Electronic-integration compatible photonic integrated circuit and method for fabricating electronic-integration compatible photonic integrated circuit
RE048379 ยท 2021-01-05
Assignee
Inventors
Cpc classification
G02B6/43
PHYSICS
G02F1/0311
PHYSICS
International classification
G02B6/00
PHYSICS
H01L21/00
ELECTRICITY
G02B6/43
PHYSICS
Abstract
An electronic-integration compatible photonic integrated circuit (EIC-PIC) for achieving high-performance computing and signal processing is provided. The electronic-integration compatible photonic integrated circuit comprises a plurality of electronic circuit structures and a plurality of photonic circuit structures. The electronic and photonic circuit structures are integrated by a process referred to as monolithic integration. An electronic circuit structure includes one or more electronic devices and a photonic circuit structure includes one or more photonic devices. The integration steps of electronic and photonic devices are further inserted into standard CMOS process. The photonic circuit structures and the electronic circuit structures are integrated to form the electronic-integration compatible photonic integrated circuit device.
Claims
.[.1. A method for integrating one or more photonic devices on a substrate with a semiconductor device, wherein the semiconductor device comprises one or more electronic devices, the method comprising: a. creating a first optical cladding layer on a first substrate; b. creating a first optical core layer on top of the first optical cladding layer, wherein the first optical cladding layer and the first optical core layer form a first optical waveguide; c. creating a first optical-fiber coupling lens on the first optical waveguide; d. creating a first optical device, wherein an optical beam in the first waveguide propagates substantially in the first optical core layer; e. optically connecting the first optical device to the first waveguide; f. creating a first interspaced dielectric structure layer on top of the first optical core layer; g. creating a second optical core layer on top of the first interspaced dielectric structure layer; and h. creating a second optical device with optical beam propagating substantially in the second optical core layer; i. mounting a first optical fiber on the first substrate to facilitate a coupling of light beam energy between the first optical fiber and the first optical waveguide..].
.[.2. The method of claim 1, wherein the first substrate is a silicon-on-insulator wafer, the first optical cladding layer is a silicon dioxide layer, and the first optical core layer is a silicon-on-insulator wafer..].
.[.3. The method of claim 1, wherein the first optical-fiber coupling lens is a superlens..].
.[.4. The method of claim 1, wherein the first optical device is one of an optical waveguide, an optical wavelength multiplexer, an optical wavelength demultiplxer, an optical grating, an optical beam splitter, a polarization beam splitter, an optical isolator, a polarization rotator, an optical interferometer, an optical modulator, an optical ring resonator, an optical disk resonator, an optical curved reflector, an optical mirror, an optical amplifier, an optical detector, a laser, a light-emitting device, a nonlinear-optical device, a photonic transistor, a optical harmonic frequency generator, and an all-optical device..].
.[.5. The method of claim 1, wherein the first substrate comprises a trench to support the optical fiber..].
.[.6. The method of claim 1, wherein the first substrate comprises a V-groove to support the optical fiber..].
.[.7. The method of claim 1, wherein the optical beam enters the second optical device through a tapered waveguide structure made from the second optical core layer..].
.[.8. The method of claim 1, wherein the optical beam exits the second optical device through a tapered waveguide structure made from the second optical core layer..].
.[.9. The method of claim 1, wherein the second optical core layer is fabricated by wafer bonding..].
.[.10. The method of claim 1, wherein the second optical core layer is fabricated by local-area wafer bonding..].
.[.11. The method of claim 10, wherein the steps a-e are performed by using a process used in fabricating electronic devices..].
.[.12. The method of claim 1, wherein the electronic devices are one of an electronic transistor, an electronic diode, a resistor, a capacitor, or an inductor..].
.[.13. The method of claim 1, wherein the electronic devices form a CMOS integrated circuit..].
.[.14. The method of claim 1, wherein the semiconductor device comprises one or more Complementary Metal-Oxide-Semiconductor (CMOS) devices, the method further comprising: a. growing a first dielectric layer and a second dielectric layer on a nano-waveguide and a disk; b. creating a deposition layer on the first dielectric layer and the second dielectric layer; c. creating a source region and a drain region by implanting N+ dopants and P+ dopants on the deposition layer, wherein an integrated structure is formed..].
.[.15. The method of claim 14, wherein the one or more Complementary Metal-Oxide-Semiconductor (CMOS) devices are integrated by using a standard CMOS fabrication process..].
.[.16. The method of claim 1, wherein the second optical core layer is a III-V compound semiconductor layer..].
.[.17. The method of claim 1 further comprising performing Quantum Well Intermixing (QWI) on the second optical core layer..].
.[.18. A semiconductor device manufactured by the method as claimed in claim 14..].
.[.19. The semiconductor device of claim 18 further comprising a plurality of electronic circuit structures..].
.[.20. The semiconductor device of claim 19, wherein the plurality of electronic circuit structures further comprise at least one of an active electronic device and a passive electronic device..].
.[.21. The semiconductor device of claim 18 further comprising a plurality of photonic circuit structures..].
.[.22. The semiconductor device of claim 21, wherein the plurality of photonic circuit structures further comprise at least one of an active photonic device and a passive photonic device..].
.Iadd.23. A method for integrating one or more photonic devices on a substrate, the method comprising: a. providing a substrate with one or more electronic devices on one or more connected or disjointed electronic device areas on a first substrate; b. providing an area of the first substrate having a layer structure with a first optical cladding layer and a first optical core layer, wherein part of the first optical cladding layer and the first optical core layer form a first optical waveguide; c. providing a first optical device; d. optically connecting the first optical device to the first waveguide; e. creating a second optical core layer above the first waveguide via a local-area wafer bonding process comprising: i. creating a first material layer structure on a second substrate, wherein the first material layer structure comprises plurality of material layers in which at least one of the plurality of material layers comprises a second optical core layer, ii. attaching a handling fixture to the first material layer structure, iii. removing the second substrate, leaving the handling fixture attached to the first material layer structure without the second substrate such that one or more areas on the first material layer structure, that will not be used for fabricating photonic devices and will later cover over the first electronic device areas during the wafer bonding process, are pre-etched away with indentation; and f. bonding the first material structure to the top surface of the first waveguide either directly or via an additional dielectric layer between the first material layer structure and the first optical core..Iaddend.
.Iadd.24. The method of claim 23, further comprising creating the first optical device along with the first optical core layer such that an optical beam in the first optical device propagates substantially in the first optical core layer..Iaddend.
.Iadd.25. The method of claim 24, wherein the first optical device is one of an optical waveguide, an optical wavelength multiplexer, an optical wavelength demultiplexer, an optical grating, an optical beam splitter, a polarization beam splitter, an optical isolator, a polarization rotator, an optical interferometer, an optical modulator, an optical ring resonator, an optical disk resonator, an optical curved reflector, an optical mirror, an optical amplifier, an optical detector, a laser, a light-emitting device, a nonlinear-optical device, a photonic transistor, an optical harmonic frequency generator, and an all-optical device..Iaddend.
.Iadd.26. The method of claim 23, further comprising creating a second optical device along with the second optical core layer such that an optical beam in the second optical device propagates substantially in the second optical core layer..Iaddend.
.Iadd.27. The method of claim 23 wherein the first optical cladding layer is air or dielectric material..Iaddend.
.Iadd.28. The method of claim 23, wherein one or more electronic devices or one or more photonic devices are created on the substrate or any of the optical layers..Iaddend.
.Iadd.29. The method of claim 28, wherein the electronic devices are one of an electronic transistor, an electronic diode, a resistor, a capacitor, or an inductor..Iaddend.
.Iadd.30. The method of claim 28, wherein the electronic devices form a CMOS integrated circuit..Iaddend.
.Iadd.31. The method of claim 28, wherein the semiconductor device comprises one or more Complementary Metal-Oxide-Semiconductor (CMOS) devices, the method further comprising: a. growing a first dielectric layer and a second dielectric layer on a nano-waveguide and a disk; b. creating a deposition layer on the first dielectric layer and the second dielectric layer; and c. creating a source region and a drain region by implanting N+ dopants and P+ dopants on the deposition layer, wherein an integrated structure is formed..Iaddend.
.Iadd.32. The method of claim 31, wherein the one or more Complementary Metal-Oxide-Semiconductor (CMOS) devices are integrated by using a standard CMOS fabrication process..Iaddend.
.Iadd.33. A semiconductor device manufactured by the method as claimed in claim 31..Iaddend.
.Iadd.34. The semiconductor device of claim 33 further comprising a plurality of electronic circuit structures..Iaddend.
.Iadd.35. The semiconductor device of claim 34, wherein the plurality of electronic circuit structures further comprise at least one of an active electronic device and a passive electronic device..Iaddend.
.Iadd.36. The semiconductor device of claim 33 further comprising a plurality of photonic circuit structures..Iaddend.
.Iadd.37. The semiconductor device of claim 36, wherein the plurality of photonic circuit structures further comprise at least one of an active photonic device and a passive photonic device..Iaddend.
.Iadd.38. The method of claim 23, wherein the first substrate is a silicon-on-insulator wafer, the first optical cladding layer is a silicon dioxide layer, and the first optical core layer is the top silicon layer of a silicon-on-insulator wafer..Iaddend.
.Iadd.39. The method of claim 23, wherein the first substrate comprises a trench to support an optical fiber..Iaddend.
.Iadd.40. The method of claim 23, wherein the first substrate comprises a V-groove to support an optical fiber..Iaddend.
.Iadd.41. The method of claim 23, wherein the optical beam enters the second optical device through a tapered waveguide structure made from the second optical core layer..Iaddend.
.Iadd.42. The method of claim 23, wherein the optical beam exits the second optical device through a tapered waveguide structure made from the second optical core layer..Iaddend.
.Iadd.43. The method of claim 23, wherein the steps a-d are performed by using a process used in fabricating electronic devices..Iaddend.
.Iadd.44. The method of claim 23, wherein the second optical core layer is a III-V compound semiconductor layer..Iaddend.
.Iadd.45. The method of claim 23 further comprising performing Quantum Well Intermixing (QWI) on the second optical core layer..Iaddend.
.Iadd.46. The method of claim 23, wherein a first coupling lens further couples light between an optical fiber and a waveguide on the substrate..Iaddend.
.Iadd.47. The method of claim 46, wherein the first optical-fiber coupling lens is a superlens..Iaddend.
.Iadd.48. The method of claim 46, wherein the first coupling lens is a superlens and the superlens has a width WSPL, height HSPL and length LSPL fabricated on top of the waveguide by depositing multi-layers of alternating dual materials..Iaddend.
.Iadd.49. The method of claim 48, wherein the dual materials are Silicon (Si) and Silicon Dioxide (SiO2)..Iaddend.
.Iadd.50. The method of claim 48, wherein the superlens dimensions are WSPL=20 m, HSPL=15 m, and LSPL=20 m..Iaddend.
.Iadd.51. The method of claim 49, wherein the thickness ratio of Silicon to Silicon Dioxide is varied such that the effective index of the superlens is n=3.5 at the bottom of the lens near the surface of the waveguide and is at least partially of silicon, and the effective index of the superlens is n=1.5 near the top of the superlens and is at least partially of silicon dioxide..Iaddend.
.Iadd.52. The method of claim 46, wherein the optical fiber is an optical fiber with a full-width-half maximum (FWHM) mode size diameter of 8 m..Iaddend.
.Iadd.53. The method of claim 51, wherein the index variation of the superlens is such that the mode from the waveguide will expand in the lens such that the strong lensing effect in the superlens results in an output beam of a larger diameter..Iaddend.
.Iadd.54. The method of claim 53, wherein the wavefront is a flat wavefront at the output surface of the superlens..Iaddend.
.Iadd.55. The method of claim 46, wherein the optical fiber is held from below by a structure having a V-groove or a trench configuration, fabricated on the substrate that holds the optical fiber at the right height with respect to the lens optical center to achieve maximum coupling of light beam energy between the optical fiber and the waveguide on chip..Iaddend.
.Iadd.56. The method of claim 55, wherein the structure is a trench of rectangular shape or of a shape adapted to hold the fiber in steady position..Iaddend.
.Iadd.57. The method of claim 55 wherein additional materials between substrate and fiber hold the fiber in place..Iaddend.
.Iadd.58. The method of claim 23, wherein the Bonding process is one that involves at least the following steps: a. making a material layer structure on top of a wafer substrate, the material layer including an etch-stop layer first made on the substrate; b. making a structure containing the second optical core layer on top of the etch-stop layer; c. bonding a handling wafer to the second optical core layer side of the wafer, exposing the substrate; d. etching the exposed substrate by an etching process that stops at the etch-stop layer, resulting in a new exposed surface; e. initiating a wafer bonding process by pressing the new exposed surface onto the material at the top surface of the first optical core layer under pressure and elevated temperature; f. removing the handling wafer resulted in the second optical core layer above the first optical core layer..Iaddend.
.Iadd.59. The method of claim 58, wherein the wafer substrate is indium phosphide (InP)..Iaddend.
.Iadd.60. The method of claim 59, wherein the etch-stop layer in claim 59 is InGaAsP compound semiconductor material..Iaddend.
.Iadd.61. The method of claim 58, wherein the second optical core layer is an active optical material layer..Iaddend.
.Iadd.62. The method of claim 61, wherein the second optical core layer contains semiconductor quantum wells..Iaddend.
.Iadd.63. The method of claim 58, wherein the etched stop layer and active optical material layer are grown epitaxially on top of an InP substrate..Iaddend.
.Iadd.64. The method of claim 58, wherein the handling wafer is bonded to the second optical core layer side of the wafer using Benzocyclobutene (BCB) polymer..Iaddend.
.Iadd.65. The method of claim 58, wherein the etching process of step (d) of claim 58 is a selective chemical etching process that etches an InP substrate material at a rate faster than etching an etch-stop InGaAsP layer material..Iaddend.
.Iadd.66. The method of claim 65, wherein the chemical etching is done by hydrochloric acid (HCl)..Iaddend.
.Iadd.67. The method of claim 58, wherein after the wafer bonding, resulted in a Direct Double Optical Layer structure for an Electronic Integration Compatible Photonic Integrated Circuit (DDOL-EIC-PIC)..Iaddend.
.Iadd.68. The method of claim 58, wherein a thin interspaced dielectric layer is grown on the new exposed surface of step (d) and a thin interspaced dielectric layer is also grown on the top surface of the first optical core..Iaddend.
.Iadd.69. The method of claim 68, the interspaced dielectric layer is silicon nitride..Iaddend.
