WAFER-LEVEL TESTING PROCEDURES FOR OPTO-ELECTRONIC CHIPS

Description

Technical field

[0001] The invention relates to a method by which the functionality of electrical and optical components or circuits of a chip can be tested simultaneously at the wafer level in a wafer prober. Such a method is known in kind from US 2011 / 0279812 A1.

[0002] The invention relates to the testing and qualification of chips with photonic integrated circuits (PICs) at the wafer level. Unlike conventional, purely electrically integrated circuits (ICs), PICs integrate optical functionalities in addition to electrical circuitry.

[0003] In the manufacturing of integrated circuits (ICs), for example using CMOS technology, tests and measurements are performed at various stages of the process to monitor the manufacturing process and to conduct quality control. One established test is the electrical wafer-level test after the wafer has been prepared. This test identifies functional and non-functional chips, records them in a wafer map, and thus determines the yield. Functional chips are also referred to as Known Good Dies (KGDs). When the wafer is separated into individual chips, the non-functional chips are then sorted out. The test equipment required for the wafer-level test consists of wafer probers and wafer testers with associated contact modules (also called probe cards). The contact module connects the interfaces of the wafer tester to the individual interfaces of the chips on the wafer mounted on the wafer prober.In principle, the contact module can be designed to contact only one chip or several simultaneously. It is also not strictly necessary for the chips to still be embedded in the wafer for contacting. To contact multiple chips on a wafer simultaneously or sequentially, the chips simply need to have a fixed and defined orientation relative to each other.

[0004] Test equipment for testing purely electronic chips (semiconductor chips with ICs) has been optimized and diversified over decades in order to qualify high volumes of diverse ICs with high throughput in order to optimize costs.

[0005] PICs are typically manufactured using the same established semiconductor processes, such as CMOS technology. The very low production volumes of PICs compared to ICs mean that semiconductor fabrication facilities usually only perform process characterization tests, but not functional tests of the PICs. Functional characterization is the responsibility of the end customer and is often performed on sawn chips. The test equipment used employs independent, separate electrical and optical contact modules and is not optimized for throughput. State of the art

[0006] Testing PICs at the wafer level requires coupling light into and out of the PIC plane, typically using integrated grating couplers as coupling points, as described in the literature "Grating Couplers for Coupling between Optical Fibers and Nanophotonic Waveguides" (D. Taillaert et al., Japanese Journal of Applied Physics, Vol. 45, No. 8A, 2006, pp. 6071-6077). The grating couplers can be functional components within the chip or sacrificial structures on the wafer, e.g., in the groove or on adjacent chips.

[0007] In accordance with the state of the art, fiber-optic systems are used for wafer-level testing, as described in the technical literature: "Test-station for flexible semi-automatic wafer-level silicon photonics testing" (J. De Coster et al., 21st IEEE European Test Symposium, ETS 2016, Amsterdam, Netherlands, May 23-27, 2016. IEEE 2016, ISBN 978-1-4673-9659-2). These systems include a fiber-optic module that couples light into and out of the chip's coupling points via individual optical fibers. To ensure repeatable optical coupling, the optical fibers must be aligned to the coupling points with sub-micrometer precision, at distances of up to a few micrometers. This is only possible with the aid of high-precision actuators, such as a combination of hexapods and piezoelectric elements. Secondly, each individual optical coupling requires a time-consuming, active adjustment process designed to achieve maximum coupling efficiency.

[0008] This means that existing wafer-level test systems are characterized by Sequential, time-consuming contacting of all optical coupling points of a chip one after the other; that is, parallel contacting of all optical coupling points of a chip is not possible or only possible to a very limited extent, and parallel contacting of multiple chips is not possible at all. Device-specific special solutions are required, meaning that conventional wafer probers can only be converted with complex and costly modifications and are then no longer usable, or only to a limited extent, or only after time-consuming conversion for wafer-level testing of ICs. Separate, not permanently connected electronic and optical modules; that is, both must be held and adjusted separately.

[0009] From the aforementioned US 2006 / 0109015 A1, an optoelectronic contact module (probe module) for testing chips (device under test - DUT 140) with electrical and optical inputs and outputs is known. The contact module provides an interface between a test apparatus (ATE) and the device under test (DUT) and is equipped with electrical contacts (electrical probes), optical contacts (optical probes), optical elements, and combinations thereof to transmit signals to and from the DUT and to redistribute these signals for an interface with the test apparatus.

[0010] Regarding the optical inputs and outputs, it is disclosed that these are created via optical elements located on the contact plate and / or the redistribution plate and are tuned to various coupling mechanisms, e.g., free radiation, quasi-free radiation, or waveguides. Diffractive and refractive elements are specified as suitable optical elements for this purpose. It is also disclosed that a photodetector or a light source can be arranged directly at the interface to the device under test (DUT), which then constitutes the optical input or output on the contact plate.

[0011] Furthermore, the aforementioned US 2006 / 0109015 A1 teaches that for optical coupling implementations via free-beam or quasi-free-beam links, where the optical signal is guided through a free space between the optical element and the interface to the DUT, the optical signal is focused or collimated to achieve high coupling efficiency of the transmitted signal. Signal coupling here is therefore subject to the concept of coupling the signal as completely as possible.

[0012] According to an embodiment of the aforementioned US 2006 / 0109015 A1, the optical and electrical signal lines (optical and electrical distribution network) are implemented on separate redistribution plates. It is proposed to route the electrical signals from the DUT to the edge regions of the contact plate, so that the electrical signals are coupled into the first redistribution plate located above the contact plate at the edge region. This allows an opening to be formed in the first redistribution plate, in which only the electrical signals are redistributed, through which the optical signals are routed to a separate second redistribution plate located above it.

