WO2022075924A1

WO2022075924A1 - Long-wave infrared waveguide device

Info

Publication number: WO2022075924A1
Application number: PCT/SG2021/050608
Authority: WO
Inventors: Yiming Ma; Yuhua CHANG; Weixin LIU; Chengkuo Lee
Original assignee: National University Of Singapore
Priority date: 2020-10-09
Filing date: 2021-10-08
Publication date: 2022-04-14

Abstract

A long-wave infrared, LWIR, waveguide device, a method of fabricating an LWIR waveguide device, a method of fabricating a waveguide device, a waveguide device fabricated using the method, and a waveguide device. The method of fabricating an LWIR waveguide device comprises the steps of providing a substrate; providing a waveguide on the substrate, the waveguide configured to propagate light in the LWIR; and providing a detector element on the substrate, the detector element coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide.

Description

LONG-WAVE INFRARED WAVEGUIDE DEVICE

FIELD OF INVENTION

The present invention relates broadly to a long-wave infrared, LWIR, waveguide device, a method of fabricating an LWIR waveguide device, a method of fabricating a waveguide device, and a waveguide device.

BACKGROUND

Any mention and/or discussion of prior art throughout the specification should not be considered, in any way, as an admission that this prior art is well known or forms part of common general knowledge in the field.

The long-wave infrared (LWIR) spanning between about 6-14 pm is an important wavelength range in the electromagnetic spectrum. It not only encompasses an atmospheric transparency windows (about 8-14 pm) which is essential for remote sensing and thermal imaging, but also overlaps with the primary absorption bands of numerous biochemical bonds as well as the fingerprint region (about 7-20 pm), both of which are of prime interest to absorption-based sensing. The intrinsic molecular selectivity of absorption spectroscopy enables multiplexed sensing without the need of sensor surface functionalization.

SUMMARY

In accordance with a first aspect of the present invention, there is provided a long-wave infrared, LWIR, waveguide device comprising: a substrate, a waveguide on the substrate and configured to propagate light in the LWIR; and a detector element on the substrate and coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide.

In accordance with a second aspect of the present invention, there is provided a method of fabricating an LWIR waveguide device comprising the steps of: providing a substrate, providing a waveguide on the substrate, the waveguide configured to propagate light in the LWIR; and providing a detector element on the substrate, the detector element coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide. In accordance with a third aspect of the present invention, there is provided a method of fabricating a waveguide device, the method comprising the steps of: forming a waveguide in a membrane suspended on a first substrate; and transferring the membrane onto an acceptor substrate; wherein transferring the membrane comprises using a stamp with an array of protrusions formed on a surface thereof, the array of protrusions extending between comers of the stamp to as to engage the membrane at multiple points away from the corners of the stamp.

In accordance with a fourth aspect of the present invention, there is provided a waveguide device fabricated using the method of the third aspect.

In accordance with a fifth aspect of the present invention, there is provided a waveguide device comprising: a waveguide formed in a material layer made from a first material on an acceptor substrate free from the first material; wherein a lateral size of the material layer is larger than 1mm x 1mm.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will be better understood and readily apparent to one of ordinary skill in the art from the following written description, by way of example only, and in conjunction with the drawings, in which:

Figure 1 shows a schematic drawing illustrating a LWIR integrated waveguide photonic system according to example embodiments.

Figure 2 shows a schematic drawing illustrating a non-limiting example of a flip-chip bonding process for use in an example embodiment.

Figure 3(a) schematically illustrates the evanescent field detection principle of a waveguide sensor according to an example embodiment.

Figure 3(b) shows a graph illustrating that the adsorption of CO2 by polyethylenimine (PEI) leads to a few absorption peaks, according to an example embodiment.

Figure 3(c) shows a graph illustrating that some gases and VOCs themselves provide characteristic absorption peaks in 6-8 pm, according to an example embodiment.

Figure 4 shows a schematic diagram illustrating the fabrication of LWIR suspended Si waveguide devices, according to an example embodiment.

Figure 5(a) shows a schematic diagram illustrating the fabrication process of LWIR SOCF waveguide-integrated PdSe2 photodetectors according to an example embodiment. Figures 5(b) shows an image of a PDMS stamp comprising a pyramid array, according to an example embodiment.

Figure 6(a) shows a schematic diagram illustrating the fabrication process of PDMS stamp according to an example embodiment.

Figure 6(b) shows a schematic diagram illustrating the fabrication process of integrating graphene photodetectors with waveguides according to an example embodiment.

Figure 7(a) shows the SWG pattern of the grating coupler according to an example embodiment.

Figure 7(b) shows a SWG with bending portions, according to an example embodiment,

Figure 7(c) shows a SWG Y-junction, according to an example embodiment.

Figure 7(d), a SWG directional coupler, according to an example embodiment.

Figure 8(a) shows the measured coupling efficiency spectrum of a grating coupler, according to an example embodiment.

Figure 8(b) shows the measured propagation loss of the waveguides, according to an example embodiment.

Figure 8(c) shows the measured bending loss of the waveguides bends, according to an example embodiment.

Figure 9(a) presents the performance of the cascaded Y-junction according to an example embodiment.

Figure 9(b) shows the imbalance spectrum measured at the last stage of the cascaded Y- junction, according to an example embodiment.

Figure 9(c) shows the self-normalized transmission in the transmitted port (T/I) and coupled port (X/I) of a single directional coupler with varying coupling lengths L_c at 6.48 pm wavelength, according to an example embodiment.

Figure 9(d) shows the contour map of coupling efficiency of a single directional coupler against the wavelength and L_c, according to an example embodiment.

Figure 10(a) shows a membrane used in a SOCF waveguide devices according to an example embodiment.

Figure 10(b) shows a zoom-in illustrating the supports are designed to be trapezoidal and break at their shorter edge when the membrane is picked up, according to an example embodiment.

Figure 11(a) shows the optical microscope (OM) image of waveguides with different lengths fabricated on a 2.4 mm x 2.4 mm membrane for cutback measurement of propagation loss, according to an example embodiment. Figure 11(b) shows the optical microscope image of waveguides with different bend numbers fabricated on a 2.4 mm x 2.4 mm membrane for cutback measurement of bending loss, according to an example embodiment.

Figure 11(c) shows the optical microscope image of cascaded Y-junctions fabricated on a 2.4 mm x 2.4 mm membrane for cutback characterization of Y-junction, according to an example embodiment.

Figure 11(d) shows the SEM image of waveguide with SWG cladding, according to an example embodiment.

Figure 11(e) shows the SEM image of waveguide bend, according to an example embodiment.

Figure 11(f) shows the SEM image of Y-junction, according to an example embodiment.

Figure 11(g) shows the broadband spectrum of propagation loss in 6.3-7.1 pm. The inset shows representative cutback measurement results at 6.55 pm, according to an example embodiment.

Figure 11(h) shows the broadband spectrum of bending loss in 6.3-7.1 pm. The inset shows representative cutback measurement results at 6.55 pm, according to an example embodiment.

Figure 1 l(i) shows the broadband spectrum of Y-junction transmission in 6.3-7.1 pm. The insets show representative cutback measurement results at 6.55 pm, according to an example embodiment.

Figure 12(a) shows a schematic side view drawing illustrating that PdSe2 holds a unique puckered pentagonal crystal structure.

Figure 12(b) shows a schematic top view drawing illustrating that PdSe2 holds a unique puckered pentagonal crystal structure.

Figure 13(a) shows a PdSe2 flake field-effect-transistor, FET, fabricated on a Si substrate according to an example embodiment.

Figure 13(b) shows angular-resolved polarized Raman spectroscopy graph in parallel configuration to determine the two in-plane lattice axes of the FET according to an example embodiment.

Figure 13(c) shows the photoresponse along a-axis is -1.3 times stronger than that along b-axis for the FET, according to an example embodiment.

Figure 14(a) shows the optical microscope image of the fabricated PdSe2 flake photodetector integrated with a SOCF LWIR SWGWG waveguide system, according to an example embodiment.

Figure 14(b) presents the zoom-in view of the PdSe2 photodetector according to an example embodiment. Figure 14(c) shows the Raman spectra of the PdSe2 flake in parallel configuration with polarization along a- and b-axes, respectively, in the detector according to an example embodiment.

Figure 14(d) shows that the flake thickness is confirmed to be 67 nm by AFM, according to an example embodiment.

Figures 14(e) shows the cross-sectional transverse electric field distribution of the 67 nm PdSe2 flake on the SWGWG waveguide at 6.51 pm wavelength, according to an example embodiment.

Figures 14(f) shows the zoom-in view from Figure 14(e) at the PdSe2 layer/flake, according to an example embodiment.

