1. Introduction

With increasing consumption of fossil fuels and the urgency of climate change and global warming, clean, renewable and sustainable energy source is greatly desired. Among others, hydrogen (H₂) is considered as a highly potential yet highly debated candidate for substitution of fossil fuels [1–3]. Aside from the potential as a fuel to power vehicles, hydrogen is already widely utilized in fertilizer, food and petrochemical industries, as well as in biomedical applications such as disease diagnosis [4]. However, the wide flammability limit (4-75 volume percent in air) and low ignition energy of hydrogen raises great safety concerns and puts hurdles for its large-scale applications. Therefore, detection of hydrogen leakage at an early stage is of critical importance at all stages of production, storage, transportation and utilization of hydrogen. To date, mainstream commercial hydrogen sensors are electrical ones that rely on detection of resistance change of the sensing element, which is predominately made from palladium (Pd). On the one hand, electrical sensors offer good sensitivity and very low detection limit (on the order of several parts per million, ppm) [5–9]. On the other hand, they pose potential danger of ignition of hydrogen/air mixture due to possible sparks generated by the electrical component, which could be devastating in a hydrogen rich environment. In contrast, optical hydrogen sensors that allow remote readout are generally considered safer and more favorable in a flammable atmosphere, as they are electricity-free at the sensing site and thus eliminate the danger of spark-generation [10–12].

Plasmonic hydrogen sensors [13–25], a subcategory of optical hydrogen sensors, have attracted much research interest in recent years due to their compactness and good tunability in operation wavelength, which allows them to be integrated with low-cost monochromatic illumination and detection devices [26]. Plasmonic nanostructures, made mainly from noble metals such as Au and Ag, support resonant collective free-electron oscillations on their surfaces when the wavelength of incident light matches their geometries, known as localized surface plasmon resonance (LSPR). However, most plasmonic metals are chemically inert and are thus not proper sensing element for hydrogen. Pd is a plasmonically “poor” but catalytically active transition metal widely used as the sensing element in hydrogen sensors, due to its good selectivity to hydrogen and thermodynamically suitable and fully reversible hydride formation properties [27], which also receives increasing research interests as a model system of first-order phase transition at the nanoscale [28–32]. Pd can absorb a substantial amount of hydrogen at the interstitial sites of its crystal lattice to form a diluted solid solution of hydrogen at low concentrations (α-phase) and a solidified hydride-phase at high concentrations (β-phase, PdHx, x≈0.6) [33]. Through combination of plasmonic metal (mainly Au and Ag) and Pd, researchers have previously demonstrated various plasmoinc hydrogen sensors, which can be basically divided into two categories, direct and indirect sensors. Direct plasmonic hydrogen sensors utilize Pd as the sensing element and optical signal transducer at the same time [13,17,34,35], while indirect sensors employ the intensity-borrowing scheme where plasmonic nanoantennas made from noble metals probe the property change of a Pd nanoparticle locating nearby [14,15,20–22], whose scattering spectrum in the visible range is usually very broad and featureless and does not allow for efficient hydrogen sensing [14]. Aside from these pioneering and successful demonstrations, challenges remain for the real-world application of plasmonic hydrogen sensors. One important aspect is that most plasmonic metal such as Au and Ag is not CMOS-compatible and thus not accepted by semiconductor foundries, which limits their feasibility in low-cost and large-scale applications. Moreover, present plasmonic/optical hydrogen sensors generally demonstrate a detection limit on the level of 0.5% volume concentration or even higher [11,12,15–18,20,22,23,26,34,36–38], with some exceptional cases of 0.1% [13,24,25,35,39,40], which is still too high compared with the lower flammability limit of 4% and is much higher than that of their electrical counterparts.

To tackle the above-mentioned challenges, we propose and present in this work a plasmonic hydrogen sensor based on high-density arrayed palladium-aluminum (Pd-Al) and aluminum- palladium (Al-Pd) hybrid nanorods. While Pd is the sensing element, Al here plays the role of plasmonic metal, i.e., to enhance the plasmon resonance of the hybrid nanorods in the visible range. Compared to Au and Ag, Al is cheap, abundant, easy to process and more importantly, together with Pd, CMOS-compatible [41,42], thus allowing mass-production of the proposed sensor in modern semiconductor foundries with low cast. Moreover, Al is able to support LSPR from the visible to ultraviolet spectral range and thus receives increasing attention in recent years as an alternative plasmonic metal. Researchers have demonstrated many nanophotonic applications based on Al, such as ultraviolet plasmonics [43–45], nonlinear plasmonics [46], photocatalysis [47] as well as plasmonic structural colors [48–50]. In this work, adapting the ideas of subwavelength plasmonic structural color pixels [48–50], we intentionally arrange the hybrid nanorods into high-density regular arrays with elaborately designed periodicities and dimensions. By doing so, the interplay between the periodicities and dimensions of the nanorods allows us to engineer their optical properties, ensuring high optical contrast for spectroscopic measurement. Moreover, compared to the case of random arrays typically fabricated by colloidal-hole lithography method with larger and random inter-particle spacing [13,20,34,36] and to the case of single-particle antenna-probe geometry [14,15], the high-density regular arrays offer much more sensing sites and open surfaces of palladium to hydrogen per unit area, a key factor to push down the detection limit and to speed-up the response process [9]. In addition, the hybrid nanorods proposed here have a stacked geometry, requiring only one lithography step to fabricate [15,36,51], which releases the requirement of complex multi-step lithography process used to define the lateral antenna-probe geometries [14]. Experimentally, the fabricated nanorod arrays have unambiguously detected hydrogen at a volume concentration of 40 ppm (0.004%) in nitrogen (N₂) carrier gas, i.e., one thousandth of the lower flammability limit, which is comparable to and even better than some state-of-the-art electrical hydrogen sensors [5–8].

2. Experimental section

2.1. Sample fabrication

The nanorod arrays were fabricated through a standard electron-beam lithography (EBL, JEOL JBX6300FS) procedure. First, cleaned silicon substrate was spin-coated with photoresist (Poly-methyl methacrylate (PMMA), Allresist) and then exposed to electron-beam to define the rods. After development, the substrate was covered sequentially by a 20 nm-thick Al (plasmonic metal) and a 20 nm-thick Pd (hydrogen sensitive material) film through thermal evaporation in the same coater chamber (SKY, BOX- RH400). After a subsequent lift-off process, the Al-Pd nanorod arrays were obtained. Keeping all other procedures the same, only reversing the evaporation order of Al and Pd materials results in the other Pd-Al hybrid nanorod configuration.

