Post-synthetic treatments on Ni x Mn x Co 1 − 2 x (OH) 2 for the preparation of lithium metal...

Materials Research Bulletin 46 (2011) 1878–1886

Post-synthetic treatments on NixMnxCo1 � 2x(OH)2 for the preparation oflithium metal oxides

I. Rodrigues, J. Wontcheu, D.D. MacNeil *

Departement de chimie, Universite de Montreal, Montreal QC H3T 1J4, Canada

A R T I C L E I N F O

Article history:

Received 19 March 2011

Received in revised form 19 July 2011

Accepted 25 July 2011

Available online 30 July 2011

Keywords:

A. Oxides

B. Chemical synthesis

C. Thermogravimetric analysis (TGA)

C. Electrochemical measurements

D. Energy storage

A B S T R A C T

A series of hydroxides NixMnxCo(1�2x)(OH)2 for x = 0.00–0.50 were prepared. These hydroxides were used

as the precursors in the synthesis of electrochemical active lithiated mixed metal oxides,

LiNixMnxCo(1 � 2x)O2. The traditional co-precipitation method was used to synthesize the hydroxides

and the effect of different post-synthetic treatments were tested. The solutions after co-precipitation of

the hydroxides were heated under hydrothermal or microwave assisted hydrothermal conditions at

180 8C. All samples were analyzed with X-ray diffraction (XRD), scanning electron microscopy (SEM) and

electrochemical measurements. We observed that the hydroxides undergo oxidation to an oxyhydroxide

phase as the stoichiometry varies during their synthesis and with post-synthetic treatments. As the

concentration of Ni and Mn increases in the sample, a mixture of both hydroxide and oxyhydroxide

phases is obtained. SEM images demonstrate a sintering effect on the hydroxide particles after post-

synthetic treatment, while XRD measurements on these samples show an increase in crystallinity and

reduced turbostratic disorder. The oxides synthesized from these precursors demonstrate similar

electrochemical performance with one another.

� 2011 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Materials Research Bulletin

jo u rn al h om ep age: ww w.els evier .c o m/lo c ate /mat res b u

1. Introduction

LiCoO2 has been the most widely used cathode material forrechargeable lithium-ion batteries due to its ease of fabrication,high energy density and excellent cycle life. However, it presentselevated production costs due to the use of expensive and rarecobalt metal, in addition, a number of incidents have raisedconcerns about its safety [1–3]. There have been numerous reportson lithium mixed metals oxides (LiNixMnxCo(1 � 2x)O2) as anemerging commercial cathode material for the replacement ofLiCoO2 [3–10]. One of the difficulties presented by these mixedmetal oxides is the low particle density obtained by the traditionalsynthesis method. The precursors of the oxides, mixed metalhydroxides (NixMnxCo(1 � 2x)(OH)2), are typically prepared througha co-precipitation method that consists of precipitating a mixtureof metal salts within a basic solution [5]. This precipitationtypically leads to a material with a small particle size and low tapdensity. In the second reaction step, the hydroxides are subse-quently oxidized with a lithium salt at high temperatures in air toform the lithium metal oxide [3–10].

Multiple strategies during the co-precipitation of the hydro-xides as well as during their oxidation have been pursued,

* Corresponding author. Tel.: +1 514 343 7054; fax: +1 514 343 7586.

E-mail address: dean.macneil@umontreal.ca (D.D. MacNeil).

0025-5408/$ – see front matter � 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.materresbull.2011.07.032

including novel synthetic procedures, cation substitutions, andmetal doping [12–15]. Typically, the low tap density of theresultant oxide can be traced back to the low tap density of theprecursor hydroxide and thus some researchers have focused onimproving the properties of the precursor hydroxide, which shouldresult in an improved oxide [13,14]. Lithium mixed metal oxides(Li[NixLi(1/3 � 2x/3)Mn(2/3 � x/3)]O2) with a substantial increase inparticle size (and higher packing density) were obtained by heatingthe isolated hydroxides precursors (NixMn(1 � x)(OH)2) directlyafter precipitation [13]. These precursors after heat treatmentresulted in denser oxides compared to non-treated hydroxideprecursors. The higher packing density of the oxide leads to moredense electrode films and to batteries with more energy densitycompared to batteries using oxides from a non-treated hydroxideprecursor.

In this report, (NixMnxCo(1 � 2x)(OH)2) powders were pre-pared using two different post-synthetic treatment methods.The first one consists of a hydrothermal treatment within anauto-clave at 180 8C for either 5 or 24 h and the second methodis a microwave assisted hydrothermal procedure at 180 8C for15 min. Each hydroxide was subjected to these elevatedtemperature and pressure treatments in solution immediatelyafter co-precipitation. The impact of these processes on thephase composition of the hydroxides and oxides powders, on thephase morphology and on the electrochemical properties will bediscussed in this work.