Figure 15(a) shows the optical microscope and SEM images of a grating coupler GC I, according to an example embodiment.

Figure 15(b) shows the optical microscope and SEM images of grating coupler GC II, according to an example embodiment.

Figure 15(c) shows the optical microscope and SEM images of a Y-junction, according to an example embodiment.

Figure 15(d) shows a graph illustrating the on-chip insertion losses from the input grating couplers to the input port of the PdSe2 photodetector, according to an example embodiment.

Figure 16(a) shows a graph illustrating the repeatable photocurrent with stable and reversible photoresponse in the device according to an example embodiment.

Figure 16(b) shows a graph illustrating that the photocurrent increases linearly with the increasing incident power, according to an example embodiment.

Figure 16(c) shows a graph illustrating that the waveguide integration according to an example embodiment provides more than 11 times enhancement of the responsivity in photocurrent.

Figure 16(d) shows a graph illustrating that the photocurrent decreases with increasing temperature, according to an example embodiment.

Figure 17(a) plots the dark current and the sum of shot noise and Johnson noise calculated from the dark current for the waveguide-integrated PdSe2 photodetector according to an example embodiment.

Figure 17(b) shows the current noise power density spectrum measured by lock-in amplifier, according to an example embodiment.

Figure 17(c) presents the measured frequency response, showing a 3-dB bandwidth of 2.74 kHz, according to an example embodiment. Figure 17(d) shows a graph illustrating that the responsivity decreases with increasing wavelength, and correspondingly, the calculated noise equivalent power, NEP, increases, according to an example embodiment.

Figure 18(a) shows a schematic illustration of the suspended Si waveguide gas sensing platform, according to an example embodiment.

Figure 18(b) shows an optical image of the suspended Si spiral waveguide, according to an example embodiment.

Figure 18(c) shows a zoom-in view of the sensing waveguide surrounded by toluene molecules as indicated by the square box in Figure 18(a).

Figure 18(d) shows the absorption spectrum of toluene in 6-8 pm wavelength range.

Figure 18(e) shows a schematic illustration of the gas sensing testing setup according to an example embodiment.

Figure 19(a) shows the optical absorbance of toluene at its absorption peak of 6.65 pm versus toluene concentration, according to an example embodiment.

Figure 19(b) shows the response and recovery characteristic cycle curve of the suspended Si platform to alternative injection between pure N2 and 75 ppm toluene-N2 dilution, according to an example embodiment.

Figure 19(c) shows the response time, according to an example embodiment.

Figure 19(d) shows the recovery time, according to an example embodiment.

Figure 20(a) shows an artist’s impression of the envisioned on-chip EWIR spectroscopic sensor featuring transfer-printed SOCF multichannel waveguides integrated with graphene photodetectors and zoom-in view of the waveguide-integrated graphene photodetector with GSG electrode configuration, according to an example embodiment.

Figure 20(b) shows a sketch of the on-chip sensing mechanism using toluene as an example, according to an example embodiment

Figure 20(c) shows a band diagram of the graphene photodetector at zero bias, according to an example embodiment.

Figure 21(a) shows a graph illustrating power absorption ratios of graphene at different widths and thicknesses of the signal electrode, according to an example embodiment.

Figure 21(b) shows a graph illustrating power absorption ratios of metal at different widths and thicknesses of the signal electrode, according to an example embodiment.

Figure 21(c) shows graphene absorptances at different widths and thicknesses of the signal electrode, according to an example embodiment. Figure 21(d) shows total absorptances at different widths and thicknesses of the signal electrode, according to an example embodiment.

Figure 21(e) shows a schematic illustration of the simulated structure, according to an example embodiment.

Figure 21(f) shows the simulated electric field IE_yl distribution in the xy plane, according to an example embodiment.

Figure 21(g) shows the simulated electric field IE_yl distribution in the xz plane, according to an example embodiment.

Figure 21(h) shows the simulated electric field IE_yl distribution in the yz plane, according to an example embodiment.

Figure 22(a) shows an OM image of the graphene photodetector integrated with SOCF waveguide featuring two grating couplers with different designs for broadband characterization, according to an example embodiment.

Figure 22(b) shows a false-colored SEM image of the waveguide-integrated graphene photodetector, according to an example embodiment.

Figure 22(c) shows the temporal photoresponses under different incident powers, according to an example embodiment, where the inset shows the results under low incident powers.

Figure 22(d) shows the photocurrent and responsivity as a function of incident power, according to an example embodiment.

Figure 22(e) shows the spectral responsivity, according to an example embodiment.

Figure 22(f) shows the measured and predicted frequency response, according to an example embodiment.

Figure 22(g) shows the spectral current noise power density, according to an example embodiment with the inset showing the spectral NEP.

Figure 23(a) shows an OM image of the graphene photodetector integrated with SOCF folded waveguide, according to an example embodiment.

Figure 23(b) shows a falsecolored SEM image of the waveguide-integrated graphene photodetector, according to an example embodiment.

Figure 23(c) shows the photoresponse of the waveguide-integrated graphene photodetector under alternating injection of pure N2 and 0.72% toluene-N2 dilution into the vicinity of the folded waveguide, according to an example embodiment.

Figure 23(d) shows the noise and LoD as a function of the incident power to the waveguide- integrated graphene photodetector, according to an example embodiment, with the inset showing the noise as a function of the signal measured by the MCT detector. Figure 24 shows a flow chart illustrating a method of fabricating an LWIR waveguide device, according to an example embodiment.

Figure 25 shows a flow chart illustrating a method of fabricating a waveguide device, according to an example embodiment.

DETAILED DESCRIPTION

Embodiments of the present invention provide a LWIR integrated waveguide photonic system featuring the integration of waveguide devices on substrate and the integration of light source and detectors with the waveguide devices for LWIR on-chip photonic sensing. Example embodiments for each component and integration process are described. According to example embodiments, LWIR waveguide devices are provided in a broad wavelength range on both suspended silicon (Si) and silicon-on-calcium-fluoride (SOCF) platforms. Various functional building blocks, including grating couplers, Y-junctions and directional couplers, are described according to example embodiment, showing good broadband operation capability. In addition, LWIR waveguide-integrated photodetectors according to example embodiments are provided through integrating palladium diselenide (PdSe2) and graphene photodetectors with SOCF waveguides. Moreover, on-chip absorption sensing of toluene according to an example embodiment is demonstrated through integrating a graphene photodetector with a SOCF folded waveguide. The fabrication and broadband characterization of waveguides, various functional building blocks, and waveguide-integrated photodetectors, and the demonstration of on-chip absorption sensing using a photodetector-integrated waveguide sensor according to example embodiments are provided for the first time for LWIR Si photonics, to the best of the inventors’ knowledge. The example embodiments show the potential of the realization of LWIR integrated waveguide photonic system for various on-chip sensing applications.

System Configuration according to example embodiments

The LWIR integrated waveguide photonic system 100 according to an example embodiment is schematically illustrated in Figure 1. The system 100 is based on the integration of waveguide devices with substrate 104 and the integration of light source 106 and detectors 108 with the waveguide devices. The light emitted from the light source 106 is coupled into the waveguide 110 by a grating coupler 112. Various functional building blocks such as Y-junction 114, directional coupler 116, and multi-mode interferometer (MMI) can be employed for light splitting and routing, generally indicated at numeral 118. The light is guided to multiple waveguide channels generally indicated at numeral 120 in the sensing region for multiplexed sensing, where numerous waveguide structures can be utilized. For example, spiral waveguides 122 and folded waveguides 124 can be used to increase the waveguide length (compared to a straight waveguide 126) while reducing the overall footprint. Various effects can be leveraged to enhance the light-matter interaction so as to improve the sensing performance, such as cavity enhancement by ring resonator and slow light effect by photonic crystal waveguide. In the example shown in Figure 1, a microfluidic channel system 128 is provided to guide and confine the analytes, e.g., gases, for their interaction with the waveguide channels 120. At the output, the light is coupled into detectors e.g. 108, which are integrated with waveguides e.g. 130 or grating couplers e.g. 132, to convert the optical sensing signals to electrical signals.

