DE102014011268A1 - Optical defect detection - Google Patents

Abstract

An improved device for the detection of defects (7, 7 ') on an object (1), in particular of defects (7) in the form of microholes (7') in a material web (1 ') or a film (1' ') draws inter alia by the following features: at least one primary light source (L, L1) is provided, by means of which a light beam can be generated, with a sensor and / or detection device with sensor or detector elements (SES, SE) with an evaluation device ( AE) for evaluating the defect location (7), - the device with the evaluation device (AE) c) comprises a balancing circuit for the sensor or detector elements (SES, SE) in such a way that a difference between one of a sensor or detector element (SES , SE) and at least one measurement signal generated by an adjacent or referenced sensor or detector element (SES, SE) generates a balanced signal for defect location evaluation, and / or d) comprises a measurement setup such that the defect sites (7, 7 ') pass through the beam or focus waists (2w0) imaged onto the surface of the object (1) to be examined at a speed (v), a defect location -dependent measurement signal (convolution) by a in the beam path to the defect point (7, 7 ') arranged sensor and / or detector element (SES, SE) from the entrance of the defect point (7, 7') in the jet or Fokustaille (2w0 ) until leaving the jet or focus waist (2w0).

Description

Background of the invention

The invention relates to an optical detection method for measuring defects (defects) in and on materials. By measuring techniques applications on a variety of moldings, but especially on material webs (polymer or metal foils, paper or textile webs), coatings and permeable or porous materials can be covered. A special application of measurement technology relates to so-called separator films and membranes.

With these separator films, production-related pinholes may occur across the film thickness (th), which in the worst case could lead to battery failure.

To detect these holes, there are two main methods in the market, namely optical and electrostatic detection.

State of the art

Defect detection systems are widely represented in the market (Schenk, OCD, ISRA, Futec, etc.). In principle, a moving object is irradiated with a light source, usually an LED line source, and the re-inflected or transmitted light is detected by means of a line camera, usually a CCD or CMOS array, via a photo-optic system. If a defect occurs, there is a deviation in the gray values compared to a reference. The analysis of the measured values then takes place against faulty patterns from a teaching procedure. The resolution of the conventional detection systems depends on the working width (AB) - with moving material webs such as plastic films in the transverse direction (TD) transverse to the withdrawal direction of the material web - the number and size of the Arraypixel or the number of cameras used, and in Direction of the web (MD) from the clock frequency of the array. Currently available in the civilian market are cameras (Teledyne) with a pixel size (pitch) of 3.5 μm, 16 k pixels and a data transmission rate (line rate) of up to 140 kHz.

The method is suitable for high-contrast defects and materials with different optical properties between base material and defect. In addition, the material or web should not absorb too much or scatter. Currently, defects larger than or equal to 50 μm are measured by the method, in the experimental range the limit is currently 14 μm, the defects themselves are qualified with synchronized microscopy.

Small signal changes, such as microholes, porosities or micro-scratches in front of a strongly transmitting or scattering background, can not be detected due to overdriving of the array elements. The biggest problem is the diffraction of the very small imperfections (below 50 microns), which causes the intensity distribution of the Airy pattern decreases sharply and is no longer detectable before the background signal. The image brightness is proportional to the diameter of the pinhole to the fourth power. At half the diameter, therefore, the amount of light decreases by four times, the area of the diffraction disk increases fourfold. The luminous flux density is thus 16 times lower in the pixel.

Another problem is the line lights, since the discrete arrangement of the LED results in a periodic intensity pattern.

The US Pat. No. 7,224,447 describes a tipping paper detecting device in which the polarization change of the scattering web or material web is utilized with respect to the polarization of the laser beam passing through the microhole. With a polarizing beam splitter and various geometric images, a distinction can be made between direct light and scattered light.

Similar sensors, but without polarization measurement, can be found in the DE 102 51 610 and US 2006 010 847 in which a light source irradiates the web via an optical system, and a single sensor, usually a photodiode (PIN), detects the signal deviation with respect to a reference via focus optics. In order to obtain a sufficiently large working width (AB), several of these sensors are lined up in an overlap.

In addition, reading devices are made of scanning devices ( WO 2002/73 173 ; DE 38 50 871 ; DE 101 24 943 ) are known in which a long contact image sensor (CIS contact image sensor) is in or almost in contact with the material or web. Frequently, a micro-optic (GRIN, etc.) is applied between the web and the line sensor to facilitate or improve the adjustment and focusing.

All of these sensors have the disadvantage that they can not resolve the signals of diffractive defects with respect to the background signal.

MD resolution according to the prior art

The resolution in the transport direction depends on the readout rate (read rate - line rate l_r) and the speed of the object v or Defective. Assuming defects of 7 [μm] are to be detected, this is possible with a read rate of 70 [kHz] (line rate - raster time approx. 14 [μs]) up to MD speeds of 30 [m / min] ( 14 ).

In an idealized representation ( 15 ) is a defect in a 1: 1 image of z. B. 7 [μm] size at time t0 before the line array (here with pitch 3.5 [μm]) or the CIS imaged. At a withdrawal speed of the web and thus a defect thereon of z. B. 30 [m / min], the defect is detected at a read rate (line rate) of 70 [kHz] at time t1 = t0 + 14 [μs] ( 15 ).

Minor defects (in 15 indicated by a smaller circle) could then fall between the line rate grid and thus not be recognized.

Realistically, however, it has to be taken into account that images> 1 are used and usually no coherent light sources are used. Especially for defects smaller than 50 [μm], the diffraction effects must be considered.

By diffraction at a defect z. B. in the form of a micro hole creates a strong expansion of the image (Airy disc). Instead of a point of light, as required by geometrical optics, Fraunhofer diffraction yields a diffraction disk (Airy disk) with a diameter DA = 2.44 λf / DP, where λ is the wavelength of the irradiated monochromatic light, f is the focal length and DP designates the (average) diameter of the defect of an imaging lens.

In the case of small defects, even when a coherent light source is used, the focus is no longer sharp, but a widened Airy pattern is created whose image brightness also decreases with the fourth power of the decreasing radius. Instead of the pixels n and n + 1 in the example given, DA / Pitch [pixels] are now illuminated with greatly reduced intensity.

