SCALABLE CONTINUOUS-FLOW PROCESSES FOR MANUFACTURING PLASMONIC NANOMATERIALS PRASANNA GANESAN KRISHNAMURTHY (B.E., UNIVERSITY OF MUMBAI, INDIA) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF ENGINEERING DEPARTMENT OF CHEMICAL AND BIOMOLECULAR ENGINEERING NATIONAL UNIVERSITY OF SINGAPORE 2013 DECLARATION I hereby declare that the thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which have been used in the thesis. This thesis has also not been submitted for any degree in any university previously. _________________ Prasanna Ganesan Krishnamurthy 16th September 2013 ii Acknowledgements The two years spent as a masters student at NUS have been a fantastic learning experience. Gladly taking me under his wing, and guiding me with a gentle push and giving all the freedom one could ask for, my advisor Dr.Saif Khan made these formative years in research truly a memorable experience. Looking at everyday problems that research throws at you with scientific rigor, and that contagious enthusiasm and excitement, Saif inspired us all. I thank him sincerely for all the opportunities and his constant encouragement and support, and teaching me how to appreciate research and science. I highly appreciate his endeavor for fostering such a rich and dynamic learning environment in the research group. I was fortunate to have had the opportunity to work with Dr. Taifur Rahman during my first year in NUS; a fantastic mentor who was also a great colleague and friend. The Khan Lab environment was a one filled with a fun and friendly vibe and I thank all the people behind it. Thank you Pravien for all the help, advice, suggestions, ideas and for all our discussions on anything and everything. Working with AJ and listening to all his innovative (and many times crazy) new ideas for startups and fixes for experiments, and a new perspective on everything was fun and great food for thought. It was great knowing and working with Zahra, Abu, Dominik, Reno, Arpi, Josu, Swee Kun, Suhanya, Sophia and all the undergraduates in the lab; thank you guys for all the fun times in the lab and during our outings. Zahra, Reno and Arpi, it was a pleasure knowing you guys and thanks for all the good times. Zita, your presence as a friend and well wisher was a great support. I thank you for all the help and your friendship. I thank Arghya, Maninder, Prhashanna, Bhargav, Rajnish,Shruthi, Neerja for their friendship. Thank you Romil and Bharat for being great roommates and making home a fun place to go back to at the end of the day. Many thanks to Kriti for her company and great Indian food. I can’t thank you enough Aditi, you have been my greatest support. Words cannot express my gratitude towards my parents, without whose support, love and sacrifice, I couldn’t have been here. iii Contents Prologue ................................................................................................................... vi List of Figures ........................................................................................................ viii List of symbols ......................................................................................................... xi Nanoparticles and Nanoshells .................................................................................... 1 1.1 Gold nanoparticles and nanoshells .................................................................... 1 1.2 Gold Nanoshell applications ............................................................................. 5 1.2.1 Imaging ..................................................................................................... 6 1.2.2 Therapy ..................................................................................................... 7 1.2.3 Drug delivery ............................................................................................. 8 1.2.4 Bioassays ................................................................................................... 8 1.2.5 Harnessing Solar energy ............................................................................ 9 1.3 Nanoshell synthesis and challenges ................................................................ 10 1.4 Overview........................................................................................................ 15 Microfluidics and Nanomaterials ............................................................................. 16 2.1 Rise of Microfluidics ...................................................................................... 16 2.1.1 High Surface to Volume ratios ................................................................. 17 2.1.2 Mass Transport ........................................................................................ 17 2.1.3 Low consumption volumes ...................................................................... 18 2.1.4 The numbers game ................................................................................... 18 2.2 Microfluidics for nanomaterials synthesis ....................................................... 21 2.2.1 Single phase microfluidic methods ........................................................... 21 2.3 Microfluidics and Gold nanoshells ................................................................. 28 2.3.1 Liquid phase reagents............................................................................... 28 2.3.2 Gaseous Reagents .................................................................................... 30 2.4 The path ahead ............................................................................................... 31 Microfluidics for nanomaterials synthesis using reactive gases ................................ 32 3.1 Introduction .................................................................................................... 32 3.2 Experimental Details ...................................................................................... 34 3.3 Synthesis of gold nanoparticles....................................................................... 38 3.4 Gold nanoshell synthesis ................................................................................ 40 3.5 Dynamical tunability of particle morphologies ............................................... 42 3.6 Gas-Liquid mass transfer ................................................................................ 45 iv 3.7 Overview........................................................................................................ 49 Scaling-up nanomaterials synthesis .......................................................................... 51 4.1 Introduction .................................................................................................... 51 4.2 Scale-up in Microfluidics ............................................................................... 53 4.3 “Milli-fluidics” ............................................................................................... 55 4.4 Concept and development............................................................................... 57 4.5 Experimental Details ...................................................................................... 59 4.6 Results and Analysis ...................................................................................... 61 4.6.1 Effect of droplet morphology and flow velocity on product quality .......... 63 4.7 Conclusion ..................................................................................................... 68 Epilogue .................................................................................................................. 70 Thesis Contributions ............................................................................................ 70 Future directions .................................................................................................. 71 Bibliography ............................................................................................................ 72 v Prologue Nanotechnology and nanomaterials attract a tremendous amount of interest from academia and industry with over US$ 67.5 billion spent by the world governments for nanotechnology based research over the past decade. As of 2011, this figure stands at about US$ 10 billion per year worldwide and is increasing. The past decade saw the development of numerous applications of nanomaterials ranging from sensing and biological assays to cancer treatment, bio-imaging, drug delivery and solar energy harvesting. Most applications especially like optical and biomedical require tight control over nanoparticle sizes and shapes. As nanomaterials based research reaches maturity, the translation of such discoveries to real world technologies is limited by the ability to synthesize speciality nanomaterials with careful control over morphologies in large quantities. Microfluidics has emerged as a promising tool for controlled synthesis of nanomaterials, but has mostly been limited by the complexity of operation, costs and extremely small throughputs. This thesis aims to develop a microfluidic method for the facile synthesis of nanomaterials with dynamic control of morphologies, and the further scale-up such a system for achieving higher production rates, maintaining the tunability and control possible in microreactors. This work focuses on the synthesis of a special class of nanomaterials: plasmonic gold-silica nanoshells. Moving away from traditionally used liquid phase reducing agents, a gaseous reagent: carbon monoxide is used in this work. A novel droplet microfluidic method using a parallel channel configuration is first developed for the easy and safe integration of a toxic gas like CO in the reactor. Exploiting the exquisitely controlled gas-liquid contacting made possible using this method and the enhanced transport and mixing in microscale droplets, broad tunability of morphologies (gold coverage over silica) and hence optical properties is vi demonstrated. Finally, building upon the parallel channel concept for introducing gases into the system, a scaled-up millifluidic method is developed. Optimizing the process parameters, upto 25 times increase in production rates compared to the chip based method is demonstrated with excellent control over nanoparticle morphology. This work aims to pave the path ahead for scale-up of microfluidic methods for the synthesis of nanomaterials with complex and fast chemistries. vii List of Figures Figure 1.1: (a) Theoretically calculated optical resonances of metal nanoshells over a range of shell thicknesses. (b) Plasmon response of a nanoshell due to the interaction between a sphere and a cavity plasmon. (c) Tunable optical properties of gold nanorods by changing the aspect ratios - Top: TEM images of nanorods with increasing aspect ratios from left to right; Bottom left: Colours exhibited by aqueous nanorods suspensions; Bottom Right: Corresponding UV-Vis spectra of the nanorods suspensions. Figure 1.2: (a) Theoretically calculated optical resonances of metal nanoshells over a range of shell thicknesses. (b) Plasmon response of a nanoshell due to the interaction between a sphere and a cavity plasmon. (c) Tunable optical properties of gold nanorods by changing the aspect ratios - Top: TEM images of nanorods with increasing aspect ratios from left to right; Bottom left: Colours exhibited by aqueous nanorods suspensions; Bottom Right: Corresponding UV-Vis spectra of the nanorods suspensions. Figure 1.3: (a) Schematic of nanoparticle-enabled solar steam generation (b) The tuning of the absorption cross section of the gold nanoshells to overlap the solar spectral irradiance. (c) Plasmonic light-trapping geometries for thin-film solar cells – a: Trapping of light scattered by plasmonic particles at the surface of solar cell; b: Light trapping by the excitation of localized surface plasmons in metal nanoparticles embedded in the semiconductor; c: Light trapping by the excitation of surface plasmon at the metal/semiconductor interface. Figure 1.4: (a) Schematic of the steps involved in nanoshell synthesis (b) Top: Schematic of reduction of gold ions over the surface of silica in the presence of gold seeds; Bottom: Different stages of growth of the gold shell over the silica core. Figure 2.1: (a) Schematic of the single phase multistage microreactor system for gold and silver nanoparticle synthesis and TEM image of the gold nanoparticles synthesized in this system using sodium borohydride. (b) Schematic of the microreactor system for multistep synthesis of iron oxide-silica nanoshell structures and TEM images showing the nanoshells obtained. (c) Photograph of the siliconpyrex microreactor for gold nanoparticle synthesis. The photo also shows the deposition of gold on the channel walls after usage. Figure 2.2: (a) Schematic of reagent injection and droplet generation in the device used for gold nanorods synthesis and TEM images of gold nanorods of different aspect ratios obtained. (b) Photograph of the device used for iron oxide nanoparticle synthesis showing generation of droplets containing the two reagents and downstream coalescence by electric actuation. TEM image shows the iron oxide nanoparticles synthesized using this method. viii Figure 2.3: (a) Schematic of generation of foams in the microchannels for gold nanoshell synthesis using liquid phase reducing agent. Alongside is a TEM image of the gold nanoshells synthesized. (b) Schematic of generation and merging of droplets and bubbles to form compound droplets for nanoshell synthesis using carbon monoxide. On the side is a TEM image of the nanoshells obtained using this method. Figure 3.1: Concept of membrane-based droplet microfluidic device for materials synthesis with a reactive gas. Figure 3.2: Schematic of the experimental setup. Insets: (i) stereomicroscopic image of aqueous (AQ) droplet formation in fluorinated oil (FO), (ii) aqueous droplets in the parallel channel network; carbon monoxide in the gas channel is dosed into aqueous droplets in liquid channel through the intervening PDMS membrane. Scale bars: 300 µm. Figure 3.3: UV-Vis absorbance spectrum of colloidal gold synthesized using the microfluidic device. Insets: TEM image of gold nanoparticles and measured particle size distribution. Figure 3.4: Deposition of colloidal gold on the wall of liquid microchannel running in parallel to CO channel. Figure 3.5: Gold-silica core-shell particles obtained by changing only the silica volume fraction in the reagents at a fixed residence time of 60 sec: TEM images of 230±20 nm silica particles with: (a) nano-islands (fs=4 x10-5 in 0.36 mM K-Gold), (b) almost coalesced nanoislands (fs= 3.4 x10-5 in 0.37 mM K-Gold). (c) complete nanoshell (fs= 2.8 x10-5 in 0.37 mM K-gold). (d) Ensemble UV-Vis spectra for all particles (a)-(c). Figure 3.6: (a) Ensemble optical absorbance spectra for samples from 5 batch experiments using fs= 2.8 x10-5 in 0.37 mM K-gold (b) Ensemble optical absorbance spectra for 6 samples collected during 6 hours of continuous microfluidic synthesis using fs= 3.4 x10-5 in 0.37 mM K-Gold Figure 3.7: Gold-coated silica particles obtained by changing droplet residence times at fixed gold and silica concentrations. TEM images of 230±20 nm silica particles (fs=2.8x10-5 in 0.37 mM K-Gold) with: (a) pre-attached gold seeds, and (b)-(d) varying degrees of gold coverage. (e) Ensemble UV-Vis spectra for all particles (a)(d) Figure 3.8: Gold-silica core-shell particles obtained by changing the residence times with a fixed fs=3.1 x10-5 in 0.37 mM K-Gold : TEM images of 230±20 nm silica particles with: (a) sparse nano-islands, (b) nanoislands, (c) almost coalesced nanoislands (d) complete nanoshell. Figure 3.9: (a) Plot showing fractional coverage (Fc) of gold over the silica cores of particles (measured by digital image analysis of TEM micrographs in MATLAB TM) ix synthesized using two different but fixed silica particle concentrations (fs=2.8 x10-5 and fs=3.1 x10-5 in 0.37 mM K-Gold). (b) Plot of absorbance maxima in the UV-vis spectra against residence time for particles synthesized using two different but fixed concentrations (fs=2.8 x10-5 and fs=3.1 x10-5 in 0.37 mM K-Gold) of silica particles. Figure 4.1: Concept of tube-in-tube based droplet milli-fluidic reactor for materials synthesis with a reactive gas. Figure 4.2: Schematic of the experimental setup consisting of infusion pumps, crossflow droplet generator, stainless steel (SS) outer and PTFE inner tubes. The outer tube is pressurised with carbon monoxide. The whole setup is within a fume hood. Figure 4.3: Gold-silica core-shell particles obtained by changing only the silica volume fraction in the reagents at a fixed residence time of ~ 1 minute: TEM images of 230±20 nm silica particles with: (a) nano-islands (fs=2.265 x10-5 in 0.37mM Kgold), (b) almost coalesced nanoislands (fs= 1.7x10-5 in 0.37mM K-gold). (c) complete nanoshell (fs= 1.13 x10-5 in 0.37mM K-gold). (d) Ensemble UV-Vis spectra for all particles (a)-(c). Figure 4.4: Microscopic images of droplet formation using the cross flow fitting. (a)(d) show increasing droplet lengths with increasing aqueous flow rates keeping oil flow rates constant. Figure 4.5: Gold-coated silica particles obtained by changing droplet velocity(v) and residence times(τ)( at fixed gold and silica concentrations. TEM images of 230±20 nm silica particles (fs= 1.13 x10-5 in 0.37mM K-gold): (a) v=5.3 mm/sec & τ =113 sec (b) v=7.4 mm/sec & τ =81 sec (c) v=8.4 mm/sec & τ = 71 sec (d) v=9.5 mm/sec & τ =63 sec x List of symbols τ Residence time (sec) fs Volume fraction of silica particles in the aqueous phase U Flow velocity (m/sec) µ Viscosity of the continuous phase (N-sec/m2) γ Interfacial tension between the continuous and dispersed phase (N/m) ρ Density of the continuous phase (Kg/m3) D Diffusivity (m2/sec) H Henry’s Law constant (mol/m3-Pa) CS Solubility of gas in the liquid phase (M) Pe Peclet number Re Reynolds number Ca Capillary number tD Diffusion time (sec) tm Mixing time (sec) k La Volumetric mass transfer coefficient (sec-1) Vd Volume of the droplet (m3) xi Chapter 1 Nanoparticles and Nanoshells Silica-Gold nanoshells and nanoislands are a special class of plasmonic nanomaterials that have garnered a lot of interest in the past decade. Beginning with an introduction to metallic nanomaterials the advantages of such gold nanoshells are discussed. Some of the important applications of gold nanoshells are reviewed. Finally the conventional techniques for synthesizing these core-shell structures and the challenges involved are discussed to better understand the need for developing alternative microfluidic techniques for their synthesis, which is the focus of the forthcoming chapters. 1.1 Gold nanoparticles and nanoshells Nanomaterials today have far flung applications from medicine 1-3 to enhancing oil recovery from reservoirs4, 5 and have tremendous attention from the research community as well as industry6-8. Gold based metallic nanomaterials and metaldielectric hybrid nanomaterials synthesis will be the focus of this work. Metallic nanoparticles suspensions have fascinated men for centuries with their unique optical and surface active properties not observed in their bulk forms. From the stained glass of roman times, to the famous Purple of Cassius 9 in the seventeenth century to Faraday’s sols10, gold nanoparticles have attracted immense attention. The colours exhibited by nanoparticle suspensions was attributed to the absorption and scattering of light by the particles by Gustav Mie in 1908 11. Metals are great conductors of heat and electricity due to the overlap of the conduction and valence bands and the highly delocalized nature of the electrons. When the size of metals is 1 reduced to dimensions smaller than the mean free path of the electrons, intense absorption in the near UV and visible part of the spectrum is observed. This is due to the coherent oscillation of the free electrons on the surface of the metals at a frequency equal of that of the electromagnetic wave.12, 13 This is known as surface plasmon resonance (SPR). Noble metals especially gold and silver are unique, because their free electron densities lie in a range which makes their nanoparticles exhibit SP peaks in the visible region14. The brilliant colour of nanoparticle suspensions is due to this phenomenon. This resonant frequency has been known to depend on the type of metal, their dielectric environment and crucially on the size and shape of the nanoparticles. As the size or shape of the nanoparticles change, the change in the surface geometry causes a shift in the electric field density on the surface resulting in a change in oscillation frequency of the electrons. 