Electrochemical Capture and Release of CO2 in Aqueous Electrolytes Using an Organic Semiconductor Electrode.

Developing efficient methods for capture and controlled release of carbon dioxide is crucial to any carbon capture and utilization technology. Herein we present an approach using an organic semiconductor electrode to electrochemically capture dissolved CO2 in aqueous electrolytes. The process relies on electrochemical reduction of a thin film of a naphthalene bisimide derivative, 2,7-bis(4-(2-(2-ethylhexyl)thiazol-4-yl)phenyl)benzo[lmn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone (NBIT). This molecule is specifically tailored to afford one-electron reversible and one-electron quasi-reversible reduction in aqueous conditions while not dissolving or degrading. The reduced NBIT reacts with CO2 to form a stable semicarbonate salt, which can be subsequently oxidized electrochemically to release CO2. The semicarbonate structure is confirmed by in situ IR spectroelectrochemistry. This process of capturing and releasing carbon dioxide can be realized in an oxygen-free environment under ambient pressure and temperature, with uptake efficiency for CO2 capture of ∼2.3 mmol g-1. This is on par with the best solution-phase amine chemical capture technologies available today.

The effects of anthropogenic carbon dioxide present a challenge to scientists in many fields today. Successful sequestration and utilization technologies both rely on the key steps of, controlled capture, storage, and release of carbon dioxide (CO2). Several approaches exist in order to capture and release CO2 on demand. The most developed technology relies on amin-based capture agents which can capture CO2 efficiently and release it back via a temperature swing process.[1] These processes are complicated by the necessity of energy-intensive heating and management of large volumes of chemicals. This has inspired work on solid-state capturing materials. Porous graphitic compounds, polymeric resins were also investigated[2−6] and metal organic frameworks (MOFs)[7] have been optimized with respect to their ability of storing carbon dioxide.[8−10] A remarkable CO2 uptake efficiency of 33.5 mmol g–1 sorbent at 40 atm was achieved using Zn-based MOF.[9]

Despite having high uptake efficiencies, most of the CO2 storing materials mentioned above only perform at high pressure and/or temperature conditions for capturing and releasing of carbon dioxide. The use of redox processes leading to a direct capture and release of CO2 at low temperatures and ambient pressures is still mostly uncharted. Earlier works exploring the redox properties for capturing CO2 dates back to 1971 in which Reddy and co-workers made CO2 react with benzaniline electrochemically.[11] This was followed by studies of electrochemical carboxylation of unsaturated ketones in 1984,[12] Later, Wrighton and Mizen reported the electrochemical reaction of CO2 with 9,10-phenanthrenquinone to yield a biscarbonate dianion,[13] which then was supported via theoretical calculations.[14] Buttry and co-workers also combined the use of organic small molecules with electrochemistry and explored the CO2 capture performance of 4,4′-bipyridine and reported the step-by-step thermodynamical breakdown of cycles leading to electrochemical trapping and release of carbon dioxide.[15] The same group recently reported the use of benzylthiolate, which is generated electrochemically by reduction of benzyldisulfide, as a CO2 capturing agent.[16] We have recently built on the earlier carbonyl redox work by studying the electrochemical addition of CO2 to an industrial carbonyl pigment quinacridone. The quinacridone was deposited on a conducting substrate to form an organic semiconductor electrode which could, upon reduction, capture CO2 in acetonitrile electrolyte, which was subsequently released by oxidation. This process had an uptake efficiency of 4.6 mmol g–1 sorbent.[17] To this end, we introduce a new molecule, 2,7-bis(4-(2-(2-ethylhexyl)thiazol-4-yl)phenyl)benzo[lmn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone (NBIT) (Figure a), which is able to efficiently capture CO2 from aqueous environment and by forming a semicarbonate structure. Synthetic details can be found in the Supporting Information. The naphthalene bisimide core was chosen as an electron-deficient carbonyl pigment unit. To further lower its reduction potential to ensure reduction of the carbonyl moieties in aqueous electrolytes at potentials more positive than the onset of proton reduction, para-phenylene thiazole substituents were added. This, however, might be a trade-off because with the lowering of reduction potential might eventually affect the nucleophilic character of the enolate forming upon reduction.[18] Finally, 2-ethylhexyl termination was chosen to provide resistance to dissolution in aqueous electrolyte upon repeated cycling. NBIT gave an uptake efficiency of ∼2.3 mmol.g–1 in aqueous electrolytes (at 1 atm and 22 °C in 0.1 M Na2SO4 in water), representing a substantial step forward over our previous efforts.

