Skip to main content
NCBI home page
Chemical Science logoLink to Chemical Science
. 2026 Sep 30. Online ahead of print. doi: 10.1039/d6sc06154k

Phosphonium outperforms ammonium poly(ionic liquids) for fast ion transport polymer electrolytes

Kewei Cai a,†, Jhonatan Soto Puelles a,†, Alejandro Herranz Berzosa b, Gabriele Lingua b, David Mecerreyes b,d, Daniel Brandell c, Maria Forsyth a,b,d,✉, Fangfang Chen a,✉
PMCID: PMC13624938  PMID: 42819074

Abstract

Cationic poly(ionic liquids) (polyILs) are promising solid polymer electrolytes, yet the role of cationic charge-centre chemistry in regulating alkali–metal transport and interfacial behaviour remains poorly understood. Here, we computationally investigated a phosphonium-based polyIL at high lithium or sodium concentrations and compared these systems with previously reported ammonium-based analogues. The phosphonium systems demonstrated superior alkali–metal ion diffusion, which was consistent with their high ionic conductivities. Molecular-level analyses show that phosphonium cation chemistry does not necessarily lead to weaker isolated interactions with the anion, as demonstrated here. Rather, it is associated with weaker collective ionic interactions in the condensed phase, contributing to lower Tg, enhanced temperature-dependent free-volume growth, and faster ion dynamics. PolyIL-electrode interface modelling investigation further reveals different ion adsorbed behaviour under increasingly negative electrode polarization. Compared with the ammonium analogue, the phosphonium system exhibits significantly higher populations of Na+ and FSI− near the highest negatively charged electrode surface, which may influence the subsequent SEI formation. Collectively, these molecular insights highlight the phosphonium-based polyILs for achieving enhanced ionic conductivity and motivate future interface design strategies for tuning interfacial performance.

Poly(ionic liquid) electrolytes with phosphonium cations exhibit weaker interionic interactions, enhanced Li+ and Na+ transport, and distinct interfacial chemistry near highly negatively charged metal surfaces compared with their ammonium analogues.graphic file with name d6sc06154k-ga.webp

Introduction

Cationic poly(ionic liquids) (polyILs), also known as polymerized ionic liquids, have emerged as promising polymer electrolyte (PE) materials for solid-state alkali metal batteries due to their ability to provide efficient lithium or sodium ion transport when alkali metal salts are incorporated at high concentrations. In contrast to conventional polyethylene oxide (PEO)-based systems, polyIL-based PEs often exhibit higher metal ion transference numbers, reflecting enhanced metal ion transport relative to other ionic species. This advantage appears to be associated with their positively charged backbones, which electrostatically bind anions and suppress their mobility, ultimately enhancing the relative diffusion of charge carrier ions. This makes polyILs attractive for overcoming the long-standing limitations of low ionic conductivity and poor metal-ion transference numbers in PEs.1,2

We have demonstrated these advantages previously in poly(diallyldimethylammonium) bis(fluorosulfonyl)imide (PDADMA FSI), which has ammonium cations integrated directly into the polymer backbone. Upon the addition of alkali–metal salts, optimal ionic conductivities of 0.7 × 10−4 S cm−1 with LiFSI and 1 × 10−3 S cm−1 with NaFSI salt were observed at 80 °C, accompanied by moderately high Li+ and Na+ transference numbers over 0.5.1,2 These electrolytes also demonstrate stable cycling in lithium and sodium symmetric cells at 80 °C, highlighting their potential for lithium- and sodium–metal batteries. However, further increasing ionic conductivity at lower temperatures (<80 °C) remains a challenge for enabling practical ambient-temperature operation.

Recently, various strategies have been explored to enhance the properties of polyILs through both experimental and computational studies.3 One strategy is to employ mixed anions to improve ionic conductivity. For example, incorporating LiFTFSI into PDADMA FSI increases the ionic conductivity to 1.1 × 10−4 S cm−1 at 80 °C at a 1 : 2 polyIL-to-salt molar ratio, representing a 1.6-fold enhancement relative to the corresponding 1 : 1.5 LiFSI system.4 Similarly, introducing NaTFSI into PDADMA FSI at a 1 : 2 polyIL-to-salt ratio enhances ionic conductivity below 50 °C compared with the single-FSI system at the same salt loading; however, this enhancement diminishes at higher temperatures.5

On the other hand, molecular modelling suggests that Li+ and Na+ transport in polyIL-based systems can be tuned through polycation design. For example, introducing a flexible methoxyethyl side chain enhances both Li+ and Na+ mobility in poly(diallyl methoxyethyl ammonium) bis(trifluoromethanesulfonyl)imide (PN1, 2O1 TFSI), likely through increased volumetric expansion and additional ether-oxygen coordination sites that facilitate Na+ migration without strongly trapping it.6 Experimentally, the protonation form of PDADMA+, i.e. PDADMAH+, has also shown increased ionic conductivity with lithium salt,7 further supporting polycation modification as an effective strategy for enhancing ion transport in polyIL electrolytes.

