Intercalated Organic Redox-active Anions for Enhanced Capacity of Layered Double Hydroxides

A layered double hydroxide of the MgAl type was prepared using a coprecipitation method similar to that described by the usual adapted procedure. ¹²¹ 100 ml of deionized and decarbonated water was used and the preparation was carried out in a 250 ml reactor under a nitrogen atmosphere. Anthraquinone-2-sulfonate was dissolved in this media at 80 °C and the pH was adjusted to 10 with 0.1 M NaOH solution following the mole ratio between AQS and Al³⁺ cations of n_AQS = 4n_Al. 25 ml of the aqueous metal chlorides solution (0.333 mol l⁻¹ in MgCl₂ and 0.167 mol l⁻¹ in AlCl₃) was added dropwise to the reactor at 80 °C and under magnetic stirring for 3 h. The pH of the reaction mixture was kept constant at 9.5 by adding a 1 M NaOH solution. After complete addition of the metal salts, the reaction mixture was left to age at 80 °C for 24 h. The precipitate was then centrifuged and washed three times with hot water. The solid was further dried in an oven for 24 h at 40 °C. The X-ray diffraction spectra of the obtained solid are displayed in S1.

X-ray diffraction analyses were performed using a theta−theta PANalytical X'Pert Pro diffractometer equipped with a Cu anticathode (λKα1 = 1.540598 Å, λKα2 = 1.544426 Å) and a X'Celeretor detector. For the phase identification and refinement of the unit cell parameters of the series samples, patterns were recorded in the range of 2−80°/2θ with a step size of 0.325°/min three times.

Scanning electron microscopy (SEM) images were recorded using a JSM-7500F field-emission scanning electron microscope operating at an acceleration voltage of 3 kV and magnifications of ×1 K, ×20 K, and ×50 K or using a Zeiss MERLIN scanning electron microscope with a magnification of ×10 K and an acceleration voltage of 20 kV. Samples to be imaged were mounted on conductive carbon adhesive tabs and coated with a gold thin layer.

The composite working electrodes used in this study have been prepared as described in Ref. 121 and consisted of 60% active material (LDH, MgAlAQS), 30% conductive additive carbon black (Super Graphite, Superior Graphite Co.) and 10% binder (PTFE, Sigma Aldrich, 60% solution in water). The prepared electrodes resulted in an area of 0.5 cm² with a thickness of around 100 μm and an average mass loading of 9.2 mg cm⁻². They were pressed into a stainless steel current collector in the form of a grid under 500 MPa. For the UV/vis experiments, smaller electrodes with an average area of 0.04 cm² and an average mass loading of 18 mg cm⁻² have been used.

The investigated electrochemical cells were assembled in a 3-electrode beaker type setup, where the working electrode was the LDH composite electrode, the counter electrode was a platinum wire and an Ag/AgCl (3 M NaCl) electrode was used as the reference electrode. All electrodes were submerged into 10 ml of 1 M sodium acetate (NaCH₃COO) aqueous electrolyte.

Electrochemical tests were performed using a multichannel potentiostatic-galvanostatic workstation (BioLogic Science Instruments, VMP3, operated under ECLab software) at room temperature. Before all electrochemical experiments, the cells were set to equilibrium by leaving at open circuit for 1 h. Cyclic voltammetry (CV) measurements were recorded between 0 and –1.2 V vs Ag/AgCl at a scan rate of 1 or 2 mV s⁻¹. Constant current galvanostatic charge/discharge experiments were performed between 0 to—1.2 V vs Ag/AgCl (3 M NaCl) with a current of 0.43 A g⁻¹ (4 C).

Electrochemical CV tests for UV/vis analysis were performed in the same conditions using a single channel potentiostat-galvanostat (PalmSens, PalmSens 4).

UV/vis analysis was carried out using a UV/Vis/NIR spectrophotometer (PerkinElmer LAMBDA 1050) in transmission mode. The absorption maximum of AQS in the electrolyte was located at 330 nm, and absorption spectra were measured in the range between 300–360 nm.

