Neutron and X-ray Surface Scattering Reveals the Morphology of Soft Matter Thin Films

Neutron s X-rays

Doctoral Thesis in Engineering Mechanics

Neutron and X-ray Surface Scattering

Reveals the Morphology of Soft

Matter Thin Films

CALVIN BRETT

Neutron and X-ray Surface Scattering

Reveals the Morphology of Soft

Matter Thin Films

CALVIN BRETT

Doctoral Thesis in Engineering Mechanics KTH Royal Institute of Technology Stockholm, Sweden 2020

Academic Dissertation which, with due permission of the KTH Royal Institute of Technology, is submitted for public defence for the Degree of Doctor of Philosophy on Friday the 11 December 2020, at 10:15 a.m. in F3, Lindstedsvägen 26, Stockholm.

TRITA-SCI-FOU 2020:40 ISBN 978-91-7873-717-8

Neutron and X-ray Surface Scattering Reveals the

Mor-phology of Soft Matter Thin Films

Calvin Brett

Department of Engineering Mechanics, Royal Institute of Technology SE-100 44 Stockholm, Sweden

Wallenberg Wood Science Center, Royal Institute of Technology SE-100 44 Stockholm, Sweden

Deutsches Elektronen Synchrotron (DESY) DE-22 607 Hamburg, Germany

Abstract

The last decades have been overshadowed by reports about the seemingly endless increase use of fossil-based resources. With the development of new products, our mindset is changing so that we more and more need to consider sustainability in our daily lives. Furthermore, smarter devices are indispensable in our world and daily life, and these are expected to be smaller and smaller in size.

To support the transition from fossil-based to sustainable materials, we need to develop knowledge of new materials. Within this thesis project, the aim has been to understand the thin-film properties of sustainable materials and to develop methodologies to measure these. As sustainable template material wood-based nanocellulose was chosen as a bio-degradable representative with favourable favourable physical properties, such as lightweight, transparency, and flexibility. These properties make nanocellulose a perfect candidate for future advanced applications in thin-film organic solar cells, supercapacitors, or sensors. Nanocellulose comprises only a part of such a device, and hence the relevant functional materials and their combinations have to be studied to reveal the in-teraction between multiple material components on the final device performance. As the nanoscale, or even ˚Angstrom scale, governs the macroscopic physical properties, it is crucial to understand the materials in detail. Ergo, neutron and X-ray surface-sensitive scattering methods were applied to study nanoparticle deposition layering kinetics and the effects of environmental changes, which revealed the morphology of the resulting nanoporous nanocellulose thin films. The knowledge was used to infiltrate water-soluble intrinsic conductive polymers into these nanopores, which serves as a model for transparent but conductive templates for organic electronics. By changing the environment of the films through humidity cycling, the impact of the environment during a real-life application could be illustrated. Neutron scattering experiments also showed that the cellulose-conductive polymer composite (or hybridmaterial) changes irreversibly during humidity cycling while the pure nanocellulose films show fully reversible properties.

Furthermore, the thermal decomposition of silver nitrate deposited on nanocellulose was studied to understand the nanofibrils’ impact on the synthesis of nanoparticles. The transparency allowed in situ studies of the synthesis

demonstrated routes for minimal material usage concepts for surface synthesis processes. It was also discovered that the process allows for band-gap tuning, which can be directly be applied in organic solar cells to tailor the band-gap to be adapted and hence increasing the efficiency. The morphological properties, as studied using X-rays and neutrons, were correlated to macroscopic properties by measuring wettability, surface topography, spectroscopy, or conductivity to examine the full materials application possibilities. Neutron and X-ray scattering methods are complementary and wisely combined, thus allowed pioneering studies of bio-based sustainable nanocomposites leading to advanced functional material concepts that support the development of devices using less fossil-based materials.

Keywords: Cellulose, soft matter, X-ray scattering, neutron scattering, wetta-bility, spectroscopy.

Neutron- och r¨

ontgenytspridningsmetoder avsl¨

ojar

morfo-login hos tunna filmer av mjuka material

Calvin Brett

Teknisk Mekanik, Kungliga Tekniska H¨ogskolan

SE-100 44 Stockholm, Sverige Wallenberg Wood Science Center, Kungliga Tekniska H¨ogskolan

SE-100 44 Stockholm, Sverige

Deutsches Elektronen Synchrotron (DESY) DE-22 607 Hamburg, Tyskland

Sammanfattning

Den ¨okande anv¨andningen av fossilbaserade resurser representerar en global samh¨allsutmaning som kan m¨otas genom utveckling av nya materialkoncept. V˚art samh¨alle fylls ¨aven av smarta produkter, vilka b˚ade f¨orv¨antas bidra till ¨

okad h˚allbarhet och att kontinuerligt bli mindre i storlek. F¨or att m¨ojligg¨ora denna utveckling och samtidigt minska anv¨andningen av fossilbaserade material beh¨over vi ¨oka v˚ar kunskap om nya biobaserade material och deras anv¨andning. M˚alet med denna avhandling ¨ar en ¨okad insikt i egenskaperna hos tunna biobaserade filmer, och att utveckla metoder f¨or att karakterisera dessa egen-skaper. Utg˚angspunkten har varit att anv¨anda vedbaserad nanocellulosa, en biobaserad och f¨ornyelsebar r˚avara som ¨ar biologiskt nedbrytbar, har mycket goda mekaniska egenskaper, och kan anv¨andas f¨or framst¨allningen av material-koncept med l˚ag densitet, transparens och flexibilitet. Detta g¨or nanocellulosa till en perfekt grundkomponent i framtida avancerade biobaserade tunnfilmsap-plikationer s˚asom organiska solceller, superkondensatorer eller sensorer, vilket m¨ojligg¨ors genom inblandning av komponenter som ger ¨onskad funktion. Olika materialkoncept har d¨arf¨or studerats med syftet att ¨oka f¨orst˚aelsen f¨or inter-aktionerna mellan olika komponenter och hur dessa p˚averkar prestanda och funktion.

De makroskopiska egenskaperna av ett material ges av dess hierarkiska struktur, fr˚an ˚Angstr¨oms- och nanometerskalan och upp till variationer p˚a mikrometer och millimeterskalan. Ergo har fokus varit p˚a att studera kinetik relaterat till framst¨allning av filmer uppbyggda genom ytdeponering av nanopar-tiklar, inkluderande studier av effekter av omgivande atmosf¨ar. Dessa studier har genomf¨orts med hj¨alp av ytselektiva neutron- och r¨ontgenspridningsmetoder och har resulterat i goda insikter i morfologin hos tunna nanopor¨osa nanocellulosa-filmer. Vidare har ¨aven effekter av att fylla dessa nanoporer med vattenl¨osliga ledande polymerer studerats, d˚a detta ¨ar en intressant funktionaliseringsv¨ag som m¨ojligg¨or tillverkning av transparenta och ledande templat f¨or organisk elektronik. I detta sammanhang studerades ¨aven effekter av fuktvariationer i omgivningen, vilket ¨ar en viktig fr˚agest¨allning f¨or m˚anga biobaserade material och cellulosabaserade material i synnerhet. Med hj¨alp av neutronspridnings-experiment kunde skillnader mellan rena nanocellulosafilmer och filmer fyllda

reversibelt under fuktcykling, till skillnad fr˚an filmerna med ledande polymerer, vars struktur f¨or¨andrades irreversibelt. Slutligen studerades ¨aven temperatu-rens inverkan p˚a syntesen av nanopartiklar som initieras genom deponering av silvernitrat p˚a nanocellulosafilmer, d¨ar materialens transparens till¨at in-situ stu-dier av processen, spektroskopiska egenskaper samt plasmoniska effekter. Detta visade att nanocellulosafilmer ¨ar ett lovande templat f¨or ytsyntesprocesser, d¨ar det s¨arskilt observerades att de olika processparametrarna direkt kan kopplas till det observerade bandgapet, kunskaper som kan anv¨andas f¨or att skr¨addarsy organiska solceller s˚a att dessa blir effektivare.

Nyckelord: Cellulosa, polymera material, materialprocesser, r¨ontgenspridning, neutronspridning.

List of Appended Papers

Paper 1 Calvin J. Brett, Stefano Montani, Matthias Schwartzkopf, Rolf A. T. van Benthem, Johan F. G. A. Jansen, Gianmarco Griffini, Stephan V. Roth & Mats K. G. Johansson, 2020. Revealing structural evolution occuring from photo-initiated polymer network formation. Commun. Chem. 3 (1), 88.

