Novel spherical microporous carbon as anode material for Li-ion batteries

Novel spherical microporous carbon as anode material for Li-ion batteries

Solid State Ionics 152 – 153 (2002) 43 – 50 Novel spherical microporous carbon as anode material for Li-ion batteries Qin...

498KB Sizes 0 Downloads 8 Views

Solid State Ionics 152 – 153 (2002) 43 – 50

Novel spherical microporous carbon as anode material for Li-ion batteries Qing Wang, Hong Li, Liquan Chen, Xuejie Huang* Laboratory for Solid State Ionics, Institute of Physics, Chinese Academy of Sciences, P.O. Box 603, Beijing, 100080, China Accepted 18 January 2002

Abstract Hard carbon spherules (HCS) with micropores were prepared by a hydrothermal method. It has perfect spherical morphology, controllable monodisperse particle size and smooth surface. XRD and Raman spectra show that HCS is nongraphitizable. The reversible capacity of HCS is about 430 mA h/g in using as anode material for lithium ion batteries, and the cyclic performance of HCS is excellent. The kinetic characteristics of HCS are better than MCMB. In addition, the Coulombic efficiency of HCS has been improved by surface modifications such as CVD of acetylene and coating of tetraethoxysilane (TEOS) on the surface of HCS. Furthermore, pinning of the nanosized SnSb alloy particles on the surface of HCS hinders the electrochemical aggregation of alloy particles effectively during charge/discharge cycles. Consequently, the cyclic performance and reversible capacity are much enhanced. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Li-ion batteries; Anode; Spherical carbon; Microporous; Composite

1. Introduction Li-ion batteries are currently the best portable energy devices for the consumer electronics. Recent development of the Li-ion batteries has been achieved by using selected carbon or graphite materials as anode [1 –3]. The performances of Li-ion batteries, e.g., capacity, voltage profile and cycling life, depend strongly on the microstructure of the anode materials. To get higher energy and power densities, a significant attention has been paid to investigation of key parameters of carbons which influence electrochemical properties. *

Corresponding author. Tel.: +86-10-82649046; fax: +86-1082649050. E-mail address: [email protected] (X. Huang).

Carbon materials have large varieties in microstructure, texture, crystallinity and morphology, depending on their preparation processes and the precursor materials. Carbon with spherical morphology has been proved to be competent in using as anode material for Li-ion batteries owing to its high packing density, low surface-to-volume ratio and maximal structural stability, etc. One example is the commercial graphitized mesocarbon microbeads (MCMB). In addition, it is evident that the micropores within the carbon can supply extra capacity [4,5]. Hard carbon spherules (HCS) with micropores have been prepared by hydrothermal method [6]. It has perfect spherical morphology, controllable monodisperse particle size and smooth surface. In this paper, we will introduce the novel HCS material, including its structural characteristics and electro-

0167-2738/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 2 7 3 8 ( 0 2 ) 0 0 6 8 7 - 2


Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

Fig. 1. Schematics of the possible formation process of dewatering sugar spherules through emulsion polymerization mechanism. (A) Dewatering of sugar; (B) formation of spherical micelles; (C) growing of the nuclei.

chemical behaviors. Furthermore, surface modification and combination with nano-SnSb alloy of HCS will be also reported in detail.

ization at high temperatures. The detailed process has been reported in a previous paper [6]. The carbonization was carried out under argon atmosphere at 1000 and 2500 jC for 2 h, respectively (the corresponding samples were denoted as HCS10 and HCS25).

2. Experimental 2.1. Material preparation and characterization HCS was prepared by hydrothermal method. Here, sugar was selected as the precursor. The typical preparation process can be described simply in two steps, dewatering at low temperatures and carbon-

Fig. 2. SEM image of HCS heat-treated at 1000 jC. Diameter of the spherules demonstrated here is about 6.5 Am.

Fig. 3. Powder XRD patterns of HCS heat-treated at 1000 jC (HCS10) and 2500 jC (HCS25), respectively.

Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50


Fig. 6. Variation of DLi with the potential of an HCS10 electrode. The inset shows that of a graphitized MCMB electrode.

tion of other samples will be discussed in corresponding sections. Fig. 4. Raman spectra of HCS10 and HCS25.

