Switching between Local and Global Aromaticity in a Conjugated Macrocycle for High‐Performance Organic Sodium‐Ion Battery Anodes

Abstract Aromatic organic compounds can be used as electrode materials in rechargeable batteries and are expected to advance the development of both anode and cathode materials for sodium‐ion batteries (SIBs). However, most aromatic organic compounds assessed as anode materials in SIBs to date exhibit significant degradation issues under fast‐charge/discharge conditions and unsatisfying long‐term cycling performance. Now, a molecular design concept is presented for improving the stability of organic compounds for battery electrodes. The molecular design of the investigated compound, [2.2.2.2]paracyclophane‐1,9,17,25‐tetraene (PCT), can stabilize the neutral state by local aromaticity and the doubly reduced state by global aromaticity, resulting in an anode material with extraordinarily stable cycling performance and outstanding performance under fast‐charge/discharge conditions, demonstrating an exciting new path for the development of electrode materials for SIBs and other types of batteries.


Introduction
Aromatic organic compounds hold great promise for becoming the next generation of battery electrode materials owing to their low-cost, environmentally benign, and recyclable nature. [1] Although lithium-ion batteries (LIBs) have been greatly successful for various applications,n ext-generation materials are desirable to reduce the dependence on toxic heavy metals and lithium as well as to increase the freedom in structure and property tuning. [2] Sodium-ion batteries (SIBs) are am uch praised alternative to LIBs,b ut the anode material conventionally used for LIBs,graphite,is inactive for SIBs, [3] which is due to the thermodynamically unfavorable insertion of sodium ions.T odate,the small group of aromatic organic compounds found to be suitable as SIB anode materials largely consists of sodium carboxylates. [4] While promising specific capacities were achieved with these compounds,m ost of them suffered from significant degradation issues when tested under fast-charge/discharge conditions or for long-term cycling. Fundamentally new concepts are needed for solving these issues and for designing stable, high-performance organic SIB anode materials.
Recent fundamental studies of conjugated macrocycles indicate that global (anti)aromaticity effects in organic compounds can lead to promising redox properties for battery applications. [5] Shinokubo et al. investigated ac oncept that employed global aromaticity to stabilize the doubly reduced or oxidized state of ac onjugated macrocycle in order to achieve good redox properties for applications as electrode material in LIBs. [6] Thep orphyrinoid, which they used, can switch between an antiaromatic neutral state,f eaturing am acrocyclic conjugated system of [4n] p-electrons,a nd an aromatic doubly reduced (or oxidized) state of [4n+ +2] pelectrons,o beying Hückelsr ule.T he charged states of the porphyrinoid were stabilized effectively by the global aromaticity,b ut the concept suffers from the inherent destabilizing effect of the global antiaromaticity in the neutral state. Generally,antiaromatic compounds lack meaningful practical applications owing to their inherent instability.T he porphyrinoid used by Shinokubo et al. required steric protection by bulky substituents to counterbalance the destabilizing effect and achieve sufficient stability for testing in batteries.
Herein we explore the concept of switching between al ocally aromatic neutral state (instead of ag lobally antiaromatic neutral state) and ag lobally aromatic doubly reduced state,o vercoming the inherent issues of antiaromaticity and providing an exciting new design concept for organic electrode materials.Instead of bulky substituents (as used for stabilizing the porphyrinoid), we introduced vinylene bridges along the [4n] p-electron system of an antiaromatic conjugated macrocycle known as [24]annulene (Figure 1a left;[ 24]annulene substructure indicated by bold bonds), creating locally aromatic phenylene units with [4n+ +2] p-electrons that we expected to dominate the structure. Surprisingly,a lthough the conjugated macrocycle resulting from this design process, [ 2.2.2.2]paracyclophane- 1,9,17,25tetraene (PCT), was subject to intense fundamental studies in the 1970s and 1980s, [7] our study is the first to explore the concept of switching between local and global aromaticity and to investigate the benefits of the macrocyclic structure for applications as electrode material.

