JP4500729B2 - Surface shape measuring device - Google Patents

Description

  The present invention relates to a surface shape measuring apparatus for precisely measuring the shape of an optical element such as a lens or mirror having a free-form surface, or a mold for molding the optical element.

  FIG. 15 shows a surface shape measuring apparatus according to a conventional example disclosed in Patent Document 1, and has the same shape placed on the surface of the mold base 101a in close contact with the mold shaft portion 101b of the lens mold 101. The three positioning balls 111 are measured by scanning in the X and Y directions with the measuring probe 103. After calculating the vertex coordinates of each positioning sphere 111, the plane formed by each vertex coordinate and the center of the lens mold 101 are calculated based on the perpendicular bisector of each two vertex coordinates. After obtaining the measurement data of the shape of the mold transfer surface 101c, the inclination of the mold transfer surface 101c with respect to the previous plane and the center of the mold transfer surface 101c with respect to the center axis of the lens mold 101 are obtained. Calculate eccentricity.

  FIG. 16 is a flowchart for explaining the above steps. First, three positioning spheres are installed (S101), and the surface of the positioning sphere is measured using a measuring probe (S102). Next, the Z coordinate of the spherical surface with respect to the measured XY position is calculated (S103), the difference is calculated (S104), the root mean square (RMS) value is calculated (S105), and the RMS value is reduced. (S106), the process returns to S104 until the RMS value is sufficiently small (S107), and repeats (S107). When the RMS value is sufficiently small, the process returns to S102 and repeats until all three spheres are measured (S108). Next, vertex coordinates of each sphere are calculated (S109), and the plane including the three vertices and the center position of the mold are calculated (S110). Next, the surface of the mold transfer surface is measured using a measurement probe (S111). Next, the Z coordinate of the design shape with respect to the XY position of the measured point is calculated (S112), the difference is calculated (S113), the RMS value is calculated (S114), and coordinate conversion is performed so that the RMS value becomes small ( S115), the process returns to S113 until the RMS value becomes sufficiently small and repeats (S116). When the RMS value becomes sufficiently small, the shape of the mold transfer surface is calculated (S117), and the inclination and eccentricity of the mold transfer surface are calculated. Calculate (S118), (S119).

Patent Document 2 discloses a three-dimensional surface shape measuring apparatus that measures the shape of a free-form surface and a relative position with respect to the position-mark sphere with respect to an object to be measured having three or more position mark spheres and a free-form surface. A configuration is disclosed in which a probe that traces a surface to be measured and measures a shape and a non-contact sphere center position measuring means that measures the center position of a position mark sphere are fixed to a movable member that can be moved to the position. The non-contact sphere center position measuring means measures a position in two directions perpendicular to the optical axis for each position mark sphere by an optical system for measuring a light spot position, and a position in the optical axis direction by an optical system for detecting a focal position. Measure. That is, the center position of each position mark sphere is measured three-dimensionally by the non-contact sphere center position measuring means having an autofocus function, and the measurement shape of the free-form surface by the probe is corrected, so that the mounting posture of the object to be measured is corrected. It eliminates the effects of changes and ambient temperature.
JP 2001-133239 A JP 2004-77144 A

  However, the conventional example disclosed in Patent Document 1 has the following problems.

  (1) Measurement takes time. Since the surface of the three positioning spheres is scanned using the measurement probe and the center position of the sphere is calculated from the measured coordinates, the three spheres must also be measured using the measurement probe. , It takes time to measure. That is, in addition to the measurement time of the mold surface to be measured, it takes a measurement time of the surface of the crushed ball.

  In particular, the measurement of the sphere center position by such probe scanning takes time. In addition, since it is necessary to measure the position of a three-dimensional sphere, information on a plurality of cross-sections is required instead of one cross-section, and a long probe scanning distance is required. The wasted time required for acceleration / deceleration cannot be ignored.

  (2) The measurement result is affected by the mounting posture of the object to be measured. The position of the mold can be measured by measuring the vertices of the three positioning balls, but the position of the sphere apex is changed when the mounting posture of the mold that is the measurement object and the three balls is integrated. to be influenced. For example, as shown in FIG. 17, assuming that the radius of the positioning sphere 111 is R and the inclination angle is θ, the displacement of the sphere apex is Rsin (θ), so that R is 10 mm, and θ is 5 mm. It will also become. Therefore, it is a precondition that the three spheres are almost parallel on the XY plane of the apparatus, which is a great obstacle to implementation.

  (3) The influence of environmental temperature change is large. The linear thermal expansion coefficient of the object to be measured is not always small. This is especially true for molds. Even if the environmental temperature is controlled to be constant and thermal deformation is suppressed, the effect cannot be made zero. In the conventional example, since measurement time is required, the measurement error increases due to the influence of environmental temperature change.

  (4) Since the life of the three positioning balls is short, the measurement accuracy is deteriorated. In order to measure the positions of the three positioning spheres, the surface is traced with a contact type probe, so that it is conceivable that the surface is slightly worn.

  On the other hand, since the positioning sphere must be measured every time the shape of the object to be measured is measured, the number of measurements tends to increase. For this reason, wear cannot be ignored, and the measurement error of the positioning sphere gradually increases. Although it is conceivable to change the positioning sphere to a new one every time it is measured, it is not only uneconomical but also difficult to replace when used with adhesive fixation.

