Method and apparatus for optically aligning integrated circuit devices

[0001] This application hereby claims priority under 35 U.S.C. 119 to U.S. Provisional Patent Application No. 60/465,316, filed on 24 Apr. 2003, entitled, “A Light Emission Receiving Alignment Technique,” by inventors Robert J. Bosnyak and Robert J. Drost (Attorney Docket No. SUN-P9595PSP).

[0002] This invention was made with United States Government support under Contract No. NBCH020055 awarded by the Defense Advanced Research Projects Administration. The United States Government has certain rights in the invention.

[0003] 1. Field of the Invention

[0004] The present invention relates to integrated circuit devices. More specifically, the present invention relates to a method and an apparatus for optically aligning integrated circuit devices.

[0005] 2. Related Art

[0006] Advances in semiconductor technology presently make it possible to integrate large-scale systems, including tens of millions of transistors, into a single semiconductor chip. Integrating such large-scale systems onto a single semiconductor chip increases the speed at which such systems can operate, because signals between system components do not have to cross chip boundaries, and are not subject to lengthy chip-to-chip propagation delays. Moreover, integrating large-scale systems onto a single semiconductor chip significantly reduces production costs, because fewer semiconductor chips are required to perform a given computational task.

[0007] Unfortunately, these advances in semiconductor technology have not been matched by corresponding advances inter-chip communication technology. Semiconductor chips are typically integrated onto a printed circuit board that contains multiple layers of signal lines for inter-chip communication. However, signal lines on a semiconductor chip are about 100 times more densely packed than signal lines on a printed circuit board. Consequently, only a tiny fraction of the signal lines on a semiconductor chip can be routed across the printed circuit board to other chips. This problem is beginning to create a bottleneck that continues to grow as semiconductor integration densities continue to increase.

[0008] Researchers have begun to investigate alternative techniques for communicating between semiconductor chips. One promising technique involves integrating arrays of capacitive transmitter plates and receiver plates onto semiconductor chips to facilitate inter-chip communication. If a first chip is situated face-to-face with a second chip so that transmitter plates on the first chip are capacitively coupled with receiver plates on the second chip, it becomes possible to transmit signals directly from the first chip to the second chip without having to route the signal through intervening signal lines within a printed circuit board. Other possible chip-to-chip communication techniques involve using optical fibers to couple light energy between chips, or using wire loops to couple magnetic fields between chips.

[0009] However, it is not a simple matter to align semiconductor chips with sufficient accuracy to facilitate chip-to chip communication. Conductive plates must align to couple electric fields. Optical fibers must align to couple light energy. Wire loops must align to couple electric fields. This alignment process must be extremely precise because these conductive plates, optical fibers and wire loops are generally a small fraction of the size of a semiconductor die, and in some cases can approach the smallest possible feature size that can be manufactured on a semiconductor substrate.

[0010] Hence, what is needed is a method and an apparatus for aligning semiconductor chips with sufficient accuracy to facilitate inter-chip communication.

[0011] One embodiment of the present invention provides a system that facilitates aligning a first semiconductor die with a second semiconductor die, wherein the first semiconductor die and the second semiconductor die are arranged active face to active face. Note that the active face contains circuitry for communicating between semiconductor dies. The system starts by generating light on an active face of the first semiconductor die. The system then collimates the light within the active face of the first semiconductor die to form a first beam of light which is projected onto the second semiconductor die. Next, the system receives the first beam of light on an active face of the second semiconductor die and determines a position of the first beam of light on the active face of the second semiconductor die. Finally, the system determines an alignment of the second semiconductor die relative to the first semiconductor die based on the determined position of the first beam of light.

[0012] In a variation on this embodiment, the system generates light on the active face of the first semiconductor die and collimates the light into a second beam of light, wherein the second beam of light is at an angle to the first beam of light. Next, the system receives the second beam of light on the active face of the second semiconductor die, and determines a position of the second beam of light on the active face of the second semiconductor die. Finally, the system determines a distance between the first semiconductor die and the second semiconductor die by analyzing the position of the second beam of light on the active face of the second semiconductor die compared to the position of the first beam of light on the active face of the second semiconductor die.

[0013] In a variation on this embodiment, the light waves are generated by a Zener diode, a light emitting diode (LED), a vertical cavity surface emitting laser (VCSEL), or an avalanche breakdown PN diode.

[0014] In a variation on this embodiment, the first beam of light is collimated by shining the first beam of light through annuli created on metal layers on the first semiconductor die.

[0015] In a further variation, the shape of the annuli can include squares, circles, or octagons.

