2011年，富锂材料，Li[Li0.2Mn0.56Ni0.16Co0.08]O2 with improved rate capability

Journal of Power Sources 196 (2011) 4821–4825 Contents lists available at ScienceDirect Journal of Power Sources journa l homepage: www.e lsev ier .com Short communication Synthe th Li[Li0.2M e c J. Li, R. Kl S. P Institute of Phy a r t i c l Article history: Received 14 O Received in re 17 December 2 Accepted 3 Jan Available onlin Keywords: Lithium-ion ba Cathode mate Li[Li0.2Mn0.56N NMC Cycling perfor Rate capability Ni0.16 s bee and partic bviou capa hiev 1. Introduction Layered transition metal oxides have been investigated exten- sively as ca LiCoO2 is th very good hand, it als and safety pure electri capacity an point in the regard, the Co, Ni, etc.) materials in high capaci bers of this 0.6Li2MnO3 (also indi widely inv an initial tial range Li[Li0.2Mn0. because of ∗ Correspon E-mail add electrode polarization and improve the activation of the Li2MnO3 component, even if present in a small amount [13,14]. Neverthe- less, the long-term cycling performance of this latter material is 0378-7753/$ – doi:10.1016/j. thode materials for lithium-ion batteries. There into, e most important commercial material because of its electrochemical performance. However, on the other o suffers some drawbacks, such as high cost, toxicity, problems, which inhibit its further use in hybrid and c vehicles [1–3]. The identification of a cheaper, higher d safer layered cathode materials has been the focusing study of cathode materials in the last decade. In this solid solutions of layered Li2MnO3 and LiMO2 (M=Mn, have been shown as promising candidates for cathode lithium ion batteries since they exhibit a relatively ty, low cost and improved safety [2,4–11]. Two mem- family, Li[Li0.2Mn0.6Ni0.2]O2, which can be indicated as ·0.4LiMn0.5Ni0.5O2, and Li[Li0.2Mn0.54Ni0.13Co0.13]O2 cated as 0.6Li2MnO3·0.4LiMn1/3Ni1/3Co1/3O2), are estigated already. Usually, these materials deliver capacity of about 250mAhg−1 within the poten- extending from 2.0 to 4.8V [8–12]. Between them, 54Ni0.13Co0.13]O2 always shows better performance the presence of cobalt that significantly reduce the ding author. Tel.: +49 251 8336026; fax: +49 251 8336032. ress: stefano.passerini@uni-muenster.de (S. Passerini). still not very satisfactory. In addition, it still contains 13mol% of Co, which certainly represents an issue with regard to cost. Finally, most of the works on these materials report the electrochemical performance at very low current rates (such as C/20), which is of no help to ascertain their capability of matching with the requirement of today’s lithium ion batteries. In this work, we report on the synthesis and characterization of a novel cathode material with a substantially lower Co content. The material has a general formula of Li[Li0.2Mn0.56Ni0.16Co0.08]O2, which can also be seen as the solid solution of Li2MnO3 and LiMn0.4Ni0.4Co0.2O2 in 6:4 molar ratio. The structure, physical properties and electrochemical performance are reported in the following, with a special attention to the rate capabilities at high current rates. 2. Experimental Li[Li0.2Mn0.56Ni0.16Co0.08]O2 was synthesized by a solid-state reaction method from lithium hydroxide hydrate (LiOH·H2O Aldrich >98%) and manganese–nickel–cobalt hydroxide precursor [15]. This precursor was prepared by co-precipitating the aqueous solution of the three metal acetate salts (Mn, Ni, and Co; Aldrich >98%) in a stoichiometric ratio of 56:16:8, with lithium hydrox- ide. After extensive rinsing with distilled water, the precipitate see front matter © 2011 Elsevier B.V. All rights reserved. jpowsour.2011.01.006 sis and electrochemical performance of n0.56Ni0.16Co0.08]O2 with improved rat öpsch, M.C. Stan, S. Nowak, M. Kunze, M. Winter, sical Chemistry, University of Muenster, Corrensstr. 