Ternary Ionogel Electrolytes Enable Quasi-Solid-State Potassium Dual-Ion Intercalation Batteries

Sustainable battery materials and chemistries are required to complement the growing demand for renewable energy from solar farms, wind mills, and hydroelectric power stations which are characterized by either demand fluctuations or periodic supply interruptions. To keep a balance between supply and demand at all times, the intermittent nature of renewables should be leveled using stationary batteries deploying abundant, inexpensive, and nontoxic materials. Current stationary batteries rely heavily on expensive transition metals (e.g., nickel– cadmium or nickel–metal hydride batteries, and vanadium redox flow batteries), toxic elements (e.g., nickel– cadmium and lead–acid batteries), or require high temperature to operate (e.g., sodium–sulfur batteries operating at 300 350 C). In the interest of avoiding toxic and expensive minerals, there is a pressing need for sustainable battery materials that can provide comparable performance and cycle life. In this regard, the dual-ion battery (DIB) concept has emerged as a promising chemistry for future energy storage applications. In contrast to the “rocking chair” model in lithium-ion batteries, the energy storage mechanism in an archetype of a DIB is underpinned by the simultaneous intercalation of cations and anions from the electrolyte into, respectively, negative and positive electrodes containing graphite. In a lithium DIB, anion intercalation and extraction require an operating voltage window ranging from 3 to 5.2 V versus Liþ/Li with the extent of reversibility ultimately depending on the type of graphite, electrolyte-salt concentration, type of electrolyte solvent, type of anion in the electrolyte, working temperature, and amount of electrolyte in the cell. Reported gravimetric capacities for half-cells generally vary between 80 and 140mAh g 1 with discharge voltages averaging 4.5 V. On the basis of these metrics, lithiummetal DIBs can be optimized to deliver cell-level energy density and specific energy above 200Wh L 1 and 100Wh kg , respectively, better than lead–acid batteries (50–80Wh L 1 or 20–55Wh kg ) and comparable with Ni–metal hydride batteries (150–220Wh L 1 or 50–70Wh kg ) or Na–S batteries (150–300Wh L 1 or 80–150Wh kg ). Using graphite as the anion-hosting electrode, a wide selection of materials can be utilized in the negative electrode, which allows for increasingly diverse electrode– electrolyte combinations in the design of DIBs. In addition, the fact that the negative and positive electrodes host different ions during cell operation relaxes the requirements on the type of electrolytes used. Thus, a stationary battery based on a DIB technology presents significant strategic advantages such as the likelihood of replacing transition metal-containing cathodes with graphite or organic materials and a broad choice of electrolytes. These benefits are, in turn, anticipated to facilitate the transition to beyond Li-ion technologies harnessing more abundant Naþ-, Kþ-, Ca2þ-, or Al3þ-based resources. To maximize the energy density of KDIBs, the graphite cathode should ideally be paired with a K metal or graphite electrode. The use of K metal in KDIBs ensures unlimited supply of Kþ to A. Kotronia, K. Edström, D. Brandell, H. D. Asfaw Department of Chemistry-Ångström Laboratory Uppsala University Lägerhyddsvägen 1, Box 538, 75121 Uppsala, Sweden E-mail: antonia.kotronia@kemi.uu.se; habtom.desta.asfaw@kemi.uu.se


