Microfluidics at high-intensity X-ray sources: from microflow chips to microfluidic liquid jet systems

Microfluidics at high-intensity X-ray sources: from microflow chips to microfluidic liquid jet systems DOCTORAL THESIS (Dissertation) to be awarded the degree of Doctor rerum naturalium (Dr. rer. nat.) at the Faculty of Biology, Chemistry and Earth Sciences, University of Bayreuth submitted by Dipl. Chem. Martin Trebbin from Hamburg Bayreuth, 2013 The work described in this thesis was carried out at the Institute of Physical Chemistry at the University of Hamburg (August 2009 to September 2010) and the Department of Physical Chemistry I at the University of Bayreuth (October 2010 to September 2013) under the supervision of Prof. Stephan Förster. This is a full reprint of the dissertation submitted to obtain the academic degree Doctor of Natural Sciences (Dr. rer. nat.) and approved by the Faculty of Biology, Chemistry and Geosciences of the University of Bayreuth. Acting dean: Prof. Dr. Rhett Kempe Thesis submitted: 13.09.2013 Date of Scientific Colloquium: 19.12.2013 Doctoral Commitee: Prof. Dr. Stephan Förster (1st reviewer) Prof. Dr. Andreas Fery (2nd reviewer) Prof. Dr. Hans-Werner Schmidt (Chairman) Prof. Dr. Thomas Scheibel "If you want better results, you have to ask yourself better questions.” - unknown Table of contents Summary…………………………………………………………….….….………….13 Zusammenfassung ………………………………….……………………………. 15 1 Introduction …………………………………………………...…………………….17 1.1 Motivation ………………………………………………………………………….17 1.2 Microfocus X-ray sources ……………………………………………………....17 1.3 Introduction to microfluidics …………………………….……………………. 20 2 Microfluidic devices as experimental platforms for X-ray studies ….. 25 2.1 Introduction ………………………………………………………………………. 25 2.2 X-ray compatible microfluidic device types ……………………………..…. 29 2.3 Outlook: new processing techniques and device materials ……………... 34 2.4 Microfluidic liquid jet devices for next generation X-ray sources ….…. 35 2.5 Conclusions ……………………………………………………………………......39 3 Theoretical Fundamentals ………………………………………………… 41 3.1 Self-organization of amphiphiles ……………………………………….….…. 41 3.2 Fluid dynamics fundamentals for microfluidics …………………………... 43 3.2.1 Navier-Stokes equations ……………………………………………….…… 43 3.2.2 The No-Slip condition ……………………………………………………… 45 3.2.3 Convection and Diffusion …………………………………………………... 46 3.3 Solution of non-linear problems ………………………………………………48 3.3.1 Fluid structure interaction ………………………………….……………….48 3.3.2 Shear forces and Non-Newtonian fluids ………………………….….….… 48 3.3.3 Two-phase systems and liquid jets …………………………………………. 49 3.4 Small angle X-ray scattering ………………………………………………….. 53 3.4.1 Preface ……………………………………….….…….….….……………… 53 3.4.2 Introduction ………………………………………………………………… 53 3.4.3 The scattering vector …………………………………………………..….... 54 3.4.4 The scattering pattern ………………………………………………………. 55 3.4.5 Form factor …………………………………………………….……………. 58 3.4.6 Structure factor ……………………………………………………………... 59 3.4.7 Bragg reflexes ……………………………………………………………….. 60 3.4.9 Model-based order analysis ………………………………………………… 63 3.4.10 Particle orientation distribution …………………………………………... 64 4 Methods and Techniques ………………………………………………….. 71 4.1 Photolithography ………………………………………………………………… 71 4.2 Molding Materials ………………………………………………………………. 75 5 Summary and Strategy …………………………………………………….. 79 quantifiziert. Die Strömungsorientierung von zylindrischen, Wurmund Fibrillen-artigen Strukturen ist von zentraler Bedeutung für die viele Prozesse, wie z. B. der Faserherstellung, den Spritzguss-Verfahren oder dem Fluss von Zellen und Proteinen durch dünne Kapillaren. Ein weiteres Beispiel für geschlossene Kanäle demonstriert die hohe Probeneffizienz mikrofluidischer Rillen-Chips. Diese sind nur wenige Millimeter groß, scher-induzierend und benötigen lediglich 2-5 µl Probe für die Röntgenanalyse eines scher-orientiertern PolymerNanokomposit Materials. Der zweite Typ von Mikrofluidik-Chips basiert auf einer offenen Düsen-Geometrie und produziert Flüssigkeits-Strahlen mit Mikrometer-Druchmessern (0.9 bis 5 µm) bei sehr geringen Flussraten (150 to 1000 µl h-1) unter Normaldruckoder Vakuum-Bedingungen. Die vorgestellten mikrofluidischen Flüssigkeits-Strahl-Chips basieren auf dem Prinzip der gas-dynamischen virtuellen Düse (GDVN), welches der zuverlässigen und verstopfungsfreien Betrieb über lange Zeiten ermöglicht. Darüber hinaus sind diese Chips einfach und mittels softlithographischer Techniken herstellbar, welches ein präzises und reproduzierbares Mikrokanal-Design ermöglicht, dass kritisch für die Optimierung von Flüssigkeits-Strahlen bei kleinen Flussraten ist. Diese Design-Kontrolle wird demonstriert durch die einfache Integration zusätzlicher mikrofluidischer Elemente, wie z. B. ein Strahlim-Strahl-Strömungsfokussierung oder dichter Gruppen vieler Düsen, ohne zusätzliche Herstellungsschritte. Das mikrofluidische Flüssigkkeits-Strahl System wurde ebenfalls mittels einer Reihe mikroskopischer Methoden untersucht (s. oben) sowie CFDSimulationen. Zusammen mit der Variation experimenteller Parameter sowie Kanalgeometrien, haben diese Untersuchungen zu einem besseren Verständnis des fluiddynamsichen Verhaltens von Flüssigkkeitsstrahlen in mikrofluidischen Chips sowie zu der Kontrolle des Strahldruchmessers und des Tropfenabbruch-Typs geführt. Diese erwähnten Vorteile (Zuverlässigkeit, geringer Probenverbauch, etc.) und die offene Kanalgeometrie machen dieser mikrofluidischen Flüssigkeits-Strahl Systeme höchst relevant für die Etablierung als Probenumgebung an Röntgen-FEL. Diese Einrichtungen generieren Röntgenblitze, die ultrakurz (fs-Bereich) und enorm intensiv sind und mit denen ein ganzes Streubild mit nur einem Lichtpuls aufgenommen werden kann. In diesem Prozess explodiert die Probe und wird in ein leuchtendes, ca. 60 000 K-heißes Plasma verwandelt. Deshalb sind statische Proben oder geschlossene Flusssysteme die diesen Röntgenquellen der nächsten Generation inkompatibel, was die Wichtigkeit des Ansatzes offener Flüssigkeits-Strahlen unterstreicht. Der wissenschaftliche Weitblick des mikrofluidischen Konzepts und der lithographischen Mikrofabrikation wurde ebenfalls demonstriert, indem zusätzlich Lösungen für andere Röntgen-freie Anwendungen gefunden wurden. Die Beispiele hierfür umfassen CFDsimulationen für nicht-lineare wissenschaftliche Problemstellungen eines Sprühtrockners, genauer dessen interne Fluid-Struktur-Interaktion, oder die Gestaltung & Herstellung von Mikrostrukturen für den Mikro-Kontaktdruck, also das mikrometer-genaue Stempeln, kugelförmiger Polyelektrolyt-Bürsten. Ein weiteres Beispiel demonstriert die Kombination von Nanotechnologie und Mikrostrukturierungs-Techniken für Entwicklungen in Richtung sensorischer Anwendungen. 16 1 Introduction 1.1 Motivation A wide range of nature’s fundamental growth principles, like for example self-assembly or nucleation & growth, are governed by processes on the nanometer to micrometer length scale and their understanding is of great interest for modern material sciences and condensed matter physics. 1-3 The investigation of these elemental principles requires the combination of advanced analysis techniques that extend into the relevant length scales with sample environments that are capable of controlling the experimental physical and chemical conditions with very high precision. From an analysis point of view, X-ray scattering techniques and microscopic methods offer a variety of tools to study these processes with great detail on all the relevant length scales. With the current developments in synchrotron technology and with the advent of free electron lasers (FEL), very exciting possibilities arise such as time-resolved experiments at highly brilliant synchrotron sources and serial femtosecond X-ray nanocrystallography at Xray free electron lasers. 4-7 With respect to the sample environment, the interdisciplinary scientific field of microfluidics is predestined for fundamental investigations because it focuses on the precise control and manipulation of very small fluid volumes in micrometer-sized geometries. 8 The combination of microbeam X-ray scattering, microscopy and microfluidics is currently being developed into a powerful experimental methodology suitable for the in situ investigation of nanostructures, particle alignment and the in situ study of kinetics. 9-14 This progress is enabled by creating X-ray compatible microflow chips and microfluidic liquid jet devices that allow new scientific approaches compared to traditional experiments on the macroscale because the sample’s environmental conditions, like small scale concentration gradients or flow fields on the micron-scale, can be tailored to particular scientific questions. 9,13,15 The following chapters will provide an overview over this combination of technologies by covering its origins & challenges, the theoretical fundamentals as well as the involved methods and techniques. 1.2 Microfocus X-ray sources Small-angle X-ray scattering (SAXS) allows to analyze the nanometer-sized internal structure of a sample. However, X-ray sources with traditional collimation systems are only capable to deliver beams which are typically millimeter-sized. This range of X-ray spot sizes prevents the precise analysis of small or heterogeneous samples due to the signal averaging over the illuminated area. 17 This situation improved tremendously with the greatly increased peak brilliance of state-ofthe-art synchrotrons and with the advent of specialized techniques for the generation of microfocused X-rays. 4,16 Due to the pioneering work at dedicated focused X-ray beamlines, a variety of focusing principles has been developed that involve components like KirkpatrickBaez (KB) crossed mirrors or -multilayers, capillaries, Fresnel optics, wave guides or compound refractive lenses (CRL) and together with advances in high-precision positioning systems, these techniques now enable X-ray foci down to the nanometer range. 11,16-26 Consequently, samples can now be analyzed with a much higher spatial resolution and greater precision. This capability allows to distinguish between differently structured domains in a material of which the scattering signal previously had been averaged. However, the focus spot size is not the only thing to consider for experiments with focused X-ray beams. It is also very important to consider the divergence of the beam as well as the overall photon flux. 20 The relation between these three parameters can be imagined as a triangle because all of them are interrelated. For example, it is possible to generate very smallsized X-ray beams using CRL-focusing optics. This also increases the photon flux at the focus position tremendously. However, this is only possible at the cost of a higher beam divergence and consequently a loss of resolution. It is also possible to have a small beam with low divergence for an increased resolution, but this might only be possible at the cost of photon flux which results in longer measurement times. If photon flux and divergence are both important, the minimum X-ray spot size has to be larger with a decrease of spatial resolution, although the beam size would still be in the low micrometer range. An example overview over the spot sizes and its influence on divergence and flux for different focusing techniques is shown in Fig. 1. Figure 1 Comparison of different X-ray microfocusing techniques and their spot sizedependent influence on flux and divergence. The focusing types include capillaries (top), Kirkpatrick-Baez (KB) crossed mirrors (middle) and compound refractive lenses (CRL, bottom). (Figure from 20, Copyright IOP Publishing) 18 Today’s dedicated microfocus X-ray beamlines at 3rd generation synchrotrons, like for example P03 (PETRA III at DESY, Hamburg, Germany) or ID13 (ESRF, Grenoble, France), also take the focusing distance from the optics to the sample into account. This leads to a lower divergence at a given focus spot size while a high photon flux is maintained. In the end, it is a question of the optimum combination of focusing settings for a given sample or experimental setup. In case of a microfluidic X-ray experiment, the maximum divergence of the microfocused beam is dictated by the channel height which typically ranges between 50 µm and 100 µm. In most of our microfluidic SAXS experiments, the X-ray beam size was adjusted to spot sizes around 10 µm by 10 µm to guarantee a low divergence and maintain a high photon flux. An example which clearly demonstrates the benefits of the high spatial resolution of microfocused X-ray beams is the analysis of thin celluloseor high-performance polymer fibers27-29 A fiber’s structure can be mapped in great detail which also provides information about the fibers internal structure. Another very interesting example for a biological sample system are spider silk fibers. 22,30-34 These in vitro and in vivo studies revealed how the spider silk fiber’s nanostructure changes during the spinning process, during its elongation or under the influence of and how this affects the fiber’s tensile strength and micro-structural properties. The smaller X-ray beam size is also important for grazing-incidence small-angle scattering (GISAXS) because the microfocus-illuminated area is much smaller compared to conventional X-ray beams, enabling better spatial scanning resolution. 35 Another example where the use of a microfocused X-ray beam greatly improves the spatial resolution is SAXS microtomography. 36-40 This measurement technique reveals the internal three-dimensional structure of a sample, like for example a high performance polymer fiber, by rotating it during the detailed mapping with the X-ray beam. Due to the Nyquist-Shannon sampling theorem and the microfocused X-rays that enable the mapping of much smaller volumes, the sample scans require much less images (and therefore shorter scanning times) for a given resolution compared to tomographic scans with larger beams. 41,42 The ongoing development of X-ray sources leads to increasingly brilliant and intense beams. 4 This can lead to new problems concerning the sample: the maximum X-ray dose before the sample degrades. Biological samples or soft matter are just two examples of sample systems which are susceptible to radiation damage. 43 A protein crystallographic case study demonstrated that it is possible to increase the sample’s resilience to beam damages through freezing. 44 Obviously, this cryogenic approach is impractical for liquid samples or solutions that freeze below the targeted temperature, but it is a very useful method for protein crystals. However, this study also revealed that the successful collection of a crystallographic data set is only possible up to a certain X-ray dose because otherwise the sample degrades before the data set is obtained. 44 This maximum dose dictates a minimum crystal size in the micrometer 19 range which is very limiting for the protein-structure determination at traditional X-ray sources like synchrotrons or lab sources. The required minimal crystal size could be reduced by the use of microfocused X-rays in combination with motorized sample handling that enable precise crystal scans, but it is still an extremely challenging task to grow ’large’ micrometersized protein single crystals in the first place. 43,45 One possible approach to avoid the beam damage for liquid samples or solutions is to use a continuous sample stream minimizing the sample residence time in the volume of the X-ray focus. This approach requires a sample environment that allows to control liquids with very high precision and reproducibility on the micrometer scale. If this was not the case, the possibly high sample consumption rates of macroscopic flow systems would make the required sample amounts impractical. This is where the combination of microfocused X-ray beams and X-ray compatible microfluidic devices becomes important and demonstrates its strengths. 9,46-48 The challenges and benefits of X-ray compatible microfluidic sample environments as well as first examples of experiments will be discussed in further detail in the following chapter. Another approach for overcoming the maximum dose limit of solid and liquid samples becomes available with the advent of X-ray free electron lasers (XFEL). 49-51 This next generation of X-ray sources solves the radiation damage problem by generating highly intense femtosecond X-ray pulses and applying the principle of ‘diffraction before destruction’. 6,52 Each generated X-ray pulse is so intense that a full diffraction pattern is collected with this single shot while its pulse length is so short (femtosecond range) that the diffraction pattern is generated before the sample is destroyed by Coulomb explosion. Consequently, the successful collection of diffraction patterns requires a fresh sample with each single pulse. However, the realization of an efficient and reliable way of generating a continuously replenishing sample stream is a very challenging task. 53,54 This is where the microfluidic liquid jet devices, which are introduced in more detailed in chapter 7.3, offer great potential as a sample environment for XFELs and other pulsed laser experiments due to the flexible design control and fast fabrication routines of these devices. 1.3 Introduction to microfluidics Microfluidics has become today’s platform technology for the precise control and manipulation of fluids on the micrometer scale. 8 The term ‘microfluidics’ originates from the combination of microfabrication techniques such as soft lithography with the fluid dynamics on the micrometer scale. 10,55-57 Accordingly, small amounts of fluids ( to liters) are handled and manipulated using channels with dimensions of tens to hundreds of micrometers. 58 Hence, the biggest advantage of microfluidics lies in the micrometer dimensions of the channels and the related fluid dynamic implications for the samples such as laminar flow or diffusion-based mixing. These fundamental physical properties at the micron scale enabled the evolution of a variety of microfluidic tools. As an example, these tools can be highly 20 beneficial for the precise study of nucleation and growth processes and the fast and efficient screening of experimental conditions, like i.e. pH, ionic strength, species compositions, shear forces, cosolvents and concentration. 58 In the beginnings of this technology, the microfabrication of these small channels has been strongly influenced by the field of microelectromechanical systems (MEMS) that involves electronic circuits, sensors and micromechanical components. 58-61 This field offers a rich portfolio of available glassand silicon-related fabrication techniques which stimulated the early development of microfluidics. 58 Today, the most used microfluidic device fabrication technique is soft lithography which is the combination of soft materials such as polydimethylsiloxane (PDMS) with photolithography. 55,56,62,63 These fabrication methods and more device materials will be discussed in greater detail in the chapters 2 and 4. The closely related terms ‘miniaturized total analysis systems’ (µTAS) or ‘lab on a chip’ originate from the 1990s. 64 They describe the concept of combining the elements of microchemical ‘factories’ on a small, single chip which can incorporate functional elements such as pumps, valves, mixers, switches, heaters, multiplexers, electrodes and sensors. 8,58,65-83 Consequently, the ‘lab on a chip’ concept aims towards the increase of mobility and the reduction of energy consumption, waste production and ultimately production costs by eliminating the need for traditional laboratory equipment. 64 This idea has been demonstrated for complicated chemical reactions and complex microchannel networks that combine multiple functional elements on a single chip. 67,79 The concept of a micro reaction plant on a chip has also been demonstrated for complex reactions like the living anionic polymerization of block copolymers with direct on-chip DLS analysis of the resulting micelles. Another complicated reaction on a chip has been demonstrated for the synthesis of 18F-labeled organic compounds that are used in positron emission tomography (PET). 58,81,84,85 Additionally to the already-mentioned features of microfludics that include low sample consumption, the beneficial features of microfluidics also include the integration of functional elements on a chip which enable small device footprints. Further, the fabrication costs are typically small, the waste production is minimized and it is also possible to run exothermic reactions while maintaining temperature control. 86,87 This great temperature control is enabled by to the small amounts of reacting mass combined with the high surface to volume ratio of the microchannel network. As a consequence, safe operation is guaranteed while the uniform heat transfer also gives great control over the reaction kinetics. 86,87 Further examples include the production of microparticles and nanoparticles with a large diversity of morphologies and physicochemical properties with respect to size, shape, surface charge and amphilicity. 88-96. Although the volumes of the handled fluids are typically small, the massive parallelization of microfluidic devices offers the potential of upscaling the processes to industrial scales. 97 21 Microfluidic technology also offers many advantages when it comes to sample analysis and, consequently, today’s list of developed applications for microfluidic platforms is manifold. 58 For example, sample analysis related demonstrations include “separations coupled to mass spectroscopy, high-throughput screening in drug development, bioanalyses, examination and manipulation of samples consisting of a single cell or a single molecule“.58,74,96,98-101 Further, applications include processes such as free-flow electrophoresis or blood sample analysis which have been improved and miniaturized. 102,103 Together with the above-described functional elements, like valves and pumps, combinatoric experiments and high-throughput reaction screenings became possible 67,79,104,105 As an example, Quake et al. developed methods for the microfluidic large scale integration which is the microfluidic analogue to the technological jump from single transitors to microprocessors in electrical engineering. 67,104 Through microfluidic valves, pumps, and multiplexers, this technology enables combinatorics and high-throughput screenings (HTS) for single cell analysis, deoxyribonucleic acid (DNA) synthesis, digital polymerase chain reaction (dPCR), genome sequencing, as well as and large scale genomics and proteomics. 80,106-117 Furthermore, this HTS-approach also allows to find and optimize protein crystallization conditions while only requiring very small amounts of sample. 118-121 Figure 2 Illustration of the microfluidic large scale integration concept. 67,79,104 (A,B) The flow and mixing of samples with nanoliter volumes are controlled using small valves and pumps. Accordingly, the resulting device footprint is very small (C) and the computer-controlled devices can handy very complex tasks such as the high-throughput screening of fluorescencebased single-cell assays (D). (Images from 67(A,B, Copyright Science), 79 (C, D, Copyright Nature)) 22 Another example for microfluidic condition screening and combinatorics is the crystallization of single protein crystals. A wide range of microfluidic tools have been developed for this purpose due to this field’s great importance for medicine and the life sciences. 