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–1886 1879

2. Experimental

2.1. Preparation

Co(NO3)2�6H2O (98%), Ni(NO3)2�6H2O (98%), Mn(NO3)2�6H2O(98%) and LiOH�H2O (98%) (Aldrich) were used as startingmaterials and all solutions were prepared in distillated anddegassed water. NixMnxCo1 � 2x(OH)2 (x = 0.00, 0.05, 0.15, 0.30,0.45 and 0.50) were first prepared by the traditional co-precipitation method [5]. A 0.4 M solution containing the mixmetal nitrates with the desired stoichiometries was slowlydripped into a stirred basic solution of LiOH (1.2 M) using a pumpdelivering the metal solution at �3 mL/min. After delivering themetal solution and rinsing with water to ensure completedelivery of all metal salts, three different routes were developedfor the subsequent preparation of the lithium metal oxide. Thefirst route used the hydroxides isolated directly from the co-precipitation after rinsing with distilled water. The second routeconsisted of a post-synthetic hydrothermal treatment in whichthe aqueous solution containing the hydroxides was transferredinto a Teflon container and placed within a sealed digestivevessel (Parr). The vessel was then placed in an oven at 180 8C for 5or 24 h. In the third route a microwave assisted hydrothermaltreatment was applied. Here, the solution was sealed in closedTeflon liners, which were placed in a turntable for uniformheating within a microwave digestion system (MARS5, CEM). Thesystem operated at a frequency of 2.45 GHz and a power of1200 W. The temperature of the microwave was ramped rapidlyto 180 8C and kept under these hydrothermal conditions for15 min. In all cases the precipitate after treatment was rinsedseveral times with distilled water and dried overnight under dryair. The final lithiated oxide, LiNixMnxCo1 � 2xO2, was prepared bymixing the precursors hydroxides with an excess (3%) amount ofLiOH. After pelletizing the solid mixture it was heated in air at500 8C for 3 h, ground, a new pellet formed and heated at 900 8Cfor 3 h followed by a quench cooling (between large copperplates).

Fig. 1. XRD profiles of NixMnxCo(1 � 2x)(OH)2 (bottom black line = as prepared by co-

microwave assisted post-synthetic hydrothermal treatment). (For interpretation of the r

this article.)

2.2. Characterization

The crystalline phases of samples were determined by an X-raydiffractometer (XRD, Bruker D8 Advance) using Cu Ka radiationwith a step size of 0.0258 and step time of 15 s in a range of 15–608.Scanning electron micrographs (SEM) were carried out on a HitachiS-4300 microscope. Thermogravimetric analysis (TGA) measure-ments were performed under a flowing atmosphere of dry air witha TA Instrument thermogravimetric analyser (SDT600) at 15 8C/min from room temperature to 500 8C.

Electrochemical evaluations were performed by combining theoxide with 10% of a conductive carbon (Super-P Li, Timcal) and 10%polyvinylidene difluoride (PVDF, 5.5% in N-methylpyrroldinone(NMP)) with an excess of NMP to make a slurry. The slurry (80%active) was then deposited on a carbon coated Al foil using a doctorblade. The slurry was then dried at 70 8C and electrodes 13 mm indiameter were cut for cell assembly in standard 2032 coin-cellhardware (Hohsen) using a single lithium metal foil as bothcounter and reference electrode and a Celgard 2200 separator. Cellswere assembled in an argon-filled glove box using 1 M LiPF6 inethylene carbonate (EC)/diethyl carbonate (DEC) (1:3 by vol)electrolyte (UBE). Electrochemical evaluations were performed bycharging and discharging between 2.2 and 4.2 V (or 4.5 V) using acurrent of 5 mAh/g for the first 5 cycles and a current of 30 mAh/gfor the next 50 cycles at 30 8C on a BT-2000 electrochemical station(Arbin).

3. Results and discussion

3.1. Mixed metal hydroxides

A full range of NixMnxCo(1 � 2x)(OH)2 hydroxides (x = 0.00–0.50)were prepared by the traditional co-precipitation method [5] andthen subjected to two different post-synthetic treatments at 180 8C.Heating the dry isolated hydroxides at an elevated temperature hasbeen previously shown to be an efficient method towards moredense hydroxide particles [13]. The post-synthetic treatments

precipitation; middle red line = after 5 h hydrothermal and top blue line = after

eferences to colour in this figure legend, the reader is referred to the web version of

706050403020

*

(110

) (120

)

(130

)

*

x = 0.50

x = 0.4 5Inte

nsity

(a.u

.)