TD resolution according to the prior art

Suppose a 1: 1 mapping to an array is present ( 15 ), then a defect at time t0 in the transverse or TD direction over at least one, but usually over several pixels as a gray value over the line array. At the time t0 + line rate (sampling or read-out rate), another gray value pattern is mapped. Depending on the read-out rate (line rate), the overlay of the pattern can detect the defect and the defect position in the machine direction and transverse direction, at least as long as the read rate (line rate) is faster than the take-off speed of the product web. Each pixel is thus read out at the read rate (line rate). The same applies to an image in which a partial area of the working width ABa is mapped to the array length array_l with ABa: array_l> 1. If line arrays are used, then for reasons of cost, focus optics are usually used to map a larger TD area to the relatively short array. Now that a larger portion of a pixel falls on a pixel, the signal / noise ratio deteriorates.

• A: Auflösung in TD-Richtung [μm] mit 2048 [Pixel] Kamera
• B: Auflösung in TD-Richtung [μm] mit 4096 [Pixel] Kamera
• C: Auflösung in TD-Richtung [μm] mit 8192 [Pixel] Kamera
• D: Auflösung in TD-Richtung [μm] mit 16384 [Pixel] Kamera.

In 16 For this purpose, a representation of the maximum resolution [μm] in the transverse direction in relation to the scanning range in [mm] is shown with respect to the four variants

• A: Resolution in TD direction [μm] with 2048 [pixel] camera
• B: Resolution in TD direction [μm] with 4096 [pixel] camera
• C: Resolution in TD direction [μm] with 8192 [pixel] camera
• D: Resolution in TD direction [μm] with 16384 [pixel] camera.

At the in 17 In the case of a 16 k [pixel] array and a pitch of 3.5 [μm], the result is that the lines start at a 1: 1 mapping (array_l = pixel · pitch) an array length, and with a 1: 1 image a TD length of about 57 [mm].

In order to make the readout or sampling rate (line rate) as large as possible, the pitch size (ie the distance measure) and thus the array length is reduced.

For cost reasons, an image larger than 1: 1 is selected, which has the consequence, however, that the intensity distribution of a small defect is mapped to fewer pixels or a smaller partial area of a pixel, and the intensity per pixel decreases greatly relative to the background, and so that the detection difficult.

Together with the diffraction at small defects, the signal is hardly distinguishable from the noise. The same applies to contact image sensor (CIS).

In other systems with focus optics, a particular work area ABa is mapped onto a single sensor Rr. Many of these individual detectors Rr are then lined up along the working width (AB) in succession or overlapping. Occur z. B. several defects within ABa at the same time, then increases the Rr signal height accordingly.

Is in this sensor arrangement, the carrier material, ie the material web is not transparent or not reflective, then the defect detection depends mainly on the sensitivity of the sensor.

If the material is partially transmissive or reflective, then the background signal is amplified or detected in addition to the defect signal, so that the signal is difficult to separate from the background.

Signal to background and signal to noise ratio

• – Eine schmale Laserlinie einzusetzen. Das Signal-/Hintergrundverhältnis wird umso schlechter, je größer die angestrahlte Fläche ist.
• – Eine Wellenlänge für den Strahler (Laser) zu verwenden, bei der die zu überprüfende Materialbahn insbesondere in Form einer Folie stärker absorbiert.
• – Streulicht auszublenden.
• – Polarisationsänderung der Folie berücksichtigen.
• – Zusätzliche Eigenschaften der Objekte ausnutzen.

In order to improve the signal-to-background ratio and thus the signal-to-noise ratio, the following solution possibilities are available:

• - Use a narrow laser line. The signal / background ratio gets worse, the larger the illuminated area.
• - To use a wavelength for the radiator (laser), in which absorbs the material to be tested web more particularly in the form of a film.
• - hide stray light.
• - Consider polarization change of the foil.
• - exploit additional properties of the objects.

The simplest arrangement is when the object, in particular a web, is transparent only for defects, or reflect only the defects in light or dark field arrangement. Then there is a good contrast ratio and the detection limit is determined by the size and characteristics of the defect, the withdrawal speed and the bandwidth of the detection.

Is the material or web z. As an aluminum foil or coated with aluminum translucent material web, then z. B. absorbs visible light. Defects in the form of z. B. Microholes can then be detected theoretically up to the resolution limit of the light used.

The practical problem is to distinguish the low signal of the microhole (pinhole) from the noise signal. If the difference is large enough, the pinhole signal can be clearly distinguished by mathematical methods. However, if diffraction or scattering occurs at the defect, no distinction is possible even if the irradiated intensity is greatly increased. The main problem with prior art line systems is that the Airy pattern is mapped onto many cells and thus can no longer be distinguished from the noise or background signal. In addition, the geometric image of the defect on the detector via a focus optics, which collects the entire room light and contributes to the noise signal or background signal.

If it is detected in reflection or dark field arrangement, then the reflection properties of the defect compared to the carrier material have to be considered.

Small signal on a large surface

The situation becomes even more difficult when the web is transmitting and / or scattering. Then the signal to background and signal / noise ratio increases significantly ( 17 ). In addition, it must be avoided that the sensors saturate. An increase in the illumination intensity causes only an increase of the background signal and a deterioration of the signal / noise ratio.

As far as that goes 17 referenced, which represents a background / noise signal RS (not proportional to its size), and at a measuring point on the other hand, only slightly protruding defect location signal DS, which in 17 is shown separately again in an enlarged view. The detection time MD-t, TD-t in the withdrawal (MD) or in the transverse direction (TD) is reproduced on the x-axis, whereas the amplifier signal AS is drawn on the Y-axis.

In addition, if the illumination intensity is different over the corresponding partial path, then the pixel n is illuminated differently than the next adjacent pixel n + 1.

Object of the invention

It is therefore the object of the invention to overcome the disadvantages of the prior art and to provide an improved apparatus for optical defect detection.

The object is achieved according to the features specified in claim 1. Advantageous embodiments of the invention are specified in the subclaims.

