JP2008111891A - Optical element holding device and exposure apparatus provided with the same - Google Patents

Abstract

An optical element holding device capable of exhibiting high positioning accuracy even when a power supply is restarted, and an exposure apparatus including the same.
An optical element holding device having a driving mechanism 71 for driving a lens LE in the optical axis direction and a lens position detecting means 72 for measuring the position of the lens LE. The lens position detecting means 72 comprises: It has an incremental vertical displacement detection sensor 722 that measures the relative position of the lens LE in the optical axis direction, and an absolute vertical displacement detection sensor 721 that measures the absolute position of the lens LE.
[Selection] Figure 3

Description

  The present invention relates to an optical element holding device having a driving unit that drives an optical element such as a lens by a piezo actuator and the like, and a position measuring unit that measures the position of the optical element, and an exposure apparatus including the same.

  More specifically, the present invention relates to an optical element holding device capable of exhibiting high positioning accuracy even when the power supply is restarted, and an exposure apparatus including the same.

  Conventionally, as a holding device for an optical element having a drive unit that drives an optical element and a position measurement unit that measures the position of the optical element, a technique using two types of position adjustment devices in the position measurement unit has been disclosed. (For example, refer to Patent Document 1).

  In the holding device described in FIG. 13 of Patent Document 1, two-degree-of-freedom position adjustment mechanisms capable of driving an object in the central axis direction and the tangential direction are provided at three locations along the circumferential direction. Yes. In the XYZ orthogonal coordinate system where the optical axis of the optical element is the Z-axis and the direction perpendicular to the Z-axis is the XY plane, the degree of freedom of driving the optical element is 6. The optical element can perform X-axis direction translation, Y-axis direction translation, Z-axis direction translation, rotation around the X axis (tilt), rotation around the Y axis (tilt), and rotation around the Z axis (rotation).

  In addition, the holding device is provided with three position measuring units that measure the position of the optical element along the circumferential direction. The position measuring unit uses a biaxial capacitive sensor that detects the displacement between the optical axis direction of the optical element and the tangential direction of the outer peripheral portion.

  On the other hand, in the holding device described in FIG. 2 of Patent Document 1, position adjustment mechanisms that drive the optical elements in the optical axis direction using piezoelectric actuators are provided at three locations along the circumferential direction. With the position adjusting mechanisms provided at these three locations, the optical element can be translated in the Z-axis direction, rotated around the X axis (tilt), and rotated around the Y axis (tilt).

Moreover, the position measurement part of an optical element is provided in three places along the circumferential direction. This position measuring unit is an optical linear encoder that detects the displacement of the optical element in the optical axis direction.
JP 2003-337272 A

  However, in the holding device disclosed in FIG. 13 of Patent Document 1 described above, a capacitance sensor is used in the position measurement unit, so that the detection resolution, the dynamic range, the linearity of the detection output signal, and the like are interference types. There was a problem that it was inferior to the displacement meter. The capacitance sensor can detect an absolute position.

  Further, since the holding device disclosed in FIG. 2 of Patent Document 1 described above uses an optical linear encoder that detects the displacement of the optical element in the optical axis direction, only the relative movement amount of the subject is detected. There was a problem that I could not. If only the relative movement amount can be detected, the absolute position of the optical element cannot be recognized when the apparatus is restarted after the power is turned off, and the optical element can be returned to the coordinate value before the power is turned off. Can not.

  In Patent Document 1, the holding device in FIG. 2 and the holding device in FIG. 13 hold optical elements separately and independently.

  Accordingly, an object of the present invention is to provide an optical element holding device and an exposure apparatus including the same that can exhibit high positioning accuracy even when the power supply is restarted.

  In order to achieve the above object, an optical element holding device according to one aspect of the present invention includes a drive unit that drives an optical element in an optical axis direction and an inclination direction with respect to the optical axis, and a position that measures the position of the optical element. An optical element holding device having a measurement unit, wherein the position measurement unit measures a position of the optical element in an optical axis direction and a tilt direction with respect to the optical axis. The first measuring unit measures the relative position of the optical element, and the second measuring unit measures the absolute position of the optical element.

  An exposure apparatus provided with such an optical element holding device also constitutes one aspect of the present invention.

  Further objects and other features of the present invention will become apparent from the preferred embodiments described below with reference to the accompanying drawings.

  According to the present invention, it is possible to measure both the absolute position and the relative position of an optical element, and to provide an optical element holding device capable of exhibiting high-precision positioning accuracy even when the power supply is restarted. An exposure apparatus can be provided.

  Hereinafter, an aspect of the present invention will be described with reference to the accompanying drawings. In addition, in each figure, the same code | symbol is attached | subjected about the same component and the overlapping description is abbreviate | omitted.

  FIG. 1 shows an exposure apparatus provided with a first embodiment of the optical element holding apparatus of the present invention. The exposure apparatus 1 is a so-called scanner type exposure apparatus that scans a reticle R under slit illumination and performs an exposure operation while scanning the wafer WF in synchronization with the reticle R.

  In this exposure apparatus 1, an XYZ orthogonal coordinate system is defined as shown on the upper side in the drawing. That is, the optical axis direction of the projection optical system 2 is defined as the Z axis, the scanning direction of the reticle R and the wafer WF is defined as the X axis, the direction orthogonal to the X axis in the apparatus plane is defined as the Y axis, and the apparatus vertical direction is defined as the Z axis.

  The exposure apparatus base 3 has a bottomed cylindrical main body 31 and a lens barrel mount 32 for accommodating and fixing a part of the projection optical system 2 in the main body 31. A damper means 33 is provided between the main body 31 and the lens barrel mount 32 in order to prevent vibrations on the exposure apparatus installation surface from being transmitted to the projection optical system 2.

  A wafer stage WF1 for adjusting the position of the wafer WF in the optical axis direction, that is, focus adjustment, is provided at the bottom of the main body 31. The wafer stage WF1 moves the wafer WF in the X-axis direction in synchronization with the scanning operation of the reticle stage R1. Wafer stage WF1 can also be adjusted in position in the Y-axis direction in order to improve the overlay accuracy of the reticle image. The wafer stage WF1 is provided with a spherical mirror WF2 used when measuring the optical aberration of the projection optical system 2. Note that a photosensitive agent is applied to the wafer WF.

  A reticle stage R1 on which a reticle R serving as an original plate of a semiconductor circuit pattern is placed is placed on a reticle mount R2 that covers the upper side of the projection optical system 2. The reticle mount R2 is for fixing the reticle stage R1 to the lens barrel mount 32.

  On the lens barrel mount 32, the illumination unit 4 is provided. The illumination unit 4 includes a laser light source unit 41, an exposure illumination means 42 used when illuminating the reticle R, an aberration measurement illumination means 43 used during an aberration measurement process, an optical path switching mirror 44, It has. On the lens barrel mount 32, a main body control means 8 for controlling the operation of the entire exposure apparatus is provided adjacent to the laser light source unit 41.

  Aberration measuring means 5 for measuring the optical aberration of the projection optical system 2 is provided on the reticle mount R2. The aberration measuring means 5 includes an interferometer 51, a half mirror 52, a reflecting mirror 53, and a collimator lens 54. Among these, the reflection mirror 53 and the collimator lens 54 are mounted on the reticle stage R1 so as to be able to advance and retreat with respect to the light beam incident portion of the projection optical system 2.

  A lens control unit 6 is provided on the lens barrel mount 32, and the lens control unit 6 controls various optical elements of the projection optical system 2 according to a predetermined control flow. This control enables fine adjustment of the optical system such as optimization of the optical aberration of the lens. The fine adjustment of the optical system is performed, for example, when the projection optical system 2 is assembled or when the exposure apparatus 1 is delivered to a delivery destination.

  The projection optical system 2 is configured by stacking a plurality of lens units 7 that are optical element holding devices, and the lens unit 7 includes a drive unit that drives the lens, and a position measurement unit that measures the position of the lens. ,have.

  FIG. 2 shows the structure of the lens unit 7 in a plan view. The drive unit for driving the lens is provided in three places as the drive mechanism 71 in an annular fixed lens barrel 73 at intervals of 120 degrees around the central axis. On the other hand, a position measurement unit that measures the position of the lens is provided between the drive mechanisms 71 along the circumferential direction as the lens position detection means 72. The lens position detection means 72 detects the optical axis direction (Z-axis direction) displacement of the lens frame 74 and the radial direction displacement orthogonal to the optical axis.

  The lens position detection means 72 includes an absolute vertical displacement detection sensor 721 that is a second measurement unit, an incremental vertical displacement detection sensor 722 that is a first measurement unit, and an incremental lateral displacement detection sensor 723 that is a third measurement unit. It is configured.

  The absolute vertical displacement detection sensor 721 is a device that measures the absolute position of a lens that is an optical element, and projects a detection laser beam in the optical axis direction (Z-axis direction) of the lens. The incremental type vertical displacement detection sensor 722 is a device that measures the relative position in the optical axis direction of the lens or in the tilt direction with respect to the optical axis. Similar to the absolute type vertical displacement detection sensor 721, the optical axis direction (Z-axis direction) of the lens. A detection laser beam is projected onto the screen. The incremental lateral displacement detection sensor 723 is a device that measures the relative position of the lens in a direction orthogonal to the optical axis direction, and projects a detection laser beam in the radial direction of the lens.

  The coordinate system of the lens unit 7 is the same as the XYZ orthogonal coordinate system of the exposure apparatus 1 described with reference to FIG. That is, the same direction as the optical axis of the lens is the Z axis, and the radial direction of the lens is the XY plane.