Paper 2 Joakim Engstr¨om*, Calvin J. Brett*, Volker K¨orstgens, Peter M¨ uller-Buschbaum, Wiebke Ohm, Eva Malmstr¨om & Stephan V. Roth, 2020. Core-Shell Nanoparticle Interface and Wetting Properties. Adv. Funct.

Mater. 30, 1907720. *equal contribution.

Paper 3 Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Marc Gensch, Lucas P. Kreuzer, Volker K¨orstgens, Martin M˚ansson, Henrich Frielingshaus, Pe-ter M¨uller-Buschbaum, L. Daniel S¨oderberg & Stephan V. Roth, 2019. Water-Induced Structural Rearrangements on the Nanoscale in Ultrathin

Nanocellulose Films. Macromolecules 52, 4721-4728.

Paper 4 Calvin J. Brett, Wiebke Ohm, Bj¨orn Fricke, Tim Laarmann, Volker K¨orstgens, Peter M¨uller-Buschbaum, L. Daniel S¨oderberg & Stephan V. Roth. Nanocellulose-Assisted Thermally-Induced Growth of Silver Nanoparticles for Optical Applications. Under revision.

Paper 5 Calvin J. Brett, Ola K. Forslund, Elisabetta Nocerino, Lucas P. Kreuzer, Tobias Wiedmann, Lionel Porcar, Norifumi Yamada, Nami Matsub-ara, Martin M˚ansson, Peter M¨uller-Buschbaum, L. Daniel S¨oderberg & Stephan V. Roth. Humidity-induced Nanoscale Restructuring in PE-DOT:PSS and Cellulose reinforced Bio-based Organic Electronics. In preparation.

Paper 1 Part of experimental work, major part of planning, evaluation, supervi-sion of experimental work and writing.

Paper 2 Part of experimental work, major part of planning, evaluation and writing.

Paper 3 All experimental work, major part of planning, evaluation and writing. Paper 4 All experimental work, major part of planning, evaluation and writing. Paper 5 All experimental work, major part of planning, evaluation and writing.

Other publications

The following papers, although related, are not included in this thesis. Qing Chen, Calvin J. Brett, Andrei Chumakov, Marc Gensch, Matthias Schwartzkopf, Volker K¨orstgens, L. Daniel S¨oderberg, Anton Plech, Peng Zhang, Peter M¨uller-Buschbaum & Stephan V. Roth, 2020. Layer-by-layer Spray-coating of Cellulose Nanofibrils and Silver Nanoparticles for Hydrophilic Interfaces. Under revision.

Wei Cao, Shanshan Yin, Martina Plank, Andrei Chumakov, Matthias Opel, Wei Chen, Lucas P. Kreuzer, Julian E. Heger, Markus Gallei, Calvin J. Brett, Matthias Schwartzkopf, Artem A. Eliseev, Evgeny O Anokhin, Lev A. Trusov, Stephan V. Roth & Peter M¨ uller-Buschbaum, 2020. Spray Deposited Anisotropic Ferromagnetic Hybrid Polymer Films of PS-b-PMMA and Strontium Hexaferrite Magnetic Nanoplatelets. Under revision.

Iuliana Ribca, Marcus E. Jawerth, Calvin J. Brett, Martin Lowoko, Matthias Schwartzkopf, Andrei Chumakov, Stephan V. Roth & Mats Johansson, 2020. Exploring the effects of different crosslinkers on lignin-based thermoset properties and morphologies. Under revision.

Zhaoyang Yao, Fuguo Zhang, Yaxiao Guo, Heng Wu, Lanlan He, Zhou Liu, Bin Cai, Yu Guo, Calvin J. Brett, Yuanyuan Li, Chinmaya Venugopal Srambickal, Xichuan Yang, Gang Chen, Jerker Widen-gren, Dianyi Liu, James M. Gardner, Lars Kloo & Licheng Sun, 2020. Conformational and Compositional Tuning of Phenanthrocarbazole Based Dopant-Free Hole-Transport Polymers Boosting Performance of Perovskite Solar Cells. J. Am. Chem. Soc. 142 (41), 17681-17692.

Marcus E. Jawerth, Calvin J. Brett, C´edric Terrier, Per T. Lars-son, Martin Lawoko, Stephan V. Roth, Stefan Lundmark & Mats Johansson, 2020. Mechanical and Morphological Properties of Lignin-Based Thermosets. ACS Appl. Polym. Mater. 2 (2), 668–676.

Xia Sheng, Yuanyuan Li, Taimin Yang, Brian J. J. Timmer, Tom Willhammar, Ocean Cheung, Lin Li, Calvin J. Brett, Stephan V. Roth, Biaobiao Zhang, Lizhou Fan, Yaxiao Guo, Xiaodong Zou, Lars Berglund & Licheng Sun, 2020. Hierarchical micro-reactor as electrodes for water splitting by metal rod tipped carbon nanocapsule self-assembly in carbonized wood. Appl. Catal. B 264, 118536.

Leila Josefsson, Xinchen Ye, Calvin J. Brett, Jonas Meijer, Carl Olsson, Amanda Sj¨ogren, Josefin Sundl¨of, Anton Davydok, Maud Langton, ˚Asa Emmer & Christofer Lendel, 2020. Potato Protein Nanofibrils Produced from a Starch Industry Sidestream. ACS Sustainable Chem. Eng. 8 (2), 1058–1067.

Fuguo Zhang, Zhaoyang Yao, Yaxiao Guo, Yuanyuan Li, Jan Bergstrand, Calvin J. Brett, Bin Cai, Alireza Hajian, Yu Guo,

Roth, Lars Kloo & Licheng Sun, 2019. Polymeric, Cost-Effective, Dopant-Free Hole Transport Materials for Efficient and Stable Perovskite Solar Cells. J. Am. Chem. Soc. 141 (50), 19700–19707.

Marc Gensch, Matthias Schwartzkopf, Wiebke Ohm, Calvin J. Brett, Pallavi Pandit, Sarathlal K. Vayalil, Lorenz Bießmann, Lucas P. Kreuzer, Jonas Drewes, Oleksandr Polonskyi, Thomas Strunskus, Franz Faupel, Andreas Stierle, Peter M¨uller-Buschbaum & Stephan V. Roth, 2019. Correlating Nanostructure, Optical and Electronic

Properties of Nanogranular Silver Layers during Polymer-Template-Assisted Sputter Deposition. ACS Appl. Mater. Interfaces 11 (32), 29416–29426. Tomke E. Glier, Lewis Akinsinde, Malwin Paufler, Ferdinand Otto, Maryam Hashemi, Lukas Grote, Lukas Daams, Gerd Neuber, Ben-jamin Grimm-Lebsanft, Florian Biebl, Dieter Rukser, Milena Lipp-mann, Wiebke Ohm, Matthias Schwartzkopf, Calvin J. Brett, Toru Matsuyama, Stephan V. Roth & Michael R¨ubhausen, 2019. Functional Printing of Conductive Silver-Nanowire Photopolymer Composites. Sci. Rep. 9 (6495).

Nian Li, Lin Song, Lorenz Bießmann, Senlin Xia, Wiebke Ohm, Calvin J. Brett, Efi Hadjixenophontos, Guido Schmitz, Stephan V. Roth & Peter M¨uller-Buschbaum, 2019. Morphology Phase Diagram of Slot-Die Printed TiO2 Films Based on Sol–Gel Synthesis. Adv. Mater. Interfaces 6 (1900558).

Wiebke Ohm, Andr´e Rothkirch, Pallavi Pandit, Volker K¨orstgens, Peter M¨uller-Buschbaum, Ramiro Rojas, Shun Yu, Calvin J. Brett, Daniel L. S¨oderberg, & Stephan V. Roth, 2018. Morphological properties of airbrush spray-deposited enzymatic cellulose thin films. J. Coat. Technol. Res. 15, 759–769.

Conferences

Part of the work in this thesis has been presented at the following international conferences. The presenting author is underlined.

Oral contribution:

Calvin J. Brett, Mehmet Girayhan Say, Isak Engquist & Magnus Berggren, Daniel L. S¨oderberg & Stephan V. Roth. In situ R2R Spray Deposition of Flexible Nanocellulose/PEDOT:PSS-based Supercapacitors Studied Using Surface Sensitive X-ray Scattering. MRS Spring/Fall Meeting & Exhibit. Virtual, 2020.

