Standard Article

You have free access to this content

Reactive Distillation

  1. Michael Sakuth1,
  2. Dieter Reusch2,
  3. Ralf Janowsky3

Published Online: 15 JAN 2008

DOI: 10.1002/14356007.c22_c01.pub2

Ullmann's Encyclopedia of Industrial Chemistry

Ullmann's Encyclopedia of Industrial Chemistry

How to Cite

Sakuth, M., Reusch, D. and Janowsky, R. 2008. Reactive Distillation. Ullmann's Encyclopedia of Industrial Chemistry. .

Author Information

  1. 1

    Sasol Solvents Germany GmbH, Moers, Germany

  2. 2

    Degussa AG, Marl, Germany

  3. 3

    Degussa AG, Mobile, Alabama, United States

Publication History

  1. Published Online: 15 JAN 2008

Chemistry Terms

Choose one or more boxes to highlight terms.

Abstract

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading

The article contains sections titled:

1.Introduction
2.Mathematical Modeling of Reactive Distillation Processes
2.1.Equilibrium-Based Models
2.2.Rate-Based Models
3.Design of Reactive Distillation Processes
3.1.Procedures for Process Design Studies
3.2.Flow sheet for Process Development
4.Industrial Applications
4.1.Commercial Packing Structures
4.2.Industrial Catalytic Distillation Processes

1 Introduction

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading

Reactive distillation (RD) is a process in which a catalytic chemical reaction and distillation (fractionation of reactants and products) occur simultaneously in one single apparatus. Reactive distillation belongs to the so-called “process-intensification technologies”. From the reaction engineering view point, the process setup can be classified as a two-phase countercurrent fixed-bed catalytic reactor.

In the literature this integrated reaction – separation technique is also known as catalytic distillation (CD) or reaction with distillation (RWD). According to [1], CD is a process in which a heterogeneous catalyst is localized in a distinct zone of a distillation column. RD is the more general term for this operation, which does not distinguish between homogeneously or heterogeneously catalyzed reactions in distillation columns. RWD is a trademark of the Koch Engineering Company for reactive distillation technology that uses their KataMax packing structures. A brilliant overview on the current status of RD technologies, modeling, industrial applications, etc., can be found in [2].

The present article exclusively deals with RD processes that operate with a heterogeneous catalyst system, i.e., CD technology. The most important advantage of CD technology for equilibrium-controlled reactions is the elimination of equilibrium limitation of conversion by continuous removal of products from the reaction mixture. It is the application of Le Chatelier′s principle to displace the chemical equilibrium by increasing the concentrations on the one side of the reaction, i.e., the reactants, and decreasing it on the other, i.e., the product side. The chemical composition at this equilibrium point can be calculated by means of the Gibbs energy of reaction at a given temperature. Activities must be used to recalculate the composition from the equilibrium constant (i.e., the molar fractions of the components).

Usually, a partially converted reaction mixture, close to chemical equilibrium, leaves the fixed-bed reactor section and enters the CD column in the fractionating zone to ensure the separation of products from feedstock components. The fractionated unconverted feedstock components enter the catalytic section in the CD column for additional or total conversion. The catalyst packing zone is installed in the upper or lower-middle part of the column, with normal distillation sections above and below.

CD technology has several advantages over conventional operating methods, such as a fixed-bed reactor connected to a fractionating column, in which the distillate or bottoms have to be recycled after further separation steps for a total overall conversion.

Apart from increased conversion, the following benefits can be obtained [1]:

  • The most important benefit of CD technology is the lower capital investment, because two process steps can be combined and carried out in the same device (so called “process intensification”). Such integration leads to lower costs for pumps, piping and electrical instrumentation.

  • If CD is applied to an exothermic reaction, the reaction heat can be used to vaporize part of the surrounding liquid, which represents three fundamental advantages: The maximum temperature in the structured catalytic packing is limited to the boiling point of the reaction mixture, so that the danger of hot spots is reduced significantly (so-called “Siedekühlung”). Also, extremely simple and reliable temperature control is achieved. In addition, the integration of reaction heat in the distillation process leads to energy savings by reducing reboiler duty.

  • Product selectivity can be improved owing to fast removal of reactants or products from the reaction zone. Thus, the probability of consecutive reactions, which may occur in the conventional operation mode, is generally lowered.

  • If the reaction zone in the CD column is located above the feed point, poisoning of the catalyst can be avoided. This leads to longer catalyst lifetime compared to the conventional mode of operation.

  • The possibility to break azeotropes in the vapor - liquid equilibrium, because reactants or products can act as entrainers or because the azeotropes can simply disappear.