  FIG. 3 shows a form in which a lens frame 74 is attached to the lens unit 7 of FIG. FIG. 4 shows a cross section taken along line IV-IV in FIG. As shown in these drawings, the fixed barrel 73 of the lens unit 7 has a cross-sectional shape in which an inverted L-shaped standing portion is formed on the upper surface of a flat plate. The lens LE is held by a lens frame 74, and the lens frame 74 is fixed to the drive mechanism 71 by screws, bolts, or the like.

  More specifically, a flange portion 74a is formed on the upper surface of the lens frame 74 that houses the lens LE, and the flange portion 74a is fastened to the displacement output portions of the three sets of drive mechanisms 71. A lens frame displacement detection target mirror 75 is attached to the flange portion 74a.

  When the target mirror 75 is attached in this way, the detection laser beam projected from the absolute vertical displacement detection sensor 721 or the incremental vertical displacement detection sensor 722 is reflected by the target mirror 75. The amount of displacement of the target mirror 75 can be detected from this reflected light information.

  In the above configuration, when the three sets of drive mechanisms 71 are driven by an equal amount, the lens LE can be driven to translate in the optical axis C direction (Z-axis direction). Further, by providing a predetermined difference in the driving amount of the driving mechanism 71 provided at three places, the lens LE can be tilted in the θX and θY directions shown in FIG.

  At this time, when predetermined coordinate transformation is performed on the output signals of the absolute vertical displacement detection sensor 721 and the incremental vertical displacement detection sensor 722 and the result is fed back, the translation amount and tilt drive amount of the lens LE are closed loop. Be controlled.

  Further, from the output signal of the incremental lateral displacement detection sensor 723, the shift amount of the reticle image accompanying the parallel decentering of the lens LE in the X axis direction and the Y axis direction can be calculated. Then, by adding the influence of this calculation result to the driving amount of the wafer stage WF1 shown in FIG. 2, the alignment error of the reticle image due to the eccentricity of the lens LE can be eliminated.

  5 to 8 show the detailed structure of the drive mechanism 71. 5 is a plan view of the drive mechanism 71, FIG. 6 is a side view of the drive mechanism 71, and FIG. 7 is a cross-sectional view taken along line VII-VII in FIG. FIG. 8 is a perspective view showing a displacement takeout member and a direction changing member, which will be described later. In FIG. 8, illustration of various screws is omitted.

  And the local coordinate system UVW of the drive mechanism 71 is prescribed | regulated as follows. That is, the U axis is an axis along the radial direction of the lens LE, the V axis is an axis along the radial direction of the lens LE and is an axis perpendicular to the U axis, and the W axis is an axis parallel to the optical axis C of the lens LE. It is.

  As shown in these drawings, the drive mechanism 71 includes a displacement take-out member 711, a direction changing member 712, a laminated piezoelectric actuator 713, and a fastening member such as a screw connecting these members, and a fixed barrel 73. (Described in FIG. 6).

  As the piezo actuator 713, a rod body in which disk-shaped piezoelectric elements and electrodes are alternately stacked is housed in a sealed cylindrical container, and one end of the rod body is projected outside the container. Then, the elongation of the entire length in the U-axis direction increases substantially in proportion to the applied voltage. Further, the piezo actuator 713 is connected to the displacement extraction member 711 using a piezo adjustment screw 714.

  Then, the displacement of the piezo actuator 713 in the U-axis direction is transmitted to the direction changing member 712 via the displacement extracting member 711, and the lens frame 74 is moved up and down along the W-axis direction. The dimensional error of the piezo actuator 713 can be corrected by adjusting the screwing depth of the piezo adjusting screw 714 with respect to the displacement extracting member 711.

  The displacement extraction member 711 is manufactured by performing wire electric discharge machining on a plate-shaped metal block having a predetermined thickness as a base material. And a screw hole is formed in a predetermined location using a drilling machine.

  The direction changing member 712 is manufactured by performing wire electric discharge machining on a plate-shaped metal block having a predetermined thickness as a base material, similarly to the manufacturing of the displacement extracting member 711. At this time, the direction change link 712a and the support link 712b are formed. And a screw hole is formed in a predetermined location using a drilling machine.

  Next, the assembly procedure of the drive mechanism 71 will be described. First, in the perspective view of FIG. 8, the both side arms of the displacement extraction member 711 are inserted into the left and right vacant spaces of the direction changing member 712 and fastened using screws. Next, as shown in FIG. 5, the piezo actuator 713 is mounted in the space surrounded by the displacement extraction member 711, and the piezo receiving link 713 a that is the output end of the piezo actuator 713 is pushed leftward using the piezo adjustment screw 714. . Then, the displacement extraction member 711 is brought into pressure contact with the piezo receiving link 713a. This completes the mounting of the piezo actuator 713. Finally, the drive mechanism 71 is installed on the annular fixed lens barrel 73 to complete the assembly.

  9 and 10 show a link mechanism of the displacement extraction member 711 and the direction changing member 712. FIG. FIG. 9 schematically shows the displacement extracting member 711, and FIG. 10 schematically shows the direction changing member 712. Here, the operation principle of the drive mechanism 71 will be described with reference to FIGS. 9 to 10 with reference to FIGS.

  When a predetermined voltage is applied to the two electrode terminals of the piezo actuator 713, the total length L of the piezo actuator 713 extends by dL in the U-axis direction. Then, as shown in FIG. 9, one piezo receiving link 711b1 is displaced by dU1 = −dL / 2, and the other piezo receiving link 711b2 is displaced by dU2 = dL / 2. Then, the displacement extraction link 711a configured to be rotatable about the elastic hinges H11 and H21 rotates slightly around the W axis. As a result, the connecting link 711c1 is displaced by dU3 = − (3/2) dL, and the connecting link 711c2 is displaced by dU4 = (3/2) dL.

  The displacement of the connecting links 711c1 and 711c2 in the U-axis direction also displaces the horizontal links 712c1 and 712c2 of the direction changing member 712 as shown in FIG. Then, the direction conversion link 712a arranged so as to have an orientation of about 45 degrees with respect to the U axis rotates slightly, and the lens frame drive link 712d is raised by dW in the W axis direction.

  As described above, the lens frame drive link 712d is slightly displaced in the W-axis direction as the piezo actuator 713 is extended. The lens frame drive link 712d is displaced only in the W-axis direction, and the U-axis direction and the V-axis direction. There is no displacement. This is because support links 712e are connected to the left and right sides of the lens frame drive link 712d.

  The support link 712e restrains the displacement of the lens frame drive link 712d in the U-axis direction, and the support link 712f restrains the displacement of the lens frame drive link 712d in the V-axis direction. The support link 712f is disposed at the end of the horizontal links 712c1 and 712c2 on the center side of the direction changing member. And the displacement of the horizontal links 712c1 and 712c2 in the U-axis direction is not restricted, and the displacement in the V-axis direction is restricted. Therefore, the motion of the horizontal links 712c1 and 712c2 that do not have a V-axis direction component is transmitted to the lens frame drive link 712d via the direction conversion link 712a.

  With the above configuration, the region near the screw hole 712g (described in FIG. 5) on the lens frame drive link 712d is displaced only in the W-axis direction, and the displacement in the U-axis direction and the V-axis direction is restricted. Therefore, the lens frame 74 can be accurately displaced in the W-axis direction. Moreover, the displacement magnification rate of the piezoelectric actuator 713 can be suitably set by appropriately selecting the intervals between the elastic hinges H11, H12, H13, H21, H22, and H23.

  On the other hand, as shown in FIG. 10, the lens frame drive link 712d is supported by the support links 712e on both sides, but the link 712d is partially thin (see FIG. 6). 712d1 can be elastically deformed in the twist direction around the U axis. Further, the direction changing link 712a that restrains the lower side of the lens frame driving link 712d also has a slight degree of freedom of elastic deformation in the V-axis direction.

  Therefore, as shown in FIG. 7, the lens frame driving link 712d can be slightly twisted around the U axis centered on the point P at the center of the cross section of the support link 712e. That is, the lens frame drive link 712d is capable of translational movement in the W-axis direction and twisting movement around the U-axis. This twisting motion causes the lens frame 74 to tilt.

  11 to 13 show the detailed structures of various sensors provided in the lens position detecting means 72 shown in FIG. FIG. 11 shows an absolute vertical displacement detection sensor 721. As the absolute type vertical displacement detection sensor 721, for example, a triangulation type laser length measuring device that observes the return position of the laser beam projected toward the target mirror 75 that is the subject and detects the position of the target mirror 75 is used. It can be used suitably. Further, as an example, the one disclosed in JP-A-10-267648 can be used. Hereinafter, a schematic configuration and position detection principle of the absolute type vertical displacement detection sensor 721 will be described.

  The absolute vertical displacement detection sensor 721 includes a laser light source 72a that emits a semiconductor laser for measurement, a collimator lens 72b, and a quadrant light receiving means 72c that can detect the position of the center of gravity of a received light beam. The quadrant light receiving means 72c is an optical position sensor (PSD: Position Sensitive Detector), but a charge coupled device image sensor can also be suitably used as the light receiving means. The sensor control circuit 72d performs driving of the laser light source 72a, processing of signals output from the quadrant light receiving means 72c, various calculations, and the like.

