Towards Integrated Plasmonic Gas Sensors in the MWIR

Optical measurement approaches have proven to provide intrinsic selectivity and the sensitivity, required for the development of integrated gas sensors. In an ongoing project, we work towards a Si-photonics non-dispersive infrared gas sensor and are investigating the possibility of the incorporation of IR-plasmonic materials, which could allow an increase in sensitivities and reduce the size of such sensors. Here we present the basic concept and discuss in some detail first results concerning fabrication and characterization of the plasmonic properties.


Introduction
Integrated environmental sensing for personal health-care monitoring is a topic of increasing interest and has triggered much research towards full integrated sensor solutions. In this context, optical spectroscopic measurement approaches in the infrared region can provide intrinsic selectivity and sensitivity, as required for the development of integrated gas sensors. In an ongoing project, we work towards a Si-photonics non-dispersive infrared gas sensor and are investigating the possibility of the incorporation of IR-plasmonic materials, which could allow an increase in sensitivities and reduce the size of such sensors. Here, we will first present the overall idea, which consists of the combination of pillar photonic crystal waveguides with plasmonic elements to provide maximal interaction with gaseous analytes [1]. Then, we describe the characterization of the very first test structures, which were fabricated. Reflectivity measurements on grating structures allow the detailed characterization of the plasmon resonances, which can also be related to theoretical estimations and FEM simulations.
The basic concept, which we investigate here in the context of a miniaturized sensitive integrated gas sensor, is based on the idea of combining plasmonic propagation bound to a conductive surface with propagation within a pillar type photonic crystal (PhC) structurebased waveguide [2], which enables slow group velocity, wavelength selectivity and strong interaction with the gas to be measured. The conceptual structure is shown in Figure 1.
The photonic crystal waveguide is based on dielectric pillars made from Si. Since the gas can freely penetrate between the pillars, strong overlap between the photonic mode and the analyte is expected. Furthermore, the slower light propagation effectively increases interaction time and thus sensitivity. However, detailed simulations reveal that Eng. Proc. 2021, 6, 90 2 of 4 conventional two-dimensional dielectric PhC waveguides would require a prohibitively high aspect ratio to efficiently confine the guiding mode in the vertical direction. By combining a PhC waveguide and surface plasmon polaritons (SPPs), the proposed system efficiently confines the optical mode vertically while benefiting from the lateral confinement enabled by PhC structures.
Eng. Proc. 2021, 6, 90 2 of 4 and the analyte is expected. Furthermore, the slower light propagation effectively increases interaction time and thus sensitivity. However, detailed simulations reveal that conventional two-dimensional dielectric PhC waveguides would require a prohibitively high aspect ratio to efficiently confine the guiding mode in the vertical direction. By combining a PhC waveguide and surface plasmon polaritons (SPPs), the proposed system efficiently confines the optical mode vertically while benefiting from the lateral confinement enabled by PhC structures. Within this work, we wanted to demonstrate the feasibility of fabrication as well as to characterize possible plasmonic materials to be used.

Fabrication of the Structures
In the initial tests, two structures were realized. On the one hand, fabrication of the pillar type PhCs was established and on the other hand we fabricated simple test structures, to be able to validate the SPP-properties of different plasmonic materials.
The proposed structures were fabricated on 8-inch silicon (Si) substrates in the cleanroom facilities of Infineon Technologies Austria AG in Villach. An oxide layer with a thickness of about 2 μm was deposited to decouple the waveguide from the substrate. Then a doped polycrystalline Si layer was deposited via low-pressure chemical vapor deposition (LPCVD) on top. Afterwards the structures were etched using a standard Bosch etch process [2]. For the pillars, the Si layer had a height of roughly 4 μm and the etch process was performed over the whole thickness. For the grating structures, a Sithickness of 600 nm was chosen and by varying the etching times, gratings with different depths were prepared. For tests with metals, the plasmonic material with a thickness of about 100 nm was added by a sputtering process to ensure good sidewall coverage.
More details on the fabrication process are given in [3]. Representative SEM images of the structures are shown in Figure 2. Within this work, we wanted to demonstrate the feasibility of fabrication as well as to characterize possible plasmonic materials to be used.

Fabrication of the Structures
In the initial tests, two structures were realized. On the one hand, fabrication of the pillar type PhCs was established and on the other hand we fabricated simple test structures, to be able to validate the SPP-properties of different plasmonic materials.
The proposed structures were fabricated on 8-inch silicon (Si) substrates in the cleanroom facilities of Infineon Technologies Austria AG in Villach. An oxide layer with a thickness of about 2 µm was deposited to decouple the waveguide from the substrate. Then a doped polycrystalline Si layer was deposited via low-pressure chemical vapor deposition (LPCVD) on top. Afterwards the structures were etched using a standard Bosch etch process [3]. For the pillars, the Si layer had a height of roughly 4 µm and the etch process was performed over the whole thickness. For the grating structures, a Si-thickness of 600 nm was chosen and by varying the etching times, gratings with different depths were prepared. For tests with metals, the plasmonic material with a thickness of about 100 nm was added by a sputtering process to ensure good sidewall coverage.
More details on the fabrication process are given in [4]. Representative SEM images of the structures are shown in Figure 2.
Eng. Proc. 2021, 6, 90 2 of 4 and the analyte is expected. Furthermore, the slower light propagation effectively increases interaction time and thus sensitivity. However, detailed simulations reveal that conventional two-dimensional dielectric PhC waveguides would require a prohibitively high aspect ratio to efficiently confine the guiding mode in the vertical direction. By combining a PhC waveguide and surface plasmon polaritons (SPPs), the proposed system efficiently confines the optical mode vertically while benefiting from the lateral confinement enabled by PhC structures. Within this work, we wanted to demonstrate the feasibility of fabrication as well as to characterize possible plasmonic materials to be used.

