Influence of Polymorphism on the Electrochemical Behavior of Dilithium (2,3-Dilithium-oxy)-terephthalate vs. Li

Organic electrode materials offer obvious opportunities to promote cost-effective and environmentally friendly rechargeable batteries. Over the last decade, tremendous progress has been made thanks to the use of molecular engineering focused on the tailoring of redox-active organic moieties. However, the electrochemical performance of organic host structures relies also on the crystal packing, like the inorganic counterparts, which calls for further efforts in terms of crystal chemistry to make a robust redox-active organic center electrochemically efficient in the solid state. Following our ongoing research aiming at elaborating lithiated organic cathode materials, we report herein on the impact of polymorphism on the electrochemical behavior of dilithium (2,3-dilithiumoxy-)terephthalate vs. Li. Having isolated dilithium (3-hydroxy-2-lithium-oxy)terephthalate through an incomplete acid-base neutralization reaction, its subsequent thermally induced decarboxylation mechanism led to the formation of a new polymorph of dilithium (2,3-dilithium-oxy-)terephthalate referred to as Li4-o-DHT (β-phase). This new phase is able to operate at 3.1 V vs. Li+/Li, which corresponds to a positive potential shift of +250 mV compared to the other polymorph formerly reported. Nevertheless, the overall electrochemical process characterized by a sluggish biphasic transition is impeded by a large polarization value limiting the recovered capacity upon cycling.


Introduction
Lithium-Ion Batteries (LIBs) are well-known to dominate the field of electrochemical storage technologies from portable-electronic market to electric vehicles. They obviously meet the requirements in terms of energy density, cyclability, and safety, but they also bring increasing challenges regarding scarcity of resources [1][2][3] and recyclability [4,5] due to the emerging large-scale application markets as highlighted by our group for several years [6]. Among the various alternative chemistries, the use of electrodes made of electroactive organic compounds could limit the demand on metal-based raw materials as well as mitigate the environmental burden of current LIBs [7,8]. Indeed, interesting assets like the abundancy of their constitutive atoms all over the world, their mild synthesis conditions or simpler recycling solutions can be put forward in their favor. In the past fifty years many promising organic electrode materials have been identified for possible applications in various cell configurations giving rise to an abundant and very interesting literature on the topic (see for example Refs. [7,[9][10][11][12][13][14][15][16]). Note that pioneering works were reported in this field as early as 1997 by M. Armand and co-workers by demonstrating the possible Li insertion in oxocarbons to reach specific capacities higher than 500 mAh g −1 [17], paving the way for further studies on lithiated discrete molecules.
For the last decade, our group has been focusing on lithiated positive organic electrodes especially by working on the robust and efficient dihydroxyterephthalate redoxactive backbone (Figure 1). We first reported dilithium (2,5-dilithium-oxy)-terephthalate (Li 4 -p-DHT or (Li 2 )(Li 2 )-p-DHT) as lamellar host electrode material [18], which displays an interesting stability upon cycling at an average potential of 2.55 V vs. Li + /Li although limited to half of the theoretical capacity in our experimental conditions. Its full capacity (i.e., 2-electron reaction; 223 mAh g −1 at 0.1 C) was later demonstrated by Chen group when prepared as graphene-supported nanosheets [19,20]. The study of its ortho regioisomer, namely, dilithium (2,3-dilithium-oxy)-terephthalate (Li 4 -o-DHT or (Li 2 )(Li 2 )-o-DHT), enabled a potential gain of +300 mV reaching 2.85 V vs. Li + /Li [21] thanks to the mitigation of stereoelectronic effects at the molecular level [22]. More recently, keeping the lamellar arrangement in the solid state, the cation substitution on the carboxylate functional groups was also found to be highly effective to tune the redox potential of the lihium phenolate redox-active moiety [23,24]. Thus, the reversible electrochemical lithium extraction occurs at 3.45 V vs. Li + /Li in magnesium (2,5-dilithium-oxy)-terephthalate (Mg(Li 2 )-p-DHT) due to the high ionic potential value of Mg 2+ ion that influences the Li-O bonds causing changes in the electronic distribution in the ligand (stabilization of the distribution of p-electron density) [23].
possible Li insertion in oxocarbons to reach specific capacities higher than 500 mAh [17], paving the way for further studies on lithiated discrete molecules.