[0013] In summary, the aforementioned US 2006 / 0109015 A1 presents numerous ideas for how a contacting module, which is justifiably divided into a contacting plate and a redistribution plate (e.g., due to wear and tear of the mechanical contacts for electrical signal transmission), could be additionally equipped with optical signal lines. However, it completely disregards the fact that the tolerances possible for the mechanical contact of the electrical inputs and outputs of the contacting module to the device under test (DUT) are not transferable to the optical inputs and outputs.

[0014] While the transmission of a consistently identical electrical signal only requires mechanical contact between needles on the contacting module and contact pads on the DUT, which can be ensured within a comparatively large positional tolerance of a few µm in all three spatial directions, the quality of optical signal transmission is affected by a much smaller deviation from a target position, in the sub-µm range.

[0015] If, as described in the aforementioned US 2006 / 0109015 A1, the coupling efficiency of the optical signal is optimized by collimating or focusing the optical beam, the entire contacting module must be adjusted with high precision in the sub-micrometer range. Otherwise, the adjustment-dependent repeatability of the measurement is insufficient for the described applications. This, in turn, means that the contacting module cannot utilize the adjustment tolerances for electrical contacting typical of conventional electrical wafer probers, which are in the range of a few micrometers in the X, Y, and Z directions. Complex and expensive custom wafer prober solutions are required, including various actuating elements such as piezoelectric actuators and linear axes or hexapods, to align the contacting module with high precision relative to the device under test (DUT).

[0016] Another critical point is that a so-called overdrive of typically a few tens of micrometers in the Z-direction is set for precise electrical alignment of the needles. This means that after the needles make initial contact with the electrical contact pads, the contact module is moved an additional amount in the Z-direction to ensure reliable electrical contact. Wear and deformation of the needles are generally compensated for by adjusting the overdrive during operation. However, with simple collimation or focusing of the optical beam, as described in the aforementioned US 2006 / 0109015 A1, the working distance in the Z-direction must only vary by a few micrometers to achieve repeatable coupling. Therefore, this type of optical coupling is incompatible with established electrical contacting methods.

[0017] US Patent 2011 / 0279812 A1 discloses a contact module for testing chips with electrical and optical inputs and outputs. The chip is mounted on a movable carrier, which allows it to be roughly aligned with the contact module. This rough alignment is sensor-controlled, based on position monitoring of the chip or its alignment marks. Fine alignment of the chip is performed in two steps. In the first step, it is checked whether the electrical inputs and outputs are in contact with the contact module. For this purpose, the chip is suctioned onto the contact module so that the chip's electrical inputs and outputs make contact with the module's electrical contacts. A test signal is used to verify successful contact, and if contact is faulty, it is corrected by repeating the rough alignment. In the second step, the optical inputs and outputs are aligned.The chip's optical inputs and outputs can receive or transmit focused or collimated beams with adjusted aperture and focus position. The contact module's optical inputs and outputs feature variable optics that can generate beams focused perpendicular to the surface of the optical inputs and outputs, with adjustable axial and lateral focus position and beam aperture. For adjustment, the variable optics include at least one optical element with a variable focal length and / or at least one movable optical element. Axial focus adjustment is achieved based on distance measurements using additional distance sensors or intensity measurements using an optical test signal.Lateral adjustment of the focus position is based on intensity measurements. A test beam, already focused on the chip's surface, is moved in a scanning motion relative to the optical inputs and outputs until it is optimally coupled into these inputs and outputs. The time required for this can be reduced by initially performing the scanning motion with a larger focus diameter for the test beam. After fine-tuning, the chip is tested using specific electrical and optical test signal sequences. When testing multiple identical chips, parts of the coarse and fine adjustments can be saved and reused.

[0018] German patent DE 102018108283 A1 discloses an electro-optical printed circuit board for contacting photonic integrated circuits, in which an optical beam path passes through the printed circuit board. A disadvantage is the low achievable precision of the optical coupling between the printed circuit board and the chip under test.

[0019] WO 2019 / 029765A1 discloses a position tolerance-insensitive contacting module for contacting optoelectronic chips. A disadvantage is that the alignment of the contacting module to the chip under test may be suboptimal.

[0020] From DE10 2013 223034 A1 an optical coupling device for the optimal positioning of a fiber with respect to a grid coupler on a DUT is known, for example in an on-wafer test procedure, although no electrical coupling is described.

[0021] US patent 2021 / 0096176 A1 discloses an optical test system for a device under test (DUT) in which a scan is performed with an optical test tip. A disadvantage is that the optical test tip can easily break off, and simultaneous optical and electrical coupling is difficult to achieve.

[0022] US Patent 7,378,861 B1 discloses an optical alignment method for wafer-level testing of optoelectronic chips. A disadvantage is the independent electrical and optical alignment, which requires at least two independent positioning systems. Furthermore, semi-automatic testing of silicon photonics devices under test (DUTs) on a wafer is known (DE COSTER JEROEN ET AL: "Test-station for flexible semiautomatic wafer-level silicon photonics testing", 2016 21st IEEE EUROPEAN TEST SYMPOSIUM (ETS), IEEE, May 23, 2016 (2016-05-23), pages 1-6, XP032928385, DOI: 10.1109 / ETS.2016.7519306). This involves optically contacting the coupling grids of the DUT using waveguides.

[0023] For testing electronic chips at the wafer level, it is common practice to use a first camera to determine the spatial orientation of the probe tips and a second camera to determine the spatial orientation of the contact pad centers. Control signals are then generated from the resulting relative orientation of the probe tips to the contact pads. These signals are used to control a positioning stage, positioning the contact pad centers perpendicularly below the probe tips before the probes are brought into contact with the pads. Object of the invention

[0024] The object of the invention is to find a method for testing optoelectronic chips arranged on a wafer with electrical interfaces in the form of contact pads and fixedly arranged optical interfaces in the form of optical deflection elements with a specific coupling angle, which is based on proven methods for testing electronic chips. Solution to the task

[0025] This problem is solved by a method according to claim 1 for testing opto-electronic chips arranged on a wafer with electrical interfaces in the form of contact pads and optical interfaces fixedly arranged therein in the form of optical deflection elements. Description

[0026] The method according to the invention is used for testing optoelectronic chips arranged on a wafer.