In the present work, we extend this investigation to alternative polycation chemistries beyond ammonium-based systems. Phosphonium-based poly(ionic liquid)s represent a promising alternative that can outperform their ammonium analogues. In parallel experimental work, poly(diallyldimethylphosphonium bis(fluorosulfonyl)imide) (PDADMP FSI) with a LiFSI salt at 1 : 1.5 molar ratio was shown to exhibit a significant increase in ionic conductivity, with 1.5 × 10−3 S cm−1 achieved at 80 °C, nearly 17-fold higher than that of the corresponding ammonium-based analogue 1 : 1.5 PDADMA FSI : LiFSI.8 In contrast to this experimental study, which mainly focused on the experimental characterisation of PDADMP FSI/LiFSI and a comparison with the mixed-anion LiTFSI system, the present work focuses on how cationic charge-centre chemistry, i.e. ammonium versus phosphonium, regulates collective ionic interactions, structural dynamics, alkali–metal coordination and transport, and interfacial organization in polyIL electrolytes. We first compare neat and lithium salt-containing systems for which experimental data are available. We then extend the investigation to sodium salt-containing systems to predict Na+ transport in PDADMP FSI. We also investigate the factors contributing to fast ion dynamics in PDADMP FSI. Finally, we examine electrolyte–electrode interfaces through preliminary interface modelling, focusing on the interfacial electrolyte composition adsorbed on the negatively charged gold surface and its response to increasing negative surface polarisation, an aspect that remains largely unexplored for salt-containing polyIL electrolytes. Therefore, this provides a molecular basis for understanding and tuning the interfacial composition, which could guide future interface engineering of polyIL-based electrolytes, as previously demonstrated in ionic liquid-based electrolytes.9

Results and discussion

Comparison of neat polyIL systems with ammonium and phosphonium cationic centres

The nature of the cationic charge centre plays a critical role in determining electrolyte properties. For example, quaternary phosphonium-based ionic liquids typically exhibit lower viscosities, higher ionic conductivities,10 and greater thermal stabilities11 than their ammonium counterparts. Neat polyILs incorporating phosphonium or ammonium cations via polymer side chains have been investigated previously, with phosphonium-based systems generally exhibiting higher anion conductivities than their ammonium analogues.12,13 More recently, we developed polyILs in which phosphonium cationic charge centres are integrated directly into the polymer backbone for the first time, and demonstrated that these materials exhibit significantly higher ionic conductivities in the presence of a high concentration of lithium salts compared with their ammonium analogues.8 To elucidate the role of the cationic charge centre in governing ionic interactions and metal ion transport properties, we integrated molecular dynamics (MD) simulations with DFT calculations to provide detailed molecular-scale insight into underlying mechanistic factors. The neat polyILs are first investigated in this section.

The Tg of the two neat polyIL systems were calculated and compared via cooling simulations performed under isothermal–isobaric (NPT) conditions over a range of temperatures, with a cooling rate of 1 K ns−1 (simulation details in SI). The Tg values were determined from the volume–temperature (V–T) relationship by identifying the intersection point of two linearly fitted regions with distinct thermal expansion coefficients (Fig. 1a and Table S1). It should be noted that our theoretical models use equally short polymer chains, which do not reproduce the molecular weights and dispersities of the experimental polymers. These differences should therefore be considered when quantitatively comparing the absolute experimental and MD-derived Tg values. Nevertheless, the two polyIL systems were simulated using directly comparable models, allowing a controlled comparison of the relative differences in Tg.

Fig. 1. (a) Volume–temperature (V–T) profiles obtained from cooling simulations for PDADMA FSI and PDADMP FSI. The simulated Tg values were determined from the intersection of linear fits to the glassy and rubbery regimes. (b) DFT binding energy to separate two pairs of PDADMA FSI or PDADMP FSI in a complex. (c) Radial distribution functions (RDF) calculated between polycations (through nitrogen/phosphorus atom) and anions (through nitrogen atom). (d) Spatial distribution function (SDF) calculated for polycation centre surrounding the anion (FSI−). (e and f) Rotational correlation function (RCF) profiles of two selected vectors in anion (S Created by potrace 1.16, written by Peter Selinger 2001-2019 O bond) and cation (N–C bond).

Fig. 1

Open in a new tab

The MD-derived Tg of PDADMA FSI was approximately 379 K, which is markedly higher than that of PDADMP FSI at 263 K. The experimental Tg of PDADMA FSI is 393 K, in good agreement with the MD prediction. However, the experimental Tg of PDADMP FSI could not be reliably determined from the DSC trace because no clearly resolved glass-transition step was observed.8 Nevertheless, comparison with the corresponding TFSI analogues supports the prediction that PDADMP FSI has a substantially lower Tg. The experimental Tg of PDADMA TFSI is 389 K, close to that of PDADMA FSI, and significantly higher than that of PDADMP TFSI at 313 K.8 Consistently, MD-derived V–T analysis using the same protocol predicts a much higher Tg for PDADMA TFSI (392 K)6 than for PDADMP TFSI (267 K).8 The same trend observed for both the experimental and MD-derived Tg values of the TFSI systems therefore supports the credibility of the lower Tg predicted by MD for PDADMP FSI.

