Multi-Electron Fused Redox Centers in Conjugated Aromatic Organic Compound as a Cathode for Rechargeable Batteries

The straightforward synthetic approach to TAPQ is depicted in Scheme 1.¹² It started from commercially available 1,2-phenylediamine and 2,5-dihydroxy-1,4-benzoquinone which were mixed together and stirred at melting state for 5 hours. Then the produced 5,14-dihydroquinoxalino[2,3-b]phenazine was converted to 5,7,11,14-tetraaza-6,13-pentacenequinone by oxidation with potassium dichromate in sulfuric acid.

Scheme 1. Synthesis of 5,7,12,14-tetraaza-6,13-pentacenequinone.

The electrode was firstly prepared by spreading the slurry of 40 wt% TAPQ, 50 wt% conductive carbon (Super P), and 10 wt% poly-amide imide (PAI) onto the aluminum foil. Then it was dried overnight in the 60°C vacuum oven and punched to circular pieces with a diameter of 14 mm. Electrochemical cells were assembled in an Argon-filled glove box by employing the lithium foil as the counter electrode and the glass fiber as the separator. To inhibiting the dissolution of the reduced organic electrode, a high-salt-concentration electrolyte, that is 4M solution of LiTFSI in tetraethelyneglygol dimethylether (TEGDME), was selected as the electrolyte. Galvanostatic charge-discharge measurement was carried out on a Land CT 2001A battery test system at room temperature. Cyclic voltammetry (CV) tests were performed on an electrochemical workstation CHI660B.

The structure of raw material and electrode was characterized by X-ray diffraction (XRD) (BrukerD8 Advance, Germany) using Cu-Kα radiation (λ = 0.1540 nm) at 40 kV, 40 mA. X-ray photoelectron spectroscopy(XPS) measurement was conducted by a PHI 5300C electron spectroscopy for chemical analysis (ESCA) system with a monochromatic Mg-Kα X-ray source. Fourier transform infrared spectra were recorded with pellets made of active power scraped from the electrode and KBr on an FTIR (Nicolet iS10). For all the ex situ measurement, the electrodes were obtained by disassembling the coin cells at three states (as-prepared, fully discharged to 1.4V and fully recharged to 3.5V in the first cycle). Then the electrodes were rinsed with dimethylether (DME).

Geometry optimization and energy calculation of the molecules were performed using Gaussian 09 quantum chemistry package¹³ based on the Unrestricted B3LYP method¹⁴ and the basis set of 6-31+G(d). The polarizable continuum model (PCM) was employed, with static relative permittivity and dynamic relative permittivity set to 7.9 for solvent TEGDME.¹⁵

The obtained product was characterized by ¹H NMR and MALDI-TOF mass spectroscopy and was found to be consistent with the proposed structure (Fig. 1). The crystal structure of TAPQ was obtained from Cambridge Crystal Structure Database (CCDC: 761663).¹² It has a coplanar structure and hence tends to stack due to the strong π-π interactions. The interlayer distance between two neighboring molecules is 3.32 Å, which facilitates the charge transfer among the molecular layers. Moreover, intermolecular hydrogen bonds can be found between C–H and N/O of two parallel molecules, which forms a three-dimensional network and produces a channel for lithium ions (Fig. 2).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The cyclic voltammetry (CV) properties were initially evaluated between 1.4 and 3.5V vs. Li/Li⁺ at four different scan speeds and the results (Fig. 3a) clearly point to the multiple redox reactions. Four pairs of oxidation/reduction peaks display at 3.0/2.7V, 2.9/2.3V, 2.3/2.0V, 2.0/1.7V, corresponding to the four lithiated processes respectively. For various scanning rates, the peak currents constantly rise along with the increased scan speeds. What's more, there is a linear relation between the peak currents and the square root of sweep rates for both the positive and negative scans, indicating that the redox process of TAPQ is diffusion-controlled (Fig. 3b). For each specific scan rate (Fig. 3c), the curves in three subsequent cycles overlapped quite well without any polarization, correlated with the high reversibility of the cell.

