JP2009052989A - Optical measuring apparatus for microscopic object - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an optical measuring apparatus for microscopic objects which can perform accurate measurement practically by absorbing the measurement errors caused by noise. <P>SOLUTION: The optical measuring apparatus for a microscopic object comprises a multi-wavelength laser light generating means and a photographic means for photographing a plurality of interference images by observation light and object light, while changing the optical path length of an optical system. The optical measuring apparatus measures the phase of a part in which the brightness of each point of an interference image for each wavelength of the plural-wavelength laser light changes and measures the size of an object under measurement by a combination of phase values. By having the measured phase value divided for each predetermined measurement section of the object under measurement and analyzed by a conversion table in which the combination of sections are represented by section numbers, the sizes exceeding the wavelength of laser light used are measured; combination of the section numbers of the conversion table is set rough, in advance; and by having the combination of the vacant section number interpolated by a neighboring combination number, the errors in the measured phase value are absorbed and corrected. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a configuration of an optical measuring device for a minute object that optically measures the size of the minute object.

  Conventionally, the steps between predetermined positions of a minute object and the three-dimensional shape, for example, the steps between nanometer-level patterns and the height of minute mechanisms in a manufacturing line such as a semiconductor or liquid crystal display production line or a micromachine are optically measured. 2. Description of the Related Art As a measurement apparatus for a minute object to be measured, for example, a white light source is used as a light source and measurement is performed using interference of white light from the white light source (see, for example, Patent Document 1).

  Since the white light source is a light source of synthetic light, it generates light having a number of wavelengths. Therefore, when light of many wavelengths from the white light source is caused to interfere on the measurement object as described above, the position where the light strengthens due to the interference is a very short section on the object.

  Therefore, the optical system is moved up and down while imaging the measurement object with a solid-state imaging means such as a CCD camera to change the optical path length, and the light quantity becomes maximum at each point in the image of the imaged measurement object. If the position on the optical system at this time is obtained, it is possible to measure the three-dimensional shape such as the concave-convex shape of the surface of the minute object to be measured.

  Now, for example, FIG. 12 shows a specific configuration of a conventional minute object measuring apparatus employing such a configuration.

  In FIG. 12, reference numeral 1 denotes a minute measurement object at the nano level, 2 denotes a position on the optical system path from the CCD camera 6 facing the measurement object 1, and reflection from the measurement object 1. An objective lens 3 for condensing light is used to change the optical path length by minutely driving the objective lens 2 in the focus direction (optical axis direction) on the optical system path. A beam splitter that includes a mirror 4a and that irradiates the measurement object with the light from the white light source 5 through the objective lens 2, and the beam splitter 6 converges the reflected light from the measurement object 1 through the objective lens 2. 4 is a CCD camera that receives an image of the measurement object 1 by inputting to the light receiving element portion via 4.

  However, the above-described conventional measuring device, in principle, needs to be positioned on the optical system path with a measurement accuracy of nanometer units, and for example, when trying to measure the height of a measurement object of several tens of micrometers, Several hundreds of positioning steps are required, and a considerably long measurement time is required.

  Therefore, in the above conventional measurement apparatus, the higher the required measurement accuracy and the wider the required measurement range, the more measurement time is required, and the measurement efficiency is extremely poor, and the production efficiency of semiconductor production lines and the like is improved. It is an obstacle.

In order to solve this problem, for example, as shown in FIG. 1, first, second, and third laser light generating means L ₁ , L ₂ , L ₃ that generate laser light having a plurality of different wavelengths are used. And first, second and third imaging means C ₁ , C ₂ and C ₃ corresponding to the first, second and third laser light generating means L ₁ , L ₂ and L ₃ , The laser beams emitted from the first, second, and third laser beam generating means L ₁ , L ₂ , and L ₃ merge at the first and second beam splitters B ₁ and B ₂ . After that, the third beam splitter irradiates the laser beam by being divided into two laser beams, that is, irradiation light to the measurement target object and reference light to the mirror.

