Journal of Power Sources 9798 (2001) 4751

High-capacity carbons for lithium-ion batteries prepared from rice husk

George Ting-Kuo Fey*, Chung-Lai Chen
Department of Chemical Engineering, National Central University, Chung-Li 32054, Taiwan, ROC
Received 6 June 2000; accepted 4 December 2000

Abstract

We have prepared carbon anode materials by pyrolysis of rice husk (RH). Through X-ray powder diffraction, BrunauerEmmettTeller
(BET) measurement, and ICP-AES analysis, the effects of acidbase pre-treatment, pyrolysis temperature, and the use of a proprietary pore-
genic agent on the cell capacity of these materials were investigated. The carbonaceous materials made from RH treated with a proprietary
agent showed an extraordinarily high reversible capacity of 1055 mAh/g. To our best knowledge, this new hard carbon material possesses
the highest reversible capacity ever reported for any carbon anode materials of lithium-ion batteries. # 2001 Elsevier Science B.V. All
rights reserved.

Keywords: High-capacity carbon; Lithium-ion batteries; Rice husk

1. Introduction many years from those both inside and outside the eld of
agriculture.
The search for low-cost and high-capacity carbon anode The major constituents of RH are silica (2025%) and
materials is now arousing interest worldwide, stimulated by cellulose, which yields carbon when pyrolyzed under inert
increasing demand for lithium-ion batteries used to power atmosphere [9]. Silicon-doped hard carbons were reported to
portable electronic devices such as notebook computers and improve cell capacity and cycle behavior [10,11]. Accord-
phones [1] and for vehicle propulsion in zero-emission ingly, we consider RH to be a potentially good low-cost
vehicles (ZEVs) [2]. High-capacity carbons refer to materi- source for high-capacity amorphous carbon containing
als with capacities greater than that of graphite (372 mAh/g). micro-amounts of silicon.
Several series of hard carbons used as anodes of lithium-ion The main objectives of this preliminary work are: (a) to
batteries have been prepared by pyrolysis of natural or develop a procedure for preparing high-capacity hard carbon
agricultural precursors such as sugar [35], cotton [6] and from RH; (b) to optimize the process variables such as
coffee beans [7]. The capacity of such materials is critically pyrolysis temperature and pore-genic agents; (c) to attempt
dependent on pyrolysis conditions and precursor sources. To preparation of hard carbons with high surface area and well-
lower the cost of raw materials, the selection of precursor has developed porosity by carbonization with pore-genic agents.
been shifted to natural or agricultural residues in this work.
Rice husk (hereafter abbreviated as RH) is a form of
agricultural biomass and a major by-product in the rice 2. Experimental
milling industry. The estimated world-wide RH production
was about 100 million tons according to a 1996 report by the The raw material, RH obtained from a local rice mill, was
Food and Agricultural Organization [8]. Disposal of such thoroughly washed with water to remove adhering soil and
vast amounts of RH has been one of the main problems clay, and dried at 1108C in an oven overnight. The dried husk
facing the rice milling industry. The traditional method of (ca. 30 g) was reuxed with 400 ml of 3N HCl solution for
economical disposal, namely burning, faces growing public 1 h. After leaching, the husk was thoroughly washed with
opposition and legal regulation. The utilization and manage- distilled water until the ltrate was free from acid. The
ment of rice crop residues has received more attention over leached husk was then dried at 1108C for 2 h. After drying,
the husk was rst transferred into a quartz tube, which was
then purged with argon for 30 min. The tube was initially
*
Corresponding author. Tel.: 886-3-425-7325; fax: 886-3-425-7325. heated at 1508C for 1 h to remove water from the husk and
E-mail address: [email protected] (G.T.-K. Fey). then heated at a 58C/min rate to reach pyrolysis temperature

0378-7753/01/$ see front matter # 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 7 7 5 3 ( 0 1 ) 0 0 5 0 4 - 3
48 G.T.-K. Fey, C.-L. Chen / Journal of Power Sources 9798 (2001) 4751