  (5) Unable to deal with errors that occur during positioning ball measurement. For example, if the positioning sphere is not attached due to human error, the measurement probe tries to contact the place where the sphere should be, but there is no sphere there, and the object to be measured, jig, etc. There may be other things. In such a case, there is a possibility of damaging the probe and the object to be measured.

  In other words, it is only possible to measure whether or not the positioning sphere, that is, the spherical surface is properly set there, using a probe. Therefore, since the probe must be brought close to the object to be measured and the surface is actually traced, there is an inevitable risk of the probe colliding with the above-described problem.

  Moreover, the surface shape measuring apparatus disclosed in Patent Document 2 has the following unsolved problems.

  (1) Since an autofocus function for detecting a focus position for measuring the position of the position mark sphere in the optical axis direction is required, the optical system becomes complicated. And the sensitivity of the autofocus used for detecting the focal position in the optical axis direction is generally low when the NA (numerical aperture) of the objective lens is small, and the measurement accuracy is deteriorated. This is because a lens having a small NA has a wide focal depth, and a lens having a small NA has a low sensitivity even if the focal position is slightly shifted.

  (2) The position of the optical surface on the semiconductor device cannot be measured. That is, in some micromachine applications using light, there are those in which a fine lens or mirror is formed on a semiconductor device. For example, an optical switch that moves a minute mirror or lens. It is necessary to measure the position of such a fine optical element with reference to the semiconductor device. However, in the three-dimensional measurement according to the conventional example, a cross-shaped position mark used as an alignment mark of the semiconductor device is used. difficult.

  The present invention has been made in view of the above-mentioned unsolved problems of the prior art, and detects a position mark serving as a reference for shape measurement using a simple and high-sensitivity optical system, and attaches an object to be measured. It is intended to provide a highly reliable surface shape measuring device that has little effect on measurement accuracy even if the posture changes or the environmental temperature changes. .

In order to achieve the above object, a surface shape measuring apparatus of the present invention is a surface shape measuring apparatus for measuring the shape of an object to be measured using three position mark spheres as a position reference, and is held by a holding means. A probe that measures the surface of the object to be measured, a moving body that moves the probe in three dimensions with respect to the object to be measured, and a light beam from the light source that is focused on each position mark sphere and reflected from the position mark sphere. to detect light, Starring based on different two sets of optical position detecting means which is deployed in the optical axis direction, position information of the reflected light detected Ri by the prior SL 2 set of optical position detection means with each other It possesses a calculating means for performing calculation, and based on the two sets of position information of the reflected light detected by the optical position detector, a straight line passing through the center of the position mark ball parallel to the optical axis direction Obtained by the computing means, the two obtained By determining the center of the vertical line at the position where the line is closest, and calculates the center position of the position mark sphere.

  Compared with the case where the surface of the position mark serving as a reference for shape measurement is traced with a probe, the measurement time is significantly reduced. By shortening the measurement time, the environmental temperature change within the measurement time is also reduced, and an improvement in measurement accuracy can be expected. In addition, since the position of the position mark can be measured in a non-contact manner, it is possible to avoid an increase in measurement error due to wear and a shortening of the life.

  Further, two sets of two-dimensional position information obtained from different optical axis directions are processed by two detection optical systems using a position sensor or the like that receives light reflected from the surface of each position mark. Since the position of the position mark is obtained three-dimensionally, the position mark detection optical system does not require an autofocus function. That is, it does not require an autofocus function as in the case where the center position of the position mark sphere is directly measured three-dimensionally. Therefore, it is possible to use a detection optical system with extremely high sensitivity and low cost.

  Furthermore, there is an advantage that the present invention can be easily applied not only to the position mark sphere but also to a planar position mark on a semiconductor device.

  Embodiments of the present invention will be described with reference to the drawings.

1 to 7 show a first embodiment. As shown in FIG. 1, the surface shape measuring apparatus according to this embodiment attaches a workpiece W ₁ to a jig 2 that is a holding means on a base 1. Are three position mark balls 3 is attached to the jig 2. The jig 2 fixes each position mark ball 3 and holds the object to be measured W ₁ in a detachable manner. An X slider 5, a Y slider 6, and a Z slider 7 are provided on a base 4 integrated with the base 1 so as to be movable in three axial directions of X, Y, and Z, respectively, and are connected to a control device 8. X1 laser length measuring machine 9a is a probe 10 in the Z-slider 7 as a mobile body moves in three dimensions with respect to the measured object _{W 1} for measuring the position, X2 laser measuring machine 9b, Z1 laser measuring machine 9c, Although not shown, a Y1 laser length measuring device and a Y2 laser length measuring device are provided in a direction perpendicular to the paper surface. The Z slider 7 is provided with a contact type probe 10 having a tip sphere 10a. This probe 10 has a known parallel leaf spring structure, for example, and outputs a signal according to the displacement of the leaf spring. The signal is connected to the control device 8 via the probe control amplifier 11. Further, the Z slider 7 is provided with light spot position measuring optical systems 12a and 12b which are two sets of optical position detecting means ( detecting optical systems ) for measuring the center position of each position mark sphere 3 from different optical axis directions. Are connected to the control device 8 via the position measuring amplifiers 13a and 13b, respectively.

