WO1997042537A1

WO1997042537A1 - Peak processing method for confocal optical apparatuses

Info

Publication number: WO1997042537A1
Application number: PCT/JP1997/001517
Authority: WO
Inventors: Takeshi Okamoto; Masato Moriya
Original assignee: Komatsu Ltd.
Priority date: 1996-05-02
Filing date: 1997-05-02
Publication date: 1997-11-13
Also published as: JPH09297012A

Abstract

A peak processing method for a confocal optical apparatus which has a plurality of confocal optical systems arranged two-dimensionally, and a light source of a high coherence, comprising the step of determining the height of a peak by sampling in the direction of the height thereof an actual output waveform Vs of Z-V profile, in which a periodic vibration waveform Vc of the above-mentioned light source is superposed on a theorical confocal output waveform Vt, characterized in that the sampling period is an integral multiple of the height Zt of the actual output waveform Vs of the Z-V profile, and the value of the height of the peak of this intermittently outputted Z-V output is found.

Description

Description Peak processing method in confocal optical device

The present invention provides a confocal optic for use in a three-dimensional shape inspection device that inspects a three-dimensional shape, for example, a shape of an object to be measured, such as a solder bump for IC mounting, whose approximate surface shape is known, at a high speed. The present invention relates to a peak processing method in a so-called tandem-type confocal optical device in which a plurality of confocal optical systems are arranged in a two-dimensional direction, particularly in a device. Background art

This type of confocal optical device is configured as shown in FIG. The light from the light source 1 is reflected by the mirror 40, becomes parallel light via the magnifying lenses 2a and 2b, and enters the hologram 3 as reference light. The hologram 3 reproduces light equivalent to a point light source emitted from each pinhole position of the pinhole array 4 in which pinholes are two-dimensionally arranged by diffracting the reference light.

This reconstructed light is projected onto the object (measured object) 6 via the first objective lens 5a, is reflected and scattered by the object 6, and the reflected light is reflected by the first objective lens 5a, The light passes through the hologram 3 and is focused on the pinhole array 4 via the second objective lens 5b. Note that Fig. 1 represents the light at one pinhole position as a representative.

FIGS. 2, 3 and 4 show the surface of the object 6 in the confocal optical system. With respect to the positional relationship in the optical axis direction (z direction), how the reflected light forms an image near the pinhole array 4 depending on the position of the focal point of the first objective lens 5a of the projected light This is shown. According to this, as shown in FIG. 3, the reflected light passes through the pinhole 4a of the pinhole array 4 only when the focal point and the surface of the object 6 match (focus). In other words, as shown in FIG. 2, when the light spot of the beam is behind the reflecting surface (surface) of the object 6 (back pin), or as shown in FIG. 4, when it is before the reflecting surface (front). At the pin, the reflected light is blocked by the pinhole array 4 and can hardly pass through, and a so-called light receiving aperture function is performed.

By utilizing this characteristic, the amount of reflected light passing through the pinhole 4a while moving the object 6 in the optical axis direction (Z direction) can be reduced by the first and second relay lenses as shown in Fig. 1. The position where the maximum amount of light is obtained is the surface of the object 6 by making the light incident on the two-dimensional photodetector array 8 via 7a and 7b, that is, the object The surface position of 6 can be measured. This is called peak processing. FIG. 1 has the confocal optical system described with reference to FIGS. 2 to 4 and has a two-dimensional arrangement of the pinholes 4a, so that while moving the object 6 in the Z direction, If the amount of reflected light passing through 4a is measured and peak-processed, the surface shape of the object 6 corresponding to each pinhole 4a can be measured. In practice, the first and second objective lenses 5a and 5b are both composed of a telecentric system (also referred to as an afocal system or a tandem arrangement optical system). Instead of moving 6 in the Z direction, move the first objective lens 5a in the Z direction and measure. The light passing through the pinhole 4a forms an image on a photodetector array 8 for detecting two-dimensional light via the first and second relay lenses 7a and 7b, and the individual pinholes are formed. The light passing through 4a is imaged and measured on an independent photodetector. The controller 9 controls the XY position (offset position in the Z direction if necessary) of the stage 10 on which the object 6 is placed, determines the measurement visual field, and moves the first objective lens 5a in the Z direction. Then, the measured value of the photodetector array 8 is read out while performing the peak processing while detecting the position in the Z direction, and the result is displayed, output or recorded.

