Project topics and materials
Get Complete Project Material File(s) Now! »
LITERATURE
Apart from the catalytic process itself, several mass transfer steps can influence the rate and/or selectivity of a solid catalysed gas-liquid reaction as illustrated in Figure 2.1. In packed bed gas-liquid reactors, it is well known that flow configuration and characteristics have an important effect on these transfer processes. Especially for trickle-bed reactors operating in the trickle flow regime, these processes can have an intricate interaction.
Trickle flow is characterised by gravity-driven liquid flow over a packed bed with gas continuous flow. The liquid trickles down the packing, giving rise to (possibly) incomplete wetting, liquid velocity profiles and liquid maldistribution. This morphology of liquid flow directly influences the mass transfer steps depicted in Figure 2.1:
• The area available for liquid-solid mass transfer is directly affected by the fraction of external particle area contacted by the flowing liquid.
• The area available for gas-solid mass transfer is directly affected by the fraction of external particle area contacted by the flowing liquid. Gas-solid mass transfer is generally regarded as being fast enough not to be rate-determining.
• Liquid velocity and liquid velocity profiles will affect gas-liquid and liquid-solid mass transfer coefficients.
• The geometry and the extent of fractional wetting affect intraparticle diffusion (Yentekakis and Vayenas, 1987).
• Maldistribution of the liquid can cause parts of the catalyst bed to be almost completely dry or almost completely flooded. The former gives rise to bed-scale incomplete catalyst utilisation if the liquid reagent is non-volatile, or the formation of hot spots if reaction can occur in the gas-phase (Sedriks and Kenney, 1972). The latter can result in part of the bed becoming completely deprived of the gaseous reagent (Ravindra et al., 1997b). In this thesis, the interaction between liquid flow morphology and the different mass transfer steps is collectively termed liquid-solid contacting. The primary focus is on liquid-solid mass transfer and internal diffusion as affected by liquid flow morphology.
Since the study of partial wetting is integral to all work, the study focuses mainly on trickle flow, where partial wetting occurs and mass transfer effects are of major importance.
Several other flow regimes are possible for gas-liquid downflow in trickle-bed reactors. Flow maps (Satterfield, 1975; Gianetto and Specchia, 1992; Sie and Krishna, 1998) and correlations (Fukushima and Kusaka, 1977; Larachi et al., 1999) can be used to determine the flow regime applicable to a specific reactor. In terms of pilot and industrial trickle-bed reactors, the trickle and pulsing flow regimes are the most important (Gianetto and Specchia, 1992). An excellent hydrodynamic description of pulse flow as a hybrid between trickle flow (at the high interaction boundary) and dispersed bubble flow is provided by Boelhouwer et al. (2002).
Wetting efficiency
The first wetting efficiency correlation that was used in TBR studies/modelling is based on packed column data (Puranik and Vogelpohl, 1974). Most wetting efficiency data are derived from a tracer response measurement technique, based on the effect that intraparticle diffusion has on a tracer response curve. The technique was first proposed by Colombo et al. (1976) and later streamlined by Mills and Dudukovic (1981). Based on a Thiele modulus argument related to the work of Dudukovic (1977), the wetting efficiency is taken as the square root of the apparent effective diffusivities measured for trickle- and single-phase liquid flow. A more complete theoretical development of the effect of partial wetting on a tracer response curve was given by Ramachandran et al. (1986), but was never used for the measurement of wetting efficiency. The work was however used by Julcour-Lebigue et al. (2007), for a theoretical validation of the usual tracer technique. The major disadvantage of the tracer wetting efficiency measurement method is that it is model-based.
For the purpose of correlating wetting efficiency with liquid and gas properties and operating conditions, the tracer technique has always been the most important tool for data generation. Other important wetting efficiency measurement methods are the dissolution method (Specchia et al., 1978; Lakota and Levec, 1990; Gonzalez-Mendizabal et al., 1998), colorimetry and reaction methods. Pironti et al. (1999) proposed a wetting efficiency measurement method based on pressure drop, which was later shown to be inaccurate by Baussaron et al. (2007).