[0027] The chips feature electrical interfaces in the form of contact pads and correspondingly fixed optical elements. Optical elements can be, for example, passive optical elements, optoelectronic actuators and / or optoelectronic sensors and / or electro-optical modulators. The optical elements include optical interfaces in the form of optical deflection elements with a specific coupling angle α. The deflection elements can be designed, for example, as grating couplers or mirrors. The specific coupling angle α can represent the angle that an optical signal, or its central ray, forms with a perpendicular to the wafer. The perpendicular can be in the z-direction. The specific coupling angle can typically be greater than 0° and less than 25°. A common value for the specific coupling angle α might be, for example...The angle of deflection is 11.6° and is advantageous for the optical fibers with a corresponding wedge-shaped cross-section used for coupling in the final application of the chip. Advantageously, the beams deflected by several, and particularly advantageously by all, optical deflection elements can be parallel. If the optical elements are designed for beam bundles, the parallelism can be determined with respect to the central rays of the beam bundles. Then the coordinate system can be chosen such that the y-components of the deflected beams vanish, i.e., that the deflected beams can lie in xz-planes.

[0028] In the method according to the invention, the wafer is received by a positioning table which is adjustable in the x, y, and z directions of a Cartesian coordinate system relative to a contacting module and rotatable about the z-axis. Advantageously, the x and y coordinates can be adjusted by means of the positioning table. Advantageously, the z-coordinate can also be adjusted by means of the positioning table. Alternatively or additionally, the adjustability in the z-direction can be provided by means of a vertical lifting and lowering device of the contacting module. The rotatability about the z-axis can advantageously be provided by means of a rotatable positioning table. Alternatively or additionally, the rotatability about the z-axis can be provided by means of a rotating device of the contacting module. The contacting module has electrical interfaces in the form of needles with needle tips associated with the contact pads.Furthermore, the contacting module has optical interfaces assigned to the optical deflection elements.

[0029] Any deviation of the x-travel path or y-travel path of the positioning table from the x-direction or y-direction of the chip can be compensated for, for example, by means of the aforementioned rotation about the z-axis or by means of a rotation matrix in the sense of a rotation as a linear coordinate transformation (so-called alias transformation in the xy-plane) when controlling the travel paths of the positioning table.

[0030] In a first alignment step, the wafer is positioned towards the contact module such that the needle tips are arranged in a first position (x1, y1, z1), spaced apart in the z-direction, above a predetermined location on the associated contact pad of a first chip. The contact module maintains a first distance a to the first chip in the z-direction. This first distance a is greater than a maximum distance I between the needle tips and the contact module. The maximum distance of the needle tips to the contact module can also be considered as an effective needle length in the z-direction in the mechanically unloaded (i.e., force-free) state. The predetermined location can be a specific xy position, i.e., a point in an xy plane. The predetermined location can be, but need not be, the center of the contact pad. The needles can, for example, be arranged perpendicular (i.e.,The needles arranged in the z-direction, so-called "vertical needles," may be designed and may have only a slight spring effect. Advantageously, the needles can be designed as spring-loaded vertical or inclined needles. Particularly advantageously, the needles can be designed as cantilever needles. Cantilever needles can be those needles that are arranged in a resilient manner like a bending beam. Preferably, cantilever needles can be arranged such that, with respect to their bending beam section, they form an angle greater than 60° to the z-axis in the unloaded (force-free) state, particularly preferably greater than 70°, and most preferably greater than 80°. The first alignment step can be carried out, for example, using one or more cameras, with which the first position can be located.Alternatively, manual positioning or passive positioning by striking the wafer edge against designated stops can also be performed.

[0031] Starting from the first position (x1, y1, z1), a second position (x2, y2, z2) is determined in a second alignment step by moving the positioning table to the contacting module in the z-direction at a second distance b, where the needle tips are in contact with the contact pads, so that electrical signals can be transmitted via the respective corresponding interfaces and optical signals can be transmitted via the respective corresponding optical interfaces.According to the invention, in the second alignment step, the optical deflection elements are aligned relative to the optical interfaces present on the contacting module by deflecting the positioning table in scan positions (xs[i], ys[i], zs[i]) within a scan area relative to the x and / or y coordinates of the first position in the x and / or y direction and / or adjusting it in the z direction, and at the scan positions P[i] = (xs[i], ys[i], zs[i]) an optical signal is coupled via at least one of the optical interfaces of the contacting module and one of the optical deflection elements, wherein the second position (x2, y2, z2) is defined by the fact that the optical signal is coupled with a maximum degree of coupling, the degree of coupling being determined by means of at least one electrical signal transmitted via at least one electrical interface.If several optical interfaces are coupled in the second alignment step, the average of the coupling degree from the several optical interfaces can be used for further analysis.

[0032] The scan area can generally be understood as the spatially defined region in which the scan takes place. As a special case, the scan area can be two-dimensional, and as a special case of this special case, linear. Mathematically, the scan area can be defined as the smallest convex polyhedron containing all scan positions (xs[i], ys[i], zs[i]). The scan area can advantageously be two-dimensional; in this case, the polyhedron can be degenerate into a polygon. The index i can be used as a running index i=1, 2 ... imax for the sequential numbering of the scan positions. This allows the temporal sequence of the scan positions to be predetermined. However, the scan positions can also be determined adaptively. This can mean that the position P[i+1] is only determined during the scan's execution, taking into account the coupling degrees at the scan positions P[1] to P[i].The scan field can be understood as a z-projection of the scan area onto the xy-plane of the wafer. The scan field can generally be two-dimensional (i.e., planar), as a special case one-dimensional (i.e., linear), and as a special case of the special case zero-dimensional (i.e., point-like). The xy-extent of the contact pads can be understood as their geometric extent in the x and y directions. A scan field smaller than the xy-extent of the contact pads can mean that the needle tips do not leave their respective contact pads during the scan. In a special implementation, the scan can be linear, corresponding to a one-dimensional scan area. In a special case of this implementation, the scan can be linear in the z-direction. Then the scan field can be point-like. This can have the advantage of enabling scanning even with very small contact pads.

