lable at ScienceDirect

Journal of Cleaner Production 239 (2019) 118029

Contents lists avai

Journal of Cleaner Production

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

Self-recuperative high temperature co-electrolysis-based methanolproduction with vortex search-based exergy efficiency enhancement

Yus Donald Chaniago a, 1, Muhammad Abdul Qyyum a, 1, Riezqa Andika b, Wahid Ali c,Kinza Qadeer a, Moonyong Lee a, *

a School of Chemical Engineering, College of Engineering, Yeungnam University, Gyeongsan, Republic of Koreab Department of Chemical Engineering, Faculty of Engineering, Universitas Indonesia, Depok, 16424, Indonesiac Department of Chemical Engineering and Technology, Jazan University, Jazan, 45971, Saudi Arabia

a r t i c l e i n f o

Article history:Received 12 March 2019Received in revised form9 August 2019Accepted 12 August 2019Available online 13 August 2019

Handling editor: Giorgio Besagni

Keywords:Solid oxide electrolyzer cellCo-electrolysisSelf-heat recuperationVortex searchExergy analysisMethanol

* Corresponding author.E-mail address: mynlee@yu.ac.kr (M. Lee).

1 These authors contributed equally to this work.

https://doi.org/10.1016/j.jclepro.2019.1180290959-6526/© 2019 Elsevier Ltd. All rights reserved.

a b s t r a c t

The reduction of greenhouse gas emission via the transformation of carbon dioxide into methanol resultsin several secondary benefits including the production of a valuable by-product that can be used forenergy storage and as a fuel source. As such, this is a promising approach for mitigating climate change.Methanol production via the co-electrolysis process using solid oxide electrolyzer cells is an efficacioussolution to the issue of excess electricity storage in the context of renewable energy and carbon dioxideutilization. However, this process is an energy-intensive and temperature-sensitive method, mainly dueto the requirement of high-temperature electrolysis. In this context, this study investigates and evaluatesthe potential for overall performance improvement by minimizing energy consumption and increasingmethanol production using self-heat recuperation technology. The newly developed vortex searchstrategy was employed to achieve the maximum potential benefit from retrofitted recuperators. Detailedexergy analysis was performed for the process and the evaluation of its performance. The findingsrevealed that the electrochemical system for co-electrolysis has the highest exergy destruction rate. Byemploying the vortex search approach, the exergy loss of the energy process system can be reduced by61.7% with a total reduction of the exergy loss of 15.9%, while improving methanol production anddecreasing distillation reboiler duty. The simple solution of self-recuperation with optimization that wasutilized in this study is a flexible approach that can be directly applied to the improvement of co-electrolysis and methanol synthesis.

© 2019 Elsevier Ltd. All rights reserved.

1. Introduction

Fossil fuel use has increased atmospheric carbon dioxide (CO2)emission levels in recent years, leading to problems related to theenvironment and public health including climate change and airpollution. This major issue of CO2 emission has resulted in theglobal monitoring of this gas (Witze, 2018). Moreover, there hasbeen substantial efforts to limit the carbon footprint of the world'spopulation by emphasizing the use of renewable energy such ashydropower (Bello et al., 2018), biogas energy (Haider et al., 2019),and solar photovoltaic systems (Parida et al., 2011) as powersources. In addition, other technologies have been exploited to

facilitate the improvement of CO2 electroreduction into useful fuels(Gao et al., 2016) and the transition to solar, wind, and geothermalenergy (Mekonnen et al., 2016). However, not all renewable energyalternatives results in a reduction of CO2 emissions. In the worstcase, a renewable energy fails to offset the emitted CO2, e.g., biofuelproduction (DeCicco et al., 2016). Renewable energy intermittencyis another drawback, e.g., in May 2016, in Germany, electricitysupply far exceeded demand and the price of electricity becamenegative, and consumers were paid for their electricity consump-tion (Andika et al., 2018). Thus, there is a need for systems that canuse excess electrical energy, in anticipation of positive renewablegrowth towards 2023 (Gielen et al., 2019). In addition, CO2 is anessential material for many industrial processes (Pierantozzi,2003). To address the aforementioned issues, several promisingsolutions have been proposed in which CO2 emissions are reducedand converted into useful products, energy storage and fuels, thatare both environmentally and economically beneficial

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 1180292

(Koytsoumpa et al., 2018). The most promising systems for CO2reduction involve the conversion of this gas into valuable products(Cu�ellar-Franca and Azapagic, 2015).

Among the various possible products, methanol (MeOH) is ofparticular interest because it can be utilized as a transport fuel(Verhelst et al., 2019). In addition, it can be further processed toproduce an alternative to conventional diesel via MeOH dehydra-tion (Sabour et al., 2014) or can be used as an energy carrier forhydrogen (H2) (Rosen and Scott, 1988), while also serving as achemical for electricity storage (Kauw et al., 2015). Moreover,MeOH is considered to be a versatile chemical intermediate that iswidely used in the manufacture of other chemicals and can be usedto produce a wide variety of products used in daily life (Bozzanoand Manenti, 2016). MeOH production from CO2 is a favorable,clean technique (Riaz et al., 2013) with the integration processfacilitated by accessing excess electricity (Kauw et al., 2015). Thisavoids the need for CO2 seclusionwhich is a very expensive process(Olah et al., 2009). To convert CO2 to MeOH that utilizes surpluselectrical energy, the power to the MeOH (PtM) concept caneffectively use the excess electricity and concurrently incorporatethe usage of CO2 as feedstock.

To realize the PtM concept, H2 production is required via aprocess that utilizes excess electricity. Co-electrolysis is aneconomically competitive process that is capable of achievingvarious functions including the production of H2 from an abundantsource (such as water) and implementation using various renew-able electric sources (Fu et al., 2010). Solid oxide electrolyzer cells(SOEC) are electrolyzers that can undergo a co-electrolysis process.In comparison with other electrolysis devices (polymer electrolytemembrane (PEM) electrolysis and alkaline water electrolysis), SOECis much more efficient because of their thermodynamic and kineticadvantages with respect to water splitting (Chen et al., 2014).Another advantage of SOEC in the PtM concept is that it is capableof producing carbon monoxide (CO). MeOH synthesis is conductedmainly with CO; thus, CO production is required in addition to H2and CO2. Furthermore, the reaction is 100 times faster and catalystdurability is enhanced in the presence of CO2 (Wender, 1996). Dueto these advantages, SOECs that can produce CO and H2 and con-trolling their composition could play a key role in MeOH synthesisfrom a CO2 source. In addition, the superior energy conversion ef-ficiency of SOEC is still under investigation (Im-orb et al., 2018).Widespread application of this technology will require significantprocess improvements together with performance analysis toensure that the demonstrated potential is realized. In addition tothese advantages, compared to PEM and alkaline water electrolysis,the SOEC-based co-electrolysis process requires considerably moreheating energy but has a lower electricity demand (Bhandari et al.,2014). This is because co-electrolysis process involves high-temperature processes between the energy source and the elec-trolysis process (Uhm and Kim, 2014). This requires additionalprocessing, which in turn is an issue that affects the realization ofthe process objectives. Thus, the biggest concerns are the use ofappropriate materials including the catalyst (Kim et al., 2018), andthe recovery of heat (G€otz et al., 2016). Another concern is theamount of energy generated from the MeOH synthesis processduring the high-temperature exothermic reaction that creates theneed for a heat recovery system.

Techno-economic and off-design analysis of SOEC (Reznicek andBraun, 2018) and dynamic temperature control of SOEC (Wanget al., 2019a) for co-electrolysis are currently under investigationstudied. However, these studies only consider standalone co-electrolysis processes and do not specifically target the composi-tion ratio in the co-electrolysis process product that is required forMeOH production in the PtM concept. Incorporating the co-electrolysis of CO2 and H2O into MeOH production is vital to

MeOH production from CO2 that involves excess electrical power.This is because attempts at co-electrolysis improvement willbenefit MeOH production as a synergistic effect. Another processfor MeOH production from CO2 via the co-electrolysis process wasevaluated by employing SOEC using a heat exchanger network(HEN) in the integrated process from CO2 capture to MeOH syn-thesis and utility (Al-Kalbani et al., 2016). However, this approach isdifficult to implement in simulation or practice due to the use ofnumerous heat exchangers. Further, it is difficult to analyze theimportant variables that affect MeOH production for further pro-cess improvement. Thus, a simpler and efficient process configu-ration is required for progressive improvement of co-electrolysisand MeOH integrated system configuration.