.Iadd.70. The method of claim 58, wherein in step (e), wafer bonding process is initiated by pressing the thin interspaced dielectric layer on the new exposed surface of step (d) against the interspaced dielectric layer on the top surface of the first optical core under external pressure and elevated temperature..Iaddend.
.Iadd.71. The method of claim 70, wherein the wafer bonding, results in an Indirect Double Optical Layer structure for an Electronic Integration Compatible Photonic Integrated Circuit (IDDOL-EIC-PIC)..Iaddend.
.Iadd.72. The method of claim 58, wherein in step (e) of claim 58, the surfaces to be bonded is subjected to the plasma treatment to create hydrophilic surfaces before bonding..Iaddend.
.Iadd.73. The method of claim 72, wherein the plasma treatment is achieved with 1-minute exposure to oxygen plasma..Iaddend.
.Iadd.74. The method of claim 70, wherein the interspaced dielectric layer surfaces to be bonded is subjected to the plasma treatment to create hydrophilic surfaces before bonding..Iaddend.
.Iadd.75. The method of claim 74, wherein the plasma treatment is achieved with 1-minute exposure to oxygen plasma..Iaddend.
.Iadd.76. The method of claim 58, wherein in step (e), the bonding is achieved at an elevated temperature above 150 C. and less than 400 C..Iaddend.
.Iadd.77. The method of claim 58, wherein the handling wafer is InP wafer and is selectively etched away using HCL in step (f)..Iaddend.
.Iadd.78. The method of claim 23, wherein the wafer bonding involves a second wafer made with the second optical core layer with a second surface..Iaddend.
.Iadd.79. The method of claim 78, wherein the top exposed surface of the first optical core layer on the substrate is the first surface, and wafer bonding involves pressing the second surface of the second optical core layer onto the first surface of the first optical core layer under pressure and elevated temperature..Iaddend.
.Iadd.80. The method of claim 79, wherein before the pressing of the two surfaces, certain areas of the second surface that are not later be fabricated with photonic devices are pre-etched away with indentations resulting in a new second surface in such a way that the during the pressing down process, the new second surface touches only the area designated as photonic areas on the substrate wafer and not the electronic areas on the substrate wafer..Iaddend.
.Iadd.81. The method of claim 80, wherein the wafer bonding process resulted in local-area wafer bonding in which one or plurality of areas designed as the photonic areas are bonded with the second optical core layer..Iaddend.
.Iadd.82. The method of claim 23 wherein an additional step is added that includes covering the photonic areas with protective materials during the high-temperature electronic fabrication or CMOS process, and then uncover the protected photonic areas by removing the photonic-area protective materials so that a fresh silicon surface is uncovered for performing the wafer bonding..Iaddend.
.Iadd.83. The method of claim 82 wherein the protective material is In2O3..Iaddend.
.Iadd.84. The method of claim 62 wherein a quantum-well intermixing process to shift the bandgap energy of the quantum well is performed before the wafer bonding..Iaddend.
.Iadd.85. The method of claim 84 wherein the quantum well intermixing is low-energy temperature-assisted ion-implantation quantum well intermixing (LETAI-QWI)..Iaddend.
.Iadd.86. The method of claim 62 wherein a quantum-well intermixing process to shift the bandgap energy of the quantum well is performed after the wafer bonding..Iaddend.
.Iadd.87. The method of claim 86 wherein the quantum well intermixing is low-energy temperature-assisted ion-implantation quantum well intermixing (LETAI-QWI)..Iaddend.
.Iadd.88. The method of claim 23, wherein the second optical core layer is made of material selected from the group consisting of: Indium Phosphide (InP), Gallium Arsenide (GaAs), Indium Gallium Arsenide (InGaAs), Indium Gallium Arsenide Phosphide (InGaAsP), Indium Aluminum Gallium Arsenide (InAlGaAs), Aluminum Arsenide (AlAs), Aluminum Gallium Arsenide (AlGaAs), Indium Gallium Aluminum Phosphide (InGaAlP), Indium Gallium Phosphide (InGaP), Gallium Nitride (GaN), Aluminum Nitride (AIN), Gallium Aluminum Nitride (GaAlN), Gallium Phosphide (GaP), Aluminum Phosphide (AlP), Aluminum Antimonide (AlSb), Gallium Antimonide (GaSb), Zinc selenide (ZnSe), Zinc Sulphide (ZnS), Cadmium Sulphide (CdS), Silicon Carbide (SiC), Silicon Germanium (SiGe), Indium Gallium Antimonide (InGaSb), or Indium Antimonide (InSb), Germanium (Ge), Silicon-Germanium (SiGe), or two or more of the combinations thereof..Iaddend.
.Iadd.89. The method of claim 23, wherein the dielectric material layer including the first optical cladding layer and the first optical core layer are made of material selected from the group consisting of: tantalum pentoxide (Ta2O5), zirconium oxide (ZrO2), niobium pentoxide (Nb2O5), hafnium oxide (HfO2), zinc oxide (ZnO), germanium oxide (GeO2), lead oxide (PbO), yttrium oxide (Y2O3), aluminum oxide (Al2O3), silicon carbide (SiC), titanium carbide (TiC), titanium nitride (TiN), chromium nitride (CrN), carbon nitride (CN), carbon boride (CB), aluminum nitride (AlN), zinc selenide (ZnSe), barium fluoride (BaF2), magnesium fluoride (MgF2), Diamond like Carbon (DLC), Benzocyclobutene (BCB), cyclized transparent optical polymer (CYTOP), or a polymer of imide monomers (Polyimide), or two or more of the combinations thereof..Iaddend.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) The preferred embodiments of the invention will hereinafter be described in conjunction with the appended drawings, provided to illustrate and not to limit the invention, wherein like designations denote like elements, and in which:
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
(52)
DESCRIPTION OF EMBODIMENTS
(53) Overview of EIC-PIC:
(54) The present invention provides a method for the realization of Photonic/Nanophotonic Integrated Circuits that are compatible with electronic integration, referred to as electronic-integration compatible Photonic/Nanophotonic integrated circuit that will be referred to as EIC-PIC. An EIC-PIC shall have photonic devices integrated on a substrate using fabrication processes that are compatible with electronic integration processes.
(55) Electronic Device Materials:
(56) Electronic devices are typically fabricated on a layer of semiconductor material made up of silicon, GaAs, InP, or GaN substrate. The most commonly used material is Silicon. For illustration but not limitation, we will focus our discussion on the case of Silicon electronics integrated with photonic devices. Generalization to the utilization of other materials for electronics other than Silicon will be obvious to those skilled in the art.
(57) Photonic Device Materials and Optical Wavelength:
(58) For illustration but not limitation, the photonic/nanophotonic devices are illustrated for operations at .sub.opt=1550 nm wavelength range. Generalization of the photonic devices for operation at other wavelength range will be obvious to those skilled in the art by scaling the device dimensions somewhat proportional to the wavelength of operation, by utilizing passive optical materials transparent at the wavelength of operation, and by utilizing active optical material that can provide optical gain at the wavelength of operation.
(59) Basic EIC-PIC Structure:
(60) A basic wafer structure for an EIC-PIC provided by the present invention is illustrated in
(61) In an exemplary embodiment, the thin layer of electronic material 1001 is also transparent to the optical wavelength of operation .sub.opt forming the first optical core layer and the dielectric layer 1011 has a refractive index that is lower than the refractive index of first optical core layer 1001, enabling layer 1001 to be used as a transparent optical waveguide core surrounded below by lower waveguide cladding 1011 and above by air as the upper waveguide cladding, both the upper and lower waveguide claddings have lower refractive index than waveguide core layer 1001. As shown in
(62) Single-Optical-Layer EIC-PIC:
(63) For the purpose of illustration but not limitation, as shown in
(64) In an exemplary embodiment, the first substrate layer 1021 is the silicon substrate of a silicon-on-insulator (SOI) wafer, the first lower optical cladding layer 1011 is the silicon dioxide layer on top of the silicon substrate in the SOI wafer, the first optical core layer 1001 is the top silicon layer of the SOI wafer, and the first upper optical cladding layer 1031 is a silicon dioxide layer.
(65) Input and Output Lens:
(66) In an exemplary embodiment of the invention, the SOL-EIC-PIC consists of at least one input or output optical-fiber coupling optics (i.e. it can have only one or more input coupling optics without any output coupling optics or only one or more output coupling optics without any input coupling optics. It can also have both input coupling optics and output coupling optics)
(67)
(68) The superlens 4030 or 4050 is further shown in
(69) In an exemplary embodiment, the dual materials are Silicon (Si) and Silicon Dioxide (SiO.sub.2). The superlens 5030 dimensions are W.sub.SPL=20 m, H.sub.SPL=15 m, and length L.sub.SPL=20 m. The thickness ratio of Silicon and Silicon Dioxide is varied such that the effective index of the superlens is close to n=3.5 at the bottom of the lens near the surface of waveguide 5020 and will be mainly of silicon (i.e. it will have thicker layers of silicon versus silicon dioxide). The effective index of the superlens is close to n=1.5 near the top of the superlens and will be mainly of silicon dioxide (i.e. it will have thicker layers of silicon dioxide versus silicon). The optical fiber 5010 is an optical fiber with a full-width-half maximum (FWHM) mode size diameter of about 8 m. The optical wavelength of operation .sub.OPT is 1550 nm. The index variation of the superlens 5030 with the design algorithm shown in U.S. Pat. No. 7,643,719 will be such that the mode from the nano waveguide 5020 will expand in the lens such that the strong lensing effect in the superlens results in an output beam of about 8 m in diameter over less the 20 m of propagation through the superlens, and the wavefront is a flat wavefront at the output surface of the superlens. The 8 m optical beam diameter at the output surface of the superlens with flat beam wavefront matches the fiber beam diameter of 8 m. Thus, the superlens 5030 enables efficient coupling of the light from the nano waveguide 5020 to the optical fiber 5010 or from the optical fiber 5010 to nano waveguide 5020 due to the reciprocity of light propagation as is known to those skilled in the art.
(70) In an exemplary embodiment, the input or output optical fiber is held from below by a V-groove or a trench structure fabricated on the substrate 1021 to match the size of the fiber and hold it at the right height with respect to the lens optical center to achieve maximum coupling of light beam energy between the optical fiber and the waveguide on chip. This is illustrated in
(71) Photonic or Nanophotonic Device:
(72) In an exemplary embodiment of the invention, as shown in
(73) The vertical optical mode confinement of photonic/nanophotonic device is due to the high refractive index core layer 1001 that confines the light energy to propagate mainly in core layer 1001. The horizontal confinement and propagation geometry of the optical beam or beams will depend on the device as described below. The horizontal geometry is defined by lithography (photolithography or E-Beam lithography etc) and by vertically etching down the core layer 1001, as is well know to those skilled in the art.
(74) An exemplary embodiment of an optical waveguide is a straight or curved waveguide 6000 shown in
(75) An exemplary embodiment of an optical wavelength multiplexer or an optical grating is an ultra-compact wavelength multiplexer/demultiplexer, also called the Super-Compact Grating or the Compact Curved Grating (hereinafter referred to as UCDeMux 7000 as shown in
(76)
(77) In one exemplary embodiment based on poly-silicon gate, the vertical cross section of Tunable Ring Resonator 8000 is shown in
(78) In another exemplary embodiment based on transparent conducting oxide (TCO) gate, the vertical cross section of Tunable Ring Resonator 8000 is shown in
(79) Optical beam splitter, optical polarization beam splitter, optical isolator, polarization rotator, optical interferometer, optical phase shifter, optical modulator, optical ring resonator, optical disk resonator, optical curved reflector, optical mirror, nonlinear-optical device, photonic transistor, optical harmonic frequency generator, all-optical device, are some of the examples of common integrated optical devices that are well known to those skilled in the art. For example, an optical interferometer can be a Mach Zehnder interferometer with an input beam splitter splitting the wave into two waveguiding arms and recombine them at another beam splitter. They can be fabricated in a way similar to the case of the optical waveguide by patterning the planar pattern of the two beam splitters/combiners and two waveguide arms using lithography and then etch down the silicon layer 1001 to form the required pattern on the chip. Optical phase shifting or modulation can make use of the refractive index change of silicon with temperature by fabricating a heating element to shift the refractive index of the waveguide in the device involved. Optical phase shifting or modulation can also make use of the refractive index change of silicon with electron or hole carrier density by fabricating an electrical structure as is known to those skilled in the art to inject electrons or holes into the silicon to shift the refractive index of the waveguide in the device involved. All-optical device can be realized utilizing the nonlinear optical absorption in silicon to change the phase or absorption of an optical beam by another optical beam as is known to those skilled in the art.
(80) Optical amplifier, laser, light-emitting device, and Optical detector are more specialized devices. In the case of silicon as layer 1001, optical amplifier and optical detector cannot be easily realized on layer 1001 and will require the Double-Optical-Layer EIC-PIC (DOL-EIC-PIC) technique to be described below. In the case of GaAs, GaN, InP, as layer 1001, these materials can provide optical gain and absorption and can be used to realize detectors or amplifiers in layer 1001.
(81) Fabrication:
(82) The Single-Optical-Layer EIC-PIC can be fabricated by using an electronic device fabrication compatible process (e.g. CMOS compatible process) to be described below to realize at least one coupling lens and at least one optical device on the chip.
(83) Direct Double-Optical-Layer EIC-PIC:
(84) For the purpose of illustration but not limitation, as shown in
(85) In an exemplary embodiment, the first substrate layer 1021 is the silicon substrate of a silicon-on-insulator (SOI) wafer, the first lower optical cladding layer 1011 is the silicon dioxide layer on top of the silicon substrate in the SOI wafer, the first optical core layer 1001 is the top silicon layer of the SOI wafer, the adjacent optical layer 10101 is made of III-V semiconductor materials (e.g. InGaAs/InGaAsP/InP or GaAs/AlGaAs material systems), and the first upper optical cladding layer 10201 is a silicon dioxide layer.
(86) Multiple Active Areas:
(87) In one embodiment, the adjacent optical layer 10101 is placed on multiple sub-areas on the chip as illustrated by the side and top views in
(88) Input or Output Lens:
(89) In an exemplary embodiment of the invention, the DDOL-EIC-PIC consists of at least one input or output optical-fiber coupling optics (i.e. it can have only one or more input coupling optics without any output coupling optics or only one or more output coupling optics without any input coupling optics. It can also have both input coupling optics and output coupling optics).
