JPS63201508A - Optical measuring method for surface roughness - Google Patents

Abstract

PURPOSE:To remove noise generated owing to dust and a fine projection by passing a signal through a low-pass filter and correcting it when the quantity of light detected by all photodetectors becomes abnormal. CONSTITUTION:Light emitted by a laser diode 5 becomes parallel luminous flux by passing through a lens 6 and a beam splitter 7 and is converged by a collimator lens 1 on the surface of a body 9 to be measured. Its reflected light passes through the beam splitter 7 and is split into two by a half-mirror 10 and incident on critical angle prisms 2a and 2b, and they are measured by two-split photodetectors 3a and 3b, and 3c' and 3d', whose output unbalance is measured to measure the roughness of the object surface 9. The sum of the outputs of all the photodetectors is compared by a comparator 15 with a reference value Sth, so that a selecting circuit selects and outputs a signal passed through a delay circuit 20 when the sum is larger or the signal passed through the low-pass filter 20 when not.

Description

[Detailed Description of the Invention] (Industrial Application Field) 5. This invention provides an optical surface roughness measurement method for measuring the roughness of a precisely machined surface such as an old magnetic disk master or a semiconductor substrate surface. Regarding.

(Prior art and its problems) As this type of optical surface roughness measurement method, the optical micro displacement sensor (hereinafter referred to as rHIP) described below is used.

A method using SSJ) is well known.

FIG. 4 is a schematic diagram showing the principle of this HIPO8S, in which a critical angle prism 2 is placed on the side of an objective lens 1, and a parallel beam of light is directed from a light source (not shown) through the objective lens 1 onto the surface of the object to be measured. configured to irradiate.

At this time, the surface of the object to be measured is at the focal position of the objective lens 1 (B
), the reflected light from the surface of the object to be measured passes through the objective lens 1 and becomes a parallel beam of light.

The critical angle prism 2 is set so that its reflecting surface is at an angle close to the critical angle with respect to such a parallel reflected light beam, and is oriented at an angle close to the critical angle with respect to the optical axis of the optical path from the objective lens 1 to the critical angle prism 2. A pair of seven-point detectors 3a and 3b are arranged at symmetrical positions.

In the above case where the surface of the object to be measured is located at the focal position of the objective lens 1, the parallel reflected light beam from the surface of the object to be measured is totally reflected by the reflection surface of the critical angle prism 2, and the two photodetectors 3a, 3b receive the same amount of reflected light. Therefore, the two photodetectors 3a, 3
b output a.

At this time, the output of the differential amplifier 4 at the next stage, which receives the signal b and outputs the difference as a shape signal of the object to be measured, becomes 0''.

On the other hand, when the object to be measured is at a position A close to the objective lens 1 side from the focal position B, the reflected light passing through the objective lens 1 becomes a divergent light beam as shown in FIG. Of this diverging light flux, the incident angle of the light that enters above the optical axis of the reflective surface of the critical angle prism 4 is smaller than the critical angle, so it is not totally reflected, and some light is transmitted from the reflective surface to the critical angle prism 4. transmitted to the outside. On the other hand, since the angle of incidence of the light incident on the lower side of the optical axis of the reflective surface of the critical angle prism 4 is larger than the critical angle, it is totally reflected and transmitted to the photodetectors 3a and 3b. Therefore, the amount of light incident on one photodetector 3a decreases, and the output (ab) of the differential amplifier 4 takes a negative value.

Conversely, when the object to be measured is located at a position C away from the focal position B with respect to the objective lens 1, the reflected light passing through the objective lens 1 becomes a convergent light beam and reaches the reflection surface of the critical angle prism 40. The relationship between the incident angles of the incident light is the opposite of the above-mentioned case of the diverging light flux, and the amount of light incident on the other photodetector 3b decreases, and the output (a-b) of the differential amplifier 4 becomes a positive value. becomes. In this way, the differential amplifier 4 obtains an analog displacement output (+)/(-) in which the position of the surface of the object to be measured turns the focal position B "on". This output (a-b)
furthermore, the total amount of incident light on the photodetectors 3a and 3b (a
+b) is removed to give a standardized value (a-b)/(a + b) that is not affected by changes in illuminance of the light source, etc., and the surface roughness of the object to be measured is measured based on this value. It will be done.