For the last decade, our group has been focusing on lithiated positive organic ele trodes especially by working on the robust and efficient dihydroxyterephthalate redo active backbone (Figure 1). We first reported dilithium (2,5-dilithium-oxy)-terephthala (Li4-p-DHT or (Li2)(Li2)-p-DHT) as lamellar host electrode material [18], which displays interesting stability upon cycling at an average potential of 2.55 V vs. Li + /Li although li ited to half of the theoretical capacity in our experimental conditions. Its full capacity (i 2-electron reaction; 223 mAh g −1 at 0.1 C) was later demonstrated by Chen group wh prepared as graphene-supported nanosheets [19,20]. The study of its ortho regioisom namely, dilithium (2,3-dilithium-oxy)-terephthalate (Li4-o-DHT or (Li2)(Li2)-o-DHT), en bled a potential gain of +300 mV reaching 2.85 V vs. Li + /Li [21] thanks to the mitigation stereoelectronic effects at the molecular level [22]. More recently, keeping the lamellar rangement in the solid state, the cation substitution on the carboxylate functional grou was also found to be highly effective to tune the redox potential of the lihium phenola redox-active moiety [23,24]. Thus, the reversible electrochemical lithium extraction occu at 3.45 V vs. Li + /Li in magnesium (2,5-dilithium-oxy)-terephthalate (Mg(Li2)-p-DHT) d to the high ionic potential value of Mg 2+ ion that influences the Li-O bonds causi changes in the electronic distribution in the ligand (stabilization of the distribution of electron density) [23]. Recently, complementary investigations performed on the ortho-regioisomer bac bone led us to synthesize an intermediate compound, namely, dilithium (3-hydroxy lithium-oxy)terephthalate or (Li2)(Li,H)-o-DHT ( Figure 1). Unexpectedly, after subseque thermal treatment in neutral atmosphere, a polymorph of the previously describ (Li2)(Li2)-o-DHT (or Li4-o-DHT) compound [21] was recovered by intermolecular therm rearrangement (retro-Kolbe-Schmitt-type reaction). In this context, we report herein th complementary study focused on the ortho-regioisomer backbone by explaining the k Recently, complementary investigations performed on the ortho-regioisomer backbone led us to synthesize an intermediate compound, namely, dilithium (3-hydroxy-2-lithiumoxy)terephthalate or (Li 2 )(Li,H)-o-DHT ( Figure 1). Unexpectedly, after subsequent thermal treatment in neutral atmosphere, a polymorph of the previously described (Li 2 )(Li 2 )-o-DHT (or Li 4 -o-DHT) compound [21] was recovered by intermolecular thermal rearrangement (retro-Kolbe-Schmitt-type reaction). In this context, we report herein this complementary study focused on the ortho-regioisomer backbone by explaining the key chemical reactions leading to the synthesis of (Li 2 )(Li,H)-o-DHT and (Li 2 )(Li 2 )-o-DHT (new polymorph), respectively. The electrochemical behavior measured in Li half-cell for the latter are also described and compared to data reported earlier [21].

Preliminary Observations
In a former study [21], we reported, for the first time, the synthesis of the (Li 2 )(Li 2 )-o-DHT material performed in solventless conditions thanks to the occurrence of the retro-Kolbe-Schmitt reaction by controlled heating under inert atmosphere of the dilithium dicarboxylate (Li 2 )(H 2 )-o-DHT precursor. The latter was obtained by acid-base reaction when adding lithium carbonate in an aqueous suspension of 2,3-dihydroxyterephthalic acid (H 4 -o-DHT). To complete this pioneer study, we recently tried to synthesize (Li 2 )(Li 2 )-o-DHT by direct four-proton neutralization in methanol (MeOH)/lithium methoxide (MeOLi) medium as previously reported by Chen group for the preparation of the para-regioisomer, namely, (Li 2 )(Li 2 )-p-DHT [20]. Note that such a chemical approach was also successfully applied by our group to synthesize the tetralithium salt of tetrahydroxybenzoquinone (THQ, 2,3,5,6-tetrahydroxy-1,4-benzoquinone) in 2009 [25]. In practice, 4.4 equivalents of MeOLi in methanol were added dropwise to a suspension of H 4 -o-DHT and subsequently stirred overnight; the solid phase was then recovered by filtration. However, several analyses performed on the as-obtained powder led to the conclusion that an incomplete lithiation reaction systematically occurred by producing the (Li 2 )(Li,H)-o-DHT) compound (Scheme 1), which is an intermediate between the (Li 2 )(H 2 )-o-DHT and (Li 2 )(Li 2 )-o-DHT. chemical reactions leading to the synthesis of (Li2)(Li,H)-o-DHT and (Li2)(Li2)-o-DHT (new polymorph), respectively. The electrochemical behavior measured in Li half-cell for the latter are also described and compared to data reported earlier [21].

Preliminary Observations
In a former study [21], we reported, for the first time, the synthesis of the (Li2)(Li2)-o-DHT material performed in solventless conditions thanks to the occurrence of the retro-Kolbe-Schmitt reaction by controlled heating under inert atmosphere of the dilithium dicarboxylate (Li2)(H2)-o-DHT precursor. The latter was obtained by acid-base reaction when adding lithium carbonate in an aqueous suspension of 2,3-dihydroxyterephthalic acid (H4-o-DHT). To complete this pioneer study, we recently tried to synthesize (Li2)(Li2)o-DHT by direct four-proton neutralization in methanol (MeOH)/lithium methoxide (Me-OLi) medium as previously reported by Chen group for the preparation of the para-regioisomer, namely, (Li2)(Li2)-p-DHT [20]. Note that such a chemical approach was also successfully applied by our group to synthesize the tetralithium salt of tetrahydroxybenzoquinone (THQ, 2,3,5,6-tetrahydroxy-1,4-benzoquinone) in 2009 [25]. In practice, 4.4 equivalents of MeOLi in methanol were added dropwise to a suspension of H4-o-DHT and subsequently stirred overnight; the solid phase was then recovered by filtration. However, several analyses performed on the as-obtained powder led to the conclusion that an incomplete lithiation reaction systematically occurred by producing the ( Although methoxide ion is a strong base, this partial neutralization of the catechol moiety can be explained by the existence of an intramolecular hydrogen bond once the first proton is removed (Figure 1), which does not exist in the para-regioisomer. Therefore, the remaining proton is better stabilized. This induces an increase of the pKa2 value relative to the second OH group as commonly reported in the literature [26,27]. Consequently, the use of a stronger base is being required to fully neutralize the o-DHT ligand. However, this (Li2)(Li,H)-o-DHT intermediate compound having not being reported to date, we decided to go further to complete our knowledge on this materials family.