The electrochemical behavior of MgAlAQS electrodes in 1 M NaCH₃COOH aqueous electrolyte during cyclic voltammogram (CV) measurements at a scan rate of 2 mV s⁻¹ in the potential range from 0 to −1.2 V vs Ag/AgCl is shown in Fig. 4. The evolution between the first and second CV are shown in Fig. 4A, since these are of particular interest in this experimental series. The highest current density in this graph is measured for the first reduction of [AQSO₃]⁻ to [AQSO₃]³⁻ at - 1.03 V vs Ag/AgCl with a value of 2.01 A g⁻¹. The related charge determined for this first reduction process is 99 mAh g⁻¹, which is very close to the theoretical capacity of MgAlAQS of 107 mAh g⁻¹ (93%). The corresponding oxidation peak of the first CV cycle is obtained at −0.48 V vs Ag/AgCl with a significant loss in maximum current density and charge at 1.54 A g⁻¹ and 52 mAh g⁻¹. The coulombic efficiency of this first reduction (charge) and oxidation (discharge) couple is close to 50%. This indicates a large hysteresis phenomenon occurring at the working electrode already in the first cycle. In the second cycle, a further decrease in the charge related to both redox peaks, specifically for the reduction process, is also observed. The related capacities are 48 mAh g⁻¹ and 37 mAh g⁻¹ for reduction and oxidation processes respectively, with a coulombic efficiency of 72%.

Zoom In Zoom Out Reset image size

The loss in capacity with increasing number of cycles is further illustrated in Fig. 4B. After 200 cycles, the degradation results in a charge of 7 mAh g⁻¹ for the reduction and 3 mAh g⁻¹ for the oxidation (42% coulombic efficiency). The shape of the CV curves towards the more negative potentials suggests an increased contribution from the hydrogen evolution reaction (HER) with increasing cycle number, which could explain the drop in coulombic efficiency.

The results shown in Fig. 4, present a continuous decrease in capacity, with a pronounced drop on the first cycle. Since nearly all the electroactive AQS molecules in the electrode respond at the first reduction (close to 93% of the theoretical capacity), it suggests that the successive reduction reactions involve AQS molecules which remain electrochemically active, demonstrating that the redox reaction in Fig. 2B is reversible. However, it also appears that a second electrochemical reaction, the hydrogen evolution reaction progressively increases with the number of cycles, probably due to an electro catalytic activation of the stainless steel current collector. The onset of the HER on stainless steel in alkaline solution has been reported to be around −0.28 V vs RHE (i.e. around −0.48 V vs Ag/AgCl/KClsat). ^127,128

The rapid degradation of the MgAlAQS electrode in terms of specific capacity for the charge/discharge process at a current density of 0.43 A g⁻¹, which corresponds to a 4 C rate (full charge or discharge of the electrode within 15 min if the full capacity is utilized), is clearly depicted in Fig. 5. A specific capacity of 100 mAh g⁻¹ is obtained for the first reduction (charge) of the LDH composite electrode, with a potential plateau at −0.78 V vs Ag/AgCl (Fig. 5A), which matches perfectly with the results recorded in the CVs depicted in Fig. 4. However, upon the first re-oxidation (discharge), the active material delivers a specific capacity of only 44 mAh g⁻¹ (ƞ_c = 44%) at −0.6 V vs Ag/AgCl. Therewith, it is obvious that also in the charge/discharge process of MgAlAQS a pronounced decrease in specific capacity occurred right after the first reduction of the material and indicates again a strong loss of redox active sites in the working electrode. In the second charge, a specific capacity of 46 mAh g⁻¹ is measured which is close to the previous discharge capacity. This confirms that the AQS activity lost during the first charge is no longer available for the following cycles. The second discharge delivers 37 mAh g⁻¹ (ƞ_c = 80%) with a loss which is proportionally lower, suggesting that the remaining redox active sites are slightly more stable in the structure of the LDH during cycling.

Zoom In Zoom Out Reset image size

The further cycling behavior of the MgAlAQS electrodes in 1 M NaCH₃COOH in H₂O is shown in Fig. 5B for 200 cycles at the same current density (0.43 A g⁻¹, 4 C). After the significant decrease in specific capacity during the first charge, a continuous decrease is measured. After only 20 cycles, the specific capacity of the working electrode is reduced to 21 mAh g⁻¹ and further decreases to very low last cycle capacities for charge and discharge processes of 6 and 5 mAh g⁻¹ after 200 cycles, respectively. Apart from the first cycle, the coulombic efficiency ranges from 80% to 90%. The values depicted in Fig. 5B are shown more detailed in S2.

