Design and simulation of ethane recovery process in an extractive

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

Design and simulation of ethane recovery process in an extractivedividing wall column

Yadollah Tavan a,*, Shahrokh Shahhosseini b, Seyyed Hossein Hosseini c

aNational Iranian Gas Company (NIGC), Tehran, Iranb Process Simulation and Control Research Laboratory, School of Chemical Engineering, Iran University of Science and Technology, P.O. Box 16765-163,Narmak, Tehran, IrancChemical Engineering Department, Faculty of Engineering, Ilam University, 69315-516 Ilam, Iran

a r t i c l e i n f o

Article history:Received 9 May 2013Received in revised form1 March 2014Accepted 3 March 2014Available online 14 March 2014

Keywords:Dividing wall columnDistillationSeparationSimulationAzeotropic processDesign

a b s t r a c t

Separation of CO2 from hydrocarbons in the natural gas is complicated due to the existence of anazeotrope between ethane and CO2 at the cryogenic temperatures. The key issues to break this azeotropeare high investment costs for the unit equipments and the associated high energy requirements.Accordingly, an innovative process based on the dividing-wall column (DWC) technology is designedusing short-cut methods and relevant rigorous simulations. The energy demand and some environ-mental factors such as CO2 removal efficiency and CO2 emission reduction are studied for the conven-tional and DWC processes. It is found that the process including DWC is a better choice than theconventional one from economical and environmental point of views. Remarkably, this technology re-duces the energy demand up to 51.6%.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The increase in global energy demand has led to widespreadinvestigations on alternative sources of primary energy even at themost remote areas of the earth. Natural gas is the most sought, afterliquid fuel source, due to its cleaner combustion and less flue gasemission into the atmosphere (Alfadala and Al-Musleh, 2009).Natural gas contains impurities such as carbon dioxide, hydrogensulfide, carbon disulfide, mercaptans and sometimes traces ofcarbonyl sulfide. The removal of acid gases, H2S and CO2 from gasstream is essential due to environmental, operational and healthreasons (Maddox, 1982). Generally, the acid gas pipeline specifi-cations are 4.0 ppmH2S and 2 vol% of CO2 with the dew point of lessthan 263 K at 4500 kPa. Since H2S is extremely corrosive and toxic,it is removed from the gas before its consumption. Apart frommeeting customer’s contract specifications and successful lique-faction process, removal of CO2 from natural gas at high pressurehas currently become a global issue (Tavan and Hosseini, 2013a).Despite several researches done on CO2 capturing in chemical

processes (Sun and Smith, 2013; Harkin et al., 2010; Câmara et al.,2013), the existence of the minimum boiling CO2-ethane azeo-trope in natural gas process could causes certain problems. Highconcentrations of carbon dioxide in natural gas occur when carbondioxide is used for enhanced oil recovery. An azeotrope betweenethane (C2H6) and CO2 complicates separation of CO2 from naturalgas. Accordingly, using natural gas liquid (NGL) as extractivecomponent, Lastari et al. (2012) proposed low temperature distil-lation process in a series of distillation columns. The system gen-erates high pressure CO2, pure ethane and some amounts of NGL.However, the conventional extractive distillation process typicallyincludes two serial distillation columns. The main disadvantage ofthis separation process is its high capital investment and highamount of energy required to fulfill the desired purification.Therefore, to overcome this drawback, advanced intensification andintegration process techniques such as thermally coupled distilla-tion columns, dividing-wall columns (DWC), heat-integrateddistillation columns and reactive distillation (RD) were employed(Yildirim et al., 2011). In a DWC, themiddle section of a single vesselis split into two sections by inserting a vertical wall into anappropriate position of the column (Bravo-Bravo et al., 2010;Gutiérrez-Guerra et al., 2009). The DWCs have attracted moreattention in the chemical industries recently due to separation of

* Corresponding author. Tel./fax: þ98 21 88912525.E-mail address: [email protected] (Y. Tavan).

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier .com/locate/ jc lepro

http://dx.doi.org/10.1016/j.jclepro.2014.03.0150959-6526/� 2014 Elsevier Ltd. All rights reserved.

