UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Conjugated polymers in mesoporous hosts Pattantyus-Abraham, Andras Geza 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2004-902498.pdf [ 34.46MB ]
Metadata
JSON: 831-1.0061179.json
JSON-LD: 831-1.0061179-ld.json
RDF/XML (Pretty): 831-1.0061179-rdf.xml
RDF/JSON: 831-1.0061179-rdf.json
Turtle: 831-1.0061179-turtle.txt
N-Triples: 831-1.0061179-rdf-ntriples.txt
Original Record: 831-1.0061179-source.json
Full Text
831-1.0061179-fulltext.txt
Citation
831-1.0061179.ris

Full Text

C O N J U G A T E D P O L Y M E R S I N M E S O P O R O U S H O S T S by A N D R A S G E Z A P A T T A N T Y U S - A B R A H A M B . Sc. , Queen's Univers i ty , 1996 B . Sc. , Queen's Univers i ty , 1997 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department o f Chemis t ry) W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A September 2003 © Andra s G e z a Pat tantyus-Abraham, 2003 Abstract The subject o f this thesis is the synthesis and characterization o f composite materials based on electroluminescent conjugated polymers in mesoporous hosts. These materials were studied w i t h the goal o f p roduc ing a structure i n w h i c h the electrical properties on encapsulated conjugated po lymer chains cou ld be measured. Towards this goal , both the creation o f thin film hosts w i t h oriented and ordered mesopores and new methods for the incorporat ion o f polymers into mesoporous hosts are described, a long w i t h characterization techniques for showing the po lymer distr ibution on the nanometre scale. The preparation o f the conjugated po lymer poly( 1,4-phenylene vinylene) ( P P V ) inside the 3.1 n m channels o f a hexagonally ordered mesoporous s i l i ca material , M C M - 4 1 , is described. The M C M - 4 1 surface was first derivat ized w i t h an organic base. Subsequent introduct ion o f monomer d issolved in ethanol resulted i n base-initiated polymer iza t ion in the pores o f M C M - 4 1 . A pore size reduct ion o f 0.3 n m was seen i n the composite material by ni trogen physisorpt ion. Electron-energy loss spectroscopy ( E E L S ) showed that the composi te had a distinct loss signal related to the Tt-electron system on the polymer . Energy-fi l tered t ransmission electron microscopy ( E F T E M ) wi th 200 k e V electrons showed that the po lymer was evenly distributed throughout the composi te material through mapp ing o f the 7i-electron losses near 6 e V . A po lymer mass content o f ~8 % indicated the presence o f approx. 6 po lymer chains in each pore. The photophysical properties o f P P V inside the composi te were found to be s imi lar to bu lk P P V . For the preparation o f mesoporous thin f i lms wi th channels oriented normal ly to the surface, three literature approaches were investigated: the self-assembly o f mesoporous s i l i ca films w i t h the S B A - 2 structure, the thermal oxidat ion o f FeO /Si02 f i lms, and the anodic oxidat ion o f a l u m i n u m substrates. The latter approach, carried out at l o w temperature, is shown to y i e l d a lumina f i lms w i t h the desired pore structure. F i l m s w i t h a pore size o f 4 ± 1 n m are created w i t h an applied potential o f 15 V i n 1.2 M sulfuric ac id ( in 1:1 water:methanol) at -39 ° C . E E L S and E F T E M were appl ied to the analysis o f composite materials created b y the adsorption o f a thin layer o f poly[2-methoxy-5-(2 ' -e thylhexyloxy)- l ,4-phenylene vinylene] ( M E H - P P V ) on the surface o f porous a lumina membrane w i t h 60 n m pore diameter. The measurements were carr ied out w i th a 200 k e V electron beam travel l ing paral lel to the pores o f the host membrane. The TC-electron losses o f the po lymer cou ld not be discerned in this i i geometry. Strong surface losses were present at 8, 13 and 18 e V . A long-range loss mode not associated w i t h bu lk or surface losses appeared at 7.0 e V , up to 30 n m from the pore surface. B o t h these loss modes interfered w i t h the detection o f the 7t-electron losses and the po lymer dis tr ibut ion cou ld not be conf i rmed. The or ig in o f the long-range loss mode was identif ied as the Cherenkov effect. E E L S w i t h 120 k e V electrons shifted the peak energy o f this loss mode to 8.3 e V , w h i c h indicated a dependence on the electron speed. Samples wi th different pore diameters but a f ixed interpore spacing also showed shifts i n the peak posi t ion. Theoret ical mode l l ing o f the loss spectrum o f a cy l ind r i ca l pore suggested that these observations arise from the interaction o f the generated Cherenkov radiation w i t h the nearby pores i n the membrane. T h i s introduces the poss ibi l i ty o f s tudying photonic nanostructures b y E E L S . Different methods for in t roducing the conjugated po lymer into an oriented porous host are explored. The idea o f creating surface-grafted conjugated polymers on s i l i con substrates through step polymer iza t ion is investigated; as p r o o f o f concept, a surface-grafted d i m e r is synthesized through the Wadswor th -Horne r -Emmons reaction. It is further shown that s imple centrifugation o f a po lymer solution, w h i l e a l l owing solvent evaporation, provides a sufficient d r i v i n g force for polymer insert ion into the host. Th i s composi te is investigated by electron microscopy . E E L S and E F T E M analysis were also appl ied to u l t ramicrotomed thin sections o f this material , w i t h the electron beam perpendicular to the pore axis. The results showed that relativist ic effects m a y also be important i n this geometry, effectively mask ing the dis tr ibut ion o f the 7t-electron losses associated w i t h the polymer . Possible routes to the preparation o f a l ight-emit t ing device ( L E D ) based on porous a lumina films are described. The use o f the under ly ing a l u m i n u m substrate as electron-injecting electrode was investigated but devices prepared i n this manner d i d not show electroluminescence. The formation o f porous a lumina films o n conduct ive substrates such as s i l i con , i n d i u m t in oxide and gold was also investigated and the encountered experimental difficult ies are reported. m Table of Contents Abstract i i Table o f Contents i v L i s t o f Tables v i i L i s t o f Figures v i i L i s t o f S y m b o l s and Abbrev ia t ions x i Acknowledgemen t s x i v CHAPTER 1 Introduction 1 1.1 Conjugated Po lymers : Organ ic Conductors and Semiconductors 2 1.1.1 Organic M o l e c u l e s w i t h Conjugated Tt-Electron Systems 3 1.1.2 Conjugated M o l e c u l e s i n the S o l i d State 4 1.1.3 Organ ic Conjugated Po lymers 4 1.2 Luminescence i n Conjugated M o l e c u l e s 7 1.3 Elect roluminescence i n Conjugated Po lymers 8 1.3.1 Elec t ro luminescence Processes i n Conjugated Po lymers 8 1.3.2 Elec t ro luminescence Ef f i c i ency 10 1.3.3 A n Ideal D e v i c e Structure 11 1.4 Encapsulated Conjugated Po lymers 13 1.4.1 Approaches to Encapsula t ion 13 1.4.2 Ordered Porous Mater ia l s as Encapsulants 14 1.4.3 Or iented Mesoporous S i l i c a Encapsulant 15 1.4.4 L i q u i d Crys t a l Encapsulant 17 1.4.5 C l a y Encapsulant 20 1.4.6 C y c l o d e x t r i n Encapsulant 21 1.4.7 Literature Summary 23 1.5 Thesis S u m m a r y 23 References 25 CHAPTER 2 Characterization of Nanocomposite Materials 31 2.1 Scanning Elec t ron M i c r o s c o p y 32 2.2 Transmiss ion Elec t ron M i c r o s c o p y ( T E M ) 33 2.2.1 T E M Sample Preparation 34 2.3 H i g h Reso lu t ion C h e m i c a l A n a l y s i s 37 2.3.1 X - R a y Photoelectron Spectroscopy 38 2.3.2 Energy-Dispers ive X - R a y Spectroscopy and Scanning A u g e r M i c r o s c o p y . 39 2.3.3 E lec t ron Ene rgy -Loss Spectroscopy ( E E L S ) 4 0 2.4 Genera l Pr inc ip les o f E E L S 4 0 2.5 E E L S Instrumentation 43 2.6 Approaches to E E L S Data A c q u i s i t i o n 45 2.7 Quantitative A n a l y s i s o f E E L S Spectra 46 2.8 A p p l i c a t i o n o f E E L S to Organ ic Mater ia l s 47 2.9 C o n c l u s i o n 48 CHAPTER 3 A PPV/MCM-41 Composite Material... 51 3.1 Ordered Porous Host Mater ia l s 52 iv 3.1.1 Zeol i tes 52 3.1.2 Mesoporous Mater ia l s 53 3.2 Character izat ion o f M C M - 4 1 Mater ia l s 56 3.2.1 Dif f rac t ion Techniques 57 3.2.2 Phys isorp t ion 58 3.2.3 Other Techniques 59 3.3 Po lymer i za t ion w i t h i n M C M - 4 1 60 3.3.1 P P V i n M C M - 4 1 60 3.4 Exper imenta l Results 62 3.5 D i scuss ion 71 3.5.1 Thermogravimetr ic A n a l y s i s 71 3.5.2 Phys isorp t ion Data . . ! 72 3.5.3 X - r a y and Neu t ron Dif f rac t ion 73 3.5.4 E E L S and E F T E M 74 3.5.5 U V / V i s Absorbance and Photoluminescence 76 3.6 C o n c l u s i o n 76 Exper imenta l Detai ls 78 References 80 CHAPTER 4 Preparation of Mesoporous Thin Film Host 84 4.1 A l i g n m e n t o f Porous T h i n F i l m s 85 4.1.1 Interface-Induced A l i g n m e n t 86 4.1.2 F ie ld- Induced A l i g n m e n t 88 4.1.3 Or iented Porous T h i n F i l m s by Other Approaches 90 4.2 Further Investigation o f S B A - 2 Mesoporous S i l i c a F i l m s 92 4.3 Further Investigation o f the F e O / S i 0 2 Sys tem 94 4.4 Porous A l u m i n a F i l m s ( A n o d i c A l u m i n u m O x i d e ) 99 4.4.1 Pore W a l l Structure 101 4.4.2 Pore G r o w t h Processes 102 4.4.3 Pore Lat t ice Format ion 103 4.4.4 Preparation o f O p t i m a l Host from Porous A l u m i n a 105 4.4.5 Barr ie r Laye r T h i n n i n g 108 4.5 C o n c l u s i o n 109 Exper imenta l Detai ls 110 References 113 CHAPTER 5 Characterization of a Porous Alumina/MEH-PPV Composite Material 118 5.1 E E L S Samples 120 5.2 M E H - P P V L o w - L o s s Spectra and Z e r o - L o s s Peak R e m o v a l 121 5.3 Porous A l u m i n a L o w - L o s s Spectra 124 5.4 Porous A l u m i n a / M E H - P P V Compos i t e Spectra , 129 5.5 C o n c l u s i o n 132 Exper imenta l Detai ls .( 132 References 134 CHAPTER 6 Aloof Cherenkov Effect in Porous Alumina 135 6.1 The Cherenkov Effect 135 6.2 Further Measurements 139 6.3 Effect o f P r imary B e a m Energy 139 v 6.4 M o d e l l i n g o f the A l o o f Cherenkov Effect 144 6.4.1 M o d e l l i n g o f 197 k e V Data 145 6.4.2 M o d e l l i n g o f 117 k e V Data 149 6.4.3 C o m p a r i s o n w i t h Exper iment 149 6.5 Effect o f Pore Diameter o n the Cherenkov Peak 151 6.6 C o n c l u s i o n 156 Exper imenta l Detai ls 157 References 158 CHAPTER 7 Polymer Guest Incorporation 160 7.1 Internal P o l y m e r Synthesis 160 7.1.1 Surface-Graft Po lymer i za t ion o f Conjugated Po lymers 161 7.1.2 Der iva t iza t ion o f Porous S i l i c o n Surfaces 165 7.1.3 Wadswor th -Horne r -Emmons Reac t ion on S i l i c o n Surface 167 7.2 Externa l P o l y m e r Synthesis 169 7.2.1 Po lymer A d s o r p t i o n L o a d i n g 170 7.2.2 V a c u u m (Fil trat ion) L o a d i n g 170 7.2.3 Centr i fugal L o a d i n g 171 7.3 Preparation o f Centr i fuged Samples 172 7.4 Character izat ion o f Centr i fuged Samples 174 7.4.1 Scanning Elec t ron M i c r o s c o p y 174 7.4.2 Scanning Elec t ron M i c r o s c o p y 174 7.4.3 Transmiss ion Elec t ron M i c r o s c o p y 178 7.4.4 Energy-f i l tered Transmiss ion Elec t ron M i c r o s c o p y 180 7.4.5 S T E M / E E L S 183 7.5 C o n c l u s i o n : 184 Exper imenta l Detai ls 185 References 189 CHAPTER 8 Comments on Device Fabrication 193 8.1 D e v i c e Structure 193 8.2 Dev ices from Porous A l u m i n a F i l m s on A l u m i n u m F o i l 195 8.3 Dev ices from T h i n F i l m s on Conduc t ing Substrates 198 8.3.1 A l u m i n u m F i l m Depos i t i on 199 8.3.2 Porous A l u m i n a / S i l i c o n 200 8.3.3 Porous A l u m i n a / I T O 201 8.3.4 Porous A l u m i n a / G o l d 203 8.4 C o n c l u s i o n 203 Exper imenta l Detai ls 204 References 205 v i List of Tables Table 1.1 E lec t r i ca l conduct iv i ty o f po lymers and some c o m m o n materials 6 Table 1.2 Properties o f the ideal host mater ia l for conjugated polymers 13 Table 2.1 Current spatial resolut ion l imi ts o f chemica l analysis techniques 38 Table 3.1 I U P A C classif icat ion o f porous materials b y pore size 52 Table 3.2 Parameters for ca lcula t ion o f P P V mass fraction i n M C M - 4 1 72 Table 3.3 Po lymer mass fraction F for Npolymer chains per pore i n M C M - 4 1 72 Table 4.1 Pub l i shed parameters for self-ordered porous a lumina growth 103 Table 6.1 Cherenkov condi t ion for c o m m o n T E M beam energies 138 Table 6.2 Cherenkov loss peak parameters for ordered samples 152 List of Figures Figure 1.1 Butadiene, a s imple conjugated molecule 3 Figure 1.2 Conjugated polymers 5 Figure 1.3 Structure o f a s imple conjugated polymer-based electroluminescent device 8 Figure 1.4 Elect roluminescence processes i n a s imple po lymer device 9 Figure 1.5 Ideal device structure consis t ing o f oriented and isolated po lymer chains 12 Figure 1.6 Incorporation o f M E H - P P V into oriented mesoporous s i l i ca 16 Figure 1.7 (a) Structure o f l i q u i d crystal mesogen, (b) structure o f lyotropic l i q u i d crystal w i t h po lymer guest 18 Figure 1.8 High-temperature convers ion o f a water-soluble precursor po lymer to P P V 18 Figure 1.9 Water-soluble P P V derivat ive 20 Figure 1.10 Preparation o f M E H - P P V / C l a y composi te mater ial 20 Figure 1.11 Encapsulat ion o f conjugated po lymer chain w i t h (3-cyclodextrin 22 Figure 2.1 Truncated p y r a m i d geometry o f epoxy-embedded sample for sectioning b y u l t ramicro tomy 36 Figure 2.2 The smal l angle cleavage ( S A C ) technique appl ied to a thin f i l m 36 Figure 2.3 E lec t ronic exci ta t ion and de-excitat ion mechanisms i n a so l id 38 Figure 2.4 Geomet ry o f (a) elastic, (b) inelastic, inner-shell and (c) inelastic, outer-shell scattering events i n v o l v i n g an electron and a carbon atom 41 Figure 2.5 P r i n c i p a l features o f an E E L spectrum 42 Figure 2.6 Schematic o f pos t -co lumn Gatan Imaging Fi l ter on a T E M 44 Figure 2.7 Illustration o f chemica l analysis o f a two-component sample by S T E M / E E L S over a set o f points and E F T E M over the who le image 46 Figure 3.1 C h e m i c a l structure o f po ly ( l ,4 -pheny lene v iny lene) ( P P V ) 52 Figure 3.2 Examples o f pore topologies w i t h (a) 1-D, (b) 2 - D and (c) 3 - D connect ivi ty 53 Figure 3.3 Stages i n the formation o f M C M - 4 1 54 Figure 3.4 Transmiss ion electron micrograph o f M C M - 4 1 material obtained us ing Ci6H33(CH3)3NCl surfactant, showing hexagonal lattice spacing and w a l l thickness 55 Figure 3.5 N i t rogen (o ) adsorption and (•) desorption isotherms for M C M - 4 1 58 Figure 3.6 The G i l c h route to P P V starting from dichloro-p-xylene . 61 Figure 3.7 Synthetic scheme for the preparation o f P P V / M C M - 4 1 hyb r id mater ia l . . 62 Figure 3.8 (a) Thermogravimetr ic analysis o f T B A O H - t r e a t e d M C M - 4 1 , (b) nitrogen adsorption isotherm and B J H pore distr ibution for empty and T B A O H - t r e a t e d M C M - 4 1 64 Figure 3.9 (a) Powder X - r a y diffraction pattern and (b) neutron scattering data for M C M - 4 1 , P P V / M C M - 4 1 sample 1 and sample 2 65 Figure 3.10 (a) Thermogravimet r ic analysis o f P P V / M C M - 4 1 , and (b) B J H pore distr ibution o f empty M C M - 4 1 , P P V / M C M - 4 1 sample 1 and sample 2 67 Figure 3.11 Spectral signature o f P P V / M C M - 4 1 and M C M - 4 1 b y (a) F T - I R and (b) E E L S 68 Figure 3.12 E F T E M images for (a) M C M - 4 1 and (b) P P V / M C M - 4 1 69 Figure 3.13 (a) R o o m temperature U V / V i s absorbance and photoluminescence o f P P V / M C M - 4 1 composite, (b) temperature-dependent photoluminescence spectra o f P P V / M C M - 4 1 70 Figure 4.1 2 - D hexagonal p a c k i n g o f surfactant micel les i n aqueous solution onto a graphite surface 86 Figure 4.2 Surfactant templates for (a) M C M - 4 1 and M C M - 4 8 structures (Ci6H 33N(CH 3)3X) and (b) S B A - 2 structure ( C i e H s s N C C H s M C ^ N C C H s ^ X : ) , where X = B r or C l 87 Figure 4.3 The two extreme possibi l i t ies for surfactant micel le al ignment i n a porous support: (a) i n the axia l d i rec t ion and (b) i n the circumferential d i rect ion 88 Figure 4.4 Orientat ion o f mesoporous channels b y electro-osmotic f l ow 90 Figure 4.5 X - r a y diffraction pattern o f S B A - 2 f i lm g r o w n on m i c a before and after calcinat ion. 93 Figure 4.6 Preparation o f th in f i l m w i t h oriented channels from a FeO:Si02 film 95 Figure 4.7 T E M image o f cross-section o f FeO:Si02 film on glass (a) after oxidat ion, before etching, (b) after etching and Pb-staining. The cross-section was prepared by the S A C technique 96 Figure 4.8 E D X spectrum obtained from cross-sections o f F e O : S i 0 2 f i lms: (a) o x i d i z e d (some A u contaminat ion is apparent), (b) etched to remove Fe203, (c) Pb-stained 98 Figure 4.9 S E M images o f porous a lumina f i lm produced at 40 V i n 0.3 M oxal ic ac id , us ing Masuda ' s two-step approach: (a) top v i e w , (b) cross-section 99 Figure 4.10 A simple e lectrochemical c e l l for anodizat ion o f a l u m i n u m substrates 100 Figure 4.11 Structure o f porous a lumina f i lms g r o w n on a l u m i n u m : (a) geometry o f pore pack ing , (b) cross-section showing barrier layer o f thickness ~ L / 2 at bot tom o f pores 101 Figure 4.12 Preparation o f ful ly-ordered porous a lumina f i l m by two-step anodizat ion 104 Figure 4.13 Porous a lumina samples anodized at (a) 20 ° C , (b) -8 ° C and (c) -40 ° C at 15.0 V i n 1.2 M H 2 S 0 4 (1:1 F f 2 0 : M e O H ) , w i t h result ing pore size distr ibutions 106 Figure 4.14 Effect o f temperature on pore diameter for samples anodized at 15.0 V i n 1.2 M sulfuric ac id 107 Figure 5.1 Adso rp t ion o f thin layer o f M E H - P P V onto porous a lumina host, showing (a) empty host, and polymer-coated host i n (b) p lan v i e w and (c) as thin section for T E M 119 Figure 5.2 Porous a lumina film after soaking i n M E H - P P V solution, seen i n cross-section, as shown by (a) l ight mic roscopy (b) fluorescence mic roscopy 119 Figure 5.3 S E M image o f cross-section o f sample for E E L S experiments 121 Figure 5.4 T E M image o f M E H - P P V film supported by a lacey carbon g r id 122 Figure 5.5 Low- los s spectrum o f M E H - P P V thin film, also showing var ious approaches to r e m o v i n g the zero-loss peak: raw data, matr ix deconvolut ion, Four ier - log deconvolut ion, power law fit over 1.3 - 2 . 0 e V 123 Figure 5.6 L o w - l o s s spectrum o f pore i n porous a lumina film, showing collected data and data w i t h zero-loss peak removed by a power l aw fit over 1.3 to 2.0 e V 125 Figure 5.7 S T E M dark-f ie ld image o f 1.4 u m thick porous a lumina film, showing the locat ion o f l ine a long w h i c h loss spectra were acquired at 2 n m intervals 127 v i i i ( Figure 5.8 Representative low-loss spectra for porous a lumina f i l m : near pore centre, ~7 n m from pore w a l l , and at w a l l 128 Figure 5.9 Energy-f i l tered images o f 0.2 u\m porous a lumina film 128 Figure 5.10 (a) S T E M dark-f ield image o f porous a l u m i n a / M E H - P P V composite , showing the l ine a long w h i c h E E L spectra were acquired 130 Figure 5.11 C o m p a r i s o n o f low-loss spectra near pore axis and at ~7 n m from the pore w a l l , for porous a lumina and porous a l u m i n a / M E H - P P V composite , respect ively 131 Figure 5.12 C o m p a r i s o n o f low-loss spectra o f empty and M E H - P P V - t r e a t e d porous a lumina, nearest to pore w a l l 131 Figure 6.1 Geomet ry o f Cherenkov radiation due to an electron t ravel l ing (a) through a m e d i u m and (b) near a m e d i u m w i t h dielectric function e(co) 137 Figure 6.2 Die lec t r ic function o f a lumina 137 Figure 6.3 (a) S T E M dark-f ield image, showing pore used for E E L S analysis at 197 k e V (b) E E L S spectra acquired over the pore diameter w i t h (3 = 0.34 mrad 141 Figure 6.4 (a) S T E M dark-f ield image, showing pore used for E E L S analysis at 117 k e V (b) E E L S spectra acquired over the pore diameter w i t h (3 = 1.5 mrad 142 Figure 6.5 C o m p a r i s o n o f low-loss spectra at 197 and 117 k e V pr imary beam energies: (a) ax ia l (5 = 0 nm) and intermediate (s = 22 nm) spectra, and (b) wa l l -g raz ing (s = 28 nm) spectra 143 Figure 6.6 C o m p a r i s o n o f (a) experimental electron beam w i t h convergence angle a and (b) theoretical mode l 145 Figure 6.7 Theoret ical loss distr ibution for 200 k e V electrons t ravel l ing ( M o d e l A ) d o w n a single pore i n a lumina, and ( M o d e l B ) d o w n a cy l indr ica l pore o f outer rad i i 61 , 94 and 127 n m . The inner pore radius is 29 n m 147 Figure 6.8 Theoret ical loss dis tr ibut ion for 200 k e V electrons t ravel l ing d o w n a cy l ind r i ca l hole w i t h 0, 6 and 12 neighbour ing pores ( M o d e l C ) , showing the loss probabi l i ty for (a) axia l (5 = 0 n m ) and (b) near-wal l (s = 27 nm) trajectories. The pore radius is 29 n m and the spacing is 90 n m 148 Figure 6.9 Effect o f electron veloci ty on the loss probabi l i ty function illustrated us ing M o d e l B .150 Figure 6.10 C o m p a r i s o n between theoretical and experimental spectra at different impact parameters s: (a) 0 n m , (b) 22 n m , (c) 28 n m , us ing the 1+6 cyl inder mode l (radius 29 n m , spacing 90 nm) 150 Figure 6.11 S T E M dark field images showing geometry o f ordered porous a lumina membranes produced b y a two-step anodizat ion at 40 V i n 0.3 M oxal ic acid: (a) 62 n m , (b) 74 n m , (c) 84 n m diameters; (d) lower magnif ica t ion image showing size o f ordered domains 152 Figure 6.12 Cross-sect ion o f ordered porous a lumina membrane 153 Figure 6.13 (a) Cherenkov peak shift for a fixed lattice spacing (105 nm) w i t h different diameters: 62 n m , 74 n m , and 82 n m at 197 k e V and (o) 62 n m at 117 k e V ; (b) losses d o w n to 2 e V revealed by subtraction o f the zero-loss peak (diameter 62 n m , 197 k e V ) 155 Figure 7.1 Surface-init iated anionic po lymer iza t ion o f M E H - P P V 163 Figure 7.2 Step po lymer iza t ion o f surface-grafted P P V b y the W H E reaction 164 Figure 7.3 Porous s i l i con der ivat izat ion b y cathodic electrografting 164 Figure 7.4 Protect ion o f aldehyde i n 4-ethynylbenzaldehyde. „ .'. 165 Figure 7.5 Synthetic route to conjugated d imer on porous s i l i con surface 166 Figure 7.6 F T - I R spectra after (a) cathodic electrografting, (b) deprotection o f aldehyde and (c) W H E reaction on porous s i l i con substrate 168 ix Figure 7.7 Photoluminescence o f derivat ized porous s i l i con : excitat ion spectrum and emiss ion spectrum o f W H E reaction product; emiss ion o f deprotected aldehyde 168 Figure 7.8 C h e m i c a l structure o f poly[2-methoxy,5-(2 ' -e thylhexyloxy)- l ,4-phenylene vinylene) ( M E H - P P V ) 169 Figure 7.9 (a) V a c u u m - d r i v e n po lymer infil tration into a porous membrane, (b) S E M image o f po lymer i n 200 n m pores o f A n o p o r e membrane 171 Figure 7.10 Centr i fugal po lymer load ing into a porous a lumina f i l m 172 Figure 7.11 (a) Centrifuge rotor assembly w i t h two substrate holders, (b) detail o f substrate holder f rom above and (c) f rom inside, showing O - r i n g seal 173 Figure 7.12 S E M image o f cross-sections o f centrifuged samples, showing (a) po lymer over layer and (b) some po lymer penetration into pores 175 Figure 7.13 A n a l y s i s o f po lymer penetration into porous a lumina b y X P S and S E M 175 Figure 7.14 S E M images o f centrifuged samples, observed from bottom w i t h the host etched away : 176 Figure 7.15 X P S results from bot tom o f empty porous a lumina and centrifuged sample 177 Figure 7.16 T E M images o f thin sections o f (a) empty porous a lumina host and (b) centrifuged sample ; 179 Figure 7.17 (a) T E M image and (b) E E L S o f th in section o f empty porous a lumina host: zero-loss peak, porous a lumina before and after zero-loss peak subtraction, a l u m i n u m 180 Figure 7.18 Unf i l te red ( T E M ) and energy-filtered (5, 25 e V ) images o f empty porous a lumina and centrifuged M E H - P P V / p o r o u s a lumina composite 182 Figure 7.19 (a) S T E M image o f thin section o f centrifuged sample. The contrast has been increased i n the inset to show the polymer tubules, (b) E E L S associated w i t h d r i l l i ng i n a lumina and a po lymer tubule 184 Figure 8.1 Ideal conjugated po lymer device components and their assembly: (a) po lymer insertion into host, (b) cathode evaporation and (c) anode deposi t ion 194 Figure 8.2 D e v i c e fabrication sequence from porous a lumina f i l m on a l u m i n u m fo i l : (a) anodizat ion o f a l u m i n u m fo i l , (b) barrier layer th inning by potential reduction, (c) po lymer introduct ion by centrifugation, (d) I T O deposit ion b y R F sputtering, (e) contact lead bond ing w i t h si lver epoxy and epoxy embedding o f upper surface, (f) a l u m i n u m fo i l r emova l b y chemica l etching, (g) a l u m i n u m cathode deposi t ion b y thermal evaporation. 196 Figure 8.3 S E M images o f porous a lumina f i l m on a n-type s i l i con wafer w i t h the barrier layer e l iminated b y the potential reduct ion method 197 Figure 8.4 S E M images o f porous a lumina f i l m , showing e l iminat ion o f barrier layer by rapid potential reduct ion 197 Figure 8.5 D e v i c e fabrication sequence from an a l u m i n u m f i lm on a conduct ive substrate: (a) in i t ia l a l u m i n u m f i l m on substrate, (b) porous a lumina film growth, (c) f inal porous a lumina f i l m w i t h barrier layer removed , (d) complete device w i t h conjugated po lymer sandwiched between cathode and anode 198 Figure 8.6 S E M images o f h i l locks on porous a lumina f i lms prepared from R F sputtered a l u m i n u m f i lms, showing (a) top surface and (b) cross-section 199 Figure 8.7 Defects i n porous a lumina films anodized from electron-beam evaporated a l u m i n u m f i l m , showing (a) S E M image o f cross-section o f a film anodized at 20 ° C i n 1.2 M sulfuric ac id and (b) T E M image i n plan v i e w o f a film anodized at -39 ° C 201 Figure 8.8 S E M image o f porous a lumina f i l m on ITO-coated glass substrate, showing an area where most o f the I T O was consumed 202 List of Abbreviations and Symbols Abbreviation/Symbol Description 0 degree, unit of angle a pore radius, n m a beam convergence semi-angle A Angstrom, unit o f length, 1 0 " 1 0 m A M P E L A d v a n c e d Mater ia ls and Process Engineer ing Laboratories P spectrometer col lect ion semi-angle p ratio o f particle speed to the speed o f light i n v a c u u m B E T Brunauer-Emmet t -Tel ler B J H Barret-Joyner-Halenda c speed o f light in vacuum, 3.00 x 10 8 m s"1 C C o u l o m b , unit o f charge ° C degrees Cels ius , unit o f temperature C C D charge-coupled device c m centimetre, unit o f length, 10"2 m d pore spacing dhki interplanar spacing D pore diameter D C direct current e dielectric function Ci real part o f the dielectric function £ 2 imaginary part o f the dielectric function E D X energy dispersive x-ray spectroscopy E E L S electron energy loss spectroscopy E F T E M energy-filtered transmission electron mic roscopy e.g. (exempl i gratia, Lat in) for example et al. (et a l i i , Lat in) and co-workers etc. (et cetera, La t in) and others e V electron-Volt , unit o f energy, 1.602x 10" 1 9 J ex situ (Latin) outside, external F I B focused ion beam F T - I R Fourier transform infrared F W H M full w i d t h at h a l f - m a x i m u m g gram, unit o f mass G I F Gatan imag ing filter h hour, unit o f t ime, 3600 s i.e. ( id est, Lat in) as i n in situ (Latin) inside, internal I R infrared I T O Ind ium T i n O x i d e I U P A C International U n i o n o f Pure and A p p l i e d Chemis t ry J Joule, unit o f energy, k g m 2 s"2 K degrees K e l v i n , unit o f temperature k e V ki lo-e lec t ron-Vol t , 1.602x10" 1 6 J k g k i log ram, unit o f mass, 10 3 g A, wavelength L pore spacing L E D l ight-emit t ing device m metre, unit o f length M molar , unit o f concentration, m o l L " 1 M C M M o b i l C o m p o s i t i o n o f Matter M E H - P P V poly[2-methoxy,5-(2 ' -ethylhexyloxy)-1,4-phenylene vinylene] m i n minute, unit o f t ime, 60 s m m o l m i l l i m o l , 1 0 " 3 m o l m o l mole, constant o f value 6.02 x 1 0 2 3 mrad mi l l i rad ian , 10"3 rad ms mi l l i second , unit o f t ime, 1 0 3 s N newton, unit o f force, k g m s"2 n - D n-d imens ional n m nanometre, unit o f length, 10~9 metre p pressure p0 standard pressure x i i p H hydrogen ion concentration on logar i thmic scale, - log io [H + ] P L photoluminescence N M R nuclear magnetic resonance P P V poly( 1,4-phenylene vinylene) dc Cherenkov angle Q magnitude o f scattering vector rad radian, unit o f angle, % rad = 1 8 0 ° ref. reference R F radio-frequency R I E reactive ion etching R P M revolutions per minute s second, base unit o f t ime s impact parameter, n m S A C T small-angle cleavage technique sat saturated solut ion S B A Santa Barbara seem surface cubic centimetre per minute, unit o f f low rate S E M scanning electon mic roscopy S T E M scanning t ransmission electron microscopy S T M scanning tunnel l ing microscopy t thickness o f adsorbed layer T Tesla , unit o f magnetic flux density T B A tetrabutyl a m m o n i u m T E M transmission electron mic roscopy T G A thermogravimetr ic analysis T H F tetrahydrofuran Tor r unit o f pressure, 1 Tor r = 1 m m H g = 133 k g m" s" U V / V i s ul traviolet /visible V V o l t s , J C" 1 X P S x-ray photoelectron spectroscopy c o particle energy i n electron-Volts wt. weight x i i i Acknowledgements I w o u l d l ike to acknowledge foremost m y supervisor, Prof. M i k e W o l f , for a l l the help and support he p rov ided dur ing the course o f this thesis. M y fe l low Canadians have also supported me generously through the Na tu ra l Sciences and Engineer ing Research C o u n c i l . Th i s w o r k drew on the expertise o f many talented staff members , students and professors at U B C . and elsewhere. A number o f graduate students helped m y w o r k i n important ways. The members o f the W o l f group were generous and tolerant o f m y ways, and I w o u l d l ike to thank D r . O l i v i e r C lo t and D r . Cerr ie Rogers especially for tutoring m e on the finer points o f organic synthesis. D r . Kat ja Rademacher , K e r i K w o n g , G l e n K u r o k a w a and Josh E d e l also w o r k e d on s imi lar projects and their collaboration was very useful. Dar ren B r o u w e r and J i m Sawada from the Fyfe group helped w i t h the analysis o f m y porous materials. A m o n g the departmental staff, I w o u l d l ike to thank especial ly the m e c h shop staff -B r i a n Snapkauskas, C e d r i c N e a l , Des Lovr i t y , R o n M a r w i c k , K e n L o v e and R a z v a n N e a g u - for their w o r k on m y project and their friendship. B r i a n D i t chbu rn p rov ided custom glassware wi th superior helpfulness. A n d r e W o n g , from the Facul ty o f Dentistry, taught m e u l t ramicrotomy o f hard materials and helped w i t h the very early electron mic roscopy work . M a r y Mager , from Meta l s and Mater ia ls Engineer ing , p rov ided s imi lar help w i t h sample preparation and electron microscopy . E la ine H u m p h r e y and Garnett Huygens , from the B i o i m a g i n g facil i ty, ran a great facil i ty w i t h fabulous microscopes and p rov ided lots o f advice as w e l l . Jeff Y o u n g and M u r r a y W . M c C u t c h e o n , in the Phys ics Department, carr ied out the variable-temperature luminescence studies reported i n chapter 3. The most important results o f this thesis can a l l be traced back to the expert help o f Prof. G i a n l u i g i Bot ton , w i t h w h o m I started the w o r k on electron energy-loss spectroscopy. The in i t ia l x i v help he p rov ided p roved invaluable for comple t ing this work . Further on, R i c h a r d H u m p h r e y at the Un ive r s i t y o f Calgary helped w i t h measurements, and Prof. K a r e n K a v a n a g h at S i m o n Fraser Unive r s i ty gave access to her new microscope for the most important measurements. The investigation into the Cherenkov effect was helped immense ly by a collaboration w i t h Profs . A lbe r to R ivacoba , F . Javier G a r c i a de A b a j o and Nerea Zaba la at E u s k a l H e r r i k o Unibersi tatea i n Spain . The i r theoretical insight into the p rob lem guided the investigation and established the signif icance o f the results. O n the fabrication side, A l i n a K u l p a , D o u g W o n g and J i m M a c k e n z i e assisted i n various ways. I part icularly appreciated the great help A l Schmalz provided , and his friendship just as m u c h . D r . K e n W o n g and D r . P h i l i p W o n g carried out a l l the X - r a y photoelectron spectroscopy on m y samples. Last ly, I w o u l d l ike to acknowledge the Green Col lege commun i ty for three ful f i l l ing (and thesis-delaying!) years, w h i c h have marked me as m u c h as any experience i n m y life. x v CHAPTER 1 Introduction The three challenges o f materials chemistry are synthesis, characterization and understanding. A l l o f these have seen rapid progress in the past decades, largely due to advances in experimental methods and instrumention for investigating materials on the nanometre scale. Th i s fol lows from the importance o f interactions on this scale: the properties o f al l materials, from atom clusters to the bulk , whether homogeneous or not, are governed by interactions on the nanometre level . Unders tanding and exploi t ing these interactions has thus become centrally important in materials chemistry. Consequently, the development o f new tools for structural characterization on this scale is o f equal importance. F o r crystalline materials, the tools for complete characterization on the atomic level have been available for a long t ime through X - r a y diffraction. Y e t this technique only provides an average picture o f the material and may overlook structural features w h i c h may dominate material function, such as interfaces and defects. Furthermore, amorphous materials and materials w i t h structure on a different scale are also excluded from investigation by diffraction. Structural analysis o f such materials can only be done by h igh resolution microscopy, and this area remains i n development for organic materials. T h i s thesis describes the preparation o f composi te materials consis t ing o f conjugated polymers w i th in mesoporous hosts. In this case, the material properties o f interest - charge transport and luminescence i n the conjugated po lymer - cou ld not be investigated further without first establishing the detailed structure o f the composi te . A s this structure var ied on the 1 nanometre scale, the characterization was not t r iv ia l and the development o f the appropriate techniques became the focus o f this work . In a larger context, this reflects the need for proper material characterization before property studies can be undertaken. The methods fo l lowed in this w o r k should be readily applicable to other nanostructured systems wi th conjugated po lymer components. In this chapter, the properties o f conjugated polymers are introduced along wi th a descript ion o f their m a i n applicat ion i n l ight-emitt ing devices. The mot iva t ion for in t roducing such polymers into porous host materials is discussed, and the literature on this developing field is rev iewed. F ina l ly , the scope o f this thesis is presented along w i t h a short summary o f each chapter. 1.1 Conjugated Polymers: Organic Conductors and Semiconductors The optical and electrical properties o f molecular crystals o f conjugated molecules have been the subject o f study since the 1940 ' s , 1 w h i l e the study o f conjugated polymers began i n the 1970's . Organic polymers w i t h an electron system delocal ized over the complete chain are o f interest for a number o f reasons, inc lud ing the novelty o f charge transport in polymers , theoretical interest in 1-D conductors, the useful mechanica l properties o f po lymer ic materials, and their s imple and inexpensive processing. V e r y significant breakthroughs have occurred i n the study o f conjugated polymers over the past 30 years. These polymers have been used as the active material in a number o f applications, inc lud ing sensors , 2 ' 3 t ransistors, 4 " 6 l ight-emit t ing dev ices , 7 " 9 photovoltaic d e v i c e s , 1 0 and l a s e r s . 1 1 ' 1 2 Electroact ive polymers are also p lay ing a r o l e 1 3 i n the development o f molecular e l e c t r o n i c s . 1 4 ' 1 5 1.1.1 Organic Molecules with Conjugated 7t-Electron Systems In many organic molecules , interesting and useful electrical properties arise from the presence o f delocal ized 7t-electron systems. These systems are referred to as conjugated %-electron systems. F r o m the organic chemis t ' s point o f v i ew, such a system is recognized in a chemica l structure as a sequence o f alternating single and double bonds between carbon atoms (Figure 1.1(a)). Electrons on other atoms, such as oxygen, sulfur and nitrogen, can also participate in a conjugated system. F r o m the phys ica l chemist ' s point o f v i e w , a conjugated n -electron system can be defined as a set o f adjacent, parallel , half-fil led p electron orbitals on a molecule (Figure 1.1(b)). In the framework o f molecular orbital theory, these p orbitals combine to form molecular orbitals that are spread over the chain o f interacting atoms (Figure 1.1(c)). A r o m a t i c molecules are a special case where 4 N + 2 conjugated 7i-electrons form a r ing , w h i c h imparts addit ional stability to the molecule . The highest occupied molecular orbital ( H O M O ) is a n bonding orbital , and the lowest unoccupied molecular orbital ( L U M O ) is a n ant ibonding (or 7t*) orbital . In large conjugated electron systems, the bond ing orbitals have a quasi-continuous set o f energy levels, w h i c h is referred to as the valence band. Ana logous ly , the ant ibonding orbitals form the conduct ion band. M o s t importantly, the electrons i n the conjugated system are delocal ized over the extent o f the Figure 1.1 Butadiene, a s imple conjugated molecule: (a) chemica l structure, (b) p-orbi tals fo rming conjugated system, (c) lowest energy (o f four) 7t molecular orbital . 3 conjugated bonds. A conjugated molecule may therefore act as a pathway for charge transport i f a charge carrier is introduced into the H O M O or L U M O . Th i s can be accompl ished by either o x i d i z i n g or reducing the molecule , either chemica l ly (also k n o w n as doping) or electrochemically. 1.1.2 Conjugated Molecules in the Solid State The electrical properties o f a material depend largely on the distr ibution o f available energy states above the highest occupied state. T h i s is the energy difference between the conduct ion band and the valence band, w h i c h determines the amount o f energy required to promote an electron f rom the H O M O to the L U M O . The size o f the band gap is used to classify a material as a conductor, semiconductor or insulator. In conjugated molecules , the band gap is determined by the size o f the conjugated system, and in most cases it falls in the range o f semiconductors. The electronic properties o f conjugated molecules in the sol id state (i.e., as a material) are altered by the effects o f interactions between neighbour ing molecules. E lec t ronic processes, such as charge transport, are then a combina t ion o f intra- and intermolecular processes. Fo r smal l molecules , intermolecular processes are necessarily important. Fo r larger molecules , in particular polymers , the relative importance o f these processes varies from material to material . 1.1.3 Organic Conjugated Polymers Since the 1970's , molecules w i t h long extensions o f conjugated 71-electrons have been the subject o f scientific pursuit, as a result o f interest in their electrical and opt ical properties. The concept o f organic molecules as molecular wires has seen significant d e v e l o p m e n t . 1 6 ' 1 7 In 4 H A B C O R F G F i g u r e 1.2 Conjugated polymers : (a) polyacetylene, (b) polypyrrole , (c) polythiophene, (d) polyani l ine , (e) po ly( l ,4 -phenylene) , (f) po ly( l ,4 -phenylene v inylene) ( P P V ) , (g) po ly[2-methoxy,5-(2 ' -e thylhexyloxy)- l ,4-phenylene vinylene] ( R = M e , R ' = 2-ethylhexyl) ( M E H -P P V ) . particular, polymers w i t h conjugated u-electron systems extending along the whole length have been investigated. These are semiconductors in the prist ine state. The simplest such polymer , polyacetylene (Figure 1.2(a)), was reported to be a conductor when doped by Shi rakawa, Heeger and M a c D i a r m i d in 1 9 7 7 . 1 8 The doped po lymer exhibits an electrical conduct ivi ty that can reach metal l ic levels (Table 1.1). The conduct ivi ty o f polyacetylene was an enormously important scientif ic d iscovery that earned the N o b e l P r i z e for chemistry i n 2000 . The physics associated w i t h the excited states on the po lymer chain also proved to be very r i c h . 1 9 The incorporation o f aromatic subunits (benzene, thiophene, pyrrole, etc. - see Figure 1.2) a l lows control over the structure and properties o f the polymer . These aromatic subunits may be chemica l ly modi f i ed to adjust their electrical and optical p roper t i e s , 2 0 and to impart other 5 T a b l e 1.1 Elec t r ica l conduct ivi ty o f polymers and some c o m m o n materials. Conduc t iv i ty P o l y m e r Other Mater ia ls (S c m ' 1 ) 10 5 doped polyacetylene, or copper, i ron 1 doped polypyrrole , graphite, polythiophene, etc. doped s i l i con 10" 5 fr-arcs-polyacetylene i nd ium, t in , s i l icon ew-polyacetylene water 1 0 - i o d iamond l O " 1 5 polythiophene, polypyrrole ny lon 1 0 - 2 0 Tef lon quartz desirable properties such as solubil i ty. The incorporation o f metal centres into the po lymer chain is also an area o f active research but falls outside the scope o f this thesis. In general, the incorporat ion o f functional units that interact w i t h the conjugated 7r-system can be used to make the conduct ivi ty sensitive to the presence o f external s t i m u l i . 