In this paper, we systematically investigated tailoring bolometric properties of a proposed heat-sensitive TiOx/Ti/TiOx tri-layer film for a waveguide-based bolometer, which can play a significant role as an on-chip detector operating in the mid-infrared wavelength range for the integrated optical gas sensors on Ge-on-insulator (Ge-OI) platform. As a proof-of-concept, bolometric test devices with a TiOx single-layer and TiOx/Ti/TiOx tri-layer films were fabricated by varying the layer thickness and thermal treatment condition. Comprehensive characterization was examined by the scanning transmission electron microscopy (STEM), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS) analyses in the prepared films to fully understand the microstructure and interfacial properties and the effects of thermal treatment. Quantitative measurements of the temperature- and time-dependent resistance variations were conducted to deduce the minimum detectable change in temperature (ΔTmin) of the prepared films. Furthermore, based on these experimentally obtained results, limit-of-detection (LoD) for the carbon dioxide gas sensing was estimated to demonstrate the feasibility of the proposed waveguide-based bolometer with the TiOx/Ti/TiOx tri-layer film as an on-chip detector on the Ge-OI platform. It was found that the LoD can reach ∼3.25 ppm and/or even lower with the ΔTmin of 11.64 mK in the device with the TiOx/Ti/TiOx (47/6/47 nm) tri-layer film vacuum-annealed at 400 °C for 15 min, which shows great enhancement of ∼7.7 times lower value compared to the best case of TiOx single-layer films. Our theoretical and experimental demonstration for tailoring bolometric properties of a TiOx/Ti/TiOx tri-layer film provides fairly useful insight on how to improve LoD in the integrated optical gas sensor with the bolometer as an on-chip detector.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
To date, significant advancements in the optical gas sensor at the mid-infrared (mid-IR) spectral range have been made, which has attracted great attention for chemical and biological sensing applications due to the strong absorption of various gases at the mid-IR region . Especially for the integrated optical gas sensors that can be manufactured using the mature complementary metal-oxide-semiconductor (CMOS) compatible processes , over the past few years, a variety of integrated photonics platforms based on the Ge have been proposed, which has a broad optical transparency window up to 15 μm , such as Ge-on-Si [4,5], Ge-on-Si-on-insulator (Ge-on-SOI) [6,7], Ge-on-silicon nitride [8,9], SiGe on Si , and Ge-rich SiGe .
Recently, Ge-on-insulator (Ge-OI)-based integrated photonics platforms at mid-IR region have been extensively explored [12–14], which could offer realization of strong optical confinement resulting in the tight bending radius and compact waveguides owing to the larger refractive index difference between Ge core and insulator (e.g., SiO2, Y2O3, BaF2, and CaF2, etc.) as a dielectric underneath the Ge [13,15] compared to the other reported platforms. Furthermore, Ge-OI structure has been of great interest about Ge CMOS transistors for better device performance than conventional Si transistors stemming from its higher carrier mobility and lower carrier effective mass , which can be integrated with the photonic devices, exploiting monolithic 3-D (M3D) integration technologies [17,18] for realizing versatile functionalities in a single chip. Consequently, it should be noted that Ge-OI-based integrated optical gas sensor would be an ideal candidate for high-performance on-chip gas detection with the abovementioned compelling strengths of the Ge-OI structures.
From a practical perspective, in order to realize the ultra-compact and highly-complete Ge-OI-based optical gas sensor, monolithic integration of the mid-IR light source and detector, which can be operated at high temperature (at least room temperature), is necessary. Especially for mid-IR detection, however, conventional photodetector requires cooling to a quite low temperature to guarantee the operation with a reasonable noise level. To avoid this, there have been substantial investigations for thermal detectors as an alternative approach to conventional photon detectors. Bolometers are a type of thermal detectors exploiting the photothermal conversion , that is, temperature variation by heat generation resulting from the absorbed incident light alters the electrical resistance of the heat-sensitive bolometric material, which can be measured by applying bias voltage or current. Apart from the photon detectors, due to the inherent operating principle of the bolometer, which is thermal-to-electrical conversion, both wavelength-independent and room temperature operation could be obtained for the various mid-IR sensing applications. The temperature coefficient of resistance (TCR) characterizes the extent of the resistance variation with an increase/decrease in temperature, defined as20], germanium silicon oxide (GexSiyO1-x-y) , germanium-tin alloy (Ge1-xSnx) , vanadium oxide (VOx) , and vanadium tungsten oxide (VWOx) . Particularly, in recent years, titanium oxide (TiOx) has been extensively explored as a suitable material due to its great physical and chemical properties [25–28]. Moreover, it was found that the structural and electrical properties of the TiOx films are influenced by the deposition parameters such as deposition method, substrate temperature, and annealing condition by manipulating the oxygen stoichiometry and the crystalline phase in the films [28–30].
Meanwhile, in consideration of improving the limit-of-detection (LoD), which is one of the crucial figure-of-merits (FoMs) for the gas sensors, it is essentially required to optimize the overall bolometric properties (i.e., resistivity, TCR, and noise characteristics), not simply to increase the TCR, for adopting the bolometer as an on-chip detector in the optical gas sensor. However, to the best of our knowledge, so far, no report has been published concerning tailoring bolometric properties for optical gas sensing applications. One of the dominant performance parameters associated with the LoD is the minimum detectable change in temperature (ΔTmin) of the designed bolometer, which quantifies the thermal sensitivity of the detector. Here, we note that the ΔTmin of TiOx-based bolometer could be readily tailored by employing the multi-layer synthesis techniques in the proposed TiOx/Ti/TiOx tri-layer films depending on the thickness of each layer and the thermal treatment conditions.
In this paper, the relationship between the bolometric properties and the deposition parameters in the heat-sensitive TiOx/Ti/TiOx tri-layer films for the bolometer with varying the layer thickness and thermal treatment conditions is investigated in detail, which can play a significant role as an on-chip detector for the integrated optical gas sensors on Ge-OI platform. Furthermore, LoD would be estimated based on the simulation and experimental results to demonstrate the feasibility of the proposed on-chip detector with TiOx/Ti/TiOx tri-layer film in the Ge-OI-based optical gas sensor, considering carbon dioxide (CO2) as a target gas. We believe that these systematic investigations could provide useful guidance on improving the LoD of the integrated optical gas sensor with the bolometer as an on-chip detector.
2. Concept of integrated optical gas sensor
2.1 Ge-on-insulator (Ge-OI)-based integrated optical gas sensing platform
A variety of approaches have been suggested for the operating principles of integrated optical gas sensors , such as gas sensors based on Raman spectroscopy , refractive index unit , cavity ring-down spectroscopy , photothermal spectroscopy , and infrared absorption spectroscopy . Figure 1(a) shows the conceptual configuration of a Ge-OI-based fully integrated optical gas sensing platform exploiting infrared absorption spectroscopy. For on-chip detection of multiple target gases, as shown in Fig. 1(a), light is coupled from an integrated mid-IR light source (e.g., quantum cascaded laser  and interband cascaded laser ) with a passive spectrometer (e.g., arrayed waveguide grating ) to waveguides (e.g., slot  and spiral waveguide ) where light-gas interaction occurs. The mid-IR light source with a spectrometer can be replaced by a tunable mid-IR laser source  for a more densely integrated sensor system. Considering the slot waveguides with air clad, there would be a strong interaction between the confined light in the slot region and the target gas, which results in the absorption of incident light depending on the amount of the gas concentration. Afterward, the optical power of the transmitted light could be measured by the detectors, as shown in Fig. 1(a), which might be arranged after the slot-to-strip converters .
