3] - American Chemical Society

Jan 3, 2012 - Material Science and Engineering Program, University of California Riverside, Riverside, ... 600 °C), L333 crystals are quite small (ca...
3 downloads 0 Views 3MB Size
Article pubs.acs.org/crystal

Crystal Growth of Li[Ni1/3Co1/3Mn1/3]O2 as a Cathode Material for High-Performance Lithium Ion Batteries Jianxin Zhu,† Thinh Vo,‡ Dongsheng Li,‡,# Richard Lu,‡ Nichola M. Kinsinger,‡ Laj Xiong,§ Yushan Yan,§ and David Kisailus*,†,‡ †

Material Science and Engineering Program, University of CaliforniaRiverside, Riverside, California 92521, United States Department of Chemical and Environmental Engineering, University of CaliforniaRiverside, Riverside, California 92521, United States



ABSTRACT: Li[Ni1/3Co1/3Mn1/3]O2 (L333) has been successfully synthesized through a two-step coprecipitation and annealing method. Crystal phase and growth behavior were monitored by X-ray diffraction (XRD), while particle morphologies were observed by SEM and TEM. The crystal growth mechanisms during the annealing process were interpreted, and the growth exponent and activation energy for crystal growth were calculated. At low temperatures (e.g., 600 °C), L333 crystals are quite small (ca. 14 nm). Significant crystal growth occurs above 750 °C with crystallite diameters increasing to 850 ± 10 nm after 3 h at 1000 °C.



neous material even at a temperature as low as 750 °C, while Guo et al.9,10 used a solid-state pyrolysis reaction method for doping other elements effectively or coating carbon to increase the conductivity. Although these mixed metal oxides have been synthesized by a number of routes, there is limited information about details of the growth of the particles during synthesis. Consequently, a thorough understanding of the growth of these materials is of fundamental importance, not only for its intrinsic scientific interest but also for its technological significance.15,16 Here, we have successfully utilized a two-step synthesis method involving coprecipitation and annealing to investigate the crystal growth behavior of Li[Ni1/3Co1/3Mn1/3]O2 displaying the α-NaFeO2 structure. The information revealed by this study will help develop guidelines to control the size and morphological features more precisely and will help create design strategies for cathodes that will enable efficient lithium ion transport and extended lifetime batteries.

INTRODUCTION LiCoO2 has been used as a cathode material for commercially available lithium-ion batteries due to its facile preparation and excellent cycling stability.1 However, the toxicity, high cost of cobalt, and charging limitations prohibit its long-term viability. In 2001, Ohzuku and Makimura2 introduced an alternative cathode material, Li[Ni1/3Co1/3Mn1/3]O2 (L333), because it combines the advantages of high specific capacity, low cost, long cyclic life, and both structural and thermal stability.3−6 Thus, it proves to be a promising candidate cathode in battery applications. Some of the factors that limit the performance of these materials (as well as other cathodes) are related to relatively large particle sizes, inhomogeneous particle size distributions, and low packing densities. The electrochemical performance of these materials has been reported to be dependent upon structure, composition, and voltage range during charge/discharge. Zhang et al.7 reported that smaller particle size and good homogeneity of materials may mainly account for the improved electrochemical properties of the L333. Therefore, the microstructure of the cathode material is a key factor in its performance, because it controls the packing density, diffusion rates, and pathways. The particle size and distributions are strongly dependent on the synthesis methods and annealing processes. Although lower temperature synthetic routes may yield smaller particles, which shorten diffusion path lengths, there is a reduction in the tap density. Lee et al.5 synthesized larger L333 particles (i.e., 10 μm) using coprecipitation and annealing to yield nearly uniform particles that had a tap density of 2.39 g/cm3. Additional synthesis processes8−11 and structural characterization methods12−14 of L333 materials have been utilized. Cho et al.8 reported the carbonate coprecipitation method, which could yield homoge© 2012 American Chemical Society