  In the above configuration, the laser light emitted from the laser light source 72a becomes a substantially parallel light beam via the collimator lens 72b, enters the target mirror 75, is reflected by the target mirror 75, and then is received by the four-divided light receiving means 72c. Incident on the surface. The quadrant light receiving means 72c has a function of outputting a signal corresponding to the position of the center of gravity of the light receiving spot. Therefore, the center position of the light receiving spot when the target mirror 75 moves in the Z-axis direction together with the lens frame 74 can be calculated by the sensor control circuit 72d. Based on the calculation result, the displacement of the lens LE in the Z-axis direction is detected.

  Here, since the triangulation type laser length measuring device can detect the absolute distance to the subject, the power to the absolute type vertical displacement detection sensor 721 is once cut off, and when the power is supplied again, it is on the Z axis. The Z coordinate of the lens LE relative to the reference point can be detected with sub-μm accuracy.

  FIG. 12 shows an incremental vertical displacement detection sensor 722. As the incremental type vertical displacement detection sensor 722, for example, an interference type laser length measuring device can be preferably used. The interference type laser length measuring device combines the reflected light from the target mirror 75 that is the subject and the reflected light from the total reflection mirror 72e in the incremental type vertical displacement detection sensor 722 to interfere with each other. It is a device that measures the amount of movement of the subject by counting the wave number of. As an example of the interference type laser length measuring device, a position detection device disclosed in Japanese Patent Laid-Open No. 11-94514 can be used. Hereinafter, the schematic configuration and position detection principle of the incremental type vertical displacement detection sensor 722 will be described.

  The incremental type vertical displacement detection sensor 722 includes a laser light source 72a that emits a semiconductor laser for measurement, a collimator lens 72b, a four-divided light receiving means 72c, and a sensor control circuit 72d. Further, the incremental type vertical displacement detection sensor 722 includes a total reflection mirror 72e, a non-polarization beam splitter 72f, a polarization beam splitter 72g, an interference optical element 72h, and a light absorption means 72j.

  In the above configuration, the laser light emitted from the laser light source 72a is shaped into a loose condensed light beam (shown as a parallel light beam in FIG. 12) via the collimator lens 72b, and enters the non-polarizing beam splitter 72f. Then, 50% of the incident light beam passes through the non-polarizing beam splitter 72f and reaches the light absorbing means 72j, where it is absorbed. Further, the remaining 50% of the incident light beam is reflected by the non-polarization beam splitter 72f, travels to the left, and enters the polarization beam splitter 72g.

  The S-polarized light beam reflected by the surface of the polarizing beam splitter 72g is incident on the target mirror 75, reflected by the target mirror 75, re-reflected by the surface of the polarizing beam splitter 72g, and passes through the original optical path. Return to the non-polarizing beam splitter 72f. On the other hand, the P-polarized light beam transmitted through the polarization beam splitter 72g enters the total reflection mirror 72e, is reflected by the total reflection mirror 72e, and then re-transmits through the polarization beam splitter 72g and returns through the original optical path.

  With the above structure, the S-polarized reflected light from the target mirror 75, which is the moving object, and the P-polarized reflected light from the total reflection mirror 72e are combined by the polarization beam splitter 72g. Then, 50% of the combined light passes through the non-polarizing beam splitter 72f and enters the interference optical element 72h. The interference optical element 72h includes, in order from the left incident surface side, a quarter-wave plate, an aperture, a phase diffraction grating having a staggered grating structure, and a four-division polarizing plate arranged with the polarization directions shifted by 45 degrees from each other. , Is arranged.

  Here, the combined light is linearly polarized light that is orthogonal to each other, but is converted to circularly polarized light that is opposite to each other by the ¼ wavelength plate in the interference optical element 72h. Next, only a central light beam having a large light quantity passes through the aperture, and is amplitude-divided into four light beams by a phase diffraction grating having a staggered grating structure, and passes through a four-part polarizing plate. The quadrant light beam generated in the above process is converted into interference light whose brightness timing is shifted by 90 degrees in terms of phase. Then, the four light beams are respectively incident on the light receiving portion of the four-divided light receiving means 72c, and the four output signals are processed by the following method.

  First, a DC component is removed by extracting the differential of two signals having a phase difference of 180 degrees from each other, and the resulting signal is used as an A-phase signal. Similarly, by extracting the differential of the remaining two signals having a phase difference of 180 degrees from each other, the DC component is removed, and the resulting signal is used as the B-phase signal. Since the A-phase signal and the B-phase signal have a phase difference of 90 degrees, when this is displayed on an oscilloscope, a circular Lissajous waveform is obtained (see FIG. 18).

  Here, when the change in the optical path length of the measurement light reciprocating between the incremental type vertical displacement detection sensor 722 and the target mirror 75 coincides with the light source wavelength, the interference signal changes by one period. That is, when the subject moves a distance that is half the wavelength of the light source, the interference signal changes by one period. Therefore, if the light source wavelength of the measurement laser light is λ = 800 nm, the interference signal also changes by one cycle every time the target mirror 75 moves by 400 nm, and the Lissajous waveform makes one round along the circular locus. Therefore, the amount of movement of the subject can be measured in units of 400 nm by measuring the wave number of the interference signal by a known method. Further, if the interference signal is divided (divided) using an electrical phase dividing device, the amount of movement of the subject can be detected with sub-nm accuracy.

  FIG. 13 shows an incremental lateral displacement detection sensor 723. The above-described incremental type vertical displacement detection sensor 722 detects the amount of vertical movement of the target mirror 75. On the other hand, the incremental lateral displacement detection sensor 723 detects the lateral movement amount of the target mirror 75, that is, the radial movement amount of the lens LE. Therefore, the incremental vertical displacement detection sensor 722 is an internal optical element. The arrangement is different. However, since the detection principle of both sensors is the same, description is abbreviate | omitted.

  FIG. 14 shows a control circuit for controlling the exposure operation and lens driving operation of the semiconductor exposure apparatus 1 of FIG. As shown in the figure, the control circuit comprises a main body control means 8 for controlling the operation of the entire exposure apparatus and a lens control means 6 for controlling the position of the lens. The main body control means 8 has a main body CPU 81, to which a mount control means 82, an illumination control means 83, a reticle stage control means 84, a wafer stage control means 85, and an interferometer control means 86 are connected.

  The mount control means 82 controls the vibration isolation operation of the lens barrel mount 32. The illumination control unit 83 controls the illumination mode and light quantity of the illumination unit 4. The reticle stage control means 84 performs drive control of the reticle stage R1. Wafer stage control means 85 controls the drive of wafer stage WF1. The interferometer control means 86 performs measurement control of the interferometer 51.

  The lens control means 6 has three lens CPUs 61, and this lens CPU 61 is provided for each lens unit 7 shown in FIG. That is, the same number of lens CPUs 61 as the number of lens units 7 incorporated in the projection optical system 2 are provided. Each lens CPU 61 is connected with three drive mechanism control means 62, and each drive mechanism control means 62 controls the drive mechanism 71 shown in FIG. The drive mechanism control means 62 incorporates a piezo driver 71a and a piezo element 71b, and drives the piezo actuator 713 shown in FIG.

  The lens CPU 61 is connected with three lens position detecting means 72. The lens position detection means 72 is provided with an absolute vertical displacement detection sensor 721, an incremental vertical displacement detection sensor 722, and an incremental lateral displacement detection sensor 723.

  In the above configuration, the lens CPU 61 communicates with the main body CPU 81, drives and controls the drive mechanism 71 based on a predetermined program, drives the lens toward the target position, and various aberrations of the projection optical system 2 are minimized. The drive mechanism 71 is controlled to be

  FIG. 15 shows a control block of the lens unit 7. The lens target position calculation block B81 in the main body CPU 81 calculates the Z-axis direction drive target value Z0 and the tilt drive target values θX0, θY0 of the lens reference point L0, and transmits the results to the lens CPU 61. Then, the target value is input to the addition point P1 in the lens CPU 61.

  At the addition point P1, feedback signal values Zm, θXm, θYm, which will be described later, are subtracted from the target values Z0, θX0, θY0, and the residual signals are input to the compensation circuit block B71. In the compensation circuit block B71, a PI compensator is used to improve control stability. The signal that has passed through the compensation circuit block B71 is input to the coordinate transformation block B72.

  In the coordinate conversion block B72, the residual signal is converted into a piezo drive voltage command value corresponding to the Z-axis direction drive target value Z0 of the drive mechanism 71. Here, if the drive target values of the lens LE are Z0, θX0, and θY0, the drive displacements W1, W2, and W3 of the drive mechanism 71 (see FIG. 6) can be calculated by multiplying by a predetermined conversion matrix. That is, they are related by the following formula 1.

  Therefore, the coordinate transformation block B72 may perform transformation according to the following expression 2 in order to output a voltage value.

  Here, δZ0, δθX0, and δθY0 are control residuals, VP1, VP2, and VP3 are control command voltage signals to the piezo driver 71a, and CV is a conversion coefficient for converting a displacement amount into a control voltage.

  When the control command voltage signals VP1, VP2, and VP3 calculated by Expression 2 are input to the piezo driver 71a, the piezo element 71b expands by a predetermined amount. The lens frame 74 is given drive displacements W1, W2, and W3 via a direction changing portion 71c including a displacement extraction member 711 and a direction changing member 712 of the drive mechanism 71, and the lens frame 74 and the lens are moved to desired positions. Move to.

  Then, the three sets of lens position detection means 72 measure the amount of displacement of the target mirror 75 (see FIGS. 11 to 13) fixed to the lens frame 74, and a total of nine types of measurement signals are output to the sensor output origin defining block. Input to B73. Here, the nine types of measurement signals are a mixture of an absolute position signal, which is a non-incremental displacement signal, and an incremental displacement signal. However, in the description of FIG. 16, the vertical displacements H1, H2, and H3 that are absolute displacement components in the optical axis C direction and the horizontal displacements S1, S2, and S3 that are absolute displacement components in the lens radial direction are considered.