• • die Bandbreite BW der Sensorik in MD-Abzugsrichtung der Materialbahn zu erhöhen, und damit die Abzugsgeschwindigkeit zu erhöhen;
• • Defekte zu erkennen, die der Fraunhofer oder Fresnel Beugung unterliegen;
• • Defekte in oder auf Objekten zu erkennen, die den Messstrahl zumindest teilweise transmittieren, streuen bzw. reflektieren;
• • Defekte zu erkennen, die ein geringes Signal-Rauschverhältnis aufweisen;
• • Defekte zu erkennen, die ein hohes Untergrundsignal aufweisen;
• • Ungleichmäßigkeiten der Beleuchtungsquelle auszugleichen.

The invention is based on a system and a method for measuring defects in order to:

• • to increase the bandwidth BW of the sensor in the MD take-off direction of the material web, and thus to increase the take-off speed;
• • detect defects that are subject to Fraunhofer or Fresnel diffraction;
• • detecting defects in or on objects that at least partially transmit, scatter or reflect the measuring beam;
• • detect defects that have a low signal-to-noise ratio;
• • detect defects that have a high background signal;
• • To compensate for unevenness of the illumination source.

In principle, the defect scans the beam waist (2w0) of a focused line laser L. This convolution, together with the high bandwidth of the scanning sensor and the knowledge of the take-off speed, gives a signal which makes it possible to draw conclusions about the type of defect, the shape and size.

In particular, in the context of the invention, a referencing of the measurement signal relative to the background signal during operation is possible. In systems according to the prior art, a referencing z. B. against a defect-free object 1 made in a particular area. This background signal is then electronically referenced. If the optical properties of the carrier material change during operation, then the actual defect signal can not be detected.

The invention will be explained in more detail with reference to drawings. In detail:

1 FIG. 2: shows a schematic plan view with respect to a measuring arrangement provided in the context of the invention for the detection of defects in a moving object, in the form of a material web in the embodiment shown; FIG.

2 : one along the direction of the arrow 9 in 1 reproduced representation of the embodiment according to 1 (omitting the rollers provided below the material web);

3 : A partial plan view of the embodiment according to 1 in enlarged reproduction;

4 a schematic plan view of a sensor arrangement, for example, using a plurality of photodiodes, which also serve to detect the background and / or scattered light signal, in order to achieve a better result on this basis with respect to a received defect signal;

5a FIG. 2 is a first schematic representation of the advancing movement of the material web to be examined and a small defect point displaced thereon with a movement speed in order to explain the principles and the mode of operation of the present invention; FIG.

5b a second schematic representation with respect to the advancing movement of the material web to be examined and thereon, located at a movement speed displacing large defect to explain the principles and the operation of the present invention;

6 a double-row measuring device having a plurality of PIN photodiodes;

7 : a schematic measuring arrangement in side view parallel to the object path plane;

8a a schematic enlarged fragmentary cross-sectional view parallel to the object path plane using a diaphragm device;

8b : a partial plan view of the embodiment according to 8a ;

9 a schematic arrangement of slanted detector and sensor elements;

10 : a schematic representation with respect to the relationships and results in obliquely positioned detector and / or sensor elements;

11 : a representation similar to 10 with respect to a modified measurement situation;

12 : a repeat presentation similar to 11 with regard to a further modification;

13 : a schematic representation of a measurement and evaluation structure using evaluation electronics;

14 : a graph relating to the resolution in the direction of movement of the object to be examined;

15 : a schematic representation of how a defect point to be examined moves on a moving object to clarify the limit values with regard to the expected defect position signal;

16 a further diagram with respect to the resolution in the transverse direction with respect to the scan area; and

17 : a partial reproduction of the underground and / or noise signal to be determined and a defect signal located above it which only slightly lifts off.

The invention will be described below first with reference to a basic measuring arrangement with reference to 1 and 2 explained.

In the optical detection, a material web 1 (For example in the form of a plastic film 1'' ) or another object 1 moved in the direction of the withdrawal direction MD through a device. The deduction direction is hereinafter referred to briefly as MD (Machinery Direction). Of course, the measured object can be fixed and the measuring arrangement can be moved. The basic measurement arrangement is therefore analogous to existing systems.

In the embodiment shown in the schematic plan view according to 1 For example, runs a material web 1 in the form of a plastic film 1'' via two offset in the MD direction offset from each other arranged rollers 3a and 3b whose axes of rotation 3'a and 3 'b are aligned transversely and in particular perpendicular to the withdrawal direction MD. Here are in the figures, insbesondre in 1 the deduction direction is abbreviated as TD by a corresponding arrow or vector MD and the transverse direction (transverse direction) extending therefrom. The thickness direction th is perpendicular to the plane and thus perpendicular to the plane E of the web 1 , wherein the perpendicular to the plane extending thickness vector th in 1 is shown as a circle. The thickness direction is referred to as abb abbreviated.

About the transverse direction TD of the material web 1 are one or more light sources L (for example, LED line elements, line laser) and sensor elements S above or below the material web 1 positioned in transmission or reflection arrangement (light or dark field arrangement). The same applies of course to other moldings, if they are to be examined accordingly. Occurs a defect 7 (Defect site 7 ), then a signal deviation is detected. The assignment of the defect in the transverse direction TD happens through the sensor location. The assignment in MD is done by an encoder signal, z. For example, a rotary encoder can be attached to a roller. In 2 is a side view of the arrangement according to the arrow 9 in 1 reproduced, so a representation parallel to the material web level E.

The concept of the direction of detection in the thickness direction th of the material web 1 is that a light source L, which may consist of one or more units, on the web 1 for example in the form of a plastic film 1'' in transmission or reflection, with both, collinear, waveguide configuration or dark field arrangement.

The lightwave L along the working width (preferably in the TD direction) consists of one or more individual light sources L1 stacked along the AB. It is off 3 also to see that the light source drawn there works in the form of a laser with a laser line length Llength. In other words, a plurality of such light sources and in particular lasers must be positioned side by side, ie offset transversely relative to one another, in order to cover and check the entire web width. With regard to the light sources and in particular of the laser, L or line light source L is also used to this extent by a transverse or TD light source.