Figure 1

(a) Chemical structure of NBIT. (b) Cyclic voltammetry (CV) of NBIT film on a glassy carbon working electrode (WE). Electrolyte was 0.1 M Na2SO4 in water. Pt foil was used as counter electrode (CE), whereas Ag/AgCl (3 M KCl) was utilized reference electrode (CE). Inset: Uptake of CO2 upon electrochemical reduction of NBIT. (c) Cyclic stability of NBIT under N2.

Cycling between 0 and −1 V in 0.1 M Na2SO4 in water under nitrogen atmosphere (N2), NBIT film first yields a broad double-peak with its minor maximum around −0.80 V and its major maximum around −0.85 V (Figure S1). Upon subsequent cycling NBIT shows two redox peaks (Figure b, red curve). At around −0.70 V the compound shows a quasi-reversible one-electron reduction and at around −0.85 V a second one-electron reversible reduction is observed. These peaks are distinctive of many other carbonyl pigments and have been reported previously.[19−23] The molecule can withstand 20 cycles under N2 atmosphere without showing any change in its electrochemical characteristics and up to 30 cycles with a 30% decrease in current density (Figure c). When the electrolyte solution is saturated with CO2 and the potential is cycled, a complete loss of the peaks is observed (Figure b blue curve). The stability under CO2 atmosphere is different than that of N2. The molecule can withstand 30 times cycling of the potential (Figure S2). The word stability here refers to the stability of the pigment as a compact film on the surface. Due to their nature carbonyl pigments tend to get solubilized upon reduction or oxidation. In the case of N2 atmosphere the molecule becomes soluble upon repetitive cycling under reducing conditions. We can observe the reduced molecule in solution upon UV–vis measurement.[24]

This phenomenon was observed previously for another carbonyl pigment, Quinacridone, and published by our group[17] signaling that the binding of CO2 to the carbonyl group, to form a carbonate-like structure, is taking place. Reduction of quinacridone and related carbonyl pigments is known to take place at the carbonyl functional groups. The same behavior of these oxygens accommodating negative charges is expected for NBIT. To support this, Gaussian 09 was employed for basic quantum-chemical calculations using a density functional theory (DFT)/ HF-based method with the B3LYP hybrid functional[25] and the 3-21g basis set. For simplicity only the naphthalene bisimide core with phenyl rings attached to the nitrogen atoms was considered for the geometry optimization and frequency calculation. The calculation converged and no imaginary frequencies were obtained. The graphic representation of the electronic frontier orbitals shows, that the lowest unoccupied molecular orbital (LUMO) is distributed over the large π-system of the naphthalene core including the carbonyl groups, whereas the highest occupied molecular orbital (HOMO) is more localized on the imide nitrogens and the attached phenyl substituents (Figure S3).

To assess the electrochemical nature of the disappearance of the redox peaks, we conducted electrochemical impedance spectroscopy (EIS) on NBIT films (Figure ). EIS was performed after three CV cycles to ensure quasi-reversible conditions. Spectra were collected in the potential range between 0 and −1.2 V. A step size of 0.2 V was applied. Each potential was kept constant for 5 min to ensure steady state conditions before the impedance measurement, ranging from 500 kHz to 100 mHz with a peak amplitude of ±10 mV. Measurements were performed under Ar and CO2 saturated conditions.

Figure 2

(a) CV of NBIT film on a glassy carbon electrode (WE) with colored points indicating the potentials of EIS data acquisition shown in b and c. Electrolyte was 0.1 M Na2SO4 in water. Nyquist plots of NBIT film on a glassy carbon electrode (b) under Ar and (c) under CO2 respectively.

From the EIS data, we conclude that the ohmic resistance of the electrolyte solution (Rs) and the double layer capacitance (Cdl) are almost equal in all measurements, whereas the charge transfers resistance (Rct) changes substantially. Ohmic resistances from 15.5 Ω to 17.0 Ω are measured for the NBIT film under Ar and under CO2, respectively, whereas the Cdl is found to range from 26 to 71 μF cm–2. The charge transfer resistance (Rct) of the NBIT film reduction is strongly dependent on the atmosphere present, and it decreases significantly with the applied decreasing potential. Although under Ar atmosphere, at high potentials between 0.0 V and −0.6 V, the Rct is expected to be very large and could not be measured within the given frequency range, it is found to be 2864 Ω at −1.0 V and 537 Ω at −1.2 V, respectively. Under CO2 atmosphere the Rct changes from 3213 Ω at −0.6 V to 746 Ω at −1.0 V and finally to 96 Ω at −1.2 V.