Then, the reflected light hitting the measurement object is again merged with the reflected light from the mirror M by the _third beam splitter B ₃ to form an interference image, and the fourth and fifth beam splitters B ₄ , B are formed. ₅ , the first, second, and third imaging means C ₁ , C ₂ , and C ₃ are used for imaging.

In front of each of the first to third imaging means C ₁ , C ₂ , C ₃ , optical filters F _{1 to} F ₃ having characteristics corresponding to the wavelengths of the respective laser beams are mounted, and the respective laser beams. An interference image for each wavelength is taken.

  On the other hand, the measurement object is placed on a predetermined stage ST, and the stage ST on which the measurement object is placed is driven by a piezo element or the like to change the optical path length as desired. When the optical path length is changed in units of several tens of nm, the state of interference between the reflected light from the material and the reference light changes, so that the brightness of the image changes. This change changes like a sine wave as shown in FIG. 2 according to the change amount of the optical path length, and one wavelength of the change of the optical path length becomes a period.

  Since the optical path length varies from place to place according to the unevenness of the surface of the measurement object, the way in which the brightness changes changes from place to place, and appears as a phase difference of a sine wave-like brightness change from place to place. Therefore, when this phase difference is measured as shown in FIG. 3, the difference in height can be calculated.

  However, since the sine wave whose brightness changes returns to the original in one period of the wavelength, the difference can be measured only up to the length of one wavelength with the laser light source of one wavelength after all.

However, as described above, when laser light sources L ₁ , L ₂ , and L ₃ having a plurality of wavelengths are used and the phases are individually measured, the change in optical path length depends on the wavelength. As shown, the relationship with the phase of the change in optical path length shifts between the two laser light sources. Therefore, if the phases measured by a plurality of light sources having different wavelengths are combined, it is possible to measure more than the wavelength of one light source.

  In the measurement of a minute object by such a phase shift method using laser beams of a plurality of wavelengths, for example, as shown in FIG. 5, the phase θ of the sine wave of the brightness change is changed by moving the optical path length by every ΔL. It is coded for each ΔL, and a length of one wavelength or more is measured by a combination of the codes.

  For example, when the light source 1 is evaluated in five levels 0-4 with the light source 1 and is evaluated in four levels 0-3 with the light source 2, there are 20 kinds of combination codes, so the wavelength is four times the wavelength of the light source 1. Can be measured. Therefore, for example, by preparing a phase code conversion table as shown in FIG. 6, measurement exceeding the wavelength of the laser beam becomes possible.

  In addition, the measurement of the phase of the brightness change for each light source can be performed by dividing the wavelength of the light source into several tens of wavelengths regardless of the length to be measured. Therefore, the conventional measurement apparatus using white light interference can be used. Compared to this, the measurement time can be greatly shortened.

“Vision Technology Practical Workshop” VIEW 2006, Proceedings, pages 1-7 “Latest Trends in Ultra-Precision 3D Shape Measurement: Development of Innovative Algorithms in White Interference and Its Applications”

  However, when preparing the above phase code conversion table, it is possible to deal with measurement errors due to phase code fluctuations, measurement accuracy during phase measurement, noise, etc. when the wavelength of light is not an integral multiple of the optical path length movement. However, unless these points are solved, it is difficult to put it into practical use as a measuring device.

  The present invention has been made in response to such a technical problem. By limiting the use area on the above-described phase code conversion table, the optical object of the minute object that can absorb the above measurement error can be obtained. The object is to provide a measuring device.

  In order to achieve the same object, the present invention is configured with the following problem solving means.

(1) The invention of claim 1 The present invention relates to a laser light generating means for generating laser light having a plurality of wavelengths including observation light and target light, and an object for measuring observation light and target light from the laser light generating means. An optical system that guides the imaging device in an interferable state, an optical path length varying unit that varies an optical path length of the optical system, and an observation light and an object light while changing the optical path length of the optical system by the optical path length varying unit Imaging means for capturing a plurality of interference images, and measuring the phase of the portion where the brightness of each point in the image of the interference image for each wavelength of the laser light of the plurality of wavelengths changes, An optical measuring device for a minute object configured to measure the size of a measured object by a combination of phase values obtained by dividing the measured phase value into predetermined measurement sections of the measurement object, Phase conversion that represents a combination of sections using section numbers By analyzing with a table, it becomes possible to measure the size exceeding the wavelength of the laser beam to be used, and the combination of section numbers in the above phase conversion table is coarsened in advance, and the unused section number As for the combination of, the error of the measured phase value is subjected to absorption correction by interpolating with a combination number in the vicinity of the same section.

