|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 EXPANSION OF THE POLYMER-INDUCED LI QUID-PRECURSOR (PILP) PROCESS TO NON-CALCIUM BASED SYSTEMS By SARA JENSEN HOMEIJER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008
2 2008 Sara Jensen Homeijer
3 To my family, especially my loving husband, Br ian, who has been my inspiration and rock through this long process, and my parents, w ho have always encouraged and supported me.
4 ACKNOWLEDGMENTS First, I would like to tha nk Dr. Laurie Gower, my adviser and committee chair, for all of her help and guidance. The freedom she allowed me to delve into different types of research and to explore several different opportunities, both at UF and at Hewlett Pack ard, was invaluable. I would like to also acknowledge the rest of my supervising committee for their guidance and support: Dr. Anthony Brennan, Dr. Wolfgang Sigmund, Dr. David Norton and Dr. Joanna Long. I also wish to thank Dr. Mec holsky and Dr. Batich, both former committee members, for their great ideas and guidance during my early work. I have had the great honor of working with many wonderful colleagues during my time here. Much of this work would not have been possible without the help of several fantastic undergraduate research students, especially Richar d Barrett, Tamas Werner, Andreas Baur, Brittany Bales, Brittni Pitts, Mark Haupt and Gare th Strowbridge. I also would like to thank the members of the Biomimetics Research Group esp ecially Dr. Matthew Olszta, Dr. Fairland Amos and Dr. Yi-yeoun Kim for being such great mentors and inspirations. My time here would have been much less interesting (both in terms of research and entertainment) without Mark Bewernitz, my friend and lab mate. Thank you for all your crazy ideas and lively discussions. I also wish to thank my lab mates Sang Soo J ee, Chih-Wei Liao, Palanikkumaran Muthiah, and Dr. Taili Thula, as well as my former lab mates, Dr. Rajendra Kuma r, Dr. Xingguo Cheng, Dr. Lijun Dai and Dr. James Mellman for all of their help, support and great ideas. I would like to thank the Major Analytical Instrumentation Ce nter (MAIC), for the use of their equipment. I would especially like to thank Kerry Siebein for all of her help (and mentorship) on the TEM and SEM, Andrew Gerger for help with AFM. The Cell and Tissue Core Laboratory (CTAC) at th e McKnight Brain Institute, and especially Doug Smith, also deserve my appreciation for the help with and use of their confocal microscope. I also thank the
5 National Science Foundation and the University of Florida for their generous financial support through the Graduate Research Fellowship Program and the Name d Presidential Fellowship. Last, (but certainly not least) I thank my family, especially my husband, Brian, my parents, Glenn and Emily Jensen, and my brother, Kevin Jensen; and my friends, especially Marie and Kevin Kane, and Heather and Chad Macuszonok without poker nights, wine and Orlando weekends, I dont know that I could have lasted this long and the IMG research group, who so kindly adopted me.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.........................................................................................................................9 ABSTRACT...................................................................................................................................13 CHAP TER 1 INTRODUCTION..................................................................................................................15 2 BACKGROUND.................................................................................................................... 18 Biomineralization.............................................................................................................. .....18 Biomineralization Mechanisms....................................................................................... 18 Fibrous Biominerals........................................................................................................21 Vertebrate Enamel.................................................................................................... 21 Sea Urchin................................................................................................................ 23 Biomimetic Mineralization..................................................................................................... 27 Polymer-Induced Liquid-Precursor (PILP) Process........................................................ 27 Mesocrystal Assembly.....................................................................................................28 One-Dimensional Materials.................................................................................................... 30 Functional Materials........................................................................................................ 30 Vapor-Liquid-Solid Mechanism..............................................................................30 Solution-Liquid-Solid (SLS) Mechanism................................................................ 31 Other Methods..........................................................................................................32 Minerals....................................................................................................................... ....32 Carbonate Minerals............................................................................................................. ....37 3 GROWTH OF NANOFIBROUS BA RIUM CARBONATE ON CALCI UM CARBONATE SEEDS...........................................................................................................54 Introduction................................................................................................................... ..........54 Materials and Methods...........................................................................................................57 Materials...................................................................................................................57 Preparation of Seed Substrates................................................................................. 57 Growth of Barium Carbonate Precipitates............................................................... 57 Characterization.......................................................................................................57 Results and Discussion......................................................................................................... ..58 Conclusions.............................................................................................................................63
7 4 THE POLYMER-INDUCED LIQUID-PRECURSOR (PILP) PROCESS IN THE NON-C ALCIUM BASED SYSTEMS OF BARIUM AND STRONTIUM CARBONATE........................................................................................................................70 Introduction................................................................................................................... ..........70 Materials and Methods...........................................................................................................73 PILP Droplet Collection........................................................................................... 73 Dynamic Light Scattering........................................................................................74 Crystal Morphology Experiment.............................................................................. 74 Results and Discussion......................................................................................................... ..75 Conclusions.............................................................................................................................83 Supporting Information..........................................................................................................84 5 MECHANISITIC STUDY OF THE GROWTH OF MINERAL FIBERS IN THE PRESENCE OF POLYACRYLIC ACID .............................................................................. 95 Introduction................................................................................................................... ..........95 Materials and Methods...........................................................................................................98 Materials...................................................................................................................98 Preparation of Seed Substrates................................................................................. 98 Fiber Synthesis......................................................................................................... 99 Characterization.......................................................................................................99 Results and Discussion......................................................................................................... 101 Conclusions...........................................................................................................................112 6 CONCLUSIONS AND FUTURE WORK ........................................................................... 121 Conclusions...........................................................................................................................121 Future Work..........................................................................................................................124 APPENDIX A STRONIUM CARBONATE COATINGS ON BARIUM CARBONATE FIBERS ...........125 Introduction................................................................................................................... ........125 Materials and Methods.........................................................................................................126 Barium Carbonate Fiber Synthesis.........................................................................126 Strontium Carbonate Coating................................................................................. 126 Characterization.....................................................................................................127 Results and Discussion......................................................................................................... 127 Conclusions and Future Work.............................................................................................. 128 LIST OF REFERENCES.............................................................................................................131 BIOGRAPHICAL SKETCH.......................................................................................................141
8 LIST OF TABLES Table page 2-1. List of various biominerals, their formulas, the organism in which they are found, and their functions...........................................................................................................39 2-2. Several crystallographic and physical constants for various group IIA m etal carbonate minerals.............................................................................................................51 5-1. Experim ental conditions for fiber synthesis.................................................................... 114
9 LIST OF FIGURES Figure page 2-1. Examples of biominerals................................................................................................... .41 2-2. Hierarchical assembly of enamel structure........................................................................ 42 2-3. Sea urchin tooth ............................................................................................................... ..43 2-4. Fixed and etched polished surfaces reveal th at there is a very thin organic sheath surrounding each fiber....................................................................................................... 44 2-5. The Polym er-Induced Liquid-Precursor (PILP) process...................................................45 2-6. Morphologies of various m inerals ob tained through the PILP process............................. 46 2-7. Synthetic mesocrystal ........................................................................................................47 2-8. Main m echanisms for mesocrystal assembly..................................................................... 48 2-9. Vapor-Liquid-Solid m echanism for Silicon whisker growth............................................. 48 2-10. Solution-L iquid-Solid Mechanism fo r growth of III-V semiconductors........................... 49 2-11. Solution P recursor-Solid (SPS) mechanism...................................................................... 50 2-12. CaCO3 fibers grown via the SPS mechanism.................................................................... 51 2-13. Formation mechanism for BaSO4 nanofiber formation in the presence of polyacrylate and block copolymers........................................................................................................ 51 2-14. BaSO4 fiber bundles and BaCrO4 individual fibers grown using double hydrophilic block copolymers............................................................................................................... 52 2-15. SrCO3 fibers grown in the pres ence of poly(acrylic acid)................................................. 52 2-16. Hydroxyapatite nanorod assembly aided by fluidic, am orphous calcium phosphate........ 53 2-17. Typical crystal morphologies of carbonate minerals......................................................... 53 3-1. Barium carbonate fibers grown on different substrates, but with sim ilar fibrous morphology........................................................................................................................65 3-2. Energy dispersive spectroscopy analysis of BaCO3 fibers growing off of a calcite rhomb. ........................................................................................................................ ......66 3-3. Scanning electron microscopy and transmission electron microscopy images of BaCO3 fibers grown under different conditions................................................................ 66
10 3-4. X-ray diffraction results for fibers grown on calcite rhom bohedra seeds......................... 67 3-5. Scanning electron microscopy and transmission electron m icroscopy analysis of various fiber morphologies................................................................................................68 3-6. Scannign electron microscopy and transmission electron m icroscopy analysis of larger twisted fiber aggregates........................................................................................... 69 4-1. Experimental set-up for BaCO3 PILP droplet coll ection experiment................................ 85 4-2. Polarized light microscopy images (with gypsum -plate ) of BaColO3 and SrCO3 PILP phase collected from the reac tion solution at early time points................................ 86 4-3. Polarized light microcsopy and scanning electron m icroscopy images of BaCO3 structures produced via the PILP process.......................................................................... 87 4-4. Polarized light microscopy and scanning electron m icroscopy images of SrCO3 structures produced via the PILP process.......................................................................... 88 4-5. Barium and strontium carbonate fibers.............................................................................. 89 4-6. X-ray diffraction results for various c onditions showed less resolved, broadened peaks for all conditions relative to the control reaction (without polym er) for both BaCO3 and SrCO3..............................................................................................................90 4-7. Transmisson electron microscopy images of some BaCO3 and SrCO3 morphologies produced via the PILP process........................................................................................... 91 4-8. Polarized light microscopy and scanning electron m icroscopy analysis of SrCO3 fibers ........................................................................................................................ ......92 4S-1. Scanning electron microscopy images of all experimental conditions for BaCO3............93 4S-2. Scanning electron microscopy images of all experimental conditions for SrCO3.............94 5-1. Fiber morphology from previous experim ents................................................................114 5-2. CaCO3 fibers grown via the PILP process....................................................................... 115 5-3. Fibers grown on freshly cleaved geologic calcite. ........................................................... 115 5-4. BaCO3 fibers growing off of a film of BaCO3 that has delaminated from a calcite rhomb...............................................................................................................................116 5-5. Atromic force microscopy height im ages of individual fibers of BaCO3, CaCO3 and SrCO3. ............................................................................................................................117 5-6. Transmission electron micrographs of BaCO3, CaCO3 and SrCO3 fibers, demonstrating the changes in orienta tion along the length of mineral fibers..................117
11 5-7. Confocal microscopy images of mineral fibers grown with PAA tagged with 5-BMF .. 118 5-8. Proposed mechanism of m ineral fiber formation............................................................. 119 5-9. High resolution transmissi on electron m icroscopy imag es showing and aragonite CaCO3 fiber and a SrCO3 fiber........................................................................................120 A-1. Scanning electron m icroscopy images of BaCO3 fibers partially coated with SrCO3.....130 A-2. Transm ission electron microscopy analysis of coated fiber............................................130
12 LIST OF ABBREVIATIONS ACC Amorphous Calcium Carbonate ACP Amorphous Calcium Phosphate AFM Atomic Force Microscopy EDS Energy Dispersion Spectroscopy GOx Glucose Oxidase PAA Poly (acrylic acid) pAsp Poly (aspartic acid) PILP Polymer-Induced Liquid-Precursor PLM Polarized Light Microscopy SEM Scanning Electron Microscopy SPS Solution-Precursor-Solid TEM Transmission Electron Microscopy XRD X-ray Diffraction
13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EXPANSION OF THE POLYMER-INDUCED LI QUID-PRECURSOR (PILP) PROCESS TO NON-CALCIUM BASED SYSTEMS By Sara Jensen Homeijer December 2008 Chair: Laurie B. Gower Major: Materials Science and Engineering Biologically produced hard tissues, often refe rred to as biominerals, are hierarchical composite structures formed of mineral and or ganic matrix. Often, the mineral phase has a fibrous morphology. The in vitro formation of minerals with the non-equilibrium morphologies found in natural systems has been studied for many years. Our group has proposed the polymerinduced liquid-precursor (PILP) process, which uses polyanionic additives as process-directing agents, as a close mimic for how morphological control may be achieved in biomineralization. Using the PILP process, we have successfully recreated many of the morphological features of biominerals. In this work, the PILP process is expanded into the non-calcium based systems of barium and strontium carbonate, to demonstrat e the non-specificity of organic-inorganic interactions in this process, and the form ation mechanism for minerals with fibrous morphologies is explored. First, fibrous barium carbonate (BaCO3/witherite) crystals 50 nm in diameter and several microns in length were grown on calcium carbonate (CaCO3) seeds at temperatures as low as 4C. The BaCO3 fibers were deposited onto calcite rhombs or CaCO3 films using the polymer-induced liquid-precursor (PILP) process, which was induced with the sodium salt of polyacrylic acid (PAA). Fibers were succe ssfully grown on calcite seeds of various
14 morphologies. These fibers displayed single-crystalline SAED diffraction patterns, but after examining high-resolution TEM lattice images, it was revealed that the fibers were in fact made up of nanocrystalline domains. We postulate that these nanocrystalline domains are well aligned due to a singular nucleation event (i.e., each fiber propagates from a single nucleation event on the seed crystal) with the nanocrystalline dom ains resulting from stresses caused by dehydration during crystallization of the hi ghly hydrated precursor phase. Next, barium and strontium carbonate with various non-equilibrium morphologies were grown in the presence of poly(acrylic acid) sodium salt to induce the polymer-induced liquidprecursor (PILP) process. Previously, the PILP process had only been demonstrated for calcium based systems, such as calcium carbonate and p hosphate. In this report, evidence of a liquidphase amorphous precursor for both barium and strontium carbonate is presented, and these phases were used to synthesize various unique morphologies in the transformed crystals, including films, fibers and cones. Finally, the formation mechanism of these mineral fibers, of both BaCO3 and SrCO3, as well as CaCO3 grown via the PILP process was investig ated in depth. The nanogranular texture of these fibers, first discovered in the BaCO3 system, was confirmed to also be present in the CaCO3 and SrCO3 fibers. Fluorescence studies were done to determine the distribution of occluded polymer throughout the fibers, and AFM, SEM and TEM were employed to study the morphology and structure of these fibers, and a mechanism for fiber formation is described. These findings demonstrate that the PILP pro cess is non-specific a nd applicable to many different ionic salt crystal systems.
15 CHAPTER 1 INTRODUCTION Biologically produced hard tissues, often refe rred to as biom inerals, are hierarchical composite structures formed of mineral and orga nic matrix. The organic matrix, composed of proteins and/or polysaccharides, often consists of an insoluble phase (such as collagen in bone) as well small quantities of water soluble polyanionic proteins. So me classic examples of these types of materials include the brick and mortar structure of mollusk nacre and the lamellar structure of bone. In some cases, the mineral phase has a fibrous morphology. For example, in vertebrate teeth, polycrystalline bundles of hydroxyapatite (HAP) rods are surrounded by a matrix of polycrystalline HAP and a small amount of insoluble proteins and polysaccharides. These complex hierarchies impart great st rength, hardness and fracture toughness to the composites. The in vitro formation of minerals with the non-e quilibrium morphologies found in natural systems has been studied for many years. Ou r group has proposed the polymer-induced liquidprecursor (PILP) process, which uses polyanionic additives as process-directing agents, as a close mimic for how morphological control may be achieved in biomineralization. Using the PILP process, we have successfully recreated ma ny of the morphological features of biominerals, including films1-3, helices1, molded crystals4 and fibers5,6. In this work, the PILP process is expanded into the non-calcium based systems of barium and strontium carbonate, to demonstrate the non-specificity of organic-inorganic intera ctions in this process, and the formation mechanism for minerals with fi brous morphologies is explored. This study contains a literature review on the field of biomineralization, with an emphasis on fibrous biominerals, and the use of biomimetic processes to grow non-biological materials.
16 Inorganic one-dimensional material s are also discussed, specifically in regards to various growth mechanisms and synthesis techniques. This review can be found in Chapter 2. Early in the work on fibrous biominerals in our group, experiments were conducted in the CaCO3 system using Ba2+ and Sr2+ in low amounts as impurities, to mimic the inclusion of various cations in biominerals. As a control experiment, BaCl2 was substituted for CaCl2 in the fiber growth experiment. To our surprise, fibe rs were found to nucleate on the calcite seed crystals. This is especially surprising, as BaCO3 has an aragonitic crystal structure. It was assumed that the fiber formation mechanism migh t be similar to the solution-precursor-solid (SPS) mechanism previously proposed for the CaCO3 fibers, yet differences in fiber morphology led us to question the validity of this hypothesis. This led to a fu ll investigation of the growth of BaCO3 fibers on calcite seeds, whic h is reported in Chapter 3. Next, we expanded the PILP process into th e barium and strontium carbonate systems. This work is detailed in Chapter 4. Following the study of the BaCO3 fibers on seed crystals (detailed in Chapter 3), we wished to confirm that the PILP process was responsible for their formation, as well as expand our studies into the SrCO3 system. First, evidence of a fluidic, amorphous precursor phase was found, which subseque ntly crystallized into birefringent BaCO3 or SrCO3 structures. Next, the effects of ba rium and strontium concentration, polymer molecular weight and polymer concentration on the morphology of the resultant product were studied. A wide range of morphol ogies were found, including spheruli tes, films, fibers, and selfreplicating fibrous cone stru ctures. The differences in the morphologies of the BaCO3 fibers described in Chapters 3 and 4, the SrCO3 fibers described in Chapter 4, and the CaCO3 fibers previously reported by our group6, inspired a more detailed mech anistic study of mineral fiber formation via the PILP process.
17 The mechanism for the formation of fibrous mi nerals is examined in greater detail in Chapter 5. Several mechanisms for fibrous minera l formation in the presence of anionic polymer additives have been proposed by several researchers, including the Solution-Precursor-Solid, or SPS, mechanism proposed by our group6, mesocrystal assembly7-9, colloidal aggregation10, and amorphous phase-aided alignment11 (all described in de tail in Chapter 2). In order to clarify which mechanism(s) are relevant to the CaCO3, BaCO3 and SrCO3 fibers synthesized via the PILP process, several studies were carried out. First, detailed polarized light microscopy (PLM), scanning electron microscopy (SEM) and transmi ssion electron microscopy (TEM) analysis was done to document the morphology, microstructure a nd nanostructure of these fibers. Atomic force microscopy (AFM) studies were performed to gain high-resolution information about the fiber surfaces. Various seed substrates were us ed to determine whether epitaxy plays a role in fiber formation. Finally, fluorescently tagged polyacr ylic acid was used to map the final location of occluded polymer in the fully formed fibers, and a new mechanism for fiber formation is proposed. Chapter 6 revisits and expands upon the conclusi ons generated in the previous chapters. Finally, Appendix A outlines a process for coating the fibrous minerals with another material, resulting in a double-layer fiber.
18 CHAPTER 2 BACKGROUND Biomineralization Hierarchical com posites comprised of an organic matrix and mineral crystals with complex morphologies are abundant in biological systems. Some of the main minerals that have been utilized by nature are listed in Table 2-1. Some examples include bone, which is comprised of aligned platelets of hydroxyapatite embedded in a collagen matrix12, and sea urchin teeth, which are composed of crystalline calcite rods embe dded in a microcrystalline calcium carbonate matrix13. Vertebrate teeth also make use of a fi brous microstructure, where interpenetrating hydroxyapatite (HAP) rods form within a matrix of insoluble proteins and polysaccharides. The organic phase is subsequently degraded and removed as the biomineral matures14. Figure 2-1 shows several examples of biom inerals, which illustrate the complex, nonequilibrium morphologies created by nature. Figure 2-1A is an SEM image of calcite rods from the rib reg ion of a sea urchin14. Figure 2-1B is an SEM image of the calcitic skeleton of a coccolith, a unicellula r form of algae15. Figure 2-1C illustrates the brick-and-mortar structure of mollusk nacre. This biom ineral composite, made of single-crystalline ara gonite platelet bricks with intercalcated organic mortar is renowned for its high fracture toughness and resistance to crack propagation16. The amorphous silica exoskeletons of various species of diatoms are shown in Figure 2-1D. The exoskeleton of each spec ies of diatom has a unique morphology that persists over several generations17. Biomineralization Mechanisms The study of biom ineralization has become increasingly popular as sc ientists strive to understand how nature forms these elaborate inor ganic materials. Biomineralization produces minerals which, in addition to possessing unique and elaborate morphol ogies, also generally
19 have higher fracture toughness than their geologically formed counter parts. This is due to the incorporation of bioorganic additives or impurities, like proteins, lipids, or ionic salts. These additives also have the ability to modi fy mineral formation (polymorph, phase, etc)17. According to Lowenstam and Weiner18, biomineralization can e ither biologically induced or biologically controlled. Bi ologically induced mineralization often occurs as a result of a perturbation in the local environment due to the presence of an organism, which does not necessarily gain any benefit from the mineralizati on, and typically leads to crystals that are not regulated, and exhibit more conventional faceted habits. Biologically controlled mineralization is often of greater interest, because the mech anisms by which organisms are able to direct mineralization are still somewhat poorly understood. Several key points, however, are well understood. First, many biominerals are formed in sealed compartments, or vesicles, creating a mini-reactor for controlled reactions. This allo ws the organism to regulate ionic strength, pH, and concentration and type of additives, among other factors. Some examples of vesicle boundary materials include lipid bilayers, polymer ized insoluble macromolecules, compartments between cells, compartments be tween cells and substrates and compartments between cells and macromolecules. Control over the composition of the mother liquid, or crystallization solution, is vital to the biomineralization proce ss. Additives, such as ionic impurities and macromolecules, play a dramatic role in biomin eralization. For example, sea urchin spine has been shown to be comprised of high magne sium-calcite, with occluded macromolecules present19. It is thought that these additives inhibit crystal nucleation and modify crystal growth. An early theory on role of macromolecules and proteins in biomineralization was the molecular recognition theory20. The inspiration for this theory was the preferred crystallographic orientations of several biomin erals, including the aragonitic pl ates in nacre, which are well
20 aligned with the c-axis of the platelet oriented normal to the face of the platelet, and bone, where the 002 direction of the hydroxyapatite platelets is aligned paralle l to the long axis of the collagen fibrils in which they are embedded. Re peat units within macromolecules are believed to ionically bind with small mineral clusters, possi bly below the critical cluster size, lowering the interfacial energy of the cluster. This lowers the activation energy for nuc leation, thus stabilizing the early nuclei. Provided that the organic matr ix has preexisting organi zation (i.e. charged or polar repeat units in a regular a rray), and that these repeat units are arranged in a complementary way to the nucleating crystal (i.e. negatively charged repeat units in the polymer are spaced to interact with positively charged species on the mineral surface), the macromolecules can effectively template mineral nucleation. The re cognition is chemical and steric, similar in concept to the specific lock-and-key type recogn ition between enzymes and their substrates. By controlling the nucleatio n of the early-phase biomineral the resulting orientation and morphology are controlled. Recently, the idea that biominerals form vi a an amorphous precursor phase has gained favor as amorphous phases have been discovered in an increasing number of organisms. Beniash et al.21 discovered that sea urchin spines form via amorphous calcium carbonate, and Weiss et al.22 discovered that the aragonite in bivalve la rvae also forms via amorphous calcium carbonate. Vertebrate bone has also been shown to form via an amorphous tran sition from amorphous calcium phosphate, ACP, according to Weiner23. In addition, several organisms produce stable amorphous minerals, including numer ous organisms from the taxon Crustacea, as well as several species of plants, in cystoliths24. In a recent article, Navrotsky25 proposes that biomineralization proceeds through hydrated amorphous nanoclusters or particles due to surf ace energy reduction. Hydrated and less ordered
21 phases have a lower surface energy than their mo re stable polymorphs, lowering the activation energy for nucleation. If the thermodynamic stab ility differences between various polymorphs (i.e. vaterite vs. aragonite vs. cal cite vs. ACC) are small, a cross over in stability can occur due to surface energy for small particle si zes. In biological systems, control over both nucleation and crystal growth are necessary to produce minerals with the elaborate morphologies and sometimes non-thermodynamically stable polymorphs (i.e. aragonite in nacre). This requires a balance between thermodynamic driving force and kinetics. Amorphous precursors release a great deal of the thermodynamic drive, and the particles can be further stabilized by interactions with macromolecules in solution. They hydrated amorphous phases are still reactive enough to crystallize, albeit at slow, c ontrolled rates, allowing for the ch aracteristic molded morphology found in many biominerals, according to Navrots ky. In contrast, our group has argued that a kinetically formed amorphous phase can have flui dic character when it is induced by anionic polymers, thus enabling molded non-equilibrium mo rphologies, as will be discussed further in a later section. The vast array of mineralizing organisms makes a complete review of biomineralization difficult. Therefore, two biomin erals with a fibrous morphology, vertebrate teeth and sea urchin teeth, will receive additional fo cus in the following section. Fibrous Biominerals Vertebrate Enamel Enam el is the hardest component of the vertebrate body, made of 96% mineral. Ameloblasts are responsible for the formation of enamel. Enamel is made up of thin, closely packed, long fibers of hydroxyapatite. These fibers bend and twist along their length, and are surrounded with a thin layer of inter-rod enamel. This material is also hydroxyapatite, but has a different orientation than the main enamel fibers26.