Journal of Cleaner Production 72 (2014) 222e229

Author's personal copy

more components in one single distillation unit, thereby savingboth energy and capital costs. The theoretical studies have shownthat DWCs could lead to at least 30% reduction in energy costscompared to conventional schemes (Sangal et al., 2012; Gómez-Castro et al., 2011). It is notable that DWC technology is notlimited to ternary separations; it can also be used in azeotropicseparations, extractive distillation, and reactive distillation (Kissand Suszwalak, 2012a).

Although several researches have conducted about extractivedividing-wall columns (Ignat and Kiss, 2012; Sangal et al., 2014;Wuet al., 2013), the process being studied is different from conven-tional extractive distillation columns. In conventional extractivecolumns a third component is added to the system and solvent lossin the product streams requires a make-up stream; while in thepresent study, the solvent is a mixture of propane and heaviercomponents (NGL) in which the solvent stream is quite similar tothe light key (ethane). These distinct features of the process leads tosome convergence problems. Furthermore, the present study hassome advantages such as existence of no water in the solventstream and accordingly non-corrosive behavior of the solvent ascompared with conventional extractive processes. In addition, inorder to reduce the energy requirements and number of trays in theextractive distillation process of separating the CO2/ethane azeo-trope, possibility of using DWC is examined by HYSYS3.1 (www.aspentech.com) for the first time using top-wall configuration.Furthermore, the rates of the interconnecting streams are opti-mized in order to minimize energy requirement of the process.Eventually, energy requirements and some environmental param-eters of the novel DWC (improved) process and the conventionalone are compared with each other to find amore beneficial process.

2. Simulation

2.1. Thermodynamic analysis of the extractive column

Several strategies have been used in industries in order toseparate the azeotropic mixtures. Some of them require addition ofa third chemical component for shifting the vaporeliquid equilib-rium such as extractive distillation, which uses a higher boilingsolvent and azeotropic heterogeneous distillation for entrainingchemical component. Another method for breaking azeotropes,which does not require addition of a third component, is pressureswing azeotropic distillation, wherein two columns operate at twodifferent pressures (Doherty and Malone, 2001; Luyben, 2013). Thethermodynamic analysis should be done prior to choosing the bestmethod for separating the azeotropes. For this purpose, a systemcontaining CO2/ethane and n-pentane as an agent of extraction isconsidered. CO2 and ethane are dissimilar molecules and havedifferent boiling points of �78 and �88 �C and molecular weightsof 44 and 30 g/mol, respectively. These molecules have a strongrepulsion towards each other which leads to existence of a mini-mum boiling azeotrope in the system as shown in Fig. 1(a, b). Fromthe figure, the azeotrope compositions are 0.67 and 0.64 at 2400and 1500 kPa, respectively. Therefore, the relative volatility of CO2-ethane azeotrope does not significantly change with pressure andconfirms that the pressure swing distillation is not appropriate forthe present study. Fig. 1b clearly indicates that the phase envelopeof the system drastically changes with n-pentane mole fraction.Furthermore, addition of n-pentane decreases CO2 freezing tem-perature to�75.9 �C (preventing formation of solid CO2). Therefore,using extraction distillation is a promising process for the presentsystem. The residue curve of CO2/ethane/n-pentane can be helpfulin elucidating the distillation part. Accordingly, the residue curvemap of the mixture at the pressure of 2400 kPa is illustrated inFig. 2. As shown in Fig. 2, the ternary mixture presents a single

Fig. 1. The properties of CO2-ethane azeotrope process in terms of (a) binary diagramand (b) phase envelope.

Fig. 2. The residue curve map of CO2, ethane and normal pentane.

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229 223

Author's personal copy

binary azeotrope at �19.11 �C with CO2 (mole fraction of 0.67) andethane without any liquid phase splitting. This figure also showsthe residue curve lines from the lowest azeotrope boiling point upto the highest one (n-pentane). The results indicate that extractivedistillation can be used in presence of a miscible mixture and ahomogeneous azeotrope. Moreover, this azeotrope limits theattainable separation in a single column and addition of the secondcolumn is mandatory in order to cross the azeotrope.