  The change in the section number on the phase conversion table moves diagonally and linearly on the phase conversion table according to the change in the size of the measurement object. Then, when the end of the same phase conversion table is reached, the next combination of section numbers appears in a straight line in parallel with the location shifted by the wavelength difference of the plurality of laser beams combined therefrom.

  Therefore, if the area between these two straight lines is not filled with a combination of different section numbers, and the area belongs to the closer of the two straight lines, the wavelength difference between the section numbers will be n times on the phase conversion table. If there is a distance, it is possible to absorb a change in measurement error for n / 2.

  In this way, by limiting a predetermined use area portion on the phase conversion table to be used, an empty area is provided between two combined straight lines, and data is interpolated by inserting a combination number on a nearby straight line there. By doing so, it is possible to absorb the fluctuation of the measurement value due to the above-described measurement error or the like.

(2) Invention of Claim 2 This invention is characterized in that the phase conversion table in the invention of Claim 1 has an n-dimensional structure corresponding to the number of laser light sources to be used.

  The phase conversion table of the configuration of the invention of claim 1 is a secondary table when two sets of laser light sources of two wavelengths are used, and three-dimensional when three sets of laser light sources of three wavelengths are used. Although it is a table, the basic measurement interpolation process is the same and can be processed in the same way.

  As a result of the above, according to the present invention, it is possible to absorb the measurement error due to the variation of the phase code when the wavelength of the light is not an integral multiple of the movement amount of the optical path length, the measurement accuracy at the time of phase measurement, the influence of noise, etc. Accurate measurement is possible, and practical use is possible.

  FIGS. 1-11 has shown the structure of the optical measurement apparatus of the micro object based on the best embodiment of this invention provided with the above characteristics.

For example, as shown in FIG. 1, the apparatus includes first, second, and third laser light generating means (semiconductor lasers) L ₁ , L ₂ , L ₃ for generating a plurality of laser beams having different wavelengths, First, second and third imaging means (CCD cameras) C ₁ , C ₂ and C ₃ corresponding to the first, second and third laser light generating means L ₁ , L ₂ and L ₃ ; First, second, third, fourth, and fifth beam splitters B ₁ , B ₂ , B ₃ , B ₄ , and B ₅ , a reference mirror RM that forms reference light, and a rotation that rotates the reference mirror RM Driving means D, measurement object (micro object) W, nano stage ST on which the measurement object W is mounted, and optical path length for changing the optical path length by driving the nano stage ST by a predetermined micro distance in the optical axis direction. Variable means (piezoelectric drive type piezo actuator) A, and a plurality of first, second and third laser light generating means _1, L _2, L ₃ a plurality of the laser light is first wavelength emitted from, after merging with the second beam splitter B _1, B ₂ portions, the third beam splitter B ₃ parts by the object W The laser beam is divided into two laser beams, i.e., an irradiation light to the reference mirror and a reference light to the reference mirror RM.

Then, the reflected light hitting the measuring object W is merged with the reflected light from the reference mirror RM again by the third beam splitter B ₃ to form an interference image, and the fourth and fifth beam splitters B _4. , B _5, and images are taken by the first, second and third imaging means C ₁ , C ₂ and C ₃ . Before each of these imaging means C ₁ , C ₂ , C ₃ , first, second, and third optical filters having characteristics corresponding to the laser light wavelengths of the laser light sources L ₁ , L ₂ , L ₃ , respectively. F ₁ , F ₂ , and F ₃ are attached to capture an interference image for each wavelength of each laser beam.