As previously mentioned, the thermal detectors, especially for the bolometer, based on the photothermal conversion, can play a great role in mid-IR detection in the optical gas sensor as an alternative solution for the conventional photon detectors. Additionally, a waveguide-based bolometer would be a suitable configuration, which can be monolithically integrated with the proposed optical gas sensing platform for its on-chip detector. Until now, there have been few demonstrations of the waveguide-based bolometer, which were both based on the SOI platform, with the amorphous silicon and metallic nanostructures acting as the bolometric material and light absorber, respectively [42,43]. As a consequence, as shown in Fig. 1(b), we suggest a waveguide-based bolometer with TiOx/Ti/TiOx tri-layer film as an on-chip detector for integrated optical gas sensors on the Ge-OI platform. For the absorption of incident light resulting in a heat generation, a heavily doped (p+) Ge waveguide within the bolometer region is adopted, exploiting the mechanism of free-carrier absorption in Ge , which could provide a quite simple structure of the waveguide-based bolometer without introducing any additional nanostructures. The extent of free-carrier absorption, which results in heat generation, varies markedly depending on the type of dopant, level of doping concentration, and wavelength of the light. For the holes in Ge, an increase in the absorption coefficient is greater than that of the electrons in Ge, and it becomes much larger for the higher doping concentration .
2.2 Simulation for waveguide-based bolometer
In order to verify the feasibility of the proposed waveguide-based bolometer with a thermally sensitive TiOx-based bolometric material, heat simulation was performed. First of all, to model the heat source from the absorbed optical power within the p+ Ge waveguide region, we simulated the heating effect of the photothermal conversion in the designed waveguide structure, which supports only the fundamental TE mode, using 3-D finite-difference time-domain (FDTD) simulation (Lumerical FDTD solutions). It was considered that an input light with an optical power of 10 mW at 4.23 μm due to the strong absorption peak of CO2 at that wavelength was incident into the bolometer region, which is the p+ Ge waveguide. Subsequently, the optical absorption data obtained from the FDTD simulation was imported as a heat source to the heat simulation (Lumerical HEAT solutions) and the temperature response within the bolometric material was calculated.
As a proof-of-concept, the p+ Ge waveguide with a doping concentration of 1019 cm-3 (absorption coefficient was calculated as ∼873.85 cm-1 with the Ref. to ) on the Ge-OI platform with the buried oxide of 2-μm-thick SiO2 was designed with the geometrical parameters of 500 nm, 3 μm, and 10 μm for its thickness, width, and length, respectively, as depicted in Figs. 2(a) and 2(b). Also, an insulator of 20-nm-thick Al2O3 for hindering any undesirable current flow through the Ge waveguide, bolometric material of 100-nm-thick TiO2, and electrode of Ti/Au (40/60 nm) was arranged on the top of the p+ Ge waveguide region. The electrode separation was set to be 5 μm, since the bolometric test devices used for the experimental measurements in this work have the electrode separation of 5 μm, as shown in the inset of Fig. 9(a). The bolometric material was assumed as a 100-nm-thick TiO2 single-layer film for simplicity in modeling. Figure 2 shows the results of heat simulation, which represent a thermal efficiency (K/mW) of the designed waveguide-based bolometer, defined as the amount of temperature increase depending on the incident optical power. The background environment temperature and convective heat coefficient of air were set to be 293 K and 10 W/(m2·K), respectively. The thermal properties of the materials used in this simulation are indicated in Table 1. An increase in temperature is observed along the given waveguide, as shown in Fig. 2(a), and the generated heat is highly localized within the p+ Ge waveguide region. After that, the heat is conducted to the TiO2 layer, as shown in Fig. 2(b) indicating the cross-section view along the propagation direction. It was found that the thermal efficiency of the designed waveguide-based bolometer was 9.312 K/mW, considering the average temperature rise inside the TiO2 layer at this simulation. Taking account of the overall simulation results, it could be a viable and effective approach for on-chip mid-IR detection to adopt the waveguide-based bolometer with a TiOx-based bolometric material and p+ Ge waveguide region for the light absorption.
3. Materials and fabrication
3.1 Deposition of films
Figure 3 shows the fabrication process flow of the bolometric test devices with TiOx single-layer (as a reference) and TiOx/Ti/TiOx tri-layer film. A lightly Ga-doped p-type Ge(100) substrate with a concentration of ∼1×1016 cm-3 was prepared with a standard wafer cleaning process with acetone, methanol, and isopropyl alcohol. We used the lightly doped Ge substrate for the convenience of the fabrication process. However, ion implantation should be included to form the heavily doped (p+) Ge waveguide region for the heat generation resulting from the light absorption, considering the fabrication of the waveguide-based bolometer on the Ge-OI platform. We deposited 20 nm-think Al2O3 as an insulator layer at 250°C by atomic layer deposition (ALD) after pre-treatment of the sample in a diluted 1:10 HF and deionized (DI) water solution for 30 seconds. It was deposited to prevent unwanted leakage currents flowing through the Ge substrate and localize the current flow within the bolometric materials, TiOx single-layer, and TiOx/Ti/TiOx tri-layer film in this work.
Subsequently, the proposed bolometric materials consisting of TiOx single-layer (∼100 nm) or TiOx/Ti/TiOx tri-layer film (49/2/49 nm, 48/4/48 nm, and 47/6/47 nm; total thickness of ∼100 nm) were prepared by electron-beam (e-beam) evaporation. Each TiOx and Ti layer was deposited from Ti3O5 (99.99% purity) and Ti (99.995% purity) source, respectively, as a starting material. During deposition, the temperature inside the chamber was kept at the range between 30°C-40°C. After the deposition, some of the samples were annealed in the vacuum atmosphere (∼8×10−3 Torr) for 15 min with varying the temperature of 300, 400, and 500°C. Here, the layer thickness and thermal treatment conditions were varied for the systematic investigation about tailoring bolometric properties of the given structures. For the formation of the electrode, we deposited 40-nm-thick Ti metal contact and 60-nm-thick Au electrode onto the patterned samples by e-beam evaporation, followed by a lift-off technique for the removal of the photoresist mask. The total thickness as 100 nm of the films was determined to easily compare the ratio of the thickness of the metallic Ti layer in the prepared films. But, it can be tailored for the desired level of resistance of the fabricated device, considering some issues for the thicker film. The thicker the total thickness of the film with the identical thickness ratio of each TiOx and Ti layer, the more thermal energy is required for the sufficient interdiffusion of metallic Ti atoms during the thermal treatment process. Moreover, the heat generated from the p+ Ge waveguide region might not be uniformly conducted inside the prepared films, and consequently, the thermal efficiency of the proposed waveguide-based bolometer would be lower. The notation for fabricated bolometric test devices with different geometries and annealing conditions used in this paper is represented in Table 2.
