Current methods for CO2 capture and concentration (CCC) are energy intensive due to their reliance on thermal cycles, which are intrinsically Carnot limited in efficiency. In contrast, electrochemically driven CCC (eCCC) can operate at much higher theoretical efficiencies. However, most reported systems are sensitive to O2, precluding their practical use. In order to achieve O2 stable eCCC, we pursued the development of molecular redox carriers with reduction potentials positive of the O2/O2- redox couple. Prior efforts to chemically modify redox carriers to operate at milder potentials resulted in a loss in CO2 binding. To overcome these limitations, we used common alcohols additives to anodically shift the reduction potential of a quinone redox carrier, 2,3,5,6-tetrachloro-p-benzoquinone (TCQ), by up to 350 mV, conferring O2 stability. Intermolecular hydrogen-bonding interactions to the dianion and CO2-bound forms of TCQ were correlated to alcohol pKa to identify ethanol as the optimal additive, as it imparts beneficial changes to both the reduction potential and CO2 binding constant, the two key properties for eCCC redox carriers. We demonstrate a full cycle of eCCC in aerobic simulated flue gas using TCQ and ethanol, two commercially available compounds. Based on the system properties, an estimated minimum of 21 kJ/mol is required to concentrate CO2 from 10% to 100%, or twice as efficient as state-of-the-art thermal amine capture systems and other reported redox carrier-based systems. Furthermore, this approach of using hydrogen-bond donor additives is general and can be used to tailor the redox properties of other quinones/alcohol combinations for specific CO2 capture applications.
Physical Sciences - Article
Oxygen Stable Electrochemical CO2 Capture and Concentration through Alcohol Additives
https://doi.org/10.21203/rs.3.rs-1139257/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Avoiding the most severe climate effects from anthropogenic carbon dioxide (CO2) emissions requires the advancement of CO2 capture and concentration (CCC) technology. Currently, most approaches to CCC use thermal swings, which are energetically inefficient and expensive.1 There are several advantages for using electrochemical methods (eCCC) over thermal-swings. These include independence from Carnot limitations to achieve theoretical efficiencies of up to 100%, operation at ambient temperatures, modular scalability for point source applications, and the use of increasingly economical renewable electricity.2–4
A common approach to eCCC is the use of redox carriers.4–6 Redox carriers have two stable oxidation states, shown as R and Rn– in Scheme 1. In the reduced state (Rn–), the carrier has a high binding constant (K1,CO2) to facilitate CO2 capture from dilute streams. In the oxidized state (R), the carrier has a low binding constant (K2,CO2) allowing for CO2 release and concentration. Several classes of redox-active carriers have been investigated for eCCC applications including bipyridines,7–9 thiols,10 and quinones.11–18 However, eCCC systems generally degrade from aerobic input streams because the reduced carriers react with oxygen (O2) resulting in unproductive carrier oxidation and the generation of superoxide, which can cause destructive radical reactions with the carrier, solvent, or electrolyte (red reaction in Scheme 1).19,20 Since oxygen is present in flue gas and atmospheric CO2 sources, practical eCCC methods must overcome this limitation.
Aerobic stability is possible if E1/2(R/ Rn–) (Ecap in Scheme 1) is positive of the O2/O2●– reduction potential. The range of this reduction potential in a few different solvents is demarcated by the dashed black lines in Figure 1 (experimental measurements are shown in Figure S1). The second key parameter for a redox carrier is its CO2 binding constant (K1,CO2), which must also be optimized for the application. For example, in order to attain >90% capture efficiency (>90% of incoming CO2 is captured by the solution in a single pass), log(K1,CO2) must be greater than ~3.2 and ~5.5 for capture from flue gas and atmospheric concentrations, respectively.11,12,22 The minimum value of log(K1,CO2) needed for flue gas or atmospheric capture is also shown as the dashed green and blue lines, respectively in Figure 1. The shaded green and blue regions indicate the working regimes required for O2 stable eCCC from flue gas or atmospheric resources.
Among reported redox carriers for eCCC, quinones have shown particular promise, with greater CO2 binding constants at milder potentials compared to other organic redox carriers (Figures 1 and S2).4,7–10 Under anaerobic conditions, quinones have performed eCCC from concentrations of less than 1% to greater than 90%.21,23 Quinones are also currently produced at large scales, economical, and easily modified through functionalization. Electronic structure modifications through functionalization of quinones results in a linear free energy relationship (LFER) between the binding constant (log(K1,CO2)) and reduction potential (E1/2) (Figure 1). As a result, all previously explored redox carriers fall outside of the required working regimes for aerobic flue gas and atmospheric capture applications (Figure 1, Table S1).
As the two key properties for a redox carrier cannot be independently tuned through conventional functionalization, we pursued the use of intermolecular hydrogen bonding interactions through alcohol additives to break the LFER. Our studies demonstrate that common alcohol additives result in beneficial changes to the two key properties of a redox carrier – reduction potential and CO2 binding constant. We also describe how hydrogen-bonding interactions are optimizable through the pKa of the alcohol additive. Using this approach, we demonstrate efficient O2-stable eCCC from flue gas concentrations using a commercially available quinone and alcohol. 24–26
The use of alcohol additives described herein provides a facile approach for tuning the redox carrier properties into desirable ranges that are not accessible through traditional molecular functionalization. This approach can be applied to optimize redox carrier properties for different eCCC applications using easily accessible quinone and alcohol combinations.
