Oxygen-Functionalized Polyacrylonitrile Nanofibers with Enhanced Performance for Lithium-Ion Storage.

Functionalization and morphological construction can promote lithium-ion storage performance of organic polymers. In this contribution, exceptional lithium ion storage performance is empowered to porous polyacrylonitrile (PAN) nanofibers via the integration of template-assisted electrospinning technology and thermal treatment. It is found that the atmosphere adopted during the annealing process controls the storage behaviors of Li+. Impressively, the samples annealed in air present competitive capacities, rate capabilities, and a stable lifetime, compared with other counterparts (PAN powders and PAN fibers treated in N2). Such enhancement in performance is attributed to the enriched oxygen-based functionalities (mainly C=O group) which guarantee a high specific capacity and the porous structure which facilitates the transportation of Li+ and electrons to improve the rate capability. It is envisioned that such morphology control and surface functionalization open up new horizons in the development of organic electrode materials with enhanced lithium-ion storage performances.

Introduction

Organic materials have been considered as potentially promising candidates for the new generation of “green batteries” with ever-improving electrochemical performances.[1,2] In general, polymers with active redox sites or groups demonstrate more excellent stability than small molecules during the cycling measurements, benefitting from their negligible dissolution in organic electrolytes.[3,4] However, polymers also suffer from the following drawbacks including the low ionic and electronic conductivity and the limited active sites, leading to unsatisfactory capacity and poor rate capability for lithium-ion storage.[5,6] Hence, to further optimize the storage performance, the functionalization of polymers with improved conductivities of electrons and lithium ions and enriched active redox sites for Li+ accumulation/extraction is highly desirable but still challenging.

In general, polyacrylonitrile (PAN) is considered as a structural polymer and widely used as a membrane support for LIBs or precursors for carbon materials because of its comfortable mechanical hardness, satisfactory biocompatibility, excellent chemical inertness, high melting point, and low permeability.[7−10] Typically, for the preparation of carbon materials, a preliminary thermal treatment is always conducted at relative low temperatures (280–350 °C) under different atmospheres (N2 or air) to motivate the molecular crosslinking with enhanced mechanical strength.[11,12] Despite concentrating on structural evolution during this pretreatment process, the research on the discrepancies in surface characteristics as well as the resulting performances in lithium-ion storage is scarcely concerned.[13,14]

Meanwhile, rational design and construction of the morphology of electrode materials (especially inorganic materials) have been evidenced as an effective strategy to further enhance the electrochemical performances in LIBs. Among these designed strategies for morphologies, nanofibers decorated with pores have been paid tremendous attention because of the following merits: (1) the porous structures with exposed higher surface areas and abundant active sites will provide electrodes with a larger specific capacity and (2) the one-dimensional (1D) nanostructure will short the diffusion distance of Li+ ions and speed up the transfer of electrons, eventually enhancing the rate capability and prolonging the cycling life of devices.[13−15]

Based on the above considerations, we develop an integrated strategy, which combines the template-assisted electrospinning technology with the subsequent heat treatment to fabricate porous PAN nanofibers with surface functionalities (Scheme ). Benefiting from the abundant oxygen functionalities and porous structure, the as-obtained porous PAN nanofibers present a competitive capacity, an excellent rate capability, and an extended cycling life span (418 mA h g–1 at 50 mA g–1 after 300 cycles), compared with the original PAN precursors.