(90) Photonic or Nanophotonic Device:
(91) In an exemplary embodiment of the invention, as shown in
(92) In an embodiment of the invention, the first optical core layer 1001 is silicon which silicon cannot provide optical gain. Thus in the case of active optical devices such as optical amplifiers, lasers, or light-emitting devices, the adjacent optical layer 10101 can then be made of an optical material with gain such as compound semiconductor material. Thus the direct Double-Optical-Layer structure can be used to achieve a wide range of active integrated optical devices that involve optical gain.
(93) In an exemplary embodiment of device 10900 shown in
(94) Fabrication:
(95) The DDOL-EIC-PIC can be fabricated using an electronic device fabrication compatible process (e.g. CMOS compatible process) to be described below to realize at least one coupling lens and at least one optical device on the chip.
(96) Tapered Waveguide Structure:
(97)
(98) In an embodiment of the invention, as shown in
(99) In another embodiment of the invention, as shown in
(100) The height of layer 1001 is t.sub.EM, and the height of layer 10101 is tt.sub.Ph.
(101) In an exemplary embodiment, waveguide core 1001 is silicon with thickness t.sub.EM=300 nm, and top waveguide layer 10101 is InGaAs quantum well structure with three InGaAs quantum wells interspaced with InGaAsP barrier layers and adjacent optical layer 10101 has a refractive index around n.sub.AOL=3.4 and a total thickness tt.sub.Ph=150 nm. The taper dimensions are WW.sub.INI-UP=WW.sub.FIN-DN=100 nm, WW.sub.INI-DN=WW.sub.FIN-UP=400 nm, LL.sub.TAP-UP=LL.sub.TAP-DN=30 m, W.sub.UP=400 nm, W.sub.WG=400 nm.
(102) The various methods and systems described above for monolithic integration of active and/or passive photonic (and/or nanophotonic) devices in a way that is compatible with electronic integration will become clear with detailed descriptions of the methods and systems for the various exemplary EIC-PICs described below.
(103) Indirect Double-Optical-Layer EIC-PIC:
(104) For the purpose of illustration but not limitation, as shown in
(105) In another exemplary embodiment, at least one dielectric layer (first upper optical cladding layer) 11201 with thickness t.sub.UOTM and refractive index n.sub.UCL is further placed on top of layer 11101. Layer 11051 and layer 11101 with optional layer 11201 together is referred to as active-layer structure 11400. In one embodiment, layer 11051 is Silicon Nitride and layer 11101 is InGaAs quantum well structure with one or more InGaAs quantum wells inter-spaced with InGaAsP barrier material to provide optical gain for the wave coupling from layer 1001 to layer 11101. As is well known to those skilled in the art, there are other gain material structures one can use to achieve optical gain.
(106) In an exemplary embodiment, the first substrate 1021 layer is the silicon substrate of a silicon-on-insulator (SOI) wafer, the first lower optical cladding layer 1011 is the silicon dioxide layer on top of the silicon substrate in the SOI wafer, the first optical core layer 1001 is the top silicon layer of the SOI wafer, the first interspaced dielectric structure layer 11051 is a silicon nitride layer, the second optical core layer 11101 is made of III-V semiconductor materials (e.g. InGaAs/InGaAsP/InP or GaAs/AlGaAs material systems), and the first upper optical cladding layer 11201 is a silicon dioxide layer.
(107) Up/Down Coupling Tapered Waveguide in Indirect Double-Optical-Layer EIC-PIC:
(108) In an embodiment of the invention, the coupling between optical beam propagating in the second waveguide core layer 11101 and optical beam propagating in the first waveguide core layer 1001 is achieved using tapered waveguide vertical coupling structure as discussed below.
(109)
(110) In another embodiment, the coupling of optical beam from propagating in layer 11101 down to propagating in layer 1001 is achieved by using an optical vertical down-coupling tapered waveguide structure 11106. 11106 is a tapering waveguide section in layer 11101 tapering from an initial with W.sub.INI-DN to a final width W.sub.FIN-DN in a taper length L.sub.TAP-DN. The waveguide width of layer 1101 W.sub.WG-UP and waveguide width of layer 1001 W.sub.WG-DN can be equal or different.
(111) The height of layer 1001 is t.sub.EM, the height of layer 11101 is t.sub.Ph, and the height of layer 11105 is t.sub.DIE.
(112) Tapered Waveguide Coupling Structure Design:
(113) As is well known to those skilled in the art, if the first waveguide core layer 1001 has a propagating refractive index n.sub.E, lower than the propagating refractive index n.sub.Ph of the second waveguiding core layer 11101 with width W=W(Z) at location Z, where Z is spatial displacement along the direction of propagation as shown in
(114) By tapering the waveguide 11105 so that W(Z) changes with Z in layer 11101, one can change the propagating index n.sub.Ph so that it matches the value of n.sub.EI. At the region 11104 in
(115) As is well known to those skilled in the art, the exact shape of the taper is not so important as to have one substantial length of the tapering section where n.sub.Ph=n.sub.EI. Hence the taper can have arbitrary curvilinear shape the way it changes its horizontal width W(Z) as a function of spatial coordinate distance Z along the direction of wave propagation. As is well known to those skilled in the art, when this substantial length is equal about one coupling length, there will be efficient transfer of light energy between the top and the bottom waveguiding core layers 11101 and 1001 (typically >90%). Thus, the taper's width typically will go from a width approximately equal to or larger than half the wavelength in the top waveguiding core 11101 to a width smaller than half the wavelength in the top waveguiding core 11101. The tapering length L.sub.TAP-UP or L.sub.TAP-DN is chosen so that a substantial length of the taper achieves near-resonant waveguide energy coupling as described above.
(116) Once the optical beam energy is transferred from the lower waveguide core 1001 to the upper waveguide core 11101 and propagates into a straight waveguiding section with constant width, the energy will stay in 11101. This is because n.sub.Ph(W) will become higher than n.sub.EI (due to the increased width at the taper end connecting to the straight waveguiding section of 11101) so light energy will only guide in the upper waveguiding core 11101 until it meet with the down-coupling tapered waveguide.
(117) In one exemplary embodiment, lower waveguide core layer 1001 is silicon with thickness t.sub.EM=200 nm, upper waveguide core layer 11101 is InGaAs quantum well structure with five InGaAs quantum wells interspaced with InGaAsP barrier layers with the respective material bandgap energies and thicknesses shown in
(118)
(119) Multiple Active Areas:
(120) In one embodiment, the active-layer structure 11400 is placed on multiple sub-areas on the chip as illustrated by the top view and cross-sectional view in
(121) Input or Output Lens:
(122) In an exemplary embodiment of the invention, the IDDOL-EIC-PIC consists of at least one input or output optical-fiber coupling optics (i.e. it can have only one or more input coupling optics without any output coupling optics or only one or more output coupling optics without any input coupling optics. It can also have both input coupling optics and output coupling optics)
(123) Photonic/Nanophotonic Device:
(124) In an exemplary embodiment of the invention, the IDDOL-EIC-PIC consists of at least one photonic/nanophotonic device 11040, that can be an optical waveguide, an optical wavelength multiplexer, an optical wavelength demultiplxer, an optical grating, an optical beam splitter, a polarization beam splitter, an optical isolator, a polarization rotator, an optical interferometer, an optical modulator, an optical ring resonator, an optical disk resonator, an optical curved reflector, an optical mirror, an optical amplifier, a laser, a light-emitting device, an optical detector, a nonlinear-optical device, a photonic transistor, an optical harmonic frequency generator, an all-optical device, or other photonic/nanophotonic devices well known to those skilled in the art.
(125) In an embodiment, layer 1001 is silicon which silicon cannot provide optical gain. Thus in the case of active optical devices such as optical amplifier, laser, or light-emitting device, up-down vertical coupling structure is used to transport beam from layer 1001 to layer 11101. Layer 11101 can then be made of an optical material with gain such as compound semiconductor material. Thus the Indirect Double-Optical-Layer structure can be used to achieve a wide range of active integrated optical devices that involve optical gain.
(126) In an exemplary embodiment of device 11900 shown in
(127) Fabrication:
(128) The IDDOL-EIC-PIC can be fabricated using an electronic device fabrication compatible process (e.g. CMOS compatible process) to realize at least one coupling lens and at least one optical device on the chip to be described below.
(129) The various methods and systems described above for monolithic integration of active and/or passive photonic (and/or nanophotonic) devices in a way that is compatible with electronic integration will become clear with detailed descriptions of the methods and systems for the various exemplary EIC-PICs described below.
(130) Indirect Triple-Optical-Layer EIC-PIC:
(131) For the purpose of illustration but not limitation, as shown in
(132) Exemplary CMOS Process:
(133) In order to describe an exemplary electronic device and integrated circuit fabrication process, we will use the example of a CMOS process. Other electronic device fabrication methods have substantial similarity with a CMOS fabrication process. For example, the process always involve some steps of ion implantation and annealing at high temperature (1000 C.), and some steps of oxide growth at high temperature (1000 C.) Thus, the CMOS process is chosen for the purpose of illustration and not limitation. The purpose of illustration is to show how the photonic device integration process steps can be inserted to be made compatible with the high-temperature steps in an electronic device fabrication process. The details of the electronic device fabrication process are not essential and those skilled in the art will know how to adopt the exemplary embodiments to apply to different electronic device fabrication process accordingly.
(134) As is known to those skilled in the art, a CMOS process for fabricating electronic devices can deviate somewhat from the exemplary typical CMOS process shown here. Such deviations typically will not affect the insertion of the photonic device processes described that are to be done at certain main process points in the CMOS process or an electronic device and integrated circuit fabrication process. The main process points are well illustrated by the exemplary typical CMOS process shown.
(135) A typical CMOS process involves the following main steps:
(136) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(137) Step 2 ** Insertion point H1 (Covering of the photonic-device area with protective material layer).
(138) Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
(139) Step 4: Grow SiO.sub.2 on Si (1000 C.).
(140) Step 5: N-well ion implant
(141) Step 6: Wafer annealing (1000 C., 15-30 mins)
(142) Step 7: ** Insertion point H2
(143) Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask. (at around 1000 C.)
(144) Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.)
(145) Step 10: **Insertion point H3
(146) Step 11: Grow gate oxide (TOX)
(147) Step 12: Threshold adjust implant
(148) Step 13: Polysilicon deposition on TOX
(149) Step 14: **Insertion point H4
(150) Step 15: CVD oxide on Polysilicon
(151) Step 16: ** Insertion point H5
(152) Step 17: N+ source and drain implant
(153) Step 18: ** Insertion point H6
(154) Step 19: P+ source and drain implant
(155) Step 20: Annealing 900-1100 C.
(156) Step 21: Oxide spacer creation around gate
(157) Step 22: Silicidation (conductive metal silicide on gate/drain/source)
(158) Step 23: ** Insertion point L1
(159) Step 24: ** Insertion point L2
(160) Step 25: ** Insertion point L3
(161) Step 26: ** Insertion point L4
(162) Step 27: ** Insertion point L5
(163) Step 28: Thick 500 nm oxide for metal masking
(164) Step 29: Drain, gate, source and photonic device metal contact via multilevel metals
(165) More specifically, the steps are described in detail below. These steps include:
(166) Step 1: Start with an SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(167) Step 2: ** Insertion point H1. Will be used for depositing protective materials to cover of the photonic-device area.
(168) Step 3: An optional step involving implanting P dopants on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
(169) Step 4: A thin film of SiO.sub.2 dielectric layer is placed on a silicon (Si) wafer by heat treatment process involving a temperature of 1000 C. (Initial Oxidation). The wafer is patterned and the oxide is etched to form oxide mask with open window for N-Well oxidation.
Step 5: N-type dopants are implanted on the silicon wafer layer to form N-wells (N-Well implant).
Step 6: A process of Wafer Annealing is carried out to activate the implanted dopants. The process requires heat treatment involving a temperature of 1000 C. (N-Well drive-in). The oxide mask is then etched away. The structure formed after wafer annealing process is considered as a first substrate layer.
Step 7: ** Insertion point H2 for fabrication of photonic device.
Step 8: A first dielectric layer of SiO.sub.2 and a second dielectric layer of Si.sub.3N.sub.4 are deposited (at around 1000 C.) over the structure to form a mask for the growth of field oxide.
Step 9: A process of channel-stop ion implantation is carried out. The process involves creating a photo resist layer over the SiO.sub.2 and Si.sub.3N.sub.4 layers that are then etched to form a mask. After performing channel stop field implant, the photo resist layer is removed and Si.sub.3N.sub.4 mask is also removed. and a thick oxide layer is grown (field oxidation) on the exposed portions of the silicon at 1000 C., thus producing thick Field Oxide (FOX). FOX grows in the area where Si.sub.3N.sub.4 layer is absent.
Step 10: ** Insertion point H3 for fabrication of photonic device.
Step 11: Gate Oxide is grown on the areas where FOX is absent (gate oxidation). The growth of Gate Oxide acts as a gate dielectric (TOX).
Step 12: The process of growing TOX is followed by a threshold-adjust implantation to adjust native threshold voltage (threshold implant). The native threshold voltage is adjusted by implanting a thin layer of dopants near the surface of the structure.
Step 13: A polysilicon layer is deposited as a deposition layer on top of the gate dielectric (TOX) (polysilicon deposition).
Step 14: **Insertion point H4.
Step 15: A process of Chemical Vapor Deposition (CVD) is performed to place a CVD oxide layer on the polysilicon deposition layer. In CVD, the structure formed is exposed to one or more volatile precursors, which react and/or decompose on the surface to produce a desired deposit.
Step 16: **Insertion point H5.
Step 17: The patterning process is followed by N+ ion implantation to form N+ source region and N+ drain region (N+ Source/Drain).
Step 18: ** Insertion point H6.
Step 19: P+ dopants are implanted over source and drain terminals to form P+ source region and P+ drain region (P+ Source/Drain).
Step 20: A process of annealing is performed at 900-1100 C. to activate the implanted dopants.
Step 21: An oxide space is created around gate terminal, such that the deposition of the silicide during silicidation process will not short the gate region with source region or drain region.
Step 22: A process of silicidation is performed. The silicidation process involves depositing conductive material on gate region or source region or drain region through Chemical Vapor Deposition (CVD) process. The structure formed after the process of silicidation is considered as an integrated structure.
Step 23: ** Insertion point L1.
Step 24: ** Insertion point L2.
Step 25: ** Insertion point L3.
Step 26: ** Insertion point L4.
Step 27: ** Insertion point L5.
Step 28: Metal masking process is performed after the etching of superlens. A 500 nm thick oxide layer is used for the process of metal masking.
Step 29: The last step involves creation of metal contacts for drain region, source region, gate region and metal contacts for photonic devices using multi-level metals.
