Triphenylamine-Contained Multiporous Polyimide: Its Preparation and Application as Anode for Lithium-Ion Storage

Organic-based electrode materials as the next-generation sustainable Li-ion batteries (LIBs) have attached widespread attention due to the characteristics of multiple redox-active centers. Organic compounds possess high structural flexibility, which allows to tailor their redox properties by tailoring technology of molecular engineering [1–3]. As a result, it has made great efforts to explore the various organic electrode materials for rechargeable batteries, including organosulfur [4,5], organic free radicals [6–8], conjugated carbonyls [9,10], and conducting polymers [11,12]. Imide compounds are one kind of carbonyl compounds, and because of their high theoretical specific capacity and fast redox reaction characteristics, imide compounds and their polymers are becoming a promising organic electrode materials in the future energy storage field [13].

Small-molecular imide compounds always suffer from severe dissolution into aprotic electrolytes, which causes serious capacity degradation and reduced utilization efficiency of carbonyl active centers. Comparatively, polyimides (PIs) can largely suppress the solubility problem because of its high molecular weight, which results in the enhanced cycling performance, making it attractive as the organic electrode [13–15]. Redox mechanisms of PIs are mainly attributed to two steps of one-electron reversible redox processes from carbonyl groups of dianhydrides in the polymer, which accompanies with production of both anion radical and dianion, respectively [16]. Typically, the condensation reaction of diamine monomers and dianhydride moieties is mainly way for obtaining polyimide, and the obtained polyimide as the cathode exhibits better electrochemical performances than their parent monomeric materials. Currently, various polyimides such as pryomellitic dianhydride (PMDA), 1,4,5,8-naphthalenetetracarboxylic (NTCDA), and 3,4,9,10-perylenetetra-carboxylic dianhydride have been prepared by condensation reaction, and their applications as the electrode materials for the energy storage had demonstrated the remarkable attractive prospects [17–24].

To achieve high-performance organic-based batteries, recently, the constructed conjugated polymers with nanostructure are recently becoming the focus of attention. Various porous polymers have been extensively researched as the anode and cathode electrode materials in energy storage application [25–28]. Among them, polyarylimide-based conjugated porous polymers have been intensively studied recently as novel organic cathode material because of their designable chemical structure, stable polymer molecular skeleton, and controllable nanopore morphology [29] Usually, porous polyimide can be prepared by choosing of different imide centers [30] and linkages [13,31–34], which makes its preparation more diverse. The constructed porous polymers possess high surface area, which provides abundant redox-active sites exposing to electrolyte, leading to fast ion transport among the organic electrodes. As a result, it results in a positive influence on the cell performances such as the rate performances and the improved specific capacity [35–39]. However, polyimide derivatives among the most studies are usually applied as organic cathode materials in LIBs or sodium ion batteries [35–39], and its exploration as the anode of the energy storage cell is rarely involved.

Therein, through the polycondensation reaction between NTDA and/or PMDA and tris(4-aminophenyl)amine (TAPA) respectively, two triphenylamine-contained polyimides were prepared. Contributed to the introduction of triphenylamine as the linkage, the resultant triphenylamine-based polyimide displayed an improved surface area as well as the well-formed porous structure. Explored as the anode in lithium ion batteries, the resultant triphenylamine-based polyimide showed an improved reversible specific capacity and remarkably rate performances. To our knowledge, it has less been reported for the porous polyimide as the anode of rechargeable batteries with the remarkably electrochemical performances.

Tris(4-nitrophenyl)amine, PMDA, and naphthalene-1,4,5,8-tetracarboxylic dianhydride (NTDA) were obtained from Aladdin Chemical Reagent Co., Ltd. Isoquinoline, m-cresol, tetrahydrofuran, and other reagents were reagent grade and used as received unless otherwise stated.

Tris(4-aminophenyl)amine was synthesized based on a modified procedure as reported in the literature [40,41]. A total of 0.5 g of tris(4-nitrophenyl)amine was first dissolved in 30 mL of N,N-dimethylformamide (DMF) solvent, 0.2 g of Pd/C as catalyst (5 wt%) was then added slowly, and finally 4 mL hydrazine hydrate was added slowly into above mixture. The reaction mixture was kept reaction at 80 °C under nitrogen atmosphere for 20 h. The obtained crude product was recrystallized and washed with ethanol for three times to obtain white crystal product with a yield of 43%. Fourier transform infrared (FTIR) (KBr, cm⁻¹) (Fig. S1): 3406, 3336, 1618, 1502, 1594, 1331, 1261, 1117, and 829. ¹H NMR (400 MHz, dimethyl sulfoxide (DMSO), ppm) (Fig. S2) : 6.84 (m, 6H, Ar-H), 6.57 (m, 6H, Ar-H), and 4.72 (m, 6H, −NH₂).

