Nanostructured titanium oxide as active insertion material for negative

Nanostructured titanium oxide as active insertion material for negative electrodes in Li-ion batteries Marcus Fehse

To cite this version: Marcus Fehse. Nanostructured titanium oxide as active insertion material for negative electrodes in Li-ion batteries. Other. Université Montpellier II - Sciences et Techniques du Languedoc, 2013. English. .

HAL Id: tel-01018975 https://tel.archives-ouvertes.fr/tel-01018975 Submitted on 7 Jul 2014

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche fran¸cais ou étrangers, des laboratoires publics ou privés.

!"#$

' ( )( * + ,+ ' ( (, + ,+ ( 0( + ,+ / ' ( ( * + + ( 3( 3 + (+ 35 ( )( + (+ / ' ( )( + ,+

% &

- ., ./

.

'

, , 1 &

2*

.,

3

4 ' 2 2

Affidavit I hereby confirm that I prepared this thesis independently, by exclusive reliance on the literature and tools indicated therein. Furthermore, I state that this thesis has not been submitted to any other examination authority neither in this current nor in an altered form.

Montpellier, Marcus Fehse

“Science cannot solve the ultimate mystery of nature. And that is because, in the last analysis, we ourselves are part of nature, and therefore part of the mystery, we are trying to solve." Max Planck

Acknowledgement It is needless to say that such a comprehensive and complex work as a PhD thesis, can not be achieved without the help of numerous people. In the following I will try to list everyone that was involved in this thesis. Due to the vast number of contributors I’m almost certain that the list will be incomplete. I apologize for this inconvenience an assure you of my gratefulness. To begin with thanks to François Fajula, and Deborah Jones who gave me the opportunity to conduct this work in the group of ICG-AIME. Thanks also to all members of the jury particularly to the rapporteurs who spared no efforts to read this manuscript thoughtfully and critically. Furthermore I want to express my gratitude to the CNRS as well as SAFT for the financial support throughout my thesis. Special thanks to my contact persons Cécile Tessier and Florent Fischer for their help and guidance, for the supply of materials and chemicals as well as the valuable comments and suggestions from industrial perspective. I particularly appreciated the opportunity to go to the SAFT research facilities in Bordeaux to profit from in depth knowledge and cutting-edge equipment for electrode formulation. Scientific research is teamwork and I was blessed to find in the group of ICG-AIME helpful, welcoming and supportive colleagues. Therefore I want to thank all members of the ICG-AIME group. Thanks to Jean-Claude Jumas and Josette Olivier-Fourcade for their support in the beginning of this project and their contribution throughout this thesis. Special thanks to Cyril Marino who dedicated a lot of his time to ensure my fast familiarization with lithium ion battery research. In recognition to his unresting ambitions to spice my daily routine with meridional flair, I thank José for his companionship and for his disposition to lend a helping hand wherever necessary. Thanks David for supporting me in the office and good luck for your thesis. I also like to thank Ali Darwiche for his kindness and support and I wish him all the best for his future career, wherever it may lead him. In particular, I want to thank Bernard Fraisse and Julien Fullenwarth who play a crucial role in the functioning of this research group. Their keen perception for technical problems of all sorts, their commitment and helpfulness are the backbone of the day-to-day scientific work in this group. Notably, thanks to Julien for the XRF characterizations and to Bernard for his support during operando XAS and XRD measurement. Also thanks to Anne-Marie and Cathy, the secretaries, who do a great job in relieving the researchers from the burden of administration and paperwork. Other members of the group, I would like to thank are: P.E. Lippens for conducting DFT calculations on Nb-doped TiO2 , David Bourgogne for measuring numerous Raman spectra as well as Moulay Sougrati for his help at the XAS beamline. Also i would like to thank Sara Cavaliere and Iuliia Savych for the fruitful collaboration involving electrospinning and doping of TiO2 . My gratitude also to

i

ii

Julien Hannauer who joined the team one year ago and is working ambitiously to improve the operando IR cell as well as adopt it for other spectroscopic methods. Thanks also to Elia, Vincent, Surya, Sara, Fred and the others of the “late-lunch-group”. I’m grateful to have encountered such great colleagues who became my friends, I’m sure we will keep in touch. During my PhD thesis i had the opportunity to be mentor of two master students which was an enriching experience, both on personal and scientific level. I appreciate your effort and help Remi Grosjean and Matthieu Toiron and wish you good luck for the future. Naturally many people from other research groups and facilities have conduced to the success of this thesis. I especially want to thank Wanjie Zhang and Herve Martinez from IPREM group at Pau for conducting XPS measurements on post mortem electrodes and their support in analysing and interpreting the data. Also to Sophie Cassaignon from CMCP Collége de France for the HRTEM micrographs and collaboration on further development of hydrothermal synthesis. As well as David Maurin and Jean-Louis Bantignies from L2C, Université Montpellier II for stimulating discussions and support in setting up an operando IR spectroscopy technique. The collaboration with the ICG-CTMM group of Marie-Liesse Doublet allowed me to get to know the value and potential of theoretical approaches. Thanks to Mouna Ben Yahia not only for carrying out the calculations and thereby supplying important information to understand lithium insertion process but also for her patience and endurance in explaining theoretical models and methods to me. Thanks to the DESY-Hamburg, Germany and the Elettra-Trieste, Italy synchrotrons for granting us beamtime to conduct XAS measurements as well as for the financial and technical support. Thanks to Myrtille for proofreading and the french translation. As this acknowledgment draws to an end I want to thank two special persons without whom this work wouldn’t have been possible. Lorenzo and Laure, aka “LoLa” were much more than my bosses; they were my teacher, colleague, mentor, role model and friend at the same time. I especially value their profound scientific knowledge and professional working attitude while at the same time sharing laughs and fun. Thanks for three great years. Last but not least, thanks to my friends and family, who have despite of the great distance, kept close contact to me and given me all the support that I needed. Thanks also to the the city of Montpellier and its people who have grown dear to my heart, we shall meet again.

Contents 1 Introduction 1.1 Lithium Ion Batteries . . . . . . . . . . 1.2 Negative electrode materials . . . . . . . 1.3 Performance enhancement . . . . . . . . 1.3.1 Nanostructuring . . . . . . . . . 1.3.2 Porous structures . . . . . . . . . 1.3.3 Doping and composite formation 1.4 Objectives . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

2 Synthesis 2.1 Hydrothermal synthesis . . . . . . . . . . . . . . . . . . . 2.1.1 Experimental details . . . . . . . . . . . . . . . . 2.1.2 Synthesis parameters - dependencies and influences 2.2 Sol-gel synthesis . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Conventional petri dish . . . . . . . . . . . . . . . 2.2.2 Electrospinning . . . . . . . . . . . . . . . . . . . . 2.2.3 Summary and Discussion . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

3 Lithium insertion sites and mechanism 3.1 Influence of phase composition and morphology on electrochemistry 3.2 Influence of textural properties on electrochemical cycling . . . . . . 3.2.1 Conventional sol-gel . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Ultrasound assisted sol-gel . . . . . . . . . . . . . . . . . . . . 3.2.3 Template assisted . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Electrospun anatase nanofibers . . . . . . . . . . . . . . . . . 3.2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Operando measurements . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 In situ XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2 In situ XAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Lithium insertion mechanism . . . . . . . . . . . . . . . . . . . . . . 3.5 Lithium insertion sites in TiO2 (B) . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

1 2 4 7 8 10 12 12

. . . . . . .

15 15 15 17 23 24 28 30

. . . . . . . . . . . . .

33 33 37 38 39 39 40 41 45 46 49 58 58 61

iii

Contents

iv

4 Improvement of performance: Doping and Composites 4.1 Substitutional doping by Nb via electrospinning . . . . . . . . . . . . . . . 4.1.1 Effects of Nb dopant on structure and morphology of electrospun nanofibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Doping of TiO2 (B) via hydrothermal synthesis . . . . . . . . . . . . . . . 4.3 Coating and decoration of titanium oxide . . . . . . . . . . . . . . . . . . 5 Improvement of performance: Cell assembly, Electrolyte 5.1 Cell assembly . . . . . . . . . . . . . . . . . . . 5.2 Electrode formulation . . . . . . . . . . . . . . 5.2.1 Carbon additive . . . . . . . . . . . . . 5.2.2 Binder . . . . . . . . . . . . . . . . . . . 5.3 Electrolyte . . . . . . . . . . . . . . . . . . . . 5.3.1 Electrolyte degradation . . . . . . . . . 5.3.2 Adjusting solvent composition . . . . . . 5.3.3 Role of the lithium ion salt . . . . . . . 5.4 Charge transport limitations in electrodes . . .

and Additives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

65 65 66 77 81 85 85 85 87 90 93 94 97 98 101

6 Conclusion and Perspectives 105 6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 6.2 Prospects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Bibliography

111

List of Figures

137

List of Tables

145

A Abbreviations

147

B Characterization methods B.1 Electrochemical cycling . . . . . . . . . . . . . . . . . B.1.1 Galvanostatic cycling . . . . . . . . . . . . . . B.1.2 Cyclic voltammetry . . . . . . . . . . . . . . . B.2 Atomic absorption spectroscopy . . . . . . . . . . . . . B.3 X-ray photoelectron spectroscopy . . . . . . . . . . . . B.4 X-ray diffraction . . . . . . . . . . . . . . . . . . . . . B.4.1 Crystallite size determination . . . . . . . . . . B.4.2 Lattice parameter determination . . . . . . . . B.5 Raman . . . . . . . . . . . . . . . . . . . . . . . . . . . B.6 Surface Area and Porosity Analysis . . . . . . . . . . . B.7 Transmission electron microscopy . . . . . . . . . . . . B.8 Scanning electron microscopy, Energy dispersive X-ray

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . scattering

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

149 149 149 150 150 150 151 151 152 153 153 154 155

Contents

B.9 X-ray fluorescence . . . . . . . B.10 X-Ray absorption spectroscopy B.10.1 Experimental details . . B.10.2 Fitting parameters . . .

v

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

155 155 157 157

C Computational methods 161 C.1 Calculation of electronic band structure of Nb-doped anatase . . . . . . . 161 C.2 Calculation on lattice structure and lithium insertion sites of TiO2 (B) . . 161

1 Introduction The strong increase in primary energy demand within the last 100 years, driven by quadrupling of world population as well as the tripling in energy per capita consumption [1, 2], has been primarily shouldered by the combustion of fossil fuels, see Fig. 1.1.

Figure 1.1: World energy demand history and forecast, showing contributing energy sources. Figure taken from [3].

However, the awareness of the depletability of fossil fuels, their contribution to the green house effect as well as their negative impacts on economy (external trade deficit) and geopolitics (dependence on unstable regions, trouble spots) have given rise to new technologies, such as renewable and nuclear energy in the last decades. The latter, however bears an unsolved problem of nuclear waste disposal as well as uncontrollable security risk, as dramatically shown by nuclear melt down in Fukushima in March 2011. Therefore the importance of developing and expanding the proportion of discontinuous, decentralized renewable energy sources, such as solar and wind power as well as alternative and environmentally benign fuels for automotive applications, is on the agenda of many countries today [4]. Energy storage plays a key role within this scenario and has therefore become increasingly important, whereat electrochemical energy storage in lithium ion batteries (LIB1 ) takes a sublime position [5].

1

Lithium ion battery

1

1 Introduction

2

1.1 Lithium Ion Batteries Almost 30 years ago Yoshino et al. presented a new secondary battery system C/LiCoO2 based on the preceding work of Goodenough and Whittingham [6–8] on the “rocking chair system”1 . It was adapted and introduced commercially 1991 by SONY and Asahi Kasei/Toshiba [9]. This battery system, which became worldwide known as Lithium ion battery (LIB) was one of the biggest milestone in the development of electrochemical energy storage since Alessandro Volta presented the first voltaic pile in 1800.

Figure 1.2: Comparison of gravimetric and volumetric energy density of various secondary battery systems. Figure taken from [10].

Originally conceptualized to feed small portable electronic devices, they spread quickly, due to their superior storage capacity (see Fig. 1.2), long cycle life and high charging currents to become indispensable for modern high technology society [11]. Furthermore, lithium ion batteries do not show a memory effect, feature low self discharge [12] and can operate in a wide temperature range of typically -20 to 60 ◦ C, which contribute to their superior role. Today LIB are at the dawn of conquering new markets such as automotive application, backbone for smart power grids and even implementation in the aviation market [13]. This triumph is enabled by a consistent progress in research and development of the lithium ion battery, which has led to doubling its capacity (≥200 W h/kg) while at the same time mass production cut cost by over 80% to less than 400 $/kW h within the last two decades [11, 13–16]. A Ragone plot showing the evolution of power and energy density of the lithium ion battery during the last 25 years, given different milestone examples of cathode materials, is presented in Fig. 1.3. Despite this impressive development, LIB are still quite far from reaching benchmarks imposed by automotive industry, which demand gravimetric energy density of 500 W h/kg,

1

This metaphor is used to picture the back and forth movement of lithium ions between negative and positive electrode during charge and discharge

1.1 Lithium Ion Batteries

3

Figure 1.3: Specific energy vs. specific power of cathode materials for lithium ion battery systems and their year of introduction, taken from [16].

volumetric density of 1000 W h/l at a cost of 100 $/kW h. In order to close the immense gap between energy density of petrol (2000 W h/l) and of a LIB (300 W h/l), innovative materials and concepts are needed. It is evident that, although some incremental improvements can be made throughout engineering and optimizing electronics, new redox couples and concepts will be necessary in order tackle this immense challenge [5, 17]. All electrochemical cells have a mutual assembly, which comprises two electrodes; the positive and the negative, an electrolyte and other ancillary constructional components. In Figure 1.4 the typical setup of a lithium ion battery is schemed. During discharge the graphite lattice, which is hosting the intercalated Li+ , is oxidized and the lithium ions migrate through the ion conducting electrolyte to the positive cathode. The latter can be categorized mainly in three types, namely lithium metal oxide (e.g. LiCoO2 ), transition metal phosphates (e.g. LiFePO4 ) or spinels (e.g. LiMn2 O4 ). In order to make batteries usable for electric work, the electrons are conducted through an external circuit. While recharging, electric current is forced in the opposite direction and electrical energy is hence converted to chemical throughout the formation of charged products (e.g. LiC6 and Li1−x CoO2 ). In contrast to other battery technologies, many different materials can be used to host Li+ , both at the anode and the cathode. However, from the wide spectrum of potential redox reactions only few meet the high demands to reach the state of commercial application.

1 Introduction

4

Figure 1.4: Schematic assembly of standard lithium ion battery in discharge mode using graphite as negative and LiMO2 as positive electrode material. Figure taken from [16].

1.2 Negative electrode materials There are three main groups of lithium ion negative electrode materials, their properties and reaction mechanism are presented in Fig. 1.5 [18]. Conversion type materials undergo a strong modification of crystal structure upon Li+ uptake, which is usually connected to a large volume expansion. Moreover they often reveal strong hysteresis in electrochemical cycling curve, which reduces the energy efficiency. Conversion materials are based on binary transition metal compounds Ma Xb (M = transition metal Ni, Co, Fe; X = O, S, F, P, N...) [19]. Their reversible reaction with lithium leads to full reduction of the transition metal to the metallic state enabling a large lithium uptake by the formation of a Lin X matrix embedding the transition metal nanoparticles. Alloying-type materials inherit even higher theoretic capacities (up to 4200 mAh/g) by multi-step alloying with Li+ yielding compositions as high as Li4.4 M (M = Si, Ge, Sn) but featuring even higher volume changes. The immense volumetric expansion upon lithiation is the main drawback for their application as electrode materials [5]. Even after decades of research, leading to nanostructured materials and implication of sophisticated additives and binders to buffer these volume changes, they do not meet life span, security, and cycle stability requirements [18]. As a consequence, the third group, which comprises the intercalation materials is currently dominating the LIB market. The storage in this type of material is based on the topotactic1 incorporation of Li+ into an open host whereupon none or only minor structural modifications occur. In order to compensate the charge, a reduction of the host lattice occurs upon cation intercalation. These reaction characteristics result in reduced theoretical capacity compared to conversion and alloying materials [21], but at the same time provide intrinsic advantages such as high reversibility, good cycling

1

A transition in which the crystal lattice of the product phase shows one or more crystallographically equivalent, orientational relationships to the crystal lattice of the parent phase [20]

1.2 Negative electrode materials

5

stability and long cycle life, which make them the materials of choice for most of today’s application.

Figure 1.5: Schematic presentation of reaction mechanism and properties of three main electrode material types for use in lithium ion batteries. Black circles: voids in the crystal structure, blue circles: metal atoms, yellow circles: lithium ions. Figure taken from [18].

The most widely applied insertion type negative electrode material is graphite. Li+ are accommodated within a layered structure leading to a slight expansion of ≈16 %, see Fig. 1.4. The maximum uptake of Li is LiC6 , corresponding to a theoretic capacity of 372 mAh/g. Besides its good cycling properties the dominant role of carbon as insertion type material is furthermore based on its low cost, its abundance and its environmental benignancy. Mentionable endeavors to further enhance the power density throughout implication of carbon nanofibers [22, 23] and graphene [24–26], have been carried out. However, graphite’s low atomic density limits the volumetric energy storage density. Furthermore, the risk of lithium plating leading to fatal short circuit especially at elevated currents make the search for alternatives indispensable [27, 28]. Aggravatingly, the compulsory formation of SEI1 at the graphite surface to prevent electrolyte co-intercalation reduces coulombic efficiency and may hamper charge transport [29]. An intensively studied and to some degree already commercialized candidate to replace graphite are titanium oxide based insertion materials. Just like graphite, titania features environmental benignancy and is suitable for low cost mass production. More importantly, its characteristics comprise excellent lithium insertion properties, low self-discharge rate and high chemical stability [30]. The electrochemical mechanism is based on the redox couple Ti+4 /Ti+3 at a potential between 1.3 and 1.9 V vs. Li/Li+ , depending on the polymorph [31]. This provides an increased safety upon cycling as well as the option to 1

Solid electrolyte interface

6

1 Introduction

replace the cost intensive Cu by Al as the current collector [32]. Compared to carbon, spinel type Li4 Ti5 O12 (LTO) is considered a zero-strain material, meaning that almost no change of volume occurs during electrochemical cycling. This leads to excellent cycling stability and long cycle life, even at high cycling rates [33–35]. Furthermore, its flat operating potential (≈ 1.5 V vs. Li/Li+ ) and easy and cost effective synthesis have promoted LTO to take a leading role among the titanium oxide based electrode materials for industrial research and consumer applications, for example by Altairnano [36], or SCiB battery by Toshiba [37]. While the intrinsic disadvantage of insulating character could be somewhat overcome by doping, incorporating conductive second phase, etc. [35] its low theoretical capacity of 175 mAh/g is a major drawback in the hunt for high energy density LIB. TiO2 theoretical capacity is almost double with 336 mAh/g (corresponding to the insertion of one mole of Li per mole TiO2 ), which makes it an appealing alternative to the LTO. Pioneering work on lithium insertion into different TiO2 polymorphs was performed by Zachau [38]. The anatase phase, which was for long time considered the most electroactive polymorph [39] has a comparatively complex insertion mechanism. It is subdivided in three steps: solid solution, biphasic lattice transformation from tetragonal to orthorhombic system and subsequent monophasic lithium insertion reaction, which is accompanied by 4 % lattice expansion[40]. Compared to LTO, it has an elevated insertion potential plateau of 1.8 V vs. Li/Li+ . More recently the monoclinic TiO2 bronze, often referred to as TiO2 (B) phase, which is the least dense (3.64 g/cm3 [39]) among the common TiO2 polymorphs, has gained much attention. It crystallizes in the C2/m space group, which implies the existence channels in þb and in þc direction. This open channel framework can act as insertion sites as well as diffusion paths, and is crucial for high performance intercalation material. TiO2 (B) has the same elevated theoretic capacity as anatase, but at a reduced insertion potential of ≈ 1.5 V vs. Li/Li+ , which increases the total energy density of the complete battery 1 . In their recent publication, Aravindan et al. have impressively demonstrated the long term stability and excellent capacity retention of TiO2 (B), in full as well as in half cell set ups [41]. Moreover an increasing number of studies claim the superiority of TiO2 (B) phase over anatase due to fast pseudocapacitive and open channel storage rather than slow solid-state diffusion controlled lithium storage [40, 42–48]. Despite of the attention received, little detailed information on the insertion mechanism of TiO2 (B) and on the origin of the pseudocapacitive storage is known. There is a vivid discussion in literature among theoretical groups on the stability of lithium insertion sites but also on the lithium insertion mechanism, which was nicely covered by Morgan and Madden [49]. In their study, they question the common assumption that the insertion occurs primarily into the open b-channels. Also the role of channels as backbone for fast lithium ion transportation is challenged by measurements of relatively poor rates of

1

Total energy density of a battery is defined by the product of capacity and the output voltage

1.3 Performance enhancement

7

self-diffusion of Li+ in TiO2 (B) nanowires [50]. Wilkening et al. argument therefore, that the reason for the observed superior rate capability is rather linked to high surface area and unique specific surface energy. Understanding the pseudocapacitive storage systems is however of great interest, since it could hold the solution for new fields of application, as this hybrid of traditional batteries and double layer capacitor storage combines high energy density with high power properties, see Fig. 1.6 [51]. Fundamental research looking at the processes occurring during lithium insertion and de-insertion into this material by implying operando techniques 1 is necessary to elucidate the storage mechanism before utilization in practical applications.

Figure 1.6: Idealized voltage and differential capacity (δC /δV) profiles for basic battery, pseudocapacitor and capacitor energy storage. Pseudocapacitor is hereby the name used for a intermediate between high capacity battery and high rate capacitive energy storage, Fig. taken from [51].

1.3 Performance enhancement Unfortunately, TiO2 (B) and titania polymorphs in general bear some intrinsic disadvantages. The main drawback is their poor electronic and ionic conductivity resulting from lack of electrons in the conduction band (3d0 ) in combination with a wide band gap (≥3 eV )[52] as well as low solid state diffusion coefficient of lithium (DLi ≈10−13 cm2 /s)[31]. 1

A methodology wherein the characterization of a material is conducted, while it is subject to operating conditions

1 Introduction

8

Another well known handicap is the relatively high irreversible capacity loss of TiO2 upon first cycle [51]. This is linked to adsorbates on TiO2 surfaces as well as to the catalytic behavior of TiO2 , which challenges electrolyte stability. There are numerous efforts to overcome these obstacles, which will be encompassed in the following paragraphs.

1.3.1 Nanostructuring When regarding the Einstein diffusion equation, see Eq.1.1, reducing the grain size of a material seems like the intuitive solution to overcome the problem of slow solid diffusion. The diffusion path length d is proportional to the square of diffusion time t while diffusion coefficient DLi can be considered as constant at a defined temperature.

t=

d2 2DLi

(1.1)

A gedankenexperiment puts this into evidence, revealing that charging time of a particle can be reduced from 300 to 3 s if the size is reduced from 100 to 10 nm, considering an average diffusion coefficient of 10−13 cm2 /s [31]. Numerous experimental results have confirmed this relationship showing that nanostructuring can significantly enhance lithium insertion into TiO2 , see Tab. 1.1. Shin et al. presented a comparative study showing galvanostatic lithium insertion curves into TiO2 anatase with different nanostructures, see Fig. 1.7. It is salient that by a reduction of particle size and increase of specific surface area an incorporation of Li+ far beyond the bulk threshold value of 0.5 Li (165 mAh/g) can be achieved. Particle size reduction extends the plateau of biphasic insertion mechanism, while surface area and open porosity network promote especially the slope like insertion for ≥0.5 Li. While the first is being attributed to reduced diffusion path length, the latter is mostly linked to interfacial storage [53]. Despite the great achievements that have been obtained via nanostructuring regarding total capacity and rate capability, new challenges arise from this approach. The ambivalence of this technique lies in the fact that reduction of particle size does not only lead to shortened diffusion paths, but also strongly influences the electrochemical properties of the material. With decreasing particle size, the surface and interface properties become more influential/pronounced as their absolute share grows. This can be reflected by multiple effects such as; • Insertion/de-insertion potential: Particle size has influence on the reactivity and hence the electrochemistry of a particle. A distribution in particle size is therefore often accompanied by the flat insertion/ de-insertion plateau becoming indistinct. • Solubility: Solubility of reacting phases is a function of particle size, a reduction in particle size can significantly enhance solubility and hence promote solid solution

1.3 Performance enhancement

9

Figure 1.7: First galvanostatic discharge curves of commercial anatase with two different grain sizes (200 and 5-10 nm) and nanoporous anatase. Fig. taken from [53].

insertion mechanism [54]. • Defect chemistry Concentration of defects influence the conductivity since charge carriers (holes, electrons) are created for charge compensation. An effect on the insertion sites and mobility of lithium ions is also likely [55]. • Catalytic behavior: Since catalysis is a surface process, the total share and nature of the surface has influence on the catalytic ability. The presence of highly active TiO2 , which is well known as catalyst for numerous applications, could influence the stability of electrolyte components. • Formation of surface sites: With increased share of surface, the number of surface sites for lithium storage becomes more important. Their electrochemical insertion properties (potential, reversibility, energetics), are likely to differ significantly from bulk sites. In a comprehensive study, Wang and coworkers have studied the effect of decreasing particle size of crystalline anatase on the nature of energy storage. They found that a decrease in particle size leads to increased pseudocapacitive share. However, the overall capacity gain was moderate [56]. Another aspect is the long term stability of the material. The nanostructured scaffold needs to remain stable in order to guarantee enhanced cycling properties throughout the whole cycle life of the battery. Preventing agglomeration and irreversible structural

10

1 Introduction

changes, leading to a deterioration of the capacities have to be circumvented. None of the nanostructured TiO2 samples, presented in Tab. 1.1 reveals a capacity retention that would satisfy industrial demands (≥ 99.9%) and result in tolerable capacity loss over the battery lifetime. With regard to the increased surface storage for nanostructured materials, Shin et al. point out that the interfacial storage is, differently from the thermodynamic stable bulk intercalation, a non-equilibrium phenomena. It is therefore likely to be subject to capacity loss during the cycle life due to material deterioration [53]. It is also noteworthy that nanostructuring lowers the volumetric density since the packing density of nanostructured materials is generally lower than bulk materials.

1.3.2 Porous structures Besides the reduction of diffusion length and creation of surface sites storage, there are further requirements to promote cycling properties, which cannot be addressed by simple nanostructuring. In order to allow fast transfer of charge carriers, a high contact area of active insertion material and electrolyte is aspired. Furthermore, the nanostructured material should have an open scaffold structure, which allows the homogeneous diffusion of the with electrolyte throughout the whole electrode. In order to assure a constant goodness, the structure should be stable for long time and many cycles. A possible solution to approach these requirements is by furnishing the active TiO2 material with an adequately tailored pore system, e.g. see Tab. 1.1. According to Dylla et al., porosity can enhance both, overall storage capacity and high rate capacity due to improved lithium ion coupled electron transfer kinetics, faster diffusion of Li+ through the interconnected TiO2 network, and increased electrode/electrolyte contact area [51]. Findings by Shin et al. are in line with these results, reporting extremely high rate capability of 46 mAh/g at 60C for hierarchical nanoporous Anatase [53]. Ren et al. point out that size, uniformity in shape and size, and interconnectivity of the pores is of vital importance for the electrochemical cycling properties. They give reason that a mesoporous system is more suitable to fulfill these requirements than a nanoporous one, and report over 300 mAh/g in the first cycle corresponding to the theoretic formation of Li0.96 TiO2 [67]. Liu and co-workers reported also on lithium insertion properties of mesoporous TiO2 (B) microspheres. They claim that benefits of this distinct geometry lie in the ability to increase volumetric energy density, due to efficient microsphere packing, while maintaining the nanostructure needed for fast surface lithiation kinetics [45]. More recently, they reported porous nanosheet TiO2 (B) morphology, which showed impressive rate capability with over 200 mAh/g at 10C over 200 cycles [68]. In spite of the significant enhancement that has been achieved via formation of mesoporous TiO2 , their application in mass scale is still hindered by complex and costly synthesis methods. Establishing a template and surfactant free synthesis method for mesoporous TiO2 is therefore of great interest. Furthermore, the issue of adequate pore size distribution, total pore volume as well as the electrolyte stability have to be regarded

TiO2 (B)

TiO2 (B)

anatase TiO2 (B) 95% TiO2 (B) +5% anatase 87% TiO2 (B) +13% anatase TiO2 (B)

anatase

anatase

anatase

6

30, 5 thick

9-11

15-20

256 200

126 200

216

237

315

190

140

7

227 205

230

220

270

275

190

20

54

67

34

111

34 50

84

336

66

66

170

rev. capac- current rate ity [mAh/g] [mAh/g]

200 -

123

-

80

135

particle/ crys- surface tallite size [nm] area [m2 /g] 100-300, thick170 ness 5

13 20-40 nm 30 nm, nanoribbons length 1-2 µm mesoporous mi- diameter 1 µm, crospheres 5-10 nm grains nanoparticles 20-25 2 µm length needle-like 250 10, thickness nanotubes 2.5

mesoporous (C16-TiO2 ) spherical hollow microspheres + nanotubes mesoporous hollow spheres nanoparticles nanowire

nanosheets

anatase

anatase

morphology

phase

93%(n=2-80)

95%(n=2-50)

86%(n=2-10)

72%(n=2-10)

82%(n=2-10)

76%(n=2-60) 90%(n=2-100)

80%(n=40)

68%(n=2-500)

91%(n=2-30)

93% (n=1-50)

92% (n=2-100)

capacity retention after n cycles

Table 1.1: Physical and electrochemical properties of TiO2 , data taken from [28]

[66]

[65]

[64]

[45]

[63]

[62] [44]

[61]

[60]

[59]

[58]

[57]

ref

1.3 Performance enhancement 11

12

1 Introduction

more closely.

1.3.3 Doping and composite formation Doping and composite formation are often pursued approaches in the quest to overcome the intrinsically low conductivity of TiO2 (10−12 to 10−7 S/cm), which is the main drawback of titania based insertion materials. Its 3d0 electronic configuration and the bandgap of ≈ 3.3 eV make it a typical insulator. However, TiO2 electronic band structure is particularly sensitive not only to punctual defects but also to dopants. Greatest achievements regarding tailoring of band gap via reducing particle size, introducing defects and dopants have been made in the field of photocatalysis, aiming at adjusting the absorption window to the extent of turning the colour of TiO2 from snow-white to black [69–72]. One of the most widely used approaches is the doping with aliovalent dopants, which leads to the formation of electrons or holes depending whether donor or acceptor cation is chosen [73, 74]. Studies show that already small amounts of dopant can provoke a significant increase in conductivity. Tailoring the defect chemistry of TiO2 to adequately adjust the electronic properties was primarily applied to improve the photocurrent of TiO2 [75–77] but lately more and more battery research groups make use of it in order to overcome the low conductivity [78–80] Another defect chemistry related approach to boost conductivity of TiO2 has been demonstrated by Shin et al.. Instead of introducing foreign atoms to create charge carriers and vacancies, TiO2 was treated in reductive atmosphere at elevated temperature in order to create oxygen vacancies and additional electrons. They report that such intrinsically doped TiO2 has enhanced lithium insertion properties due to improved electronic charge transport, although the relationship between conductivity enhancement and rate capability is non linear [55]. From the economical viewpoint, using soft chemistry synthesis methods, which offer the possibility of facile one-pot doping is an interesting solution to overcome the drawbacks of TiO2 . Hence, studying the feasibility of utilizing hydrothermal and sol-gel synthesis method for facile doping is of great interest, along with elucidating the influence of a dopant on the electrochemical and physical properties of the material.

1.4 Objectives The first objective of this thesis is to establish a simple, low-cost synthesis method for TiO2 with facile upscale option. Furthermore, the possibility of tailoring texture, morphology and phase composition of the TiO2 will be elaborated and investigated, regarding its effects on electrochemical cycling properties. In this respect, the synthesis and electrochemical viability testing of nanostructures and mesoporous scaffolds will be emphasized. Once the synthesis method has been set up, doping, formation of composite, and coating shall be attempted to adjust the insertion properties. The leitmotiv of these synthesis and sophisticated material tailoring endeavors, is the creation of a TiO2 based

1.4 Objectives

13

negative electrode material, which has a superior capacity than commercially available LTO (Ctheo =175 mAh/g) while at the same time featuring equally high cycling stability and rate capability. Another main objective of this work is to deepen the understanding of the internal mechanism and processes upon lithium insertion in TiO2 , paying special attention to the lesser-known TiO2 (B) phase and the proclaimed pseudocapacitive storage mechanism. For this, characterization by spectroscopic and diffraction techniques under operando conditions shall be considered.