There are three important constraints for applying CD technology to catalytic chemical reactions:

  • The use of CD technology is only possible if the temperature window of the vapor – liquid equilibrium is equivalent to the reaction temperature. By changing the column operating pressure, this temperature window can be altered.

  • The flexibility in the operating temperature of a CD column is not only restricted by the fact that two phases are required for the distillation process. Also the thermal stability of the catalyst can limit the upper operating temperature.

  • Because of the necessity of wet catalyst pellets, the chemical reaction must take place entirely in the liquid phase.

  • As it is very expensive to change the catalyst in the structured packing of a CD column, only catalysts with a long lifetime are suitable for this process.

As a “fourth constraint” CD technology is somehow difficult to model mathematically, which complicates the scale-up from technical plant scale to full production scale as well [2].

In the literature, it can be found that endothermic reactions are not suitable for the CD technology, because the reaction heat condenses part of the vapor stream. Although endothermic reactions require more reboiler duty and therefore exhibit no large energy savings, there are no restrictions with regard to the application of this technology [3].

Chemical reactions, which may benefit from CD technology, should fulfill the above-mentioned criteria in general. Reactions of this type include, for example, etherifications, esterifications, transesterifications, hydrations, hydrolysis, condensations, hydroisomerizations, oligomerizations, alkylations, transalkylations, and selective hydrogenations.

An excellent overview of the current status of published applications is given in [2].

2 Mathematical Modeling of Reactive Distillation Processes

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading

2.1 Equilibrium-Based Models

Multicomponent separation processes, such as normal distillation processes, have been modeled by using the equilibrium-stage concept for a century. Therefore, early works on reactive distillation also used the equilibrium-stage model to simulate reactions with superimposed distillation.

The principal assumption of the equilibrium-stage model is that the vapor and the liquid stream that leave the stage are in thermodynamic equilibrium. In most real distillation columns, of course, the residence time is too short to reach total equilibrium. For this reason, efficiencies have been introduced into the model (e.g., Murphree efficiency, vaporization efficiency, etc.) to account for the nonideal behavior. MESH (i.e., Material balance, Equilibrium relationship, Summation of all substances, and enthalpy balance H) equations are used to simulate conventional distillation columns ([RIGHTWARDS ARROW]Distillation, 1. Fundamentals – Multi-Component Mixtures).

To introduce the chemical reaction superimposed to the distillation further equations are needed to simulate reactive distillation processes ([RIGHTWARDS ARROW]Reaction Columns – Mathematical Model). The simplest way to consider chemical reactions is to use the equilibrium constant Ki, but many reactions are not fast enough to reach chemical equilibrium in one theoretical stage. Therefore, it is often necessary to use kinetic expressions, which describes the reaction rate as a function of temperature and concentration activities of each component of the reaction scheme.

2.2 Rate-Based Models

To describe the real phenomena of mass transfer in distillation processes in more detail, a second-generation nonequilibrium model was developed by Krishnamurthy and Taylor [4]. A nonequilibrium stage (Fig. 1) represents either a single tray or a section of packing in a column.

original image

Figure 1. Nonequilibrium based stage model without feed and side streams

Contrary to the equilibrium-based stage model, this rate-based model treats the vapor and liquid phases separately and combines them by means of a mass transfer rate and a heat transfer rate through the interfacial area. The total mass transfer rates are:

mathml alt image(1)
mathml alt image(2)

where NjV or NjL are the vectors of molar fluxes of the vapor or liquid stream, respectively, ctj are the molar densities, kj are the matrices of mass transfer coefficients, aj is the specific interfacial area, and NtjV or NtjL are the total mass transfer rate of the vapor or liquid stream, respectively [5].

At the interface of each stage j phase equilibrium (E equation, see above):

mathml alt image(3)

and summation equations (S equations, see above) for each component i:

mathml alt image(4)

as well as

mathml alt image(5)

must be fulfilled.

To simulate the reactive distillation process, an expression for the reaction rate rij = f (temperature, concentrations/activities) must be added. For catalytic distillation the chemical reaction is taken into account in the material balance of the bulk liquid phase

mathml alt image(6)

where sjL is the liquid side-withdrawal and fijL is the feed flow rate of component i to stage j in the liquid phase. For many catalytic distillation processes, liquid – catalyst mass transfer must also be considered.

The reaction heat ΔHr which is produced (exothermic reaction) or consumed (endothermic reaction) is considered in the heat balance of the liquid phase (see Fig. 1).