Scheme 1

Schematic Illustration of the Formation Process of Porous PAN/PEG-X-Air Nanofibers

Results and Discussion

The facile synthesis procedure is illustrated in Scheme . First, a homogeneous dimethylformamide (DMF) solution containing PAN and polyethyleneglycol (PEG) with different mass ratios was electrospun to form nanofibers. After that, these PAN/PEG nanofibers were subjected to a heat treatment under different atmospheres (Air: PAN/PEG-Air, N2: PAN/PEG-N2). During this thermal treatment, PAN was transformed into a different ladder-shaped conductive polymer in air or nitrogen atmosphere as verified in previous literature studies.[16,17] In the meantime, the decomposition and phase transformation of PEG lead to the formation of a porous and cross-linking structure.[18,19]

To elucidate the thermal stabilities of PAN and PEG under different atmospheres, thermogravimetric analysis (TGA) was conducted, as shown in Figure S1 (Supporting Information). It is observed that PEG presents a lower decomposition temperature than PAN. For example, under the air condition, PEG begins to lose weight at 240 °C. In contrast, PAN remains stable even at the temperature of 400 °C. Under the protection of N2, PEG and PAN will be more stable than those in air. These discrepancies in thermal stabilities inspire us to treat samples at 280 °C.

To investigate the morphology evolution, the scanning electron microscopy (SEM) images of the samples were taken. As shown in Figure S2a,b (Supporting Information), after electrospinning, both PAN and PAN/PEG-1 are present as uniform nanofibers with diameters of 200–400 nm. After being annealed in air or N2, the obtained PAN-Air, PAN-N2, PAN/PEG-1-Air, and PAN/PEG-1-N2 maintain their original morphologies as compared in Figures a,b and S2c,d (Supporting Information). Similarly, the fibrous structures are also observed for PAN/PEG-0.5-Air, PAN/PEG-2-Air, PAN/PEG-0.5-N2, and PAN/PEG-2-N2, as displayed in Figure S2e–h (Supporting Information). To further investigate their microstructures, transmission electron microscopy (TEM) measurements were conducted. As displayed in Figure c, PAN/PEG-1-Air exhibits a smooth surface, different from the rough surface of PAN/PEG-1-N2 in Figure d. Such discrepancies in surface morphology might be due to the difference in atmosphere for thermal treatment. In air, the attendance of oxygen speeds the cyclizing reaction, the dehydrogenation, as well as the further oxidation of PAN, resulting in the formation of the plied and interlaced structures with a smooth surface characteristic.[20] Differently, under the protection of nitrogen, the cyclizing and oxidation reactions are inhibited and the carbonization occurs dominantly, forming the rough surfaces and porous structures. The high-resolution TEM (HRTEM) images of PAN/PEG-1-Air and PAN/PEG-1-N2 imply their porous nature, as demonstrated in Figure e,f.

Figure 1

SEM, TEM, and HRTEM images of PAN/PEG-1-Air (a,c,e) and PAN/PEG-1-N2 (b,d,f).

To further reveal their porosities, nitrogen adsorption/desorption analysis was carried out, as illustrated in Figure S3 and Table S1 (Supporting Information). All of them display type IV sorption isotherm curves, confirming the abundant existence of mesopores in the structure (Figure S3a,c, Supporting Information). For example, the pore size distribution (PSD) of PAN/PEG-Air indicates that the diameters of pores mainly centered at 2–5 and 5–10 nm, implying a hierarchical structure (Figure S3b, Supporting Information). As summarized in Table S1, the heat treatment can adjust the porosities of samples under any atmosphere. Moreover, it is observed that the samples treated under N2 present higher surface areas than those samples treated in air. This discrepancy in structure implies the occurrence of different reactions, assisted with the TGA and TEM observation.

Furthermore, to investigate the elemental composition of the obtained samples, the element analyses were carried out. As shown in Table S2 (Supporting Information), the content of oxygen increases, when the samples are treated in air (PAN-Air and PAN/PEG-1-Air). Comparatively, when the samples are annealed under nitrogen protection, the carbon and nitrogen elements are dominant. Finally, a uniform elemental distribution in PAN/PEG-1-Air is further elucidated via SEM and the correlated elemental energy-dispersive X-ray (EDX) mapping (Figure S2i–l, Supporting Information).

