A Technoeconomic Model Coupled with HYSYS to Analyze the Electroreduction of Carbon Dioxide to Ethanol

This research develops a technoeconomic analysis to study the profitability of ethanol production from CO2 electroreduction. A HYSYS simulation is used to calculate the separation costs, a challenge in previous models available in literature. The profitability of a 10 000 kg per day CO2 electroreduction plant to produce ethanol is studied. An optimization of the pressure swing adsorber and distillation tower, which greatly influence the total cost of the plant (≈20% of total cost), is carried out, obtaining a total cost of separation of £1.94 ×$\ \times \ $ 106. The study demonstrates that reducing the voltage applied to values around 0.5 V, i.e., by increasing the pH up to 12 makes the process economically feasible with a current density over 15 mA cm−2. It also shows that the process is economically feasible using a current density of 5 mA cm−2 if the electricity cost from renewable sources drops to 2.0 ×$\ \times \ $ 10−2 £ kWh−1. Finally, it is proved that if catalyst stability is not considered, some catalysts currently available in the literature can be used with a positive economic income. The results of this research show that the industrial electroreduction of ethanol can be feasible and can attract interest in the industrial adoption of the technology.


Introduction
Anthropogenic increase in carbon dioxide emissions threatens the future of the planet. It is reported that 203 Gt of CO 2 is recycled every year through the carbon cycle, which helps to maintain a sustainable balance in the atmosphere. [1] However, due to human activities, over 30 Gt of additional CO 2 are consistently added to the atmosphere yearly, resulting in an unsustainable increase in concentration. [1] In 2018, more than 50% of global CO 2 DOI: 10.1002/adts.202200782 production came from the use of fossil fuels in the power industry or other industrial combustion. [2] The COVID-19 pandemic temporarily affected the CO 2 emissions, but the reliability of fossil fuels continues to increase due to energy requirements and the emissions increased after the pandemic. [3] However, the society has become more conscious of the negative effect of the CO 2 emission on the environment and many governments have established objectives to reach net-zero carbon emissions. [4] To address the current fuel crisis, a possibility is to utilize the CO 2 and convert it into fuels to create a more sustainable energy profile. [5] The conversion of CO 2 to ethanol meets these criteria as it is a renewable fuel with an energy density of 1,366.8 kJ mol −1 , making it a suitable alternative to fossil fuels. Carbon dioxide can be removed from the atmosphere and transformed into valuable chemicals mimicking the behavior observed in plants in a process known as artificial photosynthesis. While the process of producing complex hydrocarbons similar to those produced by plants is currently impossible to replicate, it is plausible to produce light liquid hydrocarbons such as formic acid, methanol, or ethanol. [6] One promising type of artificial photosynthesis is based on producing electricity through solar photovoltaic systems that can be used to convert CO 2 . [7] The electroreduction technologies avoid energy-intensive steps as they can be performed under ambient conditions. [5,8] However, challenges remain with this technology and to prove it is a worthy investment, it needs to be assessed on an industrial scale to determine economic feasibility. [9] Technoeconomic models can be used to assess the profitability of CO 2 electroreduction chemical plants using different economic principles such as general technoeconomic and gross margin -which calculate the profit per kilogram and hour, respectively -and net present value (NPV), which calculates the profit over the span of a project. [10] Ruiz-López et al. reported ethanol together with ethylene as liquid products with larger market interest. Ruiz-López et al.'s article also highlighted the scarce number of studies focused on the economic assessment of CO 2 electroreduction due to the complexity of the analysis. [11] A reoccurring challenge across the models developed to evaluate the production of ethanol from electroreduction of CO 2 is the calculation for the separation costs. The separation costs have also been described as a vital element in models to calculate the overall profitability of a plant to produce other liquid products like formic acid. [12] In previous models, Orella et al. used generalized correlations to account for the separation costs. [13] This generalization means that the type of separation equipment used and the operation cost of separation are not accurately accounted for in this model. Verma et al. and Jouny et al. also identified the separation costs as a point for further work after applying their models. [10,14] The incertitude about the cost of separation has ended with the models reaching different conclusions about the profitability of producing ethanol.