Component and Integration Solutions according to example embodiments

For the LWIR photonic material platform, silicon-on-insulator (SOI) can be considered according to example embodiments because of the mature complementary metal-oxide- semiconductor (CMOS) fabrication technology and the abundant infrastructures built in near- and mid-wave infrared. The strong SiO2 absorption is circumvented through locally removing the buried oxide (BOX) and suspending the Si waveguides using subwavelength gratings (SWGs). Although this platform possesses advantages of simple process and good scalability, the suspended Si waveguide structures can cause additional complications to the integration of light source and detectors. Therefore, a transfer printing technique is developed according to other example embodiments to integrate Si waveguide devices with a rigid transparent substrate, which will be described in detail with reference to Figure 5 below. In one example embodiment, calcium fluoride (CaF2) is chosen as the substrate material because of its wide transparency window up to 8 pm and low refractive index of ~1.4. Therefore, silicon-on- calcium-fluoride (SOCF) platform according to an example embodiment offers both a wide bandwidth covering the whole transparency window of Si and a high index contrast.

The potential light sources for integration according to example embodiment can be microheater, laser, and frequency-down-conversion component. The potential detectors for integration according to example embodiments can be thermopile, bolometer, and other MEMS radiation detectors. Flip-chip bonding technique can be used to integrate light source and detectors with waveguide devices, as illustrated in Figure 2. Figure 2 serves as a non-limiting example for the illustration of flip-chip bonding process. Here, firstly Au pads 200, 202 are deposited and patterned on the surfaces of both Si wafer 204 and thermopile chip 206. Next, the thermopile chip 206, with Au pads also formed thereon in corresponding positions to the Au pads 200, 202, is flipped, aligned with, and bonded on top of the Si wafer 204 via thermocompression bonding. Finally, epoxy 208 is coated on the surface of the Si wafer 204 and then microheater 210 is bonded onto the epoxy 208

Emerging 2D materials are regarded as a promising alternative solution for LWIR photodetection on waveguide devices, according to example embodiments. The layered lattice structures enable their direct monolithic integration with waveguide devices using a transfer printing process, which will also be described in detail with reference to Figure 5 below. 2D materials that can be utilized for LWIR photodetection according to example embodiments can include graphene, black phosphorus (BP), black arsenic -phosphorus (b-AsP), molybdenum disulfide (M0S2), palladium diselenide (PdSe2), platinum diselenide (PtSe2), etc. Among them, PdSe2 is a promising candidate for LWIR photodetection because of its narrow indirect bandgap of 0.05 eV corresponding to a cut-off wavelength of -23 pm and high stability in ambient air. The gapless nature of graphene endows it with high-speed broadband photoresponse from ultraviolet to terahertz while also resulting in a large dark current when bias is applied, reducing the detection sensitivity. Zero-bias operation is thus preferred to reduce the dark current as well as the power consumption. Sensing Applications according to example embodiments

Because the LWIR overlaps with the primary absorption bands of numerous biochemical bonds as well as the fingerprint region, the LWIR integrated waveguide photonic system according to example embodiments possesses promising potential for various absorption-based sensing applications. Figure 3(a) schematically illustrates the evanescent field detection principle of waveguide sensors. The characteristic absorption wavelengths are determined by the vibration frequencies of the functional groups. Consequently, different functional groups of unlike biochemical molecules possess different characteristic absorption bands, providing intrinsic molecular selectivity to absorption spectroscopy. Therefore, multiplexed sensing can be performed without the need of sensor surface functionalization. Taking suspended Si or SOCF platform according to example embodiments, the usable wavelength range in the LWIR is about 6-8 pm, which covers the characteristic absorption peaks of various common gases and volatile organic compounds (VOCs). For example, the adsorption of CO2 by polyethylenimine (PEI) leads to a few absorption peaks, as shown in Figure 3(b). Some gases and VOCs themselves provide characteristic absorption peaks in 6-8 pm, as shown in Figure 3(c).

System Demonstration according to example embodiments

LWIR waveguide devices in a broad wavelength range from 6.3 to 7.1 pm on both suspended Si and SOCF platforms for use in example embodiments where fabricated. Various fabricated functional building blocks, including grating couplers, Y-junctions and directional couplers, demonstrated high performance. In addition, LWIR waveguide-integrated PdSe2 and graphene photodetectors for use in example embodiments were demonstrated on the SOCF platform. Furthermore, toluene absorption sensing on the suspended Si platform and on-chip absorption sensing of toluene through integrating a graphene photodetector with a SOCF folded waveguide are demonstrated in example embodiments.

Fabrication Processes according to example embodiments

As shown in Figure 4, the fabrication of LWIR suspended Si waveguide devices according to an example embodiment starts from a commercially available 8" SOI wafer 400 with 1.5 pm thick device layer 402 and 3 pm thick BOX layer 404. The subwav elength grating waveguide (SWGWG) patterns 406 are defined by e-beam lithography (EBL) using ZEP-520A resist 408 and then transferred to the Si device layer 402 by SF6/C4F8 deep reactive-ion etching (DRIE). The wafer 400 is then submerged into a 1:5 hydrofluoric acid (HF) bath for 3.5 h, to ensure removal of the BOX 404 beneath, in particular under the widest devices, in the example embodiments described the grating couplers e.g. 112, 132 (Figure 1). Thanks to the high mechanical robustness of the 1.5 pm thick SWGWGs 407, the wafer 400 can be dried using nitrogen gun blowing after the HF bath. The resist 408 is only removed by oxygen plasma etching after the HF etching in order to protect the Si top surface.

Figure 5(a) illustrates the fabrication process of LWIR SOCF waveguide-integrated PdSe2 photodetectors according to an example embodiment, which also serves as an example of utilizing a transfer printing technique for the integration of waveguide devices with new acceptor substrate and the integration of 2D material photodetectors with waveguide devices, according to example embodiments. The fabrication also starts from an 8" SOI wafer 500 with 1.5 pm thick device layer 502 and 3 pm thick BOX layer 504. The SWGWG patterns 506 and membrane patterns 508 are simultaneously defined by EBL using ZEP-520A resist 510 and then transferred to the Si device layer 502 by SFe ZUFs DRIE. The wafer 500 is then submerged into a 1:5 HF bath for 3.5 h, to ensure removal of the BOX 504 beneath the widest devices. Because of the relatively fragile structures of the large suspended membranes, the wafer 500 is dried using critical point dryer. The membranes 512 are then transfer printed using a transfer station consisting of a microscope, a micromanipulator, and a sample stage. A polydimethylsiloxane (PDMS) stamp 514 with pyramidal bumps e.g. 516 is brought down until it is in contact with the membrane 512, then retreated quickly to pick up the membrane 512 and break the Si supports e.g. 513. The donor SOI wafer 500 is then removed and replaced by a CaF2 acceptor substrate 518. The stamp 514 is then lowered to bring the membrane 512 onto the CaF2 acceptor substrate 518 surface. By retreating the stamp 514 slowly, the membrane 512 is printed onto the CaF2 acceptor substrate 518 surface because of the larger adhesive force therebetween compared to between the membrane 512 and the stamp 514. The residual ZEP- 520A resist 510 is removed by firstly acetone bath then oxygen plasma etching. PdSe2 flakes e.g. 520 are first mechanically exfoliated from a bulk crystal 522 by a tape and then transferred onto a PDMS stamp 524 on a glass slide (not shown). Next, a selected PdSe2 flake 520 is transfer printed onto the SWGWG 528 using the transfer station. Electrodes are patterned by LaserWriter with AZ 1512 HS as resist 530. Subsequently, 10 nm Ti and 100 nm Au 532 are deposited by e-beam evaporation, followed by lift-off in acetone to form the metal contacts 534, 536. It is worth noting that this transfer printing technique according to example embodiments is capable of integrating virtually any photonic device materials (e.g., silicon nitride, germanium, aluminum nitride) with virtually any substrate materials (e.g., sapphire, diamond, chalcogenide glasses). As a result, heterogeneously integrated anything-on-anything photonics can be achieved. It is also worth noting that this transfer printing technique is able to integrate fundamentally any 2D materials (e.g., graphene, black phosphorus) with waveguide devices based on nearly any photonic device materials for on-chip optoelectronic integrations.

With reference to Figures 5(a) and (b), in this example embodiment, the PDMS stamp 514 comprises a pyramid array 538, in contrast to for example using pyramids only at the four comers of the stamp 514. It was found by the inventors that when using a PDMS stamp with pyramids only at the four corners, the PDMS surface between the pyramids at the corner can stick to the Si membrane when the membrane is large and soft, thus the membrane size that can be transferred is limited. In contrast, having the pyramid array 538, or generally an array of protrusions, extending between comers of the stamp to as to engage the membrane at multiple points away from the comers of the stamp can advantageously enable transfer of larger and soft membranes.

