Patent Application
20100284027
This invention relates to the field of measurement techniques, and particularly to an optical system and method for measurement of parameters of so-called Through-Silicon Vias (TSV).
Packaging technology for Integrated Circuits (IC) in the semiconductor industry undergoes increased development in order to satisfy a need for miniaturization and/or mounting reliability. Wafer level processing (WLP) techniques have been developed to allow various features of IC packages to be formed within a wafer before the wafer is diced. For instance, certain WLP techniques are used to form device interconnection features together with other wafer processing steps, thereby avoiding the need to form wire bonding after IC chips are diced.
In general, such WLP techniques allow IC package manufacturing processes to be streamlined and consolidated. Moreover, WLP techniques can generally be performed in parallel on a plurality of IC chips arranged in a matrix on the wafer, thereby allowing a plurality of IC chips to be formed and tested while still in a wafer stage. By performing WLP techniques in parallel across a plurality of IC chips, IC package manufacturing throughput is increased and the total time and cost required to fabricate and test IC packages is decreased accordingly. In addition, by forming features such as device interconnections at the wafer level, the overall size of IC packages can be reduced.
One of the WLP techniques used to form device interconnections involves the formation of a through silicon via. A through silicon via (TSV) is usually formed by creating a hole (via-hole) through a semiconductor substrate and/or various material layers formed on the substrate, and then forming a penetration electrode in the hole. The penetration electrode may be connected to internal features of an IC chip such as signal terminals, data transmission lines, transistors, buffers, and so on. In addition, the penetration electrode may be connected to features external to the IC chip, such as a PCB, via an external terminal.
Depending on the type of process in use, via-holes could be formed in different layered stacks of materials in wafers and other substrates. The holes are typically formed by etching based on Reactive Ion Etching (RIE) or laser drilling by ablation. The ion etching can be performed by a variety of processes optimized for materials, etch rate, sidewall slope, smoothness and other parameters. A well known method of etching is the Bosch method which is based on alternating steps of semi-isotropic etching and deposition.
In the so-called “via-first approach”, the holes are first formed in silicon, by etching using an etch mask and photoresist and/or other harder mask materials. The mask materials layers are relatively thin layers on thick silicon. If the via-holes are formed by laser-drilling, no etch mask material is used, the holes being formed only in the thick silicon. Via-hole diameters can range from the one micron scale up to tens of microns, with depth to diameter aspect-ratios going from 5 or less up to 30 or more. The minimum pitch that can be implemented between the holes is a critical parameter for minimizing the distance that electrical signals have to cover between the vertically stacked ICs. The minimum pitch is usually a small multiple of the minimum hole diameter.
In the so-called “via-last approach”, the holes are formed in the wafer backside through the silicon, until coming up to the conductive material on the wafer frontside. In this case, in addition to the possible etch mask materials on the silicon as in the via-first, the bottom of the hole (i.e. near to the wafer frontside) is formed in a different, possibly conducting, material such as copper. The layered stack can thus include possible masking layers, silicon sidewalls and a conductive bottom.
An additional option for the via-last approach is based on etching or laser drilling through the full stack of materials on the wafer frontside including dielectric insulating materials. The bottom of the hole is deep in the silicon substrate. The layered stack can include in this case, possible masking layers, sidewalls of various dielectric materials, sidewalls of silicon and bottom of silicon. In via-last, the holes are usually formed with dimensions and pitch (e.g. tens of microns) larger than in the via-first, usually targeted to be connected with the underside of previously formed copper pads on the front surface of the wafer.
The present invention provides a novel optical system and method for obtaining at least one of a cross-section profile, depth, width, slope, undercut and other parameters of via-holes by non-destructive technique. It should be understood that to control etch processes and subsequent coating processes, via-holes profile has to be determined. For example, controlling the depth of the holes is critical for reaching a correct vertical location. The ability to effectively coat the inside of the holes, usually with barrier layers, and subsequently fill them with a conductive material, depends on the geometrical profile of the holes.
To control the profile cross-section of the holes, a physically destructive cleavage or etch is usually performed. The cross-section is then exposed to imaging by optical-microscope, scanning-electron-microscope or scanning-ion-microscope.