22 Cui et al.27 recently described the hierarchical structure of human enamel as being comprised of seven levels, from the nano to micro scale. The first level consists of hydroxyapatite crystals, which form mineral nanofibr ils, the second level. In the third level, the nanofibrils always align lengthw ays, aggregating into fibrous polycrystalline bundles. These bundles then grow to form thicke r, splayed bundles in level four, then prism/interprism continua in level five. At the micro scale, prisms a ssemble into prism bands, level six, which present different arrangements across the thickness of the en amel layer, level seven. This is shown in Figure 2-2. Analogous to bone, the HAP in enam el is aligned in the 002 direction along the long axis of the nanofibrils. The HAP nanofibers cr ystals are surrounded by organic material, similar to the sheath found in sea urchin skeletal elements, which is di scussed further in the following section. The individual fibers were found vi a SEM to be 30-40 nm in diameter. These nanofibers then aggregate together in parall el alignment, lengthwise, to form thicker polycrystalline bundles 80-130 nm in diameter. These then further aggregat e into thicker fibers, ~800 nm in diameter, that assemble in two distinct orie ntations to form prisms and interprism continua27. The fibers are laid down by ameloblasts via the Tomes process, and are oriented such that the c-axis of the HAP crystal is parallel (up to 2.2 misorientation) to the di rection of the enamel. Some of these crystals traverse the entire th ickness of the enamel, and are serpentine in morphology26. In addition, Nakahara and Kakei28 assert that during the ea rliest stages of enamel crystallization, the first minera l deposited is amorphous, and surrounded by an organic envelope. This observation, however, is rarely discussed, and the crystals are usually assumed to form via the conventional crystallizati on pathway. Robinson et al.29 did an in depth AFM study of
23 individual enamel rods, and found that they are relatively smooth at neut ral pH, but upon etching with lower pH solutions, bands and spherical doma ins ~40 nm in diameter were revealed. Using a functionalized tip (either ca rboxyor hydroxyl-terminated), ar eas of relatively high or low friction were mapped, which correspond to charge density on the surface of the crystal. This methodology revealed the band and do main structure in higher reso lution, as the areas between the bands and spheres were relatively low fr iction regions for hydroxylated tips (i.e. less negatively charged). This could point to a relatively higher concentration of occluded organic material, or a less homogeneous crystal structure. The exact formation mechanism of the ini tial nanofibers is unknown; however, we speculate that the SPS mechanism c ould play a major role. This id ea, with further discussion on sea urchin, is explored in furt her detail in the next section. Sea Urchin Echinoderm or sea urchin, skeletons are made up of magnesian calcite (~5% MgCO3), with 0.1 wt% occluded glycoproteins30. The initial mineral depos its are observed during the larval stage, when spicules are formed. Prim ary mesenchyme cells (PMCs) are responsible for spicule growth, fusing together to form the comp artments, called syncytial cords, where spicule growth takes place. The initial deposit is a ~1 m calcite crystal with typical rhombohedra shape. From this initial seed crys tal, three smooth radii grow along the a-axes of the crystal, which in turn form the triradiate syncytium. As the spicule grows, a cluster of these cells is observed at the tip. The rate of spicule elongatio n is independent of the number of PCMs present at the tip, and is approximately 5-13 m/hr. The entire spicule has a smooth, curved surface and diffracts as a single crystal in x-ray diffraction, and has a single-crysta lline extinction pattern in polarized light microscopy. In a addition, a significant amount of A CC was found by Beniash et al.21 in the center of the spicules
24 and a series of concentric rings are observed in high resolution light and transmission electron microscopy. Sea urchin spicules were f ound to be comprised of many well-aligned nanodomains, on the order of 30-50 nm, surrounded by organic material, using both synchrotron studies31 and high resolution TEM32. Sethmann et al.32, who conducted an in-depth TEM analysis of spicules, suggest th at the nano-clustered gr anular growth struct ure results from the basic building block particles. They also propose two possible scenarios to e xplain the single crystalline behavior of the overall structure. Th e first hypothesis is that the clustered crystals may form by oriented aggregation of preformed cal cite nano-particles, similar to the mesocrystal assembly8 crystallization mechanism explained in detail in a later section. The second hypothesis is that spicule growth proceeds via the attachment of ACC particles to the preexisting crystalline biomineral, or by the formation of gelatinous films of CaCO3 and proteins on the biomineral surface. Crystal lization of precursor phase would subsequently proceed from the biomineral surface with protein in tercalation in between the semi-coherent crystal domains. This mechanism is similar to the PILP process2, described in detail in a later section. Similar CaCO3 gelatinous coatings were previously reported by Sethmann33 for CaCO3 mineralization in the presence of pAsp. They also speculate that small amounts of ACC may be present between the crystal nanodomains. The presence of ACC in larval spicules has led to the speculation that calcium and carbonate are transported to the growth front in the form of amorphous calcium carbonate, which has been found in vesicles in side spicule generating cells.34 This hypothesis is supported by a study by Politi et al.,35 who examined regenerating adult urch in spines. Etching with DI water after 4 days of regeneration re vealed a 100-200 nm thick layer of ACC on the outside of the regenerating region. Over time, the ACC on the outside of the spin e is transformed into calcite.
25 The newly regenerated spine also diffracts as a si ngle crystal, meaning that the new spine section grew via isoepitaxy off of the existing, fractured spine. Perhaps most significantly, Ma et al.13 have recently shown that th e teeth of the sea urchin are also formed via a transient amorphous precurso r phase. Similar to the rest of the mineral component of sea urchins, teeth are comprised of high-magnesium calcite (up to 40 mol%). Each urchin has five teeth, which have a T shap ed cross-section. The overall structure of the tooth is shown in Figure 2-3. The teeth are composed of a com plex arrangement of single crystalline calcite plates, needles, and the high magnesium calcite micro-crystals, and behave as two single crystals in polarized light. The tooth has been used as a system for studying mineralization processes in organisms, because it is continuously renewed due to wear on the tip, and thus always contains all stag es of mineralization. The tooth consists of three regions: the plumula, which is the growth regi on, the shaft, which attaches the tooth to the jaw, and the sharp, chewing end. The plumula is enclosed by a singl e layer of epithelial cells which are responsible for continual growth of the tooth. The first el ements formed at the tip of the plumula, the primary plates are attached to lamellar-needle complexes at the basal net. The needles jut out from the primary plates at an an gle. As the needles extend into the keel region, they grow in both diameter and length13,36,37. This is illustrated in Figure 2-3. Similar to the spicules of sea urchin larvae, the fibers of th e tooth are also surrounded by a th in sheath of organic material36. Evidence of these sheaths surroundin g individual fibers both in the early stage and prism stage is shown in Figure 2-4. The resultant microstructure is com prised of long, apparently singlecrystalline needles of high-magnesium calcite embedded in a polycrystalline magnesian calcite matrix.
26 Ma et al.13 found that freshly extracted LNCs from st age I were isotropic in polarized light, indicating that they are compri sed of amorphous material. Afte r soaking in water, the LNCs became birefringent, indicating crystallinity. LN Cs extracted from Stage II were immediately birefringent. Individual needle s from stage I, which are ~150 nm in diameter, were also shown to be amorphous in TEM, as they lacked a distinct diffraction pattern or lattice fringes. At Stage II, however, the needles had a si ngle crystalline diffraction pattern and lattice spacings that corresponded to single crystalline calcite. The growth direction was found to be the 102. In an etching experiment, the core of Stage II needles was etched away, indicating that the shell of the needle is crystalline calcium car bonate, while the interior is ACC. By the time the needles reach Stage III, and grow in diameter to ~3m the entire thickness is etch ed evenly, indicating that the entire thickness is crystalline13. These results are similar to earlier studies on spicule formation21 and spine regeneration35, indicating that the entire sk eleton of the sea urchin forms via a transient amorphous phase. The amorphous phase is considered to be stabilized by the proteins and macromolecules associated with biomineralization. In this resp ect, biomineralization is similar to the PILP process (which will be described further below), in that both processes utilize soluble acidic macromolecules to stabilize an amorphous precu rsor phase. The membrane found by Beniash et al.21 to border the forming spicules, as well as the concentric rings found within the spicules, could be macromolecules that were excluded during crystallization, similar to the transition bars found in calcite film formation via the PILP process2. Similarly, spherical nanodomains, proposed by Sethmann et al.32,38 to be evidence of a precursor pha se in spicules, are also found in dental enamel29, where the more soluble, less charge-d ense regions between the domains could correspond to areas with occluded proteins and/or poor crystallinity. Similar spherical domains
27 have been observed in aragonitic39 and calcitic3 films generated by the PILP process. The complex, unique morphologies pr oduced via biomineralization have inspired a new area of research which will be discussed in the following section. Biomimetic Mineralization Natures ability to cons truct amorphous mi nerals, unique, non-equilibrium morphologies and complex, hierarchical composites in an aqueous environment under ambient conditions have inspired many researchers to mimic these proces ses in the laboratory setting. Biomimetic mineralization utilizes the insigh t gained from the study of biom ineralization to engineer new materials, and includes the repl ication of both the biomineraliz ation process and/or the end product. The biomimetic approach has led to the us e of both soluble additives and insoluble matrices similar to those found in biological systems to modify the nucleation, growth and morphology of inorganic materials under mild processing conditions. In addition to the typical minerals found in biology, these biomimetic pr inciples can also be applied to other, nonbiological materials. Many techniques have been developed in recent years, including the use of scaffolds, such as collagen40 or molded compartments4, templates, such as self-assembled monolayers10,41,42 and Langmuir monolayers39,43 as well as the a ddition of soluble macromolecules to direct crystal growth1,2,44,45. Two major biomimetic crystallization mechanisms will be discussed in greater detail in this dissertation: the Polymer-Induced LiquidPrecursor (PILP) process, and mesocrystal assembly. Polymer-Induced Liquid-Precursor (PILP) Process The use of polymers as crystal growth modi fiers has been extensively studied in many crystal systems, including calcium carbonate1-3,46, silica47, calcium phosphate, and barium carbonate5,45,48-50. Our research group is interested in biomimetic crystallization in particular, and have proposed the polymer-induced liquid-precu rsor (PILP) process as a close mimic for
28 how morphological control may be achieved in biomineralization2,4,40. The PILP process is illustrated in Figure 2-5. In this process, a crystalliz ation solution, consisti ng of anionic polym er and a group IIA soluble salt (commonly CaCl2, BaCl2 or SrCl2), is placed in a closed container with ammonium carbonate. Ammonium carbonate decomposes at room temperature to CO2, ammonia and water. CO2 diffuses into the crystallization solution, acting as a supplier for CO3 2ions. The anionic polymer sequesters cations from solution, which in turn attract carbonate ions, leading to a liquid-liquid phase separation. The ion and polymer-rich phase, or the PILP phase, then collects on the bottom of the petri dish, onto a glass cover slip or other suitable substrate. There, individual droplets coalesce to form an amorphous film, which then pseudomorphically transforms into a crystalline phase through the exclusion of wate r and polymer. Generally, some polymer remains occluded within the crystalline mineral In some cases, other non-equilibrium morphologies are formed, such as nanofibers5,6, helices1 and molded crystals4 are also synthesized. Examples of these various morphologies are shown in Figure 2-6. Mesocrystal Assembly As defined by Colfen and Antionetti8, mesocrystals are made of aligned nanoparticles. Often, mesocrystals appear to be single crystalli ne on the macro scale. For example, sea urchin spines, which have a single extinction direc tion in polarized light, are now described as mesocrystals because study at high magnification shows recognizable individual building blocks. They are often recognized after fusion by the inclusion of organic material in the crystal. Mesocrystal formation mechanisms have been proposed as an analog for biomineralization, due to the sim ilar appearance of biogenic minerals and synthetic mesocrystals. Sea urchin spines51,52, spicules32, and aragonite tablets in nacre53 all exhibit well-aligned nanocrystalline domains, which have organic ma terial both occluded within and surrounding the crystals. In addition to natural minerals, synthetic mesocrystals ha ve also received a great deal
29 of attention lately. Several examples of synthetic mesocrystals are shown in Figure 2-7. In all of these exam ples, individual nanoparticles have self-assembled into highly ordered, aligned structures with a common crystallographic orientation. Three main mechanisms for mesocrystal formation are shown in Figure 2-8. The first possibility, in which nanoparticles are aligned via directional, physical fields requires that the nanoparticles be anisotropic. Som e examples of the forces that may cause this type of mesocrystal assembly include polarization forces as well as electric, magne tic and dipole fields. The anisotropy could either be a property of th e material itself, for ex ample a dipole moment along one crystallographic axis or oppositely char ged faces of a crystal, or be induced through the adsorption of an additive. Identical faces of the crystal may also orient towards each other when in close contact via van der Waals attraction. In the second mechanism, Figure 2-8B, mineral bridges conn ect individual nanoparticles. According to Oaki and Im ai53, in the presence of a polymer, nanoparticle growth is quenched by adsorption onto the surface. The bridges then nuc leate at defect sites, and a new nanocrystal grows on the bridge. This process repeats itself, resulting in a me socrystal. This mechanism is discounted by Colfen and Antonietti8 on the basis of kinetics and thermodynamic arguments. Rather, they argue that the bridges form between oriented nanoparticles through a dissolution/recrystallization m echanism. In addition, this mechanism also applies when amorphous intermediates are present. When there is a high concen tration of amorphous intermediates, for example, PILP droplets, previo usly nucleated structures may continue to grow by the addition of colloidal par ticles from solution. Upon attachment, the amorphous particle is restructured to match the crystallographic orientation of the underlying substrate.
30 In the third possibility, Figure 2-8C, particle alignment occurs via a liquid-crystal like assem bly process. In this mechanism, anisotropi c particles are constrained, and as they continue to aggregate, entropic forces cause the alignment of the particles. One-Dimensional Materials The synthesis of one-dimensional m aterials is especially attractive to the scientific community due to their unique proper ties. In this dissertation, some of the main methods used to fabricate functional materials, such as semic onductors and magnetic materials, will first be discussed, followed by several methods used to fabricate one-dimensional minerals. Functional Materials Vapor-Liquid-Solid Mechanism The Vapor-Liquid-Solid Mechanism VLS, was first proposed by Wagner and Ellis54 in 1964, to describe growth of single-crystalline Si whiskers on silicon substr ate. In this method, gold is used as a catalyst. A small droplet of me tal is placed on Si substrate and heated so that a solid-solution of Si and Au forms. Vapor precursor is added, whic h diffuses into the Au droplet and becomes supersaturated, which then crystallizes out onto substr ate. The droplet of Au then is displaced upwards as that localized region be comes crystalline, and the process continues. This process is illustrated in Figure 2-9. Unidirectional growth results from the difference between sticking coefficients of the atoms from the vapor phas e and the atoms on the liquid and solid surfaces. The liquid surface captures pr actically all the impinging vapor phase atoms, while the solid surfaces reject almost all of th ese atoms because the sticking coefficients are orders of magnitude smaller. Thus, the rate of axial growth of the crystal fed by the liquid will exceed its lateral growth rate by orders of magnitude, leading to the apparent unidirectional growth.
31 Many authors have added to the knowledge ba se of the VLS mechanism over the years, including Givargizov55, who in 1975 calculated many of the kinetic and thermodynamic limitations of VLS growth. He found that the growth rate of whiskers or fibers formed with the VLS mechanism was proportional to the fiber diameter, and that there is a critical diameter for fiber growth, which depends on supersaturation and temperature. He also determined that the rate limiting step in VLS growth is the incorpora tion of ions from the molten droplet into the crystal lattice. R ecently, Gosele et al.56 determined that the size limit of the molten droplet is a function of the vapor pressures of the metal and substrate ions in the vapor phase, and that the whisker critical dimension depends on the solu bility of the whisker components in the metal droplet, and on droplet size. Th e fiber diameters possible via th is mechanism range from 100s of nm to m, due to thermodynamic constraint s on the droplet size. Morales and Lieber57 were able to overcome this using laser ablation to create catalyst droplets of small diameter, which resulted in Si and Ge nanowires with diameters as low as 3 nm. Recently, this group has expanded this concept, using monodisperse Au nanoparticles as catalysts to make nanowires with monodisperse diameters58. Solution-Liquid-Solid (SLS) Mechanism In 1995, Trentler et al.59 proposed the solution-liquidsolid mechanism for III-IV semiconductor wire growth. In this system, pol ycrystalline fibers of various semiconductors, including InP, InAs and GaAs, were synthesized at low temperatures (<203C) in organic solvents. A metallic catalyst with a melting temp erature below the solvent boiling point, such as In, Sn, Bi or Ga, is used to generate the nanowire through the decomposition of organometallic precursors. Surfactants in the solution act to passivate the growing nanowire surface, which further restricts growth to one dime nsion. This process is shown in Figure 2-10.
32 Other Methods Various other m ethods have also been used to fabricate one dimensional materials, including solvothermal synthesis, and templated growth, which will be touched on briefly in this dissertation. Solvothermal synthe sis does not use catalyst particles. Instead, a solvent is mixed with metal precursors and crystal growth regulat ing or templating agents are heated in an autoclave to maintain high temperature and pressure. The growth mechanism is not clear, however some authors speculate that oriented attachment of nanocryst als may be responsible60. Template-directed synthesis utilizes a scaffold fo r material synthesis. An example of a possible template includes nanoscale channels within mes oporous materials, such as porous alumina and track-etch polycarbonate membranes. This me thod is attractive because the size of the resultant rod, whisker or fibe r can be tailored by adjusting the size of the template. The nanoscale channels are filled w ith either the reactant containing solution, a sol-gel, or an electrochemical method, to produce a variety of materials, including metals, semi-conductors, conductive polymers and oxides61. Minerals Various other m ethods have also been used to fabricate one-dimensional mineral materials, including hydrothermal methods and templated gr owth, as well as several biomimetic methods, including the Solution-Precurso r-Solid (SPS) mechanism, which takes advantage of the amorphous, fluidic precursor phase generated by the PILP process, as described below. Hydrothermal methods, utilizing reverse micelle s and microemulsions have been used to produce several minerals, including titania62, hydroxyapatite63,strontium carbonate64, calcium carbonate, barium carbonate and calcium sulfide65. Although the use of a spatially constrained reactor makes this methodology somewhat biom imetic, the high temperatures and pressures used in this method of processing preclude it from being categorized in this way.
33 Previously, our group has found that th e PILP process can lead to CaCO3 product with a fibrous morphology, which we suggested form s via a Solution-Precursor-Solid (SPS) mechanism, due to certain features of the mo rphogenesis that appear analogous to the VaporLiquid-Solid (VLS) and Solution-Liquid-Solid (SLS) mechanisms6. In the two semiconductor processing methods, one-dimensi onal growth is regulated through a flux droplet, and this appeared to be the case for our newly proposed SPS mechanism54,59. In VLS and SLS processes a molten metal flux droplet is cr eated in a high temper ature environment. In contrast, the SPS mechanism utilizes ambient temperatures and aqueous conditions in which the reactant flux presumably is a liquid-phase mineral precursor generated by the PILP process. The proposed SPS mechanism is illustrated in Figure 2-11 and utilizes conven tional calcite crystals (of rhom bohedral habit) as the substrate, or seed crystals upon which the calc ite fibers are grown. In all three methods, a bobble head is observed at the tips of the fibers, which is a remnant of the solidified flux droplet (metal or mineral) that had served to concentrate the reactants into a droplet for the one-dimensional growth of the fi ber. The morphology of the fibers is also indicative of crystallization via a liquid-phase pr ecursor because the fiber growth pattern is not always straight, as one would e xpect from the more traditional solution crystallization process, but instead are curved or serpentine. Simila r serpentine growths are reported for the VLS and SLS mechanisms. In our prior report, it was observed that seed crystals with high defect texture yielded a larger number of fibers, which seems to suggest that the PILP phase is accumulated at the defects until reaching a size sufficient for actin g as a flux droplet (i.e., a larger size droplet may allow it to remain fluidic for fiber forma tion, while nanosized drop lets seem to solidify rapidly upon interaction with the substrate).