Introduction
Sustainable battery materials and chemistries are required to complement the growing demand for renewable energy from solar farms, wind mills, and hydroelectric power stations which are characterized by either demand fluctuations or periodic supply interruptions. [1]To keep a balance between supply and demand at all times, the intermittent nature of renewables should be leveled using stationary batteries deploying abundant, inexpensive, and nontoxic materials. [2,3]Current stationary batteries rely heavily on expensive transition metals (e.g., nickelcadmium or nickel-metal hydride batteries, and vanadium redox flow batteries), toxic elements (e.g., nickelcadmium and lead-acid batteries), or require high temperature to operate (e.g., sodium-sulfur batteries operating at 300À350 C). [3,4] In the interest of avoiding toxic and expensive minerals, there is a pressing need for sustainable battery materials that can provide comparable performance and cycle life.In this regard, the dual-ion battery (DIB) concept has emerged as a promising chemistry for future energy storage applications.In contrast to the "rocking chair" model in lithium-ion batteries, the energy storage mechanism in an archetype of a DIB is underpinned by the simultaneous intercalation of cations and anions from the electrolyte into, respectively, negative and positive electrodes containing graphite. [5,6]In a lithium DIB, anion intercalation and extraction require an operating voltage window ranging from 3 to 5.2 V versus Li þ /Li with the extent of reversibility ultimately depending on the type of graphite, electrolyte-salt concentration, type of electrolyte solvent, type of anion in the electrolyte, working temperature, and amount of electrolyte in the cell. [7,8]Reported gravimetric capacities for half-cells generally vary between 80 and 140 mAh g À1 with discharge voltages averaging 4.5 V. [6,9] On the basis of these metrics, lithium metal DIBs can be optimized to deliver cell-level energy density and specific energy above 200 Wh L À1 and 100 Wh kg À1 , respectively, better than lead-acid batteries (50-80 Wh L À1 or 20-55 Wh kg À1 ) and comparable with Ni-metal hydride batteries (150-220 Wh L À1 or 50-70 Wh kg À1 ) or Na-S batteries (150-300 Wh L À1 or 80-150 Wh kg À1 ). [6,10]Using graphite as the anion-hosting electrode, a wide selection of materials can be utilized in the negative electrode, which allows for increasingly diverse electrodeelectrolyte combinations in the design of DIBs.In addition, the fact that the negative and positive electrodes host different ions during cell operation relaxes the requirements on the type of electrolytes used.[13] These benefits are, in turn, anticipated to facilitate the transition to beyond Li-ion technologies harnessing more abundant Na þ -, K þ -, Ca 2þ -, or Al 3þ -based resources. [6,14,15]o maximize the energy density of KDIBs, the graphite cathode should ideally be paired with a K metal or graphite electrode.The use of K metal in KDIBs ensures unlimited supply of K þ to sustain long cycle life.[17] Kravchyk et al. reported a highly functioning KDIB composed of natural graphite and potassium metal electrodes in 5 M KFSI electrolyte, for which nearly 100 mAh g À1 discharge capacity was obtained with a mean discharge voltage of 4.7 V versus K þ /K. [16]A possible candidate to replace K metal electrode is the potassium graphite intercalation compound (KGIC) with a maximum composition of KC 8 and a theoretical capacity of %279 mAh g À1 below 0.5 V versus K þ /K in carbonate-based electrolytes. [18][21] It is worth noting, however, that the risk of forming potassium metal dendrites and excessive reaction with the electrolyte make potassium intercalated graphite unsafe to use in practical KDIBs.Therefore, safer electrode materials must be developed as alternatives to graphite and K metal.Layered transition metal dichalcogenides (TMDCs) [22] and transition metal oxychalcogenides (TMOCs) [23] have been recently identified as suitable hosts for K þ storage due to their large interlayer gaps.For instance, 2H-MoS 2 exhibits an interlayer spacing of %6.15 Å which is large enough for reversible K þ intercalation at potentials between 0.5 and 2 V versus K þ /K and therefore eliminates the risk of K plating. [24,25]MoS 2 reversibly forms the K 0.5 MoS 2 compound, which can deliver a maximum specific capacity of %80 mAh g À1 over extended cycling; therefore highlighting its potential for practical applications. [25]n a DIB, the electrolyte serves not only as an ion-conductor but also as a reservoir of ions required for electrochemical energy storage in the positive and negative electrodes.Thus, mostly highly concentrated electrolytes (HCEs) are used in DIBs along with thick and very porous separators. [6]The majority of HCEs contain unusually high amount of salt (3-5 mol L À1 ) dissolved in carbonate solvents which are prone to degradation beyond 4.5 V. [26] Under such circumstances, cell damage is inevitable as the salt and solvent molecules undergo oxidative degradation resulting in poor coulombic efficiency and graphite exfoliation in the long run. [6,12]Replacing liquid electrolytes with gel polymer electrolytes (GPEs) can potentially mitigate parasitic reactions involving free solvent molecules. [27,28]The use of GPEs is beneficial as it can help minimize parasitic reactions, increase thermal stability, lower risk of short-circuit failure, reduce electrolyte leakage, and create compact cell design. [29]There are few examples of GPEs proposed for use in semi-solid-state DIBs.A PVdF-HFP membrane consisting of up to 8% Al 2 O 3 nanoparticles was gelled in 1 M NaPF 6 in EC-DMC-EMC and applied as an electrolyte in Sn-graphite DIB for which a discharge capacity of approximately 100 mAh g À1 was obtained at 500 mA g À1 . [28]A quasi-solid-state sodium metal DIB demonstrated by Xu et al. was fabricated by in situ polymerization of ethoxylated pentaerythritol tetraacrylate (EPTA) monomer in 0.5 M NaPF 6 in a 1:1:1 mixture of propylene carbonate (PC), ethylmethyl carbonate (EMC), and fluoroethylene carbonate solvent (FEC) with 4 wt% 1,3-propanesultone (PS) additive. [30]The GPE, which was reportedly stable up to 5.5 V versus Na þ /Na, helped realize a DIB that could deliver 93 mAh g À1 reversible capacity at 10 mA g À1 after 100 cycles and 78 mAh g À1 reversible capacity at 100 mA g À1 over 1000 cycles at %87% retention.Du et al. used a GPE composed of a potassium perfluorinated sulfonate resin soaked in 1 M KPF 6 in EC-DEC to design a KDIB featuring a graphite negative electrode and an organic positive electrode based on 3,4,9,10-perylenetetracarboxylic 3,4:9,10-dianhydride (PTCDA). [31]The semi-solid-state graphite-PTCDA battery could deliver up to 90 mAh g À1 at 800 mA g À1 with an output voltage of approximately 2.2 V. Another type of GPE that is of particular interest in DIBs is the so-called ionogel electrolyte (IGE).A ternary system, the IGE consists of a liquid phase composed of a salt dissolved in ionic liquid (IL) and a solid phase based on poly (ionic liquid) (polyIL) with cationic backbone and mobile anions. [32]The polyIL creates a mechanically stiff matrix for the liquid phase, while the IL acts as a plasticizer and a solvent for the salt which is anticipated to promote high ionic conductivity and to provide a reservoir of cations and anions needed for cell operation. [28,32,33]Different types of aprotic IGEs have been studied as candidates for application in Li-and Na-ion batteries [34,35] and supercapacitors, [36] and can be cheaper alternatives to HCEs in DIBs.In this article, we report a full-cell KDIB based on MoS 2 negative electrode and graphite positive electrode in a ternary ionogel electrolyte (t-IGE) consisting of potassium bis(fluorosulfonyl)imide (KFSI), 1-butyl-1-methylpyrrolidinium bis(fluorosulfonyl)imide (Pyr 14 FSI), and poly[diallyldimethylammonium bis(trifluoromethanesulfonyl)imide] (pDDA-TFSI).The selection of the t-IGE components aimed to fulfill the criteria of 1) environmental friendliness (pDDA-Cl can be synthesized in water and precipitated to pDDA-TFSI), 2) safety and high oxidative stability (ensured by the incorporation of FSI and TFSI anions that do not hydrolyze readily), and 3) practically meaningful performance at near-room temperature (achieved due to the use of the IL plasticizer and the advantageous concept of anion mixing). [8]he following reactions take place at the positive and negative electrodes.
Negative electrode Positive electrode In the equations depicted above, x and n stand for the number of moles of electrons or ions exchanged at the interface, and the number of moles of graphite in the positive electrode.The physicochemical properties such as thermal stability, glass transition temperature, and ionic conductivities were investigated for varying compositions of KFSI/Pyr 14 FSI/pDDA-TFSI.The performance of the t-IGEs was investigated in prototype pouch cells in half-and full-cell configurations and compared with KDIBs making use of a reference HCE, 5 M KFSI in EC-EMC-DMC solvent mixture.Initially, the kinetics of FSI intercalation in graphite were studied in half-cells (vs metallic K) at 40 C and the most promising ionogel was further tested in full cell format (MoS 2 vs graphite).Performance metrics including the initial discharge capacity and rate capability are promising, and indicate the feasibility of the chosen approach for future applications.