Both Figs. 4 and 5 show a strong loss of the redox capacity in CV and constant current cycling especially in the first reduction. In order to investigate this electrode degradation, the surface of pristine and cycled MgAlAQS composite electrodes have been observed by SEM to identify possible changes of the electrode, which could have caused a decline in specific capacity. However, the SEM images of pristine and cycled MgAlAQS electrodes in Figs. 6A and 6B do not reveal any degradation during cycling, since no difference in the composite material can be depicted.

Zoom In Zoom Out Reset image size

This together with the low coulombic efficiencies obtained of the electrodes indicate that the absence of structural electrode film failure, and instead that unwanted side processes during the cycling are most likely responsible for the degradation of the active material. Furthermore, since neither an apparent change of the electrodes nor additional peaks or plateaus in the cycling experiments were observed, a possible explanation is the dissolution of electro-active AQS anions into the electrolyte. To understand this process and to find evidence for such hypothesis, in situ UV/vis measurements were carried out to identify changes in the electrolyte during cycling.

The dissolution of AQS shown by an UV/vis experiment in an in situ beaker type glass cell with a MgAlAQS working electrode after 40 min of CV at 1 mV s⁻¹ (one forward sweep = 20 min, 40 min for a full cycle) between 0 V and −1.2 V vs Ag/AgCl (3 M NaCl) and another MgAlAQS electrode in the same setup during 40 min at open circuit potential (OCP) are compared in Fig. 7A. The quantity of AQS in the aqueous electrolyte was calculated from the increase in the absorption maximum of anthraquinone at 330 nm. The quantity in % refers to the amount of AQS remaining in the LDH host structure vs the initial amount of AQS present in the WE. In the first 10 min, no significant dissolution is observed neither during OCP nor CV measurement. After 10 min, once the potential reaches a critical value on the CV experiment, the evolution of the two curves differs significantly. After 40 min at OCP no significant amount of AQS was found in the electrolyte (100% remaining in the electrode), but a strong dissolution was observed after 10 min of the CV reduction sweep. At this point, a steep decrease of the AQS in the WE is calculated until a plateau is reached at 25% after 25 min of CV, i.e. a strong increase of AQS in the electrolyte is observed up to 75% of the initial amount intercalated in the electrode (Fig. 7A). At the end of the CV, after 40 min, the quantity of AQS in the electrode does not change significantly in the last 15 min of the experiment (i.e. during the oxidation sweep), and stays at 25%.

Zoom In Zoom Out Reset image size

It is therefore clear that during a CV measurement much more AQS is dissolved in the electrolyte in a very short time in comparison to the amount observed during the OCP test. Figure 7B emphasizes this fact, since here even after 19 d at OCP, there is still 95% of residual AQS in the electrode material, which is much more AQS left in the electrode compared to the CV test.

Based on these results, it is obvious that due to the potential changes during the CV experiment, the majority of the redox active AQS is released from the electrode in the time range between the 10th and the 25th minute. To elucidate this process, Fig. 8 overlaps the potential at the electrode, the time and the current (I vs t) from 0 V to −1.2 V vs Ag/AgCl at a scan rate of 1 mV s⁻¹ with the evolution of the remaining AQS in the working electrode.

Zoom In Zoom Out Reset image size

The monitored loss of active sites vs the potential at the MgAlAQS based electrode is shown in Fig. 8. The two plots highlight that the dissolution of redox active AQS is caused by the reduction of [AQSO₃]⁻ to [AQSO₃]³⁻ at the potential range between −0.6 V and −0.9 V vs Ag/AgCl (between minute 8 and 15). The appearance of reduced [AQSO₃]³⁻ in the electrolyte starts within a short delay after the reduction of the [AQSO₃]⁻ anions at 11 min and continues until a plateau is reached at 25 min, which corresponds to the end of the reduction sweep. From there, the following oxidation sweep does not lead to an increase in AQS in the electrode. The final amount of residual AQS in the WE amounts to 25% after only one CV cycle. This means 75% is dissolved in the electrolyte and no longer accessible for efficient charge storage for the subsequent cycles. It also shows that the remaining 25% reduced AQS that are still in the electrode can be re-oxidized as indicated from the oxidation peak around −0.65 V vs Ag/AgCl. The redox reaction displayed in the unfolded CV of Fig. 8, results in a coulombic efficiency of 23%, which nicely matches the displayed loss in AQS anions after the first reduction. Therefore, it appears that none of the dissolved AQS reacts at the surface of the electrode. In repeated experiments with the same conditions, other values for the remaining percentage of AQS have been obtained (S3). However, in all tests the same trend and the same potential for the release of AQS anions was observed.