119,120,122-128 These examples show that it is possible to generate screening libraries for the automated crystallization of enzymes, proteins and other substances under defined conditions in droplets of individually addressable, on-chip microcompartments while only requiring very small amounts of sample. All these examples demonstrate that microfluidics offers great control over fluids, reactions and experimental conditions by taking advantage of the fluid dynamics on the micrometer scale that enable laminar flow and diffusive mixing. 10 In combination with the abovementioned microfocus X-ray techniques, many new experimental opportunities arise which would not be possible with conventional macroscopic systems. However, the transfer of microfluidic technology to X-ray experiments is technically very challenging due to the X-ray compatibility of the different device materials. In this ongoing transfer process, a variety of fabrication approaches and multiple device types have been developed. The next chapter will review these available X-ray compatible device types and describe the studies that have been performed at microfocus X-ray sources. 23 24 2 Microfluidic devices as experimental platforms for X-ray studies 2.1 Introduction The combination of microfluidics and microfocused X-ray beams is a relatively young field of research that started with early experiments by Pollack et al. during the uprise of microfocused X-rays. 46-48 These sample environments have been used for kinetic folding studies of biological systems such as RNA or proteins. 46-48 Hence, this technology-combining approach links the benefits of small X-ray microbeams with the ones from microfluidics and, therefore, allows to study the sample system’s response in situ and in operando with high precision and control. 9-12 The resulting opportunities as well as the first approaches and current developments of microfluidic sample environments for X-ray experiments will be reviewed over the course of this chapter. As mentioned above, the turbulence-free flow conditions within the microchannels offer great control over the chemical and physical conditions. These experimental parameters, like concentration, pH, shearand extensional forces, can be controlled precisely and reproducibly by adjusting flow, mixing and the microchannel geometry according to the sample system of interest. For example, reaction kinetics are diffusion-based due to the laminar flow. These fluid dynamic conditions enable detailed scans of the mixing area with microfocused X-ray beams and results in observable reaction time scales that extend over range from zero to several seconds. 46-48,129-132 Further, the earlier-mentioned problem of radiation damage (chapter 1.2) can be tackled by using sample environments that replenish the sample continuously and therefore reduce the X-ray dose to avoid beam damage of the sample. This is particularly important for fastdegrading systems, like for example biological samples. While the dose barrier for these samples is about 200 photons for X-rays at an energy of 12 keV, the X-ray dose in a typical microfluidic continuous flow experiment is reduced to doses on the order of 0.1 photons due to the very short sample residence times in the exposed volume. 9,44,49 The experimental setup can be optimized even further by tuning the X-ray energy to maximize the transmission through the device’s microchannel material and the liquid therein. 133 Next to the capability of successfully measuring fast-degrading and X-ray sensitive samples, microfluidic devices enable the study of precious samples that are only available in small amounts, like i.e. membrane proteins or deoxyribonucleic acid (DNA). 46-48,130-132 Traditional experiments could prevent their analysis due to the limited sample availability and the experiment’s minimum required sample quantity while microfluidic devices are highly sample efficient and consume only very small volumes in the range of microor even nanoliters per hour. 67,79,118,134 25 Since the soft lithographic replication using polydimethylsiloxane (PDMS) is so easy to learn, it has become one of today’s most used fabrication technique for microfluidic devices. 58 During the microstructure replication the curing PDMS shrinks less than 1% which results in very precise replicas that can be used for example as stamps in micro contact printing. 151-154 These open replicas can also be sealed to create closed microchannels. This is typically achieved by activating the PDMS using air plasma and then binding it covalently to glass slides which results in very pressure resistance microchannels. 57 This combination of PDMS that is bonded glass is typically the material combination of choice when it comes to microscopy-related applications of microfluidics due to its excellent optical transparency from 240 nm to 1000 nm and its low toxicity. 58 Additionally, the elasticity of PDMS can be controlled by the ratio between oligomer the cross-linker while this material’s gas permeability is beneficial for cell cultures or microevaporation. 9,15,58 PDMS also has two significant drawbacks: the limited compatibility with solvents and the device fouling from the unspecific adsorption of biomolecules. 9,34,147 However, PDMS offers a wide range of surface modification possibilities which enables to minimize these drawbacks or avoid the negative effects all together. 155-159 The modification routines include sol-gel glass coatings for improved solvent resistance, layer-by-layer deposition of polyelectrolytes for permanently hydrophilic channel walls, the covalent deposition of fluorinated repellants against device fouling and UV-controlled photochemistry or grafting reactions on the channel surface for wettability tuning. 155-159 PDMS can also be used for the measurement for SAXS directly if the devices are very thin. This has been demonstrated at the Diamond light source in for the orientation analysis of lamellae in microchannels. 160 Thin PDMS-based devices also allow microevaporation that can be used for the generation of concentration gradients in non-flowing samples. 15 Although showing a stronger background signal, these devices’ material homogeneity typically allows to extract the sample’s signal through background subtraction. However, this process can decrease the signal quality and eventually the ability to evaluate the obtained scattering patterns. This decreased signal-to-noise ratio of the sample can result from factors like the substraction-induced reduction of the signal and added detector readout noise from combining multiple files. 161 Additionally to the X-ray background signal, the solvent compatibility of this material is also very limited. 147 A much better solvent compatibility is provided by glass-only-based microfluidic devices. As an example, glass capillary devices with tube-in-tube geometries have been successfully operated at synchrotrons for the study of in situ spider silk fiber formation. 22 This device type also offers the benefit of reduced or no wall clogging due to the coaxial liquid sheath of the outer capillary. 162 Further, the glass capillaries can also be etched down to thicknesses around 50 µm for improved X-ray transmission. The main downside of this device type is its fabrication procedure because it is complex, involves precise manual capillary alignment skills and lastly, the design variations of the channel geometry are limited. 22,162 32 When it comes to X-ray transmission experiments, glass and PDMS or their combination is counterproductive because of the material’s own strong small angle scattering signal. 163 Alternatively, microfluidic devices can be directly fabricated though laser ablation. 141 The microstructures are written directly into Kapton film which is then sealed with another Kapton film to yield the closed microchannel. The resulting devices are very thin, solvent resistant and they show good mechanical properties. Further, Kapton is an excellent material for X-ray applications because of its high radiation resistance and low absorption. It has also been shown that the laser ablation approach can be applied to other materials such as poly(methyl methacrylate) (PMMA), polystyrene (PS) and cyclic olefin copolymers (COC). 164 However, the major drawback of this laser-based approach is the low machining speed which can lead to long processing times due to the typically wide-spreading microchannel patterns. 141 Another variant of stable microfluidic devices with Kapton as the only window material have been described by Pfohl et al. 23,130,139,140 In these examples, open microchannels are spark eroded into stainless steel plates which are then sealed at the top and bottom using selfadhesive Kapton film. The minimum channel size is restricted to about 60-100 µm due to the resolution of the spark eroding technique. 9 While this minimum channel size can be sufficient for many microfluidic experiments, the main drawback of this technique is the limited design flexibility because it is only possible to create relatively simple structures like straight lines. 9 This design flexibility can be increased by choosing a channel wall material that is suitable for rapid prototyping. 57 A widely used process is called soft lithography and it involves microchannel templates that are fabricated by microstructuring a photoresist on a silicon wafer using UV lithography. 56 A thin layer of moldable material, such as PDMS, is then casted on this template by doctorblading or spin coating to generate open microchannels. These channels are then sealed with Kapton films from top and bottom, similar to the steelbased devices that have been described above. 13,130,131,139,165 This combination offers great design flexibility through rapid prototyping while maintaining good X-ray properties by using Kapton windows. However, the X-ray signal could be influenced by the adhesive layer of the sealing kapton tape and the use of PDMS also limits the solvent compatibility of these devices. 147 A different routine that takes advantage of moldable materials and rapid prototyping is the fabrication of microfluidic devices made of an UV-curable optical adhesive by Nordland (NOA81). 14,132,166,167 Originally used for the glueing of optical components such as lenses, this thiol-ene-based material is cured by a radical mechanism that allows surface chemistry modification and which is initiated by a UV-sensitive initiator. 149,168 This material also allows to fabricate very thin devices with window thicknesses of a few tens to hundreds of microns while the small angle background signal is also much lower compared to PDMS. 9,132,135 Furthermore, it provides a much better resistance to a wide range of solvents compared to 33 PDMS. 147,149 It has also been demonstrated that this material is suitable for the fabrication of microfluidic three-dimensional flow geometries by aligning and sealing two microstrucured layers. 132 Each side is fabricated by multilayer lithography which yields a flow focusing geometry that minimizes or prevents clogging of the sample to the microchannel wall. 9,132,162 However, the exact alignment of the two halves with micrometer precision can be very challenging. A general downside of this material is the observable beam damage of this material, especially for higher X-ray intensities, which can be observed as a brownish spot on the yellow-white translucent material. However, this beam damage does not interfere with the flowing samples on the timescale of typical microfluidic experiments. 9,132 2.3 Outlook: new processing techniques and device materials New ways to fabricate microfluidic devices arise with the advent of new materials, new processing techniques or the recombination of existing components. As mentioned earlier, so far no material or routine has been studied yet that matches all criterions simultaneously, but many good routines have been describe for the specific experiments. Materials which allow rapid prototyping, like PDMS or NOA81, can take advantage the great design flexibility that is enabled through soft lithography. 57 Other rigid or hard materials, in terms of material properties or ease of processing (i.e. silicon, glass, Kapton), usually rely on fabrication processes that can be complicated or do not allow the same design flexibility as soft lithography. The relatively young combination of soft lithography with casting-, (microinjection) moldingand hot-embossing-approaches offer multiple processing paths for the fabrication of microfluidic devices. 169-174 This makes a wider range of (polymeric) materials accessible to microfluidics and offers great potential for the fabrication of microfluidic devices with material properties tailored the specific experimental needs while maintaining the great design flexibility of rapid prototyping. As an example, rapid-protoyping-based PDMS-stamps and - templates can be used to emboss the microchannel structures into a wide range of thermoplastic materials such as polystyrene (PS), poly(methyl methacrylate) (PMMA), cyclic olefin copolymers (COC) and THV (fluorinated terpolymer: tetrafluoroethylene, hexafluoropropylene and vinylidene fluoride). 148,150,175-182 Further, the use of fabrication techniques that are based on microinjection molding or hot embossing can greatly increase the production speed of microfluidic devices to industrial scales due to the relatively fast embossing step in the template replication process. 163,169,170,172,183,184 Next to Kapton with its great properties for X-ray applications, a number of X-ray compatible window materials have been tested including polymethylmethacrylate (PMMA), cyclic olefin copolymers (COC), negative photo resist SU-8 (MicroChem) or polypropylene (PP). It is possible to fabricate thin (ca. 250 µm or less) and microstructured films out of these materials of which the background scattering is very similar to air. 9,163 From these 34 examples, thermoplasts like PMMA or COC can easily be structured by softor hot embossing while SU-8 can be microstructured directly by UV-lithography. 9,184,185 Multiple uses of COC as a material for microfluidic devices have been demonstrated. 184 COC-based microfluidic devices also offer a great potential when it comes to the integration of functional elements into the device. These integrated features of COC devices are manifold and include Au-electrodes for electrochemistry, elements for electrochromatography, or PDMS-based control layers for the screening of lyotropic phases. 186-188 Further, the surface chemistry of COC-based devices can be modified by UV-grafting procedures and therefore optimized for the particular experimental needs. 189 Among the X-ray-related examples is a computer-coupled high throughput screening setup by Arleth et al. that uses microfluidic devices for the fast SAXS analysis of protein-folding under the influence of changing buffer conditions. 176,179 Further, the automatization potential of microfluidic devices is demonstrated by using X-ray-CDs that are combined with automated sample positioning at synchrotron beamlines (PETRA III, DESY, Hamburg). These ‘SAXS-LabDiscs’ are rotational microfluidic devices that are based on centrifugal flow principles for the combinatoric mixing and screening of samples. 125,180,190-194 It has also been shown that COC-based microfluidic devices are suitable for microfluidic grazing-incidence small-angle X-ray scattering (GISAXS) experiments. 35 However, a downside if this polymeric material is the limited solvent compatibility and, based on its thermoplastic nature, the lack of resistance against higher temperatures. 163,177,180,184,189,195 As described above, the advent of new techniques and materials offers new paths for the fabrication of microfluidic devices that are tailored to the experimental requirements. A different device fabrication approach will be discussed in the next section with the generation of liquid jets that avoid any materials in the X-ray beam path. 2.4 Microfluidic liquid jet devices for next generation X-ray sources Today’s developments in synchrotron technology continually push the peak brilliance of the X-rays, enabling fast measurements as well as time-resolved in situ experiments. 7 Consequently, radiation-induced sample degradation and the maximum X-ray dose start to become the limiting factors for experiments at these facilities. As described above, one way avoid the dose problem is to scan different spatial position with microfocused beams, or use (microfluidic) continuous flow systems. However, each of these alternatives is limited to certain sample types. 45,142 Another path for overcoming the X-ray dose problems becomes available with the advent of highly intense and ultrashort-pulsed X-ray free electron lasers (XFEL). 6,49 These free35 electron laser sources currently generate X-ray pulses at rates up to 120 Hz (LCLS, SLAC, Stanford, USA) while future facilities will even generate up to ca. 27,000 pulses per second that are bundled in 10Hz bunch trains (European XFEL, Hamburg, Germany, in 2015/2016). These X-ray pulses are so intense that a full diffraction pattern of nanometersized crystals is collected within a single shot and that the X-ray beams “are capable of destroying anything in their path”. 6,50,196 During the illumination with a single light pulse, the sample explodes and turns into a glowing plasma (ca. 60,000 K). 196. Due to the ultrashort femtosecond pulse length however, the X-ray pulses outrun the explosion process, ‘freezing’ the atom positions in space. 49,50 An analogue example from the macroscopic world for this ‘freezing’ principle is the highspeed flash photography work done by Harold Edgerton (MIT, Cambridge, USA). For example, his photograph “Cutting the Card Quickly” (1964, 1 µs exposure) shows how a bullet-separated play card levitates in air at its original position. This concept of ‘diffraction before destruction’ is of great importance for the study of ultrafast processes and the characterization of a wide range of X-ray sensitive samples. 49 As an example, the structure determination of transmembrane proteins is important for medicine and the life sciences because this information can help to understand the signalling of cells or mechanisms of diseases. 6 However, the structure determination of proteins is a very challenging task because these and many other biological samples are only available in small amounts and/or tend not to form single crystals of sufficient dimensions for traditional X-ray (micro-)crystallography. 6 It is much easier to grow nanometer-sized protein crystals, but these cannot be analyzed at tradionional X-ray sources due to their small size and the discussed dose limit. The relation between the average intensity of a diffraction peak and the crystal volume is approximately proportional. 45 Hence, the required beam intensity of a 20 µm crystal is 1000-fold higher than for a 200 µm crystal. 45 Theoretical and experimental results show that the critical dose for a successful structural evaluation is mainly dictated by the crystal size, despite cryogenic attempts of measuring protein crystals that are cooled below 100 K. 45 Recently, femtosecond X-ray protein nanocrystallography has been demonstrated for the high-resolution characterization of photosystem I&II and the model protein lysozyme, showing that the X-ray dose limit can be overcome as described above. 6,197-199 As a consequence of the enormous X-ray pulse intensity, statically mounted samples or experimental environments for flowing samples that are based on closed geometries are incompatible with these 4th generation X-ray sources. Therefore, special sample environments are required that deliver the samples in mid-air and under vacuum conditions while being as sample efficient as possible. 53,200 These kinds of sample environments for XFELs and how microfluidics offers great potential for the adequate delivery of samples will be discussed in this chapter. Currently, continuously replenishing sample streams are generated using aerodynamic lens particle injectors or liquid jets using a glass-based capillary-in-capillary design. 53,200 Since the 36 glass capillaries are widely used and suitable for liquid samples, this chapter will focus only on these. The principle for the generation of liquid jet is based on a gas sheath which shapes a liquid stream and has first been demonstrated in a plate-orifice geometry. 201,202 Later, this concept has been transferred to glass capillaries which run essentially clogging free due to their gas-dynamic virtual nozzle design (GDVN). 53 In other words, the pressured gas forms a liquid stream and avoids any wall contact of the liquid. 53 This results in a very stable and reliable system for the generation of nanoor micrometer-sized liquid jets that require only small amounts of sample (down to ca. 100 µl h-1). 203,204 The main drawback of this glass capillary design is the complex fabrication process that involves steps like flame polishing of the tip, its grinding as well as alignment of the inner and outer capillary. 53,203-205 These steps require the manual skills and attention of a lab worker which intrinsically results in geometric variations of the nozzles. It is therefore very hard or impossible to exactly reproduce a targeted design or even automate the process. This is a major issue because the generated fluid dynamics of the liquid jet and, hence, minimum flow rates and liquid jet diameters strongly depend on the geometric and experimental parameters. 205-207 The fabrication procedure is one key point where microfluidic devices shine, due to their fast and easy fabrication and highly reproducible design which is based on established softlithographical techniques. 55-57 In this thesis, microfluidic chip-based devices are presented that produce liquid jets with µm-diameters (20 down to 2 µm, or even 940 nm) at very low flow rates (down to 150 µl h-1) under atmospheric or vacuum conditions. These microfluidic liquid jet devices are also based on the gas dynamic virtual nozzle (GDVN) design which enables reliable and essentially clogging-free jetting over long periods of time. 53 The flexibility in microchannel design control is demonstrated by the easy integration of additional microfluidic features, such as jet-in-jet flow focusing, which could enable new in situ experiments at XFELs, or dense arrays of multiple adjacent liquid jet nozzles on a single device, without the need of additional production steps. Hence, the potential of simplifying and up-scaling the fabrication of micro-nozzles for the generation of liquid jets is demonstrated. The microfluidic liquid jet system is highly relevant for the establishment of microfluidics at XFELs because these devices deliver the sample continuously, reliably and efficiently in atmoshperic or under vacuum conditions, as illustrated in Fig.6. 