The present invention is related generally to optical interconnects using microLEDs, and more particularly to optical interconnects with a flexible waveguide.
Computing and networking performance requirements are seemingly ever-increasing. Prominent applications driving these requirements include data center servers, high-performance computing clusters, artificial neural networks, and network switches.
For decades, dramatic integrated circuit (IC) performance and cost improvements were driven by shrinking transistor dimensions combined with increasing die sizes, summarized in the famous Moore's Law. Transistor counts in the billions have allowed consolidation onto a single system-on-a-chip (SoC) of functionality that was previously fragmented across multiple ICs. However, Moore's Law appears to be reaching its limits as shrinking feature sizes below 10 nm results in decreasing marginal performance benefits with decreased yields and increased per-transistor costs.
Beyond these limitations, a single IC can only contain so much functionality, and that functionality is constrained because the IC's process cannot be simultaneously optimized for different functionality, e.g., logic, DRAM, and I/O. Increasingly, improving system performance is dependent on implementing very high bandwidth interconnects between multiple ICs.
Unfortunately, compared to the on-chip connections, today's chip-to-chip connections are typically much less dense and require far more power (for example normalized as energy per bit). These inter-IC connections are currently significantly limiting system performance. Specifically, the power, density, latency, and distance limitations of interconnects are far from what is desired.
New interconnect technologies that provide significant improvements in multiple performance aspects are highly desirable. It is well-known that optical interconnects may have fundamental advantages over electrical interconnects, even for relatively short interconnects of <<1 meter. Unfortunately, implementation of optical interconnects for inter-IC connections may face a host of problems. Included in these problems is that of coupling light from one IC to another IC. Electrical interconnect technology for inter-IC communications at a substrate or circuit board level may be relatively well-developed. The same may not be as true for optical interconnect technology for inter-IC communications, particularly for high-throughput applications that preferably do not negatively impact existing modes of electrical interconnections.
Aspect of some embodiments may utilize a high-speed array of LEDs coupled to an array of waveguides, where the waveguides may be formed on a shaped or flexible substrate in a largely one-dimensional array, where the LEDs may operate at speeds greater than 1 Gb/s. In some embodiments, the waveguides may have layers that are thin to allow tight bending of the assembly. In some embodiments, the waveguides may be bent from the horizontal to the vertical to allow coupling to an array of emitters or detectors that may be mounted face-up on the circuit board. In some embodiments, the waveguide assembly may have a non-rectangular shape, which allows the connection at the two ends to be arbitrarily positioned. In some embodiments, the flexible substrate may comprise metallic and dielectric layers, where some of the metallic layers may be patterned to create electrical transmission lines with well-defined impedance characteristics. In some embodiments, the optical waveguides layers may comprise dielectric layers in a stack of metallic and dielectric layers, where some of the metallic layers may be patterned to create electrical transmission lines with well-defined impedance characteristics. Some embodiments provide a parallel optical interconnect, comprising: a first optical transceiver including a plurality of first microLEDs and a plurality of first photodetectors, the first photodetectors monolithically integrated on a first silicon integrated circuit and the first microLEDs on the first silicon integrated circuit; a second optical transceiver including a plurality of second microLEDs and a plurality of second photodetectors, the second photodetectors monolithically integrated on a second silicon integrated circuit and the second microLEDs on the second silicon integrated circuit; a flexible substrate; and a plurality of waveguides on the flexible substrate, the waveguides being part of an optical link between the first optical transceiver and the second optical transceiver.