28/30, 48149 Muenster, Germany e i n f o ctober 2010 vised form 010 uary 2011 e 14 January 2011 ttery rial i0.16Co0.08]O2 mance a b s t r a c t The high voltage layered Li[Li0.2Mn0.56 Li2MnO3 and LiMn0.4Ni0.4Co0.2O2, ha temperature annealing at 900 ◦C. XRD constitutedof small andhomogenous capability. After the initial decay, no o different rates. Steady-state reversible at 5C and 110mAhg−1 at 20C were ac 2.5 and 4.8V at 20 ◦C. / locate / jpowsour e high voltage cathode material apability asserini ∗ Co0.08]O2 cathode material, which is a solid solution between n synthesized by co-precipitation method followed by high SEM characterizations proved that the as prepared powder is les (100–300nm),whichare seen toenhance thematerial rate s capacity fading was observed when cycling the material at cities of 220mAhg−1 at 0.2C, 190mAhg−1 at 1C, 155mAhg−1 ed in long-term cycle tests within the voltage cutoff limits of © 2011 Elsevier B.V. All rights reserved. Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 下划线 Administrator 下划线 Administrator 高亮 Administrator 高亮 Administrator 下划线 Administrator 下划线 4822 J. Li et al. / Journal of Power Sources 196 (2011) 4821–4825 Fig. 1. Riet Li[Li0.2Mn0.56N was dried u was then m milling for5 was anneal for 24h. The liquidnitro The com tents in th coupled pla ARCOS, Am terized by the Bruker 90◦. Lattice with GSAS with the he (SEM, EVO® solid phase the AccuPy (USA). Electrod position of (TIMCAL), a onto Al foil 2mgcm−2 for 30 s) wa reference e swagelok c electrolyte. current rat between 4. tester (USA couple. All expe reproducibi 3. Results Fig. 1 sho of Li[Li0.2M (9.9%) demo the low inte the crystal l ture with s and (104) p and (110) p EM pictures of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 material at different magnifica- ) ×5k; (b and c) ×50k. xagonal latticeparameterswere refined tobea=2.8560(1) A˚, 83(2) A˚, and volume=100.65(2) A˚3, with a c/a ratio of 4.99. veld refinement results for the XRD pattern of i0.16Co0.08]O2 material by GSAS. nder vacuum at 120 ◦C overnight. The dried material ixed with a stoichiometric amount of LiOH·H2O by ball h in thezirconia jar.After suitablegrinding, themixture ed in air at 480 ◦C for 5h, and then, as a pellet, at 900 ◦C final material was obtained by quenching the pellet in gen in order to “freeze” the layered solid-solutionphase. position in terms of lithium and transition metal con- e active material was determined by the inductively sma optical emission spectrometry (ICP-OES, SPECTRO etek, Germany). The crystalline structure was charac- X-ray diffraction (XRD) using the Cu K� radiation on D8 Advance (Germany) in the 2� range from 10◦ to parameters were determined by Rietveld refinement software. The particle size distribution was evaluated lp of a high resolution Scanning Electron Microscopy MA 10 microscope, Zeiss). The BET surface area and density were measured by using the ASAP 2020 and c II 1340 from Micromeritics Instrument Corporation es were prepared by casting the slurry, with the com- 80wt% active material (by weight), 10wt% Super P nd 10wt% PVDF (Kynar® FLEX 2801, Arkema Group), . The electrode active material mass loading was about while the thickness (after pressing at 3–4 tons cm−2 s 10�m. With metal lithium foil as the counter and lectrodes, the cathode electrodes were assembled into ells with the 1M LiPF6 in 1:1 EC:DMC solution as the Cellswere cycled galvanostatically at different constant Fig. 2. S tions. (a Thehe c=14.2 es (nominal capacity =200mAhg−1, 1C=200mAg−1) 8V and 2.5V at 20 ◦C using Maccor series 4000 battery ). All potential reported in this work refer to the Li/Li+ riments, including synthesis, were duplicated to check lity. and discussion ws theXRDpattern and theRietveld refinement results n0.56Ni0.16Co0.08]O2. The values of �2 (4.926) and Rwp nstrate a satisfactory refinement.With the exclusion of nsity reflections, in particular thosewithin 20◦ and 25◦, attice was approximated as signed to the layered struc- pace group of R3¯m [3], the intensity ratio of the (003) eaks (equal to 1.11), and the clear splitting