Calvin J. Brett, Daniel L. S¨oderberg, Mats K. G. Johansson & Stephan V. Roth. In situ X-ray study on Roll-to-Roll spray deposited nano cellulose thin films. COSI conference. Noordwjik, Netherlands, 2019.

Calvin J. Brett, Stephan V. Roth & Daniel L. S¨oderberg. R2R spray deposition of CNF thin films studied real-time using surface sensitive scattering methods. TAPPI NANO. Chiba, Japan, 2019.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. GISAS study of spray deposited metal precursor ink on a cellulose template. ACS Spring Meeting. Orlando, USA, 2019.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. GISAXS / WAXS self-assembly study of spray deposited metal precursor inks on bio-based polymers. GISAS conference. Gyeongju, Korea, 2018.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. In situ self-assembly study of spray deposited

metal-organic decomposition inks on bio-based polymers. MACRO conference. Cairns, Australia, 2018.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. Large-scale spray deposited conductive nanoparticle on bio-based thin films. ETCC/COSI conference. Amsterdam, Netherlands, 2018. Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. Tailored wetting of charged cellulose nano fibrils on thin films. ACS Spring Meeting. Orlando, USA, 2018.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. In situ growth study during spray deposition of bio-based materials. DPG Spring Meeting. Berlin, Germany, 2018.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. In situ growth studies on charged cellulose nanofibril thin films. GRC Conference. Boston, USA, 2017.

Calvin J. Brett, Wiebke Ohm, Matthias Schwartzkopf, Pablo Mota-Santiago, Daniel Severin, Nigel Kirby, Christina Trautmann,

adjustable nanometer sized ion tracks in Polydimethylsiloxane. DPG Spring Meeting. Dresden, Germany, 2017.

Poster contribution:

Calvin J. Brett, Daniel L. S¨oderberg & Stephan V. Roth. Tailored functionalization of nanocellulose thin films using grazing incidence X-ray and neutron scattering. GISAXS 2019. Hamburg, Germany, 2019.

Calvin J. Brett, Daniel L. S¨oderberg & Stephan V. Roth. GISAS study of spray deposited nanocellulose templates. RACIRI 2019. Kaliningrad, Russia, 2019.

Calvin J. Brett, Daniel L. S¨oderberg & Stephan V. Roth. Self-assembly and Structural Morphology Study of Thin Nanocellulose Templates. DESY User Meeting. Hamburg, Germany, 2019.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. Study of sprayed bio-based thin films. DESY User Meeting. Hamburg, Germany, 2018.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. In situ growth studies on charged cellulose nano fibril thin films. Kolloid Tagung. Munich, Germany, 2017.

Calvin J. Brett, Nitesh Mittal, Wiebke Ohm, Daniel L. S¨oderberg & Stephan V. Roth. In situ growth studies on charged cellulose nano fibril thin films. COSI Conference. Noordwjik, Netherlands, 2017.

Abbrevations

2D Two-dimensional

3D Three-dimensional

4D Four-dimensional

AFM Atomic force microscopy

Ag Silver

Ag2O Silver oxide

AgCl Silver chloride

AgNO3 Silver nitrate

BFGS2 Broyden-Fletcher-Goldfarb-Shanno algorithm C2H6O2 Ethylene glycol

CAM Contact-angle measurements CH2I2 Diiodmethane

CNF Cellulose nanofibres

Co Cobalt

Coh. elas. Coherent elastic

Corr. Correlation CPP CNF/PEDOT:PSS sample Cr Chromium Cu Copper D2O Deuterium oxide DE Germany

DMA Dynamic mechanical analysis

DMAEMA N,N-dimethylaminoethyl methacrylate DWBA Distorted wave Born approximation EELS Electron energy loss spectroscopy e.g. Exempli gratia / for example

Fe Iron

FE-SEM Field-emission scanning electron microscopy

FF Form factor

FR France

FTIR Fourier-transform infrared spectroscopy

GISANS Grazing incidence small-angle neutron scattering GISAS Grazing incidence small-angle scattering

GISAXS Grazing incidence small-angle X-ray scattering GIWAXS Grazing incidence wide-angle X-ray scattering GSL Gnu Scientific Library

H2O Water HCl Hydrochloric acid Hg Mercury HT High Tg resin JP Japan xvi

Abbreviation Definition

LED Light-emitting diode LNMG Log-normal Mie-Gans LT Low Tg resin

MiNaXS Micro- and Nanofocus X-ray Scattering Beamline PETRA III

Mo Molybdenum

NaClO Sodium hypochlorite NaOH Sodium hydroxide NP Nanoparticle

NR Neutron reflectometry

OWKR Owens, Wendt, Rabel, and Kaelble PANI Polyaniline

Part. Particle

PBMA poly(butyl methacrylate)

PDMAEMA Poly(2-dimethylaminoethy methacrylate)

PEDOT:PSS Poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) PiLC Rasberry Pi Logic Controller

PMMA Poly(methyl methacrylate) PP PEDOT:PSS sample QR-code Quick Response code R2R Roll-to-roll

RAFT Reversible addition–fragmentation chain transfer RH Relative humidity

RT Room temperature

SANS Small-angle neutron scattering SAXS Small-angle X-ray scattering SDD Sample-to-detector distance SEM Scanning electron microscopy

Si Silicon

SiN Silicon nitride SiO2 Silica

SLD Scattering length density Solv. Solvent

TEM Transmission electron-microscopy TEMPO (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl TOF time-of-flight

GB United Kingdom US United states of America

USANS Ultra small-angle neutron scattering USAXS Ultra small-angle X-ray scattering UV Ultraviolet light

UV-Vis Ultraviolet to visible light WANS Wide-angle neutron scattering WAXS Wide-angle X-ray scattering

Xe Xenon

XRR X-ray reflectometry

Contents

Abstract v

Sammanfattning vii

List of Appended Papers ix

Abbrevations xv

Part I - Overview and Conclusions

Chapter 1. Introduction 1

Chapter 2. Scientific Context 5

2.1. Deposition Techniques 5

2.1.1. Spray Deposition 5

2.1.2. Spin Coating 6

2.2. Thin Film Characterization 7

2.2.1. Neutron/X-ray Interaction with Matter 7

2.2.1.1. Scattering 7

2.2.1.2. Absorption 11

2.2.1.3. Reflectometry 12

2.2.1.4. Small- and Wide-angle Neutron/X-ray Scattering 16 2.2.1.5. Grazing Incidence Small-angle Neutron/X-ray Scattering 19

2.2.1.6. Simulation 22

2.2.1.7. X-ray and Neutron Sources 22

2.2.2. Atomic Force Microscopy 25

2.2.3. Scanning Electron Microscopy 25

2.2.4. UV-Vis Spectroscopy 27

2.2.5. Contact Angle Measurements 30

2.3. Materials 31

2.3.1. Silicon and Silica 32

2.3.3. Latex 34

2.3.4. PEDOT:PSS 34

2.3.5. Photo-active Polymer Resins 35

2.3.6. Silver and Silver Nitrate 36

Chapter 3. Instrumentation 37

3.1. Lab Spray Deposition Setup 37

3.2. R2R Spray Deposition Setup 38

3.3. UV-Polymerization Setups 40

3.4. NR/GISANS Cyclic Charge/Discharge Setup 41

Chapter 4. Results and Discussions 42

4.1. Paper 1 - Structural Evolution during UV Polymerization 42 4.2. Paper 2 - Tailoring of Wettability by Core-shell Nanoparticles 49 4.3. Paper 3 - Restructuring Induced by Water in CNF Films 55 4.4. Paper 4 - CNF-assisted Growth of Silver Nanoparticles 62 4.5. Paper 5 - Humidity Influence CNF-based Organic Electronics 71

Chapter 5. Conclusions 77

Chapter 6. Outlook 79

Acknowledgements 80

Bibliography 83

Part II - Papers

Paper 1. Revealing Structural Evolution Occurring From

Photo-Initiated Polymer Network Formation 97

Paper 2. Core–Shell Nanoparticle Interface and Wetting

Properties 119

Paper 3. Water-Induced Structural Rearrangements on the Nanoscale in Ultrathin Nanocellulose Films 143 Paper 4. Nanocellulose-Assisted Thermally-Induced Growth of

Silver Nanoparticles for Optical Applications 175 Paper 5. Humidity-induced Nanoscale Restructuring in

PEDOT:PSS and Cellulose reinforced Bio-based Organic

Electronics 201

Part I

Chapter 1

Introduction

In times where the world is changing due to the effects of global warming, we need to take action and mitigate. Thus research is imperative to address these developments and new ideas emerge to handle the human-made environmental crisis.[1] Nowadays fossil-based synthetic polymers are found in the majority of our daily products.[2] However, there are means to improve by reducing and by completely replacing these fossil-based materials with sustainable alternatives.[3] A significant research effort has been performed over the last decade to facilitate the use of wood-based nanocellulose as a generic nanoparticle with remarkable physical properties.[4–7] The ability to deconstruct wood into the fundamental components such as lignin, hemicellulose and nanocellulose, offer multiple platforms for new material and liquid concepts, where applications range from biofuels to complex nanocomposites with controlled anisotropic properties.[8–10] The full understanding of the fundamental structure and composition of wood offers the ability to mimic nature, and reassembly of nature’s nanoparticles in different ways to be used in various application areas.[11] Although the use might be simple, the understanding is complex since these nature-based materials are highly heterogeneous. Hence, this PhD thesis aims to develop methodologies for studies of nanocellulose, and also more generally soft matter materials for sustainable applications.

