Open Access
Add to BibSonomy
Abstract
The efficient investigation of integrated photonic and superconducting circuitry relies on the flexible design of masks for the lithography steps. While commercial software packages for high-end design are commonly employed, for scientific users, such software systems are often financially exclusive and thus custom solutions are needed. Here we present a flexible open source Python framework that allows mask generation of integrated circuitry by easy-to-learn Python scripting. The framework is designed to facilitate the design of new photonic building blocks, since it allows defining the geometry by reusing existing parts or direct design using geometric objects. Through the use of existing and user-defined building blocks, complex integrated circuits can be created in a convenient fashion. We illustrate the capabilities of the framework by realizing hybrid nanophotonic-superconducting circuits, as well as hybrid 2D-3D nanophotonic circuits through multi-step nanofabrication. Because all design parameters can be defined by the user, the framework is not limited to a particular platform and can rapidly be adapted for new applications.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Integrated photonics and Silicon photonics in particular has been a developing field for the last decades. Photonic circuits allow the realization of complex optical functionality in a chip-scale framework and hold promise for overcoming limitations of their complementary metal oxide semiconductor (CMOS) electronic counterparts in specific applications. Although significant progress has been made in the realization of photonic circuitry in the past years, the photonic community is still lagging behind in terms of complexity and thus would benefit from the electronic design automation (EDA) tools which are available to the CMOS electronics community. In consequence, powerful tools have been developed to push silicon photonic circuits out of small university labs with the goal to make the platform more accessible to a greater target audience. Through alliances like Jeppix [1] or ePIXfab [2], so called Multi-Project Wafer (MPW) runs have become accessible, which allow the user to design and build their own photonic integrated circuits (PICs) without the need to own fabrication facilities. To assist the design-process for such “fab-less” users, institutions which offer MPW runs have developed Process-Design-Kits (PDKs), which are software packages containing the layout for pre-designed and already characterized photonic devices. The users can then create their own photonic circuits by combining the devices in the PDK and even simulate the photonic performance to some extent.
However, being a “fab-less” user comes with the downside of being dependent on the devices in the PDK and, more importantly, on the processes provided by the respective foundry. Testing new material platforms or implementing additional process steps is therefore challenging, because doing so would directly influence the designs of the other users on an MPW run. In addition, MPW runs can be costly and exhibit delivery times of the fabricated circuits on chip of several months. Hence, especially smaller university groups rely on either their own electron-beam (Ebeam) or optical lithography capabilities, or on cooperation partners. For such groups, using the above-mentioned software for PICs is not appropriate, since they usually develop completely new devices and new processes tailored to the possibilities of in-house fabrication.
Although Software solutions like the Synopsys PIC Design Suite [3] allow for the scripting of new devices in addition to pre-defined PDKs, these design environments need to be licensed and therefore are often costly as well. To circumvent these issues, in 2012 researchers developed IPKISS [4], a parametric design and simulation tool based on the free and easy to learn computer language Python [5]. This framework was put under general public license (GPL) and is therefore accessible free of charge. However, further development of the freely accessible package was discontinued around 2013, being only maintained in part within the commercially available software suite Luceda Photonics [6].
In this work, we present an alternative open source solution, GDSHelpers [7], a Python based, easy-to-learn software package, which allows the user to quickly generate arbitrary photonic and other devices. In addition to that, we provide an online documentation of our package [8] with a tutorial, which explains the usage of the package to the users. Python was chosen as the programming language for GDSHelpers because it is widely recognized as a convenient and straight-forward, yet very powerful, computer language. Because mask definition for lithography typically consists of implementing series of polygons describing the outer borders of the structures of each layer, the focus of this work was put on polygon generation. Such operations have traditionally been investigated in the context of computational geometry and efficient computer libraries already exist [9,10]. We therefore leverage the GEOS library [11] via the Shapely Python library [12] for performing geometric operations like calculating rotated version of shapes or unifying overlapping geometric objects. For converting the generated mask to a universal GDSII-/OASIS-file, which are two of the standardized formats for mask definition and often used for electron beam lithography (EBL), we employ gdspy [13] / gdscad [14] / fatamorgana [15] and provide a converter for shapely objects. Because our focus is the design of integrated photonic devices with parametrized properties, GDSHelpers is a purely script-based framework without graphical user interface (GUI). Because we use Python, the open nature of this language allows for easy integration of third-party tools like Blender [16] or MEEP [17] into our GDSHelpers framework.