2 Synthesis In this chapter, different synthesis methods studied and developed throughout my PhD thesis are presented. As basic requirements, cost effectiveness, low toxicity, simplicity and possible upscale were aspired. The characterization techniques that were used throughout this thesis are described in detail in the Appendix, see B.

2.1 Hydrothermal synthesis One of the most widely used methods for synthesizing TiO2 -based materials is under hydrothermal conditions. This procedure is in agreement with the basic requirements stated above as it comprises a facile one-pot method relying mostly on abundant and cheap commercial TiO2 and alkali metal base as the starting compounds. Hydrothermal synthesis proves to be a versatile approach to produce a broad variety of morphologies without the help of templates or surfactants. Also, it bears the possibility of simple composite formation and doping [81–85]. Pioneering work on hydrothermal synthesis of anatase nanotubes was done by Kasuga et al. [86]. Later, this method was adopted by Armstrong et al. for hydrothermal preparation of TiO2 (B) nanowires [43].

2.1.1 Experimental details A detailed flow chart showing synthesis steps and structural processes of the preparation of the materials studied in this work is presented in Fig. 2.1. Throughout the hydrothermal synthesis study different upscale factors were applied, therefore we will give ratios and concentrations rather than absolute values. A TiO2 anatase-rutile mixture from Umicore was used as starting material. TiO2 was added to a strongly alkaline solution (NaOH pellets Sigma Aldrich) and homogenized by magnetic stirring. In this step, hydrated TiO6 octahedral-shaped ions are formed, see Fig. 2.2. Subsequently the milky white suspension is transferred into a Teflon-lined stainless steel autoclave, shut tight and left 72 h at 150 ◦ C in a Memmert UNE400 oven under air. Under such severe conditions, Ti-O-Ti bonds are broken, and structural rearrangement occurs in a dissolution-precipitation mechanism with the formation of sodium hydrogen titanates, with general formula Na2x Tiy Ox+2y *(H2 O)n [43, 87]. After cooling naturally to room temperature, the obtained white sorbet-like suspension was washed to eliminate the remaining sodium hydroxide. This step is followed by washing with 0.1 M HCl solution for ion-exchange, leading to the formation of hydrogen titanate with a layered structure. According to literature, it is in during the acid washing step that one dimensional

15

2 Synthesis

16

Figure 2.1: Flow chart of hydrothermal synthesis.

nanostructured TiO2 is formed throughout protonation and formation of hydrous titania (Ti-OH) [88, 89]. In a final washing step, the pH is brought to 7 by washing-off excessive acid with water and filtering. The obtained sample is then dried overnight in a Büchi oven at ≈100 ◦ C to remove structural and absorbed water molecules and form amorphous or nanocrystalline TiO2 polymorphs. After grinding and sieving with a 50 µm mesh, a fine white powder is obtained, which contains according XRF1 measurements ≤0.7 at% of impurities (mainly Na). This powder is subject to a heat treatment of 4 h @450 ◦ C under air. Main purposes of this final step is to remove remaining structural water molecules, and enhance the crystallization of TiO2 particles.

Figure 2.2: Formation of octahedral-shaped TiO6 ions upon dissolving. Figure based on [90].

1

X-ray fluorescence

2.1 Hydrothermal synthesis

17

2.1.2 Synthesis parameters - dependencies and influences There are numerous publications on how synthesis parameters such as choice and concentration of the TiO2 precursor [91–93], synthesis duration and temperature [43, 89, 94–96], choice, duration of application and strength of acid [89–91, 97] and base [43, 93, 95] influence morphology, crystallinity, phase composition, particle size, etc. of hydrothermally produced TiO2 . The effects of some of these parameters were thoroughly studied also in this work, as resumed in the following paragraphs. Alkalinity and powder concentration Throughout a comprehensive survey of hydrothermal synthesis conditions we were able to define dependencies and trends, at a defined synthesis duration and temperature, regarding morphology and phase composition of the obtained titania. An overview of the different morphologies that could be obtained throughout facile adjusting of the initial solid-to-liquid ratio1 in the mother suspension is given in Fig. 2.3. It shows the nitrogen physisorption curves and TEM2 pictures of the three main morphology groups. All physisorption isotherm follow the same pattern, which is defined as type II curve by IUPAC 3 [98], which connotes monolayer and multilayer adsorption on non-porous materials. The very small opening of hysteresis at pressure above p/p0 ≈ 0.8 and extending up to p/p0 ≈ 1 states that only large pores are present, which are not all filled. This is in agreement with the TEM pictures, which show no small pores inside the particles. The large pores correspond to the spaces between the aggregates. The nanoparticle morphology shows the highest content of such inter-particle voids. The strong difference of the quantity of absorbed nitrogen is directly linked to the specific surface area, which ranges from 250 m2 /g for sheet, 190 m2 /g for nanoparticles and 50 m2 /g for rod structure. Experimentally we find, that for high amounts of titania corresponding to a log(Na/Ti) value below 0.9, a sheet structure is favored, while the decrease in solid-to-liquid ratio towards a log(Na/Ti) value of 1 leads to elongated nanoparticles with average diameter of ≈15 nm. As we further reduce the amount of titania in the mother suspension, these rods grow bigger, and at log(Na/Ti) = 1.3 are composed almost entirely of ribbons or rods with an average diameter of 40-60 nm. The latter is the most commonly found morphology for these synthesis conditions, and has been published by many groups before [30, 43, 94, 99, 100]. To the best of our knowledge, sheet structures are however seldomly obtained throughout this kind of synthesis. Usually the dehydration during drying process is coupled to the formation of cage and tubular structures [101–103]. Presumably a high solid-to-liquid ratio suppresses this formation and hence the intermediate micrometer scale TiO2 sheets are preserved. 1 2 3

refers to the ratio of solid TiO2 powder suspended in NaOH solution Transmission electron microscopy IUPAC: International Union of Pure and Applied Chemistry; is recognized as the worlds authority in developing standards in chemistry

2 Synthesis

18

Figure 2.3: Nitrogen physisorption curves and TEM pictures of three main morphologies; sheets, nanoparticles and rods synthesized throughout adequately adjusting solid-to-liquid ratio (powder concentration) in the mother suspension.

Differently from the morphology, phase composition seems to be principally dependent on the concentration of the alkali metal solution. The XRD1 patterns of three selected samples with different phase compositions are compared in Fig.2.4. All diffraction peaks can be attributed to TiO2 reflecting the absence of crystalline impurities. The low signalto-noise ratio denotes a partial crystallization, while the broadened peaks indicate very small crystallite sizes. These circumstances in combination with the fact of coinciding Bragg peaks of anatase (ICSD2 9852) and TiO2 (B) (ICSD 171670) [104] make Raman an indispensable characterization method for distinguishing and quantifying these phases. The phase ratios were obtained by fitting the area under the Raman peaks shown in Fig. 2.5, according to a quantification method already detailed in the literature [63]. The Anatase-to-TiO2 (B) ratio ranges from anatase rich 80/20 over even mix 40/60 to TiO2 (B) rich 10/90 sample. We found that at 150◦ C, a NaOH concentration lower than 11 M produces mainly anatase, whereas higher concentrations lead to samples rich in TiO2 (B), which is in good agreement with findings by Armstrong et al.[43]. Analogously to the XRD pattern no peaks from impurities could be identified in the Raman spectra. It should be noted that the XRD and TEM study of the samples before the final heat treatment (not shown) reveals that the phase composition is globally preserved, displaying that the

1 2

X-ray diffraction Inorganic crystal structure database


19

Figure 2.4: XRD pattern of three biphasic TiO2 mixtures containing different ratios of anatase and TiO2 (B) phase.

latter step only produces a dehydration of the solid and does not influence the structure. Furthermore, we like to point out the difficulties of obtaining pure phases by this synthesis method while maintaining the morphology. This limitation arises from an interdependence of the two main synthesis parameters, solid-to-liquid ratio of the suspension and alkalinity.

Figure 2.5: Raman spectra of three biphasic TiO2 mixtures containing different ratios of anatase and TiO2 (B) phase.

High Resolution Transmission Electron microscopy (HRTEM1 ) analysis provides further

1

High resolution transmission electron microscopy

20

2 Synthesis

insight into the fine microstructure and crystallinity of the TiO2 mixed-phase samples. While for the anatase-rich sample extended crystalline zones are evident, crystalline areas in TiO2 (B) rich sample are much smaller and less well defined, (see Fig. 2.6 (a) and (b) respectively). This is in line with the reduced signal-to-noise ratio for TiO2 (B) rich samples in XRD patterns. All visible crystalline zones of the anatase rich sample shown in Fig. 2.6(a) have a interplanar distance d of ≈ 0.355 nm corresponding to the (101) plane, which is the dominant exposed plane of the anatase phase [105]. For the TiO2 (B) rich sample, see Fig. 2.6(b), we found three different interplanar distances being (i) 0.587, (ii) 0.357 and (iii) 0.627 nm corresponding to distance between two different (200),(110), and (001) planes, respectively. It is interesting to notice that the latter is the base while 200 and 110 are the side facets of the pseudohexagonal prism, which is the proposed equilibrium shape of TiO2 (B) in calculation by Vittadini et al.[106]. No inhomogenity or phase segregation of TiO2 (B) and anatase was perceived in the micrographs.

Figure 2.6: HRTEM micrograph of (a) anatase-rich and (b) TiO2 (B)-rich sample with magnified cutouts for interplanar distance measurement.

In summary, it can be stated that the here presented hydrothermal synthesis method is a facile, and low-cost approach to reproducibly synthesize highly pure TiO2 polymorph mixtures with a wide range of morphology. We demonstrated that while the alkalinity is mainly influencing the phase composition in terms of anatase/TiO2 (B) ratio, the solid-toliquid ratio (Na/Ti ratio) affects primarily the morphology of the resulting titanate. These trends are condensed in Fig. 2.7, which features TEM images of the variety of different morphologies obtained sheets, nanoparticles and rods as well as structural models of the


21

tetragonal TiO2 anatase and monoclinic TiO2 (B) phase. This figure shows imposingly that a wide range of morphologies and phase compositions can be achieved throughout the simple adjustment of two synthesis parameters of the hydrothermal synthesis. This approach for tailoring TiO2 is an attractive alternative to the use of template or surfactants, which are often costly, toxic, hazardous or are difficult to remove without residue.

Figure 2.7: Trends and dependencies found for hydrothermal synthesis of regarding phase composition and morphology of titania. Increasing dilution (Na/Ti ratio) leads from a) sheets to b) elongated nanoparticles and finally to c) developed nanorods/nanoribbons while increasing the pH promotes TiO2 (B) formation.

The effect of morphology and phase composition on the electrochemical properties will be discussed in Sec. 3.1. Choice of base and starting material, reaction time Within the master thesis of Remi Grosjean, which was conducted in a collaboration between the Collège de France, Paris and the ICG-AIME Montpellier, the influences of hydrothermal synthesis parameters on the properties of the obtained TiO2 were further investigated. The most important finding was surely that by replacing NaOH by KOH for the alkaline treatment, but maintaining similar molarity, highly pure TiO2 (B) could be obtained. In Fig. 2.8(a) XRD pattern of TiO2 (B) rich and TiO2 (B) pure are compared, the latter showing no sign of crystalline anatase or other phases. However, a more reliable tool to distinguish TiO2 modifications is by Raman spectroscopy, the spectra of NaOH solution derived anatase-TiO2 (B) mixtures are compared to the KOH solution derived TiO2 (B)-

22

2 Synthesis

pure sample in Fig. 2.8(b). One can clearly see that ratio of TiO2 (B)-to-mutual peak ratio is inverted for the TiO2 (B) pure sample which confirms the high purity of TiO2 (B) phase. Whether this increased yield in TiO2 (B) phase, which expresses an enhanced phase stability, is linked to reduced hardness of the KOH base and/or the increased size of K+ ion compared to Na+ imposing sterical constraints, cannot be answered with certainty at this moment. However, in spite of several washing steps potassium impurities of ≈1-2% in the final product could not be avoided compared to ≤0.5% of Na for NaOH-based synthesis. It remains however nebulous whether the increased amount of impurities is a cause or a consequence of the enhanced TiO2 (B) formation. Understanding this connection demands further in depth research. Since the samples containing TiO2 (B) pure phase were only obtained towards the end of this thesis the characaterization of these samples remain incomplete. Furthermore, it was shown that the phase composition of the hydrothermal synthesis product is independent of the phase composition of the titanium oxide precursor. This is in line with the aforementioned dissolution-reprecipitation mechanism of the starting TiO2 into the final alkali-hydrogen titanates.

Figure 2.8: (a) Comparison of diffraction patterns of TiO2 (B)-rich and TiO2 (B)-pure sample and (b) Raman spectra of Anatase/TiO2 (B) mixtures derived from NaOH solution and TiO2 (B)-rich derived from KOH solution.

Another noteworthy effect is the reduction in yield of TiO2 (B) when synthesis time is reduced from 72 to 48 h indicating that the dissolution and formation of hydrogen titanates is a time extensive process. The feasibility of using the hydrothermal synthesis for simple one pot doping of TiO2 with other transition metals by adding the wanted amount of an adequate precursor was also investigated and will be further discussed in Sec. 4.2.

2.2 Sol-gel synthesis

23

2.2 Sol-gel synthesis Sol-gel is a wet-chemical, cost-effective, low-temperature method in which an oxide network is obtained throughout the polycondensation reaction of molecular precursor in solution. The general reaction of the sol-gel process is schemed in Fig. 2.9.

Figure 2.9: Scheme of sol gel process.

This technique allows tailoring of the composition and the microstructure throughout controlling the precursor chemistry and processing conditions. Numerous publications demonstrate the feasibility of this method for facile synthesis of nanometer sized crystalline TiO2 powder of high purity at relatively low temperature [107]. Furthermore, this method features the option of synthesizing porous TiO2 with and without the aid of templates [108, 109] or surfactants [110, 111]. Literature states that extra capacity as well as enhanced rate capability can be obtained with porous titania electrode materials, which is mostly attributed to easy accessibility of storage sites in combination with higher Li+ diffusion rates [109, 112, 113]. Another asset of the sol-gel method is the facile doping possibility by adding an adequate precursor solution to the mother sol. This method is an efficient way to tailor the electronic structure of the host material [114–117]. Coupling the sol-gel process with diverse processing techniques such as spin-coating, electrospinning, dip-coating, etc. expands even further the versatility of this synthesis method [118–120]. Recent years have witnessed a growing development of electrospun TiO2 -based electrodes for lithium-ion batteries, fuel cells and other conversion and energy storage devices [119, 121–124]. Therefore we have focused on two main routes of sol-gel synthesis. Firstly a conventional “petri dish” approach and secondly electrospinning aided sol-gel synthesis, which are presented in the following paragraphs.

2 Synthesis

24

2.2.1 Conventional petri dish The conventional petri dish approach is the standard precipitation of a bulk gel material using only common lab equipment, such as glassware, magnetic stirrer and ultrasonic bath. We followed synthesis method as proposed by Yu et. al [125], using 9.7 ml titanium tetraisopropoxide (TTIP) (97% Aldrich) as the Ti-precursor, which was mixed with 0.92 ml acetic acid (Sigma-Aldrich) to control hydrolysis rate and 20 ml unhydrous ethanol as solvent. With the help of the training student Matthieu Toiron from IUT Montpellier, three different sol-gel processes were pursued using these reactants. • conventional: Precursor solution added dropwise into distilled water under stirring • sonicated: Dropwise addition of precursor solution into water while applying high energy ultrasound to the sol-gel solution • template assisted: Triblock copolymer P123 is added as template to the precursor solution and added dropwise to distilled water under stirring A mutual process of 3 h drying in glass oven under vacuum at 100 ◦ C and subsequent 1 h heat treatment at 400 ◦ C in air completes the synthesis process. The three different sol-gel syntheses lead to TiO2 anatase with distinct morphologies. The characterization of the three obtained materials are shown in the following paragraphs. Conventional Petri dish sol-gel synthesis without the aid of template or ultrasound lead to homogeneous white powder with a very faint shade of grey. We suspected therefore the presence of impurities from the synthesis. Raman reveals presence of carbon, which was quantified by XRF to ≈ 5 %. XRD shows no second phase besides anatase crystallites with an average size of 17 nm. TEM image reveals spherical particles of ≈ 20 nm, in line with the XRD analysis, which agglomerate to form bigger spherical particles of 70-200 nm diameter. SEM1 confirms that an agglomerate diameter of around 200 nm is indeed the most common. Fig. 2.11(a) shows the nitrogen physisorption curve revealing typical type II curve, which implies the presence of only macropores but no mesopores, see Appendix B for more details. This is confirmed by the pore size distribution, which shows only few pores with diameter of 50 nm, 2.11(b). The specific surface area of 26 m2 /g derives mostly from external surface. Ultrasound assisted sol-gel synthesis Homogeneous white powder was obtained by this method, with no impurities and merely some traces of rutile besides the main phase anatase. 1

Scanning electron microscopy


25

Figure 2.10: SEM and TEM image of conventional sol-gel synthesis sample.

Figure 2.11: a) Physisorption isotherm and b) poresize distribution of conventional sol-gel synthesis sample.

The spherical particles found in TEM picture, see Fig. 2.12, have an average diameter of ≈ 13 nm, which is in good agreement with the crystallite size found via Williamson Hall plot1 . One can furthermore see, that these crystallites tightly agglomerate to form much bigger particles. From SEM micrograph we can infer that those agglomerates are somewhat spherical and can grow up to 1 µm in diameter. Nitrogen physisorption isotherm, see Fig. 2.13 a) resembles a type IV curve, which means that monolayer adsorption as well as capillary condensation is occurring. The latter is expressed by a strong hysteresis and is the fingerprint of mesoporous materials. In Fig. 2.13 the absorption isotherm as well as the pore size distribution based on density functional theory DFT are shown. A specific surface area of 100 m2 /g, which is composed of external as well as internal surface was calculated using BET2 theory. An average pore 1 2

An approach to evaluate crystallite size based on the effect of XRD line broadening due to finite crystal size and micro strain, see Appendix B.4.1 Brunauer-Emmett-Teller

2 Synthesis

26

Figure 2.12: SEM and TEM image of ultrasonic assisted sol-gel synthesis.

width of around 6 nm is found, which is considered small mesoporosity according to IUPAC1 [98]. The creation of small mesopores throughout application of ultrasound results in doubling of the pore volume per gram of material.

Figure 2.13: a) Physisorption isotherm and b) poresize distribution of ultrasound assisted sol-gel synthesis TiO2 .

Sol-gel synthesis of TiO2 using triblock copolymer P123 As a structure directing agent we used nonionic triblock copolymer Pluronic 123, which consists of two symmetrical hydrophilic poly(ethylene oxide) (PEO) blocks and one hydrophobic poly(propylene oxide) (PPO) block, see Fig. 2.14. Depending on external conditions (concentration, temperature), it forms spherical or cylindrical micelles with hydrophobic core (PPO) and hydrated outer shell (PEO) [126]. HO(CH2 CH2 O)20 (CH2 CH(CH3 )O)70 (CH2 CH2 O)20 H

1

International union of pure and applied chemistry


27

Figure 2.14: Schematic drawing of triblock copolymer P123, consisting of two hydrophilic PEO blocks, which are connected via hydrophobic PPO block. Figure taken from [127].

Keeping all other reactants equal we added 3.7 mg of P123 following a procedure proposed by Zhao et al. for mesostructured meso-macroporous anatase [108]. By titrating ≈90 ml at 25◦ C we assured to be far above the proposed critical micelles concentration of P123 of 4*10−3 wt% [128] which is precondition for forming porous scaffolds. After the heat treatment, a homogeneous bright white powder was obtained, showing no sign of residues or impurities by visible inspection, Raman or XRD. X-ray diffraction reveals almost pure anatase phase with some traces of brookite, and an average crystallite size of about 9 nm was determined, using the aforementioned Williamson-Hall approach. TEM pictures show spherical particles with average diameter of 10 nm see Fig. 2.15, and therewith slightly smaller than the sonicated spheres. More striking is the difference in agglomeration of the particles. The sample derived from template assisted sol-gel shows less agglomeration and more open spaces between particles. SEM confirms that particles form less compact and much smaller agglomerates.

Figure 2.15: SEM and TEM image of sol-gel synthesis using P123 as template.

28

2 Synthesis

Nitrogen physisorption indicates mesoporosity and a significant specific surface area, see Fig. 2.16. In detail, we obtain a specific surface area of 144 m2 /g and average pore width of 13 nm, principally in the form of interparticle voids. Such porosity caused by the stacking of particles is referred to as textural porosity [129], and can be observed in Fig. 2.15. Besides enlarging the pore size, the application of template also leads to quadrupling of pore volume per gram of TiO2 .

Figure 2.16: a) Physisorption isotherm and b) poresize distribution of template assisted sol-gel synthesis TiO2 .

2.2.2 Electrospinning Within the wide range of synthesizable nano-morphologies via sol-gel, 1-D nanostructures such as nanotubes and nanofibers are particularly interesting for application in LIB. Their large surface-to-volume ratio provides for enlarged contact of electrode and electrolyte, while their vectorial electron (along the long dimension) and Li+ (along the lateral direction) transport properties can enhance the low ionic and electronic conductivity of TiO2 [80, 130, 131]. Among the methods for generating 1D nanostructures, electrospinning is a simple and versatile approach for preparing nanofibers and nanotubes of polymers [132], composites [133] and ceramics [134] with controlled and reproducible diameters, particle size and shape as well as phase composition. Furthermore, it is a low temperature, cost effective approach, offering facile doping and particle embedding via implementation of suitable precursors [121, 135, 136], which will be discussed in more detail in Chapter 4. To obtain the nanofibrous TiO2 , the carrier polymer solution based on polyvinyl pyrrolidone (average Mw ≈ 1,300,000 g/mol, Aldrich) in absolute ethanol (puriss., SigmaAldrich) is mixed with a solution containing 0.52 ml of titanium(IV)isopropoxide (97%, Aldrich), and 1 ml of acetic acid (Sigma-Aldrich). Electrospinning of the final solution is carried out in air at room temperature with a standard syringe and a grounded collector plate configuration. The distance between the needle tip and the collector plate is 10 cm, the applied voltage 15 kV and the flow rate 0.5 ml/h. The as-prepared fibers were calcined


29

Figure 2.17: Scheme of electrospinning in which precursor solution is fed constantly through a syringe into strong magnetic field leading to a continuous fiber, which is collected on a plate.

in air at 500 ◦ C with a heating rate of 5 ◦ C/min for 6 hours in order to remove the carrier polymer. The set-up of an electrospinning apparatus is schemed in Fig. 2.17. In Fig. 2.18 (a) TEM and (b) SEM micrograph of electrospun TiO2 fibers are shown. The contrast differences in TEM picture can be attributed to porous structure as well as grains with different crystallographic orientation. The SEM picture reveals clearly that the axial dimension of the fibers is nanoscale (≈80±25 nm) while there longitudinal extension is in the micrometer range.

Figure 2.18: (a) TEM and (b) SEM micrograph of electrospun TiO2 nanofibers.

From Fig. 2.19 it is apparent that the dominant phase of the electrospun nanofibers is anatase. While in the diffraction pattern minor contribution of rutile (ICSD 1781510) as well as traces of brookite (ICSD 30380) can be identified, all the peaks in Raman spectra at 144, 197, 397, 516 and 639 cm−1 are attributed to anatase modes E∗g , Eg , B1g , A1g / B1g and Eg respectively [137]. Neither of the principal peaks of rutile 447 cm−1 nor brookite 124 cm−1 are found. This discrepancy is most probably due to the lower sensitivity of Raman spectroscopy compared to XRD. From the physisorption isotherm shown in Fig. 2.20, a specific BET surface of 44 m2 /g was calculated. The adsorption and desorption curve form a slight hysteresis resembling a type IV curve, which agrees with the mesoporous character of the sample. The pore size

2 Synthesis

30

Figure 2.19: (a) XRD pattern and (b) Raman spectra of electrospun TiO2 nanofibers.

distribution, shown in Fig. 2.20(b), is in line with this observation, stating a contribution of over 40 % of the total pore volume by mesopores (2-30 nm), which can be attributed to pores inside the nanofibrous structure. The other part is issuing from large mesopores (≥40 nm) and macropores, which can be attributed to interparticle spacings in between the nanofibers.

Figure 2.20: (a) Hysteresis of physisorption isotherms and (b) incremental and cumulative pore size distribution for electrospun TiO2 nanofibers.

2.2.3 Summary and Discussion In order to facilitate their comparison, key parameters of phase composition and specific morphology of the different samples that were obtained throughout adjusting the conventional “petri dish” and electrospinning sol-gel synthesis methods are summarized in Tab. 2.1. The table contains also features of hydrothermally synthesized anatase rich sample as reference material.

100 m2 /g

144 m2 /g

44 m2 /g

meso(5-20 nm) & macrop- mostly macroporous (≤40 nm), orous, 1.2 cm3 /g 3 0.75 cm /g

mesoporous ≈13 nm, 0.4 cm3 /g

180 m2 /g

15±3 nm

17±2 nm

9±2 nm

⊘

20%

14±5 nm

nanorods ≈10 nm

anatase, traces of ru- anatase, tile TiO2 (B) ≤1% Na

Electrospun

porous nanofibers, ⊘ ≈80 nm

template assisted Anatase, traces of brookite -

hydrothermal synthesis

particles ⊘ particles ⊘ little ≈15 nm, agglom- ≈13 nm erates ≤1 µm agglomeration

Anatase, traces of rutile -

sonicated

few big mesopores, small mesopores 0.06 cm3 /g ≈6 nm, 0.1 cm3 /g

specific sur- 26 m2 /g, external surface face area

pore width & volume

phase comonly anatase position residue 5 % carbon particles ⊘ ≈20 nm smooth morphology agglomerates ⊘ ≈ 200 nm crystallite 17±3 nm size

conventional

Sol-gel synthesis

Table 2.1: Summary of sol-gel synthesized samples

2.2 Sol-gel synthesis 31

32

2 Synthesis

Tab. 2.1 shows imposingly the variety of different morphologies and porosities that can be obtained through simple modification of the sol-gel process. In contrast to hydrothermal method, we succeeded in producing mesoporous TiO2 without the aid of template or surfactant simply by applying ultrasounds. However, in order to tailor the poresize in a wide range, the use of template seems to be inevitable. Furthermore, the use of a template appears to be beneficial for avoiding agglomeration of particles, hence leading to a more open structure and to an increase in specific surface area. Electrospinning leads to homogeneous, almost pure anatase nanofibers with a narrow distribution of the diameter and elevated specific pore volume, while specific surface area remains comparatively small. The electrochemical cycling performance of these samples is evaluated in Sec. 3.2

3 Lithium insertion sites and mechanism In this chapter, the role of intrinsic characteristics of TiO2 such as local and long-range structure, morphology and texture are discussed regarding its lithium insertion properties. The techniques used for the characterization of electrochemical and structural properties are explained in detail in the Appendix B.

3.1 Influence of phase composition and morphology on electrochemistry From the variety of samples synthesized using hydrothermal synthesis method (see Sec. 2.1), three samples were chosen for intensive electrochemical investigation. These samples feature different phase composition but similar nanorod morphology. An anatase rich sample (80/20), an even mix composition sample (40/60) and a TiO2 (B) rich sample (10/90) with specific surface areas of 180, 70 and 50 m2 /g respectively, were tested. In Fig. 3.1 the galvanostatic curve of the first two cycles at cycling rate C/201 are pictured for a) TiO2 (B) rich sample (10/90), b) even mix (40/60) and c) anatase rich sample (80/20). It is noteworthy to say that electrochemical cycling was carried out using EC2 :PC3 :3DMC4 1M LiPF6 in "half cell“ configuration with TiO2 based electrode as positive (cathode) and lithium metal as negative (anode), see Appendix B.1. When comparing the electrochemical cycling curves, one notices that TiO2 (B) rich sample shows a continuously decreasing "S"-shaped slope, which indicates a solely solid solution storage mechanism, while the anatase rich sample reveals a more complex multistep insertion mechanism. For both systems, the galvanostatic curves agree well with those published in the literature ([63, 138, 139] and [100, 140, 141], respectively). The absence of any sharp insertion plateaus is well reflected by the presence of broad insertion peaks in the derivative of TiO2 (B) rich sample, see Fig. 3.1(a). Besides the dominant characteristic twin peak at ≈1.5 V a minor reduction peak just below 1.8 V is observed, which coincides with the insertion peak of Li+ into the anatase phase. The fact that no corresponding de-insertion peak is found makes the origin of this peak nebulous. While Armstrong et al. have attributed it to insertion in nanosized TiO2 (B) [44] it is not reported for micro

1 2 3 4

cycling rate of 1C corresponds to the insertion of 1M lithium per 1M TiO2 , equivalent to a current of 336 mA. Ethylene carbonate Propylene carbonate Dimethyl carbonate

33

34

3 Lithium insertion sites and mechanism

sized TiO2 (B) [142, 143]. The insertion sites and insertion mechanism of the TiO2 (B) are discussed in more detail in Sec.3.5 and Sec. 3.4, respectively, at the end of this chapter. In the anatase rich sample, see Fig. 3.1(c) firstly a topotactic insertion (x≤0.1) based on solid solution mechanism takes place, which extension is largely defined by the particle size [54]. This is followed by a characteristic plateau at roughly 1.8 V expressing the first order transition from tetragonal anatase to orthorhombic Li0.5 TiO2 up to about 0.35 lithium. The insertion de-insertion reaction plateaus are nicely reflected by sharp and well defined peaks in the derivative at 1.78 and 1.84 V , respectively. Further lithium insertion is characterized by a slope-like insertion corresponding to continued insertion into the new formed orthorhombic phase. The increased share of monophasic slope compared to biphasic plateau insertion mechanism is a phenomena often observed for nanostructured materials [54, 139, 144]. It can be related to various effects; to surface strain-induced free energy variations resulting in change of Li+ insertion site energetics, a particle size dependent potential leading to distribution rather than single voltage plateau, furthermore to energetic changes on the interface of the coexisting phases, as well as a reduction of miscibility gap [51, 145]. The even mix sample, see Fig. 3.1(b) shows similar behaviour but with a shorter plateau compared to the anatase-rich sample. These differences are in agreement with the fact that, although anatase and TiO2 (B) are chemically similar, lithium insertion and de-insertion reaction follow different mechanisms. The capacity loss between first and second cycle is a frequently seen phenomena for TiO2 materials. For the three samples here investigated an irreversible capacity of ≈20 % was found regardless of their specific surface area differences or phase phase composition. This low coulombic efficiency is attributed in the literature either to the electrolyte decomposition induced formation of a solid electrolyte interphase (SEI) and/or filling of irreversible lithium sites in the TiO2 structure [44, 63, 140, 144, 146, 147]. Experiments to investigate the origin of this phenomena are indicating a degradation of certain electrolyte components occurring at potential values far above SEI formation potentials within their supposed stability window, see Sec. 5.3.1. This might be related to the catalytic effect of high specific surface area nanostructured TiO2 . In their study Brutti and coworkers showed that a surface treatment of TiO2 (B) nanowires can significantly reduce the extent of the capacity loss during the first cycle [66] therefore it is assumed that the contribution of Li+ trapping plays only a minor role. See also Sec. 5.3.1 for more detailed information on irreversible capacity and electrolyte decomposition. To evaluate the rate capability and capacity retention of the samples several consecutive cycling runs with the same cell while progressively increasing the cycling rate were carried out, which leads to a step function behaviour (not shown). In general the capacity retention is somewhat reduced during the first 20 cycles at C/20 but increases and stabilizes at high value upon further cycling, the reason for this are probably the aforementioned parasitic reactions, which are more influential at slower cycling rate. The fact that same capacity retention is found when returning to low cycling speed C/20 is in line with this argumentation. To facilitate the comparison of their electrochemical storage ability the

3.1 Influence of phase composition and morphology on electrochemistry

35

Figure 3.1: (left:) Potential as function of mole of lithium inserted per mol TiO2 for the first two galvanostatic cycles @ C/20 and (right): corresponding derivatives of titania nanorod samples with anatase-to-TiO2 (B) ratio of (a) 10/90 (b) 40/60 and (c) 80/20.

mean capacities of the 10th cycle as function of cycling rate for the three samples with different phase composition are plotted in Fig. 3.2. This graph reveals an inverse relation between cycling speed and capacity obtained. Although this is a mutual trend for three investigated samples, significant differences were found. The mutual trend of reducing capacities as cycling speed is increased, can be attributed to growing kinetic limitations. At the low cycling rate of C/20, capacity values of different phase composition lie between 200 and 220 mAh/g. However, the capacities obtained for anatase rich and even mixture samples decrease more rapidly than the capacities of TiO2 (B) rich sample, once the cycling rate is accelerated. At 5 C, anatase rich and even mixture sample reach only ≈ 70 mAh/g and 95 mAh/g, respectively, while for the TiO2 (B)-rich sample a capacity of 120 mAh/g is obtained. This result is in line with the general idea that the anatase phase is less capable of lithium ion insertion and de-insertion at elevated rates than the bronze phase [42, 45, 46, 139]. Such superior performance of TiO2 (B) is in agreement with values of literature 190 mAh/g at 0.65 C and 100 mAh/g at 6C [44], 202 mAh/g at 0.65 C and 140 mAh/g at 6 C [148] as well as 175 mAh/g at C/3

36


Figure 3.2: Evolution of capacities for samples of high (80/20), medium (60/40) and low content (10/90) content of Anatase as function of cycling rate.

and 75 mAh/g at 6C [63]. Naturally these values can only be taken as a rough orientation for comparison since key factors such as electrode formulation and preparation, specific surface area and electrode loading are staggering. In order to further elucidate this phenomena, we have studied the derivative curves of the galvanostatic cycling, which allows one to determine not only the exact electrochemical potential of lithium ion insertion and de-insertion reaction of the anatase and TiO2 (B) phase, but also estimate their intensity (peak area). In Fig. 3.3 two derivative curves of an even mixture sample (40/60) cycled a) at C/20 and b) 5C are presented. Characteristic twin peaks of TiO2 (B) at ≈1.5 V are identified as well as the anatase contribution at ≈1.8 V [149].