The assembly of materials and the resulting physical properties can always be related to the hierarchical structure and thus to the nanoscale.[12] In 1914 Max von Laue was awarded the Nobel prize in physics for the discovery of X-ray diffraction caused by crystals.[13] In the year after, the scientists Bragg (father and son) were awarded the physics Nobel prize for the X-rays structure determination of how crystal planes are assembled.[13] Their methodologies are widely applied today to study any kind of materials from rocks to proteins and can be used to study not only crystalline but also amorphous materials with respect to resolving nanoscale packing and to follow and further predict the macroscopic properties.[14, 15] In 1964 Dorothy C. Hodgkin received the chemistry Nobel prize for the structural determination of biochemicals (namely penicillin and vitamin B12) by X-rays.[13] Her work also showed that X-ray structural determination needs to be complemented with competences and methods from other areas to prepare samples, perform experiments and to

understand the overall effects on the structure, thus often relying on close collaborations. Such collaborations have fostered the growth of the scattering community over the years, which in turn has resulted in this PhD thesis, were physicists, chemists and engineers have worked in a team sharing common goals.

For structural determination, it is essential to understand how X-rays interact with matter. X-rays consist of photons that interact with the electron cloud around the atomic nucleus, and the scattering depends on the number of electrons.[16] Thus, when studying lighter elements such as hydrogen or carbon, it can be hard to achieve contrast using X-rays. However, as a further development, anomalous scattering addressing the element-specific absorption edges can be used.

In 1994 the physics Nobel prize was awarded Bertram N. Brockhouse and Clifford G. Shull for the development of neutron scattering and diffraction techniques.[13] In contrast to X-rays, the neutron interacts with the nucleus itself. The scattering is as a result of this not depending on the number of electrons, and the contrast can be tailored to study, e.g. complex polymeric composite materials. Both X-rays and neutrons have a wavelength similar to inter-atomic distances, which allows us to probe the atomic structure. Additionally, neutron energies are in the region of elementary excitation and allow studies of dynamics in materials.[16, 17] Neutrons are also used due to their weak interactions, and high transparency in shielded sample environments were X-rays cannot penetrate. This gives possibilities to study materials, e.g. under pressure, chemical environments or inside devices during operations.

Our smart society started in the early 21st _{century with, e.g. smartphones,}

fitness trackers, small batteries and flexible solar cells which is a part of the future for our society. Regarding the development of these technologies, the goal is on one hand to support the development of new applications, and on the other hand to increase sustainability by reusing, reducing and replacing of fossil-based materials. Therefore, the design and fabrication of such devices must be reviewed, addressing a wide range of research ranging from catalysis, photo-voltaics, photonics to thermoelectrics.[18–20] Within this PhD thesis, grazing incidence neutron and X-ray scattering techniques have been applied to study the evolution of buried structural features in situ (”live”) during processing. These techniques are crucial for studying the nanoscale evolution of materials in situ during, e.g. external triggers, growth, operation and degradation in thin films (< 500 nm).[21] Grazing incidence small-angle X-ray scattering (GISAXS) allows time-resolved measurements were growth dynamics on surfaces can be an-alyzed.[22] The technique is applied to study fabrication methods as sputtering, spin coating and spray deposition or in operando measurements in organic solar cells and sensors.[23–28] As a complement, grazing incidence small-angle neu-tron scattering (GISANS) has the potential of providing contrast enhancement routes to facilitate studies of polymer micelle formation or water permeation in materials.[27, 29] Additionally, time-of-flight neutron scattering can be used to

1. Introduction 3 study depth-dependent morphology simultaneously allowing measurements of morphological kinetics at different depths of different materials.[30–32]

Within the first project of this thesis, GISAXS was used in combination with different laboratory-scale measurements to correlate time-dependent structure morphology to polymerization effects. The polymerization conversion was correlated to structures in the material which grow during UV polymerization. We observed how the polymer resin mixture contributed to the film formation and how structural sizes are depending on the initial radius of gyration and glass-transition temperature.[33]

In the second project, spray deposited latex nanoparticles were studied under thermal annealing on silicon surfaces by X-ray scattering. Latex nanoparticles of different sizes and formulations were distributed with low surface coverage on the surface, and the wettability was correlated to the thermally tailored nanoparticle structure. We speculate here that the polymer chains align in such a way that the latex core polymer contributes predominantly to the wetting after highest temperature annealing.[28]

In the third project, surface charge modified nanocellulose was spray de-posited on a model substrate to study the impact of surface charge on the self-assembly. GISAXS was used in combination with atomic force microscopy to correlate internal morphology and the surface topography. Wettability studies showed that with increasing surface charge, the contact angle could be tailored and the surface roughness tuned. GISANS experiments on a selected sample showed that humidity swells and changes the internal structure fully reversible inside the thin film. [27]

The fourth project used the nanocellulose template from project three, onto which silver nitrate was deposited as a silver precursor ink on top. X-ray scattering and diffraction were used to determine the structure evolution of the ink during thermal decomposition of the precursor. The silver salt degraded to metallic silver, and the size distribution was followed during the thermal annealing. In situ UV-Vis spectroscopy measurements revealed the plasmonic behaviour of the nanoparticles created on the surface. We concluded that the thermal assisted growth on cellulose is beneficial and could lower the use of silver for optical applications. The in situ study further revealed that the energy band-gap of the resulting silver nanoparticles could be tailored and so allow for potential use in organic solar cells.

In the fifth project, the nanoporosity of the nanocellulose template was used to permeate a conductive polymer inside the thin film. This model system is used to study the impact of the stiff and rigid nanocellulose fibres on the bionanocomposite under cyclic humidity changes. Herein, we used grazing incidence neutron scattering and reflectometry in combination with conductivity measurements to understand the structural evolution in changing environments on the conductivity. This study revealed that previously shown fully reversible cyclability of the pure nanocellulose film does not apply to the composite material. Here, the water reassembles the morphology and so changes the

overall conductivity of the thin film. This knowledge is crucial and needs to be considered when such material components are to be used in applications exposed to variations in humidity such as e.g. organic solar cells.

Chapter 2

Scientific Context

This chapter covers the scientific aspects of this PhD thesis. Spray and spin coating deposition techniques are introduced followed by thin film characteriza-tion techniques. This development of characterizacharacteriza-tion the methodologies for bio-based thin films is one of the main aims of this thesis, hence this section is quite extensive. It covers commonly used surface characterization techniques such as reflectometry, small- and wide-angle scattering as well as their corre-sponding grazing incidence implementations. Complementary techniques as atomic force microscopy (AFM), scanning electron microscopy (SEM), UV-Vis spectroscopy and contact angle measurements are also described.

