Yidan
Song
a,
Yuanrui
Gao
b,
Hongren
Rong
a,
Hao
Wen
a,
Yanyong
Sha
a,
Hanping
Zhang
a,
Hong-Jiang
Liu
*b and
Qi
Liu
*ac
aSchool of Petrochemical Engineering, Jiangsu Key Laboratory of Fine Petrochemical Engineering, Advanced Catalysis and Green Manufacturing Collaborative Innovation Center, Changzhou University, 1 Gehu Road, Changzhou, Jiangsu 213164, P. R. China. E-mail: liuqi62@163.com
bDepartment of Chemistry, College of Science, Shanghai University, No. 99 Shangda Road, Shanghai, 200444, P. R. China. E-mail: liuhj@shu.edu.cn
cState Key Laboratory of Coordination Chemistry, Nanjing University, Nanjing, Jiangsu 210093, China
First published on 11th January 2018
Through the covalent bonding between naphthalenediimide diamine (NDIDA) and graphene oxide (GO), we synthesize NDIDA-functionalized graphene oxide (NDIDA-GO). The as-synthesized NDIDA-GO is characterized by powder X-ray diffraction, Fourier transform infrared spectroscopy, thermogravimetric analysis, Raman spectroscopy, X-ray photoelectron spectroscopy, scanning electron microscopy, transmission electron microscopy, and the Brunauer–Emmett–Teller surface area analysis. As a cathode material for lithium-ion batteries, within a voltage window of 4.5–1.5 V, NDIDA-GO exhibits a high specific capacity, good cyclic stability, and rate capability, keeping a specific capacity of 240 mA h g−1 after 50 cycles at 50 mA g−1. This work provides an effective route for the development of high-performance organic-based cathode materials for lithium-ion batteries.
Over the past several decades, many organic compounds have been investigated as the cathode materials for their potential use in LIBs.12,20,21 Conductive organic polymers,22 organosulfur compounds,23 free radical compounds,24,25 and carbonyl compounds26,27 are four major classes. Among them, the carbonyl compounds are considered as the most promising candidate as the cathode materials for LIBs because of their fast kinetics and high capacity.28,29 Unfortunately, in most cases, these cathode materials usually exhibit a low cycling performance and fast capacity decay owing to their severe dissolution in the organic electrolyte and a poor electrical conductivity.30 To solve the dissolution problem, many strategies are proposed including polymerization of small molecule carbonyls,30,31 employing the solid-state electrolytes,32,33 and optimizing the molecular structure.34
Besides those strategies, constructing the carbon-supported organic composites is a good route, which may overcome the two main issues mentioned above simultaneously.35–38 Graphene and graphene oxide (GO)-based materials, as the carbon-based materials, have attracted a considerable interest within the scientific community due to their excellent electrical and mechanical properties, as well as their applications in many fields, such as LIBs, solar batteries, sodium-ion batteries, and the composite materials etc.39–44
Herein, we focus our attention on 1,4,5,8-naphthalenetetracarboxylic dianhydride (referred as NTCDA herein). As a polycarbonyl compound, NTCDA can be used as a suitable electrode for LIBs owing to its high theoretical specific capacity of 400 mA h g−1 (considering that NTCDA has four electrons to transfer).30 To overcome the dissolution problem of NTCDA and guarantee the efficient electron transfer in the electrochemical reactions, we designed and synthesized naphthalenediimide diamine-functionalized graphene oxide (NDIDA-GO) as a new organic cathode material for LIBs via the amide formation between NDIDA and GO (Fig. 1). Consequently, the resultant NDIDA-GO exhibited a higher specific capacity (240 mA h g−1 after 50 cycles) when compared to the LiFePO4 cathode material (with a theoretical capacity of 170 mA h g−1), excellent cyclability, and better rate capability. Importantly, this convenient and elegant method can be extended to assemble GO with many other multiple carbonyl compounds to find the higher capacity and better cyclability organic cathode materials to promote the development and application of the LIBs.