A way to simulate and estimate separation costs in chemical plants is by using the energy and material balance results reported on HYSYS-based simulations. [12,15] There are examples in the literature of the simulation of distillation columns to separate hydrocarbons and alcohols for this purpose. Hashmi conducted an economic analysis to separate a mixture of heptanetoluene, comparing two distillation processes and finding that extractive distillation was economically more feasible than conventional distillation. [16] Tehlah et al. used HYSYS as a platform to calculate the cost of crude oil distillation units and used it to optimize the process. [17] The simulation of the separation process for ethanol purification has been studied in bioethanol production plants. [18] Focusing on the cost of alcohol separation, Zubir et al. simulated and studied the cost of separating alcohol mixtures. [19] In this work, we integrated a technoeconomic model with HYSYS simulations for accurate calculation of the capital and operating costs in a plant for ethanol production from the electroreduction of CO 2 . In particular, we intend to improve and optimize the accountability of the separation costs due to distillation in the ethanol purification step and the recycling of the unreacted CO 2 . To the best of our knowledge, this is the first study integrating both tools and we intend to offer a new and more accurate platform to evaluate the ethanol synthesis cost using electroreduction methods. The model obtained can be directly applied in analyzing the profit of electrocatalytic processes, such as water electrolysis to produce hydrogen, CO 2 electroreduction, or electrolytic wastewater treatment.

Aspen HYSYS Simulation
References and screen images from Aspen HYSYS are used and reprinted with permission from Aspen Technology, Inc. AspenTech, Aspen HYSYS and the AspenTeach leaf logo are trademarks of Aspen Technology, Inc. All rights reserved.
HYSYS was used to model a plant to treat ≈10 000 kg per day of CO 2 . A plant like this would be able to treat around 4000 CO 2 tonnes per year, which were generated by small energy plants producing around 200 MW h −1 from gas fuel. [20] A process flow diagram of the plant is included in Figure 1. The model flow sheet used in HYSYS and information on each of the streams can be found in Section S1 (Supporting Information). The stream "Fresh CO2" was fed to the "CO2 Mixer" block, where it was mixed with the recycled gas stream. The "Fresh CO2" stream was assumed to contain only CO 2 and its thermodynamic condition and mole flow rate were inputted. A conversion reactor module (named "Reactor") was used to simulate the reaction inside the cell. The hydrogen ions H + were the charge carrier within the cell. They were transported through the electrolyte to the cathode side, where they reacted with CO 2 to produce e − . The cells of electrocatalytic reactors could not be modeled in HYSYS; therefore, the overall reaction was specified in the "Reactor" block. The "Reactor" block assumed that the whole reactor was a black box with no differentiation for the anode and cathode reactions. The CO 2 would be recirculated in the reactor until the total CO 2 conversion reached the desired value, which was one of the parameters studied in this work. The Set function "Condition Set" in Aspen HYSYS was used to set the electrolytic cell reactor at isothermal conditions by specifying the temperature of the outlet "Gas Stream" stream to be the same as the temperature of the feeds. Similarly, the stream "Fresh H2O" containing pure water was fed to the "H2O Mixer" block, where it was mixed with the recycled gas stream from the distillation column, and its thermodynamic condition and mole flow rate were inputted to maintain the gas/water stoichiometric molar ratio of 2:3 in the feeds to the electrolytic cell reactor.
The gas product stream ("Gas Stream") was treated to achieve CO 2 purity of at least 99% before recycling using pressure swing adsorption (PSA) technology. Since a PSA could not be modeled in Aspen HYSYS, a Component Splitter (named "Gas Separation Unit") represented the separation unit. The Component Splitter allowed for the calculation of the material and energy balances used to calculate the amount of adsorber and the cost of the unit. A general distillation was chosen for methanol separation from the liquid product stream ("Liquid Stream"). The "Heat Exchanger 1" was included to raise the temperature of the "Liquid Stream," while lowering the bottom product temperature to enhance the separation within the distillation column. The bottom product was further cooled down by the "Heat Exchanger 2" to the operating temperature of the electrolytic cell reactor of 25°C.