The present invention enables determining one or more parameters of at least one via-hole in a structure by providing an optical system comprising inter alia an illumination system for producing at least one light beam and directing it on a sample in a region of the structure containing at least one via-hole. The light reflected from the via-hole (or holes), is collected by a detection system comprising a detector. The detection system is configured and operable to collect a pattern of light reflected from the illuminated region, the light pattern being indicative of one or more parameters of the via-hole. The pattern of the reflected light incident on the detector is analyzed by a control system connected to the detection system and one or more parameters of the hole(s) are found. The control system comprises a memory utility for storing a predetermined theoretical model comprising data representative of a set of parameters describing via-holes reflected pattern, and a data processing and analyzing utility configured and operable to receive and analyze image data indicative of the detected light pattern and determine one or more parameters of the via-hole. It should be noted that all the subsequent description relating to the measurement of a single hole is applicable to multiple holes whether placed in arbitrary locations or in a repetitive array structure.
In some embodiments, the detection system comprises a light sensitive surface located in a far field relation or in a Fourier relation with respect to a sample surface.
It should be understood that the incident light beam on the sample undergoes different reflections in the hole depending on at least some of the following parameters: the incident-beam's tilt from normal, hole opening shape, aspect-ratio, sidewall slope, sidewall slope variation, bottom rounding, surface roughness, surface absorption, surface coating etc.
The light pattern reflected from the sidewalls effectively describes the shape and parameters of the sidewalls and enables their analysis. For example, if the bottom of the hole has a high degree of rounding, the light beam reflected from the bottom effectively spreads into a two-dimensional fan which can probe the sidewalls, even when the incident beam is normal to the surface of the sample. If the bottom of the hole is essentially flat, controlled tilt of the incident beam is required to obtain information on the sidewalls.
In some embodiments, the illumination system comprises a plurality of light sources. The light sources may emit at different angles of incidence on a sample surface, the detection system collecting a plurality of light patterns corresponding to different angular positions. At least two of the plurality of light sources can produce light beams at different wavelengths.
The illumination system may comprise a light source associated with an aperture configured to shape the angle incident on a sample surface.
In some embodiments, the illumination system comprises an array of LEDs or lasers.
The illumination system may comprise an objective lens configured to focus a plurality of beams on the sample, the array being located in the back-focal plane of the objective lens, the back-focal plane having a Fourier relation or a far field relation with the sample surface plane. The array may also be located at a plane having a Fourier relationship or far field relationship with the sample surface plane. The array may be imaged to a plane having a Fourier relationship or far field relationship with the sample surface plane. The array may be arranged in a grid-like pattern.
The one or more parameters of the via-holes may be selected from the followings: geometrical profile, cross-section profile, depth, width, slope, undercut, hole opening shape, aspect-ratio, sidewall slope, sidewall slope variation, bottom rounding, surface roughness, surface absorption, surface coating of the via-holes.
In some embodiments, the system is configured and operable for determining parameters of multiple holes distributed in a sample in arbitrary locations or in a repetitive array arrangement.
The detection system may comprise a lens configured to collect the pattern of the reflected light and to image the pattern onto the detector.
The system may comprise at least one of polarizer or spatial filter configured to block a portion of the light pattern reflected from the region surrounding the via-hole.
According to another broad aspect of the present invention, there is provided a method for determining parameters of at least one via-hole. The method comprises providing data indicative of an initial theoretical profile of at least one via-hole having a set of parameters and a model image based on the theoretical profile; the set of parameters describing via-holes reflected pattern and being based on morphological characteristics of an image of a sample to be analyzed; illuminating a region of a sample containing at least one via-hole; collecting a pattern of light reflected from the region and being indicative of one or more parameters of the via-hole; receiving and comparing an image data indicative of the detected light pattern and the model image and determining at least one parameter of the via-hole in the sample.
In some embodiments, the method comprises determining a degree of correlation between the image data indicative of the detected light pattern and the model image and when the degree of correlation is beyond a predefined range, generating as new set of parameters based on the differences between the images.