Reduction of TCQ Under an N2 Atmosphere and effect of added alcohols. Cyclic voltammetry (CV) of 2,3,5,6-tetrachloro-p-benzoquinone (TCQ) in the absence of hydrogen-bonding interaction under an N2 atmosphere exhibits two reversible, one-electron reductions in polar aprotic solvents, such as dimethylformamide (DMF, black trace Figure 2A), dimethylsulfoxide, (DMSO, Figure S3), acetonitrile (CH3CN), and benzonitrile (PhCN).11–13,27 Similar to the studies reported by Linschitz and Gupta,27 the addition of weakly acidic alcohols (pKa(H2O) ≥ 12.5) to solutions of TCQ in DMF or DMSO under an N2 atmosphere shifts the second reduction event (E(TCQ●–/ TCQ2–)1/2) to more positive potentials, while the first reduction event (TCQ/ TCQ●–) remains unaffected (Figure 2a). Eight alcohol additives (2, 2, 2-trifluoroethanol (TFE), 2, 2, 2-tribromoethanol (TBE), ethylene glycol (EG), 2-methoxyethanol (2-ME), ethanol (EtOH), hexanol (HexOH), 2-propanol (i-PrOH), and tert-butanol (t-BuOH)) were used in this study and all resulted in anodic shifts in potential, with some in excess of 350 mV (Figure 2a). For all alcohols, the magnitude of the anodic shift increases with alcohol concentration; however, at each concentration the shift is larger for alcohols with lower pKa values (Figure S4-S6, and Table S2 pKa values are shown in Table 1). Both reduction events remain reversible with alcohol additives at all scan rates measured (ν = 10-1000 mV/sec, Figure S7 and Table S3). Cyclic voltammograms of chemically prepared [TBA]2[TCQ2–] in DMF with 2M ethanol are nearly identical to that of TCQ in the same solvent mixture, with no evidence of decomposition after several hours (Figure S8).
Hydrogen-bonding interactions with EG, t-BuOH, and EtOH and the reduced TCQ species were further investigated using UV-visible spectroscopy and spectroelectrochemistry (UV-vis SEC). UV-vis SEC of TCQ in the absence of hydrogen-bond donors in DMF indicates that the quinone (TCQ), radical anion (TCQ●–), dianion (TCQ2–), and doubly-protonated dianion (TCQH2) have distinct absorbance features (Figure S9). The addition of alcohols results in a blue-shift (decrease in wavelength) in the absorbance of TCQ2- (Figure 2b), while the absorbance spectra of TCQ, TCQ●-, and TCQH2 are unaffected (Figure S10, S11). Further, this change in λmax for TCQ2- is larger for alcohols with lower pKa (Figure 2b). However, while the shift in the absorption spectra of TCQ2- indicates that the alcohols interact with the dianion, they do not correspond to TCQH2, signifying TCQ2- is not being protonated.28 UV-visible spectra of chemically synthesized [TBA]2[TCQ2–] has the same features and do not change over several hours (Figure S9). The spectroscopic, electrochemical, and spectroelectrochemical data all indicate that the alcohols hydrogen-bond to TCQ2–, resulting in anodic shifts to the E(TCQ●–/ TCQ2–)1/2 potential ‒ in some cases positive of the O2/O2●– reduction potential in DMF ‒ without resulting in protonation or decomposition.
CO2 Binding to TCQ Dianion and effect of alcohol additives.
Under an N2 or CO2 atmosphere, no changes occur in the UV-vis SEC absorption spectra of TCQ and singly reduced TCQ●–, confirming that neither species reacts with CO2. In contrast the absorption spectra of TCQ2– is significantly blue-shifted in the presence of CO2 (dotted versus solid traces, Figure S11), indicating CO2 binding. Similar to other quinones, TCQ2– reacts with CO2 to form an aryl carbonate species (Scheme 2)12,13, which we have characterized by 13C NMR (Figure S13)18 and Fourier-transform infrared (FTIR) spectroscopy (Figure S14).15,18
UV-vis SEC studies with added ethylene glycol, tert-butanol, and ethanol also result in no changes to the electronic absorption spectra between an N2 and CO2 atmosphere for TCQ and TCQ●–. However, the absorption corresponding to [TCQ(CO2)]2– is further blue-shifted in the presence of alcohol additives (Figure 3b). Of the three alcohols, only ethylene glycol appears to protonate TCQ2–to form TCQH2, which was not observed under an N2 atmosphere. The electronic absorption spectrum of chemically synthesized [TBA]2[TCQ2–] in 2M ethanol in DMF under CO2 is nearly identical to TCQ2– generated electrochemically under the same conditions (Figure S15).