Pressure drop effects were not considered for the simulation of this plant.

Technoeconomic Model
Critical variables, such as Faradaic efficiency and current density, are commonly used in literature to evaluate the performance of electrocatalytic systems. [21] These key performance indicators are utilized to build the model and examine the profitability of existing catalysts in the literature. A description of the two key figures of merit -Faradaic efficiency and current density -is included in Section S2 (Supporting Information).

Model Assumptions
The model operated within the following assumptions that led to the creation of the block diagram in Figure 2.
1) The model was assumed to operate for 350 days per year, with 15 maintenance days, for 5 years. [14] 5 years payback is a common criterion for attractive technologies, so this period was selected for the time of operation. [22] It will also operate at 0.116 kg s −1 (10 000 kg per day), a suitable operating condition as defined by the literature. [14] 2) The reactor is constantly operating during this time period with no decrease in productivity. The reactor also operates at a steady state, which means no replacements or adjustments to the production rate are considered. While the durability of the catalyst is of concern, this assumption is used in literature and reduces the need for additional user input. [13] 3) The system is modeled on a lumped basis, so spatial variation is not considered within the reactor for mass or heat transfer. The costs used in the model are overestimated in literature to consider any effect of the mass or heat transfer due to the reactor geometry. [13] 4) The reactor operates isothermally and no temperature variation is considered. This can be assumed as an advantage of electroreduction compared to other CO 2 reduction methods because there are no steps which require a high energy input or high temperature. [5] 5) The costs per electrode area are derived as per the alkaline water electrolyzer system following examples in the literature. [14] The production of hydrogen through electrolysis provides a good basis for fixed costs per area due to its alkaline reaction conditions and similar scale. 6) The reactor contents: reactants, products, and electrolytes undergo total recycling until the desired conversion is completed, as shown in Figure 2. The total reaction time is beyond the scope of the simplified model. A low liquid accumulation means total recycling must be used to achieve ideal conversion. Therefore, the electrolyte is recycled with the total reactor contents and separated via distillation as waste, as assumed in the literature. [14] The model consists of a carbon dioxide and water feed into the reactor with a total recycle proceeding this. Once the conversion reaches the desired amount, it is fed into the separation units as simulated with Aspen HYSYS, which consists of a pressure swing adsorber and distillation tower. The carbon dioxide and water are recycled back into the reactor and this results in an ethanol stream of around 96 wt% as per the distillation specification.

Capital Costs
The capital cost has been divided into reactor cost (C rea ), cost of the electrode (C ele ), cost of the catalyst (C cat ), cost of separation (C sep ), and cost of coolers and heat exchanger (C coo ). www.advancedsciencenews.com www.advtheorysimul.com C rea , C ele , and C cat are calculated using Equations (1)-(3), respectively [10,13] where C rea is the capital cost associated to the reactor cost, C elec is the capital cost due to the electrode, and C cat is the catalyst capital cost. F rea is the fixed cost associated with the reactor per electrode area (11 531.4 £ m −2 ), F ele is the fixed cost associated with electrodes per electrode area (7 977.69 £ m −2 ), F cat is the fixed cost associated with the catalyst (0.0232 £ kg −1 ), W cat is the catalyst loading (10 mg m −2 ), F GDL is the cost of gas diffusion layer per electrode area (43 £ m −2 ), and A is the electrode area (m 2 ). [10,13,23] Equation (4) was formulated to calculate the electrode area, A, as the costs are dependent on this value according to Equations (1)-(3). The electrode area is related to production rate, Faradaic efficiency, and current density and the process to obtain Equation (4) is shown in Section S3 (Supporting Information) whereṀ p,rea is the production rate of ethanol in the reactor (kg s −1 ) obtained from the simulation, F is the Faradaic constant (C mol −1 ), FE is the Faradaic efficiency (%), e − p is the number of electrons exchanged in the reaction, MW p is the molecular weight of ethanol (kg mol −1 ), j is the current density (A m −2 ), and Y p is the stoichiometric coefficient.