2.2. Optical reflection measurements under relatively high hydrogen concentration

White light reflection spectra of fabricated nanorod arrays were measured via a micro-spectroscopic system (Nikon ECLIPSE LV150N) equipped with a home-built sample stage, which functioned also as the gas cell for latter gas sensing experiments. White light from a halogen lamp was introduced by a microscope objective (Nikon, 20 × , NA = 0.4) to reach the nanorod arrays through a glass window of the sample stage, with polarization controlled parallel to the long-axis of the nanorods. Reflected light was collected by the same objective operated in bright-field mode and sent into a spectrometer (Andor Kymera 193i, Newton 920). Inlet of the sample stage was connected to a gas-mixing and mass flow control system (Sensirion, SFC5300), which allowed hydrogen and nitrogen to be mixed at precise concentrations with a steady flow rate. Outlet of the sample stage was connected to the exhaust system of the building. After purging the gas cell with pure N₂ with a flow rate of 1 L/min for sufficient time (> 3 min), reflection spectra under pure N₂ environment were taken. Similarly, reflection spectra under 3% H₂ in N₂ were taken after purging the gas cell with 3% H₂ in N₂ at a flow rate of 1 L/min for more than 3 minutes. Under pure N₂ environment, moving away from the nanorod arrays, a reference spectrum was recorded from bare areas on the same substrate, which was utilized throughout the work to obtain the normalized reflectance spectra shown below. Due to normalization, a reflectance value greater than one means that the reflectance from the nanorod array is higher than that from the flat SiO₂/Si surface, at corresponding wavelengths.

2.3. Hydrogen sensing experiments

To perform hydrogen sensing experiments, H₂ with various concentrations in N₂ carrier gas was flowed over the sample substrate while continuously monitoring reflection spectra of the nanorod arrays. Concentration of hydrogen was periodically varied from 40 to 5000 ppm (0.004%-0.5%) in a preprogrammed sequence with details shown in Table 1 of appendices. A relatively high flow rate of 4000 sccm was chosen to reduce the time delay for the gas mixture to flow from the valve of the mass flow controller along a ca. 2 m-long pipe to reach the sample, so that the change in hydrogen concentration and the response of the nanorod arrays is better synchronized. Actually, hydrogen sensing experiments under smaller flow rates such as 2000 sccm were also performed, which demonstrated nearly identical results except that the sensor response was a little bit delayed with respect to the change of hydrogen concentration. White light reflection spectra were recorded continuously every 10 seconds (8 s exposure + 2 s interval) while the nanorod arrays were exposed to the above-mentioned gas-flow sequences. Peak positions of the reflection spectra, obtained by employing a centroid method that calculates the center of mass of the reflection peak [21,52], were used as the readout. Each round of hydrogen sensing experiments took roughly three hours which contains more than three complete gas-flow cycles.

Table 1. Gas-flow sequence with various hydrogen concentrations.

View Table

3. Results and discussion

The proposed high-density regular array of Al-Pd hybrid nanorods is schematically shown in Fig. 1(a), with Al rods at the bottom and Pd rods on the top. Pd-Al hybrid nanorods have a reversed stacking order. We keep the widths of all the nanorods to be 40 nm and the thicknesses of the Al and Pd layer to be both 20 nm while varying the periodicities in x- and y-directions (Px, Py) as well as the lengths (L) of the nanorods. Periodicities (Px, Py) are constrained to be no larger than 400 nm to ensure subwavelength spacing among the nanorods in the visible range, and the length (L) of the nanorods varies from 75 to 375 nm. The nanorod arrays lie on top of a 500 μm-thick silicon substrate, with a 100 nm-thick natively oxidized SiO₂ spacer layer. Such a silicon substrate with oxidization layer is purposely chosen for two reasons: 1. CMOS-compatibility, so that all the materials involved (Pd, Al, SiO₂ and Si) are CMOS-compatible; 2. to avoid high damping caused by direct contact of plasmonic nanorods and silicon [53]. Figure 1(b) depicts an exemplary SEM image of one fabricated Al-Pd nanorod array with Px = Py = 250 nm and L = 200 nm, demonstrating good homogeneity of fabricated nanorods. Each array has a 12 × 12 μm² footprint and forms a square-shaped color-patch. As can be seen from a typical bright-field reflective optical microscope image in Fig. 1(c), different colors appear in accordance with the different geometrical parameters (Px, Py, L) of the nanorod arrays [48–50]. The unpatterned areas in between the color-patches show featureless gray color, corresponding to the reflection from a flat SiO₂/Si surface under white light illumination. Except for some nanorod arrays with shorter length and larger spacing (e.g., the ones at the lower-left corner of Fig. 1(c)), the contrast between the color patch and the unpatterned area is in general rather high, which benefits their superior performance as optical hydrogen sensors, as is evident in following characterizations. Then, fabricated sample substrate was put into a home-built microscope sample stage (Fig. 1(d)). Such a home-built micro-spectroscopic instrument equipped with mass flow control system allows us to continuously perform optical measurements with precisely controlled concentration and flow rate of gas-mixture to characterize the optical response and sensing capabilities of the nanorod arrays.

 

Fig. 1 Proposed high-density hybrid nanorod array plasmonic hydrogen sensor and the platform for hydrogen sensing experiments. (a) Schematic illustration of the proposed Al-Pd (bottom layer: Al) hybrid nanorod array on top of a 500 μm-thick Si substrate, separated by a 100 nm-thick SiO₂ oxidization layer. (b) SEM image of a typical fabricated Al-Pd nanorod array (Px = Py = 250 nm, L = 200 nm), demonstrating good homogeneity. Scale bar: 200 nm. (c) Bright-field reflective optical microscope image of some fabricated Al-Pd nanorod arrays, which appear as various color patches under white-light illumination. All the color patches have the same periodicity in x-direction, i.e., Px = 250 nm. From left to right column, the periodicity in y-direction decreases from 300 nm to 100 nm in steps of 50 nm, and from top to bottom row, the length of nanorods decreases from 225 nm to 125 nm in steps of 25 nm. Scale bar: 10 µm. (d) Artistic illustration of the optical characterization and gas sensing setup. A home-built sample stage allows gas-mixture with precise concentration to flow across the sample while continuously monitoring the white light reflection spectra.

