HETEROGENEOUS INTEGRATED PHOTONIC PLATFORM WITH InGaP LAYER

InGaP layers with butt coupling and mode conversion techniques address the challenge of efficient optical coupling in heterogeneous integrated lasers, enabling high-performance PICs with reduced losses and scalable manufacturing.

JP2024161334AActive Publication Date: 2024-11-19ネクサスフォトニクスインコーポレイテッド
View PDF 8 Cites 0 Cited by

Patent Information

Application Number
JP2023113077
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-07
Filing Date
2023-07-10
Publication Date
2024-11-19
Estimated Expiration
2043-07-10

AI Technical Summary

Technical Problem

Current heterogeneous integrated lasers face challenges in achieving efficient optical coupling between dissimilar materials, particularly at short wavelengths, due to large differences in refractive indices, leading to complex and costly manufacturing processes and high optical losses.

Method used

The use of InGaP layers for active devices in combination with butt coupling and mode conversion techniques to facilitate high-yield bonding and robust semiconductor processing, enabling efficient optical coupling between active and passive waveguides.

Benefits of technology

This approach allows for high-performance, scalable manufacturing of photonic integrated circuits (PICs) that operate at short wavelengths with reduced optical losses and improved manufacturing yield, supporting a broadband wavelength range and higher optical powers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024161334000001_ABST
    Figure 2024161334000001_ABST
Patent Text Reader

Abstract

To provide a heterogeneous laser and a PIC that operate around and below the GaAs bandgap wavelengths and use butt coupling and tapered structures to promote efficient coupling between active and passive waveguides to provide better performing active devices.SOLUTION: A heterogeneous device 100 includes a passive waveguide structure 108 and an active waveguide structure. The passive waveguide structure is attached to a substrate 105 and includes a dielectric layer. The active waveguide structure is attached on the top of the passive waveguide structure and includes a quantum well layer on top of an InGaP layer 101-1. The InGaP layer provides an n-contact function.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to photonic integrated circuits, and more particularly, certain embodiments of the present invention relate to improved performance of heterogeneously integrated lasers and active components having waveguides operating at short wavelengths using optically coupled dissimilar materials. [Background technology]

[0002] A photonic integrated circuit (PIC) or integrated optical circuit is a device that integrates multiple photonic functions, similar to an electronic integrated circuit. The major difference between the two is that photonic integrated circuits provide the functionality for information signals carried on optical carrier waves. The most commercially available material platform for photonic integrated circuits is indium phosphide (InP), which allows the integration of a variety of optically active and passive functions on the same chip. While many current PICs are realized on the InP platform, there has been significant research in the past decade into using silicon rather than InP for the realization of PICs, due to some of the superior properties of silicon and its superior processing power, which leverages the investments already made for electronic integrated circuits.

[0003] The biggest drawback in utilizing silicon for PICs is that it is an indirect bandgap material making it difficult to provide an electrical pump source. This problem is typically solved by assembling a PIC containing two or more chips made from dissimilar materials in separate processes. Such an approach is difficult due to the need for very fine alignment, increasing packaging costs and limiting scaling. Another approach to solving the bandgap problem is to bond two dissimilar materials together and process them together, removing the need for precise alignment during bonding of larger parts or full wafers of dissimilar materials and allowing for mass production. In this disclosure, we use the term "hybrid" to describe the first approach, which involves precise assembly of separately processed parts, and we use the term "heterogeneous" to describe the latter approach, which involves bonding two materials together and then processing the bonded result to define waveguides and other components of interest.

[0004] To transmit optical signals between dissimilar materials, the heterogeneous approach utilizes tapers whose dimensions are gradually decreased until the effective modal refractive indices of the dissimilar materials match, and there is efficient power transfer. This approach generally works well when the materials have small differences in refractive index, as is the case for silicon and InP. When there are larger differences in effective refractive index, for example between SiN and GaAs (and their ternary and quaternary systems), the requirements on the taper tip dimensions become prohibitive, limiting efficient power transfer. In particular, extremely small taper tip widths (on the order of tens of nanometers) may be necessary to provide good coupling. Achieving such dimensions is complex and cost prohibitive.

[0005] While InP and silicon-based PICs address many current needs, they have some limitations; notably the fact that the operating wavelength range is limited by material absorption, which increases losses, and that there is a limit to the maximum optical intensity, and therefore optical power, that a PIC can handle. To address these limitations, SiN, SiNO x Alternative waveguide materials such as LiNbO3, TiO2, Ta2O5, AlN or others have been considered. Generally, such dielectric waveguides have higher band gap energies that provide better high power handling and transparency at shorter wavelengths, but generally such materials also have lower refractive indices. For example, SiN with a band gap of about 5 eV has a refractive index of about 2, AlN has a band gap of about 6 eV and a refractive index of about 2, and SiO2 with a band gap of about 8.9 eV has a refractive index of about 1.44. In comparison, GaAs has a refractive index >3. This makes the tapered approach difficult.

[0006] The alternative hybrid approach suffers from the drawbacks already mentioned above, namely the need for precise alignment and the corresponding complex packaging and scaling limitations.