In order to fully understand the microstructure of the proposed TiOx/Ti/TiOx tri-layer film and the interface properties of TiOx/Ti and Ti/TiOx with a fundamental insight into the influence on a thermal treatment, the scanning transmission electron microscopy (STEM), and energy-dispersive X-ray spectroscopy (EDX) line scan analyses were conducted. The specimens were prepared with the focused ion beam (FIB) technique.
Figures 4(a) and 4(d) show the cross-sectional STEM images of the fabricated devices with TiOx/Ti/TiOx (47/6/47 nm) tri-layer films, which are as-deposited (T3-0) and vacuum-annealed at 400°C for 15 min (T3-2), respectively. As shown in Fig. 4(a), the upper and lower TiOx layer appears as dark, and the metallic Ti layer as bright within the prepared TiOx/Ti/TiOx tri-layer film, which is observed in more detail at the enlarged STEM image with the abrupt material interfaces of TiOx (Ti) and Ti (TiOx) layers shown in Fig. 4(b). The microstructure of the device was also confirmed by the EDX elemental mapping patterns of Ti (red), O (magenta), Al (green), and Ge (cyan), as shown in Fig. 4(b). Figure 4(c) depicts the measured EDX line scan profiles showing the atomic percent of Ti, O, Al, Ge, and C atoms along the line drawn in Fig. 4(b), where the metallic Ti layer is clearly distinguishable from the upper and lower TiOx layer within the prepared tri-layer film, due to the abrupt changes in the atomic percent of Ti and O at the interface between TiOx (Ti) and Ti (TiOx) layers.
For the annealed sample (T3-2), however, an interdiffusion of metallic Ti atoms towards the upper and lower layer was observed, as shown in Figs. 4(d)–4(f). As depicted in Figs. 4(d) and 4(e), no seemingly obvious interface was observed within the obtained tri-layer film from the STEM images. It is also in agreement with the results of EDX line scan profiles of Fig. 4(f), showing no abrupt change in the atomic percent of Ti and O, contrary to the as-deposited sample (T3-0). Moreover, the interdiffusion of the metallic Ti atoms can contribute to the formation of bonding with the defective oxygen (Ti-O bonding) within the TiOx/Ti/TiOx tri-layer film depending on the thermal treatment conditions, which is further discussed in the results of X-ray photoelectron spectroscopy (XPS) analysis shown in Fig. S3 in Supplement 1. Also, it should be noted that the polycrystalline structures were observed in the annealed sample at the high-resolution STEM images as shown in Figs. 4(g) and 4(h), which is consistent with the results of X-ray diffraction (XRD) analysis in Fig. S1 in Supplement 1, and the correlation between the amorphous-to-polycrystalline phase transition and the electrical resistivity in the prepared films will be discussed later in more detail in Fig. 8.
4. Experiments and results
4.1 Temperature dependence of resistance
In order to characterize the sheet resistance (Rsh) and specific contact resistivity (ρc) in the prepared bolometric test devices, circular transmission line method (CTLM) patterns with the gap spacing (d) of 5, 10, 15, 20, 40, and 60 μm, as depicted in Fig. 5(a), were fabricated. The total resistance values of the CTLM patterns depending on the gap spacing were measured by the four-point-probe system with Keithley’s 4200-SCS parameter analyzer with varying the substrate temperature from 293 K to 343 K in 5 K step (ΔT = 5 K) using a thermo-electric cooler (TEC) controller to precisely control the temperature. As observed in Figs. 5(b)–5(i) and Fig. S4 of Supplement 1, the slope of linearly fitted curves determining the Rsh declines steadily in all of the fabricated devices with an increase in substrate temperature, regardless of the material composition and annealing conditions. It implies that the proposed TiOx/Ti/TiOx tri-layer films possess the negative TCR (-%/K) irrespective of the deposition parameters, same as the TiOx single-layer film which has been widely studied [26–30]. However, the extent of the resistance variation with an increase/decrease in the substrate temperature depends on the fabricated devices, which indicates that the bolometric properties of the TiOx/Ti/TiOx tri-layer film can be simply tailored depending on the material composition and thermal treatment conditions. The experimental results of the temperature-dependent total resistance values with varying gap spacing for TiOx/Ti/TiOx (49/2/49 nm, 48/4/48 nm) tri-layer films are displayed in Fig. S4 in Supplement 1. From these measurements, we extracted the ρc values (see Supplement 1, Fig. S5) and Rsh in the prepared bolometric test devices.
Figure 6 represents the temperature-dependent Rsh in the fabricated devices. As mentioned above, it is obvious that the values of Rsh vary inversely with the change of temperature in all of the fabricated devices, which confirms that the obtained films have a typical semiconducting behavior following the Arrhenius relationship , expressed as2), the activation energies ΔE were extracted from the slope of linear characteristics, i.e., Arrhenius plots (ln(Rsh) vs. 1000/T), as shown in Fig. 7.
Here, from Eq. (3), it is clear that the absolute value of TCR is proportional to the activation energy level.
Figure 8 represents the experimental results of Rsh and TCR at 293 K depending on the thermal treatment conditions in the fabricated devices. As shown in Fig. 8(a), there is a clear downward trend in Rsh values with an increment in the thickness of the Ti layer, regardless of the thermal treatment conditions in the prepared films. Consequently, in light of this consistent trend, increasing the ratio of the Ti layer within the TiOx/Ti/TiOx tri-layer film reduces the Rsh value of the film. Also, from the results in Fig. 8(a), we noted that the TiOx/Ti/TiOx tri-layer films possess much lower Rsh values even after the sufficient thermal treatment, which are several orders of magnitude, compared to the TiOx single-layer films. In consequence, we assumed that a minor amount of remaining metallic titanium content (i.e., Ti0+ state) and multiple-valence titanium oxide with the presence of Ti2+, Ti3+, and Ti4+ states  might be co-existing in the TiOx/Ti/TiOx tri-layer film, implying that the mixed-phase of Ti and TiOx in the prepared films, contrary to the TiOx single-layer films which predominantly contain the Ti4+ state, as confirmed in the results of XPS analysis (see Supplement 1, Fig. S2). In order to identify the presence of mixed-phase in the prepared TiOx/Ti/TiOx tri-layer films, XPS depth-profile analysis was carried out. Despite the observed reduction and degradation during Ar+-ion etching with the alteration of chemical state and material composition, it was reasonable to identify the mixed-phase of Ti and TiOx in the prepared films (see Supplement 1, Fig. S3 for detailed information about XPS depth-profile analysis). As a consequence, the relative ratio of multiple-valence states and metallic Ti content in the titanium oxide films could contribute greatly to the bolometric properties of the films and the LoD for the optical gas sensor, as will be discussed later.
Variation of Rsh and TCR depending on the thermal treatment conditions is expressed in Figs. 8(b)–8(e), which confirms that the bolometric properties are strongly dependent on the deposition parameters and thickness of the Ti layer and shows the correlation between Rsh and TCR. For the case of TiOx single-layer film, as shown in Fig. 8(b), it was found that the annealed samples have lower Rsh values compared to the as-deposited sample and Rsh values are on the decline as the annealing temperature increases. It might be due to the amorphous-to-polycrystalline phase transition and the further improved crystallinity in TiOx thin film with increasing the annealing temperature , as confirmed in the results of XRD analysis in Fig. S1 in Supplement 1. In addition, as the annealing temperature increases, the lower values of TCR were obtained, which results from the decreased ΔE values  (summarized in Table S1 in Supplement 1) extracted from Arrhenius plots shown in Fig. 7(a).