The lower Tg of phosphonium system compared to ammonium system has also been reported in polyIL with cationic centre in the side chains by Long and coworkers.12 They postulated that the higher conductivity and lower Tg observed in phosphonium systems, compared with their ammonium analogues, arises from the electronic structure of the phosphonium cations. These cations feature a positively charged phosphorus centre, in contrast to the neutral to slightly negative nitrogen centre in ammonium cations. The phosphorus centre is accompanied by modestly negative adjacent carbon atoms that partially screen the positive centre and weaken ion-pairing with the counter anion.12 Our simulations support this charge picture in both the MD force field charge parameters and the DFT-derived ChelpG atomic charges. The OPLS_AA force field14 was used for MD simulations, assigning charges of +1.326 and −0.042 to the phosphorus and nitrogen atoms, respectively, before scaling charge by a factor of 0.7. The DFT-derived ChelpG atomic charges for these two atoms in the optimised polycation–FSI ion pairs (Fig. S1) are −0.003 for N and +0.617 for P, respectively. However, the positive phosphorus centre does not weaken FSI− binding in an isolated ion pair. DFT calculations show that the PDADMP–FSI ion pair has a binding energy 3–8 kJ mol−1 more negative than that of PDADMA–FSI, depending on the DFT functional used (Fig. S1). This indicates that isolated ion-pair binding alone cannot explain the lower Tg of the phosphonium system.

Because the electrolyte behaviour in the condensed polymeric phase is governed by collective and cooperative interactions rather than by isolated ion pairs, we further examined cluster models comprising multiple interacting ions to account for condensed-phase effects (Fig. 1b and S1). In these cluster models, the ion dissociation energies in PDADMP–FSI clusters, including both a single FSI− coordinated to two polycations or two interacting polycations–FSI ion pairs, were lower than those of the corresponding PDADMA–FSI clusters. For example, dissociation of one polycation from an FSI-bridged polycation ion pair required 3 kJ mol−1 less energy in the phosphonium system (Fig. S1(b)), while separation of two interacting PDADMP–FSI ion pairs required approximately 10 kJ mol−1 less energy than in the ammonium analogue (Fig. 1b). These results indicate that the reduced Tg of the phosphonium-based polyIL is associated with a weaker collective ionic interaction network in the condensed phase, rather than with weaker isolated polycation–anion interaction.

The following analysis at 353 K focuses on the dynamic behaviour of the two neat polyILs. The mean square displacement (MSD) profiles (Fig. S2) show faster FSI− anion diffusion in the PDADMP system, as indicated by the higher slope of the MSD vs. time plot. Spatial distribution function (SDF) analysis further reveals a larger and more continuous distribution of phosphonium cationic centres around FSI− anions compared with the corresponding ammonium nitrogen centres (Fig. 1d). This behaviour is indicative of enhanced cation reorientation motion and/or faster FSI− rotational motion, either of which could give rise to the observed characteristics. Consistently, rotational correlation function (RCF) analysis of selected cation and anion vectors (S Created by potrace 1.16, written by Peter Selinger 2001-2019 O in anions, N–C or P–C and C–C in cations) shows enhanced rotational mobility in the PDADMP system, as evidenced by the more rapid decay of the RCF profiles in Fig. 1e, f and S3. Taken together, these results indicate enhanced dynamical flexibility in the phosphonium-based system relative to the ammonium analogue. This difference is consistent with their different physical states at 353 K: PDADMP FSI is in a rubbery state while PDADMA FSI is near or in a glassy state.

Comparison of PDADMA and PDADMP systems at high lithium and sodium salt concentration

In this section, we investigate alkali–metal ion transport in PDADMP FSI systems using representative polyIL-to-salt molar ratios. Previous studies on PDADMA FSI showed that when the salt concentration exceeds that of the polyIL repeat unit, the Tg significantly decreases and ionic conductivity increases, with optimal conductivities observed at the 1 : 1.5 polyIL-to-salt molar ratio for LiFSI and 1 : 2 for NaFSI.1,2 Interestingly, NMR measurements showed even higher D7Li and D19F diffusivities at the 1 : 2 LiFSI salt concentration,1 indicating faster Li+ and FSI− motion in the amorphous phase of the 1 : 2 system. In addition, in our recent experimental study of PDADMP FSI/LiFSI system, the 1 : 1, 1 : 1.5 and 1 : 2 compositions were also examined.8 Although the 1 : 1.5 composition is experimentally important for the lithium system, the 1 : 1 and 1 : 2 compositions were selected here to represent lower and higher salt-loading regimes that bracket the optimal composition, while also enabling direct comparison with the sodium system, for which the 1 : 2 composition was the main focus. These simulations therefore provide molecular-level insight into the concentration-dependent changes in ion coordination and transport across both LiFSI and NaFSI systems.

The Tg was simulated first (Fig. S4a and Table S1), showing a decrease upon the addition of either lithium or sodium salts to PDADMP FSI. This Tg depression highlights a cooperative plasticization effect induced by the high salt concentration,15 consistent with experimentally reported trends for the PDADMA FSI system (Table S2). For both lithium- and sodium-containing systems, the simulations further indicate that PDADMP-based electrolytes exhibit consistently lower Tg values than their corresponding PDADMA-based counterparts. However, when comparing lithium and sodium systems, simulations show contrasting results to experimental Tg for PDADMA systems, where sodium systems exhibit higher Tg than lithium systems. This discrepancy likely reflects limitations in the current force field parametrization or Tg estimation methodology and indicates that further refinement may be required to capture subtle cation-specific effects. Nevertheless, the simulations successfully reproduce the qualitative influence of salt concentration and polycation chemistry on electrolyte Tg.

Ion transport was next examined using mean-squared displacement (MSD) analysis (Fig. 2a) for a qualitative comparison between the PDADMA and PDADMP systems. Self-diffusion coefficients were also estimated from the MSD data (Table S3). For both lithium and sodium systems, PDADMP-based electrolytes consistently exhibit higher Li+/Na+ and FSI− mobility than their PDADMA counterparts.