2 ' 3 ' 6 T h e structures o f conjugated polymers in solut ion and in the bulk can vary significantly. O n short length scales, the chain structure is usually planar for op t imal conjugation. O n a longer scale, the po lymer may deviate significantly from the ideal o f a molecular wi re and exist i n a coi led-up configuration. T h i s structure is diff icult to determine i n the so l id state but may be investigated in solution by dynamic light-scattering measurements . 2 2 The conduct ivi ty o f doped polymers was found to degrade significantly through environmental e x p o s u r e . 2 3 ' 2 4 A s undoped semiconductors, they are somewhat more stable and this area has become the m a i n focus o f w o r k in the past decade. 6 1.2 Luminescence in Conjugated Molecules The process o f light emiss ion by a molecule in an excited state is termed luminescence, and many conjugated molecules are h ighly luminescent. The excited state may be a spin singlet or a spin triplet, whi le the ground state is normal ly a singlet in organic molecules . The transition from the excited state to the ground state can occur through both radiative and non-radiative processes. A radiative decay process in w h i c h spin angular momentum is conserved (also cal led a spin-a l lowed transition, e.g. between two singlet states or two triplet states) is termed fluorescence. Th i s is a rapid process, w i t h a typical lifetime on the nanosecond scale. A sp in-forbidden radiative transition (e.g. from a triplet to a singlet state) is m u c h slower (on the microsecond scale) in organic molecules and is called phosphorescence. The singlet ground state causes fluorescence to be the normal emiss ion process in organic molecules. A typica l non-radiative decay process takes the molecule to a very h ighly excited vibrat ional level o f the ground state, and the excess vibrat ional energy is dissipated eventually as heat. Th i s mechanism may dominate i f the radiative pathway is s low; for this reason, phosphorescence is not usually observed in organic conjugated molecules . The ini t ia l excited state can be generated in a number o f different ways , and this is used to dis t inguish different types o f luminescence: photoluminescence from optical excitation, electroluminescence from electrical excitation, and chemiluminescence from chemica l reaction. These modes o f excitation may each be important for different applications; they may also differ significantly in the number o f singlet and triplet states that they ini t ia l ly create. F o r conjugated polymers , the most important o f these processes is electroluminescence. 7 1.3 Electroluminescence in Conjugated Polymers Electroluminescent materials are o f special importance in display technology. Electroluminescence is described as the emiss ion o f light f rom condensed matter under the action o f an electric field.25 T h i s phenomenon was reported for smal l conjugated organic molecules in 1 9 8 7 . 2 6 It was first reported for a conjugated polymer , po ly( l ,4 -phenylene vinylene) ( P P V ) , i n 1990 by Fr iend and c o - w o r k e r s . 2 7 A s imple polymer-based device is shown in Figure 1.3: the active po lymer layer is sandwiched between a metal l ic cathode and a transparent anode. W h i l e the performance o f these devices has been continously improved , the m a i n barrier to commerc ia l iza t ion has been device degradat ion. 6 Conjugated polymers are very susceptible to photo-oxidat ion; the ox id i zed molecules then serve as low-energy traps for electrons, w h i c h reduces the luminescence e f f i c i e n c y . 2 8 ' 2 9 Other routes to device degradation also e x i s t . 3 0 ' 3 1 1.3.1 Electroluminescence Processes in Conjugated Polymers The technological importance o f electroluminescence has made understanding the various electronic processes in these devices an important objective. The mechan i sm o f electroluminescence in polymers has been wide ly investigated and rev iewed in d e t a i l . 9 ' 1 9 ' 2 5 The Cathode: aluminum or calcium Polymer: -100 nm thick Anode: indium tin oxide on glass Light emission Figure 1.3 Structure o f a s imple conjugated polymer-based electroluminescent device. 8 Elect r ic field • Conduc t ion band ^ >, OA u» O C L i g h t m ^ ^ ^ ^ , \;\: • • Valence band A n o d e Po lymer Cathode F i g u r e 1.4 Electroluminescence processes in a s imple po lymer device. key processes, illustrated in Figure 1.4, are charge injection, transport, recombinat ion and de-excitation. U n d e r the action o f the applied electric field, the po lymer is ox id i zed at the anode ( ind ium tin oxide ( ITO) or gold) , w h i c h introduces a posi t ive charge by r emov ing an electron from the H O M O (valence band). Th i s species is referred to as a hole. A t the cathode, the po lymer is reduced, w h i c h places an electron into the L U M O (conduct ion band). The w o r k function o f the cathode material is chosen to match the po lymer L U M O level as closely as possible; a l u m i n u m and ca l c ium are used c o m m o n l y . I f there is substantial mismatch , the injection process does not proceed efficiently. It fol lows that the devices do not conduct under reverse bias. These charges migrate towards each other through the material due to the applied field through a combinat ion o f intrachain and interchain transport. W h e n a hole and an electron meet w i t h i n the polymer , they may recombine to create either a singlet or triplet excited state. A s discussed above, the singlet state w i l l usually decay radiatively, comple t ing the electroluminescence process. The triplet state usually decays non-radiatively; it may be 9 converted to a singlet state through the addi t ion o f a s e n s i t i z e r ; 3 2 ' 3 3 the triplet state energy may also be converted to light through p h o s p h o r e s c e n c e . 3 4 ' 3 5 The excited state may be local ized on one po lymer chain or spread over adjacent chains. In the latter case, it has been found that the radiative l ifetime is substantially l o n g e r , 2 2 ' 3 6 w h i c h can then favour non-radiative processes. 1.3.2 Electroluminescence Efficiency The power efficiency o f an electroluminescent device is very important for many applications, as it determines the brightness and power consumpt ion o f the device. Important progress has been made i n i m p r o v i n g this efficiency fo l lowing the discovery o f electroluminescence in conjugated p o l y m e r s . 9 M a n y factors affect the power efficiency, and here only the internal quantum efficiency is discussed. Th i s is defined as the number o f photons created per injected e lect ron. 9 The overal l internal quantum efficiency is determined pr inc ipa l ly by the photoluminescence y i e ld and the singlet y i e ld . The photoluminescence efficiency, w h i c h can be measured separately, places an upper l imi t on electroluminescence efficiency. Op t i ca l excitation o f the po lymer creates an excited state identical to the one produced by the charge recombinat ion process. S ince opt ical excitation creates only singlet states, this a l lows the relative importance o f radiative and non-radiative processes to be investigated. The presence o f interchain interactions is important in this respect, as an excited state wi th a s low radiative decay can be more susceptible to non-radiative quenching. In the device, the higher mobi l i ty o f holes relative to electrons in the po lymer causes recombinat ion to occur close to the cathode interface. T h i s interface is k n o w n to have many chemica l defects at w h i c h non-radiative decay can o c c u r , 3 7 ' 3 8 w h i c h reduces the photoluminescence efficiency relative to the pristine material . 10 In electroluminescence, the singlet y i e ld corresponds to the fraction o f excited states generated as singlets. The creation o f the non-radiative triplet states dur ing recombinat ion reduces this y ie ld . Based on s imple spin statistics, the singlet spin fraction is expected to be one quarter, and this was long bel ieved to be the l imi t i ng factor o f organic electroluminescent devices. However , some recent e x p e r i m e n t a l 3 9 " 4 2 and t h e o r e t i c a l 4 3 ' 4 4 results have indicated that the probabili t ies o f singlet and triplet formation may not be the same, and that singlet creation is favoured in po lymer ic materials (as opposed to smal l molecules) . Hence the singlet fraction may approach unity in conjugated polymers . M o r e sophisticated device structures provide higher efficiency: mult i - layer devices are used to eliminate the problem o f higher hole m o b i l i t y 4 5 and the injection processes can be improved by op t im iz ing the cathode m a t e r i a l 4 6 or by modif ica t ion o f the anode interface by ultrathin po lymer l a y e r s . 4 7 The complex chemistry o f evaporated metal contacts has l imi ted the understanding o f processes at the cathode in te r face . 4 6 1.3.3 An Ideal Device Structure The electronic processes in conjugated po lymer devices are difficult to study due to the amorphous nature o f the po lymer film. The disorder in these films gives rise to complicated 2-and 3 - D phenomena despite the 1-D nature o f the po lymer chain . It is then o f scientif ic interest to develop systems where interchain processes are el iminated, thereby a l l owing the intrachain properties to dominate. Th i s amounts to electrical isolation o f ind iv idua l po lymer chains. It is clear that complete isolation o f each chain w o u l d el iminate a l l charge transport, and therefore a functional device can only be achieved i f each po lymer chain is in electrical contact w i t h both electrodes (Figure 1.5). A n ordered and oriented encapsulant is necessary to achieve such a structure. A n ideal device design w o u l d also include chemica l ly wel l-def ined interfaces to 11 < 5 n m F i g u r e 1.5 Ideal device structure consis t ing o f oriented and isolated po lymer chains. the cathode and anode, such that the charge injection processes could be studied more r igourously. Th i s area has already seen important progress through the study o f conjugated molecules self-assembled on metal i n t e r f a c e s . 1 7 ' 4 8 The goal o f measur ing single po lymer chain electrical properties thus serves to introduce the theme o f encapsulation. A list o f desirable properties may be generated based on the premise that the host material should be present only to induce the desired ordering o f the conjugated po lymer guest and otherwise not interfere w i t h measurements being made on the guest (Table 1.2). In practice, it must be recognized that host-guest interactions cannot be entirely avoided, and may be difficult to account for. * 12 1.4 Encapsulated Conjugated Polymers Encapsula t ion can be generally defined as the preparation o f materials in w h i c h there is a reduced degree o f interaction between the guest molecule and its surroundings. There are two fundamental motivations behind the goal o f p roduc ing encapsulated conjugated polymers . First , as discussed above, the study o f single po lymer chains isolated in the so l id state is expected to reveal the fundamental photophysical and electrical behaviour o f the conjugated po lymer chain , without any effects due to aggregation w i t h other po lymer c h a i n s . 4 9 Second, a suitable encapsulant w o u l d provide protection against environmental agents, most importantly oxygen, w h i c h w o u l d a l low devices to operate for longer periods o f t ime under ambient condit ions. In this context, encapsulation impl ies the reduction o f chemica l and electrical interactions o f the po lymer chains. 1.4.1 Approaches to Encapsulation The photophysics o f isolated chains can be studied by dispers ing the po lymer in an inert Table 1.2 Properties o f the ideal host material for conjugated polymers . Property Basis 1. A l i g n e d channels normal to the substrate 2. N a r r o w pores (< 5 nm) 3. Op t i ca l transparency 4. E lec t r ica l insulator 5. C h e m i c a l l y inert 6. Eas i ly prepared as th in film (controlled thickness, defect-free) 7. G o o d "analytical contrast" A l l o w good electrical transport M i n i m i z e number o f po lymer chains per pore, ideally only one per pore (guaranteed i f diameter is < 2 nm) A l l o w optical characterization o f po lymer guest Measure only guest electrical properties, M i n i m i z e interchain processes M i n i m i z e interactions w i t h the po lymer guest A l l o w l ight-emitt ing device fabrication Ease o f characterization po lymer m a t r i x . 5 0 ' 5 1 However , this approach does not preclude effects due to interactions among 13 different segments o f the same chain, i f the po lymer is in a co i led configuration. Poss ible phase segregation o f the po lymer is also an issue in these materials, and cannot be excluded without detailed microstructural analysis o f the dispersed polymer . S imi la r ly , dilute po lymer solutions can be studied. In this way , the intrachain hole mobi l i ty has been measured for isolated po lymer c h a i n s 5 2 and the effect o f interchain interactions on photophysical behaviour has been d e t e r m i n e d . 2 2 These approaches to encapsulation are useful but afford a degree o f encapsulation w h i c h prevents direct electrical measurements on the isolated chains. Furthermore, a degree o f disorder remains on the level o f the po lymer chain conformation. Fo r this reason, novel approaches to encapsulation are desirable and one can look to ordered porous materials. M a n y o f the requirements o f Table 1.2 can be satisfied by porous materials based o n s i l i ca and a lumina . 1.4.2 Ordered Porous Materials as Encapsulants The synthesis o f conjugated polymers w i t h i n ordered host materials has been reviewed extensively i n recent y e a r s , 2 1 ' 5 3 ' 5 4 and a b r i e f overv iew is g iven here. Further details on various host materials are g iven i n chapters 3 and 4. In the late 1980 's , the ini t ia l w o r k on the synthesis o f polyacetylene wi th in the channels o f zeolites (crystalline, microporous aluminosi l icate materials) was carried out by B e i n and E n z e l . They incorporated p o l y p y r r o l e , 4 9 po ly th iophene , 5 5 and p o l y a n i l i n e 5 6 into different zeolites. Th i s w o r k was then extended to produce conjugated polymers w i t h i n the channels o f a mesoporous s i l i ca m a t e r i a l . 5 7 " 5 9 These results provided the first indications about the behaviour o f conjugated polymers in a confined environment: mic rowave conduct ivi ty measurements indicated that the confined polymers could be more conduct ive than in the b u l k . 6 0 The 14 composi te materials were i n powder form, w h i c h precluded achieving any macroscopic orientation o f the samples. M a r t i n e t al. developed oriented, encapsulated materials based on porous a lumina and track-etch m e m b r a n e s . 6 1 " 6 3 The pore diameters in these membranes were i n a l l cases larger than 10 n m , w h i c h a l lowed for convenient preparation o f po lymer microtubules but d i d not a l low any po lymer confinement effects to be observed. Substantial progress towards the goal o f encapsulating conjugated po lymer chains into oriented hosts was reported i n the literature dur ing the course o f this thesis. The w o r k by the groups o f Tolbert and Schwarz at the Univers i ty o f Cal i forn ia , Santa B a r b a r a 6 4 " 6 8 and the w o r k by the group o f G i n at the Univers i ty o f Ca l i fo rn ia , B e r k e l e y 6 9 " 7 5 represent the most significant developments in this field. They prepared host-guest systems where the conjugated polymer guest was encapsulated in very narrow channels, and their evidence suggested that interactions between different po lymer chains were el iminated. Th i s was accompl ished through investigation o f the t ime-resolved opt ical properties o f the composite material . Howeve r , the goal o f measur ing the electrical properties o f single conjugated po lymer chains direct ly has not been achieved to date; the nearest result describes the hole mobi l i ty on single conjugated po lymer chains in s o l u t i o n . 5 2 These and other relevant results are reviewed here to provide a perspective on the field o f encapsulated conjugated polymers . 1.4.3 Oriented Mesoporous Silica Encapsulant 15 The use o f a large external magnetic f ield to al ign the channels o f a surfactant-templated mesoporous s i l i ca material was pioneered by F i rouz i e t a l . 1 6 Th i s method may be used to produce sol id samples w i th an overal l channel orientation parallel to the appl ied magnetic field, w i th a lattice spacing o f 3.5 n m and a diameter o f 2.2 n m . 7 7 It was then shown that a soluble P P V derivative, poly[2-methoxy,5-(2 ' -e thylhexyloxy)- l ,4-phenylene vinylene] ( M E H - P P V , Figure 1.2(g)), cou ld be introduced into the channels o f such a host fo l lowing proper chemica l functionalization o f the channel surfaces (diameter reduced to ~ 1.7 n m ) . 6 4 The po lymer loading was s imply accompl ished by p lac ing the host in a heated po lymer solution for some t ime (Figure 1.6), then fo l lowing an op t imized solvent wash ing sequence to m a x i m i z e removal o f external polymer . The ini t ia l evidence for po lymer incorporation was based on measurements o f the fluorescence polarizat ion. It was recognized that at least some unencapsulated po lymer was present in larger cavities in the host material , but its effect cou ld be m i n i m i z e d through selective oxidat ion. It was further argued on geometrical considerations (the polymer pack ing radius is 0.8 - 0.9 n m in the sol id state) that only one po lymer chain could be present in each pore. Investigation o f the photoluminescence polar izat ion dynamics indicated that ini t ia l excitation energy local ized on the unencapsulated po lymer migrated qu ick ly to the lower-energy interacting polymer chains isolated polymer chains F i g u r e 1.6 Incorporation o f M E H - P P V into oriented mesoporous s i l ica . 16 encapsulated c h a i n s . 6 5 - 6 6 Further energy migrat ion along the encapsulated chain was a s lower process. The measurement o f transient absorption dynamics i n the femtosecond regime a l lowed the encapsulated po lymer to be dist inguished very clearly, based on comparisons wi th M E H -P P V i n dilute solution and as a thin f i l m . Overa l l , these results p rov ided very compe l l ing arguments for the presence o f isolated M E H - P P V chains in the pores o f the s i l i ca host. The single chains were not observed directly in the host channels, but the photophysical behaviour o f the composite indicated that chain isolation had been achieved. The determination o f the mic rowave conduct ivi ty o f the sample, w h i c h w o u l d give a value for the conduct ivi ty o f a single conjugated po lymer chain , was reportedly under w a y 6 4 but has not yet been publ ished. 1.4.4 L i q u i d C r y s t a l E n c a p s u l a n t The use o f lyotropic l i qu id crystals as a host material was pursued by Smi th e t a l , us ing a polymer izable mesogen to form an inverse hexagonal matr ix (Figure 1 .7) . 6 9 They ini t ia l ly used a water-soluble P P V precursor (Figure 1.8), w h i c h was expected to segregate into the aqueous phase o f the l i qu id crystal. C ros s - l i nk ing o f the matr ix stabil ized the l i qu id crystal host, after w h i c h the P P V precursor cou ld be converted to the final form by heating under vacuum. The interpore distance was 4.0 n m and the pore diameter ~1.5 n m . The resulting composi te showed a blue-shifted photoluminescence spectrum relative to bulk P P V . There was a large increase i n the photoluminescence intensity, and the absolute photoluminescence efficiency was reported as 3 0 % , substantially larger than for other samples i n the same study ( 5 - 2 0 % ) 7 3 but comparable to other values reported for bu lk P P V ( 2 7 % ) . 7 8 17 F i g u r e 1.8 High-temperature conversion o f a water-soluble precursor po lymer to P P V . The photoinduced absorption spectrum was also shifted to higher energy and showed an excitation intensity dependence consistent w i th interchain energy transfer i n h i b i t i o n . 7 4 Photoluminescence-detected magnetic resonance measurements showed that the composite material behaved s imi lar ly to conjugated po lymer dispersed in an inert m a t r i x . 7 2 The femtosecond photoluminescence dynamics were not investigated, though this w o u l d have provided the best evidence for po lymer chain isolat ion. The difficulty w i th a l l the above comparisons is that it is not k n o w n whether the po lymer precursor convers ion step proceeds in the composite in the same manner as in the bulk . F i g u r e 1.7 (a) Structure o f l i qu id crystal mesogen, (b) structure o f lyotropic l i qu id crystal w i th po lymer guest. 18 Subsequent w o r k w i t h a water-soluble P P V d e r i v a t i v e 7 9 (Figure 1.9) a l lowed more definit ive experiments: the po lymer could be extracted from the composite after synthesis and the photophysical properties o f the encapsulated and free po lymer chains could be compared d i r e c t l y . 7 5 T h i s showed that the local dielectric environment o f the matr ix , w h i c h is h ighly polar due to the carboxylate groups o f the mesogen, had little effect on the observed differences. It was also argued (without direct evidence, due to l ow po lymer content in the composite) that the effective conjugation length o f the po lymer was unaffected by the encapsulation process. Th i s left the isolation o f the po lymer chains as the only remain ing factor responsible for the observed differences. However , it is not clear whether the behaviour o f this P P V derivative can be general ized to P P V itself. The composite material cou ld also be prepared as an oriented film, w i t h the channels reportedly running perpendicular to the substrate, 7 1 by pressing the l i qu id crystal between two glass plates. The alignment was induced by the interactions at the glass / l iquid crystal interface. However , no direct evidence o f this al ignment was provided . W i t h such an oriented film, it was c la imed that a nearly ideal device structure was produced. The electrical properties o f this device structure were investigated, and it was found that both the po lymer and the stacked benzene rings o f the matr ix conducted electricity. Electroluminescence was detected and it was found to originate from both the matr ix and the polymer . Thus the electroluminescence behaviour o f the isolated chains cou ld not be investigated by this approach. 19 COO"Na+ COO"Na+ F i g u r e 1.9 Water-soluble P P V d e r i v a t i v e . 7 9 1.4.5 C l a y E n c a p s u l a n t Recently, O . O . Park and co-workers reported the successful intercalation o f M E H - P P V between the layers o f an organoclay material (Figure 1 . 1 0 ) . 8 0 ' 8 1 The incorporation o f some polymer was evidenced by an increase in the spacing o f the clay layers, as determined by X - r a y diffraction. N o further details o f the structure o f the composite were reported. W h i l e the layered structure o f the organoclay does not a l low ful l isolation o f the po lymer chains, the experimental results seemed to indicate that there was a large enhancement o f photoluminescence intensity (18x; absolute efficiencies were not reported). The absolute photoluminescence efficiency o f M E H - P P V is k n o w n to be 1 0 - 1 5 % ; 7 8 therefore the c la imed increase may be considered somewhat suspect. Electroluminescent devices cou ld be fabricated us ing this composi te material as the active F i g u r e 1.10 Preparation o f M E H - P P V / C l a y composite material . 20 layer between I T O and a l u m i n u m electrodes. The electroluminesence efficiency o f these devices was enhanced. Th i s increase was attributed to the confinement o f the charged species. The hole mobi l i ty in the composite was also reduced relative to bu lk M E H - P P V , w h i c h may have led to a better balancing o f the hole and electron distr ibut ion. Surpr is ingly , the device operated equally w e l l in reverse bias, indicat ing that the choice o f electrode material was not affecting device operation. However , this fact was not discussed in any detail. Clear ly , this material exhibi ted some very novel behaviour, but the l imi ted characterization d i d not a l low the or ig in o f this behaviour to be identified. Some further w o r k has been carried out on the incorporation o f P P V into layered host m a t e r i a l s . 8 2 ' 8 3 1.4.6 C y c l o d e x t r i n E n c a p s u l a n t A n d e r s o n et al. have shown that luminescent conjugated molecules encapsulated in cyclodextr in have substantially enhanced p h o t o s t a b i l i t y . 8 4 ' 8 5 They have extended this approach to polymers by synthesizing water-soluble conjugated po lymer chains threaded inside stacked P-cyclodextr in rings (Figure 1 .11) . 8 4 Sufficiently large capping groups at the end o f the po lymer chain prevent dethreading. Th i s produces po lymer chains w h i c h are very t ightly encapsulated. Di rec t investigation o f thin f i l m morphology by atomic force microscopy a l lowed the encapsulated po lymer strands to be resolved, w h i c h indicated that there was substantial reduction in interactions between chains. The absolute photoluminescence efficiency o f the po lymer chain was shown to i n c r e a s e 8 6 by 3 - 4x but the photostability o f these polymers was not discussed. The encapsulation provided by these macrocycles is not complete, w h i c h sti l l a l lows the intermolecular charge transport necessary for device o p e r a t i o n . 8 6 Encapsula t ion was shown to increase the electroluminescence efficiency by 2 - 5x . A g a i n device stability was not discussed. 21 HO F i g u r e 1.11 Encapsula t ion o f conjugated po lymer chain w i t h p - c y c l o d e x t r i n . 8 6 T h i s approach to encapsulation appears very p romis ing and w i l l hopefully be subject to further study. The current results clearly indicate that this approach is useful but have not y ie lded any further insight into the electroluminescence processes themselves. A s noted i n the literature, one l imi ta t ion o f this approach is the mob i l i t y o f the macrocycle a long the po lymer chain , w h i c h can cause a non-uni form distr ibut ion o f the encapsulan t . 8 6 Short, encapsulated ol igomers prepared i n this fashion cou ld also be grafted to a conduct ing surface, and such a configuration w o u l d be ideal for electrical measurements on single po lymer chains. Th i s has not been reported to date. 22 1.4.7 Literature Summary The w o r k o f N g u y e n e t a l . 6 5 p rov ided the clearest results on the properties on encapsulated conjugated polymers . The results on l i qu id crystal encapsu la t ion 7 5 were important but further investigation o f the encapsulated po lymer w o u l d be needed to show the effect o f chain isolation on electronic excitations. The w o r k on clay e n c a p s u l a t i o n 8 0 ' 8 1 was interesting but d id not provide m u c h more understanding o f the electronic processes i n conjugated polymers . The w o r k on P-cyclodextrin encapsu la t ion 8 6 is important because a functional device was demonstrated, and further insight might be gained from photophysical studies. These works illustrate col lect ively that encapsulated conjugated polymers exhibi t nove l behaviour, i n particular enhanced luminescence efficiencies. Howeve r , the cause o f this behaviour is difficult to identify without a detailed analysis o f the nanostructure o f the material — and incorrect conclusions might be d rawn i f such analysis is neglected. It is also clear that the ideal device structure, consis t ing o f al igned and isolated conjugated po lymer chains, remains to be reported. 1.5 Thesis Summary This thesis describes the synthesis and characterization o f conjugated po lymer guests in mesoporous host materials. The p r inc ipa l mot ivat ion o f the w o r k was to create materials and device structures w h i c h w o u l d further the understanding o f electronic processes w i t h i n conjugated polymers . However , it was recognized early on that little understanding can be gained from materials w h i c h are not fully characterized. A change in material properties cannot be rat ionalized i f the structure o f the material is not k n o w n in detail, w h i c h is w e l l illustrated by the w o r k on M E H - P P V / c l a y composites discussed above. T h i s realization guided this w o r k 23 towards the applicat ion o f h igh resolution characterization techniques to conjugated po lymer composite materials. The uni fy ing theme o f the final w o r k is then characterization o f such composi te materials on the nanometre scale. In chapter 2, the m a i n characterization techniques used for analysis o f composi te materials are reviewed, w i th an emphasis on electron mic roscopy techniques. The most important o f these is electron energy-loss spectroscopy ( E E L S ) and energy-filtered t ransmission electron microscopy ( E F T E M ) . The applicat ion o f E E L S to the analysis o f organic materials is reviewed. A general introduction to ordered porous host materials is presented i n chapter 3, and the literature on introduction o f polymers into such hosts is reviewed. Th i s is fo l lowed by a study o f the incorporation o f P P V into a mesoporous s i l i ca composite material . C h e m i c a l analysis o f the composite material w i t h nanometre resolution by E E L S and E F T E M is used to show directly that the conjugated po lymer is present inside the pores o f the host. Chapter 4 is devoted to the preparation o f oriented porous thin films towards the goal o f preparing an ideal electroluminescent device structure. Porous a lumina membranes prepared at l o w temperatures are presented as very good candidates for preparing such a structure. The ini t ia l analysis o f conjugated polymer/porous a lumina composite materials is described in chapter 5. The E E L S and E F T E M measurements were carried out parallel to the pores o f the host. The analysis was hindered by spectral features due to surface effects in E E L S . Furthermore, the samples showed a further loss at a large distance from the sample surface that cou ld not be associated wi th a surface p lasmon. Chapter 6 deals w i t h this unexpected loss at large distance from the sample surface. The Cherenkov effect is identified as the source o f this spectral feature. The experimental and theoretical results show that the spectroscopy o f the samples on the loca l scale was affected by the large scale structure o f the sample, due to the radiative nature o f the Cherenkov effect. 24 Chapter 7 discusses further ways o f preparing polymer/porous a lumina f i l m composites and their characterization. Init ial w o r k in the direct ion o f preparing conjugated ol igomers grafted to a s i l i con surface is described. The use o f different d r i v i n g forces for composite preparation is discussed, and the centrifugal force is investigated in detail. T h i n sections o f the samples are investigated by E E L S and E F T E M . Low-energy losses i n the a lumina host at 200 k V are found to interfere w i th the polymer-specif ic losses. Chapter 8 discusses some o f the possible approaches towards ideal device fabrication. A l t h o u g h the overal l goal o f preparing an ideal device structure has not been realized, the elements for preparing such a device and the tools for its structural characterization have been developed. References 1. Pope, M . ; Swenberg , C . E . Electronic Processes in Organic Crystals; Clarendon Press: O x f o r d , 1982. 2. H igg ins , S. J . Chem. Soc. Rev. 1997, 26, 247. 3. Goldenberg , L . M . ; B r y c e , M . R . ; Petty, M . C . J. Mater. Chem. 1999, 9, 1957. 4. H o r o w i t z , G . Adv. Mater. 1998, 10, 365. 5. Sirr inghaus, H . T . N . F . R . H . Science 1998, 280, 1741. 6. C a c i a l l i , F . Philos. Trans. R. Soc. London, Ser. A 2000, 358, 173. 7. Heeger, A . J . Solid State Commun. 1 9 9 8 ,1 0 7 , 673 . 8. Rees, I. D . ; Rob inson , K . L . ; H o l m e s , A . B . ; Towns , C . R . O . ; D e l l , R . MRS Bull. 2002, 27, 4 5 1 . 9. A k c e l r u d , L . Prog. Polym. Sci. 2003, 28, 875. 25 10. N u n z i , J . M . CR. Phys. 2002, 3, 523. 11. Kranze lb inder , G . ; L e i s i n g , G . Rep. Prog. Phys. 2000, 63, 729. 12. Scherf, U.; R i eche l , S.; L e m m e r , U.; Mahr t , R . F . Curr. Opin. Solid State Mater. Sci. 2001, 5, 143. 13. Saxena, V . ; Malho t ra , B . D . Current Applied Physics 2003, 3, 293 . 14. Tour , J . M . Acc. Chem. Res. 2000, 33, 791 . 15. Tour , J . M . Molecular electronics: commercial insights, chemistry, devices, architecture and programming.; W o r l d Scient i f ic : L o n d o n , 2003 . 16. B u m m , L . A . ; A r n o l d , J . J . ; C y g a n , M . T . ; Dunbar , T . D . ; B u r g i n , T . P . ; Jones, L . ; A l l a r a , D . L.; Tour , J . M.; Weiss , P . S. Science 1996, 271, 1705-1707. 17. Reed , M . A . ; Z h o u , C ; M u l l e r , C . J . ; B u r g i n , T . P . ; Tour , J . M . Science 1997, 278, 252-254. 18. Shi rakawa, H . ; L o u i s , E . J . ; M a c D i a r m i d , A . G . ; Ch i ang , C . K . ; Heeger, A . J . Chem. Commun. 1977, 578. 19. Greenham, N . C ; Fr iend , R . H . Semiconductor device physics of conjugated polymers; S o l i d State Phys ics ; A c a d e m i c Press: N e w Y o r k , 1995; Vol . 49 , p. 1-149. 20. B u r n , P . L . ; Kraft , A . ; Baigent, D . R . ; Bradley , D . D . C ; B r o w n , A . R . ; F r iend , R . H . ; G y m e r , R . W . ; Ho lmes , A . B . ; Jackson, R . W . J. Am. Chem. Soc. 1993, 115, 10117-10124. 21. C a r d i n , D . J . Adv. Mater. 2002, 14, 553. 22 . N g u y e n , T . Q . ; D o a n , V . ; Schwartz , B . J . J. Chem. Phys. 1 9 9 9 , 1 1 0 , 4068. 23. Pekker , S.; M o r i n , A . ; Berniere , F . Polymer Communications 1990, 31, 75. 24. K o b r y a n s k i i , V . M . Polym. Degrad. Stab. 1991, 33, 387. 25. K a l i n o w s k i , D . J. Phys. D: Appl. Phys. 1999, 32, R l 79. 26 . Tang , C . W . ; V a n S lyke , S. A . Appl. Phys. Lett. 1987, 51, 913-915. 26 27. Burroughes , J . H . ; Bradley , D . D . C ; B r o w n , A . R . ; M a r k s , R . N . ; M a c k a y , K . ; F r iend , R . H. ; Burns , P . L . ; Ho lmes , A . B . Nature 1990, 347, 539-41 . 28. Papadimit rakopoulos , F . ; Kons tad in id i s , K . ; M i l l e r , T. M . ; O p i l a , R . ; Chandross , E . A . ; G a l v i n , M . E . Chem. Mater. 1994, 6, 1563-8. 2 9 . Rothberg, L . J . ; Y a n , M . ; Papadimit rakopoulos , F . ; G a l v i n , M . E . ; K w o c k , E . W . ; M i l l e r , T . M. Synth. Met. 1996, 80, 41-58. 30 . Scott, J . C . ; Kau fman , J . H . ; B r o c k , P . J . ; D iP i e t ro , R . ; Sa lem, J . ; Go i t i a , J . A . J. Appl. Phys. 1996, 7 9 , 2745. 3 1 . K i m , J . S.; H o , P . K . H . ; M u r p h y , C . E . ; Baynes , N . ; Fr iend , R . H . Adv. Mater. 2002,14, 206. 3 2 . H e , G . F . ; Chang , S. C ; C h e n , F . C ; L i , Y . F . ; Y a n g , Y . Appl. Phys. Lett. 2002, 81, 1509. 3 3 . B a l d o , M . A . ; Thompson , M . E . ; Forrest, S. R . Nature 2000, 403, 750. 34. K a n , S. D . ; Zhang , J . Y . ; L i u , X . D . ; Shen, F . Z . ; M a , Y . G . ; Shen, J . C . Synth. Met. 2003, 135, 253 . 35. Welter , S.; Brunner , K.; Hofstraat, J . W . ; D e C o l a , L . Nature 2003, 421, 54. 36. Samuel , I. D . W . ; Rumbles , G . ; C o l l i s o n , C . J . ; F r iend , R . H . ; Mora t t i , S. C ; Ho lmes , A . B . Synth. Met. 1997, 84, 497. 37. Becker , H.; Burns , S. E . ; Fr iend , R . H . Phys. Rev. B: Condens. Matter 1997, 56, 1893-1905. 38. Dannetun, P . ; Loegd lund , M . ; Fah lman , M . ; B o m a n , M . ; Stafstroem, s.; Salaneck, W . R . ; Lazza ron i , R . ; Fredr iksson, C ; Bredas, J . L . ; et a l . Synth. Met. 1993, 55, 212-17. 39 . Cao , Y . ; Parker , I. D . ; Y u , G . ; Zhang , C ; Heeger, A . J . Nature 1999, 397, 414. 4 0 . W i l s o n , J . S.; Dhoot , A . S.; Seeley, A . J . A . B . ; K h a n , M . S.; K o h l e r , A ; Fr iend , R . H . Nature 2001, 413, 828. 41. Wohlgenannt , M.; Vardeny , Z . V . J. Phys.: Condens. Matter 2003, 15, R 8 3 . 27 42. Wohlgenannt , M . ; J iang, X . M ; Vardeny , Z . V . Synth. Met. 2003, 137, 1069. 43. Karabunar l iev , S.; Bit tner , E . R . J. Chem. Phys. 2 0 0 3 , 1 1 9 , 3988. 44 . Karabunar l iev , S.; Bit tner, E . R . Phys. Rev. Lett. 2003, 90, N o . 057402. 4 5 . Baigent , D . R . ; Greenham, N . C ; Gruener, J . ; M a r k s , R . N . ; Fr iend , R . H . ; Mora t t i , S. C ; Ho lmes , A . B . Synth. Met. 1994, 67, 3-10. 46. B r o w n , T . M . ; F r iend , R . H . ; M i l l a r d , I. S.; Lacey , D . J . ; But ler , T. ; Burroughes , J . H . ; C a c i a l l i , F . J. Appl. Phys. 2003, 93, 6159-6172. 47 . H o , P . K . H . ; K i m , J . S.; Burroughes , J . H . ; Becker , H . ; L i , S. F . Y . ; B r o w n , T . M . ; C a c i a l l i , F.; F r iend , R . H . Nature 2000, 404, 4 8 1 . 48 . C y g a n , M . T. ; Dunbar , T . D . ; A r n o l d , J . J . ; B u m m , L . A . ; Shedlock, N . F . ; B u r g i n , T . P . ; Jones, L.; A l l a r a , D . L.; Tour , J . M.; Weiss , P . S. J. Am. Chem. Soc. 1998, 120, 2721-2732. 49 . B e i n , T. ; E n z e l , P . Angew. Chem. Int. Ed. Engl. 1989, 28, 1962. 50 . Hel le r , C . M . ; C a m p b e l l , I. H . ; L a u r i c h , B . K . ; Smi th , D . L . ; Bradley , D . D . C ; B u r n , P . L . ; Ferraris, J . P . ; M u l l e n , K . Phys. Rev. B: Condens. Matter 1996, 54, 5516-5522. 51. Hagler , T . W . ; Pakbaz , K.; V o s s , K . F . ; Heeger, A . J . Phys. Rev. B: Condens. Matter 1991, 4 4 , 8652-66. 52. Hoofman , R . J . O . M . ; D e Haas, M . P . ; Siebbeles, L . D . A . ; W a r m a n , J . M . Nature 1998, 3 9 2 , 54-56. 5 3 . O z i n , G . A . ; C h o m s k i , E . ; Khusha lan i , D . ; Mac lach l an , M . J . Curr. Opin. Colloid Interface S c i . 1998, 3 , 181-193. 54. M o l l e r , K . ; B e i n , T . Chem. Mater. 1998,10, 2950-2963. 55. E n z e l , P . ; B e i n , T . Chem. Commun. 1989,18, 1326. 56 . E n z e l , P . ; B e i n , T . / . Phys. Chem. 1989, 93, 6270 . 57. W u , C . - G . ; B e i n , T. Science 1994, 264, 1757-1759. 28 58. W u , C . - G . ; B e i n , T. Conducting polymer wires in mesopore hosts; Studies in Surface Science and Catalysis; Elsevier : A m s t e r d a m , 1994; V o l . 84, p . 2269-2276. 59. W u , C . - G . ; B e i n , T. Science 1994, 266, 1013-1015. 60 . Esnouf , S.; Beuneu , F . ; E n z e l , P . ; B e i n , T . Phys. Rev. B: Condens. Matter 1997, 56, 12899-12904. 6 1 . C a i , Z . H . ; M a r t i n , C . R . J. Am. Chem. Soc. 1989, 111, 4138 . 62 . M a r t i n , C . R . ; V a n D y k e , L . S.; C a i , Z . H . ; L i a n g , W . B . J. Am. Chem. Soc. 1 9 9 0 , 1 1 2 , 8976. 63 . M a r t i n , C . R . Chem. Mater. 1996, 8, 1739-1746. 64. W u , J . J . ; Gross , A . F . ; Tolbert , S. H . J. Phys. Chem. B 1999, 103, 2374. 65 . N g u y e n , T . Q . ; W u , J . J . ; D o a n , V . ; Schwartz , B . J . ; Tolbert, S. H . Science 2000, 288, 652. 66. Tolbert , S. H . ; W u , J . J . ; Gross , A . F . ; N g u y e n , T . Q . ; Schwartz , B . J . Microporous Mesoporous Mater. 2001, 44, 445. 67. Schwartz , B . J . ; N g u y e n , T. Q . ; W u , J . J . ; Tolbert , S. H . Synth. Met. 2 0 0 1 , 1 1 6 , 35. 68 . N g u y e n , T . Q . ; W u , J . J . ; Tolbert , S. H . ; Schwartz , B . J . Adv. Mater. 2001,13, 609. 69. Smi th , R . C ; Fischer, W . M . ; G i n , D . L . J. Am. Chem. Soc. 1997, 119, 4092-4093. 70 . G i n , D . L . ; Smi th , R . ; D e n g , H . ; Le i s i ng , G . Synth. Met. 1 9 9 9 , 1 0 1 , 52. 7 1 . Markar t , P . ; Zojer , E . ; Tasch , S.; Smi th , R . ; G i n , D . L . ; Le i s i ng , G . Synth. Met. 1 9 9 9 , 1 0 2 , 1155. 72 . Zojer , E . ; Markar t , P . ; L i s t , E . J . W . ; Graupner , W . ; Smi th , R . ; L e i s i n g , G . ; Shinar , J . ; G i n , D. L. Synth. Met. 1999, 102, 1270. 73 . L i s t , E . J . W . ; Markar t , P . ; Graupner , W . ; L e i s i n g , G . ; Partee, J . ; Shinar , J . ; Smi th , R . ; G i n , D . L . Opt. Mater. 1999,12, 369. 74. Gadermaier , C ; L i s t , E . J . W . ; Markar t , P . ; Graupner , W . ; Partee, J . ; Shinar , J . ; Smi th , R . ; G i n , D . ; Le i s i ng , G . Synth. Met. 2000, 111, 523 . 29 75. G i n , D . L . ; Y o n e z a w a , K . Synth. Met. 2 0 0 1 , 1 2 1 , 1291. 76. F i r o u z i , A . ; Schaefer, D . J . ; Tolbert , S. TL; S tucky, G . D . ; C h m e l k a , B . F . J. Am. Chem. Soc. 1 9 9 7 , 1 1 9 , 9466-9477. 77. Tolbert , S. H . ; F i r o u z i , A . ; Stucky, G . D . ; C h m e l k a , B . F . Science 1997, 278, 264-268. 78. Greenham, N . C ; Samuel , I. D . W . ; Hayes , G . R . ; Ph i l l i p s , R . T. ; Kessener, Y . A . R . R . ; Mora t t i , S. C ; Ho lmes , A . B . ; F r iend , R . H . Chem. Phys. Lett. 1995, 241, 89-96. 79 . Wagaman , M . W . ; Grubbs , R . H . Macromolecules 1997, 30, 3978. 80. Lee , T . W . ; Park, O . O . ; K i m , J . J . ; H o n g , J . M . ; K i m , Y . C . Chem. Mater. 2001 , 1 3 , 2217 . 81 . Lee , T . W . ; Park, O . O . ; Y o o n , J . ; K i m , J . J . Synth. Met. 2001, 121, 1737. 82. Lee , H . C ; Lee , T . W . ; Park, O . O . Opt. Mater. 2003, 21, 187. 83. Lee , H . C ; Lee , T . W . ; L i m , Y . T. ; Park, O . O . Appl. Clay Sci. 2002, 21, 287. 84. Taylor , P . N . ; O 'Conne l l , M . J . ; M c N e i l l , L . A . ; H a l l , M . J . ; A p l i n , R . T. ; Ande r son , H . L . Angew. Chem. Int. Ed. Engl. 2000, 39, 608. 85. C r a i g , M . R . ; Hutchings , M . G . ; Clar idge , T . D . W . ; Ande r son , H . L . Angew. Chem. Int. Ed. Engl. 2002, 41, 1769-1772. 86. C a c i a l l i , F . ; W i l s o n , J . S.; M i c h e l s , J . J . ; D a n i e l , C ; S i lva , C ; Fr iend , R . H . ; Sever in , N . ; Samor i , P . ; Rabe , J . P . ; O 'Conne l l , M . J . ; Taylor , P . N . ; Ande r son , H . L . Nature Mater. 2002,1, 160-164. 30 CHAPTER 2 Characterization of Nanocomposite Materials The chemica l analysis o f aperiodic structures at h igh spatial resolution is p rov ing to be a central challenge o f nanoscience. A s new methods are devised to assemble materials into structures on the nanometre scale, it is becoming more important to show by means as direct as possible that the desired phys ica l and chemica l structures are obtained. W i t h the advent o f field-emission electron microscopes, the atomic force microscope ( A F M ) and the scanning-tunnel l ing microscope ( S T M ) , topographical characterization at nanometre resolution has become a relatively routine experiment. Howeve r , chemica l analysis at the nanometre scale is st i l l far from reaching the same level o f s impl ic i ty , and there are instrumental l imi ts on the achievable resolut ion . 