The method may comprise varying at least one illuminating parameter and collecting a plurality of patterns of reflected light corresponding to the variation and comparing each image data indicative of the detected light patterns with a corresponding model image. The illumination parameter(s) is/are selected from the followings: angle of incidence of a light beam on the sample surface, wavelength of incident light beam, divergence angle of incident light beam, rotation angle between an illumination source illuminating the region of the sample and the sample surface, polarization of the incident light beam.
The method may comprise varying the angle of incidence of the light beam on the sample surface and detecting the reflected light pattern as function of different incident angles, being indicative of the morphology of the shape of the hole.
In other embodiments, the method comprises varying the divergence angle of incident light beam and detecting the reflected light pattern as function of different divergence angles, varying the resolution of the reflected light patterns and the accuracy of the parameters of the via-holes.
In other embodiments, the method comprises varying the wavelength of incident light beam and detecting the reflected light pattern as function of different wavelengths, being indicative of the depth and the pitch of scalloping profile of the walls' hole.
The parameters of the theoretical profile may be selected from at least one of the following parameters: hole opening shape, aspect-ratio, sidewall slope, sidewall slope variation, bottom rounding, surface roughness, surface absorption, surface coating.
The method may comprise rotating the plane of incidence of the light beam in relation with the plane of the sample surface by rotating at least one of a sample, a light source and an aperture located in front of the light source and a polarizer.
The method may comprise characterizing thickness distribution and geometrical profile of a coating layer deposited in the via-hole by performing measurements before and after the coating, and by analyzing the differences between the measurements.
In some embodiments, the method comprises segmenting the image data indicative of the detected light pattern into different profiles using the morphological characteristics of the image. The morphological characteristics are selected from symmetry of the image in relation to at least one incidence angle, narrow-angle spread of reflection angles around the incident angle indicative of side-wall angles, number of rings in the reflected pattern of light indicative of the aspect ratio of the hole, portion of the reflected pattern of light at high angles indicative of a slight bottom rounding, reduced or hazy light pattern as indicator of surface roughness.
The method may comprise collecting a pattern of light scattered from the region and analyzing an image data indicative of the scattered light pattern indicative of the shape and the aspect ratio of the holes.
In order to understand the invention and to see how it may be carried out in practice, embodiments will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:
Reference is made to
Reference is made to
The incident light beam on the sample undergoes different reflections from interfaces in the hole depending on at least some of the following parameters: the incident-beam's tilt from normal, hole opening shape, aspect-ratio, sidewall slope, sidewall slope variation, bottom rounding, surface roughness, surface absorption, surface coating etc. In some embodiments, to enable the analysis of at least one of the parameters of the via-holes, measurements have to be performed by varying at least one parameter of the above list, and the difference between the measurements is used to determine the value of at least one corresponding parameter.
In one embodiment, the angle of the incident light beam is varied and the measurements are performed at different angles. The illumination system 102 may comprise multiple light sources (e.g. small-area sources) at various off-axis angles enhancing parameter sensitivity and/or simplifying the analysis. The range of off-axis angles for obtaining optimal information content depends on the shape of the hole and especially on the aspect-ratio of the hole. The larger the tilt from normal incidence, the larger the number of reflections the light rays undergo within the structure.
The multiple light sources can be operable simultaneously or separately. The multiple light sources can be of the same wavelength or of different wavelengths. For example, in order to collect the information from different angles simultaneously, the illumination system may comprise multiple light sources of different colors associated with a color CCD camera.
The shaping of the incidence angles on the sample may also be produced by a large area source associated with an aperture. This aperture can be static, scanned or switchable. The light source can thus be scanned through a range of angles, and a series of images can be captured at successive angular positions. This configuration can be useful inter alia if the examined structure has sharp angular variations in wall angle, especially re-entrant profiles, resulting in sharp variations in the reflected light pattern versus the incidence angle. Alternatively, the scanning can be achieved by using an extended wide-band light source, placing a spatially graded spectral filter in front of it, and capturing images through a time variable spectral filter, such as a filter wheel, synchronized with image capture on the detector.
The illumination system 102 can also be implemented by an array of small sources e.g. an individually addressable Light Emitting Diode (LED) array, illuminating a relatively large region of the structure. For simplifying the determination of the parameters of the via-holes, such array can be placed in a plane having a far-field or Fourier relation to the sample surface and conjugates with the light sensitive detector surface.