2.2. Simulation of extractive process

Fig. 3a illustrates a direct sequence of extractive distillationcolumns (conventional scheme) used for separation of CO2 fromethane. The process contains two distillation columns, two con-densers and two reboilers. In the first column, a fraction of NGLstream is served as high entrainer which preferentially carriesethane to the next column. The CO2 and ethane streams (mainproducts) are drawn from the top of the columns with composi-tions of 95 and 99.9 mol %, respectively. The bottom stream of thesecond column that contains higher hydrocarbons is divided intotwo parts, one of which is pumped back into the first column forbreaking the azeotrope. For simulation of the conventional processillustrated in Fig. 3a, two “Set” blocks are added to the system inorder to control the solvent/feed flow ratio (S/F) and rate of recy-cling stream. The conventional process is simulated by HYSYS 3.1with Peng-Robinson property model for the vaporeliquid equilib-rium of the system (Torres-Ortega et al., 2013). The two serial col-umns containing 50 ideal trays operate at 2400 kPa with a 100 kPapressure drop. The number of stages, optimum inlet stages and feedrates are extracted from the Lastari et al. (2012) study. The runspecifications of CO2 purity and the temperature of the condenserin the top stream of the first column are 95% and �14 �C, respec-tively. Furthermore, the condenser temperature of �5.5 �C andethane purity more than 99% are used in the top stream of thesecond column. The input data and simulation results are listed inTables 1 and 2.

The process under study is a challenging and complex simula-tion case and some considerations are essential during its designand optimization. Fig. 4 shows the temperature and compositionprofiles of CO2/ethane/propane along with the first and the seconddistillation columns. This figure shows that the required specifi-cations are satisfied by themodel. The concentrations of ethane andCO2 increase oppositely along the first column. In addition, theconcentration of CO2 is negligible in the second column and theNGL stream contains lower amounts of CO2 and ethane.

Fig. 3. The diagrams of (a) extractive distillation and (b) the dividing wall column(improved) processes: (symbols used in Fig. 3b: RCY: “Recycle” operator; SET: “Set”operator; E: Heat exchanger; TEE: Flow splitter; 1, 2, 3, 4, Vap and Liqq: Materialstreams; Qa, Qcc, n and Qr: Energy streams).

Table 1Input data and simulation results of the conventional and DWC systems.

Stream Feed Solvent Conventional DWC

CO2 Ethane NGL CO2 Ethane NGL Vap Liqq

PropertyFlow rate (mol/s) 3800 3024 1288 1720 792.1 1280 1772 760 3033 8545Pressure (kPa) 2415 2410 2400 2400 2600 2400 2400 2600 2500 2500Temperature (�C) 30 40 �13.05 �5.53 96 �13.38 �5.99 90.07 33.01 30.36Mol%CO2 32.25 0 94.96 0.14 0 95.42 0.21 0 0 0Ethane 46.23 0.50 3.30 99.66 0.50 0.14 99.01 0 83.3 51.29Propane 7.53 32.63 1.74 0.2 0 32.86 3.86 0.17 30.89 8.13 15.24i-C4 7.47 35.93 0 0 35.84 0.58 0 36.43 5.89 18.00n-C4 3.29 15.46 0 0 15.79 0 0 16.54 1.90 8.00i-C5 2.09 10.3 0 0 10.03 0 0 10.39 0.58 5.00n-C5 1.1 4 5.18 0 0 4.98 0 0 5.75 0.20 2.47

Table 2Comparison of energy demands for the conventional and DWC systems.