Design of Ternary Ionogel Electrolytes
The fabrication procedure followed to prepare the t-IGE films is schematically shown in Figure 1a.In all t-IGEs, the KFSI to pDDA-TFSI weight ratio was fixed at 30:70, while the ratio between Pyr 14 FSI and pDDA-TFSI was varied from 10:90 to 50:50.Increasing the Pyr 14 FSI to pDDA-TFSI ratio greatly influenced the mechanical properties and appearance of the ionogel films; while all films remained freestanding after evaporating the acetonitrile solvent, the ionogels with a low pyr 14 FSI: pDDA-TFSI ratio (10:90 and 20:80) proved to be rigid and brittle as compared with the films with a higher Pyr 14 FSI content.For compositions varying from 30:70 to 50:50, the resulting t-IGE films were soft, somewhat sticky and optically transparent, possibly due to the plasticizing effect of Pyr 14 FSI.Typically, the 40:60 ionogel film shown in Figure 1a was observed to be flexible and transparent at room temperature.The 50:50 Pyr 14 FSI: pDDA-TFSI ratio appeared to be an upper limit, as handling of the films without causing mechanical damage proved difficult.

Microstructures of Ternary Ionogel Electrolytes
Different film morphologies and microstructures were observed depending on the composition of the ionogels, as can be seen in the top and cross-sectional scanning electron microscopy (SEM) images in Figure 1b-d and in Figure S1-S3, Supporting Information.The top-view and cross-sectional micrographs of the 10:90 t-IGE with an approximate thickness of 350 μm (Figure 1b,d) revealed a rather homogeneous surface with largely smooth texture.It is likely that this type of morphology arose from the confinement of Pyr 14 FSI in pDDA-TFSI, which is in accordance with similar observations in other ionogel systems. [37]The confinement of the ionic liquid in discrete domains, surrounded by a continuous pDDA-TFSI phase, could explain the rigidity of the ionogel films.The SEM images for the 40:60 ionogel film shown in Figure 1c,e revealed a clear morphological change as a second type of domain with more granular texture emerged along with the smooth surface observed for the 10:90 t-IGE film.These domains were visibly porous and exhibited lamellar morphology, suggesting the formation of a continuous percolating network of pyr 14 FSI phases.Furthermore, the cross section of the 40:60 t-IGE, as shown in Figure 1e, appeared to be significantly thinner %150 μm and rugged, most likely as a result of the higher amount of liquid used to prepare the film.The elemental energy-dispersive X-ray spectroscopy (EDS) maps of the cross section of a graphite electrode infiltrated with the 40:60 t-IGE, as shown in Figure 1f, verified homogeneous distribution of the three components of the electrolyte throughout the porous graphite electrode, all the way down to the Al current collector.The SEM images showing the microstructures of the rest of t-IGE films (20:80, 30:70, and 50:50 compositions) can be found in Figure S1-S3, Supporting Information.

