Reactive Dividing-Wall Columns How to Get More with Less Resources?

Chem. Eng. Comm., 196:1366 1374, 2009 Copyright # Taylor & Francis Group, LLC ISSN: 0098-6445 print=1563-5201 online DOI: 10.1080/00986440902935507 Reactive Dividing-Wall Columns How to Get More with Less Resources? ANTON A. KISS, J. J. PRAGT, AND C. J. G. VAN STRIEN AkzoNobel Chemicals, Research and Technology Center, Arnhem, The Netherlands This work presents a novel integrated reactive-separation design based on a dividing-wall column (DWC) applied to an industrial case study within AkzoNobel Chemicals. To the best of our knowledge this is one of the first reported industrial applications of a reactive DWC. Due to changing market conditions, one of the by-products in a plant became more economically attractive than the main product. However, the design of the existing plant does not allow an increase of the by-product production rate at the cost of the main product. To solve this problem we developed a novel integrated design that combines reaction and separation into a feasible reactive DWC that allows 35% savings in capital and 15% savings in energy costs. This article describes the novel reactive DWC design, presents the rigorous simulation results, and makes a comparison with the base case alternative. Keywords Energy savings; Overcoming equilibrium; Reactive distillation; Dividing-wall column Introduction The steep changes in oil and gas prices are boosting the fortunes of companies offering energy-saving technologies. One such technology is the dividing-wall column (DWC) for distillation. Due to its many advantages, distillation is still the major separation process used in the chemical processing industry. However, one important drawback is its considerable energy requirements distillation can generate more than 50% of plant operating cost (Taylor et al., 2003). Process intensification aims at significant capital and energy savings, as well as environmental benefits, by integrating different phenomena or operations, as, for example: reactive separations, dividing-wall columns, heat integrated reactors or columns (Dudukovic, 2009). Along with reactive separations, there is also the possibility to integrate two different separation units. Conventionally, a ternary mixture can be separated via a direct sequence (most volatile component is separated first), indirect sequence (heaviest component is separated first), or distributed sequence (mid-split) consisting of two or three distillation columns (Figure 1). This separation sequence evolved via the Petlyuk column configuration (Petlyuk et al., 1965) consisting of two fully thermally coupled distillation columns (Figure 2, left), to the concept of dividing-wall column (Schultz et al., Address correspondence to Anton A. Kiss, AkzoNobel Chemicals, Research and Technology Center, Velperweg 76, 6824 BM, Arnhem, The Netherlands. E-mail: Tony.Kiss@ akzonobel.com 1366

Reactive Dividing-Wall Columns 1367 Figure 1. Separation of a ternary mixture via direct, indirect, and distributed sequence. 2002). The name DWC is given because the middle part of the column is split into two parts by a wall (Figure 2, right). Feed, typically containing three or more components, is pumped into one side of the column facing the wall. Deflected by the wall, the lightest component A flows upward and exits the column as top distillate, while the heaviest component C drops down and is withdrawn from the bottom of the column. The intermediate boiling component B is initially entrained up and down with both streams, but the fluid that goes upward subsequently separates and falls down on the opposite side of the wall. Similarly, the amount of B that goes toward the bottom separates and flows up to the back side of the wall, where the entire B product is recovered by a side draw stream. DWC is very appealing as it can separate three or more components in a single distillation tower, thereby eliminating the need for a second unit, thus saving the cost of building two columns and cutting operating costs by using a single condenser and Figure 2. Petlyuk column with prefractionator vs. dividing-wall column (DWC).