In some embodiments the waveguides comprise waveguide cores on cladding on the flexible substrate. In some embodiments the waveguides form a one-dimensional array of waveguides. In some embodiments the waveguides form two or more layers of waveguides. In some embodiments the waveguides have a numerical aperture between 0.2 and 0.7. In some embodiments the waveguides are multimode waveguides for wavelengths between 400 nm to 500 nm. In some embodiments the flexible substrate is comprised of a laminate of dielectric and metallic layers. In some embodiments at least some of the metallic layers are patterned to form electrical transmission lines. In some embodiments at least two of the layers of waveguides are separated by a metal layer. In some embodiments the metal layer is patterned to form electrical transmission lines. In some embodiments the first silicon integrated circuit includes first drivers for the first microLEDs and first electrical receiver circuitry for the first photodetectors, and the second silicon integrated circuit includes second drivers for the second microLEDs and second electrical receiver circuitry for the second photodetectors. In some embodiments a first end of the waveguides is in a first connector housing, the first connector housing including a first alignment feature, and wherein a second end of the waveguides is in a second connector housing, the second connector housing including a second alignment feature. In some embodiments the first alignment feature comprises a first connector alignment hole and the second alignment feature comprises a second connector alignment hole. In some embodiments the first connector housing is on a first standoff on the first silicon integrated circuit and the second connector housing is on a second standoff on the second silicon integrated circuit. In some embodiments the first standoff includes a first standoff alignment hole and the second standoff includes a second standoff alignment hole, with a first alignment pin in the first connector alignment hole and the first standoff alignment hole, and with a second alignment pin in the second connector alignment hole and the second standoff alignment hole.
These and other aspects of the invention are more fully comprehended upon review of this disclosure.
Some embodiments use LED with structures that can be modulated at high speeds due to their short carrier lifetime. In some embodiments, these LEDs are used in optical interconnects where the communication interconnect may comprise one or a plurality of optical communication links in the form of a parallel optical interconnect. In some embodiments, the very low power and high density of LED-based parallel optical interconnects are used in mobile platforms such as mobile phones, tablets, and laptop computers. In some embodiments, the LED-based parallel optical interconnects have very low power and high density, which may be preferable in those mobile platforms, where both power and space may be at a premium.
In some embodiments, for example as illustrated in
In some embodiments, a microLED is distinguished from a semiconductor laser (SL) as follows: (1) a microLED does not have an optical resonator structure; (2) the optical output from a microLED is almost completely spontaneous emission, whereas the output from a SL is dominantly stimulated emission; (3) the optical output from a microLED is temporally and spatially incoherent, whereas the output from a SL has significant temporal and spatial coherence; (4) a microLED is designed to be driven down to a zero minimum current, whereas a SL is designed to be driven down to a minimum threshold current, which is typically at least 1 mA.
In some embodiments, a microLED is distinguished from a standard LED by (1) having an emitting region of less than 10 μm×10 μm; (2) frequently having cathode and anode contacts on top and bottom surfaces, whereas a standard LED typically has both positive and negative contacts on a single surface; (3) typically being used in large arrays for display and interconnect applications.
In some embodiments, each microLED used in a parallel optical interconnect is driven with a current in the range of 10 uA to 500 uA. In some embodiments, the per-bit energy consumed by each lane of a parallel optical interconnect is in the range of 0.05 pJ/bit to 1 pJ/bit.
In some embodiments, the waveguides are made from polymer. In some embodiments, each waveguide comprises a core surrounded by a top cladding and bottom cladding, where the cladding has a lower index of refraction than the core. In some embodiments, each waveguide core has an approximately rectangular cross-section. In some embodiments, there may be multiple layers of waveguides.
As shown in
For relatively short reach interconnects (<1 meter) used in a typical mobile platform, a high numerical aperture is desired with a relatively large index difference between the core and the cladding. This allows the waveguides to capture more of the light from the LEDs. In some embodiments, the waveguides have an numerical aperture (NA) in the range of 0.2 to 0.7.
In some embodiments, the waveguide core material is comprised of a higher index polymer or epoxy such as SU8, while the lower and upper waveguide is comprised of a lower index polymer or epoxy such as PMMA. There are other materials that may be appropriate. In some embodiments of polymer waveguides, fluorination or other materials are used to tailor the refractive indices of the cladding and/or core layers.