Equipment Conventional DWC

Column1 Column2 DWC Main column

PropertyTotal trays 50 50 20 40Condenser duty (kW) 95,782 181,471 18,096 88,981Reboiler duty (kW) 592,44 197,272 e 150,855Total duty (kW) 155,026 378,742 18,096 239,835

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229224

Author's personal copy

3. Results and discussion

3.1. Simulation of extractive DWC unit

A flow diagram of extractive DWC that is consisted of twocondensers, a reboiler and a main column containing sieve trays, isshown in Fig. 3b. Such configuration for a DWC has previously beendescribed by the researchers (Yildirim et al., 2011; Kiss andSuszwalak, 2012a,b; Tavan and Hosseini, 2013b) and usuallycalled top-wall configuration. In the DWC system, the solvent isseparated as a single bottom product, while two other distillateproducts (CO2 and ethane) are extracted from each side of the maincolumn. Since there is no off-the-shelf DWC unit in the currentcommercial process simulators, two coupled columns are used inHYSYS 3.1, as thermodynamically equivalent of the extractive DWC.The liquid coming out of the bottom of the prefractionator (PF) isfed into the main column (MC), while a vapor stream is withdrawnfrom the MC and fed into the bottom of the PF (also known as thevapor split). Two distinct routes are used for decomposing the DWCsystem into shortcut columns as shown in Fig. 5(a, b). These figuresare self-explanatory and the results reveal that the stages numberof 20 and 40 are accounted for the PF and MC, respectively.Therewith, tray rating feature of HYSYS is used to calculate pressuredrops in the columns. The “Vap” stream is drawn from the 20thstage of theMC column and introduced to the bottom tray of the PF.The bottom stream of the PF column is also sent to the 20th stage ofthe MC. Since the PF section has a condenser, CO2 purity in thedistillate stream (95%) is used as input data. Moreover, the MC has 3degrees of freedom, therefore, the condenser temperature of �6 �Cand ethane purity in the distillate and bottom stream of 99.99 and0.1%, respectively, are used as the input parameters for easierconvergence of the solution. It is worth mentioning that the results

of the base simulation and shortcut design are used as the initialestimates for the feed tray locations and vapor splitting. In thisstudy, the design steps and implementation of the DWC are similarto the conventional process. The detailed information about theDWC simulation can be found elsewhere (Premkumar andRangaiah, 2009). The simulation results of the DWC unit are alsolisted in Tables 1 and 2.

Fig. 6(a, b) shows the composition profiles in both sides of thewall. As can be seen in the figure, high purities of the valuableproducts are obtained in the extractive DWC. Furthermore, thereare large amounts of CO2 and propane in the MC and the ethaneconcentration increases gradually along the column. In the PF col-umn, concentration of ethane increases, while CO2 concentrationdecreases along the first column and the same trend can beobserved in extractive sequence columns. A comparison betweenthe conventional scheme, which includes two sequence columnsand the proposed DWC, is made. It is found that energy saving of13.7% is possible by using DWC unit. This value is roughly equiva-lent to the corresponding term obtained by Kiss and Suszwalak(2012a) for bioethanol dehydration process. It should be notedthat 13.7% energy saving is considerably low due to the fact that theoptimized parameters of the conventional scheme proposed byLastari et al. (2012) was used in this study and the DWC system isnot in its optimized state, yet. Consequently, the DWC system isoptimized in the next section in order to make a fair comparison.

3.2. Optimization of extractive DWC unit

Hereafter, it is assumed that energy optimization of a DWC withfixed number of trays is equivalent to energy minimization of thewhole DWC process since the energy consumption strongly de-pends on the interconnection between the liquid and vapor rates

Fig. 4. The temperature and composition profiles of CO2/ethane/propane along the (a)first and (b) second distillation columns.

Fig. 5. Decomposing of the dividing wall column system into shortcut columns basedon (a) removal of ethane and (b) removal of CO2: (symbols used in figures: QC1, QC2and QC3: Energy streams for condensers; Qr1, Qr2 and Qr3: Energy streams forreboilers; Up-MC, Down-MC and PF: Shortcut columns).

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229 225

Author's personal copy

(Delgado-Delgado et al., 2012). Hence, optimization of the inter-connecting streams leads to an optimal design. In the present study,the effect of the interconnecting stream rates, “Vap” and “Solvent”streams, on the energy demands are investigated by fixing thenumber of ideal stages for the DWC unit. Since the “solvent” and“Vap” streams have significant effects on the composition profilesand duties in the extractive DWC, sensitivity analysis must be car-ried out first in order to determine initial values for the “solvent”and “Vap” streams (Xia et al., 2012). The results of sensitivityanalysis are used for optimization by HYSYS process software af-terwards. Furthermore, it is tried to keep the final product streamsat the industrial condition (CO2> 95mol % and ethane>99mol % inthe top streams).