CO2 binding at TCQ2– is also evident by comparing the cyclic voltammograms of TCQ recorded under N2 and CO2 atmospheres. While the first reduction to TCQ●– is identical under both conditions, the second reduction features an anodic shift in E1/2 in the presence of CO2, indicating an ErErCr event, or two reversible electron transfer events followed by a rapid reversible chemical step (CO2 binding, Figure 3a).29,30 The addition of alcohol additives under CO2 results in a further anodic shift in the second reduction event, E(TCQ●–/ TCQ2–)1/2 (Figure 3a). The magnitude of the anodic shift correlates with increasing concentration and decreasing pKa of the alcohol (Figures 3, S16, and S17). These results indicate that the alcohols also hydrogen-bond to the CO2 adduct of TCQ2– ([TCQ(CO2)]2–). Chemically synthesized [TBA]2[TCQ2–] in the presence of ethanol under a CO2 atmosphere feature CVs that are almost identical to solutions of TCQ under the same conditions, even after sitting in solution for several hours (Figure S18). CVs and electronic absorption spectra of a [TBA]2[TCQ(CO2)2–] solution after addition of ethanol are indistinguishable from those when CO2 is added to an ethanol-containing solution of [TBA]2[TCQ2–], which verifies that the same product is formed and establishes that CO2 can insert into the hydrogen-bonds formed between ethanol and TCQ2– in solution.
Once formed, TCQ(CO2)2– can be oxidized, resulting in the loss of bound CO2. UV-vis SEC with solutions of [TBA]2[TCQ2–] in pure DMF or 2M EtOH in DMF show quantitative conversion into TCQ●– or TCQ upon one or two-electron oxidation under both N2 and CO2 atmospheres (Figure S19 and SI for experimental details). These features are also present in the CVs of chemically synthesized [TBA]2[TCQ2–] recorded in the presence or absence of EtOH under N2 (Figure S8) or CO2 atmosphere, when the species is TCQ(CO2)2–(Figure S18).
The anodic shift of E(TCQ●–/ TCQ2–)1/2 under CO2 versus N2 atmosphere can be used to calculate the CO2 binding constant (K1,CO2) of TCQ2– using equation 1 (see additional Equation 1 discussion in the SI).31
In equation 1, R is the gas constant, T is temperature, F is Faraday’s constant, and n is the number of electrons being passed in the redox event (n = 1 for E(TCQ●–/ TCQ2–)1/2). The number of CO2 molecules that are bound is represented by the term q (which was previously determined to be one for TCQ).12 E°′ and E1/2 are the half-wave potential in the absence of CO2 and in the presence of a known CO2 concentration in solution ([CO2]), respectively. Using this method, the log(KCO2) of TCQ2– is 3.7 ± 0.2 in the absence of hydrogen-bond donors in DMF.12
Comparison of the E(TCQ●–/ TCQ2–)1/2 under N2 and CO2 atmospheres with alcohol additives were used to measure KCO2 according to equation 1. KCO2 steadily decreases (decreasing ΔE) with increasing concentrations of 5 of the 8 alcohols investigated (trifluoroethanol, tribromoethanol, ethylene glycol, 2-propanol, or tert-butanol) in both DMF and DMSO (Figure S20, S21). Thus, even though these alcohols shift E°′(TCQ●–/ TCQ2–) into the desired aerobic operating regime (green region, Figure 1), K1,CO2 also decreases. As a result, these alcohols do not perform better than other functionalized quinones without hydrogen-bonding (Figure 1) for flue gas capture applications. However, three of the alcohols (ethanol, hexanol, and 2-methoxyethanol) exhibit the opposite behavior, with increased values of KCO2 at higher alcohol concentration (Figure S22). The increased CO2 binding affinities at milder potentials afforded by these additives shifts TCQ into the desired operating regime for flue gas capture (black square vs. orange star, Figure 1), successfully breaking the LFER to make them viable candidates for eCCC from flue gas concentrations containing O2.
Spectrophotometric experiments were used to independently verify the electrochemically-derived KCO2 values. Titration of [TBA]2[TCQ2–] with CO2 was monitored using electronic absorption spectroscopy. [TBA]2[TCQ2–] quantitatively converts into TCQ(CO2)2– upon addition of 1 equivalent of CO2 in pure DMF and 2M ethanol in DMF solutions (Figure S23). For each of the titrations, KCO2 was calculated from the disappearance of the absorption corresponding to TCQ2– using the Benesi-Hildebrand method (Figure S23).32,33 From this data, TCQ2– has values of log(KCO2) = 2.97 ± 0.04 and 3.4 ± 0.2 in pure DMF and 2M ethanol in DMF, respectively. While these values are slightly lower than those measured using voltametric methods (log(KCO2) = 3.7 ± 0.2 and 4.3 ± 0.2 for DMF and 2M ethanol in DMF, respectively) they confirm the trend found using electrochemical methods. Importantly, the CO2 binding constant measured by both methods in the presence of ethanol is sufficient for capture from flue gas concentrations.
Optimizing the Hydrogen-bonding Interactions with Reduced TCQ Species
In order to advance the development of air-stable eCCC carriers, it is important to understand why addition of ethanol, hexanol, and 2-methoxyethanol break the LFER for TCQ, while the other alcohols (trifluoroethanol, tribromoethanol, ethylene glycol, 2-propanol, and tert-butanol) do not. Figure 4 shows a general reaction coordinate diagram for quinone reduction and binding to one molecule of carbon dioxide based on an EEC mechanism.12,13 Species A is the doubly reduced quinone dianion and species B is its CO2 adduct. The interaction of a hydrogen-bond donor lowers the energy of both species A and B. The magnitude in which the two key redox carrier properties (E1/2 and K1,CO2) change relies on the relative energy changes of A and B. If the energies of species A and B are reduced by the same degree, there should be little to no effect on KCO2, as it is directly dictated by the energy difference between the two species, however the reduction potential will be anodically shifted. As a result, proper tuning of intermolecular hydrogen-bonding interactions with both species A and B is required to break the LFER shown in Figure 1 and permit stability under aerobic conditions.
