Guangda
Li
,
Liqiang
Xu
*,
Qin
Hao
,
Meng
Wang
and
Yitai
Qian
Key Laboratory of Colloid and Interface Chemistry, Shandong University, Ministry of Education, School of Chemistry and Chemical Engineering, Jinan, 250100, P. R. China. E-mail: xulq@sdu.edu.cn; Fax: +86 531 88366280; Tel: +86 531 88364543
First published on 2nd November 2011
Carbon nanocages (CNCs) with diameters of about 200∼500 nm have been synthesized by a simple method. The electrochemical properties of the CNCs as anode materials were evaluated by discharge/charge measurement and electrochemical impedance spectroscopy. Results showed that the CNCs displayed excellent cycling performance and good rate capability with no noticeable capacity fading up to 50 cycles at current densities of 100, 300 and 500 mA g−1. Their electrochemical properties were significantly improved after annealing treatment of the CNCs at 600 °C. For example, the CNCs(2)-annealed exhibited much better electrochemical performance with a high reversible capacity of 520 mAh g−1 after 50 cycles at current density of 100 mA g−1. A discharge capacity of 380 mAh g−1 can be obtained after 50 cycles at high current density of 500 mA g−1. The results of the Raman, thermal gravimetric and electrochemical impedance analysis indicated that the graphitization degree, electronic conductivity and charge-transfer rate of the CNCs have been improved after annealing treatment.
In the last few decades, carbon nanomaterials such as carbon nanotubes, carbon nanofibers, graphene and composites based on these nanostructures have been widely investigated because nanoscale carbon materials have a relatively large number of lithium ion insertion sites on their surface, and charge-transfer resistances at the interface between the electrolyte and electrode materials are expected to be small.5–13 These materials exhibit improved storage capacity but most anodes of pure carbon nanomaterials still suffer from the capacity decay in the discharge/charge cycles.
At present, few reports focus on the 3D hollow carbon nanomaterials such as hollow carbon nanospheres and hollow carbon nanocages (CNCs) in lithium-ion batteries. 3D carbon nanomaterials can be readily used for improving the rate capability of the lithium-ion batteries because the solid-state diffusion length is short and their relatively large surface area can also benefit the charge-transfer rate. In particular, 3D hollow carbon nanomaterials provide several advantages for applications in lithium-ion batteries: electrode materials with hollow structures can accommodate lager volume changes during the discharge/charge process, and hollow structures can provide extra space for Li ion storage.14,15 Tien et al.16 synthesized uniform carbon spheres and the as-prepared anode exhibited a stable capacity above 330 mAh g−1 after 130 cycles. Lee et al.17 synthesized a 3D porous carbon material by a sol–gel process. The porous carbon electrode showed a superior rate capability compared to spherical carbon and bulk carbon. Wang et al.11 reported that the CNCs/graphene have been prepared by catalytic decomposition of xylene on MgO supported Co and Mo catalyst in supercritical CO2 at a pressure of 10.34 MPa and temperature ranging from 650 to 750 °C. The reversible capacity of the CNCs could achieve 574 mAh g−1 at a current density of 100 mA g−1 after 60 discharge/charge cycles. However, fabricating 3D hollow carbon nanomaterials through a simple method to take advantages and restrain shortcomings is a challenge work for developing high-performance lithium-ion batteries.
In this study, we report a simple strategy to synthesize high-yield CNCs and Fe3O4/CNCs as anode materials for high-performance lithium-ion batteries. The as-obtained CNCs have diameters of 200∼500 nm. The CNCs and Fe3O4/CNCs display excellent cycling performance and good rate capacity. It is found that their electrochemical properties can be significantly improved after the annealing treatment at 600 °C for 10 h. To the best of our knowledge, there are few research reports about the CNCs for lithium-ion batteries.
