Pressure-Induced Solid–Solid Phase Transitions in Barocaloric Organic Ionic Plastic Crystals

Abstract

Barocaloric (BC) materials represent a highly promising family of solid-state refrigerants that offer real potential to replace environmentally damaging gas refrigerants. The rational design of BC materials relies on understanding solid–solid phase transitions (SSTs) under pressure. Organic ionic plastic crystals (OIPCs) emerge as promising BC candidates, but direct characterization of their pressure-dependent phase behavior remains a challenge. Herein, we employ high-pressure broadband dielectric spectroscopy (HP-BDS) as a tool to map the phase diagrams of three OIPCs: 4-ethyl-4-methylmorpholinium bis­(fluorosulfonyl)­imide ([C₂mmor]­[FSI]), tetraethylammonium bis­(fluorosulfonyl)­imide ([N₂₂₂₂]­[FSI]), and tetraethylammonium bis­(trifluoromethanesulfonyl)­imide ([N₂₂₂₂]­[TFSI]). We demonstrate that the real part of the complex dielectric permittivity (ε′) is uniquely sensitive to SSTs in OIPCs, which are driven by changes in dipolar reorientation. By performing both isobaric cooling/heating and isothermal compression/decompression experiments, we directly determine the pressure sensitivity (dT_SS/dP) of the SSTs and observe the transition hysteresis in terms of temperature and pressure, finding both properties to be strongly ion-dependent. Furthermore, hysteretic behavior at high pressure is found to be unique to each OIPC, despite similarities in ion structure. In particular, the pressure hysteresis directly observed for [C₂mmor]­[FSI] is lower than that predicted from quasi-direct (isobaric) measurements, indicating the kinetic delay is smaller for a pressure change than a temperature change, which would further improve the already excellent BC metrics of this OIPC.

Introduction

The barocaloric (BC) effect is the thermal change observed in a material upon the application and removal of hydrostatic pressure. BC materials have garnered substantial research attention in recent years − due to their promise of providing an energy-efficient, low global warming potential and solid-state alternative to the volatile hydrofluorocarbon refrigerants ubiquitous in conventional liquid–vapor refrigeration systems. To develop this application, there is a need for new materials with high-entropy (ΔS_ss) transitions that are highly sensitive to pressure, with transition temperatures (T_ss) in subambient regions targeted for cooling applications. Unfortunately, a vast majority of the literature materials exhibiting large or “colossal” BC effects (ΔS_ss > 100 J kg^–1 K^–1) have transition temperatures above ambient, which are of limited practical use. Among the most promising emerging families of BCs are organic ionic plastic crystals (OIPCs), which often exhibit high ΔS_ss transitions below room temperature.

OIPCs are salts comprised of anions and organic cations that are characterized by their propensity to exist in locally disordered, solid “plastic” phases prior to their melting transitions, above which they form ionic liquids. The plastic phases are separated from lower temperature crystalline phases by solid–solid phase transitions that typically involve an onset of localized dynamic disorder of the ions while their centers of mass occupy fixed lattice sites. − The orientational and/or conformational freedom of ions in the plastic phases of OIPCs leads to a range of interesting thermophysical properties that can be exploited for various energy applications. For example, high defect concentrations can facilitate fast-ion conduction, which is ideal for solid-state electrolyte applications. − Alternatively, significant latent heat transitions can be exploited for thermal energy storage. It has recently been recognized that OIPCs with high entropy change transitions can display a substantial increase in free volume at the disordering transition, which makes their transition temperatures, T_ss, highly sensitive to pressure. This underpins their emerging application as pressure-responsive, solid BC refrigerants.

The phase transition and thermophysical properties of OIPCs are highly dependent on the nature of the cations and anions used, and slight structural modifications to either ionic component can often produce pronounced and as-yet unpredictable effects on the material’s properties. − For example, for two structurally analogous OIPCs based on an N-alkyl-N-ethylpyrrolidinium cation paired with the bis­(fluorosulfonyl)­imide ([FSI]⁻) anion, the ethyl-substituted analogue (i.e., [C₂epyr]­[FSI]) displays an ionic conductivity over an order of magnitude higher than that of the methyl analogue ([C₂mpyr]­[FSI]), attributed to the enhanced rotational freedom of the ethyl substituent in the plastic phase. Furthermore, for the N-methyl-N-isopropyl pyrrolidinium ([C_i3mpyr]⁺) cation, the OIPC formed when paired with the [FSI]⁻ anion displays a ΔS_ss approximately double that of the OIPC formed when paired with the structurally similar bis­(trifluoromethanesulfonyl)­imide ([TFSI]⁻) anion (92 J kg^–1 K^–1 and 190 J kg^–1 K^–1 respectively), and the volume change is ∼30% larger.