In contrast, the use of a recuperator (either for cooling orheating) represents a simpler design and can retrofit the targetprocess that requires energy recovery to increase process efficiency(Qyyum et al., 2019b). The larger the temperature difference be-tween the cold stream and the hot stream input of a recuperator,the higher is the recuperator's effectiveness (Cao et al., 2016).However, the following issues must be addressed for the effectiveuse of recuperators in a system that involves co-electrolysis andMeOH synthesis: (i) the recuperator design is affected by theminimum internal temperature approach (MITA). Thus, the targettemperature determination using the reasonable specific MITAwithin the recuperator process is essential. (ii) The product of co-electrolysis consequently influences the MeOH production pro-cess; thus, high-temperature recovery via the inclusion of addi-tional processing for efficiency improvement must satisfy thespecific composition ratio of the co-electrolysis product. The issueof finding the best solution for process variables can be resolved viathe appropriate implementation of optimization (Wang et al.,2019b). Process optimization (either design or operational) isconsidered as one of the most important and critical steps duringthe scaling up of or improvement of any process, with the aim ofachieving maximum performance at minimal cost.

The objective of this study is to design a simple efficient processvia the development and implementation of rigorous optimizationfor application towards the progressive improvement of co-electrolysis and the MeOH integrated system configuration,including process performance evaluation. In this regard, self-recuperation-assisted co-electrolysis and the MeOH synthesisprocess are introduced. They were simulated and analyzed usingAspen HYSYS V10 and rigorously optimized via a newly developedsingle-solution, metaheuristic, vortex search (VS) approach bylinking the MATLAB 2017b with HYSYS model. The single-solution-based VS algorithm was developed by Do�gan and Ӧlmez (Do�ganand €Olmez, 2015). The dominant characteristic of this algorithmis the adaptive interval (step) size phenomenon that significantlyimproves the functionality of the search mechanism. The VS algo-rithm program balances neighborhood weak and strong locality todetermine an optimal solution. Various optimization methods havebeen implemented with respect to renewable energy (Ba~nos et al.,2011). To the best of the authors’ knowledge, the VS algorithmprogram is the first approach to be implemented in co-electrolysisand MeOH production systems. To evaluate the retrofitted inte-grated process performance, exergy analysis, which is an appro-priate scientific concept for energy studies, can be applied toachieve sustainable development (Bilgen and Sarıkaya, 2015). Theanalysis provides good insight into the best approach for achievingsystem sustainability in an efficient manner.

This study is limited to energy recovery on pre- and post-hightemperature co-electrolysis and MeOH synthesis processes toachieve energy efficiency and MeOH production improvement in aPtM process that utilizes electricity from a cleaner source. Thus, adetailed model for MeOH synthesis is required to present the

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 3

optimization effect. However, establishing a complex SOEC modelis not the aim of this study because during intensification in a large-scale process, transforming a model from complex to simple hassignificant benefits in terms of its simulation (O'Brien et al., 2007).

2. Base case and proposed optimal retrofitting scheme

Given that the continuous process of the co-electrolysis andMeOH production system is still under development, there arenumerous unresolved challenges, which presents an opportunityfor the improvement and refinement of this approach. This workuses the configuration process (Al-Kalbani et al., 2016) inwhich theCO2 absorber is excluded by rearranging the recycle stream beforeentering the SOEC unit as a base case. Further, the base case isretrofitted with a recuperator on a pre and post SOEC and a MeOHreactor, followed by optimization in to achieve the maximum po-tential benefit. The performance comparison between the base andoptimal retrofitted cases was analyzed via exergy analysis. Thereare three main stages in the process of MeOH production fromcaptured CO2: CO and H2 production via co-electrolysis process thatutilizes SOEC (CO and H2 unit), MeOH synthesis, and MeOHpurification.

Fig. 1. Base case process with

The base case is presented in Fig.1. There are three key aspects ofthe process: CO and H2 unit dfrom stream 1 and 2 to stream 16;MeOH synthesisdfrom stream 17 until stream 30, including recy-cling at stream 34; MeOH purificationdfrom F-5 until C-1. In thisprocess, the CO2 feed stream or stream 2 was assumed to be 95mol% with water as the remainder at 1000 kmole =h, and was intro-duced to compressor K-1 to increase the desired operating pres-sure. The H2O stream was introduced to pump (P) with the samepressure as K-1. H2Owas then vaporized, using heater E-1, to obtainsaturated steam or vapor fraction¼ 1. These two feeds were com-bined in the mixer and then heated using heater E-2 until anappropriate temperature was achieved to maintain the water gasshift (WGS) reaction at approximately 800 �C, before entering theco-electrolysis process. In a simulator flowsheet, the co-electrolysisprocess that utilized SOEC is represented as Co-E. The base case ofthe Co-E duty without a recuperator (HX) was 206,175.79 kW.

The ratio of H2, CO, and CO2 (or SðH2=COÞ ¼ H2�CO2COþCO2

¼ 2), whichhas been characterized in several studies, was implemented in thisstudy. The initial split ratio, T-1, for the return stream to the co-electrolysis process was 0.2. We recycled H2O from stream to M-1and the H2O feed at stream 1was automatically adjusted to be 2191kmole =h to maintain the SðH2=COÞ ratio at stream 16. The SOEC

out a recuperation unit.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 1180294

product from stream 16 was then introduced to the MeOH reactor(R-1), via a series of compressors until a pressure of 70.93 bar wasachieved, with each compressor pressure ratio below 3. H2O is thenconsecutively removed using flash drums F-2, F3, and F4. Stream 26was combined with recycled stream 34 in a stream 34/stream 26ratio of 3.643. The Inlet stream 27 was fed at 265 �C (Al-Kalbaniet al., 2016) and the gases associated with stream 29 were cooledto 40 �C using cooler E-8. The purge gas from the reactor was set as1.00% mole/mole. The final process was MeOH purification. Pre-liminary purification of MeOHwas achieved using two simple flashdrums, F-5 and F-6, until the pressure was reduced to 2 bar. Sub-sequently, a distillation column (C-1) was used to purify MeOH inthe liquid phase of the top product at a reasonable MeOH streamproduct temperature.

SOEC implementation for co-electrolysis with the electro-chemical reduction reaction is energy-intensive and MeOH syn-thesis is performed at a high temperature. This requires additionalwork to realize efficiency improvement. To address these issues,two recuperators were installed, before Co-E and R-1, as shown inFig. 2. The H2O feed was pre-heated by E-1 before being vaporizedby recuperator-1 (HX-1). This is because the enthalpy associatedwith the Co-E output stream is insufficient to directly transformH2O to steam. The E-1 output vapor fraction composition variable,which represents E-1 enthalpy consumption, is affected by thetemperature variation in HX-1. Another issue is the determinationof the appropriate temperatures for the hot and cold stream out-puts from HX-1, within the optimal process constraints of MITA.

Fig. 2. Proposed self-heat recuperatio

After heating in HX-1, steam and CO2 were introduced to anexothermic reactor or WGS to maintain the desired temperaturebefore entering Co-E. For MeOH synthesis, before entering R-1, thestream was introduced to recuperator-2 (HX-2) to utilize the heatoutput from R-1 due to the moderately high temperature of thehighly exothermic reaction. Direct temperature adjustment in HX-2with specific MITA may be unfavorable to the conversion in R-1;thus, to maintain or improve the performance and efficiency of R-1,heater E-6 and cooler E-7 were used before and after R-1, respec-tively. The final stepwasMeOH purification using flash drums and adistillation column.

2.1. Base case high-temperature co-electrolysis process

A typical SOEC consists of a cathode, an anode, and an electro-lyte, as shown in Fig. 3. The electrolyte is an ion-conductive, denseceramic layer, sandwiched between a porous cathode and anode.The principles of SOEC used for electrolysis and co-electrolysis arepresented in Fig. 3(a) and b, respectively.

As a co-electrolyzer, SOEC converts CO2 and H2O into CO, H2, andO2, according to the following reactions at the cathode or fuelelectrode (Kim et al., 2015)

H2OðgÞ þ2e�/H2ðgÞ þ O2� DH800�C ¼ 151:8 kJ=mol (1)

CO2ðgÞ þ2e�/COðgÞ þ O2� DH800�C ¼ 185:9 kJ=mol (2)

n process for retrofitting scheme.