Zoom In Zoom Out Reset image size

The voltage plateaus observed in the galvanostatic charge/discharge profile (Fig. 3d) completely correspond with the CV results. With regard to the TAPQ/Li battery at 2C, it delivered a reversible capacity of 274.6mAh/g in the first cycle and still rendered 164.5 mAh/g after 50 cycles, yielding both a great initial capacity and a high capacity retention of 60%. The dramatic capacity decay was only observed in the first five cycles and the subsequent stabilized capacity approximately corresponds to two electrons reaction. We speculate that although a C=O bond is shared with two adjacent C=N bonds to chelate with lithium cations, one carbonyl group can only stably accommodate one lithium ion because of the steric hindrance effect, just like the phenomenon demonstrated in other quinone-based organic electrode batteries.^16,17 For this reason, only half of its maximum capacity is highly reversible. Much less capacity loss was detected afterwards, and it is mainly due to the dissolution of the organic active material in the electrolyte. Although the electrolyte with ultrahigh salt concentration was selected to decrease the solubility of reduced product, it could not avoid the dissolution completely.

The rate capacities are also investigated (Figs. 4a, 4b) and the battery shows the maximum capacities of 320.0 mAh/g, 290.0 mAh/g, 274.6 mAh/g, 257.2 mAh/g and 186.1 mAh/g under the current levels of 0.1C, 0.5C, 1C, 2C and 5C, respectively. (Theoretical capacity is set at 515.5 mAh/g.) However, for a higher current rate, the reversible capacity is higher as the reaction time per cycle is shorter, and less reduced product dissolute. What's more, it is noted that the cell of TAPQ electrode can obtain a higher capacity compared with the electrode of a similar structure 1,4,5,8-tetraaza-9-10-anthraquinone (TAAQ).¹⁸ Two more benzene rings in TAPQ make it a molecule of higher weight, yet a bigger conjugated backbone enable it to support more charges and accommodate more lithium ions.¹⁹ In addition, the maximum discharge/charge capacity of the TAPQ/Li cell is 320.0/309.0 mAh/g, that is to say, the electrode could take up and release about four lithium atoms per molecule at most. This can be explained by lithium ions correlate with the organic compound via chelating with the carbonyl group and four neighboring conjugated -N = groups. Studies of similar structures can also account for this explanation.^18,20 More interestingly, the cell showed a remarkable cycle performance under a high current loading of 10C, which could maintain a specific capacity of about 150 mAh/g for over 300 cycles with a stable coulombic efficiency of nearly 100% (Fig. 4c).

Zoom In Zoom Out Reset image size

To begin with, we make use of the computational studies to help us get a preliminary understanding of the redox mechanism. Firstly, we did calculations for three situations of [TAPQ]Li and the geometry optimization results confirm that lithium will not singly bound with any atoms. (Fig. 5a) Nitrogen and oxygen atoms work as collaborative centers during the lithium intercalation. They can form a stable five-membered chelate ring when react with alkyl metal ions. This arrangement has been widely used in the coordination chemistry.²¹ Here we take advantage of this model to produce electrochemistry reactions and the formation of chelate cannot only facilitate the reduction progress but stabilize the molecule structure as well.¹⁸ Secondly, we compared the Gibbs Free Energy for the three possible configurations of [TAPQ]Li₂ and found that two lithium atoms prefer to react with the nitrogen and oxygen of the same side so as to obtain the lowest energy (Fig. 5b). For [TAPQ]Li₃ and [TAPQ]Li₄, there are only one possibility of the structure and we find that the planar structure of TAPQ is being kept when reduced. We also noted that Gibbs Free Energy deviations of molecules suddenly elevate after inserting four lithium atoms (Fig. 6, Table I), so the lithiation will be more difficult afterwards. This can be explained by the broken planar structure and decreased stability of the product molecules. The result is also in concern with the mostly four electrons transforming of TAPQ electrode we observed. In addition, we calculated the theoretical potentials of TAPQ upon lithiation and match them with the four reduced processes. (Table II) The redox potentials can be corresponded with the CV curves, though polarization should be considered to some extent. The complete redox mechanism is depicted in Fig. 5c and more details are shown in Table III.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