A comprehensive understanding of the solid–solid phase behavior of OIPCs under elevated pressure is the next step for advancing their application as BC materials. High-pressure dielectric spectroscopy (HP-DS) is a uniquely powerful technique for probing the dielectric properties of OIPCs under various temperature (T) and pressure (P) conditions. This technique is frequently employed to monitor the dynamics in ionic liquids and molecular glass-formers − across a wide range of time scales, from localized dipolar reorientations to long-range ion transport. Its exceptional sensitivity to changes in molecular mobility makes it an excellent probe for various phase transitions. For instance, HP-DS can track crystallization kinetics in real-time, as the progressive freezing of molecular mobility during nucleation and growth is directly reflected in the evolution of dielectric permittivity and conductivity. , Beyond crystallization, dielectric spectroscopy is also highly effective at detecting liquid–liquid transitions , and isomerization processes, where subtle structural reorganizations manifest as distinct changes in relaxation behavior. Consequently, HP-DS emerges as an indispensable, nondestructive method for mapping the complex high-pressure phase diagrams of OIPCs and quantifying the pressure sensitivity (dT_ss/dP) of their SSTs, a critical BC parameter that defines the pressure requirements of their associated cooling systems when used as refrigerants.

Herein, we have chosen three model OIPCs of different chemical architectures of cation and anion (Figure ) – two with a common [N₂₂₂₂]⁺ cation and two with a common [FSI]⁻ anion – to investigate the influence of ion structure on their dielectric properties and phase behavior at various thermodynamic T-P conditions. The solid–solid phase transitions of the examined OIPCs were induced by isobaric cooling/heating and isothermal compression/decompression, the latter performed in a pressure range of 0.1–700 MPa. To characterize the BC properties of the OIPCs, dielectric spectroscopy was employed, and the permittivity representation was used to monitor the solid–solid phase transitions. The obtained data provides a link between the molecular structure of the ions and the pressure sensitivity of the phase transitions, as well as the degrees of both temperature and pressure hysteresis, advancing the development of a predictive framework for next-generation barocaloric OIPCs.

1.

Chemical structures and abbreviations of the OIPCs studied here. [C₂mmor]: 4-ethyl-4-methylmorpholinium; [N₂₂₂₂]: tetraethylammonium; [FSI]: bis­(fluorosulfonyl)­imide; [TFSI]: bis­(trifluoromethanesulfonyl)­imide.

Materials and Methods

1-Bromoethane (≥99%, Sigma-Aldrich), 4-methylmorpholine (≥99.5%, Sigma-Aldrich), lithium bis­(fluorosulfonyl)­imide (99.9%, Solvionic), potassium bis­(fluorosulfonyl)­imide (99.9%, Suzhou Fluolyte Co), lithium bis­(trifluoromethanesulfonyl)­imide (99.9%, Solvionic), dichloromethane (HPLC grade, ChemSupply or ≥99.8%, Supelco), tetraethylammonium bromide (99%, Sigma-Aldrich), ethanol (99.5%, ChemSupply), ethyl acetate (99.5%, ChemSupply), acetonitrile (≥99.9%, Sigma-Aldrich), and dimethyl sulfoxide-d₆ (99.96%, Sigma-Aldrich) were all used as received.

Synthesis of OIPCs

[C₂mmor]­[FSI], [N₂₂₂₂]­[TFSI] and [N₂₂₂₂]­[FSI] were synthesized by the group at Deakin University following previously published procedures. , A brief summary of the synthesis protocols is provided below.

[C₂mmor]­[FSI]

[C₂mmor]Br was first synthesized by the dropwise addition of 15.54 g (0.154 mol) of 1-bromoethane to a stirring solution of 14.435 g (0.145 mol) 4-methylmorpholine in acetonitrile at 50 °C under an N₂ atmosphere. The reaction mixture was heated at 50 °C overnight (approx 15 h) and then at reflux (75 °C) for 5 h. The reaction mixture was cooled to room temperature and the crystals of [C₂mmor]­Br were collected by vacuum filtration. Ethyl acetate was added to the filtrate and the resulting crystals were collected by vacuum filtration. The solids were combined and dried under vacuum for 1 h before being recrystallized in an ethanol/ethyl acetate solution twice. The recrystallized product was dried under vacuum at 50 °C for ∼2 days (Yield: 61%). 6.668 g (0.0318 mol) of the [C₂mmor]Br product and 6.012 g (0.0321 mol) of LiFSI were then each dissolved in 10 mL of water. The two solutions were combined and stirred at room temperature for ∼3 h. The [C₂mmor]­[FSI] product was extracted with DCM (3 × 20 mL). The DCM solution was washed with water (3 × 5 mL) and concentrated by rotary evaporation before being dried under vacuum at 50 °C for upward of 2 days, yielding [C₂mmor]­[FSI] as a white solid (yield: 60%). ¹H NMR (400 MHz, DMSO-d ₆): δ 3.92–3.89 (m, 4H), 3.49 (q, 2H, J = 7.3 Hz), 3.38 (t, 4H, J = 5.0 Hz), 3.09 (s, 3H), 1.25 (t, 3H, J = 7.2 Hz). ¹³C NMR (100.6 MHz, DMSO-d ₆): δ 60.27, 59.90, 58.93, 45.81, 7.26. MS: ES⁺(m/z): [M]⁺ calcd for [C₂mmor]⁺, 130.12; found, 129.8. ES^–(m/z): [M]⁻ calcd for [FSI]⁻, 179.92; found, 179.9.