(170) The steps labeled as H1 to H6 are insertion points for the fabrication steps of the photonic devices that are compatible with high temperature of 1000 C. used in the electronic device or integrated circuit fabrication process typified by the CMOS process illustrated. The steps labeled as L1 to L5 are insertion points for the fabrication steps of the photonic devices that are not compatible with high temperature of 1000 C. used in the electronic fabrication or CMOS process and have to be done after all those high-temperature process steps used in the electronic fabrication or CMOS process.
(171) We will illustrate more specifically how the fabrication steps of EIC-PIC are compatible with the abovementioned CMOS process. To do so, we will illustrate them with the fabrication of EIC-PIC on SOI (Silicon on Insulator) wafer. It is done for the purpose of illustration and not limitation. We will illustrate the process steps for the cases of Single-Optical-Layer EIC-PIC (SOL-EIC-PIC), Direct Double-Optical-Layer EIC-PIC (DDOL-EIC-PIC), and Indirect Double-Optical-Layer EIC-PIC (IDDOL-EIC-PIC), respectively.
(172) Exemplary CMOS Compatible Fabrication Steps for SOL-EIC-PIC Chip with Passive Photonic Device on Silicon:
(173) The top view of an exemplary Single-Optical-Layer EIC-PIC (SOL-EIC-PIC) chip fabricated on SOI wafer is shown in
(174) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(175) Step 2 ** Insertion point H1 is used. It involves depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material must have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: ** Insertion point H2 (unused)
Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.)
Step 10: **Insertion point H3 (unused)
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: **Insertion point H4 (unused)
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5 (unused)
Step 17: N+ source and drain implant
Step 18: ** Insertion point H6 (unused)
Step 19: P+ source and drain implant
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1 is used: The photonic-area protective material is removed. Lithographically pattern the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and device pattern using typical dry or wet etching technique known to those skilled in the art.
Step 24: ** Insertion point L2 is not used.
Step 25: ** Insertion point L3 is not used.
Step 26: ** Insertion point L4 is used for depositing the multi-layer materials for the superlens structure.
Step 27: ** Insertion point L5 is used for etching of the superlens, which is performed after the superlens deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(176) This is not the only possible process, as an alternative, the passive photonic device and waveguides can also be fabricated during H1 as it will be compatible with the subsequent high-temperature process steps as shown by the steps below.
(177) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(178) Step 2 ** Insertion point H1 is used. First, it involves lithographically pattern the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and device pattern using typical dry or wet etching technique known to those skilled in the art. It ends with depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices have been or will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material must have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: ** Insertion point H2 (unused)
Step 8: Grow SiO.sub.2 and Si3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.).
Step 10: **Insertion point H3 (unused)
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: **Insertion point H4 (unused)
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5 (unused)
Step 17: N+ source and drain implant
Step 18: ** Insertion point H6 (unused)
Step 19: P+ source and drain implant
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1 is not used.
Step 24: ** Insertion point L2 is not used.
Step 25: ** Insertion point L3 is not used.
Step 26: ** Insertion point L4 is used. The photonic-area protective material is removed. Multi-layer materials for the superlens structure is deposited.
Step 27: ** Insertion point L5 is used for etching of the superlens is performed after deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(179) Exemplary CMOS Compatible Fabrication Steps for SOL-EIC-PIC Chip with Passive and Active Photonic Device on Silicon:
(180) The top view of an exemplary Single-Optical-Layer EIC-PIC (SOL-EIC-PIC) chip with passive and active photonic devices fabricated on SOI wafer is shown in
(181) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(182) Step 2 ** Insertion point H1 is used. First, it involves lithographically patterning the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and ring or disk modulator device pattern using typical dry or wet etching technique known to those skilled in the art. It ends with depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices have been or will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material much have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: **Insertion point H2 (not used).
Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.).
Step 10: **Insertion point H3. Open optical modulator top from the protective materials deposited in Step 2, P adjustment implant into modulator
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: **Insertion point H4. A polysilicon or n-Transparent Conducting Oxide (n-TCO) or a combination of n-TCO and polysilicon layer is deposited as a deposition layer on top of the gate dielectric (TOX) (unused). Certain part of the deposition layer is subsequently removed by etching, in a way such that only portions of the deposition layer over the edges of the modulator are maintained and rest of the deposition layer is removed.
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5. A pattern for N+ implant is created over the silicon based modulator. As an exemplary process, the process of patterning involves a process of depositing a negative photo resist and a process of photolithography. After photolithography all portions to receive n-F implants are exposed. N+ implant is then performed to achieve N+ dopant for the modulator. The area is covered by resist again to block it from the next implant step for transistor device.
Step 17: N+ source and drain implant for transistors.
Step 18: ** Insertion point H6. The N+ source region and the N+ drain region are protected by placing a resist cover over the implanted N+ source region and N+ drain region and then opening the silicon based modulator for implanting P+ dopants. P+ implant is then performed to achieve P+ dopant for the modulator. The area is covered by resist again to block it from the next implant step for transistor device.
Step 19: P+ source and drain implant for transistors.
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1 is not used.
Step 24: ** Insertion point L2 is not used.
Step 25: ** Insertion point L3 is not used.
Step 26: ** Insertion point L4 is used. The photonic-area protective material is removed. Multi-layer materials for the superlens structure is deposited.
Step 27: ** Insertion point L5 is used for etching of the superlens is performed after deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(183) Exemplary CMOS Compatible Fabrication Steps for DDOL-EIC-PIC Chip with Passive Photonic Device on Silicon and Active Photonic Device on Compound Semiconductor:
(184) The top view of an exemplary Direct Double-Optical-Layer EIC-PIC (DDOL-EIC-PIC) chip with passive and active photonic devices fabricated on SOI wafer is shown in
(185) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(186) Step 2 ** Insertion point H1 is used. First, it involves lithographically pattern the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and device pattern using typical dry or wet etching technique known to those skilled in the art. It ends with depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices have been or will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material much have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: ** Insertion point H2 (unused)
Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.).
Step 10: **Insertion point H3 (unused)
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: **Insertion point H4 (unused)
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5 (unused)
Step 17: N+ source and drain implant
Step 18: ** Insertion point H6 (unused)
Step 19: P+ source and drain implant
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1: The photonic-area protective material is removed. Compound semiconductor wafer bonding with the Silicon surface of the SOI substrate is carried out between the compound semiconductor surface and the silicon surface of the SOI substrate. The process of wafer bonding is further explained in detail.
Step 24: ** Insertion point L2: Quantum Well Intermixing (QWI) process is followed by the wafer-bonding process. The QWI process involves implanting ions into a number of quantum wells using thin layer of oxides as an implant mask. The QWI process can also be performed before the wafer bonding process, depending on situation.
Step 25: ** Insertion point L3: After QWI, during a process of compound semiconductor device (e.g. InP/InGaAsP device) fabrication, a protective cover material is applied over the compound semiconductor device to protect the compound semiconductor device from the superlens processes.
Step 26: ** Insertion point L4 is used for depositing the multi-layer materials for the superlens structure.
Step 27: ** Insertion point L5 is used for etching of the superlens is performed after deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(187) Exemplary CMOS Compatible Fabrication Steps for IDDOL-EIC-PIC Chip with Passive Photonic Device on Silicon and Active Photonic Device on Compound Semiconductor:
(188) The top view of an exemplary Indirect Double-Optical-Layer EIC-PIC (IDDOL-EIC-PIC) chip with passive and active photonic devices fabricated on SOI wafer is shown in
(189) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(190) Step 2 ** Insertion point H1 is used. First, it involves lithographically pattern the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and device pattern using typical dry or wet etching technique known to those skilled in the art. It ends with depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices have been or will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material much have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: ** Insertion point H2 (unused)
Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.).
Step 10: **Insertion point H3 (unused)
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: **Insertion point H4 (unused)
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5 (unused)
Step 17: N+ source and drain implant
Step 18: ** Insertion point H6 (unused)
Step 19: P+ source and drain implant
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1: The photonic-area protective material is removed. Compound semiconductor wafer bonding with the Silicon surface of the SOI substrate is carried out between a Si.sub.3N.sub.4 layer deposited on the integrated structure and an Si.sub.3N.sub.4 layer deposited on the silicon surface of the SOI substrate wafer. The process of wafer bonding is separately explained in detail.
Step 24: ** Insertion point L2: Quantum Well Intermixing (QWI) process is followed by the wafer-bonding process. The QWI process involves implanting ions into a number of quantum wells using thin layer of oxides as an implant mask. The QWI process can also be performed before the wafer bonding process, depending on situation.
Step 25: ** Insertion point L3: After QWI, during a process of an compound semiconductor device fabrication, a protective cover material is applied over the compound semiconductor device to protect the compound semiconductor device from superlens processes.
Step 26: ** Insertion point L4 is used for depositing the multi-layer materials for the superlens structure.
Step 27: ** Insertion point L5 is used for etching of the superlens is performed after deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(191) Exemplary CMOS Compatible Fabrication Steps for IDDOL-EIC-PIC Chip with Passive/Active Photonic Device on Silicon and Passive/Active Photonic Device on Compound Semiconductor:
(192) The top view of an exemplary Indirect Double-Optical-Layer EIC-PIC (IDDOL-EIC-PIC) chip with passive and active photonic devices fabricated on SOI wafer is shown in
(193) Step 1: Start with SOI wafer with undoped or lightly doped top silicon layer for the SOI wafer.
(194) Step 2 ** Insertion point H1 is used. First, it involves lithographically pattern the silicon waveguides and passive silicon devices using typical lithography processes known to those skilled in the art such as photolithography or ebeam lithography. The silicon is then etched down to form the waveguide and device pattern using typical dry or wet etching technique known to those skilled in the art. It ends with depositing thick layer of photonic-area protective material such as a few micron-thick silicon oxide or other materials (e.g. In.sub.2O.sub.3 etc) to cover the silicon area that photonic devices have been or will be fabricated. The actual material and thickness to be used is depending on the later ion-implantation process. The material much have a thickness thick enough to stop most of the high-energy implanting ions to reach into the silicon layer in this area. This is to prevent crystal damage or optical loss caused by the potential implanting ions.
Step 3: P implant on most of the substrate area except the area of the SOI wafer intended to form low-loss silicon optical waveguide.
Step 4: Grow SiO.sub.2 on Si (1000 C.)
Step 5: N-well ion implant
Step 6: Wafer annealing (1000 C., 15-30 mins)
Step 7: Insertion point H2 is used. A pattern for creating a nano-waveguide and a silicon-based ring or disk modulator is created on the first substrate layer. The nano waveguide is typically 0.3 m thick.
Step 8: Grow SiO.sub.2 and Si.sub.3N.sub.4 for field oxide mask (at around 1000 C.).
Step 9: Channel-stop implant, grow field oxide (FOX) (at around 1000 C.).
Step 10: Insertion point H3 is used. Portions of SiO.sub.2 and Si.sub.3N.sub.4 dielectric layers placed over the modulator are removed to open the modulator for p-type adjustment of dopants.
Step 11: Grow gate oxide (TOX)
Step 12: Threshold adjust implant
Step 13: Polysilicon deposition on TOX
Step 14: Insertion point H4 is used. A polysilicon or n-Transparent Conducting Oxide (n-TCO) or a combination of n-TCO and polysilicon layer is deposited as a deposition layer on top of the gate dielectric (TOX).
Step 15: CVD oxide on Polysilicon
Step 16: ** Insertion point H5 (unused)
Step 17: N+ source and drain implant
Step 18: ** Insertion point H6 (unused)
Step 19: P+ source and drain implant
Step 20: Annealing 900-1100 C.
Step 21: Oxide spacer creation around gate
Step 22: Silicidation (conductive metal silicide on gate/drain/source)
Step 23: ** Insertion point L1: The photonic-area protective material is removed. Compound semiconductor wafer bonding with the Silicon surface of the SOI substrate is carried out between an Si.sub.3N.sub.4 layer deposited on the integrated structure and an Si.sub.3N.sub.4 layer deposited on the silicon surface of the SOI substrate wafer. The process of wafer bonding is separately explained in detail.
Step 24: ** Insertion point L2: Quantum Well Intermixing (QWI) process is followed by the wafer-bonding process. The QWI process involves implanting ions into a number of quantum wells using thin layer of oxides as an implant mask. The QWI process can also be performed before the wafer bonding process, depending on situation.
Step 25: ** Insertion point L3: After QWI, during a process of an compound semiconductor device fabrication, a protective cover material is applied over the compound semiconductor device (e.g. InP/InGaAsP device) to protect the compound semiconductor device from superlens processes.
Step 26: ** Insertion point L4 is used for depositing the multi-layer materials for the superlens structure.
Step 27: ** Insertion point L5 is used for etching of the superlens is performed after deposition. During the process of deposition and etching of superlens, a protective cover material is maintained over the rest of the structure.
Step 28: Thick 500 nm oxide for metal masking.
Step 29: Drain, gate, source and photonic device metal contact via multilevel metals.
Step 30: Mounting an optical fiber on the substrate to facilitate a coupling of light beam energy between the optical fiber and an optical waveguide on the substrate through a coupling optics (e.g. the superlens);
(195) The exemplary embodiments above are for the purpose of illustration and not limitation. For example, while illustrated for compound semiconductor (e.g. InP/InGaAsP system,) based devices, they can be modified for other materials for fabricating active photonic devices. Also, it is well known to those skilled in the art that there are more than one ways for the CMOS process steps. The main idea here is the division of the CMOS process steps into two categories: the steps before the last high-temperature process step (high-temperature steps are steps involving temperature above 400 C.), and the steps after the last high-temperature process step. The insertion points for the fabrication of the photonic devices can also be categorized into two categories: the high-temperature insertion points before the last high-temperature CMOS process step (high-temperature steps are steps involving temperature above 400 C.), and the low-temperature insertion points after the last high-temperature process step. The main idea is that the wafer bonding or local-area wafer bonding of the photonic layer structure shall be done after the last high-temperature CMOS process step at the low-temperature insertion points. This is because the photonic layer structure usually contains compound semiconductor materials that can decompose under temperature higher than approximately 500 C. for a relatively long time period of longer than a few minutes. The V-groove or trench that holds the superlens can be prefabricated on the substrate such as the silicon substrate of the SOI wafer used for the fabrication. It can also be fabricated during one of the processing steps such as while the silicon waveguides or devices are fabricated or just before the optical fiber mounting step. The fabrication of the V-groove usually involves the use of KOH etchant to etch long the crystallographic planes to a predefined depth with high accuracy as is well known to those skilled in the art or by dry etching using ICP (inductively coupled plasma) or RIE (reactive ion etching) to etch down the silicon using metal mask, a process known to those skilled in the art, to form a pre-defined deep trench. The optical fiber can be mounted and held using UV epoxy or solder.