Ethanol (15 ml) and ferrous oxalate (1.5 g) were added into a stainless-steel autoclave of 20 ml capacity. Other reaction parameters were similar with above mentioned processes. The obtained samples in the autoclave were washed with absolute ethanol and distilled water for several times. After that, the Fe3O4/CNCs composite materials were obtained. These composites were washed with hydrochloric acid for several times, and then CNCs also could be obtained. These samples were denoted as CNCs(2) and Fe3O4/CNCs(2). The detailed experiment process and growth mechanism of this sample can be found in our previous report.18
The as-obtained samples (CNCs(1), CNCs(2), and Fe3O4/CNCs(2)) were annealed at 600 °C in a tubular furnace under following argon atmosphere for 10 h. After this step, these annealed samples are denoted as CNCs(1)-annealed, CNCs(2)-annealed, and Fe3O4/CNCs(2)-annealed, respectively.
![]() | ||
Fig. 1 (a) XRD patterns of the CNCs(1) before treated with hydrochloric acid and (b) after being washed with hydrochloric acid. |
The morphology of the as-prepared samples is further characterized by SEM and TEM. The SEM and TEM images in Fig. 2a and b show hollow polygonal morphology of the CNCs(1) which indicate that large quantity of the CNCs(1) were obtained by this simple method. Most of the CNCs(1) have diameters ranging from 200 to 500 nm and the wall thickness is about 70 nm. The wall of a carbon nanocage is composed of amorphous or disordered structure through HRTEM observation (see Fig. 2c). The disordered character is also detected by the corresponding XRD pattern. Fig. 2d and e show the TEM images of the samples with different treatment process. It can be clearly seen that the solid nickel particles are encapsulated in the graphite layers, which clearly evidenced the growth mechanism of the CNCs(1). In the initial stage of the reaction, the decomposition of the nickel oxalate occurs at relatively low temperature, which leads to the formation of small polygonal nickel nanoparticles. With the increasing reaction temperature, citric acid decomposes into carbon atoms. Then, the above mentioned as-formed nickel nanoparticles serve as nucleation center and the carbon atoms coat around the nickel nanoparticles. Hollow CNCs(1) could be obtained after the Ni particles being washed with hydrochloric acid solution. Furthermore, it is found that the CNCs(1) could also be obtained when using other carbon source such as ethanol, glucose, and so on. A schematic image of the formation process for CNCs is shown in Fig. 2f.
![]() | ||
Fig. 2 (a, b) SEM and TEM images of the CNCs(1); (c) HRTEM image of the wall structure of a carbon nanocage; TEM images of the CNCs(1) with different treatment processes: (d) before treated with acid; (e) washed with dilute hydrochloric acid for ∼2 h. (f) A schematic illustration of the formation process of the CNCs: (A) polygonal nickel nanoparticles were formed at first; (B) carbon atoms coated around these nickel nanoparticles; (C) graphitic layers were formed subsequently; (D) CNCs obtained after acid treatment process. |
The SEM image (Fig. 3a) reveals that a large quantity of quadrangular CNCs(2) are also obtained by this approach. The length of the side of CNCs(2) ranges from 200 to 350 nm and the wall thickness is about 10∼15 nm (Fig. 3e). Fig. 3b shows a typical image of the Fe3O4/CNCs(2) composites. Fig. 3c and d reveal clearly structural characteristic of the Fe3O4/CNCs(2) composites, which display a general morphology and distinct core/shell structure characteristics. It is clearly seen that Fe3O4/CNCs(2) composites are composed of outer graphite layers and inner Fe3O4 nanoparticles. Trace amount of Fe (0.35 wt%) was also detected in the final CNCs(2) according to the ICP test after being washed with acid.