[0016] In a further variation, the angle of the first beam of light can be determined by creating the annuli in the metal layers such that a line through the centers of the annuli is set at a desired angle from the perpendicular axis of the active face of the first semiconductor die.

[0017] In a variation on this embodiment, the first beam of light is received on the active face of the second semiconductor die by a p-n junction photo-detector.

[0018] In a variation on this embodiment, the system further comprises a plurality of photo-detectors on the active face of the second semiconductor die for receiving the first beam of light.

[0019] In a further variation, the plurality of photo-detectors are arranged in a vector.

[0020] In a further variation, the plurality of photo-detectors are arranged in an array.

[0021] In a variation on this embodiment, the system further comprises generating light on the active face of the first semiconductor die and collimating the light into a third beam of light, wherein the third beam of light is not coplanar with the first and second beams of light. Next, the system receives the third beam of light on the active face of the second semiconductor die, and determines a position of the third beam of light on the active face of the second semiconductor die. Finally, the system determines an orientation of the first semiconductor die and the second semiconductor die in all six degrees of alignment by analyzing the positions of the first, second, and third beams of light on the active face of the second semiconductor die

[0032] The following description is presented to enable any person skilled in the art to make and use the invention, and is provided in the context of a particular application and its requirements. Various modifications to the disclosed embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the present invention. Thus, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

Semiconductor Dies

[0033] FIG. 1 illustrates two semiconductor dies in accordance with an embodiment of the present invention. One embodiment of the present invention provides a mechanism that aligns semiconductor die 102 to semiconductor die 104 to facilitate inter-chip communication between semiconductor die 102 and semiconductor die 104. As is illustrated in FIG. 1, semiconductor dies 102 and 104 are arranged active-face to active-face so that communication features on semiconductor dies 102 and 104 are aligned with each other. In one embodiment of the present invention, the alignment mechanism is optical. Hence, during the alignment process, light is emitted from semiconductor die 102 and received by semiconductor die 104.

[0034] In one embodiment of the present invention, two forms of emitting and receiving arrays are used. A first form, uses light collimated perpendicular to the die surface to detect the X-Y position. A second form uses light aligned at an angle (which is compensated for dissimilar refractive indexes) to detect the separation (Z) between the dies. This technique improves upon visual alignment and prior electrical alignment techniques because the expected resolution is only limited by the wavelength of the emitted light and the ability to collimate that light. Moreover, this technique allows large chip separation and uses less die area than existing alignment techniques.

Metal Layer Arrangements

[0035] FIG. 2 illustrates two possible metal layer arrangements for collimating light in accordance with an embodiment of the present invention. Note that in FIG. 1, semiconductor die 102 contains a sparse array of light emitters, and semiconductor die 104 contains a larger array of light receivers. Moreover, semiconductor die 102 uses patterns in the metal layers to collimate light. For example, annuli stack 200 comprises vertically stacked annuli in the metal layers on semiconductor die 102. Note that annuli stack 200 collimates light generated by diode 204 at right angles to the surface of semiconductor die 102. The receiving array on semiconductor die 104 detects the x-y position of the emitter array by scanning the state of the light intensity over the larger array.

[0036] Another set of emitting and receiving arrays sets the collimation at an angle so that the emitters and receivers yield the separation distance Z between the semiconductor dies. The light can be angled by offsetting the centers of the annuli in the metal layers. For example, in FIG. 2 light emitted from diode 206 is collimated by passing it through annuli stack 202, which is arranged at an angle from the perpendicular axis of semiconductor die 102. Furthermore, the angle of collimation from the perpendicular axis is chosen to compensate for the change in the index of refraction between the dielectric of the dies and the interstitial medium between the dies. Note that the angle between the dies is derived from two X-Y positions located at different points on the receiving die.

[0037] The emitters can be one of many possible light sources. Zener diodes, light emitting diodes (LED), vertical cavity surface emitting laser (VCSEL), avalanche breakdown PN diodes, etc. Furthermore, these sources can be coherent or non-coherent. Note that VCSELs yield non-coherent light at currents less than the lasing current. One possible emitter for standard CMOS technology is avalanched diode light emission because no extra processing step are required to make this emitter. A simple photo-detector receiver, such as a p-n junction photo-detector, is adequate for detecting position.

Metal Layer Structures

[0038] FIG. 3A illustrates metal layer annuli arranged to produce a 90-degree beam of light in accordance with an embodiment of the present invention. Stacked apertures in the metal layers collimate the generated light. More specifically, the aperture can be formed by annuli stacked with their centers set at the desired angle from the perpendicular axis of the die surface. In another example, each metal layer may have two metal strips separated by the minimum metal separation of 0.6 microns for a 0.4 micron technology. In each alternate metal layer, strips are aligned at right angles to each other. The stacked aperture will be 0.6 microns or wavelength of orange light. An array of 64 receivers (8 by 8) will take up about 100 square microns for a 0.4 micron technology. Note that an array of one emitter is sufficient to determine the local X-Y position. However, an array of a few of emitters can be used to implement a vernier scheme for interpolating the X-Y position. More advanced technology will allow even further resolution in a smaller space.