136 37 Figure 6 Combining the liquid jet principle with microfluidics. (A) Illustration of the building block principle of functional microfluidic tools that can be combined and stacked using the microfluidic liquid jet device principle. (B) The X-ray beam hits the liquid jet in this illustration of the experimental setup. Microfluidic devices are advantageous because each liquid jet device contains a dense array multiple microfluidic GDNV-nozzles which enable fast nozzle changes and, consequently, reduce (expensive) downtimes at the X-ray free electron lasers. (Image adapted and extended from 6) 38 2.5 Conclusions In conclusion, the combination of microfluidics with microflocused X-rays is a valuable experimental methodology for the study of fast in situ experiments. While this field is still emerging, a wide range of device types is already available. The on-going development of microfabraction techniques and advent of materials for the production of X-ray compatible microfluidic devices add to the great potential and this technique’s future applications. Additionally, the variety of other X-ray imaging and spectroscopic techniques could extend the experimental opportunities of microfluidics at X-ray sources even further. Future developments of microfluidic systems and highly brilliant X-ray sources, such as synchrotrons or XFELs, could soon lead to the fundamental understanding of nucleation and growth processes or integration of the high-throughput screening of proteins that yields full threedimensional as well as dynamics information about these species or even whole cells; with important insights for the natural and life sciences. 39 40 3 Theoretical Fundamentals 3.1 Self-organization of amphiphiles One of nature’s fundamental principles is self-assembly. 208 While technical systems are typically organized by men, the self-organization of natural systems is based on internal processes on the very small scale. This bottom-up approach begins with single functional molecules that build up and organize themselves to larger structural elements. The selforganization of a system induces properties that it did not have before, like i.e. order, and this process can be described by the theory of spontaneous symmetry breaking. 208,209 For example, this mechanism can be observed in nature with the formation of lipid double membranes of cells or the self-assembly of micelles and liquid crystals. 209-211 For molecules to be able to form such assemblies of higher order, their structure and functional elements need to meet certain molecular prerequisites. 208 These molecules need to possess (and unite) both long range repulsive and short range attractive forces in order to be capable of forming structured domains, as illustrated in Fig.7. Figure 7 Illustration of amphiphile building blocks with long range repulsive and short range attractive forces and their self-assembly. (from 208, Copyright WILEY-VCH) Long range repulsive forces can appear as Coulomb repulsion, chemical incompatibility or hydrophobic interactions while covalent bonds or the local conservation of electroneutrality are examples of short range attractive forces. 208 These prerequisites are met by molecules such as lipids or amphiphilic block copolymers of which the latter can form a wide variety of superstructures as illustrated in Fig8. 1,208 Continuous phases are formed in bulk or in lyotropic phases, and colloids like spherical or cylindrical micelles or vesicles are typically formed by microphase separation in dilute solutions. 1,208 41 3.3 Solution of non-linear problems For linear systems, the input is proportional to the output. Most systems which are of interest for science and their applications are non-linear due to the complexity of nature. Their successful simulation is of great interest because precious resources such as material costs, man-power and development time can be saved. Further, these systems can lead to interesting non-linear scientific problems of which a few examples will be discussed below. 3.3.1 Fluid structure interaction As mentioned earlier, PDMS is a soft polymeric material that deforms elastically under force, like i.e. high pressures. Depending on the application, the mechanical properties of this elastomer can also be beneficial or a down side. On the one hand, the deformation under high pressures and high flow rates can result in rounded channels and reduced the wall contact of a hydrodynamically focused liquid stream. 137 On the other hand, this pressure-induced deformation (at >2-3 bar) can alter the microchannel geometry which could make the prediction of flow condition more complicated. With the help of computer-based CFD simulations (COMSOL Multiphysics v4.2a) it is possible to solve these problems numerically based on the finite element method. This approach couples the microchannel geometry deformation with the fluid flow field which are incrementally affecting each other in a nonlinear way. This routine results in the dynamically stable state of this problem. The theoretical background as well as the solved models, which are good agreement with the real experimental demonstrations, are described in one of the publications of this thesis (see chapter 7.5). 137 3.3.2 Shear forces and Non-Newtonian fluids Another non-linear problem occurs when one has to predict the flow of non-Newtonian fluids. These fluids change viscosity under the influence of forces like shear or elongation. Generally speaking, the viscosity of non-Newtonian fluids can increase (shear thickening, dilatant) or decrease (shear thinning, pseudoplastic). The wormlike polymeric micelle solutions which have been used in this thesis are examples of shear thinning non-Newtonian fluids. Their change of viscosity under shear can be described by the Cole-Cole- (or Cross- ) equation which is given by: 247-249 with the zero shear viscosity , the high-shear viscosity , the internal relaxation time and the power law exponent characterizing the shear thinning between and .13,135 The fluid flow of non-Newtonian fluids can now be calculated by coupling this equation with the above-described Navier-Stokes equations of an incompressible fluid and running the FEM-based CFD-simulation. As described in the publications of this thesis (see chapter 7.1 & 7.2), the simulated flow and predicted shear and extensional forces are also in good agreement with the experimental results (SAXS, µPIV, polarization microscopy) and can be 48 applied i.e. for the prediction of perpendicular particle orientation in confined geometries. 13,135 3.3.3 Two-phase systems and liquid jets Another example of a non-linear system, which has been studied in this thesis, is the fluid flow of two-phase systems (see chapter 7.3). 136 The gas-dynamic virtual nozzle principle is a two-phase flow system which uses a pressured-gas sheath for the generation of liquid jets. 53 This design prevents wall contact of the liquid and nozzles of this type run essentially clogging-free while consuming only small amounts of sample. 53,54 The non-linearity of this system lies in the coupled flow fields of each fluid. When the liquid enters the nozzle geometry, where the gas is already flowing, the liquid’s surface shape is affected by the gas flow. At the same time, its presence alters the gas flow dynamics, etc. The theoretical background and the simulation are described in a paper of this thesis. 136 In this paper, a time-resolved model enables to begin the simulation with easily definable starting conditions. The CFD-simulation of this coupled interaction incrementally leads to a stable equilibrium state and a stable gas-shaped liquid jet. 54,207 The resulting simulated liquid jet shape and diameter are found to be in good agreement with the experimental results from high speed video microscopy. 136 This CFD-model allows the very detailed analysis of the whole system, including jet shape, pressure gradients, shear rates, velocities at every simulated position of each fluid. If one only needs to estimate the liquid jet diameter, there is also an alternative, analytical approach which has been described for a plate-orifice configuration, as illustrated in Fig.11. 201 Assuming cylindrical coordinates for the axis of the liquid jet, the cusplike meniscus at the inlet is pulled towards the nozzle orifice by the pressure gradient that is generated by the gas stream. 201 This pressure difference and tangential viscous stress of the gas sheath leads to the formation of a thin liquid thread with the radius as illustrated in Fig.11. 201,206 The averaged momentum equation for this case is given by: 201 with with the flow rate , the liquid pressure and the surface tension stress . The liquid evaporation will be neglected as well as the viscous extensional term which is negligible compared to the kinetic energy term. This holds true for many flow rates of stable liquid threads. 201 Thus, the above averaged momentum equation can be simplified to: 49 Figure 11 Illustration of a general plate-orific nozzle design for the generation of liquid jets and the involved geometric parameters. The liquid exits the capillary with the diameter at a flow rate of . It is shaped by the pressured gas sheath ( , left & right arrows) along the path towards the nozzle and as it passes the nozzle with the diameter . The diameter and the stability of the resulting liquid jet depend on the given geometric parameters as well as other fluid dynamic properties such as fluid density , viscosity or surface tension . (Image from 206, Copyright American Institute of Physics). Assuming high pressure gradients towards the nozzle and that the confining nozzle aperture is of the order or thinner than its diameter, this equation can be integrated. 201 This integration yields a simple and universal expression for the jet diameter that is given by This formula is independent of the geometrical parameters (like i.e. inletand outlet diameters, inlet-to-oulet-distance, etc.), liquid-gas surface tension and liquid and gas viscosities. 201 The validity of this expression has been demonstrated experimentally for plateorifice configurations and glass capillary setups. 201,206 Our experimental results show that there is also a good agreement of this formula with the liquid jet diameters that are generated in microfluidic gas-dynamic virtual nozzles (see chapter 7.3). 136 When it comes to the lowest possible flow rate however, the nozzle shape has a decisive effect. 206 It has been demonstrated that the minimum flow rate for stable jetting can be controlled by adjusting the geometry. 206 For three different nozzle diameters ( ), the following graph (Fig.12) shows the minimum flow rates ( ) which are plotted against the distance ( ) between the liquid inlet and the nozzle aperture. This graph clearly shows that this -ratio is very sensitive. Consequently, a precise and reproducible nozzle design control is essential for the optimization of liquid jet system and their integration as a sample environment at high intensity X-ray sources. 6 50 Figure 12 The minimum flow rate for stable jetting of water at different combinations of nozzle distance and nozzle diameters with 100 µm (circles), 200 µm (squares) and 400 µm (triangles). The pressure difference was set to 250 mbar. (Image from 206, Copyright American Institute of Physics). The regimes for stable liquid jets (steady jetting) can be mapped out using Reynoldsand Weber number diagrams as illustrated in Fig.13. 206 The breakup transition lines depend on the specific nozzle geometries & fluid properties and are therefore only valid for a specific case, as indicated in Fig.13 bottom. The careful mapping of a scenario’s jet-breakup parameter combinations allows to identify its breakup type. In case of a plate-orifice configuration, the following equations can be used for the parameter conversion: 206 and with with the density of the fluid , the flow rate , the jet radius , the viscosity , the surface tension and the pressure difference . These equations can also be combined in a radius-independent form if the exact value of the pressure difference is known: and As described above, the Reynolds number describes the ratio between inertial and viscous forces while the Weber number is the ratio between the inertia of the fluid compared to its surface tension. 10,213 A stable jet is observed if the fluid outruns the instabilities convectively. 201,206,207,250,251 If the gas flow is too high for a given liquid stream for example, the jet turns into a spray which corresponds to a local&global stabilityto local&global instability-transition. This ‘right-to-left’ transition is marked as a blue intersected line in Fig.13. 201,206,207,250,251 51 If the gas pressure at a given flow rate is lowered slowly (‘top-to-bottom’), the breakup corresponds to a local stabilityto local instability-transition while being globally stable. In this case a continuous droplet train at a constant frequency is observed with a steady liquid column in the range of the nozzle opening. This breakup type is marked by the red dotted line in Fig.13 and is also known as the Leib-Goldstein limit. 250,251 Further lowering of the pressure finally leads the global instability (Fig.13 bottom). 201,206,207,250,251 Under certain conditions when the nozzle geometries become very small, as in microfluidic liquid jet devices and very small jets, the relative influences of the fluid’s surface tension and of the shear from fast-flowing gas streams on the liquid surface increase. Hence, the underlying assumptions (see momentum equation discussion above) are not neglectable anymore and the Reand We-number conversions can become inaccurate. 201,206 It is therefore preferrable to rely on qualitative results for the identificaltion of breakup types. This can be achieved by recording the jet breakup transition using highspeed cameras. This has been demonstrated in the attached paper on microfluidic liquid jet systems, i.e. for the columnlength of liquid jets or droplet trains in the local instability regime (see chapter 7.3). 136 Figure 13 Reynoldsand Weber number diagrams and jet breakup types at minimum flow rates. The liquid jets are observed in the steady jetting regime. The red dotted line describes the Leib-Goldstein transition (local stability to local instability). 250,251 The blue intersected line marks the global stability to instability transition. The top diagram marks the different jet breakup transition regions. Spraying can typically be observed in the global instability regime while a continuous droplet streams can be an example for the local instability breakup type. The bottom graph illustrates how the curves in the Re-We-space shift with changing nozzle geometries. (Images from 206, Copyright American Institute of Physics). 52 3.4 Small angle X-ray scattering 3.4.1 Preface Small angle X-ray scattering (SAXS) is one of today’s most important experimental techniques for the characterization of soft condensed matter and colloidal systems. This chapter will give an overview over its theoretical background and describe the fundamental X-ray scattering principles. A more detailed description can be found in the original, cited literature on which this chapter is based. 12,252-263 3.4.2 Introduction The scattering of electromagnetic waves is a very helpful tool for the characterization of colloids and polymers. The fundamental setup of a scattering experiment is illustrated in Fig. 14. Fig. 14 Illustration of the fundamental elements of a scattering experiment. The X-ray source (X) emits light which passes a collimation system (C). The X-rays then hit the sample (S) and the scattering pattern is recorded using a detector (D) while the primary beam is absorbed by the beamstop (B). (Image adapted from 258) The X-rays, which should be as monochromatic as possible, are emitted from a source (X), like a synchrotron or rotating anode. The X-rays pass a collimation system (C) before they hit the sample (S). This interaction causes the sample’s electrons to resonate and the induced dipoles emit secondary waves of the same frequency. This scattering process is considered to be elastic because the incident photons have the same energy as the scattered ones. The scattered X-rays are coherent and the phases of these secondary waves differ from each other due to the different spatial positions of the scattering electrons. Consequently, interference of these secondary waves occurs and the resulting scattering patterns are then recorded using 2D digital detectors (i.e. Pilatus, FReLoN, MarCCD) or alternative recording techniques such as 1D counting devices or image plates. The interference of the scattered X-rays and therefore the scattering pattern is characteristic for the given sample’s structure. The fundamental principle of the waves’ interference can be described by the Bragg equation. 253 A scattering reflex, which is an intensity maximum, can be found where the waves’ path difference is an integer multiple of the wavelength which leads to constructive 53 interference. This principle is illustrated in Fig. 15. which clearly shows that the path difference is determined by the scattering angle and the distance of the scattering planes . Fig. 15 Illustration of the Bragg equation with the incident X-rays and , the lattice distance and the half scattering angle . The path difference of the wave is given by . (Image adapted from 263) The resulting scattering pattern of a sample is an angle-dependent intensity distribution that is characteristic for the sample’s structure. Given by the range of the (small) scattering angles of this method, the size range of structures that are studied by SAXS typically lies between 1 nm and 100 nm. This is also why this experimental technique has become very popular for the characterization of sample systems such as polymers, colloids, soft condensed matter and nanomaterials. 3.4.3 The scattering vector For a better understanding of the interference phenomena one can imagine the interaction between the X-ray beam and two scattering electrons at the positions and . The vector which describes the distance between them is given by as illustrated in Fig. 16. Fig. 16 Phase relation between the two scattering centers and and the geometric construction of the scattering vector . (Image adapted from 258) All interferences of waves that originate from scattering events sum up with respect to their amplitude and phase. Due to the equality of electrons when it comes to X-ray scattering, only 54 the location-dependent phase difference needs to be considered which is given by This equation contains the wave vectors which are given by for the incident wave, while the scattered wave is described by with the unit vectors ( ) in -direction. The absolutes of these vectors are given by Considering the angle of the X-rays , the wave vectors can be combined to construct the scattering vector which is given by The geometric representation of this relation is shown in Fig. 16. The phase shift can be described by the multiplication of this vector with . The scattering curve is received by plotting the measured intensity against the scattering vector’s absolute value which is given by The absolute value of can also be expressed as which is based on the following equation where the relation between , and has a similarity to the Bragg equation. 3.4.4 The scattering pattern The coherent scattered X-ray waves interfere with each other and the amplitude of the resulting wave is given by with the scattering length . For SAXS the scattering length of an electron is given by255 55 with the elementary charge , the electron’s mass and the speed of light . Consequently, the scattering length of an atom with the ordering number is given by The cumulative scattering of a sample is the sum of all scattered waves. Therefore, the collective scattering amplitude is received by integration and the scattering length of a single pair of scattering centers is replaced by the density distribution of all scattering centers . This is described by the following formula. Since this equation’s mathematical form is a Fourier transformation, the scattering amplitude and the scattering center density distribution are a pair of Fourier transforms. Further, this equation links the real space (with the vector ) to the reciprocal space (with the scattering vector ). However, the scattering amplitude is experimentally not accessible because only the scattering intensity is measured. Their relation is defined as In other words, the scattering intensity is the time-averaged square of the absolute value of the scattering amplitude . It is time-averaged, as indicated by the pointed brackets , because the measurement is long compared to the system’s dynamics. Another relevant mathematical operation is the convolution of two functions and it is given by It is commonly known that a convolution of and in the real space corresponds to the multiplication of their Fourier transforms and in the reciprocal space: 263 The application of this convolution theorem on the complex scattering amplitude and the scattering intensity yields As briefly described above, the scattering amplitude and the scattering center density are a pair of Fourier transforms. The above convolution theorem also shows that this is also true for the scattering intensity and the pair correlation function . The scattering intensity results from the square of the absolute value of the scattering amplitude while the pair distribution function results from the self-convolution of the scattering center density . 56 In the form of a Fourier transform can also be written as264 which is the spatially averaged intensity of a statistically isotropic system without any long range order, such as dilute particles. Here, is the auto-correlation function which is given by and the following equation highlights the relation between the pair distribution function and the auto-correlation The following Fig. 17 illustrates these relations between real and reciprocal space graphically. Fig. 17 Graphical representation of the mathematical operations that link the scattering amplitude , the scattering center density , the scattering intensity and the pair correlation function . (Image adapted from 258) The Fourier-transformation is fully reversible in both directions while the square of the absolute value and the self-convolution are not. The wanted scattering center density is not extractable from the experimentally measured scattering intensity because the phase information is missing; this is also known as the phase problem. This is why two different approaches have been developed to receive the scattering center density . The so-called indirect method is based on the modeling of the scattering center density using spline functions. 265-270 The splines are transformed to the measurement space and fitted to the scattering curve. The desired scattering center density is finally received by the deconvolution of these fitted spline functions. 