  Here, the movement amount and coordinate definition of each part of the lens frame 74 will be described with reference to FIG. As shown in the drawing, the flange portion 74a of the lens frame 74 that holds the lens LE is fastened to the displacement output portions of the three sets of drive mechanisms 71 (described in FIG. 2) arranged at intervals of 120 degrees. Then, assuming that the vertical drive displacements of the three sets of drive mechanisms 71 are W1, W2, and W3, the three portions of the flange portion 74a move along the optical axis C direction by W1, W2, and W3, respectively.

  Further, a lens position detecting means 72 (described in FIG. 2) is disposed between the drive mechanisms 71 in the flange portion 74a. Assuming that the vertical displacements of the three sets of position detecting means are H1, H2, and H3 and the horizontal displacements are S1, S2, and S3, as shown in the figure, at three locations of the flange portion 74a, H1, H2, and H3. , S1, S2, and S3 are detected.

  FIG. 17 defines lens coordinates for controlling the position of the lens LE. The center of the incident surface (the vertex of the first lens surface) when the lens LE is placed at the design reference position is defined as the coordinate origin. Define the XYZ Cartesian coordinate system. Here, the Z axis is the optical axis C, the X axis is the radial axis, and the Y axis is the radial axis orthogonal to the X axis. The center of the incident surface of the lens LE is a lens position representative point at the time of lens position control, that is, a lens reference point L0.

  When the three sets of drive mechanisms 71 are driven with the same drive displacement (W1 = W2 = W3), the lens reference point L0 translates in the Z-axis direction, and the amount of displacement is represented by Zm. Further, when the drive mechanism 71 is driven so that the displacement amounts of the three sets of drive mechanisms 71 are different, the lens LE tilts around the X axis or the Y axis according to the relative change in the drive amount of the three sets of drive mechanisms 71. Displace. In the figure, the amount of tilt displacement is represented by θXm and θYm.

  In general, when the lens LE is tilt-driven, the tilt center does not coincide with the lens reference point L0, and the translational (lateral shift) displacement in the Y-axis direction with the tilt around the X-axis is about the Y-axis. A translational (lateral shift) displacement in the X-axis direction occurs with the tilt. These lateral shift displacements are represented as Ym and Xm, respectively.

  Next, returning to FIG. 15, the coordinate conversion block B74 converts the six types of displacement information calculated in the above-described process into the position coordinates of the lens LE. Here, the pre-conversion measurement values H1, H2, H3, S1, S2, and S3 of the lens LE and the coordinate values (conversion values after conversion) Xm, Ym, Zm, θXm, θYm, and Rm of the lens reference point L0 These can be converted into each other using a predetermined conversion matrix. That is, they are related by the following Expression 3.

  Here, a total of five converted measurement values Xm to θYm are as described above. Rm is a value related to the diameter dimension of the lens frame 74. In the case of the present embodiment, some of the coefficients included in the conversion matrix are substantially zero, and are expressed as the following Expression 4.

  That is, the Z-axis translational displacement Zm and the tilt displacements θXm, θYm of the lens are calculated from the vertical displacements H1, H2, H3 that are absolute displacements in the Z-axis direction, and are absolute displacements in the lens radial direction. It does not depend on the horizontal displacements S1, S2, S3. Further, the displacements Xm, Ym in the direction orthogonal to the Z axis of the lens LE and the diameter dimension Rm of the lens frame 74 are calculated by the horizontal displacements S1, S2, S3 which are absolute displacements in the lens radial direction, and are in the Z axis direction. It does not depend on the vertical displacements H1, H2 and H3 which are absolute displacements. In other words, the displacement component in the direction having the degree of freedom of control is detected by the sensor having the detection sensitivity in the Z-axis direction, and the displacement component in the direction having no degree of freedom of control has the detection sensitivity in the direction orthogonal to the Z-axis. It is detected by the sensor it has.

  Here, vertical displacements H1, H2, and H3 before conversion and various displacements Zm, θXm, and θYm after conversion are displacements obtained by actively driving the optical element. On the other hand, horizontal displacements S1, S2, S3 before conversion and various displacements Xm, Ym, Rm after conversion are secondary displacements that cannot be actively controlled.

  Of the six converted measurement values obtained in the above-described process, three pieces of information necessary for position control of the lens LE are Zm, θXm, and θYm. Therefore, the control residual, which is the difference from the target values Z0, θX0, θY0, can be input to the compensation circuit block B71 by inverting this value at the addition point P1. Then, by re-driving the piezo actuator 713 so as to eliminate the control residual, the lens LE can be positioned with high accuracy.

  On the other hand, the measured six pieces of displacement information Xm, Ym, Zm, θXm, θYm, Rm are transmitted to the wafer stage position correction block B82 of the main body CPU 81. Of these displacement information, for example, Xm and Ym are used to reduce the overlay error of the reticle image. This is because the reticle image superposition error occurs when the lens LE generates lateral shift displacements Xm and Ym by a tilt operation.

  Further, the lateral shift displacements Xm and Ym move the reticle image formed on the wafer WF, and also change the aberration of the projection optical system. Therefore, by correcting the scan displacement amount of the wafer stage WF1 based on Xm and Ym, the positioning accuracy of the reticle image on the wafer WF can be improved, and the overlay error can be reduced.

  Further, for example, Rm is used for detecting a failure of the lens unit 7 and a change with time of the dimension. Since the diameter of the lens frame 74 is essentially invariable, it can be determined whether or not the lens unit 7 has failed by determining whether or not the measured value Rm has changed significantly. Further, if it is determined whether or not Rm has changed by a minute amount, it is possible to detect a change in size with time. When Rm changes by a minute amount, it is possible to estimate thermal expansion of the lens LE and the lens frame 74, long-term deformation due to metal aging, or the like.

  Next, a method for adjusting each sensor constituting the lens position detecting means 72 provided in the lens unit 7 will be described. Here, adjustment refers to waveform shaping of the periodic signal output from the incremental displacement detection sensor, so-called Lissajous adjustment, and adjustment that defines the measurement reference point of the incremental displacement detection sensor based on the output value of the absolute detection displacement sensor. .

  First, the principle of Lissajous adjustment will be described. Lissajous adjustment is adjustment for shaping the waveform of each sine wave as a pre-process when calculating a displacement signal from a plurality of sine waves having different phases. As shown in FIG. 18, the horizontal axis represents the A-phase output of the interference signal calculated by the sensor control circuit 72d, and the vertical axis represents the B-phase output of the interference signal. Since the A-phase signal and the B-phase signal are sine waves having the same output amplitude and a phase difference of 90 degrees, the trajectory of both phases is ideally circular. However, if there is a manufacturing error of each optical member in the sensor or a difference in sensitivity between the four-divided light receiving parts, the amplitudes of both AB phases differ or the phase difference deviates from 90 degrees. A simple elliptical trajectory. When this elliptical locus is electrically divided, a division error occurs and an accurate displacement signal cannot be obtained. In order to avoid this problem, the Lissajous adjustment means (not shown) included in the sensor control circuit 72d is adjusted, and the waveform is shaped (Lissajous adjustment) so that the Lissajous waveform becomes a circle (solid line in the figure).

  Next, reference point scanning and a reference point defining method of the vertical displacement detection sensor will be described. 19A to 19D, the horizontal axis represents the position of the lens LE in the Z-axis direction, and the vertical axis represents the displacement measurement value of the target mirror 75 in the Z-axis direction, that is, the sensor output value. FIG. 19A shows a case where the position of the lens LE in the Z-axis direction (lower limit position) coincides with the designed base position Hbtm1 when the drive mechanism 71 is powered off.

  In this state, when a voltage is applied to the piezo actuator 713 of the three sets of drive mechanisms 71 to drive the lens LE in the positive direction of the Z axis, the output value Habs of the absolute type vertical displacement detection sensor 721 changes as indicated by a broken line. To do. Here, in order to position the lens LE at the optical origin that is an ideal design position, that is, when the sensor 721 outputs a signal corresponding to the origin when it reaches the center of the movable range, it is mechanical or electrical in advance. Assume that the sensor 721 is adjusted.

  As described above, the sensor 721 can detect the absolute position of the subject, but the linearity of the detection output signal is relatively inferior to the incremental vertical displacement detection sensor 722 that is an optical interferometer. That is, the output waveform (dotted line) of the absolute type vertical displacement detection sensor 721 passes through the origin 0 like Habs, but is non-linear compared to FIG.

  On the other hand, the incremental vertical displacement detection sensor 722 is superior to the absolute vertical displacement detection sensor 721 in terms of detection resolution and detection output signal linearity. However, although the amount of change from when the power is turned on to the sensor 722 is output, the absolute position of the subject cannot be detected, so the output value Hinc is as shown by the solid line in FIG.

  Therefore, if the output value Hinc0 of the incremental vertical displacement detection sensor 722 when the lens LE passes the origin matches the output value Habs0 of the absolute vertical displacement detection sensor 721, the reference point position (measurement) of the detection sensor 722 is measured. Standards). By the above operation, the origin of the incremental type vertical displacement detection sensor 722 is defined by the reference point Habs0 of the absolute type vertical displacement detection sensor 721, and the absolute position of the subject can be detected substantially.