Figure 3.3: Derivative curves of an even mix sample at a) C/20 and b) 5C. Characteristic lithium insertion and de-insertion peaks of the two coexisting phases are marked.

Two things are salient when comparing the derivatives at different cycling rates: firstly, the proportional changes of the peaks, which correspond to the reaction of anatase and of the TiO2 (B) phase. In fact, while anatase peaks are striking at C/20, they are faint

3.2 Influence of textural properties on electrochemical cycling

37

at 5C. Secondly, a smaller shift of the peak positions for TiO2 (B) than for anatase upon increase of cycling rate is found, which is shown in Fig. 3.4 for an even mixture sample (60/40). The potential shift was calculated by subtracting the reduction potential (Li+ insertion) from the oxidation potential (Li+ de-insertion), taken from the derivative curve. This indicates that anatase compared to TiO2 (B) bears higher internal resistance, which is in line with the superior rate capability of TiO2 (B) stated in Fig. 3.2. This leads us to the conclusion that anatase phase is less suitable for high performance cycling than TiO2 (B) phase.

Figure 3.4: Shift redox potential as function of cycling rate for the anatase and TiO2 (B) phase component of even mixture(40/60) sample including standard deviation.

Although chemically equal, anatase and TiO2 (B) feature different electronic and crystal structure, which determine not only the diffusion rates but also the position, accessibility and energetics of lithium insertion sites [38, 46, 68, 99, 140, 150, 151]. In this lies the reason for the strong deviation between redox potentials of lithium insertion and de-insertion of the anatase and TiO2 (B) phase, as well as the distinct differences in polarization of the two phases.

3.2 Influence of textural properties on electrochemical cycling As briefly mentioned in Sec. 2.2, numerous studies have been published on the enhancement of electrochemical properties of TiO2 by adequately tailoring specific morphology and texture. Special attention has been drawn on developing porous structures, as such combine several advantages, which are promising to boost power density of TiO2 as lithium insertion material. Firstly, an open structure provides intimate intersection of participating components and buffers volume changes upon cycling, secondly small particles sizes effectively reduce the diffusion lengths, and thirdly, increased specific surface area enables fast surface diffusion as well as pseudocapacitive storage [45, 109, 112, 113, 152– 154]. In order to investigate the influence of textural properties on electrochemical cycling

38


of TiO2 anatase, the three different conventional ”petri dish“ sol-gel synthesized anatase samples (conventional, ultrasound assisted, template assisted) were used to prepare film electrodes from a NMP based slurry as described in Sec. 5.2.

3.2.1 Conventional sol-gel The galvanostatic curve obtained for the sample synthesized by sol-gel without the use of surfactant nor ultrasound at cycling rate C/20 with corresponding derivative curve is presented in Fig. 3.5(a). It shows the biphasic plateau followed by a slope like curve typical of the anatase phase. Besides the characteristic insertion and de-insertion peaks at ≈1.8 V we find also a cathodic peak, which emerges after 20 cycles at ≈1.45 V . The fact that no corresponding distinct anodic peak can be attributed, eliminates TiO2 (B) or brookite phase as the possible cause [43, 155]. This peak would also comply with the insertion of lithium into the rutile phase, but a corresponding de-insertion in a wide range around 2 V [156] is not observed either. This example demonstrates the great potential of electrochemical cycling technique for identifying and distinguishing TiO2 phases, complementary to diffraction and spectroscopic techniques.

Figure 3.5: a) Galvanostatic cycling curve with corresponding derivative and b) rate capability of TiO2 anatase from conventional sol-gel synthesis.

The rate capability is shown in Fig. 3.5(b). The maximum capacity values of around 150 mAh/g are below those for anatase rich samples prepared by hydrothermal synthesis, which achieved almost 200 mAh/g. Moreover, the values at high cycling rate (5C) fall behind those of hydrothermally synthesized anatase. Noteworthy is the low capacity retention at cycling rate 3C. A slope like behaviour with loss of 30 % of original capacity is found, see Fig. 3.5(b). The analysis of the corresponding derivative (not shown) reveals a drastic flattening of the potential peaks, indicating a loss of storage capacity. This could be related to instability of the morphology, forming bigger agglomerates as cycling rate is raised, or to increased kinetic limitations, which reduces the share of material actively contributing to the storage of lithium ions.


39

3.2.2 Ultrasound assisted sol-gel For the sonicated sol-gel sample, galvanostatic curve with corresponding derivative as well as the capacities for progressively increased cycling rates are presented in Fig. 3.6. The galvanostatic curve is similar to that of the preceding samples. It clearly depicts the insertion and de-insertion of lithium ions into anatase. Additionally we observe an event upon first discharge (lithium insertion) at ≈ 2.5 V leading to significant irreversible lithium consumption. The influence of water was discarded as the samples were dried under vacuum at 200 ◦ C in glass oven and then transferred directly into glove box without being exposed to ambient air. Also the influence of residues or impurities to this systematic phenomena can be excluded. We therefore assume that irreversible lithium consumption is probably linked the specific morphology of micropores, evoking a parasitic reaction or trapping of Li+ . Since significant amounts of rutile were found besides dominant anatase phase for this sample in XRD and Raman, we expect to find contribution of both phases in the electrochemical signature of this sample. In the first insertion curves (depicted in black), see Fig. 3.6(a), we observe a broad flat peak which upon cycling transforms into a sharp peak (red curve) at 1.45 V , which can be attributed to the rutile phase. The reason for this change in Li+ insertion mechanism could be the gradual growth of the particle size leading to a changed and more clearly defined lithium insertion potential. The capacities obtained for this sample are even lower than those for first synthesis. Only 120 mAh/g are achieved at the lowest cycling rate (C/20), and ≈20 mAh/g at the highest cycling rate (5C). This reduced cycling performance might be related to inauspicious morphology as well as to the increased share of rutile. Rutile is known to be an unfavorable phase for the Li+ insertion, unless it is elaborately nanostructured [40, 147]. In fact, Kubiak et al. [157] claim an irreversible biphasic reaction of rutile to LiTiO2 at 1.4 and 1.1 V , which limits its abilities to serve as lithium ion host. This is in line with our observation of the absence of distinct anodic peak of the rutile phase.

3.2.3 Template assisted Electrochemical cycling of template sol-gel TiO2 is illustrated in Fig. 3.7. The galvanostatic curve shows as before mostly anatase character. However, in the derivative curve two peaks at 1.5 and 1.75 V are identified and attributed to the cathodic and anodic insertion reaction of the small portion of brookite [155], respectively. The capacities obtained for this sample are the highest of all the sol-gel TiO2 materials and are comparable with those of hydrothermally synthesized anatase. For slow cycling rate C/20 values slightly above 180 mAh/g were achieved and at high cycling rate 5C ≈65 mAh/g.


40

Figure 3.6: a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of TiO2 anatase from ultrasound assisted sol-gel synthesis.

Figure 3.7: a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of TiO2 anatase from template assisted sol-gel synthesis.

3.2.4 Electrospun anatase nanofibers The electrode films were prepared using the conventional procedure by grinding the nanofiber mats, adding carbon black and PVdF1 , dissolving the mixture in NMP2 , and tape casting it on a copper foil. The galvanostatic electrochemical cycling curve and corresponding derivative for electrospun TiO2 nanofibers is presented in Fig. 3.8(a). The curve reveals typical insertion characteristics of Li+ in anatase phase, being the plateau 1 2

Polyvinylidene fluoride N-Methyl-2-pyrrolidone


41

at 1.8 V followed by a slope like insertion. Noteworthy is the relatively high irreversible lithium insertion during the first cycle of almost 0.25 Li. The vault of the derivative curve of the first cycle between 1.5-1.2 V indicates the presence of non-reversible parasitic reactions, which are not present in the second cycle. We can therefore attribute the capacity loss during first cycle to these reactions. Furthermore, we observe an increase of insertion potential of 0.03 V from the first to the second cycle. Such a reduction of overpotential for the lithium insertion reaction has been previously observed, however, not as pronounced as for this sample. We assume that it is related to surface morphology transformations during the first lithium insertion, facilitating the charge transfer. Upon further cycling, no parasitic peaks are identified and the position of insertion and de-insertion potential remain constant; it is however apparent that the extension of the derivative peaks decrease. Since the area under the derivative peaks is a measure for the amount of electrons exchanged, this indicates a decrease in total capacity. This reduced capacity retention at C/20 is clearly visible in the fig. 3.8(b). The overall capacity values and rate capability is relatively low and in the same range as that of the conventional sol-gel samples, see Sec. 3.2.1.

Figure 3.8: a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of electrospun TiO2 anatase nanofibers.

3.2.5 Conclusion Several trends and dependencies can be deduced by carefully analysing and comparing the above presented results, which are summarized in Tab. 3.1. Comparing conventional and sonicated samples, we observe that the increase of specific surface area does not necessarily yield to an increase of capacities (150 vs. 120 mAh/g @ C/20); nonetheless, it does improve rate capability (20 vs. 30 mAh/g @ 5C), see Fig. 3.9. We can therefore note that mesoporous structuring does not automatically lead to extra capacity. A comparison between ultrasound and template assisted sample shows that size and confinement of

42


the pores are equally important, which is in agreement with findings by Shin et al. [53]. Increased average pore diameter and specific pore volume lead to a much enhanced total capacity (183 mAh/g @ C/20) as well as improved rate capability (65 mAh/g @ 5C). In this regard, the most beneficial effect of the structure aiding template is preventing agglomeration of small particles and formation of larger structures as for conventional and sonicated sample see Fig. 2.10 and 2.12, respectively. Another noteworthy property is the particle size, which reduces by a factor of two for the petri dish sol-gel samples on going from the conventional to the ultrasound assisted to the template assisted material. Wagemaker et al. demonstrate in their study that decreasing anatase particle size is beneficial for solubility of Li+ in TiO2 and strengthens the solid solution domains, which hence leads to increased capacities [54]. Side effect of enhanced cyclability for these materials is claimed and attributed to facilitation in accommodating of structural changes[144]. Compared to the hydrothermally synthesized TiO2 anatase, even the best performing of the sol-gel prepared samples shows inferior electrochemical properties (183 vs. 199 @ C/20 and 65 vs. 75 @ 5C). This can be attributed to the three times higher average pore volume, although its mostly made up by macroporous voids, in combination with the increase in specific surface area. Concluding, these results underline the importance of an open, nanostructured and stable confinement for insertion materials providing enlarged contact of electrode and electrolyte in combination with reduced Li+ diffusion path lengths, which enable the participation of maximum share of available active material to the electrochemical cycling process.

Figure 3.9: Reversible capacity at slow (C/20) and high (5C) cycling rate as function of specific surface area for differently synthesized TiO2 anatase samples. Showing general trend of increasing capacity with increase of specific area, with deviating of the sonicated sample.

The case of the electrospun material is somewhat ambiguous. We observe comparatively


43

low rate capability and capacities, see Fig. 3.9, which lie in the range of the conventional petri dish sample, 140 and 150 mAh/g @ C/20 together with 10 and 20 mAh/g @ 5C, respectively. This is astonishing as the electrospun material features increased specific surface area (44 vs. 26 mAh/g) and even more importantly much higher specific pore volume (0.75 vs. 0.06 cm3 /g). These are assets that should provide for potential high power energy storage. The fact that this is not observed for the electrospun material could be related to the undesired reactions observed in electrochemical cycling during first cycle. The deposition of decomposition products leads to formation of a passivation layer hampering cycling, for more details see Sec. 5.3.1. Also the lack of optimization of electrode formulation can have an enormous influence. A unique morphology, such as the nanofibers, requires an optimization of processing and electrode formulation to unveil its potentials. Due to the limited amount of material available, we could not carry out such an optimization within the framework of this PhD thesis, but this could be an interesting point to develop in a future study of such systems. The importance of electrode formulation will be further discussed in in Sec. 5.2.

3 Lithium insertion sites and mechanism 44

Sol-gel synthesis sonicated

20%

hydrothermal synthesis

anatase, TiO2 (B) ≤1% Na

anatase, traces of rutile -

template assisted anatase, traces of brookite -

anatase, traces of rutile -

porous nanofibers, nanorods ⊘ ≈10 nm ⊘ ≈80 nm

Electrospun

Table 3.1: Morphological and electrochemical properties of sol-gel synthesized samples

conventional phase comonly anatase position residue

spherical particles spherical particles ⊘ ≈15 nm, ag- ⊘ ≈13 nm little glomerates ≤1 µm agglomeration

17±2 nm

morphology

9±2 nm

meso5-20 nm and macrop- mostly macroporous orous ≤40 nm , 1.2 cm3 /g 0.75 cm3 /g

5 % carbon particles ⊘ form ≈20 nm smooth spherical agglomerates ⊘ ≈ 200 nm

17±3 nm

mesoporous ≈13 nm 0.4 cm3 /g

180 m2 /g

199 @C/20, 75 @ 5C

44 m2 /g

10

144 m2 /g

140 @C/20, @5C

low coulombic efficiency at slow cycling rates C/20 & C/5, retention low at 3C 183 @ C/20, 65 @ 5C

highest rate capability & coulombic high irreversibility efficiency, capac- 1st cycle, low reity retention un- tention @ C/20 stable

15±3 nm

crystallite size

few big mesopores, small mesopores 0.06 cm3 /g ≈6 nm, 0.1 cm3 /g

14±5 nm

pore width & volume

100 m2 /g

low rate capabil- high irreversibility ity, low retention 1st cycle, low rate at 3C capability

specific sur- 26 m2 /g, external surface face area electrochemical cycling

20

150 @C/20, @5C

120 @ C/20, 30 @ 5C

capacity [mAh/g]@ cycling rate

3.3 Operando measurements

45

3.3 Operando measurements In order to evaluate the potential and properties of TiO2 as insertion material, it is crucial to elucidate and understand the root foundation of its electrochemical cycling characteristics. Therefore we developed and carried out a variety of operando characterization measurements to study the internal processes occurring upon lithium insertion and de-insertion. In order to couple electrochemical cycling with other characterization techniques, a special type of electrochemical cell has been designed [158], see Fig. 3.10. This cell allows acquisition of diffraction or spectroscopic data during electrochemical cycling. It uses a Beryllium window, which also works as the current collector. The sturdy stainless steel body in combination with the plunger and Teflon joints provide for the tightness and proper contact of the individual battery components. The cell can be used in reflection and transmission mode as indicated by the arrows 1-2 and 3-4, respectively see Fig. 3.10 (b).

Figure 3.10: Customized in situ cell for transmission (full arrows) and reflection mode (empty arrows) characterization techniques [158].

The focus of our work lies in studying the insertion and de-insertion mechanism of Li+ in the monoclinic TiO2 (B) phase, since comparatively few analyses have been published that concern its profound understanding, while in the same time a growing number of publications evoke its superior cycling properties. For comparative reason, we carried out operando measurements also on the anatase phase. Since this is a much more studied phase, it allowed us to refer our results to those obtained in previous studies.

46


3.3.1 In situ XRD Experimental details In situ XRD measurements were carried out with TiO2 (B) rich, self supported electrode films in reflection mode using the cell presented in Fig. 3.10. Self supported films containing an elevated share of PVdF as binder were subject to electrochemical cycling at slow rate C/20. XRD patterns were taken in the angle range 20-50◦ with an acquisition time of ≈ 1 h per pattern. The scan range and time per step are chosen as comprise between acqusition time and Li+ insertion rate. In situ XRD of TiO2 (B) The result of an operando measurement consisting of consecutive discharge (Li+ insertion), charge (Li+ de-insertion) and discharge is presented in Fig.3.11. The peak at 46◦ is attributed to hexagonal Be originating from Be window, which could be used to calibrate the diffraction patterns.

Figure 3.11: In situ XRD measurement upon consecutive discharge (Li+ insertion) and charge (Li+ de-insertion) of TiO2 (B) rich self supported film, showing reversible peak shift of (110),(003),(601) and (020) all attributed to the TiO2 (B) phase, while peak at 46◦ results from Be window.

The evolution of diffraction patterns shows gradual peak shifts upon lithium insertion, which proves to be reversible upon lithium de-insertion. No new peaks occur, which is agreement with the general idea of monophasic insertion mechanism. The main changes are observed for the (020) and (110) hkl planes at 48.5◦ and 24.9◦ respectively, but also the (60¯ 1) at 44.5◦ undergoes position change. The low crystallinity of the TiO2 (B) rich samples and the set up induced short acquisition


47

time led to diffraction patterns that are not suitable for a complete Rietveld refinement. For this reason, a simple peak matching procedure was used to follow the changes in the diffraction pattern by adequately adjusting the lattice constant without structural optimization, using the computer program PowderCell 1 . In doing so, we could follow the evolution of the lattice parameters as a function of amount of inserted lithium, which is shown in Fig. 3.12. By accumulating multiple diffraction patterns of the fully lithiated and non lithiated state, we obtained sufficiently good pattern to perform a full pattern matching using FullProf [160]; the obtained data points are shown as hollow markers in Fig. 3.12 while the fitted xrd pattern can be found in the Appendix Fig. B.3.

Figure 3.12: Lattice expansion of TiO2 (B) upon lithium insertion as derived from XRD peak shift using PowderCell. Hollow markers are based on full pattern matching of accumulated diffractograms before and after lithium insertion.

During the first lithium insertion, we observe a strong increase of a and b of 2.9 and 4.5 % respectively, while c can be considered as constant taking into account the experimental error. This anisotropic expansion is in agreement with findings by Armstrong et al. [161]. Applying the full pattern matching method, based on accumulated diffractograms before and after lithium insertion, we find similar trend of anisotropic expansion, whereas the expansion in a and b is somewhat reduced. Notably we obtained values of 1.73 and 3.61 %, respectively. These discrepancy to the results obtained by PowderCell are understandable, since the refinement takes into account a change of the angle β of almost 1.5 %, while the angle was kept fix in PowderCell analysis.

1

Software to visualize crystal structure and generate Bragg reflections based on crystallographic data [159]

48


In situ XRD of Anatase For comparative reason, the in situ XRD pattern for consecutive insertion and de-insertion of Li+ into anatase rich sample is shown in Fig. 3.13. Due to higher crystallinity of the anatase phase, better defined patterns were obtained, which allow one to follow the three processes of lithium insertion into anatase. The first stage is solid solution based insertion into tetragonal, which does not inflict changes of lattice parameter or formation of second phase up to x≤0.08 Li, see Fig. 3.14. How much of this lithium is consumed by parasitic reactions instead of being intercalated can not be answered with certainty at this point.

Figure 3.13: In situ XRD measurement upon consecutive discharge (Li+ insertion) and charge (Li+ de-insertion) of anatase rich sample.

This process is followed by the biphasic insertion reaction consisting of the transformation of tetragonal to orthogonal phase, which is accompanied by emerging peaks at 23.9, 39.9 and 44.4◦ corresponding to hkl planes (011), (004) and (020) of the lithiated Li0.5 TiO2 phase. Thirdly, monophasic lithium insertion occurs, which is resembled by slope-like insertion curve. This process is expressed by peak shifting of 39.9◦ (004) and 47.7◦ (200) corresponding to a lattice expansion of ≈3.7%, which is line with values reported in literature [39, 162]. Interestingly, no formation of tetragonal Li1 TiO2 is observed even though insertion exceeds 0.7 Li; similar results were obtained for lithium insertion in nanostructured anatase by Kim et al.[163] in the range of 1-3 V and insertion of up to 1.1 Li. The phenomena observed during in situ XRD measurement are in very good agreement with the proposed three domain insertion mechanism of anatase presented in Sec. 3.1. Comparing patterns from the beginning of insertion (non lithiated state) and at end of charge (after one complete cycle), we can see that initial status is recovered, indicating that the observed lattice and phase transformation are reversible during the first cycle.


49

Figure 3.14: (a)In situ XRD patterns and (b) electrochemical cycling curve upon first lithium insertion into anatase rich sample; showing three main domains of lithium insertion; I. x≤0.08 solid solution without structural changes, II. 0.08≤x≤0.48 transformation of tetragonal to orthorhombic phase and III. x≥0.45 solid solution uptake into orthorhombic structure linked to a gradual variation of the lattice parameters.

3.3.2 In situ XAS Experimental details The same in situ cell as for in situ XRD was used for in situ XAS1 measurements performed at DESY Hamburg, Germany and Elettra- Trieste, Italy. The set up for measuring the XAS spectra in transmission mode, with the in situ cell comprised between the ionization chambers aligned in the X-ray beam, is pictured in Fig. 3.15. As stated in appendix (see Sec.B.10), the XAS signal can be divided into two main portions carrying complementary information, and interpreted using two different methodologies. On the one hand, the X-ray absorption near edge structure (XANES), which deals with the narrow energy range around the absorption edge containing information mainly on electronic structure, and on the other hand, the extended X-ray absorption fine structure (EXAFS), which is concerned with the wide energy range extending up to 1000 eV beyond absorption edge. This range bears information on the local structure around the absorbing atom, as emitted electrons are scattered at next neighbours, leading to interference patterns. For the operando measurements, the samples were cycled at cycling rate C/20, and simultaneously about three X-ray absorption spectra per hour in the range of 4800-6100 eV were recorded.

1

X-ray absorption spectroscopy


50

Figure 3.15: In situ experimental setup at the XAS beamline at DESY Hamburg in transmission mode.

In situ XAS of TiO2 (B) In Fig. 3.16(a), the modification of the Ti K-edge (4966 eV ) and pre-edge upon lithium insertion, along with (b) the shift of K-edge position throughout reduction and subsequent oxidation are presented. While the changes of the pre-edge are attributed to the filling of Ti 3d orbitals and to modification of the coordination of titanium at the center of the TiO6 octahedra, the shift of the K-edge is directly linked to a change of the oxidation state of the absorbing Ti atom. Upon lithium insertion, Ti+IV is reduced to Ti+III , which is expressed by a shift of the absorption edge to lower energies. This trend is inverted upon Li+ de-insertion. It is noteworthy that this process is largely reversible although at the end of charge the starting point of pristine TiO2 is not completely recovered. This could be an indication of parasitic reactions leading to formation of stable lithium compounds or of partial Li+ trapping in irreversible sites. As mentioned above, besides the XANES1 region there is also the EXAFS2 region, which comprises valuable information on the local structure of the examined sample. The evolution of the Fourier transform (FT), EXAFS signal is presented in Fig. 3.17, schematically indicating the approximate positions of the next neighbors to the absorbing central Ti atom. These spectra are not phase corrected, therefore the Ti-O distances do not comply with the later reported bond lengths. It is important to note, that EXAFS fitting allows only a certain number of free parameters because of possible correlations between them in the case of close shells concerning similar neighbours. This makes it impossible to obtain 1 2

X-ray absorption near edge structure Extended X-ray absorption fine structure


51

Figure 3.16: a) Evolution of Ti K-edge upon Li insertion and b) reversible shift of Ti K-edge position during reduction and subsequent oxidation of TiO2 (B) rich sample.

a unique valid solution, and therefore it is necessary to group the contributions of atoms of same kind and similar distance in shells.

Figure 3.17: Fourier transform and EXAFS signals of TiO2 (B) rich sample as function of lithium inserted, dotted lines are roughly indicating position of contributing next neighbours of Ti central atom.

From this we can derive the modification of bond lengths upon lithium insertion as plotted in 3.18 for (a) Oxygen and (b) Titanium. For comparison, the EXAFS results are accompanied by chosen bond lengths, derived from in situ XRD. The latter were calculated assuming a simple lattice expansion without structural optimization, and can therefore only provide some orientation. We obtained best fitting results using two oxygen shells: one containing 5 oxygen atoms at 1.9 and a second comprising one oxygen at 2.4 Å.

52


Three fits of real space contribution for different partition of next neighbour oxygens, namely 6-0, 5-1 and 4-2 are compared to experimental results before and after lithiation in Fig. 3.19 (a) and (b), respectively. While before the insertion of lithium the 5-1 and 4-2 fitting model have fairly similar R-factor, it is evident that at the end of lithium insertion (corresponding to 0.62 Li) the 5-1 repartition is the oxygen distribution model that suits best the experimental found data.

Figure 3.18: Bond length modification upon lithium insertion into TiO2 (B) rich sample derived from FT EXAFS signal (hollow markers) and from XRD(filled markers) for (a) oxygen and (b) titanium next neighbors.

Figure 3.19: Comparison of experimental and fitted next neighbour real space contribution considering the repartition of next neighbour oxygens in 6-0, 5-1 and 4-2.

These findings are in agreement with calculations by Ben Yahia et al. [164] who claim existence of two distinct Ti environments within the TiO2 (B) structure. In detail, they report on a difference in the degree of distortion of the six-fold coordination. One of the titanium environments is stronger distorted featuring a particularly long Ti-O distance


53

of one the six surrounding next oxygen neighbours. This site can thus be considered almost as a five-fold square pyramidal coordinated environment named Ti1, which is quite different to the less distorted six fold Ti2 environment, see Fig. 3.20.

Figure 3.20: Coordination of Ti within TiO2 (B) structure revealing two different Ti environments; strongly distorted square pyramidal five-fold Ti1 and less stronly distorted octahedral six-fold coordinated Ti2. Calculated Ti-O distances are marked in red and measured ones in blue, Fig. taken from [164].

Upon lithium ion insertion smooth, gradual expansion of bond length is found for the inner Ti-O bonds shell of the Ti centers. The plot of the XRD derived bond lengths follow the same slope, which is in line with the idea that the XAS derived plot can be considered as an average sum of these individual Ti-O bond lengths. For the outer oxygen shell, the XAS and XRD derived bond length goes hand in hand showing that the bond length remains more or less constant up to about 0.4 Li. Once lithium insertion exceeds this value, a strong increase is found for the XAS derived signal, which is in discrepancy to the XRD derived bond length. We suppose therefore the existence of a two regime process: first process involving a mere expansion of the lattice, whereas the second process inflicts a significant distortion of the coordination shell. The existence of a two regime process with a threshold value at ≈0.4 lithium is further supported by the observation of two peaks close to 1.5 V forming the characteristic TiO2 (B) twin peak, see Fig. 3.1(a). For the Ti-Ti bonds, which are presented in Fig. 3.18(b), the gradual expansion found by EXAFS is largely in agreement with the results from XRD within the error bars. Only the middle shell of XAS shows slight divergence to the XRD derived ones at elevated lithium insertion values. The expansion of bond lengths upon lithium insertion is highly reversible upon lithium de-insertion (not shown), just like the alteration of the crystal lattice. The evolution of the Debye-Waller factor σ 2 , which can be seen as a measure of disorder of the system, shows a similar trend to that of the shift of K-edge position, smooth gradual increase is observed upon lithium insertion, see Fig. 3.21. This indicates that the intercalation of Li+ inflicts a reduction of order of the system, which can also be seen by decreasing intensity and definition of next neighbour contributions, see 3.17. Upon Li+ de-insertion, the initial status is largely retrieved, implying good reversibility.

54


Figure 3.21: Evolution of Debye-Waller factor for Oxygen and Titanium next neighbour shells upon lithium insertion into TiO2 (B) rich sample

As mentioned before, despite its appealing properties for LIB, few experimental studies focusing on elucidating internal mechanism and processes upon cycling of TiO2 (B) have been published. To the best of our knowledge, the only existing XAS study on lithium insertion in TiO2 is an ex situ study by Okumura et al. [99, 165]. Their results are however somewhat contradictory to ours. On the one hand, they show that for nanosized TiO2 (B) wires ( 100

3.4

4.9

200

≈80 (EC/DMC)

5.8

10.7

LiBF4

93.9

151.9

LiPF6 ’s dominance as lithium ion salt is a result of its evenly balanced, good compliance regarding the divers requirements. In the last decade, more and more efforts have been directed to replace this all-rounder, since its sensibility towards moisture and elevated temperature became known to the scientific community [224, 226]. The instability of the electrolyte during electrochemical cycling and the resulting problems will be further discussed in the following.

5.3.1 Electrolyte degradation The subject of electrolyte stability towards anodic oxidation, cathodic reduction and other parasitic reaction involving parts of the cell assembly, has accompanied the development of LIB since its very beginnings [5]. However, ever since nanostructuring of electrode materials is pursued vigorously for performance enhancement (see Sec. 1.3), this issue has gained new importance. Nanostructuring brings in its wake not only an increase of active surface sites relative to total mass, but also surface strain, which affects the free energy and might create favourable conditions for decomposition [51, 181]. In electrochemical cycling of nanostructured TiO2 , we observe two phenomena, which

5.3 Electrolyte

95

are linked to chemical transformation of the electrolyte: firstly, a significant irreversible capacity upon first discharge, and secondly, a smaller but continuous irreversible lithium consumption upon consecutive cycling, which leads to poor coulombic efficiency. These phenomena are indicated as I and II in Fig. 5.11, respectively.

Figure 5.11: (a) Galvanostatic cycling curve with corresponding derivative, which shows first two cycles in black and 20th cycle in red and (b) rate capability of TiO2 anatase from template aided sol-gel synthesis. Effects of chemical transformation of electrolyte are marked with green arrows: I irreversible capacity during first cycle, and II continuous irreversible lithium consumption upon consecutive cycling at slow rate C/20.

The irreversible capacity during first cycle is always detected regardless of the cycling rate, while the continuous irreversible lithium consumption is only visible at very slow cycling rate C/20. Most groups working on titania based anodes avoid electrochemical cycling at such slow rates, and hence this phenomena is rarely addressed. We identified undesired reactions in the lower voltage region as origin of this capacity loss. Therefore, raising the lower potential window barrier from 1 to 1.2 V was one of the first measures taken, which brought some improvements without sacrificing too much of the reversible stored capacity, see Fig. 5.12. Our findings are in agreement with experimental and theoretical studies stating reduction potential of EC, DMC and PC in the range of 0.9-1.3 V [222, 225]. However, there are also quite recent studies, which deny electrolyte decomposition at potentials ≥1V [100, 138], and propose poor conductivity and filling of irreversible sites as reasons for the incomplete lithium recovery. The use of TiO2 as active insertion material brings along further risks of undesired reactions. These reactions appear to be linked to the intrinsic nature of this material, since electrodes based on other materials, that have been prepared and cycled in the same manner, do not reveal such behaviour. On the one hand, TiO2 features well known catalytic properties, which can lead to destabilization of electrolytes, and on the other hand, the TiO2 surface carries always chemi- or physisorbed H2 O and OH groups. Upon contact with lithium, they form irreversibly strong Ti-O-Li bonds accounting for the main part of the capacity loss upon first cycle, while the liberation of proton is likely to initiate electrolyte reduction [227].

96

5 Improvement of performance: Cell assembly, Electrolyte and Additives

Figure 5.12: Electrochemical cycling curve and corresponding derivative for TiO2 (B) rich sample cycled at C/20 with (a) limited potential window 1.2-2.2 V , and (b) wide potential window 1-2.5 V , indicating parasitic reactions feature with green arrow.