271,272 Another approach is the model-based or direct method. This approach uses a given structure with a known scattering center density that is Fourier-transformed to receive the 57 with This profile shape smoothly transitions from a Lorentz peak shape at very small -values to a Gauß shape at large values of The peak width is determined by the domain size which is given by The correlation function is given by which also describes the deviation, known as the Debye-Waller factor, from the ideal lattice position based on the following equation In this formula is the mean square deviation and is the nearest neighbor distance. 3.4.10 Particle orientation distribution In the above discussion, the orientation of anisotropic particles, such as cylindrical micelles or wormlike micelles, has been assumed to be isotropic. This results in ring-like, isotropic scattering patterns on two-dimensional detectors which can simply be represented as radially averaged scattering curves. If the particles in a sample are oriented however, the resulting scattering patterns are also anisotropic and both components of the scattering vector ( and ) need to be considered in the analysis. A good experimental example for anisotropic scattering patterns are small angle X-ray scattering studies of shear-oriented wormlike micelles in small microfluidic channel geometries, as illustrated in Fig. 22. 64 Fig. 22 (A) The sample is flowing though the microchannel geometry and passes a narrow section (red box). (B) 3D illustration of the X-ray microbeam that passes the microchannel of flowing wormlike micelles (hexagonal closest packing, determined by SAXS) and the resulting SAXS-pattern which is captured using a digital detector (Pilatus 300K, Dectris). The parallel or perpendicular orientation of the wormlike micelles is controlled by varying the experimental parameters, such as the channel geometry, the flow rate or the sample concentration.(Adapted from 13, Copyright PNAS) The wormlike micelles in the microchannel are oriented according to the flow-induced shear and extensional forces which are controlled by the experimental parameters such as microchannel geometry, flow speed and particle concentration. Due to the highly reproducible flow conditions, this setup enables the correlation between structuraland orientation information from SAXS studies with fluid dynamic studies from other methods like high speed video analysis, particle image velocimetry or polarization microscopy. Instead of splitting the scattering vector into its components ( and ) it is more beneficial to use the polar coordinate system. 252 Here, the scattering vector is described by its absolute value and the angle between cylinder axis and the scattering vector. This results in the following equation for the scattering intensity of oriented cylindrical micelles258 with the form factor , the structure factor , the fraction of micelles with the angle and the number of micelles . The form factor is further defined by 65 The distribution function describes the cylindrical micelle orientation with the angle between the cylinder axis with the base vector and the director which defines the direction of the shear field. The expressions and are yielded from the factorization of the scattering amplitude into its cross sectionand length-contribution. The pointed brackets indicate the averaging across the corresponding size distributions of cylinders and radii. As described above, it is possible to express for typical bock copolymers, that have a core-shells-structure and a density-profile of , by using hypergeometric functions. is given by with and the ratios of the radii ( ) and densities ( ) compared to the shell. These are given by The form factor calculation requires a definition for the relation between the angles and , as shown in Fig. 23. Fig. 23 Three-dimensional illustration of the vector sphere. (Image from 252,258, Copyright Elsevier). 66 In this figure the director is determined by the angles and . The base vectors of the cylinder axes are positioned on a cone that is directed towards . Therefore, the range of angles of is given by and the integration of the function leads to The vector is calculated by using the rotation matrix based on the following equation The vectors and are given by and The rotation matrix is defined by with Lastly, is given by The orientation of anisotropic particles can be described by a range of distribution functions that are given by 67 with the parameter which can take on values between zero and infinity. The graphs of these different distribution functions are shown in Fig. 24. Fig. 24 The graphs of different distribution functions. (Image from 252, Copyright Elsevier). Among these functions, the Onsager and Maier-Saupe distributions can be pointed out because they are very important for the description of particle orientations in lyotropic and thermotropic liquid-crystalline systems. The distribution functions are normalized using the following factor The resulting mean deviation angle between cylinders and the director, which can take on values between 0° and 90°, is given by A very general function is the Laguerre distribution which changes from a Gauß function (at 68 ) to a Heaviside function (at very large ). The order parameter can take on values between 0 and 1. For a known distribution function it is given by The above equations are valid for diluted systems and with raising concentrations it becomes necessary to take the structure factor into account as described by van-der-Schoot284 with the cylinder concentration and with 69 70 4 Methods and Techniques 4.1 Photolithography Early materials that have been used for the fabrication of microfluidic devices include glass or silicon. 78 These materials were used because a wide range of microstructuring techniques were readily available from the field of semiconductor technology and because these materials offer resistance against a wide range of solvents. While glass is additionally optically transparent, silicon is opaque which limits microscopic applications for the latter material. A downside of both materials is the need for expensive clean room environments during their processing and the use of aggressive chemicals which makes it an expensive and resource intensive process. 57,285 The fast and effective fabrication of microfluidic devices in a short time became possible through the progress in the areas of photoand soft lithography. 56 Polymer-based materials offer more application friendly properties which is why they have gained importance during the past years. 230,286 One example for these are photoresists like SU-8 which are used in photolithography. 287 This technology enables the creation of microstructures based on specific technical drawings which are created using computer aided design (CAD) software. By using the computer-designed photo masks, photo resists are selectively exposed for the generation of microstructures. 56 These microstructures are then replicated using soft materials such as polymers to create microstructured stamps or microfluidic devices which is why this technology is called soft lithography. 55 Due to the precise control over design features and the shortened system optimization feedback loop, this process enables (simulation-based) rapid protoyping. 57 Furthermore, the replication templates, also called masters, can be re-used multiple times without quality loss enabling mass production and low fabrication costs. 288 The microstructuring of the photo resist happens by selective exposure using photo masks. 57 In case of a negative photoresist like SU-8, the non-cross-linked areas of the photoresist remain soluble and will be removed in the development stage of the lithographic process. The development bath typically contains 80 to 100% 2-methoxy-1-methylethyl acetate solution and is, similar to the photo resists, commercially available (mr-DEV 600, Microchem). Only the insoluble cross-linked areas of the photoresist remain on the substrate, typically a polished flat silicon wafer, and will be used as a master template for the subsequent replication steps. 55,56 The resolution of soft lithography is limited by diffraction and, hence, dependent on the wave length of the light source. 56 The range of available photo resist materials is only small due to the specific processing requirements. 71 Figure 25 Illustration of the rapid prototyping concept which describes the short designfeedback-loop that enables the fast optimization of microfluidic systems. (Image adapted and modified from 56) Figure 26 Fabrication of SU-8 master templates. A polished silicon wafer is spin coated using the negative photo resist SU-8. Using selective UV-exposure with an emulsion film mask, the microchannel geometry (i.e. mixing cross for hydrodynamic focusing) is received afer the development process which removes the non-cross-linked material. (Image from 13, Copyright PNAS) The photo resist EPON SU-8 is a material that is widely used for the fabrication of microfluidic devices. Developed by Shell Chemicals, this highly-functionalized monomer (69% in -butyrolactone) is capable of intermolecular cross-linking which enables the formation of three-dimensional structures with nanometer resolution. 289-292 Figure 27 Structural formula of EPON SU-8. 222,293,294 72 The mechanism of this cross-linking is a cationic ring opening polymerization (ROMP) of the epoxide that is started by photo-initiated aryl sulfonium salts. The lithographic process involves a UV-exposure step that leads to the formation of fluoroantimonic acid ( ). This is a strong Lewis acid that is formed in the presence of triarylsulfonium hexafluoroantimonate (3.3% solution in propylene carbonate, 4-Methyl-1,3dioxolan-2-one) and proton donors like the organic solvent. 295-297 This reaction mechanism is illustrated in Fig.28. Figure 28 Photochemical pathways during the generation of fluoroantimonic acid ( ).297,298 Since the photo initiator is consumed in an alternative reaction path during the UVexposure, it is critically important to adjust the light dose accordingly. Otherwise, the required initiator is not available in sufficient amounts for the acid-induced epoxide crosslinking during the subsequent “post exposure bake” step, which is the heating of the exposed photoresist. The reaction mechanism of the cationic ring opening polymerization of the SU-8 photo resist is shown in Fig.29. Figure 29 Catalytic reaction mechanism during the cross-linking of SU-8. 299 The epoxide group is protonated and the nucleophilic attack of the hydroxyl group of another SU-8 molecule leads to the opening of the epoxide ring. The catalyst is regenerated by the elimination of a proton and the cycle repeats itself which leads to a cross-linked network. 73 A key benefit of the microstructuring technique called soft lithography is its precise control of the microchannel geometries. It is therefore possible to combine different experimental techniques in a complementary way. Consequently, the exact reproduction of microchannel geometries is crucial for the combined use of different experimental techniques. Accordingly, microfocus SAXS-studies, microscopic experiments, such as polarization microscopy or microparticle image velocimetry (µPIV), and computational fluid dynamics (CFD) simulations have been used in an integrated approach that has been applied in the successful explanation of the perpendicular re-orientation effect (see chapter 7.1). We could show by polarization microscopy, µPIV-experiments and additional CFD simulations that the perpendicular orientation of wormlike micelles is the result of the interplay between the xorienting shearand y-orienting extensional forces; with x along the flow direction and y along the widening of the tapering. The re-orientation of anisotropic particles has been investigated in further detail in the following paper (see chapter 7.2). The established system of just-described complementary analysis methods has been applied for a systematic parameter screening, including the tapering ratios (-width, -length), the flow speeds and concentrations. Further, the rheological parameters have been varied systematically using CFD-simulation that are based on the sample’s experimentally measured rheological parameters. The system’s response to these parameter variations has been analyzed and quantified carefully. A relative ranking of the described system-controlling parameters could be derived from the combined experimental and theoretical results. The results of this work (chapters 7.1 & 7.2), and the perpendicular re-orientation effect in general, are of great importance for application that require orientation control, such as injection molding, fiber spinning or processing of composite materials. The second (chapter 7.2) paper also describes an improved version of the closed-channel Xray compatible microfluidic device that is based on a recently published paper. 132 These devices are made of the UV-curable adhesive NOA81 (see chapter 2.1 and 4.2) as the device material. However, the published design also has two weak points. The first is related the tubing-interface which glues PDMS onto the flat NOA81-device, punches a hole through the complete structure and seals it again the self-adhesive Tape. This reduces the pressure resistance as well as the solvent compatibility of the device (see chapter 2.1). The second flaw is the lack of the device’s height control. Since a correct background subtraction is essential for SAXS-experiments, especially for dilute or weakly-scattering samples, a homogenous device height is of great importance, as is the overall material thickness of the microfluidic device. The improved fabrication of NOA81-devices is described in the experimental section while its illustration is presented in the supplemental section of this paper (chapter 7.2). The first improvement involves the integration of bridging structures that are used of controlling the device height by adjusting the photoresist layer height through spin-coating. This leads to very even and extremely thin and stable microfluidic chips, reducing the X-ray background 80 signal and improving the signal transmission. The second improvement is related to the tubing interface. During the molding or replication step, a tubing is punched through the (PDMS- )microstructure and serves as a temporary template, enabling the direct connection of the tubing to the finished replica. As a consequence, the solvent compatibility of NOA81 is fully maintained and many polar and unpolar solvents, which are incompatible with PDMS, can now be used in this purely NOA81-based devices without issues. Additionally the pressure resistance of these devices is improved and can even be enhanced by glueing tubing to the device. Further, these improved devices run more reliably, have a smaller device footprint enabling denser arrays of microchannels, better optical properties and lower background signal compared to PDMS-Kapton-devices. Above all, the fabrication is easier due to the precise height control and much quicker due to the fast UV-exposure times (under one minute) compared to tens of minutes, or even hours, for thermally cured PDMS. Another microfluidic device that is also made of this material (NOA81) is described in chapter 7.4. This variation is closed and not actively flowing, but therefore only very small sample amounts are required for the SAXS-measurement of sheared samples. With only 25 µl of fluid volume, the nanoparticle lyotropic gel sample is applied to a millimeter-sized patch of the microstructured grid (14 µm spacing) while being sheared in the process. The microgrid is then sealed against evaporation using Kapton tape. Hence, this device demonstrates the high sample efficiency and ease of use when it comes to experiments at synchrotrons. Additionally, this work was essential for the improvement and optimization of soft lithographic fabrications techniques related to NOA81 that lead to the successful fabrication of high-resolution NOA81-based X-ray compatible microflow devices which are presented in chapter 7.2. The next paper describes a microfluidic liquid jet system (see chapter 7.3). As an example of an open microfluidic sample environment, it is designed for experiments at highly brilliant Xray sources or X-ray free electron lasers. The liquid jet is generated based on the gas-dynamic virtual nozzle design (GDVN) which represents the current state of the art at XFELs. 6 For this reason this microfluidic device runs essentially clogging-free and highly reliable over long periods of time. 53 The goal of this paper was to create a microfluidic liquid jet design that is easy and fast to fabricate, since the current glass capillary fabrication is a complex and manually challenging procedure. The creation of glass-based nozzles requires a skilled producer because the fabrication involves manual steps like grinding, flame-polishing and alignment of the capillaries. In contrast, the here presented microfluidic liquid jet nozzles can easily be replicated by using standard PDMS-device fabrication steps which are very fast to learn. The microfluidic devices further offer the benefit of a parallelized nozzle design which enables to create complex jet-in-jet-focusing geometries or complete arrays of multiple nozzle simultaneously in one fabrication sequence. Consequently the device footprint is very small which enables fast nozzle changes by simply switching to the adjacent nozzles. 81 The fabrication of these microfluidic liquid jet devices became possible due to our advances in multi-layered soft lithography. This enables the fabrication of 3D-microchannels by sealing two matching halves of microstructured PDMS that are treated with an air plasma. Since the micrometer-precise alignment would be manually challenging, the replication templates already incorporate alignment structures by design. These snap-in alignment structures could be integrated through the multi-layered microstructure design and they work similarly to Lego® building blocks, locking the microstructures of the two PDMS-halves in their right position during the sealing step. A key benefit of this microfluidic device over the state of the art glass capillaries is the highly reproducible design that is enabled through soft lithography. This micrometer-precise control over the nozzle geometries is critically important for the optimization of the liquid jets, especially at small flow rates. As described in chapter 2.4, the geometric parameters strongly affect the minimum flow rate for stable jetting which is the key factor for the high sample efficiency of these devices. Taking advantage of this design control, the relevant parameters, such as liquid flow rate and pressure difference, were varied to study the liquid jet dynamics. We found that the microfluidic liquid jet diameters can adjusted with great control by varying the pressures & flow rates and that they are in good agreement with available analytical expressions for the prediction of jet diameters in plate-orifice geometries. 201 This has also been verified by in-situ environmental scanning electron microscopy of the liquid jet exiting the nozzle. Furthermore, the variations of the above parameters allowed to control the jet breakup type. These jet breakup transitions have been studied using highspeed video microscopy and we also found that the stable jet’s column length can be controlled over a wide length scale. The jet shape could be predicted successfully by using time-resolved non-linear 3D CFDsimulations that describe the two-phase flow of fluids. The simulated shape of this theoretical liquid jet is in good agreement with the experimental microscopic results. The jet’s 3D shape was also studied by using confocal laser scanning microscopy. Since this study involved an enhanced design that incorporated a jet in jet hydrodynamic focusing geometry, before the liquid jet is shaped by the gas sheath, the fluid flow within the liquid jet could also be studied using in these experiments. Interestingly, we find that the flow of the two dye solutions (rhodamine B and fluorescein) is inverted and the inner focused liquid stream can be observed at the outside of the liquid jet as it exits the nozzle geometry. This inversion might originate from the inversion of speeds at the outer boundaries of the liquid that lead to the generation of vortices countering the liquid flow. 207 The velocity profile is parabolic inside the microchannel, therefore fastest in the liquid stream’s center and slowest close to the walls, this scenario changes quickly as the liquid exits the microchannel. Suddenly, the fastest liquid velocity can be found at the outside where the fast-flowing gas is in contact with the liquid. This kind of diffusion-controlled mixing inside the liquid jet could 82 enable new kinds of experiments at XFELs, such as the femtosecond-pulse diffraction of in situ nucleation and growth processes. The microfluidic concepts and lithographic microfabrication techniques have also been used for creating solutions for applications that are not related to X-ray experiments. As presented in chapter 7.5, a CFD-model has been created which describes non-linear interaction between the fluid and the microchannel walls under high pressures and high flow rates. The CFD-simulations showed how PDMS-microchannels with high aspect ratios deformed under these unusual conditions towards a rounded shape. This round shape was in good agreement with the microscopic images. We further found that this round shape of microchannel reduces the wall contact of flowing species which were mixed by hydrodynamic focusing before entering this round passage. Therefore, clogging of the device can be strongly reduced or eliminated. The creation of this CFD-model have lead to deeper insights about the simulation software and its capabilities and laid a foundation for the other non-linear CFD-models described above. Soft lithography is essential for the work of the presented papers for which the foundation has been laid with the following two papers. The paper described in chapter 7.6 describes the combination of nanotechnology and microstructuring techniques for developments towards sensoric applications. This work required the creation of deep, high aspect ratio, SU-8 microchannels. The optimization the parameters during the UV-lithographic process lead to better-defined and improved microchannel wall geometries for all microfluidic devices of this thesis. The work described in chapter 7.7 focuses on increasing the soft lithographic resolution of microstructured templates and an optimized replication thereof. This optimization resulted in very small and well-defined letters of high resolution that served as a micro contact printing stamp for the defined deposition of spherical polyelectrolyte brushes. This increased lithographic resolution lead to the capability of fabricating very small microfluidic features, such as narrow taperings (chapter 7.1 & 7.2), narrow grids (chapter 7.4) and small nozzles (chapter 7.3). Hence, the lithographic developments of these two papers played a crucial role for the successful application of soft lithography in the other papers by increasing the quality of high aspect ratio microchannel walls as well as the increasing resolution of very small microstructures. The following chapter will describe the authors’ contribution to these papers, followed by the papers in the subsequent chapter. 