Fabrication of the Structures
In the initial tests, two structures were realized. On the one hand, fabrication of the pillar type PhCs was established and on the other hand we fabricated simple test structures, to be able to validate the SPP-properties of different plasmonic materials.
The proposed structures were fabricated on 8-inch silicon (Si) substrates in the cleanroom facilities of Infineon Technologies Austria AG in Villach. An oxide layer with a thickness of about 2 μm was deposited to decouple the waveguide from the substrate. Then a doped polycrystalline Si layer was deposited via low-pressure chemical vapor deposition (LPCVD) on top. Afterwards the structures were etched using a standard Bosch etch process [2]. For the pillars, the Si layer had a height of roughly 4 μm and the etch process was performed over the whole thickness. For the grating structures, a Sithickness of 600 nm was chosen and by varying the etching times, gratings with different depths were prepared. For tests with metals, the plasmonic material with a thickness of about 100 nm was added by a sputtering process to ensure good sidewall coverage.
More details on the fabrication process are given in [3]. Representative SEM images of the structures are shown in Figure 2.

Plasmonic Characterization
Characterization of the plasmonic properties was done with reflective measurements on the grating structures. The schematic setup is shown in Figure 3a. The beam of a Quantum Cascade laser (QCL, MIRcatTM, DRS Daylight Solution), which was tunable in the range around 4.2 µm, was guided to the grating sample and the intensity of the reflected beam was measured with an MCT detector. The sample was mounted on a rotation stage, to adjust the incoming angle of the beam. The laser was linearly polarized with the polarization perpendicular to the grooves of the grating.

Plasmonic Characterization
Characterization of the plasmonic properties was done with reflective measurements on the grating structures. The schematic setup is shown in Figure 3a. The beam of a Quantum Cascade laser (QCL, MIRcatTM, DRS Daylight Solution), which was tunable in the range around 4.2 μm, was guided to the grating sample and the intensity of the reflected beam was measured with an MCT detector. The sample was mounted on a rotation stage, to adjust the incoming angle of the beam. The laser was linearly polarized with the polarization perpendicular to the grooves of the grating. Measurements were typically performed for different angles in the range from 24-30° and over a wavelength range of 4.0 μm-4.3 μm. The reflected intensities were referenced to the reflection spectrum obtained from a flat Au-coated Si substrate.

Results and Discussion
The measurements reported here were performed on a shallow grating with a depth of 50 nm, coated with a 100 nm Ag-layer. Results are shown in Figure 4. A clear resonance dip from the plasmon resonance can be observed, the position of which varies continuously with the reflection angle. Measurements were typically performed for different angles in the range from 24-30 • and over a wavelength range of 4.0 µm-4.3 µm. The reflected intensities were referenced to the reflection spectrum obtained from a flat Au-coated Si substrate.

Results and Discussion
The measurements reported here were performed on a shallow grating with a depth of 50 nm, coated with a 100 nm Ag-layer. Results are shown in Figure 4. A clear resonance dip from the plasmon resonance can be observed, the position of which varies continuously with the reflection angle.
On the long wavelength side, starting at a wavelength from around 4.2 µm, the data are quite noisy, which is caused by the absorption band of CO 2 , which peaks at around 4.26 µm. Since measurements were performed in ambient air, with an overall pathlength of about 100 cm, the absorption caused by CO 2 is significant. Nevertheless, in this region it is possible to follow the SPP resonance dip.
The position of the resonance as a function of incidence angle is shown in Figure 4b and closely follows the theoretical prediction except for a constant offset, probably caused by uncertainties in the zero calibration of the angle. Figure 4c shows a zoom on the SPP resonance for the case of 26 • incidence angle. It has a width of less than 5 nm. This is in good agreement with simulations, which predict about 2-3 nm width for a grating depth of 50 nm and which can also well describe the shape of the signal.
Overall, the results predict good performance for Ag layers in the mid-IR range. In addition, more detailed investigations including different metals and structures have meanwhile also been performed [5,6]. enced to the reflection spectrum obtained from a flat Au-coated Si substrate.

Results and Discussion
The measurements reported here were performed on a shallow grating with a depth of 50 nm, coated with a 100 nm Ag-layer. Results are shown in Figure 4. A clear resonance dip from the plasmon resonance can be observed, the position of which varies continuously with the reflection angle.

Conclusions
We have reported the very first results in the context of a novel integrated sensing concept, which combines PhC waveguides with SPP propagation. Reflection measurements on Ag-coated grating test structures revealed narrow SPP resonances, which is in good agreement with simulation, indicating favorable properties for mid-IR plasmonic sensors. We are confident that approaches incorporating plasmonic structures will significantly extend the range of possibilities in the field of integrated infrared sensors.
Funding: This work was performed within the PICASSO-project funded by the BMK in the framework of the program "Produktion der Zukunft" (Prj. Nr. 871417).