Results and Discussion
PCT was obtained in as ingle step from low-cost starting materials terephthalaldehyde and p-xylylenebis(triphenylphosphonium bromide) by aW ittig reaction (see the Supporting Information for details). Purification of the crude product by preparative gel permeation chromatography (GPC) afforded PCT as ap ure,b right yellow powder in yields of 13 %. As ascalable alternative purification method, sublimation instead of GPC proved to be feasible at moderately reduced pressure (approx. 0.4 mbar) and elevated temperature (240 8 8C). Similar yields of 11 %were obtained by sublimation, with no differences in purity according to 1 HNMR measurements (Supporting Information,Figures S1 and S2). Such as imple,l ow-cost synthesis and purification is rarely achieved for conjugated macrocycles but is important to facilitate preparation on the scale required for battery electrodes.
As af irst step in assessing our molecular design, we investigated its ability to stabilize the neutral state of PCT by determining the thermal properties of the compound. Thermogravimetric analysis (TGA) revealed high thermal stability,w ith the decomposition starting at approx. 290 8 8Ca nd a5%mass loss at 317 8 8C(Supporting Information, Figure S5). Differential scanning calorimetry (DSC) showed am elting point at 245 8 8C( Supporting Information, Figure S6). Neither in the solid state,n or in solution, were any stability issues observed by 1 HNMR measurements.I nc ontrast, [24]annulene (obtained in three steps in an overall yield of approx. 0.1 %) was reported to decompose on attempted melting point determination and almost fully decomposed at room temperature within 24 hours. [8] Theh igh stability of PCT compared to its parent compound [24]annulene confirmed the intended effect of creating locally aromatic units along the [4n] p-electron system on the stability of the compound.
To investigate the effect of the molecular design on the ring currents,wecarried out anisotropy of the induced current density (ACID) calculations [9] as anext step.T he ACID plot of neutral PCT shows slightly disturbed diatropic (aromatic, clockwise) currents on the four phenylene units (Figure 1b  left). No global ring current was observed along the [24]annulene substructure.H ence,a ccording to these calculations, neutral PCT can indeed be regarded as composed of locally aromatic phenylene units with [4n+ +2] p-electrons connected by vinylene units,w ith no significant antiaromatic contribution from the global [4n] p-electron system of the [24]annulene substructure.T his observation is in accordance with the experimentally determined 1 HNMR signals at 7.32 ppm (phenylene) and 6.42 ppm (vinylene). Furthermore,l owtemperature 1 HNMR measurements showed no indication of antiaromatic character on cooling to 193 K( Supporting Information, Figure S4).
Tw ofold reduction of PCT to the corresponding dianion PCT 2À (Figure 1a right) drastically changes the ring current flow.T he ACID plot of PCT 2À shows as trong diatropic ring current along the [24]annulene substructure (Figure 1bright), indicating that the two additional electrons delocalize over the macrocyclic substructure and create ag lobally aromatic [4n+ +2] p-electron system that obeys Hückelsr ule.T he diatropic ring current mainly flows along the perimeter of the macrocycle,c onfirming conclusions previously drawn from 1 HNMR measurements of the dianion, which experimentally indicated its globally aromatic nature. [7d] No local ring currents were observed in the ACID plots of PCT 2À .T he plots effectively visualize the transition from al ocally aromatic neutral state to ag lobally aromatic doubly reduced state.Upon further reduction, the calculations predict astrong paratropic (antiaromatic,c ounter-clockwise) ring current in the tetraanion PCT 4À ,f ollowed by recovery of the diatropic ring current in the hexaanion PCT 6À (Supporting Information, Figure S11). Similar effects are predicted for the corresponding cations (Supporting Information, Figure S12).