  On the other hand, the measurement target W is placed on the nanostage ST as described above, and the nanostage ST on which the measurement target W is mounted is driven at the nano level by the optical path length varying means A, and the optical path length is set. It is changed every desired distance ΔL. For example, when the optical path length is changed in units of several tens of nanometers, the interference state between the reflected light from the measurement object W and the reference light changes, so that the brightness of the image changes.

  For example, as shown in FIG. 2, the brightness change of the image changes in a sine wave shape according to the change amount of the optical path length, and one wavelength of the change of the optical path length becomes a period. Since the optical path length varies from place to place according to the unevenness of the surface of the measuring object W, as is apparent from FIG. 3, the way of changing light and dark also varies from place to place. Appears as a phase difference. Therefore, when this phase difference is measured, the difference in the height h of the measuring object W can be calculated.

Since the sine wave whose brightness changes returns in one period of the wavelength, the difference can be measured only up to the length of one wavelength with a laser light source of one wavelength. However, as described above, if a laser light source having a plurality of wavelengths is used and the phase is individually measured, the change in the optical path length depends on the wavelength. _1/2, lambda _2/2 of the two laser light sources ₁ (L _1), between 2 (L _2), coming shifted as shown in FIG 4, for example. Therefore, if the phases measured by the plurality of laser light sources 1 and 2 are combined, it is possible to measure more than the wavelength of one light source. Note that the horizontal axis X in FIG. 4 indicates the amount of phase advance.

  In the measurement of a minute object by the phase shift method using light sources having a plurality of wavelengths, for example, as shown in FIG. 5, the phase θ of the sine wave of the brightness change is changed every ΔL by moving the optical path length every ΔL. Encode and measure the length of one wavelength or more by combining the same code value.

  When the light source 1 is evaluated in five stages 0-4 with respect to the wavelength of the laser light, and the light source 2 is evaluated in four stages 0-3, since there are 20 kinds of combination codes, the light source 1 is four times the wavelength Can be measured. For example, by using the phase code conversion table as shown in FIG. 6 described above, measurement exceeding the wavelength of the laser beam becomes possible. In addition, the measurement of the phase of the brightness change for each light source can be performed by dividing the wavelength of the light source into several tens of wavelengths regardless of the length to be measured. Therefore, the conventional measurement apparatus using white light interference can be used. Compared to this, the measurement time can be greatly shortened.

  However, as described above, in the preparation of the phase code conversion table, when the wavelength R of the light is not an integer multiple of the movement amount ΔL of the optical path length, It is necessary to deal with measurement errors due to the measurement accuracy and noise, otherwise it cannot be put into practical use.

  Therefore, the following configuration and creation method are employed in the phase conversion code table in this embodiment.

  That is, in the same phase code conversion table, when a change in the optical path length at which one nanostage ST operates is ΔL, a phase code (section number) divided by ΔL is attached for each light source wavelength L. That is, when the wavelength of the laser beam is L, the phase code has a code of 0−L / ΔL.

  In the example of the combination of two sets of two wavelengths of laser light in FIG. 6, a value of 0-37 is assigned to the laser light of wavelength 1, and a value of 0-43 is assigned to the laser light of wavelength 2. When the end of the table is reached with the first linear combination from (0,0) to (37,37), the next linear arrangement starts from (0,38) and (5,43) Until then, the next ... and so on.

  By the way, when changing in this way, if a height code is sequentially added from 0, the height code continues until the entire table of FIG. 6 is filled, and in this case, 38 × 44 = 1672. Can be coded. Therefore, the height up to 1672 × ΔL can be measured.

  As a specific method for preparing the phase code conversion table, the nanostage ST is moved for each of 1672 × ΔL by the optical path length varying means to obtain individual codes at that time, and the two lasers are obtained. The number n of movements by the optical path length varying means A at that time may be put in the combination position with light.

  However, this method has a problem that a measurement error occurs when there is a variation in phase code or a measurement error in the phase itself when the wavelength R of light is not an integral multiple of the movement amount ΔL of the optical path length.

  Therefore, in this embodiment, for example, as shown in FIG. 7, the combinations of the section numbers in the phase code conversion table of FIG. 6 are coarsened in advance (the measurement range has a width). The combinations of the vacant section numbers are interpolated with the combination numbers in the vicinity of the section to absorb and correct phase code fluctuations and measurement errors.