Fig. 3. Principles of electrolysis (a) and co-electrolysis (b) in a SOEC (Andika et al., 2018).

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 5

The O2�O ions are selectively transferred through the solid elec-

trolyte and combined to form O2 molecules at the anode triple-phase boundary (Zheng et al., 2017). An electrochemical reactionoccurs at both electrodes, and both the reverse water gas shift(RWGS) and WGS thermochemical reactions take place on thesurface of the electrode (Kim et al., 2015). Co-electrolysis is morecomplicated than the electrolysis H2O or CO2 due to the WGS andRWGS reaction (Zheng et al., 2017). The thermochemical reactionsof RWGS and WGS can be expressed as shown in Equation (3), andthe visualization of co-electrolysis on SOEC is shown in Fig. 4.

CO2ðgÞ þH2ðgÞ%COðgÞ þ H2OðgÞ DH800�C ¼ 36:82 kJ=mol

(3)

Compared to PEM and alkaline water electrolysis, high-temperature SOEC has a lower cell voltage of 0.95e1.3 V and itsthermodynamic and kinetic advantages lead to higher efficiencycompared to that of other devices at 82% LHV, or equal to 3.7 kWh/Nm3 (Andika et al., 2018).

This study does not focus on the complex rigorous SOEC model.Currently, reported SOEC models are complicated, especially whenintegrated with other process models. (Im-orb et al., 2018). To

Fig. 4. Co-electrolysis p

model Co-E into a simulation, a simple electrolyzer can be repre-sented by a one-dimensional model in HYSYS (KOH et al., 2010) andAspen Plus (Im-orb et al., 2018), which is performed by using aconversion or stoichiometry reactor. This reactor is represented byreactor model C in the HYSYSmodel. However, the thermochemicalRWGS andWGS reactions occur in parallel on the electrode catalyst(Ni) or cathode (Ni, 2012). Therefore, additional reaction modelingis necessary. This reaction can be validly modeled using a simpleequilibrium reactor on the cathode side (O'Brien et al., 2007). Toobtain the desired inlet temperature before entering the co-electrolysis process, an equilibrium reactor or reactor model Ethat supports the structure of the shift reaction is applied (McKellaret al., 2010). The scheme of the HYSYS models for Co-E with re-actions that occur in the SOEC is presented in Fig. 5. In the simu-lation, the Peng-Robinson thermodynamicmodel is used for the co-electrolysis process.

The temperature range also played a significant role. Theadvantage of a higher temperature is a trade-off with materialstability, and the favorable temperature range for co-electrolysis onSOEC has been reported to be 600e1000 �C (Zheng et al., 2017). Inthis study, the SOEC temperature was maintained at ~800 �C. In the

rocess (Ni, 2012).

Fig. 5. The Co-E represented using a one-dimensional co-electrolysis HYSYS model that is equivalent to the co-electrolysis process.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 1180296

SOEC, the sweep gas was used to remove O2 from the anode; asmore sweep gas is blown through the SOEC, the SOEC efficiency isreduced (Barelli et al., 2017). The amount of sweep gas was calcu-lated using the ratio between the SOEC feed and the sweep gas flowthat resulted in a CO2 and H2O conversion of 100% and 80%,

VNðTPÞ¼1

2FðTP � TRÞ�y2;O2

� y1;O2

�hy2;H2

ðTPÞ � y1;H2

i, ðTP

TR

ðy2;O2y1;O2

ðy2;H2 ðTPÞ

y1;H2

DGR;H2OðTÞ þ RT ln

1� yH2

� y0;CO2� yN2

yH2y0:5O2

!dyH2

dyO2dT (6)

respectively (McKellar et al., 2010). The ratio between the SOECfeed and the sweep gas flow was approximately 1.03� 109. In thebase case process, the cathode product or stream 10 was cooled byE-3 to 40 �C to transform H2O to the liquid phase. The separation ofH2O from the other gases took place in flash drum F-1, where thecomposition ratio of H2, CO, and CO2 controls the MeOH synthesisprocess.

The electricity consumption of the SOECs is represented by thepower usage of the reactor, which can be calculated using Equation(4). In this case, the total power supplied to the conversion reactorrepresents electricity consumption. The electrical power requiredby the SOEC unit is expressed in (4), where _W is the electrical po-wer supplied to the electrolyzer (W), Vop is the average cell voltageduring operation (V), I is the total current (A), i is the currentdensity (A/m2), and Acell is the total active area of the co-electrolysiscells in the SOEC stack (m2) (Fu et al., 2010).

_W ¼Vop I ¼ Vop i Acell (4)

Vop is written as a function of the current density ((5)), where VNis the mean Nernst potential of the operating cell (V) and ASR is themean area-specific resistance of the co-electrolyzer stack (U m2),which can be experimentally obtained (Fu et al., 2010). VN accountsfor the variation in the gas composition and the temperature acrossthe operating cell, and can be obtained from an integral form of thesteam-hydrogen-based (or the CO2eCOebased) Nernst equation

(Equation (6)).

Vop ¼VN þ i$ ASR (5)

In (6), R is the molar gas constant (8.314 Jmol�1 K�1), TP is theproduct temperature at the co-electrolyzer outlet (K), and TR is thereactant temperature at the co-electrolyzer inlet (K). The mole-fraction subscripts 0, 1, and 2 refer to the cold inlet, hot co-electrolyzer inlet, and hot co-electrolyzer outlet states, respec-tively (O'Brien et al., 2007).

VN can also be calculated by simply averaging the Nernst po-tentials at the inlet and outlet of the cell (Equation (7)), where F isthe Faraday's constant (96,485 Cmol�1); and pO2 ;R and pO2 ;P are thepartial pressures of oxygen for the reactant gas at the co-electrolyzer inlet (kPa) and the product gas at the co-electrolyzeroutlet (kPa), respectively (Fu et al., 2010).

VN ¼ � 0:5�RTR4F

lnpO2;R þRTR4F

lnpO2;P

�(7)

2.2. Base case MeOH production process

MeOH can be synthesized via different reactions routes(Senatore et al., 2018). The overall MeOH production reactions canbe categorized as exothermic synthesis, and occur in parallel withthe endothermic RWGS reaction (Equations (8), (9), and (10)).

Table 1Parameters for the kinetic model (Bussche and Froment, 1996).

Parameter (unit) Value

A BffiffiffiffiffiffiffiffiKH2

p(bar�1) 0.499 17,197

KH2O (bar�1) 6.62� 10�11 124,119KH2O

K8K9KH2

3453.38 e

k05aK

02K3K4KH2

(mol s�1 kg�1 bar�2) 1.07 36,696k

01(mol s�1 kg�1 bar�1) 1.22� 1010 �94765

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 7

CO2ðgÞ þ3H2ðgÞ%CH3OHðlÞ þ H2OðgÞ : DH ¼ �87kJ=mol ð25�CÞ(8)

COðgÞ þ2H2ðgÞ%CH3OHðlÞ : DH ¼ �128kJ=mol ð25�CÞ(9)

COðgÞ þH2OðgÞ%CO2ðgÞ þ H2ðgÞ : DH ¼ þ41 kJ=mol ð25 �CÞ(10)

The MeOH synthesis reactor is among the most important as-pects of MeOH production. The target of the present work is tominimize energy consumption in this system and to improveMeOH synthesis. Therefore, a detailed model of MeOH synthesis isrequired for optimization. There are numerous MeOH synthesiskinetic rate equations in the literature (Bozzano and Manenti,2016). The kinetic equation based on the leading industry cata-lyst, Cu/ZnO/Al2O3, as originally proposed by Bussche and Froment,was used in the current study for the MeOH reactor (Bussche andFroment, 1996). This equation was chosen for two reasons. Firstly,it keeps the process simple; the catalyst used by Bussche and Fro-ment is similar to that used in a Lurgi reactor, for which the reactordimensions have been proven to be applicable (Chen et al., 2011).Secondly, the reactor dimensions are not a concern in this study. Byusing the same catalyst as Bussche and Froment, the same kineticscan be used and can be optimized for the MeOH process in thisstudy. The model is based on the Langmuir-Hinshelwood-Hougen-Watson (LHHW)model (Equation (11) and (12)). Parameters for thekinetic model is provided in Table 1.