(196) Furthermore, while we assume the photonic layer structures involve certain semiconductor materials, it can be various other types of materials for fabricating active photonic devices including but not limited to: Indium Phosphide (InP), Gallium Arsenide (GaAs), Indium Gallium Arsenide (InGaAs), Indium Gallium Arsenide Phosphide (InGaAsP), Indium Aluminum Gallium Arsenide (InAlGaAs), Aluminum Arsenide (AlAs), Aluminum Gallium Arsenide (AlGaAs), Indium Gallium Aluminum Phosphide (InGaAlP), Indium Gallium Phosphide (InGaP), Gallium Nitride (GaN), Aluminum Nitride (AlN), Gallium Aluminum Nitride (GaAlN), Gallium Phosphide (GaP), Aluminum Phosphide (AlP), Aluminum Antimonide (AlSb), Gallium Antimonide (GaSb), Zinc selenide (ZnSe), Zinc Sulphide (ZnS), Cadmium Sulphide (CdS), Silicon Carbide (SiC), Silicon Germanium (SiGe), Indium Gallium Antimonide (InGaSb), or Indium Antimonide (InSb), Germanium (Ge), Silicon-Germanium (SiGe) with various compositions of the constituent materials and the combinations thereof.
(197) A number of materials may also be used as the dielectric layer(s). They include but are not limited to oxides, nitrides, carbides, such as tantalum pentoxide (Ta.sub.2O.sub.5), zirconium oxide (ZrO.sub.2), niobium pentoxide (Nb.sub.2O.sub.5), hafnium oxide (HfO.sub.2), zinc oxide (ZnO), germanium oxide (GeO.sub.2), lead oxide (PbO), yttrium oxide (Y.sub.2O.sub.3), aluminum oxide (Al.sub.2O.sub.3), silicon carbide (SiC), titanium carbide (TiC), titanium nitride (TiN), chromium nitride (CrN), carbon nitride (CN), carbon boride (CB), aluminum nitride (AlN), zinc selenide (ZnSe), barium fluoride (BaF.sub.2), magnesium fluoride (MgF.sub.2), Diamond like Carbon (DLC). In another embodiment, the dielectric is a polymer or organic material such as Benzocyclobutene (BCB), cyclized transparent optical polymer (CYTOP), and a polymer of imide monomers (Polyimide).
(198) Thus, the process described here is for illustrative purpose only and it does not limit the scope of the invention in any way. Different combinations of these steps and other similar steps are contemplated and are within the scope of the present invention.
(199) The steps discussed are a combination of standard CMOS process and additional CMOS compatible steps that aid in the fabrication of both Si and compound semiconductor (e.g. InP/InGaAsP) based photonic devices. Steps 2, 24, 25, 26, and 27 are the additional CMOS compatible steps according to various embodiments of the invention. Steps 16, 18 listed are potentially reserved to insert additional steps for the fabrication of photonic devices are not used in this example.
(200) The silicon based photonic devices are fabricated on the Silicon-On-Insulator (SOI) wafer together with the CMOS components using CMOS wafer-level process. The steps 1 to 20 are high temperature processes. Step 21, 22, 28, 29 are the beginning of the CMOS metal-interconnect process. The compound semiconductor (e.g. III-V InP/InGaAsP) wafer bonding begins either after step 20 or at step 23. The figures illustrated are assumed to have compound semiconductor (e.g. III-V InP/InGaAsP) wafer bonding after step 23. This is to ensure that the properties of compound semiconductor (e.g. III-V InP/InGaAsP) wafer do not change because of high temperature involved in steps 1 to 20. Step 5 is illustrated in
(201) Exemplary CMOS Compatible Fabrication Steps for EIC-PIC Chip with Triple or more Optical Layers:
(202) Those skilled in the art will know how to generalize from two optical layers to three or more optical layers by inserting more wafer bonding steps to bond the third top layer to the second middle layer in a way similar to the bonding of the second optical layer to the first optical layer, and the fabrication of more tapered waveguide couplers for vertical coupling of light between two optical layers.
(203) Wafer Bonding and Local-Area Wafer Bonding:
(204) Part of the process steps mentioned above involve local-area wafer bonding resulting in thin-film transfer of active optical materials on a substrate. Exemplary embodiments for wafer boding and local-area wafer bonding processes are illustrated below.
(205) Wafer Structure to Transfer:
(206) For the purpose of illustration, we assume the compound semiconductor layer to be transferred involves InP/InGaAsP epitaxial layer structures often used for the fabrication of active photonic devices, such as optical amplifiers and lasers as are well known skilled in the art. The exemplary embodiments for a compound semiconductor (e.g. InP/InGaAsP) Wafer based Bonding and Etch Back are as follows: An epitaxial layer structure shown in
(207) N-Side Down for Near-Epi-Wafer-Smooth Bonding Surface:
(208) After the epitaxial growth, an InP handling wafer (HW) is bonded to the epi-wafer through a Benzocyclobutene (BCB) polymer as shown by Step 2 in
(209) Local-Area Wafer Bonding to SOI:
(210) As described earlier, the thin silicon nitride layer is deposited on top of the silicon surface of the SOI wafer designated for compound semiconductor devices, as shown by Step 5 in
(211) The exemplary embodiments above are for the purpose of illustration and not limitation. For example, it is well known to those skilled in the art that there are more than one ways to achieve wafer bonding. The main steps, however, is the involvement of pressure and temperature treatment to press two wafers together and achieve atomic-level bonding. Subsequently, a selective etching step is performed to result in the transfer of the active optical thin film (e.g. the compound semiconductor thin film) on a substrate (e.g the SOI substrate). The selective etching can be wet chemical etching or dry selective etching using machines such as Reactive-Ion Etcher (RIE) or Inductively-Coupled Plasma Etcher (ICP), and the use of appropriate etching gasses.
(212) Also for example, the main steps of the local-area wafer bonding illustrated involve a handling wafer with pre-etched areas to match the electronic areas so that many subareas of the active materials (e.g. the compound semiconductor active layer) can be transferred to the substrate wafer (e.g. the SOI wafer) in a single effort without having the bonding wafer touching the electronic area. Another step include covering the photonic areas with protective materials during the high-temperature electronic fabrication or CMOS process, and then uncover the protected photonic areas by removing the photonic-area protective materials so that a fresh silicon surface will be uncovered for performing the wafer bonding. For such protective process steps, see for example Step 2 and Step 23 of the two sections describing CMOS Compatible Fabrication Steps for Exemplary IDDOL-EIC-PIC Chip. As mention there, a good protective material to use is In.sub.2O.sub.3 material that is resistive to most fluorine chemistries that etches SiO.sub.2 and are commonly used in electronic device fabrication processes, but can be removed by Chlorine chemistries sledome used in electronic device fabrication processes. Other protective materials include a combination of In.sub.2O.sub.3 and metallic films or certain polymer materials. Other steps may be specific to the type of materials being bonded. Thus, the process described here is for illustrative purpose only and it does not limit the scope of the invention in any way. Different combinations of these steps and other similar steps are contemplated and are within the scope of the present invention.
(213) Exemplary Embodiments of EIC-PIC with Multiple Photonic Device Components and Photonic Sub-System on Chip:
(214) Above, we have illustrated the essential exemplary structure and fabrication processes for various EIC-PICs. Below, we will provide more exemplary EIC-PIC based integrated circuits with multiple device components, resulting in photonic subsystem on chip with active and passive photonic devices integrated with electronic circuits. The simplest case of an EIC-PIC has:
(215) (1) A fiber input coupler or a fiber output coupler or both;
(216) (2) At least one integrated photonic or nanophotonic device connected to the fiber input coupler or the fiber output coupler with an optical waveguide; and
(217) (3) An optional electronic device or electronic integrated circuit on the chip.
(218) Electronic-Integration Compatible Photonic/NanoPhotonic Integrated Circuit (EIC-PIC) is a photonic integrated circuit or a nanophotonic integrated circuit fabricated to be compatible with an electronic integrated circuit. However, this provides a benefit of simple fabrication processes, thus providing compatibility with electronic integrated circuit manufacturing line. The presence of an electronic circuit is optional. Applications of EIC-PIC include, but are not limited to, optical receivers used in fiber optic communication.
(219) The performance is enhanced by integrating photonic devices and optical fiber connections with electronic microprocessor chip. This results in an increase in the speed of data input and output from the electronic microprocessor chip. The applications are capable of performing very fast computation on chip as compared to the current microprocessor chips. However, the current microprocessors are capable of performing fast computation at a data rate of 3-10 Gigabits per second (1 Gigabit/sec=10.sup.9 bits/sec), with the difference that the microprocessor chips suffer from the slow electrical connections to devices outside the chip through an electrical wire. The use of electrical transmission line may increase electrical data transmission speed to tens of Gigabit/sec, however the cost of electrical transmission line is high.
(220) Further, enhancement is achieved by combining high data transmission rate of photonic system with the high computation speed of electronic integrated circuit. The data transmission speed of photonic system is enhanced by using multiple optical wavelengths (40-100 different wavelengths), each wavelength being capable of carrying high data rate of 2.5-40 Gigabits/sec. Such a photonic data transmission system is commonly referred to as Dense Wavelength Division Multiplexing (DWDM) system. For example, with 40 wavelengths, each operating at 40 Gigabit/sec, a total aggregated data rate of 1.6 Terabits/sec (1 Terabit/sec=1012 bits/sec) is achieved. To carry this huge data into and out of the electronic microprocessor chip, only single optical fiber is sufficient. Considering the advantage provided by the optical fiber, the electrical wires are replaced by an optical DWDM data transmission system on an electronic chip to enable integrated ultra-high-speed data transmission in-to or out-of the electronic circuits using optical fibers. The photonic data transmission system integrated on EIC-PIC is therefore referred to as integrated chip-level optical network (ICON). An EIC-PIC with ICON is referred to as EIC-PIC+ICON.
(221) The ICON on EIC-PIC results in increasing the input and output data rates from and to an electronic microprocessor, thus enabling ultra-high-speed data transmission between two or more electronic integrated circuits. Such a photonic data transmission link is referred to as an inter-chip optical interconnect. ICON also helps to increase data transmission speed within the electronic chip. In that case, photonic waveguides and devices on the chip are used to transmit information from one part of the electronic circuit to another part of the electronic circuit on the same substrate. Such a photonic data transmission link is called an intra-chip optical interconnect.
(222) The two exemplary EIC-PIC devices are illustrated below in accordance with the preferred embodiments of the present invention. It will be apparent to a person skilled in the art that there are many different possible configurations for an EIC-PIC device in terms of the photonic integrated circuit geometry, the electronic integrated circuit geometry, the specific photonic and electronic devices employed in the circuits, and the fabrication method thereof. These exemplary devices are therefore shown for the purpose of illustration and they do not limit the scope of the present invention in any way.
(223) First Exemplary EIC-PIC-ICON Device:
(224)
(225) The first part is an Input Part 102. The Input Part 102 includes a fiber input coupler 108, and a Photonic Integrated Circuit (PIC). The fiber input coupler 108 is integrated on a chip. An optical beam, with N wavelength channels, propagating in the input optical fiber, is connected to the coupler 108. Each of the N wavelength channels carries encoded signals as intensity modulation of the wavelength channel.
(226) A Photonic Integrated Circuit (PIC) is included to perform optical DWDM data demultiplexing and optical to electrical data conversion. The PIC includes a coupling optic 110 to couple light into a waveguide 112a that guides the optical beam to a wavelength de-multiplexer 114. The wavelength de-multiplexer 114 splits the light in N wavelength channels as N outputs from the demultiplexer 114. Each of the N outputs of the demultiplexer 114 is connected to an integrated optical amplifier 116 through an optical waveguide 112b to amplify the optical power.
(227) The second part is the interior part 104. The Interior Part 104 consisting of an integrated electronic circuit area includes:
(228) (1) N optical-to-Electrical Data Conversion Lines using photodetectors are used to convert optical power to electrical signals. The optical amplifier 116 output from the Input Part 102 is connected to an integrated photodetector 118 through another optical waveguide 112c. The photodetector 118 converts the optical signal in that wavelength channel to a stream of electrical signals. Therefore, there are N photodetectors 118 to produce N number of electrical signals.
(229) (2) N electrical data lines 120a are used to transfer the electrical data from the N photodetectors 118 to an electronic circuit 122.
(230) (3) The electronic circuit 122 receives the electrical data from the N electrical data lines 120a to process the electrical data streams that are received.
(231) (4) There are M electrical data output lines 120b from the electronic circuit. The M electrical data streams are converted to M optical data streams by modulating the optical outputs from M semiconductor lasers 124.
(232) The third part is a Fiber Output Port 106 incorporating Electrical-to-Optical Data Conversion, DWDM Wavelength Multiplexing, and Optical Fiber Output. Further, the Optical Fiber Output includes a Photonic Integrated Circuit (PIC) and a wavelength multiplexer.
(233) The Photonic Integrated Circuit (PIC) is configured to perform electrical to optical data conversion and DWDM optical data multiplexing. This PIC comprises M integrated semiconductor lasers 124 with intensity directly modulated by the M electrical data output lines 120 from the electronic circuit 122. The M lasers 124 have different emission wavelengths. Each wavelength is tuned to one of the standard DWDM wavelength channels. The intensity modulation of each of the M lasers 124 can be achieved by modulating the injection current to the M lasers 124 for data rates up to about 10 Gigabit/sec. For higher data rates (>10 Gigabit.Math.sec), an integrated modulator such as an electro-absorption modulator or an electro-optic modulator is used by connecting the modulator to the output of the M lasers 124 with an integrated waveguide. The outputs of the M modulated lasers 124 have different DWDM wavelengths and are sent to an integrated wavelength multiplexer 126 with M inputs through a waveguide 112d. The wavelength multiplexer 126 combines the M different wavelength channels to produce a single output. The output of the wavelength multiplexer is a single optical beam.
(234) The output beam from the wavelength multiplexer 126 is transmitted to an integrated fiber coupler 128 through an integrated waveguide 112e. The optical coupler 128 couples the output beam from the output waveguide to an output optical fiber 130.
(235) The integrated photonic devices in the EIC-PIC include, but are not limited to, optical input coupler, optical wavelength demultiplexer, optical amplifier, photodetector, optical waveguide semiconductor laser, optical modulator, optical wavelength multiplexer, and optical output coupler. It will be apparent to a person skilled in the art that the devices may be regular photonic devices or nano-photonic devices capable of performing the same function, which further provide much smaller device size.