The flowchart (steps S _{1 to} S ₉ ) in FIG. 8 shows a specific method for configuring the phase code conversion table in the present embodiment configured from such a viewpoint.

That is, in the first step S ₁ , the measurement object W such as a flat mirror is placed on the nanostage ST described above, and the stage ST is moved every predetermined measurement interval (optical path length variable interval) ΔL. The desired measurement range is moved, and an image for the measurement range is taken and recorded for each movement amount ΔL.

Next, in step S _2, to obtain a brightness in a stable position of the screen, determining the phase code Ci at that position. That is, for each image captured as described above, the phase is measured using the images before and after it to obtain the phase code Ci. Here, the phase code Ci is, for example, a value obtained by dividing the phase position θi of the laser beam obtained by a method such as normalized correlation with the standard pattern by the measurement accuracy Δθ.

  In this way, when the phase code Ci at each height is obtained for each laser light source, the height code is put into the combination position of the phase code in the phase code conversion table of FIG. 10 (the coarse phase code of FIG. 10). Write as long as you can write to free space in the conversion table).

  The phase code conversion table is configured in n dimensions depending on the number n of laser light sources to be used. The example of a two-dimensional table with two light sources is as shown in FIG. 10 above, and the example of a three-dimensional table with three light sources is as shown in FIG. 11. The number of table frames is one period for each laser light source. It is a value divided by a constant value Δθ determined from the measurement accuracy.

In step S _4, obtains a point that combines a phase code Ci obtained in step S ₂ in such a phase code conversion table, when connecting therebetween, for example, as shown in FIG. 7, each point phase code When a combination of Ci is viewed as a coordinate value, it is in principle a straight line (see the two-dot chain line portions of A, B, and C in FIG. 7). That is, since the combination codes are arranged on a straight line, the position on the same phase code conversion table is regarded as a coordinate, an approximate straight line formula is calculated, and the error is corrected for each straight line (correction is repeated).

In this case, the combination points of the phase codes Ci on the conversion table are arranged in a straight line although they are linear. Therefore, since some errors there also the measurement accuracy, in step S ₅ subsequent example, by calculating an approximate straight line by a least square method and the linear interpolation for all combinations positions of all the frames on the straight line Insert a height code and linearly interpolate the height code to enter an intermediate height.

However, when a height code on a straight line is inserted during this interpolation operation, the position where the code in the range of ± n in the n-dimensional space defined as the error range is inserted from the position where the height value is entered (the position of the ± n section). if already contains a predetermined value other than the straight line, as shown in step S _6, where it ends the interpolation operation, it deletes the data of the straight line.

The above steps S _{3 to} S ₆ are repeated until it is confirmed in step S ₇ that assignment to each column is no longer possible.

If then assign as described above is completed, after putting a height code in a linear shape as shown in step S _8, the empty frame, filled with high code that is closest to the frame. For this, for example, thickening processing in image processing can be applied. For this reason, in this process, first, a space of ± 1 before and after the vacant position in the previous space of the conversion table is examined, and if there is already a height code, it is moved, otherwise it is left vacant. That is, an empty frame is searched for on the phase code conversion table of FIG. If there is a free frame, for example, a 3 × 3 two-dimensional window as shown in FIG. 9 is applied to the frame in the case of the two-dimensional table of FIG.

Now, for example, when as a frame is empty frame of A ₀ in the middle of a two-dimensional window (3 × 3) of FIG. 9, examine the window position other than the frame containing the value to the middle and immediately before, there If the height code has already been entered, the process of moving that code and leaving it unoccupied is repeated many times until there is no empty frame (step S ₉ ). Then, when there is no empty frame in the in-phase code conversion code table, the process is terminated.

  As a result of these processes, in the case of the two-dimensional table of FIG. 7, a phase code conversion table as shown in FIG. 10 is finally obtained.