  The base 1 is provided with a column 14 that supports a position measurement standard, and is provided with a reference mirror 15 that is fixed to the column 14 and serves as a reference in the X direction. The reference mirror 15 is used as a target for the X1 and X2 laser length measuring instruments 9a and 9b. Further, a reference mirror 16 that is fixed to the column 14 and serves as a reference in the Z direction is provided. This reference mirror 16 is used as a target of the Z1 laser length measuring instrument 9c. Similarly, a reference mirror fixed to the column 14 and serving as a reference in the Y direction is provided. This reference mirror is used as a target of the Y1, Y2 laser length measuring device.

  The output signals of these laser length measuring instruments, i.e., the distance from the reference mirror, are connected to a data processing device (not shown), and data processing is performed as position information on the surface to be measured.

  Next, one light spot position measuring optical system 12a will be described with reference to FIG. The light spot position measuring optical system 12a divides the light emitted from the semiconductor laser 17a into two directions by the half mirror 18a. The light that has passed through the half mirror 18a strikes the shielding plate 19a and is absorbed. On the other hand, the light reflected by the half mirror 18 a is reflected by the mirror 20 a and the optical axis is bent in the direction of the position mark sphere 3. This light becomes a light beam Ba which is a focused light beam by passing through the objective lens 21a. At this time, the positions of the mirror 20a and the objective lens 21a are adjusted in advance so that the focal position of the light beam Ba is directly below the contact probe 10 having the tip sphere 9. Although it is difficult to make this adjustment very precisely, the adjustment error at this time can be corrected. This method will be described later using a flowchart.

  As shown in FIG. 2, when the focal position of the light beam Ba and the center position of the position mark sphere 3 substantially overlap, the light beam Ba is reflected by the surface of the position mark sphere 3 and travels backward in the optical path. A position sensor 23a that detects a two-dimensional position of a point image that is an optical image (light spot) is provided at a position where the light beam Ba passes through the half mirror 18a and is focused. On the other hand, the light reflected by the half mirror 18a returns to the direction of the semiconductor laser 17a as unnecessary light, but a small hole is formed in the vicinity of the exit of the semiconductor laser 17a so as not to be scattered by surrounding articles and become unnecessary stray light. It absorbs by making contact with the opened second shielding plate 24a.

  The four signals from the position sensor 23a are passed through the two subtraction circuits 25a and 26a of the arithmetic means provided in the control device 8, and two difference signals are obtained. This signal represents the position of the light spot. However, this is affected by fluctuations in the output of the semiconductor laser 17a as the light source. That is, if the output of the semiconductor laser 17a is multiplied by α, the four outputs of the position sensor 23a are also uniformly multiplied by α, and naturally the difference signal is also multiplied by α. Therefore, all the four signals from the position sensor 23a are added by the adder circuit 27a, and the difference signal is divided by the divider circuits 28a and 29a. Since the difference signal multiplied by α is divided by α, the influence of the output fluctuation of the semiconductor laser 17a can be eliminated. The light spot position signals representing the two-dimensional position of the position mark sphere 3 thus obtained are denoted by φ1a and φ2a.

  Further, the output of the adder circuit 27a is connected to the error determination circuit 30a. When the output of the adding circuit 27a is low, there is a possibility that light does not hit the position sensor 23a properly. Therefore, it can be determined whether or not an error state has occurred by determining whether or not this signal is below a predetermined value. The error determination circuit 30a compares the output of the adder circuit 27a with a preset signal level, and outputs an error signal if it is low. This error signal is assumed to be e12a.

Next, it will be described with reference to FIGS. 3 and 4 that the signals φ1a and φ2a represent the position in two directions perpendicular to the optical axis Va of the light beam Ba as the two-dimensional position information of the center of the position mark sphere 3. FIG. The light emitted from the semiconductor laser 17a is condensed by the action of the objective lens 21a located at the distance D1, and focused at E1 at the distance D2. At this time, from the lens formula, if the focal length of the objective lens 21a is f1, the following formula is established.
1 / f1 = 1 / D1 + 1 / D2

This image E1 is reflected by the spherical surface of the position mark sphere 3 at the distance D3. If the radius of the position mark sphere 3 is R, the spherical surface may be considered to have the same effect as a concave lens having a focal length of R / 2, so the point image is mapped to a position E2 that is separated by a distance D4. That is, the following expression is established.
2 / R = 1 / D3 + 1 / D4

  Now, since the case where the focal position E1 and the center position of the position mark sphere 3 substantially coincide is considered, D3 and D4 are substantially the same distance.

This point image again passes through the objective lens 21a at the distance D5 and forms a focal point E3 on the position sensor 23a at the distance D6. That is, the following expression is established.
1 / f1 = 1 / D5 + 1 / D6

  Consider the case where the spherical surface of the position mark sphere 3 is displaced by δ in the direction perpendicular to the optical axis Va in such an optical system arrangement. As shown in FIG. 4, the state at this time is the same as the point image E1 moves δ for the position mark sphere 3, and therefore the position E2 moves by δ * (D4 / D3 + 1) with respect to the optical axis Va.

Furthermore, since the objective lens 21a maps to the focal point E3, the movement δ1 of the focal point E3 is expressed by the following equation.
δ1 = δ * (D4 / D3 + 1) * D6 / D5

  For example, if D3 = D4 and D6 = D5, then 2δ. That is, the two-dimensional movement of the position mark sphere 3 in a plane perpendicular to the optical axis Va can be detected by doubling. The signals are φ1 and φ2.