We hypothesized that the drop in KCO2 with increasing alcohol concentration for trifluoroethanol, tribromoethanol, ethylene glycol, 2-propanol, and tert-butanol was due to disproportionate stabilization of TCQ2– versus [TCQ(CO2)]2– (species A vs B in Figure 4). To evaluate this hypothesis, the strength of the hydrogen-bonding interactions between TCQ2– and [TCQ(CO2)]2– with each alcohol were quantified using cyclic voltammetry. The reduction of TCQ●– to TCQ2–, and consequent hydrogen-bond formation with n number of hydrogen-bond donors (HB), can be represented by the following chemical steps:
From this series of equilibria and the Nernst equation, the equilibrium constant of the dianion with n molecules of hydrogen-bond donor, HB (KHB(2–)), can be calculated using equation 227 (more detail on equation 2 can be found in the SI).
The difference between KHBCO2 and KHBN2, represented as ΔLog(KHB(2–)) (where ΔLog(KHB(2–)) = KHBN2 – KHBCO2), effectively measures how much more (or less) the energy of TCQ2– is lowered by hydrogen-bonding interactions compared to [TCQ(CO2)]2– (displayed graphically in the reaction coordinate diagram shown in Figure 4). If a hydrogen-bond donor lowers the energy of [TCQ(CO2)]2– more than TCQ2– (ΔKHB(2–) < 0), CO2 binding will be more favorable and KCO2 will increase as the concentration of hydrogen-bond donor in solution increases. Conversely, if the interaction with TCQ2– is too strong(ΔLog(KHB(2–)) > 0), then CO2 binding will be less favored, and KCO2 will decrease.
Table 1 lists values for KHB(2–) and n under N2 (KHBN2) or CO2 (KHBCO2) atmosphere, as well as ΔLog(KHB(2–)) for each of the alcohols investigated (plots used in the determination of each KHB value are shown in Figures S24-S27). As shown in Figure 5A, there is a linear correlation between log(KHBN2)and the pKa34–37 of the alcohol additive, where more acidic alcohols have larger values of KHBN2. Likewise, a plot of log(KHBCO2) versus pKa34–37 shows a similar linear relationship, albeit with a shallower slope (Figure 5B). The relationship between ΔLog(KHB(2–)) and pKa is not linear, but is in fact v-shaped (Figure 5C). ΔLog(KHB(2–)) decreases with increasing pKa up until a certain point (pKa ~ 16), after which ΔLog(KHB(2–)) starts increasing. Three of the alcohols investigated (ethanol, hexanol, and 2-methoxyethanol) have ΔLog(KHB(2–)) values less than zero. This result is consistent with changes in KCO2 with increasing concentration of each of the alcohols investigated. Verifying our hypothesis, the correlation between ΔLog(KHB(2–)) and KCO2 indicates that the hydrogen-bonding interactions of each alcohol with [TCQ(CO2)]2– and TCQ2– needs to be finely tuned to promote CO2 binding at milder potentials. Thus, the alcohols with ΔLog(KHB(2–)) < 0 have pKa’s around 16, while alcohols with either higher or lower pKa have ΔLog(KHB(2–)) > 0, although it is possible that the higher values of ΔLog(KHB(2–)) for 2-propanol and tert-butanol are not a result of pKa, but increased steric hindrance.
Table 1. Equilibrium constants for TCQ2– and TCQ(CO2)2– in DMF and DMSO under CO2 and N2 atmosphere with various alcohol hydrogen-bond donors.
|
|
pKa (H2O)† |
Log(KHBN2) (n) |
Log(KHBCO2) (n) |
ΔLog(KHB(2-)) |
Log(KCO2) (ΔE1/2 ~250 mV) |
|||||
|
tert-butanol (t-BuOH) |
18.0 |
1.1 (1.8)
|
0.51 (1.0) |
0.57 |
(<< 2.4)* |
|||||
|
2-propanol (i-PrOH) |
17.0 |
2.2 (2.2) |
1.9 (1.8) |
0.29 |
(<3.7)* |
|||||
|
|
|
|
|
|
|
|||||
|
Hexanol (HexOH) |
16.1 |
2.6 (2.3) |
3.2 (3.0) |
–0.58 |
(>4.5)* |
|||||
|
|
|
|
|
|
|
|||||
|
Ethanol (EtOH) |
15.9 |
3.4 (2.3) |
4.0 (2.5) |
–0.63 |
4.4 |
|||||
|
|
|
|
|
|
|
|||||
|
2-methoxyethanol (2-ME) |
15.7
|
2.4 (2.3) |
2.9 (2.7) |
–0.53 |
(>4.1)* |
|||||
|
|
|
|
|
|
|
|||||
|
Ethylene Glycol (EG) |
14.2 |
6.9 (2.5) |
4.4a (2.4) |
6.4 (2.6) |
3.6a (2.4) |
0.52 |
0.72a |
3.3 |
3.3a |
|
|
|
|
|
|
|
|
|||||
|
Tribromoethanol (TBE) |
13.4 |
5.5 (2.3) |
3.7 (1.8) |
1.8 |
2.5 |
|||||
|
|
|
|
|
|
|
|||||
|
Trifluoroethanol (TFE) |
12.5 |
7.4 (3.2) |
4.9a (2.2) |
5.6 (2.6) |
3.0a (1.5) |
1.8 |
1.9a |
2.4 |
2.0a |
|
† pKa values obtained from references 34–37.
