Rechargeable lithium (Li) batteries are the primary energy storage technology,1–3 widely used in portable electronics and electric vehicles.4,5 Commercial Li+ batteries depend on electrolytes composed of free Li salts dissolved in carbonate-type solvents. the cost-effectiveness and high conductivity of this liquid electrolyte solution have resulted in its wide usage. However, batteries with organic liquid electrolytes suffer from safety problems, such as leakage, flammability, and explosion.6–9 In addition, the limited capacity of Li+ batteries is not conducive to the long-term operation of large-size electronic devices or vehicles. Recently, increasing attention has been focused on Li metal batteries (LMBs), which exhibit a very high ideal capacity of 3860 mAh g−1. However, the highly active Li metal anode poses safety concerns, such as the reduction of Li+ to Li0 in the form of dendrites on its surface. These dendrites can penetrate the separators and reach the cathode, causing cell failure. Many researchers have employed various methods to overcome these shortcomings, such as designing a three-dimensional (3D) anode substrate,10 building an artificial solid electrolyte interphase,11 introducing additives to the electrolyte,12 and using solid-state electrolytes.13 Among these methods, the use of solid-state electrolytes is a promising approach because it can mitigate safety concerns and enable the development of high-density cells.14–16
Solid-state electrolytes for LMBs can be classified into two major categories: inorganic and polymer. Inorganic electrolytes include oxides,17,18 sulfides,19 thin-film-type solid-state electrolytes,20 and hydride-21 and halide-type materials.22 These inorganic electrolytes can address the majority of the shortcomings of the Li metal anode, including its safety concerns. However, inorganic electrolytes have numerous drawbacks, such as their instability in an ambient atmosphere, interfacial resistance resulting from their inability to make conformal contact with electrodes, and a decrease in actual capacity due to their electrochemical incompatibility with the Li metal anode.23,24 Compared with inorganic electrolytes, polymer electrolytes are more versatile and adaptable.25,26 Researchers can design the building blocks of polymer electrolytes and control their connectivity. In addition, polymer electrolytes are lightweight, stable, cost-effective, and easy to process for manufacturing purposes. Typical polymer examples include poly(ethylene oxide),27 poly(methyl methacrylate),28 polyvinylidene fluoride,29 and polyacrylonitrile.30 Ionic polymers, which contain anionic moieties, such as imidazolates, are of particular interest. The ring structure of imidazolates, the conjugate base of imidazole, allows facile functionalization and tunability, provides thermal stability, and offers a broad electrochemical window.31,32 Zhang et al.33 reported that polymers containing imidazole moieties coordinate with Lewis acids. The introduction of imidazole improved the solubility of polymers and resulted in a Li+ conductivity (σLi+) of up to 1.30 × 10−3 mS cm−1. Nikolaev et al.34 introduced a series of electron-deficient imidazole ligands into a polymer to investigate the relationship between the imidazole moiety and σLi+. The electron-deficient imidazole heterocycles reduced the total σLi+ but increased the transference number (tLi+) from 0.42 to 0.48. Because of the inherent amphoteric behavior of imidazole, it can both accept and donate protons.31,35,36 Although these first-generation polymer electrolytes have been extensively studied, they exhibit low σLi+ and tLi+.
