Spinel-Li3.5 + xTi5O12 coated LiMn2O4 with high surface Mn valence for an enhanced cycling performance at high temperature

Electrochemistry Communications 31 (2013) 92–95

Contents lists available at SciVerse ScienceDirect

Electrochemistry Communications journal homepage: www.elsevier.com/locate/elecom

Spinel-Li3.5 + xTi5O12 coated LiMn2O4 with high surface Mn valence for an enhanced cycling performance at high temperature J. Yao a,⁎, C. Shen a, P. Zhang a, C.A. Ma a, D.H. Gregory b, L. Wang a,⁎ a State Key Laboratory Breeding Base of Green Chemistry-Synthesis Technology, College of Chemical Engineering and Material Science, Zhejiang University of Technology, Hangzhou, Zhejiang, PR China b WestCHEM, School of Chemistry, University of Glasgow, Glasgow G12 8QQ, UK

a r t i c l e

i n f o

Article history: Received 4 March 2013 Accepted 11 March 2013 Available online 17 March 2013 Keywords: LiMn2O4 Spinel Li4Ti5O12 Coating Mn valence

a b s t r a c t Li3.5 + xTi5O12 and Li4Ti5O12 layers coated LiMn2O4 were prepared via sol–gel route, with subsequent annealing at 750 °C. XRD studies showed coating did not change the structure of LiMn2O4. TEM studies showed coating layers were formed around LiMn2O4 particles, both corresponding spinel LiTiO phase. XPS studies showed the valence of surface Mn for the Li3.5Ti5O12-coated-LiMn2O4 increased after annealing. This sample also demonstrated a favorable cycling performance. © 2013 Elsevier B.V. All rights reserved.

1. Introduction At the origin of capacity fading mechanism in LiMn2O4 cathodes is the well-established Jahn–Teller effect of high spin Mn 3+ [1]. As a result, during the (de)lithiation process, the average valence of Mn atoms, particularly at the surface, would be easily reduced to less than + 3.5, due to the accumulation of Li + [2], thus causing local structural damage leading to unfavorable cycling performance. Therefore, a higher oxidation state of surface Mn would be crucial to maintain the stability for the LiMn2O4. Another approach that has been explored is surface coating by using oxides, e.g. ZrO2, SiO2, and Al2O3 via a sol–gel method as previously reported [3,4]. Li4Ti5O12 is a potentially competitive coating candidate, as it possesses a higher chemical diffusion coefficient (10 −6 cm 2 s −1) than that of LiMn2O4 (10 −10–10 −12 cm 2 s −1), and shares the same spinel structure making it structurally compatible with the targeting LiMn2O4. Li4Ti5O12 is a zero-strain material during the insertion/extraction of lithium ions. Li4Ti5O12 also has a high thermodynamic stability [5]. And therefore, the successful combination of abovementioned two advantages would greatly improve the cycling ability of LiMn2O4. However, no report on producing spinel LiTiO coated LiMn2O4 with high surface Mn valence was found. Herein we report a novel in-situ strategy to prepare a spinel LiTiO coated LiMn2O4, where Mn atoms at the surface possess a higher formal oxidation state than +3.5. To realize this, we used a Li deficient spinel Li3.5Ti5O12 (rather than Li4Ti5O12) to coat the LiMn2O4. With an ⁎ Corresponding authors. Tel.: +86 571 88320611; fax: +86 571 88320832. E-mail addresses: [email protected] (J. Yao), [email protected] (L. Wang). 1388-2481/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.elecom.2013.03.014

appropriate annealing treatment (which is also a necessary step to form LiTiO coating), Li+ ions at surface of LiMn2O4 may diffuse from Li saturated region (of LiMn2O4) into the relatively Li unsaturated region of Li3.5Ti5O12. Such Li diffusion would not likely occur for the Li saturated Li4Ti5O12 coated LiMn2O4. As a result, the oxidation state of Mn at the surface of the spinel manganese is increased due to the departure of Li ions. Using this strategy, we could obtain a spinel LiTiO coated LiMn2O4 with higher oxidation state for the surface Mn atoms, and both factors contributed greatly to the improved high temperature cycling performance.

2. Experiment section 2.1. Sample preparation LiMn2O4 powder was synthesized via the solid-state reaction of MnO2 and Li2CO3 [6]. To coat LiMn2O4 with Li4Ti5O12 and Li3.5Ti5O12, LiCH3COOLi·2H2O and Ti(C4H9O)4 with a stoichiometric cationic ratio (e.g. 4:5 and 3.5:5) were dissolved in a solution contained ethanol and distilled water to form a clear solution. Then citric acid was added into the solution with stirring to obtain a sol. Then the as-prepared LiMn2O4 was homogeneously dispersed into coordinated solution. After hydrolyzing for 5 h, the resulting precipitate was filtered, washed several times with distilled water and dried at 120 °C for 2 h. Finally, the powder product was annealed in nitrogen at 750 °C for 5 h. For comparison, pristine LiMn2O4 was also heated at 750 °C for 5 h, and sample obtained is denoted as SA-1. The Li4Ti5O12 coated sample is denoted as SA-2 and Li3.5Ti5O12 coated sample is denoted as SA-3.