In this paper, we apply GDSHelpers for the design of exemplary classes of photonic library elements and demonstrate parameterized photonic components in hybrid systems. In particular, we showcase the co-integration of nanophotonic with superconducting devices which relies on multiple fabrication steps which necessitate the aligned exposure of different patterns. Furthermore, we demonstrate co-integration of planar photonic devices with 3D nanostructures. We employ superconducting devices for realizing waveguide integrated single photon detectors and related devices, which are often connected to electrical contact pads. Since 3D nanoprinting has become mature, 3D elements are now being co-integrated within traditional nanophotonic circuits to provide free-form shaping and thus optical functionality which is challenging to implement in a planar fashion. Within GDSHelpers we provide both the connection points to planar circuits and the alignment structures needed for carrying out additive manufacturing.
The GDSHelpers framework is organized into sub-packages. The largest one is the parts sub-package, which contains the photonic, electronic, superconducting and combined components and the classes needed for the generation of alignment structures, text and optical codes. The sub-package layout allows for organizing multiple devices in a convenient way, most notably for parameter sweeps. It further allows the user to see the exposure fields for e-beam lithography in order to avoid stitching marks at critical structures. The helpers sub-package collects several smaller functions, such as the conversion of the pattern for positive resist exposure or the automatized distribution of holes for under etching. Using the sub-package geometry, the layout can be converted to a GDSII-file for mask generation or electron beam exposure (EBL). The export sub-package allows to export the structure to a blender-file and also to directly render the structure to a 3D-image. Together, the framework provides a rich suite of design tools for flexibly creating advanced photonic circuitry.
2. Pattern generation
In GDSHelpers each layer of a device is described by an object from the Shapely library, which is either a Shapely polygon, e.g. a list of the coordinates of the shell and lists of the coordinates of the holes inside the polygon, or a Multi-polygon, which is a collection of several polygons. This approach allows the definition of arbitrary shapes by either directly specifying the coordinates of the shells and holes of the polygons, or by combining previously defined building blocks. For facilitating the generation of these shapes, we have defined various building blocks in the sub-package parts, which allow convenient parameter-sharing between the users of these devices.
The most fundamental and widely used part in the design of photonic circuits with GDSHelpers is the waveguide. Waveguides are used to interconnect photonic library elements on chip and thus perform the optical analogous function of an electrical wire in CMOS circuits. A waveguide starts at a given position with a certain angle and has a defined width. This combination of position, angle and width, in the following is called a port, as it is usually the connection point between different photonic parts of the design. Even though the initial intention of the design of such a waveguide was a nanophotonic circuitry, this library element is not limited to optical applications, but has also proven to be useful for realizing e.g. electrical connections or for the generation of features for other devices.
After the creation of the waveguide, various segments can be attached (see Fig. 1), e.g. a simple straight segment, a bend or a more advanced Bezier curve, allowing for straight-forward and convenient design of smooth connections between different components in the photonic circuit. This design framework allows to automatically route individual segments to other ports and thus to address arbitrary connection points within a large-scale layout.
Fig. 1. Composition of a waveguide, which is generated at a certain port (red arrow). The waveguide consists of four segments which are added successively (connecting ports are represented by green arrows). Finally, the ending port (blue arrow) of the waveguide is extracted, allowing to attach another photonic building block at the end.