2.1. Deposition Techniques

2.1.1. Spray Deposition

Sprays or atomization is a process where a liquid is forced through a small orifice generating a stream of droplets.[34] Sprays are important for many application ranging from combustion processes, air condition, drying, fire suppression, to coatings.[34, 35] The droplet size can be tailored for the different applications from sub-micrometer to millimeter sizes.[36] Within this thesis small droplets sizes of up to 5 µm were used, and by spraying differently sized droplets surfaces can be nanostructured, see Fig. 2.1.[27, 35, 37] As a deposition technique spraying is widely applied in industry as well as in laboratories for research and development.[34, 36, 38] Another advantage of spray deposition is that it can produce large-scale homogeneous films, especially were techniques like dip coating fail due to their long contact times with the coated surface.[27, 39] Furthermore, is is a non-equilibrium layering technique which can give very low roughness stratified films.[40] It has been shown that spray deposition can be used to fabricate sensors [41], transistors [42], photo diodes [43, 44], organic solar cells [45, 46] and supercapacitors.[47] X-ray scattering has also been applied to study spray deposition of colloidal polystyrene nanoparticles and thickness studies of bacterial cellulose.[35, 48–50] Template assisted spray deposition, or spraying at a grazing incidence angle can be used to control alignment of anisotropic nanoparticles as nanocellulose or silver nanowires.[51–53]

Figure 2.1: Schematic of the spray or atomization process. A liquid jet is ejected from a small orifice and breaks up in three regimes: dense regime (water jet), dilute regime (heterogeneous droplet sizes), very dilute regime (fog, homogeneous small droplets). All experiments within this thesis are performed in the later regime.

2.1.2. Spin Coating

Spin coating is a deposition technique were a liquid film is spread on a rotating surface by the centrifugal force induced by rotation. Due to its low complexity and low cost it has been widely applied by industry. Specifically, it is used by the semiconductor industry to apply photo resists prior to lithography, where the possibility to tune the thickness simply by the rotation speed is a great advantage over many other methods.[54–56] Spin coating can also be used to level out uneven surfaces.[57] However, spin coated films may show striation patterns caused by thickness fluctuations induced by the solvent.[58] Hence, the solvent selection is crucial for successful deposition, and if wisely chosen, also allows layering of nanoparticles or polymers within the deposited liquid.[59] Off-centered spin-coating has been used to deposit nanostructured films with internal gradients, e.g. for transistors.[60, 61] The drying and self-assembly has been studied using in situ surface-sensitive X-ray scattering to yield the inner morphology during the process.[62] Recent developments show that spin coating can also be used for epitaxial film growth were the crystal substrate defines the in-plane and out-of-plane orientations.[63] In comparison of these techniques (spray and spin coating) they both yield highly stratified films on a time scale of seconds, nanostructuring can be done with well-defined droplets or off-centering of the sampling surface. However, there are two obvious drawbacks with spin coating compared to spray coating, first it is a batch process where the size of substrates which are in the centimetre range, limiting the applicability for large-scale surfaces. Secondly, for spin coating only flat substrates can be coated

2.2. Thin Film Characterization 7 whereas for spray deposition non-level 3D shape can be coated.[34, 36, 38, 55, 64]

2.2. Thin Film Characterization

To understand material properties on the macroscale, typically the hierarchical structure from the nanoscale to microscale needs to be understood. This can be achived by adoption of measurement techniques to extract the physical properties addressed. The available techniques vary massively in complexity, ranging from optical microscopy mapping real space to scattering methods where the reciprocal space is mapped and evaluated. To obtain the full picture of the studied material it is typically necessary to use complementary tools to gain the full picture and to correlate the hierarchical structure from the ˚Angstr¨om level up to the macroscopic physical properties. Typical lab-based tools such as atomic force microscopy (AFM) and scanning electron microscopy (SEM) are often needed to pre-evaluate samples to facilitate the understanding that can be gained by analysing the reciprocal space in neutron or X-ray scattering experiments. Within this thesis the interaction between waves (or particles) with matter is studied, using their scattering, absorption, refraction and reflection properties. This yields depth dependent knowledge of the materials density and morphology, which is correlated with optical and other functional properties such as wettability.

2.2.1. Neutron/X-ray Interaction with Matter 2.2.1.1. Scattering

Scattering effects are present for all interaction between travelling particles (also described as travelling waves) interacting with matter. For studying the nanoscale of materials, X-rays have a long-standing history and are widely accessible in labs and large-scale facilities. X-rays can be expressed as electro-magnetic waves with wavelength, λ, in the ˚Angstr¨om-nanometer region, which is related to the wave number k = 2π_λ. The relation between the wavelength λ and the photon energy E in keV is given by

λ[˚A] =hc E =

12.398

E[keV] (2.1)

In classical physics, one assumes a planar wave impinging on atoms or matter. The planar wave is than scattered, resulting in a spherical wave. For X-rays, photons interact with the electron cloud of atoms, giving radially irradiated the scattered wave. Contrary to X-rays, neutrons are not scattered by the electron cloud but instead interact with the nuclei. Neutrons nuclear particle have a mass of mn = 1.675 · 10−27kg and a lifetime of tn = 886 s.

Furthermore, the neutron is electrically neutral but has a magnetic moment of µ = γµn with γ = −1.913 beeing the gyromagnetic ratio and the nuclear

antiparallel to its s = 1/2 spin, and the de-Broglie wavelength of a moving neutron with velocity v can be described as λ = _m2π~

nv. The wave number k is calculated similarly as for X-rays, the wave vector ~kn = m_~n~v. The kinetic

energy of the neutrons can be derived as En = ~

2_k2

2mn

(2.2) Typical energy ranges and their corresponding names for X-rays and neu-trons can be found in Table 2.1. For this thesis project cold neuneu-trons as well as hard X-rays were used to probe material structure.

Table 2.1: Representation of the energy and wavelength regimes and name conventions for neutrons and X-rays. Adapted and combined.[16, 17, 65]

Neutrons Wavelength [˚A] Energy [meV] Ultra cold > 40 < 0.05

Cold 2.4 − 40 0.05 − 14

Thermal 0.6 − 2.4 14 − 200

Hot 0.3 − 0.6 200 − 1000

Epithermal < 0.3 1000 − 1 · 106

X-rays Wavelength [˚A] Energy [keV] Soft ≈ 120 − 2.5 ≈ 0.1 − 5 Hard ≈ 2.5 − 0.05 ≈ 5 − 250

The scattering process is referred to as elastic if the incident and exiting wavelengths of the wave are equal. In the quantum mechanical description, the incoming wave has a momentum and an energy, which can be transferred to the scattering object, resulting in that the exiting wave has a lower energy and thus longer wavelength. This is referred to as inelastic scattering, e.g. Compton scattering.[16]

The differential scattering cross-section dσ_dΩ
describes incident photons or neutrons Φ0passing through a unit area per second are scattering under a

solid angle ∆Ω, and how many photons or neutrons that can be detected at a distance R on the detector Iscattered

 dσ dΩ  = Iscattered Φ0∆Ω (2.3) The total cross-section is the integral over all scattering angles defined as σX,total= 4πr20· (2/3) = 0.665 · 10−24cm2. In case of neutrons the differential

cross section of one nucleus can be written as dσ dΩ
= b

2

j and the total scattering

cross section σn,total= 4πb2j.[66]

The change in phase after the scattering event in the origin and in the distance ~r using ~k and ~k0 _{for the wavevector before and after the scattering}

2.2. Thin Film Characterization 9

∆φ(~r) = (~k − ~k0_{) · ~r = ~q · ~r (2.4)}

here ~q = ~k − ~k0 _{is the scattering vector. In the elastic scattering event}

|~k| = |~k0_{| so that the scattering vector simplifies to}

~

q = 2|~k| sin θ = 4π

λ · sin θ (2.5)

The volume element d~r at ~r adds −r0ρ(~r)d~r to the scattered field with the

phase factor ei~q·~r and thus the scattering length of the single atom is defined as − r0f0(~q) = −r0

Z

ρ(~r)ei~q·~rd~r (2.6) where f0(~q) is the atomic form factor and r0= e

2

4π0mc2 = 2.82 × 10

−5_{˚A =}

2.82 fm the radius of the electron. In the limit ~q → 0 the number of electrons in the atom Z can be deduced from the atomic form factor f0_{(~q → 0) = Z. The}

scattering length density for X-rays from an atom with atomic number Z is simply Z ∗ r0. The contrast between different materials can then be directly

related to the number of electrons for X-rays, whereas for neutrons this is completely different, see Figure 2.2. In the far-field, when ~q → ∞, the scattering elements in the volume diverge in phase and so the atomic form factor goes to zero f0_{(~q = ∞) = 0. The presented classical description until here does not}

describe the full process of the scattering event. The electrons are bound in the atom and have discrete energy levels. Resonant scattering occurs when the incident energy matches the element specific atomic absorption edges. When the energy is much higher than these edges the atomic form factor is f0 = 0. The complex form factor if00 includes the phase differences, so the dissipation of energy and the absorption of the media.