Fig. 2 (a) FT-IR spectra of NDIDA, GO, and NDIDA-GO. (b) XRD spectra, (c) Raman spectra, and (d) XPS spectra of GO and NDIDA-GO. High-resolution XPS spectra of C 1s (e) and O 1s (f) for NDIDA-GO. |
The functionalization of GO with NDIDA was further examined by Raman spectroscopy (Fig. 2c). As presented in the Raman spectra, GO exhibits the D band at 1348 cm−1 and the G band at 1602 cm−1. The D band corresponds to a breathing mode of A1g symmetry and the G band can be assigned to the E2g vibration mode of the sp2 C atoms. The Raman spectrum of NDIDA-GO shows the similar spectral feature with that of GO, which is the D band at 1333 cm−1 and the G band at 1598 cm−1. Meanwhile, the band-intensity ratio (ID/IG) increases from 0.87 to 1.15, indicating that the structure distortion arises from the covalent bonding interaction between GO and NDIDA.48,49 The shift of the Raman peak should be attributed to the charge transfer between GO and the newly grafted NDIDA moieties.49 The X-ray photoelectron spectroscopy (XPS) technique was applied to examine the elemental composition of the sample. Fig. 2d exhibits the XPS spectra of GO and NDIDA-GO for comparison. Based on the calculation, the ratio of the intensity of the C 1s peak to that of the O 1s peak in the spectrum of NDIDA-GO is larger than that for GO, also indicating the successful bonding between NDIDA and GO. Furthermore, the N content increases from 0.74 atomic% for GO to 4.27 atomic% for NDIDA-GO due to the amidation reaction between GO and NDIDA (Table S1†). The changes of the intensity of the N 1s peak intensities for NDIDA-GO and GO also further confirm this fact (Fig. S1†). The elemental composition analysis indicates that the contents of C, N, and O atoms in NDIDA-GO are 72.95, 4.27, and 22.77 atomic%, respectively (Table S1†). Based on these data, it can be concluded that there is one NDIDA group per 43 carbon atoms on a GO sheet. Accordingly, there is about 35.8 wt% of NDIDA in NDIDA-GO. The high-resolution XPS spectra of C 1s and O 1s for NDIDA-GO are shown in Fig. 2e and f, respectively. As seen from these figures, the spectrum of C 1s can be divided into five peaks located at 283.6, 285.3, 285.7, 286.4, and 287.3 eV, corresponding to C–C bonding (sp2 carbon), C–C bonding in the defect graphite lattice (sp3 carbon), C–N bonding, C–O bonding, and CO bonding, respectively.48,49 The spectrum of O 1s can be divided into two peaks centered at 531.7 and 530.9 eV, attributing to C–O bonding and CO bonding, respectively.48,49 Besides, according to the reported spectra, the binding energy of N 1s for NDIDA is 401.01 eV,47 while the binding energy of N 1s for NDIDA-GO is shifted to 398.76 eV, as shown in Fig. S1.† This fact further verifies the linking between NDIDA and GO. The formation of the amide groups can increase the electron density on the N atoms of NDIDA-GO, resulting in the decrease of the binding energy of N 1s.51 The similar shifts have been observed for the NDI-TFP polymer.47
TGA was used to further investigate the presence of the functional groups in NDIDA-GO. The TGA curves of GO, NDIDA, and NDIDA-GO are shown in Fig. S2.† It can be seen that GO shows a little mass loss of about 2.5% at the temperature under 100 °C and about 16.5% at the temperature lower than 200 °C. The two mass losses should be ascribed to the removal of the adsorbed water and the decomposition of the oxygen-containing functional groups on the GO structure, respectively.46 In comparison, the TGA curve of NDIDA-GO shows nearly 27% mass loss at the temperature lower than 200 °C, resulting from the removal of the adsorbed water and the pyrolysis of the NDIDA grafted onto GO and the oxygen-containing functional groups on GO. It is worthwhile to mention that the mass loss of NDIDA is lower than that of GO and NDIDA-GO at the temperature lower than 580 °C. This phenomenon may be related to the hydrogen bonding and π–π interactions between the NDIDA molecules, which results in the better thermal stability of NDIDA. After the reaction of NDIDA and GO, since no original hydrogen bonding and π–π interactions exist, the thermal stability of NDIDA-GO is lower than that of NDIDA at the temperature under 580 °C.
The microstructure and morphology of NDIDA-GO were investigated by field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM). The FESEM image of NDIDA-GO (Fig. 3a, S3a and b†) reveals that the NDIDA-GO sample is composed of a large number of nanosheets and each sheet has a wrinkled and folded structure, indicating that it also consists of many ultrathin sheets. Compared with the morphology of GO (Fig. S3c and d†), the morphology of NDIDA-GO shows no obvious change. From the TEM image of NDIDA-GO (Fig. 3b), a multilayer structure can be observed, further confirming that it is composed of many ultrathin sheets.