  FIG. 19C shows a case where the position of the lens LE in the Z-axis direction (base position, that is, the lower limit position) is located at Hbtm2 that is slightly deviated from the design value Hbtm1 when the power to the drive mechanism 71 is turned off. . Here, as a possibility that the base position of the lens is shifted, an initial length change due to creep of the piezo actuator 713, an initial length change due to replacement of the piezo actuator 713, or a preload adjustment amount change is assumed.

  In this state, a voltage is applied to the piezo actuator 713 to drive the lens LE in the positive direction of the Z axis. Then, the output value Habs of the absolute type vertical displacement detection sensor 721 is different from that in FIG. 19A at the output start point, but the subsequent output signal coincides with FIG.

  On the other hand, the incremental type vertical displacement detection sensor 722 outputs a change from the time when the power is turned on. At this time, if the base position of the lens LE is deviated for the above-described reason, the output value Hinc indicated by the solid line is This is different from the output value Hinc indicated by a thin line (corresponding to the solid line in FIG. 19A).

  However, as described above, when the output value Hinc0 of the incremental vertical displacement detection sensor 722 when the lens LE passes the origin and the output value Habs0 of the absolute vertical displacement detection sensor 721 are matched, the reference point of the sensor 722 is obtained. A position is defined. The result is shown in FIG. 19D, and the origin of the incremental type vertical displacement detection sensor 722 coincides with that in FIG.

  That is, by performing the above adjustment, the reference point of the incremental type vertical displacement detection sensor 722 is defined, and the sensor 722 has a function as a high accuracy absolute type vertical displacement detection sensor 721.

  Next, the displacement that occurs secondary when the lens LE is tilt driven will be described. FIG. 20 shows a form before the lens LE is tilt-driven around the X axis. FIG. 21 is an enlarged view of the vicinity of the incremental lateral displacement detection sensor 723 of FIG.

  Here, in order to tilt-drive the lens LE counterclockwise around the X axis, in the three sets of drive mechanisms 71 in FIG. The output part of the lower left drive mechanism 71 is driven and controlled in the positive direction of the Z axis. This will be explained with reference to FIG. 16. When the driving mechanism 71 is controlled by assigning -δW to the driving displacement W1 and + 2δW to the driving displacements W2 and W3, translation in the Z-axis direction does not occur at the lens reference point L0. Only a tilt around the X axis occurs in LE.

  When the tilt center axis of the lens LE at this time is indicated by a point Pc in FIG. 20, the tilt center axis Pc is a plane passing through the upper surface (see FIG. 8) of the lens frame drive link 712d of the three sets of drive mechanisms 71. It substantially coincides with the axis that is above and intersects the optical axis C. The reason will be described below.

  In FIG. 16, when the three sets of drive mechanisms 71 have the same Z-axis direction drive displacement, the lens frame 74 translates in the Z-axis direction. On the other hand, when the Z-axis direction drive displacements of the three sets of drive mechanisms 71 are different from each other, the lens frame drive link 712d (described in FIG. 10) of each drive mechanism 71 translates in a different amount in the Z-axis direction, resulting in a lens. The frame 74 tilts. By this tilting operation, the lens frame driving link 712d to which the lens frame 74 is fastened is also forcedly inclined by an angle equal to the tilt angle.

  At this time, as described with reference to FIG. 10, the support link 712e for restraining the lens frame drive link 712d has a structure capable of being twisted and deformed around the U axis. Therefore, the lens LE can be tilted. Therefore, the tilt center axis Pc of the lens LE in FIG. 20 is in the vicinity of the intersection between the surface including the thin portion 712e1 (described in FIG. 6) of the support link 712e and the optical axis C. Here, the Z-axis direction height of the surface including the thin portion 712e1 is substantially equal to the height of the fastening surface between the lens frame drive link 712d and the lens frame 74.

    That is, when the lens LE in the lens unit 7 is tilt-driven, there is no fixed tilt center as defined by a bearing or a pivot mechanism, but a portion with low torsional rigidity is formed in a part of the drive mechanism 71. Thus, the tilt center axis Pc of the lens LE is defined.

  Next, the lateral shift amount when the lens LE is tilted will be described. 20 and 24, RR is a line segment connecting the tilt center axis Pc and the displacement measurement point P11 of the incremental lateral displacement detection sensor 723 in the target mirror 75. Point P12 is the intersection of a horizontal line passing through the tilt center axis Pc and a vertical line passing through the displacement measurement point P11. A right triangle PcP11P12 is assumed by these three points Pc, P11, and P12.

  RS is the Y-axis direction component of the line segment RR, and HS is the Z-axis direction component of the line segment RR. Α is a relative angle between the horizontal line passing through the tilt center axis Pc and the line segment RR. In FIG. 21, the distance in the Y-axis direction between the reflecting surface of the target mirror 75 onto which the measurement beam of the sensor 723 is projected and the tilt center axis Pc is defined as RS.

  FIG. 22 shows a cross section when the lens LE is tilted in the counterclockwise direction by an angle β. FIG. 23 shows an enlarged view of the vicinity of the incremental lateral displacement detection sensor 723 of FIG. Along with the tilt displacement of the lens LE, the target mirror 75 is displaced in the positive direction of the Z axis and is also displaced by a small amount in the negative direction of the Y axis (right direction in FIGS. 22 and 23). Therefore, the right triangle PcP11P12 defined in FIG. 24 moves to become a right triangle PcP21P22. Therefore, the Y-axis direction component RS between the measurement beam irradiation part P21 and the tilt center axis Pc increases by δRS.

  Next, the lateral shift amount δRS will be described. In FIG. 24, a right triangle PcP11P12 indicated by a solid line represents before tilt driving, and a right triangle PcP21P22 indicated by a broken line represents after tilt driving. However, in FIG. ing. Here, ∠P11PcP12 = α, and the following formula 5 is established.

  Further, assuming that the line segment connecting the vertex P11 before tilt and the vertex P21 after tilt is ARC, the following formula 6 is established when the tilt angle β is small.

  If the intersection of the horizontal line passing through the point P11 and the vertical line passing through the point P21 is P3, the distance δRS between the point P11 and the point P3 is the lateral shift amount in the Y-axis direction of the target mirror 75 caused by the tilt drive of the lens LE. The following formula 7 holds.

  However, γ = ∠P12P11P21 = ∠P3P21P11. Then, assuming that the tilt angle β is a negligible angle with respect to other angles including α, and paying attention to the base angle of the isosceles triangle PcP11P21 in FIG. , Holds.

  Therefore, Expression 8 is substituted into Expression 7, and further using Expression 5 and Expression 6, the following Expression 9 is obtained.

  That is, when the Z-axis dimension between the detection beam position P11 of the incremental lateral displacement detection sensor 723 and the tilt center axis Pc of the lens LE is HS, the tilt drive angle is β, and the lens LE is tilt driven, the displacement (δRS) is detected. A signal is generated at the sensor 723.

  On the other hand, assuming that the light source wavelength of the sensor 723 is λ, an interference signal corresponding to one cycle of the Lissajous waveform described in FIG. 18 is obtained when the distance between the sensor 723 and the target mirror 75 changes by λ / 2. Therefore, in order to tilt drive the lens LE to generate an interference signal in the sensor 723 and perform the Lissajous adjustment using this, it is necessary to satisfy the following Expression 10.

  From the formulas 9 and 10, the following formula 11 is derived.

  As an example of specific numerical values, if λ = 800 (nm) = 8 × 10 −4 (mm) and β = 1 × 10 −4 (rad) are substituted into Equation 11, HS> 4 (mm) is obtained. . Specifically, an interference type laser length measuring device having a light source wavelength of λ = 800 (nm) is used, and the practical maximum tilt angle in the drive mechanism 71 is β = 1 × 10 −4 (rad). The lens unit 7 is used. Then, the measurement beam position of the length measuring device is separated by 4 (mm) or more in the Z-axis direction with respect to the tilt center axis Pc of the lens LE. Lissajous adjustments can be made.

  On the other hand, when the lens LE is tilt-driven around the axis parallel to the Y axis by the same tilt angle β, the lateral displacement amount in the X-axis direction observed by the upper right and upper left lens position detecting means 72 in FIG. √3 / 2≈0.87 times the δRS calculated by Therefore, the Z-axis direction separation amount HS calculated by Expression 11 needs to be about 1.15 times, that is, 4.6 (mm) or more. Further, in order to accurately perform the Lissajous adjustment described with reference to FIG. 18, it is desirable that the interference signal has about two cycles. Therefore, in consideration of these circumstances, it is desirable to secure a separation amount that is at least twice as large as the separation amount HS calculated by Expression 11.

  FIG. 25 shows a flowchart of a sensor initialization routine for executing the adjustment process of the various detection sensors 721, 722, 723. First, in step S101, power supply to the various detection sensors 721, 722, and 723 included in the lens position detection unit 72 is started, and the lens position can be measured. In this step, since the three sets of drive mechanisms 71 are not yet operated, the lens LE is in the base position, that is, the lowest position in the movable range in the Z-axis direction.

  Next, in step S102, the three sets of driving mechanisms 71 are driven at a constant speed to move the lens LE upward by a predetermined amount in the Z-axis direction. In step S103, during the constant speed driving, the Lissajous adjustment of the three incremental vertical displacement detection sensors 722 is performed based on the principle described with reference to FIG. In step S104, the lens LE is returned to the lowest point in the Z-axis direction, that is, the base position. In step S105, the three sets of drive mechanisms 71 are driven by an equal amount to move the lens LE upward in the Z-axis direction to the drive range upper limit position. In step S106, reference point scanning of the absolute type vertical displacement detection sensor 721 is performed based on the method described in FIG. 19 during execution of step S105. In step S107, the reference point of the incremental type vertical displacement detection sensor 722 is defined.

