SUPPPORTING INFORMATION

Transcription

1 SUPPPORTING INFORMATION Dry Reforming of Methane on Rh Doped Pyrochlore Catalysts: A Steady-State Isotopic Transient Kinetic Study Felipe Polo-Garzon a, Devendra Pakhare b, James J. Spivey c and David A. Bruce a* a Department of Chemical and Biomolecular Engineering, 7 Earle Hall, Clemson University, Clemson, SC 29634, USA b Pyrochem Catalyst Company, 161 Decimal Drive, Jeffersontown, KY 40299, USA c Department of Chemical Engineering, Louisiana State University, Baton Rouge, LA 70803, USA * Corresponding author. Phone: address: dbruce@clemson.edu S1 Experimental details S1.1 Catalyst synthesis The catalyst used in this work, a 2 wt% Rh-substituted lanthanum zirconate (LRhZ) pyrochlore, was provided by collaborators 1. This catalyst was synthesized by the modified Pechini Method 2 using lanthanum nitrate hexahydrate (La(NO3)3 6H2O), zirconium oxynitrate (ZrO(NO3)2 xh2o) and rhodium nitrate (Rh(NO3)3 2H2O) as metal precursors. In the synthesis, the dissolved metal precursors were combined. Later, the complexing agent citric acid solution and the polymerizing agent ethylene glycol were added. The solution was stirred at 70 C to form a gel and then, the polyesterification reaction was promoted at 0 C. The resulting resin is later calcined to form the pyrochlore material. The detailed synthesis procedure has been reported earlier 1. S1.2 Catalyst characterization Previous work performed extensive characterization of the LRhZ catalyst 3 by means of X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS) and temperature programmed reduction (TPR); concluding that the Rh dopant was located within the lattice of the pyrochlore crystal phase. In the present work, Scanning Transmission Electron Microscopy (STEM) and Energy Dispersive X-ray spectroscopy (EDX) were performed to ascertain catalyst morphology and the dispersion of Rh on the surface, respectively. SEM and EDX were performed using the 1

2 Hitachi Scanning Transmission Electron Microscope HD2000, with a field emission source and a resolution of 0.24 nm at 200 kv. The BET (Brunauer Emmett Teller) surface area of the catalyst was analyzed by N2 physisorption using a Micromeritics ASAP 2020 system g of LRhZ were degassed under a vacuum of 10-3 mmhg at 90 C for 10 h; after this, the temperature was ramped to 300 C at a rate of 10 C/min and held for 1 h before N2 physisorption data were collected at 77 K. S1.3 Kinetic measurements Reactor set-up and catalyst reduction Catalytic measurements were carried out in a straight tube quartz microreactor (4 mm i.d., 6.35 mm o.d.). A 9 mg catalyst sample was added to the reactor and held in place using quartz wool above and below the catalyst. The lower portion of the reactor tube was filled with quartz beads (2 mm diameter) so as to maintain the vertical position of the catalyst bed at a fixed height. Reactor heating was provided by a furnace (Applied Test Systems, Inc.) controlled by a programmable temperature controller, and the temperature in the catalytic bed was appropriately calibrated with respect to the set point value in the temperature controller for the furnace. Before catalyst testing, the sample was first reduced at 800 C and 2 psig using 73.9 cm 3 /min of 26.9 % (v/v) H2 (UHP) in Helium (industrial grade) for 1 hour (after heating to that temperature at a rate of approximately 35 C/min). Then, 65 cm 3 /min of He was used to purge the reactor system for min before DRM experiments at 800 C and 2 psig were carried out. The total flow rate of gaseous feed to the reactor was kept constant at 74.6 cm 3 /min and contained 64.8 cm 3 /min of 95.1% He + 4.9% Ar (Airgas), 4.9 cm 3 /min of CH4 (instrument grade, Airgas) and 4.9 cm 3 /min of 2 (instrument grade, Airgas). Note that all listed gas flow rates in this text are at standard temperature and pressure conditions, i.e., not reactions conditions. Initial DRM reactions were always at 800 C just after the reduction with H2 because the species created during reaction further reduced the surface of the catalyst. After steady-state was reached at 800 C, the catalyst temperature was then changed (if needed) to the desired reaction temperature. The product gases exiting the reactor were sent to the mass spectrometer (MS) (Pfeiffer Vacuum) for analysis via 1/16-inch stainless steel capillary tubing. The MS was connected to a computer for high-speed continuous data acquisition using the software Balzers Quadstar 422 version