Fig. 2. (a and b) Mean square displacement of alkali metal ions and FSI− anions for (a) PDADMA and PDADMP systems with LiFSI or NaFSI at a polyIL-to-salt molar ratio of 1 : 2, and for (b) PDADMA and PDADMP systems with LiFSI at polyIL-to-salt molar ratios of 1 : 1 and 1 : 2. (c) RDFs between the polycationic charge centre (N for ammonium and P for phosphonium) and Na+, and snapshots to illustrate 200 ps trajectories of a selected inner-layer and outer-layer Na+. (d) Fraction of inner-shell and outer-shell Na+ in PDADMA and PDADMP 1 : 2 systems. (e) Average residence time of outer-shell Na+ in PDADMA and PDADMP 1 : 2 systems.

Fig. 2

Open in a new tab

For the lithium systems, where experimental measurements are available for comparison, Li+ and FSI− in PDADMP electrolytes display self-diffusion coefficients nearly one order of magnitude higher than those in the corresponding PDADMA systems (Table S3). Consistently, the experimentally measured conductivity is 17 times higher.8 However, for the sodium systems, the enhancement in ion diffusion is also observed but is less pronounced, with PDADMP exhibiting an approximately two-fold increase relative to PDADMA.

The effect of salt concentration on ion diffusion was further examined for the lithium-containing systems at two polymer-to-salt ratios of 1 : 1 and 1 : 2, which can be directly compared with available experimental data. The change in ion dynamics observed in the MD simulations is consistent with experimental trends. Fig. 2b shows a decrease in the MSD of both Li and FSI when increasing the salt content from 1 : 1 to 1 : 2, which is consistent with the experimentally measured decrease in ionic conductivity at 1 : 2 relative to 1 : 1.8 In contrast, for the PDADMA FSI system, an increase in MSD was observed from 1 : 1 to 1 : 2, supported by NMR-measured diffusivity results.1 This consistency in the lithium salt systems provides good evidence of the model's ability to capture the key structure–property relationships of these materials, providing a basis for extending the model to explore transport behaviour in the sodium systems and to generate experimentally testable predictions.

It is noteworthy that the highest ionic conductivity for PDADMA and PDADMP electrolytes is achieved at a polyIL : LiFSI molar ratio of 1 : 1.5. Notably, the 1 : 1.5 PDADMP electrolyte exhibits a superior conductivity of 1.5 × 10−3 S cm−1at 80 °C and retains a high ambient-temperature conductivity of 1.5 × 10−4 S cm−1 at 30 °C,8 significantly enhancing its practical applicability for all-solid-state lithium batteries. In contrast, the 1 : 2 composition shows the formation of a new crystalline phase, likely driven by excess salt, which decreases conductivity in both systems, but to different degrees. The anion coordination environment was analysed in Table S4 to quantify the excess-salt environment. Three types of FSI− coordination environments were considered: FSI–polycation represents a salt-uncoordinated polyIL environment; FSI–Li represents an excess-salt environment; and Li–FSI–polycation reflects a co-coordinating environment. The excess salt environment remains minor, involving less than 3% of FSI−, in both polyILs. This fraction is slightly higher in PDADMP FSI (2.66%) than in PDADMA FSI (2.26%), which may contribute to a greater decrease in ionic conductivity.

Although the enhancement in ion diffusion in the PDADMP FSI system relative to the PDADMA FSI analogue is less pronounced for sodium systems than for lithium systems, the overall high Na+ diffusivity observed in our MD simulations nonetheless suggests that phosphonium-based polyILs remain strong candidates for practical applications. This is particularly relevant given that conductivities on the order of 1.0 × 10−3 S cm−1 have been already demonstrated for the 1 : 2 PDADMA FSI/NaFSI system with PVDF as mechanical reinforcement at 80 °C.2 Next, further analysis was conducted focused on sodium-containing systems to provide more insight into Na+ mobility with two polycation chemistries.

The polyIL/salt system differs from conventional polymer electrolytes, where alkali metal cation typically coordinates directly with polar groups on the polymer backbone. In the present system, the backbone interacts with FSI− anions, whereas metal cations are co-coordinated by FSI− anions. Therefore, the spatial distribution of metal cations relative to the polycation framework is interesting for understanding ion transport in the polymer matrix. The RDFs were calculated between the cationic centres of PDADMA (through N) or PDADMP (through P) and Na+, showing two peaks within 1 nm. These peaks are assigned to inner-layer and outer-layer Na+, respectively, relative to cationic charge centre (Fig. 2c).

The population of Na+ in two environments was first quantified, with 92% and 89% of Na+ residing in the inner layer for PDADMA and PDADMP, respectively (Fig. 2d), indicating only a modest difference between the two systems. The dynamics of Na+ in the two layers were then examined by calculating short-time MSD over 200 ps. As shown in Fig. S5, outer layer Na+ shows overall higher MSD values than inner layer Na+, as illustrated by the Na+ trajectory snapshots shown in Fig. 2c. This observation is also supported by our previous analysis, which showed that Na+ located farther away from the cationic charge centre exhibits higher diffusion.2

Finally, since Na+ dynamically migrates between the inner- and outer-layer environments, the longer the Na+ remains in the outer layer, the greater its contribution to overall Na+ diffusion. The average time in outer layer (Fig. 2e) differs markedly between the two polyILs, increasing from 74 ps in PDADMA to 105 ps in PDADMP. This indicates that Na+ tends to remain longer in the weakly-bound outer layer environment in PDADMP before returning to the less mobile inner layers, therefore contributing to the enhanced Na+ diffusion.

