JPH0562922B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an apparatus for measuring the surface roughness of a surface to be measured, such as an ultra-precision machined surface, by scanning the surface with light.

(Prior art and its problems) In magnetic disks, optical disks, semiconductor wafers, and the like, it is necessary to process the surfaces with ultra-precision processing and to measure the surface roughness after processing with high precision. Mechanical stylus-type roughness meters are usually used to measure surface roughness, but if the surface to be measured is soft, the measurement may scratch the surface. Therefore, it is desirable to use a non-contact surface roughness measuring device. A typical non-contact surface roughness measuring device of this type is an optical device, and FIG. 11 shows an example of a device using a Mireau type interference mechanism. In the figure, light L given from a light source 1 is reflected by a half mirror 2, condensed by a condensing lens 3, and reaches a Mireau type interference mechanism 4. This Mireau type interference mechanism 4 has a light shielding part 6 in a part of the transparent plate 5.
A beam splitter (half mirror) 7 is further arranged in parallel to the transparent plate 5.

A portion of the light incident on the Mireau-type reflection mechanism 4 continues downward in the figure, is reflected by the surface to be measured 9 of the object to be measured 8, and becomes reflected light LS. This reflected light LS passes through the Mireau-type reflection mechanism 4 in the opposite direction (upward) and reaches the condenser lens 3.

On the other hand, the remaining part of the light that reaches the beam splitter 7 from the condenser lens 3 via the transparent plate 5 is reflected on the surface of the beam splitter 7, and then also on the surface of the transparent plate 5 (reference plane 10). reflected light
Becomes LR. Furthermore, beam splitter 7
After being reflected, the light passes through the transparent plate 5 and proceeds toward the condenser lens 3.

Therefore, the light I traveling upward in the figure through the condensing lens 3 becomes interference light between the light LS reflected on the surface to be measured 9 and the light LR reflected on the reference surface 10. The optical path difference between these two types of reflected lights LS and LR is determined by the distance required for the round trip between the beam splitter 7 and the reference surface 10 and the distance required for the round trip between the beam splitter 7 and the surface to be measured 9. This is the difference between the required distance. Therefore, each of these distances is
When written as DR, DS, the value of (DR-DS) and the irradiation light L
A phase difference determined by the relationship between the two reflected lights LR and LS occurs between the two reflected lights LR and LS, and the intensity of the interference light I changes depending on the phase difference. Therefore, the interference light I
Detecting the intensity with the optical sensor 11 changes the distance between the beam splitter 7 and the surface to be measured 9, and therefore the height of the portion of the surface to be measured 9 that is irradiated with light. Then, the roughness of the surface to be measured 9 is measured by translating the object to be measured 8 in the horizontal direction in the drawing and thereby scanning the entire surface to be measured 9 with light.

By the way, in such a surface roughness measuring device,
Distance between beam splitter 7 and surface to be measured 9
Since the intensity change of the interference light I with respect to DS becomes a periodic change whose distance period is the wavelength λ of the irradiation light L, the upper limit for roughness measurement is 1/2 of the wavelength λ (0.3μ for visible light).
m). In addition, since the detection accuracy of interference light is 1/20 to 1/30 of the intensity change, it ends up being 0.03 to 0.015μ.
m is the limit of detection accuracy.

Therefore, in order to perform measurements with high precision that exceeds this limit, the entire Mirow interference mechanism 4 is moved vertically by a distance corresponding to λ/4 using a PZT (piezo element), etc., to correct the phase difference. A technique has been proposed in which the angles are increased by 90°, 180°, and 270° from the original values, and measurements are taken for each of them. In this case, let F ₀ be the detection intensity in the initial positional relationship, and let F ₉₀ , F ₁₈₀ , and F ₂₇₀ be the detection intensities after adding each of the above phase differences, then F ₀ = G ₁ + G ₂ cosφ...( 1) F ₉₀ =G ₁ +G ₂ cos(φ+90゜) =G ₁ +G ₂ −sinφ ……(2) F ₁₈₀ =G ₁ −G ₂ cosφ ……(3) F ₂₇₀ =G ₁ +G ₂ sinφ …… (4) holds true. However, φ is the initial phase difference,
G ₁ and G ₂ are a DC component and an AC component, respectively.