Porous crystalline polymers, such as covalent organic frameworks (COFs), have recently attracted considerable attention as solid-state electrolytes because of their large accessible porous surface area, uniform spatial distribution, and designable structure.37–39 Hu et al. synthesized a benzimidazole-based ionic COF that exhibited a considerably high σLi+ of up to 7.2 mS cm−1 and a low activation energy of 0.10 eV.31 The weak interaction between Li+ and imidazolate and the well-defined two-dimensional COF structure contributed to the superior Li+ conduction performance. However, the crystalline nature of the COF impeded its use as a battery electrolyte. Jeong et al. reported that a Li-coordinated sulfonate COF exhibited an ionic σLi+ of 2.7 × 10−5 S cm−1, a tLi+ of 0.9, and an activation energy of 0.18 eV.40 The synthesized COFs demonstrated reversible and stable Li plating/stripping cycling under solvent-free conditions. Another important class of polymers is porous organic polymers (POPs), which comprise porous and amorphous networks. POPs are widely used in diverse engineering fields, such as gas separation, water purification, proton conductors for fuel cells, and electrochemical energy storage.41 The advantages of POPs are similar to those of COFs, including a large surface area, a tunable pore size, a designable skeleton, and abundant functional groups.42,43 When ionic functional groups are installed in the porous space of POPs, the resulting ionic POPs (iPOPs) can conduct ions. Moreover, the amorphous nature of POPs with minimal contact resistance can solve the interfacial problems frequently encountered with solid-state batteries.44 The mechanisms through which iPOPs strike a balance among high conductivity, low surface resistance, and good processibility when functioning as solid electrolytes remain to be determined. In particular, the introduction of rigid organic linkers can generate free volumes within the polymer network, facilitating Li+ transport. For instance, rigid aromatic triptycene can contribute to a high internal molecular free volume. Schon et al.45 reported the first example of a highly stable 3D perylene diimide–triptycene framework used as the cathode material for a Li+ battery, which exhibited a capacity of 75.9 mAh g−1 and retained 88.2% of the capacity after 200 cycles. Triptycene played a crucial role in the construction of 3D frameworks in the cathode. However, the use of liquid electrolytes may cause safety problems and the low capacity may limit their practical application.
No study has used triptycene POPs as solid electrolytes. Thus, we hypothesized that anionic 3D triptycene amorphous networks would serve as excellent solid-state Li+ conductors. In addition, we speculated that the combination of ring resonance from imidazolates and homoconjugation from the triptycenes would result in the generation of charge-delocalized networks. This phenomenon would reduce the anionic charge in the network and facilitate the diffusion of Li+ through the network, resulting in improved conduction. Thus, we prepared imidazolate-functionalized triptycene POPs with phenoxides as bridging struts. In this study, we present Li+-coordinated triptycene-based imidazolate POPs (Li+@Trp-Im-O-POPs), which exhibited an σLi+ of 4.38 mS cm−1, a tLi+ of 0.95, and a cell capacity of 114 mAh g−1. In this study, cells were constructed using a LiFePO4 cathode. We synthesized Li+@Trp-Im-O-POPs through the condensation reaction between 2,3,6,7,14,15-hexaaminotriptycene (HAT) and 2,5-dihydroxyterephthalaldehyde (DHTA). These compounds were further deprotonated to coordinate Li+ (Fig. 1; see Supplementary Information for synthesis details). In addition, we prepared control samples from HAT and terephthalaldehyde (TA), denoted as Li+@Trp-Im-POPs, which exhibited a σLi+ of 2.42 mS cm−1, a tLi+ of 0.98, and a cell capacity of 112 mAh g−1. Although the conduction properties of Li+@Trp-Im-POPs were poorer than those of Li+@Trp-Im-O-POPs, these values were still higher than those obtained from ceramic materials (oxides and sulfides), indicating the efficacy of our design strategy, which leverages the porous nature of triptycene and ion conduction from imidazolate through the charge-delocalized network. Furthermore, the Li|Li symmetrical cell, equipped with Li+@Trp-Im-O-POPs as the solid-state electrolyte, was stable. Moreover, Li-plating/stripping experiments could be performed for over 500 h without substantial attenuation. The Li|iPOPs|LiFePO4 cell had an initial capacity of 114 and 112 mAh g−1 for Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs, respectively. In addition, the Li+@Trp-Im-O-POP- and Li+@Trp-Im-POP-based batteries retained 86.7% and 85.8% of the capacity after 200 cycles, respectively.
We synthesized triptycene-based iPOPs through the condensation reaction.46 In brief, DHTA (or TA) solution in DMF was gradually added to homogeneous HAT solution in DMF at −30 ℃ under nitrogen. Subsequently, the temperature was increased to 130 ℃ under air. The yellow, powdery, insoluble triptycene-based polymers were synthesized with yields of 73% and 77% for Trp-Im-O-POPs and Trp-Im-POPs, respectively (Fig. 1; see Supplementary Information for details). Further treatment of the polymers with n-BuLi resulted in the deprotonation and introduction of Li+ to imidazolates and phenoxides, leading to Li+-coordinated iPOPs (Li+@iPOPs; specifically, Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs). The excess n-BuLi present in the iPOPs was removed using dry hexane.
