DE102010019668B4 - Method for determining the topography of a surface of an object - Google Patents

Abstract

Method for determining the topography of a surface (20) of an object (11) in which a light bundle emanating from a light source (1) is collimated via a collimator mirror (4), with the collimated light beam projecting a pattern onto the surface (20) and by means of a camera with a camera lens, at least two images (26, 27a, 27b) of an intensity distribution of the light beam reflected by the surface (20) are detected at at least two different focus settings in a plane (24) spaced from the surface (20). having,
wherein the pattern is generated via a structured mask (2), which is arranged in the beam path of the light beam, and
wherein the topography is determined from a comparison of the at least two images (26, 27a, 27b) taking into account the focus settings.

Description

Technical application

The present invention relates to a method for the clear and reference-free determination of the topography of an optical radiation-reflecting surface of an object, in which a pattern with a collimated light beam is projected onto the surface and an intensity distribution of the light beam reflected by the surface is detected in a plane containing a Distance from the surface.

The determination of the surface topography of objects is necessary above all in technical areas in which the surface quality, in particular the flatness or structure of the surface, plays an essential role for the application or for further processing steps. Examples include the surfaces of wafers in semiconductor technology or the surface quality of a workpiece after polishing.

State of the art

Topographies of reflective surfaces can be determined by means of plane wavefronts. For this purpose, the surface is illuminated with a flat wavefront. Deviations of the reflecting surface from an ideal plane are introduced into the wavefront during reflection. The reflected wavefront is then analyzed for these changes.

In the so-called Makyoh method, an image of the intensity distribution of a parallel light beam reflected on the surface is detected in a plane which is at a distance from the surface. From this intensity distribution a deviation of the surface topography from an ideal plane can be determined. An example of such a technique shows F. Riesz "A quantitative approach to Makyo (magicmirror) topography", Journal of Crystal Growth 210 (2000), pages 370-374 ,

From the US 6,545,764 B1 a modified technique for determining the topography of a surface is known in which two images of the intensity distribution at two different distances from the surface are detected and evaluated. The evaluation is carried out by a complex iterative process.

From the DE 602 12 987 T2 Another method for determining the topography of a surface of an object is known. In this technique, a pattern is projected onto the surface by means of a collimated light beam and the intensity distribution of the light beam reflected at the surface is detected at a distance from the surface. By using a flat reference surface instead of the surface of the object, another image is recorded in the same way. By comparing the two images, the surface topography of the object can then be determined on the basis of the distortion of the pattern. This is based on the 1a illustrating the surface, the reference plane and the beam path of a single light beam in the reflection. The dashed line represents the reference surface 21, the oblique solid line shows an example of an uneven surface 20 of the object to be examined. In the figure, an incident light beam 22 of the light beam is shown, which corresponds, for example. With a reference point of the pattern. After reflection at the surface, the intensity distribution of the light beam is detected in the plane 24 indicated in the figure, which has a distance L from the surface. The distance Δd between the black and white points in the figure represents the deviation of the reflected light beam from the perpendicularly incident light beam in the detection plane. If the viewed light beam corresponds to a reference point in the pattern, the distance Δd corresponds to the distance of the imaged reference points 23, 25 in the two pictures. With the aid of the distance L between the surface and the detection plane, the local inclination angle α of the surface of the object relative to the reference plane can then be determined: $$tG2\mathrm{\alpha \; =}\frac{\Delta d}{L}\text{,}$$

Using this equation, the topography profile of the surface of the object can then be calculated.

However, the implementation of this method is not free of problems. Thus, the accuracy of determining the topography is limited by the quality or flatness of the reference surface.

For optimal representation of the images also the surface of the object and the surface of the reference mirror must be aligned exactly perpendicular to the incident light beam. During the adjustment, however, there is often a slight deviation of the normals of the reference surface and the object surface from the incident light rays. 1b Fig. 12 illustrates a case where the reference surface 21 has an inclination of α _Ref ≠ 0. The shift of the reference point between the two images is therefore smaller or larger by δ _Ref than that which, in the ideal case, is the 1a is shown. This leads to an error of δ _Ref / Δd in the calculation of the topography profile.

For an optimal calculation of the topography, the surface of the object and the reference surface must also be in the same plane. 1c illustrates a case in which the reference and object surfaces have different distances to the detection plane, so lie on different levels in height. This leads to a different distance L for the two measurements. The difference in the planes ΔL = LL _Ref changes the shift between the reference points in the images of the ideal case of 1a on δ _ref . This also leads to an error of δ _Ref / Δd in the calculation of the topography profile.