(236) Second Exemplary EIC-PC-ICON Device:
(237)
(238) The Input Port 400 with multiple Optical Fiber Input Parts 100Ca-100Ch and Input Parts 142 and 144 is further illustrated in conjunction with
(239) The Interior Part 200 receives optical beams from the input port 400 and performs various functions, including Optical-to-Electrical Data Conversion and data processing with electronic integrated circuits.
(240) The Output Port 402 with multiple Optical Fiber Output Parts 300a-300h and Output Parts 312 and 316 performs various functions including the functions of Electrical-to-Optical Data Conversion, DWDM Wavelength Multiplexing, and providing optical Output.
(241) The Input Port:
(242) The Input Port 400 of the EIC-PIC+ICON device 100B (hereinafter referred to as electronic-photonic device or EP device) has multiple optical fiber input parts 100a-100h and other input parts 142 and 144. An optical fiber input part 100C of the multiple optical fiber input parts 100a-100h is as illustrated in conjunction with
(243) The optical fiber input part 100C includes a fiber 132, a coupling optics 134, a wavelength multiplexer 136 and plurality of active device 138. The light beam travels through the fiber 132 of each of the optical fiber input part 100C. The light beam from the fiber 132 is coupled to coupling optics 134. The coupling optics 134 couples light beam from the fiber 132 into the wavelength multiplexer 136 through a passive waveguide layer 160 (not shown in the figure). The passive waveguiding layer 160 has been further described in detail in conjunction with
(244) In accordance with an embodiment of the present invention, the superlens is deposited on a substrate. The structure of superlens 134 comprises multiple layers of two alternating materials with different refractive indices and having a thickness substantially less than its optical wavelength of operation. The thickness ratio of the two layers defines an effective index of the superlens 134 around that layer. The refractive indices of the superlens 134 can thus be varied in a direction vertical to the substrate to fit any refractive index profile vertically. An illustrative design of the refractive index profile for the superlens 134 is described below.
(245) Hereinafter, the superlens coupling optics 134 will be interchangeably referred to as the SLCO-SmfSiw 134. The SLCO-SmfSiw 134 could also be referred to as the integrated ultra-compact superlens coupling optics for passive-aligned multiport coupling between waveguides and single-mode fibers 132. In one embodiment of the invention, the SLCO-SmfSiw 134 couples light into 0.3 m thick passive silicon wave guiding layer 160 located on an SOI substrate. In another embodiment of the invention, the SLCO-SmfSiw 134 can be 10 m thick. In some cases, the output beam power of the SLCO-SmfSiw 134 is 100-160 W and practical coupling loss is 1-3 dB.
(246) In one embodiment of the invention, SLCO-SmfSiw 134 is used for coupling short wavelengths of light in an optical fiber communication wavelength range of 1.5 m. The advantages of using the SLCO-SmfSiw 134 will be apparent to a person skilled in the art due to its integration capability and its ultra compact size. In another embodiment of the invention, thin-film photonics requires coupling of light into a thin film having a thickness of 0.2-0.3 m for operation at an optical wavelength of 1.5 m. This can be achieved using the SLCO-SmfSiw 134, wherein the integrated superlens SLCO-SmfSiw 134 has one half of the parabolic profile having a refractive index layer at the surface of a wafer. In some cases, the superlens SLCO-SmfSiw 134 has a varying refractive index in the range of n.sub.Si3.6 for silicon (Si) to n.sub.SiO21.45 for silicon dioxide (SiO.sub.2). In some other cases, the refractive index is graded down to a value of around 1.5 at the top of the superlens SLCO-SmfSiw 134, whereas the refractive index at the bottom has a value nearly equal to that of the waveguide. The grading down of the refractive index results in a rapid vertical expansion of the beam emerging from the silicon waveguide, which is further bent using the focusing power of the superlens 134. The output 140a of SLCO-SmfSiw 134 would be resulting in a vertically plane wavefront within a short propagation distance. The beam size of the output 140a is therefore enlarged to a value mating the large mode size of an optical fiber of typically around 5-8 m. Further, there is no reflection from the superlens-waveguide interface as there is no sharp refractive-index step. In an embodiment of the present invention, the coupling into fiber is greater than fifty percent and less than one hundred percent.
(247) The output beam 140a from one of the lenses of the SLCO-SmfSiw 134 is further wavelength demultiplexed through the multiplexer 136. In an embodiment of the present invention, the multiplexer 136 is a silicon ultra-compact wavelength multiplexer/demultiplexer (hereinafter referred to as Si-UCDeMux 136). (The general design of ultra-compact wavelength multiplexer/demultiplexer is given in patents Curved grating spectrometer, U.S. Pat. No. 7,283,233 and Integrated signal manipulator for manipulating optical signals, application Ser. No. 11/451,797, and is herein incorporated by reference.) In an embodiment of the invention, Si-UCDeMux 136 has a channel spacing of 5 nm. Si-UCDeMux 136 could also be referred to as Silicon Ultra-Compact Wavelength Multiplexer/DeMultiplexer, based on Ultra-Large-Angle-Grating for High Bandwidth Data Communications through WDM. In some cases, the output beam power at the demultiplexer 136 output is 25-80 W and practical loss is less than 3-6 dB.
(248) Ultra-Large-Angle-Grating is an optical component of the Si-UCDeMux 136 with a specific regular pattern, which splits light into multiple beams travelling in different directions depending on their wavelength channels. Each direction goes to an active device 138. In another embodiment illustrated in conjunction with
(249) One of the output beams 140b-140f from the Si-UCDeMux 136 is further coupled to the active layer of an active device 138 through a tapered waveguide coupler 602 (not shown in the figure). Tapered waveguide coupler 602 has been further described in detail in conjunction with
(250) PT-HE-WC-A-R 138 are photon transistors and require powering laser sources. The photon transistor PT-HE-WC-A-R 138 acts as an optical switching device. The wavelength converters of PT-HE-WC-A-R 138 are used to convert the input wavelengths to a common wavelength. The common wavelength is used by the light beams in the interior part of the EP device 100B. Further, it also provides an optical gain (G) to boost the input signal. In some cases, the optical gain, G is 10. PT-HE-WC-A-R 138 therefore, serves the function of both a pulse regenerator and an optical isolator. The speed of operation is enhanced by a combination of low switching power and use of saturation and stimulated emission. To keep the switching power low, a linear gain is used. The gain is also used to affect interference of coupler which results in a switching action.
(251) GAMCI based devices include a pair of energy-up photonic transistor (EUPT) and energy-down photonic transistor (EDPT). The pair works sequentially to provide multiple functionalities. The figure of merit provided by GAMCI based photonic transistor is much better than those provided by semiconductor amplifiers. The main function of the pair is to provide wavelength conversion, pulse regeneration and power amplification without broadband spontaneous-emission noise, polarization conversion to TE polarization for on-chip polarization-sensitive devices and optical isolation preventing feedback to the input optical fiber.
(252) In some embodiments of the present invention, the output beams 140g-140k from PT-HE-WC-A-R 138 have a power in the range of 250-800 W. The output beams 140g-140k are further sent to the interior part of the device with a power to sustain a loss greater than 10 dB before detection, to enable the propagation of the output beams 140g-140k to a considerable distance. Also, in another embodiment of the present invention, the output beams 140g-140k can be split into multiple beams to reach multiple destinations in the EP device 100B.
(253) In an embodiment of the invention, the beam from the coupler SLCO-SmfSiw 134 is sent through an exemplary route to a wavelength filter 142. In another embodiment of the present invention, the wavelength filter 142 is a silicon ring-resonator tunable wavelength filter (Si-R-TWF) 142. The Si-R-TWF 142 filters the wavelengths by picking a particular wavelength. Therefore, a single fiber channel of a particular wavelength could be selected. The output of Si-R-TWF 142 is further sent to a high-speed photo-detector (HS-PD) 212 (not shown in the figure). The HS-PD 212 is located in the interior part of the EP device. The Si-R-TWF 142 is further described in detail in conjunction with
(254) The Interior Part:
(255) The output beams from the input port 400 of the EP device are sent to different parts in the interiors of the EP device. Four different scenarios receiving the beams from the input are discussed below:
(256)
(257) 1. Optical Clock Distribution 202: The signal beam (i.e. the output beam) 140g-140k received from the input part 100C is further sent to multiple parts of the EP device for optical clock distribution 202 of signals. In an embodiment of the invention, the optical clock distribution 202 is such that each clock signal lies within a range of 1-2 mm with a similar value of tolerance, of any transistor circuit. The beam propagates for a considerably long distance in the device. The clocked signal distribution will be apparent to a person skilled in the art.
(258) 2. All-Optical Computing (AOC)/Signal Processing Chip 204: The signal beam received from the input port 100 is further sent to a circuit referred to as an all-optical chip 204. The all-optical chip 204 performs required ultra-fast all-optical computing before detecting and sending it to the output part. In some cases, the all-optical chip 204 also performs signal processing before detecting and sending it to the output part 300. All-optical computing will be apparent to a person skilled in the art.
(259) 3. Direct Detection 206: The signal beam is detected and converted to electrical signals. The electrical signals can be used to drive the electronic circuitry.
(260) These three scenarios as described above are used for direct data transport in the EP device. An integrated optical amplifier 214 provides a boost to weak signals propagating over a long distance. In an embodiment of the present invention, the optical amplifier 214 is an InP semiconductor optical amplifier for nano-waveguide (SOA-Nw). Narrow width of SOA-Nw 214 allows the SOAs to operate at low current. In some cases, the SOA-Nw 214 operates with current in the range 20-40 mA. In other cases, the width of SOA-Nw 214 can be 1.5 m. The current is approximately five times lesser than a conventional SOA.
(261) In an embodiment of the invention, to convert optical signals into electrical signals, an integrated photodetector 212 is used. The photodetector 212 is an InP high-speed PIN-diode Photodetector 212. Hereinafter, the photodetector 212 is interchangeably referred to as an InP high-speed PIN-diode photodetector (HS-PD) device 212. PIN diode in the InP high-speed PIN-diode photodetector 212 provides a working efficiency at wavelengths of about 1550 nm for a speed greater than 20 Gb/s. The significance of using PIN detector is the ease provided during integration with planar photonic circuits.
(262) 4. Interior Optical Interconnects 208: The signals from the electronic sub-circuitry are sent optically from one part of the EP device to another part by using interior optical interconnects 208.
(263) To serve the purpose of transmission of signal from part of the EP device to another, in some cases, an on-chip transmitter-receiver pair is required. A directly modulated laser 210 is used to power multi-destination interior optical interconnects for waveguides for multiple locations. The directly modulated laser 210 has small size, low threshold, reasonably high output, and high direct modulation speed. In an embodiment of the invention, directly modulated laser 210 provides high power efficiency and is referred to as low-threshold high-performance nano-laser (LTHP-NL) 210. The output power of LTHP-NL 210 is greater than 2 mW. The light beam from the LTHP-NL 210 is coupled to a silicon layer. In another embodiment of the invention, the LTHP-NL 210 splits into multiple destinations for point-to-point signal transport within the device, providing a peak output of 60-120 W.
(264) The Output Port:
(265) The output optical beams and output electrical signals from the interior part of the EP device are sent to different parts in the output port 402. The output port has multiple output parts 300a-300g of the EP device as per the requirements of the circuitry and the application in which the EP device is used. The output from the EP device can be fed to an electronic device or a photonic device. In some embodiments of the invention, a multiple optical fiber output port 402 with 10 optical fibers is used. The multiple optical fiber output port 402 is described in detail in conjunction with
(266)
(267) The output part 300 includes a grating laser 302, electro-optic modulators 304, a multiplexer/demultiplexer 306, coupling optics 308 and an optical waveguide 318. An external laser chip 312 and coupling optics 314 for vertical emission are separately provided in the output port 402. The optical beam path in the output port involves a grating laser 302 and the coupling optics 308 to bring the light beam in from the beam output ports of Interior Part 200, which is further provided to passive optical waveguides 318a-318e. The typical electrical signal path in the output port involves an electrical wire 320 to bring the electrical signals in from the electrical signal output wire 222 of Interior Part 200. With reference to
(268) From Electronics at output port: To transport a large amount of data out of the device, a specific grating laser 302 is used. The grating laser 302 is referred to as stepper-lithography compatible multiple-wavelength WDM curved grating laser (SC-WDM-CGL) 302. SC-WDM-CGL 302 provides multiple beams with different wavelengths having a pre-defined spacing. In one embodiment of the present invention, SC-WDM-CGL 302 provides 5 beams for 5 wavelength channels having a spacing of 5 nm. Further, SC-WDM-CGL 302 can be fabricated using stepper lithography. It will be apparent to a person skilled in the art that a stepper is a device used in the manufacture of integrated circuits (ICs) and has operation similar to that of a photographic enlarger. The process of lithography is a process used in micro-fabrication to selectively remove parts of a thin film or the bulk of a substrate. The grating pitch of few m and a compact size provide a high-quality laser with wavelength determined by the grating, resulting in a relatively smaller WDM laser.
(269) The beams from SC-WDM-CGL 302 are coupled to a silicon layer. Output from laser 302 is sent to a passive waveguide 318. In some cases, the coupling loss is less than 0.5-1 dB. The beam is further sent to silicon high-speed electro-optic modulators (Si-HS-EOMs) 304. Si-HS-EOMs 304 convert the electrical data to optical beam. It will be apparent to a person skilled in the art that electro-optic modulator 304 is an optical device having a signal-controlled element displaying electro-optic effect, which can be used for modulation. Electro-optic modulator 304 can be used in the invention to modulate light. In some cases, the output of Si-HS-EOM is 0.5-1 mW with a loss of 3-6 dB. Si-HS-EOMs 304 provide a modulation rate of 20 Gb/s.
(270) The output beams 318f-318k modulated by Si-HS-EOMs 304 are sent to a silicon ultra-compact wavelength multiplexer (Si-UC Mux) 306 that combines wavelengths to a single output beam. In an embodiment of the invention, the channel spacing provided by Si-UC Mux 306 is 5 nm. The power output is 125-500 W and a loss of less than 3-6 dB.
(271) The output beam from the Si-UC Mux 306 is further coupled to an optical fiber 310 through an integrated coupler 308. In an embodiment of the invention, the integrated coupler 308 is referred to as Superlens Coupling Optics for Single-Mode Fiber to Silicon Waveguide (SLCO-SmfSiW) 308 and is functionally similar to the SLCO-SmfSiW 104. A peak pulse power of 60-400 W from the fiber 310 could be achieved. The required power of greater than 200 W could be achieved within this range.