15 Manipulation of the frequency and hence the wavelengths at which particles absorb and scatter light has been done for years by mainly varying their sizes and shapes and their dielectric surroundings. Metallic nanomaterials especially gold of different shapes and sizes have found applications in diagnostics and therapy16-18, assays19-20, bio-imaging21, drug delivery22, nanomedicine23, 24 , biosensing25, photo-catalysis26 and more. Bottom-up methods (starting from molecular level to desired structures) are widely used for the controlled synthesis of metallic nanomaterials. The mechanism of formation of nanoparticles usually involves two steps: nucleation followed by growth27. Most metallic nanoparticle syntheses (leaving aside kinetically controlled growth mechanisms28 of anisotropic particles) are characterized by very fast nucleation and growth kinetics. Tight control over the nucleation and growth stages is necessary29, 30 to obtain monodisperse sizes and shapes of such nanoparticles. 2 The limitation here is that the tunability of the plasmon resonance of solid gold nanoparticles is restricted as the resonance frequencies shift to longer wavelengths only with increasing particle sizes and the plasmon response to size change is found to be weak.31 Capitalizing on the fact that the plasmon resonance also depends on the dielectric constant of the surrounding medium, the attachment of solid nanoparticles to larger beads or supports was found to give unique optical properties. In the late 1990s, the Halas group at Rice University introduced a new type of hybrid particle where a dielectric nanoparticle: silica was decorated with gold nanoparticles32. This new breed of metallodielectric nanoparticles known as gold nanoshells exhibited enhanced optical properties compared to solid nanoparticles of equivalent size.33 The equivalent plasmon response of a nanoshell was attributed to the interaction of plasmons, characteristic of a sphere and a cavity.34 The unique property was the facile tunability of the optical response by changing the physical characteristics. Unlike gold nanoparticles gold nanoshells could be used to access a wide range of wavelengths. By changing the control parameters like size of the silica core, thickness of the shell and coverage of gold over the silica surface, SP bands of these particles could be shifted continuously from wavelengths ranging from the visible to the near-infrared regions of the electromagnetic spectrum31. This kind of functionality provided a great amount of freedom to design such particles and tailor them for specific applications. Spherical gold or silver nanoparticles could not be used to achieve this kind of facile optical response. Other nanoparticle geometries that displayed such shape based tunability were obtained when anisotropy was introduced, the most popular among which has been gold nanorods.31 The plasmon in nanorods could be tuned by varying their aspect ratios. The aspect ratio defines the two resonance frequencies corresponding to the interaction of light with the longitudinal and transverse 3 dimensions of the nanorods.12 Nanorods are produced by introducing shape inducing agents like surfactants during the growth step of the synthesis.35, 36 By adhering to specific facets of the nuclei the surfactants do not allow further deposition and growth in those facets thereby introducing anisotropy.37, 38 Nanorods and this family of nanomaterials have been useful for various applications especially for bio-imaging and therapy.18, 39 Nanorods are mostly synthesized using CTAB (cetyl trimethylammonium bromide), a surfactant which as mentioned earlier acts as a shaping and importantly as a capping agent. Due to the cytotoxicity of CTAB, applications of nanorods are currently limited.40 Also since they are highly surfactant stabilized, the loss or removal of these surfactant layers or just storing for prolonged time periods causes them to lose their shape. Nanoshells on the contrary are structurally stable, and have been found to be benign and biocompatible and can be further surface functionalized with a variety of substrates for different applications: biological and otherwise. Ever since their conception by Halas and group, this family of core-shell structured nanomaterials have received much attention and fuelled intense fundamental and applied research. 4 Fig 1.1: (a) Theoretically calculated optical resonances of metal nanoshells over a range of shell thicknesses.32 (b) Plasmon response of a nanoshell due to the interaction between a sphere and a cavity plasmon.31 (c) Tunable optical properties of gold nanorods39 by changing the aspect ratios - Top: TEM images of nanorods with increasing aspect ratios from left to right; Bottom left: Colours exhibited by aqueous nanorods suspensions; Bottom Right: Corresponding UV-Vis spectra of the nanorods suspensions. 1.2 Gold Nanoshell applications The facile ability to tune the absorption frequency of gold nanoshells has been their greatest advantage. The ability to tune the resonance of nanoshells near the infrared part of the spectrum combined with their bio-compatibility and their ease of bioconjucation make them ideal for a variety of biomedical applications41, 42. The two 5 major focus areas as far as application of nanoshells go have been bio-imaging and photo thermal therapy. The near-infrared part of the spectrum called the “water window” is the region where the physiological transmissivity of tissues and blood is the highest and where they act transparent letting light to penetrate through them with minimal heating and scattering limited attenuation. Light in this spectral region has been observed to penetrate more than 1cm into the tissues without any observable damage43. 1.2.1 Imaging Conventionally organic fluorescent dyes and NIR dyes such as indocyanine green have been tested and used as contrast agents for imaging cancer cells. Even though optical imaging using such dyes is cheap, it is limited by weak optical signals and lack of contrast between the tumorous and surrounding benign tissues. Gold nanoshells owing to their excellent absorbance properties and biocompatibility have been used effectively for such imaging purposes44. Compared to a dye like indocyanine, gold nanoshells have been found to be more than a million times more absorbent towards NIR light. Thus excellent contrast can be achieved using such nanomaterials. Also the structural stability of nanoshells makes them less susceptible to in-situ chemical or thermal degradation than conventional dyes. The ability to functionalize the nanoshell surfaces easily with bio-molecules or conjugate with antibodies and biomarkers make selective uptake possible. Thus the nanoparticles attach only to the targeted malignant tissues and not the surroundings thereby making contrast based imaging easier (Fig 1.2a). Halas and West have demonstrated conjugation of nanoshells with anti-HER2 breast cancer biomarkers for in-vitro imaging of the cancerous tissues45. 6 1.2.2 Therapy Non invasive thermal ablation of tumors involves the administration of a minimal but lethal dose of heat to the target tissues. Conventionally this is done by placing probes in the interstices or between cavities. Ultrasound, Laser, RF and microwave therapies which are extracorporeal have also been demonstrated. However all such therapies suffer from the same limitation: non-specific delivery of heat resulting in death of surrounding tissues. Tumor specific heat treatment has been done by heating iron oxide nanoparticles using altering magnetic fields. However for effective treatment this method needed large quantities of iron nanoparticles to be delivered to the tumor sites. Gold nanoshells can be easily conjugated with biomolecules so that they specifically attach to the target cells and can be used for high contrast imaging as discussed in the previous section. Since nanoshells are intense absorbers (a million times stronger than conventional dyes), when NIR light of the resonant wavelength of the nanoshells is administered, the amount of heat generated locally due to the conversion of the electron kinetic energy to thermal energy is capable of destroying the surrounding tissues. Due to the leaky nature of cancer tissues, nanoparticles smaller than 400nm have enhanced permeability and retention within such cells, making therapy more effective and localised45. Halas and West have demonstrated extracorporeal photo thermal ablation of tumors in-vitro as well as in-vivo (in mice). Using PEG functionalised NIR absorbing gold nanoshells and NIR lasers, they were able to induce localised cell death confined to the nanoshell treatment area (Fig 1.2a). Interestingly incubation of these cells with PEG coated nanoshells without any NIR treatment resulted in no cell mortality thereby proving the non-cytotoxic nature of gold nanoshells43. 7 1.2.3 Drug delivery The same principle of heat generation by absorption at resonant wavelengths has been exploited for controlled release of drugs. Drug-laden temperature sensitive hydrogel embedded with nanoshells, when exposed to light, experience temperature increase due to the heat generated at the surface of the nanoshells thereby causing the hydrogels to rupture and release the contents. Acrylamide based reversible polymeric hydrogels embedded with gold-gold sulphide nanoshells that absorb at the NIR part of the spectrum have been demonstrated for controlled drug delivery46. When exposed to NIR laser the hydrogels collapse thereby releasing the contents. Due to the sudden collapse the drugs experience convective release into the environment. When the NIR source is stopped, the gels swell back in sometime thereby stopping drug release. Fig 1.2: (a) Combined imaging and therapy of SKBr3 breast cancer cells using HER2-targeted nanoshells.45 The top row shows the increase in contrast during imaging while using target specific functionalized nanoshells. The middle row shows the highly localized cell death using anti-HER2 functionalized nanoshells. Bottom row shows the silver staining assessment to determine nanoshell binding (dark spots) (b) Gold nanoshells for blood immunoassays UV-vis spectrum of disperse nanoshells and spectrum of nanoshells/antibody conjugates following addition of analyte.47 1.2.4 Bioassays Nanoshells due to their strong optical signals have been used for immunoassays to determine target analyte concentrations in blood. In conventional optical immunoassays carried out under visible light, several purification steps are involved since a variety of biomaterials present in the sample absorb visible light. Nanoshells 8 that absorb in the near IR part of the spectrum can be conjugated with antibodies that interact with the specific analyte of interest. In the presence of the analyte the nanoshells in the blood aggregate to form dimers. The formation of dimers causes a red shift in the plasmon response of the suspension (Fig 1.2b).47 When conducted using near-IR light, the optical signals by other biomolecules present in the blood is minimal and careful quantification of the target analyte can be carried out. 1.2.5 Harnessing Solar energy Due to the greatly tunable nature of the optical properties of nanoshells they have been recently even used to more efficiently harness solar energy. By tailoring the Fig 1.3: (a) Schematic of nanoparticle-enabled solar steam generation 49 (b) The tuning of the absorption cross section of the gold nanoshells to overlap the solar spectral irradiance. 49 (c) Plasmonic light-trapping geometries for thin-film solar cells – a: Trapping of light scattered by plasmonic particles at the surface of solar cell; b: Light trapping by the excitation of localized surface plasmons in metal nanoparticles embedded in the semiconductor; c: Light trapping by the excitation of surface plasmon at the metal/semiconductor interface.50 shape and size of gold nanoshells and creating a mixture of these particles such that they absorb strongly across the solar spectrum, maximum utilization of sunlight can 9 be done (Fig 1.3b).48 Halas and group have used this technology for developing solar steam generators. The intense heat generated at the surface of the nanoparticles when light is incident upon them is used to evaporate the surrounding water (Fig 1.3a). 49 Such tailored nanoshells have also been tested to be used in photovoltaic cells in order to increase absorption efficiencies and also induce light trapping (Fig 1.3c). 50 1.3 Nanoshell synthesis and challenges The synthesis of silica-gold core shell nanostructures is carried out by growing nanometre scale gold film onto colloidal silica pre-seeded with small gold nanoparticles. The silica nanoparticles are synthesized separately, using the now standard one step Stober / modified Stober process51 or the seeded growth process for synthesizing silica particles of larger sizes with high monodispersity52. In order to decorate the silica surface with gold nanocrystals, its surface is first modified with functional molecules to enhance coverage of gold. The surface is modified commonly by using various silanes53, 54. The most commonly used APS (3-aminopropyltriethoxy silane) is a bi-functional organic molecule having an ethoxy group on one end and NH group on the other. The ethoxy group forms a bond with the OH terminated silica surface thereby making the surface now NH terminated. Gold nanoparticles in the range of 2-4 nanometres are synthesized by reducing chloroauric acid by tetrakishydroxymethylphosphonium chloride (THPC) which also acts as a capping agent55. These THPC gold particles are negatively charged56. When introduced into a solution containing the APS coated silica particles, these gold nanoparticles attach to the surfaces due to electrostatic attraction. In order to grow a shell onto the seeded surface, a gold plating solution is used. The plating solution is an aged mixture of chloroauric acid and potassium carbonate. Addition of potassium carbonate increases 10 the pH to about 7.5 and results in the hydrolysis of gold chloride to form various gold chloride – hydroxide species, the most dominant being AuCl3(OH)- , AuCl2(OH)2AuCl(OH)3-. This speciation of chloroauric acid has been shown to favour the size controlled gold nanoparticle synthesis. However at higher pH (pH> 10), Au(OH)4species is seen to dominate which has been found to have lower tendency to be reduced therebye affecting nanoparticle synthesis and growth. To initiate the growth, the seeded silica particles are added to the plating solution in the presence of a reducing agent. Fig 4: (a) Schematic of the steps involved in nanoshell synthesis54 (b) Top: Schematic of reduction of gold ions over the surface of silica in the presence of gold seeds; Bottom: Different stages of growth of the gold shell over the silica core.32 Here the Au3+ ions from the plating solution are reduced to Au0 onto the silica surface, catalysed by the pre-seeded gold nanoparticles on the lines of electroless plating. This causes the existing gold nanoclusters on the surface to grow and coalesce to form nano-islands and eventually form a continuous gold shell. The nanoshell growth method was first demonstrated by the Halas group using sodium borohydride as the reducing agent 53. However NaBH4 being a very strong reducing agent, it gave rise to secondary nucleation thereby causing non-homogenous growth as well as colloidal gold formation in the surrounding liquid phase. Graf and 11 van Blaaderen later adopted the use of hydroxylamine hydrochloride as the reducing agent which resulted in good nanoshells54. Halas later demonstrated the use of formaldehyde with ammonium hydroxide for synthesizing silver nanoshells that was adopted even gold nanoshell synthesis57. This method in particular gave smooth nanoshell morphologies. Recently Halas and group extended the use of carbon monoxide, a gas that has been previously used for a variety metal nanomaterials synthesis58-61, for the synthesis of gold nanoshells.62 Here instead of formaldehyde or hydroxylamine hydrochloride, CO acts as a reducing agent providing the necessary electrons to reduce Au+3 to Au0. CO dissolves in the aqueous solution containing the reagents and reacts with water thereby providing the necessary free electrons. CO( g )  H 2O  CO2(aq )  2e  2 H  CO( g )  2 H 2O  HCO3  2e  3H  The reduction of Au+3 from the salt to Au is a two step process. The Au+3 is initially reduced to AuCl2− and then to Au0. AuCl4  2e  AuCl2  2Cl  AuCl2  e  Au 0  2Cl  During the gold shell growth, CO adsorbs on the surface of the existing gold nanoparticles and reacts with water to give carboxyhydroxyl intermediates which then dissociate to give carbon dioxide and free electrons. These free electrons are then taken up gold chloride- hydroxide species physisorbed on the nanoparticles and are reduced to Au0. The existing seed gold nanoparticles, as mentioned earlier, behave like a catalyst by acting as a conduit for electron flow from CO to the gold salt as the gold deposition and growth process progresses60. 12 The morphology of nanoshells has been reported to be influenced by the reducing agent used and using CO was found to produce thinner shells62 than all other liquid phase reagents; a quality that is highly desirable from the point of view of plasmon tunability (thinner shells lead to resonance at longer wavelengths). Even though liquid phase reagents have been used widely for materials synthesis gaseous phase reagents are gaining popularity due to their inherent advantages; they do not degrade with time as is common with their liquid analogues, produce simple by-products, and are easy to separate from the liquid-phase reaction mixture. Nanomaterials synthesis invloves careful control over reagent concentrations, but many liquid phase reagents lose activity or degrade over time thereby rendering them useless and causing variations in the experimental conditions; for example sodium borohydride (NaBH4) has to be kept cooled before usage to avoid hydrogen gas evolution causing reduction in activity. However gaseous reagents are highly stable and can be stored for extended periods without any loss of activity. The other major advantage is the ease with which the gaseous reagents can be separated. In wet chemical methods the addition of excess reducing agent is not uncommon to ensure complete reaction. The separation of the excess reagents then becomes a post processing challenge. Reagents like formaldehyde and hydroxylamine hydrochloride that are used commonly for gold nanoshell synthesis, are highly toxic, making their careful removal imperative when the end product is needed for biomedical applications. While using gaseous reducing agents like CO, by-products like CO2 and the excess gas escape into the ambient atmosphere making a separation step unnecessary. Carbon monoxide has also been known to adsorb easily onto gold surfaces, which has raised concerns about the safety of the gold nanoparticles synthesized thereof. However supported gold nanoparticles have been established as an effective catalyst for oxidation of carbon monoxide even 13 at low temperatures63, thus making the likelihood of residual carbon monoxide bound to gold nanoshell surfaces highly remote. The ease of separation of gaseous reagents also makes rapid reaction quench possible which is difficult while using liquid phase reagents. Due to the high dependence of optical properties on the size, shape and morphology of such nanoparticles tight control over these characteristics becomes essential. Synthesis of such gold nanoshells using electroless plating is usually characterized by rapid autocatalytic reaction kinetics. Duraiswamy and Khan, in a diffusion controlled environment estimate shell growth rates upto 200nm/sec with the growth going to completion within seconds64. Growing uniformly thick shells on all particles to get monodisperse populations thus requires rapid reagent dispensing and homogenization. Carrying out such syntheses involving fast kinetics in batch scale methods using stirred vials or reactors, makes the process highly mass transfer controlled. The mixing induced by traditional stirrers proves to be insufficient and slow resulting in pockets of inhomogeneities within the reaction mixture. This causes different extents of shell growth on particles at different places within the reactor resulting in polydisperse product populations65. Improper mixing also leads to nucleation and growth of gold nanoparticles outside the silica surface in the surrounding liquid phase. Valuable gold precursors get wasted as a result and the product requires an additional post-processing / cleanup step. It is therefore not surprising that even in single-phase solution based methods, nanoshells with control over morphology are synthesized only in small volumes using conventional flask-based batch processes, with limited reproducibility and often insurmountable issues with scale-up. The use of a gaseous reducing agent in conventional batch synthesis, while conferring several advantages as noted above, exacerbates the problem of reagent addition and homogenization, as 14 an additional inter-phase mass transfer step comes into the picture. Scale-up of such sensitive syntheses for large scale production thus becomes highly challenging especially if gaseous reagents are used. Moreover, safety concerns are associated with the usage of reactive gases such as CO in the laboratory or pilot plant scale because of their high flammability and/or acute toxicity. 