EXPERIMENTAL SECTION

L333 powders were prepared using a two-step method (i.e., coprecipitation and annealing). The precursor to L333, (Ni1/3Co1/3Mn1/3)(OH)2, was synthesized at room temperature through a coprecipitation method similar to those previously reported.5 Briefly, stoichiometric amounts of nickel, cobalt, and manganese nitrates (99.5% purity, Aldrich Chemical Co.) were thoroughly dissolved in deionized water to achieve a homogeneously mixed 100 mM solution. NaOH (2 M) was added (under stirring) to this solution to modify the pH to 10, thus inducing precipitation. The resultant suspension was subsequently stirred at 600 rpm under a Received: May 4, 2011 Revised: December 7, 2011 Published: January 3, 2012 1118

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123

Crystal Growth & Design

Article

nitrogen atmosphere at room temperature for 2 h. The (Ni1/3Co1/3Mn1/3)(OH)2 precipitates were washed with DI water 3 times and dried under vacuum at 70 °C overnight. The (Ni1/3Co1/3Mn1/3)(OH)2 precursor was then mixed with a 5% excess of LiOH·H2O and subsequently ground with a mortar and pestle. The mixture was placed in an alumina boat and annealed in a preheated tube furnace at different temperatures (i.e., 600−1000 °C) for specific durations (3−24 h) in air, which also helped in the oxidation of Co and Mn. Phase identification was determined by X-ray diffraction analysis (XRD, Philips X’Pert) using Cu Kα radiation. From the resulting XRD patterns, crystallite diameters were calculated based on the Scherer formula.17 Particle sizes and morphologies were observed using a scanning electron microscope (SEM, FEI XL30) at 10 kV accelerating voltage. The SEM was equipped with an energy-dispersive X-ray spectrometer (EDS) that was used to confirm sample stoichiometry (based on examination of five different regions per sample). A transmission electron microscope (TEM, FEI T20), operated at 200 kV, was used to identify crystallite size, morphology, and phase. Thermal and mass analyses were performed using a thermal gravimetric analyzer/differential scanning calorimetry (TGA/DSC, Mettler Toledo) from room temperature to 1000 °C at a heating rate of 10 °C/min to investigate material behavior during annealing.

Figure 2. XRD patterns of Li[Ni1/3Co1/3Mn1/3]O2 heated for 12 h at (a) 200, (b) 300, (c) 400, (d) 500, (e) 600, (f) 700, (g) 800, and (h) 900 °C.



maximum (FWHM) of the XRD peaks. The reported known indication of an ordered hexagonal layered structure,1,18,19 which is confirmed by the presence of doublet peaks for the (006)/(102) at 37° (2θ) and the (108)/(110) reflections at 65° (2θ), is present for the samples heated at or above 700 °C. This is an indicator of the ordering of the L333 layered structure that enables the reduction of cationic mixing during intercalation and deintercalation of lithium ions inside the host matrix, leading to an increase in the electrochemical performance of this cathode material.19 Lattice parameters were calculated from the refinement of the XRD data, which are shown in Table 1. The ratios of c/a for

RESULTS AND DISCUSSION XRD. Figure 1 depicts an XRD pattern of the (Ni1/3Co1/3Mn1/3)(OH)2 precursor after coprecipitation

Table 1. Calculated Structure Parameters for Synthesized Li[Ni1/3Co1/3Mn1/3]O2 parameters

700 °C

800 °C

900 °C

a (Å) c (Å) c/a

14.133 2.852 4.966

14.158 2.851 4.969

14.176 2.853 4.969

all the samples are larger than 4.899, which confirms the layered structure based on work by Ngala et al.20,21 They reported that the c/a ratio of the lattice parameter is a direct measurement of the deviation of the lattice from a perfect cubic close-packed lattice. Since the ideal cubic close-packed lattice has a c/a value of 4.899, a larger c/a ratio indicates a more ordered layered structure. For all of our samples (Table 1), the ratios are larger than 4.960, and thus, they have the well-ordered layered structure. In addition, according to the Wang et al.,22 the R factor (R = I102 + I006/I101) is related to the hexagonal ordering (i.e., the lower R value, the better hexagonal ordering). All of our samples have an R value less than 0.45, which is indicative of a good hexagonal ordering of the lattice. TG−DSC. The thermochemical properties (Figure 3) of synthesized L333 materials were investigated by thermogravimetric analysis and differential scanning calorimetry (TGA/ DSC). As the specimen is heated, there is an endothermic reaction at 90 °C accompanied by a ∼13% weight loss, which can be attributed to water desorption (consistent with the observed mass loss). The exothermic peak observed at 259 °C

