JP3146591B2 - Reference surface shape measurement method and shape measurement system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an interferometer system used in an interferometer system for causing a reflected wavefront from an object to be measured to interfere with a reference wavefront caused by a reference surface and obtaining the shape of the first object to be measured based on the interference fringe pattern. The present invention relates to a method for measuring the shape of a reference plane.

[0002]

2. Description of the Related Art When an optical surface shape error is measured by an interferometer, a relative measurement with a reference surface is generally performed. The surface accuracy (referred to as PV) of the reference surface is PV = λ / 10
To λ / 20 (in the case of a He-Ne laser, λ = 633n
m), and more accurate measurement requires measurement of the absolute shape of the reference plane.

[0003] As a conventional measuring method of this kind, there is the following method.

One method is shown in FIGS. This shape measuring system for determining the surface shape of a collimating lens (Fizeo lens) M1, which is a reference surface, includes a laser light source 1, an image sensor 2, a semi-transparent mirror 3, and collimating lenses L and M1. And The reflected wavefront of the lens M2 and the reference wavefront of the collimating lens M1 are caused to interfere with each other, and the shape error of the collimating lens M1 is measured based on the interference fringe pattern. The measurement is performed according to the following procedure.

[0005] For simplicity of description, Front
The optics aberration is assumed to be zero.

(1) Measurement is performed in the state shown in FIG. At this time, the rotational positions of the collimating lenses L and M1 and the test lens M2 are set to 0 °. The shape error measured at this time is: F0 (shape error of collimating lens M1 at 0 °) + H0 (test lens M at 0 °)
2 shape error).

(2) Measurement is performed in the state shown in FIG. 2B (only the lens to be tested is rotated by 180 ° around the optical axis). At this time, the rotational positions of the collimating lenses L and M1 are 0 °. The shape error measured at this time is
F0 (shape error of collimating lens M1 at 0 °) + H180 (shape error of test lens M2 at 180 °). This data is calculated by 180 °
By rotating it, F180 (the shape error of the collimating lenses L and M1 at 180 °) + H0 (0
(Error of the test lens M2 in °).

(3) The measurement is performed in the state shown in FIG. 3 (the lens M2 is removed and the mirror-4 is placed at the focal position of the collimating lenses L and M1). At this time, the rotational positions of the collimating lenses L and M1 are 0 °. The shape error measured at this time is only the shape error of the collimating lens M1, and F0 (shape error of the collimating lens M1 at 0 °) + F180 (180
(Shape error of the collimating lens M1 in °).

(4) By calculating the difference between the data of (1) and (2), F0 (the shape error of the collimating lens M1 at 0 °) −F180 (the error of the collimating lens M1 at 180 °) Shape error). By adding this result and the data of (3), collimation
The shape error F0 of the focusing lens M1 is obtained.

The problem with this method is that the lens to be inspected can be precisely adjusted while maintaining the alignment of the entire system accurately.
That is, it must be turned 80 ° and a mirror must be installed.

Another method for determining the shape error of the collimating lens M1 is the "wavefront averaging method". This measurement method uses a test lens 42 having a bright F-number,
The wavefront measurement is performed a plurality of times by moving the area to be measured on the test lens 42 without moving the collimating lens 41. By averaging the measurement data, it is possible to reduce the influence of the shape error of the test lens 42 on the measurement and obtain the surface shape of the collimating lens 41.
FIG. 4 is a diagram illustrating the principle of this measurement method. Equation 1 is obtained by averaging the measurement data WK.

[0012]

(Equation 1)

When the number of measurements n is increased while changing the surface shape WTK of the lens 42 to be measured,
Since the term approaches zero, the surface shape WR of the collimating lens 41 is determined.

This method uses the NA of the collimating 41.
Prepare a lens 42 having a larger NA, and use the lens 42
Is shifted laterally with respect to the optical axis, so that the wavefront is averaged and the shape error of the reference spherical surface of the collimating lens 41 is obtained.

[0015]

In the above-mentioned prior art, there is a problem that the shape of the object to be measured is limited.

