Collective photothermal effect of Al2O3‐supported spheroidal plasmonic Ru nanoparticle catalysts in the sunlight‐powered Sabatier reaction

Plasmon catalysis is an interesting technology concept for powering chemical processes with light. Here, we report the use of various Al2O3‐supported Ru spheroidal nanoparticles as catalyst for the low‐temperature conversion of CO2 and H2 to CH4 (Sabatier reaction), using sunlight as energy source. At high loadings of Ru spheroidal nanoparticles (5.9 % w/w), we observe a sharp increase in the rate of the sunlight powered reaction when compared to the reaction in dark at the same catalyst bed temperature. Based on our results we exclude plasmon coupling as cause, and attribute the rate enhancement to collective photothermal heating of the Al2O3‐supported Ru nanoparticles.

Plasmon catalysis is an interesting technology concept for powering chemical processes with light. Here, we report the use of various Al 2 O 3 -supported Ru spheroidal nanoparticles as catalyst for the low-temperature conversion of CO 2 and H 2 to CH 4 (Sabatier reaction), using sunlight as energy source. At high loadings of Ru spheroidal nanoparticles (5.9 % w/w), we observe a sharp increase in the rate of the sunlight powered reaction when compared to the reaction in dark at the same catalyst bed temperature. Based on our results we exclude plasmon coupling as cause, and attribute the rate enhancement to collective photothermal heating of the Al 2 O 3 -supported Ru nanoparticles.
Driven by the global aspiration to achieve a CO 2 -neutral society, the conversion of CO 2 to the energy carrier CH 4 via the Sabatier reaction has recently experienced its renaissance. [1] This reaction offers the prospect of a reduction of anthropogenic CO 2 emissions and long-term storage of renewable energy, producing the safe and easy-to-handle energy carrier methane that possesses a high gravimetric storage density. Conventionally, the Sabatier reaction (CO 2 + 4 H 2 !CH 4 + 2 H 2 O) [2] is performed in dark and catalysed by supported metal nanoparticles (e. g. Ni, Ru, Rh on supports such as alumina), and requires thermal activation at temperatures between 300°C and 500°C. [3] Instead of conventional heating, sunlight would be an appealing and sustainable alternative for powering this process. To pave the way to shift from conventional thermally activated processes to sustainable, solely sunlight-powered ones, plasmon-enhanced chemical reactions have very recently become a topic of intense research. [4] Plasmon catalysis makes use of the localized surface plasmon resonance (LSPR) of metal nanocatalysts for light harvesting. Upon illumination of a plasmonic nanocatalyst a coherent electron oscillation occurs in these particles, which dephases and generates hot electrons. [5] These can then either transfer into an electron accepting orbital of a nearby adsorbate, or thermalize resulting in an increased temperature of the catalyst. Because of the electronic oscillation, the plasmonic catalyst may also behave as an electromagnetic dipole and emit light coherently at the same frequency. [4a] While part of this emitted light is scattered to the far field, the other is concentrated at the metal surface. Ergo, plasmonic nanoparticles can be efficient sources of electrons, heat and light, and as such boost the activity of chemical reactions, influence their selectivity and/or provide a tool for spatial and temporal control. Depending on their type of metal, size, and shape, plasmonic nanocatalysts can harvest a particular part of the sunlight spectrum. [6] Combining plasmonic nanoparticles with several sizes and shapes within a single catalyst could be exploited to harvest the energy from the entire solar spectrum reaching the earth's surface.