To further study the surface functionalities of these samples, the Fourier transform infrared spectroscopy (FTIR) spectra were recorded, as shown in Figure . For the original PAN fibers, the peaks located at 2930, 2240, and 1451 cm–1 can be ascribed to the stretching vibration of −CH2–, stretching vibration of the −C≡N group, and bending vibration of −CH2–, respectively.[21] After annealing in air (PAN-Air and PAN/PEG-1 Air), the strength of the stretching vibration of −C≡N (2240 cm–1) weakens and the vibration of −C=N at 1590 cm–1 appears, implying the formation of a ladder-shaped conductive polymer.[17−20] Impressively, the overlapped absorption peaks at the interval of 1750–1520 cm–1 can be fitted into three individual peaks, which can be ascribed to the symmetric stretching vibration of C=O (1670 cm–1), C=C (1620 cm–1), and C=N (1590 cm–1), respectively. Besides, two characteristic peaks of C–H (1370 cm–1) and C=C (810 cm–1) are also observed, indicating the cyclization. Similarly, the characteristic peaks of cyclization C=N (1590 cm–1), C–H (1378 cm–1, 1260 cm–1), and C–C (1060 cm–1) are also found for PAN-N2 and PAN/PEG-1-N2. However, the representative peak of C=O (1670 cm–1) and C=C (1620 cm–1) is negligible, indicating the difficulty of the dehydrogenation and oxidation under such inert atmosphere. A weak peak at 1113 cm–1 assigned to the −CH2–O–CH2– symmetric stretching vibration and two characteristic peaks at 949 and 844 cm–1 corresponding to the −CH2–O–CH2– in-plane deformation vibration are observed in PAN/PEG-1-N2, which can be attributed to the residue of PEG within the framework under an inert atmosphere. In short, these differences in spectra discover both the variation in surface functionalities and the occurrence of different pyrolytic reactions.[17−23]

Figure 2

FTIR spectra of PAN, PAN-Air, PAN/PEG-1-Air, PAN-N2, and PAN/PEG-1-N2.

To further verify the composition variations between PAN/PEG-1-Air and PAN/PEG-1-N2, the X-ray photoelectron spectroscopy (XPS) measurements were adopted, as demonstrated in Figure . Three peaks are indexed to C 1s (285.3 eV), N 1s (399.3 eV), and O 1s (532 eV), respectively (Figure a,e). It can be noted that a higher percentage of oxygen in PAN/PEG-Air is found than that in PAN/PEG/-1-N2, accordant with the elemental analysis (Table S2, Supporting Information). After deconvolution, the high-resolution scan of C 1s can be de-convoluted into three different species: C=C (C–C, C–H) at 284.6 eV, C=N (C–N, C–O) at 286.5 eV, and C=O at 288.2 eV, respectively, as displayed in Figure b,f.[14] For the N 1s spectrum, the peaks can be fitted into two peaks at 398.7 eV for C=N and 399.9 eV for C–N, respectively, as demonstrated in Figure c,g.[21] More impressively, the high-resolution scan of the O 1s spectrum (Figure d,h) can be divided into two peaks at 531.1 eV for C=O and 533.3 eV for C–O, respectively. A weak peak of C=O in PAN/PEG-1-N2 might arouse from the residual solvent of DMF. Quantitatively, the content of C=O in PAN/PEG-1-Air (63.8%) is much higher than that of PAN/PEG-1-N2 (31.6%), consistent with the element analysis and FTIR data and account for the potential capability improvement in lithium-ion storage.

Figure 3

Wide XPS surveys and the high-resolution XPS scans of C 1s, N 1s, and O 1s for PAN/PEG-1-Air (a–d) and PAN/PEG-1-N2 (e–h).

