1 Introduction
Carbon nanotubes (CNTs) as superior matrix materials for functional nanostructures have been widely applied in various fields including sensors [
1–
2], catalysis [
3–
5] and electrochemical energy storage [
6–
8] due to their special features such as high aspect ratio, large surface area, high electrical conductivity, and strong mechanical strength. Typically, the hybridization of CNTs with functional nanoparticles (NPs) always results in improved performances by means of the homogenous dispersion of these active NPs and the high conductivity and flexibility to maintain the integrated stability of the hybrid nanostructures [
9–
15].
In electrochemical energy storage fields, lithium-ion batteries (LIBs), as one of the most popular power source, have dominated the market of portable electronic products, especially for laptops and cell phones [
16–
17]. However, with the advent of new high-power electricity products such as electric vehicles, commercial LIBs with graphite as the anode material have fallen behind their demand for rechargeable battery with high energy density and power density [
18]. Electrode materials play a key role in the performances of the batteries. Therefore, in order to improve the electrochemical performances, it is necessary to exploit more efficient anode materials for LIBs. Transition metal oxides with high theoretical specific capacity, which based on the lithium storage mechanism of reversible conversion reaction, have become the promising anode active materials for LIBs [
19–
22]. Among them, Fe
3O
4 has attracted more attention due to its higher theoretical specific capacity of 922 mA·h·g
−1, large natural abundance and low toxicity [
21,
23]. However, pristine Fe
3O
4 as anode material suffers large volume expansion and severe aggregation due to the reversible conversion reaction during the repeated lithiation/delithiation, which results in rapid decay of cycling capacity and poor rate performance [
24–
25].
Introducing CNTs as the matrix of metal oxides can not only alleviate the volumetric expansion, but also increase the electric conductivity of the whole electrode material, which is beneficial to maintain the stable cycling ability and high rate performance of the electrodes [
26–
30]. Although there have been some reports about Fe
3O
4/CNT composite as anodes for LIBs [
31–
33], there are still some deficiencies in material preparation methods rooting from the very inert surface of CNTs. To improve the hydrophilcity, the surface engineering seems unavoidable. Usually, the strategies that using mixed organic solvent systems and the addition of various surfactants are employed to decrease the surface tension [
34–
35], and the functionalization through acid treatment or graft modification is another effective route to enhance the interaction between CNTs and active matter-contained species [
8,
26,
33]. Even though prominent effects on the improvement of the association between active NPs and CNTs have been achieved, the massive usage of organic solvents, surfactants and corrosive acids will exert more pressure to environment obviously. Therefore, the development of facile and environmentally friendly methods for the preparation of active NP/CNT composites is very necessary.
To avoid the unfavorable affects which have to be faced in wet chemistry routes, in this study, we report a facile and green gas-phase strategy basing on the sublimation-pyrolysis behavior of ferrecene. With the experimental design that confining the mixture of ferrocene and CNTs in a closed reactor, and controlling the temperature-rising process, a carbon-coated Fe3O4 nanoparticle/carbon nanotube (Fe3O4@C/CNT) composite is received, and ultrafine dimension and homogenous dispersion of Fe3O4 NPs on CNTs are realized simultaneously during the one-step reaction process. Benefiting from the solution-free synthetic strategy, the one-dimensional (1D) CNTs intertwine together to form a three-dimensional (3D) network structure. The special porous configuration is very helpful for convenient electrolyte diffusion and short Li-ion transport pathways. As a result, the Fe3O4@C/CNT composite exhibits substantially enhanced electrochemical performances, especially the excellent rate capability (rate capacity retention reaches to 54.5% at 6000 mA·g−1) and ultra-stable cycling ability. In addition, the relationship between microstructure and electrochemical performances, as well as the kinetic behavior are investigated to well understand the advantages of this facilely constructed Fe3O4@C/CNT hybrid structure.
2 Experimental
2.1 Preparation of the Fe3O4@C/CNT composite
The Fe3O4@C/CNT composite was prepared by a pyrolysis of the mixture of CNTs and ferrecene in an enclosed reactor and the following heat treatment. The details of the experimental process are as follows. Firstly, 0.05 g of CNT powder and 0.20 g of ferrocene were placed in a 100 mL steel reactor. The reactor was then caped tightly and placed into a muffle furnace. After rising the temperature to 450 °C at a slow rate of 3 °C·min−1 and maintaining for 3 h, the as-received black powder was heat-treated at 700 °C for 2 h under Ar atmosphere to removal the impurity and improve the crystallinity of Fe3O4 phase. Finally, the Fe3O4@C/CNT composite was obtained.