Yuxiang and Xien [6] used this nonequilibrium stage model with simplified expression of the reaction kinetics based on concentrations to simulate the synthesis of MTBE, which they investigated experimentally in a column filled with catalyst granules in cloth pockets. The results of their simulation agreed fairly well with the experiments. However, Sundmacher and Hoffmann [7] investigated MTBE formation over a broad concentration range and found that the macrokinetics could be correctly described only by using liquid-phase activities for the intrinsic reaction kinetics and liquid – solid mass transfer.

The equilibrium-based stage model is an excellent pragmatic approach to simulate reactive distillation processes, especially those systems with homogeneous reactions. This model is part of all available process simulators (e.g., RADFRAC in ASPEN-PLUS, Aspen Technology, Inc.). For a detailed study of the superimposed distillation and interactions between reaction with mass transfer a rate-based approach is usually preferred (e.g., RATEFRAC in ASPEN-PLUS, Aspen Technology, Inc.). Due to its complexity and the fact that only limited data are available on mass-transfer coefficients, this approach does not find a broad usage. Furthermore, small deviations in these coefficients can have a severe negative impact on the simulation results [8].(see Fig. 2)

original image

Figure 2. Model complexity in simulation of reactive distillation systems

3 Design of Reactive Distillation Processes

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading

3.1 Procedures for Process Design Studies

Only a few systematic methods are available for the process design of multifunctional units such as reactive distillation columns. One method is the concept of reactive distillation lines with reactive azeotropes by Buzad and Doherty [9], Bessling et al. [10] and Stichlmaier et al. [11]. A brief description of this method is given below. More details can be found in [2].

Figure 3 shows the transformation of coordinates for the equilibrium reaction A + B ⇌ C. This transformation reduces the equilibrium concentration of component C to a single point on the line of the transformed coordinates A and B. Mathematically one can define a chemical equilibrium reaction by Equation 7:

mathml alt image(7)

where νi are the stoichiometric coefficients of the reactants A, B, etc., and the desired products P, etc. The transformed liquid and vapor coordinates for an arbitrary component i, i.e., Xi and Yi, are given by Equation 8 [12].

mathml alt image(8)
mathml alt image(9)

where R is a chosen reference component. R can be reactant or a product component. The stoichiometric coefficient νT in these equations is defined as:

mathml alt image(10)
original image

Figure 3. Transformation of coordinates for a equilibrium reaction A + B ⇌ C.

Both the vapor and liquid molar fractions of component P are set to zero, i.e., XP = YP = 0. According to Equation 8, the transformed composition coordinates fulfill the unity constraints:

mathml alt image(11)

By using these transformed coordinates for the feed components A and B, the mass balance and operating lines of a distillation column can be set up in the same way as for a system without a reaction. It reveals that feasible top and bottom compositions have to meet the lever rule ([RIGHTWARDS ARROW]Distillation, 1. Fundamentals – Binary Mixtures) through the feed concentration and an appropriate distillation line, as shown in Figure 4. This is the same procedure as for conventional nonreactive systems.

original image

Figure 4. Application of the concept of reactive distillation lines. As mass balance must be fulfilled, the feed concentration XF is located on a straight line between the transformed concentrations XD and XB.

Figure 5 shows reactive distillation lines for the equilibrium reaction A + B ⇌ C in the presence of an inert component D. Note that product C is no longer necessary, because of the transformed-concentration coordinates between A and B (see Eq. 8). For a given feed concentration F, two process layouts can be derived. In design type I, component C decomposes, and B can be separated from A and D. With design type II, component C is formed, and the inert component D is separated from a mixture of A and B, which is in chemical equilibrium with C. If the equilibrium constant of the reaction A + B ⇌ C is increased, a reactive azeotrope appears along line AB (Fig. 6). According to Ung and Doherty [13] this reactive azeotrope is characterized by identical vapor and liquid transformed-concentration coordinates (Eq. 12).

mathml alt image(12)
original image

Figure 5. Reactive distillation lines for an equilibrium reaction A + B ⇌ C in the presence of an inert component D (LB = low-boiling, MB = medium-boiling, HB = high-boiling).

original image

Figure 6. Reactive distillation lines for an equilibrium reaction A + B ⇌ C in the presence of an inert component D with a reactive azeotropic point (LB = low-boiling, MB = medium-boiling, HB = high-boiling).

In this case the reactive azeotrope and the inert component D are nodes; components A and B are saddles. Therefore, in contrast to Figure 5, it is not possible to realize design type I (decomposition of C). Only design type II, i.e., the formation of C and separation of D at the top and A, B, and C at the bottom of the RD column, is feasible. The concept of reactive distillation lines with reactive azeotropes can also be extended to several other reaction types and to nonideal reactive distillation systems as well.

