1. IntroductionThe concept of Li-ion batteries has achieved an impressing technological and commercial success over the last two decades. In spite of many R&D works devoted to the development of new electrode materials and electrolyte components, the core chemistry of the vast majority of commercially accessible Li-ion cells has not changed essentially. In particular, the prevailing type of existing technologies is based on liquid electrolytes, typically being a solution of a lithium salt (LiPF6, etc.) in a mixture of organic carbonates (ethylene carbonate, diethyl carbonate, etc.). Batteries with liquid electrolytes have certain inherent drawbacks which include the necessity of using rigid battery casings and, more importantly, potentially lower safety due to the existence of volatile flammable solvents. In spite of great efforts, attempts to prepare fully solid state Li-ion battery (employing dry polymer electrolyte) have not been successful so far, mainly resulting from too low conductivities showed by polyether based lithium salt complexes. In the mid-90’s researchers from Bellcore

Corporation developed the first viable Li-ion battery technology that can be regarded as quasi-solid. The concept involved entrapping a standard liquid electrolyte in a microporous polymeric matrix/separator made of PVdF/HFP copolymer [1-4]. Because in this assembly the separator material is identical to the binder used in the electrodes, the whole unit can be integrated together by hot-pressing, which makes it possible to enclose the electrode-separator sandwich in elastic, lightweight pouches. In order to achieve better electrolyte uptake by the separator, fine silica powders are dispersed as the filler in the polymeric matrix; thus composite polymer-ceramic membranes are obtained. Since then, PVdF/HFP-ceramic filler-liquid electrolyte systems of this type have been developed by many researchers [5-13]. In particular, a considerable number of different submicron- and nano-sized ceramic fillers have been examined, including especially SiO2, TiO2, Al2O3, ZrO2, and many others (refer to review works [14-18] for overviews of the developments in ceramic fillers for composite polymer electrolytes). The principal role of dispersed ceramic powder in gel systems is to increase the conductivity by

Central European Journal of Chemistry

* E-mail: walkowiak@claio.poznan.pl

aInstitute of Non-Ferrous Metals Branch in Poznań Central Laboratory of Batteries and Cells, 61-362 Poznań, Poland

bPoznań University of Technology, Institute of Chemical Technology and Engineering, 60-965 Poznań, Poland

Mariusz Walkowiaka,*, Monika Osińskaa, Teofil Jesionowskib, Katarzyna Siwińska-Stefańskab

Synthesis and characterization of a new hybrid TiO2/SiO2 filler for lithium conducting gel electrolytes

Research Article

Abstract:

© Versita Sp. z o.o.

Received 22 March 2010; Accepted 28 August 2010

Keywords: Li-ion battery • Polymer gel electrolyte • PVdF/HFP • Functionalized ceramic filler

This paper describes the synthesis and properties of a new type of ceramic fillers for composite polymer gel electrolytes. Hybrid TiO2-SiO2 ceramic powders have been obtained by co-precipitation from titanium(IV) sulfate solution using sodium silicate as the precipitating agent. The resulting submicron-size powders have been applied as fillers for composite polymer gel electrolytes for Li-ion batteries based on poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF/HFP) copolymeric membranes. The powders, dry membranes and gel electrolytes have been examined structurally and electrochemically, showing favorable properties in terms of electrolyte uptake and electrochemical characteristics in Li-ion cells.

Cent. Eur. J. Chem. • 8(6) • 2010 • 1311-1317DOI: 10.2478/s11532-010-0110-3

1311

Synthesis and characterization of a new hybrid TiO2/SiO2 filler for lithium conducting gel electrolytes

increasing the liquid electrolyte uptake; whereas in “dry” polymer electrolytes, the same final effect is reached by reducing the polymer crystallinity and providing fast conduction pathways for ions.