1.4 Overview As batch scale methods faced such challenges for nanoshell synthesis and otherwise, microfluidics emerged as an effective platform for continuous synthesis of such particles where the inherently high surface to volume ratios and enhanced mass and heat transport properties gave the advantage of better control over the population size and particle morphology. It also came with the promise of easy scale up using parallelization and pile-up for large scale commercial synthesis of nanomaterials that may be required in the future. Microfluidics for nanomaterials synthesis and specifically for nanoshell synthesis becomes the focus of the next chapter. 15 Chapter 2 Microfluidics and Nanomaterials The rise of gold nanoshells as a new kind of material and a superior counterpart to solid nanocrystals was discussed in the previous chapter. Controllable and reproducible synthesis of such metallodieletric nanoparticles in small batch volumes let alone in large scale was recognised as the major challenge. Microfluidics as a platform for nanomaterials synthesis is the focus of this chapter. Starting from a brief discussion of the fundamentals and features of microfluidics, nanomaterials synthesis methods developed using this platform, with a focus on metallic and semiconductor particles are reviewed. Finally the microfluidic techniques currently established for nanoshell synthesis and their limitations are discussed thereby setting the stage for understanding the motivation behind and novelty of the current work described in the following chapters. 2.1 Rise of Microfluidics The challenges faced in reproducible synthesis of gold nanoshells described in the previous chapter are common to a variety of other nanomaterials especially metallic and semiconductor particle chemistries with fast kinetics and harsh conditions. With increasing difficulty in carefully synthesizing complex nanomaterials the community turned its attentions towards Microfluidics. Microfluidics refers to the technology of handling fluids confined in channels with cross sectional dimensions in the size range of micrometres. The origins of miniaturized flows can be traced back to the development of gas-phase chromatography (GPC) and high-pressure liquid chromatography (HPLC) based microanalytical methods which involved analysis of minute samples in small capillaries66,67. With the development of the 16 photolithography techniques for microelectronics microelectromechanical systems (MEMS) and increasing need for better and efficient analysis methods for the then booming field of genomics and molecular biology, microfluidics as a field was born in 199068. Introduced as Micro Total Analytical System (µ-TAS) where different steps or unit operations for analysis could be incorporated together onto a small device, the next decade saw its development as a platform for high resolution chemical and biological assays69. Even though µ-TAS and microfluidic chips were inspired by the microelectronics industry the rapid change in the fundamental physics with miniaturization was greater and more apparent in the former. The characteristic features of flows in such dimensions and its unique advantages are discussed below70, 71. 2.1.1 High Surface to Volume ratios One of the most obvious but important effect of miniaturization is the huge increase in surface area to relative volume by several orders of magnitude. The high surface to volume ratio results in efficient heat and mass transfer in such flows since more interface is available for transfer to occur. Thus creation and homogenization of chemical or temperature gradients is faster. 2.1.2 Mass Transport Mass transport or mixing for either laminar or turbulent flows ultimately depend son molecular diffusion. The characteristic diffusion time scale τ = L2/D where L is the diffusion length and D is the diffusivity. When the length scale is brought down from centimetres or meters to micrometers the time needed for homogenization drastically falls thereby giving microfluidic flows the mass transfer advantage. 17 2.1.3 Low consumption volumes The fluids handled in microfluidics are in the nanolitre or microlitre volume range. This is greatly advantageous for expensive chemical and biological assays where reagent volumes available are extremely small and valuable. This also enables safe handling of toxic or hazardous reagents. This is one of the main reasons for the development of numerous point of care devices using the microfluidic platform, that can be carried anywhere due to the small physical footprint and use very low sample volumes. 2.1.4 The numbers game Microfluidic flows can be characterized by certain dimensionless numbers that describe the interplay of different physical phenomena in these length scales. I. Reynolds number: Laminar flows The flows in microchannels are highly laminar and operate at very low Reynolds numbers. Reynolds number which describes the ratio of inertial to viscous forces is given as Re  dU   Where d is the channel dimension, U is the velocity, ρ is the fluid density and µ the viscosity. With micrometer sized dimensions and velocities in the range of centimetres/sec, the Reynolds number for liquid flows can range between the orders of 10 -6 to 10. This shows that the inertial forces are irrelevant in this flow regime. This has enabled highly ordered laminar flow patterns like co-flowing streams where transport is only due to diffusion to be used for studying separations and careful molecular level analysis. 18 II. Peclet Number:Mixing The peclet number describes the relative importance of convection to diffusion. Pe  Uw D Where U is the flow velocity, w is the cross sectional length of the microchannel and D is the diffusivity. Peclet numbers in microchannels usually lie in the range of 10 to 10 5. Thus convective transport is almost always faster. Laminar flows with high Pe, where only molecular diffusion is dominant has been used for laminar flow patterning where confinement between the co-flowing laminar flows is important; for careful separation based on diffusivities of the analytes72,73 and even studying reaction kinetics74. The Tmixer and H-mixer type microfluidic devices were widely popular for such studies. Some applications like studying kinetics of fast reactions or nanomaterials synthesis require extremely rapid mixing of different streams containing reagents. Single phase flows in microchannels are laminar in nature and the turbulent convective mixing observed in macroscale flows is absent. Here, as mentioned earlier, diffusive mixing is slow compared to convection of materials along the channel. For such flows, the distance along the channel that is required for complete mixing to occur (U l2/D) is directly proportional to the Peclet number and increases with higher Pe. In order to induce fast mixing transverse flows need to be generated. Ultimately the distance over which diffusion must occur for homogenization must be reduced. For laminar flows this is achieved by chaotic advection where fluid elements are stretched and folded exponentially enhancing mixing. The staggered herringbone mixer75 with asymmetric grooves patterned on the channel walls generates transverse components in the flow inducing chaotic mixing. Using chaotic advection, the distance over which complete mixing occurs scales logarithmically with Peclet number and not linearly. This 19 method also drastically reduces axial dispersion commonly observed in the parabolic flow front in laminar flows. III. Capillary Number:Multiphase flows The capillary number describes the ratio of viscous to interfacial stresses. Ca  U  Here, U is the flow velocity, µ is the viscosity and γ is the interfacial tension. Capillary numbers in microchannel flows is less than one. In the case of multiphase oil-water flows it ranges between 10 -3 to 10-2. Thus interfacial forces dominate in microfluidics. With high surface to volume ratios and dominant capillary forces, there exist a lot of opportunities for carefully manipulating flows in such channels. Functionalization of channel walls to establish wettability gradients for generating flows and inducing flows through channels by capillary forces have been demonstrated70. But the most interesting and widely used application of manipulating surface forces has been for generating high stable multiphase flows: immiscible liquid-liquid flows and gas-liquid flows76-78. The droplet microfluidic platform has gained a lot of attention and has been used for applications ranging from kinetic studies79 to crystallization80. Here one of the liquids which preferentially wets the wall forms the continuous phase and the other liquid (usually the reagent) forms the dispersed phase. Compartmentalization and isolation of nanolitre volumes into such emulsions is the major advantage over single phase flows81. Droplet flows in microchannels also experience shear induced recirculatory flows (vortices) which highly enhance mixing within them82. Faster mixing by chaotic advection can also be induced by flowing droplets through winding channels. With complete isolation each droplet can act as well stirred reaction flask. This feature has been utilised for studying kinetics of fast reactions 83. 20 Single phase flows suffer from insufficient solely diffusion based mixing. They also face the problem of axial dispersion due to the parabolic velocity profiles. The micromixer modules available to speed up mixing and channel patterning techniques to induce chaotic advection are complicated and require difficult fabrication steps. All these problems make homogeneous flows unsuitable for conducting fast reactions. Also in material synthesis, contact of the reagents with the channel walls causes fouling of the reactor walls. Droplet microfluidics circumvents all these problems. 2.2 Microfluidics for nanomaterials synthesis The enhanced mass and heat transfer characteristics, high surface to volume ratios and the ability to exercise precise control over the reactant flow, contact and mixing led to the pursuit of microfluidics as a platform for synthesis of nanomaterials. Over the past decade, numerous research groups around the world have demonstrated synthesis of a variety of nanomaterials ranging from co-block polymers and hydrogels to quantum dots and gold nanoparticles84-86. The section reviews some of those works. The focus here will be on inorganic nanomaterials, especially metallic nanoparticle synthesis. 2.2.1 Single phase microfluidic methods I. Semiconductor nanoparticles Single phase laminar flow methods were adopted in the initial days of nanomaterial synthesis using microfluidics. One of the first demonstrations of this was by the deMello and group in Imperial College London for the synthesis of cadmium sulphide nanoparticles (CdS quantum dots)87. They employed a glass-silicon-glass sandwich device with channels etched on both sides of the silicon wafer. The two reagents sodium sulphide and cadmium nitrate entered through the two inlets on two sides of the reactor. After the inlet, the channels were split into 16 parallel channels to induce 21 mixing after which the two reagents from the 16 channels meet each other, mix, react and converge into one channel and exit the reactor. In order to avoid aggregation sodium polyphosphate was added to the cadmium precursor before entering the reactor. II. Oxide nanoparticles Ali-Abou Hassan and group used the homogeneous continuous flow method for the synthesis of iron oxide- silica nanoshell type structures88. They used a set of three Fig 2.1: (a) Schematic of the single phase multistage microreactor system for gold and silver nanoparticle synthesis and TEM image of the gold nanoparticles synthesized in this system using sodium borohydride.91 (b) Schematic of the microreactor system for multistep synthesis of iron oxidesilica nanoshell structures and TEM images showing the nanoshells obtained. 88 (c) Photograph of the silicon-pyrex microreactor for gold nanoparticle synthesis. The photo also shows the deposition of gold on the channel walls after usage.89 22 coaxial silica-capillary in PDMS–Glass microreactors (Fig 2.1b) for conducting the three different steps involved in the process. In the first step, functionalization of the pre-synthesized Fe2O3 with APTES molecules is performed. Mixing of this particle suspension with the silica precursor TEOS was done in the second reactor. In the final reactor ammonium hydroxide was introduced to the above mixture to facilitate the formation of a silica shell over the iron oxide nanoparticles. In all the three reactors, 3D co-axial flows were used, where one of the reagents is injected co-axially with the other reagent and mixing occurs by flow focusing. Magnetic and fluorescent stable oxide nanoshells could be synthesized using this method. III. Gold nanoparticles The microfluidic method was adopted for gold nanoparticle synthesis many times by different groups. Due to the rapid nucleation kinetics, shape and size control were a big problem in conventional synthesis and microreactors were adopted to overcome this challenge. Wagner and Kohler were one of the first to try gold particle synthesis 89. In their first attempt they utilized a simple silicon-pyrex reactor (Fig 2.1c) with serpentine channels and three inlets for the three reactants. Small gold nanoparticles (seeds) in the 12nm range produced off chip by the citrate reduction method were used as one of the reactants. Ascorbic acid (reducing agent) and the gold seeds first meet near the inlet and flow through the channels. Midway, the third gold chloride stream joins the main channel and the reaction occurs through the course of the remaining length of the channel. 15 to 24 nm gold nanoparticles were produced using this method. The same group demonstrated a method for synthesis directly from the precursors instead of starting from gold seeds90. Here they used a silicon-pyrex based chip where 23 intermittently the channel split into two and then quickly converge, so as to increase the mixing between the two reagents. Ascorbic acid was again used as the reducing agent. Polyvinyl pyrrolidone was used a stabilizing agent to prevent product aggregation. 5 to 50 nm gold particles were synthesized in this case with narrow size distributions. Recently the same group demonstrated a general multiple stage microreactor system for synthesis of any type of metal nanoparticle91. They built their system using a series of 3-4 connected reactors (Fig 2.1a) with a similar split and recombine type channel configuration to ensure good mixing. By introducing a water stream, downstream from the main inlet but before the reaction occurred they could control the concentration as needed. They used sodium borohydride as the reducing agent which was introduced in the third reactor after the starting reagents have undergone sufficient mixing in the first two reactors. They demonstrated synthesis of 4-7nm particles of gold, silver and copper using this method. 2.2.2 Droplet Microfluidics Even though the single phase flow methods achieved better size control and distributions compared to batch methods, all the methods faced a common problem of severe reactor fouling. Since heterogeneous nucleation is more favored than homogeneous nucleation, the exposure of the reactants directly to the walls induced nucleation and growth on the walls, thereby clogging the channels severely and affecting the final product quality. Droplet microfluidics where isolation and better mixing of the reagents were possible was adopted widely for nanomaterials synthesis. Some of the prominent 24 demonstrations are described in this section. Gas-liquid segmented flow techniques are not highlighted here. I. Semiconductor nanoparticles One of the first demonstrations using droplets was for the synthesis of cadmium sulphide and cadmium selenide nanoparticles by Ismagilov and group 92. Using a PDMS based reactor they show the synthesis of CdS nanoparticles and also a multistep synthesis of CdS-CdSe core shell particles. The aqueous reagents were completely isolated from the channel and were mixed rapidly due to the winding channel present immediately downstream from the inlets. The mixing was calculated to be achieved in 5 ms after which the residence time of the droplets could be controlled by changing flow rates to give different particle sizes. Downstream near the outlet another inlet was provided where a quench stream could be injected into the droplets in a synchronized fashion to stop the reaction and control the further growth. Instead of a quench stream another reagent sodium selenide was introduced into the droplets which resulted in CdS-CdSe core-shell particles. The method resulted in no fouling of the reactor and they were able to produce quantum dots of different sizes in a highly controlled manner. The above method was at room temperature but CdSe nanoparticles with crystalline morphologies need high temperatures upto 250oC. Since PDMS cannot withstand such temperatures an easier tube based method was demonstrated by deMello and group93. Using a robust PTFE based capillary microreactor they synthesized CdSe nanoparticles in a droplet platform. The PTFE reactor was immersed in a oil bath and was maintained at 250 oC to facilitate the reaction. Using the same reactor setup they also demonstrated the synthesis of silver and titanium oxide nanoparticles at high temperatures. 25 II. Metal Oxides Iron oxide nanoparticles are used for applications like magnetic resonance imaging and for developing high density storage media. They are usually synthesized by coprecipitation of Fe+2 and Fe+3 salts in the presence of a base like ammonium hydroxide. The reaction is known to be rapid and using homogeneous flows in microreactors will cause immediate clogging of the channels. Using droplets can overcome the homogenization and fouling problem but using conventional droplet formation techniques will lead to clogging at the channel inlet where the reagents meet. Frenz and Griffiths developed a droplet microfluidic technique for synthesis of such iron oxide nanoparticles94. They used a PDMS based reactor with electrodes fused into them. They generated surfactant stabilized droplets containing the two reagents at different inlets. Downstream the two channels converge, but due to the presence of surfactants the two droplets don’t coalesce automatically. As they flow together, an electric field is applied and the two droplets coalesce (Fig 2.2b) and mix instantaneously to form iron nanoparticle precipitates. Electrocoalescence was thus effectively used to control the droplets and the ensuing reaction. III. Gold nanoparticles Anisotropic gold nanoparticles like nanorods are of much interest due to their applications in imaging and catalysis as we have seen earlier. Duraiswamy and Khan demonstrated a droplet microfluidic method for size and shape sensitive synthesis of such particles95. Using a PDMS based reactor they used the seed mediated growth method using CTAB as a surfactant for the synthesis of nanorods, spheres and other dog-bone type structures. The gold seeds are synthesized off-chip and were introduced as one of the reagents. The reagents meet the continuous phase at the T-junction and form droplets (Fig 2.2a). The shape (aspect ratios) of the gold nanorods could be 26 controlled on demand by varying the inlet silver-ion concentrations which acts as a shape influencing agent. Gold-Silver-Gold multiple shell type structures were recently synthesized using the droplet microfluidic platform using a capillary reactor by the Kohler group96. These shell type structures display highly enhanced optical properties compared to nanoparticles of the same size. Gold seeds in the size range of 10 nm were synthesized off-chip using sodium citrate as the reducing agent. These seeds along with silver solution and ascorbic acid were introduced into the reaction tube via a Tjunction. They formed droplets in the continuous phase where a silver shell was formed over the gold particle. Downstream they introduce another reagent (gold ions) in a synchronized manner into the flowing droplets. Here the gold ions mix into the Fig 2.2: (a) Schematic of reagent injection and droplet generation in the device used for gold nanorods synthesis and TEM images of gold nanorods of different aspect ratios obtained.95 (b) Photograph of the device used for iron oxide nanoparticle synthesis showing generation of droplets containing the two reagents and downstream coalescence by electric actuation. TEM image shows the iron oxide nanoparticles synthesized using this method. 