The surface area and porosity of NDIDA-GO were characterized by the nitrogen adsorption–desorption technique. The adsorption–desorption isotherm is shown in Fig. 4. The isotherm for NDIDA-GO exhibits a typical type-III sorption isotherm, indicating the existence of macropores. The surface area and total pore volume of the NDIDA-GO sample are 14.15 m2 g−1 and 0.026 cm3 g−1, respectively. The average pore diameter is calculated with the BJH method to be 24.42 nm. The macropores originated from the interspaces of the nanosheets. Thus, an abundant porous structure provides an excellent condition for the ion storage and diffusion. The surface area, total pore volume, and average pore diameter of the GO sample are 11.38 m2 g−1, 0.01 cm3 g−1, and 7.58 nm, respectively (Fig. S4†), less than those of the NDIDA-GO sample.
Fig. 4 Nitrogen adsorption–desorption isotherm for the synthesized NDIDA-GO. The inset: pore size distribution. |
The cycling performances of the NDIDA-GO, GO, and NDIDA electrodes were measured at 50 mA g−1 in the range of 1.5–4.5 V versus Li/Li+. As shown in Fig. 5b, the first discharge capacity of the NDIDA electrode is only 30.6 mA h g−1 and decreases to 3.8 mA h g−1 after 50 cycles, indicating that it has a poor Li storage behavior due to its low conductivity and solubility in the electrolyte. For GO, it displays the discharge capacity of 98 mA h g−1 after 50 cycles, meaning that it has a poorer Li storage performance. On the contrary, the NDIDA-GO electrode shows a good cycling stability along with a high capacity. The initial discharge capacity is 240 mA h g−1. After 5 cycles, the NDIDA-GO electrode shows a stable capacity. After 50 cycles, the discharge capacity of the electrode maintains at 240 mA h g−1, which is 100% of the discharge capacity in the first cycle.
The value is superior or comparable to that of some previously reported carbon-based cathodes,37,38 organic cathodes, and organic/carbon composite cathodes.35,52,53 The improvement of the electrochemical performances of the NDIDA-GO electrode should be attributed to its electrical conductivity improvement and insolubility in the electrolyte. The solubility experiment of the NDIDA electrode and the NDIDA-GO electrode verifies that NDIDA from the NDIDA electrode dissolves easily in the electrolyte, while NDIDA-GO is insoluble (Fig. S5†) due to the combination of NDIDA and GO via the covalent bonding. As shown in Fig. S6,† the high reversible capacity may mainly originate from the reversible reaction between the CO and epoxide functional groups in NDIDA-GO and Li+ ions,12,54 where the conjugated carbonyl groups from NDIDA joins a multistep reduction during the discharge process and then the as-formed alkoxide groups are reoxidized in the subsequent charge process, showing the potential of NDIDA-GO as a future cathode material in LIBs.
The battery was further cycled at different current densities to study the rate performance. As shown in Fig. 5c, the average reversible capacities of the NDIDA-GO electrode obtained at 100, 200, 500, and 1000 mA g−1 are about 180, 153, 112, and 101 mA h g−1, respectively. It is worth pointing out that the capacity can be returned to 181 mA h g−1 when the current density is back to 100 mA g−1 after such higher current cycling, revealing that the NDIDA-GO electrode has a better rate capability.