Besides basic segments, a fully parametrized curve is also possible, allowing for free design of the path of the waveguide as well as the width of the waveguide. An example device using this possibility for generating a long waveguide with a small footprint is a waveguide spiral, which can be used e.g. as a delay line or for spontaneous four wave mixing [18] in a nanophotonic circuit.
Using this interconnection strategy, photonic circuits can be created from individual elements as illustrated in Fig. 2. Different building blocks are placed at user defined positions or at certain ports of previously defined photonic components using the make_at_port method. After placement within a larger layout, they are subsequently connected automatically with waveguides (see Fig. 2), which allows for a fast design of integrated circuitry. The automatic calculation of the waveguide trajectory facilitates parameter-sweeps, since the routing approach adapts to a changing building block size. Furthermore, the implementation allows to control the waveguide width along the trajectory using a user-defined function.
Fig. 2. Concept of our design framework. (a) Individual components (surrounded by rectangles), here a ring-resonator, a y-splitter, and four grating couplers, can be placed at user-defined positions in a conceptual way. (b) Subsequently the ports (represented in (a) by circles at the edges of the parts) of the devices are connected automatically by waveguides e.g. Bezier-curves. GDSHelpers also provides capability for labelling devices as shown in the (b).
In the following we present frequently used nanophotonic building blocks which are already implemented in GDSHelpers and also introduce elementary superconducting building blocks.
2.1 Power splitters
The basic design framework implements three different types of waveguide splitters. A convenient and frequently used solution is the Y-Splitter, consisting of two different waveguides which are bent in opposite directions, starting from the same port. The device splits the incoming light adiabatically into two outgoing waveguides. This splitter is usable for designing both nanophotonic and superconducting circuitry.
Alternatively, multimode-interference-couplers (MMI couplers) [19] are implemented, also allowing for power splitting in equal parts. In comparison to the Y-splitters these devices show higher tolerance to fabrication errors, as they do not depend on the exact fabrication of a small gap and therefore are also better suited for higher refractive index platforms like SOI. For flexibility, in GDSHelpers the MMI coupler supports a freely chosen number of input and output ports and optionally calculates the appropriate taper positions. These elements are shown in Fig. 3.
Fig. 3. Shapes of various nanophotonic splitters implemented in GDSHelpers. (a) Y-splitter. (b) 2 × 2 Multi-mode-interference-splitter (MMI). (c) Directional coupler.
For obtaining a tailored splitting ratio in nanophotonic circuitry directional couplers are implemented. These devices consist of two parallel waveguides, which are bent away from each other before and after the coupling region in order to achieve a specifically designed splitting ratio. This way the splitting ratio can be chosen freely by varying the coupling length or the size of the coupling gap continuously.
As an application example, using the aforementioned splitters, on-chip interferometers can be realized. Exemplary Mach-Zehnder-interferometers (MZI) based on Y-splitters and MMI couplers are therefore included in the framework. MZI devices can be employed e.g. for modulating an input signal by placing a controllable phase shifter in one arm, for filtering optical signals or for precise measurements of the insertion loss of an on-chip optical device-under-test inserted in one arm of the interferometer.
2.2 Fiber-to-chip couplers
Coupling light from an optical fiber to an integrated waveguide is one of the fundamental challenges for integrated nanophotonic circuits. Therefore, focusing grating couplers [20–22] are implemented as an elementary building block in GDSHelpers, allowing for the scattering of light from the chip out-of-plane, where single fibers or a fiber-array can be positioned. Within the library different coupling geometries can be chosen depending on the given material system or the wavelength of interest. The possibility to freely choose the design parameters allows for easy adaption to a desired photonic platform and wavelength range.
As an alternative to the coupling via a grating, edge couplers can be easily realized with GDSHelpers. The possibility to change the width of the waveguide continuously along the direction of propagation enables convenient implementation of inverse tapers, which are commonly used for edge-coupling to fibers [23]. These waveguide tapers can be attached to planar waveguides using the port scheme outlined above. In addition, they provide convenient connection points to 3D photonic components as described in the following.