In the transmission or reflection direction or in a dark field arrangement is also a detection device, which is also referred to below as a sensor or sensor device S. This sensor device can consist of one or more units Ss which emit the light emitted by the TD light source L and the material web 1 transmitted or reflected light detected. In principle, it is possible to refer to the light emitted by the laser and the transmitted light against each other, or to an additional light source.

The assignment of the defect in the direction of rotation via a timing and / or sensor, the MD movement of the object 1 detected. By means of an MD position detection device MD-E operating in the withdrawal direction MD, for example in the form of a marking system discussed below, the position, ie the position or the location of the detected defect or defect, can be detected and / or identified.

In order to keep the position fluctuation of a web low, known measures for stabilization, such as. B. with closely spaced rolls, air cushion, etc. are performed.

In order to reduce the ambient light and possibly light conduction in the carrier material, corresponding beam traps can be used, so that background and / or scattered radiation should be absorbed as far as possible.

Based on a simple experimental setup of the 2 and 4 the problem should be explained and the inventive solution to be pointed out.

A) Case studies with good signal / noise, background ratio

A laser L1 radiates in the normal vector of TD-MD plane E to an object 1 , and the transmitted light strikes a sensor or detector element SEs, which in the imaginary experiment should consist of five individual photodiodes PIN_A. The axis of symmetry of the laser should define the optical axis OA which hits the active area SE_AF of the PIN_An in the center.

a) Assuming the static case, in which exactly one defect, z. B. a micro hole 7 (pinhole), symmetrically centered to the active area SE_AF of a PIN_A, in an otherwise opaque object 1 , is depicted. For larger defects 7 the diffraction can be neglected and the sensor receives this through the web 1 passing or reflected light. So this PIN sees the max. Intensity of the beam source, which passes through the einhole or the intensity maximum of the diffraction. The intensity of the background signal in a web z. As an aluminum foil or a material coated with Al material web is low because z. B. visible light is absorbed and the rest of the laser radiation does not pass through the object. The defect detection then largely depends on the sensitivity of the sensor.

Significant, however, is the increasing diffraction phenomenon with smaller defects.

Due to the Airy relationship is subject to the defect 7 strong diffraction, the intensity returns to the fourth power of the decreasing half radius each, and the transmitted light undergoes a broad spatial intensity distribution at an Airy angle AW.

Suppose a defect occurs along a parallel line to the TD by OA z. B. at the PIN_An on.

The intensity at the PIN_An decreases sharply while the intensity of the surrounding PIN increases. In particular, when the active areas of the PIN are close to each other, such as the z. B. is the case with line arrays. The farther the sensor element is from the object, the stronger the effect becomes.

a) To make this simple static case detectable, the following methods can be used:

• - referencing by clocking (eg switching on and off) the laser source;
• - the use of a diaphragm PFS (see 7 ) between the object 1 and the sensor element SEs reduces the diffracted light and thus the background signal;
• The use of arrays with a smaller pitch size (small pitch) acts just like a diaphragm; but a 1: 1 mapping to the array is necessary;
• - furthermore, the distance could be reduced;
• - An increase in the laser power usually brings the opposite effect, since the diffracted radiation increases and also the noise signal of the laser increases; and
• - Between the photodiode PIN_An and at least one of the surrounding photodiodes (PIN), the difference signal can be formed with a balanced circuit; by this step, the background signal z. B. the photodiode PIN_An + 1 electronically subtracted from the signal of the photodiode PIN_An, and the defect is thereby resolvable; This innovative step proves to be particularly favorable.

b) static case with at least two defects along the TD axis passing through OA.

Assuming that each defect is symmetrical in the middle of PIN_An and PIN_An + 1, then the same detection method as for a defect can be used. Only the referencing must be against one or more other reference PIN, here z. B. the PIN_Aq, PIN_An - 1 and PIN_Am be made.

According to the invention, all PINs with the highest possible bandwidth are queried in parallel. The evaluation routine thus detects an increased signal at the two PINs compared to the other. In the simplest case, the local resolution in TD is determined by the size of the active area of the individual sensor (eg the PIN or the pixels). Are z. For example, if two defects are mapped to the active area of the PIN_An, then it simply increases its signal level and simulates a major defect. Remedy here creates the inventive skew of the sensors, as explained in the discussion of the dynamic cases.

c) detection of a defect in the direction of withdrawal or withdrawal MD

• c1) werden die PIN der Reihe n wie ein Array mit der Abtastrate l_r abgetastet, dann erhält man die gleiche Auflösung wie beim Stand der Technik diskutiert. Der wesentliche Unterschied besteht darin, dass die aktiven Flächen der PIN größer als die Pitch-Größe eines Arrays sind.
• c2) Nimmt man an, dass alle PIN mit einer Bandbreite zumindest >> l_r (also mit einer erfindungsgemäßen Bandbreite, die größer ist als die herkömmliche Ausleserate) ausgelesen werden können, dann erhält man die drei Signale der PIN_Aq, n, m. Der Defekt tastet die Lichtquelle ab, die 3 zeitlichen Signale kommen dadurch zustande, dass die aktiven Flächen voneinander getrennt sind.

Assuming that all active areas are uniformly illuminated, and the movement of a defect is parallel to the MD direction by CA at the speed v, the following cases are obtained depending on the sampling rate and method.

• c1), the PINs of row n are sampled like an array with the sample rate l_r, then the same resolution is obtained as discussed in the prior art. The main difference is that the active areas of the PIN are larger than the pitch size of an array.
• c2) Assuming that all PINs with a bandwidth at least >> l_r (ie with a bandwidth according to the invention which is greater than the conventional readout rate) can be read out, then one obtains the three signals of PIN_ Aq, n, m. The defect scans the light source, the three temporal signals are due to the fact that the active surfaces are separated from each other.

d) Basic principle Scanning the beam or focus waist with the defect

The basic principle is based on the 5a and 5b be clarified. A portion of the film of the working width ABa parallel to the transverse extension TD is irradiated with a line laser of length Llength and the beam waist 2w0 in the focal plane E. Occurs the edge of a defect at time t0 with the speed v in the focus area (as in the case of a small-sized defect spot (small micro hole 7 ' ) in 5a is shown), then the corresponding sensor element SE should provide a signal. The signal should then be detected until the defect has completely left the focus area again. In this assumed ideal case, therefore, the signal duration t2 = t1 + DP / v where DP is the diameter of a detected defect 7 for example in the form of a microhole 7 ' and v represents the relative direction of movement of the object relative to the line laser. In practice, the size of the defect is then determined via the half-width of the detector signal, a mathematical deconvolution and a calibration pattern for sensor adjustment. In practice, the sensor usually only responds within the beam waist 2w0, therefore it is necessary here to work with calibration patterns in a first setting ( 5a ).