On the contrary, as shown in Figs. 8(c)–8(e), it was found that the changing tendency of Rsh with varying the annealing temperature in the TiOx/Ti/TiOx tri-layer film differs markedly from the trend in TiOx single-layer film. In order to interpret the relationship between Rsh and annealing temperature, both (i) crystallinity improvement and (ii) interdiffusion of the metallic Ti atoms, which can contribute to the formation of Ti-O bonding within the TiOx/Ti/TiOx tri-layer film, should be considered (see Supplement 1, Fig. S3 shows evidence for the interdiffusion of the metallic Ti atoms with the formation of Ti-O bonding).
For the TiOx/Ti/TiOx tri-layer films annealed at 300°C, slight decreases of Rsh were consistently observed compared with the as-deposited samples in all of the devices (T1-1, T2-1, and T3-1). It indicates that the amorphous-to-polycrystalline phase transition would be a predominant factor influencing a rise in the electrical conductivity rather than the interdiffusion of the metallic Ti atoms, which could result in the formation of bonding with the defective oxygen within the prepared films decreasing the electrical conductivity, or there might be not sufficient thermal energy for the formation of Ti-O bonding. As the samples were annealed at 400°C, steep rises in Rsh were observed in all of the devices (T1-2, T2-2, and T3-2). The increased Rsh is mainly originated from the predominance of interdiffusion of the metallic Ti atoms with the formation of Ti-O bonding as increasing the annealing temperature, rather than the contribution of crystalline properties. In contrast, it appears that the enhanced crystalline structures dominantly affect a decrease in the electrical resistivity for the annealing temperature of 500°C (T1-3 and T2-3), with the exception of the sample with the 6-nm-thick Ti layer (T3-3). The Rsh value of the T3-3 sample remains on an upward trend since the interdiffusion of the metallic Ti atoms with the formation of Ti-O bonding would be still ongoing due to its relatively thicker Ti layer. Additionally, as shown in Figs. 8(c)–8(e), for the TiOx/Ti/TiOx tri-layer films, a variation in the TCR accords closely with the aspects of change in Rsh values except for the samples of T1-3 and T2-3, which could be an effective approach to possess a relatively high absolute value of TCR (T1-3; -2.955%/K, T2-3; -2.574%/K) while reducing the electrical resistivity. This improvement in the TCR might be explained by compensation of the oxygen vacancy donor defects from the titanium acceptor defects within the prepared films during thermal treatment at the high annealing temperature, which can cause a downward shift of the Fermi level with the increased ΔE value , and further studies for such an opposite trend are required to comprehend these results more fully.
In light of these experimental results shown in Fig. 8, it can be concluded that increasing the annealing temperature clearly contributes to the interdiffusion of the metallic Ti atoms with the formation of Ti-O bonding within the TiOx/Ti/TiOx tri-layer films, as confirmed in the STEM with EDX line scan profiles shown in Fig. 4 and the results of XPS depth-profiling shown in Fig. S3 in Supplement 1. Moreover, indeed by increasing the thickness of the Ti layer, higher thermal energy should be required to further promote the formation of Ti-O bonding. Hence, it is worth noting that not only the thermal treatment condition but also the thickness of the Ti layer plays a crucial role in tailoring bolometric properties of the TiOx/Ti/TiOx tri-layer films.
4.2 Time dependence of resistance
To assess the ΔTmin of the designed bolometer with the TiOx single-layer and TiOx/Ti/TiOx tri-layer films, which is strongly associated with the LoD of the optical gas sensor, time-dependent resistance variations were measured in the fabricated bolometric test devices with the pad separation of 5 μm, as shown in the inset of Fig. 9(a).
Prior to the time-dependence measurement, the current-voltage (I-V) characteristics were examined. Figure 9 shows the I-V characteristic curves for Ti/Au contacts on the TiOx single-layer and TiOx/Ti/TiOx tri-layer films in the fabricated devices with the thermal treatment, which were measured while sweeping the voltage from -1 V to +1 V with 0.01 V step at 293 K. As shown in Fig. 9, it was found that the linear I-V characteristics were observed in most of the devices with TiOx/Ti/TiOx tri-layer films, indicating Ohmic behaviors were obtained at the interface between the TiOx/Ti/TiOx tri-layer film and the Ti/Au electrode with comparatively low ρc values. The lowest ρc value obtained was 1.199×10−5 Ω·cm2 in the device with TiOx/Ti/TiOx (48/4/48 nm) tri-layer film annealed at 300°C (T2-1). On the contrary, samples with the TiOx single-layer films (S-1, S-2, and S-3) and TiOx/Ti/TiOx (49/2/49 nm) tri-layer films annealed at 400°C (T1-2) and 500°C (T1-3) exhibited nonlinear I-V characteristics with the particularly large ρc values (Supplement 1, Fig. S5), which implies that there would be Schottky contact with a relatively large barrier height  between the prepared bolometric material and the Ti/Au electrode.
The smallest amount of change in temperature that can be detected by measurement with the designed bolometer can be quantified as the ΔTmin (K), which can be determined by the resistance fluctuation at a given temperature and the resistance-temperature (R-T) characteristics. Figure 10 describes schematically how to experimentally estimate the ΔTmin in the fabricated devices. While the resistance values of a bolometric material at an environment of fixed temperature are measured for a certain time, there would be a fluctuation of the measured resistance with a standard deviation (±σ) in the resistance noise distribution, which can be converted to a temperature variation, i.e., ΔTmin, based on the R-T characteristics. In other words, it is not a single point measurement, but rather a temporal measurement of the resistance noise. Hence, we conducted the time-dependent measurements of resistance variation to assess the ΔTmin in the fabricated devices. The 1000 points of resistance values are measured with a 40 Hz rate at a constant temperature of 293 K with the TEC controller. It is pretty reasonable to deduce the ΔTmin of the prepared bolometric materials with our experimental procedures since the thermal change of the given system occurs at a much larger timescale . Also, we opted for the operation bias voltage as 0.5 V for the DC measurements, which is a relatively low value applied between the two electrodes to prevent the undesirable self-heating and minimize the electrical power consumption. Meanwhile, the samples without thermal treatment (as-deposited samples) were excluded from these measurements due to the substantial resistance fluctuation, which might be ascribed to a large amount of intrinsic defect and the poor interface characteristics within the prepared films .
The measurement results of time-dependent resistance values in the fabricated devices with TiOx/Ti/TiOx (47/6/47 nm) tri-layer films at different annealing temperatures are shown in Fig. 11 (see Figs. S6, S7, and S8 in Supplement 1 for the results of TiOx (100 nm) single-layer, TiOx/Ti/TiOx (49/2/49 nm) tri-layer, and TiOx/Ti/TiOx (48/4/48 nm) tri-layer films, respectively). The average resistance values (m) are represented in each time-dependence measurement, as shown in Figs. 11(a), 11(c), and 11(e). Subsequently, we plotted histograms of the dataset from these measurements showing a deviation from the average resistance value of each point, which can be approximated by a standard normal (Gaussian) distribution with the corresponding ±σ range. Thus, the measured resistance noise was represented as a value of σ/m (%) from the histograms shown in Figs. 11(b), 11(d), and 11(f).