To experimentally investigate the reduction of PCT and confirm that aglobal [4n+ +2] p-electron system is formed, we conducted cyclic voltammetry (CV) measurements in solution. Previous CV measurements on ahanging mercury drop electrode indicated that the reduction of PCT in DMF is ar eversible two-electron process. [7c] This conclusion was drawn from the small difference of the cathodic and anodic peak potentials (DE p )o f3 0mV, which is below the thermodynamic limit for ao ne-electron process (57 mV at 25 8 8C). Although we found the reduction to be chemically and practically reversible (Supporting Information, Figure S13, first 30 cycles), it was not thermodynamically reversible on our glassy carbon electrode (DE p = 61 mV). This renders confirmation of the reduction stoichiometry (that is,t he number of electrons transferred per molecule) from DE p or the slope of steady-state voltammograms difficult (also as these parameters are influenced by electrode process kinetics and Ohmic drop). Therefore,wedecided to follow adifferent approach to confirm the stoichiometry and prepared PCT solutions of known concentrations in 0.1m NBu 4 PF 6 in N,N-dimethylformamide (DMF), propylene carbonate (PC) and 1,2-dichloroethane (DCE) for CV using ap latinum disc ultramicroelectrode (UME) of 25 mmd iameter (Supporting Information, Figure S14). Based on the diffusion limited current of the reduction in these measurements,weestimated the diffusion coefficients of PCT in the different solvents, assuming at wo-electron stoichiometry (see the Supporting Information for details). Thediffusion coefficients were then used to estimate the hydrodynamic radii of PCT (0.51 nm in DMF and PC,0 .39 nm in DCE), which were in good agreement with the optimized (non-spherical, cylinder-like) geometry of PCT (radius:c a. 0.65 nm, height:c a. 0.25 nm; Supporting Information, Figure S9), confirming the twoelectron mechanism of the reduction and the formation of ag lobal [4n+ +2] p-electron system in all tested solvents.Like the reduction, the oxidation of PCT was found to be at woelectron process in all three solvents (Supporting Information, Figure S16), but the oxidation was chemically reversible only in DCE (with an estimated redox potential of the PCT/ PCT 2+ couple of 0.77 Vvs. ferrocene/ferrocene + (Fc/Fc + ), see the Supporting Information for details).
Forestimating the kinetic parameters of the reduction, we recorded cyclic voltammograms of the same PCT solutions on alarger platinum disc electrode of 2mmdiameter and fitted simulated voltammograms to the measured voltammograms ( Figure 2). This allowed us to estimate standard electron transfer rate constants (k 0 )of1.8 10 À3 cm s À1 (DMF and PC) and 1.0 10 À3 cm s À1 (DCE) as well as electron transfer coefficients (a)o f0 .17 (DMF), 0.32 (PC) and 0.29 (DCE) (see the Supporting Information for details). Ther ate constants correspond to thermodynamically quasi-reversible cases. [10] Thek inetics are faster than the electroreduction of Li + in DMF (k 0 = 4.7 10 À4 cm s À1 ), [11] but slower than thermodynamically reversible cases,f or example,f errocene oxidation in acetonitrile (k 0 = 8.4 cm s À1 ). [12] Evaluation of the redox potential of the PCT/PCT 2À couple revealed as olvent-dependent potential as low as À2.29 Vv s. Fc/Fc + in DCE. Chemically reversible reduction of aromatic organic compounds at such low potential is very unusual but highly beneficial for application as an anode material in batteries.Alow reduction potential increases the difference between the redox potentials of cathode and anode  Table S2. Other fitted parameters:a)uncompensated resistance (R u ): 100 W,c apacitance of the electric double layer at the working electrode-electrolyte interface (C dl ): 50 mF; b) R u = 100 W, C dl = 3 mF; c) R u = 1000 W, C dl = 10 mF. Temperature: 293.2 K. and, hence,e nables higher operating voltage in full cells. While we attribute the chemical reversibility of the reduction to the stabilization of PCT 2À by global aromaticity/ delocalization of the charges in the macrocycle,t he low redox potential is the result of another exciting feature of the molecular design:P CT is ap ure hydrocarbon with no heteroatoms or functional groups that would shift the redox potential to higher values by withdrawing electron density from the conjugated p-electron system. Such functional groups or heteroatoms are usually required to store carrier ions in the reduced form;the molecular design of PCT solves this conundrum. Forc omparison, the porphyrinoid mentioned in the introduction showed ar edox potential of À0.90 Vv s. Fc/Fc + for the first reduction and À1.67 Vv s. Fc/Fc + for the second reduction (in CH 2 Cl 2 ), [6] which results in alower operating voltage of the full-cell (compared to PCT) but also in av oltage profile with an unfavorable slope over alarge potential range.
