Master thesis - Nanologica AB

Towards Mesoporous Titania Microspheres bySupramolecular Self-assembly

Master’s Thesis in Engineering Nanoscience

Martin A. Olsson

Lund 2009

Department of Materials Chemistry Lunds tekniska högskolaLunds universitet Lunds universitetSE-581 83 Lund, Sweden 581 83 Lund

Towards Mesoporous Titania Microspheres bySupramolecular Self-assembly

Master’s Degree Project in Materials Chemistryat the Ångström Laboratory

Martin A. Olsson

Supervisors: Dr. Alfonso E. Garcia-BennettNanotechnology and Functional Materials,dept. Engineering Sciences, Uppsala University.Nanologica AB, Uppsala Science Park

Prof. Reine WallenbergDept. Materials Chemistry, Lund Institute of Technology

Examiner: Prof. Staffan HansenDept. Materials Chemistry, Lund Institute of Technology

Lund, August 26th, 2009

Avdelning, InstitutionDivision, Department

Division for Chemical EngineeringDepartment of Materials ChemistryLund UniversitySE-581 83 Lund, Sweden

DatumDate

2009-08-26

SpråkLanguage

2 Svenska/Swedish2 Engelska/English

2

�

RapporttypReport category

2 Licentiatavhandling2 Examensarbete2 C-uppsats2 D-uppsats2 Övrig rapport2

�

URL för elektronisk version

http://theses.lub.lu.se/undergrad/

ISBN—

ISRN

Serietitel och serienummerTitle of series, numbering

ISSN—

TitelTitle

Porös Titandioxid genom Supramolekylär SjälvorganiseringTowards Mesoporous Titania Microspheres by Supramolecular Self-assembly

FörfattareAuthor

Martin A. Olsson

SammanfattningAbstract

The increasing demand for electricity with a larger population of our planet forcessolutions to ensure a future energy source for humanity. The climate change de-mands that this energy source is renewable. Such an energy source is the sunlightthat hits our planet every day, which exceeds the present world consumption ofenergy per year. Dye-sensitized solar cells (DSSCs) are attractive light-harvestingdevices since their design is suitable for large-scale mass production comparedto all other present solar cell devices. DSSCs would be highly suitable as roof-top solar collectors if their conversion efficiency was increased. However, there isroom for improvement of the power conversion efficiency of dye-sensitized solarcells. The high accessible surface area to sensitize titania is a viable route to im-provement in the conversion efficiency of DSSCs. Employing the sol-gel methodof producing titanium dioxide causes an inexpensive method for producing nanos-tructured titania nanomaterials with high surface area. The results in this thesisshow that the sol-gel method with supramolecular self-assembly templating is aplausible approach of generating high surface area TiO2, since there is an evidenttemplating effect in titania. Decreasing the pore size has been investigated forthe self-assembly of folate molecules in solution to achieve mesoporous titania bylower reaction rates. The results from this thesis show that formation of micro-spheres of titania by spray-drying increases the surface area by a factor of two. Inthis thesis the titania does not reach the goals for application in dye-sensitizedsolar cells due to constraints on a pH > 7 which causes too high reaction ratesand low crystallinity. The main conclusion is that acidic conditions for titaniasynthesis is essential to the anatase crystallization and therefore the methods em-ployed except chelation will improve the titanium dioxide and hence the efficiencyof dye-sensitized solar cells. Materials prepared towards the end of this thesis,show that the relationship between surface area, crystallinity and dye adsorptioncapacity is not evident. High dye adsorption capacities were achieved by materialsprepared by an atrane route which shows 28% of the dye adsorption capacity ofthe commercial titania P-25.

NyckelordKeywords Sol-Gel, Mesoporous, Titania, Chelation, Atrane, Acetylacetone, Folic acid, DSSC

AbstractThe increasing demand for electricity with a larger population of our planet forcessolutions to ensure a future energy source for humanity. The climate change demandsthat this energy source is renewable. Such an energy source is the sunlight that hitsour planet every day, which exceeds the present world consumption of energy peryear. Dye-sensitized solar cells (DSSCs) are attractive light-harvesting devices sincetheir design is suitable for large-scale mass production compared to all other presentsolar cell devices. DSSCs would be highly suitable as roof-top solar collectors iftheir conversion efficiency was increased. However, there is room for improvementof the power conversion efficiency of dye-sensitized solar cells. The high accessiblesurface area to sensitize titania is a viable route to improvement in the conversionefficiency of DSSCs. Employing the sol-gel method of producing titanium dioxidecauses an inexpensive method for producing nanostructured titania nanomaterialswith high surface area. The results in this thesis show that the sol-gel method withsupramolecular self-assembly templating is a plausible approach of generating highsurface area TiO2, since there is an evident templating effect in titania. Decreasingthe pore size has been investigated for the self-assembly of folate molecules insolution to achieve mesoporous titania by lower reaction rates. The results from thisthesis show that formation of microspheres of titania by spray-drying increases thesurface area by a factor of two. In this thesis the titania does not reach the goalsfor application in dye-sensitized solar cells due to constraints on a pH > 7 whichcauses too high reaction rates and low crystallinity. The main conclusion is thatacidic conditions for titania synthesis is essential to the anatase crystallization andtherefore the methods employed except chelation will improve the titanium dioxideand hence the efficiency of dye-sensitized solar cells. Materials prepared towardsthe end of this thesis, show that the relationship between surface area, crystallinityand dye adsorption capacity is not evident. High dye adsorption capacities wereachieved by materials prepared by an atrane route which shows 28% of the dyeadsorption capacity of the commercial titania P-25.

v

Acknowledgments

This master’s thesis project has been financed by Nanologica AB. I want to thankDr. Alfonso Garcia-Bennett in particular for giving me this opportunity to work ina state of the art research environment in the field of mesoporous materials and forfor his advice. I want to thank Prof. Reine Wallenberg for his supervision. I wantto acknowledge the generous advice for UV-vis spectroscopy by the Nanologicaemployee Dr. Roberto Hanoi. I want to thank Ph.D.-student Rambabu Atlurifor sharing his office with me at the division for Nanotechnology and FunctionalMaterials at the Ångström Laboratory.

vii

Contents

1 Introduction 31.1 Aims of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.2 Related work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Background 52.1 Dye-Sensitized Solar Cells . . . . . . . . . . . . . . . . . . . . . . . . 52.2 Crystal structure of titania . . . . . . . . . . . . . . . . . . . . . . . 62.3 Sol-gel chemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.3.1 The effect of synthesis parameters . . . . . . . . . . . . . . . 102.4 Mesoporosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.5 Self-assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.6 Hydrolysis retardation . . . . . . . . . . . . . . . . . . . . . . . . . . 152.7 Industrial drying of powders . . . . . . . . . . . . . . . . . . . . . . . 152.8 Characterization techniques . . . . . . . . . . . . . . . . . . . . . . . 17

2.8.1 Scanning electron microscopy . . . . . . . . . . . . . . . . . . 172.8.2 X-ray diffraction (XRD) . . . . . . . . . . . . . . . . . . . . . 182.8.3 N2−sorption . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.8.4 Thermogravimetric measurements . . . . . . . . . . . . . . . 192.8.5 Ultraviolet-visible spectroscopy . . . . . . . . . . . . . . . . . 19

2.9 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3 Experimental 233.1 Synthesis of TiO2 by folate route (NFM-2) . . . . . . . . . . . . . . 233.2 Slowing down the hydrolysis rate . . . . . . . . . . . . . . . . . . . . 23

3.2.1 Acetylacetone route (NFM-2-AC) . . . . . . . . . . . . . . . . 233.2.2 Atrane route (NFM-2-AT) . . . . . . . . . . . . . . . . . . . . 24

3.3 Calcination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.4 Spray-drying (NFM-2-SD) . . . . . . . . . . . . . . . . . . . . . . . . 253.5 Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.5.1 XRD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.5.2 SEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.5.3 N2-sorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273.5.4 Thermogravimetric measurements . . . . . . . . . . . . . . . 273.5.5 UV-vis absorbance . . . . . . . . . . . . . . . . . . . . . . . . 27

ix

x Contents

4 Results and discussion 294.1 Designed materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5 Conclusions 39

6 Future work 41

List of AbbreviationsDSSC : Dye Sensitized Solar CellSEM : Scanning Electron MicroscopyFA : Folic acidAPES : (3-aminopropyl)triethoxysilaneTTIP : Titanium(IV) isopropoxideCSDA : Co-Structuring Directing AgentFWHM : Full Width at Half MaximumXEDS : X-ray energy dispersive spectrometryTG : Thermogravimetric measurementsTEAH3 : triethanolamine (2,2’,2”-nitriloethanol)NFM-2 : Titania prepared by folate routeTG : Thermogravimetric measurementSE : Secondary electron

