Raphael L.
Streng
a,
Sergei
Vagin
bc,
Yuejie
Guo
a,
Bernhard
Rieger
c and
Aliaksandr S.
Bandarenka
*ab
aPhysics of Energy Conversion and Storage, Department of Physics, Technische Universität München, James-Franck-Str. 1, 85748 Garching, Germany. E-mail: bandarenka@ph.tum.de
bCatalysis Research Center TUM, Ernst-Otto-Fischer-Straße 1, 85748 Garching, Germany
cWACKER-Lehrstuhl für Makromolekulare Chemie, Department of Chemistry, Technische Universität München, Lichtenbergstraße 4, 85748 Garching, Germany
First published on 21st March 2024
Polyimides such as poly(naphthalene four formyl ethylenediamine) (PNFE) demonstrate promising electrochemical properties as anode materials for aqueous sodium-ion batteries. They display high specific capacities and good cyclability. Still, an accurate physical model describing the exact charge storage mechanism in these electroactive polymers needs to be explored in more detail. This work aims to address this gap through an impedance study to elucidate the mechanisms of reduction and oxidation processes of polyimide electrodes. An electrochemical impedance model is also proposed, which enables good fitting of the impedance data at all potentials. This model confirms the commonly accepted two-stage enolization reaction as the main reduction path and also unveils a previously unrecognized single-electron pathway. A complete model was identified by including this novel pathway, which enabled a better understanding of these materials.
Redox-active organic materials can be a sustainable, inexpensive, and high-performance alternative to inorganic compounds that often contain scarce or polluting elements.26,27 Polyimides (PIs) have emerged as a promising anode material for aqueous metal-ion batteries due to their low electrode potential, high cycling stability, and superior specific capacity.28,29 In addition, they exhibit remarkable mechanical strength, thermal resistance, and chemical stability and are fully compatible with aqueous electrolytes.24
Whereas the exceptional electrochemical performance of PIs has been reported several times,28–30 a more detailed examination of the underlying charge transfer processes should be further performed. Ultimately, understanding the charge storage mechanism in PIs is of great importance to unlock their full potential and accelerate their application in practical systems. Electrochemical impedance spectroscopy (EIS) is a non-destructive, non-invasive, and fast technique that can reveal charge transfer mechanisms and characterize degradation as well as mass transport processes.31 By applying a low amplitude sinusoidal voltage or current signal over a certain frequency range, EIS can fully extract any information on the impedances in the cell. However, to understand the data, it is essential to interpret the spectrum with an accurate physical model, often expressed for simplicity in terms of equivalent electric circuits (EECs).32 Thus, for each electrode, complete EECs that describe the physical processes in a meaningful way need to be created. These models can then be combined to gain information on the full battery.
This work provides new insight into the redox mechanism of the electrode material poly(naphthalene four formyl ethylenediamine) (PNFE, see Fig. 1) by subjecting it to an in-depth EIS study.
Poly(naphthalene four formyl ethylenediamine) was prepared according to literature procedure with minor modification.28 In brief, 1,4,5,8-naphthalene tetracarboxylic dianhydride (NTCDA, 2.68 g, 10 mmol) and ethylenediamine (EDA, 0.6 g, 10 mmol) were mixed with 20 ml of degassed N-methyl pyrrolidone (NMP) in a Schlenk flask. The flask was equipped with an air-cooled condenser and an overpressure valve on top, and the reaction mixture was heated overnight in a sand bath at a bath temperature set to 215 °C. After cooling, the dark precipitate was filtered off and washed thoroughly with NMP and subsequently with water. It was dried in vacuo with a stepwise increase of drying temperature (room temperature, 120 °C and 210 °C) to a constant vacuum of 0.08 mbar yielding 2.7 g (92%) of grey-brownish powder.
Elemental analysis: calcd for an infinite polyimide (C16H8N2O4), %: C 65.76, H 2.76, N 9.59; calcd for an octamer, %: C 65.10, H 3.03, N 10.52; found, %: C 65.13, H 2.91, N 10.32.
CV (cyclic voltammetry) and EIS (electrochemical impedance spectroscopy) measurements were conducted using a BioLogic VSP-300 Potentiostat. To prevent any disturbances caused by the impedance of the reference electrode, a small shunt capacitor was connected between the reference electrode and an additional platinum wire, which was wrapped around the Luggin capillary and ended near it.