Figure 1. XRD pattern of [Ni1/3Co1/3Mn1/3](OH)2 precursor.

under N2 at room temperature. Indexing of this diffraction pattern confirms the formation of the metal hydroxide structure.18 All peaks were indexed on the basis of the hexagonal Mn(OH)2 structure (JCPDS 12-0696). The metal hydroxide precursor was mixed with LiOH·H2O and annealed at different temperatures. Figure 2 shows diffraction patterns of Li[Ni1/3Co1/3Mn1/3]O2 annealed in air at increasing temperatures for 12 h durations. Indexing of these diffraction patterns confirmed the formation of the layered L333 structure. All peaks from L333 were indexed on the basis of the hexagonal α-NaFeO2 structure with the R3̅m space group (JCPDS 82-1495). XRD analysis of samples annealed as low as 200 °C for 12 h indicated the presence of crystalline Li[Ni1/3Co1/3Mn1/3]O2. However, small quantities of both Li2MnO3 and LiMnO2 were also detected. After annealing at 600 °C, the impurity phases were no longer detected. Above 700 °C, significant crystal growth occurs, as indicated by the reduced full width at half1119

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123

Crystal Growth & Design

Article

Figure 3. TG−DSC of the mixture of [Ni1/3Co1/3Mn1/3](OH)2 and LiOH·H2O.

region II (750−900 °C) is larger. Finally, above 900 °C (region III) significant grain growth is observed. These three regions were divided based on the different slope of the grain growth curve, increasing from 0.05 ± 0.01 in region I to 1.54 ± 0.02 in region II, and finally up to 6.09 ± 1.11 in region III. This increased grain growth behavior in regions II and III is partially attributed to the chemical potential gradient of the lithium ion between the surface and the interior of the particles due to lithium oxide evaporation (as observed in our thermal analyses). This may also act as a driving force for mass transport, which results in increased particle size. Similar effects were observed by Antolini27 on the grain growth of LiNiO2 at high temperatures. Figure 6 depicts bright-field TEM micrographs of Li[Ni1/3Co1/3Mn1/3]O2 annealed at 800 °C for different durations. The crystal diameters increase from 57 ± 3 nm (0.5 h) to 169 ± 11 nm (24 h) and demonstrate good uniformity. The micrographs also confirm the single-crystalline nature of the particles observed by SEM.

is due to the decomposition of hydroxyls,23 and correspondingly the crystallization process is initiated with concurrent dehydration reactions. This can be seen from the corresponding diffraction pattern of the sample heated at 200 °C for 12 h. At the same time, there is another endotherm at 445 °C, which is related to LiOH melting. 24 Furthermore, the third endothermic peak at 750 °C is probably due to the completion of the ordering of the hexagonal lattice. This is similar to work reported by Delmas et al., who suggested that synthesis of LiNiO2 systems at temperatures higher than 700 °C produced a 2D-type cationic ordered structure,25 and also by Jouybari et al., who reported a required annealing temperature of 700 °C during their synthesis of LiNi0.8Co0.2O2.26 This is concurrent with observations from XRD analyses, which depicts an ordered layered structure at and above 700 °C. No additional reactions are observed at higher temperatures; however, there is a small weight loss (∼2.5%) from 800 to 1000 °C likely attributed to the evaporation of Li2O.23 Microstructural Analysis. Figure 4 depicts SEM micrographs of Li[Ni1/3Co1/3Mn1/3]O2 specimens annealed in air at different temperatures for 3 h. As mentioned previously, we observe significant crystal growth between 800 and 1000 °C. Specimens heated to 600 °C demonstrate a nanocrystalline nature (i.e., crystal diameters less than 20 nm, as determined by transmission electron microscopy and X-ray diffraction). These crystals grow in size with increased temperature up to 250 nm by 900 °C and eventually to 850 nm by 1000 °C. The EDS spectra of the annealed powder (800 °C, 3 h, placed on a silicon wafer) in Figure 4 shows the presence of Ni, Co, and Mn in a 1:1:1 ratio with O, C, and Si (wafer) present. No additional impurities were detected. The combination of XRD and EDS results confirmed that the final products were Li[Ni1/3Co1/3Mn1/3]O2. Figure 5 shows the increase in particle diameter (as measured by both SEM and TEM) with temperature temperature for samples annealed for 3 h. It is clear that there are three growth regions between 600 and 1000 °C. Region I, from 600 to 750 °C, exhibits minimal crystal growth, while the growth rate in