Synthesis, Characterizations and Thermal Behavior of (Li2)(Li,H)-o-DHT
Taking into account our preliminary observations, the new series of experiments consisted in using 3 equivalents of MeOLi only to produce the titled compound. A white-grey powder was recovered after reaction (16 h, 30°C in Ar filled glove box), subsequently washed, and characterized by using several characterization techniques. The presence of unaltered o-DHT ligand in the compound was readily confirmed by both liquid 1 H and 13 C NMR measurements ( Figure S1) after a derivatization reaction based on a full reprotonation step using H2SO4 as performed in former studies [18,21]. The as-produced organic salt was also characterized by thermal analyses (TG-DSC-MS) and by Fourier-transform infrared spectroscopy (FT-IR). Thermal analyses performed under argon ( Figure 2) Although methoxide ion is a strong base, this partial neutralization of the catechol moiety can be explained by the existence of an intramolecular hydrogen bond once the first proton is removed (Figure 1), which does not exist in the para-regioisomer. Therefore, the remaining proton is better stabilized. This induces an increase of the pK a2 value relative to the second OH group as commonly reported in the literature [26,27]. Consequently, the use of a stronger base is being required to fully neutralize the o-DHT ligand. However, this (Li 2 )(Li,H)-o-DHT intermediate compound having not being reported to date, we decided to go further to complete our knowledge on this materials family.

Synthesis, Characterizations and Thermal Behavior of (Li 2 )(Li,H)-o-DHT
Taking into account our preliminary observations, the new series of experiments consisted in using 3 equivalents of MeOLi only to produce the titled compound. A whitegrey powder was recovered after reaction (16 h, 30 • C in Ar filled glove box), subsequently washed, and characterized by using several characterization techniques. The presence of unaltered o-DHT ligand in the compound was readily confirmed by both liquid 1 H and 13 C NMR measurements ( Figure S1) after a derivatization reaction based on a full reprotonation step using H 2 SO 4 as performed in former studies [18,21]. The as-produced organic salt was also characterized by thermal analyses (TG-DSC-MS) and by Fourier-transform infrared spectroscopy (FT-IR). Thermal analyses performed under argon ( Figure 2) indicated that the compound is methanol-free since the first weight loss (∆m = 23%) occurring beyond 300 • C induced a CO 2 evolution (i.e., a thermal degradation of the pristine compound and not a desolvation process). The FT-IR spectrum ( Figure S2) confirmed well this hypothesis because the presence of MeOH molecules in the crystal structure should notably induce a broad O-H vibration band at 3300-3200 cm −1 [20], which was not visible in the present case. However, a weak vibration band located at ≈3700 cm −1 was detected, which can be assigned to the remaining OH group in the ligand. Finally, the proposed C 8 H 3 Li 3 O 6 chemical formula was confirmed by multi-elemental analyses (C, H, Li), as well as by thermal analysis performed under pure O 2 ( Figures S3 and S4).
Inorganics 2022, 10, x FOR PEER REVIEW indicated that the compound is methanol-free since the first weight loss (Δm curring beyond 300 °C induced a CO2 evolution (i.e., a thermal degradation of t compound and not a desolvation process). The FT-IR spectrum ( Figure S2) conf this hypothesis because the presence of MeOH molecules in the crystal structu notably induce a broad O-H vibration band at 3300-3200 cm −1 [20], which was in the present case. However, a weak vibration band located at ≈3700 cm −1 wa which can be assigned to the remaining OH group in the ligand. Finally, the C8H3Li3O6 chemical formula was confirmed by multi-elemental analyses (C, H, as by thermal analysis performed under pure O2 ( Figures S3 and S4). We decided to further investigate the thermal behavior of (Li2)(Li,H)-o-DH first weight loss observed by TG measurement starting at ≈300 °C is associated dothermic phenomenon, which is not expected in the case of an irreversible th radation. In fact, we previously experienced that such an endothermic pheno companied by CO2 release could be ascribed to the formation of a new phase by thermal rearrangement according to the retro-Kolbe-Schmitt reaction [18,21]; th being known for a long time for preparing alkali salts of β-hydroxy acid (BHA) [ fore, the thermal properties upon heating of (Li2)(Li,H)-o-DHT were studied by ture-resolved X-ray powder diffraction (TRXRPD) collected from 25 to 500 °C u 0.2 °C.s −1 (Figure 3). First, the recorded data show that the starting material is c We decided to further investigate the thermal behavior of (Li 2 )(Li,H)-o-DHT since the first weight loss observed by TG measurement starting at ≈300 • C is associated to an endothermic phenomenon, which is not expected in the case of an irreversible thermal degradation. In fact, we previously experienced that such an endothermic phenomenon accompanied by CO 2 release could be ascribed to the formation of a new phase by concerted thermal rearrangement according to the retro-Kolbe-Schmitt reaction [18,21]; this reaction being known for a long time for preparing alkali salts of β-hydroxy acid (BHA) [28]. Therefore, the thermal properties upon heating of (Li 2 )(Li,H)-o-DHT were studied by temperature-resolved X-ray powder diffraction (TRXRPD) collected from 25 to 500 • C under N 2 at 0.2 • C.s −1 (Figure 3). First, the recorded data show that the starting material is crystallized and stable until 265-270 • C in agreement with the thermal analysis data. Beyond this temperature, (Li 2 )(Li,H)-o-DHT undergoes a phase transition with the appearance of new Bragg peaks especially within the 20-30 • 2θ-range while typical diffraction peaks of the pristine compound vanish. A further heating leads to a complete decomposition process with the progressive formation of crystallized Li 2 CO 3 with amorphous carbon as expected in such pyrolysis experimental conditions. Inorganics 2022, 10, x FOR PEER REVIEW 5 of 16 With the aim to isolate and characterize this new phase as well as the possible occurrence of a retro-Kolbe-Schmitt thermal reaction, powder of (Li2)(Li,H)-o-DHT was heated under controlled pyrolysis conditions and further studied.