[N₂₂₂₂]­[TFSI]

7.147 g (0.034 mol) [N₂₂₂₂]Br and 9.568 g (0.035 mol) LiTFSI were each dissolved in ∼15 mL water. The two solutions were combined and stirred at room temperature for ∼3 h. The resulting white crystals were collected by vacuum filtration before being dissolved in DCM (20 mL) and washed with water (3 × 5 mL). The DCM solution was concentrated by rotary evaporation and dried under high vacuum for ∼2 days at 50 °C to give [N₂₂₂₂]­[TFSI] as a white crystalline solid (yield: 76%). ¹H NMR (400 MHz, DMSO-d ₆): δ 3.21 (q, 8H, J = 7.3 Hz), 1.16 (tt, 12H, J_HH = 7.3 Hz, J_HN = 1.9 Hz). ¹³C NMR (100.6 MHz, DMSO-d ₆): δ 51.84, 7.54. MS: ES⁺(m/z): [M]⁺ calcd for [N₂₂₂₂]⁺, 130.16; found, 129.1. ES^–(m/z): [M]⁻ calcd for [TFSI]⁻, 279.92; found, 280.0.

[N₂₂₂₂]­[FSI]

6.79 g (32.3 mmol) [N₂₂₂₂]Br and 7.15 g (32.6 mmol) KFSI were each dissolved in 20 mL water. The solutions were combined and left to stir overnight at RT, during which time a white precipitate formed. DCM (50 mL) was added to the solution which was again left to stir overnight at RT. The DCM layer was then isolated and washed with water (5×). The solution was then concentrated by rotary evaporation and dried under high vacuum for ∼2 days to give [N₂₂₂₂]­[FSI] as a white solid (yield: 80%). ¹H NMR (400 MHz, DMSO-d ₆): δ 3.20 (q, 8H, J = 7.2 Hz), 1.15 (tt, 12H, J_HH = 7.2 Hz). MS: ES⁺(m/z): [M]⁺ calcd for [N₂₂₂₂]⁺, 130.16; found, 130.1. ES^–(m/z): [M]⁻ calcd for [FSI]⁻, 179.92; found, 179.9.

Methods

Dielectric Spectroscopy (DS)

For dielectric measurements, the examined samples were placed between two stainless steel plates of a capacitor (10 mm in diameter), electrically separated by a Teflon spacer (0.1 mm thickness). Ambient-pressure dielectric measurements of the OIPCs were performed using a Novo-Control GmbH Alpha dielectric spectrometer. The dielectric spectra ε′(f) were initially measured in the frequency range 10⁶–10^–1 Hz on cooling and subsequent heating, with a temperature step of 5 °C for all examined samples. The obtained results are presented in Figure S1 in the Supporting Information. In this graph, the liquid–solid and solid–solid transitions are easily detectable at every single frequency. However, to determine the transition temperature precisely, a single-frequency measurement at 300 kHz was performed on cooling and heating at a rate of 5 °C/min. This frequency was selected because the measurements require less than a second per data point (compared to the full-frequency spectra (10⁶ Hz to 10^–1 Hz) that require ∼15 min to acquire), enabling continuous monitoring of the dielectric response during dynamic temperature scan and thus precise determination of the onset temperature of the solid–solid transition.

Following the initial characterization at atmospheric pressure, dielectric measurements under high-pressure conditions were performed. The schematic visualization of the high-pressure experimental setup is presented in Figure b. The capacitor with OIPC, previously used for ambient-pressure dielectric measurements, was equipped with wires for electrical contact to the impedance analyzer and was tightly wrapped with Teflon tape.

2.

(a) Schematic representation of the experimental procedure applied to the OIPCs at ambient and elevated pressure. (b) The high-pressure setup is connected to an appropriate impedance analyzer for monitoring the dielectric response of the OIPCs at various T-P thermodynamic conditions. Note that the capacitor with the examined OIPC is protected from the surrounding pressure-transmitting medium by Teflon tape.

To exert pressure, a capacitor with OIPC was placed in the chamber and compressed using a nonpolar, noncorrosive pressure-transmitting liquid. We used Plexol 201 as the pressure-transmitting medium (supplied from Sigma-Aldrich, (bis­(2-ethylhexyl) sebacate)). Here, we note that the pressure-transmitting medium does not come into contact with the examined sample. The Teflon wrapping serves as a physical barrier, isolating the sample from the surrounding fluid and preventing contamination. Consequently, the applied pressure-transmitting medium does not alter the SSTs under the experimental conditions. Furthermore, high-pressure measurements were performed on OIPCs in the solid state (below their melting points). Therefore, the sample did not flow during the experiments, and the electrode–sample contact remained stable throughout isothermal compression–decompression cycles. After the high-pressure measurements, the Teflon tape was removed and the capacitor with the sample was measured again using an ambient-pressure setup. The consistency of the dielectric characteristics of the OIPC observed before and after the high-pressure experiments confirms that the pressurized fluid does not penetrate the Teflon wrapping and that the measured dielectric response originates exclusively from the OIPC sample.