2.3 Integration of planar circuits with 3D-photonic components
Even though planar nanophotonic circuitry offers a widespread range of optical components for system design, the inherent restriction to two-dimensional shapes (or 2.5-dimensional by using multiple layers) limits the possibilities for some applications. In particular, waveguide intersections and optical vias are difficult to implement with planar structures. 3D direct laser writing (DLW) provides the means to overcome these limitations. Since waveguiding nanostructures written by DLW are compatible in size to planar waveguides, efficient optical interfaces between both technologies can be created.
To facilitate the integration of 3D nanophotonic components with pre-fabricated planar waveguide devices, the GDSHelpers library offers the possibility to add tapered waveguides at ports, for which also markers are generated automatically or can be set by hand. These structures provide transition ports to interconnect planar waveguides with 3D waveguides which then provide the means to optically address 3D photonic components. The alignment markers on the planar substrate allow the detection of the underlying coordinate system by computer-vision. Furthermore, arbitrary points can also be declared as ports for 3D-hybrid integration in order to allow vertical or angled connection to 3D building blocks,.
When using hybrid integration, in addition to the GDSII-file, a specification file will be created which contains the positions of the alignment markers and the structures, as well as additional information about the interconnection points. These synchronized files allow for automatized fabrication of hybrid-structures, such as three-dimensional-polymer-couplers [24,25] and other three-dimensional structures, including inter-chip-photonic-wire-bonds [26,27].
2.4 Nanophotonic cavities
In order to enable resonant enhancement on chip or to realize strong light matter interaction, photonic cavities are a method of choice. Currently, in GDSHelpers we have implemented two types of cavities, namely one-dimensional photonic crystal cavities and ring resonator devices.
In the library class based on one-dimensional photonic crystals, a waveguide with holes at periodic distances is implemented. Because the distance between the holes in the cavity center can be varied, a defect state can be introduced, which makes the cavity useful for efficient coupling to single photon sources or for applications as a narrowband filter. For optimizing the geometry, e.g. for achieving higher quality-factors, this feature allows fine-tuning of parameters such as the hole-tapering percentage or the waveguide-shape.
In addition to photonic nanobeam cavities, integrated ring resonators can be attached easily to waveguides, as shown in the photonic circuit in Fig. 2. These devices can be used for nonlinear experiments necessitating field enhancement, including four-wave-mixing [28] or parametric-down-conversion [29], or as waveguide coupled optical filters for certain wavelengths. When a second bus waveguide is attached to the ring, the ring resonator geometry can also be used as an add-drop-filter.
2.5 Superconducting devices
Besides purely photonic devices, also library elements for hybrid photonic-superconducting circuits and electronic devices are included in the design framework. Currently, the library comprises designs for three different devices, including single photon detectors [30–32] in the form of a nanowire placed on top of a photonic waveguide, structural elements for electrically driven single photon emitters such as carbon-nano-tubes [33,34] and the purely electronic nanocryotron (nTron) [35]. The nTron is a current sensitive, nanowire-based superconducting electrothermal device, which allows suppression of superconductivity in the nanowire by applying an additional current to a connected bottleneck. Using this current-control mechanism it can be operated as a comparator and thus allow to realize logical elements or digital amplifiers for detector signals. These components are connected to on-chip electrodes and thus provide electrical access to the circuits in addition to optical access via coupling structures.
Our fabrication process of these multi-layer structures relies on multi-step EBL. Thus, alignment structures are typically combined with the layout. The alignment markers are placed in close vicinity to the fabrication location for accurate positioning during EBL. This way each device can be precisely addressed and fabricated with high precision. The library supports different types of marker structures and electronic connectors for superconducting devices.