* ΔE1/22 of 250 mV was not reached at [ROH] ≤ 3 M.
a Determined from measurements taken in DMSO.
In addition to alcohol pKa, solvent identity also affects KHBN2 and KHBCO2. Measured values of KHB(2–) in DMSO are several orders of magnitude lower than those obtained in DMF (TFE and ethylene glycol, Table 1). Lower values of KHB(2–) in DMSO likely arise from its larger solvent donor number compared to DMF.38 This would result in stronger hydrogen-bonds formed between the alcohol and solvent molecules that are more difficult to disrupt to form hydrogen-bonding interactions with TCQ2– or [TCQ(CO2)]2–. This hypothesis is further supported by prior studies, which report KHBN2 values for TCQ2– with ethanol and TFE that are significantly larger in acetonitrile and benzonitrile (which have lower donating abilities than DMF or DMSO)38 than what we measured in DMF and DMSO.27
From the alcohols studied, ethanol is the most promising candidate for use in an eCCC system utilizing TCQ as the redox carrier, due to its favorable values of KHBCO2, KHBN2, and ΔLog(KHB(2–)). Ethanol’s large KHBN2 means that that E(TCQ●–/TCQ2–)1/2 is shifted significantly positive without a large concentration of alcohol. For example, at 2M ethanol concentration, E(TCQ●–/ TCQ2–)1/2 is shifted over 230 mV positive than in the absence of an alcohol (black traces, Figure 6). Importantly, this potential is 200 mV positive of E(O2/O2●–)1/2 (potential at half-maximum current = –1.30 V vs. [Fe(C5H5)2]+/0 at a glassy carbon electrode in pure DMF, or –1.20 V vs. [Fe(C5H5)2]+/0 in DMF containing 2M ethanol, Figure S1 and Figure 6 (below)), thereby avoiding undesirable losses in faradaic efficiency from O2 reduction or carrier decomposition from superoxide generation.19,20 Additionally, KHBCO2 is larger than KHBN2 for ethanol, resulting in a negative ΔLog(KHB(2–)) which indicates that CO2 binding is not weakened by hydrogen-bonding interactions, but is in fact strengthened by them. Altogether, these properties indicate that when combined with a hydrogen-bond donor such as ethanol, TCQ becomes an effective redox carrier for eCCC from aerobic streams at flue gas concentrations.
Bulk CO2 Capture and Release Studies
The combined redox and CO2 binding properties of TCQ2– with ethanol additives prompted us to investigate whether this system could be used to capture and release CO2 from flue gas concentrations in the presence of O2. A closed system was used to complete CO2 capture and release from 10% CO2 sources. The electrolysis was performed using a sealed H-cell similar to the one depicted in Figure 7A (and Figure S29), where the CO2 concentration of the working cell headspace was periodically monitored through an IR CO2 analyzer in a closed loop using a small pump that circulated the headspace. Each cycle was initiated by sparging a solution of chemically synthesized [TBA]2[TCQ2–] in the working compartment with simulated flue gas to form TCQ(CO2)2–. An oxidizing potential was then applied to the working cell solution to release bound CO2 and form TCQ●–. TCQ was used as a sacrificial oxidant in the counter cell to balance charge and limit undesirable crossover effects. Once oxidation was complete, the working cell solution was then reduced to reform TCQ2– and capture the CO2 released in the previous step.
In the absence of ethanol, CO2 capture and release was tested with TCQ using simulated aerobic (87:10:3, N2:CO2:O2 (v/v)) and anaerobic (90:10, N2: CO2 (v/v)) flue gas mixtures. Experimental details and results are described in the SI. Although capture and concentration are observed in both cases, decomposition occurs upon reduction of the carrier, which prevents re-capture of the CO2 released.
When the same capture cycle experiment is performed in the presence of 2M ethanol, the system completes the entire electrochemical capture cycle. Under a 89:8:3, N2:CO2:O2 (v/v) atmosphere, a 50 mM solution of [TBA]2[TCQ2–] in 2M ethanol in DMF captured and concentrate CO2 from 8.0% to 36.3% (v/v), passing 38.4 Coulombs of charge during the oxidation (Figure 7B). The initial and final headspace CO2 concentration and charge passed corresponds to a 95% Faradaic efficiency (FE) and 98% yield of released CO2 based on the moles of TCQ2– in solution.
The solution was then reduced once more until a total charge of –37.8 Coulombs was passed. The CO2 concentration in the headspace decreased from 36.3% to 15% (v/v), resulting in a 73% FE for the re-capture step.
