Coke Levels Research Articles

Evaluation of Modes of Catalyst Deactivation by Coking for Cumene Cracking over Zeolites

The deposition of coke deactivates a catalyst by the suppression of active sites and/or by choking of pores. The so-called Constant-Coke Arrhenius Plot technique can assess the magnitudes of these two modes of deactivation using reaction experiments under process conditions. For cumene cracking over REY zeolite, a TGA has been used as a continuous-flow microreactor to monitor coke levels, with activity measured by an on-line GC. Using modifications described previously for small-pore catalysts, we show that both temperature and coke level influence the relative importance of site suppression and pore choking in the deactivation of the zeolite. At relatively high temperatures, site suppression is predominant at all coke levels studied here. At lower temperatures, pore choking is more important at low levels of coke, with site suppression increasing in importance at the higher coke levels. The results are consistent with a physical model developed.

Examination of de-coking of promoted (Co, Ni) Mo/γ-Al 2O 3 catalysts by X-ray photoelectron spectroscopy

Promoted (Co, Ni)-Mo/γ-Al 2O 3 hydrodesulphurization catalysts have been decoked using hydrogen gasification. The catalysts were exposed to a benzene, toluene and xylene (BTX) feed for one week until a significant level of coke had been built up on the surface. The catalysts were then decoked by gasifi- cation at 400°C in hydrogen. The decoking was monitored using the X-ray photoelectron spectroscopy C/Al ratio. After 6 h gasification there was a tenfold decrease in the C/A1 ratio indicating a significant amount of decoking. Furthermore, the initial activity was regained as determined by the conversion of a standard BTX feed.

Upgrading of an Australian coal-derived liquid. 2. Characterisation of Ni-Mo/Al 2O 3 catalysts

A suite of ten Ni-Mo/Al 2O 3 catalysts of varying active species loadings, previously tested for their ability to upgrade an Australian coal-derived liquid, were characterised by a range of techniques. This included surface area and porosity measurements, XRD, TPR and acidity analysis. The coke deposition during upgrading was also measured by TGA. It was found that those catalysts of highest initial activity with regards to phenol removal and aromatic hydrogenation (1.9 wt% Ni/7.9 wt% Mo to 3.7 wt% Ni/15.1 wt% Mo) had high surface areas (> 170 m 2/ g) and pore radii of near 50 Å, were amorphous in character, possessed both metal-support and metal-metal interactions and had intermediate acid site densities. These catalysts were also found to undergo more rapid coke deposition, with coke levels of up to 11% by weight of the spent catalyst being found, resulting in a loss of surface area and porosity. It has been previously argued that the catalysts of highest initial activity possess an optimum active phase dispersion on the alumina support. Evidence from a number of techniques was found to indicate that the catalysts of low loadings (below 3.7 wt% Ni/15.1 wt% Mo) had well dispersed active phase.

Piperidine hydrogenolysis on a commercial hydrocracking catalyst: III. The effects of zeolite unit cell size, catalyst sulfur content, and coke deposition on catalyst activity and deactivation