Download Full Size | PDF

First, reflection spectra of Pd-Al and Al-Pd nanorod arrays were examined under the condition of pure N₂ environment and N₂ with relatively high H₂ concentration (3% H₂ in N₂). As an example, normalized reflectance of a Pd-Al nanorod array (Pd at bottom, Px = 250 nm, Py = 150 nm, L = 225 nm) is displayed in Fig. 2(a). With pure N₂ as the environmental gas, Pd is in its original metallic state. Due to the specific periods in both x- and y-directions as well as the length of the nanorods, the reflection spectrum shows a low reflectance dip around 480 nm and a reflection peak roughly centered at 710 nm (solid blue curve). Under ambient condition (1 atm pressure and 300 K), when replacing the gas environment of pure N₂ with 3% H₂ in N₂, which is higher than the critical concentration for the phase transition of Pd to take place, Pd is fully hydrogenated into PdHx [15,34,36]. Upon the phase transition, volume expansion due to lattice strain and change of dielectric property of Pd occurs accordingly, which in turn results in a spectral shift. As can be seen from the dashed blue curve in Fig. 2(a), reflection spectrum of PdH-Al nanorods shows an obvious red shift and a decrease in reflectance. Finite element simulation results (FEM, COMSOL Multiphysics) in Fig. 2(b) of the same nanorod array demonstrate a good agreement with the experimental results, when artificially changing the material from Pd to PdH with both volume expansion and change in dielectric property of Pd considered. Similar behavior is also observed for the case of Al-Pd nanorod arrays (Al at the bottom), and Fig. 2(c) depicts a typical result for an Al-Pd nanorod array with the same geometrical parameters as the Pd-Al nanorods in Fig. 2(a). Under pure N₂ environment, the nanorod array shows relatively high reflectance in the blue wavelength range (< 450 nm) with a reflection dip around 550 nm and a broad reflection plateau for wavelength greater than 750 nm (solid-green curve). When purging the gas cell with a gas-mixture of 3% H₂ in N₂, Pd undergoes the same phase transition, and the reflection spectrum of the Al-PdH nanorod array changes accordingly. Aside from an overall red shift, a pronounced decrease in reflectance above the wavelength of 570 nm is observed (dashed-green curve). Simulated spectra in Fig. 2(d) agree well with experimental results as well. However, there are still some observable distinctions between experiments and simulations for both cases, appearing mainly as the detailed line-shapes of the simulated spectra. Such discrepancies could arise from dimension deviations of fabricated nanorods from the design (periodicities in both directions, length and width of nanorods, see Fig. 6 in appendices), surface roughness of fabricated nanorods (especially when Al is on the top) compared to the perfect shapes in simulations, as well as the difference of dielectric constants of real materials (Si, SiO₂, Al, Pd and PdH) to the tabulated values used in simulations [33,54]. Taking all the above-mentioned factors into account, one should in principle be able to obtain simulated spectra that match better with the measured ones. Nevertheless, our simulations have caught the main characteristics of the optical behavior of the hybrid nanorod arrays with hydrogenation of Pd, i.e., a spectral red shift plus a decrease in reflectivity.

 

Fig. 2 Optical response of hybrid nanorod arrays under pure N₂ and 3% H₂ in N₂ gas environment. Experimentally measured (a) and numerically simulated (b) normalized reflectance spectra of a Pd-Al nanorod array (Px = 250 nm, Py = 150 nm, L = 225 nm) under pure N₂ (corresponding to Pd, solid blue curves) and 3% H₂ in N₂ (corresponding to PdH, dashed blue curves) gas environment. (c) and (d) The same as (a) and (b), but for an Al-Pd nanorod array with the same dimension.

Download Full Size | PDF

Below the critical hydrogen concentration where Pd is always in the α-phase, Pd has no hysteresis effect on the absorption and release of hydrogen [15,19,23], rendering this concentration range more suitable for a sensor. Therefore, it is more interesting to examine the performance of the nanorod arrays at very low hydrogen concentrations, and extensive hydrogen sensing experiments were carried out. During each hydrogen sensing measurement, we observed some long-term drifts of measured signal, which could originate from instabilities of the light source and the optical setup as well as variations in ambient temperature. To exclude the influence of long-term drift, one can employ the cross-polarization self-referencing drift-correction scheme introduced by Langhammer et al [55]. Since in our case the long-term drifts did not overwhelm the sensing signals, we fitted the baselines of measured time traces of spectral shifts and subtracted them from the raw data to highlight the sensing response, with typical results presented in Fig. 3 (see Figs. 7 and 8 in appendices for original time trace of centroid wavelength and the procedure of baseline subtraction, as well as additional hydrogen sensing results).

 

Fig. 3 Typical hydrogen sensing performance of a Pd-Al (a) and an Al-Pd (b) hybrid nanorod array. Blue and green circles correspond to measured data points, and the blue and green curves are just lines connecting the data points. Red bars correspond to the hydrogen concentration. Baselines are subtracted from original time traces of centroid spectral shifts to highlight sensing responses. Step-wise hydrogen sensing signals as well as the responses at 40 ppm, indicated by the black arrows, are clearly visible.

Download Full Size | PDF

Figure 3(a) plots the time trace of centroid wavelength shifts of a Pd-Al nanorod array under the above-mentioned gas-flow sequence. The Pd-Al nanorod array examined here has a geometrical parameter set of Px = 350 nm, Py = 150 nm, L = 250 nm. At a first glance of Fig. 3(a), it is noticed that the nanorod array demonstrates apparent step-wise sensing signals (spectral shifts) in accordance with the increase of hydrogen concentration. During a roughly three hours measurement, three complete gas-flow cycles are accomplished, and the response of the nanorod array shows three clear periods as well. More importantly, at the very lowest hydrogen concentration of 40 ppm at the beginning of each gas-flow cycle, we repeatedly see a clear and unambiguous shift of more than 0.5 nm, as indicated by the black arrows in Fig. 3(a). This is a proof of detection of hydrogen at a concentration (0.004%) three orders of magnitude smaller than its lower flammability limit (4%). Moreover, from the distinct and reliable responses of ≳0.5 nm at 40 ppm and the noise level of ≲ ± 0.15 nm, it is inferred that an even lower detection limit is achievable. As a matter of fact, the lowest concentration of hydrogen that we can achieve (40 ppm) during our experiment is limited by our present mass flow control system and could be in principle lower, if a better mass flow controller were utilized. Thus, 40 ppm is in principle not the physical limit of our nanorod array sensors. Similarly, in Fig. 3(b), sensing response of an Al-Pd nanorod array is presented, whose geometrical parameters are Px = 300 nm, Py = 100 nm, L = 175 nm. Like the case of Pd-Al nanorods, the Al-Pd nanorod array also demonstrates obvious and reliable sensing responses to various H₂ concentrations. Moreover, at the lowest hydrogen concentration of 40 ppm that we can achieve, Al-Pd nanorod array exhibits clear sensing signals on the order of 0.5 nm as well, indicated by the black arrows in Fig. 3(b). However, with a closer look at the two time traces of sensing response to the same gas-flow sequence, one can notice a major difference between the sensing behaviors of Pd-Al and Al-Pd nanorods: the amount of spectral shifts under the same hydrogen concentrations are different for the two configurations. Despite the different sensing behaviors of the two nanorod configurations, which will be analyzed in detail in following paragraphs, both Pd-Al and Al-Pd nanorod arrays can respond to hydrogen at a concentration as low as 40 ppm, which is much lower than that of most plasmonic hydrogen sensors and is among the lowest reported so far [13–26,34–40,51,55,56]. It has to be pointed out that not all the fabricated nanorod arrays demonstrate such superior hydrogen sensing capability, i.e., while some arrays respond to very low hydrogen concentration (40 ppm), some respond only at higher concentrations (100 or 200 ppm). The reason for such a behavior is still under investigation and could be explained by a correlation study between the hydrogen sensing performance and the details of each corresponding nanorod array, which is currently underway and beyond the scope of this work at the moment.