Next, a manufacturing process of the hologram 3 will be described with reference to FIG. The light source 11 is a coherent light source such as a laser, and the light emitted from the light source 11 is split into two lights by a beam splitter 12 into two lights, each of which is a hologram. It becomes the light source for reference light and object light of 3. If the light from the light source 11 exhibits linearly polarized light characteristics, the polarization direction of the linearly polarized light is rotated by rotation of the first half-wave plate 13a, and the beam splitter 12 is turned on. By using a polarizing beam splitter, the division intensity ratio is set to a desired value.

The reference light and the object light generated by the wavefront splitting by the beam splitter 12 are divided into first, second and third and fourth magnifying lenses 14a, 14b and 14c, The light is enlarged by 14 d and incident on the hologram 3 and the pinhole array 4 respectively. The light transmitted through the pinhole array 4 is diffracted by 4a at each pinhole, becomes light equivalent to a point light source, is converted into parallel light by the objective lens 5b, and is converted into object light by the hologram 3. Incident. By adjusting the second and third half-wavelength plates 13b and 13c, the polarization directions of the reference light and the object light are changed. The direction is set to the desired direction (generally the same direction), and preparation for hologram exposure is completed.

6 to 8 show a first other confocal optical device that does not use a hologram. FIG. 6 shows a first conventional type shown in Japanese Patent Application Laid-Open No. Hei 4-26959 and US Pat. No. 5,239,178. In this case, the light from the light source 1 is magnified by the magnifying lens 2 and enters the pinhole array 4, and the light diffracted by each of the pinholes 4a passes through the beam splitter 15 and The light is projected onto the object 6 by the second and first objective lenses 5b and 5a.

Then, the light projected and reflected and scattered on the object 6 passes through the first and second objective lenses 5a and 5b in reverse, enters the beam splitter 15, and is reflected there. It forms an image on the photodetector array 8. FIG. 7 is a second alternative confocal optical device shown in U.S. Pat. No. 4,806,004. In this case, the light from the light source 1 is magnified by the magnifying lens 2 and enters the pinhole array 4, and the light diffracted by the pinhole 4a is transmitted to the second and first objective lenses 5b and 5a. Therefore, the light is projected on the object 6.

The light projected on the object 6 and reflected and scattered is condensed on the pinhole array 4 having the function of a light receiving aperture through the objective lenses 5a and 5b. The light passing through each pinhole 4 a is reflected by the half mirror 41, and forms an image on the photodetector array 8 on a one-to-one basis via the relay lens 7. In this configuration, the pinhole array 4 that creates a point light source for projection and the pinhole array 4 that serves as a light receiving aperture have the same structure. However, light must be incident from behind pinhole array 4. Therefore, stray light such as light R reflected by the mask of the pinhole array 4 is prevented from reaching the detector array 8 by some method. In the second other confocal optical device, the pinhole 4a does not correspond one-to-one with the pixels of the detector array 8, and therefore, the pinhole array 4 is scanned in the XY plane. The confocal optical system is referred to as a tandem scanning confocal optical system.

FIG. 8 is a tandem type of the same type as the second confocal optical device described in Japanese Patent Application Laid-Open No. H11-50393 and U.S. Pat. No. 4,927,254. 2 shows a scanning optical system. In this method, the pinhole array 4 employs a so-called nip code disk (NipkowDisc) in which the pinholes 4a are arranged in a spiral shape on a disk, and is rotated. By rotating the disk-shaped pinhole array 4, scanning is performed to obtain images between the pinholes 4a, 4a.

In the confocal optical devices having the above-described configurations, from the viewpoint of reducing aberrations in the optical path and improving measurement accuracy, a laser beam having strong coherence may be used as the reference light source. is there. As described above, the amount of light V detected by the photodetector array 8 when measuring the position of the surface of the object 6 using the reference light by the laser light is, as described above, the distance between the confocal optical system and the object. The distance, that is, the amount of light at the highest position (in-focus position) changes depending on the surface height Z of the object 6, and the amount of light reflected at the highest position (focusing position) becomes maximum. Is measured.

A theory showing how the amount of light V changes with respect to the surface height Z of the above object 6. The theoretical confocal output waveform Vt has a gentle mountain shape in the output Z-V profile shown in FIG. On the other hand, the light source at this time has a periodic vibration waveform Vc whose amplitude changes depending on the light amount V because it is laser light. As a result, the actual output waveform Vs input to the photodetector array 8 becomes a waveform in which the periodic oscillation waveform Vc is superimposed on the theoretical confocal output waveform Vt.