[0033] A coupling coefficient k[i] can be determined at each of the scan positions P[i]. The second position (x2, y2, z2) can be an interpolated maximum position of the coupling coefficients determined at the scan positions. Alternatively, the scan position at which the maximum coupling coefficient was measured can be used as the second position without interpolation. At the end of the second alignment step, the determined second position (x2, y2, z2) can be approached, but this is not mandatory.

[0034] According to the invention, in a third alignment step, the first of the chips is brought to a third position (x3, y3, z3) by first setting the x3 and y3 coordinates of the positioning table and then moving it in the z-direction to a third distance c, which is smaller than the second distance b, wherein the needle tips are in contact with the contact pads with a predetermined pressure force, wherein the third position (x3, y3, z3) was calculated from the second position and the coupling angle α of the relevant optical deflection element for the third distance c before the third alignment step.

[0035] The third position (x3, y3, z3) can be determined by the fact that, for the third distance assigned to coordinate z3, a maximum degree of coupling exists at the point with coordinates x3, y3. The third distance c can be a predetermined optical working distance at which the needles bear against the contact pads with a predetermined pressure force.

[0036] The third position can be calculated, for example, as follows. Without loss of generality, the beam deflected by a deflecting element at the specific coupling angle α can lie in an xz plane (i.e., perpendicular to y). Then the third position (x3, y3, z3) can be determined as x3 = x2 + (z3 - z2) tan α and y3 = y2, where the coordinate z3 can be determined by the presence of the specified working distance c. As described below, the required working distance c (and thus the coordinate z3) can change over time.

[0037] In the third position, electrical or optical signals can then be transmitted via the respective corresponding interfaces for testing the opto-electronic chip.

[0038] Advantageously, a position difference resulting from the first position and the second position of the first of the chips is stored as an offset and taken into account for the adjustment of further chips after they have been positioned in a first position relative to the contacting module.

[0039] In particular, to take into account the wear of the needle tips, it is advantageous if the optical working distance c, at which the needles are in contact with the contact pads with a predetermined pressure force, is monitored and the third position is corrected if the pressure force changes.

[0040] By determining the change in the optical working distance c over the long term over the service life of the contacting module, it is advantageously possible to replace the needles of the contacting module with new needles during the process if the optical working distance c falls below a predetermined minimum distance.

[0041] By advantageously overexposing the respective interface with the optical signals during coupling, positional tolerances of the optical interfaces of the chips are compensated for.

[0042] It is even more advantageous if the optical signals transmitted via the respective assigned interfaces exhibit a tophat distribution of their radiation intensity when coupled in.

[0043] Advantageously, the third alignment step can follow directly after the second alignment step. This minimizes processing time. It is also advantageous to include a fourth alignment step between the second and third alignment steps, in which the positioning table is moved to a fourth distance d in the -z direction (i.e., in the negative z direction), where this fourth distance d is greater than the maximum distance I between the needle tips and the contact module. This reduces the risk of scratching the pads.

[0044] Advantageously, in the second alignment step, the scan positions (xs[i], ys[i], zs[i]) can lie in an xy-plane (x, y, z2). The scan positions Ps[i] can then be (xs[i], ys[i], z2).

[0045] It is also advantageous in the second alignment step for the scan positions (xs[i], ys[i], zs[i]) to lie in a yz plane (x2, y, z). The scan positions Ps[i] can then be (x2, ys[i], zs[i]). The scan positions can advantageously be chosen such that z is monotonically increasing, i.e., z[i+1] ≥ z[i] ∀ i ∈ {1, 2, ..., imax-1}.

[0046] It is also advantageous in the second alignment step for the scan positions (xs[i], ys[i], zs[i]) to lie on a line (x2, y2, z). The scan positions Ps[i] can then be (x2, y2, zs[i]). In this case, alignment with respect to y can be omitted.

[0047] Advantageously, at least one of the optical elements of a chip can comprise an electro-optical sensor and / or an electro-optical actuator and / or an electro-optical modulator. An electro-optical element can be electrically contacted via two electrical interfaces. It is also possible for one of the electrical interfaces to be replaced by a substrate contact (ground contact). When using such a ground contact, one electrical interface of the electro-optical element may be sufficient for the process. For example, the anode of an electro-optical element can be connected to ground and only the cathode connected via one of the electrical interfaces, or vice versa.

[0048] Advantageously, the electro-optical sensor can be designed as a photodiode or a phototransistor. In the second alignment step, the coupling coefficient can then be determined by coupling light via the optical interface of the contact module into the deflecting element associated with the optical sensor and measuring a photocurrent via the corresponding electrical interfaces. The coupling coefficient can be determined from the ratio of the photocurrent to the light power output by the contact module at the interface.

[0049] Advantageously, the electro-optic actuator can be configured as a light-emitting diode (LED) or a laser diode. A light-emitting diode can be a conventional LED or a superluminescent diode. In the second alignment step of an electro-optic actuator, the coupling coefficient can be determined by supplying an operating current to the actuator via electrical interfaces, causing it to generate light. The generated light is directed to the optical interface of the contact module via the associated deflection element, and the amount of light received is measured. The coupling coefficient can be determined from the ratio of the photocurrent to the light power output by the optical element at the interface. The operating current of the optical element can be used as a measure of the optical light power.

