Bernard L. Meyers

A multitechnique characterization of dealuminated mordenites

Mordenite zeolites were altered both physically and catalytically by dry thermal treatments, acid leaching at mild temperatures, sequential thermal/acid treatments, and hydrothermal steaming. Good agreement was demonstrated between four diverse analytical procedures, X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), 29Si and 27Al nuclear magnetic resonance (NMR), and NH3 temperature-programmed desorption (TPD) and the extent of dealumination. The complementary nature of 29Si and 27Al NMR in characterizing the Al content over broad concentrations was demonstrated. Both XRD and NMR showed that dry heat treatments alone expelled as much as 70% of the aluminum from the framework of the fresh mordenite, but XPS showed no change in the surface aluminum concentration. Strong zeolite acidity as measured by TPD of NH3 above 290 °C was nearly completely destroyed or made inaccessible by dry thermal treatment. This acidity was restored by acid leaching which removed the detrital aluminum, making the remaining tetrahedral aluminum sites accessible to NH3. Acidity decreases faster than residual structural aluminum content, possibly because a progressively larger fraction of the structural aluminum is in partially inaccessible four-member rings. Unlike the results obtained with NH3, fresh mordenites had very little strong acidity when measured by a larger molecule such as t-butylamine. But dealumination by HCl increased the accessibility of such acid sites to these larger molecules. The thermally dealuminated mordenite was hydrothermally dealuminated and analyzed by 27Al NMR and XPS. Roughly 40% of the remaining tetrahedral aluminum is lost from the structure upon steaming at 500 °C for 70 h. Here, XPS detects modest increases in surface aluminum concentration which were not observed for the heat-treated samples. This shows that steam is required for expelled aluminum to migrate to the zeolite surface. Acid dealumination of these mordenites was shown by N2 adsorption to develop maximum micropore volume and to minimize external zeolite surface area. The changes observed do not fully explain the aging characteristics during methanol conversion exhibited by deeply dealuminated mordenites; however, changes in catalyst selectivity toward higher-molecular-weight products can be explained by increases in the micropore volume following acid dealumination.

Metal-support interactions in precipitated, magnesium-promoted cobaltsilica catalysts

Abstract To define cobalt-silica interactions, precipitated, Mg-promoted catalysts were prepared by the cobalt nitrate-sodium carbonate reaction, without support and in the presence of silica supports. The major component of the unsupported Co catalyst was tentatively identified as Co2CO4 on the basis of elemental analyses and the IR carbonate absorptions at 1500 and ≈ 1400 cm−1. Upon heating in air at 360°C, the unsupported catalyst converted to Co3O4 according to IR and XRD analyses. TPR studies indicated easy reduction at 300-400°C, with one peak for the Co2CO4 and the Co(OH)2 components. A magnesium-promoted unsupported catalyst showed similar chemical properties and nearly identical IR spectrum; however, significantly different reducibility characteristics were observed by TPR. The composition of the silica-supported catalysts varied between two extremes. In the case of unreactive silica, IR and EM analyses showed a blend of the unsupported catalyst with the silica. In the other extreme, with reactive silicas, all the cobalt converted to cobalt silicates, as indicated by an Si O stretching vibration at 1034 cm−1 and by the absence of carbonate absorptions. TPR traces of the blend-type catalysts were similar to their unsupported analogues. The cobalt silicates required temperatures > 700°C for reduction according to TPR. The TPR of the intermediate-type catalysts that were, in part, blends of the unsupported catalyst with the silica and, in part, cobalt silicates showed cobalt silicate reductions starting above 400°C. This may be due to the catalytic effect of the cobalt that was reduced in the 300-400°C range. Electron microscopy, in combination with energy-dispersive X-ray analyses, further confirmed the cobalt-silica reaction. Cobalt silicates appeared as a “growth,” often with a filamentous structure, leaving other areas of the silica completely intact. The formation of cobalt silicates by a solid-solute reaction is discussed.

Flocculation mechanism: charge neutralization and bridging.