2.2 Materials characterization
X-ray diffraction (XRD) (Bruker D8 Advance) and Raman spectroscopy (Renishaw RM2000 Confocal) measurements were applied to confirm the phase composition and structure of the samples. Scanning electron microscopy (SEM) (JSM-7001F, JEOL) with energy dispersive X-ray spectroscopy (EDS) and high-resolution transmission electron microscopy (HRTEM) (JEM 2010, JEOL) were used to observe the morphology and microstructure of the samples. The thermogravimetric (TG) analysis (TG/DTA 6300) was applied to determine the content of Fe3O4 in the as-prepared Fe3O4@C/CNT composite.
2.3 Electrochemical measurements
The electrochemical performances were measured by assembling CR2016 coin-type cells in a glove box filled with Ar gas. The working electrode was prepared by spreading the mixture of the Fe3O4@C/CNT composite, carbon black and polyvinylidene fluoride (PVDF) with the weight ratio of 8:1:1 on Cu foil. The average loading of active materials on the electrode slices is about 0.84 mg·cm−2. The counter electrode was pristine lithium metal foil. The separator was polypropylene film with micropores. The electrolyte was a 1.0 mol·L−1 LiPF6 dissolved mixture of ethylene carbonate and dimethyl carbonate (1:1, v/v) with vinylene carbonate (VC) (5 vol.% VC) as an additive. The galvanostatic discharge/charge tests were performed on a CT2001A Battery Testing System (LAND, Wuhan) in the voltage range of 0.01–3.0 V (vs. Li/Li+) at room temperature. The cyclic voltammetry (CV) profiles and electrochemical impendence spectroscopy (EIS) results were obtained by using a CHI 660E electrochemical workstation with a scan rate of 0.25 mV·s−1 and a frequency range of 100 kHz to 0.01 Hz.
3 Results and discussion
Ferrocene as an organo-transition metal compound with aromatic property exhibits an interesting physicochemical feature that begins to sublimate and pyrolyze at 100 and 400 °C above, respectively. Its pyrolysis products are carbon and Fe
0 or Fe
xO
y which mainly depend on the content of oxygen in surrounding atmosphere [
36–
37]. So it is feasible to load iron oxide NPs on the surface of CNTs by controlling the sublimation-pyrolysis behavior of ferrocene through a careful temperature-elevating process and the reaction atmosphere adjustment. Figure 1 illustrates the particular preparation process of the Fe
3O
4@C/CNT composite. In the airtight reactor with a certain amount of air, ferrocene powder sublimates and gradually diffuses among the whole space of reactor along with the temperature raises from room temperature to around 400 °C. When the temperature reaches 400 °C above, the gas-phase ferrocene begins to decompose, accompanying by the generation of carbon-coated Fe
3O
4 NPs due to the insufficient oxygen supply [
38]. During the pyrolysis and oxidation process of ferrocene, Fe
3O
4 NPs are expected to be anchored on the surface of CNTs to form the Fe
3O
4@C/CNT composite.
Fig.1 Schematic illustration of the construction process for the Fe3O4@C/CNT composite. |
Full size|PPT slide
The XRD patterns of the as-prepared Fe
3O
4@C/CNT composite and the pristine CNTs in the 2
θ range of 10°–80° are shown in Fig. 2(a). The diffraction profile for the Fe
3O
4@C/CNT composite is similar to that of the pristine CNTs, and the additional diffraction peaks at 18.3°, 30.5°, 35.5°, 37.1°, 42.9°, 65.7°, 70.9° and 73.9° can be indexed to (1 1 1), (2 2 0), (3 1 1), (2 2 2), (4 0 0), (5 3 1), (6 2 0) and (5 3 3) planes of Fe
3O
4 cubic crystals (JCPDS No. 19-0629), respectively, indicating the obtained composite contains both the features of CNTs and Fe
3O
4 phase. In addition, the broad peaks belonging to Fe
3O
4 diffraction implies the ultrafine dimension of the formed Fe
3O
4 NPs [
39]. Figure 2(b) shows Raman spectra of the pristine CNTs and Fe
3O
4@C/CNT composite. The two samples display similar Raman curves that the distinctive D band at 1346 cm
−1 (disordered carbon) and G band at 1575 cm
−1 (ordered graphitic carbon) are observed [
32,
40]. However, the characteristics for the Fe
3O
4 phase in the low Raman shift region are not detected, which could be due to the homogenous dispersion and amorphous texture of Fe
3O
4 NPs. The intensity ratio between D-band and G-band (
ID/
IG) is enhanced relative to that of pristine CNTs, suggesting the decreased degree of graphitization due to the formation of an amorphous carbon coating on the surface of Fe