Ti−O−Hsurface + Li+ −−→ Ti−O−Li + H + In a recent study Brutti et al. have shown that irreversible capacity reduces significantly, when TiO2 surface is treated prior to cell assembly [66]. This increase in coulombic efficiency gives evidence that it is indeed the surface group reactions that are responsible for the capacity loss rather than the filling of irreversible lithium sites. To further elucidate the reactions involved in the degradation of electrolyte, we studied the C1s XPS of a pristine electrode, of an electrode that had been soaked in the electrolyte without applying any current, and of electrode film that was recovered after more than 100 cycles in swagelok cell, see Fig. 5.13. All electrodes are using carbon black and PVdF as additive and binder, repectively, as well as EC:PC:3DMC 1M LiPF6 as the electrolyte. On the pristine electrode, we can identify a main peak at around 284.6 eV , which is composed of the C-H and Carbon black (CB) component [228]. Furthermore, a shoulder at 286.5 eV , which can be attributed to CO component, as well as less intense COO and carbonate at 288.5 and 291 eV , respectively, are found. Upon soaking the electrode in the electrolyte, only the reduction of intensity of the CB contribution is observed. After cycling, however, the spectra changes significantly. We observe clearly an emerging of the CO2− 3 peak, which is accompanied by a shift to higher energies. We suppose that, for the post mortem cell, the peak depicts mainly the presence of Li2 CO3 , whereas for the pristine and soaked samples other carbonate species are responsible for the peak at ≈290 eV . At the same time, we observe a vanishing of the carbon black signal, which suggests that the surface of the electrode is covered by a sufficiently thick layer (≥5 nm) of parasitic reaction products, with Li2 CO3 being one of the main components.

5.3 Electrolyte

97

Figure 5.13: Evolution of C1s XPS peak of film electrode using PVdF as binder and LiPF6 as lithium ion salt for (a) pristine (b) electrolyte soaked (c) post mortem electrode.

5.3.2 Adjusting solvent composition Additive - Vinylene carbonate The use of VC originates from its elevated reduction potential ≈1.4 V [225], which ensures that a significantly thick protective SEI layer has been formed around the carbon, before any PC cointercalation can occur. This asset of directing and promoting SEI formation could be a useful tool to stabilize and reduce the continuous lithium consumption, once a protective layer has been formed during the first cycle(s). In electrochemical tests, coulombic efficiency improved, as can be seen from cycling curve, but at relatively high cost regarding capacity retention, see Fig. 5.14. In detail, we observe a much increased capacity loss upon first cycle, which expresses the growth of SEI. Within the following 10 cycles coulombic efficiency drastically increases to values above 97% and remains stable, independent of the cycling rate. Unfortunately, the consolidation of coulombic efficiency goes hand in hand with a steady loss in overall capacity. We assume that the SEI layer blocks the access to the TiO2 particles for lithium ions, and hence a

98


growing share of the electrode material becomes inactive. This is also reflected by the rate capability decrease, since at high cycling rate 5C only 67 vs. 98 mAh/g is achieved with and without VC, respectively, see Fig. 5.14(b).

Figure 5.14: (a) Electrochemical cycling curve at C/20 and (b) capacities obtained for progressively increased cycling rate for TiO2 (B) rich sample with electrolyte containing 2%VC additive.

We can conclude that the overall effect of VC is only partially beneficial for the cycling performance of TiO2 based electrode materials. It is however imaginable that, by lowering VC concentration, a thickness optimization of SEI layer could be obtained that would prevent continuous electrolyte reduction, while at the same time allowing fast charge transfer. Role of Propylene carbonate The electrolyte component of propylene carbonate (PC) features a high decomposition potential [225], and moreover recent publications show that PC is dispensable without direct negative influences for electrochemical cycling [173, 229, 230]. The series of cycling of experiments with PC free electrolyte shows indeed no deterioration regarding key features such as rate capability and capacity retention. However, also no improvement regarding the parasitic reactions and high irreversible lithium consumption during cycling at slow rate C/20 are found. We conclude therefore that PC is not the major cause of this undesired event.

5.3.3 Role of the lithium ion salt As stated in the introduction of this section, LiPF6 is one of the most widely used electrolyte salts due to is high solubility and high mobility. However its instability and the resulting cascade of degradation reactions are of growing concern. It is well known that even at room temperature an equilibrium exists LiPF6(s) −−→ PF5(g) + LiF

5.3 Electrolyte

99

whereby the production of the gaseous PF5 drives this reaction continuously to the right. This reaction is displaced to the right side by the increase in temperature and the presence of moisture [223]. An interesting alternative to LiPF6 is LiBF4 . It presents a higher ion mobility, but due to its reduced dissociation rate, the overall conductivity is inferior to that of LiPF6 [231]. In Fig. 5.15, the comparison of rate capability for two representative samples, (a) cycled using LiBF4 and (b) using LiPF6 , is presented. While the first cycle shows a similar gap between charge and discharge capacity for both samples, the coulombic efficiency shows opposing trends for the following cycles. For the sample cycled with LiBF4 , the coulombic efficiency shows slow but continuous improvement, whereas for the sample cycled with LiPF6 coulombic efficiency strongly decreases after having reached a maximum after 4 cycles. This is mainly due to a rapid growth of discharge capacity, while charge capacity remains almost constant. A look at the evolution of the derivative curve reveals not only the formation of a shoulder on the left of the cathodic peak at ≈1.4 V , but also a rise of parasitic events at lower voltage. Neither of these features are found in the derivative of LiBF4 sample. This is in line with the higher coulombic efficiency that is obtained. It must, however, be stated that samples cycled with LiBF4 seem to have slightly lower rate capability.

Figure 5.15: Rate capability for progressively increased cycling rates and corresponding derivative of representative TiO2 (B) rich samples cycled with (a) LiBF4 and (b) LiPF6 .

Electrochemical cycling carried out at 60 ◦ C on TiO2 (B) rich samples using EC:PC:3DMC

100


1M LiPF6 showed pronounced parasitic reactions, which are reflected by an elevated irreversible lithium consumption compared to ambient temperature (25 ◦ C). Post mortem visible inspection revealed strong colouring of electrolyte, as well as disintegration of whatman glassfiber separator, see Fig. 5.16(a). Replacing glassfiber by polymer based celgard separator brought only some improvement of the coulombic efficiency. As Fig. 5.16(b) reveals, the celgard separator remains intact under the severe conditions, but a strong colouring of electrolyte solution is still visible. Only by exchanging the electrolyte solute LiPF6 for LiBF4 an improvement in capacity retention and coulombic efficiency could be achieved. We conclude therefore that increased temperature promotes the degradation of LiPF6 , leading to the formation of LiF. In the presence of moisture on the TiO2 surface (see Sec. 5.3.1), this may lead to the formation of HF, with detrimental consequences for the electrode and cell components.

Figure 5.16: Post mortem inspection of electrodes cycled with EC:PC:3DMC 1M LiPF6 using (a) whatman glassfiber and (b) celgard polymer separator.

In order to deepen the understanding on the role of the solute in the degradation reactions upon electrochemical cycling, XPS characterization on ambient temperature cycled post mortem electrodes was carried out. In Fig. 5.17, the C1s (a, b) and the F1s (c, d) of the two different electrodes are compared. For the C1s, we see, besides a narrowing of the main peak composed of CB and C-H, a much reduced contribution from CO3 at 290 eV for the LiBF4 sample. The F1s XPS spectra of LiPF6 and LiBF4 cycled sample are best fitted using three components, a major one at 685.6 eV attributed to ionic LiF [220], and two weaker ones at 688.5 and 687.0 eV attributed to C-F in PVdF and in LiPF6 or LiBF4 , respectively [66, 221, 232]. For the sample cycled using LiBF4 , the peak at 688.5 eV is much less intense while the intensity of the middle peak is increased. Leroy et al. showed in an extensive XPS study on surface film formation on graphite electrodes that, in the potential

5.4 Charge transport limitations in electrodes

101

window 3-4.2 V , LiBF4 produces less LiF than LiPF6 , due to the higher stability against hydrolysis. Unfortunately we can not draw such a conclusion from our spectra, since both reveal a significant LiF peak [232].

Figure 5.17: XPS spectra of post mortem electrodes cycled using LiPF6 (a,c) and LiBF4 (b,d).

Summarizing, we can infer that dissociation of LiPF6 and LiBF4 occurs at values well above 1.2 V , even though the reaction process and products are different from each other. We found considerably more carbonate species for the LiPF6 cycled electrode than for the one cycled with LiBF4 , while significant LiF deposition was found in both cases. We have seen that the electrolyte degradation can have detrimental consequences for both electrochemical cycling stability and coulombic efficiency, due to formation of passivative layer as well as to electrode deterioration. Therefore, we attribute the enhanced coulombic efficiency to the higher stability of LiBF4 solute against hydrolysis compared to LiPF6 .

5.4 Charge transport limitations in electrodes In his work on cathode materials Gaberšček introduced a simple method to gain insight information on transport limitation in electrodes [233]. He suggested that the transport process can be subdivided into three processes; A transport to the particle (external transport) B entering of charge carriers into active particles

102


C diffusion within the insertion material particle (internal transport) which are illustrated in Fig. 5.18(a). The curve obtained by plotting the resistance of the electrode, which is derived by Ohm’s law using polarization and applied current over the active mass loading, allows conclusion about the transport limiting process. In Fig. 5.18(b), the slopes obtained for internal transport limited (F), external transport limited (S), and intermediate case (M) are plotted. It is noteworthy that this theory does not allow one to make a distinction between electronic and ionic charge carriers, only the sum of transport resistance can be exploited. This is insofar problematic, as electronic and ionic conductivity are likely to be independent and diverging in magnitude.

Figure 5.18: (a) Schematic division into three transport process in electrodes and (b) dependency of electrode resistance on loading, the three curves illustrate the cases; F internal transport limited, S external transport limitation and M intermediate case transport.

In Fig. 5.19, the calculated resistance for film electrodes based on TiO2 (B) rich sample containing 8 % carbon nanotubes (CNT) and 12 % PVdF as additive and cycled at rate C/20 using EC:PC:3DMC 1M LiBF4 are presented. At low and medium active mass loading (≤ 2.6 mg/cm2 ), the data points follow a steep exponential declining slope resembling the F-curve. According to Gaberšček’s model, this corresponds to internal transport limitation. However, at higher loading we observe increasing resistance, which suggest a linear slope similar to the S curve. Such behaviour indicates external charge transfer limitation. Nevertheless, the low number of data points at elevated loading puts this assumption on a weak basis. Concluding, we can state that the proposed internal resistance limitation at low and medium loadings is in line with the previously discussed low charge transfer properties of TiO2 , see Sec. 1.3. If the increased resistance at loadings ≥3 mg/cm2 were confirmed, this would mean that, for the electrode production based on this active insertion material with industrial relevant loadings one should focus on improving intra-electrode wiring, rather than on enhancing intrinsic TiO2 charge transport capability. In the literature, one dimensional carbon is the material of choice for providing and effective conductive network via coating, composite formation or physical mixing [59, 214, 234, 235].

5.4 Charge transport limitations in electrodes

103

Figure 5.19: Electrode resistance as function of active mass loading of TiO2 (B) rich electrodes using carbon nanotubes (CNT) as additives cycled at C/20 with EC:PC:3DMC LiBF4 .

6 Conclusion and Perspectives 6.1 Conclusion The principal objectives of this thesis were (i) establishing of a soft chemistry method for low cost TiO2 production with facile upscale possibility, and (ii) evaluate the possibility and study the feasibility of nanostructuring, formation of porous scaffolds as well as doping, to enhance the electrochemical cycling performances of TiO2 . We have demonstrated that hydrothermal synthesis method offers the possibility for synthesizing highly pure nanostructured TiO2 . Throughout a comprehensive study, we were able to single out the influence of alkalinity (molarity of NaOH mother solution) and solid-to-liquid ratio (ratio of TiO2 powder in NaOH solution) on the physico-chemical properties of the final product. While alkalinity governs mostly the phase composition favouring the formation of either anatase or TiO2(b), the solid-to-liquid ratio influences mainly the final morphology. By adequately adjusting these synthesis parameters, we have been able to synthesize samples within a wide range of nanostructures from nanosheets, to nanoparticles to nanoribbons, and to tune the phase composition from anatase to TiO2 (B) rich phases. Our galvanostatic electrochemical measurements agree with the literature, showing that the insertion mechanism of Li+ in anatase differs significantly from that of TiO2 (B). While the insertion of lithium in anatase occurs through a complex multistep insertion process, involving solid-solution, biphasic phase transformation and monophasic insertion, the insertion of lithium in TiO2 (B) implies the existence of a single monophasic insertion process. Furthermore, the average insertion potential of TiO2 (B) (characteristic twin peak at ≈1.5 V ) is lower than that of anatase, which has its main insertion plateau at 1.8 V . All hydrothermally synthesized TiO2 samples, regardless of their phase composition showed reversible cycling capacities above the typical capacity of LTO (165 mAh/g [218]) at slow cycling rates C/20 (17 mA/g) and C/5 (67 mA/g). As cycling rate increases, however, the margin between capacity values of TiO2 and LTO are reduced. At elevated cycling rate 5C (1680 mAh/g), only TiO2 (B) rich sample shows comparable high rate capability to LTO reported values [236]. The significantly higher rate capability of TiO2 (B) over anatase rich samples is a striking phenomena, which is also reflected by an elevated polarization of anatase compared to that of TiO2 (B). This performance discrepancy is in line with results from cyclic voltammetry, which shows that, while anatase has a purely faradaic bulk intercalation mechanism, governed by slow solid-state diffusion, TiO2 (B) lithium insertion has a significant share of fast surface pseudocapacitive storage. Regardless of the fact that the anatase rich phase reveals three times higher specific surface area. The observation of such different electrochemical behaviour for two so similar TiO2 phases, led

105

106

6 Conclusion and Perspectives

us to undertake further investigation toward elucidating internal processes and mechanism of the lithium insertion process. Particularly, the insertion process of TiO2 (B) is of great interest since few experimental studies are available for this polymorph. Operando XRD studies on TiO2 (B) are in line with the proclaimed monophasic insertion process, revealing no formation of new phases upon lithium insertion. An anisotropic lattice expansion in þa and þb direction is found, which corresponds to a total volumetric expansion of ≤5%. For anatase, our results fitted well with the three step mechanism that was deducted from electrochemical cycling curve. Firstly, no structural changes occur up to 0.08 Li corresponding to the solid solution domain. Upon further lithium insertion, the tetragonal anatase is transformed in to orthogonal phase, which is subsequent expanding under further lithium insertion. The formation of tetragonal LiTiO2 was not observed. The reduction of Ti+IV to Ti+III , going along with the insertion of lithium into the TiO2 lattice, can be nicely followed by XANES. For TiO2 (B), we found a largely reversible linear shift of edge position indicating that TiO2 (B) is lithiated gradually. Furthermore, XAS allows the investigation of local structure behaviour under operando conditions via the exploitation of EXAFS signal. We obtained best fitting results for the Ti coordination using an inner and outer oxygen shell. This implies that the TiO6 octahedron is largely distorted, so that the existence of two differently coordinated Ti environments, notably octahedral and square pyramidal, best describe the TiO2 (B) structure. This is strong contrast to the Anatase phase, which shows a narrow distribution of the Ti-O bond lengths indicating the prevailing of an symmetrical octahedral coordination. Upon lithium insertion into TiO2 (B), the distance to the outer oxygen is increased once lithium insertion exceeds 0.45 Li, which suggests the presence of a two regime mechanism. This could be coupled to the characteristic twin peak of TiO2 (B) in galvanostatic cycling curve at ≈1.5 V . We conclude that lithium insertion in TiO2 (B) is a purely monophasic process, which inflicts relatively little change of global structure of the host lattice. For the local structure, however, a reduction of coordinational order is indicated by the decreasing pre-edge peaks as well as by the broadening of oxygen distance distribution upon lithium insertion. For the anatase phase, the important structural changes during the biphasic transformation are nicely reflected by EXAFS results. In this insertion domain from 0.08 to 0.45 Li, a significant increase of outer oxygen shell bond length and Debye-Waller factor, reflecting the distortion of the octahedral coordination, are recorded. For lithium insertion exceeding 0.45 Li, this trend does not continue, and we observe stabilization of the values, which implies that a different insertion mechanism is prevailing. First calculation results by Mouna Ben Yahia are in agreement with the literature, stating the presence of three distinct lithium insertion sites in the TiO2 (B) structure. Furthermore, it was shown that these sites feature different energetic stability. Preliminary results are somewhat contradictory but they indicate that the C site in the channel along þb is not necessarily the preferred site. However, further work is needed to elucidate this question. We showed that by using sol-gel synthesis coupled with ultra sound, TiO2 anatase with mesoporous structure can be obtained without any surfactant or template. However,

6.2 Prospects

107

electrochemical performance enhancement was only achieved when using a template, which led to much increased specific pore volume and increased size of mesopores. We conclude therefore that adequately adjusting the size, distribution and connectivity of the pores is vital for the electrochemical performance enhancement. Especially preventing agglomeration of particles by providing a stable hierarchical open scaffold structure is a key point in achieving high rate capability. Both synthesis methods used in this thesis (hydrothermal and sol-gel), offer the possibility of facile one-pot doping by adding an adequate precursor to the mother solution. We found that introduction of aliovalent cation of donor (Nb+V ) or acceptor (Fe+III ) type leads to substitution of Ti, and does not inflict any significant global or local structural changes of the anatase and TiO2 (B) lattice, respectively. On the example of Nb-doped anatase, we were able to show that the electronic structure is distinctly modified from insulator towards metallic-type conduction, which results in increased rate capability for doped samples compared to non-doped samples of similar morphology. Another major part of my PhD thesis was concerned with electrolyte stability. The observation of strong parasitic reactions, leading to large irreversible capacities, came as a surprise to us, since from the literature no SEI formation or electrolyte decomposition were to be expected above 1 V . Since these phenomena are little studied but are of crucial importance for proper functioning of titania based electrode materials, intensive efforts were undertaken to elucidate its origins and develop countermeasures. Post mortem investigation reveals electrolyte colouring as well as notable amounts of carbonate and fluoride species deposition on the electrode film, which suggest the decomposition of electrolyte at elevated potentials. In the derivative of electrochemical cycling curves, we could attribute specific features growing with the number of cycles at potentials lower than 1.4 V to these irreversible parasitic reactions. We have carried out various test to elucidate the source of this issue, and have circled out the autodecomposition and the reaction with traces of surface water of the lithium ion salt LiPF6 as one of the main causes. The parasitic reactions could be largely decreased by replacing LiPF6 with more stable and less reactive LiBF4 . The role of PVdF binder in these parasitic reactions is less salient, but we were able to show that coulombic efficiency was improved when replaced by CMC. The coulombic efficiency improvements that the use of VC as electrolyte additive brings, are strongly outweighed by the loss in capacity retention and rate capability. Summarizing, we can state that, regardless of the insights gained on decomposition reactions and the mitigation that was achieved throughout different countermeasures, many questions and problems concerning electrolyte stability for titania based electrode materials remain unsolved.

6.2 Prospects Due to the limitation of time, many tasks, projects, and ideas remain incomplete or unrealized within my PhD thesis. Thus this section is dedicated to the possible continuation

108

6 Conclusion and Perspectives

of this work. We demonstrated that IR1 spectroscopy allows the identification of both active insertion materials and electrolyte simultaneously. However, low signal to noise ratio and superposition of vibrational modes of TiO2 make the attribution challenging. We anticipate the use of other spectroscopic methods under operando conditions such as UV-vis and Raman, in order to monitor material and electrolyte behaviour upon electrochemical cycling simultaneously. A suitable operando cell using a quartz window has been recently developed by Julien Hannauer, and is currently being tested regarding its feasibility. Another interesting project would be the XAS operando investigation of the pseudocapacitive storage mechanism in TiO2 (B). As cycling rate increases, the storage becomes more reliant on this fast surface diffusion controlled process. Modern quick XAS allows acquisition of XAS spectra of several hundreds of eV within seconds. Hence, fast mechanism and processes can be monitored, which in the case of TiO2 (B) could give further insights on the pseudocapacitive storage mechanism. From what we know today about insertion mechanism of TiO2 (B), we would expect to see decreasing of EXAFS and XANES changes during lithium insertion as cycling rate is increased due to decreasing contribution of slow solid-diffusion controlled faradaic lithium storage. We have shown that adjusting sol-gel synthesis parameters and implying templates is a worthwhile approach to enhance cycling performance of TiO2 , but there is strong need to optimize the pore size distribution, the pore shape and the specific pore volume. Hence, future material tailoring should aim at a stable, interconnected, homogeneous pore network, which allows fast charge transport, electrolyte penetration and prevents agglomeration to improve the electrochemical cycling properties. It is needless to say that a trade-off between enhancement of rate capability and a decrease in coulombic efficiency due to electrolyte decomposition is to be expected for such high surface area nanoporous materials. Furthermore, we like to point out that every significant change in the nanostructure of the active insertion material demands for laborious adjustment of electrode formulation. While doping with different dopants and concentration was successful and valuable approach to overcome the intrinsic low electronic conductivity, the attempts to form a homogeneous coating were not satisfying. In order to achieve a thin and evenly distributed film, the use of sophisticated surface treatment, aiming at functionalizing TiO2 surface groups such as Ti-OH should be proceeded. This method is especially interesting, since an optimized coating could not only serve as conductive network, but also as protective layer against electrode deterioration. Coupling of experimental data with theoretical methods bears great potential for studying complex systems such as lithium insertion mechanism, as it allows one to reduce the complexity of the real system to an ascertainable model. This model can then be used to elucidate position, accessibility and energetics of insertion sites, as well as to study

1

Infra-red

6.2 Prospects

109

the evolution of bond lengths, coordination and lattice parameter upon lithium insertion. However, the calculation processes are time consuming, which lead to the fact that in this thesis only first results of calculations performed by Mouna Ben Yahia could be presented. Additional calculations aiming at monitoring the evolution of electronic structure upon lithium insertion, by modelling the Ti K-edge of TiO2 (B) at different stages of lithiation, carried out by Frédéric Lemoigno, were still in progress when this manuscript went to press.

Bibliography [1] International Energy Agency IEA. www.iea.org [2] Crossette, Barbara: United Nations Population Found. www.unfpa.org. Version: 2011 [3] Orr, Lynn: Changing the world’s energy system / Standford University. – Forschungsbericht [4] TOLLEFSON, Jeff: Charging up the future. In: Nature 456 (2008), Nr. 7221, 436–440. http://cat.inist.fr/?aModele=afficheN&cpsidt=20860939. – ISSN 0028–0836 [5] Scrosati, Bruno ; Garche, Jürgen: Lithium batteries: Status, prospects and future. In: J. Power Sources 195 (2010), Nr. 9, 430–2419. http://dx.doi.org/10.1016/j.jpowsour. 2009.11.048 [6] Mizushima, K ; Jones, P C. ; Wiseman, P J. ; Goodenough, J B.: LixCoO2 (0â¤xâ¤1) : A new cathode material for batteries of high energy density. In: Solid State Ionics 3-4 (1981), August, Nr. 0, 171–174. http://dx.doi.org/http://dx.doi.org/10.1016/ 0167-2738(81)90077-1. – DOI http://dx.doi.org/10.1016/0167–2738(81)90077–1. – ISSN 0167–2738 [7] Whittingham, M S.: The Role of Ternary Phases in Cathode Reactions. In: Journal of The Electrochemical Society 123 (1976), Januar, Nr. 3, 315–320. http://dx.doi.org/10. 1149/1.2132817. – DOI 10.1149/1.2132817 [8] Thackeray, M M. ; David, W I F. ; Bruce, P G. ; Goodenough, J B.: Lithium insertion into manganese spinels. In: Materials Research Bulletin 18 (1983), April, Nr. 4, 461–472. http://dx.doi.org/http://dx.doi.org/10.1016/0025-5408(83)90138-1. – DOI http://dx.doi.org/10.1016/0025–5408(83)90138–1. – ISSN 0025–5408 [9] Yoshino, Akira: The Birth of the Lithium-Ion Battery. In: Angewandte Chemie 51 (2012), Februar, 5798–5800. http://dx.doi.org/10.1002/anie.201105006. – DOI 10.1002/anie.201105006. – ISSN 1521–3773 [10] Landi, Brian J. ; Ganter, Matthew J. ; Cress, Cory D. ; DiLeo, Roberta A. ; Raffaelle, Ryne P.: Carbon nanotubes for lithium ion batteries. In: Energy Env. Sci. 2 (2009), 638–654. http://dx.doi.org/10.1039/B904116H. – DOI 10.1039/B904116H [11] Buchmann, Isidor: Version: 2013

www.batteryuniversity.com.

www.batteryuniversity.com.

[12] Nishi, Yoshio: Lithium ion secondary batteries; past 10 years and the future. In: J. Power Sources 100 (2001), Nr. 1-2, 101–106. http://dx.doi.org/10.1016/S0378-7753(01) 00887-4

111

112

Bibliography

[13] Mukherjee, Rahul ; Krishnan, Rahul ; Lu, Toh-Ming ; Koratkar, Nikhil: Nanostructured electrodes for high-power lithium ion batteries. In: Nano Energy 1 (2012), Juli, Nr. 4, 518–533. http://dx.doi.org/10.1016/j.nanoen.2012.04.001. – DOI 10.1016/j.nanoen.2012.04.001. – ISSN 22112855 [14] Gartner, John (.: The EV Conundrum: Uncertain Resale Value Complicates Li-ion Battery Market. www.smartgridnews.com. Version: 2009 [15] Sankey, Paul (Deutsche B. ; Clark, David T. (Deutsche B. ; Micheloto, Silvio (Deutsche B.: The End of the Oil Age. 2011 and beyond: A reality check / Deutsche Bank Securities. Version: 2010. http://green.autoblog.com/2011/ 01/06/deutsche-bank-li-ion-battery-cost-forecast-per-kwh/ (December 2010). – Forschungsbericht. – Elektronische Ressource [16] Goodwin, David (ABSL Power S.: Li-ion battery technology / ABSL Power Solutions. Version: 2010. http://www.abslpower.com (July). – Forschungsbericht. – Elektronische Ressource. – 1–16 S [17] Bruce, Peter G. ; Scrosati, Bruno ; Tarascon, Jean-Marie: Nanomaterials for rechargeable lithium batteries. In: Angewandte Chemie (International ed. in English) 47 (2008), Januar, Nr. 16, 2930–46. http://dx.doi.org/10.1002/anie.200702505. – DOI 10.1002/anie.200702505. – ISBN 1334463808 [18] Palacin, M R.: Recent advances in rechargeable battery materials: a chemists perspective. In: Chem. Soc. Rev. 28 (2009), 2565–2575. http://dx.doi.org/10.1039/B820555H. – DOI 10.1039/B820555H [19] Cabana, Jordi ; Monconduit, Laure ; Larcher, Dominique ; Palacín, M R.: Beyond intercalation-based Li-ion batteries: the state of the art and challenges of electrode materials reacting through conversion reactions. In: Advanced materials (Deerfield Beach, Fla.) 22 (2010), September, Nr. 35, E170–92. http://dx.doi.org/10.1002/adma.201000717. – DOI 10.1002/adma.201000717. – ISSN 1521–4095 [20] McNaught, A.D. ; Wilkinson, A.: IUPAC Compendium of Chemical Terminology 2nd ed. (the "gold book"). 2. Blackwell Scientific Publications. http://dx.doi.org/10.1351/ goldbook. http://dx.doi.org/10.1351/goldbook. – ISBN 0–9678550–9–8 [21] Wu, Hui ; Cui, Yi: Designing nanostructured Si anodes for high energy lithium ion batteries. In: Nano Today 7 (2012), 414–429. http://dx.doi.org/10.1016/j.nantod.2012.08.004 [22] Lee, Seung W. ; Yabuuchi, Naoaki ; Gallant, Betar M. ; Chen, Shuo ; Kim, Byeong-Su ; Hammond, Paula T. ; Shao-Horn, Yang: High-power lithium batteries from functionalized carbon-nanotube electrodes. In: Nature nanotechnology 5 (2010), Juli, Nr. 7, 531–7. http: //dx.doi.org/10.1038/nnano.2010.116. – DOI 10.1038/nnano.2010.116. – ISSN 1748– 3395 [23] Kim, C. ; Yang, K. S. ; Kojima, M. ; Yoshida, K. ; Kim, Y. J. ; Kim, Y. a. ; Endo, M.: Fabrication of Electrospinning-Derived Carbon Nanofiber Webs for the Anode Material of Lithium-Ion Secondary Batteries. In: Advanced Functional Materials 16 (2006), Dezember, Nr. 18, 2393–2397. http://dx.doi.org/10.1002/adfm.200500911. – DOI 10.1002/adfm.200500911. – ISSN 1616301X

Bibliography

113

[24] Yoo, EunJoo ; Kim, Jedeok ; Hosono, Eiji ; Zhou, Hao-shen ; Kudo, Tetsuichi ; Honma, Itaru: Large reversible Li storage of graphene nanosheet families for use in rechargeable lithium ion batteries. In: Nano letters 8 (2008), August, Nr. 8, 2277–82. http://dx.doi. org/10.1021/nl800957b. – DOI 10.1021/nl800957b. – ISSN 1530–6984 [25] Wu, Zhong-Shuai ; Zhou, Guangmin ; Yin, Li-Chang ; Ren, Wencai ; Li, Feng ; Cheng, Hui-Ming: Graphene/metal oxide composite electrode materials for energy storage. In: Nano Energy 1 (2012), Januar, Nr. 1, 107–131. http://dx.doi.org/10.1016/j.nanoen.2011. 11.001. – DOI 10.1016/j.nanoen.2011.11.001. – ISSN 22112855 [26] Su, Fang-Yuan ; He, Yan-Bing ; Li, Baohua ; Chen, Xue-Cheng ; You, Cong-Hui ; Wei, Wei ; Lv, Wei ; Yang, Quan-Hong ; Kang, Feiyu: Could graphene construct an effective conducting network in a high-power lithium ion battery? In: Nano Energy 1 (2012), Mai, Nr. 3, 429–439. http://dx.doi.org/10.1016/j.nanoen.2012.02.004. – DOI 10.1016/j.nanoen.2012.02.004. – ISSN 22112855 [27] Dolle, MickaeÌl ; Sannier, Lucas ; Beaudoin, Bernard ; Trentin, Michel ; Tarascon, Jean-Marie: Live Scanning Electron Microscope Observations of Dendritic Growth in Lithium/Polymer Cells. In: Electrochemical and Solid-State Letters 5 (2002), Nr. 12, A286. http://dx.doi.org/10.1149/1.1519970. – DOI 10.1149/1.1519970. – ISSN 10990062 [28] Reddy, M V. ; Subba Rao, G V. ; Chowdari, B V R.: Metal oxides and oxysalts as anode materials for li ion batteries. In: Chemical reviews 113 (2013), Juli, Nr. 7, 5364–457. http://dx.doi.org/10.1021/cr3001884. – DOI 10.1021/cr3001884. – ISSN 1520–6890 [29] Liu, Chang ; Li, Feng ; Ma, Lai-Peng ; Cheng, Hui-Ming: Advanced Materials for Energy Storage. In: Adv. Mater. 22 (2010), Nr. 8, E28–E62. http://onlinelibrary.wiley.com/ doi/10.1002/adma.200903328/abstract [30] Choi, Min G. ; Lee, Young-Gi ; Song, Seung-Wan ; Kim, Kwang M.: Lithium-ion battery anode properties of TiO2 nanotubes prepared by the hydrothermal synthesis of mixed (anatase and rutile) particles. In: Electrochimica Acta 55 (2010), August, Nr. 20, 5975–5983. http: //dx.doi.org/10.1016/j.electacta.2010.05.052. – DOI 10.1016/j.electacta.2010.05.052. – ISSN 00134686 [31] Kavan, Ladislav: Electrochemistry of titanium dioxide: some aspects and highlights. In: Chemical record (New York, N.Y.) 12 (2012), Februar, Nr. 1, 131–42. http://dx.doi.org/ 10.1002/tcr.201100012. – DOI 10.1002/tcr.201100012. – ISSN 1528–0691 [32] Mancini, M. ; Nobili, F. ; Tossici, R. ; Marassi, R.: Study of the electrochemical behavior at low temperatures of green anodes for Lithium ion batteries prepared with anatase TiO2 and water soluble sodium carboxymethyl cellulose binder. In: Electrochimica Acta 85 (2012), Dezember, 566–571. http://dx.doi.org/10.1016/j.electacta.2012.08.115. – DOI 10.1016/j.electacta.2012.08.115. – ISBN 0737402296 [33] Ohzuku, Tsutomu ; Ueda, Atsushi ; Yamamoto, Norihiro.: Zero-strain insertion material of Li[Li1/3Ti5/3]O4 for rechargeable lithium cells. In: Journal of the Electrochemical Society 142 (1995), Nr. 5, S. 1431–1435. http://dx.doi.org/10.1149/1.2048592. – DOI 10.1149/1.2048592. – ISSN 0013–4651

114

Bibliography

[34] Ariyoshi, Kingo ; Yamato, Ryoji ; Ohzuku, Tsutomu: Zero-strain insertion mechanism of Li[Li1/3Ti5/3]O4 for advanced lithium-ion (shuttlecock) batteries. In: Electrochimica Acta 51 (2005), November, Nr. 6, 1125–1129. http://dx.doi.org/10.1016/j.electacta.2005. 05.053. – DOI 10.1016/j.electacta.2005.05.053. – ISSN 00134686 [35] Yi, Ting-Feng ; Jiang, Li-Juan ; Shu, J ; Yue, Cai-Bo ; Zhu, Rong-Sun ; Qiao, Hong-Bin: Recent development and application of Li4Ti5O12 as anode material of lithium ion battery. In: J. Phys. Chem. Solids 71 (2010), Nr. 9, S. 1236–1242 [36] altairnano. http://www.altairnano.com/ [37] http://www.scib.jp/en/. http://www.scib.jp/en/ [38] Zachau-Christiansen, B ; West, K ; Jacobsen, T: Lithium insertion in different TiO2 modifications. In: Solid State Ionics 30 (1988), 1176–1182. http://www.sciencedirect. com/science/article/pii/0167273888903529 [39] Yang, Zhenguo ; Choia, Daiwon ; Kerisit, Sebastien ; Rosso, Kevin M. ; Wang, Donghai ; Zhang, Jason ; Graff, Gordon ; Liu, Jun: Nanostructures and lithium electrochemical reactivity of lithium titanites and titanium oxides: A review. In: J. Power Sources 192 (2009), Nr. 2, 588–598. http://dx.doi.org/10.1016/j.jpowsour.2009.02.038 [40] Fröschl, T ; Hörmann, U ; Kubiak, P ; Kučerová, G ; Pfanzelt, M ; Weiss, C K. ; Behm, R J. ; Hüsing, N ; Kaiser, U ; Landfester, K ; Wohlfahrt-Mehrens, M: High surface area crystalline titanium dioxide: potential and limits in electrochemical energy storage and catalysis. In: Chemical Society reviews 41 (2012), August, Nr. 15, 5313–60. http://dx.doi.org/10.1039/c2cs35013k. – DOI 10.1039/c2cs35013k. – ISSN 1460–4744 [41] Aravindan, V. ; Shubha, N. ; Cheah, Yan L. ; Prasanth, R. ; Chuiling, W. ; Prabhakar, Rajiv R. ; Madhavi, S.: Extraordinary long-term cycleability of TiO2-B nanorods as anodes in full-cell assembly with electrospun PVdF-HFP membranes. In: Journal of Materials Chemistry A 1 (2013), Nr. 2, 308. http://dx.doi.org/10.1039/c2ta00078d. – DOI 10.1039/c2ta00078d. – ISSN 2050–7488 [42] Zukalova, Marketa ; Kalbac, Martin ; Kavan, Ladislav ; Exnar, Ivan ; Graetzel, Michael: Pseudocapacitive Lithium storage in tio2(b). In: Chem. Mater. 17 (2005), 1248–1255. http://pubs.acs.org/doi/abs/10.1021/cm048249t [43] Armstrong, A. R. ; Armstrong, Graham ; Canales, Jesús ; Bruce, Peter G.: TiO(2)-B nanowires. In: Angewandte Chemie (International ed. in English) 43 (2004), April, Nr. 17, 2286–8. http://dx.doi.org/10.1002/anie.200353571. – DOI 10.1002/anie.200353571. – ISSN 1433–7851 [44] Armstrong, A. R. ; Armstrong, Graham ; Canales, Jesus ; Bruce, Peter G.: TiO2(B) nanowires as negative electrodes for rechargeable lithium batteries. In: J. Power Sources 146 (2005), Nr. 1-2, 501–506. http://dx.doi.org/10.1016/j.jpowsour.2005.03.057 [45] Liu, Hansan ; Bi, Zhonghe ; Sun, Xiao-Guang ; Unocic, Raymond R. ; Paranthaman, M P. ; Dai, Sheng ; Brown, Gilbert M.: Mesoporous TiO(2) -B Microspheres with Superior Rate Performance for Lithium Ion Batteries. In: Advanced materials (Deerfield Beach, Fla.)