Further, in other conventional techniques, non-uniformity of the lens to be inspected is necessarily required.

An object of the present invention is to use a wavefront averaging for generating a constant pattern and extract the pattern (wavefront generation extraction method), thereby alleviating the restriction on the averaging of the lens to be measured. Is to provide the law.

According to the present invention, there is provided a wavefront reflected from an object to be measured ,
To interfere with the reference wavefront by the reference plane, by the interference fringe pattern, a shape measuring method of the reference plane used in the interferometer system for determining a shape error of the measurement object, before
Serial measurement first an alternative to the object (hereinafter, referred to as a first dummy) of obtaining the first rotation mean shape about the first rotation axis, the first dummy or second
Of one of the dummy, the determination of the second rotational average shape for the second rotation axis, said first rotation mean shape before
More and serial second rotation mean shape, the rotationally symmetric element, the first rotation averaged shape or be determined Te with <br/> the second rotation averaged shape, the first rotating average shape and the second
A is either one of the rotational average shape of the rotary
The one used to determine the symmetric component and the rotational symmetry component
Than the partial, this consisting of determining the shape error of the reference plane
And features.

[0019]

According to the present invention, the true rotationally symmetric component (RS) and the non-rotationally symmetric component (A) which define the shape error of the first dummy are defined below.
S) into a true rotationally symmetric component (R
After obtaining S), the shape error of the reference plane is obtained. True rotationally symmetric component (RS), non-rotationally symmetric component (A
The definition of S) will be described with reference to the schematic diagram of FIG. In FIG. 5, H indicates a distribution diagram of the shape error of the first dummy. While rotating the first dummy, measured values at respective points on the first dummy are obtained, and an average of the measured values is a rotationally symmetric component indicated by RS. The non-rotationally symmetric component denoted by AS is H-
Defined by RS.

The procedure for obtaining the shape error of the reference plane is as follows.

(1) Obtain an interference fringe pattern OR. This is the sum of the reference plane shape error F0, the rotationally symmetric component RS, and the non-rotationally symmetric component AS.

(2) The first rotationally symmetric component is obtained by averaging the surface shape data obtained by measurement while rotating. This is because the shape error F0 of the reference plane and the rotationally symmetric component (R
S1).

(3) Displace the first rotation axis. While either one of the first dummy and the second dummy is rotated at least one rotation around the first rotation axis after the displacement, the surface shape data obtained by the measurement is averaged for the second dummy. Find the rotationally symmetric component. This is the sum of the reference plane shape error F0 and the rotationally symmetric component (referred to as RS2).

(4) From the first rotationally symmetric component and the second rotationally symmetric component, a rotationally symmetric component that does not include a shape error of the reference plane (this is called a true rotationally symmetric component, and this is the RS1 or RS2). Is determined for the first rotationally symmetric component or the second rotationally symmetric component.

(5) At least one of the first rotationally symmetric component and the second rotationally symmetric component obtained in (2) and (3) (when the second dummy is used, the first rotationally symmetric component is used). From the component or the second rotationally symmetric component used to determine the true rotationally symmetric component) and the true rotationally symmetric component, the shape error of the reference plane is determined.

[0026]

FIG. 1 shows a Fizeau-type interferometer system which is a shape measuring system according to the present invention. In the present embodiment, a substantially true sphere is prepared, and the accuracy of the rotation axis is ensured so that fringe-scanning fringes are formed. This Fizeau-type interferometer system rotates a control unit 8, a laser light source 1, an image pickup device 2, a semi-transparent mirror 3, collimating lenses (Fizeo lenses) L and M1, and a damper 10. First
, And an α stage 7 as a moving means. The control unit 8 controls the image pickup device 2 and the laser light source 1,
And exchange data with the processing unit 15. The processing unit 15 includes a first rotationally symmetric component detecting unit 81, a second rotationally symmetric component detecting unit 82, a true rotationally symmetric component detecting unit 83, and a shape error detecting unit 84, as shown in FIG. Have. The processing unit 15 has a memory (not shown) for storing programs and data, a CPU (not shown). The CPU and the memory include a first rotationally symmetric component detecting unit 81, a second rotationally symmetric component detecting unit 82, a true rotationally symmetric component detecting unit 83, and a shape error detecting unit 84 shown in FIG. Perform a function.