34 Mesocrystal assembly has also been propos ed as a formation mechanism for several minerals, including BaSO4 and BaCrO4 66,67 and BaCO3 45. Polyacrylate (as in PILP) and doublehydrophilic block copolymers, which are surface active, were used to facilitate the fiber formation. Many of these fibers were bundled together, branching out from a single nucleation point. Similar to the SPS mechanism, the fi rst stage of fiber formation in these Ba2+ based systems, as proposed by Qi et al. in 200168 for BaSO4, is the formation of amorphous, polymerrich, nano-sized colloidal partic les in solution. Both of the po lymers used here sequester Ba2+ in solution (as in the PILP process), thereby causing the nucleation of the amorphous BaSO4 phase to occur in the Ba2+ rich regions around the polymer chains. This mechanism is shown in Figure 2-13. Initially, am orphous nanoparticles are nucleat ed in solution, and are stabilized by the polymers present. Some particles stick to the reaction vessel or other heterogeneous nucleation site. Other nanoparticles in solution bind to the immobilized particle aggregating into a colloidal cluster, presumably due to van der Waal s attraction. Next, crystallization begins to occur within the aggregates, particles fuse togeth er to reduce surface energy, and lattice energy is reduced by structural rearrangement. Polymer in solution now binds to specific (positively charged) crystal faces. The authors believe that this selective adsorption l eads to an electrostatic multipole field, which drives the further oriented attachment of crystalline aggregates. Secondary nucleation then can cause the formati on of fiber bundles. In another study from the Colfen group67, isolated BaCrO4 fibers were fabricated by introducing a charged colloidal species (microcapsules of PSS/ PAH: poly(styrene sulfonate, sodium salt)/polyallylamine hydrochloride) to the reaction solution, which se rved as a heterogeneous nucleation point for fiber formation. This charged su rface in solution reduces the rate of side-nucleation, which is
35 believed to cause the cone-like fi ber bundles described for the BaSO4 system. Examples of BaSO4 and BaCrO4 fibers from the Colfen group are shown in Figure 2-14. Balz et al.10 synthesized SrCO3 fibers in the presence of pol yacrylic acid (PAA) on a self assembled monolayer (SAM) template. SEM and AFM images of these fibers are shown in Figure 2-15. These authors propose two different possible fiber form ation mechanisms. The first is that PAA chains aggregate into long strands that deprotonate, bind with Sr2+ in solution, and adsorb onto the SAM surface. These PAA stra nds are then proposed to serve as a template for fiber growth. The other is that PAA/SrCO3 aggregates form in solu tion, then deposit onto the SAM. This is suggested because 10-20 nm crystalline SrCO3 particles are observed in solution after only 10 min of reaction time. PAA on the surface of the particles may interact and crossconnect to form fiber like aggregates. In suppor t of this theory, the au thors cite AFM images showing a colloidal surface texture consisting of pa rticles ~30 nm in diameter, the lack of an overall preferred crystallographi c orientation and the interwoven, serpentine appearance of the fibers after drying. Recently, Tao et al.11 have produced mineral structures which appear to form via a mesocrystal-type assembly process, with the ai d of an amorphous precursor that they suggest may have fluidic character. Here, various soluti on additives, including PAA, glycine, glutamic acid and amelogenin were used in the crystal lization of hydroxyapatite. Using glycine, the authors were able to produce rods of HAP, w ith a mosaic structure (common crystallographic orientation, but with areas of disorder) 5 nm in width and 80 nm in length. These rods were formed by the assembly of nanoparticles. The initi al (5 nm diameter) nanoparticles were formed in the presence of PAA (Mw 2000 g/mol) within a few minutes of the start of the reaction, and were surrounded by a thin layer of amorphous calcium phosphate (ACP). These particles were
36 isolated by filtration and drying, th en resuspended into a solution containing Gly or Glu. They were observed to aggregate into larger, unstable colloidal clusters ~30 nm in diameter. In the presence of Gly, these nanopartic les reorganized into linear ch ains, with the crystalline HAP particles linked together by ACP. At this point, the en capsulated nanocrystals were randomly oriented. The ACP between the nanoparticles then crystallized, and, according to these authors, the nanoparticles jiggled around in the amorphous phase to a lign themselves in a common direction. The resultant product was a nanorod ma de of single-crystalline HAP, with a thin coating of ACP. The assembled nanoparticles, crystallization of th e surrounding ACP to HAP and resultant single-crysta lline rod are shown in Figure 2-16. Over a pe riod of 2 m onths, these nanorods assembled into longer, ~100 nm long crystal bundles, similar in morphology to vertebrate enamel fibers. Both the SPS mesocrystal assembly and the ACP alignment mechanism of Tao11 take advantage of amorphous mineral precursor partic les formed/stabilized in solution by acidic polymers. Qi et al.68 and Balz et al.10 argue in favor of selectiv e adsorption of polymers onto specific mineral faces as the driving force for fiber formation. Tao et al.11 argue that the fluidic amorphous ACP enables the assembly and realig nment of primary nanoparticles in a linear assembly, after which it crystallizes, yielding si ngle-crystalline rods. In our previous work6, we have argued that an amorphous mineral flux dropl et directs fiber formation in an mechanism analogous to the VLS and SPS mechanisms, also hi ghlighted in this chapter. In all cases, polyacrylic acid with a low mo lecular weight (2000-5000) was us ed, so it is likely that the mechanism for fiber formation in these experiments is similar. The formation of mineral fibers or rods via an amorphous precursor route has been demonstrated by various authors in several mine ral systems. Although th e mechanism of fiber
37 formation may be slightly different for each mineral, we believe that many commonalities, such as the importance of an amorphous precursor and/or the assembly of primary nano-scale particles, are present. Clarification of this mechanism is the focus of the following chapters of this dissertation. Carbonate Minerals Group IIA m etal carbonates are of great interest to the biomimetic research community. Calcium carbonate, CaCO3, is a biomineral that is major co mponent in invertebrate structural components17. Barium and strontium carbonate are also of interest to the biomimetic community because even though they are not found in biominer als, they serve as interesting model systems for biomimetic mineralization due to their similar chemistry with CaCO3, and that they only form a single polymorph, which is aragonitic. For our purposes, these other two carbonates seemed like a reasonable place to start for dete rmining if the PILP process could be extended into non-biological materials. Ta ble 2-2 describes the different cr ystal structures of calcium, barium and strontium carbonate. They typical morphologies of CaCO3 (calcite and aragonite), BaCO3 and SrCO3 are shown in Figure 2-17 Barium carbonate mineralization has received in creased interest recently due to several important applications, such as the production of barium salts, pigm ent, optical glass, ceramic, electric condensers, and barium ferrite69, and as a precursor for the production of superconductor, piezoelectric and ceramic materials70. Barium carbonate, BaCO3, has a single polymorph, witherite, which is an aragonitic crystal structure, and is subject to twinning71. BaCO3 with interesting morphologies has been produced using a number of different methods. Long, polycrystalline and mosaic fibers have been synthesized in the presence of double hydrophilic block copolymers45,49. Globular aggregates, twisted sheets and helicoidal filaments have been grown in sodium metasilicate gels48. Batch crystallizers have b een employed to produce several
38 morphologies72: reverse micelles have produced long fibers of barium carbonate73; and several mixed solvent and/or template methods have also been employed to stabilize different barium carbonate morphologies74,75. Another IIA metal carbonat e, strontium carbonate, SrCO3, or strontianite, has also recently received increased attention from biomimetic res earchers. Strontianite is the only polymorph of SrCO3, and like witherite, has an aragonitic crystal structure71. Strontium carbonate has many applications, for example in cathode ray tubes fo r televisions and computer monitors, use in fireworks and pyrotechnics, as an additive for specialty glass, a component in ferrite magnets, and as a precursor for various st rontium compounds, including SrTiO3, which is an important piezoelectric material76. Recently, SrCO3 has been used in several biosensor77 and phosphor78 applications. A variety of SrCO3 morphologies have been recently prepared, including spheres, rods, whiskers and ellipsoids64, fibers10, ribbons79, needles80, wires81 and hexahedral ellipsoids82 via several methods, including reverse micelles64,80,83, solvothermal methods64,82,84, self assembled monolayers10,85 at liquid-liquid interfaces79,86.
39 Table 2-1. List of various biominerals, their formulas, the organism in which they are found, and their functions17,18. Mineral Class Name Formula Organism/Function Calcite CaCO3 Algae exoskeletons Trilobites eye lens Aragonite CaCO3 Mollusks exoskeleton Fish gravity device Corals skeleton87 Vaterite CaCO3 Ascidians spicules Amorphous Calcium Carbonate CaCO3 Plants Ca2+ Storage Mollusks16, Spicules21 skeletal precursor IIA Metal Carbonates Strontianite SrCO3 Corals Impurity88 Calcium Phosphate Hydroxyapatite Ca10(PO4)6(OH)2 Vertebrates endoskeletons, teeth, Ca2+ storage Whewellite CaC2O4H2O Plants Ca2+ storage Calcium Oxalate Weddelite CaC2O4H2O Plants Ca2+ storage Gypsum CaSO4 Jellyfish larvae gravity device Barite BaSO4 Algae gravity device IIA Metal Sulfides Celestite SrSO4 Acantharia cellular support Silicon dioxide Silica SiO2nH2O Algae exoskeletons Magnetite Fe3O4 Bacteria magnetotaxis Chitons teeth Goethite -FeOOH Limpets teeth Lepidocrocite -FeOOH Chitons teeth Iron Oxides Ferrihydrite 5Fe2O3H2O Animals and plants Fe protein storage
40 Table 2-2. Several crystallograp hic and physical constants for va rious group IIA metal carbonate minerals. Formula Name Symmetry71 Space Group71 Unit Cell Parameters ()71 Ksp CaCO3 Calcite Trigonal R 3c a = 4.9896 c = 17.0610 3.36-9 Aragonite Orthorhombic P mcn a = 4.9611 b = 7.0672 c = 5.7404 6.0-9 Vaterite Orthorhombic P bnm a = 4.1300 b = 7.1500 c = 8.4800 1.23-8 BaCO3 Witherite Orthorhombic P mcn a = 5.3126 b = 8.8958 c = 6.4284 2.58-9 SrCO3 Strontianite Orthorhombic P mcn a = 5.0900 b = 8.3580 c = 5.9970 5.60-10
41 A D C B Figure 2-1. Examples of biominerals. A) Sea urchin spine, comprised of calcite rods in an amorphous calcium carbonate matrix (B) Calcite skeleton of E. Inxleyi type A coccolith Scale bar = 1 m. C) Aragonite br ick and mortar structure from red abalone shell D) Elaborate silica skeletons of various diatoms Images reprinted with permission from: A) Olszta et al.14 2003 Informa Healthcare. B) Sprengel15 C) Li et al.16 2004 American Chemical Soci ety D) www.micrographia.com
42 Figure 2-2. Hierarchical assembly of enamel structure. Reproduced with permission: Cui et al.27 2007 Wiley-VCH Verlag GmbH & Co. KGaA.
43 Figure 2-3. Sea urchin tooth. Left: A whole sea urchin tooth. Mi ddle: The growth stages of the sea urchin tooth. Right: the st ages of formation of the pl umula (stages I, II and III) and keel in the shaft (stage IV). Plumula stage I: two primary plates (PPs) at the forming tip; Plumula stage II: Lamellar-needle complexes (LNCs) form at the edge of the primary plates. They are oriented out of the plane of the pr imary plate to which they are attached. Plumula st age III: The needles of the LNC grow in both length and diameter. Shaft (IV): Long needles extend from the PPs into the keel of the tooth shaft. Reproduced with permission: Ma et al.13 2007 Wiley-VCH Verlag GmbH & Co. KGaA.
44 Figure 2-4. Fixed and etched polished surfaces rev eal that there is a very thin organic sheath surrounding each fiber. A) Organic sheaths surrounding small fibers in the stone region. B) Organic sheaths surrounding larger fibers (prisms) from the mid-region of the tooth. Reproduced with permission: Wang et al.36 1997 Royal Society of London.
45 CaCl2+ pAsp NH4CO3 CO2 Ca2+Ca2+Ca2+CO3 2-CO3 2-CO3 2Ca2+Ca2+Ca2+CO3 2-CO3 2-CO3 2-Ca2+Ca2+CO3 2-CO3 2Glass coverslip Bottom of petri dish Figure 2-5. The Polymer-Induced Liquid-Precursor (PILP) process. Initially, CO2, generated by the decomposition of ammonium carbonate, di ffuses into a crystall ization solution of anionic polymer (poly aspart ic acid, poly acrylic acid) and a group IIA metal salt (commonly CaCl2, BaCl2 or SrCl2). The anionic polymer sequesters cations from solution, which in turn attrac t carbonate ions. This le ads to a liquid-liquid phase separation. The ion and polymer-rich phase, or the PILP phase, then collects on the bottom of the petri dish, onto a glass cover slip. There, individual droplets coalesce to form an amorphous film (shown here as magenta, to represent the appearance in polarized light microscopy with a gypsum wa ve plate). Over time, this film spreads and the edges smooth out. Patches within the isotropic film become birefringent (orange and blue patches) as crystal tablets nucleate and spread across the precursor. The resultant film retains the morphology of the precursor, although the resultant phase is crystalline.
46 Figure 2-6. Morphologies of various minerals obtained through the PILP process. A) SrCO3 nanofibers. B) CaCO3 helix C) Patterned CaCO3 film on a self assembled monolayer D) Sea urchin spine replica made of molded CaCO3. Images reproduced with permission: B) Gower and Tirrell1 1998 Elsevier. C) Kim et al.3 2007 American Chemical Society. D) Cheng et al.4 2006 American Chemical Society.
47 Figure 2-7. Synthetic mesocrys tals. A) Anatase nanoparticles. B) HRTEM of anatase particles, showing single-crystalline nature. C) BaCO 3 helical mesocrystal D) Calcite selfsimilar mesocrystals. Images reprodu ced with permission: A&B) Penn and Banfield62 1999 Elsevier. C) Yu and Colfen45 2005 MacMillan Publishers, Ltd: Nature Materials. D) Xu et al.89 2008 Wiley-VCH Verlag GmbH & Co. KGaA.
48 A) B) C) Figure 2-8. Main mechanisms for mesocrystal assembly (adapted from Colfen and Antonietti8). A) Particle alignment via physical fields (i.e. electrostatic, magnetic dipole, etc.). Orange arrows indicate mutual alignment. B) Epitaxial alignment between nanoparticles via mineral bridges (orange). C) Alignment via special constraints. Silicon Substrate Au-Si Liquid Alloy Vapor <111>Silicon wire Vapor Figure 2-9. VLS mechanism for Silicon wh isker growth, adapted from Wagner and Ellis54. A droplet of molten gold is placed on a sili con substrate, and a vapor precursor is pumped in. The silicon whisker then grow s out of the Au-Si so lid solution particle, perpendicular to the substrate, with the molten particle remaining on the tip.
49 M, E Crystalline MEGrowth DirectionFlux Droplet SolutionLiquidSolidR3M + EH33 RH Figure 2-10. Solution-Liquid-Solid Mechanis m for growth of III-V semiconductors, adapted from Trentler, et al.59 M and E are group III and group V elements respectively; R3M and EH3 represent precursor salts in solution.
50 Figure 2-11. Solution precursor-solid (SPS) mechanism (adapted from Olszta et al.6). A) Mineral precursor droplets ar e created when the polyanion ic polymer induces liquid liquid phase separation in th e crystallizing solution. B) The PILP precursor droplets adsorb onto calcite substrate seed crystals possibly accumulating at surface defects. Solidified droplets are black, while liquid dr oplets are white. C) After some droplets have adsorbed onto the surface, other fluidi c droplets from the solution coalesce with subsequent droplets and form the primar y flux droplet that leads to onedimensional growth, shown on the right side of C. Fiber growth continues as long as the flux droplet remains fluidic, as indicated by the arrow. At the same time, droplets falling through the solution may wick to th e solidified droplets on the surface of the seed and form a coating on the solid subs trate (left side). D) The flux droplet solidifies, often leaving a bobble remnant on the tip.
51 Figure 2-12. CaCO3 fibers grown via the SPS mechanism. A) Fibers growing off of a calcite rhomb seed substrate. The fibers appear to have an iso-epitaxial relationship with the underlying crystal. B) Bobbles on the tips of the fibrous growths (arrows), which are thought to be remnants of the PILP flux droplets. Reproduced with permission: Olszta et al.6 2004 American Chemical Society Figure 2-13. Formation mechanism proposed by Qi et al.68 for BaSO4 nanofiber formation in the presence of polyacrylate and block c opolymers 1) Polymer stabilized amorphous particles. 2) Heterogeneous nucleation a nd aggregation of amorphous particles into colloidal clusters. 3) Crystallization of cl usters, selective adsorption of polymers onto charged faces. 4) Selective adsorption causes an electrostatic field, which further drives oriented aggregation of crystalline cl usters from solution (5 and 6). 7) This occurs in multiple locations at the same time. 8) Secondary nucleation leads to the formation and widening of bundles and cones. Reproduced with permission: Qi et al.68 2001 Wiley-VCH Verl ag GmbH & Co. KGaA.
52 Figure 2-14. BaSO4 fiber bundles (A) and BaCrO4 individual fibers (B) grown using double hydrophilic block copolymers. Images repr oduced with permission: A) Yu et al.66 Copyright (2003) American Chemi cal Society. B) Yu et al.66 2003 Wiley-VCH Verlag GmbH & Co. KGaA Figure 2-15. SrCO3 fibers grown in the presence of poly(acrylic acid) A&B) SEM images. C) AFM image. Reproduced with permission: Balz et al.10 2005 American Chemical Society.
53 Figure 2-16. Hydroxyapatite nanorod assembly aided by fluidic, amorphous calcium phosphate. A) Previously formed nanospheres are aligned into the linear chains ACP (white arrows cements the HAP nanoparticles together. B) The phase transformation of calcium phosphates and the newly crysta llized HAP domains (circles) in the ACP phase are illustrated between the two HAP nanocrystallites. C) Final linear assemblies are single HAP crystals. Re produced with permission: Tao et al.11 2007 American Chemical Society. Figure 2-17. Typical crystal morphologies of carbonate minerals. A) BaCO3 sheaves of wheat morphology B) CaCO3 (calcite) typical rhombohedra C) SrCO3 spherulite Scale bar = 100 m
54 CHAPTER 3 GROWTH OF NANOFIBROUS BARIUM CARBONATE ON C ALCIUM CARBONATE SEEDS Fibrous barium carbonate (BaCO3/witherite) crystals 50 nm in diameter and several microns in length were grown on calcium carbonate (CaCO3) seeds at temperatures as low as 4 C. The BaCO3 fibers were deposited onto calcite rhombs or CaCO3 films using the polymerinduced liquid-precursor (PILP) process, which was induced with the sodium salt of polyacrylic acid (PAA). The structure and morphology of the resultant fibers were investigated using scanning electron microscopy (SEM), transmissi on electron microscopy (TEM), selected-area electron diffraction (SAED), and polarized light microscopy (PLM ). Fibers were successfully grown on calcite seeds of various morphologies, with a range of barium concentrations, and PAA molecular weight and concentrati on. Two categories of fibers were grown: straight and twisted. Both types of fibers displayed single-crystalline SAED diffrac tion patterns, but after examining high-resolution TEM lattice images, it was revealed that the fibers were in fact made up of nanocrystalline domains. We postulate that these nanocrystalline domains are well aligned due to a singular nucleation event (i.e., each fiber propagates from a si ngle nucleation event on the seed crystal) with the nanocrystalline domains resul ting from stresses caused by dehydration during crystallization of the highly hydr ated precursor phase. These BaCO3 fibers grown on calcite substrates further illustrate the robustness and non-specifi city of the PILP process. Introduction Hierarchical com posites comprised of an organic matrix and mineral crystals with complex morphologies are abundant in biological systems. Some examples include sea urchin teeth, which are composed of crystall ine calcite rods embedded in an amorphous calcium carbonate matrix. Vertebrate teeth also make use of a fibrous microstructure, where interpenetrating hydroxyapatite (HAP) rods form within a matrix of insoluble proteins and polysaccharides. The
55 organic phase is subsequently degraded and removed as the biomineral matures14. Minerals with fibrous morphology have been produced in vitro using several techniques, including hydrothermal processes using reve rse micelles and microemulsions62-65, the biomimetic solutionprecursor-solid (SPS) mechanis m reported in our prior work6, using block copolymers as templates45,50, and self assembly of bacteriophage90. These methods have been used to synthesize fibers of a variety of material s including calcium, barium and strontium carbonate6,45,64,65,73,91, hydroxyapatite63,68,92 and barium sulfate68. The reverse micelle technique has successfully produced single-crystalline fibers of a variety of materials. The harsh chem icals and high temperatures required ( 80C), however, make this type of process less useful for biom imetic applications, whic h capitalize on the benign processing conditions that allow for incorporation of biological components. Both the block copolymer and SPS mechanisms utilize aqueous so lutions with polyanionic additives and low temperatures for crystallization; however, the SPS mechanism is capable of producing fibers that appear to be single crystalline6, while the block copolymer appro ach produces fibers made of nanocrystalline self-assembled aggregates45. Recently, Wang, et al.50 synthesized witherite nanofibers with diameters of 40-140 nm diam eter and several millimeters in length using phosphonated double-hydrophilic block copolymers. Th ese fibers were found to be composed of ~20 nm witherite nanoparticles. Our group is interested in biomim etic crystallization processes2,14 and has proposed that the polymerinduced liquid-precursor (PILP) process may play a fundamental role in the morphogenesis of calcitic biominerals. We have recently expanded our studies to determine if the PILP process can be induced in other mine ral systems, such as calcium phosphate, and as reported here, barium carbonate. In the PILP pro cess, described in detail in our previous report2,
56 a highly hydrated mineral precursor is induced by using short-chai ned, polymeric processdirecting agents, such as the polyanionic salt of polyaspartic aci d (poly-Asp) or poly(acrylic acid) (PAA). In this process, the negatively ch arged polymer sequesters a high concentration of ions and inhibits crystal nucleation such that liquid liquid phase separation of droplets of a metastable precursor is formed within the cr ystallizing solu tion. These nanoscopic droplets, which can grow to a size of several microns in diameter, usually settle on the substrate and coalesce into amorphous films. The films subs equently transform th rough solidification and crystallization into crystalline mineral films com posed of calcite, vaterite or aragonite, in the case of the calcium carbonate system. In addition, we have also found that the PILP process can lead to CaCO3 product with a fibrous morphology, whic h we suggested grow via a SolutionPrecursor-Solid (SPS) mechanism, which was de scribed in detail in Chapter 2 of this dissertation, and shown in Figure 2-10. Here, the non-specificity of the PILP process wa s examined to see if fibers of one mineral, barium carbonate, could be grown on seeds of anot her mineral, calcite, which would indicate if epitaxy is an essential co mponent of the fiber formation proce ss (as in the VLS mechanism). It is assumed that epitaxy is not feasible since the two minerals differ in both composition and lattice structure. The witherite fibers, which have an aragonitic crys tal structure with an orthorhombic lattice, were surpri singly found to nucleate on calcite seed crystals, which have a hexagonal lattice structure. In addition, it was found that CaCO3 films deposited by the PILP process could also be used as suitable substrates to seed the formation of fibers. In accordance with our previous findings6, the calcium carbonate PILP films, which have a higher defect density than solution grown calcite rhombs, were found to yield a high density of fibers when used as seed substrates for fiber formation.
57 Materials and Methods Materials All chem icals were reagent grad e and used without further puri fication. Stock solutions of CaCl2H2O (Sigma), BaCl2 (Fisher) and polyacrylic acid so dium salt (Aldrich; Mw = 5100 and 8000 g/mol) were filtered three times through 0. 22 m Millipore syringe tip membrane filters before use. Preparation of Seed Substrates Calcite seeds were grow n on glass cover slip s overnight at room temperature in a 60 mm OD Petri dish filled with 3 mL of a 45 mM CaCl2 (dihydrate, Sigma) so lution via vapor diffusion of the decomposition products of (NH4)2CO3 (Sigma) (NH3, CO2 and H2O). Calcium carbonate film seeds were grown on glass cover slips overn ight in a 60 mm OD Petri di sh filled with 3 mL of a 6 mM CaCl2 (dihydrate, Sigma) solution, with 50 g/ml PAA (Mw = 8000 g/mol, Sigma) additive. Substrates were wash ed in water and ethanol and dried using nitrogen prior to use. Growth of Barium Carbonate Precipitates Calcium carbonate substrates were placed in a solution of BaCl2 (1.5, 3 or 6 mM) and PAA (MW = 5100 or 8000 g/mol, 0, 10, 50, 100 and 1000 g/ml) at 4C for 1 week or at room temperature (20C) for 24 hours in the presence of NH4CO3. At the conclusion of the experiment, the samples were washed in wa ter and ethanol and dr ied with nitrogen. Characterization The precipitates m orphology and size were ex amined using polarized light microscopy (Olympus BX60), scanning electron microsc opy (JEOL 6400, JEOL 6335f), and transmission electron microscopy (JEOL 2010f). Chemical composition was analyzed using energydispersive x-ray spectroscopy, and crystal stru cture using selected area electron diffraction (JEOL 2010f).