Then, from equations (1) to (4), we get φ=tan ^-1 F ……(5) F≡(F ₂₇₀ −F ₉₀ )/(F ₀ −F ₁₈₀ ) ……(6), but ( Since equations 5) and (6) are equations for calculating the phase difference φ (and therefore the height of the surface to be measured 9) from the four detected intensity values, the error is statistically reduced,
Measurement accuracy can be improved.

In this way, the measurement accuracy of the apparatus shown in FIG. 11 can be improved, but this is done on the premise that the positional relationship between the measuring apparatus (particularly the optical system) and the object to be measured 8 remains unchanged. However, in reality, the positional relationship (interval) between the object to be measured 8 and the apparatus changes over time due to various external vibrations. and,
In order to prevent this, a large vibration isolation table must be provided, which limits the installation location of the device.
Furthermore, if the object to be measured 8 moves slightly up and down depending on the straightness of a mechanism (for example, a slide mechanism) for translating the object to be measured 8 for scanning,
This vibration is reflected in the measurement results.

For this reason, in cases such as nanometer-order roughness measurement, even if improvements are made to improve accuracy as described above, in reality, various vibrations can cause problems between the surface to be measured and the measurement system. Due to the varying spacing, there was a problem in that it was difficult to perform highly accurate roughening.

(Object of the Invention) The present invention is intended to overcome the above-mentioned problems in the prior art, and provides a surface roughness measurement on which highly accurate roughness measurements can be made even if the distance between the surface to be measured and the measurement system varies. The purpose of the present invention is to provide a roughness measuring device.

(Means for Achieving the Object) In order to achieve the above-mentioned object, the present invention targets a surface roughness measuring device that scans a surface to be measured and measures the surface roughness of the surface to be measured. ) A first distance that has a sensor surface positioned at a predetermined location and optically detects the relative distance relationship between the center of the scanned area on the surface to be measured and the sensor surface at that point in time; (b) another sensor surface positioned in a predetermined positional relationship with respect to the sensor surface of the first distance sensor, and a region near the center and the another sensor surface; (c) a second distance sensor that optically detects an average relative distance relationship between the central portion and the neighboring area based on the detection results of the first and second distance sensors; and calculation means for calculating and calculating a relative unevenness relationship between the regions, and calculating the surface roughness of the surface to be measured based on the relative unevenness relationship.

(Example) A. Configuration of Example FIG. 1 is a diagram showing a surface roughness measuring device 100 which is an example of the present invention. This device 1
00 includes a support mechanism 20 for supporting and moving the object to be measured 8, and a sensor head 30 including an optical system and a detection system, which will be described later. Among these, the support mechanism 20 has an electric slide table 22 that moves in translation in the X direction in the figure on a fixed base 21.
performs the above-mentioned translational movement by the driving force of the linear actuator 23.

On the electric slide table 22, there are three
PZTs 24a to 24c are fixedly arranged in a substantially equilateral triangular shape, and a sample stage 25 is mounted on these PZTs 24a to 24c. The object to be measured (sample) 8 is placed on this sample stage 25 . The PZTs 24a to 24c individually expand and contract in the Z direction in the figure in response to a PZT control signal given from the PZT controller 52. As a result, the sample table 25 (and therefore the surface to be measured 9 of the object to be measured 8) has a variable attitude and position in the normal direction n (height in the Z direction in the illustrated example). There is. For this reason, these PZT
24a to 24ct and the like constitute a variable movement mechanism that changes the attitude and normal position of the surface to be measured 9.

On the other hand, the sensor head 30 includes a condensing lens 31 and a Mireau type interference mechanism 32 at the lower end of a hollow cylindrical sensor tube 45. However, in this embodiment, as shown in FIG. 2, a lens (Mirow type objective lens) whose lower surface is flat is used as the condensing range 31, and this flat surface is used as the reference surface 33. Further, a light shielding layer W is embedded near the center of this reference surface 33. Furthermore, the half mirror 34 performs the same function as the beam splitter 7 shown in FIG.