Bibliography

115

23 (2011), August, Nr. 30, 3450–3454. http://dx.doi.org/10.1002/adma.201100599. – DOI 10.1002/adma.201100599. – ISSN 1521–4095 [46] Koudriachova, M V.: Role of the surface in Li insertion into nanowires of TiO2(B). In: Surf. Interface Anal. 42 (2010), Nr. December 2009, S. 1330–1332. http://dx.doi.org/10. 1002/sia.3335. – DOI 10.1002/sia.3335 [47] Kim, Bo-Ra ; Yun, Kang-Seop ; Jung, Hee-June ; Myung, Seung-Taek ; Jung, Sang-Chul ; Kang, Wooseung ; Kim, Sun-Jae: Effect of anatase phase on electrochemical properties of the TiO2(B) negative electrode for lithium-ion battery application. In: Current Applied Physics (2013), Januar, 1–4. http://dx.doi.org/10.1016/j.cap.2013.01.012. – DOI 10.1016/j.cap.2013.01.012. – ISSN 15671739 [48] Fehse, M. ; Fischer, F. ; Tessier, C. ; Stievano, L. ; Monconduit, L.: Tailoring of phase composition and morphology of TiO2-based electrode materials for lithium-ion batteries. In: Journal of Power Sources 231 (2013), Juni, 23–28. http://dx.doi.org/10. 1016/j.jpowsour.2012.12.058. – DOI 10.1016/j.jpowsour.2012.12.058. – ISSN 03787753 [49] Morgan, Benjamin J. ; Madden, Paul a.: Lithium intercalation into TiO2(B): A comparison of LDA, GGA, and GGA+U density functional calculations. In: Physical Review B 86 (2012), Juli, Nr. 3, 035147. http://dx.doi.org/10.1103/PhysRevB.86.035147. – DOI 10.1103/PhysRevB.86.035147. – ISSN 1098–0121 [50] Wilkening, M. ; Heine, J. ; Lyness, C. ; Armstrong, A. ; Bruce, P.: Li diffusion properties of mixed conducting TiO2-B nanowires. In: Physical Review B 80 (2009), August, Nr. 6, 064302. http://dx.doi.org/10.1103/PhysRevB.80.064302. – DOI 10.1103/PhysRevB.80.064302. – ISSN 1098–0121 [51] Dylla, Anthony G. ; Henkelman, Graeme ; Stevenson, Keith J.: Lithium Insertion in Nanostructured TiO(2)(B) Architectures. In: Acc. Chem. Res. 46 (2013), Februar, Nr. 5, 1104–1112. http://dx.doi.org/10.1021/ar300176y. – DOI 10.1021/ar300176y. – ISSN 1520–4898 [52] Babu, Veluru J. ; Nair, Appukuttan S. ; Peining, Zhu ; Ramakrishna, Seeram: Synthesis and characterization of rice grains like, Nitrogen-doped TiO2 nanostructures by electrospinning-photocatalysis. In: Materials Letters 65 (2011), Oktober, Nr. 19-20, 3064–3068. http://dx.doi.org/10.1016/j.matlet.2011.06.035. – DOI 10.1016/j.matlet.2011.06.035. – ISSN 0167577X [53] Shin, Ji-Yong ; Samuelis, Dominik ; Maier, Joachim: Sustained Lithium-Storage Performance of Hierarchical, Nanoporous Anatase TiO2 at High Rates: Emphasis on Interfacial Storage Phenomena. In: Advanced Functional Materials 21 (2011), September, Nr. 18, 3464– 3472. http://dx.doi.org/10.1002/adfm.201002527. – DOI 10.1002/adfm.201002527. – ISSN 1616301X [54] Wagemaker, Marnix ; Borghols, Wouter J H. ; Mulder, Fokko M.: Large impact of particle size on insertion reactions. A case for anatase Li(x)TiO2. In: Journal of the American Chemical Society 129 (2007), April, Nr. 14, 4323–7. http://dx.doi.org/10. 1021/ja067733p. – DOI 10.1021/ja067733p. – ISSN 0002–7863

116

Bibliography

[55] Shin, Ji-yong ; Joo, Jong H. ; Samuelis, Dominik ; Maier, Joachim: Oxygen-Deficient TiO2âδ Nanoparticles via Hydrogen Reduction for High Rate Capability Lithium Batteries. In: Chem. Mater. 24 (2012), 543–551. http://pubs.acs.org/doi/abs/10.1021/cm2031009 [56] Wang, J. ; Polleux, J. ; Lim, J. ; Dunn, B.: Pseudocapacitive Contributions to Electrochemical Energy Storage in TiO2 (Anatase) Nanoparticles. In: Journal of Physical Chemistry C 111 (2007), Oktober, Nr. 40, 14925–14931. http://dx.doi.org/10.1021/jp074464w. – DOI 10.1021/jp074464w. – ISSN 1932–7447 [57] Chen, Jun S. ; Tan, Yi L. ; Li, Chang M. ; Cheah, Yan L. ; Luan, Deyan ; Madhavi, Srinivasan ; Boey, Freddy Yin C. ; Archer, Lynden A. ; Lou, Xiong W.: Constructing Hierarchical Spheres from Large Ultrathin Anatase TiO2 Nanosheets with Nearly 100% Exposed (001) Facets for Fast Reversible Lithium Storage. In: J. Am. Chem. Soc. 132 (2010), Mai, Nr. 17, 6124–6130. http://dx.doi.org/10.1021/ja100102y. – DOI 10.1021/ja100102y. – ISSN 1520–5126 [58] Saravanan, Kuppan ; Ananthanarayanan, Krishnamoorthy ; Balaya, Palani: Mesoporous TiO2 with high packing density for superior lithium storage. In: Energy & Environmental Science 3 (2010), Nr. 7, 939. http://dx.doi.org/10.1039/c003630g. – DOI 10.1039/c003630g. – ISSN 1754–5692 [59] Das, Shyamal K. ; Bhattacharyya, Aninda J.: Influence of Mesoporosity and Carbon Electronic Wiring on Electrochemical Performance of Anatase Titania. In: Journal of The Electrochemical Society 158 (2011), Juni, Nr. 6, A705. http://dx.doi.org/10.1149/1. 3574028. – DOI 10.1149/1.3574028. – ISSN 00134651 [60] Chen, Jizhang ; Yang, Li ; Tang, Yufeng: Electrochemical lithium storage of TiO2 hollow microspheres assembled by nanotubes. In: J. Power Sources 195 (2010), Nr. 19, 6893–6896. http://dx.doi.org/10.1016/j.jpowsour.2010.04.005 [61] Zhang, Feng ; Zhang, Yu ; Song, Shuyan ; Zhang, Hongjie: Superior electrode performance of mesoporous hollow TiO2 microspheres through efficient hierarchical nanostructures. In: Journal of Power Sources 196 (2011), Oktober, Nr. 20, 8618–8624. http://dx.doi.org/10. 1016/j.jpowsour.2011.06.006. – DOI 10.1016/j.jpowsour.2011.06.006. – ISSN 03787753 [62] Reddy, M. V. ; Valerie Teoh, X. W. ; Nguyen, T. B. ; Michelle Lim, Y. Y. ; Chowdari, B. V. R.: Effect of 0.5 M NaNO3: 0.5 M KNO3 and 0.88 M LiNO3:0.12 M LiCl Molten Salts, and Heat Treatment on Electrochemical Properties of TiO2. In: Journal of The Electrochemical Society 159 (2012), Januar, Nr. 6, A762. http://dx.doi.org/10.1149/2. 077206jes. – DOI 10.1149/2.077206jes. – ISSN 00134651 [63] Beuvier, Thomas ; Richard-Plouet, Mireille ; Mancini-Le Granvalet, Maryline ; Brousse, Thierry ; Crosnier, Olivier ; Brohan, Luc: TiO2(B) nanoribbons as negative electrode material for lithium ion batteries with high rate performance. In: Inorganic chemistry 49 (2010), September, Nr. 18, 8457–64. http://dx.doi.org/10.1021/ic1010192. – DOI 10.1021/ic1010192. – ISSN 1520–510X [64] Wessel, Claas ; Zhao, Liang ; Urban, Sven ; Ostermann, Rainer ; Djerdj, Igor ; Smarsly, Bernd M. ; Chen, Liquan ; Hu, Yong-Sheng ; Sallard, Sébastien: Ionic-liquid synthesis route of TiO(2) (B) nanoparticles for functionalized materials. In: Chemistry

Bibliography

117

(Weinheim an der Bergstrasse, Germany) 17 (2011), Januar, Nr. 3, 775–9. http://dx.doi. org/10.1002/chem.201002791. – DOI 10.1002/chem.201002791. – ISSN 1521–3765 [65] Saito, Morihiro ; Murota, Yosuke ; Takagi, Mikihiro ; Tajima, Makoto ; Asao, Takayuki ; Inoue, Hidemi ; Tasaka, Akimasa ; Inaba, Minoru: Improvement of the Reversible Capacity of TiO_2(B) High Potential Negative Electrode. In: Journal of the Electrochemical Society 159 (2012), S. A49. http://dx.doi.org/10.1149/2.051201jes. – DOI 10.1149/2.051201jes. – ISSN 00134651 [66] Brutti, Sergio ; Gentili, Valentina ; Menard, Hervé ; Scrosati, Bruno ; Bruce, Peter G.: TiO2-(B) Nanotubes as Anodes for Lithium Batteries: Origin and Mitigation of Irreversible Capacity. In: Advanced Energy Materials 2 (2012), März, Nr. 3, 322–327. http://dx.doi.org/10.1002/aenm.201100492. – DOI 10.1002/aenm.201100492. – ISSN 16146832 [67] Ren, Yu ; Hardwick, Laurence J. ; Bruce, Peter G.: Lithium Intercalation into Mesoporous Anatase with an Ordered 3D Pore Structure. In: Angew. Chem. Int. Ed. 49 (2010), Nr. 14, S. 2570–2574 [68] Liu, Shaohua ; Jia, Haiping ; Han, Lu ; Wang, Jiulin ; Gao, Pengfei ; Xu, Dongdong ; Yang, Jun ; Che, Shunai: Nanosheet-Constructed Porous TiO(2) -B for Advanced Lithium Ion Batteries. In: Advanced materials (Deerfield Beach, Fla.) 24 (2012), Mai, 3201–3204. http://dx.doi.org/10.1002/adma.201201036. – DOI 10.1002/adma.201201036. – ISSN 1521–4095 [69] Chakrapani, Vidhya ; Tvrdy, Kevin ; Kamat, Prashant V.: Modulation of electron injection in CdSe-TiO(2) system through medium alkalinity. In: Journal of the American Chemical Society 132 (2010), März, Nr. 4, 1228–9. http://dx.doi.org/10.1021/ja909663r. – DOI 10.1021/ja909663r. – ISSN 1520–5126 [70] Naldoni, Alberto ; Allieta, Mattia ; Santangelo, Saveria ; Marelli, Marcello ; Fabbri, Filippo ; Cappelli, Serena ; Bianchi, Claudia L. ; Psaro, Rinaldo ; Dal Santo, Vladimiro: Effect of nature and location of defects on bandgap narrowing in black TiO2 nanoparticles. In: Journal of the American Chemical Society 134 (2012), Mai, Nr. 18, 7600–3. http://dx.doi.org/10.1021/ja3012676. – DOI 10.1021/ja3012676. – ISSN 1520–5126 [71] Chen, Xiaobo ; Liu, Lei ; Yu, Peter Y. ; Mao, Samuel S.: Increasing Solar Absorption for Photocatalysis with Black Hydrogenated Titanium Dioxide Nanocrystals. In: Science 331 (2011), Februar, Nr. 746, 746–749. http://dx.doi.org/10.1126/science.1200448. – DOI 10.1126/science.1200448. – ISSN 1095–9203 [72] Tao, Junguang ; Luttrell, Tim ; Batzill, Matthias: A two-dimensional phase of TiOâ with a reduced bandgap. In: Nature chemistry 3 (2011), Mai, Nr. 4, 296–300. http: //dx.doi.org/10.1038/nchem.1006. – DOI 10.1038/nchem.1006. – ISSN 1755–4349 [73] Sheppard, L. R. ; Dittrich, T. ; Nowotny, M. K.: The Impact of Niobium Surface Segregation on Charge Separation in Niobium-Doped Titanium Dioxide. In: The Journal of Physical Chemistry C 116 (2012), Oktober, Nr. 39, 20923–20929. http://dx.doi.org/10. 1021/jp3065147. – DOI 10.1021/jp3065147. – ISSN 1932–7447

118

Bibliography

[74] Baumard, J. F. ; Tani, E.: Electrical conductivity and charge compensation in Nb doped TiO2 rutile. In: The Journal of Chemical Physics 67 (1977), Nr. 3, 857. http: //dx.doi.org/10.1063/1.434952. – DOI 10.1063/1.434952. – ISSN 00219606 [75] Yang, Min ; Jha, Himendra ; Liu, Ning ; Schmuki, Patrik: Increased photocurrent response in Nb-doped TiO2 nanotubes. In: Journal of Materials Chemistry 21 (2011), Nr. 39, 15205. http://dx.doi.org/10.1039/c1jm12961a. – DOI 10.1039/c1jm12961a. – ISSN 0959–9428 [76] Sheppard, L R. ; Bak, T ; Nowotny, J: Electrical properties of niobium-doped titanium dioxide. 1. Defect disorder. In: The Journal of Physical Chemistry B 110 (2006), November, Nr. 45, 22447–22454. http://dx.doi.org/10.1021/jp0637025. – DOI 10.1021/jp0637025. – ISSN 1520–6106 [77] Furubayashi, Yutaka ; Hitosugi, Taro ; Yamamoto, Yukio ; Inaba, Kazuhisa ; Kinoda, Go ; Hirose, Yasushi ; Shimada, Toshihiro ; Hasegawa, Tetsuya: A transparent metal: Nb-doped anatase TiO2. In: Applied Physics Letters 86 (2005), Nr. 25, 252101–3. http: //dx.doi.org/10.1063/1.1949728. – DOI 10.1063/1.1949728. – ISSN 00036951 [78] Ali, Zahid ; Cha, Seung N. ; Sohn, Jung I. ; Shakir, Imran ; Yan, Changzeng ; Kim, Jong M. ; Kang, Dae J.: Design and evaluation of novel Zn doped mesoporous TiO2 based anode material for advanced lithium ion batteries. In: Journal of Materials Chemistry 22 (2012), Nr. 34, 17625. http://dx.doi.org/10.1039/c2jm33315e. – DOI 10.1039/c2jm33315e. – ISSN 0959–9428 [79] Wang, Yude ; Smarsly, Bernd M. ; Djerdj, Igor: Niobium Doped TiO2 with Mesoporosity and Its Application for Lithium Insertion. In: Chemistry of Materials 22 (2010), Dezember, Nr. 24, 6624–6631. http://dx.doi.org/10.1021/cm1020977. – DOI 10.1021/cm1020977. – ISSN 0897–4756 [80] Ryu, Myung-hyun ; Jung, Kyu-nam ; Shin, Kyung-hee ; Han, Kyoo-seung ; Yoon, Sukeun: High Performance N-Doped Mesoporous Carbon Decorated TiO2 Nanofibers as Anode Materials for Lithium-Ion Batteries. In: Journal of Physical Chemistry C 117 (2013), 8092–8098. http://pubs.acs.org/doi/full/10.1021/jp400757s [81] Hirano, Masanori ; Dozono, Hayato ; Kono, Takeshi: Hydrothermal synthesis and properties of solid solutions and composite nanoparticles in the TiO2 - SnO2 system. In: Materials Research Bulletin 46 (2011), September, Nr. 9, 1384–1390. http://dx.doi.org/ 10.1016/j.materresbull.2011.05.016. – DOI 10.1016/j.materresbull.2011.05.016. – ISSN 00255408 [82] Hirano, Masanori ; Ichihashi, Yoshiko: Phase transformation and precipitation behavior of niobium component out of niobium-doped anatase-type TiO2 nanoparticles synthesized via hydrothermal crystallization. In: Journal of Materials Science 44 (2009), September, Nr. 22, 6135–6143. http://dx.doi.org/10.1007/s10853-009-3848-2. – DOI 10.1007/s10853– 009–3848–2. – ISSN 0022–2461 [83] Yang, Zunxian ; Du, Guodong ; Guo, Zaiping ; Yu, Xuebin ; Chen, Zhixin ; Guo, Tailiang ; Liu, Huakun ; Guoa, Tailiang: TiO2(B)@carbon composite nanowires as anode for lithium ion batteries with enhanced reversible capacity and cyclic performance. In: J. Mater.

Bibliography

119

Chem. 21 (2011), Nr. 24, 8591–8596. http://dx.doi.org/10.1039/c0jm03873c. – DOI 10.1039/c0jm03873c. – ISSN 0959–9428 [84] Choi, Min G. ; Lee, Young-Gi ; Song, Seung-Wan ; Kim, Kwang M.: Anode properties of titanium oxide nanotube and graphite composites for lithium-ion batteries. In: J. Power Sources 195 (2010), Nr. 24, 8289–8296. http://dx.doi.org/10.1016/j.jpowsour.2010. 06.104 [85] Cheng, Yu H. ; Huang, Yizhong ; Kanhere, Pushkar D. ; Subramaniam, Vishnu P. ; Gong, Dangguo ; Zhang, Sam ; Highfield, James ; Schreyer, Martin K. ; Chen, Zhong: Dual-Phase Titanate/Anatase with Nitrogen Doping for Enhanced Degradation of Organic Dye under Visible Light. In: Chem. Eur. J. 17 (2011), 2575–2578. http: //onlinelibrary.wiley.com/doi/10.1002/chem.201002256/full [86] Kasuga, Tomoko ; Hiramatsu, Masayoshi ; Hoson, Akihiko ; Sekino, Toru: Formation of titanium oxide nanotube. In: Langmuir 7463 (1998), Nr. 5, 3160–3163. http://pubs.acs. org/doi/abs/10.1021/la9713816 [87] Kasuga, Tomoko ; Hiramatsu, Masayoshi ; Hoson, Akihiko ; Sekino, Toru ; Niihara, Koichi: Titania Nanotubes Prepared by Chemical Processing. In: Adv. Mater. 11 (1999), Nr. 15, 1307–1311. http://onlinelibrary.wiley.com/doi/10.1002/(SICI) 1521-4095(199910)11:15%3C1307::AID-ADMA1307%3E3.0.CO;2-H/abstract [88] Hernández-Alonso, M D. ; García-Rodríguez, S ; Sánchez, B ; Coronado, J M.: Revisiting the hydrothermal synthesis of titanate nanotubes: new insights on the key factors affecting the morphology. In: Nanoscale 3 (2011), Mai, Nr. 5, 2233–40. http: //dx.doi.org/10.1039/c1nr10081e. – DOI 10.1039/c1nr10081e. – ISSN 2040–3372 [89] Xu, Yuanmei ; Fang, Xiaoming ; Xiong, Jian ; Zhang, Zhengguo: Hydrothermal transformation of titanate nanotubes into single-crystalline TiO2 nanomaterials with controlled phase composition and morphology. In: Mater. Res. Bull. 45 (2010), Nr. 7, 799–804. http://dx.doi.org/10.1016/j.materresbull.2010.03.016 [90] Yin, Hengbo ; Wada, Yuji ; Kitamura, Takayuki ; Kambe, Shingo ; Murasawa, Sadao ; Mori, Hirotaro ; Sakata, Takao ; Yanagida, Shozo: Hydrothermal synthesis of nanosized anatase and rutile TiO2 using amorphous phase TiO2. In: Journal of Materials Chemistry 11 (2001), Nr. 6, 1694–1703. http://dx.doi.org/10.1039/b008974p. – DOI 10.1039/b008974p. – ISSN 09599428 [91] Hernández-Alonso, María D. ; García-Rodríguez, Sergio ; Suárez, Silvia ; Portela, Raquel ; Sánchez, Benigno ; Coronado, Juan M.: Highly selective one-dimensional TiO2based nanostructures for air treatment applications. In: Applied Catalysis B: Environmental 110 (2011), November, 251–259. http://dx.doi.org/10.1016/j.apcatb.2011.09.009. – DOI 10.1016/j.apcatb.2011.09.009. – ISSN 09263373 [92] Nguyen, KC ; Cha, HG: The Axis-Oriented, Anatase TiO2 Single Crystals with Dominant {001} and {100} Facets. In: Crystal Growth & Design 11 (2011), 3947–3953. http://pubs. acs.org/doi/abs/10.1021/cg200815t

120

Bibliography

[93] Lim, Ying Wen L. ; Tang, Yuxin ; Cheng, Yu H. ; Chen, Zhong: Morphology, crystal structure and adsorption performance of hydrothermally synthesized titania and titanate nanostructures. In: Nanoscale 2 (2010), 2751–2757. http://pubs.rsc.org/en/content/ articlelanding/2010/nr/c0nr00440e [94] Yoshida, Ryuhei ; Suzuki, Yoshikazu ; Yoshikawa, Susumu: Syntheses of TiO2(B) nanowires and TiO2 anatase nanowires by hydrothermal and post-heat treatments. In: J. Solid State Chem. 178 (2005), 2179–2185. http://dx.doi.org/10.1016/j.jssc.2005.04.025 [95] Bavykin, Dmitry V. ; Parmon, Valentin N. ; Lapkin, A ; Walsh, Frank C.: The effect of hydrothermal conditions on the mesoporous structure of TiO 2 nanotubes. In: J. Mater. Chem. 14 (2004), 3370–3377. http://pubs.rsc.org/en/content/articlelanding/2004/ jm/b406378c/unauth [96] Testino, Andrea ; Bellobono, Ignazio R. ; Buscaglia, Vincenzo ; Canevali, Carmen ; D’Arienzo, Massimiliano ; Polizzi, Stefano ; Scotti, Roberto ; Morazzoni, Franca: Optimizing the photocatalytic properties of hydrothermal TiO2 by the control of phase composition and particle morphology. a systematic approach. In: Journal of the American Chemical Society 129 (2007), März, Nr. 12, 3564–75. http://dx.doi.org/10.1021/ja067050+. – DOI 10.1021/ja067050+. – ISSN 0002–7863 [97] Zhou, Weijia ; Du, Guojun ; Hu, Peiguang ; Li, Guohong ; Wang, Dongzhou ; Liu, Hong ; Wang, Jiyang ; Boughton, Robert I. ; Liua, Duo ; Jiang, Huaidong ; Liu, Duo: Nanoheterostructures on TiO2 nanobelts achieved by acid hydrothermal method with enhanced photocatalytic and gas sensitive performance. In: J. Mater. Chem. 21 (2011), Nr. 22, 7937–7945. http://dx.doi.org/10.1039/c1jm10588d. – DOI 10.1039/c1jm10588d. – ISSN 0959–9428 [98] Sing, K. S. W.: Reporting Physisoprtion Data for Gas/Solid Systems with Special Reference to the Determination of Surface Area and Porosity. In: Pure Appl. Chem. 54 (1982), Nr. 11, 2201–2218. http://195.37.231.82/publications/pac/pdf/1982/pdf/5411x2201.pdf [99] Okumura, Toyoki ; Fukutsuka, Tomokazu ; Yanagihara, Asuki ; Orikasa, Yuki ; Arai, Hajime ; Ogumi, Zempachi ; Uchimoto, Yoshiharu: Electronic and local structural changes with lithium-ion insertion in TiO2-B: X-ray absorption spectroscopy study. In: Journal of Materials Chemistry 21 (2011), Nr. 39, 15369–15377. http://dx.doi.org/10. 1039/c1jm11335f. – DOI 10.1039/c1jm11335f. – ISSN 0959–9428 [100] Xu, Jinwei ; Jia, Caihong ; Cao, Bin ; Zhang, W F.: Electrochemical properties of anatase TiO2 nanotubes as an anode material for lithium-ion batteries. In: Electrochim. Acta 52 (2007), 8044–8047. http://dx.doi.org/10.1016/j.electacta.2007.06.077 [101] Zhu, Guan-Nan ; Wang, Cong-Xiao ; Xia, Yong-Yao: Structural transformation of layered hydrogen trititanate (H2Ti3O7) to TiO2(B) and its electrochemical profile for lithium-ion intercalation. In: Journal of Power Sources 196 (2011), März, Nr. 5, 2848–2853. http: //dx.doi.org/10.1016/j.jpowsour.2010.07.023. – DOI 10.1016/j.jpowsour.2010.07.023. – ISSN 03787753 [102] Casarin, Maurizio ; Vittadini, Andrea ; Selloni, Annabella: First principles study of hydrated/hydroxylated TiO2 nanolayers: from isolated sheets to stacks and tubes. In: ACS

Bibliography

121

nano 3 (2009), Februar, Nr. 2, 317–24. http://dx.doi.org/10.1021/nn800608n. – DOI 10.1021/nn800608n. – ISSN 1936–086X [103] Feist, Thomas P. ; Davies, Peter K.: The soft chemical synthesis of TiO2(b) from layered titanates. In: Journal of Solid State Chemistry 101 (1992), 275–295. http://dx.doi.org/ 10.1016/0022-4596(92)90184-W [104] Inorganic Crystal Structure Database (ICSD), FIZ Karlsruhe [105] Lazzeri, Michele ; Vittadini, Andrea ; Selloni, Annabella: Structure and energetics of stoichiometric TiO2 anatase surfaces. In: Phys. Rev. B 63 (2001), März, Nr. 15, 155409. http://dx.doi.org/10.1103/PhysRevB.63.155409. – DOI 10.1103/PhysRevB.63.155409 [106] Vittadini, Andrea ; Casarin, Maurizio ; Selloni, Annabella: Structure and Stability of TiO2-B Surfaces: A Density Functional Study. In: J. Phys. Chem. C 113 (2009), Nr. 44, 18973–18977. http://pubs.acs.org/doi/abs/10.1021/jp9073009 [107] Macwan, D P. ; Dave, Pragnesh N. ; Chaturvedi, Shalini: A review on nano-TiO2 sol-gel type syntheses and its applications. In: J. Mater. Sci. 46 (2011), 3669–3686. http: //link.springer.com/article/10.1007/s10853-011-5378-y [108] Zhao, Jianquan ; Wan, Peng ; Xiang, Jing ; Tong, Tong ; Dong, Lei ; Gao, Zhinong ; Shen, Xinyu ; Tong, Hua: Synthesis of highly ordered macro-mesoporous anatase TiO2 film with high photocatalytic activity. In: Microporous and Mesoporous Materials 138 (2011), Februar, Nr. 1-3, 200–206. http://dx.doi.org/10.1016/j.micromeso.2010.09.002. – DOI 10.1016/j.micromeso.2010.09.002. – ISSN 13871811 [109] Kubiak, P ; Geserick, J ; Hüsing, N ; Wohlfahrt-Mehrens, M: Electrochemical performance of mesoporous TiO2 anatase. In: J. Power Sources 175 (2008), Nr. 1, 510–516. http://dx.doi.org/10.1016/j.jpowsour.2007.09.044 [110] Das, Shyamal K. ; Gnanavel, M. ; Patel, Manu U. M. ; Shivakumara, Chikkadasappa ; Bhattacharyya, Aninda J.: Anamolously High Lithium Storage in Mesoporous Nanoparticulate Aggregation of Fe3+ Doped Anatase Titania. In: Journal of The Electrochemical Society 158 (2011), Nr. 12, A1290. http://dx.doi.org/10.1149/2.029112jes. – DOI 10.1149/2.029112jes. – ISSN 00134651 [111] Collins, Andrew ; Carriazo, Daniel ; Davis, Sean a. ; Mann, Stephen: Spontaneous template-free assembly of ordered macroporous titania. In: Chemical communications (Cambridge, England) 44 (2004), März, Nr. 5, 568–569. http://dx.doi.org/10.1039/ b315018f. – DOI 10.1039/b315018f. – ISSN 1359–7345 [112] Jung, Hun-Gi ; Oh, Sung W. ; Jin, Ce ; Jayaprakash, N ; Sun, Yang-Kook: Mesoporous TiO2 nano networks: Anode for high power lithium battery applications. In: Electrochem. Comm. 11 (2009), 756–759. http://dx.doi.org/10.1016/j.elecom.2009.01.030 [113] Guo, Yu-Guo ; Hu, Yong-Sheng ; Sigle, W ; Maier, Joachim: Superior Electrode Performance of Nanostructured Mesoporous TiO2 (Anatase) through Efficient Hierarchical Mixed Conducting Networksâ . In: Adv. Mater. 19 (2007), Nr. 16, 2087–2091. http: //onlinelibrary.wiley.com/doi/10.1002/adma.200602828/full