To control the temperature during the measurements, the pressure chamber was placed inside the environmental chamber (Weiss Technik, model WT 110/70), which operated over the temperature range from −70 °C to +80 °C. The chamber was equipped with a programmable temperature controller, enabling isothermal holds with temperature stability of ±1.3 °C. The temperature was continuously monitored using a calibrated Pt100 resistance temperature sensor positioned inside the pressure chamber, ensuring accurate measurement of the sample temperature under all experimental conditions. The compression and decompression experiments of the OIPCs were performed at a constant temperature with a constant pressure rate. Since compression inherently increases the sample temperature, we selected 14 MPa/min as the standard rate to minimize this perturbation and maintain isothermal conditions for all measurements reported.

As presented in Figure a, the highest-temperature solid phase (immediately below the melting point) is designated as phase I. Consequently, upon cooling/compression, phase I transforms to lower-temperature phases, which are sequentially labeled phase II, phase III, etc. Thus, phase I represents the most disordered plastic crystalline state, while phases II and III represent progressively more ordered crystalline phases. This nomenclature is applied consistently to all three OIPCs studied herein.

Nuclear Magnetic Resonance Spectroscopy (NMR)

NMR spectra were recorded at 298 K on a Bruker Avance III NMR spectrometer equipped with a 9.4 T magnet and 5 mm TBO probe, operating at 400.13 MHz (¹H). Chemical shifts (δ) are reported in parts per million (ppm) and were referenced to the residual solvent signals (¹H, ¹³C) or from the solvent block (²H) signal according to IUPAC recommended secondary referencing method and the manufacturer’s protocols (¹⁹F).

Mass Spectrometry (MS)

Analysis was carried out at the School of Chemistry, Monash University. A 5 μL injection volume of test compound was used in conjunction with a 3 min isocratic gradient of acetonitrile with 0.1% formic acid over a 3 min run time an Agilent 1290 Infinity II (Santa Clara, CA, USA). Analysis was conducted on an Agilent 6546 QTOF MS system (Santa Clara, CA, USA) with a dual ESI source. The MS was operated in positive or negative mode using the following conditions: nebulizer pressure 45 psi, drying gas flow-rate 10 L/min, gas temperature 300 °C, capillary voltage 4000/-4000 V, fragmentor 170 V and skimmer 65 V. The instrument was operated in the extended dynamic range mode with data collected in m/z range 100–1500.

Results and Discussion

According to common practice, the electric conductivity σ*­(f) formalism is the most widely used approach for presenting dielectric data from ion-containing systems. , However, for monitoring the solid–solid phase transition in the examined organic ionic plastic crystals (OIPCs), the real part of the complex dielectric permittivity (ε′) was deemed the optimal observable. The rationale for this choice lies in the fundamental nature of the solid–solid phase transition, which predominantly involves a change in dipolar orientation and molecular symmetry, rather than long-range ion hopping/diffusion. Further, short-range dynamic disorder is expected to have the greatest influence on barocaloric properties through its effect on free volume, which also underpins ion diffusion in OIPCs. , Since ε′ directly probes a material’s polarizability and alignment of permanent dipoles under an electric field, it is uniquely suited to capture these changes. Conductivity, on the other hand, is only sensitive to changes in translational ion transport. Hence, while σ*­(f) is an effective probe for monitoring the crystallization of isotropic liquids, it can be insensitive to subtle molecular changes during the solid–solid phase transitions (SSTs) of interest here.

Variable Temperature Measurements

Figure a illustrates the temperature evolution of the static dielectric constant ε′ of the OIPCs recorded at a frequency of 300 kHz at ambient pressure conditions. The lowest temperature phases (i.e., phase II of [C₂mmor]­[FSI] and phase III of the ammonium-based materials) all display low values of dielectric permittivity (ε′ ≈ 2), indicating highly ordered crystalline phases with no residual rotational disorder. In other words, after crystallizing into these phases, the cations and anions are relatively static and can no longer reorient in response to the alternating electric field used in the dielectric measurement. For all OIPCs, heating from the low-temperature regime results in a monotonic increase of the dielectric constant, which is interrupted by step increases in ε′, consistent with the occurrence of first-order, disordering phase transitions. Note that the transition temperatures identified as the onset of ε′(T) increase are consistent with those determined by differential scanning calorimetry (DSC) in refs and (see Table for comparison).

3.

(a) Temperature evolution of the dielectric constant for the examined OIPCs determined at ambient pressure (300 kHz) on cooling (blue symbols) and subsequent heating (green symbols) at 5 K/min. Blue and green arrows indicate solid–solid phase transition temperatures (T_ss) determined by the onset method. Phase regions are labeled I, II, and III according to the convention defined in the text (phase I = highest-temperature plastic phase). The thermal hysteresis ΔT_hys (vertical arrows) represents the temperature difference between the transitions on heating and cooling. (b) Pressure evolution of ε′ at selected isothermal conditions measured on compression (blue symbols) and decompression (green symbols). Arrows indicate solid–solid phase transition pressures (P_ss) determined by the onset method. Phase regions are labeled as in panel (a). The pressure hysteresis ΔP_hys (horizontal arrows) represents the pressure difference between the transitions on compression and decompression. (c) Phase diagrams determined for [C₂mmor]­[FSI], [N₂₂₂₂]­[FSI] and [N₂₂₂₂]­[TFSI] (from left to right). Solid lines are linear fits to the experimental data points, with slopes (dT_ss/dP) reported in Table . Red stars on the T_SS-P lines, denoting the phase II-phase I transition of [C₂mmor]­[FSI] and phase II-phase III of [N₂₂₂₂]­[TFSI], are taken from high-pressure differential thermal experiments presented in ref . Gray regions indicate the hysteresis ranges, where two solid phases can exist depending on how the high-pressure experiment was performed (compression/decompression).