3 The tubing connecting the reactor and MS analysis system were maintained at 0 C, and the inlet assembly to the mass spectrometer was held at 70 C to avoid water condensation. The MS recorded data approximately every 0.6 s, which was sufficiently fast to measure transients in species concentrations as well as approximate values for surface residence times at the highest temperature studied. Steady-state isotopic transient kinetic analysis (SSITKA) set-up The masses followed with respect to time in the mass spectrometer were 44 (2), 28 (), 15 (CH4), 2 (H2), 45 ( 2), 29 ( ), 17 ( CH4) and 40 (Ar). The MS signal used for masses 28 () and 29 ( ) were adjusted so that the contribution from 2 and 2 ionization, respectively, are appropriately subtracted. Although water is produced, its mass (18) was not followed in order to optimize the recording speed of the mass spectrometer, and in work reported by Pakhare et al. 3 on the same catalyst it is shown that the quantity of water produced from the Reverse Water Gas Shift (RWGS) reaction is small, and RWGS takes place up to approximately 700 C. SSITKA measurements were performed using two different labeled gases. First, labeled methane ( CH4, 99 atom % C, Aldrich) was used. Therefore, a switch between 64.8 cm 3 /min Ar/He (95.1 % He+ 4.9 % Ar (Airgas)) with 4.9 cm 3 /min CH4 and 64.8 cm 3 /min He with 4.9 cm 3 /min CH4 was made, while holding the 2 flow constant at 4.9 cm 3 /min. Second, labeled carbon dioxide ( 2, 99 atom % C, <3 atom % 18 O, Aldrich) was used. Therefore, a switch between 64.8 cm 3 /min Ar/He (95.1 % He % Ar (Airgas)) with 4.9 cm 3 /min 2 and 64.8 cm 3 /min He with 4.9 cm 3 /min 2 was made, while holding the CH4 flow constant at 4.9 cm 3 /min. The switch was made using a Valco 2-position valve with an electric actuator without disturbing any reaction condition (i.e., the total gas flow rate and reaction pressure were kept constant at 74.6 cm 3 /min and 2 psig during the switch). In addition, two back pressure regulators in the system were used to keep the pressure of the flow going through the reactor constant at 2 psig before and after the switch. The gas-phase holdup for the reaction system was measured using the dilute Ar tracer in the He carrier gas. The volume that species have to travel and is inherent to the system was calculated to be approximately 4.4 cm 3 using the normalized Ar signal. 3

4 The SSITKA apparatus used in this study (see Figure S1) readily enabled the effects of switching between labeled and unlabeled gases to be quantized, which in turn provided information about the rate of carbon migration through different species on the catalyst surface at steady-state reaction conditions. Surface kinetic parameters, such as the average surface residence time, surface concentration of intermediates and turnover frequency were determined following the SSITKA formalism 4. Figure S1 Schematic diagram of the SSITKA reaction system. Normalized transient responses SSITKA is an experimental technique that provides information about surface kinetic parameters, such as, surface residence time, surface concentration of intermediates and turnover frequencies during steady-state reaction conditions; information which is practically inaccessible computationally or using any other surface analysis technique. 4

5 After reaching steady-state reaction conditions, the switch from an unlabeled reactant to a labeled one leads to the replacement of unlabeled surface intermediates by the labeled ones, without altering the steady-state conditions of the reaction (assuming the molecular mass and vibrational characteristics of the labeled and unlabeled species are similar). The speed with which the surface intermediates are replaced gives information about the surface residence time and the concentration of intermediates on the surface. As the switch from unlabeled to labeled reactants provokes a transient decay of the unlabeled product species and a transient rise of the labeled ones, the normalized step decay transient response for a product P is defined as: P P F (t) r (t) / r ss (1) P Where, the normalized step decay ( F(t)) is the ratio between the transient reaction rate during the switch (r P (t)) and the reaction rate at steady-state conditions (rss). Therefore, it can be calculated as, P F(t) y(t) y y y ss ss where y(t) is the mole fraction of a reactor effluent species as a function of time (t), yss is the mole fraction at steady-state, which is just before the switch at t = 0, and y is the mole fraction long after the switch, when the concentration of labeled species completely replaces that of the unlabeled species. The responses recorded by the MS detector must be corrected for the contribution of gas-phase hold up in the system. This correction is achieved by measuring the outlet concentration profile of a dilute inert gas (argon). In this work, the system gas-phase hold up is measured using an argon tracer that is included with the unlabeled reactant streams. 4,5 5