The cyclic voltammetry (CV) curves of the second cycle for the NDIDA-GO, GO, and NDIDA electrodes tested in the voltage range of 1.5–4.0 V at 1 mV s−1 are shown in Fig. 5d. Compared with NDIDA and GO, NDIDA-GO presents a significantly improved gravimetric current, indicating that the electrical conductivity of NDIDA-GO is higher than that of NDIDA and GO. This result should be ascribed to the covalent combination between NDIDA and GO via the amide formation and the possible partial reduction of GO in the process of the amidation reaction.50 The removal of some oxygen-containing functional groups from GO can result in the electrical conductivity increase of NDIDA-GO. Evidently, two pairs of the redox peaks can be observed in the CV curve of the first cycle of NDIDA (Fig. S7†), corresponding to the lithiation and delithiation of two conjugated carbonyl groups.55,56 But in the CV curve of the second cycle, only a pair of the redox peaks can be clearly observed (Fig. 5d). From the CV curve of NDIDA-GO, we only see a pair of the relatively broad redox peaks at around 2.4 and 2.70 V. The formation of the broad peaks is relative to the multistep reaction processes,57,58 which is also confirmed by the charge/discharge curves without obvious plateaus of NDIDA-GO (Fig. 5a). In addition, from the CV curve of GO, the redox peaks are not seen (Fig. 5d). Fig. S8† is the CV curves of the NDIDA-GO electrode from the first to the fourth cycle. Obviously, the cycles of CV for the NDIDA-GO electrodes show the similar features during lithiation and delithiation after the second cycle. This means that NDIDA-GO has the excellent cycle stability. Because the NDIDA-GO electrode has the wide voltage window as well as the rapid charge–discharge curves without obvious plateaus, NDIDA-GO may be used as an electrode material in lithium-ion hybrid supercapacitors.59,60
Electrochemical impedance spectroscopy (EIS) of the NDIDA-GO electrode was performed in the frequency range from 100 Hz to 0.01 Hz for investigating the kinetics of electrochemical reactions on the electrode.
As shown in Fig. 6, each Nyquist plot is composed of two partially overlapped semicircles and a straight line in the low-frequency range. The semicircle in the high-frequency range should be assigned to the formation of the SEI film (Rs) and the other semicircle can be ascribed to the charge-transfer process (Rct). An equivalent circuit model is presented in the inset of Fig. 6. Here, R1 stands for the internal resistance containing the battery component resistance and the electrolyte solution resistance, Rs represents the SEI passivating film resistance, Rct is the charge transfer resistance, while W, Q1, and Q2 represent the Warburg impedance and CPE constant phase elements, respectively.61,62 For a better comparison, the simulated values of EIS obtained by Zview software are displayed in Table 1. It can be observed that all R1 values after 1 and 50 cycles are small, 6.3 and 3.1 Ω, respectively; the Rs and Rct values after 50 cycles are 81.4 and 52.8 Ω, respectively, higher than those corresponding values after 1 cycle. This result may be ascribed to the loss of some active materials on the current collector in the process of charge–discharge.63 The phase angles after 1 and 50 cycles at the low-frequency range are all larger than 45°, showing that the Li+ ions have better mobility.57 For comparison, the Nyquist plot of the NDIDA electrode after 1 cycle is presented in Fig. S9.† As seen in Fig. S9,† only a large semicircle appears in the Nyquist plot, which is attributed to the combination of Rs and Rct (Rs+ct). The Rs+ct value of the NDIDA electrode is 4366 Ω, which is much higher than the Rs and Rct values of the NDIDA-GO electrode, indicating that the migration kinetics of Li+ ions through the SEI film is enhanced and the velocity of the charge transfer reaction in the NDIDA-GO electrode becomes faster compared to those with the NDIDA electrode. As shown in Fig. S10,† the R1, Rs, and Rct values of the GO electrode are also larger than those of the NDIDA-GO electrode, meaning that the migration ability of Li+ ions in the NDIDA-GO electrode is also stronger than that in the GO electrode.
Sample | R 1 | R s | CPE1-T | CPE1-P | W 1-R | R ct | CPE2-T | CPE2-P | W 2-R |
---|---|---|---|---|---|---|---|---|---|
After 1 cycle | 6.3 | 35.9 | 4.550 × 10−3 | 0.46 | 0.92 | 17.3 | 2.5 × 10−4 | 0.62 | 0.62 |
After 50 cycles | 3.1 | 81.4 | 2.700 × 10−5 | 0.67 | 261 | 52.8 | 456.8 | 5.28 | 1.00 × 10−2 |
Footnote |
† Electronic supplementary information (ESI) available: The XPS spectra of N 1s for NDIDA-GO and GO, TGA curves of GO, NDIDA, and NDIDA-GO, FESEM images of NDIDA-GO and GO, N2 adsorption–desorption isotherm for GO, CV curves of NDIDA and NDIDA-GO, Nyquist plot of the NDIDA and proposed electrochemical Li storage mechanism for the NDIDA-GO electrode. See DOI: 10.1039/c7se00543a |
This journal is © The Royal Society of Chemistry 2018 |