1. Characteristics of the Solid–Solid Phase Transitions in the Examined OIPCs .

[C2mmor][FSI]
	T [°C]	ΔS [J/molK]	dTSS/dP [°C/100 MPa]	ΔV [cm3/g]
I → II	–25b	N/A	13.6 ± 0.6	N/A
II → I*	11*, 11*	84*	10.8 ± 0.5*	0.0286 ± 0.0019*
Melt → I	130	N/A	N/A
I → Melt*	129*	8*	N/M
[N2222][FSI]
Melt → I	70	N/A	N/A	N/A
I → Melt*	68*, 70*	5*	N/M
I → II	35	N/A	17.9 ± 0.7
II → I*	42*, 39*	29*	17.9 ± 0.9*	0.0155 ± 0.0022*
II → III	–15	N/A	19.7 ± 0.8
III → II*	–1, −0.7	15*	20 ± 0.8*	0.0090 ± 0.0010*
[N2222][TFSI]
Melt → I	100	N/A	N/A	N/A
I → Melt*	105*, 104*	25*	N/M
I → II	50	N/A	28.2 ± 0.9
II → I*	53*, 52*	3*	28.2 ± 1.1*	0.0021 ± 0.0004*
II → III	–15	N/A	13.4 ± 0.6
III → II	5c*, 4.3*	67*	21 ± 0.9*	0.034 ± 0.004*

^aΔS_SS denotes the entropy changes during the SST determined from calorimetric data presented in refs and . dT_SS/dP coefficients were determined experimentally using dielectric spectroscopy. ΔV denotes the volume change during the SST calculated from the CC equation.

^bDetermined from BDS with the standard uncertainty of ±1.3 °C.

^cDetermined from DSC. Reverse transitions and their corresponding parameters are denoted as stars (*).

^dN/A in ΔS column values - not available in literature; N/M – not measured experimentally.

Closer inspection of the dielectric data recorded within the plastic crystalline phase I reveals relatively high values of ε′ for [C₂mmor]­[FSI] and [N₂₂₂₂]­[FSI] (ε′ = 8–16) and a lower dielectric permittivity for [N₂₂₂₂]­[TFSI] (ε′ = 3.5–5). This indicates that in the plastic phase I of the [FSI]-based compounds, there is still large-scale collective alignment of ions. This can be rationalized by differences in charge distribution across the [FSI]⁻ and [TFSI]⁻ anions. [FSI]⁻ has a more localized negative charge on the sulfonyl groups, which could lead to stronger and more directional Coulombic interactions with the [N₂₂₂₂]⁺ cation, including in the disordered phases. In contrast, [TFSI]⁻ has extensive charge delocalization over the −CF₃ groups, which reduces the effective charge density and likely weakens directional interactions in the disordered phase. The lower melting transition entropies (ΔS_m) of the [FSI]-based compounds compared to [N₂₂₂₂]­[TFSI] (Table ) suggest that some collective alignment of the ions persists in the melted phases of the [FSI]-based salts.

A clear correlation is observed when comparing the magnitudes of the changes in ε′ with values of ΔS_ss for the corresponding SSTs. From Figure a, it is evident that there is no substantial difference in dielectric constant ε′ between phases I and II of [N₂₂₂₂]­[TFSI], while the phase II to phase III transition is much more pronounced in ε′(T) behavior. This is consistent with ΔS_ss being large for the III–II transition and markedly lower for the II–I transition (see Table ). A similar relationship between the size of the discontinuous jump in ε′ and ΔS_ss at the SSTs is observed for [N₂₂₂₂]­[FSI], but with the larger permittivity and entropy changes during the II–I transition. This correlation arises because the step-change in ε′ at the transition reflects the gain in dipolar polarizability when ions acquire rotational and conformational freedom. For [N₂₂₂₂]­[TFSI], the negligible Δε′ at the II–I transition indicates that the anion already possesses significant conformational disorder in phase II, so little additional freedom is gained upon entering phase I. Conversely, the large Δε′ at the III–II transition signals the unlocking of this conformational flexibility, consistent with its dominant entropy change. For [N₂₂₂₂]­[FSI], the rigid anion forces the majority of disorder to appear at the II–I transition, where both Δε′ and ΔS_ss are the largest. This indicates that the key mechanisms dictating ΔS_ss in OIPCs are closely correlated with the dielectric permittivity and thus the number of relaxing elements and their dipole moments.