[0007] A recent approach to the problem discussed above was presented in US Patent 10,859,764, which uses butt coupling in combination with a mode converter to allow heterogeneous processes to be used without the need for extremely small taper widths. While it provides good performance when operating at longer wavelengths, it becomes more difficult when operating at shorter wavelengths, mainly due to limitations related to the contact layers. In many native lasers (e.g., commercial InP or GaAs lasers made on the same type of substrate), the current typically spreads from top to bottom, so thick cladding can be used to control losses from the contact layers, even if those layers are absorbing at the operating wavelength. In a typical heterogeneous integration, the current typically spreads laterally from top to bottom, so the mode has a large overlap with at least one of the contact layers. For heterogeneous silicon integration, see e.g. Komljenovic et al., “Photonic integrated circuits using heterogeneous integration on silicon” (DOI: https: / / doi.org / 10.1109 / JPROC.2018.2864668), and for heterogeneous silicon nitride integration, see e.g. Tran et al., “Extending the spectrum of fully integrated photonics to submicrometre wavelengths” (DOI: https: / / doi.org / 10.1038 / s41586-022-05119-9). The former heterogeneous integration approach is typically used to cover the wavelength range from 1250 nm to 1700 nm, while the latter approach, including butt coupling and mode conversion, is used to cover the wavelength range from about 975 nm to 1050 nm and can operate down to about 900 nm.

[0008] One proposed solution is to increase the cladding thickness so that the bottom contact is farther away from the optical mode, as discussed in U.S. Patent Application Serial No. 18,071,239, but this approach may require a more complex coupling structure (typically with more than one layer) and a more complicated overall process, while still resulting in relatively high coupling losses between the gain / active and passive structures. One of the main reasons for the higher losses is the large vertical offset of the centers of the optical modes in the active and passive structures. This large offset also limits the ability to use the passive structures to evanescently couple into the active structure and define periodic structures, which is a common method of making distributed feedback (DFB) lasers.

[0009] Many applications require light sources and PICs that operate at wavelengths below 900 nm. In contrast, GaAs-based gain material systems offer good performance down to wavelengths of 630 nm. GaAs lasers typically utilize GaAs contacts, whose bandgap is about 870 nm, which introduces very high losses when the optical mode of the laser operating near or below the bandgap wavelength has overlap with the contact layer. To provide transparency at shorter wavelengths, Al x Ga 1-x Although As is sometimes used as a cladding and can be used as a contact, high Al concentrations have significant difficulties in many aspects of heterogeneous lasers: it is difficult to process, can result in higher losses, has reliability concerns, is very difficult to make good contacts with (heat budget limitations in heterogeneous lasers due to differences in thermal expansion coefficients between the substrate and the bonded material), and is difficult to bond with high yields. For all these reasons, it is clear that there is a need for high performance heterogeneous lasers that can operate below the GaAs band gap wavelengths. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] U.S. Pat. No. 10,859,764 [Patent Document 2] U.S. Patent Application Serial No. 18,071,239 [Non-patent literature]

[0011] [Non-Patent Document 1] Komljenovic et al. “Photonic integrated circuits using heterogeneous integration on silicon” (DOI: https: / / doi.org / 10.1109 / JPROC.2018.2864668) [Non-Patent Document 2] Tran et al. “Extending the spectrum of fully integrated photonics to submicrometre wavelengths” (DOI: https: / / doi.org / 10.1038 / s41586-022-05119-9) Summary of the Invention [Problem to be solved by the invention]

[0012] The present invention is directed to heterogeneous lasers and PICs that operate around and below the GaAs bandgap wavelengths and use butt coupling and tapered structures to facilitate efficient coupling between active and passive waveguides to provide better performing active devices. In particular, the embodiments described below relate to the detailed design of laser structures that support high yield bonding, high performance contacts, and robust semiconductor processing with high uniformity and yield. [Brief description of the drawings]

[0013] [Figure 1]1 illustrates a device according to one embodiment of the present invention shown in cross-sectional end view. [Diagram 2] The effect of contact layer thickness on several confinement factors in an example heterogeneous laser is shown and three example mode shapes are provided. [Diagram 3] 1 shows cross-sectional views from six edges of a device according to some embodiments of the present invention and illustrates processing steps for making the same. [Figure 4] 1 shows a cross-sectional view of a device according to some embodiments of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] Described herein are embodiments of a platform for the realization of photonic integrated circuits using wafer bonding and deposition of dissimilar materials, with improved optical coupling through the use of mode conversion and butt coupling schemes. More specifically, certain embodiments of the invention relate to detailed designs of laser structures that support high yield bonding, high performance contacts, and robust processing with high uniformity and yield when operated at short wavelengths.

[0015] In the following detailed description, reference is made to the accompanying drawings which form a part of this specification, in which like numerals designate like parts throughout and which show by way of example embodiments in which the subject matter of the present disclosure may be practiced. It is to be understood that other embodiments may be utilized and structural or logical changes may be made without departing from the scope of the present disclosure. Therefore, the following detailed description should not be taken in a limiting sense, and the scope of the embodiments is defined by the appended claims and their equivalents.