C sep is calculated as the addition of the cost of: 1) The distillation column considering the cost of the shell (C shell ) and the internal plates (C internal ) from Equations (5) and (6) [24] 2) The condenser (C cond ) and the reboiler (C reb ) are calculated with Equations (7) and (8) [24] 3) The adsorption column (C c ), calculated with Equation (9), and the adsorbent (C zeo ), calculated with Equation (10) [25] log In the previous equations, F mat is the fixed cost associated with the construction material of the column (8 £ kg −1 for carbon steel), F 0ST is the fixed cost associated with the condenser and reboiler per heat transfer surface area (430 £ m −2 for carbon steel), F zeo is the fixed cost associated with the adsorbent (0.221 £ kg −1 ). [24] The total shell weight of the distillation column, W V , is the sum of the weight of the vessel and the supporting skirt. It can be written in terms of column height, diameter, and wall thickness, as in Equation (11) [24] where ac is the construction material density (7,850 kg m −3 ), D c is the column diameter (m), H c and H s are the height of the column and the supporting skirt, respectively (assumed 2 m), e c and e s are the wall thickness of the column and the supporting skirt, respectively (assumed 0.005 and 0.01 m). The heat transfer surface area of the condenser and reboiler can be obtained from Equations (12) and (13) where Q cond and Q reb are the heat duty of the condenser and the reboiler obtained in the simulation, respectively (kJ s −1 ), U is the general heat exchanger coefficient (kJ s −1°C−1 m −2 ) and ΔT lm is the log mean temperature difference (°C). The condensation temperature is assumed to be constant at T C = 30°C, the cooling water temperature 20°C, and U c is 500 kcalh −1°C−1 . According to Didier and Perez, U R ΔT lm ≈ 30 000 kcal h −1 . [24] A typical PSA process is transient, multistep, multipressure, and cyclic. It uses two or more identical columns (physically and operationally) that operate out of phase with each other to allow an overall continuous feed to the process. [25] In this research, the PSA unit is assumed to include two identical adsorption columns. Each column has a fixed bed of solid zeolite adsorbent. Since H 2 O in the vapor product adsorbs strongly on most known adsorbent for CO 2 , it is normally removed selectively before CO 2 capture. [26] For simplification, H 2 O is assumed to be removed from the gas before entering the PSA unit, and this additional expenditure is ignored. In addition, the PSA process may compress the feed for separation, which would increase its temperature. [26] This means that a compressor, followed by a cooler, is required upstream of the adsorption columns. These units have not been included in the simulation for simplification, as they are not necessary for the simulation to converge.
According to Susarla et al., a gas feed flow of about 2.8 kg s −1 gave the minimum capture energy for a single column of 6 m length and 2 m diameter. [25] Therefore, the dimension of the adsorption columns used in this system is sized accordingly based on the mass flow rate of the vapor product. The column is assumed to be cylindrical, so its volume can be obtained from Equation (14) where D ad and H ad are diameter and height of the adsorption column, respectively (m). www.advancedsciencenews.com

www.advtheorysimul.com
Assuming each bed operates for t c = 2 h between regeneration, the amount of zeolite required is given by Equation (15) M zeo =Ṁ CO 2 × t C q (15) whereṀ CO 2 is the molar flow rate of CO 2 recovered (kmol s −1 ) obtained from the simulation and q is the loading (5 mol CO2 kg zeo −1 ). [26] Finally, the cost of the cooler and heat exchanger (C coo ) is calculated with Equation (16) and considering that both heat exchangers have the same area of heat exchange [27] where A HE is the area of heat exchange obtained from the simulation (m 2 ). All the costs were updated to 2021 prices to reflect the inflation rate.