  After executing step S107, in step S108, the lens LE is once returned to the base position, that is, the lowest point. Next, in step S109, the lens LE is moved to a design reference position (optical origin), that is, the position of the origin 0 shown in FIG. In step S110, the lens LE is tilted by a predetermined amount around the X axis and the Y axis. In step S111, the Lissajous adjustment of the three incremental lateral displacement detection sensors 723 is performed in order while executing step S110.

  In step S112, the lens LE is moved again to the design reference position (optical origin), that is, the position of the origin 0 shown in FIG. In step S113, reference points for the three incremental lateral displacement detection sensors 723 are defined. In the present embodiment, an absolute lateral displacement detection sensor is not provided. The reason is as follows. The lens frame 74 constrained by the three sets of drive mechanisms 71 cannot be actively controlled in the X-axis direction and the Y-axis direction orthogonal to the Z-axis, but instead has rigidity in the X-axis direction and the Y-axis direction. This has the advantage of high position reproducibility.

  That is, the position reproducibility in the X-axis direction and the Y-axis direction is good when the lens LE is returned to the optical origin position based on the output signals of the absolute type vertical displacement detection sensor 721 and the incremental type vertical displacement detection sensor 722. . Accordingly, by defining the output value of the incremental lateral displacement detection sensor 723 when the lens LE is controlled at the optical origin position as the origin, the absolute absolute value in the lateral direction can be obtained without the absolute lateral displacement detection sensor. It is possible to define the position with high accuracy. Then, the sensor initialization routine is completed.

  With the above flow, the Lissajous adjustment of the incremental vertical displacement detection sensor 722 and the incremental lateral displacement detection sensor 723 and the origin specification of the lens LE using the absolute vertical displacement detection sensor 721 are performed. Then, it becomes possible to accurately detect the translational displacement of the lens LE in the Z-axis direction and the tilt displacement around the X-axis and the Y-axis.

  The flow shown in FIG. 25 is performed when the assembly of the lens unit 7 is completed, and is operated for final adjustment when the semiconductor exposure apparatus 1 shown in FIG. 1 is completed, and for manufacturing a semiconductor wafer. It is executed as appropriate during the operation.

  According to the first embodiment described above, the incremental vertical displacement detection sensor 722 that measures the relative position of the lens LE in the optical axis C direction of the lens LE, and the absolute vertical displacement detection sensor that measures the absolute position of the sensor 722. 721. For this reason, both the absolute position and the relative position of the lens LE can be measured, and high-precision positioning accuracy can be exhibited even when the power supply is restarted.

In addition, since the incremental lateral displacement detection sensor 723 that measures the relative position of the lens LE in the direction orthogonal to the optical axis C direction is provided, the tilt displacement amount of the lens LE can be calculated. Further, by driving the lens LE to be tilted by the driving mechanisms 71 provided at three locations, the alignment error of the reticle image can be eliminated.
(Sensor modification 1)
In the embodiment described above, the example using the absolute type vertical displacement detection sensor 721 has been described. This is a triangulation type sensor that can quantitatively detect and output a position from a predetermined reference point as a reference point detection sensor in the Z-axis direction. In addition, for example, a photoelectric switch type photosensor that determines whether or not the lens LE that is an optical element exists in the vicinity of the reference point may be used.

  FIG. 26 schematically shows a cross section in the vicinity of the lens position detecting means 72. As shown in the figure, a vertical origin detection sensor 721A different from the absolute type vertical displacement detection sensor 721 of the above-described embodiment is fixed to the flat portion of the fixed barrel 73. The sensor 721A is a photoelectric switch type photosensor, and has a light projecting means 72k made of a light emitting diode and a light receiving means 72m made of a silicon photodiode. The light projecting means 72k and the light receiving means 72m are disposed so as to face each other with respect to the light projecting means 72k with the detection recess 72n interposed therebetween. The drive control and signal processing of the light projecting means 72k and the light receiving means 72m are performed using a sensor control circuit 72p.

  A target mirror 75 is attached to the flange portion 74a of the lens frame 74, and a slit member 76 is attached to the target mirror 75 along the vertical direction. The slit member 76 is disposed between the light projecting means 72k and the light receiving means 72m, and measures the change in the position of the lens frame 74 by moving the slit member 76 by driving the lens frame 74 as a holding frame. can do. More specifically, the slit 76a is formed in one portion of the insertion portion inserted into the detection recess 72n. In the state where the lens LE is positioned at the design reference point (optical origin), the center of the through hole of the slit 76a is an axis connecting the center of the projected light beam of the light projecting means 72k and the center of the light receiving surface of the light receiving means 72m. Adjustments are made in advance so as to match.

  Next, the reference point scanning and reference point defining method of the vertical origin detection sensor 721A will be described with reference to FIG. In FIG. 27A to FIG. 27D, the horizontal axis is the position of the lens LE in the Z-axis direction. The vertical axis represents the sensor output value of the vertical origin detection sensor 721A or the incremental vertical displacement detection sensor 722. FIG. 27A shows a case where the Z-axis direction position (lower limit position) of the lens LE coincides with the designed base position Hbtm1 when the power to the drive mechanism 71 is turned off.

  In this state, when a voltage is applied to the piezo actuators 713 included in the three sets of drive mechanisms 71 to drive the lens LE in the positive direction of the Z axis, the output value Habs of the vertical origin detection sensor 721A changes as indicated by a broken line. Here, when the lens LE reaches the optical origin that is the ideal position in design, that is, the center of the movable range, the detected light beam central axis of the sensor 721A coincides with the center of the slit 76a, and the output signal of the sensor 721A The sensor position is adjusted in advance so as to maximize the.

  Therefore, by driving the subject in one direction and measuring the output value of the sensor 721A, it can be recognized whether or not the subject exists in the vicinity of the reference position, and the subject is determined from the differential value of the output signal of the sensor 721A. It can be detected whether or not the sample has passed the reference position. However, when the subject is greatly separated from the reference position, the sensor 721A cannot obtain information on the position of the subject.

  On the other hand, the incremental type vertical displacement detection sensor 722 has excellent linearity and detection resolution of the detection output signal, but cannot detect the absolute position of the subject, and the change from when the sensor 722 is turned on is detected. In order to output, the output value Hinc changes like a solid line. Therefore, the reference value position of the incremental vertical displacement detection sensor 722 is made by matching the output value Hinc0 of the incremental vertical displacement detection sensor 722 when the lens LE passes the origin with the output value Habs0 of the vertical origin detection sensor 721A. Is defined. The result is shown in FIG. 27 (b). By the above operation, the origin of the incremental vertical displacement detection sensor 722 is defined by the reference point Habs0 of the vertical origin detection sensor 721A, and the absolute position of the subject can be substantially moved. Detection is possible over the entire range.

  FIG. 27C shows a case where the position of the lens LE in the Z-axis direction (base position, that is, the lower limit position) is located at Hbtm2 that is slightly deviated from the design value Hbtm1 when the power to the drive mechanism 71 is turned off. Here, as a possibility that the base position of the lens LE is displaced, an initial length change due to creep of the piezo actuator 713, an initial length change due to replacement of the piezo actuator 713, a preload adjustment amount change, and the like are assumed. When a voltage is applied to the piezo actuator 713 in this state to drive the lens LE in the positive direction of the Z axis, the output value Habs of the vertical origin detection sensor 721A is different from that in FIG. The subsequent output signal has the same shape as the broken line in FIG.

  On the other hand, the incremental type vertical displacement detection sensor 722 outputs a change from when the power is supplied to the sensor 722. If the base position of the lens LE is deviated for the above-described reason, an output value Hinc indicated by a solid line is output. Is different from the output signal in the case of FIG. 27B shown by a thin line. However, as described with reference to FIG. 27A, the output value Hinc0 of the incremental vertical displacement detection sensor 722 when the lens LE passes the origin is calibrated with the output value Habs0 of the vertical origin detection sensor 721A. A reference point position of the sensor 722 is defined. The result is shown in FIG. 27D, and the origin of the incremental type vertical displacement detection sensor 722 coincides with that in FIG.

That is, by performing the above adjustment, the reference point of the incremental type vertical displacement detection sensor 722 is defined, and the sensor 722 has a function as a high accuracy absolute type vertical displacement detection sensor.
(Sensor modification 2)
In the above-described embodiment, the example using the incremental displacement detection sensors 722 and 723 has been described. The sensors 722 and 723 were Michelson interferometers. In addition, for example, a linear encoder that diffracts coherent light incident on the diffraction grating and measures the movement amount of the diffraction grating by counting interference fringes formed by the diffraction light may be used.

  FIG. 28 shows a main part in the vicinity of the lens position detecting means. A first target 761 and a second target 762 are fixed to the lower surface (shown in FIG. 26) of the flange portion 74a of the lens frame 74 (not shown). The first target 761 is a transparent thin plate having a plane perpendicular to the radial direction of the lens LE, and has a diffraction grating 761a and an absolute code track 761b. Here, the diffraction grating 761a is a diffraction grating in which grooves extend in the tangential direction of the lens outer periphery (lens edge), in other words, grooves and peaks are alternately arranged in the optical axis direction of the lens. Also, the absolute code track 761b is a line in which the light shielding portion and the transparent portion are arranged in accordance with a predetermined rule along the optical axis direction, and a plurality of rows are arranged in the tangential direction of the outer periphery of the lens to form a code pattern. .