[0016] The description may use perspective descriptions such as top / bottom, inside / outside, over / under, etc. Such descriptions are merely used for ease of discussion and are not intended to limit the application of the embodiments described herein to a particular orientation. The description may use the phrases "in one embodiment" or "in an embodiment," each of which may refer to one or more of the same or different embodiments. Furthermore, terms such as "comprising," "including," and "having" used with respect to the embodiments of the present disclosure are synonymous.

[0017] For purposes of this disclosure, the phrase "A and / or B" means (A), (B), or (A and B). For purposes of this disclosure, the phrase "A, B, and / or C" means (A), (B), (C), (A and B), (A and C), (B and C), or (A, B, and C).

[0018] The term "coupled with," along with its derivatives, may be used herein. "Coupled" may mean one or more of the following: "Coupled" may mean that two or more elements are in direct physical, electrical, or optical contact. However, "coupled" may also mean that two or more elements are in indirect contact with one another but still cooperate or interact with one another, and one or more other elements are coupled or connected between the elements that are said to be coupled to one another. The term "directly coupled" means that two or more elements are in direct contact on at least a portion of their surfaces. The term "butt-coupled" is used herein in its ordinary sense to mean "end-on" or axial coupling, where there is minimal or zero axial offset between the elements in question. For example, the axial offset may be slightly greater than zero when some thin intervening layer is formed between the elements, such as a thin coating layer that is typically used to provide a highly reflective or anti-reflective function. Note that the axes of two waveguide structures or elements do not have to be collinear for them to be accurately described as butt-coupled. In other words, the interface between the elements does not have to be perpendicular to either axis, for example in the case of an angled interface. No adiabatic changes occur between butt-coupled structures.

[0019] The terms "active device", "active structure" or other "active" elements, parts, components may be used herein. A device or part of a device that is called active can generate, amplify, modulate and / or detect light using electrical contacts. This is in contrast to what we mean by "passive device", "passive structure" or other "passive" elements, whose primary function is to confine and direct light and / or provide splitting, combining, filtering and / or other functions commonly associated with passive devices. Some passive devices can provide functions that overlap with the functionality of active devices, such as, for example, phase tuning implemented using thermal effects that can provide modulation. An absolute distinction between "active" and "passive" based purely on material composition or device structure should not be assumed. For example, a silicon device can be considered active under certain modulation conditions or detection of low wavelength radiation, but is considered passive in most other circumstances.

[0020] FIG. 1 is a schematic cross-sectional end view of a heterogeneous integrated photonic device 100 optimized for operation at short wavelengths. In some embodiments, the wavelength of operation is less than 900 nm. The exemplary cross-section includes a substrate 105, which can be any substrate suitable for semiconductor and dielectric processing, such as Si, InP, GaAs, quartz, sapphire, glass, GaN, silicon-on-insulator, or other materials known in the art. In the embodiment shown, a layer of a second material 104 is deposited, grown, transferred, bonded, or otherwise attached to the top surface of the substrate 105 using techniques known in the art. The main purpose of layer 104 is to provide an optical cladding to material 102 (described below) and to form an optical waveguide, if necessary. Optical waveguides are commonly achieved by confining an optical wave by placing a higher refractive index core between two lower refractive index layers. In some embodiments, layer 104 is omitted and the substrate 105 itself serves as the cladding.

[0021] Layer 102 is deposited, grown, transferred, adhered, or otherwise attached on top of layer 104, if present, and / or on top of substrate 105 using techniques known in the art. The refractive index of layer 102 is higher than the refractive index of layer 104, if present, or higher than the refractive index of substrate 105, if layer 104 is not present. In one embodiment, the material of layer 102 is selected from the group consisting of SiN, SiNO x , TiO2, Ta2O5, (doped)SiO2, LiNbO3, and AlN. In some embodiments, other common dielectric materials may be used for layer 102. In other embodiments, high bandgap semiconductor materials may be used for layer 102. High bandgap semiconductors may be defined by bandgap energy greater than 1.5 eV (or bandgap wavelength shorter than 830 nm). Examples of high bandgap semiconductor materials are GaN, AlGaN, AlGaAs (with Al concentration >5%), etc., and are typically characterized by bandgap wavelengths shorter than 700 nm. Either or both of layers 104 and 102 may be patterned, etched, or redeposited to tailor their functionality (defining a waveguide, splitter, coupler, grating, or other passive component) as is common in the art (not shown in FIG. 1). Layer 102 corresponds to what is referred to as a passive layer or passive structure in the remainder of this disclosure, including the claims.

[0022] The refractive index of layer 108 is lower than that of layer 102, and layer 108 is above layer 102 and below layer 101 and layer 403 (not visible in FIG. 1, but described below with reference to FIG. 4). Layer 108 serves to planarize the patterned surface of layer 102, as further shown in FIG. 4. In some embodiments, the planarity of the top surface of layer 108 is provided by chemical mechanical polishing (CMP) or other etching, chemical and / or mechanical polishing methods. In other embodiments, for example when the material of layer 108 is spin-on glass, polymer, photoresist or other suitable material, the planarity is provided by the intrinsic nature of the method by which layer 108 is deposited. Planarization may be controlled to leave a desired, typically very low thickness layer on top of layer 102 (as shown in FIG. 1) or to remove all material above the level of the top surface of layer 102 (not shown). Where layer 108 remains on top of layer 102, the target thickness is in the range of 10 nm to hundreds of nm, with the actual thickness being between zero and hundreds of nm greater than the target thickness due to non-uniformity of the planarization process. In some embodiments, a spin-on material is used to planarize and then etched back, resulting in improved overall uniformity across the wafer compared to typical CMP processes. The top surface of layer 108 is typically characterized by a low surface roughness that makes it suitable for bonding. In some cases, the roughness is <5 nm RMS, and in yet other embodiments, the roughness is <1 nm RMS. It is generally known that the bonding yield improves as the surface roughness is reduced. A good CMP process can achieve a surface roughness of less than 0.5 nm RMS or lower (better).