Silica has been by far the most extensively studied type of ceramic filler. In a great number of works, various silica powders have been proven to improve polymer and polymer gel electrolyte characteristics such as specific conductivity, electrochemical window, interfacial stability, mechanical stability and lithium transference number [19-21]. Titanium oxides constitute the other type of filler with proven ability to enhance some of the above mentioned properties [22-24]. Because of the close analogy of structures and chemistries, it is possible, in the course of an appropriate synthetic route, to combine SiO2 and TiO2 into one composite material. Thus, the motivation of this work was to examine the properties of hybrid TiO2/SiO2 filler for composite polymer electrolytes. It may be expected that certain favorable properties of both SiO2 and TiO2 will be reinforced and a synergic effect will be detected. This paper describes the synthesis and preliminary results for a new type of filler for PVdF/HFP-based composite gel electrolytes which is a hybrid TiO2/SiO2 oxide submicron powder that is both pristine and additionally surface functionalized with methacryloxy and vinyl groups.

2. Experimental Procedure

2.1. Synthesis and characterization of TiO2- SiO2 hybrid powdersTiO2/SiO2 hybrids were obtained from a solution of titanium(IV) sulfate using a sodium silicate solution via the co-precipitation emulsion method. The emulsifiers were the non-ionic surfactants from the groups of NP3 and NP6 (nonylphenylpolyoxyethyleneglycol ethers with mean oxyethylenation degree 3 and 6 respectively). Modification of TiO2/SiO2 hybrid surfaces was carried out using two alkoxysilanes: 3-methacryloxypropyltrimethoxysilane (U-511 and vinyltrimethoxysilane (U-611), both purchased from UniSil. Two concentrations of the silane coupling agent were applied (10 or 20 weight parts per 100 weight parts of TiO2/SiO2). The process of the powder surface modification was conducted using the “dry technique”. The solution contained 10 or 20 weight parts of silane coupling agent per 100 weight parts of TiO2/SiO2 and 20 cm3 of the solvent, resulting in a 4:1 (v/v) mixture of methanol and water which was introduced into the 500 cm3 capacity reactor also charged with 40 g of the powder sample and a dosing solution of the organic modifying agent. The contents were mixed for 1 h to

Figure 1. Structural characterization of the hybrid TiO2-SiO2 ; a) EDX spectrum, b) XRD pattern, c) SEM image, d) volumetric particle size distribution

1312

M. Walkowiak et al.

ensure complete homogenization of oxide hybrids with the solution of modifying compound. The solvent was then evaporated by distillation. Subsequently, TiO2/SiO2 composites were dried at the temperature of 105°C in a stationary drier for 2 h.

Particle size distribution was determined using a Zetasizer Nano ZS (Malvern Instruments Ltd.) using the non-invasive back light scattering technique (NIBS). Scanning electron microsope (SEM) images and EDX spectrum of the ceramic powder were obtained using a Zeiss VO 40 apparatus. BET surface area was determined with analysis using ASAP 2020 instrument (Micromeritics Instruments Co.). XRD pattern was obtained using TUR–M62 diffractometer.

2.2. Preparation and characterization of dry membranesThe PVdF-HFP gels were prepared according to a method similar to the so-called Bellcore process (two-step method). PVdF/HFP copolymer (Kynarflex, Atofina), was added to acetone together with dibutyl phthalate (DBP) and the filler powder. The weight ratio of filler to copolymer was 1:10. The mixtures were stirred vigorously for several hours followed by ten minutes of ultrasonic shaking. The solutions were cast on a glass plate covered by a Petri dish and left for slow solvent evaporation. The dry membranes were subjected to SEM observations (cross-sections were obtained by freezing the specimens in liquid nitrogen and fracturing). The resulting plasticized membranes were subsequently immersed in diethyl ether and left overnight under stirring to extract DBP. The extraction stew was followed by drying at 60°C under vacuum.