With the computational results obtained in hand, XPS was then employed to explore the changes of chemical bonding in TAPQ during the battery cycling. In the N1s spectrum(Fig. 7a), the as-prepared electrode showed a peak at 399.6 eV assigned to C=N-C bond.²² After discharging, the original peak diminished and an additional peak with a lower binding energy at 397.5 eV evolved, which could be identified as N-Li bond.²⁰ This indicates that the C=N-C bond also participate in the lithiation process with the formation of lithium-chelate. For the recharged state of the cell, the C=N-C peak recovered and replaced the N-Li peak, implying the reversible electrochemistry of the battery. The O1s spectrum (Fig. 7b) confirmed that the peak of C=O bond (532.2 eV) in the pristine electrode shifted to the position of the C-O bond (531.0 eV) when discharged and then returned to the initial state after the cell was recharged, which are about the same as the commonly observed.²³ The two spectra above testify that nitrogen and oxygen atoms work together to coordinate with the lithium ion, which is in agreement with the previous DFT calculations. In addition, as XPS can only detect the surface characteristic of the sample, we can infer that the surface substance of the electrode is highly reversible during the cycling process. The irreversibility of the cell is largely owning to the incomplete reaction of inner part of the electrode as well as the dissolution problem.

Zoom In Zoom Out Reset image size

Ex situ FTIR analysis was also performed to verify the changes in molecule structures (Fig. 7c). For the discharged state of the battery, the peak at 1197 cm⁻¹ comes into being, corresponding to the emerging of the C-O bond, while the intensity of the C=O peak (1693 cm⁻¹) decreases. This implies that the C=O bond takes part in the redox reaction of TAPQ, forming a chelation between the C-O bond and lithium atoms. This is consistent with XPS data. When the cell was recharged, the vibration mode of C=O bond relatively strengthened while the C-O bond become weak but did not totally disappear as the as-prepared electrode shown, giving rise to the slight irreversibility of the battery.

Alteration in the crystal structure was also investigated by comparing the ex situ X-ray diffraction patterns of the electrodes retrieved in three different cycle conditions (Fig. 6d). We found that the initial crystal structure partially lost upon the first discharge and remained as this afterwards. As we can see in the pattern of the two cycled states, there are only two peaks can be detected obviously. One peak is at 9.7°, corresponding to the (002) plane, which also exist in the as-prepared state but the intensity is weaker. The other is at 13.3°, which is newly evolved and slightly different from the (012) plane in the pristine. Therefore, the crystallinity being kept is owing to the relatively strong intermolecular force of π-π stacking, providing electrochemical activity to the reaction during the battery cycling. While the losing part may be attributed to the distortion of weak hydrogen bonds among π stacks in the lithium interaction progress, in consistent with the poor cycle performance in the former several cycles. The observed amorphousness of the crystalline is very common for organic active materials.^8,20,24

The possibility of utilizing TAPQ electrode is not limitedly based on LIB, it is also electrochemically active to sodium or potassium ion batteries (Fig. 8). The charge-discharge profiles reveal that there are two detectable pairs of voltage plateaus at 2.6/2.3 V and 1.6/1.3 V vs. Na/Na⁺ and also two pairs of voltage plateaus displayed at 2.9/2.5 V and 1.8/1.4 V vs. K/K⁺. This corresponds well with the redox potentials in lithium battery (3.0/2.7 and 2.1/1.7 V vs. Li/Li⁺). The similar curves indicate a similar redox mechanism of alkali ions in TAPQ electrode. While the cycling performance and capacity in sodium and potassium batteries are not as good as lithium battery. The limitation could be attributed to the non-optimized electrolytes and the larger radii of sodium and potassium ions compared with lithium. All in all, the insensitiveness to select the cation of counter electrode makes the TAPQ electrode more useful and its prospect more attractive.

Zoom In Zoom Out Reset image size