In all cases, cooling of the OIPCs from a high-temperature regime results in a drop in ε′, indicating a crystallization transition from an isotropic liquid to a plastic crystalline phase I. Upon a further decrease in temperature, the dielectric constant exhibits a monotonic reduction interrupted by sharp discontinuities marking the temperatures of ordering (crystallization) transitions, confirming the reversibility of the SSTs. However, a thermal hysteresis (ΔT_hys) is observed, with the SSTs occurring at lower temperatures on cooling than on heating. The most pronounced hysteresis is observed for the phase I–II transition in [C₂mmor]­[FSI] (ΔT_hys = 36 ± 2 °C), which is consistent with this transition having the highest ΔS_ss among those studied.

On the other hand, for the ammonium-based OIPCs, ΔT_hys is practically negligible for the phase I–II transitions and substantially larger for the II–III transitions (see Figure a and Table for exact values). This is an interesting observation, as one might predict that a more disordering (higher ΔS_ss) transition would require greater supercooling to initiate crystallization during cooling. Hysteresis, i.e., supercooling, is not desirable in BC materials because it requires higher pressure to drive the transition reversibly (and therefore a higher pressure during the cooling cycle in application). The minimum pressure requirement of the system (P _rev) is directly related to ΔT_hys through P _rev = ΔT_hys/[dT _ss/dP], hence low ΔT_hys and high dT_ss/dP are targeted in BC OIPCs. It is worth noting that, as well as being related to a material’s intrinsic properties, ΔT_hys is a kinetic phenomenon and is thus also influenced by experimental parameters, such as scan rates and sample volumes. Nevertheless, as materials with low or no hysteresis are targeted for this application, it is highly valuable to understand the relationships between ion structures and their disorder modes and the propensity of the OIPCs to supercool. Considering there is no distinct correlation between the magnitude of ΔS_ss and ΔT_hys in these analogous ammonium salts, it can be derived that the [TFSI]⁻ anion is contributing to the enhanced hysteresis observed for [N₂₂₂₂]­[TFSI]. This might be attributed to the more complex conformational landscape of this anion, as rotation of the CF₃ groups allows for various combinations of cis-trans configurations. Thus, upon cooling, the anion may be trapped in a higher energy conformation that is kinetically stabilized, hindering the onset of the exothermic transition into the lower energy structure and thus increasing the observed hysteresis.

Variable Pressure Measurements

While cooling is the most straightforward method for inducing a phase transition to a higher density (usually more ordered) phase, isothermal compression offers another approach. In contrast to isobaric cooling, which affects ion dynamics through a reduction in temperature and the corresponding increase in density, compression at constant temperature operates solely by reducing free volume and stabilizing the higher-density phase. This distinction makes isothermal compression an indispensable tool for understanding solid–solid transitions in ionic liquids and related materials. These conditions also best reflect those of the BC application. Therefore, in the next step, we performed a series of isothermal compression–decompression scans on the OIPCs and monitored the evolution of the dielectric constant as a function of pressure. Specifically, for [C₂mmor]­[FSI], compression at seven different temperatures was performed (T = −15, −10, 0, 10, 25, 40, and 50 °C), while high-pressure experiments of [N₂₂₂₂]­[FSI] and [N₂₂₂₂]­[TFSI] were performed in 10 °C increments in the range of 10–80 °C. Note that for experiments conducted within the hysteretic temperature range, decompression was followed by heating to room temperature to revert to phase I (see Figure a), as escaping the metastable phase II state would otherwise require the application of negative pressure.

Representative results of ε′(P) scans obtained along the compression and decompression routes at different temperature conditions are presented in Figure b. As expected, isothermal compression has a fundamentally similar effect on the dielectric constant to isobaric cooling, i.e., a clear drop in ε′ is observed at the solid–solid transitions. On the other hand, decompression results in an increase in dielectric permittivity at the SSTs, consistent with the ambient pressure heating scans. Precisely, compression of [C₂mmor]­[FSI] at 50 °C results in a sharp reduction in ε′ – being a signature of phase I-phase II transitionat 556 MPa. On the other hand, an increase in ε′(P) indicative of the phase II–I transition is observed on decompression at a critical value of 364 MPa. Consequently, a pressure hysteresis (ΔP_hys) of 192 ± 17 MPa is observed between the compression and decompression pathways of [C₂mmor]­[FSI] at 50 °C. From a closer inspection of the decompression scan of [C₂mmor]­[FSI] performed at 50 °C, an interesting feature appears: an initial increase in ε′(P) associated with the phase II → phase I transition is followed by an unexpected decrease before the final increase. This feature is reproducible across multiple decompression rates and is absent at lower temperatures (≤30 °C). We attribute this behavior to a pressure-induced conformational trapping of the [C₂mmor]⁺ cation, followed by a kinetically delayed relaxation upon decompression. The morpholinium cation possesses significant conformational flexibility due to ring puckering (chair, boat, twist-boat) and rotation of the ethyl and methyl substituents, creating a rich conformational landscape. At elevated temperatures, the cation possesses a broad distribution of conformers. Under high pressure, the equilibrium shifts toward more compact conformers that pack efficiently in the dense phase II. Upon decompression, the material initially transforms to phase I while retaining the high-pressure-stabilized conformational distribution. As pressure decreases further, conformational relaxation toward the equilibrium low-pressure distribution becomes kinetically accessible. This relaxation involves reorientation of the cations and changes in ring conformation, temporarily disrupting dipolar alignment and causing a transient decrease in ε′. The absence of this feature at lower temperatures is consistent with reduced conformational mobility, preventing relaxation on the experimental time scale. Notably, this behavior is not observed in the symmetric tetraethylammonium-based OIPCs, highlighting the unique role of the asymmetric morpholinium cation in enabling pressure-induced conformational trapping.