rCH3OH ¼k

05aK

02K3K4KH2

pCO2pH2

h1� �1�K*

1�

pH2OpCH3OH

.p3H2

pCO2

�i�1þ �KH2O

�K8K9KH2

�pH2O

�pH2

þ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiKH2

pH2

p þ KH2OpH2O3

rCH3OH ¼0:91 exp

�36696RT

�PCO2

PH2� 3:46� 1010 exp

��1þ 3453:38 exp

�0RT

�PH2OPH2

þ 0:499 exp�17197RT

��PH2

0:5 þ 6

rRWGS ¼1:04� 1010 exp

��94765

RT

�PCO2

� 0:97� 108 exp

1þ 3453:38 exp�

0RT

�PH2OPH2

þ 0:499 exp�17197RT

� �PH2

0:5 þ 6:6

rRWGS¼k

01pCO2

�1�K*

3�pH2OpCO

�pCO2

pH2

�1þ�KH2O

�K8K9KH2

�pH2O

�pH2

þ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiKH2

pH2

p þKH2OpH2O

(12)

The equilibrium constants for K*1 and K*

3 are as follows (Graafet al., 1990):

log10K*1 ¼

3066T

� 10:592 (13)

log101�K*3 ¼ �2073

Tþ 2:029 (14)

The reactor dimensions, catalyst data, and pressure drop(approximately 3 bar) were as established from the Lurgi reactor(Chen et al., 2011), with detailed equations, pressures, and reactionrates, as from mol

kgcatS(Bussche and Froment, 1996). The MeOH syn-

thesis model by Bussche and Froment has already been adopted in aprevious study on process simulation (Van-Dal and Bouallou, 2013)and (Ghosh et al., 2019). In this study, the MeOH reactor (R-1) inFigs. 1 and 2 was modeled using a plug flow reactor (PFR) in theHYSYS simulator.

To make the original form of Bussche's equation equivalent tothe heterogeneous catalytic reaction expression used in HYS-YS,equations (11) and (12) were modified, taking into account thereaction rate unit in HYSYS, kg.mol/m3s. Thus, the reactor voidfraction significantly affects the kinetics of MeOH synthesis becauseHYSYS uses a solid density definition on the interface for the PFRand Langmuir kinetics. The MeOH reactor void fraction wasassumed to be 0.285, which is similar to that used for a Lurgireactor. Modified equations for MeOH synthesis and the RWGSreaction were then established (Equation (15) and (16)). A detailedmodel derivation and validation are presented in the supportinginformation. To calculate the pressure and temperature of the gas inrelation to the kinetic rate and gas fugacity of the MeOH synthesis,the Soave-Redlich-Kwong equation of state (Soave,1972) is suitable(Graaf et al., 1986). This equation was utilized in the simulation.

(11)

�21933RT

�PH2OPCH3OH

P2H2

:62� 10�11 exp�124119

RT

�PH2O

�3kgmolem3

R:s(15)

��55125

RT

�PH2OPCOPH2

2� 10�11 exp�124119

RT

�PH2O

kgmolem3

R:s(16)

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 1180298

For MeOH synthesis, reactors are typically operated in thetemperature range of 220e269 �C (Ganesh, 2014) and at highpressures of approximately 50e100 bar (Van-Dal and Bouallou,2013). In the present study, the MeOH reactor was operated at70.93 bar and 265 �C. The reactor pressure drop was fixed at 3 bar(Chen et al., 2011) and the CO2 conversion was approximately 40%in R-1.

2.3. Base case MeOH purification

MeOH and unreacted reactants were separated using a sepa-rator. The unreacted components were recycled. Subsequently, theMeOH was purified using the separator tank and distillation col-umn. The target MeOH purity and recovery in the distillation col-umn were 99.5mol %. For the purification unit, a near-optimumMeOH distillation preliminary design was utilized; moreover, ashortcut columnwith a reflux ratio of 1.3 times the refluxminimumwas used (Chaniago and Lee, 2018). The last separation of the basecase process consisted of 49 theoretical stages of distillation topurify the MeOH in the liquid phase of the top product with areasonable MeOH stream product temperature of approximately36.3 �C. To calculate the activity coefficient that affects the vapor-liquid equilibrium in the separation that involves water, the non-random two-liquid model (NRTL) was utilized because it is suit-able for separation. In the optimization process, the column stageswere manually minimized and the column reflux ratio was opti-mized using the VS algorithm.

3. Exergy analysis

An energy evaluation to determinate the benefit of the sus-tainable energy process for green technology may be inadequate.An exergy analysis was therefore performed to determine theimproved efficiency and thermodynamic loss due to the irrevers-ible increase in entropy (Hajjaji et al., 2012). This analysis addressesthe loss (destruction) of available useful energy in the system. Byconsidering the 1st and 2nd laws of thermodynamics, the conceptcan be understood and the tool can be developed i.e., exergyanalysis, which is widely employed to analyze the use of energybased on quantitative and qualitative perspectives. Generally, thechange in exergy (DEх) is the sum of the physical (ph) and chemical(ch) exergies (Ghannadzadeh et al., 2012), and is expressed asshown in (17) where the units for DExph and DExch are kJ =h.

DEх¼DExph þ DExch (17)

The physical exergy ðExphÞ of the unit can be defined as themaximum amount of work obtained when a stream of matter istaken reversibly in a physical process. The physical exergy is givenby (Venkatarathnam, 2008):

DExph ¼ðh� hoÞ � To ðs� soÞ (18)

where h and ho are the enthalpies at the examined flow andreference, respectively; s and so are the entropies at the examinedflow and reference, respectively.

The chemical exergy ðExch) can be explained and determined asthe maximumwork that can be obtained frommatter by processesthat involve heat and the exchange of substances in the samesystem. The chemical exergy is expressed by (Im-orb et al., 2018):

DExch ¼X

yiexch;i0 þ RT0X

yilnyi (19)

where yi is the fraction of component i and exch;i0 is the standardchemical exergy of component i.

For a system that includes electrical and mechanical work, theexergy of the work is given as:

ExW ¼ W (20)

Exergy analysis of individual components present in the samesystem can be performed using the following equation:

Exloss ¼Xi

Exloss;i (21)

where, Exloss;i can be calculated as:

Exloss;i ¼Xi

Exin;i �Xi

Exout;i (22)

4. Optimization algorithm

The non-optimized process with a recuperator can be opti-mized using vortex search optimization (VSO) after exergy anal-ysis. VSO is a meta-heuristic type algorithm and is superior toother heuristic types such as pattern search (PS) and randomsearch (RS) algorithms. It has the advantage of using an adaptableinstead of a fixed step size during search tasks. This algorithmoptimizes the exploration and exploitation of the search region toprovide the best results. Therefore, several highly non-linearprocesses have been successfully optimized using the VSOapproach (Qyyum et al., 2019a). In VSO, the initial guess solutionsearch space can be calculated using the mean value of the boundsprovided, as follows:

mo ¼ upper boundþ lower bound2

(23)

Further, n candidate solutions are randomly generated using amultivariate Gaussian methodology (Equation (24)), where d is thedimension, z is the d � 1 vector of a randomly generated variable, mis the d� 1 vector of the sample mean (taken as the center), and n isthe covariance matrix.

pðzjm; yÞ ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffið2pÞd

���y���r exp

� 12ðz� mÞT ðz� mÞ

y

!(24)

n can be computed from y ¼ s2:½I�d�d, where s2 represents thevariance of the distribution and I represent the d � d identity ma-trix. The initial standard deviation (s0) can be calculated as follows:

soð¼ roÞ ¼ maxðupper boundÞ �minðlower boundÞ2

(25)

In this case, s0 can also be considered as the initial radius (r0)and is selected to have a large value initially, for full coverage of thesearch space. The VSO flow diagram is shown in Fig. 6. The radiusstep-size, which is one of the main features of VSO is vital to theadjustments that are performed. A further explanation of the VSalgorithmwith an emphasis on the details of its implementation forthe optimization of highly non-linear and complex systems isprovided in the literature (Do�gan and €Olmez, 2015).

The constrained were handled for the optimization problem bythe exterior penalty function method (EPF) (Qyyum et al., 2018),which involves folding the constraints into the objective function.The reformulation of this problem is represented by Eqn. (26) andthe minimum and upper bounds for each decision variable are

Fig. 6. Vortex search optimization flowchart.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 9

defined in Equation (27).