J. Yao et al. / Electrochemistry Communications 31 (2013) 92–95

2.2. Characterization The phase compositions of the samples were identified by powder X-ray diffraction (XRD, PANalytical, X'Pert Pro) using Cu Kα radiation (λ = 1.5408 Å). The microstructure and morphology of products were investigated by scanning electronic microscopy (SEM, Hitachi S-4700) and transmission electronic microscopy (TEM, Tecnai G2 F30 S-Twin). X-ray photoelectron spectroscopy (XPS, Kratos AXIS Ultra DLD) was employed to probe the surface for Mn and Ti species and to determine their oxidation states. Energy-dispersive X-ray spectroscopy (EDS) was used in conjunction with TEM to determine the elemental composition of products from both area and point scans.

2.3. Electrochemical testing The as-prepared materials were evaluated in a CR2032-type coin cell assembled in an argon filled glove box. The active materials, acetylene black and polyvinylidene fluoride (PVDF) were ground in 1-methyl-2-pyrrolidinone (NMP) solution in a weight ratio of 80:12:8 to form slurry, which was then casted onto an aluminum foil current collector and dried at 120 °C for 10 h under vacuum. The thickness of electrode composite is ca. 60 μm. The separator is Celgard 2400. Li metal as the counter electrode and an electrolyte of 1 mol L −1 LiPF6 in a solvent mixture of ethylene carbonate (EC), dimethyl carbonate (DMC) and ethylmethyl carbonate (EMC) (1:1:1, v/v/v) were used. Galvanostatic charge–discharge tests were performed over a voltage range of 3.3–4.3 V (vs. Li/Li +) with a Land CT2001A battery test system at 55 °C. Electrochemical impedance spectroscopy (EIS)

93

was recorded with a Solartron Impedance/gain phase analyzer (model SI 1260) coupled to a potentiostat (SI 1287) over a frequency range from 10 −2 Hz to 10 5 Hz at the open circuit voltage at room temperature.

3. Results and discussion HRTEM studies of the prepared samples revealed that layers with a thickness of ca. 3 nm were coated onto the bare surface of LiMn2O4 (Fig. 1A). For both coated samples, the bulk LiMn2O4 material component shows good crystallinity with lattice fringes extending to the grain boundary. For instance, the distances between the neighboring lattice fringes are approximately 0.4767 nm and 0.2505 nm which correspond to the d-spacings for the (111) and (311) planes respectively in the Li–Mn–O spinel phase. EDX results for SA-2 and 3 from selected areas indicate that the Ti to Mn ratio is about 2:98. The coated layer in SA-3 exhibits lattice fringes with spacings of 0.251 nm and 0.212 nm, corresponding well to the d-spacings for the (311) and (400) planes respectively for a LiTiO spinel phase (Fig. 1A). All of the diffraction peaks corresponded to a spinel structure (space group Fd-3m (no. 227)) and in good agreement with ICSD PDF card no. 35-0782. No LiTiO phase was detected indicating that there is no structural change of surface-modified LiMn2O4. As shown in the inset of Fig. 1B, the LiMn2O4 spinel phase diffraction peaks exhibit a discernible shift to lower 2θ following the annealing process at 750 °C, indicating the increase in the cubic lattice parameter. Rietveld refinement using the Reflex module of Materials Studio found the sample SA-3 is with the largest a value of 8.2653(3) Å. Indeed, Xia et al. found that the lattice

(A) (a)

(b)

(c)

Intensity a.u.

111

Intensity a.u.

Intensity a.u.

(B)

35.9 36.0 36.1 36.2 36.3 36.4

43.7 43.8 43.9 44.0 44.1 44.2

2 (degree)

2 (degree)

(311)

(400)

311

400

222

440 331

333

531

SA-3 SA-2 SA-1

10

20

30

40

50

60

70

80

2 (degree) Fig. 1. (A) HRTEM images of (a) SA-1, (b) SA-2 and (c) SA-3. (B) Powder XRD patterns of the pristine (SA-1) and coated (SA-2, SA-3) LiMn2O4 samples.