Both the pyromellitic-based polyimide (N1) and the naphthalene-1,4,5,8-tetracarboxylic-based polyimides (N2) were polymerized in the similar reaction conditions. Typically, polycondensation reaction was carried out as follows: A solution of TAPA (0.41 g, 1.38 mmol) and PMDA (0.46 g, 2.07 mmol) or NTDA (0.55, 2.07 mmol) in m-cresol (20 mL) was reacted in nitrogen atmosphere with ice water bath for ∼4 h. Next, the reaction was heated up to room temperature (∼20 °C) with several drops of isoquinoline as catalyst introduced. The reaction was then performed at room temperature for 8 h and at 100 °C, 140 °C, and 180 °C for 4 h, respectively. Finally, the reaction temperature was further heated up to 200 °C, and the reaction was kept at this temperature for another 12 h. During the reaction process, the produced water was constantly blown out of the system by nitrogen flow. After natural cooling, the precipitated solid product was filtrated and then washed with methanol as a solvent for several times. Finally, the obtained solid was purified for 24 h using soxhlet apparatus with tetrahydrofuran as a solvent, and the product was vacuum dried at 180 °C for 24 h. Yield was 64% for N1 and 78% for N2. The detail synthesis processes are shown in Scheme 1.

Thermogravimetric analyses (TGA) were carried out using a STA 449C (Germany Chi Instrument Manufacturing Co., Ltd.), in which the running temperature was from room temperature to 800 °C at a heating rate of 10 °C/min in nitrogen atmosphere. Hydrogen Nuclear Magnetic Resonance (¹H NMR) spectra was performed using an AVANCE III 500 MHz spectrometer (Bruker, Switzerland). Fourier transform infrared Spectrometer spectra of the polymers were measured on a NEXUS 470 spectrometer (American Thermal Power Company) with KBr pellets. Scanning electron microscopy (SEM) measurements of the samples was observed using a Hitachi SU8010 scanning electron microscope (Hitachi, Japan). Brunner–Emmet–Teller (BET) measurements were carried out with a Surface Area and Porosity Analyzer (Micromeritics, ASAP2020). A Thermo-Finnigan Flash EA-1112 (CE, Italy) instrument was applied for elemental analysis.

For electrochemical characterization of active materials, a film coating method was first applied to prepared the electrodes, in which a mixture slurry containing the prepared active materials (50 wt%), Super P (40 wt%), and poly(vinylidene fluoride) (PVDF) as a binder (10 wt%) in N-methyl-2-pyrrolidone was coated on Cu collector foils and then vacuum dried at 70 °C for 24 h. The loading mass of the composite on the current collector was about 2.3 mg with 1.15 mg of the active materials. After that, a CR 2032 type cell was assembled with the aforementioned prepared electrodes as the cathode, lithium foil as the anode, a Celgard 2300 type pf polypropylene microporous membrane as the separator, and 1M LiPF₆ dissolved in ethylenecarbonate (EC) and dimethylcarbonate (DMC) (EC/DMC = 1:1 v/v) as the electrolyte. All assembled process for battery was performed in a glove box, which filled argon with a less than 0.1 ppm of oxidant content and water content. The assembled battery was kept in the stat of rest for 12 h to permit the resulted composite electrode permeated fully by the electrolyte. A LAND CT2001A setup was applied to carry on the charge–discharge measurements, and the measurement was at different constant current density in 0.01–3.0 V versus Li/Li⁺ in room temperature. CHI 660E electrochemical working station was applied to measure both cyclic voltammograms (CVs) and electrochemical impedance spectroscopy (EIS), respectively. CV testing was carried out using the assembly CR 2032 type cell with a scanning rate of 1 mV/s in the scanning potential from 0.01 V to 3.0 V. EIS tests were performed using the assembled cell, which was measured in open circuit voltage state of 3.0, 1.5, and 0.01 V, respectively, with the frequency change from 0.01 Hz to 100 Hz.