To evaluate the electrochemical properties, the galvanostatic charge–discharge measurements were tested at a current density of 50 mA g–1 within a potential range of 0.01–3.0 V (vs Li/Li+). As depicted in Figure a, the PAN/PEG-1-Air sample presents insertion/extraction capacities of 1029/590, 492/488, 444/438, and 418/405 mA h g–1 in the 1st, 10th, 50th, and 100th cycles, with corresponding Coulombic efficiencies (CEs) of 57.3, 99.2, 98.5, and 99%, respectively. The irreversible capacity loss in the first cycle might be ascribed to the solid electrolyte interface (SEI) formation and some irreversible side reactions. After that, no significant capacity loss is observed in subsequent cycles even after 300 cycles, indicating an excellent reversible capacity and cycling stability of PAN/PEG-1-Air. In addition, the cycling performances of other samples are also investigated, as shown in Figures b and S4 (Supporting Information). In general, it is observed that the samples annealed in air (PAN-Air: 322 mA h g–1; PAN/PEG-0.5-Air: 359 mA h g–1; PAN/PEG-1-Air: 410 mA h g–1; PAN/PEG-2-Air: 371 mA h g–1) present higher capacities than other counterparts (PAN: 83 mA h g–1; PAN-N2: 64 mA h g–1; PAN/PEG-0.5-N2: 231 mA h g–1; PAN/PEG-1-N2: 211 mA h g–1; PAN/PEG-2-N2: 195 mA h g–1). Such enhancement in capacities might originate from the high density of oxygen-based functionalities, consistent with the FTIR and XPS analysis. In addition, it has to be noted that the porosities induced by the decomposition of PEG also facilitate the mass and electron transfer. In contrast, the samples treated in N2 depict inferior capacities, especially for those templated by PEG. Considering their similar porosities, it is deduced that the recession in capacities might result from the scarcity of oxygen-containing groups. In addition, the decreased capacities are also correlated with the incomplete decomposition of PEG under nitrogen, which cover the active sites for lithium storage, in accordance with FTIR observation. Finally, the discharge capacities of PAN-P-Air (237 mA h g–1) and PAN-P-N2 (155 mA h g–1) are also significantly lower than those of electrospun nanofibers, further proving that the morphology design is one of the main factors affecting the capacity. Moreover, the rate performances of the samples are also investigated (Figure c). Typically, PAN/PEG-1 shows the best rate capabilities among them (720, 493, 389, 311, and 254 mA h g–1) at the current densities of 20, 40, 60, 80, and 100 mA g–1, respectively. Furthermore, as summarized in Table , the PAN/PEG-1-Air nanofibers also demonstrate a fast and high accessibility for lithium intercalation/deintercalation than other reported materials.

Figure 4

(a) Discharging–charging curves of PAN/PEG-1-Air, (b) cycling performances of PAN/PEG-1-Air, PAN, PAN-Air, PAN-N2, and PAN/PEG-1-N2 samples as well as corresponding CE of PAN/PEG-1-Air at 50 mA g–1 for 300 cycles, (c) rate performance of these samples at various current densities, (d) Nyquist plots of these electrodes, (e) cyclic voltammograms of PAN/PEG-1-Air, and (f) cyclic voltammograms of PAN/PEG-1-N2 between 0 and 3 V at a scan rate of 0.1 mV s–1.

Table 1

Comparison of PAN/PEG-1-Air with Other Reported Carbonyl-Based Organic Polymers in Terms of Lithium-Ion Storage

carbonyl-based organics	rates or current density (mA g–1)	cycles	reversible Capacity (mA h g-1)	capacity retention %	reference number
Li-lawsone	0.5 C	1000	277	99 (3rd cycle)	(5)
LiDHAQS	1 C	500	220	73.3	(6)
PDHBQS	250	500	135	59	(24)
PSB	10	100	175	90	(25)
humate	100	100	491.9	91.6	(26)
PMMA	0.2 C	150	196.8	73.5	(27)
PDA	500	580	1414	93	(28)
PEDOT:PSS	0.1 C	200	266	93	(29)
VG 8/G	100	200	272	100	(30)
TPB	0.2 C	100	223.2	91.4	(31)
PAN/PEG-1-Air	50	100	418.5	85 (10th cycle)	our work