Operation Costs
The operation consists of electricity (C op,elec ), utilities (C U,cond , for condenser and C U,reb , for reboiler), and material costs (C mat ). The electricity costs were calculated using Equations (17)-(20) [13] C op,elec = P elec ⋅ ( where I is the current (A), V is the voltage (V), P elec is the price of electricity (3.9 × 10 −5 £ Wh −1 ), P W is the cost associated with the energy consumption of the condenser (2.54 × 10 −7 £ kJ −1 ), and P S is the cost associated with the energy consumption of the reboiler (6.67 × 10 −6 £ kJ −1 ). [28] The condenser and reboiler duties (Q cond , Q reb ) and inlet flow rate of CO 2 and H 2 O (Ṁ CO 2 ,in ,Ṁ H 2 O,in ) can be obtained from the simulation. F CO2 and F H2O are the cost of CO 2 (0.01088 £ kg CO2 −1 ) and water (0 £ kg H2O −1 ), respectively. is the operation period (5 years). The energy consumption of the PSA (Q ads ) unit can be estimated using Equation (21) [13,29] Q ads = P CO2 ⋅Ṁ CO 2 (21) where P CO2 is the energy consumption per unit CO 2 recovered (MJ kg CO2 −1 ),Ṁ CO 2 is the mass flow rate of CO 2 recovered (kg h −1 ) obtained from the HYSYS simulation.
The costs associated with the energy consumption of the condenser and the reboiler were updated to 2021 prices to reflect the inflation rate.

Profit Calculations
To calculate the income (INC) made by electroreduction considered in the model, the equation in Equation (22) was used. The market price of ethanol is used to calculate the income from the process whereṀ p,out is the flow rate of ethanol obtained in the process and calculated with the simulation (kg s −1 ) and the F P is the market price of ethanol (0.72 £ kg −1 ). [14] The profit value was updated to 2021 prices to reflect the inflation rate.

Distillation Column Optimization
The plant flow sheet includes two separation operation units: 1) an adsorption column for the separation of the CO 2 not reacted and that is recirculated to the reactor; 2) a distillation column that separates the liquid products of the reaction.
Distillation columns are commonly used for the purification of ethanol. [30] The azeotrope between water and ethanol limits ethanol purification using distillation at ≈96% (by mass). [31] Using a different technology, such as membrane separation, would be necessary to reach higher purification of ethanol needed for certain uses. However, this would increase the costs of separation. In this work, the study was limited to the use of a distillation column and, in the first stage, it evaluated the effect of different parameters on the cost of the column to optimize it.
The number of plates in the column was optimized using HYSYS based on the total cost of the distillation column and the mass fraction of ethanol obtained. The number of plates in the column will determine the ability of the column to reach a good separation between water and ethanol. In principle, a higher number of plates will ensure the maximum purification is reached but will also increase the size of the column, having a direct impact on the capital costs. The total cost (including capital and operation cost) of the column at different plate numbers is included in Figure 3A. It can be observed that the column shows the lowest cost when using 20 plates. The cost when using 15 plates is slightly more expensive. The cost of the column dramatically increases when using 5 plates or more than 35 plates. When using 5 plates, the simulation uses a high amount of energy in both the condenser and the reboiler to ensure good separation of ethanol. Therefore, while the capital cost is the minimum, the operation cost makes this configuration unviable. The operation costs remain approximately constant when the number of plates is 15 or higher due to the azeotropic mixture. However, the capital costs make the column more expensive as more plates are added. Therefore, 20 plates were selected as the optimum configuration. Figure 3A also shows how the mass fraction of ethanol obtained with the distillation column increases as more plates are used. The mass fraction of ethanol is 1.5% less when using 15 plates than when using 40. However, this value can be optimized when modifying other column parameters, allowing it to reach a good purification at the minimum cost.
The second parameter optimized was the entry point to the column. The entry point helps to improve the degree of purification of ethanol. Figure 3B shows how the percentage of ethanol achieved increased when the liquid fraction to separate was introduced closer to the reboiler. In Figure 3, plate 20 refers to the bottom of the column (closer to the reboiler) and plate 1 to the top of the column (closer to the condenser).