A half mirror 36 that partially reflects the light from the monochromatic light source 35 is provided above the condenser lens 31, and a light shielding plate 37 is provided above the half mirror 36. The center of this light shielding plate 37 has a predetermined size (for example, a diameter
A pinhole 38 of 10μm) is drilled,
Above this pinhole 38 is a photomultiplier tube 3.
9 is positioned. Further, on the lower surface of the light shielding plate 37, three silicon photodetectors 40 are arranged concentrically around the pinhole 38.
a to 40c are fixed at approximately equal angular intervals. These photo detectors 40a to 40c
has a relatively large sensor surface and has a fixed positional relationship with respect to the sensor surface of the photomultiplier tube 39.

As will become clear from the description below, among the above optical configurations, the Mireau type interference mechanism 32,
The pinhole 38 and the photomultiplier tube 39 connect the center P of the area A of the surface to be measured 9 that is currently irradiated with light (FIG. 3) and the light-receiving surface (sensor surface) of the photomultiplier tube 39. ) is formed as a detection optical system of a "first distance sensor" that detects the distance relationship between the two. In addition, Mireau type interference mechanism 3
2 and photodetectors 40a to 40c, three neighboring regions surrounding the center P in FIG.
A detection optical system of a "second distance sensor" is formed that averages the relative distance relationship between each of R _a to R _c and the sensor surfaces of the photodetectors 40a to 40c. That is, in this embodiment, the Mireau type interference mechanism 32 is used as both the first and second distance sensors.

Next, the signal processing system will be explained. The signal processing system of this device 100 is roughly divided into two systems, one of which is a system that calculates surface roughness from detected data. This system consists of (a) a photomultiplier tube 39 and a photodetector 4;
The calculations described later are performed on the respective detection outputs (interference signals) S _p , S _a -S _c of 0a to 40c, and thereby the optical path difference signals D _p , D _a -
Interference signal processing circuits 61p, 61a to 61b that generate D _c (described later), (b) the optical path difference signal
Adder 62 for calculating the sum of D _a to D _c , (c) Adder 6
a divider 63 that divides the output of 2 by "3", and
(d) A subtracter 64 is provided for subtracting the output of the divider 63 from the optical path difference signal D _p . The reason for performing each of these operations will be explained in detail later. The output of the subtracter 64 (local unevenness signal) is given to the computer 50, where it is accumulated and also given to the plotter 65. By measuring the entire surface to be measured 9 by scanning to be described later, the measured roughness state is recorded on the plotter 65.

On the other hand, other signal processing systems are
This is for controlling the expansion and contraction of the PZTs 24a to 24c. In this system, the detection outputs (interference signals) of photodetectors 40a to 40c S _a
~S _c is taken in and a drive control signal is output to the PZT controller 52. In this embodiment, this function is realized by the computer 50.
Further, the computer 50 outputs a scanning control signal for scanning movement to the linear actuator power supply circuit 53.

In the illustrated example, hardware circuits are partially used in the signal processing system for ease of understanding, but it is of course possible to perform all functions of these circuits using software.

B Operation of Embodiment Next, the operation of this device 100 will be explained focusing on the characteristic parts of the embodiment.

First, the light irradiation area A on the surface to be measured 9 and the reference surface 33 are irradiated with light L from the light source 35,
Two types of reflected light LS and LR are generated as a result.
The optical principle from (FIG. 3) to obtaining the interference light I is the same as that of the conventional example. However, as already explained, the light irradiation area A is a larger area than the area (center P) where the unevenness is being detected at that time.

Of the interference light I, the reflected light LS from the center P
The light containing the light is given to the photomultiplier tube 39 through the pinhole 38, and is converted in the photomultiplier tube 39 into a photoelectric conversion signal (interference signal) S _p according to its intensity. Also, of the interference light I,
Light including reflected light from each of the neighboring regions R _a to R _c in FIG.
0a to 40c and photoelectrically converted.