The described loss of active sites in the WE is caused by a charge imbalance after the reduction (charging) of the active material. In the discharged state of [Mg₂Al(OH)₆]⁺[AQSO₃]⁻ x 2 H₂O, [AQSO₃]⁻ anions compensate the positive charge of the [Mg₂Al(OH)₆]⁺ layers. In the charged electrode, however, this equilibrium is no more valid due to the presence of [AQSO₃]³⁻. This specie demands three times more positive charge than [AQSO₃]⁻, which can, however, not be provided by the LDH layers. Therefore, to restore the charge equilibrium, [AQSO₃]³⁻ anions need to leave the LDH structure. In theory, this process should result to the release of 66% of the initial quantity of AQS anions out of the electrode to yield to the charged LDH active material [Mg₂Al(OH)₆]⁺([AQSO₃]³⁻)_0.33. This means 33% of AQS is in the LDH electrode after reduction and 66% is dissolved in the electrolyte. In reality, release percentages between 60 and 90% have been measured after 1 CV cycle (40 min) in repetitive experiments (S3). This deviation is possibly explainable by the interaction of the WE with acetate anions from the electrolyte, which are present in abundance, and which can be exchanged with [AQSO₃]³⁻ upon cycling the electrode. Additionally, during oxidation and the corresponding polarization towards more positive potentials, the re-intercalation of [AQSO₃]³⁻ is statistically unlikely, since vastly more acetate anions are available in the electrolyte for this process. This release of AQS anions out of the electrode into the electrolyte explains the evident decrease in specific capacity measured after the first reduction of the electrode.

This means that, although we demonstrated the redox activity of AQS in the structure of MgAlAQS, the system is by nature not stable during charging and more efforts must be dedicated to the stabilization of the capacity upon cycling, mainly by preventing electroactive molecules to escape from the LDH interlayer space.

In this study, we proved the usability of redox active anions intercalated in between LDH layers for energy storage application. In these experiments, the specific capacity of the first reduction/charge of AQS anions in MgAlAQS was measured at the potential of −0.78 V vs Ag/AgCl with a total charge capacity of 100 mAh g⁻¹, which is close to the theoretic capacity and therewith highlights the complete electrochemical activity of the anions immobilized in the LDH interlayer spacing.

However, during all electrochemical experiments, insufficient cycling stability of MgAlAQS was observed, which is caused by the dissolution of charged [AQSO₃]³⁻ into the electrolyte after the first reduction, which has been evidenced by in situ UV/vis experiments. In this process, the charged anion [AQSO₃]³⁻, unlike the discharged [AQSO₃]⁻, causes a charge imbalance in the LDH structure and therewith is forced out of the interlayers of the LDH working electrode. With the release of [AQSO₃]³⁻ for charge compensation, these active sites are not accessible for charge storage during subsequent cycles. In order to exploit the redox activity of AQS anions, however, this species has to be reduced, which will in any case cause the described intrinsically low cycling stability of the presented active material [Mg₂Al(OH)₆]⁺[AQSO₃]⁻ x 2 H₂O.

Therefore, it is necessary to search for other intercalated LDHs, with redox active anions, which show less dissolution during charging or ways to prevent the dissolution of the active species into the electrolyte. We will investigate this approach in future studies. Nevertheless, with the findings of this work a vast variety of possible combinations of LDH structure and redox active groups is opened to be investigated for energy storage application.

This research was funded by the French Research Agency ANR AAPG2020 "LaDHy," ANR-20-CE05–0024–01. Labex STORE-EX (ANR-10-LABX-76–01) is also acknowledged for financial support.