Electrophoresis measurements and electron-microscope observations with two model colloids and a polymeric flocculant show zeta-potential changes and details of floc structure. Fibers of the flocculant extend radially from the particle surface and vary in thickness from 20 to 300 angstroms. Both charge neutralization and bridging may function simultaneously.

Multitechnique analysis of a deactivated resid demetallation catalyst

Abstract A spent resid demetallation catalyst has been characterized using kinetic analysis, N 2 adsorption, X-ray photoelectron spectroscopy (XPS or ESCA), elemental analysis, X-ray diffraction (XRD), electron microprobe analysis, and CO chemisorption. During the demetallation process Ni, V, Fe, S, and C deposited from the resid feed onto the catalyst surface leading to its deactivation. The catalyst accepted about 100% of the original catalyst weight as metals and coke. Surface areas dropped by 30–65% and pore volumes were reduced by 60–85%. The deposition of metals, sulfur, and coke takes place preferentially at the entrances of pores, causing pore mouth plugging. This effect can be semiquantitatively described by the constraint index which is obtained by division of the N 2 adsorption/desorption hysteresis loop area by pore volume. All elements exhibit preferential deposition on the catalyst extrudate surfaces as evidenced by XPS and by electron microprobe extrudate cross-section analyses. Vanadium and nickel sulfides are formed on the working catalyst surface. The sulfides have been identified as V 3 S 4 and Ni 3 S 2 or Ni 2 S. Upon air exposure these sulfides oxidize slowly on the surface. Nickel and vanadium deposition through the reactor bed parallel each other with a maximum deposition at about 17% of bed depth, whereas the carbon profile is relatively uniform. The top section of the catalyst bed, containing 1.5 times the metals but 0.9 times the carbon compared to the bottom section, was 30–50% more active for both V and Ni removal.

Hydrothermal dealumination of faujasites

Abstract Two faujasites (USY and REY) were hydrothermally treated at temperatures between 400 and 850 °C in 100% steam and up to 65 h to induce dealumination. The Al expulsion from the zeolite lattice was followed by X-ray diffraction (XRD) and by 29 Si and 27 Al nuclear magnetic resonance (NMR) spectroscopy. Both techniques showed large changes in crystal structure in less than 30 min while crystallinity lossses were kept below 15% at temperatures up to 500 °C. Excellent agreement in the rate of dealumination was found between XRD and 29 Si NMR measurements. The Al expulsion from the zeolite lattice is accompanied by Al diffusion through the channel network to the zeolite particle surface. The surface enrichment by Al was quantitatively measured by X-ray photoelectron spectroscopy (XPS or ESCA). The dealumination and Al migration processes are both logarithmic with time and a linear correlation between them was found at all temperatures for up to 65 h of steaming. The diffusion of Al to the particle surface takes place only in the presence of steam and the migrating species is speculated to be a hydroxylated Al ion. These studies were complemented by secondary ion mass spectrometry (SIMS) and scanning transmission electron microscopy (STEM). SIMS analysis demonstrated a very large enrichment of Al in the top 30 A of steamed zeolites. STEM measurements of Si Al profiles on 800 A-thick microtomed sections confirmed the SIMS data, but indicated broader, less steep changes in the Si Al ratio across zeolite particles. Because dealumination causes both structural changes (shrinkage of unit cell) and chemical changes (decrease in acidity) in a zeolite, it has profound implications on catalyst performance and on catalyst deactivation. For instance, dealuminated faujasites exhibit drastically reduced coking rates in the cumene cracking test reaction which will be discussed in some detail.