To further scrutinize the difference of sensing behaviors between the two configurations of hybrid nanorods, two important parameters from their sensing signals, response time (τ_Rise) and maximal shift (Δλ_m), are extracted for a quantitative comparison. Maximal shift Δλ_m is defined as the difference between the maximum and the minimum of the centroid spectral shift under a certain hydrogen concentration, and the response time τ_Rise is defined as the time needed for the sensor response to reach 90% of the maximal shift. According to such definition and extracting these parameters from each peak of the signal trace for the two sensor configurations, we obtained and plotted τ_Rise and Δλ_m versus hydrogen concentration (in logarithmic scale) in Fig. 4. As can be seen from Fig. 4(a), for the concentration range involved (40-5000 ppm), sensors with Pd-Al configuration demonstrate overall a longer response time than Al-Pd sensors. Concretely, for the case of Pd-Al sensor, its response time decreases monotonically from ≈108 s to ≈52 s when hydrogen concentration increases from 40 to 5000 ppm. Similarly, the response time of Al-Pd sensor decreases from ≈35 s to ≈18 s when hydrogen concentration increases from 40 to 200 ppm. However, beyond 200 ppm, response time of Al-Pd sensor decreases no more and stabilizes around 18 s with further increase of hydrogen concentrations, which indicates to some extent the physical limit of the present sensor. Although a response time of 50-110 s is relatively too long, a response time around 20 s is reasonable for a practical sensor, especially at very low hydrogen concentrations. The error bars come from statistics of 9 (or 12) data points (3 data points/time trace × 3 times of measurements; 12 data points for 40 and 100 ppm as there are 4 data points available from each measurement). Looking at the maximal spectral shifts in Fig. 4(b), two distinct behaviors for the two sensor configurations are clearly observed. Except for the lowest hydrogen concentration of 40 ppm, where the spectral shift is nearly identical for both cases (0.5 nm), the overall response of Pd-Al sensors is greater than that of Al-Pd ones. With the increase of hydrogen concentration, both Pd-Al and Al-Pd nanorod sensors show an increase in maximal shift, but with different amplitude. Obviously, maximal shift of Pd-Al sensors increases faster than Al-Pd ones. In the end, the Pd-Al sensor reaches a final shift close to 7 nm while the Al-Pd sensor reaches only 2 nm at 5000 ppm. From the comparisons above, it is clear that the Pd-Al nanorod array exhibits, under the same hydrogen concentration, larger spectral shift, while the Al-Pd configuration possesses a shorter response time in the concentration range we have examined here. These results suggest that in applications where a more pronounced spectral shift is pursued, the Pd-Al configuration should be utilized, while the Al-Pd configuration is a better choice for cases where a shorter response time is required.

 

Fig. 4 Comparison of the sensing behavior of the two sensor configurations. Response time τ_Rise (a) and maximal shift Δλ_m (b) of the nanorod array sensors for the Pd-Al (blue circle) and Al-Pd (green diamond) configuration under different hydrogen concentrations. Error bars come from statistics of 9 (12) data points at each concentration. Blue and green curves connect the data points and serve as a guidance to the eye.

Download Full Size | PDF

Following the same reasoning in the work by Strohfeldt et al, physical insights to understand the observed different sensing behaviors of the two configurations can be obtained. As previously pointed out by Strohfeldt et al in a similar stacked system [36], the stacking order and the resulting material asymmetry matters, and the material in direct contact with the substrate is found to play a dominant role in the hybrid system in the sense of sensitivity. Thus, in our case, for Pd-Al nanorods, since Pd layer is at the bottom, it renders the whole system more Pd-like, and due to the electric field concentration around the corners of Pd, the sensitivity of Pd-Al nanorods to hydrogen (slope of the curve in Fig. 4(b)) and the absolute values of spectral shift is higher than that of Al-Pd nanorods. Regarding the response time, it is mainly determined by the overall splitting and diffusion rate of hydrogen molecules on exposed Pd surface. Since the Al-Pd configuration has more Pd surfaces exposed to the gas environment compared to the Pd-Al configuration, providing more sorption sites of hydrogen molecules, it is comprehensible that the Al-Pd nanorods demonstrate shorter response times than Pd-Al nanorods. The response time may be further shortened through elaborate design of the sensor geometries to expose more Pd surfaces to the gas environment.

So far, we have experimentally demonstrated sensing of hydrogen at a concentration down to 40 ppm. However, it is quite counter-intuitive that such simple Pd-Al and Al-Pd hybrid nanorod arrays could achieve such a low detection limit. After all, the materials and the geometry used are often seen in this area. Yet, several unique characteristics of the hybrid nanorod array sensors presented here have to be pointed out, which account for the unprecedentedly low detection limit. First, on the material perspective, aluminum is not only a rising alternative plasmonic metal which attracts increasing research attention due to its abundance, low-cost, CMOS-compatibility and its widely tunable LSPR property from UV to NIR, but also has been demonstrated as a photocatalyst which is able to dissociate hydrogen under light illumination [47]. Moreover, Pd, as a canonical transition metal catalyst, has proven to be able to add photocatalysis property, when driven by the “forced plasmons” induced by a plasmonic antenna near-by in an “antenna-reactor” configuration [57,58]. In both cases, hot carriers generated by plasmon decay play an important role, which are under intensive investigations recently [59–61]. Especially in the “antenna-reactor” geometry, the Al nanoantenna focuses far-field light onto the surface of catalytic Pd and induces an optical polarization inside the Pd nanoparticle, the so-called “forced plasmon”, which generates hot carriers and promote hydrogen dissociation when it decays. Similarly, in our case, since the Al rod and Pd rod are in direct contact, it is even easier for the electrons inside the Pd layer to oscillate at optical frequencies under the drive of incident light and the drive of Al plasmons, compared to the case of near-field coupling in the “antenna-reactor” geometry. This is proven by numerical simulation results of the charge distributions shown in Fig. 5, as well as by the results in the work of Strohfeldt et al [36]. Figures 5(a) and (b) show the surface charge distribution of the typical Pd-Al and Al-Pd nanorods under light illumination at wavelength of 700 and 750 nm, respectively, whose geometric dimensions are the same as that in Fig. 2. It is clear that the layered hybrid nanorod behaves like a single rod, i.e., charges in the two layers oscillate in phase for both cases of Pd-Al and Al-Pd nanorods. This conclusion is further confirmed when examining the electric field distribution in a cross-sectional view as shown in Figs. 5 (c) and (d). It is also worthy to point out that charges in the two layers of both Pd-Al and Al-Pd nanorods always oscillate in-phase during one complete optical cycle, not only at the moment depicted in Fig. 5, as confirmed by numerical simulations. Thus, from the material perspective, it is reasonable to argue that hot carrier mediated photocatalytic splitting of hydrogen molecule takes place on our sensor surface, which dramatically facilitates the detection of hydrogen at extremely low concentrations and accelerates the response time. A more detailed and thorough investigation of the role of photocatalytic effects of Al and Pd would aid to understand the mechanism and dynamics of the hybrid nanorod sensors, which is however beyond the scope of the present work.