[0023] Layer 101 is bonded on top of at least a portion of the top surface of the corresponding passive waveguide or passive waveguide structure (108, 102). The bond can be a direct molecular bond or an additional material, such as a polymer film, can be used to promote the bond, as known in the art. The bonding material, if used, must have reasonably low loss at the wavelength of operation since the optical mode 150 has some overlap with layer 108. Layer 101 constitutes what is generally referred to as an active device and is made of materials including, but not limited to, GaAs and InP based ternary and quaternary materials. Layer 101 is multilayered in some embodiments and as shown in FIG. 1, and includes sublayers that provide optical and electrical confinement and electrical contacts, as known in the art for active devices. In the embodiment shown in FIG. 1, layer 101 includes five individual sublayers, 101-1, 101-2, 101-3, 101-4, and 101-5, all of which can include additional sublayers. Sublayers of layer 101 in some embodiments provide vertical confinement (top / bottom in FIG. 1), while lateral confinement (left / right in FIG. 1) is provided by at least one etch as known in the art for active devices and is explained in more detail with the aid of FIG. 3.

[0024] Layer 101 corresponds to what is referred to in the remainder of this disclosure, including the claims, as the active layer, active waveguide, or active waveguide structure.

[0025] Sublayer 101-1 includes a highly doped InGaP n-doped contact layer. In some embodiments, the thickness of the InGaP n-doped contact layer is between 50 nm and 500 nm. The doping is typically 1e18 or greater to provide good contact, but can be lower in some embodiments. The doping can be constant, piecewise constant, or graded, or a combination of any or all of these. In some embodiments, typically when the thickness of the InGaP n-doped contact layer is greater than 150 nm, the doping can be higher at the top (towards the contact) and lower at the bottom (towards the bond interface with 108 or 102). This can result in reduced internal losses. Contacts 109a and 109b are laterally offset from optical mode 150, so that they do not contribute to optical losses. In some embodiments, sublayer 101-1 also includes a superlattice realized by at least two thin layers with different properties. The superlattice can help prevent dislocation propagation or otherwise promote more robust laser performance. In some embodiments, the superlattice is realized with thin layers of InGaP and AlInP. In some embodiments, low strain is introduced into the superlattice layers to improve the dislocation filtering effect. The typical thickness of each superlattice layer is less than 10 nm, but can be up to 50 nm thick. In many embodiments, the superlattice prefers InGaP as the final layer (at the bottom) due to its excellent stability and processability. In some embodiments, the InGaP layer of sublayer 101-1 can be directly bonded to layer 102 / 108 without the need for additional thin layers.

[0026] The bandgap wavelength of InGaP with 51% Ga is about 650 nm, allowing low optical losses for lasers operating at about 660 nm or longer. With an increasing Ga content the bandgap wavelength can be lowered even further, down to 580 nm with a Ga concentration of 70%.

[0027] The thickness of the InGaP contact layer 101-1 (including the highly doped regions) is an optimization parameter because it affects the series resistance of the active device, it affects the confinement in the quantum wells (as will be explained in more detail with the aid of FIG. 2), and it affects the process control requirements when the contacts are etched open (as will be explained in more detail with the aid of FIG. 3).

[0028] Sublayers 101-2 and 101-3 define what is referred to as an active width (see also representation 330 in FIG. 3 showing width 337). In some embodiments, the active width can be the same as the mesa width (width 327 defined by patterned sublayers 101-4 and 101-5, as seen in representation 320 in FIG. 3), but in the illustrated embodiment, active width 337 is larger than mesa width 327 in order to reduce overlap of optical mode 150 with the etched sidewalls of active region 101-3. Active region 101-3 includes quantum well, quantum dot, pn junction and / or pin junction layers, along with optional separate-confinement heterostructure (SCH) layers. In the remainder of this disclosure, including the claims, the term "quantum well" is used for simplicity to include quantum dot layers, pn junction layers, pin junction layers and quantum well layers and / or combinations of two or more of them. In some embodiments, as will be explained in detail with the aid of FIG. 3, the top layer of sublayers 101-3 is InGaP and may serve as an etch stop layer to improve the uniformity of the etch and the active devices.

[0029] Optional sublayers 101-2 may include what are commonly known as n-cladding and / or n-waveguide layers. In both cases, the layers of sublayer 101-2 include materials that are optically transparent at the operating wavelength, with background doping typically smaller than that of the contact regions. In some embodiments, the doping may be 1e18 or less. In other embodiments, the doping may be constant, piecewise constant, graded, and / or some or all combinations. In some embodiments, such layers may be realized by InGaP ternary and / or AlInGaP quaternary semiconductors, both of which may be grown on a GaAs substrate with high quality. In other embodiments, n-cladding and / or n-waveguide functionality may also be provided by barrier layers and / or SCH layers in sublayer 101-3.