Among these, photo detectors 40a to 4
Since the light-receiving surface of 0c has a somewhat larger area than the cross-sectional area of the pinhole 38, the photoelectric conversion signals (interference signals) S _a to S _c of the photodetectors 40 a to 40 c are generated in the neighboring area.
The signal reflects the average height of the unevenness inside each of R _a to R _c . In other words, in the case of the surface to be measured 9 with unevenness as shown in FIG. 4, the average _heights of the neighboring regions R _a to R _c are
Interference signal S _a reflecting H _c (H _c is omitted)
S _c will be obtained.

Interference signals S _p , S _a to S _c can be obtained by arranging the optical system and detection system in this way, but before starting the actual roughness measurement, first, the computer 5
0 to the PZT controller 52 to once extend the PZTs 24a to 24c within a range of approximately the wavelength λ of the light L. At this time,
The computer 50 performs the above expansion control while taking in the interference signals S _a to S _c . By that
Observe the change in the interference signal S _i (i=p, a to c) with respect to the change in length of the PZTs 24a to 24c (therefore, the overall vertical change in the surface to be measured 9).
As illustrated in FIG. 5, this change is sinusoidal. Through this process, the maximum value S _inax and the minimum value S _inio (i=p, a
Find ~c).

After that, the PZT24a is moved to the position where each interference signal S _a to _{S c} for the neighboring regions R _a to R _c becomes the average value S _a0 , S _b0 , S _c0 of the respective maximum value and minimum value.
~ 24c is expanded and contracted.

These routines are illustrated in FIG. 6. Next, in the step of positioning the PZTs 24a to 24c so that the interference signals S _a to _{S c} have their average values (median values),
S _a0 , S _b0 , S _c0 and the interference signal S _a at that point,
Assuming that the deviations from S _b and S _c are Δ _a , Δ _b , and Δ _c , for PZT24a, Δ _0a = (Δ _a + Δ _b + Δ _c )/3+α (Δ _b + Δ _c −2Δ
_a )...(7) is the deviation parameter, and so that this deviation parameter Δ _0a is zero or a value close to it,
Drive PZT24a. Other PZT24a, 2
For 4c, use the subscripts a, b, and c in equation (7) that are changed cyclically. Further, α is a tilt correction coefficient set in advance.

By doing this, the surface to be measured 9
is adjusted to a height corresponding to point Q in FIG. To do this, point Q or this
At another point Q' whose phase is shifted by an integral multiple of 360°, the interference signal due to the height change of the surface to be measured 9 is
This is because the rate of change in S _i (the differential coefficient in the graph of FIG. 6) is the largest, improving measurement accuracy. Note that the relationship between the direction of change in the height of the surface to be measured and the sign of change in the interference signal S _i is reversed from the above, but the heights correspond to points Q ₁ and Q ₂ in Fig. 5. Adjustments may be made.

These points Q, Q', Q ₁ and Q ₂ are the optical path length of the incident light L until it is reflected by the surface to be measured 9 and returns, and the optical path length of the incident light L until it is reflected by the reference surface 33 and returned. This corresponds to the height of the surface to be measured such that the difference with the length (that is, the optical path difference between the reflected lights LS and LR) is an odd multiple of 1/4 of the wavelength λ of the light L. In other words, at these points Q, Q', Q ₁ , and Q ₂ , the mutual phase difference between the reflected lights LS and LR is an odd multiple of 90°. However, even if the distance corresponds to the vicinity of points Q, Q', etc. (for example, about ±0.2 times the amplitude of the interference signal), a considerable value can be obtained as the rate of change.

In this way, the PZTs 24a to 24c are adjusted to heights corresponding to the same phase difference, and the surface to be measured 9 and the reference surface 33 are optically parallel to each other. However, "optically parallel" means that when the respective reflected lights are caused to interfere, the phases of each part of the cross section of the light beam are the same within one wavelength. Therefore, as in a modified example (FIG. 10) to be described later, the surface to be measured 9 and the reference surface 3
3 may be geometrically non-parallel. By achieving such optical parallelism, the optical path difference between the three interference signals S _a to S _c is equal to the wavelength λ
This avoids a situation where measurements are performed with a deviation of an integer multiple of .