Diffraction angle (2Theta)

x = 0.0 0*

* NixMnxCo(1- 2x)OOH

24h Hydrother mal

Fig. 2. XRD profiles of NixMnxCo(1 � 2x)(OH)2 after post-synthetic hydrothermal

treatment for 24 h showing the development of the NixMnxCo(1 � 2x)OOH structure.

400300200100Temperature (°C)

80

90

100

Wei

ght l

oss (

%)

x = 0.30

a)

b)

Fig. 3. TGAmeasurements(15 8C/min,dryair) forNixMnxCo(1 � 2x)(OH)2 aspreparedby

traditional co-precipitation (a) and after post-synthetic hydrothermal treatment (b).

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–18861880

carried out in this work, hydrothermal and micro-wave assistedhydrothermal, are unique in that for the first time the heating step isapplied to the hydroxides within the synthetic solution, withoutisolation and drying of the product. These treatments also providehigh pressure, which presents an additional variable that could leadto an improved material.

Fig. 1 shows the X-ray diffraction pattern of all samples withinthe series NixMnxCo(1 � 2x)(OH)2 (x = 0.00 to x = 0.50) before andafter the two post-synthetic treatments. Introducing Ni and Mn intothe sample (increasing x) results in significant changes to thediffraction patterns: from well-defined diffraction peaks (x = 0.00) tointense peak broadening (x = 0.05). This change in diffraction profiletowards small and broad Bragg peaks can be attributed to theoxidation of the sample into oxyhydroxides [14]. As Ni and Mnare introduced in the series, the pristine CdI2 structure becomesunstable and the development of a turbostratic phase contributes tothe change in the diffraction pattern as reported by Jouanneau andDahn [11]. This change in diffraction profile towards broad Braggpeaks with lower intensities as well as the appearance of new peaksat lower diffraction angles can be attributed to the oxidation of thesample into oxyhydroxides [14]. Kosova et al. [16] reportedpreviously that the interslab distance within the hydroxide withhigh Mn content increases as a result of the intercalation of anionsand water within the layer to compensate the added positive chargeleading to samples with very low crystallinity. Additionally to thepeak broadening we observed with larger Mn content, an interestingcolour differentiation was also observed among our NixMnx-

Co(1 � 2x)(OH)2 phases. Immediately after synthesis, the sampleswith a high value of x were light pink in colour. After rinsing withdistilled water the colour turned to a persistent light brown. vanBommel et al. attributed this colour change to an oxidative processupon heating the hydroxides in air [14]. The samples containing alow concentration of Ni and Mn (low value of x), do not demonstratea colour change, maintaining their pink colour throughout therinsing procedure, leading to non-oxidized samples. Thus, there is anincrease in the possibility of oxidation of the hydroxide as theamount of Ni and Mn is increased in the hydroxides [14,17]:

2NixMnxCoð1 � 2xÞðOHÞ2þ 1/2O2 ! 2NixMnxCoð1 � 2xÞOOH þ H2O

The oxyhydroxide phase is present in some of the samples as-prepared via co-precipitation, as well as after the two post-synthetic treatments. As the samples are treated at highertemperature during the post-synthetic treatment, the Bragg peaksbecame sharper and more intense indicating a more well definedoxyhydroxide phase. The exposure to elevated temperatureseliminates the structural defects that are readily apparent atlower temperatures and cause broadening of the diffractionpattern. Keeping the reaction at elevated temperature for a longerperiod of time (traditional hydrothermal compared to microwave)increases the formation of the oxyhydroxide phase. This is clearlyshown in Fig. 1 where more intense and well-defined Bragg peaksare observed for the hydrothermally treated samples (red line inFig. 1) compared to samples treated with the microwave (blue linein Fig. 1). This improvement in crystallinity at elevated tempera-ture is also apparent with the non-oxidized hydroxide samples(low values of x in NixMnxCo(1 � 2x)(OH)2).

Fig. 2 shows the XRD diffraction patterns obtained fromsamples that were hydrothermally treated for 24 h (x = 0.00,0.45 and 0.50). We can clearly see a well-defined diffraction peak at�368, which originates from the oxyhydroxide phase [17]. Twoother small peaks at �288 and 448 are also visible and these arerelated to the oxyhydroxide phase. If we compare the XRD patternshown in Figs. 1 and 2 for the composition with x = 0.00, it is clearthat a pure single hydroxide phase can be prepared via thetraditional co-precipitation technique, while after hydrothermal