The above Vt, Vc, Vs are obtained by the following equations.

V t = F (Z)

V c = ∑ (A n X F n (Z) x S i n (2 TT X Z / (M n x λ) + C n))

V s = V t + V c

Where: S is the dominant wavelength of the coherent light source

A n: Coefficient of vibration component

F n (Z): coefficient of vibration component

M n: Positive integer including 0 or reciprocal of integer excluding 0

(1 / N)

C n: phase of vibration component

In FIG. 9, only one vibration component is described to simplify the description. In the following description, for the same reason, one component is shown, but there is essentially no difference in the function and effect of the present invention even with the n component.

Figure 9 shows that the Z-V output wavelength at the theoretical confocal is superimposed by an integral multiple of the dominant wavelength λ of the light source with strong coherence, or 1 / integer. ing. In particular, in a device that performs peak processing by moving the objective lens 5a, M n = 1/2, and Zt = i / 2.

In such a confocal optical device, the object 6 (or the first objective lens 5a) is slowly moved in the direction of the optical axis of the optical system, and the output at this time causes the VZ profile. When the peak amount of the file is processed and the mere light amount Z at the peak portion is output, the detection is performed using the actual output waveform Vs on which the periodic vibration waveform Vc is superimposed on the surface height Zn of the object 6 to be detected. The height Z s is output, and the detection error of | Z s — Z n | occurs.

If the Z-V characteristic shown in Fig. 9 above is output by intermittent sampling as the object 6 moves, this sampling period is the period Zt of the periodic vibration waveform Vc. If not synchronized, the above output error is further enlarged.

An object of the present invention is to provide a peak processing method in a confocal point optical device for solving the above-mentioned problems in the conventional technology. Disclosure of the invention

In order to achieve the above object, a first aspect of the present invention provides:

A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic vibration waveform Vc of the above light source is superimposed on the theoretical confocal output waveform Vt. — In a peak processing method in a confocal optical device in which a real output waveform Vs of a V profile is sampled in a height direction to obtain a height of a peak portion,

The sampling period is an integral multiple of the period Zt of the actual output waveform Vs of the Z-V profile, and the intermittently output Z-V output This is a peak processing method in a confocal optical device in which a value of a height of a peak portion is obtained.

According to this configuration,

Z—When the actual output waveform Vs of the V profile is sampled in the height direction, the sampling interval is an integer multiple of the period Zt of the actual output waveform Vs. The effect of the vibration of the actual output waveform V s appears evenly at any sampling position.

A second aspect of the present invention provides:

A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic vibration waveform Vc of the above light source is superimposed on the theoretical confocal output waveform Vt. In a peak processing method for a confocal optical device in which a real output waveform Vs of a Z-V profile is sampled in a height direction to obtain a height of a peak portion, the Z-V profile is In the height direction, the light amount at intervals of an integral multiple of the period Zt of the actual output waveform Vs is integrated at each interval, and the height value of the peak portion is obtained for each integrated Z-V output. A peak processing method in a focusing optical device.

A third aspect of the present invention,

A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic vibration waveform Vc of the above light source is superimposed on the theoretical confocal output waveform Vt. The peak processing method in a confocal optical apparatus in which the actual output waveform Vs of the Z-V profile is sampled in the height direction to obtain the height of the peak portion, the Z-V profile In the height direction of the signal, the light quantity at intervals sufficiently larger than the period Zt of the actual output waveform Vs is integrated, and each integral Z — V output This is a peak processing method in a confocal optical device in which a value of a height of a peak portion is obtained for a force.

According to these configurations,

By integrating the actual output waveform V s over an integral multiple of the period Z t of the oscillation or at a period sufficiently larger than the period Z t, the oscillation of the actual output waveform V s is smoothed. You.

Further, in the above configuration,

The step-moving means is used as a moving means for sampling the actual output waveform Vs of the Z-V profile in the height direction, and the feed step amount of the step-moving means is actually output. It is preferable to set 2 to 10 steps for the period Zt of the waveform Vs.

According to this configuration,

The movement in the height direction within the integration section is performed in steps of 2 to 10 in each integration range.