The first rotating means 20 rotates the dummy 10 about a rotation axis perpendicular to the optical axis. The α stage 7 serving as a moving unit is configured to move the first rotating unit 20 around the measuring optical axis by 90 degrees.
Rotate °. FIG. 11 shows the details of the first rotating means 20 and the α stage 7. In FIG. 11, reference numeral 1100 denotes a through hole provided in a cylindrical shape so as to pass eccentrically through the center of the dummy 10 in order to prevent the wavefront incident on the dummy 10 from being reflected on the back surface. In order to prevent back surface reflection, a method may be used in which a true sphere is made of colored glass that absorbs a measurement wavelength, in addition to the method of providing a through hole.

Then, the wavefront reflected by Dami-10 interferes with the reference wavefront of the collimating lens M1 to be measured, and the shape error of the collimating lenses L and M1 is measured by the interference fringe pattern. It is.

In the above configuration, the light emitted from the laser 1 enters the semi-transparent mirror 3. A part of this light is reflected by the reference surface 16 of the Fizeau lenses L and M1, returns to the semi-transparent mirror 3, and goes straight upward. Thus, the light enters the image sensor 2.

On the other hand, a part of the light incident on the Fizeo lenses L and M1 is converted into an appropriate spherical wave by the Fizeo lenses L and M1 and is incident on the dummy 10. Then, the light is reflected here, again passes through the Fizeau lenses L and M1, returns to the semi-transparent mirror 3, is bent upward, and enters the image sensor 2.

At this time, if the position of the dummy 10 is adjusted and the shape of the dummy 10 and the shape of the spherical wave generated by the Fizeo lenses L and M1 substantially match each other, there is not enough space on the image sensor 2. An interference fringe of roughness is observed. The observed interference fringe gives information on the deviation between the shape of the Dami-10 and the spherical wave, that is, information on the wavefront aberration, and one fringe is exactly equal to the deviation of half the wavelength λ of the light from the laser light source 1. ing.

[0032] Therefore, when the shape of the dummy -10 is close to spherical, Fizeau by analyzing the interference fringe pattern becomes what is suitable roughness of the interference fringes throughout - lens M
1 can be collectively measured.

The measurement is performed according to the following procedure. FIG.
(A case where α = 90 ° is easy to understand will be described.)
FIG. 6 shows the positions on the dummy and the Fizeo lens M1 for a part of the data obtained as a result of the measurement.

(1) The optical axes of the collimating lenses L and M1 coincide with the central axis of the dummy 10. In this state,
While the dummy 10 is rotated once by the first rotating shaft 20, the surface shape data at positions 71 and 72, which are parts thereof, shown in FIG. The first rotationally symmetric component detecting means 81 obtains a first rotationally symmetric component by averaging the surface shape data obtained by measurement while rotating the bearing 14. This is the collimating lens M
1 is the sum of the shape error F0 and the rotationally symmetric component RS1.

(2) The first rotation axis 20 is rotated by 90 ° around the axis 73 in FIG. 7 by the α stage 7 for moving the optical axis.
Rotate and move. The second rotationally symmetric component detecting means 82 takes the surface shape data at the position 73 shown in FIG. 7 while rotating the dummy 10 at least once around the first rotation axis 20 after the movement. The second rotationally symmetric component is obtained by averaging the surface shape data obtained by the measurement. This is the sum of the shape error F0 of the collimating lens M1 and the rotationally symmetric component RS2. (1) First rotating shaft 20
Is perpendicular to the axial direction of the first rotating shaft 20 in (2), and the obtained patterns are perpendicular to each other.