2.6 Additional features and support capabilities
For the fabrication of the nanophotonic/-electrical devices not only negative-tone-resist (“write-what-should-stay”) is used, but also positive tone resist (“write-what-should-go-away”) is often employed for mask-fabrication. Yet, the design of library elements for negative-tone-resist is often more convenient. Because with positive resist the desired pattern needs to be inverted, the GDSHelpers framework offers the automatic function convert_to_positive_resist, which inverts the structure, while keeping the written area as small as possible for short writing times by only applying a thin buffer around the structures (see Fig. 4). This is achieved by first inflating the contours of the original structure, followed by a polygon simplification step combined with polygon subtraction of the negative resist shapes.
Fig. 4. Conversion of a negative-tone shape for use with positive resist. The original shape of the grating coupler is inflated, to define the outer borders of the exposed area. For saving computational time and mask file size, the created shape is simplified. Finally, the original shape is subtracted and the pattern for positive resist remains.
For realizing free-standing devices which require wet under-etching after lithography and nanostructuring, the function create_holes_for_under_etching is included. This library function generates an array of openings for wet release at a desired distance from the nanostructures. This approach can be used for allowing wet etchants to flow through the openings under the waveguide structures, allowing for easy design of substrate-detached circuitry.
For post-fabrication navigating on the chip using optical imaging, two support possibilities are provided by the library. Simple text can be added to the layout, allowing for fast identification of each on-chip device by the researcher or a visual inspection system. In addition, also the generation of QR-codes and different types of alignment markers are included in the library, allowing for identification of the devices by computer-vision. In particular this feature facilitates automatized characterization of the devices after nanofabrication.
3. Pattern export
By combining the nanostructures outlined above into full layouts with many circuit elements, the final shape for each lithography layer can be created and internally represented by a Shapely object. Using the function convert_to_layout_objs, this object can be converted to a gdspy-/gdsCAD/fatamorgana-shape, which can be further processed using the gdspy-/gdsCAD/fatamorgana-library. This way the whole layout is finally converted to the GDSII/OASIS-format, which are standard formats used for electron-beam-lithography. Due to GDSHelpers’ inherent nature of utilizing abstraction layers, the user does not directly interact with the above-mentioned libraries, instead just has to choose the output file type (GDSII / OASIS). The software then conveniently takes care of the differences in the underlying programming interfaces. Because this function can in principal convert arbitrary Shapely-objects, the converter is not limited to nanophotonic/superconducting applications but can be used for other research applications which utilize the GDSII/OASIS-format. Due to the open-source-nature of the library allowing the addition of new building blocks, the GDSHelpers library can be easily extended for additional fields of research.
Because it is often desirable to obtain 3D rendered images of the fabricated devices, the GDSHelpers library offers a function for export to the 3D modeler Blender [16], which converts the Shapely-object to an internal blender object. This support function allows for convenient three-dimensional rendering of the integrated structures.
4. Full chip layouts
Using GDSHelpers, complex devices and layouts can be created in a convenient way by assembling multiple parts. Because usually many such devices are integrated on the same chip, the GDSHelpers framework allows to arrange them conveniently in a layout, e.g. for parameter scans.
Exemples of fabricated devices which were designed using GDSHelpers are shown in Fig. 5. The micrographs show a) waveguide integrated superconducting nanowire single photon detectors (SNSPDs). In the lower part of the picture the photonic circuitry consisting of waveguides, MMI-splitters and grating couplers is shown. In the upper part of the picture gold contacts pads are visible. These are connected to a superconducting nanowire on top of the waveguide; In Fig. 5b) nTrons and tunable MZIs are shown. In the middle the markers which were used for precision alignment of the gold-layer to the superconductor-layer are shown. At the bottom part a photonic alignment circuit consisting of two grating couplers connected by a waveguide is visible. The image in Fig. 5c) shows a micromechanical phase shifter. To the upper side it is connected to gold contact pads. In the lower area the photonic circuitry is shown, which embeds the phase shifter in an MZI for precision measurements of phase shift and loss. In Fig. 5d) an electronically controlled thermo-optic phase shifter is presented. Above these, gold circuits were fabricated for individual heating of the spirals for precision tuning of the transmission spectrum of the MZI. Because in each device different materials were used for waveguides, electrodes and detectors, the layouts consist of multiple layers. The alignment for the successive electron-beam or optical exposures steps is done using markers. If positioning accuracy in the range of a micrometer is satisfactory, only global markers at the edges of the pattern are used. This is the case for the presented tunable MZI in Fig. 5d, where the dimensions of both the spiral waveguides and the gold circuitry are rather big and have therefore relaxed alignment demands.