Material-suffixAT : Atrane routeAC : Acetylacetone routeSD : Spray-drying route

Chapter 1

Introduction

There has been a long search for a low-cost solar cell since the early 90s [1–5] thatcan replace and compete with solid state solar cell devices based on silicon. Plantcells and bacteria uses photosynthesis to harvest light energy. Using light-harvestingmolecules for making solar cells was firstly developed by Michael Grätzel et al. [1].A DSSC uses light-harvesting molecules to make a semiconductor sensitive towardslight i.e. dye-sensitization of the semiconductor. Titanium dioxide is one of the moststudied materials for DSSCs as the sensitized semiconductor. [3] The high amountof research on titania as a suitable electrode material is due to its wide band gapand electron transfer properties of one of its crystal structures, namely anatase [6].A detailed description of a DSSC will be given in the following chapter. DSSCs hasthe advantage of low-cost processing compared to other solar cell devices [7]. Thedemand for electrical grade silicon for the electronics industry is undoubtedly high.The largest cost of the conventional pn-junction solar cells comes from processing.For the DSSCs, the largest part of the cost comes from the dyes which roughly costsabout $750 per gram. The drawback of DSSCs is that the manufacturing processhas not yet really been improved much since the original design. The highest powerconversion efficiency that has been achieved is 11% for non-transparent DSSCs. [8]The conversion efficiency is determined by several parameters e.g. the thickness ofthe titania electrode [9], surface area of the titania [2] and the dye absorbance spectraamongst others [10], the surface morphology [11] and the amount of inter-particleconnections [7] that causes increased conductivity. Synthesis of titania is achievedat low temperature with weak dissociative acids or bases. Sol-gel chemistry occursvia nucleophilic attack during acidic, neutral and basic [12] conditions for salts andtransition metal alkoxides. [13;14] Titania has been recognized for its potential as anelectrode material due to its photoinduced electronic transfer properties associatedwith the anatase phase. [1] Other applications of the anatase are in catalysis [15;16],lithium-ion batteries [17] and cosmetics [18]. Therefore the titania has been intenselystudied the past years for applications such as: DSSCs, self-cleaning coatings [19],antibacterial tooth implants [19], sensors [20] and depolluting layers [20]. This master’sthesis is concerned with the design of titania for DSSCs with the use of folic acid asnanoscale pore forming agents. Previously, mesoporous titania has been prepared

3

4 Introduction

for DSSCs giving mixed results in efficiency in similar cell designs e.g. mesoporoustitania by surfactant routes has given 5% [10] and 8% [21]. When mesoporous titania issintered during processing the mesopores are blocked and the surface area decreases.This thesis will describe a novel route to slow down hydrolysis under basic conditionsfor titanium alkoxides. The thesis elaborates upon titania materials obtained duringthis thesis work. The thesis starts off in the application of titanium dioxide inDSSCs as a motivation for an in-depth sol-gel chemistry review. This is followed bya brief general extension of sol-gel chemistry to mesoporous materials. Self-assemblyand its application for synthesis of mesoporous materials is described. The literaturereview of the thesis covers mostly synthesis but also processing of the material.The characterization with the procedures used are presented as well, since some ofthe techniques are not conventional in nanoscience. Detailed procedures are notintended to be exhaustive but merely increase the reproducibility of the work.

1.1 Aims of the thesis• To investigate a novel self-assembly synthesis route for making a new titaniabased material.

• Investigation of the pore size control in titania by supramolecular self-assembly.

• To investigate an industrial drying processing method of the titania materialby laboratory bench-scale method for obtaining solid spherical microparticlesof titania for use as electrode material in DSSCs.

• To compare the dye adsorption of standard commercial available titaniapowder used for DSSCs (Degussa P-25) [22;23] and the synthesized titaniamaterial.

1.2 Related workThere are several scientific reports available on the sol-gel synthesis of mesoporoustitania. [19] However, since there are a lot of patents on the use of surfactant routesit has become difficult to commercialize those routes further [24;25]. Since sol-gelchemistry is a broad field with a lot of different precursors, a general treatment isnot the scope for a master’s thesis project. A full treatment of sol-gel chemistryis covered by Brinker (1990) [13]. Transition metal oxides are covered by Livageet al. (1988) [14]. Since the synthesis conditions are very important in the case ofsol-gel chemistry, two different materials are not often comparable with the samesynthesis conditions. The work in this thesis is based on the recent synthetic resultsof self-assembly templating by Garcia-Bennett et al. (2009). The following chapteris a brief review based on and around this literature.

Chapter 2

Background

2.1 Dye-Sensitized Solar CellsA DSSC is an electrochemical cell. The cell design is depicted in Figure 2.1(a). Theprinciple of a DSSC is a sandwich structure that consists of two transparent glasslayers each coated with an electrode material. The whole cell is denoted as F-dopedITO//nc-TiO2//N719//electrolyte//Pt//F-doped ITO, where nc-TiO2 is referredto as nanocrystalline titania. One of the electrodes is a titania layer and the otheris a transparent conducting oxide (TCO), usually Indium Tin Oxide, ITO. Thetitania electrode is soaked in a dye (Figure 2.1(b)) during processing causing thedye to chemisorb to the surface of the titania by carboxylate linkages. Between thetwo electrodes there is an electrolyte usually I – /I –3 which serves as a redox pair.The dye adsorbs UV-light and causes a charge separation [26] of electrons and holesin the ground and excited redox states. The electrons are injected to the titaniaconduction band through the carboxylate-titanium linkage. The injected electronsfrom the dye is replaced by electrons from the oxidation of the electrolyte redoxmolecule iodide. The electrons are transported through the titania network to aback contact and performs work in an external circuit. The electrons return to thecounter electrode and take place in the reduction of triiodide. There is a decreasingenergy path through the cell, hence it is thermodynamically possible. In general theefficiency reaches 10% when sintered nanocrystalline TiO2 is used. Novel designs forthe electrode material are TiO2-nanotubes [27], ZnO nanowires [28;29] with conversionefficiency 0.5-1.5 % [30] and microspheres of nanocrystalline ZnO [30] with conversionefficiency 5.4% [30]. The latter design will be further elaborated upon for titania inthis master’s thesis.

5

6 Background

Figure 2.1: (a) The principle of a dye-sensitized solar cell. (b) The most widely used dye forDSSCs, namely N719.

The oxidation and reduction reactions taking place in the DSSC are

Oxidation: 3 I− − 2 e− → I−3Reduction: I−3 + 2 e− → 3 I−

2.2 Crystal structure of titaniaTitania is a wide band gap semiconductor with three polymorphs [31]. The threecrystal structures are rutile, brookite and anatase. Since brookite is not relevant tothis master’s thesis it is excluded from this description. The rutile crystal structureis shown in 2.2(a)-(b) whereas the anatase crystal structure is depicted in Figures2.2(c)-(d). Both rutile and anatase are tetragonal minerals. Rutile can be viewed asremoving every other row of octahedra from an NiAs crystal structure. Comparedto the NiAs crystal structure which is ionic, the TiO2 is covalently bonded. Thestructure has a Ti:O coordination of 6:3 and is nearly close-packed. [32] The titaniumatom is coordinated by six oxygen in a slightly distorted octahedron. Viewing thestructure from the perspective of the oxygen, one oxygen atom is surrounded bythree planar titanium atoms at the corners of a triangle. The structure can alsobe viewed in terms of chains of octahedra sharing corners and opposite edges [32].Anatase is body centred. The crystal structure also has a glide plane along thec-axis in the ab-plane. The TiO6 octahedra are corner sharing in the plane in analternating pattern. The space group for rutile is the tetragonal P42/mnm whereP means a primitive lattice, 42 describes how the elements are screwed along afour-fold axis 90◦ with translation half of the lattice vector, the m and n indices aremirror plane and a n-glide plane , respectively. An n-glide plane is a translationhalf the lattice vector and mirroring of a diagonal of a face for the unit cell. Foranatase the space group is I41/amd, where I means a body centred lattice, 41

2.3 Sol-gel chemistry 7

means a screw axis where a rotational symmetry occurs by 90◦ with four-foldsymmetry and translation one fourth of the lattice vector. [6;32] The letters a andd are glide planes where a is a translation 1/2 of the lattice vector along only thea-axis and d is a glide plane along a fourth of a space diagonal or face of the unit cell.

Figure 2.2: (a) The rutile structure (b) The rutile structure in the [110]-projection. (c) Theanatase crystal structure viewed from the [001]-projection. (d) The anatase crystal structureviewed from the [100]-projection. The titanium atoms are illustrated with the larger spheres andthe polyhedra. Oxygen atoms are represented with the smaller spheres. [33;34]

2.3 Sol-gel chemistryPreparing titanium dioxide in the simplest possible manner by the sol-gel methodis straightforward and comprises adding a highly reactive titanium alkoxide towater, which causes a white precipitate. Templating titanium dioxide on thenanoscale by supramolecular self-assembly is however not as straightforward. Inorder to control the reactions occurring, a deeper understanding is appropriate.The two types of reactions that take place for sol-gel chemistry are hydrolysisand condensation. Polycondensation causes formation of colloids which readilypolymerize to precipitates in solution if the hydrolysis rate is too fast. [13] It is theinitial hydrolysis step that is the rate determining step in gel formation and thegelation rates is increased by the addition of spectator solvent. [35]Sol-gel chemistry isa subfield of inorganic and materials chemistry. Sol-gel chemistry has been recentlyre-visited due to the discovery of mesoporous materials. The sol-gel chemistry usesprecursor molecules that usually are in the form of alkoxides but halogen derivativesare also common but even more reactive [13]. Some typical precursor alkoxides aredepicted in Figure 2.3. Precipitation occurs by the same theory as proposed byFlory [13;14] on the general polymerization, extending the definition from organicpolymers to inorganic. Even if the inorganic framework is polymerizing it is not this

8 Background

fact that causes gelling, but the polymerization of colloids. The meaning of a colloidin the sense of sol-gel chemistry is strictly an oligomeric entity. The transition metalalkoxides are very reactive due to the electronegative alkoxide groups. This makesthe titanium alkoxides highly unstable to nucleophilic attack which is the mainreaction for the alkoxide chemistry. Therefore titanium dioxide readily precipitatesupon contact with merely water. [14] The general differences for silica precursorsand titania precursors can be summarized as:

• Silica has been the most studied material system, but the sol-gel chemistry ofsilica is somewhat different from titania. The titanium atom, in the precursoralkoxides, is more electropositive than the silicon atom. [13]

• Titanium alkoxides exhibit various stable coordinations. When the alkoxideis coordination unsaturated i.e. has a stable coordination different from theoxidation number, it is able to undergo coordination expansion by olation andalkoxy-bridging. [13]

Figure 2.3: (a) Tetraethylorthosilicate (b) Titanium(IV) ethoxide (c) Titanium(IV) isopropoxide.