While the above RDF and inner-/outer-layer analyses provide local insight into the Na+ coordination environment and layer-dependent mobility, the overall higher diffusivity in the PDADMP system is consistent with its weaker collective ionic interaction network, which may facilitate local coordination rearrangements and ion motion. This is supported by energy analysis in Fig. 3. At the DFT level, we analysed the energy required to dissociate either a polycation or a NaFSI ion pair from a PolyIL-NaFSI cluster. Both energies are lower in the PDADMP system, as indicated by Fig. 3a. The average interionic interaction energies were also calculated from MD simulations between polycations and anions or between metal ions and anions in Fig. 3(b and c). For both types of interactions, the total energies (including van der Waals and coulombic contributions) are less negative in PDADMP than in PDADMA. The softer ionic network is also supported by the volume expansion analysis. The volume expansion at a given temperature above Tg (i.e., ΔV = V(T) − V(Tg)) was evaluated at Tg-scaled temperatures, derived from the V–T profiles (Fig. S4b). The PDADMP system shows a higher volume expansion rate than PDADMA, which may also contribute to greater free-volume generation and facilitate ion movement. Collectively, the PDADMP system exhibits weaker average collective ionic interactions together with lower Tg, greater temperature-dependent volume expansion, altered Na+ spatial distributions, and faster ion dynamics. The weaker ionic interactions may reduce the energetic cost of local coordination and network rearrangements and thereby facilitate ion motion.

Fig. 3. (a) DFT calculated energy for dissociating either a polycation or a NaFSI ion pair from a polyIL-NaFSI complex. (b and c) Average interionic interaction energies, including van der Waals and coulombic contributions and their sum, calculated from bulk-phase MD simulations for (b) polycation–anion interactions and (c) metal cation–anion interactions in PDADMA and PDADMP systems with 1 : 2 LiFSI and NaFSI salt.

Fig. 3

Open in a new tab

Comparison of the interfacial ion composition for PDADMA and PDADMP polyIL electrolytes and Na+ transport across interface

In this final section, we present a computational investigation of how the two polyIL electrolyte families respond to increasing negative polarisation of a metal electrode surface. These conditions are relevant to the metal deposition (or charging) environment in a metal battery system, providing initial molecular-level insight into the organisation and interactions of the electrolytes at negatively charged electrode interfaces. This information is critical as it is closely related to the initial formation of solid electrolyte interphase (SEI). The relevant modelling studies for polymer electrolytes are comparatively recent and limited to PEO systems,16 yet it remains largely unexplored for polyIL-based electrolytes.

An interfacial model was employed in which the electrolytes were confined between two parallel gold surfaces with applied surface charge densities of 0, ±6 and ±12 µC cm−2. Gold was selected as a chemically stable and well-defined model metal surface, allowing the effect of negative electrode polarisation on electrolyte organisation and interfacial interactions to be examined without the additional complexity of surface reconstruction or spontaneous chemical reactions associated with more reactive anode. Simulation details are provided in the supplementary information, and the investigation is based on the sodium systems.

Fig. 4(a and b) present the average numbers of Na+, FSI− and polycation units adsorbed on the neutral or negatively charged surface in each polyIL system, while representative snapshots of the adsorbed electrolyte layer at different surface charge densities are given in Fig. 4(c), (d) and S6 (the second replica simulation). Surprisingly, although the concentration of Na+ is twice that of the polycation units and the polycations are incorporated into macromolecule chains, the sodium salt species are not dominantly adsorbed on the metal surface. In contrast, polycationic units are considerably more abundant than Na+ at all surface polarisation conditions in the PDADMA system, albeit to varying extents. FSI− is the most abundant species at the uncharged surface. Its population decreases as the surface charge density becomes more negative, from 0 to −6 and −12 µC cm−2, but remains higher than that of Na+ at all charge densities. Na+ population remains low, decreasing slightly from 12 at 0 µC cm−2 to 10 at −6 µC cm−2, before increasing to 20 at the highest −12 µC cm−2.

Fig. 4. Average number of electrolyte species adsorbed on the anode for (a) PDADMA FSI/NaFSI and (b) PDADMP FSI/NaFSI at different surface charge densities of 0, −6, and −12 µC cm−2. Snapshots of the adsorbed electrolyte species at different charge densities are presented for (c) PDADMA and (d) PDADMP systems. Color codes: FSI− (red sticks), Na+ (green balls), polymer (aqua), N in ammonium (blue), and P in phosphonium (yellow).

Fig. 4

Open in a new tab

The PDADMP system also shows a similar pattern at surface charge densities of 0 and −6 µC cm−2. However, distinctive behaviour is observed at −12 µC cm−2, where the population of Na+ and FSI− significantly increases, exceeding that of phosphonium polycation units. These results suggest that the two polyILs exhibit closely comparable interfacial structural responses to moderate negative polarisation but diverge under stronger negative surface polarisation.

Notably, the significantly higher interfacial populations of both Na+ and FSI− in PDADMP at −12 µC cm−2 could suggest an interfacial chemistry with a greater amount of salt aggregates at higher current densities in experiment. This may favour the formation of salt-derived inorganic SEI species that influence electrochemical performance, as widely reported in recent research.17 This will be explored further in future experimental work. In addition, although most FSI− at the metal surface are coordinated with Na+, some are coordinated only with the polycations at 0 and −6 µC cm−2 in both systems. This coordination environment is not observed at −12 µC cm−2, suggesting that strong negative polarisation reorganises the interfacial coordination structure of FSI− and may consequently affect its reduction behaviour.