3O
4 NPs.
Fig.2 (a) XRD patterns and (b) Raman spectra of the Fe3O4@C/CNT composite and the pristine CNTs. |
Full size|PPT slide
In order to investigate the morphology and microstructure of the Fe3O4@C/CNT composite, SEM and TEM observations are carried out. As shown in Fig. 3(a), it is found that the CNTs form a 3D network by interweaving with each other, which is beneficial to fast electron/ion transformation and convenient electrolyte contact for active matter. It is noted that no obvious agglomeration can be identified apart from a few bright dots that adhere to the surface of CNTs (Figs. 3(b) and 3(c)). To reveal the detailed composition of the distinct particles (bright dots), the EDS analyses in different regions are performed. The result (Fig. 4(a)) reflects that both of the bright dot region (denoted as A) and the CNT region (denoted as B) are composed of the elements of Fe, C and O, and the respective content of these elements in the two regions is nearly consistent as well, implying the homogeneous dispersion of Fe3O4 NPs on the surface of CNTs. The favorable dispersity of the Fe3O4 NPs is also confirmed by SEM mapping measurement (Fig. 4(b)), in which the regions belonging to Fe and O elements overlap well with a uniform distribution on the CNT region. The TEM observation with different magnifications reveals more information about the microstructure feature of the Fe3O4@C/CNT composite. As shown in Figs. 3(d)–3(f), it can be intuitively observed that the Fe3O4 NPs with a size of less than 10 nm are firmly anchored on the surface of CNTs and coated by a very thin carbon coating, which are beneficial to alleviate the volume expansion effectively during the lithiation/delithiation processes to great extent. In addition, the HRTEM image (Fig. 3(f)) manifests that the lattice fringes belonging to the Fe3O4 NPs are very vague. It is indicative of the poor crystallinity of the Fe3O4 NPs, and this observation is consistent with the results of XRD and Raman measurements above.
Fig.3 (a)(b)(c) SEM, (d)(e) TEM and (f) HRTEM images of the Fe3O4@C/CNT composite with different magnification. |
Full size|PPT slide
Fig.4 (a) EDS and (b) SEM mapping of the Fe3O4@C/CNT composite. |
Full size|PPT slide
The relative contents of Fe
3O
4 and CNT in the Fe
3O
4@C/CNT composite are determined by the TG analysis shown in Fig. 5. One can find that the initial decomposition temperature of the pristine CNTs is ~520 °C, while that for the Fe
3O
4@C/CNT composite decreases to around 355 °C. The distinct difference is mainly due to the formation of the amorphous carbon coating on the surface of Fe
3O
4 NPs, since the amorphous carbon is easier to be oxidized in air than the graphitic carbon [
41]. The TG curve of the Fe
3O
4@C/CNT composite shows that the weight loss mainly occurs in the range of 400–600 °C, and the weight remains unchanged at 24% after the temperature is above 600 °C, which corresponds to the weight of the residual Fe
2O
3 phase resulting from the oxidation of Fe
3O
4 in air. So it can be calculated out that the weight percentage of Fe
3O
4 in the Fe
3O
4@C/CNT composite is about 24.8%.
Fig.5 The TG curve of the Fe3O4@C/CNT composite. |
Full size|PPT slide
The electrochemical performance of the Fe
3O
4@C/CNT electrode and the pristine CNT counterpart electrode are evaluated in CR2016 coin-type half cells with lithium foil as the counter electrode. Figures 6(a) and 6(b) show CV curves of the two electrodes in the voltage range of 0– 3.0 V. It is found that the Fe
3O
4@C/CNT electrode displays a similar profile to that of the pristine CNT electrode. The analogous broad cathodic peaks at around 1.5 and 0.7 V appear in the first cycle, which are ascribed to the irreversible side reactions between Li
+ and electrolyte molecules and the formation of a solid electrolyte interface (SEI) film. It also reflects the similar chemical state of the surface of CNTs before and after the loading of these carbon-coated Fe
3O
4 NPs. In addition, the pair of reversible cathodic/anodic peaks at 0.02 V/0.25 V and the anodic peaks at 1.1 and ~2.4 V can be attributed to the insertion/extraction process of Li
+ upon graphite layers, as well as the possible reactions with defects and other active sites [
42]. In the subsequent CV scans of the Fe
3O
4@C/CNT electrode, it is noted that a new reduction peak that is absent in the corresponding curves of the pristine CNT electrode emerges at ~0.75 V, and is indicative of the reaction of Fe
3O
4 with Li
+: Fe
3O
4 + 8Li
+ + 8e
−→ 3Fe
0 + 4Li
2O. Correspondingly, the broad anodic peaks at ~1.62 and ~1.86 V are resulted from the stepwise oxidation process of Fe
0 to Fe
2+ and Fe
3+, respectively [
43].