For the determination of solvent uptake ability, round pieces cut from carefully dried membranes were weighed then immersed in a container with propylene carbonate (PC, anhydrous, 99.7%, Aldrich). In predefined amounts of time the membrane pieces were removed from the container, blotted lightly to remove excess liquid, weighed and immediately returned to the container. Liquid phase uptake was recorded as increasing changes in membrane weight due to membrane swelling:

[%]1000

0 ×−

=m

mmincreaseweightMembrane t

where: m0 - weight of dry membrane, mt - weight of membrane after a given time of swelling.

2.3. Electrochemical characteristics of gel electrolytesGel electrolytes were prepared by immersing round pieces (10 mm in diameter) of dry membranes for the period of 1 hour in the electrolyte consisting of a 1 M solution of LiPF6 (99.99%, Aldrich) in a 1:1 w/w mixture of ethylene carbonate (EC, anhydrous, 99%, Sigma-Aldrich) and dimethyl carbonate (DMC, anhydrous, 99%, Sigma-Aldrich). This action is referred to as activation of the dry membranes.

Conductivities were determined in two-electrode Swagelok-type cells with stainless steel electrodes (St│gel│St, where St refers to stainless steal) at several temperatures (0–70ºC) on the basis of impedance spectra obtained by means of PARSTAT 2263 (Princeton Applied Research) impedance analyzer in the frequency range 100 kHz – 1 Hz. Typically each measurement was

Figure 2. The TiO2/SiO2 surface functionalization using alkoxysilane: a) hydrolysis of 3-methacryloxypropyltrimethoxysilane, b) condensation of U-511 silane with inorganic filler, c) hydrolysis of vinyltrimethoxysilane and d) TiO2/SiO2 hybrid filler condensation mechanism with U-611 modifier.

1313

Synthesis and characterization of a new hybrid TiO2/SiO2 filler for lithium conducting gel electrolytes

repeated several times to ensure good reproducibility of results. The cells were thermostated during measurements in a climatic chamber (Vötsch).

Anodic stabilities have been studied in Li│gel│Pt cells by means of linear sweep voltammetry technique (scan rate 10 mV/s, potentials swept from the rest potential up to 6 V vs. Li/Li+, PARSTAT 2263 apparatus).

Functioning of the anode/electrolyte assembly was examined using anodic half-cells (Li│gel│graphite) with SL-20 graphite (Superior Graphite Co.) as the active anode material. Galvanostatic charging/discharging was performed at the current density of 10 mA/g between 0 and 2 V vs. Li/Li+. The graphite active mass was bound with the PVdF binder (Fluka, 5% of the overall electrode mixture) and coated onto copper foil current collectors using a “doctor blade” technique.

3. Results and Discussion

3.1. Characterization of ceramic powdersThe synthesized hybrid powder particles can be regarded as SiO2 matrices in which domains of TiO2 are

embedded. According to this hypothesis, the process of their formation involves local rearrangement of titanium coordination from octahedral to tetrahedral, although this view requires future experimental confirmation. EDX spectrum of a pristine TiO2/SiO2 (sample ST), presented in Fig. 1a, confirms clearly that the powder consists of both Ti and Si species and allows for the assessment of relative amounts of silica and titania in the powder (70.2 and 29.8%, respectively). XRD pattern (Fig. 1b) reveals that the TiO2 component is in the form of crystalline rutile and anatase phases; whereas the SiO2 component is fully amorphous. The hybrid powder (see SEM image in Fig. 1c) has a narrow particle size distribution profile with a maximum at ca. 600 nm (Fig. 1d).

The original material was subsequently functionalized by grafting methacryloxy or vinyl groups on their surfaces. Fig. 2 diagrams synthetic routes for the case of methacryloxy and vinyl functionalization. Thus two kinds of functionalized TiO2-SiO2 are formed. In addition, both in the case of methacryloxy and vinyl functionalization two concentrations of the appropriate silane coupling agent in the reaction mixture has been applied leading to two different degrees of surface coverage. This produces four different functionalized fillers for further

Figure 3. SEM images of the cross-sections of the dry membranes containing the studied fillers: a – no filler, b – ST, c – STM10, d - STM20, e – STV10, f – STV20; white bars correspond to 10 µm.