83 References 1. M. Antonietti and S. Förster, Advanced Materials, 2003, 15, 1323–1333. 2. S. Förster, Topics in Current Chemistry, 2003, 226, 1–28. 3. S. Förster, B. Berton, H. P. Hentze, E. Krämer, M. Antonietti, and P. Lindner, Macromolecules, 2001, 34, 4610–4623. 4. I. Robinson, G. Gruebel, and S. Mochrie, New J. Phys., 2010, 12. 5. T. Narayanan, Curr Opin Colloid in, 2009. 6. H. N. Chapman, P. Fromme, A. Barty, T. A. White, R. A. Kirian, A. Aquila, M. S. Hunter, J. Schulz, D. P. DePonte, U. Weierstall, R. B. Doak, F. R. N. C. Maia, A. V. Martin, I. Schlichting, L. Lomb, N. Coppola, R. L. Shoeman, S. W. Epp, R. Hartmann, D. Rolles, A. Rudenko, L. Foucar, N. Kimmel, G. Weidenspointner, P. Holl, M. Liang, M. Barthelmess, C. Caleman, S. Boutet, M. J. Bogan, J. Krzywinski, C. Bostedt, S. Bajt, L. Gumprecht, B. Rudek, B. Erk, C. Schmidt, A. Hömke, C. Reich, D. Pietschner, L. Strüder, G. Hauser, H. Gorke, J. Ullrich, S. Herrmann, G. Schaller, F. Schopper, H. Soltau, K.-U. Kühnel, M. Messerschmidt, J. D. Bozek, S. P. Hau-Riege, M. Frank, C. Y. Hampton, R. G. Sierra, D. Starodub, G. J. Williams, J. Hajdu, N. Timneanu, M. M. Seibert, J. Andreasson, A. Rocker, O. Jönsson, M. Svenda, S. Stern, K. Nass, R. Andritschke, C.-D. Schröter, F. Krasniqi, M. Bott, K. E. Schmidt, X. Wang, I. Grotjohann, J. M. Holton, T. R. M. Barends, R. Neutze, S. Marchesini, R. Fromme, S. Schorb, D. Rupp, M. Adolph, T. Gorkhover, I. Andersson, H. Hirsemann, G. Potdevin, H. Graafsma, B. Nilsson, and J. C. H. Spence, Nature, 2011, 470, 73–77. 7. D. H. Bilderback and P. Elleaume, Journal of Physics B: …, 2005. 8. H. A. Stone, A. D. Stroock, and A. Ajdari, Annu. Rev. Fluid Mech., 2004, 36, 381–411. 9. S. Köster and T. Pfohl, Mod. Phys. Lett. B, 2012, 26, 1230018. 10. T. Squires and S. Quake, Rev. Mod. Phys., 2005, 77, 977–1026. 11. J. Als-Nielsen and D. McMorrow, Elements of Modern X-ray Physics, Wiley, Chichester, 2011. 12. O. Glatter and O. Kratky, Academic Press, 1982. 13. M. Trebbin, D. Steinhauser, J. Perlich, A. Buffet, S. V. Roth, W. Zimmermann, J. Thiele, and S. Förster, PNAS, 2013, 110, 6706–6711. 14. S. M. Taheri, S. Fischer, M. Trebbin, S. With, J. H. Schröder, J. Perlich, S. V. Roth, and S. Förster, Soft Matter, 2012, 8, 12124. 15. L. Daubersies, J. Leng, and J.-B. Salmon, Lab Chip, 2013, 13, 910. 16. C. Riekel, P. Engström, and C. Martin, J. of Macromolecular Sc., Part B, 1998, B37, 587– 599. 17. R. J. Davies, M. Burghammer, and C. Riekel, Synchrotron Radiation in Natural Science, 2006, 5, 96–99. 18. S. V. Roth, R. Döhrmann, M. Dommach, M. Kuhlmann, I. Kröger, R. Gehrke, H. Walter, C. Schroer, B. Lengeler, and P. Müller-Buschbaum, Rev. Sci. Instrum., 2006, 77, 085106. 84 19. A. Timmann, R. Döhrmann, T. Schubert, H. Schulte-Schrepping, U. Hahn, M. Kuhlmann, R. Gehrke, S. V. Roth, A. Schropp, C. Schroer, and B. Lengeler, Rev. Sci. Instrum., 2009, 80, 046103. 20. C. Riekel, M. Burghammer, and R. Davies, IOP Conf. Ser.: Mater. Sci. Eng., 2010, 14, 012013. 21. A. Buffet, A. Rothkirch, R. Döhrmann, V. Körstgens, M. M. Abul Kashem, J. Perlich, G. Herzog, M. Schwartzkopf, R. Gehrke, P. Müller-Buschbaum, and S. V. Roth, J. Synchrotron Rad (2012). 19, 647-653 [doi:10.1107/S0909049512016895], 2012, 1–7. 22. A. Martel, M. Burghammer, R. Davies, E. DiCola, P. Panine, J.-B. Salmon, and C. Riekel, Biomicrofluidics, 2008, 2, 024104. 23. A. Otten, S. Köster, B. Struth, A. Snigirev, and T. Pfohl, J Synchrotron Rad, 2005, 12, 745–750. 24. A. Snigirev, V. Kohn, I. Snigireva, and B. Lengeler, Nature, 1996. 25. B. Struth, AIP, 2004, vol. 705, pp. 804–807. 26. C. G. Schroer and B. Lengeler, Phys. Rev. Lett., 2005, 94, 054802. 27. M. Müller, C. Riekel, R. Vuong, and H. Chanzy, Polymer, 2000, 41, 2627–2632. 28. C. Riekel and R. J. Davies, Curr Opin Colloid in, 2005, 9, 396–403. 29. R. J. Davies, M. Burghammer, and C. Riekel, Macromolecules, 2007, 40, 5038–5046. 30. A. Martel, M. Burghammer, R. J. Davies, E. Di Cola, C. Vendrely, and C. Riekel, J. Am. Chem. Soc., 2008, 130, 17070–17074. 31. C. Riekel, M. Rossle, D. Sapede, and F. Vollrath, Naturwissenschaften, 2004, 91, 30–33. 32. C. Riekel, B. Madsen, D. Knight, and F. Vollrath, Biomacromolecules, 2000, 1, 622–626. 33. C. Riekel, M. Müller, and F. Vollrath, Macromolecules, 1999, 32, 4464–4466. 34. M. E. Kinahan, E. Filippidi, S. Köster, X. Hu, H. M. Evans, T. Pfohl, D. L. Kaplan, and J. Wong, Biomacromolecules, 2011, 12, 1504–1511. 35. J. F. Moulin, S. V. Roth, and P. Müller-Buschbaum, Rev. Sci. Instrum., 2008, 79, 015109. 36. N. Stribeck, A. A. Camarillo, U. Nöchel, C. Schroer, M. Kuhlmann, S. V. Roth, R. Gehrke, and R. K. Bayer, Macromolecular Chemistry and Physics, 2006, 207, 1139–1149. 37. N. Stribeck, U. Nöchel, and A. Almendarez-Camarillo, Macromolecular Chemistry and Physics, 2008, 209, 1976–1982. 38. J. M. Feldkamp, M. Kuhlmann, S. V. Roth, A. Timmann, R. Gehrke, I. Shakhverdova, P. Paufler, S. K. Filatov, R. S. Bubnova, and C. G. Schroer, Phys. Status Solidi (a), 2009, 206, 1723–1726. 39. C. G. Schroer, M. Kuhlmann, S. V. Roth, R. Gehrke, N. Stribeck, A. AlmendarezCamarillo, and B. Lengeler, Appl. Phys. Lett., 2006, 88, 164102–164102–3. 40. M. Kuhlmann, J. M. Feldkamp, S. V. Roth, and C. G. Schroer, 2008, pp. 570–573. 41. J. M. Feldkamp, Dissertation, 2009, 1–136. 42. A. C. Kak and M. Slaney, IEEE Press, 1988. 43. A. Gonzalez and C. Nave, Acta Crystallogr D Biol Crystallogr, 1994, 50, 874–877. 44. R. Henderson, Proceedings of the Royal Society B: Biological Sciences, 1990, 241, 6–8. 85 45. S. Cusack, H. Belrhali, A. Bram, M. Burghammer, A. Perrakis, and C. Riekel, Nat. Struct. Biol., 1998, 5, 634–637. 46. L. Pollack, M. Tate, N. Darnton, J. Knight, S. Gruner, W. Eaton, and R. Austin, PNAS, 1999, 96, 10115–10117. 47. L. Pollack, M. W. Tate, A. C. Finnefrock, C. Kalidas, S. Trotter, N. C. Darnton, L. Lurio, R. H. Austin, C. A. Batt, S. M. Gruner, and S. Mochrie, Phys. Rev. Lett., 2001, 86, 4962–4965. 48. R. Russell, I. S. Millettt, M. W. Tate, L. W. Kwok, B. Nakatani, S. M. Gruner, S. Mochrie, V. Pande, S. Doniach, D. Herschlag, and L. Pollack, Proceedings of the National Academy of Sciences, 2002, 99, 4266–4271. 49. R. Neutze, R. Wouts, D. van der Spoel, E. Weckert, and J. Hajdu, nature, 2000, 406, 752–757. 50. L. Lomb, T. Barends, S. Kassemeyer, A. Aquila, S. Epp, B. Erk, L. Foucar, R. Hartmann, B. Rudek, D. Rolles, A. Rudenko, R. Shoeman, J. Andreasson, S. Bajt, M. Barthelmess, A. Barty, M. Bogan, C. Bostedt, J. Bozek, C. Caleman, R. Coffee, N. Coppola, D. DePonte, R. B. Doak, T. Ekeberg, H. Fleckenstein, P. Fromme, M. Gebhardt, H. Graafsma, L. Gumprecht, C. Hampton, A. Hartmann, G. Hauser, H. Hirsemann, P. Holl, J. Holton, M. Hunter, W. Kabsch, N. Kimmel, R. Kirian, M. Liang, F. R. N. Maia, A. Meinhart, S. Marchesini, A. Martin, K. Nass, C. Reich, J. Schulz, M. M. Seibert, R. Sierra, H. Soltau, J. C. Spence, J. Steinbrener, F. Stellato, S. Stern, N. Timneanu, X. Wang, G. Weidenspointner, U. Weierstall, T. White, C. Wunderer, H. Chapman, J. Ullrich, L. Strüder, and I. Schlichting, Phys. Rev. B, 2011, 84, 214111. 51. A. Barty, C. Caleman, A. Aquila, N. Timneanu, L. Lomb, T. A. White, J. Andreasson, D. Arnlund, S. Bajt, T. R. M. Barends, M. Barthelmess, M. J. Bogan, C. Bostedt, J. D. Bozek, R. Coffee, N. Coppola, J. Davidsson, D. P. DePonte, R. B. Doak, T. Ekeberg, V. Elser, S. W. Epp, B. Erk, H. Fleckenstein, L. Foucar, P. Fromme, H. Graafsma, L. Gumprecht, J. Hajdu, C. Y. Hampton, R. Hartmann, A. Hartmann, G. Hauser, H. Hirsemann, P. Holl, M. S. Hunter, L. Johansson, S. Kassemeyer, N. Kimmel, R. A. Kirian, M. Liang, F. R. N. C. Maia, E. Malmerberg, S. Marchesini, A. V. Martin, K. Nass, R. Neutze, C. Reich, D. Rolles, B. Rudek, A. Rudenko, H. Scott, I. Schlichting, J. Schulz, M. M. Seibert, R. L. Shoeman, R. G. Sierra, H. Soltau, J. C. H. Spence, F. Stellato, S. Stern, L. Strüder, J. Ullrich, X. Wang, G. Weidenspointner, U. Weierstall, C. B. Wunderer, and H. N. Chapman, Nature Photon, 2011, 6, 35–40. 52. J. Ullrich, A. Rudenko, and R. Moshammer, Annu Rev Phys Chem, 2012, 63, 635–660. 53. D. P. DePonte, U. Weierstall, K. Schmidt, J. Warner, D. Starodub, J. C. H. Spence, and R. B. Doak, J. Phys. D: Appl. Phys., 2008, 41, 195505. 54. A. M. Gañán-Calvo, D. P. DePonte, M. A. Herrada, J. C. H. Spence, U. Weierstall, and R. B. Doak, Small, 2010, 6, 822–824. 55. Y. Xia and G. M. Whitesides, Annual review of materials science, 1998, 28, 153–184. 56. G. M. Whitesides and Y. Xia, Angew. Chem. Int. Ed., 1998, 37, 550–575. 86 57. D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984. 58. G. M. Whitesides, Nature, 2006, 442, 368–373. 59. C.-M. Ho and Y.-C. Tai, Annu. Rev. Fluid Mech., 1998, 30, 579–612. 60. A. Groisman, M. Enzelberger, and S. Quake, Science, 2003, 300, 955–958. 61. J. Thiele, University of Bayreuth, 2011. 62. G. M. Whitesides and A. D. Stroock, Physics Today, 2001, 54, 42. 63. A. Stroock and G. M. Whitesides, Electrophoresis, 2002, 23, 3461–3473. 64. A. Manz, N. GRABER, and H. M. WIDMER, Sensors and Actuators B: Chemical, 1990, 1, 244–248. 65. A. R. Abate, M. B. Romanowsky, J. J. Agresti, and D. A. Weitz, Appl. Phys. Lett., 2009, 94, 023503. 66. A. R. Abate, T. Hung, P. Mary, J. J. Agresti, and D. A. Weitz, Proceedings of the National Academy of Sciences, 2010, 107, 19163–19166. 67. T. Thorsen, S. J. Maerkl, and S. R. Quake, Science, 2002, 298, 580–584. 68. Y.-C. Tan, V. Cristini, and A. P. Lee, Sensors and Actuators B: Chemical, 2006, 114, 350– 356. 69. W. Xu and S. J. Muller, Lab Chip, 2011, 11, 435–442. 70. T. Pfohl, Analyse, Manipulation und Aggregation von biologischen Makromolekülen im mikrofluidischen Scherfluss, De Gruyter, 2010. 71. T. Franke and A. Wixforth, Phys. Unserer Zeit, 2007, 38, 88–94. 72. A. R. Abate, J. J. Agresti, and D. A. Weitz, Appl. Phys. Lett., 2010, 96, 203509. 73. D. Therriault, S. R. White, and J. A. Lewis, Nat. Mater., 2003, 2, 265–271. 74. S. K. Sia and G. M. Whitesides, Electrophoresis, 2003, 24, 3563–3576. 75. C. Xi, D. L. Marks, D. S. Parikh, L. Raskin, and S. A. Boppart, Proceedings of the National Academy of Sciences, 2004, 101. 76. D. S. Kim, S. H. Lee, T. H. Kwon, and C. H. Ahn, Lab Chip, 2005, 5, 739. 77. S. Hardt, K. S. Drese, V. Hessel, and F. Schönfeld, Microfluid Nanofluid, 2005, 1, 108– 118. 78. S. Quake and A. Scherer, Science, 2000, 290, 1536–1540. 79. J. W. Hong and S. R. Quake, Nature biotechnology, 2003, 21, 1179–1183. 80. J. W. Hong, V. Studer, G. Hang, W. F. Anderson, and S. R. Quake, Nature biotechnology, 2004, 22, 435–439. 81. C. Lee, G. Sui, A. Elizarov, C. Shu, and Y. Shin, Science, 2005, 310, 1793–1796. 82. J.-C. Baret, O. J. Miller, V. Taly, M. Ryckelynck, A. El-Harrak, L. Frenz, C. Rick, M. L. Samuels, J. B. Hutchison, J. J. Agresti, D. R. Link, D. A. Weitz, and A. D. Griffiths, Lab Chip, 2009, 9, 1850. 83. H. Shafiee, M. B. Sano, E. A. Henslee, J. L. Caldwell, and R. V. Davalos, Lab Chip, 2010, 10, 438. 84. T. Q. Chastek, K. Iida, E. J. Amis, M. J. Fasolka, and K. L. Beers, Lab Chip, 2008, 8, 950. 87 85. K. Iida, T. Q. Chastek, K. L. Beers, K. A. Cavicchi, J. Chun, and M. J. Fasolka, Lab Chip, 2009, 9, 339. 86. C. Hany, H. Lebrun, C. Pradere, J. Toutain, and J.-C. Batsale, Chemical Engineering Journal, 2010, 160, 814–822. 87. H. Pennemann, P. Watts, S. J. Haswell, V. Hessel, and H. Lowe, Org. Process Res. Dev., 2004, 8, 422–439. 88. S. Seiffert and D. A. Weitz, Polymer, 2010, 51, 5883–5889. 89. E. Tumarkin and E. Kumacheva, Chem. Soc. Rev., 2009, 38, 2161. 90. C. Priest, A. Quinn, A. Postma, A. N. Zelikin, J. Ralston, and F. Caruso, Lab Chip, 2008, 8, 2182. 91. Y. Morimoto, W.-H. Tan, Y. Tsuda, and S. Takeuchi, Lab Chip, 2009, 9, 2217–2223. 92. Y. Song, J. Hormes, and C. S. S. R. Kumar, Small, 2008, 4, 698–711. 93. Y. Song, H. Modrow, L. L. Henry, C. K. Saw, E. E. Doomes, V. Palshin, J. Hormes, and C. S. S. R. Kumar, Chem. Mater., 2006, 18, 2817–2827. 94. A. Jahn, J. Reiner, W. Vreeland, and D. DeVoe, Journal of Nanoparticle Research, 2008. 95. P. Kallio and J. Kuncova, TEKES Technology review, 2004. 96. P. S. Dittrich and A. Manz, Nat Rev Drug Discov, 2006, 5, 210–218. 97. W. Li, J. Greener, D. Voicu, and E. Kumacheva, Lab Chip, 2009, 9, 2715. 98. R. S. Ramsey and J. M. Ramsey, Anal. Chem., 1997, 69, 1174–1178. 99. J. Pihl, M. Karlsson, and D. T. Chiu, Drug Discov. Today, 2005, 10, 1377–1383. 100. A. R. Wheeler, W. R. Throndset, and R. J. Whelan, Analytical …, 2003. 101. A. A. Werdich, E. A. Lima, B. Ivanov, I. Ges, M. E. Anderson, J. P. Wikswo, and F. J. Baudenbacher, Lab Chip, 2004, 4, 357. 102. R. T. Turgeon and M. T. Bowser, Anal Bioanal Chem, 2009, 394, 187–198. 103. I. K. Dimov, L. Basabe-Desmonts, J. L. Garcia-Cordero, B. M. Ross, A. J. Ricco, and L. P. Lee, Lab Chip, 2011, 11, 845–850. 104. J. Melin and S. R. Quake, Annual review of biophysics and biomolecular structure, 2007, 36, 213–231. 105. M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer, and S. R. Quake, Science, 2000, 288, 113–116. 106. Y. Huang, P. Castrataro, C.-C. Lee, and S. R. Quake, Lab Chip, 2006, 7, 24. 107. C. Zhang, J. Xu, W. Ma, and W. Zheng, Biotechnol. Adv., 2006, 24, 243–284. 108. E. A. Ottesen, J. W. Hong, S. R. Quake, and J. R. Leadbetter, Science, 2006, 314, 1464–1467. 109. I. P. G. Marshall, P. C. Blainey, A. M. Spormann, and S. R. Quake, Applied and Environmental Microbiology, 2012, 78, 8555–8563. 110. S. J. Pamp, E. D. Harrington, S. R. Quake, D. A. Relman, and P. C. Blainey, Genome Res., 2012, 22, 1107–1119. 111. J. Wang, H. C. Fan, B. Behr, and S. R. Quake, Cell, 2012, 150, 402–412. 112. F. E. Dewey, S. Pan, M. T. Wheeler, S. R. Quake, and E. A. Ashley, Circulation, 2012, 125, 931–944. 113. D. Pushkarev, N. F. Neff, and S. R. Quake, Nature biotechnology, 2009, 27, 847–850. 88 114. H. C. Fan, Y. J. Blumenfeld, Y. Y. El-Sayed, J. Chueh, and S. R. Quake, YMOB, 2009, 200, 543.e1–543.e7. 115. R. A. White, P. C. Blainey, H. C. Fan, and S. R. Quake, BMC Genomics, 2009, 10, 116. 116. M. Meier, R. Sit, W. Pan, and S. R. Quake, Anal. Chem., 2012, 84, 9572–9578. 117. M. Meier, R. V. Sit, and S. R. Quake, Proceedings of the National Academy of Sciences, 2013, 110, 477–482. 118. C. L. Hansen, E. Skordalakes, J. M. Berger, and S. R. Quake, Proceedings of the National Academy of Sciences, 2002, 99, 16531. 119. C. Hansen and S. R. Quake, Current opinion in structural biology, 2003, 13, 538–544. 120. B. Zheng, J. D. Tice, and R. F. Ismagilov, Advanced Materials, 2004, 16, 1365–1368. 121. C. L. Hansen, S. Classen, J. M. Berger, and S. R. Quake, J. Am. Chem. Soc., 2006, 128, 3142–3143. 122. B. Zheng, J. Tice, L. Roach, and R. Ismagilov, Angewandte Chemie (International ed. in English), 2004, 43, 2508. 123. J. D. Ng, P. J. Clark, R. C. Stevens, and P. Kuhn, Acta Cryst (2008). D64, 189-197 [doi:10.1107/S0907444907060064], 2008, 1–9. 124. J. Leng and J.-B. Salmon, Lab Chip, 2009, 9, 24. 125. S. Emamzadah, T. J. Petty, V. De Almeida, T. Nishimura, J. Joly, J.-L. Ferrer, and T. D. Halazonetis, Acta Cryst (2009). D65, 913-920 [doi:10.1107/S0907444909021489], 2009, 1– 8. 126. L. Li and R. F. Ismagilov, Annu. Rev. Biophys., 2010, 39, 139–158. 127. S. L. P. A. S. P. S. K. N. A. P. J. A. K. S Guha, 2011, 1–3. 128. M. Maeki, S. Yoshizuka, H. Yamaguchi, M. Kawamoto, K. Yamashita, H. Nakamura, M. Miyazaki, and H. Maeda, Anal Sci, 2012, 28, 65. 129. J. B. Knight, A. Vishwanath, J. P. Brody, and R. H. Austin, Phys. Rev. Lett., 1998, 80, 3863–3866. 130. T. Pfohl, A. Otten, S. Köster, R. Dootz, B. Struth, and H. M. Evans, Biomacromolecules, 2007, 8, 2167–2172. 131. S. Köster, H. M. Evans, J. Y. Wong, and T. Pfohl, Biomacromolecules, 2008, 9, 199–207. 132. M. E. Brennich, J.-F. Nolting, C. Dammann, B. Nöding, S. Bauch, H. Herrmann, T. Pfohl, and S. Köster, Lab Chip, 2011, 11, 708. 133. C. Reich, M. B. Hochrein, B. Krause, and B. Nickel, Rev. Sci. Instrum., 2005, 76, 095103. 134. T. Pfohl, F. Mugele, R. Seemann, and S. Herminghaus, ChemPhysChem, 2003, 4, 1291– 1298. 135. M. Trebbin, S. With, A. Mark, C. Hanske, A. Buffet, G. Santoro, S. Yu, J. Perlich, S. V. Roth, and S. Förster, in preparation, 2013. 136. M. Trebbin, K. Krüger, D. DePonte, J. Schulz, S. V. Roth, H. N. Chapman, and S. Förster, in preparation, 2013. 137. J. Thiele, M. Windbergs, A. R. Abate, M. Trebbin, H. C. Shum, S. Förster, and D. A. Weitz, Lab Chip, 2011, 11, 2362. 89 288. M. E. Sandison and H. Morgan, J. Micromech. Microeng., 2005, 15, S139–S144. 289. K. Y. Lee, J. Vac. Sci. Technol. B, 1995, 13, 3012. 290. M. M. A. A. K. Mohamed, 2010, 1–26. 291. J. Simčič, P. Pelicon, Z. Rupnik, M. Mihelič, A. Razpet, D. Jenko, and M. Maček, Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 2005, 241, 479–485. 292. C. G. Khan Malek, Microelectronics Journal, 2002, 33, 101–105. 293. J. Zhang, K. Tan, G. Hong, L. Yang, and H. Gong, J. Micromech. Microeng., 2001, 11, 20–26. 294. J. Zhang, K. Tan, and H. Gong, Polym. Test., 2001, 20, 693–701. 295. Y. Wang, M. Bachman, C. Sims, G. Li, and N. Allbritton, Langmuir, 2006, 22, 2719– 2725. 296. J. Dektar and N. Hacker, J. Am. Chem. Soc., 1990, 112, 6004–6015. 297. MicroChem NANOTM SU-8 Organic Resin Solution, MSDS, 12.12.2001, www.cise.columbia.edu/clean/msds/su-8resist.pdf (1.07.2013), 2009. 298. Y. Wang, J.-H. Pai, H.-H. Lai, C. E. Sims, M. Bachman, G. P. Li, and N. L. Allbritton, J. Micromech. Microeng., 2007, 17, 1371–1380. 299. Guerin, L. The SU-8 homepage, http://www.geocities.com/guerinlj/ (17.07.2008). 2009. 300. X. Yan and P. Gu, Computer-Aided Design, 1996. 301. E. Sachs, S. Allen, H. Guo, and B. Banos, 1997. 302. E. Sachlos and J. T. Czernuszka, Eur Cell Mater, 2003. 303. C. M. Cheah, C. K. Chua, C. W. Lee, C. Feng, and K. Totong, Int J Adv Manuf Technol, 2004, 25, 308–320. 304. A. Khademhosseini, PNAS, 2006, 103, 2480–2487. 305. C. J. Campbell, S. K. Smoukov, K. J. M. Bishop, E. Baker, and B. A. Grzybowski, Advanced Materials, 2006, 18, 2004–2008. 306. P. J. Kitson, M. H. Rosnes, V. Sans, V. Dragone, and L. Cronin, Lab Chip, 2012, 12, 3267–3271. 307. K. B. Anderson, S. Y. Lockwood, R. S. Martin, and D. M. Spence, Anal. Chem., 2013, 85, 5622–5626. 308. V. Dragone, V. Sans, M. H. Rosnes, P. J. Kitson, and L. Cronin, Beilstein J Org Chem, 2013, 9, 951–959. 309. Y. Liao, J. Song, E. Li, Y. Luo, Y. Shen, D. Chen, Y. Cheng, Z. Xu, K. Sugioka, and K. Midorikawa, Lab Chip, 2012, 12, 746. 310. A. D. Campo and C. Greiner, J. Micromech. Microeng., 2007, 17, R81–R95. 311. S. Jeon, J. U. Park, R. Cirelli, S. Yang, C. E. Heitzman, P. V. Braun, P. J. A. Kenis, and J. A. Rogers, Proceedings of the National Academy of Sciences, 2004, 101, 12428. 312. J. A. Lewis and G. M. Gratson, Materials Today, 2004, 7, 32–39. 313. Y. L. Zhang, Q. D. Chen, H. Xia, and H. B. Sun, Nano Today, 2010. 314. B. Zimmer, Nanoscribe GmbH, 2013, 1–14. 315. MSDS: Sylgard(R) 184 silicone elastomer ket (base), DOW Corning, v1.5, 2008, 2008. 316. MSDS: Sylgard(R) 184 silicone elastomer ket (curing agent), DOW Corning, v1.8, 2007, 2007. 96 317. H. Finkelmann and G. Rehage, Macromol. Rapid Commun., 1980, 1. 318. B. Krucke, H. Zaschke, S. Kostromin, and V. Shibaev, Acta Polym., 1985, 36. 319. 2007, 1–3. 320. E. Delamarche, A. Bernard, H. Schmid, A. Bietsch, B. Michel, and H. Biebuyck, J. Am. Chem. Soc., 1998, 120, 500–508. 321. G. M. Whitesides, E. Ostuni, S. Takayama, X. Jiang, and D. Ingber, Annu. Rev. Biomed. Eng., 2001, 3, 335–373. 322. N. B. Cramer, J. P. Scott, and C. N. Bowman, Macromolecules, 2002, 35, 5361–5365. 97 98 6 Individual contribution to joint publications The results which are presented in the scientific papers of this thesis are based on the collaborative work with others. The contribution of every co-author will be specified in further detail below and the corresponding author will be indicated with an asterisk (*). Chapter 7.1 This work is published in PNAS, 2013, 110, 6706–6711 and featured as a “Technical Highlight” in DESY Photon Science 2011 and on the front pages of the DESYand Uni Bayreuth web pages; entitled: "Anisotropic particles align perpendicular to the flow direction in narrow microchannels". by Martin Trebbin, Dagmar Steinhauser, Jan Perlich, Adeline Buffet, Stephan V. Roth, Walter Zimmermann, Julian Thiele*, and Stephan Förster*. (shared corresponding authorship) I performed most of the experiments & data analysis, created the non-linear non-Newtonian flow CFD-simulation model; and wrote the manuscript together with Julian Thiele and Stephan Förster. Dagmar Steinhauser and Julian Thiele helped with the experiments at the synchrotron beamlines at which Jan Perlich, Adeline Buffet and Stephan V. Roth provided the technical support. Walter Zimmermann was involved in scientific discussions and helped with the correction of the manuscript. Further, Stephan Förster supervised the project. Chapter 7.2 This work is prepared as a manuscript that is ready for submission; entitled: "Microfluidic SAXS for the high-throughput screening and correlation of complex fluid behavior with structural information". by Martin Trebbin, Sebastian With, Andres Mark, Christoph Hanske, Adeline Buffet, Gonzalo Santoro, Shun Yu, Jan Perlich, Stephan V. Roth, and Stephan Förster*. I performed most of the experiments & data analysis, created the non-linear non-Newtonian flow CFD-simulation model and wrote the manuscript. Sebastian With was involved in the device fabrication and helped with the experiments at the synchrotron beamlines at which Adeline Buffet, Gonzalo Santoro, Shun Yu, Jan Perlich, Stephan V. Roth provided the technical support. Andreas Mark was involved in the µPIV-experiments and their analysis. Christoph Hanske provided the tracer particles for the µPIV-experiments. Stephan Förster supervised the project and corrected the manuscript. 