Due to the low fraction of ethanol (18% by mass) obtained in the reactor, the most efficient separation was achieved when the liquid products obtained in the reactor were introduced at the bottom of the column (plate 18). However, introducing the liquid products for separation in plate 16 decreases the operation costs as the energy necessary in the condenser and the reboiler is lower. Therefore, plate 16 was selected as the entry point, as the difference in the mass fraction obtained was just 0.2%.
Finally, the size of the column was optimized. The plate spacing would determine the length of the column. Common values for this parameter are from 0.1 to 1 m. [32] An intermediate value of 0.55 m was used in previous calculations. Figure 3C shows the cost and mass fraction of ethanol obtained at different plate spacing values based on the HYSYS simulation. The operation cost remains very similar for all the spacing values. However, the capital costs increase as the column becomes larger. The value selected for future simulations based on this study was 0.1 m. The degree of separation remained constant for all the plate spacing values studied and is reasonable considering the scale of the process.
The effect of the diameter of the column was also evaluated using HYSYS and shown in Figure 3D. The mass fraction of ethanol obtained only varies slightly with the diameter of the col-umn. Similar to the effect of the plate spacing, the operation costs barely change, while the main effect is observed in the capital costs, which increase as the diameter becomes larger. A good degree of separation is achieved and the costs are lower with small diameters due to the scale of the process. For this reason, 0.5 m was selected as an optimum value to keep a good ratio of diameter/length. [33] Around 0.4 M£ is saved after optimization of the parameters in the range studied.

Effect of Adsorption Column in the Cost of the Plant
Beside the distillation column, the rest of the separation costs are due to the adsorption column, which also strongly depends on the cost of the adsorbent. In this research, zeolite with a cost of 221 £ t −1 was selected for the calculations. [34] The amount of zeolite and the size of the adsorber will be based on the amount of CO 2 unreacted from the reactor.
The electroreduction of CO 2 is performed via electrocatalysis, a combination of electrolysis and catalysis. Reactions (R1)-(R3) take place in the electrolytic cell Reactions 1 and 2 happen in the anode and the cathode, respectively. In the simulation, the whole reactor was considered a black box with no separation of the anode and cathode. Instead, the overall reaction, presented as Reaction 3, was used as an input in HYSYS.
Despite the current conversions of the reaction being below 10%, in the calculation of the next sections, it will be considered that the mixture of CO 2 and water was recirculated, achieving superior conversions. [14,35] However, this factor affects the separation costs, so the influence of conversion values in the simulation was evaluated. The results are included in Figure 4.
The separation costs (including the cost of the distillation column) are strongly influenced by the conversion achieved in the reactor. They are up to 5 times more for efficiencies under 10% achieved with the current electrochemical technologies.
Both the capital and the operation cost increase with the lower efficiency. In the case of the capital costs, the smaller the amount of unreacted CO 2 , the smaller the quantity of adsorber needed. The cost of the adsorber needed to recover the unreacted CO 2 makes the process extremely expensive for the operation cost. However, the decrease in efficiency leads to slightly higher costs for the distillation column operation.
The lower efficiencies also affect the profitability of the process as the mass flow of ethanol obtained after the distillation column decreases dramatically with efficiencies under 5%.
On the other hand, in this section, it was not considered that a recirculation step inside of the reactor to achieve a higher conversion of CO 2 would increase the volume and the cost of the reactor exponentially, as the concentration of the reactants would decrease each recirculation. For this reason, it was accounted that a 50% conversion of CO 2 was achieved in the reactor for future calculations. This value was selected previously in the literature. [14] Furthermore, the separation costs only slightly increase when the CO 2 conversion drops from 99% to 50% of CO 2 conversion and the mass flow rates and mass fraction of ethanol obtained after the distillation column almost do not vary with CO 2 conversions >50%.
In Figure 4B, it can be seen that the main cost of separation is due to the distillation column operation, followed by the electric cost of the PSA. The capital cost only represents 7% of the total costs of separation. The capital cost of the PSA is negligible for the cost of zeolite selected in this research, but it is very variable and dependent on the cost of the adsorbent.