The findings of Fourier transform infrared spectroscopy (FTIR), Brunauer–Emmett–Teller (BET) surface area measurement, thermal gravimetric analysis (TGA), powder X-ray diffraction (PXRD), scanning electron microscopy (SEM), and X-ray photoelectron spectroscopy (XPS) indicated the success of the condensation reaction and the porous and amorphous nature of iPOPs. The FTIR findings of Trp-Im-O-POPs and Trp-Im-POPs confirmed the formation of imidazole rings (Fig. 2a). The aldehyde peaks of TA and DHTA at approximately 1681 cm−1 and the N–H stretching bands of HAT at 3000–3100 cm−1 disappeared after the reaction. New vibration peaks at approximately 1629, 3383, and 3209 cm−1 were attributed to C=C/C=N stretching, N–H stretching, and free N–H and hydrogen-bonded N–H, indicating the formation of imidazole rings. The new peaks at 1613 (C=N in Trp-Im-POPs), 1623 (C=N in Trp-Im-O-POPs), 1432, 1437, and 1497 cm−1 can be attributed to the vibrations of benzimidazole rings.46–48 The BET sorption experiment at 77 K revealed the porosity of the iPOPs (Fig. 2b). The calculated BET surface areas of Trp-Im-O-POPs and Trp-Im-POPs were 192.7 and 205.4 m2 g−1, respectively. The reduction in the BET surface area of Trp-Im-O-POPs may be due to the presence of two hydroxy groups on the linker aromatic rings. The hysteresis of the N2 adsorption isotherm presumably resulted from the flexible nature of organic polymers. In addition, the BET surface areas of Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs were 6.2 and 7.0 m2 g−1, respectively. The introduction of Li+ into the iPOPs led to a decrease in the BET surface area. The synthesized iPOPs exhibited high thermal stability (Fig. 2c). TGA experiments revealed that both Trp-Im-POPs and Trp-Im-O-POPs began to decompose at approximately 300 °C. After Li+ coordination, the thermal stability of both polymers decreased slightly. Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs began to decompose at approximately 220 and 300 °C, respectively. The residual weight fractions of Li+-coordinated polymers at 700 °C were 30.9% and 15.5% for Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs, respectively. These values were higher than those of the base POPs, indicating the introduction of Li+ into the base POPs. PXRD results revealed the amorphous nature of both iPOPs (Fig. 2d). Furthermore, SEM images demonstrated a pure phase polymer with particle sizes ranging from 0.10 to 0.23 μm for both base iPOPs (Fig. 2e, f). No prominent change was noted in the morphology of the Li+-coordinated iPOPs (Fig. S4). XPS results revealed that Li+ coordination was successful (Figs. 2g-i and S5). The shifts in the N 1s and O 1s spectra for the Li+@Trp-Im-O-POPs indicated that the Li+ coordinated with the hydroxide groups and the benzimidazole ring (Figs. 2g-h and S6). In addition, a downshift occurred in the N 1s spectrum of the Li+@Trp-Im-POPs after the introduction of Li+ (Fig. 2i). The results of inductively coupled plasma mass spectrometry indicated that the Li+ loading in the Li+@Trp-Im-POPs and Li+@Trp-Im-O-POPs was 4.2 and 6.1 wt%, respectively (Fig. S7a).