The obtained hard carbon spherules were characterized by powder X-ray diffraction (Rigaku B/max2400 X-ray diffractometer equipped with Cu radiation), scanning electron microscope (Hitachi S-4000), JEOL 200CX transmission electron microscope and Raman spectra (Renishow 1000 micro-Raman Spectrometer with an excitation line 632.8 nm). Prepara-

2.2. Electrochemical measurement A typical carbon electrode was prepared by coating slurries of HCS (or other samples) and polyvinylidene fluoride (PVDF, 5 wt.%) dissolved in N-methyl pyrrolidone on copper foil substrates. After coating, the sheet was dried under vacuum at 120 jC for 8 h. The counter and reference electrodes were made of Li foil, and the electrolyte was the solution of 1 M LiPF6

Fig. 5. The first two charge/discharge cycles of a Li/HCS10 cell. The Coulombic efficiency is 73% during the first cycle.


Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

Fig. 7. Discharge profiles for an HCS10/LiCoO2 battery at two rates.

dissolved in a 50/50 vol.% mixture of ethylene carbonate (EC) and diethyl carbonate (DEC). A threeelectrode glass cell was assembled in an argon-filled glove box. The electrochemical measurements were performed by using an electrochemical workstation (CH Instruments, Model 660A). The testing procedure included charge/discharge experiment and potential relaxation measurements. The latter technique has been described in a previous paper [7].

3. Results and discussion 3.1. General features of HCS Different from the formation process of MCMB at a high temperature, of which the associated mesogens are arranged spherically throughout the liquid-crystallike mesophase pitch due to minimum surface energy

Fig. 9. SEM images of the surface modified HCS. (A) HCS – CVD; (B) HCS – SiOx.

[8], the formation of dewatering sugar spherules is presumably more close to the conventional emulsion polymerization mechanism of colloidal spheres [9].

Fig. 8. Schematics shows the filling of nanopores with deposited acetylene molecules. (A) Adsorption and pyrolysis of acetylene molecules on the pore wall; (B) carbonization and deposition of pyrolyzed acetylene molecules on the pore wall (see Ref. [14]).

Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

Presumably, as the autoclave is heated to a certain temperature, the adjacent sugar molecules in the solution start to dehydrate and polymerize (polycondensation reaction) together. Consequently, a kind of amphiphilic compound with a larger hydrophobic alkyl and hydrophilic hydroxyl is formed, as the model shown in Fig. 1. With the polymerization of sugar molecules, as the concentration reaches the critical micelle concentration (cmc), the amphiphilic compound forms spherical micelles, of which the hydrophobic groups make up of the core of the micelles, and the hydrophilic hydroxyls occupy the surface. Following the nucleation process, the surface hydroxyls continue to combine with the nearest free molecules and dehydrate; consequently, the spherical micelle particles begin to grow up. As the sugar molecules are exhausted, the growth of the spherical particles stops. It can be seen from the image of scanning electron microscope shown in Fig. 2 that the HCS particles have perfect spherical morphology and smooth surface, whereas the widely used graphitized MCMB has sphere-like shape and rough surface. After graphitizing at 2500 jC, morphology of the obtained HCS25 has hardly changed. The particle size of HCS could be


controlled by changing the preparation condition. In the case of Fig. 2, the diameter of a typical HCS is about 6.5 Am. The structure of HCS was characterized by XRD and Raman spectra. The XRD patterns of HCS10 and HCS25 were shown in Fig. 3. Clearly, HCS10 (heat treatment at 1000 jC) is disordered carbon. The d002 is about 3.9 nm. In addition, the ‘‘tail’’ at low Bragg angle region indicates porous structure of the material [10]. We also performed N2 adsorption experiment. The BET area of HCS10 is about 400 m2/g. Therefore, there must be many micropores within the spherules. As evidenced from the XRD pattern of HCS25, treatment at higher temperature for instance at 2500 jC results in partly graphitized carbon containing microcrystalline regions. According to Scherrer equation, the crystallite size along c-axis (Lc) increases with reduction of FWHM of 002 peaks after graphitization. The Raman spectra of HCS10 and HCS25 were shown in Fig. 4. Commonly, the 1575 cm  1 line (G-mode) is assigned to the E2g species of the infinite crystal of graphite, whereas the 1355 cm  1 (D-mode) line is attributed to a particle size effect [11]. It is known that the intensity ratio of IG/ID is proportional to the crystallite size along the a-axis

Fig. 10. The first two charge/discharge cycles of a Li/HCS – SiOx cell. The Coulombic efficiency is 82.5% during the first cycle.


Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

(La). It is clear from Fig. 4 that La is reduced after graphitization. Hence, the obtained spherical carbon is a kind of hard carbon. The conclusion is also proved by HRTEM and the following electrochemical results.