Identification of the New Phase and Proposed Thermal Rearrangement Mechanism
To isolate this new phase in sufficient quantity, 600 mg of the (Li2)(Li,H)-o-DHT salt was gently annealed at 290 °C during 20 h in a glass oven (Büchi B-585 glassoven Kugelrohr-drying model) placed inside a glove box. As reported earlier [18,21], a thermal gradient can be readily obtained within this glass oven making the separation of the volatile organic compounds (potentially formed) by a simple evaporation/crystallization process possible. Preliminary observations with the naked eyes showed (i) the sample turned from a light gray color into beige upon heating, and (ii) that white and well-shaped small crystals were collected inside the condensation compartment of the glass oven. The latter were readily identified as pure crystallized catechol by single-crystal XRD analysis (monoclinic P21/c crystal structure [28]).
The recovered annealed sample, which had lost 23% in weight, was then characterized by several techniques. While SEM images of the pristine particles of (Li2)(Li,H)-o-DHT revealed agglomerations of submicrometer-sized platelets which were a few tenths of a nanometer thick (Figure 4a), the annealed particles exhibited numerous holes and voids (Figure 4b,c) in good agreement with the departure of both CO2 and catechol molecules. Interestingly, liquid 1 H and 13 C NMR measurements after derivatization reaction confirmed again the presence of unaltered o-DHT ligand in this new phase ( Figure S5). With the aim to isolate and characterize this new phase as well as the possible occurrence of a retro-Kolbe-Schmitt thermal reaction, powder of (Li 2 )(Li,H)-o-DHT was heated under controlled pyrolysis conditions and further studied.

Identification of the New Phase and Proposed Thermal Rearrangement Mechanism
To isolate this new phase in sufficient quantity, 600 mg of the (Li 2 )(Li,H)-o-DHT salt was gently annealed at 290 • C during 20 h in a glass oven (Büchi B-585 glassoven Kugelrohrdrying model) placed inside a glove box. As reported earlier [18,21], a thermal gradient can be readily obtained within this glass oven making the separation of the volatile organic compounds (potentially formed) by a simple evaporation/crystallization process possible. Preliminary observations with the naked eyes showed (i) the sample turned from a light gray color into beige upon heating, and (ii) that white and well-shaped small crystals were collected inside the condensation compartment of the glass oven. The latter were readily identified as pure crystallized catechol by single-crystal XRD analysis (monoclinic P2 1 /c crystal structure [28]).
The recovered annealed sample, which had lost 23% in weight, was then characterized by several techniques. While SEM images of the pristine particles of (Li 2 )(Li,H)-o-DHT revealed agglomerations of submicrometer-sized platelets which were a few tenths of a nanometer thick (Figure 4a  Powder XRD (PXRD) pattern analysis of the recovered sample ( Figure 5a) confirm that the new phase observed by TRXRPD beyond 290 °C (Figure 3) was successfully i lated thanks to this controlled thermal treatment. Beyond the color difference not above, this collected PXRD pattern did not fit with that previously reported for Li4-o-DH [21] ( Figure S7) suggesting that a new polymorph of this tetra-lithiated salt was obtain through this synthesis pathway, hereafter referred to as Li4-o-DHT (β-phase). Note th the formerly described form obtained from thermal rearrangement of (Li2)(H2)-o-DHT now referred as Li4-o-DHT (α-phase). The comparison of the FT-IR spectra for these tw phases showed many similarities as expected with polymorphs ( Figure S8) by displayi the same typical vibrations bands at similar wavenumber values (νas(COO − ), ν(C= νs(COO − ) or ν(C-OLi)), although additional bands were noticed too. In addition, the we defined PXRD pattern showed that the β-phase exhibits a very good crystallinity co pared to the α-phase making further crystallographic analyses possible. After pattern dexing in a monoclinic lattice and space group determination (P21/m), several attemp were made to solve the crystal structure of the β-phase but we were only able to locate t o-DHT unit in the lattice (Figure 5b) and not the four lithium atoms. This is due to t difficulty of finding a coherent oxygen environment around the lithium atoms, which c adopt a wide variety of coordination polyhedra, from highly distorted tetrahedra to re ular octahedra [29]. In addition, the presence of a possible preferred orientation in t collected PXRD pattern, suggested by the plate-like crystallites observed by SEM (Figu 4b,c), makes it difficult to fully resolve the structure. Nevertheless, refinement of the c parameters of the PXRD pattern was successfully achieved in the P21/m space group w a = 6.8039 (6)   Powder XRD (PXRD) pattern analysis of the recovered sample (Figure 5a) confirmed that the new phase observed by TRXRPD beyond 290 • C (Figure 3) was successfully isolated thanks to this controlled thermal treatment. Beyond the color difference noted above, this collected PXRD pattern did not fit with that previously reported for Li 4 -o-DHT [21] (Figure S7) suggesting that a new polymorph of this tetra-lithiated salt was obtained through this synthesis pathway, hereafter referred to as Li 4 -o-DHT (β-phase). Note that the formerly described form obtained from thermal rearrangement of (Li 2 )(H 2 )o-DHT is now referred as Li 4 -o-DHT (α-phase). The comparison of the FT-IR spectra for these two phases showed many similarities as expected with polymorphs ( Figure S8) by displaying the same typical vibrations bands at similar wavenumber values (ν as (COO − ), ν(C=C), ν s (COO − ) or ν(C-OLi)), although additional bands were noticed too. In addition, the well-defined PXRD pattern showed that the β-phase exhibits a very good crystallinity compared to the α-phase making further crystallographic analyses possible. After pattern indexing in a monoclinic lattice and space group determination (P2 1 /m), several attempts were made to solve the crystal structure of the β-phase but we were only able to locate the o-DHT unit in the lattice (Figure 5b) and not the four lithium atoms. This is due to the difficulty of finding a coherent oxygen environment around the lithium atoms, which can adopt a wide variety of coordination polyhedra, from highly distorted tetrahedra to regular octahedra [29]. In addition, the presence of a possible preferred orientation in the collected PXRD pattern, suggested by the plate-like crystallites observed by SEM (Figure 4b,c), makes it difficult to fully resolve the structure. Nevertheless, refinement of the cell parameters of the PXRD pattern was successfully achieved in the P2 1 /m space group with a = 6.8039 (6)  Finally, the CO2 evolution observed by TG-DSC-MS (Figure 2) was then quan in a new series of dedicated experiments using micro gas chromatography. In prac few mg of (Li2)(Li,H)-o-DHT was placed in a glass ampoule, under argon atmosphere heated at 300 °C at 3.5 °C.min −1 , which is a similar heating rate compared to both th analyses and TRXRPD measurements making the control of the pyrolysis reaction p ble. The weight loss of the sample due to CO2 release in the gas phase was measured 10.9 ± 1.0%. Note that the use of much higher heating rates favors the thermal degrad reaction pathways leading to higher CO2 releases ( Figure S9).