The variation trends of m and σ/m in all of the fabricated devices depending on the annealing temperature obtained from the time-dependence measurements shown in Figs. 11 and Figs. S6, S7, and S8 in Supplement 1 are expressed in Figs. 12(a) and 12(b), respectively. As we can see, the resistance noise (σ/m) shows a similar trend with the obtained m and ρc values, as described in Figs. 12(a) and Fig. S5 in Supplement 1, respectively. It might be ascribed to the causes of resistance fluctuation such as the resistance, contact resistivity, interface characteristics of the film, and the number of carriers contributing to the conduction process [54–56]. At last, we deduced the ΔTmin (mK) from the resistance noise (σ/m) and TCR of each prepared film shown in Figs. 12(b) and 8, respectively, which was calculated by dividing the σ/m (%) value by the absolute value of TCR (%/K). Figure 12(c) shows the results of ΔTmin evaluation at 293 K of the prepared films. It was found that the ΔTmin could reach 11.64 mK in the device with TiOx/Ti/TiOx (47/6/47 nm) annealed at 400°C (T3-2), which is ∼7.7 times lower compared to the lowest case among the devices with TiOx (100 nm) single-layer film (S-2), 89.73 mK.
4.3 Estimation of limit-of-detection (LoD)
In order to demonstrate the feasibility of the proposed waveguide-based bolometer with TiOx/Ti/TiOx tri-layer film as an on-chip detector monolithically integrated with the optical gas sensor on the Ge-OI platform, the LoD was estimated based on the experimental results in the above previous chapters. In this paper, we consider carbon dioxide (CO2) as a target gas that contains a strong absorption peak at the wavelength of ∼4.23 μm. The CO2 is one of the major greenhouse gas molecules causing global warming, which has been a severe threat to the world’s environment over the past decades . Thus, accurate and high-resolution CO2 gas sensing is essentially required, especially for the integrated optical gas sensors.
The optical absorption is dependent on the concentration of the target gas (CO2 in our case). The relationship between output optical power P (W) and the concentration of target gas C (mol·L-1) can be determined by Beer-Lambert’s law  which is expressed asSupplement 1.
The sensitivity of the optical gas sensor S is determined by the ratio between the variation in the optical power resulting from the gas absorption and in the concentration of target gas . It could be calculated by differentiating Eq. (4) with respect to C and given by
Here, the sensitivity could be improved by lowering the value of αprop. Moreover, there is an optimal length of the slot waveguide Lopt, which maximizes the value of S, expressed as
An optimal length Lopt = 1.404 cm can be calculated from Eq. (6) under low gas concentration assumption (∼100 ppm), considering the parameters as Γ = 56.69%, αprop = 3 dB/cm (0.691 cm-1), and εgas = 9200 mol-1·L·cm-1 for CO2 at 4.23 μm and at atmospheric pressure . Since the propagation losses for the slot waveguide on the Ge-OI platform at the mid-IR range have not been measured yet, we assumed the value of αprop = 3 dB/cm due to the propagation losses of a few dB/cm in the Ge-based waveguide for mid-IR integrated photonics that have been reported so far .
The minimum detectable output optical power Pmin which could be measured by the given detector of the optical gas sensor is defined by the difference between the optical power for the case of C = 0 and Cmin, written by4) and (7), expressed as
It should be noted that LoD is physically limited by the performance of the detector which determines the value of Pmin. For the proposed waveguide-based bolometer with TiOx single-layer or TiOx/Ti/TiOx tri-layer films as the given detector in the optical gas sensor, Pmin could be expressed as
The LoD for CO2 detection can thus be estimated at the wavelength of 4.23 μm using Eqs. (8) and (9) for the proposed optical gas sensor on the Ge-OI platform with the waveguide-based bolometer with TiOx single-layer or TiOx/Ti/TiOx tri-layer films as the on-chip detector. For this analysis, P0 = 10 mW and SNR = 3, which is commonly adopted 3-σ criterion [60,61], are used. From the previous heat simulation described in Fig. 2, it could be found that the value of ηth for the designed waveguide-based bolometer was 9.312 K/mW. At last, we thus decided to introduce the experimentally obtained values of ΔTmin which are shown in Fig. 12(c) for the LoD estimation in the optical gas sensor. For the fabricated bolometric test device of T3-2, TiOx/Ti/TiOx (47/6/47 nm) tri-layer film vacuum-annealed at 400°C for 15 min, the value of LoD was estimated as ∼3.25 ppm (Cmin = 1.3514×10−7 mol/L) considering the ΔTmin = 11.64 mK (at 293 K). Here, it has ∼7.7 times lower value compared to the best case among the devices with TiOx single-layer film (S-2), which is ∼25.13 ppm (Cmin = 1.0452×10−6 mol/L) considering the ΔTmin = 89.73 mK (at 293 K). The values of LoD for all of the fabricated bolometric test devices are plotted in Fig. 13. It is of great importance to note that the trends in LoD represent almost identical with the ΔTmin shown in Fig. 12(c), which suggests that tailoring ΔTmin of bolometer in the optical gas sensors plays a substantial role in determining the LoD.
Furthermore, in light of our experimental substantiation, the bolometric properties could be simply tailored depending on the material composition and thermal treatment condition in the bolometric materials of TiOx/Ti/TiOx tri-layer film. In other words, it is possible to design an on-chip detector for the optical gas sensors having the desired level of requirements such as LoD, electrical resistance, and electrical power consumption in the gas sensor considering a type of selected target gas and its larger embedded system with a readout integrated circuits (ROICs). In addition, the annealing temperature should be chosen in consideration of the device fabrication on the ROIC and the processing temperature limit for the CMOS compatibility. In this work, the fabricated bolometric test device of T3-2, TiOx/Ti/TiOx (47/6/47 nm) tri-layer film vacuum-annealed at 400°C for 15 min, shows the lowest value of LoD (ΔTmin), which is a reasonable level of temperature for the considerations. Also, if we deposit the TiOx/Ti/TiOx tri-layer film by atomic layer deposition (ALD) technique with the precisely programmed procedure, the material composition within the films could be readily customized at a much lower processing temperature. Consequently, in consideration of the overall results for the LoD estimation in the waveguide-based bolometer with the proposed TiOx/Ti/TiOx tri-layer film, we believe that it could be practically utilized as an on-chip detector for various optical gas sensing applications. Moreover, aside from the Ge-OI-based integrated photonics platform, the proposed waveguide-based bolometer can be easily adopted as an on-chip mid-IR detector on the other integrated photonics platform for optical gas sensing due to its simple structure.
Further improvements of LoD could be achieved by enhancing thermal efficiency of the waveguide-based bolometer through (1) engineering the geometrical parameters of the waveguide and electrode, (2) optimizing doping profile within the p+ Ge waveguide, and (3) adding air-trench structure below the p+ Ge waveguide by buried oxide etching resulting in the increased capability of thermal isolation. And it could be also enhanced by (4) minimizing the propagation losses in the designed slot waveguide with the sensitivity enhancement and (5) introducing some additional plasmonic nanostructures integrated with the bolometer to further increase the absorption in the highly localized region [42,43].