Fig. 5. Micrographs of various fabricated devices designed using GDSHelpers. (a) SEM-Micrograph of an SNSPD. (b) In the upper part of the picture four gold contacts connected to an nTron are visible. (c) An integrated micromechanical phase shifter is shown in the center region of the picture. (d) A photonic circuit consisting of two MZIs with each containing a spiral in each arm.
For the other shown devices, namely an SNSPD in Fig. 5a, an nTron in Fig. 5b and a micromechanical phase shifter in Fig. 5c, high positioning accuracy is critical. Therefore, for these devices not only the global markers were used, but also local markers (visible in the pictures as small gold rectangles) for precision alignment for each device.
Besides combining nanophotonic, electronic and superconducting circuitry, the GDSHelpers framework also allows to attach inverse tapers to waveguides and to add markers for optical computer-vision to the chip. Using these structures, 3D photonic components can be added to a planar layout after nanostructuring. All necessary data for a subsequent DLW-step is automatically written into a specification file during creation of the layout. This file allows to navigate on the chip using computer-vision as well as calculating the position of the inverse tapers in order to write structures at these positions.
In Fig. 6 two examples for this hybrid integration scheme are shown. In Fig. 6a 3D-polymer out-of-plane coupling structures [25] were attached to the previously fabricated silicon nitride waveguides. In Fig. 6b 3D-polymer waveguides connecting two inverse tapers are depicted. In both cases the 3D-polymer components were aligned using computer vision and the specification file generated by GDSHelpers. Arbitrary components can thus be designed in 3D and subsequently be integrated additively to an existing chip.
Fig. 6. SEM-micrographs of nanophotonic circuitry with precision aligned 3D-polymer-structures. (a) Several devices which use 3D-polymer structures in order to couple light from a waveguide into the fiber and vice versa. (b) Nanophotonic circuitry integrated with 3D-polymer connections between planar waveguides.
5. Conclusion
We presented an open source design framework for integrated nanophotonic/superconducting structures, which allows rapid creation of complete integrated circuitry. The framework is implemented in the easy-to-learn programming language Python such that new users are able to get familiar with the framework in a short amount of time. Because the toolkit is open source, programming of new devices is readily accommodated, because already existing devices can be used as template code. Furthermore, the GDSHelpers library allows collaborative development, because other users of the framework can contribute their building blocks to the main framework. We anticipate that this approach will lead to easy parameter-sharing between different research groups, because the implemented parts will become experimentally reproducible this way. Over time, we hope to support a growing user base and therefore also diversify the available building blocks implemented in this framework.
Funding
Horizon 2020 Framework Programme (780848); H2020 European Research Council (724707); Deutsche Forschungsgemeinschaft (PE 1832/6-1, 1832/6-2).
Acknowledgements
This research was supported via the European Union's Horizon 2020 research and innovation program (780848, Fun-COMP project), the European Research Council (724707, PINQS project) and the Deutsche Forschungsgemeinschaft (DFG) (project PE 1832/6-2). We acknowledge support from the Open Access Publication Fund of the University of Muenster.