The waveguides may be multimode at the wavelength of operation, which in some embodiments is in the range of 400 nm-500 nm. In some embodiments, the waveguide core transverse dimensions are in the range of 10 um to 50 um. In some embodiments, the waveguide transverse dimensions are in the range of 50 um to 100 um. In some embodiments, the spacing between waveguide cores is in the range of 3 um to 50 um. For instance, an array of 50 waveguides, each with a 40 um core width and spaced 10 um apart would be 2.5 mm wide. If each waveguide carries 4 Gb/s, the 50-lane parallel optical interconnect would carry 200 Gb/s of data—all typical numbers useful in computing applications.
The flex circuit may have two or more metallic layers 415a-c that are continuous along a length of the cross-section of the flex circuit and may have the dielectric layers 413a-b therebetween. In some embodiments, an inner metallic layer 415a may be attached to the bottom cladding of the waveguide array. A first dielectric layer 413a may be between and fill the space between the inner metallic layer and an intermediate metallic layer 415b. In some embodiments, a second dielectric layer 413b may exist between the intermediate metallic layer and an outer metallic layer 415c. In some embodiments, the metallic layers, or some of them, may have metallic patterns 423 within the dielectric layers and in between the continuous metallic layers to form high-speed electrical transmission line structures with well-defined impedance, such as striplines, microstrips, and coplanar waveguides. The metallic layers and patterns may also comprise various power and ground plane structures.
In some embodiments, the metallic layers, or some of them, may have metallic patterns 513, or otherwise electrical transmission line structures, and such metallic patterns may be embedded between the layers of the waveguide array and in between the waveguide cores 515. In some embodiments, the optical waveguide regions, including the waveguide cores 515 and/or the cladding layers 517a-d, may also serve as dielectric regions that are part of electrical interconnect structures such as striplines, microstrips, and coplanar waveguides. The benefits of this structure are (1) higher area density of interconnects; (2) fewer layers needed to be fabricated to support a given optical and electrical connectivity.
In some embodiments of a parallel optical interconnect, optical waveguides on a flexible substrate are optically coupled to an array of optoelectronic components comprising a plurality of microLEDs and/or photodetectors. In some embodiments, the waveguide cores 615 of the optical waveguides may be formed on and in-between claddings 617. In some embodiments, the photodetectors are monolithically integrated on a silicon integrated circuit with electrical receiver circuitry, possibly comprising transimpedance amplifiers, other receiver circuitry, and/or digital logic circuitry. In some embodiments, the microLEDs are fabricated on another substrate and then transferred onto a silicon integrated circuit (IC). In some embodiments, the silicon IC comprises transmitter circuitry such as drivers for the LEDs and/or other analog and digital circuitry. In some embodiments, the same integrated circuit comprises both transmitter and receiver circuitry.
Some embodiments include components to increase or possibly maximize the efficiency in coupling light from each microLED to its respective waveguide. In some embodiments, and as shown in
Some embodiments include features that may further enhance the usefulness of the use of the flexible optical transmission medium. For example, in some embodiments, there is one or more layer of waveguides on each of the two surfaces of the flexible substrate. In some embodiments, there is a mirror formed in the connector or in each waveguide that causes the light to turn 90° without requiring a tight bend. In some embodiments, there may be no bend in the waveguides, with the emitters and detectors mounted vertically on the board, for example as shown in
In some embodiments of a parallel optical interconnect, each waveguide supports a bidirectional link, where there is a both a microLED and photodetector at each end of each waveguide. In some further embodiments, a microLED is mounted on top of a larger photodetector. In some further embodiments, the parallel optical interconnect is used in “half-duplex” mode such that either the microLED or the detector at each end of each waveguide can be electrically activated to set the direction of transmission. Mechanically, there are a host of methods of aligning the connectors to the optical device with various latching mechanisms.
Although the invention has been discussed with respect to various embodiments, it should be recognized that the invention comprises the novel and non-obvious claims supported by this disclosure.
This application claims the benefit of U.S. Provisional Patent Application No. 63/272,106, filed on Oct. 26, 2021, the disclosure of which is incorporated by reference herein.
Number | Date | Country | |
---|---|---|---|
63272106 | Oct 2021 | US |