Hence, combining these the atomic form factor can be written as with emphasis on the energy E dependency

f (~q, E) = f0(~q) + f0(E) + if00(E) (2.7) The energy dependency play a role when the energy of the impinging X-rays is equal to or below the binding energy of the electrons, altering the scattering cross-section. In case of hard X-rays, this is usually constant and can be neglected in the following. In case of molecules, the scattering based on multiple atoms must be evaluated as the sum over all involved atomic form factors (j)

Fmolecule(~q) =X

j

fj(~q)ei~q·~rj (2.8)

Crystalline materials show periodic features as described by Bragg’s law

Figure 2.2: Scattering lengths for coherent bc (red circles) and incoherent bi

(black squares) neutron scattering events and the atomic number Z (blue line) which is proportional to the X-ray scattering length. Highlighted is protium (1

1H, cyan) and deuterium (D =21H, magenta).[67]

where n is the number of constructive interferences, θ is the angle of incidence and d is the distance between the lattice planes. This equation does allow the retrieval of the position of the constructive interference but not the intensity. Combining the lattice properties and the molecule form factor one can find a solution for the crystal form factor

Fcrystal(~q) =X j fj(~q)ei~q·~rj · X n ei~q· ~Rn _(2.10) here the first term denotes the unit cell structure factor and the second term the lattice sum and Rn are the lattice vectors that define the lattice. The

lattice sum P

ne

i~q· ~Rn _{are phase factors on the unit circle in the complex plane.} The lattice sum only diverges if the scattering vector follows

~

2.2. Thin Film Characterization 11 where the integer is N the number of unit cells. The lattice vector ~Rnis defined

as

~

Rn= n1~a1+ n2~a2+ n3~a3 (2.12)

where ~aiis the lattice base-vectors and ~niare integers. Using the reciprocal

lattice and the corresponding lattice vectors

~a∗₁= 2π a~2× ~a3 ~ a1· ( ~a2× ~a3) , ~a∗₂= 2π a~3× ~a1 ~ a1· ( ~a2× ~a3) , ~a∗₃= 2π a~1× ~a2 ~ a1· ( ~a2× ~a3) (2.13) hence the reciprocal space is defined by:

~

G = h~a∗₁+ k~a∗₂+ l~a∗₃ (2.14) Where h, k, l are integers called Miller indices. By combining the real space lattice vector in equation 2.12 and the reciprocal space lattice vector in equation 2.14 one finds

~

G · ~Rn= 2π(hn1+ kn2+ ln3) = 2π · integer (2.15)

Hence ~q = ~G, which shows that scattering only occurs if ~q is equal to a reciprocal lattice vector. This prerequisite is also called the Laue condition, which is equivalent to Bragg’s law in scattering. This means that crystals do scatter solely in distinct regions in the reciprocal space, which allows a complete determination of the crystal structure. This feature is used in diffraction or wide-angle scattering. It should be noted that only single scattering events have been allowed here, i.e. no multiple scattering within the crystal. This is called the kinematic Born approximation.[16, 68]

2.2.1.2. Absorption

Absorption by a material takes place when the impinging wave (particle) gets absorbed by an atom and the energy is transferred to an electron that leaves the now ionized atom. This process is referred to as photoelectric absorption process described for the case of light by Heinrich Hertz in 1887 and mathematically explained by Albert Einstein using the particle description of light.[69, 70] The absorption can be described by a linear absorption coefficient µ, and by definition the µdz is the attenuation in an infinitesimal thick sheet dz in the thickness direction z. The intensity can than be described as

− dI = I(z)µdz (2.16)

which can be written in differential form dI

I(z) = −µdz. (2.17)

By requiring that I(z = 0) = I0 this yields the Beer-Lambert law

The absorption coefficient is proportional to the atomic number density ρatom

and to the absorption cross-section σabs

µ = ρatom· σabs=

 ρmassNA

M 

· σabs (2.19)

using the Avogadro number NA, the mass density ρmass and the molar

mass M . When an inner atomic electron is ejected from the atom it leaves a hole in the shell which is subsequently filled with an electron from an outer shell by emission of an photon with energy equal to the difference between the shells. This process is known as fluorescence. Another process to fill the hole is by a so-called Auger electron. The Auger electron gets released from the atom when the energy of the filled first electron hole is not emitted as photon (fluorescence) but released as free electron. The latter processes are not part of this thesis work.

2.2.1.3. Reflectometry

As previously discussed X-ray photons can interact with matter and they can refract at the interfaces of different media. The refractive index for X-rays is defined as

n = 1 − δ + iβ. (2.20)

By definition the refractive index in vacuum is nvacuum= 1 and usually ∆n ≈

10−6 which does implies very small refraction angles. As the real part of the equation 2.20 is less than unity, this means that the phase velocity is larger than the velocity of light in the media. However, the law of relativity is not violated as the group velocity, which contains the information, still travels slower than the speed of light.

The refractive index for neutrons n ≈ 1 −λ2

2πρ~bc, for X-rays ρ~bc is given by

electron density multiplied by classic electron radius ρer0.

δ is the real dispersive and β the imaginary absorptive term. It can be written in the form of the atomic form factors

δ = 2πρatomf

0_{(q → 0)r} 0

k2 (2.21)

the wavevector k = 2π/λ shows the dependency of the incoming X-ray wave-length and the atomic form factor as described above f0_{(q → 0) = Z. The}

absorptive term β is defined as

β = µ/2k (2.22)

were µ is the absorption coefficient as defined in equation 2.19. Following Snell’s law (Willebrord Snellius (1580–1626)), it is possible to relate the incident grazing angle α to the refracted grazing angle α0 as

cos α = n cos α0 (2.23)

and setting α = αc yields total external reflection, α0= 0◦. It should be noted

2.2. Thin Film Characterization 13 against the sample surface. As the refracted angles are small, the cosine can be simplified using a taylor series expansion, which together with equation 2.20 and by neglecting the absorption leads to

αc= √ 2δ = √ 4πρr0 k (2.24)

The critical angle ends up in the range 0.01◦ to 0.5◦ in the hard X-ray regime of in the range of 0.5◦ to 1◦ with cold neutrons.[16, 21, 27, 35, 71] The total external reflection occurring when the incident angle is below this critical angle leads to an evanescent wave propagating parallel to the interface in the refracted medium. The intensity decays on the nanometer scale in the medium. However, for the case when the incident angle is above the critical angle, the material penetration depth can be several micrometers, see Figure 2.3. This feature is not an abrupt change in penetration depth but is material dependent and highly tunable, which provides surface sensitivity or bulk sensitivity. This is central for methods such as grazing incidence small-angle scattering.

The incident, reflected and transmitted wavevector, ~kI, ~kR, ~kT, with

corre-sponding amplitudes aI, aR, aT needs to be evaluated to gain full knowledge

on the reflected intensity. Snell’s law and Fresnel equations must be derived requiring that the wave and its derivative is continuous at the interface z = 0. Hence, the amplitudes and their corresponding wavevector can be written as

aI+ aR= aT, aI~kI+ aR~kR= aT~kT (2.25)

by applying that the wavenumber in vacuum is given by k = |~kI| = |~kR| and

inside the medium as nk = |~kT|. Considering the components of ~k parallel and

perpendicular to the surface this yields

aIk cos αi+ aRk cos αi= aT(nk) cos αf (2.26)

− (aI− aR)k sin αi= −aT(nk) sin αf (2.27)

were αi and αf are the incoming and exit angles from the surface respectively.

Using Snell’s law from equation 2.23 and the critical angle from equation 2.24, again expanding the cosine due to the small angle approximation, we get

α2i = α2f+ 2δ − 2iβ = α2f+ α2c− 2iβ. (2.28)

By combining now equation 2.25 and equation 2.27 we obtain neglecting absorbtion β aI− aR aI+ aR = nsin αf sin αi ≈ αf αi (2.29) From this, the Fresnel equation for the reflectivity amplitude r and the transmittivity amplitude t can be written as

r ≡aR aI = αi− αf αi+ αf , t ≡aT aI = 2αi αi+ αf (2.30)

with the corresponding intensities are

R = |r|2, T = |t|2 (2.31)

The penetration depth Λ where the intensity drops with 1/e of the X-rays in the material can be expressed after Dosch et al.[72] as

Λ =√ λ 2π(li+ lf ) (2.32) with li,f = v u u t(α2 c− α2i,f+ s  (α2 i,f− α2c)2+  λµ 2π  (2.33) where µ is the absorption coefficient and λ the X-ray wavelength.