The diffusion of lithium between buckled graphene sheets is faster than that in the micropores evidently (Fig. 6). In addition, we also compared the chemical diffusion coefficient of lithium in HCS10 with that in

3.2. Electrochemical properties The charge/discharge profiles of HCS10 are shown in Fig. 5. The reversible capacity is about 430 mA h/g, and little hysteresis can be seen. From the charge and discharge curves, two parts can be distinguished: one is the slope region at higher voltage (>0.1 V), and the other is the flat region at lower voltage ( < 0.1 V). They should correspond to two different insertion/ extraction mechanisms of lithium ions [4,12,13]. Unlike graphitized carbon, generally insertion/extraction of lithium in the buckled graphene sheets and edges occurs at higher voltage, and lithium filling in micropores occurs at lower voltage. Recently, Stevens and Dahn [5] have further confirmed the mechanism of lithium filling into nanopores in pyrolyzed glucose near 0 V (vs. Li/Li + ) by means of in situ small-angle X-ray scattering technique. Using potential relaxation technique [7], the cell was first discharged or charged to a certain voltage, which corresponds to certain lithium content in the electrode. Then, the anodic or cathodic current was kept at zero, and the changes of open circuit voltage were recorded with time using chronopotentiometry technique. From Fick’s second law and Nernst equation, the variation of potential with time can be expressed as h u  u  i p2 ˜ ln exp l F  1 ¼ ln N  2 D ð1Þ host t RT L According to the thickness of the electrode L (here select the average radius of HCS) and potential at infinite time ul, we can evaluate the chemical diffusion coefficient of lithium in the HCS from the variation of potential with time and Eq. (1). The variation of DLi with potential is shown in Fig. 6. With the falling of potential, DLi decreases continuously till the potential tends to 0.1 V or so. As the potential is lower than 0.1 V, the value of DLi shows almost no change. However, further insertion of lithium at full lithiated state of HCS below 0.0 V results in steep decrease of DLi. Thus, the variation indicates three types of electrochemical processes.

Fig. 11. SEM images of SnSb alloy pinned HCS. Morphology of HCS – SnSb (A), (B) before cycling and (C) after cycling.

Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

graphitized MCMB. Obviously, the chemical diffusion coefficient of lithium in HCS10 above 0.1 V is larger than that of graphitized MCMB from the inset of Fig. 6. It is reasonable since the graphene layer spacing of HCS is larger than that of graphitized MCMB. The larger value of DLi in the HCS suggests good rate capability. Fig. 7 shows the discharge profiles of an HCS/LiCoO2 cell at different current density. The discharge capacity obtained under 2.5 C is still more than 90% of that under 0.5 C. Therefore, this HCS material can stand high rate charge/discharge and is promising for electric vehicle utilization. 3.3. Surface modification of HCS Like most amorphous carbonaceous materials, this kind of material also presents large irreversible capacity during the first charge/discharge cycle (Coulombic efficiency is about 73% at the first cycle). The formation of SEI on outer surface and within micropores of HCS may play an important role. Surface modification is applied to solve this problem. Firstly, in order to fill and block the open nanopores as seen in Fig. 8, HCS was treated by the CVD of acetylene at 900 jC as the preparation of molecular sieve carbon [14]. However, the obtained sample was covered by


loose carbon soot (see SEM image in Fig. 9(A)), which may bring extra capacity loss. In addition, we have noticed recently that there is no trace of SEI formed on Si particles during charge/discharge processes according to their voltage profiles, HRTEM [15 – 17] and unpublished FTIR results. It may be related to a layer of amorphous silica. Therefore, we further treat the sample with tetraethoxysilane (TEOS). It can be seen in Fig. 9(B) that after hydrolyzing and heat treatment, the reacted TEOS covered the surface of HCS. The charge/discharge profiles of the treated samples were shown in Fig. 10. Obviously, the presence of SiOx has greatly improved the efficiency compared with the original HCS. However, the improvement of efficiency during the first cycle also brings the decrease of reversible capacity in the meantime, namely, a reduction from 430 to 360 mA h/g compared with pristine HCS. 3.4. Alloy –HCS composite Nanosized alloy particles, such as Sb, SnSb and Si, possess very high Li-storage capacity, but they show serious electrochemical agglomeration [16 – 20]. It is one of the key factors leading to capacity fading [20]. Pinning nanosized alloy on carbon surface has been

Fig. 12. Charge/discharge curves of a Li/HCS – SnSb cell.