1368 A. A. Kiss et al. reboiler. Compared to the conventional distillation design arrangements, the DWC offers the following benefits:. Reduced number of equipment units compared to conventional configurations.. Lower energy consumption compared to (in)direct separation sequences.. High thermodynamic efficiency due to reduced remixing effects.. High purity for all three or more product streams reached in only one column. In fact, using dividing-wall columns can save up to 30% in the capital invested and up to 40% in energy costs (Schultz et al., 2002; Isopescu et al., 2008). Moreover, retrofitting conventional column systems to DWC is also a feasible option (Premkumar and Rangaiah, 2009). However, note, that using DWC requires a match between the operating conditions of the two standalone columns, and due to its design limitations, the main weakness of DWC is its inflexibility to changes in the nature of the feed (Becker et al., 2001). Several other successful examples of integrated processes can be found among reactive separations that combine reaction and separation steps in a single unit, such as reactive distillation (Taylor and Krishna, 2000; Kenig et al., 2004; Kiss et al., 2008; Olanrewaju and Al-Arfaj, 2008; Fang and Liu, 2007). Nevertheless note that such integration requires a match of the reaction and separation conditions. Compared to conventional reactor-distillation sequences, the integrated reactivedistillation design brings several advantages such as:. Increased selectivity via suppression of secondary reactions. Increased conversion due to overcoming equilibrium limitations. Reduced energy consumption via in situ heat integration. Avoidance of hot spots due to liquid evaporation. Ability to separate close boiling components. Breaking of azeotropes by chemical reaction DWC and reactive distillation are both improvements of traditional distillation units, but at the same time they correspond to two different ways of integration: separation-separation and reaction-separation, respectively (Mueller and Kenig, 2007). The incentives of these integrated units could be further enhanced if they are combined via an additional integration step. The resulting unit, called a reactive dividing-wall column (RDWC), has a highly integrated configuration that consists of one condenser, one reboiler, reactive zones, a prefractionator, and the main column together in a single-shell distillation setup. RDWC offers an alternative to conventional reactive distillation towers or multicolumn arrangements, with potential for significant cost savings (Becker et al., 2001; Schultz et al., 2002; Barroso-Mu~noz et al., 2007; Hernandez et al., 2009). Regardless of their increased complexity, RDWC systems are still controllable by applying an efficient plant-wide control strategy (Wang and Wong, 2007; Wang et al., 2008). Process Description This article describes a novel integrated design project applicable to an AkzoNobel Chemicals plant. To the best of our knowledge this is one of the first industrial applications of a RDWC reported in literature up to date (Kiss et al., 2007; Sander et al., 2007; Wang et al., 2008; Hernandez et al., 2009).

Reactive Dividing-Wall Columns 1369 The process involves a relatively complex, fast chemical equilibrium of 10 species, denoted below by letters A J, and sorted in descending order of volatility, A being the most volatile and J the heaviest component. Due to the homogeneous catalyst, the reactions take place everywhere in the system, provided that the appropriate reactants are present. Therefore, as not all kinetic parameters were available for this system, the reactions were modeled as fast equilibrium reactions, as follows: 1. A þ J $ C þ H (main reaction) 2. B þ H $ C þ E 3. D þ H $ C þ I 4. B þ E $ A þ F 5. F þ J $ 2G Conventionally, the reactor outlet mixture (F1: ABCDEHI) is separated in a series of distillation columns. Most of the streams are recycled back to the reactor, while component H is purified (min. 98.5%) and afterwards sold as the main product. However, due to market demand changes, the by-product C became more economically attractive than the main product H. Therefore, the production focus needs to change from the main product H to component C, one of the by-products. The problem is that the current plant design is not suitable for producing more of component C at the cost of the main product H. Moreover, the obvious option of adding another reactor and two distillation columns for this production switch was discarded due to unavailable floor area and the high investment costs involved. To solve this problem we investigated a base case design alternative, namely a two-column configuration that uses a reactive distillation column (RDC), followed by a conventional distillation column (DC). Fortunately, the operating parameters, such as temperature and pressure, are similar in these two columns. Therefore the design was further integrated into a reactive dividing-wall column (RDWC) setup that combines the two columns of the base case into one distillation unit. Results and Discussion The conceptual design of the columns was performed using graphical stage composition lines (Matthias et al., 2004) and stage-to-stage methods for reactive distillation column design (Daza et al., 2004), although novel methods have been recently published (Sotudeh and Shahraki, 2007). Considering the presence in the system of a nonideal vapor containing polar, solvating, or associating components, at low to moderate pressures, the nonrandom two liquid equation of state (NRTL-HOC) with Henry s law was selected as the property model (Gutierrez-Antonio et al., 2008). Note that three binary homogeneous azeotropes are present in the system (C-E, D-E, C-I), as illustrated by the residue curve maps shown in Figure 3. However, in an integrated RD setup, these azeotropes can be broken by the chemical reactions previously described. The flowsheet of the base case design (RDC þ DC) is shown in Figure 4. This sequence has two columns, two reboilers, and two condensers, and it requires a great deal of piping and floor area that are not available in the existing plant. However, the advantage of this setup is its flexibility, as the columns can operate at different pressures. Figures 5 and 6 show the composition and temperature profiles in the RD and DC columns. The top product of the first column is a mixture of most volatile