Thermal Stability and Ionic Conductivities of t-IGEs
The thermal stability of the ionogels was assessed using thermogravimetric analysis (TGA) in air, as shown in Figure 2a.The decomposition of the pristine pDDA-TFSI powder started at %270 C, while onset temperature for the KFSI/Pyr 14 FSI/ pDDA-TFSI t-IGEs was found at %207 C. The occurrence of mass loss at such high temperatures for both pDDA-TFSI and the ionogels might suggest a very low water content, which is a prerequisite for the electrolytes to be used in K-based electrochemical devices.The lower onset temperature of decomposition observed for the ionogel decomposition was largely due to the presence of the 1-butyl-1-methylpyrrolidinium cation in the ionic liquid, which is more susceptible to thermal degradation than its polymerized counterpart in pDDA-TFSI.It must, however, be emphasized that the thermal stability of the ionogels by far surpasses that of conventional liquid electrolytes.The operational temperature range of electrolytes based on EC-DMC, for example, is restricted by the low boiling point of the solvents (89 C for DMC).In the presence of LiPF 6 , the most commonly used salt in commercial lithium-ion batteries, electrolyte decomposition is further accelerated due to the formation of highly reactive PF 5 species. [38]Even in cases where less reactive salts (e.g., LiTFSI) and higher salt concentrations are used, mass loss is still initiated at far lower temperatures in liquid carbonate electrolytes (%100-150 C). [39] The excellent thermal stability of the ionogels  À1 .In each plot, the spectra were normalized with respect to the most intense band.Note that the t-IGE compositions (10:90-50:50) correspond to the weight ratios between the ionic (pyr 14 FSI) and polyionic liquid (pDDA-TFSI) components.
was further confirmed using differential scanning calorimetry (DSC) measurements.As shown in Figure 2b, the normalized DSC heating traces (À80-190 C, first cycle) exhibited two broad but distinct endothermic peaks, the first being situated between À60 and À25 C, which coincided rather well with reported values for the T g of similar systems studied within the context of Liion batteries. [34,40]Deviations from the results in these publications can be ascribed to the use of potassium instead of lithium salts and also to slight differences in the relative amount of salt, ionic liquid, and poly(ionic liquid).Similar to what was reported for Li þ , K þ may have the ability to strongly interact with the ionic liquid and poly(ionic liquid), thus partially counteracting plasticizing effects, thus explaining the trends observed in Figure 2b.The second endothermic feature was located between 100 and 150 C and its origin is not yet determined.Possible reasons include phase transitions, such as the melting of KFSI (%100 C for pure KFSI), in case salt aggregates exist in the sample.
The ionic conductivity measurements in Figure 2c revealed, as anticipated, that the ionic conductivity in t-IGE films increased with increasing content of IL component.Practically, meaningful conductivities (0.1-1.0 mS cm À1 ) were observed for the 30: 70-50:50 compositions, within a temperature range of 30-80 C.This is in good agreement with the ionic conductivities reported in works investigating the Li-analogue of the studied ternary ionogel, which also resulted in 0.1-1 mS cm À1 between RT and 80 C. [34,40] The 50:50 and 40:60 t-IGE exhibited similar conductivities (% 0.1 mS cm À1 ) at an operating temperature of 40 C. The remaining ionogels resulted in conductivities below 0.1 mS cm À1 and were thus not expected to deliver satisfactory performance in near-room temperature KDIBs, which was further reflected in the electrochemical tests discussed in the following section.It must, however, also be noted that, even though the t-IGEs did not show signs of degradation up until 207 C, the conductivity measurements performed on the 40:60 and 50:50 compositions exhibit a rather unstable behavior at elevated temperatures.This is likely due to the initiation of material flow, which could also be aggravated by the high and nonhomogeneous pressure exerted in the coin cell.
Fourier transform infrared spectroscopy (FTIR) measurements (Figure 2d and S4, S5, Supporting Information) provided a means of probing interactions between the pDDA-TFSI polymer backbone, the KFSI salt, and the pyr 14 FSI ionic liquid at a molecular level.A tabular summary of the observed bands and their assigned modes is shown in Table S1, Supporting Information, which proved to be in good agreement with literature reports. [41]In short, the spectra of the IGEs were, to a large extent, the direct superposition of the components present in pDDA-TFSI and bands originating mostly from the vibrational modes of the FSI anion in the IL.Shifts were observed for a few bands, including a band related to the skeletal movement of the quaternary ammonium moiety in pDDA which shifted from 1178 to 1172 cm À1 .Other bands that shifted include the asymmetric stretch of the sulfonyl moiety in TFSI, which moved from 1133 to 1136 cm À1 .Lastly, another peak shifted toward higher wavenumbers was detected from 2872 cm À1 in pDDA-TFSI to 2881 cm À1 in the 50:50 t-IGE.Such peak shifts suggested the slight rearrangement of pDDA-TFSI and dipole-dipole interaction with the ionic liquid, leading to the formation of the ionogel.

Electrochemical Stability of t-IGEs and Performance in Half-Cell KDIBs
The stability of the 40:60 t-IGE upon reductive conditions was tested by cycling K versus Al.The first cyclic voltammograms (Figure 3a) measured at 0.05 mV s À1 indicated that the IGE was stable down to 0 V versus K þ /K.In addition, the reversible plating and stripping of K metal on the Al-substrate could be observed at potentials between À0.5-0.5 V versus K þ /K, which confirmed that no such processes would be present in the MoS 2 versus graphite dual-ion cells.The oxidative stability of the 40:60 t-IGE was tested in a similar manner (in a K-Al cell) between 3.0 and 5.0 V versus K þ /K (Figure 3b).The maximum anodic current decreased from 0.1 μA cm À2 for the first cycle to %0.02 μA cm À2 for the third cycle.The areal current was substantially lower as compared with the values reported for ionic liquid-based electrolytes (e.g., 1 mA cm À2 for 1 M LiFSI in Pyr 14 FSI), [26] which indicated that the gel electrolyte suppressed parasitic reactions.This finding is particularly encouraging, as most of the state-of-the-art electrolytes available for DIBs (and in particular FSI-based alternatives) tend to cause extensive damage to the Al current collector during long-term cycling. [42]After the first cycle, the onset potential for oxidation shifted from 3.0 to approximately 4.7 V versus K þ /K as parasitic reactions involving impurities from the electrolyte or the Al current collector decreased or ceased to occur.In addition, as Al is known to itself be reactive at the high potentials needed for the operation of DIBs, the anodic stability of the 40:60 t-IGE was also assessed versus a positive electrode consisting of glassy carbon, which resulted in a similar behavior (see Figure S6, Supporting Information).Proof-of-concept half-cells were constructed using K metal disks as negative electrodes and graphite as positive electrodes and cycled at 40 C. The results, shown in Figure 3c-f, demonstrated that the ionic conductivity of the electrolyte influenced the performance of the KDIBs.With increasing ionic liquid content, the voltage hysteresis in the galvanostatic curves decreased, while the specific capacity increased from 6 mAh g À1 for the 10:90 t-IGE to almost 60 mAh g À1 for the 40:60 t-IGE during the initial discharge process.The measurements helped decide that the 40:60 t-IGE was the most promising for practical application at near-room temperature, and it was selected for further investigation in full dual-ion cells.