Q. Wang et al. / Solid State Ionics 152 – 153 (2002) 43–50

proved to be an effective way to hinder the electrochemical agglomeration [19,20]. The unique surface structure of HCS makes it an excellent carrier for the dispersion of nano-alloy. In this section, a similar preparation process was employed [19,20]. Morphology of the SnSb –HCS composite with a SnSb content of 30 wt.% was shown in Fig. 11(A) and (B). Clearly, most of SnSb alloy particles with the average diameter around 100 nm are dispersed on the surface of HCS particles uniformly and separately. The typical voltage profiles shown in Fig. 12 indicate that different electrochemical reactions occurred on the SnSb –HCS composite. Besides the insertion/extraction of Li into the hard carbon, a slope ranged from 1.1 to 0.8 V appears at the first discharge curve and disappears in the following cycles during Li insertion process, which mainly attributes to the decomposition reaction caused by a small amount of surface oxide on nano-SnSb. The plateaus at 0.8 V in the discharge curve and at 1.1 V in the charge curve are related to the alloy reaction of Li with Sb (since some Li –Sb alloy has little electrochemical reversibility, the plateau during extraction is shorter than that of insertion). The multistep Li – Sn alloy reactions should drop to the voltage region ranged from 0.7 to 0.2 V. After a discharge/charge cycle, the SnSb particles are still separated on the surface of HCS (see Fig. 11(C)). However, for the pure nano-SnSb alloy material, it has lost its particle shape and merged into a bulk [20]. Therefore, the goal of hindering electrochemical agglomeration of alloy particle is achieved. Consequently, the cyclic performance of the composite is greatly improved compared with pure SnSb alloy. In this case, all elements in the composite materials are active for Li storage. The measured value is 480 mA h/g, which is higher than pure HCS.

4. Conclusions The HCS material prepared by hydrothermal method has perfect spherical morphology, controllable monodisperse particle size and smooth surface. XRD and Raman spectra indicate that the obtained carbonaceous material is nongraphitizable. When using as negative electrode material for Li-ion batteries, the HCS shows rather large reversible capacity. Moreover, the chemical diffusion coefficient of lithium in

HCS is larger than that in graphitized MCMB. The lithium ion battery using HCS as negative electrode shows good rate ability. In addition, covering of SiOx on the surface of HCS has been proved to be an effective way to the improvement of efficiency of the HCS material. Moreover, pinning nano-alloy particles on carbon surface tightly not only enhances the dimensional stability of alloy particles during Li insertion and extraction, but also integrates the other advantages of carbon and nano-alloy.

Acknowledgements This work is supported by NSFC (No. 59972041) and National 863 key program.

References [1] M. Winter, J.O. Besenhard, M.E. Spahr, P. Novak, Adv. Mater. 10 (1998) 725. [2] S. Flanfrois, B. Simon, Carbon 37 (1999) 165. [3] M. Endo, C. Kim, K. Nishimura, T. Fujino, K. Miyashita, Carbon 38 (2000) 183. [4] E. Peled, V. Eshkenazi, Y. Rosenberg, J. Power Sources 76 (1998) 153. [5] D.A. Stevens, J.R. Dahn, J. Electrochem. Soc. 147 (2000) 4428. [6] Q. Wang, H. Li, L. Chen, X. Huang, Carbon 39 (2001) 2211. [7] Q. Wang, H. Li, X. Huang, L. Chen, J. Electrochem. Soc. 148 (2001) A737. [8] H. Marsh, M. Martinez-Escandell, F. Rodriguez-Reinoso, Carbon 37 (1999) 363. [9] Y. Xia, B. Gates, Y. Yin, Y. Lu, Adv. Mater. 12 (2000) 693. [10] W. Xing, J.S. Xue, J.R. Dahn, J. Electrochem. Soc. 143 (1996) 3046. [11] F. Tuinstra, J.L. Koenig, J. Chem. Phys. 53 (1970) 1126. [12] Y. Han, J. Yu, G. Park, J. Lee, J. Electrochem. Soc. 146 (1999) 3999. [13] J.R. Dahn, T. Zheng, Y. Liu, J.S. Xue, Science 270 (1995) 590. [14] T. Kyotani, Carbon 38 (2000) 269. [15] H. Li, G. Zhou, D. Yu, X. Huang, L. Chen, Z. Zhang, Appl. Phys. Lett. 73 (1999) 677. [16] H. Li, X. Huang, L. Chen, G. Zhou, Z. Zhang, D. Yu, Y. Mo, N. Pei, Solid State Ionics 135 (2000) 181. [17] H. Li, X. Huang, L. Chen, Z. Wu, Y. Liang, Electrochem. Solid-State Lett. 2 (1999) 547. [18] H. Li, G. Zhu, X. Huang, L. Chen, J. Mater. Chem. 10 (2000) 693. [19] L. Shi, H. Li, Z. Wang, X. Huang, L. Chen, J. Mater. Chem. 11 (2001) 1501. [20] H. Li, L. Shi, W. Lu, X. Huang, L. Chen, J. Electrochem. Soc. 148 (2001) A915.