The alkoxy group OR where R is alkyl or an aryl-group, is an electron donat-ing group and stabilizes the highest oxidation state of the titanium atom in themolecule. A silicon atom (14Si) has the orbital structure 1s22s22p63s23p2 whereasa titanium atom (22Ti) is described by an orbital structure 1s22s22p63s23p64s23d2.The lone d-electron pair of the titanium causes a very high reactivity in compari-son to the p-electrons of silicon. The consequence is that titanium alkoxides aremuch more reactive than its corresponding silicon alkoxides. Titania thereforeprecipitates by hydrolysis and condensation of titanium alkoxides. The mecha-nism for hydrolysis and condensation is nucleophilic substitution. The hydrolysistakes place by nucleophilic addition of a water molecule to the titanium alkoxideat neutral conditions. This leads to a transition state with an intermediate withcoordination number five [36], and produces a good leaving group shown in Figure 2.4.

The leaving group can be deduced by a partial charge model. In the non-catalyzedreaction a proton transfer is essential to take place in the transition state. Aproton is transferred from the incoming water molecule to the leaving alkoxy-groupproducing the parent alcohol as the leaving group in the neutral reaction mechanismshown in Figure 2.4. [14] The meaning of ’parent’ in this sense is the correspondingalcohol to the alkoxy-groups. The molecular species with the largest negativepartial charge is the nucleophile. The partial charge of a molecular species is

2.3 Sol-gel chemistry 9

Figure 2.4: The reaction formulas under neutral conditions pH=7. [13;14;36].

related to its reactivity. When two atoms combine in a reaction a charge transferoccurs. The electronegativity is linear with the atom charge. The charge transferstops when the electronegativities of the moleuclar species becomes equal to themean electronegativity. This is the principle of electronegativity equalization andcorresponds to the electronic corresponding principle of chemical equilibrium inthermodynamics. [14]. In the absence of a catalyst, hydrolysis and condensation istaking place by a SN -mechanism according to nucleophilic addition and a consecutiveproton transfer in the transition state and the production of a leaving group whichreceives the proton as depicted in Figure 2.4. The reactions depend on the strengthof the nucleophile, the electrophilicity of the metal atom and the partial charge of ofthe leaving group, δ(molecule) . The general reaction is that an alkoxy-group wherethe partial charge is negative, attacks the metal atom where the partial chargeis positive. Silicon alkoxides are not very reactive with water (pH=7) comparedto the transition metal alkoxides which causes highly exothermic reactions. Thepartial charge of the titanium metal atom in Ti(OEt)4 is +0.63 and for Si inSi(OEt)4 it is +0.32. A minimum value of the rate of hydrolysis can be estimated tok = 1mmoldm−3s−1 at pH=7 for Si(OR)4 which is five orders of magnitudes lessthan for Ti(OR)4. [13] Acid or base catalysts can influence both the hydrolysis andcondensation rates, but also the structure of the condensed product. [14] Acids serveto protonate nucleophilic alkoxide groups in the reaction mechanism, enhancingthe reaction kinetics by producing good leaving groups and eliminate the need forproton transfer in the transition state. Therefore hydrolysis goes to completion ifsufficient water is added. [13]

Under basic conditions the hydroxyl ion acts as a nucleophile towards titaniumalkoxides and no proton transfer is needed in the transition states [37]. For example

10 Background

Figure 2.5: The reaction mechanisms under acidic conditions. The acidic hydrolysis is relatedto a SN2-reaction and the condensation is a SN1-type of reaction.

the hydrolysis of titanium isopropoxide is not generating any isopropanol leavingmolecules under basic conditions compared to acidic conditions. Increasing thesynthesis temperature causes a faster hydrolysis and condensation, in general, dueto lowering of the energy barrier for the reaction. In general under a large pH rangefor acidic conditions the hydrolysis is fast and the condensation is slow. Underbasic conditions the hydrolysis is slow and condensation is fast. By dissolvingthe transition metal alkoxide in different media, different nucleophilic interactionsoccurs due to the polarity of the solvent. For example, the titanium(IV) ethoxidetakes the form of an oligomer in non-polar solvents. This is due to the coordinationnumber of 5-6. This type of alkoxy bridges are is stable to hydrolysis and thereforeslows down the hydrolysis rate, but causes the formation of unwanted gels instead.When dissolving titanium ethoxide in ethanol the alkoxide is instead monomericdue to solvent interaction instead. Metal alkoxides are often dissolved in organicsolvent before hydrolysis is performed. As a general rule, dilution should lead tolower alkoxy association but the polarity of the solvent has to be taken into accounte.g. Ti(OEt)4 remains trimeric in non-polar solvent such as hexane, but the sameis not true for using polar solvent such as ethanol. This is due to the nucleophilicityof the polar solvent causing solvation of the oligomer. The polar solvent is thusexpected to be associated with the alkoxide instead.

2.3.1 The effect of synthesis parametersThe changes upon pH of the sol-gel chemistry of titania under acidic and basic condi-tions is explained by adsorption and desorption of hydronium ions or hydroxide ionsrespectively at titanium and oxygen-sites. [38] The time for reaching the equilibriumof adsorption of hydroxide ions to Ti(OH)4 is a few hours at room temperature

2.4 Mesoporosity 11

and is accelerated at higher temperatures. [38] Neutral electrolytes also can enhancethe anatase transformation due to shielding the effect of the attractive forces fromtitanium-atoms to hydroxide ions under basic conditions. Increasing the adsorptionof hydroxide ions to TiO2 nuclei is the reason for the declining nucleation rate ofanatase TiO2 with basic pH. [38] The synthesis temperature determines the phasetransformation temperature during calcination. For a lower water-to-alkoxide ratio,organic components are easier to eliminate and the grain size is smaller explainedby less degree of polymerization. [39]

Figure 2.6: The reaction mechanisms under basic conditions.

2.4 MesoporosityThe discovery of the first synthetic zeolite ZSM-5 was done by what today is theExxon Mobil Corporation. But it was not until the discovery of the first mesoporoussilica MCM-41 that the wide research into the mesoscale porosity (2-50 nm) begun in1992 [40]. If the pores are below 2 nanometres they are called micropores and above50 nanometres they are called macropores. It was mesoporous silica with the use ofcationic surfactants that was firstly discovered in [40]. An alkaline route renders thesilica negatively charged, whereas the surfactant is positively charged. This leads toa charge interaction between a micellar phase above a critical micelle concentration.It is owing to this charge interaction that there is a pore formation. The frameworkcondenses around an already formed surfactant liquid crystalline phase. Several dif-ferent mesoporous titania routes have been reported since then such as the use of theanionic surfactant Dodecyl sulphate [41], the neutral surfactant Dodecylamine [42],the cationic surfactant Cetyltrimethylammonium bromide [43], the neutral poly-

12 Background

meric surfactant P123 [44] and a non-surfactant β-cyclodextrin/urea-mixture [45].The use of derivatives of nucleotides as templates has only been reported in silica [46].

Figure 2.7: (a) Folic acid molecule. (b) Tetrameric self-assembly unit of folate molecules [46].

Figure 2.8: The sequence of self-assembly with stacking into columns and into an hexagonalphase with p6mm symmetry.

2.5 Self-assemblySupramolecular self-assembly is defined as the process by which a molecular speciesforms spontaneously from its monomer molecules. Common interactions are aro-matic interactions, van der Waals forces and hydrogen bonding. It is possible toanalyze the interactions based on combined perpendicular T-geometry, face-to-faceinteractions and offset interactions of π − π-stacked systems [47;48]. Net favourableπ − π interactions are not due to attractive π − π interactions but occur due toattractive interactions between π-electrons and the σ-framework i.e. π − σ interac-tions that are larger than the repulsion of the π−π interactions. [49]. Garcia-Bennettet al. showed that folic acid, see Figure 2.7(a), self-assemble and can be utilizedfor pore formation with a silica framework, using (3-aminopropyl)triethoxysilanefor inducing the self-assembly [46]. The fact that folic acid derivatives self-assemblein salt solution has been reported previously [50;51]. The self-assembly with ionsare best with potassium ions for guanosine derivatives and sodium ions for folic

2.5 Self-assembly 13

acid which utilizes cation-pi interactions [52–54]. The self-assembly used in sol-gel byGarcia et al. [46] of folate is somewhat different from the surfactant self-assemblysince no hexagonal phase exists but is formed in-situ. The self-assembly rely on thehydrogen-bonding of one of the tautomers of the pterin-group. The pterin-group isthe nucleotide-derivative of guanine. Four folate pterin-groups with two hydrogenbonds per group arrange the folate molecules into tetramers, shown in Figure 2.7(b).The tetramers stack themselves as face-to-face [46;55] in a beautiful way as chiralcolumns with the glutamate groups of the folic acid sticking out from the columns,see Figure 2.8. The distance between tetramers in the stacks is 0.33-0.34 nm forsodium folates. The sodium is thought to be present between the tetramers stabiliz-ing the structure. Both carboxylates of folate molecules interact with APES. Theco-structure directing agent (3-aminopropyl)triethoxysilane is depicted in Figure2.11. It acts as a base causing production of OH – catalysing hydrolysis of thealkoxide. But it also associates with its protonated positive amine-group to thenegative deprotonated carboxylic group of the folate-stacks. The alkoxy groups of(3-aminopropyl)triethoxysilane can then contribute in hydrolysis and condensationof an inorganic network. The tetrameric stacks produce a hexagonal phase withina framework. The stacks formed by the discoid tetramers are depicted in Figure2.8. The APES is critical to the pore condensation as it will cause condensation tooccur at the surface of the folate-stacks. The pore condensation is depicted for asingle folate molecule of a tetramer in Figure 2.9(a) and a corresponding pore issketched in Figure 2.9(b).

Figure 2.9: (a) Condensation of a mesopore under basic conditions. The titania has a negativelycharged surface.(b) The Figure is a sketch of a plausible pore caused by one tetramer of afolate-stack. When the as-synthesized titania is calcined the organic is burned.