In contrast to reductions in solution, counterions (sodium ions in SIBs) need to be able to insert into the solid-state anode material during the reduction (charging) process.T o assess the capability of PCT to host sodium ions,weanalyzed its solid-state packing as anext step.X-ray diffraction (XRD) analysis of single crystals grown from acetic acid solution revealed an extensively disordered structure with two overlapping orientations (termed orientation Aa nd Bh ere) occurring in ar atio of ca. 56:44 (Supporting Information, Figure S17). Interestingly,wefound large voids in the crystal packing,w hich were not occupied by solvent molecules. Assuming that one of the two PCT molecules in the unit cell adopts orientation Aa nd the other molecule orientation B and placing ap robe of the radius of Na + (1.02 )o n ar egularly spaced grid (0.1 spacing) in this unit cell to identify empty space,w ee stimated that 5.8 %( 66.68 3 )o f the unit cell is empty space large enough to hold sodium ions (see Figure 3a). This equals the volume of 15.0 Na + per unit cell or 7.5 Na + per PCT molecule,e asily providing space for the two counterions per molecule expected to insert upon twofold reduction. Assuming an A-only (B-only) orientation, the same analysis indicated that 4.1 %(8.1 %) of the unit cell volume can hold sodium ions (Supporting Information, Figure S19). With aprobe of the radius of Li + (0.76 ), these values increased to 15.5 %(AB), 14.1 %(A-only) and 17.0 % (B-only;S upporting Information, Figure S20).
However,the XRD pattern of the PCT powder used later for electrode preparation (as obtained after purification by GPC), did not match our single-crystal data. Luckily,t he pattern did match with ap olymorph reported in the 1970s (CCDC 1229545 [7b] )( Supporting Information, Figure S26a). Analysis of the crystal structure of this polymorph by the method described above revealed similarly large void volumes of 4.9 %( Na + ;F igure 3b)a nd 14.9 %( Li + ;S upporting Information, Figure S22) of the unit cell. Although the authors reported that the single crystal of this polymorph was obtained from acetic acid solution, powder XRD of ground crystals grown from acetic acid matched well with our newly reported crystal structure (Supporting Information, Figure S26b).
XRD analysis of as ingle crystal that we obtained by sublimation confirmed that large voids are ageneral property of solid-state PCT;v oid volumes of 5.5 %( Na + ;F igure 3c) and 14.9 %( Li + ;S upporting Information, Figure S25) were found in this polymorph. Forc omparison, we analyzed the crystal structure of 1,4-distyrylbenzene (CCDC 921998 [13] ), astructurally related linear compound. Theanalysis revealed significantly smaller void volumes of 0.8 %( Na + ;F igure 3d) and 3.1 %(Li + ;Supporting Information, Figure S27), corroborating our theory that the macrocyclic geometry inhibits dense packing and provides voids for counterions,facilitating the intermolecular diffusion of ions and preventing unfavor- able volume expansion during the charging process.Furthermore,i na ll of the polymorphs,t he distances between macrocycles were found to be larger than 3.7 (Supporting Information, Figures S18, S21, and S24), which was predicted to be the minimum spacing between layers required for sodium insertion (sodiation) in carbon materials. [14] Given this promising combination of properties,wewere excited to finally assess PCT as an anode material in SIBs.As afirst step in the assessment, CV was conducted in the range of 0.01-2.0 Va tas can rate of 1mVs À1 to confirm the electrochemical activity of solid-state PCT in the anode potential range (Figure 4a). Apair of reduction and oxidation peaks was observed at 0.4 and 0.5 Vvs. Na/Na + ,respectively, and these potentials are similar to those of hard carbons,the most popular anode materials in SIBs,w hich undergo reduction involving delocalized p orbitals for sodium ion storage. [15] Thepeak current density increased during the first few cycles,p resumably due to some interfacial activation. From as o-called b-value analysis (Supporting Information, Figure S28), which is based on the reduction and oxidation peak current densities in CV measurements at different scan rates,w ec oncluded that the sodium ion storage with PCT is mainly diffusion-controlled, as the b-values of the reduction and oxidation peaks were 0.56 and 0.64, respectively.A subsequent galvanostatic test at ac urrent density of 200 mA g À1 (Figure 4b)s howed that PCT exhibits as pecific capacity of 133 mAh g À1 at this current density,which is close to its theoretical capacity of 131 mAh g À1 corresponding to twofold reduction and storage of two Na + ions per PCT molecule.Inthis galvanostatic test, the voltage monotonically decreased in the range of 0.8-0.01 V, which is in contrast to the voltage profile of hard carbons that show two distinct sodium ion storage regimes (interlayer storage between the graphene layers and adsorption within the micropores). [15b] Thet wo regimes result in steep and sloppy slopes in the voltage profile,respectively,making the operating voltage of hard carbons less definitive.