![]() | ||
Fig. 3 (a) SEM image of the CNCs(2); (b) TEM image of the Fe3O4/CNCs(2) composites; (c, d) HRTEM image of the area marked with a rectangle in (b); (e) HRTEM image of the wall structure of a carbon nanocage. |
Raman spectra have been used to investigate the graphitization degree of the carbon materials.19 The Raman spectra of the CNCs(1) (Fig. 4a and b) show that there are two strong peaks centered at ∼1350 (D-band) and ∼1580 cm−1 (G-band). The G-band at 1580 cm−1 corresponds to the Raman-allowed optical mode E2g, which is closely related to the vibration in all sp2-band carbon atoms in a two-dimensional hexagonal lattice. The D-band at 1350 cm−1 could be assigned to the vibrations of carbon atoms with dangling bands in planar terminations of disordered graphite. Therefore, the intensity ratio of D-band and G-band (ID/IG) is usually used to determine the disorder degree in graphite layers. The higher ratio of ID/IG indicates more defects in graphite layers.20,21 It is worth noting that the ID/IG of the CNCs(1)-annealed shows a noticeable decrease compared to that of the CNCs(1). This result reveals that the graphitization degree of the CNCs(1) were improved after annealing treatment. The improved graphitization degree and quality of the CNCs is very important to the performance of the CNCs for lithium-ion batteries.
![]() | ||
Fig. 4 Raman spectra of the (a) CNCs(1) and (b) CNCs(1)-annealed. |
The graphitization degree and quality of the CNCs is evaluated by the TGA measurements.22,23Fig. 5a shows the TGA curves of the CNCs(1) and CNCs(1)-annealed. It is found that the CNCs(1)-annealed are more stable under high-temperature air oxidation than those of CNCs(1). The combustion temperature of the CNCs is remarkably increased after an annealing treatment. Compared with the CNCs(1) (∼500 °C), the combustion temperature of CNCs(1)-annealed is increased ∼100 °C. The similar phenomenon and tendency can also be found in Fig. 5b. It is known that the defect sites lead to a decrease in the thermal stability of the carbon materials at elevated temperature. High annealed temperature can facilitate the removal of functional groups and heal the defects in graphite layers.24 This further proves that the graphitization degree and structure stability can be significantly improved after annealing treatment. The high graphitization degree and low defects play important roles in improving the cycling performance of the carbon anode materials in lithium-ion batteries.
![]() | ||
Fig. 5 TGA curves of the CNCs: (a) CNCs(1) and CNCs(1)-annealed; (b) CNCs(2) and CNCs(2)-annealed. |
The electrochemical properties of the CNCs for lithium-ion battery applications are evaluated using galvanostatic discharge/charge cycles. Discharge/charge cycles are carried out in the potential window of 0.01–2.5 V versusLi. The voltage versus capacity and capacity versus cycle number curves for the CNCs and Fe3O4/CNCs at different current densities of 100, 300 and 500 mA g−1 are shown in Fig. 6. The first discharge capacities are 734 mAh g−1 for CNCs(1), 875 mAh g−1 for CNCs(1)-annealed, 808 mAh g−1 for CNCs(2), 1050 mAh g−1 for CNCs(2)-annealed, 1442 mAh g−1 for Fe3O4/CNCs(2) and 1580 mAh g−1 for Fe3O4/CNCs(2)-annealed at current density of 100 mA g−1. It can be observed that the cycling performance and rate capability of the samples have been significantly improved after annealing treatment at 600 °C for 10 h. Because the graphitic carbon have a high electronic conductivity and stabilized structure compared with that of disordered carbon.25 The capacity fades rapidly with high irreversible capacity in the first cycle. The first capacity fading mainly attribute to the decomposition of the electrolyte and the formation of the solid electrolyte interphase (SEI) layer at the surface of the CNCs.26,27 The irreversible capacity of carbon materials at the initial cycles is related to the reaction of lithium with active sites and functional groups (such as –CO or –OH etc.) on the electrode materials.28,29 Furthermore, a relatively volume change during the cycles can result in structural instability and capacity fading.1 During the subsequent cycles, the CNCs(1) and CNCs(1)-annealed present stable performance even at high current density of 500 mA g−1. It can be seen that the reversible capacities of the CNCs(1) and CNCs(1)-annealed are about 335 mAh g−1 and 380 mAh g−1 at current density of 100 mA g−1 after 50 cycles.