Using the wave vector one can rewrite the Fresnel coefficients in a more useful way for reflection and for performing scattering experiments using

Q = 2k sin α ≈ 2kα (2.34)

Qc= 2k sin αc≈ 2kαc (2.35)

and the dimensionless value for the case of only one incoming beam can be written as q ≡ Q Q0 ≈ 2k Qc α (2.36)

The Fresnel coefficients can then be written as r(q) = q − q 0 q + q0, t(q) = 2q q + q0 (2.37) where Qc= 2k √

2δ. Fresnel reflection at large qz equals to R(qz) ∝ q−4z which

also is known as Porod law. The reflectivity, transmittivity and the penetration depth is plotted in Figure 2.3 for most materials used in this thesis.

The listed Fresnel coefficients are here only evaluated for an single interface. For the case of a multilayer, multiple reflections and transmitted terms must be taken into account. For a single layer thin film the Fresnel reflection can be written as

rsingle=

r0,1+ r1,2p2

1 + r0,1r1,2p2

(2.38) where p is the phase factor that can be derived with p2j = ei∆jQj. In absence of

multiple reflections, the Fresnel coefficient for the reflection and transmission at the interface (j, j + 1) can be written as

rj,j+1= Qj− Qj+1 Qj+ Qj+1 , tj,j+1= 2Qj Qj+ Qj+1 (2.39) To evaluate more than a single layer with two interfaces the exact recursive method by Parratt needs to be introduced, where a stack of N layers is assumed, with the indices of refraction for each layer given by nj = 1 − δj+ iβj, and layer

2.2. Thin Film Characterization 15 thicknesses ∆j respectively, see Figure 2.4. The wave vector transfer for the

j ’th layer can be written as

Qj =

q

Q2_{− 8k}2_δ

j+ i8k2βj (2.40)

Figure 2.3: Fresnel reflectivity R(q), Fresnel transmittivity T (q) and the pene-tration depth Λ calculated for nanocellulose (CNF), silver (Ag), PEDOT:PSS, silica (SiO2), silicon (Si) for hard X-ray with a energy E = 13 keV. Adapted

Layer 0 j …j+1 N-1 N N+1 Interface 0,1 j-1,j … j,j+1 N-1,N N,N+1 j-1 …2 1 1,2 … Δj nj Δ_N nN Δ₁ n₁ qz k0,1 k0,1’ ϴ0 k1,2’ ϴ0 ϴ₁ z x

Figure 2.4: Schematic representation of the reflection on a multilayer material with j layers, with a thickness of each layer ∆j and a corresponding refractive

index of nj.

The reflectivity from the last layer and the substrate can be written as

r0_{N,N +1}= QN − QN +1 Nj+ QN +1

(2.41)

Thus the N ’th layers reflectivity can be derived recursively

rN −1,N =

r_{N −1,N}0 + r_{N,N +1}0 p2_N

1 + r0_{N −1,N}r0_{N,N +1}p2_N (2.42) By using this method, the reflectivity from each individual layer of a stack can be calculated as well as the total reflectivity. However, this calculations do not include roughness or interfaces with gradients, which needs further reading.[16] 2.2.1.4. Small- and Wide-angle Neutron/X-ray Scattering

When performing small-angle X-ray or neutron scattering (SAXS/SANS) studies, only small wave vector or small scattering angles are considered. In the following part the focus is on SAXS, however the differences to SANS will be highlighted.

2.2. Thin Film Characterization 17 The SAXS scattering intensity can be written as

ISAXS(q) = f2 Z V ρatomei~q~rndVn Z V ρatome−i~q~rmdVm =     Z V ρsldei~q~rdV     2 (2.43) where ρsld= f · ρatom is obtained by multiplying by the radius of the electron

r0with the scattering length density (SLD).

For the case of coherent elastic neutron scattering the cross section is defined as dσ dΩ|coh.elas. =       X j bjei~q~rj       2 (2.44)

where rj is the nuclear coordinates. The scattering sum can be evaluated as an

integral X j bjei~q~rj ≈ Z V ρbei~q~rdV (2.45)

with the neutron SLD defined as the sum over coherent scattering length densities bj in a given volume V0. The unit cell in diffraction is defined as

ρb= 1 V0 X j∈V0 bj (2.46)

giving the SANS cross section dσ dΩ|SAN S = N S(~q)     Z V ρbei~q~rdV     2 (2.47) where N the number of identical nanoparticles and S(~q) the structure factor.[17, 65]

For cases were nanoparticles do not have any inter-particle interactions, see Figure 2.6, one can study the SAXS intensity in the very dilute regime by subtracting the intensity of the particle and the solvent as follows

ISAXS(q) = (ρsld,part.− ρsld,solv.)2

    Z V ei~q~rdV     2 = ∆ρ2V_p2|F (~q)|2, with F (~q) = V_p−1 Z Vp ei ~Q~rdVp (2.48)

where ∆ρ = (ρsld,part.−ρsld,solv.) is the SLD difference between the nanoparticles

(part.) and the surrounding solvent (solv.). The introduced particle form-factor F (~q) depends on the particle morphology such as size and shape, see section 2.2.1.1. Analytic solutions can be derived for simple structures. Spherical and cylindrical form-factors are used in this thesis and are therefore presented in

the following. The spherical form factor can be derived as F (~q) = Vp−1 Z Vp ei ~Q~rdVp= Vp−1 Z R 0 Z 2π 0 Z π 0 ei ~Q~rsin θr2dθdφdr = 4π Vp Z R 0 sin qr qr r 2_{dr = 3} sin qr − qr cos qr q3_r3  ≡ 3J1(qr) qr (2.49)

with J1 the first order Bessel function and R the particle radius. An oriented

cylinder form factor can be written after Guinier as,[73] F (q, β) = 2(∆ρ)Vc sin (1 2qL cos β) 1 2qL cos β ·J1(qR sin β) qR sin β (2.50)

where the volume of the cylinder Vc= πR2L is spanned by the length L and

the radius R. In case of randomly oriented cylinders the form factor needs to be adjusted to

F2(q) = Z π/2

0

F2(q, β)sin(β)dβ (2.51)

For cases of unknown particle form-factors one usually start by observing qr → 0, where the so-called Guinier analysis can be applied.[73] From the decrease of the intensity in this regime the size of arbitrary shaped particles can be estimated. As a result the radius of gyration or Guinier radius is introduced

R2_G= V_p−1 Z

Vp

r2dVp (2.52)

For the case of spherical particles, the Guinier radius can be directly be retrieved from the SAXS intensity as

ISAXS(q) ≈ ∆ρ2V_p2e−q2R2G/3 _(2.53) So far only the dilute particle regimes have been considered, without inter-particle interaction. When these are present the so-called structure factor S(q) has to be considered in the derivation in addition to the form factor F (q), for the particles with the intensity written as

ISAXS(q) = ∆ρ2V_p2|F (~q)|2S(q) (2.54) Many models can be defined for the structure factor, but it should be noted that if q → ∞ the structure factor is S(q) → 1.[74–77] There are good compilations of structure factors in literature.[78, 79] Within the scope of this thesis a one-dimensional and a two-one-dimensional paracrystalline model have been applied. The inter-particle interactions can have on a regular lattice long range order as e.g. in crystals. In case of self-assembled materials mostly the long-range order is gradually distorted with distance which can be described with a paracrystalline model, see Figure 2.5. The probability to find a particle at a position on the lattice vector a or b is defined by the corresponding probability distribution Pa

2.2. Thin Film Characterization 19

Figure 2.5: Schematic representation of the one-dimensional and two-dimensional paracrystal. The one-dimensional paracrystal is marked with the dashed red box, the full figure depicts the two-dimensional paracrystal. The circles represent the area where the probability (P ) to find particles is maximum. Adapted from R´emi Lazzari [79].

Z Pa,b(~r)d2r = 1, Z ~ rPa(~r)d2r = ~a, Z ~rPb(~r)d2r = ~b (2.55)

If we define the Fourier transformation of the probability distribution as Pa,band assume that the positions in a and b are independent the structure

factor S(~q) can be written as S(qxy) = Y a,b real 1 + Pa,b(qxy) 1 − Pa,b(qxy)  (2.56)

The probability distributions are commonly described mathematically with Gaussian, Lorentzian or pseudo-Voigt functions.