To qualitatively assess whether orientation-dependent Au–polycation interactions could contribute to the different interfacial populations, we calculated the adsorption energies of representative polycation units on the gold surface. Adsorption energies were calculated using density functional theory for 40 configurations of both PDADMA and PDADMP cations on the gold surface (Fig. S7). For most comparable configurations, the two cations show similar adsorption energies. Fig. 5 displays the eight configurations for which the adsorption energies between the PDADMA and PDADMP differ by at least 0.5 eV. In five out of eight configurations, PDADMA cations exhibit stronger adsorption energy than the PDADMP cations, and three of these configurations have the ring orientated parallel to the surface. In contrast, in the remaining three configurations (34, 38, and 39), for which PDADMA shows weaker adsorption, the ring is orientated vertically to the surface. From the MD snapshots, the parallel ring orientation appears to be dominant, for which PDADMA exhibits stronger adsorption than PDADMP. The isolated-unit DFT calculations therefore suggest that orientation-dependent Au–polycation interactions could contribute to the observed differences in interfacial populations. However, these calculations do not reproduce the electrostatic and screening environment of the charged condensed-phase interface and therefore cannot establish the origin of the interfacial partitioning observed in the MD simulations. The interfacial composition is also governed by collective effects that are not fully captured by isolated-unit adsorption calculations, including ion coordination, polymer connectivity and chain flexibility, surface coverage, configurational entropy and surface polarization. The slower ionic dynamics of the PDADMA system may further limit Na+ redistribution towards the surface over the simulation timescale, although thermodynamic differences in interfacial partitioning may also contribute.

Fig. 5. (a) Adsorption energies between PDADMA/PDADMP cations and gold, calculated via density functional theory. Top and lateral views of optimized conformations are presented for (b) PDADMA and (c) PDADMP.

Fig. 5

Open in a new tab

The interfacial structures were further studied for both polyILs through number density profiles along the z direction for all electrolyte species from two replicate interface simulations (Fig. S8 and S9). The PDADMA and PDADMP systems show broadly similar patterns. Within 2 nm of the surface, Na+ and FSI− present four distinct density peaks, whereas polycations show only one pronounced peak close to the surface. These results suggest that the polycations are concentrated in the innermost interfacial region, while the salt species form a multilayered structure extending several nanometres from the surface.

We also compared the Na+–FSI− (via N) coordination structure through RDFs in the first three layers and in the bulk at different negative polarisations (Fig. S10 and S11). Three overlapping subpeaks between 0.2 and 0.6 nm are assigned to Na–NFSI, Na–OFSI bi-dentate, Na–OFSI mono-dentate coordination geometries, respectively. Their relative intensities vary among the three interfacial layers with the RDF of the third layer closely resembling that of the bulk phase. Those of the first and second layers show slight differences, indicating that the electrode mainly perturbs Na+–FSI− coordination within the two innermost layers. The Na+–FSI− coordination number remains above 5 in the bulk region and in the second and third interfacial layers, but decreases to around 4.3 in the first layer adjacent to the uncharged surface and to around 3 near the negatively charged surface (Fig. S12). This indicates that the electrode surface progressively reduces the Na+ coordination shell, with a stronger effect under negative polarisation. Analysis of the second-layer composition further shows that Na+ and FSI− are the predominant species here, with only a small number of polycation units present (Fig. S13).

Conclusions

In this work, we apply computational methodology to demonstrate faster alkali–metal ion diffusion in phosphonium-based cationic poly(ionic liquids) compared with the ammonium analogue, highlighting this new polyIL as a promising solid polymer electrolyte for both lithium and sodium batteries. We identified several molecular features associated with the enhanced ion mobility, including the overall weaker dissociation energies related to the interconnected polyIL-salt network, a larger outer-shell Na+ population (i.e. Na+ located further from the cationic charge centre) and longer times that Na+ stays in the mobile outer-shell in the phosphonium system, and a higher volumetric expansion rate in the phosphonium system relative to the ammonium system, consistent with weaker internal interactions.

Together, these features are consistent with and may contribute to the higher Na+ mobility and conductivity observed in the phosphonium-based polyIL. In addition, we investigated the interfacial chemistry of the sodium salt-containing electrolyte near uncharged and negatively charged electrode surfaces and observed a distinct interfacial composition at a high negative surface polarisation of −12 µC cm−2. The phosphonium system exhibits a higher interfacial salt population than the ammonium analogue. Although the Na+ concentration is twice that of the polycation repeat units, a substantial polycation population remains adsorbed at the electrode surface, while the second interfacial layer is predominantly salt-rich. Along the surface-normal direction, one pronounced polycation density peak is observed near the electrode, whereas three to four distinct layers are identified for Na+ and FSI−. These insights advance our understanding of polyIL-based electrolytes and interfaces, motivate further experimental investigations and offer new avenues for the design of polyIL-based polymer electrolytes for solid-state batteries.

Author contributions

F. C. supervised the project, contributed to funding acquisition, methodology development, result analysis, and manuscript drafting. K. Cai and J. S. Puelles conducted simulations, contributed to methodology development, result analysis and manuscript drafting. D. Mecerreyes, D. Brandell, G. Lingua and A. H. Berzosa participated in data discussion and manuscript revision. M. Forsyth co-supervised the project and participated in data discussion and manuscript revision.