99 Chapter 7.3 This work is prepared as a manuscript that is ready for submission; entitled: "Microfluidic Liquid Jet System with compatibility for atmospheric and high-vacuum conditions". by Martin Trebbin, Kilian Krüger, Daniel DePonte, Stephan V. Roth, Henry N. Chapman, and Stephan Förster*. I performed most of the experiments & data analysis, created the non-linear two-phase flow CFD-simulation model and wrote the manuscript. Kilian Krüger was involved in the device fabrication, scientific discussions and helped with the experiments. Daniel DePonte, Stephan V. Roth and Henry N. Chapman were involved in scientific discussions and helped with the correction of the manuscript. Further, Daniel DePonte was involved in the ESEMexperiments. Stephan Förster supervised the project and corrected the manuscript. Chapter 7.4 This paper is published in Soft Matter 2012, 8, 12124; entitled: "Lyotropic phase behavior of polymer-coated iron oxide nanoparticles". by Sara Mehdizadeh Taheri*, Steffen Fischer, Martin Trebbin, Sebastian With, Jan H. Schröder, Jan Perlich, Stephan V. Roth and Stephan Förster*. I designed and fabricated the grid-based static microfluidic device using soft lithography. Sara Mehdizadeh Taheri performed most of the experiments & data analysis and wrote the manuscript with Stephan Förster. Further, I helped, together with Sebastian With, Jan H. Schröder with the experiments at the synchrotron beamlines at which Jan Perlich, Stephan V. Roth provided the technical support. Steffen Fischer was involved in the synthesis of polymers and scientific discussions. Stephan Förster supervised the project and corrected the manuscript. Chapter 7.5 This paper is published in Lab Chip, 2011, 11, 2362 and featured in Chemistry World; entitled: "Early development drug formulation on a chip: Fabrication of nanoparticles using a microfluidic spray dryer". by Julian Thiele, Maike Windbergs, Adam R. Abate, Martin Trebbin, Ho Cheung Shum, Stephan Förster, and David A. Weitz*. I developed the non-linear FEM-based CFD-simulation model and contributed to the simulation-related writing of the manuscript. Julian Thiele performed most of the experiments and wrote the manuscript. Adam Abate was involved in scientific discussions. Maike Windbergs performed the spray experiments in bulk and was involved in scientific discussions. Ho Cheung Shum conducted the SEM analysis of the drug. Stephan Förster 100 corrected the manuscript. David Weitz supervised the project. Parts of this work have been submitted for patenting. Chapter 7.6 This paper is published in Nanotechnology 2011, 22, 305303; entitled: "Freestanding films of crosslinked gold nanoparticles prepared via layer-by-layer spincoating". by Hendrik Schlicke, Jan H. Schröder, Martin Trebbin, Alexey Petrov, Michael Ijeh, Horst Weller and Tobias Vossmeyer*. I designed and fabricated the microstructures using computer aided design and photolithpgraphy, was involved in their microscopy as well as the scientific discussion. Jan H. Schröder performed most of the experiments and wrote the manuscript together with Tobias Vossmeyer. Hendrik Schlicke and Alexey Petrov were involved in the experiments related to the films and the scientific discussions. Michael Ijeh was involved in ligand synthesis. Horst Weller corrected the manuscript. Further, Tobias Vossmeyer supervised the project. Chapter 7.7 This paper is published in Z. Phys. Chem. 2012, 226, 569–584; entitled: "Adsorption of spherical polyelectrolyte brushes: from interactions to surface patterning". by Christoph Hanske, Johann Erath, Christin Kühr, Martin Trebbin, Christian Schneider, Alexander Wittemann, and Andreas Fery*. I produced a specially designed stamp for micro contact printing using soft lithography. Christoph Hanske performed adsorption experiments, the micro contact printing, analyzed these experiments, was involved in scientific discussions, and wrote the manuscript. Johann Erath performed the AFM interaction measurements, analyzed these experiments, was involved in scientific discussions, wrote parts of the manuscript, and corrected the manuscript. Christin Kühr synthesized the SPBs. Christian Schneider synthesized the model particles for SPBs. Alexander Wittemann developed the synthesis protocol for the SPBs, was involved in scientific discussions, wrote parts of the manuscript, and helped correcting the manuscript. Andreas Fery analyzed the results, helped with discussions, and corrected the manuscript. 101 102 7 Publications 103 104 7.1 Anisotropic particles align perpendicular to flowdirection in narrow microchannels Martin Trebbin,1 Dagmar Steinhauser,2,3 Jan Perlich,4 Stephan V. Roth,4 Walter Zimmermann,5 Julian Thiele,6* and Stephan Förster 1* 1) Physical Chemistry I, University of Bayreuth, D-95447 Bayreuth, Germany 2) Max-Planck-Institute for Dynamics and Self-Organization, D-37073 Göttingen, Germany 3) Deutsches Institut für Kautschuktechnologie, D-30519 Hannover, Germany 4) HASYLAB/DESY, D-22607 Hamburg, Germany 5) Theoretical Physics I, University of Bayreuth, D-95447 Bayreuth, Germany 6) Radboud University Nijmegen, Institute for Molecules and Materials, NL-6525 AJ, Nijmegen, The Netherlands The flow orientation of anisotropic particles in narrow channels is of importance in many fields ranging from the spinning and molding of fibers to the flow of cells and proteins through thin capillaries. It is commonly assumed that anisotropic particles align parallel to the flow direction. When flowing through narrowed channel sections one expects the increased flow rate to improve the alignment. Here we show by microfocus synchrotron X-ray diffraction and polarized optical microscopy for the first time that, after passing a narrow channel section, anisotropic colloidal particles align perpendicular to the flowdirection. We find this to be a general behaviour of anisotropic colloids, also observed for disk-like particles. The perpendicular alignment is stable, extending throughout the remaining part downstream the channel. We show by micro particle image velocimetry and finite element computational fluid dynamic simulations that the perpendicular orientation is due to the velocity field having large perpendicular gradients in the expansion zone after the narrow section. Shear-thinning, a typical property of anisotropic particles, promotes perpendicular extensional and orientation. Our discovery has important consequences when considering the flow orientation of polymers, micelles, fibers, proteins or cells through narrow channels, pipes or capillary sections. An immediate consequence for the production of fibers is the necessity for realignment by extension in flow direction. For fibrous proteins, reorientation and stable plug-flow are likely mechanisms for protein coagulation. 105 If reorientation in perpendicular direction in a channel expansion zone is a general property of anisotropic colloids, it should have its cause in the hydrodynamic flow pattern. Using micro particle image velocimetry we determined the velocity profile ),( yxv in the narrow section as shown in Fig. 4. Entering the channel section, there is first a contraction zone with planar extensional flow in flow (x-) direction, followed by an expansion zone with planar extensional flow in perpendicular (y-) direction. Fig. 4 A shows the polarized optical micrograph of the channel section indicating zones with parallel (blue) and perpendicular (yellow) flow orientation. Fig. 4B shows the measured flow velocity, obtained from velocimetry measurements of added 3.3 µm diameter tracer particles. As expected, the flow velocimetry is largest in the narrow section of the channel. Fig. 4 C shows the measured velocity profiles across the channel at position I before entering the contraction zone, and at position III after the expansion zone. The velocity profiles are both non-parabolic, a consequence of the shearthinning, non-Newtonian flow behavior of the micelles. Both are hydrodynamically stable states, yet having different velocity profiles. The pre-tapering velocity profile at position I has a broad, but clearly noticeable maximum, whereas the post-tapering velocity profile at position III is completely flat, indicating plug flow. The velocity profiles can be switched back and forth in subsequent narrow channels as shown in Fig. S5 (Supporting Information). Fig. 4 D shows the measured velocity components x v and y v across the channel at position II in the expansion zone. From the velocity components the shear rates xy v∇= γ ! and the extensional rates yy v−∇= ε ! can be calculated and are displayed in Fig. 4 E. We observe that in the middle part of the cross section the extensional rate ε ! is either larger, or at least of comparable magnitude to the shear rate γ ! . Extensional flows are much more effective in orienting and aligning anisotropic particles compared to shear flows. [20,21] They lead to reorientation if the extensional rates become comparable to the shear rates. With an internal relaxation time of the cylindrical micelles of 400≈ τ s as determined from the rheological measurements and extensional rates of 2≈ ε ! s-1 (see Figure 4 E), values of the Deborah number are 1800 >>== ετ ! De . Under these conditions the micelles are highly susceptible for flow-induced alignment. In Fig. 4 F we mapped regions where 14.0/ ≥ γε !! in yellow, and regions where 14.0/ < γε !! in blue for comparison with the flow birefringence pattern in Fig. 4 A. We note that the near-zero values of y v in the regions before the contraction zone and after the expansion zone of the channel lead to some scatter of the data. Yet, we observe that by choosing a threshold of 14.0/ = γε !! , regions of high extensional rates in Fig. 4 F agree well with regions of perpendicular orientation in Figs. 3 C and 4 A. At the channel walls shear flow dominates such that 14.0/ < γε !! and micelles remain oriented in flow-direction as observed experimentally. The beginning of the sharp rise of the shear rate γ ! close to the channel wall (see Figure 4 E) defines a relatively sharp transition with a stable interface between zones of perpendicular and parallel cylinder orientation. 112 To distinguish features of the flow pattern specifically related to particle anisotropy from features related to just channel geometry, we performed numerical computational fluid dynamic (CFD) simulations to calculate shear rates and extensional rates in the contraction/expansion zone. The calculations were done for Newtonian liquids, but also for non-Newtonian, shear-thinning liquids such as solutions of wormlike micelles. Shear-thinning was accounted for by measuring the shear-rate dependent viscosity of the micellar solution using a cone-plate rheometer and fitting the measured flow curve to the Cross equation (see Supporting Information). This equation well describes the measured data und serves to parameterize the flow curve in terms of its highand low-shear viscosity, the relaxation time and a power-law exponent, which can be implemented in the CFD-simulations. Details of the simulations and the experiment are described in the Supplementary Information. Figure 5 A shows the calculated velocity field ),( yxv for a shear-thinning solution in the contraction/expansion zone for a typical channel geometry and flow rate used in the experiments. Figure 5. Velocity field ),( yxv for cylindrical micelles calculated by CFD simulations. Upper panel (A): Calculated velocity in the wide and narrow channel section. (B): Calculated velocity profiles )(yv x in the pre-tapering zone I (!!!) and post-tapering zone III (!!!). (C): Velocity profiles )(yv x (!!!) and )(yvy (!!!) in the expansion zone indicated in (A). (D) Shear rate )(y γ ! (!!!) and extensional rate )(y ε ! (!!!) in the expansion zone. (E) Map of the ratio γε !! / in the wide and narrow channel section. In the orange regions 14.0/ > γε !! , whereas in the blue regions 14.0/ < γε !! . The resulting color map shows good agreement with the polarized optical micrographs in Fig. 3A, the X-ray intensity map in Fig. 3 C, and the measured velocity map in Fig. 4 F. 113 Fig. 5 B shows the calculated velocity profiles across the channel at position I before entering the contraction zone, and at position III after the expansion zone. The velocity profiles are both non-parabolic, with an almost constant flow velocity in the central part of the channel and a strongly decreasing flow velocity close to the channel walls, a consequence of the shear-thinning, non-Newtonian flow behavior. This is different for Newtonian fluids where the flow velocity has a continuously varying parabolic profile as shown by micro particle image velocimetry and CFD calculations in the Supporting Information. Figure 5 C shows the calculated velocity components x v and y v , and Figure 5 D the corresponding shear rate xy v∇= γ ! and extensional rate yy v−∇= ε ! along the line across the expansion zone indicated in Figure 5 A. We observe, as in the experimental data in Fig. 4, that over the major central part of the cross section the extensional rate ε ! is larger or at least of the same order of magnitude as the shear rate γ ! . Figure 5 E shows the calculated ratio γε !! / over the contraction/expansion zone with a color scale adjusted such as yellow color indicates the zone where 14.0/ > γε !! , whereas blue color indicates the zone with 14.0/ < γε !! . A comparison with Figure 2, 3A, 3C, and 4 A shows that choosing a threshold of 14.0/ = γε !! also in the calculations nearly quantitatively reproduces the yellow zone with perpendicular alignment and the blue zone with parallel alignment in our experiment. CFD simulations show that with decreasing flow rate and diameter of the tapered cross-section the area of the perpendicular oriented zone increases, which is in agreement with our experimental observations. The calculations do not reproduce the difference observed in the two stable preand post-tapering velocity profiles shown in Fig. 4 C. To account for this difference the effect of anisotropic colloids on the shear field has to be modeled in more detail, which in the present calculations has only indirectly been accounted for via the resulting shear-thinning behavior. In conclusion, we show that cylindrical micelles orient either parallel to the flow direction or, after passing through a narrow channel section, perpendicular to the flow direction. Both orientations are stable downstream the channel. Experiments with cylindrical micelles of different type and recent literature indicate that the reorientation in perpendicular direction is generally occurring for anisotropic cylindrical and disk-like colloids. The perpendicular orientation is caused by the velocity field having large perpendicular gradients in the expansion zone after the narrow section. Shear-thinning, a typical property of anisotropic particles, promotes perpendicular extensional and orientation. This phenomenon has important implications when considering the flow orientation of polymers, fibers, proteins or cells through narrow sections such as dies, molds or tapered capillaries. An immediate consequence for the production of fibers is the necessity to subsequently apply extensional forces to re-align polymers of fibrils in flow-direction for optimal fiber mechanical properties. For fibrous proteins reorientation and stable plug-flow are mechanisms for protein or cell coagulation with possible relations to thrombosis. [22] Current experiments indicate that perpendicular flow-orientation can be utilized to orient cylindrical micelles perpendicular to surfaces, which is of relevance for applications involving electrical or thermal transport perpendicular to a surface such as in hybrid solar cells. 114 Materials and Methods Fabrication of microfluidic devices: The microchannel master of the microfluidic device was fabricated using optical lithography.[23] The microchannel network was designed in AutoCAD 2011 and printed on a mask foil with an UV-absorbent ink (Zitzmann GmbH). An inverse black-white image of the device design is shown in Figure 1A. To pump fluids into the device, inlet ports are interfaced with tubing. Their punch location is surrounded with polygons that scatter light, making it easy to see and accurately punch the corresponding PDMS replica that are fabricated using soft lithography, as described in detail in the supplemental part to this publication.[24,25] Stable tubing interfaces are an important prerequisite for long-term in-situ scanning experiments at the synchrotron beam line. Preparation of cylindrical micelle solution: Poly(isoprene-b-polyethylene oxide) (PI110-PEG198, mean Mw 16,200 g mol-1) was prepared by sequential living anionic polymerization, yielding a block copolymer with narrow polydispersity MW/Mn = 1.02, where Mw and Mn are the weightand number-averaged molecular weights. The synthesis and characterization of PI-PEG is described in detail elsewhere.[26] Poly(ethylenebutylene-b-polyethylene oxide) (PEB39-PEO102, mean Mw 7,700 g mol-1, MW/Mn = 1.06) was obtained from EVONIK and lyophilized before use. The dry polymers were dissolved in Millipore-quality water with a resistivity of 18.1 MΩ cm-1 before use. The solutions were homogenized using an UltraTurrax T8 (IKA Werke GmbH) and stored to allow the copolymer to swell in the water for three weeks at room temperature. Before the microfluidic experiments, the solutions are filtered through a PTFE filter with 5 µm pore size. Device operation at the beamline: The experiments were performed at the beamline BW4 and P03 at HASYLAB/DESY. The microfluidic device was connected to high precision syringe pumps (Cetoni GmbH, Nemesys system) and positioned in the X-ray beam. After collecting the necessary background data of an empty microchannel, the syringe pumps are set to typical flow rates of 32.4 µL h-1 corresponding to a mean stream velocity of 360 µm s-1. After 15 min. of equilibration time, measurements along the flow direction are performed with a microfocused X-ray beam at a wavelength of λ = 0.1381 nm. At both beamlines the beam was 20 µm in width and 30 µm in height. X-ray scattering patterns were recorded with step sizes of 70 µm at a distance of 3.128 m behind the microfluidic device using a Pilatus 300K detector (Dectris Ltd.) with a pixel size of 172 µm by 172 µm. The integration time is 240 s. 115 References [1] Yang HH, Allen SR (2000) in Advanced fiber spinning technology, (Ed: Nakajima T), Woodhead Publ., Abington, England, Ch. 6. [2] Cuculo JA, Hotter JF, Zhou Q (2001) in Structure Formation in Polymeric Fibers, (Ed: D. R. Salem), 1st edn., Hanser Gardner Pubns., Ch. 3. [3] Bouxsein NF, Hirst LS, Li Y, Safinya CR, Samah ZA, MacDonald NC, Pynn R (2004) Alignment of filamentous proteins and associated molecules through confinement in microchannels. Appl Phys Lett 85:5775. [4] P. Butler P (1999) Shear induced structures and transformations in complex fluids. Curr Opin Colloid Interface Sci 4:214. [5] Squires TM, Quake SR (2005) Microfluidics: fluid physics at the nanoliter scale. Rev Mod Phys 77:977. [6] Stone HA, Stroock AD, Ajdari A (2004) Engineering flows in small devices: microfluidics toward a lab-on-a-chip. Annu Rev Fluid Mech 36:381. [7] Barrett P, Faucon M, Lopez J, Cristobal G, Destremaut F, Dodge A, Guillot P, Laval P, Masselon C, Salmon JB (2006) X-ray microfocussing combined with microfluidics for on-chip X-ray scattering measurements. Lab Chip 6:494. [8] Martin HP, Brooks NJ, Seddon JM, Terrill NJ, Luckham PF, Kowalski AJ, Cabral JT (2010) Complex fluids under microflow probed by SAXS: rapid microfabrication and analysis. J Phys Conf Ser 247:012050. [9] Li S, Liu N, Chan-Park MB, Yan Y, Zhang Q (2007) Aligned single-walled carbon nanotube patterns with nanoscale width, micron-scale length and controllable pitch. Nanotechnol 18:455302. [10] Hesse HC, Beck R, Ding C, Jones JB, Deek J, MacDonald NC, Li Y, Safinya CR (2008) Direct imaging of aligned neurofilament networks assembled using in situ dialysis in microchannels. Langmuir 24:8397. [11] Rammensee S, Slotta U, Scheibel T, Bausch AR (2008) Assembly mechanism of recombinant spider silk proteins. Proc. Natl. Acad. Sci. U.S.A. 105:6590. [12] Sun B, Sirringhaus H (2006) Surface tension and fluid flow driven self-assembly of ordered ZnO nanorod films for high performance field effect transistors. J Am Chem Soc 128:16231. [13] Dimalanta ET, Lim A, Runnheim R, Lamers C, Churas C, Forrest DK, De Pablo JJ, Graham MD, Coppersmith SN, Goldstein S, Schwartz DC (2004) A microfluidic system for large DNA molecule arrays. Anal Chem 76:5293. [14] Cates ME, Candau SJ (1990) Statics and dynamics of worm-like surfactant micelles. J Phys Cond Mat 2:6869. [15] Förster S, Konrad M, Lindner P (2005) Shear thinning and orientational ordering of wormlike micelles. Phys Rev Lett 94:017803. [16] Richtering W (2001) Rheology and shear induced structures in surfactant solutions. Curr Opin Colloid Interface Sci 6:446. 116 [17] Waton G, Michels B, Steyer A, Schosseler F (2004) Shear-induced demixing and shearbanding instabilities in dilute triblock copolymer solutions. Macromolecules 37:2313. [18] Gao C, Kulkarni SD, Morris JF, Gilchrist JF (2010) Direct investigation of anisotropic suspension structure in pressure-driven flow. Phys Rev E 81:041403. [19] Cromer M, Cook LP, McKinley GH (2011) Pressure-driven flow of wormlike micellar solutions in rectilinear microchannels. J Non-Newtonian Fluid Mech 166:180. [20] Singh AP, Rey AD (1995) Computer simulation of dynamics and morphology of discotic mesophases in extensional flows. Liquid Crystals 18:219. [21] Oliveira MSN, Alves MA, Pinho FT, McKinley GH (2007) Newtonian fluid flow through microfabricated hyperbolic contractions. Exp Fluids 43:437. [22] Liu Q, Mirc D, Fu BM (2008) Mechanical mechanisms of thrombosis in intact bent microvessels of rat mesentery. J. Biomechanics 41:2726. [23] Nguyen NT, Wereley S (2002), Fundamentals and Applications of Microfluidics, 1st edn., Artech House, Ch. 3. [24] Xia Y, Whitesides GM (1998) Soft lithography. Annu Rev Mater Sci 28:153. [25] Quake SR, Scherer A (2000) From microto nanofabrication with soft materials. Science 290:1536. [26] Förster S, Krämer E (1999) Synthesis of PB-PEO and PI-PEO block copolymers with alkyl lithium initiators and the phosphazene base t-BuP4. Macromolecules 32:2783. 117 118 Supporting Information Fabrication of Kapton-PDMS-Kapton microfluidic devices (Figure S1): A negative photoresist (SU-8 50, Microchem Co.) is spin-coated onto a silicon wafer. A mask aligner (Süss Mikro Tec) is used to impart the microchannel structure into the photoresist. We optimize the master device fabrication to obtain microchannels with a very uniform height of 100 µm. Although PDMS is widely applied to replicate the microchannel master device using soft lithography, PDMS scatters and absorbs X-rays. To fabricate X-ray compatible microfluidic devices involving PDMS, we modified the conventional fabrication procedure, based on the work of Evans and Dootz.[S1, S2] After pouring PDMS pre-polymer (Sylgard 184, Dow Corning) on the master device, excess pre-polymer is removed from the master device with a razor blade. The remaining PDMS is cured and a small piece of selfadhesive polyimide tape (DuPont™ Kapton®) is used to cover the area of interest of the microchannel network including the curved and tapered microchannel sections. A second layer of PDMS is cured onto the previous layers. The PDMS replica is removed from the master device and inlet ports are punched into the polymer using a biopsy punch needle (Harris Uni-Core™ 0.75 mm). The bottom of the device is sealed with Kapton tape and a window is cut into the top PDMS layer. Thus, the microchannels in the area of interest are solely sealed with X-ray transparent Kapton tape. Figure S1. Fabrication of microfluidic devices with X-ray analysis capability based on PDMS. (A) Master device fabrication using photolithography. (B) Fabrication of X-ray transparent microfluidic Kapton-PDMS-Kapton sandwich devices. (B1) PDMS is poured on a master device, (B2) and excess PDMS is cut-off the microchannel structure. (B3) Selfadhesive Kapton tape is used to seal the area of interest, (B4) and a second layer of PDMS is grafted on top. (B5) The PDMS replica is peeled off the master device, and inlet ports for fluids are added. (B6) The bottom is sealed with Kapton tape, (B7, B8) before a window is cut into the top PDMS layer with similar dimensions as the Kapton window in B3. 119 Computational Fluid Dynamics Simulations (CFD): The fluid dynamics calculations are based on the Navier-Stokes equations assuming an incompressible fluid, i.e. ρ = const. [S3], ρ∇∙!=0 ρ ∂! ∂!+ρ!∙∇!=∇∙−!!+!∇!+∇!!