In collaboration between the chemical industry and different European universities (EU-research project “Brite-Euram”), two computational tools have been developed for the design of RD processes, namely SYNTHESIZER and DESIGNER. SYNTHESIZER is based on the above-mentioned analogy between normal distillation processes and RD processes. It investigates the feasibility of RD for an existing process during re-engineering or in the planning stage for a new process. DESIGNER gives a more precise insight into the RD process with respect to column parameters (e.g., column diameter, etc.). It contains calculation procedures with equilibrium-based stage models as well as rate-based models for a specific process design [14].

3.2 Flow sheet for Process Development

Figure 7 shows a flow sheet as a conceptual basis for the development of a CD process.

original image

Figure 7. Flow sheet for the process development of a CD technology (CSTR = continuous stirred tank reactor)

(*: see Chapter Introduction)

The key feature of this flow sheet is the feasibility study followed by pilot plant experiments (optional) and a basic engineering concept. After using the above-mentioned tools for process synthesis and design (see Section Procedures for Process Design Studies), the parameters of a kinetic model must be determined in an appropriate experimental setup.

Simulation studies with a simple equilibrium-based stage model, such as RADFRAC (ASPEN-PLUS), should reveal the optimal feed point to the CD column, the operating conditions of the CD column, and the position of the reaction zone in the CD column. As a first approximation, these studies can be carried out by using equilibrium constants to incorporate the chemical reaction into the column (less complex model simulation, see Fig. 2.). For a more detailed investigation, it is recommended to use an appropriate activity-based reaction expression with the determined kinetic parameters in a conventional simulator.

In pilot plant experiments, which are carried out in a CD column with an inner diameter of 50 – 100 mm, the simulation studies are examined. As the process is iterative, these experiments are needed to check and to improve the simulations.

As the last step of the feasibility study, all obtained results are fed into a final simulation to decide whether an RD or conventional setup (e.g., reaction section followed by separation steps with internal recycle streams) should be applied. Preferably, a rate-based model, such as for example RATEFRAC (ASPEN-PLUS), is used to check whether any mass-transfer limitations interfere with the removal of reactants. Such limitations can cause undesired consecutive reactions in the reaction zone.

As it is reported by some authors [2] it should be mentioned that predictive scale-up procedures from lab-scale experiments (50 – 80 mm inner diameter column) to pilot-plant scale (≈300 mm inner diameter column) already reveal some evident deviation in conversion, e.g., for methyl acetate synthesis or MTBE decomposition in a CD column. The origin of this effect is up to now not fully understood.

The possible reasons for the observed deviations, discussed in [2], could be:

  • reduced reaction rates due to incomplete catalyst wetting,

  • mass-transfer limitations, or

  • maldistribution, etc.

Sensitivity studies of different CD column parameter can help to understand the order of magnitude of the deviations on the overall conversion, if scale-up calculations are made. These studies can facilitate improvement of the reliability of such simulations.

4 Industrial Applications

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading

4.1 Commercial Packing Structures

All commercial CD packing structures have in common that the catalyst is located in a distinct zone of a distillation column. Apart from the problem of positioning these packing structures, adequate contact between catalyst surface and liquid phase must be ensured.

Nowadays, in most industrial processes macroporous acidic ion-exchange resins are used, but any other solid catalyst pellet can be placed in the pockets of a structured packing. As the entire catalyst structure must be removed from the distillation column to exchange the catalyst, one evident prerequisite for the catalyst is a long lifetime (more than about three years). Catalyst replacement means shutting down whole plants which results in a significant loss of costs and operating time.

CD Tech Catalyst Bales. This type of spatial configuration in a distillation column was first developed and commercialized by CR & L (Chemical Research & Licensing Company) in 1980 [15]. It is now licensed by CD Tech (Chemical Distillation Technologies), a joint venture between CR & L and ABB Lummus Crest. In this technology the catalyst is supported in a fiberglass cloth, which is wrapped with stainless steel wire mesh. The demister wire mesh has two tasks. First, it stabilizes the fiberglass packing; second, it provides the void space necessary for distillation. These fiberglass cloths are rolled into catalyst bales and stacked on sieve trays in the column (Fig. 8).

original image

Figure 8. CD Tech catalyst bales

KataMax and KataPak Technology. Analogous catalytic packing configurations were developed in 1991/1992 independently by Koch Engineering Company as KataMax technology [16] and by Gebrüder Sulzer [17] as KataPak technology. Here the catalytic section in the CD column has a conventional corrugated structure, wire-cloth distillation packing which contains appropriate catalyst pellets for the desired reaction. The solid catalyst is held in an envelope of meshed and crimped screen, sealed at the edges. These envelopes are stacked together in modular blocks, so that the crimp of one envelope has a 90 degree angle of inclination to the crimp of the next. In this architecture, channels are formed for the rising vapor flow and descending liquid flow. At the intersections, they provide intensive mixing of vapor and liquid phases and radial distribution in a block. It is reported [18] that the performance parameters of the KataMax packing (i.e., the hydraulic and mass transfer efficiencies) are similar to those of standard distillation devices (e.g., the Flexipac packing). At liquid loadings of 10 – 50 m3m−2h−1 catalyst contact efficiency exceeded 75 – 90 %. Catalyst contact efficiency is the ratio of the rate constant of the catalyst in the packing to the rate constant of the catalyst in a stirred tank reactor for a simple liquid-phase first-order reaction at the same temperature.