Fig. 7a shows the influence of interconnecting stream (“Vap”flow rate) on the total energy requirement that is summation of allduties of the reboiler and condensers, for the DWC unit. This figureshows that flow rate of “Vap” stream has a critical impact on theprocess duty; consequently, “Vap” stream is an important param-eter in DWC design that should be optimized. As can be seen inFig. 7a, increasing “Vap” flow rate leads to a subtle decrease in totalduty till the flow rate reaches up to 4000 mol/s, which attributes toreduction in the condenser duty of the PF. By increasing the flowrate of “Vap” stream above the minimum point (4000 mol/s), en-ergy consumption increases significantly due to the increase inreflux ratio of the MC in order to attain the desired purification ofethane in the product stream, as shown in Fig. 7b. Fig. 7b also showsthat by increasing the “Vap” rate, the mole fraction of CO2 in the topproduct of PF column decreases from 0.97 to 0.84. This confirmsthat major part of the CO2 content appears in theMC. Therefore, theoptimum flow rate of “Vap” stream needed to attain a desired CO2purity in the system is 3000 mol/s; hence the total energy demandis minimized.

Fig. 8 shows the influence of the solvent flow rate on total en-ergy requirement of the DWC unit by fixing stage numbers of the

columns and “Vap” flow rate (3000 mol/s). It is found that thesolvent flow rate is a crucial design parameter that should also beoptimized. From Fig. 8a, when the solvent rate is increased from300 up to 6000 mol/s, duties of the units increase due toenhancement of the columns stream rates.When the solvent rate isincreased, concentrations of CO2 and ethane in the product streamsrise exponentially in the system and consequently, required refluxratio decreases as depicted in Fig. 8b. As a result, addition of solventto the system exhibits an advantage (reduction in reflux ratio) and adisadvantage (increase in duties) simultaneously. Therefore, thesolvent amount should be determined optimally not only todecrease both duty and reflux ratio, but also to achieve the indus-trial specification of the product streams. In the solvent rate of3000e3300 mol/s, all desired specification of the product streams(CO2 > 95 mol % and ethane> 99 mol % in the top streams) aresatisfied and with further increase in the solvent rate, more energyis required due to increase of the column rates. Accordingly, theoptimum rate of the solvent stream is found to be 3024 mol/s.

Another simulation was carried out using optimal values ofinterconnecting streams (solvent rate of 3024 mol/s and “Vap” flowrate of 3000 mol/s) and the energy results are listed in Table 2. Itshould be note that in order to minimize energy requirements inthe whole DWC unit, the sequential quadratic programming (SQP)method implemented in HYSYS is also used to confirm the opti-mized results tabulated in Table 2. The target is minimization ofenergy demand using the interconnecting flows as search variableswhile assuming the pure CO2 and ethane at the top of the columnsas a constraint. During performance tests of this method, thenumber of trays remained constant. The optimal result of SQPmethod clearly confirms earlier findings presented in Table 2.Moreover, the mole balance of the components in the systems ispresented in Table 3. It can be inferred from Table 2 that 51.6%

Fig. 6. The results of the preliminary design of dividing wall column system in terms of(a) composition profiles of CO2/ethane/propane along the PF and (b) the MC.

Fig. 7. Finding the optimal rate of “Vap” stream using (a) energy demand and (b)product specification.

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229226

Author's personal copy

reduction in energy demand is possible for the optimal design ofthe extractive DWC configuration compared with the optimaldesign of conventional extractive distillation columns. This is in afair agreement with the findings of other researchers of DWCschemes (Yildirim et al., 2011; Delgado-Delgado et al., 2012; Pre-mkumar and Rangaiah, 2009).