Fig.6 CV curves of (a) the pristine CNT electrode and (b) the Fe3O4@C/CNT electrode. Galvanostatic discharge/charge curves of (c) the pristine CNT electrode and (d) the Fe3O4@C/CNT electrode. |
Full size|PPT slide
The typical discharge/charge profiles of the Fe
3O
4@C/CNT and pristine CNT electrodes at 100 mA·g
−1 are presented in Figs. 6(c) and 6(d). No obvious discharge and charge plateaus are observed for the pristine CNT electrode, and the profiles are consistent with the results reported else [
32]. In contrast, the Fe
3O
4@C/CNT electrode exhibits very different electrochemical behavior that the gentle discharge slope at ~0.8 V and the steep charge plateau in the range from 1.5 to 1.9 V reflect the redox reactions between Fe
3O
4 and Li
+, which is in accordance with the CV results. It also can be seen that the CV and discharge/charge profiles overlap very well since the third cycle, indicating that the stable electrochemical redox reversibility is quickly built up after the initial lithiation-induced structural rearrangement and the stable SEI film formation. Benefiting from the rational hybridization of CNT with Fe
3O
4 NPs, the Fe
3O
4@C/CNT electrode releases high initial discharge and charge capacities of 971 and 576 mA·h·g
−1, respectively, resulting in an initial Coulombic efficiency of 61.4%. The large initial capacity loss of 395 mA·h·g
−1 is due to the irreversible decomposition of electrolyte and the interface reactions with Li
+ to form the SEI layer [
44]. In contrast, the pristine CNT electrode only relieves a charge capacity of 204 mA·h·g
−1 with a much lower initial Coulombic efficiency of 41.1%. These results signify that the gas-phase process during the loading of Fe
3O
4 NPs improves the surface state of electrode material and avoid the occurrence of some irreversible side reactions with Li
+.
The cycling performances of the Fe
3O
4@C/CNT electrode and the pristine CNT electrode are shown in Fig. 7(a). For the pristine CNT electrode, even though the reversible capacity is low, the cycling stability is excellent without any evident capacity decay upon cycling, demonstrating its suitability when used as a matrix material. After the combination with Fe
3O
4 NPs, the reversible capacity elevates greatly owe to the high theoretical specific capacity of the Fe
3O
4 active phase. With the discharge/charge cycling, it presents slight capacity decay in the first four cycles. The initial capacity decay mainly results from the partial degradation of unstable microstructure units and the gradual formation of the SEI film [
45]. Afterwards, a gradual capacity increase appears in the following cycles. This phenomenon is very common in transition metal oxide based anode materials, for which however the mechanism is still not very clear, and the possible reasons can be mainly ascribed to the material pulverization induced active site increase [
46], the interfacial Li-ion storage and the polymeric/gel-like film formation with a so-called “pseudocapacitance-type” effect during the lithiation process [
44,
47]. After 100 cycles at a current density of 100 mA·g
−1, a high charge capacity of 861 mA·h·g
−1 (0.72 mA·h·cm
−2 for areal specific capacity) is received, which is 49.5% higher than that of the first cycle, and a stable Coulomibic efficiency of 98% above also can be kept with the continuous lithiation/delithiation processes. The excellent cycling performance would stem from the microstructure advantage of the Fe
3O
4@C/CNT electrode. The CNT matrix with robust mechanical strength and flexibility effectively buffers the volume fluctuation and provides convenient charge transfer pathway, and the firm association with matrix and homegeous dispersion of the carbon-coated ultra-small Fe
3O
4 NPs further alleviate the volume effect and are beneficial to the fast e
−/Li
+ transportation. To verify the structural stability of the Fe
3O
4@C/CNT composite, the post morphology observation for the tested electrode is carried out. As can be seen from Fig. 8, the cycled electrode still maintains the active NPs/CNT configuration with the coverage of a uniform SEI film. It seems that, even though the partial pulverization of the active NPs occurs after suffering the cycling process, these fragments still firmly anchor on the surface of CNTs, and the integrated hybrid structure are well maintained without obvious collapse or destruction, which guarantees the stable electrochemical performance.