1 The development o f characterization techniques for composi te materials based on porous inorganic hosts was an ongoing challenge w i t h i n our research group. The central goal was to establish the po lymer distr ibution w i t h i n the composi te material on the nanometre scale. Op t i ca l techniques for characterization are pushed to their present-day l imi t by confocal fluorescence microscopy, w h i c h is a w i d e l y used tool for establishing the distr ibution o f fluorescent materials w i t h 0.1 u.m resolution in 3 - D . Th i s can be readily used for many conjugated polymers i n optical ly transparent hosts. A recently developed technique, near-field scanning optical m i c r o s c o p y , 2 can be used to map fluorescence in 2 - D w i t h even higher resolution. B u t for w o r k at resolution better than 10 n m , electron microscopy is used for 31 topographical and chemica l analysis. T h i s is introduced be low and is fo l lowed by a review o f h igh resolution chemica l analysis techniques. T w o forms o f electron mic roscopy are most relevant to current research in materials science: scanning electron microscopy ( S E M ) and t ransmission electron mic roscopy ( T E M ) . S E M is a more wide ly used technique due to its ever - improving s impl ic i ty , w h i l e T E M is often necessary when details in the nanometre range are important. 2.1 Scanning Electron Microscopy The S E M is a very versatile instrument for the characterization o f surface topography and material compos i t i on . 3 A n energetic beam o f electrons ( typically 1 to 30 k V ) is rastered across the surface o f the specimen, and an image is constructed by detecting electrons emitted from the surface. I f an electron interacts elastically w i th an atomic nucleus in the sample and is returned out, it is said to be backscattered. These electrons provide some sensitivity to the atomic number o f the sample but are emitted in relatively l ow y ie ld . A backscattered electron detector may be used to obtain some elemental contrast in a sample but w i t h l imi ted resolution. I f an electron undergoes a number o f inelastic processes i n the sample before being re-emitted out, it is cal led a secondary electron. These are emitted in a y i e ld approaching unity and are largely independent o f material composi t ion . The intensity o f secondary electron emiss ion observed at each point is modulated by the local sample structure, as more electrons can escape from protrusions than from depressions. Th i s a l lows a topographical image to be formed. The secondary electron detector normal ly used in S E M presents a surface image w i t h a point resolution o f - 1 . 5 n m at accelerating voltages from 10 to 20 k V (Hi tach i S-4700 f ield-emission S E M , U B C Elec t ron M i c r o s c o p y L a b ) . L o w e r accelerating voltages may be used to reduce charging effects and to increase surface detail, at the expense o f resolution. 32 For insulat ing samples, a conduct ive coating must be appl ied to prevent charge bui ld-up on the surface, as low-voltage imaging does not provide sufficient resolution. T h i s is usually accompl ished by sputtering a th in coating o f gold-pal ladium onto the sample. Th is coating is adequate for most S E M w o r k but begins to show some structure at higher magnificat ions (above 100,000x) , w h i c h was the usual operating regime for this work . Other coating materials (pla t inum, ch romium) are smoother and are more suitable for h igh resolution w o r k but were not readily available for use at U B C . In the S E M , the contrast between an inorganic host material and an organic guest is not very large, especially wi th a conduct ive coat ing on the surface. Some inferences may be made f rom changes in the geometry (e.g., complete filling o f pores). 2.2 Transmission Electron Microscopy In the T E M , electrons are transmitted through a very thin sample . 4 The contrast is p rov ided by the abil i ty o f the sample to scatter electrons, w h i c h is largely a product o f thickness and atomic number. The accelerating voltage (beam energy) is typical ly set at 200 k V for materials science, in order to observe thicker specimens, w i t h some loss o f contrast over operating at 80 k V (as used for b io logica l samples). In general, a T E M is required to investigate structures w i t h detail be low 5 - 1 0 n m . Sub-nanometre resolution is readily achievable on standard T E M instruments. The achievable information l imi t on a modern field-emission T E M at 200 k V is 0.12 n m (Tecnai F 2 0 T E M , Nano- Imag ing Faci l i ty , S F U ) . A T E M can also be operated in scanning mode ( S T E M ) , where it functions s imi lar ly to an S E M , but w i t h a higher resolution and the choice o f detecting transmitted electrons (bright-field imag ing for unscattered electrons, dark-field imaging for strongly scattered electrons) or secondary electrons. 33 The large electron-beam energy causes damage to the sample over t ime. Porous s i l i ca and a lumina films tend to be fairly sensitive to beam damage in the T E M , as has been observed for the mesoporous material M C M - 4 1 . 5 > 6 Organ ic materials are s imi la r ly sensitive. Organic materials scatter electrons in the T E M only weakly , due to the l o w atomic number o f carbon. The simplest approach to i m p r o v i n g the contrast o f organic materials is heavy-metal staining, w h i c h is w i d e l y used for b io log ica l specimens. The v i n y l i c carbons in po ly( l ,4 -phenylene vinylene) ( P P V ) and its derivatives are readily stained by o s m i u m tetroxide. The phenyl group may also be stained by ruthenium tetroxide. 7 T h i s should provide better contrast i n the T E M because o f the increased electron scattering by the heavy nucle i . Howeve r , in situations where the amount o f po lymer is l imi ted , this may st i l l not provide sufficient analytical contrast, and the use o f more special ized techniques for chemica l analysis is required. 2.2.1 T E M S a m p l e P r e p a r a t i o n The central drawback o f the T E M is the need for electron-transparent samples. S m a l l particles may be investigated directly, but bu lk samples must be thinned or sectioned to be observed. The useful thickness range is a function o f the accelerating voltage and composi t ion ; for w o r k at 200 k V , a thickness be low 200 n m is usually required. W h i l e sample preparation techniques are w e l l established, they are generally t ime-consuming. The observation o f porous films in the plan geometry, that is l ook ing d o w n the channels, is straightforward and not sensitive to film thickness, p rov ided free-standing films can be prepared. The film cross-sections are more difficult to image, as they must be prepared in the form o f th in sections less than 200 n m thick i n order to be electron-transparent. The preparation o f these th in sections is not entirely t r iv ia l for hard materials. The convent ional technique for cross-section preparation o f such materials is d i m p l i n g and ion-34 m i l l i n g . M a n y samples are glued to form a sandwich structure, w h i c h is then sawed into 1 m m slices. A 3 m m disk is cut from this slice, ground to 100 urn thickness, and then d impled in the centre unt i l the sample is - 2 5 u m thick at the centre. The sample is then ion-mi l l ed using an argon ion beam unt i l a smal l hole is formed. The cross-sections are finally observed in the T E M along the edges o f the mi l l ed hole. Th is process is evidently tedious and suffers from artifacts introduced by the i o n - m i l l i n g process (amorphisation, preferential removal o f elements). The state-of-the-art in thin-section preparation involves the focused ion-beam ( F I B ) technique. It is executed inside an S E M w i t h a beam o f ga l l ium ions: the sample is m i l l e d away at a precisely k n o w n location unt i l the desired section is obtained. T h i s approach is also prone to p roduc ing damage artifacts. Fo r the duration o f most o f this work , there was no ready access to the proper instrument - only one existed i n Canada outside o f industrial research laboratories. In late 2002 , a new F I B instrument was installed at S F U . A s a s imple alternative, the small-angle cleavage ( S A C ) technique was developed by M c C a f f r e y for routine cross-section preparation o f thin f i lms for T E M . 8 " 1 0 There are two major requirements for successful execution o f the technique: a substrate that cleaves readily and a th in f i lm (< 300 nm) w i t h good adhesion to the substrate. The sample is cleaved at a shal low angle (< 3 0 ° ) to form a sharp wedge (Figure 2.2). The last micrometre o f the wedge near the tip is then sufficiently th in for observation i n cross-section by T E M . W i t h practice, S A C samples can be produced w i t h i n ha l f a day, w h i c h is substantially faster than the d i m p l i n g technique. Howeve r , the technique is less useful for f i lms that are not homogeneous in the plane o f the substrate, as the gradual change in thickness o f the wedge makes it difficult to visual ize films wi th addit ional structure in the depth o f the cross-section. Fo r such films, u l t ramicrotomy and focused-ion beam m i l l i n g are more suitable approaches to thin section preparation. Nevertheless, the S A C technique is very useful for v i e w i n g cross-sections without artifacts f rom the sect ioning process. 35 Figure 2.2 The small angle cleavage (SAC) technique applied to a thin film: the tip of the final specimen will often be thin enough over a sufficient length for TEM observation if a < 30°. Ultramicrotomy is commonly used to prepare cross-sections of biological specimens. The specimen is first embedded in epoxy resin, then trimmed with a razor blade or glass knife to form a truncated pyramid around the area of interest (Figure 2.1). After careful cleaning and washing to remove any small particles, the pyramid is sectioned from bottom to top using a Thin film Substrate Sectioning direction Figure 2.1 Truncated pyramid geometry of epoxy-embedded sample for sectioning by ultramicrotomy. The resulting sections float into a water bath where they are collected with TEM grids. 36 d iamond knife; the result ing thin sections float into a water bath, where they may be collected us ing a T E M gr id . The sections are typical ly 20 to 60 n m thick. Th i s technique is most readily applicable to soft materials, but in fact can be used to section even the hardest m a t e r i a l s . 1 1 ' 1 2 Cross-sections o f self-assembled mesoporous s i l i ca films on m i c a and graphite were produced by u l t r a m i c r o t o m y . 1 3 ' 1 4 Furneaux e t a l . successfully used ul t ramicrotomy to obtain th in sections o f porous a lumina films for T E M . 1 5 2.3 High Resolution Chemical Analysis The different approaches to spatially resolved chemica l analysis are reviewed here, fo l lowed by a more detailed descript ion o f the technique that was used in this work , electron energy-loss spectroscopy ( E E L S ) w i t h i n the T E M . The S T M , w h i c h can give chemica l information at very h igh resolution, is not discussed here as it is not applicable to insulat ing materials. These and other emerging techniques, especially scanning probe techniques, are described i n the proceedings o f a recent workshop on nanoscale spectroscopy. 2 M o s t o f the techniques for chemica l analysis rely on ini t ia l ly exc i t ing inner-shell (core) electrons by an energetic beam o f electrons or X- rays (Figure 2.3) and then detecting the results o f the de-excitation process. The outer-shell electrons also provide less direct information for chemica l analysis under certain circumstances. The spatial resolution is l imi ted , i n a l l cases, by the in i t ia l excitation vo lume. The resolution l imits o f the techniques discussed here are summar ized in Table 2.1. 37 Vacuum level Fermi level Core electrons Valence electrons 4 — • - - • — • -Light Auger X-Ray Inner-shell Exci ta t ion Outer-shell Exci ta t ion De-exci tat ion Figure 2.3 Elect ronic excitation and de-excitation mechanisms i n a sol id, (adapted from ref. 1) Table 2.1 Current spatial resolution l imits o f chemica l analysis techniques. Technique Sampling Depth (55») Lateral Resolution (nm) 10 6 ( 3 0 r ) > 10 1000 10 1 X - r a y photoelectron spectroscopy Scanning A u g e r microscopy Energy-dispers ive X - r a y spectroscopy i n S E M Energy-dispers ive X - r a y spectroscopy in T E M Elect ron energy-loss spectroscopy in T E M - 1 . 0 - 1 . 0 1000 100 100 + for a synchrotron X-ray source 2.3.1 X-Ray Photoelectron Spectroscopy Monochroma t i c X- rays are used as the excitation source for X - r a y photoelectron spectroscopy ( X P S ) ; these result in the emiss ion o f photoelectrons and A u g e r electrons from the sample. The kinet ic energy o f the ejected electrons is measured, y ie ld ing the core electron energies. The relative abundance o f any element may be readily calculated, and some information on the chemica l state o f the element may also be obtained. A l t h o u g h the lateral resolution is l imi ted by the beam diameter (-1 m m for conventional sources; - 3 0 n m for 38 synchrotron s o u r c e s 1 6 ) , this technique is very surface-sensitive as a result o f the l imi ted mean free path o f the generated photoelectron, w h i c h is on the order o f 1 n m . However , the adsorption o f atmospheric hydrocarbons on the surface can affect the analysis o f carbon-containing samples. W i t h the development o f more refined X - r a y optics, this technique may come to be very important for h igh resolution analysis, as it induces less sample damage than electron m i c r o s c o p y . 1 6 2.3.2 Energy-Dispersive X-Ray Spectroscopy and Scanning Auger Microscopy The electron beam w i t h i n an S E M can be used as a h ighly focused source o f excitation ( d o w n to ~1 n m diameter for a f ield emiss ion source) for the generation o f X- rays and A u g e r electrons. These are collected and analysed to form the basis for energy-dispersive X - r a y ( E D X ) analysis and scanning A u g e r microscopy ( S A M ) , respectively. In the case o f bu lk samples, the X - r a y signal used for E D X originates from a relatively large domain surrounding the point excited by the electron beam, due to the penetration depth o f the energetic electrons and the low absorption o f X- r ays . T h i s large excitation vo lume effectively reduces the spatial resolution to 1 u m but also a l lows for a thin conduct ive f i lm to be applied to the surface i f the sample is insulating. O n the other hand, only A u g e r electrons released near the surface can escape and be detected (as w i th photoelectrons), w i t h the result o f higher lateral resolution ( -10 nm) than E D X but w i t h the requirement that the sample be conductive (or very thin wi th a conduct ive backing) to avoid charging effects. W h e n coupled to a S T E M , E D X may also be used to quantify the compos i t ion o f thin film samples. Since the sample must be sufficiently thin to transmit electrons, the excitation v o l u m e for X - r a y s is greatly reduced and the lateral resolution is l imi ted by scattering o f the electron beam wi th in the thin section, effectively reaching 10 n m . In practice, sufficient material 39 must also be present to obtain reasonable count ing statistics, w h i c h l imits the m i n i m u m sample thickness that may be used. B e a m damage to the sample m a y also affect the results i f a particular component is being removed at a higher rate. 2.3.3 Electron Energy-Loss Spectroscopy Fina l ly , the distr ibution o f energy losses incurred by the transmitted electron beam in a T E M may also be measured to y i e ld elemental composi t ion . T h i s approach belongs to a fami ly o f techniques described generally as electron energy-loss spectroscopy. A l t h o u g h it is l imited to th in samples, it provides higher resolution (~1 nm) and sensit ivity (as few as 1 to 10 atoms can be detected 1 ) than a l l the techniques listed above (see Table 2.1). It is ideally suited for the analysis o f nanocomposi te materials and is described in more detail be low. 2.4 General Principles of EELS The standard reference for E E L S in the T E M is the book by Ege r ton . 1 A more recent publ ica t ion by B r y d s o n 1 7 covers some new developments and focuses on the experimental aspects o f E E L S . In general, the interactions between a t ravel l ing electron and the components o f a so l id are termed scattering events (Figure 2.4). I f the electron is scattered from an atomic nucleus, the process occurs without the electron los ing any significant amount o f energy. T h i s elastic scattering can cause large deviations in the trajectory o f the electron, to the extent where it may be backscattered out o f the so l id , or smaller deviations (e.g., diffraction for crystalline materials). Some electrons w i l l also excite col lect ive oscil lat ions o f the atoms in the sol id (phonons); these occur at l o w energies i n the m e V range and are not dist inguishable from the unscattered and elastically scattered electrons i n the T E M (though they may be i n other forms o f 40 B Figure 2.4 Geometry o f (a) elastic, (b) inelastic, inner-shell and (c) inelastic, outer-shell scattering events i n v o l v i n g an electron and a carbon atom, (adapted from ref. 1) E E L S wi th higher energy resolution). These electrons are detected by the spectrometer as the zero-loss peak, the w id th o f w h i c h gives the energy resolution o f the instrument. A typica l loss spectrum is shown i n Figure 2.5. Inelastic scattering occurs when the electron interacts w i t h either the inner-shell or outer-shell electrons o f an atom (Figure 2.4). The scattering from inner-shell electrons produces an energy loss characteristic for each element, w i th a value typical ly between 50 and 2000 e V . In the loss spectrum, these appear as an ionizat ion edge (e.g., at 284 e V for carbon Is electrons) and are usually employed for elemental analysis. The fine structure o f the ionizat ion edge is also used to derive addit ional information about the chemica l state o f the element. There has been extensive w o r k on mode l l i ng these e d g e s , 1 ' 1 8 but since they play no role in this w o r k they are not discussed any further. The excited state produced by inelastic scattering can also manifest i tself as a collect ive excitation o f the outer-shell electrons, generally referred to as a p lasmon. O n longer t ime scales, the energy o f the p lasmon is distributed over many electrons, but al l the energy may be carried by a single electron on short t ime scales, w h i c h makes such excitations possible even in 41 c Zero-loss peak Inner-shell ionization edges 1 1 1 r 100 200 300 Energy ( eV) r 400 500 F i g u r e 2.5 P r inc ipa l features o f an E E L spectrum, (adapted from ref. 17) insulators. 1 Typ ica l ly , p lasmons i n v o l v i n g both a and n electrons appear between 5 and 50 e V for most materials. A r o m a t i c organic materials also show a distinct 7T.-7T.* p lasmon around 7 e V . T h e outer-shell electrons that are excited in these bu lk plasmons are difficult to mode l f rom first pr inciples , especially i n non-metals. They may be more readily treated by consider ing the response o f the whole sol id to the t ravel l ing electron, w h i c h is very s imi la r to the response to a passing photon. The latter is g iven by the complex dielectric function s(oS). I f the sample thickness is k n o w n accurately, the low-loss spectrum may be used to calculate s(co) for the material and compare it to data from optical measurements . 1 9 The correlation between the two is sufficient to a l low the use o f either one to predict the other, and as such E E L S complements and extends opt ical techniques for determining £(co). 42 The bulk plasmons occur w i th in the bu lk o f the material and must be dist inguished from excitations that appear at interfaces, w h i c h are termed surface plasmons. Geometr ies where the electron beam interacts extensively w i t h the surface o f the material can lead to important surface p lasmon effects, typical ly be low 20 e V in energy. Together, the bulk and surface plasmons dominate the low-loss spectrum (< 50 e V ) . P lasmons o f both types can be excited from substantial distances from the sample: up to 8.0 n m for the bu lk p lasmon and over 12 n m for surface p l a s m o n s . 2 0 The fa l l -of f is however exponential and reasonable count ing statistics are obtained up to h a l f these distances. The intensity o f the loss peaks relative to the total spectrum area is a function o f sample thickness. Furthermore, both types o f plasmons can lead to mul t ip le scattering events per electron (plural scattering), leading to the appearance o f mul t ip le peaks in the loss spectrum w i t h a Po isson distr ibut ion. In this case, the single scattering distr ibution can be recovered by deconvolut ing the spectrum (see chapter 5). The beam energies typical ly used in T E M lead to electrons w i t h relativistic speeds, w h i c h may also lose energy through the Cherenkov effect. Th i s loss mode proved to be very important for the samples that were investigated here and is discussed in more detail i n chapter 6. 2.5 E E L S Instrumentation The schematic o f a modern T E M wi th E E L S capabili t ies is shown i n Figure 2.6. The type o f electron source determines the brightness, energy spread and S T E M probe size; a f ie ld-emiss ion source is usually employed for materials science w o r k . The beam is accelerated to the desired energy (usually 200 k V for materials science), and magnetic lenses are used to form the probe. In T E M mode, the probe is diffuse and i l luminates a large part o f the sample; the convergence semi-angle a is also smal l , w h i c h is equivalent to nearly parallel i l luminat ion . The 43 T E M ~^Ss> Probe forming optics Electron source Sample G I F Energy filtering slit CCD detector F i g u r e 2.6 Schematic o f post -column Gatan Imaging Fil ter on a T E M . transmitted electrons are transferred by further magnetic lenses to the Gatan imag ing filter ( G I F ) . 2 1 The G I F consists o f a magnetic p r i s m spectrometer coupled to a charge-coupled device ( C C D ) detector by mult ipole magnetic lenses. In imag ing mode, the T E M image is formed on the C C D ; i f energy filtering is required, a slit is inserted after the spectrometer to select electrons w i t h the desired energy loss. In spectroscopy mode, the loss spectrum is projected on the C C D . In S T E M mode, the probe is focused to a spot that can be as smal l as 0.2 n m in diameter. T h i s strong focusing entails a larger a (>10 mrad). The image is recorded us ing a dark-field detector ( w h i c h detects electrons that are strongly scattered by the sample), thereby a l l owing the m a i n electron beam to continue into the G I F where the loss spectrum is measured. Thus it is straightforward to perform E E L S experiments w i t h the microscope i n S T E M mode: the electron 44 probe is scanned across the sample to produce the image; the probe may then be rel iably posi t ioned in the location o f interest on the sample to record a loss spectrum. 2.6 A p p r o a c h e s to E E L S D a t a A c q u i s i t i o n The opt imal approach to chemica l analysis by E E L S depends on the complexi ty o f the loss spectrum. In T E M mode, the energy o f the image-forming electron beam may be filtered us ing a slit o f set w i d t h (energy-filtered T E M , E F T E M ) . E lementa l distributions may be imaged directly by fi l tering on the appropriate spectral feature (Figure 2.7). Th i s approach is generally useful when a l imited number o f well-understood components need to be mapped over a large area. The energy resolution is set by the slit w i d t h and is usually 1 e V or greater. The spatial resolution is essentially the same as the T E M , but the magnif icat ion is l imi ted by the presence o f t h e s l i t t o ~ 1 0 0 , 0 0 0 x . In S T E M mode, loss spectra may be acquired at specific points, lines or areas (Figure 2.7). T h i s a l lows detailed examinat ion o f specific areas on the sample. The spatial resolution is l imi ted to ~1 n m by the electron probe size and scattering w i t h i n the sample; the effect o f the latter may be l imi ted by restricting the col lect ion angle. The energy resolution in S T E M mode is determined by the energy spread o f the electron beam. W i t h a field-emission electron source, this spread is about 0.6 to 0.9 e V , depending on the current output o f the electron source. The process o f acqui r ing a complete loss spectrum for each point in an area o f interest is cal led spectrum imaging . T h i s a l lows detailed processing o f the spectra over the whole image and is useful in cases where it is not possible to separate the requisite spectral information by E F T E M . However , this process can be fairly t ime-consuming. 45 E , F i g u r e 2.7 Illustration o f chemica l analysis o f a two-component sample by S T E M / E E L S over a set o f points and E F T E M over the whole image. 2.7 Q u a n t i t a t i v e A n a l y s i s o f E E L S S p e c t r a E E L S and E F T E M are most c o m m o n l y used to provide qualitative analysis o f material composi t ion . Quantif icat ion o f the elemental or phase composi t ion is less straightforward. A n accurate thickness map must be first generated, to correct for the thickness dependence o f the losses. P lura l scattering effects must be also be el iminated. The scattering cross-section o f each component, at a specific ionizat ion edge or p lasmon, must be determined through calculat ion or reference samples. A complete discussion o f the quantification process is provided by Eger ton . 1 The focus o f this w o r k was on the qualitative level and no attempts were made to quantify material composi t ion by these techniques. 46 2.8 Application of EELS to Organic Materials B o t h E E L S and E F T E M have been used to investigate many organic materials. In many cases, special precautions are taken to l imi t the radiation damage to the samples, e.g., by us ing a diffuse beam, col lect ing over many different areas, and keeping the sample at cryogenic temperatures. There is also st i l l debate over whether a diffuse beam (i.e., E F T E M ) or a focused probe (i.e., S T E M ) causes the least amount o f damage to organic ma te r i a l s . 2 2 It has become evident that each material must be studied separately to understand its beam stability. The first important study o f organic materials by E E L S , by Isaacson, investigated thin f i lms o f the nucleic ac id bases . 2 3 The complex dielectric function e(co) was obtained wi th a resolution o f ~0.25 e V . The der ived absorption coefficient showed very good agreement w i t h U V absorption data, provided that the electron dose was carefully l imi ted to m i n i m i z e damage. M o r e recently, b io logica l cel l-staining chromophores have also been mapped at h igh resolution by filtering on the low-loss peaks o f the c h r o m o p h o r e s . 2 4 The resolution l imi t o f this technique was estimated to be 1.6 n m , based on the edge sharpness o f larger s t ructures . 2 5 Carbon nanotubes have been extensively studied by E E L S , 2 6 " 2 8 as they are fairly stable to the beam. The low-loss spectra show the expected % p lasmon around 4 to 6 e V and the bulk p lasmon near 23 e V . E F T E M has proved to be a very useful tool for the study o f po lymer b l e n d s . 7 - 2 9 The ionizat ion edges o f minor component elements (nitrogen for polyamides , oxygen for poly(methyl methacrylate), sulfur for poly(phenylene sulfide)) may be used wi th some effectiveness to determine po lymer d i s t r i b u t i o n s . 3 0 ' 3 1 B u t certain phases can show trace oxygen contaminat ion (e.g., polybutadiene) that complicates the identification process. Po lymers w i t h aromatic groups, such as polystyrene, can be readily identified i n a blend due to the dist inctive n-%* p lasmon at 7 e V . 3 2 The polymers may also be stained in various ways to increase contrast; this comes wi th the 47 advantage o f phase stabilization by the c ross - l ink ing o f chains and the disadvantage o f altered chemica l states and possible s h r i n k a g e . 3 2 These reports clearly indicate that conjugated polymers can be readily dist inguished wi th E E L S and E F T E M by the dist inct ive n p lasmon o f aromatic rings; analysis us ing the carbon K -edge is also possible when there is sufficient material present. The analysis o f the low-loss spectrum requires accurate removal o f the zero-loss peak; this is discussed in more detail in chapter 5. 2.9 C o n c l u s i o n E E L S and E F T E M are both h ighly suited for h igh resolution characterization o f nanostructured conjugated polymers . Th i s applicat ion is successfully demonstrated in chapter 3 on a conjugated polymer/mesoporous s i l i ca composite, and less successfully in chapters 5 and 7 on a conjugated polymer/porous a lumina composite . These latter results showed h o w the sample geometry and relativistic effects can play a large role in the low-loss spectrum, m a k i n g chemica l analysis less straightforward. 48 References 1. Eger ton, R . F . Electron energy-loss spectroscopy in the electron microscope; 2 n d ed.; P l e n u m Press: N e w Y o r k , 1996. 2. Watanabe, Y . ; H e u n , S.; Salvia t i , G . ; Y a m a m o t o , N . , Eds . Nanoscale spectroscopy and its applications to semiconductor research; Springer. B e r l i n , 2 0 0 0 3. Golds te in , J . I. Scanning electron microscopy and X-ray microanalysis; 3rd ed.; K l u w e r A c a d e m i c / P l e n u m Publ ishers : N e w Y o r k , 2003 . 4. Ful tz , B . Transmission electron microscopy and dijfractometry of materials; Springer: N e w Y o r k , 2001 . 5. O z k a y a , D . ; Thomas , J . M . ; Shephard, D . S.; Maschmeyer , T . ; Johnson, B . F . G . ; Sankar, G . ; O l d r o y d , R . Inst. Phys. Conf. Ser. 1997, 153, 403 . 6. B lanford , C . F . ; Carter, C . B . Microsc. Microanal. 2003, 9, 245 . 7. Correa , C . A ; Hage, E . Polymer 1999, 40, 2171 . 8. M c C a f f r e y , J . P . Ultramicroscopy 1991, 38, 149. 9. W a l c k , S. D . ; M c C a f f r e y , J . P . Thin Solid Films 1997, 308, 399 . 10. Suder, S.; Faunce, C . A . ; Donne l ly , S. E . Thin Solid Films 1997, 304, 157. 11. M a l i s , T . F . ; Steele, D . Ultramicrotomy for materials science; Spec imen preparation for Transmiss ion Elec t ron M i c r o s c o p y o f Mater ia l s II; Mater ia ls Research Society: Pi t tsburgh, P A , U . S . A . , 1990, p. 3. 12. Quintana, C . Micron 1997, 28, 217. 13 . Y a n g , H . ; K u p e r m a n , A . ; C o o m b s , N . ; Mamicheafara , S.; O z i n , G . A . Nature 1996, 379, 703. 14. Y a n g , FL; C o o m b s , N.; Soko lov , I.; O z i n , G . A . J. Mater. Chem. 1997, 7, 1285. 1 5 . Furneaux, R . C ; Thompson , G . E . ; W o o d , G . C . Corros.Sci. 1978,18, 853. 49 16. Spence, J . C . H . ; H o w e l l s , M . R . Ultramicroscopy 2002, 93, 213 . 17. B r y d s o n , R . Electron Energy Loss Spectroscopy; B I O S , i n association w i t h the R o y a l M i c r o s c o p y Society: O x f o r d , 2 0 0 1 ; V o l . 48 . 18. Eger ton, R . F . ; M a l a c , M . Ultramicroscopy 2002, 92, 47 . 19. Danie ls , J . ; Festenberg, C . v . ; Raether, H . ; Zeppenfeld, K . Springer Tracts Mod. Phys. 1970, 54, 78. 20 . M u l l e r , D . A . ; S i l cox , J . Ultramicroscopy 1995, 59, 195. 2 1 . Gatan; Pleasanton, C A , U . S . A . (www.gatan .com). 22 . Var lo t , K . ; M a r t i n , J . M . ; Quet, C . Micron 2001, 32, 3 7 1 . 23 . Isaacson, M . J. Chem. Phys. 1972, 56, 1803. 24 . Ottensmeyer, F . P . ; D a v i s , J . A . ; Heng , Y . M . ; Barfels , M . M . G . Microbeam Analysis 2000; Institute o f Phys ics Conference Series; Institute o f Phys ics Pub l i sh ing ; Br i s t o l ; no. N u m b e r ; p . 181. 25 . Barfels , M . M . G . ; J iang, X . G . ; H e n g , Y . M . ; Arsenaul t , A . L . ; Ottensmeyer, F . P . Micron 1998, 29, 97. 26. K u z u o , R . ; Terauchi , M . ; Tanaka, M . Jpn. J. Appl. Phys. 1992, 31, L I 484. 27 . B u r s i l l , L . A . ; Stadelmann, P . A . ; Peng, J . 1.; S., P . Phys. Rev. B 1994, 49, 2882. 28 . Stephan, O . ; K o c i a k , M . ; Henrard , L . ; Suenaga, K . ; Gloter , A . ; Tence, M . ; Sandre, E . ; C o l l i e x , C . J. Electron. Spectrosc. Relat. Phenom. 2001, 114, 209 . 29 . D u Chesne, A . Macromol. Chem. Phys. 1999, 200, 1813. 30 . H o r i u c h i , S.; Yase , K . ; Ki t ano , T. ; Higash ida , N . ; Oug i zawa , T . Polym. J. 1997, 29, 380. 3 1 . H o r i u c h i , S.; Ishi i , Y . Polym. J. 2000, 32, 339 . 32. Var lo t , K . ; M a r t i n , J . M . ; Quet, C . Polymer 2000, 41, 4599 . 50 CHAPTER 3 A PPV/MCM-41 Composite Material M a n y o f the host properties desired for opt imal conjugated po lymer encapsulation can be found in self-assembled porous inorganic materials. These materials have been o f longstanding scientific and industrial interest, due to their large internal surface area, large sorption capacity, thermal stability and catalytic activity towards smal l molecules . 1 M o r e recently, as instruments for structural and chemica l analysis on the nanometre scale have become commonplace , porous materials have been investigated as templates for novel nanostructured materials. A b r i e f overv iew o f the properties and applications o f self-assembled inorganic porous materials is g iven here, w i t h an emphasis on the mesoporous material M C M - 4 1 . M C M - 4 1 , in its pure s i l i ca form, possesses many o f the properties o f an ideal host for conjugated polymers : it has narrow and al igned channels, w i t h optical ly transparent and electrically insulat ing wal ls . Th i s made it attractive to us for an ini t ia l study on polymer/host composi te material synthesis and characterization. The literature already contained important w o r k on the synthesis o f p o l y m e r / M C M - 4 1 composite materials, in particular w i t h non-luminescent conjugated polymers . The preparation and characterization o f a new composite material i n v o l v i n g M C M - 4 1 as a host for the luminescent conjugated po lymer p o l y ( l , 4 -phenylene vinylene) ( P P V , Figure 3.1) is described in this chapter. 51 ro4 Figure 3.1 C h e m i c a l structure o f p o l y ( l ,4-phenylene vinylene) ( P P V ) . 3.1 Ordered Porous Host Materials Porous materials are categorized according to their pore size into three classes: microporous , mesoporous and macroporous (Table 3.1). Fo r the purpose o f po lymer encapsulation and orientation, the materials o f interest w o u l d have an ordered pore structure w i t h a pore diameter be ing near the boundary between the microporous and mesoporous classes (~2 nm) . Table 3.1 I U P A C classification o f porous materials by pore s i ze . 1 Pore Diameter Des ignat ion < 2 n m microporous 2 - 50 n m mesoporous > 5 0 n m macroporous 3.1.1 Zeolites The major class o f ordered microporous materials is the zeolite f ami ly . 1 Zeolites are crystalline aluminosil icates possessing a w i d e variety o f pore structures, i nc lud ing one, two and three d imens iona l pore networks (Figure 3 .2) . 2 A number o f naturally occur r ing zeolites are k n o w n but a large number o f synthetic zeolites have been discovered since the 1940's . The synthesis usually consists o f hydrothermal crystal l izat ion o f a reactive gel at elevated temperature i n a sealed vesse l . 3 The inorganic zeolite f ramework condenses into a structure determined by the templat ing (or structure-directing) agents, w h i c h can include water molecules , 52 A B C F i g u r e 3.2 Examples o f pore topologies w i th (a) 1-D, (b) 2 - D and (c) 3-D connectivity. organic cations and salts. F o l l o w i n g the synthesis, the trapped templating agent can be removed by calcinat ion or ion-exchange to expose the pore structure. The crystalline nature o f zeolites a l lows very precise structural characterization us ing X - r a y diffraction and solid-state nuclear magnetic resonance. The well-defined structure also leads to h igh selectivity in catalytic reactions. The largest pore size reported to date in zeolites is on the order o f 1.2 n m in diameter . 4 These zeolites can accommodate smaller polymers (e.g. polystyrene 5 ) but not larger polymers (in particular conjugated polymers w i t h so lub i l i z ing side-chains). A further compl ica t ion arises from the smal l crystallite size exhibi ted by zeolites: the preparation o f continuous thin films is not readily possible, although recent results in this area appear p r o m i s i n g . 6 For these reasons, mesoporous materials are more useful as general hosts for conjugated polymers. 3.1.2 M e s o p o r o u s M a t e r i a l s Mesoporous host materials are desirable for many applications, but were not available wi th wel l -def ined pore structures unt i l 1992, when B e c k et al. reported the discovery o f the M 4 1 S family o f mesoporous aluminosi l icate ma te r i a l s . 7 ' 8 Whereas zeolites are templated by 53 smal l organic molecules , M 4 1 S materials were shown to be templated by l iqu id crystalline phases formed by straight-chain surfactants (Figure 3.3). O f particular interest in this family was the si l iceous material M C M - 4 1 , w h i c h shows uni form 1-D channels packed i n a 2 - D hexagonal array (Figure 3.4). Th i s simultaneously endows the material w i t h a large surface area (1000 m 2 g" 1, almost a l l internal) and a large pore vo lume (0.8 c m 3 g" 1). The straight channels are ideal for po lymer guest accommodat ion , though 180° defects in channel direction are possible. The pore wal l s usually consist o f s i l i ca but other oxides can be readily incorporated into the framework Th i s general approach to templat ing has since been appl ied to the synthesis o f many new porous mater ia ls . 9 The synthesis o f M C M - 4 1 proceeds from a mixture o f water, surfactant, s i l i ca source and an ac id or base catalyst; this forms a gel that is then heated in a sealed container. The spacing o f the hexagonal phase can be readily altered by the choice o f surfactant chain length and the addi t ion o f organic swel l ing agents. 8 Var ia t ions on the s i l i ca s o u r c e , 1 1 p H o f the m i x t u r e , 1 2 , c o u n t e r - i o n s 8 ' 1 3 and reaction tempera ture 1 4 also affect the f inal structure. Thus pore diameters from 1.5 to 10 n m may be obtained; recent developments w i t h po lymer ic templating agents have A B C D E F i g u r e 3.3 Stages i n the formation o f M C M - 4 1 : (a) surfactant mice l le w i t h the hydrophobic chains in the center and the polar head groups l in ing the outside, (b) cy l indr ica l surfactant mice l le , (c) hexagonal array formed cooperatively by surfactant micel les and silicate species, (d) condensed material (as-made M C M - 4 1 ) , (e) calcined material, (adapted from ref. 1 0 ) 54 F i g u r e 3.4 Transmiss ion electron micrograph o f M C M - 4 1 material obtained us ing C i 6 H 3 3 ( C H 3 ) 3 N C l surfactant, showing hexagonal lattice spacing and w a l l thickness, (courtesy o f G . Bot ton, C A N M E T / N a t u r a l Resources Canada) extended this range to 50 n m . 1 5 > 1 6 A t the complet ion o f the synthesis, the organic templating agent is normal ly removed by calcinat ion, a process that also induces further condensation o f the s i l ica framework through dehydration. Solvent e x t r a c t i o n 1 7 and supercri t ical f luid e x t r a c t i o n 1 8 have also been shown to be effective in r emoving the templat ing agents. The final form o f the material is then a powder w i t h a typica l particle size o f 1 u m . The formation o f M C M - 4 1 thin f i lms has also been extensively investigated; it is discussed in chapter 4. The h igh degree o f order in the M 4 1 S family o f materials derives from the cooperative arrangement o f the cy l indr ica l surfactant micel les and the silicate ions in the l iqu id crystalline phase. Th i s long-range order i n the pack ing o f the channels can be observed by X - r a y diffraction. However , the s i l i ca framework i tself is amorphous and microporous, and its inner surface has many slightly different chemica l sites. Thus M C M - 4 1 is referred as a zeol i t ic material to reflect that is an aluminosi l icate material w i th only long-range ordering. The loss o f 55 crystall inity, when compared to the zeolites, makes characterisation more difficult; this is offset, i n the context o f guest incorporation, by the benefits o f tunable pore diameters. Mesoporous sil ica-based materials may be readily modi f i ed since there are many accessible surface hydroxy groups: it has been found that 2 6 - 3 0 % o f al l S i atoms in M C M - 4 1 bear a surface hydroxy g r o u p . 1 9 The surface derivatization can proceed by wel l -developed silane c h e m i s t r y . 8 ' 2 0 A d d i t i o n a l functionality can be achieved by substituting other elements into the f r a m e w o r k , 2 1 and us ing organical ly-modif ied framework sou rces . 2 2 This f lexibi l i ty and the large internal surface area make M C M - 4 1 an attractive host material , and the creation o f composi te materials based on M C M - 4 1 has been reviewed r e c e n t l y . 2 2 ' 2 3 In many studies, it has not always been evident that M C M - 4 1 provides any substantial advantage over disordered porous materials, except poss ibly i n its somewhat larger specific surface area. However , the wel l-def ined channels are evidently necessary for the role o f template for nanometre-scale wires and ordered encapsulation. 