[0050] Advantageously, the electro-optical modulator can be designed as a Mach-Zehnder interferometer. In this case, the coupling coefficient can be measured via two optical interfaces. Conveniently, the interferometer can be switched into a forward state by applying a voltage to the corresponding electrical interfaces.

[0051] Advantageously, at least one optical actuator can be used as a sensor in the second alignment step. For example, an LED or laser diode can be operated as a photodiode. This can have the advantage that no operating current needs to be passed through the electrical contacts, but only a comparatively small photocurrent to determine the coupling degree. Examples of implementation

[0052] The invention will be explained in more detail below using exemplary embodiments and the accompanying drawings. The drawings show: Fig. 1a shows a chip of a first embodiment arranged in a first position relative to the contacting module after the first alignment step. Fig. 1b shows the first embodiment during the second alignment step. Fig. 1c shows the first embodiment after the third alignment step. Fig. 2a shows a chip of a second embodiment arranged in a first position relative to the contacting module after the first alignment step. Fig. 2b shows the second embodiment during the second alignment step. Fig. 2c shows the second embodiment after the third alignment step. Fig. 3 shows the fourth step of a modification of the second embodiment. Fig. 4 shows a scanning procedure during the second alignment step. Fig. 5 shows the corresponding scan field.

[0053] In a method according to the invention, optoelectronic chips 1 arranged on a wafer with electrical interfaces in the form of contact pads 1.1 and fixedly arranged optical interfaces in the form of optical deflection elements 1.2, e.g., grating couplers or mirrors, are tested with a specific coupling angle α. The specific coupling angle α represents the angle that an optical signal or its central ray forms with a perpendicular to the chip 1. It is typically greater than 0° and less than 25°. A common value for the specific coupling angle α in air is, for example, 11.6° and refers to the optical fibers with a corresponding wedge-shaped tip used for coupling in the final application of the chip.

[0054] In this process, a wafer is picked up by a positioning table 3, which is adjustable in the x, y, and z directions of a Cartesian coordinate system relative to a contacting module 2 and rotatable about the z-axis, similar to a prior art method. The contacting module 2 has electrical interfaces 2.1 and optical interfaces 2.2 that can be assigned to the chips 1. The electrical and optical interfaces of the chips 1 are manufactured in different process steps within the wafer assembly, so that while they exhibit only small positional tolerances relative to each other, the arrangements formed by the electrical interfaces can have tolerance deviations compared to the arrangements formed by the optical interfaces, particularly from wafer to wafer. The same applies to the tolerances of the contacting module. Furthermore, the coupling angle α can be subject to tolerances.In particular, the coupling angle can differ from wafer to wafer due to tolerances.

[0055] In a first alignment step, the wafer is positioned at the contact module 2 such that the electrical interfaces on the contact module 2, in the form of needles 1.1, are arranged in a first position (nominal position) perpendicular to predetermined points 1.4, which correspond to the centers 1.3 of the contact pads 1.1, of a first chip 1. During alignment, the contact module has an adjustment distance a in the z-direction, greater than the free length I of the needles 2.1 to the chip 1, so that no contact can occur between the tips of the needles 2.1 and the contact pads 1.1. This corresponds to a first position (x1, y1, z1) with a z-coordinate z1. See also Fig. 1a ,For the sake of simplicity, as in the other figures, the illustration of the adjustment in the y-direction has been omitted here. Accordingly, positioning paths during individual adjustment steps are only shown as positioning paths in the x-direction. In a non-figurative variation of this embodiment, the predetermined positions are arranged at a distance from the centers of the contact pads.

[0056] The first adjustment step is advantageously performed according to a well-established routine. One camera measures the needles 2 by focusing on their tips, while a second camera measures the contact pads 1.1 of the chip 1. Both cameras are previously referenced to each other using a normal. This allows for the precise calculation of the optimal position (nominal position) of the needles 2 relative to the contact pads 1.1, and thus the positioning of the contact module 2 to the chip 1. This is usually done via regression and extrapolation of the measured values. Furthermore, a corresponding contact pad 1.1 is expected for each needle 2 located. The user cannot typically intervene in this routine. Similarly, alternative structures such as alignment marks cannot usually be used for positioning.The cameras not only correct the xy position and any rotation around the z-axis, but also determine the z-position. The reference points for the needles 2.1 are their needle tips 2.3.

[0057] It is clear to those skilled in the art that, due to positional tolerances between the needle tips 2.3 and the centers of the contact pads 1.1, it is not possible to position all needle tips exactly above the centers of the contact pads 1.1 simultaneously. Ultimately, a position is adjusted where the mean deviation is smallest. However, the positional tolerance of the electrical interfaces relative to each other is negligibly small compared to the positional tolerance of the arrangement of the electrical interfaces relative to the arrangement of the optical interfaces of a chip 1. This is primarily due to the fact that the electrical and optical interfaces are manufactured sequentially using different process steps. The same applies to the tolerances of the contact module. Furthermore, the coupling angle α may be subject to tolerances. In particular, the coupling angle may differ from wafer to wafer due to these tolerances.

[0058] Since the production of the electrical interfaces or the production of the optical interfaces for all chips of a wafer takes place in one process sequence, the positional deviation between the arrangements of the electrical interfaces and the arrangements of the optical interfaces of individual chips of a wafer is at least approximately the same.

[0059] After the arrangement of the electrical contacts of chip 1 (pins 2.1) is aligned with the electrical contacts of contacting module 2 (contact pads 1), the actual position of the optical interfaces of the chip to the optical interfaces on contacting module 2.2 deviates from a target position due to various reasons, see Fig. 1a .