treatment for 24 h (Fig. 2) a small peak at a diffraction angle of�368 is observed indicating the oxidation of some of the sample tooxyhydroxide. The extent of oxidation is increased with increasetreatment time at elevated temperature. It can be confirmed by thechange of colour of samples hydrothermally treated for 24 h. After24 h of treatment all samples were light brown even to the lowervalues of x, demonstrating the extension of the oxidative process. Asimilar trend is observed for the other hydroxide samples.Interestingly, the amount of the oxyhydroxide phase in each ofthe samples in Fig. 2 is similar, which suggest that the extent ofoxidation for each sample upon extended hydrothermal treatmentis limited to the same amount. It is possible to conclude that thehydrothermal treatment favors the oxidation of the sample into anoxyhydroxide phase. It is possible that the addition of a mildchemical reductor during the hydrothermal reaction at elevatedtemperature could lead to less oxidation of the hydroxide into theoxyhydroxide phase.

The amount of oxidation within these samples can be exploredvia thermogravimetric experiments [11]. TGA on all samples wereperformed under a flowing atmosphere of dry air and the weightloss for each sample was compared. It is expected that moreoxidized samples will demonstrate a lower weight loss due to the

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–1886 1881

sample already being partially oxidized. For clarity, we presentonly the TGA results for the sample as prepared by the traditionalco-precipitation and after the hydrothermal treatment (5 h) withx = 0.30 in Fig. 3. The TGA profile can be separated into twodifferent regions. The first includes the mass loss by each sampleup to 250 8C and is attributed to the release of absorbed water inthe material. The mass loss above 250 8C can be attributed mainlyto the oxidation of samples into the oxide phase and thedecomposition of interlayer ions contained within the structureof hydroxides. Sample as prepared by co-precipitation had a massloss of roughly 8% above 250 8C, while the same sample buthydrothermally treated after synthesis had mass loss of onlyroughly 4%. This decrease in mass loss after hydrothermaltreatment supports our view that the hydroxides oxidizes to anoxyhydroxide phase during post-synthetic hydrothermal treat-ment and this was supported by our XRD investigation describedpreviously (Figs. 1 and 2).

A structure refinement was carried out for the hydrothermallytreated samples by a Rietveld analysis [18–21] using the integratedpowder diffraction software TOPAS version 3.0 [22]. Clearly thehydrothermal samples in Fig. 1 show the presence of bothhydroxide and oxyhydroxide phases, therefore a two phaseRietveld refinement was performed. The background was interpo-lated linearly between selected points and the shape of thereflections was modeled with a pseudo-Voigt function, whilepreferred orientation was treated using March’s function. Theatomic coordinates of the oxygen atoms were refined withoutconstraint. The starting values for the atomic positions were thoseof the Co(OH)2 structure, in space group P3m1 (No. 164) and theCoOOH structure, in the space group P63/mmc (No. 194) [23]. Therewere no corrections performed for absorption. The latticeparameters of the hydroxide phase obtained for samples treatedunder hydrothermal conditions which demonstrate both hydrox-ide and oxyhydroxide characteristics are presented in Table 1. Theresults of Table 1 agree well with the results described above inthat the amount of oxyhydroxide phase increases as the Ni and Mncontent of the sample increases. Additionally one can see a netincrease on the lattice parameters of the NixMnxCo(1 � 2x)(OH)2

phases due the introduction of larger Mn2+ cations.Scanning electron microscopy (SEM) was used to investigate the

effect of post-synthetic treatments on the particle size andmorphology of each of the samples. While the morphology of allthe samples is similar, an increase in particle size with post-synthetictreatment is visible. Fig. 4 demonstrates a small sintering effect afterhigh temperature treatment. Comparing particles size from Fig. 4a,as-prepared by co-precipitation, to Fig. 4b and c after post-synthetictreatment, we can see that Fig. 4b (hydrothermally treated sample),has slightly larger particles. This sintering effect at elevatedtemperatures is not uncommon and leads to an increase incrystallinity, readily apparent in the diffraction profiles shown inFig. 1. As the sample sinters the disorder within the structure isrelieved and there is the development of more well defined Braggreflections (see Fig. 1). This development in Bragg reflectionsincreases as the sintering or time at elevated temperature increases.

