Universal Scaling and Design Rules of Hydrogen-Induced Optical Properties in Pd and Pd-Alloy Nanoparticles

Abstract

Hydride-forming metal nanoparticles sustaining localized surface plasmon resonance have emerged as prototypical material to study the fundamentals of hydrogen-induced phase transformations. They have also been proposed as signal transducers in next-generation hydrogen sensors. However, despite high current interest in hydrogen sorption by nanomaterials in general and such sensors in particular, the correlations between nanoparticle size, shape, and composition, the amount of hydrogen absorbed, and the obtained optical response have not been systematically experimentally studied. Focusing on hydrogenated Pd, PdAu- and PdCu-alloy nanoparticles, which are of particular interest in hysteresis-free plasmonic hydrogen sensing, we find that at practically important Au/Pd and Cu/Pd ratios the optical response to hydrogen concentration is linear and, more interestingly, can be described by a single universal linear trend if constructed as a function of the H/Pd ratio, independent of alloy composition. In addition to this correlation, we establish that the amplitude of optical signal change is defined solely by the spectral plasmon resonance position in the non-hydrogenated state for a specific nanoparticle composition. Thus, it can be maximized by red-shifting the LSPR into the NIR spectral range via tailoring the particle size and shape. These findings further establish plasmonic sensing as an effective tool for studying metal-hydrogen interactions in nanoparticles of complex chemical composition. They also represent universal design rules for metal-hydride-based plasmonic hydrogen sensors, and our theoretical analysis predicts that they are applicable not only to the H/Pd/Au or H/Pd/Cu system investigated here but also to other H/Pd/Metal combinations.

Metal–polymer hybrid nanomaterials for plasmonic ultrafast hydrogen detection

“…In both backgrounds, down to the lowest H 2 concentration experimentally attainable, that is, 10 ppm and 20 ppm, respectively, a Δλ peak response is clearly discernible. This places our system among the most sensitive hydrogen sensors reported so far under simulated application conditions in air 29,30 , notably with significant potential to further enhance the sensitivity by, for example, tailoring the coating layer material and its thickness, and by optimizing the nanoparticle dimensions 22 . To this end, by extrapolating the LoD based on the experimentally derived noise in flow conditions ( Supplementary Fig.…”

“…Specifically, we chose a Pd 70 Au 30 alloy, as well as a pure Pd control. The chosen alloy composition is optimal because, at a lower Au fraction, sensor accuracy is compromised by hysteresis 20,21 (that is, the readout is non-specific as it depends on the hydrogen pressure history), while at a higher Au fraction the sensitivity is reduced 4,22 . We nanofabricated the alloy and Pd nanostructures with average particle dimensions of 190 nm (diameter) and 25 nm (height) according to the 'hole-mask colloidal lithography' method since particles of this size are structurally very stable and exhibit a distinct plasmon-related light-absorption peak.…”

“…Following exposure to 1,000 mbar of hydrogen, this peak further broadens and redshifts due to the change in both volume and permittivity of the metal nanoparticles induced by hydrogen population of the interstitial lattice sites 15 . This LSPR peak shift, Δλ peak , which is a linear function of the hydrogen concentration inside the nanoparticles expressed as the H/Pd ratio 22 , then serves as the basis for our nanoparticle-polymer hybrid optical hydrogen sensors.…”

“…These results contrast with earlier work reporting on nearly doubled hydrogen solubility in metal-organic framework-coated Pd nanoparticles 26 . To address this in more detail, we note that the sensitivity of a plasmonic hydrogen sensor is dictated by the spectral position of the plasmonic resonance, λ peak , in the non-hydrogenated state 22 . Adding a polymer coating spectrally redshifts λ peak (cf.…”

“…Data collection was carried out with QSoft (Biolin Scientific). The absorbed hydrogen content in the nanoparticles was calculated from the QCM frequency shift via the Sauerbrey equation 45 and analysis of the nanoparticle surface coverage based on SEM images of the crystal, as described in detail elsewhere 22 . All measurements were carried out at 30 °C.…”