Figure 4. Solvent uptakes measured for the examined dry membranes as a function of time immersed in the liquid

Table 1. Acronyms and basic description of the synthesized ceramic fillers

Filler acronym

Description Surface functionality

ST TiO2/SiO2 not functionalized, hydroxyls

STM10 TiO2/SiO2, modified with U-511 (10 w. parts)

methacryloxy and hydroxyls

STM20 TiO2/SiO2, modified with U-511 (20 w. parts)

methacryloxy

STV10 TiO2/SiO2, modified with U-611 (10 w. parts)

vinyl and hydroxyls

STV20 TiO2/SiO2, modified with U-611 (20 w. parts)

vinyl

1314

M. Walkowiak et al.

testing. Table 1 contains symbols of all the described fillers along with their basic descriptions including the prevailing type of surface functionality. BET surface areas of the powders are given in Table 2.

3.2. Characterization of dry membranesAll powders have been applied as fillers in composite polymeric separator membranes. The average thickness of the membranes was ca. 100 µm. SEM images of the membrane cross-sections (Fig. 3) have been taken in order to assess their porosities. Visual inspection of these cross-sections produce interesting conclusions concerning the effect of the presence of filler and its surface functionality on the membranes’ internal structures. First, it appears that membrane without filler has very poor porosity whereas adding “bare” TiO2/SiO2 (sample ST) develops the internal surfaces significantly (compare Figs. 3a and 3b). Going further, when comparing two pairs of images: c-d and e-f, it is striking that increasing the extent of surface functionalization apparently leads to less developed porosity. This applies both to methacryloxy and vinyl type of functionality.

Fig. 4 illustrates how the membrane weights increase upon immersion in propylene carbonate. The solvent uptakes after 24 hours of swelling are also collected

in Table 2. All of the membranes with functionalized TiO2/SiO2 absorb solvent extremely well and much better than the membrane without any filler. The maximum uptake is also achieved quickly (less than one hour). The observed solvent uptake in the order of 324-338% are among the highest observed in the literature. Quite surprisingly, nonfunctionalized TiO2/SiO2 membranes (sample ST) absorb the solvent slowly with the maximum uptake being low. The results of visual observations are generally in poor agreement with scanning electron microscopy. Apparently, the existence of functionalized ceramic surfaces is of decisive importance for fast and voluminous liquid phase absorption.

3.3. Characterization of gel electrolytesDry membranes have been activated with liquid electrolyte. The resulting gel electrolytes have been subjected to conductivity measurements. In Fig. 5, specific conductivities are plotted as a function of temperature and values for 25 °C are also reported in Table 2. The conductivity data are in good agreement with solvent uptake results. Gels containing functionalized fillers exhibit the highest conductivities approaching 10-3 S cm-1. It is noteworthy that the highest conductivities have been measured for gels with moderately

Figure 5. Temperature dependencies of ionic conductivities determined for the gel electrolytes with various fillers, presented as Arhenius plots

Figure 6. Linear sweep voltammetry curves recorded on platinum electrode for the gel electrolytes with the examined fillers

Table 2.

Filler acronym BET surface area / m2 g-1 A24 / % σ25 / S cm-1 Qrev / mAh g-1

No filler N/A 251 2.3×10-4 285ST 36 91 0.4×10-4 34STM10 10 325 7.9×10-4 311STM20 4 324 4.5×10-4 308STV10 12 328 8.5×10-4 320STV20 5 338 4.3×10-4 311

Selected numerical data of the studied powders and membranes; A24 – solvent uptake (in percentage terms) of dry membranes after 24 hours of soaking, σ25 – specific conductivities of the gel electrolytes at 25°C, Qrev – reversible capacity of graphite in half-cells with gel electrolytes containing the examined fillers

1315

Synthesis and characterization of a new hybrid TiO2/SiO2 filler for lithium conducting gel electrolytes

functionalized fillers (samples STM10 and STV10 - 7.9×10-4 and 8.5×10-4 S cm-1 at 25°C, respectively). Gels without any filler features exhibit somewhat lower conductivity (2.3×10-4 S cm-1). Particularly low conductivities are measured for the gel without functionalized TiO2/SiO2 (ST), 0.4 × 10-4 S cm-1 at 25°C which is understandable in light of the solvent uptake results.

