Integrating CO2 capture with electrochemical conversion using amine based capture solvents as electrolytes

Carbon dioxide (CO 2 ) is currently considered as a waste material due to its negative impact on the environment. However, it is possible to create value from CO 2 by capturing and utilizing it as a building block for commodity chemicals. Electrochemical conversion of CO 2 has excellent potential for reducing greenhouse gas emissions and reaching the Paris agreement goal of zero net emissions by 2050. To date, Carbon Capture and Utilization (CCU) technologies (i.e. capture and conversion) have been studied independently. In this communication, we report a novel methodology based on the integration of CO 2 capture and conversion by the direct utilization of a CO 2 capture media as electrolyte for electrochemical conversion of CO 2 . This has a high potential for reducing capital and operational cost when compared to traditional methodologies (i.e. capture, desorption and then utilization). A novel mixture of chemical and physical absorption solvents allowed for the captured CO 2 to be converted to formic acid with faradaic efficiencies up to 50 % and with carbon conversion of ca. 30 %. By increasing the temperature in the electrochemical reactor from 20 °C to 75 °C, the productivity towards formic acid increased by a factor of 10, reaching up to 0.7 mmol·m -2 ·s -1 . The direct conversion of captured CO 2 was also demonstrated for carbon monoxide formation with faradaic efficiencies up 45 %.


Introduction
The abrupt increase of CO2 emissions into the atmosphere due to the use of fossil fuels creates adverse effects in the environment 1,2 . It is imperative that measures are implemented to drastically lower these emissions in pursuance of the Paris agreement. To date, many strategies have been contemplated with the ambition of achieving this purpose. Carbon capture and sequestration (CCS) [3][4][5] , where CO2 is captured from industrial sources and stored underground, is and will remain, an important approach to diminishing atmospheric CO2 concentrations. Carbon capture can be considered a mature technology and has been used in a broad variety of industrial applications. Despite the high technology readiness of the technology, extensive studies are still being carried out to improve its economical feasibility and reduce the environmental footprint. For example, solvents which demand lower energy input, lead to less corrosion, and which undergo minimal thermal or oxidative degradation are being sought [6][7][8] . However, sequestration of captured carbon dioxide is still under debate due to its mid-term solution character, such as its limitation to areas where CO2 can be stored underground and its unprofitability 9,10 . An approach with a strategic long-term vision is CO2 capture and utilization (CCU), where the captured CO2 is used as a feedstock for the production of valuable chemicals. This approach creates value for CO2, leading to new value chains, and forming a stepping stone towards a circular carbon economy while realizing a sustainable way of reducing atmospheric CO2 concentration. To date, studies have approached CO2 capture 5,6,[11][12][13][14] and conversion [15][16][17][18][19][20] independently, whereas the strategy of this study was the integration of these two processes.
Electrochemical conversion of CO2 can be considered an efficient strategy for CO2 utilization due to the mild operating conditions and the possibility of tailor-made reaction pathways 17 . Preliminary techno-economic analysis has determined that the most economically viable products from CO2 are those where only 2 electron transfers are involved, such as carbon monoxide or formate 21 . Nevertheless, several hurdles need to be overcome in order to industrialize the electrochemical conversion of CO2 22 . One of the most significant challenges is related to the poor solubility of CO2 23 in the electrolyte, leading to mass transfer limitations which limit current densities and reduce overall reaction efficiencies 24,25 . Several strategies to increase the productivity of added-value chemicals from CO2 have been investigated, such as the use of gas diffusion electrodes (GDE) [26][27][28] , the use of elevated pressures 29,30 and the use of non-aqueous electrolytes 31,32 in order to increase the availability of the gas reactant at the electrochemical surface. However, from a complete system approach, the CO2 capture step is still required for the overall process. Therefore, this study proposed an integrated approach where CO2 capture and conversion are combined by using the capturing media as an electrolyte for the electrochemical conversion process. The use of CO2 capture solvents as electrolytes can overcome the above-mentioned limitations of a conventional electrochemical system by allowing the absorbed CO2 to be feed into the electrolyzer at high concentration, thereby decreasing mass transfer limitations and increasing overall process efficiency 33 . Additionally, this integrated approach offers the benefit of removing the CO2 purification step -a desorption column -from in the conventional capture process. CO2 desorption is energy intensive and therefore expensive, so substituting this with a direct (electrochemical) conversion step leads to an improvement of the economic feasibility of the overall integrated process 34,35 . Jens et. al. assessed cost saving by integrating CO2 capture with utilization and concluded that with an upstream higher than 30 mol % CO2, 46% savings in energy demand can be achieved when compared to a process without integration 36 . Figure 1 shows a schematic representation of the two possible routes for CCU technologies: conventional decoupled CO2 capture and conversion, and integrated CO2 capture and conversion. In the frame of CCU, the integration of CO2 capture with CO2 conversion must be addressed at the early development stage since a synergy between them is of paramount importance. Therefore, the selection of a capture medium should be appraised together with the compatibility assessment of that medium as electrolyte for electrochemical conversion. Finding suitable solutions for both capture and conversion processes will lead to a more robust system for the reduction of CO2 emissions.
A limited number of studies have addressed the use of CO2 capturing media for further electrochemical conversion. Chen et al. studied the electrochemical conversion of CO2 in a monoethanolamine (MEA) containing solution 37 . However, a constant feed of gaseous CO2 was required as reduction of the captured CO2 species was not observed. Diaz et al. reported, for the first time, the reduction of CO2 to CO in switchable polarity solvents which can also be used for CO2 capture 38 . However, this study showed low faradaic efficiencies towards CO (ca. 20 %), partially due to the aqueous nature of the solvent.
This study proposes a novel methodology that integrates CO2 capture and conversion using a mixture of chemical and physical CO2 absorption solvents as electrolytes. This presents the benefit of the abatement of continuous CO2 feeding as a source of free CO2. More specifically, it comprises the reduction of CO2 towards formic acid, glycolic acid, oxalic acid and carbon monoxide in a mixture of 2-amino-2-methyl-1-propanol (AMP) and propylene carbonate (PC) solution. The presented technology comprises the in situ liberation of the captured CO2 in the electrolyzer by increasing the temperature of the electrolyte in the vicinity of the electrode. This technology utilizes the inherent elevated temperature that is reached in electrolyzers at industrial scale due to the ohmic losses of the electrochemical system, thereby benefiting from what at first consideration might be perceived a drawback in the process.