A group of fresh, deactivated, and deactivated-regenerated commercial hydrocracking catalysts was characterized using piperidine hydrogenolysis as a probe reaction at 301C, hydrogen partial pressures of about l6 atm, and initial concentrations of piperidine ranging from 4.14 × 10 −3 11.86 × 10 −3 g mol/liter. The first part of the study examined the effects of the zeolite unit cell size on the piperidine hydrogenolysis activity and catalyst deactivation and revealed that over the range of unit cell size examined (24.35-24.56 Å), adsorption equilibrium constants associated with the acidic and metallic catalyst functions decreased while reaction rate constants increased with increasing unit cell size. The increase of acidic function activity with unit cell size is explained in terms of the total number of potential acid sites present per unit cell. The increase of metallic function activity with unit cell size is suggested to reflect mostly changes of the metallic function upon regeneration. Different types of sites are proposed to exist on the metallic catalyst function in order to account for the results. An inverse relationship between catalyst deactivation and unit cell size was generally observed for both catalyst functions. These results point to the zeolite unit cell size as a possible parameter for correlating catalytic activity as well as selectivity, especially for the acidic catalyst function. The second part of the study examined the effects of wt% sulfur and wt% carbon at low levels of sulfur and coke on the piperidine hydrogenolysis activity and showed that the regenerated catalysts with constant zeolite unit cell size (≈24.40 Å and low levels of coke (<0.8 wt%) had similar acidic function activities. The acivity correlation of the metallic function on these regenerated catalysts with low levels of sulfur (<0.4 wt%) also indicated the presence of two different types of metal sites. One showed a decrease in activity with increasing sulfur content and a possible dependency on the regeneration procedure while the activity of the other site was invariant. The activity of the metallic and acidic functions of the catalyst was also examined at high levels of sulfur and coke and constant unit cell size. For this study a series of deactivated samples was used. The metallic function activity of these catalysts was found to decrease with the commercial time on stream. Between 2.2 and 2.6 wt% sulfur, the activity of the metallic function decreased linearly with the sulfur content on the catalyst. At higher levels of sulfur (>2.5 wt%) the deactivated catalysts had similar metallic function activities. Between 2 and 6 wt% carbon, the activity of the acidic function decreased linearly with the carbon content on the catalyst. At low levels of coke (< 1.5 wt%), the effect of the unit cell size on acidic activity prevailed. The wt% carbon on the deactivated as well as the regenerated catalysts was found to have no effect on the catalyst deactivation rates observed during the reactions.

Activity, selectivity, and deactivation of high-sodium HY zeolite

The effect of deactivation on the activity and selectivity of (very high) Na + content HY zeolite catalysts was studied in a pulse microreactor equipped with a thermogravimetric analysis unit (TGA). Cumene was used as the reactant. The Na + level studied was from 3.2 to 11.8% of Na 2O. As the Na + content decreases, the acidity of the zeolite increases and deactivation decreases. The effect of coke level on selectivity to benzene, toluene, ethylbenzene, dimethylbenzene, ethyltoluene, cymene, and diisopropylbenzene depends on the Na + content. For example, the selectivity to benzene, the major aromatic formed by the fresh catalyst, decreases with increasing coke level and Na + level, while the selectivity to ethyltoluene increases under the same conditions. A maximum in selectivity with coke level is obtained for most other products. At intermediate Na + levels (around 7% Na 2O), ethylbenzene shows a maximum in selectivity while cymene passes through a minimum. The maximum coke deposition and the coke level for the maximum incremental coke deposition are two important parameters that distinguish conversion and selectivity patterns.

Xe NMR and Ar quasi-equilibrium sorption of coked H-Y zeolite

Acidic, HY zeolite was partially coked with propylene at 250 and 475°C. Standard micropore volume determination by N 2 physical sorption indicates that nearly all of the coke was confined to the micropores with little change in the larger mesopores. Xe NMR chemical shifts increased linearly with a decrease in the micropore volume for the coke samples. However, initial coke deposition on HY increased the Xe chemical shift beyond the changes predicted by the decrease in micropore volume. The increase in chemical shift due to coke was also found to be dependent on the aromaticity of the coke. Coke formation at 250°C (65% spz carbon) increased the chemical shift by 6.5 ppm while coke formed at 475°C (>95% spz carbon) increased the chemical shift by 8.5 ppm. The size of the pore apertures, i.e., the opening to the alpha cage, was determined by low-pressure Ar adsorption isotherms. As the level of coke increased, the size of the alpha cage aperture decreased by as much as 0.07 nm. A bimodal distribution in the pore apertures was clearly resolved on some catalysts at higher coke levels. One pore size was characteristic of apertures essentially free of coke (0.72 nm); while, the other, smaller aperture was more severely restricted due to coke deposits (0.58 nm). The inability of Xe NMR to detect two pore environments observed by Ar sorption may result from rapid exchange of Xe between the two sites. This implies that the distance between the two pore environments is small and that the two pore environments are uniformly distributed throughout the catalyst particle.