The dissolution method compares dissolution rates for trickle flow with dissolution rates in liquid-full operation at the same interstitial velocities. The disadvantage of this method is that liquid-full operation measurements are required over the whole flow range and that a good estimate is needed for liquid holdup (interstitial velocity). Also, it is not completely certain whether or not single-phase liquid flow and trickle flow has the same mass transfer characteristics at the same interstitial velocities. A second disadvantage is that non-porous or slightly porous soluble packing material is used, contrary to the porous catalysts encountered in trickle-bed reactors.
Colorimetry makes use of colourant in the liquid to colour particles in the bed where they were in contact with the liquid. The bed is dismantled, and the particles are then examined through optic methods such as photography and subsequent image processing.
The major advantage of colorimetry is that it is direct and no model or assumptions are needed. Also, more information can be extracted about catalyst wetting other than the average wetting efficiency. The liquid flow pattern should however be stable so that the wetting efficiency is not overestimated. The major disadvantage is that the method is destructive and requires bed re-packing after each experimental run. Other aspects of and the possible pitfalls in the colorimetric evaluation of wetting efficiency is discussed in Chapter 3, which reports a colorimetric study to obtain the distribution of particle wetting. Until recently, very little colorimetric wetting efficiency data were available in the literature, with the first data for trickle-bed reactors reported by Lazzaroni et al. (1988).
Thereafter, Ravindra et al. (1997a) performed a colorimetric study to describe trickleflow morphology rather than to measure wetting efficiencies. Recently, Baussaron et al. (2007) generated an extensive amount of colorimetric wetting efficiency data for several fluids, expanded by Julcour-Lebique et al. (2009) to propose a colorimetry-based wetting efficiency correlation.
In terms of direct applicability, wetting efficiency measurements from reactor studies are arguably the most important. After all, an important reason for the study of wetting efficiency or any hydrodynamic parameter is to understand the possible effects it has on reactor performance. The biggest disadvantage for reaction methods as a tool for parametric studies of wetting efficiency, is that most reaction studies are specific to a certain system with specific reagents and catalyst combined under specific operating conditions. Also, the existing reaction methods are based on some reactor model that require estimates for either external mass transfer or particle kinetics or both. Specific reaction-based wetting efficiency studies are discussed in more detail in section 2.3.
Figure 2.2 shows wetting efficiency correlations developed from data that were obtained with different measurement methods. Predicted wetting efficiencies are applicable to the reaction system that is employed in the study that is reported in Chapter 5. Also included is the prediction of the artificial neural network (ANN) of Larachi et al. (2001). It is based on almost all the known wetting efficiency data published before the correlation was developed, and can therefore be recommended for trickle-bed reactors where there are no experimental studies at the exact reactor conditions. However, empirical correlations can fail to extrapolate accurately, and it is often recommended that correlations that were developed from data generated at conditions close to the conditions of interest should be used (Dudukovic et al., 2002). To reduce sensitivity to extrapolation, Iliuta and Larachi (1999) developed a semi-theoretical model which integrates liquid holdup, pressure drop and wetting efficiency.
1 Introduction
2 Literature
2.1 Liquid-solid contacting: A short historical overview
2.2 Measurement and correlations for solid-liquid contacting
2.3 Reactor studies and liquid-solid contacting
2.4 Conclusions
3 Visualisation of wetting morphology
3.1 Experimental
3.2 Results and discussion
3.3 Summary
4 Effectiveness factors for partially wetted catalysts
4.1 Numerical method
4.2 Monodispersed particles
4.3 Eggshell particles
4.4 Summary
5 Liquid-solid contacting in a pilot reactor
5.1 Finding an applicable reaction system
5.2 Reaction system characteristics
5.3 Pilot studies
5.4 Conclusions
6 Closing remarks