94 existing mixture and reaction ensues to form gold-silver-gold type nanoshells. Using this method they demonstrated synthesis monodisperse populations of 50nm sized multi-shell structures 27 2.3 Microfluidics and Gold nanoshells With silica-gold nanoshells gaining prominence, microfluidic techniques were also adopted for synthesis of such hybrid nanoparticles. However unlike semiconductor or gold nanocrystals, not many demonstrations exist for nanoshells synthesis using this platform. The methods developed until now are discussed in detail in this section. The main theme of the thesis is to develop efficient methods that can provide control over product and can be eventually scaled-up. These are the factors that are kept in mind while reviewing the current methods. 2.3.1 Liquid phase reagents The first demonstration of gold silica nanoshells synthesis in a microfluidic platform was shown by Duraiswamy and Khan64. They used a liquid phase reducing agent (hydroxylamine hydrochloride) for the synthesis. Using a PDMS based chip reactor, with silicone oil as the continuous phase they generated plugs containing the aqueous reagents. The reducing agent and the gold plating solution containing the silica particles flow in separate streams and meet at the inlet junction. The reagents get rapidly homogenized due to the intense internal re-circulatory flows in droplets. Sharp particle population sizes with well defined gold covered silica particles were obtained using this method. But due to the extremely fast kinetics, as the reagents meet, instantaneous reaction caused fouling and blockage at the inlet over a period of time. The particles generated outside the droplets at the inlet eventually travelled downstream causing fouling of the entire channel. To overcome this problem, an inert gas like nitrogen was introduced into the channel. With careful manipulation of the gas liquid flow rates, they were able to generate regular gas liquid alternate flows (Fig 2.3a). Such three phase flows with ordered alteration of gas and liquid cells dispersed in the immiscible third phase are called foams. The injected gas periodically clears the 28 aqueous streams from the T-junction, thereby preventing reagent buildup and nanoparticle deposition. The technique enabled synthesis of particles with different gold coverages with very good control over sizes compared to conventional batch techniques. The only disadvantage in this method is the manipulation of the fluid properties and generation of stable and regular gas liquid flows; which needs expertise and cannot be done intuitively. The operational issues here with respect to generation of flow patterns and need for continuous monitoring highly affect the scalability of a sensitive system like this. Sebastian and Santamaria recently demonstrated a microfluidic method for synthesis of gold silica nanoshells97. To make process operation simpler and to facilitate scaleup, they performed the various steps involved in nanoshell synthesis from colloidal silica synthesis to gold shell growth, in the microreactor. To keep things easier they employed TEFLON tubing as the reaction tube instead of fabricated chips and used single phase, non-segmented flows for synthesis. In order to achieve proper mixing, the reagents are flown through a commercially available micromixer before entering the reaction tubes. In the micromixer, the reagents split into many streams at the inlet thereby achieving mixing by focussing of the streams at the outlet. The functionalized silica core particles were first synthesized in the reactor. The gold seed particles were prepared separately in batch. The seeding of the silica with the gold nanoparticles was however done in the microreactor. Finally using formaldehyde as the reducing agent, the gold shell growth step was conducted in the reactor. This offers a very easy method for nanoshell synthesis that can be adopted by anyone without expertise in microfluidics to obtain particles better than those obtained using batch methods. However the method faces the age old problem of reactor fouling. Especially the metallic micromixer walls will be subjected to severe deposition due to 29 heterogeneous nucleation. Also the use of single phase flows in micrometer sized tubing will have mass transfer and mixing limitations. The authors also report problems of colloidal gold synthesis i.e. nucleation and growth outside the silica surfaces, which will need post processing steps. In this method, moving away from the droplet platform gives great operational benefits, but not without compromises on the product quality. 2.3.2 Gaseous Reagents The use of liquid phase reagents for nanoshell synthesis in a microfluidic platform is limited to these two demonstrations. Gaseous reducing agents as previously mentioned are more stable and are easier to separate, but suffer from homogenization problems when done in the batch scale. A microfluidic demonstration for synthesis of gold nanoshells using carbon monoxide was done by Duraiswamy and Khan65. Fig 2.3: (a) Schematic of generation of foams in the microchannels for gold nanoshell synthesis using liquid phase reducing agent. Alongside is a TEM image of the gold nanoshells synthesized. 64 (b) Schematic of generation and merging of droplets and bubbles to form compound droplets for nanoshell synthesis using carbon monoxide. On the side is a TEM image of the nanoshells obtained using this method.65 Using a PDMS microfluidic device similar to their previous work described above, they demonstrated a three phase system where the reagent containing aqueous slugs 30 and the carbon monoxide (reducing agent) bubbles form compound droplets and flow in the microchannel. The droplets and the gas bubbles are formed in separate Tjunctions which merge together downstream in another T-junction (Fig 2.3b). The gas and the liquid compartments are separated by a thin film of oil. The carbon monoxide then diffuses through this oil film into the aqueous droplet where it gets homogenized by the re-circulatory flows. Unlike their previous method where the gas and liquid compartments are form at a single T-junction, here the reducing agent and the reagents are formed separately in order to prevent immediate contact and reaction thereby causing deposition at the T-junction. Monodisperse populations of gold nanoshells and nanoislands of varying coverages could be obtained in this method using a gaseous reducing agent. The main shortcoming here is again the complications involved in operating the system. The proper operation of the method depends on the synchronised formation of compound droplets. Operation of such a system would need constant monitoring and parallelization for scale would be extremely challenging considering the non linear nature of the dynamics of bubble-droplet systems. 2.4 The path ahead Microfluidics and its advantages, and how they were leveraged for nanomaterials synthesis were the focus areas of this chapter. The current state of the art as far as nanoshells synthesis using microfluidics was also discussed here. Now equipped with the knowledge of the technology available and the current challenges in implementing microfluidics for nanoshells synthesis, the next chapter introduces a new microfluidic method for using reactive gases for nanoshells synthesis to overcome these challenges. 31 Chapter 3 Microfluidics for nanomaterials synthesis using reactive gases The development of a continuous flow multi-phase microfluidic method for gold nanoshells synthesis using a reactive gas forms the crux of this chapter. Building upon the discussion in the previous chapter, a method that exploits the enhanced transport properties of droplet microfluidics and the benefits of a gaseous reducing agent and yet is easy to operate and eventually scale-up is presented here. Using a parallel channel configuration, the chip based method presented here, enables safe and easy integration of carbon monoxide a know reducing agent for metal ions for the synthesis of nanoshells. The exquisitely controlled gas-liquid contacting made possible by this method, enables tunable gas-liquid contact times, allowing precise regulation of the dosage and mixing of CO into the droplets and thereby the extent of the reaction without any need for reaction quench. Thus the tunability of the nanoshell morphologies is achieved not only by varying initial reagent concentrations but also by simply changing the residence times. Starting from the concept and development of the method, the chapter then focuses on operation and control, after which synthesis of gold nanoparticles and nanoshells is discussed and finally ends with a discussion on the transport phenomena and the salient features of the system. 3.1 Introduction In the previous chapter, gold-nanoshell synthesis using wet chemical batch methods and the need for microfluidic continuous flow techniques were discussed. The continuous flow methods for gold nanoparticle and gold nanoshell synthesis have the disadvantage of insufficient mixing, axial dispersion and RTD problems and most 32 importantly wall deposition, which ultimately leads to reactor clogging and breakdown. Droplet microfluidics with its compartmentalization and shear induced convective mixing, overcomes the above mentioned challenges and has been effectively used for synthesis of gold based nanoparticles. The use of reactive gases for nanomaterials synthesis has gained popularity58-61 owing to the inherent advantages of stability and ease of separation, however is challenging to implement for large scale synthesis involving sensitive and fast reactions. Even though droplet microfluidics has been preliminarily used for nanoshell synthesis using gaseous reagents, it comes with its own complexities in operation making scale-up challenging. Gaseous reagents have been previously integrated into microfluidic reactors especially for organic synthesis in many ways 98-100. Gas-liquid segmented flow has been widely popular due to the enhanced mixing and mass transfer found in such flows 101. An alternative way of introducing gases into the chip that has been recently used is the parallel flow design having adjacent gas and liquid reactant channels separated by membranes102, 103 . Such designs have been used for organic synthesis, creating concentration gradients along a microchannel104, oxygenation of blood in artificial lungs104, and deoxygenation for enhanced photo-stability in single-molecular study106. Instead of the dual channel design with the two channels placed one above the other, a parallel channel configuration with the channels adjacent to each other is used here 107. Such a design makes fabrication easier and also reduces the possibility of collapse of the structure during operation. In the chip, a thin gas permeable membrane physically separates a microchannel carrying gaseous CO from another channel carrying a segmented flow of monodisperse reagent-filled aqueous drops and immiscible oil (Figure 1). A poly(dimethylsiloxane) (PDMS) based microfluidic reactor is used in 33 Figure 3.1. Concept of membrane-based droplet microfluidic device for materials synthesis with a reactive gas. which a segmented train of droplets containing the aqueous reagents is first created at a T-junction, and then transits through a parallel-channel network situated downstream of the T-junction (Figure 2). CO flows through the parallel gas channel and it diffuses through the thin PDMS membrane (200 µm) separating both the channels and into the droplets. The CO that diffuses into each droplet gets rapidly homogenized due to intense internal convection within translating droplets 82. Due to the high permeability of PDMS to CO, it offers no significant resistance to mass transfer of CO in the system108, 109 3.2 Experimental Details Materials: 3-aminopropyl tris(trimethylsiloxy)silane (APTTS, 99%), hydrogen tetrachloroaurate(III) trihydrate (HAuCl4, 99.99%), tetrakis(hydroxymethyl) phosphonium chloride (THPC, 80% in water), sodium hydroxide (NaOH, Reagent grade, 97%), acetonitrile (CH3CN, HPLC grade 99.9%), potassium carbonate (K2CO3, 99.99%), Octadecafluorodecahydronaphthalene (Perfluorodecalin, Fluka, mixture of cis and trans, 95%) from Sigma-Aldrich Co. Ltd., Singapore, were all used 34 as obtained without any further purification. De-ionized water and glassware washed in aqua regia and rinsed thoroughly in water were used for all experiments. Synthesis of amino-functionalized silica spheres: 230 nm silica particles were synthesized by the Stöber method51. The surfaces of silica spheres were functionalized with APTTS in refluxing acetonitrile. Amino-functionalized silica particles were washed with acetonitrile, ethanol and DI water by sonication and several cycles of centrifugation.53 Gold-seeded silica particles: 2-3 nm colloidal gold particles used for seeding the amino-functionalized silica was made according to the procedure of Duff et al. 55 Under rapid stirring 1.5 mL of 0.2 mM aqueous NaOH and 1.0 mL of 67.7 mM aqueous THPC were sequentially added to 45.5 mL of DI water followed by addition of 2 mL of aqueous 25 mM HAuCl4. The solution immediately turned brown-red indicative of the colloidal gold formation. 150 µL of amino-functionalized silica particles suspension with 20ml of colloidal gold solution was shaken for 8 hours at room temperature and centrifuged to separate the gold-seeded silica particles. These particles were washed three times with water by cycles of centrifugation to remove the unattached colloidal gold particles.53, 54 K-Gold Solution: 3 mL of 1 wt% HAuCl4 was taken with 50 mg K2CO3 in 200 mL DI water (1.8mM) and aged for 1 day to produce a colorless gold plating solution, socalled K-gold, containing gold hydroxide ions.110 Microfabrication: Microfluidic device patterns were fabricated onto silicon wafers by standard photolithography using negative photoresist SU-8 2050.111 The silicon masters were subsequently used to mould polydimethylsiloxane (PDMS) replicas. PDMS base and curing agent (Sylgard 180 kit from Dow Corning) were mixed in the 35 10:1 ratio, degassed and poured over the silicon master. After ensuring that no trapped air bubbles are present in the poured mixture, it is left to cure at 70oC for 2 ½ hours. They mould is then removed and carefully peeled off the master. Holes are punched at the channel inlets to insert the tubings. The channels and inlets are carefully cleaned by scotch tape to ensure no dust is present. The PDMS mould is then bonded to a PDMS spin-coated glass slide using air plasma treatment. The device is then left untouched for a few hours to ensure proper bonding without air gaps. The device inlets are cleaned once again after which the device is plasma treated and the inlet PEEK tubings are inserted immediately. The inlet tubings were fixed in position by using a 5 minute Epoxy glue. The chip is then placed in the 100 oC oven for 24 hours. The final step helps in making the channel walls hydrophobic so as to ensure complete wetting by the continuous phase. The microchannels have rectangular crosssection and the details of the channel dimensions are: width and depth of the liquid and gas channel are 300 μm and ~124 μm and 100 μm and ~124 μm respectively, and the parallel liquid and gas channels are separated by 200 μm. Devices with lengths of the parallel channel section of 91.5 cm and 21.8 cm were used for the experiments. Microfluidic device setup and operation: Carbon monoxide is a highly toxic gas. All experiments are carried out in a wellventilated fume hood. A schematic of the microfluidic device and experimental setup is shown in Figure 3.2. Syringe pumps (Harvard, PHD 2000) were used to deliver perfluorodecalin (‘FO’) from three syringes and a mixture of gold-seeded silica particles and K-gold solution (‘AQ’) from another syringe to the microfluidic device. Carbon monoxide was delivered from a cylinder equipped with a pressure gauge through a PTFE tube leading into the on-chip gas-inlet. The reagent concentration was varied in terms of 36 the volume fraction of gold-seeded silica particles (‘fs’) in AQ (‘fs’ is the total volume of the silica spheres/total solution volume). The PDMS membrane was pre-saturated by flowing CO in the gas channel for at least 1 hour. For synthesizing gold nanoparticles, a device with the parallel channel section length of 91.5 cm was used. Operating at flow rates of AQ: 3 µL/min and total FO: 9 µL/min, the residence time of the reagents in the channel was 170 seconds. For the experiments with different goldseeded silica volume fractions in AQ (fs varied from ~ 2.8 x 10 -5 to 4 x 10-5), a device of length of 91.5 cm was used with a constant flow rate of AQ: 5 µL/min and total Figure 3.2. Schematic of the experimental setup. Insets: (i) stereomicroscopic image of aqueous (AQ) droplet formation in fluorinated oil (FO), (ii) aqueous droplets in the parallel channel network; carbon monoxide in the gas channel is dosed into aqueous droplets in liquid channel through the intervening PDMS membrane. Scale bars: 300 µm. 37 FO: 30 µL/min, corresponding to a residence time of 60 seconds. The residence times were calculated using the superficial velocity of the droplets in the channel and verified by high-speed microscopic observation. For the experiments with a fixed fs and varying residence times, a 21.8 cm long device was utilized. To obtain different residence times the flow rates used were: AQ: 3 µL/min and total FO: 4 µL/min; AQ: 20 µL/min and total FO: 150 µL/min; AQ: 20 µL/min and total FO: 200 µL/min; AQ: 30 µL/min and total FO: 260 µL/min. Thus residence times ranging from 70 seconds to ~2 seconds were achived. The carbon monoxide gas pressure was maintained at 10 psig. At the liquid outlet, aqueous colloidal solution was readily phase-separated from fluorinated oil. Carbon monoxide gas was vented out from the gas outlet. Sample collection and analysis The outlet from each microfluidic device was connected via fluoropolymer (FEP) tubing to a 2 mL centrifuge tube. Approximately 500 µL of the aqueous sample was collected in the tube for every experimental condition. Since the PFD is heavier than water the aqueous phase can be easily separated by simple decantation. The separated aqueous phase was then gently centrifuged to remove the remaining PFD. The purified sample was centrifuged and washed with water 2 to 3 times. Aqueous samples containing nanoparticles were analyzed with a UV-vis spectrometer (Shimadzu UV-2450). A drop of this sample was placed onto a 200 mesh formwar protected copper grid and allowed to dry overnight. The copper grid was then analyzed using TEM (JEOL 2010, accelerating voltage 200 kV). 3.3 Synthesis of gold nanoparticles The parallel channel device is first tested for the synthesis of colloidal gold particles by CO-assisted reduction of aqueous HAuCl4. Droplets of aqueous 0.3 mM HAuCl4 were generated in FO and transported in parallel to the CO-filled channel (CO 38 pressure = 10 psig). The residence time of the droplets in the reactor is around 3 minutes, thereby giving sufficient time for reduction of gold ions to form gold nanoparticles. The pink aqueous phase product at the outlet is then analyzed by UVVis spectroscopy and Transmission Electron Microscopy. Figure 3.3: UV-Vis absorbance spectrum of colloidal gold synthesized using the microfluidic device. Insets: TEM image of gold nanoparticles and measured particle size distribution. The UV-Vis spectrum of the products (Figure 3.3) shows an absorbance spectrum characteristic of colloidal gold. Highly monodisperse gold nanoparticles (Inset Figure 3.3) in the 10 nm range (measured from TEM micrographs) could be obtained from this method. No deposition of colloidal gold on the channel walls took place as the contents of the droplets were segregated from the wall by a thin oil film. Interestingly, in an experiment where a single phase of aqueous HAuCl4 solution was used instead of droplet flow in oil, colloidal gold was seen to immediately deposit on the walls of the liquid channel as the solution came in contact with 39 Figure 3.4: Deposition of colloidal gold on the wall of liquid microchannel running in parallel to CO channel. CO (Figure 3.4). This observation illustrates the superior operational flexibility of the droplet flow over single flow regime for solid nanomaterials synthesis. 