Proposed system
The proposed integrated CO2 capture and conversion system comprises of three main steps: absorption of CO2 in the chemical capture solvent, liberation of the absorbed CO2 inside the electrolyzer, and electrochemical conversion of the in-situ liberated CO2.
The most developed and extensively utilized capture solvents for CO2 are those based on thermally regenerable solvents such as MEA or AMP, among others 39 . In the capture step, CO2 reacts exothermically with non-hindered primary and secondary amines to form a carbamatespecies according to Eq 1. In the presence of water, due to the steric hindrance of AMP, bicarbonate is formed according to Eq 2.
In the absence of water, for each mol of AMP used, 0.5 mol of carbamate is formed, resulting in a theoretical CO2 absorption of 22 g/L when a 1 M AMP solution is used 40 . This CO2 loading capacity is considerably higher than the absorption capacity of aqueous solutions at ambient conditions (ca. 1.5 gr/L).
In a conventional CO2 capture unit, the enriched CO2 solution is introduced into a stripper where, by the addition of thermal energy, the complexation is reversed, leading to the liberation of CO2 and the generation of lean capture solvent. However, in the proposed technology, the solvent with the carbamate/bicarbonate species is feed into the cathodic side of an electrochemical reactor where the captured CO2 is desorbed in the vicinity of the electrode surface, as a result of the equilibrium reaction (eq 1). In order to promote the in situ liberation of CO2, the electrochemical reactor is maintained at an elevated temperature ( ~75 °C). This mild temperature increase can easily be reached inside industrial size electrochemical reactors due to the associated ohmic losses of the electrochemical process which translate into heat.
The liberated CO2 is then electrochemically converted to formic acid, carbon monoxide or oxalic acid on the cathode according to Eq. 3-5, thereby shifting the equilibria in equation (1) back to the regeneration of the amine.
2 H2O à O2 + 4H + + 4 e - (6) According to these equations, the presence of protons is essential for CO2 conversion. This can arise from two different sources. Firstly, in the trace amounts of water in the catholyte as a result of proton transfer through the cationic exchange membrane (CEM) that separates anolyte and catholyte compartments. The water content before and during electrolysis was measured with titration and it was found to increase with time due to diffusion through the membrane (See section 1 in the supporting information). Higher water content was measured in the catholyte when electrolysis experiments were carried out at 75 °C versus 20 °C (See Figure S.1 in the Supporting Information). Secondly, the protonated amine counterion which is formed in situ with the carbamate species. This was proposed by Chen et.al. 37 for CO2 reduction in MEA according to Eq. 7. However, it should be noted that if protons are abundant in solution, the competitive Hydrogen Evolution Reaction (HER) might be expected to dominate and the selectivity towards the desired products might be negatively affected.
Following reaction, the regenerated amine solution with residual unconverted CO2 exiting the electrolyzer can then be introduced in the absorber column to capture CO2, thereby recycling the capture solvent. It is important to note that in the proposed system where a liquid product is formed (e.g. formic acid) prior recycling of the capture solvent a separation and purification step is needed in order to remove the CO2 reduction product and traces of water from the solvent.