Reference is made to
The divergence angle of the incident beam on the sample can also be varied. A relatively small divergence angle can result in sharper, more distinguishable features on the detector to the detriment of reduced light pattern intensity and thus increased measurement integration time. A larger divergence angle reduces the measurement time but can degrade the resolution of features in the image on the detector and thus reduce the ability to accurately differentiate parameters of the hole profile. The divergence angle can be optimized during the measurement set-up process by varying the divergence angle to the largest possible where the resolution in the image is limited by small scale scattering effects and/or the resolution of the optical system.
As described above, the detector 104 is located in the optical system 100 at a position which enables capturing the light pattern reflected and scattered from the sample surface. The detector 104 can be placed at a distance from the sample in order to capture a far-field image of the light pattern. Alternatively, an optical element can be placed in front of the detector 104 to enable more flexibility in locating the detector. This can preferably enable construction of a better optimized or more compact optical system.
A lens 210 can be used to collect the reflected light pattern and image it onto the light sensitive surface of the detector 104. Optionally the lens may be configured so that the image on the detector is of a Fourier relation to the sample.
In some embodiments, a polarizer can be located in the optical path before or after the light beam interaction with the sample. A polarizing beam splitter/combiner can be used. Alternatively, two crossed polarizers can be located in the optical path before and after the sample to enable blocking of the light pattern reflected directly from the region of the structure surrounding the hole.
In other embodiments, the plane of incidence can be rotated in relation to the axis of the sample (azimuth). For non-circular holes, e.g. holes with rectangular openings, the rotated plane of incidence can be used to obtain information on the via-hole(s) shape. The rotation can be achieved by rotating the sample, and/or the source and/or an aperture in front of the source. The rotation can be combined with possible rotation of a polarizer(s). For holes having a substantially circular opening, the rotated plane of incidence can be used if the holes are closely spaced in a grid-like arrangement and the orientation of the diffraction pattern of the holes arrangement can give additional information for the analysis.
In some embodiments, a spatial filter (e.g. field stop) is included in the optical path at a position which is optically conjugated to the sample surface to enable limiting the lateral extent of the measured light beam interaction with the sample, the spatial filter being located either in the incident beam path or in the optical path between the sample and the detector.
In some embodiments, an additional imaging detector can be added to the optical system at a location conjugate to the plane of the sample surface to enable verification of the measurement location. For example, a CCD camera can be added to provide an image of the sample surface to enable alignment on a preferred measurement site by means of pattern recognition. The illumination for the pattern recognition can be performed by separate means or by utilizing the incident light beam used for the hole measurements.
Due to the large scale size of the hole dimension (diameter), e.g. multi-micron range, with respect to the wavelength of the light beam, the incident light beam can be incoherent. This depends on the coherence length of the light source used and the numerical aperture and other parameters of the illumination system. A system using incoherent illumination is not sensitive to the arrangement of holes whether in a random or in a grid-like pattern and is not subject to speckle effects. Low coherence or to incoherence can simplify the modeling of the reflection from the sample due to the absence of interference effects.
On the other hand, coherence can provide additional information on the structure due to such interference effects. Additionally, using a highly collimated light source with small extent in the Fourier-plane or far-field relation, reduces the possible smearing of the reflection pattern.
Moreover, the reflected pattern of an incoherent beam, with wavelength considerably smaller than the measured structure, provides only angle-based information, which is scale invariant, thus enabling the determination of the shape of the hole but not of its absolute size.
If the lateral coherence length is larger than the diameter of the hole and/or the longitudinal coherence length is larger than the depth of the hole, then the reflected pattern can contain diffraction and interference effects. Based on knowledge of the wavelength and materials of the structure, these effects can enable determination of absolute geometrical values and not just relative ones.
If the illumination system produces at least two incoherent light beams at different wavelengths (e.g. using at least two different light sources), the pattern of the reflected light collected from each wavelength does not provide any additional information, except for the wavelength dispersion of reflectivity. When coherence effects appear, they are stronger for longer wavelengths of dimensions closer to the dimensions of the hole. To reduce the effect of the coherence on the reflection pattern, a light source with broader wavelength range can be used. A wideband coherent light source such as a super continuum light source (e.g. fiber-laser) can also be used if high intensity is required.