(from 500 to 9008C) for another hour. The black residues 9008C to obtain an optimum or practical temperature, which
(about 8 g) were ground as carbon product. In the case of would result in the largest capacity for lithium in RH
base treatment, the acid-free husk was reuxed with carbons. Three series of carbon samples were prepared with
400 ml of NaOH solution (0.1, 0.3, or 4N) for 1 h. After the use of acid, base and pore-genic agent. The d0 0 2 value is
leaching, the husk was washed with distilled water until the interlayer distance between graphene sheets. The empiri-
the ltrate was free from base and the process continued cal parameter, R, is dened in [13], and measured as the ratio
following the normal procedure as described above. When of the peak count rate at the (0 0 2) peak divided by the
a pore-genic agent was used, the husk was immersed in the background level at the same angle. It is used to estimate the
solution of the agent for 1 day with sufcient stirring. After fraction of single-layers in hard carbon samples randomly
pyrolysis, the carbonaceous product was then treated with positioned among groups of single-layers, bilayers and
3N HCl solution to remove metallic ingredients present in trilayers. The value of R depends on the single layer fraction.
the RH. Liu et al. indicated that samples with lower R, having more
Powder XRD measurements were made with a Siemens single layers, generally gave larger specic capacities. The
D-500 diffractometer equipped with a diffracted beam hydrogen to carbon atomic ratio of the samples (H/C atomic
monochromator including D-5000 data treatment software ratio) was calculated by taking the ratio of hydrogen to
and Cu Ka radiation of 1.5406 A . The sample was set up on carbon weight percentages (obtained from the results of
a glass holder in the diffractometer. Data were collected CHN elemental analysis) and multiplying by 12 (carbon
between 5 and 808 (2y) at a 0.058 resolution. The specic atomic weight). Zheng et al. showed that large hysteresis is
surface area of the carbon sample was measured by nitrogen highly correlated to the high hydrogen content or H/C
gas adsorption and the BrunauerEmmettTeller (BET) atomic ratio [14].
calculation in a Micromeritics ASAP-2010 surface analyzer. The XRD patterns of the RH carbons pyrolyzed in the
A mixture of 30 mol% (m/o) nitrogen in 70 m/o helium was temperature range between 500 and 9008C are shown in
used as a carrier gas. The hydrogen and carbon contained in Fig. 1. All the peaks have broad shapes indicating highly
the disordered carbons were measured with a Heraeus CHN- disordered carbons [15]. Only the (0 0 2) Bragg peak due to
rapid elemental analyzer. the stacking of carbon layers appears signicant. The prole
Carbon electrodes were fabricated on a weight basis of of the (0 0 2) peak of these materials changes little as the
90% carbon, 8% polyvinyldiuoride (PVDF) and 2% carbon temperature is increased. However, the interlayer distance
black (added to enhance the electronic conductivity). Excess (d0 0 2) increases with increasing pyrolysis temperature. As
NMP was used until the paste achieved a smooth syrupy listed in Table 1, the interlayer distance increases from
viscosity. The paste was then spread uniformly on copper 3.770 A for carbon pyrolyzed at 5008C to 4.100 A for carbon
foil and NMP evaporated at 1108C for about 1.5 h. The pyrolyzed at 9008C. The increase in d0 0 2 may be ascribed to
coated foil was compressed into 110130 mm. Then, the foil the introduction of some interstitial atoms into the initial
was cut into a circle with a diameter of 13 mm. Coin-type- pristine graphite structure [16]. Since silica is one of the
2325 cell hardware used for the cycling tests was similar to major constituents (2025%) in RH, silicon could be the
that described in [12]. A microporous lm (Celgard 3501 as interstitial atoms between the aromatic planes of carbon
the separator) wetted with electrolyte (1 M LiPF6 dissolved atoms. In this work, the quantitative results from induc-
in a 50/50 v/v% mixture of EC and DEC) was sandwiched tively-coupled plasma-atomic emission spectroscopy (ICP-
between the carbonaceous cathode and lithium foil anode. AES) analysis indicated that the silicon content in our
The coin cell assembly was carried out in an argon-lled carbon products without any base treatment pyrolyzed
glove box where both moisture and oxygen concentration between 500 and 9008C was about 1516%, whereas the
were about 1 ppm. A Maccor S-4000 Battery Test System silicon content in the carbon products with base treatment at
controlled the cell cycling tests. The cells were rst dis- various NaOH concentrations pyrolyzed at 7008C was less
charged using a constant current of 37.2 mAh/g (0.18C rate) than 0.01% beyond the detection limit of ICP-AES method.
until the voltage dropped to 0.005 V (versus lithium metal), The H/C atomic ratio decreases monotonically for each of
then the cells were charged to 3 V. the samples as they are heated. For instance, the H/C atomic
ratio decreases from 0.68 to 0.31 as the pyrolysis tempera-
ture increased from 500 to 9008C.
3. Results and discussion Fig. 2 displays XRD patterns of the RH-based carbons
pyrolyzed at 7008C. Prior to pyrolysis, the precursors of
Table 1 describes the RH carbon samples prepared in this these carbon samples were treated with various concentra-
preliminary work, and contains the following items: (a) the tions (0.1, 0.3, or 4N) of NaOH solutions to remove silica
sample code, (b) pyrolysis temperature, (c)(e) process from RH. The d0 0 2 values of RHC06 to RHC08 in Table 1
variables such as the use of acid, base and pore-genic agent, show that the interlayer distance decreases with increasing
(f)(h) XRD results, (i) H/C atomic ratio, (j) silicon content, NaOH concentration. This is good evidence that the
(k) BET surface area, and (l) reversible capacity. The decrease in silicon content due to the treatment of high
pyrolyzing temperature, Tpyr, was changed from 500 to NaOH concentration results in a decrease in d0 0 2. The
Table 1
Summary of the samples prepared from rice husk in this work