Morphologies of N1 and N2 are initially characterized by SEM measurement. As shown in Fig. 1, the pristine N1 is made up of small micro-size particles (≈700–1000 nm), which further second agglomerates to form the bigger aggregation Figs. 1(a) and 1(c). The second-aggregated morphology structure with small particles is in favor of the permeation of the electrolyte molecular into the polymer matrix, and it, as a result, will benefit to the full utilization of active material and consequently to the improved electrochemical and battery performance. As a comparison, N2 presents a plate-like structure Figs. 1(b) and 1(d), indicating that N2 polymer possesses a well film-formed properties. Furthermore, the porous structure of N1 and N2 is measured by nitrogen adsorption analyses at 77.3 K. The two polymers show the different nitrogen adsorption/desorption isotherms, which implies the polymer possess the different pores characteristics Fig. 1(e). N1 at the relative low pressure of P/P₀ < 0.001 displays a rapid nitrogen uptake, which implies the polymer possess a more micropore structure containing in materials, while N2 shows a less micropore structure, corresponding to a relatively low nitrogen uptake at the same pressure range. Based on International Union of Pure and Applied Chemistry (IUPAC), N1 and N2 belong to a typical adsorption behavior of I and II. In addition, a slight hysteresis loop is observed for N1, which persists to the low pressures upon the desorption branch. It is possible due to elastic deformation or swelling behavior of polymer. For the N2, a comparative large hysteresis loop is observed consisted in both the low pressures and the high pressures, indicative of the wide pore distribution in the N2. The measured BET surface areas are 738 m²/g for N1 and 456 m²/g for N2, respectively. The significantly different BET surface area for triphenylamine-contained polyimide implies that the different imide centers of N1 and N2 have remarkably affected on the aggregation behavior of the obtained polymers. The pore size distributions of the polymers are also calculated Fig. 1(f), in which N1 exhibits a pore size range in 1.7–4.2 nm, while N2 mainly shows a wide porous distribution with the hierarchical pores from 1.8 to 20 nm. Also, a 0.53 cm³/g of pore volume is observed for N1, which is larger than that of N2 (0.31 cm³/g). The obviously improved surface area and pore volume for the pyromellitic-contained polyimide (N1) will provide full contact of active centers of electrode material with electrolytes during the electrochemical process. Therefore, it is anticipated that the high-energy storage performances could be obtained for N1.

The chemical structure for the obtained N1 and N2 polyimides is further analyzed by FTIR spectra. After polymerization, two strong absorptions at ∼1717–1722 cm⁻¹ and ∼1631–1654 cm⁻¹, assigned to typical symmetric and asymmetric vibrations of C=O group, are displayed in the pyromellitic-based polyimide (N1) and the naphthalene naphthalene-1,4,5,8-tetracarboxylic-based polyimides (N2), indicative of the formed imide ring in the resulted polymer Fig. 2(a). Still, the C–N–C stretching vibration band was observed at 1345–1137 cm⁻¹. In addition, it is observed that a broad band at around 3400 cm⁻¹ occurs for both N1 and N2, which is probably due to the overlapping effect of absorption peaks for both the amino group in triphenylamine and the –OH of H₂O. FTIR indicates that the triphenylamine-based polyimides of N1 and N2 have been successfully polymerized by condensation reaction of triphenylamine monomers with NTDA and PMDA, respectively. However, a weak absorption should-peak at 1776–1788 cm⁻¹ is observed in two polymers, which implies that the condensation polymerization is not very complete, and there is a small amount of un-reacted end groups of anhydride and amine groups remaining in both polyimides. Elemental analysis for the prepared polymers has also been measured, and the results are presented in Table 1. As can be seen, the measured contents of C, O, and N element are almost consistent with theory calculation values, indicating that polyimide polymers of N1 and N2 have been prepared. For further analysis, it is found that the measured contents of C and N elements are slightly larger than that of theory calculation values, while measured content of O element is smaller than that of theory calculation value, which indicates that N1 and N2 contain much of triphenylamine units in polymers. Furthermore, the calculation H content based on the measured C, N, and O content are 2.71 and 3.01 wt%, respectively, which is slightly higher than that of the theory values, implying of some polyamide acids occurring in the N1 and N2, due to incomplete imidization reaction during high-temperature polymerization. Furthermore, the thermal stabilities of the resulted porous polyimide (N1 and N2) are investigated comparatively, and TGA measurements are conducted under N₂ atmospheres with the temperature change from room temperature to 800 °C Fig. 2(b). Two samples exhibit the similar thermal degradation characteristics. In detail, the weight loss for two samples is almost ignorable with the temperature increasing up to over 150 °C, indicative of excellent heat resistance. Then, the first thermal decomposition process occurs from ∼150 °C to until 510 °C, with about 17 wt% of weight loss for N1 and 20 wt% of weight loss for N2, respectively. Following it, the second thermal degradation occurs up to 800 °C, with the 55.3 wt% content of carbonized residue (char yield) for N1 and 48.2 wt% for N2, respectively. It indicates that the pyromellitic-based polyimide (N1) exhibits the more thermal stable than the naphthalene-based polyimides (N2).