  In FIG. 4, a position shifted from the focal point E3, which is the image forming position, by a distance D7 is considered. Since the position is deviated from the imaging position, the spot size is large at this position. However, the deviation δ2 of the center position is larger than δ1. That is, the sensitivity can be increased.

  Since the other light spot position measuring optical system 12b has the same configuration, the description thereof is omitted. As two sets of two-dimensional position information by the two light spot position measuring optical systems 12a and 12b, signals corresponding to movements perpendicular to the optical axis Va are obtained according to movements perpendicular to the optical axes Va and φ1a and φ2a. These signals are φ1b and φ2b.

  A method for determining the three-dimensional position of the position mark sphere 3 from these four signals, φ1a, φ2a, φ1b, and φ2b will be described with reference to FIG. The signals φ1a and φ2a represent positions where the position mark sphere 3 is separated from the optical axis Va in a direction perpendicular to the optical axis Va. Therefore, the signals φ1a and φ2a are parallel to the optical axis Va and ) Through the straight line DVa. Similarly, since the signals φ1b and φ2b represent positions where the position mark sphere 3 is separated from the optical axis Vb in a direction perpendicular to the optical axis Vb, the signals φ1b and φ2b are parallel to the optical axis Vb and pass through the spherical center. DVb can be determined. If measurement can be performed without error, the two straight lines DVa and DVb should intersect. This is because both straight lines should pass through the center of the same position mark sphere 3.

  When the four signals φ1a, φ2a, φ1b, and φ2b are zero, the straight lines DVa and DVb indicate that they coincide with the optical axes Va and Vb. Therefore, the center of the position mark sphere 3 is the optical axes Va and Vb. Indicates that it is coincident with the intersection.

  However, it is natural to consider that the two straight lines DVa and DVb do not intersect due to the influence of electrical noise mixed in the four signals φ1a, φ2a, φ1b, and φ2b and the effects of assembly errors. Therefore, as shown in FIG. 5, the midpoint c of the vertical line drawn to the position where the two straight lines DVa and DVb are closest to each other is output as the center position (three-dimensional position) of the position mark sphere 3.

  Further, a positional deviation A that is a distance between the two straight lines DVa and DVb is also calculated. The positional deviation A is a distance that is essentially zero if the above-described error is small. Therefore, it represents an error included in the measurement result. If the amount of this error, that is, the positional deviation A is larger than a predetermined value, it is determined as an error.

  The laser length measuring devices 9a to 9c and the like of the apparatus of FIG. 1 measure and correct the position and orientation of the Z slider 7, and the method will be described with reference to FIG. For the three reference mirrors fixed to the base 1 and orthogonally arranged as described above, the lengths measured by each laser length measuring device are X1, X2, Y1, Y2, and Z1. Z1 measures the distance between the Z reference mirror 16 and the probe 10. Also, the difference between the mounting height of the laser length measuring device that measures X1 and Y1 and the mounting height of the laser length measuring device that measures X2 and Y2 is L2, and the difference from that to the mounting height of the probe 10 is L3. The difference from there to the center position of the tip sphere 10a of the probe 10 is L4, and the distance from there to the focal position of the light beams Ba and Bb is L5. These distances L2 to L5 are lengths determined by the component dimensions of the measuring apparatus. L1 is obtained by subtracting L2 and L3 from the measured value Z1.

In this configuration, the measurement coordinates of the center position of the tip sphere 10a of the probe 10 can be calculated by the following equations (1a) to (1c).
Probe sphere X coordinate = −X1− (X2−X1) * (L2 + L3 + L4) / L2
... (1a)
Probe sphere Y coordinate = −Y1− (Y2−Y1) * (L2 + L3 + L4) / L2
... (1b)
Probe sphere Z coordinate = −Z1−L4 (1c)

Similarly, the measurement coordinates of the focal positions of the light beams Ba and Bb can be calculated by the following equations (2a) to (2c).
Focal position X coordinate = −X1− (X2−X1) * (L2 + L3 + L4 + L5) / L2
... (2a)
Focal position Y coordinate = -Y1- (Y2-Y1) * (L2 + L3 + L4 + L5) / L2
... (2b)
Focus position Z coordinate = −Z1−L4−L5 (2c)

  This coordinate position is not affected by the attitude error of the Z slider because it is measured with the reference mirror of the three surfaces as the position reference.

Next, the operation of the surface shape measuring apparatus of FIG. 1 will be described based on the flowchart shown in FIG. First, the workpiece W ₁ is detachably attached to the jig 2 to which the three position mark balls 3 are attached. Next, it is determined light spot position measurement optical system 12a, and 12b, and whether there is a need to position the probe 10 for measuring the free-form surface of the workpiece W ₁ (step S1). This is necessary when, for example, the device is first manufactured or when the probe 10 is replaced and the probe position information is lost. Further, as described above, the measurement positions of the light spot position measurement optical systems 12a and 12b, that is, the positions of the light beams Ba and Bb are roughly aligned so that they are directly below the probe 10, but this is very precisely performed. It is difficult to adjust. In this sense, first, it is necessary to align the light spot position measuring optical systems 12a and 12b with the probe 10. However, once alignment is performed, alignment is not necessary.