It is worth mentioning that the Petlyuk column is generallymore efficient than other thermally coupled schemes and a DWC ispractically identical to a Petlyuk column if the heat transfer acrossthe column wall is neglected or the wall is insulated. Accordingly,Fig. 9 that shows the temperature difference between both sides ofthe wall is prepared. As can be seen in this figure, the difference isless than 20 �C. Therefore, it can be assumed that there is no heattransfer between two sides of the wall and the modeled system isthermodynamically equivalent to the Petlyuk column. Moreover,Fig. 10 shows the composition profiles in both sides of the wall. Thefigure shows that products with high purity are obtainable in theDWC configuration. As a main result, using a DWC is an interestingand suitable choice in extractive CO2-ethane process for the goal ofreducing operating and investment costs.

3.3. Comparison between alternatives

In order to make further comparison between the conventionaland DWC processes, in addition to energy demand, environmentalfactors such as efficiency of CO2 removal and CO2 emission reduc-tion are investigated in the present research. Subsequently, resultsof energy demand and environmental impacts analyses arecompared with each other.

3.3.1. CO2 removal efficiencyRemoval efficiency (h) is defined as percentage of CO2 in the gas

stream that is removed during absorption operation. It should benoted that removal efficiency implies absorption performance ofthe system (Tavan et al., 2014). The removal efficiency for CO2 issimply determined from the difference between the amounts ofCO2 entering the column and the corresponding term leaving eachstage of the column, which can be expressed by the followingequation:

h ¼"1�

�YCO2

;nth1� YCO2

;nth

� 1� YCO2;in

YCO2;in

!#(1)

where yCO2;in and yCO2 ;nth stand for mole fractions of gas-phase CO2entering to the column and leaving from each tray, respectively.Fig. 11a displays CO2 removal performance for the introduced sys-tems. In addition, Fig.11b also shows distribution of overall removalefficiency of the processes. It is clearly shown in Fig. 11a thatremoval efficiency of the conventional process is quite higher thanthe DWC in themiddle trays of the absorption column. For the DWCprocess, higher removal efficiencies are observed in the lowersection of the absorption column. Additionally, equal removal ef-ficiencies are observed in the upper section of the absorption col-umn, which indicates the same performances of the processes.

3.3.2. Estimation of CO2 emission reductionDuring fuel combustion, air is assumed to be in excess to ensure

complete combustion, so that no carbon monoxide is formed. Theamount of emitted CO2, [CO2] Emiss (kg/s), is related to the energyequivalent of the fuel, Q Fuel (kW), in the heating device, as follows(Gadalla et al., 2005; Tavan et al., 2014):

½CO2�Emiss ¼�QFuelNHV

��C%100

�a (2)

where a (¼3.67) is the ratio of the molar masses of CO2 and C, whileNHV, which is equal to 39,771 (kJ/kg), stands for net heating value

Fig. 8. Finding the optimal rate of “Solvent” stream using (a) energy demand and (b)product specification.

Table 3Mole balance of the components in the proposed systems.

Component CO2 Ethane Propane i-C4 n-C4

PropertyConventional processInput (mol/s) 1225.50 1756.74 286.14 283.86 125.02Output (mol/s) 1225.49 1756.66 286.14 283.89 125.07DWC processInput (mol/s) 1225.50 1756.74 286.14 283.86 125.02Output (mol/s) 1225.17 1756.25 286.04 284.29 125.07

Fig. 9. Temperature profile of the dividing wall column system.

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229 227

Author's personal copy

of heavy oil fuel with a carbon content of 86.5%.The flame tem-perature of a boiler is lower than the flame temperature of afurnace, because combustion heat is removed immediately to thesteam. However, the same theoretical flame temperature of 1800 �Cmay still be used. A stack temperature of 160 �C is also used in thecalculations. The energy equivalent of fuel can be calculated asfollows.

QFuel ¼QProc

lProcðhProc � 419Þ TFTB � TO

TFTB � Tstack(3)

where lProc (kJ/kg) and hProc (kJ/kg) are the specific latent heat andspecific enthalpy of steam delivered to the process, respectively,and TTFB (�C) is the flame temperature of the boiler flue gases. Theabove equation is obtained from the simple steam balance aroundthe boiler required to relate the energy equivalent of fuel in theboiler to provide a heat duty of QProc. The boiler feed water isassumed to be at 100 �C with a specific enthalpy of 419 kJ/kg. FromFig.11b, it is evident that when the base case (conventional process)emits one unit of CO2 (kg/s) to the environment, the improvedprocess (DWC) emits much lower CO2, by 0.59. Therefore, it can beseen that 41% reduction in carbon emission is possiblewith the newprocess and it comes as no surprise that the DWC alternative is inthe pole position with the lower carbon footprint.