If one assumes, on the other hand, that a major defect 7 with a larger sized micro hole 7 ' present (as in 5b is shown), then extends the signal duration, the half-width and the signal height accordingly, since in this embodiment, more light hits the sensor. The corresponding conditions are in 5b regarding the 102 shown.

In contrast, lies an equally large dimensioned thin point 7 as a defect 7 before, then reduces the signal level in this case, because less light is transmitted. Instead of the receiver signal 102 On the other hand, one obtains a lower intensity receiver signal in intensity 101 , as in 5b is shown. Suitable calibration patterns and teaching methods can be used to determine the type of defect, the defect size and the defect shape.

e) several defects in MD and TD direction

Are several defects on a line parallel to MD, the z. B. goes through OA, then the defects can be distinguished by the time sequence.

If several defects lie on a line parallel to TD, the z. B. goes through OA, then occur at the same time t0 to the PIN of the row n signals that differ by their signal duration and height depending on the defect. Meeting z. B. two defects at the same time on the sensor elements, d. H. the PIN_An, then increases accordingly the signal.

B) Case studies and inventive solutions for highly transparent or strongly scattering, strongly diffracting defects or strongly scattering objects

The resolution of the sensor depends on the transmission or reflection of the carrier material, ie, for example, the material web 1 from. The more transparent or reflective the material is, the larger the background signal becomes and the lower the background signal-to-noise ratio ( 17 ).

If it is not possible to increase this ratio, the signal can no longer be detected. So it is z. B. possible in the prior art, the wavelength of the light source to change so that they on an absorption band of the object 1 meets or other physical properties, such. B. Polarization properties of the object 1 exploited.

Subject to the defect 7 strong diffraction, then the intensity goes back to the fourth power of the decreasing half radius and the passing light experiences a broad spatial intensity Airy distribution. The intensity along the optical axis OA thus continues to decrease with respect to the background signal.

If the intensity of the laser source also fluctuates, the signal is no longer detectable. These fluctuations may be due to the state of the art (/ PCD Hobbs: Building Electro-Optical Systems, Wiley 2009, ISBN 978-0-470-40229-0 / and OCT optical coherence tomography) are compensated via a reference detection of the split laser beam. Usually basic circuits are sufficient for / DG Houser, E. Garmiere: Balanced Detection technique to measure small changes in transmission, Applied Optics, 33 (6), (1994) /out. Likewise, commercial circuits according to Newport, Thorlabs, New Focus, etc. can be used.

Ba) Assuming the experimental situation of a), then depending on the transmission or Reflection or scattering for defects <50 microns to consider the diffraction.

Assuming the beam source L or L1 has a sufficiently large extent, then the background signal is at least from another PIN photodiode z. For example, PIN_An + 1 was added in the vicinity of PIN_An.

As an inventive feature, the referencing of at least two detector elements (in this case the PIN_An with the surrounding PIN_An + 1) is used by a balancing circuit. As already mentioned, laser fluctuations are compensated by balanced photodiodes (Hobbs), although the principle has not hitherto been used for the detection of defect sites.

With the principle circuit according to the prior art to a so-called. Balanced detection (Balanced or auto-balanced photo-diode detection), it is then possible to better resolve the defect signal from the background signal and the partial diffraction intensity. The closer the individual detection elements are to each other, the better the effect becomes. In principle, both discrete photoelements such as arrays, photodiodes, CIS, etc., and position sensitive photosensitive devices (PSD position sensitive devices) with multi-channel resolution can be used.

If a defect in the photodiode PIN_An detected, then you get, depending on the bias of the PIN, z. B. a positive signal. The photodiode PIN_An + 1 detects the background signal and provides a negative signal when this PIN is driven in reverse polarity as the PIN_An. So if there is no defect, then the photodiodes PIN_An, n + 1 see the same background signal, the difference between the two signals is zero. If the difference signal of PIN_An and PIN_An + 1 is used in a balanced detection, then even the smallest difference signals can be detected.

Bb) Several defects on TD

Assuming that two defects occur symmetrically in the middle of PIN_An and PIN_An + 1, then the same detection method as for a defect can be used. Only the referencing software must be against one or more other reference PIN, here z. B. the PIN_Aq, PIN_An - 1 and PIN_Am be made.

If there is a defect at PIN_An and at PIN_An + 1an, then one receives a positive or negative signal of the respective PIN, the signal level of which depends on the size of the respective defect. The difference signal would thus indicate a defect for the PIN at which a higher signal is applied. In this case, therefore, the respective difference signals between the PIN_An or PIN_An + 1 are formed with another PIN_Reference.

According to the invention, all PINs with the highest possible bandwidth are queried in parallel. The evaluation routine thus detects an increased signal at the two PINs compared to the other. Of course it is also possible to set a PIN_Reference. Since the individual sensor elements are stacked above the AB, at least one PIN is provided for the overlap. It therefore sees the same signal as the leading PIN, but later.

Bc) Detection of defects in MD

One of the main problems in the detection of small defects is the high background signal of the transmitting or reflecting or scattering object. The aim is thus to keep the background signal as low as possible in order to be able to resolve the defect signal. As stated earlier, small defects with systems that operate at a relatively slow sampling rate can no longer detect the defects.

To keep the background signal as small as possible, a focused line laser is selected with a laser waist, on the one hand gives a small background signal to the sensor elements, but on the other hand is still wide enough to allow the folding of the defect with the laser waist within the bandwidth of the sensor.