of the (108) eaks suggest an ordering structure of R3¯m single phase. The weak r Li2MnO3-li ordering of [17]. The ICP transition m in good agr The crystall formula [1] and (104). of the pure conditions The mo cles and p microscopy that the p particles. T eflections are known to originate from the monoclinic ke (C2/m) super lattice [16], which correspond to the the Li+, Ni2+, and Mn4+ ions in the transition metal layer -OES analysis indicated the molar ratio of Li and the etal ions to be Li:Mn:Ni:Co=1.21:0.53:0.17:0.09, i.e., eement with the composition of the starting reactants. ite sizewas calculated fromXRDby using the Scherrer’s for the three main reflection peaks, e.g. (0 03), (1 01) The average result is 89nm which is smaller than that LiMn0.4Ni0.4Co0.2O2 prepared with the same annealing [18]. rphology of the Li[Li0.2Mn0.56Ni0.16Co0.08]O2 parti- ressed electrode, investigated by scanning electron , is shown in Fig. 2. The SEM image in Fig. 2a shows owder contains aggregates of primary, round sharp hese primary particles, which are better observed in Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 J. Li et al. / Journal of Power Sources 196 (2011) 4821–4825 4823 Fig. 3. Com Li[Li0.2Mn0.56N Fig. 2b, are 300nm. Sin Li ions, the are expecte electrode is is compose material an the prepare BET, is 7.3 mercial LiM Europe Gm from the so (7.31m2 g− 100–300nm is expected enhancing t cially the h the compar area materi 4.78m2 g−1 a worse hig material in larger parti cycle and 61 87.2mAhg− material. Th better perfo Fig. 4 Li[Li0.2Mn0. sities. As e exhibited tw of two diffe shifted to h plateau, wh almost the on the curr of the LiM to the oxid capability o of the Li ion ered in the of 111.5mA 4.4/4.5V du loss of oxyg irst ch at di ycling performance of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrodes charged at discharged at different rates. The tests carried out in the voltage range 2.5V and 4.8V at 20 ◦C. sformed in a MnO2-like phase. The capacity delivered in cond plateau strongly decreases as the current increases Ahg−1 at 0.1C against 105mAhg−1 at 20C) thus indicating is process has rather slow kinetics. capacity delivered during the first discharge is obvi- resulting from the insertion of lithium in both the 4Ni0.4Co0.2O2-like regions and the MnO2-like regions. parison of high rate (20C) cycling performance of i0.16Co0.08]O2 electrodes with different particle sizes. homogenous and have a small size between 100 and ce small particles present shorter diffusion path for insertion and de-insertion of the Li in this material d to be faster [1,19,20]. The SEM image of pressed shown in Fig. 2c. It can be seen that the electrode d of a very homogenous mixture of primary active d Super P. In addition, the specific surface area for d Li[Li0.2Mn0.56Ni0.16Co0.08]O2 powder, as detected by 1m2 g−1, which is much higher than that of com- n1/3Ni1/3Co1/3O2 material (0.61m2 g−1, TODA Kogyo bH). In addition, the average particle size calculated lid phase density of 4.30 g cm−2 and BET surface area 1), is 190nm, which is in very good agreement with the particle size detected by SEM. Such a high surface area to favor the charge transfer process of lithium ions thus he electrochemical performance of this material, espe- igh rate capability. This is well shown in Fig. 3 where ison of the rate performance for different BET surface als is shown. The material having a BET surface area of (corresponding to 290nm average particle size) shows h rate (20C) performance than the small particle size terms of both capacity and cycling stability. In fact, the cle size material delivered only 119mAhg−1 at the 1st .6mAhg−1 after 300 cycles against 145.4mAhg−1 and 1, respectively, delivered by the smaller particle size is result proved that the smaller particle will lead to Fig. 4. F as cycled Fig. 5. C 0.1C and between be tran this se (220m that th The ously LiMn0. rmance at high rate cycling. compares the first