  The second target 762 is a flat transparent thin plate arranged perpendicular to the first target 761. A diffraction grating 762 a is formed on the second target 762. Here, the diffraction grating 762a is a diffraction grating having grooves extending in the lens optical axis direction, in other words, grooves and peaks alternately arranged in the radial direction of the lens.

  In order to detect the absolute position of the first target 761 in the Z-axis direction, a vertical displacement detection sensor 77A is provided. A concave portion is provided on the upper surface of the sensor 77A, and a first target 761 is inserted into the concave portion. To detect the displacement of the first target 761 in the Z-axis direction, for example, an encoder disclosed in Japanese Patent Laid-Open No. 9-126723 Is used. The detection principle will be briefly described.

  Part of the laser light emitted from the sensor enters the diffraction grating 761a, and the transmitted diffracted light reenters the sensor 77A. Then, the diffracted light is synthesized by a predetermined method in the sensor 77A, and interference fringes are created. Here, when the diffraction grating 761a is displaced in the Z-axis direction, the state of the diffracted light changes and the interference fringes in the sensor 77A move. Therefore, the photoelectric sensor that measures the interference fringes corresponds to the brightness of the interference fringes. A periodic signal (incremental signal) is output.

  On the other hand, the other part of the laser light emitted from the sensor 77A enters the absolute code track 761b, and the transmitted light returns to the absolute code light receiving element group arranged inside the sensor 77A. Then, the absolute position of the absolute code track 761b in the Z-axis direction can be detected by processing the output signal of the element group.

  Further, a lateral displacement detection sensor 77B is provided in order to detect a position in a direction intersecting the optical axis of the lens, that is, a relative movement amount of the second target 762 in the lens radial axis direction. A concave portion is provided on the upper surface of the lateral displacement detection sensor 77B, and the second target 762 is inserted therein. Then, the relative displacement in the lens radial direction is detected. As the sensor 77B, a sensor in which the absolute displacement detection function is omitted from the vertical displacement detection sensor 77A is used. Therefore, when the second target 762 moves in the lens radial direction, a periodic signal (incremental signal) corresponding to the movement amount is output.

  Note that the period of the incremental signal output from the optical interferometer of the above-described embodiment depends on the light source wavelength, but the period of the incremental signal output from the linear encoder of (Sensor Modification 2) is that of the diffraction grating. Depends on the grating pitch. Therefore, by using a diffraction grating having a sub-μm grating pitch, detection resolution similar to that of an optical interferometer can be obtained.

  Therefore, by counting the periodic signals, the movement amounts of the first target 761 and the second target 762 can be detected with a resolution of the lattice pitch order. Further, by electrically dividing the periodic signal, the detection resolution can be increased to the sub-nm order. When such a linear encoder is used, the effective area of the diffraction grating is enlarged, so that even if the coherence distance (coherence length) of the encoder light source is not sufficient, highly accurate position detection can be performed over a long stroke. .

  In the first embodiment described above, the optical elements constituting the projection optical system are all the lenses LE. However, in the second embodiment, the projection optical system includes the first to fourth lens groups LE1 to LE4 and the first to first lens groups. This is a catadioptric system composed of third mirrors M1 to M3. The driven object supported so as to be movable is a mirror. Further, the optical element of the first embodiment has three degrees of freedom in translation in the Z-axis direction and tilt displacement around the X-axis and Y-axis, but in the second embodiment, around the predetermined axis. Only tilt drive is allowed.

  FIG. 29 shows a projection optical system of an exposure apparatus provided with the second embodiment of the optical element holding device of the present invention. This projection optical system has a mirror unit MU including a reticle R, first to fourth lens groups LE1 to LE4, first to third mirrors M1 to M3, and a third mirror M3. Each of the first to fourth lens groups LE1 to LE4 includes a plurality of lenses. Since the components other than the projection optical system are the same as those in the first embodiment, description thereof is omitted.

  30 to 32 show a mirror unit MU for driving the third mirror M3 among the optical elements constituting the projection optical system of FIG. FIG. 30 is a plan view in which the third mirror M3 and the mirror frame 74A which is a holding frame are removed. FIG. 31 is a plan view when the third mirror M3 and the mirror frame 74A are attached. FIG. 32 shows a cross section taken along line XXXII-XXXII in FIG.

  In the first embodiment, since the optical element to be driven is supported and driven by a plurality of sets of drive mechanisms 71, a plurality of degrees of freedom in translation in the Z-axis direction and tilt around an axis orthogonal to the Z-axis are provided. Had. On the other hand, in the second embodiment, the third mirror M3 as an optical element to be driven is fixedly held on the mirror frame 74A, and the mirror frame 74A is supported by two pivot members 78. Yes.

  Here, the pivot member 78 is fixed to the plane portion of the fixed barrel 73 at a position facing each other across the Y axis, and the uppermost surface thereof and the flange portion 74a of the mirror frame 74A are coupled by screws. A hinge portion having a small cross-sectional area is provided at the intermediate portion of the pivot member 78, and the uppermost surface can be swung by a slight angle by elastic deformation of the hinge portion. Accordingly, the mirror frame 74A fastened to the two pivot members 78 is free to move only in the θX direction.

  Further, a set of drive mechanisms 71 is installed at an intermediate portion between the two pivot members 78, that is, on the Y axis plus side in FIG. Therefore, by applying a voltage to the piezo actuator 713 of the drive mechanism 71 and driving the fastening portion with the mirror frame 74A in the Z-axis direction, the mirror frame 74A is tilt-driven in the θX direction.

  On the other hand, one mirror position detection means 72A is arranged at a position facing the drive mechanism 71 across the optical axis C, and a position measurement assisting unit is provided at two positions separated by 120 degrees around the optical axis C. A certain mirror position detecting means 72B is arranged. The former mirror position detection means 72A includes an absolute vertical displacement detection sensor 721, an incremental vertical displacement detection sensor 722, and an incremental lateral displacement detection sensor 723, as in the first embodiment.

  The latter mirror position detection means 72B has an incremental vertical displacement detection sensor 722 as a first measurement unit and does not have an absolute vertical displacement detection sensor 721 as a second measurement unit. Further, the mirror position detection means 72B has an incremental lateral displacement detection sensor 723 which is a third measurement unit.

  The reason is that the drive direction of the mirror frame 74A is only in the θX direction and the degree of freedom is 1. Therefore, in order to recognize the absolute angle of the tilt angle from the reference position, only one absolute position detecting means is sufficient. Because. The reason why three incremental lateral displacement detection sensors 723 are provided as in the first embodiment is as follows. That is, the third mirror M3 has no degree of freedom in the X-axis and Y-axis directions, but is slightly in the X-axis and Y-axis directions due to the slight deflection of the pivot member 78 when the mirror frame 74A is tilt-driven. This is to detect the displacement.

  The positional relationship between the tilt center of the mirror frame 74A and the three incremental lateral displacement detection sensors 723 is arranged so as to satisfy the relationship described with reference to FIGS. It should be noted that the third mirror M3 is only allowed to tilt around the X axis and cannot be tilted around the Y axis. Therefore, it is necessary to calibrate all the incremental lateral displacement detection sensors 723 by tilt driving around the X axis.

  At this time, in the two mirror position detection means 72B in FIG. 31, the distance in the horizontal direction (Y-axis direction) between the tilt center and the measurement unit of the target mirror 75 is about half that of the first embodiment. Therefore, the same effect as that of the first embodiment can be obtained by setting the distance HS between the measurement unit of the target mirror 75 and the tilt center in the Z-axis direction to about twice that of the first embodiment.

  FIG. 33 shows a flowchart of a sensor initialization routine for executing the adjustment process of each sensor used in the second embodiment. First, in step S201, power supply to each sensor included in the mirror position detection means 72A and 72B is started to make a measurement possible state. In step S201, since the drive mechanism 71 is not yet operated, the third mirror M3 is at the base position, that is, one end of the tilt movable range around the X axis.

  In step S202, the drive mechanism 71 is driven at a constant speed, and the third mirror M3 is driven at a constant speed around the X axis by a predetermined amount. In step S203, the Lissajous adjustment of the three incremental vertical displacement detection sensors 722 and the three incremental lateral displacement detection sensors 723 is performed based on the principle described with reference to FIG. Do it. In step S204, the third mirror M3 is returned to the lowest point in the tilt direction, that is, the base position. In step S205, the drive mechanism 71 is driven again at a constant speed, and the third mirror M3 is driven at a constant speed up to the upper limit of the drive range in the θX direction.

  In step S206, reference point scanning of the absolute type vertical displacement detection sensor 721 included in the mirror position detection means 72A is performed based on the method described in FIG. 19 while executing step S205. However, in the second embodiment, since the third mirror M3 cannot be translated in the Z-axis direction, the lens position shown on the horizontal axis in FIG. 19 is interpreted by replacing the tilt position of the mirror. Next, in step S207, the reference points of the three incremental type vertical displacement detection sensors 722 are defined.

  After step S207 is executed, in step S208, the third mirror M3 is once returned to the base position, that is, the lowest point of the tilt drive range. Next, in step S209, the third mirror M3 is driven to the design reference position (optical origin), that is, the position of the origin 0 shown in FIG. In step S210, reference points for the three incremental lateral displacement detection sensors 723 are defined. After executing step S210, the sensor initialization routine is completed.