In order to keep the diffraction angle and thus the spatial distribution through the defect as small as possible, the sensor element SE _{S is brought} close to the bottom of the object, and there is a diaphragm with an opening on the order of the laser waist 2w0, attached (see 8a and 8b ). Of course, suitable imaging optics can also be interposed. In this case, the aperture can be provided with an opening in the order of magnitude corresponding at least to the laser waist 2w0.

With a high bandwidth BW of the sensor system with the condition BW >> l_r, it is possible that the defect scans the laser waist 2w0. The obtained signal is mathematically a convolution of the defect with the laser waist, together with the feed rate v can then be concluded on the size of the defect. In addition, if the signal is digitized, then you can use the signal level and the waveform on the type and shape of the defect infer. In principle, the same teaching methods as in array systems according to the prior art are used here.

C) Summary of the inventive features with multiple defects in the MD and TD directions

Referencing the photodiode PIN_An with the surrounding photodiode PIN or a reference photodiode PIN through a balancing circuit:

Signaling of each sensor sub-element, z. B. PIN_An and PIN_An + 1 with high bandwidth, and simultaneous difference of at least two PIN. The difference is formed between two series-connected PIN, each of which is supplied with reverse biased voltage.

Reduction of the background signal:

The size of the active surface of the detector element or the optional focus optics is selected so that at least the focus width of the light source can be imaged thereon. This can be achieved by suitable measures such. B. screens are guaranteed. The area of the active surface should be selected so that as little stray light or diffracted light strikes the active surface, in order to avoid overloading of the sensors or to reduce a high background signal.

bandwidth:

Compared with the prior art, in the invention, each detector element Rr, r ∈ N, with its maximum bandwidth BW and not with the line rate (line rate) l_r read (as was previously possible only in the prior art).

Folding:

In principle, the defect scans the focus or beam waist. By mathematically unfolding the intensity profile with the feed rate, the mean diameter of the defect can be determined.

Embodiments and basic structure of a detection unit

A corresponding basic structure of a detection unit will be described below with reference to FIG 3 . 6 and 7 explained in greater detail.

It shows 7 a light source L1, which has a focus optics FO _in on the object 1 radiates, so that the beam or focus waist 2w0 is close or ideally in the object plane E, so for example in the case of a material web 1' in the plane E of this material web 1' , because in this object plane E are the defect points to be located 7 especially in the form of microholes 7 ' , Coherent light sources L1 whose power can be regulated are preferably used, and which can optionally be used in pulse mode. Preferably, line lasers of the beam length Llength with a low coherence angle and a uniform intensity profile are used.

So z. B. a web 1' also in the form of a plastic film 1'' Only a slight variation in position (that is to say fluctuations perpendicular to the object plane E) can be carried out or executed, additionally guidance and stabilization units SLTin, out can be provided above and below the object, preferably over the entire width of the material web to be examined 1' extend. The guidance and stabilization units, which may only be provided on one or both sides of the material web plane E, may additionally be designed as jet traps in order to reduce the ambient light and / or the light conducted through the object.

The object should be able to be passed through this sandwich-type guiding and stabilizing units without generating any friction or damage.

The receiver units SE and / or the optional focus optics units FO _out on the receiver side are as close as possible to the rear side of the object to be examined 1 (Web 1' ) after the actual object 1 to place. This is to prevent that the Airy area widens greatly and the intensity of the subsurface is no longer separate.

The corresponding optics and / or sensor units are preferably integrated in the stabilization device and an optionally provided light trap, as described in detail in FIG 7 can be seen.

In the exemplary embodiment shown, it is thus also possible for the material web to be examined to be tested on the receiver side 1' an optional imaging optics FOLG be provided, which may include, for example, photoconductive elements (GRIN, Selfoc etc.).

Subsequently, one or more receiver-side optional optical devices PFS may also be provided, for example diaphragms, polarization devices, filters, devices for reducing light scattering, etc.

Finally, the mentioned receiver units SE are provided, into which the light emitted by the light generator, in particular in the form of a laser, is received, for example by the frequently mentioned photodiodes, which are arranged on the same optical axis (-line) as the light sources.

With a corresponding structure, therefore, a detection device could come into play, which comprises a plurality of juxtaposed photodiodes PIN in a row extending in the transverse direction TD, as described, for example, with reference to FIG 6 is shown (where in 6 a variant described in more detail below with two adjacent rows of sensor elements SE in the form of photodiodes PIN_A, PIN_C etc. is shown).

In the exemplary embodiment shown, a further laser-side optional device PFOin is subsequently provided in the beam direction of the focus optics FOin, for example in the form of a polarizer, a filter, etc.

In the presentation according to 7 is the laser L1 subsequently inserted into the beam path an optional beam splitter SGU, z. B. may consist of a deflection or prism means, a scanning unit, etc., whereby the laser beam is split. As a result, in two laser beam planes LE1 and LE2 in the withdrawal direction MD offset line regions on the guided through the laser beam at the speed v material web 1' be checked and scanned. Therefore, the entire explained transmitter-side optical device FOin, PFOin as well as the receiver-side optical device FOLG, PFS are designed so that the same conditions exist for both laser beam phenomena LE1 and LE2. Accordingly, two rows of sensors SEn and SEm are arranged side by side, which form the sensor device. The optional dual laser plane option is chosen to minimize the cost of the laser source and to ensure overlap of the active areas with the discrete design of the sensor elements SE (single PIN arrangement and non-integrated array design).

Preferably, a line laser L1 with a small beam waist (2w0) and a line length of the same intensity (beam convergence <1 °) is used. The intensity profile is focused on the object by the focus optic FOin 1 displayed.

Are discrete sensors such. For example, PIN photodiodes, CIS sensors (contact image sensor based on CMOS technology), CMOS sensor, CCD sensors, etc., it depends on the immediate distance of each active area and the Size of the defect to be detected 7 Whether a second detector row LE2 must be provided next to the first detector row LE _1, in each case, as sensor unit SE in order to ensure complete detection (FIG. 6 ). Are z. B. commercially available PIN photodiodes used, then either two light sources must be used, or it is a beam splitter or optical switching unit SGU provided, as described above. For this purpose, elements according to the prior art, such as optical beam splitters, fiber couplings, AO, PEM, etc. are used.