1370 A. A. Kiss et al. Figure 3. Residue curve map of the systems C-D-E and C-E-I. components A, B, and C. The second column separates A þ B in the top and C as bottom product. The temperature profiles in these columns show small differences, suggesting reactive DWC as a logical choice. Note that the dimensionless temperature is calculated by dividing the temperature on a specific stage by the maximum temperature of all columns (T stage =T max ), namely the reboiler temperature of the reactive DWC, the alternative described soon after. In addition to the base case, we considered the more integrated design that combines reaction and separation into one reactive DWC (Figure 7). Note that modeling reactive distillation and DWC is nowadays possible using state-of-the-art rate-based models. (Kenig et al., 2004; Mueller and Kenig, 2007) However, due to the absence of a reactive DWC unit in AspenONE AspenPlus, this was simulated using two coupled rigorous RadFrac (reactive-) distillation units, the thermodynamic equivalent of reactive DWC. The reactive DWC setup consists of only one shell, one reboiler, and one condenser and requires less piping and floor space than the base case. However, the column diameter is relatively larger than the diameter of the columns presented in the base case. The key factor that allows such an integration of two columns into one unit is the similar pressure and temperature conditions in the standalone columns. Figure 4. AspenPlus flowsheet of the two-column distillation sequence.

Reactive Dividing-Wall Columns 1371 Figure 5. Composition profiles in the two columns (base case). The reactive DWC column has 22 stages in total, with the feed located on stage 8, and liquid side-draw from stage 12. Out of these stages, 3 of them are located above and below the dividing wall, for the common rectifying and stripping sections, respectively. Figure 8 shows the liquid composition and the temperature profiles in the RDWC. Chemical reactions take place only on the feed side and bottom sections of the column, where the light components are separated from the heavy ones. A small amount of component H is added on top of the feed location in order break the azeotropes and to push the mid-boiling and heavy components (D J) to the bottom of the column. Moreover, the formation of heavy components F and G (waste by-products) is avoided by adding an extra feed stream of light component A in the Figure 6. Temperature profiles in the two columns (base case).

1372 A. A. Kiss et al. Figure 7. Reactive dividing-wall column alternative and AspenPlus flowsheet. bottom of the column. Reactant A consumes the heavier component F and avoids the parallel conversion of F into by-product G, according to the following reactions: A þ F $ B þ E F þ J $ 2G The product side of the column performs only the separation of product C from B, with no reactions taking place here, as A, B, and C do not react with each other, according to reaction Equations (1) (5). Main product C is collected as a high-purity side stream from the product-side of the column. Note that component C has a high purity on a large range of stages (Figure 8, left), thus the column is very robust and able to cope with disturbances in feed flow rate and composition. This is also one of the reasons to have a large number of stages on the product removal side of the column, similar to the feed side. Figure 8. Composition and temperature profiles in RDWC (C1=C2 feed=product side).

Reactive Dividing-Wall Columns 1373 The temperature differences between the feed- and product-side of the reactive DWC are reasonably small, the maximum difference being less than 25 C (Figure 8, right), hence it can be relatively easily achieved in practice. Note that compared to the base case, the height of the reactive DWC remains the same as that of the RD but the diameter is slightly larger. The economics of the RDWC was compared against the base case process. Rigorous calculations of the equipment costs and investment were performed using the AspenTech ICARUS process evaluator. Since the RDWC makes use of two columns in one shell (DWC) and only one reboiler and one condenser, the investment costs are lower than that of the base case. Note that the equipment cost includes: equipment and setting, piping, civil and electrical, structural steel, instrumentation, insulation, paint, and manpower. For the RDWC case the total investment is 35% less than that of the base case, due to the need for only one condenser and reboiler. In addition, about 15% less energy is required since the mid-boiling product C is evaporated only once, with no remixing effect in the reactive DWC. Conclusions The industrial case study presented in this work shows that equilibrium limitations can be overcome and high purity products can be obtained by integrating reaction and separation into a reactive dividing-wall column. Basically, the reactive DWC unit integrates a reactive distillation column with a conventional distillation tower. The key factor that allows this integration is the similar pressure and temperature profiles in the two standalone columns of the base case design (RD þ DC). The reactive DWC setup offers similar or better performance than the base case design. Moreover, due to the robust design, the column copes well with disturbances in both feed flow rate and composition. Compared to the base case design using two standalone distillation columns, the reactive DWC alternative developed for this industrial application allows 35% savings in capital costs and 15% savings in energy costs, respectively. References Barroso-Mu~noz, F. O., Hernandez, S., Segovia-Hernandez, J. G., Hernandez-Escoto, H., and Aguilera-Alvarado, A. F. (2007). Thermally coupled distillation systems: Study of an energy-efficient reactive case, Chem. Biochem. Eng. Q. J., 21, 115 120. Becker, H., Godorr, S., and Kreis, H. (2001). Partitioned distilliation columns Why, when & how, Chem. Eng., 108, 68 74. Daza, O. S., Pérez-Cisneros, E. S., Bek-Pedersen, E., and Gani, R. (2004). Graphical and stage-to-stage methods for reactive distillation column design, AIChE J., 49, 2822 2841. Dudukovic, M. P. (2009). Challenges and innovations in reaction engineering, Chem. Eng. Commun., 196, 252 266. Fang, Y. J. and Liu, D. J. (2007). A reactive distillation process for an azeotropic reaction system: Transesterification of ethylene carbonate with methanol, Chem. Eng. Commun., 194, 1608 1622. Gutierrez-Antonio, C., Iglesias-Silva, G. A., and Jimenez-Gutierrez, A. (2008). Effect of different thermodynamic models on the design of homogeneous azeotropic distillation columns, Chem. Eng. Commun., 195, 1059 1075. Hernandez, S., Sandoval-Vergara, R., Barroso-Mu~noz, O. F., Murrieta-Due~nas, R., Hernandez-Escoto, H., Segovia-Hernandez, J. G., and Rico-Ramírez, V. (2009). Reactive