In the next step of [C₂mmor]­[FSI] analysis, the temperature of the SST (T_ss) was plotted vs pressure in Figure c (where ΔP_hys is the horizontal displacement between the SST temperatures at each T condition). From this graph, it can be observed that ΔP_hys occurs at every examined T condition, and the magnitude of ΔP_hys increases with decreasing temperature. On the other hand, when constant pressure is considered, ΔT_hys (being the vertical displacement between the SST temperatures at each P condition) decreases with increasing pressure. Consequently, the slope of the T_SS-P line characterizing compression of [C₂mmor]­[FSI] is steeper than that of the decompression pathway, which is quantified by a higher dT_SS/dP coefficient of the phase I–II transition (dT_SS(I–II)/dP = 13.6 °C/100 MPa) when compared to the reverse (II–I) transition (dT_SS(II–I)/dP = 10.8 °C/100 MPa). Note that the obtained T_SS-P line denoting the phase II–I transition of [C₂mmor]­[FSI], which represents the thermodynamic (endothermic) transition, is in good agreement with that derived from high-pressure differential thermal analysis experiments (red stars in Figure c).

As presented in Figure b and c, the ammonium-based OIPCs display more complex pressure behavior than [C₂mmor]­[FSI] due to the multiple solid–solid transitions observed in these materials. Consistent with ambient pressure data, squeezing of [N₂₂₂₂]­[FSI] results in a significant drop of ε′ upon the phase I–II transition, and a much smaller decrease in dielectric permittivity is associated with the subsequent transition to phase III. Conversely, a larger decrease in ε′ is observed for the phase II–III transition in [N₂₂₂₂]­[TFSI] and that of the I–II transition is noticeably smaller. This suggests that a similar degree of dipolar freedom exists in phase I and II of [N₂₂₂₂]­[TFSI] at both ambient and elevated pressures. On the other hand, the phase I–II transition arrests the majority of the dipolar reorientation in [N₂₂₂₂]­[FSI]. This correlates well with the ΔS of the II–I transitions, which is larger for [N₂₂₂₂]­[FSI] than for [N₂₂₂₂]­[TFSI] (see Table ).

At first glance, Figure c reveals that all three OIPCs display distinctly different pressure behavior. This is best illustrated by comparing the shapes of the gray-shaded areas in Figure c, which indicate the existence of the metastable phase I/II region for [C₂mmor]­[FSI] and the metastable II/III regions for [N₂₂₂₂]­[FSI] and [N₂₂₂₂]­[TFSI]. This indicates a unique effect of P and T on hysteresis for each of the OIPCs, with hysteresis decreasing with increasing P and T for [C₂mmor]­[FSI], remaining relatively constant for all P and T conditions for [N₂₂₂₂]­[FSI], and increasing with P and T for [N₂₂₂₂]­[TFSI]. Interestingly, both [N₂₂₂₂]­[FSI] and [N₂₂₂₂]­[TFSI] display negligible hysteresis in their I–II transitions (compression and decompression scans overlap) and noticeable hysteresis in their II–III transitions, despite the II–III transition being the dominant transition in terms of ΔS_SS in [N₂₂₂₂]­[FSI]. This suggests there is no clear correlation between the magnitude of ΔS_SS and ΔP_hys/ΔT_hys in these OIPCs.

The combination of isobaric and isothermal data provides a unique opportunity to compare observed (and practically relevant) values of ΔP_hys with those calculated from experimentally observed values of ΔT_hys and dT_ss/dP (which is the approach most commonly used in the BC field). Usefully for the application, direct observations of ΔP_hys for [C₂mmor]­[FSI] are lower than those calculated from the “quasi-direct” (isobaric) measurements. For example, for [C₂mmor]­[FSI] a ΔT_hys of 36 ± 2 °C is observed during cooling at ambient pressure, which would predict a ΔP_hys of 330 MPa (via ΔP_hys = ΔT_hys/[dT_ss/dP]), higher than the observed value of 260 MPa. It therefore appears that the phase transformation kinetic delay is smaller during a pressure change than during a temperature change. This may reflect, among other effects, the almost instantaneous effect of pressure on the structure (related to the material’s speed of sound), whereas thermal conduction delays the reduction in the internal temperature during cooling. This is supported by relatively consistent values of ΔP_hys observed for [C₂mmor]­[FSI] under different compression rates (data not shown). Interestingly, while a lower ΔP_hys is also observed for [N₂₂₂₂]­[FSI] than calculated from ΔT_hys [dT_ss/dp] (78 MPa vs 90 MPa), a larger ΔP_hys is observed for [N₂₂₂₂]­[TFSI] than predicted per the above method (96 MPa vs 69 MPa). This indicates that multiple factors contribute to this difference, including chemical structure.