  When three sets of laser light sources having three wavelengths are used, a three-dimensional conversion table as shown in FIG. 11 can be obtained. In this case, the basic processing is the same, and straight lines are shown in FIG. The method of filling a space that is drawn spatially also forms a 3 × 3 × 3 three-dimensional window (combination of three sets of windows in FIG. 9) (the straight line in FIG. 11 is (Omitted).

  According to the above configuration, in the minute height measurement device based on the combination of phases of interference images from a plurality of laser light sources having different wavelengths, the phase measurement value includes some errors or the phase measurement value is Even when some coding errors occur due to not being an integral multiple of the coding width, it is possible to perform accurate and stable measurement that can be used in practice by absorbing these errors.

It is a figure which shows the basic composition of the optical measurement apparatus of the micro object based on the best embodiment of this invention. It is explanatory drawing which shows the brightness change (sine wave-like change) of the measurement image by the change of the interference state of the reflected light from the measurement object corresponding to the change of the optical path length in the same apparatus, and a reference light. It is explanatory drawing which shows that the difference of the height of a measurement object can be measured from phase difference (DELTA) (theta) corresponding to the brightness change (sine wave-like change) of the measurement image in the same apparatus. It is explanatory drawing which shows that the wavelength or more can be measured by combining a some laser beam, since the brightness change by the optical path length change of the some laser beam in the same apparatus is decided by the wavelength of a laser beam. Description of an encoding method in the case where the optical path length is moved for each ΔL in the same apparatus, and the phase θ of the sine wave of the brightness change is coded for each ΔL, and the length of one wavelength or more is measured by the combination of the codes. FIG. It is a figure which shows the structure of the phase code conversion table which converts the phase code into height in the same apparatus. It is a figure which shows the basic composition of the n-dimensional table (secondary table) corresponding to the number n (n = 2) of the laser light sources for enabling it to cope with some measurement errors in the apparatus. It is a flowchart which shows the formation method of the final phase code conversion table in the same apparatus. It is a figure which shows the two-dimensional window (3x3) used in the formation process of the phase code conversion table in the apparatus. It is a figure which shows the example of the two-dimensional table at the time of 2 light sources and 2 wavelengths from which the vacant frame was lost on the table in the formation process of the phase code conversion table in the apparatus. It is a figure which shows the example of the three-dimensional table at the time of 3 light sources and 3 wavelengths from which the empty frame was lost on the table in the formation process of the phase code conversion table in the apparatus. It is a figure which shows the structure of the optical measurement apparatus of the micro object by the conventional white light source.

Explanation of symbols

L ₁ ~L ₃ the first to third laser light sources, B ₁ .about.B ₅ first to fifth beam splitter, C ₁ -C ₃ first to third CCD camera, F ₁ to F ₃ Are the first to third optical filters, RM is the reference light mirror, D is the rotation drive means, ST is the nanometer stage on which the measurement object W is placed, and A is the optical path length variable means.

Claims (2)

  Laser light generation means for generating laser light having a plurality of wavelengths including observation light and target light, and optics for guiding the observation light and target light from the laser light generation means to the imaging device in a state where they can interfere with each other via a measurement object An optical path length varying means for varying the optical path length of the optical system, and an imaging means for capturing a plurality of interference images of the observation light and the target light while changing the optical path length of the optical system by the optical path length varying means. Measuring the phase of the portion where the brightness of each point in the interference image for each wavelength of the laser beams of the plurality of wavelengths changes, and determining the size of the measurement object by combining the measured phase values. An optical measuring device for a minute object to be measured, wherein the measured phase value is divided into predetermined measurement sections of a measurement object, and a phase conversion table in which combinations of the sections are represented by section numbers. By analyzing by In addition to making it possible to measure beyond the wavelength of the laser light to be used, the combination of section numbers in the above-mentioned phase conversion table is coarsened in advance, and the combination of the unused section numbers is the same section. An optical measurement apparatus for a micro object characterized in that an error of a measured phase value is corrected and corrected by interpolating with a combination number in the vicinity of.   The optical measurement apparatus for a minute object according to claim 1, wherein the phase conversion table has an n-dimensional structure corresponding to the number of laser light sources to be used.

Images