94

J. Yao et al. / Electrochemistry Communications 31 (2013) 92–95

SA-1

Background Mn4+ Mn3+

2p 3/2

4+

Raw Intensity

2p 3/2

Tetrahedral coordination

SA-3

3+

Mn /Mn =47.1/52.9%

SA-2 2p 3/2 SA-2

4+

SA-3

(A)

2p 3/2

(B)

SA-1

3+

Mn /Mn =47.5/52.5%

456

457

458

459

460

461

B.E.(eV) Mn4+/Mn3+=65.3/34.7%

638 640 642 644 646 648

B.E.(eV) Fig. 2. Deconvoluted profile of (A) Mn 2p3/2 and (B) Ti 2p3/2 XPS spectra of the surface of SA-1, 2 and 3.

constants of LiMn2O4 increase with decreasing Li content [7]. We speculate that, for our SA-3 sample (Li3.5Ti5O12 coated LiMn2O4), part of Li at LiMn2O4 surface may diffuse from Li saturated region to the Li unsaturated Li3.5Ti4O12. If this is the case, the departure of Li ions at LiMn2O4 surface would consequently result in an increase of Mn oxidation state in original surface LiMn2O4. To verify this hypothesis we further perform the XPS analysis to check the oxidation state of surface Mn valence. The oxidation state of the manganese at the surface was determined from XPS data by the curve fitting of the Mn 2p spectral peaks [8]. The experimental peak shape for Mn 2p3/2 was modeled by employing multiple-splitting patterns derived for Mn3+ and Mn4+ at binding energies of 641.7, and 642.7 eV from the standard compounds Mn2O3 and MnO2, respectively and Fig. 2A shows the fit of the models to the experimental spectra for SA-1, SA-2 and SA-3 respectively. The surface of the SA-1 and SA-2 samples consists of almost equal amounts of Mn4+ and Mn3+. By contrast, however, SA-3 exhibits and Mn4+: Mn3+ ratio of ca. 65.3:34.7%. The difference in the surface Mn4+: Mn3+ ratio may be due to the Li+ extraction from the host LiMn2O4 surface structure. This is because SA-2 has lithium-saturated coating layer (of Li4Ti5O12), whereas the surface of the SA-3 is lithium-deficient (Li3.5Ti5O12). Considering the XPS spectra for titanium in more detail (Fig. 2B), for Ti in SA-3, The Ti\O bond is weakened (from 458.3 eV to 457.9 eV), possibly due to the lithium ion diffusion to form spinel

4.4 120

Discharge capacity / mAh g-1

4.2

Potential / V vs. Li+/Li

4.0 3.8 3.6 3.4 3.2 3.0

(A)

SA-1 SA-2 SA-3

2.8 2.6

100 80 60 40

SA-1 SA-2 SA-3

20

(B)

0 0

20

40

60

80

100

0

120

10

20

Specific Capacity / mAh g-1

40

50

-160

-50 SA-2 -40 SA-3 -30 SA-1

-20

(C)

-10

SA-1

-140

Z'' imaginary (ohm)

Z'' imaginary (ohm)

30

Cycle number

-120 -100

SA-2

-80 -60

SA-3

-40 -20

(D)

0

0 0

20

40

50

60

Z' real (ohm)

80

100

120

0

50

100

150

200

250

300

350

400

Z' real (ohm)

Fig. 3. (A) Voltage profiles of pristine and coated spinel LiMn2O4 samples at 1C between 3.3 and 4.3 V in coin-type half cells at 55 °C. (B) Discharge capacities (with error bars) of the pristine and coated spinel LiMn2O4 samples vs. the cycle number at 1C, the representative electrochemical impedance spectra of pristine and coated LiMn2O4 electrodes at (C) 5th and (D) 50th cycles, the inserted is the equivalent circuit used to fit the EIS.