G.T.-K. Fey, C.-L. Chen / Journal of Power Sources 9798 (2001) 4751
(a) Sample (b) Tpyr (c) Acid leaching (d) NaOH (e) Pore-genic (f) 2y (g) d0 0 2 (h) X-ray (i) H/C (j) Si (k) Surface (l) Capacity (mAh/g)
(8C) (3N HCl) (N) agent ratio (0 0 2) )
(A ratio (R) atomic ratio content (%) area (m2/g)

Prior to After First cycle Tenth cycle

pyrolysis pyrolysis

D C D C

RHC01 500 v 23.58 3.770 1.67 0.68 14.8 199 7 2 2 2

RHC02 600 v 23.46 3.789 2.24 0.53 15.7 277 605 279 235 238
RHC03 700 v 22.41 3.964 2.02 0.48 15.9 277 691 320 289 279
RHC04 800 v 21.75 4.083 2.07 0.46 15.0 294 637 279 240 234
RHC05 900 v 21.66 4.100 2.34 0.31 16.2 205 207 89 131 128
RHC06 700 v 0.1 23.05 3.855 1.96 0.46 267 444 165 134 132
RHC07 700 v 0.3 23.31 3.813 1.88 0.34 371 819 463 293 292
RHC08 700 v 4 23.37 3.803 1.85 0.32 <0.01 404 571 354 265 261
RHC09 500 v 1:1 23.04 3.857 1.92 0.45 1296 1190 457 608a 554a
RHC10 500 v 1:3 23.68 3.754 2.00 0.42 1597 2259 853 858a 793a
RHC11 500 v 1:5 23.95 3.713 2.10 0.37 1403 2374 1055 1128a 1051a
a
Data were taken at the end of the fifth cycle; D: discharge; C: charge.

49
50 G.T.-K. Fey, C.-L. Chen / Journal of Power Sources 9798 (2001) 4751

Fig. 1. XRD patterns of RH carbons pyrolyzed at various temperatures.

effect of NaOH treatment on d0 0 2 is clearly seen in the fact

that the interlayer distance decreases from 3.964 A without
Fig. 3. Voltage profiles for cells with RH carbon samples pyrolyzed at
base treatment to 3.8553.803 A with base treatment. The H/
various temperatures.
C atomic ratio decreases slightly after base treatment,
especially in higher concentrations.
The effect of Tpyr from 500 to 9008C on the capacity of Li/ 280 mAh/g at the tenth cycle. During the base treatment, only
carbon coin cells was studied. The voltage proles of these samples reuxed with NaOH at higher concentrations made
samples are shown in Fig. 3, except for the RHC01 sample slight improvements in capacity. The voltage proles are
pyrolyzed at 5008C which showed virtually no capacity as shown in Fig. 4.
indicated in Table 1. The optimum Tpyr was 7008C in terms When pretreated using a proprietary pore-genic agent of a
of the discharge and charge capacity of the rst cycle, 691 and higher concentration prior to pyrolysis, the resulting hard
319 mAh/g, respectively. The reversible capacity was about carbon products (RHC10 and RHC11) showed a tremendous