The cyclic voltammograms (CVs) were recorded at different scan rates ranging from 1 mV s−1 to 50 mV s−1 between 0 V vs. SSC and −1.1 V vs. SSC. Staircase electrochemical impedance spectroscopy was carried out in the same potential range in steps of 100 mV. Before recording the spectrum at each step, the potential was held for 10 minutes to ensure equilibrium conditions, which are vital for impedance measurements. A 10 mV AC amplitude was applied to the electrode with frequencies ranging from 1.00 MHz to 250 mHz. After the validity of the data was verified by Kramers–Kronig analysis, the spectra were fitted using the EIS Data Analysis 2.1 software.35
Fig. 2(C) shows the TGA curve of PNFE. The small initial mass loss is caused by the evaporation of residual water within the active material. Decomposition of the polymer can only be observed at a temperature of approximately 500 °C which confirms the excellent thermal stability of PNFE. To analyze the morphology of the PNFE electrode scanning electron microscopy (SEM) was conducted. It reveals a relatively even distribution of polymer agglomerates, which are typically less than 1 μm in size, as shown in Fig. 2(D–F). Together with the Nafion™ binder and the carbon black, the active material forms a porous film with pore sizes in the nm to μm range. To further analyze the surface structure of the PNFE electrodes, XPS spectra were recorded. The survey scan depicted in Fig. 2(G) shows characteristic peaks for carbon, nitrogen, oxygen, and fluorine. Carbon and oxygen are part of the active material and the binder, whereas nitrogen and fluorine can be attributed to PNFE and NafionTM, respectively. In the C 1s spectrum (Fig. 2(H)), multiple peaks can be observed. The sharp peaks at 284.8 eV, 286.1 eV, 289.2 eV, and 292.4 eV correspond to C–C, C–N, CO, and CF2 bonds, respectively. The weaker peaks at 284.3, 287.4, and 293.8 are attributed to CC, O–CO or N–CO, and CF3. In addition, the extended delocalized electrons in the aromatic rings of PNFE result in a π–π* satellite structure at 291.5 eV. These results indicate that the active material and the binder are both present at the surface of the electrode, which is necessary to ensure mechanical stability as well as electrolyte access.
The electrochemical properties as well as the rate dependence of PNFE were investigated by measuring CVs at various scan rates. A typical CV recorded at 10 mV s−1 is shown in Fig. 3(A). Two separate pairs of reduction and oxidation peaks can be clearly distinguished. The oxidation peaks lie at −0.68 V vs. SSC and −0.39 V vs. SSC respectively. The reduction peaks are close at −0.87 V vs. SSC and −0.61 V vs. SSC implying a reversible redox reaction. This double-peak structure supports the two-electron reduction mechanism which is commonly reported for PNFE and similar polyimides.36,37 As shown in Fig. 4, the reduction leads to a two-step enolization reaction of two opposite carbonyl groups. In the first step, a radical anion is formed followed by the formation of the di-anion which allows it to adsorb two sodium ions. It should be noted that this is the generally accepted scheme for ion adsorption at carbonyl-based organic electrode materials. Since there are four carbonyl groups, one may expect further reduction by up to four electrons. Whereas this is generally possible, it only occurs at much lower potentials and thus below the hydrogen evolution reaction (HER) threshold.38 The relatively wide expansion of the peaks in the x-direction suggests repulsive interactions between the two redox centers preventing the resolution of the two separate electron transfers. Therefore, a full understanding of the redox mechanism is only accessible by employing additional techniques. Here we use EIS to reveal the underlying processes.
Fig. 4 Reduction mechanism of PNFE. In the first step, a radical anion is formed, followed by the generation of the dianion in the second step. Further reduction (3 or 4 electrons) is possible but only occurs at much lower potentials.38 |
Fig. 3(B) shows the CVs of PNFE at different scan rates between 1 and 50 mV s−1. The peak-to-peak separation ΔEp–p increases with increasing scan rate, which is mostly linked to the uncompensated resistance stemming from the electrolyte's ionic conductivity. However, it becomes clear that the charge transfer mechanism is highly reversible as ΔEp–p drops to 59 mV at a scan rate of 1 mV s−1 for both pairs of peaks. By relating the peak current ipeak to the scan rate ν it is possible to deduce information on the processes limiting the charge transfer and, therefore, determine the current response in the CVs:
(1) |
The total transferred charge within one cycle can be calculated by integrating the CVs. Fig. 3(D) shows that a maximum specific discharge capacity of ca. 153 mA h g−1 is measured at the slowest scan rates. As indicated by the dotted line, this value is significantly lower than the theoretical specific capacity of PNFE for a two-electron charge transfer. This observation is in accordance with previous studies on various polyimides, which consistently report capacities considerably below the theoretically expected value.28–30,37 This implies that the redox mechanism of PNFE is not sufficiently described by the simple two-electron transfer shown in Fig. 4.