  When positioning, first, a point cloud when the surface of the position mark sphere 3 is traced by the probe 10 is measured (step S2). The center position is calculated from the point group on the surface using the least square method or the like (step S3). The position obtained as a result of the calculation is defined as Pa.

  Here, it should be noted that this point Pa is not the apex of the sphere as in the conventional example. Unlike the apex position, the center position (sphere center position) of the sphere should be the same point no matter what direction it is measured. In other words, measuring the position of the position mark sphere from any direction gives the same result.

  The two-dimensional position of the position mark sphere 3 in a plane perpendicular to the optical axes Va and Vb is measured by the optical spot position measuring optical systems 12a and 12b. The X, Y, and Z sliders 5, 6, and 7 are moved using the control device 8 so that the position Pa of the position mark sphere measured by the probe 10 and the focal positions of the light beams Ba and Bb are substantially matched (step S4). At this position, it is determined whether any one of the error signals e12a and e12b described in FIG. 2 has been output (step S16). If an error is output, it means that the position mark sphere 3 is not at the assumed position or the semiconductor laser 17a or the like has failed. An error is displayed (step S17), and the operation stops. On the other hand, if there is no error, the signals φ1a, φ2a, φ1b, and φ2b can be considered as correct measurement values.

  These four signals φ1a, φ2a, φ1b, and φ2b represent deviations in the vertical direction of the light beams Ba and Bb with respect to the optical axes Va and Vb. The position data c is calculated, and the X, Y, Z sliders 5, 6, and 7 are finely adjusted using the control device 8 so that the position data c is zero, that is, coincides with the intersection of the two optical axes Va and Vb. (Step S5). In other words, each slider is finely adjusted so that the signals φ1a, φ2a, φ1b, and φ2b approach zero.

  After fine adjustment, the position X of the position mark sphere 3 is determined from the positions and postures of the X, Y, and Z sliders 5 to 7 by the method described with reference to FIG. 6 and the equations (2a), (2b), and (2c). , Y and Z positions (three-dimensional positions) are obtained, and the positional deviation A of the position mark sphere 3 is calculated by the method described with reference to FIG. At this time, the four signals φ1a, φ2a, φ1b, and φ2b should be sufficiently small by fine adjustment of the X, Y, and Z sliders 5 to 7, but adjustment errors that cannot be adjusted still remain. The position data c is calculated from the four signals φ1a, φ2a, φ1b, and φ2b, the adjustment residual is converted into an error in the X, Y, and Z directions, and the X, Y, and Z positions of the center of the position mark sphere 3 are corrected. If it does, it is still precise. The X, Y, and Z positions are set as Pb (step S6).

  The positional deviation A becomes large when the optical axes Va and Vb of the two light beams Ba and Bb do not coincide with each other or when the error mixed in the four electrical signals φ1a, φ2a, φ1b, and φ2b is large. Therefore, if the positional deviation A is larger than a predetermined value, it is determined as an error (step S22), an error is displayed (step S23), and the process is stopped.

  Next, a position difference d = Pa−Pb is calculated and used as a position correction amount. This represents the position correction amount between the probe 10 that measures the free-form surface and the light spot position measurement optical systems 12a and 12b that measure the position mark sphere 3 (step S7).

  Next, all the position mark spheres 3 are measured by the light spot position measuring optical systems 12a and 12b. Each measurement is the same as the procedure described above. A case where the nth position mark sphere is measured will be described. First, the focal positions of the light beams Ba and Bb are moved to the center position of the nth position mark sphere (step S8). Since the approximate position is known from the design drawing of the jig 2, the drawing for attaching the jig 2 to the measuring device, etc., the X, Y, Z sliders 5 to 7 can be moved to that position. At this position, it is determined whether any one of the error signals e12a and e12b described in FIG. 2 is output (step S18). If an error is output, it means that there has been some trouble such as the position mark sphere 3 not being in the assumed position, so that an error is stopped and an error is displayed on the display device such as a computer screen (step S19). ),Stop. On the other hand, if there is no error, the signals φ1a, φ2a, φ1b, and φ2b can be considered as correct measurement values.

  Next, since the four signals φ1a, φ2a, φ1b, and φ2b represent deviations in the vertical direction with respect to the light beams Ba and Bb, as described above, the position data c of the position mark sphere 3 is obtained by the method described with reference to FIG. And the X, Y and Z sliders 5 to 7 are finely adjusted using the control device 8 so that the position data is zero, that is, coincides with the intersection of the two optical axes Va and Vb (step S9).

  The center position (X, Y, Z position) of the nth position mark sphere at this time is calculated and set as Pn (step S10). At this time, the position correction amount d obtained earlier is added. As a result, the position Pn is converted to the position as measured by the contact probe 10. In addition, compared with the case of measuring with the probe 10, it is not necessary to trace the surface, so the measurement time is considerably short.

  Similarly to the previous case, when the positional deviation A is larger than a predetermined value, it is determined as an error (step S20), an error is displayed (step S21), and the process is stopped.