Lying with the object 1 or the optical defect specific optical properties that can improve the signal / background ratio, additional optical elements are used. Absorbs the object 1 or the defect at a certain wavelength, then z. B. edge or bandpass filter can be used. The same applies to any polarization properties that can be implemented with appropriate optics. Such a procedure is known in the art and shown in detail in the prior art.

All these measures are also used on the receiver side to improve the signal-to-noise background ratio.

Are z. B. standard PIN photodiodes used, then the active surface z. B. by a diaphragm (in 7 . 8a and 8b denoted by the general designation FOLG), because the scattered light increases the background signal. This is for example in a representation parallel to a film plane E in 8a and in a partial plan view in FIG 8b played. For active areas of small pitch sizes z. As with arrays, this component can also be omitted. For small pitch surfaces, on the other hand, an array adjustment device is often required to hit the focus line.

If the zero balance between PIN_An and PIN_An + 1 is not done electronically, z. B. at the active area of the PIN_An the zero balance with an ND filter PFS be made.

As optional optics also filters and / or polarization optics can be provided on single or all active surfaces.

Instead of the recipient z. B. PIN photodiodes directly under the object 1 To arrange, it is possible to direct the signal via optical fibers to the receiver. For this purpose, state-of-the-art optics such as fiber optics, GRIN, collection and magnification optics, etc. are used.

Inclination of the sensors to increase the TD resolution

The individual detector elements SEs, s ε N, are staggered overlapping to cover the entire working width AB. For most applications, it is sufficient for the location of the defect for the to know respective active area of a sensor within the sensor element. Is z. B. a sensor element SEs 10 PIN with respective active lengths of 2 mm, then sufficient in practice, this TD resolution of 2 mm for quality control.

For a higher TD resolution, the sensor elements SE can be tilted, as in an excerpt plan view of 9 and in a graph according to 11 is reproduced.

The simplest case with individual discrete sensor elements, such as single PIN photodiodes, arrays, CIS, etc. is present if at time t ₀ one or more defects occur at the same location x (t ₀ ) along in the transverse direction TD. Now runs z. B. the movie 1'' at a speed v by the sensor S with the receiver device R _r , then the defect at x (t ₁ ), then x (t ₂ ), ... x (t _n ) is detected in sequence first. An MD-directional encoder and temporal discrimination may cause the TD location of the defect 7 be determined (taking all suitable measures for the exact position determination of a defect in the direction of movement MD of the object 1 , ie in the withdrawal direction MD of the material web shown here 1'' are suitable; Reference is also made in this regard to known methods and devices for this purpose) ( 10 ).

Referencing relative to the background signal occurs as with detector elements that are staggered parallel to TD. PIN_An is referenced to PIN_An + 1 or to PIN_Referenzj.

If several defects coincide with the sensor line, then at time t = t ₁ there are no differences in the signal sequence, which is the same as those with sensor elements SE aligned parallel to the transverse direction TD. Only the spatial resolution in the transverse direction TD can be improved. The corresponding conditions are in the graph according to 11 played.

If sensor elements with a long active surface z. B. PSD, elongated PIN or light-guiding elements used, then results in the impact of multiple defects at time t1, a larger signal height. The spatial resolution is then obtained by a second sensor element, which then also serves as a reference element for the background signal. From the illustration according to 12 the corresponding conditions can be taken.

data collection

One possible setup for data collection is in 13 played. This in 13 to be seen electronic evaluation device AE includes various functional and / or assemblies, as in 13 are shown schematically in their structural interconnection and structure.

The photodiodes are preferably operated reverse-biased. For a balanced operation, the difference is formed by two PINs (variants: cathode of PIN_An and the anode of PIN_Cn, K of PIN_An and A of PIN_An + 1, K of PIN_An and A of PIN_Reference, etc.), and a transimpedance amplifier (TIA) fed. Of course, it should also be possible to take a difference signal to other PIN photodiodes or a reference sample online. When overlapping two individual detector elements, one PIN of the row n is for safety's sake ever matched with a PIN of row m. If these two PIN photodiodes give a zero signal, they can, for. B. serve as a reference for the other PIN photodiodes.

To ensure this, a multiplexer can interconnect a single PIN against each other or with respect to a reference sample.

These balanced signals are fed to an A / D converter to achieve the required bandwidth must be used here in parallel operation. The control work is done by an FPGA, which can be adapted to the different process processes and applications.

The data is read out via FIFO buffer to the computer.

13 shows an example of the electronic evaluation, in which - as based on 6 is shown - in addition to the actual laser measurement line n, the optionally provided second laser measurement line m is provided. Therefore, the in 13 shown electronic evaluation device quasi also shown in its double structure, namely with one device DSP with associated multiplexer and an analog / digital converter A / D and a device FPGA for the parallel processing of the signals including the mentioned FIFO latches. Would you like the optional second laser measurement line m in 6 or in 13 omit the structure would be halved in this respect. However, of course, a computer must always be provided to control and evaluate the entire device.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 7224447 [0008]
• DE 10251610 [0009]
• US 2006010847 [0009]
• WO 2002/73173 [0010]
• DE 3850871 [0010]
• DE 10124943 [0010]

Cited non-patent literature

• PCD Hobbs: Building Electro-Optical Systems, Wiley 2009, ISBN 978-0-470-40229-0 [0088]
• DG Houser, E. Garmiere: Balanced Detection Technique to Measure Small Changes in Transmission, Applied Optics, 33 (6), (1994) [0088]

Claims (19)