FðX; p; rh; ghÞ¼ f ðX; pÞ þ PðX; p; rh; ghÞ (26)

XLi �Xi � XU

i i ¼ 1;2; :::n (27)

In (26), rh and rg are the penalty constants (multipliers), and PðX;p; rh; rgÞis the penalty function, expressed in (28).

P�X; p; rh; rg

: rh

Xlk¼1

hkðX; pÞ2�

þ rg

Xmj¼1

gk�max

n0; gjðX; pÞ

o�2�(28)

hk represents the equality constraint and gk is the inequalityconstraint. The present study does not include any equality con-straints; hence, rh is assumed to be zero. HYSYS and MATLAB wereinterfaced using component object model technology for this

Table 2VSO objective functions, decision variables, lower bound (lb), upper bound (ub), and constraints.

Step Process Objective Function Decision Variables Constraints

1 CO and H2 unitProcess

Co-electrolysis duty 1. Split ratio T-1 for stream return to co-electrolysis process(lb¼ 0.1; ub¼ 0.3)2. WGS T-In (lb¼ 750 �C; ub¼ 800 �C)

1.H2 � CO2

COþ CO2¼ 2

2. HX-1 MITA z18e20 �C3. Co-Electrolysis Temperature (750e850 �C)

2 MeOH SynthesisProcess

Max MeOH production (Bottomproduct F-5)

1. R-1 T-In (lb¼ 221 �C; ub¼ 269 �C)2. R-1 T-Out (lb¼ 220 �C; ub¼ 240 �C)3. HX-2 Cold Stream T-Out (Stream 33)(lb¼ 199 �C; ub¼ 230 �C)

1. R-1 T (221e269 �C)2. R-1 T-In> R-1 T-Out3. HX-2 MITA z 8e10 �C4. Reactor Recycle Ratio¼ 3-7

3 MeOH Distillation(C-1)

Reboiler Duty Reflux Ratio (lb¼ 0.1; ub¼ 0.5) 1. MeOH Purity¼ 99.5 Mole %2. MeOH Recovery¼ 99.5 Mole %

Fig. 7. Optimization range for retrofitted case.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 11802910

optimization process. In the proposed study, among the key con-straints, feasible and realistic MITA values of approximately 20 �Cfor fluid temperatures exceeding 300 �C and 10 �C for low fluidtemperatures were chosen (Harvego and McKellar, 2011). The de-tails of the HYSYS andMATLAB linking procedure is provided in the

supporting information.This study focused on improving the overall energy and exergy

efficiency of MeOH production using retrofitted recuperators. Theprocess attempts to optimize the single objective function for eachpart of the process; hence, there are three objective functions

Table 3Equations for physical exergy loss calculation for equipment associated with co-electrolysis and MeOH production.

Equipment Exph�loss (kJ/h)

Compressor (K) Exloss ¼ ðmÞðExin � ExoutÞ� WPump (P) Exloss ¼ ðmÞðExin � ExoutÞ� WCooler exchanging heat with ambient (E) Exloss ¼ ðmÞðExin � ExoutÞPhase separator (F) Exloss ¼ ðmÞExin � ðmÞExLiq � ðmÞExVapThrottle valve (JT valve) Exloss ¼ ðmÞðExin � ExoutÞMixer (M) Exloss ¼

P ðmÞExin � ðmÞðExoutÞTee (T) Exloss ¼ ðmÞðExinÞ�

P ðmÞExoutRecuperator (HX) Exloss ¼

P ðmÞExin �P ðmÞExoutWater gas shift reactor (WGS) Exloss ¼ ðmÞðExin � ExoutÞ� ðDHRÞCo-Electrolysis unit (Co-E) Exloss ¼ ðmÞðExin � ExoutÞþ ðDHRÞMeOH reactor (R-1) Exloss ¼ ðmÞðExin � ExoutÞMeOH column (C-1) Exloss ¼ ðmÞðExin � ExoutÞþ ðQRÞ

Table 4Equations for chemical exergy loss calculation for equipment associated with co-electrolysis and MeOH production.

Equipment Exch�loss (kJ/h)

WGS; Co-E; R-1 Exloss ¼ ðmÞðP yiexch;i0 þ RT0P

yilnyiÞin � ðmÞðP yiexch;i0 þ RT0P

yilnyiÞout

Table 5Standard molar chemical exergies from (Im-orb et al., 2018)and (De Alwis et al., 2009).

Component exch;i0 ðkJ=molÞ

H2O(g) 9.50O2 3.97CO2 19.48CO 274.71H2 236.09CH3OH 754.69

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 11

involved in the optimization (Table 2). The range of optimization isillustrated in Fig. 7.

5. Results and discussions

Exergy analysis was performed to identify the exergy loss in allindividual equipment items in the process. In this work, exergyanalysis was performed for each piece of equipment engaged in co-electrolysis and MeOH production. The physical exergy was calcu-lated for every stream, while the chemical exergy was only calcu-lated for equipment where reactions occurred. The equations usedfor physical and exergy loss calculations in the equipment that areassociated with co-electrolysis and MeOH production are providedin Tables 3 and 4, respectively.

The standard molar chemical exergies for the investigatedcomponents are summarized in Table 5. The calculated physical andchemical exergy losses for the base case are listed in Table 6.

Optimal results were obtained after the three optimization steps

Table 6Physical and chemical exergy losses in equipment associated with base case co-electroly

Equipment Exph�loss ð� 105ÞkJ =hWGS �1669.28Co-E 9130.18R-1 270.53C-1 488.29S (F) 30.05S (M, T and JT valve) 87.82S (K and P) 94.65S (E) 2785.20Whole Process ð� 105ÞkJ =h 11,217.65

listed in Table 2 were applied. For co-electrolysis, H2O was auto-matically adjusted during the optimization process, while main-taining SðH2=COÞ ¼ 2, 2181 kmol =h H2O, as required for 1000kmol =h CO2. There was not a significant reduction of the electrol-ysis power; it decreased by only 1195.82 kW from 206,175.79 kW,due to the fixed conversion reaction and hard constraints. However,the optimization achieved its purpose in terms of finding theapplicable variable of HX. Another important decision variable wasthe stream return to co-electrolysis from T-1. The ratio of thestream at T-1 determined the amount of H2O required, whichinfluenced the co-electrolysis process. The optimal T-1 for thestream return to co-electrolysis was determined to be 0.15. A sig-nificant change was obtained for the R-1 temperature, which waslower than that of the base case R-1. The optimal recycling ratio was3.722, which resulted in more MeOH production.

The MeOH reaction for this process was based on CO. Thecompositions of CO and MeOH during the reaction are compared inFig. 8. The optimized solution alongside the reactor shows adecrease in the CO composition and an increase in the MeOHcomposition for given reactor lengths. The final process, MeOHpurification, was improved with fewer stages and a lower reboilerduty (Table 7c).

After optimizing the retrofitted case, the system was analyzedbased on exergy analysis. The calculated physical and chemicalexergy losses for the optimal case with the recuperators are shownin Table 8.

The expressions for the physical and chemical exergy losses inTables 6 and 8 were utilized. By applying the recuperators, theexergy destruction via the coolers, E (air/water-based), could be

sis and MeOH production.

Exch�loss ð� 105ÞkJ =h Total ð� 105ÞkJ =h0.53 �1668.75�0.23 9129.95�0.11 270.420.00 488.290.00 30.050.00 87.820.00 94.650.00 2785.20

Fig. 8. CO and MeOH composition profiles in the MeOH reactor for the optimal case.

Table 7cOptimal findings for the proposed MeOH distillation streams and equipmentschemes (Step 3).

MeOH Distillation (C-1) Base C-1 Optimal C-1

Temperature e Top (�C) 36.3 48.5Pressure - Top (bar) 1.01 1.01Pressure - Bottom (bar) 2.03 2.03Mole Flow (kmole/h) 936.74 940.19Vapor Fraction 0.00 0.00MeOH Recovery (mole frac) 0.995 0.995MeOH Purity (mole frac) 0.995 0.995Theoretical Stages 49 30Duty (kW) 14,613.64 13,559.67

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 11802912

significantly reduced (Fig. 9).The detailed exergy losses of each equipment for the

recuperator-integrated optimal process are shown in Fig. 9. Theoverall exergy loss was reduced to 9425.57 ð� 105ÞkJ =h, or 15.9%,compared to the non-optimal process. Fig. 9 shows that approxi-mately 82% of the exergy loss occurred in the co-electrolysis unit.The remaining processes accounted for 18% of the exergy loss;however, the majority of this loss occurred in the coolers i.e., 49% of18%. The optimal system with self-heat recuperation significantlyreduced the exergy loss from 2785.20 ð�105ÞkJ =h to 1067.84ð�105ÞkJ =h for the overall heat transfer of the heat exchanger.