[0030] Sublayer 101-4 is a p-cladding layer, i.e., a cladding layer that helps reduce internal losses by controlling the overlap between optical mode 150 and metal 110, and between optical mode 150 and sublayer 101-5, which provides the p-contact functionality described below. Sublayer 101-4 may include additional layers, such as bandgap smoothing layers, etch stop layers, graded layers, etc., to provide improved performance or to facilitate more robust manufacturing as known in the art of semiconductor device design and manufacturing. Sublayer 101-4 is typically doped with a doping level between 1e16 and 2e18. The thickness of sublayer 101-4 varies depending on the wavelength of operation and the refractive index difference between the p-cladding layer and sublayer 101-3. In some embodiments, the thickness of the p-cladding layer is greater than 400 nm. Optimization of the doping and thickness balances keeping the series resistance through the p-cladding layer low while keeping the optical loss acceptably low. It should be noted that p-doped materials typically have higher optical loss than n-doped materials, and a thinner layer of p-cladding allows for more overlap of modes with the p-contact layer 101-5, described below. The doping of the sublayer 101-4 can be constant, piecewise constant, graded, and / or a combination of all of these. Both the optimization of the contact resistance and the optimization of the optical loss can be performed by commercially available numerical solvers or other simulation / computational techniques including analytical expressions and approximations. In some embodiments, the p-cladding can be realized by InGaP ternary and / or AlInGaP quaternary semiconductors.

[0031] Sublayer 101-5 is a p-contact layer. In some embodiments, sublayer 101-5 comprises a heavily p-doped GaAs layer. In some embodiments, the doping is greater than 5e18. In yet other embodiments, the doping is even higher, reaching 1e19 or greater. A GaAs layer (with a bandgap wavelength longer than the laser operating wavelength) may be used for this contact, since the relatively thick p-cladding keeps the overlap of mode 150 and sublayer 101-5 very small. In other embodiments, heavily p-doped (greater than 5e18) InGaP may also be used for the p-contact layer. In some embodiments, the thickness of the contact layer is between 50 nm and 250 nm.

[0032] Contacts 109a, 109b and 110 (as part of sublayers 101-1 and 101-5) provide ohmic contacts to the respective semiconductor contact layers. In some embodiments, contact 110 may be realized as a standard metal contact to heavily p-doped GaAs as is common in the GaAs semiconductor laser field. In the case of a heavily p-doped InGaP contact layer, an exemplary metallization for an ohmic contact is: 1. Ti / Pt / Au: Ti(<50nm), Pt(<100nm), Au(>100nm), annealing below 400℃, annealing time below 120 seconds It could be. Contacts 109a and 109b require special consideration due to limitations regarding the use of doped InGaP layers and typical thermal limitations of heterogeneous systems that limit activation / rapid thermal annealing temperatures to less than 450° C. or lower. The choice of metal stack and annealing conditions are the two most important factors. Some examples of InGaP ohmic contacts with activation temperatures less than 400° C. that are suitable for the present invention are as follows: 2. Pd / Ge / Au: Pd(<20nm), Ge(<200nm), Au(>100nm), annealing temperature <375℃, annealing time <120s 3. Ge / Ni / Ge / Au: Ge(<20nm), Ni(<50nm), Ge(<100nm), Au(>100nm), annealing below 375℃, annealing time below 120s. The numbers in parentheses suggest optimal thickness limits for certain metal layers, but it will be apparent to one skilled in the art that ohmic contacts can result from a variety of modifications. All contacts can be connected to pad metals (not shown) through vias to facilitate efficient current and / or voltage control of active devices.

[0033] The upper cladding layer 107 may be made of any material, including but not limited to polymer, SiO2, SiN, SiNO x etc. In some embodiments, the same material is used for layers 107 and 108. In some embodiments (not shown), the cladding function of layer 107 may be provided by multiple depositions and multiple materials, for example both may provide cladding and passivation of the active devices.

[0034] The dashed arrows 160 illustrate a typical current injection path through the active region from the top to the side of the device, illustrating one of the key differences between heterogeneous integration and intrinsic (monolithic) photonic platforms. As shown in Figure 1, the mode has a large overlap with at least one of the contact layers (the n-contact in the embodiment shown in Figure 1). Figure 2 provides additional details in this regard.

