In Situ X-ray Microscopy Reveals Particle Dynamics in a NiCo Dry Methane Reforming Catalyst under Operating Conditions

Herein, we report the synthesis of a γ-Al2O3-supported NiCo catalyst for dry methane reforming (DMR) and study the catalyst using in situ scanning transmission X-ray microscopy (STXM) during the reduction (activation step) and under reaction conditions. During the reduction process, the NiCo alloy particles undergo elemental segregation with Co migrating toward the center of the catalyst particles and Ni migrating to the outer surfaces. Under DMR conditions, the segregated structure is maintained, thus hinting at the importance of this structure to optimal catalytic functions. Finally, the formation of Ni-rich branches on the surface of the particles is observed during DMR, suggesting that the loss of Ni from the outer shell may play a role in the reduced stability and hence catalyst deactivation. These findings provide insights into the morphological and electronic structural changes that occur in a NiCo-based catalyst during DMR. Further, this study emphasizes the need to study catalysts under operating conditions in order to elucidate material dynamics during the reaction.


Sample imaging
STXM experiments for the particle in Figure 1 were performed at the PolLux beamline of the Paul Scherrer Institute, while the particle in Figure 2 was measured at the Hermes beamline at SOLEIL synchrotron radiation facility. [1][2] The spatial resolution of the used zone plates was 35 nm. During the STXM measurements, the sample is raster-scanned in the x-and y-directions, and the transmission from a selected number of points is converted into their respective Optical Density (OD) maps. The resulting 2D image contains information regarding the elemental distribution, oxidation state, and the local coordination. In the stack mode, a series of the spectral elemental composition maps are collected in a selected energy range, including the pre-edge, edge and post-edge regions. As an example, to generate the elemental composition maps, shown in Figure S8, Co and Ni L-edges (Figures S10 and S11) were measured for each of the calcined particles. The STXM data obtained at the Ni L-edge illustrates that there are some regions in each of the three nanoparticles which have higher concentrations of Ni. In addition, by investigating the Co L-edge, we observe that Co is not evenly distributed though the particles and it appears that there are some Co-deficient regions. Next, to determine the relative distributions of the elements, spectral images of Ni and Co in the same particle were overlaid to obtain the Red, Green, and Blue (RGB) maps. To obtain the RGB maps we have to specify which element corresponds to red, green and blue. Since only two elements (Ni and Co) were probed during this study, blue and red were assigned to Co (thus resulting in violet), while green was used for Ni. Areas where Co and Ni have an equal concentration result in white areas on the maps (Figure 1). The Al K-edge of the support was not probed suring STXM measurements due to the low absorption cross section of Al K-edge and the thickness of the support. For the in situ STXM measurements, surface micromachined nanoreactors were employed. [3][4][5][6] The samples were loaded into the nanoreactors by drop casting the catalyst suspension on the inlet of the nanoreactors. Ex situ reference samples were prepared by directly drop casting the samples onto thin Si3N4 membranes.

Data processing
The acquired data were processed by using the aXis2000 software package (http://unicorn.mcmaster.ca/aXis2000.html) and OriginPro 9.0.

Nanoreactor
The nanoreactors employed for the in situ measurements were similar those utilized in references 3-5 and were produced by NanoInsight. These are silicon-based miniaturized chemical reactors realized by means of microfabrication techniques as described in reference 6. The nanoreactor is shown in Figure S13, it consists of a silicon-rich silicon nitride (SiNx) microchannel integrated on the surface of a silicon (Si) rectangular support. The microchannel has a length of 4.6 mm, a width of 0.3 mm, and a height of only 4.5 µm. To prevent its deformation under pressurization and to increase its strength, the top and bottom part of the microchannel are connected together by an array of micropillars. This structure allows pressures up to 9 bar to be reached. A microheater spiral made of a thin film of molybdenum (Mo) is integrated on the top-side of the central part of the microchannel. The microheater is equipped with four contacts to allow its powering while enabling the accurate reading of the temperature of the heated area. By flowing a current into the microheater its temperature and that of the central part of the microchannel can be increased from room temperature up to a maximum of 750 °C. As specified by the supplier, the accuracy of the set temperature is 2 °C for temperatures up to 350 °C, 4 °C for temperatures between 350 °C and 600 °C, and 7 °C for temperatures between 600 °C and 750 °C. An array of 41 circular windows each having a diameter of 6 µm, is integrated between the windings of the microheater spiral. The same number of windows is integrated on the bottom part of the microchannel and aligned to the top ones. The windows are made of SiNx and have a thickness of only 18 nm making them highly transparent to X-rays. The Si support is removed in a square area of 1 mm x 1 mm making the central part of the microchannel and the microheater suspended. This configuration provides the thermal isolation needed for the functioning of the microheater while allowing the transmission of soft X-rays. Finally, two small square-shaped inlet/outlet apertures are created by removing the Si support in correspondence of the ends of the microchannel. The inlet/outlet apertures are employed for flowing the liquid containing the nanoparticles and for the pressurization of the microchannel by means of gasses.

Experimental details STXM holder
Custom-made holders were designed and produced by NanoInsight to interface the nanoreactor and to allow its mounting on the interferometrically controlled piezoelectric stage of the Pollux-PSI and of the HERMES-Soleil STXM chambers. Each of the holders is provided with a housing slot and a retaining lid for the mounting and fixing of the nanoreactor. To allow transparency both the holder and the lid are provided with holes in correspondence of the central part of the microchannel. The holder integrates two gas channels ending in the housing slot in correspondence of the inlet/outlet apertures of the nanoreactor. Here a leak-tight connection is achieved by means of O-rings. The holders are equipped with four spring-loaded probe-needles connecting to the four contact-pads of the microheater. For the experiments, the gas channels of the holder were connected to flexible PEEK tubing. A gas feedthrough flange was employed to connect to the PEEK tubing from the external side of the STXM chamber. In a similar way, electrical wires were soldered to the probe-needles and brought to the external side of the STXM chamber by means of a flange provided with D-sub-9 electric feedthroughs ports.