The operation from step S8 is repeated until all the position mark spheres 3 are measured, and when it is finished, the operation proceeds to the next (step S11). One orthogonal coordinate system is determined from the three ball center positions (step S12). There are several methods. For example, the Z axis is taken in a direction perpendicular to the plane defined by three points (measured sphere positions), the line connecting the two points is taken as the X axis, and the remaining Y The axis can be defined as an axis perpendicular to the X and Z axes. Since the orthogonal coordinate system defined in this way is fixed to three spheres, it is called an object coordinate system. When viewed from this measured object coordinate system, the measured object W ₁ including the free-form surface to be measured is always at the same position. This is because the object to be measured W ₁ is fixed to the jig 2 fixed to the position mark sphere 3 as described at the beginning.

Next, the free-form surface, which is the surface of the object to be measured W ₁ , is traced by the probe 10 to measure a point group on the free-form surface (step S13). All the point groups are coordinate-converted into the measured object coordinate system (step S14). Since the point group obtained in this way uses the three position mark spheres 3 as the position reference, the same point group can be obtained no matter where the object to be measured W ₁ is installed or in any posture. Next, a difference from the design shape, that is, an error shape is calculated using the least square method (step S15).

  According to the method described above, the measurement time can be shortened. This is because the measurement time for three balls is shortened. When tracing the surface of a sphere using a probe as in the conventional example, since three-dimensional positional information is required, a plurality of cross-sectional traces are required, and thus a long trace length is required. In addition, when tracing with repeated reciprocation, the time required for acceleration / deceleration cannot be ignored. On the other hand, in the method according to the present embodiment, the X, Y, and Z axes are moved so that the position mark sphere is approximately at the center position of the optical axis, and then fine adjustment is performed according to the signals φ1a, φ2a, φ1b, and φ2b. Can determine the position of the ball. Since it is not necessary to trace the surface of the sphere, the time can be significantly reduced.

  Since the center position of the sphere is measured rather than the position of the top of the sphere, it is not affected by the position and orientation of the object to be measured. This is because the sphere is used in such a manner that the same point is pointed regardless of the center position measured from any direction.

  In addition, since it is not necessary to obtain an autofocus signal, the NA (numerical aperture) of the lens may be small. If the lens NA is small, the distance between the lens and the position mark sphere can be increased, so that the arrangement of the optical system is simplified and the degree of design freedom can be greatly improved.

  Further, since the measurement time of the sphere is shortened, the entire measurement time is also shortened. Since errors such as temperature deformation that changes with time can be relaxed, measurement accuracy can be improved as a result.

  In this embodiment, a position sensor is used for each detection optical system. However, the two-dimensional point position of the light spot may be detected using an image pickup device such as a CCD.

  Further, although the probe for measuring a free-form surface has been described as a contact type probe, the same applies to a non-contact type probe.

  As the light source, for example, a light emitting element such as an LED may be used instead of the semiconductor laser. However, the point closer to the point light source is preferable because the spot at the focal position becomes smaller and the light spot position can be easily detected.

  The mirror is used to change the direction of the optical axis and obtain an oblique light beam, but this may be omitted, or instead of bending the optical axis once by the mirror, a mirror is added and the optical axis is changed. It may be bent. By adding a mirror in this way, the degree of freedom in designing the arrangement of optical elements can be improved.

  This example has the following effects.

  (1) Since it is not necessary to trace the surface of the position mark sphere as in the conventional example, it is possible to expect a significant reduction in measurement time.

  (2) Since the center of the sphere is measured instead of the top of the sphere, the same measurement result can be obtained regardless of the position and orientation of the object to be measured. This is because the center of the sphere is the same point when measured from any direction.

  (3) Since the measurement time can be shortened, the environmental temperature change within the measurement time is also reduced, and the measurement accuracy is improved.

  (4) Since the center of the position mark sphere can be measured in a non-contact manner, there is no need to worry about wear of the sphere, and there is no increase in error due to wear, so that measurement accuracy can be improved.

  (5) Since it is non-contact, the life of the position mark sphere can be extended.

  (6) Since there is no contact, there is no increase in measurement error due to wear, so accuracy is improved.

  (7) By setting the measurement position of the position mark sphere on the extension line of the probe axis, the influence of the position and posture error of the moving member accompanying the probe movement can be minimized as in the case of the probe.

  (8) Since it is not necessary to optically detect the position in the optical axis direction, a simple optical system in which the autofocus function is omitted can be used.

  (9) The reliability of the measurement can be improved by monitoring the optical axis deviation and interrupting the measurement if it is large.

  (10) Since it is not necessary to use autofocus, the NA of the lens of the detection optical system may be small. As a result, the distance between the lens and the object to be measured can be increased.

  (11) It is possible to automatically detect an abnormal state such as when the position mark sphere is dropped and to stop the apparatus.

  (12) It is possible to automatically detect an error and stop the apparatus safely even for an error that has been difficult to deal with in the past, such as a wrong part being attached instead of the position mark sphere.

  (13) By detecting an abnormality quickly, the influence of the error can be minimized.

  (14) Since an error can be detected and stopped safely, it is also effective in preventing accidents. By preventing accidents, the operating rate of production equipment is improved and production costs are reduced.

  FIG. 8 shows the second embodiment, which is different from the first embodiment only in the light source portion and the half mirror portion. Laser light emitted from the semiconductor laser 37 that is a common light source is split into two by an optical fiber 41a and an optical fiber 41b that are splitting means. The optical fiber 41a is connected to the optical fiber emitting end 40a, from which laser light is emitted. Since the same applies to the optical fiber 41b, the description thereof is omitted.