[0035] FIG. 2 shows some exemplary simulation results for designing an optimal heterogeneous laser structure using InGaP contact layers. Representation 200 shows the change in the confinement factor in the quantum well layer (part of sublayer 101-3), the InGaP bottom n-contact (part of sublayer 101-1), and the GaAs top p-contact (part of sublayer 101-5) as the thickness of the InGaP layer 101-1 is varied between 50 nm and 500 nm. Note that in the plot the confinement is specified as a percentage. In this exemplary laser structure, the confinement in the InGaP n-contact is about 1% at a contact thickness of 150 nm, such a thickness can provide good series resistance. This proves the importance of using low-loss, high bandgap contact materials. The impact of the p-contact is very small since the overall confinement is less than 1e-5 and close to 1e-6, which results in a small contribution to the internal loss from the GaAs contact layer. Views 220, 240 and 260 show example mode shapes (as contour lines) for example laser structures with InGaP bottom contact thicknesses of 50 nm, 150 nm and 250 nm, respectively. The impact of the n-contact thickness on the mode shape and on controlling both the internal losses and the confinement in the quantum wells is significant. Fine optimization of the optical modes to optimize the performance of the active device can be easily performed using commercially available solvers with standard material libraries.

[0036] 3 shows six views (300, 310, 320, 330, 340, and 350) corresponding to cross sections of an integrated device of the type described above after several exemplary steps in operations carried out to fabricate the device. Operations to fabricate a device need not always include all of the depicted functions, operations, or actions, nor do they need to include them in the exact order shown by the sequence of views 300 through 350.

[0037] However, in the exemplary case, the result of the work specific to a heterogeneous platform with InGaP layers will first be represented as 300. Here, suitable passive structures 302 realized on the substrate (105), bottom cladding (104), passive waveguides (102) and planarization layer (108) as shown and explained in more detail with the aid of FIG. 1 have a die or wafer of active layer 301 bonded on them, as is common in heterogeneous processes. The steps to achieve this are similar to those described, for example, in US Pat. No. 10,641,959 or other heterogeneous integration references. The bonded active layer 301 includes five sublayers 101-1 to 101-5 as explained with the aid of FIG. 1, as well as the substrate 309 on which these five sublayers are grown. There may be additional grown layers between the five sublayers and the substrate to promote better material quality (e.g. buffer layers) and / or to promote substrate removal (e.g. etch stop layers).

[0038] The operation can then proceed to the next step shown in display 310. Here, the substrate 309 is removed, leaving the active layer 311 containing only the five sublayers as described above. The removal process can utilize a combination of mechanical grinding / lapping / polishing, chemical mechanical polishing, dry etching and / or wet etching. In some embodiments, a layer of InGaP is used as a selective layer to facilitate improved process control. In some embodiments, the substrate is mechanically thinned to less than 50 μm, and then the final substrate removal is performed by selective etching, resulting in a very high uniformity of the substrate removal process. With the removal of the substrate, only a thin layer containing the five sublayers remains on the top surface of the passive structure (302) wafer. In some embodiments, the remaining layer is less than 5 μm thick, allowing the use of high quality photolithography for subsequent processing. The alignment between the passive structures defined in 302 before bonding and the active structures defined in subsequent steps is very good (less than 250 nm or better) due to the use of common lithographic alignment marks on the passive structure (302) wafer. This is one of the key advantages of heterogeneous integration in terms of device uniformity.

[0039] In this exemplary case, the process has a width W M 327) has been performed. Such etching may utilize dry and / or wet etching and may also utilize an etch stop layer to improve uniformity across the wafer. The remaining active layer 321 is not etched in this step. In some embodiments, InGaP is used as a selective etch stop layer in etching the mesa. In some embodiments, the mesa width is between 1 μm and 5 μm. In yet other, typically high power designs, the mesa width is greater than 5 μm and may be thicker up to 100 μm or more.

[0040] Next, the process is A3, the reader may proceed to view 330, which shows a cross section after etching of active portion 334 (at 337) has been performed. Such etching may utilize dry and / or wet etching, and may also utilize an etch stop layer to improve uniformity across the wafer. The remaining active layer 331 (including sublayer 101-1 as described with the aid of FIG. 1) is not etched in this step; it includes an InGaP layer to provide an etch stop, and also includes contacts as described in connection with FIGS. 1 and 2. In some embodiments, active width 337 is at least 2 μm greater than mesa width 327. In other embodiments (not shown), active width 337 may be substantially the same width as mesa width 327.

[0041] The process may then proceed to representation 340, which shows a cross section after metal contact 345 has been deposited. In some embodiments, the n-contact (bottom contact in representation 340) must be deposited after the active etch once the n-doped InGaP layer 341, corresponding to 101-1 in Figure 1, is exposed, while the deposition of the p-contact (top contact in representation 340) can be done earlier (e.g., after substrate removal in representation 310, or after mesa etch in representation 320). As previously explained in connection with Figure 1, the metallization for the p-doped and n-doped layers is different.

[0042] Finally, the process may proceed with depositing cladding 359 (as described in connection with FIG. 1 and layer 107) before opening the vias and depositing metal pads 356. The result is shown in display 350.

[0043] Further processing of various dielectric and / or semiconductor layers, and / or addition and processing of electrical contacts, vias, and index matching layers may be performed as known in the art, which may include heaters, passivation, etching trenches, forming light blocking structures, etc.

[0044] The process takes advantage of the InGaP layer to provide etch selectivity and high bandgap contacts with low optical loss. By appropriate dry etch conditions (gas selection, plasma power, etc.) or wet etchants (chemical selection), InGaP can be selectively removed from different III / V compounds (such as GaAs) or vice versa, providing a high degree of process control and uniformity.