14 Background

Figure 2.10: The relationship of the pH to the self-assembly process at room temperature. Notethe neutral amine-group of the folate above pKa = 8.3. Below the pH of 3.5 the amine-group ofthe folic acid is positively charged and both carboxylic groups of the glutamic acid moiety are notdeprotonated. Below the pH of 5 the α-carboxylate is formed. Below the pH of 8.3 the γ-carboxylicgroup is deprotonated. Above the pH of 8.3 the amine of the pterin-group is deprotonated. Theneutral amine group of the pterin moiety is critical to the Hoogsten-type hydrogen bonding for theformation of tetramers.

Folate forms at a pH above 3.5. [56] Above the pH of the pKa = 8.3, the amine onthe pterin-group of folate is deprotonated. [46;57] The amine takes part in Hoogsten-type hydrogen bonding [49] causing the formation of the tetrameric unit. Belowthe pKa = 9.8 of the base APES the amine of (3-aminopropyl)triethoxysilane isprotonated [46]. The synthesis conditions are therefore in the approximate pH rangeof 8.3 < pH < 9.8. The reason for using a CSDA is that the titania is negativelycharged under basic conditions. The folate is also negatively charged hence an elec-trostatic interaction causes the possibility to condense a framework surrounding thefolate stacks. The relationship with pH of the folate is similar to the titration curveof glutamic acid. The scheme with relation to the pKa-values is shown in Figure 2.10.

Figure 2.11: (3-aminopropyl)triethoxysilane. The co-structuring agent (CSDA).

2.6 Hydrolysis retardation 15

The crystal structure of titania that is obtained by synthesis is determined primarilyby the preparation. The crystallite size of anatase can be controlled by adjustingthe aging time of the dry gel. The rutile content can be controlled by the wet gelaging time. Moreover, the surface structure also contributes to the nucleation ofrutile. [58]

Figure 2.12: Organic reaction mechanism of 1-isopropoxytitanatrane.

2.6 Hydrolysis retardationSelecting the proper solvent for the reaction can determine the precipitation of a sta-ble colloid. [13;14;20] Sol-gel reactions of TTIP has all led to immediate precipitationin water and organic solvent. This is not the case if a chelating agent is used. [59]As was stated in section 2.3 the hydrolysis rate of titanium alkoxides are very high.The reaction rate can be decreased by lowering the functionality i.e. the amount ofavailable alkoxy-groups that can be hydrolysed can be decreased by addition of achelating agent. [14] Increasing the coordination number, causes a higher stability tonucleophilic substitution. [13;14] Displacing the equilibrium can be made with theparent alcohol in excess in the synthesis solution under acidic conditions. [13] In termsof the available theory on sol-gel chemistry the most straightforward way to obtaina lower hydrolysis rate would be to decrease the partial charge of the titanium atomin the precursor. [13] The partial charge of titanium isopropoxide was emphasizedin section 2.3 where it was compared to the silica precursor tetraethylorthosilicate.The decrease of only the functionality can be obtained with acetylacetone. Acety-lacetone has been shown to stabilize the anatase crystal structure and prevent rutileformation up to 1000◦C. [60] The replacing ligands are however not larger than e.g.isopropoxide groups, hence not blocking coordination sites. [61] The sol-gel reactionof acetylacetone modified titanium alkoxides causes acetylacetone to remain boundto titania even when hydrolysis is performed with excess of water. [62] Lowering thepartial charge and the functionality can be obtained by the use of an atrane complex.Triethanolamine has been used to slow down the hydrolysis rate of titanium dioxidesuccessfully. [38]

2.7 Industrial drying of powdersThis section will deal with an industrial method for large-scale drying of titaniapowders. Spray-drying is a drying method that has been considered as a bottom-

16 Background

Figure 2.13: Spray-dryer setup. The droplets produced is in the micron range. (a) feed solution(b) peristaltic pump (c) nozzle column (d) nozzle (e) spray cylinder (f) outlet (g) cyclone (h)vessel (i) filter

up approach since it causes the possibility of designing functional materials withnanoscale properties of architectures in the size of microns from nanoparticles. [63]Spray-drying is an already used method of preparing large amounts of dried powdersfor the food industry [63;64] e.g. as dairy powders [63;64], for the chemical industry [64]and the pharmaceutical industry [64–68]. The method has been increasingly usedfor preparation of materials for controlled release of drugs in drug-delivery. [69;70]This is even the method for making instant coffee. [71] Spray-drying has shown toproduce micron-sized particles which are collapsed [72], donought-shaped, dispersedspherical shells [63;72;73], dispersed [73;74] and agglomerated [73] spherical solids insilica. Spray-drying has also been used to spray-dry mesoporous silica successfullyinto microspheres using surfactant [73]. Moreover, collapsed [75], spherical shells [75;76]and solid spheres [77] have been reported for titanium dioxide. Nanocrystallinetitania microspheres of an approximate size of 10 µm has shown to have a surfacearea of 560 m2/g even without any template for mesoporosity. [73] Even mesoporoustitania microspheres of size 8.3 µm that are solid have been reported but witha low surface area of 50.1 m2/g and smooth surface texture [78]. It is thereforedesirable to investigate spray-drying of nanocrystalline titania but also to investigateif mesoporosity can give higher surface area.The synthetic conditions contributes to high temperatures due to the water contenti.e. outlet temperatures needed is at least 100◦C at (d) in Figure 2.13. A lowertemperature is expected to improve the morphology of the particles. A donoughtshape form due to higher evaporation rates in the centre of the droplet, which isdue to the higher temperature in the droplet centre. The formation of a crust inthe beginning of the drying and the migration of dispersant to the surface causesthe crust formation that results in particle shells.

The spray-dryer used is depicted in Figure 2.13 with letters corresponding to thedifferent parts of the equipment. The gas supply is attached to the N2-inlet andis heated up by an electric heater. The feed solution (a) is pumped with a simple

2.8 Characterization techniques 17

peristaltic pump (b). The feed solution is heated up in the nozzle tube (c) by thehot gas. The sprayed solution leaves the nozzle at (d) and travels through the spraycylinder (e). The aerosols travel to a cyclone (g) to separate from the gas streamand the product is obtained in a vessel(h). The outgoing air is filtered and anaspirator pumps air through the system. [79] Solvent circulates in a closed loop. Theinert gas with solvent vapour is cooled hence the solvent condenses into a bottle. [79]The cleaned gas stream is pre-heated and flows back into the spray dryer. [79] Ingeneral, hollow microspheres are produced from well-dispersed sols whereas if thesols are partially aggregated they produce solid particles [80].

2.8 Characterization techniques

2.8.1 Scanning electron microscopyCharacterization of the material morphology is performed with scanning electronmicroscopy (SEM). A scanning electron microscope uses electrons instead of photonsto produce an image. To produce electrons an electron gun is used. For highresolution SEM field emission sources are used exclusionary. The lenses used tofocus the electron beam are magnetic and therefore has hysteresis built in whichcauses astigmatism. Bombarding a conducting sample with high energy electrons,can be used to produce signals that contain morphological information. Secondaryelectrons (SEs) that are knocked out from core electron shells have a low kineticenergy and can be collected with a detector that attracts the electrons with apositive potential. This way of detecting the SEs causes an artificial light in theimages. Also due to the bombarding of electrons, X-rays are produced from theexcited volume when electrons fall back from their respective excited states. X-rayenergy dispersive spectroscopy can be used to analyse the characteristic X-raysproduced.

Figure 2.14: The image shows a LEO1550 scanning electron microscope equipped with a highresolution SE-detector called InLens within the pole piece.(a) N2-liquid dewar (b) Secondaryelectron detector (c) Column of lenses (d) IR-camera (e) XEDS detector that can be inserted into the chamber.

18 Background

2.8.2 X-ray diffraction (XRD)Characterization of the crystal structure is performed with the use of X-ray radiation.The X-rays have wavelengths in the order of atomic distances in crystal structuresi.e. in the order of 0.1 nm. X-ray diffraction is described by the Bragg law

nλ = dhkl sin θhkl (2.1)

Figure 2.15: Image of a Bragg-Bretano Siemens D5000 X-ray diffractometer geometry in lockedcoupled mode; the x-ray tube and detector are aligned at the same angle. The order of the X-raypathway is (a) X-ray tube (b) divergence slit (c) illuminated spot (d) anti-scattering slit (e) filterslit for calibration of the equipment (f) detector slit (g) detector. In (h) a sample holder is shownand in (i) a sample holder lid. In (j) a titania specimen prepared by capillarity effects is shown.

where dhkl is the spacing between sets of crystal planes with indices hkl, θhkl is halfthe angle between the incoming light beam and the outgoing ray. This law describesthe constructive and destructive interference of the radiation when incident onthe set of crystal planes. The crystallite size is described by the Scherer equationwhich is equation 2.2. This equation states qualitatively that the crystallite sizedecreases with a broad Full-Width-at-Half-Maximum (FWHM). The larger FWHM,the smaller crystallite size. The opposite is also true i.e. that for a high intensitywith a small FWHM has a large crystallite size.

Dhkl = Kλ

βhkl cos θhkl(2.2)

The value n in equation 2.1 is an integer and λ is the wavelength of incoming rays.The Dhkl is the crystallite thickness, the K-factor takes a value between 0.94 and0.98 for perfect cubic and perfect spherical crystallites, respectively. All angles arein radians as well as the peak width βhkl and in θ(◦). For anatase the crystallite sizeis calculated from the 101 peak since it has the highest abundance in the crystallites.

Reflections at low-angles < 6◦are equivalent to order at distances corresponding tomesoscale order. Low-angle XRD cannot measure mesoporosity, only the quality ofthe mesoporosity in terms of the orderity. The difficulties with measurements atlow angles is that the background noise from the sample holder increases, it is alsopossible that artefacts from the detector appear. Zero noise silicon wafer sampleholders are preferred for these measurements.