Results and discussion
Cyclic voltammetry measurements were carried out to evaluate the viability of CO2 reduction in the capture media. Figure 2 shows the cyclic voltammograms performed on a Pb electrode in PC solution in the presence and absence of capture media (AMP), and in the presence of captured CO2, at different temperatures.

Figure 2. Cyclic voltammograms of Pb electrode in a 0.7 M TEACl in propylene carbonate solution in the absence of 2amino-2-methyl-1-propanol (AMP) (black line), in the presence of 1 M AMP (red line), in the presence of CO2 captured at 1 bar and 15 °C in 1M AMP (yellow line), in the presence of CO2 captured in 1 M AMP after removal of free CO2 with N2 and at 15 °C (green line), in the presence of CO2 captured in 1 M AMP after removal of free CO2 with N2 and at 45 °C (blue line) and in the presence of CO2 captured in 1 M AMP after removal of free CO2 with N2 and at 75 °C (purple line). Inset displays a zoom in between -1.6 V and -2.5 V vs Ag/AgCl.
When CO2 is added to the electrolyte, a reduction current which was not present in the blank is observed with an onset potential of -2 V vs Ag/AgCl. However, if the electrolyte is flushed with nitrogen gas to remove non-complexed CO2, the current density profile is very similar to that measured for the solution without CO2 (blank). This indicates reduction barely occurs, suggesting that the carbamate/bicarbonate species are not themselves directly reduced. Interestingly, when the temperature of the electrolyte is raised to 45 °C, a reduction current is observed with an onset potential of -2 V vs Ag/AgCl. This result suggests that when the temperature is increased, the desorbed CO2 from the carbamate species undergoes electrochemical reduction. When the temperature of the electrolyte is increased to 75 °C, a considerably higher current is observed with an onset potential of -1.8 V vs Ag/AgCl. The higher current observed at 75 °C compared to 45 °C can be explained by increased CO2 desorption at the higher temperature. It is important to mention that the CO2 preloaded electrolyte forms a biphasic system at ambient temperatures, while at temperatures above 60 °C the system is monophasic. The higher solubility of carbamates at higher temperature and the lower viscosity of the solution enhances the mass transfer.
In order to shed light onto the nature of the formed reduction products and to better comprehend the effect of the temperature during the electrochemical reduction of the captured CO2, chronoamperometric studies were carried out at a constant potential of -2.5 V vs Ag/AgCl on a Pb electrode at different temperatures ( Figure 3).
In accordance with the cyclic voltammetry results, the current density measured during electrolysis substantially increased when the electrolyte temperature was increased, rising from -3.9 mA/cm 2 at 15 °C to -25 mA/cm 2 at 75 °C.  In order to assess the importance of the organic solvent (propylene carbonate) as opposed to water, a control experiment was carried out on a 1M AMP aqueous solution (see Figure  40 mA/cm 2 ). However, low faradaic efficiencies towards formic acid were observed (ca. 13 % after 1 h of electrolysis and ca. 3 % after 5 h of electrolysis). The high current densities can be attributed to the competitive reduction reaction, namely HER, due to the increased temperatures promoting this reaction when water is present. As such, exchanging the organic solvent with water results in a significant negative effect on the selectivity of the reaction, reducing the efficiency of formic acid formation by ca. 70%.
The CO2 concentration (α) in the solvent mixture was measured before and during electrolysis using the so-called BaCO3 precipitation method as described by Li et. al. 42 and Santos et. al. 43 . Figure 4 shows the CO2 loading (moles of CO2/moles AMP) as a function of time at different electrolysis temperatures. The maximum CO2 loading achieved prior to reaction was ca. 0.35 mol CO2/mol AMP, decreasing over time as a result of the desorption of CO2 during the equilibrium reaction (1). The CO2 loading decreases sharply in time as the temperature of the electrolyte is increased, indicating a faster CO2 desorption rate at higher temperatures. After 5 h, the CO2 loading is ca. 10% of the initial value when electrolysis is performed at 75 °C and ca. 93.5 % of the initial value when performed at 15 °C. The conversion of the liberated CO2 is shown in Figure 5a and it was calculated as: /0 1 is the moles of residual captured CO2 after a given electrolysis time.
The amount of CO2 liberated (Figure 5b) at 75 °C after 5 h is approximately 7 times higher than at 15 °C. Correspondingly, the amount of CO2 converted after 5 h of electrolysis at 75 °C is 8 times higher than at 15 °C ( Figure 5c). Importantly, the highest conversion after 5 hours of electrolysis (ca. 30 %) was achieved when the electrolyte was kept at 45 °C (Figure 5a) due to there being a better balance between the amount of CO2 liberated and that being converted. It is important to note that the electrolysis was carried out in an open system, therefore, all the desorbed CO2 that didn't react was vented out of the system. Thereby, closing the system and operating it at elevated pressures, would allow non-converted CO2 to be absorbed again increasing the conversion.
The increase of temperature required to desorb the CO2 from the capture solvent might at first seem an extra operational cost for the system, however, when electrolysis is carried out at higher current densities and/or in a stack reactor, the inherent ohmic losses 44 associated with the electrochemical reaction are converted into heat. These are usually seen as a disadvantage, however, with the current integrated CO2 capture and conversion strategy, they can now be advantageously utilized to improve the electrolysis process.