Device for detecting defects ( 7 . 7 ' ) on an object ( 1 ), in particular defects ( 7 ) in the form of microholes ( 7 ' ) in a material web ( 1' ) or a movie ( 1'' ), with the following features - at least one primary light source (L, L1) is provided, by means of which a light beam can be generated, - by means of the at least one light source (L, L1) the object ( 1 ) with a sensor and / or detection device with sensor or detector elements (SE _S , SE) for detecting the emitted from the at least one light source (L, L1) light beam, after this from the zu examining object ( 1 ) was reflected in reflection arrangement or the object to be examined ( 1 ) in the transmission direction, and - there is also an evaluation device (AE) for evaluating the defect location ( 7 ), characterized by the following further features - the device with the evaluation device (AE) a) comprises for the sensor or detector elements (SE _{S ,} SE) a Balanzierschaltung such that under difference between one of a sensor or detector element (SE _S , SE) and at least one measurement signal generated by an adjacent or referenced sensor or detector element (SE _{S ,} SE) generates a balanced signal for defect location evaluation, and / or b) comprises a measurement setup such that the defect sites ( 7 . 7 ' ) the beam or focus waists (2w0) which are incident on the surface of the object to be examined ( 1 ) are traversed at a speed (v), wherein a defect-dependent measurement signal (convolution) by a in the beam path after the defect point ( 7 . 7 ' ) arranged sensor and / or detector element (SE _S , SE) from the entrance of the defect site ( 7 . 7 ' ) is generated in the beam or focus waist (2w0) until leaving the beam or focus waist (2w0). Apparatus according to claim 1, characterized in that the device is constructed so that the sensor or detector elements (SE _S , SE) with at least 60%, preferably at least 70%, 80%, 90% and preferably with 100% of their maximum bandwidth be read out. Device according to Claim 1 or 2, characterized in that the at least one further sensor or detector element (SE _S , SE) provided for a sensor or detector element (SE _S , SE) is arranged directly adjacent. Device according to one of claims 1 to 3, characterized in that the device and the associated measuring arrangement and / or the sensor or detector elements (SE _S , SE) for the detection of defects ( 7 . 7 ' ) are formed such that the beam or focus waist (2w0) of the light beam of the beam or focus waist (2w0) of the light beam imaged by the at least one light source (L, L1) at the location of the sensor or detector element (SE _S , SE) by the corresponding sensor or detector element (SE _S , SE) is measurable. Device according to one of claims 1 to 4, characterized in that for the measurement of a defect ( 7 . 7 ' ) sensor or detector elements (SE _S , SE) is assigned a diaphragm device which is designed with respect to the aperture or aperture so that the maximum beam or focus waist (2w0) via the sensor or detector elements (SE _S , SE) the received light beam is measurable. Device according to one of claims 1 to 5, characterized in that the at least one light source (L, L1) and the associated transmission and / or measuring device for detecting the at least one light source (L, L1) emitted light beam relative to the examining object ( 1 ) is movable in a direction of movement (MD) with a relative object speed (v), and that the detector and / or sensor elements (SE _S , SE) provided for measuring the background and / or scattered radiation in the direction of movement (MD) and / or sensor elements (SE _S , SE) upstream or downstream, which are used to measure a defect location ( 7 ) are provided. Device according to one of claims 1 to 6, characterized in that the sensor or detection device or the sensor or detector elements (SE _S , SE) for the detection of defects ( 7 . 7 ' ) comprises at least two in the direction of movement (MD) offset from each other and transverse or perpendicular thereto in the transverse direction (TD) extending measuring ranges or series of measurements (n, m). Apparatus according to claim 7, characterized in that the two transversely extending (TD) transverse or perpendicular to the relative direction of movement (MD) extending measuring ranges comprise two separate light sources (L, L1), preferably in the form of two laser arrays. Apparatus according to claim 7, characterized in that the at least two measuring areas comprise only one laser arrangement as a light source (L, L1), wherein further in the radiation path of the at least one light source (L, L1) emitted light beam, a radiation divider is arranged, whereby two in the direction of movement (MD) mutually offset detection areas or lines can be generated. Device according to one of claims 7 to 9, characterized in that for the sensor or detector elements (SE _S , SE) based on their overall size have a smaller active area, wherein in a series (n) sensor or detector elements (SE _S , SE) are arranged offset to one another in the transverse direction (TD) relative to the sensor or detector elements (SE _S , SE) arranged in another preferably adjacent row (m), preferably by half the dimension of the transverse extension width of the sensor or detector elements (SE _S , SE), in such a way that the active areas of the sensor or detector elements (SE _S , SE) overlap in the transverse direction (TD). Device according to one of claims 1 to 10, characterized in that for the at least one light source (L, L1), a line laser is provided which generates a fanned light beam in a plane which transversely and in particular perpendicular to the relative object speed (v ) moving object ( 1 ) runs. Device according to one of claims 1 to 10, characterized in that as the light source (L, L1) preferably a laser or a line laser is provided, the beam or Fokustaille (2w0) has a size which is smaller than the respective extent of the active Surface of the sensor or detector element (SE _S , SE). Device according to one of claims 1 to 12, characterized in that there is further provided a light trap for reducing the background and / or scattered radiation. Device according to one of claims 1 to 13, characterized in that an object stabilization device is provided, through which the object ( 1 ) is moved with its object speed (v), about which a change in position transversely and in particular perpendicular to the direction of movement (MD) of the object ( 1 ) is prevented. Device according to one of claims 1 to 14, characterized in that depending on the optical properties of the object to be examined ( 1 ) Edge or bandpass filters, polarization devices are provided. Device according to one of claims 1 to 15, characterized in that the sensor or detector elements (SE _S , SE) have a greater longitudinal extent than transverse extent, and that these sensor or detector elements (SE _S , SE) relative to the relative direction of movement ( MD) of the object ( 1 ) are oriented at an angle greater than 0 ° and less than 90 °, preferably at an angle between 10 ° to 80 °, 20 ° to 70 °, 30 ° to 60 °, 40 ° to 50 ° and in particular by 45 °, and that the sensor or detector elements (SE _S , SE) are arranged so that their active area overlap in the transverse direction TD. Device according to one of the preceding claims, characterized in that the sensor or detector elements (SE _S , SE) comprise discrete sensors, in particular PIN photodiodes, CIS sensors, CMOS sensors, arrays or CCD cameras. Device according to one of claims 1 to 16, characterized in that the sensor or detector elements (SE _S , SE) comprise position-sensitive photo elements (PSD), in particular with a multi-channel resolution. Device according to one of the preceding claims, characterized in that the electronic evaluation device (AE) is designed so that above the data of the sensor or detector elements (SE _S , SE) are processed and evaluated in parallel operation.

Images