2.8 Characterization techniques 19

2.8.3 N2−sorptionA typical equipment for measuring surface area and pore size is shown in Figure2.16. For a mesoporous material, filling the pores with N2-gas will primarily lead tomonolayer coverage. The next layer will then be physisorbed to the first layer at ahigher pressure. At a high enough pressure the pores will be filled with gaseousN2 that condenses to liquid N2. When lowering the pressure isothermally, it takesmore energy to fill the pore than to evaporate the liquid N2. [81] This can be seenas a relative pressure drop in desorption and hence hysteresis in the sorption data.

Figure 2.16: A Micromeritics ASAP 2020 equipment for N2-sorption measurements.(a) degasstations (b) reference tubes with thermos filled with liquid nitrogen (c) analysis station withthermos filled with nitrogen (d) glass balloon for N2-sorption measurements (e) isothermal jacket.

2.8.4 Thermogravimetric measurementsA thermogravimetric analyser is depicted in Figure 2.17(a). The thermogravimetricanalysis (TG) is used to measure weight-loss of a sample by heating at a constantrate inside a furnace (b). The furnace contains a delicate balance (c) which isable to record the weight of the sample. For porous materials the weight-loss fromburning the organic template can be deduced from the change in weight within thecorresponding temperature interval where the organic phase is burned.

2.8.5 Ultraviolet-visible spectroscopyChromophores have the ability to absorb light. This is owing to the potential wellsformed by the delocalization of electrons. The electrons can therefore be excitedto higher energy levels. The electronic states can be divided into singlet statesand triplet states. When electrons fall back from excited singlet states it causesfluorescence for singlet state transitions and phosphorescence for electrons fallingback to triplet states. Ultraviolet-visible spectroscopy is the opposite of measuringthe fluorescence i.e. measuring the transmittance of light and hence the absorptionby excitation in the ultraviolet-visible region. The UV-vis spectroscopy has beenused to measure the light absorbance as a function of adsorption of dye to titaniumdioxide.

20 Background

Figure 2.17: The figure shows a thermogravimetric analyser to measure weight loss upon heatinga sample. (a) A thermogravimetric analyser (b) a furnace (in cavity) (c) a delicate balance insidethe furnace (d) air flow meter (e) N2-flow meter.

Figure 2.18: The figure shows the box that is a UV-vis spectrometer.

2.9 MethodsThe method to achieve the aims of this thesis have been synthesis of titanium dioxideby the sol-gel method and characterization thereof. There are several differentavailable pathways including chemical vapour deposition, inert gas condensation [82],oxidation-hydrothermal synthesis of metallic titanium, non-hydrolytic [83] sol-gelsynthesis of anatase at low temperature and hydrolytic [84] synthesis of titaniafrom titanium alkoxides and titanium chlorides [85]. The reason for choosing thehydrolytic synthesis lies in the affordability [86] of the method for up-scaling withwidely researched titania precursors. The fact that this master’s thesis has theintention of investigating an extension of a silica route for Nanologica AB to thesynthesis of mesoporous titania has come with certain challenges. The largestchallenges is the basic conditions of the folate route [46] that comes from the use ofthe folic acid as a template. This comprises a narrow pH range that is essentialto have a templating effect. [46] Moreover, the reasons for using π − πstacking forpreparing mesoporous titania is that a hierarchical supramolecular self-assemblyroute has not yet been reported for titania in the literature. In addition, folicacid have interesting electronic properties [87]. The semiconducting properties ofguanine-based nanowires has been reported to conduct charge [88]. A general methodcan be extended to a range of transition metal oxides that are interesting as supports

2.9 Methods 21

in heterogeneous catalysis. A novel folate route [46] with potential is investigated forobtaining control of mesoporosity. Characterization has been made of crystallinity,morphology, organic weight-loss, surface area and adsorption.

In chapter 2, the reasons for synthesizing titania was stated for the use as adye-sensitized electrode. Several precursors can be used for the preparation oftitanium dioxide. Common precursors used are Ti(OR)4 where R=ethyl [38], iso-propyl [12;38;89;90] and butyl [39;91]. In this thesis, titanium isopropoxide was chosenbecause it is a fluid at room temperature compared e.g. titanium ethoxide. Tita-nium ethoxide is decisive because one might think that this precursor will behaveas tetraethylorthosilicate (TEOS), which has become a standard in the literaturefor synthesis of mesoporous silica [46]. As was stated in section 2.3 the titaniumisopropoxide has more chemical similarities with TEOS owing to the fact that the co-ordination number equals the oxidation state. APES can be used as a co-structuringagent and as the base. [46] The basic conditions are needed for the charge-interactionof APES and folate. The recipes used is similar to and modifications of a folate routeproposed by Garcia-Bennett et al. [46] using TEOS as precursor. In the synthesis,the reacted solution is unstirred overnight. Keeping the solution unstirred was asuitable way for storage in oven, at synthesis temperature, decreasing the batch timein oil-bath. That means a higher quantity of batches can be made consecutively. Inthe following chapter the synthesis of titanium dioxide will be described.

There are a couple of ways to consider for lowering the hydrolysis rate in thesynthesis. By capping the titanium with a nitrogen atom the nucleophilicity islowered under basic conditions. [13] Since titanatranes have a high viscosity, an1-isoporpoxytitanatrane was diluted in polar solvent to make it possible to pipettefrom a stock solution according with section 2.3. The following folate route wasprepared at room temperature with the use of the 1-isopropoxytitanatrane inethanolic solution as precursor. Addition of solvent causes dilution of the alkoxideprior to mixing which causes a slower reaction [84]. A lower water content also lowersthe hydrolysis rate of titanium alkoxides. [84] In order to commercialize a mesoporoustitania powder a spray-drying technique that causes smaller monodispersed particlesis needed for use in DSSCs. Spray-drying has been reported to be a viable route toobtaining spherical shaped particles from slurry and was described in section 2.7.

Chapter 3

Experimental

3.1 Synthesis of TiO2 by folate route (NFM-2)All synthesis of NFM-2 materials was carried out by the author. Titanium(IV)isopropoxide (TTIP) was used as a precursor. The (3-aminopropyl)triethoxysilane(APES) was used as the base. The templating agent was folic acid (FA). All chemicalswere purchased from Sigma-Aldrich and used as received. Safety precautions weretaken due to the high reactivity of the base and the TTIP. The chemicals wereprepared in the molar ratio of 0.13 FA/ 1 TTIP/0.33 APES/ 230.5 H2O [92]. Anamount of 1.21 g of folic acid was added to 87.6 ml distilled water in a plasticbottle. The aqueous folic acid yellow solution was stirred in a closed flask vigorouslyat 60◦C on a Heidolph MR3002 heater/stirrer. The base APES was added atonce using a pipette. The colour change upon addition of APES was from darkyellow to bright yellow. The precursor TTIP was added in a shot of 6.0 mL.The (3-aminopropyl)triethoxysilane was added firstly and then the titanium(IV)isopropoxide within 15 seconds. A rule of thumb was to add the precursor whenthe solution changed colour. The synthesis was heated and stirred for about 20 minand was then left unstirred for 18 h. The reacted solution was vacuum filtered andwashed with about 30 ml of distilled water.

3.2 Slowing down the hydrolysis rateSlowing down the hydrolysis rate was investigated through preparing chelatedtitanium(IV) isopropoxide with acetylacetone and triethanolamine for use in thesynthesis of NFM-2-AC and NFM-2-AT, respectively.

3.2.1 Acetylacetone route (NFM-2-AC)The recipe was only changed by increasing the complexity by preparing the precursor.Instead of adding the TTIP precursor directly to the synthesis it was chelated withacetylacetone. An amount of 6 mL acetylacetone was stirred together with TTIP

23

24 Experimental

for 10 minutes. The solution turned colour from being transparent to being brightyellow.

3.2.2 Atrane route (NFM-2-AT)

Folic acid was added to a plastic bottle and was dissolved in distilled water. Theprepared titanatrane was added in a shot. The solution was stirred at roomtemperature for about 5 minutes. The solution was clear yellow with no precipitationoccurring. The flask with the containing solution was placed in the oven at thesynthesis temperature 60 ◦C. A phase separation of solvent and the wet powdercould be seen after aging in the oven. The powder was a paste and the watercontent was high even after filtering. The paste was dried in the oven at 40◦C.An 1-isopropoxytitanatrane was prepared by organic synthesis with the reactiondepicted in Figure 3.1. Preparation was firstly investigated by rotoevaporation at40◦C to collect the isopropanol formed. It was found that this is an expensive way,since a large amount is evaporated in the vacuum line. Instead isopropanol wassmelled during evaporation without vacuum in fume cupboard and then a vacuumline attached. The preparation of the titanatrane used 30 mL TTIP added to around-bottom flask.

Figure 3.1: The reaction scheme for the synthesis of 1-isopropoxytitanatrane.

Figure 3.2: Setup for organic synthesis of 1-isopropoxytitanatrane. A vacuum line is used forevaporation of isopropanol.

3.3 Calcination 25

Figure 3.3: A Nabertherm B150 oven used for calcination. The oven can heat up to 2000◦C.

3.3 Calcination

The procedure used for calcination was cleaning ceramic boats with ethanol andwhich were dried in oven at 100◦C. A temperature ramp of 10 h for reaching 550◦Cwas used. Samples were heated isothermally for 6 h. The calcination was performedwith a Nabertherm B150 oven. The synthesis temperature was chosen to 60◦C toenhance anatase transformation during calcination.

3.4 Spray-drying (NFM-2-SD)

A Büchi Mini Spray-Dryer B-290 was used. The glassware of the spray-dryer wascleaned thoroughly for prevention of possible contamination of processed powder.A feed solution was prepared by normal synthesis as described in section 3.1. Thespray-dryer was started-up and the aspirator was set to 50 %. The N2-carrier gaswas switched on and the oxygen sensor was controlled so that the apparatus wasevacuated with air and filled with N2-flow. The flow was controlled with respect toa pressure of 20 bar. The inlet temperature was typically set to around 200◦C toreach an outlet temperature of 100◦C. During spray-drying with the pumping rateset too high the outlet temperature dropped.