Effect of the capture solvent concentration
The concentration of the capture solvent (AMP) was varied between 0 M and 3 M in order to understand its influence on the electrolyte solution during CO2 reduction. Figure 6 shows the current density, productivity and faradaic efficiency towards formic acid, glycolic acid and oxalic acid, obtained on a Pb electrode during electrolysis at -2.5 V vs Ag/AgCl at 75 °C, as a function of the concentration of AMP in PC, in a CO2 pre-saturated solution of the two. The current density increased from -10 to -25 mA/cm 2 when the concentration of AMP was increased from 0 M to 1 M, reaching a plateau between 1 and 2 M AMP, and decreasing to -20 mA/cm 2 in 3M AMP. The maximum faradaic efficiency (ca. 50%) and productivity (ca. 0.7 mmol . m -2. s -1 ) of formic acid was observed when a concentration of 2 M AMP in PC was used. Interestingly, increasing capture solvent concentration also leads to a decrease in current densities, faradaic efficiency and productivity of formic acid. This lower efficiency may be explained due to the higher viscosity of the 3M AMP solution, giving rise to mass transfer limitations, as well as the lower conductivity of the solution, potentially leading to current limitations (See Table 1 in the Supporting Information). The higher viscosity might be explained due to the higher levels of carbamate due to the higher loading capacity of the solution and the low solubility of the carbamate species in PC.
Importantly, the faradaic efficiency and the productivity of oxalic acid is drastically increased when low levels/no AMP are present in the electrolyte solution. The highest faradaic efficiency towards oxalic acid was 80%, measured after 1 hour of electrolysis in the absence of AMP. However, the faradaic efficiency towards oxalic acid decreased over time reaching ca. 20 % after 5 hours of electrolysis due to the decrease in available CO2 (see Figure S.3 in the Supporting Information). While the results shown in Figure 6 are an average of five hours of electrolysis, faradaic efficiencies varied with time. Detailed information about their evolution over time as a function of the concentration of AMP and water content can be found in section 3 of the Supporting Information.
In general terms, it is clear that the nature of the electrolyte solution and the presence or absence of capture solvent significantly influences the product distribution.

Formic acid and CO formation in a continuous flow cell
In order to make the first steps towards validating the proposed integrated CO2 carbon and conversion methodology for larger-scale production, on route to commercial implementation, the technology was transferred from batch to a scalable, (semi-)continuous reactor concept. Electrochemical reduction of a CO2 saturated 1M AMP in PC solution was carried out using an electrochemical flow cell in which the electrolyte was continuously circulated through the reactor (See Figure 7a). In order to promote desorption of CO2 in the vicinity of the electrode, a heating chamber was placed on the back of the metallic electrode, maintaining an electrode temperature of 75 °C during reaction. The results of these experiments (See Figure 7 b and c), using a Pb electrode, are similar results to those obtained in batch mode (see Figure 6). Only a slight increase in current density and a slight decrease in faradaic efficiency towards formic acid were observed. More details can be found in Section 4 in the Supporting information.
Additionally, a different cathode material (Au) was investigated for CO2 reduction in the (semi-)continuous reactor concept, with the goal of extending the methodology to include the production of alternative products. It is known that in aqueous solutions, Au selectively reduces CO2 to CO 45 . During electrochemical reduction in a 1M AMP in PC solution preloaded with CO2 using a Au cathode, carbon monoxide was observed as the main product. Figure 7 d and e show the current density and the faradaic efficiency towards CO during electrolysis at different potentials. The highest faradaic efficiency towards CO (ca. 45%) was obtained when the electrolysis was carried out at -1.6 V vs Ag/AgCl with a current density of ca. -15 mA/cm 2 .
Further research to optimize the system, including the electrolyte composition and process conditions, is currently being carried out in our laboratories. Nevertheless, these results indicate the viability of implementing the proposed integrated CO2 capture and conversion system in a continuous process and shows its potential for the reduction of captured CO2 towards added-value chemicals such as formic acid or CO.