58 Polarized Light Microscopy (PLM) Analysis: Samples were examined using an Olympus BX60 polarized optical microscope with a first-order red (gypsum) -plate. Both calcite and witherite are birefringent when illuminated with crossed-polars, and appear bright yellow or blue when using the gypsum plate, in contrast to th e bright magenta background that is seen for optically isotropic substances, such as the gla ss slide, or the amorphous mineral precursor. All samples were first rinsed gently with water, th en with ethanol and drie d with nitrogen prior to viewing. Scanning Electron Microscopy (SEM) Analysis : The samples were dried under vacuum overnight, fixed to an aluminum stub using doubl e-sided copper tape, and then sputter coated with Au/Pd. The samples were then examined with either a JEOL 6400 SEM at an accelerating voltage of 15 kV or a JEOL 6335F FEGSEM inst rument at an accelerating voltage of 10 kV, both equipped with energy dispersive spectrometers (EDS). Transmission Electron Microscopy (TEM) Analysis: To examine the fibrous BaCO3, the samples were first scratched with a razor blade to dislodge some of the fibers from the CaCO3 substrate. A small aliquot of ethanol was then dispensed onto the scratched area and immediately drawn up using a micropipette. The removed aliquot was then dropped onto a 200 mesh copper TEM grid coated with carbon. The sample was examined on a JEOL 2010f transmission electron microscope at 100 kV in bright field (BF) and selected area el ectron diffraction (SAED) modes. Results and Discussion Barium carbonate fibers of high aspect ratio were grown on calcium carbonate seeds under a range of reaction conditions. Fibers were fo rmed on seeds of both calcite rhombs and PILP generated calcium carbonate thin films, for various concentrations of BaCl2, and PAA, as well as different reaction temperatures, tim es, and PAA molecular weight. Figure 3-1 shows PLM and
59 SEM images of fibers grown on bot h calcite seed crystals and CaCO3 film composed of domains and patches of single-crystalline calcite. In the PLM image of Figure 3-1A, the birefringent calcite seed crystals can be read ily identified as the bright and da rk rhombohedral shaped crystals (~60 m arrows). The fibers are growing both o ff the surface and between the rhombs, as shown at higher magnification in the SEM image ( Figure 3-1B). It appear s as though the bundles of fibers were stretched between the calcite rhom bs by capillary forces during the drying process. For the film seeds, the bright blue and yellow patches in the PLM image ( Figure 3-1C) correspond to the calcium carbonate film seeds (arrows), but the fibers are not readily identifiable at this low magnifica tion. In the SEM image, however, it is apparent that there are a large number of fibers randomly arranged in a mesh-like structure, which corresponds to the cloudy brownish regions in Figure 3-1C. We consider the calc itic film s to have a high defect texture due to the edges of the films, which are around 500 nm thick, as well as the fact that these films often express a variety of crystallographic faces that are le ss thermodynamically stable than the 104 faces found on the calcite rhombs. We wished to confirm that the fibers s een growing off of the calcite rhombs in Figure 3-1 were in fact com prised of barium carbonate and not calcium carbonate with barium impurities. In previous work, it was found that Sr2+, Mg2+ and Ba2+ can be incorporated into CaCO3 fibers6. Using the Oxford EDS system on the JEOL SE M 6400, we gathered spectra in two locations, shown in Figure 3-2. The first spectrum was from the area o ver the rhomb, which was also covered with fibers. The second spectrum was fr om an area comprised solely of fibers growing off of the rhomb. In the first sp ectra, the elements Ca, Ba, C and O are all present, consistent with both BaCO3 and CaCO3 phases that were present. The Au and Pd peaks are from the thin conductive coating applied to the sample prior to SEM imaging to prevent charging. In the
60 second sample, the Ca peak is greatly depressed, the Ba peak is larger compared to the Au and Pd peaks, which are assumed to remain roughly c onstant across the entire sample. This evidence supports the assumption that the fibers are comprised of BaCO3, but, because EDS is not a quantitative technique, furt her analysis was needed. As can be seen in the SEM and TEM images of Figure 3-3, the diameter of the fibers is 63 8 nm as determined from a sampling of 100 fibe rs, and appears to be relatively consistent along the length of the fibers. Electron diffractio n, as shown in SAED pa ttern in the inset of Figure 3-3B, showed a consistently single-crystalline spot patter n along the length of the fiber, with d-spacings that correspond to the ther m odynamically stable phase of BaCO3, witherite, confirming the results of the EDS study (Figure 3-2). This phase determination was also supported by XRD, Figure 3-4, which showed peaks from both the c alcite seeds and witherite phases. For most reaction conditions, two varieties of fibers were found, straight and twisted. Figure 3-5A is an SEM image of several fibers. A typical straight fiber, labeled S, and twisted fiber, labeled T, are marked. These two types of fibers were analyzed using the TE M, shown in Figure 3-5B. The straight fiber in this figure was fractured duri ng sample preparation. In this im age, texture variations within the fibers are ap parent. The twisted fiber can be distinguished in TEM by the light and dark stri ations along the length of the fiber as the twist creates thicker and thinner regions across the electron beam. Latti ce images of both fibers (the straight fiber in Figure 3-5C, the twisted fiber in Figure 3-5D) show that the fibers are not single crystalline, as was indicated by the single-cry stalline dif fraction pattern in Figure 3-3B, but are in fact made up of m any nanosized crystalline domains of approximate ly 5 nm in diameter. In the straight fiber,
61 the domains appear to be roughly spherical, whil e in the twisted fiber, they are elongated and slightly larger, ~10-15 nm. In addition to the straight and twisted fibe rs previously described, in some cases larger fibers, which appear to be comp rised of high-aspect-ratio nanoparticles assembled into belts or ribbons, were found. The SEM and TEM images of these fibers are shown in Figure 3-6A and Figure 3-6B, respectively. The selected-area di ffraction p attern of a segment of the aggregate fiber ( Figure 3-6B, inset) does not exhibit distinct diffraction spots, but instead is com prised of arcs. This indicates a well-ali gned, polycrystalline structure. The aforementioned single-crystalline diffraction pattern of the thinner fibers is especially surprising in light of the HR-TEM images in Figure 3-5C) and D), which shows nanodomains within th e structure of fibers that appeared si ngle crystalline in electr on diffraction and PLM. Recently, Yu et al.45 synthesized helical barium carbonate fibers ( Figure 2-7C) that were quite sim ilar to the aggregate fibers shown in Figure 3-6A, B, in that they both appear to be com posed of aggregated or self-assembled nanocrystals. In that paper, the fibers were shown via high resolution SEM to be composed of ~20 nm doma ins. The domains in Yus fibers correspond well to the apparent domain size of our aggregate fibers. The twisted fibers shown in Figure 35A, B, and D also appear to be composed of elongated, s ingle-crystalline domains that are well aligned. The domain size is smaller, however, on the order of 5 nm across. This is similar to the spherical domains in our straight fibers, which se em to be similar to the spherical domains in calcitic sponge spicules, as shown by Sethmann et al.32 with high resolution TEM. These authors suggested a correlation between these domains and the in vitro observations of nanocluster mediated growth of calcite in the presence of polyaspartic acid33. In the PILP system, films that are smooth at the micron scale can be seen by AFM to also be composed of nanoclusters3,
62 similar to those reported by Sethmann et al.32 33 Our studies find that P AA and polyaspartic acid behave very similarly in generating PILP nanodrople ts, so the similarity in nanocluster features with the other reports are not surprising. It is not clear if fiber formation in this BaCO3 system proceeds via the SPS process because there is no evidence of a bobble tip on these fibers, which is a remnant of the flux droplet that leads to one-dimensional growth6. However, these fibers are very small, which means the flux droplet would also be small, and could conceivably taper at the tip as the droplets in solution decline in concentration. Alternatively, the elongated particles and helical nature of the fibers shown in Figure 3-6 suggest a m esocrystal type assembly, similar to that observed by Yu et al.45. The aggregate fibers grown by Yu et al. had an obvious polycry stalline nature observable under SEM, which is similar to our large aggregate fibers, except that th ey were larger in diameter. The twisted fibers shown in Figure 3-5 also have some elongated doma ins under high resolution TEM, but lack the more readily apparent polycrystalline appearance and diffraction pattern. Given the similarity in processing conditions for the BaCO3 to our prior CaCO3 system, and the lack of an explanation for how the simple PAA additive could lead to an organized mesocrystal assembly, we currently favor the SPS explanation6, but plan further mechanisti c studies to verify. In our system, the single crystalline diffrac tion patterns, which persisted even in the twisted fibers across the bends (re sults not shown), must be due to a high degree of alignment of the nanocrystalline domains along the axis of the fibe r. One explanation is that the fiber was first grown as an amorphous precursor phase, as previously postulated for the CaCO3 system6, which then crystallizes into wither ite via a pseudomorphic transf ormation (retaining the fibrous morphology). We have observed in other studies that contraction occurs during the amorphous to crystalline transformation as the waters of hydration and po lymer are excluded, leading to a
63 considerable amount of lattice st rain (observed as a gr adual shift in crysta llographic orientation of PILP deposited films)2. This would explain how a crystal that emanates from a singular nucleation event can result in a mosaic texture, which we believe arises from the breakdown of the lattice stresses caused by th e dehydration of the precursor phase. As the nanodroplets coalesce into continuous films or fibers, a sm all amount of polymeric impurities may become entrapped at the boundaries, leading to a nanoclust er texture within an overall single-crystalline structure. This theory is also compatible with the theory hypothesized by Sethmann et al.33 for the mechanism of growth of nanoclustered calcite. In that study, a model system for biomineralization was studied in situ using atomic force microscopy. Calcite seed crystals were placed in a solution containing po lyaspartate, calcium chloride and sodium carbonate solutions. It was observed that a transient, amorphous, ge l-like phase was first observed, which then transformed into nanoclusters of calcite on the order of 100 nm in size. Wang et al.50 also observed similar spherical nanosized domains of ~20 nm on the surface of BaCO3 fibers grown in the presence of double-hydrophi lic-block-copolymers. Similar nanosized domains observed in sea urchin spines and sponge spicules appear to be single-crystal line structures as well31,32,52,93, yet have been described51,52 as being iso-oriented mosaic cr ystals due to the mosaic texture observed by HR-TEM and diffraction analysis. In terestingly, these biominerals have now been shown to grow from an amorphous precursor pathway21,22,24,94,95, whose growth is presumably modulated by the anionic proteins ex tracted from the biominerals. Conclusions The PILP process com bined with the SPS meth od of generating mineral fibers, which was originally confined to the CaCO3 system, has been demonstrated to be viable for BaCO3. This process is especially interesting in that the BaCO3 fibers differ in both composition and structure than the seeds, which suggests the possibility of using this non-specific process for fiber
64 formation in other inorganic systems. We also show that films can be used as seed substrates for fiber formation, and the presumed high de fect density of the PILP-formed CaCO3 films yielded a dense fibrous mesh of BaCO3. This correlates with our previous findings that suggest that defects stimulate fiber nucleation. The fibers, which appear single crystalline when examined optically and by diffraction, show a nanodomain lat tice texture, even though they likely grow from a singular nucleation event; therefore, we consider the fibers to be singl e crystals with mosaic texture. In addition, the nanodomain texture may be related to the nanoclusters that have been observed in a variety of PILP formed cr ystals, and correlates well with the textures observed in biologically formed minerals.
65 1 m 100 m 1 m 100 m Figure 3-1. Barium carbonate fibers grown on different substrates, bu t with similar fibrous morphology. (A&C) PLM and (B&D) SEM imag es of fibers grown on calcite rhomb seeds (A&B) and on calcium carbonate film seeds (C&D). The arrows indicate calcite seed crystals.
66 10 m Ca Ba Ba Pd Au O C Figure 3-2. EDS analysis of BaCO3 fibers growing off of a calcite rhomb. A B100 nm A B100 nm Figure 3-3. SEM (left) and TEM (ri ght) high magnification images of BaCO3 fibers grown under different conditions. A) SEM of isolated fibers from the mesh of fibers growing off of CaCO3 films: 24 hours, room temper ature, film seed, 6 mM BaCl2, 50 g/ml PAA, MW = 8000 g/mol (also shown in Figure 3-1C,D); B) TEM and SAED pattern (inset) of a fiber extracted from a rhomb seed crystal: 1 week, 4C, calcite seed, 6 mM BaCl2, 10 g/ml PAA, MW = 5100 g/mol (also shown in Figure 3-1A,B); both conditions result in fibers with diam eters of 50-100 nm, lengths of 10s of microns, and in this case, a gr owth direction .
67 15 20 25 30 35 40 45 50W C C C C C C W W W W W WW 2 (deg)Arbitrary Units 15 20 25 30 35 40 45 50W C C C C C C W W W W W WW 2 (deg)Arbitrary Units Figure 3-4. X-ray diffraction results for fibers grown on calc ite rhombohedra seeds 10 g/ml PAA (Mw 5100 g/mol) at 4Cfor 7 days, and 6 mM BaCl2 (top, red) and 3 mM BaCl2 (bottom, blue). All the peaks in both cases can be attributed to either the calcite seeds or witherite fibers.
68 Figure 3-5. SEM and TEM analysis of vari ous fiber morphologies. A) SEM image of both straight and twisted fibe rs, labeled S and T. B) High resolution TEM image of the two types of fibers shown in A. C) La ttice image of a straight fiber, S, which shows spherical, approximately 5 nm domains (several individual domains circled). D) Lattice image of twisted fiber, T, showing elongated domains approximately 15 nm in diameter and 50 nm in length.
69 Figure 3-6. SEM and TEM analysis of larger twisted fiber aggreg ates. A) SEM image of larger twisted fibers, approximately 200 nm in diameter, composed of elongated, parallel aligned nanodomains, approximately 20 nm in diameter. B) TEM image of aggregate fibers shown in E. Inset: the correspondi ng selected area diffraction pattern, showing arcs instead of diffraction spots, indica ting an aligned polycrystalline texture.
70 CHAPTER 4 THE POLYMER-INDUCED LIQUID-PRECURS OR (PILP) PROCESS IN THE NONCALCIUM BASE D SYSTEMS OF BARI UM AND STRONTIUM CARBONATE Barium and strontium carbonate with various non-equilibrium morphologies were grown in the presence of poly(acrylic acid) sodium sa lt to induce the polymer-induced liquid-precursor (PILP) process. Previously, the PILP proce ss had only been demonstrated for calcium based systems, such as calcium carbonate and phosphate. In this report, evid ence of a liquid-phase amorphous precursor for both barium and stront ium carbonate is presented, and these phases were used to synthesize various unique morphologies in the tr ansformed crystals, including films, fibers and cones. These findings demonstr ate that the PILP process is non-specific and applicable to many different ionic salt crystal systems. Introduction The field of biom ineralization and its syntheti c counterpart, biomimetic mineralization, has been very active in recent years. Specificall y, the use of polymers as crystal growth modifiers has been extensively studied in many cr ystal systems, including calcium carbonate1-3,46, silica47, calcium phosphate96, barium sulfate44,66-68 and barium carbonate5,45,48-50. These techniques include the use of doublehydrophilic block copolymers45,49,50,97, templates42,48 and acidic polymeric additives1-4,33,39, as examples. Our groups work in biomimetic crystallization has led to the hypothesis that the polym er-induced liquid-precursor (PILP) process may serve as a close mimic for how morphological control can be achieved in biomineralization2,4,40. In light of the recent evidence that shows that many CaCO3 biominerals are formed via a transient amorphous precursor, we hypothesize that th e acidic proteins found associat ed with biominerals may follow a pathway that is similar to what we observe in our in vitro model system. In this case, the polymer additive is not considered to be a structur e-directing agent, as in selective adsorption to
71 specific crystallographic faces, but rather acts as a process-directing agent, which brings about the kinetically preferred multi-step crystallization pathway. We have recently expanded our studies to determine if the PILP process can be induced in other mineral systems, such as calcium phosphate96, and as reported here (and previously)5, barium carbonate. In the PILP process, describe d in detail in our previous report for the CaCO3 system2, short-chained, polymeric process-directing agents, such as the polyanionic salt of polyaspartic acid (poly-Asp) or poly(acrylic acid) (PAA), transform the conventional crystallization process (i.e., nucleation and growth) into a prec ursor process that induces a highly hydrated amorphous phase. The negati vely charged polymer sequesters a high concentration of ions and inhibits crystal nucleation, allowing liquid liquid phase separation of droplets of a metastable precurs or to form within the crystallizing solution. These nanoscopic droplets, which in some cases, can grow to a size of several microns, usually settle on the substrate and coalesce into amorphous films. These films subsequently transform, through solidification and crystalliza tion, into crystalline minera l films of cal cite, vaterite1, or aragonite39, in the calcium carbonate system. Depending on how the precursor droplets are deposited, other non-equilibrium morphologies can be fo rmed as well, such as nanofibers5,6, helices1, templated films3, and molded crystals4. Barium carbonate mineralization has received in creased interest recently due to several important applications, such as the production of barium salts, pi gment, optical glass, electric condensers and as a precursor for the production of superconductors such as barium ferrite69, piezoelectric and ceramic materials70. Barium carbonate has been produced with interesting morphologies using a number of different met hods. For example, long, polycrystalline and mosaic fibers have been synthesized in the presence of double hydrophilic block copolymers45,49.
72 Globular aggregates, twisted shee ts, and helicoidal filaments have been grown in sodium metasilicate gels48. Reverse micelles have produced long fibers of barium carbonate73, and several mixed solvent and/or template methods ha ve also been employed to stabilize different barium carbonate morphologies, i.e. nanorods via a mixed solvent method75, and peanuts, rods, ellipsoids and dumbbells with a polyvinylpyrole template method74. In our previous work, described in Chapter 3, we discovered that BaCO3 fibers could be grown on calcite seed crystals, which we ha d presumed occurred via the same solutionprecursor-solid (SPS) mechanism we had reported on for CaCO3. The SPS mechanism relies on a PILP flux droplet, in analogy to the molten me tal flux droplet that provides one-dimensional growth in the vapor-liquid-solid (VLS) and solution-liquid-solid (SLS) systems5. On the other hand, Colfen and coworkers have grown BaCO3 fibers using a racemic block copolymer, and they suggest that the fibers are form ed by a mesocrystal assembly mechanism45. An important distinction between these tw o proposed mechanisms for mineral fiber formation would be the fluidity of the precursor phase. Therefore, in order to establish that a fluidic amorphous precursor is generated in our BaCO3 system, and to demonstrate that the PILP process is valid for crysta l systems other than CaCO3 and calcium phosphate (CaP), we methodically investigated the use of poly(acrylic acid) as a polymeric proc ess-directing agent for barium carbonate, without the confounding issue of the presence of calcite seed crystals which were used in our prior study. Evidence of the PILP process mi ght include PILP droplets in solution, films of coalesced droplets on the substrate, or molten morphologies, as found in our previous work2-4. Another group IIA, alkaline earth carbonate, strontium carbonate (SrCO3), was also investigated for potential compatibility with the PILP process. Both BaCO3 (witherite) and
73 SrCO3 (strontianite) have only one polymorph, whic h in both cases, has an aragonitic crystal structure71. Strontium carbonate has many applications for example in cathode ray tubes for televisions and computer monitors, use in fi reworks and pyrotechnics, as an additive for specialty glass, a component in ferrite magne ts, and as a precursor for various strontium compounds, including SrTiO3, which is an important piezoelectric material76. Recently, SrCO3 has been used in several biosensor77 and phosphor78 applications. A variety of SrCO3 structures, including spheres, rods, whiskers and ellipsoids64, fibers10, ribbons79, needles80, wires81 and hexahedral ellipsoids82 have been prepared via several methods, including reverse micelles64,80,83, solvothermal methods64,82,84, self assembled monolayers10,85 and at liquid-liquid interfaces79,86. The expansion of the PILP process into two non-calcium based systems, BaCO3 and SrCO3, further shows the versatility and robustness of the PILP process. One of the long range goals of biomimetic research is to take general principl es that are learned from biomineral systems and apply them to non-biological materials; thus th e non-specificity of this unusual mineralization process is appealing. Materials and Methods PILP Droplet Collection To generate and collect PILP dropl ets, a 15 m l solution of 6 mM BaCl2 or SrCl2 and 100 g/ml poly(acrylic acid) (PAA) (Aldrich, Mw 5100 g/mol) was placed into a 25 ml glass vial. Approximately 1.5 g of NH4CO3 was crushed and then placed in Publix plastic wrap, and placed below the lid of the vial. The vial was capp ed, and the reaction left at room temperature. When the solution became cloudy, a sample of the solution was pipetted out and placed on a freshly cleaned glass slide and examined w ith polarized light microscopy (PLM). The experimental procedure is shown in Figure 4-1.
74 Dynamic Light Scattering Light scattering experiments were carried out using a NanoSightTM LM20 dynamic light scattering instrument, equipped with Nanoparticle Tracking Anal ysis (NTA) software. Samples were prepared using the same procedure descri bed for the PILP droplet collection experiment, with concentrations of 6 mM BaCl2 and 20 g/ml PAA (Aldrich, Mw 8000 g/mol). The reaction was allowed to proceed for 2 minutes, and then so lution was pipetted out from near the air-water interface and analyzed. The ammonium carbonate powder was removed to quench the reaction. Crystal Morphology Experiment Crystallization was carried out using the amm onium carbonate vapor diffusion method, as described previously2. Specifically, glass cover slips clean ed using Nochromix were placed in a 33 mm polystyrene petri dish (F alcon) containing 3 ml of crys tallization solution or a 100 mm polystyrene petri dish (Falcon) with 4 compartments, each containing 5 ml of crystallization solution, composed of BaCl2 or SrCl2 (Aldrich) and poly(acrylic acid) (Aldrich), with MW of 5100, 8000 or 15000 g/mol, in nano pure water (res istivity: 18.1 ). All reagents were used as received without further purificati on. The petri dishes were covere d with parafilm, and placed in a chamber which contained freshly ground ammonium carbonate (Sigma), also in a petri dish, covered by parafilm. Four needle holes were pu nched into the film cove ring the crystallization solution, and eight in the ammonium carbonate covering. The r eactions were run at room temperature (~ 25C) for four days, at which time the glass cover slips were removed from solution, gently rinsed by dipping in water and ethanol to remove excess salt solution, and air dried. Samples were first characterized by pol arized light microscopy using an Olympus BX60 polarized light microscope (PLM) with a first-order red (gypsum) -plate, and then electron microscopy using a JEOL 6335f field emission s canning electron microscope (SEM) at 15 kV, and a JEOL 200CX transmission electron microscope (TEM) at 200 kV; and crystal phase
75 determined by an Philips APD 3720 x-ray diffractometer (XRD), step size 0.02 1.250 sec/step over a 2 range of 18-60 Results and Discussion Evidence of BaCO3 and SrCO3 PILP formation is shown in Figure 4-2. In Figure 4-2A, an am orphous film made of partia lly coalesced droplets of BaCO3 precursor was extracted after 30 minutes of reaction from the vial as described in the first experimental section above. In Figure 4-2B, taken after two hour s of reaction, a sim ilar film appears birefringent, indicating that the amorphous precursor phase has crystallized. Simila rly, a partially birefringe nt film of partially coalesced droplets of SrCO3 was imaged after 2. 5 hours of reaction ( Figure 4-2C). After 3 hours, Figure 4-2D, a similar film appears fully birefr ing ent, and spherulites and small crystalline droplets are also found. The presence of early stage PILP droplets in the BaCO3 system was also confirmed via dynamic light scattering. This da ta can be found in the supporting information. Several different BaCO3 morphologies were produced by varying the concentration and molecular weight of PAA. Figure 4-3 shows examples of th e m orphologies found. A complete gallery of images of all the experimental c onditions can be found in the supporting information, Figure 4S-1. Four main types of BaCO3 morphologies were observed. At low concentrations of PAA, spherulitic aggregates were the main component ( Figure 4-3A, E). At increasing concentrations of PAA, fibers tens of m icrons in length and le ss than 500 nm in diameter were found which appear to be of the same morphology as the fibers described in Chapter 3. In addition, cone-shaped fibrous bundles were f ound at higher concentrations which closely resemble the BaSO4 cone-shaped particles described by Qi et al.68 and Yu et al.66 ( Figure 2-14). Like Qi and Yus bundles, these al so appear to have a single ex tinction direction, and, at higher molecular w eight PAA (see supporting information), a self-replicating struct ure is formed. At
76 the highest PAA concentrati ons investigated here, thic k, continuous films of BaCO3 formed at the air-water interface. Th e film patches shown in Figure 4-3 are of fractured film pieces that settled on th e bottom of the petri dish. Several different SrCO3 morphologies were produced by varying the concentration and molecular weight of PAA. Figure 4-4 shows examples of th e m orphologies found. A complete gallery of images of all the experimental c onditions can be found in the supporting information, Figure 4S-2. Four main types of SrCO3 morphologies were observed. At low concentrations of PAA, either cone-like aggregat es or spherulitic aggregates were the main component (Figure 44A, B, E, and F). The cone-like structures found in Figure 4-4A are not birefringent in their central region, but are shown via SEM ( Figure 4-4E) to be continuous. The lack of birefringence is caused by the orientation of th e cone-like structures, in that th e iso tropic axis is perpendicular to the plane of the substrate. The small amount of birefringence visible in some parts of the cones is due to either lattice stra ins of slight changes in orienta tion. At increasing concentrations of PAA, fibers tens of microns in length and le ss than 300 nm in diameter were found that appear to be similar in morphology to the BaCO3 fibers ( Figure 4-3C, G), although these fibers appear more twiste d, as compared to the relatively straight BaCO3 fibers. At the highest PAA concentrations and low PAA Mw, thick, continuous films of SrCO3 formed at the air-water interface. The fibrous BaCO3 and SrCO3 crystals shown in Figure 4-3B and Figure 4-4C appear to grow random ly off of an underlying aggr egate or film, as can be seen in Figure 4-5. The PLM im ages of both the BaCO3 and SrCO3 fibers in Figure 4-5A&C clearly show that the fibers are growing out into solution off of an aggregate or central nucleation point on the gla ss slide. This is confirmed by the SEM imaging ( Figure 4-5B&D). The fibers in the SEM images appear as
77 dense mats on the surface because they collapse during drying due to capillary forces. In both cases, the fibers appear to nucleate from a si ngle site, be that an aggregate, as in Figure 4-5A&D, or a sm aller object on the slide, as in Figure 4-5B and the smaller fibrous regions in Figure 4-5C. However, the seed was not di rectly added, as in our CaCO3 fiber studies; but probably resulted from an earlier deposit. All of the morphologies shown in Figure 4-4 and Figure 4-5 and in the supporting inform ation were shown via XRD to be the phase witherite, which is the only phase of BaCO3, or strontianite, the only phase of SrCO3. This information is shown in Figure 4-6. As compared to the control reaction, the peaks fr om the reactions containing P AA are broadened. An amorphous hump is visible in many of the conditions, but it cannot be said whether this is from remnant amorphous mineral, or simply due to the gla ss slide on which the sample was mounted. The broadened peaks are generally considered indi cative of small crystal size, according to the Scherer equation, although lattice di stortions are also a possibility with this precursor system. The fiber and film morphologies were also examined in the TEM, as shown in Figure 4-7. Figure 4-7A illustrates the granular texture of the BaCO3 films. As seen in Figure 4-7B&C, two distinct m orphologies of fibers we re present, a straight fiber ( Figure 4-7B), which has a singlecrysta lline SAED pattern (inset ), and an aggregate fiber, wh ich has a polycrystalline pattern (inset) ( Figure 4-7C). This pol yc rystalline pattern in Figure 4-7C is an arc pattern, rather than continuous rings, which indicates that there is a prefe rred orientation of the BaCO3 crystals. Both types of fibers were also reported in our previous work on BaCO3 fibers grown on calcite seed crystals, detailed in Chapter 3. A granular SrCO3 film is shown in Figure 4-7D. This film appears to be porous and is less continuous than the BaCO3 film analyzed in the TEM, although both displayed a polycrystalline texture in SAED. The SrCO3 fiber shown in Figure 4-7E&F is
78 of similar size to the BaCO3 fiber in Figure 4-7B, however, it displays an arc pattern in electron diffraction. The arcs in this pattern are m uch less broad than in the pattern of the BaCO3 aggregate fiber, which indicates that the fiber is comprised of well-aligned domains. This was confirmed via dark field analysis, Figure 4-7F. This dark field image was constructed using the 021 beam, and reveals the nanodomain structure of the fiber. In light of the nanodomain texture found in the BaCO3 fibers in Chapter 3, the nanodomain texture of the SrCO3 fibers is not surprising; however, the domains appear to be less well aligned, as shown by the arc SAED pattern, as compared to the seemingly single-crystalline SAED pattern of the straight BaCO3 fiber in Figure 4-7B. The nanodomain texture of the BaCO3 fibers was only revealed via HRTEM analysis in our prior report. The SrCO3 films also appeared to have a more coarsegrained texture, as seen in the TEM and SEM images in Figure 4-7D and Figure 4-4H, respectively. Minerals with fibrous m orphology have been previously reported in both the BaCO3 5,45 and SrCO3 10 systems, which were suggested as be ing formed by a process of oriented attachment73 and mesocrystal assembly,45 in the case of BaCO3, and via templating by polymer chains,10 in the case of SrCO3. Similar to our previous report on CaCO3 fibers, we believe that the BaCO3 and SrCO3 fibers grown in this study also form via a Solution-Precursor-Solid (SPS) mechanism.6 Although seed crystals were not used in this study to in itiate the fibers, it appears that small particles, films, or aggregates formed first, which then initiated the fiber formation process (i.e., collection of a flux droplet which causes one-d imensional growth). The BaCO3 fibers shown in our Chapter 3, which were nucleat ed on calcite seed crystals, were found to be comprised of many ~5-20 nm nanocrystalline do mains in HRTEM, in spite of exhibiting a single-crystalline SAED pattern. On the other hand, our groups first report on calcite fibers
79 formed by the SPS mechanism discussed the sing le crystalline nature of the fibers, as demonstrated by the SAED spot patterns, and a singular extinction direct ion in cross-polarized light. It is possible, however, th at the calcite fibers may also have nanocrystalline domains, as they were too thick to be examined usi ng high resolution TEM analysis. The nanodomain texture further reinforces the biomimetic nature of the PILP process, as a nanogranular texture has also been shown to exist in s ponge spicules and sea urchin spines51,52,98. These biominerals are now known to be form ed via an amorphous precursor pathway, but it remains to be determined if the amorphous pha se has fluidic character associated with the PILP process (we argue that this provides the co alescence needed to build such coherent singlecrystalline structures). The na nogranular texture of the urchin was surprising because a single optical extinction direction and spot diffraction pattern in XRD are the trademarks of this biomineral. Similarly, the SrCO3 fibers grown in this study appear to behave optically as single crystals, as shown in Figure 4-8A, yet have an arc pattern in SAED (Figure 4-7E). At higher m agnification, however, it can be seen that the longer SrCO3 fiber in Figure 4-8A exhibits a gradual shift in orientation near the end. In addition, a colloidal surface text ure is also observed at high m agnification in SEM, Figure 4-8B. This suggests that the SrCO3 fibers, like the BaCO3 fibers, and possibly CaCO3 fibers, grown via the SPS mechanism are in fact mesocrystals. According to Colfen7, mesocrystals can form via several mechanisms ( Figure 2-8), the most fa mous being the oriented assembly of discreet nanoparticles that are aligned via directional, physical fields, which requires that the nanoparticles be anisotropic. Some examples of the forces that may cause th is type of mesocrystal assembly include polarization forces, as well as electric, magnetic and dipole fields. The an isotropy could either be a inherent property of the material itself, for example a dipole moment along one crystallographic axis or oppositely
80 charged faces of a crystal, or be induced through the adsorption of an additive. Identical faces of the crystal may also come into alignment when brought in cl ose contact via van der Waals attractions. A second formation mechanism, wh ere entropic forces cause the alignment of growing anisotropic particles in a confined space, has also been proposed. A third mechanism for mesocrystals arises from mineral bridges th at connect individual nanoparticles and cause a common crystallographic alignment. This mechanism applies when amorphous intermediates are present, and seems most pertinent to our sy stem. When there is a high concentration of amorphous intermediates, previously nucleated stru ctures may continue to grow by the continued addition of colloidal particles from solution, whic h would be PILP droplets in our system. Upon attachment, the amorphous particle is restructured as it crystallizes via isoepitaxy to match the crystallographic orientation of th e underlying crystal substrate. Mesocrystal formation via the assembly of amorphous, liquid-like colloidal droplets was reporte d by Ma et al. for DL-alanine99, and is discussed in a recent book by Colfen and Antonietti8. Mesocrystal textures have been found in a wide variety of biological systems, including sea urchin spines51,52, spicules32, and aragonite tablets in nacre53, which all exhibit well-aligned na nocrystalline domains that have organic material both occluded within and surrounding the crystals. The third mesocrystal formation mechanism can explain the mesocrystal texture, but does not address what leads to the initial formation of the anisot ropic assembly. The proposed SPS growth mechanism helps to build upon the meso crystal scenario. In the SPS mechanism, the amorphous precursor is thought to co llect to a size that forms a flux droplet, or catalyst, for fiber growth. This flux droplet provide s a means for generating the anis otropic collection of precursor phase, which is otherwise difficu lt to argue being that neither the amorphous nanoparticles or the polyaspartate additive seem to provide a polarizin g force that could delive r anisotropic assembly.