[0060] Firstly, the position of the optical interfaces on chip 1 deviates from its target position in the x, y, z, and around the z-direction in a statistically variable manner for each wafer, while the deviations around the x and y directions can be assumed to be constant due to the alignment of the positioning stage 3 after the position of chip 1, and thus of the wafer, has been fixed (systematic deviations). Furthermore, the coupling angle can deviate from a target angle.

[0061] Secondly, there is a deviation in the position of the optical interfaces on the contacting module 2.2 from a target position relative to the needles 2.1, depending on the assembly accuracy of the optical module, which is an integral part of the contacting module. This applies not only to positional deviations in the x, y, and z directions, but also to tilting around the z, x, and y directions. All six parameters are assembly-related, systematic deviations.

[0062] Thirdly, variable deviations occur over the lifetime of contact module 2. This is due to mechanical wear and tear, resulting in changes in the shape of the needle tips, as well as possible bending and the consequent changes in the fit.

[0063] Using camera measurement, the positioning table 3 effectively corrects the position of the entire contacting module 2 to the chip / wafer accordingly - and thus also changes the position of the arrangement of the optical interfaces of the contacting module 2.2 to the arrangement of the optical interfaces on the chip 1 due to the specific coupling angle of the respective optical deflection element, in particular a grating coupler.

[0064] Starting from the first position (x1, y1, z1), a second alignment step is performed (see here). Fig. 1b )A second position (x2, y2, z2) is determined. For this purpose, the positioning table (3) is first moved to the contacting module (2) in the z-direction to a second distance (b) at which the needle tips (2.1) are in contact with the contact pads (1.1). This ensures that electrical signals can be transmitted via the respective corresponding interfaces and optical signals via the respective corresponding optical interfaces. Now, in the second alignment step, the optical deflection elements (1.2) are aligned relative to the optical interfaces (2.2) on the contacting module (2). In a scan field (4, 5) smaller than the xy extent of the contact pads (1.1), the positioning table (3) is deflected to scan positions Ps[i] = (xs[i], ys[i], z2) within a scan area (4, 5) relative to the x and / or y coordinates of the first position in the x-direction (4) and y-direction (5). The scan thus takes place in an xy-plane.In further non-figurative variations of the example, the scan can also include delivery in the z-direction; for example, the scan can be performed in a yz-plane.

[0065] At the scan positions (xs[i], ys[i], zs[i]) an optical signal is coupled via at least one - here two - of the optical interfaces of the contacting module (2.2) and one of the optical deflection elements (1.2).

[0066] The second position (x2, y2, z2) is defined by coupling the optical signal with a maximum coupling degree, where the coupling degree is determined by means of at least one electrical signal transmitted via at least one electrical interface (1.2, 2.1). For a photodiode, two of the three depicted electrical interfaces are used. Alternatively, one of the two interfaces used can be replaced by a substrate contact of the wafer. If several optical interfaces are coupled simultaneously, the average coupling degree from several interfaces can be used for further analysis.

[0067] In the diagram, the second position is located at the x-coordinate x2, which is different from x1. See also... Fig. 1b . The second position does not need to be approached; it is sufficient if it is determined, for example, by means of regression from the coupling degrees at the scan positions.

[0068] In a third alignment step, the first of the chips (1) is moved to a third position (x3, y3, z3) by first setting the x3 and y3 coordinates of the positioning stage and then moving it in the z-direction to a third distance c, which is smaller than the second distance b. The needle tips (2.3) then contact the contact pads (1.1) with a predetermined pressure, as shown in Fig. 1cThis is illustrated. This is achieved by selecting an optical working distance smaller than the free length of the needles I. The difference is called overtravel or overdrive. The needles can, for example, bend or flex elastically. Elastic bending is exaggerated here to illustrate the effect. In the second alignment step, a smaller overtravel is selected than in the third alignment step. The second distance b is larger than the third distance c. This protects the contact pads during scanning.

[0069] Overtravel ensures reliable electrical contact between the needles and the contact pads (low contact resistance). After the initial light contact of the needles with the contact pads, the wafer is moved upwards by a few tens of micrometers in the z-direction during the third alignment step. This achieves two things. First, it breaks through any oxide surfaces that may be present, resulting in a reproducible, low-resistance contact. Second, overtravel generates a constant contact pressure from the needles, as the additional travel causes them to compress and exert a pressure on the contact pad. This pressure varies depending on the needle type used, but it can be estimated at approximately 0.03 N per needle. Overtravel is typically specified in multiples of MIL (1 mil = 1 / 1000 inch = 0.0254 mm).

[0070] In a contact module where the needle arrangement has a fixed position relative to the optical interfaces, the overtravel value must be considered during the final chip alignment. This ensures that, in the contacted state, an optical working distance is maintained between the contact module's optical interfaces and the chip, maximizing the coupling of the optical signals. The third position (x3, y3, z3) is calculated from the second position and the coupling angle (α) of the relevant optical deflection element (1.2) for the third distance (c) prior to the third alignment step. This allows the x and y coordinates of the third position to be set first, followed by adjusting the z coordinate to the desired working distance c. This prevents the needles from scratching the pads.

[0071] To adjust the optical working distance c, the chip is moved to the third position. In this position, the final alignment state is achieved, in which both the electrical and optical interfaces of the chip and contact module are optimally aligned. This ensures the best possible measurement capability of the optical signal flows (maximum position of optical coupling), while the electrical signal flow is also present. To test the chip, electrical and optical signals are then passed through their respective corresponding interfaces.

[0072] The adjustment path from the first position (nominal position) to the third position (maximum position of optical coupling) represents an offset (in Fig. 1c, represented as a comparison of coordinates x3 to x1 and x2, and as coordinate z3) which is advantageously stored and taken into account when aligning all further chips on this wafer. That is, the nominal position of the positioning stage is offset by the corresponding difference x3-x1 in the x-direction (see Fig. 1c ) and the difference y3-y1 in the y-direction (not shown) is corrected. However, this requires a corresponding shift of the needles over the contact pads, which is still tolerable.