Even though the measurements under Ar saturation can be well-explained by a single R/C element in series with the Rs (Figure S6a), two R/C elements have to be used, in series with the Rs, to fit the spectra under CO2 saturation at −0.8 V and below, indicating the presence of two time constants (Figure S6b). The first one may be attributed to the interface between the glassy carbon support and the NBIT film, while the second one is indicative for the interface between the NBIT film and the electrolyte. Because of the high resistance of the NBIT film under Ar atmosphere, the second interface to the electrolyte is not resolved with the performed EIS measurement. In general, the overall Rct decreases with applied negative potential under both, Ar and CO2 atmosphere, possibly through improved electronic conductivity of NBIT and most likely due to cathodic charge transfer at ≤ −1.0 V (Figure S6b). A detailed list of all elementsand their corresponding values used to fit the spectra, can be found in Table S1.

In order to confirm the assignment of the carbonate structure, as the species arising upon reduction and reaction of NBIT with CO2, schematically shown in the inset of Figure b, we performed in situ attenuated total reflection Fourier transform infrared spectroelectrochemistry (ATR-FTIR-SPEC) measurements on the NBIT films cast on germanium (Ge) reflection element. In this configuration, the Ge crystal acts as both, WE and the reflection element, whereas a Pt plate and a silver chloride-coated silver wire act as CE and quasi RE respectively (Figure ).

Figure 3

(a) IR spectrum of NBIT film under different applied potentials. (b) Schematic representation of the spectroelectrochemical cell.

Figure a shows the IR characteristics of NBIT under a constant flow of CO2 saturated Na2SO4 electrolyte (flow rate = 3 mL min–1) at different applied potentials. Prior to applying an electrochemical potential an IR spectrum was recorded as the baseline. As the potential of the WE reached −0.8 V, where the first reduction was observed according to the CV measurements, changes in the IR spectrum started to become observable. Evidence for formation of a secondary-alcohol functionality can be observed around ∼3700 cm–1, which decreases to yield a peak around ∼3300 cm–1 corresponding to carboxylic acid-like – OH functionality. This behavior is attributed to the formation of carbonate-like structure upon addition of CO2 to the now-activated carbonyl group. Moreover, the peak around ∼1670–1680 cm–1, corresponding to the imide carbonyl, decreases as a new peak at 1277 cm–1 rises, exhibiting formation of carbonate-like structure.[24,26]

Finally, we determined the uptake efficiency of NBIT, which is the figure of merit for CO2 capture agents, of NBIT to be ∼2.3 mmol CO2 g–1. The amount of CO2 was determined by gas chromatography (GC). After the CO2 is electrochemically captured by the NBIT film from carbon dioxide saturated solution, the electrolyte solution was purged with N2 for 120 min to remove residual CO2 from the solution. Upon oxidation of the film captured CO2 is released. The release of CO2 can be observed at around 0.24 V with a new redox peak and it was seen that the NBIT regained its characteristic peaks (Figure S4). A 2 mL headspace sample was taken with a gastight glass syringe and injected into GC. The amount of NBIT molecules in the film was measured spectrophotometrically as well as electrochemically. The part of the film that was exposed to electrochemical treatment was dried and then dissolved into a UV–vis cuvette using chlorobenzene as the solvent. Absorbance measurement was performed, and the amount of NBIT on the surface was found as 22.8 nmol using molar extinction coefficient, which was determined via a calibration curve (Figure S5). Integration of total charge during one CV measurement can be used to estimate the amount of electrochemically active NBIT on the surface.[27,28] We found this value to be 19.1 nmol, which is in good agreement with the obtained value via spectroscopy. The faradaic efficiency of the process is calculated via dividing the amount of captured CO2 (3.49 × 10–8 mol) with the amount of electrons used for the capture which is 3.82 × 10–8 mol. With this we reached a faradaic efficiency of ∼91% for the process.

In summary, we showed the performance of a naphthalene bisimide derivative, NBIT, as an electrochemical CO2 capturing agent. NBIT achieved a competitive uptake efficiency of 2.3 mmol CO2 g–1, which can be advantageous over state-of-the-art amine based capturing agents in terms of operating conditions although the uptake efficiency is not as high (∼8 mmol g–1).[29] EIS measurements indicate that the process is taking place on the surface of the electrode instead of the bulk of the electrolyte suggesting that CO2 is being anchored in the semiconductor film. In situ IR spectroelectrochemistry supported this, with evidence for formation of a carbonate-like species. In contrast to many carbon capture processes, NBIT can realize the capture and release of CO2 in ambient pressure and temperature. This underscores that carbonyl pigments are organic semiconductors,[17,23,30] with suitable stability for aqueous electrochemical applications, and are thus a promising materials class.