Usually the catalyst content of KataMax or KataPak lies between 20 and 25 vol. %. For KataMax a HETP value (height equivalent to a theoretical plate) between 0.4 and 0.6 m can be used for simulation. For KataPak the HETP is 1.0 m at normal F factors ([RIGHTWARDS ARROW]Distillation, 2. Equipment – Operation Region of Tray Columns). The specific surface area of KataMax is 210 m2/m3 and of KataPak, 85 or 125 m2/m3 (KataPak-S 170.Y or KataPak-S 250.Y, respectively).

The benefits of this type of packing are a very good distribution of liquid and vapor phases at a low pressure drop, efficient contact of the reactants with catalyst pellets and instantaneous distillative removal of reactants. Therefore, it combines excellent separation performance with an efficient mass and heat transfer for chemical reaction.

MultiPak Technology. This structured catalyst packing was developed by the company Montz in 1998 in cooperation with the University of Dortmund [19]. In MultiPak technology the stacks are built up of alternating layers of meshed, but not crimped, catalyst envelopes and crimped structured distillation sheets. From the viewpoint of reaction and fractionating behavior, it has the same performance as KataMax or KataPak. The advantage of MultiPak is the identical geometric structure in an 80-mm laboratory packing and a 300- or 1000-mm pilot plant packing. Therefore, scaleup from pilot plant experiments to industrial plants is much easier. It is claimed that a defined flow pattern of vapor and liquid is achieved by this construction. This has advantages when resins are used as catalyst. As there is only liquid phase in the catalyst envelopes, the pellets are completely wetted under process conditions, and there is no swelling or shrinkage due to alternating contact with vapor and liquid. The disadvantage is a higher pressure drop along the structured packing.

Multichannel Packing [20]. Details on this type of catalyst packing were published in 2003 by BASF. The point of departure to develop these packings was to improve trickle-bed reactor performance by installing corrugated distillation packings inside the reactor and pouring catalyst particles into them. The catalyst particles are brought loose into the packing cavities, distributed under the influence of gravity. By alternating packing layers of high and low specific surface areas (so called “catalyst barrier layer”) the catalyst could be held in a specific area of the CD column. The advantage of these packings is their simplicity and the ease of catalyst removal in case of deactivated catalyst. As the way of introducing the catalyst into a specific column segment is similar to a normal trickle-bed reactor, the authors describe also a simpler modeling in case of scale-up studies.

However, if the catalyst size distribution is not 100 % uniform, some particles may enter the lower part of the CD column. Especially with reversible reactions, which can only be performed in a specific segment of a CD column setup (e.g., MTBE synthesis), the undesired reverse reaction could occur in the bottom, if catalyst particles are present there.

Reactive Rings. The reactive rings of VEBA Oel [21] are based on packing bodies such as Raschig rings or Berl saddles. On the outer and inner surface of these bodies, macroporous ion exchange resins are bonded chemically or physically. For example, the active Raschig rings are prepared by using an open-celled sintered glass, which is impregnated with a mixture of styrene, a long chain alkane, p-divinylbenzene, and a radical starter (AIBN: azobisisobutyronitrile). After the polymerization procedure, the resin is sulfonated with chlorosulfonic acid to build up the active catalytic sites.

These types of reactive rings have been used in an etherification process in a pilot plant column [7, 22].

Other Technologies. Another interesting, but as yet not commercialized, technology for immobilizing catalyst pellets in a distinct zone of a CD column is to incorporate a fixed-bed in the downcomer of a conventional tray column [23].

4.2 Industrial Catalytic Distillation Processes

Catalytic distillation can be used in catalytic chemical reactions which are limited by a chemical equilibrium. There are various reactions that satisfy this criterion, but only for etherification, esterification, and alkylation (synthesis of ethylbenzene or cumene) is this technology applied on an industrial scale. An overview of further possible applications can be found in [1, 2, 22].