By compiling all these experimental results, a peculiar retro-Kolbe-Schmitt rea seems to occur by controlled pyrolysis of (Li2)(Li,H)-o-DHT to produce a new polym of Li4-o-DHT. Scheme 2 summarizes the proposed thermal concerted rearrangement tion that involves 4 equivalents of (Li2)(Li,H)-o-DHT). Finally, the CO 2 evolution observed by TG-DSC-MS (Figure 2) was then quantified in a new series of dedicated experiments using micro gas chromatography. In practice, a few mg of (Li 2 )(Li,H)-o-DHT was placed in a glass ampoule, under argon atmosphere, and heated at 300 • C at 3.5 • C.min −1 , which is a similar heating rate compared to both thermal analyses and TRXRPD measurements making the control of the pyrolysis reaction possible. The weight loss of the sample due to CO 2 release in the gas phase was measured to be 10.9 ± 1.0%. Note that the use of much higher heating rates favors the thermal degradation reaction pathways leading to higher CO 2 releases ( Figure S9).
By compiling all these experimental results, a peculiar retro-Kolbe-Schmitt reaction seems to occur by controlled pyrolysis of (Li 2 )(Li,H)-o-DHT to produce a new polymorph of Li 4 -o-DHT. Scheme 2 summarizes the proposed thermal concerted rearrangement reaction that involves 4 equivalents of (Li 2 )(Li,H)-o-DHT).
The next study was then focused on the electrochemical storage properties of the β-phase vs. Li to compare its performance towards the α-phase previously reported [21].

Electrochemical Behavior of Li 4 -o-DHT (β-Phase) as Active Electrode Material vs. Li
The chemical stability of the β-phase in air was first assessed by FT-IR spectroscopy thanks to a series of spectra recorded at different exposure times ( Figure S10). While the α-phase oxidizes rapidly in air (solid-state autoxidation reaction [21]), the β-form appears stable with no obvious color change. The main modification of the FT-IR spectra deals with the moisture uptake as shown by the rapid O-H stretching band appearance located in the 3000-3500 cm −1 region. This stability in contact with dioxygen implies that the new polymorph could electrochemically desinsert/insert Li ions in the solid state at more than 3 V vs. Li + /Li [18,21]; it should be recalled that an average operating potential limited to 2.85 V vs. Li + /Li was recorded for the α-phase. reaction pathways leading to higher CO2 releases ( Figure S9).
By compiling all these experimental results, a peculiar retro-Kolbe-Schmitt reaction seems to occur by controlled pyrolysis of (Li2)(Li,H)-o-DHT to produce a new polymorph of Li4-o-DHT. Scheme 2 summarizes the proposed thermal concerted rearrangement reaction that involves 4 equivalents of (Li2)(Li,H)-o-DHT).

Scheme 2. Proposed thermal concerted retro-Kolbe-Schmitt reaction involving (Li2)(Li,H)-o-DHT
by controlled pyrolysis reaction to produce Li4-o-DHT (β-phase). The by-products, catechol and carbon dioxide, are released in the gas phase. Note that heterolytic cleavage of bonds have been considered for the proposed mechanism without direct evidence.

Scheme 2. Proposed thermal concerted retro-Kolbe-Schmitt reaction involving (Li 2 )(Li,H)-o-DHT by controlled pyrolysis reaction to produce Li 4 -o-DHT (β-phase)
. The by-products, catechol and carbon dioxide, are released in the gas phase. Note that heterolytic cleavage of bonds have been considered for the proposed mechanism without direct evidence.
The electrochemical storage properties of the new polymorph as active electrode material were performed vs. Li in Swagelok ® -type cells according to a similar procedure to that reported for the α-phase testing [21] to make comparisons easier. After a preliminary screening, typical electrochemical measurements consisted of galvanostatic cycling tests within the 3.4−2.8 V vs. Li + /Li potential range recorded at a cycling rate of one e − /Li + exchanged per Li 4 -o-DHT unit in 10 h (denoted 1 Li + /10 h).