81 This is analogous to th e VLS and SLS mechanisms55,59, which also lead to anisotropic onedimensional growth that is limited in dimension by the size of the flux droplet. This droplet, which is molten metal in these systems, is enriched with reactants that become saturated. By analogy, the PILP flux droplet in our SPS mechanism is inherently saturate d with ionic species; but the fibers do not necessarily crystallize as th ey grow from the flux droplet, but could simply solidify into the solid amorphous phase. An amorphous fiber could subsequently crystallize into a highly oriented mesocrystal through mineral br idges along the preformed fiber. The polymer content is so low in our system that we have not generally considered the connections as mineral bridges per say, but rather have described the long-range crystallographi c order as being caused by crystal growth across the co alesced droplets of amorphous pr ecursor. Thus, the nanodomain texture may simply be a remnant of the colloidal droplets that were either partially solidified, or capped by small amounts of polymeric impurity. Another potential cause of nanodoma in texture within an otherwis e coherent crystal is that there could be a breakdown of th e lattice structure from the cons iderable shrinkage that must occur during dehydration of the hydrated amorphous precursor. Shrinkage defects have been observed in mineral films deposited by the PILP process, such as biaxial strains1. For the case of the fibers here, there appears to be a gradual shift in birefringence along the length of the SrCO3 fibers ( Figure 4-8A), which may be related to the gradual buildup of latti ce strain as the crystallization proceeds over such large di stances along th e length of the fibers. The mechanism proposed here is markedly diffe rent from the mechanism proposed by Balz et al.10 for SrCO3 fibers grown in the presence of P AA on a self-assembled monolayer template ( Figure 2-15). These authors propose that PAA chains, when in the presence of Sr2+, aggregate into long strands and adsorb onto the SAM surf ace. These PAA strands are then proposed to
82 serve as a template for fiber growth. The evid ence used by these authors in support of this mechanism includes a colloidal surface texture on th e fibers (particles ~30 nm in diameter), the interwoven, serpentine appearance of the fibers on the substrate after drying, the presence of 0.2% organic content in the fibers found via DSC, the lack of preferred fiber orientation, and the selective adsorption of material on to the SAM only when both PAA and Sr2+ are in solution. We note that the evidence cited in support of their mechanism is also compatible with the SPS mechanism. The colloidal surface texture has be en observed in many PILP generated minerals, including the SrCO3 fibers shown in this report, and CaCO3 films of both calcite and aragonite3,39. In addition, the serpentine growth of CaCO3,, BaCO3 and SrCO3 fibers is both observed here and has been previously reported5,6,14. Organic material is also known to be occluded in crystalline struct ures grown via the PILP process, as evidenced by DSC100 and fluorescence observation101. PILP droplets have been shown by Kim et al.3 to preferentially deposit on carboxy-terminated SAMs, which would lead to the type of QCM data shown by Balz, where the deposition of ma terial onto the SAM surface was only observed when both PAA and Sr2+ are in solution. In addition, a preferred orientat ion of the fibers is not expected in the SPS mechanism, as the initial crystallite, which may be randomly or iented relative to the fiber growth direction, would dictate the overall orientati on of the resulting fiber. Alt hough we have seen some fibers that appear to have an epitaxial relationship with the seed substrate6, this does not seem to be a necessity, as in the VLS/SLS systems. Given that bent CaCO3 fibers can still be single crystalline, this seems to suggest that crys tallization often procee ds across a preformed amorphous fiber. Therefore the seed substrate do es not necessarily cause the crystallization of
83 the fiber, but somehow triggers the collection of the precursor phase into a flux droplet, which then catalyzes one dimensional fiber growth. In the mechanism proposed by Balz et al.10, it is not clear how a small number of PAA chains, which do not generally have a great deal of structure in ionic solutions, especially at low molecular weight (as in this re port), would be capable of the large scale assembly needed in order to template mineral fiber formation. In particular, the authors claim that the ligand stabilized aggregates agglomerate at random, so it is not clear how this mechanism could lead to such highly anisotropic growth, particularly since the building bloc ks do not appear to be highly anisotropic, as judging by their AFM images. Ther efore, we believe that it is more likely that these polymers behave as described in our previous reports1-3,6,14,39,43, sequestering ions in solution, leading to the formation of a highly co ncentrated amorphous phase which then acts as a flux droplet for one-dimensional fiber growt h. Although our studies cannot yet confirm the SPS mechanism for fiber growth, they do support th e possibility of formation of a fluidic flux droplet, which is known to be a mechanism leadin g to anisotropic fibrous structures. Our goal here was to demonstrate the PILP process in ot her non-calcium based systems, and more focused studies are needed to fully resolve the mechan ism(s) leading to these interesting fibrous structures. Conclusions Barium and strontium carbonate are both importa nt materials, both in their own right and as precursors to functional materials such as barium or strontium titanate. Novel aqueous methods of producing unique mor phologies of barium and strontium carbonate, such as fibers and films, are of great interest and importance. This work has demonstrated the robustness and versatility of the PILP process, and its exciti ng applications to new material systems.
84 Supporting Information Light scattering data an d full galleries of BaCO3 and SrCO3 results are shown in this section. Light scattering results also showed the development of small PILP droplets at an early stage in the reaction for BaCO3. After only 2 minutes of reaction time at room temperature, particles as small as 10 nm were detected. At th is time, 50% of the particles in solution were under 200 nm in size, and the average size being 250 nm. After allowing the solution to incubate (without additional CO2 diffusion) for 25 min, the average pa rticle size grew to 340 nm, with 50% of the particles sized under 380 nm. A video of the particles in solu tion can be found in the supplemental information after 2.5 min of react ion, and no incubation time, is shown in the supporting information, at th e end of the dissertation.
85 6 mM BaCl2100 g/mL PAA (NH4)2CO3 Time Ba2+Ba2+Ba2+CO3 2-CO3 2-CO3 2Ba2+Ba2+Ba2+CO3 2-CO3 2-CO3 2-Ba2+Ba2+CO3 2-CO3 2Figure 4-1. Experimental set-up for BaCO3 PILP droplet collection experiment. 15 mL of 6 mM BaCl2, 100 g/ml PAA (Mw = 5100 g/mol) is pl aced in a 25 mL glass vial. Crushed ammonium carbonate is placed in PublixTM cling wrap. The plastic wrap containing the ammonium carbonate is placed into the vial, which is then capped. The reaction then takes place at room temperature until the solution becomes cloudy, signifying the formation of PILP droplets (inset). At this point, the cap is removed and an aliquot of the PILP containing so lution is removed and placed on a glass slide for PLM analysis.
86 AB CD Figure 4-2. PLM images (with gypsum -plate) of BaCO3 and SrCO3 PILP phase collected from the reaction solution at early time points. A) An amorphous film-like patch made from coalesced BaCO3 PILP droplets extracted afte r 30 min of reaction at room temperature. B) A crystalline film from BaCO3 PILP phase at 2 hours reaction time. C) Partially crystalline SrCO3 film that appears to be co mprised of partially coalesced droplets (2.5 hours reaction time). D) Crysta lline film, crystalline spherulite (with Maltese cross), and many small droplets of SrCO3 (3 hours at RT).
87 AB D C EF H G 10 m 10 m 10 m 10 m Figure 4-3. PLM (A-D) and SEM (E-H) images of BaCO3 structures produ ced via the PILP process using 6 mM BaCl2 and PAA with molecular wei ght of 5100 g/mol. Scale bar in SEM images is 10 m; in PLM images, 100 m. A, E) Large, spherulitic aggregate structures, as well as small s pherulites with typi cal sheath of wheat morphology; 10 g/ml PAA. B, F) Many sma ll diameter fibers, hundreds of m in length, 50 g/ml PAA. C, G) Horsetails th at have a single extinction direction in PLM, 100 g/ml PAA. D, H) Grainy film s formed at the air/ water interface, 500 g/ml PAA. When viewed in zoom mode one can see a dense-packed array of acicular crystals that appear to be emerging from the grains in the film.
88 200 nm 200 nm 1 m 1 m ABCD EFGH Figure 4-4. PLM (A-D) and SEM (E-H) images of SrCO3 structures produced via the PILP process using 6 mM SrCl2 and PAA with molecular weight of 5100 g/mol or 8000 g/mol. Scale bar in PLM images is 100 m. A, E) Cone shaped, self-similar structures, 10 g/ml PAA, Mw 8000 g/mol. SEM scale bar = 100 nm. B, F) Large spherulitic aggregates, 50 g/ml PAA, Mw 5100 g/mol. SEM scale bar = 1 m. C, G) Many small diameter fibers, hundreds of m in length, 50 g/ml PAA, Mw = 8000 g/mol. SEM scale bar = 100 nm. In the PLM image, the dark patches at the top right are composed of a thick mat of fi bers that had collapsed upon drying. The SrCO3 fibers are much smaller than the BaCO3 fibers shown at the same magnification in Figure 4-3B above. D, H) Thick rough film s with a coarse-grain structure formed at the air/water inte rface, 500 g/ml PAA, Mw 5100 g/mol. SEM scale bar = 1 m.
89 AB DC D C Figure 4-5. Barium and strontium car bonate fibers. A) PLM image of BaCO3 fibers in solution growing off an aggregate (dar k central patch). The fibers are curved and serpentine, and appear to grow at random orientations and angles off of the aggregate out into the solution. B) SEM image of fibers, which appe ar to have nucleated off of two regions near the center of the image, and then grown randomly outward. They collapse down to a dense mesh due to capillary forces during drying. C) PLM of SrCO3 fibers, which appear to nucleate both o ff of aggregates (two large patches in the center of the image) and in smaller, isolated patches on the glass slide. D) SEM image of the fibers, which nucleated off of something in the central area in the image, which is now covered by fibers due to drying.
90 Figure 4-6. XRD results for various conditions showed less resolved, broadened peaks for all conditions relative to the control reaction (without polymer) for both BaCO3 (left) and SrCO3 (right). All conditions contained either 6 mM BaCl2 or 6 mM SrCl2, and PAA with Mw = 8000 g/mol was used for all conditions containing polymer. Conditions: 1) control reaction with no PAA; 2) 10 g /ml PAA; 3) 50 g /ml PAA; 3) 100 g /ml PAA; 4) 500 g /ml PAA; 5) 500 g /ml PAA.
91 Figure 4-7. TEM images of some BaCO3 and SrCO3 morphologies produced via the PILP process. A) High magnification image of BaCO3 film, which corresponds to Figure 4-3D. Scale bar = 200 nm. B) TEM im age of BaCO3 fibers, which correspond to Figure 4-3B. Inset: SAED pattern with d-spac ing s that match that of witherite. Scale bar = 500 nm. C) High magnification image of aggregate BaCO3 fiber. Scale bar = 500 nm. D) High magnification image of SrCO3 film, which corresponds to Figure 44D. E) Bright field and F) Da rk field im age of a single SrCO3 fiber. Dark field image constructed from 021 reflection (inset, E. d-spacings match those of strontianite). S cale bar = 100 nm.
92 100 nmAB Figure 4-8. A) PLM analysis of SrCO3 fibers. This polarized optical micrograph (with gypsum wave-plate) shows two birefringent fibers in the center, oriented 90 to each other. The largely uniform colors of the fibe rs suggest a uniform crystallographic orientation. However, upon closer inspection of the longer fiber (and others), there appears to be a gradual shift in orienta tion along the length of the fiber. B) SEM micrograph of SrCO3 fibers with colloi dal surface texture.
93 Figure 4S-1. SEM images of all experimental conditions for BaCO3. A) PAA Mw = 5100 g/mol. B) PAA Mw = 8000 g/mol. C) PAA Mw = 15000 g/mol. For the 10 g/ml PAA, spherulitic aggregates were the ma in reaction product. This morphology was also found for 50 g/ml PAA (MW 15000 g/mol). Fibers with a small (<100 nm) diameter were found for 50 g/ml PAA (Mw 5100 and 8000 g/mol), while larger, polydisperse fibers were found for Mw 8000, 100 g/ml condition. Cone-like fibrous bundles were the product for Mw 15000, 100 g/ml as well as 50 g/ml, Mw 15000. Films of BaCO3 were present for the highest con centration of PAA (500 g/mL) and 5100 g/mol molecular weight. A mixture of spherulites, films and thick fibers appeared for the 500 g/ml, 8000 and 15000g/mol molecular weight conditions. Scale bars = 10 m.
94 Figure 4S-2. SEM images of all experimental conditions for SrCO3. A) PAA Mw = 5100 g/mol. B) PAA Mw = 8000 g/mol. C) PAA Mw = 15000 g/mol. For the 10 g/ml PAA, spherulitic aggregates were the main reaction product for Mw 5100 and 15000. For Mw 8000 g/mol, self-similar cones were found. Spherulites were also found for the 50 g/ml PAA (MW 5100 and 15000 g/mol). For 50 g/ml Mw 8000, patches of small fibers were found. At 100 g/ml PAA, thicker fibers were present for Mw 8000 and 15000 g/mol, while films of SrCO3 were present for both 100 and 500 g/ml PAA, Mw 5100 g/mol. A mixture of small particles, sp herulites and some granular films were found at 500 g/ml PAA, Mw 8000 and 15000 g/mol. Scale bars = 10 m.
95 CHAPTER 5 MECHANISITIC STUDY OF THE GROWTH OF MINERAL FIBERS IN THE PRESENCE OF POL YACRYLIC ACID The formation mechanism of mineral fibers grow n via the PILP process is investigated in three crystal systems: BaCO3, SrCO3 and CaCO3. CaCO3 fibers with aragonitic crystal structure were synthesized for the first time. AF M, SEM and TEM were employed to study the morphology and structure of the fibers generated in this study, and the na nogranular texture of these fibers, first discovered in the BaCO3 system, was confirmed to also be present in the CaCO3 and SrCO3 fibers. Fluorescence studies were done to determine the distribution of occluded polymer throughout the fibers, and a mechanism for fiber formation is proposed. Introduction Biom inerals with fibrous morphology are prevalen t in nature; examples include sea urchin teeth and vertebrate enamel. Recently, several researchers13,22,24,32,35,94,102 have shown that many biominerals form via an amorphous precursor phase. Of particular relevance here, Ma et al.13 have recently shown that the teeth of the s ea urchin are formed via a transient amorphous precursor phase. These results are similar to earlier studies on sea urchin spicule formation21 and spine regeneration35, indicating that the entire skeleton of the sea urchin forms via a transient amorphous phase. Our group is interested in biomim etic crystallization processes2,14 and has proposed that the polymerinduced liquid-precursor (PILP) process may play a fundamental role in the morphogenesis of calcitic biominerals. In the PILP process, described in detail in our previous report,2 a highly hydrated amorphous mineral precurs or is induced by using short-chained, polymeric process-directing agents, such as the polyanionic salt of polyaspartic acid (poly-Asp) or poly(acrylic acid) (PAA). In our previous work, detailed in Chapter 3 and 4, the nonspecificity of the PILP process was demonstrated via the synthesis of complex structures of
96 barium and strontium carbonate, including films, fi bers and horsetails. In addition, witherite fibers, which have an aragonitic crystal structur e with an orthorhombic lattice, were found to nucleate on calcite seed crystals, which have a hexagonal lattice stru cture. In our initial report, where we also found that the PILP process can lead to CaCO3 product with a fibrous morphology, which suggested growth via a So lution-Precursor-Solid (SPS) mechanism6. In addition to the SPS mechanism, several other techniques have been successfully employed to synthesize mineral fibers. The reve rse micelle technique has successfully produced single-crystalline fibers of a variety of materials. The harsh chemicals and high temperatures required ( 80C), however, make this type of proce ss less useful for biomimetic applications, which capitalize on the benign proc essing conditions that allow fo r incorporation of biological components. In addition, biomineral systems are able to regulate w ith precise control the organization of their fibrous elements, which is a capability lacking in materials engineering. Mesocrystal assembly has also been propos ed as a formation mechanism for several fibrous minerals, including BaSO4, BaCrO4 66,67 and BaCO3 45 ( Figure 2-14). Polyacrylate (as in PILP) and double-hydro philic block copolymers, which are surface active, were used to facilitate the fiber formation. Many of these fibers were bundled together, bran ching out from a single nucleation point. Often, this bundle morphology wa s self-replicating, and large superstructures composed of many bundles of nanofib ers were the result. This pro cess is described in detail in Chapter 2. Both the block copolymer and SPS mechanisms utilize aqueous solutions with polyanionic additives and low te mperatures for crystallization; however, the SPS mechanism is capable of producing fibers that appear to be single crystalline6, while the block copolymer approach produces fibers made of nanoc rystalline self-assembled aggregates45.