As described above, when the light source of the tandem confocal optical device has strong coherence, for example, when a laser beam is used, the main wavelength of the light source; A 1 / integer vibration phenomenon appears, and the peak processing of the simple Z-V characteristic is inaccurate and cannot be applied to fine measurement and inspection. However, according to the method of the present invention, the vibration phenomenon of the light source is averaged. As a result, high-precision measurement and inspection can be performed, and can be adequately applied to fine measurement and inspection that could not be conventionally applied. BRIEF DESCRIPTION OF THE FIGURES

BRIEF DESCRIPTION OF THE DRAWINGS The present invention is illustrated in the following detailed description and attached drawings illustrating embodiments of the invention. Some aspects will be better understood. The embodiments shown in the accompanying drawings are not intended to specify the invention, but merely to facilitate explanation and understanding.

In the figure,

FIG. 1 is an explanatory diagram showing the configuration of a tandem-type confocal optical device. Figure 2 is an explanatory diagram showing the imaging state of the reflected light near the pinhole.

FIG. 3 is an explanatory diagram showing an image forming state of the reflected light near the pinhole.

Fig. 4 is an explanatory diagram showing the image formation state of the reflected light near the pinhole.

FIG. 5 is a diagram illustrating the configuration of an optical system that exposes a hologram. FIG. 6 is a structural explanatory view showing a tandem confocal optical device having another structure.

FIG. 7 is a configuration explanatory view showing a tandem confocal optical device having another configuration.

FIG. 8 is an explanatory diagram showing a tandem confocal optical device having another configuration.

FIG. 9 is a diagram showing a ZV profile of an actual output waveform. FIG. 10 is a diagram showing sampling for an actual output waveform in the first embodiment of the peak processing method according to the present invention.

FIG. 11 is a block diagram showing a peak processing device.

FIG. 12 is a flowchart showing the peak processing step.

FIG. 13 is a diagram showing a second embodiment of the present invention.

FIG. 14 is a diagram showing a third embodiment of the present invention. BEST MODE FOR CARRYING OUT THE INVENTION

A first embodiment of the present invention will be described with reference to FIG. The waveform shown in FIG. 10 is the actual output waveform V s. Focusing on the periodicity of the waveform, V 1, P 1, P 2, P3, P [pi in the output value sampled Ngushi, each sampling height Z sl, Z s 2, Z s 3, ···, obtain Z s _n. Then, Ri by the respective sampling height Z s _n to the this the peak processing, obtaining a detection surface height Z h of the object 6 '. The detected surface height Z h ′ is averaged with the periodic vibration waveform of the actual output waveform V s and approximated to the actual desired surface height Z h.

Since such measurement is continuously performed from the lowest part to the highest part of the object 6, the detection surface height Z h ′ is a relative height from the lowest part. It is obtained and is substantially the same as the surface height Zh to be detected.

The above-mentioned sampling height is an output value corresponding to a plurality of positions in the height direction of the object 6, and this output value is sampled, for example, by the emission of the light source 1 in the confocal optical device shown in FIG. This is done by using the shutter 16 provided on the side, or by blinking the light source 1. The means for moving the object 6 or the first object lens 5a for sampling in the optical axis direction (Z direction) is the same as that of the conventional one, and any method is applicable.

By setting the sampling period equal to or an integral multiple of the oscillation period Zt of the actual output waveform Vs, the influence of the oscillation of the actual output waveform Vs can be even at the sampling position. Will come out. FIG. 11 shows a processing device for performing peak processing of each detection value by the above-mentioned sampling. Reference numeral 20 denotes a stage controller on which the object 6 is placed and which can be moved in the X, Y, and Z directions.10 is a stage controller that controls the movement in each of the above directions, and 21 denotes the above stage. This is a Z-axis encoder that detects the amount of movement of 10 (or the first objective lens 5a) in the Z-axis (height direction). Reference numeral 23 denotes an A / D converter for A / D converting an analog detection value from the photodetector array 8, and the detection value converted into a digital value by the A / D converter is a comparator 24. Is compared with the memory values P11, P12, ... stored in the peak value memory 25, and if it is higher than that value, it is input to the data sector 26. Input to Z-value memory 27. Reference numeral 28 denotes a display device, 29 denotes a CPU, and 30 denotes processing memory.

Fig. 12 shows the flow when one sampling is performed by the above processing device. First, the memory of the Z-value memory 27 is cleared, and the indices i, j, and k in the X, Υ, and Z directions are cleared and initialized. Top 1). Then, for example, the increment of k, i, j in one pixel of each pixel (corresponding to each detection value Dll, D12,...) Of the photodetector array 8 in FIG. (Steps 2 and 3), read out the detected value Dij of one pixel in the photodetector array 8 (Step 4), and then read the detected value Dij in correspondence with the peak value memory 25. The memory value Pij to be compared is compared with the comparator 24 (step 5). As a result, if the detected value is larger than the memory value, the data is transferred from the comparator 24 to the data selector 26. The signal is input, and the detected value of the Z-axis encoder 21 at this time is written to the Z-value memory 27 to perform peak processing (Step 6). The above peak processing is performed on all pixels (L × M = measurement field) of the photodetector array 8 (step 7), and this is performed on all sampling positions in the Z direction (step 8).