Another problem with using only an image of the intensity distribution of a light beam reflected at the surface of the object is that ambiguities may arise due to the large amount of information in the image. 2 shows an example in which two different surface profiles 20a, 20b and four light beams 22, corresponding to four reference points, are shown. The upper part of the figure shows the position of the imaged reference points 25 in the captured image 26, which is identical here for both surfaces with respect to the position of the four reference points. In this constellation, the two inner light beams each deviate from the vertical in such a way that they coincide after reflection for both surfaces 20a, 20b in the same point of the detection plane 24. Therefore, the inclination of the object surface can be misinterpreted from the image and thus the topography can be calculated incorrectly.

Both an undesired tilting of the object surface or the reference surface and a different height of object and reference surface can lead to errors in the calculation of the topography profile. Furthermore, the quality of the reference surface or of the mirror used for this purpose is a limiting factor with regard to the accuracy of the determination of the topography. An ambiguity in the picture can lead to a wrong interpretation of the topography. To date, these problems can only be solved in part by using a flat reference surface with the smallest possible deviations from an ideal plane, and adjusting the object surface and the reference surface in each case with very high accuracy. However, this is time consuming and expensive.

The DE 10 2005 007 244 A1 describes a method for detecting the shape of reflective surfaces, in which a light beam emanating from a light source is split over a microlens array into a plurality of light needles which strike the surface in parallel alignment with one another. In the method, images of the intensity distribution of the light needles reflected from the surface are detected at two different distances to determine the surface shape from a comparison of the two images taking into account the known position and beam direction of the reflected light needles.

The object of the present invention is to provide a method for determining the topography of a surface of an object, with which these problems are largely avoided, so that the accuracy of the measurement no longer depends on the surface quality of a reference surface and the most accurate adjustment and ambiguity the evaluation be avoided.

Presentation of the invention

The object is achieved by the method according to claim 1. Advantageous embodiments of the method are the subject of the dependent claims or can be found in the following description and the embodiments.

In this case, the light beam reflected by the surface of the object is directed onto a camera which has a camera lens. The images captured with the camera are processed in a computer. For this purpose, a suitable evaluation algorithm can be used which on the one hand identifies reference points in the recorded images and on the other hand from the displacement of these reference points into the respective ones Images the topography of the surface according to equations which apply to arrangements with shifted image reception plane. At least two images with at least two different focus settings are captured. The two images are then evaluated with the respective parameters L ₁ and L ₂ known from the different focus settings

The accuracy of determining the topography of the surface also depends on the structure of the pattern used. Preferably, a dot matrix is used as a pattern, wherein the individual points of the dot matrix serve as reference points whose relative displacement in the at least two images is determined in order to calculate therefrom the topography of the surface. Of course, however, other, in particular more complex, patterns or a combination of dots and other geometric shapes can be used as a pattern. The pattern is preferably generated via a structured mask, which is arranged in the beam path of the light beam.

The method is not limited to the acquisition of two images at two distances or two angles of inclination. Rather, even more images can be recorded at further distances or angles of inclination and used for the evaluation.

The proposed method is suitable for surface profile determination of objects with reflective surfaces. By eliminating the use of a reference surface and reducing adjustment requirements, the process is simplified and the accuracy of topography determination improved. In this case, all surfaces reflecting optical radiation, such as, for example, semiconductor wafers, reflective metal surfaces or mirrors, can be measured. The method is also suitable for the measurement of structured surfaces, such as, for example, surfaces of integrated circuits or of MEMS components.

list of figures

• 1 eine schematische Darstellung der Verhältnisse bei der Vermessung gemäß dem Stand der Technik;
• 2 eine weitere schematische Darstellung der Verhältnisse bei der Vermessung gemäß dem Stand der Technik;
• 3-5 Verhältnisse zur Herstellung von Gleichungen, die für Anordnungen mit verschobener Bildempfangsebene gelten;
• 6 eine beispielhafte Anordnung zur Durchführung des vorgeschlagenen Verfahrens; und
• 7 Verhältnisse bei einem geneigten Objekt.