97 Other authors have also proposed formation mech anisms for fibrous minerals. Balz et al.10 synthesized SrCO3 fibers in the presence of PAA on a self assembled monolayer template. These authors propose two different possible fiber formation mechanisms. The first is that PAA chains aggregate into long strands when deprotonated and bound with Sr2+ ions, which then adsorb onto the SAM surface. These PAA strands ar e then proposed to serve as a template for fiber growth. The other is that PAA/SrCO3 aggregates form in solution, and then deposit onto the SAM. Recently, Tao et al.11 have produced mineral structures which appear to form via a mesocrystal-type assembly process, with the aid of an amorphous phase postulated to have liquid-like character. Here, various solution ad ditives, including PAA, glycine, glutamic acid and amelogenin were used in the crystallization of hydroxyapatite. These authors propose that HAP nanoparticles, produced in th e presence of PAA, and have a HAP core/ACP shell structure, align in linear chains in the pr esence of Gly. The alignment is dramatically enhanced with amelogenins, the primary protein in dental enamel. The partic les in the chain are randomly oriented and linked by the surround ing ACP. Over time, the AC P transforms to HAP as the primary particles jiggle around in the amorpho us phase to align themselves in a common direction. Both the SPS mesocrystal assembly and the ACP alignment mechanism of Tao11 take advantage of amorphous mineral precursor partic les formed/stabilized in solution by acidic polymers. Qi et al.68 and Balz et al.10 argue in favor of selectiv e adsorption of polymers onto specific mineral faces as the driving force for fiber formation. Tao et al.11 argue that some fluidic character of the ACP may enable the a ssembly and realignment of primary nanoparticles in a linear assembly, after which it crystallizes, yielding single crys talline rods. In our previous work6, we have argued that an amorphous mineral flux droplet directs fiber formation, in an
98 mechanism analogous to the VLS and SPS mechanisms also highlighted in Chapter 2. In all cases, polyacrylic acid with a lo w molecular weight (2000-5000) was us ed, so it is likely that the mechanisms for fiber formation in these experiments are similar. The formation of mineral fibers or rods via an amorphous precursor route has been demonstrated by various authors in several mine ral systems. Although th e mechanism of fiber formation may be slightly different for each mineral, we believe that many commonalities, such as the importance of an amorphous precursor and/or the assembly of primary nano-scale particles, are present. Here, we will exam ine the formation mechanism for mineral fibers formed with the aid of PAA as a process di recting agent by studying three systems: barium, strontium and calcium carbonate. Materials and Methods Materials All chem icals were reagent grad e and used without further puri fication. Stock solutions of CaCl2H2O (Sigma), BaCl2 (Fisher), SrCl2 (Sigma) and polyacrylic ac id sodium salt (Aldrich; Mw = 5100 and 8000 g/mol) were filtered three times through 0.22 m Millipore syringe tip membrane filters before use. The fluorophore 5-Bromomethyl fluorescein (5-BMF, Molecular Probes, Eugene, Oregon) was dissolved in di methylformamide (Fluka) to be used for fluorescently labeling the polymer ic process-directing agent. Preparation of Seed Substrates Calcite seeds were grow n on glass cover slip s overnight at room temperature in a 60 mm OD Petri dish filled with 3 mL of a 45 mM CaCl2 (dihydrate, Sigma) so lution via vapor diffusion of the decomposition products of (NH4)2CO3 (Sigma) (NH3, CO2 and H2O). Substrates were washed in DI water prior to use.
99 Fiber Synthesis Crystallization was carried out using the amm onium carbonate vapor diffusion method, as described previously.2 Specifically, glass cover slips clean ed using Nochromix, or cover slips with calcite seeds, or freshly cleaved geologic calcite (Ward Natural Science, Rochester, NY), were placed in a 35 mm polysty rene petri dish (Falcon) cont aining 3 ml of crystallization solution, or a 100 mm dish with 4 compartments, each containing 5 ml crys tallization solution. The reactions consisted of CaCl2, BaCl2 or SrCl2 (Aldrich) solutions with poly(acrylic acid) (Aldrich), with MW of 5100 or 8000 g/mol, in nanopure wa ter (resis tivity: 18.1 ). The concentrations found to be best for fiber formation are given in Table 5-1. All reagents were used as received without further purification. The petr i dish es were covered with parafilm, and placed in a chamber which contained freshly ground ammoni um carbonate (Sigma), also in a petri dish, covered by parafilm. Four needle holes were pu nched into the film cove ring the crystallization solution, and eight in the ammonium carbonate covering. The r eactions were run at room temperature (~ 25C) or 4C for a period of time ranging from three to seven days, at which time the glass cover slips were removed from solution, gently rinsed by dipping in water and ethanol to remove excess salt solution, and air dried. Characterization The precipitates m orphology and size were ex amined using polarized light microscopy (Olympus BX60), scanning electron microscopy (JEOL 6335f), transmission electron microscopy (JEOL 200CX, JEOL 2010f) and atom ic force microscopy (Digital Instruments Dimension 3100). Chemical composition was analyzed using energy-dispersive x-ray spectroscopy, and crystal struct ure using selected area electr on diffraction (JEOL 200CX, JEOL 2010f).
100 Scanning Electron Microscopy (SEM) Analysis: The samples were fixed to an aluminum stub using double-sided copper tape or colloidal graphite, and then sputter coated with Au/Pd or carbon. The samples were then examined with a JEOL 6335F FEGSEM instrument at an accelerating voltage of 15 kV, equipped with an energy dispersive spectrometer (EDS). Transmission Electron Microscopy (TEM) Analysis: To examine the fibers, the samples were first scratched with a razor blade to dislodge some of the fibe rs from the cover slip or seed. A small aliquot of ethanol was then dispensed on to the scratched area and immediately drawn up using a micropipette. The removed aliquot was then dropped onto a 200 mesh copper TEM grid coated with lacy carbon. The sample was examined on a JEOL 200CX transmission electron microscope at 200 kV in bright field (BF) and selected area el ectron diffraction (SAED) modes, and on a JEOL 2010 TEM at 200 kV. Atomic Force Microscopy: Fiber samples dried on glass cover slips were examined using a Nanoscope 3 scanning probe microscope (Digital In struments). Scan rates ranged from 2 to 5 Hz with 512 sampling points per scan line at room temperature in air. Fluorescence Studies: Polyacry lic acid (Sigma, Mw = 5100 or 8000 g/mol) was incubated with 5-BMF (in DMF) for 1 hour at 60C in ambe r vials to tag some of the carboxylate groups. Ratios of 100:1 and 1000:1 COOH groups: 5-BMF mo lecules were used. Tagged PAA solutions were stored at 4C, wrapped in aluminum fo il to reduce exposure to light. Fiber growth experiments were carried out as previously described, with the addition of wrapping the reaction vessels in aluminum foil to reduce light exposur e. The concentration of PAA used in the crystallization solutions was increased ( Table 5-1) to account for decreased activity of the PAA (due to steric hindrance, decreased electrostatic charge, etc). These conditions are given in Table 5-1. Im aging was done on a Leica TCS SP2 Laser Scanning Confocal Microscope (Leica
101 Microsystems, Heidelberg), using LCS (Leica Confocal Software) Version 2.61, build 1537. A 488 nm laser was used for excitation, a 500 535 nm emission band to detect the fluorescence, and a scan speed of 400 image lines per second. Results and Discussion The first task of this study was to make sure the fluorophore probe did not disrupt the reaction, and that we could therefore reproduce the BaCO3, CaCO3 and SrCO3 fibers that were reported in Chapters 3 and 4, and, in the case of CaCO3, in our groups previous work6. For reference, examples of these morphologies are given in Figure 5-1. The BaCO3 and SrCO3 fibers had similar morphologies to previous experiments. When reproducing the CaCO3 of Olszta et al.6, both with and without fluorophore, we found a new morphology, in addition to the previously reported fibers; the CaCO3 fibers were found to have two distinct morphologies, the thick straight calcite fibers si milar to those reported by Olszta6 (see Figure 5-1B), and thin curved fibers. These results (for fibers grown without fluorophore) are shown in Figure 5-2. The curvy fibers ( Figure 5-2A-C) are smaller ( 200-600 nm ) in diameter than the straight fibers (500 nm 1.5m ) ( Figure 5-2D-F), and they nucleated on th e edges of films and aggregates on the glass slide, rather than nucleating off of the calcite seed rhombs. The curvy fibers, when examined in TEM ( Figure 5-2C) had an obvious granular te xture, and SAED analysis (inset) revealed an arc pattern, indicat ing that the individual dom ains are not random and reasonably well aligned. The most surp rising find, however, was that the d-spacings measured from this SAED pattern matched those of aragonite, not calcite, as was the case in the larger fibers ( Figure 5-2F, inset) and for Olsztas fibers6. The calcite fibers are ge nerally too thick to obtain diffraction from the main middle region. The la rger, straight fibers have a faceted surface texture, and spot SAED pattern for a region along the edge of the fiber ( Figure 5-2F, inset). The
102 facets on the surface are approximately para llel, which, according to Yue and Meldrum103, indicate a single crystal (for the regions where correspondence is retained). This premise is supported by the spot SAED pattern, whic h is indicative of a single crystal. In previous discussions about the formation mechanism, one area that we questioned was the role (or lack thereof) of the seed crystals. For this purpose, we used freshly cleaved geologic calcite crystals, which have an extremely flat surface, as seeds for fiber growth. In this way, we were able to visualize whether or not the fibers grow directly off of a defect site on the seed crystal, as speculated by Olszta6, and if they are epitaxially (or pseudo-epitaxially, in the case of BaCO3 and SrCO3) related to the seed. In the case of the VLS and SLS mechanisms, upon which our proposed SPS mechanism was based, epitaxy plays an important role in stimulating the fiber growth. However, we believe that it is possibl e that in the SPS mechanism, an amorphous fiber may be grown first (via solidificat ion of PILP phase in the flux dropl et rather than crystallization from reactants in the molten metal flux droplet), which then crystallizes. This idea is based on the observation that some fibers were found to re tain uniform crystallogr aphic orientation, even across bends in the fiber. The results of this study are shown in Figure 5-3. In all three systems, the fibers appear to grow off of a film that has deposited on the surface of th e calcite crystal. It had been observed in prior studies that PILP droplets wi ll preferentially adsorb to the mineral surfaces instead of the glass cover slip. In Figure 5-3A, the BaCO3 fibers are emanating from a thick, rough film. For the CaCO3 fibers, Figure 5-3B (these are calcite f ibers, like those in Figure 5-2D), a rough thinner film is evident from the roug h, colloidal surface texture. For SrCO3, Figure 5-3C, a patchy film is present on the surface of the rhomb, and many protrusions and fibers are emanating from this area. The dark, smooth patch in the lower left corner is an area with little to
103 no film. In addition, dehydration and the addition of thermal ener gy from the electron beam can cause the underlying film to delaminate from the underlying calcite rhomb, as shown in Figure 54. These findin gs also help explain why the calcite fibers often appear to have an epitaxial or orientational relationship with the underlying rhomb substrate. As shown in Figure 5-1B and Figure 5-2D, many of these fibers appear to grow in the sam e direction, somewhat perpendicular to the rhomb surface. In the case of BaCO3 and SrCO3, the film that nucleates on the calcite surface is unlikely to crystallize in the same crystallographic orientation as the underlying crystal, as they do not have th e same crystal structure. Thes e films are probably granular and polycrystalline ( Figure 4-7) typical of the aragonitic phase, which commonly breaks down into a spherulitic texture, and would ther ef ore have a variety of orientations. Assuming that the fibers take on the orientation of the unde rlying film, a preferred orientati on of the fibers would not be likely, as the underlying granular film would not have a dominant or ientation. Alternatively, in the case of CaCO3, as the amorphous PILP phase deposits onto the existing calcite crystal, it would likely crystallize via iso-epitaxy. When the fibers then nucleate off of the film, they too would have the same epitaxial re lationship with the calcite seed. However, the calcite fibers do not always have this apparent directionality with the seed, and more ofte n grow into solution in more arbitrary orientations. Granular or colloidal surface textures have been cited by several researchers as evidence of a colloidal aggregation formati on mechanism for mineral fibers10,32. Evidence of a colloidal texture for SrCO3 fibers has been presented previously, in Chapter 4. To compare the surface textures of all three fiber materials, AFM was used. These results are presented in Figure 5-5. The BaCO3 fiber ( Figure 5-5A) has the smoothest surface text ure of the three fiber ty pes studied.
104 It is not completely smooth, however, and has a layered or striated appearance. The CaCO3 fiber, Figure 5-5B, (an aragonitic fibe r like the one pictured in Figure 5-2B) has a rough, som ewhat platy surface, which may be caused by surface recrystallization, while the SrCO3 fiber ( Figure 5-5C) has a colloidal surface, as previo u sly reported. The colloidal texture and arc pattern in SAED of the SrCO3 fibers might be explained by its relatively low solubility (SrCO3:105.6010SPK compared to BaCO3: 92.5810SPK calcite: 93.3610SPK and aragonite: 96.010SPK ). BaCO3 and SrCO3 fibers were formed using the same concentration of Ba2+ or Sr2+ (6 mM). Since SrCO3 is less soluble, the PILP droplets formed in solution are less stable due to the higher relative supersaturat ion of the crystallization solution. Therefore, the SrCO3 PILP droplets are more likel y to be solid amorphous or to have begun to crystallize when the fibers are forming, leading to a collo idal surface texture and arc SAED pattern. The coherence of the calcite fiber is surprising, when compared to the arc pattern and platy surface of the aragonite fiber, given that aragonite has a hi gher solubility. However, calcite typically forms large single crystals in solution, as opposed to aragonite, which usually forms smaller, needlelike crystals, usually in the form of a spherul ite. An analogous situation has been described by Yue and Meldrum103, who studied the single crystallinity of various minerals formed in a hydrogel compartment, and found that the natural size of the crystal und er the given growth conditions dictated the coherence of the final molded crystal. Ther efore, crystals that naturally form larger single crystals woul d be expected to produce larger, more coherent fibers. Since SrCO3, BaCO3 (both aragonitic) and aragonite are a ll needle-like and very susceptible to twinning104, especially along the 110planes105, the natural crystal size is smaller, which likely has a stronger influence on the overall coheren ce of the fiber than sol ubility, especially as these fibers have been found to form only ove r a narrow range of reaction conditions.
105 In Chapters 3 and 4, we have stated that both BaCO3 and SrCO3 are composed of well aligned domains, as they have either an appare ntly single-crystalline spot pattern in SAED (BaCO3) or an arc pattern in SAED (SrCO3) (Figure 4-7). In this chap ter, we found that, like the fibers of Olszta et al.6, the larger calcite fibers ( Figure 5-2F) have a spot pattern in SAED, indicating either a single crysta l or a fiber of extrem ely well aligned domains. The aragonitic fiber ( Figure 5-2C) has an arc pattern in SAED, similar to the SrCO3 fibers described in Chapter 4. In all instances, however, pref erential alignment of the domains was evident. This enabled us to track the changes in orientati on along the length of the fiber, which would tell us if there is a gradual shift in orientation, as was prev iously observed via po larized light for SrCO3 in Chapter 4, for all crystal systems. These results are shown in Figure 5-6. Not surprisingly, given the spot SAED pattern, the BaC O3 fiber showed the least amount of orientation change over the area examined. For the aragonitic CaCO3 fibers, Figure 5-6B, the fiber growth direction changed along the length of the fiber, following the contours of the fiber. In the case of the calcite fibers, which were straight and too thic k to do SAED analysis through the cen ter, they did exhibit a consistent orientation along the edge of the fiber, which is suppor ted by the parallel facets ( Figure 5-2). Interestingly, the growth direction of the left part of the SrCO3 fiber ( Figure 5-6C) continues, with som e slight devi ation, across a 115 bend in the fi ber. The shift in orientation observed here and in Chapter 4 ( Figure 4-8) could be the result of the su bstantial shrinkage that occurs during dehydration of the hydrated amorphous precursor, a nd are likely related to the gradual buildup of lattice strain as the crystallization proceeds over such large distances along the length of the fibers. Similar shri nkage defects such as biaxial strains1 have been observed in mineral films deposited by the PILP process.
106 To differentiate between seve ral different formation mech anisms proposed for mineral fiber formation6,8,10,11, a fluorescent tag was used during fiber mineralization. The results of this experiment are shown in Figure 5-7. To en sure that the signal was from tagged polymer occluded within the mineral fibers, and not fr om tagged polymer adsorbed onto the surface, a control study was done. In this st udy, a sample of fibers from a previous experiment was soaked in a tagged-polymer solution of th e same concentration of those us ed to grow the fibers for 1.5 hours, then rinsed and imaged in the same wa y as the other samples. The fluorescent signal (Figure 5-7D) is greatly diminished in the control sample, confirming that the signal from the other images is due to polymer occluded within the fibers. A the bobble tip was only observed in the calc ite fibers synthesized by Olszta et al.6 (Figure 2-12), and was not observed in the BaCO3 or SrCO3 system; therefore, more conclusive evidence in support of the SPS mechanism6 would be if we were to see limited fluorescent signal from the main part of the fiber, but to see a st rong fluorescence at the end of the fiber, where the proposed flux droplet, or its remnant bobble tip, or its remnant, would be located. Conversely, we would expect the fluorescent signal to remain relatively co nstant throughout the length and width of the fibers if a colloidal aggregation mechanism was responsible, assuming the polymer is attached to the colloids. In all three cases, BaCO3 (Figure 5-7A) calcite (Figure 5-7B) and SrCO3 (Figure 5-7C), the fibers are uniformly bright along their entire length. This result lends credence to a colloidal aggregation theory of fi ber formation, such as the colloidal aggregation mechanism proposed by Balz et al.10 or the oriented attachment/mesocrystal assembly mechanism proposed by Colfen et al.44,99. On the other hand, we belie ve that in the absence of anisotropic particles, and the non-specific bindin g character of the polymer, such an oriented assembly mechanism is not sufficient for explaining the highly anisotropic growth.
107 With the evidence gathered in this study, a new theory for fiber formation via the PILP process has been developed, which encompasse s some aspects of each of the previously suggested mechanisms, including colloidal aggregation and SPS, yielding a more complete explanation of the colloidal te xture and assembly mechanism. An illustration of the proposed mechanism is shown in Figure 5-8. Initially, PILP droplets form in solution as the negatively charged PAA m olecules sequester ions (in this illustration, Ca2+, but the same process happens for Ba2+ and Sr2+). In our fiber experiments, a substrate was placed at the botto m of the petri dish to nucleate the fibers. This substrate could incl ude a plain glass cover slip, a cover slip with calcite rhombs for seeds or fres hly cleaved geologic calcite. As stated by Gower and Odom2, the PILP droplets, which are fluidic, deposit on the underlying substrate and coalesce. PILP structures tend to preferentially form on the calcite seeds, when present. When calcite seeds are not used, PILP films and aggregates are typicall y found on the glass cover slip. This stage is referred to as Stage 1 in Figure 5-8. In Stage 2, P ILP droplets in solution conti nue to fall through solution and preferentially deposit onto the existing structures formed in St age 1. Protrusions and irregularities in the substrate will preferentially a ttract PILP droplets from solu tion due to their higher surface energies. This leads to raised structures, wh ich further raises the surface energy due to the increased surface roughness and sharpness of the protrusion tip. This high energy site on the tip could lead to an auto-catalytic effect for pref erential adsorption of dropl ets at the tip of the protrusion, which with time, will turn into and extend the tip of the fiber in a one-dimensional fashion. The preferential nuclea tion of fibers to high energy site s, such as the edges of PILP films106 and defect sites on seed crystals5,6, has been shown in several of our previous reports. Although the newly adsorbed particles could be either PILP liquid droplets or amorphous
108 nanoparticles, once they have attached onto the tips of the fibers, they coalesce with the underlying structure, creating a new high energy site at the tip, eff ectively creating a sticky end for anisotropic growth. This eff ect would presumably be enhanced by the fluidic character of the PILP droplets. This growth proce ss is described in Stage 3. It is important to note that at any point in the crystallization process, all thr ee stages can be happeni ng simultaneously. In Stage 3, many PILP droplets stick to the amorphous (liquid or solid) sticky tip of existing fiber structures, effectively limiting the diam eter of the fiber to th e size of the tip of the protrusion, which likely is an agglomerate of PI LP droplets, which are still liquid or partially solidified amorphous nanoparticles. This collecti on of droplets is still analogous to the flux droplet described in the SPS mechanism, in th at a highly supersaturated phase accumulates selectively at the growing fiber tip and thus restricts the minerali zation to only occur within this limited region, leading to the one-dimensional growth for which the VLS/SPS/SPS mechanisms are recognized. Judging from the variability in co lloidal textures of the fibers, the flux droplet may contain a bit of both liquid PILP droplets and solid amorphous nanoparticles, where partially solidified droplets may create nanograins suspended within the flux droplet. This could explain why bobble tips are sometimes observe d, but not in all cases, as there may be differing degrees of fluidity of the amorphous precursor for different crystal systems/phases. The main difference between the SPS mechanism and our new mechanism is that we no longer believe that the flux droplet is a molten/liquid that regulates the influx of reactants into the fiber, but that it may be composed of an aggr egate or suspension of nanograins of the amorphous precursor. In Chapter 4, we speculated that the fibers may be mesocrystals due to their colloidal surface texture and arc SAED pa tterns. According to Colfen7, mesocrystals can form via several
109 mechanisms, including the oriented assembly of discreet nanoparticle s that are aligned via directional, physical fields, often via the interactions of adsorb ed additives, and via mineral bridges that connect individua l nanoparticles and cause a co mmon crystallographic alignment ( Figure 2-8). Thus, a mesocrystalline structure can for m by a variety of mechanisms. A mechanism for the crystallizat ion of DL-alanine, proposed by Ma, Colfen and Antonietti99, is most similar to the mechanism we propose in Figure 5-8. According to Colfen, when a high concentration of a morphous intermediates is pr esent, previously nucleated structures may continue to grow by the continue d addition of collo idal particles from so lution (which would be PILP droplets in our system). Upon attachment the amorphous particle is restructured as it crystallizes via isoepitaxy to match the crysta llographic orientation of the underlying crystal substrate. The long-range crystallographic order in our system is caused by crystal growth across the coalesced droplets of amorphous pr ecursor. Thus, the nanodomain texture may simply be a remnant of the colloidal droplets that were either partially solidified, or capped by small amounts of polymeric impurity. This would lead to the so-called mineral bridges described in some reports; but the alternative pe rspective on the same process is that this is simply entrapped polymer. There could also be a breakdown of the lattice structure from the considerable shrinkage that must occur duri ng dehydration of the hydrated amorphous precursor, which is another potential cause of nanodomain texture within the otherwise coherent crystal. Shrinkage defects have been observed in minera l films deposited by the PILP process, such as biaxial strains1. Balz et al.10 ( Figure 2-15) also proposed a colloi d al aggregation mechanism for SrCO3 fiber formation. One difference between our mechanism and that of Balz is that we believe that the particles formed by the PILP process still have some liquid-like character, which allows them
110 to coalesce with the existing st ructure before solidifying and cr ystallizing. Another primary difference is that Balz et al.10 propose that PAA chains in solu tion aggregate into long strands that serve as a template for fiber growth. In contrast, we believe th at PAA functions as a process-directing agent, sequester ing ions in solution and stabilizi ng a fluidic liquid phase, which then forms fibers through preferential depositi on of droplets or nanopar ticles on the sticky ends that are initiated by high energy pr otrusions (fiber tips, surface ir regularities, etc). Although it is reasonable that the anionic polymer could assemb le nanoparticles into lin ear aggregates (as has been observed in Taos report11), in our system, the fibers are only found on the seed substrates. It is difficult to envision why the polymer templating effect would only occur on a seed substrate, therefore, our ne wly proposed mechanism accounts for this seemingly important contribution of the seed substrate. In addition, we believe the st icky end, or flux droplet, is needed to provide the very uniform diameter of these fibers, where more general aggregation mechanisms based on polarity or polymer adsorption would lead to a more statistical distribution of colloidal aggregates, and thus variable widths th roughout any given fiber. This would be particularly true in the mineral systems examin ed here, which do not form polar crystals. In addition, the adsorption behavior of polyaspartate and polyacrylate is generally considered to be non-specific, and the nanoparticles are not anisotropic in shape or chemistry, so the driving force for such pronounced anisotropic growth is diffic ult to rationalize as an oriented attachment phenomenon. We believe that many biominerals with comp lex morphologies are formed via a fluidic amorphous precursor pathway, which provides the coalescence needed to bu ild coherent, singlecrystalline structures. However, many of these biominerals have recently been found to have a nanogranular texture. For example, sea urchins, which have a single opti cal extinction direction
111 and spot diffraction pattern in XRD have a nanogranular texture32,52. As reported in Chapter 4, some SrCO3 fibers appear to behave optic ally as single crystals; Olszta et al. also showed calcite fibers that have a single extinction direction14. Optical studies of this type were difficult for the BaCO3 system due to the small size of the fibers, but these fibers have a single-crystalline diffraction pattern in SAED. When examined us ing high resolution TEM, however, these fibers were shown to have a nanogranular structure. Similarly, the single extinction di rection for SrCO3 did not translate into a single-crystalline SAED pattern, but rather an arc pattern was observed. The calcite fibers gr own here and previously by Olszta6 are too thick to do detailed SAED or high resolution TEM (for lattice imaging), although it is re asonable to expect that these fibers may also have a nanogra nular texture since the reaction conditions are so similar. HRTEM images of SrCO3 and the aragonitic CaCO3 fibers, shown in Figure 5-9, illustrate semicoherent dom ain texture of the CaCO3 aragonite and SrCO3 fibers. In both cases, although there are regions of disorder, the latt ice fringes appear to traverse nearly the entire region viewed. Recently, Tao et al.11 have produced rods of hydroxyapatite (HAP) with a mosaic structure (i.e. common crystallographic orientation, but with areas of diso rdered structures) which appear to form via a mesocrystal-type assembly proces s, which they suggest may be aided by a fluidic amorphous precursor ( Figure 2-16). In this system, ~5 nm di ameter nanoparticles were formed in the presence of PAA that were surrounded by a thin layer of amorpho us calcium phosphate (ACP), which were observed to aggregate into larger, unstable colloidal clusters, ~30 nm in diameter, over time. In the presence of Gly, th ese nanoparticles reorganized into linear chains, with randomly oriented, crystall ine HAP particles linked together by ACP. The ACP between the nanoparticles then crystalliz ed, and, according to these aut hors, the nanoparticles jiggled around in the amorphous phase to align themselv es in a common direction. The resultant
112 product was a nanorod made of single-crystalline HAP, with a thin coating of ACP. These authors do not address the reason why the na noparticles, surrounded by amorphous material, align to form a linear structure. It is inferred that somehow the interactions with Gly in solution would induce polarity in the syst em and cause alignment, sim ilar to mesocrystal assembly, except the particles are randomly oriented during initial assembly, and the surrounding amorphous material is isotropic. The coalescence of the amorphous phase into a single fiber, however, is similar to the mechanism proposed here for mineral fiber formation. In our system, small solidified domains, or even nanocrystal nuclei, may be present in the PILP droplets as they coalesce to form the fibers, which may explai n some of the nanodomain texture: as the amorphous phase crystallizes via isoepitaxy, sm all crystalline regions within the larger amorphous area would become trapped, leading to islands of randomly oriented material in an otherwise coherent lattice. Similar semi-coherent lattice stru ctures have been observed in numerous studies on sea urchin spicules32 and adult skeleton52, and were also observed by Sethmann et al.33 for calcite growth from a gelatinous coating. As Tao et al.11 have postulated, a small amount of motion may be available to these small nuclei, allowing them to realign somewhat with the surrounding cr ystal. Alternatively, the highly hydrated amorphous phase may allow for dissolution and/or recrystallization of the particles, such that only a very small amount of misaligned domains remain in the final, crystalline fiber. Th e mechanism of mineral fiber formation proposed here explains many of the phenomena observed in the BaCO3, CaCO3 and SrCO3 fibers, including the various surface textures spot or arc SAED patterns, shifts in crystallographic orientation along th e lengths of the fibers and na nogranular texture in HRTEM. Conclusions The formation mechanism of BaCO3, CaCO3 and SrCO3 mineral fibers formed via the PILP process was studied via SEM, TEM, AFM and confocal microscopy. Aragonitic CaCO3
113 fibers were synthesized for the first time. SEM analysis revealed that the fibers do nucleate directly off of calcite seed substrates, but rather nucleate off of films or aggregates (of the same material as the fibers) made in the early stag es of the PILP process, and therefore do not necessarily nucleate with an epit axial (or pseudo-epitaxial) relationship with the supplied seed substrate. AFM analysis revealed that the BaCO3 fibers have a relatively smooth surface compared to CaCO3 (aragonite), which was platy, and SrCO3, which was colloidal. These structures were confirmed via HRTEM. TEM studies also revealed that these fibers can have different types of crystallographic orientation shifts, where there is a shift in orientation along the length of the straight BaCO3 fibers, that the orientat ion of the aragonitic CaCO3 fibers follows the direction of the fiber, and that the orientation of the SrCO3 fibers, while exhibiting small shifts, a relatively uniform orientation persists across bends in the fiber. Although a few fibers were examined for each, a more thorough statistical analysis would be required to see if these were consistent trends for the different minera l types. A fluorescence study was also conducted, where it was observed that PAA was occluded within the fibers. Thes e results supported a colloidal aggregation mechanism. With this new information, a new mechanism for fiber formation was developed that, similar to the SPS mechanism, envisions an auto-catalytic functionality at the fiber growth front.