(3) The true rotationally symmetric component detecting means 83
From the first rotationally symmetric component and the second rotationally symmetric component, a rotationally symmetric component that does not include a shape error of the collimating lens M1 (this is called a true rotationally symmetric component, which is RS1 or RS2). , The first rotationally symmetric component. This will be described with reference to FIG. Since the rotationally symmetric component on each ring shown in FIG. 7 has the same value,
, RS1A + F0, RS1B + F0 (first rotationally symmetric component) and RS2A + F0, RS2B + F0 (second rotationally symmetric component). The first rotationally symmetric component and the second
When the difference between the rotationally symmetric components is taken, the shape error of the collimating lens M1 disappears, and the values of the intersections A, B, C, and D become as shown in FIG. With reference to the value of the first rotationally symmetric component at the center O, the value on the arc OE is a constant RS
Since the common portion of 1A is included, all the values of the second rotationally symmetric component are determined.

Using the fact that the values of the second rotationally symmetric components are the same, the point B and the point C can be obtained by using the values on the arcs OB and CF instead of the arc OE. The value of the rotationally symmetric component stops even if the arc EF 'is excluded.

Further, as another method for obtaining the unknown value, RS
1A, RS1B, RS2A, RS2B, and RS3B, RS2A, R based on RS1A, utilizing the fact that there are substantially 3 equations for A, B, C, D values.
The value of S2B can also be determined.

Thus, the rotationally symmetric component of the shape error is determined. This operation may be performed for the entire test lens 10.

(4) The shape error detecting means 84 calculates the shape error of the collimating lens M1 from the first rotationally symmetric component obtained in (1) and the true rotationally symmetric component.

In FIG. 7, the reason why the first rotation axis is 90 ° with respect to the measurement optical axis is as follows. Rotation by the first rotation axis is performed at an angle (β
°), as shown in A view, the 91 after β rotation, which corresponds to 72 in the case of 90 °, has a curvature, which complicates the analysis.

Therefore, as shown in FIG. 7, the first rotation axis is rotated by α ° while being perpendicular to the measurement optical axis, and after extracting the pattern on the center line of the A view, the correction by α ° is sufficient. Will be.

Of course, the first rotation axis is 90
It is possible to measure even if it is not °, but if the pattern is kept in a straight line when averaging, analysis becomes easier.

Also, in the case of FIGS.
Since measurement is performed at only three points (0 °, 180 °, mirror installation), alignment requirements are severe, but in the case of the present invention, since 360 ° is divided into a large number, alignment requirements at each position are required. Has the effect that the final accuracy is improved over the prior art.

In FIG. 1, after the dummy is rotated about the first rotation axis, the first rotation axis is not rotated about the measurement optical axis, but the α stage (the second rotation axis is used). ), The dummy may be rotated one or more times, and the rotationally symmetric component obtained by this may be used as the second rotationally symmetric component.

FIG. 10 shows an embodiment in which the dummy is swung without being rotated. The swinging method is suitable for a case where the rotation must be limited. The case where the rotation must be finite is, for example, a case where a concave mirror is used as a dummy as shown in FIG. 10, and when rotated around the first rotation axis as shown in FIG. In this case, no interference occurs. The third rotating shaft corresponds to the first rotating shaft 20 in FIG. 1, and the rotating shaft 106 corresponds to the rotating shaft 73 in FIG. 7.

FIG. 10 shows only the Fizeau lens 100 and the dummy 101. Reference numeral 102 denotes a state when the dummy 101 is swung. The interference optical system is the same as in FIG. The procedure of the measurement is as follows.

(1) While swinging the dummy to the left and right around the third rotation axis 105 (by positioning the arc PS and then the arc QR at a position corresponding to the arc KL), The first swing component is determined by averaging the surface shape data obtained by the measurement. Points S and Q are arc PS = arc QR =
This is a point defined by the arc KL.

(2) As shown in FIGS. 10C and 10D, the center 107 of the cat's eye of the collimating lens
Is provided with a mirror 104, the direction of the measurement optical axis is changed to a direction perpendicular to the swing direction, and the dummy is moved to a position where the moved measurement light is reflected.