Note that such adjustment processing may be performed only once before performing the actual roughness measurement, but if the surface to be measured 9 is gently undulating,
There may be a case where the height of the surface to be measured 9 deviates from a predetermined range centered on the point Q as the scanning progresses due to the electric table 22 being tilted. In order to deal with such cases, it is necessary to use the interference signals S _a to S _c and the threshold values shown in Figure 5.
It is desirable to constantly compare the interference signals S _{a to S TH} with S TH and perform the above adjustment each time the interference signals S a to S _c exceed the range of ±S _TH .

Also, the maximum value obtained as above
The values of S _inax and minimum value S _inio (i=p, a to c) are sent to the computer 50 via a signal line (not shown) for the purpose of conversion processing (described later) into an optical path difference signal during actual measurement. 1 to the interference signal processing circuits 61p, 61a to 61b shown in FIG.

Next, the actual roughness measurement process will be explained. Note that the following processing is also shown in the flowchart of FIG. First, the linear actuator 23 shown in FIG. 1 is driven to sequentially move the electric table 22 in the x direction, thereby starting optical scanning. Captured interference signals S _p , S _a ~
S _c is the interference signal processing circuit 61p, 61a in FIG.
~61c, respectively. These interference signal processing circuits 61p, 61a to 61c process the optical path difference signal D _i (i=p, a to c) as follows (8) to (10).
It is determined by performing a performance based on the formula.

D _i = (λ/2π) sin ^-1 (K _1i / K _2i ) ……(8) K _ii ≡2S _i − (S _inax + S _inio ) ……(9) K _2i ≡S _inax −S _inio …… (10) (i=p, a to c) The fact that the above equations (8) to (10) are formulas for calculating the optical path difference from the value of the interference signal means that equations (8) to (10) can be Solving for _i , S _i [(S _inax −S _inio ) sin (2πD _i /λ) + (S _inax + S _in
io )]/2...(11) This can be easily understood. Of these, (S _inax - S _inio )/2 is the amplitude, and (S _inax + S _inio )/2 is the amount of movement of the origin.

The optical path difference signal D _i (i=
p, a to c) are processed by the adder 62, the divider 63, and the subtracter 64 as unevenness signals about the center P: Z _p =D _p -(D _a +D _b +D _c )/3...( 12) becomes. As can be seen from this equation, the unevenness signal Z _p
has a value expressing the relative unevenness relationship (positional relationship) between the center P and the neighboring regions R _a to R _c . In other words, the optical path difference signals D _a to D _c are signals that reflect the average heights H _a to H _c within the regions R _a to R _c in each of the neighboring regions R _a to _{R c} , but these is averaged to obtain the average height of the entire light irradiation area A, and the level of the optical path difference signal D _p at the center P is determined based on this average height.

Therefore, even if the surface to be measured as a whole moves up and down due to vibration of the object to be measured 8, and each optical path difference signal D _i (i=p, a to c) shifts accordingly, ( Since these are mutually canceled by Equation 12), it is possible to know the exact unevenness state. Then, this unevenness signal Z _p
is taken into the computer 50, and the relationship between the level and the scanning position is recorded on recording paper by the plotter 65. By performing this process on each position of the surface to be measured 9 by scanning, the roughness of the surface to be measured 9 is measured.

C Data Example FIG. 8 is a graph showing the unevenness signal Z _p (FIG. 8 c) measured by the above embodiment together with the optical path difference signals D _p and D _a (FIG. 8 a, b). However, the horizontal axis is time, and in this example of Figure 8,
Object to be measured 8 to check the influence of external vibrations
is stationary (that is, no scanning is performed). Then, at time _t1 in the figure, an impact force is applied to forcibly vibrate the object to be measured 8.

As can be seen from this figure, the optical path difference signals D _p and D _a (the same goes for D _b and D _c ) fluctuate considerably due to such vibrations, but the unevenness signal Z _p corresponding to the difference between them fluctuates considerably. There is almost no disturbance.

Furthermore, in Figure 9, which is the result of measuring the surface roughness by actually scanning the mirror-finished surface 9 to be measured, it can be seen that minute irregularities are captured by the irregularity signal Z _p . . However, the 8th
In the figure, t ₂ to _{t 3} is a period during which substantial measurements are performed by scanning (translational movement).