charge–discharge profiles of 56Ni0.16Co0.08]O2 cathode at different current den- xpected from the XRD characterization, the material o plateaus during the first charge, due to the existence rent lithium de-insertion processes [6]. Both plateaus igher potential with the increase of current. In the first ich is located at 3.8–4.4/4.5V, all electrodes delivered same capacity (about 110–120mAg−1) independent ent used. This plateau is associated to the delithiation n0.4Ni0.4Co0.2O2-like regions [21] that corresponds ation of Ni2+ →Ni4+ and Co3+ →Co4+. Since the rate f these two processes is improved by the participation s in the transition metal layer [22], the charge deliv- first plateau is very close to the theoretical capacity hg−1. The second plateau, which is located above ring charging, is widely accepted to originate from the en from the layered Li2MnO3 lattice [6], which might Fig. 5 Li[Li0.2Mn0. and discha within the capacities a in Table 1. able to de at 0.2C and Table 1 Capacity of Li Charge rate w Cycle 1st 30th 80th 100th arge and discharge profiles of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrodes fferent rates in the voltage range between 2.5V and 4.8V at 20 ◦C. compares the capacity retention of 56Ni0.16Co0.08]O2 electrodes when charged at 0.1C rged at different current rates (0.2C, 1C, 5C and 20C) cut-off voltages of 2.5V and 4.8V (20 ◦C). The delivered t the 1st, 30th, 80th and 100th cycles are also listed In the first discharge process, the electrodes were liver capacities of 271.4mAhg−1 and 230.6mAhg−1 1C, respectively. Very interestingly, at 5C and 20C [Li0.2Mn0.56Ni0.16Co0.08]O2 electrodes discharged at different rates. as always 0.1C. Discharge capacity (mAhg−1) 0.2C 1C 5C 20C 271.5 230.6 195.3 155.4 225.1 195.4 164.6 113.7 217.4 192.8 155.9 110.4 214.6 190.7 155.0 109.4 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 下划线 Administrator 高亮 Administrator 高亮 4824 J. Li et al. / Journal of Power Sources 196 (2011) 4821–4825 rates, the electrodes were still able to deliver 195.3mAhg−1 and 155.4mAhg−1, respectively, which corresponded to 72% and 57% of the 0.2C delivered capacity. This comparison clearly indicates the promising high rate capability of the material. However, the discharge capacities decreased in the next several cycles inde- pendent on the current used in the test. At the 30th cycle, the capacity dropped to 225.1mAhg−1 at 0.2C (i.e., 82.9% of the 1st cycle capacity), and 113.4mAhg−1 at 20C (i.e., 73.1% of the 1st cycle capacity). However, after the 30th cycle the capacity fade was much less significant at all rates. It is proved that capacities as high as 217.4mAhg−1 after 80 cycles at 0.2C rate, and 190.7mAhg−1, 155.5mAhg−1 and 109.4mAhg−1 after 100 cycles at 1C, 5C and 20C, respectively, were obtained. The comparison with the values obtained at the 30th cycle indicates that only a minor capacity loss, i.e., less than 8mAhg−1, took place. This proves that this material is a good candidate for high discharge rate electrodes even though the capacity retention in the initial 20 or 30 cycles still needs to be improved. Nevertheless, it is important to notice that the capacity delivered by our material during the 100th discharge at 1C rate very well compare with that of the Li–Ni–PO4 coated 0.5Li2MnO3·0.5LiMn1/3Ni1/3Co1/3O2 which is reported to have the better cycle stability at high rate [23]. Unfortunately, direct comparison with long term high rate test of other materials from the same family is not possible so far till now due to the lack of literature data. In fact, most of the reports present results which are limited in both the cycle number and rate capability. To the best our knowledge, our material has shown the best cycle stability during long-term test. Because of the above mentioned high discharge rate perfor- mance, the material was then tested at high charge current. Fig. 6 shows thedelivered capacities of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 