The total cost of separation after optimization is £1.95 × 10 6 , which is in the same order of magnitude as the cost calculated using Kunnakorn et al. and Grande et al. approaches (£2.46 × 10 6 ), as in Equation (23) [36] where F PSA is the cost of the adsorption per kilogram of CO 2 separated (0.04 £ kg CO2 −1 ),Ṁ CO 2 is the mass flow rate of CO 2 separated (kg h −1 ) obtained from the simulation, F Dis is the cost of distillation per kilogram of ethanol separated (0.15 £ kg ethanol −1 ), andṀ p,out is the mass flow rate of ethanol separated in the distillation column (kg h −1 ) obtained from the simulation. The cost has been updated to 2021 values to consider the inflation.

Effect of the Adsorption Column in the Cost of the Plant
This section investigates how different reaction operation parameters affect the cost and profitability of the plant, based on a technoeconomic model. Figure 5A shows the income of the plant when the reactor works at different Faradaic efficiencies and current densities. At first, the model shows a loss in income across all Faradaic efficiencies for the process. Even at very high Faradaic efficiencies, the process plant does not obtain any positive income. The income values reach an asymptotic value and do not show any profit even at the highest current density studied (200 mA cm −2 ; not shown in the graph).
The current density is demonstrated to be a key factor in the profitability of the plant. In Figure 5B, it is shown the highest costs as a function of the current density when the Faradaic efficiency is 99%. This graph reveals that the highest cost came from the electricity operation costs, which shows that improvements in the efficiency of the catalyst can have a great influence on the profitability of the plant at current densities over 10 mA cm −2 . The second factor that influences the total cost of the plant is the reboiler. This cost is fixed as per the HYSYS simulation carried out in the previous sections and justifies the need for a thorough optimization of the separation units in the plant. The values of www.advancedsciencenews.com www.advtheorysimul.com  the cost of the reactor and the electrodes are very significant at low current densities, demonstrating the importance of finding stable catalysts able also to work at high current densities.
The electricity cost from Figure 5B can also be adjusted using process parameters. The electricity cost was calculated with Equation (17). The current could not be altered directly to reduce electricity costs because it is a dependent variable and not an input to the model in Equation (17). However, the voltage can be varied to improve the profitability of the plant. Figure 6A shows the effect of the voltage used in the process. Voltage describes the energy required to make the current flow and the electroreduction reaction to occur. [37] Higher values of voltage needed for the reaction to occur will increase the cost of the operation cost of the reactor. The process became economically viable at 0.5 V, according to the model, and a current density of 120 mA cm −2 . Voltage had originally been given as 0.74 V needed for the reaction to occur at pH = 7.
The voltage value required for the reaction to happen is not the same for all electrochemical cell conditions and, by increasing the pH of the cell, the voltage required can be decreased. A relationship between the pH and the cell voltage can be derived from the Nernst equation for this reaction, as shown in Equation (24) where E ′ 0 is the standard potential (−1.15 V) and pH is the value of pH. The variation of income with pH is illustrated in Figure 6B considering Equation (24). A higher pH favors the profitability of the plant. However, increasing the pH favors ethylene production, a competing reaction pathway to ethanol. [38] While under high pH conditions, the morphology of the catalyst could also be altered, affecting the stability and decreasing the Faradaic efficiency of ethanol production. [5] Figure 6B demonstrates that for the process to be optimized in profitability, shown by a positive value, the current density must be increased to around 8 mA cm −2 or above. At more mild pH conditions, pH < 12, the current densities needed must be over 10 mA cm −2 .