The overall performance characteristics in this work are summarized in Table S1 in Supplement 1.
5. Conclusion and outlook
In conclusion, we have established the relationship between the bolometric properties and the deposition parameters in the heat-sensitive TiOx/Ti/TiOx tri-layer films for a bolometer, which can play a great role as an on-chip detector operating in the mid-IR range for the Ge-OI-based integrated optical gas sensors. The bolometric test devices with TiOx single-layer and TiOx/Ti/TiOx tri-layer films were fabricated by varying each layer thickness and the thermal treatment condition for a systematic investigation about tailoring bolometric properties to improve the LoD in the gas sensors. The prepared films were comprehensively characterized by the STEM-EDX, XRD, and XPS analyses to fully understand the microstructure and interfacial properties and explore the effects of thermal treatment, resulting in the alteration of the crystalline properties and interdiffusion of metallic Ti atoms with the formation of Ti-O bonding within the TiOx/Ti/TiOx tri-layer films. Subsequently, we conducted the measurements of temperature- and time-dependent resistance variation to deduce the ΔTmin of the prepared films. Furthermore, the LoD in the proposed Ge-OI-based gas sensing platform was estimated, considering the CO2 as a target gas, based on the experimentally obtained ΔTmin values in this work. It was found that the value of LoD could reach ∼3.25 ppm with the ΔTmin of 11.64 mK in the device with TiOx/Ti/TiOx (47/6/47 nm) tri-layer film vacuum-annealed at 400°C for 15 min, showing ∼7.7 times lower value compared to the best case of TiOx single-layer film. And it can be further improved by increasing the thermal efficiency of the given waveguide-based bolometer and minimizing the propagation losses in the designed slot waveguide.
Our theoretical and experimental investigations about tailoring bolometric properties in the proposed TiOx/Ti/TiOx tri-layer film from a perspective of the gas sensing applications provide valuable insight on how to improve LoD in the integrated optical gas sensor with the bolometer as an on-chip detector. Furthermore, our work also suggests the feasibility of the Ge-OI-based integrated optical gas sensing platform with a waveguide-based bolometer for realizing the ultra-compact and high-performance on-chip gas sensor with potentially extending an operation range to a variety of gases (e.g., CH4, NOx, CO, and NH3, etc.), as well as enabling a real-time multi-gases detection due to the wavelength-insensitivity of the bolometer.
Ministry of Trade, Industry and Energy (20012263); National Research Foundation of Korea (2020R1F1A1052718); Korea Institute of Science and Technology (2E31372); BrainKorea 21 (BK21) FOUR.
The EDA Tool was supported by the IC Design Education Center.
The authors declare no conflicts of interest.
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
See Supplement 1 for supporting content.
1. R. Soref, “Mid-infrared photonics in silicon and germanium,” Nat. Photonics 4(8), 495–497 (2010). [CrossRef]
2. J. Wu, G. Yue, W. Chen, Z. Xing, J. Wang, W. R. Wong, Z. Cheng, S. Y. Set, G. S. Murugan, X. Wang, and T. Liu, “On-chip optical gas sensors based on group-IV materials,” ACS Photonics 7(11), 2923–2940 (2020). [CrossRef]
3. G. Z. Mashanovich, M. Nedeljkovic, J. Soler-Penades, Z. Qu, W. Cao, A. Osman, Y. Wu, C. J. Stirling, Y. Qi, Y. X. Cheng, L. Ried, C. G. Littlejohns, J. Kang, Z. Zhao, M. Takenaka, T. Li, Z. Zhou, F. Y. Gardes, D. J. Thomson, and G. T. Reed, “Group IV mid-infrared photonics [invited],” Opt. Mater. Express 8(8), 2276–2286 (2018). [CrossRef]
4. Y.-C. Chang, V. Paeder, L. Hvozdara, J.-M. Hartmann, and H. P. Herzig, “Low-loss germanium strip waveguides on silicon for the mid-infrared,” Opt. Lett. 37(14), 2883–2885 (2012). [CrossRef]
5. M. Nedeljkovic, J. S. Penadés, C. J. Mitchell, A. Z. Khokhar, S. Stanković, T. D. Bucio, C. G. Littlejohns, F. Y. Gardes, and G. Z. Mashanovich, “Surface-grating-coupled low-loss Ge-on-Si rib waveguides and multimode interferometers,” IEEE Photonics Technol. Lett. 27(10), 1040–1043 (2015). [CrossRef]
6. U. Younis, S. K. Vanga, A. E.-J. Lim, P. G.-Q. Lo, A. A. Bettiol, and K.-W. Ang, “Germanium-on-SOI waveguides for mid-infrared wavelengths,” Opt. Express 24(11), 11987–11993 (2016). [CrossRef]
7. S. Radosavljevic, N. T. Beneitez, A. Katumba, M. Muneeb, M. Vanslembrouck, B. Kuyken, and G. Roelkens, “Mid-infrared Vernier racetrack resonator tunable filter implemented on a germanium on SOI waveguide platform [invited],” Opt. Mater. Express 8(4), 824–835 (2018). [CrossRef]
8. W. Li, P. Anantha, S. Bao, K. H. Lee, X. Guo, T. Hu, L. Zhang, H. Wang, R. Soref, and C. S. Tan, “Germanium-on-silicon nitride waveguides for mid-infrared integrated photonics,” Appl. Phys. Lett. 109(24), 241101 (2016). [CrossRef]
9. W. Li, P. Anantha, K. H. Lee, H. D. Qiu, X. Guo, S. C. K. Goh, L. Zhang, H. Wang, R. A. Soref, and C. S. Tan, “Spiral waveguides on germanium-on-silicon nitride platform for mid-IR sensing applications,” IEEE Photonics J. 10(3), 1–7 (2018). [CrossRef]
10. L. Carletti, P. Ma, Y. Yu, B. Luther-Davies, D. Hudson, C. Monat, R. Orobtchouk, S. Madden, D. J. Moss, M. Brun, S. Ortiz, P. Labeye, S. Nicoletti, and C. Grillet, “Nonlinear optical response of low loss silicon germanium waveguides in the mid-infrared,” Opt. Express 23(7), 8261–8271 (2015). [CrossRef]
11. J. M. Ramirez, V. Vakarin, C. Gilles, J. Frigerio, A. Ballabio, P. Chaisakul, X. Le Roux, C. Alonso-Ramos, G. Maisons, L. Vivien, M. Carras, G. Isella, and D. Marris-Morini, “Low-loss Ge-rich Si0.2Ge0.8 waveguides for mid-infrared photonics,” Opt. Lett. 42(1), 105–108 (2017). [CrossRef]