Alternatively or additionally, in the absence of absolute hole size information in the reflected pattern, an additional channel can be implemented as part of the system to measure the width of the opening of the hole, for example using a high magnification imaging channel. Thus the combination of the information from the two channels can provide both the shape and size information.
To determine the depth of the hole, coherent illumination at normal incidence can be used and the interferences between the bottom reflection and the surrounding region of the hole can be measured. Combining absolute measurements of certain parameters, (e.g. hole dimensions such as depth, diameter, etc.) with relative measurements of the full hole cross-section (e.g. determination of the shape/geometry/profile, etc. of the hole), enables providing of the full cross-section measurement in absolute terms.
Moreover, phase measurement of the reflection can provide additional information on the hole profile (e.g. depth profile). This can be performed by splitting off a part of the incident beam prior to the interaction with the sample and causing interference of the split-off incident beam with the reflected beam. The split-off beam could be expanded in order to cover the whole detector light sensitive surface and thus produce phase sensitive intensity measurements.
For all types of light sources, a calibration procedure might to be performed to characterize the light pattern collected on the detector when reflected from a known reflectivity sample such as a single-crystal silicon wafer. This calibration can be performed on a periodic basis depending on the stability of the illumination system source and the optical system construction.
Reference is made to
If the analysis is carried out as part of an ongoing measurement flow, the initial theoretical profile can be based on the output result of a previous measurement. This can reduce the number of iterations required to obtain the desired fit.
The generation of a set of parameters can be based on analyzing the general morphology of the measured image. The morphological characteristics of the measured image can be selected from at least one of the following: the symmetry of image in relation to the incidence angle(s); the narrow-angle spread of reflection angles around the incident angle indicative of side-wall angles, the number of rings in the reflected image which is indicative of the aspect-ratio of the hole, the portion of the reflected light pattern at relatively high angles which is indicative of a slight bottom rounding, the reduced or hazy light pattern as indicator of surface roughness and others.
In some embodiments, multiple measured images are obtained, for example from multiple incidence angles of light sources. The multiple measured images induce the creation of a corresponding multiplicity of model images generated from the theoretical profile and a comparison metric is defined accordingly. The fit is then carried out in parallel for all the images.
It should be noted that the image produced from the theoretical profile can be based on geometrical ray tracing calculations or diffraction based calculations based on physical optics or on a combination of both. In order to reduce the calculation time, a series of theoretical profiles can be pre-calculated and stored in the memory utility. The profiles can be generated on the basis of a theoretical range of process conditions or based on characterization of actual samples. The theoretical range of process conditions can be the input of the control system, together with the parameters of measurement conditions of the optical system. The control system thus provides a range of possible theoretical profiles. These theoretical profiles are subsequently used to generate images and the images are stored in a database as illustrated in the figure.
The calculation time can also be reduced by recognizing morphological characteristics of the image (i.e. pattern recognition) and segmenting the image into specific profile families. The segmentation can be performed using at least one the morphological characteristic.
These morphological characteristics can also be used to recognize process excursions, e.g. reduced intensity of the light pattern in the image could be an indicator of enhanced scalloping formed for example by a Bosch-type etching process. A process alarm illustrated in 314 can be raised based on such a parameter.
An adaptive mode of measurement can be implemented, especially when the analyzing of samples of unknown hole profile is performed. In this adaptive mode of measurement the illumination system parameters are varied while successively capturing measurement images until a sharp image with strongly recognizable features is found. The parameters which can be varied include, but are not limited to: angle of incidence, rotation angle, wavelength, divergence angle. The recognizable features include, but are not limited to: under-filled rings, spots, high-angle signal rings, rings with cusps, bands dependent on the number of sidewall reflections.
Reference is made to
Reference is made to
Bottom rounding, due to its more varied angle content, causes rays to reflect in a much larger angle range than a flat bottom. The larger the area of non-flat bottom (e.g. bottom rounding), the larger the percentage of rays reflected out of the above described rings. The rays are reflected into a range of angles of order tens of degrees. The full reflection angular range can be analyzed or the reduced angular range image of the main sidewall and bottom reflections can be analyzed separately. Reference is made to
Reference is made to
The separation of effects due to the bottom rounding from those due to the sidewall slopes can be achieved by fitting the image to an image based on a theoretical model. If the parameter separation is difficult, measurements can be carried out at additional angles of incidence.