 

Fig. 5 Simulation results to show that the Pd-Al and Al-Pd layered hybrid nanorods behave like a single nanorod. Surface charge distributions of (a) Pd-Al and (b) Al-Pd hybrid nanorod at wavelengths of 700 and 750 nm, respectively, whose dimensions are the same as that in Fig. 2. Electric field distributions of (c) Pd-Al and (d) Al-Pd hybrid nanorod under the same condition as (a) and (b). Color plots represent the norm of electric field, while arrows indicate local electric field at each position. It is clear that charges in the two layers oscillate in-phase in both cases (a.u. = arbitrary unit).

Download Full Size | PDF

Second, on the structure and arrangement perspective, these hybrid nanorods were fabricated by high precision electron beam lithography method, which is highly reproducible and results in smaller particle-to-particle variation (see the SEM image in Fig. 1(b)), compared to nanoparticles synthesized wet-chemically or fabricated through colloidal-hole lithography method. The latter two cases result in a relatively broad distribution of particle size and shape, which broadens the characteristic spectral feature of plasmon resonance, and this inhomogeneous broadening will hamper the detection of very small spectral shift at extremely low hydrogen concentration, e.g., 40 ppm. In our case, since the nanorod arrays demonstrate good homogeneity which form patches with quite uniform color (see Fig. 1(c)), the imhomogeneous spectral broadening is not so severe that it in principle provides a better sensitivity from a perspective of ensemble measurement.

The last but not least, from an instrumental and technical perspective, the precise temperature control of the substrate during measurement, the effort to keep illumination and detection optics as stable as possible, as well as the centroid spectra analysis method [52], which is quite sensitive to small shifts of the barycenter of a spectrum, all contribute to the achievement of the extremely low detection limit of hydrogen in this work.

4. Conclusions

We have fabricated CMOS-compatible high-density Pd-Al and Al-Pd hybrid nanorod arrays and demonstrated their superior competency as a hydrogen sensor. The nanorods were patterned by EBL into regular arrays with subwavelength spacing, in a way similar to produce plasmonic structural color pixels. Actually, such a patterning procedure could be replaced by cost-effective and up-scalable methods such as interference lithography [62] or nano-imprint lithography. Reflection spectrum of both Pd-Al and Al-Pd nanorod array demonstrates an obvious red shift upon phase transition of Pd to PdH, under relatively high hydrogen concentration (3%). More importantly, at a concentration of 40 ppm, both nanorod array configurations demonstrate unambiguous sensing signals (≳0.5 nm) with reasonable response times (tens of seconds). The experimentally demonstrated detection limit here is much lower than that of most plasmonic hydrogen sensors so far and is comparable to their state-of-the-art electrical counterparts. The detection limit is believed to be even lower, estimated from the measured signal-to-noise ratio. Different sensing behaviors are found for the two geometrical configurations, where Pd-Al nanorods provide larger spectral shift while Al-Pd ones exhibit shorter response time, and the different sensing behavior can be understood by considering the material asymmetry of the hybrid nanorods. Several possible factors are discussed that contribute to the experimentally observed low detection limit. For a practical sensor, it is meaningful to examine the sensor performance under realistic conditions, with the presence of humidity, O₂ and/or CO. Towards real-world application of the plasmonic hydrogen sensor presented here, one can further optimize the component proportion ratio between Al and Pd to strike a better balance between spectral shift and response time, utilize low-cost lithography method other than EBL while keeping the nanorods in high-density regular arrays, as well as further engineer the geometries to shorten the response time to several seconds.

Appendices

A. Details of gas-flow sequence

Each gas-flow cycle consists of alternations of 200 s pure N₂ purge and 200 s flow of H₂ + N₂ mixture, expect for the last pure N₂ purge, which is 400 s long. Total time of a gas-flow cycle is 3200 s, and the gas-flow cycle was repeated during measurements. Concentrations of H₂ are 40, 100, 200, 500, 1000, 2000 and 5000 ppm, respectively, as shown in the following table.

B. Geometrical deviations of fabricated nanorods from the design

 

Fig. 6 SEM characterization of a typical Pd-Al nanorod array (design: Px = 400 nm, Py = 300 nm, L = 350 nm, W = 40 nm) demonstrates minor but observable deviations of periodicities and dimensions of fabricated nanorods from the design as well as the surface roughness of the nanorods, especially when Al layer is on the top.

Download Full Size | PDF

C. Long-term drift and the procedure of baseline subtraction

White light reflection spectrum was recorded every 10 seconds (8 s exposure + 2 s interval) during the ca. three hours hydrogen sensing measurement. Centroid method was applied to each measured reflection spectrum to get the centroid wavelength of the reflection peak, represented by the individual data points in Fig. 7(a) (open blue circles). All the data points constitute the original time trace of centroid wavelengths under various hydrogen concentrations. Subsequently, through finding the local minima of the original time trace, a baseline (thick blue curve) is obtained, which represents the long-term drift during the three hours measurement. As indicated by the black arrows in Fig. 7(a), it is clear that the long-term drift during the measurements do not overwhelm even the smallest original sensing signals under the hydrogen concentration of 40 ppm. The baseline is then subtracted from the original time trace to highlight the sensing response (i.e., the centroid shift), as shown in Fig. 7(b) (the same one as Fig. 3(a) in the main text).

 

Fig. 7 Original sensor response and the procedure of baseline subtraction. (a) Measured original time trace of the centroid peak wavelengths of the reflection spectrum under various hydrogen concentrations (open blue circles). Through finding the local minima, a baseline (thick blue curve) is created and then subtracted from the original time trace. After subtraction, sensing signal (time trace of centroid shift) is obtained (b).

Download Full Size | PDF

D. More hydrogen sensing results

More time traces of original centroid wavelengths from hydrogen sensing experiments of hybrid nanorod arrays with different geometrical parameters. The Pd-Al and Al-Pd nanorod arrays have the same periodicities (Px, Py) as the one shown in Fig. 3 of the main text, respectively, but with a different rod length L. All sensing experiments suffer from some long-term drifts, which, however, do not overwhelm the sensing signals. As indicated by the black arrows in Fig. 8, sensing of hydrogen at a concentration of 40 ppm is clearly visible, illustrating good reproducibility of the sensing response.