Full-Cell MoS 2 -Graphite KDIBs Using Highly Concentrated Electrolyte: A Reference Study
In Figure 4, results from galvanostatic tests on prototype graphite-graphite and MoS 2 -graphite KDIBs with a concentrated liquid electrolyte (5 M KFSI in EC:DMC:EMC (1:10:10 v/v)) are presented, which should be perceived as a reference system.With a reversible capacity of 250 mAh g À1 and above and working voltage 0À0.5 V versus K þ /K, graphite can be convenient to design compact and high energy density KDIBs.As shown in Figure 4b, the dual-graphite cell could deliver a discharge capacity of %99 mAh g À1 with CE of %73% for the second cycle with a mean voltage of %4.5 V.The initial charge and discharge capacities were %166 and %93 mAh g À1 , respectively, for which the iCE was %56%, demonstrating that there were significant irreversible reactions.It is worth noting that potassium-intercalated graphite compounds are known to be very reactive in carbonate-based electrolytes, and the iCE values associated with the first cycle are mostly lower than 80% [43] and as a result the irreversible capacity loss of the KDIB during the first cycle is partly due to the negative electrode.The galvanostatic curves for a 3-electrode cell with MoS 2 and graphite electrodes shown in Figure 4c indicated that the first cycle was characterized by an iCE of %54% due to an interplay of irreversible reactions at both electrodes.In the second cycle, the reversible capacity was %75 mAh g À1 with an average voltage %3.5 V and a CE of nearly 85%, which was higher than what was observed for the dual-graphite cell.Further rate tests were conducted on the full cells, as shown in Figure 4d,e.The MoS 2 -graphite cells could provide about 80À90 mAh g À1 at specific currents of 5 to 40 mA g À1 in a working voltage range from 0.5 to 4.7 V versus K þ /K.At higher applied currents, namely, 80, 160, and 300 mA g À1 , the capacities obtained were, respectively, %69, 46, and 17 mAh g À1 .The performance metrics were on a par with the most successful K-based dual-ion batteries, [16] while the use of MoS 2 helped avoid the risk of potassium metal plating associated with graphite electrodes (a stage-I graphite intercalation compound [GIC] is formed close to 0 V versus K þ /K in 2-electrode half-cells). [44]4.3.Design of Quasi-Solid-State MoS 2 -Graphite KDIBs Using t-IGEs Following the electrochemical characterization of the MoS 2graphite system in a reference HCE, demonstrator quasi-solidstate full-cells were constructed employing the IGEs.Two different approaches were pursued in the design of the MoS 2graphite KDIBs: cells with only the graphite electrodes infiltrated with the t-IGEs and cells with both MoS 2 and graphite electrodes infiltrated with the t-IGEs.The electrolyte volume was varied to achieve optimum cell performance.As shown in Figure 5, the removal of the problematic K interface led to considerably improved performance in full cells (MoS 2 vs graphite), with the initial discharge capacity (normalized to the mass of graphite) reaching 80 mAh g À1 for semi-solid-state devices cycled at 40 C with the 40:60 IGE.The results should be encouraging as they compared well with capacities reported for similar KDIBs operating entirely on liquid electrolyte.Thus, this work highlights the potential use of t-IGEs in future practical KDIBs.[13,17,20,45] The solution casting method normally leads to films with a thickness of %200 μm.To decrease the thickness, an improved electrolyte design was attempted by infusing the 40:60 t-IGE into a 40 μmthick plain-weave glass fiber fabric.Overall, an electrolyte thickness reduction of 64% was achieved in this study.It is anticipated that the glass fiber fabric not only helps reduce the electrolyte thickness, but will also provide a physical barrier to eliminate risks of short-circuit and boost the mechanical performance of the KDIB.The impact of the thickness of the t-IGEs can be seen in the galvanostatic cycling in Figure 5b-d.The 60 and 120 μmthick t-IGEs reinforced with glass fiber fabrics showed discharge capacities lower than the completely freestanding, %170 μmthick t-IGE, implying that the amount of electrolyte in the KDIBs might not be sufficient to sustain cell functioning.A theoretic estimation of the capacity and specific energy (discussed in Section S4, Supporting Information) implied similar trend.However, a continuous capacity fading was observed after extended cycling, with all cells decreasing to 40 mAh g À1 over 20 cycles, and the coulombic efficiency remained below 90% for all cells, as shown in Figure 5d.The poor coulombic efficiency (in particular during the first cycle) and the subsequent capacity fading could arise from structural degradation of the layered MoS 2 electrode, and electrolyte starvation as only the graphite electrode was infiltrated with the ionogels.
To take advantage of the beneficial characteristics of the glass fiber fabrics, we further looked into the fabrication of more compact KDIBs with both electrodes infused with the t-IGE in the presence of the fabric separators, as shown in Figure 6a.A selection of the galvanostatic curves for charge-discharge test at 20 to 160 mA g À1 is shown in Figure 6b.Accordingly, the discharge capacities and associated coulombic efficiencies varied from %60 mAh g À1 and 91% at 20 mA g À1 , %49 mAh g À1 and 96% at 40 mA g À1 , %36 mAh g À1 and 97% at 80 mA g À1 to %25 mAh g À1 and 99% at 160 mA g À1 .After long-term cycling at the specified currents, the cell was able to recover to its stable initial capacity, i.e., after 70 cycles the discharge capacity was %60 mAh g À1 .It is worth noting that the capacity increased continuously over 45 cycles stabilizing at %60 mAh g À1 .Such behavior could be indicative of increased utilization of the MoS 2 active material as it underwent interlayer expansion and became more accessible to the ionic liquid.All in all, the potential value of using ternary IGEs in KDIBs was demonstrated, which included safety, processability, and electrochemical property considerations.Conclusive evidence was obtained to show that the t-IGEs were suitable for use in in KDIBs, as they exhibited high conductivities, a large electrochemical stability window, and delivered reversible discharge capacity comparable with KDIBs based on concentrated liquid electrolytes. [13,19,20,46]There is a lot of room for improving the performance of the t-IGEs in the future by incorporating various salt and cosolvent additives that form stable solid electrolyte interphase on the negative electrode and cathode-electrolyte interface layer on the positive graphite electrode.