DISCUSSION

Based on SEM and TEM observations, it appears that there are three separate growth regimes (highlighted in Figure 5) during the L333 annealing process. In order to investigate the crystal growth mechanisms within these regions, we calculated the crystal diameters of the specimens either through X-ray diffraction measurements (i.e., via the Scherer equation17 utilizing the full width at half-maximum, FWHM) or through direct measurements of crystals in SEM images (at crystal diameters greater than 100 nm). We plot the ln(crystal diameter) vs ln(time) for samples annealed at different temperatures, which are shown in Figure 7. By fitting these curves, we obtain a linear relationship between ln(time) and ln(crystal diameter) which, through theory developed by Lifshitz, Slyozov, and Wagner (LSW),28,29 describes the growth of the average crystallite size, G, with time, t, by eq 1: 1120

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123

Crystal Growth & Design

Article

Figure 4. SEM micrographs of Li[Ni1/3Co1/3Mn1/3]O2 annealed at (a) 600, (b) 700, (c) 800, (d) 900, (e) 950, or (f) 1000 °C for 3 h with an EDS spectrum (bottom) from sample c.

Dn − D0n = kt

Since D0 is the initial grain diameter, it is found that when ln D is plotted versus ln(time), a straight line is obtained (Figure 7), as stated above. We obtain both 1/n and ln k values from the slope and intercepts of the observed straight line, respectively. Frequently, the slope of the curve plotted in this way is less than 0.5, usually falling between 0.1 and 0.5. This may be due to a small initial crystallite size, D0 (in our case, it was calculated to be 9 nm from L333 formed at 400 °C) or due to the presence of inclusions or solute segregation, which inhibits grain growth, especially at low temperatures.31 Table 2 depicts the values of the rate-controlling mechanism exponent, 1/n, and the crystal growth rate, ln k. As expected, we can see with the increase of temperature, the values of both 1/n and ln k increase.

(1)

Here, D0 is the initial crystallite size, k is a temperaturedependent parameter of the crystal growth rate, and n is an exponent that describes the rate-controlling mechanism. As a generalization, eq 1 can be rearranged and written as eq 230−32

D − D0 = kt 1/ n

(2)

Taking the natural log of eq 2, we obtain eq 3:

ln(D − D0) = ln(k + 1)/n ln t

(3) 1121

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123

Crystal Growth & Design

Article

Figure 5. Crystal diameters of L333 annealed for 3 h at temperatures from 600 to 1000 °C. Three separate growth regions are observed.

As mentioned previously, we observe three primary growth regions in our annealing schedule. In region 1 (i.e., 600−750 °C), there is almost no grain growth as well as a small ratecontrolling mechanism exponent 1/n data (1/n = 0.2). Here, we suspect that the crystal growth mechanism at this temperature range would likely be due to surface diffusion. In regions II and III, we observe different grain growth behavior.

Figure 7. Relationship between crystal diameter and annealing time for samples annealed from 700 to 1000 °C.