[0045] 4 is a schematic cross-sectional view of an integrated photonic device 400 utilizing butt coupling and mode conversion for efficient coupling between dissimilar materials. Functional layers 401 to 408 correspond to functional layers 101 to 108 as described in connection with FIG. 1 (unless expressly defined differently).

[0046] Layer 408, whose refractive index is lower than that of layer 402, overlies layer 402 and underlies layers 401 and 403 (described in more detail below).

[0047] Layer 401 is bonded onto the top surface of at least a portion of a corresponding passive waveguide structure, including layer 402 and optionally including layer 408. Layer 401 constitutes what is generally referred to as an active device, and includes sublayer 401a (including sublayers 101-2, 101-3, 101-4 and 101-5 as described with the aid of FIG. 1) and sublayer 401b, including sublayer 101-1 as described with the aid of FIG. 1. 401b includes a heavily doped InGaP layer used to provide contacts, while being optically transparent at the wavelength of operation. The overlap of optical mode 450 (guided inside the active device) with sublayer 401b is large (as also shown in FIG. 2), and in some embodiments, the evanescent tail of the optical mode extends to layer 402 / 408. Thus, the periodic structure in 402 / 408 may provide optical feedback for realizing a distributed feedback (DFB) laser or the like.

[0048] Efficient coupling of optical mode 450 supported in active structure 401 to optical mode 453 supported in the passive structure for which layer 402 provides the core is facilitated by layer 403 and, if present, layer 406. Optional layer 406 acts primarily as either an anti-reflection coating or a highly reflective coating at the interface between layers 401 and 403. Layer 403, in some embodiments, acts as an intermediate waveguide that accepts the profile of the optical mode supported in the waveguide for which layer 401 provides the core (illustrated by line 450), efficiently captures it as mode profile 451, and gradually transmits it to mode profile 452 and finally to 453. Mode profile 453 is efficiently coupled to the waveguide for which layer 402 provides the core.

[0049] The use of intermediate layer 103 is advantageous in that it can be made of a material with a very high refractive index (such as GaAs / InP and its ternary and quaternary systems) that supports operation in the short wavelength range, e.g., SiN, SiNO x This greatly improves efficient transmission between the low refractive index materials in layer 402, such as TiO2, Ta2O5, (doped)SiO2, LiNbO3 and AlN.

[0050] Differences between the optical modes supported by the waveguides in layers 401 and 402, respectively, may or may not be evident from observation of the mode profiles, but less than 100% mode shape overlap along with a non-zero vertical offset (in FIG. 4) between modes 450 and 453 may result in significant optical loss (in the absence of intermediate layer 403). In some embodiments, losses up to 2 dB may be deemed acceptable, but losses greater than that may not be. In other cases, a loss level of 5 dB may be the selected criterion for acceptability. The function of layer 403 is to keep the optical coupling loss due to incomplete mode overlap and vertical offset (between modes 450 and 453) below whatever is determined to be an acceptable level in a given application.

[0051] In some embodiments, layer 408 is not present and layers 401 and 403 are both disposed on top of patterned layer 402, 401 by adhesion and 403 by various deposition techniques. In such embodiments, there is no planarization step.

[0052] It will be understood that these exemplary embodiments teach only a few examples of heterogeneously integrated lasers and active components operating at short wavelengths utilizing the present invention, and that many other similar arrangements can be envisioned. Furthermore, such lasers and active components can be combined with multiple other components, such as various filtering elements, amplifiers, monitor photodiodes, modulators, and / or other photonic components to provide additional functionality or better performance.

[0053] Embodiments of the present invention provide many advantages. The integrated platform enables scalable fabrication of PICs made from multiple materials that provide higher performance and / or the ability to operate over a broad wavelength range. Additionally, the platform can handle higher optical powers compared to typical Si- or InP-waveguide-based PICs.

[0054] The present invention utilizes a process flow that typically consists of wafer bonding of components of compound semiconductor material on a carrier wafer with dielectric waveguides (as described with the aid of Figures 1, 3 and 4) followed by semiconductor manufacturing processes as known within the art. This allows precise definition of the optical alignment between active and passive waveguides, typically via a lithography step, eliminating the need for precise physical alignment. Said lithography-based alignment allows for scalable manufacturing using wafer-scale technologies.

[0055] It will be appreciated that optical coupling between modes in the active and passive layers is reciprocal, so that, taking Figure 4 as an example, a structure may be configured to facilitate the conduction of light from region 401 to region 402, but it may also facilitate conduction in the reverse direction, i.e., from region 402 to region 401. It will be appreciated that multiple such transitions, with no limit on the number or direction, may be realized on an appropriately configured PIC.

[0056] Other approaches have relied on die attachment of fabricated optically active devices to passive waveguides, which requires very tight alignment precision that typically exceeds what a typical die bonder can provide, an aspect that limits the throughput and optical coupling performance of the process.

[0057] In some embodiments, the active devices create hybrid waveguide structures with dielectric layers that can be used to create wavelength selective components, e.g., formed inside a laser cavity, e.g., for distributed feedback (DFB) lasers or similar components.

[0058] The optical device embodiments described herein may be incorporated into a variety of other devices and systems, including, but not limited to, various computing and / or consumer electronic devices / appliances, communication systems, computing systems, medical devices, sensors and sensing systems.