To confirm the charge storage mechanism and delocalization of the electrons in the p-conjugated system of the macrocycles,w ec onducted ex situ X-ray photoelectron spectroscopy (XPS) analysis of pristine,s odiated, and desodiated electrodes.I nt he C1sb ranch (Supporting Information, Figure S29a), the C = Cbond disappeared upon sodiation and the CÀCbond was shifted to alower energy on account of the reduction of PCT engaging the p orbitals of the conjugated macrocycle. [16] Thes odiation was also clearly reflected in the appearance of ap eak at 1072.1 eV in the Na 1s branch (Supporting Information, Figure S29b). Besides the observed sodiation, the formation of the solid electrolyte interphase (SEI) layer was detected in the C1sa nd F1s branches for both the sodiated and desodiated states;t he peak of the sodiated electrode at 288.2 eV in the C1sbranch (Supporting Information, Figure S29a) can be assigned to Na 2 CO 3 ,w hereas the peak at 287.4 eV of the desodiated electrode can be assigned to COOR. In asimilar context, NaF was detected as an SEI component at 684.2 eV in F1sbranch (Supporting Information, Figure S29c). After desodiation, the peak corresponding to the CÀCb ond was mostly restored, although the peak assigned to the C=Cbond was not as much, presumably due to the SEI formation.
Forevaluating the electrochemical stability of PCT during sodium (de)insertion, we finally conducted cycling and rate capability tests using electrodes with 30 wt %o fP CT as the active material. As shown in Figure 4c,P CT showed extraordinarily stable cycling performance without any capacity fading at all over 500 cycles when measured at 2C (1 C = 100 mA g À1 ), confirming that our molecular design effectively stabilized the electrode in both the neutral and reduced state and that solubility of the compound is not an issue (see also the Supporting Information, Figure S30). As the initial changes in the CV measurements,weattributed the observed gradual initial capacity increase to interfacial activation, but an in-depth analysis is required to clarify further. Stable cycling performance was also observed when testing electrodes with ahigh weight content (50 wt %) of PCT (Supporting Information, Figure S31). Therate capability test (Figure 4d) further revealed outstanding performance under fast-charge/ discharge conditions.PCT exhibited capacity retentions of 81, 64, 48, and 27 %w ith respect to its initial capacity of 148 mAh g À1 at 1C when the C-rate was increased to 2C, 5C,1 0C,a nd 20 C, respectively.T he corresponding voltage profiles are presented in the Supporting Information, Figure S32. When the C-rate was returned to 1C,t he capacity was recovered to 161 mAh g À1 ,v erifying the very robust nature of PCT under high C-rates.T he higher capacity at 1C than the theoretical capacity of PCT is attributed to some capacity of the conductive agent denka black (50 wt %) used in the electrode (the voltage profiles and capacity retention of denka black are provided in the Supporting Information, Figure S33).

Conclusion
Our results show that stabilizing the charged state of conjugated macrocycles by global aromaticity is av ery effective strategy to obtain high-performance organic battery electrode materials,ifthe stability of the neutral state is also considered. Designing macrocycles that are globally aromatic in the charged state in such away that their neutral state can be stabilized by local aromaticity,aswedid on the example of [2.2.2.2]paracyclophane- 1,9,17,, can prevent destabilizing global antiaromaticity in the neutral state and result in highly stable organic compounds for battery electrodes.A sar esult of this molecular design, the compounds can switch between as table locally aromatic and as table globally aromatic state and, thus,s how excellent redox properties even without introducing functional groups or heteroatoms,a sw ec ould demonstrate in cyclic voltammetry (CV) measurements of PCT.Assessment of PCT as an anode material in sodium-ion batteries (SIBs), where the exceptionally low reduction potential of À2.29 Vv s. ferrocene/ferrocene + (Fc/Fc + )o ft he compound is of particular benefit, confirmed that the molecular design concept can afford organic electrode materials with excellent performance under fast-charge/discharge conditions and without capacity fading over hundreds of cycles.T he assessment also revealed that the two-electron nature of the reduction has abeneficial effect on the voltage profile of the electrode.W ecan further conclude from our results that the macrocyclic geometry of PCT leads to voids in the solid-state packing capable of hosting sodium ions,w hich is considered to facilitate the insertion of ions during the charging process and may further explain the excellent performance of the material as an SIB anode.
Thes tepwise approach in assessing PCT set out in this work can serve as at emplate for designing and assessing conjugated macrocycles for battery electrodes,with relatively simple anisotropy of the induced current density (ACID) calculations giving ag ood indication of the capability of the compounds to switch between local and global aromaticity before experiments are performed.