3.2 Characterization of MCM-41 Materials The characterization techniques described in chapter 2 are direct ly applicable to M C M -4 1 , especially h igh resolution t ransmission electron microscopy ( T E M ) , electron energy-loss spectroscopy ( E E L S ) and energy-filtered t ransmission electron microscopy ( E F T E M ) , w h i c h a l low local chemica l characterization. A number o f other techniques are used to measure material-averaged properties. W h i l e there is general agreement on the formation mechan i sm for M C M - 4 1 , there is widespread disagreement on its exact structure. Th i s seems to be due to the fact that the mesoporous 2 - D hexagonal phase can be readily made under a variety o f different condit ions us ing different starting materials, and the exact characterization o f the w a l l structure is difficult . 56 3.2.1 D i f f r a c t i o n T e c h n i q u e s Diffraction is used to identify the presence of an ordered phase. The interplanar spacing dhki associated with a plane identified by the Mil ler indices hkl must satisfy the Bragg condition: 2dhkls\n0 = nA, « = 1,2,3,... (Eq. 3.1) For a hexagonal system, diffraction cannot be observed between planes when / is odd, or (h + 2k) = 3n, where n is an integer. However, / = 0 for a 2-D system, so peaks are expected for all the planes (100), (110), (200), (210), etc. The 2-D interplanar spacing dm is related to the lattice constant as: A - a ahk0 ~ 4 / , 2 , 2 N (Eq. 3.2) - ( h + h k + k ) v H ' Powder X-ray diffraction is routinely used to characterize M C M - 4 1 materials. Neutron scattering has also been used to investigate the pore structure in more d e t a i l . 2 4 ' 2 5 While X-ray powder diffraction data is usually plotted against diffraction angle (26), neutron scattering data is presented as a function of the magnitude of the scattering vector, Q: ^ 4;rsinf? , ^ „ „N Q = (Eq.3.3) A Diffraction techniques may be used to investigate the presence of molecular guests within the channels of M C M - 4 1 i f there is contrast matching between the guest and the host. In the case of X-ray diffraction, sufficient electron density to match the pore walls is obtained for halogenated organic compounds. 2 6 For neutron scattering, contrast matching has been reported for 59% deuterated benzene in M C M - 4 1 2 7 Therefore, not all guests can be readily detected using diffraction techniques. 57 3.2.2 P h y s i s o r p t i o n Physisorpt ion is w i d e l y used to characterize the surface and pore structure o f mesporous ma te r i a l s . 2 8 Adsorbates o f different size and chemica l functionality are used to probe the structure and reactivity o f the material surface. Mic roporos i ty is generally probed us ing argon wh i l e mesoporosity is investigated w i t h nitrogen. The nitrogen adsorption isotherm can be separated into three regimes: (1) micropore fd l ing and monolayer formation, (2) mult i layer formation, and (3) capil lary condensation and further superficial adsorption (Figure 3.5). Reg ime (2) yields the total surface area o f the sample, w h i c h is usually determined us ing the Brunauer-"O C D - Q s_ o CO T J < C D E o > 600 500 400 300 200 100 B E T W G A n a l y s i s © o o CD G B J H A n a l y s i s o 0.0 0.2 0.4 0.6 0.8 1.0 P/Pn F i g u r e 3.5 N i t rogen (o) adsorption and ( • ) desorption isotherms for M C M - 4 1 . The lack o f hysteresis is characteristic o f M C M - 4 1 materials. The points used for total surface area ( B E T ) and pore size distr ibution ( B J H ) analysis are indicated. 58 Emmet t -Tel le r ( B E T ) a p p r o a c h . 2 9 R e g i m e (3) is usually analysed to determine the pore size distr ibution, w h i c h is the most important and also most debated property o f mesoporous materials. The classic determination o f pore size distributions from adsorption isotherms is that o f Barrett, Joyner and Ha lenda ( B J H a n a l y s i s ) ; 3 0 recently more sophisticated approaches have been shown to y i e ld better absolute agreement w i t h other techniques for determining size d i s t r i b u t i o n s . 3 1 ' 3 2 In the B J H approach, the thickness o f the adsorbed nitrogen layer for a g iven relative pressure is g iven by the Hark ins -Jura e q u a t i o n 3 3 wi th the fo l lowing empi r i ca l parameters calibrated for M C M - 4 1 materials: 3 4 0.1 -10.3968 60.65 0 . 0 3 0 7 1 - l o g 2-+ 0.3 nm ( E q . 3.4) However , there is no single unambiguous technique for measur ing this distr ibution for mesoporous materials. The variations in the structure o f M C M - 4 1 due to differing synthesis condi t ions make it diff icul t to compare results between different studies i n the literature, and lead to contradictory conclusions about the pore s t r u c t u r e . 1 6 - 2 4 ' 3 5 ' 3 6 Nevertheless, M C M - 4 1 behaves as an ideal adsorbent and can be used as a reference m a t e r i a l . 3 4 In this work , B J H analysis is used to quantify the shift in the capil lary condensation point in the nitrogen adsorption isotherm. 3.2.3 O t h e r T e c h n i q u e s Thermogravimetr ic analysis ( T G A ) is used to reveal the organic content o f composite materials based on M C M - 4 1 . The host i tself shows no mass loss up to 1000 ° C . The thermal degradation processes for an organic material occur at specific temperatures and this may be 59 used to determine the mass content o f specific organic components in the composite . The derivative o f the T G A data is used to emphasize the presence o f different degradation processes, and these can be fitted satisfactorily by Gaussian functions. T h i s a l lows the mass content o f each process to be determined more accurately. Fourier-transform infrared spectroscopy (FT- IR) can be used to identify the organic components through their characteristic molecular vibrations, as the s i l i ca matr ix is most ly transparent to infrared radiation. 3.3 Polymerization within MCM-41 Reports on po lymer inc lus ion i n M C M - 4 1 appeared i n the m i d 1990's , i n v o l v i n g both conjugated and insulat ing polymers . In situ polymer iza t ion wi th in the channels o f M C M - 4 1 was reported ini t ia l ly by W u and B e i n . 3 7 Oxida t ive polymer iza t ion o f anil ine was achieved by first condensing the monomer from the vapour phase into the host channels. The loaded host was then soaked in an aqueous solution o f oxidant, w h i c h produced encapsulated po lymer chains. Ac ry lon i t r i l e was po lymer ized s imi lar ly , us ing instead a solution o f radical i n i t i a t o r . 3 8 Unge r e t a l . prepared other free-radical initiated polymers in M C M - 4 1 : polystyrene, poly(methyl methacrylate) and po lyv iny lace ta t e . 3 9 The monomers were loaded into M C M - 4 1 through vapour exchange, and the gas-phase radical initiator was subsequently diffused into the loaded M C M -4 1 . 3.3.1 PPV in MCM-41 The introduction o f P P V into M C M - 4 1 must also proceed through in situ synthesis because o f the insoluble and infusible nature o f the polymer . The polymer iza t ion o f P P V can be 60 carried out by numerous rou te s , 4 0 the simplest being a base-initiated condensation k n o w n as the G i l c h r o u t e 4 1 (Figure 3.6). However , this route was not exploitable w i t h i n the M C M - 4 1 host: wh i l e the monomer could be readily loaded into M C M - 4 1 through subl imat ion, the subsequent introduction o f a base o f sufficient strength was not possible. Aqueous bases attacked and dissolved the host, and non-aqueous bases so lub i l i zed the monomer and extracted it f rom the pores before polymer iza t ion could occur. Other polymer iza t ion routes were investigated and an elegant solution to this p roblem was found through the w o r k o f K u m a r et al.,42 w h o had prepared P P V wi th in the pores o f V y c o r , a disordered porous glass. Th i s approach used a more reactive monomer , xylylene bis(tetrahydrothiophenium chloride) , w h i c h could be po lymer ized by deprotonated surface hydroxy groups. In s imi la r fashion, M C M - 4 1 was converted to a basic fo rm by deprotonating its surface hydroxy groups wi th a non-aqueous base ( tetrabutylammonium hydroxide ( T B A O H ) in methanol) and isolated. Th i s activated form o f M C M - 4 1 thus contained the ini t iat ing base w i t h i n its channels. Po lymer confinement occured through the rapid polymer iza t ion o f the monomers wi th in the host channels (Figure 3.7). The key evidence to each step in this process is based on F T - I R spectroscopy, T G A , nitrogen physisorpt ion, E E L S and E F T E M . X - r a y and neutron diffraction experiments p rov ided no addit ional information on the composite materials. NBuOK/THF F i g u r e 3.6 The G i l c h route to P P V starting from dichloro-p-xylene. 61 F i g u r e 3.7 Synthetic scheme for the preparation o f P P V / M C M - 4 1 hybr id material . O n l y one cha in is shown in the pore for clarity. 3.4 E x p e r i m e n t a l Resu l t s The pure s i l i ca M C M - 4 1 host was synthesized fo l lowing a literature p rocedu re . 4 3 Characterizat ion w i t h X - r a y diffraction and nitrogen physisorpt ion indicated substantial variations from batch to batch in the lattice constant (4.3 to 4.5 nm) and B J H pore diameter (3.1 to 3.6 nm) . Thus comparisons were only made w i t h materials prepared in the same batch. The B E T surface area was (1.0 ± 0.1) x l O 3 m 2 g" 1, and the total pore vo lume was 0.8 c m 3 g"1. The F T -I R spectrum o f the empty host showed absorption bands for the surface hydroxy groups from 3700 to 3000 cm" 1 , the s i l i ca f ramework from 1100 to 600 cm" 1 , and adsorbed water at 1700 cm" 1 . 62 F o l l o w i n g T B A O H treatment o f M C M - 4 1 , T G A showed 10% water content and a 3 8 % mass loss between 100 and 3 0 0 ° C ; this was ascribed to the decomposi t ion o f the T B A counter-ion (Figure 3.8(a)). The mass content o f T B A suggests that 1 5 % o f a l l S i atoms in M C M - 4 1 had an associated w i t h a T B A counter-ion. The presence o f the counter-ion reduced the B J H pore diameter by 1.2 n m (Figure 3.8(b)). The polymer iza t ion step then yie lded a bright green powder . Excess monomer and possible side-products could be easily separated from the powder by wash ing dur ing filtration. The rma l convers ion o f the po lymer under v a c u u m led to a fluorescent ye l low powder . X - r a y diffraction o f the resulting P P V / M C M - 4 1 composi te indicated that the sample order remained (Figure 3.9(b)). The result o f neutron scattering experiments carr ied out by L . Fan , Z . T u n and J . Y o u n g on M C M - 4 1 and two polymer-conta ining samples is shown i n Figure 3.9(b). The strong scattering peak at Q « 0.16 is in agreement w i th the X - r a y data. The presence o f po lymer in the sample d id not seem to alter this peak substantially. 63 100-f 50 0.30 ^ 0.25 g CD -1 0.20 2. 5L H0.15 I co 0.10 § - 0.05 Q - 0.00 ->—i—1—i—1—I—1—I—'—i—1—I—'—I—1—I—1—I—r 0 100 200 300 400 500 600 700 800 900 1000 Temperature ( ° C ) Relative Pressure, p/p0 0.0 0.2 0.4 0.6 0.8 1.0 0.05 600 500-o> 400-I C O E "S 300 •a l_ O CO " 200 CD E _2 > 100 OA B o o o o o o o o o ' 1.5 , o ° TBA/MCM-41 D ~ 2.2 nm a 0.18 TJ o —i CD < O c 3 CD 3 Cd C Q 3 Pore Diameter (nm) F i g u r e 3 . 8 (a) Thermogravimetr ic analysis o f T B A O H - t r e a t e d M C M - 4 1 , (b) (o) , ( • ) nitrogen adsorption isotherm and (A) , ( A ) B J H pore size distr ibution for empty and T B A O H - t r e a t e d M C M - 4 1 , respectively. 64 F i g u r e 3.9 (a) Powder X - r a y diffraction pattern and (b) neutron scattering data for ( • ) M C M -4 1 , P P V / M C M - 4 1 ( • ) sample 1 and ( • ) sample 2. 65 T G A o f the composite material (Figure 3.10(a)) showed an ini t ia l water desorption at 4 0 ° C , fo l lowed by a 1% weight loss at 2 5 0 ° C , w h i c h is most l ikely due to residual T B A in the material that d i d not decompose dur ing the thermal convers ion process. The po lymer degradation began at 3 5 0 ° C and showed at least two different decomposi t ion processes, at 5 2 5 ° C and 7 3 0 ° C . Samples consistently showed a po lymer content o f 8 ± 2 wt. % . A n a l y s i s o f the 2 2 1 nitrogen adsorption isotherm (not shown) indicated a reduced B E T surface area ( 8 . 5 x 1 0 ' mV) and pore vo lume (0.51 ± 0.04 cm 3 g" ' ) . The B J H pore diameter was observed to decrease by 0.3 ± 0.1 n m (Figure 3.10(b)). The F T - I R spectrum (Figure 3.11(a)) showed new bands characteristic o f P P V . The band at 3025 cm" 1 is associated w i t h the C - H stretch in fnms'-vinylene, wh i l e the bands at 1517 and 1422 cm" 1 are due to 1,4-phenylene r ing stretching m o d e s . 4 4 The distr ibution o f po lymer was investigated by E E L S and E F T E M . The low loss spectra o f the empty and polymer-f i l led M C M -41 are compared in Figure 3.11(b). The zero-loss peak has been removed by careful subtraction o f a reference peak, revealing that only P P V / M C M - 4 1 has losses be low 8 e V . E F T E M images acquired at 0, 6 and 12 e V w i t h a 2 e V w i n d o w were used to image the distr ibution o f the losses (Figure 3.12). The U V / V i s absorbance spectrum o f the composite showed an onset at 500 n m and a peak at 420 n m , wh i l e the photoluminescence ( P L ) spectrum showed peaks at 517, 547 and 590 n m (Figure 3.13(a)). The temperature-dependent P L spectra o f encapsulated P P V were measured in collaboration wi th M . M c C u t c h e o n and J . Y o u n g , us ing an excitation wavelength o f 386 n m (Figure 3.13(b)). 66 100 4 98 H CD 0.008 0.007 - | 0.006-'o> 0.005 • 0 .004 -0 . 0 0 3 -0.002 -0.001 -E c E <u E O > CD i O 0. 0.000-4 0.12 4 0.10 03 4 0.08 -4 0.06 ^ . o CO CO 4 0.04 0.02 4 0.00 o o i—•—i—1—r 100 200 300 400 500 600 700 800 900 1000 Temperature (°C) B PPV/MCM-41 D~2 . 9nro D - 2 . 8 nm MCM-41 D-3.1 nm 1.5 2.5 3.0 3.5 4.0 Pore Diameter (nm) - i — 4.5 5.0 F i g u r e 3.10 (a) Thermogravimetr ic analysis o f P P V / M C M - 4 1 , and (b) B J H pore size distr ibution o f ( • ) empty M C M - 4 1 , P P V / M C M - 4 1 (A) sample 1 and (o) sample 2. 67 3000 2500 2000 Wavenumber (cm1) 8000 1500 0 10 20 30 40 50 60 Energy (eV) F i g u r e 3.11 Spectral signature o f (—) P P V / M C M - 4 1 and (—) M C M - 4 1 , us ing (a) F T - I R and (b) E E L S . 68 A B F i g u r e 3.12 E F T E M images for (a) M C M - 4 1 and (b) P P V / M C M - 4 1 , filtered wi th a 2 e V w i n d o w centered on the g iven energies. The 6 e V image reveals the polymer distr ibution in the composite material; the contrast has been enhanced on the inset. Note the presence o f the lacey carbon support in (a), w h i c h also shows a strong n-ic p lasmon. The ma in scale bar is 50 n m and the inset scale bar is 20 n m . 69 I—1—I—1—I—1—I—'—I—'—I—1—I—1—I—1—I—1—I 250 300 350 400 450 500 550 600 650 700 Wavelength (nm) l — • — i — 1 — i — 1 — i — « — i — 1 — i — • — i — • — r — i — • — i — • — i — 1 — i — 1 — i — • — i — 1 — i — 1 — i — 480 510 540 570 600 630 660 690 Wavelength (nm) F i g u r e 3.13 (a) R o o m temperature (—) U V / V i s absorbance and ( ) photoluminescence o f P P V / M C M - 4 1 composite , (b) temperature-dependent photoluminescence spectra o f P P V / M C M -4 1 . 70 3.5 D i s c u s s i o n The presence o f P P V is clearly indicated by the U V / V i s absorbance, fluorescence and F T - I R spectrum o f the composite sample. The central issue in the characterization o f this material is whether the po lymer chains actually reside in the channels o f M C M - 4 1 , or possibly form a th in coat on the outside o f the host particles. 3.5.1 T h e r m o g r a v i m e t r i c A n a l y s i s The T G A data for P P V / M C M - 4 1 (Figure 3.10(a)) clearly indicate that the thermal behaviour o f the po lymer is altered from its unencapsulated form, w h i c h shows only one degradation process at 5 4 0 ° C . 4 5 The higher temperature degradation process at 7 3 0 ° C has also been observed for P P V encapsulated i n m o n t m o r i l l o n i t e . 4 6 Th i s strongly suggests the presence o f some encapsulated polymer . The lower degradation process might be associated wi th unencapsulated polymer . A s imple mode l for the composite material can be used to determine the approximate po lymer mass content for a g iven number o f po lymer chains per channel in M C M - 4 1 . W i t h the assumptions that (1) the po lymer chains have no conformational defects, (2) they are oriented paral lel to the channels, (3) the effect o f end-groups can be ignored, and (4) the po lymer on the external surface is m i n i m a l , the mass fraction F o f po lymer in M C M - 4 1 for a g iven number o f chains Nper pore can be calculated, us ing parameters g iven in Table 3.2: J V - 2 . 5 7 x l Q - 2 2 ( E q . 3.5) ~ 1.90 x 1 0 ' 2 0 +N- 2.57 x l O " 2 2 71 Table 3.2 Parameters for calculat ion o f P P V mass fraction in M C M - 4 1 Property Value Source P P V repeat length 0.66 n m Ref. 47 P P V repeat unit mass 102.1 g mol" 1 M C M - 4 1 w a l l density 2.2 g cm" 3 amorphous Si02 M C M - 4 1 lattice constant , 4.3 n m X - r a y diffraction M C M - 4 1 pore diameter 3.1 n m N2 physisorpt ion Values o f F for smal l values o f N are g iven in Table 3.3. Th i s clearly shows that even one po lymer chain per pore w o u l d represent a substantial mass loading o f the sample. The experimental mass content o f 8% suggests that the pores o f M C M - 4 1 contain 6 po lymer chains each on average. I f assumption (4) is inva l id , and some o f the po lymer ( -50%) is externally located as suggested by the T G A , each channel contains 3 polymer chains on average. Table 3.3 Po lymer mass fraction F for Npolymer chains per pore in M C M - 4 1 , us ing E q . 3.5. N 1 2 4 6 8 10 F(%)± 1% 1 3 5 8 10 12 3.5.2 Physisorption Data The reduction o f the B J H pore diameter o f the composi te material , albeit smal l , was observed reproducibly and is significant. The B J H method assumes a cy l indr ica l geometry for the pores and does not provide any information about the arrangement o f the po lymer chains w i t h i n the pores; one possible interpretation is that the po lymer is present as a th in layer on the wal l s o f the pores. A s the P P V backbone is planar, a po lymer chain can lie flat against the pore w a l l and produce a relatively smal l change in the B J H pore diameter. The observed differences in pore diameter reduction could suggest differences in po lymer conformat ion wi th in the pores, leading to a different effect on the isotherm. The smal l amount o f residual T B A ions observed by 72 T G A (1 wt. %) w o u l d be expected to reduced the B J H pore diameter by 0.03 n m , based on the effect o f 38 wt. % T B A content. I f po lymer iza t ion occurred strictly in solut ion, due to proton transfer from the solvent or trace water and consequent migra t ion o f the basic species into the bulk solution, the po lymer might be present strictly as a layer coating the outside o f the M C M - 4 1 particles. The effect on the nitrogen adsorption isotherm w o u l d then be very different: i f the pores were fully b locked , clearly no substantial adsorption w o u l d occur, w h i c h is not the case. Par t ia l blockage o f the pores by a superficial layer w o u l d not alter the observed capil lary condensation point, but w o u l d introduce hysteresis into the isotherm, w h i c h was also not observed. A n external layer that d i d not impede the adsorption process w o u l d not reduce the observed pore diameter, but s imply reduce the specific surface area and pore vo lume due to the presence o f addit ional , non-porous mass in the sample. The relative decrease i n the specific surface area (15%) and pore vo lume (36%) is greater than the addit ional mass (8%) and also in agreement w i t h the relative decrease expected for a cy l indr ica l geometry ( d V / V = 2 d A / A ) . 3.5.3 X-ray and Neutron Diffraction The X - r a y diffraction data indicated that the po lymer ized sample retained the overal l order o f the host material , w i th no substantial shift i n the diffraction peak posit ions. N o attempt was made to interpret the changes in the peak intensities. The neutron scattering results showed only smal l differences between the empty M C M - 4 1 and the P P V / M C M - 4 1 composite. M o d e l l i n g o f M C M - 4 1 coated w i t h a 0.1 nm-th ick coating o f material contrast-matched to the pore wal ls has been shown to decrease the ampli tude o f the m a i n scattering peak and shift the relative posi t ion o f the smaller p e a k . 4 8 The smal l difference observed in the m a i n peak ( Q « 0.16) between the two P P V / M C M - 4 1 samples does 73 not correlate w i t h the B J H pore diameters, and the smaller peak (Q « 0.28) is not sufficiently resolved on a l l samples to determine any shift relative to the m a i n peak. The difference in scattering cross-section between P P V and M C M - 4 1 cannot be determined without further mode l l i ng and it is not possible to state what the effect o f a smal l amount o f P P V i n the pores w o u l d have on the diffraction pattern. A s they stand, these results are not in disagreement w i th the hypothesis o f a smal l number o f chains present in the pores. 3.5.4 E E L S and E F T E M N o r m a l T E M investigation o f the M C M - 4 1 and P P V / M C M - 4 1 particles revealed no readily v i s ib le differences. A substantial surface layer o f po lymer w o u l d have been v i s ib le as an amorphous (structure-free) layer at the particle edges. Fo r 1 u m particles, an external polymer mass content o f 4 % w o u l d give rise to a uni form external layer o f ~7 n m (assuming a po lymer density o f 1 g c m " ) A l s o , no bulk polymer particles were observed. The electron energy-loss spectrum o f P P V / M C M - 4 1 particles clearly showed the loss attributable to the 71-71* p lasmon at 6-7 e V , w h i c h was absent from the empty M C M - 4 1 sample. Losses be low 5 e V were due to the opt ical absorption o f P P V , w h i c h has an onset at 2.5 e V and a peak near 3 e V 4 9 A t energies above 8 e V , the spectra are dominated by the broad bulk p lasmon o f the s i l i ca matr ix centered at 22 e V , i n agreement w i th the value observed for a form o f mesoporous s i l i c a . 5 0 The po lymer bulk p lasmon w o u l d be expected to appear at a s imi lar e n e r g y 5 1 and could not be dist inguished. The shoulders v i s ib le between 10 and 20 e V are most l ike ly due to weak surface plasmons. The stability o f the material seemed very good under a diffuse beam without any special measures to l imi t beam damage. Energy filtering on the 7i-7t* p lasmon is experimental ly most pract ical as the tai l o f the zero-loss peak is substantial be low 5 e V . The 6 e V energy-filtered images o f M C M - 4 1 and 74 P P V / M C M - 4 1 may be compared to determine the po lymer distr ibution in the composite . E m p t y M C M - 4 1 does not show any substantial losses at 6 e V , and the filtered image is correspondingly dark. P P V / M C M - 4 1 , on the other hand, shows 6 e V losses distributed throughout the particle. Intensity variations are seen, and these are due to thickness variations o f the sample, as seen in the 0 e V image. There is no indicat ion o f edge brightness, w h i c h w o u l d have suggested the presence o f a thin polymer coating on the outside o f the particles. The per iodici ty o f the host material is also v is ib le i n the 6 e V image o f P P V / M C M - 4 1 . W h i l e this structure is not fully resolved, it strongly suggests that the po lymer is confined i n the channels. These results corroborate the physisorpt ion and T G A data i n suggesting that the po lymer chains are incorporated into the channels o f M C M - 4 1 . W h i l e the presence o f some external po lymer may be indicated by the T G A data, the E F T E M results indicate that this can only be in the form o f a very thin layer (< 2 nm) , corresponding to < 1 wt. % external polymer . The measurements were carried out on the same samples that were used for neutron scattering, further indicat ing that neutron scattering was not part icularly suited for investigating these samples without resorting to deuteration to match the scattering cross-section o f the host and guest. Th i s demonstrates the abil i ty o f E F T E M to reveal directly the presence o f conjugated po lymer at very h igh resolution. The previous estimate o f the resolution l imi t for mapp ing smal l conjugated molecules was based on the edge sharpness o f a molecular c r y s t a l , 5 2 whereas these results show directly that such h igh resolution is achievable. Furthermore, E F T E M was used to image directly the presence o f an organic guest in the channels o f M C M - 4 1 for the first t ime. The low number o f po lymer chains per pore also suggests that it may be possible to image a single isolated conjugated po lymer chain (perhaps dispersed in a non-conjugated po lymer ic matr ix) by this technique. 75 3.5.5 UV/Vis Absorbance and Photoluminescence The U V / V i s absorbance spectrum o f the composite is very s imi lar to that o f P P V , indicat ing a substantial degree o f polymeriza t ion. H o w e v e r , it d i d not show any o f the structure observed in the absorbance spectrum o f stretch-ordered P P V . 5 3 The P L spectrum at r oom temperature is i n close agreement w i t h literature reports. The structure observed in the spectrum is s imi lar to that o f unencapsulated P P V . The peaks are red-shifted w i t h decreasing temperature; this effect has been attributed in bulk P P V to the reduction o f torsional modes in the polymer chain that affect the effective conjugation l e n g t h . 5 3 The opt ical measurements d i d not reveal any substantially new behaviour i n the composite material . Init ial attempts at measuring P L dynamics were unsuccessful, as the sample rapid ly bleached under laser excitat ion at 386 n m . 3.6 Conclusion The experimental evidence showed that a P P V / M C M - 4 1 composi te was successfully synthesized. C h e m i c a l analysis by E E L S and E F T E M al lowed the presence o f P P V inside the channels to be established unambiguously . The po lymer mass content suggested that 3 to 6 po lymer chains resided in each pore. Howeve r , this material is less than ideal for property comparisons for a number o f reasons. The in situ synthesis leads to a polymer that cannot be fully characterized due to its insolubi l i ty . C o m p a r i s o n o f the P L dynamics wi th no rma l P P V w o u l d be difficult , due to the simultaneous change i n structure and environment, w h i c h has also been an issue in the w o r k o f G i n et al. on P P V in a lyotropic l i qu id crystal h o s t . 5 4 There is most l ike ly some po lymer coating the outside o f the M C M - 4 1 particles, w h i c h w o u l d further complicate the analysis. Th i s was also found i n the w o r k o f Tolbert et al. on M E H - P P V i n an oriented mesoporous s i l i ca host, and the 76 energy migra t ion between po lymer outside and inside the channels was s t u d i e d . 5 5 Such studies are not possible on the composite prepared here because o f the lack o f macroscopic orientation. These complicat ions can be el iminated i f the host is prepared in the form o f an oriented th in film, and i f a soluble form o f P P V is used instead. T h i s a l lows a fully characterized polymer to be used and more meaningful comparisons to be made. Therefore further efforts were directed towards the creation o f an appropriate th in film host. 77 Experimental Details M C M - 4 1 was synthesized according to a literature procedure us ing hexadecyl t r imethylammonium chloride as the surfactant . 4 3 The synthesis was carried out at 80 ° C for 2 days in a Tef lon- l ined stainless steel bomb for most samples. One large batch was synthesized in a polypropylene bottle to provide enough sample for neutron scattering. After wash ing wi th methanol and water, the collected powder was calc ined under air w i th a heating rate o f 1 ° C min" 1 to 5 4 0 ° C and held at that temperature for 6 hours. N i t rogen adsorption analysis was performed on a M i c r o m e r i t i c s A S A P 2010 instrument. Thermogravimetr ic analysis was carried out us ing a T A Instruments T G A 51 under N2 f l ow and a heating rate o f 10 ° C min" 1 . Infrared spectra were obtained from K B r pellets us ing a B O M E M M B 1 5 5 S F T - I R spectrometer. P o w d e r X - r a y patterns were collected on a R i g a k u Rotaflex rotating-anode diffractometer. Chemica l s were obtained from A l d r i c h Inc. X y l y l e n e bis(tetrahydrothiophenium chlor ide) was pur i f ied by recrystall ization from water. E thano l was dr ied over 4 A molecular sieves. A d a p t i n g the w o r k o f K u m a r et a l . , 4 2 the calc ined M C M - 4 1 was dried under vacuum at 100 ° C and then treated wi th 1 M tetrabutylammonium hydroxide in methanol under dry condit ions. Th i s mixture was left to stand for 4 h at room temperature, after w h i c h the basic M C M - 4 1 was filtered o f f and dried under vacuum. The resulting so l id was placed in a 10-20% w / w solution o f xylylene bis(tetrahydrothiophenium chloride) in dry ethanol at 50 ° C for 24 h , and then washed wi th ethanol and water to remove excess monomer and base. The bright yel low-green powder was dried under v a c u u m at room temperature; subsequent heating to 2 0 0 ° C under v a c u u m (10" 2 Torr ) for 6 h resulted i n the powder turning bright ye l low in color . 78 For T E M analysis, samples were deposited on lacey earbon-coated C u grids (Ted Pel la , Inc.) from a suspension in methanol . E lec t ron energy-loss spectroscopy and energy-filtered t ransmission electron mic roscopy were carried out on a Tecnai F 2 0 T E M equipped wi th a Gatan Imaging Filter. The accelerating voltage was 197 k V (200 k V nomina l ly , offset by 3 k V by the G I F ) . Loss spectra were recorded in T E M mode by p lac ing the particle o f interest above the G I F entrance aperture (diameter 2.0 m m ) . T h e zero-loss peak was recorded separately for subtraction by m o v i n g to an empty area on the gr id . The system energy resolution, g iven by the F W H M o f the zero-loss peak, was 0.9 e V . The energy dispersion o f the spectrometer was 0.10 e V / p i x e l , w h i c h was calibrated us ing a 50 e V wobble on the drift tube. E F T E M images were acquired w i t h a 2 e V slit. The U V / V i s absorbance spectrum was measured for P P V / M C M - 4 1 as a nujol m u l l , us ing a U n i c a m U V - 2 spectrometer. Low-temperature P L spectra were obtained by pressing the samples between quartz plates in a cryostat and us ing the frequency-doubled output o f a Ti :Sapphi re laser (386 nm) to excite the sample. The luminescence was passed through a Czerny-Turner scanning monochromator ( D i g i k r o m 242 from C V I Laser Corpora t ion , us ing only one 600 grooves mm" 1 grating) and collected w i t h a Hamamatsu R 2 2 5 7 photomult ipl ier tube. The spectra were corrected for the response o f the grating and the photomult ipl ier tube by measur ing the spectrum o f a quartz tungsten halogen lamp (Or i e l 77501 I l luminator) o f k n o w n output. 79 References 1. van B e k k u m , H . ; Flanigen, E . M . ; Jacobs, P . A . ; Jansen, J . C , E d s . Introduction to Zeolite Science and Practice;E\sevier: A m s t e r d a m , 2001 2. M c C u s k e r , L . B . Structural Aspects of Molecular Sieves; Comprehens ive Supramolecular Chemis t ry ; Elsevier Science: N e w Y o r k , 1996; V o l . 7, p. 393 . 3. Kess ler , H . Synthesis of Molecular Sieves; Comprehens ive Supramolecular Chemis t ry ; E l sev ie r Science: N e w Y o r k , 1996; V o l . 7, p . 425 . 4. F lanigen , E . M . Zeolites and Molecular Sieves: A Historical Perspective; Studies in surface science and catalysis; E lsevier : A m s t e r d a m , 2 0 0 1 ; V o l . 137, p. 11. 5. F r i s ch , H . L . ; X u e , Y . P . ; Maaref , S.; Beaucage, G . ; P u , Z . C ; M a r k , J . E . Macromolecular Symposia 1 9 9 6 , 1 0 6 , 147. 6. L a i , Z . P . ; B o n i l l a , G . ; D i a z , I.; N e r y , J . G . ; Sujaoti , K . ; A m a t , M . A . ; K o k k o l i , E . ; Terasaki , O . ; Thompson , R . W . ; Tsapatsis, M . ; V l achos , D . G . Science 2003, 300, 456. 7. Kresge , C . T . ; L e o n o w i c z , M . E . ; Ro th , W . J . ; V a r t u l i , J . ; B e c k , J . S. Nature 1992, 359, 710. 8. B e c k , J . S.; V a r t u l i , J . C ; Ro th , W . J . ; L e o n o w i c z , M . E . ; Kresge , C . T ; Schmit t , K . D . ; W u , C . T . - W . ; O l s o n , D . H . ; Sheppard, E . W . ; M c C u l l e n , S. B . ; H i g g i n s , J . B . ; Schlenker , J . L . J . A m . Chem. Soc. 1 9 9 2 , 1 1 4 , 10834. 9. L inssen , T ; Cassiers, K . ; C o o l , P . ; Vansant , E . F . Adv. Colloid Interface Sci. 2 0 0 3 , 1 0 3 , 121. 10. B e c k , J . S.; V a r t u l i , J . C . Curr. Opin. Solid State Mater. Sci. 1996,1, 76. 11. Ande r son , M . T ; M a r t i n , J . E . ; Od inek , J . G . ; N e w c o m e r , P . P . Chem. Mater. 1998, 10, 1490. 12. Edler , K . J . ; Whi t e , J . W . Chem. Mater. 1997, 9, 1226. 13. Cheng , C . F . ; Park, D . H . ; K l i n o w s k i , J . J. Chem. Soc, Faraday Trans. 1997, 93, 193. 80 14. C h e n g , C . F . ; Z h o u , W . Z . ; Park, D . H . ; K l i n o w s k i , J . ; Hargreaves, M . ; Gladden , L . F . J . Chem. Soc., Faraday Trans. 1997, 93, 359 . 15. Zhao , D . ; Feng , J . ; H u o , Q . ; M e l o s h , W . ; Frer ickson , G . H . ; C h m e l k a , B . F . ; D t u c k y , G . D . Science 1998. 16. Schuth, F . Curr. Opin. Colloid Interface Sci. 1998, 3, 174. 17. Zhang , Z . R . ; Suo, J . S.; Zhang , X . M ; L i , S. B . Acta Phys.-Chim. Sinica 1999, 15, 333. 18. K a w i , S.; L a i , M . W . Chem. Commun. 1998, 1407. 19. Landmesser , H . ; K o s s l i c k , H . ; Storek, W . ; F r i cke , R . Solid State Ionics 1 9 9 7 , 1 0 1 , 2 7 1 . 20 . Jaroniec, C . P . ; K r u k , M . ; Jaroniec, M . ; Sayari , A . J. Phys. Chem. B 1 9 9 8 , 1 0 2 , 5503. 2 1 . Sayari , A . ; L i u , P . Microporous Mater. 1997,12 , 149. 22 . O z i n , G . A . ; C h o m s k i , E . ; Khusha l an i , D . ; Mac lach lan , M . J . Curr. Opin. Colloid Interface Sci. 1998, 3, 181. 23 . M o l l e r , K . ; B e i n , T . Chem. Mater. 1998 ,10 , 2950. 24. Edle r , K . J . ; Reynolds , P . A . ; Whi t e , J . W . J. Phys. Chem. B 1 9 9 8 , 1 0 2 , 3676. 25 . Ramsay , J . D . F . Adv. Colloid Interface Sci. 1998, 77, 13. 26. Mar l e r , B . ; Oberhagemann, U . ; V o r t m a n n , S.; G ie s , H . Microporous Mater. 1996, 6, 375 . 27 . Ramsay , J . D . F . ; H o i n k i s , E . Physica B 1998, 248, 322. 28. S ing , K . S. W . Adv. Colloid Interface Sci. 1998, 77, 3. 29 . Brunauer , S.; Emmet t , P . H . ; Tel ler , E . J. Am. Chem. Soc. 1938, 60, 309. 30 . Barrett, E . P . ; Joyner, L . G . ; Halenda , P . P . J. Am. Chem. Soc. 1951, 73, 373 . 3 1 . R a v i k o v i t c h , P . I.; Hal ler , G . L . ; N e i m a r k , A . V . Adv. Colloid Interface Sci. 1998, 77, 203 . 32 . C h o m a , J . ; Jaroniec, M . ; B u r a k i e w i c z - M o r t k a , W . ; K l o s k e , M . Appl. Surf. Sci. 2002 , 196, 216. 33. Ha rk in s , W . D . ; Jura, G . J. Am. Chem. Soc. 1944, 66, 1362. 81 34. K r u k , M . ; Jaroniec, M . ; Sayar i , A . Langmuir 1997 ,1 3 , 6267. 35 . A l b o u y , P . A . ; A y r a l , A . Chem. Mater. 2 0 0 2 , 1 4 , 3391 . 36. Schacht, S.; Janicke, M . ; Schuth, F . Microporous Mesoporous Mater. 1998, 22, 485 . 37. W u , C . - G . ; B e i n , T . Science 1994, 264, 1757. 38. W u , C . - G . ; B e i n , T. Science 1994, 266, 1013. 39. L l e w e l l y n , P . L . ; C ies la , U . ; Decher , H . ; Stadler, R . ; Schuth, F . ; Unger , K . K . Zeolites and related microporous materials: State of the art 1994;; E lsevier ; no. N u m b e r ; p. 2013 . 40. H s i e h , B . R . Poly(p-phenylene vinylene)s: Methods ofpreparation and properties; P o l y m e r i c Mater ia ls Encyc loped ia ; C R C Press: B o c a Raton , F l , 1996; V o l . 9, p. 6537 . 4 1 . G i l c h , H . G . ; Wheelwr igh t , W . L . J. Polym. Sci., Part A: Polym. Chem. 1966, 4, 1337. 42 . K u m a r , D . N . ; Bhawa lka r , J . D . ; Prasad, P . N . Appl. Opt. 1998, 37, 510. 43 . Khusha l an i , D . ; K u p e r m a n , A ; C o o m b s , N . ; O z i n , G . A . Chem. Mater. 1996, 8, 2188 . 44 . Papadimit rakopoulos , F . ; Kons tad in id i s , K . ; M i l l e r , T . M . ; O p i l a , R . ; Chandross , E . A . ; G a l v i n , M . E . Chem. Mater. 1994, 6, 1563. 45 . Gagnon , D . R . ; Capis t ran, J . D . ; Karasz , F . E . ; L e n z , R . W . ; A n t o u n , S. Polymer 1987, 28, 567. 46 . O r i a k h i , C . O . ; Zhang , X . R . ; Lerner , M . M . Applied Clay Science 1999, 75, 109. 47 . Mar tens , J . H . F . ; Bradley , D . D . C ; B u r n , P . L . ; F r i end , R . H . ; H o l m e s , A . B . ; Marseg l i a , E . A. Synth. Met. 1991 ,47,301. 48. T u n , Z . ; M a s o n , P . C . Acta Crystallogr., Sect. A: Found. Crystallogr. 2000, 56, 536. 49 . Samuel , I. D . W . ; Crys ta l l , B . ; Rumbles , G . ; B u r n , P . L . ; Ho lmes , A . B . ; F r iend , R . H . Synth. Met. 1993, 54, 2 8 1 . 50. Y i n , J . S.; W a n g , Z . L . Appl. Phys. Lett. 1999, 74, 2629. 82 51. Eger ton, R . F . Electron energy-loss spectroscopy in the electron microscope; 2nd ed.; P l e n u m Press: N e w Y o r k , 1996. 52. Barfels , M . M . G . ; J iang, X . G . ; H e n g , Y . M . ; Arsenaul t , A . L . ; Ottensmeyer, F . P . Micron 1998, 29, 97. 53 . P ich le r , K . ; Ha l l iday , D . A . ; Brad ley , D . D . C ; B u r n , P . L . ; F r iend , R . H . ; H o l m e s , A . B . J . Phys.: Condens. Matter 1993, 5, 7155. 54. G i n , D . L . ; Y o n e z a w a , K . Synth. Met. 2 0 0 1 , 1 2 1 , 1291. 55. Tolbert , S. H . ; W u , J . J . ; Gross , A . F . ; N g u y e n , T . Q . ; Schwar tz , B . J . Microporous Mesoporous Mater. 2001, 44, 445. 83 CHAPTER 4 Preparation of Mesoporous Thin Film Host M i c r o p o r o u s and mesoporous materials often have anisotropic pore structures. A good example is the mesoporous material M C M - 4 1 , discussed in the previous chapter: it has straight 1-D channels arranged i n a 2 - D hexagonal lattice. Th i s anisotropy can only be exploited, however , i f the macroscopic orientation o f the material can be control led either dur ing or after synthesis. Such control over the orientation is necessary for a number o f applications, i nc lud ing photonic materials, membranes and sensors. It is also key to fabricating the ideal device where only intrachain processes can occur: proper channel orientation w i l l a l low good charge transport a long the conjugated po lymer chains i n the device, and a narrow pore diameter w i l l favour the isolation o f the po lymer chains (Figure 1.5; Table 1.2). The development o f a porous thin f i l m host w i t h oriented 1-D channels was thus o f central importance to this research effort. Notwi ths tanding a l l the requirements expressed in Table 1.2, the central challenge is the fabrication o f porous th in films w i t h 1-D channels a l igned i n the d i rec t ion o f the substrate normal . The synthesis o f films and membranes w i t h such oriented porous structures has been the subject o f continuous scientific pursu i t . 1 " 5 Impressive progress has been made i n many areas but the goal o f fabricating un i fo rm defect-free membranes w i t h very sma l l pores remains elusive. Membranes w i t h pores d o w n to 10 n m diameter are readi ly available, but the size regime be low that is not w e l l developed. A s the membranes o f interest for this w o r k lie i n this uncharted territory, it was clear at the outset that substantial effort w o u l d be required to develop the desired th in f i lm host. However , a number o f p romis ing first reports were present i n the literature, w h i c h 84 suggested that the goal was w i t h i n reach. These reports are rev iewed here and the investigations into the more p romis ing routes discussed. M a n y o f the characterization techniques applicable to analysing bu lk mesoporous samples can also be appl ied to th in f i lms. X - r a y diffraction can be used to evaluate the orientation o f the film.6 Physisorpt ion measurements are not straightforward, due to the l imi ted total surface area o f the films. The most important techniques are based on electron microscopy: scanning electron mic roscopy ( S E M ) and transmission electron mic roscopy ( T E M ) a l low h igh resolution inspection o f the film structure on the loca l level , as rev iewed i n chapter 2. 4.1 A l i g n m e n t o f P o r o u s T h i n F i l m s Self-assembly is an attractive route for materials synthesis because o f the inherent order o f the result ing materials. O n c e the appropriate condit ions for self-assembly are k n o w n , the synthesis is usual ly straightforward. Thus a lot o f effort has been devoted to the pursuit o f systems w h i c h self-assemble into porous thin films and membranes w i t h various geometries. For self-assembled mesoporous materials, the methods used to achieve alignment fall into two general categories: interface-induced alignment and field-induced alignment ( inc lud ing electric, magnetic and f low fields). These various methods are rev iewed here, and it w i l l become obvious that no ideal approach to the fabrication o f the desired membrane v i a self-assembly exists as o f yet. Some very recent w o r k on solvent evaporat ion-induced alignment does appear very p romis ing i n creating mesoporous thin films wi th channels normal to the substrate. 7 85 4.1.1 In t e r f ace - Induced A l i g n m e n t Surfactant-templated materials, i nc lud ing the members o f the M 4 1 S fami ly ( M C M - 4 1 , M C M - 4 8 ) , have been found to al ign themselves i n specific ways on a number o f interfaces. It is thought that the arrangement o f the ini t ia l layer o f surfactant micel les on the interface largely determines the al ignment o f the g r o w i n g film. M i c a and graphite induce a paral lel al ignment o f cy l ind r i ca l micel les (Figure 4.1) through fairly weak electrostatic and van der Waa l s interactions r e spec t ive ly . 8 ' 9 The result ing films st i l l have substantial in-plane disorder (direction o f channels, 180° defects), w h i c h l imi ts their usefulness. The air-water interface has also been shown to favour a s imi lar type o f o r d e r i n g . 