[0059] It will be understood that this disclosure teaches only some examples of illustrative embodiments, and that many variations of the invention can be readily devised by those skilled in the art after reading this disclosure, and that the scope of the invention is to be determined by the claims which follow. [Explanation of symbols]

[0060] 100 Heterogeneous Integrated Photonic Devices 101 layers 101-1 Sublayer, InGaP layer 101-2 Sublayer 101-3 Sublayer 101-4 Sublayer 101-5 Sublayer 102 Passive Waveguide 103 Middle Class 104 Bottom Cladding 105 Substrate 107 Upper cladding layer 108 Planarization layer 108 Passive Waveguide Structure 109a Contact 109b Contact 110 Contact 150 Light Mode 160 Current injection path 301 Active 302 Passive Structure 309 Substrate 323 Mesa 327, W M Mesa Width 334 Active Division 337, W A Active Width 345 Metal Contact 356 Metal Pad 359 Clad 400 Integrated Photonic Devices 401 layers 401a Sublayer 401b Sublayer 402 Patterned Layer 403 Middle Class 406 layers 408 layers 450, 451, 452, 453 mode profiles

Claims

1. 1. A heterogeneous device, comprising: a passive waveguide structure attached to a substrate, the passive waveguide structure including a dielectric layer; an active waveguide structure attached to a top surface of the passive waveguide structure, the active waveguide structure including a quantum well layer overlying an InGaP layer; A heterogeneous device, wherein the InGaP layer provides an n-contact function.

2. The heterogeneous device of claim 1 , wherein the InGaP layer is characterized by a thickness between 50 nm and 500 nm.

3. The heterogeneous device of claim 1 , wherein the InGaP layer is characterized by a doping concentration of 1e18 or greater.

4. 10. The heterogeneous device of claim 1, wherein the InGaP layer serves as an adhesion layer between a bottom surface of the active waveguide structure and a top surface of the passive waveguide structure.

5. the InGaP layer is a layer in a superlattice; The heterogeneous device of claim 1 , wherein the superlattice comprises a bonding interface of the active waveguide structure.

6. The passive waveguide structure is made of SiN, SiNO x , TiO 2 , Ta 2 O 5 , (doped) SiO 2 , LiNbO 3 10. The heterogeneous device of claim 1 , further comprising a dielectric layer comprising at least one of: SiN; and AlN.

7. The heterogeneous device of claim 6 , wherein the dielectric layer is characterized by a thickness between 50 nm and 600 nm.

8. 10. The heterogeneous device of claim 1, wherein metallization for contacting the InGaP layer comprises at least one of Pd, Ge, Au, and Ni.

9. the active waveguide structure includes a p-cladding layer overlying the quantum well layer; The heterogeneous device of claim 1 , wherein the p-cladding layer is characterized by a thickness greater than 400 nm.

10. 10. The heterogeneous device of claim 1, wherein a top surface of the passive waveguide structure to which the active waveguide structure is adhered is characterized by a roughness of less than 1 nm RMS.

11. The active waveguide structure has a width of 1 μm≦W M Width W ≦5 μm M 10. The heterogeneous device of claim 1 having a cross-sectional shape characterized by a mesa.

12. The cross-sectional shape is W A =W M +2 μm active width W A 12. The heterogeneous device of claim 11 further characterized by:

13. The active waveguide structure has a width of 5 μm<W M Width W M 10. The heterogeneous device of claim 1 having a cross-sectional shape characterized by a mesa.

14. 10. The heterogeneous device of claim 1, wherein the active waveguide structure includes a second InGaP layer overlying the quantum well layer and acting as an etch stop layer.

15. 1. A heterogeneous device, comprising: a passive waveguide structure attached to a substrate, the passive waveguide structure including a semiconductor layer; an active waveguide structure attached to a top surface of a semiconductor waveguide structure, the active waveguide structure including a quantum well layer overlying an InGaP layer; A heterogeneous device, wherein the InGaP layer provides an n-contact function.

16. 16. The heterogeneous device of claim 15, wherein the semiconductor waveguide structure comprises a semiconductor layer having a bandgap of 1.5 eV or greater.

17. 16. The heterogeneous device of claim 15, wherein the InGaP layer is characterized by a thickness between 50 nm and 500 nm.

18. 16. The heterogeneous device of claim 15, wherein the InGaP layer is characterized by a doping concentration of 1e18 or greater.

19. 16. The heterogeneous device of claim 15, wherein the InGaP layer serves as an adhesion layer between a bottom surface of the active waveguide structure and a top surface of a dielectric waveguide structure.

20. the InGaP layer is a layer in a superlattice; The heterogeneous device of claim 15 , wherein the superlattice comprises a bonding interface of the active waveguide structure.

Citation Information

Patent Citations

  • Semiconductor laser device

    JP2015179783A

  • Semiconductor device and manufacturing method of the same

    JP2019003973A

  • Integrated active devices with improved optical coupling to dielectric waveguides.

    JP2021513667A

  • Optical temperature measurement in photonic circuits

    JP2022060987A

  • Integrated GAAS active device with improved optical coupling to dielectric waveguides

    JP2023121105A