Conclusions
A ternary IGE enables the design of functioning MoS 2 -graphite KDIBs with promising cycle life and rate capability.The cell design is anticipated to provide compact KDIBs with improved safety and structural integrity.The ionogels were composed of KFSI, Pyr 14 FSI, and pDDA-TFSI, with the KFSI:pDDA-TFSI fixed at a weight ratio of 30:70, and the pyr 14 FSI:pDDA-TFSI weight ratio ranging between 10:90 and 50:50.The salt, ionic liquid, and poly(ionic liquid) were dissolved in acetonitrile to prepare the precursor solutions to the ionogels, which were infused under vacuum into the graphite and MoS 2 electrodes cast on aluminum current foils.All ionogel films exhibited thermal stability up to %210 C. At 40 C, the 40:60 composition resulted in a gel with conductivity of %0.1 mS cm À1 and with good processability and mechanical stiffness.Increasing the fraction of the ionic liquid was a key factor in enhancing the ionic conductivity of the IGEs and lowering interfacial resistances.Electrochemical investigations performed at 40 C in half-cells (graphite vs K), showed that the ionogels can function at least up to 5.1 V versus K þ /K.However, the lower ionic conductivity at near-room temperatures limited the performance of the 10:90À30:70 t-IGEs in half-cells and best attainable capacity was 60 mAh g À1 for the 40:60 t-IGE.Further rate performance and achievable capacity tests were performed in full-cell MoS 2 versus graphite KDIBs in which only the graphite electrodes were infused with the 40:60 t-IGE.In such cell configurations, the initial discharge capacity was 80 mAh g À1 at a specific current of 10 mA g À1 .The thickness of the electrolyte in the cells averaged %170 μm.The impact of electrolyte thickness on cell performance was studied by varying the amount of electrolyte precursor solution infiltrated into plain-weave glass fiber fabrics.Though a significant reduction was achieved in electrolyte volume, %64%, the cell performance was lower than in the absence of the glass fiber reinforcement, which was, in turn, indicative that the amount of ions in the electrolyte was lower than what was required for adequate cell performance.Following that was an effort where the 40:60 t-IGE was infused into both the negative and positive electrodes in the presence of a ply of glass fiber fabric.Significant performance enhancement was observed for tests at specific currents of 20À160 mA g À1 .For instance, a stable capacity of nearly 60 mAh g À1 was obtained at 20 mA g À1 even after the end of the rate capability test after 70 cycles.The performance of the quasi-solid-state MoS 2 -graphite KDIB was on a par with other potassium dual-graphite cells based on highly concentrated liquid electrolytes.Future works must focus on improving the longterm performance of KDIB full cells with IGE to increase the coulombic efficiency beyond 90% at low specific currents and extend the capacity retention over a greater number of cycles.