+! with the density of the fluid ρ, the pressure !, the identity matrix !, the dynamic viscosity of the fluid !, the velocity field u and the volume force F. Solutions of anisotropic particles exhibit pronounced shear-thinning. To model the resulting flow profile, we used the software package COMSOL Multiphysics v4.2a, which allows one to import CAD-designed microchannel geometries and takes into account Non-Newtonian flow behavior in computational fluid dynamics simulations. [S3] To integrate non-linear flow behavior in the CFD-model, experimental data are employed obtained by rheometry using a Bohlin Gemini 200 which was used to measure the shear-rate dependence of the viscosity. The measured flow curve can be well described by the Cross equation: [S4, S5] η=η!+η!−η! 1+(τ!γ)! where the viscosity!η is described by the zero-shear viscosity η!, the high-shear viscosity η!, the internal relaxation time τ! and the power-law exponent n characterizing the shear thinning between η! and η!. This equations well reproduces the experimental data as shown in Fig. S2. The values of the parameters obtained by fitting the equation to the measured flow-curve are then used in the CFD-calculations. Figure S2. Experimental rheological data: viscosity (η) as a function of the shear rate (γ) is described by the Cross equation. The simulations yield the velocity field ),( yxv from which the shearand extensional rates can be calculated. A mean shear rate [S3] 120 γ= 4u! !+2u!+v! !+4v! ! 2 with the corresponding xor y-components of u (velocity in x-direction) and v (velocity in ydirection) is used to map the local velocity on the microchannel. The model is solved for 1412784 finite elements and 1072212 degrees of freedom using a multifrontal massively parallel solver (MUMPS). The average element quality of the mesh is 0.9889 on a scale from 0 to 1, where 1 is the highest quality; the minimal element quality is 0.6923. Using a Windows 7 x64 machine with two quad-core Intel® Xeon® E5440processors operating at 2.83 GHz and an internal memory of 32 GB RAM. All relevant parameters, which are used in the simulations, are summarized in Table S1. Table S1. Material properties used in the simulation model. Figure 3 & S3 NonNewtonian Newtonian Flow rate vflow 32.4 µL h-1 32.4 µL h-1 Flow speed v 18.52 mm s-1 18.52 mm s-1 zero-shear viscosity η! 19522 Pa s - high-shear viscosity η! 0.1 Pa s - internal relaxation time τ! 388.5 s - power-law exponent n 0.99 - Viscosity η [S6] - 1.002∙10-3 kg m-1 s-1 Density ρ [S6] 998.2 kg m-3 998.2 kg m-3 Temperature T 293.15 K 293.15 K Calculated velocity field for water The calculated velocity field for the wormlike micelles has been discussed in the main part of the publication. Here we present the same calculations for water as a Newtonian fluid. This provides insight to what extent the shear-thinning behavior causes the observed reorientation behavior. We calculated the velocity field for water under the same experimental conditions as for wormlike micelles. The results are shown in Fig. S3, which is organized similar to Figure 5 in the main section of the manuscript for comparison. 121 Introduction Microfluidic SAXS experiment. Microfluidics enables the precise control of liquids on the nanoliter scale. 1 These very well defined flow conditions make this technology predestined for fundamental investigations at microfocused X-ray sources. The basic idea of this experimental setup is to utilize the very well defined continuous flow conditions of the microchannel to scan it with a X-ray microbeam and record the small-angle X-ray scattering (SAXS) pattern of each measured position. This mapped data then allows to get a fast and detailed overview over the flow experiment within the microchannel. The combination of microbeam X-ray scattering and microfluidics is currently being developed into a powerful experimental methodology suitable for the investigation of nanostructures, particle alignment and the in situ study of kinetics by creating X-ray compatible microflow chips and microfluidic liquid jet devices. 2-10 The sample is pumped through the microchannels of a X-ray compatible device and the microfocused X-ray beam passes the flowing sample; in this case anisotropic wormlike micelles. The scattered X-rays, which contain the structural information, are recorded using a 2D digital detector (Piltatus 1M, Dectris). The typical experimental setup at the microfocus beamline P03 (PETRA III, DESY) is shown in the supplemental information in Fig.S1. 11 Recently, microfluidic SAXS scanning experiments at the P03 and BW4 beamlines (DESY, Hamburg) revealed the striking effect, that after passing a narrow section, wormlike particles are rotated perpendicular to the flow direction, keeping this orientation over the remaining length of the channel. 2 The flow-alignment of cylindrical, wormlike or fibrous structures is central to many processing steps such as in the production of fibers, during injection molding or the flow of cells and proteins through thin capillaries. 12-16 In this paper, this perpendicular orientation will be investigated in more detail to understand the influence of experimental parameters on the orientation effect. For this task, we create Xray compatible microfluidic devices which are made of NOA81 (Nordland Optical Adhesive 81) for the SAXS analysis. 9,17-19 A redesigned NOA81-fabrication routine now allows the production of very thin devices for high X-ray transmission while also maintaining the material’s very good solvent compatibility. The resulting microfluidic device is shown in Fig.1B while its detailed fabrication routine is explained in the experimental section. This routine is soft lithography-based because it uses inverted PDMS-microchannels as a molding template. 20,21 Therefore, the high microfluidic design flexibility is maintained which is important for rapid prototyping. 1 With this given design control from soft lithography, we vary the geometric parameters of the microchannel like channel width (250 to 500 µm) tapering ratios (1:10 to 1:2.5; w. r. t. channel width) and -lengths (9:1 to 2:1; w. r. t. channel width), as illustrated in Fig.1A. Further, we vary the sample concentration (30 to 0% w/w) and flow speeds (100 to 2000 µl h-1). 128 Figure 1 (A) The CAD-based design and use of high resolution photo masks (128 kdpi, JD Photo, UK) enable precise control over the microchannel geometries. In this case the shear field is controlled by smooth taperings with varying ratios (1:10 to 1:2.5) and lengths (9:1 to 2:1) with respect to the original channel widths (250 to 500 µm) at a constant channel height of 100 µm. Further, the varied relevant experimental parameters include flow speed and sample concentration. (B) The resulting NOA81-device and a typical electron micrograph of the SU-8 replication template (tapering ratio 3:1, channel width 150 µm). Results and Discussion SAXS pattern analysis. The microfluidic SAXS scanning setup enables fast sample screening of the varying experimental conditions. Each obtained SAXS pattern of a given position contains various structural information about the cylindrical micelles, like i.e. particle size, unit cell dimensions and orientational distribution; an example SAXS pattern is shown in Fig.2. The vertically-positioned crescents in this anisotropic SAXS pattern indicate a parallel orientation of wormlike micelles in respect to the (horizontal) flow direction. Additionally, the azimuthal peak width contains information about the micelles’ orientation distribution. After the radial averaging of this SAXS pattern, the fits and projections can be calculated using the analysis software Scatter which yields further structural information about the sample. 22,23 This sample contains hexagonally closest packed micelles (PEB39-b-PEO102, 30% w/w in pure water) which are oriented parallel to the (horizontal) flow direction. Their radius is 10.0 nm with varying lengths up to the micron range and a unit cell size of 40.9 nm (1 nm displacement, Laguerre distribution with =14). 23 This software also allows the calculation of the micelle’s 2D scattering patterns based on the 3D model of hexagonally closest-packed cylinders as shown in the image inlay in Fig.2. 23 The fitting parameters are listed in the following Tab.1. 129 Table 1 List of fitting parameters for the SAXS-pattern calculation using the analysis software Scatter. 22 Parameter Value Model Hexagonally packed cylinders (P6/mm) Micelle radius , nm 10.0 Relative standard distribution , nm 0.15 Cylinder length , nm 73.2 Relative standard distribution , nm 0.1 Unit cell dimensions , nm 40.9 Displacement , nm 1.0 Radial domain size , nm 213 Radial domain size , nm 65 Distribution function type Laguerre Distribution function parameter 14 Figure 2 The obtained SAXS patterns contain various structural information of the cylindrical micelles, like i.e. particle size (r=10 nm), unit cell dimensions (40.9 nm) and orientational distribution (Laguerre distribution with =14). The vertical crescents in this anisotropic pattern indicate parallel orientation in respect to the flow horizontal direction. After the radial averaging of this SAXS pattern, the fits and projections are calculated using the analysis software Scatter. 22,23 Further, it is also possible to calculate the 2D scattering pattern, as shown in the bottom right corner, based on the corresponding 3D model of these hexagonally closest-packed cylinders. 130 Small-angle X-ray scattering is a complementary method that can be correlated to the results from other techniques such as micro particle image velocimetry (µPIV), polarization microscopy and CFD-simulations which will be discussed over the course of this paper. Generation of color-coded pixel maps from SAXS data. A fast overview over the obtained SAXS patterns of a single experiment is gained by the real-time generation of pixel maps based on the SAXS scanning locations. The SAXS pattern’s structural information is then used to color-code the individual pixels based on the averaged intensity within micelleorientation dependent regions of interest (ROI). The resulting pixel maps of different ROIs are shown in Fig.3B. The first pixel map shows the averaged intensity for all micelles orientations and gives an impression about the microchannel’s shape. The ROI for the micelle orientation parallel to the flow is shown below and the color-coded pixel maps are in good agreement with the images obtained from polarization microscopy (Fig.3A). In the SAXSbased pixelmap we observe that the parallel orientation is strongest in the beginning of the tapering, where both extensional and shear forces are applied in flow direction. Parallel micelle orientation is also found close to the channel walls where the wall shear is most dominant due to the flow of the fluid. The third ROI-based pixel map represents the areas where the micelles are rotated perpendicular to the flow due to the dominance of extensional forces which are directed vertically to the main flow direction. 2 This perpendicular orientation can be also observed well as an orange-colored area in polarization microscopy to which the SAXS scanning position can be correlated as indicated in Fig.3A. Lastly, the plus 45° and minus 45° micelle orientations are shown below and are either the result of parallel micelle alignment to the tilted microchannel walls of the tapering or due to the transition from parallel to perpendicular orientation. The effect of the tapering ratio on the perpendicular orientation. The velocity flow field in the microchannel and, hence, the shear and extensional forces are varied by changing the tapering ratios of the microchannel’s narrow section between 10:1 and 2.5:1 with respect to the microchannel width. The flow of micelles through narrow sections under the same conditions is studied using polarization microscopy, as shown in Fig.4A. We find that the perpendicular orientation, which is indicated by the orange region, has the broadest width and is most pronounced at the largest tapering aspect ratio of 10:1. Both the color-intensity and the width of the orange stream decrease with wider microchannel taperings. The same set of experiments is also studied using microfluidic SAXS of which the resulting color-coded pixel maps are shown in Fig.4B,C. These pixel maps share the same color-scale to illustrate the relative intensity of re-orientation. Similar to the observed trends using polarization microscopy, we find that the re-orientation increases with larger tapering ratios and that less perpendicularly orientated micelles are observed with lower the tapering aspect ratios. 131 Figure 3 (A) Polarization microscopic (PM) images of the setup using crossed polarizers with a quarter wave plate give an impression about the flowing micelle’s orientation: the blue color in the PM corresponds to a parallel orientation, while orange areas correspond to a perpendicular orientation of micelles in respect to the flow direction. The channel width is 500 µm with tapering narrowing down to 50 µm (10:1) over a length of 3500 µm (7:1). The micelle concentration is 30% w/w at a flow speed of 200 µl h-1. (B) This orientation is confirmed by the SAXS experiments at the P03/MiNaXS beamline which can easily be correlated with the PM because the conditions, such as flow speed, channel width, tapering ratio and length, etc. are highly reproducible in the microfluidic device. The flowing complex fluid is scanned at a given set of positions using the microfocused X-ray beam. Each obtained pixel corresponds to a scattering pattern and therefore contains the full structural information. (C) Color-coding these pixels based on the averaged intensity at a given region of interest (ROI) gives a detailed overview over the structural evolution. The basic types of orientational distribution in respect to the flow direction are isotropic, parallel, perpendicular, tilted plus 45° or minus 45°. The color bars of these pixel maps is adjusted individually to express the orientation regions more clearly. 132 Figure 4 Comparison of the influence of the tapering ratios on the micelle orientation behavior. (A) Polarization microscopic images under the same conditions as the following SAXS pixel maps (micelle concentration 30% w/w, 200 µl h-1) which show the regions of parallel (B) and perpendicular (C) orientation at different tapering ratios (10:1, 5:1, 2.5:1) with respect to the channel width of 500 µm. The color-scale is equal for all pixel maps to illustrate the relative intensity of re-orientation. 133 Analysis of the influence of the tapering ratio on the orientation distribution. The orientation distribution change along the microchannel has to be considered for the detailed analysis of the influence of tapering ratios on the intensity of perpendicular micelle orientation. This orientation information can be extracted from a series of SAXS patterns from the tapering ratio variation (Fig.4B,C). These vertical scans with microfocused X-ray beams, along the microchannel of each tapering aspect ratio, are compared in the Fig.5 providing a first qualitative analysis. Here, the first and last rows of SAXS patterns show the parallel micelle orientation close to the walls which is indicated by the vertical position of the peaks. The second and fourth rows show transition regions where the wormlike micelles change from parallel to perpendicular orientation and vice versa. This realignment appears as ‘fish’- shaped or crescent-like peaks in the SAXS patterns that originate from an asymmetrical orientational distribution along the azimuth. The scanning resolution of this transition is determined by the size of the X-ray microbeam which is 20x30 µm2 in our experiments. Therefore, the observed ‘fish’-shape could originate from a much sharper transition region that is smeared by the overlaying X-ray beam of two scanning positions at parallel and perpendicular micelle orientation. However, the color transition in polarization microscopic images suggests a smooth orientation transition. A smaller microfocused X-ray beam could be used to resolve this region in more detail and study this transitional layer in more detail. The central part of the microchannel scan is shown in the third row. We find that the micelle orientation distribution strongly depends on the tapering aspect ratio. In case of the 10:1 ratio, almost all micelles are aligned perpendicularly to the flow which is indicated by the peak in horizontal position. However, the horizontal peak intensity and, hence, perpendicular micelle orientation decreases with lower tapering ratios. In case of the 2.5:1 ratio, the SAXS pattern shows a more isotropic orientation characteristic with a preferred parallel orientation. 134 Figure 5 The detailed SAXS analysis is based on a vertical line scan across the microchannel at the tapering exit, as indicated by the long black arrow in the polarization microscopic image at the top. Below, the SAXS-patterns of womlike micelles in parallel and perpendicular orientations as well as the transitional stage are shown. Here, the ratio decreases with each column from left (10:1) to right (2.5:1). The red box marks the SAXS-patterns with are analyzed in more detail in Fig.6. 135 Next, this trend can be analyzed quantitatively by choosing a cake-like ROI of the SAXS patterns and calculating the azimuthally averaged intensity as indicated in Fig.6A. This averaging results in angle-dependent intensity curves which contain the peak intensities, - widths and -rotation angles of the parallel and perpendicular micelle orientations at the specific scan position in the microchannel. These curves at the central scan positions (see mark in Fig.6A) for all three taperings are combined in a single graph (Fig.6C) which gives an overview over the peak intensity changes. This graph reveals that the overall measured intensity stays constant and that only the orientation distribution changes with the tapering ratio. The peak of the parallel orientation at this central scan position (Fig.6C) decreases with a raising tapering ratio while the perpendicular orientation peak raises accordingly. For a better general overview, this extraction process of angle-dependent intensity curves (see Fig.6C) is performed for the complete series of SAXS patterns. The results of the three vertical scans are then combined in individual 2D plots which are shown in Fig.6B where the color represents the extracted peak intensities. These 2D color plots now allow to differentiate between two regions of orientation intensity which are based on the parallel or perpendicular micelle orientation. Similar to the described orientation trends from above, we find that the relative peak intensities change with the tapering ratio where the most pronounced perpendicular orientation is found in the central part of the channel with a tapering ratio of 10:1. In case of the 5:1 ratio, the non-parallel orientation is weaker, but still clearly visible and has about the same intensity level as the parallel orientation. This trend continues with the 2.5:1 tapering ratio where the non-parallel orientation merely reveals itself as an isotropic orientation distribution with a preferred parallel orientation. These micelle re-orientation shifts can also be quantified by the intensity ratio which we define as This ratio describes the relation between the summed peak intensities of micelles with perpendicular orientation ( from 61 to 117°) with the combined peak intensity sums of parallel and perpendicular micelles ( with from −37 to 40°). This measure allows to correlate the relative peak intensities from the SAXS experiments with the orange/bluecontrast observed in polarization microscopy. We find that a value of = 0.1 corresponds to the blue-colored regions with a preferred parallel micelle orientation, i.e. in proximity to the channel walls. The weakly orange-colored region at low tapering ratios (2.5:1, see Fig.4A) can be observed at = 0.25 which corresponds to an isotropic micelle distribution. Regions with a more pronounced orange color which can be found at higher tapering ratios (5:1 & 10:1, see Fig.4A) indicate that the orientation distribution is dominated by a perpendicular micelle alignment with -values > 0.35. The observed trends can also be verified by microparticle image velocimetry (µPIV) in the following section of this paper. 136 Figure 6 Extended overview over the micelle orientation distribution analysis from the SAXS pixel maps in Fig.4B,C. (A) By azimuthally averaging a cake-like slice from the SAXS image, it is possible to study and quantify the micelle’s orientation distribution. (C) The resulting angle-dependent peak intensity curves reveal the influence of the microchannel’s tapering ratio on the orientation distribution. (B) This effect can be studied in more detail by extracting the azimuthally averaged curves for all positions across the microchannel directly after the tapering and for different tapering ratios. The results are combined in 2D color maps of the angle-dependent peak intensity which give an overview over the micelle orientation changes across the microchannel that are quantified by the intensity ratio . 137 of wormlike micelles, numerical CFD-simulations that are based on experimental rheometric data are in good agreement with the results from small-angle X-ray scattering, polarization microscopy and microparticle velocimetry. 2 Since the current system is based on the same block copolymer (PEB39-PEO102, Evonik), we assume the CFD-simulations to be also valid for predicting the general trends of the fluid’s flow when the rheometric sample parameters are varied theoretically. The viscosity change of the micelle solution is described by the ColeCole- (or Cross-) equation which is given in the experimental section of this paper. From this formula, the zero shear viscosity as well as the internal relaxation time are varied in a series of CFD-simulations. For this, three different values for each parameter are chosen while all other parameters are kept constant, as summarized in Tab.1. An overview over the rheological curves that result from these parameter combinations are shown in Fig.10A. The different colors correspond to the varying zero shear viscosities while the symbol shapes change with varying internal relaxation times. These variations are to represent the rheological changes of a sample with different concentrations or types of anisotropic particles to study the general trends. Although similar rheological property-changes could also be achieved and studied experimentally, this CFD-simulation-based approach is easier to control and less time-consuming. These different rheological curves for defined parameter variations are then used in the CFD-model to describe the fluid’s rheometric properties. 