It should be noted that the reflected light beam from the sample surface outside the hole, which is essentially the specular reflection, is much stronger and more spatially concentrated than the reflected light beam from the hole. Moreover, the reflected light beam signal from the hole is much stronger and more spatially concentrated than the scattered light beam signal from the hole. A large dynamic range is therefore required from the detector to enable capture and measurement of all the reflected information simultaneously. Additionally, it can be useful to insert a spatial filter (e.g. mask) into the reflected beam that blocks only the region of the specular beam to eliminate the specular reflection or the region of the near-axis specular beam to eliminate the scattered light beam signal, reflected from flat areas outside the hole, thus enabling sufficient detector integration time without suffering from saturation of the reflected beam. If the weaker high angle reflections and scatter need to be analyzed, it is possible to block the near-axis reflections e.g. by providing appropriate blocking aperture stop.
In order to filter out the light reflected from the flat areas outside the hole, it is also possible to polarize the incident light and to provide a polarizer rotated at 90 degrees in the path to the detector. This can reduce the dynamic range required from the detector. The light collected by the detector will then only be skew rays undergoing reflections in the hole, having a direction out of the polarized plane of incidence.
It should be noted that if the hole has a non-flat bottom, the reflected light pattern becomes more complex owing to the continuous gradient of the reflection angle at the bottom. Then, the image has to be collected and analyzed with a large dynamic range, blocking the specular reflection from outside the hole. Therefore, for non-flat bottom hole, larger detector integration times or higher intensity incident light are then required.
Reference is made to
It should be noted that the shape of the hole opening is usually circular but the present invention can be used to characterize profiles of holes with various openings including square, elliptical, square with rounded corners as well as holes with a square opening tapering into a circle or rounded square at the bottom.
Moreover, due to the shallow incidence angle on the sidewalls, the reflected light from the hole can be at least partially polarized. Intentionally polarizing the incident light beam can enable extraction of ellipsometric effects of the sidewall reflections. The ellipsometric effects can be utilized for measurement of coatings on the sidewalls including both transparent and thin metallic coatings. The absorption of multiple reflections on the sidewalls by the substrate or coating materials causes a stepwise reduction of the intensity of some of the reflected light pattern depending on the number of reflections the rays undergo. This stepwise reduction of intensity is an additional parameter that can be modeled and used for obtaining information on the hole profile.
It should be noted that relatively small sharp features within the hole add diffraction effects to the reflected light pattern. This can affect the required angular divergence of the incident beam. Scanning the incident beam angle can enable more detailed sensing of such diffraction effects and determination of their size and location within the cross-section.
Increased background noise and loss of light patterns intensity or contrast are due to scattering induced by roughness of the sidewalls and of the bottom of the hole. Due to the fact that the scattered beam emanates, at least partially, from the bottom of the hole, the intensity of the scattered beam is higher close to the vertical axis of the hole. Therefore, the spatial profile of the scattered beam contains information on the shape of the hole opening and the aspect-ratio. It should be noted that certain techniques typically used for the formation of the etched holes can cause systematic sidewall roughness, e.g. scalloping caused by the well known Bosch etch process as illustrated in
For small hole-openings in the range of the wavelength dimension, the reflected light pattern contains diffraction effects. The theoretical model is then based on the convolution of the geometrical reflection from the sample surface outside the hole with the diffracted light from the shape of a single hole-opening. A light beam having large lateral coherence relative to the diameter of the hole can be used and a far field pattern of the reflection from the region of hole-opening can be analyzed to obtain the absolute size. If the holes are arranged in dense periodic array, additional diffraction effects could appear depending on the effective lateral coherence length. The diffraction image can be analyzed and the absolute size of the hole-opening can be obtained based on the knowledge of geometry of the system and the wavelength of the light beam.
Reference is made to
| Number | Date | Country | Kind |
|---|---|---|---|
| 188029 | Dec 2007 | IL | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/IL08/01599 | 12/10/2008 | WO | 00 | 6/7/2010 |