3.4 Gold nanoshell synthesis For the synthesis of silica-gold nanoshells, a protocol established by Halas group for batch synthesis is adopted. 250(±20) nm pre-seeded silica particles suspended in gold salt solution (‘K-gold’ which acts as the Au+3 source) as the aqueous phase (AQ) and perfluorodecalin as the fluorinated oil FO is used and CO flow at 10 psig is maintained in the gas channel. The residence time of the droplets within the parallel channel section is about 60 seconds. CO reduces the gold ions within the aqueous droplets onto existing gold seeds on silica spheres. Tuning of plasmon resonances can be achieved stoichiometrically by varying the relative amounts of gold ions, reducing agent or gold-seeded silica particles in the aqueous solution, thus regulating the extent of gold coverage on silica. By varying the amount of gold ions relative to the number of silica particles (defined as the volume fraction fs of gold-seeded silica particles) in the aqueous feed to the microfluidic reactor, different gold coverages on the silica cores are obtained. For example, where gold ions are limiting, silica particles covered with either disperse or almost coalesced nanoislands (Figure 3.5 a & b) were controllably produced with 40 monodisperse gold coverage for both cases. By further lowering fs, silica particles with complete gold shell were obtained with high monodispersity (Figure 3.5c). The plasmon band of the core-shell particles with increasing gold coverage shifts progressively towards longer wavelengths (Figure 3.5d). Thus, monodisperse coreshell particles with tunable surface coverage and therefore tunable plasmon resonance were synthesized in this continuous platform. Figure 3.5: Gold-silica core-shell particles obtained by changing only the silica volume fraction in the reagents at a fixed residence time of 60 sec: TEM images of 230±20 nm silica particles with: (a) nanoislands (fs=4 x10-5 in 0.36 mM K-Gold), (b) almost coalesced nanoislands (fs= 3.4 x10-5 in 0.37 mM KGold). (c) complete nanoshell (fs= 2.8 x10-5 in 0.37 mM K-gold). (d) Ensemble UV-Vis spectra for all particles (a)-(c). In order to test the reliability of the platform the device was operated continuously over a six hour period, during which samples were collected at one hour intervals. The quality of the nanomaterial synthesized was examined by UV-Vis spectroscopy. Gratifyingly, all six samples showed overlapping absorbance spectra (Figure 3.6) 41 indicating identical surface coverage. In order to compare the current system with batch methods, nanoshells synthesis was conducted in batch (3 ml each) by bubbling CO gas into the reagents in a glass cuvette. The same experiment was repeated 5 times and products were analysed using UV-Vis spectroscopy. The results clearly show 5 different batches produce core-shell materials with significantly different coverage (evident from the varied absorption maxima between 750 to 850 nm). Some batches even showed presence of gold nanoparticles along with the nanoshells. This illustrates the challenges in reliability and reproducibility in conventional nanosynthesis, especially when a reactive gas-phase is involved. Figure 3.6: (a) Ensemble optical absorbance spectra for samples from 5 batch experiments using fs= 2.8 x10-5 in 0.37 mM K-gold (b) Ensemble optical absorbance spectra for 6 samples collected during 6 hours of continuous microfluidic synthesis using fs= 3.4 x10-5 in 0.37 mM K-Gold 3.5 Dynamical tunability of particle morphologies In the previous case, at all reagent concentrations , rapid microfluidic mass transport ensured that a residence time of ~60 s is sufficient to dose each droplet with enough CO to reduce all the existing Au3+ onto silica. Thus the final gold coverage, and therefore plasmon resonance, at nominally ‘high’ residence times is pre-determined by the initial [silica]/[Au3+] ratio in the aqueous feed. From the standpoint of ‘on demand’ nanoparticle engineering, tuning nanoparticle morphology just by varying 42 process parameters (i.e. flow rates and residence times) with feed solutions of fixed composition needed to be explored. Such processes based tuning, by operating in a CO-limited regime at shortened gas-liquid contact times well below 60 s by varying only droplet residence times and thereby regulating the extent of gold coating for fixed silica/[Au3+] ratio is presented here. For the experiments, the same batch of 250(±20) nm pre-seeded silica particles as in the previous section is used with the CO pressure maintained again at 10 psig. Short gas-liquid contact times in the 1-5 s range are used for these experiments by appropriately tuning the relative flow rates of the aqueous and oil phases to access high droplet speeds. Silica particle volume fraction fs = 2.8 x 10-5 and aqueous gold concentration ~0.37 mM are used, such that they are just sufficient to ensure complete coverage of all particles by a thin layer of gold at high residence times (> 10 s). The system yielded particles (Figure 3.7(b)-(d)) with tunable reduced fractional gold coverage ranging from ~0.65 – 0.85 (Figure 3.9(a), obtained by digital image analysis of TEM micrographs) with sub-second time resolution in the ~1-5 s range. Slightly higher silica volume fractions (fs = 3.1 x 10-5) at the same gold concentration yielded similar tunability of surface coverage, albeit with lower final coverage due to the gold limiting condition in this case (Figure 3.8 & Figure 3.9(a)). The UV-Vis signatures of the particles obtained are shown in Figure 3.7(e) & 3.9(b), and as seen in the previous section the absorbance peaks red shift with increasing coverage. These results highlight the remarkable tunability of plasmon resonances over more than 300 nm (starting from the 520 nm plasmon peak of the seeded silica) with small changes in gas-liquid contact time in the 1-5 s range (Figure 3.7(a) & 3.9(b)). Only a trace amount (if any) of free colloidal gold (verified by UV-Vis and TEM characterization) was observed in the reaction mixture. The absence of free colloidal gold could be due 43 Figure 3.7: Gold-coated silica particles obtained by changing droplet residence times at fixed gold and silica concentrations. TEM images of 230±20 nm silica particles (fs=2.8x10-5 in 0.37 mM K-Gold) with: (a) pre-attached gold seeds, and (b)-(d) varying degrees of gold coverage. (e) Ensemble UV-Vis spectra for all particles (a)-(d) to the controlled administration of reducing reagent, the thermodynamic favorability of heterogeneous nucleation over homogeneous nucleation and the rapid autocatalytic nature of nanoshell synthesis60. Thus, particles produced in this process are 44 monodisperse, with uniform gold coverage and do not require any further cleaning or purification steps. Figure 3.8: Gold-silica core-shell particles obtained by changing the residence times with a fixed fs=3.1 x10-5 in 0.37 mM K-Gold : TEM images of 230±20 nm silica particles with: (a) sparse nanoislands, (b) nanoislands, (c) almost coalesced nanoislands (d) complete nanoshell. The ability to exert temporal control over the gas-liquid contacting by controlling the residence time of the droplets, allowing careful regulation of the amount of CO infused into the aqueous droplets is due to the inherently accelerated mass transfer in such microscale systems – a capability that is simply unrealizable at the macroscale. The next section discusses the salient aspects of mass transfer and mixing in such systems. 3.6 Gas-Liquid mass transfer The remarkable tunability of gold coverage arises from the ability to precisely dose and mix limited amounts of CO into the aqueous droplets at short residence times. 45 Figure 3.9: (a) Plot showing fractional coverage (Fc) of gold over the silica cores of particles (measured by digital image analysis of TEM micrographs in MATLABTM) synthesized using two different but fixed silica particle concentrations (fs=2.8 x10-5 and fs=3.1 x10-5 in 0.37 mM K-Gold). (b) Plot of absorbance maxima in the UV-vis spectra against residence time for particles synthesized using two different but fixed concentrations (fs=2.8 x10-5 and fs=3.1 x10-5 in 0.37 mM K-Gold) of silica particles. The synthesis of metal-dielectric core shell nanoparticles such as gold/silver-silica nanoshells is carried out using electroless plating which involves the autocatalytic deposition of gold on colloidal silica pre-seeded with gold nanoparticles. Such electroless plating reactions are known to be extremely fast, making the process mass transfer controlled. Since the membrane is thin and highly permeable to gas, it offers no significant resistance to the mass transfer of CO. The membrane is pre-saturated by flowing CO in the gas channel for at least 1 hour in our experiments. Thus, the transport of CO from the membrane wall to the droplets represents the rate-limiting step for the growth of gold seeds on the silica particles. To support this assumption a comparison of the membrane mass transfer and the liquid phase mass transfer resistances is provided. Under steady state conditions the flux across the membrane and from the wall to liquid phase will be equal. A simple mass transfer model can be written to determine the flux (mol/m2-sec). Membrane side flux: 46 N P  PGAS  Pi  z ...... (1) Since the kinetics is rapid, assuming instant consumption of CO within droplets, liquid side flux is given as N  kL  CS  0  ...... (2) Where P is the permeability of CO through PDMS, PGAS is the pressure in the gas channel (10 psig), Pi is the pressure at the solid-liquid interface and CS is the solubility of CO in the aqueous phase. A relation between the two can be given using Henry’s law Pi  CS H ...... (3) Henry’s law constant (H) for carbon monoxide and water is 9.9 x 10-6 mol/m3-Pa.112 Adding equations 1 and 2 and using relation 3 we get  1 z N    PGAS  kL H P  N PGAS 1 z  kL H P ...... (4) This equation is of the form Flux= Driving force/Resistance Thus the total resistance can be written as the sum of the membrane and liquid phase resistances RTOTAL  1 z  kL H P 47 The permeability of CO through PDMS has been reported to be in the range of 400 barrer.113, 114 In SI units this is equal to 1.34 x 10-13 mol-m/m2-s-pa. Typical volumetric wall-to-liquid mass transfer coefficients (kLa) are tremendously accelerated in such microscale segmented flows, and are in the range of ~ 0.1 s -1.115 Here a is the area per unit volume available for mass transfer. Using a for the specific geometry of our system, we obtain kL= 3.93 x 10-6 m/s. Using these values the two individual resistances (m2-s-pa/mol) are determined. Here, the liquid phase resistance is found to be an order of magnitude higher than that of the membrane. Therefore, the transport of CO from the membrane wall to the droplets can be assumed to represent the rate-limiting step for the growth of gold seeds on the silica particles. Mixing is extremely rapid within the aqueous droplets due to intense internal convection; the mixing time tm (~ 0.1 s) decreases with increasing droplet speed116 and is smaller by at least an order of magnitude compared with the residence times used. It is also smaller by at least two orders of magnitude when compared to the characteristic time for CO homogenization via pure diffusion (tD ~ w2/DCO = 45 s, where w = 300 m is the width of microchannel and DCO = 2 x 10-9 m2/s is the aqueous diffusivity of CO) – a fact that highlights the advantage of using moving microscale droplets for multiphase synthesis. Thus the CO that enters the droplets from the wall gets rapidly homogenized and is consumed extremely fast in the ensuing reaction. Thus a simple mass transfer model that assumes gas-saturated walls and instantaneous mixing and reaction of CO within the aqueous drops, predicts that the number of moles of CO dosed into each droplet varies linearly with time as NCO = kLaCsVdt ~ 2.5 48 x 10-12t, where Vd is the volume of each droplet (~ 25 nL), Cs is the solubility of CO in water (~ 1 mM)117 and kLa is a mass transfer coefficient (~0.1 s-1). Now, since at least 14 x 10-12 moles of CO are required for complete reduction of ~9 x 10 -12 moles of Au3+ in each droplet (at ~0.37 mM), it can be inferred that CO is the limiting reagent in the 1 - 5 s range. Furthermore, the uniformity of gold coverage on silica highlights the rapid and complete mixing of CO within moving microfluidic droplets at residence times that are much smaller than characteristic diffusion time tD (~45 s). However these are order-of-magnitude estimates, and represent a lower bound on the amount of CO entering the droplets. CO can also enter the aqueous droplets via the oil phase; however, since this involves a two-step mass transfer process (wall-to-oil and oil-to-aqueous), a reasonable assumption is made that the dynamics are slower than the direct wall-to-aqueous transfer of CO. 3.7 Overview A continuous microfluidic method for nanomaterials synthesis using reactive gases was presented in this chapter. Exquisite temporal control of gas-liquid contacting and mass transfer was enabled by a hybrid droplet-based microfluidic platform. Such controlled gas-liquid contacting and gas dosage enables dynamic tuning of the morphology (and plasmon resonance) of the nanomaterials simply by varying the residence times of the droplets in the microchannel. Highly mono-disperse populations of silica-gold nanoshells and nano-islands were synthesized using this system with excellent control over morphologies and thereby the optical properties. Such truly ‘on-demand’ nanoparticle engineering is nearly impossible to realize in conventional gas-liquid batch chemistry or in solution-based continuous processes, and is the first demonstration of its kind. Using liquid phase reducing agents in continuous as well as batch systems, would require a rapid reaction quench to 49 terminate the progress of the plating reaction at such short residence times, and would necessitate, for example, rapid introduction and mixing of a large amount of quench fluid into the reactor. This is a very challenging requirement that almost completely precludes operation in the short contact time regime, and therefore cannot enable simple, robust and tunable nanoparticle engineering by varying operating parameters such as flow rates. This platform can also be extended to other contacting schemes such as iterative gas dosing at regular intervals or sequential dosing of multiple reactive gases into the reaction mixture for aiding multi-step materials or organic synthesis. Having successfully developed a microfluidic platform for facile nanoshell synthesis, the next chapter focuses on one of the main themes of the thesis: scale-up and how it can be achieved by building upon the current method. 50 Chapter 4 Scaling-up nanomaterials synthesis The previous two chapters discussed the development of microfluidic methods for the synthesis of speciality plasmonic nanomaterials. With a robust and easy microfluidic method that allows facile control over particle synthesis now developed, the next step is the scale-up of the process to achieve higher throughputs that in the future can achieve commercial scale production. Scale up is one of the main objectives of the thesis and the work done in this chapter aims to achieve this and includes some discussion on the finer aspects of scale-up. Moving away from the conventionally employed parallelization and numbering up of microfluidics reactors, the chapter will introduce an alternative and easier way where just an order of magnitude increase in the length scales brings about increased productivity and tremendous improvements in the economics and ease of operation. 4.1 Introduction Nanotechnology today is a multi-billion dollar industry118, 119 with more money and resources being pumped in for nanomaterials based fundamental and applied research. Having garnered interest for over a hundred years, with advancement in techniques for synthesis and characterization, the last two decades have seen a great leap in the state of the art of nanoscience. Plasmonic and semiconductor nanomaterials especially have been the forerunners, gaining much attention with exciting discoveries leading to applications ranging from sensing and biological assays to cancer treatment, bio-imaging, drug delivery and solar energy harvesting 17,31 . With the demand for nanomaterials with highly precise properties increasing, for fundamental 51 and applied studies, the last decade saw microfluidics fill in the gap where conventional methods with their inconsistencies failed. With successful demonstrations of continuous and reproducible production of a variety of nanomaterials the lab on a chip platform was a game changer for materials synthesis. Droplet microfluidics in particular gained attention, owing to the enhanced transport and mixing within droplets and the absence of wall deposition, for careful metallic and semiconductor NPs synthesis usually involving complex and fast chemistries However, today even as nanomaterials based research attains maturity, the commercialization and hence the outreach of such discoveries has been highly limited120. One of the major challenges that the field faces in translating the findings and laboratory demonstrations into applied technologies in the real world, is the large scale, reproducible, cost effective and sustainable synthesis of such nanomaterials 120, 121 . Applications for example of plasmonic nanoparticles in next generation photovoltaic cells48, high efficiency solar water heaters49 and photo catalysts122 that have been recently demonstrated will need larger volumes of nanoparticles. The commercially available nanoparticles are produced using batch methods where scaleup becomes highly challenging for difficult chemistries. Due to the economies of scale, the difficulty in quality control during scale-up and the niche demand until now, the costs involved for such syntheses is usually large thereby escalating the prices. Even though microfluidics has been a viable platform, it is limited by the inherently small production rates restricted to microscale volumes and the difficulty of fabrication and operation. A successful scale-up demonstration using batch methods has been for gold nanorods by the groups in UIUC and Rice University123, 124 and has been commercialized and is sold by a start-up company named Nanopartz. However materials involving very fast kinetics and harsh reaction conditions have not been 52 mass produced and the supply is highly limited. With the lack of scaled-up processes that are commercially viable microfluidics faces the challenge to be a relevant technology. The future of nanotechnology and that of microfluidics for synthesis of nanomaterials lies in the development of high throughput, scaled up methods that achieve the same level of control over size and morphology. 4.2 Scale-up in Microfluidics The traditionally adopted technique for scale-up in microfluidics has been parallelization or pile up of microreactors. This usually involves the operation of 100s or 1000s of microreactors simultaneously or stacked together in the form of modules with flow splitters to increase throughputs. Multiplexing or parallel processing has helped develop technologies for Point of Care devices that perform bio-assays 125-127 and PCR128 and provide detection of multiple targets with high resolution of results with very low consumption of the sample129. Fluidigm from Stanford University has been one of the pioneers in making this kind of technology commercially available. Numbering-up has been widely used for increasing throughputs in chemical synthesis that involve harsh reaction conditions or sensitive reactions for pharmaceutical applications. Stacked-up silicon or metal based chip microreactors have been demonstrated for liquid phase reactions such as amide formation hazardous reactions such as fluorination131 and ozonolysis 132 130 and gas-liquid . Such reactor modules have also been demonstrated for pharmaceutical synthesis involving organometallic reactions that have been adopted by the industry.133 Numbered-up capillary microreactor systems have similarly been demonstrated for reactions such as polymerization involving highly reactive species. 