Figure 6(a) illustrates the PDMS stamp fabrication process in this example embodiment. The mold is fabricated from a <100> Si wafer with 300 nm SiO2. The sample is firstly spin-coated with AZ 1514 at 6500 rpm and patterned into an array of squares with 45 pm period and 15 pm width. SiO2 is etched by Ar/CHFa at 60 °C. The etching rate is around 100 nm per minute. Then <100> Si is wet etched by 30% potassium hydroxide (KOH) at 80 °C for 30 min to create the inverse pyramids with SiCh layer as a hard mask. Magnetic stirring is recommended otherwise air bubbles may block SiCh opening from KOH solution. Afterward, SU8 3050 is spin-coated at 1000 rpm for approximately 120 pm thickness and exposed. The mold is left in the air to age for one day to facilitate demolding. Before PDMS can be poured in, the SU8 surface is further passivated by merging mold into detergent solution. The mixing ratio of PDMS base and agent is 5:1 for better elasticity. The regular 10:1 mixing ratio leads to soft microtips on the stamp, which will not bounce back to the initial state and the membrane cannot be transferred. Finally, PDMS is cured at 65 °C on a hot plate for 2 hours and peeled off.

The supporting beams e.g. 513 (Figure 5(a)) are full etched in this example embodiment. The supporting beam dimensions and space is preferably optimized to achieve a better trade-off between the stiffness to support membrane and the possibility of break during transfer.

Accordingly, embodiments of the present invention are able to transfer significantly larger membrane than 1mm x 1mm, for example, but not limited to, 2.4mm x 2.4mm.

Figure 6(b) illustrates the fabrication process to integrate graphene photodetectors with SOCF waveguides according to an example embodiment, which also serves as an example of a modified transfer printing technique for the integration of 2D material photodetectors with waveguide devices, according to example embodiments. The SOCF waveguides are prepared using the same process as illustrated by the top half part of Figure 5(a). Compared with the bottom half part of Figure 5(a), the process shown in Figure 6(b) is more complicated but provides higher success rate of integrating 2D materials with waveguide devices. In Figure 6(b), the graphene transfer process starts from a silicon transfer substrate spin-coated with PMGI and PMMA. Graphene flakes are mechanically exfoliated onto the transfer substrate using typical scotch tape. After identifying the flake with suitable size and thickness by optical microscopy and Raman spectroscopy, the flake is isolated by removing a circle of resist around it using tweezer scratching. Then the PMGI layer within the circle is slowly dissolved by MF319 solution, which is delivered to the circular scratch by tweezer leveraging the capillary effect. The transfer substrate is then gently dipped in DI water with a small tilt angle. The PMMA membrane carrying the chosen graphene flake is spontaneously separated from the silicon substrate and floats on the DI water because of the surface tension. The membrane is lifted out of the water using a washer and slowly dried by the light from the optical microscope. The transfer station is used to align the graphene flake with the waveguide on the target substrate. Once in contact, heat is applied to the target substrate to melt the PMMA from the washer onto the waveguide. The target substrate is then baked at 130 °C for 1 h to ensure good adhesion between graphene and waveguide. The PMMA residue is removed by acetone. Electrodes are patterned by e-beam lithography with PMMA as resist. Subsequently, 1 nm Ti and 70 nm Au are deposited by e-beam evaporation, followed by lift-off in acetone.

Characterization of Suspended Silicon Waveguide Devices according to example embodiments

Figure 7 shows the optical microscope and scanning electron microscope (SEM) images of the fabricated suspended-SWG-based Si waveguide devices according to example embodiments. The SWG pattern of the grating coupler 600 in Figure 7(a) is uniform rectangular holes arranged in a periodic array in both orthogonal axes. P_x = 3.9 pm and P_y = 3.8 pm are the period of the SWG along x- and y-axis, respectively. f_x = 0.70 and f_y = 0.70 are the filling factors of holes along x- and y-axis, respectively. In Figure 7(b), W = 2.5 pm is the waveguide e.g. 602 width. ASWG = 1 pm is the SWG period, smaller than the Bragg period of -1.15 pm. WSWG = 2.95 pm is the width of the SWG cladding. Lsi = 0.33 pm is the Si width in the SWG cladding. The bending radius of the waveguide bends e.g. 604 is 30 pm. In Figure 7(c), the x- and y- spans of the Y-junction e.g. 606 are 307 and 288 pm, respectively. In Figure 7(d), Directional couplers e.g. 608 with a fixed coupling gap of 0.5 pm and different coupling lengths are fabricated.

Figure 8(a) shows the measured coupling efficiency spectrum of the grating coupler 600 (Figure 7(a)) and corresponding polynomial fitting. The grating coupler achieves a maximum coupling efficiency of -7.05 dB at the central wavelength of 6.63 pm. The 1- and 3-dB bandwidths are 176 and 304 nm, respectively. Cutback method is employed to measure the propagation and bending losses of the waveguides/bends e.g. 602, 604. The results are shown in Figure 8(b) and (c), respectively. The propagation loss shows an average and standard deviation of -4.3 and 0.39 dB/cm, respectively. The bending loss presents an average of -0.056 dB/90° and a standard deviation of 0.012 dB/90°.

Figure 9(a) presents the performance of the cascaded Y-junction of Figure 7(c). The slopes extracted from linear fittings show low insertion losses of -0.43 ± 0.18, -0.09 ± 0.11, 0.01 ± 0.21, and -0.03 ± 0.25 dB at 6.55, 6.58, 6.61, and 6.64 pm wavelength, respectively. Figure 9(b) shows the imbalance spectrum measured at the last stage of the cascaded Y-junction. The imbalance is with an average of -0.04 dB and a standard deviation of 0.38 dB. Figure 9(c) shows the self-normalized transmission in the transmitted port (T/I) and coupled port (X/I) of a single directional coupler with varying coupling lengths L_c at 6.65 pm wavelength. The data are fitted well with a sine- squared function according to the coupled mode theory, signifying good performance of the directional coupler according to an example embodiments. The contour map of coupling efficiency against the wavelength and L_c is plotted in Figure 9(d). At different wavelengths, optimized coupling efficiencies could be achieved in directional couplers with suitable L_c. This contour map serves as a design guideline for directional coupler built on the presented suspended Si platform according to example embodiments.

The measurements were performed using a setup similar to the one shown in Figure 18(e) and described below, i.e., using off-chip commercial laser and detector.

Characterization of Silicon-on-Calcium-Fluoride Waveguide Devices according to example embodiments

The SOCF waveguide devices according to an example embodiment are patterned within a membrane in order to be transfer printed, as described above with reference to Figures 5 and 6. In an example embodiment, the membrane 900 is designed to be 2.4 mm x 2.4 mm in area, as shown in Figure 10(a). To ensure uniform HF access beneath the membrane, an array of holes e.g. 902 with 2 pm diameter and 25 pm separation are additionally patterned across the membrane 900. Over the areas where the waveguides e.g. 904 and grating couplers e.g. 906 are present, the holes are omitted. The supports e.g. 908 are designed to be trapezoidal and break at their shorter edge when the membrane 900 is picked up, as shown in a zoom-in portion in Figure 10(b).

In order to reduce the scattering loss, the waveguide e.g. 904 width is increased to 3.1 pm in an example embodiment. WSWG = 3.35 pm and Lsi = 0.38 pm are adopted in an example embodiment. The bending e.g. 911 radius is 35 pm. As shown in Figures 11(a), (b), and (c), waveguides with different lengths, waveguides with different bend numbers, and cascaded Y- junctions are fabricated on 2.4 mm x 2.4 mm membranes, respectively. The input and output grating couplers (e.g. 1000/1002, 1004/1006) are separated by at least 1.4 mm in example embodiments to eliminate light coupling between the input and output fibers through surface reflection. Figure 11(d) shows the SEM image of waveguide with SWG cladding, according to an example embodiment. Figure 11(e) shows the SEM image of waveguide bend, according to an example embodiment. Figure 11(f) shows the SEM image of Y-junction, according to an example embodiment.

Using cutback method, the propagation loss is measured with an average and standard deviation of -4.64 and 0.51 dB/cm, respectively, as shown in Figure 11(g). The average and standard deviation of the bending loss are -0.054 and 0.005 dB/90°, respectively, as shown in Figure 11(h). As shown in Figure 11 (i), the transmission is -3.42 ± 0.27 dB/port, confirming a broadband 50:50 light splitting ratio and a low insertion loss of ~0.42 dB. The losses are comparable with those of the above described suspended devices according to example embodiments. This might be because the reduced sidewall scattering loss by wider waveguide width is compensated by the additional scattering loss at the interface between Si and CaF2.

