reference intensity is an example in which the intensity at one point is used, but the reference intensity may be determined by using the average value of the reflected light intensity in a certain area centered on the extreme position. Using the above method, the reflected light intensity ratio explained in FIG. 3 can be obtained from the ratio of the extreme value intensity and the reference intensity.
If the ratio of reflected light intensity is small, a 1% variation in the reflected light intensity ratio will result in a variation within 0.02 μm when converted to dimensions, so when detecting the reflected light intensity and performing A/D conversion as explained in Figure 1. A/
The D converter 10 converts 10 bits into 0.01μ.
Dimensions can be calculated with an accuracy of m or less.

Next, reflected light intensity pattern determination will be described.

FIG. 5 shows various examples of V-type and W-type reflected light intensity patterns. The curve 50 in FIG. 5(a) is an example of a normal V-shaped pattern, where the minimum portion 501 is monotonically decreasing, and the right side 506
The part increases monotonically. A line 504 is a slice level for binarizing the V-pattern curve 50. Since the intensity change of the curve 50 of the reflected light intensity pattern changes according to a certain pattern, the pattern width 1 at the slice level 1504 set for the minimum part 5010 intensity level has a certain width and is binarized. When this happens, the double saliva pattern shows one falling and one rising change. In such a case, it is determined to be normal.

The curve 51 in FIG.
In the example of the reflected light intensity pattern when the light is attached to the entire surface of 9, the intensity level of the minimum part 5100 decreases, and at the slice level 511 set for the minimum intensity level, the intensity level of the minimum part 5100 decreases.
The width of the binarization pattern when it is converted into a value is also thick (it becomes thick).The change in the binarization pattern is once for both falling and rising edges, as in the case of FIG. 5(A).

In this way, even if the change in the binarization pattern is normal, if the extreme value intensity level and the binarization pattern width are abnormal, it can be determined that the pattern is abnormal. In this example, it is assumed that the reflectance of dust is smaller than the reflectance rm of the dimension measuring section 109. On the other hand, if the reflectance of the dust is thick (i.e., larger than the reflectance rs of the base material portion 110), the reflected light intensity pattern will have an upwardly convex peak.The curve 52 in FIG. Measuring section 10
9 is an example of a reflected light intensity pattern when minute dust is attached to the end of the lens 9.

In this example, there are two minimum parts 520 and 521, and the intensity of each minimum part is at the same level as the minimum intensity level according to the dimension to be measured. line 52
2 is the slice level for converting into 2 (W), and the change in the binarization pattern when binarizing is different from the previous case,
It can be determined that this is an abnormal pattern since it shows two falling and two rising changes.

As described above, in the V-type reflected light intensity pattern, normal and abnormal patterns can be easily distinguished by examining the extreme value intensity level and the binarized pattern.

In the above explanation, the slice level for performing binarization judgment was set to a single level, but if you set multiple slice level values and perform binarization judgment, pattern judgment Improves accuracy.

The curve 53 in FIG. 5) is an example of a normal W-shaped pattern, which has two minimum parts 530, 561 and one maximum part 562, and in particular, the minimum intensities of the two minimum parts are equal. Fifth
A curve 54 in the figure (E) is an example of an abnormal W-shaped pattern, and the intensity levels of the two minimum portions 540 and 541 are different.

In this case, like the case in Figure 5 (c), there is minute dust attached, but the pattern can be easily determined by comparing the intensity levels of the two minimal parts. be. The waveform 55 in FIG. 5 (f) is an example of an abnormal W-shaped pattern, and as in the case of FIG.
．． This results in a complex pattern in which not only 551.552 but also a plurality of maximum portions 556.554 exist.

In this example, pattern determination is possible by detecting the number of times the position has an extreme value. Here, in order to detect the extreme value position, it is sufficient to perform differentiation or difference processing on the pattern 55 and detect a point where the differential value becomes O.

By the method described above, it is determined whether the reflected light intensity pattern is normal or abnormal, and dimensions are calculated for the normal pattern.

Considering the reflected light intensity ratio, pattern determination, and the range of dimension values to be measured as explained above in FIGS. 3 and 5, there are combinations of desirable settings for the distance between the peaks of the two condensed beams.

