Direct Observation of Ni Nanoparticle Growth in Carbon-Supported Nickel under Carbon Dioxide Hydrogenation Atmosphere

Understanding nanoparticle growth is crucial to increase the lifetime of supported metal catalysts. In this study, we employ in situ gas-phase transmission electron microscopy to visualize the movement and growth of ensembles of tens of nickel nanoparticles supported on carbon for CO2 hydrogenation at atmospheric pressure (H2:CO2 = 4:1) and relevant temperature (450 °C) in real time. We observe two modes of particle movement with an order of magnitude difference in velocity: fast, intermittent movement (vmax = 0.7 nm s–1) and slow, gradual movement (vaverage = 0.05 nm s–1). We visualize the two distinct particle growth mechanisms: diffusion and coalescence, and Ostwald ripening. The diffusion and coalescence mechanism dominates at small interparticle distances, whereas Ostwald ripening is driven by differences in particle size. Strikingly, we demonstrate an interplay between the two mechanisms, where first coalescence takes place, followed by fast Ostwald ripening due to the increased difference in particle size. Our direct visualization of the complex nanoparticle growth mechanisms highlights the relevance of studying nanoparticle growth in supported nanoparticle ensembles under reaction conditions and contributes to the fundamental understanding of the stability in supported metal catalysts.

Section S1. Catalytic results Table S1. Results from catalytic test showing the GC peak concentrations of inert SiC and of 0.5 mg catalyst (sieve fraction of 38-75 µm) during exposure to CO2 hydrogenation conditions at 450 °C and 1 bar (top rows) and in the absence of CO2 (replaced by N2, bottom rows). Even at the high GHSVs used, CO2 was converted to CO (main product) and CH4. In absence of CO2, the CH4 concentration was only <1% compared to the test where CO2 was present, showing that during CO2 hydrogenation, the formed CH4 mainly originated from CO2 conversion instead of support methanation. Gas  Section S2. Exposure to 8 e -A -2 s -1 for 10 s every 2 min Figure S1. TEM images of the data shown in Figure 1 before the start of the actual experiment: A) Directly after in situ reduction of the sample at 300 °C (5% H2/Ar). Subsequently the gas was switched to H2:CO2 = 4. B) After heating the sample to 400 °C, C) After heating the sample to 450 °C. D-F) TEM images as also shown in Figure 1A-C without highlights of the nanoparticles. For the details about the exact protocol indicating the different times, we refer to Figure S18. Figure S2. TEM images at different moments in time of the full region of Ni/GNP during the experiment where the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm, 4:1 H2:CO2 and an electron dose of 8 es -1 A -2 every 2 min for ~10 s. The two analyzed regions are highlighted (the analysis of region 1 is depicted in Figure 1 and of region 2 in Figure S2).

Figure S3
. Analysis of region 2 in Figure S2. A) Transmission electron microscopy images of Ni/GNP acquired after t = 2, 20, 30 and 50 min. The perimeter of all observed particles is highlighted in yellow. B) Average projected particle area and D) average respective particle diameter calculated from the projected particle areas assuming spherical shape as a function of time. In grey the standard deviation of the measurements is shown. To investigate beam effects, the experiments were repeated on a TEM chip the in situ TEM holder without exposure to the beam (red squares) and on a TEM chip in a fixed bed reactor (green circles). D) Total number of particles in field of view. All dashed lines are added as guide for the eye. During the experiment, the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm 4:1 H2:CO2 and an electron dose of 8 es -1 A -2 every 2 min for ~10 s.
Section S3. Exposure to pure Ar Figure S4. TEM images under pure argon flow at different moments in time of Ni/GNP during the experiment where the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm, Ar and an electron dose of 20 es -1 A -2 every 2 min for ~10 s. Barely any particle growth was observed during this experiment, however the support was not stable. This was possibly caused by traces of oxygen-containing contaminations, burning the carbon and feasibly oxidizing the nanoparticles, the formation of a mild plasma under the electron beam 1 , or remaining presence of water causing edging of the outer layers of the carbon structures under electron irradiation. 2 Section S5. Exposure to 20 e -A -2 s -1 for 10 s every 2 min Figure S5. TEM images at different moments in time of the full region of Ni/GNP during the experiment where the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm, 4:1 H2:CO2 and an electron dose of 20 es -1 A -2 every 2 min for ~10 s. The two analyzed region is highlighted and the analysis is depicted in Figure S6. Figure S6. Analysis of the highlighted region in Figure S5. A) Transmission electron microscopy images of Ni/GNP acquired after t = 2, 20, 30 and 50 min. The perimeter of all observed particles is highlighted in yellow. B) Average projected particle area and D) average respective particle diameter calculated from the projected particle areas assuming spherical shape as a function of time. In grey the standard deviation of the measurements is shown. To investigate beam effects, the experiments were repeated on a TEM chip the in situ TEM holder without exposure to the beam (red squares) and on a TEM chip in a fixed bed reactor (green circles). D) Total number of particles in field of view. All dashed lines are added as guide for the eye. During the experiment, the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm 4:1 H2:CO2 and an electron dose of 20 es -1 A -2 every 2 min for ~10 s.

Section S4. Overview beam damage check
Section S6. Continuous exposure to 20 e -A -2 s -1 Figure S7. TEM images at different moments in time of the full region of Ni/GNP during the experiment where the imaged areas were exposed 1 bar, 450 °C, 0.4 sccm, 4:1 H2:CO2 and a continuous electron dose of 20 es -1 A -2 . The analysis of the two highlighted regions is shown in Figure S8.  Figure S7. A,E) Transmission electron microscopy images of Ni/GNP acquired after t = 2, 20, 30 and 50 min. The perimeter of all observed particles is highlighted in yellow. B,F) Average projected particle area and D,G) average respective particle diameter calculated from the projected particle areas assuming spherical shape as a function of time. In grey the standard deviation of the measurements is shown. To investigate beam effects, the experiments were repeated on a TEM chip the in situ TEM holder without exposure to the beam (red squares) and on a TEM chip in a fixed bed reactor (green circles). D,H) Total number of particles in field of view. All dashed lines are added as guide for the eye. During the experiment, the imaged areas were  Figure S18) and the chip containing the sample was analyzed in a TEM inspection holder under vacuum (top row). Subsequently the holder was reassembled and the experiment was continued. At the end of the experiment (t = 52 min), the holder was disassembled again and the same areas of the chip were imaged in the inspection holder (middle row). The bottom row shows the histograms of the counted nanoparticles. The results of the 4 regions were used to determine the average particle diameter that is shown in Figure 2 in the manuscript and Table S2. The results of these 3 regions were used to determine the average particle diameter that is shown in Figure 2 in the manuscript and Table S2.
Section S8. Additional analysis particle movement During the experiment, the imaged area was continuously exposed to an electron dose of 20 e -A -2 s -1 . This particle moved via both the fast hopping mechanism and later the slower movement.