122

Bibliography

[114] Wang, Yuping ; Ren, Jiankun ; Liu, Guoqing ; Peng, Panying: Synthesis and characterization of iodine ion doped mesoporous TiO2 by solâgel method. In: Materials Chemistry and Physics 130 (2011), Oktober, Nr. 1-2, 493–499. http://dx.doi.org/10.1016/j. matchemphys.2011.07.020. – DOI 10.1016/j.matchemphys.2011.07.020. – ISSN 02540584 [115] Ruiz, Ana M. ; Dezanneau, G ; Arbiol, J ; Cornet, A ; Morante, Joan R.: Insights into the Structural and Chemical Modifications of Nb Additive on TiO2 Nanoparticles. In: Chem. Mater. 16 (2004), 862–871. http://pubs.acs.org/doi/abs/10.1021/cm0351238 [116] Kang, Soon H. ; Kim, Hyun S. ; Kim, Jae-Yup ; Sung, Yung-Eun: Enhanced photocurrent of nitrogen-doped TiO2 film for dye-sensitized solar cells. In: Mater. Chem. Phys. 124 (2010), Nr. 1, 422–426. http://dx.doi.org/10.1016/j.matchemphys.2010.06.059 [117] Hu, Shaozheng ; Li, Fayun ; Fan, Zhiping ; Chang, Chein-Chi: Enhanced photocatalytic activity and stability of nano-scaled TiO2 co-doped with N and Fe. In: Applied Surface Science 258 (2011), Oktober, Nr. 1, 182–188. http://dx.doi.org/10.1016/j.apsusc.2011.08.028. – DOI 10.1016/j.apsusc.2011.08.028. – ISSN 01694332 [118] Liu, Jinming ; Zhao, Xiaoru ; Duan, Libing ; Cao, Mengmeng ; Sun, Huinan ; Shao, Jifeng ; Chen, Shuai ; Xie, Haiyan ; Chang, Xiao ; Chen, Changle: Influence of annealing process on conductive properties of Nb-doped TiO2 polycrystalline films prepared by sol-gel method. In: Applied Surface Science 257 (2011), September, Nr. 23, 10156–10160. http: //dx.doi.org/10.1016/j.apsusc.2011.07.009. – DOI 10.1016/j.apsusc.2011.07.009. – ISSN 01694332 [119] Cavaliere, Sara ; Subianto, Surya ; Savych, Iuliia ; Jones, Deborah J. ; Rozière, Jacques: Electrospinning: designed architectures for energy conversion and storage devices. In: Energy & Environmental Science 4 (2011), 4761–4785. http://dx.doi.org/10.1039/c1ee02201f. – DOI 10.1039/c1ee02201f. – ISSN 1754–5692 [120] Wang, Moo C. ; Lin, Huey J. ; Yang, Tien S.: Characteristics and optical properties of iron ion (Fe3+)-doped titanium oxide thin films prepared by a solâgel spin coating. In: Journal of Alloys and Compounds 473 (2009), April, Nr. 1-2, 394–400. http://dx.doi.org/10.1016/ j.jallcom.2008.05.105. – DOI 10.1016/j.jallcom.2008.05.105. – ISSN 09258388 [121] Cavaliere, Sara ; Subianto, Surya ; Chevallier, Laure ; Jones, Deborah J. ; Rozière, Jacques: Single step elaboration of size-tuned Pt loaded titania nanofibres. In: Chemical Communications 47 (2011), Juni, Nr. 24, 6834–6836. http://dx.doi.org/10.1039/c1cc11716e. – DOI 10.1039/c1cc11716e. – ISSN 1364–548X [122] Bauer, Alex ; Chevallier, Laure ; Hui, Rob ; Cavaliere, Sara ; Zhang, Jiujun ; Jones, Deborah ; Rozière, Jacques: Synthesis and characterization of Nb-TiO2 mesoporous microsphere and nanofiber supported Pt catalysts for high temperature PEM fuel cells. In: Electrochimica Acta 77 (2012), August, 1–7. http://dx.doi.org/10.1016/j.electacta. 2012.04.028. – DOI 10.1016/j.electacta.2012.04.028. – ISSN 00134686 [123] Luo, Wei ; Hu, Xianluo ; Sun, Yongming ; Huang, Yunhui: Surface modification of electrospun TiO2 nanofibers via layer-by-layer self-assembly for high-performance lithium-ion batteries. In: Journal of Materials Chemistry 22 (2012), Nr. 11, 4910. http://dx.doi.org/ 10.1039/c2jm15197a. – DOI 10.1039/c2jm15197a. – ISSN 0959–9428

Bibliography

123

[124] Zhu, Peining ; Wu, Yongzhi ; Reddy, M. V. ; Sreekumaran Nair, A. ; Chowdari, B. V. R. ; Ramakrishna, S.: Long term cycling studies of electrospun TiO2 nanostructures and their composites with MWCNTs for rechargeable Li-ion batteries. In: RSC Advances 2 (2012), Nr. 2, 531–537. http://dx.doi.org/10.1039/c1ra00514f. – DOI 10.1039/c1ra00514f. – ISSN 2046–2069 [125] C. Yu, Jimmy ; Zhang, Lizhi ; Yu, Jiaguo: Rapid synthesis of mesoporous TiO2 with high photocatalytic activity by ultrasound-induced agglomeration. In: New Journal of Chemistry 26 (2002), April, Nr. 4, 416–420. http://dx.doi.org/10.1039/b109173e. – DOI 10.1039/b109173e. – ISSN 11440546 [126] Booth, Colin ; Attwood, David: Effects of block architecture and composition on the association properties of poly(oxyalkylene) copolymers in aqueous solution. In: Macromolecular Rapid Communications 21 (2000), Juni, Nr. 9, 501–527. http://dx.doi.org/10. 1002/1521-3927(20000601)21:93.0.CO;2-R. – DOI 10.1002/1521– 3927(20000601)21:93.0.CO;2–R. – ISSN 1022–1336 [127] Hatakeyama, Masaomi ; Faller, Roland: Coarse-grained simulations of ABA amphiphilic triblock copolymer solutions in thin films. In: Physical chemistry chemical physics : PCCP 9 (2007), September, Nr. 33, 4662–72. http://dx.doi.org/10.1039/b702241g. – DOI 10.1039/b702241g. – ISSN 1463–9076 [128] Wanka, G. ; Hoffmann, H. ; Ulbricht, W.: Phase Diagrams and Aggregation Behavior of Poly(oxyethylene)-Poly(oxypropylene)-Poly(oxyethylene) Triblock Copolymers in Aqueous Solutions. In: Macromolecules 27 (1994), Juli, Nr. 15, 4145–4159. http://dx.doi.org/10. 1021/ma00093a016. – DOI 10.1021/ma00093a016. – ISSN 0024–9297 [129] Vu, Anh ; Qian, Yuqiang ; Stein, Andreas: Porous Electrode Materials for LithiumIon Batteries - How to Prepare Them and What Makes Them Special. In: Advanced Energy Materials 2 (2012), September, Nr. 9, 1056–1085. http://dx.doi.org/10.1002/ aenm.201200320. – DOI 10.1002/aenm.201200320. – ISSN 16146832 [130] Xia, Y. ; Yang, P. ; Sun, Y. ; Wu, Y. ; Mayers, B. ; Gates, B. ; Yin, Y. ; Kim, F. ; Yan, H.: One-Dimensional Nanostructures: Synthesis, Characterization, and Applications. In: Advanced Materials 15 (2003), März, Nr. 5, 353–389. http://dx.doi.org/10.1002/adma. 200390087. – DOI 10.1002/adma.200390087. – ISSN 09359648 [131] Barth, Sven ; Hernandez-Ramirez, Francisco ; Holmes, Justin D. ; Romano-Rodriguez, Albert: Synthesis and applications of one-dimensional semiconductors. In: Progress in Materials Science 55 (2010), August, Nr. 6, 563–627. http://dx.doi.org/10.1016/j. pmatsci.2010.02.001. – DOI 10.1016/j.pmatsci.2010.02.001. – ISSN 00796425 [132] Reneker, DH ; Yarin, AL ; Zussman, E ; Xu, H: Electrospinning of nanofibers from polymer solutions and melts. In: Advances in Applied Mechanics 41 (2007), 44–197. http: //dx.doi.org/10.1016/S0065-2156(06)41002-4. – DOI 10.1016/S0065–2156(06)41002–4 [133] Sawicka, Katarzyna M. ; Gouma, Perena: Electrospun composite nanofibers for functional applications. In: Journal of Nanoparticle Research 8 (2006), Juli, Nr. 6, 769–781. http: //dx.doi.org/10.1007/s11051-005-9026-9. – DOI 10.1007/s11051–005–9026–9. – ISSN 1388–0764

124

Bibliography

[134] Dai, Yunqian ; Liu, Wenying ; Formo, Eric ; Sun, Yueming ; Xia, Younan: Ceramic nanofibers fabricated by electrospinning and their applications in catalysis, environmental science, and energy technology. In: Polymers for Advanced Technologies 22 (2011), März, Nr. 3, 326–338. http://dx.doi.org/10.1002/pat.1839. – DOI 10.1002/pat.1839. – ISSN 10427147 [135] Li, Dan ; Xia, Younan: Fabrication of Titania Nanofibers by Electrospinning. In: Nano Lett. 3 (2003), Nr. 4, 555–560. http://pubs.acs.org/doi/abs/10.1021/nl034039o [136] Nam, Sang H. ; Shim, Hee-Sang ; Kim, Youn-Su ; Dar, Mushtaq A. ; Kim, Jong G. ; Kim, Won B.: Ag or Au Nanoparticle-Embedded One-Dimensional Composite TiO2 Nanofibers Prepared via Electrospinning for Use in Lithium-Ion Batteries. In: ACS Appl. Mater. Interfaces 2 (2010), Nr. 7, 2046–2052. http://pubs.acs.org/doi/abs/10.1021/am100319u [137] Ricci, PC ; Carbonaro, CM ; Stagi, Luigi: Anatase-to-Rutile Phase Transition in TiO2 Nanoparticles Irradiated by Visible Light. In: The Journal of Physical Chemistry C 117 (2013), 7850–7857. http://pubs.acs.org/doi/abs/10.1021/jp312325h [138] Armstrong, A R. ; Armstrong, G ; Canales, J ; García, R ; Bruce, Peter G.: Lithium-Ion Intercalation into TiO2-B Nanowires. In: Adv. Mater. 17 (2005), Nr. 7, 862–865. http://onlinelibrary.wiley.com/doi/10.1002/adma.200400795/full [139] Ren, Yu ; Liu, Zheng ; Pourpoint, Frédérique ; Armstrong, A. R. ; Grey, Clare P. ; Bruce, Peter G.: Nanoparticulate TiO2(B): An Anode for Lithium-Ion Batteries. In: Angewandte Chemie 124 (2012), Februar, Nr. 9, 2206–2209. http://dx.doi.org/10.1002/ ange.201108300. – DOI 10.1002/ange.201108300. – ISSN 00448249 [140] Lafont, U ; Carta, D ; Mountjoy, G ; Chadwick, A V. ; Kelder, E M.: In Situ Structural Changes upon Electrochemical Lithium Insertion in Nanosized Anatase TiO2. In: J. Phys. Chem. C 114 (2010), Nr. 2, 1372–1378. http://pubs.acs.org/doi/abs/10.1021/ jp908786t [141] Aldon, L. ; Kubiak, P. ; Picard, A. ; Jumas, J.C. ; Olivier-Fourcade, J.: Size Particle Effects on Lithium Insertion into Sn-doped TiO2 Anatase. In: Chemistry of materials 18 (2006), Nr. 6, 1401–1406. http://pubs.acs.org/doi/abs/10.1021/cm051445v [142] Brohan, Luc ; Marchand, ReneÂ´: Properties physiques des bronzes MxTiO2 (B). In: Solid State Ionics 9â10 (1983), 419–424. http://dx.doi.org/10.1016/0167-2738(83) 90269-2. – DOI 10.1016/0167–2738(83)90269–2. – ISSN 0167–2738 [143] Brousse, Thierry ; Marchand, René ; Taberna, Pierre-Louis ; Simon, Patrice: TiO2 (B)/activated carbon non-aqueous hybrid system for energy storage. In: Journal of Power Sources 158 (2006), Juli, Nr. 1, 571–577. http://dx.doi.org/10.1016/j.jpowsour.2005. 09.020. – DOI 10.1016/j.jpowsour.2005.09.020. – ISSN 03787753 [144] Sudant, Guillaume ; Baudrin, Emmanuel ; Larcher, Dominique ; Tarascon, Jean-Marie: Electrochemical lithium reactivity with nanotextured anatase-type TiO2. In: Journal of Materials Chemistry (2005), Nr. 1, 1263–1269. http://dx.doi.org/10.1039/b416176a. – DOI 10.1039/b416176a. – ISSN 0959–9428

Bibliography

125

[145] Wagemaker, Marnix ; Mulder, Fokko M.: Properties and Promises of Nanosized Insertion Materials for Li-Ion Batteries. In: Acc. Chem. Res. 46 (2013), Februar, Nr. 5, S. 1206–1215. http://dx.doi.org/10.1021/ar2001793. – DOI 10.1021/ar2001793. – ISSN 1520–4898 [146] Brutti, Sergio ; Gentili, Valentina ; Reale, Priscilla ; Carbone, Lorenzo ; Panero, Stefania: Mitigation of the irreversible capacity and electrolyte decomposition in a LiNi0.5Mn1.5O4/nano-TiO2 Li-ion battery. In: Journal of Power Sources 196 (2011), November, Nr. 22, 9792–9799. http://dx.doi.org/10.1016/j.jpowsour.2011.08.022. – DOI 10.1016/j.jpowsour.2011.08.022. – ISSN 03787753 [147] Pfanzelt, M. ; Kubiak, P. ; Fleischhammer, M. ; Wohlfahrt-Mehrens, M.: TiO2 rutileâAn alternative anode material for safe lithium-ion batteries. In: Journal of Power Sources 196 (2011), August, Nr. 16, 6815–6821. http://dx.doi.org/10.1016/j.jpowsour. 2010.09.109. – DOI 10.1016/j.jpowsour.2010.09.109. – ISSN 03787753 [148] Li, Jianming ; Wan, Wang ; Zhou, Henghui ; Li, Jingjian ; Xu, Dongsheng: Hydrothermal synthesis of TiO2(B) nanowires with ultrahigh surface area and their fast charging and discharging properties in Li-ion batteries. In: Chemical communications (Cambridge, England) 47 (2011), März, Nr. 12, 3439–41. http://dx.doi.org/10.1039/c0cc04634e. – DOI 10.1039/c0cc04634e. – ISBN 1062760360 [149] Ishii, Yosuke ; Kanamori, Yusuke ; Kawashita, Takehiro ; Mukhopadhyay, Indrajit ; Kawasaki, Shinji: Mesoporous carbonâtitania nanocomposites for high-power Li-ion battery anode material. In: Journal of Physics and Chemistry of Solids 71 (2010), April, Nr. 4, 511– 514. http://dx.doi.org/10.1016/j.jpcs.2009.12.024. – DOI 10.1016/j.jpcs.2009.12.024. – ISSN 00223697 [150] Arrouvel, Corinne ; Parker, Stephen C. ; Islam, M S.: Lithium Insertion and Transport in the TiO2(b) Anode Material: A Computational Study. In: Chem. Mater. 21 (2009), Nr. 20, 4778–4783. http://pubs.acs.org/doi/abs/10.1021/cm900373u [151] Nuspl, Gerhard ; Yoshizawa, Kazyunari ; Yamabe, Tokio: Lithium intercalation in TiO2 modifications. In: J. Mater. Chem. 7 (1997), Nr. 12, 2529–2536. http://dx.doi.org/10. 1039/A703935B. – DOI 10.1039/A703935B [152] Guo, Yu-Guo ; Hu, Yong-Sheng ; Maier, Joachim: Synthesis of hierarchically mesoporous anatase spheres and their application in lithium batteries. In: Chemical Communications (2006), Nr. 26, 2783. http://dx.doi.org/10.1039/b605090e. – DOI 10.1039/b605090e. – ISSN 1359–7345 [153] Ren, Yu ; Ma, Zhen ; Bruce, Peter G.: Ordered mesoporous metal oxides: synthesis and applications. In: Chemical Society reviews 41 (2012), Juli, Nr. 14, 4909–27. http: //dx.doi.org/10.1039/c2cs35086f. – DOI 10.1039/c2cs35086f. – ISSN 1460–4744 [154] Wang, K. ; Wei, M. ; Morris, M. a. ; Zhou, H. ; Holmes, J. D.: Mesoporous Titania Nanotubes: Their Preparation and Application as Electrode Materials for Rechargeable Lithium Batteries. In: Advanced Materials 19 (2007), Oktober, Nr. 19, 3016–3020. http:// dx.doi.org/10.1002/adma.200602189. – DOI 10.1002/adma.200602189. – ISSN 09359648

126

Bibliography

[155] Reddy, M. A. ; Kishore, M. S. ; Pralong, V. ; Varadaraju, U. V. ; Raveau, B.: Lithium Intercalation into Nanocrystalline Brookite TiO[sub 2]. In: Electrochemical and Solid-State Letters 10 (2007), Nr. 2, A29. http://dx.doi.org/10.1149/1.2400203. – DOI 10.1149/1.2400203. – ISSN 10990062 [156] Reddy, M. A. ; Kishore, M. S. ; Pralong, V. ; Caignaert, V. ; Varadaraju, U.V. ; Raveau, B.: Room temperature synthesis and Li insertion into nanocrystalline rutile TiO2. In: Electrochemistry Communications 8 (2006), August, Nr. 8, 1299–1303. http: //dx.doi.org/10.1016/j.elecom.2006.05.021. – DOI 10.1016/j.elecom.2006.05.021. – ISSN 13882481 [157] Kubiak, P ; Pfanzelt, M ; Geserick, J ; Hörmann, U ; Hüsing, N ; Kaiser, U ; Wohlfahrt-Mehrens, M: Electrochemical evaluation of rutile TiO2 nanoparticles as negative electrode for Li-ion batteries. In: J. Power Sources 194 (2009), Nr. 2, 1099–1104. http://dx.doi.org/10.1016/j.jpowsour.2009.06.021 [158] Leriche, J B. ; Hamelet, S ; Shu, J ; Morcrette, Mathieu ; Masquelier, C ; Ouvrard, Guy ; Zerrouki, M ; Soudan, P ; Belin, Stéphanie ; Elkaïm, E ; Baudelet, F: An Electrochemical Cell for Operando Study of Lithium Batteries Using Synchrotron Radiation. In: J. Electrochem. Soc. 157 (2010), Nr. 5, A606–A610. http://jes.ecsdl.org/content/ 157/5/A606.short [159] Kraus, W. ; Nolzeb, G.: POWDER CELL â a program for the representation and manipulation of crystal structures and calculation of the resulting X-ray powder patterns. In: Journal of Applied Crystallography 29 (1996), Juni, Nr. 3, 301–303. http://dx.doi.org/10. 1107/S0021889895014920. – DOI 10.1107/S0021889895014920. – ISSN 00218898 [160] Rodriguez-Carvajal, Juan: FULLPROF: A Program for Rietveld Refinement and ... Determination by Neutron Powder Diffraction. In: J. Physica B 192 (1993), S. 55–69 [161] Armstrong, A R. ; Arrouvel, Corinne ; Gentili, Valentina ; Parker, Stephen C. ; Islam, M S. ; Bruce, Peter G.: Lithium Coordination Sites in LixTiO2(B): A Structural and Computational Study. In: Chem. Mater. 22 (2010), 6426–6432. http://pubs.acs.org/ doi/abs/10.1021/cm102589x [162] Krol, Roel van d. ; Goossens, Albert ; Schoonman, Joop: Spatial Extent of Lithium Intercalation in Anatase TiO 2. In: The Journal of Physical Chemistry B 103 (1999), August, Nr. 34, 7151–7159. http://dx.doi.org/10.1021/jp9909964. – DOI 10.1021/jp9909964. – ISSN 1520–6106 [163] Kim, Hyejung ; Kim, Min G. ; Cho, Jaephil: Unique Structural Changes of Three Dimensionally Ordered Macroporous TiO2 Electrode Materials during Electrochemical Cycling. In: Adv. Energy Mater. 2 (2012), S. 1425–1432. http://dx.doi.org/10.1002/aenm.)). – DOI 10.1002/aenm.)) [164] Ben Yahia, Mouna ; Lemoigno, Frédéric ; Beuvier, Thomas ; Filhol, Jean-Sébastien ; Richard-Plouet, Mireille ; Brohan, Luc ; Doublet, Marie-Liesse: Updated references for the structural, electronic, and vibrational properties of TiO2(B) bulk using first-principles density functional theory calculations. In: The Journal of Chemical Physics 130 (2009), Mai,

Bibliography

127

Nr. 20, 204501–11. http://dx.doi.org/10.1063/1.3130674. – DOI 10.1063/1.3130674. – ISSN 1089–7690 [165] Okumura, Toyoki ; Fukutsuka, Tomokazu ; Yanagihara, Asuki ; Orikasa, Yuki ; Arai, Hajime ; Ogumi, Zempachi ; Uchimoto, Yoshiharu: Nanosized Effect on Electronic / Local Structures and Specific Lithium-Ion Insertion Property in TiO2 B Nanowires Analyzed by X-ray Absorption Spectroscopy. In: Chem. Mater. 23 (2011), 3636–3644. http://pubs.acs. org/doi/abs/10.1021/cm200902r [166] Dalton, AS ; Belak, AA ; Ven, Anton Van D.: Thermodynamics of Lithium in TiO2 (B) from First Principles. In: Chemistry of Materials 24 (2012), 1568–1574. http://pubs.acs. org/doi/abs/10.1021/cm203283v [167] Joly, Yves ; Cabaret, Delphine ; Renevier, Hubert ; Natoli, Calogero: Electron Population Analysis by Full-Potential X-Ray Absorption Simulations. In: Physical Review Letters 82 (1999), März, Nr. 11, 2398–2401. http://dx.doi.org/10.1103/PhysRevLett. 82.2398. – DOI 10.1103/PhysRevLett.82.2398. – ISSN 0031–9007 [168] Luca, Vittorio ; Hanley, TL ; Roberts, NK ; Howe, RF: NMR and X-ray absorption study of lithium intercalation in micro-and nanocrystalline anatase. In: Chemistry of materials 11 (1999), Nr. 12, 2089–2102. http://pubs.acs.org/doi/abs/10.1021/cm990007j [169] Fehse, M. ; Cavaliere, S. ; Lippens, P. E. ; Savych, I. ; Iadecola, A. ; Monconduit, L. ; Jones, D. J. ; Rozière, J. ; Fischer, F. ; Tessier, C. ; Stievano, L.: Nb-doped TiO2 Nanofibers for Lithium Ion Batteries. In: The Journal of Physical Chemistry C 117 (2013), Juni, 13827–13835. http://dx.doi.org/10.1021/jp402498p. – DOI 10.1021/jp402498p. – ISSN 1932–7447 [170] Parlebas, J.C. ; Khan, M.A. ; Uozumi, T. ; Okada, K. ; Kotani, A.: Theory of many-body effects in valence, core-level and isochromat spectroscopies along the 3d transition metal series of oxides. In: Journal of Electron Spectroscopy and Related Phenomena 71 (1995), März, Nr. 2, 117–139. http://dx.doi.org/10.1016/0368-2048(94)02262-3. – DOI 10.1016/0368–2048(94)02262–3. – ISSN 03682048 [171] Farges, François ; Brown, Gordon ; Rehr, J.: Ti K-edge XANES studies of Ti coordination and disorder in oxide compounds: Comparison between theory and experiment. In: Physical Review B 56 (1997), Juli, Nr. 4, 1809–1819. http://dx.doi.org/10.1103/PhysRevB.56. 1809. – DOI 10.1103/PhysRevB.56.1809. – ISSN 0163–1829 [172] Lindström, Henrik ; Södergren, Sven ; Solbrand, Anita ; Rensmo, Hå k. ; Hjelm, Johan ; Hagfeldt, Anders ; Lindquist, Sten-Eric: Li + Ion Insertion in TiO 2 (Anatase). 2. Voltammetry on Nanoporous Films. In: The Journal of Physical Chemistry B 101 (1997), September, Nr. 39, 7717–7722. http://dx.doi.org/10.1021/jp970490q. – DOI 10.1021/jp970490q. – ISSN 1520–6106 [173] Zhu, Kai ; Wang, Qing ; Kim, JH ; Pesaran, AA: Pseudocapacitive Lithium-Ion Storage in Oriented Anatase TiO2 Nanotube Arrays. In: The Journal of Physical (2012), 2–6. http://pubs.acs.org/doi/abs/10.1021/jp301884x

128

Bibliography

[174] Bard, Allen J. ; Faulkner, Larry R.: Electrochemical methods : fundamentals and applications - 2nd edition. 2nd editio. John Wiley & Sons, Inc.. – 850 S. http://fora.aa. ufl.edu/docs/47/20Sept11/UCC_20Sept11_ABE5XXX.pdf. – ISBN ISBN 0–471–04372–9 [175] Hao, Bo ; Yan, Yong ; Wang, Xiaobo ; Chen, Ge: Synthesis of Anatase TiO_2 Nanosheets with Enhanced Pseudocapacitive Contribution for Fast Lithium Storage. In: ACS applied materials & interfaces 5 (2013), Juli, Nr. 13, 6285–91. http://dx.doi.org/10.1021/ am4013215. – DOI 10.1021/am4013215. – ISSN 1944–8252 [176] Panduwinata, Dwi ; Gale, Julian D.: A first principles investigation of lithium intercalation in TiO2-B. In: Journal of Materials Chemistry 19 (2009), Nr. 23, 3931. http://dx.doi. org/10.1039/b902683e. – DOI 10.1039/b902683e. – ISSN 0959–9428 [177] Patel, S.K.S. ; Gajbhiye, N.S.: Room temperature magnetic properties of Cu-doped titanate, TiO2(B) and anatase nanorods synthesized by hydrothermal method. In: Materials Chemistry and Physics 132 (2012), Januar, Nr. 1, 175–179. http://dx.doi.org/10.1016/j. matchemphys.2011.11.020. – DOI 10.1016/j.matchemphys.2011.11.020. – ISSN 02540584 [178] Yang, Xiujuan ; Teng, Donghua ; Liu, Bingxue ; Yu, Yunhua ; Yang, Xiaoping: Nanosized anatase titanium dioxide loaded porous carbon nanofiber webs as anode materials for lithiumion batteries. In: Electrochemistry Communications 13 (2011), Juli, Nr. 10, 1098–1101. http: //dx.doi.org/10.1016/j.elecom.2011.07.007. – DOI 10.1016/j.elecom.2011.07.007. – ISSN 13882481 [179] Feng, Chuanqi ; Tang, Jing ; Zhang, Chaofeng ; Zeng, Rong: Synthesis and Electrochemical Properties of TiO2/C Nano-Fiber Composite. In: Nanoscience and Nanotechnology Letters 4 (2012), April, Nr. 4, 430–434. http://dx.doi.org/10.1166/nnl.2012.1336. – DOI 10.1166/nnl.2012.1336. – ISSN 19414900 [180] Zhang, Xiang ; Kumar, Palaniswamy S.: Electrospun TiO2-Graphene Composite Nanofibers as Highly Durable Insertion Anode for Lithium-Ion Batteries. In: The Journal of Physical Chemistry C 116 (2012), 14780–14788. http://pubs.acs.org/doi/abs/10.1021/ jp302574g [181] Manthiram, A. ; Vadivel Murugan, A. ; Sarkar, A. ; Muraliganth, T.: Nanostructured electrode materials for electrochemical energy storage and conversion. In: Energy & Environmental Science 1 (2008), Nr. 6, 621–638. http://dx.doi.org/10.1039/b811802g. – DOI 10.1039/b811802g. – ISSN 1754–5692 [182] Yuan, Tao ; Zhao, Bote ; Cai, Rui ; Zhou, Yingke ; Shao, Zongping: Electrospinning based fabrication and performance improvement of film electrodes for lithium-ion batteries composed of TiO2 hollow fibersâ . In: Journal of Materials Chemistry 21 (2011), 15041–15048. http://dx.doi.org/10.1039/c1jm11483b. – DOI 10.1039/c1jm11483b. – ISSN 0959–9428 [183] Jin, Ming ; Zhang, Xintong ; Pu, Hongting ; Nishimoto, Shunsuke ; Murakami, Taketoshi ; Fujishima, Akira: Photochromism-based detection of volatile organic compounds by W-doped TiO(2) nanofibers. In: Journal of Colloid and Interface Science 362 (2011), Oktober, Nr. 1, 188–93. http://dx.doi.org/10.1016/j.jcis.2011.06.041. – DOI 10.1016/j.jcis.2011.06.041. – ISSN 1095–7103

Bibliography

129

[184] Park, Ju-Young ; Lee, Jae-Hyeok ; Choi, Do-Young ; Hwang, Cheol-Ho ; Lee, Jae-Wook: Influence of Fe doping on phase transformation and crystallite growth of electrospun TiO2 nanofibers for photocatalytic reaction. In: Materials Letters 88 (2012), Dezember, 156–159. http://dx.doi.org/10.1016/j.matlet.2012.07.080. – DOI 10.1016/j.matlet.2012.07.080. – ISSN 0167577X [185] Bingham, Sonny ; Daoud, Walid a.: Recent advances in making nano-sized TiO2 visiblelight active through rare-earth metal doping. In: Journal of Materials Chemistry 21 (2011), Nr. 7, 2041–2050. http://dx.doi.org/10.1039/c0jm02271c. – DOI 10.1039/c0jm02271c. – ISSN 0959–9428 [186] Han, Hyungkyu ; Song, Taeseup ; Bae, Jae-Young ; Nazar, Linda F. ; Kim, Hansu ; Paik, Ungyu: Nitridated TiO2 hollow nanofibers as an anode material for high power lithium ion batteries. In: Energy & Environmental Science 4 (2011), 4532–4536. http: //dx.doi.org/10.1039/c1ee02333k. – DOI 10.1039/c1ee02333k. – ISSN 1754–5692 [187] Archana, Panikar S. ; Jose, Rajan ; Jin, Tan M. ; Vijila, Chellapan ; Yusoff, Mashitah M. ; Ramakrishna, Seeram: Structural and Electrical Properties of Nb-Doped Anatase TiO2 Nanowires by Electrospinning. In: Journal of the American Ceramic Society 93 (2010), Dezember, Nr. 12, 4096–4102. http://dx.doi.org/10.1111/j.1551-2916.2010.04003.x. – DOI 10.1111/j.1551–2916.2010.04003.x. – ISSN 00027820 [188] An, HyeLan ; Ahn, Hyo-Jin: Fabrication of wrinkled Nb-doped TiO2 nanofibres via electrospinning. In: Materials Letters 93 (2013), Februar, 88–91. http://dx.doi.org/10. 1016/j.matlet.2012.11.066. – DOI 10.1016/j.matlet.2012.11.066. – ISSN 0167577X [189] Guidi, V ; Carotta, MC: Effect of dopants on grain coalescence and oxygen mobility in nanostructured titania anatase and rutile. In: The Journal of Physical Chemistry B 107 (2003), 120–124. http://pubs.acs.org/doi/abs/10.1021/jp013572u [190] Baddour-Hadjean, Rita ; Pereira-Ramos, Jean-Pierre: Raman Microspectrometry Applied to the Study of Electrode Materials for Lithium Batteries. In: Chem. Rev. 110 (2010), Nr. 3, S. 1278–1319 [191] Lejon, Christian ; Österlund, Lars: Influence of phonon confinement, surface stress, and zirconium doping on the Raman vibrational properties of anatase TiO2 nanoparticles. In: Journal of Raman Spectroscopy 42 (2011), November, Nr. 11, 2026–2035. http://dx.doi. org/10.1002/jrs.2956. – DOI 10.1002/jrs.2956. – ISSN 03770486 [192] Perego, Céline ; Revel, Renaud ; Durupthy, Olivier ; Cassaignon, Sophie ; Jolivet, Jean-Pierre: Thermal stability of TiO2-anatase: Impact of nanoparticles morphology on kinetic phase transformation. In: Solid State Sci. 12 (2010), Nr. 6, 989–995. http: //dx.doi.org/10.1016/j.solidstatesciences.2009.07.021 [193] Song, Haiyan ; Sjå stad, Anja O. ; Fjellvå g, Helmer ; Okamoto, Hiroshi ; Vistad, Ørnulv B. ; Arstad, Bjø r. ; Norby, Poul: Exfoliation and thermal transformations of Nb-substituted layered titanates. In: Journal of Solid State Chemistry 184 (2011), Dezember, Nr. 12, 3135–3143. http://dx.doi.org/10.1016/j.jssc.2011.08.039. – DOI 10.1016/j.jssc.2011.08.039. – ISSN 00224596