To explore the kinetics of the lithiation/delithiation process, the electrochemical impedance spectroscopies (EISs) were studied (Figure d). The Nyquist plots for these electrodes feature with the depressed semicircle in the high-to-medium frequency region, an inclined straight line in the low-frequency region. The intercept on the Z′ axis in the high-frequency region is ascribed to the electrolyte resistance (Re); the size of the semicircular reflects the charge-transfer resistance (Rct) in the electrode reaction; and the inclined line in the low-frequency range represents the Warburg impedance (Zw) related to lithium diffusion in the solid.[6,20] On the basis of the simulation (Table S3, Supporting Information), PAN/PEG-1-Air presents much smaller Rct and Zw values (166.6 and 138.4 Ω cm–2) than PAN (574.3 and 269.6 Ω cm–2), PAN-Air (236.4 and 169.5 Ω cm–2), PAN-N2 (288.3 and 198.7 Ω cm–2), and PAN/PEG-1-N2 (453.2 and 231.2 Ω cm–2). These superiorities might derive from the higher content of oxygen in the framework and shorter diffusion distance for Li+ ions and electron in structure, ensuring higher reversible capacity and better rate capability in performance. To reveal the electrochemical storage mechanism, cyclic voltammograms were recorded (Figure e). An irreversible peak at around 0.31 V in the initial cathodic sweep of PAN/PEG-1-Air might be attributed to the occurrence of side reactions. The cathodic peaks at 1.18 V can be ascribed to the reduction of the Li+-active functional groups in PAN/PEG-1-Air and the intercalation of Li+ ions, while the subsequent oxidation peaks located at 1.32 V can be ascribed to the oxidation of the Li+-active functional groups in PAN/PEG-1-Air and the deintercalation of Li+ ions. Differently, the coincided cyclic voltammetry (CV) curves with weakened and broadened peaks imply a reversible capacitive performance because of the formation of solid electrolyte interface (SEI) and the reversible insertion of lithium ions during the second and third cycle. Comparatively, the CV plots of PAN/PEG-1-N2 embrace a smaller area with a lower current density than that of PAN/PEG-1-Air, illustrating a lower specific capacity of PAN/PEG-1-N2 (Figure f).

To further probe the Li+ storage mechanism, ex situ FTIR spectra and XPS were recorded (Figure ). As recorded in Figure a, the peak at 862 cm–1 strengthened/weakened and positively/negatively shifted, implying the deformation and recovery of C=O in PAN/PEG-1-Air during the reversible accumulation and release process for Li+.[4,11] In addition, the appearance/disappearance of the peak at 1427 cm–1 also indicates the reversible formation of C–O–Li for the Li+-active group of C=O in the discharging/charging process.[32] However, no obvious variations which can be assigned to C=N groups are observed, indicating negligible contribution of C=N groups in skeleton. Moreover, as displayed in Figure b, different from the graphite-like characteristic peak at 284.4 eV, all the C 1s spectra demonstrate a shift to high binding energy area, indicating the occurrence of crosslinking reactions with abundant residual oxygen containing functionalities in samples. The C 1s spectra of the pristine electrode could be deconvoluted into four peaks, which might belong to C=C/C–C/C–H (284.6 eV), C=N/C–N/C–O (286.1 eV), C=O (287.2 eV), and −CF2 (binder, 292.4 eV).[20−24] After Li+ intercalation in the discharging process, the full width at half-maximum (fwhm) of the main peak broadened from 1.2 to 2.3 eV, accompanied with the percentage increase of C–O–Li (286.1 eV) and the weakening of C=O (287.2 eV). Reversibly, the fwhm of the main peak recovered from 2.3 to 1.35 eV with the percentage decrease of C–O–Li (286.1 eV) and the increase of C=O (287.2 eV) during the delithiation process. In addition, the O 1s peak also varied during the lithium insertion/extraction process. Accompanied with the lithiation of PAN/PEG-1-Air, the peak of O 1s shifted negatively toward low binding energy area, suggesting a decrease in the electron density of oxygen atoms (Figure c).[33] In short, both the ex situ FTIR results and the ex situ XPS measurements evidence the primary contribution of C=O groups to the reversible storage of PAN/PEG-1-Air.