In the literature, many catalysts are examined under KHCO 3 electrolyte conditions. [39] Electrolytes such as KHCO 3 often act as pH buffers which means they increase the bulk pH while decreasing the local pH (area near the catalyst). [38] This provides the high pH environment preferred for minimizing electricity costs in the bulk while maintaining a local pH influenced by current density, which avoids the negative impacts on ethanol production of catalyst inefficiency and ethylene favoring. [38] The last parameter investigated to evaluate the profitability of the plant is the cost of electricity. In the previous calculations, a value of 3.9 × 10 −2 £ kWh −1 was used. The viability of the plant  will rely on cheap electricity generation of electricity from renewable resources. The lowest cost of generation recorded for electricity from renewable sources was 2.0 × 10 −2 £ kWh −1 for offshore wind in the USA. [40] A sensitivity analysis was performed on the system with a pH of 12 and a Faradaic efficiency of 99% to show the impact of this price range. The results of this are shown in Figure 7. It can be seen that the drop in electricity cost from renewable energy to 2.0 × 10 −2 £ kWh −1 would permit a profitable operation at current densities over 5 mA cm −2 .
After the evaluation of the voltage, pH, and electricity price, a final evaluation of the income at different Faradaic efficiencies was carried out. In this case, the value for the electricity price was taken from the UK Government's prediction of the cost of energy generation from renewables for optimal conditions (2.8 × 10 −2 £ kWh −1 ). [41] The pH was chosen to be 12 to simulate the basic conditions which could be generated by the concentrations of electrolyte in the reactor. Figure 8 shows the results generated for these conditions.
The general technoeconomic analysis showed a decrease in cost with an increase in current density. Figure 8 also shows the current state of some catalysts investigated for ethanol production from electroreduction of CO 2 . [42] It demonstrates that there exist catalysts that can work in conditions that could generate positive income, such as the works of Song et al. and Liu et al., making the process economically attractive. [42] In addition, a NPV calculation was also used to study the profitability of the plant. The technique considers a capital investment analysis of the project, which takes into account the return on your investment compared to what it would expect from saving that money conventionally (accounting for inflation). Therefore, it provides a good indication of whether a project is worth investing in, factoring in the time delay of seeing returns on that investment. It can be summarized with Equation (25) NPV = Expected profit generated −value of the amount invested today (25) The NPV model found that ethanol could only be achieved at over 300 mA cm −2 and 70% Faradaic efficiency. [14] These results are in the same order, but slightly more restrictive than those shown in Figure 8. The difference can be due to the NPV model considering conditional saving rather than a positive income.

Conclusion
The combination of Aspen HYSYS simulations with a technoeconomic analysis provided an overview of the profitability of a CO 2 electroreduction plant to produce ethanol. The technoeconomic model demonstrated the economic feasibility of producing ethanol from CO 2 using current electrocatalyst technology, assuming that the catalyst is stable at pH 12 and the electricity cost is below 2.8 × 10 −2 £ kWh −1 . The electricity operation costs were found to be the major cost of the process and can be cut down by finding ways to reduce the reaction voltage under 0.5 V when the current density is over 15 mA cm −2 . This can be achieved by working at higher pH. The drop in electricity cost from renewable energy to 2.0 × 10 −2 £ kWh −1 will benefit the use of a higher range of catalysts working at current densities over 5 mA cm −2 .
Separation costs were also shown as a major cost in a CO 2 electroreduction plant (£1.94 × 10 6 ). The greater precision of the separations process using HYSYS meant this article was able to provide accurate values for separation costs and prevented overestimation. The conversion achieved in the reactor greatly influences the separation cost, especially for operation separation costs, which account for more than 90% of the cost if conversion achieved in the reactor is 50%. Finally, it was demonstrated that optimizing the distillation column parameters is essential to reduce costs and achieve adequate ethanol purification, achieving up to a 0.4 M£ saving in the plant studied by working with a column with 20 plates, a diameter of 0.5 m, a plate spacing of 0.1 m, and using the plate 16 for the inlet.
Further research recommendations should focus on integrating the cost of the deactivation of the catalyst in the technoeconomic model. Additional research is necessary to find catalysts able to achieve higher conversion and Faradaic efficiencies at lower current densities and high pHs. These advances would reduce the electricity cost. Investigating technologies to improve CO 2 and O 2 , as well as water and ethanol separation, will also bring the technology closer to commercialization.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.