Table 1Lattice parameters, unit cell volume and oxyhydroxide phase present in

NixMnxCo(1 � 2x)(OH)2 after post-synthetic hydrothermal treatment for 5 h.

x in

NixMnxCo1 � 2x(OH)2

a (A) c (A) V (A3) Oxyhydroxide

phase (%)

0.00 3.181 4.651 40.70 0.15

0.05 3.188 4.678 40.86 3.75

0.15 3.193 4.680 40.94 3.77

0.30 3.214 4.685 41.03 6.03

0.45 3.244 4.711 41.25 7.22

0.50 3.279 4.716 41.57 8.35

The particle sintering effect under extended hydrothermal treatmentis seen clearly when a sample (x = 0.00) is treated under hydrother-mal conditions for 24 h. Fig. 5 presents a comparison of this samplewith the un-treated sample and shows around 100� increase inparticle size after 24 h of hydrothermal treatment. In addition, theparticle size of the sample shown in Fig. 5b has a similar particle sizeto that presented previously in the literature where the dryhydroxide samples were treated at temperatures in excess of1000 8C [13]. This increase in the primary particle size of theprecursor should result in an increase in the particle density of thesynthesized oxide. Fig. 6 demonstrates the effect of Ni and Mnconcentration on the particles size and morphology of thehydroxides prepared by co-precipitation before any post-synthetictreatment. As the Ni and Mn content within the sample increases(x ! 0.50), the particle size demonstrates a significant decrease. Thiscan be attributed to the turbostratic disorder present [11] in sampleswith high content of Ni and Mn. In the X-ray diffraction patterns(Fig. 1) small broad peaks, indicative of turbostratic disorder, arevisible as x increases to 0.50. In the macroscopic scale, this disordertends to produce a sample with small particle size. We have shown

Fig. 4. SEM images of NixMnxCo(1 � 2x)(OH)2 (x = 0.15) (a) as prepared by co-

precipitation, (b) after hydrothermal for 5 h and (c) after microwave assisted post-

synthetic hydrothermal treatment for 15 min.

Fig. 5. SEM images of NixMnxCo(1 � 2x)(OH)2 (x = 0.00) submitted to (a) 5 h and (b) 24 h of post-synthetic hydrothermal treatment (main images have a scale bar of 3 mm while

the scale bar of the insert is 30 mm).

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–18861882

SEM images related to the samples as-prepared by co-precipitationonly, but the same trend is observed for both post-synthetic treatedhydroxides.

3.2. Lithium mixed metal oxides

For application within lithium-ion batteries, these mixedmetal hydroxides need to be oxidized into lithiated oxides. It hasbeen found that the morphology of the precursor hydroxide hasa significant effect on the ability to produce the optimal dense,spherical lithiated oxides [13,14]. Thus, it is important to fullyinvestigate various synthesis and treatment methods on thehydroxides such that one can obtain dense oxides. Dense oxideswill produce dense electrodes for use in lithium-ion batteries.

Fig. 6. SEM images of NixMnxCo(1 � 2x)(OH)2 as prepared by

The lithiated oxides were prepared by reacting the precursorhydroxides with a slight excess of LiOH (3%) in air at 500 8C for3 h followed by 900 8C for 3 h with quench cooling for each step.Fig. 7 shows the XRD patterns of the oxides produced from thevarious hydroxide precursors shown in Fig. 1. The heattreatments imposed to the hydroxides precursors do not seemto affect the structure or crystallinity of the oxides as Fig. 7shows a similar pattern for all samples. For LiNixMnxCo1 � 2xO2

(x = 0.05) prepared from precursors without post-synthetictreatment there is an additional peak around 448 that seemsto be related to development of an oxide impurity. Nevertheless,there is a smooth shift of the Bragg peaks as a function of the Niand Mn content. Table 2 presents the lattice parameters ofthe oxides, from the co-precipitation, hydrothermally and

co-precipitation for the indicated concentration of x.

50403020 50403020

x = 0.00

Inte

nsity

(a.u

.)x = 0.15 x = 0.30

x = 0.45 x = 0.50

Diffraction angle (2 Theta)

x = 0.05

Fig. 7. XRD profiles of LiNixMnxCo(1 � 2x)O2 (bottom black line = from the precursor hydroxide as prepared by co-precipitation; middle red line = from precursor after

hydrothermal (5 h) and top blue line = after microwave assisted hydrothermal treatment). (For interpretation of the references to colour in this figure legend, the reader is

referred to the web version of this article.)

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–1886 1883

microwave assisted hydrothermal treated samples indexedusing the a-NaFeO2-type structure (trigonal R3m). As expected,there is an increase in both the a and c lattice parameters withincreasing Ni and Mn concentration. According to previousworks [24] a small amount of Ni2+ occupies the 3a site when apartial substitution of Co for Ni is performed. For the Rietveldanalysis of our XRD data, we therefore assume that the 3a site isoccupied by a mixture of Li and Ni with the assumption thatLi(3a) + Ni(3a) = 1, where Li(3a) is the fraction of Li on the 3a siteand Ni(3a) the fraction of Ni on the 3a site. The concentration ofthe two cations on the 3a sites is highlighted in Table 2. Theconcentration of manganese in the final oxide phase wasestimated from the Rietveld analysis and it was found that therewas only a small deviation between the calculated and expected