Anodic stability is an important characteristic of an electrolyte designed for Li-ion batteries. Fig. 6 shows linear sweep voltammetry curves recorded for the studied gel electrolytes using a platinum electrode. All the gels are practically stable up to about 4.5 V (vs. Li/Li+). The differences in the gels are quite negligible but become pronounced well above 5.5 V where intensive oxidation of the gel components begins. Gels with functionalized TiO2-SiO2 seem somewhat less resistant to anodic oxidation whereas the gel without filler is the most stable.

Lithium intercalation/deintercalation in anodic half-cells with graphite working electrodes can be regarded as an ultimate performance test of a gel electrolyte. Reversible capacities of the graphite anode have been determined and collected in Table 2. Reversible capacity is identified in this work with lithium deintercalation capacity in the first cycle and can be regarded as the capacity practically attainable from a graphite anode after the first charging (formation) step. Maximum theoretical reversible capacity (372 mAh g-1) is determined by the graphite crystal structure and the composition of the formed Li-graphite intercalation compound (LiC6). Practical capacities observed for natural and synthetic graphites (300 – 340 mAh g-1) are somewhat lower and are a result of specific ionic transport conditions, including ionic mobility, current density, etc.). Thus, ion transport properties of the applied electrolytes are of key significance for the reversible capacity of a Li-ion battery anode. As can be seen in Table 2, capacity of graphite in the cell with gel electrolyte without any filler is moderate (285 mAh g-1). Contrary to this, analysis of nonfunctionalized TiO2/SiO2 (ST) leads to a drastic deterioration of the graphite anode performance (34 mAh g-1). This can be explained on the basis of the solvent uptake and conductivity measurements. Apparently, this particular gel does not provide a medium for sufficient lithium cation transport, especially at the applied current density. On the other hand, application of functionalized TiO2/SiO2 (independent of the type of

functional groups and the degree of surface coverage) leads to very significant capacity enhancements (308-320 mAh g-1). Functional groups on the filler grain modify morphological and structural characteristics of the gels in such a way that they promote ionic transport throughout the membrane (by enhancing porous structure and liquid medium trapping). This in turn directly influences the measured capacities by minimizing the lithium transport overpotential.

4. ConclusionsA new type of filler for polymer gel electrolytes has been described. The paper reports synthesis, structural properties and preliminary electrochemical results for hybrid TiO2/SiO2 ceramic powders. New ceramic fillers greatly facilitate solvent uptake by dry PVdF/HFP membranes, but only if they had been surface modified. Nonfunctionalized TiO2/SiO2 exerts an adverse effect. The state of the ceramics’s surface chemistry seems to be an extremely important factor as far as its performance as a filler in gel electrolytes. Lack of any functionalization leads to very poor behavior in terms of solvent uptake, specific conductivity and lithium intercalation into graphite. Good ionic transport properties showed by gels with functionalized TiO2/SiO2 appear to come at the expense of slightly worse anodic stability. Here, not functionalized TiO2/SiO2 gels performs better and are the most resistant to anodic oxidation. Nevertheless, the differences are in fact negligible from a practical point of view. All the gel electrolytes are essentially stable up to about 4.5 V. The functionalized hybrid powders have proven to perform well as separators in Li-ion anodic half-cells allowing for effective lithium intercalation into graphite with reversible capacities well exceeding 300 mAh g-1. Hybrid TiO2/SiO2 powders with a moderate degree of surface functionalization can be regarded as promising filler materials for composite polymer-ceramic gel electrolytes used in Li-ion batteries.