Electrochemical impedance spectra were recorded at different potentials within the potential range in which the redox activity was observed in the CVs to obtain further insights into the mechanism. The fitting of these spectra requires a physically meaningful EEC model. Fig. 5(A) shows the equivalent circuit describing the two-stage mechanism which is depicted in Fig. 4. This circuit model can be derived from the so-called Randles circuit modeling faradaic processes controlled by diffusion and kinetics.39 The Randles circuit consists of the electrolyte uncompensated resistance RU, the double layer capacity Cdl, the charge transfer resistance Rct, and a semi-infinite Warburg element W accounting for diffusion. In this case, the diffusion limitations are assumed to arise due to local sodium-ion deficiencies within the pores of the slurry, as shown in Fig. 6. Though the concentration of sodium ions in the bulk electrolyte is very high, the movement through the smaller pores is likely diffusion-limited. In addition, the delocalization of the charge over the conjugated structure of the imide molecule affects mass transport as well. To account for the two stages of the mechanism, an adsorption capacitance Ca,2 and resistance Ra,2 are added to the Randles circuit.34 Hereby, the model reflects the fact that the reduction of the first carbonyl group determines the second step of the redox process, which unquestionably takes place after the transfer of the first electron. Thus, the two stages are conjugated, and Ca,2, and Ra,2 are connected in the circuit accordingly. The impedance linked to the contribution of the film on the graphite substrate to the overall impedance response is accounted for by Cgeo and Rgeo.
In other words, this model describes a conjugated two-electron transfer with reversible stages, which corresponds to the mechanism shown in Fig. 4. It should be noted that the capacitative contribution could also be modeled with constant phase elements (CPEs) instead of capacitors. However, the direct physical meaning of the model is not clear, especially when multiple CPEs are used. In the case of the PNFE electrode, the fit quality does not improve significantly when Cdl or Cgeo are replaced by CPEs. Furthermore, upon fitting, the exponent of the CPEs is very close to 1.00, indicating nearly ideal capacitative behavior. The resulting EEC depicted in Fig. 5(A) can be used to fit the data with high accuracy at potentials more positively than −300 mV.
Fig. 7 shows the impedance spectra at different potentials ranging from −300 mV vs. SSC to −1000 mV vs. SSC in Nyquist plots as well as the dependence of the phase shift Θ on the frequency. At −300 mV vs. SSC, the processes are fairly modeled by the two-stage mechanism. However, the spectra recorded at more negative electrode potentials within the region where faradaic processes take place do not support this model. At −400 mV and below, an additional contribution can be observed, leading to a different curvature in the transition region between high and low frequencies. This effect is even more distinctive in the Θ vs. frequency representation, which displays a different shape at these potentials. The exact origin of this is outside the scope of this manuscript as it does not influence the fitting statistics significantly.
Fig. 7 Fitted EIS spectra of PNFE at different potentials vs. SSC. The graphs on the left show the Nyquist plots (A), (C), (E), (G) and (I). The graphs on the right show the dependence of the phase shift Θ on the frequency (B), (D), (F), (H) and (J). For fitting the EEC in Fig. 5(B) was used. |
Simultaneously to the primary two-stage conjugated mechanism, a one-electron single-stage mechanism was found to take place. This process can be modeled by a charge transfer resistance Rct,1 and an adsorption capacitance Ca,134 that are connected in parallel to the elements representing the two-stage mechanism. By including these parameters, the spectra can be fitted with extremely high accuracy for all potentials, as shown in Fig. 7. This pathway is most likely linked to the formation of the stable radical anion after reduction by a single electron. This observation indicates that a certain portion of the polymer is not further reduced after this step. Reasons for this may be polymer–polymer or polymer–substrate interactions. Due to this effect, the experimental specific capacity of PNFE is expected to be lower than the theoretical specific capacity for a two-electron charge transfer as parts of the polymer only accept one electron. Therefore, the observed capacity losses serve as an important verification of the suggested EEC.
The fitting parameters and their respective errors for all potentials are given in Table S1 (ESI†). It shows that the charge transfer resistance of the two-stage conjugated pathway is significantly lower than the charge transfer resistance of the single-stage pathway. At all potentials, Rct,1 is 5–10 times higher than Rct,2, with the difference getting smaller towards lower potentials. Therefore, the two-electron mechanism is determined to be dominating, whereas the single-electron mechanism is mostly suppressed. At −300 mV vs. SSC, the difference is the largest, which is why the spectrum can also be fitted with the simpler circuit that is depicted in Fig. 5(A). However, at lower potentials, the ratio decreases, and the contribution of the single-stage mechanism becomes significant.
The introduction of the electrochemical equivalent circuit (EEC) proposed in this study enables highly accurate and minimally error-prone fitting of impedance spectra for polyimide electrodes. Moreover, the implications extend beyond this specific material, as analogous organic electrode materials relying on enolization reactions can likely be similarly modeled. Consequently, the outcomes of this research offer a potent analytical tool for comprehending the processes within carbonyl-based electrode materials and batteries employing such electrodes.
Beyond its fundamental contributions, this work paves the way for more comprehensive EIS analyses, opening up novel avenues for monitoring degradation effects, cell aging, and various processes in a non-destructive manner. The knowledge gleaned from this study promises to be instrumental in advancing the understanding of organic electrode materials, ultimately contributing to the optimization and longevity of electrochemical systems.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ya00037d |
This journal is © The Royal Society of Chemistry 2024 |