![]() | ||
Fig. 6 The first discharge/charge voltage curves and cycling performance at current densities of 100 mA g−1, 300 mA g−1 and 500 mA g−1: (a) CNCs(1); (b) CNCs(1)-annealed; (c) CNCs(2); (d) CNCs(2)-annealed; (e) Fe3O4/CNCs(2); (f) Fe3O4/CNCs(2)-annealed. |
Fig. 6c and d show the discharge/charge curves and cycling performance of the CNCs(2) and CNCs(2)-annealed. The CNCs(2)-annealed anode exhibits a good cycle stability and large reversible capacity. Although the discharge capacity dropped from 1050 mAh g−1 to 480 mAh g−1 after 10 cycles, the discharge capacity of the CNCs(2)-annealed anode gradually restored during the subsequent cycles, even after 50 cycles the anode remained a capacity of ∼520 mAh g−1 at current density of 100 mA g−1. More importantly, this sample exhibits a much better rate capability. After 50 cycles, the discharge capacity of this sample can be stably retained at about 380 mAh g−1 at high current density of 500 mA g−1.
Fig. 6e and f show the discharge/charge curves of the Fe3O4/CNCs(2) composite electrodes in the first cycle. In the first discharge cycle, both of them display a long voltage plateau at 0.5–0.75 V, and followed by a sloping curve down to the cut off voltage of 0.01 V, indicating a typical characteristic of voltage trends for the Fe3O4 electrode.19,30 The initial discharge capacity of Fe3O4/CNCs(2) is about 1442 mAh g−1 at a current density of 100 mA g−1, and the capacity is about 380 mAh g−1 after 50 cycles. However, at the large current density of 500 mA g−1, the capacity is only about 250 mAh g−1 after 50 cycles. The discharge capacity of the Fe3O4/CNCs(2) is poor as compared with that of other samples. After annealing treatment, the discharge capacity can reach 450 mAh g−1 at a current density of 100 mA g−1 and 320 mAh g−1 at a current density of 500 mA g−1 after 50 cycles. The improved cycling performance and rate capability can be attribute to the high electronic conductivity and stabilized structure. The stabilized graphitization layer can effective restrict the volume effect of the Fe3O4 nanoparticles.25
The excellent cycling performance and high-rate capability of the CNCs are ascribed to the unique hollow structure, which can improve the electron transportation, electrolyte penetration and shortened Li ion diffusion pathway in the discharge/charge cycles.11
In order to further compare the electrochemical performance of the typical pure carbon nanomaterials, some relevant information are summarized in Table 1. It is found that the first discharge capacities of the different CNCs have not been obviously improved when comparing with other pure carbon nanomaterials. But the as-prepared CNCs show both good cycling performance and enhanced high-rate performance. The reversible capacity after 50 cycles for CNCs(2)-annealed at current density of 100 mA g−1 is 520 mAh g−1, which is much larger than many previous reported pure carbon nanomaterials.