After optimizing the self-recuperative, high-temperature, andco-electrolysis-based MeOH synthesis process, the energy-savingopportunities corresponding to exergy efficiency enhancementcan be illustrated via analysis of the composite curves (temperature

Table 7aOptimization of CO and H2 unit for proposed stream and equipment schemes (Step 1).

CO and H2 unit Process 1 2 E-1

Temperature (�C) 30.0 43.9 102.5Pressure (bar) 1.01 1.80 1.11Mole Flow (kmole/h) 2181.00 1000.00 2239.Vapor Fraction 1.00 1.00 0.86Duty (kW) 25,43

CO and H2 unit Process 9 14 15

Temperature (�C) 780.0 103.0 101.0Pressure (bar) 1.11 1.10 1.04Mole Flow (kmole/h) 1563.73 3804.63 1418.Vapor Fraction 1.00 1.00 1.00Duty (kW)Ratio

Table 7bOptimal findings for the proposed MeOH synthesis streams and equipment schemes (St

MeOH Synthesis Process 32 33 E-6 3

Temperature (�C) 75.9 204.0 225.0 2Pressure (bar) 70.93 70.93 70.93 7Mole Flow (kmole/h) 13,993.18 13,993.18 13,993.18 1Vapor Fraction 1.00 1.00 1.00 1Duty (kW) 2395.38

differences between composite curves (TDCC), and temperature-heat flow between composite curves (THCC)) within the heat ex-changers, HX-1 and HX-2. Fig. 10(a) and (b) present the TDCC andTHCC curves for HX-1, while Fig. 11(a) and (b) present the TDCC andTHCC curves for HX-2, respectively.

TDCC analysis provides information on the internal approachtemperatures along the length of the recuperators. The recuper-ators can be more effective when the approach temperature (peakof the TDCC) lies between the defined range. TDCC analysis alsoaddresses the impact of all the design variables on the overallperformancewith respect to the available energy utilization. In HX-1, the peak of the TDCC curve varied from 20.0 �C to 98.0 �C, whichultimately led to different temperature gradients inside theexchanger and explains the exergy losses. However, the peakremained above 20.0 �C, which could be further reduced byemploying a more robust strategy.

Exergy losses can also be evaluated by observing the distancebetween the THCC curves. Fig. 10(b) shows that the warmer andcolder ends of recuperator HX-1 strictly follow the constrainedapproach temperature i.e., 20.0 �C. However, in the temperaturerange of 250.0e550.0 �C, the distance between the compositecurves is greater, leading to a higher approach temperature thatultimately results in a higher exergy loss from this region. This largedistance between the composite curves shows that furtherimprovement is still possible.

6 7 8

51.1 102.5 800.01.11 1.11 1.11

72 1563.73 2239.72 2239.721.00 0.86 1.00

7.43

Co-E E-2 T-1

830.0 40.0 40.01.10 1.01 1.01

68 3804.63 561.890.98 1.00

204,979.97 �2681.99Stream 8¼ 15%

ep 2).

4 35 36 R-1 E-7

25.0 221.0 110.6 221.0 40.00.93 67.93 67.93 67.93 67.933,993.18 12,102.78 12,102.78 12,102.78 12,102.78.00 1.00 0.97 1.00 0.91

25,207.51 �15957.16

Table 8Physical and chemical exergy losses in equipment associated with optimal proposed co-electrolysis and MeOH production with recuperators.

Equipment Exph�loss ð� 105ÞkJ =h Exch�loss ð� 105ÞkJ =h Total ð� 105ÞkJ =hWGS �1576.26 0.51 �1575.74Co-E 8998.29 �0.21 8998.09R-1 239.98 �0.11 239.88C-1 487.81 0.00 487.81S (F) 30.18 0.00 30.18S (M, T and JT valve) 83.20 0.00 83.20S (K and P) 94.32 0.00 94.32S (E) 991.72 0.00 991.72HX-1 62.32 0.00 62.32HX-2 13.79 0.00 13.79Whole Process ð� 105ÞkJ =h 9425.57

Fig. 9. Exergy loss comparison for major equipment items, with and without the recuperator.

Fig. 10. Optimal curves for recuperator HX-1: (a) TDCC, (b) THCC.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 13

The HX-2 THCC curves behaved differently, with only thewarmer end following the constrained approach temperature, i.e.,10 �C (Fig. 11(b)). The colder side of HX-2 results in compositecurves with large separation distances, indicating that the exergyloss in this region is higher compared to that of the warmer side.The HX-2 TDCC curves peaked at a maximum of 35.0 �C and variedbetween 10.0 �C and 35.0 �C (Fig. 11(a)). This is still significantlyhigher than the specified value, i.e., 10.0 �C. However, analysis of thecomposite curves for the proposed recuperators revealed that it ispossible to achieve further minimization of the distance betweenthe TDCC composite curves, as well as the peaks. This couldsignificantly improve the overall performance of the proposed self-

recuperative high-temperature co-electrolysis based-MeOH syn-thesis process.

6. Conclusions

This study presents a progressive implementation of the supe-rior heuristic VS algorithm for innovative optimization to addressthe heat recovery issues associated with co-electrolysis andexothermic MeOH reactors to improve the process efficiency,resulting in cleaner MeOH production. The base case exergy anal-ysis revealed that the highest exergy destruction rate was associ-ated with the process involving thermochemical reactions, with a

Fig. 11. Optimal curves for recuperator HX-2: (a) TDCC, (b) THCC.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 11802914

total exergy loss of approximately 82%. The remaining processesinvolved 18% exergy loss, most of which occurred at the coolers i.e.,49% of 18%. The VS metaheuristic approach searched for optimalvalues of performance parameters for high-temperature co-elec-trolysis-based MeOH production with appropriate operationalMITA of the self-heat recuperation. It was successfully imple-mented in the retrofitted case. The exergy loss from the energytransfer process system was significantly reduced, by 61.7% andachieved an overall exergy efficiency enhancement of 15.9%. Spe-cifically, The VSO could reduce the required energy for steamgeneration at E-1 and the co-electrolysis power. The exergy loss inthe co-electrolysis process was reduced from 9129.95 ð�105ÞkJ =hto 8998.09 ð� 105ÞkJ =h. Optimum gas recycling from T-1 wasachieved. This is considered to be among the influential variables inco-electrolysis that affects H2O requirements and co-electrolysisperformance under self-heat recuperation enhancement. The H2Orequirement was reduced to 2181 kmol =h from 2191 kmol =h tomaintain SðH2=COÞ while increasing the energy efficiency. Optimi-zation of the self-heat recuperation sequence for co-electrolysisand MeOH synthesis not only improved the process exergy effi-ciency, but also increased MeOH production from 936.53 to 940.23kmol =h, with reduced the C-1 reboiler duty.

These results suggested that the improved performance ofmethanol production in the proposed recuperated process could beachieved via the implementation of appropriate optimization andby exploiting the maximum potential benefit of the retrofittingrecuperators. This implies that the proposed solution of self-recuperation with the implementation of appropriate optimiza-tion is a potential solution to the problem of CO2 emission utiliza-tion, which is occasionally hindered by process complexity.Moreover, this method is a flexible approach that can be directlyimplemented for the enhancement of co-electrolysis and MeOHsynthesis in a system that requires heat recovery to improve theefficiency and final product.

Declaration

The authors declare no competing financial interest.

Acknowledgement

This work was supported by the 2019 Yeungnam UniversityResearch Grant. This research was also supported by the BasicScience Research Program through the National Research Founda-tion of Korea (NRF) funded by the Ministry of Education(2018R1A2B6001566) and the Priority Research Centers Programthrough the National Research Foundation of Korea (NRF) funded

by the Ministry of Education (2014R1A6A1031189).

Appendix A. Supplementary data

Supplementary data to this article can be found online athttps://doi.org/10.1016/j.jclepro.2019.118029.