Using iPOPs with a high degree of Li+ coordination, we examined the electrochemical properties of the Li+@iPOPs to evaluate their suitability as solid electrolyte materials. First, Li+@Trp-Im-O-POP or Li+@Trp-Im-POP powders were uniformly ground and pressed in a mold to obtain a solid pellet for electrochemical testing. Then, the samples were solvated with propylene carbonate (PC; 20 wt.% in the pellet), and their ionic conductivity was measured using electrochemical impedance spectroscopy at temperatures ranging from room temperature (r.t.) to 80 °C. The resulting curves exhibited Nyquist behavior with a semicircular curve at high frequencies. These were used to calculate the σLi+ of the materials. We investigated the effect of the amount of PC on σLi+. The optimal amount of PC for both Li+@iPOPs was 20 wt.% (Fig. S8a and S8b; unless otherwise mentioned, samples with 20 wt.% PC were used in all the subsequent studies). In the presence of an optimal amount of PC, the Li+@iPOPs exhibited high σLi+ values of 4.38 and 2.42 mS cm−1 at r.t. for Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs, respectively (Fig. 3a). Among imidazolates, only Li+@Trp-Im-POPs displayed a high σLi+ of 2.42 mS cm−1, indicating the effective transport of Li+ along the anionic triptycene networks.
In addition to the conduction of Li+ along the imidazolates, additional negative charges on the phenoxides resulted in a higher σLi+ of 4.38 mS cm−1, which was 1.81 times higher than the value for Li+@Trp-Im-POPs. We measured the conductivity of the base POPs as the control group. The Trp-Im-O-POPs and Trp-Im-POPs had negligible conductivity (Fig. S7b), suggesting that the coordination of mobile Li+ with the backbones of the iPOPs is necessary for conductivity. In addition, the σLi+ of both Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs increased with the temperature (Fig. S8c and S8d). The activation energy (Ea) was calculated from the linear Arrhenius plot. The activation energy of the Li+@Trp-Im-O-POPs (0.627 eV) was slightly lower than that of the Li+@Trp-Im-POPs (0.643 eV; Fig. 3b). The lower activation energy indicated that Li+ migrated across the Trp-Im-O-POP channel more rapidly and was less dependent on temperature. Moreover, the tLi+ value is another crucial parameter for evaluating ion conductivity in the solid electrolytes. A higher tLi+ value is conducive to the continuous transfer of Li+ selectively and efficiently. The results of the Bruce–Vincent–Evans method indicated that the tLi+ values were 0.95 and 0.98 for the Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs, respectively (Fig. 3c and 3d). Both Li+@iPOPs had considerably higher σLi+ values than did other iPOPs (Table S1 and Fig. 3e). To the best of our knowledge, the σLi+ of 4.38 mS cm−1 is the highest reported value for iPOPs and is comparable to that of the most efficient ionic COFs. Our comparisons were limited to only all- or quasi-solid-state conductors containing plasticizers less than or equal to 20 wt.%.31,51 To date, the single-ion conductive tLi+ of 0.95 and 0.98 are the highest reported values. Furthermore, linear sweep voltammetry measurements indicated that both Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs demonstrated a broad electrochemical window of up to 4.6 and 4.4 V, respectively (Fig. 3f). These results indicate that our Li+@iPOPs are promising materials as solid electrolytes for LMBs.
In our previous study, we demonstrated that homo-conjugated triptycenes with charge-delocalized aromatic rings could facilitate the diffusion of backbone charges, thereby enhancing the mobility of counter ions.58 In this study, we observed a similar phenomenon: the degree of negative charges on the imidazolates and phenoxides decreased throughout the iPOP network, leading to a high σLi+ in the iPOP networks. This phenomenon of charge delocalization is similar to the high mobility of Li+ when paired with bis(trifluoromethanesulfonyl)imide (TFSI), where the six fluorine atoms reduce the charge of imide, resulting in highly conductive Li+. This proven method of charge reduction/delocalization is beneficial for ion conduction because strong electrostatic charges would otherwise immobilize the counter ions. To examine charge diffusion, we performed a series of density functional theory (DFT) calculations (refer to Methods in Supplementary Information for details). For simplicity, the polymers were reduced to representative repeating units with hydrogen terminations (Fig. S9-S10). On the basis of electrostatic potential mapping, we determined that the charge was initially concentrated on the imidazolate (−5.85 eV). This charge dispersed when imidazole was fused to triptycene (−4.79 eV). Conjugation with a benzene ring further reduced the charge to −4.46 eV (Fig. 4a-c). The same trend of charge delocalization was observed for phenoxides. Phenoxides conjugated to two imidazolates had a charge of −15.16 eV. This charge decreased to −11.24 eV when imidazolates were fused to triptycenes (Figs. 4d-e and S3a-b), indicating the importance of introducing triptycene units for facilitating charge diffusion. We calculated the reduction in charge from phenoxides to imidazole-fused triptycenes to examine the charge diffusion phenomenon (Figs. S10 and S11). To estimate the charge carried by each atom in these structures, we performed Bader charge analysis based on DFT calculations (Tables S2–S8 and Figs. S12–S18). These results align with those of electrostatic potential mapping.