Pressurization of the nanoreactor
A custom-made gas mixing and injection system designed and produced by NanoInsight was employed to flow the employed gas mixtures into the nanoreactor. The system makes use of mass flow meters and a backward pressure controller to mix up to four different gasses. A forward pressure controller is employed to withdraw the gas mixture from the mixing section and to inject it into the inlet of the nanoreactor with the desired pressure. During the in situ experiments the gas mixtures were flown into the nanoreactor by pressurizing the inlet at 2.0 bar for the DMR mixture, and at 3.0 bar or 6.0 bar for the 5% H2 in Ar mixture. At the same time vacuum was applied at the outlet of the nanoreactor by means of a scroll pump. In this configuration the pressure in the central part of the microchannel (reaction zone) is the mean value of the pressure difference between the inlet and the outlet of the nanoreactor. This is due to the fact that the microchannel is symmetric and that its flow conductivity is very low compared to that of the tubing connected at its inlet/outlet apertures. The outlet pressure was measured with a pressure gauge installed at the exit of the tubing connected at the outlet of the nanoreactor. This pressure changed with the pressure applied at the inlet of the nanoreactor and with the composition of the gas mixture. In any case, during the in-situ experiments the outlet pressure was always lower than 10 mbar. Therefore, the pressure in the reaction zone of the nanoreactor can be considered simply as the half of that applied at the inlet, that is 1.0 bar for the DMR mixture, and at 1.5 bar or 3.0 bar for the 5% H2 in Ar mixture

Powering and temperature measurement
The temperature in the reaction zone of the nanoreactor was increased by flowing a current through the Mo microheater. This was done while monitoring its electrical resistance, which provided a measure of the average temperature. For this purpose, a Keithley 2611 source meter was employed in a four-wire sensing configuration. For each temperature, the current was increased stepwise until the corresponding resistance was reached. The nanoreactors employed during the experiments were calibrated on a hotplate beforehand to generate a resistance-temperature table. Figure S1. Synthesis pathway to produce NiCoO/g-Al2O3 catalyst by means of the reverse micellar method. Figure S2. The RGB composition maps of the reduced (500 °C) NiCo particles shows that increasing the pressure from 1.5 bar to 3.0 does not influence particle morphology. For the particle reduced at 750°C, the morphology is changes only slightly when increasing the pressure from 1.5 bar to 3.0 bar. Figure S3. a, b) Overview STEM image of the calcined NiCoO/g-Al2O3 catalyst prior to the activation step. Both images illustrate that the catalyst is composed of small nanoparticles that have coalesced into large agglomerates with a cubic structure. Figure S4. XRD profile for the fresh, reduced and spent NiCoO/g-Al2O3 DMR catalyst. The green squares and orange triangles indicate that the calcined catalyst contains NiO and Co3O4 as active species. After exposure to DMR conditions the formation of coke is confirmed by the peak at approximately 26 ° (black diamond). Figure S5. Bulk EDX analysis on the NiCoO/g-Al2O3 reveals that the Ni:Co ratio in this sample is close to 9:1. Figure S6. TPR profile of the NiCoO/gAl2O3 catalyst measured from 30 °C to 800 °C. This profile indicates that at temperatures higher ~550 °C both Ni and Co are fully reduced the metallic forms.

Equation S1
Yield values during the activity tests were calculated by using eq S1: In this equation n corresponds to the molar flow rate of the species and v to the stoichiometric coefficient. The obtained yield values are in fractions.

Table S1
The conversion/yield values after exposure to DMR condition for 15 hours are shown in the Equation S2-3 The water yield was calculated by using the hydrogen and oxygen balance shown in the equations below respectively.
The formed water resulted in an incomplete balance, since this was not detected analytically. An Excel macro was employed to add H 2 O until the sum of the H-and O-balances amounted to 200 %, which closed both individual balances reliably. Figure S7. Activity measurements for the NiCoO/g-Al2O3 DMR catalyst. Conversion/Yield fractions during the DMR experiment under 7% CH4 and 9.5% CO2 in N2 and a flow rate of 490 mL/min. The oven temperature was set to 800 °C to achieve a temperature of ~750 °C at the reactor bed position. The degrees of conversion after 15 h were stabilized at 80% and 70% for CH4 and CO2, respectively. The corresponding yields for H2 and CO were ~70% and ~73%, respectively. Figure S8. a) Elemental composition maps of three different NiCoO particles which show the distribution of Ni (in green) and Co (in violet) for: a,d,g) the freshly calcined particles, b,c,h) following the reduction step at 500 °C under 5% H2 in Ar, where the Ni and Co oxides are reduced, c,f,i) at 750 °C and under 5%H2 in Ar flow where the Ni and Co Al2O4 are fully reduced, and j,k) after exposure to DMR conditions for 1 minute and 120 minutes, respectively. Figure S9. Elemental composition maps of the calcined NiCo particle showing that Co and Ni are mixed throughout the particles. However, there are some areas with high Ni content while other regions are Co deficient. Figure S10. Co L3-edge spectra of the calcined NiCoO/g-Al2O3 sample and during the reduction/activation process. At temperatures higher than 350 °C the Co gets fully reduced into metallic Co. Figure S11. Ni L3-edge spectra of the calcined NiCoO/g-Al2O3 sample and during the reduction/activation process. At temperatures higher than 500 °C the Ni gets fully reduced into metallic Ni.