Second example: The first peak-to-peak distance CW is set to two types: 0.1 and the second peak-to-peak distance CW2 is set to 0.3 (both shown as standardized values). The first and second optical deflections are performed and detection is performed in a V-shaped pattern state,
Pattern judgment is performed in one of the states, for example, CW, "0"
．． Do it with 1.

In the case of a normal pattern, if the difference between the first and second dimension values of the two V-shaped patterns is within a set range, the average value of both is determined as the dimension. On the other hand, if the dimensional difference between the two is larger than the set range, the second peak-to-peak distance is further set to, for example, CWs" 0.5 to perform optical deflection, and the third dimension value in this case is For example, if the difference between the second and third dimension values is within the set range, then the average value of the two is taken as the dimension.Also, if the difference between the three dimensions is If it is outside the setting range, either re-measure from the beginning or display NG.

This example is effective when the size is relatively large.

Third example: The first peak-to-peak distance CW1 is 0.1, and the second peak-to-peak distance is cw, = Q, 5, CWs.
0.8, the light is deflected three times, and detection is performed in the pattern states of ■-shape, U-shape, and W-shape. Pattern determination is performed under the condition of CW3=0.8, and in the case of normality, the dimension values obtained under the conditions of CW1=0.1 and CW2=0.5 are compared. If the difference in dimension values is within the set range, the average of the two is taken as the dimension. If it is outside the setting range, the dimension value under the condition of CW = 0.8 is calculated and a comparison is made between the three, but the following processing is the same as that described in the second embodiment. This example is effective when the size is relatively small.

Next, FIG. 6 shows an example of a flowchart of the contents of the present invention based on the method described in the third embodiment. Step 600 sets the first peak-to-peak distance CW1 to, for example, 0.1, and step 602 sets the first peak-to-peak distance CW1 to 0.1.
10 was deflected in advance to detect the reflected light intensity, and as shown in Fig. 4 (
A waveform as shown in b) is obtained and used as a reference pattern, and the reflected light intensity is stored in step 604. This step 60
2 and 604 are performed to obtain the reference intensity. Step 606: First peak-to-peak distance CW.

A first light deflection is performed at , and a first reflected light intensity pattern is detected and stored. Step 608 calculates the extreme value intensity V of the first reflected light intensity pattern. , detect. Step 610 is the extreme value intensity V detected in step 608,
8 to calculate the approximate value of the dimension and set the second peak-to-peak distance when performing the next second light deflection.If the area is determined to be relatively small in size, for example, C
Set W2=0.5, cw3=o, s. Step 6
Step 12 stores the reflected light intensity pattern obtained when light deflection is performed in the state cw, = o, s, and step 614 stores the reflected light intensity pattern obtained when light deflection is performed in the state cw, = o, s. . In step 616, the shape of the third reflected light intensity pattern obtained in step 614 is determined, and if the pattern is abnormal, NG is displayed in step 618.

If the pattern is normal, step 620 and step 61
Extreme value intensity ■6□ of the second reflected light intensity pattern obtained in step 2.

Detect. Step 622 detects the deflection voltage value at the extreme position of the first reflected light intensity pattern obtained in step 606, and step 624 detects a position corresponding to the deflection voltage value from the reference pattern stored in step 604. Detect the reference strength of

In step 626, the peak-to-peak distance cw0. cw.

The reflected light intensity ratio of each state is calculated from the extreme value intensity V 608 s V 62° obtained in the state and the reference intensity, and in step 628, the reflected light intensity ratio is converted to a dimension and two dimension values are calculated. do. In step 660, the two dimensional values are compared, and if the dimensional difference is within the set range, step 66
Step 2 determines the average of the two dimension values as the dimension. If the dimensional difference is outside the set range, step 664 is reached;
The extreme value intensity ■634 of the reflected light intensity pattern in the state of the peak-to-peak distance CW obtained in step 4 is detected ~ step 636
In step 668, the reflected light intensity ratio is calculated, and in step 668, the dimension in the state of the peak-to-peak distance CW is calculated. In step 640, the three types of dimensions obtained are compared, and if two of the dimensions are within the set range, then in step 642, the average value is calculated and determined as the dimension. If all three types are outside the setting range, NG is displayed in step 644.