Conclusions
This study demonstrates the feasibility of an integrated CO2 capture and conversion system, where a mixture of chemical and physical absorption solvents used for CO2 capture can be used as electrolyte in an electrochemical reactor to efficiently convert CO2 into formic acid and CO.
This novel methodology is based on the liberation of the captured CO2 inside the electrochemical reactor by increasing the temperature of the system. An increase of temperature is an inherent trait of an electrochemical system, since the ohmic losses associated to the process lead to an increase of temperature. When a mixture of 1 M AMP in PC was used as CO2 capture media and afterwards as electrolyte, CO2 conversion towards formic acid was achieved with faradaic efficiencies of ca. 45 % and productivity of 0.56 mmol/m 2 ·s when the process was carried out at 75 °C. The productivity of the integrated process significantly improved when increasing the temperature from 15 °C to 75 °C. This increase in productivity is associated to the preponderant liberation of CO2 at elevated temperatures being 8 times higher at 75 °C than at 15 °C.
The selection of an adequate electrolyte composition is essential for the selective formation of the desired product. For instance: substituting the physical solvent (PC) for water lead to a 70 % decrease of formic acid production due to the higher production of hydrogen in aqueous media, while electrolysis in the absence of the capturing media (AMP) lead primarily to oxalic acid formation instead of formic acid. Importantly, doubling the concentration of the capture solvent to 2 M showed an improvement on the efficiency reaching 50 % and an improvement of the productivity of formic acid which reached 0.7 mmol/m 2 ·s. However, higher concentrations of AMP led to a decrease in the formic acid productivity and efficiency, possibly due to mass transfer limitations attributable to higher electrolyte viscosity. Thereby, the nature of the electrolyte and its relative composition play a significant role in the product distribution of the CO2 reduction reaction.
The proposed methodology was brought one step forward into the scaling-up process and the validation of the electrochemical conversion in a semi-continuous system using an flow reactor was achieved for formic acid and for CO. The highest Faradaic efficiency towards CO (ca. 45 %) was achieved at -1.6 V vs Ag/AgCl with a current density of ca. 10 mA/cm 2 .
This study presents a positive perspective in the implementation of the CCU technology by integrating CO2 capture and conversion processes towards added value chemicals, such as formic acid or carbon monoxide.

Experimental
A three-electrode electrochemical H-cell and modular microflow cell (ElectroCell) were used to perform the electrochemical conversion of CO2 for batch and flow experiments, respectively. The H-cell consisted of two compartments, while the flow cell had one extra compartment behind the cathode plate for heating purposes. When working in the H-cell, the working electrode was a Pb coil (10 cm 2 , Alfa Aesar, 99.9%) and the counter electrode was Pt coil (30 cm 2 , Alfa Aesar, 99.9%). When working in the continuous flow cell, the working electrode was Pb foil (10 cm 2 , Alfa Aesar, 99.9 %) for formic acid production and Au foil (10 cm 2 , Alfa Aesar, 99.9%) for CO formation. The counter electrode was a platinized Titanium plate (10 cm 2 , ElectroCell). The reference electrode used in both cells was a leak-free Ag/AgCl electrode (Innovative Instruments LF-1-100). The anode and cathode compartments in both cells were separated by pre-treated cation exchange membrane (Nafion-117). Prior to every experiment, the working and counter electrode were treated. Pb was treated electrochemically in a 0.5 M aqueous H2SO4 by applying -1.8V vs Pt as counter electrode for 500 s. The Pt coil was flame annealed, followed by quenching in Mili-Q water. Au foil was mechanically polished with alumina slurry, followed by sonication to remove trace alumina particles. All the glassware and cell components were rinsed with acetone and dried before using.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupportinginformationIntegratedCO2captureandconversionusingaminebasedsolvents.docx