6 S2 Parameters calculated from SSITKA Concentration of surface intermediates The average surface residence time (τavg) is calculated as the area between the normalized transient responses of the product, F (t), and that of the inert gas tracer, F Ar (t). Ar avg 0 F (t) F (t) dt This area corresponds to the average time that carbon-containing species (since C labeled gases were used), which are in the reactive pathway to form, spend on the surface. The steady-state reaction rate for carbon monoxide ( r ss ) is calculated as: r ss. PV 1 RTroom mcat where, P is the reactor pressure, V. is the volumetric flow of produced, R is the ideal gas constant, Troom is the ambient temperature (temperature at which the volumetric flows are measured) and mcat is the mass of the catalyst in the reactor. Finally, the concentration of surface intermediates (N ) is calculated from the rate of production of and the average surface residence time as follows 4, N r ss avg From N, the coverage of surface intermediates (intermediates per active site) can be obtained as follows, cover age N N A Av BET A m 6

7 The BET surface area was 8.72 m 2 /g, and the area per metal atom (including Rh, Zr and La) used for the calculations was 1.28 x10-19 m 2 /atom, which corresponds to the area per metal atom assuming the plane predominantly exposed at the surface of the catalyst is the (111). In our earlier computational work 6, the plane (111) was concluded as the most thermodynamically stable and the catalytically active for DRM over LRhZ pyrochlores. Furthermore, even though the planes (011) and (001) were found less thermodynamically stable, they present similar area per metal atom (1.05 x m 2 /atom for (011) and 1.48x10-19 m 2 /atom for (111)) as the plane (111), and averaging the area per metal atom for these 3 planes gives 1.27x10-19 m 2 /atom, a value very close to the one of the plane (111) alone. Before conducting any experiment, the catalyst was reduced to achieve greater exposure of the metal atoms at the surface as detailed in the section Reactor set-up and catalyst reduction. 7

8 Figure S2 Reaction energies for elementary steps in the main reaction pathway for DRM on the LRhZ pyrochlore surface as identified by DFT methods. A* means that species A is adsorbed on the catalyst surface. DFT data depicted in this figure are taken from 6. 8

9 2E C 2 CH 4 Ar 1E 09 2 CH 4 2E 09 1E Relative CH time (s) 4 + CH C H 2 + Ar Figure S3 DRM products transient response following a reactant switch from 2 to 2 using the LRhZ pyrochlore catalyst at 650 C and 2 psig. GSHV = 65,333 cm 3 /gcat/h. Individual responses (top graph) and cumulative responses (both labelled and unlabelled species - bottom graph). 9

10 As seen in Figure S3, after the isotopic switch, the CH4+ CH4, 2+ 2, H2 and + signals do not fully recover their intensity, despite reaching very close to the original value. We do not attribute this drop in all cumulative signals (both labelled and unlabelled species) to a kinetic isotope effect, but rather to a change inherent to reading of the mass spectrometer. 5E 09 4E 09 3E 09 2E 09 1E C H 2 + CH 4 + CH Ar Figure S4 Cumulative responses (both labelled and unlabelled species) of DRM products following a reactant switch from 2 to 2 using the LRhZ pyrochlore catalyst at 800 C and 2 psig. GSHV = 65,333 cm 3 /gcat/h. Bottom graph is a zoomed-in image of the top graph. 10

11 2E C E 09 Ar CH 4 + CH 4 2E 10 1E C H 2 + Ar Figure S5 Cumulative responses (both labelled and unlabelled species) of DRM products following a reactant switch from 2 to 2 using the LRhZ pyrochlore catalyst at 500 C and 2 psig. GSHV = 65,333 cm 3 /gcat/h. Bottom graph is a zoomed-in image of the top graph. 11

12 I 2E 09 1E C 2 CH 4 Ar CH4 4 H2 H 2 Ar Ar 2 2 II 2E 09 1E C 2 CH 4 Ar 2 2E 10 1E C H CH4 4 H2H 2 Ar Ar 2 2 2E 10 1E C H Figure S6 MS Signals for labelled and unlabelled species involved in an isotopic switch from 2 to 2 at steady state conditions for DRM using the LRhZ pyrochlore catalyst at 2 psig, GSHV = 65,333 cm 3 /gcat/h ; and temperature: I) 450 C and II) 500 C. Bottom graphs are zoomed-in images of the top graphs. References 1. Pakhare, D.; Schwartz, V.; Abdelsayed, V.; Haynes, D.; Shekhawat, D.; Poston, J.; Spivey, J. J. Catal. 2014, 316, Pechini, M. P. US Patent US Patent No. 3,330,697, Pakhare, D.; Wu, H.; Narendra, S.; Abdelsayed, V.; Haynes, D.; Shekhawat, D.; Berry, D.; Spivey, J. Appl. Petrochem. Res. 20, 3, Goodwin Jr., J. G.; Hammache, S.; Shannon, S. L.; Kim, S. Y. Taylor and Francis: 2006; pp Verykios, X. E. Appl. Catal. A: Gen. 2003, 255, Polo-Garzon, F.; He, M.; Bruce, D. A. J. Catal. 2016, 333,