Acknowledgements

The work has been financially supported by the European Regional Development Fund in the frameworks of the Innovative Economy Programme 2007-2013, project No POIG.01.01.02-00-015/09.

1316

M. Walkowiak et al.

J.-M. Tarascon, A.S. Gozdz, C. Schmutz, F. Shokoohi, P.C.Warren, Solid State Ionics 86–88, 49 (1996)M. Doyle, J. Newman, A.S. Gozdz, C.N. Schmutz, J.-M. Tarascon, J. Electrochem. Soc. 143, 1890 (1996)A. Du Pasquier, P.C.Warren, D. Culver, A.S. Gozdz, G.G. Amatucci, J.-M. Tarascon, Solid State Ionics 135, 249 (2000)A. Du Pasquier, F. Disma, T. Bowmer, A.S. Gozdz, G. Amatucci, J.-M. Tarascon, J. Electrochem. Soc. 145, 472 (1998)M. Caillon-Caravanier, B. Claude-Montigny,D. Lemordant, G. Bosser, J. Power Sources 107, 125 (2002)X. He, Q. Shi, X. Zhou, C.Wan, C. Jiang, Electrochim. Acta 51, 1069 (2005)M. Walkowiak, A. Zalewska, T. Jesionowski,D. Waszak, B. Czajka, J. Power Sources 159, 449 (2006)M. Walkowiak, A. Zalewska, T. Jesionowski, M. Pokora, J. Power Sources 173, 721 (2007)W. Pu, X. He, L. Wang, C. Jiang, C. Wan, J. Membr. Sci. 272, 11 (2006)N.T. Kalyana Sundaram, A. Subramiania, J. Membr. Sci. 289, 1 (2007)A. Subramiania, N.T. Kalyana Sundaram, A.R. Sathiya Priya, G. Vijaya Kumar, J. Membr. Sci. 294, 8 (2007) M. Stolarska, L. Niedzicka, R. Borkowska, A. Zalewska, W. Wieczorek, Electrochim. Acta 53, 1512 (2007)

A. Zalewska, M. Walkowiak, L. Niedzicki, T. Jesionowski, N. Langwald, Electrochim. Acta, 55, 1308 (2010)B. Kumar, L.G. Scanlon, J. Power Sources 52, 261 (1994)E. Quartarone, P. Mustarelli, A. Magistris, Solid State Ionics 110, 1 (1998)A.M. Stephan, K. S. Nahm, Polymer 47, 5952 (2006)J. Zhou, P.S. Fedkiw, Solid State Ionics 166, 275 (2004)M. Walkowiak, In: S.S. Zhang (Ed.), New concepts of composite polymer-ceramic electrolytes for Li-ion batteries, Advanced Materials and Methods for Lithium-Ion Batteries (Transworld Research Network, Trivandrum, India, 2007) chapter 15M. Caillon-Caravanier, B. Claude-Montigny, D. Lemordant, G. Bosser, J. Power Sources 107, 125 (2002)X. He, Q. Shi, X. Zhou, C. Wan, C. Jiang, Electrochim. Acta 51, 1069 (2005)M. Wachtler, D. Ostrovskii, P. Jacobsson, B. Scrosati, Electrochim. Acta 50, 357 (2004)R. Miao, B. Liu, Z. Zhu, Y. Liu, J. Li, X. Wang, Q. Li, J. Power Sources 184, 420 (2008)K.M. Kim, N.-G. Park, K.S. Ryu, S.H. Chang, Electrochim. Acta 51, 5636 (2006)Z.H. Li, H.P. Zhang, P. Zhang, G.C. Li, Y.P. Wu, X.D. Zhou, J. Membrane Sci. 322, 416 (2008)

References

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

1317