(3) At this time, while maintaining the positional relationship between the third rotation axis and the dummy 101, the third rotation axis and the dummy 101 are rotated by 180 ° around the rotation axis 106 and moved. . The surface shape data obtained by measurement while oscillating around the third rotation axis in the same manner as in (1) is averaged, and further added to the first oscillating component to obtain a third oscillating component. Obtain the swing component.

(4) After returning the third rotating shaft and the dummy 101 to the state of (1), the third rotating shaft is changed to the rotating shaft 1
Rotate 90 ° around 06 and move.

(5) The surface shape data obtained by measuring while oscillating around the third rotation axis after moving the dummy 101 in the same manner as in (1) is averaged to obtain the second data. Is obtained.

(6) Similarly to (2), a mirror 104 is provided at the center 107 of the cat's eye of the collimating lens, and the direction of the measurement optical axis is changed to the direction perpendicular to the swing direction.
The dummy is moved to a position where the moved measuring light is reflected.

(7) At this time, while maintaining the positional relationship between the third rotation axis and the dummy 101, the third rotation axis and the dummy 101 are rotated by 180 ° around the rotation axis 106 and moved. . The surface shape data obtained by measuring while oscillating around the third rotation axis after moving the damper 101 in the same manner as (1) is averaged, and further, the second oscillating component and Addition is performed to obtain a fourth swing component.

(8) From the third oscillating component and the fourth oscillating component, an oscillating component that does not include a shape error of the collimating lens (this is referred to as a true oscillating component) is converted to a third oscillating component. The swing component and the fourth swing component are obtained.

(9) The shape error of the collimating lens is determined from the third swing component and the true swing component.

Also in this embodiment, the rotation around the rotation axis 106 can be other than 90 °.

As shown in FIGS. 10B, 10C, and 10D, the positional relationship between the dummy and the third rotation axis before and after changing the direction of the measurement optical axis is as follows. And 1
It is necessary to be in a state of being rotated by 80 °. That is, in the state (b) → (c), (d), the dummy 101 becomes 180
It is necessary to rotate the data and superimpose the data. At this time,
At the point K, the superimposition of both data is performed by the operation of (1) (referred to as the 0 ° direction), the PQ, and the operation of (3) at (18).
RS, so sum = PR + SQ (the SQ part overlaps) and has a centrally symmetric shape,
For any point between KL, the scan area is constant. That is, the same thing as the rotationally symmetric component is obtained.

However, at this time, if a simple rotation of 180 ° around the rotation axis 106 in the state of (b), the corresponding point on the dummy 101 moves to the line symmetric position, so that (c) and (d) )
The problem is solved by using cat's eye reflection.

As described above, an even pattern can be created with this arrangement.

In the above description, the rotation of the θ stage is performed by dividing 360 ° in accordance with the roughness of the shape error of the Fizeo lens. However, when the shape error becomes large, the division must be finely divided. Must. For this reason, when the shape error is large, the division is performed, the measurement is not repeated, and the data is stored in the CCD during one rotation of the θ stage.
Data may be accumulated. This is to integrate the data with time, and then an average may be taken.

Although the present embodiment has been described with reference to a Fizeau-type interferometer system, the present invention is not limited to this, and an interferometer called an amplitude division type, ie, a Michelson interferometer and a Twyman-Green interferometer. The same can be applied to an interferometer.

[0063]

As described above, according to the present invention, it is possible to provide a method of measuring the shape of a reference plane which does not require non-uniformity of data to be averaged.

[Brief description of the drawings]

FIG. 1 is a block diagram of a Fizeau-type interferometer system according to the present invention.

FIG. 2 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the related art.

FIG. 3 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the related art.

FIG. 4 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the related art.

FIG. 5 is a principle diagram of a method for measuring the sphericity of a collimating lens according to the present invention.

FIG. 6 is a block diagram of a processing unit according to the present invention.

FIG. 7 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the present invention.

FIG. 8 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the present invention.

FIG. 9 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to the present invention.

FIG. 10 is an explanatory diagram of a method for measuring the sphericity of a collimating lens according to a second embodiment of the present invention.

FIG. 11 is a detailed view of the first rotating means 20 and the α stage 7 according to the present invention.