(272)
(273) The other input part of the input port 400 is a tunable wavelength filter 142. In an embodiment of the present invention, the tunable wavelength filter is referred to as Silicon Ring-Resonator Tunable Wavelength Filter (Si-R-TWF) 142. In Si-R-TWF 142, a silicon MZI resonator having a nano waveguide and a micro-ring or micro-disk resonator is used to achieve enhancement in speed and to maintain a low value of RC constant. The advantages of ring shape in a ring resonator will be apparent to a person skilled in the art. The resonator has a whispering gallery mode. In an embodiment, the resonator has a diameter less than 0.5 m. The whispering gallery mode provides an advantage by allowing lower electrode to be placed in the inner side of the ring/disk, by propagating around the edge of the ringdisk. This allows thicker core on the side, which leads to lower resistance for lower electrode. Further improvement is achieved by introducing Transparent Conducting Oxide (TCO) material as the upper gate electrode. TCO provides an effective waveguide cladding with fast evanescent decay of the guided field, the reason being a low refractive index n less than a value of 1.7 at 1550 nm. By placing top electrode on the top of a thin layer of TCO, a lower resistance is achieved, resulting in a much enhanced speed. The other advantages provided by TCO are simple fabrication steps. The TCO to be selected should have a low enough optical absorption at 1.5 m to avoid a high optical propagation loss coefficient . The desirable loss should be less than thirty percent. The use of Silicon MZI structure with micro-disk or ring resonators is incorporated. The structure provides a function of tunable wavelength filter 142 along with a fast optical switch. By thermally or electrically tuning the index, and therefore the resonant ring wavelength, filtering and switching can be achieved in a precise manner. The output of the wavelength filter 142 is provided to interior part of the device.
(274) According to yet another embodiment of the invention, the beam from the coupler is sent through another exemplary route to an optical switch 144. In an embodiment of the present invention, the optical switch 144 is a silicon high-speed optical switch (Si-HS-OS) 144. Hereinafter, the optical switch 144 is interchangeably referred to as Si-HS-OS 144. The Si-HS-OS 144 is designed to provide on-chip all-optical signal routing. The switching speed provided by all-optical switch (Si-HS-OS) 144 is less than 0.1 ns. The output beam power offered by Si-HS-OS 144 is in the range of 50-80 W.
(275)
(276) In an alternate route, an external InP WDM laser chip 312 emitting multiple wavelengths with 4 mW power per wavelength channel provides a WDM laser. The chip 312 is flip-chip bonded to the silicon wafer. In an embodiment of the invention, the coupling is done using Silicon wafer with light coupled into silicon layer through a superlens coupling optics for external chip to silicon waveguide coupling. Hereinafter, the coupler is interchangeably referred to as referred to as Superlens Coupling Optics on Si for External Silicon Chip (SLCO-EcSiw) 312. A practical coupling loss is less than 1-3 dB.
(277) From an external photonic chip: In one scenario, a light beam 316 from interior part of the device is coupled to an optical fiber or through a wavelength multiplexer 312. In another scenario, the light signals from a silicon waveguide are emitted out vertically to provide vertical emission through a coupling optics device 314 for vertical coupling to external device. In an embodiment of the invention, a coupling optics device 314 is referred to as Superlens Coupling Optics on Si for Vertical Emission (SLCOSiwVe) 314. Hereinafter, a coupling optics device 314 is interchangeably referred to as SLCO SiwVe 314. The external device is flip-chip bonded onto the silicon platform and coupled to a waveguide. The waveguide is referred to as a silicon waveguide through a superlens based external-chip optical coupler (SLCO-EcSiw) 312.
(278)
(279)
(280) General Specifications of the Devices and Structures:
(281) The details described here are for illustrative purpose only and they do not limit the scope of the invention in any way. For example, the waveguide can be made of any dielectric material (for example InP, GaAs, GaN, Titanium Dioxide, Silicon Dioxide, Silicon Nitride) as long as waveguiding function can be achieved as is known to those skilled in the art, and need not be silicon. The various powers are for illustration only and the optical power in these devices can be of any value as long as the device or waveguide can withstand it as is well known to those skilled in the art. The wavelength does not have to be around 1550 nm and can be any optical wavelength ranging from deep ultra violet to far infrared, depending on the materials used as is known to those skilled in the art. The waveguide and device dimensions are in general, linearly proportional to the optical wavelength used, as is known to those skilled in the art. The connection between any two devices, when not specifically stated, is generally achieved by using optical waveguide. Each device can be optional depending on the photonic function to be achieved. For example, at the input part, the photon transistor is optional (that is, it can be absent and a waveguide can be replaced in lieu of the device) if the function is wavelength demultiplexing. This can be achieved by the ultra-compact grating alone and the pulse regeneration function of the photon transistor is not essential. At the output part, the optical modulator is optional (that is, it can be absent and a waveguide can be replaced in lieu of the device) if the function is to produce intensity modulation of the laser light. This can be achieved by directly modulating the current to the semiconductor laser. In the input part, output part, or interior part, the optical amplifier is optional (that is, it can be absent and a waveguide can be replaced in lieu of the device), as long as no intensity amplification is needed. This can be replaced by the use of low-loss waveguide, higher-efficiency photodetector, or higher-power laser. As is discussed above, the EIC-PIC may be configured with different set of devices in different ways and can be as simple as one device with one input port or a single device with one output port to perform a single function or as many devices as needed with many input parts and many output ports to perform a plurality of functions.
(282) Furthermore, the device performing the optical wavelength converter or pulse regeneration or pulse shaping function does not have to be the photon transistor described and it can be replaced by other devices performing the same photonic function, such as Semiconductor-Optical-Amplifier (SOA) based wavelength converter and pulse regenerator or pulse shaper. Likewise, the wavelength multiplexing and wavelength demultiplexing devices do not have to be the ultra-compact gratings and they can be replaced by array-waveguide grating (AWG) as long as the wavelength multiplexing and demultiplexing functions can be achieved. The laser does not have to be the grating laser described and can be other semiconductor lasers such as Distributed Feedback (DFB) Laser, Bragg-Reflector (DBR) Laser, Sample-Grating Distributed Feedback Laser (SGDBR), Fabry Perot (FP) Laser, Ring Laser, Microdisk Laser, Microloop Mirror Laser, and it can be replaced by wavelength tunable or wavelength selectable or fixed wavelength laser, as long as the lasing light can be produced. The modulator does not have to be the silicon modulator described and it can be replaced by other silicon modulator, organic modulator, InP modulator, GaAs modulator, as long as the intensity modulator function can be achieved. The ring resonator does not have to be single ring and it can be replaced by multiple ring resonator or other tunable filter as long as the wavelength filtering function can be achieved. The optical input or output coupler does not have to be superlens, and it can be replaced by grating coupler, tapered waveguide coupler, or other optical beam coupling devices. The up-down coupling device does not have to be the tapered waveguide coupler, and it can be replaced by an up-down coupling device as long as energy can be coupled from the passive layer to the active layer. Many of these other devices, with proper design, can also be fabricated using the fabrication processes described here.
(283) Fabrication Method:
(284) The fabrication part of the device through monolithic integration process includes monolithic integration of active and/or passive photonic devices and their process insertion into the CMOS process and monolithic integration of compound semiconductor (e.g. III-V semiconductor) active and/or passive devices and their process insertion into the CMOS process.
(285) Monolithic integration provides a number of advantages. It provides best way to minimize any trade-offs in performance. Further, monolithic integration on a substrate like silicon provides a better heat-conducting platform, resulting in improved power efficiency of the EP device components including electronic transistors and lasers. In addition, the power requirement is also lowered and an increase in the output power is observed.
(286) Monolithic integration reduces the need for pick-and-place each component, bonding of components and other assembly processes necessary in hybrid assemblies, including those that place discrete elements on a silicon substrate. The integration of compound semiconductor (e.g. III-V semiconductor) devices on silicon is based on three important CMOS compatible processes. These processes include:
(287) 1. Wafer-level bonding at low temperature, with N-side down technique for near-epi-wafer-smooth bonding surface. In wafer processing, monolithic integration of the photonic and electronic structures and process requires the development of highly skilled process techniques. During the manufacturing process, the management of thermal budgets and surface planarity issues flow must be done to eliminate or minimize adverse effects on all of the integrated circuit elements on the chip.
(288) Furthermore, wafer bonding of compound semiconductor (e.g. III-V semiconductor) devices to silicon is a low-temperature process. The process involves silicon-nitride (Si.sub.3N.sub.4) bonding with good bond strength, good thermal conduction, and thermal matching to InP and silicon. In the process of wafer bonding, InP is introduced only after the heat treatment procedure to avoid any adverse effects caused by heat treatment, which may lead to a change in the properties of InP, which results in a difference in the functionality provided by InP.
(289) 2. Re-growth free III-V integration of passive, active and WDM broad band devices based on CMOS-compatible quantum well intermixing (QWI) technique. The QWI technique will be apparent to a person skilled in the art. The QWI technique is developed with a low loss and wide spectral shift of a value greater than 120 nm. The process provides better passive-active sections having multiple bandgap energies within a few steps for a wide variety of integrated devices, ranging from lasers, modulators, to detectors. This results in a wide wavelength range coverage (>40 nm) of WDM devices.
(290) 3. Tapered-waveguide up-down coupling technique to provide up-down coupler to steer light between the silicon and the compound semiconductor (e.g. III-V semiconductor) layer. Such tapered up-down coupler has a good tolerance to fabrication process variation. Efficient coupling of the light beam between compound semiconductor (e.g. III-V semiconductor) layer and Si is achieved using tapered waveguides. This approach does not need precise dimensions control as it is based on phase matching that will self-adjust to occur at slightly different part of the tapering section in case there are dimensional variations. TE/TM balanced coupling can be engineered.
(291)
(292)
(293) In an embodiment of the invention, the InP based active layer comprises an InP layer 156 with total thickness of 300 nm and at the center of this layer is a quantum well structure 150 with five InGaAs quantum wells having thickness of 8 nm for each quantum well spaced by 10 nm of InP. The InP layer 148 above the quantum well structure is n-doped and the InP layer 152 below the quantum-well layer is p-doped. The n and p doping enable this active layer to be electrically pumped to achieve desired optical gain. The top Ohmic contact layer 154 is n-doped InGaAs and the bottom Ohmic contact layer is p-doped InGaAs. The passive waveguiding layer 160 is of silicon, with a thickness of 300 nm, the intermediate layer 162 is of silicon nitride with a thickness of 300 nm, the lower cladding layer 164 is of silicon dioxide, and the substrate 166 is of a silicon substrate. The passive waveguiding layer 160 is also referred to as waveguide core 160.
(294) The waveguide core 160 is on top of a transparent lower waveguide cladding 164. The lower waveguide cladding 164 is on top of a substrate 166. The waveguide core layer 160 material is a transparent material with high refractive index nCOR to guide the optical wave. The transparent lower cladding material layer 164 has a refractive index nCL lower than that of the waveguide core 160. As is well known to those skilled in the art, the waveguide core or cladding material can be chosen from semiconductors (Si, InP, GaAs, GaN etc), dielectrics (TiO.sub.2, SiO.sub.2, Si.sub.3N.sub.4), and other transparent materials. The substrate 166 provides mechanical rigidity and can be of any mechanically strong material; including semiconductor, polymer, dielectrics, metal, ceramic etc.
(295) According to an embodiment of the present invention, the optical beam has optical wavelengths around 1.5 micrometers (1.5 m). The waveguide core layer 160 is of silicon, which is transparent at the wavelength of 1.5 m. The silicon waveguide core layer 160 has a refractive index nSi=3.5, and the thickness of layer 160 is 300 nanometers (300 nm). Furthermore, the cladding material layer 164 is of silicon dioxide (SiO.sub.2) with refractive index n.sub.SiO2=1.5 and a thickness of 300 nm or larger. The substrate material 166 is silicon (Si). In an embodiment of the invention, the waveguide core 160, lower waveguide cladding 164, and substrate 166 are formed by a Silicon-On-Insulator (SOI) substrate. In some cases, the wavelength of the input beam is around 1.5 m. In an embodiment of the invention, coupling of beam from waveguiding layer 160 to the active layer structure 156 is achieved by the up-down tapered waveguide coupler 602a and the beam energy efficiency can be as high as ninety five percent. In another embodiment of the invention, it is preferred that high-frequency CMOS applications have thin buried oxide (BOX) of 100 nm-500 nm and the top silicon layer thickness of 5 nm-200 nm.
(296) For photonic applications, the silicon layer thickness needs to be thicker, which could be managed by using epitaxial growth of silicon at the photonic device area. However, it is a difficult task to change the BOX thickness. For a given BOX thickness, there is a minimal silicon thickness needed to confine TE/TM losses to a value lesser than 1/cm. For example, for a BOX of 300-500 nm, a good silicon thickness to be used would be 0.3 m.
(297)
(298) The steps discussed are a combination of standard CMOS process and additional CMOS compatible steps that aid in the fabrication of both Si and compound semiconductor (e.g. InP) based photonic devices. Steps 4, 7, 10, 11, 13, 15, 20, 21, 22, 23, and 24 are the additional CMOS compatible steps according to various embodiments of the invention.
(299) The silicon based photonic devices are fabricated on the Silicon-On-Insulator (SOI) wafer together with the CMOS components using CMOS wafer-level process. The steps 1 to 18 are high temperature processes. Step 18 is the beginning of the CMOS metal-interconnect process. The compound semiconductor (e.g. III-V InP) wafer bonding begins either after step 17 or at step 19. The figures illustrated are assumed to have compound semiconductor (e.g. III-V InP) wafer bonding after step 19. This is to ensure that the properties of compound semiconductor (e.g. III-V InP) wafer do not change because of high temperature involved in steps 1 to 18. Steps 2, 4, 7, 10, 14, 16, 20, 22, 24, and 26 are illustrated in
(300) i) InP/InGaAsP Wafer Bonding and Etch Back: An epitaxial layer structure is grown on top of an InP substrate (InP-S). In some cases, the epitaxial layer further requires a 0.2 m thick InGaAsP n-F doped etch-stop and n-contact layer (ES-NC), which is grown on the InP substrate. This is followed by a thin cladding layer (CL1) of InP/InGaAsP materials having a thickness of 0.2 m. The use of n+ doped InGaAs as N+ contact may lead to a high optical absorption due to the smaller bandgap energy. Furthermore, in some other cases, a thin waveguiding core layer (WC1) of 0.25 m is grown. WC1 layer includes approximately 5-10 quantum wells within the layer. There is a 1.5 m thick cladding layer (CL2) on the top of WC1. After CL2, a thin InGaAs P+-doped p-contact layer (PC) is then grown.