3.4. Final comparison

By comparing energy demand of the conventional and DWCprocesses, it is evident that the DWC process reduces energy de-mand by 51.6%. In addition, the novel proposed DWC needs lownumber of trays compared to the conventional process. Based on

these findings and data about CO2 emission reduction and removalefficiencies, it is concluded that the novel DWC process can beconsidered as a serious alternative candidate for the CO2-ethaneazeotropic process.

4. Conclusion

In the current research, the use of DWC for extractive CO2-ethane azeotropic process is demonstrated through rigorous sim-ulations. It is found that the rates of interconnecting streams have asignificant impact on total energy demand and column specifica-tions. Therefore, in order to make a fair comparison with the con-ventional sequence distillation columns, the optimal values ofinterconnecting rates are determined based on sensitivity analysisand the SQP method. The simulation results indicated that byraising the Vap rate, the mole fraction of CO2 in the top product ofDWC column decreases from 0.97 to 0.84 and a high amount of CO2appears in the main column. In addition, escalation of “Vap” flowrate leads to enhancement in energy consumption rate due tohigher reflux ratio. Accordingly, the simulation results show thatthe optimum rate of “Vap” stream is 3000mol/s. The addition of thesolvent to the system exhibits an advantage (reduction of refluxratio) and a disadvantage (increase in duties) at the same time.Therefore, the solvent amount is determined not only to decreaseboth duty and reflux ratio, but also to achieve the industrial spec-ifications of the product streams. The results indicate that optimalvalue of the solvent rate is 3024mol/s. The results clearly show thatDWC process is feasible and the novel proposed DWC reduces en-ergy demand by 51.6% and carbon footprint by 41%. In addition, thenovel proposed DWC needs low number of trays compared with

Fig. 10. Optimal results of the improved system in terms of (a) temperature profile and(b) composition profiles.

Fig. 11. Comparison between the proposed processes in terms of (a) CO2 removal ef-ficiency and (b) environmental effects.

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229228

Author's personal copy

conventional process. Therefore, the use of DWC for extractive CO2-ethane azeotropic process is an interesting and suitable choice.

Acknowledgment

Useful comments from anonymous reviewers and the subjecteditor are acknowledgedwhich led to improvements in the originalversion of the paper.

References

Alfadala, H.E., Al-Musleh, E., 2009. Proceedings of the 1st Annual Gas ProcessingSymposium, Qatar.

Bravo-Bravo, C., Segovia-Hernández, J.G., Gutiérrez-Antonio, C., Duran, A.L., Bonilla-Petriciolet, A., Briones-Ramírez, A., 2010. Extractive dividing wall column:design and optimization. Ind. Eng. Chem. Res. 49, 3672e3688.

Câmara, G., Andrade, C., Júnior, A.S., Rocha, P., 2013. Storage of carbon dioxide ingeological reservoirs: is it a cleaner technology? J. Clean. Prod. 47, 52e60.

Delgado-Delgado, R., Hernández, S., Barroso-Muñoz, F.O., Segovia-Hernández, J.G.,Castro-Montoy, A.J., 2012. From simulation studies to experimental tests in areactive dividing wall distillation column. Chem. Eng. Res. Des. 90, 855e862.

Doherty, M.F., Malone, M.F., 2001. Conceptual Design of Distillation Systems.McGraw-Hill, New York, United States.

Gadalla, M.A., Olujic, Z., Jansens, P.J., Jobson, M., Smith, R., 2005. Reducing CO2emissions and energy consumption of heat-integrated distillation systems.Environ. Sci. Technol. 39, 6860e6870.

Gómez-Castro, F.I., Rodríguez-Ángeles, M.A., Segovia-Hernández, J.G., Gutiérrez-Antonio, C., Briones-Ramírez, A., 2011. Optimal design of multiple dividing wallcolumns. Chem. Eng. Technol. 34, 2051e2058.