Anisotropy of Palladium Diselenide for use in a detector according to example embodiments

In comparison to most highly symmetrical planar hexagonal transition-metal dichalcogenides (TMDs) and graphene, PdSe2 holds a unique puckered pentagonal crystal structure, as shown in Figures 12 (a) and (b). The structure is made up of single layers of PdSe2 stacking along the c-axis and held together mainly by van der Waals forces. From the top view in Figure 12 (b), it is seen that the monolayer PdSe2 crystals are composed entirely of pentagonal rings. In a unit cell 1200 of the monolayer, each Pd atom coordinates with four Se atoms, and the two adjacent Se atoms make a covalent bond, forming a square backbone lattice. There is a slight difference between the lattice constants along a- and b-axis. The symmetry of PdSe2 is orthorhombic, which is similar to the puckered hexagonal black phosphorus. This endows PdSe2 with a unique anisotropy on account of its in-plane low symmetry.

It has been reported that PdSe2 shows anisotropic photoresponse and the strongest photoresponse occurs along either a- or b-axis. To investigate the anisotropic photoresponse of PdSe2 in the LWIR region, a PdSe2 flake field-effect-transistor, FET, is fabricated on a Si substrate according to an example embodiment, as shown in Figure 13(a). Angular-resolved polarized Raman spectroscopy in parallel configuration is performed to determine the two inplane lattice axes, as shown in Figure 13(b). The Raman spectrum consists of four obvious peaks Ag¹, A_g ², Big², and A_g ³ at 139.6, 201.3, 218.8, and 253.6 cm^-1. The Big² Raman intensity shows a variation period of 90° and is completely forbidden along the a- and b-axes. Both the Ag¹ and A_g ³ Raman intensity have a variation period of 180°, with maximum along the a-axis and the minimum along the b-axis. The anisotropy of the A_g ³ peak is much weaker than that of the Ag¹ peak. Therefore, the PdSe2 crystal orientations can be conveniently determined by firstly finding out the polarization where the Bi_g ² peak vanishes then distinguishing the a- and b-axes according to the Ag¹ peak intensity. Next, the photoresponses under EWIR light illumination polarized along a- and b-axes are measured. When the light is polarized along a- axis, electrodes 2 and 4 are used, while electrodes 1 and 3 are for polarization along b-axis (compare figure 13(a)). As illustrated by Figure 13(c), the photoresponse along a-axis is -1.3 times stronger than that along b-axis. At 6.3 pm wavelength and 1 V bias, the device according to an example embodiment achieves a responsivity of 0.62 mA/W with light polarized along a-axis.

The measurements were performed using an off-chip commercial laser. The light out from the laser was guided onto the PdSe2 FET through the free-space reflection by a set of mirrors. The polarization of the light was controlled by a halfwave plate.

Characterization of Waveguide-Integrated Palladium Diselenide Photodetector according to example embodiments

Figure 14(a) shows the optical microscope image of the fabricated PdSe2 flake photodetector integrated with a SOCF EWIR SWGWG waveguide system, according to an example embodiment using the method described above with reference to Figure 5. Two grating couplers with different designs (GC I and GC II) are used to couple light with broader bandwidth into the waveguide system. The two waveguides 1400, 1402 connected with the two grating couplers 1404, 1406 are combined through a Y-junction e.g. 1408 and then guided to the integrated PdSe2 photodetector 1410. The waveguide 1412 is divided into two using a Y- junction 1414 and then connected to two grating couplers 1416, 1418 of the same designs as those at the input via two waveguides 1420, 1422, in order to couple the residual light out for the assistance to optical characterizations. Figure 14(b) presents the zoom-in view of the PdSe2 photodetector 1410. According to the above-studied anisotropy, the a-axis of the PdSe2 flake is aligned with the TE mode polarization in the waveguide to maximize the photoresponse. The two lattice axes are confirmed using the above-mentioned method. Figure 14(c) shows the Raman spectra of the PdSe2 flake 1424 in parallel configuration with polarization along a- and b-axes, respectively. The flake 1424 thickness is confirmed to be 67 nm by AFM (Figure 14(d)). Figures 14(e) and (f) show the cross-sectional transverse electric field distribution of the 67 nm PdSe2 flake 1424 on the SWGWG waveguide 1412 at 6.51 pm wavelength and the zoom-in view at the PdSe2 layer/flake 1424, respectively, providing an intuitive view of the light-PdSe2 interaction.

Figure 15(a-c) show the optical microscope and SEM images of GC I (1404, 1416), GC II (1406, 1418), and Y-junction (1408, 1414), respectively. Using the nomenclature of Figure 6 (a), for GC I, P_x = 3.35 pm, P_y = 3.40 pm, f_x = 0.80, and f_y = 0.72. For GC II, P_x = 3.44 pm, P_y = 3.62 pm, f_x = 0.78, and f_y = 0.72. The x- and y-spans of the Y-junction are 307 and 288 pm, respectively. The PdSe2 flake was transferred with its input edge exactly aligned with the midline of the waveguide 1412. Thanks to the symmetry of the waveguide system, the on-chip insertion losses from the input grating couplers to the input port of the PdSe2 photodetector according to an example embodiment (shown in Figure 15(d)), which is used for incident power calibration, can be conveniently extracted by halving the total on-chip insertion losses measured before the PdSe2 transfer process. At different wavelengths, the light is inputted from the grating coupler with lower loss.

The photoresponse of the waveguide-integrated PdSe2 photodetector 1410 according to an example embodiment was then characterized at 6.51 pm wavelength and 5V bias. The repeatable photocurrent generation as shown in Figure 16(a) reveals stable and reversible photoresponse in the device. It is noted that the photocurrents are measured using lock-in amplifier technique. Thus, the rise and fall times read from the temporal responses are largely affected by the time constant used on the lock-in amplifier and could not accurately reflect the response speed of the photodetector. Photoresponses under different incident powers are measured. As shown in Figure 16(b), the photocurrent increases linearly with the increasing incident power. Correspondingly, the responsivity is nearly constant (64.86-73.60 pA/W) over the measurable power range. The photoresponse of the device under free-space illumination is also measured. As illustrated in Figure 16(c), no obvious photoresponse is observed under the same incident power as that illuminated through the waveguide shown in Figure 16(a). Under a higher incident power, photoresponse is observed and the responsivity is calculated to be 5.63 pA/W, less than 1/11 of that of illumination through the waveguide. In other words, the waveguide integration according to an example embodiment provides more than 11 times enhancement of the responsivity. Photoresponses at different temperatures were also measured. The photocurrent decreases with increasing temperature, as shown in Figure 16(d).

Figure 17(a) plots the dark current and the sum of shot noise and Johnson noise calculated from the dark current for the waveguide-integrated PdSe2 photodetector 1410 according to an example embodiment. Figure 17(b) shows the current noise power density spectrum measured by lock-in amplifier. The spectrum is parallel to the 1/f reference line, implying the noise is dominated by the 1/f noise under low signal modulation frequency. Figure 17(c) presents the measured frequency response, showing a 3-dB bandwidth of 2.74 kHz. The corresponding rise/decay time can be estimated as 1 /2TT 3CIB and is shorter than 58 ps. Photoresponses at different wavelengths were measured under a signal modulation frequency of 774 Hz, right before the response begins to decay. As shown in Figure 17(d), the responsivity decreases with increasing wavelength, from 103.02 pA/W at 6.33 pm to 36.28 pA/W at 6.68 pm, because of the weaker absorption at longer wavelength. Correspondingly, the calculated noise equivalent power, NEP, increases from 0.43 to 1.23 pW/ Hz^1/2. Demonstration of LWIR absorption sensing on suspended Si waveguide platform according to example embodiments

As illustrated in Figure 18(a), a LWIR gas detection platform 1800 according to another example embodiment comprises of two identical spiral waveguides 1802, 1804, a power splitter (Y-junction) 1806, and grating couplers e.g. 1808, providing simultaneous on-chip sensing, and calibration of analytes. The power splitting function can also be realized by some other building blocks such as multimode interferometer and directional coupler in different embodiments. Two spiral waveguides 1802, 1804 are separately covered by above-hanging chambers for gas feeding (not shown in the schematic graph). With the sensing waveguide 1804 purged with mixture analytes and the reference waveguide 1802 purged with atmosphere air, in principle, one can decouple the sensing signal from external influences, including environmental absorption and temperature fluctuation. Thus, this platform according to an example embodiment can provide more robust sensing results, particularly in the low- concentration range. The devices were fabricated from an SOI wafer 1810 with a 1.5-pm-thick Si device layer and a 3-pm thick BOX layer. The device fabrication consists of one stage lithography and Si deep reactive-ion etching (DRIE) to define the photonic structures, followed by 1:5 hydrogen fluoride etching to locally remove the underneath BOX layer. The inset of Figure 18(a) shows a cross-sectional scanning electron microscope (SEM) image of the fabricated suspended waveguide, suggesting that the BOX beneath the waveguide is completely removed and the etching sidewalls are vertical and smooth. Figure 18(b) shows an optical microscope image of the spiral structure (compare 1802, 1804 in Figure 18(a) according to an example embodiment, with its sensing length L of 28.4 mm. Figure 18(c) displays the zoom-in graph of the spiral sensing area 1804 of Figure 18(a). Here, the toluene molecules e.g. 1812 are uniformly distributed around the waveguide at SWG cladding as well as upper/lower air cladding and interact with the evanescent field of the guided light, resulting in additional absorption The light transmission intensity will be severely weakened at the characteristic absorption peaks of the toluene molecules e.g. 1812 compared with those low-absorption wavelengths, as depicted in Figure 18(d), which shows the absorption spectrum of toluene in the range of 6-8 pm and its maximum absorption peak at 6.65 pm wavelength.