  The light emitted from the optical fiber output end 40a enters the polarization beam splitter 42a. The P-polarized light component is transmitted by the polarization beam splitter 42a and is absorbed by the shielding plate 19a. On the other hand, the S-polarized light component is reflected by the polarization beam splitter 42a, reflected by the mirror 20a, and the optical axis is bent in the direction of the position mark sphere 3. The light is converted into circularly polarized light by the quarter-wave plate 43a, and this light becomes a light beam Ba by passing through the objective lens 21a. At this time, the positions of the mirror 20a and the objective lens 21a are adjusted in advance so that the focal position of the light beam Ba is directly below the contact probe 10 having the tip sphere 10a. Although it is difficult to make this adjustment very precisely, the adjustment error at this time can be corrected.

  As shown in FIG. 8, when the focal positions of the light beams Ba and Bb and the center position of the position mark sphere 3 substantially overlap, the light beam Ba is reflected by the surface of the position mark sphere 3 and returns to the original optical path. Go. When passing through the objective lens 21a and passing through the quarter-wave plate 43a, it is converted again into linearly polarized light, but the direction of polarized light is changed, and P-polarized light is obtained. This light is reflected by the mirror 20a and enters the polarization beam splitter 42a. This P-polarized light passes through this optical element and enters the position sensor 23a. At this time, there is no light reflected by the polarization beam splitter 42a, that is, no S-polarized component. For this reason, there is little loss of light quantity. Further, the second shielding plate for absorbing the reflected unnecessary light is not necessary.

  In the case where only the half mirror is used as in the first embodiment, since the light passes through the half mirror twice, only at least a quarter of the light reaches the position sensor. On the other hand, in this embodiment, the light incident on the position sensor can be doubled by using the polarization beam splitter. Further, although the semiconductor laser generates heat, the semiconductor laser can be freely arranged by using an optical fiber. Furthermore, since it is divided into two using an optical fiber distributor, only one semiconductor laser is required. Since the heating element at the measurement part leads to a large measurement error for the measurement apparatus, the disposition of the semiconductor laser is a great merit in improving accuracy.

  In this embodiment, in addition to the effects of the first embodiment, there are the following effects.

  (1) By branching one light source, the number of light sources can be reduced and the apparatus cost can be reduced.

  (2) Since light is guided by an optical fiber, it is not necessary to take a component space around the light source, for example, a semiconductor laser or a component space of its drive circuit. As a result, it can be configured in a small size.

  (3) Since the light source can be arranged far away from the measurement region by using the optical fiber, there is no fear of temperature rise due to heat generation of the light source. As a result, measurement errors due to temperature rise can be reduced and measurement accuracy can be improved.

  FIG. 9 shows a third embodiment, which differs from the first embodiment in that the focal positions of the light beams Ba and Bb are deviated from the center of the probe 10. Since other configurations and operations are the same as those of the first embodiment, the description thereof is omitted.

  This embodiment has the following effects in addition to the first embodiment.

  (1) Since the optical system can be arranged without being constrained by interference with the contact type probe, it can be made compact. In the first embodiment, it is necessary to increase the focal length of the objective lens so as not to interfere with the probe. However, in this embodiment, a small optical system can be configured using a lens having a short focal length.

  (2) Since there is no need to dispose an obstacle such as a prism in the vicinity of the contact type probe, the possibility of collision between the object to be measured and the measuring device is reduced. Therefore, design restrictions on the device under test can be relaxed.

  On the other hand, the slider position / orientation correction described with reference to FIG. 6 does not function sufficiently. This is because the posture correction is possible only on the probe axis in FIG. When deviating from the axis, the influence of the attitude error around the vertical Z-axis, that is, the so-called yawing error, becomes larger among the slider attitude errors. Therefore, in this embodiment, X, Y, and Z sliders with small yawing errors are required.

  10 and 11 show Example 4, FIG. 10 is a top view thereof, and FIG. 11 is a bird's eye view. In the first embodiment, the two light beams Ba and Bb are opposed to each other and fixed to the Z slider 7, but in this embodiment, the two light beams Ba and Bb are bent and arranged as shown in the drawing. The position sensors 23a and 23b that detect the light spot position output displacement in a direction perpendicular to the light beams Ba and Bb. In other words, there is no sensitivity in the directions of the light beams Ba and Bb, and there is maximum sensitivity in the direction perpendicular to the directions. Therefore, if the two light beams Ba and Bb are arranged so as to cross each other at right angles, the sensitivity of the other can be maximized even in the direction in which the sensitivity of one of them disappears. Therefore, there is no place where the position detection sensitivity becomes low in all directions, and the measurement accuracy is improved.

  This condition can be expressed as an arrangement in which the inner product of the vectors in the directions of the optical axes Va and Vb of the two light beams Ba and Bb becomes zero.

  Further, as shown in the top view of FIG. 10, the Z-axis for holding the probe 10 and the position sensors 23a and 23b can be made smaller than in the case where the probe 10 and the contact type probe 10 are opposed to each other.

  FIG. 12 shows a fifth embodiment. This is a configuration that is easy to assemble by integrating optical components with respect to the first embodiment. Two half mirrors 18a and 18b are bonded and fixed, and two prisms 50a and 50b are bonded to form an integral structure. For bonding the optical elements, for example, an ultraviolet curable resin is used. Since the operation is the same as that of the first embodiment except for the arrangement of the optical system, the description is omitted.