Figure 5

Ex situ FTIR (a) and the high-resolution XPS scans of PAN/PEG-1-Air electrodes during the charging–discharging process [(b): O 1s, (c): C1s].

Conclusions

In summary, the functionalized porous PAN nanofibers are successfully synthesized via an integration of electro-spinning technology and annealing treatment. More impressively, the samples calcined in air present appealing capacities and excellent rate capabilities with extended cycling life spans, compared with other counterparts (the original PAN nanofibers and those samples treated under a N2 atmosphere) and other references. It is evidenced that such improvements originate from the enriched large amount of C=O groups, which ensure a high specific capacity, and a 1D porous fibrous structure, which shortens the diffusion distance of Li+ ions and speeds up the transfer of electrons, eventually enhancing the rate capability and prolonging the cycling life of devices. It is envisioned that this work will push forward the functionalization of organic polymers to enhance the performance of electrochemical energy storage via surface engineering and morphology control.

Experimental Section

Synthesis of Materials

The synthesis of porous PAN nanofibers was conducted via an electrospinning technique, followed by heat treatment under different atmospheres (Scheme ) Typically, 0.8 g of PAN (average Mw 150,000) and a certain amount of PEG (average Mw 2000) (0.4, 0.8, and 1.6 g) were mixed in 9 mL of DMF to form a homogeneous solution. Then, the solution was transferred into an injector to electro-spin the fibers. After that, the obtained fibers were annealed under air or nitrogen at 280 °C for 2 h. According to the mass ratio between PAN and PEG and the atmosphere used, the obtained samples were named PAN/PEG-X-Air (N2) (X = 0.5: PAN/PEG = 2:1, 1: PAN/PEG = 1:1, 2: PAN/PEG = 1:2). For comparison, the pure PAN nanofibers were also annealed in air or N2 with a similar protocol, which were named PAN-Air and PAN-N2, respectively.

Electrochemical Measurements

The electrochemical performances of samples were investigated by the assembly of the half-cell (CR2032), in which a certain amount of electrode materials (65 wt %), Super P (30 wt %), and PTFE (5 wt %) were homogeneously mixed and pressed on copper foils to form the working electrodes, and lithium foils work as both counter and reference electrodes. In addition, the electrolyte consists of 1 M LiPF6 in dimethyl carbonate /ethylene carbonate (1:1, v/v) and the separators are Celgard 2400 membranes. The galvanostatic charge–discharge measurement was performed in a voltage range from 3 to 0 V on a NEWARE battery testing system. The CV curves were recorded on a CHI 630D electrochemical analyzer in the voltage range 3–0 V at a scanning rate of 0.1 mV s–1. EIS was measured on a PARSTAT 2273 electrochemical work station in the frequency range from 100 kHz to 10 mHz.

Materials Characterization

TGA was performed on a PerkinElmer Diamond TGA4000 apparatus with a heating rate of 10°C min–1 under a flowing nitrogen or air. FTIR was performed on a Thermo Nicolet (Nicolet 5700). XPS analysis was measured on an ESCALAB250xi (Thermo Fisher Scientific), and the X-ray photoelectron spectrometer was equipped with an Al Ka achromatic X-ray source. The morphologies were observed using a scanning electron microscope (JSM 6701F) with element mapping by EDX. TEM images were observed using a FEI Tecnai G2 F20 s-twin field emission electron microscope operating at 200 kV accelerating voltage. Nitrogen adsorption and desorption isotherms were measured at 77 K using a Quadrachrome Adsorption Instrument. PSD curves were obtained from the gas-sorption measurement data by using the density functional theory method.