Regeneration of a PVOZn butane oxidation catalyst using chlorine containing hydrocarbons

Abstract A fresh V P O Zn (vanadyl phosphate) catalyst has a maximum molar yield of maleic anhydride of 52% (at 450°C) which eventually becomes a spent catalyst having a maximum yield of 44% (at 400°C). This catalyst can be regenerated by passing carbon tetrachloride (in an air or nitrogen carrier gas) over its surface at 300–450°C. The products in the effluent of the reactor are consistent with the decomposition of CCl4, (CO, CO2, VCl4, unknown chlorine compounds) and with part of the catalyst surface being “stripped” off (VCl4, O2, unknown phosphorus containing compounds). Twenty-four hours after the CCl4 regeneration, the catalytic properties of the regenerated and fresh catalysts are identical i.e. the conversion, selectivities, yields (as a function of temperature) as well as other physical properties - surface areas, XRD patterns, and bulk P/V ratios. The experimental data is consistent with the CCl4 selectively reacting with the poisoned sites and removing the vanadium from them by forming VCl4. A regenerated catalyst gives a XRD pattern consisting of a mixture of the catalyst precursor and the fresh catalyst indicating a bulk rearrangement (presumably topotactic). The regeneration procedure can be practiced a large number of times without a detrimental effect on the catalytic performance.

Activity decline and diffusional changes in a magnesium-promoted cobalt-silica catalyst

A Co-Mg/diatomaceous earth catalyst was prepared by slow precipitation. Transmission electron microscopy (TEM) indicated two separate phases: small needles of a Co-Mg phase and the large particles of the diatom fragments. Nitrogen sorption measurements indicated much higherpore volume (0.32 cm3g) for the catalyst than for the control catalyst made by fast precipitation (0.10 cm3g). The catalyst activity in the Fischer-Tropsch synthesis at 155 kPa declined at an unusually fast rate. The decline was accompanied by an increase of the methane selectivity and a decrease of the Schulz-Flory growth factor. Hydrogen treatments temporarily restored the catalyst activity but did not arrest the declines. Increasing the H2:CO feed ratio for the deactivated catalyst not only restored the high activity but also stabilized the catalyst. An increase of the methane selectivity and a small decrease of the growth factor also accompanied these changes. The changes in catalyst activity and product selectivity correlate well with the anticipated diffusional changes from gas phase diffusion to diffusion through liquids which have been deposited in the pores during the course of the reaction. It is inferred that the diffusional changes cause significant changes in the steady-state concentrations of the reaction intermediates and these, in turn, affect the reaction rate (conversion) and the chain termination (growth factor).

Cobalt Reducibility and Magnesium Promotion in Silica-Supported Fischer-Tropsch Catalysts

Publisher Summary It is known from earlier work that the addition of structural promoters—such as MgO and ThO2—made the reduction of the cobalt more difficult. Anderson has suggested that CoO and MgO may form a solid solution and the reduction to Co metal may be inhibited by a kinetic or thermodynamic effect. In a more recent study, Sexton et al. demonstrated by X-ray photoelectron spectroscopy (XPS) and by temperature-programmed reduction (TPR) techniques that MgO and ThO2 can form new phases with the CoO and are more resistant to reduction than CoO alone. Recent TPR studies have also shown that the coprecipitation of small amounts of magnesium with the cobalt has resulted in higher temperature requirements to achieve the complete reduction of the cobalt. It was also found that during the preparation of magnesium-promoted catalysts using the pH-controlled precipitation technique, Co silicates can form, which are not easily reduced to metallic cobalt.

Electron energy loss studies of propylene interactions in US-Y faujasites

There has been considerable interest in the interactions of hydrocarbons in the Y faujasites and the nature of the formation of coke. From the literature there is a question as to which site (Al or Si) is active in the formation of the residual carbonaceous material left in the molecular sieve after reaction with a hydrocarbon. In order to address this question, the authors have undertaken a study of the interaction of a simple probe molecule (propylene) with two series of ultrastable Y faujasite as a function of reaction temperature from 40-500C. The techniques involved in the study are core electron energy loss spectroscopy (CEELS), X-ray photoelectron spectroscopy (XPS), and {sup 13}C nuclear magnetic resonance (NMR) using cross-polarization magic-angle spinning (CPMAS). In this note they will focus on the CEELS results and the new insight that it provides into this interaction in light of the information from the NMR and XPS data for this system.