12. J. Kang, M. Takenaka, and S. Takagi, “Novel Ge waveguide platform on Ge-on-insulator wafer for mid-infrared photonic integrated circuits,” Opt. Express 24(11), 11855–11864 (2016). [CrossRef]
13. S. Kim, J.-H. Han, J.-P. Shim, H.-j. Kim, and W. J. Choi, “Verification of Ge-on-insulator structure for a mid-infrared photonics platform,” Opt. Mater. Express 8(2), 440–451 (2018). [CrossRef]
14. S. Xu, K. Han, Y.-C. Huang, K. H. Lee, Y. Kang, S. Masudy-Panah, Y. Wu, D. Lei, Y. Zhao, and H. Wang, “Integrating GeSn photodiode on a 200 mm Ge-on-insulator photonics platform with Ge CMOS devices for advanced OEIC operating at 2 μm band,” Opt. Express 27(19), 26924–26939 (2019). [CrossRef]
15. D. Marris-Morini, V. Vakarin, J. M. Ramirez, Q. Liu, A. Ballabio, J. Frigerio, M. Montesinos, C. Alonso-Ramos, X. Le Roux, S. Serna, D. Benedikovic, D. Chrasdtina, L. Vivien, and G. Isella, “Germanium-based integrated photonics from near- to mid-infrared applications,” Nanophotonics 7(11), 1781–1793 (2018). [CrossRef]
16. H.-R. Lim, S. K. Kim, J.-H. Han, H. Kim, D.-M. Geum, Y.-J. Lee, B.-K. Ju, H.-J. Kim, and S. Kim, “Impact of bottom-gate biasing on implant-free junctionless Ge-on-insulator n-MOSFETs,” IEEE Electron Device Lett. 40(9), 1362–1365 (2019). [CrossRef]
17. S.-H. Kim, S.-K. Kim, J.-P. Shim, D.-M. Geum, G. Ju, H.-S. Kim, H.-J. Lim, H.-R. Lim, J.-H. Han, S. Lee, H.-S. Kim, P. Bidenko, C.-M. Kang, D.-S. Lee, J.-D. Song, W. J. Choi, and H.-J. Kim, “Heterogeneous integration toward a monolithic 3-D chip enabled by III-V and Ge materials,” IEEE J. Electron Devices Soc. 6, 579–587 (2018). [CrossRef]
18. D.-M. Geum, S. K. Kim, S. Lee, D. Lim, H.-J. Kim, C. H. Choi, and S.-H. Kim, “Monolithic 3D integration of InGaAs photodetectors on Si MOSFETs using sequential fabrication process,” IEEE Electron Device Lett. 41(3), 433–436 (2020). [CrossRef]
19. J. W. Stewart, N. C. Wilson, and M. H. Mikkelsen, “Nanophotonic engineering: a new paradigm for spectrally sensitive thermal photodetectors,” ACS Photonics 8(1), 71–84 (2021). [CrossRef]
20. C. Vedel, J.-L. Martin, J.-L. Ouvrier-Buffet, J.-L. Tissot, M. Vilain, and J.-J. Yon, “Amorphous-silicon-based uncooled microbolometer IRFPA,” Proc. SPIE 3698, 276–283 (1999). [CrossRef]
21. A. Ahmed and R. N. Tait, “Noise behavior of amorphous GexSi1-xOy for microbolometer applications,” Infrared Phys. Technol. 46(6), 468–472 (2005). [CrossRef]
22. M. Abdel-Rahman, M. Alduraibi, M. Hezam, and B. Ilahi, “Sputter deposited GeSn alloy: a candidate material for temperature sensing layers in uncooled microbolometers,” Infrared Phys. Technol. 97, 376–380 (2019). [CrossRef]
23. B. Wang, J. Lai, H. Li, H. Hu, and S. Chen, “Nanostructured vanadium oxide thin film with high TCR at room temperature for microbolometer,” Infrared Phys. Technol. 57, 8–13 (2013). [CrossRef]
24. N. Chi-Anh, H.-J. Shin, K. Kim, Y.-H. Han, and S. Moon, “Characterization of uncooled bolometer with vanadium tungsten oxide infrared active layer,” Sens. Actuators, A 123-124, 87–91 (2005). [CrossRef]
25. Y. A. K. Reddy, B. Ajitha, Y. B. Shin, I.-K. Kang, and H. C. Lee, “Influence of passivation layer on thermal stability of Nb:TiO2-x samples for shutter-less infrared image sensors,” Infrared Phys. Technol. 100, 52–56 (2019). [CrossRef]
26. M. Y. Tanrikulu, H. R. Rasouli, M. Ghaffari, K. Topalli, and A. K. Okyay, “Atomic layer deposition synthesized TiOx thin films and their application as microbolometer active materials,” J. Vac. Sci. Technol., A 34(3), 031510 (2016). [CrossRef]
27. I. Yadav, S. Jain, S. S. Lamba, M. Tomar, S. Gupta, V. Gupta, K. K. Jain, S. Dutta, and R. Chatterjee, “Effect of growth and electrical properties of TiOx films on microbolometer design,” J. Mater. Sci.: Mater. Electron. 31(9), 6671–6678 (2020). [CrossRef]
28. Y. A. K. Reddy, I.-K. Kang, Y. B. Shin, and H. C. Lee, “Bolometric properties of reactively sputtered TiO2-x films for thermal infrared image sensors,” J. Phys. D: Appl. Phys. 48(35), 355104 (2015). [CrossRef]
29. Y. A. K. Reddy, Y. B. Shin, I.-K. Kang, H. C. Lee, and P. Sreedhara Reddy, “Enhanced bolometric properties of TiO2-x thin films by thermal annealing,” Appl. Phys. Lett. 107(2), 023503 (2015). [CrossRef]
30. Y. A. K. Reddy, Y. B. Shin, I.-K. Kang, and H. C. Lee, “Substrate temperature dependent bolometric properties of TiO2-x films for infrared image sensor applications,” Ceram. Int. 42(15), 17123–17127 (2016). [CrossRef]
31. S. A. Holmstrom, T. H. Stievater, D. A. Kozak, M. W. Pruessner, N. Tyndall, W. S. Rabinovich, R. A. McGill, and J. B. Khurgin, “Trace gas Raman spectroscopy using functionalized waveguides,” Optica 3(8), 891–896 (2016). [CrossRef]
32. K. Li, J. Li, Y. Song, G. Fang, C. Li, Z. Feng, R. Su, B. Zeng, X. Wang, and C. Jin, “Ln slot photonic crystal microcavity for refractive index gas sensing,” IEEE Photonics J. 6(5), 1–9 (2014). [CrossRef]
33. W. Li, Y. Han, Z. Chen, H. Jiang, and K. Hamamoto, “Amplifier-assisted CRDS (cavity ring-down spectroscopy) toward compact breath sensing,” Jpn. J. Appl. Phys. 58(SJ), SJJD01 (2019). [CrossRef]
34. A. Vasiliev, A. Malik, M. Muneeb, B. Kuyken, R. Baets, and G. Roelkens, “On-chip mid-infrared photothermal spectroscopy using suspended silicon-on-insulator microring resonators,” ACS Sens. 1(11), 1301–1307 (2016). [CrossRef]
35. N. Koompai, P. Limsuwan, X. Le Roux, L. Vivien, D. Marris-Morini, and P. Chaisakul, “Analysis of Si3N4 waveguides for on-chip gas sensing by optical absorption within the mid-infrared region between 2.7 and 3.4 µm,” Results Phys. 16, 102957 (2020). [CrossRef]