The sensing performance of the example embodiment is characterized by the setup depicted in Figure 18(e), which can be divided into the optical characterization module 1814 and the gas regulation module 1816. In the optical characterization module 1814, a continuous-wave quantum cascade laser 1818 is employed for light-emitting. A pair of LWIR fibers 1820, 1822 is used to access the sensing platform through the on-chip grating couplers. The fine alignment between the fibers 1820, 1822 and the grating couplers is performed with a pair of six-axis alignment stages 1824, 1826 and a sample stage 1828, and using a microscope 1839. Before being focused by a ZnSe lens 1830 into the input fiber 1820, the light is firstly polarization- controlled by a halfwave plate 1832 and modulated by a chopper 1834a, b. The postchip modulated light is converted to an electrical signal by liquid nitrogen cooled MCT detector 1836. This electrical signal is amplified by a pre amplifier before sent to a lock-in amplifier 1838 to enhance the signal-to-noise ratio. The optical path is sealed and purged with clean dry air to minimize the optical absorption of ambient environment during the test. In the gas regulation module 1816, nitrogen (N2) is selected as the buffered gas with its overall flow rate well controlled by a mass flow controller 1840 at 2.0 L/min. This buffered gas is divided into two flows, with one pumped into 99.5% toluene solution to generate a toluene-N2 mixture, and the other remaining as pure N2. After that, the two flows re-mix, and the toluene concentration in the dilution is calibrated by a commercial sensor 1842 before being pumped into the gas feeding chamber hanging above the chip 1844. The concentration of toluene in the sensing region is precisely and dynamically controlled by regulating the valves 1846, 1848 in the two flows, with a wide tuning range from several ppm to several thousand ppm.

As described above, the concentration of toluene in the sensing area is precisely controlled by the adjustable valves in two gas flows and calibrated by a commercial sensor (compare Figure 18(e)). To characterize the absorbance under certain toluene concentrations, nitrogen (N2) and toluene-N2 dilutions were alternately injected. Figure 19(a) shows the testing results of the toluene absorbance at 6.65 pm versus toluene concentration ranging from 144 to 1114 ppm. The absorbance is calculated based on the average transmission change in three response and recovery characterization cycles. A linear fit is employed to the measured data, and a sensitivity of 2.8 x 10^-5/ppm is extracted. The R-square of the fitting result is 0.995, showing a good linear response behavior. To further explore the limit of detection (LoD) of the example embodiment, absorbance responses to alternate injection between pure N2 and 75 ppm toluene-N2 dilutions are presented in Figure 19(b). The results experimentally show an LoD down to 75 ppm, which is limited by the off-chip noise floor and leaves a large room for improvement. Figure 19(c) and (d) show the response and recovery stage extracted from Figure 19(b), respectively, which are fitted with Boltzmann curves and indicate a response time of 0.8 s and a recovery time of 3.4 s with regard to 75 ppm toluene detection. It was verified that the flow rate is fast enough and has negligible influence on the response and recovery times. The temporally asymmetric behavior of the example embodiment could be explained by the dynamic interaction between the analyte and the waveguides. More specifically, as toluene is in the liquid phase at room temperature, it is unavoidable that a fraction of the vapor will condense into fine droplets on the surface of waveguides. The liquid-solid interaction between two materials is rather strong due to the large surface free energy of Si. This makes the absorption of toluene easier than desorption, and as a result, a longer recovery time than the response time. Despite this, the response/ recovery time of our sensor is still fast enough for real-time on-site monitoring applications.

It is noted that the measurements in Figure 19 are from one (sensing) waveguide. As mentioned above, the environmental absorption is minimized by sealing the optical path and purging with clean dry air. The temperature fluctuation is minimized by the air-conditioning system of the lab.

Using the platform with a reference waveguide as shown in Figure 18(a), one will be able to obtain more robust sensing results without the need of the above-mentioned off-chip controls. Design of waveguide-integrated graphene photodetector according to example embodiments

Figure 20(a) schematically illustrates an on-chip LWIR spectroscopic sensor 2000 according to an example embodiment, which features transfer-printed SOCF multichannel waveguides e.g. 2002 integrated with graphene photodetectors e.g. 2004 employing the ground- signalground (GSG) electrode configuration. With the light at different wavelengths propagating in the multiple waveguide channels e.g. 2002, the surrounding analyte molecules e.g. 2006 extensively interact with the evanescent field. As a result, the light transmission intensity, and thus the graphene photoresponse, is significantly weakened at the molecular absorption peaks compared with other wavelengths, as depicted in Figure 20(b). Figure 20(c) portrays the zerobias operation mechanism of the graphene photodetector e.g. 2004. Because of different doping levels in the metal-covered (p-doped) and uncovered (p+-doped because of adsorbates from the ambient air) parts of graphene, junctions with built-in gradients of electrostatic potential and Seebeck coefficient are formed near both sides of the middle signal electrode 2008 (Figure 20(a)) and the graphene flake 2009 (Figure 2(a)). These metal-doped graphene junctions not only exist at the metal/graphene interfaces but also extend tens of nanometers into the graphene channels, enabling effective separation and collection of photogenerated electron-hole pairs. The waveguides e.g. 2002 in an example embodiment are transfer-printed SOCF waveguides fabricated using a process as described above with reference to Figures 5(a) and (b), with a 2.4 mm x 2.4 mm membrane size.

In addition to enabling zero-bias photocarrier separation and collection, the middle signal electrode 2008 also provides plasmonic enhancement. The electric field near the signal electrode 2008 is enhanced by two mechanisms. The first mechanism is plasmon-induced field localization, which is a result of the continuity of boundary conditions in Maxwell equations. At the interface of two materials, electrical displacements on both sides should be equal, D\ = D . Because D = sE, the electric field E will be much larger in the material with a smaller dielectric constant a. The second mechanism is plasmonic resonance. With a properly designed width of the signal electrode 2008, the transverse component E_y of the waveguide mode in Si underneath (from the SOCF waveguides e.g. 2002) can be efficiently coupled to the plasmonic resonance of the metallic electrode. Such a resonance is localized surface plasmon resonance and leads to electric field enhancement as two hot spots at both sides of the signal electrode 2008. The enhanced electric field overlaps with the metal-doped graphene junctions, leading to a significant performance improvement of the graphene photodetector. The geometric parameters of the signal electrode 2008 have a decisive impact on photodetector performance. A trade-off needs to be considered between plasmonic enhancement and metal absorption. The graphene and metal absorptions are simulated using the finite-difference time-domain (FDTD) method.

The length of transferred graphene flake on waveguide is 33 pm in an example embodiment. Figures 21(a) and (b) show how the width Wsignai and thickness IM of signal electrode 2008 affect the ratios of graphene and metal, respectively, absorptions over the total absorption. A wider and thinner signal electrode makes the metal absorption more dominant, reducing the graphene absorption ratio PaG/Patotai. This trend agrees with reported results. For a higher responsivity, graphene is desired to absorb more than metal, necessitating a narrower and thicker signal electrode, which, however, also decreases their absorptions. To have a better design guideline, in Figures 21(c) and (d), the graphene and total, respectively, absorbed powers are compared with the incident power. The total absorptance Patotai/Pin increases with Wsignai, reaching 90% at 1.7 pm but drops dramatically when it is further widened. This is because plasmonic resonance is strongest only when Wsignai matches working wavelength, leading to highest local electric field enhancement. Correspondingly, the graphene absorptance Pad Pin is also improved from ~43 to ~55% with Wsignai increasing from 0.3 to 1.7 pm. With Wsignai narrower than 1.7 pm, the graphene and total absorptances can also be enhanced by reducing the metal thickness. Pao/Pin reaches maximum and is nearly constant when Wsignai and HM range from 1.3 to 1.7 pm and from 50 to 90 nm, respectively. Wsignai and HM were chosen to be 1.5 pm and 70 nm, respectively, in an example embodiment for better fabrication tolerance. Figure 21(e) shows a schematic illustration of the simulated structure, according to an example embodiment. Figure 21(f) shows the simulated electric field IE_yl distribution in the xy plane, according to an example embodiment. Figure 21(g) shows the simulated electric field IEyl distribution in the xz plane, according to an example embodiment. Figure 21(h) shows the simulated electric field IE_yl distribution in the yz plane, according to an example embodiment.