  In this embodiment, two position sensors 23a and 23b can be arranged side by side. Therefore, since this electric component can be mounted on one circuit board, assembly is easy. Also, the two semiconductor lasers 17a and 17b can be arranged side by side. Therefore, since this electric component can be mounted on one circuit board, assembly is easy. In addition, the configuration in which the optical components are directly bonded and fixed leads to the omission of the components for fixing the optical components, so that the number of components can be reduced as a whole and the assembly can be easily performed.

  FIG. 13 shows a sixth embodiment, which is obtained by replacing the position sensor of each light spot position measuring optical system with a camera which is an imaging means, as compared with the first embodiment.

  Two sets of cameras 60a and 60b are provided for the position mark sphere 3, and are connected to the image processing devices 61a and 61b, respectively. The image processing devices 61a and 61b can automatically calculate from the image of the position mark sphere 3 where the position mark is captured in the range captured by the cameras 60a and 60b. A technique for measuring the position of an image by this image processing is known and applied to a non-contact type three-dimensional coordinate measuring apparatus and the like. The image processing device 61a outputs signals φ1a and φ2a, which are two-dimensional position information obtained from the image of the position mark sphere 3, and an error signal e12a when image processing has failed. Similarly, the image processing device 61b outputs signals φ1b and φ2b which are two-dimensional position information of the position mark sphere 3 and an error signal e12b when image processing has failed. These pieces of information are the same as the information output by the optical system described in the first embodiment, and the subsequent processing is also the same, and thus the description thereof is omitted.

  In this embodiment, since a camera is used instead of the position sensor, a position mark other than the position mark sphere can also be used. For example, as shown in FIG. 14, a cross-shaped position mark 63 for alignment provided on a semiconductor device which is an object to be measured can be used. The cross-shaped position mark 63 can be easily manufactured on the substrate using a semiconductor lithography technique.

  That is, the free curved surface of the optical element is measured with a contact type probe, and the position mark on the semiconductor device is measured with two sets of cameras. Therefore, the shape of the fine optical element formed on the semiconductor device can be measured with reference to the cross-shaped position mark which is a position reference for alignment of the semiconductor device.

1 is a schematic diagram showing a surface shape measuring apparatus according to Example 1. FIG. It is a figure explaining the optical system which measures the position mark sphere of the apparatus of FIG. It is a schematic diagram which shows the optical path of the optical system of FIG. It is a schematic diagram which shows the shift | offset | difference of the light spot by the optical system of FIG. It is a figure explaining the method of calculating the three-dimensional position of a position mark sphere from four signals. It is a figure explaining the coordinate position measuring method by Example 1. FIG. 3 is a flowchart illustrating an operation according to the first embodiment. FIG. 6 is a diagram illustrating an optical system of Example 2. 6 is a schematic diagram showing Example 3. FIG. 10 is a top view showing Example 4. FIG. 10 is a bird's eye view showing Example 4. FIG. FIG. 10 is a diagram for explaining a fifth embodiment. FIG. 10 is a diagram illustrating Example 6. It is a figure explaining a cross-shaped position mark. It is a figure explaining a prior art example. It is a flowchart explaining the operation | movement of a prior art example. It is a figure explaining the problem of measuring the vertex position of a position mark ball | bowl by a prior art example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Base 2 Jig 3 Position mark ball | bowl 5 X slider 6 Y slider 7 Z slider 8 Control apparatus 9a, 9b, 9c Laser length measuring device 10 Probe 10a Tip ball | bowl 12a, 12b Optical spot position measurement optical system 17a, 17b, 37 Semiconductor laser 18a, 18b Half mirror 21a, 21b Objective lens 23a, 23b Position sensor 41a, 41b Optical fiber 42a Polarizing beam splitter 60a, 60b Camera 61a, 61b Image processing device 63 Position mark

Claims (5)

A surface shape measuring device for measuring the shape of an object to be measured using three position mark spheres as a position reference,
A probe for measuring the surface of the object to be measured held by the holding means;
A moving body that moves the probe in three dimensions relative to the object to be measured;
Two sets of optical position detection means arranged in different optical axis directions in order to focus the light from the light source on each position mark sphere and detect the reflected light from the position mark sphere;
Have a, arithmetic means for performing arithmetic on the basis of the position information before Symbol two sets by the optical position detector of Ri detected reflected light,
Based on the position information of the reflected light detected by the two sets of optical position detection means, straight lines passing through the center of the position mark sphere parallel to the optical axis direction are obtained by the calculation means, and the two straight lines thus obtained are obtained. A surface shape measuring apparatus , wherein the center position of the position mark sphere is calculated by obtaining the center of a vertical line at a position closest to .   2. The surface shape measuring apparatus according to claim 1, wherein each optical position detection means is a detection optical system that detects a two-dimensional position of a point image by reflected light of the position mark sphere by a position sensor.   2. The surface shape measuring apparatus according to claim 1, wherein each of the optical position detection means is a detection optical system that picks up an image of the position mark sphere with an image pickup means.   4. The surface shape measuring apparatus according to claim 1, wherein an error signal is generated based on outputs of the two sets of optical position detecting means.   5. The two sets of optical position detection means include a common light source and a dividing means for dividing light generated from the light source into the two sets of optical position detection means. The surface shape measuring apparatus according to claim 1.

Images