In the temperature range from 800 to 900 °C, we determined that n = 3, indicating that the growth is controlled by the

Figure 6. Bright-field TEM micrographs of Li[Ni1/3Co1/3Mn1/3]O2 annealed at 800 °C for (a) 0.5, (b) 1, (c) 3, and (d) 6 h. Fast Fourier transforms are inserted in panels a and c. 1122

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123

Crystal Growth & Design

Article

(4) Cho, T. H.; Park, S. M.; Yoshio, M. Chem. Lett. 2004, 33, 704− 705. (5) Lee, M. H.; Kang, Y.; Myung, S. T.; Sun, Y. K. Electrochim. Acta 2004, 50, 939−948. (6) Yabuuchi, N.; Ohzuku, T. J. Power Sources 2003, 119, 171−174. (7) Zhang, L.; Wang, X.; Muta, T.; Li, D.; Noguchi, H.; Yoshio, M.; Ma, R.; Takada, K.; Sasaki, T. J. Power Sources 2006, 162, 629−635. (8) Cho, T. H.; Park, S. M.; Yoshio, M.; Hirai, T.; Hideshima, Y. J. Power Sources 2005, 142, 306−312. (9) Guo, J.; Jiao, L. F.; Yuan, H. T.; Li, H. X.; Zhang, M.; Wang, Y. M. Electrochim. Acta 2006, 51, 3731−3735. (10) Guo, R.; Shi, P. F.; Cheng, X. Q.; Du, C. Y. J. Alloys Compd. 2009, 473, 53−59. (11) Kim, S. K.; Jeong, W. T.; Lee, H. K.; Shim, J. Int. J. Electrochem. Sci. 2008, 3, 1504−1511. (12) Cahill, L. S.; Yin, S. C.; Samoson, A.; Heinmaa, I.; Nazar, L. F.; Goward, G. R. Chem. Mater. 2005, 17, 6560−6566. (13) Choi, J.; Manthiram, A. J. Electrochem. Soc. 2005, 152, A1714− A1718. (14) Idemoto, Y.; Matsui, T. Solid State Ionics 2008, 179, 625−635. (15) Atkinson, H. V. Acta Metall. 1988, 36, 469−491. (16) Weaire, D.; Rivier, N. Contemp. Phys. 1984, 25, 59−99. (17) Leroy Alexander, H. P. K J. Appl. Phys. 1950, 21, 137−142. (18) Sun, Y. K.; Myung, S. T.; Kim, M. H.; Prakash, J.; Amine, K. J. Am. Chem. Soc. 2005, 127, 13411−13418. (19) Reimers, J. N.; Dahn, J. R.; Greedan, J. E.; Stager, C. V.; Liu, G.; Davidson, I.; Vonsacken, U. J. Solid State Chem. 1993, 102, 542−552. (20) Ngala, J. K.; Chernova, N. A.; Ma, M.; Mamak, M.; Zavalij, P. Y.; Whittingham, M. S. J. Mater. Chem. 2004, 14, 214−220. (21) Kim, G. H.; Myung, S. T.; Bang, H. J.; Prakash, J.; Sun, Y. K. Electrochem. Solid State Lett. 2004, 7, A477−A480. (22) Luo, X. F.; Wang, X. Y.; Liao, L.; Gamboa, S.; Sebastian, P. J. J. Power Sources 2006, 158, 654−658. (23) Li, H.; Chen, G.; Zhang, B.; Xu, J. Solid State Commun. 2008, 146, 115−120. (24) Grigorova, E.; Mandzhukova, T. S.; Khristov, M.; Yoncheva, M.; Stoyanova, R.; Zhecheva, E. J. Mater. Sci. 2011, 46, 7106−7113. (25) Delmas, C.; Menetrier, M.; Croguennec, L.; Saadoune, I.; Rougier, A.; Pouillerie, C.; Prado, G.; Grune, M.; Fournes, L. Electrochim. Acta 1999, 45, 243−253. (26) Jouybari, Y. H.; Asgari, S. J. Power Sources 2011, 196, 337−342. (27) Antolini, E. Nuovo Cimento Soc. Ital. Fis., D 1998, 20, 1735− 1743. (28) Lifshitz, I. M.; Slyozov, V. V. J. Phys. Chem. Solids 1961, 19, 35− 50. (29) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89−94. (30) Hillert, M. Acta Metall. 1965, 13, 227−238. (31) Greenwood, G. W. Acta Metall. 1956, 4, 243−248. (32) Higgins, G. T. Met. Sci. 1974, 8, 143−150. (33) Huang, F.; Zhang, H. Z.; Banfield, J. F. Nano Lett. 2003, 3, 373− 378. (34) Kisailus, D.; Choi, J. H.; Lange, F. F. J. Cryst. Growth 2003, 249, 106−120. (35) Schoeman, B. J.; Sterte, J.; Otterstedt, J. E. Zeolites 1994, 14, 568−575. (36) Dannenberg, R.; Stach, E.; Groza, J. R.; Dresser, B. J. Thin Solid Films 2000, 379, 133−138. (37) Gil, F. X.; Rodriguez, D.; Planell, J. A. Scr. Mater. 1995, 33, 1361−1366. (38) Kirchner, H. O. Metall. Trans. 1971, 2, 2861−2864. (39) Speight, M. V. Acta Metall. 1968, 16, 133−135. (40) Wagner, C. Z. Elektrochem. 1961, 65, 581−591.