Characterization of waveguide-integrated graphene photodetector according to example embodiments

Figure 22(a) shows the optical microscope (OM) image of the fabricated waveguide-integrated graphene photodetector 2200 according to an example embodiment. GC I and GC II are combined through a Y junction to couple the broadband light into the graphene photodetector 2200. The residual light is split by another Y-junction and coupled out by another pair of GC I and GC II to facilitate the optical alignment. Figure 22(b) presents the false-colored SEM image of the photodetector 2200. The photodetector performance is characterized at zero bias. A lock- in amplifier technique is utilized to improve the signal-to-noise ratio (SNR). The input laser beam is modulated by an optical chopper at 227 Hz. The lock-in amplifier collects the photocurrent at the same frequency. The incident power onto the photodetector 2200 is calibrated according to the insertion loss from the input grating coupler to the photodetector 2200. This loss is obtained with the assistance of a simultaneously fabricated reference chip with an identical waveguide structure but without photodetector integrated. Because the photodetector 2200 is integrated on the center of the whole symmetric waveguide structure, this loss can be conveniently extracted by halving the total on-chip insertion losses measured by the reference chip.

Figure 22(c) shows the temporal responses under different incident powers ranging from 0.21 to 107.64 pW at 6.51 pm. The inset shows the zoom-in view of photoresponses under low incident powers. The highly repeatable photocurrent generation reveals stable and reversible photoresponse in the device according to an example embodiment. The photocurrents are extracted from Figure 22(c) and plotted as a function of the incident power in Figure 22(d). It is seen that the photocurrent almost increases linearly with increasing incident power. Correspondingly, the responsivity (responsivity = 7_Ph/F) is power-independent and calculated to be ~7.8 mA/W. Photoresponses were then measured at different LWIR wavelengths. As shown in Figure 22(e), the device according to an example embodiment possesses broadband photoresponse from 6.3 to 7.1 pm. As expected, the responsivity does not show significant wavelength dependence. The frequency response was examined by adjusting the optical chopper modulation frequency, which is up to 10 kHz. Within this measurable range, the photoresponse does not show any noticeable degradation, implying a far larger bandwidth of our device, as illustrated in Figure 22(f). Because the photocurrent is generated near the graphene/metal interfaces by photovoltaic (PV) and photothermoelectric (PTE) effects, the 3 dB bandwidth of the device according to an example embodiment is predicted to range from 10 to 500 GHz.

The noise of the device according to an example embodiment was further measured using a lock-in amplifier in dark condition. Because zero-bias operation eliminates the dark current- induced shot noise, the noise mainly consists of 1/f noise and Johnson noise. As shown in Figure 22(g), at frequencies below ~100 Hz, the measured noise power density spectrum is parallel to the 1/f reference line, indicating the noise is dominated by the 1/f noise. At higher frequencies, Johnson noise becomes dominant, which is calculated from the measured resistance R of the device. Because the device according to an example embodiment can operate at speeds well beyond 100 Hz, it is not significantly affected by the large 1/f noise at low frequencies. The noise equivalent power (NEP) as a function of wavelength is calculated and plotted in the inset of Figure 22(g). The NEP ranges from 2.3 to 5.0 nW/Hz^1/2 across the 6.3-7.1 pm wavelength range.

Demonstration of on-chip LWIR absorption sensing according to example embodiments

In order to evaluate the feasibility of utilizing this waveguide-integrated photodetector according to an example embodiment for on-chip LWIR absorption sensing, another graphene photodetector 2300 based on the same concept was integrated with a SOCF folded waveguide 2302, as shown in Figure 23(a). The folded geometry offers a long light-matter interaction length of 1.1 cm within the small footprint of the membrane. As shown in Figure 23(b), this photodetector 2300 also employs GSG electrode configuration. The signal electrode width is optimized using the same FDTD simulation method according to the AFM measured flake thickness as for the photodetector 2000 described above. Light of 6.65 pm wavelength from a tunable LWIR laser is modulated at 227 Hz and grating-coupled to the folded waveguide 2302 sensing element.

Figure 23(c) presents the measured photoresponse of the waveguide-integrated graphene photodetector 2300 when pure nitrogen (N2) and 0.72% toluene-N2 dilution are alternately injected into the vicinity of the folded waveguide. A clear and repeatable drop of the photocurrent is observed under the injection of toluene. The 0.72% toluene leads to ~3.86% photocurrent decrease. The waveguide sensor follows Beer’s law. According to the 0.72% toluene sensing result, the absorption coefficient of pure toluene is calculated to be 69.71 cm^-1. It is also read from Figure 23(c) that noise is about 0.87% of the photocurrent. Adopting the common criterion of SNR = 3, an LoD of 0.48% is derived. The measured noise mainly consists of graphene photodetector noise and laser power fluctuation. The laser power fluctuation is assessed using a free-space liquid-N2-cooled MCT detector, as plotted in the inset of Figure 23(d). The slope of the linear fitting suggests that the laser power fluctuation occupies 0.012% of the signal. Therefore, most of the measured 0.87% noise is attributed to the noise of the graphene photodetector, which is constant and does not scale up with the signal level. By optimizing the loss from the laser to the graphene photodetector or increasing the laser power, the incident power to the graphene photodetector could be increased in different embodiments, making the laser power fluctuation more dominant. As a result, with the increase of incident power, the noise percentage could decrease and a lower LoD is expected, as illustrated in Figure 23(d). In addition to the demonstration of toluene sensing, it is worth highlighting here that the presented system according to an example embodiment would have broad applicability in LWIR spectroscopic sensing due to its broadband behavior and the inherent selectivity of LWIR absorption spectroscopy.

Similar to Figure 18E, the light out from an off-chip commercial laser was guided through the fiber and coupled into the waveguide by grating coupler, for the measurements in Figures 22 and 23.

Figure 24 shows a flow chart 2400 illustrating a method of fabricating an LWIR waveguide device, according to an example embodiment. At step 2402, a substrate is provided. At step 2404, a waveguide is provided on the substrate, the waveguide configured to propagate light in the LWIR. At step 2406, a detector element is provided on the substrate, the detector element coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide.

The LWIR waveguide device may comprises the LWIR waveguide device of any of the example embodiment described above with reference to Figures 1 to 23.

Figure 25 shows a flow chart 2500 illustrating a method of fabricating a waveguide device, according to an example embodiment. At step 2502, a waveguide is formed in a membrane suspended on a first substrate. At step 2504, the membrane is transferred onto an acceptor substrate; wherein transferring the membrane comprises using a stamp with an array of protrusions formed on a surface thereof, the array of protrusions extending between comers of the stamp to as to engage the membrane at multiple points away from the corners of the stamp.

Forming the waveguide in the membrane may comprise forming supporting beams connected to the membrane on the first substrate by full etching.

The dimensions and/or number of the supporting beams may be chosen based on a trade-off between stiffness to support the membrane on the first substrate and possibility of breaking of the support beams during transfer to the acceptor substrate.

The acceptor substrate may comprise one or more of a group consisting of CaF2, BaF2, KBr, and chalcogenide glasses. The waveguide may be configured to propagate light in the LWIR.

The waveguide device may be configured as a sensor device.

The waveguide may comprise one or more of a group consisting of silicon, germanium, and aluminum nitride.

As described above, In example embodiments, LWIR waveguide devices for operation in a broad wavelength range are provided on both suspended Si and SOCF platforms. Various functional building blocks, including grating couplers, Y-junctions and directional couplers, are provided according to example embodiments with high performance. In addition, LWIR waveguide-integrated photodetector according to example embodiments were provided through an integrating PdSe2 photodetector with SOCF waveguide. The demonstration and broadband characterization of waveguides, various functional building blocks, and waveguide- integrated photodetector according to example embodiments are for the first time provided for LWIR Si photonics, to the best of the inventors’ knowledge. The results show that embodiments of the present invention can be employed for various on-chip sensing applications including:

1) trace-gas detection;

2) environmental monitoring;

3) industrial process control;

4) medical diagnostics;

5) homeland security;

6) missile guidance.