[0073] This procedure only needs to be performed once per wafer or at longer intervals.

[0074] Furthermore, this offset value, i.e. the difference x3-x1 as well as the difference y3-y1 and its changes over time, can be observed, thus allowing statements about the wear and associated changes of the needles.

[0075] To increase the accuracy of the offset determination, the raster scan can also be performed on several chips of the wafer before the start of the inspection of all chips of the wafer and the results can be averaged.

[0076] The required overtravel, i.e. the difference cI, can change over the lifetime of the contacting module due to "run-in" or wear of the needles, thus reducing the working distance c.

[0077] Direct monitoring of the optical working distance is important to ensure the defined optical coupling properties between the contacting module and the chip (measurability) and to prevent collisions of the optical interfaces of the contacting module and chips (a few 10-100µm distance in operation).

[0078] For this purpose, a distance sensor permanently integrated into the contacting module can be used, e.g., a capacitive distance sensor. This enables monitoring of the actual optical working distance. In combination with active control, the working distance can be actively adjusted by moving the positioning table in the z-direction, and if necessary, a hard stop can be programmed for when a minimum working distance is not reached, e.g., to prevent a collision due to operator error.

[0079] Typically, a change in the needle tip positions is to be expected during the initial commissioning of the contact module (run-in behavior). This can be anticipated by multiple contact simulations before adjusting the first chip (pre-aging), thereby reducing the required offset value.

[0080] Typically, all optical interfaces of the chips, embodied by grating couplers, have coupling angles with the same angular magnitude and orientation, so that a change in the optical working distance Δb results in a correspondingly equal relative change ΔX of the optimal coupling position for all optical interfaces.

[0081] Typically, the change in the optical working distance Δb is in the range of <50 µm. For example, a change in the optical working distance of 10 µm and a coupling angle α of 11.6° in the xz-plane would result in a change in the optimal coupling position Δx of 2 µm in the x-direction. A change in the optical working distance Δb of 20 µm would result in a change in the optimal coupling position of Δx of 4 µm in the x-direction. This can still be compensated for by the correction described above; that is, the positioning stage is readjusted in the x-direction via a control command, assuming that the needles still reliably contact the contact pads.

[0082] The accuracy requirements for the adjustment steps can be reduced if the optical signals used to test the chip via the respective corresponding interfaces overexpose the coupling interface.

[0083] Fig. 2aFigure 2 shows a chip of a second embodiment arranged in a first position relative to the contacting module after the first alignment step. In contrast to the first embodiment, cantilever needles 2.1 are used here. In a first alignment step, the wafer is positioned towards the contacting module (2) such that the needle tips (2.3) are arranged in a first position (x1, y1, z1) spaced apart in the z-direction above a predetermined location (1.4) of the associated contact pad (1.1) of a first of the chips (1), wherein the contacting module (2) has a first distance (a) in the z-direction to the first of the chips (1), the first distance (a) being greater than a maximum distance (I) of the needle tips (2.3) to the contacting module (2). The predetermined locations 1.4 are provided here outside the centers 1.3 of the contact pads 1.1.

[0084] Fig. 2bThe second embodiment is shown during the second alignment step. Unlike the first embodiment, this one uses a yz-scan over a scan area in the y-direction 5 and z-direction 6 with a fixed coordinate x2=x1. Here, the scan can advantageously be aborted after exceeding the maximum coupling degree at z2. In a variation of the second embodiment, the scan is performed with a fixed y-coordinate y2=y1; this is a line scan. In a further variation, the scan is performed with a fixed z-coordinate in an xy-plane.

[0085] Fig. 2c The second embodiment is shown after the third alignment step. With cantilever needles, the contact force can be applied by means of elastic bending. The third alignment step takes place immediately after the second.

[0086] Fig. 3Figure 1 shows a fourth alignment step of a further modification of the second embodiment. Here, the positioning table is moved to a fourth distance d in the -z direction, where the fourth distance d is greater than the maximum distance I of the needle tips (2.3) to the contacting module (2). The fourth alignment step takes place between the second and third alignment steps.

[0087] Fig. 4This diagram shows a scan procedure for a yz-scan during the second alignment step. The abscissa represents the z-coordinate of the scan trajectory minus the z2 coordinate of the second position. The ordinate represents the y-coordinate of the scan trajectory minus the y2 coordinate of the second position. The trajectory (yz path) is shown as a thick line in the diagram, with arrows indicating its direction of travel. The scan positions are chosen sequentially at the corners of the plotted trajectory. The scan begins in the center right of the diagram at the second position (x2, y2, z2). A total of 55 scan positions are used. The z-coordinate increases monotonically in 54 steps, alternating between 1 µm and 0 µm. The y-coordinate moves in steps of -1µm, 0µm, and 1µm between y2-3µm and y2+3µm as shown. The scan range is 6µm in the y-direction and 27µm in the z-direction.The path length of 28 µm in the z-direction is equivalent to a beam displacement of 5.5 µm in the x-direction at a specific coupling angle α = 11.6°. The in . Fig. 5 The scan area shown (7) is the polygon surrounding the scan positions. It is formed as a flat surface in a yz-plane. The scan field (not shown) as a perpendicular z-projection of the scan area would be one-dimensional (linear) extending in the y-direction.