The use of ethanol as an additive to TCQ-based eCCC systems provides enhanced O2 stability, which is essential for practical eCCC. The batch-capture experiments performed with a closed H-cell system is similar to a 3-stage capture system, where the redox carrier is reduced or “activated” in the presence of CO2 before being pumped over to the anodic cell where it is oxidized to release bound CO2.3 The minimum thermodynamic requirement for this type of redox-carrier eCCC system can be estimated from the ΔE between the half-wave potentials of the carrier in the presence of its dilute CO2 inlet stream (in this case, flue gas at 10% CO2 v/v) and of its concentrated outlet stream (100% CO2).2,3,26,39 Cyclic voltammograms of TCQ in the presence and absence of 2M ethanol under 10 and 100% CO2 are shown in Figure 6. In the absence of ethanol, ΔE of the half-wave potentials (conservatively measured as 30mV from the potential at peak current) under 10% CO2 and 100% CO2 (the conditions under capture and release during reduction and oxidation, respectively) is 250 mV. This ΔE value corresponds to a minimum calculated work of 23.3 kJ/mol CO2 captured, or 24% energetic efficiency (based on a minimum theoretical requirement of 5.6 kJ/mol for a 10–100% CO2 concentration swing). At 2M ethanol concentration, ΔE drops to 220 mV between 10 to 100% CO2, lowering the calculated work required to just 21.5 kJ/mol, which raises the calculated maximum energetic efficiency to 26%, which is almost twice as efficient as any other reported eCCC method, or state-of-the-art alkanolamine-based systems.21,23–26,40 We note that these calculations derive a minimum energetic requirement from the redox carrier properties. In a full electrolytic system, other factors would contribute to the operating efficiency including cell overpotentials, carrier solubility, and CO2 solubility, which could all be optimized through cell engineering. Additionally, the CO2 binding constant of the redox carrier can be further optimized for specific CO2 concentrations to achieve even higher efficiencies.
Electrochemical carbon dioxide capture and concentration (eCCC) can achieve significantly higher energy efficiencies than current thermal methods and its modular infrastructure is well-suited for mobile or point source applications. However, eCCC systems have been plagued by oxygen instability. This study describes the use of alcohol additives to modify the properties of a quinone CO2 redox carrier through intermolecular hydrogen-bonding interactions. Optimizing these interactions through alcohol pKa and concentration results in beneficial changes to the redox properties and CO2 binding, the two key parameters of an eCCC redox carrier. With TCQ, the optimal interactions were achieved in 2M ethanol in DMF, with a maximum theoretical efficiency of 26% for concentrating a 10% CO2 stream to 100%. With this combination of commercially available compounds, we demonstrate successful completion of a full cycle of electrochemical CO2 capture and release in the presence of O2 from flue gas concentrations, making a significant advance towards practical eCCC. Because the ideal redox carrier parameters will ultimately depend on the concentration of CO2 in the targeted capture stream, this approach can be further used to generate a library of commercially available quinones and alcohol combinations optimized for application-specific, cost-effective, and scalable eCCC solutions.
Acknowledgments: This research was supported by the Alfred P. Sloan Foundation (JMB, JYY). We thank Alessandra Zito (UCI), Dr. Daniel Bim (UCLA), Prof. Anastassia Alexandrova (UCLA), Lauren Clarke (MIT), Dr. McLain Leonard (MIT), and Prof. Fikile Brushett (MIT) for helpful discussions.
Author contributions:
Conceptualization: JMB, JYY
Methodology: JMB
Investigation: JMB
Funding acquisition: JYY
Project administration: JYY
Supervision: JYY
Writing – original draft: JMB
Writing – review & editing: JYY
Competing interests: Authors declare that they have no competing interests
Materials & Correspondence: [email protected]
Data and materials availability: All data are available in the main text or the supplementary materials.
Supplementary Materials
Materials and Methods
Supplementary Text
Figs. S1 to S30
Tables S1 to S4
References (1-15)
(1) Sanz-Pérez, E. S.; Murdock, C. R.; Didas, S. A.; Jones, C. W. Direct Capture of CO2 from Ambient Air. Chem. Rev. 2016, 116 (19), 11840–11876. https://doi.org/10.1021/acs.chemrev.6b00173.
(2) Stern, M. C.; Simeon, F.; Hammer, T.; Landes, H.; Herzog, H. J.; Alan Hatton, T. Electrochemically Mediated Separation for Carbon Capture. Energy Procedia 2011, 4, 860–867. https://doi.org/https://doi.org/10.1016/j.egypro.2011.01.130.
(3) Shaw, R. A.; Hatton, T. A. Electrochemical CO2 Capture Thermodynamics. Int. J. Greenh. Gas Control 2020, 95, 102878. https://doi.org/https://doi.org/10.1016/j.ijggc.2019.102878.
(4) Rheinhardt, J. H.; Singh, P.; Tarakeshwar, P.; Buttry, D. A. Electrochemical Capture and Release of Carbon Dioxide. ACS Energy Lett. 2017, 2 (2), 454–461. https://doi.org/10.1021/acsenergylett.6b00608.
(5) Renfrew, S. E.; Starr, D. E.; Strasser, P. Electrochemical Approaches toward CO2 Capture and Concentration. ACS Catal. 2020, 10 (21), 13058–13074. https://doi.org/10.1021/acscatal.0c03639.
(6) Sharifian, R.; Wagterveld, R. M.; Digdaya, I. A.; Xiang, C.; Vermaas, D. A. Electrochemical Carbon Dioxide Capture to Close the Carbon Cycle. Energy Environ. Sci. 2021, 14 (2), 781–814. https://doi.org/10.1039/D0EE03382K.
(7) Ranjan, R.; Olson, J.; Singh, P.; Lorance, E. D.; Buttry, D. A.; Gould, I. R. Reversible Electrochemical Trapping of Carbon Dioxide Using 4,4′-Bipyridine That Does Not Require Thermal Activation. J. Phys. Chem. Lett. 2015, 6 (24), 4943–4946. https://doi.org/10.1021/acs.jpclett.5b02220.