The most relevant electrochemical features related to Li 4 -o-DHT (β-phase) are summarized in Figure 6 and compared to former data recorded for the α-phase [21]. Figure 6a shows the typical potential-composition (E − x) trace restricted to the first five cycles. Note that the use of 1 M LiClO 4 in propylene carbonate (PC) as the electrolyte was preferred in this study since slightly better cycling performance was obtained compared to the typical 1 M LiPF 6 in ethylene carbonate (EC)/dimethyl carbonate (DMC) (1:1 vol./vol.) formulation ( Figure S11). Unexpectedly, this new polymorph exhibits a reversible electrochemical deinsertion/insertion of Li + mostly characterized by a flat voltage plateau (biphasic process) like olivine LiFePO 4 or spinel Li 4 Ti 5 O 12 and not by a succession of plateaus/solid-solution domains as observed with the α-phase. This two-phase transition occurs at an average working potential of 3.1 V vs. Li + /Li (compared to 2.85 V vs. Li + /Li for the α-phase) in agreement with the stability of the β-phase in air underlined above. However, the overall electrochemical process is impeded by a large charge/discharge potential difference (∆E ≈ 120 mV). This polarization value, coupled with the appearance of overpotential activation peaks [30,31] at the beginning of the plateau both in charge and in discharge (Figure 6a, arrows), suggests a sluggish biphasic transition possibly rooted in the existence of high nucleation barrier of the biphasic reaction. This can explain the rapid capacity decrease upon the ten first cycles cycling (35% loss) prior to reaching a stabilized value of ≈75 mAh.g −1 over several dozen cycles (Figure 6b). The evolution of the apparent cell resistance (R app. ) was also monitored upon cycling but no significant change was noticed in agreement with the as-observed stable capacity retention ( Figure S12). A cycling test at 60 • C was also performed with the aim to improve the two-phase transition kinetics.
Results indicated just a slight decrease of the polarization value (∆E ≈ 100 mV) but accompanied by a faster capacity decay and a poor coulombic efficiency probably due to the occurrence of side reactions with the electrolyte at such a temperature ( Figure S13). The electrochemical discrepancies between the two polymorphs can be more acc rately visualized on Figure 6c,d, especially the positive potential shift for the β-phase The electrochemical discrepancies between the two polymorphs can be more accurately visualized on Figure 6c,d, especially the positive potential shift for the β-phase estimated to +250 mV. Note that the sole reversible and sharp peak observed on the potential vs. differential capacity curve (Figure 6d) is the typical expected trace for a biphasic process. Galvanostatic intermittent titration technique (GITT) was also exploited to get further insight about the as-observed polarization phenomena. The analysis of chronopotentiograms recorded during the open circuit potential (OCP) periods revealed that relatively long relaxation times were necessary to be closer to thermodynamic equilibrium confirming the sluggish biphasic transition. Figure 6e shows the E − x trace measured in GITT mode with an OCP period of 48 h. Basically, the symmetric profile of the GITT curve indicated similar kinetic limitations both in charge and in discharge. Therefore, the equilibrium redox potential value of 3.1 V vs. Li + /Li was verified. The biphasic electrochemical deinsertion/insertion process was further demonstrated by potentiostatic intermittent titration technique (PITT) measurements ( Figure 6f). As expected, the reversible flat voltage plateau is better defined by using this electrochemical method. More importantly, the current responses are characterized by bell-shaped curves with roughly symmetric current flow patterns upon charge and discharge meaning that the electrochemical process is not diffusion limited (Cottrellian behaviors) except at the very beginning and at the very end of the overall electrode process, as usual [32]. The related coulometric measurements demonstrate again that the reversible capacity is still limited to half of the theoretical value (∆x = 1, Q = 120 mAh.g −1 ) in our experimental conditions as observed in all our former studies on lithiated DHT-based electrode materials [18,21,23]. Although beyond the scope of this study, the electroactivity of the carboxylate functional groups was also assessed by galvanostatic cycling within the 0.05-2.0 V vs. Li + /Li potential range ( Figure S14). A large irreversible capacity was noticed during the first cycle as commonly observed with carboxylate moieties, which can be ascribed to the formation of a solid electrolyte interphase (SEI) [13]. The subsequent plateau that involves the carboxylate lithiation process was found poorly reversible with a rapid capacity loss upon cycling. However, no optimization was performed to improve this electrochemical behavior.

Redox Potential Tuning in the Solid State: From Polymorphism to Through-Space Charge Modulation
At this stage, it should be recalled that the relation between polymorphism and electrochemical insertion properties is well-documented in the literature, especially for inorganic electrode materials [33]. The key component is the crystal arrangement that differs from one form to another inducing variation in terms of ion diffusivity inside the crystal packing and/or electronic structure. For instance, the three TiO 2 polymorphs (anatase, rutile, and brookite) distinguish themselves by their Li-insertion properties leading to different polarization and working potentials (1.75, 1.8, and 1.6 V vs. Li + /Li, respectively) [34]. For Nb-based oxides, important gaps of capacity were reported between T-Nb 2 O 5 and B-Nb 2 O 5 (190 mAh.g −1 vs. 15 mAh.g −1 ) or between tetragonal and hexagonal LiNbWO 6 (200.4 mAh.g −1 vs. 20.5 mAh.g −1 ) [35,36]. Tarascon and co-workers [37] have also demonstrated that LiFeSO 4 F (tavorite-type structure) displays an average operating potential of 3.6 V vs. Li + /Li, whereas the triplite-type polymorph reacts at 3.9 V vs. Li + /Li (i.e., +300 mV as potential gain). However, similar studies on organic electrode materials are scarce to date. Recently, Vlad and co-workers [38] reported the electrochemical behavior vs. Li of a three-dimensional polymorph of the aforementioned Mg(Li 2 )-p-DHT, first reported by our group as a 2-D material [23]. Described in this case as a metalorganic framework (MOF) in the archetypical MOF-74 and referred to as Li 2 -Mg-DOBDC (DOBDC 4− = 2,5-dioxido-1,4-benzenedicarboxylate = p-DHT 4-), such a 3-D structure leads to very different electrochemical features with a large polarization effect and poor cycling performance. With the discovery of Li 4 -o-DHT (β-phase), this new study also underlines the critical importance of the crystal packing for organic electrode materials, as demonstrated with inorganic counterparts for long time.