FIG. 13 shows a second embodiment of the present invention. In this case, the object 6 or the first objective lens 5a is continuously moved in the Z direction, during which the position of the actual output waveform V s, for example, every two periods ί 1, f 2, f 3, ' , Fn, the integration of this waveform for two periods is performed, and the heights Z s 1 ′, Z s 2 ′, Ζ s at the period integration positions f 1, f 2, f 3,…, f η 3 ', ···, Ζ s η' This is subjected to peak processing to obtain a surface height Z h '.

FIG. 14 shows a third embodiment of the present invention. This is an example in which the integration cycle is asynchronous with the oscillation cycle Zt of the actual output waveform Vs. It is effective to set the integration interval at this time to be sufficiently larger than the period Zt of the actual output waveform Vs, for example, to be twice or more the period Zt of the vibration. Actually, the longer the integration range of the period Zt is, the greater the effect of smoothing becomes, but the resolution of the detection height Zs becomes worse, and the upper limit is about 10 times. This integration can be realized by using an integrating photodetector such as a CCD sensor in the photodetector array 8.

The periodic integration in the second and third embodiments can be realized by using an integrating photodetector such as a CCD sensor for the photodetection array 8. In this case, the actual output waveform V In order to smooth the effect of the vibration of s, it is desirable that the moving speed of the Z-direction moving means within the integration range be smooth or constant. As such, a DC motor, a piezo element, and a voice coil motor are preferably used. On the other hand, a stepping motor may be used for the stepless drive because it is easy to control the setting of the feed amount and the like for the stepless drive. It is effective that the number of movement steps in the integration range is 2 or more. In practice, the number of moving steps in the integration range should be between 5 and 10. The smoothing effect increases as the number of steps increases, but the moving speed in the Z direction decreases.

Although the present invention has been described with reference to exemplary embodiments, various modifications, omissions, and additions can be made to the disclosed embodiments without departing from the spirit and scope of the present invention. Is obvious to those skilled in the art. Therefore, the present invention is not limited to the above-described embodiments, but should be understood as including the range defined by the elements described in the claims and their equivalents. No.

Claims

The scope of the claims

1. A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic oscillation waveform Vc of the above light source is superimposed on the theoretical confocal output waveform V t In a peak processing method in a confocal optical device in which the actual output waveform V s of the obtained Z — V profile is sampled in the height direction to obtain the height of the peak portion, the sampling period is Z — V In a confocal optical system, the height of the peak portion is obtained for the intermittently output Z-V output, which is an integral multiple of the period Zt of the actual output waveform Vs of the profile. Peak processing method.

2. A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic vibration waveform Vc of the above light source is superimposed on the theoretical confocal output waveform Vt. In the peak processing method in the confocal optical device in which the actual output waveform V s of the obtained Z—V profile is sampled in the height direction to obtain the height of the peak portion, the height of the Z—V profile is obtained. In the direction, the amount of light at intervals of an integer multiple of the period Zt of the actual output waveform Vs is integrated, and the value of the height of the peak is obtained for each integrated Z-V output. Peak processing method in optical equipment.

3. A plurality of confocal optical systems are arranged in a two-dimensional direction, a light source with strong coherence is used as the light source, and the periodic vibration waveform Vc of the above light source is superimposed on the theoretical confocal output waveform Vt. Sampled Z — V profile actual output waveform V s is sampled in the height direction to find the peak height In the peak processing method in the confocal optical device as described above, the light quantity at intervals sufficiently larger than the period zt of the actual output waveform V s in the height direction of the Z—V profile is integrated, and each integral Z -A peak processing method in a confocal optical device, in which a value of a height of a peak portion is obtained with respect to a V output.

4. A step moving means is used as a moving means for sampling the actual output waveform V s of the Z—V profile in the height direction, and the feed step amount of the step moving means is used as the actual output waveform V. The peak processing method in the confocal point optical device according to claim 2 or 3, wherein the number of steps is 2 to 10 steps for the period Zt of s.