(8) Singh, P.; Tarakeshwar, P.; Buttry, D. A. Experimental, Simulation, and Computational Study of the Interaction of Reduced Forms of N-Methyl-4,4’-Bipyridinium with CO2. ChemElectroChem 2020, 7 (2), 469–475. https://doi.org/https://doi.org/10.1002/celc.201901884.
(9) Ishida, H.; Ohba, T.; Yamaguchi, T.; Ohkubo, K. Interaction between CO<SUB>2</SUB> and Electrochemically Reduced Species of N-Propyl-4,4′-Bipyridinium Cation. Chem. Lett. 1994, 23 (5), 905–908. https://doi.org/10.1246/cl.1994.905.
(10) Singh, P.; Rheinhardt, J. H.; Olson, J. Z.; Tarakeshwar, P.; Mujica, V.; Buttry, D. A. Electrochemical Capture and Release of Carbon Dioxide Using a Disulfide–Thiocarbonate Redox Cycle. J. Am. Chem. Soc. 2017, 139 (3), 1033–1036. https://doi.org/10.1021/jacs.6b10806.
(11) Bell, W. L.; Miedaner, A.; Smart, J. C.; DuBois, D. L.; Verostko, C. E. Synthesis and Evaluation of Electroactive CO₂ Carriers. SAE Trans. 1988, 97, 544–552.
(12) DuBois, D. L.; Miedaner, A.; Bell, W.; Smart, J. C. Chapter 4 - ELECTROCHEMICAL CONCENTRATION OF CARBON DIOXIDE. In Electrochemical and Electrocatalytic Reactions of Carbon Dioxide; Sullivan, B. P., Ed.; Elsevier: Amsterdam, 1993; pp 94–117. https://doi.org/https://doi.org/10.1016/B978-0-444-88316-2.50008-5.
(13) Nagaoka, T.; Nishii, N.; Fujii, K.; Ogura, K. Mechanisms of Reductive Addition of CO2 to Quinones in Acetonitrile. J. Electroanal. Chem. 1992, 322 (1), 383–389. https://doi.org/https://doi.org/10.1016/0022-0728(92)80090-Q.
(14) De Sousa Bulhõesw, L. O.; Zara, A. J. The Effect of Carbon Dioxide on the Electroreduction of 1,4-Benzoquinone. J. Electroanal. Chem. Interfacial Electrochem. 1988, 248 (1), 159–165. https://doi.org/https://doi.org/10.1016/0022-0728(88)85158-1.
(15) Qiao, X.; Li, D.; Cheng, L.; Jin, B. Mechanism of Electrochemical Capture of CO2 via Redox Cycle of Chlorinated 1,4-Naphthoquinone in BMIMBF4: An in-Situ FT-IR Spectroelectrochemical Approach. J. Electroanal. Chem. 2019, 845, 126–136. https://doi.org/https://doi.org/10.1016/j.jelechem.2019.05.057.
(16) Namazian, M.; Zare, H. R.; Yousofian-Varzaneh, H. Electrochemical Behavior of Tetrafluoro-p-Benzoquinone at the Presence of Carbon Dioxide: Experimental and Theoretical Studies. Electrochim. Acta 2016, 196, 692–698. https://doi.org/https://doi.org/10.1016/j.electacta.2016.02.159.
(17) Comeau Simpson, T.; Durand, R. R. Reactivity of Carbon Dioxide with Quinones. Electrochim. Acta 1990, 35 (9), 1399–1403. https://doi.org/https://doi.org/10.1016/0013-4686(90)85012-C.
(18) Mizen, M. B.; Wrighton, M. S. Reductive Addition of CO 2 to 9,10‐Phenanthrenequinone. J. Electrochem. Soc. 1989, 136 (4), 941–946. https://doi.org/10.1149/1.2096891.
(19) Boujlel, K.; Simonet, J. On the Electrochemical Reduction of α-Diketones in the Presence of Oxygen. Tetrahedron Lett. 1979, 20 (12), 1063–1066. https://doi.org/https://doi.org/10.1016/S0040-4039(01)87192-6.
(20) Jeziorek, D.; Ossowski, T.; Liwo, A.; Dyl, D.; Nowacka, M.; Woźnicki, W. Theoretical and Electrochemical Study of the Mechanism of Anthraquinone-Mediated One-Electron Reduction of Oxygen: The Involvement of Adducts of Dioxygen Species to Anthraquinones. J. Chem. Soc. Perkin Trans. 2 1997, No. 2, 229–236. https://doi.org/10.1039/A605549D.
(21) Scovazzo, P.; Poshusta, J.; DuBois, D.; Koval, C.; Noble, R. Electrochemical Separation and Concentration of <1% Carbon Dioxide from Nitrogen. J. Electrochem. Soc. 2003, 150 (5), D91. https://doi.org/10.1149/1.1566962.
(22) The Exact Number Depends on the Solvent-Dependent Henry’s Constant for CO2. For Flue Gas (10% CO2), Log(KCO2) Must Be Greater than: 3.0 (DMF), 3.2 (DMSO), or 2.9 (CH3CN). For Atmospheric Concentration, Log(KCO2) Must Be Greater than: 5.5 (DMF, DMSO) or 5 .
(23) Voskian, S.; Hatton, T. A. Faradaic Electro-Swing Reactive Adsorption for CO2 Capture. Energy Environ. Sci. 2019, 12 (12), 3530–3547. https://doi.org/10.1039/C9EE02412C.