Electrochemical performances of N1 and N2 are then evaluated, and Fig. 3(a) exhibits the obtained CV curves for N1 and N2, in which measurement was performed at a scan rate of 1 mV/s in 0.01–3.0 V. Both polyimides exhibit respectively two pairs of oxidation/reduction peaks at relative high potential and a pair of extend peaks at the low potential, indicative of a successive multi-electrons inserted/removed process from polyimides during the oxidation and reduction reaction. For N1, the occurred redox peaks at 1.92/1.64 V and 0.95/0.84 V corresponds to the sequential gain-loss process of two electrons in two carbonyl groups of the imide structure, while a pair of extend peaks at the low potential is due to redox process of the aromatic structure in the polymers. By contrast, N2 still exhibits the similar redox peak characteristics as N1 but with an obviously large difference of the redox potentials for high potential, implying of its larger electrochemical polarization process. In addition, N1 electrode, compared to N2 electrode, exhibits significantly enhanced peak intensity and larger peak areas, suggesting its highly electrochemical activity. Specially, the intensity for the extend redox peaks at the low potential is obviously larger than that of the other two pairs of redox potentials, which implies its high contribution of capacity as the anode.

Figures 3(b) and 3(c) show the charge and discharge curves of N1 and N2 at the first, second, and third cycle, respectively, measured at a constant current rate of 30 mA · g⁻¹ using the button cell. As shown in Fig. 3(b), N1 in the initial cycle exhibits two discharge voltage platforms at ∼1.75 and 1.0 V and a prolonged discharge curve at low potential, and this observed multi-electron redox character is consistent with the obtained CV results, the gain/loss of two electrons from the carbonyl groups in the imide structure of polyimides, and the other electrons from aromatic structure in the polymers. Correspondingly, its initial specific capacities for charge and discharge process are 350.2 and 670.4 mAh/g, corresponding with columbic efficiency of 52.2%. This lower coulombic efficiency is considered to irreversible side reaction during the formation of the solid electrolyte interface (SEI). In the second and the third cycle, the discharge voltage platforms at ∼1.75 V and 1.0 V and an extended discharge curve at low potential still occur but decay, meanwhile its charge/discharge-specific capacity reduce remarkably. For N2 Fig. 3(c), it demonstrates the similar curve characteristics with discharge voltage platforms at ∼1.80 V and 0.9 V and a prolong discharge curve at low potential but with even more serious capacity decay for the second and the third cycles than that of the first cycle. The charge and discharge cyclic stability of N1 and N2 is also measured at scan rate of 30 mA/g Fig. 3(d) for 100 cycles. N2 displays the acceptable cycling stability. Its discharge specific capacity decreases in the initial several cycles, which is still attributed to the irreversible formation of SEI on the surface of anode. Then, the discharge specific capacity remains relatively stable with smoothly change from 312 mAh/g of the second cycle to 297.5 mAh/g of the 100th cycle. Comparatively, N1 exhibits a different and interesting cycle characteristic, that is, a remarkably increased capacity is observed upon cycling, which has been less reported in the previous organic and inorganic electrodes materials. After 100 cycles, the discharge capacity increases from 385 mAh/g of the second cycle to a famously large value of around 420.3 mAh/g. The increase of capacity is probably due to (1) gradual activation of active centers in electrode materials during charge and discharge process and (2) added charge storage from stacked aromatic structure in N1. Also, the columbic efficiency for N1 and N2 samples are low in the initial cycles, which tends to stable values of nearly 98–99%, implying that stable SEI film has formed on the surface of the electrode in the following cycling.

The possible charge storage mechanism can be as described as followings. As shown in Fig. 4, the redox behaviors of aromatic polyimides in the first step are mainly due to the redox reaction of the occurred carbonyl groups in dianhydrides, which involves two successive electron gain/loss steps companying with occurring of the anion radical and the dianion, respectively. In such systems, Li ion is expected to form the stabilized Li enolatesis structure, which is in favor of forming the stable extended conjugation and enhances the reversibility and cycling stability of electrode materials [16]. While, at low potentials, Li ions may coordinate with the unsaturated aromatic ring to form lithiated aromatic complex, which acts as the lithium storage function with the decrease of potential, just as reported in the previous study [42]. Furthermore, the little occurred amide structure in the polymer, due to incomplete imidization reaction, may also contribute to charge storage.