134 One of the few commercial ventures to come out of this technology for chemical transformation is Velocys. They use thousands of microchannels packed into modules for carrying out gas to liquid 53 Fischer Tropsch synthesis.135 With a compact physical footprint, these reactors have been deployed in ships to get the factory to the natural gas source rather than piping the gas to the refineries. Droplet microfluidics has been the most reliable platform for nanomaterials synthesis with excellent reproducibility as discussed in the previous chapters. However parallelization for droplet based reactors can be challenging. The simple parallel channel based gas-liquid contacting method is easy to operate with a single chip, but when 10s or 100s of such chips are put together for parallel operation, the complexities involved are enormous. Droplet generation using pumps individually for each chip is not economical. High frequency droplet generating mechanisms have been demonstrated, 136-138 however careful routing of such huge libraries to 100s of reactors can be highly challenging due to the highly non-linear and stochastic behaviour of droplets in networks. Even though active methods for controlling droplet flows in networks for example using electric and magnetic actuation have been developed, 139-141 they are not economical and easy to operate. Gas-liquid flow distributors have been demonstrated for parallelized microreactors with upto 60 reaction channels in each chip, 142 however they are limited by the difficulty in fabrication, setup and operation. For single phase chemical synthesis considerable increase in throughputs can be attained by parallelization. But in speciality nanomaterials synthesis the amount of actual solids processed a.k.a solid loading per unit volume is usually very less. For example in gold nanoshell synthesis this loading per unit volume is 10-5 gm/ml. The particle loading can be increased stoichiometrically with the precursor concentrations, but it has its limitations. Increase in the particle volume fraction increases the chance of shear driven orthokinetic aggregation143 as the particles are in constant motion within moving droplets. Particle 54 laden droplet flows have also been shown to demonstrate sedimentation and aggregation. The particles within the droplets are suspended due to vortices generated in such flows. Particles have been observed to settle at the stagnation zones in the droplet rear144. With increasing particle concentrations at the rear droplet cap, the local viscosity increases thereby causing more stagnation and capturing of more particles. At high concentrations (109 to 1010 particles/ml), even nanoparticles (with low settling velocities) have been found to get segregated at the rear cap. At the cap region, due to the extremely high particle volume fraction, DLVO type aggregation sets in; this ultimately affects product quality. Thus processing capacity of the actual products is limited. Hence while using microchannel reactors with nanolitre droplets the number of reactors for high product throughput needs to be large. Scale out of microfluidic methods for nanomaterials synthesis- droplet based or single phase, by far has not been reported. 4.3 “Milli-fluidics” Millifluidics in contrast to microfluidics consists of flows confined in channels with the characteristic dimensions in the millimetre range. However this mere order of magnitude increase in the length scales brings about great advantages especially for nanoparticle syntheses where the actual solid product processability is limited. The origin of multiphase flows and reactions in millimetre sized channels can be traced back to the 1970s when catalytic converters were introduced to reduce NO X and CO emissions in cars. Such reactors due their shape and structure were called honeycombs or monoliths. Initially designed to be used for gaseous reactions, they later found use in gas-liquid heterogeneous catalytic reactions.145-147 Starting from the 1980s monolithic reactors were extensively studied and used for improving mass transfer in multi-phase reaction systems. The hydrodynamics and transport phenomena of 55 multiphase flows in monoliths have ever since been exhaustively studied and documented. 148, 149 The Taylor or the slug flow regime where gas bubbles of lengths greater than that of the tube diameter move along the capillary separated by liquid segments was widely popular due to the enhanced transport and mixing characteristics.150, 151 Taylor flows in monoliths have been demonstrated to improve greatly, numerous commercially relevant gas-liquid reactions and have been adopted by industries like AkzoNoble,152 Air Products153 and Corning Inc.154 In the past decade the use of single and multiphase capillary based millifluidic systems for chemical and materials synthesis and discovery, and intensification and improvement of existing processes has gained considerable attention. Much work has been done on studying such systems for liquid-liquid extractions and chemical synthesis. Modular flow distributors for multiphase gas-liquid flows have recently been developed where Taylor bubbles streams are generated and uniformly distributed to 8 parallel reactors for increasing throughputs.155, 156 A system of 8-10 reactors, with increase in the holding volume of millifluidic capillaries, can attain processing rates upto kilograms/hr. 155 Capillary based segmented flow reactors for scale-up of nanomaterials synthesis have started to gain attention in the community with few existing demonstrations. High throughput precipitation based synthesis of BaTiO3 and ZnO nanoparticles was shown by Aimable and group 157 in EPFL. Using a robust tubular reactor, liquid- liquid segmented flows were used for high temperature synthesis of such particles. Production rates upto 50 gm/hr of the solid product was possible using the tubular reactor. For speciality nanomaterials synthesis there weren’t any reports for scale-up until very recently. Using a droplet based tubular reactor system Krishnadasan and deMello158 showed high throughput synthesis of quantum dots CdSe nanoparticles. 56 Using a simple flow distribution system they modestly scale-out their system to 5 tubular (PTFE) reactors and were able to produce upto 145g/day of CdTe nanocrystals. 4.4 Concept and development As discussed in the previous sections, capillary based segmented flow reactors have gained popularity and have shown to exhibit excellent mixing and enhanced mass transport properties. Taking cue from this platform, a capillary tube based “millifluidic” method is employed for scaling-up synthesis of complex nanomaterials, moving away from the conventional parallelization techniques used in microfluidics. The synthesis of plasmonic silica-gold nanoshells and nano-islands introduced in the previous chapters is shown using this platform. In order to retain the advantages of using a reactive gas, the same droplet microfluidic based synthesis method using carbon monoxide, previously demonstrated in a chip is ported into the capillary platform. In order to introduce a gaseous reagent into the tube based platform, a tubein-tube parallel flow configuration [Fig 4.1 & 4.2] is used. Here a highly gas permeable polymer tubing is positioned inside another tubing which is pressurised with a gaseous reagent. The gas then permeates through the inner tubing into the liquid reagents flowing within. Such tube in tube configurations have been successfully demonstrated for Hydrogenation, Carbonylation and other such gasliquid reactions by Steven V Ley’s group159, 160 at Cambridge University. Such reactors enable simplified and safe gas delivery into the system. 57 Figure 4.1: Concept of tube-in-tube based droplet milli-fluidic reactor for materials synthesis with a reactive gas. A biphasic liquid-liquid segmented flow forms the main part of our system similar to that in our microfluidic method. The aqueous droplets are generated in the fluorinated oil which acts as the continuous phase and the droplet train flows through a highly permeable PTFE (TEFLON AF 2400) tubing. Since the continuous phase (FO) and the PTFE tubing are essentially fluorocarbons, the wetting properties are found to be excellent. The highly hydrophobic nature of TEFLON further makes the generation of water in oil slugs easier. The TEFLON tubing forms the inner part of the tube in tube setup with a stainless steel tube acting as the outer enclosure. The carbon monoxide gas is flown through the outer tube and is maintained at constant pressure. The carbon monoxide then permeates through the Teflon membrane (300um thick) into the biphasic system. By pre-saturating the membrane it is ensured that it offers little or no resistance to the mass transfer into the system. The gas enters the droplets through two routes: 1. From the wall through the thin liquid film into the droplets 2. Into the droplets via the liquid slugs. The gas that enters the droplets is quickly homogenised due to the enhanced convective mixing found in such liquid-liquid slugs even in these length scales and is consumed in the ensuing reaction. 58 4.5 Experimental Details Materials: Silica particles, THPC gold seeds, seeded silica particles and K-Gold solution were all prepared as described in the previous chapter. PTFE TEFLON tubing was purchased from OmniFit UK. Capillary reactor setup: A schematic of the experimental setup is given in Fig 4.2. PTFE (TEFLON AF 2400) tubing of 1000 micrometer ID and the standard 1/16 inch OD is used for all the experiments. The stainless steel outer tubing of 0.6 m length was fabricated in NUS. The inlet and outlet of the SS tube were tightly fit with Swagelok fittings with the PTFE tubing fitted in place using appropriate ferrules. The CO gas cylinder equipped with a pressure gauge was fitted with an additional Swagelok needle valve; the outlet of which was connected to a 100 micrometer PEEK tubing of about 1 meter in length. PTFE tubings of 250 um were used to connect the PEEK tubing to the gas inlet of the SS tube. Similarly the gas outlet was connected to a 250 um PTFE tube. Again the gas inlet and outlet, the PTFE tubing in place, were tightly fit with Swagelok fittings. Syringe pumps (Harvard, PHD 2000) were used to deliver perfluorodecalin (‘FO’) from two syringes and a mixture of gold-seeded silica particles and K-gold solution (‘AQ’) from another syringe to the reactor. UPCHURCH fittings were used to connect the inlet tubings from the syringes to the reactor. A commonly available cross flow upchurch fitting was used as the droplet generator at the reactor inlet. 59 Figure 4.2: Schematic of the experimental setup consisting of infusion pumps, cross-flow droplet generator, stainless steel (SS) outer and PTFE inner tubes. The outer tube is pressurised with carbon monoxide. The whole setup is within a fume hood. Operation: The Teflon tubing reactor is initially filled with the fluorinated oil. The outer SS tubing is then pressurised with the carbon monoxide gas and maintained at 40 PSI. Flushing of the reactor is continued for 3 hours allowing enough time for CO to permeate through the membrane and saturate it. After flushing with CO, the aqueous reagent infusion is started. The reagent concentration is varied in terms of the volume fraction of gold-seeded silica particles (‘fs’) in AQ (‘fs’ is the total volume of the silica spheres/total solution volume). At the inlet the fluorinated oil is operated at a constant total flow rate of 200 µL/min for all experiments. To study the effect of droplet velocity and morphology, for a fixed fs, the flow rate of the aqueous stream is varied from 50 to 250 µL/min. At the outlet of the tube, the aqueous phase product can easily be separated from the continuous phase as described earlier and is then analyzed using a UV-Vis spectrophotometer and characterized using a TEM. Droplet generation: Owing to the excellent wetting of PFD over the tubing walls facile droplet generation was possible. The droplets generated were separated from the wall by a thin film of 60 the fluorinated oil. In order to keep the system simple a commonly available Upchurch cross flow fitting was used as the droplet generator. Droplet generation in cross-flow configuration occurs via flow focussing. The aqueous phase flowing through the cross is squeezed by the flow of the continuous phase from the other two sides leading to pinching and breakup. Using this simple system, highly monodisperse and stable trains of slugs could be generated. Since the experiments were carried out in the fume hood imaging the flow was not possible. Flow visualization using the same aqueous reagents and PFD were carried out outside the fume hood to observe and study the flow characteristics (Figure 4.4). 4.6 Results and Analysis The synthesis of gold nanoshells and nano-islands were carried out using a procedure similar to that of the chip method shown in the previous chapter. 250(±20) nm preseeded silica particles suspended in gold salt solution as the aqueous phase (AQ) and perfluorodecalin as the fluorinated oil FO (Scheme 1 and Experimental section) is used and CO flow at 40 psig is maintained in the outer tube. To test the capability of the PTFE based capillary system to synthesize gold-silica nanoshells with tunability over morphology, experiments with systematic variations in the reagent concentrations with all other parameters kept constant were conducted. By varying the volume fraction fs (%) of gold-seeded silica particles within the droplets, with different gold coverages on the silica cores were obtained as expected. Operating where the gold ions are limiting, silica particles covered with either disperse or almost coalesced nanoislands were controllably produced (Figure 4.3 (a) and (b)). When the gold ions concentrations were maintained in slight excess, completely coated nanoshells with a thin gold layer over silica were obtained (Figure 4. 3(c)). Nanoparticle populations with monodisperse gold coverages were obtained in 61 these experiments. The UV-Vis spectra of the particles obtained is show in Fig 4.3(d). The products obtained contained only traces of colloidal gold which is apparent from the TEM images and the absence of any prominent signature of colloidal gold in the UV-Vis spectra. In these experiments, the flow rates were maintained constant such that the droplets spend about a minute in the gas-liquid contacting section. As previously discussed, considering the FIG 4.3: Gold-silica core-shell particles obtained by changing only the silica volume fraction in the reagents at a fixed residence time of ~ 1 minute: TEM images of 230±20 nm silica particles with: (a) nano-islands (fs=2.265 x10-5 in 0.37mM K-gold), (b) almost coalesced nanoislands (fs= 1.7x10-5 in 0.37mM K-gold). (c) complete nanoshell (fs= 1.13 x10-5 in 0.37mM K-gold). (d) Ensemble UV-Vis spectra for all particles (a)-(c). extremely fast kinetics of such electron transfer reactions, the residence time provided was sufficient for complete growth of shell over the silica particles. Thus the variation 62 in the coverages of gold over silica was merely due to changes in the silica to [Au +3] ratios. Using a 0.6 m length reactor and operating at about 250 ul/min of the aqueous phase, the volume processed using this system (15 ml/hr) is about 25 times higher than that of the previously shown chip based method. Duraiswamy and Khan 64 using the foam based chip system could operate upto 0.5 ml/hr. Sebastian and Santamaria 97 using the tube based single phase flow setup were able to process approximately 5 ml/hr of the particle suspensions. The particle loading in all the demonstrations mentioned above was of the order of 10-5 gm/ml. In the last two methods, liquid phase reducing agents were used, thus the final product obtained would need energy intensive centrifugation based cleanup in order to remove the excessive or unreacted reagents. The presence of reducing agents like formaldehyde in the product solution has been reported to cause degradation of the nanoshells suspensions. Due to the excellent wetting of PFD over the PTFE walls, no fouling was observed in the tubing after repeated usage for over a month. Due with the structural stability of PTFE and absence of wall deposition, the reactor system could be used for months in a row without any need for replacement of the tubing. 4.6.1 Effect of droplet morphology and flow velocity on product quality In the capillary system, droplet generation using the cross flow fitting setup as described in the previous section was stable and reproducible. Operating at different aqueous phase flow rates with the continuous phase kept constant, droplet trains with different morphologies were obtained. For nanoshell synthesis, keeping the concentration of the reagents at the inlet the same, varying the droplet shape and velocity was found to affect the end products obtained. 63 In these experiments the aqueous phase reagent flow rates was varied from 50 µL/min to 250 µL/min with the continuous phase flow rates kept constant throughout at 200 µL/min. Runs were carried out using two different, but fixed silica to gold ion ratios. At low velocities almost spherical droplets with diameters equal to that of the tube were obtained. With increasing aqueous flow rates, the length to diameter ratio (L/D) of the droplets increased (Figure 4.4). At the highest flow rates slugs of about 2.5 times the tube diameter were generated. Figure 4.4: Microscopic images of droplet formation using the cross flow fitting. (a)-(d) show increasing droplet lengths with increasing aqueous flow rates keeping oil flow rates constant. TEM micrographs of nanoparticle populations obtained for fs= 1.13 x10-5 in 0.37mM K-gold where the gold ions are slightly in excess are shown in Figure 4.5 (a)-(d). At low flow velocities the coverage of gold over the silica particles is highly nonuniform. Free colloidal gold formed outside the silica particles was prominent here. It can be seen that as the flow velocity increases the uniformity of gold coverage over silica improves with the best results obtained at the highest tested flow rates. The samples were analyzed using UV-Vis absorbance spectroscopy the results of which 64 can be seen in Figure 4.5e. The UV-Vis spectra show a strong signature of colloidal gold at 520-530nm for products obtained at the lowest flow rates. The colloidal gold signature gets weaker with increasing flow rates and almost disappears at the highest flow rates. The plasmon band of gold nanoshells can be seen to red shift with increasing flow rates which indicates increasing coverage of gold over silica to form complete/smooth nanoshells. Figure 4.5: Gold-coated silica particles obtained by changing droplet velocity(v) and residence times(τ)( at fixed gold and silica concentrations. TEM images of 230±20 nm silica particles (fs= 1.13 x10-5 in 0.37mM K-gold): (a) v=5.3 mm/sec & τ =113 sec (b) v=7.4 mm/sec & τ =81 sec (c) v=8.4 mm/sec & τ = 71 sec (d) v=9.5 mm/sec & τ =63 sec 65 When the experiment is carried out with concentrations such that the gold ions are highly in excess, a similar trend in coverage of gold over silica and colloidal gold formation can be observed with increasing flow velocities (Figure 4.6). Here the amount of colloidal gold formed outside is higher and at the highest velocities particles with thicker shells are obtained. Figure 4.6: Gold-coated silica particles obtained by changing droplet velocity(v) and residence times(τ)( at fixed gold and silica concentrations. TEM images of 230±20 nm silica particles (fs=0.85x10-5 in 0.37 mM K-Gold) ): (a) v=5.3 mm/sec & τ =113 sec (b) v=7.4 mm/sec & τ =81 sec (c) v=8.4 mm/sec & τ = 71 sec (d) v=9.5 mm/sec & τ =63 sec. 66 In these experiments, residence time is one of the parameters that is not maintained constant while changing droplet sizes and shapes due to current experimental limitations such as fixed reactor length. However even at the highest flow rates the residence time available is about 1 minute, which as discussed previously should be sufficient for complete reaction. Hence the variations in the products can be safely assumed to be unaffected by the different residence times for different conditions. In microscale flows as demonstrated previously, the rapid mixing and homogenization ensures that CO concentration in the entire slug is uniform thus ensuring uniform growth over all the silica particles. The rapid mixing induced in liquid slugs in microchannels is due to the shear between the moving slug and the wall; shorter the width, higher is the confinement and hence greater is the mixing. 161, 162 In millimetre length scales the confinement is reduced and hence the mixing may not be as enhanced as in microchannels for the same velocities. The increase in the length scales also means increased diffusion times for homogenization. The mixing time in slugs flowing in millimetre sized channels have been determined to be faster with increasing flow velocity. 