(24) Rochelle, G.; Chen, E.; Freeman, S.; Van Wagener, D.; Xu, Q.; Voice, A. Aqueous Piperazine as the New Standard for CO2 Capture Technology. Chem. Eng. J. 2011, 171 (3), 725–733. https://doi.org/https://doi.org/10.1016/j.cej.2011.02.011.
(25) Boot-Handford, M. E.; Abanades, J. C.; Anthony, E. J.; Blunt, M. J.; Brandani, S.; Mac Dowell, N.; Fernández, J. R.; Ferrari, M.-C.; Gross, R.; Hallett, J. P.; Haszeldine, R. S.; Heptonstall, P.; Lyngfelt, A.; Makuch, Z.; Mangano, E.; Porter, R. T. J.; Pourkashanian, M.; Rochelle, G. T.; Shah, N.; Yao, J. G.; Fennell, P. S. Carbon Capture and Storage Update. Energy Environ. Sci. 2014, 7 (1), 130–189. https://doi.org/10.1039/C3EE42350F.
(26) Liu, Y.; Ye, H.-Z.; Diederichsen, K. M.; Van Voorhis, T.; Hatton, T. A. Electrochemically Mediated Carbon Dioxide Separation with Quinone Chemistry in Salt-Concentrated Aqueous Media. Nat. Commun. 2020, 11 (1), 2278. https://doi.org/10.1038/s41467-020-16150-7.
(27) Gupta, N.; Linschitz, H. Hydrogen-Bonding and Protonation Effects in Electrochemistry of Quinones in Aprotic Solvents. J. Am. Chem. Soc. 1997, 119 (27), 6384–6391. https://doi.org/10.1021/ja970028j.
(28) Zhu, X.-Q.; Wang, C.-H.; Liang, H. Scales of Oxidation Potentials, PKa, and BDE of Various Hydroquinones and Catechols in DMSO. J. Org. Chem. 2010, 75 (21), 7240–7257. https://doi.org/10.1021/jo101455m.
(29) Zanello, P. Inorganic Electrochemistry: Theory, Practice and Application; The Royal Society of Chemistry: Cambridge, UK, 2003.
(30) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Application, 2nd ed.; Swain, E., Ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2001.
(31) Schmidt, M. H.; Miskelly, G. M.; Lewis, N. S. Effects of Redox Potential, Steric Configuration, Solvent, and Alkali Metal Cations on the Binding of Carbon Dioxide to Cobalt(I) and Nickel(I) Macrocycles. J. Am. Chem. Soc. 1990, 112 (9), 3420–3426. https://doi.org/10.1021/ja00165a027.
(32) Benesi, H. A.; Hildebrand, J. H. A Spectrophotometric Investigation of the Interaction of Iodine with Aromatic Hydrocarbons. J. Am. Chem. Soc. 1949, 71 (8), 2703–2707. https://doi.org/10.1021/ja01176a030.
(33) Kuntz, I. D.; Gasparro, F. P.; Johnston, M. D.; Taylor, R. P. Molecular Interactions and the Benesi-Hildebrand Equation. J. Am. Chem. Soc. 1968, 90 (18), 4778–4781. https://doi.org/10.1021/ja01020a004.
(34) Chemistry., I. U. of P. and A.; Data., C. on E.; Serjeant, E. P.; Dempsey, B.; Chemistry., I. U. of P. and A.; Data., C. on E. Ionisation Constants of Organic Acids in Aqueous Solution; Pergamon Press: Oxford; New York, 1979.
(35) Reeve, W.; Erikson, C. M.; Aluotto, P. F. A New Method for the Determination of the Relative Acidities of Alcohols in Alcoholic Solutions. The Nucleophilicities and Competitive Reactivities of Alkoxides and Phenoxides. Can. J. Chem. 1979, 57 (20), 2747–2754. https://doi.org/10.1139/v79-444.
(36) Takahashi, S.; Cohen, L. A.; Miller, H. K.; Peake, E. G. Calculation of the PKa Values of Alcohols from .Sigma. Constants and from the Carbonyl Frequencies of Their Esters. J. Org. Chem. 1971, 36 (9), 1205–1209. https://doi.org/10.1021/jo00808a010.
(37) Ballinger, P.; Long, F. A. Acid Ionization Constants of Alcohols. I. Trifluoroethanol in the Solvents H2O and D2O1. J. Am. Chem. Soc. 1959, 81 (5), 1050–1053. https://doi.org/10.1021/ja01514a010.
(38) Gutmann, V. Solvent Effects on the Reactivities of Organometallic Compounds. Coord. Chem. Rev. 1976, 18 (2), 225–255. https://doi.org/https://doi.org/10.1016/S0010-8545(00)82045-7.
(39) Clarke, L. E.; Leonard, M. E.; Hatton, T. A.; Brushett, F. R. Thermodynamic Modeling of CO2 Separation Systems with Soluble, Redox-Active Capture Species. ChemRxiv 2021. https://doi.org/10.33774/chemrxiv-2021-2dqk6.
(40) Rochelle, G. T. Amine Scrubbing for CO2 Capture. Science (80-. ). 2009, 325 (5948), 1652–1654. https://doi.org/10.1126/science.1176731.
Schemes 1-2 are in the supplementary files section.
There is NO Competing Interest.