An LoD of 75 ppm for toluene vapor sensing was experimentally achieved according to an example embodiment, with the corresponding response and recovery time of 0.8 and 3.4 s, respectively. With further optimization of the losses and reduction of the noise floor by using a high-power and low-noise light source, a lower LoD down to several ppm is expected in various example embodiments.

A LWIR heterogeneously integrated graphene/Si/CaF2 waveguide photodetectors with zero standby power consumption was provided according to an example embodiment. Through waveguide integration and plasmonic enhancement, a high and broadband responsivity of around 8 mA/W is achieved in the graphene photodetector under zero-bias operation. The waveguide integrated graphene photodetector is further utilized for on-chip absorption sensing according to an example embodiment. Detection of 0.72% toluene was experimentally demonstrated, Embodiments of the present invention can have one or more of the following features and associated benefits/adv antages:

Aspects of the systems and methods described herein, such as the signal collection and processing, for example for sensing applications, may be implemented as functionality programmed into any of a variety of circuitry, including programmable logic devices (PLDs), such as field programmable gate arrays (FPGAs), programmable array logic (PAL) devices, electrically programmable logic and memory devices and standard cell-based devices, as well as application specific integrated circuits (ASICs). Some other possibilities for implementing aspects of the system include: microcontrollers with memory (such as electronically erasable programmable read only memory (EEPROM)), embedded microprocessors, firmware, software, etc. Furthermore, aspects of the system may be embodied in microprocessors having software -based circuit emulation, discrete logic (sequential and combinatorial), custom devices, fuzzy (neural) logic, quantum devices, and hybrids of any of the above device types. Of course the underlying device technologies may be provided in a variety of component types, e.g., metal-oxide semiconductor field-effect transistor (MOSFET) technologies like complementary metal-oxide semiconductor (CMOS), bipolar technologies like emitter- coupled logic (ECL), polymer technologies (e.g., silicon-conjugated polymer and metal- conjugated polymer-metal structures), mixed analog and digital, etc.

The various functions or processes disclosed herein may be described as data and/or instructions embodied in various computer-readable media, in terms of their behavioral, register transfer, logic component, transistor, layout geometries, and/or other characteristics. Computer-readable media in which such formatted data and/or instructions may be embodied include, but are not limited to, non-volatile storage media in various forms (e.g., optical, magnetic or semiconductor storage media) and carrier waves that may be used to transfer such formatted data and/or instructions through wireless, optical, or wired signaling media or any combination thereof. When received into any of a variety of circuitry (e.g. a computer), such data and/or instruction may be processed by a processing entity (e.g., one or more processors). The above description of illustrated embodiments of the systems and methods is not intended to be exhaustive or to limit the systems and methods to the precise forms disclosed. While specific embodiments of, and examples for, the systems components and methods are described herein for illustrative purposes, various equivalent modifications are possible within the scope of the systems, components and methods, as those skilled in the relevant art will recognize. The teachings of the systems and methods provided herein can be applied to other processing systems and methods, not only for the systems and methods described above.

It will be appreciated by a person skilled in the art that numerous variations and/or modifications may be made to the present invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects to be illustrative and not restrictive. Also, the invention includes any combination of features described for different embodiments, including in the summary section, even if the feature or combination of features is not explicitly specified in the claims or the detailed description of the present embodiments.

In general, in the following claims, the terms used should not be construed to limit the systems and methods to the specific embodiments disclosed in the specification and the claims, but should be construed to include all processing systems that operate under the claims. Accordingly, the systems and methods are not limited by the disclosure, but instead the scope of the systems and methods is to be determined entirely by the claims.

Unless the context clearly requires otherwise, throughout the description and the claims, the words "comprise," "comprising," and the like are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense; that is to say, in a sense of "including, but not limited to." Words using the singular or plural number also include the plural or singular number respectively. Additionally, the words "herein," "hereunder," "above," "below," and words of similar import refer to this application as a whole and not to any particular portions of this application. When the word "or" is used in reference to a list of two or more items, that word covers all of the following interpretations of the word: any of the items in the list, all of the items in the list and any combination of the items in the list.

Claims

1. A long-wave infrared, LWIR, waveguide device comprising: a substrate, a waveguide on the substrate and configured to propagate light in the LWIR; and a detector element on the substrate and coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide.

2. The waveguide device of claim 1, comprising an LWIR light source element on the substrate and coupled to the waveguide for input of the light in the LWIR into the waveguide.

3. The waveguide device of any one of claim 2, comprising the LWIR light source element comprises a source grating coupler for coupling the light into the waveguide.

4. The waveguide device of claims 2 or 3, wherein the LWIR light source element comprises a source device integrated with the waveguide on the substrate.

5. The waveguide device of claims 2 or 3, wherein the LWIR light source element comprises a source optical fiber disposed for coupling the light into the waveguide.

6. The waveguide device of any one of claims 1 to 5, wherein the waveguide comprises a sub-wavelength grating, SWG, waveguide structure.

7. The waveguide device of any one of claims 1 to 6, wherein the waveguide is suspended on the substrate.

8. The waveguide device of claim 7, wherein the substrate comprises a silicon-on- insulator, SOI, wafer.

9. The waveguide device of any one of claims 1 to 7, wherein the substrate comprises one or more of a group consisting of CaF2, BaF2, KBr, and chalcogenide glasses.

10. The waveguide device of any one of claims 1 to 9, wherein the waveguide comprises a membrane transferred onto the substrate, the waveguide being formed in the membrane.

11. The waveguide device of any one of claims 1 to 10, wherein the detector element comprises a 2-dimensional, 2D, material on the waveguide.

12. The waveguide device of claim 11, wherein the 2D material comprises a flake transferred onto the waveguide.

13. The waveguide device of claims 11 or 12, wherein the 2D material comprises one or more of a group consisting of graphene, black phosphorus, black arsenic -phosphorus, molybdenum disulfide, palladium diselenide, and platinum diselenide.

26

14. The waveguide device of any one of claims 1 to 13, wherein the detector element comprising a detector grating coupler for detecting the light in the LWIR propagated in the waveguide.

15. The waveguide device of any one of claims 1 to 14, wherein the detector element comprises a detector optical fiber disposed for coupling to the waveguide.

16. The waveguide device of aye one of claims 1 to 14, wherein the detector element is integrated with the waveguide.

17. The waveguide device of any one of claims 1 to 16, wherein the waveguide comprises one or more of a group consisting of a grating coupler, waveguide bend, straight waveguide, Y-junction, and directional coupler.

18. The waveguide device of any one of claims 1 to 17, wherein the waveguide comprises one or more of a group consisting of silicon, germanium, and aluminum nitride.

19. The waveguide device of any one of claims 1 to 18, wherein the waveguide device is configured as a sensor device.

20. The waveguide device of claim 19, comprising a reference waveguide and a sensing waveguide for decoupling a sensing signal from external influences.

21. A method of fabricating an LWIR waveguide device comprising the steps of: providing a substrate; providing a waveguide on the substrate, the waveguide configured to propagate light in the LWIR; and providing a detector element on the substrate, the detector element coupled to the waveguide for detecting the light in the LWIR propagated in the waveguide.

22. The method of claim 21, wherein the LWIR waveguide device comprises the LWIR waveguide device of any one of claims 1 to 20.

23. A method of fabricating a waveguide device, the method comprising the steps of: forming a waveguide in a membrane suspended on a first substrate; and transferring the membrane onto an acceptor substrate; wherein transferring the membrane comprises using a stamp with an array of protrusions formed on a surface thereof, the array of protrusions extending between comers of the stamp to as to engage the membrane at multiple points away from the corners of the stamp.

24. The method of claim 23, wherein forming the waveguide in the membrane comprises forming supporting beams connected to the membrane on the first substrate by full etching.

25. The method of claim 24, wherein the dimensions and/or number of the supporting beams is chosen based on a trade-off between stiffness to support the membrane on the first substrate and possibility of breaking of the support beams during transfer to the acceptor substrate.

26. The method of any one of claims claim 23 to 25, wherein the acceptor substrate comprises one or more of a group consisting of CaF2, BaF2, KBr, and chalcogenide glasses.

27. The method of any one of claims claim 23 to 26, wherein the waveguide is configured to propagate light in the LWIR.

28. The method of any one of claims claim 23 to 27, wherein the waveguide device is configured as a sensor device.

29. The method of any one of claims claim 23 to 28, wherein the waveguide comprises one or more of a group consisting of silicon, germanium, and aluminum nitride.

30. A waveguide device fabricated using the method of any one of claims 23 to 29.

31. A waveguide device comprising: a waveguide formed in a material layer made from a first material on an acceptor substrate free from the first material; wherein a lateral size of the material layer is larger than 1mm x 1mm.