In Table 2.2 an attempt has been made to categorize the ranges in which the scattering is referred as SAXS/SANS, WAXS/WANS or even ultra-small-angle X-ray/neutron scattering (USAXS/USANS). These ranges are equivalent for grazing incidence techniques.

2.2.1.5. Grazing Incidence Small-angle Neutron/X-ray Scattering

Grazing incidence small-angle neutron/X-ray scattering, GISANS or GISAXS, is a methodology to study thin films on the same length-scales as when using SAXS or SANS at higher resolutions.[71, 81–83] Levine et al.were the first

Table 2.2: Different scattering techniques as function of sample-to-detector distance SDD and their corresponding approximate resolved structural size. Adapted from Ehlers et al.[80]

Acronym SDD [mm] Structure resolution WAXS/WANS ≈ 10 − 500 ˚_{Angstr¨om scale} SAXS/SANS ≈ 500 − 5000 ≈ 3 − 200 nm USAXS/USANS ≈ 5000 − 20000 ≈ 200 nm − 5 µm

ones applying GISAXS on thin films to study growth phenomena and self-assembly on a surface.[84, 85] As a general methodology these techniques are commonly referred to as grazing incidence small-angle scattering or GISAS, which combines small-angle scattering, the diffuse scattering of reflectometry and grazing incidence diffraction. The measurements are conducted close to the material dependent critical angle as introduced in equation 2.24 and include the SAXS or SANS typical form and structure factors. This technique allows to study the material self-assembly specifically on surfaces or in buried structures.

2ϴ

q

𝒌

𝒌

Direct

beamstop

q

q

Figure 2.6: Schematic SAXS or SANS geometry to study materials in trans-mission mode. In the shown example highly diluted gold nanoparticles with a size of 100 nm in water are measured at P03/MiNaXS beamline with a SDD = (8977 ± 2) mm and a energy of E = 13.1 keV. The schematics spans the wavevectors ~k and the corresponding ~q and the scattering angle 2Θ.

2.2. Thin Film Characterization 21 For the case of thin films, the full distorted wave Born approximation (DWBA) has to be applied to understand the measured scattering pattern. [35, 86–88]

GISANS is highly complementary to GISAXS measurements even though the relatively low neutron flux results in that only very slow time-dependent changes can be achieved. However, the neutrons allows the use of deuterium or a wise choice of the SLD of the involved materials to study complex multi-structured materials with even higher precision.[89] For example, time-of-flight GISANS (TOF-GISANS) can be used to study the polymer infiltration of pho-tovoltaic films by tracking selected penetration depths Λ using a full wavelength band, with specificity to the introduced polymers.[32] This method of depth sensitive TOF-GISANS and TOF-NR has further elucidated in literature.[30, 31, 90–93]

For grazing incidence methods such as GISAXS and GISANS, as well as for the case of grazing incidence wide-angle neutron/X-ray scattering (GI-WANS/GIWAXS), X-rays or neutrons interact with the surfaces of the samples at grazing incidence angles αi< 1◦. The scattered intensity is measured with

one or multiple two-dimensional detectors to access a wide range of exit angles αf (qz, out-of-plane) and lateral angles 2θ (qy, in-plane), see Figure 2.8.[71]

The x-axis is defined as parallel to the X-ray beam, the y-axis is perpendicular to the impinging X-ray beam along the sample surface and the z-axis is normal to the sample surface. For the grazing incidence geometry, the wave vector introduced in equation 2.4 is defined as

~ q =   qx qy qz  = ~k − ~k0= 2π λ  

cos αfcos 2θ − cos αi

cos αfsin 2θ

sin αi+ sin αf



 (2.57)

and the occurring scattering is based on differences in the refractive index as in equation 2.20. In the GISANS/GISAXS scattering pattern one can find the so-called Yoneda region αY onat the critical angle of the studied material or at

multiple positions for composites.[94] The position is found on the detector at αY on= αi+ αc relative to the direct beam on the detector.[71] The critical

angle is defined in equation 2.24. The reflected beam is found at αi = αf

however due to its high intensity this peak is mostly shielded, especially when the angles are very close to the critical angle as this implies total reflection.

This high intensity could destroy the detector. Hence, the rest of the scattering pattern is diffuse scattering analyzed used the above mentioned DWBA, see Figure 2.7. The form and structure factors, as introduced in section 2.2.1.4 still apply, however here they only include a single scattering event in the kinematic Born approximation. For the case of surface scattering the beam can be reflected before and/or after the scattering event, leading to three additional terms that needs to be considered when analyzing GISANS/GISAXS data. The form factor must hence be adapted to the DWBA by weighing the events with

their corresponding Fresnel coefficients by F (qxy) = F q q2 x+ q2y, qz  + R(αi)F q q2 x+ qy2, ( ~k0+ ~k)z  + R(αf)F q q2 x+ qy2, −( ~k0+ ~k)z  + R(αi)R(αf)F q q2 x+ q2y, qz  (2.58) In the far field were αi, αf >> αc the Fresnel coefficient R(αi) = R(αf) = 0,

hence the kinematic Born approximation is retained.

Figure 2.7: Representation of the four scattering and reflection events on particles on a surface within the DWBA. The four terms describe the wave vector qz transfer in z-direction. Adapted from R´emi Lazzari [79].

2.2.1.6. Simulation

Given the DWBA the analysis of GISANS or GISAXS data is mathematically complex and many assumptions are made during the analysis. There have been notable software developments aiming at partially or fully solving of the scattering patterns, numerically as well as analytically. IsGISAXS published in 2002 by R´emi Lazzari was the first to make GISAXS more accessible, followed by F itGISAXS. However these programs did not integrate all features that the science regarding complex materials science demands today.[79, 95] Recent analysis packages are BornAgain and Xi-Cam which both continuously evolve and adds on more features for the direct analysis of GISAXS, XRR and GIWAXS at the same time.[96, 97] BornAgain also includes spin scattering functionality for neutron scattering experiments.[97] The use of neural networks for image recognition and in scattering community the analysis of scattering patterns is also evolving quickly.[98] For Paper 4 the scattering patterns were analysed using the model we used in Paper 2-3 but also using the software package BornAgain running in script mode on the Maxwell super computer cluster at DESY (Hamburg, Germany), see section 4.4.

2.2.1.7. X-ray and Neutron Sources

Wilhelm Conrad R¨ontgen discovered 1895 X-rays while he was examining radia-tion from an electron discharge tube and detecting radiaradia-tion on a fluorescent screen. One of his first experiments was to study the absorption of the newly discovered X-rays by different materials and he found wood and paper being transparent compared to metal.[16] X-rays can be produced in different ways

2.2. Thin Film Characterization 23 commonly in the laboratory by the Coolidge tube or the rotating anode. In both cases electrons impinge on a cooled metal surface and are either deceler-ated which results in the continuous Bremsstrahlung spectrum or the discrete fluorescent radiation (Kα, Kβ). In cases where a monochromatic beam is used

one selects the fluorescent peaks. Here, the impinging electron collides with an inner electron of an atom and ejects it, which is followed by the relaxation of an outer electron to this free position resulting in element specific fluorescent peaks. Kα defines transition between L and K shell and Kβ between M and K

shell. Hence, the anode material defines the X-ray energy commonly used are Ag (Kα= 0.59 ˚A), Mo (Kα= 0.71 ˚A), Cu (Kα= 1.54 ˚A), Co (Kα= 1.79 ˚A), Fe

(Kα= 1.94 ˚A), Cr (Kα= 2.29 ˚A).

For experimental use a small beam with low divergence is usually needed, and for the cases of the Coolidge tube or the rotating anode, only a small fraction of the X-rays can be used for this giving low X-ray intensities. Especially for in situ and/or in operando measurements the requirements on beam quality such as intensity, divergence and energy tunability are high. All these required

2ϴf ɑf ɑi y z x SDDGIW AX/NS SDDGISAX/ NS Detector gaps Specular beamstop Inaccessible wedge 𝒌𝒊 𝒌𝒇 qz qy

Figure 2.8: Schematic grazing incidence small- and wide-angle scattering. The incident X-ray or neutron beam impinges on the sample surface under oblique angles, reflects, refracts and scatters. In two different sample-to-detector dis-tances (SDD) the intensity is captured for wide-angle and small-angle short and long distance. See for range determination also Table 2.2.