130

Bibliography

[194] Kuznetsov, M. V. ; Zhuravlev, Ju. F. ; Zhilyaev, V. A. ; Gubanov, V. A.: XPS study of the nitrides, oxides and oxynitrides of titanium. In: Journal of Electron Spectroscopy and Related Phenomena 1 58 (1992), Nr. 1-2, 1–9. http://dx.doi.org/10.1016/0368-2048(92) 80001-O [195] Bahl, M.K.: ESCA studies of some niobium compounds. In: Journal of Physics and Chemistry of Solids 36 (1975), Nr. 6, 485–491. http://dx.doi.org/10.1016/0022-3697(75) 90132-8 [196] Nowotny, J. ; Sunggyu Lee (Hrsg.): Oxide Semiconductors for Solar Energy Conversion TiO2. Boca Raton, Fl : CRC Press, 2012. – 414 S. – ISBN 9781439848395 [197] Wang, Peifang ; Wu, Jiancheng ; Ao, Yanhui ; Wang, Chao ; Hou, Jun ; Qian, Jin: Preparation and enhanced photocatalytic performance of Sn ion modified titania hollow spheres. In: Materials Letters 65 (2011), November, Nr. 21-22, 3278–3280. http://dx. doi.org/10.1016/j.matlet.2011.07.029. – DOI 10.1016/j.matlet.2011.07.029. – ISSN 0167577X [198] Ehinon, K. K. D. ; Naille, S ; Dedryve, R ; Lippens, P. E. ; Jumas, J.-C. ; Gonbeau, D.: Ni 3 Sn 4 Electrodes for Li-Ion Batteries : Li-Sn Alloying Process and Electrode / Electrolyte Interface Phenomena. In: Chem. Mater. 20 (2008), Nr. 2, 5388–5398. http: //pubs.acs.org/doi/abs/10.1021/cm8006099 [199] Knauth, Philippe ; Chadwick, Alan V. ; Lippens, Pierre E. ; Auer, Gerhard: EXAFS study of dopant ions with different charges in nanocrystalline anatase: evidence for spacecharge segregation of acceptor ions. In: ChemPhysChem 10 (2009), Juli, Nr. 8, 1238–46. http://dx.doi.org/10.1002/cphc.200800806. – DOI 10.1002/cphc.200800806. – ISSN 1439–7641 [200] Bouchet, R ; Weibel, A ; Knauth, P: EXAFS study of dopant segregation (Zn, Nb) in nanocrystalline anatase (TiO2). In: Chemistry of Materials 15 (2003), Nr. 9, 4996–5002. http://pubs.acs.org/doi/abs/10.1021/cm034640n [201] Hitosugi, Taro ; Kamisaka, Hideyuki ; Yamashita, Koichi ; Nogawa, Hiroyuki ; Furubayashi, Yutaka ; Nakao, Shoichiro ; Yamada, Naoomi ; Chikamatsu, Akira ; Kumigashira, Hiroshi ; Oshima, Masaharu ; Hirose, Yasushi ; Shimada, Toshihiro ; Hasegawa, Tetsuya: Electronic Band Structure of Transparent Conductor: NbDoped Anatase TiO2. In: Applied Physics Express 1 (2008), November, 111203–06. http://dx.doi.org/10.1143/APEX.1.111203. – DOI 10.1143/APEX.1.111203. – ISSN 1882–0778 [202] Ding, Yan-Huai ; Zhang, Ping ; Ren, Hu-Ming ; Zhuo, Qin ; Yang, Zhong-Mei ; Jiang, Yong: Preparation of graphene/TiO2 anode materials for lithium-ion batteries by a novel precipitation method. In: Materials Research Bulletin 46 (2011), Dezember, Nr. 12, 2403–2407. http://dx.doi.org/10.1016/j.materresbull.2011.08.046. – DOI 10.1016/j.materresbull.2011.08.046. – ISSN 00255408 [203] Yang, Zunxian ; Du, Guodong ; Meng, Qing ; Guo, Zaiping ; Yu, Xuebin ; Chen, Zhixin ; Guo, Tailiang ; Zeng, Rong: Synthesis of uniform TiO2@carbon composite nanofibers as anode for lithium ion batteries with enhanced electrochemical performance. In: Journal of

Bibliography

131

Materials Chemistry 22 (2012), Nr. 12, 5848. http://dx.doi.org/10.1039/c2jm14852h. – DOI 10.1039/c2jm14852h. – ISSN 0959–9428 [204] Cao, Fei-Fei ; Xin, Sen ; Guo, Yu-Guo ; Wan, Li-Jun: Wet chemical synthesis of Cu/TiO2 nanocomposites with integrated nano-current-collectors as high-rate anode materials in lithium-ion batteries. In: Physical chemistry chemical physics : PCCP 13 (2011), Februar, Nr. 6, 2014–20. http://dx.doi.org/10.1039/c0cp01119c. – DOI 10.1039/c0cp01119c. – ISSN 1463–9084 [205] Yang, Rong ; Zhao, Wei ; Zheng, Jie ; Zhang, Xuanzhou ; Li, Xingguo: One-Step Synthesis of Carbon-Coated Tin Dioxide Nanoparticles for High Lithium Storage. In: J. Phys. Chem. 114 (2010), 20272–20276. http://pubs.acs.org/doi/abs/10.1021/jp107396a [206] Park, Sang-Jun ; Kim, Young-Jun ; Lee, Hyukjae: Synthesis of carbon-coated TiO2 nanotubes for high-power lithium-ion batteries. In: J. Power Sources 196 (2011), 5133–5137. http://dx.doi.org/10.1016/j.jpowsour.2011.01.105 [207] Li, Huiqiao ; Zhou, Haoshen: Enhancing the performances of Li-ion batteries by carbon-coating: present and future. In: Chemical communications (Cambridge, England) 48 (2011), November, 1201–1217. http://dx.doi.org/10.1039/c1cc14764a. – DOI 10.1039/c1cc14764a. – ISBN 8129861348 [208] Zhang, Biao ; Huang, Zhen-Dong ; Oh, Sei W. ; Kim, Jang-Kyo: Improved rate capability of carbon coated Li3.9Sn0.1Ti5O12 porous electrodes for Li-ion batteries. In: Journal of Power Sources 196 (2011), September, Nr. 24, 10692–10697. http://dx.doi.org/10.1016/ j.jpowsour.2011.08.114. – DOI 10.1016/j.jpowsour.2011.08.114. – ISSN 03787753 [209] Das, Shyamal K. ; Darmakolla, Srikarrao ; Bhattacharyya, Aninda J.: High lithium storage in micrometre sized mesoporous spherical self-assembly of anatase titania nanospheres and carbon. In: Journal of Materials Chemistry 20 (2010), Nr. 8, 1600. http://dx.doi. org/10.1039/b919139a. – DOI 10.1039/b919139a. – ISSN 0959–9428 [210] Kim, Hyun-Soo ; Kong, Mingzhe ; Kim, Ketack ; Kim, Ick-Jun ; Gu, Hal-Bon: Effect of carbon coating on LiNi1/3Mn1/3Co1/3O2 cathode material for lithium secondary batteries. In: Journal of Power Sources 171 (2007), September, Nr. 2, 917–921. http://dx.doi. org/10.1016/j.jpowsour.2007.06.028. – DOI 10.1016/j.jpowsour.2007.06.028. – ISSN 03787753 [211] White, William B. ; Knight, Diane S.: Characterization of diamond films by Raman spectroscopy. In: J. Mater. Res. 4 (1989), S. 385–393 [212] Mattia, D ; Rossi, M P. ; Kim, B M. ; Korneva, G ; Bau, H H. ; Gogotsi, Y: Effect of graphitization on the wettability and electrical conductivity of CVD-carbon nanotubes and films. In: The journal of physical chemistry. B 110 (2006), Mai, Nr. 20, 9850–5. http://dx.doi.org/10.1021/jp061138s. – DOI 10.1021/jp061138s. – ISSN 1520–6106 [213] Andersson, Anna S. ; Kalska, Beata ; Häggström, Lennart ; Thomas, John O.: Lithium extraction/insertion in LiFePO4: an X-ray diffraction and Mössbauer spectroscopy study. In: Solid State Ionics 130 (2000), Nr. 1â2, 41–52. http://dx.doi.org/10.1016/ S0167-2738(00)00311-8. – DOI 10.1016/S0167–2738(00)00311–8. – ISSN 0167–2738

132

Bibliography

[214] Cao, Fei-Fei ; Guo, Yu-Guo ; Zheng, Shu-Fa ; Wu, Xing-Long ; Jiang, Ling-Yan ; Bi, RongRong ; Wan, Li-Jun ; Maier, Joachim: Symbiotic Coaxial Nanocables: Facile Synthesis and an Efficient and Elegant Morphological Solution to the Lithium Storage Problem. In: Chem. Mater. 22 (2010), Nr. 5, 1908–1914. http://pubs.acs.org/doi/abs/10.1021/cm9036742 [215] Wang, Yan ; Alsmeyer, Daniel C. ; Mccreery, Richard L.: Raman spectroscopy of Carbon Materials: Structural Basis of Observed Spectra. In: Chem. Mater. 2 (1990), Nr. 10, 557–563. http://pubs.acs.org/doi/abs/10.1021/cm00011a018 [216] Mouyane, M. ; Ruiz, J.-M. ; Artus, M. ; Cassaignon, S. ; Jolivet, J.-P. ; Caillon, G. ; Jordy, C. ; Driesen, K. ; Scoyer, J. ; Stievano, L. ; Olivier-Fourcade, J. ; Jumas, J.-C.: Carbothermal synthesis of Sn-based composites as negative electrode for lithium-ion batteries. In: Journal of Power Sources 196 (2011), August, Nr. 16, 6863–6869. http: //dx.doi.org/10.1016/j.jpowsour.2010.09.001. – DOI 10.1016/j.jpowsour.2010.09.001. – ISSN 03787753 [217] Jarvis, CR ; Macklin, WJ ; Macklin, AJ: Use of grafted PVdF-based polymers in lithium batteries. In: Journal of power . . . 98 (2001), 664–666. http://www.sciencedirect.com/ science/article/pii/S0378775301006966 [218] Kim, G.T. ; Jeong, S.S. ; Joost, M. ; Rocca, E. ; Winter, M. ; Passerini, S. ; Balducci, A.: Use of natural binders and ionic liquid electrolytes for greener and safer lithium-ion batteries. In: Journal of Power Sources 196 (2011), Februar, Nr. 4, 2187–2194. http: //dx.doi.org/10.1016/j.jpowsour.2010.09.080. – DOI 10.1016/j.jpowsour.2010.09.080. – ISSN 03787753 [219] Mancini, M. ; Nobili, F. ; Tossici, R. ; Wohlfahrt-Mehrens, M. ; Marassi, R.: High performance, environmentally friendly and low cost anodes for lithium-ion battery based on TiO2 anatase and water soluble binder carboxymethyl cellulose. In: Journal of Power Sources 196 (2011), November, Nr. 22, 9665–9671. http://dx.doi.org/10.1016/j.jpowsour.2011. 07.028. – DOI 10.1016/j.jpowsour.2011.07.028. – ISSN 03787753 [220] Dzhurinskij, B.F. ; Gati, D. ; Sergushin, N.P. ; Nefedov, V.I. ; Salyn, Ya.V.: Simple coordination compounds. An X-ray photoelectron study of certain oxides. In: Russian Journal of Inorganic Chemistry 20 (1975), S. 20307–2314 [221] Chiang, C: Nano-tube TiO2 composite PVdF/LiPF6 solid membranes. In: Solid State Ionics 175 (2004), November, Nr. 1-4, 631–635. http://dx.doi.org/10.1016/j.ssi.2003.12.039. – DOI 10.1016/j.ssi.2003.12.039. – ISSN 01672738 [222] Goodenough, John B. ; Kim, Youngsik: Challenges for Rechargeable Li Batteries. In: Chem. Mater. 22 (2010), Nr. 3, 587–603. http://pubs.acs.org/doi/abs/10.1021/cm901452z [223] Jones, Jennifer ; Anouti, Mérièm ; Caillon-Caravanier, Magaly ; Willmann, Patrick ; Sizaret, Pierre-Yves ; Lemordant, Daniel: Solubilization of SEI lithium salts in alkylcarbonate solvents. In: Fluid Phase Equilibria 305 (2011), Juni, Nr. 2, 121–126. http://dx.doi.org/10.1016/j.fluid.2011.03.007. – DOI 10.1016/j.fluid.2011.03.007. – ISSN 03783812

Bibliography

133

[224] Xu, Kang: Nonaqueous Liquid Electrolytes for Lithium-Based Rechargeable Batteries. In: Chem. Rev. Bd. 104, 4303–4417 [225] Zhang, Xuerong ; Kostecki, Robert ; Richardson, Thomas J. ; Pugh, James K. ; Ross, Philip N.: Electrochemical and Infrared Studies of the Reduction of Organic Carbonates. In: Journal of The Electrochemical Society 148 (2001), Nr. 12, A1341. http://dx.doi.org/10. 1149/1.1415547. – DOI 10.1149/1.1415547. – ISSN 00134651 [226] Zhang, S. S. ; Xu, K. ; Jow, T. R.: Study of LiBF[sub 4] as an Electrolyte Salt for a Li-Ion Battery. In: Journal of The Electrochemical Society 149 (2002), Mai, Nr. 5, A586. http://dx.doi.org/10.1149/1.1466857. – DOI 10.1149/1.1466857. – ISSN 00134651 [227] Borghols, Wouter J. H. ; LuÌtzenkirchen-Hecht, Dirk ; Haake, Ullrich ; Chan, Wingkee ; Lafont, Ugo ; Kelder, Erik M. ; Eck, Ernst R. H. ; Kentgens, Arno P. M. ; Mulder, Fokko M. ; Wagemaker, Marnix: Lithium Storage in Amorphous TiO[sub 2] Nanoparticles. In: Journal of The Electrochemical Society 157 (2010), Nr. 5, A582. http://dx.doi.org/10.1149/1.3332806. – DOI 10.1149/1.3332806. – ISSN 00134651 [228] Yu, Jiaguo ; Dai, Gaopeng ; Xianga, Quanjun ; Jaroniec, Mietek: Fabrication and enhanced visible-light photocatalytic activity of carbon self-doped TiO2 sheets with exposed {001} facets. In: J. Mater. Chem. 21 (2011), 1049–1057. http://pubs.rsc.org/en/content/ articlehtml/2011/jm/c0jm02217a [229] Dylla, Anthony G. ; Xiao, Penghao ; Henkelman, Graeme ; Stevenson, Keith J.: Morphological Dependence of Lithium Insertion in Nanocrystalline TiO2(b) nanoparticles and nanosheets. In: J. Phys. Chem. Lett. 2 (2012), 2015–2019. http://pubs.acs.org/doi/ abs/10.1021/jz300766a [230] Wang, H E. ; Lu, Z G. ; Xi, L J. ; Ma, R G. ; Wang, C D. ; Zapien, J a. ; Bello, I: Facile and Rapid Synthesis of Highly Porous Wirelike TiO(2) as Anodes for LithiumIon Batteries. In: ACS applied materials & interfaces 4 (2012), März, Nr. 3, 1608–13. http://dx.doi.org/10.1021/am2017738. – DOI 10.1021/am2017738. – ISSN 1944–8252 [231] Ue, Makoto: Mobility and Ionic Association of Lithium Salts in a Propylene Carbonate-Ethyl Methyl Carbonate Mixed Solvent. In: Journal of The Electrochemical Society 142 (1995), August, Nr. 8, 2577. http://dx.doi.org/10.1149/1.2050056. – DOI 10.1149/1.2050056. – ISSN 00134651 [232] Leroy, S. ; Martinez, H. ; Dedryvère, R. ; Lemordant, D. ; Gonbeau, D.: Influence of the lithium salt nature over the surface film formation on a graphite electrode in Li-ion batteries: An XPS study. In: Applied Surface Science 253 (2007), März, Nr. 11, 4895–4905. http://dx.doi.org/10.1016/j.apsusc.2006.10.071. – DOI 10.1016/j.apsusc.2006.10.071. – ISSN 01694332 [233] Gaberscek, Miran: Towards optimized preparation of cathode materials: How can modeling and concepts be used in practice. In: Journal of Power Sources 189 (2009), April, Nr. 1, 22–27. http://dx.doi.org/10.1016/j.jpowsour.2008.12.041. – DOI 10.1016/j.jpowsour.2008.12.041. – ISSN 03787753

134

Bibliography

[234] Fongy, C. ; Jouanneau, S. ; Guyomard, D. ; Lestriez, B.: Carbon nanofibers improve both the electronic and ionic contributions of the electrochemical performance of composite electrodes. In: Journal of Power Sources 196 (2011), Oktober, Nr. 20, 8494–8499. http: //dx.doi.org/10.1016/j.jpowsour.2011.06.022. – DOI 10.1016/j.jpowsour.2011.06.022. – ISSN 03787753 [235] Wang, Jin ; Liu, Xiao-Min ; Yang, Hui ; Shen, Xiao-dong: Characterization and electrochemical properties of carbon-coated Li4Ti5O12 prepared by a citric acid solâgel method. In: J. Alloys Compd. 509 (2011), 712–718. http://dx.doi.org/10.1016/j. jallcom.2010.07.215 [236] Yu, Seung-Ho ; Pucci, Andrea ; Herntrich, Tobias ; Willinger, Marc-Georg ; Baek, Seung-Hwan ; Sunga, Yung-Eun ; Pinna, Nicola: Surfactant-free nonaqueous synthesis of lithium titanium oxide (LTO) nanostructures for lithium ion battery applications. In: J. Mater. Chem. 21 (2011), 806–810. http://pubs.rsc.org/en/content/articlehtml/ 2011/jm/c0jm03064c [237] Hermann-Geppert, Iris: Gas sorption measurements. https://www. helmholtz-berlin.de/forschung/enma/solare-brennstoffe/analytische-methoden/ gassorptionsmessungen_en.html [238] Williamson, G.K ; Hall, W.H: X-ray line broadening from filed aluminium and wolfram. In: Acta Metallurgica 1 (1953), Januar, Nr. 1, 22–31. http://dx.doi.org/10.1016/ 0001-6160(53)90006-6. – DOI 10.1016/0001–6160(53)90006–6. – ISSN 00016160 [239] Hohenberg, P ; Kohn, W: Inhomogenous Electron Gas. In: Phys. Rev. 136 (1964), Nr. 3B, S. B864–871 [240] Yamamoto, Takashi: Assignment of pre-edge peaks in K-edge x-ray absorption spectra of 3d transition metal compounds: electric dipole or quadrupole? In: X-ray Spectrom. 37 (2008), Nr. 6, 572–584. http://onlinelibrary.wiley.com/doi/10.1002/xrs.1103/abstract [241] Perdew, Jp ; Burke, K ; Ernzerhof, M: Generalized Gradient Approximation Made Simple. In: Physical Review Letters 77 (1996), Oktober, Nr. 18, 3865–3868. http://www. ncbi.nlm.nih.gov/pubmed/10062328. – ISSN 1079–7114 [242] Blaha, P. ; Schwarz, K. ; Madsen, G.K.H. ; Kvasnicka, D ; Luitz, J.: WIEN2K An augmented plane wave plus local orbitals program for calculating crystal properties. 2001 [243] Weibel, A ; Bouchet, R ; Savin, S L P. ; Chadwick, A V. ; Lippens, P E. ; Womes, M ; Knauth, P: Local atomic and electronic structure in nanocrystalline Sn-doped anatase TiO2. In: ChemPhysChem 7 (2006), November, Nr. 11, 2377–2383. http://dx.doi.org/ 10.1002/cphc.200600370. – DOI 10.1002/cphc.200600370. – ISSN 1439–4235 [244] Lippens, PE ; Chadwick, AV: Structure and chemical bonding in Zr-doped anatase TiO2 nanocrystals. In: The Journal of Physical Chemistry C 112 (2008), 43–47. http: //pubs.acs.org/doi/abs/10.1021/jp075898u [245] Lippens, Pierre-Emmanuel ; Womes, Manfred ; Kubiak, Pierre ; Jumas, Jean-Claude ; Olivier-Fourcade, Josette: Electronic structure of the spinel Li4Ti5O12 studied by ab

Bibliography

135

initio calculations and X-ray absorption spectroscopy. In: Solid State Sciences 6 (2004), Februar, Nr. 2, 161–166. http://dx.doi.org/10.1016/j.solidstatesciences.2003.12. 001. – DOI 10.1016/j.solidstatesciences.2003.12.001. – ISSN 12932558 [246] Kresse, G. ; Hafner, J.: Ab initio molecular dynamics for liquid metals. In: Physical Review B 47 (1993), Januar, Nr. 1, 558–561. http://dx.doi.org/10.1103/PhysRevB.47.558. – DOI 10.1103/PhysRevB.47.558. – ISSN 0163–1829 [247] Kresse, G.: From ultrasoft pseudopotentials to the projector augmented-wave method. In: Physical Review B 59 (1999), Januar, Nr. 3, 1758–1775. http://dx.doi.org/10.1103/ PhysRevB.59.1758. – DOI 10.1103/PhysRevB.59.1758. – ISSN 1098–0121 [248] Blöchl, P. E.: Projector augmented-wave method. In: Physical Review B 50 (1994), Dezember, Nr. 24, 17953–17979. http://dx.doi.org/10.1103/PhysRevB.50.17953. – DOI 10.1103/PhysRevB.50.17953. – ISSN 0163–1829 [249] Franchini, C. ; Podloucky, R. ; Paier, J. ; Marsman, M. ; Kresse, G.: Groundstate properties of multivalent manganese oxides: Density functional and hybrid density functional calculations. In: Physical Review B 75 (2007), Mai, Nr. 19, 195128. http: //dx.doi.org/10.1103/PhysRevB.75.195128. – DOI 10.1103/PhysRevB.75.195128. – ISSN 1098–0121 [250] Monkhorst, Hendrik J. ; Pack, James D.: Special points for Brillouin-zone integrations. In: Physical Review B 13 (1976), Juni, Nr. 12, 5188–5192. http://dx.doi.org/10.1103/ PhysRevB.13.5188. – DOI 10.1103/PhysRevB.13.5188. – ISSN 0556–2805 [251] Dovesi, Roberto ; Saunders, Victor R. ; Roetti, Carla ; Orlando, Roberto ; ZicovichWilson, C M. ; Pascale, Fabien ; Civalleri, Bartolomeo ; Doll, Klaus ; Harrison, Nicholas M. ; Bush, Ian J. ; Others: CRYSTAL06 user manual. In: University of Torino, Torino (2006)

List of Figures 1.1 1.2 1.3 1.4

1.5

1.6

1.7 2.1 2.2 2.3

2.4 2.5 2.6

World energy demand history and forecast, showing contributing energy sources. Figure taken from [3]. . . . . . . . . . . . . . . . . . . . . . . . . Comparison of gravimetric and volumetric energy density of various secondary battery systems. Figure taken from [10]. . . . . . . . . . . . . . . . Specific energy vs. specific power of cathode materials for lithium ion battery systems and their year of introduction, taken from [16]. . . . . . . Schematic assembly of standard lithium ion battery in discharge mode using graphite as negative and LiMO2 as positive electrode material. Figure taken from [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic presentation of reaction mechanism and properties of three main electrode material types for use in lithium ion batteries. Black circles: voids in the crystal structure, blue circles: metal atoms, yellow circles: lithium ions. Figure taken from [18]. . . . . . . . . . . . . . . . . . . . . . . . . . . Idealized voltage and differential capacity (δC /δV) profiles for basic battery, pseudocapacitor and capacitor energy storage. Pseudocapacitor is hereby the name used for a intermediate between high capacity battery and high rate capacitive energy storage, Fig. taken from [51]. . . . . . . . . . . . . First galvanostatic discharge curves of commercial anatase with two different grain sizes (200 and 5-10 nm) and nanoporous anatase. Fig. taken from [53]. Flow chart of hydrothermal synthesis. . . . . . . . . . . . . . . . . . . . . Formation of octahedral-shaped TiO6 ions upon dissolving. Figure based on [90]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nitrogen physisorption curves and TEM pictures of three main morphologies; sheets, nanoparticles and rods synthesized throughout adequately adjusting solid-to-liquid ratio (powder concentration) in the mother suspension. . . XRD pattern of three biphasic TiO2 mixtures containing different ratios of anatase and TiO2 (B) phase. . . . . . . . . . . . . . . . . . . . . . . . . . . Raman spectra of three biphasic TiO2 mixtures containing different ratios of anatase and TiO2 (B) phase. . . . . . . . . . . . . . . . . . . . . . . . . HRTEM micrograph of (a) anatase-rich and (b) TiO2 (B)-rich sample with magnified cutouts for interplanar distance measurement. . . . . . . . . . .

1 2 3

4

5

7 9 16 16

18 19 19 20

137

138

2.7

2.8

2.9 2.10 2.11 2.12 2.13 2.14

2.15 2.16 2.17

2.18 2.19 2.20 3.1

3.2 3.3

3.4 3.5

List of Figures

Trends and dependencies found for hydrothermal synthesis of regarding phase composition and morphology of titania. Increasing dilution (Na/Ti ratio) leads from a) sheets to b) elongated nanoparticles and finally to c) developed nanorods/nanoribbons while increasing the pH promotes TiO2 (B) formation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Comparison of diffraction patterns of TiO2 (B)-rich and TiO2 (B)-pure sample and (b) Raman spectra of Anatase/TiO2 (B) mixtures derived from NaOH solution and TiO2 (B)-rich derived from KOH solution. . . . . . . . Scheme of sol gel process. . . . . . . . . . . . . . . . . . . . . . . . . . . . SEM and TEM image of conventional sol-gel synthesis sample. . . . . . . . a) Physisorption isotherm and b) poresize distribution of conventional sol-gel synthesis sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SEM and TEM image of ultrasonic assisted sol-gel synthesis. . . . . . . . . a) Physisorption isotherm and b) poresize distribution of ultrasound assisted sol-gel synthesis TiO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic drawing of triblock copolymer P123, consisting of two hydrophilic PEO blocks, which are connected via hydrophobic PPO block. Figure taken from [127]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SEM and TEM image of sol-gel synthesis using P123 as template. . . . . . a) Physisorption isotherm and b) poresize distribution of template assisted sol-gel synthesis TiO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scheme of electrospinning in which precursor solution is fed constantly through a syringe into strong magnetic field leading to a continuous fiber, which is collected on a plate. . . . . . . . . . . . . . . . . . . . . . . . . . (a) TEM and (b) SEM micrograph of electrospun TiO2 nanofibers. . . . . . (a) XRD pattern and (b) Raman spectra of electrospun TiO2 nanofibers. . (a) Hysteresis of physisorption isotherms and (b) incremental and cumulative pore size distribution for electrospun TiO2 nanofibers. . . . . . . . . . . . (left:) Potential as function of mole of lithium inserted per mol TiO2 for the first two galvanostatic cycles @ C/20 and (right): corresponding derivatives of titania nanorod samples with anatase-to-TiO2 (B) ratio of (a) 10/90 (b) 40/60 and (c) 80/20. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evolution of capacities for samples of high (80/20), medium (60/40) and low content (10/90) content of Anatase as function of cycling rate. . . . . Derivative curves of an even mix sample at a) C/20 and b) 5C. Characteristic lithium insertion and de-insertion peaks of the two coexisting phases are marked. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shift redox potential as function of cycling rate for the anatase and TiO2 (B) phase component of even mixture(40/60) sample including standard deviation. a) Galvanostatic cycling curve with corresponding derivative and b) rate capability of TiO2 anatase from conventional sol-gel synthesis. . . . . . . .

21

22 23 25 25 26 26

27 27 28

29 29 30 30

35 36

36 37 38

List of Figures

3.6

3.7

3.8

3.9

3.10 3.11

3.12

3.13 3.14

3.15 3.16

3.17

3.18

a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of TiO2 anatase from ultrasound assisted sol-gel synthesis. . . . . . . . . . . . . . a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of TiO2 anatase from template assisted sol-gel synthesis. . . . . . . . . . . . . . . . a) Galvanostatic cycling curve with corresponding derivative showing first two cycles in black and 20th cycle in red and b) rate capability of electrospun TiO2 anatase nanofibers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reversible capacity at slow (C/20) and high (5C) cycling rate as function of specific surface area for differently synthesized TiO2 anatase samples. Showing general trend of increasing capacity with increase of specific area, with deviating of the sonicated sample. . . . . . . . . . . . . . . . . . . . . Customized in situ cell for transmission (full arrows) and reflection mode (empty arrows) characterization techniques [158]. . . . . . . . . . . . . . . In situ XRD measurement upon consecutive discharge (Li+ insertion) and charge (Li+ de-insertion) of TiO2 (B) rich self supported film, showing reversible peak shift of (110),(003),(601) and (020) all attributed to the TiO2 (B) phase, while peak at 46◦ results from Be window. . . . . . . . . . Lattice expansion of TiO2 (B) upon lithium insertion as derived from XRD peak shift using PowderCell. Hollow markers are based on full pattern matching of accumulated diffractograms before and after lithium insertion. In situ XRD measurement upon consecutive discharge (Li+ insertion) and charge (Li+ de-insertion) of anatase rich sample. . . . . . . . . . . . . . . (a)In situ XRD patterns and (b) electrochemical cycling curve upon first lithium insertion into anatase rich sample; showing three main domains of lithium insertion; I. x≤0.08 solid solution without structural changes, II. 0.08≤x≤0.48 transformation of tetragonal to orthorhombic phase and III. x≥0.45 solid solution uptake into orthorhombic structure linked to a gradual variation of the lattice parameters. . . . . . . . . . . . . . . . . . In situ experimental setup at the XAS beamline at DESY Hamburg in transmission mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . a) Evolution of Ti K-edge upon Li insertion and b) reversible shift of Ti K-edge position during reduction and subsequent oxidation of TiO2 (B) rich sample. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fourier transform and EXAFS signals of TiO2 (B) rich sample as function of lithium inserted, dotted lines are roughly indicating position of contributing next neighbours of Ti central atom. . . . . . . . . . . . . . . . . . . . . . . Bond length modification upon lithium insertion into TiO2 (B) rich sample derived from FT EXAFS signal (hollow markers) and from XRD(filled markers) for (a) oxygen and (b) titanium next neighbors. . . . . . . . . .

139

40

40

41

42 45

46

47 48

49 50

51

51

52

140

List of Figures

3.19 Comparison of experimental and fitted next neighbour real space contribution considering the repartition of next neighbour oxygens in 6-0, 5-1 and 4-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.20 Coordination of Ti within TiO2 (B) structure revealing two different Ti environments; strongly distorted square pyramidal five-fold Ti1 and less stronly distorted octahedral six-fold coordinated Ti2. Calculated Ti-O distances are marked in red and measured ones in blue, Fig. taken from [164]. 3.21 Evolution of Debye-Waller factor for Oxygen and Titanium next neighbour shells upon lithium insertion into TiO2 (B) rich sample . . . . . . . . . . . 3.22 In situ XAS study of Ti K (a)edge and (b) pre-edge evolution upon electrochemical lithium insertion for anatase rich sample. . . . . . . . . . . . . . 3.23 Evolution of (a) Ti-O and (b) Ti-Ti bond lengths upon Li+ insertion derived from fitting insitu EXAFS data. . . . . . . . . . . . . . . . . . . . . . . . 3.24 Evolution of Debye-Waller factor for inner and outer oxygen shell upon lithium insertion in anatase. . . . . . . . . . . . . . . . . . . . . . . . . . . 3.25 Cyclic voltammetry curves at progressively increased scan rate (0.01-2.5 mV /s) for (a) anatase rich and (b) TiO2 (B) rich sample. . . . . . . . . . . . . . . 3.26 Peak current of anatase rich and TiO2 (B) rich sample as function of (a) square root of scan rate and (b) scan rate, including fitted exponential relation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.27 Crystallographic structure of monoclinic TiO2 (B) illustrating; (a) three distinct lithium insertion sites (green) A1, A2, and C, Ti atoms (blue) O atoms (red) Fig. taken from [49], and (b) two distinct Ti environments namely pyramidal coordinated Ti1 (blue) and octahedral coordinated Ti2 (red) along with equatorial and apical planes, Fig. as courtesy of Mouna Ben Yahia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.28 Comparison of experimental and calculated evolution of lattice parameters a, b, and c upon lithium insertion in monoclinic TiO2 (B) (a), (b), and (c), respectively. The calculated graphs were obtained using DFT and VASP assuming single site intercalation . . . . . . . . . . . . . . . . . . . . . . . 4.1 4.2 4.3 4.4

4.5

a) EDX image and b) elemental mapping of Ti and c) of Nb of tape casted film electrode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison of a) galvanostatic cycling curve and b) mean rate capability of non-doped and Nb-doped electrospun TiO2 nanofiber electrodes. . . . . (a) SEM and (c) TEM micrographs of Nb-doped electrospun TiO2 nanofibers and (b) distribution of fiber diameter. . . . . . . . . . . . . . . . . . . . . a) XRD diffractogram with anatase reference [ICSD 009853] and b) Raman spectra comparison of non-doped and Nb-doped TiO2 nanofibers, showing the complete survey and an enlarged cut out of most intense signal. . . . XRD diffractograms upon constant heating at 1 ◦ C/min for (a) non-doped and (b) Nb-doped TiO2 electrospun nanofibers. . . . . . . . . . . . . . .

