Laboratory fixed-bed reactor

Compared with laboratory fixed-bed reactors or conventional extruded monoliths, such a microstructured monolith is smaller in characteristic dimensions, lower in pressure loss by optimized fluid guiding and constructed from the catalytic material solely [3]. The latter aspect also leads to enhanced heat distribution within the micro channels, giving more uniform temperature profiles. 

In laboratory fixed bed reactors, Reynolds numbers based on the pellet diameter can be as low as one, leading to a relatively thick stagnant fluid film around the pellet. The following correlations can be used to calculate the transfer coefficients [8] ... 

Figure 7.19 represents schematically a way to determine experimentally whether external mass transfer can be neglected. Transfer effects do not occur when the conversion for a given space-time does not depend on the flow rate. The test is not very sensitive, however. This is caused by the small dependence of the mass transfer coefficient on the flow rate at the low Reynolds numbers prevailing in laboratory fixed bed reactors. 

Various correlations exist for mass and heat transfer coefficients in terms of dimensionless numbers. Table 8.7 surveys the most appropriate ones for laboratory fixed-bed reactors [8,17-19]. The Sherwood number, Sh, and the Nusselt number, Nu, express the ratio of total mass transfer to diffusive mass transfer, and the ratio of total heat transfer to conductive heat transfer, respectively. Values of kf and h for gases in laboratory systems range from 0.1 to 10 mf " s ... 

Herten and Froment (1968) studied the reaction on a doped vanadium pentoxide catalyst in a quasi-isothermal laboratory fixed bed reactor. Kinetic measurements were made in the temperature range 325-402 C for a wide range of feed compositions and varying residence times. The reaction scheme proposed is the same as that shown in Figure 3.13 with the exception that no maleic anhydride was isolated in the reaction products. They found no evidence of significant oxidation of phthalic anhydride to carbon oxides. The conversion of o-tolualdehyde to phthaiide is considered to be a relatively unimportant step in their model. 

S Smeds, T Salmi, D Murzin, Gas-phase hydrogenation of ethyl benzene over Ni. Comparison of dilferent laboratory fixed bed reactors, Appl. Catal A.General, 201 55 - 59,2000. 

If we want to determine the true intrinsic rate of a heterogeneously catalyzed reaction in a laboratory fixed bed reactor, we have to consider two conditions to obtain reliable kinetic data (Figure 4.11.12) ... 

Figure 2.2.1 shows the simplified sketch of the reactor used for the microactivity test. As can be seen, a fluid-bed catalyst is tested in a fixed bed reactor in the laboratory to predict its performance in a commercial fluid bed reactor. This can be done only because enormous empirical experience exists that has accumulated throughout several decades in several hundreds of reactors both in production and in laboratories. The standard states ... 

As mentioned in Section 2.2 (Fixed-Bed Reactors) and in the Micro activity test example, even fluid-bed catalysts are tested in fixed-bed reactors when working on a small scale. The reason is that the experimental conditions in laboratory fluidized-bed reactors can not even approach that in production units. Even catalyst particle size must be much smaller to get proper fluidization. The reactors of ARCO (Wachtel, et al, 1972) and that of Kraemer and deLasa (1988) are such attempts. 

Laboratory reactor for studying three-phase processes can be divided in reactors with mobile and immobile catalyst particles. Bubble (suspension) column reactors, mechanically stirred tank reactors, ebullated-bed reactors and gas-lift reactors belong the class of reactors with mobile catalyst particles. Fixed-bed reactors with cocurrent (trickle-bed reactor and bubble columns, see Figs. 5.4-7 and 5.4-8 in Section 5.4.1) or countercurrent (packed column, see Fig. 5.4-8) flow of phases are reactors with immobile catalyst particles. A mobile catalyst is usually of the form of finely powdered particles, while coarser catalysts are studied when placing them in a fixed place (possibly moving as in mechanically agitated basket-type reactors). 

Laboratory scale PQC evaluation studies are usually conducted in fixed bed reactors such as MAT, the results from which can provide a reliable and rapid means of ranking catalyst performance [4]. Depending upon the conditions employed, the effect of added ZSM-5 can also be predicted [5] and can give the same trends as those experienced in commercial reactors. For example, the effect of 2.5 wt% addition of ZSM-5 on gas oil cracking yields with Quantum 2000 is described in Table 1. In this example, a 4% reduction in gasoline yield occurs, predominantly from 105°C+ material. The L.P.G. composition indicates an enhancement of propene, butenes, and iso-butane, in agreement with commercial results and, furthermore, the relative increase in the individual butenes are similar to those reported by Schipper et al [1]. 