Determination of C/H ratios in coked catalysts using a portable neutron source

In a previous paper, it was shown that a portable neutron source may be employed to measure the coke content of a coked catalyst. In this study, the authors show how the portable neutron source may also be employed to determine the C/H ratio, i.e., n in CH{sub n}, in the deposited coke. The present study indicates how values of n may be determined from neutron transmission measurements provided that the coke level in the reactor is known. Experimental verification has been obtained for the catalytic dehydrogenation of butene-1 over a chromia-alumina catalyst, but the method should be applicable to a wide range of catalytic reactions where coke deposition is an important factor.

FCC Regenerator flow model

Catalyst flow patterns in FCC cross-flow and swirl-type dense-bed regenerators affect regenerator efficiency. A two-dimensional, pseudo-homogeneous flow model describing such flow patterns is developed. The predicted catalyst flow patterns for swirl regenerators indicate catalyst bypassing and recirculating zones. More severe bypassing is predicted for cross-flow regenerators, the severity of bypassing increases with increasing Reynolds number. The calculated flow patterns are combined with a radial mixing-cell model to account for radial mixing in fluidized beds. Radial mixing rates in commercial size regenerators counteract somewhat the effect of bypassing. Model calculations coupling coke burning kinetics with the flow model, however, show that bypassing leads to non-uniform coke levels and oxygen breakthrough. This modeling technology is used to optimize FCC regenerator operation.

Measurement of coke levels in catalysts using a portable neutron source

In a previous study, Byrne et al. measured quantitative coke concentration profiles in fixed-bed catalytic reactors in situ by the attenuation of a neutron beam aligned perpendicular to the axis of the reactor. The experiments were performed using a collimated neutron beam from the DIDO reactor at the Harwell Laboratory (UK). In this communication, the authors present some initial results on the in situ measurement of coke profiles using a portable neutron source. The advantages of such a source include its transportability and its availability to the nonnuclear industries. However, a possible problem in using a portable source is the low flux of neutrons compared to that of a nuclear reactor. In order to test the applicability of such a source over a range of coke levels on different catalysts, three reactions were studied. The first of these was the isomerization of xylenes over a silica-alumina catalyst. The second reaction studied was the dehydrogenation of butene to butadiene over a chromia-alumina catalyst. The third reaction considered was the cracking of cumene over a zeolite catalyst. While the silica-alumina catalyst used was a powder, the last two catalysts were in the form of the pellets. The reactions studied also represent differentmore » coking processes, i.e., parallel to the main reaction in the case of xylene isomerization reaction, consecutive to the main reaction in the case of butene dehydrogenation, and a complex series-parallel coking process in the case of the cumene cracking reaction.« less

Hydrotreating catalyst deactivation by coke from SRC-II oil

Samples of a CoMo/Al 2O 3 catalyst were partially deactivated with SRC-II feed in an autoclave reactor to give coked samples of 5 to 18 %C. The coked catalysts were analyzed for surface area, pore volume, coronene adsorption and diffusivity, and their catalytic activity determined for hydrodesulfurization (HDS), hydrodeoxygenation (HDO) and C-N hydrogenolysis (CNH) using model compounds. All of the above measurements decreased with increase in coke content. Property data indicate that some pores are blocked by coke and diffusivity results show narrowing of pore mouths with increasing coke content. Catalyst deactivation versus coke level was identical for HDS and HDO, but less for CNH. A simple model of coke deactivation was developed to relate activity to coke content. Coke is envisioned as forming wedge-like deposits in the catalyst pores.

Spectroscopic investigations on deactivation of zeolite catalysts during reactions of olefins

Coke formation over acidic zeolite catalysts upon reaction of olefins is investigated using in-situ IR spectroscopy, i.e. an IR cell which simultaneously serves as a fixed-bed flow reactor and is attached to a gas chromatograph. Zeolites HY, dealuminated H-MOR with two different Si/Al ratios and H-ZSM-5 are compared. The effects of the number of active sites and the pore volume of the zeolites on the initial rate of coke formation and the final coke level are studied. The final level of coke deposition reflects to some extent the pore volume of the zeolite catalyst. In all cases a considerable amount of coke formed remained inside the pore structure but another fraction was deposited onto the external surface. Both findings were confirmed by sorption measurements on samples before and after loading with coke. Possible mechanisms of deactivation through coke formation, viz, poisoning, consumption or blockage of sides are briefly considered. In-situ IR spectroscopy revealed that at most a small fraction of the active sites of HY, HMOR or HZSM-5 disappeared during coke deposition suggesting that deactivation is caused by site blocking rather than by poisoning or consumption of the acidic centres. Finally, in-situ IR spectroscopy turned out to be a useful tool for discriminating between low-temperature coke (formed below about 500 K) and high-temperature coke (deposited above 500 K).