Sample N2-flow (%) feed (wt%) pump (%) inlet (◦C) outlet (◦C)NFM-2-SD1 40 5 40 196 70NFM-2-SD2 55 5 20 204 77NFM-2-SD3 35 20 20 205 70NFM-2-SD4 35 20 20 180 82

Table 3.1: Spray-drying conditions as described in section 2.7. The specific conditions are onlyrelevant to the Büchi Mini Spray-Dryer B-290 and the corresponding industrial spray-dryer.

26 Experimental

3.5 Characterization

The characterization techniques used were briefly described in chapter 2. These arelow-angle XRD, TG, high-angle XRD, SEM, XEDS, N2-sorption, SEM and UV-visspectroscopy. All characterization has been performed by the author.

3.5.1 XRD

For a typical specimen preparation the powder sample was crushed to a finer powderby using a mortar and pestle. The sample was placed in a plastic holder with acavity for the sample. The amount of sample was approx 1 g. Powder samples wereprepared by making the powder smooth in the plastic sample holder by pressingwith a glass piece. A Bragg-Bretano Siemens D5000 diffractometer was used forboth high-angle and low-angle investigations. The scanning-routine used was lockedcoupled. The diffractometer was set up by increasing the voltage manually to 45kV working voltage and the working current to 40 mA. For low-angle XRD, thedivergence and receiving slits were set to 1.0 mm and the detector slit used was 0.2mm. For high-angle XRD the divergence and receiving slits was 1.0 mm whereasthe detector slit used was 0.6 mm. The samples were typically scanned with 2.5sec/step with an increment of 0.03◦(2θ) for 1 h runs, with 3.4 sec/step with anincrement of 0.02◦(2θ) for 2 h runs and with step size 0.014◦(2θ) and 10 sec/stepfor night runs. The magnetic specimen holder was inserted into the diffractometer.

3.5.2 SEM

A LEO1550 was used. Samples were prepared with low amount of powder for theSEM by mounting conducting carbon films on the specimen holders. The SEMwas set up to start at an intermediate voltage typically 7-10 kV. Depending onthe sample, typically a lower accelerating voltage of 2 kV was later required. Animproved secondary electron detector was typically used called InLens. The InLensdetector was used for higher resolutions at a low working distance about 2-3 mmand the normal secondary electron detector was used higher resolution than theInLens detector at higher working distance. A typical alignment of the columnwas made by first choosing the voltage and aperture to use. A standard 30 µmaperture was used. The aperture alignment was followed by stigmation at highmagnifications. This was done repeatedly several times at a higher magnificationsfor improvement of the resolution. An Oxford Instruments INCA system was usedfor XEDS. XEDS measurements were performed with a voltage higher than theelectronic transitions between the electron shells in titanium, silicon and oxygenwhere the highest voltage needed is above the Kβ peak of titanium. The workingdistance was set to 15 mm. The XEDS detector with ultra-thin window was insertedinto the chamber pointing towards the specimen. The conditions used in the INCAsoftware was an acquisition rate of 0.5 kcps, dead time of 7% and acquisition timeof 100 seconds.

3.5 Characterization 27

3.5.3 N2-sorptionThe isotherm-equipment, Micromeritics ASAP 2020, was filled up with liquidnitrogen. Glass balloons were flushed with acetone which was evaporated in anoven at 60◦C. The glass balloons were weighed firstly without sample, secondlywith sample and thirdly with frit. Samples were degassed for 18 h at 120◦C toremove any residual moisture. The degassed samples were weighed again to obtainthe dry weight. After the dry weight measurement the sample was inserted in theinstrument again to obtain the isotherm with the sample weight as input in thesoftware routine.

3.5.4 Thermogravimetric measurementsThermogravimetric measurements were performed on a Mettler-Toledo TGA/SDTA851e. A small crucible sample holder was used and cleaned in a preparedHCl bath. The crucibles were handled with pincette. The crucible was dried inan oven at 60◦C for a couple of minutes and the apparatus was tared with thesample holder inside. A small amount of sample was placed in the crucible whichwas placed in the furnace. The software was set up to run the measurements.

3.5.5 UV-vis absorbanceAbsorbance tests were done using a SHIMADZU UV-1650PC spectrometer and thesoftware UVProbe 2.21 from SHIMADZU. A dilute dye stock solution was preparedwith 1·10−4M of the dye N719. The dye was bought from DyeSol, Australia and usedas received. An amount of 0.1 g titania was added directly to prepared solution of20 mL stock-solution in glass-vials. Samples were tested normally after calcination.The stock-solution was prepared with tertiary butanol/acetonitrile in 1:1 volumeratio. The tertiary butanol/acetonitrile was also used as a blank. Solutions withanalyte were heated at 40◦C for 1 h before filtering and measurements. Scans weretaken from 190 - 450 nm with 0.2 nm sampling interval. The analyte was prepared bystirring the titania in stock solution for 24 hours before filtration and measurements.The following chapter describes the results obtained from the characterization oftitanium dioxide obtained from the synthetic routes and processing.

Chapter 4

Results and discussion

4.1 Designed materialsThe yellow powder obtained from NFM-2 synthesis varied in colour very muchwhen changing the synthetic conditions, such as temperature and pH. Below roomtemperature and at 45◦C, large precipitates formed whereas at 80◦C a paste wasobtained. SEM micrographs of NFM-2 are shown in Figures 4.4(a)-(d). From Figure4.4(a) it can be seen that NFM-2 powders consists of large particles, which havelarge-scale porosity and smaller porosity like channels and random large scale poresin the micron range. The Figure 4.4(b) shows that the porosity is macroporous.The largest pores are approximately 2-3 µm. The Figure 4.4(c) shows pore wallshave small particles on the surface. The Figure 4.4(d) shows that the nanoparticleson the pore walls have a surface texture below 50 nm that is not attributed tonanocrystallinity due to the amorphous nature of the samples.

The composition of NFM-2 was determined from XEDS-spectrometry and theaverage composition was 63.7 at% oxygen, 32.2% titanium and 4.1 at% silicon. TheXEDS-spectra is shown in Figure 4.1. The composition was therefore concluded tobe 88.6 % TiO2 and 11.4% SiO2.

Figure 4.2 shows low-angle XRD patterns of the different NFM-2 derivatives. Thecurve (a) is a reference curve for NFM-1 which is the mesoporous silica by the folateroute. The curve (b) shows only a broad peak. Curve (c) shows a typical peak of noorder at lower angles obtained for the materials NFM-2, NFM-2-AT and NFM-2-SD.

The Figure 4.3 shows six X-ray diffraction patterns that are shifted in relativeintensity. The curve (a) in Figure 4.3 shows the high crystallinity of P-25. The curve(b) in Figure 4.3 shows a high intensity 101 reflection at 25.4◦. A superpositionof 103, 004 and 112 reflections can be seen at 37.7◦. At 48.1◦ the 200 reflectionis present as a single peak and the peak 54.7◦ is a superposition of the 105 andthe 211 peaks. The 101 reflection has the highest intensity and therefore has thehighest abundance in the crystallites. Curve (e) has only a large broad peak which

29

30 Results and discussion

Figure 4.1: XEDS measurement on NFM-2-AC. Inset shows the site of interest for the mea-surement using a SE-detector at a working distance of 15 mm.

Figure 4.2: Low-angle XRD patterns. (a) NFM-1 (b) NFM-2-AC (c) NFM-2-AT and NFM-2

4.1 Designed materials 31

Figure 4.3: XRD patterns of calcined NFM-2 with synthesis temperature at 60◦C (a) P-25(b) NFM-2-d, pH=8.3 (c) NFM-2-a, pH=8.6 (d) NFM-2-c, pH=9.0 (e) NFM-2-b, pH=9.6 (f)NFM-2-AT

means that the powder is completely amorphous. Since 4.3(b) shows low degree oforder, the crystallinity is indeed low. In Figure 4.3 diffraction patterns are shownof NFM-2 with curves (b)-(e) in the order of increasing pH. It can be seen that thecrystallinity increases in the reverse order (e),(d),(c) and (b). Therefore it can bededuced that the crystallinity increases with decreasing pH. This can be explainedby the adsorption and desorption of hydroxide ions to titanium-sites at basic pHthat blocks condensation and prevents the formation of anatase with agreementwith Sugimoto et al. [38] described in section 2.3. Therefore a low pH in the availablerange for pi-pi-stacking to occur, is highly favourable. The curve (f) shows thediffraction pattern from the NFM-2-AT material which shows no large difference incrystallinity compared to NFM-2.

Channels were found in most samples to be pores in a random pattern. Largescale aggregation seems to occur upon drying. Based on the fact that the size ofa folate-stack is about 3 nm, the stacks should self-assemble into larger columns.A hypothesis was established that pore control may be achieved with hydrolysisrate. Therefore the acetylacetone route and atrane routes were designed by meansof using a chelating agent. The Figure 4.8 shows SEM micrographs of NMF-2-ACsamples. The Figures 4.5(a)-(d) shows SEM micrographs of NFM-AT samples.From Figure 4.5(a)-(b) it can be seen that the particle morphology is monolithic.


Figure 4.4: Representative SEM micrographs of NFM-2 obtained from the folate route.

From 4.5(c)-(d) it can be deduced that there is no obvious templating effect.

The Figures 4.6(a)-(d) show images from SEM imaging of NFM-SD with in-situspray-dried titania. The Figure 4.6(a) shows almost spherical shaped particles ina range of approximate sizes of 1-15 µm. The Figure 4.6(a) also shows cracks inthe particles that makes it possible to conclude that the spherical particles arehollow [75] and not solid throughout. The Figure 4.6(b) shows that the sphericalparticles are dispersed and are somewhat donought shaped, which can be spottedwith the change in contrast on the large particles. In Figure 4.6(c) the donoughtshape of the particles has increased in comparison to 4.6(a). This fact is even moreevident at higher magnification as shown in Figure 4.6(d). The spray-dried samplesdepicted in Figure 4.6(a)-(d) are thought to have the same particle size distributionowing to the droplet sizes produced by the spray-dryer nozzle.