FIG. 6 shows a specific configuration of a conventional optical system of the H I PO 35 based on the above principle, and FIG. 7 shows its signal processing system. In the optical system shown in FIG. 6, a laser beam generated from a laser diode 5 is converted into a parallel beam by a urimeter lens 6, the parallel beam is reflected by a polarizing beam splitter 7 to become linearly polarized light, and a 174-wave plate 8.
It is configured to focus light onto the surface of an object to be measured 9 through an objective lens 1 . Then, the reflected light from the surface of the object to be measured 9 passes through the objective lens 1 and 1/4 wavelength plate 8 again and enters the polarizing beam splitter 7 again. .

This reflected light has 174
Since it passes through the wavelength plate 8, its polarization direction is rotated by 90' with respect to the polarization direction of the laser light immediately after being reflected by the polarization beam splitter 7. Therefore, this reflected light passes straight through the polarizing beam splitter 7 as indicated by R in the figure. A half mirror 10 is arranged behind the polarizing beam splitter 7, and the reflected light is split into two by this half mirror 10.
The light is incident on two critical angle prisms 2a and 2b, respectively. Each of the critical angle prisms 2a and 2b is provided with two reflecting surfaces, and the incident reflected light is reflected twice by each of the critical angle prisms 2a and 2b, and the reflected light is reflected twice by each critical angle prism 2a. , 2b, the light is received by two-split photodetectors 3, 3' placed on the exit side of the photodetectors 2b, and photoelectrically converted. In this optical system, each two-part photodetector 3.3' has divided parts 3a, 3b, and 3c'.

Voltage outputs a, b, c, and d are obtained from 3d', and a+b+c+d is obtained by the processing of the signal processing system shown in Fig. 7, which corresponds to the analog displacement output in the case of the principle diagram shown in Fig. 4. It will be done.

In the signal processing system of FIG. 7, the first subtracter 11a is
One of the two-part photodetectors mentioned above.

Voltage output a output from the dividing parts 3a and 3b of 3.

The second subtracter 11b calculates the voltage output c, d output from the dividing parts 3c', 3d' of the other two-divided photo detector 3'.
This is to find the difference (c-d). and,
The outputs of these subtracters 118.11b are sent to the next stage Φ adder 1.
2, and is configured to obtain an output of (a-b)+(C-d). The adder 13 connects the divided parts 3a, 3b, 3c of each two-part photodetector 3.3'.
'.

This is to obtain the sum of voltage outputs a, b, c, and d output from 3d', and its output S2, that is, (a+
b+C+d) and the output of the adder 12 (a-b)+(
c-d) is input to the next-stage divider 14, where the output of (a-b+c-d)/(a+b+C+d), that is, the shape signal S1 corresponding to the surface shape of the object to be measured 9 is obtained. It is configured. Through this frame removal process, the shape signal S1 is extracted as a standardized value that is not affected by changes in the illuminance of the light source. The output S2 obtained by the adder 13, ie (a+b+c+d), is separately taken out as a single signal and used as a light amount monitor signal for checking changes in illuminance of the light source.

In this prior art, as described above, by incorporating two systems of critical angle prisms 2a and 2b, the inclination and reflectance of the object to be measured 9, unstable factors of the light source, etc. are optically canceled out. In addition, the critical angle prisms 2a and 2b each reflect the reflected light from the object 9 twice, thereby optically amplifying the detection sensitivity.

However, in the conventional example described above, there are factors such as dust and minute protrusions on the surface of the object to be measured 9 that scatter the reflected light, and the obtained shape signal @S1 changes to the surface shape as shown by the symbol P in FIG. Even when an uncorresponding abnormal value is shown, surface roughness is measured using this as a correct shape signal, which poses a problem of deteriorating measurement accuracy. In particular, if there is a factor that scatters the reflected light, such as the above-mentioned dust, at the maximum height of the object to be measured, the abnormal value of the shape signal S1 at this time will be directly measured as the maximum height of the object to be measured 9. Therefore, the problem becomes even more serious.

- Optical surface roughness measurement method using HI PO35 with various other improvements to improve measurement accuracy (Japanese Patent Laid-Open No. 59-90007, Japanese Patent Laid-open No. 59-90007)
No. 27207) has also been developed, but none of them takes into account the output abnormality caused by scattering factors of reflected light, such as the aforementioned dust and microprotrusions.

(Objective of the Invention) This invention is intended to overcome the above-mentioned problems in the prior art, and even when there is a factor that scatters reflected light, such as dust or minute protrusions on the surface of the object to be measured, the scattering factor can be An object of the present invention is to provide an optical surface roughness measurement method that can accurately measure the surface roughness of an object to be measured without being affected by the surface roughness.