2 Next, the resulting velocity flow fields from the simulation (not shown) are used to calculate the shear and extensional rates. As shown by the µPIV-experiments and polarization microscopy, the regions of perpendicular micelle orientation can be found where the extensional rate is comparable to the shear rate .2 Therefore, this ratio ( ) is used to color-code the simulated fluid flow inside the microchannel taperings. The resulting 2D color maps are arranged according to the parameter changes and shown in Fig.10B. An orange color in these 2D plots corresponds to a region where the micelle orientation is typically found to be perpendicular exceeding a value of 0.14. Accordingly, the blue area represents the regions of preferred parallel orientation with a ratio below 0.14. The simulated results clearly show that the sample’s zero shear viscosity has a strong influence on the width of the orange region after the tapering. The higher the initial viscosity of the sample, the more pronounced is the plug-like flow of the sample and the greater is the shielding of the central micelles against the wall-induced shear. The internal relaxation time controls the capability of the micelles in the center to change their orientation while flowing through the tapering. Since this is a continuously flowing sample, the geometric shear field of the tapering can be seen as a re-orientation impulse. The shorter the internal relaxation time, the more are the anisotropic particles capable of reacting to this impulse and change their orientation according to the strongest orienting force. Further, the shorter the internal relaxation time, the less are the anisotropic particles affected 144 by wall shear effects because their viscosity starts to decrease at higher shear rates. Consequently, rheological properties that are typical for non-Newtonian samples, like i.e. shear thinning and plug-like flow, strongly increases the tendency to find perpendicularly oriented micelles in microchannel after confined geometries. This observation is in agreement with the above-discussed results from small-angle X-ray scattering, polarization microscopy and microparticle image velocimetry. Figure 10 CFD-simulations based on theoretical variation of rheological sample parameters. (A) Curves are calculated based on the Cole-Cole-equation (see experimental section) that describe the fluid’s viscosity change under the influence of shear. (B) CFD-simulation results that show color plots of the ratio ( ) between the extensional rate and the shear rate . The orange color represents the regions where the micelle orientation is typically found to be perpendicular (SAXS, µPIV, polarization microscopy) because the mentioned ratio exceeds a value of 0.14. Accordingly, the blue area represents the regions of preferred parallel orientation with a ratio below 0.14. 145 Experimental Preparation of Cylindrical Micelle Solution. The formation of wormlike micelles is based on the self-assembly of the amphiphile block copolymer PEB39-PEO102 (lyophilized before use, mean Mw = 7,700 g mol-1, Mw/Mn = 1.06, Evonik). 25-27 The polymer is dissolved in pure water (Millipore-quality with a resistivity of 18.1 MΩ cm-1), homogenized using an UltraTurrax T8 (IKA Werke GmbH) and stored at room temperature for three weeks to allow the copolymer to swell. Prior to the use in microfluidic devices the solution was filtered through a polytetrafluorethylene filter with 5 µm pore size. Preparation of monodispere polystyrene (PS) tracer spheres for µPIV. The µPIVexperiments require to add small spherical tracer particles to the samples (4% w/w). The dispersion polymerization synthesis and characterization of these monodisperse polystyrene spheres (1.65 µm radius, PDI 1.01) is described in detail elsewhere. 28 To avoid clogging of the PS-spheres to the PDMS-microchannel walls, the particle hydrophilicity is increased by a treatment with polyacrylic acid (PAA, 1 g/l in 100mM NaCl) for 15 min under continuous shaking. 28 The excess of PAA was removed by re-suspending the particles three times with pure water (Millipore quality). Microfluidic (SAXS) device fabrication. The microchannels are designed using AutoCAD 2013 (Autodesk) and printed on a high-resolution emulsion film mask (JD Phtoto). An example of the channel geometry is shown in Fig.1A. The subsequent fabrication of master-templates for replication and PDMS-devices (sealed with glass-slides) for polarization microscopy and the µPIV-experiments are based on standard soft lithography which is described elsewhere. 20,21 This process involves spin coating the photo resist SU-8 50 (Microchem) onto polished silicon wafers (Si-Mat) for controlled channel heights (100 µm) and is illustrated in Fig.S41. The fabrication for the SAXS-compatible NOA81-devices includes an additional microstructuring step using a mask aligner (Süss Mikro Tec) for the generation of a two-layer microstructure as pointed out in Fig.S42. The first SU-8-layer enables to control the window thickness by adding an additional gap layer next to the microchannels for the experiments. The second layer contains the microchannels for the liquid flow (and the second layer of the gap structure). The final SU-8-master template is then replicated using standard PDMS-based soft lithography using Sylgard 184 (Dow Corning) as illustrated in Fig.S43. 20,21 Next, the PDMS-replica is used for microstructuring the UV-curable adhesive NOA81 (Nordland). 17,18 This method is based on a NOA-PDMS-hybrid routine that is described elsewhere, 9 but it is modified to allow an improved solvent compatibility. 18 Instead of drilling holes through the finished device and sealing it with PDMS, glue and tapes, a different approach was used. 146 First, holes are punched into the replicated PDMS microchannel-template using a biopsy needle (Harris Uni-Core). Next, the tubings which only serve as a template, are inserted into these holes before the liquid NOA81 is poured over the PDMS-template. A flat PDMS-block with the intended scan-window size is laid over a liquid NOA81-drop which spreads itself. The two-layered gap structure next to the microchannels control the device material height in this step, while the liquid NOA81-drop next to the template-tubings should have a height of ca. 2-3 mm. After UV-curing, the PDMS-block and the template tubings are removed before the open microchannels are sealed with a thin NOA81-film which was prepared in a similar fashion: no template-tubing, gap structures and flat (or alternatively 3D-microchannelstructured). The sealing of the device with a second NOA81-film by UV-curing is possible because the gas-permeability at the PDMS-NOA81-interface inhibits the complete curing. 9 This inhibition leaves a thin curable layer for the device sealing at the NOA81-surface. The resulting NOA81-based microfluidic device is very thin in the X-ray microbeam scanning area (ca. 250 µm) with a homogenous channel height and the microfluidic device has a very low X-ray background signal. Further, the increased material thickness at the inlets, which were formed by the template-tubings, allow to attach the tubings directly to the device without using an additional tubing-interface PDMS-layer as described earlier. 9 Therefore the full solvent compatibility of NOA81 (i.e. water, unpolar organic solvents, etc.) is maintained. 18 For compatibility with higher pressures and to guarantee reliable, stable tubing interfaces (i.e. for precious synchrotron beamtimes), it is suggested to glue the tubings to the device using two-component epoxy glues (Loctite, Henkel). Microscopic setups. The polarization microscopic experiments are recorded using the inverted microscope Axiovert S100 (Zeiss). The orange and blue color representation originates from using crossed polarizers and a quarter wave plate. The images are recorded using a Nikon D7000 camera. The microparticle image velocimetry setup involves an IX71 inverted microscope (Olympus) with a 100 W tungsten light that was focused onto the sensor of a Phantom v9.1 (Vision Research) high speed camera (2 µs minimum exposure, maximum rate of 2 Gpx s-1). The recorded video frames are analyzed using the open source software package JPIV which generates velocity fields by correlating the captured frames. 24 COMSOL simulation The CFD-simulations are performed in COMSOL Multiphysics 4.2a using a previously described model for non-Newtonian fluid in tapered microchannels of which the theoretical background is described elsewhere. 2 The underlying microchannel geometry is adjusted to the existing devices (250 µm channel width, 5:1 tapering-, 2:1 length ratio, 100 µm height) at a flow rate of 2000 µl h-1. The screening mainly involved a variation of the rhological properties of the sample which are described by the Cole-Cole- (or Cross-) equation: 29-31 147 with the zero shear viscosity , the high-shear viscosity , the internal relaxation time and the power law exponent characterizing the shear thinning between and . The modeling parameters are listed in the following Tab.1. Table 1 List of modeling parameters for the CFD-simulations including the rheological parameter variations. Parameter Value Flow rate , µl h-1 2000 Flow speed , m s-1 0.0222 Zero shear viscosity , Pa s 19522; 195.22; 1.9522 High shear viscosity , Pa s 0.1 Internal relaxation time , s 388.5; 38.85; 3.885 Power law exponent 0.99 Density , kg m-3 998.2 Temperature , K 293.15 Channel width , µm 250 Channel height , µm 100 Tapering ratio 10:1 Tapering length ratio 2:1 Device operation at the PETRAIII synchrotron (P03 beamline, HASYLAB/DESY). The microfluidic SAXS scanning experiments are performed at the Microand Nanofocus X-ray Scattering beamline (MiNaXS) in Hamburg, Germany. The fluid flow is controlled using high-precision syringe pumps (Nemesys, Cetoni GmbH). After the collection of the relevant background data from the microchannel and an extended flow equilibration time at a flow rate of 200 µl h-1, the SAXS-mapping scans are commenced. The microfocused X-ray beam for these scans has a wavelength of λ = 0.095 nm with a width of 20 µm and a height of 30 µm. The SAXS patterns are recorded at a distance of 4840 mm with integration time of 0.5 s using a digital detector (Pilatus by Dectris) with a pixel size of 172 µm by 172 µm with. 148 Conclusions In conclusion, we have performed a fundamental study of the re-orientation of anisotropic particles in confined geometries for a wide range of experimental conditions. We also demonstrated in great detail how multiple analysis techniques can be used and combined in a complementary way taking advantage of the highly reproducible design control provided by microfluidics. Next to the here-demonstrated correlation of fluid dynamics to structural information, the combination of complementary methods is also applicable and extendable to other methods and more experiments, such as flow field analysis for mixing experiments, following synthetic reaction kinetics and mapping nucleation and growth processes. 5,6,9,32-34 Using complementary analysis tools, including small-angle X-ray scattering, microparticle image velocimetry, polarization microscopy and CFD-simulations, we could identify design rules for predicting and controlling the perpendicular re-orientation of anisotropic particles after narrow sections. A relative ranking according to the effectiveness for perpendicular alignment of anisotropic particles is derived as: tapering ratio (y-orienting extensional rate) > concentration (plug-flow and non-Newtonian behavior) > length ratio (wall shear surface) > flow speed (wall shear intensity). The regions for perpendicular orientation can be controlled with further precision by tuning the rheological properties of the sample, i. e. like the zero shear viscosity and the internal relaxation time. Consequently, the perpendicular orientation of anisotropic particles can be maximized by increasing the tapering ratio, shortening the tapering, minimizing the flow speed and increasing the sample’s plug-flow by increasing its high zero shear viscosity and shortening its internal relaxation time or vice versa for maintaining and increasing parallel alignment. In summary, the here presented study provides conditions for controlling the orientation of anisotropic particles. A strong parallel orientation is of great importance for the spinning of fibers because its typically leads to increased crystallinity in the fiber and, hence, better mechanical performance. 12,35 However, the perpendicular orientation of particles can also be of great interest for applications that involve electrical or thermal transport perpendicular to a surface, such as in isolating materials or hybrid solar cells. 36-38 149 References 1. D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984. 2. M. Trebbin, D. Steinhauser, J. Perlich, A. Buffet, S. V. Roth, W. Zimmermann, J. Thiele, and S. Förster, PNAS, 2013, 110, 6706–6711. 3. A. Otten, S. Köster, B. Struth, A. Snigirev, and T. Pfohl, J Synchrotron Rad, 2005, 12, 745–750. 4. R. Barrett, M. Faucon, J. Lopez, G. Cristobal, F. Destremaut, A. Dodge, P. Guillot, P. Laval, C. Masselon, and J.-B. Salmon, Lab Chip, 2006, 6, 494–499. 5. T. Pfohl, A. Otten, S. Köster, R. Dootz, B. Struth, and H. M. Evans, Biomacromolecules, 2007, 8, 2167–2172. 6. H. M. Evans, R. Dootz, S. Koester, B. Struth, and T. Pfohl, Bull Pol Acad Sci-Te, 2007, 55, 217–227. 7. R. Dootz, H. Evans, S. Köster, and T. Pfohl, Small, 2007, 3, 96–100. 8. K. N. Toft, B. Vestergaard, S. S. Nielsen, D. Snakenborg, M. G. Jeppesen, J. K. Jacobsen, L. Arleth, and J. P. Kutter, Anal. Chem., 2008, 80, 3648–3654. 9. M. E. Brennich, J.-F. Nolting, C. Dammann, B. Nöding, S. Bauch, H. Herrmann, T. Pfohl, and S. Köster, Lab Chip, 2011, 11, 708. 10. S. M. Taheri, S. Fischer, M. Trebbin, S. With, J. H. Schröder, J. Perlich, S. V. Roth, and S. Förster, Soft Matter, 2012, 8, 12124. 11. A. Buffet, A. Rothkirch, R. Döhrmann, V. Körstgens, M. M. Abul Kashem, J. Perlich, G. Herzog, M. Schwartzkopf, R. Gehrke, P. Müller-Buschbaum, and S. V. Roth, J. Synchrotron Rad (2012). 19, 647-653 [doi:10.1107/S0909049512016895], 2012, 1–7. 12. T. Vad, J. Wulfhorst, T.-T. Pan, W. Steinmann, S. Dabringhaus, M. Beckers, G. Seide, T. Gries, W. F. C. Sager, M. Heidelmann, and T. E. Weirich, Macromolecules, 2013, 130705124618000. 13. H. H. Yang and S. R. Allen, Advanced Fiber Spinning Technology, Woodhead Publications, Abington, 2000. 14. J. A. Cuculo, J. F. Hotter, Q. Zhou, and Q, Structure Formation in Polymeric Fibers, Hanser Gardner Publications, 1st edn. 2004. 15. N. F. Bouxsein, L. S. Hirst, Y. Li, and C. R. Safinya, Applied Physics …, 2004. 16. K. J. Humphry, P. M. Kulkarni, D. A. Weitz, J. F. Morris, and H. A. Stone, Phys. Fluids, 2010, 22, 081703. 17. C. F. Carlborg, T. Haraldsson, K. Öberg, M. Malkoch, and W. van der Wijngaart, Lab Chip, 2010, 11, 3136. 18. P. Wägli, A. Homsy, and N. F. de Rooij, Sensors and Actuators B: Chemical, 2011, 156, 994–1001. 19. E. Sollier, C. Murray, P. Maoddi, and D. Di Carlo, Lab Chip, 2011, 11, 3752. 20. G. M. Whitesides and Y. Xia, Angew. Chem. Int. Ed., 1998, 37, 550–575. 21. Y. Xia and G. M. Whitesides, Annual review of materials science, 1998, 28, 153–184. 22. S. Förster, L. Apostol, and W. Bras, J Appl Cryst, 2010, 43, 639–646. 150 23. S. Förster, S. Fischer, K. Zielske, C. Schellbach, M. Sztucki, P. Lindner, and J. Perlich, Advances in Colloid and Interface Science, 2011, 163, 53–83. 24. JPIV v1.0: Open source software package for particle image velocimetry. Available at www.jpiv.vennemann-online.de. Accessed July 1, 2013. 25. S. Förster and T. Plantenberg, Angew. Chem. Int. Ed., 2002, 41. 26. M. Antonietti and S. Förster, Advanced Materials, 2003, 15, 1323–1333. 27. S. Förster, M. Konrad, and P. Lindner, Phys. Rev. Lett., 2005, 94. 28. C. Hanske, J. Erath, C. Kühr, M. Trebbin, C. Schneider, A. Wittemann, and A. Fery, Zeitschrift fur Physikalische Chemie, 2012, 226, 569–584. 29. K. S. Cole and R. H. Cole, J. Chem. Phys., 1941, 9, 341. 30. K. S. Cole, J. Chem. Phys., 1942, 10, 98. 31. M. M. Cross, J Coll Sci Imp U Tok, 1965, 20, 417–437. 32. L. Pollack, M. Tate, N. Darnton, J. Knight, S. Gruner, W. Eaton, and R. Austin, PNAS, 1999, 96, 10115–10117. 33. L. Pollack, M. W. Tate, A. C. Finnefrock, C. Kalidas, S. Trotter, N. C. Darnton, L. Lurio, R. H. Austin, C. A. Batt, S. M. Gruner, and S. Mochrie, Phys. Rev. Lett., 2001, 86, 4962–4965. 34. R. Russell, I. S. Millettt, M. W. Tate, L. W. Kwok, B. Nakatani, S. M. Gruner, S. Mochrie, V. Pande, S. Doniach, D. Herschlag, and L. Pollack, Proceedings of the National Academy of Sciences, 2002, 99, 4266–4271. 35. M. E. Kinahan, E. Filippidi, S. Köster, X. Hu, H. M. Evans, T. Pfohl, D. L. Kaplan, and J. Wong, Biomacromolecules, 2011, 12, 1504–1511. 36. G. Kaune, M. A. Ruderer, E. Metwalli, W. Wang, S. Couet, K. Schlage, R. Röhlsberger, S. V. Roth, and P. Müller-Buschbaum, ACS Appl. Mater. Interfaces, 2009, 1, 353–360. 37. M. Rawolle, M. A. Niedermeier, G. Kaune, J. Perlich, P. Lellig, M. Memesa, Y.-J. Cheng, J. S. Gutmann, and P. Müller-Buschbaum, Chem. Soc. Rev., 2012, 41, 5131. 38. M. A. Ruderer, S. Guo, R. Meier, H.-Y. Chiang, V. Körstgens, J. Wiedersich, J. Perlich, S. V. Roth, and P. Müller-Buschbaum, Adv. Funct. Mater., 2011, 21, 3382–3391. 151 152 Supplemental Information Figure S1 Experimental setup for capturing structural information of complex fluids in situ at the microfocus beamline P03/MiNaXS at the highly brilliant synchrotron source PETRA III (HASYLAB, DESY). 11 The typical experimental setup consist of three fundamental elements: microfocused X-ray beam, X-ray compatible microfluidic device and X-ray detector. This combination of microfluidics and X-ray microbeams is a powerful experimental technique which offers various advantages compared to non-continuous SAXS experiments. Figure S2 Polarization microscopic image for the illustration that the perpendicular micelle orientation effect is maintained after the addition of small polystyrene tracer spheres (1.65 µm radius, PDI 1.01). 153 7. Trebbin, M.; Thiele, J.; With, S.; Benecke, G.; Buffet, A.; Abul Kashem, M.; Perlich, J.; Müller-Buschbaum, P.; Roth, S. V.; Förster, S. HASYLAB User's Meeting 2012 - Satellite Meeting: Status and Perspectives of Small Angle X-ray Scattering at DESY, DESY Hamburg, January 25.-27., 2012. "Status of the microfluidic T-SAXS Project at P03/MiNaXS". 8. Trebbin, M.; Thiele, J.; With, S.; Buffet, A.; Perlich, J.; Benecke, G.; MüllerBuschbaum, P.; Roth, S. V.; Förster, S. 25th ECIS Conference (European Colloid and Interface Society) & 45th Biennial Meeting of the German Colloid Society, Technical University of Berlin, September 4.-9., 2011. "Particle-orientation control in microfluidic devices". 9. Trebbin, M.; Buffet, A.; Perlich, J.; Benecke, G.; Thiele, J.; With, S.; Körstgens, V.; Rawolle, M.; Herzog, G.; Müller-Buschbaum, P.; Roth, S. V.; Förster, S. HASYLAB User's Meeting 2011 - Satellite Meeting: Status and Perspectives of Small Angle X-ray Scattering at DESY, DESY Hamburg, January 26.-28., 2011. "Latest results from microfluidics at MiNaXS". 10. Trebbin, M.; Fischer, S.; Taheri, S.; Meyer, A.; With, S.; Thiele, J.; Förster, S. 2nd TUM-HASYLAB Colloquium “The metal-polymer interface”, DESY Hamburg, November 2.-3., 2010. "A microfluidic sample environment at a microfocus beamline - basics and perspectives". Poster presentations: 11. Trebbin, M.; Krüger, K.; DePonte, D.; Roth, S. V.; Schulz, J.; Chapman, H. N.; Förster, S. XFEL User's Meeting 2013, DESY Hamburg, January 23.-25., 2013. "Microfluidic liquid jet systems". 12. With, S.; Fürst, C.; Trebbin, M.; Chen, X.; Bartz, C.; Roth, S. V.; Förster, S. HASYLAB User's Meeting 2013, DESY Hamburg, January 23.-25., 2013. "Scanning structural evolution of lyotropic phases with microfluidics & microfocus SAXS". 13. Trebbin, M.; Hofmann, E.; Blüm, C.; Heidebrecht, A.; Lang, G.; Albrecht, G.; Küttner, M.; Freytag, A.-S.; Bargel, H.; Scheibel, T.; Förster, S. 4th Scientific Seminar of the North-Bavaria Biomaterials Alliance (NBBA), University of Erlangen, November 27., 2012. "Spider silk fiber formation in microfluidic devices". 256 14. Trebbin, M.; With, S.; Steinhauser, D.; Perlich, J.; Roth, S. V.; Zimmermann, W.; Thiele, J.; Förster, S. Microfluidics 2012, EMBL Heidelberg, July 25.-27., 2012. "Microfluidics meets microfocus SAXS: fast screening and correlation of complex fluid behavior with structural information". 15. Trebbin, M.; Thiele, J.; With, S.; Benecke, G.; Steinhauser, D.; Koerstgens, V.; Buffet, A.; Kashem, M. A.; Perlich, J.; Müller-Buschbaum, P.; Roth, S. V.; Förster, S. HASYLAB User's Meeting 2012, DESY Hamburg, January 25.-27., 2012. "Microfluidics meets microfocus SAXS: particle-orientation control studied insitu". 16. Trebbin, M.; Thiele, J.; Steinhauser, D.; Perlich, J.; Roth, S. V.; Förster, S. 1st Bonn Humboldt Award Winners' Forum of the Alexander von Humboldt Foundation - “Frontiers in Macromolecular and Material Science”, Bonn, October 12.-16., 2011. "Microfluidics meets microfocus SAXS: particle-orientation control studied insitu". 17. Trebbin, M.; Thiele, J.; With, S.; Perlich, J.; Roth, S. V.; Förster, S. Bayreuth Polymer Symposium ‘11, University of Bayreuth, September 11.-13., 2011. "Particle-orientation control in microfluidic devices". 18. Thiele, J.; Trebbin, M.; With, S.; Perlich, J.; Förster, S. Makromolekulares Kolloquium, University of Freiburg, February 24.-26., 2011. “Shear orientation in microfluidic channels”. 19. Thiele, J. ; Trebbin, M.; Förster, S. 44th Biennial Meeting of the German Colloid Society, University of Hamburg, September 28.-30, 2009. "Preparation of monodisperse block copolymer vesicles via flow focusing in microfluidics". 257 258 10 Acknowledgements I would like to express my deepest graditude to my supervisor Prof. Dr. Stephan who gave me the opportunity to work on both exciting and challenging research topics whilst continuously supporting me in every imaginable aspect. Stephan, I thank you from my heart. I would also like to thank all my friends, colleagues and coworkers at the University of Bayreuth, the University of Hamburg and at DESY. I would like to thank all the people who contributed to this work, especially all my students who worked with me on many research projects and who contributed greatly to the research in the Förster group. Lastly, I would like to thank my family and friends for their ongoing support and wisdom. I can’t tell how much I appreciate what you have done for me. 259 260 11 (Eidesstattliche) Versicherungen und Erklärungen (§ 5 Nr. 4 PromO) Hiermit erkläre ich, dass keine Tatsachen vorliegen, die mich nach den gesetzlichen Bestimmungen über die Führung akademischer Grade zur Führung eines Doktorgrades unwürdig erscheinen lassen. (§ 8 S. 2 Nr. 5 PromO) Hiermit erkläre ich mich damit einverstanden, dass die elektronische Fassung meiner Dissertation unter Wahrung meiner Urheberrechte und des Datenschutzes einer gesonderten Überprüfung hinsichtlich der eigenständigen Anfertigung der Dissertation unterzogen werden kann. (§ 8 S. 2 Nr. 7 PromO) Hiermit erkläre ich eidesstattlich, dass ich die Dissertation selbständig verfasst und keine anderen als die von mir angegebenen Quellen und Hilfsmittel benutzt habe. Ich habe die Dissertation nicht bereits zur Erlangung eines akademischen Grades anderweitig eingereicht und habe auch nicht bereits diese oder eine gleichartige Doktorprüfung endgültig nicht bestanden. (§ 8 S. 2 Nr. 9 PromO) Hiermit erkläre ich, dass ich keine Hilfe von gewerbliche Promotionsberatern bzw. - vermittlern in Anspruch genommen habe und auch künftig nicht nehmen werde. .................................................................................................... Ort, Datum, Unterschrift 261