A detailed analysis of T_SS vs P data, in terms of dT_SS/dP coefficients, reveals that the phase I–II transition of [N₂₂₂₂]­[TFSI] is more sensitive to pressure changes compared to the [FSI] analogue. This is reflected in a markedly higher dT_SS/dP­(I–II) of 28.2 °C/100 MPa for [N₂₂₂₂]­[TFSI], compared to 17.9 °C/100 MPa for [N₂₂₂₂]­[FSI]. Interestingly, the pressure sensitivity of both the I–II and II–III transitions in [N₂₂₂₂]­[FSI] is similar (dT_SS/dP (I–II)=17.9 vs (II–III)= 19.7 °C/100 MPa), while there is a difference in the dT_SS/dP coefficients characterizing the high- and low-temperature transitions in [N₂₂₂₂]­[TFSI] (dT_SS/dP (I–II): 28.2 °C/100 MPa vs (II–III)= 13.4 °C/100 MPa). In combination with the differences in ΔT_hys and ΔP_hys between [N₂₂₂₂]­[FSI] and [N₂₂₂₂]­[TFSI], these results highlight the pronounced effects that ion structures can have on the barocaloric properties of OIPCs.

The determined values of dT_SS/dP offer a unique opportunity to estimate the volume variations accompanying the SSTs using the Clausius–Clapeyron (C–C) equation, dP/dT = ΔS/(ΔV)⁻¹. A rearranged form of the C–C equation, ΔV_SS= dT_SS/dP·ΔS_SS, indicates that the volume change at the SST is directly proportional to both the magnitude of dT_SS/dP and ΔS_SS. Taking into account the values of dT_SS/dP and ΔS_SS parameters determined on decompression and heating pathways (the endothermic transitions), the highest value of ΔV_SS associated with any of the SSTs studied here is that of the III–II transition in [N₂₂₂₂]­[TFSI]. This is likely attributed to the larger size of the [TFSI]⁻ anion compared to the [FSI]⁻ anion; while both anions undergo a cis-trans (C ₁ -C ₂) disordering mode with similar energy barriers between the conformers, the bulky CF₃ groups increase the sweep volume of the [TFSI]⁻ anion during cis–trans interconversion. Thus, for OIPCs with structurally similar cations and [TFSI]⁻ or [FSI]⁻ anions, transitions that activate the cis-trans disordering mode of the anions would be predicted to have larger accompanying volume changes in the [TFSI]⁻ analogues. A large ΔV_SS is also estimated for II–I transition in [C₂mmor]­[FSI] (Table ), which has the highest ΔS_SS of any of the transitions studied. Thus, a correlation is observed between the magnitudes of ΔS_SS and ΔV_SS, indicating that the introduction of ion dynamics at the disordering transition is beneficial for increasing both of these key barocaloric properties. Indeed, the volume changes of approximately 0.03 cm³/g determined for the III–II and II–I transitions in [N₂₂₂₂]­[TFSI] and [C₂mmor]­[FSI] (Table ) are in good agreement with the values of ΔV_SS we have previously measured experimentally for these transitions. The values of ΔV_SS calculated for the III–II transition in [N₂₂₂₂]­[FSI] and the II–I transition in [N₂₂₂₂]­[TFSI] are an order of magnitude lower, consistent with the lower values of ΔS_SS for these transitions.

Conclusions

In this work, we have demonstrated that high-pressure broadband dielectric spectroscopy (HP-DS) is a robust methodology for probing solid–solid phase transitions in organic ionic plastic crystals, providing valuable insights for their application as barocaloric materials. By monitoring the temperature and pressure evolution of dielectric permittivity (ε’), we constructed detailed P-T phase diagrams for three OIPCs. The results reveal that the phase behavior and hysteresis are intimately linked to the chemical structure of the ion pair. The morphonium-based salt, [C₂mmor]­[FSI], exhibits a single, broad plastic phase with significant hysteresis that is less pronounced at elevated pressure, while the ammonium salts with different anions ([FSI]⁻ vs [TFSI]⁻) display starkly different transition sequences and pressure sensitivities. The remarkably high dT_ss/dP value for the I–II transition in [N₂₂₂₂]­[TFSI], together with the pressure-increased hysteresis in T_SS observed for the II–III transition, highlights how dramatically anion structure affects transition thermodynamics and kinetics. Finally, we found an excellent agreement between the volume changes (ΔV_ss) calculated from our dielectric-derived Clausius–Clapeyron analysis and experimentally measured values reported in the literature, confirming the quantitative reliability of our method. We therefore conclude that HP-DS is not merely a complementary technique but an indispensable one for the fundamental study and targeted development of OIPCs as BCs, as it directly probes the dynamic origins of their barocaloric properties under application-relevant conditions.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsami.6c02495.

• Additional details on dielectric experiments and full-frequency spectra recorded under ambient pressure conditions for all three examined compounds (PDF)

The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

The authors declare no competing financial interest.