163, 164 The internal recirculation patterns within droplets have also been found to be dependent on the droplet shapes, with slugs experiencing better mixing rates compared to spherical droplets.162, 163 Kashid and Agar 164 used CFD simulations to determine slug size and velocity effects on mixing. For organic and aqueous phase droplets with wall film, for a fixed droplet shape the mixing times varied between 3 to 4 (expressed in terms of time required to traverse a distance equal to how many times its own length before complete mixing is achieved) when the flow velocities were varied between 1mm/sec to 200 mm/sec. This translates to mixing times ranging from 10 sec to 10-2 sec with changing velocities for the parameter space explored. This shows the strong dependency of mixing times on flow 67 velocity. The mixing times for different droplet shapes were also studied starting from L/D 2. At low flow velocities, longer droplets were found to have faster mixing compared to spherical droplets with mixing times reducing with increasing velocities. The increase in mixing time with decrease in droplet aspect ratios was also reported by Tanthapanichakoon et.al.163 who used CFD simulations to determine mixing rates in slug flows. Using the plots from Kashid and Agar, an approximate value of the mixing time in the system used is found to be ~0.8 to 1 sec for velocities of 10 mm/sec with mixing times increasing with lowering velocities and droplet aspect ratio. The combination of changing droplet shape and increasing velocity can thus be attributed to affect the mixing rates and hence the final product quality in such mass transfer controlled systems. Estimation of mixing patterns and time scales in different droplet shapes and velocities with control over residence times and concentrations is necessary for further careful analysis of the interplay of these factors but is currently beyond the scope of this thesis due to experimental limitations [Refer Future Work Section]. However, these findings provide valuable insight into the parameters needed to be studied for optimizing production rates and quality, and further elucidate, how crucial, rapid mixing and mass transfer become while scaling-up such syntheses and the advantages of using droplet based systems. 4.7 Conclusion In this chapter a tube based “millifluidic” system is developed and demonstrated for the scaled-out synthesis of complex-speciality nanomaterials: silica-gold nanoshells. The novelty of this platform arises from its simplicity; using commonly available inexpensive components: SS tubes, Swagelok & Upchurch fittings and PTFE tubing, a highly robust system is developed which can be easily operated. Highly 68 monodisperse populations of nanoshells and nanoislands could be synthesized using this method. Using a single tubing of less than a meter in length, upto 25 times increase in throughputs compared to previously demonstrated continuous synthesis methods is achieved. By increasing the reactor length and hence residence time, the volumetric processing rate can be easily increased further with only a single reactor. However, with such great operational benefits, a modest scale-out of the current system to include 8-10 reactors running in parallel can be achieved in a facile manner enabling production rates of litres/day. 69 Epilogue With the boundaries of nanotechnology being pushed every day, the science stands at a place where new developments are not being pushed into the real world as applied technologies due to limitations in existing production methods of such complex materials. The development of scalable microfluidic techniques for controlled nanomaterials synthesis is imperative for the progress of nanotechnology as an industry. This thesis aimed to bring the exisisting microfluidic technologies for complex nanomaterials synthesis a step closer to achieving sustainable, large-scale and economical production. Thesis Contributions Plasmonic gold nanoshells synthesis was the focus of this work. The development of chip based microfluidic platform for synthesizing such nanoshells using a reactive gas like carbon monoxide was achieved first. The parallel channel configuration used in this method enabled highly controlled gas-liquid contacting, thereby making possible, never achieved before regulation in the dosage of the reactive gas into the reacting system, just by varying residence times. In doing this, the extent of the reaction could be controlled and hence the morphology of the nanoshells: coverage of gold over the silica particles. Synthesis of monodisperse populations of particles with varying gold coverages and thus optical properties was possible using this method. The two main advantages this platform gave over existing techniques were: 1. Integration of a reactive gas into the reactor in a safe and efficient manner and overcoming the conventional mass transfer problems faced in gas-liquid fast reactions. 2. Tunability of particle morphologies by varying a simple process parameter (residence times) thereby making operation easy. 70 Scale-up of the developed microfluidic system formed the second part of the work. Instead of following the conventional parallelization path, an increase by an order of magnitude in the scales of operation was pursued. A tube based system capable of using a reactive gas for synthesis was developed. Much like the parallel channel configuration used for the chip based method; a tube-in-tube configuration was used for integrating the reactive gas into the system. Using carbon monoxide as the reducing agent, high monodisperse gold nanoshells and nanoislands could be synthesized using this reactor. Achieving product qualities rivalling that of the chip method, a 25x increase in the processing volumes could be achieved. All this could be achieved using a highly simple and economical system built with easily accessible tubings and fittings. Future directions Further Scale-up 1. The throughput of the millifluidic system was tested using only a short reactor (60 cm long) with an ordinarily available PTFE membrane. Building a system with a longer reactor using a membrane with higher gas permeability can enable 3-4x increase in the processing rates. 2. Modest parallelization of the system to include a set of 8-10 reactors in the system using readily available module flow splitters can be implemented to considerable increase the throughput volumes. 3. Increasing particle loading is an alternative path to obtain higher throughputs. However this can lead to aggregation problems (as discussed in section 4.2) that affect the final product quality. Careful optimization of the parameter space to achieve this without inducing aggregation can be a promising future direction. 71 Bibliography 1. L. Zhang, F. X. Gu, J. M. Chan, A. Z. Wang, R. S. Langer and O. C. Farokhzad "Nanoparticles in medicine: therapeutic applications and developments." Clinical Pharmacology & Therapeutics 83, no. 5 (2007): 761-769. 2. S. D. Caruthers., S. A. Wickline and G. M. Lanza "Nanotechnological applications in medicine." Current Opinion in Biotechnology 18, no. 1 (2007): 26-30. 3. O. C. Farokhzad and R. Langer "Nanomedicine: developing smarter therapeutic and diagnostic modalities." Advanced drug delivery reviews 58, no. 14 (2006): 14561459. 4. D. A. DiCarlo, B. Aminzadeh, M. Roberts, D. H. Chung, S. L. Bryant and C. Huh "Mobility control through spontaneous formation of nanoparticle stabilized emulsions." Geophysical Research Letters 38, no. 24 (2011): L24404. 5. T.Zhang, D.Davidson, S. Bryant and C. Huh "Nanoparticle-stabilized emulsions for applications in enhanced oil recovery." In SPE Improved Oil Recovery Symposium. 2010. 6. L. Mazzola "Commercializing nanotechnology." Nature biotechnology 21, no. 10 (2003): 1137-1143. 7. R. Paull, J. Wolfe, P. Hébert and M. Sinkula "Investing in nanotechnology." Nature biotechnology 21, no. 10 (2003): 1144-1147. 8. G. M. Whitesides "Nanoscience, nanotechnology, and chemistry." Small 1, no. 2 (2004): 172-179. 9. L. B. Hunt"The true story of Purple of Cassius." Gold Bulletin 9, no. 4 (1976): 134139. 10. M.Faraday "The Bakerian lecture: experimental relations of gold (and other metals) to light." Philosophical Transactions of the Royal Society of London 147 (1857): 145-181. 11. H. Horvath. "Gustav Mie and the scattering and absorption of light by particles: Historic developments and basics." Journal of Quantitative Spectroscopy and Radiative Transfer 110, no. 11 (2009): 787-799. 12. M. A. El-Sayed "Some interesting properties of metals confined in time and nanometer space of different shapes." Accounts of Chemical Research 34, no. 4 (2001): 257-264. 13. M. A. El-Sayed "Small is different: shape-, size-, and composition-dependent properties of some colloidal semiconductor nanocrystals." Accounts of chemical research 37, no. 5 (2004): 326-333. 14. Y. Xia and N. J. Halas. "Shape-controlled synthesis and surface plasmonic properties of metallic nanostructures." MRS bulletin 30, no. 05 (2005): 338-348. 72 15. S. Eustis and M. A. El-Sayed. "Why gold nanoparticles are more precious than pretty gold: Noble metal surface plasmon resonance and its enhancement of the radiative and nonradiative properties of nanocrystals of different shapes." Chemical Society Reviews 35, no. 3 (2006): 209-217. 16. N. L. Rosi and C. A. Mirkin "Nanostructures in biodiagnostics." Chemical Reviews 105, no. 4 (2005): 1547-1562. 17. E. C. Dreaden, A. M. Alkilany, X. Huang, C. J. Murphy and M. A. El-Sayed "The golden age: gold nanoparticles for biomedicine." Chemical Society Reviews 41, no. 7 (2012): 2740-2779 18. E. C. Dreaden ,M. A. Mackey and M. A. El-Sayed "Beating cancer in multiple ways using nanogold." Chemical Society Reviews 40, no. 7 (2011): 3391-3404. 19. W. Zhao, M. M. Ali, S. D. Aguirre, M. A. Brook and Yi. Li "Paper-based bioassays using gold nanoparticle colorimetric probes." Analytical chemistry 80, no. 22 (2008): 8431-8437. 20. C. L. Schofield, A. H. Haines, A. Robert and D. A. Russell "Silver and gold glyconanoparticles for colorimetric bioassays." Langmuir 22, no. 15 (2006): 67076711. 21. P. Sharma, S. Brown, G. Walter, S. Santra and B. Moudgil "Nanoparticles for bioimaging." Advances in colloid and interface science 123 (2006): 471-485. 22. P. Ghosh, G. Han, M. De, C. K. Kim and V. M. Rotello. "Gold nanoparticles in delivery applications." Advanced drug delivery reviews 60, no. 11 (2008): 1307-1315. 23. E. Boisselier and D. Astruc "Gold nanoparticles in nanomedicine: preparations, imaging, diagnostics, therapies and toxicity." Chemical Society Reviews 38, no. 6 (2009): 1759-1782. 24. D.A. Giljohann, D. S. Seferos, W. L. Daniel, M. D. Massich, P. C. Patel and C. A. Mirkin "Gold nanoparticles for biology and medicine." Angewandte Chemie International Edition 49, no. 19 (2010): 3280-3294. 25. J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao and R. P. Van Duyne "Biosensing with plasmonic nanosensors." Nature materials 7, no. 6 (2008): 442-453. 26. A. Primo, A. Corma and H. García. "Titania supported gold nanoparticles as photocatalyst." Physical Chemistry Chemical Physics 13, no. 3 (2010): 886-910. 27 J. Park, V. Privman and E. Matijevic"Model of formation of monodispersed colloids." The Journal of Physical Chemistry B 105, no. 47 (2001): 11630-11635. 28. J. Xiao and L. Qi "Surfactant-assisted, shape-controlled synthesis of gold nanocrystals." Nanoscale 3, no. 4 (2011): 1383-1396. 29. T. Sugimoto "Preparation of monodispersed colloidal particles." Advances in Colloid and Interface Science 28 (1987): 65-108. 30. J. Th G. Overbeek "Monodisperse colloidal systems, fascinating and useful." Advances in Colloid and Interface Science 15, no. 3 (1982): 251-277. 73 31. H. Naomi "Playing with plasmons: Tuning the optical resonant properties of metallic nanoshells." MRS Bulletin 30, no. 5 (2005): 362-367. 32. S. J. Oldenburg, R. D. Averitt, S. L. Westcott and N. J. Halas "Nanoengineering of optical resonances." Chemical Physics Letters 288, no. 2 (1998): 243-247. 33. S. J. Oldenburg, J. B. Jackson, S. L. Westcott and N. J. Halas. "Infrared extinction properties of gold nanoshells." Applied Physics Letters 75, no. 19 (1999): 2897-2899. 34. E. Prodan, C. Radloff, N. J. Halas and P. Nordlander "A hybridization model for the plasmon response of complex nanostructures." Science 302, no. 5644 (2003): 419422. 35. N. R. Jana, L. Gearheart and C. J. Murphy "Seed‐mediated growth approach for shape‐controlled synthesis of spheroidal and rod‐like gold nanoparticles using a surfactant template." Advanced Materials 13, no. 18 (2001): 1389-1393. 36. B. Nikoobakht and M. A. El-Sayed "Preparation and growth mechanism of gold nanorods (NRs) using seed-mediated growth method." Chemistry of Materials 15, no. 10 (2003): 1957-1962. 37. G. Jinxin, C. M. Bender and C. J. Murphy "Dependence of the gold nanorod aspect ratio on the nature of the directing surfactant in aqueous solution." Langmuir 19, no. 21 (2003): 9065-9070. 38. C. J. Johnson, E. Dujardin, S. A. Davis, C. J. Murphy and S. Mann "Growth and form of gold nanorods prepared by seed-mediated, surfactant-directed synthesis." Journal of Materials Chemistry 12, no. 6 (2002): 1765-1770. 39. X. Huang and M. A. El-Sayed "Gold nanoparticles: optical properties and implementations in cancer diagnosis and photothermal therapy." Journal of Advanced Research 1, no. 1 (2010): 13-28. 40. D. H. Kim, A. Wei and Y. Y. Won "Preparation of Super-Stable Gold Nanorods via Encapsulation into Block Copolymer Micelles." ACS Applied Materials & Interfaces 4, no. 4 (2012): 1872-1877. 41.M. L. Brongersma "Nanoscale photonics: Nanoshells: gifts in a gold wrapper." Nature Materials 2, no. 5 (2003): 296-297. 42. R. Bardhan, S. Lal, A. Joshi and N. J. Halas "Theranostic nanoshells: from probe design to imaging and treatment of cancer." Accounts of Chemical Research 44, no. 10 (2011): 936-946. 43. L. R. Hirsch, R. J. Stafford, J. A. Bankson, S. R. Sershen, B. Rivera, R. E. Price, J. D. Hazle, N. J. Halas and J. L. West "Nanoshell-mediated near-infrared thermal therapy of tumors under magnetic resonance guidance." Proceedings of the National Academy of Sciences 100, no. 23 (2003): 13549-13554. 44. Y. Wang, X. Xie, X.Wang, G. Ku, K. L. Gill, D. P.O'Neal, G. Stoica and L. V. Wang "Photoacoustic tomography of a nanoshell contrast agent in the in vivo rat brain." Nano Letters 4, no. 9 (2004): 1689-1692. 74 45. C. Loo, A. Lowery, N.J. Halas, J. West and R. Drezek. "Immunotargeted nanoshells for integrated cancer imaging and therapy." Nano letters 5, no. 4 (2005): 709-711. 46. S. R. Sershen, S. L. Westcott, N. J. Halas and J. L. West "Temperature‐sensitive polymer–nanoshell composites for photothermally modulated drug delivery." Journal of Biomedical Materials Research 51, no. 3 (2000): 293-298. 47. L. R. Hirsch, J. B. Jackson, A. Lee, N. J. Halas and J. L. West "A whole blood immunoassay using gold nanoshells." Analytical Chemistry 75, no. 10 (2003): 23772381. 48. J. R. Cole and N. J. Halas "Optimized plasmonic nanoparticle distributions for solar spectrum harvesting." Applied Physics Letters 89, no. 15 (2006): 153120153120. 49. O. Neumann, A. S. Urban, J. Day, S. Lal, P. Nordlander, and N. J. Halas. "Solar Vapor Generation Enabled by Nanoparticles." ACS Nano 7, no. 1 (2012): 42-49. 50. H. A. Atwater and A. Polman "Plasmonics for improved photovoltaic devices." Nature Materials 9, no. 3 (2010): 205-213. 51. G. H. Bogush, M. A. Tracy and C. F. Zukoski "Preparation of monodisperse silica particles: control of size and mass fraction." Journal of Non-Crystalline Solids 104, no. 1 (1988): 95-106. 52. A. van Blaaderen, J. Van Geest and A. Vrij "Monodisperse colloidal silica spheres from tetraalkoxysilanes: particle formation and growth mechanism." Journal of Colloid and Interface Science 154, no. 2 (1992): 481-501. 53. S. L. Westcott, S. J. Oldenburg, T. R. Lee and N. J. Halas "Formation and adsorption of clusters of gold nanoparticles onto functionalized silica nanoparticle surfaces." Langmuir 14, no. 19 (1998): 5396-5401. 54. C. Graf and A. van Blaaderen "Metallodielectric colloidal core-shell particles for photonic applications." Langmuir 18, no. 2 (2002): 524-534. 55. D. G. Duff, A. Baiker and P. P. Edwards "A new hydrosol of gold clusters." Journal of Chemical Society, Chemical Communications. 1 (1993): 96-98. 56. W. Vogel, D. G. Duff and A. Baiker. "X-ray structure of a new hydrosol of gold clusters." Langmuir 11, no. 2 (1995): 401-404. 57. J. B. Jackson and N. J. Halas. "Silver nanoshells: variations in morphologies and optical properties." The Journal of Physical Chemistry B 105, no. 14 (2001): 27432746. 58. Y. Kang, X. Ye and C. B. Murray. "Size‐and Shape‐Selective Synthesis of Metal Nanocrystals and Nanowires Using CO as a Reducing Agent." Angewandte Chemie International Edition 49, no. 35 (2010): 6156-6159. 59. F. Dumestre, B.Chaudret, C. Amiens, P. Renaud and P. Fejes. "Superlattices of iron nanocubes synthesized from Fe [N (SiMe3) 2] 2." Science 303, no. 5659 (2004): 821-823. 75 60. L.A. Pretzer, Q. X. Nguyen and M. S. Wong. "Controlled Growth of Sub-10 nm Gold Nanoparticles Using Carbon Monoxide Reductant." The Journal of Physical Chemistry C 114, no. 49 (2010): 21226-21233. 61. J. Ren and R. D. Tilley "Shape‐Controlled Growth of Platinum Nanoparticles." Small 3, no. 9 (2007): 1508-1512. 62. B. E.Brinson, E. Bruce , J. B. Lassiter, C. S. Levin, R. Bardhan, N. Mirin and N. J. Halas "Nanoshells made easy: improving Au layer growth on nanoparticle surfaces." Langmuir 24, no. 24 (2008): 14166-14171. 63. M. C. Kung, R. J. Davis, and H. H. Kung "Understanding Au-catalyzed lowtemperature CO oxidation." The Journal of Physical Chemistry C 111, no. 32 (2007): 11767-11775. 64. S. Duraiswamy and S. A. Khan "Plasmonic nanoshell synthesis in microfluidic composite foams." Nano letters 10, no. 9 (2010): 3757-3763. 65. S. Duraiswamy "Microfluidic Processes Nanomaterials." (2011). for Synthesis of Plasmonic 66. D. R. Reyes, D. Lossifidis, P.A. Auroux and A. Manz "Micro total analysis systems. 1. Introduction, theory, and technology." Analytical Chemistry 74, no. 12 (2002): 2623-2636. 67. J. Atencia and D. J. Beebe "Controlled microfluidic interfaces." Nature 437, no. 7059 (2004): 648-655 68. A. Manz, N. Graber and H. M. Widmer. "Miniaturized total chemical analysis systems: a novel concept for chemical sensing." Sensors and actuators B: Chemical 1, no. 1 (1990): 244-248. 69. G. M. Whitesides "The origins and the future of microfluidics." Nature 442, no. 7101 (2006): 368-373. 70. H. A. Stone, A. D. Stroock and A. Ajdari. "Engineering flows in small devices: microfluidics toward a lab-on-a-chip." Annual Review of Fluid Mechanics 36 (2004): 381-411. 71. T. M. Squires and S. R. Quake. "Microfluidics: Fluid physics at the nanoliter scale." Reviews of Modern Physics 77, no. 3 (2005): 977-1026. 72. A. E. Kamholz and Paul Yager "Theoretical analysis of molecular diffusion in pressure-driven laminar flow in microfluidic channels." Biophysical Journal 80, no. 1 (2001): 155-160. 73. J. P. Brody and Paul Yager "Diffusion-based extraction in a microfabricated device." Sensors and Actuators A: Physical 58, no. 1 (1997): 13-18. 74. C. N. Baroud, F. Okkels, L. Ménétrier and P. Tabeling. "Reaction-diffusion dynamics: Confrontation between theory and experiment in a microfluidic reactor." Physical Review E 67, no. 6 (2003): 060104. 76 75. A. D. Stroock, S. KW. Dertinger, A. Ajdari, I. Mezić, H. A. Stone and G. M. Whitesides. "Chaotic mixer for microchannels." Science 295, no. 5555 (2002): 647651. 76. C. N. Baroud and H. Willaime. "Multiphase flows in microfluidics." Comptes Rendus Physique 5, no. 5 (2004): 547-555. 77. C. N. Baroud, F. Gallaire and R. Dangla. "Dynamics of microfluidic droplets." Lab on a Chip 10, no. 16 (2010): 2032-2045. 78. G. F. Christopher and S. L. Anna. "Microfluidic methods for generating continuous droplet streams." Journal of Physics D: Applied Physics 40, no. 19 (2007): R319. 79. H. Song, J. D. Tice and R. F. Ismagilov “A Microfluidic System for Controlling Reaction Networks in Time.” Angewandte Chemie International Edition 42, no. 7 (2003) 768-772. 80. B. Zheng, L. S. Roach, and R. F. Ismagilov. "Screening of protein crystallization conditions on a microfluidic chip using nanoliter-size droplets." Journal of the American Chemical Society 125, no. 37 (2003): 11170-11171. 81. D. Joshua, H. Song, A. D. Lyon, and R. F. Ismagilov. "Formation of droplets and mixing in multiphase microfluidics at low values of the Reynolds and the capillary numbers." Langmuir 19, no. 22 (2003): 9127-9133. 82. M. R. Bringer, C. J. Gerdts, H. Song, J. D. Tice, R. F. Ismagilov "Microfluidic systems for chemical kinetics that rely on chaotic mixing in droplets." Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 362, no. 1818 (2004): 1087-1104. 83. A. M. Huebner, C. Abell, W. T. S. Huck, C. N. Baroud and F. Hollfelder "Monitoring a Reaction at Submillisecond Resolution in Picoliter Volumes." Analytical Chemistry 83, no. 4 (2011): 1462. 