114 Table 5-1. Experimental c onditions for fiber synthesis. Concentration Concentration PAA Concentration main component poly(acr ylic acid)molecular weightcationic additive End product [Ba2+, Ca2+ or Sr2+] [PAA](Mw) [Sr2+] BaCO3 fibers 6 mM 50 g/ml 8000 g/mol0 mM BaCO3 fluorescent fibers 6 mM 100 g/ml 8000 g/mol0 mM CaCO3 aragonite fibers 12 mM 50 g/ml 8000 g/mol0.6 mM CaCO3 calcite fibers 12 mM 50 g/ml 5100 g/mol0 mM CaCO3 fluorescent calcite fibers 12 mM 50 g/ml 5100 g/mol0.6 mM SrCO3 fibers 6 mM 100 g/ml 8000 g/mol0 mM SrCO3 fluorescent fibers 6 mM 150 g/ml 8000 g/mol0 mM 1 m 1 m A C B Figure 5-1. Fiber morphology from previous experiments. A) BaCO3 fibers (as described in Chapters 3 & 4) B) Calcite fibers, reproduced with permissi on: Olszta et al.6 2004 American Chemical Society. C) SrCO3 fibers (as described in Chapter 4)
115 Figure 5-2. CaCO3 fibers grown via the PILP process at room temperature for 7 days. (A-C) Aragonitic fibers. (D-F) Calcitic fi bers. Crystallization conditions in Table 5-1. 100 nm 100 nm 1 mBC A Figure 5-3. Fibers grown on freshl y cleaved geologic calcite. A) BaCO3 fibers grown over 2 days reaction time at room temperature. B) CaCO3 fibers grown over 7 days reaction at RT. C) SrCO3 fibers grown over 2 days reacti on time at room temperature. Crystallization concentrations in Table 5-1.
116 Figure 5-4. Evidence of BaCO3 fibers growing off of a film of BaCO3 that coats a calcite rhomb. The film as delaminated from the underlying calcite rhomb due to dehydration and thermal energy due to the electron beam in SEM.
117 Figure 5-5. AFM height images of individual fi bers of A) BaCO3 B) CaCO3 and C) SrCO3.    200nm  2mABC500 nm      200nm   200nm 200nm  2m    2mABC500 nm Figure 5-6. TEM micrographs of BaCO3, CaCO3 and SrCO3 fibers, demonstrating the changes in orientation along the length of mineral fibers. A) BaCO3 fiber shows slight shift of  along length (orange arrows). White line added for visual reference. B) Aragonitic CaCO3 fiber where  is aligned roughly parallel to the growth direction of the fiber al ong its length. C) SrCO3 fiber which shows a semi-consistent orientation of  direction (orange arrows) along the entire length of the fiber, including across a ~115 bend. Parallel black lines added for visual reference.
118 10 m 20 m 15 m 100 m 100 m Control Tagged Optical Figure 5-7. Confocal microscopy images of mineral fibers grown with PAA tagged with 5BMF. A) BaCO3 B) CaCO3 C) SrCO3 D) Control experiment. Tagged: SrCO3 grown with BMF tagged PAA. Control: Fibers grown with untagged PAA, then soaked in solution containing BMF-tagged PAA fo r 1.5 hours to observe amount of fluorescence from surface adsorption. Optical: Optical image of same samples shown in Control illustrating that fibers were present over the entire sample, even though they are not visible in the Control image. Arrows correspond to similar structures in both images (spherulites and hor setails), which had the highest levels of tagged-PAA absorption.
119 Ca2+Ca2+Ca2+CO3 2-CO3 2-CO3 2Ca2+Ca2+Ca2+CO3 2-CO3 2-CO3 2-Ca2+Ca2+CO3 2-CO3 2Glass coverslip Time Time Stage 1: Droplets of amorphous, fluidic PILP phase form in solution and deposit on the bottom of the petri dish process, either on existing seed substrates or on the glass coverslip Stage 2: More PILP droplets in solution deposit preferentially on the existing structures due to higher surface energy and charge matching Stage 3: PILP droplets continue to deposit preferentially on the existing structures, especially on anisotropic structures due to increased polarity and surface energy. Previously amor phous (light grey) droplets coalesce and form fibrous structures, or we t onto the underlying substrate, and solidify and crystallize (black). Time Figure 5-8. Proposed mechanis m of mineral fiber formation.
120 5 nm 5 nm Figure 5-9. High resolution TEM images showing and aragonite CaCO3 fiber and a SrCO3 fiber. Arrow indicates overall fiber di rection. A) Aragonitic CaCO3 fiber with a platy structure. Lattice fringes persist from bottom right up through the majority of the thickness shown in this image. Another area in the upper le ft shows fringes that are slightly misoriented from the others. B) SrCO3 fiber. Lattice fringes are consistent through the middle area of the fiber, with few disruptions This structure breaks down along the edge, where th ere is less continuity.
121 CHAPTER 6 CONCLUSIONS AND FUTURE WORK Conclusions The addition of poly(acrylic ac id) as a water-soluble, anionic process-directing agent was implemented in the formation of various carbonat e salts via an amorphous liquid-phase precursor process. Non-equilibrium morphologies, such as barium and strontium carbonate films, horsetails and spherulites, a nd nanofibers of barium, calcium and strontium carbonate, were produced using this technique. In our first study, detailed in Chapter 3, th e PILP process was used to synthesize BaCO3 fibers on calcite seed crystals. Originally, the SPS method of generating mineral fibers was confined to the CaCO3 system6, but was then established as viable for BaCO3. This process is especially interesting in that the BaCO3 fibers are different from the calcite seeds in both composition and structure, which suggested the possi bility of using this non-specific process for fiber formation in other inorganic systems. We also showed that films can be used as seed substrates for fiber formation, and the presum ed high defect density of the PILP-formed CaCO3 films yielded a dense fibrous mesh of BaCO3. This correlates with our previous findings6 that suggest that crystal defects stimulate fiber nucleation. The fibers, which appear single crystalline when examined optically and by diffraction, show a nanodomain lattice texture, even though they likely crystallize from a singular nucleation event; therefore, we consider the fibers to be single crystals with mosaic text ure. In addition, the nanodomain texture may be related to the nanoclusters that have been obser ved in a variety of PILP formed crystals, and correlates well with the textures observed in biol ogically formed minerals, such as sea urchin spicules, spines and teeth13,21,34,35.
122 In our second study, detailed in Chapter 4, th e PILP process was explored further in the barium carbonate system, and introduced in th e strontium carbonate system. Evidence of an amorphous precursor phase was provided for bot h barium and strontium carbonate via a PLM study of PILP formation, wh ere non-birefringent (BaCO3) or weakly birefringent (SrCO3) films of apparently coalesced PILP particles were found in the early stages of reaction, while fully birefringent films were found af ter longer reaction times. The pr esence of early-stage (2.5 min reaction) PILP particles, whic h were shown to grow over time, was demonstrated via dynamic light scattering for BaCO3. Many of the same morphologies found previously in the CaCO3 system, including spherulites, fibers and films, were replicated in the BaCO3 and SrCO3 systems. The films in this study nucleated at the air-wate r interface, and were generally thicker, rougher and more granular than the calcite f ilms studied in our groups previous work1,3,39,107, and the SrCO3 films had a smaller grain size than the BaCO3 films. In addition, the BaCO3 fibers grown in this study were very similar to the fibe rs previously grown on calcite seed crystals5, while SrCO3 fibers, though of a similar size, were more serpentine and had a rough, colloidal surface texture and arc pattern in SAED. The more obvious mosaic or mesocrystal texture found in the SrCO3 fibers led us to reexamine the SPS mechanis m for fiber formation, previously proposed for the calcite fiber system6. This work is detailed in Chapter 5. The formation mechanism of BaCO3, CaCO3 and SrCO3 mineral fibers formed via the PILP process was studied in Chapter 5 us ing SEM, TEM, AFM and confocal microscopy. Fibers of BaCO3, calcite and SrCO3 were reproduced using the fluorophore-labeled polymer, and CaCO3 aragonite fibers were synthesi zed for the first time. SEM analysis revealed that the fibers do not nucleate epitaxially, or pseudo -epitaxially, in the case of BaCO3 and SrCO3, off of calcite seed substrates, but rather off of PILP films or aggregates of th e same material as the fibers,
123 which are made in the earl y stages of the reaction. AFM analysis revealed that the BaCO3 fibers have a relatively smooth surface compared to CaCO3 (aragonite), which was platy, and SrCO3, which was colloidal. These structures were confirmed via HRTEM. TEM studies also revealed that the crystallography within the fibrils is variable, where a shif t in crystallographic orientation was seen along the length of the straight BaCO3 fibers, while the orientation of the aragonitic CaCO3 fibers followed the direction of the fibe r, and that the orientation of the SrCO3 fibers, while shifting, persisted across bends in the fiber. A fluorescence study was also conducted to examine the mechanism, where it was observed that PAA became occluded within the fibers. These results supported a colloid al aggregation mechanism rather than the SPS mechanism. With this new information, a new mechanism fo r fiber formation was developed. In this mechanism, colloidal PILP droplets deposit onto a substrate, forming films and aggregates. As more droplets in solution form and fall to the bottom of the petri dish, they preferentially stick to existing PILP structures. They are especially attracted to bumps or irregularities in these structures. This leads to auto-catalyzing 1-dime nsional structures, which grow into high aspect ratio fibers. As new PILP dropl ets deposit onto the tips (the areas most likely to be amorphous and/or fluidic), they coalesce to form a si ngular fibrous structure. Solidification and crystallization likely initiates from a single nucleation event (o r relatively few), and proceeds along the length of the fiber, causing each fiber to have a single crystallographic orientation. Occluded polymer, shrinkage stress and the inclus ion of partially crysta lline nanoparticles may cause the nanodomain texture a nd deviations from perfect crystallographic alignment. Barium and strontium carbonate are both important materials, both in their own right and as precursors to functional materials such as barium or strontium titanate. Novel aqueous methods of producing unique mor phologies of barium and strontium carbonate, such as fibers
124 and films, are of great interest and importance. This work has demonstrated the robustness and non-specificity of the PILP pro cess, and its exciting applicati ons to new material systems. Future Work The PILP process has been successfully expand ed into non-calcium based systems. Other work in our group has focused on the calcium phos phate system. Another interesting route for future research would be to expand the PILP pro cess into functional materials. One interesting application would be to s ynthesize fibers of BaTiCO3 and SrTiCO3, which are both precursor materials to the piezoelectric materials BaTiO3 and SrTiO3. Additional light scattering experiments on both the BaCO3 and SrCO3 systems should also be completed. More characterization and optimization of the coated fibe rs detailed in Appendix A could also yield an interesting new technique for making 1D core -shell structures and hollow nanotubes. The expansion of the PILP process into functional ma terials could also yiel d exciting new materials for this application. To further characterize the mineral fibers, scanning near-field optical microscopy (SNOM) would be a wonderful technique. This characterization method, which allows the visualization of fluorescent species at nanometer resolution, would allow us to map the distribution of tagged process direction agents occluded within mineral fibers. This information could elucidate the formation mechanism of the mineral fibers. Cryo -TEM would also be a useful tool to examine early stage fiber formation. Using this technique, we could examine the fibers at various stages of formation, from the amorphous phase to fully crystalline; to better understand the formation and crystallization mechanisms. We would also be better able to examine the nanodomain structure of the fibers using cryo-TEM, as the cryo stage would limit beam damage to the fibers. Also, a statistical number of the various fiber types should be examined using TEM, to better understand the shifts in orientation al ong the length of the various fibers.
125 APPENDIX A STRONIUM CARBONATE COATINGS ON BARIUM CARBONAT E FIBERS Introduction Core-shell structures have recently received a great deal of interest in the research community. Typically, core-shell structures c onsist two different types of materials, for example, a metal nanoparticles coat ed with a functionalized polymer108,109, a micelle coated by a solid mineral110, block copolymer micelles111 or nanoparticles coated by SiO2 112. These structures have found applications in many fields, including drug delivery111,113, photonics114, biosensors108. One-dimensional core-shell structures are rarer, however. One example is the SiC/graphite and SiC/SiO2 core-shell fibers produced by Ye et al.115 which were grown using electrospun polyacrylonitrile fibers as templates, which were then heated to 1600C in the presence of SiO vapor, which results in a SiC core with a graphite shell. Etching can be done to remove the graphite, resulting in a SiC fiber with a thin coati ng of amorphous SiO2. In another example, Singh et al.116 produced SiO2 fibers coated in nanocrys talline diamond via microwave plasma enhanced chemical vapor deposition. These 1D core-shell stru ctures have potential applications as reinforcement for composites and in photonic devices. Hollow structured inorganic materials have recently attracted attention for their unique properties and potential applic ations, including drug-delivery113, photonics117, electronics110, sensors118 and catalysis119. Previously, template assisted synt hesis has been used to fabricate hollow structures. Template materials are often nanoparticles, and this process requires multiple steps, including template surface fictionalization and core-etching. Often, these methods result in spherical morphologies. Recently, Lu et al.120 explored making hollo w 1D structures of hematite, using a sacrificial template. However, they were able to make relatively low aspect ratio particles. Microcrystal line diamond fibers have been synthesized by depositing a diamond
126 coating on a tungsten core, follo wed by etching of the tungsten121. High aspect ratio hollow fibers have also been reported by Brei et al.122. These were piezoelectric ceramic nanoparticles embedded in a thermoplastic matrix fabr icated using an extrusion process. Materials and Methods Barium Carbonate Fiber Synthesis Crystallization was carried out using the am monium carbonate vapor diffusion method, as described previously2. Specifically, glass cover slips cl eaned using Nochromix or freshly cleaved geologic calcite (Ward Natural Scien ce, Rochester, NY) were placed in a 35 mm polystyrene petri dish (Falcon) containing 3 ml of crystallizati on solution or a 100 mm dish with 4 compartments, each containing 5 ml crystalliz ation solution, which was composed of 6 mM BaCl2 (Aldrich) and 50 g /ml poly(acrylic acid) (Aldrich), with MW of 8000 g/mol, in nanopure water (resistivity: 18.1 ). All reagents were used as received without further purification. The petri dishes were covered with parafilm, and placed in a chamber which contained freshly ground ammonium carbonate (S igma), also in a petri dish, covered by parafilm. Four needle holes were punched into the film covering the crys tallization solution, and eight in the ammonium carbonate covering. The reactions were run at room temperature (~ 25C) for three days, at which time the glass cove r slips were removed from solution, gently rinsed by dipping in water and ethanol to rem ove excess salt solution, and air dried. Strontium Carbonate Coating Cover slips containing BaCO3 fibers (synthesized as described previously) were placed into a 35 mm petri dish contai ning 3 ml solution of 6 mM SrCl2 100 g /ml PAA (Mw 8000 g/mol), and crystallization was carried out as described in the previous section.
127 Characterization The coated fibers morphology and size were examined using scanning electron microscopy (JEOL 6335f or JEOL 6400), tran smission electron microscopy (JEOL 200CX). Chemical composition was analyzed using energy -dispersive x-ray spectroscopy, and crystal structure using selected area electron diffraction (JEOL 200CX, JEOL 2010f). Scanning Electron Microscopy (SEM) Analysis: The samples were fixed to an aluminum stub using double-sided copper tape or colloidal graphite, and th en sputter coated with carbon. The samples were then examined with a JEOL 6335F FEGSEM or JEOL 6400 instrument at an accelerating voltage of 15 kV equipped with energy dispersive spectrometers (EDS). Transmission Electron Mi croscopy (TEM) Analysis: To examine the fibers, the samples were first scratched with a razor blade to dislodge some of the fibe rs from the cover slip or seed. A small aliquot of ethanol was then dispensed on to the scratched area and immediately drawn up using a micropipette. The removed aliquot wa s then dropped onto a 200 mesh copper TEM grid coated with lacy carbon. The sample was examined on a JEOL 200CX transmission electron microscope at 200 kV in bright field (BF) and selected area el ectron diffraction (SAED) modes. Results and Discussion Strontium carbonate coatings were successfully grown over a scaffold of barium carbonate fibers. The barium carbonate fi bers were similar in size, length and morphology to those reported in Chapter 3. SEM analysis was us ed to visualize the morphology of the coating ( Figure A-1). The lower magnification image re veals an incomplete coating growing over sm aller BaCO3 fibers. The BaCO3 fibers were on the order of 100 nm in diameter, and the coating had an outer diameter on the order of 300-500 nm. Higher magnification images of fibers growing at an angle to the substrate ( Figure A-1, right) reveal that the coating surrounds the f iber, and is not just a film that was deposited on top. This image also illustrates the colloidal
128 surface texture of the fiber/coating. We believe that the colloidal-looking surface texture on the thinner sections is caused by a thin film of SrCO3 deposited onto the surface of the BaCO3 fibers, which typically have a smooth surface, as described in Chapter 5. To determine the phases selected area electron diffraction was used. A single coated fiber was imaged using TEM ( Figure A-2), and the SAED pattern corresponding to the coated area was recorded (Figure A-2, inset). The circum ferential coating of SrCO3 over the BaCO3 was somewhat surprising. Our motivation to do this experiment, namely to see if fibers would continue to grow in fresh solution, was proven unattainable; however, this discovery was qui te serendipitous. We believe that, provided that the PILP phase of a given material will wet the surface of another fibrous material, the fabrication of co re-shell fibers of a variety of materials may be possible. Conclusions and Future Work BaCO3 (witherite) fibers were succe ssfully used as seeds for SrCO3 growth. SEM and EDS were done to illustrate that a thin SrCO3 film covered the entire fiber, with a thicker coating (100-150 nm thick) surroundi ng the individual BaCO3 fibers in some areas. TEM and SAED analysis proved that the core and coating phases were BaCO3 (witherite) and SrCO3 (strontianite). The BaCO3 core fiber displayed an apparen tly single crystalline spot pattern, while the SrCO3 coating was polycrystalline, but made up of well-aligned, as evinced by the arc pattern in SAED. These preliminary findings lead us to believe that this same type of process could be possible for any materi al that forms a PILP phase. Future work should include optimization of the conditions to get a uniform co ating over the fiber scaffolds. Other materials should also be explored both as the core fibers and coatings. Hollow-core mineral fibers would also be an exciting application, so dissolution studies for BaCO3, SrCO3 and CaCO3, as well as
129 the other, more soluble, fibers (i.e. electrospun polymer fibers, etc) as scaffolds should also be explored.
130 100 nm 100 nm Figure A-1. SEM images of BaCO3 fibers partially coated with SrCO3. Figure A-2. TEM analysis of coated fiber. SAED pattern (inset). Spots (labeled in orange) correspond to witherite; arcs (labeled in blue) correspond to strontianite.
131 LIST OF REFERENCES 1 Gower, L. A. and Tirrell, D. A. Calcium carbonate films and helices grown in solutions of poly(aspartate). Journal of Crystal Growth 191, 153-160 (1998). 2 Gower, L. B. and Odom, D. J. Depositi on of calcium carbonate films by a polymerinduced liquid-precursor (pilp) process. Journal of Crystal Growth 210, 719-734 (2000). 3 Kim, Y. Y., Douglas, E. P., and Gower, L. B. Patterning inorganic (caco3) thin films via a polymer-induced liquid-precursor process. Langmuir 23, 4862-4870 (2007). 4 Cheng, X. G. and Gower, L. B. Molding mineral within microporous hydrogels by a polymer-induced liquid-precursor (pilp) process. Biotechnology Progress 22, 141-149 (2006). 5 Homeijer, S. J., Olszta, M. J., Barrett, R. A. et al. Growth of nanofibrous barium carbonate on calcium carbonate seeds. Journal of Crystal Growth 310, 2938-2945 (2008). 6 Olszta, M. J., Gajjeraman, S., Kaufman, M. et al. Nanofibrous calcite synthesized via a solution-precursor-solid mechanism. Chemistry Of Materials 16, 2355-2362 (2004). 7 Colfen, H. and Antonietti, M. Mesocrysta ls: Inorganic superstructures made by highly parallel crystallization and controlled alignment. Angewandte Chemie-International Edition 44, 5576-5591 (2005). 8 Colfen, H. and Antonietti, M., Mesocrystals and nonclassical crystallization. (John Wiley & Sons Ltd, Chichester, UK, 2008). 9 Niederberger, M. and Colfen, H. Oriented attachment and mesocrystals: Non-classical crystallization mechanisms ba sed on nanoparticle assembly. Physical Chemistry Chemical Physics 8, 3271-3287 (2006). 10 Balz, M., Therese, H. A., Kappl, M. et al Morphosynthesis of st rontianite nanowires using polyacrylate templates tethered onto self-assembled monolayers. Langmuir 21, 3981-3986 (2005). 11 Tao, J. H., Pan, H. H., Zeng, Y. W. et al. Roles of amorphous calcium phosphate and biological additives in the assemb ly of hydroxyapatite nanoparticles. Journal of Physical Chemistry B 111, 13410-13418 (2007). 12 Weiner, S. and Wagner, H. D. The mate rial bone: Structure mechanical function relations. Annual Review of Materials Science 28, 271-298 (1998). 13 Ma, Y., Weiner, S., and Addadi, L. Mine ral deposition and crystal growth in the continuously forming teeth of sea urchins. Advanced Functional Materials 17, 2693-2700 (2007).