Typical materials | Current density mA g−1 | First capacity mAh g−1 | Cycle number | Remaining capacity mAh g−1 | Ref. |
---|---|---|---|---|---|
Carbon nanotubes | 57 | 1800 | 36 | 750 | 27 |
Carbon nanotubes | 75 | 310 | 20 | 320 | 31 |
Carbon nanotubes | 75 | 1000 | 30 | 200 | 32 |
Carbon nanotubes | 25 | 1200 | 50 | 450 | 33 |
Single-wall carbon nanotubes | 50 | 942 | 60 | 150 | 34 |
Ordered mesoporous carbon | 50 | 3083 | 50 | 500 | 35 |
Ordered mesoporous carbon | 100 | 3100 | 20 | 850 | 36 |
Hierarchical porous carbon | 18.6 | 710 | 20 | 220 | 37 |
Hierarchical porous carbon | 75 | 1500 | 40 | 500 | 38 |
Carbon nanosprings | 50 | 750 | 100 | 420 | 3 |
Carbon nanofibers | 100 | 430 | 30 | 220 | 5 |
Carbon nanofibers | 100 | 680 | 25 | 260 | 7 |
Carbon spheres | 50 | 950 | 135 | 400 | 16 |
Hollow carbon nanospheres | 50 | 540 | 15 | 290 | 39 |
CNCs/graphene | 100 | 1271 | 60 | 574 | 11 |
CNCs(1) | 100 | 734 | 50 | 335 | this work |
CNCs(1) | 500 | 586 | 50 | 238 | this work |
CNCs(1)-annealed | 100 | 875 | 50 | 373 | this work |
CNCs(1)-annealed | 500 | 507 | 50 | 280 | this work |
CNCs(2) | 100 | 808 | 50 | 314 | this work |
CNCs(2) | 500 | 427 | 50 | 242 | this work |
CNCs(2)-annealed | 100 | 1050 | 50 | 520 | this work |
CNCs(2)-annealed | 500 | 728 | 50 | 380 | this work |
Fig. 7 shows the Nyquist plots of the CNCs and CNCs-annealed. The depressed semicircles in the high frequency are related to the charge transfer process. The numerical value of the diameter of the semicircle on the Zre axis is approximately equal to the charge transfer resistance (Rct).40 It can be seen that there is an obvious decrease in Rct after the annealing treatment. This indicates that the electronic conductivity of the CNCs was improved after the annealing treatment.41,42 The inclined lines in the low frequency are attributed to the Warburg impedance, which is associated with lithium ion diffusion in the electrode materials (CNCs). The lithium ion diffusion coefficient could be calculated using the following equation:43
D = 0.5 (RT/AF2Cσ)2 | (1) |
Zre = Re + Rct + σω−1/2 | (2) |
![]() | ||
Fig. 7 Nyquist plots of the CNCs at open-circuit voltage. |
Samples | Re (Ω) | Rct (Ω) | σ (Ω cm2 S−0.5) | D (cm2 S−1) | io (mA cm−2) |
---|---|---|---|---|---|
CNCs(1) | 1.5 | 502.4 | 165.1 | 2.11E−15 | 5.11E−5 |
CNCs(1)-annealed | 2.2 | 109.5 | 114.4 | 4.93E−15 | 2.34E−4 |
CNCs(2) | 1.1 | 352.9 | 228.4 | 1.10E−15 | 7.28E−5 |
CNCs(2)-annealed | 1.8 | 101.4 | 47.7 | 2.52E−14 | 2.53E−4 |
It is found that the lithium ion diffusion coefficient of the electrode increased after annealing treatment. The diffusion coefficient of the CNCs(2)-annealed shows the higher mobility for lithium ions diffusion than other electrode materials.38 Furthermore, the exchange current density (io = RT/nFRct) of the CNCs(2)-annealed is higher than in other electrode materials. Therefore, the charge transfer reaction of the CNCs(2)-annealed is stronger than that of other electrode materials.44,45 It is demonstrated that the Rct obviously decreases and the lithium ion diffusion coefficient increases after annealing treatment, and therefore improves the discharge/charge performance and rate capability of the batteries.
The reason for the different capacities in lithium-ion batteries with the use of the different kinds of CNCs can also be explained by the Nyquist plots. It is obvious that the charge transfer resistance (Rct) of the CNCs(2) was smaller than that of CNCs(1), indicating the better electronic conductivity (means the higher Li ion transfer speed across interfaces between the electrolyte and active electrode materials)3,5 of the former sample. In addition, the wall thickness of the CNCs(1) was about 70 nm while that of the CNCs(2) was only about 10∼15 nm (see Fig. 2b and Fig. 3e). Thinner graphite layers are favorable for the electrolyte penetration and shortened Li ion diffusion paths in discharge/charge cycling.14 Therefore, the capacities and performances of the CNCs(2) were better than those of CNCs(1) as anode materials in lithium-ion batteries in this study.
This journal is © The Royal Society of Chemistry 2012 |