elec- trodes char cut-off volt in the 1st, 3 From Fig. 6 charge capa and 145.4m values are o ing process Fig. 6. Cycling performance of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrodes subjected to continuous cycling at different rates (0.2C, 1C, 5C and 20C). Charged and discharged at the same rate. The tests were carried out in the voltage range between 2.5V and 4.8V at 20 ◦C. The low rate test date (0.2C) is referred to the upper cycle number scale. The charge efficiency during the cycle test at 0.2C is also reported. Table 2 Capacity of Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrodes charged and discharged at dif- ferent rates. Cycle Discharge capacity (mAhg−1) 0.2C 1C 5C 20C 1st 266.8 233.6 191.3 145.4 30th 223.1 201.7 168.4 118.2 th th ot st acity delivered by the electrode charged at 20C was 54.5% delivered by the electrode tested at 0.2C. This latter value is ally coincident with the result of Fig. 5. After 30 cycles, the al maintained the similar capacity of 223.1mAhg−1 for the Fig. 7. Cycling C and 20C. The test was carried out in the voltage range between 2.5V and 4.8V at 20 ◦C. due to a temperature increase (from 20 ◦C to 25 ◦C), which affected all cells under test. ged and discharged at the same current rate within the age of 2.5V and 4.8V (20 ◦C). The delivered capacities 0th, 100th, and 300th cycles are also listed in Table 2. , it is seen that the electrodes achieved first cycle dis- cities of 266.8mAhg−1, 233.6mAhg−1, 191.3mAhg−1 Ahg−1 at 0.2C, 1C, 5C and 20C, respectively. These nly slightly lower than those obtained with the charg- at 0.1C (Table 1), thus indicating that the charge rate 100 300 does n the cap of that practic materi performance of a Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrode cycled at, alternatively, 0.1 The anomalous capacity behavior recorded between the 60th and the 80th cycles is 209.4 181.5 148.9 100.3 161.0 132.5 87.2 rongly affect the material performance. Furthermore, Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 Administrator 高亮 J. Li et al. / Journal of Power Sources 196 (2011) 4821–4825 4825 low 0.2C rate, but got even a little more for the high C-rate like 1C, 5C and 20C. Then, it is expected that this material is not only useful for high rate discharging, but also for high rate charging. As already pointed out, the unsatisfying performance here is still the capacity decrease during the initial cycles independent on the current rate. The long-term fading phenomenon after the 30th cycle, espe- cially seen in the high rate test appeared however, to be associated with kinetics rather than degradation. Fig. 7 shows the capacity vs. cycle number behavior of a Li[Li0.2Mn0.56Ni0.16Co0.08]O2 electrode subjected to consecutive set of cycles at low (0.1C) and high (20C) charge and discharge rates. This double rate test was performed by charging anddischarging the electrode at 0.1C for 5 cycles, and then at 20C (both charge and discharge) for the next ten cycles. The test wasperformedwithin the voltage rangebetween2.5Vand4.8V, by repeating the steps for several times. Considering only the low rate cycles, it is seen that the electrode behaved very similarly to the one reported in Fig. 6. It delivered 263.9mAhg−1 at the 5th cycle with only a small decrease (261.7mAhg−1) at the 16th cycle, i.e., after ten cycles at 20C. In practice, the high rate cycling did not affect much the capacity fading of the material at low rate. As a matter of fact, the electrode reached the steady-state conditions after 35 low rate cycles with a stable delivered capacity of 225mAhg−1 (0.1C). The delivered capacity during the high rate cycling also shows the same behaviors as the electrode cycled at high rate only (see Fig. 6). However, th stable trend cates that i a degradati composite e cycling. It a material can ible current 4. Conclus The laye successful s shows the originating are homoge to be very h rial. The high and good ra at 0.2