84. S. Marre and K. F. Jensen. "Synthesis of micro and nanostructures in microfluidic systems." Chemical Society Reviews 39, no. 3 (2010): 1183-1202. 85. Y. Song, J. Hormes and C.S.S.R. Kumar. "Microfluidic synthesis of nanomaterials." Small 4, no. 6 (2008): 698-711. 86. C. H.Chang, B. K. Paul, V. T. Remcho, S. Atre and J. E. Hutchison "Synthesis and post-processing of nanomaterials using microreaction technology." Journal of Nanoparticle Research 10, no. 6 (2008): 965-980. 87. J. B. Edel, R. Fortt, J. C. deMello and A. J. deMello "Microfluidic routes to the controlled production of nanoparticles." Chemical communications 10 (2002): 1136. 88. A. Abou‐Hassan, R. Bazzi and V.Cabuil "Multistep Continuous‐Flow Microsynthesis of Magnetic and Fluorescent γ‐Fe2O3@ SiO2 Core/Shell Nanoparticles." Angewandte Chemie International Edition 48, no. 39 (2009): 71807183. 77 89. J. Wagner, T. Kirner, G. Mayer, J. Albert and J. M. Köhler "Generation of metal nanoparticles in a microchannel reactor." Chemical engineering Journal 101, no. 1 (2004): 251-260. 90. J. Wagner and J. M. Köhler "Continuous synthesis of gold nanoparticles in a microreactor." Nano letters 5, no. 4 (2005): 685-691. 91. J. Wagner, T. R. Tshikhudo and J. M. Koehler "Microfluidic generation of metal nanoparticles by borohydride reduction." Chemical Engineering Journal 135 (2008): S104-S109. 92. I. Shestopalov, J. D. Tice and R. F. Ismagilov. "Multi-step synthesis of nanoparticles performed on millisecond time scale in a microfluidic droplet-based system." Lab Chip 4, no. 4 (2004): 316-321. 93. A. M. Nightingale, S. H. Krishnadasan, D. Berhanu, X. Niu, C. Drury, R. McIntyre and E. Valsami-Jones. "A stable droplet reactor for high temperature nanocrystal synthesis." Lab on a Chip 11, no. 7 (2011): 1221-1227. 94. L. Frenz, A. El Harrak, M. Pauly, S. B. Colin, A. D. Griffiths and J.C. Baret "Droplet‐Based Microreactors for the Synthesis of Magnetic Iron Oxide Nanoparticles." Angewandte Chemie International Edition 47, no. 36 (2008): 68176820. 95. S. Duraiswamy and S. A. Khan "Droplet‐Based Microfluidic Synthesis of Anisotropic Metal Nanocrystals." Small 5, no. 24 (2009): 2828-2834. 96. A. Knauer, A. Thete, S. Li, H. Romanus, A. Csaki, W. Fritzsche and J. M. Köhler "Au/Ag/Au double shell nanoparticles with narrow size distribution obtained by continuous micro segmented flow synthesis." Chemical Engineering Journal 166, no. 3 (2011): 1164-1169. 97. J. Santamaria, L. Gutiérrez, V. Sebastian, M. Arruebo and L. Gomez. "Facile synthesis of SiO2/Au Nanoshells in a Three-stage Microfluidic System." Journal of Materials Chemistry 22, no. 40 (2012): 21420-21425. 98. K. Jähnisch, M. Baerns, V. Hessel, W. Ehrfeld, V. Haverkamp, H. Löwe, Ch Wille and A. Guber. "Direct fluorination of toluene using elemental fluorine in gas/liquid microreactors." Journal of Fluorine Chemistry 105, no. 1 (2000): 117-128. 99. J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M. Ueno, T. Kitamori and S. Kobayashi "A microfluidic device for conducting gas-liquid-solid hydrogenation reactions." Science 304, no. 5675 (2004): 1305-1308. 100. P. W. Miller, N. J. Long, A. J. de Mello, R. Vilar, H. Audrain, D. Bender, J. Passchier and A. Gee "Rapid Multiphase Carbonylation Reactions by Using a Microtube Reactor: Applications in Positron Emission Tomography 11C‐Radiolabeling." Angewandte Chemie International Edition 46, no. 16 (2007): 2875-2878. 101. A. Günther, S. A. Khan, M. Thalmann, F. Trachsel and K. F. Jensen "Transport and reaction in microscale segmented gas–liquid flow." Lab Chip 4, no. 4 (2004): 278-286. 78 102. C.P. Park and D. P. Kim "Dual-Channel Microreactor for Gas− Liquid Syntheses." Journal of the American Chemical Society 132, no. 29 (2010): 1010210106. 103. R. A. Maurya, C. P. Park, J. H. Lee, and D.P. Kim "Continuous In Situ Generation, Separation, and Reaction of Diazomethane in a Dual‐Channel Microreactor." Angewandte Chemie International Edition 50, no. 26 (2011): 59525955. 104. J. de Jong, P. W. Verheijden, R. G. Lammertink and M.S Wessling "Generation of local concentration gradients by gas-liquid contacting." Analytical chemistry 80, no. 9 (2008): 3190-3197. 105. J. A. Potkay, M. Magnetta, A. Vinson and B. Cmolik "Bio-inspired, efficient, artificial lung employing air as the ventilating gas." Lab on a Chip 11, no. 17 (2011): 2901-2909. 106. E. A. Lemke, Y. Gambin, V. Vandelinder, E. M. Brustad, H.W. Liu, P. G. Schultz, A. Groisman and A. A. Deniz. "Microfluidic device for single-molecule experiments with enhanced photostability." Journal of the American Chemical Society 131, no. 38 (2009): 13610. 107. R. A. Maurya, C. P. Park and D. P. Kim "Triple-channel microreactor for biphasic gas–liquid reactions: Photosensitized oxygenations." Beilstein Journal of Organic Chemistry 7 (2011): 1158-1163. 108. S. A. Stern, V. M. Shah and B. J. Hardy "Structure‐permeability relationships in silicone polymers." Journal of Polymer Science Part B: Polymer Physics 25, no. 6 (1987): 1263-1298. 109. T. C. Merkel, R. P. Gupta, B. S. Turk and B. D. Freeman. "Mixed-gas permeation of syngas components in poly (dimethylsiloxane) and poly (1trimethylsilyl-1-propyne) at elevated temperatures." Journal of Membrane Science 191, no. 1 (2001): 85-94. 110. J. K. Young, N. A. Lewinski, R. J. Langsner, L. C. Kennedy, A. Satyanarayan, V. Nammalvar, A.Y. Lin and R. A. Drezek "Size-controlled synthesis of monodispersed gold nanoparticles via carbon monoxide gas reduction." Nanoscale Research Letters 6, no. 1 (2011): 1-11. 111. W. S. Beh, I. T. Kim, D. Qin, Y. Xia and G. M. Whitesides. "Formation of patterned microstructures of conducting polymers by soft lithography, and applications in microelectronic device fabrication." Advanced Materials 11, no. 12 (1999): 1038-1041. 112. CRC Handbook of Chemistry and Physics, ed. D. R. Lide and H. P. R. Frederikse, CRC Press Inc, Boca Raton, FL, 76th Edition, 1995. 113. B. Wilks and M. E. Rezac "Properties of rubbery polymers for the recovery of hydrogen sulfide from gasification gases." Journal of Applied Polymer Science 85, no. 11 (2002): 2436-2444. 79 114. W. L. Robb "Thin silicone membranes their permeation properties and some applications." Annals of the New York Academy of Sciences 146, no. 1 (1968): 119137. 115. J. M. Van Baten and R. Krishna. "CFD simulations of wall mass transfer for Taylor flow in circular capillaries." Chemical Engineering Science 60, no. 4 (2005): 1117-1126. 116. A. Günther, M. Jhunjhunwala, M. Thalmann, M. A. Schmidt and K. F. Jensen. "Micromixing of miscible liquids in segmented gas-liquid flow." Langmuir 21, no. 4 (2005): 1547-1555. 117. B. Mattiasson and P. Adlercreutz. "Perfluorochemicals in biotechnology." Trends in Biotechnology 5, no. 9 (1987): 250-254. 118. http://www.cientifica.com/research/market-reports/nanotech-funding-2011/ 119. National Nanotechnology initiative http://www.nano.gov/ 120. P. M. Valencia, O. C. Farokhzad, R. Karnik and R. Langer "Microfluidic technologies for accelerating the clinical translation of nanoparticles." Nature Nanotechnology 7, no. 10 (2012): 623-629. 121. M. J. Pitkethly "Nanomaterials–the driving force." Materials Today 7, no. 12 (2004): 20-29. 122. S. Linic, P. Christopher and D. B. Ingram. "Plasmonic-metal nanostructures for efficient conversion of solar to chemical energy." Nature Materials 10, no. 12 (2011): 911-921. 123. T. K. Sau and C.J. Murphy. "Seeded high yield synthesis of short Au nanorods in aqueous solution." Langmuir 20, no. 15 (2004): 6414-6420. 124. https://www.nanopartz.com/ 125. D.Dean, R. S. Turingan, H. U. Thomann, A. Zolotova, J. Rothschild, S. J. Joseph, T. D. Read, E. Tan and R. F. Selden "A Multiplexed Microfluidic PCR Assay for Sensitive and Specific Point-of-Care Detection of Chlamydia trachomatis." PloS One 7, no. 12 (2012): e51685. 126. A. H. Diercks, A. Ozinsky, C. L. Hansen, J. M. Spotts, D. J. Rodriguez and A. Aderem "A microfluidic device for multiplexed protein detection in nano-liter volumes." Analytical Biochemistry 386, no. 1 (2009): 30-35. 127. M. S. Kim, T. Kim, S. Y. Kong, S. Kwon, C. Y. Bae, J. Choi, C. H. Kim, E. S. Lee and J. K. Park "Breast cancer diagnosis using a microfluidic multiplexed immunohistochemistry platform." PloS One 5, no. 5 (2010): e10441. 128. Q. Zhong, S. Bhattacharya, S. Kotsopoulos, J. Olson, V. Taly, An. D. Griffiths, D. R. Link and J. W. Larson "Multiplex digital PCR: breaking the one target per color barrier of quantitative PCR." Lab on a Chip 11, no. 13 (2011): 2167-2174. 80 129. B. K. McKenna, J. G. Evans, M. C.Cheung, and D. J. Ehrlich. "A parallel microfluidic flow cytometer for high-content screening." Nature methods 8, no. 5 (2011): 401-403. 130. Y. Kikutani, A. Hibara, K. Uchiyama, H. Hisamoto, M. Tokeshi and T. Kitamori. "Pile-up glass microreactor." Lab on a Chip 2, no. 4 (2002): 193-196. 131. K. Jähnisch, M. Baerns, V. Hessel, W. Ehrfeld, V. Haverkamp, H. Löwe, Ch Wille and A. Guber "Direct fluorination of toluene using elemental fluorine in gas/liquid microreactors." Journal of Fluorine Chemistry 105, no. 1 (2000): 117-128. 132. Y. Wada, M. A. Schmidt and K. F. Jensen "Flow distribution and ozonolysis in gas-liquid multichannel microreactors." Industrial & Engineering Chemistry Research 45, no. 24 (2006): 8036-8042. 133.N. Kockmann, M. Gottsponer and D. M. Roberge "Scale-up concept of singlechannel microreactors from process development to industrial production." Chemical Engineering Journal 167, no. 2 (2011): 718-726. 134. T. Iwasaki, N.Kawano and J. Yoshida. "Radical polymerization using microflow system: Numbering-up of microreactors and continuous operation." Organic Process Research & Development 10, no. 6 (2006): 1126-1131. 135. A. Tonkovich, D. Kuhlmann, A. Rogers, J. McDaniel, S. Fitzgerald, R. Arora and T. Yuschak. "Microchannel technology scale-up to commercial capacity." Chemical Engineering Research and Design 83, no. 6 (2005): 634-639. 136. J. Guzowski, P. M. Korczyk, S. Jakiela and P.Garstecki. "Automated highthroughput generation of droplets." Lab on a Chip 11, no. 21 (2011): 3593-3595. 137. W. Li, E. W. K. Young, M. Seo, Z. Nie, P. Garstecki, C. A. Simmons, and E. Kumacheva. "Simultaneous generation of droplets with different dimensions in parallel integrated microfluidic droplet generators." Soft Matter 4, no. 2 (2007): 258262. 138.T. Nisisako and T.Torii. "Microfluidic large-scale integration on a chip for mass production of monodisperse droplets and particles." Lab on a Chip 8, no. 2 (2007): 287-293. 139. X. Niu, M. Zhang, J. Wu, W. Wen and P. Sheng "Generation and manipulation of “smart” droplets." Soft Matter 5, no. 3 (2008): 576-581. 140. D. R. Link, E. G. Mongrain, A. Duri, F. Sarrazin, Zh.Cheng, G. Cristobal, M. Marquez and David A. Weitz. "Electric control of droplets in microfluidic devices." Angewandte Chemie International Edition 45, no. 16 (2006): 2556-2560. 141. K. Zhang, Q. Liang, X. Ai, P. Hu, Y. Wang and G. Luo "On-demand microfluidic droplet manipulation using hydrophobic ferrofluid as a continuousphase." Lab on a Chip 11, no. 7 (2011): 1271-1275. 142. N. de Mas, A.Günther, T. Kraus, M. A. Schmidt and K. F. Jensen "Scaled-out multilayer gas-liquid microreactor with integrated velocimetry sensors." Industrial & Engineering Chemistry Research 44, no. 24 (2005): 8997-9013. 81 143. Particle deposition & aggregation: measurement, modelling and simulation. M. Elimelech, X. Jia, J. Gregory and R. Williams, Butterworth-Heinemann, 1998. 144. G. K. Kurup and A. S. Basu. "Field-free particle focusing in microfluidic plugs." Biomicrofluidics 6, no.2 (2012): 022008. 145.Hatziantoniou, Vasilios, and Bengt Andersson. "The segmented two-phase flow monolithic catalyst reactor. An alternative for liquid-phase hydrogenations." Industrial & Engineering Chemistry Fundamentals 23, no. 1 (1984): 82-88. 146. V. Hatziantoniou, B. Andersson, and N.H. Schoon "Mass transfer and selectivity in liquid-phase hydrogenation of nitro compounds in a monolithic catalyst reactor with segmented gas-liquid flow." Industrial & Engineering Chemistry Process Design and Development 25, no. 4 (1986): 964-970. 147. S. Irandoust and Bengt Andersson "Mass transfer and liquid-phase reactions in a segmented two-phase flow monolithic catalyst reactor." Chemical Engineering Science 43, no. 8 (1988): 1983-1988. 148. T. C. Thulasidas, M. A. Abraham and R. L. Cerro "Flow patterns in liquid slugs during bubble-train flow inside capillaries." Chemical Engineering Science 52, no. 17 (1997): 2947-2962. 149. C. N. Satterfield and F. Özel "Some characteristics of two-phase flow in monolithic catalyst structures." Industrial & Engineering Chemistry Fundamentals 16, no. 1 (1977): 61-67. 150. M. T. Kreutzer, P. Du, J. J. Heiszwolf, F. Kapteijn, and J. A. Moulijn "Mass transfer characteristics of three-phase monolith reactors." Chemical Engineering Science 56, no. 21 (2001): 6015-6023. 151. M. T. Kreutzer, F. Kapteijn, J. A. Moulijn and J. J. Heiszwolf "Multiphase monolith reactors: chemical reaction engineering of segmented flow in microchannels." Chemical Engineering Science 60, no. 22 (2005): 5895-5916. 152. R. A. Edvinsson , M. Nyström, M. Siverström, A. Sellin, A-C. Dellve, U. Andersson, W. Herrmann and T.Berglin "Development of a monolith-based process for H2O2production: from idea to large-scale implementation." Catalysis Today 69, no. 1 (2001): 247-252. 153. R. M. Machado, D.J. Parrillo, R. P. Boehme and R. R.Broekhuis "Use of a monolith catalyst for the hydrogenation of dinitrotoluene to toluenediamine." U.S. Patent 6,005,143, issued December 21, 1999. 154. W. Liu "Ministructured catalyst bed for gas‐liquid‐solid multiphase catalytic reaction." AIChE journal 48, no. 7 (2004): 1519-1532. 155. Al-Rawashdeh, M., F. Yu, T. A. Nijhuis, E. V. Rebrov, V. Hessel, and J. C. Schouten. "Numbered-up Gas-liquid Micro/milli Channels Reactor With Modular Flow Distributor." Chemical Engineering Journal 208 (2012): 645-655 82 156. M. Al‐Rawashdeh, X.Nijhuis, E. V. Rebrov, V. Hessel and J. C. Schouten "Design methodology for barrier‐based two phase flow distributor." AIChE Journal 58, no. 11 (2012): 3482-3493. 157. A. Aimable, N. Jongen, A. Testino, Ma. Donnet, J. Lemaître, H. Hofmann and P.Bowen. "Precipitation of Nanosized and Nanostructured Powders: Process Intensification and Scale‐Out Using a Segmented Flow Tubular Reactor (SFTR)." Chemical Engineering & Technology 34, no. 3 (2011): 344-352. 158. J. C. de Mello, A. M. Nightingale, J. H. Bannock, S. Krishnadasan, F. Omahony, S. A. Haque, J. Sloan, C. Drury and R. McIntyre. "Large-Scale Synthesis of Nanocrystals in a Multichannel Droplet Reactor." Journal of Materials Chemistry A 1, no. 12 (2013): 4067-4076. 159. M. O'Brien, N.Taylor, A. Polyzos, I. R. Baxendale, and S. V. Ley "Hydrogenation in flow: Homogeneous and heterogeneous catalysis using Teflon AF2400 to effect gas–liquid contact at elevated pressure." Chemical Science 2, no. 7 (2011): 1250-1257. 160. P. Koos, U. Gross, A. Polyzos, M. O'Brien, I. Baxendale and S.V. Ley "Teflon AF-2400 mediated gas–liquid contact in continuous flow methoxycarbonylations and in-line FTIR measurement of CO concentration." Organic & Biomolecular Chemistry 9, no. 20 (2011): 6903-6908. 161. N.Lorber, F. Sarrazin, P.Guillot, P. Panizza, A. Colin, B. Pavageau, C. Hany et al. "Some recent advances in the design and the use of miniaturized droplet-based continuous process: Applications in chemistry and high-pressure microflows." Lab on a Chip 11, no. 5 (2010): 779-787 162. M. N. Kashid and D. W. Agar "Hydrodynamics of liquid–liquid slug flow capillary microreactor: Flow regimes, slug size and pressure drop." Chemical Engineering Journal 131, no. 1 (2007): 1-13. 163. W. Tanthapanichakoon, N.Aoki, K. Matsuyama, and K. Mae. "Design of mixing in microfluidic liquid slugs based on a new dimensionless number for precise reaction and mixing operations." Chemical Engineering Science 61, no. 13 (2006): 4220-4232. 164. M. N. Kashid, I. Gerlach, S. Goetz, J. Franzke, J. F. Acker, F. Platte, D. W. Agar, and S.Turek "Internal circulation within the liquid slugs of a liquid-liquid slugflow capillary microreactor." Industrial & Engineering Chemistry Research 44, no. 14 (2005): 5003-5010. 83 [...]... capillary forces, there exist a lot of opportunities for carefully manipulating flows in such channels Functionalization of channel walls to establish wettability gradients for generating flows and inducing flows through channels by capillary forces have been demonstrated70 But the most interesting and widely used application of manipulating surface forces has been for generating high stable multiphase flows:... parabolic flow front in laminar flows III Capillary Number:Multiphase flows The capillary number describes the ratio of viscous to interfacial stresses Ca  U  Here, U is the flow velocity, µ is the viscosity and γ is the interfacial tension Capillary numbers in microchannel flows is less than one In the case of multiphase oil-water flows it ranges between 10 -3 to 10-2 Thus interfacial forces dominate... almost always faster Laminar flows with high Pe, where only molecular diffusion is dominant has been used for laminar flow patterning where confinement between the co-flowing laminar flows is important; for careful separation based on diffusivities of the analytes72,73 and even studying reaction kinetics74 The Tmixer and H-mixer type microfluidic devices were widely popular for such studies Some applications... stable multiphase flows: immiscible liquid-liquid flows and gas-liquid flows76-78 The droplet microfluidic platform has gained a lot of attention and has been used for applications ranging from kinetic studies79 to crystallization80 Here one of the liquids which preferentially wets the wall forms the continuous phase and the other liquid (usually the reagent) forms the dispersed phase Compartmentalization... Microfluidics for nanomaterials synthesis The enhanced mass and heat transfer characteristics, high surface to volume ratios and the ability to exercise precise control over the reactant flow, contact and mixing led to the pursuit of microfluidics as a platform for synthesis of nanomaterials Over the past decade, numerous research groups around the world have demonstrated synthesis of a variety of nanomaterials. .. major challenge Microfluidics as a platform for nanomaterials synthesis is the focus of this chapter Starting from a brief discussion of the fundamentals and features of microfluidics, nanomaterials synthesis methods developed using this platform, with a focus on metallic and semiconductor particles are reviewed Finally the microfluidic techniques currently established for nanoshell synthesis and their... forces is given as Re  dU   Where d is the channel dimension, U is the velocity, ρ is the fluid density and µ the viscosity With micrometer sized dimensions and velocities in the range of centimetres/sec, the Reynolds number for liquid flows can range between the orders of 10 -6 to 10 This shows that the inertial forces are irrelevant in this flow regime This has enabled highly ordered laminar flow. .. use of formaldehyde with ammonium hydroxide for synthesizing silver nanoshells that was adopted even gold nanoshell synthesis57 This method in particular gave smooth nanoshell morphologies Recently Halas and group extended the use of carbon monoxide, a gas that has been previously used for a variety metal nanomaterials synthesis58-61, for the synthesis of gold nanoshells.62 Here instead of formaldehyde... nanomaterials synthesis require extremely rapid mixing of different streams containing reagents Single phase flows in microchannels are laminar in nature and the turbulent convective mixing observed in macroscale flows is absent Here, as mentioned earlier, diffusive mixing is slow compared to convection of materials along the channel For such flows, the distance along the channel that is required for. .. emerged as an effective platform for continuous synthesis of such particles where the inherently high surface to volume ratios and enhanced mass and heat transport properties gave the advantage of better control over the population size and particle morphology It also came with the promise of easy scale up using parallelization and pile-up for large scale commercial synthesis of nanomaterials that may be ... Laminar flows with high Pe, where only molecular diffusion is dominant has been used for laminar flow patterning where confinement between the co-flowing laminar flows is important; for careful... synthesis using wet chemical batch methods and the need for microfluidic continuous flow techniques were discussed The continuous flow methods for gold nanoparticle and gold nanoshell synthesis have... droplet flow over single flow regime for solid nanomaterials synthesis 3.4 Gold nanoshell synthesis For the synthesis of silica-gold nanoshells, a protocol established by Halas group for batch