  In the second embodiment described above, the mirror position detection means 72B is provided in addition to the mirror position detection means 72A having the incremental type vertical displacement detection sensor 722 and the absolute type vertical displacement detection sensor 721. Since this mirror position detecting means 72B has an incremental vertical displacement detection sensor 722 and does not have an absolute vertical displacement detection sensor 721, the number of absolute vertical displacement detection sensors 721 can be reduced. it can.

  As mentioned above, although preferred embodiment of this invention was described, this invention is not limited to these embodiment, A various deformation | transformation and change are possible within the range of the summary.

  For example, in the second embodiment described above, the optical element is configured to be tilt driven around the tilt center axis parallel to the X axis, but may be configured to be tilt driven about the tilt center axis parallel to the Y axis. good.

  In the second embodiment described above, when the driven object is a mirror, the optical axis does not always coincide with the normal passing through the center of the reflecting surface. That is, as is clear from FIG. 29, the first mirror M1, the second mirror M2, and the third mirror M3 are not coincident with the normal line in which the central axes of the incident light beam and the emitted light beam stand at the center of the mirror. Therefore, the direction of the optical axis differs depending on the definition of the optical axis.

  On the other hand, the optimum driving direction required for these mirrors is selected from the incident light beam direction, the outgoing light beam direction, and the normal direction passing through the center of the reflecting surface in accordance with the design specifications of the projection optical system. The Therefore, when the optimum drive direction is different from the normal line at the center of the reflecting surface, the drive mechanism 71 is installed at a predetermined angle on the fixed barrel 73 so that the output shaft direction of the drive mechanism 71 coincides with the optimum drive direction. Just do it.

  In the first and second embodiments described above, the incremental lateral displacement detection sensor 723 of the optical element has detection sensitivity in the radial direction of the optical element, but has detection sensitivity in the outer tangent direction of the optical element. It may be arranged.

  In the first and second embodiments described above, the example in which the optical element is driven in the optical axis direction has been described. However, the optical element may be driven in a tilt direction with respect to the optical axis.

It is explanatory drawing which shows the exposure apparatus provided with 1st Embodiment of the holding | maintenance apparatus of the optical element of this invention. It is explanatory drawing which shows the detailed form of the lens unit of FIG. It is explanatory drawing which shows the form which attached the lens and the lens frame to the lens unit of FIG. It is sectional drawing which follows the IV-IV line of FIG. It is a top view which shows the detailed structure of the drive mechanism of FIG. It is a side view which shows the detailed structure of the drive mechanism of FIG. It is sectional drawing which follows the VII-VII line of FIG. It is a perspective view which shows a displacement extraction member and a direction change member. It is explanatory drawing which shows the link mechanism of a displacement extraction member. It is explanatory drawing which shows the link mechanism of a direction change member. It is explanatory drawing which shows an absolute type vertical displacement detection sensor. It is explanatory drawing which shows an incremental type vertical displacement detection sensor. It is explanatory drawing which shows an incremental type lateral displacement detection sensor. It is explanatory drawing which shows the control circuit for controlling the exposure operation and lens drive operation of exposure apparatus. It is explanatory drawing which shows the control block of a lens unit. It is explanatory drawing which shows the movement amount and coordinate definition of each location of a lens frame at the time of arrange | positioning a drive mechanism and a lens position detection means in the vicinity of a lens. It is explanatory drawing which shows the definition of the lens coordinate at the time of performing position control of a lens. It is explanatory drawing about a Lissajous waveform. It is explanatory drawing about the reference point scanning of a vertical displacement detection sensor, and a reference point prescription | regulation method. (A) shows the reference point scanning (no displacement) of the absolute type vertical displacement detection sensor. (B) shows reference point scanning (no misalignment) of the incremental type vertical displacement detection sensor. (C) shows reference point scanning (with position shift) of the absolute type vertical displacement detection sensor. (D) shows reference point scanning (with positional deviation) of the incremental type vertical displacement detection sensor. It is explanatory drawing which shows typically one surface of a lens unit. FIG. 21 is an enlarged view of the vicinity of the incremental lateral displacement detection sensor of FIG. 20. FIG. 21 is an explanatory diagram when the lens of FIG. 20 is driven at an angle β and tilt. FIG. 23 is an enlarged view of the vicinity of the incremental lateral displacement detection sensor of FIG. 22. It is explanatory drawing which shows the change before and behind the tilt drive of a lens mathematically. FIG. 25 is a flowchart of a sensor initialization routine for executing the adjustment process of each sensor described in FIG. 18 to FIG. 24. It is explanatory drawing which shows the sensor used for the lens position detection means which concerns on 2nd Embodiment of the holding | maintenance apparatus of the optical element of this invention. It is explanatory drawing about the reference point scanning of a vertical displacement detection sensor, and a reference point prescription | regulation method. (A) shows the reference point scanning (no misalignment) of the vertical origin detection sensor. (B) shows reference point scanning (no misalignment) of the incremental type vertical displacement detection sensor. (C) shows the reference point scanning (with positional deviation) of the vertical origin detection sensor. (D) shows reference point scanning (with positional deviation) of the incremental type vertical displacement detection sensor. It is explanatory drawing which shows the modification of an incremental displacement detection sensor. It is explanatory drawing which shows the projection optical system of the exposure apparatus provided with 2nd Embodiment of the holding device of the optical element of this invention. It is explanatory drawing which shows the detailed form of the mirror unit of FIG. It is explanatory drawing which shows the form which attached the mirror and the mirror frame to the mirror unit of FIG. It is sectional drawing which follows the CC line of FIG. It is explanatory drawing which shows the flowchart of the sensor initialization routine for performing the adjustment process of each sensor used for 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Exposure apparatus 2 Projection optical system 3 Exposure apparatus base 31 Body part 32 Lens mount 33 Damper means 4 Illumination unit 41 Laser light source unit 42 Exposure illumination means 43 Aberration measurement illumination means 44 Optical path switching mirror 5 Aberration measurement means 51 Interference Total 52 Half mirror 53 Reflecting mirror 54 Collimator lens 6 Lens control means 61 Lens CPU
62 Drive mechanism control means 7 Lens unit 71 Drive mechanism 71a Piezo driver 71b Piezo element 71c Direction changing part 711 Displacement taking member 711a Displacement taking link 711b1, 711b2 Piezo receiving link 711c1, 711c2 Connecting link 712 Direction changing member 712a Direction changing link 712b, 712e, 712f Support link 712c Horizontal link 712d Lens frame drive link 712e, 712f Support link 712d1, 712e1 Thin portion 712g Screw hole 713 Piezo actuator 713a Piezo receiving link 714 Piezo adjustment screw 72 Lens position detecting means 72A, 72B Mirror position detecting means 72a Laser light source 72b Collimator lens 72c Quadrant light receiving means 72d Sensor control circuit 72e Total reflection mirror 72f Non-polarized light Beam splitter 72g Polarizing beam splitter 72h Interference optical element 72j Light absorption means 72k Light projection means 72m Light reception means 72n Detection recess 72p Sensor control circuit 721 Absolute type vertical displacement detection sensor 721A Vertical origin detection sensor 722 Incremental type vertical displacement detection sensor 723 Incremental type Lateral displacement detection sensor 73 Fixed barrel 74 Lens frame 74A Mirror frame 74a Flange portion 75 Target mirror 76 Slit member 76a Slit 761 First target 761a Diffraction grating 761b Absolute code track 762 Second target 762a Diffraction grating 77A Vertical displacement detection sensor 77B Lateral displacement detection sensor 78 Pivot member 8 Body control means 81 Body CPU
82 Mount control means 83 Illumination control means 84 Reticle stage control means 85 Wafer stage control means 86 Interferometer control means B71 Compensation circuit block B72 Coordinate conversion block B73 Sensor output origin defining block B74 Coordinate conversion block B81 Lens target position calculation block B82 Position correction Block C Optical axis CV Conversion coefficient H1, H2, H3 Vertical displacement H11, H12, H13, H21, H22, H23 Elastic hinge LE Lens LE1-LE4 First to fourth lens group L0 Lens reference point M1-M3 Third mirror MU Mirror unit P Point P1 at the center of the cross section of the support link Pc Tilt center axis R Reticle R1 Reticle stage R2 Reticle mount Rm Diameter dimensions S1, S2, S3 Horizontal displacement VP1, VP2, VP3 Control command voltage signal W1 W2, W3 Drive displacement WF Wafer WF1 Wafer stage WF2 Spherical mirror Z0 Z-axis direction drive target value θX0, θY0 Tilt drive target value δZ0, δθX0, δθY0 Control residual Xm, Ym Lateral shift displacement Zm Z-axis direction translational displacement θXm, θYm Tilt displacement

Claims (5)

A holding device for an optical element, comprising: a drive unit that drives the optical element in an optical axis direction and a tilt direction with respect to the optical axis; and a position measurement unit that measures the position of the optical element,
The position measurement unit includes a first measurement unit and a second measurement unit for measuring a position of the optical element in an optical axis direction and an inclination direction with respect to the optical axis,
The optical element holding device, wherein the first measurement unit measures a relative position of the optical element, and the second measurement unit measures an absolute position of the optical element.   The optical element holding device according to claim 1, wherein a measurement reference of the first measurement unit is defined using an output of the second measurement unit.   3. The optical element holding device according to claim 1, wherein the position measurement unit includes a third measurement unit that measures a relative position of the optical element in a direction orthogonal to the optical axis direction. .   3. The optical element holding device according to claim 1, wherein a plurality of the position measurement units are arranged along a circumferential direction of the optical element. 4.   An exposure apparatus comprising the optical element holding device according to any one of claims 1 to 4.