The Figure 4.7(a)-(d) shows images from the NFM-2-SD material obtained fromas-synthesized slurry compared to the NFM-2-SD in-situ spray-dried material de-picted in Figure 4.6. In Figure 4.7(a) several spherical titania particles can be seentogether with some non-spherical particles. The particles have a large distributionin sizes. In Figures 4.7(b)-(d) images at high magnification are shown on threedifferent spherical particles. The Figures 4.7(b) and 4.7(c) shows particles of similarsize of 4 µm. The Figure 4.7(d) shows a spherical particle of smaller size of 1.7 µmin diameter.


Figure 4.5: SEM micrographs of the NFM-2-AT material obtained from the atrane route.

Figure 4.6: SEM imaging of morphology obtained from spray-drying. (a) NFM-2-SDA1 (b)NFM-2-SDA1 (c) NFM-2-SDA2 (d) NFM-2-SDA2 .


Figure 4.7: The spherical titania aggregates of NFM-2-SD obtained from spray-drying.(a)NFM-2-SD1 (b) NFM-2-SD2 (c) NFM-2-SD3 (d) NFM-2-SD4.

In Figure 4.8(a)-(d) SEM micrographs from the NFM-2-AC material obtained fromthe acetylacetone route is depicted. The Figure 4.8(a) shows a monolithic particlewithout macroporosity. The particles have a rough surface. The Figure 4.8(b)shows that some particles are smaller and that the monolithic particles are in factaggregated from smaller entities which also is verified from Figure 4.8(a) that showsa range of aggregate sizes 1-10 µm that are aggregated to a monolith. The Figure4.8(d) shows smaller entities of approximately 1 µm and a flake that has a textureperpendicular to the plane of the flake. The flake surface has at least one smoothsurface compared to the rough entities in close proximity. The channelled textureis thought to origin from folate-stack templating.

To obtain further information about the folic acid inside the the as-synthesizedtitania material, thermogravimetric measurements were performed. The Table 4.2shows no trend in the weight loss of folic acid and the surface area or pH.

It was realized that the pKa of the alcohol matters in the sense of retarding hydrol-ysis and condensation. The alkaline mechanism was described in section 2.3. Theformation of RO – does not promote hydrolysis since the reaction is catalysed byOH – . The base that catalyses the condensation is also OH – . Thus, if the amountof RO – increases in the solution, the hydrolysis rate will not increase. Instead itwill decrease due to displacement of the equilibrium of hydrolysis and condensation.The needed displacement is thought to be 5 orders of magnitudes in the reaction rate.


Figure 4.8: SEM imaging from acetylacetone route prepared NFM-2-AC materials.

Figure 4.9: Thermogravimetric measurements of percentage weight-loss of NFM-2 and itsderivatives.


Figure 4.10: The first derivative of weight-loss obtained from NFM-2 and its derivatives.

Figure 4.11: N2-sorption curves of the different NFM-2 materials.


Sample pH BET (m2/g) Org. w.-l. (%) D101(nm)NFM-2-d 8.3 41.5 24.9 8.0NFM-2-a 8.6 42.9 28.4 4.0NFM-2-c 9.0 - 17.3 amorphousNFM-2-b 9.6 86.2 21.2 amorphousNFM-2-AT 9.1 9.5 14.9 6.5NFM-2-AC 7.9 - 50.2 amorphousNFM-2-SD3 8.3 81.1 24.9 6.6

P-25 - 49.8 - 25

Table 4.1: Summary of characterization of all the materials.

Figure 4.12: Absorbance of light with concentration. Adsorption of dye to titanium dioxide.(a) N719 dye stock solution (b) NFM-2 (c) NFM-2-AT (d) P-25. The P-25 titania has a gooddye-adsorption attributed to super-hydrophilicity with wetting angles lower than 3◦. Since it isthe carboxylic groups of the dye that links to the titania a hydrophilic interaction is critical to theadsorption of dye. [93]

The data in Figure 4.12 shows flattened peaks that are due to the filtering of titaniain the preparation but has no direct influence on the evaluation. The Figure 4.12shows absorbance curves (b) and (c) corresponding to the developed materials withthe dye stock solution in curve (a) as the absolute reference. The absorbance forP-25 is also used for comparison depicted as curve (d). The higher amount ofadsorption to titania, the lower is the absorbance with respect to the dye stock


Sample BET (m2/g) N719-adsorption (%) Solution colourNFM-2-b 86.2 19 purpleNFM-2-AT 9.5 28 pale pink

P-25 49.8 100 transparent

Table 4.2: Summary of adsorption results.

solution. The curve (c) describes the absorbance with after exposure to NFM-1-ATwhich shows a considerable drop in absorbance at a wavelength of 313 nm. Thepeak at 313 nm corresponds to 4d − π∗ metal-to-ligand absorbance. Both thepeaks at 222 nm and 266 nm correspond to π − π∗ ligand-centred absorbance. Incomparison to the 4d− π∗ of the curves (a)-(d) the P-25 samples have the samepeak inverted. Note that NFM-2-AC was not tested due to the time frame.The onlyNFM-2 material that compares to P-25 is the NFM-2-AT which has an adsorptionof 28%, with reference to the used stock solution.

Chapter 5

Conclusions

From the results in this thesis it can be deduced that the synthesis of titaniumdioxide with folate-stacks as template has inherent limitations. The thesis has beencarried out with respect to the aims of the project description from NanologicaAB. Using alkaline conditions is not the way for obtaining highly nanocrystallineanatase compared to acidic conditions. The use of alkaline conditions for synthesisof titanium dioxide should be avoided, since those conditions will render the materialnearly amorphous. Spray-drying was shown to produce spherical particles. Thelargest impact on smaller spheres was using a low weight-fraction of as-synthesizedtitania powder in the feed solution.

39

Chapter 6

Future work

Future research should focus on the advantages of the acidic conditions of thesol-gel chemistry of titanium alkoxides that would cause nanocrystalline anataseat low calcination temperatures. Guanosine-monophosfate is the first choice of aπ − π-stacked template for a similar approach as in this thesis because it will stackunder acidic conditions. [53] More research on controlling the mechanism of formingπ− π-stacks of aromatic molecules with respect to sol-gel chemistry is needed. Alsoresearch on the in-situ mechanism of phase formation for π−π-stacks in comparisonto surfactant self-assembly is needed.

41

References

[1] B. O’Regan and M. Grätzel, Nature, 1991, 353, 737–741.

[2] C. J. Barbé, F. Arendse, P. Comte, M. Jirousek, F. Lenzmann, V. Shklover, and M. Grätzel,Journal of American Ceramic Society, 1997, 80, 3157–3171.

[3] M. Grätzel, 2004, 164, 3–14.

[4] M. Grätzel, Inorganic Chemistry, 2005, 44, 6841–6851.

[5] B. E. Hardin, E. T. Hoke, P. B. Armstrong, J.-H. Yum, P. Comte, T. Torres, J. M. J. Fréchet,M. K. Nazeeruddin, and M. Grätzel, Nature Photonics, 2009, 3, 406–410.

[6] S.D. Mo and W.Y. Ching, Physical Review B, 1995, 51, 13023–13032.

[7] Y.-S. Ko, M.-H. Kim, and Y.-U. Kwon, Bulletin of Korean Chemical Society, 2008, 29,463–467.

[8] T.P. Chou, Q. Zhang, G.E. Fryxell, and G. Cao, Crystal Growth and Design, 2007, 19,2588–2592.

[9] S.-Y. Dai and K.-J. Wang, Chinese Physics Letters, 2003, 20, 953–956.

[10] M. Zukalová, J.Procházka, A. Zukal, J.H. Yum, and L. Kavan, Inorganica Chimica, 2008,361, 656–662.

[11] D.W. Lee, S.-J. Park, S.-K. Ihm, and K.-H. Lee, Langmuir, 2006, 19, 937–941.

[12] K.M.S. Khalil and M. I. Zaki, Powder Technology, 1997, 92, 233–239.

[13] C. J. Brinker, Sol-Gel Science; Academic Press Inc., San Diego, 1990.

[14] J. Livage, Progress in Solid State Chemistry, 1988, 18, 259–341.

[15] S. Y. Choi, B. Lee, D. B. Carew, M. Mamak, F. C. Peiris, S. Speakman, N. Chopra, andG. A. Ozin, Advanced Functional Materials, 2006, 16, 1731–1738.

[16] F. D. Fonzo, C. S. Casari, V. Russo, M.F. Brunella, and A. Li. Bassi, Nanotechnology, 2009,20, 1–7.

[17] O. Wilhelm, S.E. Pratsinis, E. de Chambrier, M. Crouzet, and I. Exnar, Journal of PowerSources, 2004, 134, 197–201.

[18] Pottier et al., Journal of Materials Chemistry, 2003, 13, 877–882.

[19] C. Sanchez et al., Journal of Materials Chemistry, 2006, 16, 77–82.

[20] D. Chen, Advanced Materials, 2009, 21, 1–5.

43

44 Future work

[21] S. Ngamsinlapasathian, S. Pavasupree, Y. Suzuki, and S. Yoshikawa, Solar Energy Materialsand Solar Cells, 2006, 90, 3187–3192.

[22] T. Ohno, K. Sarukawa, K. Tokieda, and M. Matsumara, Journal of Catalysis, 2001, 203,82–86.

[23] J. F. Porter, Y.-G. Li, and C. K. Chan, Journal of Materials Science, 1999, 34, 1523–1531.