The proposed method will be explained in more detail using exemplary embodiments in conjunction with the drawings. Hereby show:

• 1 a schematic representation of the conditions in the measurement according to the prior art;
• 2 a further schematic representation of the conditions in the measurement according to the prior art;
• 3 - 5 Ratios for making equations that apply to displaced image receiving plane arrangements;
• 6 an exemplary arrangement for carrying out the proposed method; and
• 7 Conditions in a tilted object.

Ways to carry out the invention

The problems occurring in the known method of the prior art have already been based on the 1 and 2 explained in more detail in the introduction, so that these figures will not be discussed in more detail here.

In a procedure after 3 Two images of the intensity distribution of the reflected light beam at the object surface at different distances L ₁ , L _{2 are recorded} to the surface with a camera. This technique is used in conjunction with the prior art method 2 suggested ambiguity avoided. 3 shows the same situation with the two different surfaces 20a . 20b as used in conjunction with the prior art method 2 is shown. By recording in each case two images of the same surface at different distances L ₁ , L ₂ , each individual light beam of the light beam can be traced, since the difference between the two images shows the direction of displacement of the respective reflected light beam or of the corresponding reference point. In the recorded pictures 26 the two surfaces 20a . 20b at the distance L ₂ , the points of the light rays or reference points in the detection plane are identical, as in the uppermost dashed line of FIG 3 is indicated. In the two pictures 27a . 27b of the two surfaces recorded at the distance L ₁ , these points are separated in the detection plane. This is indicated by the two underlying dashed lines representing the two images 27a . 27b at the distance L ₁ represent. Thus, the determination of the topography is unique by the capture of two images at different distances L. A reference surface is not required.

The change of the distance L takes place here either by changing the focus of the camera lens or by changing the working distance, d. H. the distance between the surface and the camera lens or camera.

$$h=\frac{1}{2}\frac{\Delta d}{\Delta L}\cdot d_{Zelle}\text{.}$$

In the alternative method, in which the images are recorded at different distances from the object surface, an undesirable tilting of the object or the surface of the object can also be determined and eliminated. Such undesirable tilting of the surface 20 the object is in 4a indicated. When comparing the two images with the different distances L ₁ and L ₂ , the object tilting is shown as a systematic shift of all points of the reflected light beams or of the projected pattern in the detection plane. This systematic shift can be determined with a known pattern and thus calculated out when determining the topography. 4a shows the tilting or inclination angle α ' and the respectively measurable shift of the reference points between the two images Δd ' ₁ . Δd ' ₂ and Δd ' ₃ for the three light beams or reference points shown. The distance of the reference points was designated d _cell in this figure. After determining the systematic shift of all points of the reflected light rays or reference points in the detection plane and after correcting the data for the object tilt, the topography profile can be calculated by the remaining displacement of the reference points by the distances Δd ₁ , Δd ₂ and Δd ₃ in 4b is indicated. The calculation of the topography profile is done with the following equations $$tG2\mathrm{\alpha }=\frac{\Delta d}{L}$$

$$H=\frac{1}{2}\frac{\Delta d}{\Delta L}\cdot d_{Zelle}\text{,}$$

These equations can be seen from the schematic representation of 5 be traced. Here, ΔL = L ₂ -L ₁ and d _{cell represents} the distance between two juxtaposed incident light beams or reference points. From the shift Δd of the points of the light beams or the reference points in the detection plane and the difference ΔL of the different distances of the detection plane Surface, the angle between the perpendicularly incident and the reflected light rays is calculated. This angle is twice as large as the local inclination angle of the object surface α. With the aid of the determined local inclination angle α and the distance to the surface L ₁ or L ₂ and the distance between two adjacent light beams or reference points d _cell , the topography profile h is calculated according to the above second equation.

The greater the difference ΔL of the detection distances of two images, the greater the shift Δd of the points of the light beams or the reference points in the detection plane. The parameter Δd corresponds to the parameter Δd _pix , which is also mentioned below. Both parameters differ only in their units (unit of Δd: meter, unit of Δd _pix : pixel). With the parameter r _pix-m Δd _{pix is transformed} to Δd and back. The larger the parameter Δd, the smaller the relative error δ _pix / Δd _pix : $$\Delta d\pm {\delta }_{pix}\cdot r_{pix-m}=\Delta d\left(1\pm \frac{{\delta }_{pix}}{\Delta d_{pix}}\right)$$

The parameters δ _pix = 0.5 pixels and Δd _pix correspond to the camera error, which is twice smaller than the maximum resolution of the camera d _camera = 1 pixel, and the displacement of the points of the light beams in the detection plane. The parameter r _pix-m represents the lateral calibration of the distances d on the object surface. Therefore, for the method proposed here, the greater the difference of the distances ΔL, the more accurate the measurement.