Experimental Section
K Electrodes: Potassium (Alfa Aesar) was cut into cubes inside an Arfilled glove box (O 2 < 0.2 ppm, H 2 O < 0.2 ppm).Oil residues were removed by soaking the K cubes in cyclohexane for 5 min and any remaining surface oxide was removed with a scalpel.The K cubes were placed on top of Al-foil and pressed to a foil.K disks with a diameter of 13 mm were punched and used immediately in half-cell KDIBs.
Graphite and MoS 2 Electrodes: A total of 1.0 g of 90% KS6 graphite (Timcal from Imerys), 6%, Super P (Alfa Aesar), and 4% sodium carboxymethyl cellulose (CMC, Leclanché) was mixed with 2.7 mL of a 9:1 solution of deionized water and ethanol.The mixture was homogenized by shaking for 30 min at 25 Hz in a Retsch MM 400 shaker.The slurry was bar-coated on carbon-coated Al-foil (MTI), using a 50 μm spacing for the positive graphite electrode.The coatings were dried at ambient condition, punched into electrodes, and introduced in a glove box, where they were redried for 12 h at 120 C and under vacuum.The MoS 2 (Aldrich, whose SEM, X-ray diffraction [XRD], and Raman spectrum are shown in Figure S8-10, Supporting Information) electrodes were fabricated in a similar manner, with the only difference being the composition of the slurry (85% MoS 2 , 10% Super P, and 5% CMC) and the shaking parameters (30 min, 30 Hz).
Electrolyte Films: Anhydrous KFSI (99%, Provisco) was dried under vacuum for 24 h at 60 C, while Pyr 14 FSI (99.9%,Solvionic) was used as received, as its H 2 O content was measured below <2 ppm, through Karl Fischer coulometry.Poly[diallyldimethylammonium bis(trifluoromethanesulfonyl)imide] or pDDA-TFSI (99.5%, Solvionic) was dried under vacuum for 12 h, at 120 C. The dry materials were dissolved in anhydrous acetonitrile or ACN (Sigma-Aldrich) to prepare solutions with different compositions (see Table 1).The ratio between pDDA-TFSI and ACN was kept constant for all solutions (0.12 g pDDA-TFSI/ 2 mL ACN).The solutions were kept stirring for 1 h at 300 rpm and at ambient temperature prior to casting.
To prepare the ionogels films, the precursor solutions were cast in Teflon molds with a 20 mm diameter.The weight ratio between KFSI and pDDA-TFSI was kept to be a constant (30:70) in all instances, while the ratio between Pyr 14 FSI and pDDA-TFSI was varied from 10:90 to 50:50; see Table 1).For conductivity measurements, t-IGEs were cast onto Al-foil.In addition, the ionogels were infiltrated under vacuum into the positive graphite and MoS 2 negative electrodes (for cell assembly).In most cases, 2 mL of solution was used, which was necessary to acquire a pinhole-free film with a thickness of %200 μm.In some electrodes, a %40 μm-thick plain-weave glass fiber fabric (GF-PL-25-100, 25 g m À2 , EasyComposites) was incorporated as structural reinforcement when the thickness of the t-IGE films was reduced from typically 170 to 60 μm.To produce the glass fiber-reinforced composite, the weave glass fiber was attached on top of the graphite electrode in the Teflon molds, and impregnated with 1-1.5 mL of IGE to achieve variable film thickness.
SEM: Morphological aspects of the samples were studied with a Zeiss Merlin SEM coupled with EDS.Micrographs of the top and cross-sectional views were acquired through the in-lens and secondary electron detectors, operated in the high-resolution mode, with an acceleration voltage of 2.0 kV, a probing current of 80 pA, and at a working distance of 6.5 mm.EDS was performed in analytic mode, with an acceleration voltage of 10 kV, probing current of 1.0 nA, and at a working distance of 8.5 mm.All samples were handled under glove box atmosphere and brought inertly to the SEM.
TGA: The thermal stability of the pristine pDDA-TFSI powder and the t-IGE membranes was tested using a TA Q500 apparatus.A total of 3-5 mg of either powder or ionogel membrane was placed in the TGAcrucibles and dynamic heating was applied at a scan rate of 5 C min À1 and for a temperature range between RT and 500 C.All measurements took place in an air atmosphere, meaning that the samples were not handled under inert conditions in this case.
DSC: DSC measurements were performed on the pristine pDDA-TFSI powder and on the ionogel films, with the aid of a TA Q2000 instrument.Approximately 8 mg of each composition to be investigated was placed in a DSC crucible and sealed under an Ar-atmosphere in the glove box.The samples were initially cooled to À80 C at a rate of 5 C min À1 and subsequently heated to 200 C at a rate of 10 C min À1 .The cooling-heating cycle was repeated for all samples, and it is therefore worth pointing out that the data presented in Figure 2b correspond to the first cycle.
FTIR: IR spectra were collected in the range 650-4000 cm À1 , with a resolution of 1 cm À1 , using a Perkin Elmer Spectrum One spectrometer.As no setup allowing for inert sample handling was available, sample exposure to air was inevitable, but swift handling of all samples was ensured.
Raman Spectroscopy: Raman spectrum of the pristine MoS 2 powder was collected on a Renishaw InVia Ramanscope using a 50 mW laser with a wavelength of 532 nm.
XRD: The XRD pattern of the as-received MoS 2 was acquired with a Bruker D8 Advance diffractometer, after dispersion of the powder in ethanol and application onto a Si sample holder.The pattern was collected in the range between 5 and 90 with a step size of 0.015 .
Cell Assembly and Electrochemical Testing: Impedance spectra were recorded with a Schlumberger SI 160 instrument and a blocking electrode configuration, for a temperature range from 30 to 100 C. Measurements were performed in the frequency range 0.1 Hz-10 MHz, with a 10 mV amplitude.Debye fitting enabled the extraction of the PIL membrane resistance and therefore its conductivity.For electrochemical testing Bio-Logic MPG2-1 and Arbin BTS Cycler were used.The intercalation of FSI in graphite was initially studied using galvanostatic cycling in half-cells (K vs graphite) at a rate of C/100 (assuming %100 mAh g À1 capacity for FSI intercalation).Then, full cells (a stack thickness of %280 μm) were constructed using the MoS 2 and graphite electrodes infiltrated with the t-IGEs studied galvanostatically at various specific currents.Areal mass loadings of the graphite and MoS 2 electrodes used in the semi-solid-state full cells were roughly %1.65 and %0.92 mg cm À2 .Reference full-cells were assembled and tested similarly using HCE.In the KDIBs using 5 M KFSI EC-DMC-EMC electrolyte, the MoS 2 and graphite electrodes had mass loading of 4.68 mg (%1.5 mg cm À2 ) and 3.96 mg (%1.26 mg cm À2 ), respectively, amounting to a ratio of %1.2.In the case of the dual-graphite cells using same electrolyte, the mass ratio of the positive graphite electrode (%2.3 mg cm À2 ) to the negative graphite electrode (%1.4 mg cm À2 ) was %1.6.