[Explanation of symbols]

Reference numeral 8: control unit, 1: laser light source, 2: imaging element, 3: semi-transparent mirror, L, M1: collimating lens (Fizeo lens), 10: dummy, 7: α stage, 81: First rotationally symmetric component detecting means, 82... Second rotationally symmetric component detecting means, 83... True rotationally symmetric component detecting means, 84... Shape error detecting means.

Claims (9)

(57) [Claims] An interference between a wavefront reflected from an object to be measured and a reference wavefront caused by a reference surface is obtained.
A shape measuring method of the reference plane used in the interferometer system for determining a shape error of the measurement object, first made in place of the measurement object (hereinafter,
Determining a first rotation average shape of the first rotation axis with respect to a first rotation axis; a second rotation of either the first dummy or the second dummy with respect to a second rotation axis. Obtaining an average shape; obtaining a rotationally symmetric component for the first rotation average shape or the second rotation average shape from the first rotation average shape and the second rotation average shape; 1 rotation averaged shape and of the second rotation averaged shape comprising either one obtains the rotational symmetry component
Determining a shape error of the reference plane from the one used for the calculation and the rotationally symmetric component. 2. The shape measuring method according to claim 1, wherein the measuring object has a spherical surface to be measured, and the first and second rotation axes pass through a center of curvature of the spherical surface. Characteristic shape measurement method. 3. The shape measurement method according to claim 1, wherein at least one of the first and second rotation axes is perpendicular to a measurement optical axis. 4. The shape measuring method according to claim 1, wherein, when obtaining a rotation average shape about the rotation axis, the measurement object is rotated at least once around the rotation axis. And measuring the surface shape data corresponding to a plurality of rotation positions, and averaging the surface shape data measured corresponding to the plurality of rotation positions to obtain the rotation average shape. 5. The shape measuring method according to claim 1, wherein, when obtaining a rotation average shape with respect to the rotation axis, the measurement object is rotated at least once around the rotation axis. A time integration of the surface shape data, and averaging the time-integrated surface shape data to obtain the rotational average shape. 6. The shape measuring method according to any one of claims 1 to 5, wherein, when obtaining the rotationally symmetric component, a property of a concentric contour line of the rotationally symmetric component is obtained by:
An area in which the first rotational average shape is obtained;
A shape measurement method characterized by using the rotation average shape in a region where the rotation average shape is obtained. 7. The shape measuring method according to claim 6, wherein the rotational symmetry component is obtained by using all data of the contour lines in the overlapping region. 8. The interferometer system according to claim 1, wherein the reflected wavefront from the first measurement object interferes with a reference wavefront by the reference surface, and the interference fringe pattern determines a shape error of the first measurement object. a shape measuring system for measuring a shape error of the reference plane, the first first an alternative to the measurement object (hereinafter, referred to as a first dummy) of the first rotary shaft Means for determining a first rotational average shape with respect to a second rotational axis of either the first dummy or the second dummy; and means for determining the first rotational average with respect to a second rotational axis. Means for obtaining a rotationally symmetric component from the shape and the second rotational average shape for the first rotational average shape or the second rotational average shape; and the first rotational average shape and the second rotation. There among the average shape The rotationally symmetric component
And a means for determining a shape error of the reference plane from the rotationally symmetric component. 9. A wavefront reflected from an object to be measured and a reference wavefront caused by a reference plane interfere with each other.
A memory storing a program for obtaining a shape error of the reference plane by an information processing device in an interferometer system for obtaining a shape error of the measurement object, wherein a first memory is used instead of the first measurement object. those (hereinafter, first referred to as dummy) of a first determining a first rotation mean shape about the rotational axis processing, said first either dummy or second dummy, the second A process of obtaining a second rotation average shape with respect to a rotation axis; and a first rotation average shape or the second rotation, based on the first rotation average shape and the second rotation average shape. a process for obtaining the average shape, there is either one of the first rotation mean shape and the second rotation averaged shape, obtains the rotational symmetry component
And a process for obtaining a shape error of the reference plane from the rotationally symmetric component. .