Fig. 2. XRD patterns of RH carbons treated with various concentrations of Fig. 4. Voltage profiles for cells with RH carbon samples treated with
NaOH solutions prior to pyrolysis. various concentrations of NaOH solutions prior to pyrolysis.
 G.T.-K. Fey, C.-L. Chen / Journal of Power Sources 9798 (2001) 4751 51

values of RHC11 were 2374 and 1055 mAh/g, respectively.

At the fth cycle, the reversible capacity remained at
1051 mAh/g. To our best knowledge, both irreversible
and reversible capacity values are the highest ever reported
for any carbon anode materials of lithium-ion batteries. Even
though the irreversible capacity of RHC11 was very high
and not practical at this early stage, we believe that there is
enormous potential for future work on these new super hard-
carbon materials.

References

[1] R. Krause, Pop. Sci. 64 (1993).

[2] K.M. Abraham, Electrochim. Acta 18 (1993) 1233.
[3] W. Xing, J.S. Xue, J.R. Dahn, J. Electrochem. Soc. 143 (1996) 3046.
[4] A. Gibaud, J.S. Xue, J.R. Dahn, Carbon 34 (1996) 499.
[5] W. Xing, R.A. Dunlap, J.R. Dahn, J. Electrochem. Soc. 145 (1998)
62.
Fig. 5. Voltage profiles for cells with RH carbon samples treated with a
[6] E. Peled, V. Eshkenazi, Y. Rosenberg, J. Power Sources 76 (1998)
proprietary pore-genic agent at various weight ratios of rice husk to the
153.
agent.
[7] S. Yamada, H. Imoto, K. Sekai, M. Nagamine, in: 191st Meeting of
the Electrochemical Society (Abstract), 1997, p. 85.
[8] Production Yearbook, Basic Data Unit Statistics Division, Food and
increase in cell capacity. Fig. 5 displays the voltage versus Agricultural Organization, Rome.
capacity for Li/carbon cells containing RHC09, RHC10, and [9] C. Real, M.D. Alcala, J.M. Criado, J. Am. Ceram. Soc. 79 (1996)
RHC11 samples, whose raw husks were pretreated, prior to 2012.
pyrolysis, in a solution of a weight ratio of RH to pore-genic [10] A.M. Wilson, J.R. Dahn, J. Electrochem. Soc. 142 (1995) 326.
[11] J.S. Xue, K. Myrtle, J.R. Dahn, J. Electrochem. Soc. 142 (1995)
agent, 1:1, 1:3, and 1:5, respectively. As listed in Table 1, the
2927.
cell capacity of these samples increases with an increase in [12] J.R. Dahn, A.K. Sleigh, H. Shi, J.N. Reimers, Q. Zhong, B.M. Way,
the weight ratio of RH to pore-genic agent. Interestingly, the Electrochim. Acta 38 (1993) 1179.
use of a pore-genic agent results in a signicant increase in [13] Y.H. Liu, J.S. Xue, T. Zheng, J.R. Dahn, Carbon 34 (1996) 193.
surface area or the number of micropores. Based on BET [14] T. Zheng, Y.H. Liu, E.W. Fuller, S. Tseng, U. von Sacken, J.R. Dahn,
J. Electrochem. Soc. 142 (1995) 2581.
analysis, surface area of raw RH carbons without any
[15] J.R. Dahn, A.K. Sleigh, H. Shi, J.N. Reimers, Q. Zhong, B.M. Way,
treatment was about 20 m2/g, whereas surface area of RH Electrochim. Acta 38 (1993) 1179.
carbons with pore-genic treatment was at least 1200 m2/g as [16] M. Tidjani, J. Lachter, T.S. Kabre, R.H. Bragg, Carbon 24 (1986)
shown in Table 1. The rst discharge and charge capacity 447.