D Modification In addition, in the above example, PTZ24a to 24c
By performing control together with PZT24a, it is possible to both prevent the influence of vibration and improve accuracy.
Control by ~24c is not essential.

Also, when performing control using the PZTs 24a to 24c, the position and orientation of the reference plane 81 as the reflective surface of the reference plate 80 in the normal direction n in an apparatus using the Twyman Green interference mechanism as shown in FIG. are inserted between the reference plate support part 83 and the fixing surface 84.
It may be changed by c. However, in FIG. 10, light to the reference plane 81 is provided via a half mirror 87 and an objective lens 85. Also,
Light is applied to the surface to be measured 9 via a condenser lens 86 . Of course, both the reference surface and the measured surface may be supported by a variable support mechanism so that the positions and postures of both can be changed.

Furthermore, in the above embodiment, three neighboring regions R _a
Although the detection results for ~R _c are used, it does not limit the number of nearby regions to be detected;
There can be any number. However, in order to accurately know the spatial position and orientation of a certain area on the surface to be measured, it is preferable to set three or more neighboring areas. Furthermore, in order to reduce errors due to direction, it is preferable to set the neighboring regions so as to surround the center P at equal angles as in the above embodiment.

The present invention can be applied not only when scanning is performed by moving the surface to be measured, but also when moving the sensor head side. Furthermore, the present invention is applicable not only to devices that utilize optical interference, but also to surface roughness measuring devices that measure distance from the amount of defocus. (Effects of the Invention) As explained above, according to the present invention, since the surface roughness is measured based on the unevenness relationship based on the average height of the neighboring area, vibrations are generated in the surface to be measured and the measurement system. Even if the spacing between these changes with the addition of , highly accurate roughness measurement can be performed.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the schematic configuration of the embodiment, FIG. 2 is a diagram showing the internal configuration of the sensor head in the embodiment, FIGS. 3 to 5 are explanatory diagrams of the operation of the embodiment, and FIGS. FIG. 7 is a flowchart showing the operation of the embodiment, FIGS. 8 and 9 are graphs showing data examples, FIG. 10 is a diagram showing a modification of the present invention, and FIG. 11 is a diagram showing a conventional device. It is a layout diagram. 8...Object to be measured, 9...Surface to be measured, 24a-2
4c, 82a-82c...PZT, 31... Condensing lens, 32... Mireau type interference mechanism, 33, 81
... Reference plane, 35 ... Light source, 37 ... Light shielding plate, 3
8...Pinhole, 39...Photomultiplier tube, 40
a~40c...Photodetector, 61p, 61
a to 61c...interference signal processing circuits.

Claims (1)

[Scope of Claims] 1. A surface roughness measuring device that measures the surface roughness of a surface to be measured by scanning the surface, the device comprising: a sensor surface positioned at a predetermined location; a first distance sensor that optically detects the relative distance relationship between the center of the scanned area on the surface to be measured and the sensor surface; a second distance having another sensor surface positioned at a position in a positional relationship, and optically detecting an average relative distance relationship between a region near the center and the another sensor surface; a sensor; and a calculating means for calculating a relative unevenness relationship between the central portion and the neighboring area based on the detection results of the first and second distance sensors, A surface roughness measuring device characterized in that the surface roughness of the surface to be measured is determined based on a relationship between concavities and convexities. 2 The first and second distance sensors generate optical interference according to the optical path difference between the first and second reflected lights obtained when the measurement light is reflected by a predetermined reference surface and the measured surface, respectively. and an optical sensor that detects the interference light from the interference mechanism, and at least one of the surface to be measured and the reference surface is configured to adjust the posture and orientation of the surface by a predetermined driving means. It is supported by a variable support mechanism that can change the position in the normal direction, and based on the detection outputs of the first and second distance sensors, (a) the surface to be measured and the reference surface are (b) and the optical path difference between the first and second reflected lights is an odd multiple of 1/4 of the wavelength of the measurement light or a distance in the vicinity thereof; The surface roughness measuring device according to claim 1, further comprising position/attitude control means for applying a drive control signal to the drive means of the variable support mechanism.