Conflicts of interest

The authors declare no competing financial interest.

Supplementary Material

SC-OLF-D6SC06154K-s001
SC-OLF-D6SC06154K-s002

Acknowledgments

All authors acknowledge the financial support from the Australian Research Council through discovery project DP240101661. All computational work was supported by computational resources provided by NCI Australia, an NCRIS-enabled capability supported by the Australian Government.

Data availability

Force field files and polymer coordinate files are provided as separate supplementary information (SI). Supplementary information: details of modelling methods, extra DFT results, diffusion and rotation analysis, Tg simulation results, extra snapshots of interfacial adsorbed layers, number density profiles, and RDF analysis. Tables of Tg fitting parameters for V–T profiles, Tg data, diffusivities, and coordination environment analysis. See DOI: https://doi.org/10.1039/d6sc06154k.

References

  1. Wang X. Chen F. Girard G. M. Zhu H. MacFarlane D. R. Mecerreyes D. Armand M. Howlett P. C. Forsyth M. Poly (ionic liquid) s-in-salt electrolytes with co-coordination-assisted lithium-ion transport for safe batteries. Joule. 2019;3(11):2687–2702. [Google Scholar]
  2. Chen F. Wang X. Armand M. Forsyth M. Cationic polymer-in-salt electrolytes for fast metal ion conduction and solid-state battery applications. Nat. Mater. 2022;21(10):1175–1182. doi: 10.1038/s41563-022-01319-w. [DOI] [PubMed] [Google Scholar]
  3. Lingua G. Chen F. Kondou S. Wang X. Mecerreyes D. Forsyth M. Poly(IL)s-in-Salt Electrolytes as a Versatile Platform for Next Generation Metal Batteries. Acc. Mater. Res. 2025;6(11):1368–1378. doi: 10.1021/accountsmr.5c00215. [DOI] [Google Scholar]
  4. Kondou S. Abdullah M. Popov I. Martins M. L. O'Dell L. A. Ueda H. Makhlooghiazad F. Nakanishi A. Sudoh T. Ueno K. et al., Poly(Ionic Liquid) Electrolytes at an Extreme Salt Concentration for Solid-State Batteries. J. Am. Chem. Soc. 2024;146(48):33169–33178. doi: 10.1021/jacs.4c12616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chen F. Miguel Guerrero Mejía L. Kondou S. Cai K. Forsyth M. Microstructure and Ion Transport in Mixed-Anion Poly(ionic liquid) Electrolytes. ACS Appl. Polym. Mater. 2025;7(22):15420–15428. doi: 10.1021/acsapm.5c03005. [DOI] [Google Scholar]
  6. Cai K. Puelles J. S. Kondou S. Ravindranath S. B. Forsyth M. Chen F. Effect of Side-Chain Modifications on Salt-Concentrated Poly(Ionic Liquid) Electrolytes. Batter. Supercaps. 2025;8(12):e202500484. doi: 10.1002/batt.202500484. [DOI] [Google Scholar]
  7. Lingua G. Gallastegui A. Pairetti Y. Herranz Berzosa A. Cai K. Chen F. Mecerreyes D. Forsyth M. Protic Poly(ionic liquid)s with High Ionic Conductivity and Li+ Transport Number as Moderate Temperature Solid Electrolytes. Chem. Mater. 2025;37(2):697–708. doi: 10.1021/acs.chemmater.4c02765. [DOI] [Google Scholar]
  8. Herranz Berzosa A. Cai K. Lingua G. Mantione D. Santos l. Santino F. Forsyth M. Chen F. Mecerreyes D. Phosphonium Poly(Ionic Liquid) Electrolytes for Fast Lithium-Ion Conduction. J. Am. Chem. Soc. 2026;148(18):19074–19086. doi: 10.1021/jacs.6c02428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. (a) Makhlooghiazad F. Huang L. Byrne N. Howlett P. C. Forsyth M. Ionic Liquids Electrolytes for High Performance Sodium Batteries—Chemistry, Composition, and Interfaces. Adv. Energy Mater. 2026:e06236. doi: 10.1002/aenm.202506236. [DOI] [Google Scholar]; (b) Liu X. Xu C. Adenusi H. Wu Y. Passerini S. Development of PFAS-Free Locally Concentrated Ionic Liquid Electrolytes for High-Energy Lithium and Aluminum Metal Batteries. Acc. Chem. Res. 2025;58(3):354–365. doi: 10.1021/acs.accounts.4c00653. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Huang L. Maleki M. Salado M. Porro J. M. Petrenko V. Le Brun A. Fraysse K. S. Li H. Zhang X. Atkin R. et al., Interphase design from ionic liquid cation mixtures and multi-mode surface analysis for safe and stable Na metal batteries. EES Batteries. 2025;1(4):853–866. doi: 10.1039/d5eb00052a. [DOI] [Google Scholar]
  10. (a) Tsunashima K. Sugiya M. Physical and electrochemical properties of low-viscosity phosphonium ionic liquids as potential electrolytes. Electrochem. Commun. 2007;9(9):2353–2358. doi: 10.1016/j.elecom.2007.07.003.Scopus. [DOI] [Google Scholar]; (b) Ghatee M. H. Bahrami M. Emergence of innovative properties by replacement of nitrogen atom with phosphorus atom in quaternary ammonium ionic liquids: Insights from ab initio calculations and MD simulations. Chem. Phys. 2017;490:92–105. doi: 10.1016/j.chemphys.2017.04.005.Scopus. [DOI] [Google Scholar]