We performed Li-plating/stripping experiments to evaluate the impedance and stability between the Li+@POP electrolytes and the metallic Li interface. The time-dependent voltage profile was recorded at 10 μA cm−2, displaying polarization voltages of 30 and 100 mV for Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs, respectively (Fig. 5a and 5d). The small polarization voltages for both Li+@iPOPs indicate a low interfacial impedance between the electrolytes and Li metal. In addition, the voltage curves remained stable for up to 500 h, indicating that POP-based electrolytes had a stable interface with Li metal, preventing Li dendrite growth during such an extended cycle. Both the Li+@Trp-Im-O-POPs and Li+@Trp-Im-POPs exhibited high σLi+ and tLi+, good thermal stability, superior contact, and a stable interface with Li metal. These features make Li+@iPOPs excellent candidates for use as solid electrolytes in solid-state LMBs. We used LiFePO4 as the cathode material and Li+@POPs as the electrolyte. Both of them were paired with Li to assemble LMBs. Charge and discharge experiments were performed at a current density of 0.5 C at r.t. (with voltage ranging from 2.5 to 4 V). The initial capacities of Li+@Trp-Im-O-POP- and Li+@Trp-Im-POP-based batteries were 114 and 112 mAh g−1, respectively (Fig. 5b and 5e). The Li+@Trp-Im-O-POP- and Li+@Trp-Im-POP-based batteries retained 86.7% and 85.8% of the capacity after 200 cycles. Furthermore, the Coulombic efficiency was maintained at 99% for both batteries (Fig. 5c and 5f), indicating that no side reaction occurred at the electrolyte–electrode interface during the testing. Li+@Trp-Im-O-POP- and Li+@Trp-Im-POPs-based batteries exhibited similar performance, suggesting that cathode materials limit their capacity. However, the Li+@Trp-Im-O-POP-based batteries exhibited a higher rate capacity at elevated C rates than did the Li+@Trp-Im-POP-based batteries (Fig. S7). The superior rate performance of the Li+@Trp-Im-O-POP-based battery can be primarily attributed to the higher σLi+ of the electrolyte (4.38 mS cm−1).59
In conclusion, our findings indicated the potential of solid-state electrolytes as a viable solution to address safety problems encountered in LMBs. We designed imidazolate-based iPOPs, Li+@Trp-Im-O-POPs, by using a simple bottom-up synthetic approach. These Li+@Trp-Im-O-POPs exhibited a high σLi+ of 4.38 mS cm−1 and a tLi+ of 0.98. These favorable conductive properties result from the highly delocalized charges over the entire triptycene network, the high density of ion-conducting moieties (phenoxides and imidazolate), and the porous nature of the iPOPs. Moreover, Li+@Trp-Im-O-POPs demonstrated excellent compatibility with Li metal electrodes and Li deposition/stripping cycling performance. When we used Li+@Trp-Im-O-POPs as the electrolyte in a battery with an LiFeO4 cathode, we determined an initial capacity of 114 mAh g−1. This capacity only decreased by 13.3% after 200 cycles. In addition, Li+@Trp-Im-POP-based batteries retained more than 85% capacity after 200 cycles. Our triptycene-based iPOPs demonstrated excellent Li+ conduction and high performance when installed in an LMB as a solid-state electrolyte. This type of charge-delocalized network, mimicking the concept of LiTFSI, is promising for the design of next-generation solid electrolytes.