[0039] FIG. 3B illustrates metal layer annuli arranged to produce a 45-degree beam of light in accordance with an embodiment of the present invention. FIG. 3C illustrates additional annuli shapes in accordance with an embodiment of the present invention.

[0040] The simplest receivers in a CMOS technology are reverse-biased diodes. Light absorbed in these diodes creates electron-hole pairs in the semiconductor space charge region that produces a detectable current. This process is slow compared to normal switching speeds of CMOS circuitry but nevertheless is detectable in a time frame allowing detection of the position and re-alignment between motion and detection phases. The circuit may additionally retain state by charge storage or may have a memory element associated with each receiver site. Scanning the state of the receivers yields the relative position of the dies.

Compensating for Refraction

[0041] FIG. 4 illustrates compensating for refraction in accordance with an embodiment of the present invention. For separation measurements the index of refraction must be considered. The refractive light passing from one media to another is describe by the following formula:
N₁·sin(A₁)=N₂·sin(A₂)
where N₁, and N₂ are the indices of refraction for each media and A₁ and A₂ are the angles of the light beams. When light passes through a dense medium to a media with different optical density, the light is refracted. Hence, light passing through the layers of SiO₂ to air is refracted. For a set of two dies separated by an air-gap the differences of refracting between SiO₂/air and air/SiO₂ requires a collimated angle of 29 degrees from the perpendicular because the ratio of N₂/N₁ is about 0.66. For a set of dies separated by oil, as currently used in recent experiments, the SiO₂/oil collimating angle can be set at 45 degrees because the index of refraction for SiO₂ is 1.47 and oil is about 1.46; a lucky circumstance.
Offset Detection Circuit

[0042] FIG. 5 presents a circuit diagram for an offset detection system in accordance with an embodiment of the present invention. Light generated by emitter diode 502 on semiconductor die 102 is received by photo-detector 504 on semiconductor die 104. The output of the receiving diode is scanned by row and column using the row and column pass transistors illustrated in FIG. 5 to determine the location of the received light on semiconductor die 104. Techniques for scanning row and column are well-known in the art and will not be discussed further herein.

Photo-Detector Arrays

[0043] FIG. 6 illustrates photo-detector arrays 602-606 in accordance with an embodiment of the present invention. As is illustrated in FIG. 6, photo-detector arrays 602-606 on semiconductor die 104 are one-dimensional arrays, wherein each array receives a beam of light generated by a corresponding light source on semiconductor die 102. Note that smaller arrays require semiconductor dies 102 and 104 be close to optimal alignment before beams of light generated on semiconductor 102 are detected on semiconductor 104. Whereas, larger arrays allow detection with coarser alignments. However, larger arrays also consume more area on the receiving die.

Six Degrees of Alignment

[0044] FIG. 7 illustrates the possible six degrees of alignment between semiconductor dies 102 and 104. In one embodiment of the present invention, light beam detection and subsequent analysis allows x, y, z, θ, Ψ, and Φ alignments to be determined.

Process of Aligning Semiconductor Dies

[0045] FIG. 8 presents a flowchart illustrating the process of aligning two semiconductor dies in accordance with an embodiment of the present invention. The system starts when light is generated on semiconductor 102 (step 802). This light is collimated via structures in the metal layers on semiconductor 102 into a first beam of light which is projected onto semiconductor die 104 (step 804). Next, the incident beam of light is detected at a specific position of semiconductor 104. At the same time, the system generates light from another light source on semiconductor 102 (step 810) and collimates the light via structures in the metal layers on semiconductor 102 into a second beam of light with an angle different than the first beam of light (step 812). Next, the second beam of light is detected at a specific position of semiconductor 104 (step 814). Finally, the system determines the orientation of semiconductor 104 to semiconductor 102 based on the detected positions of the first and second beams.

[0046] Note that with three or more beams of light, it is possible to determine all six x, y, θ, z, Ψ, and Φ alignments.

[0047] The foregoing descriptions of embodiments of the present invention have been presented for purposes of illustration and description only. They are not intended to be exhaustive or to limit the present invention to the forms disclosed. Accordingly, many modifications and variations will be apparent to practitioners skilled in the art. Additionally, the above disclosure is not intended to limit the present invention. The scope of the present invention is defined by the appended claims.