References
1. X. Leijtens, “JePPIX: the platform for Indium Phosphide-based photonics,” IET Optoelectron. 5(5), 202–206 (2011). [CrossRef]
2. P. Dumon, W. Bogaerts, R. Baets, J.-M. Fedeli, and L. Fulbert, “Towards foundry approach for silicon photonics: silicon photonics platform ePIXfab,” Electron. Lett. 45(12), 581 (2009). [CrossRef]
3. I. Synopsys, “PIC Design Suite,” https://www.synopsys.com/photonic-solutions/pic-design-suite.html.
4. W. Bogaerts, P. Dumon, E. Lambert, M. Fiers, S. Pathak, and A. Ribeiro, “IPKISS: A parametric design and simulation framework for silicon photonics,” 9th Int. Conf. Gr. IV Photonics2, 30–32 (2012). [CrossRef]
5. Python Software Foundation, “Python Language Reference (python.org),” (2018).
6. “Luceda Photonics,” https://www.lucedaphotonics.com.
7. “GDSHelpers on Github,” https://github.com/HelgeGehring/gdshelpers.
8. “GDSHelpers’ documentation,” https://gdshelpers.readthedocs.io.
9. A. Fabri, G. J. Giezeman, L. Kettner, S. Schirra, and S. Schönherr, “On the design of CGAL a computational geometry algorithms library,” Softw. - Pract. Exp. 30(11), 1167–1202 (2000). [CrossRef]
10. K. Mehlhorn, S. Näher, and C. Uhrig, “The LEDA platform for combinatorial and geometric computing,” in (1997), pp. 7–16.
11. “GEOS - Geometry Engine, Open Source,” (2018).
12. S. Gillies and Others, “Shapely: manipulation and analysis of geometric objects,” (2007).
13. L. H. Gabrielli, “gdspy,” https://github.com/heitzmann/gdspy.
14. Andrew G. Mark, “gdsCAD,” https://github.com/hohlraum/gdsCAD.
15. J. Petykiewicz, “fatamorgana,” https://mpxd.net/code/jan/fatamorgana.
16. Blender Foundation, “Blender,” https://www.blender.org/.
17. A. F. Oskooi, D. Roundy, M. Ibanescu, P. Bermel, J. D. Joannopoulos, and S. G. Johnson, “Meep: A flexible free-software package for electromagnetic simulations by the FDTD method,” Comput. Phys. Commun. 181(3), 687–702 (2010). [CrossRef]
18. J. W. Silverstone, D. Bonneau, K. Ohira, N. Suzuki, H. Yoshida, N. Iizuka, M. Ezaki, C. M. Natarajan, M. G. Tanner, R. H. Hadfield, V. Zwiller, G. D. Marshall, J. G. Rarity, J. L. O’brien, and M. G. Thompson, “On-chip quantum interference between silicon photon-pair sources,” Nat. Photonics 8(2), 104–108 (2014). [CrossRef]
19. M. BŁahut and D. Kasprzak, “Multimode interference structures-properties and applications,” Opt. Appl. 34, 573–587 (2004).
20. F. Van Laere, T. Claes, J. Schrauwen, S. Scheerlinck, W. Bogaerts, D. Taillaert, L. O’faolain, V. Thourhout, and R. Baets, “Compact Focusing Grating Couplers for Silicon-on-Insulator Integrated Circuits,” IEEE Photonics Technol. Lett. 19(23), 1919–1921 (2007). [CrossRef]
21. G. Roelkens, D. Van Thourhout, and R. Baets, “High efficiency grating coupler between silicon-on-insulator waveguides and perfectly vertical optical fibers,” Opt. Lett. 32(11), 1495–1497 (2007). [CrossRef]
22. Z. Xiao, T.-Y. Liow, J. Zhang, P. Shum, and F. Luan, “Bandwidth analysis of waveguide grating coupler,” Opt. Express 21(5), 5688 (2013). [CrossRef]
23. V. R. Almeida, R. R. Panepucci, and M. Lipson, “Nanotaper for compact mode conversion,” Opt. Lett. 28(15), 1302 (2003). [CrossRef]
24. M. Blaicher, M. R. Billah, T. Hoose, P.-I. Dietrich, A. Hofmann, S. Randel, W. Freude, and C. Koos, “3D-Printed Ultra-Broadband Highly Efficient Out-of-Plane Coupler for Photonic Integrated Circuits,” Conf. Lasers Electro-Optics STh1A.1 (2018).