References

Al-Kalbani, H., Xuan, J., García, S., Wang, H., 2016. Comparative energetic assess-ment of methanol production from CO2: chemical versus electrochemicalprocess. Appl. Energy 165, 1e13. https://doi.org/10.1016/J.APENERGY.2015.12.027.

Andika, R., Nandiyanto, A.B.D., Putra, Z.A., Bilad, M.R., Kim, Y., Yun, C.M., Lee, M.,2018. Co-electrolysis for power-to-methanol applications. Renew. Sustain. En-ergy Rev. 95, 227e241. https://doi.org/10.1016/J.RSER.2018.07.030.

Ba~nos, R., Manzano-Agugliaro, F., Montoya, F.G., Gil, C., Alcayde, A., G�omez, J., 2011.Optimization methods applied to renewable and sustainable energy: a review.Renew. Sustain. Energy Rev. 15, 1753e1766. https://doi.org/10.1016/J.RSER.2010.12.008.

Barelli, L., Bidini, G., Cinti, G., 2017. Airflow management in solid oxide electrolyzer(SOE) operation: performance analysis. ChemEngineering 1. https://doi.org/10.3390/chemengineering1020013.

Bello, M.O., Solarin, S.A., Yen, Y.Y., 2018. The impact of electricity consumption onCO2 emission, carbon footprint, water footprint and ecological footprint: therole of hydropower in an emerging economy. J. Environ. Manag. 219, 218e230.https://doi.org/10.1016/J.JENVMAN.2018.04.101.

Bhandari, R., Trudewind, C.A., Zapp, P., 2014. Life cycle assessment of hydrogenproduction via electrolysis e a review. J. Clean. Prod. 85, 151e163. https://doi.org/10.1016/J.JCLEPRO.2013.07.048.

Bilgen, S., Sarıkaya, _I., 2015. Exergy for environment, ecology and sustainabledevelopment. Renew. Sustain. Energy Rev. 51, 1115e1131. https://doi.org/10.1016/J.RSER.2015.07.015.

Bozzano, G., Manenti, F., 2016. Efficient methanol synthesis: perspectives, tech-nologies and optimization strategies. Prog. Energy Combust. Sci. 56, 71e105.https://doi.org/10.1016/J.PECS.2016.06.001.

Bussche, K.M.V., Froment, G.F., 1996. A steady-state kinetic model for methanolsynthesis and the water gas shift reaction on a commercial Cu/ZnO/Al2O3Ca-talyst. J. Catal. 161, 1e10. https://doi.org/10.1006/JCAT.1996.0156.

Cao, Y., Gao, Y., Zheng, Y., Dai, Y., 2016. Optimum design and thermodynamicanalysis of a gas turbine and ORC combined cycle with recuperators. EnergyConvers. Manag. 116, 32e41. https://doi.org/10.1016/J.ENCONMAN.2016.02.073.

Chaniago, Y.D., Lee, M., 2018. Distillation design and optimization of quaternaryazeotropic mixtures for waste solvent recovery. J. Ind. Eng. Chem. 67, 255e265.https://doi.org/10.1016/J.JIEC.2018.06.036.

Chen, L., Chen, F., Xia, C., 2014. Direct synthesis of methane from CO2eH2O co-electrolysis in tubular solid oxide electrolysis cells. Energy Environ. Sci. 7,4018e4022. https://doi.org/10.1039/C4EE02786H.

Chen, L., Jiang, Q., Song, Z., Posarac, D., 2011. Optimization of methanol yield from aLurgi reactor. Chem. Eng. Technol. 34, 817e822. https://doi.org/10.1002/ceat.201000282.

Cu�ellar-Franca, R.M., Azapagic, A., 2015. Carbon capture, storage and utilisationtechnologies: a critical analysis and comparison of their life cycle environ-mental impacts. J. CO2 Util. 9, 82e102. https://doi.org/10.1016/J.JCOU.2014.12.001.

De Alwis, H.P.N.S., Mohamad, A.A., Mehrotra, A.K., 2009. Exergy analysis of directand indirect combustion of methanol by utilizing solar energy or waste heat.Energy Fuels 23, 1723e1733. https://doi.org/10.1021/ef8007129.

DeCicco, J.M., Liu, D.Y., Heo, J., Krishnan, R., Kurthen, A., Wang, L., 2016. Carbonbalance effects of U.S. biofuel production and use. Clim. Change 138, 667e680.

Y.D. Chaniago et al. / Journal of Cleaner Production 239 (2019) 118029 15

https://doi.org/10.1007/s10584-016-1764-4.Do�gan, B., €Olmez, T., 2015. A new metaheuristic for numerical function optimiza-

tion: vortex Search algorithm. Inf. Sci. 293, 125e145. https://doi.org/10.1016/j.ins.2014.08.053.

Fu, Q., Mabilat, C., Zahid, M., Brisse, A., Gautier, L., 2010. Syngas production via high-temperature steam/CO2 co-electrolysis: an economic assessment. Energy En-viron. Sci. 3, 1382e1397. https://doi.org/10.1039/C0EE00092B.

Ganesh, I., 2014. Conversion of carbon dioxide into methanol e a potential liquidfuel: fundamental challenges and opportunities (a review). Renew. Sustain.Energy Rev. 31, 221e257. https://doi.org/10.1016/J.RSER.2013.11.045.

Gao, S., Lin, Y., Jiao, X., Sun, Y., Luo, Q., Zhang, W., Li, D., Yang, J., Xie, Y., 2016. Partiallyoxidized atomic cobalt layers for carbon dioxide electroreduction to liquid fuel.Nature 529, 68e71. https://doi.org/10.1038/nature16455.

Ghannadzadeh, A., Thery-Hetreux, R., Baudouin, O., Baudet, P., Floquet, P., Joulia, X.,2012. General methodology for exergy balance in ProSimPlus® process simu-lator. Energy 44, 38e59. https://doi.org/10.1016/J.ENERGY.2012.02.017.

Ghosh, S., Uday, V., Giri, A., Srinivas, S., 2019. Biogas to methanol: a comparison ofconversion processes involving direct carbon dioxide hydrogenation and viareverse water gas shift reaction. J. Clean. Prod. 217, 615e626. https://doi.org/10.1016/J.JCLEPRO.2019.01.171.

Gielen, D., Boshell, F., Saygin, D., Bazilian, M.D., Wagner, N., Gorini, R., 2019. The roleof renewable energy in the global energy transformation. Energy Strateg. Rev.24, 38e50. https://doi.org/10.1016/J.ESR.2019.01.006.

G€otz, M., Lefebvre, J., M€ors, F., McDaniel Koch, A., Graf, F., Bajohr, S., Reimert, R.,Kolb, T., 2016. Renewable Power-to-Gas: a technological and economic review.Renew. Energy 85, 1371e1390. https://doi.org/10.1016/J.RENENE.2015.07.066.

Graaf, G.H., Scholtens, H., Stamhuis, E.J., Beenackers, A.A.C.M., 1990. Intra-particlediffusion limitations in low-pressure methanol synthesis. Chem. Eng. Sci. 45,773e783. https://doi.org/10.1016/0009-2509(90)85001-T.

Graaf, G.H., Sijtsema, P.J.J.M., Stamhuis, E.J., Joosten, G.E.H., 1986. Chemical equilibriain methanol synthesis. Chem. Eng. Sci. 41, 2883e2890. https://doi.org/10.1016/0009-2509(86)80019-7.

Haider, J., Qyyum, M.A., Kazmi, B., Zahoor, M., Lee, M., 2019. Simulation study ofbiomethane liquefaction followed by biogas upgrading using an imidazolium-based cationic ionic liquid. J. Clean. Prod. 231, 953e962. https://doi.org/10.1016/J.JCLEPRO.2019.05.252.

Hajjaji, N., Pons, M.-N., Houas, A., Renaudin, V., 2012. Exergy analysis: an efficienttool for understanding and improving hydrogen production via the steammethane reforming process. Energy Policy 42, 392e399. https://doi.org/10.1016/J.ENPOL.2011.12.003.

Harvego, E.A., McKellar, M.G., 2011. Optimization and Comparison of Direct andIndirect Supercritical Carbon Dioxide Power Plant Cycles for Nuclear Applica-tions. United States.

Im-orb, K., Visitdumrongkul, N., Saebea, D., Patcharavorachot, Y.,Arpornwichanop, A., 2018. Flowsheet-based model and exergy analysis of solidoxide electrolysis cells for clean hydrogen production. J. Clean. Prod. 170, 1e13.https://doi.org/10.1016/J.JCLEPRO.2017.09.127.