Table 2. Grain Growth Parameters at Different Temperatures temp (°C) 700 750 800 850 900 950 1000

1/n 0.21 0.20 0.35 0.35 0.39 0.44 0.48

± ± ± ± ± ± ±

0.05 0.02 0.03 0.02 0.02 0.02 0.05

ln k 1.81 2.23 3.96 4.65 4.78 5.68 5.97

± ± ± ± ± ± ±

0.13 0.06 0.07 0.05 0.06 0.05 0.11

volume diffusion of ions within the matrix. At temperatures from 900 to 1000 °C, n = 2, indicating that crystal growth is controlled by the diffusion of ions along the matrix−particle boundary.33−40 We also observe triple grain junctions at 950 °C (Figure 4e) that are indicative of a metastable equilibrium for the L333 system.28 Based on our observations, we find that the atomic diffusion across the grain boundary is a simple activated process. Thus, k can be written as

k = k 0 exp( − Ea /(RT ))

(4)

ln k = ln k 0 − Ea /(RT )

(5)

where Ea is the activation energy for the process, T is the absolute temperature, and R is the gas constant. Thus, the activation energy for crystal growth can be derived from the ln k vs 1/T plot (eq 5). We calculated the activation energy in the temperature range of 800−900 °C to be 109 ± 17 kJ/mol and in the temperature range of 900−1000 °C to be 162 ± 25 kJ/ mol. The calculated activation energies at these temperatures match the observed diffusion mechanisms within these temperature ranges.



CONCLUSIONS We have successfully synthesized Li[Ni1/3Co1/3Mn1/3]O2 through a two-step coprecipitation and annealing method, investigating the crystal growth processes at different temperatures. At temperatures below 750 °C, surface diffusion is expected to be the dominant growth mechanism based on minimal crystal growth and a small rate-controlling exponent. As the annealing temperature is increased (i.e., 800−900 °C), crystal growth is controlled by volume diffusion, while grain boundary diffusion is the predominant mechanism for the temperature range of 900−1000 °C. Based on these results, we can establish synthesis and annealing parameters to optimize crystal size and, perhaps, control over morphological features, both of which are crucial toward enhancing rate capabilities, energy densities, and cycling stability of future Li-ion batteries.



AUTHOR INFORMATION

Present Addresses #

Molecular Foundry, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States. § Department of Chemical Engineering, University of Delaware, Newark, DE 19716, United States.



REFERENCES

(1) Ozawa, K. Solid State Ionics 1994, 69, 212−221. (2) Ohzuku, T.; Makimura, Y. Chem. Lett. 2001, 642−643. (3) Shaju, K. M.; Rao, G. V. S.; Chowdari, B. V. R. Electrochim. Acta 2002, 48, 145−151. 1123

dx.doi.org/10.1021/cg200565n | Cryst. Growth Des. 2012, 12, 1118−1123