In addition, significant advances have been made in both basic and applied research which allow a smart and efficient solution to most of these problems. As an example, let us quote the development of the synthesis of novel catalytic materials with tailor-made and more suitable characteristics (stable nanocrystals, controlled hydrophobicity, better thermal and/or mechanical stability, etc.), the understanding of the complex phenomena involved in the catalytic transformation of polar molecules within zeolite micropores or the demonstration that fixed bed reactors, which have many advantages over conventional batch reactors, can be easily used, even for liquid-phase reactions and even for laboratory scale experiments. 

The effect of feed composition cycling on the time-average rate and temperature profile was explored in the region of integral conversion in a laboratory fixed bed ammonia synthesis reactor. Experiments were carried out at 400°C and 2.38 MPa over 40/50 US mesh catalyst particles. The effect of various cycling parameters, such as cycle-period, cycle-split, and the mean composition, on the improvement in time-average rate over the steady state were investigated. 

The results Illustrated by Figures 3 and 4 resemble those obtained in the Berty recycle reactor under similar conditions. The space-mean, time average rates for the fixed-bed reactor were only about 50% of those measured in the Berty reactor, because, of course the former reactor achieved conversions high enough for the back reaction to become important. The significance of these observations is that 1) CSTR and differential reactors, widely used for laboratory studies, seem to reflect performance improvements obtainable with fixed-bed, integral reactor which resemble commercial units, and 2) improvement from periodic operation are still observed even tfien reverse reactions become important. 

When appropriate material systems are not available for model experiments, accurate simulation of the working conditions of an industrial plant on a laboratory- or bench-scale may not be possible. Under such conditions, experiments on differently sized equipment are customarily performed before extrapolation of the results to the full-scale operation. Sometimes this expensive and basically unreliable procedure can be replaced by a well-planned experimental strategy. Namely, the process in question can be either divided up into parts which are then investigated separately (Example 9 Drag resistance of a ship s hull after Froude) or certain similarity criteria can be deliberately abandoned and then their effect on the entire process checked (Example 41/2 Simultaneous mass and heat transfer in a catalytic fixed bed reactor after Damkohler). 

According to Table II, much higher L/dp ratios are required in laboratory experiments than are expected from the established rule-of-thumb that L/dp > 50 is usually acceptable. For fixed-bed reactors, L/dp > 50 is indeed a sufficient requirement, provided a particle Reynolds number (NRe) is 10 or above. In laboratory experiments the particle Reynolds numbers, however, are usually much smaller, and Re < 0.1 is more the rule than the exception. The literature does not stress this point sufficiendy, which has led to confusion and many misconceptions. 

Recent investigations have shown that increasing the diameter of a fixedwaxy deposits. Plugging is also found at higlt temperature or low Hj/CO operation and is related with local nonuniformity in the catalyst temperatures [22. Kinetic measurements on laboratory-scale recycle reactors have been used to predict product distributions for pilot plant reactors and were in good agreement with experimental results 1231. 

In principle, the safest way to represent an industrial reactor on a laboratory scale is to reduce the diameter while keeping the bed length the same. In a well-designed industrial fixed-bed reactor where proper care is taken to ensure uniform distribution of feed over the cross section of the bed, there are theoretically no cross sectional differences. Hence, a more slender but equally tall test reactor would be a good representation of the commercial reactor, provided that the diameter of the test reactor is not so small that wall effects become appreciable (to be discussed later). 

A commercial fixed-bed reactor is generally operated as an adiabatic reactor since at this scale heat losses to the surroundings are generally insignificant and removal or supply of heat in the bed requires special arrangements. In laboratory reactors operated above ambient temperatures natural heat losses are quite appreciable and heaters are required to maintain temperatures at desired levels. 

Realistic analysis of fixed-bed reactor experiments requires calculation of interfacial states. Laboratory reactors are typically much shorter than full-scale units and operate with smaller axial velocities, producing significant departures of the iiiterfacial states from the measurable values in the mainstream fluid and consequent difficulties in establishing catalytic reaction models. Interfacial temperatures and partial pressures were calculated with jn and jo- and used in estimating reaction model parameters, in a landmark paper by Yoshida. Ramaswami. and Hougen (1962). Here we give an updated analysis of interfacial states in fixed-bed reactor operations for improved treatment of catalytic reaction data. 

---