Gutiérrez-Guerra, R., Segovia-Hernández, J.G., Hernández, S., 2009. Reducing energyconsumption and CO2 emissions in extractive distillation. Chem. Eng. Res. Des.87, 145e152.

Harkin, T., Hoadley, A., Hooper, B., 2010. Reducing the energy penalty of CO2 captureand compression using pinch analysis. J. Clean. Prod. 18, 857e866.

Ignat, R.M., Kiss, A.A., 2012. Integrated bioethanol separation and dehydration in anovel extractive DWC. Chem. Eng. Trans. 29, 619e624.

Kiss, A.A., Suszwalak, D.J.-P.C., 2012a. Enhanced bioethanol dehydration by extrac-tive and azeotropic distillation in dividing-wall columns. Sep. Purif. Technol. 86,70e78.

Kiss, A.A., Suszwalak, D.J.-P.C., 2012b. Innovative dimethyl ether synthesis in areactive dividing-wall column. Comput. Chem. Eng. 38, 74e81.

Lastari, F., Pareek, V., Trebble, M., Tade, M.O., Chinn, D., Tsai, N., Chan, K.I., 2012.Extractive distillation for CO2-ethane azeotrope separation. Chem. Eng. Process52, 155e161.

Luyben, W.L., 2013. Comparison of extractive distillation and pressure-swingdistillation for acetone/chloroform separation. Comput. Chem. Eng. 50, 1e7.

Maddox, R.N., 1982. In: Gas conditioning and processing, vol. 4. Campbell PetroleumSeries, Oklahoma, United States.

Premkumar, R., Rangaiah, G.P., 2009. Retrofitting conventional column systems todividing-Wall Columns. Chem. Eng. Res. Des. 87, 47e60.

Sangal, V.K., Kumar, V., Mishra, I.M., 2012. Optimization of structural and opera-tional variables for energy efficiency of a divided wall distillation column.Comput. Chem. Eng. 40, 33e40.

Sangal, V.K., Kumar, V., Mishra, I.M., 2014. Process parametric optimization of adivided wall distillation column. Chem. Eng. Commun. 201, 72e87.

Sun, L., Smith, R., 2013. Rectisol wash process simulation and analysis. J. Clean. Prod.39, 321e328.

Tavan, Y., Hosseini, S.H., 2013a. A novel application of reactive absorption to breakthe CO2-ethane azeotrope with low energy requirement. Energy. Convers.Manage 75, 407e417.

Tavan, Y., Hosseini, S.H., 2013b. From laboratory experiments to simulation studiesof methanol dehydration to produce dimethyl ether e part II: separation andsimulation. Chem. Eng. Process 73, 151e157.

Tavan, Y., Shahhosseini, S., Hosseini, S.H., 2014. Feed-splitting technique in theextractive distillation of CO2-ethane azeotropic process. Sep. Purif. Technol. 122,47e53.

Torres-Ortega, C.E., Segovia-Hernandez, J.G., Gomez-Castro, F.I., Hernandez, S.,Bonilla-Petriciolet, A., Rong, B.G., Errico, M., 2013. Design, optimization andcontrollability of an alternative process based on extractive distillation for anethane-carbon dioxide mixture. Chem. Eng. Process 74, 55e68. http://dx.doi.org/doi:10.1016/j.cep.2013.09.011.

Wu, Y.C., Hen-Chia, P., Chien, I.-L., 2013. Critical assessment of the energy-savingpotential of an extractive dividing-wall column. Ind. Eng. Chem. Res. 52,5384e5399. Www.aspentech.com (25 01 14.).

Xia, M., Yu, B., Wang, Q., Jiao, H., Xu, C., 2012. Design and control of extractivedividing-wall column for separating methylal�methanol mixture. Ind. Eng.Chem. Res. 51, 16016e16033.

Yildirim, Ö., Kiss, A.A., Kenig, E.Y., 2011. Dividing wall columns in chemical processindustry: a review on current activities. Sep. Purif. Technol. 80, 403e417.

Y. Tavan et al. / Journal of Cleaner Production 72 (2014) 222e229 229