Etherification Processes (see also [RIGHTWARDS ARROW]Methyl Tert-Butyl Ether). The application of CD technology to etherification is limited to synthesis of methyl tert-butyl ether (MTBE), ethyl tert-butyl ether (ETBE), tert-amyl methyl ether (TAME), and tert-amyl ethyl ether (TAEE).

Because of their good octane enhancing properties and low volatilities these ethers are excellent gasoline blending compounds (so-called oxygenates). Reduction of the tetraethyllead content in gasoline in the mid-1970s led to a dramatic increase in the demand for octane enhancers, with MTBE being used increasingly. Driven by the U.S. Reformulated Gasoline Program, which mandated a minimum oxygen content in gasoline of 2 %, world MTBE production has reached ca. 19 × 106 t/a in 1997 [24]. The successful use of CD technology in MTBE production led to its worldwide establishment.

It should be mentioned that end of 1997 California started to ban MTBE from the gasoline pool due to its effect as a water contaminant and its possible cause of health problems (see also [RIGHTWARDS ARROW]Methyl Tert-Butyl Ether). In mid-2006 the U.S. Environmental Protection Agency terminated the legal requirement of using MTBE in the gasoline pool [25]. The production capacities of MTBE are therefore shrinking day-by-day.

Table 1 summarizes processes for ether production based on CD technology. As shown in the simplified process flow diagram for the ETHERMAX process (Fig. 9), all processes have in common a reactor section followed by a CD column. There is no industrial process in which the whole conversion is performed in a single CD column. A partially converted mixture from the reactor section, which is nearly in chemical equilibrium, enters the CD column below the catalyst packing zone to ensure the separation of the ether from the feed stream. The catalyst packing is installed in the upper middle of the column, with normal distillation sections above and below.

Table 1. Overview of CD processes for ether production [27]
Process (licensor)Process descriptionCatalystMarket share*
  • *

     For MTBE, produced by CD technology in 1997

ETHERMAX (UOP)fixed-bed tubular reactor or adiabatic recycle reactor followed by a reactive distillation column with KataMax Packingsulfonated ion-exchange resinca. 32 %
CDMTBE & CDTAME (CD-Tech)adiabatic reactor operating at the boiling point followed by a reactive distillation column with CD Tech Balesacidic ion-exchange resinca. 58 %
CDETHEROL (CD-Tech)adiabatic fixed-bed reactor operating at the boiling point followed by a reactive distillation column with CD Tech Balestrifunctional ion-exchange resin
CATACOL (IFP)recycle reactor, fixed-bed reactor followed by a reactive distillation columnacidic ion-exchange resinca. 10 %
original image

Figure 9. Simplified flow diagram for the ETHERMAX process

Usually, the reaction of the isoalkene with methanol or ethanol is conducted in the presence of a slight stoichiometric excess of the alcohol. In addition to the advantages of a higher selectivity for the ether (lower formation of C8 and C10 dimers) and shifting of the chemical equilibrium to the product side, this leads to a secure process control. In the absence of the alcohol in the reaction zone of the CD column exothermic dimerization and oligomerization of the C4 and C5 alkenes takes place at high reaction rates. A sharp and excessive temperature rise (hot spot) causes irreversible catalyst deactivation and catalyst damage. The excess alcohol is collected in the overhead of the CD column and separated from the hydrocarbon stream by extraction with water.

In the case of MTBE production, a conversion of up to 99.9 % is possible by using this technology, if sufficient ether is present at the bottom of the column. Because of a minimum boiling azeotrope in the binary MTBE – methanol system, the ether carries the alcohol into the reaction zone of the CD column. This unusual vapor – liquid behavior and the high difference in the heat of vaporization between C4 hydrocarbons and methanol are possibly responsible for the known multiple steady states in CD column operation [22, 26].

The TAME process usually gives a lower conversion of 91 – 95 % only.

Esterification Processes. Esterification is a good example for the beneficial use of the CD technology. In conventional methyl acetate production the recovery of methyl acetate from the reactor outlet stream is complicated, because methanol forms a low boiling azeotrope with water and the separation of water from unconverted acetic acid is difficult. Due to these separation problems, the old Eastman Kodak process used a reactor coupled with eight distillation columns and an extraction column [28].

In the reactive distillation process, almost pure methyl acetate can be collected in the overhead of a single CD column at acetic acid conversions of greater than 99 %. The Eastman Kodak process uses sulfuric acid as catalyst [29], but a heterogeneous catalyst system (acid ion-exchange resin) can also be used successfully [30].