1 0 F i l m s g r o w n on S i (100) and (111) show no preferential ordering, but S i (110) does induce al ignment paral lel to the [001] d i r ec t ion , 6 indicat ing that proper registry o f the templating mesophases w i t h the interface is necessary to achieve ordering. However , a straightforward approach to predic t ing w h i c h surfaces w i l l produce proper registry has not been found yet. It is also possible to modi fy a surface to induce al ignment i n a preferred direct ion, as shown by the rubb ing m e t h o d , 1 1 w h i c h is c o m m o n l y used to a l ign l i qu id crystalline materials for d isplay applications. A thin po ly imide film is coated on the substrate and then rubbed i n a particular direct ion, w h i c h aligns the channels o f the mesoporous material parallel to the rubb ing F i g u r e 4.1 2 - D hexagonal pack ing o f surfactant micel les i n aqueous solution onto a graphite surface, after ref. 8. 86 direct ion. In-plane X - r a y diffraction shows that the distr ibution o f channel directions has a F W H M o f 2 9 ° . The orientation o f the in i t ia l template layer on the surface can also be controlled by mod i fy ing the chemica l structure o f the templat ing agent (Figure 4.2). Th i s has been shown through the use o f a two-headed quarternary a m m o n i u m salt to form a structure designated S B A -2 . 1 2 The ini t ia l reports on the structure o f S B A - 2 f i lms , based on T E M studies, suggested that there was a continuous channel no rma l to the interface. A later study us ing high-resolut ion T E M showed that the in i t ia l structure assignment was erroneous and that there was a higher degree o f connect ivi ty i n the porous structure than or ig inal ly i d e n t i f i e d . 1 3 It is therefore st i l l difficult to self-assemble a f i lm w i t h oriented channels normal to the interface. O n e plausible approach around this p rob lem is the use o f a substrate that presents vert ically-oriented interfaces, i.e., by us ing a larger porous support to nucleate the film. Porous a l u m i n a 1 4 " 1 7 (see be low) and c a p i l l a r i e s 1 8 have been used as substrates for the g rowth o f zeolites and mesoporous silicates. These oriented supports present interfaces oriented paral lel to the substrate normal (Figure 4.3). Th i s has been recently used to prepare membranes o f M C M - 4 8 , A B F i g u r e 4.2 Surfactant templates for (a) M C M - 4 1 and M C M - 4 8 structures ( C i 6 H 3 3 N ( C H 3 ) 3 X ) and (b) S B A - 2 structure ( C i e H a s N C C H s M C ^ N C C F f e ^ ) , where X = B r or C l . 87 w h i c h has an isotropic cubic structure, and therefore cannot show a preferred orientation. A supported M C M - 4 1 structure has not been reported to date, perhaps because the desired orientation is not achieved. Nevertheless, one w o u l d expect that the surface curvature o f the porous support (i.e., the inverse o f the channel radius) w o u l d have a significant effect i n favouring alignment o f the channels i n the axia l direct ion over the circumferential direct ion. A B F i g u r e 4.3 The two extreme possibi l i t ies for surfactant micel le alignment i n a porous support: (a) i n the axia l direct ion and (b) i n the circumferential direct ion. A smaller radius o f curvature R w o u l d be expected to favour the axia l orientation. 4.1.2 F i e l d - I n d u c e d A l i g n m e n t External ly appl ied fields offer an attractive approach to obtaining the desired channel al ignment i n porous materials. In many cases, the product material is only partially al igned wi th the appl ied f ield, as only an average alignment is achieved, w i th substantial deviations presumably due to thermal disorder o f the templates. It has been shown that the appl icat ion o f a large magnetic f ield ( -12 T) can be used to al ign the channels i n a mesoporous m a t e r i a l . 1 9 Th i s results from anisotropy i n the diamagnetic 88 susceptibili t ies o f the templating agents. Once the template micel les reach equ i l ib r ium w i t h the magnetic f ield (wi th 6 0 % o f the domains al igned wi th the magnetic f ield, as seen by solid-state N M R ) , condensation o f the s i l ica f ramework is achieved by a l l o w i n g gaseous HC1 to diffuse in . The resulting material was characterized b y X - r a y diffraction, w h i c h suggested that 7 8 % o f the pores were al igned i n the desired d i r e c t i o n . 2 0 Hil lhouse et al. have used a f low f ie ld to induce preferential al ignment o f a mesoporous s i l i ca f i lm g r o w n i n a capil lary t u b e . 2 1 The channels were found to be oriented i n the direct ion o f the external f low for thin films (< 200 nm) , based on observation o f the macroscopic morphology o f the film b y S E M . Th icke r films tended to lose this preferred alignment and became s imi lar to films g r o w n under static condit ions. M o r e w o r k i n this direct ion was carr ied out by K i m and Y a n g . 2 2 The flow field p rov ided by laser-induced ablation has also been used to generate al igned s t r u c t u r e s . 2 3 ' 2 4 The laser was used to create a seed film o f material oriented by the ablation o f a guest molecule , ferrocene. Further treatment w i t h the synthesis mixture then formed a continuous film wi th an orientation der ived from the seed film. The use o f both an interface and an appl ied electric field to fully control al ignment has also been demons t ra ted . 2 5 A n elastomeric m o l d was used to create micron-scale channels on a substrate between two electrodes (Figure 4.4). The electric field creates an electro-osmotic flow between the two electrodes, w h i c h is further gu ided by the channels. Thus both the flow and the presence o f the interfaces contribute to a l igning the template micel les , leading to a very h igh degree o f al ignment parallel to the flow direct ion. L o c a l i z e d Joule heating then induces condensation o f the s i l ica framework. W h i l e this approach created h igh quali ty al ignment o f the channels ( X - r a y diffraction indicated that the direct ional spread was 1.7°, substantially better than other methods for a l igning the micel les) , it w o u l d not be straightforward to fabricate thin 89 F i g u r e 4.4 Orientat ion o f mesoporous channels by electro-osmotic f low, after ref. 25 . f i lms (i.e., w i t h channel lengths from 100 to 200 nm) us ing this technique. One poss ibi l i ty may be to combine it w i th a porous support o f the type discussed above. 4.13 O r i e n t e d P o r o u s T h i n F i l m s by O t h e r A p p r o a c h e s Track-etch m e m b r a n e s 2 6 are prepared by exposing a polycarbonate or polyester f i l m to a beam o f energetic heavy ions i n a linear accelerator. The ions leave damage tracks that are roughly normal to the f i lm surface, and w h i c h are then etched chemica l ly to the desired diameter. The obtainable diameters range from 10 n m to several microns . These are available commerc ia l ly from Wha tman , Inc. under the Nucleopore trademark. W h i l e the technology behind these membranes is fairly w e l l developed, it is unclear i f they can be prepared i n the form o f thin f i lms, or obtained wi th smaller pore diameters. 90 K o n d o h e t al. reported an almost ideal porous f i l m structure, attained by o x i d i z i n g FeO:Si02 thin films w i t h a specific s toichiometry at 600 ° C . 2 7 The al ignment was thought to be induced by the direct ion o f oxygen flow, as the ox id i zed Fe precipitated out (as Fe2C>3) in co lumns perpendicular to the film surface. The Fe2C>3 co lumns cou ld be convenient ly etched out, leaving beh ind a porous s i l i ca matr ix . The result ing film had an average pore size o f 4 n m , as determined by nitrogen physisorpt ion and supported by T E M studies. Porous a lumina (anodic a l u m i n u m oxide) is also commerc ia l ly important, especially for m i c r o f i l t r a t i o n . 2 8 It is produced by anodiz ing a l u m i n u m samples i n aqueous electrolytes. U n d e r the appropriate condit ions, the result ing oxide has ver t ical pores running through it, and the pore spacing can be controlled by the appl ied potential. The exact cause o f the al ignment o f the channels is s t i l l unclear but appears to be related to both the appl ied potential and the strain created b y the expansion o f the lattice from a l u m i n u m to a l u m i n u m oxide. The al ignment o f the channels is almost perfect after extended growth and publ i shed reports indicated that the diameter m a y be adjusted f rom 20 to 500 n m . T h i c k membranes are commerc ia l ly available from Wha tman , Inc. under the A n o p o r e trademark. It m a y be possible to prepare s imi lar structures by electron-beam li thography and chemica l etching o f an appropriate substrate. A t the early stages o f this thesis, the feature size for this technique was on the order o f 50 n m , and thus too large for the purpose o f this project. Di rec t electron-beam d r i l l i ng o f a lumina or s i l i ca films us ing a field-emission T E M i n scanning mode is a very p romis ing approach for creating a smal l number o f very smal l diameter h o l e s . 2 9 Porous thin films based on the self-assembled S B A - 2 structure, the FeO/SiC»2 system and porous a lumina were in i t ia l ly j udged to be p romis ing thin film hosts. The i r synthesis and characterization was undertaken to investigate further their suitabil i ty as oriented thin film hosts for conjugated polymers . 91 4.2 Further Investigation of SBA-2 Mesoporous Silica Films Self-assembled films are attractive candidates for study as they are often h igh ly ordered and un i fo rm. A s s u m i n g that the chemistry to y i e l d the correct geometry can be found, these films w o u l d be ideal host materials for conjugated polymers . A s stated above (section 4.1), the geometry o f a film nucleated on an interface is thought to depend on the arrangement o f the very first layer o f templating agent on the interface. A s such, control o f the template-interface interactions should i n pr inc ip le a l low the preparation o f films w i t h the desired pore orientation. M o s t o f the w o r k reported i n the literature has focused on th in films related to the M C M - 4 1 structure and based on straight-chain a lky l a m m o n i u m surfactants. These adopt the w e l l -established parallel al ignment w i t h the interface i n a hexagonal structure. A different orientation can only be expected from films w i t h substantially different structures. The structure type is control led by the choice o f surfactant and p H . A smal l number o f geometr ical parameters have been found to govern the type o f structure produced b y a g iven surfactant. The hydrophobic tail length, hydrophi l ic head size and charge can a l l be adjusted to y i e l d fair ly predictable s t ruc tures . 3 0 Tolbert et a l . found that the use o f a two-headed surfactant (Figure 4.2(b)), w h i c h templates the 3 - D hexagonal S B A - 2 structure, a l lowed the synthesis o f thin films w i t h a possible channel no rma l to the interface. In this case the in i t ia l layer on the surface was found to consist o f packed hemispheres, representing the higher symmetry o f S B A -2. The proposed structure consisted o f large cages, the connect ivi ty o f w h i c h w o u l d determine the form o f the channels through the material . 92 A s such, the material appeared p romis ing and was investigated i n more depth. The reported S B A - 2 structure was synthesized on freshly-cleaved m i c a substrates. The presence o f the S B A - 2 phase was conf i rmed by X - r a y diffraction (Figure 4.5). T h e (002) peak appears strongly, indicat ing that the film is al igned w i t h the c axis perpendicular to the film. The (112) and (004) peaks appear weakly . The calc ined sample showed a shift i n the (002) peak to higher angle, indicat ing that the framework size had decreased dur ing calcinat ion. In order to investigate the microstructure o f S B A - 2 films, the samples g r o w n on m i c a were embedded i n epoxy resin and sectioned us ing a d i amond knife . However , the film d i d not survive the sect ioning process and thus no further information on the pore structure was obtained co C L O o 7 6 5 4 £ 3 CO c CD 1 2 1 0 o o C M o o Peak 29 (002) 2.29 (as made) (002) 2.59 (calcined) (112) 4.6 (004) 5.1 1X1 2 £ 8 x10 8 29 (°) F i g u r e 4.5 X - r a y diffraction pattern o f S B A - 2 film g rown on m i c a (—) before and (—) after calcinat ion. 93 i n this fashion. A report appearing at this t ime indicated that S B A - 2 d i d not have the desired connectivi ty, based on h igh resolution T E M i m a g i n g . 1 3 Th i s study o f bu lk S B A - 2 indicated that the structure g r o w n as a f d m w o u l d have straight 1-D channels paral lel to the interface and z igzag channels running perpendicular ly to the interface, a geometry not conducive to conjugated po lymer inc lus ion . In v i e w o f this information, and p r o m i s i n g developments w i t h porous a lumina f i lms, no further characterization was attempted. 4.3 F u r t h e r Inves t iga t ion o f the F e O / S i C h S y s t e m The FeO/SiC»2 system is the subject o f one report i n the l i te ra ture . 2 7 The inspirat ion for this approach reportedly comes from the observation o f elongated Fe203 structures i n natural samples. In this material, the ordering is determined by the direct ion o f oxygen diffusion as F e O is o x i d i z e d to Fe203 w i t h i n the SiC>2 matr ix . The morphology o f the resultant materials depends strongly on the relative concentration o f F e O and Si02. A t composi t ions between 70:30 and 60:40 FeO:Si02, it is reported that co lumns o f Fe203 g row perpendicular ly to the substrate w i t h i n the Si02 matr ix (Figure 4.6). These columns can then be etched out by aqueous HC1 , leaving behind the Si02 matr ix . The result ing 1-D channels i n Si02 have a diameter o f 4 n m , as determined by a Barret t-Joyner-Halenda ( B J H ) analysis o f the nitrogen physisorpt ion isotherm. W h e n the FeO:Si02 ratio is lower than 60:40, the result ing Fe203 particles are trapped i n the Si02 matr ix and cannot be etched out. H igher F e O levels lead to an ox id ized f i lm w i t h larger, disordered Fe203 structures. U s e o f the correct FeO:Si02 ratio should produce a th in f i l m host w i t h nearly ideal characteristics: porous Si02 f i lms w i t h ver t ical 4 n m channels w i t h easily control led thickness. The slight drawback o f this process is the rather h igh processing temperature ( 6 0 0 ° C ) . 94 O2 diffusion F i g u r e 4.6 Preparation o f th in film w i t h oriented channels from a FeO:Si02 f i l m . The dupl ica t ion o f these publ i shed results was attempted. U s i n g a 65:35 FeO:Si02 target, thin films were sputtered unto glass and s i l i con substrates us ing radio-frequency ( R F ) sputtering. Freshly-sputtered films were ini t ia l ly transparent w i t h a deep green colour. F o l l o w i n g oxidat ion at 600 ° C , the films turned red i n appearance, indicat ing the formation o f Fe203. Th i s colour disappeared completely upon etching w i t h 1:1 H C h H b O , suggesting that most o f the Fe203 had been removed from the thin film. These observations were i n agreement w i th the chemistry suggested by the in i t ia l reference. The microstructure o f the films at different stages o f processing was investigated by T E M . Cross-sections o f the thin films were readily obtainable by the small-angle cleavage ( S A C ) technique. 95 The cross-section o f the ox id ized , unetched film is shown i n Figure 4.7 (a). V e r y little structure is apparent, i n contrast w i th the publ ished micrographs, where columnar structures were v is ib le . The energy-dispersive X - r a y ( E D X ) analysis o f the thin film w i t h a S T E M (Figure 4.8(a)) y ie lded a Fe :S i atomic ratio o f 2.1 ± 0.2, w h i c h was wi th in the range deemed opt imal (1.9 - 2.3). X P S analysis, on the other hand, y ie lded a Fe :S i ratio o f 0.9, suggesting the surface F e O content was w e l l be low the desired level . The etched film is shown i n Figure 4.7 (b), and there is insufficient contrast to observe the pore structure. The E D X spectrum showed that the Fe203 had been fully removed (Figure 4.8(b)), indicat ing that there were no isolated Fe203 particles. Th i s suggested that the F e : S i ratio was near or above opt imal . The X P S results, w h i c h reflect the compos i t ion at the surface, may indicate that the film had a lower concentration o f Fe near the surface. Heavy-element staining o f the pore surface was attempted to improve the contrast between glass substrate F i g u r e 4.7 T E M image o f cross-section o f F e O : S i 0 2 f i lm on glass (a) after oxidat ion, before etching, (b) after etching and Pb-staining. The cross-section was prepared by the S A C technique. The scale bars are 50 n m long. 96 the pores and the surrounding matr ix . L e a d acetate is k n o w n to b i n d to surface hydroxy groups, but it d i d not improve the contrast o f the pores, despite clear indicat ion o f the presence o f P b by the E D X spectrum (Figure 4.8(c)). The nitrogen adsorption isotherm on a 200-nm thick sample (anticipated total surface area ~0.5 m 2 ) d i d not show a clear condensation step, suggesting a w i d e dis t r ibut ion o f pore sizes. The measurements were also l imi ted i n accuracy by instrumental drift. Th i s stood i n contrast w i t h the reported adsorption isotherm, w h i c h showed a clear capil lary condensation step corresponding to a 4 n m B J H diameter. W h i l e the th in films were porous, as evidenced b y the complete r emova l o f the Fe203 by etching, no wel l -def ined channels cou ld be detected by electron mic roscopy or nitrogen physisorpt ion. At tempts to contact the authors to discuss possible processing issues were unsuccessful. Further efforts to reproduce this w o r k were abandoned since porous a lumina showed more promise. 97 0.00 2.56 5.12 7.68 10.24 X - R a y Energy (keV) Figure 4.8 E D X spectrum obtained from cross-sections of F e O : S i 0 2 films: (a) oxidized ( Au contamination is apparent), (b) etched to remove Fe203, (c) Pb-stained. 4.4 Porous Alumina Films (Anodic Aluminum Oxide) A n o d i c films on a l u m i n u m have a long history o f industrial applicat ion as cosmetic and protective layers. The choice o f anodizat ion condit ions (electrolyte, temperature, potential) has a dramatic effect on the structure o f the resulting a lumina film. The porous films obtained under the appropriate condit ions were found to exhibi t a structure w h i c h fitted the requirements expressed i n Table 1.2 very closely. The porous nature o f some o f these films has been rev iewed i n depth by T h o m p s o n and W o o d . 3 1 D u e to their industrial importance, these films have been extensively studied since the 1930's. M a n y characteristics o f the films were w e l l understood by the 1980's, and it was recognized that very ordered films cou ld be produced under certain condit ions. The self-ordering phenomenon was not fully exploi ted unt i l the 1990's: i n 1993, M a s u d a e t al. reported a s imple technique for the product ion o f porous anodic a lumina wi th h igh ly ordered pores and a r m ~ m dt 41 < A & m**m ™ a t ' ^  • • • ^Km. JMIK -^ €P A ^ & ^ JI ^ A ^  • i n * g| w — Figure 4.9 S E M images o f porous a lumina film produced at 40 V i n 0.3 M oxal ic acid , u s ing Masuda 's two-step approach: (a) top v i e w , (b) cross-section. The scale bars are 100 n m long. Images courtesy o f D r . K . Rademacher . 99 corresponding narrow pore size distr ibution (Figure 4 . 9 ) . 3 2 T h i s development led to the widespread applicat ion o f porous a lumina membranes both as hosts for other m a t e r i a l s 3 3 " 3 6 and as t empla te s 3 7 for the synthesis o f t u b u l e s 3 8 and w i r e s 3 9 ' 4 0 w i th un i fo rm diameters i n the 30 to 100 n m regime. The development o f wel l -ordered large pore samples a l lowed the fabrication o f photonic crystals either directly us ing patterned porous a lumina f i l m s 4 1 ' 4 2 or by transferring the pattern to another substrate by dry chemica l e t c h i n g . 4 3 " 4 6 The anodizat ion o f other materials, most notably s i l i con , also produces very useful porous s t ructures . 4 7 Porous a lumina is prepared by anodiz ing a l u m i n u m in an ac idic electrolyte, typical ly phosphor ic , sulfuric or oxa l ic ac id . T h e anode is attached to the a l u m i n u m substrate, w h i l e the cathode usually consists o f a p la t inum mesh electrode (Figure 4.10). Unde r the appropriate condit ions, the pores are straight and hexagonally p a c k e d , 4 8 - 4 9 as shown i n Figure 4.9. The structural parameters are shown i n Figure 4 .11 : the pore diameter is D and the lattice spacing is Glass ce l l w i t h O - r i n g Figure 4.10 A s imple electrochemical ce l l for anodizat ion o f a l u m i n u m substrates. 100 L. The oxide growth process leaves a barrier layer at the bot tom o f the pores w i th a thickness on the order o f L/2. Anion-contamina ted layer Bar r ie r layer B Figure 4.11 Structure o f porous a lumina f i lms g r o w n on a l u m i n u m : (a) geometry o f pore pack ing , (b) cross-section showing barrier layer o f thickness ~L/2 at bot tom o f pores. 4.4.1 Pore Wall Structure The nature o f the pore wal l s has been studied extensively. The results o f many early studies have been rev iewed by T h o m s o n and W o o d . 3 1 The pore wal ls consist o f a relatively thick, amorphous, anion-contaminated surface layer over a more dense core o f pure a lumina . The two different zones o f a lumina can be readily dis t inguished i n T E M micrographs o f the m a t e r i a l . 3 1 ' 4 9 ' 5 0 There exists considerable variat ion i n anion incorporat ion among the different f i lms prepared us ing sulfuric, oxal ic , and phosphoric a c i d s . 5 1 U s i n g T E M and E D X , 5 0 the anion distr ibution has been shown to be un i fo rm through the less dense layer. After washing , the dis tr ibut ion o f anions i n the ce l l w a l l is lowest i n the pure a lumina core, reaches a m a x i m u m w i t h i n the less dense layer and falls o f f near the pore surface. Th i s dis tr ibut ion can be the result o f an ini t ia l ly un i fo rm anion contamination, w h i c h is altered dur ing the post-synthesis water 101 rinse by preferential removal o f anions near the pore s u r f a c e . 5 2 ' 5 3 The nature o f the anion contaminat ion may be very important for electric device applications o f porous a lumina, as the anions m a y become mobi le under h igh electric fields, causing electrical short-circuits or affecting the electrical properties o f the po lymer host. These anions m a y be removed completely by thermal treatment o f the free-standing m e m b r a n e . 5 4 " 5 9 Membranes prepared us ing oxal ic ac id as the electrolyte show the removal o f oxalate ion over 2 0 0 - 4 0 0 ° C , w i t h crystal l izat ion o f the pore wal l s occurr ing at 8 2 0 - 8 4 0 ° C . W h e n sulfuric ac id is used as electrolyte, the contaminat ing anions are removed at 9 7 0 ° C . 4.4.2 Pore Growth Processes A number o f empi r i ca l relationships have been established for porous a lumina growth: (1) the spacing o f the pores is proport ional to the appl ied potential (2.5 to 2.8 n m V " 1 ) 4 9 ; (2) the pores form an ordered lattice at certain potentials for each electrolyte; (3) the pore depth ( f i lm thickness) is proport ional to the total charge passed (proport ional to t ime under galvanostatic anodization). W h i l e these empi r ica l relationships have been thoroughly investigated, the processes w h i c h govern the growth o f the pores are not fully understood. A qualitative analysis has been presented by T h o m p s o n and W o o d . 3 1 In acidic electrolytes, pore formation must be the result o f the concurrent oxidat ion o f the a l u m i n u m to a lumina and dissolut ion o f the a lumina by the electrolyte, along w i t h a higher a lumina dissolut ion rate i n depressions. Th i s suggests that field-assisted dissolut ion is a key d r i v i n g force beh ind pore growth. A loca l ized temperature increase at the metal-oxide interface due to Joule heating may also p lay a role i n the process. The larger effect o f electrolyte temperature on the film parameters indicates that local temperature variations do not have a domina t ing role i n the process. 102 The pore structure m a y be further altered by etching w i t h phosphoric acid , w h i c h attacks the a lumina m u c h faster than sulfuric ac id or oxal ic acid . The anion-contaminated layer is also attacked faster than the the more dense cel l boundary. Typ ica l ly , the pore wal ls are etched i n 5 wt. % phosphor ic ac id at a rate o f 8 n m h" 1 at r o o m tempera ture . 6 0 4.4.3 P o r e L a t t i c e F o r m a t i o n For a g iven electrolyte, self-ordering may be obtained over a range o f appl ied potentials. Publ i shed values are shown i n Table 4 .1 . U n d e r these condit ions, the pore lattice becomes ordered w i t h a domain size between 1 and 4 p m . B e l o w this range o f potentials, there is considerable pore branching dur ing the growth process. Thus not a l l pore sizes are readily accessible i f an ordered pore lattice is desired. The method for obtaining fully ordered structures consists o f a two-step anodization p r o c e s s . 3 2 ' 4 8 - 6 1 A n in i t ia l long anodizat ion per iod is carr ied out, unt i l the pores are g r o w i n g in the ideal arrangement at the metal/oxide interface. The oxide is then stripped us ing a phosphoric ac id /chromic ac id solution, where the chromic ac id protects the a l u m i n u m substrate from dissolut ion once the oxide is r emoved (Figure 4.12). Th i s leaves a scalloped surface that is T a b l e 4.1 Pub l i shed parameters for self-ordered porous a lumina growth. Electrolyte Concentrat ion Potential Temperature Spacing Reference ( V ) ( ! Q (nm) O x a l i c ac id 0.3 M 40 1 105 49,61 Sulfuric ac id 2 0 w t % 18.7 1 50 62 Sulfur ic ac id 0.3 M 25 not reported 66.3 49 Sulfur ic ac id 2 0 w t % 15-25 1 41 - 6 8 63 Phosphor ic ac id 1 wt % 195 not reported 50J 49 103 5.8 w t % H 3 P 0 4 + 1.5 w t % C r 0 3 • 12 h , 6 0 ° C A n o d i z e F i g u r e 4.12 Preparation o f fully-ordered porous a lumina f i l m by a two-step anodizat ion process.The in i t ia l oxide layer is stripped, fo l lowed by further anodizat ion o f the substrate. anodized in a second step for the desired duration, y ie ld ing a fully ordered f i l m o f the appropriate thickness. M a s u d a has elegantly shown that ordered pore growth is initiated by the scal loped surface by preparing s imi lar nanoindentations on a l u m i n u m surfaces w i t h a patterned S i C w a f e r 6 1 and wi th an atomic force m i c r o s c o p e . 6 4 Th i s nanoidentation technique a l lowed the fabrication o f defect-free lattices on the m m scale. M a s u d a also reported that electrolyte concentration and temperature have no significant effect on the formation o f the lattice. The self-ordering o f the pore lattice has also been the subject o f several investigations but st i l l remains to be u n d e r s t o o d . 4 7 - 4 9 ' 6 2 ' 6 5 It has been shown through computer s imulat ion that ordered lattices can arise from in i t ia l pore formation on isolated defects , 5 1 g iven a f ixed interpore spacing. However , the or ig in o f the f ixed lattice spacing remains unclear. The mechanica l stress o f expansion from a l u m i n u m to a lumina has been suggested as a source o f repulsive interaction between g r o w i n g p o r e s . 6 2 N i e l s c h e t al. pointed out that a l l ordered porous a lumina films have a 10% porosity (ratio o f pore area to uni t ce l l area) independent o f the electrolyte and po t en t i a l . 4 9 These points are discussed further be low. 104 4.4.4 Preparation of Optimal Host from Porous Alumina Porous a lumina is evidently a very versatile material , as the pore size and spacing can be easily adjusted over a w ide range and the thickness o f the porous film is s imple to control . T h i n f i lms may be prepared on various substrates by anodiz ing evaporated a l u m i n u m f i lms o f the correct t h i c k n e s s . 6 0 Th i s also a l lows cross-sections for S E M to be prepared i n a convenient manner. The smallest pore size reported i n the literature is 20 n m , obtained us ing 20 wt % sulfuric ac id at 15 V . 6 3 There is a recent report o f a membrane w i t h 5 n m pore s i z e 6 6 but it was s imp ly obtained at a l o w potential without any ordering, and characterized indirect ly by gas diffusion measurements. The condit ions for the preparation o f f i lms w i t h an ordered lattice o f pores w i t h diameters be low 20 n m are not established i n the literature. However , M o s k o v i t s reported unpubl ished results at a conference showing that the pore diameter cou ld be very effectively reduced through manipula t ion o f the electrolyte tempera ture . 6 7 In order to reach temperatures w e l l be low 0 ° C , a mixture o f methanol and water was used as the s o l v e n t . 6 8 U s i n g 1.2 M sulfuric ac id i n a 3:1 mixture o f methanol and water, pore sizes d o w n to 4 n m cou ld be reached at anodizat ion temperatures o f -40 ° C . In our hands, this approach was also successful. It was found that a 1:1 methanol:water mixture was effective as the solvent d o w n to -50 ° C . W h i l e keeping the anodizat ion potential and electrolyte concentration constant, f i lms were anodized at 2 0 ° C , - 8 ° C and - 4 0 ° C (Figure 4.13). The f i lms anodized at 2 0 ° C and - 8 ° C were observed by S E M , whereas the film obtained at -40 ° C required a T E M to dis t inguish the pores. It can be seen that whi le the order ing is not perfect, the pore spacing does not vary significantly from 40 n m - the spacing is clearly fixed by the appl ied potential . The pore diameter, on the other hand, changes dramatical ly w i t h temperature: f rom 21 n m at 20 ° C to 4 n m at -40 ° C (Figure 4.14). 105 Diameter (nm) F i g u r e 4.13 Porous a lumina samples anodized at (a) 20 ° C , (b) -8 ° C and (c) -40 ° C at 15.0 V i n 1.2 M H2SO4 (1:1 H 2 0 : M e O H ) , w i t h result ing pore size distributions. The scale bar is 20 n m . 106 -1 1 1 ' 1 1 r~ -40 -20 0 20 Temperature ( ° C ) F i g u r e 4.14 Effect o f temperature on pore diameter for samples anodized at 15.0 V i n 1.2 M sulfuric ac id . The effect o f electrolyte concentration was also investigated by rais ing the concentration o f sulfuric ac id to 5.0 M . However , there was no effect on the pore size at f ixed potential and temperature. Th i s stands i n contrast w i t h the results o f Patermarakis et al, w h o reported that the square o f the pore base diameter is inversely proport ional to the proton activity at the base o f the pore, based on room temperature data obtained galvanostatically w i t h different concentrations o f sulfuric a c i d . 5 3 The samples anodized at -40°C clearly show some ordering. Th is fact a l lows some comment to be made on the proposal that the order ing o f the pore lattice is d r iven b y mechanica l stress i n the a l u m i n u m to a lumina transformation. The i r coefficients o f thermal expansion are 23 .1X10 - 6 K " 1 and 8 . 4 x l 0 " 6 K " 1 , r e spec t ive ly . 6 9 Th i s difference i n expansion coefficients w o u l d suggest that any effects due to the stress o f expansion w o u l d be very different at lower temperatures, and any process dependent on this stress w o u l d be altered. The fact that the lattice spacing remains the same despite the change i n temperature w o u l d argue against any domina t ing role for mechanica l stress in the order ing process. Further experiments w o u l d be required to validate this hypothesis fully. The drawback o f anodiz ing at lower temperatures is that the anodizat ion rate is reduced drastically. The current density declines to 10 u A cm" 2 at -40 ° C and the oxide growth rate becomes - 2 0 0 nm/24 h . W h i l e this is m u c h slower than no rma l anodization, it is not unreasonable for obtaining a f i l m w i t h a pore size that is otherwise diff icult to fabricate. Moreove r , since the spacing o f the pores is the same at r o o m temperature, the in i t ia l anodizat ion o f the two-step process for preparing highly-ordered f i lms can be carr ied out at r o o m temperature. 4.4.5 Barrier Layer Thinning The barrier layer created dur ing the anodizat ion process (Figure 4.9) must be r emoved to a l low electrical contact to the conjugated po lymer guest. The a l u m i n u m substrate for the porous a lumina film may i tself be used as the cathode material . The requirement then is for a method to remove, or at least to thin significantly, the barrier layer present between the pores and the a l u m i n u m . Phosphor ic ac id etching is the simplest approach to barrier layer removal , p rov ided the a l u m i n u m substrate is r emoved and the film etched from the b o t t o m . 3 2 Otherwise , i on m i l l i n g w i t h an argon-ion beam m a y be e m p l o y e d . 7 0 The pract ical l imi ta t ion o f this method is the diameter o f the beam i n the ion m i l l , w h i c h is on the order o f 1 m m on the instrument available at U B C ( V C R ion m i l l , Meta l s and Mater ia ls Engineer ing) . Th i s restricted the sample area w h i c h cou ld be processed easily, m a k i n g it impract ica l for device fabrication. 108 React ive ion etching ( R I E ) m a y be used attack the a lumina selectively and direct ionally, and this approach was investigated by D r . K . Rademacher i n our research group. In this approach, a boron tr ichloride/argon p lasma is used to etch away the a lumina. A carbon tetrafluoride/oxygen mixture can also be used to etch a l u m i n a , 7 1 but these gases were not readily available on the etcher at U B C . A p p l i e d from the bottom o f the sample (wi th the a l u m i n u m removed by saturated H g C k ) , this method was found to open some pores. Howeve r , the condit ions for opening the pores from the top o f the sample were not found. Barr ie r layer th inning m a y also be achieved b y reducing the potential at the end o f the anodizat ion step. This can be executed either stepwise or gradually. I f this is carr ied out s t e p w i s e , 7 2 pore-branching occurs as smaller pores nucleate at the bottom o f the or ig ina l pores; the branching leads to an inverted tree structure at the bot tom o f the anodized layer. Th i s is used i n the commercia l ly-avai lable A n o d i s c membranes to obtain 20 n m pore sizes. Better control o f the rate o f potential reduct ion leads to a smal l hole at the bottom o f each p o r e . 7 3 It should be noted that extended anodizat ion at very l o w potential (< 1 V ) eventually leads to the separation o f the a lumina membrane from the a l u m i n u m substrate. 4.5 C o n c l u s i o n It was shown that a porous a lumina host for conjugated polymers can be readily prepared w i t h the desired properties. The formation o f smal l diameter films was achieved u s ing the l o w -temperature anodizat ion method reported by M o s k o v i t s . The result ing f i lms show al igned 1-D channels w i t h a diameter o f 4 ± 1 n m , and the barrier layer cou ld be effectively thinned by a s imple potential reduction method. However , many aspects o f porous a lumina film formation are s t i l l poor ly understood, and it m a y be possible to improve the l o w temperature film growth rate by further explor ing different combinat ions o f electrolyte, temperature and anodizat ion potential. 109 Experimental Details 1. SBA-2 Films8 1 2 The 16-3-1 gemin i surfactant was synthesized according to the literature p r o c e d u r e 7 4 . T h i n f i lms wi th the S B A - 2 structure were then g r o w n on m i c a substrates. The reaction mixture was o f the fo l lowing molar compos i t ion : 1.0 H 2 0 : 0.076 H C 1 : 9 . 5 x 1 0 ^ 16-3-1: 1 .8xl0" 3 tetraethylorthosilicate. After 5 minutes o f st irring, this mixture was transferred to a Tef lon- l ined stainless steel bomb. A freshly cleaved m i c a substrate was then floated upside d o w n on the solution, and the sealed b o m b was placed i n an oven at 8 0 ° C for 16 h . After coo l ing to r o o m temperature, the m i c a substrate was retrieved and r insed w i t h d is t i l led water and dr ied i n air. The surfactant was removed by calcinat ion i n air: the temperature was s lowly raised to 540 ° C over 18 h to avoid c racking the film, and held at 5 4 0 ° C for 6 h . X - r a y diffraction patterns were obtained on a R i g a k u 6 k W powder diffractometer. N i t rogen physisorpt ion measurements were carried out on a M i c r o m e r i t i c s A S A P 2010 instrument at 77 K . Samples were embedded i n E P O N resin for u l t ramicrotomy w i t h a d i amond knife (Micros tar ) . 2. FeO:Si0 2 System A 3" sputtering target w i t h a compos i t ion o f 65:35 F e O : S i 0 2 by powder m i x i n g vo lume (Pure Tech) was used to prepare thin films on glass and s i l i con substrates by R F sputtering. Th i s process produces a p lasma i n the w o r k i n g gas (argon), w h i c h then strips material from the target and deposits it on the substrate. It is an efficient way to deposit dielectric materials. The R F power was set to literature value, 200 W , wh i l e the argon pressure was 10 m T o r r , the highest obtainable on the sputtering system being used. A sputtering t ime o f 4 h y ie lded 200 nm-th ick 110 films. The films were dark green i n colour. The oxidat ion o f the films was carried out at 6 0 0 ° C i n air, result ing i n a change o f colour to red. E t c h i n g w i t h 1:1 H C L E ^ O overnight produced a clear film. W h e n cross-sections were required, 0.1 m m - t h i c k coversl ips (#0, E . M . Sciences) were used as the substrate. The cross-sections prepared by S A C were observed w i t h a H i t a c h i H - 8 0 0 T E M equipped w i t h an X - r a y detector for E D X analysis. The analysis area for E D X was selected by us ing the microscope i n S T E M mode and scanning over the area o f interest only. Quantitative analysis o f the E D X spectra was carr ied out us ing the Z A P T E M program, y ie ld ing a compos i t ion o f 2 5 % Fe, 12% S i and 6 3 % O . X P S analysis was performed on a L e y b o l d M A X 2 0 0 system us ing monochromated A l K « radiation as the excitat ion source. The peaks were fitted us ing the X P S P E A K computer program (R. W . M . K w o k , Chinese Un ive r s i t y o f H o n g K o n g ) . The Fe 2p, S i 2p and O Is peaks were used to calculate atomic percentages: 12 % Fe, 13 % S i , 74 % O . Ni t rogen physisorpt ion measurements were carr ied out on a M i c r o m e r i t i c s A S A P 2010 instrument at 77 K . A 200 n m film was deposited onto coversl ips w i t h a total projected area o f 60 c m 2 . The approximate total surface area was 0.5 m 2 , us ing an estimated porosi ty o f 2 5 % and the reported specific surface area o f 800 m 2 g" 1. 3. P o r o u s a l u m i n a A l u m i n u m foi l (99.99+%, A l d r i c h ) was degreased i n acetone, then r insed w i t h dis t i l led water. N o electropolishing was done. A l u m i n u m thin films were prepared on n-type s i l i con substrates by electron-beam evaporation (wi th a base pressure o f 2x10" 6 Torr) from a l u m i n u m slugs (99 .999%, A l f a Aesar) i n a graphite crucible or R F sputtering (3" target, 9 9 . 9 9 % puri ty, 300 W w i t h 6 mTor r argon). I l l A n o d i z a t i o n was carried out under the condit ions g iven i n the text. M o s t samples were anodized without any st i rr ing o f the electrolyte, al though st i rr ing w i l l improve the order ing o f the pore lattice. Fo r the two-step process, the in i t ia l ly g r o w n oxide layer was stripped by immers ing the substrate i n 5.8 wt. % H3PO4 + 1.5 wt. % C r 0 3 at 50 to 60 ° C overnight. A n o d i z a t i o n was then continued at the ini t ia l condit ions. Pore -widen ing was effected by 5 % H3PO4 at 20 ° C for the g iven periods. L o w temperature anodizat ion was carr ied out i n a F T S Systems (Stone R idge , N . Y . ) M u l t i - C o o l chi l ler . Samples were coated w i t h a thin layer o f A u / P d for S E M (Hi tachi S-4700), typical ly w i t h an accelerating voltage between 10 and 20 k V and a 6 m m w o r k i n g distance. Free-standing f i lms and ul t ramicrotomed cross-sections were observed by T E M (Hi tach i H - 8 0 0 at 200 k V , H i t ach i H - 7 6 0 0 at 80 k V ) . Pore size distributions were determined us ing the S c i o n Image computer program (Sc ion Corp . , M D , www.sc ionco rp . com) or the equivalent ImageJ program (Nat iona l Institute o f Heal th, w w w . n i h . g o v ) . The pore diameters were calculated for. a c i rc le o f area equivalent to the imaged pores. A r i o n m i l l i n g was done us ing a V C R Ion m i l l w i t h 5 k V ion energy, 4 0 p A beam current and 9 0 ° angle o f incidence. React ive ion etching was done w i t h a PlasmaQuest E C R etcher. T o open a majori ty o f pores from the bot tom o f the sample, the R F power was 150 W , the mic rowave power 300 W , and the substrate temperature was 48 ° C . The gas flow rates for A r / B C l 3 / C l 2 were 20.5/20.1/2.8 seem, respectively. The etching t ime was 480 s. 112 References 1. Br inke r , C . J . Curr. Opin. Colloid Interface Sci. 1998, 3, 166. 2. Edler , K . J . ; Roser , S. J . Int. Rev. Phys. Chem. 2001, 20, 387. 3. Schuth, F . Curr. Opin. Colloid Interface Sci. 1998, 3, 174. 4. Tsapatsis, M . ; Gavalas , G . R . MRS Bull. 1999, 24, 30 . 5. C h i a n g , A . S. T. ; Chao , K . J . J. Phys. Chem. Solids 2001, 62, 1899. 6. M i y a t a , H . ; K u r o d a , K . J. Am. Chem. Soc. 1999, 121, 7618. 7. R y a n , K . M . ; Erts , D . ; O l i n , H . ; M o r r i s , M . A . ; Ho lmes , J . D . J. Am. Chem. Soc. 2003, 125, 6284. 8. Y a n g , H . ; K u p e r m a n , A . ; C o o m b s , N . ; Mamicheafara , S.; O z i n , G . A . Nature 1996, 379, 703 . 9. Y a n g , H . ; C o o m b s , N . ; Soko lov , I.; O z i n , G . A . J. Mater. Chem. 1997, 7, 1285. 10. Y a n g , H . ; C o o m b s , N . ; D a g , O . ; Soko lov , I.; O z i n , G . A . J. Mater. Chem. 1997, 7, 1755. 11. M i y a t a , H . ; K u r o d a , K . Chem. Mater. 1999,11 , 1609. 12. Tolbert , S. H . ; Schaffer, T . E . ; Feng , J . L . ; Hansma , P . K ; Stucky, G . D . Chem. Mater. 1997, 9, 1962. 13. Z h o u , W . Z . ; Hunter, H . M . A . ; Wr igh t , P . A . ; G e , Q . F . ; Thomas , J . M . J. Phys. Chem. B 1 9 9 8 , 1 0 2 , 6933 . 14. Tsa i , T . G . ; Sh ih , H . C ; L i a o , S. J . ; Chao , K . J . Microporous Mesoporous Mater. 1998, 22, 333. 15. C h u , L . ; Ande r son , M . A . J. Memhr. Sci. 1 9 9 6 , 1 1 0 , 141. 16. N i s h i y a m a , N . ; U e y a m a , K ; Matsukata , M . Microporous Mater. 