  11. Shimada J. Tsunashima K. Ue M. Iwasaki K. Tsuda T. Kuwabata S. Kanematsu H. Hirai N. Kogo T. Ogawa A. Physical and Electrochemical Properties of Ionic Liquids Based on Quaternary Phosphonium Cations and Carboxylate Anions as Electrolytes. ECS Trans. 2017;75(52):105. doi: 10.1149/07552.0105ecst. [DOI] [Google Scholar]
  12. Hemp S. T. Zhang M. Allen Jr M. H. Cheng S. Moore R. B. Long T. E. Comparing Ammonium and Phosphonium Polymerized Ionic Liquids: Thermal Analysis, Conductivity, and Morphology. Macromol. Chem. Phys. 2013;214(18):2099–2107. doi: 10.1002/macp.201300322. [DOI] [Google Scholar]
  13. Chen M. White B. T. Kasprzak C. R. Long T. E. Advances in phosphonium-based ionic liquids and poly(ionic liquid)s as conductive materials. Eur. Polym. J. 2018;108:28–37. doi: 10.1016/j.eurpolymj.2018.08.015. [DOI] [Google Scholar]
  14. (a) Doherty B. Zhong X. Gathiaka S. Li B. Acevedo O. Revisiting OPLS Force Field Parameters for Ionic Liquid Simulations. J. Chem. Theory Comput. 2017;13(12):6131–6145. doi: 10.1021/acs.jctc.7b00520. [DOI] [PubMed] [Google Scholar]; (b) Jorgensen W. L. Maxwell D. S. Tirado-Rives J. Development and Testing of the OPLS All-Atom Force Field on Conformational Energetics and Properties of Organic Liquids. J. Am. Chem. Soc. 1996;118(45):11225–11236. doi: 10.1021/ja9621760. [DOI] [Google Scholar]
  15. Forsyth M. Porcarelli L. Wang X. Goujon N. Mecerreyes D. Innovative Electrolytes Based on Ionic Liquids and Polymers for Next-Generation Solid-State Batteries. Acc. Chem. Res. 2019;52(3):686–694. doi: 10.1021/acs.accounts.8b00566. [DOI] [PubMed] [Google Scholar]
  16. (a) Ebadi M. Costa L. T. Araujo C. M. Brandell D. Modelling the Polymer Electrolyte/Li–Metal Interface by Molecular Dynamics simulations. Electrochim. Acta. 2017;234:43–51. doi: 10.1016/j.electacta.2017.03.030. [DOI] [Google Scholar]; (b) Asha A. S. Iroegbu J. N. Visayas B. R. B. Mayes M. Shen C. Molecular insights into the electric double-layer structure at a polymer electrolyte-electrode interface. Electrochim. Acta. 2023;446:142131. doi: 10.1016/j.electacta.2023.142131. [DOI] [Google Scholar]; (c) Rivera-Pousa A. Otero-Mato J. M. Montes-Campos H. Méndez-Morales T. Diddens D. Heuer A. Varela L. M. Ternary Solid Polymer Electrolytes at the Electrochemical Interface: A Computational Study. Macromolecules. 2024;57(9):3921–3936. doi: 10.1021/acs.macromol.3c02669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. (a) Karbak M. Yim K. Meng Y. S. Shao Y. Xu W. Weakly solvating electrolytes: a solvation-centric paradigm for rechargeable metal batteries. Chem. Sci. 2025;16(44):20694–20717. doi: 10.1039/d5sc06221g. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Cui X. Ding S. Wang Y. Teng H. Feng Y. Han X. Rong X. Xi K. Zheng Q. Lu Y. et al., Fundamentals, Status, and Prospects of Liquid Organic Electrolytes for High-Energy Sodium-Ion Batteries. Adv. Mater. 2025;38(29):e19965. doi: 10.1002/adma.202519965. [DOI] [PubMed] [Google Scholar]; (c) Wang Z. Guo X. Dong Y. Qu W. Wang K. Dong L. Yang C. Lei Z. Liang J.-Y. Low-Concentration Electrolyte Engineering for Rechargeable Batteries. Adv. Mater. 2025;37(37):2501215. doi: 10.1002/adma.202501215. [DOI] [PubMed] [Google Scholar]; (d) Lee J. Kim J. Jang W. Lee D. G. Kim H. An Y. Son J. Kang M. Lee G. Lee J. et al., A strategic tuning of interfacial Li+ solvation with ultrathin polymer layers for anode-free lithium metal batteries. Joule. 2026;10(1):102226. doi: 10.1016/j.joule.2025.102226. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SC-OLF-D6SC06154K-s001
SC-OLF-D6SC06154K-s001.pdf (3MB, pdf)
SC-OLF-D6SC06154K-s002
SC-OLF-D6SC06154K-s002.zip (21.8KB, zip)

Data Availability Statement

Force field files and polymer coordinate files are provided as separate supplementary information (SI). Supplementary information: details of modelling methods, extra DFT results, diffusion and rotation analysis, Tg simulation results, extra snapshots of interfacial adsorbed layers, number density profiles, and RDF analysis. Tables of Tg fitting parameters for V–T profiles, Tg data, diffusivities, and coordination environment analysis. See DOI: https://doi.org/10.1039/d6sc06154k.

Articles from Chemical Science are provided here courtesy of Royal Society of Chemistry

RESOURCES