(Means for Achieving the Object) In order to achieve the above-mentioned object, the optical surface roughness measuring method according to the present invention includes the following steps:
Of the reflected light from the surface of the object to be measured, the first reflected portion of the optical path from the objective lens to the critical angle prism is detected by a plurality of photodetectors located symmetrically with respect to the optical axis of this optical path, and the detection signals of these photodetectors are When measuring the roughness of the surface of the object to be measured using the difference as a shape signal of the object to be measured, the total amount of reflected light detected by the plurality of photodetectors is compared with a predetermined reference value to detect an abnormality in the total amount of reflected light. In the abnormality detection section, the shape signal is corrected by passing it through a low-pass filter.

By doing so, when there is a factor such as dust or minute protrusions on the surface of the object to be measured that scatters the reflected light, noise generated in the shape signal due to the scattering factor is removed by the low-pass filter.

(Example) In surface roughness measurement in the prior art described above, the present invention provides a method for detecting a signal between a light amount monitor signal corresponding to the total amount of reflected light detected by a photodetector and a shape signal corresponding to the surface position of the object to be measured. This method makes use of the following correlations. That is, let us compare FIG. 2(a) showing the waveform diagram of the light amount monitor signal S2 with FIG. 2(b) showing the waveform diagram of the shape signal S in the same time period as this light amount monitor signal. Also, when the level of the light intensity monitor signal is abnormally reduced due to reflected light scattering factors such as dust and microprotrusions as mentioned above, a corresponding acicular abnormal waveform will always occur in the shape signal. I understand.

Therefore, the present invention attempts to remove the abnormal waveform, that is, the noise that occurs in the shape signal in the section where the signal level shows an abnormal decrease by sampling the light quantity monitor signal.

FIG. 1 is a block diagram of a signal processing system used in an embodiment of the present invention. In this example, the sixth optical system is used as the optical system.
The prior art shown in the figure is used as is, and the signal processing system shown in FIG. The signal processing system shown in is connected.

In the signal processing system shown in FIG. 1, the comparator 15 is used to compare the level of the optical monitor signal S2 with the level of a predetermined reference signal 8th. The level of the light amount monitor signal S2 is set to a level at which a decrease in the level of the light amount monitor signal S2 can be determined. The reference signal S1h given from the reference signal generator 16 is input to the inverting input terminal of the comparator 15, and the light amount monitor signal S2 is input to the non-inverting input terminal of the comparator 15. When the value falls below 8th, the comparator 15 outputs a pulse signal.

The pulse width setting circuit 17 at the next stage of the comparator 15 is a circuit for stretching the pulse signal to a predetermined pulse width, and is composed of a mono-multivibrator or the like. Then, the pulse signal whose pulse width has been expanded by this pulse width setting circuit 17 is sent to the next stage signal selection circuit 1.
8 as a logic input to control its operation.

On the other hand, the low-pass filter 19 receives the above-mentioned shape signal S1 corresponding to the surface position of the object to be measured, and filters out the high frequency components contained in the signal, specifically, the scattering of the reflected light due to the above-mentioned dust. The noise-removed shape signal S3 output from this low-pass filter 19 is input to the signal selection circuit 18 described above. In addition to the low-pass filter 19, the shape signal S1 is also input to a delay circuit 20, and the shape signal S1 delayed by the delay circuit 20 is also input to the signal selection circuit 18. The delay time given by this delay circuit 20 is the shape signal @S
1 is set according to the time required for smoothing processing by the low-pass filter 19, and as a result, the shape signal S3 output from the low-pass filter 19 and the delay circuit 20
The shape signal S1 that has passed through is input to the signal selection circuit 18 in synchronization with the shape signal S1.

This signal selection circuit 18 is composed of a programmable amblifier having a function of selectively outputting an input signal according to a predetermined control logic input, and receives the pulse signal given from the pulse width setting circuit 17 as a logic input. In the section where this pulse signal is input, the shape signal S3 from which noise has been removed from one low-pass filter is output, and in the other sections, the shape signal S1 that has passed through the delay circuit 20 is output.

Next, the operation of the signal processing system shown in FIG. 1 will be explained.

The light amount monitor signal S2 input from the non-inverting input terminal of the comparator 15 is compared with the level of the reference signal 8th. When the level of the light amount monitor signal S2 becomes lower than the reference signal 8th during the optical scanning process of the surface of the object to be measured, that is, the total amount of reflected light received by the photodetector abnormally decreases due to scattering factors of reflected light such as dust. When this occurs, the comparator 15 outputs a pulse signal. This pulse signal is expanded to a predetermined pulse width by a pulse width setting circuit 17 at the next stage and inputted to a signal selection circuit 18.