Table 2Lattice parameters and cell unit volume of LiNixMnxCo(1 � 2x)O2, indexed using the R3m

precipitation, 5 h post-synthetic hydrothermal treatment and microwave assisted hydr

x in LiNixMnxCo(1 � 2x)O2 a (A) c (A)

Co-precipitation 0.00 2.819 14.09

0.05 2.823 14.07

0.15 2.842 14.17

0.30 2.855 14.28

0.45 2.880 14.31

0.50 2.892 14.40

Hydrothermal 0.00 2.816 14.06

0.05 2.819 14.09

0.15 2.837 14.15

0.30 2.857 14.22

0.45 2.878 14.28

0.50 2.890 14.29

Microwave hydrothermal 0.00 2.814 14.03

0.05 2.821 14.05

0.15 2.837 14.19

0.30 2.863 14.34

0.45 2.880 14.51

0.50 2.901 14.66

Mn content, thus very minimal (if any) Mn dissolution occurredduring the synthesis of the material. The Mn content within thefinal oxide phase is shown in Table 3.

Fig. 8 shows SEM images of the LiNixMnxCo1 � 2xO2 preparedfrom hydroxide precursors either as-prepared by co-precipitation(Fig. 8a), hydrothermal treated for 5 h (Fig. 8b) or microwavehydrothermal treated (Fig. 8c). Each sample had a Ni and Mnconcentration of 0.15 (x = 0.15) and this figure can be comparedwith that of the hydroxide sample, previously presented in Fig. 4.While the particle size of the precursor hydroxides (5 h ofhydrothermal and micro-wave assisted treatments) showed asmall increase with post-synthetic treatment at high temperature,the particle size of all the oxides presented in Fig. 8, regardless ofthe particle size of the precursor, are similar. It should be noted

space group for the oxide samples prepared from precursors originating from co-

othermal treatment.

V (A3) Li(3a) z(ox) Rwp RB GOF

96.80 1.000 0.235 7.25 4.25 3.01

97.31 0.981 0.240 4.69 6.52 2.85

97.43 0.957 0.243 8.21 5.04 1.99

98.05 0.950 0.248 9.26 6.21 1.68

98.67 0.926 0.237 10.02 5.06 1.98

99.23 0.912 0.234 8.11 7.25 1.35

96.69 1.000 0.246 8.65 5.02 1.67

97.11 0.983 0.239 5.69 3.56 1.16

97.32 0.951 0.231 6.79 4.99 1.77

97.82 0.948 0.239 8.14 3.89 1.09

98.74 0.914 0.240 7.56 3.44 1.19

99.10 0.905 0.238 9.31 6.45 1.86

97.02 1.000 0.236 10.15 6.08 2.07

97.22 0.985 0.229 4.21 4.25 1.10

98.01 0.949 0.246 7.16 5.36 1.45

98.11 0.942 0.242 5.16 4.16 2.09

98.97 0.912 0.239 6.54 6.01 3.10

99.82 0.909 0.237 7.02 5.34 2.11

Fig. 8. SEM images of LiNixMnxCo(1 � 2x)O2 (x = 0.15) synthesized from the

hydroxide precursors (a) as prepared by co-precipitation, (b) after 5 h

hydrothermal and (c) after microwave assisted post-synthetic hydrothermal

treatment.

Table 3Estimated Mn content in LiNixMnxCo(1 � 2x)O2 phases.

x 0.00 0.05 0.15 0.30 0.45 0.5

x0 0.00 0.045(4) 0.154(2) 0.301(2) 0.439(4) 0.497(2)

x00 0.00 0.044(2) 0.151(3) 0.302(3) 0.435(2) 0.486(3)

x is the nominal composition, x0 and x00 were obtained from the Rietveld refinement

of the X-ray data from the microwave-assisted and hydrothermal treatments

respectively (standard deviations are in brackets).

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–18861884

here that the differences seen with the hydroxide precursors inFig. 4 are much smaller than found in Fig. 5 and would be difficultto discern after the high temperature exposure required for oxideformation. Extending the hydrothermal treatment on the hydrox-ide to 24 h showed a significant increase in particle size (Fig. 5).Using this hydroxide precursor for the lithiated oxide provides a

Fig. 9. SEM images of LiNixMnxCo(1 � 2x)O2 (x = 0.00) synthesized from the hydroxide pre

after 24 h hydrothermal (scale bar 5 mm, insert scale bar 20 mm).

significant increase in particle size to the oxide. Fig. 9 compares theparticle size of the oxides (x = 0.00) synthesized from hydroxidestreated hydrothermally for either 5 (Fig. 9a) or 24 h (Fig. 9b).Clearly, the increase in the particle size of the precursor hydroxides(Fig. 5) can be maintained during oxide formation and a longerhydrothermal treatment provides a larger particle size. Theselarger oxides should lead to more dense electrodes and improvedenergy density in larger electrochemical cells. The particle size ofthe oxides from the 24 h hydrothermally treated hydroxidesamples show a similar particle size to oxides synthesized throughthe use of hydroxides treated at 1000 8C [13].

