|dc.description.abstract||Hydrogen production by catalytic partial oxidation of natural gas was investigated using tools ranging from theoretical calculations to experimental work and sophisticated characterization techniques.
Catalytic partial oxidation (CPO) was carried out in a conventional continuous flow experimental apparatus using a xed-bed reactor, and operating at 1 atm and furnace temperatures in the range from ambient to 1073 K. The feed typically consisted of a mixture of methane and air, with a CH4/O2 ratio of 2, and the average bed residence time was in the range 10-250 ms. Steam methane reforming (SMR) was carried out in the same apparatus at similar temperatures and pressure in a feed consisting of methane, nitrogen and water, with a steam to carbon ratio of 2.0-4.0.
Temperature programmed (TP) techniques, including oxidation (TPO), reduction (TPR), reaction (TPCPO) and methane dissociation (TPMD) was used to characterize catalytic properties such as ignition temperatures, the catalyst reducibility and activation energies. Dispersions from catalyst surface area measurements were compared to X-ray diffraction (XRD) techniques and electron microscopy (SEM, TEM,STEM) to obtain information on catalyst particle sizes and dispersion. X-ray photoelectron spectroscopy (XPS) provided information on the speci c catalyst surface composition, which was compared to results on the bulk structure obtained by XRD.
The effect of modifying cobalt catalysts supported on alumina was investigated by adding small amounts of Ni, Fe, Cr, Re, Mn, W, Mo, V and Ta oxides. The idea behind this work was to investigate whether the cobalt crystals were decorated, covered or encircled by a modi er and to what extent this aected catalyst performance. The choice of modi ers in this study was based on the principle that in any chemical process it may be just as important to identify groups of elements that have negative effects as identifying the best promoters. It was found that the presence of impregnated metals or metal oxides that may either form bimetallic or mixed oxide species with cobalt had a negative eect on the catalyst performance. Results from temperature programmed catalytic partial oxidation (TPCPO) indicated that deactivation of cobalt involves oxidation of the active phase.
Nickel catalysts supported on alumina were modi ed with small amounts of reducible or partially reducible oxides from Co, Fe, Cr, Mn, W, Mo, Re, Cu, Pt, Rh and Pd. It was of particular interest to study the catalysts at conditions giving a partial conversion of oxygen, and it was quite remarkable that two fairly stable modes of operation were obtained at less than 100% oxygen conversion. The product composition at low gas velocities and complete oxygen conversion was typically in accordance with thermodynamic equilibrium evaluated at bed exit gas temperatures. Arrhenius type exponential factors, or apparent activation energies, were calculated at bed exit gas temperatures based on integral reactor data from TPCPO. At low tempratures energies were found to be 77 and 54 kJ/mol, respectively. At higher temperatures the apparent activation energies were halved, suggesting a reaction transitioning from being limited by reactivity to one being limited by diffusion, which was in agreement with the evaluation based on reactor properties.
Nickel and nickel-cobalt aluminate spinel catalysts were studied in the partial oxidation and steam reforming of methane. Methane conversion during CPO over unreduced NiAl2O4 did not exceed the empty reactor conversion, and the unreduced NiAl2O4 catalyst showed no activity in SMR at 973 K furnace temperature and 1 atm.
Close to equilibrium yields were obtained over reduced catalysts. The activation energies in SMR over reduced NiAl2O4 were in accordance with what was expected (100 kJ/mol). Replacing 25% of the nickel with cobalt lowered the catalyst dispersion at high loadings, but the activation energies in steam methane reforming remained unchanged. Based on the combined XRD, XPS and STEM results the nickel metal particles appeared to be 15-25 nm, possibly partially embedded in the aluminate, and were to some degree covered by surface layers of bimetallic or Al0 species.
The effect of noble metal promoters (0.005 wt% Rh, Ru, Pd or Pt) on the activation of methane and synthesis gas production by catalytic partial oxidation was investigated on 0.5 wt% Ni/γ-Al2O3. It was shown that temperature programmed activation of methane (10 K/min) initiated above 600 K, but was rather slow until the temperature reached 950-1000 K. Adding a noble metal caused a signi cant drop in the ignition temperature during TPCPO, and the ignition temperature could be correlated to the reducibility of the noble metal oxide, as estimated by the change in free energy when reducing lattice oxygen. Methane partial oxidation over 0.5 wt% Ni catalysts, both with and without promoter, yielded high selectivity to synthesis gas (> 93%) and stable performane for 20-70 h continued operation, but synthesis gas production at temperatures below 1073 K required a promoter when the catalyst was ignited by TPCPO.
Two different reactor materials (Fecralloy, Nicrofer) were investigated in the catalytic partial oxidation of methane and propane. In this study focus was put on the stability and applicability of Nicrofer as compared to Fecralloy. It was shown that high temperature calcination of Fecralloy established a stable alumina coating, which effectively increased the surface area of the reactor providing sites for the Rh particles. The reactor performance was stable with both methane and propane feeds. The Nicrofer was wash-coated with alumina using a sol-gel technique, and SEM images of the Nicrofer reactor revealed the formation of Cr-layers and Cr-oxide structures covering the impregnated Rh particles. The Cr rich structures coating the Rh particles were found to be detrimental to the reactor performance, and the alumina wash-coating was consequently not successful at stabilizing the Nicrofer reactor material.
Catalytic partial oxidation (CPO) of methane and its mechanistic aspects were evaluated using the unity bond index-quadratic exponential potential (UBI-QEP) formalism applied to 91 elementary surface reaction equilibria. MATLAB was used in this evaluation applied to single crystal surfaces of the most active metals for the catalytic conversion of methane by oxygen or steam, including fcc(111) surfaces of Ni, Rh, Pd, Pt and Ru, as well as Ni(100). In the principal UBI-QEP domain, the clean metal surface, the calculations were in accordance with previously reported mechanisms involving CH
3 dissociation as the rate determining step in methane activation by dissociation. Large barriers to the formation of H2(g) and CO(g) were found, in fact the formation of CO(g) could be rate limiting on some catalysts, and there were low barriers to H2O(g) and CO2(g) in the presence of CO, O, H and OH on the catalyst surface. Dierences between results predicted by UBI-QEP and DFT reports in the literature were in some cases found to be signi cant on a molecular level in terms of heats of adsorption and activation energy barriers. It was shown that the UBI-QEP method is extremely sensitive to uncertainties in atomic heats of adsorption.||nb_NO