This situation calls for further efforts in terms of crystal chemistry to get high-voltage lithiated organic electrode materials and, thus, fabricate high-energy full organic or hybrid Li-ion batteries. Basically, the research efforts that have been made for the last twenty years were (logically) focused on organic molecular approaches [7], and notably on the redox potential tuning by playing with substitutions and other (intramolecular) electronic effects [22,39,40], or, in other words, thanks to the traditional "through bond" rationalization without any three-dimensional geometrical consideration concerning these bonds [22]. It was, in fact, our motivation when switching from the Li 4 -p-DHT electrode material to the ortho-regiosiomer counterpart, which enabled +300 mV as potential gain with the synthesis of the α-phase [21]. However, the crystal arrangement of the redox-active organic molecular units in the solid state can potentially do more and must be further addressed in the future by the community. In fact, the crystal framework can mitigate or enhance such intramolecular "through bond" electronic effects by "through-space" electrostatic modulation of charge distribution depending on the molecular conformation in the crystal packing [41]. A relevant example has been recently reported by Siew et al. [42] in the field of organic batteries with the tetralithium 2,5-dihydroxy-1,4-benzenediacetate that exhibits a working potential of 3.35 V vs. Li + /Li in the solid phase against 2.70 V measured in solution. Although the respective crystal structures are not known, it can be hypothesized that the potential gain observed when switching from Li 4 -o-DHT (α) to Li 4 -o-DHT (β) is related to conformational considerations. As a result, the cumulative potential gain that exceeds half a volt observed between Li 4 -p-DHT and Li 4 -o-DHT (β) combine the two effects: "through bond" and "through-space" electrostatic modulation of charge distributions.

Materials and Methods
General analytical techniques and characterizations: Liquid 1 H and 13 C NMR spectra were recorded on Bruker AVANCE III at 400 MHz and 100 MHz, respectively. Chemical shifts (δ) are given in ppm relative to TMS. DMSO-d 6 was purchased from Eurisotop Company (purity > 99.5%). For the derivatization reaction, a few drops of H 2 SO 4 were added to the o-DHT-based compounds in a sealed NMR sample tube in a glove box. Elemental analyses (C, H) were obtained by combustion analyzer using either a Flash 2000 CHN analyzer (Thermo Scientific) for H 4 -o-DHT or an UNICUBE analyzer (elementar) for the corresponding lithiated salts. The Li content was precisely determined by Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) thanks to an iCAP 6300 radial analyzer (Thermo Scientific). A mono-element solution of Li (1000 ppm, CHEMLAB) was used for the calibration. Fourier-transform infrared spectroscopy measurements were recorded on a FT-IR Vertex 70 (Bruker), and pellets were prepared by mixing the synthesized materials with spectroscopic-grade potassium bromide at 0.5% (w/w). The thermogravimetricdifferential scanning calorimetry (TG-DSC) analyses were performed with a SENSYS evo (Setaram) whereas thermal analyses coupled with mass spectrometry were performed with a STA449F3 Jupiter and QMS403C Aëolos instruments (NETZSCH). Scanning electron microscopy (SEM) images were collected on a JSM-7600F microscope (JEOL).
X-rays diffraction techniques and refinements: Temperature-resolved X-Ray Powder Diffraction (TRXRPD) experiments were performed on a D8 Advance diffractometer (Bruker) equipped with an Anton Parr Chamber HTK 1200 N high-temperature attachment. Data were collected in Bragg-Bentano geometry with a Cu-anode X-ray source operated at 40 kV and 40 mA. The Cu-Kβ radiation was filtered by means of a Ni foil. The experiment was performed under nitrogen flow with a step of 0.016724 • and an acquisition time of 1.2 s per step (heating rate of 0.2 • C·s −1 ). Single crystal X-ray diffraction measurement were performed at room temperature using a XtaLAB Synergy diffractometer (Rigaku) operating with Mo-Kα radiation (λ = 0.71073 Å). The powder XRD (PXRD) patterns were typically recorded in air at room temperature with a Bruker D8 diffractometer operating with Cu-Kα 1 radiation (λ = 1.54060 Å) and a LynxEye detector in Bragg-Brentano geometry except for capillary powder samples. In the latter case, PXRD patterns were collected in transmission mode on an INEL XRG CPS120 diffractometer operating with Cu-Kα 1 radia-tion (λ = 1.54060 Å) and equipped with a curved detector, ranging from 0 to 120 • . Pattern indexing, space group determination and structure solution by direct space methods using parallel tempering algorithm and Monte Carlo global optimization were performed in FOX software [43]. Le Bail refinements were performed in Jana2006 [44] using a pseudo-Voigt function for the peak profile. The asymmetry of the peaks was corrected with the Berar-Baldinozzi function [45]. CO 2 measurement: In order to quantify the carbon dioxide release, 5 mg of (Li 2 )(Li,H)o-DHT was placed in a glass ampoule, under argon atmosphere, and heated at 300 • C at 3.5 • C.min −1 . The gaseous atmosphere was then analyzed with a micro-gas chromatography set-up (µ-GC-R3000, SRA instrument) using ultra-high purity (99.9999%) argon and helium carrier gases. Prior to the acquisition of the chromatogram, the gas contained in the ampoule was mixed with argon at 1.5 bar for five minutes. Various gases can be identified and quantified (H 2 , O 2 , N 2 , CH 4 , CO, and CO 2 ). Only CO 2 was identified in the present work. It was then quantified. The response of the micro-gas chromatography towards CO 2 is linear in a wide concentration range. The estimated error in the gas measurement was less than 10%.