1996, 7, 299. 17. N i s h i y a m a , N . ; Park, D . H . ; K o i d e , A . ; Egashira , Y . ; U e y a m a , K . J. Memhr. Sci. 2 0 0 1 , 1 8 2 , 235. 113 18. R a i m o n d i , M . E . ; Maschmeyer , T . ; Templer , R . H . ; Seddon, J . M . Chem. Commun. 1997, 1843. 19. F i r o u z i , A . ; Schaefer, D . J . ; Tolbert , S. H . ; S tucky, G . D . ; C h m e l k a , B . F . / . Am. Chem. Soc. 1997,119, 9466. 20 . Tolbert , S. H . ; F i r o u z i , A . ; S tucky, G . D . ; C h m e l k a , B . F . Science 1997, 278, 264. 2 1 . Hi l lhouse , H . W . ; O k u b o , T. ; van E g m o n d , J . W . ; Tsapatsis, M . Chem. Mater. 1997, 9, 1505. 22. K i m , W . J . ; Y a n g , S. M . Chem. Mater. 2000, 12, 3227. 23 . M u n o z , T. ; Ba lkus , K . J . J. Am. Chem. Soc. 1999, 121, 139. 24. Ba lkus , K . J . ; Scott, A . S.; G i m o n - K i n s e l , M . E . ; B l a n c o , J . H . Microporous Mesoporous Mater. 2000, 38, 97. 25. Trau , M . ; Y a o , N . ; K i m , E . ; X i a , Y . ; Whites ides , G . M . ; A k s a y , I. A . Nature 1997, 390, 674. 26. K r o s , A ; Nol t e , R . J . M . ; Sommerd i jk , N . A . J . M . Adv. Mater. 2002,14, 1779. 27. K o n d o h , S.; Iwamoto, Y . ; K i k u t a , K . ; H i rano , S. J. Am. Ceram. Soc. 1999, 82, 209. 28. Thompson , G . E . Thin Solid Films 1997, 297, 192. 29 . Berger, S. D . ; Sal isbury, I. G . ; M i l n e , R . H . ; Imeson, D . ; Humphreys , C . J . Philos. Mag. B 1987, 55 , 341 . 30 . H u o , Q . S.; Margolese , D . I.; Stucky, G . D . Chem. Mater. 1996, 8, 1147. 3 1 . Thompson , G . E . ; W o o d , G . C . Anodic films on aluminum; Treatise o n Mater ia ls Science and Technology; A c a d e m i c Press: N e w Y o r k , 1983; V o l . 23 , p . 205. 32 . M a s u d a , H . ; N i s h i o , K . ; Baba , N . Thin Solid Films 1993, 223, 1. 33 . X u , C . X . ; X u e , Q . H . ; Z h o n g , Y . ; C u i , Y . P . ; B a , L . ; Zhao , B . ; G u , N . Nanotechnology 2002, 13, Al. 34. G a o , T.; M e n g , G . W . ; T i a n , Y . T. ; Sun , S. H . ; L i u , X . O . ; Zhang , L . D . J. Phys.: Condens. Matter 2002, 14, 12651. 114 35. Y a n g , Y . ; L i , H . Y . ; C h e n , H . L . ; B a o , X . M . Chem. J. Chin. Uni. 2002, 23, 768. 36. Hornyak , G . ; K r o l l , M . ; P u g i n , R . ; S a w i t o w s k i , T . ; S c h m i d , G . ; B o v i n , J . O . ; Karsson , G . ; Hofmeister , H . ; Hopfe , S. Chem. Eur. J. 1997, 3, 1951. 37. Hoyer , P . Curr. Opin. Colloid Interface Sci. 1998, 3, 160. 38. She l imov , K . B . ; D a v y d o v , D . N . ; M o s k o v i t s , M . Appl. Phys. Lett. 2000, 77, 1722. 39 . Routkev i tch , D . ; B i g i o n i , T.; M o s k o v i t s , M . ; X u , J . M . J. Phys. Chem. 1996, 100, 14037. 40 . Y a n g , Y . ; C h e n , H . L . ; M e i , Y . F . ; C h e n , J . B . ; W u , X . L . ; B a o , X . M . Sol id State Commun. 2 0 0 2 , 1 2 3 , 279. 4 1 . M a s u d a , H . ; O h y a , M . ; A s o h , H . ; N a k a o , M . ; N o h t o m i , M . ; Tamamura , T . Jpn. J. Appl. Phys., Part 2 1999, 38, L 1 4 0 3 . 42 . Y o k o o , A . ; N o t o m i , M . ; S u z u k i , H . ; N a k a o , M . ; Tamamura , T. ; M a s u d a , H . IEEE J. Quantum Elec. 2002, J S , 938. 4 3 . L i , A . P . ; M u l l e r , F . ; B i rne r , A . ; N i e l s c h , K . ; Gosele , U . Adv. Mater. 1999 , 1 1 , 483 . 44 . M a s u d a , H . ; Watanabe, M . ; Y a s u i , K . ; T r y k , D . ; Rao , T. ; Fu j i sh ima, A . Adv. Mater. 2000, 12, 444. 45 . N a k a o , M . ; O k u , S.; Tanaka, H . ; Shibata, Y . ; Y o k o o , A . ; Tamamura , T. ; M a s u d a , H . O p t . Quantum Mater. 2002, 34, 183. 46 . M i k u l s k a s , I.; Juodkazis , S.; Tomasiunas , R . ; D u m a s , J . G . Adv. Mater. 2001 , 1 3 , 1574. 47 . L i , F . Y . ; Zhang , L . ; Metzger , R . M . Chem. Mater. 1998, 10, 2470 . 48 . M a s u d a , H . ; Hasegwa, F . ; O n o , S. J. Electrochem. Soc. 1 9 9 7 , 1 4 4 , L 127. 49 . N i e l s c h , K . ; C h o i , J . ; S c h w i r n , K . ; Wehrspohn , R . B . ; Gosele , U . Nano Lett. 2002, 2, 677. 50. O n o , S.; M a s u k o , N . Corros. Sci. 1992, 33, 503. 51 . Randon , J . ; M a r d i l o v i c h , P . P . ; Govyad inov , A . N . ; Paterson, R . J. Colloid Interface Sci. 1995, 169, 335 . 115 52. Patermarakis, G . ; Moussoutzanis , K . ; N i k o l o p o u l o s , N . J. Solid State Electrochem. 1999, 3, 193. 53. Patermarakis, G . ; Moussoutzanis , K . ; Chandr inos , J . J. Solid State Electrochem. 2001, 6, 39. 54. M a r d i l o v i c h , P . P . ; G o v y a d i n o v , A . N. ; M a z u r e n k o , N . I.; Paterson, R . J. Membr. Sci. 1995, 98, 143. 5 5 . M a r d i l o v i c h , P . P . ; G o v y a d i n o v , A . N . ; M u k h u r o v , N . I.; R z e h v s k i i , A . M . ; Paterson, R . J . Membr. Sci. 1995, 98, 131. 56. Ozao , R . ; O c h i a i , M . ; Y o s h i d a , H . ; Ich imura , Y . ; Inada, T . J. Therm. Anal. Calorim. 2001, 64, 923 . 57. Ozao , R . ; O c h i a i , M . ; Ichimura , N . ; Takahashi , H . ; Inada, T . Thermochim. Acta 2000, 352, 9 1 . 58. Ozao , R . ; Y o s h i d a , H . ; Ichimura , Y . ; Inada, T. ; O c h i a i , M . J. Therm. Anal. Calorim. 2001, 64, 915. 59. Ozao , R . ; Y o s h i d a , H . ; Inada, T. J. Therm. Anal. Calorim. 2002, 69, 925. 60 . Crouse , D . ; L o , Y . H . ; M i l l e r , A . E . ; Crouse , M. Appl. Phys. Lett. 2000, 76, 49 . 6 1 . M a s u d a , H . ; Y a m a d a , H . ; Satoh, M . ; A s o h , H . ; N a k a o , M . ; Tamamura , T. Appl. Phys. Lett. 1997, 7 1 , 2770. 62 . Jessensky, O . ; M u l l e r , F . ; U., G . Appl. Phys. Lett. 1998, 72, 1173. 63. Su lka , G . D . ; Stroobants, S.; M o s h c h a l k o v , V . ; Borghs , G . ; Ce l i s , J . P . J. Electrochem. Soc. 2 0 0 2 , 1 4 9 , D 9 7 . 64. M a s u d a , H . ; Kanezawa , K . ; N i s h i o , K . Chem. Lett. 2002, 1218. 65 . L i , A . P . ; M u l l e r , F . ; B i rne r , A. ; N i e l s c h , K . ; Gosele , U . J. Appl. Phys. 1998, 84, 6023. 66. L i r a , H . D . L.; Paterson, R . J. Membr. Sci. 2002, 206, 375 . 67 . M o s k o v i t s , M . , Canad ian Society for Chemis t ry Conference, M a y 2000 , Calgary , A B . 116 68. Ott, J . B . ; Goates, J . R . ; Wai te , B . A . J. Chem. Thermodyn. 1979, 11, 739. 69 . L i d e , D . R . , E d . CRC Handbook of Chemistry and Physics ;CRC Press: B o c a Raton, 2000 70. A l m a w l a w i , D . ; B o s n i c k , K . A . ; O s i k a , A . ; M o s k o v i t s , M . Adv. Mater. 2000 , 1 2 , 1252. 71 . L i a n g , J . Y . ; C h i k , H . ; Yin , A . J . ; X u , J . J. Appl. Phys. 2002, 91, 2544. 72. Furneaux, R . C . ; R i g b y , W . R . ; D a v i d s o n , A . P . Nature 1989, 337, 147. 73 . K u k h t a , A . V . ; G o r o k h , G . G . ; K o l e s n i k , E . E . ; Mi tkove t s , A . I.; Taoub i , M . I.; K o s h i n , Y . A. ; M o z a l e v , A . M . Surf. Sci. 2002, 507, 593. 74. Zana , R . ; Benrraou, M.; Rueff, R . Langmuir 1991, 7, 1072. 117 CHAPTER 5 Characterization of a Porous Alumina/MEH-PPV Composite Material Compos i t e materials based on porous a lumina and conjugated polymers have been extensively investigated in our research group. A full characterization o f these materials was important for understanding their properties, and the central goal was to establish the polymer distr ibution w i t h i n the composite material on the nanometre scale. The in i t ia l experiments were a imed at investigating samples prepared by adsorbing the conjugated po lymer po ly[2-methoxy,5-(2 ' -e thylhexyloxy) l ,4-phenylene vinylene) ( M E H - P P V ) from solution onto porous a lumina (Figure 5.1). Th i s straightforward procedure y ie lded samples i n w h i c h the presence o f po lymer was evidenced by the colour and fluorescent emiss ion o f the porous a lumina layer (Figure 5.2). Confoca l fluorescence microscopy was used to establish that M E H - P P V was distributed throughout the thickness o f the porous a lumina host w i t h 0.1 urn resolution. However , the conjugated po lymer could not be located wi th higher spatial resolution by either scanning electron mic roscopy ( S E M ) or t ransmission electron microscopy ( T E M ) , o w i n g to the poor beam contrast o f the po lymer guest. The v i n y l i c carbons in M E H - P P V were readily stained by o s m i u m tetroxide, as evidenced by the disappearance o f the red-orange colour o f the sample, but this d id not improve the contrast in T E M or S E M , most l ike ly due to the l imi ted amount o f po lymer on the porous a lumina surface. 118 D ~ 60 n m F i g u r e 5.1 Adso rp t i on o f thin layer o f M E H - P P V onto porous a lumina host, showing (a) empty host, and polymer-coated host in (b) plan v i e w and (c) as thin section for T E M . F i g u r e 5.2 Porous a lumina f i l m after soaking in M E H - P P V solution, seen in cross-section, as shown by (a) l ight microscopy (b) fluorescence microscopy. The scale bar is 10 urn. Samples and images prepared by D r . K . Rademacher . 119 There were no previously publ ished studies o f porous a lumina f i lms by E E L S at the outset o f this work ; subsequently, a very relevant study o f epoxy infiltration into porous a lumina appeared in the literature. 1 In that work , the distr ibution o f carbon, a l u m i n u m and oxygen was mapped by S T E M / E E L S us ing the corresponding ionizat ion edges. The samples were also held at -134 ° C us ing a cryogenic holder to m i n i m i z e beam damage to the sample. In pr inciple , the porous a l u m i n a / M E H - P P V sample could have been mapped in the same manner as P P V / M C M - 4 1 (chapter 3) through the use o f the 7 e V % plasmon. The carbon K-edge cou ld also have been used. Nei ther o f these methods was successful but the experiments d i d produce some interesting results for further investigation. 5.1 E E L S S a m p l e s T w o geometries are possible for investigating the porous a l u m i n a / M E H - P P V composite materials by E E L S : the p lan geometry (Figure 5.1(b)) and the cross-section (Figure 5.1(c)). The cross-sections may be prepared by ul t ramicrotomy, y i e l d i n g thin sections ( - 3 0 n m th ick) o f the material . A s the samples were k n o w n to contain very little po lymer - poss ibly as little as a monolayer on the surface o f the pores - this geometry w o u l d present a very smal l amount o f po lymer to the electron beam. In chapter 7, this geometry is investigated for samples wi th larger amounts o f polymer . It was deemed advantageous to investigate these particular samples i n the plan geometry ini t ia l ly , such that the electron beam w o u l d pass d o w n the pores and interact w i t h a l l the po lymer distributed a long the length o f the pores. A number o f samples were prepared for E E L S measurements. In order to obtain the clean loss spectrum o f M E H - P P V , thin film samples were prepared by casting from T H F over water 120 and col lect ing the resulting f i lm wi th a lacey carbon gr id . The empty host material consisted o f porous a lumina membranes (0.2 and 1.4 u m thick) , and the composite material investigated was a 1.4 u m porous a lumina membrane w h i c h had been soaked in M E H - P P V . F i g u r e 5.3 S E M image o f cross-section o f sample for E E L S experiments. The scale bar is 1.0 u m . 5.2 M E H - P P V L o w - L o s s S p e c t r a a n d Z e r o - L o s s P e a k R e m o v a l The ini t ial investigation o f M E H - P P V focused on the low-loss spectrum and the dist inctive 7r-7t* p lasmon o f conjugated organic materials. The bright field T E M image o f the M E H - P P V f i l m is shown in Figure 5.4, and the collected loss spectrum is in Figure 5.5. The 7t-7i* p lasmon was seen clearly at 6.4 e V , whereas the bulk p lasmon appeared at 22 e V , as anticipated for an aromatic material . However , there were no ini t ial ly discernible features associated w i t h the onset o f optical absorption o f the po lymer chain at -3 e V , largely due to the tai l o f the zero-loss peak. For these experiments, a clean zero-loss peak was not recorded immediately, and later acquis i t ion y ie lded a peak w i t h a significantly different shape. Hence a number o f different 121 F i g u r e 5.4 T E M image o f M E H - P P V f i lm supported by a lacey carbon gr id . The scale bar is 100 n m . approaches to r emoving the zero-loss peak from the spectrum were evaluated, as the tail o f the zero-loss peak affects most o f the features be low 10 e V . The possible approaches include Four ier - log deconvolu t ion , 2 matr ix deconvolu t ion , 3 extrapolation o f the tail by a power- law fit and direct subtraction o f the zero-loss peak . 4 These procedures were implemented as V i s u a l B a s i c for E x c e l 5 scripts. Four ier - log and matr ix deconvolut ion were based on F O R T R A N programs presented by Eger ton , 2 wh i l e the power- law fit was a s imple linear least-squares regression procedure. Four ier - log deconvolut ion was found to remove the effect o f plural scattering very w e l l , but it d id introduce some artefacts near 4 e V , where the routine chooses the separation point between the zero-loss peak and the remainder o f the spectrum. The algorithm was not designed to deal wi th spectra showing substantial losses on the tail o f the zero-loss peak. Th i s cou ld have also been avoided by acquir ing a clean zero-loss spectrum separately. M a t r i x deconvolut ion d id not remove the zero-loss peak very effectively but d id remove the plural scattering smearing out the bulk p lasmon peak. A s such it was not very useful for examinat ion o f the spectrum over the 2 to 5 e V range. 122 Energy (eV) F i g u r e 5.5 Low- los s spectrum o f M E H - P P V thin film, also showing various approaches to r emov ing the zero-loss peak: ( • ) raw data, ( • ) matr ix deconvolut ion, ( A ) Four ier - log deconvolut ion, (o) power law fit over 1.3 - 2.0 e V . The inset shows the detail over 0 - 1 0 e V . Di rec t subtraction o f the zero-loss peak can be applied in a very accurate way by the use o f spline interpolation and curve-fi t t ing techniques, as has been reported recent ly . 4 Th i s method seemed to be the most reliable for revealing weak spectral features in the 2 to 5 e V range. However , it again required an accurate spectrum o f the clean zero-loss peak; the lack o f this information made it impossible to apply to the data presented here. F ina l ly , the power- law fit does not address plural scattering but does remove the ta i l o f the zero-loss peak fairly smoothly, w i th m i n i m a l artefacts in the 2-5 e V range. The possibi l i ty o f a weak shoulder feature being fitted out by this procedure cannot be ruled out, however (later 123 experiments showed that such a shoulder was present in some o f the spectra - see section 6.5). Nevertheless, this approach seemed the most reliable one and as such, the power- law fit was used to process al l the spectra, us ing least-squares linear regression to calculate the curve parameters. The implementat ion as a script a l lowed large numbers o f spectra to be processed easily. 5.3 P o r o u s A l u m i n a L o w - L o s s S p e c t r a The low-loss spectrum o f porous a lumina (Figure 5.6) showed a bu lk p lasmon at 23 e V , w h i c h is i n agreement w i t h the literature value for amorphous a l u m i n a . 6 Crysta l l ine a lumina presents a p lasmon at a slightly higher energy (26 e V ) . 2 T h i s conf i rmed the amorphous nature o f the pore walls , as reported in the literature from T E M observa t ions . 7 ' 8 A d d i t i o n a l loss modes appear be low the bulk p lasmon: one or two be low 10 e V , and one at 13 e V . The nature o f these addi t ional modes has not been fully established in the literature; it has been suggested that they are due to an a luminum(O) surface p lasmon w h i c h w o u l d arise f rom a deviat ion from the ideal a lumina s to ichiometry , 6 but this is very un l ike ly for porous a lumina . Reduc t ion o f a lumina to a l u m i n u m by the electron beam is not observed in amorphous a lumina, as a l u m i n u m atoms are removed preferentially over oxygen a toms . 9 The presence o f these addit ional loss modes, especially the ones be low 10 e V , presents a diff iculty as they w o u l d possibly mask the conjugated po lymer 71-71* p lasmon near 6 e V . Unde r these circumstances, it was necessary to study these addit ional loss modes in more detail. 124 I ' 1 ' 1 ' 1 ' 1 -0 10 20 30 40 Energy (eV) F i g u r e 5.6 L o w - l o s s spectrum o f pore in porous a lumina film, showing ( • ) collected data and (o) data w i t h zero-loss peak removed by a power law fit over 1.3 to 2.0 e V . The distr ibution o f a l l the loss modes in a pore was determined by acqui r ing loss spectra at regular intervals over a line crossing a. pore in S T E M mode (Figure 5.7). Th i s revealed a strong loss mode near the pore w a l l , some addit ional loss modes at intermediate distances from the w a l l , and one loss mode w h i c h extended at nearly constant level throughout the pore. T y p i c a l spectra for these three regimes are shown i n Figure 5.8: (a) at the w a l l , the a lumina bulk p lasmon is seen at - 2 2 e V , wi th a strong tail due to p lura l scattering; (b) at - 7 n m from the w a l l , three surface plasmons at 8, 13 and 18 e V appear, w i t h an addit ional shoulder at 7 e V ; the shoulder at 3 e V is probably an artifact due to zero-loss peak subtraction; (c) at - 23 n m from the w a l l , a strong peak at 7 e V is seen, and the surface plasmons are st i l l seen as weaker shoulders. 125 These peak posit ions are most l ike ly only accurate to about 1 e V , due to the presence o f over lapping peaks and a substantial background due to p lura l scattering. The distr ibution o f these low-loss modes can also be v isua l ized by E F T E M , as shown in Figure 5.9 for a thinner (0.2 urn) porous a lumina film. The filtered images conf i rm that the 7 e V mode is weaker but evenly distributed throughout the hole. The losses at 13 e V are seen to be confined to the neighbourhood o f the surface, and the bulk p lasmon can be seen throughout the film and jus t outside the surface. The dis tr ibut ion o f the 22 e V losses also shows the areas o f lower and higher w a l l density: the material most ly consists o f l ow density a lumina w i t h a higher density core fo rming a lattice between the pores . 8 126 F i g u r e 5.7 (a) S T E M dark-f ield image o f 1.4 u m thick porous a lumina film, showing the location o f l ine along w h i c h loss spectra were acquired at 2 n m intervals. The scale bar is 100 n m . x 0 10 20 30 4 0 50 Energy (eV) F i g u r e 5.7 (b) Deta i l o f analysis line and spatially-resolved low-loss spectra o f porous a lumina film. The zero-loss peak was removed by a power- law fit. 127 60000 Energy (eV) F i g u r e 5 . 8 Representative low-loss spectra for porous a lumina f i l m : ( • ) near pore centre, (o)~7 n m from pore w a l l , and ( A ) at w a l l . F i g u r e 5 . 9 Energy-fi l tered images o f 0.2 u m porous a lumina f i l m , us ing a 2 e V w i n d o w : (a) 0 e V , (b) 6 e V , (c) 13 e V , (d) 22 e V . The scale bar is 100 n m . 128 5.4 P o r o u s A l u m i n a / M E H - P P V C o m p o s i t e S p e c t r a The spatial dis tr ibution o f the low-loss spectra in a porous a l u m i n a / M E H - P P V composi te material was obtained in the same manner, as shown in Figure 5.10. The distr ibution o f loss modes was very s imi lar to that o f empty porous a lumina host. The same strong peak at 7 e V was observed over the centre o f the pore, but w i t h no shoulder peaks at h igher energies. A t intermediate distances, the surface plasmons were identical to those o f the empty porous a lumina. The smal l differences observed in these modes may reflect the presence o f polymer; detailed mode l l ing o f these effects w o u l d however require k n o w i n g the dielectric function o f M E H - P P V to at least 10 e V , and at present it is only k n o w n to 5 e V . 1 0 In pr inciple , this could be determined from a carefully measured M E H - P P V loss spectrum; however that was deemed beyond the scope o f this thesis. The near-wall spectra are shown in more detail in Figure 5.12. The low signal-to-noise level does not a l low any significant conclusions to be made about the presence o f the po lymer through the 7t p lasmon. Some attempts were made to f ind the carbon K-edge loss near 280 e V but without any success. Th i s approach has been shown to be successful for mapp ing epoxy penetration into porous a l u m i n a ; 1 therefore the absence o f the s ignal near 280 e V can be ascribed to the l o w amount or rapid degradation o f the po lymer in the samples investigated. 129 F i g u r e 5.10 (a) S T E M dark-f ield image o f porous a l u m i n a / M E H - P P V composite, showing the l ine along w h i c h E E L spectra were acquired. The scale bar is 100 n m . x 0 10 20 30 40 50 Energy (eV) F i g u r e 5.10 (b) Deta i l o f analysis line and spatially-resolved low-loss spectra o f porous a l u m i n a / M E H - P P V composite film. The zero-loss peak was removed by a power- law fit from 1.3 to 2.0 e V . 130 40000 \-20 30 Energy (eV) F i g u r e 5.11 Compar i son o f low-loss spectra (o,«) near pore axis and ( • , • ) at ~ 7 n m from the pore w a l l , for porous a lumina and porous a l u m i n a / M E H - P P V composite, respectively. 1000 20 30 Energy (eV) F i g u r e 5.12 C o m p a r i s o n o f low-loss spectra o f empty (•) and M E H - P P V - t r e a t e d ( A ) f o rous a lumina , nearest to pore w a l l . 131 5.5 C o n c l u s i o n A l t h o u g h the use o f the TT-TT* p lasmon to identify conjugated organic molecules by E E L S and E F T E M is w e l l established, its applicat ion to the porous a l u m i n a / M E H - P P V composite in the plan geometry was not successful. Low-energy surface plasmons and an unexpected long-range interaction at 7 e V effectively masked the region o f interest; the difficulty was compounded by the smal l amount o f po lymer present in the composite . The effect o f surface plasmons w o u l d have been m i n i m i z e d by employ ing a cross-sectional geometry instead. A larger po lymer content in the composi te w o u l d have also provided a stronger p lasmon loss peak. S u c h samples are investigated by the same methods in chapter 7. However , the most interesting feature ar is ing from these results is indeed the loss peak at 7 e V . The nature o f this interaction was studied i n more detail and is the subject o f the fo l lowing chapter. E x p e r i m e n t a l De t a i l s Porous a lumina thiri f i lms were prepared by anodiz ing 0.2 and 1.0 u m thick a l u m i n u m films evaporated onto rc-type s i l i c o n ( l l l ) wafers by e-beam evapora t ion . 1 1 The anodizat ion was carr ied out in 0.3 M oxal ic ac id at 20 ° C w i t h an applied potential o f 40.0 V , us ing a glass ce l l w i th an O - r i n g seal to the sample. U p o n complet ion o f the anodizat ion, the porous a lumina film (1.4 u m thick) had detached from the substrate, but was st i l l attached to the surrounding a l u m i n i u m film. The barrier layer was removed by etching in 5 wt. % phosphoric ac id for 40 m i n . The film was r insed w i t h dist i l led water, fo l lowed by ethanol and then dried. Some films were soaked in a solution o f M E H - P P V (0.038 wt. % in T H F ) for 48 h . A d iamond scribe was used to cut the porous a lumina film to fit a 3 m m C u T E M gr id , to w h i c h it was f ixed us ing a 132 smal l amount o f epoxy glue. M E H - P P V films were cast over water from a 0.05 wt % solution in T H F and collected w i t h a holey carbon gr id (Ted Pe l l a Inc.). O s m i u m tetroxide staining was carr ied out before embedd ing the samples epoxy for ul t ramicrotomy. Samples were exposed to o s m i u m tetroxide vapour by p lac ing them next to a drop o f 4 % aqueous solution in a covered dish for 30 m i n . Di rec t immers ion in this solution was also used. T h i n sections were obtained by ul t ramicrotomy. The samples were first sputter-coated w i t h A u / P d to prevent epoxy penetration into the porous a lumina . The embedding m e d i u m was either E P O N epoxy (various suppliers) or 3 0 2 - 3 M epoxy (Epotek, Inc.). Sect ioning was then carr ied out us ing a d iamond knife w i t h a 4 5 ° edge set w i th a 4 to 6° clearance angle, us ing water as the section col lect ing l iqu id . Elec t ron energy-loss spectroscopy was carried out on a Ph i l ip s C M 20 T E M at C A N M E T / N a t u r a l Resources Canada equipped wi th a Schottky field-emission electron gun operated nomina l ly at 200 k V , but lowered to 197 k V for spectroscopy. Measurements were made w i t h a Gatan Imaging Fil ter us ing a dispersion o f 0.05 eV/channe l on a 1024-channel detector, cover ing -5 to 45 e V losses. Spatial ly-resolved spectra were collected at regular intervals w i th an estimated probe size o f < 1 n m and an energy resolution o f 1.0 e V ( F W H M o f zero-loss peak). 133 References 1. Arayasantiparb, D . ; M c K n i g h t , S.; L ibe ra , M . J. Adhes. 2001, 76, 353. 2. Eger ton, R . F . Electron energy-loss spectroscopy in the electron microscope; 2 n d ed.; P l enum Press: N e w Y o r k , 1996. 3. S u , D . S.; Schattschneider, P . J. Microsc. 1 9 9 2 , 1 6 7 , 63 . 4. Reed , B . W . ; Sar ikaya, M . Ultramicroscopy2002, 93, 25 . 5. Microsof t ; R e d m o n d , W A . 6. Kapsa , R . ; Stara, I.; Zeze, D . ; G r u z z a , B . ; M a t o l i n , V . Thin Solid Films 1998, 317, 77. 7. Thompson , G . E . ; W o o d , G . C . Anodic films on aluminum; Treatise on Mater ia ls Science and Technology; A c a d e m i c Press: N e w Y o r k , 1983; V o l . 23 , p. 205 . 8. N i e l s c h , K . ; C h o i , J . ; S c h w i r n , K . ; Wehrspohn , R . B . ; Gosele , U . Nano Lett. 2002, 2, 677. 9. Berger, S. D . Philos. Mag. B 1987, 55 , 341 . 10. Tammer , M . ; M o n k m a n , A . P . Adv. Mater. 2002, 14, 210. 11. Crouse , D . ; L o , Y . H . ; M i l l e r , A . E . ; Crouse , M . Appl. Phys. Lett. 2000, 76, 49 . 134 CHAPTER 6 Aloof Cherenkov Effect in Porous Alumina The unexpected low-loss spectral feature observed by electron energy-loss spectroscopy ( E E L S ) in the centre o f the pores o f porous a lumina f i lms, as described in chapter 5, warranted further investigation. Several aspects o f this feature were unusual: the lack o f plural scattering, the rather large distances - up to 30 n m - f rom the pore w a l l at w h i c h it was st i l l present without very m u c h decay in intensity, and the fact that it appeared i n the spectral region usually dominated by the optical properties o f the material undergoing analysis. These characteristics suggested that the observed feature was not a surface plasmon. Pre l iminary discussions w i t h theoreticians experienced wi th mode l l ing E E L S suggested that the most plausible hypothesis for its o r ig in was rooted in the Cherenkov effect. A collaborative effort was then undertaken to validate this hypothesis. Th i s chapter describes the further experimental and theoretical results that were used to show that the Cherenkov effect is responsible for these losses. In this chapter, the energy o f the electron beam i n the t ransmission electron microscope ( T E M ) is g iven in units o f k e V , w h i c h is preferred from the theoretical standpoint. Th i s is equivalent to the experimental accelerating voltage i n the T E M . 6.1 T h e C h e r e n k o v Ef fec t The Cherenkov effect is described as the emiss ion o f radiation when a charged particle moves through a m e d i u m at a speed greater than the speed o f light i n that particular m e d i u m . 1 135 The effect was observed by Cherenkov i n 1934 , 2 in the form o f a g low i n fluids exposed to a radioactive source. The addi t ion o f quenching agents d id not alter the luminescence, suggesting that the radiation was extraordinary i n nature, and a ful l explanation o f the effect was provided by T a m m and Frank in 1937 . 3 The complex dielectric function e(co) = £i((o) + isiico) defines the interactions between a m e d i u m and electromagnetic radiation. The real part alters the wavelength o f the propagating radiation, thus reducing the speed o f propagation o f the radiation at an energy co to c I ^ j s x ( c o ) , where c is the speed o f light in vacuum. The imaginary part o f the dielectric function describes the attenuation at a g iven energy co. Thus the Cherenkov condi t ion is satisfied for electron velocit ies v w h e n s x (co)> c 2 / v 2 , w h i c h results in the emiss ion o f radiation w i t h frequency co in a h o l l o w cone wi th half-angle g iven by cost9 c = clv^sx(co) (Figure 6.1). Th i s radiation is analogous to the Shockwave produced by objects t ravel l ing faster than the speed o f sound in air. F o r 200 k e V electrons, the Cherenkov condi t ion is £\(co) > 2 .1 , w h i c h is amply satisfied for energies up to 12 e V for a lumina (Figure 6.2). The thresholds for other electron energies c o m m o n l y used in T E M are shown i n Table 6.1. Cherenkov radiation can be produced by electrons t ravel l ing in v a c u u m but passing near a material , p rov ided the Cherenkov condi t ion is satisfied w i th in the material . The radiation is then produced by the induced polar izat ion in the material and it is expected to be important up to electron-surface separations o f the order o f the range o f the evanescent field o f the electron in v a c u u m . For a g iven frequency component co, the decay constant is ~v/co = Xv/2nc, where A is the wavelength o f the Cherenkov rad ia t ion . 4 Fo r 7 e V radiation, this is approximately 20 n m . It 136 w i l l be shown be low that the Cherenkov losses are actually sensitive to the sample structure up to m u c h larger distances, due to the radiative nature o f the Cherenkov effect. i (a) B F i g u r e 6.1 Geometry o f Cherenkov radiation due to an electron t ravel l ing (a) through a m e d i u m and (b) near a m e d i u m wi th dielectric function s(co): l ight is emitted in the forward direction in a h o l l o w cone. —1—1—1—1—1—1—1—1—t— J 1 1 1 1 1 1 1 1 1 j 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1— lRe[e((o)] -V \ r \ \ \ i 1 •1 \ lm[e(o))] ~ 1 l \ I 1 1 \ X 1 I * — 1 * 1 1 \ I 1 1 \ I \ I \ \ I _ 9 § i i i i 1 i - * • i i \ \ \ \ \ N \ N \ \ \ \ \ ^ 1 i i i i 1 i r T i 1 i i i i—i—i—i—1—i—i—i—i—1—i—i—i—i— 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 co(eV) F i g u r e 6.2 Die lec t r ic function o f a lumina , after data from ref. 5. The Cherenkov condi t ion is satisfied up to 12 e V for 200 k e V electrons. Figure courtesy o f A . R ivacoba . 137 T a b l e 6.1 Cherenkov condi t ion for c o m m o n T E M beam energies. B e a m Energy P M i n i m u m s^co) ( k e V ) (v/c) 80 0.50 4.0 100 0.55 3.3 120 0.59 2.9 200 0.70 2.1 U s i n g E E L S , losses due to the Cherenkov effect have been observed in d ie lec t r ics 6 and s emiconduc to r s 6 ' 7 w i t h the electron beam passing through the material , and the accompanying opt ical emissions have been recently measured . 8 It is only recently that the " a l o o f , or "near-f i e l d " Cherenkov effect has been considered, i n a report on the loss spectra o f a lumina nanopart ic les . 9 These results showed loss features be low 10 e V for non-penetrating trajectories w h i c h could not be simulated w i t h a non-relativistic mode l , but a relativistic mode l was not presented. Porous s i l i con also shows unusual spectral features at large beam-surface separations w h i c h may be due to the Cherenkov ef fec t . 1 0 A s such, the a loo f Cherenkov effect had not been fully validated in the literature and merited further experimental and theoretical investigation. A s a point o f interest, the microscopic nature o f Cherenkov radiation is st i l l the subject o f debate. The long-standing v iewpoin t has been that the radiation is not emitted by the charged particle itself, rather by oscillations i n the m e d i u m . 1 ' > 1 2 H o w e v e r , it has been recently suggested that the radiation is i n fact better described as or iginat ing f rom the t ravel l ing c h a r g e . 1 3 Th i s debate may be purely academic, as certainly both the m e d i u m and the t ravel l ing charge are required for the phenomenon. In the theoretical models presented be low, the Cherenkov emiss ion clearly arises f rom the response o f the m e d i u m to the passing external electron. 138 6.2 Further Measurements A number o f variables could affect the 7 e V peak: pore diameter, lattice spacing, sample thickness, and pr imary electron beam energy. The pr imary beam energy was chosen as being the most indicative factor for the Cherenkov effect, as it w o u l d alter the Cherenkov condi t ion , thus induc ing a change in any loss features associated wi th Cherenkov radiation. It also addresses the concern that the observed peak is due to a finite pore length effect in the sample or an unexpected surface p lasmon. The effect o f va ry ing the pore diameter w i t h a fixed lattice spacing was also investigated. 6.3 Effect of Primary Beam Energy The effect o f p r imary beam energy was investigated by recording loss spectra w i th a lower p r imary beam energy. Measurements at 50 k e V w o u l d have been the most reveal ing, since the Cherenkov condi t ion w o u l d no longer have been satisfied and any associated losses w o u l d have disappeared altogether. T h i s was not possible due to instrumental l imitat ions, as the available Gatan Imaging Fil ter (GIF) was set up for operation at 117 and 197 k e V . The measurements were carried out on the sample used earlier (chapter 4) , a 1.4 u m porous a lumina membrane. The microscope camera length and G I F entrance aperture were var ied to investigate the effect o f the spectrometer col lect ion angle ((3) on the collected spectra. The new measurements at 197 k e V conf i rmed the reproducibi l i ty o f the peak at 7 e V in a 60 n m pore (Figure 6.3). T h e results o f the measurements at 117 k e V on the 1.4 u m film are shown i n Figure 6.4: the general form o f the loss spectra was very s imi lar to those taken at 197 k e V , but certain subtle differences were apparent when the spectra were compared (Figure 6.5). There is a smal l blue-shift o f the loss peaks to higher energies (from 7 to 8 e V ) , both at the centre o f the pore and at intermediate distances. The surface p lasmon peaks are seen to be at the 139 same energies, indicat ing that the shift is not due to cal ibrat ion error or to a change in geometry. N e a r the pore w a l l , the bulk plasmons are also seen to be in agreement at the two beam energies, again conf i rming that the shift observed in the axia l loss peak is real and caused by the change in beam energy. The col lect ion angle was var ied from 0.34 mrad to 6.0 mrad , w h i c h had very little effect on the form o f the spectrum. Th i s is in agreement w i t h the smal l scattering angles for electrons interacting wi th surface plasmons. The scattering angle o f the Cherenkov effect is also very s m a l l . 1 4 140 Energy (eV) F i g u r e 6.3 (a) S T E M dark-f ie ld image, showing pore used for E E L S analysis at 197 k e V (scale bar is 100 nm) . (b) E E L S spectra acquired over the pore diameter w i th P = 0.34 mrad . The zero-loss peak has been removed using a power law fit over 1.3 - 2.0 e V . 141 Energy (eV) F i g u r e 6.4 (a) S T E M dark-f ield image, showing pore used for E E L S analysis at 117 k e V (scale bar is 100 nm). (b) E E L S spectra acquired over the pore diameter w i th (3 = 1.5 mrad. The zero-loss peak has been removed us ing a power law fit over 1.3 - 2.0 e V . 142 3000 < ^—* co c 0 -»—' c 2000 1000 197 k V , 0 = 0 .34 m r a d 117 k V , p = 1.5 m r a d s = 22 nm 20 30 Energy (eV) 600 < 400 CO £= d) 200 B • 197 kV, p = 0.34 mrad 117 kV, (3= 1.5 mrad s = 28 nm Energy (eV) F i g u r e 6.5 C o m p a r i s o n o f low-loss spectra at 197 and 117 k e V pr imary beam energies: (a) ax ia l (s = 0 nm) and intermediate (s = 22 nm) spectra, and (b) wal l -graz ing (s = 28 nm) spectra. 6.4 M o d e l l i n g o f the A l o o f C h e r e n k o v Ef fec t The energy loss spectra were model led by N . Zabala , A . R i v a c o b a and F . J . G a r c i a de A b a j o at the Basque Count ry Unive r s i ty i n Spa in . The details o f the mathematical models are not presented here but they can be found in the l i t e r a t u r e . 1 5 ' 1 6 A s mentioned in chapter 2, the l ow energy (< 50 e V ) interactions between an electron beam and a material can be described by dielectric theory, where the dielectric function £($>) defines the response o f the material to the electric f ield o f the t ravel l ing electrons. The electrons experience energy losses due to the potential induced on the m e d i u m : the electric field o f the electron beam polarizes the m e d i u m , and this induced polar izat ion creates an electric field w h i c h acts to s low the electron beam. The losses are first calculated in the form o f the distr ibution o f energy loss probabil i t ies per unit length, dP(co)/dz. Th i s w o u l d represent the form o f the loss spectrum w i t h perfect energy resolution and wi thout any p lura l scattering, a l l o w i n g the expected losses to be seen in detail. The finite energy distr ibution o f the electron beam (i.e., the shape o f the zero-loss peak) is then taken into account through a convolu t ion process, along wi th the effect o f mul t ip le scattering, to y i e ld a direct compar ison wi th the experimental spectrum. The finite size o f the electron beam is not taken into account, and only trajectories parallel to the pore axis are considered. The spatial dependence o f the energy losses is described in terms o f the impact parameter s, w h i c h is the distance between the electron beam and the centre o f the pore (Figure 6.6). The effect o f finite beam convergence angle i n S T E M mode leads to a substantial spread o f the beam w h e n thick membranes are considered - a typica l value o f a = 10 mrad leads to a total spread o f 20 n m over 1 urn. 144 A B F i g u r e 6.6 Compar i son o f (a) experimental electron beam wi th convergence angle a and (b) theoretical mode l . 6.4.1 M o d e l l i n g o f 197 k e V D a t a The energy loss probabil i ty distr ibution for relativistic electrons t ravel l ing paraxial ly to an infinite cy l indr ica l hole in a dielectric m e d i u m was first described by Zabala et al}1 Th i s mode l a l lows the calculation o f the losses associated wi th a single hole (radius a = 29 nm) in bu lk a lumina (Figure 6.7, M o d e l A ) . N e a r the pore w a l l (s = 27 nm) , surface plasmons between 10 and 25 e V dominate the loss function. Th is is in reasonable agreement w i th the experimentally observed spectrum. Nea r the pore centre (s - 0 nm) , the loss probabil i ty is a monotonica l ly decaying curve, fa l l ing to zero at 30 e V . T h i s represents losses due to the Cherenkov effect, as no surface plasmons are excited at this distance. Th is s imple mode l thus fails to describe the experimental spectrum near the pore centre: the asymmetr ical peak at 7 e V observed for porous a lumina differs substantially from the monotonica l ly decaying curve. The observed difference suggests that the Cherenkov losses are cut o f f at 7 e V . Th i s discrepancy was assumed to arise f rom the effect o f neighbouring pores, w h i c h alters the 145 dielectric function on a local level . A l t h o u g h these pores lie beyond the range o f the evanescent field o f the electron beam, they alter the m e d i u m through w h i c h the Cherenkov radiation propagates and directly affect the loss spectrum. The effect o f the ne ighbour ing pores can be model led in a number o f ways. The simplest approach is to consider cy l indr ica l shells o f a lumina w i t h different outer rad i i , as shown in Figure 6.7 ( M o d e l B ) . Th i s mode l for the porous a lumina introduces two peaks into the distr ibution in the region o f interest. Changes in the shell thickness shift the posi t ion o f the dominant loss peak. Th i s gives an indicat ion o f w h y the Cherenkov losses are cut o f f at 7 e V : only the Cherenkov radiation w h i c h can be transmitted by the membrane gives rise to associated electron energy losses. The Cherenkov radiation is confined wi th in the cy l ind r i ca l shell , due to total internal reflection at the v a c u u m interfaces (the condi t ion for total internal reflection is guaranteed by the Cherenkov c o n d i t i o n ) . 1 6 The cy l indr ica l shell geometry leads to a quantization rule for the transmitted radiat ion, cutt ing out m u c h o f the low-energy Cherenkov radiation and the losses w h i c h w o u l d have been associated w i t h i t . 1 6 The basic physics behind the observed loss peak is therefore fairly clearly represented by this s imple shell mode l . M o r e refined representations o f the dielectric properties o f porous a lumina were also investigated, by mode l l ing the nearest neighbour ing pores directly, either w i th s ix or twelve cy l ind r i ca l nearest neighbours (1+6 or 1+12 pores). The result ing loss distributions for electrons w i t h axia l paths (s = 0 nm) and near w a l l paths (s = 20 nm) are shown i n Figure 6.8. The mode l w i t h 1+6 pores clearly provides a peak at 7 e V . T h e 1+12 mode l , w h i c h w o u l d be expected to be more accurate than the 1+6 mode l , deviates substantially by presenting a second peak at ~8 e V . H o w e v e r , the subtle differences between these models w o u l d not be discernible experimentally, due to the l imi ted energy resolution (~0.9 e V ) o f E E L S in a T E M . Losses be low 7 e V are also present in these models and this is discussed further be low. 146 7x10 CO 5 3 4 N T3 C L T 5 3 2 -Q C C -Q O 1 0 0 1 1 1 1 1 1 1 1 1 : " C) ) f \ —r—i—|—i—i—i—r—j—i—i i i MODEL A : '• \ ; N . - f / • \ / \ -a=29 nm / / \s=27 nm : X^^s=23 nm (m=0) \s=23 nm \ ^ : s=0 n n r i \ ^ A A ^ ^ ~ ~ ^ - - - - . • " l~ T r - j T - 1 - I rf- h - i — i — 10 15 20 25 30 co(eV) 35 4x10 3.5P 3 3|_b=94nm N § 2.5 t C L T 3 I 1-5 F-03 -Q H i 0.5 LU 0. - i 1 r- - l — T I 1 r b=127 nm: § 1 f. (MODEL A ) \ i Q. o Hi i i | — i i i i MODEL B b=61 nm a=29 nm s=0 nm 10 o(eV) 15 20 F i g u r e 6.7 Theoret ical loss distr ibution for 200 k e V electrons t ravel l ing ( M o d e l A ) d o w n a single pore i n a lumina , and ( M o d e l B ) d o w n a cy l indr ica l pore o f outer rad i i 6 1 , 94 and 127 n m . The inner pore radius is 29 n m . Figures courtesy o f N . Zabala , F . J . G a r c i a de A b a j o and A . R ivacoba . 