For the Sabatier process, Liu and co-workers showed that the plasmonic properties of TiO 2 -supported Rh nanospheres enable methanation of CO 2 at low temperature (< 250°C) through use of the entire solar spectrum, and claim a combination of photothermal and hot electron contribution to explain their increased activity upon illumination. [ more, the groups of Corma [8] and Ye [9] have reported supported Ni and group 8 nanocatalysts, respectively, for photomethanation of CO 2 . Corma and co-workers propose a non-thermal contribution to explain their increased activity, [8] and Ye et al. applied a high intensity light source to promote their catalyst to a temperature of approx. 300°C (bulk photothermal heating). [9] Very recently, Dai and Sun reviewed the research performed on 'the reduction of carbon dioxide on photoexcited nanoparticles of VIII group metals', and detailed all claims with respect to photothermal and hot electron contributions made for 47 different catalyst systems. [10] Recently, Lee and co-workers proposed that direct photoexcitation of CO 2 , when adsorbed to the surface of Ru nanoparticles, is possible with visible light. [11] They observed a linear relationship between the rate of the Sabatier reaction and the intensity of their light source. They claim that for silica supported Ru nanoparticles of a size between 2.6 nm and 17.1 nm the observed increase in CO 2 conversion is not related to light absorption of the nanocatalyst, but to the decreased HOMO-LUMO gap of adsorbed CO 2 . They supported their claim by DFT calculations showing that the HOMO-LUMO gap decreases from 8.5 eV to 2.4 eV upon adsorption to the Ru surface, and exclude photothermal heating, albeit without measurement of the catalyst bed temperature. [11] To the best of our knowledge, this is the only study to date claiming direct photoexcitation of CO 2 as cause for the observed rate enhancement of a metal nanoparticle catalyzed, light-powered Sabatier reaction. Also our group demonstrated an efficient sunlight-powered Sabatier reaction, driven by the LSPR of alumina-supported Ru nanorods. [12] We have demonstrated that individual Ru nanorods efficiently harvest sunlight based on their broadband LSPR. For this Ru nanorod catalyst, we identified a large 'nonthermal' contribution at a slightly elevated sunlight intensity of 8.5 kW m À 2 (8.5 sun), resulting in a high photon-to-methane conversion efficiency (PTM) of 55 % [12] over the whole solar spectrum. The PTM is the quotient of the increase in reaction rate upon illumination and the rate of incident photons. [12] It carries the same definition as the so-called 'apparent quantum yield' introduced by Liu and coworkers, [7] and quantifies the 'nonthermal' share of the reaction. In contrast to nanorods, the LSPR of spheroidal Ru nanoparticles is positioned in the UV. [12,13] Because the catalyst was capable of harvesting UV light only, the PTM achieved with spheroidal particles (5 % w/w Ru at a reactor temperature 150°C and light intensity of 8.5 sun) was much lower than for rods (same loading and conditions). [12] The difference in PTM under these conditions was 41.5 %. [12] It is an intrinsic limitation of plasmon catalysis that catalytically highly active metal nanospheres for CO 2 methanation display only a weak plasmon resonance, typically positioned in the UV. [12,13] Strong plasmonic metals like Ag, Au and Al, on the other hand do not catalyse methanation of CO 2 . Here, we report plasmonic Al 2 O 3 -supported Ru spheroidal nanoparticles (d~1 nm) as catalyst for the low-temperature conversion of CO 2 and H 2 to CH 4 . At high loadings of Ru spheroidal nanoparticles (5.9 % w/w), we observe an increase in the ratio of reaction rate upon illumination vs. in dark at the same catalyst bed temperature. Based on our results we consider photothermal heating as main contributor. We rule out plasmon coupling as cause for the observed increase in the ratio of reaction rate upon illumination vs. in dark at high Ru loadings, and attribute this phenomenon to collective photothermal heating of the Al 2 O 3supported Ru nanoparticles.
To validate whether collective effects, viz. plasmon coupling and/or collective photothermal heating, play a role in the sunlight-powered Ru-catalysed Sabatier reaction, we prepared Al 2 O 3 -supported catalysts with various loadings of Ru nanospheres. For this purpose, we impregnated pre-calcined γ-Al 2 O 3 with Ru 3 (CO) 12 , and heated the resulting material to 450°C in a tube furnace under N 2 atmosphere (SI S1). Using this preparation method, we prepared three catalysts with a Ru loading of 3.6 % w/w, 4.9 % w/w and 5.9 % w/w, as determined by inductively coupled plasma -optical emission spectroscopy (ICP-OES, SI S2). The size and shape of the Ru nanoparticles has been analysed using high angle annular dark field scanning transmission electron microscopy (HAADF-STEM, Figure 1) In the HAADF-STEM analyses, for all three catalysts small, non-agglomerated spheroidal Ru nanoparticles (nanospheres or faceted) can be observed, that are randomly distributed over the Al 2 O 3 support. The average radius of the spheroidal Ru nanoparticles was similar for all three catalysts, with a diameter of 0.8 � 0.2 nm (3.6 % w/w), 1.0 � 0.3 nm (4.9 % w/w) and 1.1 � 0.1 nm (5.9 % w/w). The average interparticle distance cannot be determined from these images, because the HAADF-STEM image is a 2D projection of the 3D catalyst (SI S3). XRD analysis indicates the presence of crystalline Ru nanoparticles on the crystalline γ-Al 2 O 3 support, although peak broadening due to the small diameter of the Ru particles makes interpretation of the XRD difficult (SI S4).