The capacity retention (capacity vs. cycle number) for theelectrodes of all oxides from Fig. 7 cycled at 30 8C between 2.2 and4.2 V are showed in Fig. 10. The first five cycles were at a rate of5 mA/g, while the remaining cycles were at a rate of 30 mA/g. Allsamples, except LiCoO2, present good capacity retention. Thecapacity of LiCoO2 cycled at 4.2 V decreases drastically withincreasing cycle number. This characteristic is typical of LiCoO2

prepared via hydroxides from co-precipitation and has beenreadily observed in the literature [6,12]. Interestingly, the LiCoO2

samples prepared from hydroxides that were subjected to a post-synthetic treatment demonstrate improved capacity retention,although commercial material (not shown) demonstrates super-ior capacity retention ability. This improvement is likely due tothe increased crystallinity and loss of turbostratic disorderobserved for the Co(OH)2 sample after high temperature post-synthetic treatment. The range in capacity values (between 150and 90 mAh/g) presented in Fig. 10 are within the values that havebeen presented previously in the scientific literature for the sameseries. Fig. 11 presents the electrochemical performance of thewhole series but using an upper cut-off potential of 4.5 V. Theseries charged to 4.5 V demonstrated an increase in capacity ofabout 20 mAh/g for each sample. The capacities presented arelower than those of similar commercial materials and this isexpected due to the un-optimized morphology of these small-scale syntheses. This can lead to issues with performance underhigh current densities and increased impedance. In addition, asmall amount of unidentified impurities could be present in thesample that has caused a lower capacity. Ultimately, the post-synthetic treatment experiments on the mixed metal hydroxides

cursors (a) after 5 h hydrothermal (scale bar 5 mm, insert scale bar 20 mm) and (b)

20 4010 30 5080

100

120

140

20 4010 30 50

80

100

120

140C

apac

ity (m

Ah/

g)

80

100

120

140

80

100

120

140

80

100

120

140

80

100

120

140

Cycle Number

x = 0.00 x = 0.05

x = 0.15 x = 0.30

x = 0.45 x = 0.50

Fig. 10. Capacity vs. cycle number for LiNixCo1 � 2xMnxO2 charged to 4.2 V. The hydroxides precursors were as-prepared by co-precipitation ( red), after hydrothermal for

5 h (x black) and after microwave assisted post-synthetic hydrothermal treatment ( blue). The charge–discharge curves consist of 5 cycles at a rate of 5 mA/g followed by 50

cycles at 30 mA/g. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

10 20 30 4080

120

160

10 20 30 40

80

120

160

Cap

acity

(mA

h/g)

80

120

160

80

120

160

80

120

160

80

120

160

Cycle Number

x = 0.00 x = 0.05

x = 0.15 x = 0.30

x = 0.45 x = 0.50

Fig. 11. Capacity vs. cycle number for LiNixCo1 � 2xMnxO2 charged to 4.5 V. The hydroxides precursors were as-prepared by co-precipitation ( red), after hydrothermal for

5 h (x black) and after microwave assisted post-synthetic hydrothermal treatment ( blue). The charge–discharge curves consist of 45 cycles at a rate of 30 mA/g. (For

interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–1886 1885

do not result in a large change in electrochemical performance ascompared to the non-treated samples except for the LiCoO2

sample prepared under co-precipitation conditions.

4. Conclusion

We have demonstrated two different post-synthetic treatmentson NixMnxCo(1 � 2x)(OH)2, via hydrothermal and microwaveassisted hydrothermal techniques, with the goal of producingmore dense hydroxides. These treatments are performed insolution right after the co-precipitation step and have been foundto increase the crystallinity and particle size of the samples as

compared to the non-treated samples. The hydroxide samplescontaining higher concentrations of Ni and Mn demonstrate anincrease in the amount of oxyhydroxide phase present in thesample and the degree of oxidation is increased with treatment atelevated temperature. Hydroxides hydrothermally treated for 24 hclearly demonstrated an increase in the concentration of theoxyhydroxide phase within the sample and a significant increase inparticle size. The particle size of the oxides prepared from thehydroxides treated for 24 h under hydrothermal conditions lead toa significant increase in the particle size compared to the untreatedsamples. These treatment methods lead to an increase in theelectrode density of lithiated oxides that should provide dense

I. Rodrigues et al. / Materials Research Bulletin 46 (2011) 1878–18861886

cathodes when produced using industrial methods for lithium-ionbatteries.