6 schematically shows an example of an arrangement for carrying out the proposed method. This is done with a light source 1 , For example, an LED or an LED array, the structure of a mask 2 on the surface of the object 11 projected. That from the light source 1 emitted light beam is in this case via a collimator mirror 4 collides and falls perpendicular to the surface of the object 11 that in a corresponding holder 5 is attached. Via a beam splitter 3 This will be the surface of the object 11 reflected light bundles on a CCD camera 8th directed a camera lens 6 having. The images captured with the camera are stored in a computer 10 processed. For this purpose, a suitable evaluation algorithm can be used which on the one hand identifies the reference points in the recorded images and on the other hand calculates the topography of the surface from the displacement of these reference points in the respective images according to the above equations. The calculator is with a controller 9 connected, in this example, the motorized holder 7 for the camera lens 6 controls. The control is carried out such that for a measurement of the surface of the object 11 the focus of the camera lens 6 is changed to capture at least two images at at least two different focus settings. The two images are then evaluated with the respective parameters L ₁ and L ₂ known from the different focus settings and the local inclination angles of the surface are determined. Thereupon a possible systematic tilting of the object surface from the data is eliminated. Subsequently, the topography of the surface is calculated from the corrected data. This is done by the evaluation program used in the computer.

In an alternative approach, the object holder becomes 5 controlled to shift the sample along the light beam. Again, at least two images with different distances of the surface to the camera 8th detected. The two images are also evaluated here in the same way as has already been explained in connection with the change of the focus settings of the camera lens.

Such an arrangement can also be used in a further procedure in which the surface of the object is tilted step by step, whereby at least two images at different angles of inclination, for example inclination angle α ₁ and inclination angle α ₂ (where α ₁ or α _{2 can} also be 0) ) to be recorded. For this, the object holder must 5 allow a defined tilt of the object. Such a gradual tilting of the object holder 5 leads to a systematic shift of the points of the light beams or the reference points in the detection plane. This shift in the two images is evaluated in the same way as when capturing the images at different distances. Correspondingly small tilting of the surface of the object is necessary since otherwise no complete image is detected. The shift of the points of the light beams or reference points in the detection plane is significantly lower than in the first alternative method. A motor-driven change in the focus of the focus lens and the controller 9 are not required here. 7 schematically shows the conditions in the recording of the two images at the same distance L ₁ but different tilting of the surface of the object.

The proposed method enables the measurement and determination of the topography of an object surface without a reference surface (eg mirror). As a result, the accuracy of determining the topography is no longer limited by the quality of a reference surface. The accuracy of the determination of the topography is - apart from the resolution of the camera used - only on the quality of the optical components of the measuring arrangement and the accuracy of the determination of the instrumental parameters L ₁ and L ₂ dependent.

LIST OF REFERENCE NUMBERS

1

light source

2

mask

3

beamsplitter

4

collimator

5

object holder

6

camera lens

7

motorized holder

8th

CCD camera

9

Controller for the camera lens

10

computer

11

object

20, 20a, 20b

Surface of the object

21

reference surface

22

beam of light

23

Reference point in the reference image

24

Input level

25

Reference point in the object image

26

Image of both surfaces

27a

Image of the first surface

27b

Image of the second surface

Claims (3)

Method for determining the topography of a surface (20) of an object (11) in which a light bundle emanating from a light source (1) is collimated via a collimator mirror (4), with the collimated light beam projecting a pattern onto the surface (20) and by means of a camera with a camera lens, at least two images (26, 27a, 27b) of an intensity distribution of the light beam reflected by the surface (20) are detected at at least two different focus settings in a plane (24) spaced from the surface (20). having, wherein the pattern is generated via a structured mask (2), which is arranged in the beam path of the light beam, and wherein the topography is determined from a comparison of the at least two images (26, 27a, 27b) taking into account the focus settings. Method according to Claim 1 , characterized in that the projected pattern has reference points (25) and the topography is determined from a relative displacement of the reference points (25) of the pattern in the images (26, 27a, 27b). Method according to Claim 1 or 2 , characterized in that an undesired inclination of the surface (20) relative to the plane (24) by the comparison of the images (26, 27a, 27b) is determined and excluded in the determination of the topography.

Images