[0088] In a non-figurative variation of the one depicted in Fig. 4 and 5 In the scan procedure shown, the y-coordinate remains fixed and the scan is performed as a line scan in the z-direction. The scan field would then be point-like. Reference symbol list

[0089] 1 Chip 1.1 Contact pad 1.2 Optical deflection element 1.3 Center of the contact pad 1.4 Predetermined position 2 Contacting module 2.1 Needle 2.2 Optical interface on the contacting module 2.3 Needle tip 3 Positioning stage 4. x- Scan area 5. y- Scan area 6. z- Scan area 7. Scan area α Coupling angle First distance Second distance Third distance Free length of the needle

Claims

1. Method for testing optoelectronic chips (1) arranged on a wafer and having electrical interfaces in the form of contact pads (1.1) and optical elements which are fixedly arranged in relation to said electrical interfaces and which comprise optical interfaces in the form of optical deflection elements (1.2) with a specific coupling angle (α), wherein the wafer is accommodated by a positioning stage (3) which is adjustable vis-à-vis a contacting module (2) in the x-, y-, and z-direction of a Cartesian coordinate system and which is rotatable about the z-axis, the contacting module (2) having electrical interfaces in the form of needles (2.1) with needle tips (2.3) assigned to the contact pads (1.1) and optical interfaces (2.2) assigned to the optical deflection elements (1.2), and • in a first alignment step, the wafer is moved toward the contacting module (2) in such a way that the needle tips (2.3) are arranged in a first position (x1, y1, z1), in each case with a spacing in the z-direction above a predetermined point (1.4) of the assigned contact pad (1.1) of a first of the chips (1), the contacting module (2) being at a first distance (a) from the first of the chips (1) in the z-direction and the first distance (a) being greater than a maximum distance (I) of the needle tips (2.3) from the contacting module (2), characterized in that • proceeding from the first position (x1, y1, z1), a second position (x2, y2, z2) is determined in a second alignment step by virtue of the positioning stage (3) being moved toward the contacting module (2) at a second distance (b) in the z-direction, in which the needle tips (2.1) abut against the contact pads (1.1) such that electrical signals are transferable via the respective interfaces assigned to one another and optical signals are transferable via the respective optical interfaces assigned to one another, with a relative alignment of the optical deflection elements (1.2) with respect to the optical interfaces (2.2) present on the contacting module (2) being implemented in the second alignment step by virtue of, in a scanning field (4, 5) smaller than an xy-extent of the contact pads (1.1), the positioning stage (3) being deflected in the x- and / or y-direction with respect to the x- and / or y-coordinates of the first position and / or being moved in the z-direction toward scanning positions Ps[i] = (xs[i], ys[i], zs[i]) within a scanning region (7), and an optical signal being coupled at the scanning positions (xs[i], ys[i], zs[i]) via at least one of the optical interfaces of the contacting module (2.2) and one of the optical deflection elements (1.2), with the second position (x2, y2, z2) being defined by virtue of the fact that the optical signal is coupled with a maximum degree of coupling, with the degree of coupling being determined by means of at least one electrical signal which is transferred via at least one electrical interface (1.2, 2.1), and • in a third alignment step, the first of the chips (1) is brought into a third position (x3, y3, z3) by virtue of the x3- and y3-coordinate of the positioning stage being initially set and said positioning stage subsequently being moved in the z-direction toward a third distance c, which is shorter than the second distance b, with the needle tips (2.3) abutting against the contact pads (1.1) with a predetermined contact force, the third position (x3, y3, z3) having been calculated from the second position and the coupling angle (α) of the relevant optical deflection element (1.2) for the third distance (c) prior to the third alignment step.

2. Method for testing optoelectronic chips (1) arranged on a wafer according to Claim 1, characterized in that the third alignment step immediately follows the second alignment step or in that a fourth alignment step is provided between the second and third alignment step and the positioning stage is moved away therein to a fourth distance d in the -z-direction, with the fourth distance d being greater than the maximum distance I of the needle tips (2.3) from the contacting module (2).

3. Method for testing optoelectronic chips (1) arranged on a wafer according to either of the preceding claims, characterized in that the scanning positions (xs[i], ys[i], zs[i]) are located in an xy-plane (x, y, z2) and / or on a line (x2, y2, z) and / or in a yz-plane (x2, y, z) in the second alignment step.

4. Method for testing optoelectronic chips (1) arranged on a wafer according to any of the preceding claims, characterized in that a positional difference arising from the first position and the second or the third position of the first of the chips (1) is stored as an offset and taken into consideration when aligning more of the chips (1) after these were positioned relative to the contacting module (2) in a first alignment step.

5. Method for testing optoelectronic chips (1) arranged on a wafer according to any of the preceding claims, characterized in that the third distance c, at which the needles (2.1) abut against the contact pads (1.1) with a specified compression force, is monitored and the third distance is corrected in the case of changes in the compression force.

6. Method for testing optoelectronic chips (1) arranged on a wafer according to any of the preceding claims, characterized in that the long-term change in the third distance c is determined over the use duration of the contacting module (2) and the needles (2.1) of the contacting module (2) are replaced by new needles (2.1) for the purposes of carrying out the method when the third distance c drops below a specified minimum distance.

7. Method for testing optoelectronic chips arranged on a wafer according to any of the preceding claims, characterized in that the optical signals transmitted via the mutually assigned interfaces swamp one of the interfaces when input coupled therein.

8. Method for testing optoelectronic chips (1) arranged on a wafer according to any of the preceding claims, characterized in that the optical signals transmitted via the mutually assigned interfaces have a top hat distribution of their radiation intensity when input coupled into one of the interfaces.

9. Method for testing optoelectronic chips (1) arranged on a wafer according to any of the preceding claims, characterized in that at least one of the optical elements of a chip comprises an electro-optic sensor and / or an electro-optic actuator and / or electro-optic modulator.

10. Method for testing optoelectronic chips (1) arranged on a wafer according to Claim 9, characterized in that the electro-optic sensor is designed as a photodiode or phototransistor and / or in that the electro-optic actuator is designed as a light-emitting diode or a laser diode or in that the electro-optic modulator is designed as a Mach-Zehnder interferometer.

11. Method for testing optoelectronic chips (1) arranged on a wafer according to either of Claims 9 and 10, characterized in that at least one optical actuator is operated as a sensor in the second alignment step.