(301) ii) N-Side Down for Near-epi-Wafer-Smooth Bonding Surface: After the epitaxial growth, an InP handling wafer (HW) is bonded to the epi-wafer through a Benzocyclobutene (BCB) polymer. Further, the epi-wafer substrate (InP-S) is etched by hydrochloric acid (HCl), which self-stops at the ES-NC due to the low etching rate of the InGaAsP layer. A thin silicon nitride layer (Si.sub.3N.sub.4InP) is then grown on top of the n-contact layer (ES-NC). In some cases, the thin silicon nitride layer has a thickness of 0.1 m. ES-NC layer has Epi-Wafer-smooth bottom surface as it is the first layer after the Epi-wafer substrate (InP-S). Therefore, the requirement for wafer bonding is accomplished, as thin Si.sub.3N.sub.4 layer provides a smooth surface.
(302) iii) Local-Area Wafer Bonding to SOI: As described earlier, the thin silicon nitride layer is deposited on top of the silicon surface of the SOI wafer designated for compound semiconductor (e.g. III-V semiconductor) devices. The pre-etched silicon waveguides are provided for achieving vertical coupling. As this is a thin layer on the original SOI wafer surface, the Si.sub.3N.sub.4Si layer will have smooth surface, meeting the requirement of wafer bonding. The wafer bonding takes place between the Si.sub.3N.sub.4 surface on the III-V InP wafer and the Si.sub.3N.sub.4 surface on the SOI wafer. To enable the wafer bonding, the silicon-nitride surfaces are subjected to plasma treatment, typically with 1-minute exposure to oxygen plasma (200 mTorr, 200 W) to create hydrophilic surfaces. Further, InP (thermal expansion coefficient .sub.InP=4.610-6/K) is more thermally matched to silicon (.sub.Si=2.610-6/K) than GaAs (.sub.GaAs=5.710-6/K) which aids in meeting the smooth surface requirement of wafer bonding. The thermal coefficient of silicon nitride is .sub.Si3N4=3.310-6/K which is in between that of silicon and InP. This temperature difference reduces temperature-induced stress. The wafers are then brought together and bonding occurs across the two silicon nitride surfaces. Bond strength of 0.16 Jm-2 is achieved by the process of annealing at 150 C. for greater than one hour. Higher bond strength of 1.1 Jm-2 can be achieved at low temperature with a annealing time of greater than 100 hours or by keeping a moderate annealing temperature of less than 400 C. The silicon nitride offers higher thermal conductivity, which is almost comparable to that of InP. It will be apparent to a person skilled in the art that the process is assumed to be an exemplary process. This process enables multiple areas to be bonded with compound semiconductor (e.g. III-V semiconductor) epitaxial layers that are subsequently used for making the active photonic devices on those areas, which is referred to as local-area wafer bonding. The process described here is for illustrative purpose only and it does not limit the scope of the invention in any way. Different combinations of these steps and other similar steps are contemplated and are within the scope of the present invention. d) High-spatial Resolution Bandgap Engineering: An ideal nano-photonics integration & device optimization technology shall have the following properties:
(303) (1) High spatial resolution: 1 m spatial resolution is required in joining active-passive sections.
(304) (2) High spectral resolution & uniformity: A wavelength of 5 nm spectral resolution for shift in gain peak with uniform shift across a 2 or 3 wafer with less than 5 nm wavelength variation.
(305) (3) Low passive section loss: The loss of the passive section is low with <1/cm.
(306) (4) High active section quality: The active section gain similar to as-grown materials.
(307) (5) Low optical reflection at joint: Joint reflection is less than 0.05% so gain greater than 1,000 is allowed.
(308) These properties are achieved through an advanced passive-active device integration technology based on quantum well intermixing (QWI). This technology is applicable to InGaAsP/InP materials at 1.3-1.55 m, which do not require materials regrowth. In an exemplary embodiment of the QWI process, the QWI process involves ion-implantation with thin layer of oxides as implant mask, therefore enabling better control of area-specific bandgap shifts. In another embodiment of the QWI process, a low-energy temperature-assisted ion-implantation quantum well intermixing (LETAI-QWI) is used. The low-energy temperature-assisted ion-implantation quantum well intermixing process is described in patent Method for shifting the bandgap energy of a quantum well layer, U.S. Pat. No. 6,878,562 and Method for quantum well intermixing using pre-annealing enhanced defects, U.S. Pat. No. 6,984,538, and are hereby incorporated by reference. In an exemplary embodiment of the LETAI-QWI process for InGaAsP/InP materials, phosphorus ions are implanted near the surface of the wafer having low ion implantation energy of 360 keV, at about 0.5-1.5 m away from the quantum well structure. The implantation process is a controlled process in order to create point lattice vacancies or point defects. These point defects are then diffused down to the quantum well region through thermal process. This results in a large shift in energy bandgap of a value greater than 120 nm. This is done by inducing inter-diffusion of atoms between the quantum wells and barrier materials of quantum wells. The well is InGaAs and phosphorus ions are diffused from the InGaAsP barrier into the well, causing the well to be effectively narrower. This narrowing of the quantum well causes a blue shift in the energy bandgap. The LETAI-QWI process is an improvement over the already existing low-temperature QWI process. The LETAI-QWI process can also be applied to both unstrained and strained quantum wells. Polarization insensitive devices could be achieved using strained quantum well.
(309) One-Step Process for Multiple Bandgaps:
(310) In another exemplary embodiment of the LETAI-QWI process, multiple bandgap energies could be achieved by fabricating SiO.sub.2 ion implant masks of differing thicknesses on the wafer. This is done by using a process referred to as photolithography having a gray-scale mask. The gray scale mask exposure results in variation in the photo-resist thickness, which is transferred into variations in the thickness of SiO.sub.2. The different SiO.sub.2 thicknesses block the ions in a different manner. This results in different ion dosages with only one implant, giving multiple bandgap energies. Therefore, 5-10 bandgap shifts could be achieved for the gain curve with basically a one-step process.
(311) Excellent Gain Materials Quality:
(312) Five different lasing wavelengths from Fabry-Perot lasers fabricated with one implant on one wafer are used. The lasing wavelength spans across a bandwidth of greater than 100 nm, with the most wavelength-shifted laser showing smaller change (of less than ten percent) in the lasing thresholds. Therefore, processed wafer could give much better gain material quality.
(313) High Spatial Resolution:
(314) The large energy shift in bandgap provides a good passive section with measured loss of less than 1/cm (4 dB/cm) including the free-carrier absorption of the p-side dopants. In an exemplary embodiment of the LETAI-QWI process, the LETAI-QWI process is capable of spatial resolution of 1 m as ion-implantation is a spatially well defined process. In case of thin-film nano-photonics, implantation is at 0.5 m away from the quantum well and expected spatial resolution is less than 0.5 m. A 0.5 m top InP will be etched away after QWI.
(315) High-quality integrated device fabrication: Examples of high performance integrated devices have been previously realized using the LETAI-QWI bandgap engineering method. Some of these examples include:
(316) (1) Polarization insensitive optical-amplifier photo-detector involving a SOA 214 integrated with 10 GHz PIN detector;
(317) (2) Integrated DWDM grating optical spectrometer (integration of grating with detectors). In case of grating spectrometer, a propagation length of 1 cm and a propagation loss of less than fifty percent could be achieved for a large passive waveguiding region is bandgap shifted by a value greater than 120 nm.
(318) The electronic-photonic device consolidates all the optical functions required in an optical transport system into a single device. Monolithic integration provides the greatest simplicity and reliability benefits when consolidating optical components into a single device. Complete photonic and electronic integration will increase subsystem functionality, robustness and design flexibility. Optical interconnects reduce electromagnetic interference and improve signal integrity better than traditional electrical-transport methods. By significantly reducing the number of fiber couplings between components, the device reduces the number of failure points and improves system reliability.
(319)
(320) In the following description, references are made to
(321) In accordance with an embodiment of the present invention, the optical wavelength demultiplexer 160 is fabricated into the waveguide core layer 160 and cladding layer 164. The optical wave in demultiplexer 136 is propagated in waveguide core layer 160. The output from the optical demultiplexer 136 is sent to the active device 138 through a vertical up-down tapered waveguide coupler that couples light from the passive waveguiding layer 160 layer to an Active Structure Layer 156. Active Structure Layer 156 is made up of optical materials that can provide amplification, absorption, or phase shifting of optical energy and can be semiconductor (InP, GaAs, GaN etc) or doped materials (erbium-doped glass or semiconductor) or any other optical material that can provide optical gain, loss, or phase shifting. Active device 138 can be powered electro-optically or all-optically, as is known to those skilled in the art. Active Structure Layer 156 is on top of an intermediate material layer 162. The intermediate material layer 162 is on top of the passive waveguiding layer 160. The active structure layer has a waveguide core 150 with a refractive index n.sub.ACT-COR. The intermediate layer 162 is any optical dielectric material characterized by a refractive index n.sub.INT. In an embodiment of the present invention, n.sub.INT has a value smaller than n.sub.ACT-COR and n.sub.COR so that the optical beam can be freely coupled from the passive layer 160 to the active layer 156 using the up-down tapered waveguide coupler 602a and from the active layer 156 to the passive layer 160 using the up-down tapered waveguide coupler 602b. In an exemplary preferred embodiment, the intermediate layer 162 is Silicon Nitride (Si.sub.3N.sub.4) with thickness of around 300 nm and refractive index n.sub.INT=1.9 and the active layer is thin-film InP based active layer with refractive index of n.sub.ACT-COR of about 3.4. The passive layer is 300 nm thick silicon with refractive index of n.sub.ACT-COR of also about 3.4. Silicon Nitride has a thermal expansion coefficient between Silicon (Si) and Indium Phosphate (InP), which will enable the thermal stress due to mismatch of the thermal coefficients between the passive and active layer to be minimized (i.e. it acts as a stress releasing layer as is known to those skilled in the art). Silicon Nitride also has high thermal conductivity compared to other dielectrics such as silicon dioxide, which will help in dissipating heat from the active devices to the silicon substrate. As is known to those skilled in the art, in one embodiment the InP based active layer comprises an InP layer 156 with total thickness of 300 nm and at the center of this layer is a quantum well structure 150 with five InGaAs quantum wells having thickness of 8 nm for each quantum well spaced by 10 nm of InP. The InP layer 148 above the quantum well structure is n-doped and the InP layer 152 below the quantum-well layer is p-doped. The n and p doping enable this active layer to be electrically pumped to achieve optical gain.
(322) The refractive index of n.sub.INT does not have to be smaller than n.sub.COR and n.sub.ACT-COR. In another embodiment, the layer 162 has a refractive index n.sub.INT that can be comparable to or larger than either n.sub.COR or nACT.sub.-COR or both or layer 162 can be entirely omitted. In this embodiment, energy is leaked from the passive layer 160 to the active layer 156 so that the light beam can interact with the active materials.
(323) Light beam from active device 138 is coupled down to the passive layer at the output and the beam at the output is transmitted to the Interior Part described below.
(324)
(325) In the following description, references are made to
(326) In an embodiment of the invention, the optical wavelength waveguide 216 is fabricated into the waveguide core layer 160 and cladding layer 164. The optical waveguide 216 is propagated in waveguide core layer 160. The output from the optical waveguide 216 is sent to the active device 214 through a vertical up-down tapered waveguide coupler 602a that couples light from the passive waveguiding layer 160 layer to an Active Structure Layer 156. In one embodiment, active device 214 acts as an optical amplifier and amplifies the light beam power. The beam traveling through Active Device 214 at active layer 166 is coupled down to the passive layer 160 using up-down tapered waveguide coupler 602b. Active Structure Layer 156 is as described above in
(327) Beam from the output of Active Device 214 at the passive layer 160 is propagated along a passive waveguide 216 at the passive layer 160. The beam is subsequently sent to active device 212 through a vertical up-down tapered waveguide coupler 602a that couples light from the passive waveguiding layer 160 layer to an Active Structure Layer 156. In one embodiment, active device 212 acts as an optical photodetector and detects the light beam power. The photodetector 212 converts the optical power to electrical power. The electrical current from the photodetector is transmitted to an electrical circuit 218a via an electrical wire 220c. The electrical circuit 218a processes the data now carried by electrical current and voltage. The electrical circuits 218a to 218d produce electrical outputs carried by electrical wires 222a to 222e. The electrical outputs are sent to the Output Port 402. In the Interior Part, there are optical beam path ways that are not terminated by a photodetector but are processed optically and the optical data after processing is sent to the output port via an optical waveguide 216k in the transparent waveguiding layer 160. The optical waveguide 216, electrical circuit 218, electrical wires 220, and electrical wires 222 are further described in detail later.
(328)
(329) To describe these figures, reference will be made to
(330) As illustrated in
(331) In accordance with an embodiment of the present invention, the optical wavelength waveguide 318 is fabricated into the waveguide core layer 160 and cladding layer 164. The optical waveguide 318 is propagated in waveguide core layer 160. The output from the optical waveguide 318 is sent to the fiber coupling optics 308 to couple light into an optical fiber 310 as shown in
(332) The typical electrical signal path in the output port involve an electrical wire 320 to bring the electrical signals in from the electrical signal output wire 222 of Interior Part 200. The electrical signal is then sent to modulate a semiconductor laser 302 through direct current modulation. Alternatively, electrical signal through electrical wire 320 is sent to modulate the laser beam from the laser 302 by modulating an optical intensity modulator 304 connected to the output of the semiconductor laser 302 as illustrated in
(333)
(334) After the fabrication process, a general cross section of an EIC-PIC 2200 fabricated is shown in
(335) General Specifications of the Devices Fabrication:
(336) The various specifications of the above exemplary device fabrication for an EP chip are given for the purpose of illustration and not limitation for a currently preferred exemplary embodiment of an EIC-PIC. For example, the CMOS process may be replaced by other electronic integrated circuit fabrication or CMOS fabrication processes as long as the low-temperature photonic active-device processes done with wafer bonding is performed after all the high-temperature processes in the electronic integrated circuit fabrication or CMOS fabrication processes. The silicon photonic device fabrication can be done after or before the high-temperature processes. The electronic integrated circuit may be necessarily based on silicon and can be based on GaN material or other semiconductor materials (InP and GaAs).
(337) While the preferred embodiments of the invention have been illustrated and described, it will be clear that the invention is not limited to these embodiments only. Numerous modifications, changes, variations, substitutions and equivalents will be apparent to those skilled in the art, without departing from the spirit and scope of the invention, as described in the claims.