Next, FIG. 7 shows a measurement example in which dimensions were actually measured using the dimension measuring device according to the present invention.

FIG. 7(A) shows the case where a conventional TV camera is used, and FIG. 7(B) shows the case according to the present invention. Dimensional clause used.

The range is from 0.3 μm to 1.1 μm, and the dimensions are measured using an electron microscope and are compared with the results.

In the case of the conventional example shown in FIG. 7(a), it can be seen that dimensions as small as about 0.7 μm cannot be measured accurately.

This is because it becomes difficult to set the slice level when performing binarization.

In the measurement according to the present invention, even dimensions up to 0.3 μm can be stably measured. The optical system used in the dimension measuring device of the present invention has a total light deflection angle of 53 μm, and the reflected light intensity is detected by changing the deflection position every 0.053 μm.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, the light intensity distribution of the probing light to the object to be measured can be easily set to an arbitrary shape, and the dimensions can be determined according to the light intensity distribution. Since dimensions are calculated by a so-called majority vote, highly accurate and reliable dimension measurements are possible.

[Brief explanation of the drawing]

FIG. 1 is a system block diagram explaining the present invention in detail;
Figure 2 shows the first embodiment of the present invention, in which the distance between the beams is 2.
A waveform diagram showing an example of the light intensity distribution and reflected light intensity pattern when changing the type, Figure 3 is a graph diagram showing the relationship between the dimensions and the reflected light intensity ratio when the distance between peaks is changed to various values,
Figure 4 is a waveform diagram showing an example of a reflected light intensity pattern to explain the method of calculating the reference intensity, and Figure 5 is a waveform diagram of reflected light to explain pattern determination of V-type and W-type reflected light intensity patterns. A waveform diagram showing an example of an intensity pattern, FIG. 6 is a flowchart diagram explaining the operation of the third embodiment according to the present invention,
FIG. 7 is a graph showing a comparison of the dimension measurement results according to the conventional example and the present invention. DESCRIPTION OF SYMBOLS 1... Laser light source, 2... 2-beam light separation means, 4... Condensing 2-beam light conversion means, 6...
- Peak-to-peak distance setting means, 7... Light deflection control means, 8... Measured object. 11...Reflected light intensity pattern storage means, 12.
...Pattern shape determining means, 13...Pattern intensity detection means, 14...Dimension calculation means, 15...Dimension comparison means, 16...・Dimension determining means, 105.106... Condensed two beam light, 109.
...Dimension measurement part, 110...Base material part. (A) Figure 2 (.) (C) (D)
Figure 3 Figure 5 (a) (b)
(8) (-) (ya・)('') Figure 6 Figure 7

Claims (1)

[Claims] a two-beam light separating means for separating laser light emitted from a laser light source into two light beams that approach each other and travel in different directions; and a light condenser that condenses the two light beams and travels along optical paths parallel to each other. A condensed two-beam light conversion means for converting into two-beam light, and first and second beams having different peak-to-peak distances between points where the light intensity of each beam constituting the condensed two-beam light is maximum. a peak-to-peak distance setting means for setting the distance; and a two-beam light separation means for irradiating the condensed two-beam light having the first and second peak-to-peak distances onto the object surface whose dimension is to be measured. a light deflection control means for controlling light deflection when first and second light deflection is performed in a preset region; Detecting reflected light from an object to create first and second reflected light intensity patterns for the entire first and second light polarizations, and storing the first and second reflected light intensity patterns. reflected light intensity pattern storage means; pattern shape determination means for determining the shape of at least one reflected light intensity pattern of the first and second anti-emission intensity patterns; and the first and second reflected light pattern intensity detection means for detecting a reference intensity and extreme value intensity of a preset area for each of the intensity patterns; a dimension calculating means for calculating a first dimension and a second dimension; a dimension comparing means for comparing the first dimension and the second dimension; and a dimension determining means for determining a dimension according to the result of the dimension comparison. A minute dimension measuring device using a laser beam, characterized in that: