The key to construct the novel high-performance I2-cathode battery lies in two aspects: electrolyte and anode. Concerning the selection of aqueous electrolyte, it should guarantee the highest conversion efficiency of I−/I0/I+, so that give full play to the potential advantages of high theoretical capacity and redox potential of the I2 cathode. That is to say, for the I−/I0 conversion, the cation in the electrolyte must reduce the conversion energy barrier. As a result, the discharge voltage could be promoted due to the reduced polarization.5 For the I0/I+ conversion, the anion in the selected electrolyte must benefit the dissociation of I+ compounds so as to accelerate the I0/I+ conversion.2, 3, 5 Regarding the anode, only when it satisfies simultaneously the demands of high structural stability, immunity to various iodine anionic species, low redox potential, high capacity and low cost, all issues of the two reported I2-cathode battery systems mentioned above can be ultimately avoided. Compared with inorganic materials, organic materials featuring structural diversity, flexible molecular stucture and low cost are a promising candidate for this novel aqueous I2-cathode battery system.30–33
Density Functional Theory (DFT) calculations were performed to predict reaction energy profiles of I−/I0 conversion with the existence of different cations within the electrolyte, for the selection of cation candidate. In addition, the dissociation energy of InX (X denotes anion of electrolyte, n denotes electron number) were compared in different anions environment, to identify the best anion. As shown in Fig. 1a, the lowest value of Gibbs free energy change (∆G) is observed when the I−/I0 conversion chemistry occurs in K+ environment. Thus, the aqueous electrolyte containing K+ should hold great potential due to the fastest reaction rate of the I−/I0 conversion.5, 20 For anions, the lowest dissociation energy of ICl is evidently observed when compared to that of other iodides, indicating the fast conversion of I0/I+ in aqueous electrolyte containing Cl− (Fig. 1b).2, 3, 5 Accordingly, the aqueous KCl electrolyte should well guarantee high performances of I2 cathode. Moreover, linear sweep voltammetry (LSV) test shows that the total operational voltage window is up to 2.72 V in saturated KCl electrolyte, which is large enough to match most aqueous batteries (Fig. S1). 12, 34–36
Then, it is necessary to determine an appropriate anode that can insert/release K+ reversibly in aqueous electrolyte. Due to the sluggish kinetics of the larger Cl− compared with K+, materials capable of intercalation/de-intercalation of Cl− were ruled out.37 As shown in Fig. 1c, organic aromatic molecules with diverse number of aromatic rings and carbonyl groups (DPPZ,38 β-PTCDA,38 PNTCDA,39 PT32), inorganic and composite materials (KTi2(PO4)3,40 [email protected]41) were compared. Clearly, PTCDI (3,4,9,10-Perylenetetracarboxylic diimide) displays the lowest redox potential and highest specific capacity, thus holding a great promise as anode in the aqueous KCl electrolyte.34, 42 X − ray diffraction (XRD) and Fourier transform infrared (FT-IR) spectroscopy were performed to determine the crystal structure and functional groups of PTCDI (Fig. S2). Besides, electrochemical performances of PTCDI electrode were investigated in saturated KCl electrolyte (Fig. S3 − 4). These results confirmed the highly reversible insertion/extraction of K+ owing to the large interplanar distance of PTCDI (Fig. S5).34, 43, 44
Based on materials screened above, the aqueous PTCDI//I2 full cell was assembled with PTCDI as anode, I2 facilely sublimated on active carbon (AC) as cathode (I2@AC, Fig. S6) and saturated KCl solution as electrolyte. Considering the theoretical capacity of I2 (422 mAh g− 1) 3, 9, 28 and PTCDI (137 mAh g− 1)43, a mass ratio of PTCDI:I2 = 3.1 is utilized. The PTCDI//I2 full cell displays a significant advantage in making full use of anode when compared with the conventional metal//I2 battery system. During the charge/discharge process, the oxidation/reduction reaction of I−/I0/I+ occurs at the I2@AC cathode and the enolization/recovery of carbonyl groups occurs at the PTCDI anode with K+ intercalation/de-intercalation (Fig. 2a-b).2, 3, 5 Hence, the chemical equations of the PTCDI//I2 full cell in saturated KCl can be described as below:
Cathode: 2I− ↔ I2 + 2e− I2 + 2Cl− ↔ 2ICl + 2e− (1)
Anode: 2PTCDI + 4K+ + 4e− ↔ 2PTCDI − 2K (2)
With the increase of cycle numbers, the initial redox peaks shifted toward positive direction gradually and finally stabilized due to the activation process (Fig. S7a).26, 45, 46 Consequently, typical CV curves of PTCDI//I2 within the voltage range of 0 − 2.4 V are presented (Fig. 2c, Fig. S7b-d). Oxidation peaks at 1.54 V, 2.24 V and reduction peaks of 1.71 V, 0.91 V, 0.72 V are clearly observed. As for the I2@AC cathode, the redox pair of 2.24 V/1.71 V corresponds to the reversible reaction of I0/I+ while the rest redox pairs correspond to the reversible transformation between I− and I0. 2, 3, 5 As for the PTCDI anode, the anodic and cathodic peaks correspond to the stepwise intercalation and de-intercalation of K+.34, 42, 44 The typical galvanostatic charge/discharge (GCD) curve displays a major plateau at 1.90 V during the discharge process (Fig. 2d). Impressively, the voltage plateau of the PTCDI//I2 full cell is of considerable superiority to that of all reported aqueous rechargeable Zn//I2,3, 5, 6, 18–21, 47 Fe//I216, Al//I2,17 H2//I2,48 and most aqueous rechargeable K+ full battery systems (ARKFBs),38–42 making it a promising prospect for high-energy output (Fig. 2e).
As shown in Fig. 2f, the cycling performance of the PTCDI//I2 full cell was further investigated by GCD mode at 40 A g− 1 and the specific capacity was calculated based on the mass of I2 in I2@AC cathode. Notably, discharge capacity of 364 mAh g− 1 was delivered after initial activation process. Such a high discharge capacity originates from the full use of I−/I0/I+ conversion reactions.2, 3, 5 Remarkably, 154 mAh g− 1 of discharge capacity was obtained after 92000 cycles at 40 A g− 1 (Fig. 2f), far exceeding all reported I2-based aqueous batteries3, 5, 6, 16–21, 47, 49–54 and most ARKFBs38, 40–42 whatever the aspect of the comparison (Fig. 2g). Such an ultra-long lifespan with high capacity at high rate is attributed to the nature of inertness to iodine anions and the layered structure of the PTCDI anode with large interplanar spacing. Apart from that, the highly reversible enolization of quinone (− C = O) to quinone salts (− C−O − M) is propitious to maintain the structural stability of the PTCDI anode during repeated cycles and thus favors the cycling performance of the full cell.34 To figure out the capacity contribution of AC, GCD of the PTCDI//AC full cell was tested and only 38 mAh g− 1 of discharge capacity was delivered, verifying the fact that capacity contribution of AC is not significant at all (Fig. S8).3
As illustrated in Fig. 2h, the rate capability of PTCDI//I2 full cell was investigated under varied current densities from 40 A g− 1 to 160 A g− 1. In detail, discharge capacities of 323, 204, 140, 116 and 104 mAh g− 1 were delivered at 40, 70, 100, 130 and 160 A g− 1, respectively. Remarkably, a discharge capacity of 322 mAh g− 1 was achieved when the current density was shifted from 160 A g− 1 to 40 A g− 1, which is 99.7% of the initial specific capacity, validating the outstanding rate capability of the full cell and in accord with the CV results in Fig. S7. As shown in Fig. 2i, the demonstrated rate capability of the full cell surpasses all reported I2-based cathode batteries3, 5, 6, 16–21, 47, 49–54 and ARKFBs38, 40–42, representing the highest level thus far. As presented in Fig. 2j, the full cell can exhibit an energy density of 434 Wh kg− 1 at a power density of 50420 W kg− 1. Moreover, an energy density of 86 Wh kg− 1 can be maintained when operated at the peak power density of 155072 W kg− 1. Furthermore, the energy density and power density of the full cell far exceed those of all reported aqueous batteries with I2-based cathode6, 14, 16, 20, 21, 27, 28, 55, 56 and most ARKFBs38, 57, possessing a very promising prospect in practical application. Accordingly, the record ultra-long lifespan, remarkable Ragone performances and high rate tolerance of the PTCDI//I2 full cell successfully demonstrate the superiority of the intrinsic inertness of the PTCDI anode to various iodine anionic species and the fast conversion of I−/I0/I+ in the saturated KCl electrolyte, addressing the major challenge of the reported aqueous I2-batteries. The results achieved promote the electrochemical performance of aqueous I2-batteries to a new epoch. Moreover, considering the multi-variety and multi-function of organic materials, the first introduction of PTCDI organic anode into aqueous I2-batteries not only enriches the family of aqueous halogen-batteries with high performance, but also brings about enormous markets for the use of halogen electrode in the near future. Furthermore, the strategy of selecting PTCDI organic material as anode provides a new sight for the construction of high-performance battery systems based on sulfur electrodes.
To identify the underlying redox chemistry of the two-electron transfer of the I2 cathode, ultraviolet-visible (UV-vis) and Raman spectroscopy were performed. Figure 3a displays the typical GCD curve of the PTCDI//I2 full cell at 40 A g− 1 marked with selected voltage points for characterizations. As illustrated in Fig. 3b, a broad absorption peak in the range from 375 nm to 494 nm corresponding to the formation of iodine molecules when charged to 2.0 V, confirming the conversion from iodide to iodine. It is noted that no characteristic peak of I3− (290 nm) was observed, indicating a direct conversion between I− and I0.2, 3, 5 A new strong peak (341 nm) corresponding to the formation of ICl interhalogens was clearly observed when charged to 2.4 V, verifying the conversion of I0 to I+. Meanwhile, the adsorption peak of iodine molecule (375 − 494 nm) vanished when charged to 2.4 V, consolidating the oxidation process from I0 to I+. Correspondingly, the characteristic adsorption peaks almost recovered to the original state during the discharge process, verifying the highly reversible conversion of I−/I0/I+.2, 3, 5 As shown in Fig. 3c, a new Raman signal (208 cm− 1) corresponding to the characteristic band of ICl interhalogens emerged when charged to 2.4 V, and vanished when discharged to 1.25 V, consolidating the reversible I0/I+ conversion and in accord with UV-vis results in Fig. 3b. 2, 3, 5, 58
To better study the energy storage mechanism of the PTCDI anode, X-ray photoelectron spectroscopy (XPS) and FT-IR were performed at different states. As illustrated in Fig. 3d, the strong characteristic peaks of K 2p emerged at the fully charged state while no signal peaks were detected at pristine state, demonstrating the successful intercalation of K+ into PTCDI. Conversely, the intensity of K 2p peaks at the fully discharged state declined sharply and almost vanished when compared with that at the fully charged state, verifying the reversible de-intercalation of K+ from PTCDI.31, 34, 42, 44 Distinct from the pristine state, a new peak (532.6 eV) corresponding to the C − O species at the fully charged state was clearly observed, verifying the conversion of C = O to C − O species (Fig. 3e).31, 32, 34, 39, 42–44, 59 During the subsequent discharge process, the relative intensity of C − O/C = O declined from 1.32 at the fully charged state to 0.12, demonstrating the reversible conversion of C − O to C = O species. Apart from that, the reversible conversion process of the PTCDI anode was also consolidated by FT-IR (Fig. 3f). It is notable that the intensity of the stretching vibration of carbonyl groups (− C = O) appearing at 1666 cm− 1 at the fully charged state is significantly weaker than that at pristine state. Meanwhile, a new and broad peak standing at 1617 cm− 1 can be clearly observed, demonstrating the conversion from carbonyl groups (− C = O) to enolate groups (− C−O).31, 32, 34, 39, 42–44, 59 Upon the subsequent discharge process, almost all the above characteristic peaks recovered to their original positions and intensities, suggesting the highly reversible conversion between carbonyl (− C = O) and enolate (− C−O) groups of the PTCDI anode.
Considering the intrinsic superiority of being inert to various iodine anionic species of the PTCDI anode and superb electrochemical performances of the PTCDI//I2 full cell in saturated KCl electrolyte, an additional saturated I2(s) was introduced into the KCl electrolyte to validate the feasibility of the PTCDI//I2 full cell in the presence of I2 in electrolyte. Herein, distinct from various strategies employed in metal//I2 batteries that try best to confine I2 in cathode to prevent iodine species from entering the electrolyte, whichever in fact cannot completely prevent, the PTCDI//I2 full cell has no need to worry about this knotty problem and even welcome the introduction of I2 into electrolyte. The operational voltage window of the saturated KCl + I2(aq) mixed electrolyte was 2.75 V, slightly broader than that of pure saturated KCl electrolyte which may ascribe to the less amount of free water when saturated I2 is dissolved in the electrolyte (Fig. S9).60 As expected, CV curves of the PTCDI electrode in the mixed electrolyte were identical to that in pure saturated KCl electrolyte, consolidating the intrinsic characteristics of inertness to various iodine anionic species of the PTCDI electrode (Fig. S10a). As a control experiment, CV measurement of the PTCDI electrode was also performed in a pH ≈ 5 dilute HCl electrolyte (the same pH value as the mixed electrolyte) and no redox peaks are observed, indicating that protons have no effect on the charge/discharge of the PTCDI electrode (Fig. S10b).
Figure 4a displays typical CV curve of the PTCDI//I2 full cell in the saturated KCl + I2(aq) mixed electrolyte. Strikingly, a cascade cell composed of two independent electrochemical processes (marked as step 1 and step 2 separately) is clearly observed. Compared with conventional batteries including only one electrochemical reaction, this cascade battery presents many advantages: a, integrating two full reactions internally and avoiding the inactive additional components needed for external connection. b, higher utilization of the reaction chamber. c, higher energy output.29 The CV curve in step 2 is in consistent with that of the PTCDI//I2 full cell in pure saturated KCl electrolyte. Apart from that, a pair of additional redox peaks in step 1 is attributed to the I2//I2 symmetric cell self-constructed in the mixed electrolyte and is discussed as follows. When using a blank graphite paper (GP, current collector in the above-discussed PTCDI//I2 batteries) as working electrode tested in the mixed electrolyte, two pairs of redox peaks corresponding to the conversion of I−/I0 and I0/I+ are apparent, which means the existence of the possibility of I2//I2 symmetric cell self-constructed in the mixed electrolyte (Fig. S11).2, 3, 5 Then, two pairs of redox peaks centered at different potentials are clearly separated (Fig. 4b). Therefore, a pair of pronounced redox peaks within the voltage range of 0 − 0.5 V is clearly detected when using blank GP as both electrodes in the full cell, demonstrating that the I2//I2 symmetric cell self-constructs successfully in the mixed electrolyte (Fig. 4c). Moreover, the voltage range of the I2//I2 symmetric cell is consistent with that of step 1 in Fig. 4a, further confirming that the redox peaks in step 1 are assigned to the I2//I2 symmetric cell. Besides, no characteristic redox peaks of the blank GP//GP full cell were detected whether in pure saturated KCl electrolyte (Fig. S12a) or in pure saturated I2(aq) electrolyte (Fig. S12b), verifying that the I2//I2 symmetric cell self-constructed only in the mixed electrolyte. Fig. S12b also substantiates the importance of Cl− in electrolyte and in accord with the theoretical calculation results in Fig. 1b.
As an indispensable part of the PTCDI//I2 cascade cell system, electrochemical performances of the I2//I2 symmetric cell (step 1) were systematically evaluated by testing the GP//GP full cell in the mixed electrolyte (Fig. S13). During the charge process, I2 molecules absorbed at the interface between the anode and electrolyte were oxidized to I+ while absorbed at the interface between the cathode and electrolyte were reduced to I− simultaneously (Fig. S13a).2, 3, 5 Reversibly, all the above oxidized and reduced states of iodine recovered to their original state individually during the discharge process (Fig. S13b). A charge plateau at 0.38 V and a discharge plateau at 0.19 V were clearly observed in the typical GCD curve of the I2//I2 symmetric cell (Fig. S13c), in accord with the CV results in Fig. 4c. A discharge areal capacity of 0.06 mAh cm− 2 can be obtained at 0.2 mA cm− 2 after 9930 cycles, exhibiting the excellent stability of the I2//I2 symmetric cell. Distinct from conventional batteries, the capacity exhibits a trend of continuous rise during cycling due to the increasing activated I2 molecules (Fig. S13d). Moreover, discharge areal capacities of 0.03, 0.016, 0.008 and 0.006 mAh cm− 2 were achieved at 2, 5, 10 and 15 mA cm− 2, respectively, indicating good rate capability (Fig. S13e). A new peak centered at 343 nm corresponding to ICl interhalogens emerged at the fully charged state and almost vanished at the fully discharged state, verifying the reversible conversion of I0/I+ (Fig. S13f).2, 3, 5 In view of a large amount of I2 contained in the mixed electrolyte, the intensity change of the broad absorption peak (375 − 494 nm) corresponding to I2 molecules was not apparent at different states.2, 3, 5