 

Fig. 8 Time traces of original centroid wavelengths for hybrid nanorod arrays with other dimensions. (a) Pd-Al nanorod array (Px = 350 nm, Py = 150 nm, L = 300 nm), (b) Pd-Al nanorod array (Px = 350 nm, Py = 150 nm, L = 275 nm), (c) Al-Pd nanorod array (Px = 300 nm, Py = 100 nm, L = 200 nm), and (d) Al-Pd nanorod array (Px = 300 nm, Py = 100 nm, L = 150 nm).

Download Full Size | PDF

Funding

Chinese Academy of Sciences Pioneer Hundred Talents Program; National Natural Science Foundation of China (51505456); Jilin Province Development and Reform Commission (2015Y028); Open Foundation of State Key Laboratory of Electronic Thin Film and Integrated Devices, University of Electronic Science and Technology of China (KFJJ201604); State Key Laboratory of Applied Optics, Changchun Institute of Optics, Fine Mechanics and Physics (CIOMP), Chinese Academy of Sciences.

Disclosures

The authors declare that there are no conflicts of interest related to this article.

References

1. J. A. Turner, C. Linkous, and N. Muradov, “Sustainable hydrogen production,” Science 305(5686), 972–974 (2004). [CrossRef]   [PubMed]  

2. L. Barreto, A. Makihira, and K. Riahi, “The hydrogen economy in the 21st century: a sustainable development scenario,” Int. J. Hydrogen Energy 28(3), 267–284 (2003). [CrossRef]  

3. A. Züttel, A. Borgschulte, and L. Schlapbach, Hydrogen as a Future Energy Carrier (Wiley‐VCH Verlag GmbH & Co. KGaA, 2008).

4. W. Shin, “Medical applications of breath hydrogen measurements,” Anal. Bioanal. Chem. 406(16), 3931–3939 (2014). [CrossRef]   [PubMed]  

5. K. W. Lin, H. I. Chen, C. T. Lu, Y. Y. Tsai, H. M. Chuang, C. Y. Chen, and W. C. Liu, “A hydrogen sensing Pd/InGaP metal-semiconductor (MS) Schottky diode hydrogen sensor,” Semicond. Sci. Technol. 18(7), 615–619 (2003). [CrossRef]  

6. K. Skucha, Z. Fan, K. Jeon, A. Javey, and B. Boser, “Palladium/Silicon Nanowire Schottky Barrier-Based Hydrogen Sensors,” Sens. Actuators B Chem. 145(1), 232–238 (2010). [CrossRef]  

7. E. Menumerov, B. A. Marks, D. A. Dikin, F. X. Lee, R. D. Winslow, S. Guru, D. Sil, E. Borguet, P. Hutapea, R. A. Hughes, and S. Neretina, “Sensing Hydrogen Gas from Atmospheric Pressure to a Hundred Parts per Million with Nanogaps Fabricated Using a Single-Step Bending Deformation,” ACS Sens. 1(1), 73–80 (2016). [CrossRef]  

8. M. Zhao, H. W. Man, C. W. Ong, N. H. Ng, and H. C. Man, “Tunability of Pd-nanogapped H2 sensors made on SiO2-coated Si micropillar arrays,” Sens. Actuators B Chem. 255, 944–951 (2018). [CrossRef]  

9. X. Li, M. Le Thai, R. K. Dutta, S. Qiao, G. T. Chandran, and R. M. Penner, “Sub-6 nm Palladium Nanoparticles for Faster, More Sensitive H2 Detection Using Carbon Nanotube Ropes,” ACS Sens. 2(2), 282–289 (2017). [CrossRef]   [PubMed]  

10. J. Villatoro, D. Luna-Moreno, and D. Monzón-Hernández, “Optical fiber hydrogen sensor for concentrations below the lower explosive limit,” Sens. Actuators B Chem. 110(1), 23–27 (2005). [CrossRef]  

11. C. Perrotton, R. J. Westerwaal, N. Javahiraly, M. Slaman, H. Schreuders, B. Dam, and P. Meyrueis, “A reliable, sensitive and fast optical fiber hydrogen sensor based on surface plasmon resonance,” Opt. Express 21(1), 382–390 (2013). [CrossRef]   [PubMed]  

12. R. J. Westerwaal, J. S. A. Rooijmans, L. Leclercq, D. G. Gheorghe, T. Radeva, L. Mooij, T. Mak, L. Polak, M. Slaman, B. Dam, and T. Rasing, “Nanostructured Pd–Au based fiber optic sensors for probing hydrogen concentrations in gas mixtures,” Int. J. Hydrogen Energy 38(10), 4201–4212 (2013). [CrossRef]  

13. C. Langhammer, I. Zorić, B. Kasemo, and B. M. Clemens, “Hydrogen storage in Pd nanodisks characterized with a novel nanoplasmonic sensing scheme,” Nano Lett. 7(10), 3122–3127 (2007). [CrossRef]   [PubMed]  

14. N. Liu, M. L. Tang, M. Hentschel, H. Giessen, and A. P. Alivisatos, “Nanoantenna-enhanced gas sensing in a single tailored nanofocus,” Nat. Mater. 10(8), 631–636 (2011). [CrossRef]   [PubMed]  

15. T. Shegai and C. Langhammer, “Hydride Formation in Single Palladium and Magnesium Nanoparticles Studied by Nanoplasmonic Dark-Field Scattering Spectroscopy,” Adv. Mater. 23(38), 4409–4414 (2011). [CrossRef]   [PubMed]  

16. A. Tittl, P. Mai, R. Taubert, D. Dregely, N. Liu, and H. Giessen, “Palladium-Based Plasmonic Perfect Absorber in the Visible Wavelength Range and Its Application to Hydrogen Sensing,” Nano Lett. 11(10), 4366–4369 (2011). [CrossRef]   [PubMed]  

17. T. Shegai, P. Johansson, C. Langhammer, and M. Käll, “Directional scattering and hydrogen sensing by bimetallic Pd-Au nanoantennas,” Nano Lett. 12(5), 2464–2469 (2012). [CrossRef]   [PubMed]  

18. A. Tittl, X. Yin, H. Giessen, X. D. Tian, Z. Q. Tian, C. Kremers, D. N. Chigrin, and N. Liu, “Plasmonic smart dust for probing local chemical reactions,” Nano Lett. 13(4), 1816–1821 (2013). [CrossRef]   [PubMed]  

19. C. Wadell, S. Syrenova, and C. Langhammer, “Plasmonic hydrogen sensing with nanostructured metal hydrides,” ACS Nano 8(12), 11925–11940 (2014). [CrossRef]   [PubMed]  