Figure 1 .
Figure 1.Preparation and microstructure of t-IGEs: a) schematic summary of preparation of t-IGEs starting from dissolving the salt, ionic liquid, and poly(ionic liquid) in acetonitrile and followed by casting in a PTFE mold and solvent evaporation to obtain a transparent film.The photos show freestanding and flexible ionogel films composed of 30 wt% KFSI, 28 wt% pyr 14 FSI, and 42 wt% pDDA-TFSI (40:60 composition).The microstructural features of selected t-IGE specimens: b) top surface SEMs of the 10:90 composition and c) the top surface SEMs of the 40:60 composition.d) Crosssectional views of the electrode and ionogels for the 10:90 composition and e) cross-sectional SEM of the electrodes impregnated with the 40:60 t-IGE.f ) EDS maps performed on the cross section of the 40:60 t-IGE show the elemental distribution.

Figure 2 .
Figure 2. Physical characterization of t-IGEs: a) TGA curves for pure pDDA-TFSI powder and t-IGEs.b) DSC of the pristine pDDA-TFSI powder and t-IGE films.c) Results from the ionic conductivity measurements performed on the t-IGEs in the temperature range 30-100 C. d) IR spectra of the t-IGE electrolytes recorded in the range of 1800-650 cm À1.In each plot, the spectra were normalized with respect to the most intense band.Note that the t-IGE compositions (10:90-50:50) correspond to the weight ratios between the ionic (pyr 14 FSI) and polyionic liquid (pDDA-TFSI) components.

Figure 3 .
Figure 3. Electrochemical stability and performance of t-IGEs in graphite-K half-cells.a) Cyclic voltammograms recorded between À0.5 and 2 V versus K þ /K and with a scan rate of 0.05 mV s À1 , demonstrating the electrochemical stability of the electrolyte upon reductive conditions and the plating and stripping behavior of K metal.b) The cyclic voltammograms recorded at 0.05 mV s À1 between 2-5 V versus K þ /K and showing the anodic behavior of Al metal disk.c-f ) Galvanostatic curves for cycles 1, 2, and 5 measured at 1 mA g À1 for KDIBs with various ionogel electrolytes, positive graphite electrodes, and K metal negative electrodes.All measurements were conducted at a temperature of 40 C. Note the difference in the scaling of the x-axis for plots (c)-(f ).

Figure 4 .
Figure 4. a) Schematic illustration of the working principle for the MoS 2 -graphite dual-ion cell.Room-temperature electrochemical testing of dual-ion intercalation battery concept in 5 M KFSI in EC/DMC/EMC (1:10:10 v/v) electrolyte: b) typical galvanostatic curves for a graphite-graphite full cell cycled at 5 mA g À1 , c) galvanostatic curves of a 3-electrode MoS 2 -graphite cell recorded at 10 mA g À1 highlighting the electrochemical potential required for FSI intercalation into natural graphite, K þ intercalation into 2H-MoS 2 and showing the overall cell chronopotentiograms for the first two cycles, d) first and second cycle galvanostatic curves of full-cell MoS 2 -graphite KDIB tested at 5 mA g À1 , and e) rate capability and cycle life tests showing extended cycling of the KDIB in (d) at 10À300 mA g À1 .

Figure 5 .
Figure 5. Galvanostatic cycling curves recorded at 10 mA g À1 and 40 C for KDIBs with ionogel electrolytes infiltrated merely in the graphite positive electrodes and cycled against MoS 2 negative electrodes: selected galvanostatic curves for the a) 40:60 IGE with a film thickness of 174 μm without glass fiber fabric separator, b) 40:60 IGE with a film thickness of 129 μm including a glass fiber fabric separator, c) 40:60 IGE with a film thickness of 63 μm including a glass fiber fabric separator, and d) extended cycle performance of the KDIB cells shown in (a)-(c).

Figure 6 .
Figure 6.Manufacturing and rate capability test on MoS 2 -graphite pouch cell KDIBs with both electrodes infused with the t-IGE in the presence of plainweave glass fiber fabrics: a) schematic description of the various parts of the KDIB, b) galvanostatic curves for the rate capability test at 20-160 mA g À1 , and c) plots of discharge capacity and coulombic efficiency over 100 cycles.Electrochemical testing was performed at 40 C.