147 F i g u r e 6.8 Theoret ical loss distr ibution for 200 k e V electrons t ravel l ing d o w n a cy l indr ica l hole w i t h 0, 6 and 12 neighbour ing pores ( M o d e l C ) , showing the loss probabil i ty for (a) ax ia l (s = 0 nm) and (b) near-wall (s = 27 nm) trajectories. The pore radius is 29 n m and the spacing is 90 n m . Figures courtesy o f N . Zabala , F . J . G a r c i a de A b a j o and A . R ivacoba . 148 6.4.2 M o d e l l i n g o f 117 k e V D a t a A t the lower pr imary beam energy, the electron veloci ty has decreased to v i e « 0.6. The Cherenkov condi t ion is thus Si(co) > 2.8, w h i c h is also satisfied up to ~11 e V for a lumina (Figure 6.2). The energy loss probabi l i ty does depend direct ly on v i e , and the theoretical spectra reflect this (Figure 6.9). A t s = 0 n m , the loss peak shifts to higher energy by - 1 . 5 e V , i n qualitative agreement w i th the experimental result. 6.4.3 C o m p a r i s o n w i t h E x p e r i m e n t The loss probabi l i ty o f the 1+6 cyl inder mode l was convoluted w i t h the experimental zero-loss peak to y i e ld a direct compar ison w i t h experiment at three representative impact parameters. (Figure 6.10). The quantitative agreement at s = 22 n m and s = 28 n m is quite good . A t s = 0 n m , the magnitude o f the Cherenkov peak is overestimated by a factor o f 2 at 200 k e V ; at 120 k e V the agreement is improved . Th i s indicates that some effects may st i l l not be fully understood, either experimental ly or theoretically. One possibi l i ty is the effect o f finite beam convergence angle in S T E M mode, w h i c h leads to a substantial spread o f the beam when thick membranes are considered. H o w e v e r , a compar i son o f the spatial dis t r ibut ion o f losses as revealed by E F T E M ( a small) and S T E M / E E L S indicates that the difference is not very pronounced. Overa l l , the agreement between theory and experiment is satisfactory, and shows that the or ig in o f the loss peaks is w e l l represented by the models presented here. 149 3 CO N T3 CL C\J > -Q CD -Q O 0.5 0.4 0.3 0.2 0.1 U 0 0 200 KeV (p=0.7) •; 120 KeV (p-0.6) 10 co(eV) MODEL B a=29 nm b=94 nm s=0 nm 15 20 F i g u r e 6.9 Effect o f electron veloci ty on the loss probabil i ty function illustrated us ing M o d e l B . (|3 = vie in this context). Figure courtesy o f N . Zabala , F . J . G a r c i a de A b a j o and A . R ivacoba . Energy loss (eV) F i g u r e 6.10 Compar i son between theoretical and experimental spectra at different impact parameters s: (a) 0 n m , (b) 22 n m , (c) 28 n m , us ing the 1+6 cyl inder mode l (radius 29 n m , spacing 90 nm) . Figure courtesy o f F . J . G a r c i a de A b a j o . 150 6.5 Effect of Pore Diameter on the Cherenkov Peak The cy l indr ica l shell m o d e l indicated that the Cherenkov loss peak shifts w i t h the thickness o f the shell , w h i c h is equivalent to changing the pore diameter whi le keeping a f ixed pore spacing. T h i s effect was investigated in more detail by preparing ordered porous a lumina membranes through a two-step anodizat ion process (chapter 4) . The lattice spacing was fixed by the anodizat ion potential (105 n m at 40 V ) , wh i l e the pore diameter was altered by etching the samples w i t h phosphoric acid. In this manner, wel l-ordered domains o f about 800 n m diameter were obtained, w i th average pore diameters o f 62, 74 and 82 n m (Figure 6.11). The S E M image o f the cross-section showed that the pores were straight (Figure 6.12). The channel diameters d id not appear to be uni form throughout the thickness o f the membrane. T h i s m a y have been an artifact due to a slight tilt in the cleavage plane or it may a consequence o f the fabrication process (see the experimental details), w h i c h might have been avoided through further opt imizat ion . The sample w o u l d s t i l l be expected to show the loss spectrum o f the smaller pore diameter w i t h less intensity. The loss peak was recorded at the centre o f the pore for each ordered membrane (Figure 6.13(a)). The loss peak is shifted to higher energy as the pore diameter is increased, from 6.7 e V at 62 n m to 7.9 e V at 82 n m . Th i s is in qualitative agreement wi th the predict ion that a thicker a lumina shell w i l l lead to a lower energy cutoff, as more o f the Cherenkov radiation is able to propagate in the membrane. Cha ng i n g the p r imary beam energy to 117 k e V again shifted the loss peak to higher energies. The intensity o f the loss peak d iminishes wi th pore diameter due to the larger distance to the pore w a l l , result ing in a smaller interaction between the electron beam and the membrane (the membrane thickness difference also has an effect, but not as substantial, see Table 6.2). The loss intensity may also decrease due to the reduced w a l l thickness (Figure 6.7). 151 T a b l e 6.2 Cherenkov loss peak parameters for ordered samples. Pore Diameter Membrane Thickness Peak Posi t ion (nm) (um) ( eV) 62 2.1 ± 0 . 3 6.7 (197 k e V ) 59 2.1 ± 0 . 3 7.8 ( 1 1 7 k e V ) 74 2.1 ± 0 . 3 7.5 82 0.85 + 0.05 7.9 I • • % 4 • • • • • • • • « • • • • I ^ ^ ^ ^ ^ ^ ^ ^ ^ j ^ ^ ^ ^ ^ ^ ^ * ^ ^ ^ ^ , w . „ „ w - _ - „ . „ F i g u r e 6.11 S T E M dark f ield images showing geometry o f ordered porous a lumina membranes produced by a two-step anodizat ion at 40 V in 0.3 M oxal ic ac id : (a) 62 n m , (b) 74 n m , (c) 84 n m diameters; (d) lower magnif icat ion image showing size o f ordered domains. The scale bars are 100 n m . 152 F i g u r e 6.12 Cross-sect ion o f ordered porous a lumina membrane. No te the pore diameters appear larger at the bottom. The scale bar is 1 u m . M o r e careful measurement o f the zero-loss peak also a l lowed the examinat ion o f losses d o w n to 2 e V . Th i s revealed that there was no sharp cut-off at the l o w energy side o f the Cherenkov loss peak (Figure 6.13(b)). Instead, the losses extend at least to 2 e V at a roughly constant value. The zero-loss peak subtraction was not sufficiently good between -2 and 2 e V to extract any further information. The precis ion o f the result is affected by noise in the spectra. The uncertainty in the calculated spectrum may be determined from the count ing statistics: the C C D detector is susceptible to shot noise, w h i c h is a Poisson process wi th an uncertainty o f ^ f N for N c o u n t s . 1 8 The uncertainty in the dark current, w h i c h is subtracted from each collected spectrum, may be estimated from the noise in the spectrum o f the zero-loss peak, w e l l away from the peak itself. These calculations suggested a noise level o f - 2 0 0 counts at 1 e V and - 1 0 0 0 counts at 0 e V , w h i c h w o u l d not account for the large fluctuations observed in the calculated spectrum. Changes in experimental condit ions (e.g., stray fields) may then be the cause o f the 1 5 3 imperfect subtraction o f the zero-loss peak, and the range o f val id i ty o f the subtraction can only be estimated from the quality o f fit at the left-hand side o f the zero-loss peak. Ove ra l l , the observed beam energy and pore radius dependence agrees w i t h the predict ions o f the shell model . It does not clarify the extent o f the long range interaction wi th the pore lattice, for w h i c h further measurements at different energies and on different samples w o u l d be necessary. A s this pore lattice effectively forms a 2 - D photonic crystal , these results introduce the possibi l i ty that E E L S cou ld be used to extract useful information from 2 - D photonic nanostructures through the Cherenkov e f fec t . 1 9 The possibi l i ty o f observing novel Cherenkov radiation effects in photonic crystals has also been the subject o f theoretical d i s c u s s i o n . 2 0 154 6000 h 4000 L 3 CO • | 2000 0 0 2 4 6 8 10 12 14 16 18 20 Energy (eV) 6000 4000 1 — T i — • — r B A - | 1 1 1 1 1 1 1 1 1 " 0 2 4 6 8 10 Energy (eV) F i g u r e 6.13 (a) Cherenkov peak shift for a f ixed lattice spacing (105 nm) w i t h different diameters: ( • ) 62 n m , ( • ) 74 n m , and ( • ) 82 n m at 197 k e V and (o) 62 n m at 117 k e V ; (b) ( • ) losses d o w n to 2 e V revealed by subtraction o f the ( A ) zero-loss peak (diameter 62 n m , 197 k e V ) . The losses between - 2 and 2 e V are not meaningful . 155 6.6 C o n c l u s i o n The 7 e V loss peak observed by E E L S i n the centre o f the pores o f porous a lumina films was found to be most l ike ly due to the Cherenkov effect. The clearest evidence is the shift in the loss peak to higher energy wi th decreasing accelerating voltage. A number o f models were investigated, and it was found that the local loss spectrum is affected by the sample structure on a m u c h larger scale, due to the radiative nature o f the Cherenkov losses. The theoretical mode l o f the loss peak at different p r imary beam energies matched the experimental data reasonably w e l l . Cer ta in aspects o f the material and geometry were not model led , and could account for the smal l discrepancies observed: the anion contamination o f the porous a lumina , and the exact geometry o f the pores neighbour ing the pore be ing investigated. M o r e direct evidence for the Cherenkov effect cou ld be obtained by detecting the generated rad ia t ion 8 but the systems in existence today detect luminescence be low 6 e V only. The nature o f these losses cou ld be further studied by fabricating and s tudying single pores, pore clusters and pore lattices w i th exactly defined geometries, such that a closer match to the theoretical models cou ld be obtained. However , the aspect ratios seen i n porous a lumina are not readily achieved us ing other fabrication approaches. Further measurements on an instrument equipped w i t h an electron monochromator (zero-loss peak F W H M ~0.1 e V ) w o u l d be needed to establish the experimental loss spectrum wi th sufficient detail to choose the correct theoretical representation o f the pore lattice. F ina l ly , it should be pointed out that the identification o f this phenomenon was most ly due to several serendipitous choices o f experimental factors: the measurement geometry, w h i c h in truth was not really useful for chemica l analysis, a th ick membrane w h i c h m a x i m i z e d the loss probabi l i ty , and a pore diameter w h i c h was sufficiently large to avoid surface p lasmon losses near the centre. 156 E x p e r i m e n t a l De t a i l s Porous a lumina thin films were prepared by anodiz ing 0.2 and 1.0 u m thick A l films evaporated onto n-type s i l i con wafers by e-beam evapo ra t i on 2 1 . The anodizat ion was carried out in 0.3 M oxal ic ac id at 2 0 ° C wi th an applied potential o f 40.0 V , us ing a glass cel l w i th an O -r ing seal to the sample. U p o n comple t ion o f the anodizat ion, the porous a lumina film had detached from the substrate, but was st i l l attached to the surrounding a l u m i n i u m film. The barrier layer was removed by etching in 5 wt. % phosphoric ac id for 40 m i n . The film was r insed w i t h dis t i l led water, fo l lowed by ethanol and then dr ied. U s i n g the same samples (chapter 5) , electron energy-loss spectroscopy was carried out on a Tecna i F 2 0 ( M e d i c a l Imaging Faci l i ty , Univers i ty o f Calgary) T E M w i t h a Schottky field emiss ion electron gun operated at nomina l ly at 120 and 200 k V . The pr imary beam energy used for spectroscopy was 117 and 197 k e V . Measurements were made w i t h a Gatan Imaging Fil ter ( G I F ) us ing a dispersion o f 0.05 eV/channe l on a 1024-channel detector, cover ing -5 to 45 e V losses. The col lect ion semi-angle was between 0.34 and 6.0 mrad; the convergence angle was not determined. The probe size was estimated to be < 1 n m w h i l e the energy resolution was 0.9 e V ( F W H M o f zero-loss peak). Measurements on ordered samples were carried out on Tecna i F 2 0 (Nano- imag ing Faci l i ty , S i m o n Fraser Univers i ty ) equipped w i t h a G I F . The dispersion was set to 0.10 eV/channel . The F W H M o f the electron probe was measured to be 0.6 n m . The energy resolution was 0.90 e V at 197 k e V and 0.87 e V at 117 k e V . Ordered samples were prepared by anodiz ing 99 .99+% a l u m i n u m fo i l ( A l d r i c h ) for 2-3 hours at 40 V in 0.3 M oxal ic acid for the first step, w i th st irring. The anodized layer was stripped overnight by immers ing in 5.8 wt. % H3PO4 + 1.5 wt. % Q O 3 at 50 to 6 0 ° C overnight. A second anodization step was carried out for 10-20 m i n under the same condit ions. The 157 resulting f i lm was washed wi th water and dried, then glued to a glass sl ide (membrane down) w i t h C rys t a lBond , a thermopolymer adhesive. The a l u m i n u m susbtrate was removed by etching w i t h a saturated H g C ^ solution, leaving the a lumina membrane attached to the glass slide w i t h the barrier layer facing up. Th i s layer was etched for 120 m i n in 5% H3PO4 w i th stirring. After wash ing wi th water and dry ing , the adhesive was removed by soaking in acetone, and the membrane inspected by T E M . The pore wal ls cou ld be further etched to increase the diameter as desired. F i l m thicknesses were determined by S E M ; in some cases there was substantial variat ion over the whole film, leading to some uncertainty in the thickness. References 1. Jackson, J . D . Classical Electrodynamics; W i l e y : N e w Y o r k , 1999. 2. Cherenkov , P . A . Dokl. Akad. NaukSSSR 1934, 2 , 451-453 . 3. T a m m , I. E . ; Frank, I. M . Dokl. Akad. NaukSSSR 1937 ,14 , 107. 4. G a r c i a de A b a j o , F . J . ; B l a n c o , L . A . Phys. Rev. B 2003, 67, N o . 125108. 5. Pa l ik , E . D . Handbook of Optical Constants of Solids; Pa l i k , E . D . , E d . ; A c a d e m i c Press: Or lando , 1985. 6. Danie l s , J . ; Festenberg, C . v . ; Raether, H . ; Zeppenfeld, K . Springer Tracts Mod. Phys. 1970, 5 4 , 78-135. 7. C h e n , C . H . ; S i l cox , J . ; V incen t , R . Phys. Rev. B. 1975 ,12 , 64-71 . 8. Y a m a m o t o , N . ; A r a y a , K . ; Toda , A . ; Sug iyama, Ff. Surf. Interface Anal. 2001, 31, 79. 9. A b e , H . ; Kurata , H . ; H o j o u , K . J. Phys. Soc. Jpn. 2000, 69, 1553. 10. W i l l i a m s , P . ; Levy -C lemen t , C ; A l b u - Y a r o n , A . ; B r u n , N . ; C o l l i e x , C . J. Porous Mater. 2000, 7, 159. 11. T y a p k i n , A . A . Physics of Particles and Nuclei 2001, 32, 506. 158 12. G i n z b u r g , V . L . Physics-Uspekhi 2002, 45, 341 . 13. Pla tonov, K . Y . ; F le i shman , G . D . Physics-Uspekhi 2002, 45, 235 . 14. Eger ton, R . F . Electron energy-loss spectroscopy in the electron microscope; 2nd ed.; P l e n u m Press: N e w Y o r k , 1996. 15. Zabala , N . ; R ivacoba , A . ; G a r c i a de A b a j o , F . J . ; Pattantyus, A . Surf. Sci. 2003, 532-535, 461-467. 16. Zabala , N . ; Pat tantyus-Abraham, A . G . ; G a r c i a de A b a j o , F . J . ; R ivacoba , A . ; W o l f , M . O . Phys. Rev. B , in press. 17. Zabala , N . ; R ivacoba , A . ; Echenique , P . M . Surf. Sci. 1989, 209, 465 . 18. Reed , B . W . ; Sar ikaya, M . Ultramicroscopy 2002, 93, 25 . 19. G a r c i a de A b a j o , F . J . ; Pat tantyus-Abraham, A . G . ; Zabala , N . ; R ivacoba , A . ; W o l f , M . O . ; Echenique , P . M . Phys. Rev. Lett. 2003, 91, N o . 143902. 20 . L u o , C ; Ibanescu, M . ; Johnson, S. G . ; Joannopolous, J . D . Science 2003, 299, 368. 2 1 . Crouse , D . ; L o , Y . H . ; M i l l e r , A . E . ; Crouse , M . Appl. Phys. Lett. 2000, 76, 49 -51 . 159 CHAPTER 7 Polymer Guest Incorporation Porous a lumina fi lms were shown to be a very close approximat ion to an ideal host material i n chapter 4: they have 4 n m pores running perpendicular to the f i l m normal , and the f i l m thickness can be easily controlled. The introduction o f the conjugated po lymer guest into the porous a lumina host must be considered w i t h i n the constraints imposed by the desired light-emit t ing device structure and the properties o f the po lymer guest. The introduction o f the po lymer guest into the porous th in f i l m host can be accompl ished first through in situ synthesis. In this direct ion, some ini t ia l w o r k on the surface-graft polymer iza t ion o f po ly( l ,4 -phenylene vinylene) ( P P V ) on a s i l i con surface is presented in this chapter. Secondly , bulk-synthesized po lymer may be introduced into the host through a variety o f methods. A novel approach, centrifugal loading, is investigated and used to prepare samples for in-depth characterization, i n particular high-resolution chemica l analysis by electron energy-loss spectroscopy ( E E L S ) and energy-filtered t ransmission electron microscopy ( E F T E M ) . 7.1 Internal Polymer Synthesis In situ synthesis o f the po lymer guest is attractive due to the relatively rapid diffusion o f monomers into the host material . M a n y conjugated polymers , such as P P V , polythiophene and polyacetylene, can only be introduced into the host in this manner, due to their insoluble and infusible nature. The groups o f M a r t i n and B e i n separately carried out p ioneer ing w o r k on 160 conjugated po lymer synthesis in porous a l u m i n a 1 and mesoporous s i l i c a , 2 respectively. Th i s approach was also exploited in the creation o f the P P V / M C M - 4 1 composite discussed in chapter 3. Recent reports describe the formation o f po ly ( l , 4 -pheny lene ) 3 and poly(2 ,5-d ie thoxyphenylene) 4 by oxidat ive coup l ing in porous a lumina membranes. C h e m i c a l vapour deposit ion o f a po lymer precursor has been shown to introduce P P V into porous a lumina as w e l l . 5 The two difficulties that face this approach is control over the degree o f po lymer iza t ion and confinement o f the po lymer iza t ion to the pore vo lume. The degree o f in situ po lymer iza t ion may differ significantly from the bulk-synthesized polymer . T h i s may be especially relevant in pores w h i c h approach molecular d imensions (< 2 nm) . Th i s difference can make meaningful property comparisons more difficult . Nevertheless, materials prepared in this manner may have useful applications in devices. Po lymer iza t ion confinement may be achieved by first loading one o f the reactants into the porous host. T h i s is exempl i f ied in the synthesis o f conf ined polyani l ine , where anil ine is first diffused into the pores . 2 The synthesis o f P P V / M C M - 4 1 discussed in chapter 3 rel ied on the abi l i ty o f the M C M - 4 1 surface to act as an ini t iat ing base when monomer was subsequently introduced. Howeve r , i f the pre-loaded species can diffuse out o f the pores, bu lk polymeriza t ion cannot be prevented entirely. A better approach is to initiate the polymer iza t ion from a substrate support ing the porous film on one side, w h i c h is k n o w n as surface-graft polymeriza t ion. T h i s last approach appears very p romis ing and some pre l iminary w o r k has been carried out in this direct ion. 7.1.1 Surface-Graft Polymerization of Conjugated Polymers Surface-graft po lymer iza t ion is a route to po lymer confinement by virtue o f po lymer growth being local ized to the end o f chains initiated from the surface. The use o f a suitable 161 surface-bound initiator is key to this route. I f the initiator is bound to a material suitable for use as an electrode, it can be guaranteed that each po lymer chain is attached to the electrode interface in a chemica l ly wel l-def ined manner. The presence o f a porous host on top o f this electrode w o u l d then y ie ld a device w i t h surface-grafted po lymer chains separated by the wal l s o f host. In this manner, on ly continuous po lymer chains w o u l d exist i n each channel o f the host. The p rob lem o f bulk polymer iza t ion is e l iminated entirely. L i v i n g polymer iza t ion is the method o f choice for preparing surface-grafted polymers . Th i s usual ly involves an anionic or radical po lymer iza t ion mechanism. There are two established l i v i n g polymer iza t ion routes to conjugated polymers : r ing-opening metathesis po lymer iza t ion ( R O M P ) 6 ' 7 and e lec t ropolymerizat ion. 8 R O M P may be used to prepare very w e l l defined polymers , and it has been demonstrated to create surface-grafted polymers from a surface i n i t i a t o r . 9 ' 1 0 Howeve r , the preparation o f the required monomer , catalyst and surface initiator were beyond the scope o f the present work . The use o f electropolymerizat ion to prepare surface-grafted conjugated polymers has been demonstrated for p o l y a n i l i n e 1 1 ' 1 2 and po ly th iophene . 1 3 In practice, e lectropolymerizat ion has the disadvantage o f producing chains wi th many chemica l defects and was not explored further for this reason. One further possibi l i ty is the G i l c h route: it has been reported that it proceeds by an anionic m e c h a n i s m 1 4 and could therefore also be initiated from a surface (Figure 7.1). Pre l iminary efforts i n this direct ion d id not y ie ld any grafted polymer , presumably because the surface ini t iat ion group was not i n sufficient concentration to compete wi th the. bulk po lymer iza t ion process. It is also possible that the surface initiator group reacts w i th ne ighbour ing initiators to deactivate the surface. 162 Figure 7.1 Surface-initiated anionic po lymer iza t ion o f M E H - P P V . Step (or condensation) polymer iza t ion was explored as an alternative route to the preparation o f surface-grafted polymers . Condensat ion reactions have been used to make w e l l -defined conjugated oligo(phenylene vinylene)s in h igh y i e l d . 8 Poly(phenylene ethynylene) derivatives o f substantial length have been prepared in this manner through repeated coup l ing o f o l i g o m e r s . 1 5 This approach was conceived to be applicable to the preparation o f P P V chains bound to a conduct ive substrate. A n example o f a synthetic route is shown in F igure 7.2, based on the Wadswor th -Horne r -Emmons ( W H E ) synthesis o f aromatic olefins from aldehydes and phosphonate esters. T h i s reaction is k n o w n to produce trans conjugated olefins in h igh y i e l d and has been applied to the synthesis o f P P V d e r i v a t i v e s . 1 6 " 1 9 It has also been demonstrated on a sol id s u p p o r t . 2 0 " 2 2 The hydrogen-terminated s i l icon surface is a convenient substrate for evaluating this route due to its we l l k n o w n c h e m i s t r y . 2 3 ' 2 4 Porous s i l i con substrates 2 5 are particularly useful as their h igh surface area a l lows the use o f t ransmission F T - I R for characterization. The most relevant reactions on this surface have been developed by the groups o f S a i l o r 2 6 " 2 8 , B u r i a k 2 9 " 3 2 and W a y n e r . 3 3 " 3 5 A m o n g these, cathodic electrografting has been used to graft functionalized alkynes and alkenes to the s i l icon surface (Figure 7 . 3 ) . 3 0 O f particular interest was the grafting 163 (EtO) 2 OP PO(OEt) 2 f-BuOK/THF j PO(OEt) 2 f-BuOK/THF Figure 7.2 Step polymer iza t ion o f surface-grafted P P V by the W H E reaction. o f phenylacetylene and 4-bromophenylacetylene through the acetylene carbon, w h i c h suggested that a fully conjugated bond to the s i l i con surface could be formed. M o s t importantly, the presence o f the aromatic halogen al lows the use o f w e l l k n o w n chemistry to further derivatize the surface: the brominated phenyl r ing could be used to form grafted oligo(phenylene) chains through the S u z u k i c o u p l i n g . 8 A s P P V derivatives were o f interest in this work , a grafted aromatic aldehyde was pursued to create surface-bound o l igo( l ,4 -phenylene vinylene) chains through the W H E reaction discussed above. The actual encapsulation o f po lymer by such surface chemistry w o u l d be accompl ished by H- •R H H H H -Si—Si—Si—Si-R H I. '. H -Si—Si—Si—Si--9 raA c m -2 Figure 7.3 Porous s i l i con derivatization by cathodic electrografting ( R = phenyl , 4-bromophenyl , e t c . ) . 3 0 164 the preparation o f porous a lumina f i lms on a s i l icon substra te . 3 6 The key step i n the process w o u l d be the successful preparation o f the hydrogen-terminated s i l i con surface, w h i c h requires removal o f the native s i l icon d ioxide film. Th i s is usually accompl ished wi th aqueous hydrogen fluoride or a m m o n i u m fluoride. Ini t ial tests indicated that porous a lumina was stable to the latter solution. React ive ion etching, c o m m o n l y used in the semiconductor industry, can also be used to remove the native oxide layer in the presence o f porous a l u m i n a 3 6 but this may not w o r k w e l l w i th a l l pore sizes. 7.1.2 Derivatization of Porous Silicon Surfaces The functionalization o f porous s i l i con substrates by cathodic electrografting was studied. 4-ethynylbenzaldehyde has the two desired functional groups: an alkyne for b ind ing to s i l icon and an aldehyde for further chemistry by the W H E reaction. A s the grafting process is reductive at the s i l i con surface, the aldehyde group was protected by convers ion to the cyc l i c acetal (Figure 7.4). A porous s i l i con substrate was then derivat ized by cathodic electrografting o f this protected aldehyde (Figure 7.5), us ing te t rabutylammonium hexafluorophosphate in methylene chloride as the electrolyte. The chemica l changes were fo l lowed by F T - I R spectroscopy. N e w C -H stretches and aromatic bands appeared after electrografting (Figure 7.6(a)); the cyc l i c acetal r\ ecu reflux Figure 7.4 Protect ion o f aldehyde i n 4-ethynylbenzaldehyde. 165 band at 943 cm" 1 cou ld also be dist inguished. The alkyne stretch was expected to appear near 2100 cm" 1 but cou ld not be conf i rmed due to the over lapping H - S i stretches. The strong S i - 0 f ramework stretch at 1100 cm" 1 indicated the presence o f substantial surface oxide . F o l l o w i n g deprotection o f the aldehyde, a strong C = 0 stretch peak appeared at 1700 cm" 1 along wi th a weaker H - C = 0 peak at 2733 cm" 1 (Figure 7.6(b)). The aromatic bands were also enhanced, w h i l e the cyc l i c acetal band disappeared. The aliphatic C - H stretches remain ing at 2 9 6 9 and 2930 cm" 1 may be due to c h e m i - or physisorbed e lec t ro ly te . 3 0 A l t h o u g h the structure o f the surface-grafted species could not be fully established, it was clear that a surface-bound aldehyde was present. o. .0 H H H H S i — S i — S i — S i — « - B u 4 N P F 6 / C H 2 C l 2 4 m A cm" 2 H I. H H S i — S i - S i — S i — O II (EtO) 2P. dilute HCl(aq) H H H H — S i — S i — S i — S i — I I H H i i — S i - S i - S i — F i g u r e 7.5 Synthetic route to conjugated d imer on porous s i l i con surface. 166 7.1.3 Wadsworth-Horner-Emmons Reaction on Silicon Surface The terminal aldehyde on the s i l i con surface was reacted w i t h 1,4-xylylenebis(diethylphosphonate) . The F T - I R spectrum o f the result ing surface (Figure 7.6(c)) showed that the C = 0 stretch at 1700 cm" 1 was completely removed and a new shoulder assigned to a P = 0 stretch had appeared at 1260 cm" 1 . The aliphatic C - H stretches around 2969 and 2933 cm" 1 were in agreement w i t h the C - H stretches observed in the phosponate ester, al though they may st i l l be due to the electrolyte. Th i s surface also exhibi ted blue photoluminescence, as shown i n Figure 7.7. The emiss ion m a x i m u m was observed at 400 n m . The excitat ion spectrum, detected at 400 n m , had a m a x i m u m at 322 n m . T h i s is in agreement wi th the literature value o f 4 -e thynyls t i lbene 3 7 and w o u l d suggest that the interaction o f the n electron system w i t h the s i l i con substrate is weak. These ini t ial results strongly suggest that the desired surface-grafted species was obtained. Howeve r , more detailed analysis w i t h mode l compounds w o u l d be required to fully conf i rm this. W h i l e this chemistry was carried out on porous s i l i con substrates, it should be readily applicable to hydrogen-terminated s i l i con wafers as we l l . Short ol igomers grafted to s i l i con w o u l d then represent a very w e l l defined junc t ion between an organic and an inorganic semiconductor . The rate o f po lymer growth can be increased by synthesizing appropriate dimers or tr imers in bulk , then us ing them to g row the surface-grafted p o l y m e r . 1 5 167 I I I I I I I " ~ l — I 1 1 1 — 250 300 350 400 450 500 550 Wavelength (nm) F i g u r e 7.7 Photoluminescence o f derivat ized porous s i l i con : ( A ) excitation spectrum (detected at 400 nm) and ( • ) emiss ion spectrum (excited at 300 nm) o f W H E reaction product; ( • ) emiss ion o f deprotected aldehyde. 168 7.2 E x t e r n a l P o l y m e r Synthes i s Soluble conjugated polymers and precursor polymers are readily prepared through bulk synthesis. Such polymers may then be inserted into the pores o f the host material in a number o f ways. Th i s process a l lows the use o f k n o w n bulk synthetic routes to control the properties o f the polymer , and direct characterization o f the polymer properties before encapsulation. Pur i f ica t ion could then remove possible undesirable by-products. Th i s is also the best approach for definite comparisons between the encapsulated and bulk polymer, as ambigui ty relating to differences in the chemica l structure is e l iminated. The central challenge lies in f inding a suitable d r i v i n g force for polymer insertion into the host material . O n e key example i n the literature i n the w o r k o f Tolbert et al, where the M E H -P P V polymer (Figure 7.8) was shown to insert i tself into a derivatized mesoporous s i l i ca material through diffusion and a d s o r p t i o n . 3 8 A second example is the use o f a pressure difference (vacuum or filtration l o a d i n g ) . 3 9 ' 4 0 A third nove l example, to date unexplored in the literature, is the use o f centrifugal forces to dr ive the po lymer into the host membrane. These latter two approaches are more useful w h e n the host pore size is larger than a few n m . O \ F i g u r e 7.8 C h e m i c a l structure o f poly[2-methoxy,5-(2 ' -e thylhexyloxy)- l ,4-phenylene vinylene) ( M E H - P P V ) . 169 7.2.1 Polymer Adsorption Loading Direc t adsorption is a very s imple method for incorporat ing po lymer into mesoporous hosts. The M E H - P P V / p o r o u s a lumina composite studied i n chapter 5 was prepared by this approach. The process is dr iven by the favourable interactions between the po lymer and the pore surface. In the case o f a lumina, both the electron-rich phenyl rings and the pendant ether oxygens can interact w i t h surface hydroxy g r o u p s . 4 1 ' 4 2 Howeve r , it is not clear i f such an approach can be used to introduce more than a monolayer o f polymer . It is most l ike ly that this process can only achieve h igh loading o f the host when the ratio o f pore surface area to pore vo lume is h igh , i.e., when the pore diameter is smal l . The pore density must be h igh as w e l l . These condit ions are met by mesoporous s i l i ca materials; on the other hand, porous a lumina can be made wi th smal l pore diameters only wi th a l ow pore density, due to the f ixed pore spacing at the anodizat ion potential for ordered pore growth (see chapter 4). Since h igh po lymer loading was desired for evaluation o f the h igh resolution characterization techniques, po lymer loading by adsorption was not investigated further. 7.2.2 Vacuum (Filtration) Loading This approach is very useful for in t roducing material into porous membranes, as has been reported recently for go ld n a n o p a r t i c l e s . 3 9 ' 4 0 The po lymer solution is d rawn through the porous membrane, and rapid solvent evaporation deposits po lymer into the pores. D u e to the large pressure difference exerted on the membrane, it must either be very th ick or supported in some form. It was found that po lymer could be readily introduced into th ick porous a lumina membranes (60 urn th ick A n o p o r e commerc ia l membranes w i t h 200 n m pore size) by s imp ly f i l ter ing a po lymer solution (Figure 7.9(a)). The substantial po lymer loading afforded by this 170 F i g u r e 7.9 (a) V a c u u m - d r i v e n po lymer infiltration into a porous membrane, (b) S E M image o f po lymer in 200 n m pores o f A n o p o r e membrane. The scale bar is 500 n m . approach a l lowed the polymer to be seen directly in the pores o f the membrane by S E M (Figure 7.9(b)). These thicker membranes could also be used as a support for vacuum loading po lymer into thinner membranes. Nevertheless, this approach was not pursued further o w i n g to its requirement o f a free-standing membrane and the relative fragility o f such membranes. 7.2.3 C e n t r i f u g a l L o a d i n g The centrifugal force may be used to drive po lymer into the pores o f the a lumina host. Centr ifugation is generally used as a separation technique for smal l particles in solution, but can also be used to create density gradients in po lymer so lu t i ons . 4 3 T w o loading processes may be possible, depending on the speed o f the centrifuge. Di rec t sedimentation o f the po lymer may be achieved at very h igh speeds (150 000 R P M ) , where the forces are on the order o f 9 x l 0 6 N . 4 3 Such centrifuges were available for use but cou ld not be readily modif ied to incorporate a vert ical ly oriented substrate. Pre l iminary experiments were carried out on a typical chemistry 171 Substrate holder Po lymer solution Solvent evaporation Porous a lumina Centr ifugal force F i g u r e 7.10 Centrifugal po lymer loading into a porous a lumina fdm. laboratory centrifuge, in w h i c h the speeds are on the order o f a few thousand R P M . It was realized that this speed may be sufficient i f the centrifugal force is appl ied wh i l e a l l o w i n g solvent evaporation, as depicted in Figure 7.10. Th i s second approach was investigated ini t ia l ly , us ing M E H - P P V dissolved i n tetrahydrofuran ( T H F ) , and was found to be effective. 7.3 P r e p a r a t i o n o f C e n t r i f u g e d Samples The experimental setup for po lymer loading is shown in Figure 7 .11: a s imple laboratory centrifuge has been modi f i ed wi th a holder for flat substrates. The holder consisted o f a threaded stainless steel barrel w h i c h could be sealed to the vert ical ly oriented substrate. The seal between the barrel o f the holder and the substrate was chosen to be a V i t o n O- r ing , as rubber O-r ings are sensitive to many c o m m o n solvents and Tef lon O-r ings were found not to provide reproducible seals. The solvent evaporation rate cou ld be control led to some extent by capping the end o f the barrel w i t h a rubber septum. A l t h o u g h porous a lumina membranes w i t h 4 n m pore diameter were available for study, membranes w i t h larger pore sizes ( typically - 2 0 nm) were used. These were more easily fabricated in large numbers for different studies, and the larger amount o f po lymer w h i c h could 172 F i g u r e 7.11 (a) Centrifuge rotor assembly wi th two substrate holders, (b) detail o f substrate holder from above and (c) from inside, showing O - r i n g seal. be introduced into these pores was important for the development o f the characterization techniques for composite samples i n general. The pore vo lume in a porous a lumina film can be calculated from the exposed substrate area (0.28 c m 2 ) and the approximate porosity (30%). For a 1 um-thick film, this amounts to - 8 n L . Comple te pore f i l l ing w o u l d then require a po lymer mass in the tens o f ug . For a po lymer solut ion o f 0.038 wt % M E H - P P V in T H F , the corresponding vo lume is on the order o f tens o f u L . Consistent results were obtained by using three consecutive centrifuging steps: (1) M E H - P P V solution, without capping the barrel, - 2 m i n (2) addit ional solvent only, w i t h a capped barrel, - 1 5 m i n (3) barrel cap removed, - 5 m i n The first step deposits a smal l amount o f po lymer onto the substrate, and the second step serves to dr ive it further into the pores and distribute it uni formly. The last step, w i t h the cap removed, ensures that the solvent is fully evaporated. 173 7.4 Characterization of Centrifuged Samples The M E H - P P V distr ibution i n the centrifuged samples was investigated us ing a number o f different techniques. V i s u a l l y , the samples showed a thicker r i ng o f po lymer deposited at the edge o f the exposed area; in some cases, smal l po lymer particles were v is ib le on the surface. The polymer distr ibution i n the centre o f the sample was usually even. 7.4.1 Scanning Electron Microscopy W i t h larger pore samples, the S E M may be used to determine the po lymer distr ibution. The backscattered electron detector was not useful at the magnificat ions o f interest (> 50 OOOx) and the scattering difference between the carbon-based po lymer and the a lumina host is not sufficient for useful contrast. The secondary-electron detector, w h i c h provides topographical information at h igh resolution, was used for this study. W h e n cross-sections o f the host material are observed by S E M , a partially fi l led pore may not stand out evidently. A completely f i l led pore w o u l d eliminate most o f the v is ib le structure in the porous f i l m . In many cases, a superficial po lymer layer could be readily dist inguished (Figure 7.12). Part o f this po lymer layer could also be seen penetrating into the pores, w h i c h indicated some pore f i l l ing . In the case o f substantial pore f i l l ing , the polymer acts as a repl ica o f the pore space. Th i s can be v isua l ized by etching away the a l u m i n u m substrate and the a lumina host, as shown i n Figure 7.13. The result o f this process is shown in Figure 7.14: elongated po lymer tubules are v i s ib le , indicat ing substantial penetration o f the po lymer into the porous a lumina host. W h e n thicker polymer- loaded samples were investigated in this fashion, the po lymer tubules were observed to have collapsed into bundles (Figure 7.14(b)). 174 F i g u r e 7.12 S E M image o f cross-sections o f centrifuged samples, showing (a) po lymer overlayer and (b) some polymer penetration into pores. The scale bars are 200 and 100 n m . X P S on this surface \ E m b e d HgCl2(sat) Conduct ive epoxy S E M on this surface \ / 5% H 3 P0 4 (aq) F i g u r e 7.13 A n a l y s i s o f po lymer penetration into porous a lumina by X P S and S E M . 175 F i g u r e 7.14 S E M images o f centrifuged samples, observed from bottom w i t h the host etched away. The scale bars are (a) 200 n m and (b) 1.0 u m (200 n m i n the inset). 176 7.4.2 X - r a y P h o t o e l e c t r o n Spec t roscopy For the purpose o f device preparation, it is necessary to create composi te films in w h i c h the po lymer penetrates the depth o f the membrane. Otherwise , electrical contact to the po lymer is not possible. Th i s cannot be determined reliably from the S E M images, but may be obtained by X - r a y photoelectron spectroscopy ( X P S ) . Th i s is a very surface-sensitive technique, because o f the l imi ted mean free path o f the generated photoelectrons. The sampl ing depth is effectively on the order o f 1 n m . B y analysing the bottom interface o f the composite , it may be used to establish whether po lymer has penetrated to the bottom o f the pores (Figure 7.13). The presence o f a carbon signal on this surface, after argon-ion sputtering to remove any adsorbed atmospheric hydrocarbons, w o u l d indicate complete po lymer penetration to the bottoms o f pores. The argon sputtering also served to thin the barrier layer at the bottom o f the pores. The result o f this experiment is shown i n Figure 7.15. The a l u m i n u m 2p (74.7 e V ) and oxygen Is (532 e V ) signals are present in both the empty and centrifuged samples. The i r O x y g e n 1 s A l u m i n u m 2p Ca rbon 1 s " i — • — i — 1 — i — 1 I—i ' 1 1 1—I n — • — i — • — i 1 r I I • I • I • I I I I I I , I I l _ l 1 1 . 1 . 1—1 5 4 0 5 3 5 5 3 0 5 2 5 8 0 7 5 7 0 2 9 0 2 8 5 2 8 0 2 7 5 B i n d i n g Energy ( eV) B i n d i n g Energy ( eV) B i n d i n g Energy ( eV) F i g u r e 7.15 X P S results from bottom o f (o) empty porous a lumina and ( • ) centrifuged sample. 177 posit ions are very close to the expected energies for a l u m i n u m oxide (74.3 and 531.1 e V , r e s p e c t i v e l y ) 4 4 ' 4 5 However , the carbon Is signal is only present in the centrifuged sample. The m a i n carbon peak, at 280.4 e V , was substantially be low the normal range (from 288 e V for oxides to 281 e V for carbides) and this was assigned to sputtering d a m a g e . 4 6 Quant if icat ion o f the amount o f po lymer present at the bottom o f the pore was not attempted. The results clearly indicate that this loading approach is effective at p roduc ing fully penetrated composite materials. 7.4.3 Transmission Electron Microscopy T h i n sections for T E M investigation were prepared by ul t ramicrotomy o f epoxy-embedded samples. P r io r to embedding, a th in layer o f A u / P d was deposited by D C sputtering on top o f the samples. For the empty host, this prevented epoxy infusion into the pores. Fo r the centrifuged samples, this layer prevented dissolut ion o f M E H - P P V into the epoxy. The thickness o f the A u / P d layer was between 20 and 50 n m . Sections were cut to thicknesses between 20 and 80 n m . A s the pore spacing is 40 n m for porous a lumina prepared at 15 V , section thicknesses on that same order w i l l contain not more than one pore. Substantially thinner sections were more difficult to prepare without induc ing substantial deformation to the porous a lumina film. Nevertheless, good quality sections as th in as 20 n m were obtained for some samples. B e a m damage to the porous a lumina host was evident i n the appearance o f c i rcular defects i n the a lumina . The stability could be improved substantially by the use o f a diffuse electron beam. U n d e r such condit ions, the sensit ivity o f a charge-coupled device ( C C D ) detector was necessary to capture images w i t h i n a reasonable amount o f t ime (< 10 s) to avo id any sample drift. 