The conceptual basis for the successful implementation of reactive distillation in methyl acetate synthesis is shown in Figure 10. There are four zones in the column, which ensure the fairly high conversion. Acetic acid is separated from methyl acetate at the top of the column (zone I). In zone II, an extractive distillation section below the acetic acid feed extracts water from methyl acetate. The reaction takes place in the middle of the column (zone III). At the bottom (zone IV) methanol is fed and stripped from descending byproduct water.

original image

Figure 10. Methyl acetate synthesis with CD technology

Alkylation Processes. CD-technology is also applied for alkylation reactions in the case of ethylbenzene and cumene manufacture [1]. As the two processes are fairly similar, only the cumene process is discussed here.

In the CDCUMENE process, CD Tech catalyst bales are filled with an acidic, wide-pore zeolite catalyst, for example, zeolite Y, β or ω [31]. The bales are stacked in the middle of a distillation column. Below the reaction zone propene is fed; benzene is passed as reflux to the top of the column. The propene concentration in the reaction zone is held fairly low to slow down a side reaction in which diisopropylbenzene (DIPB) and triisopropylbenzene (TIPB) are formed. The cumene product with impurities of ethylbenzene and n-propylbenzene, is taken from the bottom of the column. The overhead pressure of the CD column is maintained at 5 bar.

The advantages of CD technology in cumene synthesis are lower formation of oligo-isopropylbenzenes (DIPB and TIPB), higher catalyst lifetime, and a higher conversion of benzene [32]. The internal benzene recycle to the reaction zone, which is a result of the distillative cumene – benzene separation in the lower portion of the column, ensures the observed higher conversion.

4.3 Novel Application of CD with regard to Process Intensification

Distillation columns with dividing walls are used with success in some chemical processes. An EU-research project named INSERT has been started to investigate if CD technology can also be applied advantageously to such distillation columns with divided wall internals [33].

The successful use of CD technology in a divided wall column is reported in [34] for the transesterification of methyl acetate to butyl acetate (Fig. 11.). This process can be performed in one column of this type instead of using one CD column followed by two normal distillation columns.

original image

Figure 11. Transesterification of methyl acetate to butyl acetate with CD technology in a divided wall column