Fig.7 (a) Cycling performances at a current density of 100 mA·g−1, (b) rate capability, (c) rate capacity retention comparison (based on the value at 100 mA·g−1) and (d) Nyquist plots of the Fe3O4@C/CNT electrode and the pristine CNT electrode. |
Full size|PPT slide
Fig.8 TEM images of the Fe3O4@C/CNT electrode after 100 cycles. |
Full size|PPT slide
The rate capability of the Fe3O4@C/CNT electrode is also evaluated at a current density range from 100 to 6000 mA·g−1 and compared with the pristine CNT electrode, as shown in Fig. 7(b). With the stepwise increase of current density from 100 to 200, 400, 1000, 2000 and 6000 mA·g−1, high reversible capacities of 434, 395, 366, 330 and 260 mA·h·g−1 are released, respectively. More remarkably, when the current density is reset back to 100 mA·g−1, a high capacity of 682 mA·h·g−1 is received immediately, and then gradually increases to 823 mA·h·g−1 after 10 cycles. These results indicate the excellent rate ability of the Fe3O4@C/CNT electrode benefiting from its robust hybrid structure and porous feature to well withstand the drastic lithiation/delithiation and facilitate the convenient contact of active matter to electrolyte and the fast ion diffusion. The rate capacity retention comparison before and after the loading of carbon-coated Fe3O4 NPs on the CNT matrix is shown in Fig. 7(c). It can be found that the hybridization between Fe3O4 NPs and CNT results in a favorable synergistic effect on the enhancement of rate performance. Typically, the Fe3O4@C/CNT electrode still maintains high rate capacity retention of 54.5% even under a large current density of 6000 mA·g−1, while the value for the pristine CNT electrode is only 23.9%. The prominent elevation in both reversible capacity and rate performance after the combination of CNTs and Fe3O4 NPs is very attractive for the applications in high-power LIBs.
The enhancement of Li-ion diffusion kinetics and conductivity of the Fe
3O
4@C/CNT electrode is verified by the electrochemical impedance analysis. Figure 7(d) shows Nyquist plots of the pristine CNT electrode and the Fe
3O
4@C/CNT electrode before and after the initial several discharge/charge cycles. The impedance spectra profiles of both electrodes present two semi-circular arcs in high and medium frequency regions and a sloping line in the low frequency region, which correspond to the SEI film resistance, charge transfer resistance and lithium diffusion resistance within electrode host, respectively [
48–
49]. Compared with the pristine CNT electrode, the intercepts of the semicircles with the real axis for the Fe
3O
4@C/CNT electrode are much smaller, indicating lower SEI film resistance and faster charge transfer. It also can be found that, for the Fe
3O
4@C/CNT electrode, the electrode before cycling presents the smallest electric resistance, while that for the electrode after initial three cycles increases obviously due to the coverage of a SEI layer. Interestingly, it is noted that the further increase of cycle times even receives decreased impedance. This result demonstrates that the initial electrochemical activation can help to form a positive structural adjustment and gradually improved electrode surface state, and thus contributes to the enhanced electrochemical performances.
4 Conclusions
In summary, a new kind of Fe3O4@C/CNT composite has been developed through a facile gas-phase reaction route by means of the sublimation-pyrolysis behavior of ferrocene at low temperature. The ultrafine Fe3O4 NPs (<10 nm in dimension) with a thin carbon coating are found to be homogenously anchored on the surface of CNTs, resulting in greatly elevated Li-storage capacity, decay-free cycling ability and excellent rate capability. Typically, a high rate capacity retention of 54.5% is still maintained at an ultrahigh current density of 6000 mA·g−1, and a stable reversible capacity of 861 mA·h·g−1 is received after suffering 100 lithiation/delithiation processes. The superior electrochemical performances benefit from the rational porous electrode structure design with the robust CNT network as the matrix and uniformly dispersed Fe3O4 NPs as the main active materials. Furthermore, the effective gas-phase strategy illustrated in this study provides a new idea for solving the dispersion issue of functional NPs on the high inert surface of traditional carbons (e.g. CNT, graphite and active carbon), and is universal to other functional NPs/carbon application occasions, such as catalysis, electromagnetic shielding, and sewage treatment.
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}