In our previous work we have observed a strong rate enhancement for Al 2 O 3 -supported Ru nanorods when the catalyst bed temperature reached approximately 200-220°C resulting from a combination of the reactor temperature of 150°C and photothermal heating upon illumination with artificial sunlight. [12] To determine the activity of our three prepared catalysts in the sunlight-powered Sabatier process, the catalysts were tested at a catalyst bed temperature of Figure 1. HAADF-STEM of Ru nanoparticle catalysts with a Ru loading of (a) 3.6 % w/w, (b) 4.9 % w/w and (c) 5.9 % w/w, as determined by ICP-OES.. approximately 220°C, realised either through conventional heating of the reactor to this temperature, or through combined heating and illumination. In the latter case, a solar simulator was applied for illumination of the catalysts, using a light intensity of 6.2 suns (for spectrum see SI S5, Figure S7). Additional heating of the reactor was applied to increase the measured catalyst bed temperature to approximately 220°C in all three cases. This comparative study of the reaction in dark and under illumination with sunlight clearly shows that illumination yields a higher catalytic activity ( Figure 2).
The reaction rate upon illumination with 6.2 suns, all at a measured catalyst bed temperature of approx. 220°C, increased from 0.63 mol CH 4 g Ru À 1 h À 1 (3.6 % Ru w/w) and 0.77 mol CH 4 g Ru À 1 h À 1 (4.9 % Ru w/w) to 5.09 mol CH 4 g Ru À 1 h À 1 (5.9 % Ru w/ w). The ratio between the reaction rate upon illumination and in dark also increased from 2.2 (3.6 % Ru w/w) and 1.8 (4.9 % Ru w/w) to 12.3 (5.9 % Ru w/w). Based on these results, we conclude that there is a collective effect of the Ru nanoparticles which causes this boost in activity.
We have studied the catalyst with 5.9 % w/w Ru in more detail, through investigating the activity at a similar catalyst bed temperature of 210°C, achieved with different combinations of conventional heating and solar illumination (Figure 3). Reactions were performed in dark, and for the combinations 190°C reactor temperature + 1.3 suns, 172°C reactor + 3.7 suns and 150°C reactor + 6.2 suns. With an increasing share of solar illumination in the combination, the reaction rate at 210°C catalyst bed temperature increased from 0.43 mol CH 4 g Ru À 1 h À 1 in dark to 5.09 mol CH 4 g Ru À 1 h À 1 for the combination with 6.2 suns.
Upon further increase of Ru loading on our catalyst we observe an increase in Ru particle size (SI S1 and S3). Due to the increase in particle size the catalysis results using the catalyst with higher loading are not directly comparable to the results achieved with Ru loading of up to 5.9 %. However, similar trends are observed; i. e. the catalyst bed heats up upon illumination, the ratio between the reaction rate upon illumination and in dark is higher than 2, and a non-linear dependency of reaction rate on the light intensity is observed (SI S7).
To determine the cause for the collective effect upon illumination and at the higher catalyst loading (in this series 5.9 % w/w Ru), we studied the optical properties of the catalyst using UV-vis-NIR spectrophotometry. For that purpose, freshly prepared Al 2 O 3 -supported Ru catalysts powder (d~1.8 nm, 2.0 %, 5.9 % and 7.8 % as determined by ICP-OES, SI S9) was dispersed in diethylene glycol, and transmission (T) measurements were performed on the resulting stirred slurries for the wavelength range of 200 nm to 800 nm (Figure 4, displayed as relative absorption of Ru/Al 2 O 3 vs. absorption of Al 2 O 3 ). With the catalysts, we prepared slurries of the same concentration Ru, viz. 0.252 mg ml À 1 . These transmission measurements clearly show that the LSPR is positioned in the UV for all Ru loadings. Conversion-time profiles for solar methanation at 6.2 suns (ο) vs methanation in dark ( * ) using a catalyst with (a) 3.6 % w/w Ru [max. production approx. 620 mmol CH 4 g Ru À 1 ], (b) 4.9 % w/w Ru [max. production approx. 470 mmol CH 4 g Ru ] or (c) 5.9 % w/w Ru [max. production denoted by dashed line approx. 400 mmol CH 4 g Ru À 1 ]. Reaction conditions for all experiments: reaction mixture of H 2 /CO 2 /N 2 (5 : 1 : 1) [12] at 3.5 bar pressure, 300 mg of Ru/Al 2 O 3 catalyst, catalyst bed temperature approximately 220°C. Conversion-time profiles and initial rates for the solar methanation using a catalyst with 5.9 % Ru loading. Reaction conditions: reaction mixture of H 2 /CO 2 /N 2 (5 : 1 : 1) at 3.5 bar pressure, 300 mg of Ru/Al 2 O 3 catalyst, catalyst bed temperature approximately 210°C, various mixes of reactor temperature and solar illumination, viz. 215°C reactor temperature + 0 sun (*), 190°C reactor temperature + 1.3 suns (◇), 172°C reactor temperature + 3.7 suns (&) and 150°C reactor temperature + 6.2 suns (*).