Symbols and Abbreviations
aj

specific interfacial area, m2/m3

ctj

mixture molar densities, kmol/m3

E

energy transfer rate, W

fijL

feed flow rate of component i to stage j in the liquid phase, kmol/s

kj

matrix of mass transfer coefficients, m/s

kij

equilibrium ratio of component i on stage j

Lj

liquid flow rate from stage j, kmol/s

Lj−1

liquid flow rate to stage j, kmol/s

NijL

liquid phase mass transfer rate of component i, kmol/s

Nj

vector of mass transfer rates, kmol/s

Ntj

total mass transfer rate, kmol/s

NC

number of compounds

QL

liquid phase heat loss, W

QV

vapor phase heat loss, W

rij

reaction rate of component i on stage j, kmol/s

SjL

ratio of liquid sidewithdrawal

Vj

vapor flow rate from stage j, kmol/s

Vj+1

vapor flow rate to stage j, kmol/s

Xi

transformed liquid phase composition of component i

xij

mole fraction of component i in liquid phase of stage j

xijI

liquid mole fraction of component i in interface

xj

vector of liquid mole fractions

Yi

transformed vapor phase composition of component i

yij

mole fraction of component i in vapor phase of stage j

yijI

vapor mole fraction of component i in interface

yj

vector of vapor mole fractions

ΔHr

heat of reaction, J/kmol

νi

stoichiometric coefficient for component i

νT

sum of stoichiometric coefficients defined by Equation 9

Superscripts
I

Interface

L

liquid phase

V

vapor phase

Subscripts
B

bottom

D

distillate

F

feed

P

product

R

reference component

i

component number

j

stage number

t

total

Abbreviations and Acronyms
AIBN

Azobisisobutyronitrile

CD

Chemical distillation

DIPB

Diisopropylbenzene

ETBE

Ethyl tert-butyl ether

HB

High-boiling

HETP

Height equivalent to a theoretical plate

LB

Low-boiling

MB

Medium-boiling

MESH

Material balance/equilibrium condition/summation equation/heat balance

MTBE

Methyl tert-butyl ether

RD

Reactive distillation

RWD

Reaction with distillation

TAME

tert-Amyl methyl ether

TAEE

tert-Amyl ethyl ether

TIPB

Triisopropylbenzene

References

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading
  • 1
    G. G. Podrebarac, F. T. T. Ng, G. L. Rempel, CHEMTECH 1997, 37 – 45.
  • 2
    K. Sundmacher, A. Kienle (eds.): Reactive Distillation, Wiley-VCH, Weinheim 2003.
  • 3
    Hüls, EP 0 726 241, 1995 (M. Sakuth, U. Peters).
  • 4
    R. Krishnamurthy, R. Taylor, AIChE. J. 31 (1985) 449 – 456.
  • 5
    R. Taylor, R. Krishna: Multicomponent Mass Transfer, Wiley, New York 1993.
  • 6
    Z. Yuxiang, X. Xien, Trans. IchemE 70 (1992) 465 – 470.
  • 7
    K. Sundmacher, U. Hoffmann, Chem. Eng. Sci. 49 (1994) 3077 – 3089.
  • 8
    A. P. Higler, R. Taylor, R. Krishna, Chem. Ing. Sci. 54 (1999) 1389 – 1395.
  • 9
    G. Buzard, M. F. Doherty, Computers Chem. Engng. 19 (1995) 395 – 408.
  • 10
    B. Bessling, G. Schembecker, K. H. Simmrock, Ind. Eng. Chem. Res. 36 (1997) 3032 – 3042.
  • 11
    J. Stichlmaier, T. Frey, Chem. Ing. Tech. 70 (1998) 1094 – 1095.
  • 12
    J. Espinosa, P. A. Aguirre, G. A. Perez, Ind. Eng. Chem. Res. 34 (1995) 853 – 861.
  • 13
    S. Ung, M. F. Doherty, AIChE J. 41 (1995) 2383 – 2392.
  • 14
    G. Schembecker et al., Chem. Ing. Tech. 70 (1998) 1096.
    Direct Link:
  • 15
    CR & L, US 4 232 177, 1980 (L. A. Smith Jr.
  • 16
    Koch Engineering Company, EP 0 428 265, 1991 (A. P. Gelbein, M. Buchholz).
  • 17
    Gebrüder Sulzer, EP 0 396 650, 1992 (R. Shelden, J.-P. Stringaro).
  • 18
    J. L. DeGarmo, V. N. Parulekar, V. Pinjala, Chem. Eng. Prog. (1992), no. 3, 43 – 50.
  • 19
    Gebrauchsmuster No. 298 7 007.3, 1998 (A. Górak, L. U. Kreul).
  • 20
    BASF, WO 03/047747 A1, 2003 (G. Kaibel et al.).
  • 21
    VEBA Oel, US 5 244 929, 1993 (K. Gottlieb, W. Graf, K. Schaedlich, U. Hoffmann, A. Rehfinger, J. Flato).
  • 22
    K. Sundmacher, Dissertation, Universität Clausthal-Zellerfeld, 1995.
  • 23
    Koch Engineering Company, EP 0 664 721, 1997 (N. Yeoman, R. Pinaire, M. A. Ulowetz, T. P. Nace, D. A. Furse).
  • 24
    Chem. Week, Oct. 29, 1997, 44.
  • 25
    ICIS Chemical Business 2006, Mar./Apr., 32.
  • 26
    S. Hauan, T. Hertzberg, K. Lien, Ind. Eng. Chem. Res. 34 (1995) 987 – 991.
  • 27
    Hydrocarbon Processing 73 (1992) 104 – 110.
  • 28
    B. Bessling, Dissertation, Universität Dortmund, 1998.
  • 29
    V. H. Agreda, L. R. Partin, W. H. Heise, Chem. Eng. Prog. (1990) no. 2, 40 – 46.
  • 30
    J. Gmehling, J. Krafczyk, Chem. Ing. Tech. 66 (1994) 1372 – 1375.
  • 31
    G. R. Meima, M. J. M. van der Aalst, M. S. U. Samson, J. M. Garces, J. G. Lee, “Catalyst of Acid and Bases”, Proc. DGMK-Conf. (1996) 125 – 137.
  • 32
    CR & L, US 4 849 569, 1989 (L. A. Smith Jr.
  • 33
    E. Geißler, C. Großmann, S. Sander, C. Flisch, O. Ryll, H. Hasse, Chem. Ing. Tech. 78 (2006) 1282 – 1283.
    Direct Link:
  • 34
    G. Kaibel, C. Miller, T. Holtmann, H. Schoenmakers, Chem. Ing. Tech. 77 (2005) 1749 – 1758.

Further Reading

  1. Top of page
  2. Introduction
  3. Mathematical Modeling of Reactive Distillation Processes
  4. Design of Reactive Distillation Processes
  5. Industrial Applications
  6. References
  7. Further Reading
  • C. A. M. Afonso: Green separation processes, Wiley-VCH, Weinheim 2005.
  • Z. Lei B. Chen Z. Ding: Special distillation processes, 1st ed., Elsevier, Amsterdam 2005.
  • W. L. Luyben, C.-C. Yu: Reactive distillation design and control, Wiley, Hoboken, NJ 2008.
  • K. Sundmacher: Reactive distillation, Wiley-VCH, Weinheim 2003.