Therefore, the first charge step (marked as step 1) of the cascade cell based on the PTCDI//I2 in the saturated KCl + I2(aq) mixed electrolyte corresponds to the conversion of the I2//I2 symmetric cell within the voltage range of 0 − 0.5 V (Fig. 4e). During the second charge step (marked as step 2) within the voltage range of 0.5 − 2.5 V, the conversion of I−/I0/I+ occurred at the cathode and the intercalation of K+ occurred at the PTCDI anode (Fig. 4f).2, 3, 5 Overall, chemical reactions of the cascade cell can be expressed as follows:
Step 1: cathode: I2 + 2Cl− ↔ 2ICl + 2e− (3)
anode: I2 + 2e− ↔ 2I− (4)
Step 2: cathode: 2I− ↔ I2 + 2e− I2 + 2Cl− ↔ 2ICl + 2e− (5)
anode: 2PTCDI + 4K+ + 4e− ↔ 2PTCDI − 2K (6)
As presented in Fig. 4g, a major discharge plateau at 0.34 V corresponding to the I2//I2 symmetric cell was clearly observed in step 1, and the GCD curve in step 2 was identical to that of the PTCDI//I2 full cell in saturated KCl electrolyte in Fig. 2d. Moreover, these results are in accord with CV results in Fig. 4a, and consolidate the successful construction of the PTCDI//I2 cascade cell in the mixed electrolyte. A high discharge capacity of 623 mAh g− 1 was delivered at 40 A g− 1 after 6000 cycles (Fig. 4h). Impressively, the PTCDI//I2 cascade cell even achieved 105000 cycles under a higher current density of 60 A g− 1, exhibiting an outstanding cyclical stability (Fig. S14). The morphology and particle size of the PTCDI anode after cycling at 40 A g− 1 for 9200 cycles in the mixed electrolyte are almost identical to those in pristine state, consolidating the intrinsic superiority of being immune to various iodine anionic species and excellent structural stability of PTCDI (Fig. S15). As illustrated in Fig. 4i, discharge capacities of 574, 500, 376, 285 and 264 mAh g− 1 were delivered at 40, 70, 100, 130 and 160 A g− 1, respectively. Notably, a discharge capacity of 572 mAh g− 1 was delivered when the current density was shifted from 160 to 40 A g− 1, the value of which is 99.7% of that at initial 40 A g− 1, demonstrating the superb rate capability of the PTCDI//I2 cascade cell. Compared with well-known strategies (e.g., water-in-salt, liquid polymer additive)61–80 to broaden the voltage window in aqueous electrolyte, the high voltage of the PTCDI//I2 cascade cell exhibited great superiority by just utilizing the saturated KCl + I2(aq) mixed electrolyte without highly concentrated salts or a large amount of polymers (Fig. 4j). Consequently, this is a promising strategy to obtain a low-cost, eco-friendly and high-voltage aqueous battery. Therefore, distinct from the reported I2-cathode single battery, the successful construction of PTCDI//I2 cascade cell in mixed electrolyte for the first time demonstrates the feasibility of cascade cell in I2-batteries, making full use of the iodine leaked from the I2 cathode and outputting more energy as I2//I2 symmetric cell in electrolyte. Moreover, the concept of PTCDI//I2 cascade cell not only enriches the family of aqueous halogen-batteries with high performance, but also paves a new way for the construction of high-performance battery systems based on sulfur electrodes.
Inspired by the results that the conversion of I−/I0/I+ can be completed by merely using blank GP as electrode in the mixed electrolyte (Fig. S11), an aqueous cathode-free cascade cell of PTCDI//GP was successfully constructed in the same electrolyte (Fig. S16). Owing to the same reaction mechanism, reaction processes of the PTCDI//GP cathode-free cascade cell are identical to those of the PTCDI//I2 cascade cell (Fig. S16a − f). Similarly, the PTCDI//GP cathode-free cascade cell demonstrated stable cycling performance and good rate capability (Fig. S16g − i). Furthermore, a large area soft package battery of the PTCDI//GP cathode-free cascade cell with 6 cm ⋅ 8 cm was successfully assembled (Fig. 4k). The capacity retention reached 70% at 80 mA after cycling for 900 cycles, exhibiting good cyclical stability (Fig. 4l). A power density of 52 W m− 2 was obtained when an ultra-high current of 320 mA was applied, indicating the high rate tolerance of the cascade cell. Besides, a power density of 16 W m− 2 was achieved when the current was shifted from 320 to 80 mA, the value of which is almost identical to that at initial 80 mA, exhibiting good reversibility (Fig. 4m).