Kauw, M., Benders, R.M.J., Visser, C., 2015. Green methanol from hydrogen andcarbon dioxide using geothermal energy and/or hydropower in Iceland orexcess renewable electricity in Germany. Energy 90, 208e217. https://doi.org/10.1016/J.ENERGY.2015.06.002.

Kim, J., Jun, A., Gwon, O., Yoo, S., Liu, M., Shin, J., Lim, T.-H., Kim, G., 2018. Hybrid-solid oxide electrolysis cell: a new strategy for efficient hydrogen production.Nano Energy 44, 121e126. https://doi.org/10.1016/J.NANOEN.2017.11.074.

Kim, S.-W., Kim, H., Yoon, K.J., Lee, Jong-Ho, Kim, B.-K., Choi, W., Lee, Jong-Heun,Hong, J., 2015. Reactions and mass transport in high temperature co-electrolysisof steam/CO2 mixtures for syngas production. J. Power Sources 280, 630e639.https://doi.org/10.1016/J.JPOWSOUR.2015.01.083.

KOH, J., YOON, D., OH, C.H., 2010. Simple electrolyzer model development for high-temperature electrolysis system Analysis using solid oxide electrolysis cell.J. Nucl. Sci. Technol. 47, 599e607. https://doi.org/10.1080/18811248.2010.9720957.

Koytsoumpa, E.I., Bergins, C., Kakaras, E., 2018. The CO2 economy: review of CO2capture and reuse technologies. J. Supercrit. Fluids 132, 3e16. https://doi.org/10.1016/J.SUPFLU.2017.07.029.

McKellar, M.G., Sohal, M.S., Mulloth, L., Luna, B., Abney, M.B., 2010. MathematicalAnalysis of High-Temperature Co-electrolysis of CO2 and O2 Production in aClosed-Loop Atmosphere Revitalization System. United States. https://doi.org/10.2172/989895.

Mekonnen, M.M., Gerbens-Leenes, P.W., Hoekstra, A.Y., 2016. Future electricity: thechallenge of reducing both carbon and water footprint. Sci. Total Environ. 569(570), 1282e1288. https://doi.org/10.1016/J.SCITOTENV.2016.06.204.

Ni, M., 2012. An electrochemical model for syngas production by co-electrolysis of

H2O and CO2. J. Power Sources 202, 209e216. https://doi.org/10.1016/J.JPOWSOUR.2011.11.080.

O'Brien, J.E., McKellar, M.G., Hawkes, G.L., Stoots, C.M., 2007. Development andValidation of a One-Dimensional Co-electrolysis Model for Use in Large-ScaleProcess Modeling Analysis. United States.

Olah, G.A., Goeppert, A., Prakash, G.K.S., 2009. Beyond Oil and Gas: the MethanolEconomy, second ed. Wiley https://doi.org/10.1002/9783527627806.

Parida, B., Iniyan, S., Goic, R., 2011. A review of solar photovoltaic technologies.Renew. Sustain. Energy Rev. 15, 1625e1636. https://doi.org/10.1016/J.RSER.2010.11.032.

Pierantozzi, R., 2003. Carbon dioxide. In: Kirk-Othmer Encyclopedia of ChemicalTechnology. American Cancer Society. https://doi.org/10.1002/0471238961.0301180216090518.a01.pub2.

Qyyum, M.A., Chaniago, Y.D., Ali, W., Qadeer, K., Lee, M., 2019a. Coal to clean energy:energy-efficient single-loop mixed-refrigerant-based schemes for the lique-faction of synthetic natural gas. J. Clean. Prod. 211, 574e589. https://doi.org/10.1016/J.JCLEPRO.2018.11.233.

Qyyum, M.A., Qadeer, K., Lee, M., 2018. Comprehensive review of the design opti-mization of natural gas liquefaction processes: current status and perspectives.Ind. Eng. Chem. Res. 57, 5819e5844. https://doi.org/10.1021/acs.iecr.7b03630.

Qyyum, M.A., Qadeer, K., Minh, L.Q., Haider, J., Lee, M., 2019b. Nitrogen self-recuperation expansion-based process for offshore coproduction of liquefiednatural gas, liquefied petroleum gas, and pentane plus. Appl. Energy 235,247e257. https://doi.org/10.1016/J.APENERGY.2018.10.127.

Reznicek, E., Braun, R.J., 2018. Techno-economic and off-design analysis of stand-alone, distributed-scale reversible solid oxide cell energy storage systems. En-ergy Convers. Manag. 175, 263e277. https://doi.org/10.1016/J.ENCONMAN.2018.08.087.

Riaz, A., Zahedi, G., Kleme�s, J.J., 2013. A review of cleaner production methods forthe manufacture of methanol. J. Clean. Prod. 57, 19e37. https://doi.org/10.1016/J.JCLEPRO.2013.06.017.

Rosen, M.A., Scott, D.S., 1988. Energy and exergy analyses of a production processfor methanol from natural gas. Int. J. Hydrogen Energy 13, 617e623. https://doi.org/10.1016/0360-3199(88)90010-9.

Sabour, B., Peyrovi, M.H., Hamoule, T., Rashidzadeh, M., 2014. Catalytic dehydrationof methanol to dimethyl ether (DME) over Al-HMS catalysts. J. Ind. Eng. Chem.20, 222e227. https://doi.org/10.1016/J.JIEC.2013.03.044.

Senatore, A., Marino, A., Gordano, A., Basile, M., 2018. Methanol production andapplications: an overview. Methanol 3e28. https://doi.org/10.1016/B978-0-444-63903-5.00001-7.

Soave, G., 1972. Equilibrium constants from a modified Redlich-Kwong equation ofstate. Chem. Eng. Sci. 27, 1197e1203. https://doi.org/10.1016/0009-2509(72)80096-4.

Uhm, S., Kim, Y.D., 2014. Electrochemical conversion of carbon dioxide in a solidoxide electrolysis cell. Curr. Appl. Phys. 14, 672e679. https://doi.org/10.1016/J.CAP.2014.02.013.

Van-Dal, �E.S., Bouallou, C., 2013. Design and simulation of a methanol productionplant from CO2 hydrogenation. J. Clean. Prod. 57, 38e45. https://doi.org/10.1016/J.JCLEPRO.2013.06.008.

Venkatarathnam, G., 2008. Fundamental principles and processes. In:Timmerhaus, K.D., Rizzuto, C. (Eds.), Cryogenic Mixed Refrigerant Processes.Springer New York, New York, NY, pp. 1e50. https://doi.org/10.1007/978-0-387-78514-1_1.

Verhelst, S., Turner, J.W., Sileghem, L., Vancoillie, J., 2019. Methanol as a fuel forinternal combustion engines. Prog. Energy Combust. Sci. 70, 43e88. https://doi.org/10.1016/J.PECS.2018.10.001.

Wang, Yuqing, Banerjee, A., Deutschmann, O., 2019a. Dynamic behavior and controlstrategy study of CO2/H2O co-electrolysis in solid oxide electrolysis cells.J. Power Sources 412, 255e264. https://doi.org/10.1016/J.JPOWSOUR.2018.11.047.

Wang, Yinglong, Bu, G., Geng, X., Zhu, Z., Cui, P., Liao, Z., 2019b. Design optimizationand operating pressure effects in the separation of acetonitrile/methanol/watermixture by ternary extractive distillation. J. Clean. Prod. 218, 212e224. https://doi.org/10.1016/J.JCLEPRO.2019.01.324.

Wender, I., 1996. Reactions of synthesis gas. Fuel Process. Technol. 48, 189e297.https://doi.org/10.1016/S0378-3820(96)01048-X.

Witze, A., 2018. Europe eyes fleet of tiny CO2-monitoring satellites to track globalemissions. Nature 562, 176e177. https://doi.org/10.1038/d41586-018-06963-4.

Zheng, Y., Wang, J., Yu, B., Zhang, W., Chen, J., Qiao, J., Zhang, J., 2017. A review ofhigh temperature co-electrolysis of H2O and CO2 to produce sustainable fuelsusing solid oxide electrolysis cells (SOECs): advanced materials and technology.Chem. Soc. Rev. 46, 1427e1463. https://doi.org/10.1039/C6CS00403B.