36. A. Spott, J. Peters, M. L. Davenport, E. J. Stanton, C. D. Merritt, W. W. Bewley, I. Vurgaftman, C. S. Kim, J. R. Meyer, J. Kirch, L. J. Mawst, D. Botez, and J. E. Bowers, “Quantum cascade laser on silicon,” Optica 3(5), 545–551 (2016). [CrossRef]
37. W. W. Bewley, C. L. Canedy, C. S. Kim, M. Kim, C. D. Merritt, J. Abell, I. Vurgaftman, and J. R. Meyer, “Continuous-wave interband cascade lasers operating above room temperature at λ = 4.7-5.6 μm,” Opt. Express 20(3), 3235–3240 (2012). [CrossRef]
38. A. Malik, E. J. Stanton, J. Liu, A. Spott, and J. E. Bowers, “High performance 7 × 8 Ge-on-Si arrayed waveguide gratings for the midinfrared,” IEEE J. Sel. Top. Quantum Electron. 24(6), 1–8 (2018). [CrossRef]
39. F. Dell’Olio and V. M. N. Passaro, “Optical sensing by optimized silicon slot waveguides,” Opt. Express 15(8), 4977–4993 (2007). [CrossRef]
40. R. Wang, A. Vasiliev, M. Muneeb, A. Malik, S. Sprengel, G. Boehm, M.-C. Amann, I. Šimonytė, A. Vizbaras, K. Vizbaras, R. Baets, and G. Roelkens, “III–V-on-silicon photonic integrated circuits for spectroscopic sensing in the C2–C2 μm wavelength range,” Sensors 17(8), C2 (2017). [CrossRef]
41. M. A. Butt, S. N. Khonina, and N. L. Kazanskiy, “Ultrashort inverted tapered silicon ridge-to-slot waveguide coupler at 1.55 µm and 3.392 µm wavelength,” Appl. Opt. 59(26), 7821–7828 (2020). [CrossRef]
42. Y. Wu, Z. Qu, A. Osman, W. Cao, A. Z. Khokhar, J. S. Penades, O. L. Muskens, G. Z. Mashanovich, and M. Nedeljkovic, “Mid-infrared nanometallic antenna assisted silicon waveguide based bolometers,” ACS Photonics 6(12), 3253–3260 (2019). [CrossRef]
43. Y. Wu, Z. Qu, A. Osman, C. Wei, W. Cao, A. Tarazona, S. Z. Oo, H. M. H. Chong, O. L. Muskens, G. Z. Mashanovich, and M. Nedeljkovic, “Nanometallic antenna-assisted amorphous silicon waveguide integrated bolometer for mid-infrared,” Opt. Lett. 46(3), 677–680 (2021). [CrossRef]
44. M. Nedeljkovic, R. Soref, and G. Z. Mashanovich, “Predictions of free-carrier electroabsorption and electrorefraction in germanium,” IEEE Photonics J. 7(3), 1–14 (2015). [CrossRef]
45. C. J. Glassbrenner and G. A. Slack, “Thermal conductivity of silicon and germanium from 3 K to the melting point,” Phys. Rev. 134(4A), A1058–A1069 (1964). [CrossRef]
46. E. A. Scott, J. T. Gaskins, S. W. King, and P. E. Hopkins, “Thermal conductivity and thermal boundary resistance of atomic layer deposited high-k dielectric aluminum oxide, hafnium oxide, and titanium oxide thin films on silicon,” APL Mater. 6(5), 058302 (2018). [CrossRef]
47. Y.-Z. Deng, S.-F. Tang, H.-Y. Zeng, Z.-Y. Wu, and D.-K. Tung, “Experiments on temperature changes of microbolometer under blackbody radiation and predictions using thermal modeling by COMSOL multiphysics simulator,” Sensors 18(8), 2593 (2018). [CrossRef]
48. V. Linseis, F. Völklein, H. Reith, K. Nielsch, and P. Woias, “Thermoelectric properties of Au and Ti nanofilms, characterized with a novel measurement platform,” Mater. Today: Proc. 8, 517–522 (2019). [CrossRef]
49. D. Mardare, C. Baban, R. Gavrila, M. Modreanu, and G. I. Rusu, “On the structure, morphology and electrical conductivities of titanium oxide thin films,” Surf. Sci. 507-510, 468–472 (2002). [CrossRef]
50. W.-C. Peng, Y.-C. Chen, J.-L. He, S.-L. Ou, R.-H. Horng, and D.-S. Wuu, “Tunability of p- and n-channel TiOx thin film transistors,” Sci. Rep. 8(1), 9255 (2018). [CrossRef]
51. S. Wang, S. Yu, M. Lu, M. Liu, and L. Zuo, “Atomic layer-deposited titanium-doped vanadium oxide thin films and their thermistor applications,” J. Electron. Mater. 46(4), 2153–2157 (2017). [CrossRef]
52. R. T. Tung, “The physics and chemistry of the Schottky barrier height,” Appl. Phys. Rev. 1(1), 011304 (2014). [CrossRef]
53. M. Vollmer and K.-P. Möllmann, “Characterization of IR cameras in student labs,” Eur. J. Phys. 34(6), S73–S90 (2013). [CrossRef]
54. I.-K. Kang, Y. A. K. Reddy, Y. B. Shin, and H. C. Lee, “Systematic investigation on deposition temperature effect of Ni1–xO thin films for uncooled infrared image sensor applications,” IEEE Sens. J. 15(12), 7234–7241 (2015). [CrossRef]
55. D. B. Saint John, H.-B. Shin, M.-Y. Lee, S. K. Ajmera, A. J. Syllaios, E. C. Dickey, T. N. Jackson, and N. J. Podraza, “Influence of microstructure and composition on hydrogenated silicon thin film properties for uncooled microbolometer applications,” J. Appl. Phys. 110(3), 033714 (2011). [CrossRef]
56. D. S. Kim, S.-M. Park, and H. C. Lee, “Surface treatment method for 1/f noise suppression in reactively sputtered nickel oxide film,” J. Appl. Phys. 112(2), 024501 (2012). [CrossRef]
57. D. Hasan and C. Lee, “Hybrid metamaterial absorber platform for sensing of CO2 gas at mid-IR,” Adv. Sci. 5(5), 1700581 (2018). [CrossRef]
58. A. Gutierrez-Arroyo, E. Baudet, L. Bodiou, V. Nazabal, E. Rinnert, K. Michel, B. Bureau, F. Colas, and J. Charrier, “Theoretical study of an evanescent optical integrated sensor for multipurpose detection of gases and liquids in the mid-infrared,” Sens. Actuators, B 242, 842–848 (2017). [CrossRef]
59. L. Bodiou, Y. Dumeige, S. Normani, G. Louvet, P. Němec, V. Nazabal, and J. Charrier, “Design of a multimode interferometer-based mid-infrared multispecies gas sensor,” IEEE Sens. J. 20(22), 13426–13435 (2020). [CrossRef]
60. P. Su, Z. Han, D. Kita, P. Becla, H. Lin, S. Deckoff-Jones, K. Richardson, L. C. Kimerling, J. Hu, and A. Agarwal, “Monolithic on-chip mid-IR methane gas sensor with waveguide-integrated detector,” Appl. Phys. Lett. 114(5), 051103 (2019). [CrossRef]
61. E. Desimoni and B. Brunetti, “About estimating the limit of detection by the signal to noise approach,” Pharm. Anal. Acta 06(03), 1000355 (2015). [CrossRef]