J. Yao et al. / Electrochemistry Communications 31 (2013) 92–95

LiTiO phase or the less Li ions presented in the structure. There is a second peak (higher BE) at 460.2 eV near the main peak at 458.0 eV, which implies a change of chemical environment for the Ti ions in the structure [9], and indicates the possible formation of a Ti-containing spinel phase on the surface of LiMn2O4 upon higher temperature annealing. The reprehensive charge and discharge curves of uncoated and coated-LiMn2O4 under a current rate of 1C (one lithium per formula unit in 1 h) at 55 °C are shown in Fig. 3A. All of the electrodes exhibit two distinct plateaus around 4-voltage, which means that the LiTiO-coating does not fundamentally change the intrinsic electrochemical properties of LiMn2O4 such as the charge/discharge behavior [10]. After LiTiO coating, however, the polarization between the charge and discharge plateaus is increased compared to the pristine sample, indicating that the kinetics of the LiMn2O4 are indeed changed after coating. As shown in Fig. 3B is the mean capacities for prepared samples (n = 5), with error bars in the cycles. The unmodified LiMn2O4 SA-1 provided a discharge capacity of 116.2 mAh g −1 initially at a rate of 1C, while SA-2 and SA-3 exhibited slight lower discharge capacities of 112.7 mAh·g −1 and 114.4 mAh·g −1 respectively. Fig. 3B shows SA-3 is with the better cycling retention than that of SA-2, which should be attributed to high surface Mn valence [1], which leads to a stabilized surface structure. A similar result also appeared in an earlier report from Kim, where the ceramic coated spinel Li1.05Al0.05Mn1.9O4 and Li1.05Ni0.05Mn1.9O4 (prepared prior to coating) demonstrated an improved cycling performance than that of ceramic coated un-doped LiMn2O4. This could be attributed to the high Mn valence induced by Ni and Al doping [3], as all the samples were coated. The difference in our work is that we achieved the high Mn valence during one step annealing process without scarifying the capacity by foreign ion doping. EIS has been performed in an attempt to rationalize the cycle life and rate capability characteristics of the pristine and LiTiO-coated LiMn2O4 cathode materials. From the EIS spectra in Fig. 3C and D, it is evident that SA-1 displays two clear semicircles after the 5th cycle, while after the 50th cycle these two features almost combine into one semicircle and the total resistance increases considerably about 5 folds from 52 (18, which is the deviation compared to the mean value, n = 5) Ω to 260 (31) Ω, which suggests that the kinetics at the interface have changed. The mid-frequency impedance arc may result from a series of processes at the solid electrolyte interface, resulting in an increased resistance to Li +-ion transport through the surface film [11]. The EIS spectra of SA-2 show only one semicircle which is indicative of the combination of two semicircles (high and low frequencies), and the initial resistance is higher than that of SA-1. However, after the 50th cycle, the resistance for SA-2 is much lower. This illustrates that samples with coating layer can maintain

95

stable surface and SEI which make the host materials more sustainable. The spectra of SA-3 clearly show two semicircles both after the 5th and 50th cycles. The resistance increases about 47% from 62 (11) Ω to 118 (15) Ω; a significantly lower change than that seen for SA-1. This is because that spinel LiTiO coating layer can suppress the acid attack (often result of Mn loss into electrolyte) and avoid the reduction of electrolyte on the Mn sites and formation of undesired SEI layer (usually occurring below 0.7 V Li/Li +) [2,4]. It would appear, therefore, that the LiTiO-incorporated spinel cathode materials exhibit a decreased SEI resistance and suppress the reaction between the cathode surface and electrolyte. 4. Conclusion A novel surface modification method was designed to improve the structural and electrochemical performances of the spinel-structured cathode material, LiMn2O4. Using a Li deficient Li3.5Ti5O12 layer coated LiMn2O4 as precursor, a spinel LiTiO phase coated LiMn2O4 with high surface Mn valence were prepared by annealing at 750 °C. This material showed an improved cycling performance with smaller resistance compared to the Li saturated Li4Ti5O12 layer coated LiMn2O4. Acknowledgments This work was supported by Zhejiang Provincial Natural Science Foundation of China (Grant No. LQ12B01003), Zhejiang University of Technology start-up fund (Grant No.101009529), the State Key Development Program for Basic Research of china (Grant No. 2007CB216409). References [1] J.-W. Song, C.C. Nguyen, H. Choi, K.-H. Lee, K.-H. Han, Y.-J. Kim, S. Choy, S.-W. Song, Journal of the Electrochemical Society 158 (2011) A458. [2] K.A. Walz, C.S. Johnson, J. Genthe, L.C. Stoiber, W.A. Zeltner, M.A. Anderson, M.M. Thackeray, Journal of Power Sources 195 (2010) 4943. [3] J.S. Kim, C.S. Johnson, J.T. Vaughey, S.A. Hackney, K.A. Walz, W.A. Zeltner, M.A. Anderson, M.M. Thackeray, Journal of the Electrochemical Society 151 (2004) A1755. [4] T.F. Yi, Y. Xie, Y.R. Zhu, R.S. Zhu, H.Y. Shen, Journal of Power Sources 222 (2013) 448. [5] T.F. Yi, Y.R. Zhu, X.D. Zhu, J. Shu, C.B. Yue, A.N. Zhou, Ionics 15 (2009) 779. [6] J. Yao, C. Shen, P. Zhang, D.H. Gregory, L. Wang, Journal of Physics and Chemistry of Solids 73 (2012) 1390. [7] Y. Xia, M. Yoshio, Journal of Power Sources 57 (1995) 125. [8] Y.J. Wei, L.Y. Yan, C.Z. Wang, X.G. Xu, F. Wu, G. Chen, The Journal of Physical Chemistry. B 108 (2004) 18547. [9] J.C. Yu, L. Zhang, Z. Zheng, J. Zhao, Chemistry of Materials 15 (2003) 2280. [10] K.-S. Lee, S.-T. Myung, H.-G. Jung, J.K. Lee, Y.-K. Sun, Electrochimica Acta 55 (2010) 8397. [11] D. Arumugam, G.P. Kalaignan, Electrochimica Acta 55 (2010) 8709.