52

53 54 56 57 57 58

60

61

63 66 67 68

69 70

List of Figures

4.6 4.7 4.8

4.9

4.10 4.11 4.12 4.13 4.14 4.15 4.16

4.17

5.1 5.2 5.3 5.4

5.5 5.6

5.7

spectra a) Nb 3d peaks and b) Ti 2p peaks comparison of doped and non-doped electrospun TiO2 nanofibers. . . . . . . . . . . . . . . . . . . . Measured Ti K-edge of XANES spectra of non-doped and Nb-doped TiO2 . Theoretical XANES spectra at K-edge of non-doped and Nb-doped anatase TiO2 (about 3, 6 and 12 at. % of Nb). The broadened Ti p DOS is shown for comparison. The origin of energy is taken at the Fermi level. . . . . . K 2 weighted a) EXAFS signal and b) Fourier transform of non-doped and doped TiO2 nanofibers. The latter revealing position of next neighbor shells of central Ti atom for Nb-doped and non-doped TiO2 . . . . . . . . . . . . Misfit of EXAFS fit as function of extent of Nb doping of TiO2 . . . . . . . Calculated partial densities of states of NbTi15 O32 (≈6 at%). The origin of energy is taken at the Fermi level (dotted line) . . . . . . . . . . . . . . . Raman spectra of (a) Fe, (b) Ce and (c) Cu with 3% and 10% dopant in mother solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ultraviolet (UV) and visible spectra of doped and non-doped TiO2 (B) sample. XPS survey of 3% Cu-doped TiO2 (B) and high resolution spectrum of the Cu2p3/2 peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Elechtrochemical cycling curve and corresponding derivative of first two cycles and (b) rate capability of 10 %Fe doped TiO2 (B) . . . . . . . . . . (a) TEM picture of TiO2 rod with carbon coating/decoration and (b) Raman spectra of pristine and carbon coated/decorated TiO2 , indicating position of D- and G-band. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . a) Electrochemical cycling curve and corresponding derivative at cycling rate C/20 and b) rate capability for progressively increased cycling rate of a carbon decorated TiO2 with additional carbon. . . . . . . . . . . . . . .

141

XPS

Scheme of Swagelok assembly in half cell configuration vs. lithium metal. Scheme of tape casting procedure with doctor blade on suction table. . . . (a) Punched out electrode film sample, (b) film on copper foil support (c) SEM top view at 10k magnification and (d) SEM profile view . . . . . . . . Galvanostatic cycling curve and corresponding derivative of TiO2 (B) rich film electrode processed at SAFT and cycled in swagelok assembly at C/20 cycling rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Capacities of SAFT prepared TiO2 (B)rich film electrodes for subsequent cycling rates for (a) non calendered and (b) calendered sample. . . . . . . Raman spectra of carbon black (CB) and carbon nanotubes (CNT), showing D and G-band, as well as peak fit. While G-band contribution of CNT is fitted with one component, two are used for fitting CB’s G-band. . . . . . SEM picture of film electrode prepared with CNT as conductive additive at 25k magnification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

71 72

73

74 75 76 78 79 80 81

82

83 85 86 87

88 89

89 90

142

5.8

5.9 5.10 5.11

5.12

5.13

5.14

5.15

5.16 5.17 5.18

5.19

List of Figures

Comparative study of rate capabilities of film electrodes containing as additive carbon nanotubes (CNT), carbon black (CB) or an even mixture of both. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rate capability for progressively increased cycling rate for (a) CMC-water and (b) PVdF-NMP based film electrodes. . . . . . . . . . . . . . . . . . . . XPS of F1s of postmortem electrodes cycles using (a) PVdF and (b) CMC as binder. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Galvanostatic cycling curve with corresponding derivative, which shows first two cycles in black and 20th cycle in red and (b) rate capability of TiO2 anatase from template aided sol-gel synthesis. Effects of chemical transformation of electrolyte are marked with green arrows: I irreversible capacity during first cycle, and II continuous irreversible lithium consumption upon consecutive cycling at slow rate C/20. . . . . . . . . . . . . . . . . . . . . Electrochemical cycling curve and corresponding derivative for TiO2 (B) rich sample cycled at C/20 with (a) limited potential window 1.2-2.2 V , and (b) wide potential window 1-2.5 V , indicating parasitic reactions feature with green arrow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Evolution of C1s XPS peak of film electrode using PVdF as binder and LiPF6 as lithium ion salt for (a) pristine (b) electrolyte soaked (c) post mortem electrode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Electrochemical cycling curve at C/20 and (b) capacities obtained for progressively increased cycling rate for TiO2 (B) rich sample with electrolyte containing 2%VC additive. . . . . . . . . . . . . . . . . . . . . . . . . . . . Rate capability for progressively increased cycling rates and corresponding derivative of representative TiO2 (B) rich samples cycled with (a) LiBF4 and (b) LiPF6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Post mortem inspection of electrodes cycled with EC:PC:3DMC 1M LiPF6 using (a) whatman glassfiber and (b) celgard polymer separator. . . . . . XPS spectra of post mortem electrodes cycled using LiPF6 (a,c) and LiBF4 (b,d). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (a) Schematic division into three transport process in electrodes and (b) dependency of electrode resistance on loading, the three curves illustrate the cases; F internal transport limited, S external transport limitation and M intermediate case transport. . . . . . . . . . . . . . . . . . . . . . . . . Electrode resistance as function of active mass loading of TiO2 (B) rich electrodes using carbon nanotubes (CNT) as additives cycled at C/20 with EC:PC:3DMC LiBF4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

90 91 92

95

96

97

98

99 100 101

102

103

B.1 (a) Electrochemical cycling curve revealing important cycling characteristics such as reversible capacity, irreversible capacity (IC) as well as polarization and (b) first derivative showing precise position of reaction potential and sum of exchanged charge. . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

List of Figures

B.2 Linear plot of Williamson-Hall equation, intercept gives particle size D and slope corresponds to microstrain ε. . . . . . . . . . . . . . . . . . . . . . . B.3 Full pattern matching of TiO2 (B) phase before (left) and after (right) lithiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.4 Classification of isotherms according to IUPAC for: I monolayer adsorption, II first monolayer adsorption until point B then multilayer adsorption, III weak interaction between adsorbent and adsorbate. Type I,II, and III show adsorption behavior for microporous adsorbent materials, whereas type IV and V show adsorption behavior of mesoporous sorbent materials, resulting in a hysteresis. Type VI shows the cascading adsorption, which is typical for a multi-modal pore size distribution. Figure taken from [237]. . . . . . B.5 Generation of photoelectron via high energy X-ray irradiation. . . . . . . B.6 X-ray absorption intensity as function of energy for titanium atom in TiO2 , indicating main regions of spectra; pre-edge, XANES and EXAFS. . . . . .

143

152 152

154 156 156

List of Tables 1.1

Physical and electrochemical properties of TiO2 , data taken from [28] . .

11

2.1

Summary of sol-gel synthesized samples . . . . . . . . . . . . . . . . . . .

31

3.1

Morphological and electrochemical properties of sol-gel synthesized samples 44

4.1 4.2

EXAFS

Fitting Parameters for non-doped and 10% Nb-doped anatase . . . Hydrothermal precursor solution of cation dopant . . . . . . . . . . . . . .

74 77

5.1

Properties of alkyl carbonates at 25◦ C commonly used as electrolyte in LIB, extract from [223] and [224] . . . . . . . . . . . . . . . . . . . . . . . . . . Properties of solutes for LIB [224] . . . . . . . . . . . . . . . . . . . . . . .

93 94

5.2

B.1 Evolution of fitting parameters of oxygen shells upon lithiation of TiO2 (B) 158 B.2 Evolution of fitting parameters of titanium shells upon lithiation of TiO2 (B)159

145

A Abbreviations AAS BET CB CMC CNT CV DOS EC DFT DMC EDX EXAFS GGA HRTEM ICSD IR IUPAC LIB NMP NMR PC PBE PVdF SEI SEM TEM UV-vis VASP VC XANES XAS XPS XRD XRF

Atomic absorption spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Brunauer-Emmett-Teller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Carbon black . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 Carboxymethyl cellulose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Carbon nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 Cyclic voltammetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Density of states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Ethylene carbonate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Density functional theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Dimethyl carbonate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 Energy dispersive X-ray spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Extended X-ray absorption fine structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Generalized gradient approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 High resolution transmission electron microscopy. . . . . . . . . . . . . . . . . . . . . . . . . . .19 Inorganic crystal structure database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Infra-red . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 International union of pure and applied chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Lithium ion battery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 N-Methyl-2-pyrrolidone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Nuclear magnetic resonance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 Propylene carbonate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Perdew-Burke-Ernzerhof functional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Polyvinylidene fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Solid electrolyte interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Scanning electron microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Transmission electron microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Ultraviolet-visible . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 Vienna Ab initio simulation package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 Vinylene carbonate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 X-ray absorption near edge structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 X-ray absorption spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 X-ray photon spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 X-ray diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 X-ray fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

147

B Characterization methods B.1 Electrochemical cycling All electrochemical cycling measurements are carried in out in so called “half-cell” configuration, in which the titanium oxide acts as a cathode (positive electrode) with lithium metal as the counter negative electrode (anode). Hence, in this thesis the insertion into the TiO2 is termed as the discharge reaction, whereas the Li+ extraction from the oxide is called as the charge reaction.

B.1.1 Galvanostatic cycling Is a cycling technique, in which a constant current is applied to an electrode and the response of its potential is recorded. As the current is equal to flux of charge, this technique allows to obtain the electrochemical cycling curve, in which potential as function of amount of inserted lithium are plotted. However, this practice assumes that all the applied charge is used for the insertion reaction. This graph is also called the electrochemical signature since its shape and extension reflects characteristic insertion chemistry, which is material specific. It allows to extract important characteristic such as reversible capacity, nature of insertion mechanism, capacity loss, polarization etc., see Fig. B.1(a).

Figure B.1: (a) Electrochemical cycling curve revealing important cycling characteristics such as reversible capacity, irreversible capacity (IC) as well as polarization and (b) first derivative showing precise position of reaction potential and sum of exchanged charge.

Furthermore, by doing the first derivative δx/δU and plotting it versus potential, one

149

150

B Characterization methods

obtains curve shown in B.1(b), which allows a precise analysis of potential position and nature of desired and parasitic chemical reactions taking place. This is especially useful for materials that do not feature a well defined reaction plateau, such as the here shown example of TiO2 (B). For all electrochemical measurements, the cycling rate 1C corresponds to the insertion of 1 lithium per TiO2 , which is equivalent to a current of 336 mA.

B.1.2 Cyclic voltammetry Is a potentiostatic cycling technique, in which a defined potential window is traversed stepwise while current is being measured. The current is linked to the number of electrons exchanged during chemical reaction. Therefore cyclic voltammetry (CV) permits one to gain qualitative and quantitative information about electrochemical reactions. In our measurements, we used it to localize redox potential of internal processes, quantify them and study their evolution over several cycles.

B.2 Atomic absorption spectroscopy Atomic absorption spectroscopy (AAS) is analytical measure for facile and fast qualitative and quantification detection of many elements. It uses element specific absorption lines, resembling the transition of electrons to higher energy levels, for elemental identification. Quantitative measures are carried via calibrating a working curve using reference samples of known concentration.

B.3 X-ray photoelectron spectroscopy X-Ray photoelectron spectroscopy (XPS) is widely used technique, based on Einstein’s photoelectric effect, for examining surface chemistry of samples (≤5 nm). The center of interest are elemental composition and information on chemical state, but also chemical imaging and depth profile can be obtained. During the measurement, which have to be carried out under ultra high vacuum (UHV ≤10−9 mbar), the sample is irradiated with monochromatic X-rays. When these photon hit core level electron, which have element specific binding energies, their energy is transfered and leads to the emission of the core electrons from their initial state. Depending on the energy of the incident photons and the binding energy of the electron, it bears different kinetic energy when emitted. In a detector the energy and intensity of emitted photoelectrons are analyzed. Since the binding energy of the electron is specific for each element the nature but also concentration can be determined. Post mortem electrode samples were recovered in glovebox under Ar atmosphere, carefully washed in DMC, sealed air tight and sent to IPREM Pau, France where they were opened and examined without being exposed to ambient air.

B.4 X-ray diffraction

151

B.4 X-ray diffraction X-ray diffraction (XRD) is widely-used method for non-destructive structural analysis. It serves to examine the crystal structure and long-range order of condensed matter. The interference pattern, which results from the atomic planes of the specimen, is specific to each substance (fingerprint) and basis of the diffraction analysis. Therefore, XRD in combination with a catalogue of diffraction patterns is also commonly used as a chemical characterization method. Physical basis of this technique is Bragg equation; n · λ = 2d · sin(Θ)

(B.1)

whereas λ is the wavelength of X-rays, Θ the angle between incident beam and scattering planes, d the lattice spacing and n an integer. In this thesis a Philips X’Pert diffractometer was used.

B.4.1 Crystallite size determination X’Pert Highscore software by Panalytical was used throughout the thesis to analyze diffraction patterns. This software has a build in Williamson-Hall approach for line profile analysis from which information on crystal size and microstrain can be deduced based on width and shape of peak profile. An instrumental broadening correction is performed before the analysis. The facile Williamson-Hall approach was used to derive crystallite sizes from the XRD patterns. It was introduced by G.K.Williamson and his student, W.H.Hall in 1953 [238], and is based on the fact that a finite crystal size leads to a line broadening. Furthermore, it takes into account micro strain effects due to, e.g., lattice distortions or surface relaxation, which can contribute significantly to line broadening, especially for nanoscaled materials and materials with preferential crystallographic orientation growth. In this the WilliamsonHall approach differs from Scherrer equation which only takes into account crystallite size effect. In Eq. B.2 the two contribution to the line broadening are shown. From a linear plot the particle size D can be deduced from the intercept while the slope corresponds to the strain ε, see example in Fig. B.2. β = 1/D + 2 · ε · Q

(B.2)

We found that TiO2 samples with different phase composition but similar morphology have comparable strain values. Therefore throughout the thesis when comparing samples with similar morphology we limit our analysis to the comparison of grain size D.

152


Figure B.2: Linear plot of Williamson-Hall equation, intercept gives particle size D and slope corresponds to microstrain ε.

B.4.2 Lattice parameter determination To extract information on lattice parameter evolution upon lithiation we used the software PowderCell which allows a simple peak matching approach via adequate adjustment of lattice parameter a, b and, c. This method is assumes a sheer expansion of the lattice and approximates that that no structural rearrangement or angle change occurs. In order to obtain more precise information on lattice parameter full pattern matching with FullProf was performed on accumulated patterns before and after lithiation, see Fig. B.3 left and right, respectively. It is salient that quality of the fit is still limited, which is linked to the low crystallinity of the TiO2 (B) phase in combiantion with the strong Be peak at 46.5◦ as well as the presence of anatase phase.

Figure B.3: Full pattern matching of TiO2 (B) phase before (left) and after (right) lithiation.

B.5 Raman

153

B.5 Raman Raman spectroscopy exploits the share of inelastically scattered photons, resulting from the high energy, monochromatic irradiation of an atom or molecule. While most of the incident radiation is elastically scattered Rayleigh scattering a small fraction is scattered inelastically. The latter is the case, when after the excitation via absorption of a photon, the atom or molecule relaxes to a higher state (Stokes scattering) or to a lower state (AntiStokes scattering) than its initial ground state. The energy shift compared to elastically scattered photons of these Stoke-lines or Anti-Stoke-lines is specific to vibrational mode and binding couple, it can therefore be used to reliably identify molecules and compounds. The minimal preparation, the short measuring time and the non-destructiveness make it a versatile technique for progressional quality control. The ability to perform local and general analysis by modulating beam diameter, as well as the information on local structure that can be gained make Raman spectroscopy even more valuable.

B.6 Surface Area and Porosity Analysis By measuring the isotherms of N2 adsorption and desorption with a Micromeritics ASAP 2020 instrument, porous structures can be characterized and parameters such as specific surface area and differential pore volumes (or pore widths) can be determined. The measurements are performed by dosing well defined amounts of nitrogen gas into an evacuated (pinitial ≈10−4 mbar) and cooled (T = 77 K) sample tube, which contains a precisely determined amount of the porous specimen. At these low temperatures, the nitrogen molecules are physisorbed. In Figure B.4, a classification of different typical physisorption profiles is shown. According to IUPAC curve I shows simple monolayer adsorption, while II reveals first monolayer adsorption until point B, then multilayer adsorption takes place. In physisorption curve III, weak interaction between adsorbent and adsorbate occurs. While curves of type I,II, and III show adsorption behavior for microporous adsorbent materials, type IV and V show adsorption behavior of mesoporous sorbent materials, resulting in a characteristic hysteresis. Type VI shows the cascading adsorption, which is typical for a multi-modal pore size distribution. Among the most important material parameters, which can be determined from these isotherms are the cumulative pore volume, the specific surface area and the pore size distribution. The latter parameter enables one to classify porous material according to the IUPAC notation. For porous materials three groups are defined: microporous materials with pore sizes < 2 nm, mesoporous materials with an average pore size between 2 to 50 nm, and macroporous materials, which show pore sizes > 50 nm. Whereas pore size distribution and cumulative volume are calculated based on the Density-Functional-Theory (DFT) [239], the specific surface area is obtained by calculations based on the Brunauer-Emmett-Teller (BET) theory.

154


Figure B.4: Classification of isotherms according to IUPAC for: I monolayer adsorption, II first monolayer adsorption until point B then multilayer adsorption, III weak interaction between adsorbent and adsorbate. Type I,II, and III show adsorption behavior for microporous adsorbent materials, whereas type IV and V show adsorption behavior of mesoporous sorbent materials, resulting in a hysteresis. Type VI shows the cascading adsorption, which is typical for a multi-modal pore size distribution. Figure taken from [237].

Both theories are limited in their application, and include approximations and assumptions which may not reflect the reality to the whole extent. The determination of pore size distribution and cumulative volume via DFT is based on calculated density profiles, which describe the molecular density vs. distance from the wall. These density profiles are calculated for a specific adsorbate gas, in combination with a specific adsorbent scaffold, i.e. N2 gas with carbon as porous scaffold in general. Therefore, they do not consider that different combinations of materials might lead to different interactions between adsorbate and adsorbent, which would result in a change of density profiles. The BET theory was originally designed for non-porous materials with a type II isotherm. Furthermore, it is assumed that all pores are slit pores in carbon matrices due to the layer structure of the carbon source material. Therefore, absolute values determined by these theories must be considered with caution. However, they reflect the general structural character of a sample and can well be used for comparative measurements.

B.7 Transmission electron microscopy Transmission electron microscopy (TEM) is a widely used technique, in which electrons that are transmitted through a thin specimen are used as imaging species. It allows to examine structures and morphology on an atomic scale. Furthermore, the option of performing a selected area electron diffraction (SAED) allows to acquire information on local crystalline

B.8 Scanning electron microscopy, Energy dispersive X-ray scattering

155

structure.

B.8 Scanning electron microscopy, Energy dispersive X-ray scattering By irradiating a sample with electron beam, multiple signals are produced that can be exploited, using scanning electron microscopy SEM. While secondary electrons (SE), due to their low penetration depth, carry information of topography, back scattered electrons (BSE) are used for analytical purposes. The characteristic X-rays can be used for quantitative analysis. The ability to raster the sample in the x-y plane, while measuring characteristic X-rays results in a space-resolved elemental distribution map. The technique is often referred to as energy dispersive X-ray spectroscopy (EDS or EDX) when used in mapping configuration. In summary, this technique bears great potential in verifying homogeneity throughout the sample, delivering 3D topography images, as well as impurity detection.

B.9 X-ray fluorescence In X-ray fluorescence (XRF), materials are excited with high-energy, short wavelength radiation (e.g., X-rays). If the energy of the radiation is sufficient to dislodge a tightly-held inner electron, the atom becomes unstable and an outer electron replaces the missing inner electron. Consequently energy is released due to the disparity in binding energy of the inner electron orbital compared with an outer one. Since the radiation of the emitted has lower energy than the primary incident X-rays, it is termed fluorescent radiation. The energy of the emitted photon is characteristic for a specific transition between electron orbitals in a particular element, therefore, the resulting fluorescent X-rays can be used to detect the abundances of elements that are present in the sample. Throughout this thesis, a PANalytical Axios max was used, primarily for the sake of impurity determination.

B.10 X-Ray absorption spectroscopy X-Ray absorption (XAS) technique is based on the absorption of X-rays by the core shell electrons of an atom. If the irradiation energy (h ∗ ν) is higher than the binding energy of the electron, a photoelectron is created, see Fig. B.5. This phenomena is expressed by a drastic increase of the absorption intensity in the absorption spectra, shown in Fig. B.6, which is called the absorption edge. The edge is named accordingly to the the orbital of origin of the electron. The zone around the absorption edge is called X-ray absorption near edge structure or XANES. It includes the absorption edge, which lies for titanium at 4966 eV , followed by the so-called

156


Figure B.5: Generation of photoelectron via high energy X-ray irradiation.

white line 1 , as well as small pre-edge peaks. These derive from the filling of free d-orbitals, and are characteristic for transition metals such as titanium [240]. Shape and position of the edge are directly depending on the electronic structure, oxidation state, and local coordination of the absorbent atom.

Figure B.6: X-ray absorption intensity as function of energy for titanium atom in TiO2 , indicating main regions of spectra; pre-edge, XANES and EXAFS.

The third region is the Extended x-ray absorption fine structure or EXAFS, the signal deduces from the back-scattering of the photoelectrons by its neighbors. The photoelectron has both wave and particle character so that an interference of waves occurs. Depending

1

this feature in the XAS spectra reflects the s⇒p transition filling of the first free p orbitals

B.10 X-Ray absorption spectroscopy

157

on distance and shift in phase the waves either enhance or destroy each other. This can be observed in the wiggles of curve carrying the fine structure information of the material.

B.10.1 Experimental details measurements were carried out at ambient temperature on the beamlines, A (HASYLAB @ DESY, Hamburg, Germany), and XAFS (ELETTRA, Trieste, Italy). For the non-operando measurements the Ti K-edge spectra were recorded in transmission mode on homogeneous pressed pellets of adequate thickness. Fourier transformations were performed using k2 weighting and the structural parameters were determined by curve-fitting procedures using Artemis data analysis software with embedded FEFF tool for calculating scattering paths on the basis of the anatase structure. To fit Nb-doped anatase in the R-space, FEFF paths based on Nb2 O5 crystal structure, in which the central Nb atom was exchanged for Ti, were added in order to include Ti-Nb bonds in the calculations. For the operando measurements self supported electrode films with increased binder concentration were assembled in the stainless steel in situ cell, presented in Fig. 3.10. XAS

B.10.2 Fitting parameters The parameters of position and Debye Waller factor derived from fitting with Artemis software of the first two oxygen shells (5-1 repartition) and the first three Ti shells (4-2-2 repartition) are presented in Tab. B.1 and Tab. B.2, respectively.


158

Table B.1: Evolution of fitting parameters of oxygen shells upon lithiation of TiO2 (B)

Lix TiO2 (B) 0 0.04 0.11 0.15 0.19 0.23 0.26 0.30 0.34 0.37 0.41 0.45 0.52 0.56 0.60 0.63

Oxygen shells R Ti-O5 [Å] R Ti-O1 [Å] σ*10−3 [Å2 ] 1.938±0.007 2.388±0.064 8.3±1.0 1.934±0.006 2.403±0.051 8.7±0.9 1.930±0.006 2.361±0.048 8.1±0.9 1.944±0.006 2.399±0.048 9.3±0.9 1.947±0.005 2.387±0.044 11.1±0.9 1.955±0.005 2.377±0.044 11.1±0.9 1.957±0.005 2.389±0.047 11.7±1.0 1.959±0.005 2.399±0.048 11.9±1.0 1.966±0.005 2.415±0.045 11.9±0.9 1.972±0.004 2.417±0.040 11.8±0.8 1.977±0.005 2.425±0.042 12.0±0.8 1.976±0.005 2.435±0.044 11.7±0.8 1.990±0.005 2.512±0.049 12.1±0.8 1.987±0.005 2.510±0.045 10.8±0.7 1.990±0.005 2.540±0.045 10.9±0.7 2.011±0.005 2.581±0.049 12.1±0.8

R-factor 0.015 0.016 0.020 0.016 0.012 0.013 0.015 0.014 0.012 0.009 0.010 0.011 0.012 0.010 0.010 0.010

0 0.04 0.11 0.15 0.19 0.23 0.26 0.30 0.34 0.37 0.41 0.45 0.52 0.56 0.60 0.63

Lix TiO2 (B)

R Ti-Ti3 [Å] 3.080±0.016 3.082±0.012 3.090±0.015 3.090±0.014 3.106±0.011 3.118±0.012 3.125±0.014 3.123±0.014 3.131±0.013 3.140±0.012 3.147±0.012 3.146±0.014 3.140±0.015 3.153±0.016 3.153±0.016 3.156±0.013

σ*10−3 [Å2 ] 6.1±1.8 5.3±1.3 6.5±1.7 6.9±1.6 6.7±1.4 7.1±1.5 8.0±1.8 8.1±1.8 8.4±1.7 8.7±1.5 8.7±1.4 9.3±1.6 10.0±1.7 11.7±2.1 11.8±2.1 10.7±1.6

Titanium shells R Ti-Ti2 [Å] σ*10−3 [Å2 ] 3.322±0.031 6.3±3.7 3.328±0.021 4.8±2.4 3.335±0.024 4.6±2.5 3.329±0.027 7.5±3.2 3.335±0.020 6.0±2.3 3.340±0.021 6.5±2.6 3.335±0.030 8.9±3.7 3.335±0.030 9.63±3.9 3.370±0.033 11.29±4.3 3.392±0.034 12.8±4.4 3.414±0.046 13.5±4.6 3.433±0.049 15.9±6.3 3.460±0.051 16.5±7.1 3.504±0.051 17.2±8.1 3.495±0.039 14.2±5.7 3.504±0.046 17.9±7.2 R Ti-Ti2 [Å] 3.859±0.045 3.875±0.025 3.857±0.038 3.895±0.033 3.889±0.028 3.910±0.029 3.910±0.034 3.913±0.029 3.918±0.025 3.939±0.024 3.944±0.022 3.941±0.023 3.951±0.026 3.981±0.025 3.988±0.029 4.006±0.023

σ*10−3 [Å2 ] 9.6±6.2 6.7±3.5 9.3±5.4 9.8±4.8 10.1±4.1 10.6±4.3 11.0±4.8 9.9±4.1 8.7±3.0 9.9±3.1 8.7±3.0 8.2±3.1 8.9±3.5 8.5±3.4 10.2±4.1 9.3±3.2

Table B.2: Evolution of fitting parameters of titanium shells upon lithiation of TiO2 (B)

0.015 0.016 0.020 0.016 0.012 0.013 0.015 0.014 0.012 0.009 0.010 0.011 0.012 0.010 0.010 0.010

R-factor

B.10 X-Ray absorption spectroscopy 159

C Computational methods C.1 Calculation of electronic band structure of Nb-doped anatase Electronic structure calculations based on DFT and a generalized gradient approximation (GGA1 ) using an exchange-correlation potential by Perdew, Burke and Ernzerhof (PBE2 ) [241] were performed for anatase TiO2 , and three supercells of 24 atoms (NbTi7 O16 ), 48 atoms (NbTi15 O32 ) and 108 atoms (NbTi35 O72 ). The latter correspond to a Nb doping of about 12 %, 6 % and 3 %, respectively. The augmented plane-wave method + local orbital (APW+lo) as implemented in the WIEN2k code [242] was used with the muffin-tin radii Rmt(Ti)= 2.0 a.u., Rmt(O)= 1.4 a.u. and Rmt(Nb)=2.0 a.u., the semicore states: Ti 3s, Ti 3p, O 2s, Nb 4s and Nb 4p, the plane-wave cutoff: min(Rmt).max(K)=7 (where K is a reciprocal lattice vector) and the magnitude of the largest vector in the charge-density fourier expansion: Gmax=15 Ry −1/2 . The experimental values of the lattice constants were used for the calculations, but the internal atomic positions were moved in order to minimize the internal atomic forces. A similar approach was successfully used for the electronic structure calculations of Sn- [243] and Zr-doped [244] TiO2 anatase. Self-consistency was achieved with an energy tolerance of 10−4 Ry and a force tolerance of 10−3 Ry Bohr−1 . The XAS spectra were calculated from the dipole transition strengths and the partial density of states (PDOS) in the same way as previous calculations for Li4 Ti5 O12 [245]. The calculated Ti K-edge spectra were averaged over all the Ti atoms of the supercells. For comparison with experiments the calculated spectra were convoluted with a Lorentzian function (FWHM=2 eV ).

C.2 Calculation on lattice structure and lithium insertion sites of TiO2 (B) Calculations were performed within the framework of DFT using VASP [246]. Atoms are described by plane waves within the PAW (Projector augmented-wave) formalism using pseudopotentials [247, 248]. We used the GGA of PBE [241] for the exchange and correlation potential, since the hybrid B3LYP functional is not yet available for this code [249]. The convergence of the calculations was checked with respect to both, the energy cutoff (up to 800 eV ) and the k-points grid (up to a 2x6x4 Monkhorst-Pack mesh) [250], leading 1 2

Generalized gradient approximation Perdew-Burke-Ernzerhof functional

161

162

C Computational methods

to reliable converged structures. Calculations regarding the energetics of insertion sites were conducted using DFT and Crystal06 [251]. In comparison to VASP, Gaussian-type basis are used to describe the atomic orbitals. Every atomic orbital is represented by a combination of several Gaussian functions. While for core orbitals a single set is sufficient, valence orbitals require 2-3 sets of Gaussian functions.

Abstract Titania based electrode materials are promising candidates to replace widely used graphite as negative electrode material in lithium ion batteries (LIB), due to their increased safety, volumetric capacity, and high rate performance. In this thesis different low-cost synthesis approaches are evaluated to prepare nanostructured TiO2 with various phase composition and morphology. The influence of these parameters on its ability to reversibly insert lithium are studied in electrochemical measurements. In this regard we also investigated the effect of aliovalent doping and porous structures on the insertion properties of two main polymorphs of TiO2 , Anatase and TiO2 (B), revealing encouraging results in overcoming the low charge transfer, which is the main drawback of titanium oxide based materials. In order to understand the mechanism of lithium storage process of the two synthesized TiO2 phases, diffraction and spectroscopic characterization methods were carried out under operando conditions. We show that, regardless of their chemical similarity, both phases reveal very different lithium insertion processes, leading to distinct electrochemical cycling properties. Another field of interest is the adaptation of electrode components to the nanostructured TiO2 active insertion material. The choice of binder, carbon additive, and electrolyte components can have significant impact on the performance. Especially the origin and prevention of parasitic side reactions were in the focus of our work, as these pose an under estimated hindrance in the application of titania based electrode materials in LIB.

Résumé Les matériaux à base de dioxyde de titane (TiO2 ) sont des candidats prometteurs pour remplacer le graphite utilisé actuellement dans les électrodes négatives des batteries lithium-ion (LIB), du fait de leur sécurité élevée, de leur capacité volumétrique supérieure et de leurs excellentes performances à haute puissance. Dans cette thèse, différentes approches de synthèse à bas coût sont évaluées pour préparer du TiO2 nanostructuré avec différentes compositions de phase et des morphologies variées. L’influence de ces paramètres sur la capacité de TiO2 à insérer réversiblement le lithium est étudiée par des mesures électrochimiques. À cet égard nous avons également étudié l’effet du dopage aliovalent et de la morphologie poreuse sur les propriétés d’insertion du TiO2 , révélant des résultats encourageants avec notamment un transfert de charge amelioré, principale limitation des matériaux à base d’oxyde de titane. Afin de comprendre le processus de stockage du lithium des deux phases de TiO2 synthétisées, des méthodes de diffraction et de caractérisation spectroscopique ont été utilisées dans des conditions operando. Nous montrons qu’indépendamment de leur similitude de composition chimique, les deux phases révèlent des mécanismes d’insertion du lithium très différents, menant à des propriétés électrochimiques de chargedécharge très différentes. Nous avons également amélioré les performances electrochimiques en travaillant sur la formulation d’électrodes à base de TiO2 nanostructuré, en optimisant le choix des composants (additif carboné, liant, électrolyte) et le processus de préparation. De nombreuses réactions parasites électrode-électrolyte ont été mises en évidence à travers cette étude, phénomènes très peu décrits dans la littérature à ce jour.