25. H. Gehring, M. Blaicher, W. Hartmann, P. Varytis, K. Busch, M. Wegener, and W. H. P. Pernice, “Low-loss fiber-to-chip couplers with ultrawide optical bandwidth,” APL Photonics 4(1), 010801 (2019). [CrossRef]
26. N. Lindenmann, G. Balthasar, D. Hillerkuss, R. Schmogrow, M. Jordan, J. Leuthold, W. Freude, and C. Koos, “Photonic wire bonding: a novel concept for chip-scale interconnects,” Opt. Express 20(16), 17667 (2012). [CrossRef]
27. M. R. Billah, M. Blaicher, T. Hoose, P.-I. Dietrich, P. Marin-Palomo, N. Lindenmann, A. Nesic, A. Hofmann, U. Troppenz, M. Moehrle, S. Randel, W. Freude, and C. Koos, “Hybrid integration of silicon photonics circuits and InP lasers by photonic wire bonding,” Optica 5(7), 876 (2018). [CrossRef]
28. Y. Okawachi, K. Saha, J. S. Levy, Y. H. Wen, M. Lipson, and A. L. Gaeta, “Octave-spanning frequency comb generation in a silicon nitride chip,” Opt. Lett. 36(17), 3398 (2011). [CrossRef]
29. X. Guo, C. L. Zou, C. Schuck, H. Jung, R. Cheng, and H. X. Tang, “Parametric down-conversion photon-pair source on a nanophotonic chip,” Light: Sci. Appl. 6, e16249 (2017). [CrossRef]
30. O. Kahl, S. Ferrari, V. Kovalyuk, G. N. Goltsman, A. Korneev, and W. H. P. Pernice, “Waveguide integrated superconducting single-photon detectors with high internal quantum efficiency at telecom wavelengths,” Sci. Rep. 5(1), 10941 (2015). [CrossRef]
31. C. Schuck, W. H. P. Pernice, and H. X. Tang, “Waveguide integrated low noise NbTiN nanowire single-photon detectors with milli-Hz dark count rate,” Sci. Rep. 3(1), 1893 (2013). [CrossRef]
32. W. H. P. Pernice, C. Schuck, O. Minaeva, M. Li, G. N. Goltsman, A. V. Sergienko, and H. X. Tang, “High-speed and high-efficiency travelling wave single-photon detectors embedded in nanophotonic circuits,” Nat. Commun. 3(1), 1325 (2012). [CrossRef]
33. F. Pyatkov, V. Fütterling, S. Khasminskaya, B. S. Flavel, F. Hennrich, M. M. Kappes, R. Krupke, and W. H. P. Pernice, “Cavity-enhanced light emission from electrically driven carbon nanotubes,” Nat. Photonics 10(6), 420–427 (2016). [CrossRef]
34. S. Khasminskaya, F. Pyatkov, K. Słowik, S. Ferrari, O. Kahl, V. Kovalyuk, P. Rath, A. Vetter, F. Hennrich, M. M. Kappes, G. Gol’tsman, A. Korneev, C. Rockstuhl, R. Krupke, and W. H. P. Pernice, “Fully integrated quantum photonic circuit with an electrically driven light source,” Nat. Photonics 10(11), 727–732 (2016). [CrossRef]
35. A. N. McCaughan and K. K. Berggren, “A Superconducting-Nanowire Three-Terminal Electrothermal Device,” Nano Lett. 14(10), 5748–5753 (2014). [CrossRef]