Liquefaction of partially dried and oxidized coals: 3. Liquefaction results

Samples of partially dried and oxidized Belle Ayr subbituminous coal were liquefied in a recycle donor solvent (SRC-ll heavy distillate) to observe the effect of coal pretreatment on conversion. Because subbituminous coals have moisture contents typically > 25%, it would appear useful to dry these coals prior to liquefaction; however, the drying of Belle Ayr coal, either in nitrogen or oxygen-containing gases, resulted in a significant decrease in yields of liquefied coal products. The liquefaction residues recovered from these runs were examined by optical microscopy and were found to contain high levels of coke. This coke appeared to have formed by polymerization of coal-derived liquid products.

Constant-coke arrhenius plots: A diagnostic tool

The presence in coke of highly unsaturated hydrocarbons produced by unneeded side reactions and strongly adsorbed on the catalyst, is known to decrease the activity of the catalyst. If the coke molecule is adsorbed on a site hitherto active in the desired reaction, the catalyst loses activity because the site has been taken out of circulation. The presence of coke molecules adsorbed on pore walls increases the resistance to mass transfer inside the catalyst pellet, decreases the concentration of reactant in the fluid phase above an active site, and thereby decreases the rate of reaction. It is important to know which phenomenon is the predominant one in causing deactivation by coking. If both are important, it is important to know the time on stream and/or coke level at which the transition may be considered to occur. How modification of a technique recently reported can be used to evaluate separately (pseudo-) first-order intrinsic rate constants and effective diffusivities of the reactant under actual reaction conditions is shown.

The influence of nickel on the coke forming tendency of a zeolite-containing fluid cracking catalyst

A kinetic investigation of coke deposition on nickel-contaminated zeolite-based fluid cracking catalysts has been carried out in a thermo-balance. The rate of coke deposition increases with nickel content, and rises rapidly above a nickel loading of 1000 μg/g. The ultimate coke level and the apparent reaction order appeared independent of the nickel content. The results of these model tests were compared with those from a commercial micro-reactor test, and demonstrated a good qualitative similarity.

Pitch and petroleum coke additions to coke oven charges

AbstractSeveral pitch materials and a petroleum coke were added to coke oven charges in an attempt to make good metallurgical coke from Canadian coal of poor coking quality. Coal and petroleum pitches were added to a low fluid western Canadian coal of medium volatile bituminous rank, and the blends coked in a technical‐scale moveable wall test oven having a 230‐kg charge capacity. Pitches improved coke tumble test indices, the principal coke quality parameter related to blast furnace performance. Varying levels of petroleum coke were added to an eastern Canadian coal of high volatile bituminous rank, and the blends, some partially briquetted, were carbonized in a test oven. Tumble indices of coke from the partially briquetted charges approached an acceptable level. These investigations confirm that petroleum products as well as coal derivative can play a useful part in the production of a metallurgical strength coke from poor or non‐coking coals.

Effects of coking on the transport properties of H-mordenite

An experimental study of the relationships between coke content, catalyst activity, and intraparticle diffusivity is reported for cumene cracking on a commercial H-mordenite. Coking conditions were varied from 230 to 350 °C at space velocities from 0.20 to 0.65 wt/hr/wt and time on stream up to 6 hr. Effective diffusivities of the catalysts were found to decrease by over a factor of 2 during the coking process and are nonlinearly related to the coke level, independent of the temperature of coking. Under conditions of strong diffusion limitation, this change in diffusivity alters the catalytic effectiveness in proportion to the square root of the diffusivity ratio, ( D eff D 0 eff ) 1 2 .