20. S. Syrenova, C. Wadell, and C. Langhammer, “Shrinking-hole colloidal lithography: self-aligned nanofabrication of complex plasmonic nanoantennas,” Nano Lett. 14(5), 2655–2663 (2014). [CrossRef]   [PubMed]  

21. N. Li, A. Tittl, S. Yue, H. Giessen, C. Song, B. Ding, and N. Liu, “DNA-assembled bimetallic plasmonic nanosensors,” Light Sci. Appl. 3(12), e226 (2014). [CrossRef]  

22. A. Yang, M. D. Huntington, M. F. Cardinal, S. S. Masango, R. P. Van Duyne, and T. W. Odom, “Hetero-oligomer Nanoparticle Arrays for Plasmon-Enhanced Hydrogen Sensing,” ACS Nano 8(8), 7639–7647 (2014). [CrossRef]   [PubMed]  

23. C. Wadell, F. A. A. Nugroho, E. Lidström, B. Iandolo, J. B. Wagner, and C. Langhammer, “Hysteresis-Free Nanoplasmonic Pd-Au Alloy Hydrogen Sensors,” Nano Lett. 15(5), 3563–3570 (2015). [CrossRef]   [PubMed]  

24. R. Jiang, F. Qin, Q. Ruan, J. Wang, and C. Jin, “Ultrasensitive Plasmonic Response of Bimetallic Au/Pd Nanostructures to Hydrogen,” Adv. Funct. Mater. 24(46), 7328–7337 (2014). [CrossRef]  

25. K. Y. Hang, X. Zhu, X. Zhuo, R. Jiang, Z. Yang, and J. Wang, “Gold Nanobipyramid‐Enhanced Hydrogen Sensing with Plasmon Red Shifts Reaching ≈140 nm at 2 vol% Hydrogen Concentration,” Adv. Opt. Mater. 5(24), 1700740 (2017). [CrossRef]  

26. N. Strohfeldt, A. Tittl, and H. Giessen, “Long-term stability of capped and buffered palladium-nickel thin films and nanostructures for plasmonic hydrogen sensing applications,” Opt. Mater. Express 3(2), 194 (2013). [CrossRef]  

27. R. Griessen, N. Strohfeldt, and H. Giessen, “Thermodynamics of the hybrid interaction of hydrogen with palladium nanoparticles,” Nat. Mater. 15(3), 311–317 (2016). [CrossRef]   [PubMed]  

28. R. Bardhan, L. O. Hedges, C. L. Pint, A. Javey, S. Whitelam, and J. J. Urban, “Uncovering the intrinsic size dependence of hydriding phase transformations in nanocrystals,” Nat. Mater. 12(10), 905–912 (2013). [CrossRef]   [PubMed]  

29. A. Baldi, T. C. Narayan, A. L. Koh, and J. A. Dionne, “In situ detection of hydrogen-induced phase transitions in individual palladium nanocrystals,” Nat. Mater. 13(12), 1143–1148 (2014). [CrossRef]   [PubMed]  

30. S. Syrenova, C. Wadell, F. A. A. Nugroho, T. A. Gschneidtner, Y. A. Diaz Fernandez, G. Nalin, D. Świtlik, F. Westerlund, T. J. Antosiewicz, V. P. Zhdanov, K. Moth-Poulsen, and C. Langhammer, “Hydride formation thermodynamics and hysteresis in individual Pd nanocrystals with different size and shape,” Nat. Mater. 14(12), 1236–1244 (2015). [CrossRef]   [PubMed]  

31. T. C. Narayan, A. Baldi, A. L. Koh, R. Sinclair, and J. A. Dionne, “Reconstructing solute-induced phase transformations within individual nanocrystals,” Nat. Mater. 15(7), 768–774 (2016). [CrossRef]   [PubMed]  

32. A. Ulvestad, M. J. Welland, W. Cha, Y. Liu, J. W. Kim, R. Harder, E. Maxey, J. N. Clark, M. J. Highland, H. You, P. Zapol, S. O. Hruszkewycz, and G. B. Stephenson, “Three-dimensional imaging of dislocation dynamics during the hydriding phase transformation,” Nat. Mater. 16(5), 565–571 (2017). [CrossRef]   [PubMed]  

33. W. E. Vargas, I. Rojas, D. E. Azofeifa, and N. Clark, “Optical and electrical properties of hydrided palladium thin films studied by an inversion approach from transmittance measurements,” Thin Solid Films 496(2), 189–196 (2006). [CrossRef]  

34. I. Zorić, E. M. Larsson, B. Kasemo, and C. Langhammer, “Localized surface plasmons shed light on nanoscale metal hydrides,” Adv. Mater. 22(41), 4628–4633 (2010). [CrossRef]   [PubMed]  

35. Y. Shen, X. She, and C. Jin, “Mechanically Reconfigurable Pd Nanogroove Array: An Ultrasensitive Optical Hydrogen Detector,” ACS Photonics 5(4), 1334–1342 (2018). [CrossRef]  

36. N. Strohfeldt, J. Zhao, A. Tittl, and H. Giessen, “Sensitivity engineering in direct contact palladium-gold nano-sandwich hydrogen sensors,” Opt. Mater. Express 5(11), 2525–2535 (2015). [CrossRef]  

37. A. Hierro-Rodriguez, I. T. Leite, P. Rocha-Rodrigues, P. Fernandes, J. P. Araujo, P. A. Jorge, J. L. Santos, J. M. Teixeira, and A. Guerreiro, “Hydrogen sensing via anomalous optical absorption of palladium-based metamaterials,” Nanotechnology 27(18), 185501 (2016). [CrossRef]   [PubMed]  

38. N. R. Fong, P. Berini, and R. N. Tait, “Hydrogen sensing with Pd-coated long-range surface plasmon membrane waveguides,” Nanoscale 8(7), 4284–4290 (2016). [CrossRef]   [PubMed]  

39. M. Matuschek, D. P. Singh, H. H. Jeong, M. Nesterov, T. Weiss, P. Fischer, F. Neubrech, and N. Liu, “Chiral Plasmonic Hydrogen Sensors,” Small 14(7), 1702990 (2018). [CrossRef]   [PubMed]  

40. F. Gu, H. Zeng, Y. B. Zhu, Q. Yang, L. K. Ang, and S. Zhuang, “Single‐Crystal Pd and its Alloy Nanowires for Plasmon Propagation and Highly Sensitive Hydrogen Detection,” Adv. Opt. Mater. 2(2), 189–196 (2014). [CrossRef]  

41. S. Chang, S. Sen, P. Zhang, B. Gyarfas, B. Ashcroft, S. Lefkowitz, H. Peng, and S. Lindsay, “Palladium Electrodes for Molecular Tunnel Junctions,” Nanotechnology 23(42), 425202 (2012). [CrossRef]   [PubMed]  