2,3-Dihydroxyterephthalic acid (H 4 -o-DHT): 1,2-Dihydroxybenzene (2.00 g, 18.2 mmol) and potassium hydrogen carbonate (5.89 g, 58.8 mmol, 3.23 eq.) were ground with a mortar and pestle then placed in a Parr reaction system (40 mL) with dry ice (11.69 g, 265.7 mmol, 14.6 eq.) to reach 20 bars. The Parr reactor was then heated overnight to 225 • C leading to a final auto-generated pressure of 45 bars. The recovered crude material was suspended in diethyl ether (50 mL) for 30 min then filtrated. The dark green powder was dissolved in 1L of water for one hour and filtrated on Celite 535. Eventually, concentrated HCl was added to the liquid to precipitate 2,3-dihydroxyterephthalic acid. After filtration on Millipore PVdF, the creamy beige precipitate was dried under vacuum at 120 • C for 24 h resulting in pure H 4 -o-DHT (2.60g, 72%). 1  Electrochemical study: The electrochemical performance of the materials was tested vs. lithium in Swagelok ® -type cells using a Li metal disc as the negative electrode and fiberglass separators (Whatman ® ) soaked with 1 M LiClO 4 in propylene carbonate (PC) as the electrolyte unless otherwise noted. The PC solvent and the common "LP30" battery grade electrolyte (i.e., LiPF 6 1 M in EC:DMC 1:1 vol./vol.) were directly employed as received from Novolyte. The positive electrodes were prepared without binder in an argonfilled glove box by grinding a powder of Li 4 -o-DHT (β-phase) with 33 wt% carbon black (Ketjenblack EC-600JD, AkzoNobel denoted KB600 or C-NERGY SUPER C65, IMERYS denoted C65) with a pestle in a mortar. The typical positive electrode weight was ≈3.5 mg. The electrochemical measurements were performed by using a MPG system (Bio-Logic S.A., Seyssinet-Pariset, France). The apparent cell resistance (R app. ), also known as internal cell resistance, was determined by the simple current interrupt technique by using the open circuit voltage (OCV) periods integrated in the cycling program at each half-cycle.

Conclusions
In this complementary study focused on the ortho-regioisomer backbone of the DHT ligand, a novel partially lithiated intermediate (Li 2 )(Li,H)-o-DHT compound was isolated and characterized. Its controlled thermal pyrolysis allowed preparing a polymorph (βphase) of the formerly described Li 4 -o-DHT compound (α-phase). The electrochemical investigations of this new polymorph revealed an electrochemical activity vs. Li still limited to half of the theoretical capacity in our experimental conditions. Unfortunately, the electrode reaction was based on a sluggish biphasic transition process inducing a polarization effect (∆E ≈ 120 mV) combined with limited capacity recovery. Although the β-phase can still sustain a reversible capacity value of ≈75 mAh.g −1 over several dozen cycles following the first ten cycles, the overall electrochemical performance was judged lower than that observed with the α-phase. Nevertheless, this study was the occasion to notice a positive potential shift for this polymorph estimated to +250 mV compared to the α-phase, which demonstrates the critical importance of the crystal arrangement by mitigating intramolecular electronic effects by through-space charge modulation. This finding paves the way for further research by smartly tuning electronic effects at molecular level as well as in the solid state or in other words, by combining molecular engineering (organic chemistry) and crystal design (materials science). under pure O 2 (combustion) and measured in a quartz capillary due to the small amount of recovered powder; Figure S5: Typical 1H NMR (a) and 13C NMR (b) spectra measured in H 2 SO 4 /DMSO-d6 solution (derivatization reaction) of the recovered powder (new phase) after thermal treatment of (Li 2 )(Li,H)-o-DHT at 290 • C during 20 h; Figure S6. Rectilinear calibration graphs for the lithium quantification of the recovered powder after thermal treatment of (Li 2 )(Li,H)-o-DHT at 290 • C during 20 h by ICP-AES considering two atomic emission lines for Li (610.3 nm and 670.7 nm, respectively); Figure S7. Overlaid of the representative PXRD patterns for Li 4 -o-DHT (α-phase) in green and Li4-o-DHT (β-phase) in orange extracted from TRXRPD measurements at 245 • C [14] and 290 • C ( Figure 3 in the main text), respectively (* corresponds to the sample holder); Figure S8. Typical FT-IR spectrum (KBr pellet) of the recovered powder (new phase) after thermal treatment of (Li 2 )(Li,H)-o-DHT at 290 • C during 20 h (in orange); the as-obtained phase is ascribed to Li 4 -o-DHT (β-phase); Figure  S9. Chromatogram measured after heating of (Li 2 )(Li,H)-o-DHT to 300 • C at a heating rate of 60 • C.min −1 ; Figure S10. Evolution of FT-IR spectra of Li4-o-DHT (β-phase) after air exposure; Figure S11. First five cycles of a Li half-cell using Li 4 -o-DHT (β-phase) as the positive electrode material (carbon additive: 33 wt% KB600) galvanostatically cycled at 1 Li + /10 h as reported in Figure 6a