[24] J.D. Holmes, K. Wang, J.P.Hanrahan, M.A. Morris, and T.R. Spalding; Process for preparingmesoporous materials u.s. patent 2007/0166226 a1 filed apr 13, 2005.

[25] J. Y. Ying, D.M. Antonelli, and T. Sun; Methods for preparing porous metal oxides u.s.patent 5958367 filed oct 10, 1996, issued sep 28, 1999.

[26] S.Y. Huang, G. Schlichtörl, M.Grätzel A.J. Nozik, and A. J. Frank, Journal of PhysicalChemistry B, 1997, 101, 2576–2582.

[27] D. Wang, Y. Liu, B. Yu, F. Zhou, , and W. Liu, Chemistry of Materials, 2009, 21, 1198–1206.

[28] M. Law, L. E. Greene, J. C. Johnson, R. Saykally, and P. Yang, Nature Materials, 2005, 4,455–459.

[29] B. Tan and Y. Wu, Journal of Physical chemistry B, 2006, 110, 15932–15938.

[30] Q. Zhang, T. P. Chou, B. Russo, S. A. Jenekhe, and G. Cao, Advanced Functional Materials,2008, 18, 1654–1660.

[31] M. Howard, Brookite, Rutile Paramorphs after Brookite and Rutile Twins from Magnet Cove;Heldref Publications, 1999.

[32] L.E. Smart and E. E. Moore, Solid State Chemistry: An Introduction; Taylor and Francisgroup, CRC press, New York, 2005.

[33] http://wikis.lib.ncsu.edu/index.php/Anatase_and_Brookite.

[34] http://wikis.lib.ncsu.edu/index.php/Rutile.

[35] K. Ranjit, I. Martyanov, D. Demydov, S. Uma S. Rodrigues, and K. Klabunde, Journal ofSol-Gel Science and Technology, 2006, 40, 335–339.

[36] J.-K. Park, J.-J. Myoung, J.-B. Kyong, and H.-K. Kim, Bulletin of Korean Chemical Society,2003, 24, 671–673.

[37] M. Elanany, P. Selvam, T. Yokosuka, S. Takami, M. Kubo, A. Imamura, and A. Miyamoto,Journal of Physical Chemistry B, 2003, 107, 1518–1524.

[38] T. Sugimoto, X. Zhou, and A. Maramatsu, Journal of Colloid and Interface Science, 2003,259, 43–52.

[39] X.-Z. Ding, Z.-Z. Qi, and Y.-Z. He, Journal of Materials Science Letters, 1995, 14, 21–22.

[40] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, and J.S. Beck, Nature, 1992, 359,710.

[41] V. Luca, J.N. Watson, M. Ruschena, and R.B. Knott, Chemistry of Materials, 2006, 18,1156–1168.

[42] Y. Wang, S. Zhang, and X. Wu, Nanotechnology, 2004, 15, 1162–1165.

[43] A. E. C. Palmqvist, Current Opinion in Colloid and Interface Science, 2003, 8, 145–155.

[44] D. Grosso, G.J. de A.A. Soler-Illia, F. Babonneau, C. Sanchez, P.-A. Albouy, A. Brunet-Bruneau, and A. Balkenede, Advanced Materials, 2001, 13, 1085–1090.

45

[45] J.-Y. Zheng, J.-B. Pang, K.-Y. Qiu, and Y. Wu, Journal of Materials Chemistry, 2001, 11,3367–3372.

[46] Rambabu Atluri, Nilas Hedin, and A. E. Garcia-Bennett, Journal of the American ChemicalSociety, 2009, 131, 3189–3191.

[47] C. A. Hunter, Chemical Society Reviews, 1994.

[48] C. A. Hunter and J. K. M. Sanders, Journal of American Chemical Society, 1990, 112,5525–5534.

[49] L.F. Lind and I.M. Atkinsons, Self-assembly in Supramolecular Systems; 2000.

[50] G. Gottarelli, E. Mezzina, and G. P. Spada, Helvetica Chimica Acta, 1996, 79, 220–234.

[51] Y. Kamikawa, M. Nishii, and T. Kato, Chemical European Journal, 2004, 10, 5942–5951.

[52] J. C. Ma and D. A: Dougherty, Chemical Reviews, 1997, 97, 1303–1324.

[53] J. T. Davis, Supramolecular Chemistry, 2004, 43, 668–698.

[54] A. Wong, R. Ida, L. Spindler, and G. Wu, Journal of American Chemical Society, 2005, 127,6990–6998.

[55] Leonard F. Lindoy and Ian M. Atkinson, Self-assembly in supramolecular systems; RoyalSociety of Chemistry, 2000.

[56] E. A. Mironov and V.S. Nabokov, Khimiko-Farmasevticheskii Zhurnal, 1976, 10, 136–140.

[57] M. Poe, The Journal of Biological Chemistry, 1977, 252, 3724–3728.

[58] Y. Li, T.J. White, and S.H. Lim, Journal of Solid State Chemistry, 2004, 177, 1372–1381.

[59] H.-J. Chen et al., Materials Chemistry and Physics, 2007, 101, 12–19.

[60] Y. Djaoued, S. Badilescu, P.V. Ashrit, D. Bersani, P.P. Lottici, and J. Robichaud, Journal ofSol-Gel Science and Technology, 2002, 24, 255–264.

[61] Vadim et al., J. Sol-Gel Sci. Techn., 2006, 40, 163–179.

[62] A. Léaustic, Florence Babonneau, and J. Livage, Chemistry of Materials, 1989, 1, 248–252.

[63] Okuyama et al., Advanced Powder Technology, 2006, 17, 587–611.

[64] K. Masters, Spray drying handbook; John Wiley and Sons, New York, 1991.

[65] C. Freitas and R. H. Muller, European Journal of Pharmaceutics and Biopharmaceutics,1998, 46, 145–151.

[66] A. S. Rankell, H. A. Lieberman, and R.F. Schiffmann, The Theory and Practice of IndustrialPharmacy; Lea and Febiger, Philadelphia, 1986.

[67] Y.-F. Maa, P.-A. T. Nguyen, and S. W. Hsu, Journal of Pharmaceutical Sciences, 1998, 87.

[68] R. Vehring, Pharmaceutical Research, 2008, 25, 999–1022.

[69] L. Mu and S.S. Feng, Journal of Controlled Release, 2001, 76, 239–254.

[70] P. Kortesuo, M. Ahla, M. Kangas, M. Jokinen, T. Leino, L. Vuorilehto, S. Laakso, J. Kiesvaara,A. Yli-Urpo, and M. Marvola, Biomaterials, 2002, 23, 2795–2801.

[71] A. R. Mishkin and W. S. Symbolik; Spray drying process u.s. patent 3620776 filed jun 28,1968, and issued nov 1971.

46 Future work

[72] P. J. Bruinsma, A. Y. Kim, J. Lui, and S. Baskaran, Chemsitry of Materials, 1997, 9,2507–2512.

[73] H. Minoshima, Journal of Materials Chemistry, 2004, 14, 2006–2016.

[74] P. Kortesuo, M. Ahola, M. Kangas, I. Kangasniemi, A. Yli-Urpo, and J. Kiesvaara, Interna-tional Journal of Pharmaceutics, 2000, 200, 223–229.

[75] M. Iida, T. Sasaki, and M. Watanabe, Chemistry of Materials, 1998, 10, 3780–3782.

[76] K. Jinsoo, W. Oliver, and P.E. Sotiris, Journal of American Ceramic Society, 2001, 84,2802–2808.

[77] H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Cheng, and G. Q. Lu,Nature, 2008, 453, 638–641.

[78] P. O. Vasiliev, B. Faurre, J. B. S. Ng, and L. Bergström, Journal of Colloid and InterfaceScience, 2008, 319, 144–151.

[79] Buchi Mini Spray Dryer manual.

[80] E. Sizgek and J.R. Bartlett, Journal of Sol-Gel Science and Technology, 1998, 13, 1011–1016.

[81] G.-Q. Lu and X.-S. Zhao, Nanoporous Materials; Imperial College Press, London, 2004.

[82] C.-C. Wang and J.Y. Ying, Chemistry of Materials, 1999, 11, 3113–3120.

[83] M. Niederberger, M. H. Bartl, and G. D. Stucky, Chemistry of Materials, 2002, 14, 4364–4370.

[84] D. Vorkapic and T. Matsoukas, Journal of American Chemical Society, 1998, 81, 2815–2820.

[85] G. Li, L. Li, J. Boerio-Goates, and B. F. Woodfield, Journal of Materials Chemistry, 2005,127, 8659–8666.

[86] M. Z.-C. Hu, E. A. Payzant, and C. H. Byers, Journal of Colloid and Interface Science, 2000,222, 20–36.

[87] S. Gawecda, G. Stochel, and K. Szacilowski, ChemistyAn Asian Journal, Angewandte Chemie,2007, 2, 580–590.

[88] A. Calzolari, R. D. Felice, and E. Molinari, Solid State Communications, 2004, 131, 557–564.

[89] D. D. Dunuwila, C. D. Gagliardi, and K. A. Berglund, American Chemical Society, 1994, 6,1556–1562.

[90] J. Wu, S. Hao, J. Lin, M Huang, Y Huang, Z. Lan, and P. Li, Crystal Growth and Design,2007, 8, 247–252.

[91] S.B. Deshpande, H.S. Potdar, Y.B. Khollam, K.R. Patil, R. Pasricha, and N.E. Jacob,Materials Chemistry and Physics, 2006, 97, 207–212.

[92] R. Atluri, N. Hedin, and A. E. Garcia-Bennett, Journal of the American Chemical Society,2009, 131, 3189–3191.

[93] Y. Chen, E. Stathatos, and D. D. Dionysiou, Journal of Photochemistry and Photobiology A:Chemistry, 2009, 203, 192–198.

Master thesis - Nanologica AB

Technology

american ceramic

korean chemical

sensitized

american chemical

transition

nabertherm

power conversion

nias crystal