1374 A. A. Kiss et al. divided wall distillation columns: Dynamic simulation and implementation in a pilot plant, Chem. Eng. Process., 48, 250 258. Isopescu, R., Woinaroschy, A., and Draghiciu, L. (2008). Energy reduction in a divided wall distillation column, Rev. Chim., 59, 812 815. Kenig, E. Y., Pyhälachti, A., Jakobsson, K., Gorak, A., Aittamaa, J., and Sundmacher, K. (2004). Advanced rate-based simulation tool for reactive distillation, AIChE J., 50, 332 342. Kiss, A. A., Dimian, A. C., and Rothenberg, G. (2008). Biodiesel by reactive distillation powered by metal oxides, Energy Fuels, 22, 598 604. Kiss, A. A., Pragt, H., and van Strien, C. (2007). Overcoming equilibrium limitations in reactive dividing-wall columns, 17th European Symposium on Computer Aided Process Engineering Elsevier, Comput. Aided Chem. Eng., 24, 467 472. Matthias, G., Dragomir, R. M., and Jobson, M. (2004). Conceptual design of reactive distillation columns using stage composition lines, Chem. Eng. Process., 43, 369 382. Mueller, I. and Kenig, E. Y. (2007). Reactive distillation in a dividing wall column Rate-based modeling and simulation, Ind. Eng. Chem. Res., 46, 3709 3719. Mueller, I., Pech, C., Bhatia, D., and Kenig, E. Y. (2007). Rate-based analysis of reactive distillation sequences with different degrees of integration, Chem. Eng. Sci., 62, 7327 7335. Olanrewaju, M. J. and Al-Arfaj, M. A. (2008). Design and implementation of state observers in reactive distillation, Chem. Eng. Commun., 195, 267 292. Petlyuk, F. B., Platonov, V. M., and Slavinskii, D. M. (1965). Thermodynamically optimal method for separating multicomponent mixtures, Int. Chem. Eng., 5, 555 561. Premkumar, R. and Rangaiah, G. P. (2009). Retrofitting conventional column systems to dividing-wall columns, Chem. Eng. Res. Des., 87, 47 60. Sander, S., Flisch, C., Geissier, E., Schoenmaker, S. H., Ryll, O., and Hasse, H. (2007). Methyl acetate hydrolysis in a reactive divided wall column, Chem. Eng. Res. Des., 85, 149 154. Schultz, M. A., Stewart, D. G., Harris, J. M., Rosenblum, S. P., Shakur, M. S., and O Brien, D. E. (2002). Reduce costs with dividing-wall columns, Chem. Eng. Prog., 98(5), 64 71. Sotudeh, N., and Shahraki, B. H. (2007). A method for the design of divided wall columns, Chem. Eng. Technol., 30, 1284 1291. Taylor, R. and Krishna, R. (2000). Modeling reactive distillation, Chem. Eng. Sci., 55, 5183 5229. Taylor, R., Krishna, R., and Kooijman, H. (2003). Real-world modeling of distillation, Chem. Eng. Prog., 99, 28 39. Wang, S. J., and Wong, D. S. H. (2007). Controllability and energy efficiency of a high-purity divided wall column, Chem. Eng. Sci., 62, 1010 1025. Wang, S. J., Lee, C. J., Jang, S. S., and Shieh, S. S. (2008). Plant-wide design and control of acetic acid dehydration system via heterogeneous azeotropic distillation and divided wall distillation, J. Process Control, 18, 45 60.