178 A B Po lymer A l u m i n u m Porous a lumina F i g u r e 7.16 T E M images o f thin sections o f (a) empty porous a lumina host and (b) centrifuged sample (beam damage to the host is apparent). The scale bars are 100 n m , and the accelerating voltage was 80 k V . The cross-section o f the empty porous a lumina host is shown in Figure 7.16(a). The A u / P d layer is clearly vis ib le . T E M images o f a centrifuged samples are shown i n Figure 7.16(b). In this case a po lymer overlayer was v is ib le , and the A u / P d layer was attached to the embedding epoxy, w h i c h usually separated away dur ing the sectioning process. In order to establish the po lymer distr ibution, pores that cou ld be dist inguished clearly in the sections were examined. In the empty film, a thin a lumina film was associated w i t h many pores, corresponding s imply to a section w h i c h cut through part o f the pore w a l l . A s such, in the centrifuged sample, the presence o f polymer could not be dist inguished from a pore w i t h a partially sectioned wa l l . A l t h o u g h in many cases the po lymer overlayer seemed to be connected continuously w i t h the material filling the pore, the M E H - P P V distr ibution could not be established unambiguous ly by s imple examinat ion o f the T E M images. The samples were investigated further by E E L S and E F T E M . 179 7.4.4 E n e r g y - f i l t e r e d T r a n s m i s s i o n E l e c t r o n M i c r o s c o p y The approach demonstrated i n chapter 3 on P P V / M C M - 4 1 samples was applied to the analysis o f the centrifuged M E H - P P V / p o r o u s a lumina composite . It was again anticipated that the aromatic n-n p lasmon in the low-loss spectrum w o u l d reveal the po lymer distr ibution in the composite material . D u e to t ime constraints, only one sample was investigated at an accelerating voltage o f 200 k V . A 2 um-th ick porous a lumina film was used to prepare a centrifuged sample. B o t h the empty host and the centrifuged sample were ul t ramicrotomed to produce sections ~50 n m thick. The T E M image o f an empty host section is shown in Figure 7.17(a); low-loss spectra acquired on the host and on the a l u m i n u m substrate are shown in Figure 7.17(b). The a lumina bulk p lasmon loss had a m a x i m u m at 23 e V , in agreement w i th a previously reported value for amorphous a l u m i n a . 4 7 The a l u m i n u m substrate showed a surface p lasmon loss at 7 e V and bulk Energy (eV) F i g u r e 7.17 (a) T E M image (scale bar is 2 um) and (b) E E L S o f thin section o f empty porous a lumina host: ( • ) zero-loss peak, porous a lumina ( • ) before and (o) after zero-loss peak subtraction, ( A ) a luminum. 180 plasmon losses at mult iples o f 15 e V , in agreement w i t h the k n o w n a l u m i n u m p l a s m o n s . 4 8 The m i n i m u m in the a lumina spectrum occurred near 5 e V ; an attempt was therefore made to filter on this energy wi th a 2 e V slit. The energy-filtered images o f the empty and centrifuged sample are compared in Figure 7.18: both samples show substantial losses at 5 e V . The presence o f smal l tubules o f - 2 0 n m diameter in defect areas showed that some po lymer was present, but the a lumina losses at 5 e V appear to mask the po lymer distr ibution in the host. Other areas on the section were s imilar . Subtraction o f the 25 e V image (due most ly to the a lumina bulk plasmon) from the 5 e V image d i d not provide any improvement . There are two possible causes. First , it is possible that the sect ioning process has altered the centrifuged sample: the po lymer may have been displaced from the pores in the observed sections. Thus the empty and centrifuged samples appear identical . The second possible cause may be related to Cherenkov effect in the loss spectrum, as discussed in chapter 6. It appears that there are sufficient losses in a lumina be low 10 e V such that the presence o f po lymer is masked by thickness variations in the a lumina host. A previous literature report has shown by E E L S measurements at 100 k V that losses i n amorphous a lumina thin films are negligible be low 8 e V , in agreement wi th optical measurements o f the dielectric f u n c t i o n . 4 7 The results obtained here differ from this markedly since measurable losses can be observed even be low 4 e V (Figure 7.17(b)). For a 100 k V electron travel l ing through a lumina, the real part o f the dielectr ic function (£\(co)) must be above 3.3 (Table 6.1). Fo r a lumina, the value o f S\(co) is be low this threshold up to 5 e V (Figure 6.2). Th i s suggests that the use o f an accelerating voltage o f 100 k V or lower is more suitable for the investigation o f alumina-based composite materials by low-loss E E L S and E F T E M , especially for smal l amounts o f conjugated material . 181 F i g u r e 7.18 Unfi l tered ( T E M ) and energy-filtered (5, 25 e V ) images o f empty porous a lumina and centrifuged M E H - P P V / p o r o u s a lumina composite. The scale bars are 100 and 200 n m , respectively. 182 Howeve r , it cannot be excluded that some o f the features in the low-loss spectrum o f porous a lumina are due to anion-contamination o f the pore wal ls . Further study at different accelerating voltages w o u l d establish w h i c h effect is predominant. Fo r sil ica-based composi te materials (e.g. P P V / M C M - 4 1 ) , there is more f lexibi l i ty in the choice o f accelerating voltage. The dielectric function o f an a l l - s i l i ca zeolite, si l icali te, has been r epo r t ed . 4 9 These results show that £\(co) is be low 2 up to 5 e V for si l icali te and has a m a x i m u m value o f 3. Th i s suggests the losses due to the Cherenkov effect w i l l be less prominent for s i l i ca -based materials even at 200 k V , i n agreement w i t h the observed loss spectrum o f empty M C M -41 (Figure 3.11). These issues associated wi th the low-loss spectrum might have been avoided altogether by fi l tering on carbon ionizat ion edge at 284 e V instead. Th i s was not attempted because o f t ime constraints. Th i s signal is also expected to be weaker than the p lasmon loss, w h i c h w o u l d entail longer col lect ion t imes and the associated beam damage to the specimen. 7.4.5 S T E M / E E L S At tempts to measure the loss spectrum distr ibution on th in sections i n S T E M mode resulted in substantial damage to the a lumina host. The image in Figure 7.19(a) shows the damage associated wi th earlier spectrum acquisi t ions. T h e loss spectra recorded in this way showed a new peak near 9 e V (Figure 7.19(b)). Such a feature has been observed previously in a study on electron-beam hole d r i l l i ng in a l u m i n a . 5 0 It was assigned to molecular oxygen trapped i n the material, as d r i l l i ng proceeds by preferential removal o f a l u m i n u m atoms. The loss spectrum acquired direct ly on a tubule showed substantial losses be low 10 e V , suggesting that it was composed o f conjugated polymer . 183 Energy (eV) F i g u r e 7.19 (a) S T E M image o f thin section o f centrifuged sample. The contrast has been increased in the inset to show the po lymer tubules, (b) E E L S associated wi th ( • ) d r i l l i ng in a lumina and ( c ) a po lymer tubule. The zero-loss peak has been subtracted. Nevertheless, it may stil l be possible to determine the po lymer distribution in S T E M mode. T h i s has certainly been demonstrated in the core-loss r e g i m e 5 1 and may also be possible in the low-loss regime wi th careful analysis o f the spectra. Such an approach w o u l d require a cryogenic sample holder to m i n i m i z e beam damage to the a lumina host, and this was not available for this study. 7.5 C o n c l u s i o n T w o general approaches to po lymer introduction into a porous host were described. The surface-grafted synthesis o f conjugated ol igomers appears to be a p romis ing route and should be explored in more detail. The synthetic approach described here should be readily applicable to the preparation o f larger conjugated molecules on the s i l icon surface. However , its usefulness can only be conf i rmed once the structure o f the surface-grafted species is established in more 184 detail, possibly through scanning tunnel l ing microscopy. These structures w o u l d represent a j unc t ion between a molecular semiconductor and a bu lk semiconductor; the electrical behaviour o f such a junc t ion is o f fundamental interest. The use o f centrifugal force in conjunction wi th solvent evaporation has also been explored as a method to introduce a soluble conjugated po lymer in porous a lumina films. Characterizat ion o f the resulting composite clearly showed po lymer penetration. Howeve r , the po lymer could not be located w i t h h igh resolution by E E L S and E F T E M , possibly due to interference from Cherenkov losses in the a lumina host. Further experiments at different accelerating voltages w o u l d also be required i n this direct ion. Experimental Details 1. Surface Chemistry Surface derivatization o f porous s i l i con substrates fo l lowed a literature p rocedu re . 3 0 A s s i l i con is transparent to IR, F T - I R spectra were acquired i n t ransmission. Chemica l s were purchased from S i g m a - A l d r i c h , Inc. and used as-received. Ca rbon tetrachloride and ethylene g lyco l were dried over molecular sieves. Methylene chloride was dr ied by passing through activated a lumina . N M R spectra were acquired on a B r u k e r A C - 2 0 0 at 200 M H z . F T - I R spectra were recorded on a B O M E M M B 155S spectrometer. Photoluminescence spectra were acquired on a Ca ry spectrophotometer. 2-(4-ethynylphenyl) 1,3-dioxolane: Starting from 4-[(tr imethylsi lyl)ethynyl]benzaldehyde, the e thynyl group was first deprotected fo l l owing a literature p r o c e d u r e . 5 2 The aldehyde was then protected by ref luxing wi th excess ethylene g lyco l in carbon tetrachloride over ac idic a lumina for 24 h . 5 3 T L C i n 6:1 hexanes/ethyl acetate indicated a single product. The .mixture was washed twice wi th water and dried over magnes ium sulfate. Solvent removal under reduced 185 pressure y ie lded a light orange sol id . Af ter pur i f icat ion by subl imat ion, a whi te product was obtained. F T - I R ( K B r ) : 3270, 3241 ( H - C = C ) , 2890 ( H C O O ) , 2103 ( O C ) , 1424, 1384, 1222, 1074, 941 (1,3-dioxolane), 836 cm" 1 . ' H N M R (acetone-d 6): 57.4 (q, 4 H , aromatic), 5.6 (s, 1 H , benzyl ic) , 3.9 (m , 4 H , O C H 2 C H 2 0 ) , 3.5 (s, 1 H , O C H ) . M a s s spectroscopy indicated a parent M / z o f 173 ( M + - - H ) . l,4-xylylenebis(diethylphosphonate): a ,a ' -d ichloro- /?-xylene was reacted wi th 2.1 equivalents o f triethylphosphite at 140 ° C for 5 h . Excess triethylphosphite was removed under reduced pressure. The remain ing white powder was pur i f ied by recrystall ization from methylene chloride/hexanes at -20 ° C to form white needle-l ike crystals. ] H N M R (acetone-d 6): 87.2 (s, 4 H , aromatic), 4.0 (p, 8 H , M e C H 2 O P ) , 3.1 (d, 4 H , C H 2 P ) , 1.2 (t, 12 H , C H 3 ) ; 3 I P N M R : 526.2. F T -IR ( K B r ) : 2963 , 2907 ( C H ) , 1514, 1480, 1439, 1391 (aromatic), 1261 cm" 1 ( P = 0 ) . Porous silicon: substrates were prepared from n-type s i l i con wafers as reported in the literature and stored under nitrogen. F T - I R : 1100 cm" 1 ( S i - 0 f ramework) , 2240 and 2090 cm" 1 ( S i - H ) . These were present through al l the surface modif icat ions be low. Cathodic electrografting on porous silicon: The electrografting ce l l consisted o f a porous s i l icon substrate c lamped d o w n to an a l u m i n u m support w i t h a #9 glass jo in t , and capped w i t h a rubber septum. A V i t o n O - r i n g was used to form the seal to the substrate. A pla t inum mesh counter-electrode was inserted through the septum. A syringe needle was also inserted through the septum to a l low evacuation o f the ce l l . The electrografting was carried out under nitrogen atmosphere. The electrolyte, consis t ing o f 0.1 M te t rabutylammonium hexafluorophosphate and 0.2 M protected aldehyde i n methylene chlor ide, was prepared separately under nitrogen and introduced into the ce l l through the septum. A current density of ~4 m A cm" was passed through the ce l l for 3 m i n . The substrate was then washed w i t h methylene chloride, acetone, water and 186 ethanol, then dried under a stream o f nitrogen. F T - I R (not inc lud ing s i l i con surface bands): 2968, 2930, 2884 ( C - H ) , 1606, 1388, 1510 (aromatic), 943 cm" 1 (1,3-dioxolane). Aldehyde deprotection on surface: The aldehyde was deprotected by soaking in dilute HC1 for 30 m i n . The substrate was washed w i t h water and dried. F T - I R (not inc lud ing s i l i con surface bands): 2969 , 2933 , 2874 ( C - H ) , 2732 ( H - C O ) , 1700 ( C = 0 ) , 1603, 1510 (aromatic), 1307 cm" 1 (aromatic). WHE reaction on surface: The substrate, 55 m g bis(diethyl phosphonate ester) (0.15 m m o l ) and a stir bar were added to a round bottom flask and placed under nitrogen. Th i s was fol lowed by the addi t ion o f 2.0 m L T H F and 0.13 m L o f 1.0 M potassium tert-butoxide/THF (0.13 m m o l ) . A ye l low colour first developed then changed to deep orange. The mixture was refluxed for 5 h. The substrate was then washed w i t h ethanol and acetone, fo l lowed by d ry ing under streaming nitrogen. F T - I R (not inc lud ing s i l icon surface bands): 2969 , 2930 , 2869 ( C - H ) , 1510 (aromatic), 1260 cm" 1 ( P = 0 ) . Photoluminescence: E m i s s i o n m a x i m u m at 400 n m and a shoulder at 370 n m wi th excitation at 300 n m ; excitation m a x i m u m at 322 n m w i t h detection at 400 n m . 2 . V a c u u m L o a d i n g M E H - P P V was prepared accord ing to a literature p r o c e d u r e . 1 4 C o m m e r c i a l porous a lumina membranes (Anopore by Wha tman , Inc.) w i t h a nomina l pore diameter o f 200 n m on one side and 20 n m on the other side were used. The membrane thickness was 60 u m . It was placed on top o f a stainless steel tube, through w h i c h v a c u u m was appl ied. M E H - P P V polymer solution drops (0.038 to 0.8 wt % i n T H F ) were placed on top o f the membrane. Once the solution was drawn through, the membrane was removed for analysis. A Hi t ach i S-4100 f ie ld-emiss ion S E M was used for observation o f the composi te structure. 187 3. C e n t r i f u g e d Samples Porous a lumina hosts were prepared by anod iz ing a l u m i n u m fo i l samples i n 1.2 M sulfuric acid at 15.0 V or in 0.3 M oxal ic ac id at 40.0 V at room temperature, as described i n chapter 4. Some f i lms were also prepared on s i l i con substrates by anodiz ing electron-beam evaporated a l u m i n u m f i lms. These a l lowed convenient cross-section preparation by cleavage o f the s i l i con substrate. The anodizat ion t ime was generally between 2 m i n and 8 m i n , result ing i n f i lm thicknesses o f 0.5 to 2 u m . The pore wal ls were etched for 5 to 10 m i n i n 5% H3PO4, then soaked in dist i l led H2O. The samples were then r insed wi th ethanol, dr ied w i t h a heat gun and further dr ied under v a c u u m . Centrifugation was carried out i n a standard laboratory centrifuge w i t h a speed o f 1700 R P M , us ing a substrate holder described above. The approximate centrifugal force at the substrate was 3x 10 3 N . M E H - P P V was prepared according to a literature p r o c e d u r e . 1 4 A 0.038 wt % solution o f M E H - P P V in T H F (typical ly 15 - 20 u L ) was deposited into the holder by syringe and spun for 5 m i n . Pure T H F was then added (again ~20 u L ) and the holder sealed to reduce the evaporation rate. Af t e r 15 m i n o f sp inning , the seal was removed and the uncovered holder spun again for 5 m i n . A H i t ach i S-4700 field emiss ion S E M was used to observe the samples, typical ly w i th an accelerating voltage o f 20 k V for h igh resolution work . For X P S analysis o f the pore bottoms, a layer o f A u / P d was first deposited onto the top surface. Th is prevented epoxy penetration into the sample dur ing embedding o f the top surface. D u e to the need for conduct ivi ty , a s i lver-f i l led epoxy was used as the embedding m e d i u m (Epotek, Inc.). The remain ing a l u m i n u m substrate was removed by treatment w i t h HgCl2(sat)-188 A n a l y s i s was carried out on a L e y b o l d M A X 2 0 0 us ing the A l K« radiation as the excitation source. T h i n sections were obtained by ul t ramicrotomy o f samples embedded i n epoxy ( 3 0 2 - 3 M , Inc.). A d iamond knife (Micros tar , Inc.) w i t h a 4 5 ° inc luded angle was used. The knife clearance angle was set to 4 ° , and sectioning speeds as l ow as 0.2 m m s"1 were used. The th in sections were floated i n a water bath and collected either w i t h lacey carbon-coated C u grids (Ted Pel la , Inc.) or s imi lar Quantifoil-coated C u grids (Qant i fo i l M i c r o Tools G m b H ) . These grids a l lowed investigation o f the sections by E E L S without any interference from a support f i l m , due to the presence o f holes in the coatings. Init ial sample observation was carried out on a H i t ach i H - 7 6 0 0 T E M at an accelerating voltage o f 80 k V . E E L S and E F T E M were carried out on a Tecna i F 2 0 T E M equipped wi th a Gatan Imaging Filter. The accelerating voltage was 197 k V (200 k V nominal ly , offset by 3 k V by the G I F ) . Los s spectra were recorded in T E M mode by p lac ing the particle o f interest above the G I F entrance aperture (diameter 2.0 m m ) . The zero-loss peak was recorded separately by m o v i n g to an empty area on the gr id , immediately before or after spectrum col lect ion. A col lect ion t ime o f 5 s was used. Subtraction was carried out by shift ing and scal ing the zero-loss peak. E F T E M images were acquired wi th a 2 e V slit. In S T E M mode, a camera length o f 150 m m was used, and spectra were collected us ing a 50 ms d w e l l t ime. The system energy resolution, g iven by the F W H M o f the zero-loss peak, was 0.96 e V . The energy dispersion o f the spectrometer was 0.10 e V / p i x e l . References 1. M a r t i n , C . R . Acc. Chem. Res. 1995, 28, 6 1 . 2. W u , C . - G . ; B e i n , T. Science 1994, 264, 1757. 189 3. Y u , B . Z . ; Li , H . L . Mater. Sci. Eng., A 2002, 325, 215 . 4. Y u , B . Z . ; L i , M . K . ; L u , M . ; L i , H . L . Appl . Phys. A 2003, 76, 593 . 5. Kim, K.; Jin, J. I. Nano Lett. 2001, 7, 631. 6. Wagaman , M . W . ; Grubbs , R . H . Macromolecules 1997, 30, 3978 . . 7. B o y d , T . J . ; Geerts, Y . ; Lee , J . K . ; Fogg , D . E . ; Lavo ie , G . G . ; Schrock , R . R . ; Rubner , M . F . Macromolecules 1997, 30, 3553 . 8. Scherf, U . Carbon Rich Compounds Ii 1999, 201, 163. 9. Juang, A . ; Scherman, O . A . ; Grubbs , R . H . ; L e w i s , N . S. Langmuir 2001, 77, 1321. 1 0 . Harada, Y . ; G i r o l a m i , G . S.; N u z z o , R . G . Langmuir 2003 , 1 9 , 5104. 11. Li , Z . F . ; Ruckenste in , E . Macromolecules 2002, 35, 9506. 1 2 . L i u , J . ; L i n , Y . H . ; L i a n g , L . ; V o i g t , J . A . ; Huber , D . L . ; T i an , Z . R . ; Coke r , E . ; M c k e n z i e , B.; Mcdermot t , M . J . Chem. Eur. J. 2003, 9, 605 . 13 . Fabre, B . ; L o p i n s k i , G . P . ; Wayner , D . D . M . Chem. Commun. 2002, 2904. 14. Neef, C . J . ; Ferraris, J . P . Macromolecules 2000, 33, 2311 . 15. Jones, L.; S c h u m m , J . S.; Tour , J . M. J. Org. Chem. 1997, 62, 1388. 16. Strehmel, B . ; Sarker, A . M . ; Malper t , J . H . ; Strehmel, V . ; Seifert, H . ; Neckers , D . C . J . A m . Chem. Soc. 1999, 727, 1226. 17. Pfeiffer, S.; Horho ld , H . H . Macromol. Chem. Phys. 1999, 200, 1870. 18. Sugiono, E . ; Metz ro th , T. ; Detert, H . Adv. Synth. Catal. 2001, 343, 351 . 1 9 . Egbe , D . A . M . ; Corne l i a , B . ; N o w o t n y , J . ; Gunther, W . ; K l e m m , E . Macromolecules 2003, 3 6 , 5459. 20. Johnson, C . R . ; Zhang , B . R . Tetrahedron Lett. 1995, 36, 9253 . 21. Sa lv ino , J . M . ; K i e s o w , T. J . ; Darnbrough , S.; Labaudiniere , R . J. Comb. Chem. 1999, 7, 134. 190 22 . Warrass, R . ; L ippens , G . J. Org. Chem. 2000, 65, 2946. 2 3 . Bu r i ak , J . M . Chem. Commun. 1999, 1051. 24. Wayner , D . D . M ; W o l k o w , R . A . J. Chem. Soc, Perkin Trans. 2 2002, 23 . 25 . Sai lor , M . J . ; Lee , E . J . Adv. Mater. 1997, 9, 783. 26 . Song , J . H . ; Sai lor , M . J . J. Am. Chem. Soc. 1 9 9 8 , 1 2 0 , 2376. 27 . Song, J . H . ; Sai lor , M . J . Comments Inorg. Chem. 1999, 21, 69. 28 . Song, J . H . ; Sai lor , M . J . Inorg. Chem. 1999, 38, 1498. 29 . Bur i ak , J . M . ; Stewart, M . P . ; Geders, T . W . ; A l l e n , M . J . ; C h o i , H . C ; Smi th , J . ; Raftery, D . ; C a n h a m , L . T . J. Am. Chem. Soc. 1999, 121, 11491. 30. Rob ins , E . G . ; Stewart, M . P . ; Bu r i ak , J . M . Chem. Commun. 1999, 2479 . 3 1 . Schmeltzer , J . M . ; Porter, L . A . ; Stewart, M . P . ; B u r i a k , J . M . Langmuir 2002, 18, 2971 . 32. Stewart, M . P . ; B u r i a k , J . M . Comments Inorg. Chem. 2002, 23, 179. 33 . Boukher roub , R . ; M o r i n , S.; Wayner , D . D . M . ; L o c k w o o d , D . J . Physica Status Solidi a-Applied Research 2000, 182, 117. 34. Boukher roub , R . ; M o r i n , S.; Wayner , D . D . M . ; L o c k w o o d , D . J . Solid State Commun. 2001, 7 7 5 , 3 1 9 . 35 . Woj tyk , J . T . C ; M o r i n , K . A . ; Boukher roub , R . ; Wayner , D . D . M . Langmuir 2002, 18, 6081 . 36. Crouse , D . ; L o , Y . H . ; M i l l e r , A . E . ; Crouse , M . Appl. Phys. Lett. 2000, 76, 49 . 37. Glusac , K . D . ; Schanze, K . S. Polymer Preprints 2002, 43, 87. 38. W u , J . J . ; Gross , A . F . ; Tolbert , S. H . J. Phys. Chem. B 1999, 103, 2374. 39 . Hornyak , G . ; K r o l l , M . ; P u g i n , R . ; S a w i t o w s k i , T . ; S c h m i d , G . ; B o v i n , J . O . ; Karsson , G . ; Hofmeister , H . ; Hopfe , S. Chem. Eur. J. 1997, 5, 1951. 191 40 . Mar inakos , S. M . ; Brousseau, L . C ; Jones, A . ; Fe ldhe im, D . L . Chem. Mater. 1998, 10, 1214. 4 1 . Vanderbeek, G . P . ; Stuart, M . A . C ; Fleer, G . J . ; Hofrnan, J . E . Macromolecules 1991, 24, 6600. 42. Mathur , S.; M o u d g i l , B . M. Miner. Metall. Process 1998, 75, 24. 43 . E l i a s , H . G . Polymere; H u t h i g & W e p f V e r l a g : Heidelberg , 1996. 44 . Treverton, J . A. ; West , R . ; Johnson, D . ; Thornton, M. Appl. Surf. Sci. 1993, 72, 349. 45 . R e h i m , S. S. A . ; Hassan, H . H. ; A m i n , M . A . J. Appl. Electrochem. 2002, 32, 1257. 46 . P . W o n g , personal communica t ion . 47 . French, R . H . ; Mul le jans , H . ; Jones, D . J . J. Am. Ceram. Soc. 1998, 81, 2549 . 48 . Eger ton, R . F . Electron energy-loss spectroscopy in the electron microscope; 2nd ed.; P l e n u m Press: N e w Y o r k , 1996. 49. M c C o m b , D . W . ; H o w i e , A . Nucl. Instrum. Methods Phys. Res., Sect. B 1995, 96, 569. 50. Berger , S. D . ; Sal isbury, I. G . ; M i l n e , R . H . ; Imeson, D . ; Humphreys , C . J . Philos. Mag. B 1987 ,55,341. 51. Arayasant iparb, D . ; M c K n i g h t , S.; L ibe ra , M . J. Adhes. 2001, 76, 353. 52 . A u s t i n , W . B . ; B i l o w , N. ; Ke l l eghan , W . J . ; L a u , K . S. Y . J. Org. Chem. 1981, 48, 2280 . 53. K a m i t o r i , Y. ; Ho jo , M.; M a s u d a , R . ; Y o s h i d a , R . Tetrahedron Lett. 1983, 26, 4767. 192 CHAPTER 8 Comments on Device Fabrication A substantial amount o f effort was directed at fabricating a funct ioning l ight-emitt ing device ( L E D ) based on the composite materials described in this thesis. The preparation o f an L E D based on conjugated po lymer inside a porous a lumina host involves three key steps: (1) preparation o f porous thin f i l m host, (2) introduction o f the conjugated po lymer guest, and (3) appl icat ion o f electrodes. The first step was the subject o f chapter 4 and w i l l be elaborated upon here in consideration o f the third step, the need for electrodes to make an electrical device. The processes described here also assume that the conjugated po lymer insertion proceeds by centrifugation (chapter 7), w h i c h requires a host f i lm w i t h a support ing substrate. The devices prepared i n this w o r k either d i d not show any electrical conduct ivi ty or failed rapidly because o f electrical short circuits . There were no substantially new results in these endeavours but many processing issues were recognized. Th i s chapter is intended to summar ize these results and discuss possible routes to the creation o f devices based on the porous a lumina host. 8.1 Device Structure The structure o f a polymer-based ( L E D ) is constrained by the requirements o f two key processes: the field-assisted injection o f charges at the polymer-electrode interfaces, and the recombinat ion o f charges w i t h i n the po lymer (Figure 1.4). The h igh electric fields required for 193 charge i n j e c t i o n 1 ' 2 can be obtained at moderate applied potentials (< 50 V ) by m i n i m i z i n g the thickness o f the po lymer layer between the electrodes. The higher hole mobi l i ty causes charge recombinat ion to occur close to the electron-injecting electrode. Non-radia t ive quenching o f the excited state has been found to be important near this electrode, w h i c h effectively enforces a m i n i m u m device thickness o f ~90 n m . 3 A b o v e this threshold, the efficiency o f a po lymer L E D remains effectively constant w i th thickness, up to several hundred nanometres . 3 The ideal structure is shown in Figure 8.1: the central requirement for a functional device is the existence o f the contact surfaces to the anode and cathode materials. Ind ium tin oxide ( ITO) and gold are c o m m o n l y used as anode materials for M E H - P P V - b a s e d devices, whereas a number o f different low-workfunc t ion metals ( ca lc ium, silver, a luminum) may be applied as the cathode. The anodizat ion procedure for porous a lumina hosts leaves one surface immediately accessible for the applicat ion o f an electrode. The second surface is usually capped by the a lumina barrier layer, w h i c h may be over 20 n m thick (Figure 4.11). Th i s barrier layer must be F i g u r e 8.1 Ideal conjugated po lymer device components and their assembly: (a) po lymer insertion into host, (b) cathode evaporation and (c) anode deposit ion. 194 removed in all cases to create functional devices. 8.2 Devices from Porous Alumina Films on Aluminum Foil A device fabrication sequence u t i l i z ing a porous a lumina film g rown on an a l u m i n u m foi l is shown in Figure 8.2. Equivalent structures may be produced by partially anodiz ing th ick a l u m i n u m films on suitable substrates ( s i l i con wafer, glass sl ide). These substrates can be cleaved easily, w h i c h provides a convenient method for obtaining cross-sections for S E M . The barrier layer is thinned by reducing the potential at the end o f the anodizat ion process (Figures 8.3 and 8.4). Af ter po lymer introduction, an I T O anode is deposited on the po lymer by R F sputtering. A lead is attached to the I T O layer w i th si lver epoxy, and the top surface is embedded in epoxy to provide support for the device. A t this stage, the a l u m i n u m substrate can conceivably act as the electron-injecting contact for the device. Th i s was tested on several different devices but no electroluminescence was observed. A l l devices eventually failed under the applied potential. The obstacle to device operation in this form is probably the remain ing barrier layer at the bottom o f the pores. Further processing to remove this barrier layer was not possible, as the structure proved too fragile for removal o f the a l u m i n u m substrate: the host film readily separated from the deposited I T O layer dur ing the HgCl2(sat) e tching step (Figure 8.2(f)). A solution to this p roblem was not found 195 F i g u r e 8.2 D e v i c e fabrication sequence from porous a lumina film on a l u m i n u m fo i l : (a) anodizat ion o f a l u m i n u m fo i l , (b) barrier layer th inn ing by potential reduction, (c) polymer introduction by centrifugation, (d) I T O deposit ion by R F sputtering, (e) contact lead bonding w i t h si lver epoxy and epoxy embedding o f upper surface, (f) a l u m i n u m foi l r emova l by chemica l etching, (g) a l u m i n u m cathode deposit ion by thermal evaporation. 196 F i g u r e 8.3 S E M images o f porous a lumina f i lm on a n-type s i l icon wafer w i th the barrier layer el iminated by the potential reduction method. The f i lm was anodized at 15 V in 1.2 M sulfuric ac id at 2 0 ° C fo l lowed by a potential reduction to 9 V over 30 s, then a reduction to 0 V over 5 s. The scale bars are 200 n m and 100 n m , respectively. F i g u r e 8.4 S E M images o f porous a lumina f i l m , showing e l iminat ion o f barrier layer by rapid potential reduction (from 15 V to 0 V over 8 s). The scale bars are 200 n m and 100 n m , respectively. 197 8.3 Devices f r o m T h i n F i l m s o n C o n d u c t i n g Subs t ra tes T h e integrity o f the polymer/porous a lumina composi te f i l m may be better retained i f the porous a lumina host is formed directly from an a l u m i n u m f i l m on a conduct ive substrate. Substrates o f interest w o u l d include any material that can serve as anode in an L E D : s i l i con , I T O and gold . The device fabrication sequence is illustrated i n Figure 8.5. The ini t ia l step involves deposit ion o f a l u m i n u m thin f i lms. The different physical vapour deposit ion methods are reviewed below. The anodizat ion step proceeds normal ly unti l the substrate interface is reached, at w h i c h point the electrolyte may or may not react w i th the substrate material . A l u m i n u m f i lms o n s i l i con , I T O and go ld were investigated, but no suitable structures could be produced. The results and difficulties encountered are described be low. Al -ITO B F i g u r e 8.5 D e v i c e fabrication sequence from an a l u m i n u m f i lm on a conduct ive substrate: (a) in i t ia l a l u m i n u m f i lm on substrate, (b) porous a lumina f i lm growth, (c) f inal porous a lumina f i lm w i t h barrier layer removed, (d) complete device w i th conjugated po lymer sandwiched between cathode and anode. 198 8.3.1 A l u m i n u m F i l m D e p o s i t i o n The a l u m i n u m deposit ion step must also be carefully controlled for the preparation o f defect-free films. Proper c leaning o f the substrate was necessary to obtain good adhesion. A s imple protocol consis t ing o f two 10 m i n sonication steps (detergent solution, methanol) was found to be suitable for most substrates. A number o f different film deposit ion methods available at U B C were investigated. R F sputtering and electron-beam evaporation were available in the A M P E L cleanroom. D C sputtering was available i n the Department o f Physics , wh i l e thermal evaporation was available i n the Department o f Elect r ica l Engineer ing . D C sputtering o f a l u m i n u m al lows convenient deposit ion o f thicker films (> 1 u m , rate > 0.5 n m s"1). However , the setup available at U B C was found to produce contaminated films, as evidenced by gas evolut ion dur ing the anodizat ion process. It was not used for any further work . R F sputtering is normal ly used for the deposit ion o f dielectric materials. It may also be used to deposit a luminum, however h i l lock formation is k n o w n to occur (Figure 8.6). The resulting surface roughness lowers the reflectivity o f the a l u m i n u m film. The addit ion o f a smal l F i g u r e 8.6 S E M images o f h i l locks on porous a lumina films prepared from R F sputtered a l u m i n u m films, showing (a) top surface and (b) cross-section. The scale bars are 200 n m . 199 amount o f copper (~1%) is used in the semiconductor industry to prevent h i l lock formation. Th i s was not attempted here, as the effect o f copper impuri t ies on the anodizat ion step was not k n o w n . Nevertheless, R F sputtering produced continuous films w i t h good adhesion. The deposit ion rate was typical ly 0.2 n m s"1. Electron-beam evaporation is also c o m m o n l y used to deposit a l u m i n u m . The only difficulty was the presence o f pinholes in the resulting film i f the deposit ion process is too rapid. The associated surface roughness again causes a lower reflectivity in the deposited films. The anodized films then contain defects as shown in Figure 8.7. Defect-free films were achievable by keeping the deposit ion rate be low 0.5 n m s"1. Thermal evaporation was found to produce films wi th pinholes as w e l l . In this case, the effect o f deposit ion rate was not investigated i n more detail. S ince the phys ica l process is almost identical to electron-beam evaporation, it is expected that control o f the deposit ion rate w o u l d produce s imi lar pinhole-free films. 8.3.2 Porous Alumina/Silicon The possibi l i ty o f direct ly us ing s i l i con wafers as an electrode for an organic electroluminescent device has been explored in the l i terature. 4 " 6 Parker and K i m showed that both degeneratively doped n - and /?-type s i l i con wafers can be used as anode or cathode in devices made w i t h M E H - P P V . 4 W u n s c h e t a l . found that /?-type s i l i con was a better hole-injecting material than either go ld or I T O . 6 Crouse e t a l . investigated the fabrication o f porous a lumina films on s i l icon wafers for the purpose o f hexagonal pattern transfer to the wafe r . 7 They found that « - type s i l i con could be used readily as a substrate, as there were no unfavourable reactions w i t h oxal ic , sulfuric or phosphoric ac id up to potentials o f 110 V . It was observed that the barrier layer cou ld be readily removed by 200 F i g u r e 8.7 Defects in porous a lumina f i lms anodized from electron-beam evaporated a luminum f i l m , showing (a) S E M image o f cross-section o f a film anodized at 20 ° C in 1.2 M sulfuric acid and (b) T E M image in p lan v i e w o f a film anodized at -39 ° C . The scale bars are 200 n m long. a short etching step wi th 5% phosphoric acid. O n the other hand, p-type s i l icon was not suitable without an insulating coating o f s i l i con d ioxide . A s degeneratively doped s i l icon wafers were not readily available, this approach was not pursued any further but may be a p romis ing avenue for future work . 8.3.3 P o r o u s A l u m i n a / I T O The preparation o f porous a lumina films on ITO-coated glass substrates has been reported by C h u e t a / . 8 The anodizat ion process normal ly consumes the I T O layer i f it is not halted once the a l u m i n u m layer is fully ox id ized . I f anodization is halted at the right point, the I T O layer is preserved and the desired structure is obtained. These results were obtained wi th a 10 v o l . % phosphoric ac id electrolyte and 130 V . The pore size in the resulting film was 80-100 n m . Protection o f the I T O layer by deposit ion o f a th in layer o f s i l i con dioxide or a luminum oxide between the I T O and a l u m i n u m layers is also possible. One literature report indicated that a thin s i l icon d ioxide layer above the I T O layer was even beneficial to L E D efficiency, through a 201 better balance o f the hole and electron injection rates. 9 A n opt imal thickness near 2 n m was reported. At tempts were made to reproduce these results w i th a l u m i n u m deposited on ITO-coated glass and poly(ethylene terephthalate) substrates. The latter is advantageous for obtaining cross-sections by ul t ramicrotomy. However , pinholes were always present in the a luminum f i l m , as evidenced by immediate gas evolut ion (i.e., reaction o f the I T O ) at the beginning o f the anodization process. It is bel ieved that deformation o f the plastic substrate under the thermal load o f the evaporation process was the cause o f the pinholes. O n ITO-coated glass substrates, both sulfuric and oxal ic acid electrolytes (for pore sizes o f ~21 and 60 n m respectively; see chapter 4) were used to anodize the deposited a l u m i n u m f i l m . However , the anodization process, us ing the cel l depicted in Figure 4.10, d id not proceed uni formly across the substrate, w i t h the outer edge being anodized faster. Th i s produced structures in w h i c h the I T O layer was consumed at the edges (Figure 8.8), and a luminum remained in the centre. The deposit ion o f thin s i l icon d ioxide and a lumina layers (up to 4 n m in thickness) was not F i g u r e 8.8 S E M image o f porous a lumina film on ITO-coated glass substrate, showing an area where most o f the I T O was consumed. The scale bar is 200 n m . 202 found to provide any significant protection to the I T O layer dur ing anodizat ion. Th i cke r layers may have provided more protection but w o u l d then have required removal by a different process (e.g., reactive-ion etching). It is also possible that s t i rr ing o f the electrolyte w o u l d have created a more uni form anodizat ion on the substrate. L o w e r i n g the electrolyte temperature w o u l d reduce the anodizat ion rate, w h i c h may a l low better control over the end point. A s these difficult ies were not resolved, this approach d id not provide any suitable host films for further processing. 8.3.4 Porous Alumina/Gold G o l d is also c o m m o n l y used as hole-injecting electrode in polymer-based devices. However , there are two difficulties associated wi th fabricating porous a lumina films over gold . First , a l u m i n u m and gold interdiffuse readily to create an intermetallic c o m p o u n d . 1 0 Th i s proceeds rapidly at 100 ° C , w i th f i lms deteriorating wi th in 1 h. Room-temperature degradation proceeds at a s lower rate, w h i c h necessitates immediate processing o f deposited films. The second difficulty, as w i th I T O , is in the anodizat ion process: gas evolut ion occurs at the gold interface once the a l u m i n u m is consumed. A literature report indicated that it was again possible to obtain useful structures by stopping the anodizat ion at the correct t i m e . 1 1 Efforts to reproduce this result were unsuccessful due to immediate reaction evolut ion o f gas from the substrate, w h i c h suggested that pinholes were present in the a l u m i n u m film. N o further attempts were made to prepare porous a lumina hosts in this manner. 8.4 Conclusion The fabrication o f a device based on the conjugated polymer/porous a lumina composite material revealed a large number o f processing issues. The current results indicate that some o f 203 these may be overcome wi th further effort. In the case o f films on a l u m i n u m fo i l , more careful processing to remove the a l u m i n u m substrate is necessary to preserve the integrity o f the structure. A s for films on s i l i con , I T O and go ld , control o f the deposited a l u m i n u m morphology appears to be key for obtaining the desired structure without causing gas evolut ion from the under ly ing substrate dur ing anodizat ion. Dev ices on n-type s i l icon should st i l l be investigated in more detail, as it is k n o w n that it shows no reaction wi th the electrolyte. The efforts described here were a imed at relatively crude devices i n w h i c h conjugated po lymer chain confinement was not really possible. However , once the obstacles identified here are resolved, the use o f l ow temperature anodization should produce the desired structure for such confinement and a l low the study o f single chain electrical properties. Experimental Details ITO-coated glass and poly(ethylene terephthalate) substrates were obtained from Del ta Technologies , Inc. S i l i c o n wafers ( « - t y p e ) were obtained from Monsan to , Inc. Substrates were cleaned pr ior to deposit ion by sonicat ing for 10 m i n i n a detergent solution ( 1 0 % F L - 7 0 , F isher Scient if ic) and for 10 m i n i n methanol, w h i c h was fo l lowed by d ry ing in streaming nitrogen. G o l d substrates were prepared by thermal evaporation on s i l i con w i t h a thin c h r o m i u m adhesion layer (~10 nm). A l u m i n u m thin films on s i l i con , gold and I T O were prepared as described i n chapter 4. S E M images were acquired on a H i t ach i S-4700 field-emission S E M . I T O films were prepared by R F sputtering at 100 W w i t h 6 m T o r r argon (120 seem f low). Some film darkening was observed due to reduction o f metall ic impuri t ies . T h i s may be avoided by m i x i n g a smal l oxygen f low (< 0.1 seem) to the chamber dur ing deposit ion. S i l i c o n d iox ide 204 and a l u m i n u m oxide protective films were prepared by electron-beam evaporation in the same process as a l u m i n u m evaporation. For device testing, electrical contacts were attached to the I T O layer us ing si lver epoxy (Epotek, Inc.) The I T O layer was further embedded i n opt ical epoxy ( 3 0 2 - 3 M , Epotek, Inc.). Di rec t contact was made to the a l u m i n u m substrate by an alligator c l ip . Current-voltage characteristics were investigated w i t h a D C power supply and a mult imeter. The measurements were in general difficult to reproduce, presumably due to the presence o f electrical short circuits in the device that w o u l d degrade wi th t ime. Cond i t i on ing at h igh voltage (>50 V ) improved the stability somewhat but a l l devices eventually burned out. References 1. Parker , 1. D . J. Appl. Phys. 1994, 75, 1656. 2 . A r k h i p o v , V . I.; Eme l i anova , E . V. ; Tak , Y . H . ; Bassler , H . J. Appl. Phys. 1998, 84, 848. 3. Cao , Y . ; Parker , I. D . ; Y u , G . ; Zhang , C ; Heeger, A . J . Nature 1999, 397, 414. 4. Parker, I. D . ; K i m He len , H . Appl. Phys. Lett. 1994, 64, MIA. 5. Z h o u , X . ; H e , J . ; L i a o , L . S.; L u , M . ; X i o n g , Z . H . ; D i n g , X . M . ; H o u , X . Y . ; Tao, F . G . ; Z h o u , C. E.; Lee, S. T. Appl. Phys. Lett. 1999, 74, 609. 6. W u n s c h , F . ; Chaza lv i e l , J . N.; O z a n a m , F . ; Sigaud, P . ; Stephan, O . Surf. Sci. 2001, 489, 191. 7. Crouse , D . ; L o , Y . H . ; M i l l e r , A . E . ; Crouse, M . Appl. Phys. Lett. 2000, 76, 49 . 8. C h u , S. Z . ; Wada , K ; Inoue, S.; T o d o r o k i , S. J. Electrochem. Soc. 2002, 149, B 3 2 1 . 9. D e n g , Z . B . ; D i n g , X . M. ; Lee , S. T. ; G a m b l i n g , W . A . Appl. Phys. Lett. 1999, 74, 2227 . 10 . Hunter , W . R . ; M i k e s , T . L . ; Hass , G . Appl. Opt. 1972 ,1 1 , 1594. 1 1 . Y a n g , Y . ; C h e n , H . L . ; M e i , Y . F . ; C h e n , J . B . ; W u , X . L . ; B a o , X . M . Sol id State Commun. 2002, 123, 279. 205 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061179/manifest

Comment

Related Items