This is the first time that a post-synthetic treatment on thesemixed metal hydroxides precursors in solution has been reported.Previous treatments on the hydroxides were performed afterisolation and drying of the product. The particle size andmorphology of the samples as-prepared by co-precipitation andafter the various treatments were compared and we obtained asignificant increase in the primary particle size for the sampletreated under hydrothermal conditions for 24 h.

Some authors have suggested that the presence of anoxyhydroxide (NixMnxCo(1 � 2x)OOH) phase, in addition to thehydroxide phase, is visible when samples are exposed to air orelevated temperatures during their synthesis [11,16,17]. Here, weobserved a change in the degree of oxidation of the hydroxideswith a change in stoichiometry. As the value of x in NixMnx-

Co(1 � 2x)(OH)2 becomes larger the sample has a tendency ofoxidizing into oxyhydroxides. The extent of oxidation is alsoaffected by the treatments under hydrothermal conditions, wherean increase in time under hydrothermal conditions leads to anincrease in the amount of the oxyhydroxide phase.

The electrochemical performance of the oxides produced athigh temperature from the treated hydroxides (LiNixMnx-

Co(1 � 2x)O2) demonstrate a capacity and capacity retention withincreasing cycle number similar to those reported previously in theliterature. Interestingly, LiCoO2 prepared from the treated Co(OH)2

samples demonstrate improved capacity retention as compared tothose as-prepared by co-precipitation.

Acknowledgements

The authors thank NSERC and Phostech Lithium for funding thiswork under the auspices of the Industrial Research Chair program.

References

[1] G.G. Amatucci, J.-M. Tarascon, L.C. Klein, Solid State lonics 83 (1996) 167–173.[2] K. Numata, C. Sakaki, S. Yamanaka, Chem. Lett. 26 (1997) 725–726.[3] Z. Liu, A. Yu, J.Y. Lee, J. Power Sources 81–82 (1999) 416–419.[4] Z. Lu, D.D. MacNeil, J.R. Dahn, Electrochem. Solid-State Lett. 11 (2001) A191–A194.[5] Z. Lu, D.D. MacNeil, J.R. Dahn, Electrochem. Solid-State Lett. 12 (2001) A200–A203.[6] D.D. MacNeil, Z. Lu, J.R. Dahn, J. Electrochem. Soc. 149 (2002) A1332–A1336.[7] N. Yabuuchi, Y. Koyama, N. Nakayama, T. Ohzuku, J. Electrochem. Soc. 152 (2005)

A1434–A1440.[8] J.M. Paulsen, J.R. Dahn, J. Electrochem. Soc. 147 (2000) 2478–2485.[9] N. Yabuuchi, T. Ohzuku, J. Power Sources 146 (2005) 636–639.

[10] T. Ohzuku, Y. Makimura, Chem. Lett. 7 (2001) 642–643.[11] S. Jouanneau, J.R. Dahn, Chem. Matter 15 (2003) 495–499.[12] S. Jouanneau, K.W. Eberman, L.J. Krause, J.R. Dahn, J. Electrochem. Soc. 150 (2003)

A1637–A1642.[13] D.A.R. Barkhouse, J.R. Dahn, J. Electrochem. Soc. 152 (2005) A746–A751.[14] A. van Bommel, J.R. Dahn, J. Electrochem. Soc. 156 (2009) A362–A365.[15] X. Zhao, F. Zhou, J.R. Dahn, J. Electrochem. Soc. 155 (2008) A642–A647.[16] N.V. Kosova, E.T. Devyatkina, V.V. Kaichev, J. Power Sources 174 (2007) 735–740.[17] M. Butel, L. Gautier, C. Delmas, Solid State Ionics 122 (1999) 271–284.[18] H.M. Rietveld, J. Appl. Crystallogr. 22 (1967) 151–152.[19] H.M. Rietveld, J. Appl. Crystallogr. 2 (1969) 65–71.[20] D.B. Wiles, R.A. Young, J. Appl. Crystallogr. 14 (1981) 149–151.[21] R.J. Hill, C.J. Howard, J. Appl. Crystallogr. 18 (1985) 173–180.[22] F. Pertlik, Monatsh. Chem. Verwandte Teile anderer Wissenschaft. (1978) 109.[23] M. Deliens, H. Goethals, Mineral. Mag. 39 (1973) 152–157.[24] E. Wu, P.D. Tepesch, G. Ceder, Philos. Magn. B 77 (1992) 1039–1047.