On the other hand, in parallel with this, the shape signal @S1 is input to the low-pass filter 19 and the delay circuit 20, respectively, and the shape signal S1 delayed by the delay circuit 20 by the processing time required by the low-pass filter 9 and the low-pass filter 9
The shape signal S3 that has passed through is synchronized with the signal selection circuit 1 of the next stage.
8 is input. While the pulse signal is input from the pulse width setting circuit 17 to the signal selection circuit 18,
This signal selection circuit 18 outputs the shape signal S3 that has passed through the low-pass filter 9, and when no pulse signal is input, that is, when no abnormal decrease in level is observed in the light amount monitor signal S2, the shape signal S3 that has passed through the low-pass filter 9 is output. @S1 is output as shape output S4.

As mentioned above, in the section where the level of the front t-niter Th No. S2 abnormally decreases, a needle-like noise waveform occurs in the shape signal S1 correspondingly. A shape signal S3 from which high frequency components have been removed by the pass filter 9, that is, a shape signal S from which noise has been removed from the shape signal S1 is output as a correct shape signal S4.

The width of the pulse signal stretched by the pulse width setting circuit 17 is set to compensate for the difference in processing time by the low-pass filter 19 and to be sufficiently equal to the noise generation period, so that the width of the pulse signal is not output from the signal selection circuit 18. Noise is not included in the shape signal S1 that does not pass through the low-pass filter 19.

By doing this, even if abnormal noise occurs in the shape signal S1 due to dust, minute protrusions, or other factors that scatter reflected light on the surface of the object to be measured, the noise can be reliably detected from the shape signal S1. The roughness of the surface of the object to be measured is always accurately measured.

FIG. 3 shows a comparison between the waveform diagram (a) of the shape signal S1 of the object to be measured obtained by the conventional technique and the waveform diagram (b) of the shape signal S4 according to the above embodiment.

Note that (C) of the same figure shows a waveform diagram of the pulse signal output from the pulse width setting circuit 19 in the above embodiment, together with a partially enlarged view thereof. As is clear from the figure, it is known that the noise portion that showed an abnormal value in the case of the prior art is completely removed by the low-pass filter 19 in the case of the above embodiment.

(Effects of the Invention) As explained above, according to the present invention, the noise generated in the shape signal due to factors that scatter reflected light, such as dust and microprotrusions on the surface of the object to be measured, can be eliminated from the noise generated in the shape signal, which is not used in the prior art. This can be reliably removed by simply adding a simple signal processing system to the equipment used, making it possible to obtain an optical surface roughness measurement method that can be performed at low cost and with good viscosity.

[Brief explanation of the drawing]

Figure 1 is a block diagram of a signal processing system used in an embodiment of the present invention, Figure 2 is a diagram showing the correlation between the light amount monitor signal and shape signal of the object to be measured, and Figure 3 is a block diagram of the signal processing system used in an embodiment of the present invention. A waveform diagram showing a comparison of the shape signal of the object to be measured according to the example and the shape signal according to the conventional technology, FIG. 4 is a waveform diagram of the HIP
FIG. 5 is an explanatory diagram showing the operation of the HI PO35. FIG. 6 is a configuration diagram of a conventional optical system. FIG. 7 is a configuration diagram of a conventional signal processing system. The figure is a waveform diagram showing a shape signal of an object to be measured according to the prior art. 15... Comparator, 19... Low pass filter, Sl... Shape signal,
S2... Light amount monitor signal (total amount of reflected light), 8th...
・Reference signal

Claims (1)

[Claims] (1) A parallel beam of light is irradiated onto the surface of the object to be measured through an objective lens, and among the reflected light from the surface of the object, the amount of reflected light in the optical path from the objective lens to the critical angle prism is determined relative to the optical axis of this optical path. In an optical surface roughness measurement method, the roughness of the surface of the object to be measured is measured by detecting with a plurality of photodetectors located at symmetrical positions and using the difference between the detection signals of these photodetectors as a shape signal. The total amount of reflected light detected by the photodetector is compared with a predetermined reference value to detect an abnormality in the total amount of reflected light, and in the abnormality detection section, the shape signal is passed through a low-pass filter to be corrected. Optical surface roughness measurement method.