42. M. W. Knight, N. S. King, L. Liu, H. O. Everitt, P. Nordlander, and N. J. Halas, “Aluminum for Plasmonics,” ACS Nano 8(1), 834–840 (2014). [CrossRef]   [PubMed]  

43. G. Maidecchi, G. Gonella, R. Proietti Zaccaria, R. Moroni, L. Anghinolfi, A. Giglia, S. Nannarone, L. Mattera, H. L. Dai, M. Canepa, and F. Bisio, “Deep ultraviolet plasmon resonance in aluminum nanoparticle arrays,” ACS Nano 7(7), 5834–5841 (2013). [CrossRef]   [PubMed]  

44. C. W. Cheng, Y. J. Liao, C. Y. Liu, B. H. Wu, S. S. Raja, C. Y. Wang, X. Li, C. K. Shih, L. J. Chen, and S. Gwo, “Epitaxial Aluminum-on-Sapphire Films as a Plasmonic Material Platform for Ultraviolet and Full Visible Spectral Regions,” ACS Photonics 5(7), 2624–2630 (2018). [CrossRef]  

45. F. Bisio, R. Proietti Zaccaria, R. Moroni, G. Maidecchi, A. Alabastri, G. Gonella, A. Giglia, L. Andolfi, S. Nannarone, L. Mattera, and M. Canepa, “Pushing the high-energy limit of plasmonics,” ACS Nano 8(9), 9239–9247 (2014). [CrossRef]   [PubMed]  

46. M. Castro-Lopez, D. Brinks, R. Sapienza, and N. F. van Hulst, “Aluminum for Nonlinear Plasmonics: Resonance-Driven Polarized Luminescence of Al, Ag, and Au Nanoantennas,” Nano Lett. 11(11), 4674–4678 (2011). [CrossRef]   [PubMed]  

47. L. Zhou, C. Zhang, M. J. McClain, A. Manjavacas, C. M. Krauter, S. Tian, F. Berg, H. O. Everitt, E. A. Carter, P. Nordlander, and N. J. Halas, “Aluminum Nanocrystals as a Plasmonic Photocatalyst for Hydrogen Dissociation,” Nano Lett. 16(2), 1478–1484 (2016). [CrossRef]   [PubMed]  

48. K. Kumar, H. Duan, R. S. Hegde, S. C. Koh, J. N. Wei, and J. K. Yang, “Printing colour at the optical diffraction limit,” Nat. Nanotechnol. 7(9), 557–561 (2012). [CrossRef]   [PubMed]  

49. X. M. Goh, Y. Zheng, S. J. Tan, L. Zhang, K. Kumar, C. W. Qiu, and J. K. W. Yang, “Three-dimensional plasmonic stereoscopic prints in full colour,” Nat. Commun. 5(1), 5361 (2014). [CrossRef]   [PubMed]  

50. M. Miyata, H. Hatada, and J. Takahara, “Full-Color Subwavelength Printing with Gap-Plasmonic Optical Antennas,” Nano Lett. 16(5), 3166–3172 (2016). [CrossRef]   [PubMed]  

51. K. Ikeda, S. Uchiyama, T. Mai, and K. Murakoshi, “Hydrogen-Induced Tuning of Plasmon Resonance in Palladium–Silver Layered Nanodimer Arrays,” ACS Photonics 2(1), 66–72 (2015). [CrossRef]  

52. A. B. Dahlin, J. O. Tegenfeldt, and F. Höök, “Improving the instrumental resolution of sensors based on localized surface plasmon resonance,” Anal. Chem. 78(13), 4416–4423 (2006). [CrossRef]   [PubMed]  

53. Z. Dong, J. Ho, Y. F. Yu, Y. H. Fu, R. Paniagua-Dominguez, S. Wang, A. I. Kuznetsov, and J. K. W. Yang, “Printing Beyond sRGB Color Gamut by Mimicking Silicon Nanostructures in Free-Space,” Nano Lett. 17(12), 7620–7628 (2017). [CrossRef]   [PubMed]  

54. Weber and J .Marvin, Handbook of Optical Materials (CRC, 2003).

55. C. Wadell and C. Langhammer, “Drift-corrected nanoplasmonic hydrogen sensing by polarization,” Nanoscale 7(25), 10963–10969 (2015). [CrossRef]   [PubMed]  

56. J. He, N. S. Villa, Z. Luo, S. An, Q. Shen, P. Tao, C. Song, J. Wu, T. Deng, and W. Shang, “Integrating plasmonic nanostructures with natural photonic architectures in Pd-modified Morpho butterfly wings for sensitive hydrogen gas sensing,” RSC Advances 8(57), 32395–32400 (2018). [CrossRef]  

57. C. Zhang, H. Zhao, L. Zhou, A. E. Schlather, L. Dong, M. J. McClain, D. F. Swearer, P. Nordlander, and N. J. Halas, “Al-Pd Nanodisk Heterodimers as Antenna-Reactor Photocatalysts,” Nano Lett. 16(10), 6677–6682 (2016). [CrossRef]   [PubMed]  

58. D. F. Swearer, H. Zhao, L. Zhou, C. Zhang, H. Robatjazi, J. M. P. Martirez, C. M. Krauter, S. Yazdi, M. J. McClain, E. Ringe, E. A. Carter, P. Nordlander, and N. J. Halas, “Heterometallic antenna-reactor complexes for photocatalysis,” Proc. Natl. Acad. Sci. U.S.A. 113(32), 8916–8920 (2016). [CrossRef]   [PubMed]  

59. S. Mukherjee, F. Libisch, N. Large, O. Neumann, L. V. Brown, J. Cheng, J. B. Lassiter, E. A. Carter, P. Nordlander, and N. J. Halas, “Hot electrons do the impossible: plasmon-induced dissociation of H2 on Au,” Nano Lett. 13(1), 240–247 (2013). [CrossRef]   [PubMed]  

60. D. Sil, K. D. Gilroy, A. Niaux, A. Boulesbaa, S. Neretina, and E. Borguet, “Seeing Is Believing: Hot Electron Based Gold Nanoplasmonic Optical Hydrogen Sensor,” ACS Nano 8(8), 7755–7762 (2014). [CrossRef]   [PubMed]  

61. M. L. Brongersma, N. J. Halas, and P. Nordlander, “Plasmon-induced hot carrier science and technology,” Nat. Nanotechnol. 10(1), 25–34 (2015). [CrossRef]   [PubMed]  

62. S. Bagheri, N. Strohfeldt, F. Sterl, A. Berrier, A. Tittl, and H. Giessen, “Large-Area Low-Cost Plasmonic Perfect Absorber Chemical Sensor Fabricated by Laser Interference Lithography,” ACS Sens. 1(9), 1148–1154 (2016). [CrossRef]