132 14 Olszta, M. J., Odom, D. J., Douglas, E. P. et al. A new paradigm for biomineral formation: Mineralization via an amorphous liquid-phase precursor. Connective Tissue Research 44, 326-334 (2003). 15 Sprengel, C., E. Huxleyi type a, Available at http://www.naturalhistorymuseum.org.uk/hos ted_sites/ina/CODEN ET/GuideIm ages/COr fen/source/113-19.htm (2008). 16 Li, X. D., Chang, W. C., Chao, Y. J. et al. Nanoscale structural and mechanical characterization of a natura l nanocomposite material: The shell of red abalone. Nano Letters 4, 613-617 (2004). 17 Mann, S., Webb, J., and Williams, R. J. P., Biomineralization: Chemial and biochemical perspectives. (VCH Publishers, New York, NY, 1989). 18 Lowenstam, H. A. and Weiner, S., On biomineralization. (Oxford University Press, Oxford, 1989). 19 Berman, A., Addadi, L., and Weiner, S. Interactions of sea-urchin skeleton macromolecules with growing calcite crysta ls a study of intracrystalline proteins. Nature 331, 546-548 (1988). 20 Mann, S. Molecular recognition in biomineralization. Nature 332, 119-124 (1988). 21 Beniash, E., Aizenberg, J., Addadi, L. et al. Amorphous calcium carbonate transforms into calcite during sea urchin larval spicule growth. Proceedings of the Royal Society of London Series B-Biological Sciences 264, 461-465 (1997). 22 Weiss, I. M., Tuross, N., Addadi, L. et al. Mollusc larval shel l formation: Amorphous calcium carbonate is a precursor phase for aragonite. Journal of Experimental Zoology 293, 478-491 (2002). 23 Weiner, S. Transient precursor strategy in mineral formation of bone. Bone 39, 431-433 (2006). 24 Addadi, L., Raz, S., and Weiner, S. Taki ng advantage of disord er: Amorphous calcium carbonate and its roles in biomineralization. Advanced Materials 15, 959-970 (2003). 25 Navrotsky, A. Energetic clues to pathways to biomineralization: Pr ecursors, clusters, and nanoparticles. Proceedings of the National Academy of Sciences of the United States of America 101, 12096-12101 (2004). 26 Bonucci, E. ed., Calcification in biological systems. (CRC Press, Boca Raton, 1992).
133 27 Cui, F. Z. and Ge, J. New observations of the hierarchical structure of human enamel, from nanoscale to microscale. Journal of Tissue Engineer ing and Regenerative Medicine 1, 185-191 (2007). 28 Nakahara, H. and kakei, M., in Origin, evolution, and modern aspects of biomineralization in plants and animals, edited by Rex E. Crick (Springer, New York, 1990). 29 Robinson, C., Connell, S., Kirkham, J. et al. Dental enamel a biological ceramic: Regular substructures in enamel hydroxyapat ite crystals revealed by atomic force microscopy. Journal of Materials Chemistry 14, 2242-2248 (2004). 30 Wilt, F. H. Biomineralization of the spicules of sea urchin embryos. Zoological Science 19, 253-261 (2002). 31 Berman, A., Addadi, L., Kvick, A. et al. Inte rcalation of sea-urchin proteins in calcite study of a crystalline composite-material. Science 250, 664-667 (1990). 32 Sethmann, I., Hinrichs, R., Worheide, G. et al. Nano-cluster co mposite structure of calcitic sponge spicules a case study of basic characteristics of biominerals. Journal Of Inorganic Biochemistry 100, 88-96 (2006). 33 Sethmann, I., Putnis, A., Grassmann, O. et al. Observation of nano-clustered calcite growth via a transient phase mediated by organic polyanions: A close match for biomineralization. American Mineralogist 90, 1213-1217 (2005). 34 Beniash, E., Addadi, L., and Weiner, S. Ce llular control over spicule formation in sea urchin embryos: A structural approach. Journal of Structural Biology 125, 50-62 (1999). 35 Politi, Y., Arad, T., Klein, E. et al. Sea urchin spine calcite forms via a transient amorphous calcium carbonate phase. Science 306, 1161-1164 (2004). 36 Wang, R. Z., Addadi, L., and Weiner, S. Design strategies of sea urch in teeth: Structure, composition and micromechanical relations to function. Philosophical Transactions of the Royal Society of London Se ries B-Biological Sciences 352, 469-480 (1997). 37 Chen, C. P. and Lawrence, J. M. The ultr astructure of the plumula of the tooth of lytechinus-variegatus (ech inodermata, echinoidea). Acta Zoologica 67, 33-41 (1986). 38 Lam, R. S. K., Charnock, J. M., Lennie, A. et al. Synthesisdependant structural variations in amorphous calcium carbonate. Crystengcomm 9, 1226-1236 (2007). 39 Amos, F. F., Sharbaugh, D. M., Talham, D. R. et al., (2007), Vol. 23, pp. 1988-1994.
134 40 Olszta, M. J., Douglas, E. P., and Gower, L. B. Scanning electron microscopic analysis of the mineralization of type i collagen vi a a polymer-induced liquid-precursor (pilp) process. Calcified Tissue International 72, 583-591 (2003). 41 Aizenberg, J., Black, A. J., and Whitesides, G. H. Oriented growth of calcite controlled by self-assembled monolayers of functiona lized alkanethiols supported on gold and silver. Journal of the American Chemical Society 121, 4500-4509 (1999); Kuther, J., Nelles, G., Seshadri, R. et al. Templated crystallisation of calcium and strontium carbonates on centred rectangular self-assembled monolayer substrates. Chemistry-a European Journal 4, 1834-1842 (1998). 42 Aizenberg, J., Black, A. J., and Whitesides, G. M. Control of crystal nucleation by patterned self-assembled monolayers. Nature 398, 495-498 (1999). 43 DiMasi, E., Patel, V. M., Sivakumar, M. et al. Polymer-controlled growth rate of an amorphous mineral film nucleated at a fatty acid monolayer. Langmuir 18, 8902-8909 (2002). 44 Wang, T. X., Reinecke, A., and Colfen, H. In situ investigation of complex baso4 fiber generation in the presence of sodium polyacrylate. 2. Crystallization mechanisms. Langmuir 22, 8986-8994 (2006). 45 Yu, S. H., Colfen, H., Tauer, K. et al. T ectonic arrangement of b aco3 nanocrystals into helices induced by a racemic block copolymer. Nature Materials 4, 51-U55 (2005). 46 Gehrke, N., Nassif, N., Pinna, N. et al. Re trosynthesis of nacre via amorphous precursor particles. Chemistry Of Materials 17, 6514-6516 (2005); Aizenberg, J., Lambert, G., Addadi, L. et al. Stabilization of amorphous calcium carbonate by specialized macromolecules in biological and synthetic precipitates. Advanced Materials 8, 222-& (1996). 47 Foo, C. W. P., Huang, J., and Kaplan, D. L. Lessons from seashells: Silica mineralization via protein templating. Trends In Biotechnology 22, 577-585 (2004); Glawe, D. D., Rodriguez, F., Stone, M. O. et al. Polypeptid e-mediated silica growth on indium tin oxide surfaces. Langmuir 21, 717-720 (2005); Naik, R. R., Brott, L. L., Clarson, S. J. et al. Silica-precipitating peptides isolated from a combinatorial phage display peptide library. Journal Of Nanoscience And Nanotechnology 2, 95-100 (2002). 48 Bittarello, E. and Aquilano, D. Self-asse mbled nanocrystals of barium carbonate in biomineral-like structures. European Journal of Mineralogy 19, 345-351 (2007). 49 Guo, X. H. and Yu, S. H. Controlled mine ralization of barium car bonate mesocrystals in a mixed solvent and at the air/solutio n interface using a double hydrophilic block copolymer as a crystal modifier. Crystal Growth & Design 7, 354-359 (2007).
135 50 Wang, T. X., Xu, A. W., and Colfen, H. Fo rmation of self-organized dynamic structure patterns of barium carbonate crystals in polymer-controlled crystallization. Angewandte Chemie-International Edition 45, 4451-4455 (2006). 51 Blake, D. F., Peacor, D. R., and Allard, L. F. Ultrastructural and microanalytical results from echinoderm calcite implications for bi omineralization and diagenesis of skeletal material. Micron and Microscopica Acta 15, 85-90 (1984). 52 Tsipursky, S. J. and Buseck, P. R. Struct ure of magnesian calcite from sea-urchins. American Mineralogist 78, 775-781 (1993). 53 Oaki, Y., Kotachi, A., Miura, T. et al. Bridged nanocrystals in biominerals and their biomimetics: Classical yet modern crystal growth on the nanoscale. Advanced Functional Materials 16, 1633-1639 (2006). 54 Wagner, R. S. and Ellis, W. C. Vapor-liquid-solid mechanism of single crystal growth (new method growth catalysis from impurity whisker epitaxial + large crystals si e). Applied Physics Letters 4, 89-& (1964). 55 Givargizov, E. I. Fundamental aspects of vls growth. Journal of Crystal Growth 31, 2030 (1975). 56 Tan, T. Y., Li, N., and Gosele, U. Is there a thermodynamic size limit of nanowires grown by the vapor-liquid-solid process? Applied Physics Letters 83, 1199-1201 (2003). 57 Morales, A. M. and Lieber, C. M. A laser ab lation method for the synthesis of crystalline semiconductor nanowires. Science 279, 208-211 (1998). 58 Gudiksen, M. S., Wang, J. F., and Lieiber, C. M. Synthetic contro l of the diameter and length of single crystal semiconductor nanowires. Journal of Physical Chemistry B 105, 4062-4064 (2001). 59 Trentler, T. J., Hickman, K. M., Goel, S. C. et al. Solution-liq uid-solid growth of crystalline iii-v semiconductors an analogy to vapor-liquid-solid growth. Science 270, 1791-1794 (1995). 60 Rao, C. N. R., Deepak, F. L., G undiah, G. et al. Inorganic nanowires. Progress in Solid State Chemistry 31, 5-147 (2003). 61 Cao, G. Z. and Liu, D. W. Template-based synthesis of nanorod, nanowire, and nanotube arrays. Advances in Colloid and Interface Science 136, 45-64 (2008). 62 Penn, R. L. and Banfield, J. F. Mor phology development and crystal growth in nanocrystalline aggregates under hydrothermal conditions: Insights from titania. Geochimica Et Cosmochimica Acta 63, 1549-1557 (1999).
136 63 Cao, M. H., Wang, Y. H., Guo, C. X. et al. Preparation of ultrahigh-aspect-ratio hydroxyapatite nanofibers in reverse micelles under hydrothermal conditions. Langmuir 20, 4784-4786 (2004). 64 Cao, M. H., Wu, X. L., He, X. Y. et al. Microemulsion-mediated solvothermal synthesis of srco3 nanostructures. Langmuir 21, 6093-6096 (2005). 65 Kuang, D. B., Xu, A. W., Fang, Y. P. et al. Preparation of inorganic salts (caco3, baco3, caso4) nanowires in the triton x-100/c yclohexane/water reverse micelles. Journal Of Crystal Growth 244, 379-383 (2002). 66 Yu, S. H., Antonietti, M., Colfen, H. et al Growth and self-assembly of bacro4 and baso4 nanofibers toward hierarchical and repeti tive superstructures by polymer-controlled mineralization reactions. Nano Letters 3, 379-382 (2003). 67 Yu, S. H., Colfen, H., and Antonietti, M. The combination of colloid-controlled heterogeneous nucleation and pol ymer-controlled crystalliza tion: Facile synthesis of separated, uniform high-aspect-ratio single-crystalline bacro4 nanofibers. Advanced Materials 15, 133-+ (2003). 68 Qi, L. M., Colfen, H., Antonietti, M. et al. Formation of baso4 fibres with morphological complexity in aqueous polymer solutions. Chemistry-A European Journal 7, 3526-3532 (2001). 69 Macketta, J. J., Encyclopedia of chemical processing and design. (Marcel Dekker, New York, 1976). 70 Templeton, L. K. and Pask, J. A. Formation of batio3 from baco3 and tio2 in air and in co2. Journal of the American Ceramic Society 42, 212-216 (1959); Allen, B. F. F., N. M.; Lin, S. C.; Semait, R.; Luss, D.; Richardson, J. T.;, A continuous coprecipitation process for the production of 1-2-3 precu rsors. presented at the American Institute of Chemical Engineers National Meeting, 1992 (unpublished) ; Buscaglia, M. T., Buscaglia, V., and Alessio, R. Coating of baco3 crystals with tio2: Versatile approach to the synthesis of batio3 tetragonal nanoparticles. Chemistry of Materials 19, 711-718 (2007). 71 Chichagov, A. V., (Russian F oundation of Basic Research, 1997). 72 Chen, P. C., Cheng, G. Y., Kou, M. H. et al. Nucleation and morphology of barium carbonate crystals in a semi-batch crystallizer. Journal of Crystal Growth 226, 458-472 (2001). 73 Qi, L. M., Ma, J. M., Cheng, H. M. et al. Reverse micelle based formation of baco3 nanowires. Journal Of Physical Chemistry B 101, 3460-3463 (1997).
137 74 Lv, S., Li, P., Sheng, H. et al. Synthesis of single-crystalline baco3 nanostructures with different morphologies via a simple pvp-assisted method. Materials Letters 61, 42504254 (2007). 75 Ma, M. G., Zhu, Y. J., Zhu, J. F. et al. A simple route to synthesis of baco3 nanostructures in water/ethylene glycol mixed solvents. Materials Letters 61, 5133-5136 (2007). 76 Massone, J., Technology and uses of barium and strontium compounds. presented at the 5th Industrian Minerals Inte rnational Congress: Metals Bulletin, Madrid, Spain, 1982 (unpublished); Roskill; Berbenni, V., Marini, A., and Bruni, G. Effect of mechanical activation on the preparation of srtio3 and sr2tio4 ceramics from the solid state system srco3-tio2. Journal of Alloys and Compounds 329, 230-238 (2001). 77 Shi, J. J., Li, J. J., Zhu, Y. F. et al. Nanosized srco3-based chemiluminescence sensor for ethanol. Analytica Chimica Acta 466, 69-78 (2002). 78 Pan, Y. X., Wu, M. M., and Su, Q. Synt hesis of eu3+-doped calcium and strontium carbonate phosphors at room temperature. Materials Research Bulletin 38, 1537-1544 (2003). 79 Rautaray, D., Sainkar, S. R., and Sastry, M. Srco3 crystals of ribbonlike morphology grown within thermally evaporated sodium bi s-2-ethylhexylsulfosuccinate thin films. Langmuir 19, 888-892 (2003). 80 Guo, G. S., Gu, F. B., Wang, Z. H. et al. Low-temperature growth of single-crystal srco3 nanoneedles. Chinese Chemical Letters 16, 1101-1104 (2005). 81 Huang, Q., Gao, L., Cai, Y. et al. Synthe sis and characterization of strontium carbonate nanowires with a axis orientat ion and dendritic nanocrystals. Chemistry Letters 33, 290291 (2004). 82 Shi, L. G. and Du, F. L. Solvothermal synthesis of srco3 he xahedral ellipsoids. Materials Letters 61, 3262-3264 (2007). 83 Gu, F., Wang, Z. H., Han, D. M. et al Crystallization of rare earth carbonate nanostructures in the reverse micelle system. Crystal Growth & Design 7, 1452-1458 (2007). 84 Li, S. Z., Zhang, H., Xu, J. et al. H ydrothermal synthesis of flower-like srco3 nanostructures. Materials Letters 59, 420-422 (2005). 85 Kuther, J., Bartz, M., Seshadri, R. et al Crystallization of sr co3 on a self-assembled monolayer substrate: An in-s itu synchrotron x-ray study. Journal of Materials Chemistry 11, 503-506 (2001).
138 86 Reddy, S., Rautaray, D., Sainkar, S. R. et al. On the morphology of srco3 crystals grown at the interface between two immiscible liquids. Bulletin of Materials Science 26, 283288 (2003). 87 Chave, K. E., Smith, S. V., and Roy, K. J. Carbonate product ion by coral reefs. Marine Geology 12, 123-& (1972). 88 Greegor, R. B., Pingitore, N. E., and Lytle, F. W. Strontianite in cora l skeletal aragonite. Science 275, 1452-1454 (1997). 89 Xu, A. W., Antonietti, M., Yu, S. H. et al. Polymer-mediated mineralization and selfsimilar mesoscale-organized calcium car bonate with unusual superstructures. Advanced Materials 20, 1333-+ (2008). 90 Mao, C. B., Flynn, C. E., Hayhurst, A. et al. Viral assembly of oriented quantum dot nanowires. Proceedings Of The National Academy Of Sciences Of The United States Of America 100, 6946-6951 (2003). 91 Gao, Y. H., Niu, H. L., and Chen, Q. W. Preparation and characterization of baco3 single-crystalline nanowires by a solvothermal process. Chinese Journal Of Inorganic Chemistry 19, 37-40 (2003). 92 Sadasivan, S., Khushalani, D., and Ma nn, S. Synthesis of calcium phosphate nanofilaments in reverse micelles. Chemistry Of Materials 17, 2765-2770 (2005). 93 Nissen, H. U. Crystal orientation and pl ate structure in echinoid skeletal units. Science 166, 1150-& (1969); Donnay, G. and Pawson, D. L. X-ray diffraction studies of echinoderm plates. Science 166, 1147-& (1969). 94 Aizenberg, J., Weiner, S., and Addadi, L. Coexistence of amorphous and crystalline calcium carbonate in skeletal tissues. Connective Tissue Research 44, 20-25 (2003). 95 Weiner, S., Levi-Kalisman, Y., Raz, S. et al. Biologically formed amorphous calcium carbonate. Connective Tissue Research 44, 214-218 (2003). 96 Olszta, M. J., Cheng, X. G., Jee, S. S. et al. Bone structure and formation: A new perspective. Materials Science & Engineering R-Reports 58, 77-116 (2007). 97 Colfen, H. and Mann, S. Higher-order or ganization by mesoscale self-assembly and transformation of hybrid nanostructures. Angewandte Chemie-International Edition 42, 2350-2365 (2003); Gao, Y. X., Yu, S. H., Cong, H. P. et al. Block-copolymer-controlled growth of caco3 microrings. Journal Of Physical Chemistry B 110, 6432-6436 (2006). 98 Sethmann, I., Helbig, U., and Worheide, G. Octocoral sclerite ultrastructures and experimental approach to underlyi ng biomineralisation principles. Crystengcomm 9, 1262-1268 (2007).
139 99 Ma, Y., Colfen, H., and Antonietti, M. Mo rphosynthesis of alanine mesocrystals by ph control. J. Phys. Chem. B. 110, 10822-10828 (2006). 100 Dai, L., Douglas, E. P., and Gower, L. B. Compositional analysis of a polymer-induced liquid-precursor (pilp) amorphous caco3 phase. Journal of Non-Crystalline Solids 354, 1845-1854 (2008). 101 Cheng, X., Relevance of the polymer-induc ed liquid-precursor (pilp) process to biomineralization and development of bi omimetic materials Ph.D. Dissertation, University of Florida, 2005. 102 Xu, X. R., Cai, A. H., Liu, R. et al. Am orphous calcium carbonate in biomineralization. Progress in Chemistry 20, 54-59 (2008). 103 Yue, W. B., Park, R. J., Kulak, A. N. et al. Macroporous i norganic solids from a biomineral template. Journal of Crystal Growth 294, 69-77 (2006). 104 Bevan, D. J. M., Rossmaith, E., Mylrea, D. K. et al. On the structure of aragonite lawrence bragg revisited. Acta Crystallogr. Sect. B-Struct. Sci. 58, 448-456 (2002). 105 Fleming, S. D., Parkinson, G. M., and Rohl A. L. Predicting the occurrence of reflection twins. Journal of Crystal Growth 178, 402-409 (1997). 106 Kim, Y.-y. and Gower, L. B., Formation of complex non-equilibr ium morphologies of calcite via biomimetic processing. presented at the Materials Research Society, Boston, 2003 (unpublished). 107 Cheng, X. G., Varona, P. L., Olszta, M. J. et al. Biomimetic synthe sis of calcite films by a polymer-induced liquid-precurs or (pilp) process 1. Influe nce and incorporation of magnesium. Journal of Crystal Growth 307, 395-404 (2007). 108 Medintz, I. L., Uyeda, H. T., Goldman, E. R. et al. Quantum dot bioconjugates for imaging, labelling and sensing. Nature Materials 4, 435-446 (2005). 109 Nuzzo, R. G. and Allara, D. L. Adsorpti on of bifunctional orga nic disulfides on gold surfaces. Journal of the Amer ican Chemical Society 105, 4481-4483 (1983). 110 Zhu, Y. F., Shi, J. L., Shen, W. H. et al. Stimuli-responsive contro lled drug release from a hollow mesoporous silica sphere/polyelect rolyte multilayer core-shell structure. Angewandte Chemie-International Edition 44, 5083-5087 (2005). 111 Kataoka, K., Harada, A., and Nagasaki, Y. Block copolymer micelles for drug delivery: Design, characterization and biological si gnificance. Advanced Drug Delivery Reviews 47, 113-131 (2001).
140 112 Mulvaney, P., Liz-Marzan, L. M., Giersig, M. et al. Silica encapsu lation of quantum dots and metal clusters. Journal of Materials Chemistry 10, 1259-1270 (2000). 113 Pekarek, K. J., Jacob, J. S., and Mathiow itz, E. Double-walled pol ymer microspheres for controlled drug-release. Nature 367, 258-260 (1994). 114 Lopez, C. Materials aspects of photonic crystals. Advanced Materials 15, 1679-1704 (2003). 115 Ye, H. H., Titchenal, N., Gogotsi, Y. et al. Sic nanowires synthesized from electrospun nanofiber templates. Advanced Materials 17, 1531-1535 (2005). 116 Singh, M. K., Titus, E., Madaleno, J. C. et al. Novel two-step me thod for synthesis of high-density nanocrystalline diamond fibers. Chemistry of Materials 20, 1725-1732 (2008). 117 Liang, H. P., Zhang, H. M., Hu, J. S. et al. Pt hollow nanospheres: Facile synthesis and enhanced electrocatalysts. Angewandte Chemie-International Edition 43, 1540-1543 (2004). 118 Chen, J. Y., Wiley, B., Li, Z. Y. et al. Gold nanocages: Engineering their structure for biomedical applications. Advanced Materials 17, 2255-2261 (2005). 119 Jeong, U. Y. and Xia, Y. N. Photonic cr ystals with thermally switchable stop bands fabricated from se@ag2se spherical colloids. Angewandte Chemie-In ternational Edition 44, 3099-3103 (2005). 120 Lu, J., Chen, D. R., and Jiao, X. L. Fa brication, characterization, and formation mechanism of hollow spindle-like hematite via a solvothermal process. Journal of Colloid and Interface Science 303, 437-443 (2006). 121 Baranauskas, V., Ceragioli, H. J., Peterlevitz, A. C. et al. Development of tubes of microcrystalline diamond and diamond-like carbon. Thin Solid Films 398, 250-254 (2001). 122 Brei, D. and Cannon, B. J. Piezoce ramic hollow fiber active composites. Compos. Sci. Technol. 64, 245-261 (2004).
141 BIOGRAPHICAL SKETCH Sara grew up in Port Deposit, Maryland, a sm all town at the head of the Chesapeake Bay, with her parents, Glenn and Em ily Jensen, both teachers; and brother, Kevin, a law student. After graduating from Perryville High School in 1999, she attended Lehigh University in Bethehem, Pennsylvania. While at Lehigh, she was a member of the varsity field hockey team and womens rugby team. After obtaining a degree in materials science and engineering in 2003, she was accepted into the materials science and e ngineering graduate program at the University of Florida. She has worked with Dr. Laurie Gower on biomimetic mineralization for the past five years. Upon graduation, Sara will begin wo rk as a process engineer in the Imaging and Printing division of Hewlett P ackard in Corvallis, Oregon.