Rate capability is another key performance for the application of storage battery, and Fig. 5 shows the discharge specific capacity of N1 and N2 at different discharge current of 50, 100, 200, 300, 500, 700, and 1000 mA/g, respectively. N2 electrode material exhibits a generally decreased discharge capacity of 253.2, 220.9, 198.5, 169.1, 141.8, and 108.8 mAh/g at the current rate of 50, 100, 200, 300, 500, 700, and 1000 mA/g respectively, due to the enhanced polarization under the gradually increased current density Fig. 5(a). For N1, it shows an abnormal rate performance, its discharge-specific capacity increases with the current rate enhances from 50 mA/g to 100 mA/g, and then decreases with the further increased current rate. Its discharge specific capacity still remains ∼251.8 mAh/g at the rate of 1000 mA/g, which is remarkable larger than its initial one. The unusual current rate characteristic can still be ascribed to gradually activation process of electrode materials during cycling, which offsets the decrease of capacity at high current rate. As remeasured at 30 mA/g, the specific capacity of N1 returns back to 425.8 for N1 mAh/g and 300.5 mAh/g for N2, indicative of the superior capacity recovery performance for both polyimides. Figure 5(b) further presents the charge–discharge curves for both N1 and N2 electrodes at various current rates. As shown, two electrodes exhibit the relatively stable voltage plateau, especially for N1 electrode, and its discharge curve still keeps insignificant change at the high rate of 1000 mA/g, implying that it has a relatively low polarization at various current rates.

EIS for N1 and N2 have also measured at the open circuit voltage of 3.0, 1.5, and 0.01 V, respectively, as shown in Fig. 6. Two electrodes exhibit the similar characteristics with the frequency scanning from 0.01 Hz to 100 Hz, namely, it is composited of both straight line in low frequency region and semicircle in high frequency region. Therein, the diameter of the semi-circular at high frequency region represents the charge transfer impedance (R_ct) at the electrode interface, and the straight line in low frequency region indicates that diffuse control process dominates during the charge/discharge process. As shown, the diameters of the semi-circular from the curves are about 316, 36, and 1485 Ω at the open circuit voltage of 3.0, 1.5, and 0.01 V for N1 and 562, 187, and 2520 Ω at the open circuit voltage of 3.0, 1.5, and 0.01 V for N2, respectively, which indicates that N1 electrode has remarkably smaller interface charge transfer impedances than that of N2 at three open circuit voltage. Its open pore morphology as well as the resulted high specific surface area is responsible for the improved charge transfer on the surface of active material, which is the reason for the obtained high rate performances for N1 electrode.

In this study, the triphenylamine-contained polyimides were synthesized by a polycondensation of both NTDA and PMDA with TAPA, respectively. As a result, the obtained triphenylamine-contained polyimides exhibited the open porous structure and high specific surface area (738 m²/g for N1 and 456 m²/g for N2), respectively. The obtained polyimide electrodes as the anode of lithium ion batteries displayed multi-electrons redox characteristics in 0.01–3.0 V, including of the redox behaviors of aromatic polyimides and the unsaturated carbons of C₆ ring, respectively. Also, two polyimides presented an increased reversible specific capacity and an enhanced current rate performance. The cycling stability testing indicated that N2 showed the acceptable cycling stability with the 4.65% of capacity decay after 100 cycles, while N1 delivered a different cycling characteristic, in which the specific capacity increased gradually up to 420.3 mAh/g during the cycle testing. Furthermore, N1 delivered a significant rate performance, and the 251.8 mAh/g of specific capacity could still be obtained at the rate of 1000 mA/g, which makes it promising as anode of rechargeable batteries.

This research was financial supported by the Natural Science Foundation of Liaoning Province, China (Grant No. 2020-MS-232), Support Plan for Innovative Talents in Colleges and Universities in Liaoning Province (LR2017034), the National Science Foundation of China (Grant No. 51573099), and Liaoning BaiQianWan Talents Program ([2020]78) (2020921096), Scientific research project of Liaoning Provincial Department of Education (LJ2020004 and LJ2020005).

There are no conflicts of interest.

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request. The authors attest that all data for this study are included in the paper. Data provided by a third party listed in Acknowledgment. No data, models, or code were generated or used for this paper.